Landscapes and Landforms of Ethiopia: Paolo Billi Editor

World Geomorphological Landscapes
Paolo Billi Editor
Landscapes and
Landforms of
Ethiopia
Series editor
Piotr Migon, Wrocław, Poland
For further volumes:

http://www.springer.com/series/10852
Paolo Billi
Editor
Landscapes and Landforms

of Ethiopia
123
Editor
Paolo Billi
Physics and Earth Sciences
University of Ferrara
Ferrara
Italy
ISSN 2213-2090 ISSN 2213-2104 (electronic)

ISBN 978-94-017-8025-4 ISBN 978-94-017-8026-1 (eBook)
DOI 10.1007/978-94-017-8026-1
Library of Congress Control Number: 2015931921
Springer Dordrecht Heidelberg New York London

© Springer Science+Business Media Dordrecht 2015
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material
is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting,
reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval,
electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not
imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and
regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed
to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty,
express or implied, with respect to the material contained herein or for any errors or omissions that may have been
made.
Printed on acid-free paper
Springer Science+Business Media B.V. Dordrecht is part of Springer Science+Business Media

(www.springer.com)
Series Editor Preface
Landforms and landscapes vary enormously across the Earth, from high mountains to endless
plains. At a smaller scale, nature often surprises us by creating shapes which look improbable.
Many physical landscapes are so immensely beautiful that they received the highest possible
recognition—they hold the status of World Heritage Sites. Apart from often being immensely
scenic, landscapes tell stories which not uncommonly can be traced back in time for tens of
millions of years and include unique geological events such as meteorite impacts. In addition,
many landscapes owe their appearance and harmony not solely to the natural forces. For
centuries, and even millennia, they have been shaped by humans who have modified hill-
slopes, river courses and coastlines, and erected structures which often blend with the natural
landforms to form inseparable entities.
These landscapes are studied by geomorphology—‘the science of scenery’—a part of Earth
Sciences that focuses on landforms, their assemblages, surface and subsurface processes that
moulded them in the past and that change them today. To show the importance of geomor-
phology in understanding the landscape, and to present the beauty and diversity of the geo-
morphological sceneries across the world, we have launched a book series World
Geomorphological Landscapes. It aims to be a scientific library of monographs that present
and explain physical landscapes, focusing on both representative and uniquely spectacular
examples. Each book will contain details on geomorphology of a particular country or a
geographically coherent region. This volume presents the impressive geomorphic legacy of
Ethiopia which hosts many landscapes and landforms of global significance. Examples include
the otherworldly Afar Depression with its sun-baked volcanoes, high-elevation basalt plateaus,
the East African Rift valley with its splendid lakes, to name just a few. Ethiopia is also a
country where geomorphology and people have remained in particularly close association
since time immemorial. These relationships can be examined by referring to the past—as
demonstrated by the chapter on geoarcheology of Aksum, but perhaps more importantly, with
the reference to the present-day environmental problems arising from land use, soil erosion,
water resources depletion and settlement growth. A number of chapters in this book remind us
that geomorphological landscapes are not only beautiful; they are also very fragile if used
improperly.
The World Geomorphological Landscapes series is produced under the scientific patronage
of the International Association of Geomorphologists (IAG)—a society that brings together
geomorphologists from all around the world. The IAG was established in 1989 and is an
independent scientific association affiliated with the International Geographical Union (IGU)
and the International Union of Geological Sciences (IUGS). Among its main aims are to
promote geomorphology and to foster dissemination of geomorphological knowledge.
I believe that this lavishly illustrated series, which keeps to the scientific rigour, is the most
appropriate means to fulfil these aims and to serve the geoscientific community. To this end,
my great thanks go to Prof. Paolo Billi, a person long involved in geomorphological research
in Ethiopia, for agreeing to coordinate this timely volume in the series. I am also very grateful
to all individual authors who accepted invitations to contribute and, often, delivered stories
which contained original, not yet published research.
v
vi Series Editor Preface
In contrast to many other countries, Ethiopia is far less known regarding its geomorphology
at the local scale and many of its regions are still terra incognita. Therefore, it was not feasible
to strictly follow the format of previous volumes in the series and offer a wide range of site-
specific stories. However, I am sure the readers will value more general presentations of the
geomorphic environment of Ethiopia which are not only beautifully illustrated, but also
provide an updated, unique source of reference.
For me, to write the preface to the Ethiopia volume is of particular pleasure. In 2008 I was
fortunate to join the IAG-organized field trip to the Ethiopian Highlands, expertly run by Paolo
Billi, Franco Dramis and Giandomenico Fubelli (all involved in this volume), and became
fascinated with the geomorphology of Ethiopia, which lasts until nowadays. As a little evi-
dence of this fascination serves my own modest contribution to this volume, regarding the
geomorphic scenery of Aksum. After touring the country for a week I thought I knew it
reasonably well. This book in its final shape has told me how much is left to be seen.
Piotr Migoń
Contents
Part I General Introduction to the Geomorphology of Ethiopia
1 Geomorphological Landscapes of Ethiopia . . . . . . . . . . . . . . . . . . . . . . . . . 3

Paolo Billi
2 Geology of Ethiopia: A Review and Geomorphological Perspectives . . . . . . 33

Ernesto Abbate, Piero Bruni, and Mario Sagri
3 The Climate of Ethiopia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

Massimiliano Fazzini, Carlo Bisci, and Paolo Billi
4 Ethiopian Rivers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
Paolo Billi, Semunesh Golla, and Dawit Tefferra
5 Planation Surfaces and the Long-term Geomorphological

Evolution of Ethiopia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
Mauro Coltorti, Dario Firuzabadi, Andrea Borri, Pierlorenzo Fantozzi,
and Pierluigi Pieruccini
Part II Local Studies
6 Paleoglaciated Landscapes in Simen and Other High-Mountain

Areas of Ethiopia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
Hans Hurni
7 Geomorphology of the Archaeological Area of Aksum . . . . . . . . . . . . . . . . 147

Giovanni Ferrari, Rossano Ciampalini, Paolo Billi, and Piotr Migon
8 Geomorphology of the Adwa District . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163

Maria J. Machado
9 The Amba Landscape of the Ethiopian Highlands, Shaped by Rockfall . . . . 179

J. Nyssen, J. Moeyersons, J. Deckers, Mitiku Haile, and J. Poesen
10 Gully Development in the Tigray Highlands. . . . . . . . . . . . . . . . . . . . . . . . 191

A. Frankl, J. Poesen, J. Moeyersons, and J. Nyssen
vii
viii Contents
11 Tufa Dams in Tigray (Northern Ethiopia) as Late Pleistocene—Holocene

Climate Proxies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
Francesco Dramis and Giandomenico Fubelli
12 Geomorphology of Ephemeral Streams in the Kobo Basin . . . . . . . . . . . . . 213

Paolo Billi
13 Sediment Yield Variability at Various Spatial Scales and Its

Hydrological and Geomorphological Impacts on Dam-catchments
in the Ethiopian Highlands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
Nigussie Haregeweyn, Atsushi Tsunekawa, Jean Poesen, Mitsuru Tsubo,
Jan Nyssen, Matthias Vanmaercke, Amanuel Zenebe,
Derege T. Meshesha, and Enyew Adgo
14 Climatic and Hydrologic Changes in Northern Ethiopia

in the last 3,500 Years: Evidence from the Geomorphic,
Stratigraphic, and Geochemical Archives of Hayk Lake . . . . . . . . . . . . . . . 239
Massimiliano Ghinassi, Marco Benvenuti, Filippo D’Oriano, and
Marialelena Fedi
15 Rift-Related Morphology of the Afar Depression . . . . . . . . . . . . . . . . . . . . 251

Giacomo Corti, Ian D. Bastow, Derek Keir, Carolina Pagli, and Elizabeth Baker
16 Morphometric Characteristics and Hydrology

of Selected Ethiopian Rift Lakes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
Tenalem Ayenew and Merhawi GebreEgziabher
17 The Geomorphology of the Lake Region (Main Ethiopian Rift):

The Record of Paleohydrological and Paleoclimatic Events
in an Active Volcano-Tectonic Setting . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289
M. Benvenuti and S. Carnicelli
18 Water–Rock Interaction and Lake Hydrochemistry

in the Main Ethiopian Rift . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
Azeb Belete, Luigi Beccaluva, Gianluca Bianchini, Nicolò Colombani,
Massimiliano Fazzini, Chiara Marchina, Claudio Natali, and Tewodros Rango
19 The Landscape and Landforms of the Ogaden, Southeast Ethiopia . . . . . . . 323

Daniel Mège, Peter Purcell, Stéphane Pochat, and Thomas Guidat
Part III Applied Aspects
20 Geo-hazard in Ethiopia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 351

Giandomenico Fubelli and Francesco Dramis
21 Land Degradation in the Ethiopian Highlands . . . . . . . . . . . . . . . . . . . . . . 369

Jan Nyssen, Jean Poesen, Sil Lanckriet, Miro Jacob, Jan Moeyersons,
Mitiku Haile, Nigussie Haregeweyn, R. Neil Munro, Katrien Descheemaeker,
Enyew Adgo, Amaury Frankl, and Jozef Deckers
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387
Part I
General Introduction to the Geomorphology
of Ethiopia
Geomorphological Landscapes of Ethiopia
1
Paolo Billi
Abstract
This introductory chapter reports about the main geographic and geomorphologic features of
Ethiopia. The prevailing soil types are described and soil erosion data are reported and
analyzed. A brief description of the natural vegetation is provided as well. The country’s main
geomorphological landscapes regions are identified as follows: (1) the northern highlands,
including (i) the volcanic plug belt of Adwa, (ii) the central highlands and (iii) the
southwestern highlands; (2) the Rift Valley, which consists of three main portions, namely the
northern, central, and southern trunks and the Afar and Danakil depressions; (3) the southern
plateau, which consists of a northern and southern sector and includes also the Ogaden
tableland gently descending to Somalia and the Indian Ocean. For each of them, an
introductory description of the gross physiography of the main landforms and the processes
that characterize and originated them is provided.

Keywords
Ethiopia Geomorphology Landforms Landscapes
1.1 Introduction 1.2 Geography

Ethiopia is characterized by a wide variety of landscapes and
Ethiopia has a surface area of about 1,127,000 km2, that is
landforms. They were generated by a complex of tectonic,
almost twice the size of France. It takes up a large portion of
erosive, and depositional processes acting on rocks with
the inner Horn of Africa since it has no border on the Red Sea
different characteristics. The geological evolution of Ethio-
and the Indian Ocean. The Ethiopian landmass consists of a
pia, with alternating phases of orogenesis, peneplanation,
large, high elevated plateau bisected by the Rift Valley into
crustal updoming, faulting, emplacement of huge amounts of
the northwestern and the southeastern highlands, each with
lava, and deep fluvial dissection has imprinted the geomor-
associated lowlands The contrast in relief is remarkable as
phological landscapes of the country with specific charac-
land elevation ranges between −155 m of Asal Lake in the
teristics, in places unique on Earth. This introductory section
Afar depression (the lowest point in Africa) to the peak of Mt.
is a short summary of the main geographical and landscape
Ras Dejen at 4,620 m a.s.l. in the Simen Mountains (Fig. 1.1).
regions, with an outline of the most distinctive landforms
The plateau stands between 1,500 and 3,000 m a.s.l. and it is
and associated processes that characterize them.
strewn with a number of volcanoes making up high mountain
ranges, the highest of which are the Simen in the north and
the Bale mountains in the south (Fig. 1.1). The northwestern
highlands are considerably more extensive and rugged and
P. Billi (&) are divided into northern, central (centered on the Blue Nile
Physics and Earth Sciences, River catchment downstream of Tana lake), and southern
University of Ferrara,
Ferrara, Italy
sections. The southwestern portion of the plateau (known
e-mail: bli@unife.it also as Somali Plateau) is also rugged, but its elevation is
P. Billi (ed.), Landscapes and Landforms of Ethiopia, World Geomorphological Landscapes, 3

DOI 10.1007/978-94-017-8026-1_1, © Springer Science+Business Media Dordrecht 2015
4 P. Billi
Fig. 1.1 Main geomorphological regions of Ethiopia. Black triangles stand for mountain peaks and downward arrows indicate lowest points
below sea level. The study areas are marked by white rectangles and the numbers refer to the specific chapter of this volume
slightly lower (highest peak Mt. Tullu Dimtu, 4,383 m a.s.l.) Kenya border and Turkana Lake. The highest points of the
than in the northern highlands and it can be subdivided into a Rift floor are between Awasa and Shala lakes (around
northern section with the Ahmar and Garamullata mountains, 1,770 m a.s.l.) and around the Dubeta col (1,670 m s.s.l.)
a central section with the Chercher mountains, the southern (between Ziway Lake and Koka Reservoir), which forms the
part with the Bale and Haranna-Mandebo mountains and the divide between the Awash River basin and the endhoreic
Ogaden. Both the northwestern and the southeastern high- drainage systems of the Main Ethiopian Rift. The Rift bot-
lands are dissected by deep river valleys, in places as much as tom slopes gently to its lowest point at Asal Lake at −155 m
1,500 m deep (e.g., the Blue Nile gorge—Ayalew and Ya- b.s.l. in Djibouti and to the 361 m a.s.l. at Turkana Lake in
magishi 2004), and slope gently toward the Sudan lowlands the south. The floor of the Rift Valley is not uniformly flat as
and the Ogaden and Somali lowlands, respectively, reaching scattered volcanoes or volcanic systems, rising for more than
elevations as low as 500 m a.s.l. 1,000 m (e.g., Zikwala, 2,989, Boset, 2,447, Fantalé,
The Ethiopian segment of the Great Rift Valley is more 2,007 m a.s.l.) occur. Seven major lakes, all closed systems,
than 900 km long between the borders with Kenya and of tectonic or volcano-tectonic origin (Ziway, Langano,
Djibouti and its width varies between 50 and 100 km on Abijata, Shala, Awasa, Abaya—the largest one with a sur-
average. This extensive fault system is bounded to the north face area of 1,162 km2—and Chamo) are located in the main
by the Afar triangle and to the south it proceeds beyond the trunk of the Ethiopian Rift. Other lakes are found in the
1 Geomorphological Landscapes of Ethiopia 5
northernmost portion (Abe Lake, 350 km2, fed by the Awash sandstones, limestones, and the trap series volcanics. This
River) and in the Danakil depression (Afrera Lake, 70 km2) area shows marked contrasts in topography, rainfall, and
(Wood and Talling 1988). Large lakes are also present in the land use, hence soils are highly variable and many soil types
northern and central highlands. Tana Lake is by far the are present (Nitosols, Vertisols, Andosols, Lithosols, etc.).
largest in Ethiopia with its 3,600 km2, whereas the other The northeastern escarpment of Ethiopia has resulted
highland lakes such as Hayk (23 km2) (Demlie et al. 2007) from strong tectonic activity that produced rugged mor-
and Ashenge (20 km2) are much smaller. phology and significantly influenced soil characteristics.
The most notable river system is the Blue Nile and its They are, in fact, affected by severe (natural and man
tributaries, the largest of which is the Tekeze/Atbara, joining induced) erosion processes and show extreme stoniness.
the Blue Nile in the Sudan territory. Because of the general In central Ethiopia highlands, soils developed predomi-
westward slope of the western highlands, many large rivers nantly on trap basalts and subordinately, on pyroclastic
are tributaries of the Blue Nile system, which drains an rocks. Rainfall does not vary much spatially (see Chap. 3,
extensive area of their northern and central part. The Blue this volume) and soils characteristics depend mainly on
Nile, the Tekeze, and the Baro account for about half of the topography. In deeply incised river valleys, soils are thin and
country’s water outflow. Other important rivers originate have a high stoniness degree due to very high erosion rates.
from the Somali Plateau (Genale/Juba and Wabe Shebele In the eastern highlands (Chercher and Ahmar mountains),
rivers) and outflow into the Indian Ocean. Several drainage parent materials consist of Mesozoic sandstones and lime-
basins of smaller rivers are closed systems and the largest stones, flood basalts and, subordinately, of Precambrian base-
among these rivers are the Awash and Omo. ment rocks. In this area, these rocks are very mixed and, due to
strong structural influence and intensive cultivation, soils are
shallow and similar to those of the northeastern escarpment.
1.2.1 Soils To the west, where the structural influence is less marked and
rainfall increases, soils show more vertic characteristics.
Ethiopia presents a large variety of soil types. Berhanu et al. The northern portion of the Rift Valley includes semi-
(2013) classified the land mass of Ethiopia into 60 soil types desert areas with the exception of the Awash River valley
with an area coverage ranging from 1.4 to 208,882 km2. fill. Parent materials of this area include Tertiary and Qua-
Lithic Leptosols, Humic Nitisols, and Eutric Vertisols are ternary volcanic, alluvial, and colluvial deposits. The very
the major three soil types with area coverage proportion of recent age of these parent rocks and the dry weather con-
18.5, 11.9, and 10.2 %, respectively. These prominent per- dition (annual precipitation range between 300 and 500 mm)
centages are comparable with those reported by the FAO lead to soils that developed mainly on alluvial and colluvial
Soil Database (FAO 1998), though only 45 different soil materials. Eutric and Calcaric Fluvisols are common in the
types were identified by this study. alluvial plain of the Awash River.
On the base of the harmonized soil database constructed By contrast, the geology of the southern portion of the
by FAO (2009), Berhanu produced a textural classification Rift Valley is more complex. Bedrock consists mainly of
of Ethiopian soils and their relative frequencies of area Tertiary pyroclastics and Quaternary basalts but the Rift
coverage, which reveals that loam and sandy loam soils are floor includes also extensive lacustrine, colluvial, and fan
the most common soil textures. deposits. Vertic and Mollic Andosols prevail in areas
Following the UNDP/FAO (1984) report on the geo- underlain by volcanic ash and pumices, whereas Lithosols
morphology and soils of Ethiopia, a short description of the are more common where Quaternary basalts outcrop.
main soil characteristics of different physiographic regions In the Ogaden, sedimentary rocks prevail, namely lime-
can be outlined. stones, sandstones, and evaporates. In this area, precipitation
In western Ethiopia (Gamo Gofa, Ilubabor, Welega, and is scarce (commonly less than 400 mm yr−1), evapotrans-
part of Gojam and Gonder), soils develop mainly on felsic piration is high and uniform and topography is compara-
and metamorphic Precambrian rocks and flood basalts. Since tively flat, hence it is the parent material to play the most
in this part of Ethiopia the highlands slope down to the relevant role in soil differentiation and typically Gypsic and
lowlands, soils develop also on alluvial and colluvial Calcic Xerosols and Yermosols occur. In the eastern Oga-
deposits. In the highlands, the high rainfall (over den, underlain by sandstones, Cambic Arenosols prevail.
2,000 mm yr−1) is the most important factor in producing Finally, south of the Bale Mountains, flat landforms
very similar soils irrespective of the parent rocks. In western dominate and parent rocks consist mainly of deeply weath-
Illubabor, fluvial soils are found on the alluvial plain of the ered rhyolites on which deep Eutric Cambisols occur,
Baro River and its tributaries. whereas steeper slopes are underlain by Precambrian
The northern highlands are underlain by Precambrian gneisses and granites, and here Lithic phases of the Eutric
metamorphic rocks, Upper Paleozoic and Mesozoic Cambisols prevail.
6 P. Billi
1.2.2 Soil Erosion Table 1.2 Soil loss in Ethiopia

Author/s Soil loss (t ha−1yr−1)
Soil erosion in Ethiopia varies widely due to the very dif- Hurni (1988) 42
ferent physiography, land use, and soil characteristics. Sev-
Wright and Adamseged (1986) 100
eral studies have investigated the rates and causes of soil loss
Sutcliffe (1993, cited in Kappel 1996) 40
in Ethiopia, but most of quantitative data are derived from
plot studies (e.g., Hurni 1985; Soil Conservation Research Bojo and Cassells (1994, cited in 20
Kappel 1996)
Project—SCRP—Grunder 1988; Herweg and Ludi 1999) or
Stocking (1996) 165
studies were focused on restricted areas (e.g., Tegene 2000;
Nyssen et al. 2004). Bojo and Cassells (1994), however, Tamene and Vlek (2008) 14
argued that soil erosion data obtained from plot studies must Average 63
be corrected including also the delivery ratio factor in order
to assess realistic erosion rates, whereas large areas or
regional studies are mainly based on soil loss models rather vegetation removal, and erosive rain seems to play a major
than field measurements. role. Rainfall intensity of 100 mm in 24 h is calculated by
Soil loss measured in a few sites of the SCRP (Grunder Billi et al. (2015) to have an average return time of 25 year
1988) is reported in Table 1.1. These data show a large range (range 6.3–68.6 years). Land degradation is a common
of values given the different physiography, precipitation, soil feature of the Ethiopian landscape with severely eroded
characteristics, and land use of the regions considered. The areas along the rift margins (Fig. 1.2) and in the northern
plot studies by the SCRP (Grunder 1988) showed also that highlands. Other common erosion processes are gullying
the traditional conservation practices are the least effective in (Fig. 1.3) and landsliding (Fig. 1.4).
combating soil erosion. Table 1.2, on the other hand, reports Soil erosion rates obtained from direct measurement of
soil erosion data extrapolated for the whole country by dif- soil dislodgements from slopes is, however, about two
ferent authors and the data range is again large. This is orders of magnitude higher than the sediment yield data
probably due to different reference soil loss values used by obtained from river sediment flux measurements. A few
the authors. authors, in fact, report highly variable sediment yield values
Berhanu et al. (2013) calculated the erodibility factor ranging from 4 to 3,784 t km−2yr−1 (several sources, see
(K) of different soil types. Unexpectedly, they found rela- Chap. 6 this volume). Nyssen et al. (2004) developed a
tively low values as K ranges between 0.00 and 0.18. Similar simple power equation from literature data on 20 river sites
results were obtained also by Nyssen et al. (2007) in Tigray in drainage area outside the less-erodible areas in southern
and by Shiferaw (2012) in the Borena area of South Welo Ethiopia as
highlands. These conclusions indicate that in Ethiopia
soil erodibility is not an explanation for the magnitude of Ys ¼ 2;595 A0:29 ð1:1Þ
actual land degradation as the combination of steep slopes,
in which 56 % of the variability of sediment yield (Ys in
t km−2yr−1) is explained by catchment area (A in km2). This
Table 1.1 Soil loss in various parts of Ethiopia (Berehe 1996)
result is partially confirmed by field measurements of Billi
Region Soil loss (t ha−1yr−1) (2004) on the Meki River, upstream of the homonymous
Gojam 40.2–199.2 town (only a few kilometers upstream of the river outlet into
South Welo 36.5–53.8 Ziway Lake), which gave a sediment yield of about
North Shewa 152.4–214.8 60 t km−2yr−1. By contrast, field measurements by Harege-
Illubabor 18.0–135.3
weyn et al. (2008) on very small rivers in northern Ethiopia,
ranging between 0.72 and 24 km2 in catchment area, resulted
Harerge 25.5–27.8
in a much higher average sediment yield of 947 t km−2yr−1
Sidamo 41.2–49.5
(range 446–1,817 t km−2yr−1). Also these authors provided
Fig. 1.2 Severe land degradation

on the main Ethiopian Rift
escarpment near Alaba Kulito
along the Soddo-Shashemene
road. The top soil has been
completely removed by erosion
leaving a badland-type
morphology developed into in the
soft bedrock consisting of poorly
consolidated volcanic ashes and
pumices
Fig. 1.3 A deep gully incised

into the colluvial deposits
mantling the base of the Gade
Motta caldera rim (7°57′26″
N–38°39′20″E). This box or
U-shaped gully is very deep, with
steep flanks the hight of which
appreciably decreases
downstream as far as the gully
splay on the rift bottom
8 P. Billi
Fig. 1.4 Block sliding apart due

to lateral spreading from the
Amba Aradam Mountain near
Antalo (13°17′50″N–39°25′07″
E). Notice the varicolored sub-
horizontal, Jurassic Agula shales
unconformably overlain by the
Cretaceous Amba Aradam
sandstones
an interpolation but, conversely to Nyssen et al. (2004), it is soil types and climatic conditions result in a variety of
expressed by a less significant (R2 = 0.36) linear equation habitats suitable for evolution of several plant species.
Vegetation of Ethiopia, therefore, is very heterogeneous and
Ys ¼ 0:005 A þ 6 ð1:2Þ characterized by considerable endemism which is strongest
in the high mountains, in southeastern Ogaden (Soromessa
which predicts sediment yield to increase with increasing et al. 2004), Borana and Bale lowlands (Woldu 1999 in FAO
catchment area. This result confirms that in small catchments 2014). In Ethiopia, there are about 6,000 species of higher
slope erosion processes are not stationary because of local plants, of which about 10 % are endemic (Kelbessa et al.
factors, but are very effective in supplying sediment directly 1992).
into stream channels. Notwithstanding such wide variations According to White (1965), the Sahel region, which can
of data from different authors, sediment yield seems to be be characterized as wooded steppe with Acacia and Com-
rather high, especially in the northern highlands (Fig. 1.5) miphora spp. (Fig. 1.8), represents a floristically impover-
and in the Rift margins as witnessed by a few examples of ished western extension of the rich Afro-Oriental domain.
reservoirs completely filled with sediment (Fig. 1.6). Acacia species are very common in the Ethiopian Rift Valley
and other low rainfall areas but are also found in the
southern highlands where they form the basis of a traditional
1.2.3 Natural Vegetation and Land Use agricultural practice (Poschen 1986). In the arid zone plains,
bushed grassland prevails (Fig. 1.9), except some patches of
According to Ibrahim (1978), Ethiopia lies within the Sud- woodland.
ano-Zambezian phytogeographic region of Africa which The highlands of Ethiopia have climates and vegetations
comprises one of the largest formations of the continent and that vary noticeably in relation to altitude, forming part of
can be characterized as tropical, with a long dry season of the Afro-montane region. Ibrahim (1978) includes Ethiopia
4–9 months, annual rainfall between 200 and 1,500 mm, and in the Afro-Oriental domain, which covers also the lowlands
vegetation typical of Sudanian and Sahelian zones which of Tanzania, Kenya, and Somalia, and sets an altitudinal
includes steppe, savanna, and dry (Fig. 1.7) to subhumid limit of 1,100 m for this domain though it extends consid-
woodland and forest. The large differences in physiography, erably higher in the Ethiopian Rift Valley.
Fig. 1.5 A mushroom stone

witnessing severe soil erosion
north of Aksum (14°09′53″N–
38°41′58″E)
Fig. 1.6 The Aba Samule

Reservoir on the Akaki River,
south of Addis Ababa
(8°47′15″N–38°42′18″E).
The dam was constructed in 1939
but a few decades later it was
completely filled with sediment
The forest cover of Ethiopia has been declining rapidly. (Fig. 1.11) are indicated as the main causes for forest deg-
Most of the remaining forests are confined to the south and radation (Tilahun et al. 2011). According to Woldu (1999 in
southwestern parts of the country (Tilahun et al. 2011). FAO 2014), about 34 % of Ethiopia and 57 % of the land
In these areas, forests are threatened by human activities. above 1,500 m was once covered by dense forests and a
Historical documents show that Ethiopia experienced sub- further 20 % by wooded savannah. Massive deforestation has
stantial deforestation (Pankhurst 1995; McCann 1997; Dessie reduced these figures to 3.6 % of the total area (Tefera 2011)
and Kleman 2007), soil, and land degradation (Nyssen et al. (Fig. 1.12) and to 9 % of the land above 1,500 m. Widespread
2004; Nyssen et al. 2014) over the years. The need for deforestation started, particularly in the highlands, at the end
fuelwood, arable land (Fig. 1.10) and grazing areas of the nineteenth century with the expansion of agriculture.
10 P. Billi
Fig. 1.7 Example of a natural

savannah and dry woodland
forest in the southern portion of
the Rift Valley, southeast of
Konso town
Fig. 1.8 A natural Acacia forest

within the Abijata-Shala Park
(7°71′34″N–38°39′46″E)
The glorious forests of the past are witnessed by isolated, Australia to East Africa in the late nineteenth and early
huge sycamore trees that are present in the Rift and in the twentieth century and at that time the largest plantations
highlands below 2,000 m a.s.l. (Fig. 1.13) were in Ethiopia and Rwanda. In Ethiopia, the Eucalyptus
Nowadays, the most common trees of Ethiopia are the genus was introduced in 1894/1895. The purpose was to
Eucalyptus spp. (Fig. 1.14). This genus was introduced from supply fuelwood and construction timber to the new and
Fig. 1.9 Example of a bushed

grassland in “Nech Sar” (white
grass) Park near Arba Minch
(5°57′24″N–37°39′31″E)
Fig. 1.10 Cereals cultivations on

the Somali Plateau along the
Shashemene-Kofele road. This
flat area is underlain by trap
basalts and stretches eastward
from the Rift margin
growing capital city, Addis Ababa. In the 1970s, the plan-

tation area in the country was about 90,000 ha. Recent 1.3 Main Landscape Regions
estimates indicate the extension of Eucalyptus forests for
about 0.5 × 106 ha though huge numbers of trees exist in The geomorphology of Ethiopia is largely controlled by its
other land use types, such as homesteads, farm boundaries, geological structure, but weathering, erosion, and deposition
and beside roads (Dessie and Erkossa 2011). processes contributed as well in shaping the country into its
In the main Rift Valley, agricultural activity is rapidly present association of landforms and landscapes. The crustal
expanding and progressive settlement has replaced grazing evolution, characterized by a marked swelling within one of
lands with small (Fig. 1.15) to medium farms, some of the most active extensional areas of the planet, resulted in
which are mechanized (FAO 2014). three main morphostructural units, which include also
12 P. Billi
Fig. 1.11 Pasture on the western

portion of the central highlands
near Nekemte. Major landforms
of this are small volcanic plugs
and dome-shaped subvolcanic
intrusions
Fig. 1.12 Land cover

distribution by percentage of the Afro-Alpine
Ethiopian territory (Woldu 1999)
0.2 Bareland
Cultivation
13.78
26.93 Grassland
19.62 Highland Bamboo
0.67 Natural Forest
0.73 21.31 12.44 Plantation
0.02 Shrubland
Urban
0.23 3.63 0.44 Water
Wetland
Woodland
subregions with specific geomorphic features. These are surmounted by high volcanic mountains, a few of which
given as follows: (1) the western plateau that can be divided reach elevations above 4,000 m. Both main plateaus descend
into (i) the northern highlands, including the volcanic plug radially and gradually to elevations of about 500–600 m
belt of Adwa, (ii) the central highlands, and (iii) the south- around the border with Sudan and 200–400 m close to the
western highlands; (2) the Rift Valley which consists of border with Somalia. The main topographic gradient is
three main portions, namely the northern, central, and steeper (about 0.004–0.006) near the rift margins and
southern trunks and the Afar and Danakil depressions; (3) decreases to 0.001–0.002 in the distal parts where granite
the southern plateau which consists of a northern and inselbergs are a common and typical geomorphological
southern sectors and includes also the Ogaden tableland feature (Figs. 1.16 and 1.17). By contrast, the transition from
gently descending to Somalia and the Indian Ocean. the plateau margins to the rift is decidedly abrupt and
The plateaus originated by the domal uplift of the Ara- marked by stepped topography and a number of elongated
bian-Ethiopian region (Merla et al. 1979) and accumulation grabens (Fig. 1.18), with asymmetrical and unpaired horsts
of flood basalts, the thickness of which is commonly around and the bottom locally punctuated by recent trachyte plugs
1,000 m but may reach 2,000 m in some regions (Kieffer (Fig. 1.19). In places, spectacular fault-generated escarp-
et al. 2004) or even 3,000 m in the Chercher mountains ments occur (Fig. 1.20), with long and high exposed fault
(Juch 1975) (Fig. 1.1). The most elevated parts of the pla- planes, at the base of which alluvial fans are commonly
teaus stand at 3,000 m a.s.l., but in places they are found (Fig. 1.21).
Fig. 1.13 A huge sycamore tree
Fig. 1.14 Eucalyptus forest on

the Somali Plateau along the
Shashemene-Kofele road
14 P. Billi
Fig. 1.15 A small farm on the

shore of Abaya Lake, along
the Soddo-Arba Minch road
(6°30′00″N–37°46′33″E)
Fig. 1.16 Granite inselbergs,

with evidence of exfoliation
processes and products, in the
eastern side of Kassala
1.3.1 Northern Highlands low grade Proterozoic metamorphic rocks, Lower Paleozoic
granites and, subordinately Mesozoic sedimentary sequen-
The geology of the northern highlands is rather complex and ces occur, whereas the southern sector is mainly underlain
includes a variety of formations ranging in age from Pre- by volcanites of the trap series with the typical tabular
cambrian to Quaternary. Such a geodiversity provides the morphology of the flood basalts (Fig. 1.22). Continuing
northern highlands with, likely, the most varied association volcanic activity resulted in the building up of high
of landforms and landscapes. In the northernmost sector, mountain areas which include a few of the highest peaks of
Fig. 1.17 Granite inselbergs

along the road from Mega to
Moyale. Notice typical landforms
such as tors originated by
crystalline rocks weathering
Fig. 1.18 Structural basins

within the western plateau
escarpment south of Karakore,
along the Debre Birhan-Dese road
(10°24′35″N–39°56′05″E). This
small graben is asymmetric with
the western flank (on the right in
the photo) higher. The graben
floor is also slightly tilted and
dipped southward (away from the
reader)
Ethiopia. The Simen Mountains massif dominates the almost negligible slope of the area and the occurrence of
western plateau and it is the fourth highest range of Africa hard Mesozoic sedimentary formations such as the Adigrat
after Kilimanjaro, Mount Kenya and Rwenzori, with Ras Sandstone, consisting of 91 % of quartz components
Dejen (4,533 m a.s.l.) being the highest mountain peak of (Getaneh 2002), resting on more erodible glacigenic and
Ethiopia and among the first ten highest peaks of the metamorphic rocks or, like in the case of the Amba Aradam,
African continent. In the northern highlands, other notable on softer and easily erodible Agula Shales (Fig. 1.4),
high mountain peaks are the Abune Yosef (4,284 m a.s.l.), the Amba mountain geomorphology is rather common in
in the Tekeze headwaters, and a number of high mountains, Tigray (Fig. 1.24).
some ranging in elevation between 3,000 and 4,000 m, the Drainage of the northern highlands is predominantly to
most renown of which, for historical reasons, is Amba Alaji the west-northwest and includes the largest and longest
(3,438 m a.s.l.) (Fig. 1.23). Amba is a local name that is rivers of northern Ethiopia. By contrast, much smaller and
commonly used to indicate flat-topped mountains (see commonly ephemeral streams flow to the east into closed
Chap. 9 this volume). Given the tabular structure with an basins or disappear in the Danakil lowlands. All the western
16 P. Billi
Fig. 1.19 Recent (Quaternary?)

intrusions of trachytic rocks
protruding from the bottom of the
Kobo-Alamata structural basin
(12°07′39″N–39°41′15″E). View
is to the north
Fig. 1.20 Impressive fault

escarpment in the Termaber
basalts with alluvial fans at the
base near Wichale, between Dese
and Weldiya
Fig. 1.21 Alluvial fan in the

footslope of the eastern margin of
the Kobo-Alamata structural
basin (12°22′25″N–39°41′31″E)
Fig. 1.22 The typical tabular

structure of the northern
highlands with a deep canyon
incised by a tributary of the
Mereb River along the Adigrat-
Adwa road. The right-hand side
of the valley is capped by the
Paleozoic Enticho sandstones
(Abbate, personal
communication), whereas in the
opposite side the sandstones are
overlain by the Tertiary trap
basalts
Fig. 1.23 The Amba Alaji peak

(3,438 m a.s.l.) seen from the
Korem-Mekele road. The
pyramidal morphology of the
mountain top is carved into
Miocene rhyolitic ignimbrites
rivers are tributaries of the two main river systems of this of the northern highlands is mainly the result of erosion,
area, the Tekeze and the Mereb. Both these rivers flow on rather than deposition processes, as would be expected given
the bottom of deeply incised canyons (Fig. 1.22) and mark the remarkable uplift of the area that has proceeded until
the border between Eritrea and Ethiopia for a few hundreds recent times (Merla et al. 1979; Corti and Manetti 2012).
of kilometers. At the large scale, in fact, the geomorphology However, in a few places underlain by limestone rocks
18 P. Billi
Fig. 1.24 Flat-topped (amba)

mountains along the road from
Mekele to Adigrat in Tigray. This
tabular structure is formed by
sub-horizontal, Triassic Adigrat
sandstones which in places
include dark brown ferruginous/
lateritic beds, very resistant to
erosion (Enkurie 2010)
Fig. 1.25 The travertine dam

of Romanat, a few kilometers
NW of Mekele
(13°34′24″N–39°25′04″E)
(e.g., Antalo formation), such as in the Mekele outlier, thick Axum–Adwa complex is part of a larger magmatic SW–NE
deposits of travertine, forming spectacular natural dams, are belt extending for about 150 km from west of Axum to as far
present (Fig. 1.25). as Senafe in Eritrea. This magmatic complex is set along an
Among the northern highlands landscapes, the one around uplifted crustal sector overlying the Proterozoic crystalline
Adwa is unique and remarkable. It is characterized by a large basement or, in many places, its Mesozoic sedimentary cover
number of spires that rise from the eroded surface of the (Zanettin et al. 2006; Natali et al. 2013). The magmatic
basement to make a sort of stony tree forest (Fig. 1.26). The products consist mainly of an alternation of trachyte and
Fig. 1.26 The volcanic plugs of

Adwa seen from Axum
Fig. 1.27 Volcanic plugs near

Enticho
syenite plugs and domes reaching their maximum density in The plugs and domes rise from a few tens of meters to
the area around Adwa. The world renown obelisks of Aksum about 300 m above the sedimentary substrate or alkaline
were carved from the subvolcanic phonolite-syenite domes basalts (Hagos et al. 2010) (Fig. 1.27). A few of the most
that dominate the landscape around the town (see Chap. 7 this peaked and isolated pinnacles host monasteries and churches
volume). (Fig. 1.28).
20 P. Billi
Fig. 1.28 The church of St.

Pantaleon on top of a pointed
trachyte plug near Aksum
Fig. 1.29 The central highlands

north of Addis Ababa are deeply
incised by the Sodoblé River, a
tributary of the Blue Nile near
Chancho (9°26′11″N–38°38′47″
E). Notice the alternation of softer
and harder lithotypes within the
trap series on top, the sub-vertical
fault scarp parallel to the river and
cutting the lower portions of
smaller transverse divides and the
Mesozoic sedimentary formations
outcropping in the valley bottom
1.3.2 Central Highlands volcanic rocks. Only to the west, Proterozoic metamorphic
rocks crop out, whereas the Mesozoic sedimentary sequence
The central highlands are likely the most monotonous of the is exposed in the Blue Nile valley. The general morphology
Ethiopian highlands since they are almost entirely underlain is therefore that of a tabular structure gently sloping to the
by basalts of the Tertiary trap series and other, more recent west and deeply incised by the Blue Nile and its tributaries
Fig. 1.30 The flat top surface of

the central highlands north of
Addis Ababa with shield
volcanoes merged to form a
string-shaped ridge in the
background
Fig. 1.31 A sharp, sub-vertical

fault face, with a hanging river
valley, along the road between
Karakore and Kombolcha
(Fig. 1.29). The top surface is typically flat and strewn with Lake, separated by the Blue Nile which flows to the south
individual volcanic edifices or groups of a few contiguous and then turns to the west in a large bend right around Mt.
apparatuses, commonly arranged in strings along fault lines Choke. Other high mountains include the Abuye Meda
(Fig. 1.30). Some of these volcanoes may be very large and (4,000 m a.s.l.), south of Kombolcha, Amba Farit (3,975 m
stand higher on the plateau surface, rising 1,000–1,500 m a.s.l.), west of Dese, and Tulu Welel (3,200 m a.s.l.) in the
above it. This is the case of Mt. Guna (4,231 m a.s.l.) near westernmost portion of the plateau, close to the lowlands
Debre Tabor and Mt. Choke (4,154 m a.s.l.) south of Tana across the border with Sudan.
22 P. Billi
Fig. 1.32 The famous “Afar

Window” along the road from
Debre Birhan to Debre Sina
(9°50′13″N–39°44′30″E). A
small fault, transverse to the main
rifting system, makes a narrow
gap within the plateau edge and
from an elevation of about
3,100 m a.s.l. it is possible to see
the Afar lowland in the far
distance
Fig. 1.33 The rugged landscape

of the southwestern highlands.
The Omo River valley west of
Welkite (8°15′40″N–37°36′09″E)
Apart from volcanoes and their products, the northern 1.3.3 Southwestern Highlands
highlands gross geomorphology is mainly controlled by
faulting and horst and graben structures. Long, high and The southwestern highlands are less even and more rugged
commonly sub-vertical fault faces are common as well, compared to the northern and central highlands. In particular
especially close to the eastern margin (Fig. 1.31). This latter in the median part, the landscape assumes a more typical
is rather sharp and the relief contrast with the Danakil and mountainous configuration consisting of deep valleys and
Afar lowlands is remarkable (Fig. 1.32). Moving westward, mountain groups (Fig. 1.33). The north- and southeastern
the plateau elevation decreases but the general landscape sides coincide with the margins of the main and the southern
does not change appreciably, at least as far as Nekemte Ethiopian Rift Valley, respectively, where the highest
(Fig. 1.11). mountain peaks are found: Mt Gurage (3,721 m a.s.l.) in the
Fig. 1.34 The main Ethiopian

Rift valley floor near Ziway seen
from the rim of the Gademotta
caldera. The Rift bottom is
underlain by Quaternary
lacustrine, fluvio-lacustrine,
volcano-lacustrine, and volcanic
ashes
Fig. 1.35 The stepped

morphology produced by
normal faults on the Rift
margin east of Ziway Lake
(8°03′20″N–39°07′14″E)
northernmost part, Mt. Guge (4,200 m a.s.l.) west of Arba western Ilubabor (400–500 m a.s.l. on average), where
Minch and Mt. Malgudo (3,390 m a.s.l.) southeast of Jimma. wetlands are also present (Woldu and Yeshitela 2003).
To the west, the southwestern highlands maintain their
general mountainous landscape, but the valleys become
broader and the mountain and valley bottom elevation 1.3.4 The Rift Valley
decreases gradually as far as the town of Gambela, east of
which there is a quite abrupt transition, along a almost The Great Rift Valley of Africa is one of the largest tectonic
north–south alignment, from the 2,000 m a.s.l. of the area structures on Earth and one of the most attractive regions of
close to Fit Makonnen to the tropical lowlands of the the whole continent for its geomorphological relevance and
24 P. Billi
Fig. 1.36 An example of

Gilbert-type delta in the recent
shore deposits of Shala Lake
(7°28′41″N–38°38′14″E)
Fig. 1.37 The Awash River

flowing entrenched into the
Rift valley floor near Ombole
(8°23′22″N–38°45′53″E), with
the Zikwala Volcano in the
background. The river
entrenching is caused by base
level drop associated with main
Ethiopian Rift lakes shrinkage
and Afar tectonics
unique natural environments. The Ethiopian Rift consists of to the apex of the Afar triangle and its northeastern branch,
three main sections: (i) the southern portion, from the border the Danakil depression.
with Kenya to the hydrological divide between Abaya and The width of the Rift bottom is relatively constant around
Awasa Lake; (ii) the main portion, commonly indicated also 50 km (Fig. 1.34), with the narrowest section being less than
as the Great Lakes Region of Ethiopia, stretching as far as 20 km at Arba Minch. By contrast, the top width between
the Awash River; (iii) the northern portion, from the Awash the plateaus edges is more variable and depends mainly on
Fig. 1.38 A very small cinder

cone along the Nazreth-Metehara
road (8°53′57″N–39°48′37″E)
Fig. 1.39 The Fantalè shield

volcano seen from the
Awash National Park
(8°54′17″N–40°02′36″E)
the number and size of the down-faulted blocks making up deposits (Fig. 1.36). Its main geomorphic features include
the stepped morphology of the Rift margins (Fig. 1.35). It is narrow lacustrine terraces (visible especially in the Ziway–
narrowest at Arba Minch and widest at Addis Ababa where Shala lake area), river deltas, and shallow canyons
it can be as much as 120 km. The Rift margin scarps are (Fig. 1.37). River entrenchment is mainly due to the base
mainly cut into the rocks of the trap series in the southern level drop associated with the recent lakes shrinkage (see
portion and the Danakil, whereas in the Lakes Region and Chaps. 16 and 17 this volume) and the Afar lowland for-
the northern portion (namely around Dire Dawa), more mation. However, the most distinguished geomorphic fea-
recent volcanic and Mesozoic sedimentary rocks, respec- tures of the Rift bottom are volcanoes and lakes. Volcanoes
tively, prevail. punctuate the Rift floor as individual or merged edifices
The Rift floor is generally flat as it is commonly underlain ranging in size from small cinder cones (Fig. 1.38) to large
by pyroclastic material and lacustrine and fluvio-lacustrine strato-volcanoes (Fig. 1.37) or shield volcanoes (Fig. 1.39).
26 P. Billi
Fig. 1.40 A nice example

of maar west of Butajira
(8°02′52″N–38°20′58″E)
Calderas are also very common and the largest of them may conclude that water balances of the main Rift lakes are pre-
have a diameter spanning a few tens of kilometers. Calderas dominantly controlled by groundwater. A special case is
host the majority of the main Ethiopian Rift Valley great offered by fast and large expansion of Beseka Lake that
lakes (namely Langano, Abijata, Shala, and Awasa), whereas occurred in the last decades. The maps of Fig. 1.41 show
a few much smaller lakes are formed within maars (Fig. 1. clearly an impressive lake surface increase between 1979 and
40). Most of the larger lakes hold saline water (Abaya, A- 1995. The water level is still rising leading to the complete
bijata, Awasa, Beseka, Chamo, Langno, and Shala) with total drowning of both the road and the railway to Djibouti.
dissolved solids concentrations ranging from 771 ppm of Human impact is likely the main controlling factor of such a
Abaya Lake to 56,300 ppm of Beseka Lake (Wood and remarkable change in water level that started after the con-
Talling 1998; Chap. 18, this volume). The Rift lakes make up struction of a large irrigation scheme, fed by the Awash River
the base level for all the rivers draining into the Rift. With the water, for a newly set sugar plantation (Fig. 1.41).
exception of the lakes fed by the Awash River, such as Abe The Rift becomes progressively more arid as we move to
Lake, all other lakes are fed by small rivers (Chap. 16, this the Afar triangle and its northern branch, the Danakil
volume). In the recent decades, the Rift lakes have shown depression. These areas are described in detail in Chap. 15 of
contrasting hydrological trends with some lakes shrinking this volume. Here, it is worth recalling briefly that the Danakil
(e.g., Abijata) and other—Abaya, Awasa and Beseka— depression is a strip of coastal land 60–80 km wide separated
expanding (Chap. 16, this volume). Given the remarkable from the Red Sea by coastal hills draining into the inland
rate of change in water level observed (e.g., saline lakes. The Danakil depression is the hottest inhabited
Abijata −6 m, Abaya +2 m and Awasa +3 m in 20 years), not and one of the least elevated places on Earth and location of
paired by an equivalent change in rainfall in the vicinity of unique landscapes, associated with different magmatic
these lakes, it is very difficult to ascribe the lake surface activities, and evaporite deposits as much as 2 km thick
variations to climatic factors only. For Awasa Lake, forming a vast flat area from which salt domes, a few meters
Gebreegziabher (2004) attributes the lake level rise to the high, protrude (Fig. 1.42). Numerous volcanoes and volcanic
combined effects of land use and climate change, whereas landscapes occur in the Danakil, including the most active
Ayenew and Gebreegziabher (Chap. 16, this volume) and renown of them, the Erta Ale (see Chap. 15, this volume).
In the Bale mountains, we find the highest peaks of the

Somali Plateau and among the highest of Ethiopia, Mt. Tullu
Dimtu (4,383 m a.s.l.) and Mt. Batu (4,307 m a.s.l.), but
other high peaks include Mt. Kaka (4,180 m a.s.l.) and Mt.
Bada (4,130 m a.s.l.) in the Chercher range. Moving to the
north, along the plateau margin, the mountain height tends to
decrease with the highest peak of the Ahmara Mountains,
Gara Muleta, being only 3,381 m a.s.l. (Fig. 1.43).
The Somali Plateau is rather flat (Fig. 1.10) and its tabular
morphology is interrupted only by deep canyons of the main
rivers, i.e., the Wabe Shebele and the Genale-Dawa
(Fig. 1.44), and high massifs. Among the latter, the Bale
Mountains are decidedly the most prominent and consist of
few main volcanoes, that released large amounts of Oligo-
Miocene basalts and Quaternary rhyolite ignimbrites and
basalt lavas, and fluvio-lacustrine intercalations (Merla et al.
1979). The summit of the Quaternary volcanites is punctu-
ated by volcanic plugs (Fig. 1.45) and glacier erosional and
depositional features are present (see Chap. 6, this volume).
The latter are particularly evident in the Sanetti plain which
is a large plateau at an average elevation of 4,100 m a.s.l.,
punctuated by small lakes, that has been shaped by an ice
cap cover of about 180 km2 during the last glacial maximum
(Osmaston et al. 2005) (Fig. 1.46). Here, erratic boulders,
roche mountonnées, and moraines are present, whereas cir-
ques and moraine lakes (tarns) are common in the glaciated
valley that radially spread out from the northern side of the
plateau (Osmaston et al. 2005).
Evidence of glaciated landforms can be found also on the
peaks of Mts. Bada and Kaka in the southern portion of the
Chercher mountains (Osmaston et al. 2005), but is missing
in the northern, less elevated part in which the most prom-
inent landforms consist of small intermontane, structural
basins, especially along the plateau margin (Fig. 1.47).
Most of the central and southern Somali Plateau is known
as Ogaden—a large region, which extends from the Cher-
cher and Ahmar mountains (Fig. 1.43) on the northern
Somali Plateau margin to the border with Somalia and
Fig. 1.41 1:250,000 topographic maps of Beseka Lake area reporting
beyond, to the Indian Ocean. The Ogaden is a relatively flat
the state of the lake in 1979 and in 1995. Notice the huge sugar
plantation constructed after 1997 near the lake, south of Metehara and geomorphologically rather monotonous tabular land,
gently dipping to E-SE. Its main morphological features are
associated with large rivers crossing it, namely the Wabe
1.3.5 The Somali Plateau Shebele River and its left side tributaries, the Ramis, Erer,
and Fafen rivers (see Chap. 19, this volume). In the most
The physiography of the Somali Plateau is similar to the elevated northern portion of the plateau, these rivers have
northwestern highlands, but its elevation is slightly lower. incised deep valleys (Fig. 1.44) which become wider and
The highest mountain ranges, Bale, Chercher, and Ahmar shallow as the rivers approach the Somali border.
(Fig. 1.1), are aligned with the western margin and close to Though most of the Ogaden is underlain by Mesozoic
its stepped escarpment. From here, the plateau gently des- sedimentary rocks, in its northern portion, especially in the
cends to S-SE and sinks into the Indian Ocean. Jijjiga area, the flat landscape is interrupted by small
28 P. Billi
Fig. 1.42 Salt hillocks rising

from the Dallol salt plain
(14°14′16″N–40°17′51″E)
Fig. 1.43 The Gara Muleta

range seen from Kurfa Chele
(Curfacelli) village, west of Harar
(9°14′11″N–41°49′22″E)
volcanic plugs consisting of Miocene to Pleistocene basalts parallel to the course of the Wabe Shebele, east of Imeey
and rhyolites, whereas the town of Jijjiga stands on about (Merla et al. 1979). Other small, isolated basalt hills occur
100 m thick Quaternary fluvio-lacustrine deposits accumu- also in between Warder and the Somali border.
lated in a large basin. Small, flat-topped residual hills made
Acknowledgments I am grateful to all the very many farmers, peasants
of limestone resting on the main gypsum formation rise for a
and car drivers that throughout 25 years of field research all across
few tens of meters above the surrounding tableland east of Ethiopia, freely and collaboratively provided substantial help on many
the confluence between the Dacata River into the Wabe occasions. Without their hospitality and collaborative approach this
Shebele (Merla et al. 1979). Elongated basalt hills (see work would have never seen the light. Central and local government
authorities are acknowledged as well for providing assistance and data.
Chap. 19, this volume) runs southward from Jijiga and
Fig. 1.44 The Wabe Shebele

canyon near Gasera village,
northeast of Robe Bale town
(7°23′05″N–40°09′03″E)
Fig. 1.45 A number of volcanic

plugs towering on top of the Bale
Mountains, along the
Shashamane-Robe road
I would like also to thank the colleagues that shared with me the hard life finally, to Fabrizio Vannacci that faithfully assisted me with his labo-
in remote areas and the students that contributed with substantial, basic ratory, field and human skills in such a long endeavor. Renato Gerdol is
data. A special thank to Ernesto Abbate, Mario Sagri and Milvio Faz- greatly acknowledged for reviewing the section on vegetation. Many
zuoli that introduced me to the beauties of the Horn of Africa geology thanks also to Carlo Bisci for providing the base map of Fig. 1.1 and to
and geomorphology, to Giovanni Ferrari for his inspiring views and, Lorenzo Orioli for the information and photos of the Sanetti Plain.
30 P. Billi
Fig. 1.46 The Sanetti plain

within the Bale Mountains
National Park, at an average
elevation of 4,100 m a.s.l.
(photograph by L. Orioli)
Fig. 1.47 Unnamed mountains

around an intermontane structural
basin along the Asbe Teferi—
Hirna road in the northern part of
the Chercher mountains
References Billi P (2004) Sediment yield of a closed river system: the Meki River,
western margin of the Ethiopian Rift Valley. In: Proceedings of the
IAHS international symposium on sediment transfer through the
Ayalew L, Yamagishi H (2004) Slope failures in the Blue Nile basin, as fluvial system, Moscow, 2–6 Aug 2004, pp 94–100
seen from landscape evolution perspective. Geomorphology 57:95– Billi P, Alemu YT, Ciampalini R (2015) Increased frequency of flash
116 floods in Dire Dawa, Ethiopia: change in rainfall intensity or human
Berhanu B, Melesse AM, Seleshi Y (2013) GIS-based hydrological impact?. Nat Hazards 75. doi:10.1007/s11069-014-1554-0
zones and soil geo-database of Ethiopia. Catena 104:21–31 Bojo J, Cassells D (1994) Land degradation and rehabilitation in
Berehe WA (1996) Twenty years of soil conservation in Ethiopia. A Ethiopia: a reassessment. The World Bank, Environmentally
personal overview. Regional Soil Conservation Unit/SIDA, Nairobi Sustainable Development Division, AFTES Working Paper
No. 17, Washington
Corti G, Manetti P (2012) Geologia e paesaggi della rift valley in McCann JC (1997) The plow and the forest: narratives of deforestation
Ethiopia. CNR—Pacini Editore, Pisa in Ethiopia, 1840–1992. Env Hist 2(2):138–159
Demlie M, Ayenew T, Wohnlich S (2007) Comprehensive hydrological Merla G, Abbate E, Azzaroli A, Bruni P, Canuti P, Fazzuoli M, Sagri
and hydrogeological study of topographically closed lakes in M, Tacconi P (1979) A geological map of Ethiopia and Somalia.
highland Ethiopia: the case of Hayq and Ardibo. J Hydr 339:145– CNR Italy, Firenze
158 Natali C, Beccaluva L, Bianchini G, Siena F (2013) The Axum-Adwa
Dessie G, Kleman J (2007) Pattern and magnitude of deforestation in basalt–trachyte complex: a late magmatic activity at the periphery of
the South central rift valley region of Ethiopia. Mount Res Dev 27 the Afar plume. Cont Min Petrol 166(2):351–370
(2):162–168 Nyssen J, Poesen J, Moeyersons J, Deckers J, Haile M, Lang A (2004)
Dessie G, Erkossa T (2011) Eucalyptus in East Africa. Socio-economic Human impact on the environment in the Ethiopian and Eritrean
and environmental issues. Working paper FP46/E, FAO, Rome highlands—a state of the art. Earth Sci Rev 64:273–320
Enkurie DL (2010) Adigrat sandstone in Northern and Central Nyssen J, Descheemaeker K, Haregeweyn N, Haile M, Deckers J,
Ethiopia: stratigraphy, facies, depositional environments and paly- Poesen J (2007) Lessons learnt from 10 years research on soil
nology. Ph.D. thesis, Technischen Universität Berlin erosion and soil and water conservation in Tigray. Tigray livelihood
FAO (2014) Country profiles. http://www.fao.org/countryprofiles/en/ papers no. 7, Mekelle Zala-Daget Project, Mekelle University, K.U.
FAO/IGADD/Italian Cooperation (1998) Soil and terrain database for Leuven, Relief Society of Tigray, Africa Museum and Tigray
northeastern Africa and crop production zones. Land and water Bureau of Agriculture and Rural Development
digital media series # 2. FAO, Rome Nyssen J, Frankl A, Haile M, Hurni H, Descheemaeker K, Crummey D,
FAO/IIASA/ISRIC/ISS-CAS/JRC (2009) Harmonized world soil data- Ritler A, Portner B, Nievergelt B, Moeyersons J, Munro N, Deckers
base (version 1.1), FAO, Rome, Italy and IIASA, Laxenburg, Austria J, Billi P, Poesen J (2014) Environmental conditions and human
Gebreegziabher Y (2004) Assessment of the water balance of Lake drivers for changes to north Ethiopian mountain landscapes over
Awassa catchment, Ethiopia. M.Sc. thesis, International Institute for 145 years. Sci Total Env 486–486:164–179
Geo-Information Science and Earth Observation, Enschede Osmaston HA, Mitchell WA, Osmaston NJA (2005) Quaternary
Getaneh W (2002) Geochemistry provenance and depositional tectonic glaciations of the Bale Mountains, Ethiopia. J Quat Sci 20(6):593–
setting of the Adigrat Sandstone northern Ethiopia. J Afr Earth Sci 606
35(2):185–198 Pankhurst R (1995) The history of deforestation and afforestation in
Grunder M (1988) Soil conservation research in Ethiopia. Mt Res Dev Ethiopia prior to World War I. Northeast Afr Stud 2(1):119–133
8(2–3):145–151 Poschen P (1986) An evaluation of the Acacia Albida-based agrofor-
Hagos M, Koeberl C, Kabeto K, Koller F (2010) Geochemical estry in the Hararghe highlands of Eastern Ethiopia. Agrofor Syst
characteristics of the alkaline basalts and the phonolite–trachyte 4:129–143
plugs of the Axum area, northern Ethiopia. Austrian J Earth Sci 103 Shiferaw A (2012) Estimating soil loss rates for soil conservation
(2):153–170 planning in Borena, South Wollo highlands. Lamberet Academic
Haregeweyn N, Poesen J, Nyssen J, Govers G, Verstraeten G, De Vente Publishing, Ethiopia
J, Deckers J, Moeyersons J, Haile M (2008) Sediment yield Soromessa T, Teketay D, Demissew S (2004) Ecological study of the
variability Northern Ethiopia: a quantitative analysis of its control- vegetation in Gamo Gofa zone, southern Ethiopia. Trop Ecol
ling factors. Catena 75:65–76 45:209–221
Herweg K, Ludi E (1999) The performance of selected soil and water Stocking MA (1996) Soil erosion. In: Adams WM, Goudie AS, Orme
conservation measures—case studies from Ethiopia and Eritrea. AR (eds) The physical geography of Africa. Oxford University
Catena 36(1–2):99–114 Press, Oxford, pp 326–341
Hurni H (1985) Erosion—productivity—conservation systems in Sutcliffe JP (1993) Economic assessment of land degradation in the
Ethiopia. In: Proceedings 4th international conference on soil Ethiopian highlands: a case study. National Conservation Strategy
conservation, Maracay, pp 654–674 Secretariat, Ministry of Planning and Economic Development,
Hurni H (1988) Degradation and conservation of the resources in the Addis Ababa, Ethiopia
Ethiopian highlands. Mt Res Dev 8(2/3):123–130 Tamene L, Vlek PL (2008) Soil erosion studies in northern Ethiopia. In:
Kappel R (1996) Economic analysis of soil conservation in Ethiopia: Brahimo AK, Vlek PL (eds) Land use and soil resources. Springer,
issues and research perspectives. University of Berne, Berne, The Netherlands, pp 73–100
Switzerland, in association with the Ministry of Agriculture, Addis Tefera MM (2011) Land-use/land-cover dynamics in Nonno district,
Ababa central Ethiopia. J Sust Dev Afr 13(1):123–141
Kelbessa E, Demissew S, Woldu Z, Edwards S (1992) Some threatened Tegene B (2000) Processes and causes of accelerated soil erosion on
Endemic plants of Ethiopia. In: Edwards S, Asfaw Z (eds) The cultivated fields of South Welo, Ethiopia. Eastern Afr Soc Sci Res
status of some plants in parts of tropical Africa. Botany 2000: East Rev 16(1):1–22
and Central Africa NAPRECA monograph series no. 2, Addis Tilahun A, Soromessa T, Kelbessa E, Dibaba A (2011) Floristic
Ababa, pp 35–55 composition and community analysis of Menagesha Amba Mariam
Kieffer B, Arndt N, Lapierre H, Bastien F, Bosch D, Pecher A, Yirgu forest (Egdu forest) in central Shewa, Ethiopia. Ethiop J Biol Sci 10
G, Ayelew D, Weis D, Jerram AD, Keller F, Meugniot C (2004) (2):111–136
Flood and shield basalts from Ethiopia: magmas from African White F (1965) The savanna-woodlands of the Zambezian and
superswell. J Petrol 45:793–834 Sudanian Domains—an ecological and phytogeographical compar-
Ibrahim KM (1978) Phytogeographical divisions of Africa. In: Hyder ison. Webbia 19:651–681
DM (ed) Proceedings of the lst rangelands congress, Denver Wright C, Adamseged Y (1986) An assessment of the causes, severity,
Juch D (1975) Geology of the South-Eastern escarpment of Ethiopia extent and probable consequences of degradation on the Ethiopian
between 39° and 42° long. East. In: Pilger A, Rosler A (eds) Afar highlands. FAO/EHRS (Ethiopian highlands reclamation study)
depression of Ethiopia. Schweizebart, Stuttgart, pp 310–316 working paper 3. FAO/EHRS, Addis Ababa
32 P. Billi
Woldu Z (1999) Forests in the vegetation types of Ethiopia and their Wood RB, Talling JF (1988) Chemical and algal relationships in a
status in the geographical context. In: Edwards S, Demissie A, salinity series of Ethiopian inland waters. Hydrobiologia 158:29–67
Bekele T, Haase G (eds) Forest genetic resources conservation: Zanettin B, Bellieni G, Justin Visentin E (2006) Stratigraphy and
principles, strategies and actions. Institute of Biodiversity Conser- evolution of the trachy-rhyolitic volcanism of the Senafe area
vation and GTZ, Addis Ababa, pp 1–38 (Eastern Eritrean Plateau). J Afr Earth Sci 45:478–488
Woldu Z, Yeshitela K (2003) Wetland plants in Ethiopia with examples
from Illubabor, south-western Ethiopia. In: Abebe Y, Geheb K (eds)
Wetlands of Ethiopia. IUCN Blue Series, Nairobi, pp 49–58
Geology of Ethiopia: A Review
and Geomorphological Perspectives 2
Ernesto Abbate, Piero Bruni, and Mario Sagri
Abstract
The Ethiopian region records about one billion years of geological history. The first event was
the closure of the Mozambique ocean between West and East Gondwana with the
development of the Ethiopian basement ranging in age from 880 to 550 Ma. This folded
and tilted Proterozoic basement underwent intense erosion, which lasted one hundred million
years, and destroyed any relief of the Precambrian orogen. Ordovician to Silurian fluviatile
sediments and Late Carboniferous to Early Permian glacial deposits were laid down above an
Early Paleozoic planation surface. The beginning of the breakup of Gondwana gave rise to the
Jurassic flooding of the Horn of Africa with a marine transgression from the Paleotethys and
the Indian/Madagascar nascent ocean. After this Jurassic transgression and deposition of
Cretaceous continental deposits, the Ethiopian region was an exposed land for a period of
about seventy million years during which a new important peneplanation surface developed.
Concomitant with the first phase of the rifting of the Afro/Arabian plate, a prolific outpouring
of the trap flood basalts took place predominantly during the Oligocene over a peneplained
land surface of modest elevation. In the northern Ethiopian plateau, huge Miocene shield
volcanoes were superimposed on the flood basalts. Following the end of the Oligocene, the
volcanism shifted toward the Afar depression, which was experiencing a progressive
stretching, and successively moved between the southern Ethiopian plateau and the Somali
plateau in correspondence with the formation of the Main Ethiopian Rift (MER). The
detachment of the Danakil block and Arabian subcontinent from the Nubian plate resulted in
steep marginal escarpments marked by flexure and elongated sedimentary basins. Additional
basins developed in the Afar depression and MER in connection with new phases of
stretching. Many of these basins have yielded human remains crucial for reconstructing the
first stages of human evolution. A full triple junction was achieved in the Early Pliocene when
the MER penetrated into the Afar region, where the Gulf of Aden and the Red Sea rifts were
already moving toward a connection via the volcanic ranges of northern Afar. The present-day
morphology of Ethiopia is linked to the formation of the Afar depression, MER, and Ethiopian
plateaus. These events are linked to the impingement of one or more mantle plumes under the
Afro-Arabian plate. The elevated topography of the Ethiopian plateaus is the result of profuse
volcanic accumulation and successive uplift. This new highland structure brought about a
reorganization of the East Africa river network and a drastic change in the atmospheric
circulation.
E. Abbate (&) P. Bruni M. Sagri

Dipartimento di Scienze della Terra, Università di Firenze, Via La
Pira 4, 50121 Florence, Italy
e-mail: abbate@unifi.it

34 E. Abbate et al.

Keywords
Paleogeography of Ethiopia Ethiopian volcanites Planation surfaces Ethiopian highlands
Main Ethiopian Rift Afar depression
2.1 Introduction result of geodynamic and geomorphic processes which have

shaped this territory since the Oligocene. These processes,
The spectacular landscape of the Ethiopian region (Fig. 2.1) which came relatively late in the ca. one-billion-year geo-
with the typical flat-topped mountains (Fig. 2.2) (ambas) and logical history of East Africa, were triggered by the
deep-incised valleys has fascinated the European travelers impingement of a mantle plume or plumes under the Afro-
since the sixteenth century when Alvarez (1540) visited the Arabian continental crust. The plume action gave rise to
fabulous land of Priest John. This fantastic scenario is the extrusion of huge amounts of magma, uplift, and
Fig. 2.1 Digital elevation map of

Ethiopia (SRTM data) with the
main physiographic elements
2 Geology of Ethiopia: A Review and Geomorphological Perspectives 35
Fig. 2.2 Flat-topped mountains

(amba) and deeply incised valleys
in the Tigray region (northern
Ethiopia) exposing Enticho
sandstones capped by trap basalts
fragmentation of the continental crust and contributed to the and Beccaluva et al. (2009). The newly collected data made
birth of the Red Sea, Gulf of Aden, East Africa Rift valley, possible the compilation of several geological maps at two
and the adjoining Afar depression. million scale very useful for regional syntheses (Mohr 1963;
In his evocative book, Mohr (2009) reports the main Kazmin 1973; Merla et al. 1973; Tefera et al. 1996).
stages in the geological exploration of East Africa in the Information on the sedimentary successions of the Afar
nineteenth and early twentieth centuries by outstanding and Danakil Alps regions was produced by Vinassa de
scholars, such as Rüppell (1834), Johnston (1844), Munz- Regny (1931), Bannert et al. (1970), Tiercelin et al. (1980),
inger (1864), Blanford (1870), Baldacci (1891), Dainelli and Kalb et al. (1982), and Bosworth et al. (2005).
Marinelli (1912). After these pioneers, Dainelli (1943), in his The geology of the Main Ethiopia Rift has been treated,
great synthesis on eastern Africa geology, relied on his own among others, by Mohr (1962), Di Paola (1972), Woldegabriel
investigations as well as contributions by Stefanini (1933) et al. (1990), Ebinger et al. (1993), Chorowicz et al. (1994),
and Merla and Minucci (1938). Dainelli’s three volumes Boccaletti et al. (1998), Le Turdu et al. (1999), Maguire et al.
close a research cycle based on a naturalistic approach to the (2006), and Peccerillo et al. (2007). Many of these contribu-
regional geology. After the World War II, researches tions have been reviewed and summarized by Corti (2009).
resumed with a scrupulous re-examination of the available Specific attention to the sedimentary filling of the Ethiopia Rift
geological data in the light of new investigations by Mohr has been given by Street (1979), Le Turdu et al. (1999), and
(1962) and geomorphological contributions by Abul-Haggag Benvenuti et al. (2002).
(1961) and Merla (1963, and following years) (Fig. 2.3). The possible economic importance of the Proterozoic
In the wake of the plate tectonics theory, eastern Africa terranes prompted regional reconnaissance and dedicated
has become an invaluable laboratory for understanding surveys (Dainelli 1943; Mohr 1962; Beyth 1972; Kazmin
passive margin processes, continental fragmentation, and the et al. 1978; Kröner 1985; Stern 1994; and Beyth et al. 1997).
first stages of oceanization. The Afar region, as a candidate The Paleozoic to Eocene sedimentary cover of the Protero-
for the formation of a new oceanic crust, has been studied zoic terranes summarized in the general syntheses by
with particular attention using updated analytical methods Dainelli (1943) and Mohr (1962) has been thoroughly
and facilitated field exploration (e.g., Barberi et al. 1970, investigated by Dow et al. (1971), Bosellini et al. (1997,
1972; Barberi and Varet 1977; Makris and Rhim 1991). The 2001), Hunegnaw et al. (1998), Kumpulainen et al. (2006),
highland volcanites were investigated from a petrographical, and Bussert and Schrank (2007).
geochemical, and geochronological point of view by The diverse morphology of the Ethiopian region with
Zanettin and Justin-Visentin (1974), Davidson and Rex extended plains dotted by intermittent ponds and lakes,
(1980), Mohr and Zanettin (1988), Woldegabriel et al. structural corridors along the rift valley surrounded by ele-
(1990), Ebinger et al. (1993), Hofmann et al. (1997), vated plateaus, and intense climate changes since the Plio-
Rochette et al. (1998), Pik et al. (1998), Kieffer et al. (2004), cene has favored human speciation and dispersal toward
36 E. Abbate et al.
Fig. 2.3 Morphological features of the Blue Nile area, southern Afar depression, Main Ethiopian Rift, and Somali plateau. Excerpt from the
“Major landform map of Ethiopia” in Merla et al. (1979), based on a morphological map in Merla (1963)
more suitable territories. The rift valley and Afar sediments In the following paragraphs, we will summarize and com-
contain abundant paleoanthropological evidences and lithic ment on the main sedimentary, volcanic, metamorphic, and
industries connected to these exceptional conditions during structural events recorded in the Ethiopian region. They are
the first stages of human evolution. After a few investigations related to the continental fragmentation of Gondwana and
carried out before 1960, a competitive and publicized search Afro-Arabian plates and connected basin development,
for hominid fossils has been carried out in the Omo region Paleozoic glaciations, and mantle plume activity. The present-
since the end of the 1960s (e.g., Chavaillon 1971; Arambourg day morphology results from their interaction (see chapter 2.7).
1972; Leakey 1974; Walker and Leakey 1978; Johanson and
Taieb 1976; Johanson et al. 1976; Lewin 1983), and in the
Awash region (e.g., Taieb 1974; Larson 1977; Clark 1985; 2.2 The Closure of the Mozambique
Kalb 1993; Tiercelin 1986; White and Johanson 1982; Walter Ocean and the Development
1994). Recent and outstanding results include the discovery of the Ethiopian Basement
of traces of possible human precursors (Ardipithecus kad-
abba, 5.5/5.8 Ma, Haile-Selassie et al. 2009) and ancestors A Neoproterozoic crystalline basement ranging from 880 to
(A. ramidus, 4.4 Ma, White et al. 2009) followed by various 550 Ma constitutes the crustal backbone of the Ethiopian region
Australopithecus and Homo genera (e.g., Haile-Selassie et al. with wide exposures in the southern and western Ethiopia and,
2007; Johanson and Taieb 1976; Asfaw et al. 2002). to a lesser extent, in the northernmost Ethiopia (Fig. 2.4).
Fig. 2.4 Simplified geological map of Ethiopia modified from Tefera et al. (1996)
The Proterozoic terranes in Ethiopia are related to as the

East African Orogen (Stern 1994), a N–S elongated mega-
collisional structure stretching from Israel to Madagascar and
produced between West and East Gondwana by the closure
of the Mozambique ocean (Fig. 2.5). The N–S alignments of
the East African Orogen lithic components, sometimes
marked by belts of ophiolites, stand out as pronounced
present-day morphological expressions. In addition, they
influence the trends of the succeeding fragile structures.
In the north, the East African Orogen constitutes the
Arabian–Nubian Shield, and in the south the Mozambique
Belt. In northern Ethiopia, the Nubian portion of the Shield
is prevalent, with dominantly low-grade volcano-sedimen- Fig. 2.5 The East Africa Orogen squeezed between West and East
tary rocks overlain by metasediments (stromatolitic carbon- Gondwana. WA West African craton; AM Amazonian craton; S Sahara
ates and diamictites) associated with “Snowball Earth” craton; C Congo craton; K Kalahari craton; M Madagascar; I India
(Beyth et al. 2003). In southern Ethiopia, the Mozambique shield; EA East Antarctic shield; AU Australia craton. Modified from
Meert and Lieberman (2008)
38 E. Abbate et al.
Belt exposes abundant amphibolites and granulite facies trending rift basins linked to the Central Africa shear zone in
metamorphic rocks and gneiss terranes. central Ethiopia (e.g., the Blue Nile rift) (Bosellini 1989;
The tightly folded and tilted Proterozoic basement Hunegnaw et al. 1998; Gani et al. 2009) (Fig. 2.7).
underwent intense denudation following the Early Paleozoic The Jurassic marine sedimentation was preceded by
(first planation surface “PS 1” in Fig. 2.6; for a thorough deposition of the continental Adigrat Sandstones (Fig. 2.8).
review of planation surfaces, see Coltorti et al. 2007). These were deposited above the partially peneplained Tri-
Remnants of less erodible rocks, such as granites and assic surface (PS 2 in Fig. 2.6) that developed at the expense
gneisses of various age, are now inselbergs in the gently of the Permo-Triassic sediments as well as of the basement.
rugged basement landscape, e.g. close to the Ethiopia/Sudan The latter was intersected by positive structures and deeply
border (Fig. 2.3). subsiding intracratonic basins which constituted the source
areas and accumulation sags for the Adigrat Sandstones,
respectively. The Adigrat Sandstones are widespread in
2.3 The Paleozoic Glaciation Ethiopia and with correlative units in the whole of East Africa
and Arabia. In Ethiopia, they commonly rest unconformably
The first sediments above the Early Paleozoic planation on the Paleozoic and basement rocks, but in the some basinal
surface are represented by a few patchy outcrops distributed contexts (e.g., in the Ogaden basin, Hunegnaw et al. 1998)
throughout Ethiopia (Bussert and Schrank 2007) (Fig. 2.6). are probably conformable with the Permo-Triassic Karoo
Detailed studies have been carried out close to the Eritrea/ deposits. The Adigrat Sandstones are light gray or red quartz
Ethiopia boundary in the Adigrat areas where greater arenites with interbeds of conglomerates and intensely pe-
thicknesses up to 500 m are exposed (Dow et al. 1971; dogenized red mudstones (Fig. 2.8). Their thickness is vari-
Kumpulainen et al. 2006; Bussert and Schrank 2007). able even at short distances and reaches 700 m. They were
According to Bussert and Schrank (2007), Ordovician to mainly deposited in fluvial or piedmont zones, but also in
Silurian fluviatile sandstones (lower Enticho Sandstone) rest fluviolacustrine and deltaic environments (e.g., Beauchamp
beneath Late Carboniferous to Early Permian glacial fluvio/ 1977; Bosellini et al. 1997, 2001; Wolela 2008).
lacustrine deposits (upper Enticho Sandstone, and Edaga The typical facies of the Jurassic transgression is repre-
Arbi Glacials). When in contact with the basement, the sented by shallow-water Callovian to Kimmeridgian car-
glacial activity is manifested by striations, roches mou- bonates, up to 1,000 m thick, referred to as the Antalo
tonnées, grooves, and chatter marks. This reconstruction, Limestones and the Hamanlei Formation (Figs. 2.7, 2.8 and
also supported by petrographic data (Sacchi et al. 2007), 2.9). They conformably overlie the Adigrat Sandstones
could supplant a former hypothesis of an Ordovician/Silu- (Blanford 1870), and their earliest occurrences (Pliensba-
rian glaciation (Dow et al. 1971). chian/Aalenian) are found in the Ogaden (Hunegnaw et al.
Late Paleozoic fluviatile sandstones are also reported in the 1998). In the Blue Nile basin, the Adigrat Sandstones are
Blue Nile gorge (Jepson and Athearn 1964; Russo et al. 1994) followed by the Gohatsion marls and evaporites of Aalenian
and near Harar. Many deep oil boreholes in the Ogaden basin to Callovian age (Russo et al. 1994). In the Dire Dawa-Harar
have intersected Late Paleozoic sediments of possible glacial and in the Tigray area, a younger widespread marine
to fluvial environment, unconformably resting on the base- flooding is recorded since the Bathonian and in the Callo-
ment and capped by Permian to Triassic continental clastic vian/Oxfordian, respectively. Within the carbonate succes-
sediments from lacustrine, deltaic, and fluvial environment. sion, characterized by several depositional sequences, the
The latter are considered to belong to the Karoo System maximum flooding surface is marked by organic-rich marls
(Hunegnaw et al. 1998). The glacial and fluvio-glacial and shales and is time-transgressive from Oxfordian (Oga-
deposits were located at the margin of the southern ice sheet den area) to Tithonian (Tigray) (e.g., Brassier and Guleta
of the Pangea (Martini et al. 2001). 1993; Hunegnaw et al. 1998; Bosellini et al. 1997, 2001). In
northern Ethiopia, the carbonate outcrops extend to the
Mekele area and the northwestern border of the Afar
2.4 The Jurassic Flooding of the Horn depression (Figs. 2.8 and 2.9). The westernmost outcrops
of Africa occur in central Ethiopia in the Blue Nile valley south of
Choke Mt., and it is likely that the marine transgression did
Along with the initial breakup of Gondwana, an Early not cross the 36°E meridian and the 16°N parallel.
Jurassic regional marine transgression invaded the Horn of The carbonate deposits from Tigray to the Dire Dawa/
Africa from the northeast (Paleotethys) and east (India/ Harar area are truncated by an erosional surface (PS 3 in
Madagascar nascent ocean). The sea covered the NS-trend- Fig. 2.6) due to an Early Cretaceous tectonic event marked
ing block-faulted structures connected to the Karoo rift by faulting and tilting (Bosellini et al. 2001). The abruptly
system in southern Ethiopia (Ogaden region), and the NW- overlying Amba Aradam Sandstones are fluviatile and
Fig. 2.6 Schematic stratigraphic chart of Ethiopia with the major planation surfaces [PS acronyms as in Coltorti et al. (2007)]
40 E. Abbate et al.
Fig. 2.7 Paleogeographic sketches for the Horn of Africa from Permian to Early Eocene
Fig. 2.8 The reddish Adigrat

Sandstones conformably covered
by the gray Antalo Limestone
along the escarpment east of
Mekele
Fig. 2.9 The Antalo Limestones

at the foot of the escarpment east
of Mekele: horizontal beds in the
foreground and an anticline in the
background
commonly associated with lenses of quartz conglomerates thick alternation of transitional, shallow-water marine and
and red shales. They often exhibit laterites at their base. open-sea deposits including shales, carbonates, sandstones,
Their maximum thickness is 200 m, and their age deter- and evaporites (e.g., Main Gypsum, Mustahil, Cotton)
mined on the base of Orbitolina findings in the Harar region (Barnes 1976; BEICIP 1985; Hunegnaw et al. 1998). This
is Aptian to Albian (Gortani 1973; Bosellini et al. 1999). succession records many sea-level oscillations and repre-
By contrast, in the Ogaden basin, marine deposition sents the final stages of the Mesozoic flooding in the Ogaden
continues during the Cretaceous until the Turonian or pos- basin. An erosional surface truncates the Mesozoic carbon-
sibly the Campanian (Fig. 2.7). It consists of ca. 1,200-m- ate/evaporite sequence from south to north down to the
42 E. Abbate et al.
Middle Jurassic Hamanlei Formation (Merla et al. 1973). The volcanic rocks that cover most of Ethiopia have been
The overlying sediments above this erosive surface are the subdivided into five major provinces on the basis of their
Jesomma Sandstones, a few-hundred-meter-thick quartzose lithological development, type of activity, frequency of
fluvial deposit with minor conglomerates and siltstones. The volcanic centers, and age of effusion (Abbate and Sagri
age of this unit is poorly constrained and can be assigned to 1980): (1) volcanites of the northern plateau; (2) volcanites
the upper portion of the Late Cretaceous. of the southern plateau; (3) volcanites of the Somali plateau;
A shallow-water marine deposition (Auradu Limestones, (4) Afar volcanites; and (5) Main Ethiopian Rift (MER)
ca. 400 m thick) resumes at the beginning of the Paleocene volcanites.
and was connected to a new transgression recorded only in The first three groups (Merla et al. 1979) comprise the
the easternmost portion of the Ogaden basin (Fig. 2.7). It major part of the Ethiopian volcanites (Fig. 2.10). They have
lasts until the Early Eocene, and its deposits were covered by collectively been referred to as “Traps” (Blanford 1870;
the Early to Middle Eocene Taleh Evaporites which reach Kazmin 1973), a general term from an old Swedish word
250 m in thickness. meaning stairs (Fig. 2.11). Fine-grained stratoid fissural
Paleogene basalts represent the greater portion of these
volcanites. The Afar and the MER volcanites, filling two
2.5 The Cenozoic Volcanic History: megastructures related to Neogene continental fragmenta-
Floods and Volcanoes tion, have a more limited extension (Fig. 2.10) and were
referred to as Aden Series (Blanford 1870; Mohr 1962).
Concomitant with the first phases of rifting in the Afro/
Arabian plate, a period of a prolific volcanic activity took
place predominantly during the Oligocene in the Horn of 2.5.1 The Volcanites of the Northern
Africa and southern Arabia which were at that time con- Ethiopian Plateau
nected. These volcanic rocks, mainly represented by basalts
and traditionally referred to as the trap succession, have After Blanford’s (1870) attempt to subdivide the northern
estimated to have covered an area in Ethiopia not less than plateau volcanites into a lower Ashangi Group unconform-
750,000 km2 before erosion, with a total volume of ca. ably overlain by a Magdala Group, a more detailed litho-
350,000 km3 (Mohr 1983) (Fig. 2.10). Their great areal stratigraphic approach was proposed in the 1970s by
extension and volume are due to the exceptional supply of Zanettin and Justin Visentin (1973) and Gregnanin and
mantle material connected with hot plumes (Schilling 1973; Piccirillo (1974), with the distinction of the Ashangi and
White and McKenzie 1989). Aiba basalts, Alaji Rhyolites, and Termaber Basalts. These
According to Hofmann et al. (1997), the basalt activity units were incorporated in the Merla et al. (1979) and Tefera
was concentrated in the very short time of one million years et al. (1996) maps. Some authors have pointed out that the
around 30 Ma, and this rapid outpouring of a huge volume lateral heterogeneity of these volcanic rocks, the vertical
of volcanic rocks has been regarded as a possible cause of recurrence of similar lithology, and the different morpho-
climatic deterioration and ensuing mass extinction on a logical response of basalts with the same petrological or
global scale (Courtillot et al. 1988; White and McKenzie chemical characteristics (e.g., Kieffer et al. 2004) prevent the
1989; Rochette et al. 1998). adoption of the criteria commonly used for sedimentary
The volcanic succession rests above a peneplained sur- bodies. However, in the expectancy of further detailed field
face (Blanford 1869; the pre-trappean peneplanation of and petrographic studies, we prefer to maintain the distinc-
Mohr 1962; PS 4 in Fig. 2.6), marked by laterites, particu- tions proposed by the previous authors and also followed by
larly well developed in Eritrea and Tigray (Dainelli and Rochette et al. (1998) in their magnetostratigraphy analyses
Marinelli 1912; Merla and Minucci 1938) and in south- of the northern Ethiopian traps.
western Ethiopia (Davidson and Rex 1980). The laterites are The Ashangi Basalts are composed of transitional to
also present in similar contexts in Yemen (Baker et al. 1996) tholeitic olivine basalts, often highly zeolitized, alternating
and western Arabia (Overstreet et al. 1997). This extensive with subordinate tuffs (Mohr and Zanettin 1988). The flows
pedogenesis is indicative of a long period of morphological are barely evident owing to their small thickness, reduced
stability of the peneplained surface marked by low elevation, horizontal extension, and deep weathering. Their thickness
little or no vertical deformations, and sediment starvation. is from 200 to 1,000 m. The Aiba Basalts consist of well-
The laterization was active until at least 40 Ma (Ar/Ar age, developed columnar massive transitional flood-basalt flows,
Andrews Deller 2006) and can be related to the Early locally with intervening agglomerate beds. The flows are
Eocene climatic optimum. 15–50 m thick reaching in some cases almost 100 m and
Fig. 2.10 The main volcanic

provinces of Ethiopia. After
Abbate et al. (2014)
can be traced over long distance. Their total thickness peralkaline rhyolites and transitional basalts. The acidic
reaches 1,000 m. The above-described fissural basalt units terms are mainly whitish ignimbritic tuffaceous layers that
constitute the bulk of the volcanic pile of the northern can be followed for a great distance. They form typical
Ethiopia plateau, locally attaining 2,000 m in thickness. landscape elements, such as steep walls and pyramids, well
This profuse outpouring took place between 31 and 29 Ma represented in the Amba Alaji peak (Fig. 2.12). Their
(Hofmann et al. 1997; Pik et al. 1998; Ukstins et al. 2002; maximum thickness is 500 m with a decrease to nil about
Coulié et al. 2003). The Amba Alaji Rhyolites, whose 100 km west of the Afar margin. Their age ranges from
outcrops are limited to the northern portion of the northern Late Oligocene in the northern outcrops to Early Miocene
Ethiopian plateau, are an alternation of alkaline to in the south (Zanettin et al. 1974).
44 E. Abbate et al.
Fig. 2.11 Felsic and subordinate

basic volcanites of the Semien
shield volcano overlie a thick
succession of flood basalts.
Semien National Park.
Photograph courtesy of
Frances Williams
Fig. 2.12 The outstanding

pyramidal morphology of the
Amba Alaji Rhyolites at the
Amba Alaji peak
A peculiar feature of the northern Ethiopian plateau is the maps, they are grouped under the name of Termaber Basalts.
frequent occurrence of shield volcanoes, about 30 major They are made of lenticular, often zeolitized, alkali basalts
centers according to Mohr and Wood (1976), some of which with a large amount of tuffs, scoriaceous lava flows, peral-
reach as much as 100 km in diameter and 1,000–2,000 m in kaline rhyolites, and typical red paleosoils. Dike swarms and
elevation above the plateau (e.g., Semien, Guna, Choke, and acidic extrusions are present. The thickness of the Termaber
Guguftu) (Figs. 2.13 and 2.14). In the regional reviews and Basalts reaches 1,000 m close to the volcanic centers.
Fig. 2.13 Location map of the

volcanic edifices and lakes cited
in text. Volcanic edifices: 1 Alid;
2 Dallol; 3 Erta Ale; 4 Dubbi; 5
Sork Ale; 6 Ado Ale; 7 Mussa
Ale; 8 Semien; 9 Guna; 10
Choke; 11 Guguftu; 12 Tullu
Wellel; 13 Yerer; 14 Boseti Guda;
15 Fantale; 16 Egan; 17 Bora
Bericho; 18 Chillalo; 19 Badda;
20 Gademotta; 21 Alutu; 22
Chike; 23 Kecha; 24 Bale; 25
Mizan Tefari; 26 Tosa Sucha.
Lakes: a Afrera; b Ashangi;
c Haik; d Ardibbo; e Ziway;
f Langano, Abiyata, Shala;
g Abaya; h Chamo; i Chew Bahir
The ages of the Termaber shield volcanoes are Early and trachyte (syenite) plugs and domes (Prior 1900; Merla and
Middle Miocene, ranging from 23 to 11 Ma (e.g., Kieffer Minucci 1938; Mohr 1962; Hagos et al. 2010) and have been
et al. 2004), with the exception of the Semien with the base dated as Mio-Pliocene by Beyth (1972), and more recently,
of 30 Ma and the top of 19 Ma. Natali et al. (2013) determined their Ar/Ar age of 19–15 Ma.
In addition to the trap basalts and shield volcanoes, some The rocks of this volcanic complex have been quarried for
subvolcanic intrusions are an outstanding feature in the the famous ancient Axumite obelisks.
Axum–Adwa area producing a landscape of cliffs, pinnacles, On the western portion of the northern Ethiopian plateau,
and domiform hills with an east–west alignment (Le Bas and the Tana rift basin is located in a structural complex area at the
Mohr 1968; Abbate and Sagri 1980). They are basalt– junction of three grabens (Chorowicz et al. 1999). Quaternary
46 E. Abbate et al.
Fig. 2.14 Structural sketch map

of the Ethiopian and Somali
plateaus, Afar depression, and
Main Ethiopian Rift (Northern,
Central, and Southern MER) with
the main transversal tectonic line
and major volcanic edifices.
Legend: 1 Miocene volcanic
centers of Termaber unit in the
northern Ethiopian plateau; 2
Pliocene acidic domes and plugs;
3 Plio-Quaternary volcanic
centers; 4 peralkaline plugs; 5
Red Sea axis; 6 Flexures; 7
Horizontal lava flows and
sedimentary sequences; 8 tilted
lava flows and sedimentary
sequences; 9 block faulting; 10
major graben; 11 transversal
tectonic lines. A–B, C–D cross
sections shown in Fig. 2.23.
Modified from Abbate and Sagri
(1980)
olivine–basalts and subordinate phonolites cover much of the their base with acid domes, plugs, and pyroclastic deposits in
southern portion of the Tana rift (Jepson and Athearn 1964) their evolved end members.
with well-preserved lavas, plugs, and spatter cones.
The transition to the southwestern plateau volcanites is
marked by a 700-km-long and 80-km-wide east–west- 2.5.2 The Volcanites of the Southern
trending volcano-tectonic alignment (Fig. 2.14) (Addis Ab- Ethiopian Plateau
aba-Nekempt line in Abbate and Sagri 1980; Yerer-Tullu
Wellel alignment, Abebe et al. 1998) consisting of a line of For the southern plateau volcanites, there are reports of
Late Miocene to recent volcanic centers gradually shifting in limited volumes of basalts (Amaro and Gamo basalts) with
age from Late Miocene in the west to Quaternary in the east ages between 45 and 35 Ma (Davidson and Rex 1980;
(Abebe et al. 1998). These edifices have basic lava flows at Ebinger et al. 1993; George et al. 1998). They predate the
Fig. 2.15 The Jima Volcanites at

the top of the Guraghe
escarpment (Fig. 2.13). Gently
inclined rhyolite ignimbrites
incised by a channel in turn filled
by horizontally laminated
ignimbrite flows. New road on the
western flank of the MER
between Butajira and Welkite
northern Ethiopian traps by as much as 15 Ma (Rogers 2006). interbedded, particularly in the upper portion, with acidic tuff
The southern plateau traps are less thicker than those of the and loose fluvial lacustrine deposits. Two samples gave ages
northern plateau and are characterized by much thicker and of 15 and 13 Ma.
more widespread siliceous rocks. For this volcanic sequence, A few huge rhyolite plugs and domes (e.g. Mt.Egan, Mt.
resting directly on the crystalline basement or, more rarely, Mizan Tafari, Fig. 2.13) constitute a prominent and peculiar
on the Eocene basalts and Mesozoic sandstones, we follow feature of the southern Ethiopian plateau. According to their
the distinctions proposed in Merla et al. (1979). The suc- relationships with the surrounding volcanites, they are
cession begins with a hundred meters of mildly alkaline doubtfully assigned to the Pliocene (Merla et al. 1979).
basalts (Omo Basalts), capped by a thick unit, up to 1,000 m,
of rhyolites, acidic tuffs, and subordinate basalts (Jima Vol-
canites). The Omo Basalts are commonly fine grained with 2.5.3 The Volcanites of the Somali Plateau
columnar flows up to 10 m thick alternating with minor tuffs
and red paleosols. There are some sparse dates ranging Above a patch of limited extent of Jima Volcanites, the
around 30 Ma (Merla et al. 1979; Davidson and Rex 1980). succession in the Somali plateau, assigned to the Traps,
The Jima Volcanites, reaching one thousand meters of consists of the Arussi and Bale Basalts of Miocene age (K/Ar
thickness in the Omo valley, represent most of the effusives ages ranging from ca. 24 to 9 Ma, Kunz et al. 1975;
in SW Ethiopia with a wide fringe in the southern portion of Morbidelli et al. 1975; Merla et al. 1979) with variable
the Somali plateau NW of the Amaro Mts. and south of the composition from transitional to alkaline basalts (Zanettin
west–east Bonga–Goba line (Abbate and Sagri 1980). The et al. 1980) and a thickness of about 3,000 m (Juch 1975).
Jima Volcanites are mainly composed of massive, white, Rhyolitic intercalations are particularly abundant and thick in
pinkish, and gray rhyolites, comendites, and pantellerites in the middle portion of the Arussi and Bale Basalts (Juch
thick flows alternating with tuffs and subordinate basalts 1975). They become the predominant component of the
(Fig. 2.15). Reliable radiometric datings are quite scarce. At overlying Ghinir Unit alternating with basalts and fluviola-
the Omo village, toward the top of the succession, the Jima custrine intercalations. From the data supplied by Juch
Volcanites gave an age of 27 Ma (Merla et al. 1979). An (1975), it can be inferred that the Ghinir Unit attains several
overall age range from ca. 30 to 27 Ma can be assigned to this hundred meters in thickness. The radiometric ages fall in the
unit. The Wollega Basalts, resting on the basement and on the range of 6–2 Ma (Kunz et al. 1975; Morbidelli et al. 1975;
tilted Omo Basalts and Jima Volcanites, consist of 200– Merla et al. 1979). A lacustrine unit with abundant diatomites
400 m of predominant columnar alkaline basalt flows and rare sands and conglomerates (Chorora Formation,
48 E. Abbate et al.
Sickenberg and Schoenfeld 1975) is discontinuously present 1,000 km long (Fig. 2.14). It separates the southern Ethio-
beneath the Ghinir Unit at the foot of the escarpment. It is up pian plateau to the west from the Somali plateau to the east.
to 200 m thick and has yielded Late Miocene mammals Northward, the MER progressively widens out into the
(Bernor et al. 2004). complex Afar triple junction, while at its southern end, a
A peculiar morphological feature in the Somali plateau is 200–300-km tectonically disturbed area (Gofa basin and
the huge Plio-Quaternary volcanic complex of the Bale Mts. range, Baker et al. 1972) marks the transition to the Kenyan
with cones and plugs, probably connected with the Bonga– Gregory Rift in the Turkana depression.
Goba line (Figs. 2.10 and 2.14). It reaches an elevation of The volcanic history of the MER has been dealt with in
4,300 m (Mt. Batu, Quaternary, Merla et al. 1979), rests on numerous papers which often take into account limited
the Arussi and Bale Basalts, and is covered by patches of sectors of the rift. This has resulted in a proliferation of
glacial deposits. volcanic units with significant problems of correlation
among the northern, central, and southern sectors.
The MER volcanic stratigraphy was summarized by Corti
2.5.4 The Afar Volcanites (2009). He envisages a lower basalt unit with trachybasalts
and subordinate silicic flows from 11 to 8 Ma old followed
The Afar region is a quasi-triangularly shaped depressed by a widespread ignimbrite cover (e.g., Nazaret Group)
area at the intersection of the Red Sea, Gulf of Aden, and the ranging in age from 7 Ma in the northern sector to 2 Ma to
Main Ethiopian Rift (MER) (Fig. 2.1). Due to its 25-million- the south and up to 700 m thick. Most of the ignimbrite
year-long story of rifting and incipient oceanization, volca- layers are believed to have formed by catastrophic eruptions
nic rocks cover wide areas of the Afar depression (Fig. 2.10). related to the collapse of large calderas, such as the 3.5-Ma-
According to Barberi et al. (1975), the Afar volcanites old Munesa caldera now buried beneath the Ziway–Shala
can be assigned to a first stage of continental rifting which lakes (Fig. 2.13). These two units, common to the whole
lasted about twenty million years, starting from 25 Ma, and a MER, are followed by Late Pliocene basalts with pyroclas-
later stage which commenced 4 Ma during which the oce- tics fed by calderas which are limited to the northern and
anic floor in the central portion of Afar began to develop. central sectors. The subsequent Quaternary volcanic unit,
The older volcanites of the first stage include the Adolei which outcrops throughout the MER, is the Wonji Group
Basalts, Mabla Rhyolites, and Dalha Basalts (Barberi al. associated with the oblique Wonji fault belt (Mohr 1962). It
1975). They are more than 1,000 m thick and cover a time includes basalt flows and scoria cones, and large silicic
range from 26 to 6 Ma. Associated with them are alkaline and central volcanoes with calderas. These edifices and calderas
peralkaline granites aged 25–22 Ma linked to an early phase rise up to 700 m above the plain (e.g., Bora Bericho, Alutu,
of continental breakup. The most extensive volcanic sequence Gademotta, Fig. 2.13), and some of them experienced
connected with the second stage is the Plio-Pleistocene Afar phreatomagmatic activity and historical flows. They are
Stratoid Series which covers about two-thirds of the Afar referred to as en-echelon arranged magmatic segments
depression. This consists of transitional basalts, about connected to the Wonji fault belt (see later).
1,500 m thick and lies unconformably on the Dalha Basalts Off-axis magmatism is mainly concentrated on the Somali
after a phase of magmatic quiescence. Intercalated in the top plateau with huge shield volcanoes of basaltic and trachytic
of and above the Afar Stratoid Series are the transversal composition and Mio-Pliocene age (Chillalo, Badda, Chike,
volcanics and marginal rhyolitic centers (e.g., Dubbi, Ado Kecha, Figs. 2.13 and 2.16). Some of them exceed 4,000 m
Ale, Fig. 2.13). The axial volcanic ranges of Quaternary age in elevation and have a base of 30–40 km diameter rising
are a typical morphological feature of the Afar depression from the plateau level for 1,000–1,500 m. Glacial cirques
from which they rise prominently up to 1,500 m. The northern and massive moraines occur at an altitude of about 4,000 m,
range (Erta Ale) parallels the Afar axis with a NNW trend; to indicating that these high mountains were glaciated during
the south, the volcanic ranges shift gradually to WNW. They the Late Quaternary (Grove et al. 1975).
consist of fissure eruptions and shield volcanoes with basaltic
flows and alkaline to peralkaline silicic rocks. Many of them
have been active in historical times, and the Erta Ale volcano 2.5.6 The Intertrappean Beds
exhibits a spectacular lava lake.
Common intercalations in the previously described trap pile
of the Ethiopian highlands as well as in the Yemen plateau
2.5.5 The Main Ethiopian Rift Volcanites are terrestrial sediments (intertrappean beds) composed of
red clays, sands, diatomites, and lignite seams, generally a
The Main Ethiopian Rift (MER) is a NNE–SSW to N–S- few tens of meters thick, but also reaching hundreds of
trending trough 80 km wide in its central portion and meters (Abbate et al. 2014). Within the abrupt and steep trap
Fig. 2.16 The Main Ethiopian

Rift floor with a Quaternary
volcano (hill on the left) and the
faulted escarpment of the Somali
plateau (light- and dark-colored
stripes in the midground) overlain
by the huge Chillalo shield
volcano (in the background)
Fig. 2.17 The flat morphology

of the soft intertrappean
sediments (S) contrasting with the
steep slopes cut through the trap
basalts (B1, B2), the Cretaceous
sandstones (KS), and the Jurassic
limestones (JL). Slope of the
Jema river, a Blue Nile left
tributary, 100 km NNW of Addis
Ababa
escarpment, these loose sediments give rise to a pronounced The volcanic activity resumed after two/three million years
morphological break (Fig. 2.17). The intertrappean sedi- but was delayed a couple of million years in the marginal
ments occur in lenses which are sometimes continuous for areas of the trap effusion (Abbate et al. 2014). The presence
tens of km in the middle/upper portion of the traps and span of endemic proboscideans in the intertrappean sediments is
a time interval between ca. 29 and 27 Ma (Early/Late Oli- particularly significant in investigating the origin of these
gocene transition). They mark a period of volcanic quies- mammals and in clarifying the relationships between African
cence following a voluminous and rapid lava outpouring. and Eurasian faunas (Abbate et al. 2012, 2014).
50 E. Abbate et al.
2.6 The Sediments in the Afar escarpment along the Middle Awash valley (e.g., Woramso/
Depression, Main Ethiopian Rift, Mille, Hadar, Kesem/Kebena, and Chorora sub-basins). Kalb
and Adjoining Areas et al. (1982) include these continental sediments under the
Late Miocene to Pleistocene Awash Group which is more
In addition to magmatic and pyroclastic materials, the sub- than 1,000 m thick. Main lithological components are shales,
siding basins of the Afar and MER host sediments which are sands, pebbles, and diatomites with intercalated ash layers
particularly useful in reconstructing Plio-Pleistocene human and basalt flows (Fig. 2.21). The Awash Group is famed for
evolution and climatic variations as well as the paleogeog- its rich hominid content, artifacts, and mammal faunas (e.g.,
raphy and structural development of the region. Hadar, Bodo, Buri, Busidima, Daka, Sidi Koma, Sagantole,
Chorora, Adu Asa) (a rich bibliography on this matter can be
acquired from Quade and Wynn 2008).
2.6.1 The Afar Depression In central and eastern Afar, elongated graben structures
(e.g., Abhe/Gob Ad, Hanle/Dobi, Tendaho/Dubti, Gaggade,
Different types of basins occur within the Afar depression and Figs. 2.14 and 2.18) filled by Pleistocene fluviolacustrine
along its margins. They originated during successive phases deposits are related to the axial basins (Tiercelin et al. 1980). In
of tectonic deformation since the Miocene and have generated geothermal drillings in the Tendaho basin, more than 500 m of
a significant morphological response in the present-day fine-grained Pleistocene sediments have been found alternat-
topography. According to their position with respect to the ing with basalts (Battistelli et al. 2002).
Afar depression axis, we distinguish axial, peripheral, and The eastern margin of the Eritrean–Ethiopian plateaus is
marginal basins (Fig. 2.18). The first two types are located characterized by a flexure that connects the plateau to the
within the Afar depression, and the latter type characterizes Afar depression (see later). Such a deformation belt with a
the lower and upper portions of the western escarpment. This maximum width of 100 km is developed along the escarp-
distinction holds for most of the Afar, but the occurrence of ment between Asmara and Addis Ababa for 900 km
latest Oligocene sediments in the northern apex of the Afar (Fig. 2.14). Connected to this flexure are a number of tec-
triangle north of 13° Lat. N is related to an older incipient tonic depressions parallel to the escarpment and mostly
rifting with the development of a sedimentary basin which is developed along a north–south direction (“marginal basins”
not recorded southward. These older sediments, which were Mohr 1967) (Fig. 2.18). The flexure and associated depres-
the first laid down in the Afar Depression, are the Red Series sions form very conspicuous features due to their morpho-
(Bannert et al. 1970) or Danakil Formation (Brinkmann and logical continuity and pronounced expression.
Kürsten 1970; Garland 1980). They are composed of violet- The depressions are filled by Pleistocene fluviolacustrine
red to bright-red conglomerates and sands with mudstones sediments and volcanoclastics. Those situated close to the
(Fig. 2.19) locally gypsiferous, and rare freshwater gastro- plateau edge with a master fault on their western flank are
pods-bearing limestones. Alluvial fans and high-energy well developed both longitudinally and transversally (Kobo
streams with some swampy to lacustrine ponds were the main basin: 120 km by 15 km; Haik/Borkenna basin 70 km by
features of the Danakil Formation environment. Frequent 10 km). The sediment thickness in these basins is generally a
basalt flows are found intercalated in this succession. Those at few tens of meters. Those developed within the flexure are
the base and toward the top gave K/Ar ages of 24 and 4 Ma, usually smaller and characterized by their reduced or absent
respectively (Bannert et al. 1970). A thickness of about sedimentary infilling. On the contrary, basins at the base of
1,000 m is commonly assumed. The whitish gypsiferous flexure host thick sedimentary successions. From north to
Enkafala Formation, a few meters thick, unconformably south, they are Buia in Eritrea, and Garsat and Teru in
overlies the Danakil Formation and marks a marine trans- Ethiopia. These basins are filled by up to 1,000 m of the
gression in the northern Afar depression. Its marine fossils fluviolacustrine sediments of the Early-to-Middle Pleisto-
have been dated by U/Th methods at 200 to 24 ky (Lalou et al. cene Dandiero Group (Abbate et al. 2004). The Buia section
1970). Toward the center of the basin, the Enkafala Formation yielded a one-million-year old Homo erectus skull, abundant
passes transitionally into the salt formation composed of mammal faunas, and lithic industries (Abbate et al. 1998).
halite, gypsum, potash salts, and clays (Fig. 2.20). The salts
have precipitated at the surface and generate an impressive,
bright-white salt plain. A thickness of 975 m has been drilled 2.6.2 The Main Ethiopian Rift
for potash exploitation, and a thickness of 2,200 m has been
estimated by geophysical investigations. An additional area of Late Cenozoic sediment accumulation
Peripheral and axial basins occur south of 13° Lat. N is the MER where fluviolacustrine sediments with rare diat-
(Figs. 2.14 and 2.18). Fluviolacustrine sediments fill omites (Fig. 2.22) cover a large area. They represent the
peripheral basins at the foot of the western and southern Afar deposits laid down in a very wide lake which, in the past,

major Plio-Quaternary
sedimentary basins (marginal,
peripheral, central) according to
their structural setting in the
northern Ethiopia plateau
escarpment, Afar depression, and
Main Ethiopian Rift. Marginal
basins: 1 Buia; 2 Garsat; 3 Teru; 4
Kobo; 5 Hayk; 6 Borkenna.
Peripheral basins: 7 Woramso; 8
Mille; 9 Kesem-Kebena; 10
Chorora. Central basins—in Afar:
11 Abhe-Goba Ad; 12 Tendaho-
Dubti; 13 Hanle-Dobi; 14
Gaggade; in the Main Ethiopian
Rift: 15 Galana; 16 Konso-
Gardula; 17 Chew Bahir; 18
Usno; 19 Omo; 20 Kibbish; 21
Turkana. Inset map A (top left)
shows location of study area
relative to the Nubian, Arabia,
and Somali plates. Main
Ethiopian Rift within dotted lines
occupied most of the rift floor. In the northern and central part Poorly studied similar fluviolacustrine deposits occur in the
of the MER north of the Lake Shala, only a few tens of meters southern MER in the Abaya/Chamo Lake regions with a thick-
of Late Pleistocene/Holocene sediments outcrop. They ness of more than 500 m (Ebinger et al. 1993). More detailed
record at least two major phases of lake expansion separated studies have been carried out on the fluviolacustrine sediments
by a prolonged period of lowstand and erosion (Fig. 2.22) containing Australopithecus boisei, Homo erectus, and abundant
(Street 1979; Le Turdu et al. 1999; Benvenuti et al. 2002). On artifacts in the Konso/Gardula area (Katho et al. 2000). These
the basis of geophysical investigations, the presence of about sediments comprise the Konso Formation and consist of dark-
600 m of sediments has been estimated by Le Turdu et al. gray clay, red-brown silt, sand, and gravel, with frequent inter-
(1999). It has been proposed that their maximum age is as old calations of volcanic rocks and tephra. The Konso Formation is
as 500 ka (Le Turdu et al. 1999). more than 200 m thick and was deposited between 1.9 and 1.4 Ma.
52 E. Abbate et al.
Fig. 2.19 Red fluviatile

sandstones and mudstones of the
Danakil Formation with a dark
intercalation of a basalt flow,
unconformably capped by coarse-
grained Quaternary alluvial
deposits
Fig. 2.20 Highly dissected salt

crust hosting iron-rich brine
deposits near Dallol, northern
Afar depression. The mountains
in the background are the Afar
western escarpment at the edge of
the northern Ethiopian plateau
The MER terminates to the south with the Chamo and graben, all the others are half-graben with the master fault on
Konso/Gardula basins and abuts the Sagan line (Baker et al. the western side (Davidson 1983). All these basins are filled by
1972; Abbate and Sagri 1980). Through the intermediate fluviolacustrine deposits interbedded with volcanic rocks and
Chew Bahir, Usno, Omo, and Kibish graben structures, the ashes. The Omo basin, in the lower course of the Omo river, is
deformation shifts to the Kenyan Turkana rift (Fig. 2.14). With the northern extension of the Turkana rift and is particularly
the exception of the Chew Bahir, which is a symmetrical well studied owing to the occurrence of abundant vertebrate
Fig. 2.21 Cyclically arranged

lacustrine, lake margin, and
fluviatile deposits with tuff
horizons in the Hadar basin,
Middle Awash, northern Afar
(basin 8 in Fig. 2.18). This section
is adjacent to the site where the
3.2-million-year-old partial
skeleton of “Lucy” was
discovered
Fig. 2.22 Late Quaternary

lacustrine sediments consisting of
massive diatomites (whitish) and
thinly stratified, light gray
volcanoclastic deposits. A deep
erosional surface separates the
Holocene from the Late
Pleistocene successions. Central
Main Ethiopian Rift. Photograph
courtesy of Marco Benvenuti
fossils including hominids (e.g., Howell 1969; Arambourg Pleistocene Shungura formation (deHeinzelin 1983). The
1972). It is filled by one thousand meters of sediments, and development of this basin dates back to the Early Pliocene
most of them are represented by the fossiliferous Plio- (Kidane et al. 2007; McDougall and Brown 2008).
54 E. Abbate et al.
2.7 The Morphological Responses incision rate over the entire plateau, was associated with a
to the Cenozoic Geodynamic broad and regional uplift of the plateau after the impingement
Events of the Afar plume at ca. 30 Ma. The second event resulted in
the localized increase of the incision rate controlled by the
Three main physiographic provinces characterize the Ethi- buildup of the shield volcanoes at ca. 22 Ma along the Tekeze
opian region: the highlands, represented by the northern and and Blue Nile watercourses. A particularly significant
southern Ethiopian plateaus and Somali plateau, the Afar increase in the incision rate took place at ca. 11 Ma, but was
depression, and the MER (Fig. 2.1). The birth, evolution, limited to the eastern portions of the Tekeze and Blue Nile
and present-day morphology of these provinces, although catchments. It was the result of the rapid uplift of the eastern
not achieved at the same time, are linked to the uplift and margin of the plateau facing the Afar depression.
rifting of the Afro-Arabian plate since the Oligocene. Paleobotanical data can also be used to decipher the
complex history of uplift of the Ethiopian plateau. Near Lake
Tana, 130 m of fluviolacustrine sediments are interbedded
2.7.1 The Ethiopian Plateaus and their with lavas in the middle/upper portion of the traps. They are
Uplift the famous Chilga intertrappean beds renowned for their
vertebrate and flora content (among others, Unger 1866;
Using observations from along the Gulf of Aden, Dainelli Merla and Minucci 1938; Yemane et al. 1987; Kappelman
(1943) assumed an Eocene age for the main uplift of the et al. 2003; Currano et al. 2011; Abbate et al. 2014). Their
whole Horn of Africa and claimed that has occurred prior to Ar/Ar age of 27.4 Ma has been obtained from an intercalated
the trap outpouring and rifting episodes. He was also the first ash layer (Kappelman et al. 2003). The Chilga beds with their
to visualize the uplift by drawing contours of the present Guineo-Congolean wet forest trees and palynoflora devoid of
surface of the crystalline basement. gymnosperms were deposited at an altitude much lower than
The issue of the exact time of uplift is still strongly debated. the 1,950 m of their present-day elevation (Yemane
Mohr (1967) and Mohr and Zanettin (1988) assumed that the et al.1987). On the basis of this broad indication, we tenta-
major uplift of the Afro-Arabian swell occurred during the tively assume that the Chilga beds were deposited at an alti-
Pleistocene. Alternatively, Merla et al. (1979) proposed that tude not higher than 900–1,000 m. Thus, the difference from
the domal uplift started in the Oligocene with a climax at the the present-day elevation would be ca. 1,000 m and has to be
beginning of the Miocene at about 25 Ma, inferring this timing ascribed to the uplift of this plateau segment after 27 Ma. The
of events from the stratigraphy of the Neogene formations in Chilga beds also provide constraints relevant to the height of
the Somali shoulder along the Gulf of Aden. the pre-trappean surface above the sea level. Since they lie
More recently, Pik et al. (2003) suggested on the basis of above ca. 600 m of trap basalts and were deposited, as we
(U-Th)/He thermochronometry that the Ethiopian plateau assumed, at ca. 1,000 m, we deduce a modest (ca. 400 m)
has been an elevated and stable dome since the Oligocene, elevation for the pre-trappean surface. This argues against a
with its highest region along the present-day Afar escarp- pronounced pre-trappean regional doming.
ment. The regional high structure was the result of the As to the age of the uplift, the Ethiopian plateau began to
combined effects of the Afar plume impingement and rise at ca. 20 Ma according to Moucha and Forte (2011) who
associated large basalt effusions (see also Ebinger and Sleep applied a numerical model of mantle flow to reconstruct the
1998). According to Pik and coworkers, steady-state erosion uplift amount and time in East Africa.
commenced in the Blue Nile canyon as early as 25–29 Ma For the southern Ethiopian and Somali plateaus, detailed
and is still active. About 20 My ago, the drift of the Arabian morphotectonic data are scanty apart from cursory mentions
plate and a concomitant collapse along the western Afar within general papers on the Horn of Africa.
margin gave rise to the Afar depression (Pik et al. 2003). The southern Ethiopian plateau reaches its highest ele-
The morphotectonic history of the northern Ethiopian vations along the eastern margin close to the MER (over
plateau has also been assessed by Gani et al. (2007) using 3,000 m in the Guraghe region) and progressively slopes
the long-term incision rate of the Blue Nile catchment. Their westward to the Sudan lowlands and southward to the
picture proposes that starting from a broad dome with slow Turkana depression. A low elevation peneplained pre-trap-
rate of uplift from 29 to 10 Ma (phase I), a rapid rate of pean surface, continuously covered by laterites, is suggested
increase in the uplift occurred at 10 Ma (phase II) followed by Davidson and Rex (1980). Conversely, a pre-trappean
by a dramatic plateau rise at 6 Ma (phase III). doming stage is tentatively put forward by Woldegabriel
An episodic succession of tectonic events is also accepted et al. (1990) on the basis of a tilted Mesozoic sequence
by Ismail and Abdelsalam (2012) on the basis of morpho- unconformably overlain by Oligocene basalts in the Kella
tectonic analyses of the Tekeze and Blue Nile drainage sys- horst (Guraghe region), but fission-track analyses in the
tems. The first event, characterized by a low to moderate same region indicate that updoming and concomitant
denudation of the Kella horst did not begin before the Late fluvial systems. In the Somali plateau, the consequence was
Miocene (Abebe et al. 2010). a pronounced dendritic drainage toward the Indian Ocean
To these conflicting hypotheses, we add that, as in the with canyons in the upper reaches of the Juba and Webi
case of the northern plateau (Chilga), palynological data are Shebeli rivers (Fig. 2.1).
also available for the southern plateau for paleoaltitude
assumptions. Samples collected by Wolela (2007) in the
Oligocene intertrappean sediments in the Jimma region at 2.7.2 The Afar Depression and Adjacent
altitudes between 1,700 and 2,200 m proved to lack gym- Plateau Margins
nosperms. As in the northern plateau, this absence points to
an altitude much lower than that of the present day and, The Afar depression is bounded to the west by the
consequently, to an uplift substantially after the accumula- N–S-trending, 700-km-long northern Ethiopian escarpment,
tion of the main trap effusion. to the south by the W–E trending, 350-km-long Somali
The Somali plateau, another sector of the Ethiopian high- plateau escarpment, and to the east by the NNW–SSE
lands, reaches elevations of more than 3,000 m along its trending, 700-km-long Danakil/Ali Sabieh block (Fig. 2.14).
margins facing the Afar depression and the MER. It gradually Its minimum elevation reaches 120 m below sea level in the
slopes down toward the southeast (Ogaden and Indian northern sector where the axial volcanic ranges of the
Ocean). The scarce available data on the time and amount of volcanic shield complexes dominate the landscape.
its uplift derive from these margins. Juch (1980), who has As previously discussed, the northern Ethiopian and
studied a large extent of the Somali plateau escarpment, Somali escarpments are characterized by a continuous flex-
proposes a major uplift of 1,500 m younger than 2–3 Ma. Near ure (Figs. 2.14 and 2.23) (Mohr 1962; Abbate and Sagri
the margin of the plateau, this uplift was accomplished along 1969, 1980; Justin-Visentin and Zanettin 1974; Zanettin and
large normal faults cutting previous flexure-like structures. Justin-Visentin 1975; Morton and Black 1975; Juch 1975;
In a wider regional analysis, many authors, beginning with Beyene and Abdelsalam 2005) with the exception of short
Pickford (1990), have pointed out that the uplift of the East segments north of 13°N and in the Harar-Dire Dawa area
Africa plateaus caused a drastic reorganization of atmospheric which show block tilting toward Afar (e.g., Abbate and Sagri
circulation. This induced strong hydroclimatic changes and 1980; Beyene and Abdelsalam 2005).
heavy impacts on ecosystems with a trend toward more arid The onset of the escarpment formation along the western
conditions. The first event of the East African aridification, margin of Afar varies in age from north to south (Zanettin and
marked by an expansion of savanna grassland replacing the Justin-Visentin 1975): latest Oligocene (24 Ma) in the
wet forests, occurred around 8–10 or 13.5 Ma (Sepulchre northern sector, and between 13 and 8 Ma southward. The foot
et al. 2006; Bonnefille 2010; Wichura et al. 2010). of the escarpment along the whole margin was affected during
From the discussion presented above, we conclude that the Plio-Pleistocene by major normal faults dipping east.
the uplift of East Africa occurred essentially during the The westernmost escarpment of the Somali plateau
Middle/Late Miocene after the trap accumulation. By the underwent a downflexing toward Afar during the Miocene
Early/Late Oligocene transition, these volcanic successions with important normal faults activity in the Pliocene (Juch
had produced a wide bulge, particularly pronounced in the 1975). In the remaining portion of the Somali plateau
eastern portion of the northern Ethiopian plateau. At that escarpment, near Dire Dawa, block faulting becomes pre-
time, the drainage was poorly defined due to the reorgani- dominant and involves mainly Mesozoic sediments and,
zation of the African river network (Goudie 2005; Stanki- subordinately, trap basalts of latest Oligocene to Miocene age.
evicz and De Witt 2006). Beginning in the Early Miocene, Horizontal Afar Stratoid basalts of Late Miocene to Pliocene
shield volcanoes were superimposed on the northern Ethio- age seal the block-faulted structures (Juch 1975; Ethiopian
pian highland and began to control the courses of the Blue Institute of Geological Survey, Dire Dawa sheet 1985).
Nile and Tekeze rivers with their circular bases (Fig. 2.1). Both the flexural structures and block faulting, involving
The courses of these rivers became annular and stable a deformed belt 50–100 km wide, produced rugged slopes at
through pronounced erosional action enhanced by the rise of the margins of the tabular Ethiopian highlands (Fig. 2.1).
the highland. Incised valleys developed in their upper The deformation is most pronounced parallel to the margins
reaches and have gradually evolved into the present-day with closely spaced faults and dikes. Morphological features
canyons to 1,600 m deep. Starting from the Miocene, in characteristic of the edge of the northern Ethiopian plateau
addition to the updoming, the Ethiopian region experienced and associated with the marginal graben are long river
the Afar rifting and, successively, the development of the reaches parallel to the margin (from north to south, Gabala in
MER. These tectonic events affected the plateau margins and the Garsat plain, Alomata in the Kobo plain, Borkenna in the
caused further uprise due to flexural deformations. These homonymous basin) locally giving rise to endorheic lakes
marginal uplifts have firmly established the present-day (Ashangi, Haik-Ardibbo) (Fig. 2.13).
56
Fig. 2.23 Cross sections across northern (A–B) and southern (C–D) Ethiopia; location in Fig. 2.14. Modified after Merla et al. (1979) and Abbate and Sagri (1980)
E. Abbate et al.
The eastern margin of the Afar depression lies against the Ethiopian and Somali escarpments in connection with
so-called Danakil block. This continental microplate stret- intense downfaulting, subsidence, and crustal stretching.
ches from the Gulf of Zula to the Gulf of Tadjoura and is The structural complexity of the Afar depression, partic-
interposed between the Afar depression and the southern ularly evident in the central and southern sectors, is the result
Red Sea. It is composed of Proterozoic basement covered by of the interactions of geodynamic events affecting the Afro-
Mesozoic sediments, trap basalts, and southward by Neo- Arabian plate since the Early Miocene. The structural history
gene volcanites including large edifices. The Danakil block began with the anticlockwise rotation of the Danakil
reaches an elevation of 1,300 m in its northern portion and microplate (Baker 1970; Burek 1970; Sichler 1980) which
more than 2,000 m to the south (Sork Ale and Mussa Ale caused it to detach from the Nubian plate with a later
volcanoes, Fig. 2.13). No recent structural data are available northward movement (Chorowicz et al. 1999; Beyene and
for this Afar margin facing the Danakil block. According to Abdelsalam 2005). This was also facilitated by the NE drift
the geological map by Brinckmann and Kürsten (1970) and of the Arabian plate.
in Bannert et al. (1970), flexures seem to affect most of it. By ca. 10 Ma, the Gulf of Aden propagator with its active
The youthful morphology of the Afar depression is sub- spreading axis progressively invaded the southern Afar region
stantially the result of Plio-Pleistocene events connected between the Somali plateau and the southern portion of the
with the thinning of the crust and the first oceanization Danakil block. Subsequently, the extension of the MER
episodes. The magmatic and tectonic processes produced propagated northward into the southern Afar depression giv-
wide areas covered by the flood basalts of the Afar Stratoid ing rise to a full-fledged triple junction (Wolfenden et al.
Series, imposing axial shield volcanoes, and, in the central 2004). Within the Afar east-central area, rotation of small
and southern sector mainly, an intricate network of fractures blocks resulted in curvilinear grabens cut into the Afar Stratoid
and faults (CNR-CNRS 1971, 1975) (Fig. 2.14). Series (e.g., Barberi and Varet 1977; Manighetti et al. 1998).
The morphotectonic features vary from north to south.
Over a distance of 200 km at the triangular apex of the Afar
depression, between the Gulf of Zula and Lake Afrera, the 2.7.3 Morphotectonic History of the Main
Erta Ale axial shield volcano dominates an inhospitable, but Ethiopian Rift
fascinating, landscape. This segment of the Afar depression
has a maximum width of ca. 100 km along 13° Lat. N and The MER is the salient morphotectonic feature that together
hosts at its margins the exclusive occurrences of the bright- with the southern Ethiopian and Somali highlands consti-
red Miocene Danakil Formation. At its extreme north, the tutes the spectacular scenery of the southern Ethiopia land-
depression merges into the 10-km-wide Gulf of Zula boun- scape. This depressed area hosts beautiful lakes, also famous
ded by Proterozoic basement rocks and terminates at the for their bird life. They have a tectonic or volcanic origin and
Alid axial volcanic range. are variously colored, from the blue of Shala, the green of
South of 13° Lat. N, the regional picture changes Bishoftu, the whitish of Abiyata, and the pink of Langano
abruptly, probably in connection with a W/E regional tec- and Abaya, to the pearl gray of Chamo. All these lakes form
tonic alignment. The morphotectonic scenario becomes endorheic systems connected by outlet–inlet fluvial reaches
particularly complex and dominated by a series of narrow or by groundwater infiltration (Street 1979). The southern
graben (often half-graben, e.g., Abbate et al. 1995; Acocella end of this endorheic system is located in the present-day
2010) a few km wide and some tens of km long. Their ephemeral Chew Bahir lake, described at the end of nine-
sedimentary filling has been referred to as the axial basins. teenth century as a wide stretch of permanent water pool.
According to Hayward and Ebinger (1996) and Beyene and During the Late Quaternary wet periods, Lakes Abaya and
Abdelsalam (2005), three divisions can be recognized in Chamo were connected to Lake Turkana through the Sagan
central and southern Afar (Fig. 2.14). The east-central area river and Lake Chew Bahir. In turn, Lake Turkana oversp-
is affected by a diffuse rifting, trending from NNW–SSE to illed via the Akobo–Sobat rivers into the White Nile river
W–E, and in many cases forming curvilinear structures (e,g., system (Butzer et al. 1972; Street 1979). This Mediterranean
Immino graben). This area is bounded to the SW and south connection is testified by the occurrence of the typical
by the Tendaho-Goba Ad rift, a major structure with a length Nilotic Nile perch (Tilapia) in Lakes Abaya and Chamo.
of ca. 300 km and a width up to 50 km (Acocella et al. The MER is confined between the uplifted shoulders of
2008). The southeastern area is characterized by a W–E fault the southern Ethiopian and Somali highlands (Figs. 2.1 and
system parallel to the Somali plateau escarpment. In the 2.23). The escarpments can be very steep with a difference in
southwestern area, the NNE–SSW fault trend of the MER, altitude of 1,000–1,500 m from the plateau margin to the rift
which propagates in this sector of Afar, is dominant. floor (e.g., Guraghe) and narrow (5–7 km) and are marked
Toward the west and south of this complex tectonic by major border normal faults. In other places (e.g., east of
region, peripheral basins developed at the base of the Langano), the shoulder is less steep and the difference is up
58 E. Abbate et al.
to 1,000 m. Also the width of the escarpments varies reaches some thousand meters. Above the pre-Cambrian
between 10 and 20 km and is characterized by a series of basement and Mesozoic sediments, the rift is floored by ca.
steeply dipping faults with individual vertical displacement 1,000 m of syn-rift Miocene to Recent volcanoclastic and
of 200–300 m, covered by a veneer of ignimbrites (Di Paola sedimentary deposits (Cornwell et al. 2010).
1972). This early phase was followed during the Pleistocene by a
From a morphological and geological point of view, the rift-in-rift stage, in which volcanic and tectonic activity was
MER has been subdivided into three main segments: the concentrated riftward with right-stepping en-echelon mag-
northern, central, and southern (Mohr 1983; Woldegabriel matic segments, the Wonji fault belt (Mohr 1962; Gibson
et al. 1990; Hayward and Ebinger 1996; Bonini et al. 2005) 1969; Kazmin et al. 1980; Boccaletti et al. 1992; Chorowicz
(Fig. 2.14). The northern MER funnels from the Afar et al. 1994; Ebinger and Casey 2001; Wolfenden et al. 2004;
depression, where it is about 100 km wide, to the 80-km- Bonini et al. 2005; Corti 2009). In this phase, the bordering
long Dubeta Col sill (north of Ziway lake). The central faults were no longer active and the deformation was magma
MER, which is 80 km wide, includes most of the lake region assisted with diffuse diking (Ebinger and Casey 2001;
and extends southward up to the W/E Goba–Bonga line Ebinger 2005). According to Bonini et al. (1997), Boccaletti
(Fig. 2.14). This portion of the MER has an average eleva- et al. (1998), and Wolfenden et al. (2004), the first and second
tion of 1,600 m, and the lowest altitude is at Lake Abiyata phases were related to rift orthogonal and oblique extension,
(1,580 m). At the Bonga–Goba line, the southern MER respectively. However, Corti (2008) showed that both phases
narrows up to 60 km, shifts to a N–S trend, and reaches an may have resulted from a constant post-10-Ma oblique rifting
elevation of 2,000 m, decreasing southward to 1,000 m. in line with early suggestions by Gibson (1969).
From its middle portion to the south, the southern MER There is a general consensus that rifting was diachronous
bifurcates into two branches (the Lake Chamo and Galana along the whole MER. Woldegabriel et al. (1990) recog-
river rifts) separated by the 3,000-m-high Amaro horst (“a nized a 18–15-Ma rifting initiation in the southern and
small Ruwenzori” according to Mohr 1967). The southern central MER with extension in the northern MER com-
MER keeps its morphological identity until the Sagan line. mencing after ca. 11 Ma (see also Ebinger and Casey 2001;
To south and west of this, a wide basin and range structure Wolfenden et al. 2004). Bonini et al. (2005) propose a dif-
connects the MER with the Kenyan rift (Fig. 2.14). ferent picture with the onset of rifting in the southern MER
The rift floor is not uniformly flat but is occupied by between 20 and 11 Ma connected to the deformation of the
recent volcanic edifices rising some hundreds of meters Kenyan rift. At that time, no major tectonic activity was
above the plain (e.g., Fantale, Boseti Gudda, Alutu, Tosa affecting the central and northern MER sectors. At the
Sucha) and calderas (e.g., Gademota, lake Shala O’a caldera) northern end of the MER, the southward rift propagation
(Fig. 2.13). Furthermore, the strongly deformed Wonji fault started from the Afar depression and progressively affected
belt (see later) is characterized by rough and irregular mor- the northern MER in the Late Miocene (11 Ma), the central
phology with narrow uplifted blocks, valleys, lava fields, MER in the Pliocene (5.6–3 Ma), and, eventually, joined the
spatter cones, and swampy depressions. southern MER in the Late Pliocene/Pleistocene (3–0 Ma).
The origin of the MER, a continental rift with an average For the central MER, the timing is consistent with the
crustal extension rate of about 2.5 mm/a (Wolfenden et al. thermochronology data provided by Abebe et al. (2010).
2004), or 6.0 mm/a according to Corti (2009), is still a matter
of debate. It has been related to pure tensional deformation
by McKenzie et al. (1970), Di Paola (1972), and Le Pichon 2.8 Conclusions
and Franchetau (1978). A sinistral shear component has been
postulated by Mohr (1968), Gibson (1969), Kazmin (1980), After the Neoproterozoic orogenic cycle, the cratonic history
and Boccaletti et al. (1992). According to Bonini et al. of the Ethiopian region was dominated by phases of vertical
(1997) and Boccaletti et al. (1998), a sinistral oblique rifting motion and rifting recorded by various erosional cycles.
related to an E–W extension followed a pure extension They are marked by regional planation surfaces (Fig. 2.6)
orthogonal to the rift trend, whereas Chorowicz et al. (1994) which constitute geomorphological features useful in deci-
maintain that a right-lateral component of motion along the phering the different stages of the geological history of the
rift structure produced a NW-to-NNW-oriented extension. Ethiopia (see also Coltorti et al. this volume).
The MER development was characterized by an early The oldest and particularly significant planation surface
phase (Mio-Pliocene, Corti 2008) of activity of a series of (PS 1) results from the intense erosion which destroyed any
large boundary fault-formed local asymmetric basins (e.g., relief of the Precambrian orogeny across the whole of East
Abbate and Sagri 1980; Kazmin et al. 1980; Woldegabriel Africa. Some occurrences of more resistant granites and
et al. 1990; Boccaletti et al. 1992; Corti 2009) (Fig. 2.23). gneisses escaped this peneplanation and presently stand as
The total vertical displacement across the boundary faults inselbergs along the Ethiopia/Sudan boundary (Fig. 2.3).
The PS 1 records a time gap of at least one hundred million lithospheric thinnings. A different reconstruction was pro-
years (between the emplacement of the post-orogenic vided by George et al. (1998) who envisage a Kenya plume
batholiths and the deposition of the Ordovician sediments), initiating at 45 Ma and a 15 Ma younger Afar plume. More
during which planation and lateritization took place. recent authors (e.g., Furman et al. 2006; Meshesha and
The Triassic planation surface (PS 2) is mainly limited to Shinjo 2008) reconsidered the African superplume hypoth-
northern Ethiopia where it intersects a few preserved patches esis of Ritsema et al. (1999) and assumed multiple plumes
of Paleozoic fluvial and glacial deposits. This surface marks stemming from this large-scale thermal upwelling.
a relatively short period of erosion and/or non-deposition, The impact of the plume (or plumes) was concomitant
and in some basins, it is lacking with a continuous succes- with the dismembering of the Nubia/Arabian continental
sion from Paleozoic to Jurassic (e.g., in the Ogaden). block. During the Oligo-Miocene transition, the Danakil
The Cretaceous planation surface (PS 3) cuts across the block began to separate from Nubia leaving behind the Afar
Middle-to-Late Jurassic marine sediments and is most evi- region, and after a few million years, the Somali plate began
dent in northern Ethiopia where it represents a key marker to separate from Nubia with the development of the MER.
(Coltorti et al. 2007) and is overlain by continental Creta- In addition to its salient morphological features, the
ceous (Aptian–Albian) sandstones. The development of this Ethiopian region records some of the most important events
surface spans a period of a few tens of million years. of the Earth history since the Neoproterozoic:
After the Late Jurassic marine regression and deposition (i) The aggregation of the East and West Gondwana
of Cretaceous continental deposits, the Ethiopian region was continental blocks to form the Gondwana super-
an exposed land affected by tectonic deformation and continent and the development of the East Africa
regional uplifts for a period of about 70 Ma. A planation Orogen.
surface (PS 4) developed before the trap outpouring, cutting (ii) The long plate stillness of the southern portion of
the previous sedimentary successions, in some cases, down Gondwana during the Paleozoic with the erosion of
to the basement. It is recognizable over all the Ethiopian the Late Proterozoic mountain chains and develop-
region and could correspond on a continental scale to the ment of a wide planation surface.
African Surface of Burke and Gunnel (2008). The topo- (iii) The major, widespread, and long-lasting glaciation
graphical elevation of PS4 at the time of the traps outpouring during the Carboniferous to Early Permian with
was modest, a few hundred meters above sea level, Ethiopia located at the margin of the southern
according to the information provided by the palynoflora hemisphere ice sheet of Pangea.
assemblages preserved in the intertrappean sediments of the (iv) The beginning of the breakup of Gondwana and the
northern plateau (Chilga, Tana lake area) as well as the related Jurassic sea-level highstand resulting in the
southern (Jima area) Ethiopian plateaus. This is good evi- East Africa marine ingression from the Paleotethys
dence against the hypothesis of a prominent pre-trappean and the India/Madagascar nascent ocean.
updoming affecting the whole Ethiopian region. (v) The profuse and rapid basaltic activity which con-
The Tana and Jima areas also provide constraints on the tributed to the Oligocene climatic deterioration and
amount and timing of their uplift. Indications from there ensuing mass extinctions at a global scale.
suggest that uplift reached a maximum of one thousand (vi) The uplift of the East Africa plateaus which resulted
meters and was achieved after the major volcanic outpour- in a drastic reorganization of atmosphere circulation
ing. Similar relationships between uplift and time were and river pattern during the Late Neogene.
probably common to both Ethiopian plateaus. The volcanic (vii) The development of a variegated landscape with
activity, particularly prolific around 30 My, and the uplift of well-defined rift valleys which prompted the birth,
the Ethiopian region were essential in producing a bulge evolution, and radiation of the human species later
which resulted in high elevated plateaus. This new topo- impelled to colonize more suitable territories in
graphic barrier induced the reorganization of the eastern nearby continents.
Africa river network with a shifting of the regional divide
Acknowledgments Constructive commentaries and careful editorial
eastward. Rivers, originally flowing toward the Indian
suggestions by Frances Williams (Adelaide University, Australia) greatly
Ocean, were gradually captured by the Mediterranean- improved the manuscript. The authors have benefited from helpful dis-
directed paleo-Nile system (Said 1993; Goudie 2005). cussions in the field with Abebe Tsegaye (MASSA Spin off, Pisa, Italy)
As to the role of a mantle plume in the morphogenesis of and Miruts Hagos (Mekele University, Ethiopia) and in the office with
Marco Bonini, Giacomo Corti, and Federico Sani (Department of Earth
the Ethiopian region, the local impingement of a rising
Sciences, Florence University, Italy). Special thanks to World Geomor-
mantle plume has been considered as a possibility since the phological Landscapes Series Editor Piotr Migoń (University of Wroc-
1970s by Morgan (1971). Ebinger and Sleep (1998) ław, Poland) for his precious editorial assistance. We are grateful to
proposed a model with a single large plume beneath the Volume Editor Paolo Billi (Ferrara University, Italy) who prompted our
contribution and gave valuable suggestions.
Ethiopian plateau with lateral flows exploiting preexisting
60 E. Abbate et al.
References Baldacci L (1891) Osservazioni fatte nella Colonia Eritrea e pubbl. a

cura del R.Ufficio Geologico con una carta. Tipografia Nazionale
Roma 110 pp
Abbate E, Sagri M (1969) Dati e considerazioni sul margine orientale Bannert D, Brinckmann J, Käding KCh, Kretsch G, Kürsten N,
dell’altopiano etiopico nelle province del Tigrai e del Wollo. Boll Mayrhofer H (1970) Zur Geologie der Danakil Senke (Nördliches
Soc Geol It 88:489–497 Afar Gebiet), NE Aethiopien. Geol Rund 59:409–443
Abbate E, Sagri M (1980) Volcanites of Ethiopian and Somali Plateaus Barberi F, Varet J (1977) Volcanism of Afar: small-scale plate tectonics
and major tectonic lines. Atti Convegni Acc Lincei Roma 47:219– implications. Geol Soc Am Bull 88:1251–1266
227 Barberi F, Borsi S, Ferrara G, Marinelli G, Varet J (1970) Relations
Abbate E, Passerini P, Zan L (1995) Strike-slip faults in a rift area: a between tectonics and magmatology in the northern Danakil
transect in the Afar Triangle, East Africa. Tectonoph 241:67–97 depression (Ethiopia). Philos Trans R Soc Lond A267:293–311
Abbate E, Albianelli A, Azzaroli A, Benvenuti M, Tesfamariam B, Barberi F, Tazieff H, Varet J (1972) Volcanism in the Afar depression:
Bruni P, Cipriani N, Clarke RJ, Ficcarelli G, Macchiarelli R, its tectonic and magmatic significance. Tectonoph 15:19–29
Napoleone G, Papini M, Rook L, Sagri M, Tecle TM, Torre D, Villa Barberi F, Ferrara G, Santacroce R, Varet J (1975) Structural evolution
I (1998) A one-million-year-old Homo cranium from the Danakil of the Afar triple junction. In: Pilger A, Roesler EA (eds) Afar
(Afar) depression in Eritrea. Nature 393:458–460 Depression of Ethiopia. Schweizerbart, Stuttgart, pp 38–54
Abbate E, Woldehaimanot B, Bruni P, Falorni P, Papini M, Sagri M, Barnes SU (1976) Geology and oil prospects of Somalia, East Africa.
Girmay S, Tecle TM (2004) Geology of the Homo-bearing Am Assoc Pet Geol Bull 60:389–413
Pleistocene Dandiero basin (Buia region, Eritrea Danakil Depres- Battistelli A, Yiheyisb A, Calore C, Ferragina C, Abatnehb W (2002)
sion). Riv It Paleont Strat 110:5–34 Reservoir engineering assessment of Dubti geothermal field,
Abbate E, Bruni P, Coppa A, Aria D, Ferretti MP, Libsekal Y, Rook L, Northern Tendaho Rift, Ethiopia. Geothermics 31:381–406
Sagri M (2012) A new Oligocene-Miocene mammal-bearing site Beauchamp J (1977) La série sédimentaire en Ethiopie central and
from a sedimentary intercalation in the Trap basalts of central oriental. Ph.D. thesis, University Claude Bernard Lyon, 419 pp
Eritrea. Riv It Paleont Strat 118:545–550 Beccaluva L, Bianchini G, Natali C, Siena F (2009) Continental flood
Abbate E, Bruni P, Ferretti MP, Delmer C, Laurenzi MA, Hagos M, basalts and mantle plumes: a case study of the Northern Ethiopian
Bedri O, Rook L, Sagri M, Libsekal Y (2014) The Oligocene Plateau. J Petrol 50:1377–1403
intertrappean beds: regional distribution, depositional environments BEICIP (1985) Geological map of Ogaden and surrounding area, scale
and Afro/Arabian mammal dispersal corridors. J Afr Earth Sci 1:1,000,000. Ministry of Mines and Energy Addis Ababa
99:463–489 Benvenuti M, Carnicelli S, Belluomini G, Dainelli N, Di Grazia S,
Abebe T, Mazzarini F, Innocenti F, Manetti P (1998) The Yerer-Tullu Ferrari GA, Iasio C, Sagri M, Ventra D, Atnafu B, Kebede S (2002)
Wellel volcanotectonic lineament: a transtensional structure in The Ziway-Shala lake basin (main Ethiopian rift, Ethiopia): a
central Ethiopia and the associated magmatic activity. J Afr Earth revision of basin evolution with special reference to the Late
Sci 26:135–150 Quaternary. J Afr Earth Sci 35:247–269
Abebe T, Balestrieri ML, Bigazzi G (2010) The Central Main Ethiopian Bernor RL, Kaiser TM, Nelson SV (2004) The oldest Ethiopian Hipparion
Rift is younger than 8 Ma: confirmation through apatite fission-track (Equinae, Perissodactyla) from Chorora: systematics, Paleodiet and
thermochronology. Terra Nova 22:470–475 Paleoclimate. Cour Forsch Inst Senckenberg 246:213–226
Abul-Haggag Y (1961) A contribution to the physiography of Northern Beyene A, Abdelsalam MG (2005) Tectonics of the Afar depression: a
Ethiopia. University of London Athlone Press, London, 153 pp review and synthesis. J Afr Earth Sci 41:41–59
Acocella V (2010) Coupling volcanism and tectonics along divergent Beyth M (1972) The geology of central and western Tigray. Ph.D.
plate boundaries: collapsed rifts from central Afar, Ethiopia. Geol thesis, Reinische Friedrick-Wilhelm University of Bonn, 200 pp
Soc Am Bull 122:1717–1728 Beyth M, Stern R, Matthews A (1997) Significance of high-grade
Acocella V, Abebe B, Korme T, Barberi F (2008) Structure of Tendaho metasediments from the Neoproterozoic basement of Eritrea.
Graben and Manda Hararo Rift: implications for the evolution of the Precambr Res 86:45–58
southern Red Sea propagator in Central Afar. Tectonics 27 Beyth M, Avigadb D, Wetzelc HU, Matthewsb A, Berhe SM (2003)
(TC4016). doi:10.1029/2007TC002236 Crustal exhumation and indications for Snowball Earth in the East
Alvarez F (1540) Verdadeira Informação das Terras do Preste João das African Orogen: north Ethiopia and east Eritrea. Precambr Res
Indias. 458 pp 123:187–201
Andrews Deller ME (2006) Facies discrimination in laterites using Blanford WT (1869) On the geology of a portion of Abyssinia. Q J
Landsat Thematic Mapper, ASTER and ALI data—examples from Geol Soc London 25:401–406
Eritrea and Arabia. Int J Remote Sens 27:2389–2409 Blanford WT (1870) Observations on the geology and zoology of
Arambourg G (1972) Expedition internationale de rechérche paléon- Abyssinia, made during the progress of the British expedition to that
tologique dans la vallèe de l’Omo, Ethiopie, en 1967. Proc Pan Afr country in 1867–68. Macmillan, London, 487 pp
Congr Prehist 6:135–140 Boccaletti M, Getaneh A, Tortorici L (1992) The main Ethiopian Rift:
Asfaw B, Gilbert WH, Beyene Y, Hart WK, Renne PR, WoldeGabriel an example of oblique rifting. Ann Tectonicae 6:20–25
G, Vrba ES, White TD (2002) Remains of Homo erectus from Boccaletti M, Bonini M, Mazzuoli R, Abebe B, Piccardi L, Tortorici L
Bouri, Middle Awash, Ethiopia. Nature 416:317–320 (1998) Quaternary oblique extensional tectonics in the Ethiopian
Baker BH (1970) The structural pattern of the Afro-Arabian rift system Rift (Horn of Africa). Tectonoph 287:97–116
in relation to plate tectonics. Phil Trans Roy Soc London Bonini M, Souriot T, Boccaletti M, Brun JP (1997) Successive
A267:383–391 orthogonal and oblique extension episodes in a rift zone: laboratory
Baker BH, Mohr PA, Williams LAJ (1972) Geology of the eastern rift experiments with application to the Ethiopian Rift. Tectonics
system of Africa. Geol Soc Am Spec Pap 136:67 pp 16:347–362
Baker J, Snee L, Menzies M (1996) A brief Oligocene period of flood Bonini M, Corti G, Innocenti F, Manetti P, Mazzarini F, Abebe T,
volcanism in Yemen: implications for the duration and rate of Pecskay Z (2005) Evolution of the main Ethiopian Rift in the frame
continental flood volcanism at the Afro-Arabian triple junction. of Afar and Kenya rifts propagation. Tectonics 24:TC1007. doi:10.
Earth Planet Sci Lett 138:39–55 1029/2004TC001680
Bonnefille R (2010) Cenozoic vegetation, climate changes and hominid Courtillot V, Féraud G, Maluski H, Vendamme D, Moreau MG, Besse
evolution in tropical Africa. Glob Planet Change 72:390–411 J (1988) Deccan flood basalts and the Cretaceous/Tertiary bound-
Bosellini A (1989) The continental margins of Somalia: their structural ary. Nature 333:843–846
evolution and sequence stratigraphy. Mem Sci Geol Padova Currano ED, Jacobs BF, Pan AD, Tabor NJ (2011) Inferring ecological
41:373–458 disturbance in the fossil record: a case study from the late Oligocene
Bosellini A, Russo A, Fantozzi PL, Assefa G, Solomon T (1997) The of Ethiopia. Palaeogeogr Palaeoclimatol Palaeoecol 309:242–252
Mesozoic succession in the Mekele outlier (Tigrai province, Dainelli G (1943) Geologia dell’Africa Orientale. (3 vols. text, 1 vol.
Ethiopia). Mem Sci Geol Padova 49:95–116 maps). Reale Accademia Italia, Roma, 1916 pp
Bosellini A, Russo A, Schroeder R (1999) Stratigraphic evidence for an Dainelli G, Marinelli O (1912) Risultati scientifici di un viaggio nella
Early Aptian sea-level fluctuation: the Graua Limestone of south- Colonia Eritrea. Pubblicazioni Regio Istituto di Studi Superiori
eastern Ethiopia. Creta Res 20:783–791 Pratici e di Perfezionamento Firenze, 601 pp
Bosellini A, Russo A, Assefa G (2001) The Mesozoic succession of Davidson A (1983) Reconnaissance geology and geochemistry of parts
Dire Dawa, Harar Province, Ethiopia. J Afr Earth Sci 32:403–417 of Illubabor Kefa, Gemu Gofa and Sidamo, Ethiopia. The Omo
Bosworth W, Huchon P, McClay K (2005) The Red Sea and Gulf of River Project, Ministry of Mines and Energy, Ethiopian Institute of
Aden Basins. J Afr Earth Sci 43:334–378 Geological Survey Bull, 289 pp
Brassier M, Geleta S (1993) A planktonic marker and Callovian- Davidson A, Rex CD (1980) Age of volcanism and rifting in
Oxfordian fragmentation of Gondwana: data from Ogaden basin, southwestern Ethiopia. Nature 283:657–658
Ethiopia. Palaeogeogr Palaeoclimatol Palaeoecol 104:177–184 deHeinzelin J (ed) (1983) The Omo Group. Koninklijk Muesum voor
Brinkmann J, Kürsten M (1970) Geological sketch map of the Danakil Midde-Afrika, Tervuren Belgie Annalen 85:365 pp
Depression, scale 1:250,000. Bunderanstalt für Bodenforschungen, Di Paola GM (1972) The Ethiopian Rift Valley (between 7° 00′ and
Hanover 8° 40′ lat.). Bull Volcanol 36:517–560
Burek PJ (1970) Paleomagnetic evidence for an anti-clockwise rotation Dow DB, Beyth M, Hailu T (1971) Paleozoic glacial rocks recently
of the Danakil Alps, Ethiopia. Eos Trans Am Geophys Un Abs discovered in northern Ethiopia. Geol Mag 108:53–60
51:271 Ebinger CJ (2005) Continental breakup: the East African perspective.
Burke K, Gunnell Y (2008) The African erosion surface: a continental- Astronomy and Geophysics 46:2.16–2.21
scale synthesis of geomorphology, tectonics, and environmental Ebinger CJ, Casey M (2001) Continental breakup in magmatic
change over the past 180 Million Years. Geol Soc Am Mem provinces: an Ethiopian example. Geology 29:527–530
201:66 pp Ebinger CJ, Sleep NH (1998) Cenozoic magmatism through east Africa
Bussert R, Schrank E (2007) Palynological evidence for a latest resulting from impact of a single plume. Nature 395:788–791
Carboniferous-Early Permian glaciations in Northern Ethiopia. J Afr Ebinger CJ, Yemane T, Woldegabriel G, Aronson JL, Walter RC
Earth Sci 49:201–210 (1993) Late Eocene-Recent volcanism and faulting in the southern
Butzer KW, Isaac GL, Richardson JL, Washbourn-Kamau C (1972) main Ethiopian rift. J Geol Soc London 150:99–108
Radiocarbon dating of East Africa lake levels. Science 175:1069– Ethiopian Institute of Geological Survey (1985) Geological map of the
1076 Dire Dawa Sheet, scale 1:250,000, NC 37-12 Ethiopian Govern-
Chavaillon J (1971) Les habitants acheuleens de Melka-Kontoure. Proc ment Addis Ababa
PanAfr Congr Prehist 230:17–25 Furman T, Bryce J, Rooney T, Hanan B, Yurgu G, Hayalew D (2006)
Chorowicz J, Collet B, Bonavia FF, Korme T (1994) Northwest to Heads and tails: 30 million years of the Afar plume. In Yurgu G,
north-northwest extension direction in the Ethiopian Rift deduced Ebinger CJ, Maguire PKH (eds) The Afar volcanic province within
from orientation of the extension structures and fault slip analysis. the East Africa rift system. Geological Society Special Publications
Geol Soc Am Bull 105:1560–1570 No 259, pp 95–119
Chorowicz J, Collet B, Bonavia FF, Korme T (1999) Left-lateral strike- Gani NDS, Gani MR, Abdelsalam MG (2007) Blue Nile incision on the
slip tectonics and gravity induced individualization of wide Ethiopian Plateau: pulsed plateau growth, Pliocene uplift, and
continental blocks in the western Afar margin. Eclogae Geol Helv hominin evolution. GSA Today 17:4–11
92:149–158 Gani NDS, Abdelsalam MG, Gera S, Gani MR (2009) Stratigraphy and
Clark JD (1985) Découvertes effectuée dans le middle Awash en 1981. structural evolution of the Blue Nile Basin, Northwestern Ethiopia
L’Anthropologie 88:129–130 Plateau. Geol J 44:30–56
CNR-CNRS (1971) Geological map of the northern Afar, scale Garland CR (1980) Geology of the Adigrat Area. Geological survey of
1:500,000. Geotechnip, La Celle St. Claude Ethiopia, Addis Ababa, 51 pp
CNR-CNRS (1975) Geological map of the central and southern Afar, George R, Rogers N, Kelley S (1998) Earliest magmatism in Ethiopia:
scale 1:500,000. Geotechnip, La Celle St. Claude evidence for two mantle plumes in one continental flood basalt
Coltorti M, Dramis F, Ollier CD (2007) Planation surfaces in northern province. Geology 26:923–926
Ethiopia. Geomorphology 89:287–296 Gibson IL (1969) The structure and volcanic geology of an axial
Cornwell DG, Maguire PKH, England RW, Stuart GW (2010) Imaging portion of the main Ethiopian rift. Tectonophysics 8:561–565
detailed crustal structure and magmatic intrusion across the Gortani M (1973) La fauna mesocretacea degli Stati di Graua. In
Ethiopian Rift using a dense linear broad band array. Geochem Missione Geologica dell’AGIP nella Dancalia Meridionale e sugli
Geophys Geosyst 11. doi:10.1029/2009GC002637 Altopiani Hararini (1936–1938). Acc Naz Lincei Roma 4:3–98
Corti G (2008) Control of rift obliquity on the evolution and Goudie AS (2005) The drainage of Africa since the Cretaceous.
segmentation of the main Ethiopian rift. Nature Geosci 1:258–262 Geomorphology 67:437–456
Corti G (2009) Continental rift evolution: from rift initiation to Gregnanin A, Piccirillo AM (1974) Considerazioni sulle serie vulca-
incipient break-up in the main Ethiopian Rift, East Africa. Earth Sci niche e sulla struttura dell’Altopiano Etiopico centrale. Mem Mus
Rev 96:1–53 Tridentino Sci Nat 20:79–100
Coulié E, Quideleur X, Gillot PY, Courtillot V, Lefevre JC, Chiesa S Grove AT, Street FA, Goudie AS (1975) Former lake levels and climatic
(2003) Comparative K–Ar and Ar/Ar dating of Ethiopian and changes in the rift valley of southern Ethiopia. Geogr J 141:177–202
Yemenite Oligocene volcanism: implications for timing and Hagos M, Koeberl C, Kabeto K, Koller F (2010) Geochemical
duration of the Ethiopian traps. Earth Planet Sci Lett 206:477–492 characteristics of the alkaline basalts and the phonolite-trachyte
62 E. Abbate et al.
plugs of the Axum area, northern Ethiopia. Austrian J Earth Sci Kazmin V, Seife MB, Nicoletti N, Petrucciani C (1980) Evolution of
103:153–170 the northern part of the Ethiopian rift. Atti Convegni Acc Lincei
Haile-Selassie Y, Deino A, Saylor B, Umer M, Latimer B (2007) Roma 47:275–292
Preliminary geology and paleontology of new hominid-bearing Kidane T, Otofuji YI, Brown FH, Takemoto K, Eshete G (2007) Two
Pliocene localities in the central Afar region of Ethiopia. Anthropol normal paleomagnetic polarity intervals in the lower Matuyama
Sci 115:215–222 Chron recorded in the Shungura Formation (Omo Valley, South-
Haile-Selassie Y, Suwa G, White TD (2009) Ardipithecus Kadabba: west Ethiopia). Earth Planet Sci Lett 262:240–246
late miocene evidence from the Middle Awash, Ethiopia. In: Haile- Kieffer B, Srndt N, Lapierre H, Bastien F, Bosh D, Pecher A, Yirgu G,
Selassie Y, WoldeGabriel G (eds) The middle awash series. Ayalew D, Weis D, Jerram DA, Keller F, Meugniot C (2004) Flood
University of California, Berkeley, pp 159–236 and shield basalts from Ethiopia: magmas from the African
Hayward NJ, Ebinger CJ (1996) Variations in the along-axis segmen- superswell. J Petrol 45:793–834
tation of the Afar Rift system. Tectonics 15:244–257 Kröner A (1985) Ophiolites and the evolution of tectonic boundaries in
Hofmann C, Courtillot V, Féraud G, Rochette P, Yirgu G, Ketefo E, Pik the late proterozoic Arabian–Nubian shields of northeast Africa and
R (1997) Timing of the Ethiopian flood basalt event and implications Arabia. Precambr Res 27:277–300
for plume birth and global change. Nature 389:838–841 Kumpulainen R, Uchman A, Woldehaimanot B, Kreuser T, Ghirmay S
Howell FC (1969) Remains of Hominidae from Plio/Pleistocene (2006) Trace fossil evidence from the Adigrat sandstone for an
formations in the lower Omo basin, Ethiopia. Nature 223:1234–1239 Ordovician glaciation in Eritrea, NE Africa. J Afr Earth Sci 45:408–420
Hunegnaw A, Sage L, Gonnard R (1998) Hydrocarbon potential of the Kunz K, Kreuzer H, Muller P (1975) Potassium-argon age determina-
intracratonic Ogaden Basin, SE Ethiopia. J Petrol Geol 21:401–425 tions of the Trap Basalts of the SE part of the Afar rift. In: Pilger A,
Ismail EH, Abdelsalam MG (2012) Morpho-tectonic analysis of the Roesler EA (eds) Afar depression of Ethiopia. Schweizerbart,
Tekeze River and the Blue Nile drainage systems on the Stuttgart, pp 370–374
Northwestern Plateau, Ethiopia. J Afr Earth Sci 69:34–47 Lalou C, von Nguyen H, Faure H, Moreira L (1970) Datation par la
Jepson DH, Athearn MJ (1964) Land and water resources of the Blue method Uranium-Thorium des hautes niveaux de coraux de la
Nile Basin. Appendix II Geology US Dept Interior, Addis Ababa, dépression de l’Afar (Ethiopie). Rev Géo Phys Géol Dyn 12:3–8
p 221 Larson PJ Jr (1977) Matabaietu, an Olduwan site from the Afar,
Johanson DC, Taieb M (1976) Plio-Pleistocene hominid discoveries in Ethiopia. Nyame Akuma 11:6–10
Hadar, Ethiopia. Nature 260:293–297 Le Bas MJ, Mohr PA (1968) Feldspathoid rocks from the Cainozoic
Johanson DC, Splingaer M, Boaz NT (1976) Paleontological excava- volcanic province of Ethiopia. Geol Rundsch 58:273–280
tion in the Shungura Formation, lower Omo basin, 1969–1973. In: Le Pichon X, Francheteau J (1978) A plate-tectonic analysis of the Red
Coppens Y et al. (eds) Earliest man and environments in the Lake Sea-Gulf of Aden area. Tectonoph 46:369–406
Rudolph basin: stratigraphy, paleoecology, and evolution. Univer- Le Turdu C, Tiercelin JJ, Gibert E, Travi Y, Lezzar K, Richert J,
sity Chicago Press, Chicago, pp 402–420 Massault M, Gasse F, Bonnefille R, Decobert M, Gensous B, Jeudy
Johnston C (1844) Travels in southern Abyssinia through the country V, Tamrat E, Mohamed MU, Martens K, Balemwal A, Chernet T,
of Adal to the Kingdom of Shoa. Madden, London, 492 pp Williamson D, Taieb M (1999) The Ziway-Shala lake basin system,
Juch D (1975) Geology of the south eastern escarpment of Ethiopia main Ethiopian rift: influence of volcanism, tectonics, and climatic
between 39° and 42° long, East. In: Pilger A, Roesler EA (eds) Afar forcing on human formation and sedimentation. Palaeogeogr
depression of Ethiopia. Schweizerbart, Stuttgart, pp 310–316 Palaeoclimatol Palaeoecol 150:135–177
Juch D (1980) Tectonics of the southeastern escarpment of Ethiopia. Leakey REF (1974) Further evidence of lower Pleistocene hominids
Atti Convegni Acc Lincei Roma 47:407–418 from East Rudolf, North Kenya. Nature 248:653–656
Justin-Visentin E, Zanettin B (1974) Dike swarms, volcanism and Lewin R (1983) Fossil lucy grows younger, again. Science 219:43–44
tectonics of the western Afar margin along the Kombolcha-Eloa Maguire PKH, Keller GR, Klemperer SL, Mackenzie GD, Keranen K,
traverse (Ethiopia). Bull Volcanol 38:187–205 Harder S, O’Reilly B, Thibo H, Asfaw L, Khan MA, Amha M
Kalb JE (1993) Refined stratigraphy of the hominid–bearing Awash (2006) Crustal structure of the Northern Main Ethiopian Rift from
Group, middle Awash valley, Afar depression, Ethiopia. Newsl the EAGLE controlled source survey: a snapshot of incipient
Stratigr 29:21–62 lithospheric break-up. In: Yirgu G, Ebinger CJ, Maguire PKH (eds)
Kalb JE, Oswald EB, Tebedge S, Mebrate A, Tola E, Peak D (1982) The Afar volcanic province within the East African rift system.
Geology and stratigraphy of Neogene deposits, Middle Awash Geol Soc Spec Publ 259:269–291
valley, Ethiopia. Nature 298:17–25 Makris MJ, Rhim R (1991) Shear-controlled evolution of the Red Sea:
Kappelmann J, Rasmussen DT, Sanders WJ, Faseha M, Bown T, pull apart model. Tectonophysics 198:441–466
Copeland P, Krabaugh J, Fleage J, Glanz M, Gordon A, Jacobs B, Manighetti I, Tapponnier P, Gillot PY, Jacques E, Courtillot V, Armijo
Maga M, Muldoon K, Pan A, Pyne L, Richmond B, Ryan T, Selffert R, Rurgg JC, King G (1998) Propagation of rifting along the Arabia–
ER, Sen S, Todd L, Wlemann MC, Winkler A (2003) Oligocene Somalia plate boundary into Afar. J Geophys Res 103:4947–4974
mammals from Ethiopia and faunal exchange between Afro-Arabia Martini IP, Brookfield EM, Steven S (2001) Principles of glacial
and Eurasia. Nature 426:549–552 geomorphology and geology. Prentice-Hall, New Jersey, 380 pp
Katho S, Nagaoka S, Woldegabriel G, Renne P, Snow MG, Beyene Y, McDougall I, Brown FH (2008) Geochronology of the pre-KBS Tuff
Suwa G (2000) Chronostratigraphy and correlation on the Plio- sequence, Omo Group, Turkana Basin. J Geol Soc London
Pleistocene tephra layers of the Konso formation, southern main 165:549–562
Ethiopian Rift, Ethiopia. Quat Sci Rev 19:1305–1317 McKenzie DP, Davies D, Molnar P (1970) Plate tectonics of the Red
Kazmin V (1973) Geological map of Ethiopia, scale 1:2,000,000. Sea and East Africa. Nature 226:243–248
Geological Survey of Ethiopia, Addis Ababa Meert JG, Lieberman BS (2008) The Neoproteozoic assembly of
Kazmin V (1980) Transform faults in the east African Rift system. Atti Gondwana and its relationship to Ediacaran–Cambrian radiation.
Convegni Acc Lincei Roma 47:65–73 Gondwana Res 14:5–21
Kazmin V, Shiferaw A, Balcha T (1978) The Ethiopian basement, Merla G (1963) Missione geologica nell’Etiopia meridionale del
stratigraphy and possible manner of evolution. Geol Rund 67: Consiglio Nazionale delle Ricerche 1959–1960. Notizie geomorf-
531–546 ologiche e geologiche. Giorn Geol 31:1–56
Merla G, Minucci E (1938) Missione geologica nel Tigrai. Volume Prior GT (1900) On Aegirite and Riebeckite Anorthoclase rocks related
primo: La serie dei terreni. Reale Accademia, Italia Roma, 363 pp to the “Grorudite-Tinguaite” series, from the neighborhood of
Merla G, Abbate E, Canuti P, Sagri M, Tacconi P (1973) Geological Aodowa and Axuma, Abyssinia. Min Mag 12:255–273
map of Ethiopia and Somalia, scale 1:2,000,000. Consiglio Quade J, Wynn JG (Eds) (2008) The geology of early humans in the
Nazionale delle Ricerche, Roma Horn of Africa. Geol Soc Am Spec Publ 446:234 pp
Merla G, Abbate E, Azzaroli A, Bruni P, Canuti P, Fazzuoli M, Sagri Ritsema J, van Heijst HJ, Woodhouse JH (1999) Complex shear wave
M, Tacconi P (1979) A geological map of the Ethiopia and Somalia velocity structure imaged beneath Africa and Iceland. Science
and comment with a map of major landforms (scale 1:2,000,000). 286:1925–1928
Consiglio Nazionale delle Ricerche, Roma, 95 pp Rochette P, Tamrat E, Féraud G, Pik R, Courtillot V, Ketefo E, Coulon
Meshesha D, Shinjo R (2008) Rethinking geochemical feature of the C, Hoffmann C, Vandamme D, Yirgu G (1998) Magnetostratigra-
Afar and Kenya mantle plumes and geodynamic implications. phy and timing of the Oligocene Ethiopian traps. Earth Planet Sci
J Geoph Res 113. doi:10.1029/2007JB005549 Lett 164:497–510
Mohr P (1962) The geology of Ethiopia. University College of Addis Rogers NW (2006) Basaltic magmatism and the geodynamics of the East
Ababa Press, Ethiopia, 268 pp Africa rift system. In: Yrgu G, Ebinger CJ, Maguire PKH (eds) The
Mohr P (1963) Geological map of Horn of Africa, scale 1:2,000,000. Afar volcanic province within the East Africa rift system. Geological
Philip and Tacey, London Society London Special Publication, London, 259, pp 77–93
Mohr P (1967) The Ethiopian rift system. Bull Geoph Obs Addis Rüppell WE (1834) Skizze der geologischen Formation Abyssiniens.
Ababa 11:1–65 Abhandlungen Museum Senkenbergianum 1:286–288
Mohr P (1968) Transcurrent faulting in the Ethiopian rift system. Russo A, Getaneh A, Atnafu B (1994) Sedimentary evolution of the
Nature 218:938–941 Abbai River (Blue Nile) Basin, Ethiopia. N J B Geol Palaeo Mh
Mohr P (1983) Ethiopian flood basalt province. Nature 303:577–584 5:291–308
Mohr P (2009) Africa beckoning. Explorers of eastern Africa, its rift Sacchi R, Alene M, Barbieri M, Conti A (2007) On the Paleozoic
valleys and geology, in the 19th and early 20th centuries. Millbrook Tillite of the Adigrat Group (Tigrai, Ethiopia). Periodico Minera-
Nova press, Tόin An Ghárrain, Corr an Dola, Co na Gaillime, logia 76:241–251
192 pp Said R (1993) The River Nile geology, Hydrology and Utilization.
Mohr P, Wood CA (1976) Volcano spacings and lithospheric Pergamon Press, Oxford, p 320
attenuation in the Eastern Rift of Africa. Earth Planet Sci Lett Schilling JG (1973) Afar mantle plume: rare earth evidence. Nature
33:126–144 242:2–5
Mohr P, Zanettin B (1988) The Ethiopian Flood Basalt Province. In: Sepulchre P, Ramstein G, Fluteau F, Schuster M, Tiercelin JJ, Brunet
Macdougall JD (ed) Continental flood basalts. Kluwer Academic M (2006) Tectonic uplift and eastern Africa aridification. Science
Publisher, The Netherlands, pp 63–110 313:1419–1423
Morbidelli L, Nicoletti C, Petrucciani C, Piccirillo EM (1975) Sichler B (1980) La biellette danakile. Bull Soc Geol Fr 6:925–933
Ethiopian south-eastern plateau and related escarpment: K/Ar ages Sickenberg O, Schoenfeld M (1975) The chorora formation—lower
of the main volcanic events (Main Ethiopian Rift from 8° 10’ to pliocene limnical sediments in the southern Afar (Ethiopia). In:
9° 00’ lat. N). In: Pilger A, Roesler EA (eds) Afar depression of Pilger A, Roesler EA (eds) Afar depression of Ethiopia. Schweiz-
Ethiopia. Schweizerbart, Stuttgart, pp 362–369 erbart, Stuttgart, pp 277–284
Morgan WJ (1971) Convection plumes in the lower mantle. Nature Stankiewicz J, de Wit MJ (2006) A proposed drainage evolution model
230:42–43 for the Central Africa—did the Congo flow east? J Afr Earth Sci
Morton WH, Black R (1975) Crustal attenuation in Afar. In: Pilger A, 44:75–84
Roesler EA (eds) Afar depression of Ethiopia. Schweizerbart, Stefanini G (1933) Carta Geologica dell Eritrea, della Somalia e
Stuttgart, pp 55–65 dell’Etiopia alla scala di:2.000.000. Istituto Geografico Militare
Moucha R, Forte AM (2011) Changes in African topography driven by Firenze
mantle convection. Nature Geosci 4:707–712 Stern RJ (1994) Neoproterozoic (900–550 Ma) arc assembly and
Munzinger W (1864) Ostafrikanische studien. Schaffhausen, p 584 continental collision in the East African Orogen. Ann Rev Earth
Natali C, Beccaluva L, Bianchini G, Siena F (2013) The Axum-Adwa Planet Sci 22:319–351
basalt-trachyte complex: a late magmatic activity at the periphery of Street FA (1979) Late quaternary lakes in the Ziway-Shala basin,
the Afar plume. Contr Min Petrol. doi:10.1007/s00410-013-0879-0 Southern Ethiopia. Ph.D. thesis, University of Cambridge, Cam-
Overstreet WC, Stoeser DB, Overstreet EF, Goudarzi GH (1977) bridge, 457 pp
Tertiary laterite of the As Sarat mountains, Asir Province, Kingdom Taieb M (1974) Evolution quaternaire du basin de l’Awash (rift
of Saudi Arabia. Min Res Bull Saudi Arabia Direct Gen Min Res éthiopienne et Afar). Doctoral Univ Paris, 442 pp
21:1–30 Tefera M, Chernet T, Haro W (1996) Geological map of Ethiopia, scale
Peccerillo A, Donati C, Santo AP, Orlando A, Yirgu G, Ayalew D 1:2,000,000. Ethiopian Mapping Authority, Ethiopia
(2007) Petrogenesis of silicic peralkaline rocks in the Ethiopian rift: Tiercelin JJ (1986) The pliocene hadar formation; Afar depression of
geochemical evidence and volcanological implications. J Afr Earth Ethiopia. In: Frostick LE, Renaut RW (eds) Sedimentation in the
Sci 48:161–173 African Rifts. Geological Society Special Publication No 25,
Pickford M (1990) Uplift of the roof of Africa and its bearing on the pp 221–240
evolution of mankind. J Hum Evol 5:1–20 Tiercelin JJ, Taieb M, Faure H (1980) Continental sedimentary basins
Pik RD, Coulon C, Yirgu G, Hofmann C, Ayalew D (1998) The and volcano-tectonic evolution of the Afar Rift. Atti Convegni Acc
Northwestern Ethiopian plateau flood basalts: classification and Lincei Roma 47:491–504
spatial distribution of magma types. J Volc Geoth Res 81:91–111 Ukstins IA, Renne PR, Wolfenden E, Baker J, Ayalew D,
Pik R, Marty B, Carignan J, Lavé J (2003) Stability of the Upper Nile Menzies M (2002) Matching conjugate volcanic rifted margins:
drainage network (Ethiopia) deduced from (U-Th)/He thermochr- 40/39Ar chrono-stratigraphy of pre-and syn rift bimodal flood
onometry: implications for uplift and erosion of the Afar plume volcanism in Ethiopia and Yemen. Earth Planet Sci Lett
dome. Earth Plan Sci Lett 215:73–88 198:289–306
64 E. Abbate et al.
Unger F (1866) Notiz über fossile Hölzer aus Abyssinien. Sit- Wolela A (2008) Sedimentation of the Triassic–Jurassic Adigrat
zungenesberichte der Kaiserlichen Akademie der Wissenshaften sandstone formation, BlueNile (Abbai) basin, Ethiopia. J Afr Earth
Math. Naturwissenshaft Klasse, Wien 54:289–296 Sci 52:30–42
Vinassa de Regny P (1931) La geologia delle Alpi Dancale. Boll Soc Wolfenden E, Ebinger C, Yirgu G, Deino A, Ayalew D (2004)
Geol It 50:1–24 Evolution of the northern Main Ethiopian rift: birth of a triple
Walker A, Leakey REF (1978) The hominids of East Turkana. Sci Am junction. Earth Plan Sci Lett 224:213–228
239:54–66 Yemane K, Taieb M, Faure H (1987) Limnologic studies on an
Walter RC (1994) Age of Lucy and the First Family: Single crystal intertrappean continental deposit from the northern Ethiopian
40
Ar/39Ar dating of the Denen Dora and lower Kada Hadar Plateau (37° 03′E, 12° 25′N). J Afr Earth Sci 6:91–101
members of the Hadar Formation, Ethiopia. Geology 22:6–10 Zanettin B, Justin-Visentin E (1973) Serie di vulcaniti etiopiche: 1—La
White RS, McKenzie DP (1989) Magmatism at rift zones: the serie dell’altopiano etiopico centro-orientale. Boll Soc Geol It
generation of a continental margins and flood basalts. J Geoph 92:313–327
Res 94:7685–7729 Zanettin B, Justin-Visentin E (1974) The volcanic succession in central
White TD, Johanson DC (1982) Pliocene hominid mandibles from the Ethiopia. 2—The volcanics of western Afar and Ethiopian rift
Hadar Formation, Ethiopia; 1974–1977 collections. Am J Phys margins. Mem Ist Min Univ Padova 31:1–19
Anthropol 57:501–544 Zanettin B, Justin-Visentin E (1975) Tectonical and volcanological
White TD, Asfaw B, Beyene Y, Haile Selassie Y, Lovejoy CO, Suwa evolution of the western Afar margin (Ethiopia), In: Pilger A,
G, WoldeGabriel G (2009) Ardipithecus ramidus and the paleobi- Roesler EA (eds) Afar depression of Ethiopia. Schweizerbart,
ology of early hominids. Science 326:75–86 Stuttgart, pp 300–309
Wichura H, Bousquet R, Oberhäusli R, Strecker MR, Trauth MH Zanettin B, Gregnanin A, Justin-Visentin E, Nicoletti M, Petrucciani C,
(2010) Evidence for middle Miocene uplift of the East African Piccirillo EM, Tolomeo L (1974) Migration of the Oligocene-
Plateau. Geology 38:543–546 Miocene ignimbritic volcanism in the Central Ethiopan Plateau.
WoldeGabriel G, Aronson JL, Walter RC (1990) Geology, geochro- NGB Geol Palaeo Mh H 9:567–574
nology, and rift basin development in the central sector of the Main Zanettin B, Justin Visentin E, Nicoletti M, Piccirillo EM (1980)
Ethiopia Rift. Geol Soc Am Bull 102:439–458 Correlations among Ethiopian volcanic formations with special
Wolela A (2007) Fossil fuel energy resources of Ethiopia: coal references to the chronological and stratigraphical problems of the
deposits. Int J Coal Geol 72:293–314 “Trap Series”. Atti Convegni Acc Lincei Roma 47:231–252
The Climate of Ethiopia
3
Massimiliano Fazzini, Carlo Bisci, and Paolo Billi
Abstract
Several papers have been published on different issues regarding the climate of Ethiopia or of
some specific region. This presentation attempts to revise the knowledge of the climate of
Ethiopia by means of updated, longer time series and including a larger number of meteo-
stations than previous studies. Basic climatic parameters such as temperature, rainfall, relative
humidity, wind, evapotranspiration, and aridity are considered and their spatial distribution is
analyzed. The main results of such elaborations have been regionalized to obtain climatic
maps by means of geostatistical interpolation, also taking into account topogeographic
variables. These parameters were also used to update the Köppen classification of the
Ethiopian territory. Climate change is a very important issue with worrying repercussions on
agriculture and hence the social and economic development of the country. Trends of
temperature and annual, spring, and summer rains were interpolated for the last 3–5 decades.
Temperature shows a markedly increasing trend especially as regards the minimum values,
whereas annual rainfalls tend to decrease with the spring, ‘small rains’ decreasing at faster rate.

Keywords
Climate Ethiopia Temperature Precipitation Aridity index
3.1 Introduction large spatial variations in temperature and precipitation (see

Sect. 3.3). The climate of Ethiopia is therefore mainly con-
Ethiopia is among the largest countries of Africa and it is trolled by the seasonal migration of the Intertropical Con-
characterized by a wide variety of landscapes, with marked vergence Zone (ITCZ) and associated atmospheric
contrasts in relief and altitudes ranging from about 155 m circulations as well as by the complex topography of the
below sea level of Assale Lake, in the Danakil depression, to country. Landscapes with contrasting characteristics in terms
about 4,533 m a.s.l. at Ras Dejen (EMA 1988). For these of physiography and elevation, such as the highlands and the
reasons and given its geographic position close to the lowlands, experience a variety of climates from desert cli-
equator and the Indian Ocean, the country is subjected to mate to that typical of equatorial mountains.
Climate, in turn, has many obvious implications on
landforms and morphodynamic evolution of natural land-
M. Fazzini P. Billi (&)
scapes as much as on the living conditions of local people, in
Physics and Earth Sciences, a country whose economy is heavily dependent on rain-fed
University of Ferrara, Ferrara, Italy agriculture (Hadgu et al. 2013).
e-mail: bli@unife.it The relevance of climate and its variability on natural
M. Fazzini resources of Ethiopia and their potential exploitation has been
e-mail: fzzmsm@unife.it recognized since long by the first Western explorers visiting
C. Bisci the country. The establishment of a few weather stations with
School of Science and Technology, regular observations since the beginning of the twentieth
University of Camerino, Camerino, Italy

66 M. Fazzini et al.
century led Ethiopia to assume a prominent position in the (Lovett and Wood 1976), nonetheless, they are of great
Horn of Africa as far as the availability of reference meteo- interest for a preliminary description of the general climate
rological data is concerned, also within the operational of Ethiopia and surrounding regions at the nineteenth–
framework of international climate organizations. twentieth-century transition. Given their spatial distribution,
Several studies on the climate of Ethiopia have been these data can be considered suitable only for a climatic
published throughout the past two centuries (see the next point characterization and as a baseline, though some sta-
sections), but only very few of them benefited from the tions may have changed location and instrumentation, to
inherent, ongoing expansion of the time series in refining the study climate variations, if any, but do not allow to define
definition of the different climate types and their spatial the climate of Ethiopia with sufficient detail.
variability. Instead, they rather focused on specific topics,
with climate change and its implications becoming the most
investigated theme in the last decade. The aim of this chapter 3.2.2 More Recent Data and Studies
is therefore to update the knowledge of the main character-
istics of the climate of Ethiopia and to analyze the occur- In Ethiopia, most of the longest rainfall and temperature
rence of climate change trends on the base of time series records began in the 1950s and 1960s. Many other records
longer than those used in previous studies. started during the 1980s, following the impulse from the
Ethiopian Government to expand the network, partly in
response to the mid-1980s drought. In 1951, a small mete-
3.2 Previous Studies orological unit was formally established within the Civil
Aviation Department of Ethiopia, to provide meteorological
3.2.1 Historical Data information solely for flying purposes. As the need for
meteo-climatic data increased, a Meteorological Department
In sub-Saharan Africa, century long time series of meteo- was established in 1964, under the Civil Aviation Authority.
climatic data are scarce and sparse over large portions of Later, this department became the modern National Meteo-
territory (Houghton et al. 2001; Conway et al. 2004). Ethi- rological Agency (NMA), after designation as an autono-
opia is a large country, with very few well-documented long mous organization on December 31, 1980 (ENRAEMED
series of climate observations. The longest and uninterrupted 2003).
climate record for the Horn of Africa is the rainfall and Since 1984, FAO (1984) has been publishing mean
thermometric (but also wind, pressure, relative humidity, monthly agroclimatic data, including decadal rainfall data
and sunshine) data record from Asmara which began in 1890 for Ethiopia. FAO obtained these data from the National
and from Addis Ababa since 1898. These data derived from Meteorological Agency and Chernet (1982) used them for a
four intermittent sources, covering different time intervals: concise description of the Main Ethiopian Rift Valley cli-
Russian (1898–1903), Italian (1903–1904), British (1901– mate, whereas Gamachu (1988) based on them his study on
1904), and French (1898). climatic elements of mountainous regions of Ethiopia.
Some preliminary data on temperature, rainfall, relative In 1994, NOAA–NCDC put together a huge database at
humidity, and wind for Addis Ababa and Addis Alem, plus global scale, the ‘Global Historical Climate Network
sparse climatic information for a few other sites in Ethiopia, (GHCN),’ which includes daily data of several meteo-sta-
were reported by Eredia and De Castro (1914), but the tions located in Ethiopia, the source of which, however, is
quality of these data is questionable for both the short always the NMA.
duration and the methodology of measurements.
The study by Fantoli (1940) considered the meteo-cli-
matic data of three thermometric stations in Ethiopia—Addis 3.2.3 The National Atlas
Ababa (1902–1930), Gambela (1914–1932), and Harar
(1902–1918)—and other six thermometric stations, includ- In 1988, the Ethiopian Mapping Authority published the first
ing Asmara (1890–1930), located in the bordering territories, edition of the National Atlas of Ethiopia. The Atlas is likely
and 18 rain gauges, some of which were located beyond the first publication with a comprehensive description of the
modern Ethiopia. For a few of these meteorological stations, climate of Ethiopia. In fact, it includes several maps
additional meteo-climatic parameters such as atmospheric depicting the spatial distribution of a few climatic parameters
pressure, relative humidity, wind speed and direction, and such as rainfall, rainfall patterns, hail, frost, temperature,
insolation were reported. sunshine hours, climate classification, wind velocity and
These data series were used by Hurst and Black (1937), direction and the location of the meteorological stations as
Fantoli (1940, 1965), WWR (1959), and Conway et al. well. The data to prepare these climate sections of the Atlas
(2004). Though some stations have incomplete records were provided by the Ethiopian Meteorological Agency as
3 The Climate of Ethiopia 67
far as 1977, and the data analysis and summary are partly evapotranspiration, whereas the spatial and temporal distri-
based on the studies of Gamachu (1977). Though the spatial bution of temperatures have not been widely and satisfac-
distributions of climatic parameters for the three decades torily analyzed. The majority of modern studies deal with
before the mid-1970s are reported, this publication still limited areas and, commonly, are constrained by the avail-
represents a reliable reference and a milestone in the history ability of homogeneous time series of good quality. For
of climatological studies in Ethiopia. This paper draws on these reasons, modern investigations on the general charac-
the inspiring climatic maps and chapters of the National teristics of the climate of Ethiopia are missing, with the
Atlas of Ethiopia and attempts to update the analysis of exception of the National Meteorological Agency publica-
climatic parameters and climate classification considering tions (e.g., Tadege 2001, 2007) that provide a reliable
also the data of the following three decades as far as 2010. description of the climate across the whole country, though
limited to the end of the second millennium and early
beginning of the third. In the last decade, the NMA has also
3.2.4 Modern Studies expanded the number of recording stations to a remarkable
quantity and presently the entire territory of Ethiopia is
In the scientific literature, there are a number of papers on densely, though heterogeneously, covered (Fig. 3.1).
precipitation in Ethiopia and related problems, such as In the 2007 NMA report, temperature data are considered,
extreme rains, droughts, trend variations, and and for the 1951–2006 interval, increasing rates of 0.13 °C/
Fig. 3.1 Mean annual temperatures

decade and 0.37 °C/decade for mean annual minimum and analyzed the temperature and rainfall trends for the 1948–
maximum temperature, respectively, are presented. For the 2006 interval and performed a simulation projecting these
1960–2006 interval, the report by McSweeney et al. (2010) trends to 2050 under the assumption of a gradual doubling of
indicates a mean annual temperature increase of 0.28 °C per greenhouses gases. For the time series bounded to 2006,
decade and a significant increasing trend in the frequency of these authors find that air temperatures increase at the rate of
hot days and a decrease of cold days. These authors also about 0.03 °C per year across most of Ethiopia, with the
found that no statistically significant trend is recognizable exceptions of the lowlands and the northern Rift Valley,
for mean precipitation in any season. whereas rainfall trends indicate a weak increase in the arid
In a more recent study on the upper Blue Nile basin, lowlands of southeastern Ethiopia.
based on statistical and geostatistical techniques and gridded Droughts have been studied by Shanko and Camberlin
data, reconstructed from NMA weather stations and meteo- (1998)—in connection with the development of tropical
rological satellite records, Mengistu et al. (2013) found an cyclones—and by Viste et al. (2013). These latter authors,
insignificant change in temperature across the 1981–2010 analyzing rainfall data for the 1971–2011 interval, found that
interval. the drought patterns are largely influenced by the variation in
Mekasha et al. (2013) have analyzed temperature and the seasonal precipitation cycle among different areas of
rainfall extremes for 11 stations and found a general ten- Ethiopia.
dencies of increasing warm and decreasing cold extremes, Scarce information is available on other climatic parameters
whereas trends in precipitation extremes were much more such as velocity and direction of winds, air humidity, barometric
variable and increasingly inconsistent among neighboring pressure, and evapotranspiration (Enku and Melesse 2013).
stations. This study, however, does not cover the wide Though the climate of Ethiopia received a lot of attention
diversity of topography and landscapes of Ethiopia; there- from scientists and many papers have been published on
fore, its results cannot be considered as representative for the different climatic topics, given its complexity and variability
whole country. and the availability of longer time series, an updated sum-
Yet, the study of Selashi and Zanke (2004) on rainfall mary and description of the main parameters is considered
amount and rainy days changes across the country is based here of great help for those interested in the subject.
on the data from 11 meteo-stations spanning the 1965–2002
interval. These authors found no trend in the annual, kiremt
(the main, monsoon-type rainy season in the local language, 3.3 Summary Description
see the next sections) and belg (the small, spring time of the Climate of Ethiopia
rainfall in the local language) rainfall totals and rainy days
over the largest part of Ethiopia. A study on extreme rainfall The climate of Ethiopia is mainly controlled by the seasonal
and dry spell events across the same 1965–2002 interval, migration of the ITCZ and associated atmospheric circulations
based upon the same time series, has been carried out by (Beltrando and Camberlin 1993), the complex physiography
Selashi and Camberlin (2006) who found contrasting, and the marked contrast in elevation among large different
though weak, trends in extreme precipitation for both the areas of the country. The transition between lowlands and
kiremt and the belg in different geographic areas, showing a highlands is commonly very sharp, resulting in a variety of
high variability for both rainfall seasons. These results climates, from very arid to very humid typical of equatorial
contrast with a previous work by Easterling et al. (2000), mountains, with further differentiation at local scale. More-
whose findings indicate that extreme intensity is decreasing over, precipitation varies with latitude, decreasing from south
over the Ethiopian highlands. to north, whereas the meteorological framework is deeply
Other papers report climatic information on specific areas affected also by elevation and physiography, especially as
or regions of Ethiopia (e.g., Conway 2000a, b; Bewket and regards temperature distribution and anemometric character-
Conway 2007; Taye and Zewdu 2012). The paper by istics. Finally, the proximity of the Asiatic continent has to be
Bewket and Conway (2007) reports about the temporal and considered as well. In winter, in fact, the contrast between the
spatial variability of rainfall for a relatively small portion of thermal anticyclone of western Asia and Egypt and the
the country, the Amhara Region, and found no consistent equatorial low pressures determines the presence of trade
emergent pattern or trend in daily rainfall. These authors, winds blowing from northeast to southwest. These winds,
however, pointed out that the time span selected may have a relatively cool but rather dry, control the dry period (bega in
relevant influence on the results of trend analysis. Cheung the local language). In spring, the influence of southwestern
et al. (2008) arrived at similar conclusions, verifying that no winds, coming from the Congo basin, determines the season of
significant changes or trend in annual rainfall is evident at ‘little rains’ (belg in the local language) that can bring rela-
the national or watershed level in Ethiopia. Finally, Jury and tively abundant precipitation in the southern part of the
Funk (2013), using gridded data from different sources, country. In summer, the Guinean monsoon, consisting of
equatorial warm and humid winds, results in bountiful rains intervals and of the meteo-stations to be considered in this
(kiremt in the local language) which are also substantially study.
influenced by the orographic diversity mentioned above. The notable physiographic variability of the country in
Such a complex meteorological framework is reflected by many cases did not allow to fill the gaps by means of
the distribution of annual precipitation. In the Danakil ordinary correlation analysis with other stations in the
depression, it is constantly less than 250 mm but can be as vicinity, resulting in insignificant correlation coefficients
low as 50 mm. By contrast, on the highlands, 2,000 mm can (Hadgu et al. 2013). Such constraint is particularly relevant
be locally exceeded. Similar values of annual precipitation, for time series and trend analysis, especially for rainfall.
however, are recorded also in the southwestern lowlands, Notwithstanding such limitations and nonuniformity, some
likely due to a larger contribution of the spring rains. indications on long-term trends can be given anyway.
Altitude patently determines marked annual thermal Jury and Funk (2013) report about spatial variability of
gradients with sharp transitions from zones with a desert rainfall on the basis of gridded data and interpolation. Given
climate—among the hottest of the planet—to tropical, high the low density of meteorological stations and the rugged
mountain climates with minimum temperatures below zero topography of Ethiopia, this procedure does not seem to
in each month and modest annual thermal amplitude. provide a realistic pattern of rainfall distribution. Addition-
ally, the data set used by these authors includes a five sta-
tions correlation to fill data gaps, and a correlation as low as
3.4 Data Sources 0.2 has been accepted as significant. In the worst case, the
mean values of the time series are included. This procedure
As briefly introduced in the previous section, the high vari- seems, at least, questionable; in fact, it does not avert the
ability of physiography and orography of Ethiopia is reflected data limitations and introduces arbitrary conditions that are
in the complex spatial distribution of precipitation and tem- not necessarily making the data analysis stronger or more
perature as briefly introduced in the previous section. Having reliable than the use of raw data.
in mind such a complexity, large sets of data and long time Spatial distribution of rainfall has been investigated also
series for the main climatic parameters are required to produce by Viste et al. (2013) by using the raw data of the NMA, but
an updated description of the climate of Ethiopia and, possi- their study focused mainly on monthly precipitation. Other
bly, of its variations through time. For this purpose, a large authors such as Seleshi and Cumberlin (2006) investigated
number of monthly temperature and, whenever possible, daily the spatial and seasonal (bega, belg and kiremt) rainfall
rain gauge data were collected from different sources extremes, but their paper is focused on 5 days consecutive
including the Ethiopian National Meteorological Agency rains rather than daily rainfall.
(NMA), which is the main provider of field data measure- Notwithstanding the awareness of data limitations, the
ments, and other international organizations such as the approach of this paper attempts to complement the analyses
Global Historical Climatology Network (GHCN) of the of seasonal and spatial variability reported in previous
National Oceanic and Atmospheric Administration (NOAA) papers (e.g., Segele and Lamb 2005; Seleshi and Cumberlin
in the USA and the United Nations Food and Agricultural 2006; Korecha and Barnston 2007; Viste et al. 2013).
Organization (FAO), the data sets of which are substantially As an alternative to the manipulation of data beyond any
based on NMA data. The NMA produces daily rainfall data reasonable confidence and notwithstanding the existing gaps
for all the many stations that are presently operative; in 22 of and limitation and the need for the time series as longest as
them also hourly precipitation is measured, whereas FAO possible, in order to attain a certain degree of coherence with
reports rainfall data at decadal (10 days) intervals. the actual physical phenomena, the following procedure was
The temperature data of the NMA, and those of the used. Statistical tests were carried out in order to ascertain
GHCN as well, are reported as daily data for the whole the quality of data and the uniformity of the temperature time
operational time interval of first level meteo-stations and as series. The standard normal homogeneity test (Alexanders-
monthly data for lower level gauges, whereas FAO provides son 1986) and the Kolmogorov–Smirnov test were used to
only decadal data. determine the homogeneity of records, whereas the Mann-
Kendall test was used to verify the statistical significance of
trends at a 0.05 level of significance.
3.5 Data Validation and Processing Both the quality tests indicate that rainfall time series are
generally homogeneous, whereas only four out of 18 tem-
Unfortunately, the above data sets are commonly affected by perature time series proved inhomogeneous. Furthermore,
long and reiterated gaps and may include entire years of the variability of annual rainfall and temperatures over the
measurements missing. This posed several problems in the study period was analyzed by calculating the coefficient of
validation of data and in the selections of uniform time series variability (CV).
In order to provide a general picture of temperatures and Ethiopian borders and thus to improve as much as possible
precipitation variability across the whole territory of Ethio- the reliability and quality of the analyses (Fig. 3.1).
pia in the last decades and in consideration of the above- The elevation of the stations considered varies between
mentioned quality limitation of the available time series, a 10 m a.s.l. at Massawa and 2,760 m a.s.l. at Fitche. Though
simple procedure, based on the normalization of data, was marked differences are observed for the average annual and
used. For each station, the annual precipitation and the mean monthly values in the highlands—typical of equatorial high
minimum and maximum temperature of each year have been mountains—and in the northeastern lowlands, Tigray and
normalized and the Zi score calculated as follows: the Omo valley, the climate of Ethiopia is generally very
pleasant and among the most equilibrate of the African
Zi ¼ ðDi Dm Þ=sd continent since the temperatures are mitigated by the high
average elevation. Only in the Danakil Desert areas the
in which Di is the parameter annual value for the ith year, Dm temperatures reach extreme values.
is the mean value of the time series, and sd its standard The lowest mean annual temperatures are recorded at ele-
deviation. vations over 2,300–2,600 m a.s.l. (Fig. 3.1 and Table 3.1),
For each year of the time span considered, a representa- irrespective of their geographic position. At Fitche, the highest
tive Z score has been calculated by averaging the Zi scores of and coldest station, the mean annual temperature is 14.2 °C,
all the stations. In this way, a new time series of Z scores is whereas at the elevation of 2,000 m it is around 18.5 °C.
obtained. By this procedure, though not orthodox, it is In the Rift Valley, the temperatures gradually increase
possible to minimize the relative effects of gaps that are from south to north. Around the elevation of 1,500 m, they
present and almost randomly distributed within the whole are slightly over 20 °C and the 25° isotherm stands around
data set and to depict the time distribution of higher and 1,000 m a.s.l. At lower elevations, the temperatures notice-
lower than average values. Any attempt to fill the gaps ably increase and peak to 30 °C in the steppe area of Gore
through correlations with neighboring stations as suggested and to higher values in the Danakil Desert.
by Jury and Funk (2013) resulted to be unsatisfactory. On The mean maximum temperatures are particularly pleas-
the other hand, the use of normalized data and averaged ant at elevations above 1,500 m. At Fitche, they are 20,
Z scores has been considered appropriate to reduce the effect 23.5 °C at Addis Ababa and about 25 °C around the 2,000 m
of local outliers and to delineate trends that depict a general elevation. At 1,000 m, the mean values are close to 30 °C but
tendency for the entire country. This issue, however, would can be higher than 35 °C in the southwestern lowlands and,
need further detailed and more sound statistical analyses that mainly, in the steppe areas and the Danakil Desert (Fig. 3.2).
are beyond the scope of this paper. The mean minimum temperatures (Fig. 3.3) are rather
For the main climatic parameters, the data elaboration is low all across the highlands, where a few days of frost are
portrayed in colors overlying a shaded relief map. The commonly recorded every year. At Robe Bale, the station
regionalization of the data has been obtained through geo- with the lowest minimum temperatures, the mean annual
statistical interpolations considering the local elevation, lat- value is 8 °C, whereas at Addis Ababa it is 10 °C. Higher,
itude, distance from the sea and the largest lakes, and slope typically tropical values of 20–25 °C are recorded at ele-
aspect for the whole area. For temperatures, the theoretical vations below 500 m.
monthly solar radiation was considered as well. All the maps The lowest temperature ever recorded in the meteo-sta-
were interpolated and drawn using advanced GIS tools tions considered is −7 °C on the Bale Mountains (highest
(ESRI ArcGIS–ArcInfo). peak 4,377 m a.s.l.), whereas in Addis Ababa frost condi-
tions only sporadically occur.
In Ethiopia, one of the hottest areas of the planet is found:
3.6 Temperatures the Danakil depression (Vinassa de Regny 1931; Fantoli
1940). Unfortunately, recent data are not available to con-
To analyze the spatial distribution of temperature, the firm this evidence. At Dallol (130 m b.s.l.), which is con-
monthly data from 30 meteo-stations, with continuous series sidered the hottest, inhabited place of the world, Pedgley
or with negligible gaps throughout the World Meteorologi- (1967) measured a mean annual temperature of 34.7 °C,
cal Organisation (WMO) 1981–2010 reference interval, mean maximum temperatures of 41.2 °C, and a peak of
were selected (Table 3.1). Additional 17 meteo-stations, 45.7 °C for the 1960–1966 interval. According to Billi
located in the adjoining territories of Djibouti, Eritrea, (personal communication), in November 2007, temperatures
Somalia, and Sudan, were considered as well (Table 3.1) in of 42–44 °C were commonly recorded in the Samoti plain in
order to account also for the conditions right beyond the the Eritrean Danakil.
Table 3.1 Main temperature data

Meteo-stations H Tmean Tmax Tmin Ea Ed Mh Mc
(m a.s.l.) (°C) (°C) (°C) (°C) (°C)
Addis Ababa 2,354 17.0 23.5 10.4 3.1 13.1 May Dec
Al Damazin 470 27.5 35.1 19.8 7.2 15.3 Apr Sep
Alamata 1,520 22.5 30.0 14.9 6.3 15.1 Jun Jan
Alemaya 2,047 16.5 23.7 9.3 5.7 14.4 Jun Dec
Arba Minch 1,285 23.8 30.2 17.3 2.9 12.9 Mar Jul
Asab 14 30.4 35.4 25.4 9.5 10.0 Aug Feb
Asela 2,430 15.4 21.9 8.9 2.7 13.0 May Dec
Asmara 2,325 15.2 22.9 7.4 5.0 15.5 Jun Dec
Awasa 1,750 20.0 27.3 12.6 1.8 14.7 Apr Nov
Bahir Dar 1,770 19.1 26.9 11.2 4.6 15.7 May Jan
Belet Uen 198 28.6 34.8 22.4 2.6 12.4 Apr Aug
Berbera 1 30.5 35.2 25.8 13.1 9.4 Jul Jan
Borama 1,454 21.2 27.3 15.0 4.9 12.3 Mar Jul
Burao 970 22.8 29.3 16.3 6.2 13.0 May Jan
Debre Marcos 2,515 16.4 22.5 10.3 3.7 12.2 Apr Aug
Dire Dawa 1,260 25.5 31.8 19.1 7.0 12.7 Jun Jan
Djibouti 21 29.8 33.8 25.7 11.3 8.1 Jul Jan
Fitche 2,750 14.2 20.3 8.1 3.1 12.2 May Nov
Galcayo 247 30.4 34.1 26.7 9.8 7.4 Jul Feb
Gambela 484 27.6 34.8 20.4 5.3 14.4 Mar Jun
Garowe 473 30.3 33.9 26.7 9.9 7.2 Jul Dec
Gewane 627 21.5 24.9 18.1 8.0 13.8 Jun Dec
Gode 295 28.9 34.7 23.1 2.8 11.6 Mar Nov
Gonder 1,967 20.0 26.7 13.3 4.6 13.4 Apr Aug
Gore 2,002 18.9 23.9 13.9 3.7 10.0 Mar Jul
Hargeysa 1,347 22.3 27.4 17.1 5.0 10.3 May Jan
Hayk 2,030 18.3 25.8 10.7 5.8 15.1 Jun Dec
Jijiga 1,644 19.4 26.5 12.3 4.5 14.2 Aug Jan
Jimma 1,725 19.7 27.6 11.8 3.0 15.8 Apr Dec
Kassala 502 29.5 37.1 21.8 8.3 15.3 May Jan
Kebry Dehar 549 26.6 33.1 20.1 2.2 13.0 Apr Aug
Kobo 1,610 22.1 29.3 14.9 7.0 14.4 Jun Dec
Kombolcha 1,903 19.3 26.5 12.1 5.9 14.4 Jun Dec
Mandera 231 29.5 34.9 24.1 3.7 10.8 Mar Jul
Massawa 10 29.4 34.5 24.3 10.1 10.2 Jul Jan
Mekele 2,070 17.6 23.8 11.3 4.6 12.5 Jun Dec
Metahara 947 25.4 33.3 17.4 6.3 15.9 Jun Dec
Mogadishu 7 26.7 30.2 23.1 3.1 7.1 Apr Aug
Moyale 1,097 22.4 27.3 17.5 5.2 9.8 Feb Jul
Negelle 1,544 20.7 26.3 15.0 4.7 11.3 Mar Jul
Nekemte 2,080 18.3 24.0 12.6 4.1 11.4 Mar Jul
Robe Bale 2,480 14.8 21.6 8.0 2.2 13.6 Jun Dec
Sirinka 2,081 18.8 24.1 13.5 6.1 10.6 May Dec
Ziway 1,640 20.6 27.1 14 3.3 13.1 Apr Dec
H elevation, Tmean mean annual temperature, Tmax mean annual maximum temperature, Tmin mean annual minimum temperature, Ea mean annual
temperature excursion, Ed mean daily temperature excursion, Mh hottest month, and Mc coldest month
Fig. 3.2 Mean maximum annual

temperatures
Fig. 3.3 Mean minimum annual

temperatures
40
y = -0.0051x + 35.344
35 R² = 0.83
30 y = -0.0058x + 30.102
R² = 0.92
°C 25
20 Tmax
15 Tmean
10
Tmin
y = -0.0065x + 24.861
5 R² = 0.92
0
0 500 1000 1500 2000 2500 3000
H (m a.s.l.)
Fig. 3.4 Lapse rate for maximum mean and minimum mean annual temperatures
From the above considerations and Fig. 3.4, it is evident

that elevation is a main parameter influencing temperature in 3.7 Precipitation
Ethiopia and adjoining areas. The lapse rate is 5.8 °C every
1,000 m, and the mean annual 10 and 5 °C isotherms were To describe the rainfall characteristics of Ethiopia, the
estimated to be around 3,450 and 4,300 m a.s.l., respec- monthly data of 49 meteo-stations in Ethiopia and 9 in the
tively. The maximum and minimum temperature lapse rates adjoining territories of Eritrea, Djibouti, Somalia, and
are slightly different with 5.1 °C/1,000 m and 6.5 °C/ Sudan, and the daily data of 27 Ethiopian stations were used
1,000 m, respectively. (Table 3.2). The longest time series is that of Moyale cov-
Considerable differences characterize the diurnal and ering 73 years of almost uninterrupted daily data from 1935
annual temperature ranges (Fig. 3.5). The former is rather to 2008, whereas the average length of the Ethiopian time
large, especially in the central-western steppe and desert area series considered is 33 years (ranging between 20 and
(Fig. 3.6), varying between 10 and 15 °C; the latter is much 68 years), with only four stations covering a shorter interval
narrower, commonly less than 5 °C in the highlands areas (10–20 years). These latter were included as well in order to
below 2,000 m in elevation and in the tropical areas attain the larger and denser distribution of data as possible
(Fig. 3.6). In fact, sunny days prevail during most of the year for the spatial interpolations used in producing the annual
in almost the whole country, including the areas with a rainfall distribution map (Fig. 3.7).
monsoon climate; the alternation of cloudy and clear sky The annual precipitation, averaged all across the country,
conditions results in ample thermal differences (Fig. 3.5 and is 817 mm but, given the complex physiography and the
Table 3.1). By contrast, the annual thermal amplitude is different seasonal and spatial influences of the prevailing air
limited and seems to be controlled by the intertropical cir- masses and winds, a large diversity is observed among various
culation and the monsoon pattern rather than elevation. In regions. The highest annual rainfall is recorded at Gore, in the
the central-western highlands and in the Rift Valley, the western highlands at an altitude of 2,002 m a.s.l., with
hottest month is in spring and anticipates the summer, big 2,101 mm, whereas the lowest value in Ethiopia is found at
monsoon rains (kiremt), whereas in the northern and eastern Elidar with 153 mm. This rain gauge is located in the Afar
quadrants it occurs at the beginning of summer. triangle at an elevation of 423 m s.l.m., but the driest site is
In the highlands, the coldest month is in the middle of Asab, in Eritrea, with only 45 mm. Gore and Elidar are not the
winter, whereas in the southern and eastern regions, most and the least elevated stations, respectively, considered
including also high elevated areas, subjected to perturbed in this study; nevertheless, they share the extreme values of
conditions brought about by the monsoon and characterized annual rainfall. This is reflected by the poor correlation
by heavy downpours and cloudy skies, the coldest month (R2 = 0.42) between precipitation and elevation. Figure 3.7
occurs in summer. shows that higher precipitation occurs in the western part of
Fig. 3.5 Annual temperature range
Ethiopia and is partly irrespective of the station altitude. In precipitation (Fig. 3.11) may be higher than the kiremt,
Fig. 3.7 also, a westward gradient is rather evident too. Fol- contributing about 50 % of the annual rainfall such as at a
lowing these considerations, a physiographic factor (F) was few stations in the south and in Ogaden. Figure 3.9 shows
developed. It is the ratio of the meteo-station elevation (H) to also that the belg rains have little influence on the annual
L, i.e., the difference between its longitude and the 34th precipitation as they explain only 31 % of its variability.
meridian, taken as a western reference. Figure 3.8 shows that The spring rains are also the most unpredictable for their
85 % of annual precipitation variability is explained by the variation coefficient is 47 %, whereas the summer rains
physiographic parameter F = H/L. (CV = 33 %) result the most dependable for the farmers. Only
Since in two-thirds of the meteo-stations, the summer on the coast, namely at Djibouti and Massawa, the winter
rains (kiremt) contribute more than 50 % of the annual rains account for most of the annual precipitation. Along the
precipitation, with peaks of 80–85 % in the north, and 87 % Eritrean coast and in Ogaden, the largest variability of annual
of the annual precipitation variability is explained by the precipitation is found (Aseb CV = 127 %), whereas the least
kiremt (Fig. 3.9); the monsoon and its seasonal and geo- variability is observed in the central highlands (Table 3.2).
graphic pattern appear to be the most relevant factors in This observation is confirmed by the diagram of Fig. 3.12 in
controlling the spatial distribution and amount of rainfall in which the variation coefficient is well explained (R2 = 0.75)
Ethiopia (Fig. 3.10), though in some places the belg by the physiographic parameter F.
(a) Dire Dawa (Fig. 3.13). Hourly data for Mekele and Ziway show that
35 maximum intensities around 20 mm h−1, with peaks of
30 40–50 mm h−1, are rather common in these two localities.
Soil erosion is a major environmental and economic
25
problem in Ethiopia. In order to highlight the regions
°C
20 Tmin affected by severe hydro-meteoric erosion conditions, the

15
Tmax Modified Fournier Index (MFI) (Arnoldus 1980) was cal-
culated (Table 3.2) and the spatial distribution of this
10
parameter is mapped (Fig. 3.14). At two-thirds of the
5 weather stations considered, rainfall erosivity is from mod-
J F M A M J J A S O N D
erate to high and at 43 % of them it is high. The higher
(b) 35 Bahir Dar values are found in the central and part of the northern
highlands and, subordinately, in the central portion of the
30 Somali plateau margin. These areas are classified as sub-
25 jected to high and very high runoff by Berhanu et al. (2013)
20 and high erosion rates would be expected, but the occurrence
°C
Tmin
15 of soils with low erodibility, as reported by the USLE’s
Tmax
10 factor K map reported by these authors, partly mitigates this
threat.
5
0
3.8 Wind
Fig. 3.6 Mean maximum and minimum monthly temperatures at Dire
Dawa and Bahir Dar In Ethiopia, the wind regime is influenced by barometric
synoptical variations associated with the position of the
ITCZ. Winds are not strong because the Horn of Africa is
Rainfall intensity was investigated using daily data not a cyclone-genetic area and, as pointed out in Sect. 3.6,
recorded at 26 stations. The daily interval was selected temperatures are rather uniform throughout the whole year.
because, unfortunately, long time series of hourly data are not As a consequence, winds, though seldom absent, have low
available. Certainly, this is a strong constraint to sort out the velocities, and only during deep convective processes, short
regions with aggressive rains since commonly, also during but intense gusts and whirls are observed. Windiness is
the monsoon season, rainstorms are short (1–3 h) and a daily characterized by a large spatial variability due to the phys-
datum may be not sufficiently accurate to describe the rainfall iographic complexity which affects the prevailing seasonal
intensity parameter, especially in relation to the activation of air masses, especially along the Rift Valley and its margins.
soil erosion processes. Nevertheless, also using daily data, From the analysis of daily and, occasionally, 6 h data
some general results can be drawn. The mean maximum from a few meteo-stations located in the highlands, in Ti-
rainfall intensity (Imax) at each meteo-station was calculated gray, along the Rift Valley, and in Ogaden for the 2000–
by averaging the maximum values recorded in each year, 2006 interval (Fig. 3.15), it is evident that in the central-
whereas Ip is the highest value ever recorded by each rain northern part of the country, in a few sites of the Eritrean
gauge. Gambela is subjected to the highest rainfall intensities coast and the Sudan lowlands, winds from the northern
with an average of 89.4 mm in 24 h and a maximum value quadrant dominate, whereas in the southern areas, winds
ever recorded of 181.4 mm/24 h. By contrast, the lowest from the second and the third quadrants prevail throughout
values are observed in Elidar (Imax = 27.8 mm/24 h) and the whole year. On the margin of the Somali Plateau, wes-
Gewane (Ip = 31.6 mm/24 h), i.e., two rain gauges located in tern winds prevail, whereas in the far southwest of Ethiopia
the Afar depression characterized by desert and semidesert southwestern winds are more common. On the highlands,
climate conditions. Neither physiographic factor, such as the annual frequency of days with calms and light breezes
elevation or the geographic position, nor any pluviometric amounts to 12–18 % of the total.
parameter, such as annual rainfall, is able to explain the At seasonal level and particularly during the kiremt,
variability of rainfall intensity, which is likely more influ- noticeable differences from the above framework can be
enced by local orographic or morphologic factors. traced:
High intensity rains may occur in any month, but the 1. In winter, the situation described above is confirmed with
highest frequencies are recorded in April and August, i.e., in winds from the first quadrant and from northwest pre-
the middle of the spring and summer rainy seasons vailing on the central highlands and on the Somali
Table 3.2 Characteristic data of the rain gauges considered in this study
Meteo-stations H P (mm) CV (%) P (mm) P (mm) P (mm) k/a Imax Ip Tr 100 mm/ MFI
(m a.s.l.) annual annual belg kiremt bega % mm/24 h mm/24 h 24 h (years)
Adami Tulu 1,636 756.8 27 186.2 447.6 123.0 59.1 47.4 81.9 68.6 93.9
Addis Ababa 2,354 1,204.8 18 238.4 863.5 102.9 71.7 186.1
Adigudum 2,095 490.5 49 52.5 428.5 9.5 87.4 127.1
Al Damazin 470 691.0 56.0 596.0 39.0 86.3 137.3
Alamata 1,520 750.4 24 219.9 387.1 143.4 51.6 52.8 96.0 32.4 106.2
Alemaya 2,047 802.9 26 248.4 431.0 123.5 53.7 56.4 118.0 15.3 95.0
Arba Minch 1,285 931.9 17 375.9 236.6 319.4 25.4 99.7
Aseb 14 45.4 127 6.1 20.6 18.7 45.4 7.5
Asela 2,430 1,225.5 17 321.6 762.8 141.1 62.2 150.2
Asmara 2,325 512.3 34 87.8 375.7 48.8 73.3 109.7
Awasa 1,750 951.5 16 292.9 472.6 186.0 49.7 99.3
Azezo 1,966 1,153.0 146.0 914.0 93.0 79.3 220.7
Bahir Dar 1,770 1,413.2 17 105.0 1,189.8 118.4 84.2 293.7
Belet Uen 198 259.0 138.0 21.0 100.0 8.1 53.7
Bilate 1,500 744.2 23 240.3 326.9 176.9 43.9 78.9
Borciota 1,062 505.0 122.0 309.0 74.0 61.2 76.4
Burao 970 195.0 102.0 62.0 31.0 31.8 33.0
Butajira 2,000 1,118.1 19 377.7 560.6 179.8 50.1 121.1
Comar 764 920.0 90 785 45.0 85.3 180.5
Debre Marcos 2,515 1,325.5 11 208.9 970.1 146.5 73.2 204.4
Dengeco 2,111 750.5 22 249.6 429.0 71.9 57.2 53.8 98.8 21.0 103.6
Derbiga el Agemsa 1,644 678.0 239.0 340.0 99.0 50.1 80.4
Desse 2,250 1,207.1 16 244.6 789.5 173.0 65.4 54.9 94.0 47.0 189.4
Dire Dawa 1,260 638.7 27 220.2 313.6 104.9 49.1 54.8 122.3 18.1 78.3
Djibouti 21 188.0 71 41.59 17.52 128.9 9.3 24.8
Elidar 423 153.1 40 29.0 99.1 25.0 64.7 27.8 55.0 26.5
Fitche 2,750 1,125.2 12 183.0 826.0 116.2 73.4 51.4 90.9 42.8 226.2
Gambela 484 1,114.1 22 187.3 697.6 229.2 62.6 89.4 181.4 139.7
Gedaref 599 579.7 16 15.8 525.2 38.7 90.6 146.6
Gewane 627 521.9 21 135.3 268.7 117.9 51.5 29.1 31.6 71.8
Gina Ager 3,160 1,693.2 20 415.3 977.0 300.9 57.7 74.3 170.8 6.4 216.9
Goba 2,743 943.7 20 292.4 412.2 239.1 43.7 98.9
Gode 295 236.9 61 135.0 8.4 93.5 3.6 58.4 174.0 7.0 50.7
Gonder 1,967 1,131.9 18 139.4 878.4 114.1 77.6 52.4 99.1 40.9 206.7
Gore 2,002 2,101.3 23 439.7 1,261.9 399.7 60.1 62.1 107.7 24.9 247.4
Hargeysa 1,347 350.0 120.0 210.0 20.0 60.0 50.6
Hayk 2,030 1,174.0 17 280.9 717.1 176.0 61.1 62.1 132.8 16.5 176.4
Jijiga 1,644 721.3 39 249.9 329.2 142.3 45.6 75.6
Jimma 1,725 1,500.5 13 401.1 821.0 278.4 54.7 53.0 105.7 39.5 164.4
Kassala 502 278.6 43 13.0 229.0 36.6 82.2 49.0 104.9 56.2
Kebri Dehar 549 325.3 49 170.4 20.6 134.3 6.3 59.4 128.0 11.8 77.7
Kobo 1,610 725.8 22 151.5 411.8 162.5 56.7 53.2 101.5 18.3 97.7
Koka Dam 1,595 698.8 48 135.6 496.0 67.3 71.0 112.3
Kombolcha 1,903 1,043.6 15 221.3 663.1 159.2 63.5 160.9
Konso 1,053 805.6 14 358.2 153.0 294.4 19.0 55.2 96.9 28.4 86.2
(continued)
Table 3.2 (continued)

Meteo-stations H P (mm) CV (%) P (mm) P (mm) P (mm) k/a Imax Ip Tr 100 mm/ MFI
(m a.s.l.) annual annual belg kiremt bega % mm/24 h mm/24 h 24 h (years)
Kulubi 2,410 1,022.3 18 308.7 565.6 148.0 55.3 60.2 100.5 14.4 125.6
Mandera 231 286.7 56 159.4 4.5 122.8 1.6 59.6
Massawa 10 175.8 58 22.5 19.6 133.7 11.1 24.0
Mekele 2,070 598.7 27 91.3 487.7 19.7 81.5 49.4 95.5 59.6 162.6
Meki 1,660 771.2 21 185.6 482.1 103.5 62.5 103.6
Metahara 947 549.2 23 140.2 322.4 86.6 58.7 52.3 96.0 31.5 78.0
Moyale 1,097 685.0 40 352.5 63.4 269.1 9.3 68.7 167.6 6.3 104.1
Negelle 1,544 726.6 27 414.5 52.7 259.4 7.3 66.1 137.0 8.2 128.6
Nekemte 2,080 2,037.7 12 369.3 1,429.0 239.4 70.1 75.9 137.5 7.0 294.9
Robe Bale 2,480 876.3 16 281.0 380.1 215.2 43.4 43.3 112.3 8.7 92.1
Sirinka 2,081 1,128.9 14 206.1 361.3 561.5 32.0 161.8
Ziway 1,640 742.5 24 201.5 442.1 98.9 59.5 89.3
Zuquala 2,980 1,124.8 22 251.8 763.3 109.7 67.9 68.0 132.6 8.7 178.3
H elevation, P precipitation, CV variation coefficient, k/a contribution of the kiremt rainfall to the annual precipitation, Imax mean maximum rainfall intensity in 24 h, Ip
highest rainfall intensity in 24 h ever recorded, Tr return time for a rainfall intensity of 100 mm in 24 h, and MFI Modified Fournier Index Arnoldus 1980)
Fig. 3.7 Distribution of annual precipitation

2500 plateau margin, respectively. Elsewhere, southeastern

and southwestern winds prevail.
2000
2. In spring, only small changes occur. In fact, the western
P (mmyr -1 )
1500 winds rise in latitude along the Sudanese border as far as

Tana Lake, whereas in the southernmost areas they are
1000 y = 37.642x 0.5458 replaced by winds from the second quadrant. On the
R² = 0.85 Somali plateau margin and in Ogaden northeastern winds
500
predominate.
0
3. During summer, wind circulation turns eastward on the
0 200 400 600 800 1000 1200 1400 central-northern highlands, whereas such orientation is
F
not so evident in the area of Addis Ababa and in the Rift,
Fig. 3.8 Correlation between annual precipitation (P) and the and finally to northwest in Eritrea (e.g., Asmara).
physiography factor (F—see text for explanation) Southwestern winds predominate on the Sudanese low-
lands and as far as Gambela where winds from east and
southeast prevail. These winds are present also in the
2500
y = 3.7944x + 272.61 y = 2.5191x + 346.28 southern part of the country as far as the Somali plateau
R² = 0.84 R² = 0.31 margin (e.g., Harar).
2000
4. In autumn, the part of Ethiopia to the west of an ideal
Annual P (mm)
1500
Gonder-Nekemte line and the southern sectors are sub-
Kiremt
jected to western air masses, whereas in other parts of the
Bega
1000 country eastern air masses prevail with the exception of
Belg
the Harar region where northwestern winds are more
y = 1.3332x + 248.38
500
R² = 0.87
common.
On the highlands, mean wind speed is typically low
(12 km h−1) but increases in the Great Lakes Region, in
0
0 200 400 600 800 1000 1200 1400
Season P (mm) the sub-desert and bush lands of Ogaden and around the
border with Sudan, especially during the kiremt (EMA
Fig. 3.9 Correlation between annual precipitation (P) and the
cumulative precipitation of kiremt and belg main rainy seasons and
1988).
bega, the dry spell
Fig. 3.10 Distribution of the

summer, monsoon-type rains
(kiremt in the local language)
Fig. 3.11 Distribution of the little, spring rains (belg in the local language)
0.7 25
0.6
20
0.5
Frequency (%)
y =-0.128ln(x) + 0.9833
0.4
PCV
R² = 0.75 15
0.3 Imax
0.2 10 Ip
0.1
5
0
0 200 400 600 800 1000 1200 1400
F 0
Fig. 3.12 Correlation between the annual rainfall variation coefficient
(PCV) and the physiographic factor (F—see text fro explanation) Fig. 3.13 Monthly distribution of the frequency of occurrence of the
annual maximum rainfall intensity in 24 h (Imax) and of the highest
value (Ip) recorded throughout the period of recording for all the meteo-
stations considered in this study
Fig. 3.14 Spatial distribution of

rainfall erosivity expressed by the
MFI (Arnoldus 1980)
The most part of Ethiopia—and in particular the north—

45
is subjected to the influence of trade winds commonly rich in
40
water vapor. Obviously, the spatial distribution of humidity
35
is tightly correlated with precipitation. In the southwest of
30 AWASA
Ethiopia, areas with very high humidity, among the most
kmh -1
25 GORE
humid in the African continent (e.g., the great equatorial
20 GODE
lakes region), are found.
15 ADDIS ABABA
By contrast, during winter, high values are recorded in
10 GONDER
Ogaden and along the Red Sea, whereas low humidity
5
characterizes the Sudanese border belt with mean values
0 around 23 %.
The data recorded in 13 meteo-stations during the last
Fig. 3.15 Monthly variation of average wind speed for a few meteo- decade indicate that average annual relative humidity is
stations about 60 % on the highlands, in the Rift and along the Kenya
border (Fig. 3.16), but even higher values are presumed for
very high mountain areas (EMA 1988). Lower values
between 40 and 50 % are instead recorded in the northern-
3.9 Relative Humidity most portion of the Rift (e.g., Dire Dawa—Fig. 3.16).
Only inadequate data of relative humidity are available since

few stations have been measuring this parameter in the last 3.10 Cloudiness and Sunshine
decades. Notwithstanding this constraint, combining the
available data with the maps of the National Atlas (1988), Ethiopia is the country with the most extensive cloudiness in
some considerations on the spatial distribution of this Africa. Unfortunately, only very few data of Fantoli (1940)
parameter can be drawn. are available and refer to the first decades of the twentieth
90 3.11 Evapotranspiration and Aridity

80
70 Evapotranspiration is limited by soil moisture supply, sur-
60 face and atmospheric temperature and the dew point
ADDIS ABABA
threshold (Law et al. 2002). The coupling of evapotranspi-
Hr (%)
50
DIRE DAWA
40 ration, temperature, and precipitation is particularly pro-
BAHIR DAR
30 nounced in moisture-limited subtropical regions at the
DAMAZINE
20 interface between wet (monsoonal) and dry climate regimes.
10
Potential evapotranspiration (PET) was calculated fol-
0
lowing the Thorntwaite’s method, using the data from 29
J F M A M J J A S O N D meteo-stations for the 1981–2010 time range (Fig. 3.18). The
same data were used also to reckon the aridity index (Ai)
Fig. 3.16 Monthly distribution of mean relative humidity for a few
representative meteo-stations
(Fig. 3.19), known also as the desertification index (UNEP
1992), as Ai = P/PET in which P is annual precipitation.
Almost all the areas over 2,000 m a.s.l. are characterized
12 by annual PET values less than 1,000 mm, whereas pre-
ADDIS ABABA
cipitation is typically around 1,500 mm or higher. Therefore,
10 BAHIR DAR
the highlands can be classified as humid and hyper-humid in
AWASA
8 the Gore–Jimma–Tepi triangle (Fig. 3.18). As elevation
S h (hours)
MEKELE
decreases, PET commonly increases as it is observed in the
6 KEBRY DEHAR
Rift Valley. Extreme evapotranspiration conditions are
GORE
4 found in the Danakil depression where values of as much as
NEGHELLE
4,000 mm can be estimated for Dallol. This figure, combined
2 MOYALE
with inappreciable precipitation and extremely hot temper-
DAMAZINE atures, makes this area as one of the most arid on the planet.
0
ASMARA Ai values less than 0.65 are typical of drylands (Thomas
and Middleton 1994) and UNEP (1992) takes 0.65 as the
Fig. 3.17 Monthly distribution of mean sunshine hours for a few threshold value to identify desertification prone areas. Large
representative meteo-stations
portions of Ethiopia fall in this category: the whole Ogaden,
the Afar triangle, the Danakil depression, the northern por-
century. According to this author, average cloudiness is 2/10 tion of the Rift Valley, and the belt across the border with
in January and 6/10 in July, whereas clear sky days are a Sudan (Fig. 3.19).
little less frequent than mixed and cloudy ones (170 vs 195).
The maximum and minimum cloudiness occurs during the
rainy season and the boreal winter, respectively. In Ogaden 3.12 Climate Classification
and mainly in the Danakil depression, cloudiness is almost
an exceptional phenomenon. Along the Eritrean coast and As highlighted in Sect. 3.3, the climate of Ethiopia is mainly
for a few kilometers inland, night advection fogs may controlled by the seasonal migration of the ITCZ and by the
commonly form and turn into morning drizzle, integrating complex topography of the country. This combination
the scarce precipitation of this area. results in a variety of climate types and climate conditions
Sunshine is obviously complementary to cloudiness. that may vary also within a short distance (Fig. 3.20).
Monthly data of sunshine hours (Sh) measured by nine sta- A tight relation is evident between elevation and climate
tions in the last decade (Fig. 3.17) show that most of the type, especially along the plateau margins. East of this
highlands and the lowlands close to the Sudan border receive margin and in the southeastern portion of the country, a
more than 7 h of mean daily sunshine, whereas from Tigray predominantly arid and very hot climate (tropical A, Köppen
to the southeast of the country the respective value increases classification) fades, throughout a belt of climatic transition
to 10 h. During autumn and winter, 300–350 h per month are at higher elevations, into a temperate C climate, character-
commonly recorded in Tigray and Ogaden. In summer, vice ized by mild temperatures, higher than 18 °C in every
versa, in the highlands between Gonder and Addis Ababa, month, and by rainfalls with a maximum in summer (Tigray
the sunshine commonly does not exceed 50 h per month. and Somali plateau margin), in spring and a secondary

potential evapotranspiration

the aridity index (UNEP 1992)
Fig. 3.20 Map of the Köppen

climate classification
maximum in autumn (southern portion of the country) rather Over 2,600 m a.s.l. and in a few areas of the Semen and
than constant precipitation throughout the four seasons (type Bale Mountains, the climate becomes relatively cold with
AM of the southwestern areas). The thermal transition from mean monthly temperatures less than 10 °C. At elevations
Af, Aw, or BSh to the temperate C climate occurs at an higher than 3,300 m a.s.l., the climate is characterized by very
elevation ranging between 1,800 and 2,000 m a.s.l., though low temperatures and it is classified as type H which includes
in the northernmost part of the highlands the climate the peaks of the intertropical massifs. Here, vegetation con-
becomes cooler and remains arid also at high elevations (e.g. sists mainly of alpine prairie and arboreal species is absent.
Asmara, type BSh). In the areas with a temperate climate, the
temperature ranges are never large and the minimum mean
temperatures are recorded in summer, during the big, mon- 3.13 Climate Change
soon-type rains, whereas the milder months are in spring.
The prevailing climate is Cw, with a patent scarcity of pre- Climate change is one of the most investigated issues of
cipitation during the winter semester (bega). Ethiopian climate since its rain-fed agriculture is largely
In the tropical climate region with a dry season in winter dependent on the amount and regular onset of seasonal
(Aw), areas characterized by the maximum temperature in precipitation (Hadgu et al. 2013). Throughout the last cen-
spring—just before the beginning of the kiremt—and the tury, temperature and rainfall trend lines indicate a climatic
minimum temperature during the perturbed phase (July or anomaly that cannot be associated with ordinary cyclic
August)—climate sub-types Awg or Awt—are also found oscillations of these parameters.
(e.g., Jimma, Gambela, and Gonder). In the Rift Valley (e.g., In all the meteo-stations considered, an average temper-
Awasa, Ziway), these conditions are paired by a particularly ature increase of 1.1 °C is recorded for the 1981–2010
small annual range (less than 3 °C) and the climate is Awi. interval (0.04 °C per year). The same increasing rate is
In places of the southwestern portion of the country (e.g., observed for the mean minimum and maximum tempera-
Nekemte and Bonga), rainfalls are rather uniformly distrib- tures. The highest rates are recorded at Neghelle (2.7 °C in
uted in all months (cool variant of the tropical climate) and 30 years), Gonder and Robe Bale (2 °C in 30 years). Such
determine the Cfc climate. marked positive trends do not seem to be influenced by the
(a) Precipitation trends are very difficult to analyze since the

1.5
time series considered are punctuated by gaps and, some-
1.0
times, the data of 3–5 years in a row are missing. As
0.5 discussed in Sect. 3.5, any attempt to fill the gaps with a
correlation procedure failed to give reliable results. More-
Z (Tmax)
0.0
-0.5 over, the available time series span different intervals which
-1.0
start and end in different years. Nevertheless, some general,
statistically nonsignificant tendency can be depicted on the
-1.5
basis of 20 of the longest (29–46 years) time series
-2.0
selected. 12 stations out of 20 show a negative trend;
-2.5
however, in order to reduce the weight of the gaps scat-
1953
1956
1959
1962
1965
1968
1971
1974
1977
1980
1983
1986
1989
1992
1995
1998
2001
2004
2007
tered across the data set, the trend analysis has been
(b) restricted to the 1961–2009 interval and a representative
1.5
time series has been constructed by averaging the annual
1.0 precipitation at all the stations for each year. The rainfall
0.5 anomalies with respect to the long-term mean are plotted
and a negative trend of about 2.6 mm per year is observed
Z (Tmin)
0.0
-0.5
(Fig. 3.23). This diagram confirms that the 1980s and the
first decades of the twenty-first century were characterized
-1.0
by recurrent droughts that caused the spread of severe
-1.5
famines from 1983 to 1985, known as the Great Famine
-2.0 (1 mln fatalities), and another one in 2003 (Bewket and
-2.5 Conway 2007).
1953
1956
1959
1962
1965
1968
1971
1974
1977
1980
1983
1986
1989
1992
1995
1998
2001
2004
2007
The normalized data of the kiremt (Fig. 3.24 a) and belg

(Fig. 3.24 b) rains over the last five decades show that the
Fig. 3.21 Variation through time of normalized mean maximum spring rains are decreasing at a rate 1.5 times faster than the
(a) and minimum (b) temperatures. Dotted line is the 5 years moving
mean
monsoon summer rains. As discussed above, the belg rains
are less voluminous, less reliable, and poorly predictable,
and though agriculture in Ethiopia rely mainly on the sum-
local physiography nor by the expansion of the urban areas. mer rains, the decreasing trend of the spring rains may pose
In fact, at Addis Ababa, the temperature change is negligible severe constraints to the economic development of the
(0.15 °C/30 years) as are the modest decreasing rates mea- country.
sured at Mekele, Alamata, and Arba Minch (0.1–0.2 °C/ The 20 meteo-stations selected are too small in number to
30 years). sort out any significant regional pattern of change in pre-
Eight meteo-stations have longer time series starting in cipitation as neighboring stations may have contrasting
the 1950s (Fig. 3.21). The analysis of their normalized data trends (e.g., Gore and Jimma, the trend line angular coeffi-
shows a higher increasing rate as far as the 1980s, whereas cient of which is −16.2 and +1.4, respectively).
during the last 15–18 years the thermal signal becomes Very few stations have data suitable for an analysis of
stable at 0.6–0.8 °C above the semi-secular average. It is also maximum rainfall intensity variation with time. However, it
worth noticing that for these stations, the mean temperature seems that no significant change occurred during the 1961–
increase is mainly due to the contribution of the minimum 2009 interval.
temperatures that increase at a rate twice as much the Older data are available only for very few stations and
maximum temperatures. refer mainly to the study of Fantoli (1940) who worked on
Figure 3.22 shows the individual temperature changes data ranging approximately from 1890 to 1936, with dif-
recorded by the same eight meteo-stations for two periods: ferent lengths and time intervals, but for some stations the
1953–1980 and 1981–2010. At Addis Ababa, the mean beginning of the measurements is not indicated. Any com-
temperature reported by Fantoli (1940) for three decades at parison with this author’s data is therefore speculative since
the beginning of the twentieth century is the same (16.4 °C) neither information is provided about the exact location of
as that of the 1953–2010 interval, but the former have a the rain gauge nor and whether it was moved during the
much larger daily excursion, i.e., 15.2 versus 13.2 °C. operation period as discussed by Conway et al. (2004).
Mean temperature
1953-1980 1981-2010
24.7 25.5
20.5
18.1 18.8 18.6 19.5 18.8 19.3 19.1
16.0 16.9 16.1 16.8 15.6 16.6
ADDIS ABABA ALEMAYA DEBRE DIRE DAWA GORE JIMMA KOMBOLCHA NEGELLE
MARCOS
Fig. 3.22 Mean annual temperatures recorded at eight meteo-stations with long data records relative to two distinct intervals, 1953–1980 and
1981–2010
400 (a) Kiremt

1.5
300
1.0
200
Pa (mmyr-1)
0.5
100
Z (P)
0.0
0
-0.5
-100 y = -0.0057x + 0.106
-1.0
-200
y = -2.5602x + 63.964 -1.5
1961
1963
1965
1967
1969
1971
1973
1975
1977
1979
1981
1983
1985
1987
1989
1991
1993
1995
1997
1999
2001
2003
2005
2007
2009
-300
1961
1963
1965
1967
1969
1971
1973
1975
1977
1979
1981
1983
1985
1987
1989
1991
1993
1995
1997
1999
2001
2003
2005
2007
2009
(b) Belg
1.5
Fig. 3.23 Rainfall anomalies obtained by averaging the annual
precipitation of all the meteo-stations considered (Pa). Dashed line is 1.0
the 5 years moving mean
0.5
Z (P)
However, just for a qualitative reference, the differences in 0.0

annual precipitation for the 1890–1936 and the 1960–2009 - 0.5
intervals recorded by ten meteo-stations are reported in
Table 3.3. In six out of ten stations, a moderate to marked - 1.0
y = -0.0083x + 0.1604
decrease in rainfall emerges, with an average value of - 1.5
1965
1967
1969
1971
1973
1975
1977
1979
1981
1983
1985
1987
1989
1991
1993
1995
1997
1999
2001
2003
2005
2007
2009
−81.1 mm, and in two of them (Adami Tulu and Kassala), it
is very pronounced (around −176 mm). The average rainfall
increase of the other four stations is 58.4 mm, but at Gam- Fig. 3.24 Variation through time of normalized kiremt (a) and belg
bela, a maximum difference of +175 mm is observed. (b) precipitation
Also for the analysis of rainfall intensity change over a
long time, only the data of Fantoli (1940) were used. the reasons explained above, is highly questionable and is to
Unfortunately, only five stations can be used for a compar- be taken just as a mere indication. Though Fantoli (1940)
ison with the modern data of this study (Table 3.4). Two provides some data on other climate parameters such as
stations, Adami Tulu and Gore, show a modest decrease relative humidity, evaporation and wind speed, they refer to
between the two periods with −5.9 and −19.3 mm in 24 h, very few stations and any comparison with modern data is
respectively, and are paired by a small increase at Dese and totally insignificant.
Kassala, 7.0 and 15.5 mm/24 h, respectively. The most Finally, within this framework of climatic change, it is
relevant change is observed at Gambela with an increase of also worth noticing that a few areas, such as the Rift Valley,
about 100 mm in 24 h. The significance of these results, for characterized by aridity index values in the 0.65–0.8 range,
Table 3.3 Mean annual precipitation difference between the old data Table 3.4 Maximum daily rainfall comparison between the old data of
of Fantoli (1940) and the present study Fantoli (1940) and the present study
Meteo-stations 1890–1936 1960–2009 Difference 1890–1936 1960–2009 Difference
Adami Tulu 756.8 581.1 −175.7 Adami Tulu 87.8 81.9 −5.9
Addis Ababa 1,204.8 1,251.7 46.9 Dese 87.0 94.0 7.0
Asmara 512.3 491.7 −20.6 Gambela 81.3 181.4 100.1
Belet Uen 259.0 260.1 1.1 Gore 127.0 107.7 −19.3
Djibouti 188.0 126.0 −62.0 Kassala 89.4 104.9 15.5
Gambela 1,114.1 1,289.5 175.4
Gore 2,101.3 2,059.3 −42.0
High rainfall intensities in 24 h may occur every month
Kassala 500.0 322.6 −177.4 but are more common in the middle of the two main rainy
Massawa 175.8 186.0 10.2 seasons. The highest average and absolute rainfall intensities
Moyale 685.0 675.9 −9.1 are recorded at Gambela (89.4 and 181.4 mm/24 h, respec-
tively), whereas the lowest values are measured at Elidar and
Gewane in the Afar triangle. The MFI shows that rainfalls
in the last two decades, frequently experienced desertifica- are very aggressive over large part of Ethiopia; the higher
tion conditions with Ai values of less than 0.5. values of rainfall erosivity are found in the central and part
of the northern highlands and, subordinately, in the central
portion of the Somali plateau margin.
3.14 Concluding Remarks The southwest portion of Ethiopia is characterized by very
high relative humidity values, among the highest in the whole
The climate of Ethiopia, though mainly controlled by the African continent, whereas the drier areas are in the northern
position of the ITCZ, is highly variable in space and time. part of the Rift. In the last three decades, the aridity index of
The large contrasts in elevation and morphology that char- this latter area, Ogaden, Afar triangle, Danakil depression and
acterize the country are very effective in controlling local the belt across the border with Sudan, commonly decreased
climate conditions to such an extent that even sites within a beyond the critical value of 0.65; hence, these regions have to
short distance from one another may show very different be considered as desertification prone areas.
climatic conditions. In fact, even though mean, maximum Among the climatic parameters considered in this study,
and minimum temperatures show an almost constant lapse long time series were available only for temperature and
rate, no significant correlation was found for rainfall at precipitation data. The change of both maximum and mini-
national level. Annual rainfall does not change significantly mum temperature through the last three decades is rather
with elevation and seems to be more influenced by the evident with an average increase of 1.1 °C. The data analysis
geographic position of the rain gauge with respect to the of the eight weather stations with longer time series (starting
Sudanese lowlands. The introduction of the F factor, which in the 1950s) shows a marked temperature increase as far as
combines elevation with a specifically developed longitude the 1980s, whereas its rate slowed down and has become
parameter, improved remarkably the statistical explanation almost stable in the last 15–18 years. In most of the stations
of rainfall variability in terms of geographic position. considered, it is the minimum temperature, rather than the
The large orographic variability results in a wide range of maximum temperature that contributes to the mean temper-
temperatures that may be as low as −7 °C in the Bale ature increase. The former increases at a rate twice as much
Mountains and as much as 50 °C in the Danakil Desert. higher than the latter. For the eight stations with longer time
The higher rainfalls are recorded in the western high- series, the comparison between 1953–1980 and 1981–2010
lands, with 2,101 mm year−1 at Gore, and the lower in the intervals indicates that in the most recent decades, mean
Afar depression with 145 mm year−1, though even lower temperatures show an average increase of about 0.9 °C. By
values are expected for the desert areas of the Danakil. The contrast, annual precipitation is characterized by a decreas-
summer, monsoon rains account for more than 50 % and as ing trend with an average decline of about 125 mm in
much as 85 % of the annual rainfall, whereas the spring rains 49 years, with the small spring rains decreasing at a rate 1.5
are more variable in amount and predictability. time higher than the monsoon summer rains.
References Hurst H, Black P (1937) The Nile basin vol VI. Monthly and annual
rainfall totals and the number of rainy days at stations in and near
the Nile basin for the period ending 1937. Ministry of Public Works
Alexandersson H (1986) A homogeneity test applied to precipitation Press, Cairo
data. J Clim 6(6):661–675 Jury MR, Funk CR (2013) Climatic trends over Ethiopia: regional
Arnoldus HM (1980) An approximation of the rainfall factor in the signals and drivers. Int J Climatol 33:1924–1935
universal soil loss equation. In: de Boodts M, Gabriels D (eds) Korecha D, Barnston AG (2007) Predictability of June–September
Assessments of erosion. Wiley, Chichester, pp 127–132 rainfall in Ethiopia. Mon Weather Rev 135(2):628–650
Beltrando G, Camberlin P (1993) Interannual variability of rainfall in Law B, Falge E, Baldocchi D, Bakwin P, Berbigier P, Davis K, Dolman
the eastern Horn of Africa and indicators of atmospheric circula- A, Falk M, Fuentes J, Goldstein A, Granier A, Grelle A, Hollinger
tions. Int J Climatol 13:533–546 D, Janssens I, Jarvis P, Jensen N, Katul G, Mahli Y, Matteucci G,
Berhanu B, Melesse AM, Seleshi Y (2013) GIS-based hydrological Monson R, Munger W, Oechel W, Olson R, Pilegaard K, Paw UK,
zones and soil geo-database of Ethiopia. Catena 104:21–31 Thorgeirsson H, Valentini R, Verma S, Vesala T, Wilson K, Wofsy
Bewket W, Conway D (2007) A note on the temporal and spatial S (2002) Environmental controls over carbon dioxide and water
variability of rainfall in the drought-prone Amhara region of vapor exchange of terrestrial vegetation. Agric For Meteorol
Ethiopia. Int J Climatol 27:1467–1477 113:97–120
Chernet T (1982) Hydrogeology of the Lakes Region, Ethiopia. Lovett R, Wood CA (1976) Rainfall reliability in Ethiopia. Weather
Memoir N.7, the Provisional Military Government of Socialist 31:417–424
Ethiopia, Addis Ababa McSweeney C, New M, Lizcano G (2010) UNDP climate change
Cheung WH, Senay GB, Singh A (2008) Trends and spatial distribution profile: Ethiopia. http://countryprofiles.geog.ox.ac.uk
of annual and seasonal rainfall in Ethiopia. Int J Climatol 28:1723– Mekasha A, Tesfaye K, Duncan AJ (2013) Trends in daily observed
1734 temperature and precipitation extremes over three Ethiopian eco-
Conway D (2000a) Some aspects of climate variability in the north east environments. Int J Climatol. doi:10.1002/joc.3816
Ethiopian Highlands—Wollo and Tigray. SINET—Ethiop J Sci Mengistu D, Bewket W, Lal R (2013) Recent spatiotemporal temper-
23:139–161 ature and rainfall variability and trends over the upper blue Nile
Conway D (2000b) Climate and hydrology of the upper blue Nile river. river basin. Int J Climatol, Ethiopia. doi:10.1002/joc.3837
Geogr J 166(1):49–62 Pedgley DE (1967) Air temperature at Dallol, Ethiopia. Meteorol Mag
Conway D, Mould C, Bewket W (2004) Over one century of rainfall 96(1967):265–271
and temperature observations in Addis Ababa, Ethiopia. Int J Segele ZT, Lamb PJ (2005) Characterization and variability of Kiremt
Climatol 24:77–91 rainy season over Ethiopia. Meteorol Atmos Phys 89(1):153–180
Easterling DR, Evans JL, Groisman PYa, Karl TR, Kunkel KE, Seleshi Y, Camberlin P (2006) Recent changes in dry spell and extreme
Ambenje P (2000) Observed variability and trends in extreme rainfall events in Ethiopia. Theoret Appl Climatol 83:181–191
climate events. Bull Am Meteor Soc 81:417–425 Seleshi Y, Zanke U (2004) Recent changes in rainfall and rainy days in
Enku T, Melesse AM (2013) A simple temperature method for the Ethiopia. Int J Climatol 24:973–983
estimation of evapotranspiration. Hydrol Process. doi:10.1002/hyp. Shanko D, Camberlin P (1998) The effects of the southwest Indian
9844 Ocean tropical cyclones on Ethiopian drought. Int J Climatol
ENRAEMED (2003) Ethiopian natural resources and environmental 18:1373–1388
meta-database. http://clearinghouse5.fgdc.gov/enraemed/NMSA. Tadege A (ed) (2001) Initial national communication of Ethiopia to the
php United Nations Framework Convention on Climate Change (UN-
Eredia F, DeCastro L (1914) Sula climatologia dell’Etiopia. Boll Real FCCC). National Meteorological Services Agency, Addis Ababa
Soc Geog Ser V III(8):843–884 Tadege A (ed) (2007) Climate change national adaptation programme
Ethiopian Mapping Authority (EMA) (1988) The National Atlas of of action (NAPA) of Ethiopia. National Meteorological Services
Ethiopia. EMA, Addis Ababa Agency, Addis Ababa
Fantoli A (1940) Elementi preliminari del clima dell’Etiopia. Sansoni, Taye M, Zewdu F (2012) Spatio-temporal Variability and Trend of
Firenze Rainfall and Temperature in Western Amhara, Ethiopia: a GIS
Fantoli A (1965) Contributo alla climatologia dell’Etiopia. O.P.I, Roma approach. Glob Adv Res J Geogr Reg Plann 1(4):65–82
FAO (1984) Agroclimatological data for Africa, vol 1: Countries North Thomas DSG, Middleton NJ (1994) Desertification: exploding the
of the Equator. FAO Plant Production and Protection Series No. 22, myth. Wiley, Chichester
FAO, Rome UNEP (1992) Status of desertification and implementation of the
Gamachu D (1977) Aspects of climate and water budget in Ethiopia. United Nations plan of action to combat desertification. Report of
Addis Ababa University Press, Addis Ababa the Executive Director, United Nations Environment Programme,
Gamachu D (1988) Some patterns of altitudinal variation of climatic Nairobi
elements in the mountainous regions of Ethiopia. Mount Res Dev 8 Vinassa de Regny P (1931) La geologia delle Alpi Dancale. Bollett Soc
(2/3):131–138 Geol Ital 50(1):1–24
Hadgu G, Tesfaye K, Mamo G, Kassa B (2013) Trend and variability of Viste E, Korecha D, Sorteberg A (2013) Recent drought and
rainfall in Tigray, Northern Ethiopia: analysis of meteorological precipitation tendencies in Ethiopia. Theoret Appl Climatol 112
data and farmers’ perception. Academia J Env Sci 1(8):159–171 (3–4):535–551
Houghton JT, Ding Y, Griggs DJ, Noguer M, van der Linden PJ, WWR (1959) World weather records 1941–50. US Department of
Xiaosu D (eds) (2001) Climate change 2001: the scientific basis: Commerce, Weather Bureau, Washington, DC
contributions of working group I to the third assessment report of
the intergovernmental panel on climate change. Cambridge Uni-
versity Press, Cambridge
Ethiopian Rivers
4
Paolo Billi, Semunesh Golla, and Dawit Tefferra
Abstract
The scientific literature about the geomorphology and hydrology of Ethiopian rivers is very
poor, though large and socially important rivers have their source in this country. The largest
rivers deliver their waters into the Mediterranean Sea or the Indian Ocean, whereas the most of
the smaller ones have an endorheic drainage. Flow data were collected from different sources,
and the longest as possible time series were considered. A new index, the runoff concentration
index is used by analogy with the precipitation concentration index to analyze the monthly
variation of runoff. Characteristic discharges such as bankfull discharge or longer return time
discharges are calculated as well. Due to the remarkable variety of physiography of Ethiopia
landscapes and recent tectonics, rivers show different drainage network and channel
morphology. Their main features are described, and recent geomorphological changes are
analyzed. Flow data are investigated and a number of correlations with catchment parameters,
capable also to predict discharge in ungauged rivers, are reported. Runoff changes through
time do not show any significant trend, whereas the sediment yield of Ethiopian river is of the
same order of magnitude of much larger African rivers indicating a high soil erosion rate of
Ethiopian highlands.

Keywords
Hydrography Channel morphology Discharge Runoff Sediment transport
4.1 Introduction slopes. The morphological, hydrological, and hydraulic

characteristics of rivers differ in space and time and local
Rivers are among the most powerful, effective, and ubiqui- geological, climatic, and, recently, human factors make each
tous geomorphic agents, capable to shape the Earth surface river a unique physical system. Rivers are a resource and a
and to produce a large variety of landscapes. Rivers are hazard at the same time. Their political, social, economic, and
present and active all over the world, draining almost 70 % of environmental relevance are patent to everyone and both
the Earth land surface (Petts 1983) and encompassing almost industrialized and developing countries have benefited from
all environments from cold polar regions to hot drylands. their sustainable management. With the recent climate
That is the result of the deployment of a huge amount of changes and a growing human impact, rivers became central
energy that rivers use to convey runoff, generated by excess to many environmental problems and in many regions, the
precipitation, and sediment supplied by erosions processes on resulting increased risks, associated with extreme low or high
flows, are posing serious constraints forcing land managers
and decision makers to rethink the current approach to urban
P. Billi (&) and rural areas planning within a short timescale.
Physics and Earth Sciences, University of Ferrara, Ferrara, Italy Nowadays, the scientific literature on rivers, their charac-
e-mail: bli@unife.it
teristics, and processes is extensive and studies on river
S. Golla D. Tefferra morphology, sedimentology and hydrology have been carried
Ministry of Water, Irrigation and Energy, Addis Ababa, Ethiopia

90 P. Billi et al.
out in almost any region of the world (see Knighton 1993; Ocean. The river systems in between the former, Danakil,
Bridge 2003—for a comprehensive reference list). By con- Awash, Rift Valley, and Omo, are instead closed basin.
trast, information about the geomorphology and hydrology of Other small endorheic systems are found within the western
Ethiopian rivers is almost nonexistent and generally limited to main basins and are associated with the drainage into small
the main Ethiopian river, the Blue Nile (Abay in the local lakes such as Ashange or structural basins such as the Kobo–
language) (Conway 2000; Tesemma 2010). A relatively larger Alamata basin (see Chap. 19).
number of studies were instead carried out on much smaller Within the borders of Ethiopia are included also a small
streams in order to investigate other important aspects such as portion of the Mereb River headwaters (the small green area
sediment transport (Billi 2000, 2004, 2011), sediment yield close to the border with Eritrea in Fig. 4.1), the headwaters
(Haregeweyn 2008 and Chap. 13 of this volume; Vanmaerke of small ephemeral streams forming terminal fans in the
et al. 2010) and channel morphology (Billi 2007, 2008). The lowland between Dikhil and Ali Sabih within the Djibouti
scarcity of papers on Ethiopian rivers is due to many reasons, territory and some tributaries of larger ephemeral streams,
the most important of which is the lack of data. In fact, though running through Djibouti and outflowing into the Indian
the Ethiopian Ministry of Water, Irrigation and Energy has Ocean near the town of Zeylac in Somalia (the dark yellow
recently greatly improved the network of flow gauges across area close to the border with Djibouti in Fig. 4.1).
the country, the number of rivers with a long and continuous The Tekeze R. has a catchment area of 82,350 km2
record of data is still very limited. As a consequence of that (Awulachew 2007) and takes its source from Abuna Yosef
and, at the same time, aiming to attempt to supplement and to (4,190 m a.s.l.) in the portion of the northern highlands
go beyond the present data gaps, most of the recent papers on across Tigray and Amhara region that includes also the
Ethiopian rivers deal with hydrological modeling to predict Simen Mountains with the highest peak of Ethiopia, the Ras
flow regime (Collick et al. 2009; Wale et al. 2009; Legesse Dashen (4,620 m a.s.l.). The river turns around the Simen
et al. 2010; Melesse et al. 2010; Uhlenbrook et al. 2010; Di Mountains, points westward to join the Atbara R. in Sudan,
Baldassarre et al. 2011; Gebrehiwot et al. 2011), sedimenta- near the small town of Showak, and proceeds with the name
tion (Ahmed 2008; Billi and Bedri 2010; Easton et al. 2010) of Atbara as far as its confluence with the River Nile near the
and sediment transport (Haregeweyn et al. 2008 and Chap. 13 town of Atbara. Main tributaries within the Ethiopian terri-
of this volume; Steenhuis 2009; Vanmaercke et al. 2010; Billi tory are the Tirari and the Zarema rivers, whereas the
2011) or to assess the impact of climate change (Conway Angereb R. is a tributary of the Atbara (Fig. 4.2).
2000; Kim et al. 2008; Abdo 2009; Gebremicael et al. 2013), The Abay/Blue Nile is the largest river of Ethiopia. It is
land use change (Tamene 2006; Gumindoga 2010; Rientjes 1,610 km long and drains a watershed of about 176,000 km2
et al. 2011) and human impact (Batlthazar et al. 2013) on river upstream of El Diem (Kim 2008). It originates from Lake
hydrology and sediment yield. Though these studies may Tana that spills out by the town of Bahir Dar. The lake is fed
provide a useful framework to depict the main hydrological by small rivers, the largest of which are the Gilgel Abay
and sedimentation characteristics of the Abay/Blue Nile and (catchment area is 4,100 or 5,000 km2, according to Abdo
other smaller rivers, still large gaps of basic knowledge on et al. 2009 and Uhlenbrook et al. 2010, respectively, though
river processes in Ethiopia do exist. Some papers, in fact, are the first figure seems to be more appropriate) from the south,
based on old data (e.g., Gebrehiwot et al. 2011), have to cope the Rib (1,303 km2) and Gumara (1,283 km2) from the east
with insufficient data (e.g., Avery 2010) or test their models (Wale et al. 2009; Poppe et al. 2013) and the Megetch river
against very short hydrological time series. (514 km2—Wale et al. 2009) from the north. According to
The aim of this presentation is to provide an updated Kebede et al. (2006), these four major rivers contribute 93 %
picture of the hydromorphological characteristics of the of the inflow. From Lake Tana, the Abay/Blue Nile flows
Ethiopian rivers on the base of the largest as possible data southward into a deep canyon and then makes a large turn to
set, despite the data availability limitations reported above. the west pointing to the Sudan border before which it makes
another wide turn, first to the south and then to the north. The
river proceeds beyond the Sudanese border to join the White
4.2 General Setting and Hydrologic Nile at Khartoum. Within Ethiopia, the largest left tributaries
Data are the Didessa, with a watershed of about 25,800 km2 (Sima
2011) and Dabus (21,032 km2—World Bank 2008) rivers,
The hydrography of Ethiopia consists of 12 river systems whereas the largest right tributary is the Beles R.
and the ten larger ones are reported in Fig. 4.1. The western (14,200 km2—World Bank 2008). Other two important right
ones, namely Tekeze, Abay/Blue Nile, and Baro, flow into tributaries, the Dinder and the Rahad (14,891 and 8,269 km2,
the Mediterranean Sea, whereas those of the east and Oga- respectively—World Bank 2008) have their headwaters in
den, Wabe Shebele and Genale-Dawa, flow into the Indian Ethiopia, but join the Blue Nile in Sudan (Fig. 4.2).
4 Ethiopian Rivers 91
Fig. 4.1 The main hydrographic

basins of Ethiopia
Fig. 4.2 The study rivers. The

triangles indicate the flow gauges
92 P. Billi et al.
The catchment area of the Baro River is about highlands west of Addis Ababa, where the highest peak is
75,912 km2 (Awulachew 2007). Its headwaters are located 3,198 m a.s.l. In its upstream reach, the river runs down the
in the western part of the Ethiopian highlands, where their Rift Valley main escarpment and then, it reaches the Rift
elevation declines to lower values around 1,500 m a.s.l., floor where it is impounded to form the Koka reservoir.
between Guliso and Jimma. In Sudan, the Baro joins the Beyond the Koka lake, the river flows parallel to the Rift
Pibora River and becomes the Sobat River, which is a axis in a northeast direction. As the river enters the triangle
tributary of the White Nile. of the Afar depression, it slightly turns to the north, then
The Wabe Shebele River is the third longest river of eastward at the town of Tendaho and finally southward to
Ethiopia with a length of 1,150 km within the Ethiopian end up in the Abe Lake. The Awash River receives its
borders, but it has the largest catchment that, at Belet Weyne tributaries mainly from the left side, especially in the portion
in Somalia, is about 211,800 km2. Its source is in the eastern of catchment downstream of the Koka reservoir, where the
highlands, within a triangle made up by Mt. Bada, Mt. Kaka right side tributaries, coming from the eastern margin of the
and Mt. Batu (in the Bale Mountains), the highest peaks of rift escarpment, are ephemeral and commonly unable to join
which have an elevation of 4,139, 4,180, and 4,307 m a.s.l., the main river. Most tributaries are small. The largest of
respectively. The river initially flows from southwest to them are in the headwaters basin, whereas the larger one
northeast, then makes a large turn to southeast and proceeds beyond the Koka Lake is Mille River (4,467 km2).
into Somalia. As the river approaches the Indian Ocean The Omo River is an endorheic system with a catchment
coast, it turns to the south, runs for a few hundreds of area of about 79,000 km2 (Awulachew 2007). The catch-
kilometers parallel to the coastline and finally joins the Juba ment has a north–south elongated, almost rectangular shape,
River before the latter outflows into the Indian Ocean. The and its headwaters are located in the central highlands,
Wabe Shebele R. catchment is rather asymmetric with its between the Gurage Mountains and the town of Nekemte, at
main tributaries draining the left side of the catchment. They an elevation of about 2,500 m a.s.l. For the largest part of its
are the Ramis, Erer, and Fafen rivers, the sources of which is course, the river runs southward and after 760 km enters the
the eastern highlands, from Mt. Gara Muleta (3,381 m a.s.l.) Turkana Lake, which is its base level, at an elevation of
to Harar. about 360 m a.s.l.
The Genale-Dawa is the third largest river system of All the Rift Valley river systems (Fig. 4.1) drain closed
Ethiopia with a catchment of about 171,000 km2 (Awula- basins since they outflow into the large lakes on the rift floor.
chew 2007). It originates from the Bale Mountains and the The largest of these rivers are the Segen (13,612 km2), that
southeastern margin of the Rift Valley. Within the Ethiopian supplies Chew Bahir Lake in the southernmost portion of the
territory, the main tributary is the Weyb River, the source of Rift Valley, close to the Kenya border; the Bilate
which is from Mt. Batu. This tributary joins the Genale a few (5,791 km2), the source of which is in the Gurage Mountains
kilometers before this river crosses the border with Somalia, (Main Ethiopian Rift Valley); the Gidabo (3,473 km2) and
beyond which it takes the name of Juba River and, shortly the Gelana (3,411 km2) (Bilete 2009), both coming from the
downstream of the confluence of the Wabe Shebele, out- eastern margin of the southern Rift, with the latter three
flows into the Indian Ocean. rivers entering the Abaya Lake.
The Ogaden river systems (Fig. 4.1) consist of parallel,
relatively small ephemeral rivers that originate in the high-
lands south of Jijiga and from the margin of the Gulf of 4.2.1 Hydrological Data
Aden Rift. These rivers do not have any outlet into the sea
since they end up due to infiltration and evaporation around Most of flow data were obtained from the Ministry of Water,
the border with Somalia, at elevations between 400 and Irrigation and Energy of Ethiopia in the form of monthly
300 m a.s.l. and at a distance of 200–300 km from the Indian discharge data with additional information on maximum and
Ocean coastline. minimum discharge. Though Ethiopia is currently commit-
The Danakil river systems are all ephemeral and drain ted to improve its network of flow gauges, many important
endorheic basins. None of these rivers, in fact, has an outlet rivers are still ungauged, data gaps are still present and time
into the Indian Ocean. The base level of a very few of them series are discontinuous with the exception of a few rivers.
are small lakes such as Afrera and Asele, whereas the Due to high variability of catchment altitude, physiography,
majority of Danakil rivers forms terminal fans or distributary orientation and spatial rainfall distribution, any attempts to
systems on the borders or on the floor of structural basins fill the gaps by correlation method with reference rainfall
(see also Chap. 19). time series fail and the original data were used without any
The Awash river is the second longest river of Ethiopia manipulation that would only artificially and unrealistically
(about 1,200 km) and has a catchment area of about increase their validity and reliability. Additional data were
112,696 km2 (Awulachew 2007). Its source is in the taken from international organizations databases such as
UNESCO, FAO-SWALIM and RIVDIS (Vorosmarty et al. headwaters in Ethiopia but with the most downstream ref-
1998). erence flow gauge located a few kilometers beyond the
Notwithstanding these limitations, it was possible to Somali border.
compile a relatively large and complete database of river A new index, the runoff concentration index (RCI) is also
flows that covers almost the entire territory of Ethiopia and included. The RCI is introduced by analogy with the pre-
includes the most important river systems. The main cipitation concentration index proposed by Oliver (1980) in
hydrological data of the rivers considered in this study are order to define the monthly runoff variability within the year
reported in Table 4.1. In this table, the rivers/flow gauges as follows:
were divided into three categories: (1) highlands, i.e., rivers X
with the headwaters in the Ethiopian highlands and flowing RCI ¼ 100 ri2 =R2a ð4:1Þ
mainly in Ethiopia, with the most downstream flow gauge in
Ethiopia; (2) lake fed, i.e., those exiting from a lake/reservoir in which ri is the runoff volume in each individual month
or with a flow gauge located a few kilometers downstream of and Ra is the annual runoff volume. According to Oliver
it; (3) Ethiopian/Somali, i.e., those rivers with the (1980), a PCI, and by analogy a RCI, less than 10 suggests a
Table 4.1 Main hydrological characteristics of the Ethiopian rivers considered in this study
River/gauge A (km2) Ra Rau Ra RCI Qm Q1.58 Qbu Q10 Qpeak
(106 m3) (103 m3 km−2) CV (m3 s−1) (m3 s−1) (m3 s−1 km−2) (m3 s−1) (m3 s−1)
Highland
Abay/Blue Nile @ Kessie 65,784 17,136 260 0.39 22.9 543 4,548 0.069 9,373 13,681
Awash @ Hombole 7,656 1,382 180 0.29 26.9 44 357 0.047 494 497
Baro @ Gambela 23,461 11,541 492 0.13 15.0 366 1,133 0.048 1,415 1,594
Baro @ Itang 24,636 11,991 487 15.0 380
Baro @ Masha 1,653 1,872 1,132 0.19 14.9 59 208 0.126 302 357
Dabus @ Asosa 10,139 5,043 497 16.3 160
Didessa @ Dembi 1,806 189 105 0.19 21.8 6 31 0.017 87 190
Didessa @Arjo 9,981 3,960 397 18.5 128
Genale @ Chenemasa 9,273 2,911 314 0.23 11.7 92 367 0.040 630 925
Gibe @ Abelti 15,746 5,742 365 0.29 18.2 182 879 0.056 1,517 1,811
Gibe @ Asendabo 2,966 1,165 393 0.22 16.8 37 156 0.053 242 269
Gidabo @ Aposto 646 240 371 0.25 10.6 8 28 0.044 55 77
Gojeb @ Shebe 3,577 1,789 500 0.20 14.7 57 239 0.067 392 559
Ketar @ Abura 3,350 389 116 0.28 18.7 12 70 0.021 153 262
Leliso @ Adaba 135 45 331 0.26 13.7 1 9 0.067 17 27
Maribo @ Adaba 185 94 508 0.26 14.4 3 16 0.086 24 28
Meki @ Meki Town 2,433 283 116 0.40 16.8 9 58 0.024 122 218
Mille @ Mille 4,467 313 70 0.61 15.8 10 291 0.065 817 1,227
Mojo @ Mojo Village 1,264 211 167 1.03 18.5 7 89 0.071 231 346
Tekeze @ Embamadrie 45,694 5,223 114 0.35 27.0 166 1,492 0.033 2,883 3,063
Wabe Shebele @ Dodolla B.dge 1,035 218 210 0.26 12.6 7 20 0.019 121 77
Weib @ Alemkerem 3,578 386 108 0.38 14.3 12 106 0.030 190 183
Lake-fed
Abay/Blue Nile @ Bahir Dar 15,321 3,873 253 0.39 15.3 337 0.022 644 779
Awash @ Tendaho 63,485 2,178 34 0.31 10.9 430 0.007 1,065 1,687
Awash @ Melka-Sedi 21,510 1,472 68 0.27 12.1 188 0.009 373 443
Awash @ Metahara 16,417 815 50 0.58 10.9 96 0.006 245 374
Bulbula @ Kerkersitu 7,488 168 22 0.59 16.5 14 0.002 31 37
Ethiopian/Somali
Juba @ Lugh 179,520 6,063 34 0.29 11.4
Wabe Shebele @ Belet Weyne 211,800 2,132 10 0.32 12.1
A catchment area; Ra annual runoff; Rau unit annual runoff; CV variation coefficient; RCI runoff concentration index (see text for explanation); Qm mean discharge;
Q1.58 discharge with 1.58-years return time, equivalent to bankfull discharge; Qbu unit bankfull discharge; Q10 discharge with 10-years return time; Qpeak highest
discharge ever recorded
94 P. Billi et al.
uniform distribution; a value from 10 to 15 denotes a the Ketar River, joining it a few kilometers upstream of
moderately seasonal distribution; a value from 15 to 20 Ketar R. mouth into Ziway Lake (Fig. 4.4). This channel is
indicates a seasonal distribution. wider than the Ketar and, though in the vicinity of the
Q1.58 and Q10 were calculated with the Gumbel extreme Shetemata swamps it presently flows into the swamps, in the
values method only for river gauging stations with time past, it was an alternative channel of the Ketar that activated
series of at least 15 years, with the only exceptions of the during former high stands of Shetemata. Axial faulting and
Tekeze R. at Embamadrie (12 years) and the Baro R. at likely the transformation from a permanent to a seasonal
Masha (10 years). water body of the Shetemata turned the Ketar R. channel into
Sediment yield and sediment transport data are even more its present position (Corti and Manetti 2012).
discontinuous in space and time and were taken from the A more complex faulting system and tectonic deforma-
available scientific literature (see Sect. 4.5 for appropriate tion affected the Awash River course between Ombole and
citations), international organizations reports (Omuto et al. the Koka Lake in the Early Holocene. Initially, in fact, the
2009) and databases (e.g., FAO Land and Water Develop- Awash was the main river of a larger fluvial system,
ment Division—AQUASTAT). including the Meki and Mojo rivers, that outflowed into
Ziway Lake forming a large delta, presently fed only by the
Meki River. The former Awash flew southwestward, west
4.3 River Geomorphology of Ombole and Koye Lake, and received the Meki R. near
Ejersa Lele, where today this latter river makes a sharp,
4.3.1 Drainage Network angular turn to the right, before reaching Ziway Lake
(Fig. 4.5). The former connection between the Awash and
Ethiopia consists of a variety of landscapes resulting mainly the Meki is witnessed by field geomorphological evidence
from the combination of tectonics, volcanism, and fluvial of a dry, abandoned river valley and by a number of dis-
processes. The drainage network of Ethiopian rivers is a connected small streams, draining the Midrekebd ridge and
reflection of such interactions. The northeastern plateau is formerly joining the Awash R. (Sagri et al. 2008). The
deeply dissected by large perennial rivers such as the Abay/ Mojo R. joined the Awash a few kilometers upstream of
Blue Nile, the Tekeze and their main tributaries that devel- the delta in Ziway Lake, as it is presently shown by the
oped feathered and dendritic networks, whereas in the Cheleleka Lake, which maintains the former meandering
southwestern plateau, sub-parallel networks prevail (e.g., pattern of the Mojo R. (Figs. 4.5 and 4.6). Local faulting
Dawa, Genale, Weyb, Fafen and Webe Shebele rivers). and large-scale tectonic deformation then forced the Awash
Given the large occurrence of volcanoes all over Ethiopia, to flow northeastward, involving also the Mojo R. in this
the radial network is very common as well, the Zuqwala and drainage pattern rearrangement, reaching the present setting
Fantalé volcanoes being good examples (Fig. 4.3). (Sagri et al. 2008). Both the old and the present river
The rectangular drainage network is rather common in the systems show a typical rectangular drainage network lar-
Rift Valley and its margins. Here, channel pattern changes, gely controlled by axial and transverse fault systems as
due to even very recent faulting, can be easily observed. For shown by the perpendicular orientation of first and fifth
instance, in the eastern coastal plain of Ziway Lake, south of order rivers in the Meki catchment reported in Fig. 4.7
Ogolcho, a dry channel connects the Shetemata swamps with (Sagri et al. 2008).
Fig. 4.3 Examples of radial river

network around a volcano:
a Zuqwala (08° 32′ 30′′N–
38° 51′ 20′′E); b Fantale
(08° 59′ 07′′N–39° 54′ 26′′E)
Fig. 4.5 Early Holocene Awash River channel (dashed blue line)
Fig. 4.4 The old channel of the Ketar River in the plain of Ogolcho (08° 19′ 54′′N–38° 43′ 28′′E). At this stage of the river evolution, the
(08° 00′ 00′′N–38° 59′ 00′′E). The river abandoned the old channel and Awash received the Meki and Mojo rivers. Notice the presently
moved into its present position is connection with the movement of the disconnected, small tributaries of the Awash and the “meandering”
normal fault indicated by the dashed red line (modified from Corti and Cheleleka Lake, relict of the former Mojo river (modified from Sagri
Manetti 2012) et al. 2008)
Channel avulsion (Fig. 4.8) and anabranching (Fig. 4.9) and a real alluvial plain is missing or it is patchy and very
are also common in low-gradient alluvial areas or close to limited (4.11). Also on the bottom of the Rift Valley or other
and beyond the borders of the plateau swelling. structural basins, rivers commonly flow entrenched into
poorly consolidated volcanic ashes (Fig. 4.12) and fluvial–
lacustrine deposits (Fig. 4.13). This is due to tectonics that in
4.3.2 Channel Morphology many parts of Ethiopia is witnessed by vertical movements,
active also in very recent times. Rivers with a large alluvial
Three different types of headwaters are common in Ethiopia: plain are found only west of Gambela, in lowland areas
(1) volcano headwaters, stretching from an individual vol- underlain by Quaternary deposits (e.g., the Baro River), in the
cano or a volcanic ridge. They are found on the flat surface of downstream Omo River valley and similar flat lands sur-
the highland plateau and in the Rift Valley and structural rounding small lakes and swamps (e.g., the desiccated Chew
basins floor; (2) plateau headwaters. In this case, the head- Bahir Lake/swamps and the Segen River; the Awash River
waters divide is poorly defined since it is commonly marked between Beda and Yardi lakes). Other lowlands or closed,
by small hills or a few meters high ridge resulting from a structural basins are crossed by ephemeral streams that form
basalt flow; (3) escarpment headwaters, developed from the distributary systems or terminal fans.
margins of structural depressions such as the rift flanks or the Type 1 and especially type 3 headwaters bestow a moun-
shoulders of smaller structural basins. Given the physiogra- tainous geomorphology to the landscape with very steep
phy of headwaters and the recent structural setting, most of slopes and streambeds. In small mountain streams, all the
Ethiopian rivers are therefore incised into bedrock (Fig. 4.10) main geomorphological features described by Montgomery
96 P. Billi et al.
Fig. 4.6 The Cheleleka Lake

(08° 12′ 57′′N–38° 46′ 56′′E)
Fig. 4.7 Polar diagrams with

orientation of first (a) and fifth
(b) order streams in the Meki R.
catchment. The larger rivers
follow the main rift-oriented
faults, whereas the smaller, more
recent streams have a mean
transverse orientation (modified
from Sagri et al. 2008)
and Buffington (1997) characterizing these steep channels are diameter) can be easily entrained and transported (Fig. 4.16).
present (Fig. 4.14). They include step-pool sequences (Grant Evidence of transport of such large boulders as bedload is
et al. 1990), boulder berms (Carling 1987), glides (Bisson provided by their imbricated position and the formation of
et al. 1982), riffle-pool (Leopold et al. 1964) and transverse huge particle clusters (Brayshaw 1984) (Fig. 4.17). Figure 4.17
ribs (Koster 1978). Commonly, the streambed of a small also shows that large and small particles are transported
mountain stream takes up the whole valley bottom and the en mass and cobbles of 15–20 cm move in saltation as
alluvial plain is missing or consists of terraced narrow strips proved by the particle embedded between the two larger
and small patches (Fig. 4.15). The valley side slopes bound the boulders.
streambed on which braided or pseudomeandering (Hickin On the plateaus or on the treads of faulted steps, which
1972) channels develop during low flow conditions. are almost flat and characterized by low gradients, high
During floods, steep mountain streams experience high sinuosity rivers are more common. Most of them have a
shear stresses and large boulders (0.5–2.0 m in mean typical meandering pattern but are commonly entrenched
Fig. 4.8 Avulsion of the

Baro R. near Berhane Selam
(08° 10′ 16′′N–34° 07′ 00′′E)
Fig. 4.9 Anabranching of the

Baro R. close to the Sudan border
(8° 21′ 34′′N–33° 47′ 03′′E)
98 P. Billi et al.
Fig. 4.10 The bedrock channel

of the Didessa R. near Barri
(09° 01′ 50′′N–36° 09′ 18′′E)
Fig. 4.11 The Gibe River at

Abelti. In this mountain reach
a small alluvial plain is present
on the left bank. Flow is
toward the reader.
(08° 13′ 48′′N–37° 34′ 40′′E)
within shallow bedrock banks. Gravel beds (Fig. 4.18) pre- result (Fig. 4.20); (2) if it is incised, a steep mountain stream
vail over sand beds (Fig. 4.19). develops (Fig. 4.21).
As these low-gradient rivers exit the plateau/faulted The majority of the Ethiopian land consists of highlands.
blocks, their downstream pattern is determined by the mar- The larger rivers and their tributaries have dissected the
gin morphology: (1) If it is abrupt and sharp, a waterfall may plateau and formed deep and more or less narrow gorges and
Fig. 4.12 The flow gauge on the

Ketar R. near Ogolcho
(08° 01′ 58′′N–39° 01′ 08′′E). The
streambed is incised into poorly
consolidated volcanic ash
Fig. 4.13 Ephemeral stream

incised into Quaternary colluvial–
alluvial deposits at the margin of
the Danakil depression on the
way to Dallol (13° 54′ 42′′N–
40° 09′ 29′′E). Notice the three
orders of fluvial terraces on the
left. Flow is away from the reader
canyons. This is by far the most common river morphology down to the Adigrat Sandstones. The development of the
in Ethiopia, the best example of which is given by the Abay/ Abay/Blue Nile canyon is associated with the fast uplift of
Blue Nile canyon near Degen (Fig. 4.22). This spectacular the area that occurred during the Pliocene and the Pleisto-
gorge is about 1,500 m deep, with the plateau at an average cene (McDougall et al. 1975; Yemane et al. 1985), whereas
elevation of 2,500 m a.s.l. and the river bed at 1,000 m a.s.l. Ayalew and Yamagishi (2004) assume the valley incision
(Ayalew and Yamagishi 2004). The river has cut through the was initiated in the upper 150 km of the Abay/Blue Nile
trap basalt series and the Mesozoic sedimentary sequence River course before or a little after the beginning of the
100 P. Billi et al.
Fig. 4.14 A typical mountain

stream in the central highland.
Notice the small lateral bars, a
sequence of three small step-
pools and a glide upstream of
them. The stream is mainly
incised into the bedrock, the
channel is bounded by the slope
foots and only small patches of
flood plain are present
Quaternary. According to these authors, the phase of maxi- bedrock, or may have a small, stripped alluvium. The
mum river degradation occurred at the end of the Pleistocene channel is almost straight or with a low sinuosity (mainly
or at the beginning of the Holocene and valley widening imposed to the river by structural factors) and punctuated by
occurred as a result of landslides and retreating rock cliffs as alternating lateral bars (Leopold and Wolman 1957)
is now evident by the remaining tabular hills and mesas. (Fig. 4.22). A similar river and valley morphology is also
Other deep canyons are formed by the Wabe Shebele in the attained by large rivers where they cross a structural
upper portion of the Somali plateau (see Chap. 1). threshold between adjacent faulted steps, where rapids can
Notwithstanding the much larger size, these rivers share develop (Fig. 4.23), or flow on lowlands such as the Rift
their main morphological characteristics with the smaller, Valley floor. Here, recent tectonics and, primarily, the
steep mountain streams (Figs. 4.15 and 4.21). In fact, they emplacement of thick basaltic lava flows have forced the
take up the whole valley bottom, may be incised into the river to incise deep canyons (Fig. 4.24) and to form
Fig. 4.15 A mountain stream

draining a main rift escarpment
near Dese (11° 06′ 52′′N–
39° 39′ 30′′E). The streambed
takes up the whole valley bottom.
Notice the steep slopes bounding
the streambed, the occurrence of
only a small portion of flood plain
and the braided/
pseudomenadering pattern of the
low flow channel
Fig. 4.16 A steep mountain

stream draining the deep
escarpment of the Abay/Blue Nile
gorge (10° 06′ 24′′N–38° 10′ 30′′
E). The vicinity with the
headwaters favors the supply of
very coarse boulders that are
easily entrained given the steep
bed gradient and high flow energy
spectacular waterfalls (Fig. 4.25), as is the case of the Awash due to recent change in the lake base level as for instance in
River in the Awash National Park. the case of the Meki R. flowing into the Ziway Lake
Most rivers, as they enter lowland areas, develop a brai- (Fig. 4.5) or the Bilate River (Fig. 4.28) which runs south-
ded stream morphology in their proximal reaches (Fig. 4.26) ward for more than 100 km parallel to the Rift axis and form
and then, if the gradient decreases, a typical meandering a large delta into the Abaya Lake.
pattern (Fig. 4.27). By contrast, some of the rivers flowing Dryland ephemeral streams have specific morphological
on the rift bottom and entering the rift lakes may have a large characteristics that are described in more detail in Chap. 12
portion of their channel entrenched into the Rift Valley floor of this volume.
102 P. Billi et al.
Fig. 4.17 A huge pebble cluster.

Notice the small particle
interlocked in between the very
large, imbricated boulder in the
cluster stoss side and the core
boulder
Fig. 4.18 A coarse grained

meandering channel on top of the
central highlands plateau near
Debre Birhane (09° 49′ 21′′N–
39° 43′ 04′′E)
River at Embamadrie. Mean, peak, and bankfull discharge

4.4 Discharge are well correlated (R2 = 0.83, 0.90 and 0.89, respectively)
with catchment area (Fig. 4.29) whereas, as expected for the
The rivers considered in this study have very different high climatic conditions variability, no correlation was
catchment size and hydrologic characteristics. The Wabe found between Q1.58 and Qbu, nor between catchment area
Shebele river system includes both the smallest, 135 km2, (A) and Qbu, although highland- and lake-fed rivers can be
and the largest catchment, 211,800 km2, if we consider the clearly distinguished (Fig. 4.30).
flow gauge at Belet Weyne, located a few kilometers beyond The data set of this study was also used to test the
the border with Somalia (Table 4.1). As a consequence, application range of the empirical equation of Syvitski and
mean discharge (Qm) of Ethiopian rivers varies largely Milliman (2007), developed from a globally distributed
between 1 (Leliso R.) and 543 m3 s−1 (Abay/Blue Nile at database of 488 rivers, which predicts mean discharge as a
Kessie) and the same large variability is observed for function of catchment area:
bankfull discharge (Q1.58), 9–4,548 m3 s−1 (Leliso and
Abay/Blue Nile at Kessie, respectively). Bankfull discharge Qm ¼ 0:075 A0:8 ð4:2Þ
per unit catchment area (Qbu) ranges between 0.002 (Bulbula
River) and 0.126 m3 s−1 of the Baro River at Masha, where Equation 4.2 proved to predict with relative accuracy
Qbu is almost double that of the Abay/Blue Nile at Kessie (91 %) the actual mean discharge of Ethiopian rivers with
(Table 4.1) and almost four time larger than the Tekeze the exception of the Tekeze R. at Embamadrie and the
Fig. 4.19 The Weja R., a sand

bed, meandering river flowing on
top of a faulted block in the Meki
R. catchment near Koshe
(08° 01′ 15′′N–38° 31′ 07′′E).
Flow is away from the reader
Ethiopian/Somali rivers (Fig. 4.31). The strength of this

relationship is confirmed by the Spearman’s rank correlation
coefficient at a 0.01 significance level. This result indicates
that Eq. 4.2 can be used in Ethiopia to predict the mean
discharge of ungauged rivers, provided they are not located
in dryland areas. Nevertheless, a closer inspection of the
diagram of Fig. 4.31 and a more detailed analysis of data
show that Eq. 4.2 predictions are less accurate when mean
discharge is expected to be less than 20 m3 s−1. In the case of
small rivers, in fact, a different equations was derived as
follows:
Qm ¼ 0:1755 A0:51 ð4:3Þ
which is capable to explain 85 % of mean discharge

variability.
4.4.1 Runoff Volume
Annual runoff (Ra) ranges widely among the study rivers,

given their large variability in catchment area and climatic
conditions, with the lowest value (45 × 106 m3 year−1)
recorded in the smallest stream (the Leliso R. headwater
tributary of the Wabe Shebele) and the largest
(17,136 × 106 m3 year−1) measured on the Abay/Blue Nile at
Kessie (Table 4.1). Annual runoff is well correlated
Fig. 4.20 The waterfall of the Aleltu R. entering the deep canyon of
the Sodoblè R. (Muger R. catchment) on the central highland plateau (R2 = 0.87) with catchment area only for the highland-fed
near Chancho (09° 25′ 47′′N–38° 39′ 27′′E) rivers, with the datum of the Tekeze R. as an outlier
104 P. Billi et al.
Fig. 4.21 A mountain stream

draining the main Rift escarpment
of the Danakil lowland
(13° 51′ 53′′N–39° 50′ 36′′E)
Fig. 4.22 The Abay/Blue Nile

Gorge near Degen (10° 4′ 28′′N–
38° 11′ 22′′E). The river is deeply
incised into the tabular bedrock
(a), has elongated lateral bars but
no alluvial plain (b). The
streambed is bounded by foot
slopes (b)
Fig. 4.23 The rapids of the

Awash R. at Melka Konturè
(08° 42′ 14′′N–38° 36′ 27′′E)
(Fig. 4.32). Including also the Tekeze, the correlation coef- within Africa (Table 4.2), a diagram of unit annual runoff
ficient decreases to 0.75 whereas, if also the lake fed and the (Rau) versus catchment area (A) for Saharan, Sub-Saharan,
Somali rivers are included, the best fit is given by a power and Ethiopian rivers is plotted (Fig. 4.34). This diagram
function (Ra = 2.324 · A0.7), the determination coefficient of shows clearly that the Ethiopian rivers, though being among
which is only R2 = 0.63. Even poorer correlations are those with the smaller catchment area, have a unit runoff
obtained between unit runoff (Rau), i.e., the annual runoff per comparable to that of other African (especially equatorial)
unit catchment area, and the watershed area though, simi- rivers, whose catchments are as much as three orders of
larly to unit bankfull discharge, the three groups of highland, magnitude larger.
lake-fed, and Ethiopian/Somali rivers can be clearly distin- The variation coefficient (CV) values indicate that the
guished (Fig. 4.33). least inter-annual variability occurs in the Baro R. at Mash
In order to point out the potential water resources asso- and the Didessa R. at Dembi (CV = 0.19), whereas the
ciated with the runoff of Ethiopian rivers and to rank them largest variability is observed in the rift valley rivers with the
106 P. Billi et al.
Fig. 4.24 The canyon of the

Awash R. in the Awash National
Park (08° 52′ 21′′N–40° 05′ 28′′E)
Fig. 4.25 The Awash R. water

fall in the Awash National Park
(08° 50′ 32′′N–40° 00′ 42′′E)
highest values recorded in Mojo R. at Mojo Village CVs in the 0.25–0.50 range (Fig. 4.35). Therefore, though
(CV = 1.03) (Table 4.1). The annual runoff of more than the unit annual runoff (Rau) is relatively high with respect to
50 % of the studied rivers is highly variable, with a CV the catchment size (Fig. 4.34), the high values of the annual
ranging between 0.50 and 0.75, and about one quarter of the runoff CV may result in some constraints to the exploitation
sample is characterized by a medium to low variability with of river water for agriculture irrigation.
Fig. 4.26 A braided river a few

kilometers upstream of its mouth
into Abaya Lake, near Arba Minch
(06° 02′ 40′′N–37° 32′ 58′′E)
Fig. 4.27 A meandering reach of

the Awash R. upstream of the
Gora marshes (08° 23′ 20′′N–
38° 54′ 35′′E)
In order to investigate the runoff variability within a year, by Eq. 4.1 for the study rivers, although the lake-fed
by analogy with the precipitation concentration index pro- rivers were not considered since their runoff volume is
posed by Oliver (1980) to define the monthly rainfall dis- directly controlled by the lakes supply and/or reservoir
tribution in a year, the runoff concentration index (RCI) is water routing programmes, shows that more than 50 % of
used (see Eq. 4.1). them are characterized by a seasonal (37 %) and a
According to Oliver (1980), values of PCI (and by markedly seasonal (15 %) distribution of runoff, whereas
analogy of RCI) less than 10 indicate a uniform distri- only for 37 % of them the seasonality is moderate
bution, values from 10 to 15 denote a moderately seasonal (Fig. 4.36). This result is not surprising given the domi-
distribution, and values from 15 to 20 are characteristic of nant role of the monsoon type rainfalls that are concen-
a seasonal distribution. Values above 20 are associated trated in July and August all across the largest part of
with strong seasonal variations. The RCI data set obtained Ethiopia.
108 P. Billi et al.
Fig. 4.28 The Bilate River

flowing southward entrenched
into the rift valley floor. Flow is
toward the reader (07° 17′ 09′′N–
38° 04′ 22′′E)
100000 Q1.58 = 0.0589·A0.9672 1

Qp = 0.236·A0.912
R² = 0.89
R² = 0.90
10000
Q bu (m3s-1km-2)
0.1
Q (m3s-1)
1000
Qp
Highlands
100 Q1.58
0.01 Lake fed
Qm
10 Qm = 0.0076·A1.0159
R² = 0.83
1 0.001
100 1000 10000 100000 100 1000 10000 100000
A (km2 ) A (km2)
Fig. 4.29 Catchment area (A) versus characteristic discharges for the Fig. 4.30 Catchment area (A) versus unit bankfull discharge (Qbu).
study highland-fed rivers. Qp = maximum discharge ever recorded; Highland- and lake-fed rivers of this study are quite distinct. The latter
Q1.58 = discharge with 1.58-year return time equivalent to bankfull have an order of magnitude smaller unit bankfull discharge
discharge; Qm = mean discharge
2.4 million of people affected and about 17 million US$ of

4.4.2 Floods damages. The worst floods occurred in 2006, with 862
fatalities and 361,600 people affected (Alemu 2009). Sena
Though some studies have been carried out on floods in
and Woldemichael (2006) report some additional data on
Ethiopia, most of them were focused on flood forecast and
floods in different parts of the country:
flood frequency analysis, regionalization and modeling (e.g.,
Demissie 2008; Baratti et al. 2012). Very few scientific 1996 displacement of 40,000 people in Wonji and Mete-
papers and field studies dealing with recent decades floods hara due to burst of Awash River banks and
and flash floods (e.g., Alemu 2009; Billi et al. 2015) are overflow from the Koka reservoir;
available in the international literature, though these devas- 2003 100,000 people affected by Wabe Shebele river
tating events are cited in the international news, humani- flooding in Somali region;
tarian organizations, and local authorities reports. 2005 170 people died and 260,000 were displaced due to
According to EM-DAT (2013), from 1900 to 2013, 50 flooding from Wabe Shabele, Hargeysa, Bilate,
floods occurred in Ethiopia with 1,976 people killed, about Ashewa, Genale, Dawa, Fafen, Sile, and Sego rivers;
600 Table 4.2 Mean annual unit runoff volume (Rau, 103 m3 km−2) of
Ethiopian and other African rivers
500
Region/country No of data Rau Range
400
Qmp (m3s-1)
Algeria 6 147 2–486

300 Egypt 1 30
200 Morocco 4 217 27–449
Tunisia 2 23 18–28
100
North Africa 13 104 2–486
0
0 100 200 300 400 500 600
Benin 4 98 36–152
Qm (m3s-1) Burkina Faso 7 68 27–174
C.A.R. 11 337 24–683
Fig. 4.31 Very good correlation (91 %) between observed (Qm) and
predicted (Qmp) mean discharge by Eq. 4.2 (Syvitski and Milliman Cameroon 9 602 168–1,216
2007) Chad 4 226 21–544
Congo 4 426 332–511
25000 Gabon 5 765 507–1,081
Ra = 0.3004·A + 413.72 Ghana 9 159 62–288
20000 R² = 0.87
Guinea 4 1,973 314–6,509
Ra (106m3)
15000 Ivory Coast 6 375 87–653

All data Kenya 2 4,524 2,979–6,076
10000
Tekeze R. Lesotho 2 144 102–185
5000 Liberia 2 1,457 694–2,220
Malawi 2 139 102–177
0
0.E+00 2.E+04 4.E+04 6.E+04 Mali 10 213 61–377
A (km2 ) Mozambique 8 162 87–391
Fig. 4.32 Catchment area (A) explains 87 % of highland-fed rivers Niger 3 148 33–376
annual runoff (Ra) variability with the exception of the Tekeze R Rwanda 2 219 203–234
Senegal 6 140 21–311
10000 South Africa 9 37 5–122
Sudan 4 96 27–166
Rau (103m3km-2 )
1000 Tanzania 5 157 15–419

Togo 3 277 122–420
100 Highlands Uganda 5 182 31–478
Lake fed
Zaire 1 366
10 Somalia
Zambia 3 125 19–186
Zimbabwe 13 79 5–146
1
1.E+02 1.E+03 1.E+04 1.E+05 1.E+06
Sub-Saharan 143 354 5–6,509
A (km2) Ethiopia 22 290 10–1,132
Fig. 4.33 The combination of parameters such as catchment area

(A) and unit annual runoff (Rau) well distinguishes highland, lake-fed,
and Ethiopian/Somali rivers. These latter two categories of rivers have
one order of magnitude smaller unit annual runoff in spite of a
catchment area that is one to two orders of magnitude larger
Flood hydrological data measured in the field or inferred
“post event” are substantially missing (Alemu 2009; Billi
et al. 2015). The highest discharge (Qp) recorded for the
2006 the worst scenario of flood in Ethiopia, resulting in highlands fed rivers are very close to Q10, i.e., the discharge
loss of lives, damage of property and destruction of with 10-years return time (Table 4.1) and Qp is well corre-
livelihoods of tens of thousands of people in South lated with both Q10 (R2 = 0.99) and bankfull discharge
Omo, Dire Dawa and 19 in other parts of the country (Q1.58) (R2 = 0.97).
110 P. Billi et al.
10000 4.4.3 Flow Discharge Trend
Though the effects of climate change on several aspects of

R au (10 3m 3km - 2 )
1000
the physical environment have been investigated in many
100 Saharan regards, comparatively little attention has been paid to the
Sub-Saharan hydrological response of rivers throughout the last century or
10 Ethiopian the last decades (e.g., EEA 2012). The effects of climate
change on river discharge at global, continental and local
scale have been investigated by a few authors, but most of
1
1.E+02 1.E+03 1.E+04 1.E+05 1.E+06 1.E+07 the studies and reports compiled by international organiza-
A (km 2) tions are focused on future scenarios (IPCC 2007; Bates
et al. 2008) rather than on the observation and interpretation
Fig. 4.34 Catchment area (A) and unit annual runoff (Rau) for African of the changes which occurred in the past decades (Bates
rivers of different regions. Sub-Saharan rivers include a number of
rivers south of the Sahara desert (i.e., sub-Saharan s.s., equatorial, and et al. 2008). This situation is well reflected in the whole
southern Africa). Though the study Ethiopian rivers have a smaller African continent, including Ethiopia (Conway 2005; Bates
catchment area, their unit annual runoff is comparable to that of much et al. 2008).
larger rivers The Hadley Center (2005) reports a general decreasing
trend for runoff of the African rivers, whereas Conway et al.
(2009) concluded that robust identification of hydrological
change was severely constrained by data limitations for the
sub-Saharan Africa and Conway (2005) found no clear
Annual runoff CV
indication of climate change effects on the Abay/Blue Nile
4 % 0 %4 % River flow, because of uncertainty in projected rainfall pat-
terns and the influence of complex land and water man-
23 % 0.00-0.25
agement structures.
0.25-0.50 Notwithstanding such a complex situation and the lack of
0.50-0.75 continuous long time series, it is worth attempting to trace
54 % 0.75-1.00 the runoff trends across the last four decades using the
>1.00 database constructed for this study. As reported early in this
paper, both daily and monthly flow data of the study rivers
are not homogeneous in time extension and gaps occur. Any
attempt to fill the gaps by correlation analysis with rainfall
Fig. 4.35 Frequency distribution of the annual runoff variability amounts failed because of the non uniform distribution of
coefficient of the Ethiopian study rivers the rain gauges within the catchment, the occurrence of gaps
and/or long periods without rainfall data, the high diversity
in basin physiography and the spatial variability of rainfall
(Billi 2011; Billi et al. 2015). Any other manipulation of data
is considered to produce significant alteration of the actual
Runoff Concentration Index time series with errors the extent of which is comparable to
15 % those expected by using the original data set unaltered.
37 % Though the annual runoff time series of the rivers under
0-10
study range from 1957 to 2009, in order to work with a more
10-15
uniform and consistent database, only the 1968–2004
37 % 15-20 interval was considered. For the annual runoff volume of
>20 each year, the z score is calculated as follows:
z ¼ ðxi - lÞ=r ð4:4Þ
in which for a given river xi is the annual runoff (Ra) of the

Fig. 4.36 Frequency distribution of runoff concentration index (RCI) ith year, μ is the mean, and σ the standard deviation of Ra
of the study rivers time series.
1968 -2004 4.5 Sediment Transport

1.5
1
Data on sediment transport and sediment yield of Ethiopian
rivers are very scarce. The available data on specific sedi-
0.5
ment yield (SSY) are reported in a very few papers and in a
0
few cases only occasional data are reported. SSY data from
Z
-0.5 reservoir filling are reported by Tamene et al. (2006) and

-1 Haregeweyn et al. (2008). FAO (2000), Billi (2004, 2011),
y = -0.0027x - 0.1938
-1.5 Vanmaercke et al. (2010) and Haregeweyn et al. (Chap. 13
-2 this volume) studies are instead based on data obtained from
field measurements. Discontinuous field measurements on
1968
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
the Awash R. at Ombole from 1986 to 1990 are reported in
Fig. 4.37 Time series of the annual runoff Z score averaged over all detail in the FRIEND/Nile report (UNESCO 2007) and let to
the study rivers z score data. No significant change is observed for the construct the suspended sediment concentration rating curve
1968–2004 interval
for this station (Fig. 4.38). Guzman et al. (2013) present
detailed suspended sediment field measurements for three
For each year, a mean value of Z is calculated by aver- very small catchments (area ranging from 113 to 447 ha) in
aging the z values of all rivers considered and finally, the the sub-humid highlands, but their rating curves are poorly
time series of Z scores is obtained (Fig. 4.37). Figure 4.37 significant as it would be expected in such small catchments
indicates that, though a minuscule decrease results from the in which local factors, such as seasonal changes in vegeta-
interpolating line, in general, no clear trend is present; hence, tion and cultivation practices, may play a relevant role and
no evident change in annual runoff volume has occurred in introduce a high degree of randomness in the data.
the Ethiopian rivers during the 36-years interval (1968– The Ethiopian rivers considered in these studies have a
2004) considered in this study. Within this general picture, catchment area ranging from 0.72 km2 of the Adihilo R. in
obviously, there are exceptions such as marked increasing Tigray to 62,732 km2 of the Awash R. at its most down-
trends for the Abay/Blue Nile at Kessie, the Genale at stream end close to Djibouti. The SSY varies widely as well,
Chenemasa, the Ghibe at Abelti and one Ethiopian/Somali with the lowest value of 4 tkm2 year−1 of the Woyb (Genale-
river such as the Wabe Shebele at Belet Weyne, but these are Dawa river system) to 4,935 tkm2 year−1 measured on the
counterbalanced by less prominent decreasing trends for Adikenafiz reservoir, again in Tigray. The mean values of
another Ethiopian/Somali river, the Juba at Lugh, and SSY measured for 24 reservoirs and 47 river sites of Ethi-
modest negative trends for all other rivers. The negligible opia considered in this study, 78 river sites in northern
decrease in river runoff is paired by a very modest (of the Africa, 112 in the sub-Saharan Africa and the 668 world
same order of magnitude) decreasing trend of annual rainfall river data compiled by Summerfield and Hulton (1994)
across the country (see Chap. 4, this volume).
Unfortunately, runoff data of the first half of the twentieth
century (1905–1959) are available only for the Baro R. at Awash R. at Hombole
Gambela (Sutcliffe and Parks 1999). A comparison with the 400000
1977–2003 interval shows that the annual runoff has 350000 SSY = 11.165Q1.9491
decreased by 12 %. The reasons behind that are manifold but 300000 R² = 0.85
SSY (td -1)
human impact may have played an important role as 250000

observed in other African rivers (Conway et al. 2009; Mahe 200000
et al. 2013). By analogy with this finding, it could be 150000
speculated that the runoff time series of Ethiopian rivers not 100000
affected by reservoirs and spanning the past century would 50000

0
probably depict a different (decreasing?) trends conversely to 1 10 100 1,000
the general stability depicted by Conway et al. (2009) for Q (m3s-1)
several large rivers in the sub-Saharan Africa in the last
decades. Unfortunately, the lack of old river flow data is Fig. 4.38 Suspended sediment yield (SSY) rating curve for the Awash
R. at Ombole
presently a tight constraint to any adequate and reliable
analysis.
112 P. Billi et al.
without considering the African rivers, are reported in Ethiopian rivers

Table 4.3. 40 (field measurements)
The SSY data of the Ethiopian rivers, when plotted 35
against the catchment area, do not follow the typical 30
Y (10 6 t yr-1)
decaying pattern, as reported in the classic study of the 25
20
global sediment yield by Milliman and Meade (1983), and Y= 0.0005A - 0.1989
15
no significant correlation was found. By contrast, the total R² = 0.84
10
sediment yield (Y) obtained from field measurements on the 5
Ethiopian rivers varies significantly (R2 = 0.84) with catch- 0
ment area (Fig. 4.39) and is predicted by the following 0 10000 20000 30000 40000 50000 60000 70000
equation: A (km 2)
Fig. 4.39 Catchment area (A) versus sediment yield (Y)

Y ¼ 0:0005A þ 0:1989 ð4:5Þ
in which Y is expressed in 106t year−1 and A is catchment

Also Boateng et al. (2012) found a good correlation
area in km2.
(R2 = 0.98) between these two parameters for a few rivers in
Ghana, but their variability is best described by an expo-
Table 4.3 Suspended sediment yield (SSY, tkm−2 year−1) of Ethiopian nential function. Omuto et al. (2009) report a suspended
and other African and world rivers sediment rating curve with a high correlation coefficient
Region/country No of data Mean SSY Range (R2 = 0.88) for both the Juba and Wabe Shebele rivers:
Algeria 39 945.2 36–6,654
Qs ¼ 0:087Q1:96 ð4:6Þ
Morocco 36 1,173.4 100–4,620
Tunisia 1 708.0 where Qs is suspended sediment discharge in kgs−1 and Q is
Egypt 2 39.5 39–40 discharge in m3 s−1. This equation was obtained by field
North Africa 78 1,024.3 36–6,654 measurements during two short sampling campaigns in 2007
Cameroon 4 85.7 20–210 and 2008 at a few stations, including also Lugh and Belet
C.A.R. 4 5.5 3–9 Weyne that are located just a few kilometers beyond the
Chad 4 7.0 1–15
Somali border. For the Gereb Oda, a small ephemeral stream
close to Alamata in Tigray, the field measurements of Billi
Ghana 16 14.4 1–85
(2011) indicate that suspended sediment transport is best
Ivory Coast 2 15.7 9–22
described (R2 = 0.97) by an exponential law of the following
Kenya 14 2,136.0 4–19,520 type:
Lesotho 19 615.9 3–2,050
Madagascar 2 2,358.0 1,586–3,130 Qs ¼ 108 e0:12Q ð4:7Þ
Mali 2 27.3 15–40
Mozambique 3 44.1 17–80
Also the Gereb Oda data were measured during one rainy
season and for these reasons, though the correlation coeffi-
Niger 1 4,000
cient is high, Eqs. 4.6 and 4.7 must be considered as
Nigeria 14 196.1 5–483
indicative and used with caution.
Senegal 1 8.0 In order to compare the sediment yield of Ethiopian rivers
South Africa 2 495.0 100–890 with that of other African rivers, a diagram of SSY versus
Sudan 2 2,189.5 927–3,422 catchment area (A) is plotted (Fig. 4.40). From this figure, it
Tanzania 4 195.7 95–390 is evident that though the Ethiopian rivers have a compar-
Zaire 2 14.5 11–18 atively smaller catchment area, their SSY is of the same
Zimbabwe 16 150.6 1–39
order of magnitude as other, much larger rivers of the sub-
Saharan Africa. This is also the evidence of high soil erosion
Sub-Saharan 112 558.1 1–19,520
a
that is taking place in the Ethiopian highlands as reported in
World river data 668 176.0 136,000
a number of studies (e.g., Hurni 1983; Nyssen et al. 2004,
Ethiopian reservoirs 24 1,428 345–4,935 2007; Tamene and Vlek 2008).
Ethiopian rivers 47 670 4–3,784 Bedload field measurements data for Ethiopian rivers are
Ethiopia 71 935 4–3,784 very scarce and limited to the Kulfo R. in the southern Rift
a
Summerfield and Hulton (1994) data excluding the African rivers Valley (Girma and Horlachen 2004) and to the Gereb Oda in
100000 2. The physiography of headwaters, commonly located on

10000
the highland plateau or within the rift margins, and the
recent structural setting provided the most of Ethiopian
SSY (tkm -2yr -1)
1000 rivers with a pattern incised into bedrock. Wide alluvial

100 North Africa
plains are commonly missing or patchy and very limited
alluvium may be present. As a consequence, larger rivers
Sub-Saharan
10 and their tributaries have dissected the plateau and
Ethiopia
1
formed deep and more or less narrow gorges and canyons
in which the rivers show typical channel features of
0 mountain streams. Rivers with a large alluvial plain are
1 100 10000 1000000
Area (km2)
found only west of Gambela, in lowland areas at the far
margins of the structural swell, underlain by Quaternary
Fig. 4.40 Catchment area (A) versus suspended sediment yield (SSY) deposits (e.g., the Baro River), in the downstream Omo
for a number of African rivers. Sub-Saharan rivers include a number of River valley and similar flat lands surrounding small
rivers south of the Sahara desert (i.e., sub-Saharan s.s., equatorial, and
lakes and swamps.
southern Africa). Though the Ethiopian rivers have a comparatively
smaller catchment area, their SSY is of the same order of magnitude of 3. Flow data of the studied rivers show that discharge is
much larger African rivers well correlated with catchment area. In fact, catchment
area explains 83 % of mean, 89 % of bankfull, and 90 %
the Kobo Basin (Tigray) (Billi 2011). In both studies, bed- of peak discharge, respectively, whereas, given the high
load transport rate was found to be mainly controlled by unit variability of climatic and physiographic conditions, no
stream power (ω). In fact, the bedload rate (Qb) of Gereb correlation was found between bankfull and unit bankfull
Oda (Billi 2011) is well correlated (R2 = 0.71) with ω discharge, nor between catchment area and unit bankfull
through the following linear equation: discharge.
4. The equation of Syvitski and Milliman (2007) proved to
Qb ¼ 0:0567x0:097 ð4:8Þ predict with a high accuracy (91 %) the actual mean
discharge of Ethiopian rivers, with the exception of the
whereas the study of Girma and Horlachen (2004) proved Tekeze River and the Somali rivers. Both this equation
that the Bagnold’s criterion (Bagnold 1980), which is based and those developed in this study are reliable tools to
on stream power, predicts with the largest accuracy the predict characteristic discharges of ungauged rivers,
bedload of their study river. By contrast, Bagnold equation provided they are not located in dryland areas.
performs poorly with Gereb Oda bedload rates, which are 5. The unit runoff of Ethiopian rivers is comparable to that
best predicted by the empirical power formula of Martin of other African rivers with a catchment area as much as
(2003), yet based on unit stream power. The different per- three orders of magnitude larger. The flow data indicate
formance of these stream power-based equations in pre- that annual runoff volume of more than 50 % of the study
dicting bedload of the Kulfo and Gereb Oda rivers stands rivers is highly variable and, though the unit annual
likely in the ephemeral character and the finer, sandy bed runoff is relatively high with respect to the catchment
material of the latter, whereas the former is a gravelly-sand size, the high values of the annual runoff variation
bed and an almost permanent braided river. coefficient may reduce the potential exploitation of river
water for agriculture irrigation.
6. Time series of annual runoff were constructed for the
4.6 Conclusions 1968–2004 interval and the normalized data of the study
rivers averaged. The interpolating line show a very
Rivers of Ethiopia present a variety of drainage networks, modest decreasing trend indicating that no evident
channel morphologies, flow discharges, and sediment change in annual runoff volume has occurred in the
transport rates, reflecting the complex combination of tec- Ethiopian rivers during the last decades.
tonics, volcanism, physiography and climate. Geomorphic 7. The suspended sediment yield of the Ethiopian rivers is
and hydrologic data have been analyzed and led to the fol- not correlated with the catchment area nor it decreases
lowing conclusions: following a decaying pattern, as reported in the classic
1. In the Rift Valley, recent axial and minor transverse study of the global sediment yield by Milliman and
faults have markedly affected the river pattern of the Meade (1983). By contrast, the annual sediment volume
Ketar and the Awash rivers, with evidence of older of Ethiopian rivers, obtained from field measurements,
abandoned channels. varies significantly (R2 = 0.84) with the catchment area.
114 P. Billi et al.
8. Though the Ethiopian rivers have a comparatively Billi P (2007) Morphology and sediment dynamics of ephemeral
smaller catchment areas, their suspended sediment yield streams terminal reaches in the Kobo basin (northern Welo,
Ethiopia). Geomorphology 85:98–113
is of the same order of magnitude of other, much larger Billi P (2008) Flash floods, sediment transport and bedforms in the
rivers of Africa. This result confirms the high rate of soil ephemeral streams of Kobo basin, northern Ethiopia. CATENA 75
erosion that is taking place in the Ethiopian highlands as (1):5–17
reported by a number of studies (e.g., Hurni 1983; Billi P, Bedri O (2010) Sediment transport of the Blue Nile River at
Khartoum. Quatern Int 226:12–22
Nyssen et al. 2004, 2007; Tamene and Vlek 2008). Billi P (2011) Flash flood sediment transport in a steep sand-bed
ephemeral stream. Int J Sediment Res 26(2):193–209
Acknowledgments The Ministry of Water, Irrigation and Energy of Billi P, Alemu YT, Ciampalini R (2015) Increased frequency of flash
Ethiopia is acknowledged for it provided the most of the flow data used floods in Dire Dawa, Ethiopia: change in rainfall intensity or human
in this paper. This study was carried out with University of Ferrara impact? Nat Hazards 75. doi:10.1007/s11069-014-1554-0
2012 and 2013 FAR funds. Bisson PA, Nielsen JL, Palmason RA, Grove LE (1982) A system of
naming habitat types in small streams, with examples of habitat
utilization by salmonids during lowstream flow. In: Armantrout NB
(ed) Proceedings of a symposium on acquisition and utilization of
References aquatic habitat inventory information. Western Division of the
American Fisheries Society, Portland, pp 62–73
Boateng I, Bray M, Hooke J (2012) Estimating the fluvial sediment
Abdo KS, Fiseha BM, Rientjes THM, Gieske ASM, Haile AT (2009)
input to the coastal sediment budget: a case study of Ghana.
Assessment of climate change impacts on the hydrology of Gilgel
Geomorphology 138:100–110
Abay catchment in Lake Tana basin, Ethiopia. Hydrol Process
Brayshaw AC (1984) Characteristics and origin of cluster bedforms in
23:3661–3669
coarse-grained alluvial channels. In: Koster EH, Steel RJ (eds)
Ahmed AA (2008) Sediment in the Nile River system. UNESCO-IHP-
Sedimentology of gravels and conglomerates, vol 10. Canadian Soc
International Sediment Initiative, Khartoum
Petroleum Geol, Memoir, pp 77–65
Alemu YT (2009) Socio-economic impact of flooding in Dire Dawa,
Bridge JS (2003) Rivers and floodplains: forma, processes, and
Ethiopia. Unpublished MSc. thesis, International Center for Water
sedimentary record. Blackwell, Oxford
Hazard and Risk Management, Tsukuba, Japan
Carling PA (1987) Hydrodynamic interpretation of a boulder berm and
Avery S (2010) Hydrological impacts of Ethiopia’s Omo basin on
associated debris-torrent deposits. Geomorphology 1(1):53–67
Kenya’s Lake Turkana water levels and fisheries. African Devel-
Collick AS, Easton ZM, Ashagrie T, Biruk B, Tilahun S, Adgo E,
opment Bank, Tunis
Awulachew SB, Zeleke G, Steenhuis TS (2009) A simple semi-
Awulachew SB, Yilma AD, Loulseged M, Loiskandl W, Ayana M,
distributed water balance model for the Ethiopian highlands. Hydrol
Alamirew T (2007) Water resources and irrigation development in
Process 23:3718–3727
Ethiopia. Working Paper 123, International Water Management
Conway D (2000) The climate and hydrology of the upper Blue Nile.
Institute, Colombo
Geogr J 166(1):49–62
Ayalew L, Yamagishi H (2004) Slope failures in the Blue Nile basin, as
Conway D (2005) From headwater tributaries to international river:
seen from landscape evolution perspective. Geomorphology
observing and adapting to climate variability and change in the Nile
57:95–116
basin. Global Env Change 15:99–114
Bagnold RA (1980) An empirical correlation of bedload transport rates
Conway D, Persechino A, Ardoin-Bardin S, Hamandawana H, Dieulin C,
in flumes and natural rivers. Proc Royal Soc (London)
Mahe G (2009) Rainfall and water resources variability in sub-saharan
372A:453–473
Africa during the twentieth century. J Hydrometeorol 10(1):41–59
Balthazar V, Vanacker V, Girma A, Poesen J, Golla S (2013) Human
Corti G, Manetti P (2012) Geologia e paesaggi della rift valley in
impact on sediment fluxes within the Blue Nile and Atbara River
Etiopia. Consiglio Nazionale delle Ricerche, Pacini, Pisa
basins. Geomorphology 181:231–241
Demissie M (2008) Regional flood frequency analysis for upper Awash
Baratti E, Montanari A, Castellarin A, Salinas JL, Viglione A, Bezzi A
sub-basin (upstream of Koka). MSc thesis, Addis Ababa Univeristy
(2012) Estimating the flood frequency distribution at seasonal and
Di Baldassarre G, Elshamy M, van Griensven A, Soliman E, Kigobe M,
annual time scales. Hydrol Earth Syst Sci 16:4651–4660
Ndomba P, Mutemi J, Mutua F, Moges S, Xuan Y, Solomatine D,
Bates BC, Kundzewicz ZW, Wu S, Palutikof JP (2008) Climate
Uhlenbrook S (2011) Future hydrology and climate in the River
Change and Water. Technical Paper of the Intergovernmental Panel
Nile basin: a review. Hydrol Sci J 56(2):199–211
on Climate Change, IPCC Secretariat, Geneva
Easton ZM, Fuka DR, White ED, Collick AS, Biruk Ashagre B,
Belete A (2009) Climate change impact on Lake Abaya water level.
McCartney M, Awulachew SB, Ahmed AA, Steenhuis TS (2010) A
MSc thesis, Addis Abaab University, Addis Ababa
multi basin SWAT model analysis of runoff and sedimentation in
Billi P (2000) A study on the geomorphology and sediment yield of the
the Blue Nile, Ethiopia. Hydrol Earth Syst Sci 14:1827–1841
Meki river (Great Lakes Region, Ethiopian Rift Valley): preliminary
EEA (European Environment Agency) (2012) Cliamte change, impacts
results. In: Lenzi MA (ed) Dynamics of water and sediments in
and vulnerability in Europe. An indicator-based report. Report 12/
Mountain Basins, vol 20. Quaderni di Idronomia Montana Special
2012, EEA Copenhagen
Issue, pp 145–157
EM-DAT (2013) The international disaster database. http://www.
Billi P (2004) Sediment yield of a closed river system: the Meki River,
emdat.be/result-country-profile
western margin of the Ethiopian rift valley. In: Proceedings of the
FAO (2000) Global river sediment yields database. http://www.fao.org/
IAHS international symposium on sediment transfer through the
nr/water/aquastat/sediment/index.stm
fluvial system. Moscow, 2–6 Aug 2004, pp 94–100
Gebrehiwot SG, Ilstedt U, Gardenas AI, Bishop K (2011) Hydrological Melesse A, Abtew W, Dessalegne T, Wang X (2010) Low and high
characterization of watersheds in the Blue Nile basin, Ethiopia. flow analyses and wavelet application for characterization of the
Hydrol Earth Syst Sci 15:11–20 Blue Nile River system. Hydrol Process 24:241–252
Gebremicael TG, Mohamed YA, Betrie GD, van der Zaag P, Teferi E Milliman JD, Meade RH (1983) World-wide delivery of river sediment
(2013) Trend analysis of runoff and sediment fluxes in the upper to the oceans. J Geol 91:1–21
Blue Nile basin: a combined analysis of statistical tests, physically- Montgomery DR, Buffington JM (1997) Channel-reach morphology in
based models and landuse maps. J Hydrol 482:57–68 mountain drainage basins. Geol Soc Am Bull 109(5):596–611
Girma NT, Horlachen HB (2004) Investigation of performance of Nyssen J, Poesen J, Moeyersons J, Deckers J, Haile M, Lang A (2004)
sediment transport formulas in natural rivers based on measured Human impact on the environment in the Ethiopian and Eritrean
data in Kulfo River, Southern Ethiopia. In: Proceedings of the Lake highlands—a state of the art. Earth-Sci Rev 64(3–4):273–320
Abaya research symposium, vol 4. Addis Ababa, 10–12 Nov 2004, Nyssen J, Descheemaeker K, Haregeweyn N, Haile M, Deckers J,
pp 35–42 Poesen J (eds) (2007) Lessons learnt from 10 years research on soil
Grant GE, Swanson FJ, Wolman MG (1990) Pattern and origin of erosion and soil and water conservation in Tigray. Tigray Liveli-
stepped-bed morphology in high-gradient streams, western cas- hood papers no. 7, Mekelle: Zala-Daget Project, Mekelle Univer-
cades, Oregon. Geol Soc Am Bull 102:340–352 sity, K.U. Leuven, Relief Society of Tigray, Africamuseum and
Gumindoga W (2010) Hydrological impact of landuse change in the Tigray Bureau of Agriculture and Rural Development
upper Gilgel Abay River basin, Ethiopia. TOPMODEL application. Oliver JE (1980) Monthly precipitation distribution: a comparative
MSc thesis, International Institute for Geo-Information Science and index. Prof Geographer 32:300–309
Earth Observation, Enschede Omuto CT, Vargas RR, Paron P (2009) Soil erosion and sedimentation
Guzman CD, Tilahun SA, Zegeye AD, Steenhuis TS (2013) Suspended modelling and monitoring framework of the areas between rivers
sediment concentration–discharge relationships in the (sub-) humid Juba and Shabelle in southern Somalia. FAO-SWALIM Technical
Ethiopian highlands. Hydrol Earth Syst Sci 17:1067–1077 Report No. L-16, Nairobi
Hadley Centre (2005) Climate change, rivers and rainfall. Met Office, Petts GE (1983) Rivers. Butterworths, London
Exeter Poppe L, Frankl A, Poesen J, Admasua T, Dessied M, Adgoc E.,
Haregeweyn N, Poesen J, Govers G, Verstraeten G, de Vente J, Deckers J, Nyssen J (2013) Geomorphology of the Lake Tana basin,
Deckers J, Meryersons J, Haile M (2008) Sediment yield variability Ethiopia. J Maps 93: 431–437
in northern Ethiopia: a quantitative analysis of its controlling Rientjes THM, Haile AT, Kebede E, Mannaerts CMM, Habib E,
factors. Catena 75:65–76 Steenhuis TS (2011) Changes in land cover, rainfall and stream flow
Hickin ED (1972) Pseudomeanders and point dunes—flume study. Am in Upper Gilgel. Abbay catchment, Blue Nile basin—Ethiopia.
J Sci 272:762–799 Hydrol Earth Syst Sci 15:1979–1989
Hurni H (1983) Soil erosion and soil formation in agricultural Sagri M, Bartolini C, Billi P, Ferrari G, Benvenuti M, Carnicelli S,
ecosystems of Ethiopia and Northern Thailand. Mt Res Dev Barbano F (2008) Latest Pleistocene and Holocene river network
3:131–142 evolution in the Ethiopian lakes region. Geomorphology 94:79–97
IPCC (Intergovernmental Panel on Climate Change) (2007) Synthesis Sena L, Woldemichael K (2006) Disaster prevention and preparedness.
report. IPCC Secretariat, Geneva Ethiopia Public Health Training Initiative, Jimma
Kebede S, Travi Y, Alemayehu T, Marc V (2006) Water balance of Sima BA (2011) Flow regime and land cover changes in the Didessa
Lake Tana and its sensitivity to fluctuations in rainfall, Blue Nile sub -basin of the Blue Nile river, south-western Ethiopia: combining
basin. Ethiop J Hydrol 316(1–4):233–247 empirical analysis and community perception. MSc thesis, Depart-
Kim U, Kaluarachchi JJ, Smakhtin VU (2008) Climate change impacts ment of Aquatic Science and Assessment Department of Urban and
on hydrology and water resources of the upper Blue Nile River Rural Development SLU, Swedish University of Agricultural
basin, Ethiopia. IWMI Research Report 126, International Water Sciences, Uppsala
Management Institute, Colombo Steenhuis TS, Collick AS, Easton ZM, Leggesse ES, Bayabil HK,
Knighton D (1993) Fluvial forms and processes: a new perspective. White ED, Awulachew SB, Adgo E, Ahmed AA (2009) Predicting
Arnold, London discharge and sediment for the Abay (Blue Nile) with a simple
Koster EH (1978) Transverse ribs: their characteristics, origin and model. Hydrol Process 23:3728–3737
paleohydraulic significance. In: Miall AD (ed) Fluvial sedimentol- Summerfield MA, Hulton NJ (1994) Natural control of fluvial
ogy, vol 5. Canadian Soc Petroleum Geol, Memoir, pp 161–186 denudation rates in major world drainage basins. J Geophys Res
Legesse D, Abiye TA, Vallet-Coulomb C, Abate H (2010) Streamflow 99(B7):13871–13883
sensitivity to climate and land cover changes: Meki River, Ethiopia. Sutcliffe SV, Parks YP (1999) The hydrology of the Nile. IASH Spec
Hydrol Earth Syst Sci 14:2277–2287 Publ No 7, Institute of Hydrology, Wallingford
Leopold LB, Wolman MG (1957) River channel pattern: braided Syvitski JPM, Milliman JD (2007) Geology, geography and humans
meandering and straight. US Geol Surv Prof Paper 282B:39–85 battle for the dominance over the delivery of fluvial sediment to the
Leopold LB, Wolman MG, Miller JP (1964) Fluvial processes in coastal ocean. J Geol 115(1):1–19
geomorphology. Freeman, San Francisco Tamene L, Vlek PLG (2008) Soil erosion studies in Ethiopia. In
Mahe G, Lienou G, Descroix L, Bamba F, Paturel JE, Laraque A, Braimoth AK, Vlek PLG (eds) Land use and soil resources.
Meddi M, Habaib H, Adeaga O, Dielun C, Chahnez Kotti F, Springer, New York, pp 73–100
Khomsi K (2013) The rivers of Africa: witness of climate change Tamene L, Park SJ, Dikau R, Vlek PLG (2006) Analysis of factors
and human impact on the environment. Hydrol Processes 27:2105– determining sediment yield variability in the highlands of northern
2114 Ethiopia. Geomorphology 76:76–91
Martin Y (2003) Evaluation of bed load transport formulae using field Tesemma ZK, Mohamed YA, Steenhuis TS (2010) Trends in rainfall
evidence from the Vedder River, British Columbia. Geomorphology and runoff in the Blue Nile basin: 1964–2003. Hydrol Process
53:75–95 24:3747–3758
McDougall I, Morton WH, Williams MAJ (1975) Age and rate of Uhlenbrook S, Mohamed Y, Gragne AS (2010) Analyzing catchment
denudation of Trap Series basalts at Blue Nile gorge, Ethiopia. behavior through catchment modeling in the Gilgel Abay, upper Blue
Nature 254:207–209 Nile River basin, Ethiopia. Hydrol Earth Syst Sci 14:2153–2165
116 P. Billi et al.
UNESCO (2007) Flow Regimes from International Experimental and http://www.daac.ornl.gov from Oak Ridge National Laboratory
Network Data (FRIEND) of the River Nile basin. FRIEND/NILE, Distributed Active Archive Center, Oak Ridge
Final Project Report, UNESCO Regional Office, Cairo Wale A, Rientjes THM, Gieske ASM, Getachew HA (2009) Ungauged
Vanmaercke M, Zenebe A, Poesen J, Nyssen J, Verstraeten G, Deckers catchment contributions to Lake Tana’s water balance. Hydrol
J (2010) Sediment dynamics and the role of flash floods in sediment Process 23:3682–3693
export from medium-sized catchments: a case study from the semi- World Bank (2008) Tana and Beles integrated water resources
arid tropical highlands in northern Ethiopia. J Soils Sediments 10 development project. World Bank Report No: 43400-ET
(4):611–627 Yemane K, Bonnefille R, Faure H (1985) Palaeoclimatic and tectonic
Vorosmarty CJ, Fekete BM, Tucker BA (1998) Global River implications of Neogene microflora from the northwest Ethiopian
Discharge, 1807–1991, V. 1.1 (RivDIS). Data set. Available online highlands. Nature 318:653–656
Planation Surfaces and the Long-term
Geomorphological Evolution of Ethiopia 5
Mauro Coltorti, Dario Firuzabadi, Andrea Borri, Pierlorenzo Fantozzi,
and Pierluigi Pieruccini
Abstract
Four major planation surfaces (PS) characterize the Ethiopian geology and landscape. They
were modelled near or at sea level before the Ordovician (PS1), before the Late Triassic (PS2),
before the Cenomanian (PS3), and before the Oligocene (PS4). These are unconformities in the
sedimentary sequence recognizable across the entire country and the surrounding regions that
due to uplift can be locally exhumed to generate wide steps in the landscape. At the top of the
sequence, that also corresponds to the higher parts of the highlands, the flat depositional surface
of the continental flood basalts (CFB) is preserved over large areas. However, a series of deep
large palaeovalleys dissecting the CFB have been recognized on the water divides of the Afar,
the Main Ethiopian Rift (MER), the Southern Ethiopian Rift (SER), and the Somalian and
Sudan lowlands that are the main geomorphological features of the country. The palaeovalleys
testify to a radial pattern that originated from the dome created by the emplacement of the CFB.
They are easily recognizable to the south of Dire Dawa and in the northern part of the country.
They were modelled over the pre-volcanic bedrock but in most of the highlands were buried
under later volcanic products. The deep erosion of the dome and the later volcanic filling can
explain the results of the previous thermochronological investigation. Unfortunately, the
thickness of the CFB could have led to the reset of the apatite fission tracks (AFT) during the
Oligocene and it is difficult to separate the effects of deep erosion from that of uplift, also
considering that they can be closely related. In any case, a progressive incision of the thicker
part of the dome is documented soon after the CFB deposition in the Blue Nile Gorge and
during the Miocene and the Plio-Pleistocene in many other parts of the country.

Keywords
Planation surfaces Supersequence Plateau Uplift Erosion palaeovalleys
Geomorphology Ethiopia
5.1 Introduction Aden Gulf, the break-up already occurred and the drifting
stage is ongoing (Bosworth et al. 2005). The Afar area is the
Ethiopia is one of the few regions on Earth where conti- triple junction where the two more evolved branches join the
nental break-up is ongoing, generating peculiar large-scale third branch, the Main Ethiopian Rift (MER) (Mohr 1962,
geomorphological features. In the southern Red Sea and the 1967; Huchon et al. 1991; Wolfenden et al. 2004) which is a
major ca. NNE–SSE oriented tectonic depression that hosts a
series of lakes (Zway, Langano, Abijata, Shala, Awassa,
M. Coltorti (&) D. Firuzabadi A. Borri P. Fantozzi
Abaya, Chamo) (Fig. 5.1). The northern sector of the MER
P. Pieruccini has an almost regular width of ca. 90–100 km, but moving
Dipartimento di Scienze Fisiche, della Terra e dell’Ambiente, southward, it enlarges and the border faults split to create
Via di Laterina, 8, 53100 Siena, Italy the Southern Ethiopian Rift (SER) or “broad rift zone” (BRZ,
e-mail: mauro.coltorti@unisi.it
118 M. Coltorti et al.
Fig. 5.1 Digital elevation model of Ethiopia (Shuttle Radar Topography Mission 2005) with the main geomorphological features. In the SER, the
escarpment delimiting the horsts has been associated with the Western and Eastern Ethiopian Escarpment although they are in many cases fault
escarpments deeply remodelled by erosional processes. Main localities mentioned in the text: 1 Mekellé; 2 Amba Alaji; 3 Dire Dawa;
4 Kella; 5 Chorora; 6 Ziway Lake; 7 Abijata Lake; 8 Langano Lake; 9 Shala Lake; 10 Abaja Lake; 11 Chamo Lake; 12 Turmi
Ebinger et al. 2000; Pik et al. 2008)—a series of tectonic bounded by four major planation surfaces that were buried
basins, the westernmost of which continue southward gen- and later exhumed due to uplift. A clear relationship exists
erating the Turkana Lake, a part of the East African Rift between pre-CFB sedimentation and planation surfaces that
System (EARS). The Afar and the MER are sinking and act as have been modelled at sea level and later uplift, rifting and
a trap for volcanic products and continental deposits. By erosional processes. This work is aimed to show the impor-
contrast, the Nubian Plateau to the west and the Somalian tance, extent and significance of the planation surfaces and
Plateau to the east are rising (Mohr 1986; Faure 1975; Weissel discuss the relationships with the CFB depositional surface,
et al. 1995; Pan et al. 2002; Pik et al. 2006; Pizzi et al. 2006). uplift, rifting and erosional processes, especially the evolution
The uplift activated deep incision of the drainage network, of the drainage network that has led to the present-day geo-
especially to the west and east of the rift (Fig. 5.1). Incision is morphological setting.
responsible for the typical landscape of the country, made of
stepped slopes with the remains of the depositional surface of
the continental flood basalts (CFB; Trap) preserved at higher 5.2 The Sedimentary Sequence
elevations. The dissection allowed the exposure of a series of and the Planation/Unconformities
unconformities along the sides of the valley as well as along Surfaces
the margin of the rift. These correspond to buried planation
surfaces that separate one depositional cycle from another. The oldest rock, or “basement”, is made of Proterozoic
Coltorti et al. (2007) demonstrated that in northern Ethiopia, complexes that include intrusive and ultramafic rocks, highly
there were four major sedimentary cycles older than the CFB, metamorphosed rocks, schists and metasandstones
5 Planation Surfaces and the Long-term Geomorphological … 119
(Mohr 1962; Arkin et al. 1971; Aklilu et al. 1978; Merla aeolian sediments (Bosellini et al. 1997, 2001; Wol-
et al. 1979; Kazmin et al. 1980; Tefera et al. 1990). These ela 2008). The occurrence of very well sorted and rounded
rock formed at a depth of over 20 km in the crust, but before evenly laminated sandstones and quartz conglomerate indi-
the Early Palaeozoic, they were uplifted and moulded into an cates the interlayering with beach and coastal environment
extensive flat planation surface, named PS1 in the northern (i.e. the type area to the south of Mekelle). The age of the
part of the country (Coltorti et al. 2007). In fact, to the north Amba Aradam Formation is attributed to the Early Creta-
of the Mekelle outlier, this major unconformity is buried by ceous because of the age of the ca. 10-m-thick “Orbitolina
the Middle–Upper Ordovician continental and marine suc- Limestones” found in the Dire Dawa area (Bosellini
cessions (Enticho Sandstone Formation; Arkin et al. 1971), et al. 2001). A series of red buried palaeosoils are inter-
including tillites (Edaga Arbi Glacials; Beyth 1972; Dow bedded in the lower and upper part of the succession and are
et al. 1971; Saxena and Assefa 1983; Bosellini et al. 1997). also found in the other coeval formation cropping out in
The thickness of this succession does not exceed a few Horn of Africa and in Arabia (i.e. Yesomma Sandstones in
hundred metres, but this likely underestimates the original Somalia and Mukalla Fm in Yemen; Bosellini et al. 2001).
thickness because a part of it was later eroded. In fact, a Although detailed studies of these soils in Ethiopia are
second major planation surface or unconformity cuts these lacking, their red colour, some clay illuviation and Fe/Mn
rocks that are preserved only to the north of Mekelle and in precipitation features are the main indicators of pedogenic
the western side of this outlier. This is the PS2 by Coltorti processes. In the palaeosoils we observed in Tigray, the
et al. (2007) although firstly described as 1st erosional thickness is limited and does not exceed few metres. The
(peneplain) cycle by Merla and Minucci (1938). This lack of a saprolite or deep weathering horizons affecting
unconformity is buried by the Adigrat Sandstones, deposited the parent material suggests short pedogenetic intervals
in a meandering alluvial plain (Bosellini et al. 1997), which similar to the Last Interglacial palaeosoils in Italy that
are also known as the “Lower Sandstone” (Blanford 1869; evolved in ca. 20 ka (Coltorti and Pieruccini 2006). The
Merla and Minucci 1938). The proximity to the sea level is thickness of the Amba Aradam Formation does not exceed
indicated by the occurrence of beach deposits made of 200 m, but it is much less in many places. Yet, it is a
evenly laminated, well rounded and sorted quartz and minimum evaluation due to the later erosive processes.
gravels (Merla et al. 1979; Garland 1980; Tefera et al. 1990). The Amba Aradam Formation contains the youngest
The thickness varies, and as much as 700-m-thick deposits marine deposits which are preserved on the highlands up to
are preserved to the south-west of Mekelle (Beyth 1972). elevation of 2,500–2,800 m (i.e. the Amba Aradam-type area
The transition from the continental/coastal sandstones to the in Tigray) and constitute a topographic reference for the later
overlying marine Antalo supersequence is reported to be uplift (Coltorti et al. 2007).
gradual in Tigray and marked by 20- to 30-m-thick shales Another very flat planation surface or unconformity cuts
with calcarenite and sandstone intercalations (Bosellini the pre-basaltic sedimentary successions including the
et al. 1997). The age of this transgression is Late Callovian– basement. This was firstly described by Blanford (1869) and
Early Oxfordian. The Antalo supersequence in Tigray is later indicated as “pre-volcanic erosion” (Merla and Minu-
made of limestone, and shales deposited during the Late cci 1938), “pre-trappean peneplanation” (Mohr 1962) or PS4
Jurassic in a shallow water marine transgressive succession by Coltorti et al. (2007). The PS4 was buried under the CFB
(Bruni and Fazzuoli 1977; Bosellini et al. 1997). Massive (Trap volcanics of Blanford 1870). These are basaltic and
limestones characterize the lower part of the sequence, trachytes lava flows with different regional names (Arussi,
whereas in the upper part, shales with thin layers of coquina Bale, Wallega, Omo, Amba Alaji, Ashangi, Blue Nile, etc.;
limestone, calcarenites and gypsum are dominant (“Agula Merla et al. 1979; Tefera et al. 1990). The lava flows orig-
Shales” of Merla and Minucci 1938). In northern Ethiopia, inated from fissural activity and local volcanic centres that
the distribution of the sedimentary facies is also affected by a nowadays are largely eroded and only locally preserved as
series of synsedimentary normal faults oriented from skeleton volcanoes. The maximum thickness of the CFB has
WNW–ESE to E–W. The sequence and faults are cut by a been reported up 2,000 m. During this phase, a series of sills,
major angular unconformity named PS3 by Coltorti dikes and laccolites were also emplaced inside the older
et al. (2007) that, according to Bosellini et al. (2001), cor- sedimentary successions (Merla et al. 1979; Schultz
responds to a sedimentary hiatus and related erosion that et al. 2008), especially in the Antalo supersequence.
lasted ca. 20 Ma. The unconformity is buried under the Radiometric dates seem to constrain the emplacement of the
Amba Aradam Formation, also known as the “Upper CFB in the Afar area during the Oligocene, between 30.5
Sandstones” (Shumburo 1968; Getaneh 1991; Bosellini and 28.5 Ma (Baker et al. 1996; Hoffman et al. 1997;
et al. 1997, 2001), mostly made of sandstones and subor- Rochette et al. 1998; Couliè et al. 2003), while in southern
dinately conglomerate and shales of fluvial origin with local Ethiopia, it seems to have occurred around 40 Ma (Ebinger
et al. 1993; George et al. 1998). The lava flows expanded Chernet et al. 1998) and is interstratified with the Butajira
radially from the highlands locally filling shallow pale- Ignimbrites dated between 4.2 and 3 Ma (Woldegabriel
ovalleys such as the Juba in Somalia, over 500 km in dis- et al. 1990). Almost at the same time, apparently between
tance from the plateau (Carmignani et al. 1983; Abdirahim 5.6 and 5 Ma, a series of Rift Margin Silicic Centres erupted
et al. 1993). The continuous and slow basalt outpouring (Chernet et al. 1998). To the south of Addis Ababa, Pliocene
occurred over an almost flat landscape. Bohannon volcanic activity occurred along the Ambo lineament, also
et al. (1989, p. 1687) after a review of the geology of the Red known as the Addis Ababa Embayment, the westward
Sea margin stated that “the entire north-eastern part of the continuation of faults delimiting the southern margin of the
Afro/Arabian continent was almost flat and near sea level Afar. These lava flows belong to the Wechecha Formation
from Late Cretaceous to at least Early Oligocene”. In fact, and are dated between 4.3 and 3 Ma (Chernet et al. 1998),
the basalts along the Juba River, after filling the valley overlying the Addis Ababa basalts dated between 6.6 and
outflowed along the watersheds creating wide lava fields still 5 Ma. In the same area, a younger set of volcanoes (Bishoftu
partially preserved in western Somalia (Abbate et al. 1994). Fm) were active between 2.8 and 0.88 Ma (Chernet
The CFB are capped by giant ignimbrites and more acid et al. 1998). The basalts of the Bishoftu Formation are found
volcanic rocks that received local names (Jimma, Amba also on the rift floor, where the volcanic activity is included
Alaji, Magdala, etc.; Merla et al. 1979; Tefera et al. 1990; in a single unit, the Bofa Formation (EIGS 1978, 1985). In
Ayalew et al. 2002). After the emplacement of major ig- the MER, they are overlain by the volcanic Wonji Group
nimbrites that were almost synchronous with the CFB, two attributed to 3.5–1.5 Ma (Kazmin et al. 1980) and more
other events occurred at 15 and 8 Ma (Ayelew et al. 2002). recently considered younger than 1.8 Ma (Chernet
In any case, an erosional surface interrupting the et al. 1998; Boccaletti et al. 1999; Abebe et al. 2005, 2007).
emplacement of the CFB was recognized in the Blue Nile Continental sedimentary succession is interlayered in these
Gorge area (Gregnanin and Peccirillo 1974; Merla volcanic products during the Pliocene and the Pleistocene.
et al. 1979) and named “Ashangi peneplain” because it cuts
the Ashangi basalt (Blue Nile Basalts of Merla and Minu-
cci 1938). This and later unconformities deeply differ from 5.3 Methods
the previous planation surfaces because they seal a con-
trasted topography and have much more local significance. In order to describe the main geomorphologial features of
A younger phase of flood basalt emplacement (Tarmeber- the area and the relationships between the planation surfaces/
Megezez Fm of Kazmin et al. 1980) occurred during the unconformities and the geological history of Ethiopia, a
Middle Miocene (13 Ma, Mohr 1967 and Kazmin general geological map of the country was derived from the
et al. 1980; 10 Ma, Chernet et al. 1998). Kazmin et al. (1980) available regional geological map (1:2,000,000 scale; Tefera
associated the volcanic centres of these younger lava flows et al. 1990). The elevation data were derived from the SRTM
to the opening of the Afar depression. However, according DEM (Shuttle Radar Topography Mission; United States
to Chernet et al. (1998) and Wolfenden et al. (2004), the Geological Survey 2005). Unfortunately, geological vector
opening of the Afar was associated with the emplacement of data of the entire study area were not available. Therefore,
the Lower Afar Stratoid Series (LASS). These flood basalts the map of Tefera et al. (1990), once georeferenced, was
have been dated for between 4.5 and 5.6 Ma (Chernet digitalized.
et al. 1998), although an age of 0.63 Ma obtained near Dire The analysis of the geological formations (i.e. Amba
Dawa by Audin et al. (2004), if confirmed, would cast some Aradam Fm and Antalo Fm), their origin (sedimentary,
doubts on the previous chronological setting. In the south- volcanic, continental, etc.) and the presence of the main
western Afar, the LASS are interlayered with the Chorora unconformities/planation surfaces made possible to distin-
Formation, diatomite lacustrine layers bearing mammal guish four supersequences (or supersynthems) (Fig. 5.2).
fauna (Tiercelin et al. 1980; Juch 1980). They are a few tens The map of the supersequences is in WGS84 datum and
of metre thick and displaced by normal faults, but appar- UTM 38 N projection (Fig. 5.2). A series of geological cross
ently, there is no evidence of synsedimentary faulting. They sections (Fig. 5.3) were obtained from the SRTM DEM,
are still preserved almost at the top of the Somalian Plateau available for the study area at 3-arcsec resolution (ca. 90 m)
suggesting that during their deposition, at least, this part of using ESRI ArcGIS 10.2 software. The sections were
the Afar depression was still not yet developed. The LASS obtained from series of elevation points, automatically cre-
are in turn buried by ignimbrites that can be up to 300 m ated at a distance of 100 m from each other, a short distance
thick (Nazreth Group; Butajira ignimbrites and tuffs; considering the large area investigated. A vertical exagger-
Woldegabriel et al. 1990; Chernet et al. 1998). The Nazreth ation factor of 10× was adopted in order to improve the
Group is dated between 6.8 and 1.8 Ma (Bigazzi et al. 1993; visual resolution of the topography.
Fig. 5.2 Geological map with the Palaeozoic to Oligo-Miocene rocks grouped in supersynthems/supersequences bounded by planation/
unconformity surfaces. PS1 separates the basement from the Palaeozoic rocks; PS2 separates the basement and the Palaeozoic rocks from Triassic–
Jurassic rocks; PS3 separates the basement, Palaeozoic and Triassic–Jurassic rocks from Cretaceous rocks; PS4 separates the older sedimentary
rocks from Oligocene–Miocene continental flood basalts. The geological map of Tefera et al. (1990) has been used to locate sedimentary sequence
and their boundaries
southern Arabia (Kohlan Formation, Simmons and

5.4 The Planation Surfaces in Ethiopia Al-Thour 1994; Toland et al. 1994) and northern Kenya
(Mansa Guda Formation, Bosellini et al. 2001) (Fig. 5.4).
The thickness of the Adigrat Sandstones varies across the
The planation surfaces are a major feature of the Ethiopian country from a few tens of metres in the eastern regions
highlands. They are buried but well recognizable because, (south-east Ethiopia, Somalia and Yemen) up to 1,000 m in
on both sides of the plateaus, deep valley incisions, along the Blue Nile Gorge suggesting that the sandstones covered
with rift systems tectonics, allow for the observation of the a landscape made of “depressions and structural highs”
lowermost parts of the successions. The preservation of the (Bosellini et al. 2001). The difference in thickness might also
PS1 is limited to the northern part of the country, whereas its be related to the occurrence of a neglected “paraconformity”
absence to the south of Mekelle suggests that levelling of located at the base of the Antalo Formation. In fact, in many
PS2 caused erosion of the Palaeozoic rocks in the rest of the places, the Antalo Formation lies directly over the Pre-
country. The modelling of PS2 was likely preceded by a cambrian basement such as to the south-west of Mekelle
large-scale folding that generated wide synforms where the (Bosellini et al. 1997), in the Danakil Alps, and in the Dire
Palaeozoic sequences are still preserved. PS2 is a very flat Dawa area (Bosellini et al. 2001). The end of the Adigrat
surface and recognizable in most of the country because the Formation deposition is placed at the appearance of the
Adigrat Sandstones lie over the Precambrian basement in the lowermost layers of the Antalo Formation, but an exact age
rest of the country and also in Somalia (Fantozzi et al. 2002), for the sandstones is unknown, and therefore, a hiatus cannot
b Fig. 5.3 Geological cross sections of the Ethiopian supersynthems. and volcanic rocks (Miocene to present); 2 volcanic rocks (supersyn-
Location in Fig. 5.2. The absence of a supersynthem in a section them 4; Trap, Oligocene–Miocene); 4 coastal and fluvial sediments
testifies to efficacy of erosional processes during the modelling of (supersynthem 3; Amba Aradam Formation, Cretaceous); 5 coastal to
different planation surfaces. The boundaries are extrapolated, and marine deposits (supersynthem 2; Triassic–Jurassic); 6 river, coastal,
major uncertainties are due to the variation in thickness that aeolian and glacial sediments (supersynthem 1; Palaeozoic); 7 meta-
corresponds to the different importance of erosional processes, morphic and intrusive rocks (Precambrian)
especially where the sequence is buried under CFB. 1 sedimentary
Fig. 5.4 Relationship between planation surfaces and sedimentary sequence in Ethiopia and surrounding regions. PC precambrian rocks; CFB
continental flood basalts; VSC post-CFB volcanic and sedimentary sequence
be excluded. Sedimentary evidence of repeated fluctuations In this case, the “structural highs” would have existed before
of the sea level during the early stages of the Antalo marine the deposition of the younger rocks. However, in order to
transgression is also reported by Bosellini et al. (1997, make it easier to read the supersequence Map and to pinpoint
1999). The occurrence of a hiatus would be also supported the major events, all the Triassic–Jurassic sequences of
by a lateritic crust at the top of the Adigrat Sandstone For- Ethiopia were grouped into a single Adigrat–Antalo super-
mation in different parts of the Mekelle outlier (Tigray) and sequence. The Adigrat and Antalo limestones occur in the
in Eritrea (Merla and Minucci 1938; Abul-Haggag 1961; Blue Nile, buried under shales (Mugher Fm; Sagri
Bosellini et al. 1997). The marked difference in thickness of et al. 1998; Gani et al. 2009). In the Dire Dawa area, the
the Adigrat Formation might also be explained with the supersequence includes the Antalo limestones, unconform-
occurrence of tectonic movements and the formation of wide ably covered by black limestones and marls (Dire Dawa Fm),
synforms and antiforms. PS3 would have erased the suc- marly limestones and marls with shales and gypsum (Daghani
cession in the antiforms and preserved it in the synforms. Shales), and finally by reef limestones (Gildessa Fm)
Fig. 5.5 The typical Ethiopian landscape is made of Amba—flat- are Jurassic shales and marly limestones. The top of the mountain
topped mountains, such as the Amba Aradam to the south of Mekellè. results from exhumation of the PS4 because only a thin veneer of CFB
The PS3 is preserved at the base of the steeper cliff made of marine and is locally preserved at the very top (Photograph P. Pieruccini, ca.
fluvial sandstones and conglomerates. Below the unconformity there 13° 20′ 29.95″N 39° 30′ 41.61″E, looking S)
Fig. 5.6 The PS3 along the watershed separating the SER from the locally, it was also slightly remodelled during PS4 because to the north,
Omo valley. Scattered spherical quartz gravels, easily attributable to a below the CFB, the Amba Aradam is not reported (Photograph
marine origin, are found over the bedrock suggesting that the surface P. Pieruccini, ca. 5° 19′ 9.58″N 36° 34′ 46.04″E, looking E)
was covered by deposits of the Amba Aradam. It is possible that
Fig. 5.7 The exhumation of the PS3 in southern Ethiopia and the a few kilometres to the north is covered by over 1,000 m of CFB
erosion of a thick cover of the CFB is easily recognizable north of Turmi (background). (Photograph P. Pieruccini, 5° 36′ 7.92″N 36° 41′ 36.11″E,
because the surface modelled over the basement rocks (foreground) only looking N)
(Bosellini et al. 2001). The westernmost outcrops of the Uarandeb (Agula Shales-equivalent) and Gabredarre Forma-
Adigrat–Antalo supersequence in central-south Ethiopia tion, a lateral facies variation of the Gildessa Formation of
are found in the Blue Nile Gorge sequence that represents the Dire Dawa (Bosellini 1992) (Fig. 5.4). In Yemen, the equiv-
emersion of the Jurassic sediments from under the CFB and alent of the Antalo limestones is the Amran limestones around
later volcanics that cover both sides of the central and northern San’a and Suqhraq in southern Yemen, also covered by shales
MER. The Antalo limestones crop out with local facies vari- (Bosellini et al. 2001). They were recognized also in Somalia
ation in Ogaden and Somalia where they correspond to the (Fantozzi 1998; Fantozzi and Ali Kassim 2002). A tectonic
superimposed Hamanley (Antalo limestones-equivalent), phase is documented after the deposition of the Adigrat–
Fig. 5.8 The Eastern Ethiopian Escarpment (EEE) at the Afar shaped by large-scale landslides (Photograph M. Coltorti, ca.
Window, to the south of Debresina. The top of the escarpment is 9° 50′ 0.92″N 39° 44′ 23.96″E, looking E)
modelled over the CFB. The slope is dissected by river erosion and
Fig. 5.9 The Western Ethiopian Escarpment (WEE) to the north of movements (Photograph M. Coltorti, ca. 6° 25′ 16.92″N,
Chenca modelled over CFB whose depositional surface is preserved at 37° 37′ 15.77″E, looking S)
the top of the relief. The slope is affected by large-scale gravitational
Antalo supersequence that could be responsible also for the distances because the overlying Amba Aradam Formation is
minor Jurassic unconformities described in Dire Dawa district also made of strongly cemented rocks and there is no contrast
(Russo et al. 1994, 1996; Bosellini et al. 1997). to trigger processes of selective erosion and re-exhumation
PS3 affected the Adigrat–Antalo supersequence in most (Fig. 5.5).
parts of the country and in the nearby regions (Bosellini The supersequence 3, made by the Amba Aradam For-
et al. 1997, 2001). It is remarkable that the Antalo Formation, mation, crops out in large part of the country, including the
grouped together with the Adigrat Formation in the map north-eastern border of the Somalian Plateau (Beyth 1972;
(Fig. 5.2), is absent in most of the southern part of the country Hutchinson and Engels 1970; Assefa 1991; Russo et al. 1994;
where only the Adigrat Formation crops out. This would Bosellini et al. 1997; Sagri et al. 1998). It is missing in the
suggest the existence of another unconformity between these westernmost part of the Blue Nile Gorge and in the southern
two formations that is difficult to recognize in the field. part of Ethiopia. It could be inferred that it was never
Exhumed remains of the PS3 are rarely observed over long deposited in the former areas, but its presence on the western
Fig. 5.10 The Somalian Plateau and the Eastern Ethiopian Escarpment (EEE) modelled over the CFB, the Amba Aradam and the Jurassic rocks
deeply dissected by fluvial erosion (Photograph M. Coltorti, ca. 8° 13′ 34.03″N, 39° 50′ 20.00″E, looking SE)
Fig. 5.11 The Nubian or Ethiopian Plateau to the south of Chenca (Photograph M. Coltorti, ca. 6° 32′ 5.25″N, 37° 37′ 53.94″E, looking S)
side of the Turkana Lake (Merla et al. 1979; Nyagah 1997) The PS4 planation is responsible for the complete erosion
suggests its later erosion. Reworked scattered spherical quartz and/or the reduced thickness of the older supersequence and
gravels and sands are found over the basement along the new the erosion of the PS that had previously affected the Pal-
road from Jinka to Turkana Lake, where wide and flat aeozoic rocks. In fact, in many parts of the Ethiopian
watersheds largely correspond to the exhumed PS3 (Figs. 5.6 highland as well as in northern Kenya, the CFB lay directly
and 5.7). The thin layers of sands described by David- on the Precambrian basement (Merla et al. 1979; Tefera
son (1983) along the borders of the SER underlying the CFB et al. 1990). There are no elements in Ethiopia to establish
could also be attributed to the supersequence 3. This forma- the position relative to the sea level of the CFB but, as
tion crops out also in Yemen (Taouilah Fm, Simmons and Al- reported in Chap. 2, Bohannon et al. (1989) suggested that
Thour 1994; Quishin Fm, Toland et al. 1994) (Fig. 5.4), in the marine sequences preserved along the margins of the
Afar, in Ogaden, Somalia and north-western Kenia (Merla Red Sea were deposited over an almost flat landscape close
et al. 1979; Tefera et al. 1990), as well as in Sudan (Upper to the sea. In fact, in Saudi Arabia, there are also marine
Sandstones, Whiteman 1971; Yassin et al. 1984). strata interbedded within the CFB.
Fig. 5.12 The Nubian Plateau to the south of Debresina. This photograph is taken from the same locality of Fig. 5.8 (Photograph M. Coltorti, ca.
9° 50′ 0.92″N 39° 44′ 23.96″E) but looking W. Note the strong contrast of relief’s energy
Fig. 5.13 The Somalian Plateau deeply dissected by fluvial erosion (Photograph M. Coltorti, ca. 8° 10′ 35.61″N, 39° 45′ 14.73″E looking S)
later sedimentary sequence to the east. In plan view, the

5.5 Large-scale Geomorphological boundary is very irregular, with very deep indentations
Features corresponding with the major rivers and their tributaries such
as the Blue Nile or the Tekeze in the west or the Juba and
From west to east, the major geomorphological units of the Wabi Shebelle in the east (Figs. 5.1 and 5.13). This indicates
country (Figs. 5.1 and 5.2) are as follows: 1. Sudan low- a long-lasting headward retreat that eroded the original
lands; 2. Nubian Plateau; 3. Afar; 4. MER; 5. SER basins; margins of the CFB. In central Somalia, lava fields associ-
6. Somalia Plateau; and 7. Somalian lowlands. ated with the CFB are preserved at over 500 km from the
The transition from the Sudan and the Somalian low- margin of the plateau (Abdirahim et al. 1993; Abbate
lands to the Nubian Plateau and the Somalian Plateau, et al. 1994) but have been eroded in the rest of the Somalian
respectively, is a sharp boundary located in correspondence lowlands. The Blue Nile Gorge and the Tekeze valley on the
with the Western (WEE) and Eastern Ethiopian Escarpment western side are among the best examples of deep dissection
(EEE) (Fig. 5.1). These escarpments can exceed 1,000 m in involved in the retreat of the WEE.
height, and the major jump commonly develops within a few On the other hand, the boundaries of the rift-related tec-
kilometres of horizontal distance (Figs. 5.8 and 5.9). It tonic depressions (Afar, MER, SER, Figs. 5.14 and 5.15) are
delimits the top of the depositional surface of the CFB or the usually sharp and rectilinear and marked by well-expressed
later ignimbrite and volcanic rocks (Figs. 5.10, 5.11 and faulted escarpments (Mohr 1962, 1967, 1987; Wolfenden
5.12), from the basement rocks exposed in front of the et al. 2004; Pizzi et al. 2006; Abebe et al. 2007). The rivers
escarpment to the west and south-east and/or the Jurassic and have short valleys with stepped profiles, indicating a very
Fig. 5.14 The main fault escarpment to the east of Asela. In this sector, the displacement is accommodated by dozens of faults (Photograph M.
Coltorti, ca. 8° 2′ 1.45″N, 39° 6′ 1.71″E)
Fig. 5.15 The main fault escarpment along the southern margin of the Afar, to the south of Dire Dawa (Photograph M. Coltorti, ca. 9° 29′ 53.97″N,
41° 45′ 27.33″E)
recent age for rifting processes, especially in the MER and

SER. The morphology at the transition from the MER to the 5.6 The Evolution of the Drainage
Afar (Ebinger et al. 2000; Pik et al. 2008) is not much different Network and the Lesson
but more complex because the faults of the MER cut the faults from the Dire Dawa Watershed
of the south-eastern corner of the Afar (Chernet et al. 1998;
Wolfenden et al. 2004; Abebe et al. 2007). The southern and Our investigations on the watersheds (Sect. 1 in Fig. 5.16)
western borders of the Afar have a rather rectangular drainage that separate the Afar to the north from the Somalian lowlands
pattern influenced by the presence of rift-in-rift basins. The to the south of Dire Dawa allowed us to recognize two dif-
rivers have stepped profiles and are much longer than in the ferent drainage patterns. To the north of the watershed, along
MER but much shorter than those crossing the WEE and the steep main escarpment, drainage pattern is rectangular,
the EEE. In the Dire Dawa area, the margin of the Afar is with long branches trending ca. E–W, parallel to the fault
characterized by a more than 20-km-long, deeply incised and system bounding the Afar. Minor branches are normal to
terraced pediment (Pizzi et al. 2008), the distal part of which is these lineaments. The southward-running streams originate
sealed by LASS basalts dated 0.6 Ma (Audin et al. 2004). from small wind gaps hosting small lakes at ca. 2,000 m asl.
Fig. 5.16 A rectangular pattern

influenced by faults is
recognizable in the Dire Dawa
area, to the north of the
watershed, marked by Sect. 1.
A straight southward-oriented
drainage is recognizable to the
south. The latter is associated
with the continuation of a valley
that was beheaded by the
activation of the fault escarpment.
A larger palaeovalley at higher
elevation on the watershed was
already modelled over the
basement and the pre-Oligocene
sedimentary sequence (Sect. 1). It
is also well recognizable slightly
to the south from Sect. 2 where
the higher part of the slope is
modelled over the CFB. The flat
upper part of the longitudinal
profile of the river is influenced
by the original palaeovalley
before beheading, while the lower
tract is due to headward erosion
They slightly dissect the exhumed PS2 that contains also the south-west (Fig. 5.18) and ca. 3,200 on the Kundudu, a
limited patches of the Adigrat Formation (Fig. 5.17). At ca. residual relief, to the east (Fig. 5.16, Sect. 2). To the south of
2,200 m, above the smaller wind gaps, the divide shows the the watershed, the rivers are straight and their longitudinal
remnants of a much larger beheaded valleys (Fig. 5.16, profile shows a bimodal slope with the inner gradient very flat
Sect. 1 dotted line). The flanks of this large palaeovalley are and increasing in steepness southward (Fig. 5.16, Sect. 3).
actually the slopes modelled over the CFB whose deposi- The upper low-gradient profile is evidently influenced by the
tional surface reaches ca. 3,500 m on the Mt Gara Mullata to presence of the original paleovalley modelled before the
Fig. 5.17 The rectilinear southward-oriented valley to the south of the water divide of the Afar (Photograph M. Coltorti, ca. 9° 20′ 52.16″N,
41° 54′ 13.62″E looking S)
Fig. 5.18 The southern slope of the Gara Mullata, the upper part of which is modelled over the CFB. The PS4 is exhumed at the contact with the
Amba Aradam Formation (Photograph M. Coltorti, ca. 9° 12′ 7.11″N, 41° 47′ 21.26″E looking W)
beheading, whereas the lower steeper part can be related to erosion is also documented in Eritrea because CFB are
the recent headward erosion. recorded in very limited patches (Balestrieri et al. 2005).
Therefore, the plateau is made of two major geomor- The wind gaps on the watersheds around the Afar indi-
phological features at different elevations. The higher one is cate an original radial Palaeodrainage that could have orig-
the original depositional surface of the CFB, and at lower inated from the centre of the CFB now lowered under the
elevations are the remnants of palaeovalleys, up to 1,200 m Afar floor (Fig. 5.16b). Palaeovalley also occur along the
deep, that testify to the original drainage pattern before the border of the MER, although in many parts of the plateau,
activation of the rifting. Similar palaeovalleys have been the palaeomorphology is covered under younger volcanic
observed in other parts of the southern margin of the Afar, products. Along the north-western side of the Abaja Lake,
where the original watersheds are easily recognizable due to for example, the water divide with the Omo River is located
the occurrence of mesas and residual reliefs topped with the at very low elevation (palaeovalley west of 9 and 10 in
CFB at the sides of the wind gap (Figs. 5.1 and 5.16a). Fig. 5.1). Similar evidences are also found to the south of the
Palaeovalleys are also found along the margins of the Chamo Lake, on both sides of the MER, where the CFB are
Nubian Plateau and especially in northern Ethiopia. In fact, deeply dissected and the Precambrian basement crops out
to the north of the Amba Alagi (ca. 3,500 m asl), the CFB is (paleovalleys south of 12 in Fig. 5.16). The importance of
completely eroded and the watershed between the Nile river erosion of the watershed in this area comes out from the
and the Afar is located around 2,000–2,200 m. A deep abrupt reduction in thickness and locally the absence of
the CFB that, slightly to the north, on both sides of the Nile gorge while the overlying rocks have ages of 400–
Chamo Lake are over 2,000 m thick. 520 Ma has two possible explanations (Pik et al. 2003): 1.
Most probably, the so-called Addis Ababa embayment the erosion of thicker CFB and 2. the existence of a thicker
(Chernet et al. 1998) corresponds to one of these palaeoval- Mesozoic sedimentary cover before the emplacement of the
leys because there are over 1,000 m between the top of the CFB. The last explanation seems more reliable because of
preserved CFB to the north and the plateau to the south of the occurrence of the planation surfaces and a thickness of
Addis Ababa, where the surface is made of Plio-Pleistocene the Mesozoic sediments exceeding 2,000 m in the northern
volcanics (AAE in Fig. 5.16). However, we do not exclude part of the country (Getaneh 1991). An enhanced heating of
that the Addis Ababa embayment followed E–W faults along the crust during the emplacement of the CFB such as sug-
the so-called Yerer-Tellu Wellel lineament (Chernet gested by Mock et al. (1999) is not consistent with the
et al. 1998; Tsegay et al. 1998). apatite (U-Th)/He as old as 45–107 Ma (Pik et al. 2003). On
the other hand, the younger apatite (U-Th)/He ages of ca.
40 Ma found at 1,100 m below the base of the CFB in the
5.7 Uplift and Erosion Blue Nile Gorge is still older than the emplacement of the
CFB and thus indicates that the CFB never exceeded 900 m,
Many authors claimed that Ethiopian relief was created a thickness that would led to the reset of the apatite (U-Th)/
rapidly around 30 Ma ago, slightly before and synchro- He system. A sedimentary cover of at least 900 m (2,000 m
nously with the emplacement of the CFB (Baker et al. 1996; of the apatite/He PRZ—1,100 m of the thickness of the
Pik et al. 2003, 2008), when the Ethiopian lithosphere sequence above the sample) was therefore eroded before
moved over the Afar mantle plume (Hart et al. 1989; 30 Ma (end of the modelling of the PS4). Pik et al. (2003)
Schilling et al. 1992). The present elevation would be pre- inferred, on the base of thermal modelling, that in the Blue
served by the sub-lithospheric mantle upwelling and outflow Nile Gorge area, the incision of the plateau started soon after
from under Afar (Wolfenden et al. 2004). However, a strong the emplacement of the CFB, 29–25 Ma ago, while in the
uplift in later time and especially during the Plio-Pleistocene Anger River catchment, a minor tributary, 10–5 Ma ago.
is claimed by others (Merla et al. 1979; Juch 1980; Fau- However, if this occurred in an already uplifted area or this
re 1975; Mohr 1986; Weissel et al. 1995; Bohannon was the consequence of the dissection of the dome generated
et al. 1989; Pizzi et al. 2006). The amount of long-term uplift by the accumulation of the CFB is difficult to establish.
can be roughly estimated from the mean elevation of (U/Th)/He thermochronometry and apatite fission tracks
the sedimentary succession preserved in an area. In fact, the (AFT) dating of the western rift escarpment have been per-
successions are mostly made of alluvial plain and littoral formed in Eritrea, west of Asmara, but with controversial
deposits and the planation surfaces were levelled at or close results (Abbate et al. 2002; Ghebreab et al. 2002; Balestrieri
to the sea level. This also includes the PS4 because the CFB et al. 2005; Drury et al. 1994, 2006). Ghebreab et al. (2002)
expanded over an almost flat PS4 without filling deep did not find any correlation between age and elevation, with
valleys. younger ages also at the edge of the escarpment suggesting
An estimation of the plateau uplift rates of 0.015 mm/year that tectonic denudation was the major agent at work. AFT
can be inferred taking into account that the Amba Aradam ages across normal faults in the middle part of the escarpment
Formation (ca. 105 Ma) is commonly located between 2,000 indicate an old age of cooling between 23 and 16 Ma, fol-
and 2,500 m, a rate also suggested by McDougall et al. (1975). lowed by a thermal resetting event during Late Miocene and
A minimum uplift rate of 0.066–0.083 mm/year is obtained if Early Pliocene (9–3 Ma) that the authors attributed to the
we consider that the area remained at or close to the sea level rising of hot waters along faults and fractures. On the other
also during the emplacement of the CFB at ca. 30 Ma. How- hand, Balestrieri et al. (2005) found ages older than 33 Ma
ever, values of 0.5–1 mm/year have been suggested if the moving towards the plateau except for two ages found at the
movements are associated with flexural flank uplift, with a base of the escarpment that gave ca. 10 and 16 Ma, respec-
much later onset (Faure 1975; Weissel et al. 1995). tively. According to these results, the plateau surface was not
The data obtained from titanite (U-Th)/He method (Pik buried anymore under 2,000 m of sediments since 200 Ma
et al. 2003) that considered a partial retention zone (PRZ) at and in many cases also 400 Ma (close to Asmara). The
a depth 4–5 km with a gradient of 25–30 °C/km and a younger ages, but always older than 30 Ma, found along the
surface temperature of 20 °C, and apatite (U-TH)/He method escarpment seem to testify that an important erosion preceded
(PRZ at 2 km with the same gradient) suggest that a sedi- the formation of PS4, as in the Blue Nile Gorge (Pik
mentary cover with a thickness exceeding 2,000 m was et al. 2003). Along the escarpment, Balestrieri et al. (2005)
never deposited after 40 Ma in the Blue Nile Gorge and the found that AFT dating would indicate a rapid cooling during
northern part of the country. However, the youngest titanite the Early–Middle Miocene (20 Ma) and suggest a post-break-
(U-TH)/He age of 213 Ma found at the bottom of the Blue up erosion of more than 3.5–4 km on top of the escarpment.
However, slightly to the east, the upper member of the Dogali Peninsula. These authors also described sites on the eastern
Formation, dated between 15 and 6 Ma, testifies to marine side of the Red Sea where marine deposits interlayer with the
deposition and hence the existence of the Red Sea graben. CFB. However, the deposition of the volcanic products that
The occurrence of deep palaeovalley affecting the CFB are 2,000 m in places, also if occurred over a flat surface,
could easily explain the AFT date from the Kella horst, created reliefs. AFT confirm that at the periphery of the
located on the western side of the MER, 100 km to the south dome, where CFB thickness was reduced, there was no total
of Addis Ababa (Di Paola et al. 1993; Tsegay et al. 2010). In reset of the system during the Oligocene. In the Blue Nile
fact, in this area, a thickness between 1,500 and 2,300 m of Gorge, the tracks were annealed due to the burial under the
rocks was eroded after 7–6 Ma. CFB, and later, as early as 29–25 Ma, rocks started to cool
U-Th/He ages between 34 and 49 Ma have been observed again because valley erosion was already at work (Pik
in the upper part of the scarp delimiting the Hamar Range et al. 2003). Miocene erosion is documented in the SER (Pik
from the Chew Bahir in the SER (Pik et al. 2008). Ages et al. 2008), where over 2,000 m of CFB have been eroded, as
between 19 and 12 Ma are found at the bottom of the well as along some minor tributaries of the Nile (Pik
escarpment. Therefore, erosion was active before the for- et al. 2003). These authors associated the evidence in the SER
mation of PS4, but afterwards, since 19–12 Ma, a thickness to the activation of faults, but a down-cutting along lines of
of ca. 2,000 m was eroded, not necessarily due to fault weakness in the basement, later reactivated as normal faults,
activity. The later event led to the complete erosion of the would have a similar effect. A Miocene incision is docu-
over 2,000-m-thick CFB from the Hamar Range. Thus, the mented by AFT and U-Th/He analysis also along the south-
fact that the CFB are missing in many parts of the plateaux western borders of the Red Sea and the north-western side of
and that their present-day borders do not coincide with their the Afar in Eritrea (Ghebreab et al. 2002; Balestrieri
original border suggests that downwearing due to weath- et al. 2005), although the different results leave the inter-
ering and progressive removal of weathering products from pretation of subsequent drainage evolution controversial. In
the top of the CFB was coupled with much faster erosion any case, the upper member of the Dogali Fm in Eritrea lying
along the major palaeoriver systems at the edge of the Red Sea testifies to the opening of this
sector at ca. 15 Ma. In our opinion, there is still not enough
evidence to establish whether the nearby plateau was already
5.8 Conclusion elevated at that time or the uplift, as we also support, was
triggered as a consequence of flexural flank uplift (Fau-
The planation surfaces of Ethiopia are unconformities re 1975; Mohr 1986; Weissel et al. 1995).
between different depositional sequences recognizable in The recognition of beheaded palaeovalleys on the
the entire country and the surrounding regions. The country watersheds of the Afar cut over 1,200 m below the deposi-
was planated at sea level at least four times before the tional surface of the CFB allows us to establish the main
emplacement of the CFB, ca. 30 Ma. A ravinement surface location of the original radial pattern. No chronological
generated by marine erosion during a major transgression, constraints are available for the beheading of the valleys and
that is a plain of marine erosion (Davis 1899; Johnson 1930; therefore the activation of the faults in Afar. However, the
Coltorti et al. 2007), is the best explanation for the modelling recognition of hanging palaeovalleys that in the western side
of these surfaces. In fact, except PS4, the planation surfaces of the Afar and the MER are commonly buried under vol-
are covered by marine and fluvial sediments. Alternative canic products allows also to establish that the plateaux are
hypothesis of etchplanation and pediplanation (King 1975) complex geomorphological features that bear witness of the
has been already ruled out by Coltorti et al. (2007). In fact, oldest evolution of the drainage network. A large drainage
the lacking of deep weathering profiles below the planation system possibly occupied the depression, known as the
surfaces rules out etchplanation and the very flat nature of the Addis Ababa Embayment, located to the south of the capital
surfaces rules out pediplanation. The latter hypothesis would area (Fig. 5.19). If the Kella horst belongs to this drainage
also require very long periods of arid conditions and tectonic system and the complete erosion of the CFB is attributed to a
stability, a combination that it was also difficult to achieve. westward drainage system, the beheading in this area would
Moreover, there is no evidence that before or at the onset of have occurred after 3 Ma that is the age of the basalt filling
the CFB deposition, the area was elevated. This had already the palaeovalleys (Tsegay et al. 2010). This should also be
been stated by Bohannon et al. (1989) after a review of the the lower limit for the activation of border faults of the
geology of the north-eastern Africa and the Arabian central MER.
5
Planation Surfaces and the Long-term Geomorphological …
Fig. 5.19 a Present-day drainage network with location of major wind gaps corresponding to palaeovalleys hanging on the watershed of the Afar and northern MER; b Reconstruction of the
original drainage resulting from the dissection of the dome generated by the CFB. The Addis Ababa Embayment (AAE in Fig. 5.1) corresponds to a palaeovalleys now diverted towards the Omo
River Valley but was originally oriented to the Nile basin. Note that out of the Afar, we used the present-day topography that is the result of Miocene and later dissection. The reconstructed dome
has a topography at the same level or slightly higher than the highest relief of the plateau
133
Pediplanation worked during the arid phases of the with Ethiopian continental flood basalt. Geoch Cosmochim Acta 66
Quaternary at the feet of the Western and Eastern Escarp- (8):1429–1448
Baker J, Snee L, Menzies M (1996) A brief Oligocene period of flood
ment (WEE and EEE), and a pediplain over 20 km long, volcanism in Yemen: implications for the duration and rate of
nowadays terraced, has been observed along the southern continental flood volcanism at the Afro-Arabian triple junction.
margin of the Afar prior to 0.6 Ma (Pizzi et al. 2008). Earth Planet Sci Lett 138:39–55
However, at a larger scale, the major features of the Afar Balestrieri ML, Stuart FM, Persano C, Abbate E, Bigazzi G (2005)
Geomorphic development of the escarpment of the Eritrean margin,
margins, the WEE and the EEE, are to be associated with southern Red Sea from combined apatite fission tracks and (U/Th)/
headward river erosion. He thermochronometry. Earth Planet Sci Lett 231:97–110
Beyth M (1972) Paleozoic-Mesozoic sedimentary basin of Mekele
Acknowledgments We thank Maria Laura Balestrieri for the critical Outlier, northern Ethiopia. Am Ass Petrol Geol Bull 56(12):2426–
review of the paper and the useful suggestions to improve it. The paper 2439
was also realized thanks to the support received by the University of St Bigazzi B, Bonadonna FP, Di Paola GM, Giuliani A (1993) K-Ar and
Petersburg, FL, USA, NSF Project Award Abstract #1027607 “An fission track ages of the last volcano tectonic phase in the Ethiopian
Ethnoarchaeological and Archaeological Study of the Gamo Caste Rift Valley (Tullu Moye area). In: Geology and mineral resources of
System in south-western Ethiopia”. We also would like to thank in Somalia and surrounding regions. Istituto Agronomico Oltremare,
Addis Ababa Dr. Yonas Desta, General Manager of ARCCH; Wz. Firenze, Relazioni Monografie 113:311–322
Mamito, Director at the National Museum of Ethiopia; Getachew Blanford WT (1869) Observations on the geology and zoology of
Senishaw, Director of Cultural Heritage Collection and Lab Service; Abyssinia. Macmillan, London, 407 pp
and Desalegn Abebaw Andualem, Director of Cultural Heritage Boccaletti M, Mazzuoli R, Bonini M, Trua T, Abebe B (1999) Plio-
Research. quaternary volcano-tectonic activity in the northern sector of the
Main Ethiopian Rift: relationships with oblique rifting. J Afr Earth
Sci 29:679–698
Bohannon RG, Naeser CW, Dwight LS, Zimmermann RA (1989) The
References timing of uplift, volcanism and rifting peripheral to the Red Sea: a
case of passive rifting? J Geophys Res 94:1683–1939
Abbate E, Sagri M, Sassi FP (1994) Geological map of Somalia. In: Bosellini A (1992) The continental margin of Somalia. Structural
Geology and mineral resources of Somalia and surrounding regions, evolution and sequence stratigraphy. Am Ass Petrol Geol Mem
Istituto Agronomico Oltremare Firenze, Relazione Monografica, 53:185–205
Firenze Bosellini A, Russo A, Fantozzi PL, Getaneh A, Tadesse S (1997) The
Abbate E, Balestrieri ML, Bigazzi G (2002) Morphostructural devel- Mesozoic succession of the Mekele outlier (Tigre province,
opment of the Eritrean rift flank (southern Red Sea) inferred from Ethiopia). Mem Sci Geol 49:95–116
apatite fission track analysis. J Geophys Res 107 pp Bosellini A, Russo A, Schroeder R (1999) Stratigraphic evidence for an
Abdirahim MM, Ali Kassim M, Carmignani L, Coltorti M (1993) early Aptian sea-level fluctuation: the Graua Limestone of south-
Geomorphological evolution of the upper Jubba river (southern eastern Ethiopia. Cretac Res 20:783–791
Somalia). In: Geology and mineral resources of Somalia and Bosellini A, Russo A, Assefa G (2001) The Mesozoic succession of Dire
surrounding regions, Istituto Agronomico Oltremare Firenze, Rela- Dawa, Harar Province, Ethiopia. J Afr Earth Sci 32(3):403–417
zione Monografica 113:241–250 (Firenze) Bosworth W, Huchon P, McClay K (2005) The Red Sea and Gulf of
Abebe B, Coltorti M, Pizzi A (2005) Rates of late quaternary Aden Basin. J Afr Earth Sci 43:334–378
deformation along the Wonji fault belt in the Lakes Region, Main Bruni P, Fazzuoli M (1977) Sedimentological observations on Jurassic
Ethiopian Rift. Rend Boll Soc Geol It 1:41–43 and Cretaceous sequences in northern Somalia. Preliminary Report.
Abebe B, Acocella V, Korme T, Ayalew D (2007) Quaternary faulting Boll Soc Geol It 95:1571–1588
and volcanism in the Main Ethiopian Rift. J Afr Earth Sci 48:115–124 Carmignani L, Ali Kassim M, Fantozzi PL (1983) Carta Fotogeologica
Abul-Haggag Y (1961) A contribution to the physiography of northern della Regione di Gedo (Somali sud-occidentale). Quaderni di
Ethiopia. The Athlone Press, University of London, 153 pp Geologia della Somalia, Mogadiscio
Aklilu A, Amenti A, Beyth M, Dow DB, Garland CR, Temesgen H, Chernet T, Hart WK, Aronson JL, Walter RC (1998) New age
Hailu T (1978) Geological map of Adigrat area, Sheet ND 37-7 constraints on the timing of volcanism and tectonism in the northern
(1:250.000). Ministry of Mines, Geological Survey of Ethiopia, Main Ethiopia Rift-southern Afar transition zone (Ethiopia).
Addis Ababa J Vulcanol Geoth Res 80:267–280
Arkin Y, Beyth M, Dow DB, Levitte M, Temesgen H, Hailu T (1971) Coltorti M, Pieruccini P (2006) The last interglacial pedocomplexes in
Geological map of Mekele, Sheet ND 37-11 (1:250.000), Tigre the litho- and morpho-stratigraphical framework of the central-
Province. Ministry of Mines, Energy and Water Resources, northern Apennines. Quat Int 156–157:118–132
Geological Survey of Ethiopia, Addis Ababa Coltorti M, Dramis F, Ollier C (2007) Planation surfaces in northern
Assefa G (1991) Lithostratigraphy and environment of deposition of Ethiopia. Geomorphology 89(3–4):287–296
the late Jurassic-early Cretaceous sequence of the central part of Coulié E, Quidelleur X, Gillot PY, Courtillot V, Lefèvre JC, Chiesa S
northwestern Plateau, Ethiopia. Neues Jahrb Geol Paleontol (2003) Comparative K-Ar and Ar/Ar dating of Ethiopian and
Abhandunglen 182(3):255–284 Yemenite Oligocene volcanism: implications for timing and
Audin L, Quidelleur X, Coulie E, Courtillot V, Gilder S, Manighetti I duration of the Ethiopian Traps. Earth Planet Sci Lett 206:477–492
(2004) Paleomagnetism and K-Ar and 40Ar/39Ar ages in the Ali Davidson A (1983) The Omo River Project: reconnaissance geology
Sabieh area (Republic of Djibouti and Ethiopia). Constrains on the and geochemistry of part of Ilubabor, Kefa, Gemu Gofa and
mechanism of Aden ridge propagation into south-eastern Afar Sidamo, Ethiop Eth Inst Geol Surveys Bull 2:89 pp
during last 10 M year. Geophys J Int 158:327–345 Davis WM (1899) The geographical cycle. Geogr J 14:481–504
Ayalew D, Barbey P, Marty B, Reisberg M, Yirgu G, Pik R (2002) Di Paola GM, Seife-Michael B, Arno V (1993) The Kella Horst: its
Source, genesis, and timing of giant ignimbrite deposits associated origin and significant in crustal attenuation and magmatic processes
in the Ethiopian Rift Valley. In: Geology and Mineral resources of Johnson DW (1930) Planes of lateral corrasion. Science 73:174–177
Somalia and surrounding regions, Istitituto Agronomico Oltremare Juch D (1980) Tectonics of the southeastern escarpment of Afar,
Firenze, Relazione Monografica 113:311–323 (Firenze) Ethiopia. Accad Naz Lincei 47:407–418 (Rome)
Dow DB, Beyth M, Hailu T (1971) Paleozoic rocks recently discovered Kazmin V, Berhe SF, Nicoletti M, Petrucciani C (1980) Evolution
in northern Ethiopia. Geol Mag 198(1):53–59 of the northern part of the Ethiopian rift. Atti Convegni Lincei
Drury SA, Kelley SP, Berhe SM, Collier REL, Abraha M (1994) 47:275–292
Structures related to Red Sea evolution in northern Eritrea. King LC (1975) Planation surfaces upon highlands. Z Geomorph NF 20
Tectonics 13:1371–1380 (2):133–148
Drury SA, Ghebreab W, Deller AME, Talbot CJ, Berhe SM (2006) A McDougall I, Morton WH, Williams MAJ (1975) Age and rates of
comment on “geomorphic development of the escarpment of the denudation of trap series basalt at Blue Nile Gorge, Ethiopia. Nature
Eritrean margin, southern Red Sea from combined apatite fission-track 254:207–208
and (U–Th)/He thermochronometry”. Earth Planet Sci Lett 242:28–432 Merla G, Minucci E (1938) Missione geologica nel Tigrai, vol. 1 “La
Ebinger CJ, Yemane T, Woldegabriel G, Aronson JL, Walter RC serie dei terreni”. Rend Reale Acc Italia, Centro Studi per l’Africa
(1993) Late eocene-recent volcanism and faulting in the southern Orientale Italiana 3:362 pp
main Ethiopian rift. J Geol Soc London 150:99–108 Merla G, Abbate E, Azzaroli A, Bruni P, Canuti P, Fazzuoli M, Sagri
Ebinger CJ, Yemane T, Harding DJ, Tesfaye S, Kelley S, Rex DC M, Tacconi P (1979) A geological map of Ethiopia and Somalia
(2000) Rift deflection, migration, and propagation: linkage of the (1:2.000.000 scale) and comments with map of major landforms.
Ethiopian and eastern rifts, Africa. Geol Soc Am Bull 112:163–176 CNR, University of Florence, Centro Stampa Firenze, 95 pp
EIGS (Ethiopian Institute of Geological Surveys), Ministry of Mines Mock C, Arnaud NO, Cantagrel J-M, Yirgu G (1999) 40Ar/39Ar
and Energy (1978) Geological map of the Nazret Sheet _NC37-15., thermochronology of the Ethiopian and Yemeni basements: reheat-
scale 1:250,000 ing related to the Afar plume? Tectonophysics 314:351–372
EIGS (Ethiopian Institute of Geological Surveys), Ministry of Mines and Mohr PA (1962) The geology of Ethiopia. Addis Ababa University
Energy (1985) Geological Map of the DireDawa Sheet _NC37-12., Press, Addis Ababa, 268 pp
scale 1:250,000 Mohr PA (1967) The Ethiopian rift system. Bull Geophys Observatory
Fantozzi PL (1998) Transition from continental to oceanic rifting in the 11:1–65 (Addis Ababa)
gulf of Aden: structural evidence from field mapping in Somalia and Mohr P (1986) Sequential aspects of tectonic evolution of Ethiopia.
Yemen. Tectonophysics 259:285–311 Mem Soc Geol Ital 447–461
Fantozzi PL, Ali Kassim M (2002) Geological mapping in northeastern Mohr P (1987) Pattern of faulting in the Ethiopian rift valley.
Somalia (Midjiurtinia region): field evidence of the structural and Tectonophysics 143:169–179
paleogeographic evolution of the northern margin of the Somalia Nyagah K (1997) Late Paleozoic through Mesozoic cyclostratigraphy,
plateau. J Afr Earth Sci 34:25–53 sedimentology and depositional history of the Mandera Basin,
Faure H (1975) Recent crustal movements along the Red Sea Gulf Northeast Kenya and implication for hydrocarbon exploration.
of Aden coasts in Afar (Ethiopia and TFAI). Tectonophysics National Oil Corporation, Kenya, 27 pp (unpublished manuscript)
29:479–486 Pan M, Sjoberg LE, Asenjo E, Alemu A, Asfaw LM (2002) An analysis
Gani NDS, Abdelsalam MG, Gera S, Gani R (2009) Stratigraphic and of the Ethiopian Rift Valley GPS campaigns in 1994 and 1999.
structural evolution of the Blue Nile Basin, northwestern Ethiopian J Geodyn 33:333–343
plateau. Geol J 44:30–56 Pik R, Marty B, Carignan J, Lavé J (2003) Stability of the Upper Nile
Garland CR (1980) Geology of the Adigrat Area. Ministry of Mines, drainage network (Ethiopia) deduced from (U-Th)/He thermochr-
Energy and Water Resources Geological Survey of Ethiopia, onometry: implications for uplift and erosion of the Afar plume
Memoir 1:51 pp dome. Earth Planet Sci Lett 215:73–88
George R, Rogers N, Kelley S (1998) Earliest magmatism in Ethiopia: Pik R, Marty B, Carignan J, Yirgu G, Ayalew T (2008) Timing of east
evidence for two mantle plumes in one food basalt province. African Rift development in southern Ethiopia: implication for mantle
Geology 26:923–926 plume activity and evolution of topography. Geology 36:167–170
Getaneh A (1991) Lithostratigraphy and environment of deposition of Pizzi A, Coltorti M, Bekele A, Disperati L, Sacchi G, Salvini R (2006)
the Late Jurassic-Early Cretaceous sequence of the central part of The Wonji fault belt (Main Ethiopian Rift, Ethiopia): structural and
northwestern Plateau, Ethiopia. Neues Jahrb Geol Paleontol geomorphological constraints and GPS monitoring. In: Yurgu G,
Abhandunglen 182:255–284 Ebinger CJ, Macguire PKH (eds) The Afar Volcanic Province
Ghebreab W, Carter A, Hurford AJ, Talbot CJ (2002) Constraints for within the east African Rift System. Geol Soc, 259:191–207
timing of extensional tectonics in the western margin of the Red Sea (special publications)
in Eritrea. Earth Planet Sci Lett 200:107–119 Pizzi A, Pomposo G, Abebe B, Coltorti M (2008) Stile deformativo e
Gregnanin A, Peccirillo EM (1974) Considerazione sulle serie vulca- reticolo idrografico lungo la scarpata del Plateau Somalo al margine
niche e sulle strutture dell’Altopiano Etiopico Centrale. Mem Mus meridionale dell’Afar (Etiopia). Rend Online Soc Geol It 1:137–139
Tridentino Sci Nat 20:79–100 Rochette P, Tamrat E, Féraud G, Pik R, Courtillot V, Ketefo E, Coulon
Hart WK, Woldegabriel G, Walter RC, Mertzman SA (1989) Basaltic C, Hoffmann C, Vandamme D, Yirgu G (1998) Magnetostratigra-
volcanism in Ethiopia: constraints on continental rifting and mantle phy and timing of the Oligocene Ethiopian traps. Earth Planet Sci
interactions. J Geophys Res 94:7731–7748 Lett 164:497–510
Hofmann C, Courtillot V, Féraud G, Rochette P, Yirgu G, Ketefo E, Pik Russo A, Assefa G, Atnafu B (1994) Sedimentary evolution of the
R (1997) Timing of the Ethiopian flood basalt event and implications Abay River (Blue Nile) Basin, Ethiopia. Neues Jb Geol Palaontol
for plume birth and global change. Nature 389:838–841 Monatsh 5:291–308
Huchon P, Jestin F, Cantagrel JM, Gaulier JM, Al Khirbash S, Gafaneh A Russo A, Fantozzi PL, Solomon T, Getaneh A, Neri C, Russo F,
(1991) Extensional deformations in Yemen since Oligocene and the Asfwossen A, Peccerillo A, Valera P (1996) Geological map of the
Africa-Arabia-Somalia triple junction. Ann Tectonicae 2:141–162 Mekele outlier (western sheet). Italian Cooperation—Addis Ababa
Hutchinson RW, Engels GG (1970) Tectonic significance of regional University, Addis Ababa
geology and evaporate lithofacies in northeastern Ethiopia. Phil Sagri M, Abbate E, Azzaroli A, Balestrieri ML, Benvenuti M, Bruni P,
Trans Roy Soc London A267:313–329 Fazzuoli M, Ficcarelli G, Marcucci M, Papini M, Pavia G, Reale V,
Rook L, Tecle TM (1998) New data on the Jurassic and Neogene Al-Husseini MI (ed) Geo 94 The middle east Petroleum Geosci-
sedimentation in the Danakil Horst and northern Afar Depression, ences, 2, Gulf Petrol Ink Bahrain, pp 891–899
Eritrea. In: Crasquin-Soleau S, Barrier E (eds) Perythetis Memoir 3: Tsegay AT, Mazzarini F, Innocenti F, Manetti P (1998) The Yerer-
Stratigraphy and evolution of peri-Tethian platforms. Mem Mus Nat Tullu Wellelvolcanotectonic lineament: a transtensional structure in
Hist Nat Paris, 177:193–214 central Ethiopia and the associated magmatic activity. J Afr Earth
Saxena GN, Assefa G (1983) New evidence on the age of glacial rocks Sci 26:135–150
of northern Ethiopia. Geol Mag 120(6):549–554 Tsegay A, Balestrieri ML, Bigazzi G (2010) The central main Ethiopian
Schilling JG, Kingsley RH, Hannan BB, McCully BL (1992) Nd–Sr– rift is younger than 8 Ma: confirmation through apatite fission-track
Pb isotopic variations along the Gulf of Aden: evidence for Afar thermochronology. Terranova 22(6):470–476
mantle plume-continental lithosphere interaction. J Geophys Res Weissel JK, Malinverno A, Harding DJ (1995) Erosional development
97:10927–10966 of the Ethiopian plateau of Northeast Africa from fractal analysis of
Schultz RA, Mège D, Diot H (2008) Emplacement conditions of igneous topography. In: Barton CC, Pointe PRL (eds) Fractals in petroleum
dikes in Ethiopian traps. J Volcanol Geoth Res 178:683–692 geology and Earth processes. Plenum Press, New York, pp 127–142
Shumburu MM (1968) The Amba Aradom formation (formerly the Whiteman AJ (1971) The geology of the Sudan Republic. Clarendon
upper sandstone). Mobil Petroleum Ethiopia Inc., Asmara, 15 pp Press, Oxford, 290 pp
(unpublished report) Woldegabriel G, Aronson JL, Walter RC (1990) Geology, geochro-
Simmons MD, Al-Thour K (1994) Micropaleontological biozonation of nology, and rift basin development in the central sector of the main
the Amran Series (Jurassic) in the Sana’a Region, Yemen Republic. Ethiopian rift. Geol Soc Am Bull 102:439–458
In: Simmons MD (ed) Micropaleontology and hydrocarbon explo- Wolela A (2008) Sedimentation of the Triassic-Jurassic Adigrat
ration in the middle east. Chapman and Hall, London, pp 43–79 Sandstone Formation, Blue Nile (Abay) Basin, Ethiopia. J Afr
Tefera M, Chernet T, Haro W (1990) Geological Map of Ethiopia, scale Earth Sci 52:30–42
1:2.000.000. Geol Surv Ethiopia, Addis Ababa, Ethiopia Wolfenden E, Ebinger C, Yirgu G, Deino A, Ayalew D (2004)
Tiercelin JJ, Taieb M, Faure H (1980) Continental sedimentary basins Evolution of the northern Main Ethiopian rift: birth of a triple
and volcano-tectonic evolution of the Afar Rift. Accad Naz Lincei junction. Earth Plan Sci Lett 224:213–228
47:491–504 (Rome) Yassin AA, Khalil FA, El Shafie AG (1984) Explanatory note of the
Toland C, Simmons MD, Walkden GM (1994) A new sequence geological map at the scale of 1:2,000,000 of the Democratic
stratigraphic reference section for the Upper Jurassic of Yemen. In: Republic of the Sudan, 63 pp
Part II
Local Studies
Paleoglaciated Landscapes in Simen and Other
High-Mountain Areas of Ethiopia 6
Hans Hurni
Abstract
In the present-day Ethiopia, glaciated landscapes do not exist, but paleoglaciated landscapes
have been documented on a few mountain tops, which have altitudes higher than about
4,350 m asl in northern Ethiopia (Simen Mountains) and about 4,100 m asl in southern
Ethiopia (Arsi and Bale Mountains). Glaciers were associated with the Late Pleistocene cold
stages and reached as far down as 3,760 m asl in northern and 3,200 m asl in southern
Ethiopia. Bale Mountains had the most extensive Late Pleistocene glaciation, covering over
190 km2, followed by Arsi Mountains (about 85 km2). In Simen, the Late Pleistocene glaciers
covered merely 13 km2. In addition, paleo-periglacial slope deposits are found on all above-
mentioned paleoglaciated mountains and in further mountain systems which did not host
glaciers. This allows the reconstruction of the Late Pleistocene paleoclimate as being about
8 °C colder than at present (2014), much more dry, and probably without monsoon, at least in
northern Ethiopia. Most probably in the Early Holocene, the re-emergence of monsoonal rains
led to a strong erosion phase, which was followed by an extended stable phase with soil
formation, building up about 70-cm-deep A-horizons (Andosol) on the paleo-periglacial slope
deposits. These soils have been heavily degraded due to human-induced soil erosion up to
about 3800 m asl since agriculture started several decades to millennia ago.

Keywords

Late Pleistocene glaciation Paleo-periglacial slope deposits Paleoclimate Early Holocene
erosion Holocene soil formation Human-induced soil degradation
6.1 Introduction There are, however, signs of paleoglaciated landscapes on

some of Ethiopia’s highest mountains, featuring moraines
Glaciated landscapes are extremely rare in the contemporary and periglacial slope deposits. The most prominent of such
Africa and non-existent in Ethiopia, as her mountain tops are landscapes are found in the Bale Mountains (Messerli et al.
too low to exceed a permanent present-day snowline. The 1977; Osmaston et al. 2005), Mount Badda in Arsi (Potter
highest mountain, Ras Dejen, reaches 4,540 m asl, while the 1976), and in the Simen Mountains (Hurni 1981). Detecting
current snowline is estimated to be at about 4,900 m asl former moraines is quite demanding, while the paleo-peri-
(2014) and is expected to rise even higher in future decades glacial slope deposits can be seen on virtually all mountain
due to global warming. Although not all authors agree that slopes above about 3,500 m asl in northern Ethiopia, and
there was no recent glaciation in Ethiopia, this would have above about 3,200 m asl in southern Ethiopia. There are
implied colder temperatures by at least 8 °C on average. differences in altitude of these lower limits around each
mountain system, as northerly-facing slopes show the cold-
climate features more prominently than the southerly-facing
H. Hurni (&) ones.
University of Bern, Bern, Switzerland
e-mail: hans.hurni@cde.unibe.ch
140 H. Hurni
6.2 Geographical Setting For the Ethiopian lowlands, Gasse et al. (1980) studied
and Literature Quaternary landforms and paleoclimates, parallel to many
other authors. More recent studies on paleoglaciation fea-
Figure 6.1 presents an overview of Ethiopian mountain tops tures in high-mountain environments in Ethiopia include
above 4,000 m asl. The minimum height of 4,000 m asl is Osmaston et al. (2005), Zech et al. (2005), and Umer et al.
taken because mountain tops with lower altitudes do gen- (2004), including an excellent overview of Quarternay gla-
erally not show paleoglaciation features. In addition, the cation in Africa by Mark and Osmaston (2008).
extent of the Ethiopian Highlands is also given, according to In Simen, features from a paleoglacial period were observed
its agroecological belts (from Kolla, Weyna Dega and Dega by Minucci (1938), Nilsson (1949), Hoevermann (1954), Bu-
to Wurch, cf. Hurni 1998). edel (1954), Werdecker (1955), Mohr (1962), Hastenrath
In the second half of the last century, numerous investi- (1974, 1977), Messerli (1975), Williams et al. (1978), and the
gations and observations were made on paleofeatures in present author (Hurni 1981, 1982, 1986). Geologically, the
these mountain systems. For Bale, Messerli et al. (1977) Simen Mountains are composed of Trap basalts that overlie
presented an abstract of their observations, while for Arsi, Mesozoic sandstones, which in turn rest on the Precambrian
Potter (1976) described paleoglaciation on Mount Badda. basement (Kazmin 1973). The mountain tops above 3,000 m
Fig. 6.1 Overview of mountain tops higher than 4,000 m asl in the light green (Kolla and Weyna Dega, from about 1,000 m to about
Ethiopian Highlands, all with paleo-periglaciation and some with 2,500 m asl), and humid highlands are darker green to bluish (Dega to
paleoglatiation features. The country is coloured according to agro- Wurch, from about 2,500 m up to 4,540 m asl). Size of Ethiopia:
ecological belts: semi-arid lowlands are yellow to brownish (Berha, i.e. 1,106,000 km2; highlands: 575,000 km2. Source from Hurni (1998)
below about 1,000 m asl); sub-humid to humid highlands are olive to
6 Paleoglaciated Landscapes in Simen … 141
asl in Ethiopia are all remnants of ancient volcanoes, consisting “consist of firmly compacted angular blocks, which are not
of Termaber basalts and belonging to the Shield group, being sorted, not regulated in certain directions, and embedded in a
porphyritic, with tuffs and paleosols in between, connected to yellowish-brown matrix of clayey silt.” Such ridges are found
volcanic centres and dating from the Miocene to the Pliocene above 3,780 m to above 4,320 m asl, mainly on north-facing
(Merla et al. 1979). The present-day climate of Simen is slopes (from west to north-east), while only one is found on a
characterised by a unimodal rainy season from May to October, south-facing slope, which is next to the top of Mount Kidis
about 1,500 mm of annual rainfall, and about 8.5 °C average Yared (4,460 m asl), at an elevation of 4,400 m asl (Hurni
annual temperature at 3,600 m asl (adapted from Hurni 1982). 1981). Above these ridges, one can detect the catchment
Other mountain tops show similar characteristics, although areas for accumulation of ice, which usually have a concave
with some deviation in annual rainfall. form (cirque) excavated in basalt rock.
Figure 6.2 gives a section from the map 1:100,000 by Hurni
(1981) with all moraines he detected in Simen during field-
6.3 Paleoglacial Landforms in Simen work in 1974–1977, and Fig. 6.3 is a photograph of the largest
moraine found in the westerly catchment of Ras Dejen, below
6.3.1 Moraines a cirque that steeply descends into the Mesheha Valley.
Figure 6.4 presents an analysis of all paleoglacial moraines
The experienced eye will fairly easily detect former moraines observed in Simen by Hurni (1982) and includes the most
in Simen, the more since the newly built roads to the towns of important characteristics and parameters, such as the paleosnow-
Mekane Birhan in Janamora and to Dilyibza in Beyeda are line and the lower limit of the paleo-periglacial slope deposits.
passing quite near to some of them at an altitude of about The moraines in the Ras Dejen area are the longest, with
4,200 m asl, namely on Mount Bwahit (4,437 m asl), Mount ridges between 70 and 550 m, and their depth varies between
Mesarerya (4,360 m asl), and Mount Ras Dejen (4,540 m asl). 5 and 150 m. Normally, they are singular, but in a few
Moraines, according to Hurni (1986), are rubble ridges that locations, a second lateral moraine can be found behind the
Fig. 6.2 Mountain tops in Simen with paleoglaciation features of the last cold period. Features presented are moraines (full red) with ice flow
direction (black arrows); periglacial slope deposits (pale red), and snowline (red dotted). East–West distance is about 36.5 km. Source section of
map by Hurni (1981)
142 H. Hurni
Fig. 6.3 Panorama of the largest (middle) moraine of the last cold glacier was found towards the photographer, while a smaller glacier
period in the Simen Mountains, western escarpment below Ras Dejen was situated behind the moraine, having itself a small moraine on its
(4,540 m asl), seen from an altitude of about 4,200 m asl. The main right side (H. Hurni, 13 April 2013)
Fig. 6.4 Topographic analysis of

all mapped paleoglaciers and Late
Pleistocene altitudinal belts in
Simen, including the last cold
period snowline as reconstructed
by moraines and cirques using
two methods. The correlation of
snowline with slope aspect is
evident, as is the lower limit of
the periglacial belt being parallel
to the snowline. Source Hurni
(1986)
first one. In the Bwahit/Mesarerya area, moraine ridges are 6.3.2 Slope Deposits
between 70 and 250 m long and up to 15 m thick, hence
considerably smaller. Most moraines represent lateral mor- Paleo-periglacial deposits are widespread throughout the Si-
aines, often with an indication of a terminal stage. The men Mountains. These features can be best observed in the
respective cirques of the paleoglaciers are quite evident, present-day Afro-alpine grasslands between 3,600 m and
particularly within Ras Dejen north-west-facing slopes, 4,000 m asl, such as in the upper Jinbar Valley in the centre of
where rock faces and signs of overdeepening are obvious. the park, on the western slopes of Bwahit and Mesarerya
Fig. 6.5 Paleo-periglacial

deposits on a northerly-facing
slope in the Upper Jinbar Valley,
3,500–4,000 m asl. The deposits
are up to 15 m thick and were
eroded in the Early Holocene,
before soil formation built up
about 70 cm of Andosol A-
horizon over the entire landscape
(H. Hurni, September 1974)
Fig. 6.6 Andosol soil profile as

prevalent on all mountains of the
Ethiopian Highlands between
about 3,200 m asl and about
4,000 m asl, here near Gich Camp
at 3,600 m asl in Simen. Note the
black A-horizon overlaying a
compacted B-horizon, which
represents paleo-periglacial slope
deposits, with the Early Holocene
erosion surface in between. (H.
Hurni, August 1974)
mountains, and on the south-eastern slopes of Ras Dejen consists of firmly compacted angular blocks, which are not
(Fig. 6.2). Their altitudinal occurrence starts below the paleo- sorted, but regulated along or across the fall-line, and embed-
snowline, which according to the occurrence of the moraines ded in a yellowish-brown matrix of clayey silt” (Fig. 6.5). This
was determined by Hurni (1981) at about 4,250 m asl, and regulated formation of stones along the fall-line was interpreted
extends as far down as 3,600–3,400 m asl according to slope as cold-age periglacial solifluction processes.
aspect, similar to the moraines. Slope deposits, according to The thickness of the slope deposits increases downslope
Hurni (1986), are “rubble on trough-shape slopes (which) and can be up to 15 m thick. In some of the steep tributary
144 H. Hurni
Fig. 6.7 Heavily degraded

landscape in the Simen
Mountains due to century-old
agricultural land use without soil
conservation. Looking from
3,300 m asl in a northerly
direction towards the Upper
Jinbar Valley (H. Hurni, 7
February 2013)
valleys to Mesheha Valley, rubble deposits are also found on humus (up to 30 % organic matter) and contains ashes from
the valley floors and as far down as 3,000 m, and they can be volcanic eruptions and Saharan dust, and it is of more or less
up to 40 m thick (Hurni 1981). Their association with the similar depth throughout the area. The formation of such
same formation period, however, could not be clarified dur- deep soils indicates a long-lasting soil formation period
ing the fieldwork, and it may be that they belong to an earlier without much disturbance; hence, slopes must have been
Pleistocene glaciation period. densely covered by vegetation with grasses and trees during
their formation period.
6.3.3 Erosion of Slope Deposits

6.3.5 Land and Soil Degradation
The paleo-periglacial slope deposits have been incised by
gullies, which can be up to 15 m deep (Fig. 6.5). This ero- As in all of Ethiopia, the highland parts of Simen have been
sion took place before a 70-cm-thick Andosol A-horizon settled by farmers already about 2–3 thousand years ago.
was formed on top of them. According to Hurni (1986), “the The “people of Samen” were mentioned by the Greeks about
erosion phase must have taken place immediately after the 2,000 years ago (Kirwan 1972). Such agricultural activity
last cold period, before the upward movement of vegetation required deforestation, which is a form of land degradation,
regrowth took place in those belts. The climate must have because no grain will grow above the tree line, which in
been much wetter, and possibly warmer. In the climatic Simen is currently at about 3,800–4,000 m asl. With the
history of Ethiopia, the erosion phase can be inserted into the subsequent ploughing of steep slopes, soil erosion started to
phase of northward movement of the monsoons around remove the top parts of the Andosols, at a rate of about
12,000–10,000 BP.” 100 tonnes per ha per year or more during cultivation.
Within a relatively short period of time, the topsoils were
thus washed downstream. This process may have been slo-
6.3.4 Soils wed during regular fallowing periods, which were applied
traditionally. Nevertheless, over the centuries, soil degrada-
In all of Simen between about 3,000–4,200 m asl, the An- tion became devastating, as the present-day landscape of
dosol soil type is prevalent. These soils have a dark brown to Simen shows. Soil and water conservation measures were
black A-horizon, which on average is about 70 cm thick only introduced in Ethiopia as of the mid-1970s onwards (cf.
(Fig. 6.6). According to Frei (1978), this horizon is rich in Hurni et al. 2008) (Fig. 6.7).
6.4 Evolution of the Paleoglaciated common practice, while the population in Simen increased
Landscape in the Simen by a factor of 8 since the 1950s (cf. Hurni and Ludi 2000).
Mountains
The features of paleoglaciated landscape as described in 6.5 Other Paleoglaciated Mountain

Sect. 6.3 imply a sequence of processes that logically con- Tops in Ethiopia
dition each other and help interpreting them in a chrono-
logical order and building a chronology. During the Bale Mountains A short study by Messerli et al. (1977) in
glaciation period, the existence of glaciers can be inferred April 1976 showed that the Bale Mountains, which are
from the presence of moraines, and their association with the characterised by an extended high plateau above 4,000 m asl
paleo-periglacial belt appears logic, as no slope deposits were and peaks up to 4,377 m asl, were glaciated similar to the
found inside the paleoglaciated areas. The paleoglaciers are Simen Mountains, but at a much larger scale, with an ice
situated about 800 m lower than present-day glaciers would shield on the Sanetti Plateau at around 4,100 m asl and
be situated, although the latter are not existent due to insuf- individual glaciers extending into the valleys on all sides. In
ficient altitudes of the mountain tops. The climate at that time, the Tegona Valley (near the town of Goba), the valley gla-
according to Hurni (1986), appears to have been about 7 °C cier was found to have had a thickness of 250 m and
cooler, depleted in precipitation and run-off, with a tendency reaching down to an altitude of 3,200 m asl. The Bale
towards winter (November–March) precipitation, and only Mountains thus probably had the largest glaciers in Ethiopia
occasional Summer (June–August) clouding with reduced or in the Late Pleistocene, with a glaciated area of around
missing monsoons. Today, in 2014, this would be about 8 °C 190 km2 according to Osmaston et al. (2005). Zech et al.
due to global warming since 1975. The glaciers were thus not (2005) were able to date lacustrine sediments formed after
historical, as 14 °C dating inside moor deposits on the ground deglaciation on the Sanetti Plateau at 15,500–14,000 years
moraines confirmed, showing an oldest date of 4,120 ± 90 BP BP. The other two features observed in Simen, the paleo-
(sample B-3043 in Hurni 1982). The paleoclimate during the periglacial slope deposits and the post-glacial erosion, were
glaciation period is most likely associated with the Late also observed in Bale, but not studied in detail. Land and soil
Pleistocene, between may be 20,000 and 12,000 BP (cf. degradation, finally, was much less advanced than in the
Gasse et al. 1980). Simen Mountains.
The post-glaciation period, as inferred from the eroded Arsi Mountains Mount Badda (4,170 m asl) in the Arsi
slope deposits, started with intensive natural erosion, when Region was studied by Potter (1976), who found evidence of
the monsoons reached far into the Sahara. This phase ended Late Pleistocene glaciation down to an elevation of 3,650 m
when the temperature increased and vegetation cover moved asl in the west-facing valleys, with an extent of 140 km2,
upslope to the present levels (or above). With vegetation, which is probably an overestimation. Periglacial features
soil formation started to build the observed deep A-horizons were not reported. According to Umer et al. (2004), “there
covering the slope deposits, the gullies eroded into them, and are large, clear terminal moraines on Mount Bada at 3,200–
all other landscape elements between 3,000 and 4,000 m asl. 3,700 m with the glaciated area estimated to be 85 km2 with
This long and stable period must be associated with the equilibrium line altitudes (ELAs) of 3,700 m as1 on the east
Holocene period, lasting about 8,000 years, and soil for- and 3,900 m as1 on the west. This appears to represent the
mation has continued up to the present day, at least in places LGM,” the Last Glacial Maximum.
where no cultivation and associated land and soil degrada- Paleo-periglacial features on other mountains in Ethi-
tion has taken place. opia Surveys on different mountain tops in Ethiopia were
Processes leading to the present-day land degradation carried out between 1983 and 1986 and reported by Hurni
started about 2,000–3,000 years ago with widespread (1986). None of the mountains showed paleoglaciation;
deforestation, cultivation of the Andosol soils, and human- however, paleo-periglacial slope deposits, post-glacial ero-
induced soil erosion on slopes without soil protection. Soil sion, and subsequent Holocene soil formation with Andosols
degradation is widespread and still ongoing, to the extent were observed on all mountains, followed by human-induced
that human livelihoods are threatened and subsistence agri- land and soil degradation. Mountains visited included Mount
cultural practices are no longer possible for many farming Choke in Gojam (4,052 m asl), Mount Guna in South Gonder
families. Insufficient crop yields and the need to supply food (4,135 m asl), and Mount Molle (or Amba Farit) in Western
aid for a considerable portion of the year have become a Wello (4,247 m asl). Mount Abune Yosef (4,284 m asl) in
146 H. Hurni
Northern Wello near the town of Lalibela has not been visited Development and Environment (CDE). 30 pp. with map, scale 1:5
so far, but since the mountain top does not extend much into million
Hurni H, Ludi E (2000) Reconciling conservation with sustainable
the paleosnowline of the Late Pleistocene (at 4,250 m asl in development. A participatory study inside and around the Simen
Simen), paleoglaciation there is not probable either. Mountains National Park, Ethiopia. Produced with the assistance of
an interdisciplinary group of contributors. Bern: Centre for Devel-
opment and Environment. ISBN 3-906151-44-1, 476 pp
Hurni H, Debele B, Zeleke G, Ludi E, Abate S, Bantider A (2008) Land
References degradation and sustainable land management in the highlands of
Ethiopia. In: Research for development. A synthesis of NCCR
North–South research, 2001–2008. Perspectives of the Swiss
Mark BG, Osmaston HA (2008) Quaternary glaciation in Africa: key National Centre of Competence in Research (NCCR) North–South,
chronologies and climatic implications. J Quat Sci 23:589–608. University of Bern, vol 4. Geographica Bernensia, Bern
ISSN 0267-8179 Kazmin V (1973) Geological map of Ethiopia, 1:2 million. Geological
Buedel J (1954) Klima-morphologische Arbeiten in Aethiopien im survey of Ethiopia, Ministry of Mines and Energy, Addis Abeba
Fruehjahr 1953. Erdkunde VIII 2:139–156 Kirwan LP (1972) The Christian topography and the kingdom of
Frei E (1978) Andepts in some high mountains of East Africa. Axum. Geogr J 138:166–177
Geoderma 21:119–131 Merla G, Abbate E, Azzaro1i A, Bruni P, Canuti P, Fazzuoli M, Sagri
Gasse F (1975) L’evo1ution des lacs de l’Afar central (Ethiopie et T.F. M, Tacconi P (1979) A geological map of Ethiopia and Somalia
A.I.) du Plio-Pleistocene à l’Actuel, vol 3. These Université Paris (1973), scale 1:2 million. Department of Geology, University of
VI, 568 pp Florence, 95 pp
Gasse F, Rognon P, Street FA (1980) Quaternary history of the Afar Messerli B, Hurni H, Kienholz H, Winiger M (1977) Bale mountains:
and Ethiopian rift lakes. The Sahara and the Nile. Balkema, Largest Pleistocene mountain glacier system of Ethiopia. INQUA
Rotterdam, pp 361–499 Abstracts, Birmingham
Hastenrath S (1974) G1aziale und Periglaziale Formbildungen in Hoch- Minucci E (1938) Ricerche geo1ogiche nella regione del Semien. In:
Semyen, Nord-Aethiopien. Erdkunde 28:176–186 Ace R (ed) vol. I, Roma: 37–46
Hastenrath S (1977) Pleistocene mountain glaciation in Ethiopia. Mohr PA (1962) General report on an expedition to the Semien
J Glaciol 18(79):309–313 mountains. Bull Geo Obs Addis Abeba 6:155–167
Hoevermann J (1954) Ueber glaziale und “periglazia1e” Erscheinungen Umer M, Kebede S, Osmaston H (2004) Quaternary glacial activity on
in Erithrea und Nordabessinien. Veroeffentlichungen der Akadernie the Ethiopian mountains. Dev Quat Sci Part C 2:171–174
fuer Raumforschung und Landesp1anung. Abhandlung Band Nilsson E (1949) Ancient changes of climate in British East Africa and
28:87–111 Abyssinia. Geogr. Annaler 22:1–79
Hurni H (1981) Simen mountains—Ethiopia: palaeoclimate of the last Osmaston HA, Mitchell WA, Osmaston JAN (2005) Quaternary
cold period (late wurm). Paleoecology Afr 13:127–137 glaciation of the Bale Mountains, Ethiopia. J Quat Sci 20:593–606
Hurni H (1982) Simen Mountains—Ethiopia, vo1 II (in German): Potter EC (1976) Pleistocene glaciation in Ethiopia: New evidence.
climate and the dynamics of altitudinal belts from the last cold reprint: series 299, Center for Astrophysics, Cambridge, MA, 7 pp
period to the present day (Part II in co-authorship with Peter Werdecker J (1955) Beobachtungen in den Hochlaendern Aethiopiens
Staehli). Geogr Bernensia G 13, 196 pp auf einer Forschungsreise 1953/54. Erdkunde IX: 395–317
Hurni H (1986) Late quaternary in Simen and other mountains in Williams MAJ, Street FA, Dakin FM (1978) Fossil periglacial deposits
Ethiopia. Mahaney WC, Balkema AA (eds) Quaternary and in the Semien highlands, Ethiopia. Erdkunde 32:46–49
environmental research on East African mountains, Rotterdam, Zech R, Manhart A, Glaser B, Solomon, D, Zech W (2005) A high-
pp 105–120 resolution lateglacial climate record from Lake Sediments in the
Hurni H (1998) Agroecological belts of Ethiopia. Explanatory notes on Ethiopian Bale Mountains. American Geophysical Union, Fall
three maps at a scale of 1:1,000,000. Soil conservation research Meeting 2005, abstract #PP21A-1554
programme, research report 43. Addis Abeba and Bern: Centre for
Geomorphology of the Archaeological Area
of Aksum 7
Giovanni Ferrari, Rossano Ciampalini, Paolo Billi, and Piotr Migon
Abstract
Aksum is one of the most important archaeological and historical towns in Ethiopia. The
archaeological area stretches on a plateau ranging from 2,250 to 2,460 m, and the general
landscape consists of a few dome-shaped hills standing on a plateau crossed by small ephemeral
streams. The main geomorphic features are the result of the emplacement of basalt flows; the
intrusion of syenite sub-volcanic plugs; and domes, tectonics, weathering and its interaction with
various surface erosion processes. In such a context, between 700 BC and 800 AD, the rise and the
decay of the Aksumite Kingdom, one of the most known civilisations of East Africa, took place.
The core of the archaeological area of Aksum, with tall carved obelisks, is a UNESCO World
Heritage listed property. The integration of archaeological data with the analysis of soil
conservation measures and agricultural practices, adopted till present, allowed to develop a plough
marks-based methodology to assess the historical soil erosion rate and, hence, to understand
historical landscape evolution within a man-controlled environment. The occurrence of
archaeological evidence in such a characteristic geomorphological context provides an excellent
opportunity to study and understand interaction between geomorphic and anthropic processes.

Keywords
Structural landforms Geoarchaeology Soil erosion Aksum Ethiopia
7.1 Introduction Aksum (or Axum) in the north of Tigray Province used to be
the administrative centre of a flourishing state for more than
Northern Ethiopia is home to one of the most fascinating a millennium, since c. 400 BC to around 800 AD when it
ancient civilisations in the sub-Saharan Africa—the Aksu- began to decline and eventually collapsed. However, a
mite Kingdom. The now rather inconspicuous town of number of monuments and archaeological sites survived
until today, including the famous stelae carved out of local
rocks and erected to commemorate Aksumite rulers. A
G. Ferrari measure of the worldwide cultural significance of the area is
Di.P.S.A, Università di Firenze,
Piazzale Cascine 18, 50144 Florence, Italy
its inscription on the World Heritage List in 1980. This
diverse material legacy of the Aksumite civilisation, along
R. Ciampalini (&)
INRA-US InfoSol, Centre de recherche Val de Loire, 2163
with an opportunity to assess long-term interactions between
Avenue de la Pomme de Pin, 45075 Orléans, France soils, land use and soil erosion, has made Aksum a place
e-mail: rossano.ciampalini@gmail.com where geoarchaeological approach to the environment can
P. Billi be developed with a particular success. However, the
Physics and Earth Sciences, University of Ferrara, broader geomorphological context of the former Aksumite
Via Saragat 1, 44122 Ferrara, Italy capital is no less interesting. From a relatively flat surface, a
P. Migon number of flat-top domes rise (Natali et al. 2013), whose
Department of Geography and Regional, associations provide insights into structural geomorphology
Development University of Wroclaw, Wroclaw, Poland
148 G. Ferrari et al.
of syenite intrusive complexes and weathering and denuda- few scattered hills with steep to almost vertical slopes
tion patterns in basement rocks, including characteristic standing on a plateau crossed by small ephemeral streams
residual landforms. Both these topics and their mutual (Figs. 7.1 and 7.2).
interactions will be presented in this chapter. Annual rainfall ranges from 600 to 800 mm and is
characterised by high interannual variability. Precipitation is
concentrated in the main, monsoon-type rainy season, from
7.2 Geographical Setting of Aksum mid-June to mid-September (about 60–70 % of the total)
and, at a much lesser extent, in the unpredictable “small”
The town of Aksum is located in the north of Ethiopia, at rainy season from mid-March to mid-May, originated by the
roughly 14°N and 39°E, and is one of the most important development of cyclonic cells. Mean annual temperature is
archaeological and historical cities of Ethiopia. The sur- around 18 °C, with low seasonal variability, but large diurnal
rounding area stretches on a plateau ranging from 2,250 to range. The vegetation is typically a highland sparse wood-
2,460 m a.s.l., and the general physiography consists of a land, dominated by Eucalyptus brevifolia and Euphorbia
Fig. 7.1 Topography of the archaeological area of Axum, including two soil erosion assessment study sites (A and B)
7 Geomorphology of the Archaeological Area of Aksum 149
Fig. 7.2 Aerial photograph of

the area showing the flat-top of
Bet Giyorgis hill
candelabrys, though much of the area is covered only by basalts seem to overlie felsic rocks, and in some places, e.g.
sparse shrubs and grasses and many slopes are rocky or west of Aksum, basaltic dikes clearly cross-cut a felsic
almost bare. The town of Aksum itself has population of c. intrusion. Even though the absolute age of this complex is
45,000, and its downtown area is located right at the foot of unknown, compositional similarities suggest a link with
two prominent hills of Bet Giyorgis in the west and May basalt–trachyte series from the shield volcanoes overlying
Koho in the east (Fig. 7.1). the Northern Ethiopian continental flood basalts, which have
been dated as Miocene in age (Kieffer et al. 2004). Analo-
gous magmatic associations have been observed in the
7.3 Geology neighbouring Eritrean site of Senafe, where radiometric
dating indicates ages from 21 to 23 Ma (Zanettin et al.
The geology of the area is complex and includes Proterozoic 2006).
basement of crystalline rocks (mainly granites and schist), The archaeological area around Aksum is dominated by
sedimentary rocks of Mesozoic age (mainly sandstones), the nepheline syenites domes of Bet Giyorgis and Gobo Dura,
overlain by the Ethiopian Plateau flood basalts, emplaced dated for between 19 and 15 Ma (Natali et al. 2013), and the
between 31 and 29 Ma) (Hofmann et al. 1997), the later dome-shaped trachytic plug hills of May Koho, likely em-
phase basaltic rocks (17–15 Ma) and the post-trap, Oligo- placed concomitantly (Bianchini, personal communication).
cene to Miocene trachytic lavas and syenite sub-volcanic These domes stand out from a relatively flat area underlain by
intrusions in the form of plugs and domes (Hagos et al. trap basalts. On the western and eastern sides of Bet Giyorgis,
2010; Natali et al. 2013). These silica-poor volcanic to small outcrops of lacustrine deposits occur. They are tilted and
hypabyssal rocks belong to the Aksum–Adwa phonolite– in places hydrothermally affected by the emplacement of the
trachyte volcanic field that is a part of a larger regional syenite plug. In the study area, these sedimentary rocks are
magmatic district that extends for ca. 150 km from west of commonly associated with the Enticho Sandstones of Car-
the town of Aksum to the north–south belt of Adwa and boniferous to Permian age (Beyth 1972; Bussert and Schrank
covers an area of about 800 km2 (Zanettin et al. 2006; Hagos 2007; Bussert 2010; Sacchi et al. 2007) and, though it was not
et al. 2010; Natali et al. 2013). Magmatic rocks and the possible to establish a clear stratigraphy, they can be con-
eruptive centres are placed along E–W and NE–SW vol- sidered as a laterally interfingering glacigenic facies unit of
cano-tectonic trends, which conform to NE–SW striking the Edaga Arbi Formation (Bussert and Schrank 2007). Both
shear zones affecting the Pan-African basement of the area formations are unconformably overlain by the Adigrat
(Natali et al. 2013). The chronological relationships between Sandstones of Mesozoic age, mostly of continental origin, as
basalts and felsic rocks are not everywhere clear. Locally, shown by the occurrence of palaeosols, fluvial point bar
Fig. 7.3 Geomorphological sketch map of the study area
sequences and fossil wood (Beyth 1972; Merla et al. 1979; the hills and connected with the lower plain; (iii) fluvial
Bosellini et al. 1997). Blocks of sandstone and fine con- valleys incised into the etchplain; and (iv) lower clayey
glomerate, probably belonging to the Adigrat Sandstone unit, alluvial plain.
are also found on the top of Bet Giyorgis and on the northern
side of Gobo Dura.
7.4.1 Main Domes and Related Forms
7.4 Landforms and Landscapes Most of the domes protrude abruptly from the surrounding flat
area and rise above it by 200–250 m, emerging as the most
The geomorphological landscape of the archaeological area evident geomorphological forms of the region. However, in
of Aksum is characterised by a flat surface at an elevation contrast to many igneous rock domes known from elsewhere
around 2,200 m a.s.l., above which a few individual, dome- in Africa and beyond (see Thomas 1994), the domes around
shaped sub-volcanic plugs stand (Fig. 7.3). The long-term Aksum are distinguished by their remarkably flat-tops sur-
weathering and denudation enriched the morphology of rounded by steep slopes, not uncommonly cliff lines, which
these plugs, leading to the origin of a number of landforms grade into boulder-covered hillslopes further below.
associated with the typical geomorphological evolution of The flat-tops may attain considerable dimensions
crystalline rocks such as rock cliffs, castellated tors, block (Fig. 7.2). At Bet Giyorgis, the mountain-top surface is nearly
fields, boulder slopes, and mid-slope benches. In parallel, the circular, with a diameter of 2 km, enough to accommodate a
intervening flat areas between the domes have been shaped small village and extensive agricultural area. An efficient
by a variety of erosional and depositional processes. Thus, drainage and soil types with less developed vertic properties
the typical landform association in the broader Aksum area than on a floodplain have provided an ideal support for
consists of (i) domes with distinctive subordinate elements cropping since the first stages of the Aksumite Kingdom.
such as flat-tops, steep or stepped hillslopes, and boulder Outcrops of massive syenite in the form of cliffs, castel-
talus; (ii) an etchplain developed in the felsic rocks around lated spurs (Fig. 7.4) and isolated tors (Fig. 7.5) are present on
Fig. 7.4 Rock cliffs and

castellated towers above a
boulder-mantled hillslope of the
syenite plug of Gobo Dura
Fig. 7.5 Structural bench with

emerging syenite tors
the flanks of the hills. They are best developed on their north- soil-covered surfaces are dominant. These steep hillslopes are
western sides, where clusters of towers rise abruptly from the affected by three main geomorphic processes: (1) deep bed-
surrounding etchplain and connect two main relief elements, rock weathering leading to the formation of grus-type regolith
namely the etchplain and the dome tops. Otherwise, hillslopes with corestones; (2) gravity-driven downhill movement of
are less steep (6–45 %) and consist of a mosaic of bare rock boulders (corestones); and (3) creeping and soil displacement
outcrops, exposed regolith with rinded corestones (Fig. 7.6), due to gravity and water erosion.
boulder blankets and patches of soil cover. While bedrock The main feature of the lower sections of the syenite
outcrops are more frequent in steeper section, regolith- and plugs is benches, giving the slopes a step-like appearance.
Fig. 7.6 Grus-type regolith in

syenite with corestones
surrounded by weathering rinds
Fig. 7.7 A step at the foot hill of

Bet Giyorgis
Fig. 7.8 Fluvial valley incised

into basaltic substrate the north-
west of Bet Giyorgis
They are large flat areas, located at the foot of the escarp- surfaces seem to play a major role in maintaining charac-
ments, with the inner part of the treads deeply weathered and teristic footslope morphology (Wahrhaftig 1965) (Fig. 7.7).
their edges made of exposed rocks, commonly crowned with
tors. The origin of benches is not entirely clear. It is likely
that they are structure-controlled features, formed due to 7.4.2 Etchplain
selective weathering of marginal parts of the plugs. More
susceptible compartments decompose at a faster rate and In tropical and subtropical landscapes, etchplanation is the
turn into regolith, but barriers made of more resistant, and name given to a specific pathway of geomorphic evolution in
hence, less weathered parts of the intrusion prevent regolith which deep chemical weathering produces a thick saprolite
evacuation and sustain the bench. Once exposed, differential that is progressively removed by erosion (Thomas 1989,
rates of weathering between exposed and buried rock 1994). Circulating groundwaters lower the weathering front
Fig. 7.9 Floating syenite

boulders in the lower plain, with
flat-topped plugs in the
background
at various rates according to the different density of joints in metric and sub-metric boulders resulting from weathering
the substrate and lithology-controlled susceptibility to and mass movement processes active on the hill slopes.
weathering. Such etching process led to progressive differ-
entiation of relief features, including the formation of
inselbergs, steps, scarps and flat valleys. 7.4.4 Lower Plains
In the case of Bet Giyorgis hill, the structure of the plug,
with the outer parts apparently more densely jointed, favours The flat areas around the main domes have been formed by
such differentiation. In fact, the high rate of weathering in the different processes: etchplanation, flood basalt emplacement
foothill settings (sapping—Thomas 1974) and the adjacent and sedimentation of ephemeral river floods (Fig. 7.9).
areas results in the lowering and planation of the lower sur- Long-term deposition and pedogenesis produced a deep soil
faces, whereas the main elements of hilly relief are preserved. cover where vertisols are the dominant type. The lower
Along and around the plug hill base, syenite rocks undergo plains have been widely used for cropping and cereal pro-
intense chemical weathering and the soluble products are duction since the antiquity and, because of the vertic prop-
removed by the percolating waters. erties of the soil, are considered of inferior quality by the
local farmers. Notwithstanding the low gradient (0–3 %),
along almost the whole lower plain soil conservation prac-
7.4.3 Incised Valleys tices such as continuous terracing with stone bounds as high
as to 0.5 m are common. This setting allows to control the
Along the north-western margin of the Bet Giyorgis hill, the erosive energy of the monsoon intense rainfalls and to pre-
deeply weathered etchplain surface is considerably incised vent the cultivated fields from rill development.
by a few ephemeral streams. Most of the valleys in the area The lower plains, especially those south-west of the Bet
seem to have formed before the emplacement of basalts Giyorgis dome, are punctuated by scattered boulders of
associated with the second phase of the Miocene flood basalt metric size (Fig. 7.9). Field evidence indicates that their
emission (Natali et al. 2013). In fact, basalt flows appear to presence on top of the lower plain surface at a far distance
have filled most of the palaeovalleys between the plug hills from the plug hillslopes can be accounted for by a sequence
and were incised again in more recent times to achieve the of processes, involving detachment by weathering and
modern valley and river channel morphology (Fig. 7.8). This gravity displacements in which wetting and drying of the
phase of erosion resulted in deeply incised box-like streams soil may have played a significant role (Cook 1993 in
with sub-vertical bedrock banks 3–6 m high and 5–20 m Poesen and Lavee 1994), allowing for sliding of boulders on
wide (Fig. 7.8). The beds of these streams may consist of saturated muddy substrate.
Fig. 7.10 Frequency distribution

and main direction, respectively,
of the intermediate b axis of the
boulders in a site west of the Bet
Giyorgis: a vertisol (n = 104)
(valley); b saprolite substrate
(n = 74) (mid-slope); c vertisol
(n = 104) (valley); and d saprolite
substrate (n = 74) (mid-slope)
In Fig. 7.10, the frequency distribution of the intermediate, present. On the plug foothill slopes, truncated palaeosoils
b axis of the boulders and their orientation are reported, and, in places, deep saprolites buried by colluvial pro-
respectively. Two groups of boulders were considered, i.e. cesses and truncated luvisols (“makaeo” in the local
those floating on the top of saprolite at the plug hill footslope, language) are found. The low-gradient slopes around the
ideally representing the baseline of the boulder movement, hills are strongly weathered, and relict soils, probably
and those resting on the vertisols in the distal slope area. The nitisols (“keyah”), with argillic horizon and peds with
results show the mean value of the b axis of 111.6 and 94.7 cm shiny faces, are preserved under boulder cover. Soils with
for boulders located on the vertisol and the saprolite, vertic features (“bakahel”) occupy the central parts of the
respectively. The mean b axis orientations for the boulders on benches within plug hillslopes, while vertisols (“walka”)
these two different substrates are 230.6° and 263.3°, respec- (Nyssen et al. 2008; Schmid et al. 2008; French et al.
tively, i.e. almost consistent with the main slope direction 2009) are typically found on basalts, siltstones and allu-
(220°). The consistency is more evident for boulders on top of vial plains.
the vertisols and indicates that they slid down “floating” on
top of the ground surface and keeping their intermediate axis
almost parallel to the main slope dip. This pattern is observed
when particles are moved solely by gravity (Dumas and Raffy
1993), and not perpendicular to the bed gradient as it is by Table 7.1 Chronology of the Aksumite Kingdom (Fattovich 1997a;
Phillipson 2000)
contrast observed when a particle is moved by the drag of a
fluid (particle imbrication) (Briggs 1977). Historical period Year
Pre-axumite 700–400 BC
Proto-axumite 400–40 BC
7.5 Soils and Soil Erosion Assessment Early axumite 40 BC–150 AD
Classic axumite 150–350 AD
7.5.1 Soils Middle axumite 350–550 AD
Late axumite 550–800 AD
On the flat-top of Bet Giyorgis, an association of trun-
cated luvisols and shallow soils with vertic properties is Post-axumite 800–1100 AD
Fig. 7.11 Typical hillslope

terrace made for cultivation
purposes
7.5.2 Soil Use and Agriculture Fattovich et al. 2000). According to a few authors (e.g. Bard
et al. 2000; Sernicola 2008), archaeological evidence sug-
The land use of the area is primarily agriculture, and the gests that the northern slope of Bet Giyorgis has been
agricultural landscape consists mainly of level bench terraces uninterruptedly cultivated since the Early Aksumite time
with stone bunds decreasing in height according to the slope (40 BC–150 AD) (Table 7.1). The urban and rural popula-
gradient. Stony bunds to control and reduce soil erosion were tions reached their maximum during the Classic Aksumite
particularly implemented during the DERG (Coordinating period (150–350 AD), implying an expansion of land used
Committee of the Armed Forces, Police, and Territorial Army; for agricultural purposes (Sernicola 2008; Sulas et al. 2009;
1974–1987) period within the food for work programme. Sernicola and Phillipson 2011). Since at several sites, arte-
They are present on almost any slope, with the exception of the facts dating back to the different periods are found in place,
steepest ones. Structural steps, such as those on the northern it is reasonable to postulate that these locations have been
side of Bet Giyorgis, are used as natural agricultural terraces continuously cropped at least since the Early Aksumite
bounded by natural rock outcrops (kopjes) or man-made stone period.
mounds. Large grazing areas, without any conservation In Ethiopia, relevant agricultural techniques and imple-
practices, are found on gently sloping areas covered by soils ments seem to have been mainly imported from South
with vertic properties, whereas large arable fields on deep Arabia, probably before the Aksumite period (2,600–
vertisols are in the topographically lower parts of the study 2,200 BP; Butzer 1981), though stone terracing is a very old
area. All crops are rain-fed and include teff (Eragrostis tef), approach to land management since it started probably ear-
barley (Hordeum vulgare), wheat (Triticum aethiopicum) and lier than 2,500 years ago (Michels 2005). In the Aksum area,
millet (Eleusine coracana) (D’Andrea et al. 2011). Some the oldest terraced surfaces are estimated to date back to pre-
forest exclosures are present near residential and religious Aksumite period (700–400 BC) (Fattovich et al. 2000). This
areas or in marginal rocky grounds. On the steeper slopes, a conservation technique regards the slope as a whole, like in
sparse eucalyptus forest alternates with bedrock exposed by the Middle East, and considers the upper section not as an
erosion processes and fluvial dissection. enlargement of the lower cultivated area under human
The hilly land around the city of Aksum has been pressure but as an integrated part of the system, serving
extensively utilised for agriculture purposes since the different purposes (Hallsworth 1987). Thus, the slope is
antiquity and early traces of agricultural land use date back managed in the following manner: upslope terraces are
to the first millennium BC (Butzer 1981; Fattovich 1997a, b; mainly used for grazing and water harvesting, while at
Fig. 7.12 The “Maresha” ard

plough
Fig. 7.13 Plough marks features

on a boulder and various
intensities of patinas indicating
the history of exposure (after
Ciampalini et al. 2012, modified)
downslope, they are cropped (Vogel 1988) (Fig. 7.11). The are based on the intensive use of the maresha, a traditional
terraced surfaces in the higher portion of slopes do not show ard plough (Gebregziabher et al. 2006; Nyssen et al. 2000,
any sign of ploughing, confirming that this area was used 2011). The maresha seems to have been introduced in
specifically for water harvesting and grazing rather than for Ethiopia between 1000 and 400 BC by Semitic tribes from
cropping. In this context, agricultural practices were and still Yemen, or even before by Cushitic-speaking peoples from
Table 7.2 Median values of the estimated erosion (t ha−1 year−1) for ancient (Site A and B) and recent soil erosion (Site B) at two study sites
Sites Proto-axumite (since Early axumite (since Classic axumite (since Middle axumite (since Recent erosion (last
2400 year BP) 2040 year BP) 1850 year BP) 1650 year BP) 35 years)
A 2.6 3.1 3.4 – –
B – 2.5 2.8 3.1 65.8
Fig. 7.14 Geoarchaeological

legacy in the Axum area. a The
broken Great Stele in the
archaeological park in the town;
b ancient quarry at Gobo Dura,
with an unfinished obelisk; c rock
carving of Lioness at Gobo Dura
north-eastern Sudan. The plough is a pointed, steel-tipped and the site B, at which cultivation started from Early
tine fixed to a pole (Fig. 7.12). The soil, therefore, is not Aksumite to Middle Aksumite (2040–1650 year BP). The
turned over, but incised and pushed aside by two lateral latter site is particularly interesting as a major land use
wings; a handle allows the implement to be driven and lifted change occurred in 1975 when cultivation was abandoned
when an obstacle is encountered. and terrace maintenance was halted. This change produced a
significant degradation of the terraces by accelerated erosion.
The long-term soil loss calculated at these two sites ranges
7.5.3 Soil Erosion Assessment from 2.6 to 3.4 t ha−1 year−1 and for the site B, an erosion rate
ranging from 2.5 to 3.1 t ha−1 year−1 (Table 7.2). These
In the study area, soils are characterised by a large content of values can be contrasted with the 20-fold increase in erosion
rock fragments. Nepheline syenite boulders and core stones rate at the site B after the recent abandonment. The median
with weathering rinds lie on thin, buried red clayey palaeosols value of soil loss over a period of 31 years (1975–2006) is
or protrude from a saprolite mantle. These boulders are scat- 65.8 t ha−1 year−1 (Ciampalini et al. 2008, 2012).
tered across gentle slopes at the base of the domes and used to
build the bounds of the cropped terraces. During the last two
millennia, the terraces have been ploughed with the use of an 7.6 Stone Industry and Aksum Stelae
ard plough pulled by one or two oxen. Mechanical impact of
the plough on the boulders is therefore unavoidable, and they The presentation of the geomorphology of the Aksum area
are scratched on the top, if embedded in soil at a shallower would be incomplete if the geomorphic impact of quarrying
depth than tillage depth (8–16 cm), or on the sides, if they are and stone dressing is ignored. The archaeological area of
partially exposed. As soon as the farmer hears and feels the Aksum is famous for the stone-carved monuments—the
scraping, he lifts the implement, but a deep mark on the rock stelae (obelisks). The most imposing ones are displayed in
surface is already made. Plough marks are found at different the Northern Stelae Field, at the outskirts of the town and at
heights on the rocks faces, thus providing valuable informa- the foot of Bet Giyorgis hill. The highest standing intact
tion on the thickness of soil loss (Fig. 7.13). In association obelisk, dedicated to the King Ezana (reigned 330–356 AD),
with other rock weathering features such as patinas, etches, is 24 m high, but the most impressive of all is the lying and
pitting and lichen cover, plough marks can be used to broken, 33-m-long Great Stele (Fig. 7.14a). In 2008, another
reconstruct the position of old soil surfaces, and by identifying obelisk, the Rome Stele, broken into pieces in antiquity and
their occurrence, it is possible to recognise ancient cultivated taken to Italy in 1930s, was reinstated on the site. All are
lands that are presently abandoned, degraded or strongly made of local syenite and elaborately decorated. Many more
dissected (Ciampalini et al. 2008). obelisks can be found at this and other archaeological sites in
The spatial relationship between plough marks and rock the vicinity. Further, stone monuments of geoarchaeological
surface orientation gives information on the ploughing interest are numerous megalithic tombs, including a highly
direction, whereas the relative height from the top of the complex underground Mausoleum erected from huge syenite
marks belt to the present ground surface, plus the ploughing slabs in a purposefully made trench.
depth, i.e. additional 15–18 cm, is assumed as the actual soil Stone for obelisks was sought and quarried on the slopes
loss. By this method, developed by Ciampalini et al. (2008), of the Gobo Dura hill, about 3 km west of the town.
it is possible to assess soil erosion rate throughout long-time Apparently, they were the best source of rock solid enough
intervals, provided reliable dating of the cultivated surface to withstand cutting, transport and final dressing, pointing to
(e.g. by archaeological evidence, C14 and thermolumines- an excellent indigenous knowledge of syenite weathering
cence) is available. For instance, in a study plot of the Ma’ patterns, rock properties and rock behaviour. Until today,
Qono area, at the north-west slope of Bet Giyorgis, remnants of ancient quarries can be seen at Gobo Dura,
archaeological evidence in the form of old ruined building, including unfinished stelae and stones abandoned during
artefacts, pottery, iron slags and pollen analysis indicates this transport (Fig. 7.14b). The slopes of Gobo Dura feature
site has been occupied and cultivated since the Proto- further examples of interactions between rocks, landforms
Aksumite times (about 2,400 BP) (Fattovich et al. 2000). and people. One of these is the 2-m-long image of a lioness
Similarly, two sites represented by two series of agricul- carved on the site of huge monolithic syenite boulder, linked
tural terraces with plough marks features still well preserved with local legends. Chiselling the animal outline required
have been dated and investigated (Fig. 7.1). These are the site removal of the thin outer layer of rock, resulting in the
A, where the agricultural exploitation has been estimated to currently observed colour difference between the red iron-
start from Proto- to Classic Aksumite (2400–1850 year BP)
rich crust developed over long time and the carving itself Bussert R, Schrank E (2007) Palynological evidences for a latest
which lacks the crust (Fig. 7.14c). Carboniferous-Early Permian glaciation in Northern Ethiopia. J Afr
Earth Sci 49:201–210
Finally, we have to consider the amount of stone blocks Butzer KW (1981) Rise and fall of Axum, Ethiopia: a geo-archaeo-
that have been carved from the crystalline rocks of the plugs logical interpretation. Am Antiq 46:472–495
throughout the past centuries to present. These stones have Ciampalini R, Billi P, Ferrari G, Borselli L (2008) Plough marks as a
been used as construction material to build the majority of tool to assess soil erosion rates: a case study in Axum (Ethiopia).
Catena 75:18–27
the ancient buildings. Ciampalini R, Billi P, Ferrari G, Borselli L, Follain S (2012) Soil
erosion induced by land use changes as determined by plough
marks and field evidence in the Aksum area (Ethiopia). Agric
7.7 Conclusions Ecosyst Environ 146:197–208
D’Andrea AC, Richards MP, Pavlish LA, Wood S, Manzo A, Wolde-
Kiros HS (2011) Stable isotopic analysis of human and animal diets
The close vicinity of Aksum belongs to the most attractive from two pre-Aksumite/proto-Aksumite archaeological sites in
geomorphological landscapes of Ethiopia and offers a sin- northern Ethiopia. J Archaeol Sci 38(2):367–374
gular combination of natural and anthropogenic features. Dumas B, Raffy J (1993) The study of stone trails with painted tracers
on unstable slopes in the Southern French Alps: preliminary results.
The geomorphological evolution of the area is controlled by Geog Fis Din Quat 16(1):37–45
a complex combination of local tectonics, lithological Fattovich R (1997a) The peopling of the Tigray Plateau in ancient and
diversity, climate and surface processes. The main geomor- medieval times (ca. 4000 B.C.–A.D. 1500): evidences and synthe-
phic features of the area represent a classical, well-expressed sis. In: Bard KA (ed) Environmental history of human ecology on
Northern Ethiopia in the late holocene. Istituto Universitario
example of volcanic and sub-volcanic intrusions in the form Orientale, Napoli, pp 81–105
of lava flows, plugs and domes. Contemporary surface Fattovich R (1997b) Archaeology and historical dynamics: the case of
processes include deep weathering of syenite rocks, disin- Bieta Giyorgis (Aksum), Ethiopia. Annali dell’Istituto Universitario
tegration of rock cliffs along the rims of the flat-top domes, Orientale 57:46–79
Fattovich R, Bard KA, Detrassi L, Pisano V (2000) The Axum
continuing development of boulder talus, regolith and soil archaeological area: a preliminary assessment. Ist. Univ. Orientale,
erosion, as well as fluvial incision and valley widening. The Dip. Di Studi e Ricerche su Africa e Paesi Arabi. Working paper 1
development of the Aksumite Kingdom provides an addi- French C, Sulas F, Madella M (2009) New geoarchaeological
tional value to such particular geomorphological context and investigations of the valley systems in the Aksum area of Northern
Ethiopia. Catena 78:218–233
favourable conditions to investigate about landscape evolu- Gebregziabher S, Mouazen AM, Van Brussel H, Ramon H, Nyssen J,
tion and geomorphic processes during the last two millennia. Verplancke H, Behailu M, Deckersf J, De Baerdemaeker J (2006)
Historical erosion rates were calculated since the beginning Animal drawn tillage: the Ethiopian and plough, maresha: a review.
of the Aksumite Kingdom owing to the concurrence of a few Soil Tillage Res 89:129–143
Hagos M, Koeberl C, Kabeto K, Koller F (2010) Geochemical
main factors: the plough marks, the presence of artefacts and characteristics of the alkaline basalts and the phonolite—trachyte
other archaeological evidence, and the persistence of the plugs of the Axum area, Northern Ethiopia. Austrian J Earth Sci 103
cropping practices in the same plots for about two millennia. (2):153–170
There is little doubt that while the archaeological area of Hallsworth EG (1987) Anatomy, physiology and psychology of soil
erosion. Wiley, New York
Aksum has been inscribed on the UNESCO World Heritage Hofmann C, Courtillot V, Feraud G, Rochette P, Yirgu G, Ketefo E,
List as a cultural property, its geomorphological features are Pik R (1997) Timing of the Ethiopian flood basalt event and
also of major interest and international significance. implications for plume birth and environmental change. Nature
389:838–841
Kieffer B, Arndt N, Lapierre H, Bastien F, Bosch D, Pecher A, Yirgu
G, Ayalew D, Weis D, Jerram DA, Keller F, Meugniot C (2004)
References Flood and shield basalts from Ethiopia: magmas from the African
Superswell. J Petrol 45:793–834
Merla G, Abbate E, Azzaroli A, Bruni P, Canuti P, Fazzuoli M, Sagri
Bard KA, Coltorti M, Di Blasi MC, Dramis F, Fattovich R (2000) M, Tacconi P (1979) A geological map of Ethiopia and Somalia.
Environmental history of Tigray (northern Ethiopia) in the middle Consiglio Nazionale delle Ricerche, Firenze
and late holocene: a preliminary outline. Afr Archaeol Rev 17:65–85 Michels JW (2005) Changing settlement patterns in the Aksum-Yeha
Beyth M (1972) The geology of central and western Tigray. Ph.D Region of Ethiopia: 700 BC–AD 850. Cambridge Monographs in
thesis, Rheinische Friedrich-Wilhelms Universität, Bonn, W. Ger- African Archaeology. British Archaeological Reports (BAR),
many, 200 pp Oxford
Bosellini A, Russo A, Fantozzi PL, Assefa G, Tadesse S (1997) The Natali C, Beccaluva L, Bianchini G, Siena F (2013) The Axum-Adwa
Mesozoic succession of the Mekelle Outlier (Tigrai Province, basalt–trachyte complex: a late magmatic activity at the periphery of
Ethiopia). Mem Sci Geol 49:95–116 the Afar plume. Contrib Miner Petrol 166:351–370
Briggs D (1977) Sediments. Butterwoths, London Nyssen J, Poesen J, Haile M, Moeyersons J, Deckers J (2000) Tillage
Bussert R (2010) Exhumed erosional landforms of the late palaeozoic erosion on slopes with soil conservation structures in the Ethiopian
glaciation in northern Ethiopia: indicators of ice-flow direction, highlands. Soil Tillage Res 57:115–127
palaeolandscape and regional ice dynamics. Gondwana Res
18:356–369
Nyssen J, Naudts J, De Geyndt K, Haile M, Poesen J, Moeyersons J, Sernicola L, Phillipson L (2011) Aksum’s regional trade: new evidence
Deckers J (2008) Soils and land use in the Tigray highlands from archaeological survey. Azania: Archaeol Res Afr 46(2):190–204
(Northern Ethiopia). Land Degrad Dev 19:257–274 Sulas F, Madella M, French CAI (2009) State formation and water
Nyssen J, Govaerts B, Araya T, Cornelis WM, Bauer H, Haile M, Sayre resources management in the Horn of Africa: the Aksumite
K, Deckers J (2011) The use of the marasha ard plough for Kingdom of the Northern Ethiopian highlands. World Archaeol
conservation agriculture in Northern Ethiopia. Agron Sustain Dev 41:2–15
31:287–297 Thomas MF (1974) Tropical Geomorphology. A study of weathering
Phillipson DW (2000) Archaeology at Aksum, Ethiopia, 1993-7. DW and landform development in warm climates. The MacMillan Press
Phillipson Eds, London, p. XII + 538 ltd, London and Basingstoke
Poesen J, Lavee H (1994) Rock fragments in top soil: significance and Thomas MF (1989) The role of etch processes in landform develop-
processes. Catena 23:1–28 ment. I. Etching concepts and their application. Zeitschrift für
Sacchi R, Alene M, Barbieri M, Conti A (2007) On the Paleozoic Geomorphologie NF 33:129–142
Tillite of the Adigrat Group (Tigrai, Ethiopia). Periodico di Thomas MF (1994) Geomorphology in the tropics: a study of
Mineralogia 76(2–3):241–251 weathering and denudation in low latitudes. Wiley, Chichester
Schmid T, Koch M, DiBlasi M, Hagos M (2008) Spatial and spectral Vogel H (1988) Impoundment-type bench terracing with underground
analysis of soil surface properties for an archaeological area in conduits in Jabal Haraz, Yemen Arab Republic. Transactions
Aksum, Ethiopia, applying high and medium resolution data. Institute of British Geographers, N.S. 13: 29–38
Catena 75:93–101 Wahrhaftig C (1965) Stepped topography of the South Sierra Nevada,
Sernicola L (2008) Il modello d’insediamento sull’altopiano tigrino California. Geo Soc Am Bull 76:1165–1190
(Etiopia settentrionale/Eritrea centrale) in epoca Pre-Aksumita e Zanettin B, Bellieni G, Justin Visentin E (2006) Stratigraphy and
Aksumita (ca. 700 a.C.–800 d.C.). Un contributo da Aksum. evolution of the trachy-rhyolitic volcanism of the Senafe area (East
Unpublished PhD dissertation, Department of African and Arabian Eritrean Plateau). J Afr Earth Sci 45:478–488
Studies, University of Naples, Italy
Geomorphology of the Adwa District
8
Maria J. Machado
Abstract
This chapter describes the geologic and geomorphic characteristics of the Adwa district and
provides a reconstruction of the major phases of the long-term landscape evolution from the
Late Proterozoic to the Quaternary. Two landscape features of the Adwa district exemplify the
control of tectonic uplift and erosion, often in a feedback loop: the plateau landscape and river
incision. The plateau landscape developed upon metamorphic, sedimentary and volcanic
materials, with major contacts marked by planation surfaces correlative with major regional-
and sometimes continental-scale unconformities. In the southern Adwa district, the most
prominent landscape feature is a regionally extensive erosion surface bevelled across the
Precambrian basement and later buried by Palaeozoic and Mesozoic sedimentary rocks. In the
northern district, the flat-top horizontal structural plateaux are formed by the Mid-Cenozoic
upper basaltic Trap sequence, intruded by trachyte and phonolite plugs originated during the
Miocene and Pliocene volcanic activity, and a lower plateau developed over the laterised
Adigrat Sandstone Formation. The post-Pliocene landscape evolution has led to exhumation of
the former erosion surfaces and dissection of the drainage network. These stream channels
have cut large amphitheatre headwaters and flow southwards to join the regional base level of
the Weri River, a contributor to the Nile basin. The Quaternary sedimentary record in the
Adwa area is composed by consolidated carbonate Late Pleistocene rocks (travertines),
unconsolidated Pleistocene/Holocene alluvial fan deposits, Holocene fluvial deposits (river
terraces and valley fill deposits) and historical to present-day colluvial deposits.

Keywords
Northern Ethiopia Plateau landscape Erosion surfaces Valley fill deposits
8.1 Introduction shows a complex assemblage of dynamic landscapes, each

preserving a unique history of landscape evolution of the
The Adwa woreda (Amharic word for a district level divi- area.
sion) is located in the north-west part of the extensive and The general landscape traits of the Adwa area reflect the
elevated Ethiopian Plateau, largely above 1,500 m a.s.l., and relentless and unbalanced competition through time between
is built of metamorphic, sedimentary and volcanic materials. rock basement uplift, climate and the resistance of exposed
Geographically, it corresponds to the central zone (Maákelay rocks against weathering and erosion. The geomorphology,
zone) of the northern province of Tigray. The Adwa region as in the entire Ethiopian Highlands, has an evident litho-
logical and structural control showing a heterogeneous
mosaic preserving land surfaces and landforms of relative
M.J. Machado (&) antiquity. This structural landscape is characterised in the
Departamento de Geología, Museo Nacional de Ciencias north by a high surface with a low-gradient landscape at a
Naturales, CSIC, Calle Serrano, 115bis, 28006 Madrid, Spain prominent elevation of 2,200 m, which is cut by large
e-mail: machado@mncn.csic.es
164 M.J. Machado
amphitheatre headwaters that results from an evident scarp unconformably (Beyth 1972) by the metasedimentary rocks
retreat. Southwards, the incised valleys running north–south of the Tambien Group, that in the case of the Adwa area is
cut folded and faulted metamorphic rocks (volcanic and formed by four of the six formations identified in the
sedimentary) of the Neoproterozoic basement. The stream northern highlands (Tadesse et al. 1999; Miller et al. 2009;
channels trend southwards to join the regional base level of Roberts et al. 2012): the Weri Slate formation (lower), the
the Weri River at an elevation of *1,380–1,350 m. A Assem Limestone, the Tsedia Slate and the May Kinetal
remarkable feature in the geomorphology of the Adwa dis- Limestone formation. This sequence is exposed within the
trict is the presence of so-called inherited landscapes (e.g. May Kinetal synclinorium. Thrust faults trending NE–SW
denudation surfaces, volcanic calderas and regoliths), which and N–S striking faults (Fig. 8.1) affecting the Neoprotero-
are components of different past geodynamic and climatic zoic basement rocks (Tsaliet and Tambien groups) resulted
settings. The most prominent is a regionally extensive ero- from regional deformation phases (Alene et al. 2006).
sion surface bevelled across the Precambrian basement and Towards the south of Adwa region, near May Kinetal, an
later buried by Mesozoic sedimentary rocks. Cenozoic rock intrusive body can be found. Avigad et al. (2007) dated this
uplift and denudation of rocks overlying the unconformity granitoid body at *612.3 ± 57 Ma, correlating it with another
have resulted in exhumation of these ancient landforms, post-tectonic intrusive body in a nearby area next to the southern
revealing new clues to the region’s long-term landscape Eritrean border, the Mareb granite (Tadesse et al. 1997, 2000;
evolution. Avigad et al. 2003; Miller et al. 2003; Asrat et al. 2004).
The resulting physiography, therefore, consists of a het- A major continental-scale geomorphic event occurred
erogeneous array of geomorphic provinces, each featuring a between 640 and 545 Ma (Beyth and Heimann 1999; Beyth
distinctive landform assemblage that preserves unique evi- et al. 2003), and the shaping capacity of this erosive event is
dence which can be used to reconstruct the history of still evident in Adwa landscape (see Sect. 8.2). This rapid
landscape evolution. During the last decades, there has been exhumation of the Neoproterozoic bedrock, estimated for up
a growing interest in physical, environmental and palaeo- to *2 km/Ma (Ghebreab 1999), led to the origin of a lev-
environmental studies in northern Ethiopia (e.g. Virgo and elled and peneplained surface, not only in northern Ethiopia,
Munro 1978; Butzer 1981; Machado et al. 1998, 2001; but on a continental-scale, extending from north Africa
Dramis et al. 2003). Most of these works put emphasis on towards Oman (Sandler et al. 2012). Beyth et al. (2003)
the contemporary processes of environmental degradation suggest extensive erosion to be associated with post-glacial
and soil erosion (e.g. Hurni 1983; Billi and Dramis 2002; erosion. Palaeoedaphic units, whose characteristics resem-
Frankl et al. 2012) and the societal causes and consequences bling modern oxisols, dated as sub-Cambrian, were descri-
of environmental degradation (e.g. Hurni 1986; Nyssen et al. bed from several weathering profiles of the Arabian–Nubian
2001, 2004; Feoli et al. 2002). However, there is a lack of Shield (Sandler et al. 2012), in NE Africa and southern
basic understanding of the regional land systems and their Israel, indicating a change of climate in the Early Cambrian,
connection through the long-term landscape evolution. The towards warm and humid conditions. This could have
comprehensive and novel work of Coltorti et al. (2007) on enhanced fluvial as well as subaerial erosion processes,
the planation surfaces in Ethiopia constitutes an example of allowing for the formation of extensive, regional-scale
such studies. Regional studies are the basis for spatial platforms of clastic sediments.
interpolation of process-oriented studies, assessment of Lying unconformably on the top of this surface, the
natural resources and environmental planning. The present Ordovician Enticho Formation (Merla and Minucci 1938;
chapter describes the geologic and geomorphic characteris- Beyth 1972; Merla et al. 1979) is formed by a subhorizontal
tics of the Adwa district and provides a reconstruction of white, often calcareous, medium- to coarse-grained sand-
major phases of long-term landscape evolution, from ancient stone, cross-bedded at the base. The upper unit of this for-
landforms up to most recent geomorphic features linked to mation interfingers southwards in the Adwa area with the
Quaternary climatic changes and human environmental Edaga Arbi glacial deposits (Dow et al. 1971). These
history. deposits grade upwards into a massive grey shale (Indaba
T’shama). Remnants of the basal unit lie unconformably on
the top of Late Neoproterozoic/Early Cambrian peneplain, or
8.1.1 Geologic Setting on the top of yellow-weathered sediments (north of Go-
longolo). Remnants of this glacial basal unit, reworked by
The oldest rock outcrops in the Adwa area, Neoproterozoic agriculture activities in the top 70 cm, contain gravels and
in age, belongs to the Arabian–Nubian Shield, and corre- some erratic submetric blocks. Recent palaeomagnetic
sponds to the Tsaliet Group (lower) and Tambien Group studies in northern Ethiopia conducted by Kidane et al.
(Beyth 1972; Tadesse et al. 1999). The foliated metavolcanic (2013) indicate a Late Carboniferous to Early Permian age of
and greenschist facies rocks of the Tsaliet Group are overlain the Edaga Arbi glacial sediments.
8 Geomorphology of the Adwa District 165
Fig. 8.1 Geomorphology of the Adwa district area, northern Ethiopia

166 M.J. Machado
The contact between the Edaga Arbi glacial deposits and the Adigrat formation, a palaeosol (laterite) was formed
the Triassic–Jurassic Adigrat Formation is characterised by (Merla and Minucci 1938; Merla et al. 1979; Garland 1980).
an erosional discontinuity (Merla and Minucci 1938; Beyth These palaeosols are affected by a heavy separation of plinth
1972), which developed during the widespread rifting that type iron oxihydroxides affecting the Adigrat sediments, and
took place in East Africa during the Permian and Triassic in highly eroded areas, the Neoproterozoic rock basement.
(Ebinger and Scholz 2012). Although this second erosion The topographical uplift of this non-orogenic region is
phase has little geomorphological expression in the present- thought to have been initiated *30–40 Ma and linked to
day northern Ethiopian landscape (Coltorti et al. 2007), rifting resulting from the Afar mantle circulation and plume
metric palaeoincision channels mark the contact between the activity (Burke 1996; Hofmann et al. 1997; Roberts et al.
lower Adrigrat Sandstone units and the Enticho Formation in 2012), as well as flank uplift of the Main Ethiopian Rift and
the Adwa area, and the unconformable contact between this Afar depression (Davis and Slack 2002; Beyene and Ab-
Triassic–Jurassic formation with the Neoproterozoic rock delsalam 2005). The uplift and westward tilting of the rising
basement described by Bosellini et al. (1997) provide pal- Ethiopian Plateau took place in three major phases (Baker
aeophysiographic data on the Adwa and the northern Ethi- et al. 1972; Hofmann et al. 1997; Pik et al. 2003: Gani et al.
opia surface topography and geomorphic processes between 2007) of Late Eocene, Mid-Miocene and Plio-Pleistocene
the Permian and the Early Mesozoic arid period. The Adigrat age, respectively. Volcanic activity associated with uplift
red siliciclastic sandstones, with horizontal to subhorizontal and later rifting was initiated *30 Ma ago in northern and
general structure, are the only Mesozoic formation preserved central Ethiopia, with the deposition of the basaltic Trap
in the Adwa area. The basin sedimentation during this period sequence within a period of 1–2 Ma (Hofmann et al. 1997).
was controlled by a widespread transgression covering the In some areas of the northern Ethiopian Plateau, this
continent from east to south (Beyth 1972). Minor faulting sequence exceeds 2,000 m in thickness (Mohr 1962). In the
along E–W axes occurred in Tigray during the Jurassic and Adwa area, the upper basaltic Trap sequence (Kazmin 1972)
the faults were later on reactivated in Late Cretaceous (Baker overlies unconformably the laterised Adigrat formation
et al. 1972), previous to the major Cenozoic uplift of the (Merla et al. 1979), supporting flat-top horizontal structure
area. of the plateaux, however, intruded by trachyte and phonolite
A third period of widespread erosion took place during plugs (Figs. 8.1 and 8.2) originated during the Miocene and
Late Cretaceous–Early Cenozoic and was described by Pliocene volcanic activity (Beyth 1972; Hagos et al. 2010).
Mohr (1962) as the Pre-trappean peneplanation. In the The Quaternary sedimentary record in the Adwa area is
northern part of the Adwa area, along the palaeoerosive composed by consolidated carbonate Late Pleistocene rocks
terrain surface which truncates the upper and lower units of (travertines), unconsolidated Pleistocene/Holocene alluvial
Fig. 8.2 Flat-top horizontal

structure plateau (laterised
Adigrat formation) topped by an
upper plateau (basaltic Trap
sequence), intruded by trachyte
and phonolite plugs. The lower
plateau overlays unconformably
the erosion surface that cuts the
Neoproterozoic bedrock
basement (light colours)
fan deposits, Holocene fluvial deposits (river terraces and The comprehensive work of Coltorti et al. (2007) on the
infilled valley deposits) and historical to present-day collu- planation surfaces of northern Ethiopia marks a rebirth of
vial deposits. regional geomorphology and of the studies on the genetic
mechanisms of the long-term geomorphological evolution of
the area.
8.2 Geomorphological Features Although landscape shaping by erosion is an inexorable
Associated with Long-Term ongoing process in the Adwa area, the main geomorphic
Landscape Evolution features give evidence of different relationships in the past
between the two main geomorphic driving forces imprinted
Two millennia of written archives document the political and in the geological history of the area: tectonic uplift and
socio-economic history, as well as the distinctive legacy of erosion.
the Adwa district and of Ethiopia in general, as the dis- Two landscape features of the Adwa district exemplify
semination core of important worldwide food cultivars the control of tectonic uplift and erosion (often in a feedback
(Kirwan 1972; Butzer 1981; Pankhurst 1985). Yet, not until loop): the plateau landscape and river incision.
the early sixteenth century, with the first comprehensive
written description made by Francisco Alvares in 1520
(particularly focused on human settlements, climate, rivers, 8.2.1 The Plateau Landscape
routes, society and natural resources) were the main phys-
iographic elements of the Adwa region and of the Ethiopian Past, as well as contemporary travellers heading to Adwa,
Highlands described (Alvares 1540; Ross 1922a, b). Ele- either from the nearby Ethiopian civilisation pillars of Axum
ments such as incised rivers and the prominent hills (corre- (east) and Yeha (north), or from further east (Adigrat) was
sponding to the phonolite and trachyte plugs) that emerge impressed by a singular landscape of high, stepped surfaces,
from extensive, high-altitude levelled surfaces were depicted above 2,200 m a.s.l., intruded by a myriad of steep slope
for the Adwa and Axum areas. hills, corresponding to columnar basalt necks and chimneys
Not until the late nineteenth century was there an attempt as well as to trachyte and phonolite plugs that can rise as
to explain the origin and formation of the Adwa district much as 300 m above the summit surface (Fig. 8.3). These
landforms: a stepped, low-gradient high surfaces, with large slightly undulating or flat-top plateaux where, throughout the
amphitheatres in the area of scarp retreat and incised valleys central and northern Ethiopian Highlands, villages and for-
trending southwards, cutting the faulted rock basement. tresses would be located due to its difficult access (Abdul-
Blandford (1870) not only attempted a general geologic Haggag 1961; Uhlig et al. 2003) are called locally Ambas (in
description of most of the Ethiopian territory, but was also Tigrinyan and Amharic languages).
trying to explain the origin of landforms, which he attributed In the Adwa district, two types of flat-top surfaces can be
to fluvial and soil erosion. Based on field observation of fine recognised, taking into account its underlying structure, and
sediment transport on slopes and by rivers during floods rock erodibility resistance. Some plateaux are underlain by
carried out during the main rainy season, he also suggested sedimentary and volcanic Mesozoic and Cenozoic strata,
the role of concentrated rainfall in the detachment of soil/ whilst other planation surfaces cut across the metasedimen-
rock particles. tary and metavolcanic rocks of the Neoproterozoic base-
During the 1930s, extensive studies were conducted in ment. The flattened surfaces of this landscape correspond to
the area by Merla and Minucci (1938) within a broad regional-scale erosion surfaces (Merla and Minucci 1938;
research and cartographic work, focused not only on geol- Beyth 1972; Coltorti et al. 2007) and can be correlated, from
ogy but also on the main physiographic characteristics of a the stratigraphic point of view, with major regional-, and
vast area of Tigray. The work of Merla and Minucci has sometimes continental-scale (Sandler et al. 2012)
indisputably set the basis for all subsequent studies on unconformities.
regional geomorphology in the Ethiopian Highlands. Par-
ticular interest was put on the plateau geomorphology. Merla 8.2.1.1 The Plateaux
and Minucci identified three main erosion surfaces, corre- The lower plateau (*2,200–2,300 m a.s.l.) is built on a
sponding to three topographical-levelled surfaces (suggested laterised unit that tops the nearly horizontally bedded Adi-
at the time to be end-Mesozoic, sub-Miocene and end-Ter- grat Sandstone Formation. The laterite was formed on a
tiary in age) and tentatively correlated with the three major palaeosurface that truncates the upper Adigrat formation.
East African erosion surfaces. These erosion surfaces would This surface corresponds to the second oldest regional-scale
be linked with three main phases of regional uplift (Merla planation surface, described by Coltorti et al. (2007) as the
and Minucci 1938; Merla et al. 1979; Mohr 1962). Late Triassic planation.
168 M.J. Machado
Fig. 8.3 Columnar basalt necks

and trachyte and phonolite plugs
intruding the basaltic trap flows
The upper plateau (>2,300 m a.s.l.) is built on horizontal environments and are thought to be an indicator of climatic
thick upper basaltic layers of the Trap sequence, whose (Gerson 1982; Sancho et al. 1988) or anthropic changes
deposition followed the major three-phase uplift and rifting (Everard 1964; Gutiérrez and Peña 1998), or may be the
initiated *40–30 Ma. The soils of the plateau, developed result of substantial gully erosion at the talus base (Boroda
over a detrital layer, are of the Ap-Bw type (Inceptisols) or et al. 2013). The sequence of three talus flatirons could
have a marked vertic character. This detrital layer, which suggests that during the Early–Middle Pleistocene, up to
contains fine material from the weathering of basalt rocks three phases of wetter/drier climatic condition (accumula-
and trachyte gravels from later, Plio-Pleistocene volcanic tion/erosion activity) may have existed, when plateau cliff
activity, would correspond to a Quaternary period of sub- retreat would have been more active than the one observed
aerial erosion processes. during the Holocene.
Differences in erodibility between the lower, laterised
Adigrat Sandstone and the upper basaltic Trap layer, explain 8.2.1.2 The Erosion Surfaces
the steep-walled cliffs of the Lower Plateau and faster retreat The best preserved evidence of the continental-scale rapid
of the slope walls of the Upper Plateau. exhumation of the Neoproterozoic basement is located
Cliff retreat, like the drainage system draining to the Weri southwards of Adwa. This exhumation was estimated at
River (see Sect. 8.2.3), follows the NNE–SSW structures *2 km/Ma (Ghebreab 1999) and resulted in the origin of a
(striking faults affecting the Neoproterozoic basement) and levelled and peneplained surface. Remnants of this surface
could have been triggered by the inception of the contem- constitute the upper summit interfluvial surfaces (*1,900–
porary drainage network in the area, initiated after the last 2,150 m a.s.l.) of the shet’s draining towards the Weri River
(Plio-Pleistocene) phase of regional uplift and westward (Fig. 8.1). The preservation of this surface, defined by
tilting of the Ethiopian Plateau. Coltorti et al. (2007) in northern Ethiopia as the PS1 (Pre-
Active slope erosion processes and landforms, such as Ordovician planation surface), shows a strong lithological
gullies, debris flows and rock falls in the laterised Adigrat and structural control.
sandstone unit, alluvial fans, past large landslide scars and In the Adwa area, it is best preserved in Work Amba
extensive alluvial Quaternary fans at the headwaters of the (Figs. 8.1 and 8.5) cut across the carbonaceous rocks of the
Teway Ruba shet (stream or river in the local language), Tambien formation. This planation surface was eventually
indicate considerable slope erosion during the Quaternary. covered by Late Palaeozoic and Mesozoic sedimentary
Debris-covered relict slopes observed in the road that sequences (Enticho, Edaga Arbi and Adigrat formations) and
connects Adwa and Indaba T’shama (Fig. 8.4) indicate that Tertiary basalts (Trap series) rocks. After the later Plio-
slope retreat was not continuous in time. Tripartite slopes or Pleistocene uplift, this surface was exhumed as the cover
talus flatirons are common in semiarid and arid rocks were eroded and subsequently affected by erosion,
Fig. 8.4 Tripartite slopes in the Adwa area
Fig. 8.5 Work Amba erosive

surface cutting the carbonaceous
rocks of the Tambien formation
resulting in a polygenetic surface. Quaternary denudation Pleistocene volcanic materials, indicating that the current
processes allowed for the development of karst landforms drainage pattern of the area was developed only after the
and well as the developed terra rossa soils that will be most recent uplift and tilting of the Ethiopian Plateau (Ad-
described in Sect. 8.3.1. amson and Williams 1987; Gani et al. 2007). Furthermore,
Said (1993) indicated that prior to the setting of the lower
Trap series, the stream network must have been directed
8.2.2 Quaternary Alluvial Terraces eastwards, towards the Red Sea, since there is no indication
of the existence of any significant drainage to the Nile. The
In the Adwa region, the fluvial headwaters are located in the post-Pliocene landscape evolution has led to the exhumation
plateaux (on volcanic and sandstone rocks) and rivers flow of the polygenic Work Amba erosion surface and fluvial
southwards to the Weri River cutting the metamorphic rocks dissection. The most complete sequence of staircase alluvial
of the Neoproterozoic basement. Three major streams cut the fill terraces was found next to the bridge crossing the old
Adwa region, namely the Teway Ruba Shet (1,101 km2 in Adwa–Mekele road at the Weri River (Fig. 8.6) The terrace
catchment area), the Kinetal Shet (262 km2) and the Ira Shet levels can be divided into three groups (Table 8.1): (a) Weri
(212 km2), whose direction is largely controlled by the major group or older alluvium (T0 to T2); (b) Daba Tadis group
structures trending NNW–SSE. In the southern part, the (T3 to T6) and (c) the present floodplain (T7). This sequence
Weri River flows ENE–WSW cutting into the Precambrian of fill terraces is not continuous along the Weri River
structures and carving a spectacular gorge with a meandering valley and varies in terms of their number and height
pathway. Between September and June, the river discharge, above the current river bed. The youngest terraces show a
including also the Weri, is greatly reduced, and rivers may higher preservation and can be traced all along the Weri
even dry up. This river system is dissecting the Plio- valley.
170 M.J. Machado
Fig. 8.6 Fill and strath terraces

location along the Weri River
between the Gwahiro and the
Ts’edyia shet’s reaches
Table 8.1 Topographical and sedimentological characteristics of fill and strath terraces of the Weri River between the Gwahiro and Ts’ediya
reaches
Level Alt. (m) Grain size (cm) Lithology Other
modal D90
T7 +1 5–10 90 Volcanic and metavolcanic (95 %) 2 m in thickness
14–32 Quartz (5 %)
DABA TADIS group T6 +3 1–3 19 Volcanic and metavolcanic (80 %) 90-cm-thick deposit with silty sand
5–7 Quartz (15 %) matrix (5YR). At top, 1.3-m-thick,
medium- to coarse-laminated sand
Schists (5 %) well sorted. Matrix colour
10YR5.5/3.5
T5 +10 1.5–5 16 Volcanic and metavolcanic (80 %) 4.2 m in thickness. Matrix colour
6–7 Quartz (15 %) 5YR
Schist (5 %)
T4 +13/14 1–4 28 Volcanic and metavolcanic (85 %) 3.5 m in thickness
7–12 Quartz (12 %)
Schist < 3 %
T3 +20/21 1.5–5 28 Volcanic and metavolcanic (85 %) D99 boulder long axis of 62 cm
8–15 Quartz (10 %)
Schist (5 %)
WERI group T2 +24 1.5–7 56 Metavolcanics (75 %) <1 m in thickness
9–14 Schist (20 %)
Quartz (5 %)
T1 +31 1.5–7 35 Metavolcanics (70 %) 3 m in thickness. D99 boulder long
12–15 Schist (20 %) axis of 1.3 m
Quartz (10 %)
TO +35/40 Strath terrace cutting Neoproterozoic
bedrock
The alluvial deposits of the Weri group terraces (older Soil loss due to surface processes is active even in gently
alluvium) consist of gravels and boulders with metavolcanic slopping areas, and northwards, closer to the plateau foot
(70 %), schist (20 %) and quartz (10 %) lithologies. The fill area, A-R-type profiles were identified. The area closer to the
terraces of the Daba Tadis group (younger alluvium) show contact with the depression carved into the May Kinetal
gravels and boulders with the dominant lithology of meta- granitoid body presents the thickest soil depth (>1 m, vs.
volcanic and volcanic (85 %) and quartz (10–15 %), and <0.75 m of soil depth closer to the plateau contact area),
occasional schist (<5 %). The grain size of the coarser developed over the May Kinetal limestones.
fraction (D90) for the Weri group terraces (56–35 cm) is Doline-shaped depressions were identified on the Work
larger than those in the Daba Tadis terraces (16–28 cm) Amba surface (Fig. 8.1) with ones most noticeable in size
which, adopting the equation of Williams (1983) for critical present in the vicinity of the faulted granitoid contact area. A
stream power, implies a decrease of flow energy by a factor major accumulation of fine subaerial dissolution material,
2.5. The alluvium of the current floodplain consists of even fine sand detrital limestone material and sediments resulting
higher proportion of metavolcanic and volcanic gravels from the erosion of the surrounding terra rossa have evolved
(95 %) and quartz (5 %) and lacks schist gravels. In the here under conditions different from the rest of the sur-
floodplain alluvium, the size of D90 is similar to the younger rounding area. Soils developed in these sediment-filled
Daba Tadis terrace group indicating a similar flow compe- karstic depressions are rich in smectite clays, responsible for
tence. The chronology of these stream terraces is unknown, the presence of common soil cracks during the dry season.
although we can speculate that the Weri group terraces may As in the case of the Vertisols developed over the weathered
correspond to the Early to Middle Pleistocene and the Daba volcanic materials of the Trap volcanic plateaux, soils in
Tadis group to the Late Pleistocene and Holocene. these depressions are characterised by pH ranging between 7
and 8, with a good supply of bases and water storage
capacity which favour an intense agricultural use.
8.3 Geomorphological Features Other assemblages from epikarst processes due to disso-
Associated with Quaternary lution beneath a soil and vegetation cover are hohlkarren
Surface Processes microforms (according to Bögli 1978 classification). The
hohlkarren micromorphology is characterised by groove and
The Plio-Pleistocene regional uplift and westward tilting pitted morphologies formed by corrosion of water which
have enhanced denudation processes that exhumed ancient percolates from pedogenetic cover and is enriched with
landforms and deepened river valleys connected to the Weri biogenic CO2 (Sweeting 1972). Other karren types are also
River. The most characteristic landforms resulting from present on bare rock surfaces, such as rillenkarren (Fig. 8.7)
Quaternary surface processes are those produced by weath- characterised by the presence of small-scale linear solution
ering of limestones (karst and travertines), granites (tors and flutes on steeply sloping limestone surfaces, resulting from a
weathering pits—gnammas), as well as alluvial deposits combination of chemical (dissolution), biological and
accumulated on valley flat bottoms. physical (raindrop impact) processes. Their trace disappears
at the soil-ground level. At some karren outcrops, the current
ground surface is *20–30 cm below the termini of the ril-
8.3.1 Karstic Landforms and Travertines lenkarren bottom level, indicating a period of soil erosion
following the karren development, which implies that these
As previously indicated, the Work Amba erosion surface is dissolution processes are not active in the present time.
an exhumed planation surface that cuts the Neoproterozoic The development of travertines in the eastern rim of the
carbonated Tambien Group formations and the May Kinetal Work Amba surface contributes to the evidence of past
granitoid intrusion. It resulted from multiple erosion phases active karst processes. The travertines show a cascade
during the Early Palaeozoic and Mesozoic and was covered, morphology (Pentecost and Viles 1994; Pentecost 2005)
locally, by Cenozoic rocks. As a consequence of the regional formed at spring outlets with vertical walls along resistant
uplift that took place *30–40 Ma ago, the erosion of cover carbonated rock layers of the Tambien formation. The
rocks exhumed the old peneplain and led to the development accretionary nature of cascades indicates a slow moving flow
of a new pediment surface. The latter is slightly tilted south- that prevents them from being eroded. The wall travertines
westwards and was likely covered by sediments during the are relict, although lower level springs are still an important
Pliocene–Pleistocene, although currently the only indirect source of water supply for the local population. Sedimen-
evidence is the exposed relict landforms typical of a mantled tologically, they are composed by porous calcic carbonate
karst. Soils display an Ap-Bt-R or Bt-R profile type are indicating an in situ cemented plant material with a highly
reddish in colour (particularly the Bt horizon) with shallow porous framework which can be classified as phytoherm
intermittent or broken diagnostic horizons of lithic nature. framestone facies type according to Pedley’s (1990)
172 M.J. Machado
Fig. 8.7 Dissolution

morphologies (rillenkarren) on
May Kinetal Limestone rock
surfaces
classification. The travertines yielded a radiocarbon date of 8.3.3 Flat Infilled Valleys
34,500 ± 1,100 years BP (author’s data). Lake carbonaceous
material from a southern Egypt area was found to yield a Holocene alluvial deposits in the valley bottoms (Fig. 8.9)
radiocarbon dating of 36,620 ± 1,100 years BP (Brookes provide a unique opportunity for palaeoenvironmental
1993). Brookes (1993) suggested that these were formed reconstructions of the area considering both the human
under much wetter conditions and from an increase in disturbance and other external environmental stresses such
rainfall in the Ethiopian Highlands between 20 and 40 as climatic variability. Machado et al. (1998) studied the
kyears. This wetter condition period would correlate in time valley infill stratigraphy within two drainage basins, the
with the May Kinetal travertines. Teway Ruba (Adwa and Wechi sites) and the Gwahiro rivers
(May Kinetal site). This study provided a high-resolution
record of the main landscape stability and instability stages
8.3.2 Landform and Processes on Granite during the past 4,000 year: (1) three major wetter periods
Bedrock (ca. 4,000–3,500 year B.P., 2,500–1,500 year B.P. and
1,000–960 year B.P.), during which soils were formed
Other singular landforms are tors and gnammas produced on and (2) two degradation episodes (ca. 3,500–2,500 year
granite rocks. In the study area, tors of about 10 m high are B.P. and 1,500–1,000 year B.P.), during which there was an
mainly located near the contact between the limestone and the increase of sediment yield from the slopes into the valleys.
granites, where other types of contact metamorphic rocks For the past 1,000 years, and in particular since the early
(hornstones) were formed (Fig. 8.8). During the cooling phase seventeenth century, stratigraphic records together with
of the granite emplacement, a well-developed network of historic chronicles suggest an increasing aridity. Although
transversal vertical cracks was produced facilitating subsequent some of these stages seem to have affected the area of Axum
hydrothermal circulation and mineral precipitation (Twidale and the Tigray province in particular, other stages show a
1982). Other sets of fractures were formed as the granite pluton good regional correlation not only with palaeoenvironmental
was subject to pressure release close to the surface and these are proxy data from different areas within Ethiopia, but also
subhorizontal joins following the shape of the land surface. from other eastern and north-eastern African regions.
Subsequent weathering along these joints and fractures has The oldest preserved stability phase is represented by the
resulted in the formation of mushroom-shaped tors character- buried Vertisol at the Adwa site (unit 3; Machado et al.
istic for the May Kinetal area. In addition, on horizontal 1998) which was correlated with one from the confluence of
granites surface there are ubiquitous gnammas, developed as a the Midmar and the Teway Ruba rivers and dated to
result of chemical processes exacerbated by wetting and drying 3,510 ± 80 14C year B.P. This date is consistent with a
processes during the rainy season. Vertisol described by Semmel (1971) in the Ethiopian
Fig. 8.8 Tors generated by

granite weathering along joints
and fractures (May Kinetal
intrusive granitoid)
Fig. 8.9 Holocene sedimentary

valley fill sequence in the May
Kinetal site
Central Highlands and dated between 3,865 ± 105 14C year shown in a colluvial unit (unit 4) of the Adwa Site (Machado
B.P. and 3,670 ± 105 14C year B.P. The existence of wetter et al. 1998). The probably reworked orange pottery frag-
conditions within Tigray and the Ethiopian Highlands can ments collected at the base of the unit (3,747 ± 328 year B.
also be inferred from the Nile river-level records registered P. TL), and the flecks of charcoal at the top (2,230 ± 30 14C
at the Roda Nilometer. Between 1840 and 1770 BC, there year B.P.) suggest that this drier phase may have persisted
are records of high waters (Butzer 1981) preceded by a over a long time span.
period of 35 years with frequent floods (Said 1993). An A second wet phase is reflected in the formation of a
increase in sediment yield from the slopes into the valleys is Vertisol (Soil II, Machado et al. 1998), which was developed
174 M.J. Machado
between 2,470 ± 30 14C year B.P. and 1,681 ± 169 year B. severe degradation of slopes producing high sediment yield.
P. (TL). This phase may be related to other Vertisols formed An example of this phenomenon has been reported by
around 2000 14C year B.P. in Sudan (Blokhuis 1993) and is Alvares (1540) on 30 April AD 1520. During his travel from
consistent with a lacustrine phase of lake Hanlé-Dobi Massawa to Axum, Alvares witnessed a flash flood during
between 2,500 and 1,500 14C year B.P. (Gasse et al. 1980). which the previously dry streams (arroyos), reached a high-
A pollen study in the Lake Turkana (Kenya) by Umer et al. water level in only 2 h.
(1995) also detected higher values of arboreal pollen Two buried soils (Soils IV and V; Machado et al. 1998)
between 2,400 and 1,900 14C year B.P. In the Weri River were developed in fine texture sediments over a time span
(catchment area over 4,300 km2), the analysis of slack-water shorter than 100 years. In Soil V, a charcoal sample col-
flood deposits shows at least four major extreme floods, with lected at the base of the sediment yielded an age of 480 ± 30
the uppermost flood unit yielding an age of 1,790 ± 60 14C 14
C year B.P., whereas the overlaying unit contains charcoal
year B.P. Hydroclimatically, these flood events are related to with a radiocarbon age of 380 ± 30 14C year B.P. These
anomalous rainfalls in both amount and intensity during the poorly developed soils suggest that the stability of the sys-
rainy season, rather than exceptional heavy storms over a tem is irregular and short lived, either as a result of lower
small part of the basin. The soil formation and the occur- soil moisture conditions or most probably due to intensive
rence of extreme events can be interpreted as resulting from land use.
an increase in rainfall. Historically, this period coincides A general trend towards drier climatic conditions since the
with the Early Axumite period (Butzer 1981; Sutton 1989). middle sixteenth century can be inferred from the chronology
At the archaeological site of the Staele Park, Axum, Butzer of drought and famine (Pankhurst 1985; Webb and von
(1981) described colluvial and reworked cultural debris of Braun 1994) and landscape stability recorded in the alluvial
Middle to Late Axumite age (fourth to eighth century AD) fills. The major change in the environmental conditions of the
and interpreted this as resulting from slope instability due to area may have been triggered by the drought of early sev-
over-intensive land use. A clear degradational stage was also enteenth century. Drier weather conditions during the sev-
found in the valley infill deposits at the study area (May enteenth and eighteenth centuries were registered not only in
Kinetal site), with evidence of lateral debris (Machado et al. East Africa, but also in the entire east–west extent of the
1998, unit 4: 1,490 ± 60 14C year B.P.) due to intensive Sahel zone (Nicholson 1980). This drought period correlates
slope erosion processes. A period of landscape recovery and with the accumulation of ca. 150-cm-thick gravelly colluvial
stability, probably the last major one, enabled the formation deposit which, together with the pottery fragments, contains
of a well-developed Vertisol (Soil III: 980 ± 50 14C year B. also numerous boulders with marks attributed to iron-tipped
P. and 963 ± 60 14C year B.P.; Machado et al. 1998). In a plough, indicating a strong human intervention. This evi-
sequence of four flood deposit units in the Weri River, two dence points out to progressive occupation of marginal lands
individual events were dated with charcoal and cultural and steep slopes, for both cultivation and grazing.
charcoal yielding ages of 970 ± 60 14C year B.P. and The present degradation period, already perceptible in the
980 ± 50 14C year B.P. aerial photographs from 1963, was accelerated through a
For the past 1,000 years, the valley infill deposits show combination of increased variability in annual rainfall, and
two major changes in the character of the sequences: (a) an an increase in human pressure on the land. The active gravel
anomalous increase in the intensity and length of aggrada- and boulder bars, deposited within a braided system, reveal
tion phases and (b) shorter periods of soil formation giving the present fluvial dynamics of the area, related to the
rise to poorly developed soils. The aggradational sequence ephemeral nature of river discharges and frequent flash
for the past 1,000 years is up to 13 m in thickness, in contrast floods. The sharp-peaked and short-lived hydrographs may
to only 4 m of aggradation for the previous 3,000 years. also be related to the presence of a thin patchy soil and
These deposits differ from the previous aggradational phases scarce vegetation on slopes. A minimum average erosion
in their fluvial character, with gravel bars and sand sheets. rate of 21 tons/ha/year for the period from AD 1985 to 1993
These can be interpreted not only as a result of an increase in has been estimated for a 6.7 km2 watershed located in the
fluvial competence (flash floods), transporting large amounts northern part of the study area by Machado et al. (1995,
of sediments through the valley systems, but may also reflect 2001), based on the filling of an irrigation dam.
Fig. 8.10 Landscape evolution in the Adwa district area since Late Proterozoic: drivers and major morphogenetic sequences
was characterised by well-marked periods of basement uplift

8.4 Conclusions alternating with long-periods of tectonic stability and
regional erosion with formation of pedimentation surfaces
The landscape of the Adwa region has an evident litholog- (Fig. 8.10). The climate control is more evident during the
ical and structural control and shows a heterogeneous mosaic stability periods, leading to the formation of palaeosols, rock
of ancient land surfaces and landforms preserving a unique weathering, travertine formation and alluviation on flood-
landscape history of the Ethiopian Highlands. The region plains and infill valleys.
176 M.J. Machado
The long-term landscape evolution was characterised by Acknowledgments I would like to express my very great appreciation
four major morphogenetic sequences since the Late Prote- to Professor Paolo Billi, not only for his valuable constructive sug-
gestions during the planning of this chapter, but for his willingness to
rozoic era, composed by cycles of aggradation sequences give his time and patience so generously.
followed by erosion phases. The oldest sequence culminated
in a regionally extensive erosion surface bevelled across the
Precambrian basement, with an excellent development in the References
southern Adwa district (Work Amba surface). A second
phase of erosion provides the contact between the Carbon- Abdul-Haggag Y (1961) A contribution to the physiography of
iferous glacial deposits and the Triassic–Jurassic sandstones, Northern Ethiopia. University of London, Athlone Press, London
but its geomorphic expression is limited in both the Adwa Adamson D, Williams MAJ (1987) Geological setting of Pliocene
area and in many other parts of northern Ethiopia. A third rifting and deposition in the Afar depression of Ethiopia. Jour
Human Evol 16:597–610
extensive aggradation–planation cycle took place during the
Alene M, Jenkin GRT, Leng MJ, Darbyshire FDP (2006) The Tambien
Late Cretaceous-Early Cenozoic and culminated in the group, Ethiopia: an early Cryogenian (ca. 800–735 Ma) Neoprotero-
development of a plinthic palaeosol (laterite). The topo- zoic sequence in the Arabian-Nubian shield. Precambr Res 147:79–99
graphical uplift of this non-orogenic region is thought to Alvares F (1540) Verdadeira informação das Terras do Prestes João das
Indias. Original re-edition of 1889, Agência Geral das Colónias,
have been initiated *30–40 Ma, causally linked with rifting
Lisboa
resulting from the Afar mantle circulation and plume activ- Asrat A, Barbery P, Ludden JN, Reisberg L, Gleizes G, Ayalew P
ity, within three major phases of volcanic activity of Late (2004) Petrology and isotope geochemistry of the Pan African
Eocene, Mid-Miocene, and Plio-Pleistocene age. Negash Pluton, Northern Ethiopia: Mafic-felsic magma interactions
during the construction of shallow-level calc-alkaline plutons. Jour
During the Quaternary, the river system was entrenched
Petrol 45:1147–1175
into the Plio-Pleistocene volcanic materials indicating that Avigad D, Kolodner K, McWilliams M, Persing H, Weissbrod T (2003)
the current drainage pattern of the area draining to the Nile Origin of northern Gondwana Cambrian sandstone revealed by
was developed only after the most recent uplift and tilting detrital zircon SHRIMP dating. Geology 31:227–230
Avigad D, Stern RJ, Beyth M, Miller N, McWilliams MO (2007)
of the Ethiopian Plateau. The Quaternary morphogenetic
Detrital zircon geochronology of Cryogenian diamictites and Lower
sequences are controlled by changes in the base level, with Paleozoic sandstone in Ethiopia (Tigrai): age constraints on
a dominant climatic signal. Changes in base level con- Neoproterozoic glaciation and crustal evolution of the southern
trolled the terrace formation in the Weri River valley and Arabian-Nubian shield. Precambr Res 154:88–106
Baker BH, Mohr PA, Williams LAJ (1972) Geology of the eastern rift
scarp retreat with evidence of multiple phases of talus
system of Africa. Geological Society of America, Special paper 136
flatirons at the slopes. The most complete sequence of Beyth M (1972) The Paleozoic-mesozoic sedimentary basin of the
staircase alluvial fill terraces shows up to seven terrace Mekele outlier-northern Ethiopia. AAPG 56:2426–2439
treads (Weri group and Daba Tadis group), composed by Beyth M, Heimann A (1999) The youngest igneous event in the
crystalline basement of the Arabian-Nubian Shield, Timna Igneous
volcanic, metavolcanic, schist and quartz. The grain size of
Complex. Isr. J. Earth Sci. 48:113–120
the coarser fraction (D90) decreases in the youngest ter- Beyth M, Avigad D, Wetzel HU, Matthews A, Berhe SM (2003)
races indicating a decrease of flow energy by a factor of Crustal exhumation and indications for snowball Earth in the East
2.5. Surface processes in relation with the regional climate African Orogen: north Ethiopia and east Eritrea. Precambr Res
123:187–201
favoured limestone weathering during more humid condi-
Billi P, Dramis F (2002) Geomorphological investigation on gully
tions, leading to the development of karst landforms and erosion in the Rift Valley and the northern highlands of Ethiopia.
travertine accumulation at spring outlets. During the Late Catena 50:353–368
Pleistocene and Holocene, alluvial units were deposited in Beyene A, Abdelsalam MG (2005) Tectonics of the Afar depression: a
review and synthesis. J Afr Earth Stud 41:41–59
valley bottoms and these provide a detailed record of
Blandford WT (1870) Observations on the geology and zoology of
climate variability and human disturbance. Landscape sta- Abyssinia, made during the progress of the British expedition to that
bility was interpreted during three wetter periods, having country in 1867–1868. Macmillan, London
been associated with soil development (dated ca. 4,000– Blokhuis WA (1993) Vertisols in the central clay plain of the Sudan.
3,500 year B.P., 2,500–1,500 year B.P., and 1,000– Wageningen Agricultural University, Wageningen
Bögli A (1978) Karsthydrographie und physische Spelëologie: Springer
960 year B.P.). Subsequent degradation episodes (ca. Verlag, Berlin
3500–2,500 year B.P. and 1,500–1,000 year B.P.) resulted Boroda R, Matmon A, Amit R, Haviv I, Porat N, Team ASTER, Rood
in an increase of sediment yield from the slopes into the D, Eyal Y, Enzel Y (2013) Long-term talus flatirons formation in
valleys. For the past 1,000 years, and in particular since the hyperarid northeastern Negev, Israel. Quat Res 79(2):256–267
Bosellini A, Russo A, Fantozzi PL, Assefa G, Tadesse S (1997) The
the early seventeenth century, stratigraphic records together Mesozoic succession of the Mekelle outlier (Tigre Province,
with historic chronicles suggest an increasing aridity. Ethiopia). Memore di Scienze Geologiche 49:95–116
Brookes IA (1993) Geomorphology and Quaternary geology of the Kirwan LP (1972) The Christian topography and the kingdom of
Dakhla oasis region, Egypt. Quat Sci Rev 12:529–552 Axum. Geogr J 138(2):166–177
Burke K (1996) The African plate. S Afr J Geol 99:339–409 Machado MJ, Pérez-González A, Benito G (1995) Reconstrución de
Butzer KW (1981) Rise and fall of Axum, Ethiopia: a geo-archaeo- ambientes naturales y antrópicos en la región de Axum (Tigray,
logical interpretation. Am Antiq 46(3):471–495 Norte de Etiopía). In Aleixandre T and Pérez-González A (eds)
Coltorti M, Dramis F, Ollier CD (2007) Planation surfaces in Northern Reconstrucción de paleoambientes y cambios climáticos durante el
Ethiopia. Geomorphology 89:287–296 Cuaternário. CSIC Monograph 3:163–174
Davis PM, Slack PD (2002) The uppermost mantle beneath the Kenya Machado MJ, Pérez-González A, Benito G (1998) Palaeoenvironmental
dome and relation to melting, rifting and uplift in East Africa. changes during the last 4000 years in the Tigray, northern Ethiopia.
Geophys Res Lett 29(7):21-1–21-4. doi:10.1029/2001GL013676 Quatern Res 48:312–321
Dow DB, Beyth M, Hailu T (1971) Palaeozoic glacial rocks recently Machado MJ, Pérez-González A, Benito G (2001) The role of
observed in northern Ethiopia. Geol Mag 108:53–60 geomorphology and geoarchaeology in understanding land degra-
Dramis F, Umer M, Calderoni G, Haile M (2003) Holocene climate dation. In: Feoli F, Pottier D and Woldu Z (eds) Sustainable
phases from buried soils in Tigray (northern Ethiopia): comparison development of dryland areas of East Africa. EC Directorate
with the lake level fluctuations in the Main Ethiopian Rift. Quatern General XII, Science, Research and Development, pp 63–67
Res 60:274–283 Merla G, Minucci E (1938) Missione geológica del Tigray. Rendiconti
Everard CE (1964) Climatic change and man as factors in the evolution della Reale Academia d’Italia. Centro Studi per l’Africa Orientale
of slopes. Geogr J 130(1):65–69 Italiana 3:1–362
Ebinger C, Scholz CA (2012) Continental rift basins: An East African Merla G, Abbate E, Azzaroli A, Bruni P, Canuti P, Fazzuoli M, Sauri
perspective. In C Busby A Azor (eds) Tectonics of Sedimentary M, Tacconi P (1979) A Geological Map of Ethiopia and Somalia
Basins: Recent Advances. John Wiley & Sons, Chichester, 183–208 (1973), 1;2,000,000 and Comment. Consiglio Nazionale delle
Feoli E, Gallizia L, Woldu Z (2002) Evaluation of environmental Ricerche. Univeristy of Florence, Italy
degradation in northern Ethiopia using GIS to integrate vegetation, Mohr PA (1962) The Geology of Ethiopia. University Collection Addis
geomorphological, erosion and socio-economic factors. Agric Ababa Press, Ethiopia
Ecosyst Environ 91(1–3):313–325 Miller NR, Alene M, Sacchi R, Stern R, Conti A, Kröner A, Zuppi G
Frankl A, Poesen J, De Dapper M, Deckers J, Haile M, Nyssen J (2012) (2003) Significance of the Tambien Group (Tigre. N. Ethiopia) for
Gully head retreat rates in the semiarid highlands of North Ethiopia. Snowball Earth Events in the Arabian–Nubian Shield. Precambrian
Geomorphology 173–174:185–195 Research 121:263–283
Gani NDS, Gani MR, Abdelsalam MG (2007) Blue Nile incision on the Miller NR, Stern RJ, Avigad D, Beyth M, Schilman B (2009)
Ethiopian Plateau: pulsed plateau growth, pliocene uplift, and hominin Cryogenian slate-carbonate sequences of the Tambien Group,
evolution. GSA Today 17(9):4–11. doi:10.1130/GSAT01709A.1 Northern Ethiopia. Precambr Res 170:129–156
Garland CR (1980) Geology of the Adigrat area. Memor No. 1. Nicholson SE (1980) Saharan climates in historic times. In: Williams
Ministry of Mines and Energy, Addis Ababa MAJ, Faure H (eds) The Sahara and the Nile. AA Balkema,
Gasse F, Rognon P, Street FA (1980) Quaternary history of the Afar Rotterdam, pp 173–200
and Ethiopian rift lakes. In: Williams MAJ, Faure H (eds) The Nyssen J, Haile M, Poesen J, Deckers J, Moeyersons J (2001) Removal
Sahara and the Nile. AA Balkema, Rotterdam, pp 361–400 of rock fragments and its effect on soil loss and crop yield, Tigray,
Gerson R (1982) Talus relict in deserts: a key to major climatic Ethiopia. Soil Use Manag 17:179–187
fluctuations. Isr J Earth Sci 31:123–132 Nyssen J, Poesen J, Moeyersons J, Deckers J, Haile M, Lang A (2004)
Ghebread W (1999) Tectono-metamorphic history of the Neoprotero- Human impact on the environment in the Ethiopian and Eritrean
zoic rocks in eastern Eritrea. Precambr Res 98:83–105 highlands—a state of the art. Earth-Sci Rev 64:273–320
Gutiérrez Elorza M, Peña JL (1998) Geomorphology and Late Pankhurst R (1985) The History of Famine and Epidemics in Ethiopia
Holocene climatic change in northeastern Spain. Geomorphology Prior to the Twentieth Century. Relief and Rehabilitation Commis-
23:205–217 sion, Addis Ababa
Hagos M, Koeberl C, Kabeto K, Koller F (2010) Geochemical Pedley HM (1990) Classification and environmental models of cool
characteristics of the alkaline basalts and phonolite-trachyte plugs of freshwater tufas. Sed Geol 68:143–154
the Axum area, northern Ethiopia. Austrian J Earth Sci 103(2): Pentecost A, Viles H (1994) A review and reassessment of travertine
153–170 classification. Géog Phys Quatern 48(3):305–314
Hofmann C, Courtillot V, Feraud G, Rochette P, Yirgu G, Ketefo E, Pentecost A (2005) Travertine. Springer, Berlin
Pik R (1997) Timing of the Ethiopian flood basalt event and Pik R, Marty B, Carignan J, Lave J (2003) Stability of the upper Nile
implications of Plume birth and global change. Nature 389:838–841 drainage network (Ethiopia) deduces from (U-Th)/he thermochro-
Hurni H (1983) Soil erosion and soil formation in agricultural nometry: implication of uplift and erosion of the Afar plume dome.
ecosystems: Ethiopia and northern Thailand. Mt Res Dev 3 Earth and Planetary Science Letters 215:73–88
(2):131–142 Roberts EM, Stevens NJ, O’Connor PMO, Dirks PHGM, Gottfried D,
Hurni H (1986) Soil conservation in Ethiopia. Community forests and Clydes WC (2012) Initiation of the western branch of the East
Soil Conservation Development Department Ministry of Agricul- African Rift coeval with the eastern branch. Nat Geosci 5. doi:10.
ture, Addis Ababa 1038/NGEO1432
Kazmin V (1972) The Geology of Ethiopia. Ethiopian Institute of Ross ED (1922a) Early travellers in Abyssinia: part I. J R Afr Soc 21
Geological Surveys. Note Nº 821-051-12 (84):268–278
Kidane T, Bachtadse V, Alene M, Kirscher K (2013) Palaeomagnetism Ross ED (1922b) Early travellers in Abyssinia: part II. J R Afr Soc 22
of Palaeozoic glacial sediments of Northern Ethiopia: a contribution (85):5–16
towards African Permian palaeogeography. Geophys J Int, First Said R (1993) The River Nile. Geology, hydrology and utilization.
online October 3, 2013; doi:10.1093/gjl/ggt336 Pergamon Press, Oxford
178 M.J. Machado
Sancho C, Gutiérrez M, Peña JL (1988) A quantitative approach to Tadesse T, Hoshino M, Suzuki K, Iisumi S (2000) Sm-Nd, Rb-Sr, and
scarp retreat starting from triangular slope facets, central Ebro Th-U-Pb zircon ages of syn- and post-tectonic granitoids from the
Basin, Spain. Catena supplement 13:139–146 Axum area of northern Ethiopia. J Afr Earth Sci 30:313–327
Sandler A, Teutsch N, Avigad D (2012) Sub-Cambrian pedogenesis Twidale CR (1982) Granite Landforms. Elsevier, Amsterdam
recorded in weathering profiles of the Arabian Nubian shield. Uhlig S, Yiman B, Crummey D, Goldenberg G, Marrasini P, Aregay
Sedimentology 59:1305–1320 MW, Wagner E (2003) Encyclopaedia Aethiopica, vol. 1: A-C,
Semmel A (1971) Zur junquartaren klima und Reliefentwicklung in der Harrassowitz Verlag, Wiesbaden
Danakilwuste und ihren westlichen Randgebieten. Erdkunde Umer M, Bonnefille R, Johnson TC (1995) Pollen and isotopic records
25:199–209 in late Holocene sediments from Lake Turkana, Kenya. Palaeog-
Sutton JEG (1989) The history and chronology. In: Munro Hay SC (ed) eography Palaeoclimatoogy Palaeoecology 119:371–383
Excavations at Aksum: memoirs of the British Institute in Eastern Virgo KJ, Munro RN (1978) Soil and erosion features on the Central
Africa, No. 10, pp 7–26 Plateau region of Tigrai, Ethiopia. Geoderma 20:131–157
Sweeting MM (1972) Karst Landforms. MacMillan, London Webb P, von Braun J (1994) Famine and food security in Ethiopia:
Tadesse T, Suzuki K, Hoshino M (1997) Chemical Th–U total Pb lessons for Africa. Wiley, Chichester
isochron age of zircon from the Mareb Granite in northern Ethiopia. Williams GP (1983) Paleohydrological methods and some examples
J Earth Planet Sci Negoya Univ 44:21–27 from Swedish fluvial environment, I—cobble and boulder deposits.
Tadesse T, Hoshino M, Sawada Y (1999) Geochemistry of low-grade Geogr Ann 65 A:227–243
metavolcanic rocks from the Pan-African of the Axum Area,
northern Ethiopia. Precambr Res 99:101–124
The Amba Landscape of the Ethiopian
Highlands, Shaped by Rockfall 9
J. Nyssen, J. Moeyersons, J. Deckers, Mitiku Haile, and J. Poesen
Abstract
Cliff retreat occurs on the ambas or structurally determined stepped mountains of the northern
Ethiopian highlands. This chapter describes the rock fragment detachment from cliffs by rockfall,
quantifies its annual rate and identifies factors controlling rock fragment movement on the scree
slopes. It further presents a conceptual model explaining rock fragment cover at the soil surface in
these landscapes. In the May Zegzeg catchment (Dogu’a Tembien district, Tigray), rockfall from
cliffs and rock fragment movement on debris slopes by run-off and livestock trampling were
monitored over a 4-year period (1998–2001). Rockfall and rock fragment transport mainly
induced by livestock trampling appear to be important geomorphic processes. Along a 1500 m
long section of the Amba Aradam sandstone cliff, at least 80 t of rocks are detached yearly and fall
over a mean vertical distance of 24 m resulting in a mean annual cliff retreat rate of
0.37 mm year−1. Yearly unit rock fragment transport rates on scree slopes ranged between 23.1
and 37.9 kg m−1 year−1. This process is virtually stopped when exclosures are established.
A conceptual model indicates that besides rockfall from cliffs and argillipedoturbation, all factors
and processes of rock fragment redistribution in the study area are of anthropogenic origin.

Keywords
Debris slope Livestock trampling Rock fragment redistribution Subhorizontal structural
relief
9.1 Introduction retreat. Taking into account the high rates of other mass
movements (landsliding, sheet and rill erosion), fallen
The subhorizontal geological formations of the Ethiopian material is rapidly removed and cliff retreat processes are
highlands have been epirogenically uplifted over the last 25 maintained. The process immediately following rockfall
million years. This has led to important incision and cliff which occurs downslope is the further removal of rock
fragments by rolling, mainly induced by animal trampling.
Previous studies in the northern Ethiopian highlands
(Nyssen et al. 2000, 2002a; Moeyersons et al. 2006a) have
J. Nyssen (&)
also analysed the origin of the extensive rock fragment
Department of Geography, Ghent University, Krijgslaan 281 (S8),
9000 Ghent, Belgium covers. The rock fragment cover (RC) of Vertisols and soils
e-mail: jan.nyssen@ugent.be with vertic properties is clearly a result of swell–shrink
J. Moeyersons action in these soils. The rapid appearance of new rock
Royal Museum for Central Africa, 3080 Tervuren, Belgium fragments at the surface after field clearing has been related
J. Deckers J. Poesen to the active polygonal structures of the Vertisols. Rock
Department of Earth and Environmental Sciences, KU Leuven, fragments, appearing yearly at the surface of Vertisols, have
Celestijnenlaan 200E, 3001 Heverlee, Belgium been shown to belong lithologically to deposits underlying
M. Haile the Vertisols. Given this stratigraphical situation and the
Department of Land Resource Management and Environmental activity of Vertisols in the study area, the rock fragment
Protection, Mekelle University, P.O. Box 231, Mekelle, Ethiopia
180 J. Nyssen et al.
covers are thought to be squeezed up as a consequence of (Adamson and Williams 1980), and various lithologies are
argillipedoturbation accompanying swell–shrink cycles in exposed giving rise to a typical structural subhorizontal
the vertic horizon. relief with tabular, stepped landforms.
Another study in the northern Ethiopian highlands Major geological formations outcropping in the region
(Nyssen et al. 2002b) showed that the balance between lat- comprise Precambrian metamorphic rocks at the base, the
eral and vertical movements of rock fragments controls the only formations that are strongly folded and faulted and that
spatial distribution of rock fragment cover. Vertical supply were subsequently truncated by erosion. Next, the subhori-
of rock fragments to the soil surface is caused by (1) tillage- zontal formations comprise the Palaeozoic Enticho sandstone
induced kinetic sieving, bringing preferentially large rock (of fluvioglacial origin) and the Edaga Arbi tillites (Bussert
fragments (>7.5 cm) to the surface, even in the case of and Schrank 2007). The Mesozoicum is represented by the
continuous fine sediment deposition, (2) argillipedoturbation lower transgressional Adigrat sandstone (particularly cliff-
in Vertisol areas and (3) selective run-off erosion and the forming—Fig. 9.1), overlain by alternating hard and soft
development of erosion pavements. With respect to the lat- Antalo limestone layers, some 400 m thick, and by Amba
eral displacement processes, one can distinguish between (1) Aradam sandstone (Hutchinson and Engels 1970) (Fig. 9.2).
lateral transport over the soil surface by trampling and Two series of Tertiary lava flows, separated by silicified
concentrated overland flow, especially on steep slopes and lacustrine deposits (Merla and Minucci 1938; Arkin et al.
(2) rockfall from the cliffs. The latter two processes are 1971; Merla et al. 1979) cap these Mesozoic sedimentary
studied in detail in this chapter, based on observations in the rocks. As a consequence of the vertical succession of
northern Ethiopian highlands. numerous subhorizontal sedimentary rock formations as well
Gardner (1970), Lee et al. (1994), Govers and Poesen as Tertiary sills and lava flows, a subhorizontal structural
(1998) and Oostwoud Wijdenes et al. (2001) showed that the landscape of scarps and dips (Young 1987; Young and Wray
above two processes play a significant role in slope devel- 2000) has come into existence. This stepped topography,
opment of some mountain areas. Ayalew and Yamagishi resulting from the variable hardness of the different geological
(2004) also insisted on the importance of rockfall, concur- formations, is locally described as amba landscape.
rently with landsliding in shaping of the Blue Nile gorge.
Therefore, it was expected that these processes partly explain
the presence of rock-fragment-rich layers in Skeletic Regosols 9.2.2 Study Area
on debris slopes reported from the study area (Nyssen et al.
2008). Hence, in this chapter, based on an earlier publication For this study, the May Zegzeg catchment (Dogu’a Tembien
(Nyssen et al. 2006), we will (1) introduce the nature of district), a 199-ha subcatchment of Geba and Tekezze
stepped topography present so widely in northern Ethiopia, (Fig. 9.3), situated at 2,280–2,650 m a.s.l., was selected as a
(2) quantify the annual rock fragment transport rate caused by representative catchment for the northern Ethiopian high-
these processes in the northern Ethiopian highlands, (3) ana- lands. The subhorizontal geological formations in the
lyse the factors controlling rock fragment movement on scree catchment comprise layers of the Mesozoic Antalo limestone
slopes and (4) develop a conceptual model explaining RC at and Amba Aradam sandstone in the lower parts, and Tertiary
the soil surfaces based on major controlling factors. basalt flows (traps) with silicified interbedded lake deposits
in the upper parts. Quaternary deposits, consisting of allu-
vium, colluvium and tufa, are also found. The study area
9.2 The Amba Landscape comprises a typical red-black soil catena (Driesen and Dudal
1991) on basalt and Calcisols and Calcaric Regosols at the
9.2.1 Geomorphic Context foot of the limestone cliff (Nyssen et al. 2008).
The main rainy season (>80 % of total rainfall) extends from
In Ethiopia, most of peneplained Palaeozoic and Mesozoic June to September, but is preceded by 3 months of dispersed
sedimentary rocks have been concealed by Tertiary basaltic small rains. Average annual rainfall is 750 mm (Nyssen et al.
flows (Mohr 1963; Merla et al. 1979; Coltorti et al. 2007). 2005). Intense rains falling on bare soils, which have already
The dome-like uplift of the Arabo-Ethiopian region started lost most of their natural vegetation by century-long action of
during the Oligocene and had two periods of intense tectonic human society, cause severe soil erosion. Erosive rains and the
activity: in the Miocene, about 25 million years ago and in predominance of steep slopes induce a natural vulnerability of
the Plio-Pleistocene (Williams and Williams 1980). The the study area to soil erosion, despite overall low soil erodibility
elevation above sea level of the base of the basalt—about due to high clay contents and high rock fragment content. Daily
500 m in 25 million years in the southernmost areas of air–temperature variations are large (range of more than 20 °C,
Ethiopia and up to 2,500 m in the north—shows the with 5 and 28 °C as extreme range values) during the dry
importance of this uplift. All the rivers are deeply incised season, without, however, dropping below freezing point.
9 The Amba Landscape of the Ethiopian Highlands, Shaped by Rockfall 181
Fig. 9.1 The Tsaliet valley draining the Tembien highlands in Tigray flank: Edaga Arbi tillites, a relatively soft rock that is easily eroded
towards the north, seen from Ba’ati Woyane (13.670704°N, away (2); subvertical, high cliffs in Adigrat sandstone (3); alternating
39.162169°E). At the lower position, in stratigraphical order, are the hard and soft rock layers of Antalo limestone (4); again a subvertical
folded Precambrian rocks, visible at the far end, that result in an cliff of Amba Aradam sandstone (5); and finally the typical trap basalt
undulating topography (1). All subsequent subhorizontal formations are (6) that caps the sedimentary transgression–regression series
part of the amba landscape. From bottom to top in the opposite valley
After deforestation, which took place over the last amphitheatre-shaped slope section in the middle of the
4000 years (Hurni 1985; Moeyersons et al. 2006b), topsoil study catchment, and the 30-m-high and 1,327-m-long
and subsoil were removed in many places, predominantly by basalt cliff line (elevation: around 2,600 m a.s.l.) in the
water and tillage erosion. In remnant forests of the study area, upper part of the catchment (Fig. 9.3).
thick Phaeozems are found which are totally absent in nearby From 1998 to 2001, the slopes below both cliffs were
deforested areas in a comparable geomorphic context. Pres- inspected twice a year (at the end of the rainy season and in
ently, there is an active policy to reforest steep slopes, which the middle of the dry season) in order to record rockfall
is, however, not expected to completely restore the original events as indicated by (1) straight downslope scars of
climax vegetation and soils (Descheemaeker et al. 2006). damaged vegetation along the path of the rockfall; (2) bright
Field observations indicate that two geomorphic processes colour of freshly fallen and broken rocks; (3) up to 10 cm
occur on steep slope sections of the study area: (1) rock fall at deep impacts of falling rocks in cropland soil; and (4)
the end of the rainy season, inferred from the presence of information provided by shepherds about the moment,
fresh sand- and limestone blocks, up to 3 m across, on debris conditions and magnitude of the event. For each major
slopes below cliffs; (2) rock fragment transport triggered by rockfall event (Fig. 9.4), the volume of the fallen rock
livestock trampling, inferred from the noise of rolling and fragments was measured with a metre stick and horizontal
falling rock fragments, when livestock, especially goats but and vertical displacement distances with a theodolite. A
also cattle and sheep, grazes on steep slopes. weighed average displacement distance was calculated,
taking into account the volumes of transported rock frag-
ments (Fig. 9.5).
9.3 Materials and Methods
9.3.1 Rockfall Monitoring 9.3.2 Livestock Trampling Monitoring
Rockfall was monitored along the 60-m-high and 1,500-m- Medium-term monitoring of individual rock fragment
long Amba Aradam sandstone and limestone cliff line movement along steep slope sections, especially by livestock
(elevation: around 2,500 m a.s.l.), which forms an trampling, was conducted on three slope sections
Fig. 9.2 A typical amba

landscape reflects the presence of
subhorizontal lithological
structure and rapid incision. At
Guyeha, near the study area, a
prominent Amba Aradam
sandstone cliff occurs (total height
is ca. 40 m). People on the road
for scale (Photo J Nyssen)
representative for the selected catchment and its surroundings Although the aim was to use tracers similar to the rock
(Table 9.1; Fig. 9.4). Sites 1 and 2 (Fig. 9.4) were situated in fragments naturally present on the slope, it should be noted
an intensely grazed rangeland. Site 3 was located in a 5-year- that, at site 1, the basalt rock fragments were somewhat
old exclosure, where no livestock is allowed to enter, but flatter compared to the tracers (Table 9.1). The tracers were
where people come once a year to cut grass, especially for installed in July 1999. At each site, a 20–50 m long line
roofing. This last site, like the rangelands, bears some rills along the contour was materialised by a rope, following
generated by run-off, which overtops the upslope cliff. straight sections between large rocks, on which markings
Since painted rock fragments risked to be picked up by were painted. Rock fragments on the soil surface, or
shepherds, limestone rock fragments were used as tracers in embedded for less than half their volume, which were
these basalt and sandstone environments. Their three main crossed by this line, were removed and replaced by tracers of
diameters were measured and the flatness index (FI; Cailleux similar shape and size (1–4 tracers per m). In March 2001
1945) was calculated as: (20 months later), the ropes were installed again (Fig. 9.6),
tracers were recovered, and if they had moved from their
FI ¼ ðd1 þ d2 Þ=ð2d3 Þ ð9:1Þ original position, the shortest distance (which was along the
steepest slope) to the rope was measured. In rangeland,
where recovery rates were relatively low (56–67 %). That is
d1 longest diameter; probably due to (1) recent rockfall deposits which in places
d3 shortest diameter, perpendicular to d1; visibly covered the original soil surface and (2) the shep-
d2 intermediate diameter, perpendicular to d1 and d3. herds might have picked up some of the tracers by curiosity,
Fig. 9.3 Location of the May Zegzeg catchment in the northern Ethiopian highlands (after Nyssen et al. 2006)
Fig. 9.4 Location map of major

rockfall events (1998–2001) and
of livestock trampling monitoring
sites within the May Zegzeg
catchment (after Nyssen et al.
2006)
4–8 times at each site, using a 200-cm long chain with

3 × 1.5 cm links, which was placed on the surface in a
downslope direction following all irregularities of the soil
surface, while the shortest distance between its beginning
and end (Ds, in cm) was measured:
RI ¼ 200 ðlength of chain in cmÞDs ð9:2Þ
Unfortunately, grazing intensity could not be measured at

the experimental sites, and official data, organised per
municipality, could not be used because (1) they were
computed for large areas and (2) access to rangeland is also
open to livestock from Hagere Selam town and other
neighbouring villages. In a qualitative way, it can be stated
that the much degraded rangeland near Harena Village (site
2) is most intensively grazed, especially by goats. Grazing
pressure is somewhat less in the Zenako rangeland (site 1),
which is used by cattle, sheep and goats. In the exclosure
(site 3), there is no grazing.
9.3.3 Rainfall Analysis
Monthly and daily rainfall data of the nearby Hagere Selam

station were analysed to assess the representativeness of the
rain events triggering rockfall. Antecedent rain depth since
the start of the rainy season in the years with observed
rockfall was also compared to rain depths of a 20-year long
series.
Fig. 9.5 Three months after the occurrence of a rockfall event (August
1998) on the sandstone cliff. Fresh rock fragments (arrows) are easily 9.4 Results and Discussion
recognisable by their bright colour. The upper arrow indicates the place
of origin of these rocks on the cliff. Despite recovering vegetation, the
rockfall path can still be recognised on the backslope (after Nyssen 9.4.1 Yearly Rockfall
et al. 2006)
No major rockfall events were observed on the basalt cliff
since their colour contrasted with that of the surrounding during the four years of observation. Here, rockfall seems to
area. occur primarily in the form of toppling of parts of individual
At each site, the following environmental characteristics basalt columns.
were measured or assessed: mean slope gradient, mean areal Along the sandstone and limestone cliffs, four events
percentage of soil covered by short grass, long grass, shrubs were recorded, always in August, when the soils around and
or rock fragments, and that of the bare soil. The roughness in between the rocks are saturated by water (Table 9.2).
index (RI; Oostwoud Wijdenes et al. 2000) was determined Some young shepherds from the area (Fig. 9.5) observed one
Table 9.1 Characteristics of the monitoring sites for livestock trampling and experimental conditions (after Nyssen et al. 2006)
Site Lithology Mean slope Rock fragments at surface Tracers
gradient (m m−1) Mean d2 FI d2 FI Installed Recovered Recovery
cover (%) rate (%)
1. Zenako rangeland Basaltic coll 0.55 62 6.2 2.60 5.8 1.84 64 43 67
2. Harena rangeland Sandstone coll 0.72 30 5.0 2.07 5.7 1.88 81 45 56
3. Harena exclosure Sandstone coll 0.85 14 6.0 2.07 4.6 1.88 60 53 88
d2 mean intermediate diameter (cm); FI flatness index. They are based on measurements of 30–40 randomly selected rock fragments/tracers
Fig. 9.6 Harena monitoring sites

#2 and #3 for individual rock
fragment movement (July 1999).
2 rangeland; 3 exclosure.
Between both sites is a cropland
on an ancient debris flow deposit
(after Nyssen et al. 2006)
Table 9.2 Rockfall events at the sandstone and limestone cliffs (1998–2001) (after Nyssen et al. 2006)
Approx. date Lithology Total Rock density Total Horizontal displacement (m) Vertical displacement (m)
volume (m3) (kg m−3) mass (103 kg) Min Max Weighted Min Max Weighted
average average
August 1998 Sandstone 10.49 2,400 25 4.0 150.9 80.8 1.1 77.9 46.9
August 1998 Limestone 13.36 2,535 34 9.2 35.3 14.0 8.5 24.0 11.6
Mean (1998–2001) 33.10 80 37.0 24.1
major rockfall event: We saw one big rock rolling from the Antecedent rain depth since the beginning of the rainy
top of the cliff; it broke into pieces when rolling along the season and rain depth for August showed average conditions
slope. Some large rocks reached the cropland and they during three years, but in 1998, they had the second highest
broke into pieces at different sites. It was during daytime, in value in the 20-year series. Hence, it is anticipated that
August 1998. It was raining. We were not very far with our rockfall events reported here are representative for both
cattle. The villagers came quickly when they heard the noise, average and extreme rainfall conditions of the region.
because they were afraid for us. Fortunately nobody was The mean yearly rockfall of 80 t, observed along a 1,500-
hurt. Rockfall with big noise like this is exceptional. m long section (parallel to the contour) of the Amba Aradam
All rockfall events in 1998, 2000 and 2001 occurred at sandstone cliff over an average vertical distance of 24 m
the peak of the rainy season. Unlike temperate mountain (Table 9.2), should be considered as a minimum, since we
areas, where rockfall is strongly correlated with temperature did neither account for many small events involving only a
variations (Perret et al. 2006), rainfall is the main triggering few rock fragments of some kg, nor for possible extreme
factor for rockfall in the study area. Rainfall analysis shows rockfall events related to exceptionally high daily rain. An
that the maximum daily rain depths in August during the extrapolation of the observed rockfall rates to the 1,500-m-
study period correspond to 20-year average values of around long and 60-m-high cliff indicates that the cliff would retreat
40 mm day−1. Extreme daily rainfall events, such as those by at least 0.37 mm year−1 or 3.7 cm century−1. One of the
occurring in 1975 (67 mm) and 1980 (66 mm), were not few studies on sandstone cliff retreat in (sub)tropical regions
observed during the study period. (Young and Wray 2000) reported geological scarp retreat
Fig. 9.7 Tracers (white

limestone) displaced over a
distance of 6 m in 20 months
(Harena rangeland, March 2001).
The rope at the back shows the
original tracer position (after
Nyssen et al. 2006)
Table 9.3 Tracer displacement in 20 months, perpendicular to the contour (after Nyssen et al. 2006)
Site Percentage of tracers moved Displacement distance (m) of moved Mean displacement distance (m) of all tracers
tracers
Meana Standard deviation
1. Zenako range 72 1.74* ±2.37 1.25
2. Harena range 95 2.31* ±1.97 1.80
3. Harena exclosure 66 0.90# ±1.00 0.59
a
Different symbols indicate significantly different values (α = 0.1) based on unpaired Student’s t-test
rates ranging between 1.5 and 2.5 cm century−1. On the tracers is, however, significantly smaller in the exclosure
Colorado Plateau, scarp retreat rate averages 1.6 cm cen- (0.9 m) than in the rangelands (1.74 and 2.31 m) (Table 9.3).
tury−1 (Young 1985). These mean long-term values are In particular, the difference in tracer displacement distance
lower though of a similar order of magnitude as our between the rangeland and the exclosure on the Harena scree
observed short-term cliff retreat rates. slope is highly significant (α = 0.001). The percentage of
tracers moved as well as the mean distances of tracer
movement increase with increasing grazing pressure (Harena
9.4.2 Rock Fragment Movement rangeland > Zenako rangeland > Harena exclosure).
over Debris Slopes Looking at other environmental characteristics of the
three sites, further reasons for the differences in displacement
Twenty months after their placement, most tracers had left distances become clear. If mean surface roughness is similar
their original position (Fig. 9.7). In the two rangelands, 72 at the study sites (roughness index between 34 and 46),
and 95 % of the tracers had moved, but it appeared that also vegetation cover, and especially long grass cover, seems to
in the exclosure, 66 % of the tracers received an impulse be the primary explanatory factor for the observation that
which was strong enough to initiate their movement tracers moved over much smaller distances in the exclosure.
(Table 9.3). Besides exceptional illegal livestock grazing, Major explanatory factor for smaller tracer displacement
such impulses can also be caused by rock fragments falling distances in Zenako, compared to Harena, is the difference in
and rolling from the cliff, by wild animals such as hare, area of smooth surface in the latter (53 %) as compared to
jackal, hyena, caracal and porcupine (Yami et al. 2007), by the former (20 %).
run-off produced above the cliff, by forest guard and by A yearly unit rock fragment transport rate can be calcu-
people who come occasionally in the area, especially to lated, similar to the unit soil transport rate for contour
harvest grass. Mean displacement distance of the transported ploughing (Poesen et al. 1997):
Table 9.4 Calculation of the yearly unit rock fragment transport rate (Qs) for the three monitored sites (t = 1.67 year) (after Nyssen et al. 2006)
Site Mean displacement Slope gradient δ (m) d3 (m) RC (%) ρ (kg m−3) Qs (kg m−1 year−1)
distance (m) along (m m−1)
the slope of all tracers
1. Zenako range 1.25 0.55 1.10 0.034 62 2734 37.9
2. Harena range 1.80 0.72 1.46 0.035 30 2511 23.1
3. Harena exclosure 0.59 0.85 0.45 0.035 14 2511 3.3
δ horizontal component of the net mean downslope displacement distance of all tracers (including those that did not move), in the direction of the
steepest slope; d3 shortest stone diameter perpendicular to the longest diameter; RC rock fragment cover; ρ rock density
Fig. 9.8 Major factors in rock

fragment redistribution,
characterising the present day
landscape and agricultural
system. Dotted lines refer to rock
fragment movements,
schematically represented by
short arrows on the figure, and
solid lines indicate relations
between factors and processes
(bold frames) (after Nyssen et al.
2006)
9.4.3 Cliffs and Scree Slopes as Part

Qs ¼ d d3 RC q t1 ð9:3Þ of Catenas
where Qs = yearly unit rock fragment transport rate The current understanding of the detachment processes on
(kg m−1 year−1); δ = horizontal component of the net mean cliffs and transport processes over scree slopes can now be
downslope displacement distance of all tracers (including integrated with results of previous studies to analyse rock
those that did not move), in the direction of the steepest fragment redistribution along catenas on stepped mountains
slope (m); ρ = rock density (kg m−3); and t = period over in the northern Ethiopian highlands (Fig. 9.8).
which tracer displacement was monitored (year). The yearly unit rock fragment transport rate for cliffs
The yearly unit rock fragment transport rate is larger in (≥53 kg m−1 year−1) is of the same order of magnitude as
the Zenako rangeland (site 1) than in Harena (site 2) that for rangeland (38 and 23 kg m−1 year−1), but signifi-
(Table 9.4), which is due to a high rock fragment content cantly larger than the transport rate measured in an exclosure
(RC). (3 kg m−1 year−1). Corresponding rock fragment transport
coefficients (K) for rangeland (32–69 kg m−1 year−1) are evolution. Our observed sandstone cliff retreat rates are of
much larger than K for densely vegetated exclosures the same order of magnitude as long-term sandstone cliff
(3.9 kg m−1 year−1). This indicates that rocks fallen from retreat rates reported from other (sub)tropical regions, indi-
cliffs into rangeland are transported downslope mainly by cating that the anthropogenic impact on the process of rock
livestock trampling. By contrast, rockfall into exclosures is fragment detachment from cliffs is minimal.
largely stored at the upper part of the escarpment face. Rock Yearly unit rock fragment transport rates (Qs), mainly
fragments found in remnant forests on steep slopes are also induced by animal trampling, were 37.9 kg m−1 year−1 in
generally located at the foot of the cliffs. These forests act as rangeland on basalt (slope gradient S = 0.55 m m−1) and
“protection forests” against rockfall, similar to those found 23.1 kg m−1 year−1 in rangeland on sandstone colluvium
in the Alps (Stoffel et al. 2005). (S = 0.72 m m−1). Similar to sheet and rill erosion and
Downslope from cliffs and scree slopes, the overall RC at gullying (Descheemaeker et al. 2006), this process is also
the soil surface on the catena developed on basalt is large virtually stopped after exclosures are established with Qs
(55–85 %) everywhere and is not related to slope gradient equalling only 3.9 kg m−1 year−1 on a 0.85 m m−1 slope.
but to the location of mass movement bodies, as indicated by The importance of rock fragment movement on debris slopes
Nyssen et al. (2002c). In the limestone area, RC is larger on is positively correlated with grazing pressure and the areal
the steeper areas close to the cliff. Due to the combination of percentage of smooth surface, and inversely with the per-
kinetic sieving as a consequence of tillage (Oostwoud et al. centage of long grass cover.
1997) and slow deposition of fine earth at the lower side of Rock fragment redistribution on the lower part of the
the catenas, mean rock fragment size significantly increases catena is controlled by the occurrence of ancient debris flow
with decreasing slope gradient (Nyssen et al. 2002b). bodies as well as current processes such as kinetic sieving by
The vertical processes supplying rock fragments to the tillage, argillipedoturbation, development of erosion pave-
soil surface include (1) tillage-induced kinetic sieving, (2) ments by water erosion, and manual removal by farmers.
selective run-off erosion and the development of erosion Besides rockfall from cliffs and argillipedoturbation, all the
pavements and (3) vertic movements (argillipedoturbation; discussed factors and processes of rockfall redistribution in
Poesen and Lavee 1994; Nyssen et al. 2002a; Moeyersons the study area are of anthropogenic origin (Fig. 9.8).
et al. 2006a).
Acknowledgments This study was conducted in the framework of
Surface covers of large rock fragments sometimes hinder
research programme G006598.N funded by the Fund for Scientific
agriculture, and therefore, farmers remove and concentrate Research—Flanders, Belgium and of the Zala-Daget project (VLIR-
these on stone heaps (zala). However, dense rock fragment UOS, Belgium). JN was affiliated with KU Leuven while the research
covers bring also some advantage to agriculture. They was carried out. Thanks go to Berhanu Gebremedhin Abay for assistance
with all the fieldwork. Local farmers and shepherds shared their
increase infiltration rates, decrease evaporation and protect
knowledge with us. The local Agricultural Office, REST (Relief Society
topsoil against water erosion (Poesen and Lavee 1994). of Tigray) branch, and the authorities of the concerned villages and
Indeed, Nyssen et al. (2001) reported a significant negative district facilitated the research. Conceptual discussions with Andrew
relationship between RC and soil loss in the study area. A Goudie, Robert Wray and David Alexander are gratefully acknowledged.
recommendation resulting from this study is to rely on the
following farmers’ wisdom: smaller rock fragments should
never be removed from the field surface, but a limited References
number of larger rock fragments can be removed in order to
increase crop yield. Adamson D, Williams F (1980) Structural geology, tectonics and the
control of drainage in the Nile basin. In: Williams M, Faure H (eds)
The Sahara and the Nile. Quaternary environments and prehistoric
occupation in northern Africa, Balkema, Rotterdam, pp 225–252
9.5 Conclusions Arkin Y, Beyth M, Dow DB, Levitte D, Temesgen Haile TH (1971)
Geological map of mekele sheet area ND 37-11, Tigre Province,
Rockfall from cliffs and rock fragment transport on debris 1:250.000, Imperial Ethiopian Government, Ministry of Mines,
Geological Survey, Addis Ababa
slopes under rangeland, mainly by livestock trampling, Ayalew L, Yamagishi H (2004) Slope failures in the Blue Nile basin, as
appear to be important geomorphic processes in the northern seen from landscape evolution perspective. Geomorphology
Ethiopian highlands. Along a 1,500 m long section (parallel 57:95–116
to the contour) of the Amba Aradam sandstone cliff, at least Bussert R, Schrank E (2007) Palynological evidences for a latest
carboniferous-early permian glaciation in Northern Ethiopia. J Afr
80 t of rock are detached and fall annually over a mean Earth Sci 49:201–210
vertical distance of 24 m, resulting in a mean annual cliff Cailleux A (1945) Distinction des galets marins et fluviatiles. Bulletin
retreat rate of 0.37 mm year−1. The hard rock of the Amba de la Société géologique de France 15:375–404
Aradam sandstone formation and especially its ferruginous Coltorti M, Dramis F, Ollier CD (2007) Planation surfaces in Northern
Ethiopia. Geomorphology 89:287–296
cap provide structural control on the geomorphological
Descheemaeker K, Nyssen J, Rossi J, Poesen J, Haile M, Moeyer- Nyssen J, Vandenreyken H, Poesen J, Moeyersons J, Deckers J, Haile
sons J, Deckers J (2006) Sediment deposition and pedogenesis in M, Salles C, Govers G (2005) Rainfall erosivity and variability in
exclosures in the Tigray highlands, Ethiopia. Geoderma the Northern Ethiopian highlands. J Hydrol 311:172–187
132:291–314 Nyssen J, Poesen J, Moeyersons J, Deckers J, Haile M (2006)
Driesen P, Dudal R (1991) The major soils of the world. Wageningen Processes and rates of rock fragment displacement on cliffs and
Agricultural University and KU Leuven, The Netherlands scree slopes in an amba landscape, Ethiopia. Geomorphology
Gardner J (1970) Rockfall: a geomorphic process in high mountain 81:265–275
terrain. Albertan Geogr 6:15–20 Nyssen J, Naudts J, De Geyndt K, Haile M, Poesen J, Moeyersons J,
Govers G, Poesen J (1998) Field experiments on the transport of rock Deckers J (2008) Soils and land use in the Tigray highlands
fragments by animal trampling on scree slopes. Geomorphology (Northern Ethiopia). Land Degrad Dev 19:257–274
23:193–203 Oostwoud Wijdenes D, Poesen J, Vandekerckhove L, De Luna E
Hurni H (1985) Erosion—productivity—conservation systems in (1997) Chiselling effects on the vertical distribution of rock
Ethiopia. In: Proceedings of 4th international conference on soil fragments in the tilled layer of a Mediterranean soil. Soil Tillage
conservation, Maracay, Venezuela, pp 654–674, 3–9 Nov 1985 Res 44:55–66
Hutchinson RW, Engels GG (1970) Tectonic significance of regional Oostwoud Wijdenes D, Poesen J, Vandekerckhove L, Kosmas C (2000)
geology and evaporite lithofacies in northeastern Ethiopia. Phil The use of marked rock fragments as tracers to assess rock fragments
Trans Roy Soc Lond A Math Phys Sci 267:313–330 transported by sheep trampling on Lesvos, Greece. In: Foster I (ed)
Lee FT, Odum JK, Lee JD (1994) Rockfalls and debris avalanches Tracers in geomorphology. Wiley, Chichester, pp 201–220
in the Smugglers Notch area, Vermont. US Geol Surv Bull Oostwoud Wijdenes D, Poesen J, Vandekerckhove L, Kosmas C (2001)
2075:1–33 Measurements at one-year interval of rock-fragment fluxes by sheep
Merla G, Minucci E (1938) Missione geologica nel Tigrai. Volume trampling on degraded rocky slopes in the Mediterranean.
primo: la serie dei terreni. Reale Accademia d’Italia, Centro Studi Zeitschrift für Geomorphologie NF 45:477–500
per l’AOI, Roma Perret S, Stoffel M, Kienholz H (2006) Spatial and temporal rockfall
Merla G, Abbate E, Azzaroli A, Bruni P, Canuti P, Fazzuoli M, Sagri activity in a forest stand in the Swiss Prealps—A dendrogeomor-
M, Tacconi P (1979) A geological map of Ethiopia and Somalia phological case study. Geomorphology 74:219–231
(1973) 1:2.000.000 and comment. University of Florence, Italy Poesen J, Lavee H (1994) Rock fragments in top soils: significance and
Moeyersons J, Nyssen J, Poesen J, Deckers J, Haile M (2006a) On the processes. Catena 23:1–28
origin of rock fragment mulches on Vertisols: a case study from the Poesen J, Van Wesemael B, Govers G, Martinez-Fernandez J, Desmet
Ethiopian Highlands. Geomorphology 76(3–4):411–429 P, Vandaele K, Quine T, Degraer G (1997) Patterns of rock
Moeyersons J, Nyssen J, Poesen J, Deckers J, Haile M, Lang A (2006b) fragment cover generated by tillage erosion. Geomorphology
New insights in timing and steering of environmental changes in 18:183–197
Northern Ethiopia since Late-Pleistocene times: a view from the Stoffel M, Schneuwly D, Bollschweiler M, Lièvre I, Delaloye R, Myint
Highlands. Palaeogeogr Palaeoclimatol Palaeoecol 230:165–181 M, Monbaron M (2005) Analyzing rockfall activity (1600–2002) in
Mohr P (1963) The geology of Ethiopia. University College of Addis a protection forest—a case study using dendrogeomorphology.
Ababa Press, 270 p Geomorphology 68:224–241
Nyssen J, Moeyersons J, Deckers J, Haile M, Poesen J (2000) Vertic Williams M, Williams F (1980) Evolution of the Nile basin. In:
movements and the developments of stone covers and gullies, Williams M, Faure H (eds) The Sahara and the Nile. Quaternary
Tigray highlands, Ethiopia. Zeitschrift für Geomorphologie NF environments and prehistoric occupation in Northern Africa,
44:145–164 Balkema, Rotterdam, pp 207–224
Nyssen J, Haile M, Poesen J, Deckers J, Moeyersons J (2001) Removal Yami M, Gebrehiwot K, Moe S, Mekuria W (2007) Impact of area
of rock fragments and its effect on soil loss and crop yield Tigray enclosures on density and diversity of large wild mammals: the case
Ethiopia. Soil Use Manag 171:179–187 of May Ba’ati, Douga Tembien Woreda, Central Tigray, Ethiopia.
Nyssen J, Moeyersons J, Poesen J, Haile M, Deckers J (2002a) East Afr J Sci 1:1–14
Argillipedo-turbation and the development of rock fragment covers Young RA (1985) Geomorphic evolution of the Colorado plateau
on vertisols in the Ethiopian highlands. Belgeo 2:183–193 margin in west-central Arizonia: a tectonic model to distinguish
Nyssen J, Poesen J, Moeyersons J, Lavrysen E, Haile M, Deckers J between the causes of rapid symmetrical scarp retreat and scarp
(2002b) Spatial distribution of rock fragments in cultivated soils in dissection. In: Hack JT, Morisawa M (eds) Tectonic geomorphol-
Northern Ethiopia as affected by lateral and vertical displacement ogy. Allen & Unwin, London, pp 261–278
processes. Geomorphology 43:1–16 Young RW (1987) Sandstone landforms of the tropical East Kimberley
Nyssen J, Moeyersons J, Poesen J, Deckers J, Haile M (2002c) The region, Northwestern Australia. J Geol 95:205–218
environmental significance of the remobilisation of ancient mass Young RW, Wray R (2000) Contribution to the theory of scarpland
movements in the Atbara-Tekeze headwaters, Northern Ethiopia. development from observations in Central Queensland, Australia.
Geomorphology 49:303–322 J Geol 108:705–719
Gully Development in the Tigray Highlands
10
A. Frankl, J. Poesen, J. Moeyersons, and J. Nyssen
Abstract
In the Tigray highlands, gully development is linked to poverty-driven unsustainable use of
the land in a vulnerable semi-arid and mountainous environment, where intense rainfalls
challenge the physical integrity of the landscape. Over the last two centuries, three major
phases in the hydrological regime of the region could be distinguished. In the first phase,
between 1868 (or earlier) and ca. 1965, the relatively stable gully channels showed an
oversized morphology inherited from a previous period when external forcing of environ-
mental conditions caused significant channel development. In the second phase (ca. 1965–ca.
2000), increased aridity and a continued vegetation clearance accelerated dynamics of the
gully system. A sharp increase in gully headcut retreat rates, network densities and volumes
could be quantified for that period. With the widespread implementation of soil and water
conservation measures, erosion rates decreased, which announced the start of the third
hydrogeomorphic phase since ca. 2000. In 2010, about one-fourth of the gully channels were
stabilized. These hydrogeomorphic developments correspond to a gully cut-and-fill cycle in
the second half of the twentieth century and suggest that a pre-1868 cut cycle took place.

Keywords
Gully Headcut retreat Repeat photography Vertisol
10.1 Introduction (Kassas 1995). Furthermore, the resilience of the land is often
reduced by recurring droughts and severe land degradation,
In dryland environments, water availability and biomass which threatens sustainable development in these fragile
production are often restricted and confined to a short rainy environments.
season. As a result, the carrying capacity of the ecosystem is The Tigray Rural Development Study (TRDS; HTS
rapidly exceeded by human exploitation of natural resources, 1976), which investigated the state of the environment in the
especially in sub-Saharan countries like Ethiopia, with fast 1970s, concluded that land degradation was severe in the
demographic expansion and deficient exploitation techniques Tigray highlands and that natural resources were put to their
limits (Virgo and Munro 1978). Almost a decade later, the
devastating effects of drought and desertification were
A. Frankl (&) J. Nyssen
brought to a global audience with the drought that stroke the
Department of Geography, Ghent University, 9000 Ghent, region in 1984–1985. In a context of civil war, region-wide
Belgium crop failures led to massive famine and starvation. As shown
e-mail: amaury.frankl@ugent.be on historical photographs (Fig. 10.1), severe gully erosion
J. Poesen destroyed valuable land, jeopardizing in situ and down-
Department of Earth and Environmental Sciences, KU Leuven, stream agricultural production, increasing the costs of
3001 Heverlee, Belgium
transport and infrastructure construction and producing flash
J. Moeyersons floods of polluted water which threatened human health.
Royal Museum for Central Africa, 3080 Tervuren, Belgium
192 A. Frankl et al.
Fig. 10.1 Active gully channel

networks in a barren landscape of
Tigray highland (Photographs by
Jean Poesen—top, and Piotr
Migoń—middle and bottom)
10 Gully Development in the Tigray Highlands 193
Several studies reported significant gully erosion in grouped into six categories (Schumm 2005): geologic,
Ethiopia and especially addressed the Tigray highlands geomorphic, climatic, hydrologic, animal and human causes.
(Fig. 10.1; Virgo and Munro 1978; Berakhi and Brancaccio Examples that are most relevant for the case of the Tigray
1993; Nyssen et al. 2000, 2002, 2004, 2006; Billi and highlands are uplift, base level lowering, mass movements,
Dramis 2003; Munro et al. 2008; Reubens et al. 2009; Frankl increased aridity or humidity, increased mean run-off dis-
et al. 2011, 2012, 2013b, c, d). Among the first, Virgo and charge and/or peak discharge, decreased sediment load,
Munro (1978) assessed the status of soils and landscapes in animal grazing and tracking, flow diversion, urbanization
northern Ethiopia, documenting numerous cases of severe and infrastructure construction. As the geomorphic devel-
gully erosion. The importance of gully erosion in Vertisols, opment of gullies is related to alternating conditions, they
which are susceptible to soil piping, was discussed by are also referred to as non-regime channels (Schumm 2005).
Nyssen et al. (2000). Berakhi and Brancaccio (1993) and In order to efficiently transfer water and sediment via run-
Nyssen et al. (2006) studied the effect of road building on off downslope, the shape and size of gullies mainly adjust to
gully erosion risk, and Nyssen et al. (2004) investigated the peak flow discharge properties (Knighton 1998). Channel
usefulness of check dams to control gullies. Nyssen et al. shape and size may be spatially and temporally very vari-
(2006) analysed the development of four gully systems in the able, depending on local environmental characteristics (e.g.
Tigray highlands by developing a field method which is soil, vegetation), and the time that is needed to accommodate
based on how local people remember the evolution of spe- changes in run-off discharge and sediment load, i.e. reaction
cific gullies. A similar approach was used by Moges and and relaxation times (Knighton 1998).
Holden (2008) who studied gully development in southern Frankl et al. (2013c) characterized gully cross-sectional
Ethiopia. A small gully network in eastern Ethiopia was characteristics in the Tigray highlands by surveying 811
studied by Daba et al. (2003), using a time series of digital gully cross sections in various environments, reflecting
elevation models derived from small-scale aerial photo- contrasts in lithology, topography, soil, land use and climate.
graphs. An analysis of the development of gully headcuts, Gully top width (TW) varied between 0.35 and 31.90 m,
their cross sections, networks and volumes at regional scale with a median of 6.34 m. Gully depth (D) varied between
in Tigray highlands was carried out by Frankl et al. (2011, 0.20 and 12.77 m, with a median of 2.15 m, and bottom
2012, 2013b, c, d). By using repeated terrestrial photogra- width (BW) ranged between 0.10 and 19.50 m, with a
phy, aerial photographs and satellite images, these authors median of 3.00 m. The median cross-sectional area (CSA)
could define distinct phases in gully erosion development was 10.1 m2 and ranged between 0.15 and 236.5 m2. As the
since the late nineteenth century. Nigussie Haregeweyn et al. boxplots show (Fig. 10.2a), the distributions are right-
(2005, 2008) linked the presence of gullies to the rate of skewed and the variability of the observations, as indicated
reservoir sedimentation (catchment sediment yield). by interquartile range, is higher for TW (5.20) and BW (2.70)
than for D (1.79). The median TW–D ratio was 2.7, while the
median BW–TW ratio was 0.5. Note that for TW/D and BW/
10.2 Gullies and Gully Erosion TW, median and mean do not differ much as the distributions
are nearly normal. As shown in Fig. 10.2b, plotting D over
In the Oxford Dictionary of Earth Sciences (2008), a gully is TW shows wide scatter around a linear relation purged
defined as a “feature of water erosion that develops from the through the origin (0, 0).
run-off of a violent torrent that bites deeply into topsoil and Catchment area is the most important factor explaining
soft sediments”. Gully erosion can thus be associated with the variability in TW, D and CSA, as it reflects the volume
the rapid incision of valley sides or valley floors by the of water that drains to a certain point in the landscape.
erosive action of flash floods. This is also apparent from the However, the large variability in soil management and
definition of Poesen et al. (2003) who proposed that gully environmental characteristics causes the gully shape and
erosion is an “erosion process whereby runoff water accu- size to be locally very variable. As shown by Frankl et al.
mulates and often recurs in narrow channels and, over short (2013c), besides the catchment area, lithology and the
periods, removes the soil from this narrow area to consid- presence of check dams are the most important explanatory
erable depths”, a definition that is widely used in scientific factors to account for gully morphology in the Tigray
literature. Gully erosion is thus the result of disruptions in a highlands. Gully cross sections in shale-derived deposits are
previous stage of equilibrium that cause changes in the run- on average 36.7 % larger than those in deposits derived
off volumes and sediment delivered to a certain place in the from volcanics. This is mainly related to the higher erod-
landscape (Graf 1988; Knighton 1998). Causes of incision ibility of shale-derived materials and to the presence of
by concentrated overland flow are numerous and can be travertine dams. Degradation and breaching of the latter
Fig. 10.2 Cross-sectional characteristics of gullies in Tigray high- sections. Outliers larger than 20 m and 100 m2 are not displayed,
lands (Frankl et al. 2013c). a Boxplots for gully top width (TW), b plotting gully depth (D) over gully top width (TW) shows a linear
bottom width (BW), depth (D) and cross-sectional area (CSA) for 811 relation. Based on Frankl et al. (2013c)
Fig. 10.3 The deeply incised travertine dam at May Mekdan and gully network upslope (Photograph of 2010 by Amaury Frankl; 13.58°N, 39.56°E)
(Moeyersons et al. 2006; Fig. 10.3) triggers regressive

erosion and exposure of thick, fine-textured soils locked 10.3 Evidence of Past Dynamics
upslope of the dams. Check dams, which are successfully
implemented in low-active gully sections, cause the gully 10.3.1 Early Evidence of Gullying
depth to decrease by circa one-third. This makes moderately
active gully cross sections with check dams on average The oldest terrestrial photographs known for the highlands of
33.5 % smaller than those of very active gullies without Tigray are those taken during the British military expedition
check dams. to former Abyssinia in 1867–1868, with the objective to
Fig. 10.4 The gully draining the valley in 1939 (left) is low-active By 2009, the gully had extended downstream by 294 m (right). It is
(smooth cross section, vegetated) and has been partially converted to freshly incised (=high-active section) and remains active despite the
cropland (zoom). Although wood harvesting and agricultural exploi- recent soil and water conservation efforts. Original photograph (left):
tation in 1939 severely affected the environment, shrub cover on the Maugini (Istituto Agronomico per l’Oltremare, Firenze, I). Repeat
arable land (white arrow) and on the surrounding hills (black arrow) of photograph by Amaury Frankl 12.56°N, 39.25°E. Based on Frankl
this small catchment draining to Lake Ashenge is still relatively high. et al. (2011)
release a number of Europeans who had been imprisoned by In 1963–1965, gully drainage density (Dtotal) and area-
Emperor Tewodros (Sharf et al. 2003). Such historical pho- specific gully volume (Va) were still relatively low, i.e.
tographs are often the only records available on the envi- 1.86 km km−2 and 32.23 × 103 m3 km−2, respectively
ronmental status for the period 1868–1963 (Fig. 10.4), after (Fig. 10.5a–c). 48 % of the gully network was high-active.
which aerial photographs were produced for the area (Nyssen From a largely low-dynamic gully system in the 1960s,
et al. 2010). Aerial photographs taken during the Italian network expansion and increased erosion rates in the 1980s
occupation of Ethiopia in 1935–1941 have recently been and 1990s caused the drainage density and volume to peak in
discovered but were not yet analysed. In order to compare the 1994. Dtotal was then 2.52 km km−2 and Va 60 × 103 m3 km−2,
previous and current situations, historical terrestrial photo- with 93 % of the gully network being highly active
graphs were repeated in the field according to the technique (Fig. 10.5a–c). This corresponds to a soil losses by gully
of repeat photography by Hall (2001). Qualitative and erosion (SLg) of 17.6 ton ha−1 year−1 over the period 1963–
quantitative analyses of time-lapsed photographs revealed 1994. The terrestrial photographs from that period show
that in the late nineteenth and early twentieth centuries, gullies which were very active, having clear-cut walls and
gullies were common features of the northern Ethiopian transporting considerable amounts of debris (Fig. 10.1).
landscape (Frankl et al. 2011). However, most gullies seen on The average incision rate of gully cross sections was
historical photographs from that period show smooth and 0.04 m year−1, and maximum incision rates of 0.13 m year−1
vegetated cross sections (Fig. 10.4). As their size and mor- were observed in Vertisol areas (Frankl et al. 2011). At the
phology suggest, the gullies were not in equilibrium with the upper gully margins, headcuts incised upslope, while at the
prevailing conditions but were rather inherited from a pre- lower ends, debris fans were deposited and incised subse-
vious period when external forcing of environmental condi- quently. Long- to medium-term linear headcut retreat rates
tions caused significant geomorphic change (Nyssen et al. (Rl) were on average 3.8 ± 4.7 m year−1 (Frankl et al. 2012).
2009; Frankl et al. 2011). Most probably, such ancient gully This gully cut cycle, that started in the mid-1960s, lasted until
incision cycles occurred during the frequent droughts of the ca. 2000, after which a new gully fill cycle was initiated.
nineteenth century, as reported by Pankhurst (1995), leading
to famine and great mortality. These calamities would have
increased the environmental vulnerability to human inter- 10.4 Rehabilitation in the Twenty
ference and, hence, also the run-off response of the land. First Century?
10.3.1.1 1960s–1990s Cut Cycle As a result of huge efforts in environmental rehabilitation

The study of time-lapsed terrestrial photographs and aerial undertaken since the 1970s, denudation rates decreased in
photographs indicates that the recent gully incision phase in the highlands of Tigray—although remained fairly high in
Tigray highlands started around 1965, with a marked increase absolute terms—reducing the importance of gully erosion.
in gully drainage densities and volumes (Frankl et al. 2013b). Mean run-off discharge as well as sediment load and flash
Fig. 10.5 Gully cut-and-fill

cycle during the period 1963–
2010 as expressed by trends in
total drainage density (Dtotal) (a),
drainage density of the high-
active gullies (Dhigh-active) (b) and
area-specific volume development
(Va) (c). Based on Frankl et al.
(2013b)
flood peaks decreased (cf. Moeyersons 1989, 1990), causing However, strong degradation can also be the result of local
gullies to become less active and to fill in partially, espe- factors, such as different slope/channel coupling, and con-
cially when check dams were installed (Frankl et al. 2013b; nectivity or catchment scale, with large catchments having
Fig. 10.6a). When proper land management was applied, longer reaction times than small catchments. The reduced
gullies could even be transformed into a green linear oasis gully erosion rate is also apparent from the headcut retreat
which contributes to the ecological restoration of the rates studied by Frankl et al. (2012). Present-day linear
degraded area. Moreover, their resilience against the effects headcut retreat rates (Rl) are much smaller than those typical
of drought or land use changes on run-off response of the for the medium to longterm, with an average Rl of
land increased. Lower ends of gully channels became non- 0.34 ± 0.49 m year−1. However, gullying in Vertisols
active and even migrated upslope. remains very active as gully development is largely con-
This gully fill cycle is well evidenced by the analysis of trolled by soil piping. In Vertisols, present-day headcut
aerial photographs and more recent satellite images (Frankl retreat rates up to 1.93 m year−1 were recorded and large
et al. 2013b). This analysis indicates that average Dtotal and gabion (wire net structures, filled with rocks) check dams,
Va decreased to 2.20 km km−2 and 48.96 × 103 m3 km−2, which are cost- and labour-intensive, were sometimes
respectively (Fig. 10.5a–c). Even more important is the bypassed in one rainy season, forcing the gully to expand
decrease in average Dhigh-active to 1.65 km km−2, indicating laterally into the adjacent land (Fig. 10.6b).
that 25 % of the gully network had become low-active.
These findings are also supported by repeat-photography
studies, indicating that in 2006–2009, about 23 % (n = 8) of 10.5 Factors Controlling Gully
the channel cross sections were stabilized and had cross- Erosion
sectional characteristics similar to those of the pre-1963
gullies (Frankl et al. 2011). Among the other 31 gullies and 10.5.1 Hydrogeomorphic Phases
river sections studied from repeat photography, 44 % were and the Role of Land Use
highly active, whereas 23 % were in a transitional stage. and Precipitation
Strong channel degradation can be the result of a clear-water
effect, which causes the gullies in the valley bottoms to From ca. 1868 to 1965, in a first hydrogeomorphic phase,
incise. This can be facilitated by the presence of thick, gullies were mostly low-active, displaying smooth (vege-
erodible alluvio-colluvial deposits in valley bottom position. tated) cross sections. This indicates that environmental
area covered by bare soil was extensive and that the area
covered by cropland peaked. From the analysis of land use
and land cover on old terrestrial photographs, Meire et al.
(2013) also indicated a minimum in vegetation cover in the
period 1940s–1990s. Frankl et al. (2013a) showed that the
length of the crop growth period decreases with increasing
drought in the Tigray highlands, making croplands very
vulnerable to high-intensity rainfall during the summer rainy
season.
Since ca. 2000, the large-scale implementation of soil and
water conservation measures started to yield positive effects
on the environmental rehabilitation and on the stabilization
of gullies. Several studies indeed indicate that vegetation
cover and land management strongly improved in recent
decades (e.g. Gebremedhin et al. 2004; Munro et al. 2008;
Alemayehu et al. 2009; Mekuria et al. 2009; Nyssen et al.
2009; de Mûelenaere et al. 2013; Meire et al. 2013). As a
reaction to severe land degradation that stroke northern
Ethiopia in the 1970s and 1980s, environmental rehabilita-
tion programs were launched with the aim of increasing land
resilience to the effects of droughts. Biophysical conserva-
tion measures that were implemented include the following:
(1) the establishment of exclosures in critical steep-sloped
zones (Descheemaeker et al. 2006); (2) the introduction of
stone bunds (Nyssen et al. 2008) and soil trenches; and (3)
the construction of check dams in gullies (Fig. 10.6a; Nyssen
et al. 2004). At present, exclosures cover 10–15 % of the
land surface and stone bunds are found at an average density
of 57 km km−2 (Schumacher 2012). This led to greening of
the landscape in which the surface covered by bushland,
Fig. 10.6 Effectiveness of check dams as a measure to control gully forest or Eucalyptus plantation strongly increased. This
erosion. a Siltation behind gabion check dams caused the gully channel greening is partly the result of the introduction of Eucalyptus
to fill in by approximately one-third of its depth in a catchment where
slope run-off response decreased thanks to the implementation of soil trees to support the growing need of construction wood in
and water conservation measures. Notice also on the left gully bank soil decades where population is remarkably increasing. At a
pits that are dug to plant trees, b at another location, the presence of a national level, population size almost doubled, from 40
Vertisol lens in the lower soil profile (white arrow) caused the gabion
million in 1980 (Maddison 2006) to 88.4 million in 2013
check dam to be bypassed in one rainy season, forcing the gully flow to
erode the adjacent land. Photographs taken in 2009 (a) and 2011 (b) by (FAOSTAT 2013). In Tigray, population size increased from
Amaury Frankl 13.65°N, 39.21°E 3.1 in 1994 to 4.3 million in 2007, representing 6 % of the
total Ethiopian population (CSA 2008), whereas population
density increased from 63 to 86 persons km−2.
vulnerability did not yet reach a critical point for large-scale
channel expansion and degradation to occur. After 1965, a
marked transition from low- to high-active gullies took place 10.5.2 Check Dams as Soil and Water
in a second hydrogeomorphic phase. This is most probably Conservation
related to arid pulses that occurred in the 1970s and 1980s. A
similar phenomenon was observed in Senegal (Poesen et al. In addition to soil and water conservation measures imple-
2003). Such phases alter biomass production and increase mented in the gully catchments (e.g. stone bunds, soil
human pressure on land and vegetation. In order to secure trenches), the construction of loose rock or gabion check
food production, farmers were forced to cultivate steeper dams in gullies is a widely used conservation measure in the
land and overgrazing removed most of vegetation from the highlands of Tigray. Technically, the aim is to transform the
hillslopes. Analyses of region-wide land use and cover on long, uniform and steep longitudinal profile of the gully bed
the basis of Landsat imagery by de Mûelenaere et al. (2013) into successive, nearly horizontal steps (Ayres and Scoates
in the 1970s and 1980s confirmed that in 1984/1986, the 1939; Hurni 1986) by trapping sediment behind the dams.
The success of their implementation relies on the availability cracks, which can be as much as two metres deep. Run-off
of loose rocks in the immediate surroundings and the free water subsequently infiltrates into the subsoil (bypass flow)
labour programs, making such dams installation community- and drains underground. Intense subsurface erosion of the
based and cost-effective (Gebremedhin and Swinton 2003). dry, dispersive clays results in the development of soil pipes,
However, check dams commonly collapse. As indicated by which, once collapsed, may turn into gullies (Fig. 10.6b).
Nyssen et al. (2004), 39 % of loose rock dams collapse after When a Vertisol gets wet, swelling causes cracks to close
two years which is a phenomenon strongly linked to the and the Vertisol becomes almost impermeable. Conse-
drainage area and slope gradient of the soil surface next to quently, run-off production is very high, and large run-off
the gully, the product of these factors being a measure of volumes drain through pipes to the gully heads (Nyssen et al.
run-off energy. In Vertisol areas, where soil piping occurs, 2000; Frankl et al. 2012). Reducing gully expansion in
flow bypassing of check dams is commonly observed and Vertisols calls for specific measures to reduce the rates of
can even result in the collapse of large gabion check dams soil piping. As proposed by Frankl et al. (2012), introducing
during a single rainy season (Nyssen et al. 2004; Frankl et al. a subsurface geomembrane dam at gully heads can increase
2012; Fig. 10.6b). It is therefore very important that the the water storage upslope of the dam and thus reduce soil
technical instructions (BoANR 1997; Hurni 1986; Nyssen piping. This has been successfully demonstrated at May
et al. 2004) are followed when implementing new gully Ba’ati village (13.65°N, 39.21°E).
control structures and that maintenance is organized regu-
larly. Moreover, Vertisol areas with piping require specific
measures to prevent flow bypassing. More research is also 10.6 Conclusions
necessary on the extent to which check dams delay run-off
response and enhance infiltration. Fast land degradation may occur when improper land man-
agement is applied. Most dramatic is the development of
extensive and deep gully networks, which in the highlands of
10.5.3 Direct Human Intervention Tigray produce large volumes of sediment that are transported
through the gully and (ephemeral) river systems since the
By engineering the landscape through road and drainage 1960s. However, local communities have proven that this
canal construction, humans may cause important modifica- trend can be reversed. At a regional scale, since ca. 2000, gully
tions to natural run-off pathways. As shown by Nyssen et al. networks are increasingly being stabilized and the landscape is
(2002), a 6.5-km-long new road segment resulted in the re-greening. These results have to be understood within a
development of 16 new gullies (total volume of 10,034 m3) socio-economic context of strong population growth and a
and the stabilization of five small gullies (with a total volume low-level technological development, where most people rely
of 100 m3). In these newly developed gullies, linear headcut on land resources for their livelihood and where the fragility of
retreat rates (Rl) were up to 10 times greater than those the country’s economy is frequently emphasized, for example
recorded in the gullies with no changes in their catchment when climatic shocks such as drought cause severe food
area, with average Rl for the period 1994–2010 being shortages and famine. Socio-economic developments and
21.3 m year−1. Distances between culverts are commonly their relation to land degradation should therefore be moni-
large, and therefore, sites that received run-off from a rela- tored closely. With a population size which is likely to double
tively small drainage area before road construction may by 2050, Ethiopia faces immense challenges. The key is to
experience important increases in run-off because of the rehabilitate the land, as a resource base for food security and
catchment area increase. In order to avoid new gully heads ecosystem services, and to strengthen and diversify the rural
developing after building mountain roads, appropriate economy in order to make local communities less dependent
engineering works (e.g. flow energy dissipaters and/or on land resources. Such challenges are embraced by many
splitters) should be undertaken during road construction. local, national and international programmes and should
remain high on the agenda.
10.5.4 Solutions for Vertisols References

Soil piping is recognized as an important reason for the
Alemayehu F, Taha N, Nyssen J, Girma A, Zenebe A, Behailu M,
development of gullies (Valentin et al. 2005). In Vertisols, Deckers J, Poesen J (2009) The impacts of watershed management
this is related to their periodic shrinking and swelling. In a on land use and land cover dynamics in Eastern Tigray (Ethiopia).
dry Vertisol, shrinking results in the development of wide Resour Conserv Recy 53:192–198
Ayres Q, Scoates D (1939) Land drainage and reclamation. McGraw- Maddison A (2006) The world economy. OECD, Development Centre
Hill, New York Studies, Paris
Billi P, Dramis F (2003) Geomorphological investigation on gully Meire E, Frankl A, De Wulf A, Haile M, Decker J, Nyssen J (2013)
erosion in the Rift Valley and the Northern highlands of Ethiopia. Land use and cover dynamics in Africa since the nineteenth century:
Catena 50:353–368 warped terrestrial photographs of North Ethiopia. Reg Environ
BoANR (1997) Water and soil conservation, forestry development. Change 13:717–737
Manual for agricultural cadres. Lekatit 1989 (Eth. Cal.) (in Mekuria W, Veldkamp E, Haile M, Gebrehiwot K, Muys B, Nyssen J
Tigrinya). Bureau of Agriculture and Natural Resources, Mekelle (2009) Effectiveness of exclosures to control soil erosion and local
CSA (2008) Summary and statistical report of the 2007 population and community perception on soil erosion in Tigray, Ethiopia. Afr J
housing census. Population size by age and sex. Central Statistical Agric Res 4:365–377
Agency of Ethiopia, Addis Abbaba Moeyersons J (1989) La nature de l’érosion des versants au Rwanda.
Daba S, Rieger W, Strauss P (2003) Assessment of gully erosion in eastern Annales, Kon. Mus. Mid. Afr., Tervuren 19, Tervuren
Ethiopia using photogammetric techniques. Catena 50:273–291 Moeyersons J (1990) Soil loss by rainwash: a case study from Rwanda.
de Mûelenaere S, Frankl A, Haile M, Poesen J, Deckers J, Munro RN, Z Geomorph 34:385–408
Veraverbeke S, Nyssen J (2013) Historical landscape photographs Moeyersons J, Nyssen J, Poesen J, Deckers J, Haile M (2006) Age and
for calibration of LANDSAT land use/cover in the Northern backfill/overfill stratigraphy of two tufa dams, Tigray highlands,
Ethiopian Highlands. Land Degrad Dev 25:319–335 Ethiopia, evidence for Late Pleistocene and Holocene wet condi-
Descheemaeker K, Nyssen J, Rossi J, Poesen J, Haile M, Moeyersons J, tions. Palaeogeogr Palaeoclimatol Palaeoecol 230:165–181
Deckers J (2006) Sediment deposition and pedogenesis in exclo- Moges A, Holden NM (2008) Estimating the rate and consequences of
sures in the Tigray highlands, Ethiopia. Geoderma 132:291–314 gully development, a case study of Umbulo catchment in southern
FAOSTAT (2013) Statistical division of the UN food and agriculture Ethiopia. Land Degrad Dev 19:574–586
organization. http://faostat.fao.org/ Munro RN, Deckers J, Grove AT, Haile M, Poesen J, Nyssen J (2008)
Frankl A, Nyssen J, De Dapper M, Haile M, Billi P, Munro RN, Soil and erosion features of the Central Plateau region of Tigrai—
Deckers J, Poesen J (2011) Linking long-term gully and river learning from photo monitoring with 30 years interval. Catena
channel dynamics to environmental change using repeat photogra- 75:55–64
phy (Northern Ethiopia). Geomorphology 129:238–251 Haregeweyn Nigussie, Poesen J, Nyssen J, Verstraeten G, de Vente J,
Frankl A, Poesen J, Deckers J, Haile M, Nyssen J (2012) Gully head Govers G, Deckers S, Moeyersons J (2005) Specific sediment yield
retreat rates in the semiarid highlands of North Ethiopia. Geomor- in Tigray-Northern Ethiopia: assessment and semi-quantitative
phology 173–174:185–195 modelling. Geomorphology 69:315–331
Frankl A, Jacob M, Haile M, Poesen J, Deckers J, Nyssen J (2013a) Haregeweyn Nigussie, Poesen J, Nyssen J, Govers G, Verstraeten G, de
The effect of rainfall on the spatio-temporal variability of cropping Vente J, Deckers J, Moeyersons J, Haile M (2008) Sediment yield
systems and duration of the crop cover in the Northern Ethiopian variability in Northern Ethiopia: a quantitative analysis of its
highlands. Soil Use Manag 29:374–383 controlling factors. Catena 75(1):65–76
Frankl A, Poesen J, Haile M, Deckers J, Nyssen J (2013b) Quantifying long- Nyssen J, Moeyersons J, Deckers J, Haile M, Poesen J (2000) Vertic
term changes in gully networks and volumes in dryland environments: movements and the development of stone covers and gullies, Tigray
the case of Northern Ethiopia. Geomorphology 201:254–263 Highlands. Ethiopia Z Geomorph 44:145–164
Frankl A, Poesen J, Scholiers N, Jacob M, Haile M, Deckers J, Nyssen Nyssen J, Poesen J, Luyten E, Veyret-Picot M, Deckers J, Haile M,
J (2013c) Factors controlling the morphology and volume (V)— Govers G (2002) Impact of road building on gully erosion risk: a
length (L) relations of permanent gullies in the Northern Ethiopian case study from the Northern Ethiopian highlands. Earth Surf
highlands. Earth Surf Process Land 38:1672–1684 Process Land 27:1267–1283
Frankl A, Zwertvaegher A, Poesen J, Nyssen J (2013d) Transferring Nyssen J, Veyret-Picot M, Poesen J, Moeyersons J, Haile M, Deckers
google earth observations to GIS-software: example from gully J, Govers G (2004) The effectiveness of loose rock check dams for
erosion study. Int J Digital Earth 6:196–201 gully control in Tigray, Northern Ethiopia. Soil Use Manage
Gebremedhin B, Swinton SM (2003) Investment in soil conservation in 20:55–64
northern Ethiopia: the role of land tenure security and public Nyssen J, Poesen J, Veyret-Picot M, Moeyersons J, Haile M, Deckers J,
programs. Agric Econ 29:69–84 Dewit J, Naudts J, Teka K, Govers G (2006) Assessment of gully
Gebremedhin B, Pender J, Tesfay G (2004) Collective action for erosion rates through interviews and measurements, a case study
grazing land management in crop-livestock mixed systems in the from Northern Ethiopia. Earth Surf Process Land 31:167–185
highlands of northern Ethiopia. Agricul Syst 82:273–290 Nyssen J, Poesen J, Moeyersons J, Haile M, Deckers J (2008)
Hall F (2001) Ground-based photographic monitoring. Gen Tech Rep Dynamics of soil erosion rates and controlling factors in the
PNW-GTR-503. U.S. Department of Agriculture, Forest Service, Northern Ethiopian highlands—towards a sediment budget. Earth
Pacific Northwest Research Station, Portland Surf Process Land 33:695–711
HTS (1976) Tigrai Rural Development Study (TRDS). Hunting Tech- Nyssen J, Haile M, Naudts J, Munro N, Poesen J, Moeyersons J, Frankl
nical Services Ltd. Government of Ethiopia and UK Ministry of A, Deckers J, Pankhurst R (2009) Desertification? Northern
Overseas Development. Hunting Technical Services, Borehamwood Ethiopia re-photographed after 140 years. Sc Total Environ
Hurni H (1986) Guidelines for development agents on soil conservation 407:2749–2755
in Ethiopia. Soil conservation research project. Community Forests Nyssen J, Frankl A, Munro RN, Billi P, Haile M (2010) Digital
and Soil Conservation Development Department, Ministry of photographic archives for environmental and historical studies: an
Agriculture, Addis Ababa Ethiopia example from Ethiopia. Scott Geogr J 126:185–207
Graf WL (1988) Fluvial processes in dryland environments. The Berakhi Ogbaghebriel, Brancaccio L (1993) Some reflections on the
Blackburn Press, Caldwell origin and land use of pediments on the Ethiopian highlands.
Kassas M (1995) Desertification: a general review. J Arid Environ Geografia Fisica e Dinamica del Quaternario 16:101–106
30:115–128 Pankhurst R (1995) The history of deforestation and afforestation in
Knighton D (1998) Fluvial forms and processes—a new perspective. Ethiopia prior to world war I. Northeast Afr Stud 2:33–119
Hodder Education, London
Poesen J, Nachtergaele J, Verstraeten G (2003) Gully erosion and Author Biographies

environmental change, importance and research needs. Catena
Amaury Frankl Amaury Frankl (1982, Ghent) specialized in hydrogeo-
50:91–133 morphology at the Department of Geography of Ghent University (Bel-
Reubens B, Poesen J, Nyssen J, Leduc Y, Zenebe A, Tewoldeberhan S, gium). His PhD. research was on gully development and its spatiotemporal
Bauer H, Gebrehiwot K, Deckers J, Muys B (2009) Establishment variability since the late nineteenth century in the highlands of Tigray.
and management of woody seedlings in gullies in a semi-arid
environment (Tigray, Ethiopia). Plant Soil 324:131–156 Jean Poesen (1954, Fataki, D.R.C.), a full professor and head of the
Division of Geography at KU Leuven, specialized in physical geography
Schumacher M (2012) Recent trends in gully erosion as evidenced by
with a focus on soil erosion, desertification, soil and water conservation. He
repeat photography around Hagere Selam (Northern Ethiopia).
initiated and collaborated in several research projects in Ethiopia.
Unpublished Master Thesis. Institute for Geography, Technical
University of Dresden, Dresden Jan Moeyersons (1947, Willebroek) is an expert on gullying and landslides
Schumm SA (2005) River variability and complexity. Cambridge in tropical Africa and worked from many years at the Royal Museum for
University Press, Cambridge Central Africa (Belgium).
Sharf FA, Northrup D, Pankhurst R (2003) Abyssinia, 1867–1868: Jan Nyssen (1957, Sint-Martens-Voeren) has been based for many years at
artists on campaign: watercolors and drawings from the British Mekelle University (Ethiopia) where he was research coordinator for vari-
expedition under Sir Robert Napier. Tsehai Publishers, Addis ous university cooperation programmes between Belgium and Ethiopia.
Ababa Nowadays, he is a full professor at the Department of Geography of Ghent
Valentin C, Poesen J, Li Y (2005) Gully erosion, impacts, factors and University (Belgium) where he leads the Physical Geography Research
control. Catena 63:132–153 Group. He initiated and collaborated in several research projects in Ethiopia.
Virgo KJ, Munro RN (1978) Soil and erosion features of the Central
Plateau region of Tigrai, Ethiopia. Geoderma 20:131–157
Tufa Dams in Tigray (Northern Ethiopia)
as Late Pleistocene—Holocene Climate Proxies 11
Francesco Dramis and Giandomenico Fubelli
Abstract
The geomorphological–stratigraphic study of the backfill deposits of two tufa dams, Mai
Makden and Tsabati Mariam, supported by 14C dates, has provided a detailed description of
the environmental changes occurred during the Holocene. In particular, the evolution of the
investigated tufa dams points out the occurrence of some century-scale stages of tufa
deposition (10.9–9.3; 8.4–7.4; 6.6–5.5 kyrs BP) interrupted by intervals characterized by
lower or absent deposition and dam incision (9.3–8.4; 7.4–6.6 kyrs BP). Since 5.5–
2.5 kyrs BP, the deposition rates of tufa progressively declined until stream erosion incised the
dams down to the underlying bedrock. The sequence of tufa events in Tigray seems to be
parallel to the record of lake-level fluctuations in the Ethiopian Rift with the high stands
corresponding to the main deposition stages of tufa and the low stands corresponding to the
non-deposition/erosion intervals. Quite interesting, in this context, is the progressive lowering
of the lake levels after ca. 5.5 kyrs cal BP. The tufa record of Tigray also shows some
correlation with the cooling–warming stages of the Mediterranean Sea and, more in general,
with the main peaks of the Holocene global temperatures. The above relationships emphasize
the primary role of rainfall and air temperature as factors controlling the deposition of tufa.
This does not exclude human impact as an additional factor for the deposition of tufa.

Keywords
Tufa Holocene Paleoclimate Ethiopia East Africa
11.1 Introduction There is general agreement in referring the aggradation/

degradation phases of tufa dams during geological times to
Remains of tufa dams of different ages are commonly found climate controls. Warm/wet climates are believed to favour
across rivers in limestone areas (Ford and Pedley 1996). tufa aggradation because of: (1) higher concentration of
Behind such natural dams, swampy–lacustrine sedimentary biogenic CO2 in soil layers, resulting in higher rates of
sequences, consisting of clayey deposits with peaty levels limestone dissolution (Atkinson 1977; Brook et al. 1983);
and alternations of phytoclastic travertine, travertine sand (2) higher air temperature at the springs favouring water out-
layers and buried soils, may be locally observed (Pentecost gassing (Gullentops and Mullenders 1972); (3) faster
2005). The growth of tufa dams occurs where the deposition development of aquatic plants and related absorption of CO2
rate of calcium carbonate from flowing water is high enough for photosynthesis (Pedley 1990). Conversely, cold/dry cli-
to balance the stream flow erosion (Fubelli et al. 2013). mates are considered to be less favourable for CaCO3 pre-
cipitation because of reduced biological activity of soils,
F. Dramis (&) G. Fubelli

Department of Sciences, Roma Tre University, Rome, Italy
e-mail: dramis@uniroma3.it
202 F. Dramis and G. Fubelli
lower air temperatures at the springs and lesser development number of tufa dam remnants (Fig. 11.1), the age of which
of aquatic plants (Pentecost 2005). ranges from Middle Pleistocene to Holocene (Chernet and
An additional model to explain the increase/decrease of Eshete 1982; Moeyersons et al. 2006).
tufa deposition rates (Dramis et al. 1999) makes reference to Some of these dams have a spectacular appearance from
variations of thermal gradient between the ground surface the geomorphological point of view. This is the case of the
and the deep limestone aquifers induced by climate changes. Romanat dam (Fig. 11.2), dated at the Late Pleistocene
Because of the extremely low thermal conductivity of the (Moeyersons et al. 2006), and the Mai Makden dam
rocks (Vasseur et al. 1983) and the related slow penetration (Fig. 11.3) whose backfill has been the subject of study in
of thermal changes into the ground, differences of temper- this chapter. The Romanat dam (called Romanoff waterfall
ature up to several degrees and reversed thermal gradients in Google Earth) shows an impressive scenery characterized
may be produced between the ground surface and the by a 50-m-high tufa wall overlying an extensively jointed
underlying bedrock over timescales ranging from years to dolerite dyke and a great pool with several potholes in the
thousands of years in relation to thermal change magnitude Antalo Limestone (Mohr 1962). The jump, due the different
and the aquifer depth (Williams and Smith 1989). With erodibility between the two lithologies favoured the forma-
climatic changes to warmer conditions, water percolating tion of the tufa dam at this point. The Mai Makden dam
through progressively colder layers in the vadose belt as well (Fig. 11.4) is built up over horizontally layered limestone
as in the phreatic zone undergoes progressive enrichment in (Antalo Limestone; Mohr 1962) and consists of at least three
dissolved CaCO3 (Thrailkill 1968; Atkinson 1977). At the different units of phytohermal travertine (Pedley 1990)
emergence, because of higher surface temperatures, the separated by coarse gravels and weathered material, attain-
spring water becomes over-saturated with CaCO3, thus ing a total height of ca. 20 m. It is presently incised by the
inducing precipitation of tufa at waterfalls, knick points or Mai Makden River that forms a narrow gorge. Below the
on the river bed itself. Tufa aggradation may continue for a dam, there are slots and potholes carved in the limestone
long time, even if with progressively lower deposition rates, bedrock by the waterfall that exists downstream from the
till the exhaustion of the thermal disturbance in the ground. dam incision. The backfill extends upstream for more than
Opposite effects, such as deposition of carbonate in the 1 km as far as Mai Makden Village where another dam is
upper bedrock and emergence of spring waters under-satu- present.
rated with CaCO3, should be expected with the onset of a The present climate of the area, located at the extreme
climatic change to cold conditions. northern border of the ITCZ migration, at a mean elevation of
Taking into account the dropping of CO2 concentrations more than 2,000 m a.s.l., is characterized by annual air tem-
in soils due to man-made deforestation and other anthropo- peratures ranging between 15 and 25 °C and an annual rainfall
genic effects, such as changes in stream hydrology and ranging from 700 to over 1,200 mm, mostly concentrated in
increase of water turbidity, several authors (e.g. Nicod 1986; the summer (Griffiths 1962; CHP 4, this volume). The impact
Goudie et al. 1993) have stressed the effects of human of human activity on natural vegetation has been particularly
impact to explain the Late Holocene decline of tufa depo- effective in the area since the second millennium BC (Butzer
sition observed in several localities of Europe and North 1981; Fattovich 1990; Bard et al 2000). Though the present
Africa. In any case, the aggradation of tufa dams in the climatic conditions would allow Juniperus and Olea forest and
paleoclimatic record indicates a climatic regime character- mixed deciduous Juniperus woodland (Ethiopian Mapping
ized by warm–wet conditions with dense vegetation and Authority 1988), arboreal vegetation is extremely reduced,
well-developed soils. Higher rates should be expected in exception made for a few trees clustered around churches and
connection with cold to warm climate changes. On the monasteries and in the less accessible areas.
contrary, shifting of climate towards drier or cooler condi- Two stratigraphic sequences of tufa-dammed swampy–
tions would induce decreasing deposition rates and the lacustrine–alluvial deposits were investigated in Tigray
incision of dams by streams. At a more local scale, the same (Brancaccio et al. 1997; Berakhi et al. 1998; Dramis et al.
effects may be caused by man-made deforestation and rela- 2003) in detail. The first one is located downstream of the
ted soil erosion on slopes. Mai Makden village, north of Mekelle, whereas the second
crops out near the church of Tsabati Mariam, about 12 km
east of Agula (Fig. 11.5). In both cases, bedrock consists of
11.2 Tufa Dams in Tigray horizontally layered limestone (Antalo limestone, Merla and
Minucci 1938). Both the Mai Maikden and the Tsabati
Due to the widespread outcrop area of Jurassic limestone Mariam streams, which deposited the tufa dams and subse-
(Katzmin 1972), the highlands of eastern Tigray (northern quently incised them, are fed by spring waters (Chernet and
Ethiopia) are characterized by the occurrence of a large Eshete 1982).
11 Tufa Dams in Tigray (Northern Ethiopia) as Late Pleistocene … 203
Fig. 11.1 Location map
Fig. 11.2 Picturesque tufa dam

at Romanat, north-west of
Mekele. The tufa wall overlies a
dolerite dyke; a large pool is
present at the base of the
waterfall. Above the dam
remnants of an indigenous forest
can be seen
The temporal evolution (nine 14C dates) of the Tsabati 10,675 cal BP) to 2,380 ± 50 14C years BP (2,592–2,367 cal
Mariam dam backfill (Fig. 11.6) spans from Early to Late BP). The latest date refers to the uppermost organic-rich
Holocene: 9,510 ± 100 14C years BP (11,046– layer underlying a thin layer of alluvial gravels.
Fig. 11.3 A view from distance

of the Mai Makden tufa dam. The
alluvial top surface has been
formed by the last tufa
aggradation episode. The river
incision has completely cut the
tufa dam, reaching the bedrock
The time interval (fifteen 14C dates) covered by the Mai sedimentation as testified by the occurrence of desiccation
Makden dam backfill (Figs. 11.7 and 11.8) is shorter than cracks between 6,510 ± 70 14C years BP (7,485–7,346 cal
that of Tsabati Mariam (7,630 ± 80 to 3,450 ± 50 14C years BP) and 5,610 ± 70 14C years BP (6,472–6,333 cal BP)
BP) since the previous backfill has been completely eroded. (Fig. 11.9). The following evolution of the dam is charac-
However, the stratigraphic sequence here is much better terized by alternating phases of erosion and aggradation. The
exposed. Organic-rich levels and peaty layers from the lower topmost dated level, a buried Vertisol underlying a layer of
part of the backfill swampy sequence were dated between travertine sand, about 1.5 m below the summit surface of the
7,630 ± 80 14C years BP (8,521–8,382 cal BP) and backfill sedimentary sequence, yielded an age of 3,450 ± 50
4,710 ± 70 14C years BP (5,552–5,355 cal BP). The latter 14
C years BP (3,808–3,656 cal BP). Subsequently, the tufa
reading constrains the last date of the swampy environment. dam and the whole backfill were completely incised by the
A noteworthy desiccation event occurred during the swampy Mai Makden stream.
Fig. 11.4 The impressive front

of the Mai Makden tufa dam.
Three different levels of tufa
(T) separated by two fluvial/
debris flow events (F) are visible
Fig. 11.5 The Tsabati Mariam

tufa dam
U/Th dates from other tufa dams of Tigray are reported

by Moeyersons et al. (2006): 261.9 + 61.3/−38.8 kyrs BP; 11.3 Tufa Deposition in Neighbouring
44.4 ± 1.0 kyrs BP; 31.8 + 1.1/−1.2 kyrs BP; 28.7 + 1.4/ Countries
−1.5 kyrs BP; 22.5 + 0.4/−0.5 kyrs BP; 15.8 ± 1.1 kyrs BP;
and 14.1 + 0.6/−0.5 kyrs BP. However, according to these Moeyersons et al. (2006) also report U/Th and 14C dates
authors, only the last two dates, from the Mai K’arano and of tufa deposits measured by Brandt and Brook (1984)
Tsigaba dams, respectively, may be considered sufficiently (U/Th: between 7,600 and 5,000, 9,700 and 9,500, 11,800
reliable. Moreover, a date of 3,090 ± 30 14C years BP years BP) and Voight et al. (1990) (11,535 ± 105 14C years
(3,355–3,276 cal BP) provided by a charcoal sample inclu- BP—13,565–13,287 cal BP; 10,800 ± 80 14C years BP—
ded in top layer of the Tsigaba dam backfill constrains the 12,867–12,705 cal BP; 9,105 ± 105 14C years BP—10,427–
end of tufa deposition. 10,186 cal BP) in northern Somalia.
Fig. 11.6 The Tsabati Mariam

tufa dam backfill
Fig. 11.7 The Mai Makdem tufa

dam backfill showing a sequence
of swampy, lacustrine and alluvial
levels
Holocene tufa dates are reported also from Sudan by Szabo BP) were obtained in southern Yemen by Sander (2006).
et al. (1995) (9.5 ± 0.2, 7.4 ± 0.2, 5.7 ± 0.2 U/Th kyrs BP) and Finally, a very recent date (2,450 ± 110 14C years BP—2,680–
Swezey (2001) (5,395 ± 70 14C years BP—6,266–6,064 cal 2,392 cal BP) of tufa deposition, well corresponding with that
BP). Comparable dates (10,220 ± 40 14C years BP—12,060– obtained from the Tsabati Mariam backfill sequence, is
11,804 cal BP to 4,380 ± 70 14C years BP—5,175–4,894 cal reported by Swezey (2001) from Tchad.
Fig. 11.8 Toppling affecting the

Mai Makdem tufa dam backfill
Fig. 11.9 Desiccation cracks in

the lower part of the Mai Makden
tufa dam backfill
More to the north, in the Saharan Libya (Fezzan, Tadrat (Dramis et al. 2003), it results that during the Early Holocene
Acacus Mts.), a well-studied detailed sequence of Early both areas experienced climate conditions definitely wetter
Holocene tufa deposits, spanning in age from 10,030 ± 85 to than present, associated with a significant northward shift of
7,731 ± 54 U/Th years BP, was obtained by Cremaschi et al. the summer ITCZ, increase of solar insolation and a defi-
(2010). nitely stronger south-west Indian monsoon (Overpeck et al.
1996; Glennie 1998; Fleitmann et al. 2003a, b). As a con-
sequence, between ca. 10 and 7.5 kyrs BP, the currently
11.4 Paleoclimatic Significance of Tufa hyperarid regions such as the Arabic Peninsula (Sander
Dam Aggradation/Erosion Phases 2006) and the Saharan Libya (Petit-Maire and Guo 1996;
Cremaschi et al. 2010) were characterized by long-lasting
From the chronostratigraphic record of Mai Makden and rains and widespread tufa deposition.
Tsabati Mariam dam backfills, supplemented with the dates In the following times, the record of lake levels of the
provided by Moeyersons et al. (2006) from other tufa dams of Ethiopian Rift (Fig. 11.10) indicates some main episodes of
Tigray (Table 11.1), it emerges that tufa aggradation started drier climate (CHP 17, this volume) that seem to be con-
in the Late Glacial (*15.8 U/Th kyrs BP) and followed until sistent with the main phases of less (or absent) deposition or
ca. 2,380 ± 50 14C years BP (2,450–2,345 cal BP). erosion of tufa dams (Dramis et al. 2003). Quite interesting
However, the deposition of tufa was not continuous, as in this context, for its rough correspondence with the most
indicated by the occurrence of some century-scale main recent date of tufa in Tigray (and Tchad), is the progressive
intervals of less or absent deposition and dam incision, also lowering of the rift lake levels after ca. 4700 14C years BP,
testified by the presence of gravel levels and buried stream only interrupted by a short-lived high stand interval between
channels dated for 14.1–10.9; 9.3–8.4; 7.4–6.6 kyrs BP. ca. 2500 and 1500 14C years BP.
Since 5.5–3.7 kyrs BP, the geomorphological–stratigraphic The Late Pleistocene–Holocene sequence of aggradation
record of tufa dams in Tigray indicates a progressive decline phases in the tufa dams in Tigray (in particular, those
of calcium carbonate deposition rates, likely induced by the between 15.8 and 14.1, 10.9 and 9.3, 8.4 and 7.4, 6.6–
general trend of climate to drier conditions (Umer et al. 5.5 kyrs) seems to find some correlation also with the cli-
2004). mate warming stages recorded from the Mediterranean Sea
By comparing the sequence of lake-level fluctuations in floor sediments (Rohling et al. 2002) and, more generally,
the Rift (Gillespie et al. 1983; Street-Perrott and Perrott with the main peaks of the Holocene global temperature
1990; Gasse and Van Campo 1994; Gasse 2000; Umer et al. curves (Mayewski et al. 2004; Rimbu et al. 2004; Dragons
2004) with that of tufa aggradation/erosion in Tigray Flight 2012).
Table 11.1 Tufa dams dates and environmental changes in East Africa and the Mediterranean
Tigray Mai Makden dam Tigray Tsabati Tigray Mai K’Arano Somalia (cal BP) Rift lakes level Mediterranean sea
Mariam dam (cal BP) and Tsigaba dams
*15.8 U/Th BP ? Warming
*14.1 U/Th BP
Low
13,565–13,287 High
12,867–12,705
Low Cooling
11,046–10,675 10,427–10,186 High Warming
10,730–10,483
10,215–9,965
9,408–9,166
Hyatus (ca. 0.8 kyrs) Low Cooling
8,521–8,382 cal BP 8,496–8,346 High
Hyatus? (ca. 0.3 kyrs–8.2 event?) Low
8,247–8,041 cal BP 7,960–7,854 High Warming
8,012–7,867 cal BP 7,598–7,490
8,016–7,788 cal BP
7,612–7,337 cal BP
7,557–7,451 cal BP
7,485–7,346 cal BP
Hyatus (ca 0.8 kyrs) Low Cooling
6,472–6,333 cal BP 6,686–6,519 High Warming
6,423–6,292 cal BP 5,582–5,383
6,234–6,035 cal BP
6,159–5,962 cal BP
6,012–5,792 cal BP
5,926–5,775 cal BP
5,552–5,355 cal BP
Hyatus (ca. 1.7 kyrs) Low Cooling
3,808–3,656 cal BP 2,592–2,367 3,355–3,276 cal BP High Warming
Fig. 11.10 Late Pleistocene–

Holocene fluctuation of the
Ethiopian Rift lakes compared
with the aggradation phases of
tufa dams in Tigray
In this context, it is interesting to notice that Holocene BP seems to coincide with a progressive decrease of
aggradation/erosion sequences of tufa dams, definitely com- humidity in the Sahara (Petit-Maire and Guo 1996)
parable with those investigated in Tigray, have been reported paired by a cooling trend of the Mediterranean Sea;
from Central Italy (Fubelli et al. 2013; Dramis et al. 2014) and (i) the 2,592–2,367 cal BP depositional event, likely
other parts of south-central Europe (Weisrock 1986; Vaudour induced by a short-lived rise of humidity, as that reg-
1994; Žák et al. 2002; Hlaváč et al. 2003; Lespez et al. 2005; istered by the last high stand of the southern Ethiopia
Ollivier et al. 2005; Wehrli et al. 2010). From the above lakes, marks the end of tufa deposition.
considerations, it is possible to establish the following list of The relationships between climate changes and the evo-
events in East-Central Africa and the Mediterranean area: lution of tufa dams (Table 11.1) definitely confirm the
(a) tufa deposits formed between *15.8 U/Th kyrs BP and influence of water availability and temperature on the rates
12,867–12,705 cal BP (Somalia; Voight et al. 1990) in of tufa deposition. Specifically, the rapid growth of tufa
the Late Glacial period, characterized by climate dams observed in both East Africa and Europe with changes
warming in the Mediterranean Sea; a low-stand stage in from cold to warm temperatures, as at the transition from
the rift lakes after ca. 14 kyrs BP could correspond to a phases (b) to (c) or (d) to (e), and the fall of tufa deposition
short dry interval (Older Dryas?) not recorded in the rates at warm to cold transition, seem to provide some
Mediterranean Sea; support to the surface/ground temperature model proposed
(b) the 12,867–12,705 cal BP to 11,046–10,675 cal BP by Dramis et al. (1999). The exhaustion of the postglacial
interval of no deposition seems to fit well with the dry– ground/surface thermal disequilibrium, coupled with the
cold Younger Dryas; progressive aridification of climate in both East Africa and
(c) the 11,046–10,675 to 9,408–9,166 cal BP phase of tufa southern Europe (Petit-Maire and Guo 1996; Giraudi et al.
dam aggradation coincides with the rapid ‘postglacial’ 2011; Zanchetta et al. 2012), might also explain the overall
rise of sea temperature recorded at planetary level; decline of tufa deposition rates and the ultimate incision of
(d) the 9,408–9,166 to 8,521–8,382 cal BP interval of no tufa dams by rivers down to the underlying bedrock,
deposition seems to correspond to a cooling period of observed in Ethiopia, in Central Italy (Dramis et al. 2003,
the Mediterranean Sea; 2014; Fubelli et al. 2013) and in different parts of Europe as
(e) the tufa aggradation intervals recorded in Tigray well (Goudie et al. 1993).
between 8,521 and 8,382 cal BP and 7,485 and
7,346 cal BP coincide with a period characterized by an
increased wetness in Africa (Petit-Maire and Guo 1996) 11.5 Conclusions
and warm sea temperature in the Mediterranean area; in
this context, a short (8,496–8,346 to 8,247–8,041 cal The sequence of Late Pleistocene–Holocene aggradation/
BP) no-deposition interval in the Tigray sequence, even erosion phases of tufa dams in northern Ethiopia, as com-
if not well expressed, could correspond to the pared with the level fluctuations of the rift lakes and the
8.2 kyrs BP cooling event (Alley et al. 1997) registered temperature changes at the planetary level, seems to confirm
in the tufa record of Fezzan (8,619 ± 57 to 8,197 ± 56 the crucial control of water availability and surface temper-
U/Th years BP in Fezzan; Cremaschi et al. 2010) and in ature on the rates of calcium carbonate deposition. The rapid
the Mediterranean Sea; growth of tufa dams observed with cold to warm climate
(f) the 7,485–7,346 to 6,472–6,333 cal BP interval of low changes seems to provide some support to the surface/ground
deposition rates of tufa coincides with a new cooling temperature model proposed by Dramis et al. (1999). In this
stage of the Mediterranean Sea and a trend to increas- context, the exhaustion of the postglacial ground/surface
ing aridity in the Sahara (Petit-Maire and Guo 1996); thermal disequilibrium, coupled with the progressive aridi-
this interval is pointed out by the occurrence of well fication of climate, might explain the overall decline of tufa
developed desiccation cracks in the swampy sequence deposition rates and the ultimate incision of dams by streams.
at the base of the Mai Makden dam backfill; Some tufa dams in Tigray can be easily visited. The
(g) unlike Saharan Libya and Arabic peninsula, between impressive Mai Makden dam is located next to the road
6,472–6,333 and 5,552–5,355 cal BP, the monsoon connecting Mekele and Adigrat, only c. 15 km north of the
continued to keep the Tigray highlands humid enough former and well visible from it. The Romanat dam, c. 20 km
(Arz et al. 2003; Fleitmann et al. 2003a, b) to allow for to the NW of Mekele, is also easily accessible, whilst its
tufa deposition, at least in the more elevated areas; geomorphological significance is enhanced by the sur-
(h) the progressive decline of tufa deposition with alternat- rounding scenery and associated bedrock channel landforms.
ing short-lived aggradation/erosion events recorded at Above the dam and the incision, one can see the remnants of
Mai Makden between 5,552–5,355 and 3,808–3,656 cal an indigenous forest in the region and a monastery.
Acknowledgments Marek Kasprzak is acknowledged for improving Fleitmann D, Burns SJ, Neff U, Mangini A, Matter A (2003b)
Fig. 11.1. Changing moisture sources over the last 330,000 years in Northern
Oman from fluid-inclusion evidence in speleothems. Quat Res
60:223–232
References Ford TD, Pedley HM (1996) A review of tufa and travertine deposits of
the world. Earth-Sci Rev 41(3–4):117–175
Gasse F (2000) Hydrological changes in the African tropics since the
Alley RB, Mayewski PA, Sowers T, Stuiver M, Taylor KC, Clark PU last Glacial maximum. In: Alverson KD, Oldfield F, Bradley RS
(1997) A prominent widespread event 8,200 years ago. Geology (eds) Past global changes and their significance for the future.
25:483–486 Quaternary Sci Rev 19: 189–211
Arz HW, Lamy F, Päzold J, Müller PJ, Prins M (2003) Mediterranean Gasse F, Van Campo E (1994) Abrupt post-glacial climate events in
moisture source for an early-holocene humid period in the northern West Asia and North Africa monsoon domains. Earth Planet Sci
Red Sea. Science 300:118–121 Lett 126:435–445
Atkinson TC (1977) Carbon dioxide in the atmosphere of the Gillespie R, Street-Perrott FA, Switsur R (1983) Post-Glacial arid
unsaturated zone: an important control of hardness in limestones. episodes in Ethiopia have implication for climate prediction. Nature
J Hydrol 35:111–123 306:680–683
Bard KA, Coltorti M, DiBlasi MC, Dramis F, Fattovich R (2000) The Giraudi C, Magny M, Zanchetta G, Drysdale RN (2011) The Holocene
environmental history of Tigray (Northern Ethiopia) in the middle climatic evolution of Mediterranean Italy: a review of the
and late holocene: a preliminary outline. Afr Archaeol Rev 17 continental geologic data. Holocene 21(1):105–115
(2):65–86 Glennie KW (1998) The desert of SE Arabia: a product of climatic
Berakhi O, Brancaccio L, Calderoni G, Coltorti M, Dramis F, Umer M change. In: Alsharhan AS, Glennie KW, Whittle GL, Kendall
(1998) The May Makden sedimentary sequence: a reference point CGStC (eds) Quaternary deserts and climatic change. Balkema
for the environmental evolution of the Highlands of Northern Rotterdam, pp 109–116
Ethiopia. Geomorphology 23(2–4):127–138 Goudie AS, Viles HA, Pentecost A (1993) The late-holocene tufa
Brancaccio L, Calderoni G, Coltorti M, Dramis F, Ogbaghebriel B decline in Europe. Holocene 3(2):181–186
(1997) Phases of soil erosion during Holocene in the Highlands of Griffiths JF (1962) The climate of Africa. In: Russell EW (ed) The
Western Tigray (Northern Ethiopia): a preliminary report. In: Bard natural resources of East Africa. Hawkins, Nairobi, pp 77–87
KA (ed) The environmental history and human ecology of Northern Gullentops F, Mullenders W (1972) Age et formation de dépôts de tuf
Ethiopia in the Late Holocene. Studi Africanistici, ser. Etiopica 5. calcaire Holocène en Belgique. In: Macar P, Pissart A (eds)
Istituto Universitario Orientale, Napoli, pp 29–44 Processus périglaciaires étudiés sur le terrain, Comptes Rendus du
Brandt SA, Brook GA (1984) Archaeological and paleoenvironmental Symposium International de Géomorphologie, Liège-Caen 1971,
research in Northern Somalia. Curr Anthropol 25(1):119–121 67:113–135
Brook GA, Folkoff ME, Box EO (1983) A world model of soil carbon Hlaváč J, Kadlec J, Žák K, Hercman H (2003) Deposition and
dioxide. Earth Surf Proc Land 8:79–88 destruction of holocene calcareous tufa cascades in the Bohemian
Butzer KW (1981) Rise and fall of Axum, Ethiopia: a geoarcheological Karst (Czech Republic). Prace Geogr 189:225–253
interpretation. Am Antiq 46:471–495 Katzmin V (1972) Geological map of Ethiopia. EIGS, Addis Ababa
Chernet T, Eshete G (1982) Hydrogeology of the Mekele area (ND37- Lespez L, Clet-Pellerin M, Limondin-Lozouet N, Pastre J-F, Fontugne M
11). Ministry of Mines and Energy, Addis Ababa (2005) Discontinuités longitudinals des dynamiques sédimentaires
Cremaschi M, Zerboni A, Spötl C, Felletti F (2010) The calcareous tufa holocènes dans si petit vallèes de l’Ouest du Bassin Parisien,
in the Tadrart Acacus Mts. (SW Fezzan, Libya). An early Holocene l’exemple de la Mue (Basse-Normandie). Quaternaire 16(4):273–298
palaeoclimate archive in the central Sahara. Palaeogeogr Palaeocl Mayewski PA, Rohling EE, Stager JC, Karlén W, Maasch KA, Meeker
287:81–94 LD, Meyerson EA, Gasse F, van Kreveld S, Holmgren K, Lee-
Dragons Flight (2012) Holocene temperature variations. http:// Thorp J, Rosqvist G, Rack F, Staubwasser M, Schneider RR, Steig
commons.wikimedia.org/wiki/File:Holocene_Temperature_ EJ (2004) Holocene climate variability. Quat Res 62:243–255
Variations.png Merla G, Minucci E (1938) Missione geologica nel Tigrai. 1-La serie
Dramis F, Materazzi M, Cilla G (1999) Influence of climatic changes dei terreni. Reale Accademia Lincei, Centro Studi per l’Africa
on freshwater travertine deposition: a new hypothesis. Phys Chem Orientale Italiana
Earth PT A 24(10):893–897 Moeyersons J, Nyssen J, Poesen J, Deckers J, Haile M (2006) Age and
Dramis F, Umer M, Calderoni G, Haile M (2003) Holocene climate backfill/overfill stratigraphy of two tufa dams, Tigray Highlands,
phases from buried soils in Tigray (northern Ethiopia): comparison Ethiopia: evidence for late pleistocene and holocene wet conditions.
with lake level fluctuations in the main Ethiopian Rift. Quat Res Palaeogeogr Palaeocl 230:165–181
60:274–283 Mohr PA (1962) The geology of Ethiopia. Addis Ababa University
Dramis F, Fubelli G, Calderoni G, Esu D (2014) Holocene aggradation/ Press, Addis Ababa
degradation phases of tufa dams in northern Ethiopia and central Italy: Nicod J (1986) Facteurs physico-chimiques de l’accumulation des
a palaeoclimatic comparison between East Africa and Mediterranean formations travertineuses. Mediterranée 57 (1–2):161–164
Europe. Z Geomorphol. doi:10.1127/0372-8854/2014/0140 Ollivier V, Guendon J-L, Ali A, Poiron P, Ambert P (2005) Évolution
Ethiopian Mapping Authority (1988) National atlas of Ethiopia. Addis postglaciaire des environments travertineux provençaux et alpins:
Ababa noveau cadre chronologique, faciès et dynamiques morphosédi-
Fattovich R (1990) Remarks on the Pre-Axumite period in Northern mentaires. Quaternaire 17(2):51–67
Ethiopia. J Ethiop Stud 23:1–3 Overpeck J, Anderson D, Trumbore S, Prell W (1996) The southwest
Fleitmann D, Burns SJ, Mudelsee M, Neff U, Kramers J, Mangini A, Indian monsoon over the last 18,000 years. Clim Dyn 12:213–225
Matter A (2003a) Holocene forcing of the Indian monsoon recorded Pedley HM (1990) Classification and environmental models of cool
in a stalagmite from Southern Oman. Science 300:1737–1739 freshwater tufas. Sediment Geol 68:143–154
Pentecost A (2005) Travertine. Springer, Berlin Thrailkill J (1968) Chemical and hydrologic factors in the excavation of
Petit-Maire N, Guo Z (1996) Mis en evidence de variations climatiques limestone caves. Geol Soc Am Bull 79:19–46
Holocenes rapides, en phase dans les deserts actuels de Chine et de Umer M, Legesse D, Gasse F, Bonnefille R, Lamb H, Leng MJ (2004)
Nord de l’Afrique. CR Acad Sci SII 322:847–851 Late quaternary climate changes in the Horn of Africa. In: Batterbee
Rimbu N, Lohmann G, Lorenz SJ, Kim JH, Schneider RR (2004) RW, Gasse F, Stickley CE (eds) Past climate variability through
Holocene climate variability as derived from alkenone sea surface Europe and Africa. Springer, Dordrecht, pp 159–180
temperature and coupled ocean-atmosphere model experiments. Vasseur G, Bernard PH, Van de Meulebrouck J, Kast Y, Jolivet J
Clim Dyn 23:215–227 (1983) Holocene paleotemperatures deduced from borehole tem-
Rohling EJ, Casford JSL, Abu-Zied R, Cooke S, Mercone D, Thomson J, perature data. Palaeogeogr Palaeocol 43:237–259
Croudace I, Jorissen FJ, Brinkhuis H, Kallmeyer J, Wefer G (2002) Vaudour J (1994) Évolution holocène des travertins de vallée dans le
Rapid holocene climate changes in the Eastern Mediterranean. In: Midi méditéranéen français. Géogr Phys Quat 48(3):315–326
Hassan F (ed) Droughts, food and culture: ecological change and food Voight B, Gabriel B, Lassonczyk B, Ghod M (1990) Quaternary events
security in Africa’s later prehistory. Plenum Press, NY, pp 35–47 at the Horn of Africa. Berlin Geowiss Abh, RA 120:679–694
Sander KM (2006) Holocene climate and hydrologic changes recorded Wehrli M, Mitchell EAD, van der Knaap WO, Ammann B, Tinner W
in tufa and lacustrine deposits in Southern Yemen. Graduate School (2010) Effects of climatic change and bog development on
Theses and Dissertations. http://scholarcommons.usf.edu/etd/2691 Holocene tufa formation in the Lorze Valley (central Switzerland).
Fubelli G, Dramis F, Calderoni G, Cilla G, Materazzi M, Mazzini I, Holocene 20(3):325–336
Soligo, M (2013) Holocene aggradation/erosion of a tufa dam at Weisrock A (1986) Variations climatiques et périods de sédimentation
Triponzo (Central Italy). Geogr Fis Din Quat 36. doi:10.4461/ carbonatée à l’holocène – L’age des dépôts. Mediterraneée
GFDQ.2013.36.0 57(1–2):165–167
Street-Perrott FA, Perrott RA (1990) Abrupt climate fluctuations in the Williams PJ, Smith NW (1989) The frozen earth. Fundamentals of
tropics: the influence of Atlantic Ocean circulation. Nature geocryology. Cambridge University Press, Cambridge
343:607–612 Žák K, Ložek V, Kadlec J, Hladíková J, Cílek V (2002) Climate-
Swezey C (2001) Eolian sediment responses to late quaternary climate induced changes in holocene calcareous tufa formations, Bohemian
changes: temporal and spatial patterns in the Sahara. Palaeogeogr Karst, Czech Republic. Quat Int 91:137–152
Palaeocol 167:119–155 Zanchetta G, Giraudi C, Sulpizio R, Magny M, Drysdale RN, Sadori L
Szabo I, Haynes CV, Maxwell TA (1995) Ages of quaternary pluvial (2012) Constraining the onset of the holocene “Neoglacial” over the
episodes determined by uranium-series and radiocarbon dating of central Italy using tephra layers. Quat Res 78(2):236–247
lacustrine deposits of Eastern Sahara. Palaeogeogr Palaeocol
113:227–242
Geomorphology of Ephemeral Streams
in the Kobo Basin 12
Paolo Billi
Abstract
The structural basin of Kobo (northern Ethiopia) is characterised by a semi-arid, monsoon-
affected climate, and it is drained by ephemeral streams. The basin is physiographycally
asymmetric with the highest mountains along the western margin where the largest rivers
originate. The river morphology well matches the ideal model proposed by Schumm (The
fluvial system. Wiley, New York, 1977), consisting of the headwater, the main trunk channel
and the distributary system which represents the river terminus where the whole of the water
flow vanishes due to infiltration and large amounts of sediment are deposited. The main
geomorphic characteristics of the main stem and the distributary systems are described. The
study river streambed is flat, devoid of any bedforms and horizontal and planar lamination is
by far the most common sedimentary structure. Though in a few study reaches the Froude
number calculated for bankfull discharge results steadily around one, i.e. in agreement with the
extensive occurrence of the upper plane bed, the occurrence of outsized particles standing on
or protruding from the streambed, travelling for long distances on the bed surface and showing
no flow perturbation in the fine sediment in their vicinity, is interpreted by a new model of
division association capable of explaining the typical division association observed in the
study river deposits. It is characterised by large boulders rooted in a core coarse layer and the
ubiquitous occurrence of horizontal lamination is interpreted in terms of vertical distribution of
shear stress, hyperconcentrated flow and traction carpet processes.

Keywords
Ephemeral streams Drylands Distributary systems Horizontal lamination Bedload
sheets
12.1 Introduction diagnostic tools to interpret old arid river deposits as potential
reservoirs in oil and gas field exploration (North and Taylor
Ephemeral streams are a main geomorphologic feature in 1996). Ephemeral streams are generated by erratic, infrequent,
many drylands of the planet. In the last decades, these rivers high-intensity rainfalls of short duration or concentrated in a
have attracted the attention of geomorphologists and sedi- short rainy season. River flow is therefore intermittent and
mentologists for the flash flood hazards connected with their flashy and the streambed is dry for the largest part of the year.
impulsive nature (Garcia 1995; Lin 1999) and for providing Dryland rivers have distinctive features, such as the lack
a modern equivalent of depositional characteristics and of well-defined channels, high channel-width-to-depth ratio,
downstream channel narrowing, the predominance of hori-
zontally layered deposits, subtle stratification, the lack of
P. Billi (&) scour and fill sequence (Reid and Frostick 1997; Billi 2008)
Physics and Earth Sciences, University of Ferrara, and the occurrence of crescent scours (Leopold et al. 1966;
Ferrara, Italy Karcz 1968; Picard and High 1973; Billi and Tacconi 1985),
e-mail: bli@unife.it
214 P. Billi
which are not found in humid environments (Knighton and the eastern margin through a deep gorge and enters the
Nanson 1997). Moreover, many of the concepts and models Danakil depression. All these rivers are definitely ephemeral,
to describe the flow/sediment interaction, which are so rel- and direct inspection of their main geomorphological and
evant in controlling river channel morphology, are derived sedimentological characteristics is facilitated by their dry
mainly from studies on perennial rivers and proved difficult nature most of the time.
to apply to ephemeral streams. Though our knowledge on ephemeral rivers and their
The current knowledge on many issues of dryland river distributary systems has substantially increased in the last
processes, form and sediment has been summarised by decades, current conceptual morphodynamics models are
Tooth (2000a), but as pointed out by this authors, some still incomplete and derived from a few examples in a few
important aspects are still unclear and need to be studied in geographical settings. Though sand-dominated distributary
more detail. systems share only some features with terminal fans and
The geomorphology and deposits of ephemeral streams floodouts, they have not been described in detail, in spite of
have been studied mostly along their main stem and, with the patency of their role as probably the more common
the exception of a few papers (e.g. Billi and Tacconi 1985; inland river depositional system in arid and semi-arid areas.
Abdullatif 1989; Tooth 1999; Tooth 2000b), scarce infor- The main aim of this presentation is therefore to contribute
mation is available in the literature about the relationship of to fill this gap and to show yet another fascinating side of
channel morphology, sediment transport modes and sedi- Ethiopian geomorphology and landscape.
mentary structures in terminal splays and distributary sys-
tems, though they may play the most important depositional
role, especially in closed basin filling. 12.2 General Setting of the Study Area
In the structural basin of Kobo–Alamata in northern
Ethiopia, all the rivers draining the western margin end up in The Kobo–Alamata basin is an intermontane, structural basin
the basin floor where they form distributary systems, with stretching across the border between Welo and Tigray in
only one exception, the Golina River, which passes beyond northern Ethiopia (Fig. 12.1). It has an elongated, rectangular
Fig. 12.1 Study area location map. Dotted line basin watershed divide; red line, Addis Ababa—Mekele road; green squares and numbers refer to
the location of the satellite images of the corresponding figure
12 Geomorphology of Ephemeral Streams in the Kobo Basin 215
shape, reflecting its structural origin associated with the

development of the main rifting system of the area. North–
south-oriented master faults, in fact, bound the margins of the
basin and separate it from the main Ethiopian plateau to the
west, whereas the eastern ridge marks the ultimate structural
relief before the Danakil depression to the east.
The basin is filled with Quaternary alluvial deposits,
whereas both the eastern and western margins consist of
Ashenge Basalts (basalt flows with scarce tuffs) overtopped,
on the western margin, by Amba Alaji Rhyolites making up
the highest peaks of the area. In the very south-east and
north-east margins, small portions of Mesozoic sedimentary
rock formations (Antalo Limestones and Amba Aradam
Sandstones) crop out (Merla et al. 1979) (Fig. 12.2).
The lowest elevation of the basin floor is around 1,374 m
a.s.l., and it is found in a restricted area east of Kobo, taken
up by a seasonal marsh (Fig. 12.3). The highest peaks are in
the western divide at elevations of 3,400–4,100 m a.s.l.,
whereas the eastern margin is less elevated with the highest
peak at 2,409 m a.s.l. (Fig. 12.1). As a consequence of such
physiographic asymmetry, the largest rivers flow from the
Fig. 12.2 Geology of the study area. Stratigraphic units: Ja Antalo
Limestone (Callovian to Kimmeridgian). Neritic limestones and marls; western margin to the centre of the basin where they form
Ks Amba Aradam Sandstones (Upper Jurassic to Lower Cretaceous). terminal distributary channels and dry up. By contrast, the
Fluvial/littoral variegated, quartzose sandstone; Tβ Oligocene to eastern margin is drained by very short and steep streams
Miocene Ashenge Basalts. Olivine bsalts alternating with tuffs; Tσ that form alluvial fans as they enter the basin floor and small
Amba Aiba basalts (Oligocene to Miocene). Flood basalts with rare
tuffs; Mα Amba Alaji Rhyolites (Miocene). Rhyolitic ignimbrites and distributary systems beyond them (Fig. 12.4). Only one river
tuffs, trachyrhyolites and flood basalts; Qtβ Miocene to Pleistocene (the Golina R.) crosses the Kobo basin floor and overpasses
fissural basaltic lavas, subordinately, ignimbrites and rhyolites; Qa the eastern margin through a deep gorge (Agamsa gorge)
Quaternary alluvial deposits, locally terraced. Symbols: F fault line (Fig. 12.1). Beyond it, the river runs for about 100 km into
(modified from Merla et al. 1979)
Fig. 12.3 The basin floor with

the town of kobo on the left and
the least elevated area (1,374 m a.
s.l.) taken up by the seasonal
marsh close to the eastern margin
216 P. Billi
year-round since mean daily values (the only available for

this area) are commonly over 20 mm and peaks of absolute
maximum intensity in 24 h ranging from 50 to 100 mm may
occur also during the ‘small rains’ period (March–May) (see
Chap. 4) and exceptionally also in the dry winter period
(mainly December and January). Though these values are
reported as daily rainfall intensity, the core of heavier rain-
storms is much shorter and typically less than 1 h (Billi
2011). Given such highly intermittent rainfall pattern, the
streambeds of the rivers in the Kobo basin are completely
dry for most of the time and water flows only in response to
rainfalls yielding a sufficient volume of run-off.
12.3 River Morphology
With the only exception of the Golina River, which crosses

the Kobo basin floor and proceeds beyond its eastern mar-
gin, all other rivers have a similar general geomorphology.
Similar to the ideal river model proposed by Schumm
(1977), they consist of three portions: (1) the headwater;
Fig. 12.4 The alluvial fans built by steep and short streams on the
eastern margin of the basin (2) the main trunk channel and (3) the distributary system.
The headwater reaches are typically cut into bedrock or the
streambed takes up the bottom of a narrow valley filled with
the Danakil lowland where it joins the Awira River in the coarse-grained alluvium but devoid of a true alluvial plain.
Hor Mat plain at an elevation of about 700 m a.s.l. The The steepest, upstream reaches share the same morphologi-
Awira R. proceeds for further 100 km and fades out in a cal and sedimentological features observed in the mountain
large distributary system in the Gerule Plain at an elevation streams of humid and sub-humid areas (see Montgomery
of about 350 m a.s.l. and Buffington 1997, for a review). In the headwater lower
Unfortunately, no river flow data are available for any of reaches, the streambeds take up the whole valley bottom and
the rivers draining the basin, whereas two meteorological acquire a sinuous pattern, likely influenced by local geo-
stations, namely Kobo and Alamata, are located on the basin logical structures. In the western margin, the transition from
floor, near the western margin (Fig. 12.1). These two stations the headwater to the basin floor is commonly marked by a
are close to each other (only 30 km apart), at about the same narrow gorge (Fig. 12.5), downstream of which the main
elevation and in a similar physiographic position within the river trunk originates. Here, unlike what one would expect,
basin; hence, their climatic data are very similar. Mean max- no modern alluvial fan is found. This is a distinctive feature
imum and minimum temperatures are highest in June, about of the Kobo basin since all the rivers entering the basin floor
34 and 18 °C, respectively, in coincidence with high sun and from the highlands to the west do not develop an alluvial
clear sky conditions, whereas the lowest temperatures are fan. By contrast, in spite of the less elevated basin divide and
recorded in January, about 27 and 12 °C, respectively. much smaller headwater catchments, the majority of the
Mean annual rainfall ranges between 726 and 768 mm for rivers coming from the eastern side show well-developed
Kobo and Alamata, respectively, whereas potential evapo- alluvial fans (Fig. 12.4). Such a geomorphic difference
transpiration, calculated by Thornthwaite’s method, is between the two basin sides stands probably in the basin
931 mm year−1. Though the annual amount is relatively floor being slightly tilted to the east (Fig. 12.3). The longi-
large, precipitation consists mainly of few, very intense tudinal profile of the main western rivers, in fact, does not
rainstorms that are more frequent during the kiremt (the show any discontinuity or abrupt change in gradient. This
summer, monsoon-type big rains—see Chap. 4) and may implies that, with time, the river channels have adjusted to a
occur as isolated events during the dry spell. The annual progressively eastward-sloping floor. Accurate field inspec-
rainfall values are typical of sub-humid areas, and appar- tions indicate that, in some places, such river response is also
ently, a permanent baseflow would be expected in the witnessed by a chain of two or three old distributary systems,
majority of rivers. Actually, about 50 % of the annual rain each of them overstepped by the river to form another dis-
falls during July and August as typical monsoon downpours tributary system and prograding towards the opposite side of
(see Chap. 4), but precipitation intensity is high the whole the basin.
Fig. 12.5 The narrow gorge at

the transition from the headwater
to the main stem of the Dikala
River
12.3.1 The Main Stem river size, channel width expands rapidly downstream of the
headwater gorge and, though local broadenings due to par-
The main stem is found on the basin floor, and it is typically ticularly extensive bank collapse may occur, it tends to reach
straight with an overall morphology very similar to that of a maximum at the distance of 2–4 km upstream of the dis-
the arroyos of the American authors (e.g. Leopold et al. tributary system (Fig. 12.7). Typically, the main stem is
1966; Graf 1988) (Fig. 12.6). The channel, in fact, is cut into joined by tributaries only in its upstream portion where the
the Quaternary basin fill deposits and its cross section is headwater catchment of the larger rivers exceeds the range
rectangular with almost vertical banks, 4–5 m high in the of 50–100 km2 indicated by Wolman and Gerson (1978) as
upstream reaches and decreasing to about one metre at the the threshold at which channel width becomes constant.
first bifurcation of the distributary system. Irrespective of However, due to the flashy nature of floods and the long
Fig. 12.6 The straight, main

stem of the Dikala River upstream
of the town of Kobo
218 P. Billi
350 the very permeable sediment of the streambed and the banks,
Greb Oda
300 Dikala
resulting in a decrease of channel width (Dunkerley 1992;
Mersa Tooth 1999, 2000b).
250
The data reported in Fig. 12.7 are from rivers draining the
Width (m)
200 western margin, and a marked relative minimum of channel

150 width can be observed to precede the maximum width. This
100
narrowing is found at the transition between the river
expansion beyond the headwater gorge and the most
50
upstream development of alluvial terraces and can be inter-
0 preted as a knick point associated with the streambed inci-
-14 -12 -10 -8 -6 -4 -2 0 2 4
sion in the basin fill deposits following the eastward tilting
Downstream distance from datum (km)
of the basin floor. Such a structural stetting is likely
Fig. 12.7 Downstream variation of main stem channel width for, responsible for the steep gradient of the trunk river bed
Dikala and Mersa rivers. The datum coincides with the bridge on the (Table 12.1) (Billi 2007) that is by far beyond the critical
Kobo-Alamata road. The dashed lines are third (Mersa) and second
slope for sandbed streams, as are most in the study area, and
grade (Gereb Oda and Dikala), respectively, polynomial interpolation
lines close to that of the Golina, which is instead a boulder bed
river (Fig. 12.8). From Table 12.1, it can be noticed that the
main stems, the distributary channels and the Golina River
Table 12.1 Main geomorphic characteristics of the study rivers
have very similar geomorphic and sediment characteristics,
with the sole exception of the Golina bed material that is
Reach type/river Gradient D50 (mm) w/d
much coarser.
Main stem 0.0141–0.0445 0.15–2.70 43–418
Distributary 0.0109–0.0272 0.72–1.35 36–325
Golina river 0.0187 46.1–93.5 ca 300 12.3.2 The Distributary System
D50 is the particle size for which 50 % of the distribution is finer; w/d is
width/depth ratio Though the overall morphology of the study area distributary
systems is rather similar, some differences can be observed.
They include the horizontal angle of the distributary system
Negeharo Dikala Golina apex, the morphology and style of anabranching. Larger
Wunda Ashewa Gereb Oda streams split into larger distributary systems and the average
100
90
channel width of the trunk stream (W) measured in a 1–2-km-
80 sand
long reach upstream of the distributary apex, explains 91 %
70 of the apex angle (α) variability (Fig. 12.9). The total mean
width of the first two or three anabranches rooted in the
% finer
60
50 distributary apex (Wa) was found to be constantly larger than
boulders
40 W, irrespective of the trunk river size and the anabranch

30 number. Wa, in fact, ranges between 1.5 and 3.3 W. This
20
gravel
10
0 120
- 10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4
Phi units 100
80
Fig. 12.8 Bed material grain-size distribution curves of a few,
α (°)
representative ephemeral streams of the Kobo basin 60

y = 0.3297x + 27.823
40
R² = 0.91
return time of bankfull flows, which seems to characterise
20
ephemeral streams (Williams 1978; Graf 1988), rivers are
unable to effectively rework the channel morphology 0
imprinted by larger floods and oversized channel result. 0 50 100 150 200 250
Given the very high intensity of rainfall, run-off from the W (m)
basin floor sides may enter directly the river channel, thus
Fig. 12.9 Correlation between the average channel width in the reach
contributing to increase discharge and width as far as an upstream of the distributary system and the angle of the distributary
appreciable volume of water is lost by infiltration through system apex
Fig. 12.10 The distributary

system of the Harosha River. The
distributary anabranches have
comparable width, may form
secondary, smaller distributary
systems or can proceed as
individual, narrow channels with
characteristics similar to the
floodouts of Tooth (1999)
implies that, though all the anabranches are active the year- distributary system main apex (Fig. 12.10). Though Billi
round, during individual, ordinary (less than bankfull) floods, (2007) addressed such distributary styles, the factors con-
the largest proportion of the flow is conveyed only by one trolling them are not yet clear.
anabranching channel. This observation is consistent with the The distributary channels are slightly incised in the valley
theoretical predictions of Bolla Pittaluga et al. (2003) and the fill alluvium, and their bank height decreases from 1 to 2 m
findings of Mosley (1983) who pointed out that in braided upstream to zero in the distal reaches. According to Billi
rivers the main channel carries generally a high percentage of (2007), the distributary systems of the Kobo basin consist
the total discharge, ranging from 65 to 85 %. A predictive commonly of five reach units, with specific morphological
model for the equilibrium configuration of a simple ‘Y- and sedimentological characteristics. From upstream to
shaped’ bifurcation in which an upstream channel splits into downstream, they are (Fig. 12.11) (1) the main feeder
two equivalent branches has been proposed by Wang et al.
(1995). The model of Bolla Pittaluga et al. (2003), though
conceived to account for the formation of braid bars in Fig. 12.11 Sketch of the
gravel-bed rivers, can be applied also to bifurcations of dis- distributary system reach units
tributary systems. The model, in fact, depicts a condition of model: (1) main feeder channel;
(2) primary distributary channel;
unbalanced equilibrium configurations, even in the case of (3) flow expansion reach; (4)
perfectly symmetric geometry, which develops for low val- accretionary slip face; (5) run-out
ues of the sediment mobility, as it is the case of ephemeral distal reach
streams. These rivers are commonly characterised by a high
concentration of bedload transport and high sedimentation
rates, associated with the flashy nature of floods and high
infiltration rates (Billi 2008). High rates of sedimentation
may also result in a difference of bed elevation at the
upstream inlet of the anabranches and in large values of the
width/depth ratio, i.e. two typical features of ephemeral
streams (Table 12.1) (Billi 2007).
Some distributary channels run for a few kilometres as an
individual channel and fade away like the floodouts descri-
bed by Tooth (1999). Others may split again into two sec-
ondary anabranches in a tree-like structure or form a
secondary, terminal splay at a certain distance from the
220 P. Billi
Fig. 12.12 Application of the

reach units classification scheme
to the distributary system of
Gerday-Weja River. The numbers
refer to the model units of
Fig. 12.11
channel; (2) the primary distributary reach; (3) the flow centimetres) and their discharge is less than one-tenth of the
expansion reach; (4) the accretionary front; and (5) the run- main stem flow (Billi 2007). They are typically the longest
out channel. reaches within the distributary system and maintain their size
The feeder channel is the main river stem that has been as long as the flow completely vanishes due to infiltration,
described in Sect. 12.3.1 and represents the dynamic axis of splays out because of lack of confinement or vegetation
the river system (Kelly and Olsen 1993). In their upstream resistance or evaporates.
portion, the primary distributary reaches have an arroyo-like
morphology similar to that of the main stem though bank
height is smaller. Right downstream of the bifurcation, pri- 12.4 Sediment and Bedforms
mary distributary channels tend to have comparable widths
(Fig. 12.10). These channels may proceed as floodouts Bed material of the Kobo basin ephemeral streams consists
(sensu Tooth 1999, i.e. similar to the run-out channel of the mainly of medium to coarse sand and, subordinately, of fine
model presented here) or, on their turn, may bifurcate into gravelly sand with pebbles and even large boulders scattered
narrower and shallower distributary reaches. Further down- on a flat bed devoid of any bedforms or depositional bars.
stream, bank height decreases to a few centimetres and, in Actually, the streambed appears as made of shallow rhom-
practice, flow becomes unconfined forming the flow boid bars, the maximum width of which is in the order of
expansion reach (Fig. 12.12). Here, the largest proportion of 0.2–0.05 channel width. These apparent bars are poorly
flow infiltrates; hence, bedload sediment came to rest and developed, flat topped, commonly less than 50 cm thick and
deposition takes place forming a flat, wedge-shaped body lack accretionary fronts or other progradational features,
with an accretionary front. Such bed material accumulation suggesting they do not result from depositional processes but
is washed out by receding flood flows or subsequent shallow are produced by streambed dissection processes during the
flows removing the finer particles. This process makes the receding or shallower flood flows (Billi 2007, 2008).
accretionary front as the coarsest bed material unit of Smaller-scale, 40–50-mm-thick and 0.5–2.0-m-wide, over-
the whole distributary system. The flow infiltrating through lapping leaf-shaped bars/bedforms are a common streambed
the accretionary front may re-emerge downstream as clean feature. Billi (2011) observed that the thickness of these
water that recovers some excess shear stress capable of bedforms coincides with that of the bedload flux measured in
entraining the fine bed/soil particles of the distal areas and to the field on the Gereb Oda River (Fig. 12.1) during short
incise the narrow run-out channel (Fig. 12.12). These distal return time floods. This finding is consistent with the results
channels are very narrow (1–3 m) and shallow (a few obtained by the relation developed by Karim and Kennedy
Fig. 12.13 A large boulder

resting on a thin, reversely
graded, fine gravel division. No
clear evidence of both crescent
scour and wake deposit upstream
and downstream, respectively, of
the boulder can be perceived. The
spatula is about 25 cm long
(1983) to predict the thickness of sandy bedload sheets, evident flow disturbance around them (Fig. 12.13) requires
which resulted to be in the 8–65 mm range. Though more some complementary considerations and explanations. In
field and laboratory data and observations are needed, the fact, commonly, large boulders appear as floating on top of
leaf-shaped bedforms seem to be the product of moving bed massive fine gravel layers and in places, such as the main
material layers, the thickness (and likely size) of which stem and the first anabranches of the Dikala River (Fig. 12.1),
varies with flow discharge. This process was postulated by there is evidence of their movement on the streambed surface
Billi (2008) to account for the predominance of horizontal for a distance as far as 2 km (Fig. 12.14). Also Hassan et al.
lamination and the lack of small-scale bedforms, like ripples (1999) found that coarse gravel particles move during the
and dunes, on the streambed of the study area ephemeral transition regime in their study of a sandy ephemeral stream
streams. in Botswana, and field inspection after floods revealed a flat
The classical flume experiments of Simons and Rich- bed with no bedforms. In the Kobo basin study reaches,
ardson (1966) have shown that plane bed develops at the Froude numbers around one are typical at bankfull discharge
transition from lower to upper-flow regime with Froude (Billi 2008) and, though no field data of the hydraulic char-
numbers around unity. For two ephemeral streams of the acteristics of the flood vanishing flows is available, the lack
Kobo basin, Billi (2008) calculated that at bankfull flow of sedimentary structure and bedforms, such as low flow
Froude numbers are very close to one. In the flume experi- regime ripples and dunes, cannot be accounted for by a
ments of Alexander et al. (2001), antidunes were generated gradually changing flow. In fact, shallow receding flows with
under a supercritical flow and, as discharge was reduced, transport capacity enough to rework the bedforms should be
antidunes decreased in length and height and the bed became able to develop also horseshoe vortices on the upstream side
nominally plane with millimetre-high asymmetrical bed- of virtually still boulders standing on the streambed, to carve
waves as those described by Best and Bridge (1992). crescent scour holes and deposit fine wakes (Picard and High
In the study rivers, natural weavy laminae are not so 1973; Billi 1988). On the streambed of the study reaches,
evident as in the flume experiment. Their concavity is barely evidence of a late filling of the scour pockets is very
detectable in the field and they are characterised by upward uncommon and present only upstream of very large boulders
coarsening instead of normal grading as observed in the (Fig. 12.15). This suggests that both the finer material and the
laboratory (Billi 2008). Notwithstanding such differences, it boulders are transported and deposited together and the
can be accepted that horizontal laminae are generated under development of turbulent eddies, tumbling, divergent and
an upper-flow regime at the transition from plane bed to convergent flows around the large particles is inhibited.
antidune, but their extensive and ubiquitous occurrence as the From Figs. 12.13, 12.14 and 12.15, a tight connection
dominant bedform and sedimentary structure in the study between the larger boulders and the much finer, horizontally
rivers and the interaction with large boulders without any laminated sediment is evident. Billi (2008) has proposed a
222 P. Billi
Fig. 12.14 White boulders

scattered on the primary
distributary channel upstream of
Gara Lencha hill near Kobo.
These large boulders were
entrained from a bank protection
wall at the road bridge of Kobo
and travelled a few hundreds of
metres downstream on top of the
streambed
Fig. 12.15 A large boulder

rooted in a reversely graded, fine
gravel layer on the streambed of
the Wunda Ashewa River (few
kilometres South of Kobo). In this
case a small scour pocket filling
with fine sediment is present
upstream of the boulder. The
folded stick ruler is about 22 cm
long
simple model, inspired by the bipartite model of Sohn study dry rivers, the occurrence of individual large boulders
(1997), to explain the association of particles with such with basal reverse grading can be accounted for by buoyancy
diverse size and the ubiquitous occurrence of horizontal generated by vertical stratification and dispersive pressure as
lamination. The bipartite model of Sohn (1997) implies that predicted for the collision region by Sohn’s model
grain segregation is not possible in the frictional zone (Fig. 12.13 and 12.15).
because particle concentration is very high and/or the The inspection of isolated, large particles standing on the
occurrence of elutriation induced by pore fluid escaping. bed surface reveals that they rest on and/or are associated
According to these considerations, reverse grading should with a coarser (commonly fine to medium gravel) division
not occur in the lower part of the moving sediment, but in (Fig. 12.15) and the horizontally laminated and massive sand
the Kobo basin ephemeral streams deposits, the basal divi- in their vicinity is typically undisturbed. This makes up the
sion is commonly reversely graded (Figs. 12.15 and 12.16) coarse core division of the typical internal bed arrangement
and only occasionally it is massive or normally graded. observed in the study river deposits, consisting, from bottom
Though this seems to reject the hypothesis of a basal fric- to top, of the following four main divisions (Fig. 12.17): (1)
tional region and the occurrence of a bipartite flow in the the basal reversely graded or massive, fine-grained division;
Fig. 12.16 Application of the

division classification model to a
stratigraphic log of a modern
streambed cut bank of the Mersa
River. Horizontally laminated and
massive sand in the vicinity of
larger particles show no flow
perturbation features such as
scour holes or lee side fine-
grained wake. Division notation:
Cc coarse core division; Lh
horizontal lamination; MD mud
drape; RG reverse grading. The
stick ruler for scale is about 90 cm
long
Fig. 12.17 Sketch model of the

typical division arrangement
within an individual layer (from
Billi 2008—modified)
(2) the core coarse division; (3) the horizontally laminated beheaded by subsequent flood pulses and/or the following
sandy and grainy division; and (4) the receding flood flow flood and its basal division may be lacking for the onset of a
mud or, subordinately, sandy mud drape. An example of surge flood that may have deeply reworked it.
application of such a scheme to channel deposits is reported This division association model seems to conform very
in Fig. 12.16. In places, the typical division arrangement is well to the vertical shear stress distribution postulated by
not present with its complete sequence as it may be Billi (2008) within the collisional zone (Fig. 12.18) rather
224 P. Billi
from 0.01 to 0.04, which is unusually higher than the critical

slope expected for sandbed rivers.
The distributary systems on the floor of the Kobo basin
show peculiar geomorphic characteristics, different from
those described by Tooth (2000a, b) for his Australian
floodouts and those of the terminal fans studied by Parkash
et al. (1983) and Kelly and Olsen (1993). Field investiga-
tions by Billi (2007) revealed that the former are made of
five reach units, with different morphological, sedimento-
logical and hydrological characteristics: (1) the main feeder
channel; (2) the primary distributary reach; (3) the flow
expansion reach; (4) the accretionary front; and (5) the run-
out channel.
The streambed of all the study rivers is flat, devoid of any
Fig. 12.18 Vertical distribution of shear stress within hyperconcen- bedform and punctuated by individual very large particles.
trated flow (modified from Sohn 1997) In places, rhomboid bars are present, but they are only
apparent longitudinal bars since no downstream grain size
sorting nor accretionary fronts (sensu Bluck 1979) were
than to a variation in the grain size of the supplied sediment observed on their surface and in cut banks. They are sup-
as proposed by Sohn (1997). In fact, the coarse core division posed to be formed by dissection processes active during the
can be associated with the highest shear stress of the colli- receding flood flows.
sion zone. The lower massive or reversely graded division is In both the main stem and the distributary channels, 99 %
associated with lower values of shear stress at the transition of sediment structures consist of horizontal, planar bedding,
between the collisional and frictional zones (though sedi- whereas large boulders, scattered on the bed surface, show no
ment features indicating the presence of a frictional zone upstream crescent scour or downstream fine wake (Picard and
were not observed in the study river deposits). Finally, the High 1973; Billi and Tacconi 1985). This seems to indicate
upper, horizontally laminated division develops in the upper that large and fine particles are deposited at very similar flow
part of the collisional zone where, according to Sohn (1997), competence thresholds. Moreover, the large boulders are
shear stress decreases again. This division is characterised by commonly associated with or rooted into a coarse-grained,
thin bedload traction layers, and the decrease in downward reversely graded division which is the core of a typical
flux of grains, as advocated by this author, is not necessary, division arrangement making up the horizontal beds in the
given a decrease in shear stress and particle concentration in modern and older alluvium. A model to describe such an
this zone (Fig. 12.18). These considerations, the prevailing arrangement is presented. It consists of four divisions, which
geometry and upward fining of the horizontal laminae sug- from bottom to top are (1) the basal reversely graded or
gest these bedforms can be interpreted as thin bedload sheets massive, fine-grained division; (2) the core coarse division;
rather than the results of low-amplitude bed waves migration (3) the horizontally laminated sandy and grainy division; and
as postulated by Best and Bridge (1992). (4) the receding flood flow mud and sandy mud drape.
This model is based on the hypothesis of vertical shear
stress distribution postulated by Sohn (1997) for hypercon-
12.5 Conclusions centrated flows. Shear stress, in fact, is predicted to increase
from a lower value in the basal massive or reverse-graded
The ephemeral streams of the Kobo basin have a general division to a maximum coinciding with the coarse core
morphology that well matches the river system model pro- division and to decrease towards the horizontally laminated
posed by Schumm (1977). The trunk river is cut into the division on the top layer. By this model, it is possible to
Quaternary valley fill deposits, has a straight pattern and it is account for the characteristics and origin of horizontal
characterised by large width-to-depth ratios (as much as laminae that are the most common sedimentary feature of
400). Channel width tends to increase to a maximum shortly ephemeral streams. In the study rivers, the position of hor-
upstream of the first bifurcation, giving way to the distrib- izontal laminae in the division association model and their
utary system, and then, it decreases due to water infiltration upward fining fabric indicates these bedforms are to be
and splitting into downstream anabranches. Tectonics is interpreted as generated by thin bedload sheets rather than by
likely responsible for the lack of alluvial fans on the basin the migration of low-amplitude bed waves as postulated by
western margin, and the steep streambed gradient, ranging Best and Bridge (1992).
The flow energy at bankfull discharge was calculated by Karcz I (1968) Fluviatile obstacle marks from wadis in the Negev
Billi (2008) as being capable of entraining the large boulders (southern Israel). J Sedim Petrol 38:1000–1012
Karim MF, Kennedy JF (1983) Computer-based predictors for
punctuating the study river streambed, but the lack of flow sediment discharge and friction factor of alluvial streams. In:
disturbance around them and the erosionless transition of the Proceedings of second international symposium on river sedimen-
coarse core division, in which they are rooted, to the lower tation, Nanjing, China, Paper A18
and overlaying massive divisions suggest that both large and Kelly SB, Olsen H (1993) Terminal fans—a review with reference to
Devonian examples. Sed Geol 85:339–374
small particles are subjected to equal mobility conditions Knighton AD, Nanson GC (1997) Distinctiveness, diversity and
leading to the conclusion that, though buoyancy forces may uniqueness in arid zone river systems. In: Thomas DSG (ed) Arid
be active on boulders, en masse bed material transport is zone geomorphology: process, form and change in drylands. Wiley,
expected to be the prevailing bedload transport process. Chichester, pp 185–203
Leopold LB, Emmet WW, Myrick RM (1966) Channel and hillslope
Acknowledgments Research funded by the Italian Ministry of Edu- processes in a semiarid area New Mexico. USGS Professional Paper
cation, University and Research, PRIN Project 2003 n. 352-G
2003040420_003 and by National Geographic Society, Committee for Lin X (1999) Flash floods in arid and semiarid zones. UNESCO-IHP,
Research and Exploration, grant n. 8400-08. Paris
Merla G, Abbate E, Azzaroli A, Bruni P, Canuti P, Fazzuoli M, Sagri
M, Tacconi P (1979) A geological map of Ethiopia and Somalia.
Consiglio Nazionale delle Ricerche, Firenze
References Montgomery DR, Buffington JM (1997) Channel-reach morphology in
mountain drainage basins. Geol Soc Am Bull 109(5):596–611
Mosley MP (1983) Response of braided rivers to changing discharge.
Abdullatif OM (1989) Channel-fill and sheet-flood facies sequence in J Hydrol NZ 22(1):18–67
the ephemeral terminal River Gash, Kassala. Sudan Sedim Geol North CP, Taylor KS (1996) Ephemeral-fluvial deposits; integrated
63:171–184 outcrop and simulation studies reveal complexity. AAPG Bull 80
Alexander J, Bridge JS, Cheel RJ, Leclair SF (2001) Bedforms and (6):811–830
associated sedimentary structures formed under supercritical water Parkash B, Awasthi AK, Gohain K (1983) Lithofacies of the Markanda
flows over aggrading sand beds. Sedimentology 48:133–152 terminal fan, Kurukshetra district, Haryana, India. In: Collinson JD,
Best J, Bridge J (1992) The morphology and dynamics of low Lewin J (eds) Modern and Ancient Fluvial Systems. Spec Publ Int
amplitude bedwaves upon upper stage plane beds and the preser- Ass Sediment 6:337–344
vation of planar laminae. Sedimentology 39:737–752 Picard MD, High LR Jr (1973) Sedimentary structures of ephemeral
Billi P (1988) Forme di fondo grossolane. Giorn Geol 50:15–30 streams. Developments in sedimentology 17, Elsevier, Amsterdam
Billi P (2007) Morphology and sediment dynamics of ephemeral stream Reid I, Frostick LE (1997) Channel form, flows and sediments in
terminal distributary systems in the Kobo basin (northern Welo, deserts. In: Thomas DSG (ed) Arid zone geomorphology: processes,
Ethiopia). Geomorphology 85:98–113 form and change in drylands. Wiley, Chichester, pp 205–229
Billi P (2008) Flash floods, sediment transport and bedforms in the Schumm SA (1977) The fluvial system. Wiley, New York
ephemeral streams of Kobo basin, northern Ethiopia. Catena 75 Simons DB, Richardson EV (1966) Resistance to flow in alluvial
(1):5–17 channels. USGS Professional Paper 422 J
Billi P (2011) Flash flood sediment transport in a steep sand-bed Sohn YK (1997) On traction-carpet sedimentation. J Sedim Res 67
ephemeral stream. Int J Sedim Res 26(2):193–209 (3):502–509
Billi P, Tacconi P (1985) Morfologia sedimentologia e dinamica Tooth S (1999) Floodouts in central Australia. In: Miller AJ, Gupta A
fluviale di due corsi d’acqua effimeri della Somalia settentrionale. (eds) Varietis of fluvial form. Wiley, Chichester, pp 219–247
Quad di Geologia della Somalia 8:43–86 Tooth S (2000a) Process, form and change in dryland rivers: a review
Bolla Pittaluga M, Repetto R, Tubino M (2003) Channel bifurcation in of recent research. Earth-Sci Rev 51:67–107
braided rivers: equilibrium configurations and stability. Water Tooth S (2000b) Downstream changes in dryland river channels: the
Resour Res 39(3):1046–1059 Northern Plains of arid central Australia. Geomorphology 34:33–54
Bluck BJ (1979) Structure of corse-grained braided stream alluvium. Wang ZB, Fokkink RJ, De Vries M, Langerak A (1995) Stability of
Trans Roy Soc Edinb 70:182–221 river bifurcations in 1D morphodynamics models. J Hydraul Res 33
Dunkerley DL (1992) Channel geometry, bed material and inferred (6):739–750
flow conditions in ephemeral stream systems, Barrier Range, Williams GP (1978) Bankfull discharge of rivers. Water Resour Res
western N.S.W. Australia. Hydrol Process 6:417–433 14:1141–1158
Garcia CC (1995) Torrential flow frequency and morphological Wolman MG, Gerson R (1978) Relative scales of time and effective-
adjustment of ephemeral channels in South-East Spain. In: Hicki ness of climate in watershed geomorphology. Earth Surf Proc Land
EJ (ed) River Geomorphol. Wiley, Chichester, pp 169–192 3:189–208
Graf WL (1988) Fluvial processes in dryland rivers. Springer, Berlin
Hassan M, Shick A, Shaw P (1999) The transport of gravel in an
ephemeral sandbed river. Earth Surf Proc and Landforms 17:711–722
Sediment Yield Variability at Various
Spatial Scales and Its Hydrological 13
and Geomorphological Impacts
on Dam-catchments in the Ethiopian
Highlands
Nigussie Haregeweyn, Atsushi Tsunekawa, Jean Poesen,

Mitsuru Tsubo, Jan Nyssen, Matthias Vanmaercke, Amanuel Zenebe,
Derege T. Meshesha, and Enyew Adgo
Abstract
This chapter analyzes the spatial variability, impacts, and factors of sediment yield (SY) and
reservoir sedimentation rates (SRs) in the upper Blue Nile River Basin. SY data collected
using reservoir sediment surveys or runoff and suspended sediment concentration (SSC)
measurements at river gauging stations in the framework of different programs were compiled,
screened, and used for our analysis. A large spatial variation in area-specific SY (SSY),
ranging between 4 and 4,935 t km−2 year−1, was observed among catchments. This variation is
attributed to both human and environmental factors. The high SY values have drastic
consequences for the life expectancy of many reservoirs in the Ethiopian highlands: 50 % of
reservoirs risk losing their economic life within half of the design period. Moreover, the high
trapping efficiency of the reservoirs for flow and sediment led to selective deposition of
sediment fractions within the reservoir and channel stabilization and vegetation regrowth in
the downstream river reaches. Unfortunately, the availability and reliability of SY data for this
region is poor by international standards. The SY assessment initiatives taken through
institutional collaboration projects is a good start; however, such projects have limited capacity
and a short life span, so they cannot produce a sustainable solution for this important data
gap. Hence, concerted efforts on the maintenance and monitoring of existing gauging stations
on top of establishing new ones are needed to better understand the different eco-hydrological
environments in the basin.
N. Haregeweyn (&) A. Tsunekawa D.T. Meshesha M. Vanmaercke

Arid Land Research Center, Tottori University, 1390 Hamasaka, e-mail: Matthias.Vanmaercke@ees.kuleuven.be
Tottori 680-0001, Japan
M. Tsubo
e-mail: nigussie_haregeweyn@yahoo.com
Institute for Soil, Climate and Water, Agricultural Research
A. Tsunekawa Council, Pretoria, South Africa
e-mail: tsunekawa@alrc.tottori-u.ac.jp e-mail: TsuboM@arc.agric.za
D.T. Meshesha J. Nyssen
e-mail: deremesh@yahoo.com Geography Department, Gent University, Ghent, Belgium
e-mail: jan.nyssen@ugent.be
N. Haregeweyn A. Zenebe
Department of Land Resources Management and Environmental E. Adgo
Protection, Mekelle University, Mekelle, Ethiopia Bahir Dar University, Bahir Dar, Ethiopia
e-mail: amanuelza@yahoo.com e-mail: enyewadgo@gmail.com
J. Poesen M. Vanmaercke
Department of Earth and Environmental Sciences, Physical and
Regional Geography Research Group, KU Leuven, Heverlee,
Belgium
e-mail: jean.poesen@ees.kuleuven.be
228 N. Haregeweyn et al.

Keywords
Sediment yield Reservoir sedimentation Channel stabilization Blue Nile River
13.1 Introduction The first one is using direct (manual) methods such as staff
gauge readings and periodic manual runoff and suspended
Sediment yield (SY) is the net result of soil erosion and sediment concentration (SSC) sampling, while the second
sediment deposition processes. It is defined as the total sed- one is by indirect methods such as automatic data loggers/
iment outflow from a catchment, measurable at a point of a sensors for measuring water depths. Then, for both cases,
reference and for a specified period of time (Vanoni 1975). It stage-flow and flow-sediment concentration rating curves are
can be expressed in absolute terms (e.g., t year−1) or in area- established for each station in order to calculate SY or SSY
specific terms (specific sediment yield—SSY) (e.g., for the corresponding catchment.
t km−2 year−1). In total, there are 409 river gauging stations spread over
Accurate estimation of SY is important for a number of the 13 major river basins of the country (Fig. 13.1). How-
applications such as the identification of net erosion (denu- ever, only 28 of these stations have relatively well-recorded
dation) rates within catchments (Walling 1994), the design data on sediment concentration and runoff discharge
of planned hydraulic structures so that the actual lifetime of a (Redeco 2002). The same study estimated suspended sedi-
reservoir or pond can meet the requirements, delineation of ment transport of selected rivers under the project “Assess-
target areas which need soil conservation measures (Van ment and monitoring of erosion and sedimentation problems
Rompaey et al. 2005), and also better understanding of the in Ethiopia” based on the existing MoWR data for Ethiopian
impacts of past land-use or climatic changes (e.g., Ambers rivers, resulting in a database with SY and runoff estimations
2001; Vanmaercke et al. 2011). However, as in many for 230 stations, distributed over seven main river basins of
developing countries, data on catchment SY for Ethiopia are the country.
not only limited in number, but they are also often not very However, the data of these stations result from a large
reliable. The lack of a sufficient local database on SY and number of single samplings taken for different purposes,
adoptable SY models have been a problem for reservoir randomly in time and space. The majority of the data at these
designers, who have used various approaches to address stations consist of only few observations made over years or
the sedimentation problem in planning new reservoirs even decades. In addition, it was found that in most cases,
(Haregeweyn et al. 2006). This has resulted in a risky or the available hydrological information together with the data
uneconomical design of dams. on the suspended sediment loads was also restricted to a
In this chapter, we present an overview of SY monitoring limited number of instant observations taken together with
programs and SY variability at various spatial scales in the suspended sediment measurements. Due to these weak
Ethiopia. First, an overview of SY monitoring programs and points of the dataset, it was not possible to consider the
studies on SY variability carried out by the Ethiopian Min- entire data for the evaluation process (Redeco 2002). Hence,
istry of Water Resources (MoWR) as well as by own estab- only about 28 stations which have records covering wet, dry,
lished stations is discussed. Second, rates of reservoir and transition seasons and that produce a sufficient number
sedimentation for 14 dam catchments and the SY variability of observations were used for the analysis (Table 13.1). The
at various spatial scales and its impact on reservoirs are catchments have a drainage area ranging between 69 and
discussed. 11,690 km2 and a median of 248 km2.
To estimate the sediment concentration at a given
moment and location, a two-parameter regression approach
13.2 SY Monitoring Program was adopted:
by the Ethiopian Ministry of Water
Resources LogðSSCÞ ¼ Log a þ bðLog QÞ ð13:1Þ
The Ethiopian MoWR is the main responsible body for where SSC is suspended sediment concentration (mg l−1)
identifying, establishing, and managing stations designed for and Q is river discharge (l s−1).
measuring runoff and sediment discharge from major river The performance of the fitted model for each gauging
basins and their tributaries in Ethiopia (Fig. 13.1). At those station varies significantly with the coefficient of determina-
stations, the MoWR monitors water depth, stream velocity, tion (R2) ranging between 0.01 at Aleta Wondo Kolla and 0.67
runoff discharge, and sediment concentration into two ways. at Gacheb, with a median value of 0.17 (Table 13.1). Hence,
13 Sediment Yield Variability at Various Spatial Scales … 229
Fig. 13.1 River gauging stations for the main Ethiopian River Basins owned and run by the Ethiopian Ministry of Water Resources (MoWR
database); (n = 409)
much of the SSC variability remained unexplained by flow records over those collected at the Ethiopian Shelkole station.
variation; despite this, the SSY for each station (Table 13.1) The data from Shelkole station are marred with missing data
was calculated on the base of these weak regression models. throughout the period 1961–1978, while no single observa-
Therefore, the data from MoWR gauging stations need to be tion was recorded during the period 1979–1998.
supplemented with site-specific measurements for detailed Furthermore, a trend comparison between the daily
water resources planning and research purposes. observations at El Diem and Shelkole indicates an important
The lack of reliable hydrological data has been raised in association between both stations. River flow rate at El Diem
several forums in Ethiopia. For example, Humphreys et al. explained about 61 % the flow rate at Shelkole station
(1997) stressed that the planning and design of the Tekeze (Eq. 13.2 and Fig. 13.4).
medium-scale hydropower dam in Tigray, covering a

drainage area of 30,000 km2, has been constrained by QShelkole ¼ 0:72 QEldiem þ 420 R2 ¼ 0:61; n ¼ 198 ð13:2Þ
scarcity of local hydrological data and lack of appropriate
models, which forced them to depend on regional-scale data, where QShelkole and QEldiem are Blue Nile River flow rates
resulting in risky or uneconomical design of the dam. (m3 s−1) at Shelkole station in Ethiopia and at El Diem
Humphreys et al. (1997) recommended that accurate and station in Sudan, respectively.
regular measurements of flow discharge and sediment
transport should be implemented in the future.
The overall quality of the MoWR database is also dem- 13.3 Other Initiatives to Address
onstrated by a comparison of runoff discharge data from two the Sediment Yield Data Gap
adjacent gauging stations separated by an air distance of about in the Ethiopian Highlands
100 km on the Blue Nile River, but on the both sides of the
Ethiopia–Sudan border (compare Figs. 13.2 and 13.3). The Apart from the measurements of the MoWR discussed
data from the El Diem station in Sudan are superior in terms of above, there have been a few initiatives undertaken to
length of measuring period, continuity, and consistency of address the lack of reliable SY data. One notable example of
Table 13.1 Sediment yield data of selected river gauging stations (reprocessed after Redeco 2002)
Regression model: (Eq. 13.1) Catchment area Mean flow conditions
Station name Log a a b R 2
(km2) Q (l s−1) SSC (mg−1) SSY (t year−1 km−2)
Abbay River Basin
Yeda (Ambera) 2.57 369.66 0.19 0.12 125 33,136 2,589 902
Jedeb (Amanuel) 1.23 16.83 0.45 0.30 305 80,852 2,753 959
Muga (Bichena) 1.25 17.84 0.40 0.43 375 99,409 1,745 608
L-Chemoga (D. Markos) 2.56 362.24 0.09 0.04 364 96,493 1,010 352
Gudla (Dembecha) 1.22 16.62 0.37 0.47 242 64,152 968 337
Temcha (Dembecha) 1.61 40.33 0.25 0.16 406 107,627 721 251
Birr (Jiga) 1.35 22.28 0.44 0.42 978 17,860 1,697 41
Buchiksi (Metekel) 2.06 115.88 0.03 0.04 106 28,100 160 56
Dura (Metekel) 0.66 4.54 0.44 0.44 539 142,884 821 286
Ardie (Metekel) 1.56 36.37 0.21 0.07 219 58,055 348 121
Azuari (Motta) 1.07 11.86 0.47 0.33 209 55,404 1,936 674
Teme (Motta) 1.84 69.60 0.27 0.17 156 41,434 1,219 425
Fittom (Tilli) 0.90 7,930.00 0.44 0.56 282 74,755 1,042 363
Awash River Basin
Berga (Addis Alem) 1.73 54.20 0.15 0.09 248 11,633 216 13
Akaki 0.30 1.99 0.60 0.24 884 17,691 718 19
Holeta 1.81 64.27 0.16 0.06 119 5,679 250 16
Awash (Hombole) 0.55 3.56 0.62 0.40 7,656 112,355 5,042 97
Awash (Methara) 0.57 3.71 0.52 0.16 1,641 10,239 431 4
Mojo (Mojo) 2.56 365.60 0.31 0.12 1,264 13,891 6,902 100
Wonji (Wonji) 0.80 6.37 0.53 0.17 11,690 111,793 2,914 37
Baro Akobo River Basin
Gacheb (Mizan Teferi) 1.36 22.65 0.16 0.69 79 12,325 100 21
Genale Dawa River Basin
Woyib (Agarfa) 1.77 58.21 0.04 0.03 7,719 270,011 90 4
Omo-Ghibe River Basin
Sokjie (Areka) 1.10 12.65 0.47 0.18 103 23,602 1,495 450
Sheta (Bonga) 0.51 3.24 0.45 0.69 191 43,768 397 120
Amara (Sheboka) 1.57 36.90 0.27 0.20 69 15,811 497 150
Rift Valley
Aleta Wondo Kolla 2.34 220.80 −0.09 0.01 206 21,162 95 13
Bedessa (Dilla) 1.78 59.98 0.14 0.02 81 8,321 208 28
Gelana (Yirgalem) 1.17 14.86 0.32 0.08 141 14,485 322 44
Q river flow rate; SSC suspended sediment concentration; SSY specific sediment yield
such effort is the research collaboration between Belgian 13.3.1 Sediment Yield Study Based
Universities and Ethiopian Universities on some selected on Reservoir Sediment Survey
tributaries of the Tekeze and Abay (upper Blue Nile River) Programs
basins. Following the launching of different collaboration
projects, more accurate SY and/or runoff discharge data for The Zala Daget Research Project (a Flemish Inter University
24 catchments ranging from 1 to 5,000 km2 have become Council Own Initiative Cooperation Project between the
available based on reservoir sediment surveys, as well as on Catholic University of Leuven, the Belgian Royal Museum
runoff and SSC-measuring campaigns at gauging stations. for Central Africa, the Ethiopian Mekelle University and the
These initiatives are discussed below. Relief Society of Tigray) had been undertaking research
Fig. 13.2 Scatter plot showing

river flow rate data at Shelkole
station on the Ethiopian Blue Nile
River. Missing values are
indicated with the value −999
Fig. 13.3 Scatter plot showing

river flow rate data at El Diem
station on the Sudanese Blue Nile
River
Fig. 13.4 Blue Nile River flow

rate at El Diem station (Sudan)
versus Shelkole station (Ethiopia)
during the period 1965–2005
under the theme “Fighting desertification in the Tigray carried out for 13 micro-dam catchments located in a radius
Highlands, Ethiopia: lessons to be learnt from successes and of 120 km from Mekelle, Tigray’s regional capital
failures of soil erosion control measures” during the period (Fig. 13.5). The 13 reservoirs were created by constructing
2001–2007. Part of this project consisted of assessing and earth embankments to harvest seasonal runoff. The stored
modeling of reservoir sedimentation rates (SRs) and their runoff is then used later in the season for supplementing the
impacts in the north-Ethiopian highlands. For this purpose, a rain (when dry spells occur) and/or for full irrigation during
detailed survey of deposited sediment in reservoirs was the dry part of the year.
Fig. 13.5 Location of the study area (Tigray) in northern Ethiopia and 3 Sewhimeda, 4 Gereb Segen, Grashitu, Mejae, Maideli, Gum Selasa,
the studied reservoirs (modified after Haregeweyn et al. 2013). Dots 5 Adiakor, 6 Adihilo, 7 Agushella, 8 Endazoey, 9 Adikenafiz. Numbers
represent reservoirs visited, numbers with black fill represent reservoirs for without fill represent rainfall stations of the Ethiopian Meteorological
which sediment yield was measured: 1 Gindae, 2 Gereb Shegel, Authority: 1 Hawzen, 2 Yechila, 3 Wukro, 4 Mekelle airport, 5 Adigudom
The sediment thickness was measured using meter sticks sediment profile pits, and topography of the current reservoir
by opening spatially distributed 15–39 profile pits in the bottom were mapped using a theodolite (Fig. 13.6). A digital
reservoir. The deposited sediment boundary, location of elevation model (DEM) with a resolution of 1 m for the
Fig. 13.6 Part of the layout of the 26 profile pits at Gereb Shegel in Adihilo reservoir (right): (1) theodolite readings, (2) sediment
reservoir where sediment deposits are up to 3 m thick near the dam boundary, (3) profile pits, and (4) dam (modified after Haregeweyn
(left) and details of topographic mapping and sediment profile pit layout et al. 2005)
sediment thickness was created from theodolite points, tak- However, at this stage only SY data for Geba catchment
ing measured sediment thickness as z value, from which (Vanmaercke et al. 2010) are available for this discussion.
sediment volume was extracted in IDRISI® environment. In the Geba catchment, runoff and suspended sediment
More details of the reservoir survey and computation of SY discharges were monitored at ten gauging stations
values can be found in Haregeweyn et al. (2006). (Fig. 13.8). The fieldwork was carried out during the rainy
A similar reservoir sediment survey initiative was seasons (July–September) of 2004–2007. A digital pressure
reported in Haregeweyn et al. (2012) who analyzed the SR transducer (TD-diver) was installed at each station, record-
and potential mitigation strategies for the Angereb catch- ing the (water + atmospheric) pressure every 10 min.
ment in NW Ethiopia. In this case, the sediment volume was Barometric pressure transducers were installed to allow
estimated by subtracting the successive reservoir’s DEMs correction for the atmospheric pressure of the TD-divers in
generated based on point (x, y, z) data collected by bathy- these stations and their nearby stations.
metric surveys conducted in 2005 and 2007 (Fig. 13.7). In 2004, measurements were carried out two to three
times per week, whereas in 2005–2007 they were made at
least daily and also when a major change in flow depth was
13.3.2 Sediment Yield Study Based observed. More details on the instrumentation, monitoring,
on Hydrometric Stations Monitoring and sampling of flow and sediment and calculations of SY
Programs can be found in Vanmaercke et al. (2010) whereas flow
discharge data are reported by Zenebe et al. (2013).
The “Research and Capacity Building for Sustainable
Development in the Drylands of Ethiopia”—a collaboration
project between Mekelle University and Flemish Universities 13.3.3 SY Variability at Various Spatial
funded by the VLIR Institutional University Cooperation Scales and Its Controlling Factors
(IUC) Program (2003–2012)—has conducted studies on
various themes, including land resources in the Geba catch- The sediment survey on 14 reservoirs, including also Ang-
ment, an upper tributary of the Tekeze River (Fig. 13.8). ereb in the northwest Ethiopia, shows that there is a large
Within another VLIR-Own initiative project with Bahir Dar spatial variation in SSY among catchments: from 189 to
University in NW Ethiopia under the theme “Water and 2,927 t km−2 year−1, with an average SSY of 1,013
sediment budgets of Lake Tana for optimization of land (±735) t km−2 year−1 (Fig. 13.9). Haregeweyn et al. (2008)
management and water allocation (2010–2015),” SSC and reported that spatial variability of SSY for these catchments
runoff have been monitored in 16 hydrometric stations. is positively related to average catchment slope (Av_slope)
Fig. 13.7 Angereb reservoir’s bottom surface DEMs (m a.s.l) created 2005 (a) and 2007 (b). The IDs on the 2007 map represent dry
from data obtained by echo-sounder bathymetric surveys conducted in sediment bulk density sampling locations (Haregeweyn et al. 2012)
Fig. 13.8 Location of the Geba

catchment and the 10 monitored
subcatchments: Suluh (SU),
Genfel (GE), Agula (AG), Ilala
(IL), Upper Geba (UG), May
Gabat (MY), Endaselassie (EN),
Middle Geba (MG), Upper
Tankwa (UT), Lower Tankwa
(LT) (modified after Vanmaercke
et al. 2010)
Fig. 13.9 Relationship between

area-specific sediment yield
(SSY) and sedimentation rate
(SR, expressed as % of reservoir
storage loss) versus catchment
area (A) for the 14 studied dam
catchments in northern and
northwestern Ethiopia
and negatively to the proportion of catchment area treated indicates that in steeper catchments the extent of CULs is
with soil and water conservation (SWC) structures (i.e., small and the implementation of SWC activities has positive
stone bunds with or without trenches aligned parallel to the effects. Therefore, SWC activities are less applied in (rela-
contour, as well as, check dams in gullies). Moreover, the tively) less steep and in dominantly CULs compared to
same study showed that the extent of cultivated land (CUL) steeper catchments. This implies that catchments with a high
was inversely related to both slope and SWC practices which proportion of CUL are prone to erosion processes.
Fig. 13.10 Area-specific

sediment yield (SSY) in relation
to catchment area for Ethiopia
and for some selected countries
around the world (modified after
Haregeweyn et al. 2012)
Similarly, SSY of medium-sized catchments were found channels, local slopes are often very steep and sediment
to vary significantly between 497 and 6,543 t km−2 year−1 storage on the valley bottom alluvial plain, when present, is
(Fig. 13.10). Most of the sediment export occurs during a definitely limited by its very small size. This steeper topog-
few short but intense flash floods; hence, daily SY is char- raphy counteracts the fact that the probability for sediment
acterized by a very large temporal variability. There are also deposition increases with increasing sediment transport dis-
large variations in sediment concentration and grain-size tance, leading to no significant decrease in SSY (Vanmaercke
distribution which indicate changes in sediment supply et al. 2010).
during the rainy season due to the depletion of readily The magnitude and spatial variation in SSY for Ethiopian
available sediments and the development of a vegetation catchments is large by regional and global standards
cover (Vanmaercke et al. 2010). The latter study also (Fig. 13.10), which implies that land degradation and in
reported that annual SY is mainly controlled by the occur- particular soil erosion and its associated on-site and off-site
rence and magnitude of one or a few flash floods that are impacts are severe in the Ethiopian highlands.
difficult to predict since these floods mainly depend on local
rainfall patterns.
A recent regional SY study in the Blue Nile and Atbara 13.4 Impacts of Sediment Yield
River systems by Balthazar et al. (2013) reported an
important spatial variability of SSY ranging between 4 and 13.4.1 Reservoir Sedimentation
4,935 t km−2 year−1. The same study reported that 41 % of
the observed variation in SSY can be explained by surface As a result of high SYs discussed above, the SR is very high
vegetation cover (expressed as a percentage of poorly veg- in many Ethiopian reservoirs. As much as 50 % of the res-
etated areas), rainfall intensity (expressed as the Modified ervoirs (n = 14) were found affected by extreme siltation
Precipitation Index—Fournier 1960), mean annual air tem- problems so that they would lose economic life within half
perature, and the human footprint index. of the design period (Fig. 13.9). According to Haregeweyn
Overall, measured SSY of medium-sized catchments are et al. (2006), the rapid siltation is associated with (1) poor
larger than estimations from smaller catchments (Fig. 13.10). planning of the reservoirs for the expected SY during the
To some extent, this can be explained by the geomorphic design phase, which in turn is related to lack of a sufficient
characteristics of the Ethiopian highlands. The highlands are database, appropriate methodologies to predict SY, lack of
generally deeply incised by the larger river systems which skilled and experienced persons for designing dams and
give rise to a sharp increase in altitudinal differences as reservoirs and (2) poor implementation of reservoir sediment
catchment area increases. Moreover, close to the river and catchment management strategies.
Fig. 13.11 Evidence of seepage

water downstream of Mai
Mungude dam and its effect on
vegetation establishment and
stabilization of river channel in
Tigray, Northern Ethiopia
Reservoir sedimentation problems in northern Ethiopia are first and the finer particles are carried farther into the dam
relatively severe compared to other parts of Ethiopia, the East body and remain in suspension before being deposited. This
and southeast Africa region. Haregeweyn et al. (2006) can be explained by the impact of the reservoirs to create flow
reported annual SR values of 0.18–4 % for 13 reservoirs in velocities approaching zero thus attainment of still-water
northern Ethiopia. Similar studies carried out by DFID (2004) conditions (Morris and Fan 1998).
reported SR ranging between 1–3 % in Zimbabwe and 1–4 % Most of the studied reservoirs are over designed for the
in Tanzania. DFID (2004) also reported that about 15 % of the available flow and hence consequently have sediment/flow
surveyed dams were or will be filled with sediment in less trapping efficiency close to 100 % (Haregeweyn et al. 2006,
than the design period. For comparison, the world average 2012). As a result, neither sediment nor flow is supplied to
annual SR for large dams equals 0.5–1 % (WCD 2000). the river channel located downstream of the dam. So there is
no clear sign of either aggradation or degradation. Moreover,
because of excessive seepage from the reservoirs as a result
13.4.2 Impacts of Reservoir Dams of highly pervious limestone–shale–marl-intercalated geo-
on the Local Ecology and River logical formation (Berhane et al. 2013), the immediate
Geomorphology downstream valley of the dams has shown increased level of
ground water level that led to the emergence of new springs
Apart from the rapid sedimentation discussed above, most of and improvement of vegetation cover (Fig. 13.11). This
the studied reservoirs had also problems of insufficient inflow combined effect of reduced erosion and vegetation estab-
and excessive seepage (Haregeweyn et al. 2006) that have lishment resulted in the stabilization of river channels
influence on the local river geomorphology, surface flow and/ downstream (Fig. 13.12).
or groundwater recharge, and vegetation establishment.
Geomorphological adjustments upstream of a dam are
primarily concerned with sedimentation. This sedimentation 13.5 Conclusions
processes are strongly related to the texture of the sediment
and hence its dry bulk density. A spatial correlation between SY variability in the Ethiopian highlands is very large due to
sediment dry bulk density and distance from the inlet of the the large heterogeneity of biophysical and human factors.
reservoir to dam body (see Fig. 13.6) for 13 reservoirs in This study also revealed that the rate of reservoir sedimen-
Tigray, Northern Ethiopia, yielded a strong inverse relation- tation is very high: 50 % of the reservoirs were found under
ship (r2 = 0.92) between the two with a mean dry bulk density extreme siltation problem so that they would lose their
of 1.21 t m−3 and standard deviation of 0.11 t m−3 (Harege- economic life within half of the design period. The rapid
weyn et al. 2006). This implies that coarser fractions deposit siltation is associated with poor planning and poor
Fig. 13.12 The downstream

impacts of reservoirs on sediment
load and river morphology are
evidenced at the confluence of
Illala River and its tributary
Mai-Egam in Tigray, northern
Ethiopia. The Illala River is
relatively stabilized and has
permanent base flow, even in the
dry season, which is related to the
presence of a cluster of five
reservoirs located upstream. In
contrast, Mai-Egam that is joining
from the opposite slope carries
significant sediment including
bed load though it has a much
smaller catchment, but no
reservoirs upstream. Photo Jan
Nyssen®
implementation of the reservoir sediment and catchment Fournier F (1960) Climat et érosion: la relation entre l’érosion dusol par
management interventions. l’eau et les précipitations atmosphériques. Presses Universitaires de
France, France, p 203
The availability of reliable, spatially distributed SY data Haregeweyn N, Poesen J, Nyssen J, Verstraeten G, de Vente J, Govers
reflecting the different environments remains the major bot- G, Deckers J, Moeyersons J (2005) Specific sediment yield in
tleneck to understand the dynamics and drivers of SY vari- Tigray-Northern Ethiopia: assessment and semi-quantitative mod-
ability in the region. The SY database is poor by international elling. Geomorphology 69:315–331
Haregeweyn N, Poesen J, Nyssen J, De Wit J, Haile H, Govers G,
standards. The SY assessment initiatives taken through Deckers J (2006) Reservoirs in Tigray: characteristics and sediment
institutional collaboration projects is a good start; however, deposition problems. Land Degrad Dev 17:211–230
such projects have limited capacity and a short life span, so Haregeweyn N, Poesen J, Nyssen J, Govers G, Verstraeten G, De Vente J,
they cannot produce a sustainable solution for this important Deckers J, Moeyersons J, Haile M (2008) Sediment yield variability
Northern Ethiopia: a quantitative analysis of its controlling factors.
data gap. Therefore, the MoWR should concentrate efforts on Catena 75:65–76
the maintenance and monitoring of existing gauging stations Haregeweyn N, Melesse B, Tsunekawa A, Tsubo M, Meshesha D,
on top of establishing new ones needed to reflect the different Balana BB (2012) Sedimentation and its mitigating strategies:
eco-hydrological environments in Ethiopia. a case study of Angereb reservoir, northwestern Ethiopia. J Soils
Sediment 12:291–305
Haregeweyn N, Poesen J, Govers G, Verstraeten G, de Vente J, Nyssen J,
Deckers S, Moeyersons J (2013) Evaluation and adaptation of a
References spatially-distributed erosion and sediment yield model in Northern
Ethiopia. Land Degrad Dev 24:188–204
Humphreys H, Bellier C, Kennedy R, Dokin K, (1997) Tekeze medium
Ambers KR (2001) Using the sediment record in a western Oregon hydropower report, Annex B-hydrology report. Ministry Water
flood control reservoir to assess the influence of storm history and Resour Addis Ababa, p 264
logging on sediment yield. J Hydrol 244:181–200 Morris G, Fan J (1998) Reservoir sedimentation handbook design and
Balthazar V, Vanacker V, Girma A, Poesen J, Golla G (2013) Human management of dams reservoirs and catchments for sustainable use.
impact on sediment fluxes within the Blue Nile and Atbara River McGraw-Hill, New York
basins. Geomorphology 180–181:231–241 Redeco (2002) Assessment and monitoring of erosion and sedimenta-
Berhane G, Martens K, Al Farrah N, Walraevens K (2013) Water tion problems in Ethiopia. Final report, Redeco consulting, GmbH,
leakage investigation of micro-dam reservoirs in Mesozoic sedi- Hydrology Studies Department, Ministry of Water Resources,
mentary sequences in Northern Ethiopia. J Afr Earth Sci 79:98–110 Addis Ababa, Ethiopia
DFID (Department for International Development) (2004) Guidelines for Vanmaercke M, Zenebe A, Poesen J, Nyssen J, Verstraeten G, Deckers
predicting and minimizing sedimentation in small dams. Department J (2010) Sediment dynamics and the role of flash floods in sediment
for International Development, HR Wallingford Ltd., Wallingford export from medium-sized catchments: a case study from the semi-
arid tropical highlands in northern Ethiopia. J Soils Sediments Walling DE (1994) Measuring sediment yield from river basins. In: Lal
10:611–627 R (ed) Soil erosion research methods, 2nd edn. Soil and Water
Vanmaercke M, Poesen J, Verstraeten G, Maetens W, de Vente J Conservation Society, Ankeny, pp 39–80
(2011) Sediment yield as a desertification risk indicator. Sci Total WCD (World commission of Dams) (2000) Dams and development, a
Environ 409:1715–1725 new framework for decision making, Report of the World
Vanoni VA (1975) Factors determining bed forms of alluvial streams. Commission of Dams. Earth-scan publications, London
J Hydraul Div-ASCE 101(11):1435–1440 Zenebe M, Vanmaercke M, Poesen J, Verstraeten G, Haregeweyn N,
Van Rompaey AJJ, Bazzoffi P, Jones RJA, Montanarella L (2005) Haile M, Amare K, Deckers J, Nyssen J (2013) Spatial and temporal
Modeling sediment yields in Italian catchments. Geomorphology variability of river flows in the degraded semi-arid tropical mountains
65:157–169 of northern Ethiopia. Zeitschrift für Geomorphologie 57(2):143–169
Climatic and Hydrologic Changes in Northern
Ethiopia in the last 3,500 Years: Evidence 14
from the Geomorphic, Stratigraphic,
and Geochemical Archives of Hayk Lake
Massimiliano Ghinassi, Marco Benvenuti, Filippo D’Oriano,

and Marialelena Fedi
Abstract
Lake Hayk (23 km2 in surface) is located in northern Ethiopia, at an altitude of around 2,000 m
a.s.l. The Late Holocene sedimentary successions preserved along the lake margins provide a
detailed archive of the climate history of East Africa over the past 3,500 years. These
successions are represented by colluvial/stromatolitic deposits and palustrine, deltaic, and
fluvial sediments along the northern and southern coasts, respectively. During the last
3,500 years, the lake was characterized by three highstand phases, which occurred at about
3,250–3,000, 2,600–950, and 650 cal years BP—modern age (160 years BP non-calibrated).
These stages are recorded by aggradation of siliciclastic deposits both along the southern and
northern lake margins, whereas isotope data suggest that most of the stromatolites grew during
phases of intense evaporation and lacustrine contraction. The good correlation between the
Hayk Lake record and those of other late Holocene lakes of East Africa lakes allows a wide-
scale generalization of late Holocene climate variability.

Keywords
Late Holocene Climate Isotope record Stromatolites East Africa Ethiopia
14.1 Introduction 2000; Diaz et al. 2011; Xoplaki et al. 2011) which occurred
mainly after the mid-Holocene humid period (Gasse 2000,
Palaeoclimatic proxies from East African lakes (Noren et al. 2001). During the last millennium, this marked climate
2002; Viles and Goudie 2003; Mayewski et al. 2004) variability was mainly controlled by a complex interaction
highlighted the complexity of regional climatic fluctuations between ocean–atmosphere circulation determining climate
(Halfman et al. 1994; Cohen et al. 1997; Verschuren et al. gradients over the equatorial Africa (Russel et al. 2007). The
Lake Naivasha record in Kenya, representing one of the
most complete climate proxies of this time interval, shows
that the Little Ice Age in East Africa was characterized by
M. Ghinassi (&) humid conditions interrupted by three main droughts (Ver-
Dipartimento di Geoscienze, Università di Padova,
V. Gradenigo 6, 35121 Padua, Italy
schuren et al. 2000), which had a remarkable impact on
e-mail: massimiliano.ghinassi@unipd.it social and economic development.
M. Benvenuti
The coastal landforms of Hayk Lake (Fig. 14.1) in
Dipartimento di scienze della Terra, Università di Firenze, Northern Ethiopia and their clastic deposits and biogenic
V. G. La Pira 4, 50121 Florence, Italy carbonates (i.e., stromatolites) are described and interpreted
F. D’Oriano in detail for their geomorphological, stratigraphic and sedi-
GEOPHI srl, Area della Ricerca CNR Via Gobetti 101, mentological significance. Lacustrine stromatolites are
40129 Bologna, Italy common features of the East Africa Rift lakes, representing
M. Fedi an archive suitable for isotopic dating, reconstruction of
INFN Sezione di Firenze, via Bruno Rossi 1, former lake levels and paleolimnological conditions
50019 Sesto Fiorentino, FI, Italy
240 M. Ghinassi et al.
Fig. 14.1 Panorama of the Lake Hayk from the Northern coast
(Casanova 1986, 1994). This paper summarizes the results respectively (Demille 2000). Lamb et al. (2007) calculated
of a previous study (Ghinassi et al. 2012) and aims at a water balance pointing to a groundwater outflow of
extending the knowledge on Late Holocene climatic fluctu- 1.4 × 106 m3 year−1 (3 % of the total) and evaporative
ations in East Africa. losses of 33 × 106 m3 year−1 (65 % of the total).
14.3 Sedimentary Succession

14.2 Geological and Geomorphological and Chronology
Settings
Sedimentary succession of the northern (Uarababo) and
Lake Hayk is located on the eastern margin of the Ethiopian
southern (Ankarka River) margins formed in two different
Highland (Fig. 14.2a) at about 2,000 m a.s.l. Bedrock is
morphological and depositional settings. The Ankarka River
represented by the *1,000-m-thick Termaber Basalt (Merla
succession (Fig. 14.3) accumulated in a low-gradient setting
et al. 1973; Zanettin and Justin-Visentin 1973) that accu-
and is characterized by swamp and fluvio-deltaic deposits,
mulated during the stage of the Ethiopian plateau shield
whereas the Uarababo succession (Fig. 14.4) formed along
volcanism (Miocene to Early Pliocene). The lake occupies a
steep coasts and is made of colluvial fan deposits and stro-
small fault-controlled depression bounded by N–S trending
matolitic biostromes. Radiocarbon ages obtained from dif-
faults which crosscut two minor set of faults with NE–SW
ferent depositional units are summarized in Table 14.1.
and NNW–SSE trends (Fig. 14.2c).
Hayk Lake is a hydrologically closed basin with a
surface area of about 23 km2 (Fig. 14.2c). The northern 14.3.1 Fluvio-deltaic Deposits of the South-
lake margin is characterized by steep, rocky slopes, ern Coast: The Ankarka River
whereas the southern margin hosts a wide coastal plain, Succession
which is drained by the only perennial inflow to lake, the
Ankarka River (Fig. 14.2b). Since 1938, the maximum The Ankarka River succession (Fig. 14.3), up to 7 m thick, is
depth of the lake, occurring in the north-western sector, made of fluvial and coastal deposits organized into four
has decreased from 88.8 m (Baxter and Golobitsh 1970) to main, unconformity-bounded sedimentary units (Ak1–Ak4).
81 m (Lamb et al. 2007). Occasionally, following very • Unit Ak1 is up to 3 m thick, made of swampy coastal
intense precipitation, the nearby Ardibo Lake (Fig. 14.2b) mud (Fig. 14.3d), sandwiched between two vertisols,
can overflow into the Hayk Lake expanding its surface which are labeled as Vertisol A and B, respectively
catchment by about three times. The region has a subhu- (Fig. 14.3b). Coastal muds are laminated and show an
mid, tropical climate with an average annual rainfall and a alternation between millimeter-thick dark and light lam-
mean annual temperature of 1,158 mm and 18 °C, inae. The light laminae are composed of authigenic
14 Climatic and Hydrologic Changes in Northern Ethiopia … 241
channel sands (Fig. 14.3e). Distributary channel deposits

occur at an elevation of about 8.5 m a.l.l. On the whole,
these deposits yielded radiocarbon ages ranging between
1,200 and 1,000 cal year BP (Fig. 14.5).
• Unit Ak3 is about 6 m thick and consists of fluvial
gravels passing lakeward into their deltaic equivalent.
Fluvial gravels range from planar- to cross-stratified and
form amalgamated channelized units up to 1 m thick.
Fluvial deposits are organized into four subunits bounded
by unconformity surfaces of lower rank (Ak3a–Ak3d in
Fig. 14.3b). These surfaces are characterized by a marked
erosional relief (Fig. 14.3b) and are associated with
repeated phases of fluvial incision. Subunits Ak3b, Ak3c,
and Ak3d are dated about 450 cal year BP,
400 cal year BP, and modern age (160 year BP non-
calibrated), respectively (Fig. 14.5). Deltaic deposits are
poorly exposed and consist of high-angle, clinostratified
gravel and sand (Fig. 14.3c) accumulated in a Gilbert-
type delta front. The lack of continuous outcrop prevents
the fine-tuning between the deltaic gravels and subunits
Ak3a–d. The fluvial surface F3 marks the top of subunit
Ak3d (see Sect. 14.3.3) and occurs about 9–10 m a.l.l.
• Unit Ak4 consists of fluvial deposits perched along the
flanks of the modern Ankarka River valley (Fig. 14.3a)
occurring as two subunits (Ak4a–b) capped by fluvial
surfaces F1–F2 (see Sect. 14.3.3).
14.3.2 Colluvial and Stromatolitic Deposits

of the Northern Coast: The
Uarababo Succession
The Uarababo succession, about 10 m thick, is made of col-

luvial and stromatolitic deposits (Fig. 14.4) subdivided into
four unconformity-bounded sedimentary units (Ub1–Ub4).
• Unit Ub1, at least 2 m thick, is made of colluvial gravels
Fig. 14.2 a Geographic location of Lake Hayk. b Satellite image draped by 2–10-cm-thick stromatolites (Fig. 14.4e, f). The
(Google™Earth) of Lake Hayk and Lake Ardibo. c Geological sketch
map of the Hayk Basin. Isobaths in meters are from Demlie et al. (2007)
stromatolites (hereafter referred to as the “lower stromat-
olite”) standing at a maximum elevation of 5–6 m a.l.l.,
developed on colluvial gravels and, locally, on a discon-
tinuous, 0.5–2-cm-thick, sandy transgressive lag deposit
carbonate and shell debris, whereas the dark laminae are containing gastropod shells (Melanoides sp.). Charcoal
organic rich and contain abundant remains of palustrine from the uppermost part of the colluvial gravel provided a
vegetation. Coastal muds decrease in thickness landward, radiocarbon age of about 3,500 cal year BP (Fig. 14.5).
and their thinnest part is located about 5 m a.l.l. (above • Unit Ub2 is up to 6 m thick and consists of colluvial fan
lake level). Radiocarbon dating provided an age of about delta gravels (Fig. 14.4c, e) with well-developed topset
3,200 cal year BP for these muds (Fig. 14.5). (i.e., subaerial part of the fan delta) and foreset (i.e.,
• Unit Ak2 consists of fluvial sand passing lakeward into subaqueous part of the fan delta) deposits. Topsets dip
shoalwater delta deposits. Fluvial sands are medium-fine about 35° lakeward and are made of debris flow and
grained, stacked in decimeter-thick, tabular, and gently debris fall gravels with minor muddy intercalations. Fo-
inclined beds, locally interbedded with mud. Deltaic reset deposits show a similar dip and are made of debris
deposits are made of meter- to decimeter-thick, sandy flow, debris fall, and density current deposits ranging in
and muddy tabular beds (lobes) overlain by distributary grain size from sand to gravel. The topset–foreset
Fig. 14.3 Ankarka River sedimentary succession. a The northernmost sand, Gilbert delta foreset of unit Ak3. d Laminated lacustrine deposits
outcrop of the Ankarka succession. b Unconformity-bounded subunits of unit Ak1. e Shoalwater delta deposits of unit Ak2. Note the
Ak3a–d erosionally overlying the basal muddy Ak1 unit. c Gravelly distributary channels at the top of the deltaic succession
transition deposits are commonly wave-worked and and shell fragments. Upper stromatolite A and B reach
occur about 8.5 m a.l.l. A number of samples from unit the maximum elevation at about 9 and 10.5 m a.l.l.,
Ub2 indicate continuous deposition from about 2,500– respectively. Samples from the sandy transgressive lag at
1,650 cal year BP (Fig. 14.5). the base of upper stromatolites provided an age of about
• Unit Ub3 is up to 2 m thick and is made of a stromatolitic 650 cal year BP (Fig. 14.5). Consistently, samples from
carbonates (“upper stromatolite”) with intervening col- the base of upper stromatolite A and top of upper stro-
luvial gravels. The latter allow to distinguish an upper matolite B suggest that they developed between about
stromatolite A from an upper stromatolite B (Fig. 14.4b). 750 cal year BP and the modern age (Fig. 14.5).
Lakeward, the upper stromatolite B rests unconformably • Unit Ub4 includes four subunits (Ub4a–Ub4d) of wave-
on upper stromatolite A forming a 30–40-cm-thick worked gravelly sands topped by as many lacustrine
biostrome (Fig. 14.4a, c, d) due to pinching-out of the terrace surfaces (Fig. 14.6a, b).
colluvial deposits. A transgressive sandy layer (1–5 cm
thick) rich in gastropod shells occurs at the base of upper
stromatolite A. The upper stromatolite A underwent 14.3.3 Coastal Morphological Features
subaerial exposure as indicated by the brownish color Related to the Stratigraphic
(Fig. 14.4d) due to organic matter and mud infiltration Architecture
and by dissolution cavities and then filled by well-sorted
sand and gastropod shells during the following trans- Hayk Lake coastal areas are characterized, particularly along
gression. Upper stromatolite B is light yellow and con- the northern margin, by laterally continuous, flat to gently
tains two main veneers of fine- to medium-grained sand inclined (2° to 3° lakeward) surfaces ranging from 0.8 to
Fig. 14.4 Uarababo sedimentary

succession. a Biostrome formed
by upper stromatolites of unit
Ub3. Hammer for scale is 33 cm.
b Upper stromatolite of unit Ub3
consisting of two different
portions (labeled as upper strom.
A and upper strom.B) separated
by subaerial colluvial gravels.
c Biostromes of unit Ub1 and
Ub3 separated by colluvial
deposits of unit Ub2. Biostromes
of unit Ub1 consists of upper
stromatolite B resting on upper
stromatolite A without any
interbedded colluvial gravels.
d Slab of upper stromatolite (unit
Ub3). The white line marks the
surface of subaerial exposure
separating upper stromatolite A
from B. e Erosion surface
separating colluvial fan delta
deposits of unit Ub1 from those
of units Ub2. Lower stromatolites
occur at the top of unit Ub1.
f Slab of the lower stromatolite
capping unit Ub1. Knife for scale
is 11 cm
19 m a.l.l. (Fig. 14.6a, b, d). Five clusters of values can be Ankarka delta area, terraces F1, F2, and F3 make lateral
determined by plotting the elevation of these surfaces above transition into lacustrine terrace surfaces L2, L3, and L4,
the modern lake level (Fig. 14.6d). The first three clusters respectively (Fig. 14.5). Lacustrine terrace surface L1 has no
(1.8–2, 3, and 5–6 m a.l.l.) are related to benches built of fluvial equivalent.
wave-worked gravelly sand of unit Ub4, and therefore, they
can be labeled as coastal terraces (L1, L2, and L3; Fig. 14.5).
The fourth cluster (about 8.5 m a.l.l.) coincides with the 14.4 Isotopic Data from Stromatolites
brink point (i.e., topset/foreset transition) of Ub2 colluvial
fan deltas (Fig. 14.5) and, therefore, marks the final stage of Oxygen and carbon stable isotope analyses were carried out
Ub2 colluvial fan delta progradation. The fifth cluster (about on both upper stromatolites and the lower stromatolite from
9–10 m a.l.l.) is linked with a surface capping wave-worked the northern margin of the lake (Fig. 14.7a). A 3.5-cm-thick
deposits of unit Ub4 and, similarly to the first three clusters, sample from the lower stromatolite (Fig. 14.7c) provided
is interpreted as a lacustrine terrace (L4). This surface also nine sequential samples for isotope analysis. A specimen of
fits with the maximum elevation reached by the upper 17.5 cm thick of the upper stromatolite B directly over the
stromatolite B and with the surface capping fluvial deposits upper stromatolite A (unit Ub3) was finely sampled
of unit Ak3d in the Ankarka area (Fig. 14.5). (Fig. 14.7b). 26 samples from upper stromatolite A and 30
Fluvial terraces (surfaces F1–F3) occur in the distal part from upper stromatolite B were analyzed, yielding a curve
of the Ankarka River valley (Figs. 14.3a and 14.6c) at the showing frequent oscillations of the stable isotopic compo-
top of subunits Ak4a (F1), Ak4b (F2), and Ak3d (F3). In the sition. Since Hayk Lake oxygen isotope values are more
Table 14.1 Chronological data from the Ankarka River and Uarababo district successions. (A) C14 datings on charcoal samples from the
Ankarka succession; (B) C14 datings on charcoal samples from the Uarababo succession; (C) 14C data of stromatolites samples (measured
radiocarbon ages, reservoir-corrected ages, and calibrated ages—at 95.4 % confidence level—are reported. Details about the correction applied for
the reservoir age effect can be found in Ghinassi et al. (2012)
A Sample Lab.code 14C age Cal age (68.2 %) Cal age (95.4 %)
(years BP) (Cal years BP) (Cal years BP)
Ak3d HK05.4.3a BETA209892 160 ± 40 Modern Modern
Ak3d HK05.4.3b LABEC 14Fi 1503 and 170 + 36 Modern Modern
1505
Ak3c HK05.4.2 LABEC 14Fi0139 and 0144 280 ± 30 430–290 450–280
Ak3b HK05.4.1 LABEC 14Fi0138 and 350 + 50 480–320 450–310
01143
Ak2 HK05.24.1 BETA203648 1,070 + 40 1,050–930 1,060–930
HK18.2 LABEC 14Fi0401 and 0402 1,160 ± 30 1,170–1,000 1,170–980
HK2 BETA-190827 1,170 + 60 1,180–990 1,250–950
Ak1 HK3 BETA-190829 3,070 + 60 3,360–3,220 3,390–3,090
B Sample Lab.code 14C age Cal age (68.2 %) Cal age (95.4 %)
(years BP) (Cal years BP) (Cal years BP)
Ub3 HK05.28.1-10 LABEC 14F0568 and 0575 600 ± 45 645–550 660–535
HK05.28.1 LABEC 14Fi1497 and 1504 830 ± 30 765–695 790–680
Ub2 HK 33.a1 LABFC 14Fi0968 and 0963 1,810 + 40 1,815–1,710 1,860–1,620
HK05.33.b1 LABEC 14F0392 1,750 ± 40 1,715–1,610 1,810–1,550
HK31.c LABEC 14F0400 and 0405 2,220 ± 30 2,310–2,158 2,330–2,150
HK31.b LABEC 14Fi0406 and 0409 2,340 ± 60 2,470–2,210 2,700–1,740
HK31.a1 BETA209891 2,550 ± 40 2,750–2,510 2,750–2,470
Ub1 HK05.30.1 BETA203650 3,550 ± 40 3,890–3,730 3,960–3,700
C Lab.cod Measured 14C age Reservoir effect-corrected 14C age Cal age (95.4 %) Cal age (68.2 %)
(years BP) (years BP) (Cal years BP) (Cal years BP)
STR_B2 14Fi1628 600 ± 75 185 ± 100 Modern Modern
STR_B1 14Fi1623, 14Fi1625 1,450 ± 45 880 ± 80 930–680 910–730
sensitive to an evaporation-induced isotopic enrichment In the upper stromatolite A, δ18O ranges between of
(Talbot 1990; Li and Ku 1997; Yuan et al. 2011) rather than +2.99 and +6.24 ‰ and, notwithstanding a minor decrease
groundwater inflow (Lamb et al. 2007), recorded high values within the first 15 mm, it shows an overall increasing trend.
of δ18O are interpreted as indicating surface water evapo- δ13C ranges between +3.71 and +7.51 ‰ and shows a good
ration during episodes of reduced rainfall. covariance with δ18O (Fig. 14.7b). This isotope record
In the lower stromatolites, δ18O ranges between +1.68 suggests that the upper stromatolite A grew in a closed lake
and +5.05 ‰ and shows an overall increase from the base to during a stage of progressive evaporative enrichment. In the
the top of the sampled biostrome (Fig. 14.7c), indicating upper stromatolite B, δ18O varies from a minimum of
evaporative enrichment. δ13C ranges from +0.67 to +4.65 ‰ +6.06 ‰ to a maximum of +7.75 ‰ and, within an overall
and shows a good covariance with δ18O, suggesting a bal- rising trend of δ18O, a short stage of slight increase can be
ance of the dissolved inorganic carbonate in the lake water detected. δ13C fluctuates between +5.08 and +7.39 ‰, and
with atmospheric CO2 (Li and Ku 1997), hinting to a closed- the similar pattern between δ18O and δ13C (Fig. 14.7b) still
lake condition during stromatolite formation (Talbot 1990). indicates a closed lake.
Fig. 14.5 Schematic N–S cross section across the Uarababo and Ankarka sedimentary successions. Depositional environment and age of different
units are shown
The overall evaporative enrichment in δ18O documented fall, culminating in the subaerial exposure of the biostromes
by the upper stromatolites well fits the data obtained by and the development of Vertisol A in the Ankarka area.
Lamb et al. (2007) from coeval deep-lacustrine deposits On the whole, Ub1-Ak1 units record a climatically forced
cored in the SW part of the lake. regressive–transgressive cycle that finds a regional correla-
tion with other lacustrine archives. The Hayk lowstand
recorded by the Ub1 (colluvium) and Ak1 (Vertisol B)
14.5 Discussion: Lacustrine Oscillations deposits is older than 3,250 cal year BP and can be tenta-
and Comparison with Other tively correlated with a widespread dry episode centered on
Climate-Proxy Records 4,200–4,000 cal year BP (Barker et al. 2004). In East Africa,
the same event caused the regression (Fig. 14. 8b) of Lakes
The oldest part of the Hayk succession documents a lacus- Bosumtwi (Talbot and Johannessen 1992) and Ziway-Shala
trine lowstand, which was associated with colluvial sedi- (Gasse and Street 1978; Benvenuti et al. 2002), the isolation
mentation and soil formation along the northern and southern of Lake Albert (Talbot and Brendeland 2001), and a decline
coasts, respectively. The first transgression occurred at about in discharge of the White Nile (Hassan 1997). The following
3,250 cal year BP (Fig. 14.8a) and resulted in a lake level highstand, at about 3,200 cal year BP (Fig. 14.8a), is again
rising up to about 5 m above the modern level. In the Ua- recorded also (Fig. 14.8b) in Lakes Bosumtwi (Talbot and
rababo area, this transgression called a halt to colluvial sed- Johannessen 1992), Abe (Gasse and Street 1978) and Ziway-
imentation and caused the growth of the lower stromatolite Shala (Gillespie et al. 1983).
(Ub1), whereas in the Ankarka River zone, it resulted in the The lake-level fall, subsequent to this first stage and
establishment of a coastal marsh (Ak1). The evaporative possibly greater than 5 m, was followed by renewed lake
enrichment of 18O recorded in the lower stromatolites sug- flooding between 2,600 and 950 cal year BP. During this
gests that following this transgression, their development interval, the lake level rose up to about 8.5 m a.l.l.
progressed under increasing evaporation conditions (i.e., (Fig. 14.8a), providing the space for accumulation of col-
dryer climate). This trend then forced subsequent lake-level luvial fan deltas (Ub2) and deltaic deposits (Ak2) in the
Fig. 14.6 a Lacustrine terraces

L2 and related deposits, occurring
at about 3 m a.l.l. in the Uarababo
district. b Lacustrine terraces L1,
occurring at about 2 m a.l.l. and
overlying upper stromatolite
deposits (Ub3) in the Uarababo
district. c Fluvial terraces
(surfaces F2 and F2) and related
deposits along the right-hand
flank of the Ankarka River valley.
d Vertical distribution of the
morphological surfaces measured
along the lake margins
Uarababo and Ankarka areas, respectively. During the same a.l.l. (Fig. 14.8a), which determined the fluvio-deltaic
time span, also lakes Victoria (Stager et al. 2003), Abe aggradation (Ak3) in the Ankarka area, whereas the sandy
(Gasse and Street 1978), Bosumtwi (Talbot and Johannessen transgressive lag at the base of the upper stromatolite (Ub3)
1992), Ziway-Shala (Gillespie et al. 1983), and Tanganyika formed in the Uarababo area. Similarly to the previous
(Cohen et al. 1997) experienced high levels, hinting to a events, this regressive–transgressive cycle has regional
regional hydrologic signature of a moist phase (Fig. 14.8b). equivalents too. The Hayk regression at 950 cal year BP is
A further lake-level fall occurred between 950 and matched (Fig. 14.8b) in lakes Tanganyika (Cohen et al.
600 cal year BP, with a subsequent shoreline regression 1997), Naivasha (Verschuren et al. 2000), and Victoria
bringing the level at about 8–9 m lower than the previous (Stager et al. 2005) and represents the establishment of dry
highstand (Fig. 14.8a). The third lacustrine transgression climate which affected East Africa during the European
started at 600 cal year BP with a level rise of about 9–10 m Medieval Climatic Anomaly.
Fig. 14.7 a Location of the

samples showed in inset C and D.
b Variations of δ18O and δ13C in
the stromatolites of unit Ub3
(upper stromatolites). Note the
two main sandy veneers
interbedded within the biostrome.
c Changes of δ18O and δ13C in
the lowers stromatolites (unit
Ub1)
The last lacustrine rise started about 600 cal year BP and the signature of these oscillations may be represented by the
established highstand conditions which persisted until pres- two sandy veneers within the upper stromatolites B
ent (160 year BP not calibrated). This highstand phase was (Fig. 14.7b). During these two short-lived lowstands, the
punctuated by three main regressive episodes, which are upper stromatolite B may have been weakly reworked by
recorded in the Ankarka area (Ak3a–d) and in the sedi- wave winnowing witnessed by the thin sandy horizons.
mentological and isotope signatures of the upper stromato- At a regional scale, this last phase of Lake Hayk devel-
lite. Though not directly dated, the subunit Ak3a is opment corresponds to the Little Ice Age as it is synchronous
considered to have been formed during lake-level rise, (Fig. 14.8b) with a similar evolution observed in Turkana
synchronously with the formation of the sandy transgressive (Halfman et al. 1994), Abe (Gasse and Street 1978), Tang-
lag and the onset of upper stromatolite A growth along the anyika (Cohen et al. 1997), Naivasha (Verschuren et al.
northern margin (Fig. 14.9). The upper stromatolite A started 2000), and Victoria (Stager et al. 2005) lakes. The three
to develop during the lake-level rise as indicated by the δ18O short-lived episodes of Lake Hayk level fall may find a
trend of the three basal samples (Fig. 14.7b). Nevertheless, correspondence with the main droughts (Wamara, Nyaru-
its successive growth occurred under a progressive enrich- banga, and Lapanarat-Mahlatule) identified by Verschuren
ment in δ18O, likely related to a lake level lowering under et al. (2000) in Lake Naivasha (Fig. 14.9) and by Lamb et al.
dryer climate conditions (Fig. 14.7b). The lake shrinkage (2007) in a lacustrine sediment core collected in the bottom
culminated with the subaerial exposure of the upper stro- of Hayk Lake (Fig. 14.9).
matolites A and accumulation of colluvial deposits. In the The progressive lacustrine contraction, which has led to
meantime, the subunit Ak3a in the Ankarka area was the modern lake, is geochemically recorded in the uppermost
undergoing erosion. Subunits Ak3b, Ak3c, and Ak3d doc- part of the upper stromatolite B (i.e., further increase in δ18O
ument three highstands separated by two episodes of lake values) and in the development of lacustrine (L1–3) and
contraction between 450 cal year BP and modern age fluvial (F1–3) terraces. This post-Little Ice Age lake
(160 year BP not calibrated). These oscillations are not shrinking, together with the records from lakes Naivasha
evidenced in the isotope proxy of the upper stromatolites B, (Verschuren et al. 2000) and Turkana (Halfman et al. 1994),
likely because they were too rapid to be recorded within the highlights that a progressively warmer and dryer conditions
overall trend of δ18O evaporative enrichment. Nevertheless, have been established over East Africa in recent times.
Fig. 14.8 a Relative changes in

Lake Hayk level occurred during
the past 3,500 years.
b Comparison between Lake
Hayk record and proxies from
Late Holocene lacustrine
successions of East Africa
highstands and those occurred in some of the most studied

14.6 Conclusions East African lakes (e.g., Naivasha, Tanganika, Victoria,
Turkana, Abe, Ziway-Shala, Bosumtwi, and Abyiata) sup-
The Hayk Lake succession documents the main climate ports the hypothesis of their climatic origin. Furthermore, the
changes occurred in the central Ethiopian highlands over the high-frequency fluctuations documented during the late
past 3,500 cal year BP. The main episodes of increase in highstand phase in Lake Hayk can be correlated with three
rainfall which caused lacustrine rises are dated at about main drought phases (Wamara, Nyarubanga, and Lapanarat-
3,250–3,000, 2,600–950, and 650 to modern cal year BP, Mahlatule), which were previously identified by Verschuren
respectively. The correlation between such lacustrine et al. (2000) in Lake Naivasha.
Fig. 14.9 Comparison between

lacustrine oscillation documented
by units Ak3 and Ub3 (central
columns), Lake Naivasha
documentation (Verschuren et al.
2000; right-hand column) and
isotope record by Lamb et al.
(2007; left-hand column)
References Díaz HF, Trigo RM, Hughes MK, Mann ME, Xoplaki E, Barriopedro
D (2011) Spatial and temporal characteristics of climate in medieval
times revisited. B Am Meteorol Soc 92:1487–1500
Barker PA, Talbot MR, Street-Perrott FA, Marret F, Scourse J, Odada Gasse F (2000) Hydrological changes in the African tropics since the
EO (2004)Late Quaternary climatic variability in intertropical last glacialmaximum.Quat Sci Rev 19:189–211
Africa. In: Battarbee RW,Gasse F, Stickley CE (eds.) Past climate Gasse F (2001) Hydrological changes in Africa. Scince 292:2259–2260
variability through Europe and Africa.Springer, Dordrecht, pp 117– Gasse F, Street FA (1978) Late Quaternary lake-level fluctuations and
138 environments of the northern rift valley and Afar region (Ethiopia
Baxter RM, Golobitsh DL (1970) A note on the limnology of Lake and Djibouti). Palaeogeogr Palaeocl 24:279–325
Hayq, Ethiopia. Limnol Oceanogr 15:144–148 Ghinassi M, D’Oriano F, Benvenuti M, Awramik S, Bartolini C, Fedi
Benvenuti M, Carnicelli S, Belluomini G, Dainelli N, Di Grazia S, M, Ferrari G, Papini M, Sagri M, Talbot M (2012) Shoreline
Ferrari GA, Iasio C, Sagri M, Ventra D, Balemwald, A, Seifu K fluctuations of Lake Hayk (northern Ethiopia) during the last
(2002) The Ziway-Shala lake basin (Main Ethiopian Rift, Ethiopia): 3,500 years: geomorphological, sedimentary, and isotope records.
a revision of basin evolution with special reference to the Late Palaeogeogr Palaeocl 365, 366:209–226
Quaternary. J Afr Earth Sciences 35:247–269 Gillespie R, Street-Perrott FA, Switsur R (1983) Post-glacial arid
Casanova J (1986) East African Rift stromatolites. Geol Soc Spec Publ episodes in Ethiopia have implications for climate prediction.
London 25:201–210 Nature (London) 306:680–682
Casanova J (1994) Stromatolites from the East African Rift: a synopsis. Halfman JD, Johnson TC, Finney BP (1994) New AMS dates,
Phanerozoic stromatolites II. Springer, Berlin, pp 193–226 stratigraphic correlations and decadal climatic cycles for the past
Cohen AS, Talbot MR, Awramik SM, Dettman DL, Abell P (1997) 4 ka at Lake Turkana, Kenya. Palaeogeogr Palaeocl 111:83–89
Lake level and paleoenvironmental history of Lake Tanganyika, Hassan FA (1997) Holocene paleoclimates of Africa. Af Archaeol Rev
Africa, as inferred from late Holocene and modern stromatolites. 14:213–230
Geol Soc Am Bull 109:444–460 Lamb HF, Leng MJ, Telford RJ, Ayenew T, Umer M (2007) Oxygen
Demille M (2000) Hydrology, hydrogeology and hydrochemistry of and carbon isotope composition of authigenic carbonate from an
lake Haiq-Hardibo sytem. Unpublished MSc thesis, Addis Ababa Ethiopian lake: a climate record of the last 2,000 years. The
University, Ethiopia Holocene 17:515–524
Demlie M, Ayenewb T, Wohnlicha S (2007) Comprehensive hydrolog- Li HC, Ku TL (1997) δC–δ18O covariance as a paleohydrological
ical and hydrogeological study of topographically closed lakes in indicator for closed-basin lakes. Palaeogeogr Palaeocl 133:69–80
highland Ethiopia: The case of Hayq and Ardibo. J Hydrol 339: Mayewski PA, Rohling EE, Stager JC, Karle’n W, Maasch KA,
145–158 Meeker LD, Meyerson EA, Gasse F, Van Kreveld S, Holmgren K,
Lee-Thorp J, Rosqvist G, Rack F, Staubwasser M, Schneider RR, Talbot MR, Brendeland KI (2001) Strontium isotopes as palaeohydro-
Steig EJ (2004) Holocene climate variability. Quat Res 62:243–255 logical tracers in the White Nile headwaters lakes, East Africa. OOS
Merla G, Abbate E, Azzaroli A, Bruni P, Canuti P, Mazzuoli M, Sagri Transactions, AGU Fall Meeting Supplement 82, Absytacy PP21C-05
M, Tacconi P (1973) A geological map of Ethiopia and Somalia Talbot MR, Johannessen T (1992) A high resolution palaeoclimatic
(1973) 1:2,000,000 and comment with a map of major landforms. record for the last 27,500 years in tropical West Africa from the
CNR, Italy carbon and nitrogen isotopic composition of lacustrine organic
Noren AJ, Bierman PR, Steig EJ, Lini A, Southon JA (2002) matter. Earth Planet Sci Lett 110:23–37
Millennial-scale storminess variability in the northeastern United Verschuren D, Laird KR, Cumming BF (2000) Rainfall and drought in
States during the Holocene. Nature 419:821–824 equatorial East Africa during the past 1,100 years. Nature 403:
Russell JM, Verschuren D, Eggermont H (2007) Spatial complexity of 410–413
‘Little Ice Age’ climate in East Africa: sedimentary records from Viles HA, Goudie AS (2003) Interannual, decadal and multidecadal
two crater lake basins in western Uganda. Holocene 17:183–193 scale climatic variability and geomorphology. Earth Sci Rev 61:
Stager JC, Cumming BF, Meeker LD (2003) A 10,000 year high- 105–131
resolution diatom record from Pilkington Bay, Lake Victoria, East Xoplaki E, Fleitmann D, Diaz H, von Gunten L, Kiefer T (2011)
Africa. Quat Res 59:172–181 Medieval climate anomaly. Pages News 19:40 pp
Stager JC, Ryves D, Cumming BF, Meeker LD, Beer J (2005) Solar Yuan F, Sheng Y, Yao T, Fan C, Li J, Zhao H, Lei Y (2011)
variability and the levels of Lake Victoria, East Africa, during the Evaporative enrichment of oxygen-18 and deuterium in lake waters
last millennium. J Paleolimnol 33:243–251 on the Tibetan Plateau. J Paleolimnol 46:291–307
Talbot MR (1990) A review of the palaeohydrological interpretation of Zanettin B, Justin-Visentin E (1973) Serie di vulcaniti etiopiche. 1- La
carbon and oxygen isotopic ratios in primary lacustrine carbonates. serie dell’altipiano etiopico centro-orientale. Boll Soc Geol It
Chem Geol 80:261–279 92:313–327
Rift-Related Morphology of the Afar
Depression 15
Giacomo Corti, Ian D. Bastow, Derek Keir, Carolina Pagli,
and Elizabeth Baker
Abstract
The Afar Depression is a subaerial triple junction between the Nubian, Somalian and Arabian
Plates, the only place where the final stages of continental break-up can be observed on-land.
In spite of the region being hot and inhospitable, scientists have carried out fundamental work
in this unique geological setting over the last few decades. We have long-known that rifting
began on large-scale border faults that now bound the Afar Depression but what role magma
played in the development of this incipient ocean basin was not clear. However, in recent
years, it has been revealed that repeated dike intrusions together with normal faulting
accommodate extension producing a landscape dominated by spectacular fresh fault scarps
and active volcanic edifices that have been created during episodic tectonic, volcano-tectonic
and purely volcanic events. Observations from Ethiopia have fundamentally changed the way
we think about continental break-up. The challenge now is to take what we have learned and
apply it to the geological record of the rifted margins elsewhere on Earth.

Keywords
Afar Rifting Normal faulting Diking Graben Volcano
15.1 Introduction *200,000 km2 Afar is up to *300 km wide in the south

and *600 km long from south to north. Elevation drops
The Afar Depression is a triangular-shaped area of rifting at uniformly from *1,000 m in the south-west to below sea
the triple junction between the Nubian, Somalian and level in the north (Danakil depression) and in the east, where
Arabian plates (Figs. 15.1 and 15.2). Covering an area of the shores of Lake Asal, fluctuating at around 155 m below
sea level, represent the lowest subaerial point of the African
G. Corti (&) continent. Superimposed on this topography are young
Consiglio Nazionale delle Ricerche, Istituto di Geoscienze e
Georisorse, Via G. La Pira, 4, 50121 Florence, Italy volcanic landforms that only exceptionally stand 500 m
e-mail: giacomo.corti@unifi.it above the rift floor. The Afar lowland hosts one of the most
I.D. Bastow hostile environments on Earth. Maximum temperatures are,
Department of Earth Science and Engineering, Imperial College even in the coldest months, well above 30 °C and can exceed
London, South Kensington Campus, London SW7 2AZ, UK 50 °C during the summer wet season. Dallol, at the northern
D. Keir tip of Afar, has the highest average annual temperature for
National Oceanography Centre Southampton, University of an inhabited location: 34 °C between 1960 and 1966.
Southampton, Southampton SO14 3ZH, UK Rainfall is rare, averaging typically less than 200 mm per
C. Pagli year. The Afar Depression is an endorheic basin: the main
School of Geography, Earth and Environmental Sciences, waterflow into the area is the Awash River, which flows
Plymouth University, Plymouth PL4 8AA, UK
north-eastward through southern Afar where it ends in a
C. Pagli chain of interconnected lakes, the last of which is Lake
Dipartimento di Scienze della Terra, Università di Pisa, Via Santa
Maria, 53, 56126 Pisa, Italia Abhe. These saline lakes contain almost the only water in the
region.
E. Baker
Shell, Houston, TX 77079-1197, USA
252 G. Corti et al.
10 20 30 40 50
20 East African Rift

40 41 42 43 44
10 Gulf of
Zula
Jalua
0 15
Alid
Re
d
10
Da
Se
a
20 Danakil
depr 14
EA
30
R
Northern Nabro
Afar
Da
n
13 13
ak
il
Ethiopian
bl
D
oc
Plateau
k
Ado MI
IG
MH Central
12 Afar DG 12
TG Gulf of
De Tadjoura
Semera f
lf o
HG GG
Djbouti
Dessié Lake u
G den
Asal Goubhet
Bay A
11 11
Lake
Blue Nil e
Abhe Ali-Sabieh
BG
block
KK
Southern
Afar
10 AA 10
Bo ARABIAN
PLATE
Dire Dawa
AnBF
Harar
9 Somalian 9
Addis h Plateau AFRICAN
Ababa as PLATE
Aw
Africa-Arabia-Somalia
triple junction
38 39 40 41 42 SOMALIAN
PLATE
Fig. 15.1 Digital elevation model of the Afar Depression and kinematics. Black and white squares indicate major towns. AA Aleyu–
surrounding areas (data from the Shuttle Radar Topography Mission, Amoissa; Ado Ado Ale; AnBF Ankober border fault; BG Borkena
SRTM; resolution 90 m). Inset on the top left shows the extent of the graben; Bo Borama; D Dabbahu; Da Dallol; De Der’Ela graben; DG
system of rift valley composing the East African Rift. Inset on the Dobi graben; EAR Erta Ale range; GG Gaddale graben; HG Hanle
bottom right shows the plate kinematic setting of the region; dashed graben; IG Immino graben; MH Manda–Hararo rift; MI Manda-Inakir
lines indicate plate boundaries and white arrows illustrate plate rift; TG Tendaho graben
The Afar Depression is separated from the Ethiopian and Danakil depression between the Danakil block and the Ethi-
Somalian plateaus by major fault escarpments (Figs. 15.1 and opian escarpment, which connects the Afar lowland to the
15.3). Separation from the Red Sea and Gulf of Aden is Gulf of Zula. The Gulf of Tadjoura separates the Ali-Sabieh
defined less dramatically by the low relief of the Danakil block and Danakil blocks, connecting Afar to the Gulf of Aden.
(also called Danakil Alps) and the even more subdued hills of However, as is the case everywhere in Afar, young volcanic
the Ali-Sabieh block. To the north, Afar narrows in the centres prevent marine waters from flooding the region.
15 Rift-Related Morphology of the Afar Depression 253
Fig. 15.2 Tectonic setting of the 38 39 40 41 42 43 44

Afar Depression (modified after
Re
Keir et al. 2011). Solid black lines
show Oligocene–Miocene border d Arabia
faults of the Red Sea, Gulf of 15
Se 15
Aden and East African rifts. Red
a
segments show the Quaternary–
Recent subaerial rift axes, and
green triangles show Holocene
volcanoes. Dashed lines show the 14 Ethiopian 14
Tendaho–Goba’ad Discontinuity
(TGD). Gray circles show large
Plateau
earthquakes during 1973–2012
sourced from the National
Earthquake Information Centre 13 13
(NEIC) catalogue. Earthquake
focal mechanisms are from the
Global Centroid Moment Tensor
(CMT) catalogue. The black A A’
12 12
dashed line defines the
boundaries of the Danakil
microplate. The arrows show the
motion of the Danakil microplate
(McClusky et al. 2010). Cyan 11 TG Gulf of 11
lines with letters indicate the
traces of cross sections illustrated
D Aden
in Fig. 15.3
10 10
B’
20 mm/yr
B SE Plateau
Ethiopian
9 Earthquake Magnitude 9
rift
2 3 4 5 6
38 39 40 41 42 43 44
In the following sections, we describe the different anomalous uplift, this voluminous volcanism has been
physiographic provinces of Afar focusing on the morpho- related to the presence of one or more ‘traditional’ narrow
logic expression of the main tectonic–magmatic processes (e.g. Ebinger and Sleep 1998; Rogers et al. 2000) or, more
related to continental rifting and break-up. likely, a broad low wave speed thermal anomaly beneath
Ethiopia (e.g. Benoit et al. 2006; Furman et al. 2006; Bastow
et al. 2008; Ritsema et al. 2011; Hansen et al. 2012; Rooney
15.2 Western and Southern Rift et al. 2012a, b, 2013).
Escarpments The elevated plateaus are abruptly separated from the Afar
Depression by discontinuous boundary faults that give rise to
The Ethiopian, Somalian and Yemen plateaus are part of the major fault escarpments that are the continuation of the
so-called African Superswell, a wide region of anomalously margins of the Red Sea and Gulf of Aden (Wolfenden et al.
high topography comprising the East African, Arabian and 2005; Figs. 15.3 and 15.4). The faults are normally *60 km
southern African Plateaus as well as a bathymetric swell long, widely spaced and characterised by large vertical offsets
in the south-eastern Atlantic Ocean basin (Nyblade and (>1 km), accommodating a drop in elevation from 2,000 to
Robinson 1994). This anomalous topography results from 3,000 m at the ridgeline of the escarpments to less than
strong uplift during the Cenozoic, with up to *2 km rock 1,000 m in the marginal areas of the Afar Depression
uplift since *30 Ma (e.g. Pik et al. 2003). Both plateaus are (Wolfenden et al. 2005). Geochronological constraints sug-
capped with *2 km of Oligocene–Miocene thick sequences gest that the marginal faults were activated at
of flood basalts and rhyolites, with interbedded sedimentary *26–31 Ma on the western margin of Afar (i.e. Red Sea rift;
sequences (Hoffman et al. 1997). Together with the e.g. Ayalew et al. 2006; Wolfenden et al. 2005) and likely at
254 G. Corti et al.
Fig. 15.3 The western Afar

margin south of Dese (a) and in
the Ankober region (b) (Photos
D. Keir)
*33–34 Ma on the southern margin following rift activation topography of the rift margin, causing severe denudation
all along the Gulf of Aden (e.g. Leroy et al. 2012). Their creating badlands and hummocky terrain that averages 40 km
activity has been interpreted to have decreased with time in width between the floor of Afar and the crest of the present
starting from the Late Miocene, as extension-related defor- escarpments (e.g. Mohr 1978).
mation progressively migrated from the rift margins towards Different rift architecture and morphology characterise
the centre of the rift depression, reflecting a process of strain the western (Ethiopian) and southern (Somalian) escarp-
localisation as rifting proceeds to continental break-up (e.g. ments (e.g. Beyene and Abdelsalam 2005). The Ethiopian
Wolfenden et al. 2005). The crest of the escarpment is a escarpment is characterised by an overall rift-ward tilting
major drainage divide. To the west, the majority of rivers feed (Fig. 15.4), which has been interpreted as resulting from a
the Nile watershed and provide the majority of fertile sedi- first episode of down-warping of the Afar Depression fol-
ments so important for agriculture in Sudan and Egypt. To lowed by subsequent marginal faulting (Zanettin and Justin-
the east, the largely ephemeral rivers plummet down the steep Visentin 1975; Kazmin et al. 1980). This tilting is more
Fig. 15.4 Schematic geological (a) (a’)

cross sections across the western Ethiopian
and southern Afar margins (after Plateau Ethiopian
Beyene and Abdelsalam 2005). Escarpment
3000
The traces of the cross sections
are shown in Fig. 15.2 Marginal Marginal area of the
2000 graben Afar Depression
1000
W
? ? E
m asl
Km 20 40 60 80 100
(b) (b’)
Somalian Afar Stratoid
Plateau Trap Series
3000 Somalian Pre-rift Sedimentary Rocks
Escarpment Neoproterozoic Basement
2000 Marginal area of the Normal fault
? Afar Depression
1000
S ? N
m asl
Km 20 40 60 80 100
evident in the easternmost faulted blocks, where it gives rise piedmont scarp in unconsolidated materials along the
to hogback-like structures; in the internal part of the margin, escarpment of the Borkena graben with vertical differential
tilting is less developed as it is overprinted by development displacement up to 2 m, depth of 5–7 m and width at the
of marginal grabens (Figs. 15.4 and 15.5). These grabens, surface over 1 m (Gouin 1979; Fubelli and Dramis 2011).
typically 10–20 km wide and some tens of km long, are Significant seismic activity has been also recorded by the
filled with sediments of at least Pliocene–Quaternary age EAGLE network of seismic stations in the period from
(Kazmin 1972; Chorowicz et al. 1999); they represent sec- October 2001 to January 2003 along the complex Ankober
ondary features associated with major boundary fault sys- border fault system (Keir et al. 2006a, b).
tems (e.g. Corti 2012). In plan view, their en-echelon The southern (Somalian) escarpment differs from the
arrangement likely reflects a oblique component of exten- western margin of Afar in that it is characterised by a lack of
sional deformation during margin development, as also marginal basins and faulted blocks tilting away from the rift
suggested by geological analysis (Chorowicz et al. 1999). centre and a domino-style faulting that gives rise to minor half
The southern termination of the Ethiopian escarpment is at grabens only (Morton and Black 1975; Pizzi and Pontarelli
the Ankober border fault system, a major fault system that 2007) (Fig. 15.4). The margin is marked by isolated volcanic
characterises the structurally complex ‘corner’ between the centres aligned along the main structures; these volcanic
NE trending MER and the N–S trending Red Sea rift centres are more common in the south-western portion of the
(Fig. 15.2; Wolfenden et al. 2004). Although the activity of margin, close to the MER (Beyene and Abdelsalam 2005).
the marginal faults is believed to have decreased with time as The presence of numerous elbows and wind gaps on the main
a result of migration of deformation towards the rift axis, the river courses indicate a strong structural control on the local
prominent morphology of the major normal faults (with clear drainage network, with movements on faults inducing a large
morphotectonic indicators of recent activity, such as fresh number of river captures (Acocella et al. 2011). Different
scarps and triangular facets; Fig. 15.5; Chorowicz et al. from what is observed at the western escarpment, preliminary
1999) and significant seismic activity testify an ongoing geological analysis in the Harar–Dire Dawa area suggests that
tectonic activity of the major fault escarpments. The most tectonic activity along the Somalian escarpment decreased
important earthquakes in the region struck the Kara Kore significantly in the Middle–Late Pleistocene (Pizzi and Pon-
area, south of Dese, where a seismic sequence during May– tarelli 2007). However, relatively minor clusters of earth-
September 1961 produced two main shocks with ML > 6.4, quake are relatively common at the intersection between the
seven > 5.0 and more than 3,500 with ML > 3.5 (e.g. Gouin southern margin of Afar and the eastern margin of the MER,
1979). These events produced heavy alterations in the and also near Borama on the Ethiopia–Somalia border (Keir
landscape, with numerous landslides and a 12–15 km-long et al. 2006a, b).
256 G. Corti et al.
Fig. 15.5 Major normal fault bounding a marginal graben along the Ethiopian escarpment north of Dese (Photo M. Benvenuti)
northern Afar and Manda–Hararo rift and Tendaho Graben in

15.3 Rift Floor central Afar, whereas the subaerial Aden rift includes the
Asal Ghoubbet and Manda–Inakir rifts to the east (Fig. 15.6;
The floor of Afar has been shaped by *30 million years of e.g. Keir et al. 2013). A series of relatively narrow grabens
volcanic and tectonic processes that have led to the accumu- (such as the Dobi, Hanle and Gaddale grabens) in central
lation of voluminous volcanic rocks and to intense normal Afar transfer strain from the Red Sea rift to the Gulf of Aden
faulting. Although Oligocene to Early Miocene basalts (e.g. Manighetti et al. 2001). The Ethiopian Rift includes the
belonging to the trap series are preserved in limited areas at the systems of NE trending grabens and volcanic alignments that
margins of the depression (e.g. Beyene and Abdelsalam result from Nubia–Somalia rifting south of the Tendaho–
2005), the majority of volcanic rocks are Upper Miocene– Goba’ad discontinuity (e.g. Tesfaye et al. 2003).
Recent in age and formed after the young rift subsided These different sectors of the Afar Depression (northern
between the marginal faults. Volcanism is dominated by central and southern) show significant tectonic and structural
basaltic activity, whose most important manifestation is the differences, and hence are described separately in the fol-
Afar Stratoid Series, a >55,000 km2-wide and >1,500 m-thick lowing sections.
sequence consisting of basaltic lava flows and minor rhyolitic
ignimbrites erupted in the Pliocene–Pleistocene (*4–0.5 Ma)
mainly from volcanic fissures (Barberi et al. 1975; Lahitte 15.4 Northern Afar
et al. 2003; Acocella 2010). Emplacement of this basaltic
series was characterised by the highest eruption rates The morphology of northern Afar is dominated by the
(>5,000 km3/Ma) ever known for plate divergent margins and Danakil Depression, a *200–220 km-long, relatively nar-
is instead more typical of a large igneous provinces (Acocella row (50–150 km-wide) basin bound to the west by a
2010). The most recent, Quaternary–Recent volcanoes are prominent fault escarpment forming >2 km of relief at the
both composite stratovolcanoes, composed of interbedded western Afar margin and to the east by a less prominent
rhyolites, ignimbrites, andesites and basalt flows (e.g. Dab- series of faults forming a 500–1,000 m elevation rift margin
bahu), and also shield volcanoes of mainly basalts (e.g. Erta against the Danakil microplate (Fig. 15.7). Despite being on-
Ale range). Both types of volcano are the source of observed land, the basin floor is generally 50–100 m below sea level
fields of basaltic scoria cones, and fissure fed basaltic flows. with near-surface geology dominated by the a suite of
The E–NE trending transverse volcanic alignments and marine evaporite deposits (e.g. calcite, gypsum, halite and
trachytic–rhyolitic central volcanoes also characterise the sylvite). At least 2 km salt formed in the Danakil depression
Afar margins (e.g. Barberi and Varet 1977). when it was a submarine arm of the Red Sea during the
The significant volcanic activity is intimately related to Miocene. The upper part of the series deposited during
extension-related faulting and magma intrusion, which since repeated marine incursions in the Pleistocene, when lava
Quaternary–Recent times has been focused in axial magmatic flows dammed the narrow neck between the northern end of
segments (Figs. 15.2 and 15.6; e.g. Barberi and Varet 1977; the Danakil Horst and the western highlands and the sea
Manighetti et al. 2001; Keir et al. 2013). The current faulting water trapped in the depression soon evaporated. These
patterns are commonly quite complex and result from inter- incursions, the last of which occurred *30 ka (e.g. Barberi
action between the subaerial Red Sea and Aden rifts and the and Varet 1970; Bonatti et al. 1971; Hutchinson and Engels
northern termination of the Ethiopian rift (Figs. 15.1, 15.2 1972; Talbot 2008), left a classic pattern of marine reefs
and 15.6). The Red Sea includes the Danakil depression in perched at altitudes of −30 to +90 m along the former
40 42 44
39 40 41 42
A 14
Danakil C
Depression 12
B
14
Alu Dallafilla
Western
Afar Erta Ale 10
margin
Nabro D
Alaita
Danakil
13 Microplate
8
Dabbahu
Ado Ale
12 Manda-Hararo 12
Rift
Asal
Rift
Tendaho Dobe
Graben Graben
Hanle
Graben
11 11
Ethiopian
Rift
Ayelu-
Amoissa
10 10
9 9
0 20 40 120 160 80 200 km
39 40 41 42 43 44
Fig. 15.6 Structural map of the Afar Depression (modified after Keir The subaerial Red Sea rift is included in areas A (northern Afar) and B
et al. 2013). Major Pliocene–recent faults affecting the rift floor are (western Afar). The subaerial Aden rift is included in region C (central,
marked red. Recent basaltic lava flows are shaded light purple. The eastern Afar), and the northern Main Ethiopian rift is area D (southern
inset delineates the major rift system forming the Afar triple junction. Afar)
shoreline around the margins of the depression. The basin The axis of the Danakil depression is bisected by a linear,
also contains non-marine evaporites from ongoing evapo- *NNW striking, 10–20-km-wide axial volcanic range (Erta
ration of seasonal terminal saline lakes and the geothermal Ale range) which is the locus of ongoing volcanism and
pools (Talbot 2008). magmatism (e.g. Field et al. 2012; Pagli et al. 2012). This
258 G. Corti et al.
Fig. 15.7 Topography of Afar

1000
Height [m]
with major grabens and volcanic
centres labelled. Upper right 800
panel shows samples of 16˚ 600
topography on profiles A–A′ 400
across the Danakil depression and 200
Gulf of
B–B′ across central–eastern Afar. Zula 0 50 100 150 200
GT Gulf of Tadjoura Distance along A-A’ [km]
Height [m]
800
600
Jalua 400
15˚ Alid 200
0 50 100 150 200 250 300
Distance along B-B’ [km]
Dallol
Danakil
Depression A’
14˚
Alu-Dallafilla
Western
Afar margin
Erta Ale Nabro
A Alaita Danakil
13˚ Microplate 4000
B’
Dabbahu
Ado Ale
Manda-Hararo
Rift Serdo Asal
12˚ Rift 2000
Tendaho Dobi
Graben Graben GT
Hanle
Height [m]
B Graben
11˚ Ethiopian Lake 0
Borkena Rift Abhe
Graben
Kara Kore
km
Ayelu-
Amoissa
0 50 100
10˚
−2000
40˚ 41˚ 42˚ 43˚
range, unlike elsewhere in Afar, does not form a heavily shape, reflecting a control on the pathways of magma
faulted graben, but a composite system of different volcanic migration and eruption exerted by axial, rift-parallel fissures
centres with elevation of up >500 m above sea level (Fig. 15.8). Basalt flows are indeed fed from a combination of
(Fig. 15.8; e.g. Barberi and Varet 1970). The volcanic rocks *NNW-striking fissures and from edifices such as the
are largely dominated by Quaternary–Recent a′a and pa- *NNW elongate main caldera hosting Erta Ale lava lake
hoehoe basalt flows (Thurmond et al. 2006), which create the (e.g. Acocella 2006). The *1,700 × 700 m-wide, main cal-
classic shield volcano-like morphology with gentle slopes. A dera hosts a large northern crater and a smaller central crater
limited number of silicic centres create more prominent (Fig. 15.9; Acocella 2006). The central crater hosts a small pit
morphology, such as the steep-sided conical stratovolcano of crater floored by the lava lake in which lava elevation
Dalafilla that rises 300 m above surrounding lava fields SE of changes through time. It reached its highest level in February
Alu volcano. The Erta Ale range is markedly elliptical in 2010 when lava filled the pit crater; by November 2010,
a b Gada Ale c
km
0 5
c Ml 2 3 4
b
0m
0m
Alu
Alu
m
Dalafilla 20 m
0
m
Al
u
So 00 m
Dalafilla
ut
1
h
371.8 K
Borale Ale
281.6 K
1000
Erta Ale 750
Height [m]
500
Hayli Gubbi 250

Ale Bagu
0
km
0 10
Fig. 15.8 Satellite image (Source Google Earth) (a) and digital lines indicate the two en-echelon volcanic fissures, and the coloured
elevation model (SRTM data) (b) of the Erta Ale range (c) The area is the surface temperature of the erupted lava flow (from Pagli
November 2008 eruption in the Alu-Dalafilla volcanic centre. Dashed et al. 2012)
Fig. 15.9 Aerial view of the

elongate caldera of the Erta Ale
with the large northern and the
smaller central pit craters (Photo
E. Baker)
basalt overflowed into the main crater (Field et al. 2012). The *NNW-striking fissures in building the volcanic topography
larger northern crater also hosted a lava lake until early 1975. of the Erta Ale range is testified by the November 2008
The lava lakes have been persistently active for at least eruption in the Alu-Dalafilla volcanic centre, where two en-
100 years (Waltham 2010). The geometry of the main cal- echelon fissures opened south-east of Alu and erupted lava
dera, the characteristics of erupted lavas as well as the that flowed towards the east and north of Dalafilla (Fig. 15.8;
modelling of the satellite data of recent eruptions all indicate Pagli et al. 2012). InSAR data show that the eruption was
that the volcanic activity in the Erta Ale range is likely fed accompanied by up to *1.9 m of subsidence at Alu. The
from long-lived shallow (uppermost 2 km) pockets of magma deformation was caused by *1 km-wide, *10 km-long
(e.g. Pagli et al. 2012; Keir et al. 2013). The importance of deflating sill beneath the volcanic system.
260 G. Corti et al.
Fig. 15.10 Multicoloured salt

deposits, with small salt hornitos,
at Dallol (Photo D. Keir)
The axial lava flows of the Erta Ale range, which cover an 15.4.1 Central Afar
order of magnitude more surface area than similar age bas-
alts elsewhere in Afar (Bastow and Keir 2011), typically Central Afar is the complex region of mechanical interaction
flow away into the lower lying, evaporite-rich basin, creating between the on-land southern continuation of the Red Sea
a basin stratigraphy of thinly interbedded basalts and evap- Rift and the subaerial western continuation of the Aden Rift
orites (Talbot 2008), up to *5 km-thick (e.g. Makris and (e.g. Manighetti et al. 2001). The southern extent of the Red
Ginzburg 1987). Sea Rift includes the Manda–Hararo Rift and the Tendaho
The Erta Ale range drops in elevation to the north with the Graben, in the west, whereas the Gulf of Aden extends into
Dallol volcanic centre, one of the lowest subaerial volcanic Afar in the Asal-Ghoubbet and Manda–Inakir rifts, in the
vents globally. Dallol is world renowned for multi-coloured east (Figs. 15.1 and 15.7). A complex architecture made of a
salt and hydrothermal deposits, hot mineral springs, fuma- series of relatively narrow, non-volcanic grabens (such as the
roles and geysers (Fig. 15.10) which form as a consequence of Dobi, Hanle and Guma grabens) characterise the area
a complex interaction of solution and recrystallisation pro- between the Red Sea and Gulf of Aden systems, transferring
cesses driven by hydrothermal waters and rapid evaporation. extension between the two rift arms (e.g. Manighetti et al.
Dallol has been the site of a significant phreatic eruption in 2001). These three different subsectors are discussed in the
1926, which created a crater *30 m in diameter (Siebert et al. following sections.
2010). Recently, satellite remote sensing and seismicity
analysis captured the intrusion of *0.06 km3 dike accom- 15.4.1.1 Manda–Hararo Rift and Tendaho Graben
panied by a Mw 5.5 earthquake and associated fault slip along The Manda–Hararo Rift and the Tendaho Graben form a
the western flank of the rift during October 2004 (Nobile et al. series of NNW–SSE striking, 30–60 km-wide, *80 km-long
2012). The intrusion was fed by a previously unidentified basins bordered by 200–300–m-high escarpments
shallow magma reservoir beneath Dallol. The volcano lacks (Fig. 15.7). These generally lack a clear single border fault
volcanic rocks at the surface because they are covered by salt. but are characterised by numerous closely spaced relatively
The magmatic and tectonic activity also indicates that the Erta small-offset faults that displace the Stratoid Series and define
Ale range extends up to Dallol as a continuous and uniformly several rigid blocks tilted toward the rift axis (Figs. 15.11 and
trending rift branch. Activity at the Alid and Jalua, north of 15.12; e.g. Acocella et al. 2008). This inward-dipping Stra-
Dallol, has not been reported but the spacing between mag- toid blocks arranged in a domino configuration inside the
matic centres suggests that the Erta Ale Ridge further extends margins result in an overall syncline-like structure of the
to the Gulf of Zula on the Eritrean coast. basins (e.g. the central part of the Tendaho Graben; Acocella
Fig. 15.11 Structure and

Master fault
mechanism of formation of the (a) 10°
10° Depression
Tendaho Graben (modified from
axis
Acocella 2010). a Structure of the
central part of the graben. b above 5°
20° Attitude
<3
topographic section along the 10°
Hydrothermal
00
profile indicated in a; below 10° 30° site
ka
10°
10°
simplified structure (not to scale) 5°
ba
of the Tendaho Graben in cross
s
alt
section; double arrows report
s
10°
extent of block domains with 20° 5°
similar tilt. Note the domino-like Trace of
structures, with blocks profiles below
10°
characterised by a uniform tilt
10°
angle separated by faults.
c Mechanism of formation of the 15° 30°
graben as a consequence of
repeated magma withdrawal from 5°
a rift-parallel elongated reservoir
5°
20°
20° 10°
5° 5° N
15°
10°
Trace of 10° 5°
10°
profiles below 20° 10 km
Outcropping Outcropping
1 km tilted blocks tilted blocks
600 m
400 m
SW NE
10 km
Outcropping
Afar Stratoids
(0.5–2 Ma)
20°
10° rift 30°
5° axis 10°
~0° 0°
(b) 10 km
(c)
Inset C
C
et al. 2008), a geometry that has been related to rift collapse wide central axis where the youngest (0–0.2 Ma old) fissural
induced by magma withdrawal during the emplacement of basalts crop out (Lahitte et al. 2003; Acocella et al. 2008;
the Stratoid sequence (Fig. 15.11; Acocella 2010). The basins Acocella 2010). At the southern tip of the Tendaho Graben,
are filled with >1 km of lacustrine and fluvial deposits as well Lake Abhe is characterised by hundreds of travertine towers,
as by Quaternary–Recent volcanic rocks, the distribution and giving rise to several linear chains of travertine chimneys.
age of which suggest that faulting and volcanism have pro- Each tower is up to 60 m tall, formed by geothermal activity
gressively localised through the Quaternary to a *10 km- (e.g. Waltham 2010; Hussein et al. 2013).
262 G. Corti et al.
Fig. 15.12 The eastern side of

Tendaho Graben (Photo D. Keir)
The northern part of the Manda–Hararo rift is the maintain the morphology of volcanic segments (e.g. Wright
*60-km-long and *15-km-wide Dabbahu volcanic seg- et al. 2006, 2012; Keir et al. 2009; Ebinger et al. 2010; Grandin
ment, which has been the locus of a major volcano-tectonic et al. 2010).
episode during 2005–2010 (e.g. Wright et al. 2006; Rowland
et al. 2007; Grandin et al. 2009; Ebinger et al. 2010). The event 15.4.1.2 Asal-Ghoubbet and Manda–Inakir Rifts
was characterised by the emplacement of a *60 km-long, up The eastern–north-eastern edge of central Afar is character-
to 10-m-thick basaltic dike into the upper 10 km of the crust ised by a narrow zone of dense faulting that includes the two
beneath the magmatic segment during September 2005 NW-SE trending volcanic rifts of Asal-Ghoubbet, to the
(Fig. 15.13), accompanied by a minor eruption of pumice and south, and Manda-Inakir, to the north (Fig. 15.16; e.g.
ash on 26 September from the Da’Ure vent near Dabbahu Manighetti et al. 1998, 2001). These two rifts formed in the
(Fig. 15.14; e.g. Wright et al. 2006). Dike intrusion was last 1 Ma, when extension in the Gulf of Aden localised to
accompanied by near-symmetrical rift perpendicular opening axial volcanic segments breaching the Danakil Block from
of up to 8 m, with the flanks of the rift uplifted by up to 2 m, the Ali-Sabieh Block and proceeding along the Gulf of
and a 2–3-km-wide graben subsiding by 2–3 m (Figs. 15.14 Tadjoura into the Afar Depression (e.g. Manighetti et al.
and 15.15; e.g. Wright et al. 2006; Ayele et al. 2007; Grandin 1998). The *40-km-long, *15-km-wide Asal-Ghoubbet is
et al. 2009). Slip on individual normal faults was up to 3 m, the first segment of the Aden ridge to rise significantly above
with magnitude of seismic events peaking up to magnitude the sea level. Steep, inward dipping normal faults delimit the
ML = 5.6 (Ayele et al. 2007; Rowland et al. 2007). From 300–800-m-deep basin, which is subdivided in at least two
June 2006 to May 2010, the initial dike has been followed by disconnected, parallel subrifts; the rift floor is cut by dense
a sequence of 13 smaller dikes typically 1–3 m thick and arrays of open fissures roughly parallel to the normal faults
10–15 km long (e.g. Ebinger et al. 2010; Wright et al. 2012). (Fig. 15.17; Manighetti et al. 1998). Both fissures and normal
Three of these dikes reached the surface as basaltic fissural faults often display sharp and youthful morphology, with
eruptions, although the total erupted volume is a small fraction steep walls in fine white silts, suggesting development during
of that intruded into the crust (Ferguson et al. 2010). Geo- earthquakes in the last century (Manighetti et al. 1998).
physical data of this suggest that the majority of the intrusions Deformation recorded by means of geodetical and geophys-
were fed laterally from a *10-km-deep magma reservoir ical data also shows the ongoing activity of the majority of
beneath the Ado’Ale volcanic complex located at the centre of the faults (e.g. Doubre et al. 2007). Notably, the north-
the segment (Ayele et al. 2009; Grandin et al. 2009). These western subrift has faults with sharper and fresher scarps, and
episodic events of normal faulting and axial graben subsi- wider and denser fissures; this suggests that this subrift is
dence induced by lateral dike intrusion from a segment-cen- younger than the one to the south-east and that deformation
tred magma reservoir are modulated by cyclic variations in has migrated to the north-west with time (Manighetti et al.
magma supply (Medynski et al. 2013) and function to 1998).
264 G. Corti et al.
b Fig. 15.13 The top left and right panels, respectively, show the fault slip causing graben subsidence is a relatively shallow
earthquake locations reported by the National Earthquake Information phenomenon occurring mainly in the upper few kilometres of the
Center (NEIC) and focal mechanism reported by the Global Centoid Earth. The fault slip is induced by magma intrusion sourced from upper
Moment Tensor Catalog, for the September 2005 Dabbahu rifting crustal magma chambers such as Dabbahu, Gabho and the Ado’Ale
episode. Normal fault earthquakes were primarily located along the volcanic complex (not shown). Magma intrusion occured mainly at 2–
Dabbahu segment. The lower panel is a cartoon sketch illustrating that 10 km depth beneath the rift
Fig. 15.14 Top The Da’Ure vent near Dabbahu formed during a minor eruption of pumice and ash on 26 September 2005. Bottom Normal fault
formed during the volcano-tectonic event (Photos E. Baker)
Fig. 15.15 Vertical and

horizontal displacements (left and
central panels) following the
2005 Dabbahu rifting episode and
traces of the faults that were
activated during the event (right
panel). Modified after Grandin
et al. (2009)
Active volcano-tectonic activity in the area is testified by a individual faults varied between 0.15 and 0.5 m inside a
rifting event in November 1978, characterised by a major 3-km-wide zone along 10 km of the exposed area between
earthquake (Mb = 5.3) near the rift axis close to the coast of Goubhet and Asal (Ruegg et al. 1979). This deformation was
the Ghoubbet Bay and a seismic swarm that lasted for associated with basalt flows from numerous fissures and
2 months (Ruegg et al. 1979). This earthquake produced vents at the newly formed volcanic centre of Ardoukôba,
*2 m of extension and up to *70 cm subsidence, accom- located north-west of a large volcano (Fieale) in the centre of
modated by the opening of several fissures and the reacti- the Asal-Goubhet rift (Figs. 15.16 and 15.17). Mechanical
vation of normal fault scarps on the rift axis; slip on modelling of the deformation suggests that the rifting episode
266 G. Corti et al.
12.5
Im
M 2000
Al
12
Dobi
Height [m]
1000
D
Asal Gulf of Tadjoura
Ar
Tendaho
Hanle Gh 0
11.5
Ga
km
0 10 20 30
41 41.5 42 42.5
Fig. 15.16 Digital Elevation Model (SRTM data) of easternmost zone between the Asal-Goubhet and Manda Inakir segments; black
central Afar. Inset shows a tectonic sketch, with indicated the active dashed line indicates the trace of the profile shown in Fig. 15.18. Al
volcano-tectonic rift segments of Tadjoura, Asal-Goubhet and Manda Alol; Ar Ardoukoba; D Der’Ela; Ga Gaddale; Gh Goubhet; K Kam-
Inakir (after Manighetti et al. 2001). Normal faults are indicated in mourta; Im Immino; M Mak’arrasou oblique transfer zone
black; arrows indicate the relative motion of the Mak’arrasou transfer
resulted from the opening of two 4- to 8-km-long axial dikes volcanoes: Inakir to the south-east, an elongated NW trend-
at 4–5 km depth (Tarantola et al. 1979). Following the 1978 ing dome with numerous parasitic spatter cones, and Manda,
event, dike inflation continued up to 1985, accompanied by with more easterly trend and hosting the most recent lavas
seismicity and fault activity that continues today (Cattin et al. (Fig. 15.19; Manighetti et al. 1998). Swarms of normal faults
2005; Doubre et al. 2007). Overall, the style of volcano- are mostly localised north-east of the volcanic zone and form
tectonic activity in the Asal-Goubhet rift suggests that its three parallel subrifts; their topography is more attenuated
structure and morphology were likely created during epi- than in the Asal-Goubhet rift suggesting a younger devel-
sodes of lateral migration of magma along rift-parallel shal- opment of faulting in the area (Manighetti et al. 1998). Some
low dikes, fed by a magma chamber beneath the rift centre of the bounding normal faults exhibit extremely fresh mor-
(e.g. Doubre et al. 2007), as is observed at Dabbahu. phology, with light-coloured scarplets, suggesting they have
To the north–north-west, the Asal-Ghoubbet rift is con- likely ruptured in the last few hundred years (Tapponnier
nected to the Manda-Inakir rift, a volcanic range character- et al. 1990). As in the Asal-Goubhet rift, several open fissures
ised by Pleistocene–Holocene lava flows and associated (commonly more than 10 m deep and several meters wide)
normal faults (Fig. 15.16; Manighetti et al. 1998). Connec- affect the rift floor. These fissures are often sharp and steep
tion between the two rifts occurs through the Mak’arrasou and with very fresh morphology suggesting very recent
oblique transfer zone, which is composed of a network of activity; some of them cut Late Pleistocene basalt flows in the
closely spaced extensional faults and tilted blocks dipping rift floor. Most of youngest faults and fissures are concen-
towards Afar and giving rise to a large-scale flexure trated in the north-western subrift, suggesting that—likewise
(Fig. 15.18; Le Gall et al. 2011). Volcanic products in the the Asal-Goubhet rift—deformation propagated from south-
Manda-Inakir rift are associated to two large shield east to north-west (Manighetti et al. 1998). Historical
Ar
Fig. 15.17 Tectonic and simplified geological map of the subaerial section of the Asal-Goubhet Rift superimposed onto a SRTM digital elevation
model (after Doubre et al. 2007). Ar Ardoukoba; Fi Fieale caldera; SB Shark Bay caldera
Afar floor
3–1 Ma Stratoid
and Gulf basalts Lower Stratoid basalts Dahla basalts
Fig. 15.18 Structural cross section of the central part of the profile shown in Fig. 15.16. Note the network of closely spaced
Makarassou fault system, at the eastern margin of the Afar Depression extensional faults and tilted blocks giving rise to hogback-like
(after Le Gall et al. 2011). Vertical exaggeration = 4. Trace of the structures and a large-scale flexure towards Afar
volcanic activity was reported in the area, with the last 15.4.1.3 Basins Between the Red Sea and Gulf
eruption occurring in 1928 or 1929 at the south-eastern end of of Aden Systems
the Manda-Inakir rift and producing the Kammourta cinder In the central portion of Afar, between the Gulf of Aden and
cone and a lava flow (Audin et al. 1990). Red Sea systems, deformation is spread over an area
268 G. Corti et al.
with several reported surface breaks and oblique, left-lateral

faulting with displacement up to *75 cm (Gouin 1979).
Another series of destructive earthquakes, ten with 5.5 ≤
Ms ≤ 6.3, struck the Dobi graben on August 20–21 1989,
rupturing both boundary and inner floor faults over an area
about 45 km long and 15 km wide (e.g. Jacques et al. 2011).
Numerous surface breaks with complex geometry, including
fresh scarplets with vertical throws up to 30-cm-high and
open fissures up to 30 cm wide, were associated to the events
(Fig. 15.22). The two largest shocks (Ms = 6.2 and 6.3)
ruptured the southern segments of the south-western
bounding fault of the Dobi graben; a dozen other faults also
slipped along the edges of, and inside, the graben (Jacques
et al. 2011).
Overall, the complex, distributed deformation affecting
the area between the Gulf of Aden and Red Sea system
results from the overlap without direct connection of the two
major rift arms, which has been suggested to induce a
bookshelf faulting mechanism in which small rigid blocks
rotate about vertical axis along rift-parallel faults that slip
with a component of left-lateral motion (Fig. 15.23; e.g.
Fig. 15.19 NASA Earth Observatory image of central Afar. The black Tapponnier et al. 1990). The proposed block rotations were
area close to the top right corner is the recent basaltic lava field of the also previously cited as evidence that extension in Afar is
Manda volcanic centre. The narrow, curved depression in the centre of controlled by the anticlockwise rotation of the Danakil
the image is the Immino graben (IG)
block, a model called the Danakil crank-arm model (Souriot
and Brun 1992). Recent geophysical data have been used to
*120 × 100 km (Figs. 15.16 and 15.19; e.g. Manighetti et al. test these models with GPS data confirming that the Danakil
2001). This region is completely overlain by the thick series block does currently rotate anticlockwise (McClusky et al.
of the Stratoid basalts and characterised by a lack of younger 2010). However, the large component of strike slip motion
volcanism. South of *12°N, the Stratoid series is dissected on the faults in central Afar is unclear since the majority of
by a series of NW trending major normal fault systems with earthquake focal mechanisms show normal faulting (e.g.
high (100–1,000 m), steep (>70°) cumulative escarpments, Keir et al. 2013), and the deformation field constrained using
and numerous smaller, subparallel normal faults in between. InSAR and GPS also shows rift perpendicular extension in
The major border faults trend NW and define 10–20-km-wide the overlapping/en-echelon basins in central Afar with no
and 30–50-km-long subparallel basins (e.g. Dobi, Hanle, major shear component of motion (Pagli et al. 2014).
Der’Ela, Gaddale), with intervening narrow horst blocks, so
that the whole area is characterised by an overall horst-and-
graben morphology (Figs. 15.16 and 15.20). North of latitude 15.4.2 Southern Afar
*12°N, deformation becomes more complex, being char-
acterised by a network of major crosscutting active faults South of the Tendaho–Goba’ad discontinuity in the East
giving rise to curvilinear narrow grabens (e.g. Immino gra- African Rift, the morphology of the rift floor of the MER is
ben; Fig. 15.19). The basin is filled by up to >1,500-m-thick dominated by the presence of *30 km-wide, *60 km-long
sediments, mostly Quaternary–Recent lacustrine deposits and en-echelon axial volcanic segments (Figs. 15.2 and 15.24).
alluvial fans deposited during wet periods at the base of large These are characterised by strong association of active and
fault escarpments (Gasse 1991). Most of the normal faults recent volcanoes, aligned monogenetic cones and fissures,
affecting the area are characterised by a fresh morphology, and numerous normal faults striking roughly perpendicular
with a scarplet at the base of the cumulative escarpment to the direction of extension (e.g. Keir et al. 2013). The
offsetting Late Quaternary alluvial fans, thus attesting their faults of the axial volcanic segments are normally short,
recent/ongoing activity (Fig. 15.21; Manighetti et al. 2001). closely spaced and display relatively small throws (<100 m),
Several destructive earthquakes with recorded surface giving rise in places to a graben-in-graben morphology
deformation have affected the area. A major earthquake (Fig. 15.24) (Soliva and Schulz (2008). These faults are
affected the area of Serdo, between the Tendaho and Dobi characterised by steep (subvertical) scarps and are com-
grabens, in March–April 1969 (magnitude up to Ms = 6.3), monly en-echelon and linear or curved in plan view over
Fig. 15.20 Top panel The dry

alluvial plains bounded by major
normal fault escarpments of the
Dobi graben. Inset shows a
satellite image of the region.
Bottom panel A salt pan in the
Dobi Graben, where locals pump
water onto the flat basin floor and
then wait for it to evaporate.
Major boundary faults in the
background (Photos E. Baker and
D. Keir)
distances of up to a few tens of kilometres, thus delimiting (Hayward and Ebinger 1996). The axial basins are filled with
several fault-bounded blocks. Associated with the faults are both volcanic products and sequences of Pliocene–Recent
open fissures with or without vertical displacement, splay alluvial and lacustrine sediments, with rivers eroding the
patterns and complex rhomb-shaped structures (e.g. Mohr uplifted fault footwalls and producing depositional fans in
1987; Boccaletti et al. 1998; Williams et al. 2004). Structural the basins (Hayward and Ebinger 1996).
and geomorphological analyses suggest a migration of The dense fault swarms developed in the last 2 Ma (e.g.
deformation toward the axis of the volcanic segments, with Ebinger and Casey 2001) and geodetic data support that they
inner normal faults displacing recent basalt flows erupted accommodate *80 % of the present-day strain (Bilham
from rift-parallel fissures and eroded more external faults et al. 1999). Gravity and seismic imaging show extensive
270 G. Corti et al.
Fig. 15.21 A major boundary fault of the Dobi graben displacing the Afar Stratoid series, with the thick succession of basalt flows visible on the
fault footwall (Photo E. Baker). Note the presence of a fresh, light-coloured scarplet at the base of the cumulative escarpment
G
ul
fo
fA
de
n
pr
op
ag
at
or
R
ed
Se
a
pr
op
ag
at
or
Fig. 15.23 Top panel Model of bookshelf faulting, in which rift

propagation in opposite directions induces small-rigid blocks to rotate
about vertical axis along rift-parallel faults that slip with a left-lateral
component. Bottom panel Model of bookshelf faulting applied to
central Afar; the grey line indicates a hypothesised major transversal
Fig. 15.22 Ruptures associated with the 1989 Dobi seismic sequence structure (after Manighetti et al. 2001)
(after Jacques et al. 2011)
Te
n
km
da
ho
-G
0 50 100
ob
a’a
r d
Rive Di
s co
sh nt i
Awa
nuit
11° y
A
10°
Aleyu-Amoissa 4000
A’
Ethiopian
Rift
Height [m]
Fantale 2000
9°
Boseti 1000
800
600 0
Height [m]
39° 4 0 ° 400 41° 42°

200
0 50 100 150 200

Distance along A-A’ [km]
Fig. 15.24 Topography of southern Afar with major volcanic centres labelled. Bottom right panel shows a sample of topography on profile A–A′
across the southern Afar floor
magma intrusion throughout the lithosphere beneath mag- The NE–NNE striking magmatic segments related to the
matic segments (e.g. Bastow et al. 2011), with extension Nubia–Somalia opening overprint previous NW–NNW
facilitated and accommodated by a combination of magmatic striking faults, suggesting that extension in the northern
intrusion, diking and faulting (e.g. Keir et al. 2006b). MER postdates extension from initial opening of the
According to Kurz et al. (2007), the architecture and mor- southern Red Sea and Gulf of Aden (e.g. Tesfaye et al. 2003;
phology of the magmatic segments are symmetric around Wolfenden et al. 2004). However, the NW–NNW trending
central acidic strato-volcanoes, where the deformation pat- structures still accommodate some volcano-tectonic activity,
tern is characterised by small faults and a low fault density. such as the intrusion of a *N120°E striking, *6-km-long
The number of faults increases longitudinally away from the dike during May 2000 (Fig. 15.25; Keir et al. 2011). The
volcanic centres, such that maximal fault densities and the intrusion occurred beneath the *WNW striking Ayelu–
longest faults are observed at tip domains (Kurz et al. 2007). Amoissa chain of aligned volcanic centres that dominate the
This reflects an along-axis change from magmatic defor- morphology of the right stepping MER offset at *10°N
mation accommodated by diking at segment centres to (Fig. 15.25), but did not result in either volcanic eruptions or
mostly brittle deformation (faulting) at the tip domains. significant surface deformation (Keir et al. 2011).
272 G. Corti et al.
eyu Amoissa
Fig. 15.25 Digital elevation model of the Ayelu-Amoissa volcanic the dike (after Keir et al. 2011). Photo showing the Ayelu (on the right)
system. Coloured circles are epicentres related to the May 2000 and the Amoissa (the lower one on the left) volcanoes (Photo D. Keir)
intrusion event; the black dashed line indicates the surface projection of
thick sequences of basalt flows and evaporites represents a

15.5 Conclusions modern analogue for the processes responsible for the
development of volcanic rifted margins worldwide.
Afar is a spectacular example of how rifting at a complex
triple junction can create and continuously modify the
morphology of a region at different scales. Large-scale References
processes, initiated at depth in the Earth’s mantle, have
shaped the overall morphology of the area by inducing Acocella V (2006) Regional and local tectonics at Erta Ale caldera,
plateau uplift (up to *2 km) and emplacement of volumi- Afar (Ethiopia). J Struct Geol 28:1808–1820
nous flood basalt sequences at *30 Ma. The subsequent Acocella V (2010) Coupling volcanism and tectonics along divergent
subsidence of the Afar floor with respect to the surrounding plate boundaries: collapsed rifts from central Afar, Ethiopia. Geol
Soc Am Bull 122:1717–1728
elevated areas during initial Arabia–Africa rifting has cre- Acocella V, Abebe B, Korme T, Barberi F (2008) Structure of Tendaho
ated large fault escarpments that abruptly separate the rift Graben and Manda Hararo Rift: implications for the evolution of the
floor from the plateaus. Although not characterised by sig- southern Red Sea propagator in Central Afar. Tectonics 27:
nificant current deformation and in places highly eroded, TC4016. http://dx.doi.org/10.1029/2007TC002236
Acocella V, Bekele A, Coltorti M (eds) (2011) Pre‐conference
these escarpments still dominate the regional morphology. excursion guide: Tectonic Landforms and Volcanism in the
The floor of Afar is primarily shaped by the tectonic– Southern Afar. Iag regional conference 2011 geomorphology for
magmatic processes that developed since extension and human adaptation to changing tropical environments Addis Ababa,
became focused in a narrower zone of faults and volcanoes Ethiopia, 18–22 Feb 2011
Audin J, Vellutini PJ, Coulon C, Piguet P, Vincent J (1990) The 1928–
within the depression. In these rifts, continuing extension is 1929 eruption of Kammourta volcano. Evidence of tectono-
accommodated by magma emplacement and normal faulting, magmatic activity in Manda-Inakir rift and comparison with the
which gives rise to a landscape dominated by spectacular Asal rift, Afar Depression, Republic of Djibouti. Bull Volcanol
fresh fault scarps and active volcanic edifices created during 52:551–561
Ayele A, Keir D, Ebinger C, Wright TJ, Stuart GW, Buck RW, Jacques
episodic tectonic, volcano-tectonic and purely volcanic E, Ogubazghi G, Sholan J (2009) September 2005 mega-dike
events. emplacement in the Manda–Harraro nascent oceanic rift (Afar
This morphology, and its along-axis variations, reflects the Depression). Geophys Res Lett 36: L20306. http://dx.doi.org/10.
last phases of the protracted process leading to plate separation 1029/2009GL039605
Ayalew D, Marty B, Barbey P, Yirgu G, Ketefo E (2006) Temporal
and the birth of a new oceanic basin. As break-up progresses, compositional variation of syn-rift rhyolites along the western
rift-related morphology changes from the narrow fault bound margin of the southern Red Sea and northern Main Ethiopian Rift.
graben as in central Afar to the topographically prominent In: Yirgu G, Ebinger CJ, Maguire PKH (eds) The Afar Volcanic
axial volcanic range in the Danakil depression where the rel- Province within the East African Rift System. Geol Soc Spec Pub
259:121–130
atively large volumes of erupted material show that a marked Ayele A, Stuart G, Bastow I, Keir D (2007) The August 2002
increase in rate of basaltic volcanism characterises the final earthquake sequence in north Afar: insights into the neotectonics of
stages of continental break-up. Here, the accumulation of the Danakil microplate. J Afr Earth Sci 40:70–79
Barberi F, Varet J (1970) The Erta Ale volcanic range (Danakil tropical environments Addis Ababa, Ethiopia, 18–22 Feb 2011,
depression, northern Afar, Ethiopia). Bull Volcanol 34:848–917 pp 38–42
Barberi F, Varet J (1977) Volcanism of Afar: small-scale plate tectonics Furman T, Bryce J, Hanan B, Yirgu G, Ayalew D (2006) Heads and
implications. Geol Soc Am Bull 88:1251–1266 tails: 30 years of the Afar plume. In: Yirgu G, Ebinger CJ, Maguire
Barberi F, Ferrara G, Santacroce R, Varet J (1975) Structural evolution PKH (eds) The Afar Volcanic Province within the East African Rift
of the Afar triple junction. In: Pilger A, Rosler A (eds) Afar System. Geol Soc Lond Spec Publ 259:95–119
Depression of Ethiopia. Schweizerbart, Stuttgart, pp 39–54 Gasse F (1991) Tectonic and climatic controls on lake distribution and
Bastow I, Keir D (2011) The protracted development of the continent- environments in Afar from Miocene to Present. In: Katz B (ed)
ocean transition in Afar. Nat Geosci 4:248–250 Lacustrine basin exploration: case studies and modern analogs.
Bastow ID, Nyblade AA, Stuart GW, Rooney TO, Benoit MH (2008) AAPG Mem 50:19–41
Upper mantle seismic structure beneath the Ethiopian hot spot: Gouin P (1979) Earthquake history of Ethiopia and the Horn of Africa.
rifting at the edge of the African low-velocity anomaly. Geochem Int Develop Res Centre, Ottawa, Publ 118
Geophys Geosyst 9:Q12022. doi:10.1029/2008GC002107 Grandin R, Socquet A, Binet R, Klinger Y, Jacques E, de Chabalier J-
Bastow ID, Keir D, Daly E (2011) The Ethiopia Afar Geoscientific B, King GCP, Lasserre C, Tait S, Tapponnier P, Delorme A, Pinzuti
Lithospheric Experiment (EAGLE): probing the transition from P (2009) September 2005 Manda Hararo-Dabbahu rifting event,
continental rifting to incipient seafloor spreading. Geol Soc Am Afar (Ethiopia): constraints provided by geodetic data. J Geophys
Spec Pap 478:51–76 Res 114:B08404. doi:10.1029/2008JB005843
Benoit M, Nyblade A, VanDecar J (2006) Upper mantle P wavespeed Grandin R, Socquet A, Jacques E, Mazzoni N, de Chabalier J-B, King
variations beneath Ethiopia and the origin of the Afar hotspot. GCP (2010) Sequence of rifting in Afar (Manda-Hararo rift,
Geology 34:329–332 Ethiopia, 2005–2009): timespace evolution and interactions
Beyene A, Abdelsalam MG (2005) Tectonics of the Afar Depression: a between dikes from InSAR and static stress change modeling.
review and synthesis. J Afr Earth Sci 41:41–59 J Geophys Res 115:B10413. doi:10.1029/2009JB000815
Bilham R, Bendick R, Larson K, Mohr P, Braun J, Tesfaye S, Asfaw L Hansen S, Nyblade A, Benoit M (2012) Mantle structure beneath
(1999) Secular and tidal strain across the Main Ethiopian Rift. Africa and Arabia from adaptively parameterized P-wave tomog-
Geophys Res Lett 26:2789–2792 raphy: implications for the origin of Cenozoic Afro-Arabian
Boccaletti M, Bonini M, Mazzuoli R, Abebe B, Piccardi L, Tortorici L tectonism. Earth Planet Sci Lett 319:23–34
(1998) Quaternary oblique extensional tectonics in the Ethiopian Hayward N, Ebinger C (1996) Variations in the along–axis segmen-
Rift (Horn of Africa). Tectonophysics 287:97–116 tation of the Afar rift system. Tectonics 15:244–257
Bonatti E, Emiliani C, Ostlung G, Rydell H (1971) Final desiccation of Hofmann C, Courtillot V, Féraud G, Rochette P, Yirgu G, Ketefo E, Pik
the Afar rift, Ethiopia. Science 172:468–469 R (1997) Timing of the Ethiopian flood basalt event and implications
Cattin R, Doubre C, de Chabalier J-B, King G, Vigny C, Avouac J-P, for plume birth and global change. Nature 389:838–841
Ruegg J-C (2005) Numerical modelling of quaternary deformation Hutchinson RW, Engels GG (1972) Tectonic evolution of the southern
and post-rifting displacement in the Asal-Ghoubbet (Djibouti, Afar). Red Sea and its possible significance to older rifted continental
Earth Planet Sci Lett 239:352–367 margins. Geol Soc Am Bull 83:2989–3002
Chorowicz J, Collet B, Bonavia F (1999) Left-lateral strike-slip Hussein B, Chandrasekharam D, Chandrasekhar V, Jalludin M (2013)
tectonics and gravity induced individualisation of wide continental Geochemistry of thermal springs around Lake Abhe, Western
blocks in the western Afar margin. Eclogae Geol Helv 92:149–158 Djibouti. Int J Sustain Energ. doi:10.1080/14786451.2013.813027
Corti G (2012) Evolution and characteristics of continental rifting: Jacques E, Kidane T, Tapponnier P, Manighetti I, Gaudemer Y, Meyer
analogue modeling-inspired view and comparison with examples B, Ruegg JC, Audin L, Armijo R (2011) Normal faulting during the
from the East African Rift System. Tectonophysics 522–523:1–33 August 1989 Earthquakes in Central Afar: sequential triggering and
Doubre C, Manighetti I, Dorbath L, Dorbath C, Bertil D, Delmond JC propagation of rupture along the Dôbi Graben. Bull Seis Soc Am
(2007) Crustal structure and magmato-tectonic processes in an 101:994–1023
active rift (Asal-Ghoubbet, Afar, East Africa): 2. Insights from the Kazmin V (compiler) (1972) Geological map of Ethiopia (1:2,000,000).
23-year recording of seismicity since the last rifting event. Ethiopian geological survey, Ministry of Mines and Energy
J Geophys Res 112:B05406. doi:10.1029/2006JB004333 Resources, Addis Ababa
Ebinger C, Sleep NH (1998) Cenozoic magmatism in central and east Kazmin V, Seife MB, Nicoletti M, Petrucciani C (1980) Evolution of
Africa resulting from impact of one large plume. Nature 395: the northern part of the Ethiopian Rift. In: Geodynamic evolution of
788–791 the Afro-Arabian Rift System, Accademia Nazionale Dei Lincei,
Ebinger C, Casey M (2001) Continental breakup in magmatic Atti dei Convegni Lincei 47:275–292
provinces: an Ethiopian example. Geology 29:527–530 Keir D, Stuart G, Jackson A, Ayele A (2006a) Local earthquake
Ebinger C, Ayele A, Keir D, Rowland J, Yirgu G, Wright T, Belachew magnitude scale and seismicity rate for the Ethiopian Rift. Bull Seis
M, Hamling I (2010) Length and timescales of rift faulting and Soc Am 96:2221–2230
magma intrusion: the Afar rifting cycle from 2005 to present. Ann Keir D, Ebinger C, Stuart G, Daly E, Ayele A (2006b) Strain
Rev Earth Planet Sci 38:439–466 accommodation by magmatism and faulting as rifting proceeds to
Ferguson DJ, Barnie TD, Pyle DM, Oppenheimer C, Yirgu G, Lewi E, breakup: seismicity of the northern Ethiopian rift. J Geophys Res
Kidane T, Carn S, Hamling I (2010) Recent rift-related volcanism in 111(B5):B05314. doi:10.1029/2005JB003748
Afar, Ethiopia. Earth Planet Sci Lett 292:409–418 Keir D, Hamling IJ, Ayele A, Calais E, Ebinger C, Wright TJ, Jacques
Field L, Barnie T, Blundy J, Brooker RA, Keir D, Lewi E, Saunders K E, Mohamed K, Hammond JOS, Belachew M, Baker E, Rowland
(2012) Integrated field, satellite and petrological observations of the JV, Lewi E, Bennati L (2009) Evidence for focused magmatic
November 2010 eruption of Erta Ale. Bull Volcanol 74:2251–2271 accretion at segment centers from lateral dike injections captured
Fubelli G, Dramis F (2011) Coseismic surface faulting in the Kara Kore beneath the Red Sea rift in Afar. Geology 37:59–62
area (Wollo) caused by the May 29, 1961 Earthquake. In: Acocella Keir D, Pagli C, Bastow I, Ayele A, Ababa E (2011) The magma-
V, Bekele A, Coltorti M (eds) Pre‐conference excursion guide: assisted removal of Arabia in Afar: evidence from dike injection in
tectonic landforms and volcanism in the Southern Afar. Iag regional the Ethiopian rift captured using InSAR and seismicity. Tectonics
conference 2011 geomorphology for human adaptation to changing 30 (2):TC2008. doi:10.1029/2010TC002785
274 G. Corti et al.
Keir D, Bastow I, Pagli C, Chambers E (2013) The development of wave dispersion, teleseismic traveltime and normal-mode splitting
extension and magmatism in the Red Sea rift of Afar. Tectono- function measurements. Geophys J Int 184:1223–1236
physics 607:98–114 Rogers N, Macdonald R, Fitton J, George R, Smith R, Barreiro B
Kurz T, Gloaguen R, Ebinger C, Casey M, Abebe B (2007) (2000) Two mantle plumes beneath the East African rift system: Sr,
Deformation distribution and type in the Main Ethiopian Rift Nd and Pb isotope evidence from Kenya Rift basalts. Earth Planet
(MER): a remote sensing study. J Afr Earth Sci 48:100–114 Sci Lett 176:387–400
Lahitte P, Gillot PY, Courtillot V (2003) Silicic central volcanoes as Rooney T, Herzberg C, Bastow I (2012a) Elevated mantle temperature
precursors to rift propagation: the Afar case. Earth Planet Sci Lett beneath East Africa. Geology 40:27–40
207:103–116 Rooney T, Hanan BB, Graham DW, Furman T, Blichert-Toft J,
Le Gall B, Ahmed Daoud M, Rolet J, Moussa Egueh N (2011) Large- Schilling J-G (2012b) Upper mantle pollution during Afar Plume-
scale flexuring and antithetic extensional faulting along a nascent Continental rift interaction. J Petrol 53:365–389
plate boundary in the SE Afar rift. Terra Nova 23:416–420 Rooney T, Mohr P, Dosso P, Hall (2013) Upper mantle pollution
Leroy S, Razin F, Autin J, Bache F, d’Acremont E, Watremez L, Robinet during Afar Plume-Continental rift interaction. Geochim Cosmo-
J, Baurion C, Denèle Y, Bellahsen N, Lucazeau F, Rolandone F, chim Acta 102(1):6588. doi:10.1016/j.gca.2012.08.019
Rouzo S, Serra Kiel J, Robin C, Guillocheau F, Tiberi C, Basuyau C, Rowland JR, Baker E, Ebinger C, Keir D, Kidane T, Biggs J, Hayward
Beslier M-O, Ebinger C, Stuart G, Ahmed A, Khanbari K, Al Ganad I, N, Wright T (2007) Fault growth at a nascent slow spreading ridge:
de Clarens P, Unternehr P, Al Toubi A, Al Lazki A (2012) From 2005 Dabbahu rifting episode, Afar. Geophys J Int 171:1226–1246
rifting to oceanic spreading in the Gulf of Aden: a synthesis. Arab J Ruegg JC, Lépine JC, Tarantola A (1979) Geodetic measurements of
Geosci. http://dx.doi.org/10.1007/s12517-011-0475-4 rifting associated with a seismo-volcanic crisis in Afar. Geophys
Makris J, Ginzburg A (1987) The Afar Depression: transition between Res Lett 6:817–820
continental rifting and sea floor spreading. Tectonophysics 141: Siebert L, Simkin T, Kimberly P (2010) Volcanoes of the world, 3rd
199–214 edn. University of California Press, Berkeley
Manighetti I, Tapponnier P, Gillot PY, Jacques E, Courtillot V, Armijo R, Soliva R, Schultz RA (2008) Distributed and localized faulting in
Ruegg JC, King GCP (1998) Propagation of rifting along the Arabia- extensional settings: insight from the North Ethiopian Rift–Afar
Somalia boundary: into Afar. J Geophys Res 103(B3):4947–4974 transition area. Tectonics 27:TC2003. doi:10.1029/2007TC002148
Manighetti I, Tapponnier P, Courtillot V, Gallet Y, Jacques E, Gillot P- Souriot T, Brun J-P (1992) Faulting and block rotation in the Afar
Y (2001) Strain transfer between disconnected, propagating rifts in triangle, East Africa: the Danakil “crank-arm” model. Geology
Afar. J Geophys Res 106:13613–13665 20:911–914
McClusky S, Reilinger R, Ogubazghi G, Amleson A, Healeb B, Talbot CJ (2008) Hydrothermal salt—but how much? Mar Petrol Geol
Vernant P, Sholan J, Fisseha S, Asfaw L, Bendick R, Kogan L 25:191–202
(2010) Kinematics of the southern Red Sea-Afar Triple junction and Tapponnier P, Armijo R, Manighetti I, Courtillot V (1990) Bookshelf
implications for plate dynamics. Geophys Res Lett 37:L05301. faulting and horizontal block rotations between overlapping rifts in
doi:10.1029/2009GL041127 southern Afar. Geophys Res Lett 17:1–4
Medynski S, Pik R, Burnard P, Williams A, Vye-Brown C, Ferguson Tarantola A, Ruegg JC, Lépine JC (1979) Geodetic evidence for rifting
D, Blard P-H, France L, Yirgu G, Seid JI, Ayalew D, Calvert A in Afar: a brittle-elastic model of the behaviour of the lithosphere.
(2013) Magmatic cycles and development of rift topography of the Earth Planet Sci Lett 45:435–444
Manda-Hararo segment (Afar): Insights from cosmogenic 3He. Tesfaye S, Harding D, Kusky T (2003) Early continental breakup
Earth Planet Sci Lett 367:133–145 boundary and migration of the Afar triple junction, Ethiopia. Geol
Mohr P (1978) Afar. Ann Rev Earth Planet Sci 6:145–172 Soc America Bull 115:1053–1067
Mohr P (1987) Patterns of faulting in the Ethiopian Rift Valley. Thurmond AK, Abdelsalam MG, Thurmond JB (2006) Optical-radar-
Tectonophysics 143:169–179 DEM remote sensing data integrated for geological mapping in the
Morton WH, Black R (1975) Crustal attenuation in Afar. In: Pilger A, Afar Depression, Ethiopia. J Afr Earth Sci 44:119–134
Rosler A (eds) Afar Depression of Ethiopia. Schweizerbart, Waltham T (2010) Afar triangle: rift valleys and volcanoes over plate
Stuttgart, pp 55–61 divergence. In: Migoń P (ed) Geomorphological Landscapes of the
Nobile A, Pagli C, Keir D, Wright TJ, Ayele A, Ruch J, Acocella V World. Springer Science+Business Media, pp 183–190
(2012) Dike-fault interaction during the 2004 Dallol intrusion at the Williams FM, Williams MAJ, Aumento F (2004) Tensional fissures
northern edge of the Erta Ale Ridge (Afar, Ethiopia). Geophys Res and crustal extension rates in the northern part of the Main
Lett. http://dx.doi.org/10.1029/2012GL053151 Ethiopian Rift. J Afr Earth Sci 38:183–197
Nyblade AA, Robinson SW (1994) The African superswell. Geophys Wolfenden E, Ebinger C, Yirgu G, Deino A, Ayalew D (2004)
Res Lett 21:765–768 Evolution of the northern main Ethiopian rift: birth of a triple
Pagli C, Wright TJ, Ebinger CJ, Yun S-H, Cann JR, Barnie T, Ayele A junction. Earth Planet Sci Lett 224:213–228
(2012) Shallow axial magma chamber at the slow-spreading Erta Wolfenden E, Ebinger C, Yirgu G, Renne P, Kelley S (2005) Evolution
Ale Ridge. Nat Geosci 5:284–288 of a volcanic rifted margin: Southern Red Sea, Ethiopia. Geol Soc
Pagli C, Hua W, Wright TJ, Calais E, Lewi E (2014) Current plate Am Bull 117:846–864
boundary deformation of the Afar rift from a 3D velocity field Wright T, Ebinger C, Biggs J, Ayele A, Yirgu G, Keir D, Stork A
inversion of InSAR and GPS. J Geophys Res (in press) (2006) Magma-maintained rift segmentation at continental rupture
Pik R, Marty B, Carignan J, Lavé J (2003) Stability of the Upper Nile in the 2005 Afar dyking episode. Nature 442:291–294
drainage network (Ethiopia) deduced from (U-Th)/He thermochr- Wright TJ, Sigmundsson F, Pagli C, Belachew M, Hamling IJ,
onometry: implications for uplift and erosion of the Afar plume Brandsdóttir B, Keir D, Pedersen R, Ayele A, Ebinger C, Einarsson
dome. Earth Planet Sci Lett 215:73–88 P, Lewi E, Calais E (2012) Geophysical constraints on the dynamics of
Pizzi A, Pontarelli L (2007) Geometria del rift nel margine spreading centres from rifting episodes on land. Nat Geosci 5:242–250
meridionale dell’Afar: Dire Dawa (Etiopia). Rend Soc Geol It Zanettin B, Justin-Visentin E (1975) Tectonical and volcanological
4:288–289 evolution of the western Afar margin (Ethiopia). In: Pilger A, Rosler
Ritsema J, Deuss A, van Heijst H, Woodhouse J (2011) S40RTS: a A (eds) Afar Depression of Ethiopia. Schweizerbart, Stuttgart,
degree-40 shear-velocity model for the mantle from new Rayleigh pp 300–309
Morphometric Characteristics and Hydrology
of Selected Ethiopian Rift Lakes 16
Tenalem Ayenew and Merhawi GebreEgziabher
Abstract
In the Main Ethiopian Rift (MER), intra-rift faulting and associated volcanic activities resulted
in the formation of volcano-tectonic structural depressions, responsible for the formation of
many lakes on the rift floor. These lakes are bordered to the east and west by elevated
highlands where the main tributary feeder rivers originate. The MER is occupied by seven
major lakes, namely, from south to north, Chamo, Abaya, Awasa, Shala, Abijata, Langano,
and Ziway. Most of these lakes are closed basins and interconnected by groundwater, through
a NE–SW-aligned regional fault system, whereas Abijata Lake receives surface water from
Ziway and Lagano lakes by two small streams. The geomorphic characteristics and size of
these lakes changed on time scales of decades to millennia in response to different factors such
as volcanic activity, tectonics, climate change and, very recently, human impact (mainly water
withdrawal, deforestation, and irrigation). The most prominent example is the dramatic change
of Abijata, the size of which decreased dramatically in the last decades.

Keywords
Rift lakes Groundwater Hydrological budget Tectonics Ethiopia
16.1 Introduction ranges, with precipitous edges dissected by numerous rivers,

to the rift valley which is occupied by a series of lakes. The
The distribution of Ethiopian lakes is intimately linked to the altitude ranges from 4,533 m a.s.l. at Ras Dejen (Simen
geomorphological setup of the country. Ethiopia has three Mountains) in the northern highland mountains to about
major physiographic regions: the highlands (plateaus), the 120 m below sea level in the Dallol depression which is a
rift valley and the peripheral lowlands which are often wide part of the northern end of the Ethiopian Rift valley.
plains devoid of lakes. The majority of the Ethiopian lakes The Ethiopian Rift can be divided into five hydrological
are confined within the rift valley (Fig. 16.1). systems: Turkana, Lake Chew Bahir, Lake/swamps Main
The geomorphological setting is highly controlled by Ethiopian Rift (MER), Awash River basin and the Danakil
different episodes of volcano-tectonic activity that shaped basin. In these basins, there are over 15 natural lakes. The
the country in the Cenozoic, with the influence continuing MER transects the uplifted Ethiopian highlands for a dis-
till the present day. The volcano-tectonic activities resulted tance of around 1,000 km, extending from southern Afar
in the formation of a spectacular rift system and one of the across the broad basins and volcanic ranges to the southern
largest volcanic plateau of the world. There is a wide vari- watershed of Chamo Lake in the center of the country. The
ation in altitude that ranges from the highland mountain MER floor is occupied by a series of lakes that display wide
variations in chemistry and morphometric characteristics
(Ayenew 2009a, b).
T. Ayenew (&) M. GebreEgziabher In this study, emphasis is given to seven major lakes
School of Earth Science, Addis Ababa University, P.O. Box 1176, occupying the floor of the MER. These include Chamo,
Addis Ababa, Ethiopia Abaya, Awasa, Shala, Abijata, Langano, and Ziway. These
e-mail: Tenalema@yahoo.com
276 T. Ayenew and M. GebreEgziabher
Fig. 16.1 Location of Ethiopian

lakes and wetlands associated
with the Great East African rift
and adjacent highlands (key to
wetlands: 1 Turkana, 2 Chew
Bahir, 3 Chamo, 4 Abaya, 5
Awasa–Shallo, 6 Chitu, 7 Shala,
8 Abijata, 9 Langano, 10 Ziway,
11 Bishoftu crater lakes, 12
Beseka, 13 Afambo–Abe group,
14 Afrera, 15 Asele, 16 Ashenge,
17 Hayk-Ardibo, 18 Tana, 19
Wonchi–Dendi, 20 Lakes of the
Bale Mountains, 21 Ciarciar,
22 Haromaya–Adele–Finkle)
lakes occupy the rift floor in three separate hydrological et al. 2006) and form a unique system within the rift. Similar
basins: Abaya–Chamo, Awasa and Ziway–Shala (ZSB). hydrogeological setup is evident in the Kenyan rift valley
Particular emphasis is given to their morphometric charac- where major faults act as subsurface conduits connecting the
teristics and hydrology. rift lakes (Ayenew and Becht 2008).
16.2 Distribution of Lakes and Major 16.3 Overview of the Evolution,

Wetlands Geomorphology, and Climate
of the MER
The Ethiopian rift lakes are nested in en echelon fashion,
extending in the south from Chew Bahir Lake on the Ethi- 16.3.1 Evolution and Geomorphology
opia and Kenya border to the extreme north of the Danakil
Depression. Figure 16.1 shows the distribution of major The Ethiopian Rift is part of the Great East African Rift
Ethiopian lakes and wetlands. The major lakes are confined which is one of the most extensive rift systems on earth. It is
in the rift floor and are situated within different hydrological an impressive geological feature extending from the Jordan
basins; from north to south they are: (1) the lakes of the River in Middle East southward to Mozambique in southeast
Danakil basin (Asele and Afrera); (2) Awash river basin Africa, approximately 6,400 km long and 48–64 km wide.
(Abe, Gamari, Afambo, Beseka); (3) ZSB (Ziway, Langano, The Ethiopian Rift provides a unique opportunity to study
Abijata and Shala); (4) Awasa Lake basin; (5) Abaya– the transition between continental rifting to the south in
Chamo; (6) Chew Bahir and Omo basin (Turkana Lake). Kenya and new seafloor spreading to the north in Afar in the
Apart from these, there are very small crater lakes scattered onshore extension of the Red Sea and Gulf of Aden (Mac-
on the rift floor and escarpment areas. The crater lakes have kenzie et al. 2005).
small catchment areas and steep crater rim slopes. Crater Intra-rift faulting and associated volcanic activities
lakes are not included in this study. resulted in the formation of volcano-tectonic structural
In terms of surface water, most of these lakes are not depressions, responsible for the formation of many lakes in
interconnected owing to topographic and geological factors, the rift floor (Ayenew 2009a, b). The rift lakes are bordered
whereas their groundwaters are interconnected by a NE– to the east and west by elevated highlands where the main
SW-aligned regional fault system associated with tectoni- tributary feeder rivers originate. Magnificent volcanic cen-
cally active region (Ayenew 1998, 2009a, b; Alemayehu ters dot the floor of the rift valley (Fig. 16.2), separating the
16 Morphometric Characteristics and Hydrology … 277
Fig. 16.2 Ziqwala volcano,

2,989 m a.s.l., dominating the
landscape around the Koka
reservoir
different lakes. The volcanic centers are associated with remnants of large lakes evolved into separate systems as they
hydrothermal activities. In some cases, hot thermal waters receded through time (Street 1979; Chap. 17, this volume).
feed the lakes at the foot of volcanic centers and caldera rims Little is known about the evolution and the mode of origin of
(Ayenew and Becht 2008). the majority of Ethiopian lakes, whereas the late Pleistocene
The MER contains lakes of different hydrological and and Holocene hydrological and geomorphological changes
morphometric characteristics. Most lakes are aligned along of the ZSB are well documented (Benvenuti et al. 2002,
the NE trending tectonic depressions characterized by active 2005; Chap. 17, this volume). From lacustrine deposits, it
fault systems and volcanic centers originated during the appears that Chamo and Abaya were one single lake in
Cenozoic (since 65 million years ago). Figure 16.3 shows recent geological past. The eruption of central volcano and
the major lakes of the MER and selected meteorological different episodes of tectonism separated these lakes in the
stations. Pliocene. Most of the MER lakes are occupying volcano-
The general regional geological setting of the MER tectonic depressions. Some of them are the center of tec-
involves volcanism, rifting, Quaternary lake level fluctuation tonically affected caldera. Typical examples are Awasa and
and deposition of fluvial and volcano–lacustrine sediments. Shala. The volcano-tectonic setting controls the considerable
At the beginning of the Cenozoic and mainly during the variations in volume and size of these lakes. Table 16.1
Oligocene (Chap. 2 this volume), the region was deluged displays clearly this variation.
under molten lava and thick beds of explosive rocks with Most of the rift lakes are localized within a closed basin
intense tectonic activity, which initiated the formation of the fed by both perennial rivers and seasonal streams. Large
rift valley. The wide areas buried under volcanic materials, highland rivers are the source of sustained supply to the
the thickness and variety of the rock sequences and the major rift lakes. The amount and distribution of highland
prolonged duration of the eruptions make the Ethiopian rainfall strongly control the level and the size of these lakes
plateaus and the rift one of the world’s most notable volcanic (Street 1979; Ayenew 1998).
regions. Although these episodes of volcanic activity predate
the formation of the rift itself, they still played an important
role in sculpting and enriching the magnificent scenery, 16.3.2 Climate
renowned for imposing escarpments, some of them rela-
tively old and weathered while other still fresh, sharp, and Generally, the climate of the MER is semiarid to sub-humid,
steep. with distinct wet and dry seasons (Chap. 3, this volume).
The size of the lakes changed on time scales of decades to The adjacent highlands experience high rainfall and lower
millennia. Large rift lakes covered wide depressions in evaporation. During the wet season, northwesterly monsoon
closed basins in the Pleistocene wet intervals (Grove et al. winds bring high rainfall. In general, the high rainfall lasts
1975; Chap. 17, this volume). The present-day rift lakes are from June to September (Chap. 3, this volume). The dry
Fig. 16.3 Location of selected

meteorological stations and the
main MER lakes (key to major
lakes: 1 Ziway, 2 Langano, 3
Abijata, 4 Shala, 5 Awasa–Shallo,
6 Abaya, 7 Chamo)
season is from December to April. Table 16.2 shows the ultimate source of replenishment of the rift lakes in the form
long-term average rainfall and temperature in selected sta- of groundwater, river inflows, and surface runoff. Generally,
tions on the rift floor and adjacent highlands. different climatic conditions characterize the highlands, the
Average annual temperature varies from over 20 °C in the escarpment, and the rift valley. The lake level is strongly
central MER to less than 18 °C in the adjacent highlands. controlled by highland rainfall. Inter-annual variation in the
Annual rainfall varies from around 630 mm in the central lake levels is the direct reflection of the changes in the
MER to more than 1,500 mm in the southwestern and highland rainfall, with the exception of a few lakes which are
western highlands. The national average annual precipitation being used for irrigation and those influenced by irrigation
is 744 mm (NMSA 2005). The highland rainfall is the return flows and soda ash abstraction (Ayenew 2009a, b).
Table 16.1 Basic morphometric data of Ethiopian rift lakes

No. Lakes Altitude (m a.s.l.) Area (km2) Maximum depth (m) Average depth (m) Catchment area (km2) Volume (km3)
1 Ziway 1,636 440 8.9 2.5 7,414 1.5
2 Langano 1,585 230 47.9 17 2,000 38
3 Abijata 1,578 135 14.2 7.6 10,740a 0.954
4 Shala 155 370 266 86 2,300 37
5 Awasa 1,680 80 22 11 136 1.3
6 Abaya 1,285 1,140 13.1 8.6 16,342 8.2
7 Chamo 1,235 317 13 10.3 18,575a –
a
The catchment areas of Chamo and Abijata include Abaya and Ziway catchment, respectively. (Modified from Ayenew 2009a, b)
Table 16.2 Summary of long-term average monthly precipitation and temperature recorded at selected stations in the three main physiographic
regions (rift, escarpments, and highlands)
Station Location Year Jan Feb Mar Apr May Jun July Aug Sept Oct Nov Dec Annual
A. Precipitation (mm)
Asela Central MER 1982–2002 27 50 92 127 126 98 116 114 133 86 33 25 1,027
Awasa Eastern 1970–2004 61 0.1 38 113 44 104 147 172 154 158 21 50 1,061
highland
Butajira Western 1970–2004 38 67 138 127 114 121 172 174 120 46 11 14 1,140
highland
Ziway Central MER 1970–2004 17 31 56 75 72 83 145 122 86 39 2.4 4 732
B. Temperature (°C)
Arba Southern MER 1960–2002 23 24 24 24 22 22 22 22 23 23 22 22 22.7
Minch
Asela Central MER 1988–2003 20 20 21 21 20 20 19 19 19 19 18 19 19.5
Awasa Eastern 1960–2004 14 16 16 16 17 16 15 15 16 15 14 14 15.3
highland
Butajira Western 1972–2004 18 19 19 19 19 19 18 18 18 18 18 18 18.4
highland
Ziway Central MER 1970–2004 19 20 21 22 22 21 20 20 20 20 19 19 20.2
16.4.2 Hydrology
16.4 Morphometric Characteristics
and Hydrology of the MER Over the last few decades, a number of studies have been
Lakes carried out to study the hydrology of the major Ethiopian
lakes (Ayenew 1998, 2009a, b). These studies provided the
16.4.1 Morphometric Characteristics general picture of the different fluxes and water balance of
the dynamic rift valley lakes. The methodologies used to
Table 16.1 summarizes the basic morphometric characteris- quantify the water balance of each lake are addressed in
tics of the seven major lakes. Due to differences in geomor- detail in the respective referred works. Integrated hydro-
phological setting resulting from volcano-tectonic processes geological investigations involving hydrogeological map-
in the Cenozoic and Quaternary, the rift lakes display large ping, remote sensing techniques, isotope, and hydrochemical
variations in size and volume. They have also large differ- data helped to recognize the interaction of the lakes with
ences in their hydrochemistry (Chap. 18, this volume). Most groundwater and flow paths. The source of the fluxes and the
of the closed lakes are highly saline, while the open lakes are relationships between the lakes with the groundwater and the
fresh. The deepest lake in the Ethiopian rift is Shala (266 m), hydrological conditions of the three main lakes basins
while the shallowest is Ziway (less than 10 m) if we do not (Ziway–Shala, Awasa, Abaya–Chamo) are described below.
consider Chew Bahir, which used to be a vast lake when it Table 16.3 summarizes the water balance of some of the
was discovered in 1888 by Samuel Teleki, and now is partly a rift lakes. Due to the availability of pertinent data, the central
swamp and for the most part a muddy salt plain. MER lakes (Ziway, Abijata, Langano, Shala, and Awasa)
Table 16.3 Long-term average annual water balance of selected rift lakes (106m3)
Lake Area (km2) Inflow Outflow Source
Pl Ri Gi Sr El Ro Go A V
Abaya 1,162 556 Very high Very high Very high 1,900 Rare outflow Medium – – 3, 5
Chamo 551 406 High High High 900.9 Rare outflow High Ng – 3, 5
Awasa 100 106 83.1 High 83.7 132 Closed 58* Ng 82.8 6,7
Abijata 180 113 230 26.8 15 372 Closed 1.2 13 −1.4 2,4
Langano 230 186 212 135.4 Very high 463 46 18.9 – 5.5 2,4
Shala 370 232 245 237.6 40 781 Closed None Ng −26.4 2,4
Ziway 440 323 656.5 80.5 48 890 184 14.6 28 −8.6 2,4
Pl precipitation on the lake; Ri inflow from rivers; Gi groundwater inflow; Sr inflow from surface runoff; El lake evaporation; Ro outflow in river
outlets; A abstraction; V net flux; Ng negligible [Sources (1) Tessema (1998); (2) Ayenew (1998); (3) Bekele (2000); (4) Ayenew (2001); (5)
Woldemariam (2004); (6) Gebreegziabher (2005); (7) Ayenew and Gebreegziabher (2006)]
are relatively better studied. The water balance assessment 2011), whereas the expansion of Awasa Lake in the recent
was made on the basis of extensive long-term hydrometeo- years has not yet been completely understood.
rological data (National Meteorological Database from 1965 Some Ethiopian Rift lakes, particularly those located in a
to 2003) and numerical modeling (Ayenew 2001; Ayenew terminal position, have undergone significant changes
and Gebreegziabher 2006). In addition, hydrochemical and (Ayenew 2011). The major rift valley lakes show contrasting
isotope studies provided information on the subsurface lake level trends, the most extreme of which is Lake Abijata.
hydraulic links of the lakes (Darling et al. 1996; Tessema The lakes showing the most dramatic changes are those
1998; Ayenew 2003; Fikre 2006). Most of the data is where the water is used for agricultural and industrial pur-
extracted from Ayenew 2009a, b. poses. In a few cases, land degradation and neotectonic
In recent years, the size of the lakes has changed signif- movements have caused changes (Ayenew and Legesse
icantly (Fig. 16.4). The most prominent example is the 2007). The reduction of the level of Lake Abijata and the
dramatic change of Abijata Lake (Fig. 16.5). Ziway, Chamo, complete disappearance of Lake Haromaya (located in
and Abaya lakes have also shown slight changes. On the the Eastern Ethiopian highlands Fig. 16.1) are examples of
contrary, irrigation return flows resulted in expansion of the severe consequences of excessive pumping of water.
other lakes. The typical example is Beseka Lake which is Abijata’s shallow depth and terminal position make it
located in the Awash basin, north of Ziway Lake (Ayenew particularly susceptible to changes in climate and pumping.
Fig. 16.4 Lake level variation

through time. All lakes show a
moderate increasing trend with
Abijata recovering from a few
years of marked decrease
(Ayenew 2004)
Fig. 16.5 Recession of Abijata

Lake: a enhanced satellite image
showing different strandlines
representing shorelines at
different times; b reconstructed
shoreline positions at different
years
It receives most of its water from the Ziway Lake catchment due to evaporative enrichment and the supply of saline water
through the Bulbula River and a smaller amount from from hydrothermal springs. Larger rivers feed Ziway and
Langano Lake through the Horakelo River (Figs. 16.5 and Shala lakes, whereas smaller streams feed Langano Lake.
16.6). Being a closed lake, the only significant natural water Shala, Langano and Abijata lakes have also a centripetal
loss is through evaporation. In the absence of human inter- network of ephemeral streams and gullies draining the inner
ference, the lake maintained the natural balance between side of the caldera.
inflow and evaporation until the mid-1970s; that is before
irrigation activities started in the Ziway area and the estab- 16.4.2.1 Ziway Lake
lishment of the Abijata Soda Ash Factory in the mid-1980s. The main input to Ziway Lake comes from the Ketar and Meki
The reduction of the level of Abijata is clearly visible from rivers and direct precipitation. The main outputs are evapo-
old shorelines seen in satellite images that make it possible ration, discharge via the Bulbula River and groundwater
to reconstruct the size of the lake in different years leakage through faults and paleo-channels. Hydrochemical
(Fig. 16.5, but see also Fig. 17.2 in Chap. 17, this volume). and isotopic evidence reveals southward groundwater
The maximum reduction in the area, and hence the drop of migration from Ziway to Abijata (Ayenew 2003).
the level of Abijata Lake, coincided with the large-scale The annual direct precipitation over the lake is 323 mcm
water abstraction for soda production from the lake itself and (million cubic meter). Long-term average annual inflow from
for irrigation from Ziway Lake and its basin after the 1980s Ketar and Meki rivers is 392 and 265 mcm, respectively. The
(OEPO 2005). Data from a few meteo-stations in the MER remaining ungauged catchments are flat and located in the
floor and margins show also that during the last three-four rift, with low precipitation. However, during high precipita-
decades the air temperature has remarkably increased (Billi tion events overland flow reaches the lake from the northern
and Dramis 2000, Chap. 3, this volume). and western part of its catchment. The contribution from the
The ZSB is a closed basin with a catchment area of about ungauged catchments is estimated at 48 mcm annually.
13,000 km2, out of which 1,443 km2 is covered with per- Annual lake evaporation and the Bulbula River outflow
manent open water bodies (Ayenew 1998). Ziway, Langano, are 890 and 184 mcm, respectively. The annual abstraction
Abijata and Shala lakes occupy volcano-tectonic depressions of water for irrigation is around 28 mcm (Ayenew 1998).
at the center of the basin. There are also small lakes and Deribessa (2006) indicated that the groundwater outflow and
swampy areas in the rift and within its bordering escarp- pumping rate for irrigation are higher. The current abstrac-
ments (Fig. 16.6). Lakes Ziway and Langano are open, while tion from the lake and feeder rivers is expected to be over
Shala and Abijata are closed. Abijata is fed by outflows from 50 mcm. In order to attain equilibrium, leakage into the
Ziway and Langano lakes through the Bulbula and Horakelo subsurface is considered to play a relevant role, which is
rivers, respectively. Ziway Lake is the freshest of all the four consistent with the isotope and hydrochemical studies
lakes, while Shala and Abijata are highly alkaline and saline (Ayenew 2003; Deribessa 2006).
Fig. 16.6 Lakes of Ziway–Shala

basin (Ayenew 2009a, b)
16.4.2.2 Abijata Lake coming from the eastern Arsi Highlands. At least four major
The main input to the lake comes from the Horakelo and perennial rivers (Huluka, Lepis, Gedemso and Kersa) feed
Bulbula rivers and direct precipitation. The annual precipi- the lake. There are also many ungauged seasonal streams
tation on the lake accounts to 112 mcm. The long-term and springs, all joining the lake from the eastern side. By
average annual inflows from the Bulbula and Horakelo riv- extrapolating the discharge record of the gauged rivers, the
ers are 184 and 46 mcm, respectively. Surface runoff from total annual inflow from the ungauged catchments is esti-
the remaining ungauged catchments is 15 mcm. The main mated at 315 mcm. High-discharge hot springs enter into the
loss is through evaporation (372 mcm annually). The lake lake in the northern shore. These include Oitu, Bole, Tuffa,
water is being pumped for soda ash extraction. The recent and the springs east of Langano peninsula. From limited
pumping rate is unknown. In 1998, the total annual pumping field spring discharge measurements, the annual input from
rate was estimated at 13 mcm. Using groundwater flow the springs is estimated at 63 mcm.
modeling, one may infer insignificant subsurface outflow Annual lake water evaporation and outflow through the
from the lake (Ayenew 2001). The balance between total Horakelo River are 463 and 46 mcm, respectively.
inflow and outflow is positive, indicating that there is Groundwater flow models revealed the existence of large
additional inflow from groundwater, which is likely to be leakage in the southern shore, which ultimately flows into
from the northern and western sides. Lake Shala (Ayenew 2001).
16.4.2.3 Langano Lake 16.4.2.4 Shala Lake

The annual contribution from precipitation is 146 mcm. Shala Lake is a groundwater-controlled system, but rivers
There is no discharge data for the few tributary rivers and precipitation play important roles as well. Annual input
from precipitation is 232 mcm. The main tributary rivers are Gebreegziabher 2006). Shallo Lake is a remnant of a much
Dijo and Awade, which originate from the western and larger lake that included the surrounding swampy plain
eastern highlands, respectively. The latter is not gauged. The covering the floor of the caldera northeast of Awasa town
annual inflow from Dijo is 107 mcm. The recent inflow is (Telford and Lamb 1999). Nowadays, Shallo Lake is on the
quite low due to upstream diversions for local irrigation. The verge of extinction due to sedimentation and land use
southern and southeastern catchment is similar to that of changes in the catchment area.
Ketar (Ziway Lake tributary, see above) in terms of total The caldera floor is distinctly separated from the sur-
precipitation, land use, and geology though the topography rounding highlands by scarps and steep mountain slopes.
is flatter. By extrapolating the discharge data from Ketar, the The elevation difference between the floor of the caldera and
annual contribution from the Awade River could be around the bordering scarps and volcanic complexes ranges from
138 mcm. The remaining ungauged catchment generates 200 to 900 m. Rivers drain into the lakes from the moun-
little runoff since it is located in the flat lowland, with low tainous areas. The rivers from the eastern and southeastern
precipitation. If 5 % of the weighted annual precipitation highlands feed Shallo Lake throughout the year. Overflow
(785 mm) of the ungauged catchment is assumed to reach from Shallo Lake drains into Awasa Lake via the Tikur
the lake as surface runoff, the annual input will be 40 mcm. Wuha (Black Water) river. In the eastern, western, north-
Many hot springs enter into the lake in the eastern and western, and southern sides of the catchment, no perennial
southern shores. From intermittent discharge measurements, rivers reach the lake. Seasonal streams may also terminate in
annual inflow from springs is estimated at 18 mcm (Ayenew wide-open fractures before reaching the lake (Fig. 16.7).
1998). As Awasa Lake occupies a closed basin, there is no
The only loss from the lake is through evaporation, surface water outflow. The input comes from rivers, direct
accounting for 781 mcm annually. This brings the difference precipitation, and groundwater. The only gauged river is
between the total inflow and outflow to 247 mcm. The high Tikur Wuha connecting Shallo with Awasa Lake. The con-
positive residual can be explained in terms of groundwater ventional water balance estimation and hydrological models
input. Lakes Shala and Langano are certainly controlled by provided a good picture of the relative importance of the
groundwater more than any other lake in the ZSB. For Shala, different fluxes (WWSDE 2001; Ayenew and Gebreegizab-
the major groundwater input comes from Langano through her 2006). The annual input from surface runoff and pre-
faults in the northeastern shore and from the elevated areas cipitation account for 145 and 88 mcm, respectively.
in the southeast, south, and probably northwest. There is Chernet (1982) reported the flow rate of some thermal
strong evidence of groundwater inflow from the adjacent springs in the Awasa catchment. The contribution of the
Hawassa catchment (Darling et al. 1996; Ayenew and Ge- springs to the hydrological balance of Awasa Lake is likely
breegziabher 2006). The existence of large regional faults less than 5 %.
and the relative lower topographic position (1,550 m a.s.l.) The annual evaporation is 151 mcm (Geremew 2000). On
favors large volumes of groundwater to converge toward the basis of hydrological models, the net annual groundwater
Shala Lake from different directions. outflow from the lake was estimated to be 58 mcm (Ayenew
and Gebreegziabher 2006).
16.4.2.5 Awasa Lake Basin
The Awasa basin represents a large collapsed caldera bor- 16.4.2.6 The Abaya–Chamo Basin
dered by highlands to the north and east. The center of the Abaya (Fig. 16.8) and Chamo (Fig. 16.9) are marvelous
caldera is occupied by Awassa and Shallo (Cheleleka) lakes. lakes ringed by savannah plains bordered by large mountain
The elevation of Awasa Lake is 1,680 m a.s.l., representing ranges to the east and west. Abaya and Chamo lakes occupy
the culmination of the Ethiopian Rift floor level occupied by elongated tectonic grabens and are dotted with many islands.
lakes. The floor of the caldera is faulted and dotted by According to the Statistical Abstract of Ethiopia for 1967/68,
volcanic hills. It is believed that this area forms the regional Abaya Lake is 60 km long and 20 km wide with a surface
groundwater divide of the entire Ethiopian rift system. Lakes area of 1,160 km2. It has a maximum depth of 13 m with an
south of Awasa basin leak in the form of groundwater to the elevation of 1,268 m a.s.l.
south, while Awasa Lake feeds Shala Lake to the north Together with the lakes, the total area of the basin is
through regional axial faults (Ayenew 1998). 18,600 km2, with a freshwater water volume of over 12 km3.
Awasa is the smallest of all the major lakes south of the A ridge, called Angels Bridge, separates Abaya from Chamo
Awash River Basin. The catchment has no surface water Lake. Abaya is the largest lake in the Ethiopian Rift. It is
outlet. The lake is about 15 km long and 5.5 km wide with a relatively less saline and is used for fishing and recreation.
maximum depth of 22 m and covers 80 km2 (NUPI 1994). The lake is fed by the Bilate River which flows from the
The surface area of Shallo Lake is 12 km2. The elevation north and other rivers from the eastern and western high-
difference between the two lakes is around 5 m (Ayenew and lands. The main perennial rivers include Uraye, Shope, Hare,
Fig. 16.7 The Awasa Lake

catchment (1 Corbetti volcano; 2
Cheleleka hot springs; 3 Wondo
Genet Resort and hot springs)
Kola, Gidabo, and Gelana. Abaya is separated from Chamo Chamo Lake is fed by Sille and Kulfo Rivers. During
Lake by a land isthmus with a vertical offset of 65 m. extreme wet seasons, it overflows to Chew Bahir Lake
Occasionally, during high water levels of the wet seasons, through Metenfesha stream. The source of the Segan River,
the water of Abaya flows into Chamo Lake through a small which drains to Chew Bahir, lies east of the southern end of
channel. In the northern part of the basin, there is a large Chamo Lake. There is a water channel connecting Chamo
rhyolitic massif, Shire Volcano, west of Bilate River. Lake with the Segan River. The overflow occurs when the
Moderate temperature hot springs and weak fumaroles exist water level rises by about a few tens of centimeters. The
at its eastern and southern footslopes, respectively. Bodicho overflow takes place when the precipitation in the region is
hot springs are located on a fault in ignimbrite near the above long-term average. With increasing diversion of water
southeast base of Data Volcano. There are also Nech Sar from tributary rivers for irrigation, the overflow to Chew
springs near the town of Arba Minch. Dominantly, the hot Bahir and the surface water connection between Abaya and
springs emerge along the western marginal faults near the Chamo will likely cease to exist.
center of the graben. There are also three crater lakes in the northern part of the
Chamo Lake is located south of Abaya Lake, at an eleva- basin, close to the Bilate River. These are Tilo, Mecheferia,
tion of 1,235 m a.s.l. The lake is 26 km long and 22 km wide, and Budemeda lakes (Fig. 16.10). A fourth, dry crater is
with a surface area of 316 km2 and a maximum depth of 13 m located 5 km southwest of Mecheferia. These lakes are
(Woldemariam 2004). The lake water is more saline than that aligned along a NE–SW trending tectonic line, parallel to the
of Abaya Lake; however, the salinity is low as compared to the direction of the MER axis. All of them are maars or collapse
closed lakes of the ZSB. This indicates groundwater outflow structures above diatremes, extending to an igneous dyke at
to neighboring basins as in the case of Awasa Lake. greater depth (Lorenz 1986).
Fig. 16.8 Abaya Lake with the

volcanic ridge separating it from
Chamo Lake
Fig. 16.9 Chamo Lake within

the Netch Sar National Park
Comprehensive hydrological studies provided informa- watering, account for 97.6, 0.2, and 2.2 % of the total annual
tion on the water balance of Abaya and Chamo lakes (Bekele outflow.
2000, 2006; Bekele and Horlacher 2001; Woldemariam Woldemariam (2004) provided a different picture for the
2004). The main surface water input to Abaya Lake comes water balance of Abaya Lake. According to his study, the
from direct precipitation, rivers (Bilate, Uraye, Shope, Hare annual lake evaporation is 1,635 mm, which is 1,900 mcm
Kola, Gidabo, and Gelana) and springs. During wet seasons, over the 1,162 km2 total lake area. Precipitation, surface
Abaya Lake loses water to Chamo Lake. Bekele (2000) runoff and groundwater contribute 5, 14, and 81 % of the
estimated the water balance on the basis of a simple annual total inflow, respectively. The groundwater and lake
hydrological model. Precipitation and runoff contribute 27.2 evaporation account for 86 and 14 % of the annual outflow,
and 72.8 % of the total input, respectively. Evaporation, respectively. The long-term average annual precipitation of
pumping for water supply and other usage, such as livestock the area is 737 mm.
Fig. 16.10 Lakes of the Abaya–

Chamo basin
The water balance of Chamo Lake is not well understood. balance study shows that most lakes are interconnected in the
This is especially due to limited knowledge of the seasonal subsurface. The closed lakes are groundwater-dominated
surface inflow and outflow and the groundwater component. systems.
The main inputs are precipitation and inflows from Sille and Groundwater flow in the Rift is controlled by geological
Kulfo Rivers. The amount of groundwater that leaks to the structures, either directly via flows in the tensional faults or
Chew Bahir basin is unknown. Surface water outflow may through fluvial and lacustrine deposits whose pattern of
probably occur during the rainy season (Grove and Goudie occurrence is influenced by tectonics. The major ground-
1971). water conduits are faults trending parallel and sub-parallel to
the Rift axis. Based on piezometric surveying, hydrogeo-
logical field investigations, and ancillary isotope and hyd-
16.5 Conclusions rochemical data, it is concluded that most of the lakes are
hydraulically connected through rift fault systems.
Ethiopia is endowed with many natural lakes. Most of these The most important result of this study is the elaboration
lakes are confined within the Ethiopian Rift system. The of the intricate nature of the subsurface hydrology of the rift
volcano-tectonic process in the Cenozoic favored the for- system, as well as the interconnections of the lakes, whose
mation of a chain of lakes aligned along the NE–SW water balances are predominantly controlled by groundwa-
trending rift axis. These lakes have different hydrological ter. These results highlight the importance of studying the
and geomorphological settings. details of both the inflow and the outflow groundwater
The hydrology of the rift valley lakes is strongly con- components, including the installation of more piezometers
trolled by highland rainfall that feeds the lakes in the form of and seepage meters. Accordingly, the future sustainable
river flow, surface runoff, and groundwater inflow. The water management of the Rift Valley lakes basins must consider
the complex hydrological characteristics of the lakes, and the main Ethiopian rift (Ethiopia). In: Blum MD, Marriott SB,
their interrelations through the complex rift fault systems. Leclair SF (eds) Fluvial sedimentology VII, Special Publication
International Association of Sedimentologists vol 35, p 277–294
Indeed, attempting to develop and implement a water man- Billi P, Dramis F (2000) Recent climatic trends and soil erosion in some
agement plan for a given lake, without appropriately con- selected areas of Ethiopia. In: Feoli E, Pottier D, Zerihun Woldu
sidering the groundwater fluxes, will almost certainly lead to (eds) Sustainable development in drylands of East Africa, European
improper or erroneous water use practices. commission, DG XII, Science Research & Development, pp 145–
162
Acknowledgments The authors are indebted to Marek Kasprzak for Chernet T (1982) Hydrogeologica map of the lake region (with memo).
significantly improving many of this chapter figures. Ethiopian Institute of Geological Surveys, Addis Ababa
Darling WG, Gizaw B, Arusei MK (1996) Lake-groundwater relation-
ships and fluid-rock interaction in the African rift valley: isotopic
evidence. J Afr Earth Sci 22:423–431
References Deribessa A (2006) Groundwater and surface water interaction and
geoenvironmental changes in Ziway catchment. Unpublished M.Sc.
thesis, Addis Ababa University, Addis Ababa
Alemayehu T, Ayenew T, Kebede S (2006) Hydrochemical and lake Fikre S (2006) Hydrogeological system analysis in Ziway–Shala lakes
level changes in the Ethiopian rift. J Hydrol 316(1–4):290–300 area using hydrochemistry and isotope techniques, central Ethiopia.
Ayenew T (1998) The hydrogeological system of the lake district basin, Unpublished M.Sc. thesis, Addis Ababa University
central main Ethiopian rift. Published Ph.D thesis, Free University Gebreegziabher Y (2005) Assessment of the water balance of lake
of Amsterdam, The Netherlands. 259 pp Awassa catchmen, Ethiopia. M.SC thesis, International Institute for
Ayenew T (2001) Numerical groundwater flow modelling of the central Geo-Information Science and Earth Observation (ITC), Enschede
main Ethiopian rift basin. SINET: Ethiopian J Sci 24(2):167–184 Geremew G (2000) Engineering geological study of Hawassa town and
Ayenew T (2003) Environmental isotope-based integrated hydrogeo- its surroundings. M.Sc. thesis, Addis Ababa University, Addis
logical study of some Ethiopian rift lakes. J Radioanal Nucl Chem Ababa, 185 pp
257(1):11–16 Grove AT, Goudie AS (1971) Secrets of Lake Stephanie’s past.
Ayenew A (2004) Environmental implications of changes in the levels Geograph Mag 43:542–547
of lakes in the Ethiopian rift since 1970. Reg Environ Change Grove AT, Street FA, Goudie AS (1975) Former lake levels and
4:192–204 climate change in the rift valley of southern Ethiopia. Geogr J
Ayenew A (2009a) Natural lakes of Ethiopia. Addis Ababa University 141:177–194
Press, Addis Ababa, 256 pp Lorenz V (1986) On the growth of maars and diatremes and its
Ayenew T (2009b) Natural lakes of Ethiopia. Addis Ababa University relevance to the formation of tuff-rings. Bull Volcanol 48:265–274
Press, Addis Ababa Mackenzi GD, Thybo H, Maguire PKH (2005) Crustal velocity
Ayenew A (2011) Wetlands of Ethiopia: overview of environmental structure across the main Ethiopian rift: results from 2-dimensional
changes, vol I. Forum for Environment, Adis Ababa wide angle seismic modeling. Geophys J Int 162:994–1006
Ayenew T, Becht R (2008) Comparative assessment of the hydrology NMSA (2005) National meteorological database. National meteoro-
of selected Ethio-Kenyan rift lakes. J Lakes Reservoirs Res Manag logical Service Agency, Addis Ababa
13:191–196 NUPI (1994) Awassa master plan final report. National Urban Planning
Ayenew T, Gebreegziabher Y (2006) Application of a spreadsheet Institute, Addis Ababa
hydrological model for computing the long-term water balance of OEPO (2005) A review of the current status and an outline of a future
lake Awassa, Ethiopia. Hydrol Sci J 51(3):418–431 management plan for Lakes Abijiata and Ziway. Oromia Environ-
Ayenew T, Becht R, Lieshout AV, Gebreegziabher Y, Legesse D, mental Protection Office (OEPO). Unpublished technical report.
Onyando J (2007) Model-based study of the hydrodynamics of Addis Ababa, Ethiopia, 80 pp
topographically closed lakes in the Ethio-Kenyan rift: the case of Street FA (1979) Late quaternary lakes in the Ziway–Shala basin. Ph.D.
lakes Hawassa and Naivasha. J Spat Hydrol 7(1):81–100 thesis, University of Cambridge, Southern Ethiopia, 457 pp
Bekele S (2000) Water resources investigation and design guideline for Telford RJ, Lamb HF (1999) Groundwater-mediated response to
potential exploitation in limited data situation: the case of Abaya- Holocene climate change recorded by the diatom stratigraphy of an
Chamo basin. Ph.D Dissertation, TU Dresden Ethiopian Crater Lake. Quatern Res 52:63–75
Bekele S (2006) Investigation of physical and bathymetric character- Tessema Z (1998) Hydrochemical and water balance approach in the
istics of lakes Abaya and Chamo, Ethiopia, and their management study of high water level rise of Lake Beseka. M.Sc. thesis, The
implications. Lakes Reserv Res Manag 11:133–140 University of Birmingham, 90 pp
Bekele S, Horlacher HB (2001) Modelling the water balance of natural Woldemariam F (2004) Hydrogeology of Lake Abaya and Chamo
lake under limited data situation for impact assessment: the case of basin: with emphasis on the interaction between Lake Abaya and
Abaya-Chamo lakes, Ethiopia. IAHR conference, Beijing, China the surrounding aquifers. M.Sc thesis, Addis Ababa University
Benvenuti M, Carnicelli S, Belluomini G, Dainelli N, Di Grazia S, WWDSE (2001) The study of Lake Hawassa level rise. Southern
Ferrari GA, Iasio C, Sagri M, Ventra D, Atnafu Balemwald, Kebede Nations Nationalities and People Regional State. Water, Mines and
Seifu (2002) The Ziway-Shala lake basin (Main Ethiopian Rift, Energy Resources Development Bureau. Unpublished report of the
Ethiopia): a revision of basin evolution with special reference to the Water Works Design and Supervision Enterprise (WWSDE). Main
late quaternary. J Afr Earth Sci 35:247–269 report, vol II. Addis Ababa, Ethiopia, 291 pp
Benvenuti M, Carnicelli S, Ferrari G, Sagri M (2005) Depositional
processes in latest Pleistocene and Holocene ephemeral streams of
The Geomorphology of the Lake Region
(Main Ethiopian Rift): The Record 17
of Paleohydrological and Paleoclimatic Events
in an Active Volcano-Tectonic Setting
M. Benvenuti and S. Carnicelli
Abstract
This paper illustrates the main geomorphological features of the Lakes Region (Main
Ethiopian Rift) which resulted from the interplay of Late Quaternary climatic and hydrological
changes with volcanism and tectonics typical of an active continental rift. Studies carried out
over several decades demonstrated that the evolution of Late Pleistocene–Holocene fluvio-
lacustrine systems, recorded by a plethora of geomorphic and stratigraphic features, was
forced by abrupt hydro-climatic events of regional to global extent which occurred at 104–102
years scales. Besides the widely acknowledged hydro-climatic forcing, the active rift setting
concurred to regulate, through volcanism and fault activity, erosion/sedimentation rates,
geometry of the lakes basin and of the hydrographic network, and water supply to the lakes.
A volcano-tectonic imprint was left particularly during the transition from the Late Pleistocene
to the Holocene when dramatic hydrological modifications affected the region.

Keywords
Lake evolution Paleogeography Paleohydrology Climate change Rift valley tectonics
17.1 Introduction area (Corti 2009). During the last 100 ky, explosive silicic
volcanism (Abebe et al. 2007) produced surface uplift and
The Lakes Region in the Main Ethiopian Rift (MER) is a collapse and supplied huge amounts of volcaniclastic
tectonically controlled endorheic basin (Street 1979; Chernet sediments to the fluvio-lacustrine systems. Faulting (Abebe
1982) with four major lakes: Ziway, Langano, Abijata, and et al. 2007; Agostini et al. 2011) modified the basin geometry
Shala (Fig. 17.1), whose Late Quaternary morpho-strati- and the hydrographic network (Benvenuti et al. 2002; Sagri
graphic record (Fig. 17.2; Bacci 1940; Nilsson 1940; Grove et al. 2008), affecting rates of erosion/sedimentation. All
and Goudie 1971; Grove et al. 1975; Gèze 1975; Laury and these processes interplayed with climate fluctuations, result-
Albritton 1975; Gasse and Street 1978; Street 1979; Street- ing in complex hydromorphic dynamics that the strictly
Perrot 1982; Gillespie et al. 1983; Le Turdu et al. 1999; climatic hypothesis alone cannot account for. Studies carried
Benvenuti et al. 2002) is a suitable proxy for paleoclimate out in the last 20 years on the geomorphology and stratigra-
reconstructions (Coetzee and van Zinderen Bakker 1989; phy of the Lakes Region (Alessio et al. 1996; Sagri 1998;
Street and Street-Perrot 1990; Gasse and Van Campo 1994; Benvenuti et al. 2002, 2005, 2013; Sagri et al. 2008;
Thomas and Thorp 2003). Moreover, the local geomorphic Carnicelli et al. 2009) have led to a detailed reconstruction of
and sedimentary pattern must have been influenced by the the Late Pleistocene–Holocene evolution of this lacustrine
Cenozoic–Quaternary magmatic and tectonic activity of the system and to the discrimination of the roles of climate and
volcano-tectonics. This paper reviews the main results of
these studies, outlining the geomorphological evolution of
M. Benvenuti (&) S. Carnicelli the area.
Dipartimento di Scienze della Terra, Università di Firenze,
Florence, Italy
e-mail: oredep@unifi.it
290 M. Benvenuti and S. Carnicelli
Fig. 17.1 Location and

physiography of the Lakes
Region in the Main Ethiopian Rift
(inset). Red lines are the major
faults in the region extracted from
http://ethiopianrift.igg.cnr.it/data/
Main%20Ethiopian%20Rift%
20Faults%202.3.1.kmz. WFB
Wonji Fault Belt; SDZFZ
Silti-Debre Zeyt Fault Zone.
Prominent undulations transversal
to the rift trend are shown. MAW
Meki-Awash watershed
Langano, and Abijata lakes (Fig. 17.1) represent the remnants

17.2 General Setting of a larger Late Quaternary endorheic lake basin (Fig. 17.3).
Ziway Lake is fed by the Meki and Ketar rivers, draining the
The Lakes Region is located in the central portion of the Ethiopian and Somali Plateau, respectively, and forming
MER (Fig. 17.1), a 80 km wide and 700 km long NNE–SSW- deltas (Fig. 17.3). Ziway and Langano lakes are connected to
trending depression bounded by steep escarpments and Abijata Lake through the Bulbula and Horakelo rivers,
marked by a stepped morphology, resulting from a dense respectively (Fig. 17.2), whereas the southernmost Lake
swarm of NNE–SSW normal faults (Fig. 17.1). Faults cluster Shala is a separate basin, fed partly by the Dijo River
within two major belts of intense deformation which are the (Fig. 17.1) and partly by groundwater seepage from the other
Wonji Fault Belt and the Silti-Debre Zeyt Fault Zone lakes (Street 1979; Chernet 1982). The study area is char-
(Fig. 17.1). The rift floor ranges in elevation from about acterized by a marked gradient of annual rainfall from about
1,800 m in the north to about 1,580 m at the Abijata and 1,200 mm along the rift margins to about 700 mm around the
Shala lake shorelines, and it is not uniformly flat. Transverse, Ziway–Shala lakes (Sagri 1998).
low-lying undulations are common, a prominent one making
up the divide between the Lakes and Awash river watersheds
(Fig. 17.1). Magmatic activity occurred at different stages 17.3 The Late Quaternary Lakes Region:
during the Quaternary (Mohr 1962; Di Paola 1972; Corti Geomorphology and Stratigraphy
2009), as testified by inactive or hydrothermally active large of Lake-level Fluctuations
volcanoes (Alutu, Bora–Bericha, Ziqwala; Fig. 17.1) which
rise above the rift floor by as much as 1,500 m. Prominent in Following the seminal work of Street (1979), who recon-
the hydrography of the Lakes Region are the Awash and structed the main Late Quaternary fluctuations of the Ziway–
Mojo rivers, originating from the Ethiopian Plateau and Shala lakes from morpho-stratigraphic evidence, further data
feeding the Koka Reservoir while, to the south, Ziway, Shala, have been collected in the last decades, to constrain the main
17 The Geomorphology of the Lake Region … 291
Fig. 17.2 The shore changes of the northern Lake Abijata from annual 1940 to mid-1990, compiled from Sagri (1998), attest to lake
to millennial scales (Fig. 17.1 for location). The overlapping satellite oscillations at decadal scale. The 1,600 m a.s.l. marks the stillstand at
images of 2013 and 2006 (from Google Maps) outline the dramatic about 2,500 years ago when a single lake, including Abijata, Langano,
lake-level variations which occurred in recent years. Shorelines from and Shala lakes, occupied the southern portion of the basin
phases of the Lakes Region evolution (Sagri 1998; Benvenuti (Laury and Albritton 1975), by an escarpment stretching
et al. 2002). These phases, primarily defined by the maximum NE–SW (Fig. 17.4a–b). North of the Gademotta Ridge
extension attained by the lacustrine systems in specific time (Fig. 17.1), it coincides with the rift-parallel Abosa fault
intervals, are: (1) the Late Pleistocene Megalake, (2) the (Fig. 17.1; Laury and Albritton 1975), a scissor-like fault,
Reduced Lakes, (3) the Early–Mid Holocene Macrolake, (4) the scarp of which ranges in height from more than 20 m to
the Late Holocene Separated lakes. Data and evidence of the north to 5–6 m to the south. It is worth mentioning the
such evolution are shortly reported and discussed herein. occurrence of the poorly preserved terrace VI (Laury and
Albritton 1975) attesting to the highest level (above 1,700 m
a.s.l.) reached by the lakes during the Late Quaternary (see
17.3.1 Geomorphology discussion section). Terrace IV and the shoreline-free sur-
face at 1,600 m a.s.l. north of Lake Abijata represent the
Evidence of ancient lake levels is widely represented by relict of two significant Holocene lake stillstands (Grove and
terraces and stranded paleoshorelines, recognized since long Goudie 1971; Street and Gasse 1978; Street 1979). Finally,
(Nillson 1940), though not homogeneously distributed in the bathymetric surveys of Lake Ziway (AERMAP 1969)
area (Fig. 17.3). Relict stranded shorelines are mainly detected a 2 km wide and 15 km long depression, crossing
developed in the central-southern portion of the basin. Par- longitudinally the lake bottom (Fig. 17.3), interpreted as a
ticularly, north of Lake Abijata, degraded shorelines, dis- paleoriver incision (Sagri et al. 2008).
tributed between 1,600 and 1,650 m a.s.l., are separated by a
relatively flat surface, standing at 1,590–1,600 m a.s.l.
(Fig. 17.3), from fresh shorelines (Figs. 17.2 and 17.4c), 17.3.2 Late Quaternary Stratigraphy
grading toward the present lake shores.
Five main lacustrine terraces (terrace V to I, from higher Integrated analysis of erosional and depositional landforms
to lower surfaces, after Laury and Albritton 1975; Fig. 17.3), allowed to establish four unconformity-bounded strati-
which are smooth undulated surfaces separated by major graphic units (ISSC 1994; synthems 1–4; Fig. 17.5), ascri-
escarpments, occur west and north of Ziway Lake. Terrace bed to the Upper Pleistocene–Holocene. A wide collection
IV is separated from terrace V, resting above 1,670 m a.s.l. of radiocarbon dates (Haynes and Haas 1974; Geze 1975;
Fig. 17.3 Schematic

geomorphology and geology of
the Lakes Region: The map
specifically shows the main
evidence of the Late Quaternary
lake-level fluctuations
Street 1979; Gillespie et al. 1983; Le Turdu et al. 1999; • Synthem 2 is dominated by alluvial–colluvial and
Benvenuti et al. 2002; Carnicelli et al. 2009) provided a volcaniclastic deposits accumulated in the basin during the
fairly accurate chronologic calibration of these units, which last full glacial and late glacial, ca. 22,000–11,000 cal y BP
record the major stages in the lake basin evolution (Fig. 17.7b). Despite the predominance of subaerial vol-
(Fig. 17.5; Benvenuti et al. 2002). In general, these units rest canic deposits (Fig. 17.6c, e), this unit is topped by 2–3 m
unconformably over Pliocene–Early Pleistocene volcanic thick cross-bedded sands and silts (Fig. 17.6d), referred to
and volcaniclastic deposits (Abebe et al. 2005). Locally, as a lacustrine nearshore environment dominated by wave
in the area north of Langano Lake and on the slopes of the reworking of volcaniclastic deposits and standing at about
Gademotta Ridge, the Late Pleistocene–Holocene lacustrine 1,720–1,730 m a.s.l. (Laury and Albritton 1975; Benve-
succession overlies reddish-brownish alluvial–colluvial nuti et al. 2002, 2005). Synthem 2 records the Reduced
conglomerates, sandstones, and mudstones pointing to a Lakes phase, when the lakes underwent a dramatic
latest Middle Pleistocene age (Laury and Albritton 1975; shrinking (Gasse and Street 1978; Benvenuti et al. 2002).
Coltorti et al. 2002). The shoreface deposits on top of the synthem suggest that
The main characteristics of these synthems are the this phase ended with a short-lived but massive re-flooding
following: of the former lacustrine basin.
• Synthem 1 consists of colluvial, fluvio-deltaic and • Synthem 3 consists of an articulated alternation of colluvial;
lacustrine gravels, sands, and muds, as well as lacustrine fluvio-deltaic; and lacustrine gravels, sands, muds, lacus-
diatomites and volcaniclastic materials (Fig. 17.6a, b), trine diatomites, and volcaniclastic materials (Fig. 17.6a, f)
deposited during the last glacial period, ca. 100,000– that were deposited during the Early to Middle Holocene,
22,000 y BP (Fig. 17.5). This interval corresponds to the ca. 11,000–5,000 cal y BP (Benvenuti et al. 2002; Fig. 17.5).
Megalake phase, when lake waters inundated wide areas It records the Macrolake phase, during which the lake was
north of Lake Ziway rising over 1,800 m a.s.l. smaller than recorded by Synthem 1 and characterized by
Fig. 17.4 Morphological

features related to former lake
levels. a Oblique aerial view of
the northern coast of Lake Ziway
(from Bing Maps). b Annotated
picture in a showing some of the
lacustrine terraces and other
features described in the text.
TV–I: terraces after Laury and
Albritton (1975); MAW
Meki-Awash watershed.
c Ground view looking to the
west of a well-preserved
paleoshoreline (younger than
2,500 year BP) on the southern
coast of Lake Abijata. Car in the
circle for scale
frequent fluctuations (Benvenuti et al. 2002, 2005). The last 5,000 years. It records the period when the previous
maximum elevation of the lake surface during this phase single large lake separated into Ziway Lake and a
was about 1,670 m a.s.l. (Street 1979; Laury and Albritton southern lake, itself progressively splitting into the
1975; Benvenuti et al. 2002) (Fig. 17.3). present-day Abyiata, Langano, and Shala lakes (Benve-
• Synthem 4 is made of colluvial, fluvial, deltaic, and nuti et al. 2002; Fig. 17.5).
lacustrine sediments (Fig. 17.6g) accumulated over the
Fig. 17.5 Simplified geological

map of the Lakes Region (after
Benvenuti et al. 2002)
17.4 The Geomorphic and Stratigraphic allowed the formal definition of the T’ora Geosol, a pedo-
Record as a Proxy for the Late stratigraphic unit (Morrison 1977; North American Com-
Quaternary Hydro-climatic mission On Stratigraphic Nomenclature 1983) of regional
Variability significance. The Geosol components, from oldest to youn-
gest referred to as T’ora I to T’ora III, derive from a pro-
17.4.1 The Late Pleistocene–Holocene gressive conversion of volcanic glass into smectite clay
Ziway–Shala Lakes Fluctuations along an evolutionary pathway from soils dominated by
volcanic glass (Vitrands), through a Luvisol stage to nearly
Seven major highstands of the latest Pleistocene–Holocene Vertisol characters. A second relevant soil (Wer’ja paleosol;
Ziway–Shala lakes have been identified in previous studies Carnicelli et al. 2009) is found on Synthem 3 slope deposits
(Street 1979; Gillespie et al. 1983; Fig. 17.7a). Ziway–Shala above 1,670 m a.s.l. (Fig. 17.9e). This soil, locally buried by
(Zw-Sh) I was referred at before 50,000 y BP, with a a Vitrand on more recent, reworked tephra, is very similar to
maximum age likely close to 100,000 y BP (Laury and T’ora II with morphological differences mostly due to dif-
Albritton 1975; Street 1979; Bigazzi et al. 1993; Le Turdu ferent drainage conditions (Carnicelli et al. 2009). The
et al. 1999). Zw-Sh II and Zw-Sh III record highstands genesis of these paleosols indicates a marked seasonality,
predating the Last Glacial Maximum (LGM; Fig. 17.7b), with alternating, pronounced dry and humid seasons. Such a
whereas stages Zw-Sh IV–VII mark the major Holocene climate regime was likely generally moister than present.
highstands. The Early–Middle Holocene Zw-Sh IV–VI Radiocarbon dating (Carnicelli et al. 2009; Fig. 17.8a) and
highstands are considered to have reached an elevation of stratigraphic relations constrained the development of T’ora
the water surface close to the inferred basin threshold at I in a maximum time span between 19,000 and 8,300 cal y
1,670 m a.s.l., with possible overflow to the north (Street BP. T’ora II and Wer’ja soils turned out to be essentially
1979; Gillespie et al. 1983). The late Holocene Zw-Sh VII coeval, developing between 7,000 and 5,500 and between
stands at about 1,600 m a.s.l., as indicated by the morpho- 7,300 and 6,100 cal y BP, respectively. Finally, T’ora III
logical lines; after 5,000 y BP, the lakes never reached the existed as a surface soil since 5,400 cal y BP.
Early–Middle Holocene elevation and progressively reduced
their surfaces (Separated Lakes phase). Successive studies
(Alessio et al. 1996; Benvenuti et al. 2002; Carnicelli et al.
2009) confirmed the classic Street’s pattern of Holocene lake 17.4.3 Late Holocene Cut and Fill Cycles
fluctuations (Fig. 17.8a). in Ephemeral Fluvial Systems
Disconnected from Lake Levels
17.4.2 Soil Development Between The slopes between 1,720 and 1,670 m a.s.l., such as on the
the Latest Pleistocene Gademotta Ridge southwest of Lake Ziway (Figs. 17.1 and
and the Middle Holocene 17.5), consist of pre-Late Quaternary volcanic rocks and are
mantled by synthem 2–4 deposits, also resting unconformably
The hydro-climatic transitions left clear traces on the geo- on the lacustrine sediments of synthem 1 (Benvenuti et al.
morphic systems above the lakes and hence less influenced 2002, 2005). Slope and alluvial deposits within synthems 2–4
by the lacustrine base level change. An outstanding evidence include a succession of filled box-valleys (Fig. 17.9e) char-
of that is found in a composite paleosol (Morrison 1977), acterized by morphologic and depositional features similar to
developed over different volcanic and sedimentary deposits, those of modern ephemeral streams/gullies, disconnected
either included in synthem 1 and 2 or older (Carnicelli et al. from the lakes (Benvenuti et al. 2005). Independently from
2009), whereas it was never observed over synthems 3–4 their stratigraphic position, the valley fills are made up of
sediments. This soil (Fig. 17.9a) covers a wide range of stream flow, sheet flow, and mass flow deposits (Fig. 17.9e),
landforms and landscapes, always above 1,670 m a.s.l., i.e., stacked into a recurring vertical pattern. Stream flow sands and
the maximum lake level attained during the Macrolake gravels generally occur at the base of the valley fill, sharply
phase. It is made of up to three soils; the lower soil profile is overlain by sheet flow silty sands, including poorly-developed
characterized by the presence of a thick Petrocalcic horizon paleosols (Carnicelli et al. 2009; Fig. 17.9e), which form the
(Fig. 17.9b) overlain by clay accumulation horizons. One or bulk of valley fills. Mass flow deposits, related to different
two pedogenized tephra resting above these horizons bank failure mechanisms, are interbedded with alluvial
(Fig. 17.9a) typically represent the upper soil profile. In deposits, showing recurrent associations, respectively,
places, the paleosol is buried (Fig. 17.9c, d) by younger between stream flow and debris fall and sheet flow and debris
sediments; radiocarbon dating (Carnicelli et al. 2009) flow deposits (Benvenuti et al. 2005).
Fig. 17.6 The Upper Quaternary

deposits of the Lakes Region.
a Unconformable contact
between lacustrine diatomite and
volcanics of synthems 1 and
3 visible in the Bulbula River
valley, person for scale. b Detail
of wave-rippled deltaic sandstone
of synthem 1. c Pumice fall
deposits of synthem 2 in a quarry
north of Lake Ziway. d The
uppermost deposits of synthem
2 consist of cross-laminated
pebbly sandstone ascribed to delta
shoreface environments pointing
to a short-lived very high lake
level, a quarry north of Lake
Ziway. e Pumice fall deposits of
synthem 2 mantling the slopes of
the Bora–Bericha volcanic
complex (Fig. 17.1). f Typical
shell bed in the lacustrine deposits
of synthem 3 on the east coast of
Lake Shala. g Detail of synthem
4 on the east coast of Lake Shala:
Foreset beds are sharply overlain
by the bottomset of a small
backstepping delta during a rise
of the Separated Lakes
associated with tectonic structures are considered for their

17.5 Volcano-tectonic Features Related implication on the evolution of the lakes. In particular,
to the Late Quaternary landforms determining the western and northern limits of the
Development of the Lakes Holocene lakes are described. A significant portion of the
escarpment bounding the Holocene lacustrine deposits and
17.5.1 Tectonic Structures Controlling landforms (i.e., Terraces V–IV escarpment; Fig. 17.4a–b),
the Macrolake Geometry west of Lake Ziway, is associated to the Abosa fault (West
Ziway Fault in Le Turdu et al. (1999); Figs. 17.1, 17.3 and
In addition to the large evidence of erosional and deposi- 17.5). The northern boundary of the Holocene lakes is rep-
tional landforms directly related to lacustrine, alluvial, and resented by a WNW–ESE-trending rift floor undulation
slope processes, some specific morphologic features (Fig. 17.4a–b). This and other similar undulations (Fig. 17.1)
17.5.2 Explosive Silicic Volcanism During

the Reduced Lakes Phase
Synthem 2 consists mostly of volcaniclastic fall and flow

deposits that cover a large portion of the Lakes Region
(Fig. 17.6c–e). Their reworking by surface processes and the
occurrence of weakly developed paleosols (Street 1979;
Benvenuti et al. 2002, 2005) suggest subaerial deposition
during the Reduced Lakes phase. Marked grain size and
thickness gradients hint to sourcing from the main silicic
volcanoes and calderas, such as the Bora–Bericcio complex
and the Alutu volcano (Fig. 17.1), which produced large
volumes of rhyolitic and obsidian lava flows, unwelded
pumice flows, pumice falls, and ashes starting from Late
Middle Pleistocene (about 0.25 Ma, Di Paola 1972; Mohr
et al. 1980; Woldegabriel et al. 1990; Le Turdu et al. 1999;
Abebe et al. 2007). During the latest Pleistocene, huge
volcanic eruptions delivered large amounts of sediment that
mantled the desiccated lakes bottom and congested the
drainage network.
17.6 Evidence of the Evolution

of the Lakes: A Discussion
17.6.1 Geomorphological Evidence
The chronologic calibration of the lacustrine deposits asso-

ciated with the various surfaces described above coarsely
constrains the development of terraces V and VI to the Late
Pleistocene Zw-Sh I–III, corresponding to the Megalake
phase. Though not dated, the flooded palaeovalley on Lake
Ziway bottom hints to fluvial incision of an almost com-
pletely desiccated lake, an event that may be reasonably
related to the latest Pleistocene Reduced Lakes phase. A more
accurate date calibration of the lacustrine surfaces is provided
Fig. 17.7 Synopsis of the chronology of the Ziway–Shala lake for the Holocene Macrolake and Separated lakes phase
fluctuations. a Classic Street’s diagram of Ziway–Shala lacustrine
phases: All the radiocarbon dated geomorphic and stratigraphic lines of
(Fig. 17.7a). The maximum Holocene lake levels, marked by
evidence of lake level are plotted against the elevation. SD Shala the 1,670 m a.s.l. escarpment that roughly corresponds to the
Datum, Lake Shala elevation in the early 1970 (after Gillespie et al. lake overflow threshold (Laury and Albritton 1975; Street
1983). b The Zw-Sh I–VII lacustrine phases are compared with the 1979), were reached between 9,600 and 9,400 14C y BP
phases established in Benvenuti et al. (2002) and plotted on the Late
Quaternary MIS chronology (after Porter 1989)
(Zw-Sh IV) and about 5,400–5,100 14C y BP (Zw-Sh V)
(Haynes and Haas 1974; Laury and Albritton 1975). The
development of terrace IV occurred during these lake high-
reflect buried tectonic structures, transversal to the main rift stands (Fig. 17.10a). The differently preserved, stranded
fault systems, which outline an orthogonal system of trans- palaeoshorelines, visible north of Abijata Lake, record epi-
verse horsts and high-amplitude folds consistent with the sodic lake-level stillstands, modulated in a regressive trend of
extensional model proposed for MER evolution by a few lowering of a southern palaeolake (Fig. 17.10c). Since the
authors (e.g., Rosendhal et al. 1986; Boccaletti et al. 1992; escarpment separating terrace IV from terrace III tends to
Bonini et al. 1997; Corti 2009). become convex south of Ziway Lake and the minimum age
Fig. 17.8 a The calibrated

chronology of the Lake Shala
fluctuations (right axis, red curve)
and the salinity variation recorded
by diatom assemblages of Lake
Abijata sediments (left axis, black
curve, after Chalie and Gasse
2002) as proxies for major
hydrologic/climatic events during
the latest Pleistocene–Holocene
(after Carnicelli et al. 2009).
White rectangles indicate the
T’ora and Werja soil development
during the Megalake phase. Black
and white stars on the salinity
curve of Lake Abijata refer to
calibrated ages, respectively,
of streamflow and sheetflow
facies in the box-valley fills
formed during the Separated
Lakes phase on the lower slopes
of the Gademotta Ridge (see
Carnicelli et al. 2009 for details).
b Curve of temperature of
Greenland derived from GISP
data (after Alley 2000) as a
reference for the Lakes Region
events. MWP Medieval Warm
Period; LIA Little Ice Age
of terrace III ranges from 5,600 to 5,300 14C y BP (Haynes 17.6.2 Soils and paleosols
and Haas 1974; Laury and Albritton 1975), a palaeo-Ziway
lake may have started to separate from a larger southern lake The chronologic calibration of paleosols allowed to distin-
around the Middle Holocene. The palaeoshorelines facing guish distinct stages of soil formation in the stratigraphic and
Abijata Lake therefore resulted from progressive shrinking of palaeohydrologic framework. T’ora I paleosol apparently
the southern lake during the Late Holocene. The shoreline- encompasses the Reduced Lakes and Early Macrolake pha-
free surface at 1,600 m a.s.l. north of Lake Abijata is asso- ses. Since its development requires a relatively moist seasonal
ciated with the coastline of a southern lake, encompassing the climate, it is unlike that T’ora I started forming during the
modern Langano, Abijata, and Shala lakes, that still existed Reduced Lakes phase, when arid conditions dominated in the
around 2,000 14C y BP (Gillespie et al. 1983; Alessio et al. region. Also, the short-lived moist episode leading to the huge
1996) during the later Separated Lakes phase (Fig. 17.10c). lake stillstand at about 1,720 m a.s.l. during the later Reduced
Fig. 17.9 Paleosols and non-lacustrine deposits of the Lakes Region. slopes of the Gademotta Ridge. Two box-valley fills characterize
a Panoramic view on the badlands developing on the T’ora Geosol synthem 4, subdivided in different sub-units recording successive
north of Lake Ziway: in the foreground the T’ora I–II. Person for scale. stages of valley fill. The valley fills in each sub-unit consist of
b Deeply eroded T’ora Geosol exposing the petrocalcic horizon at the streamflow facies overlain by sheetflow facies, both flanked by slope
base of T’ora I. c The T’ora Geosol unconformably overlain by facies, on the whole outlining fining-upward trends. Weakly developed
synthem 4 alluvial deposits at the NE footslopes of the Gademotta paleosols (S2–S3) mark periods of stability of the valley bottom
Ridge (after Carnicelli et al. 2009). d The almost complete T’ora suitable for soil development. Persons for scale (after Benvenuti et al.
Geosol buried by synthem 4 in the same area of c; meter tape for scale. 2005; Carnicelli et al.2009)
e A section in sediments of synthems 1, 3, and 4 exposed on the ENE
Lakes phase preceded the T’ora Geosol development, as the transport capacity decreased, the bank failure deposits were
surfaces where this soil is found today were then flooded. redistributed by sheet and debris flows, leading to aggrada-
Most likely, the interval of T’ora I genesis was then the ear- tion, interbedded with paleosols, and eventual overfilling of
liest Holocene moist stages which determined the Zw-Sh IV– the valleys (Benvenuti et al. 2005). These box-valley fills are
V highstands at the onset of the Macrolake phase. T’ora II and similar to those of discontinuous ephemeral streams of the
Wer’ja paleosols fully developed during the Macrolake sub-arid SW USA (Graf 1988; Bull 1997). The Late Qua-
phase, synchronously with the Zw-Sh VI highstand at around ternary dynamics of these ephemeral streams/gullies has
1,670 meters a.s.l. (Gasse and Street 1978; Street 1979). been accounted for by an autocyclic mechanism (Schumm
Finally, T’ora III started developing at the end of the Mac- 1973) of valley backfilling and aggradation, resulting in a
rolake phase, when a seasonal climate, moister than present, typical fining-upward sedimentary succession (Packard
may have still existed in the region. 1974; Graf 1988; Bull 1997). An alternative interpretation of
the incision/aggradation disequilibrium of this kind of
streams invokes external controls, such as climatically-dri-
17.6.3 Late Holocene Cut and Fill Cycles ven hydrologic changes (Bull 1997; Waters and Haynes
2001). This latter hypothesis was positively tested by
The cyclic stacking pattern of the alluvial and slope deposits detailed radiocarbon dating of valley fill sediments and
mantling the Gademotta Ridge suggests that the early valley buried soils included in synthem 4 (Carnicelli et al. 2009;
fill phases were dominated by coarse-grained material, Fig. 17.8a). A proxy palaeohydrological record for the
transported as bedload by competent and likely permanent Separated Lakes phase is provided by the Late Holocene
water flows. Undermining of the nearly vertical banks variation of salinity of Lake Abijata, modeled from diatom
commonly determined toppling and debris fall accumulation assemblages found in the lacustrine sediment core (Chaliè
on valley bottoms. In later stages, as water discharge and and Gasse 2002). High–low salinity values approximate
Fig. 17.10 Holocene

palaeogeography of the
Ziway–Shala basin. a The
maximum Macrolake extension
reached during Zw-Sh IV–VI
highstands. b Early separation of
Lake Ziway from the southern
lake at the beginning of the
Separated Lakes phase. c The
lakes during the 2,500 years BP
highstand when a southern lake,
though reduced, still existed.
TV–III: lacustrine terraces after
Laury and Albritton (1975);
Dps degraded paleoshorelines
low–high lake levels in response to decreasing or increasing shrinking lakes may have undergone downcutting, while the
rainfall. The chronology of synthem 4 valley fills shows that others were experiencing valley backfill and aggradation
valley aggradation coincides with intervals of high lake related to reduced water discharge.
salinity (Fig. 17.8a). By contrast, stages of box-valley inci-
sion, though not directly dated, are constrained to occur
during phases of low lake salinity and higher river water 17.6.4 Climatic Forcing of the Late
supply and, hence, of larger precipitation (Carnicelli et al. Quaternary Lake-level Fluctuations
2009; Fig. 17.8a). Similarly to the modern gullies, their Late
Holocene counterparts graded at a base level disconnected The Ziway–Shala lake oscillations have been related to the
from the lakes, which explains the out of phase relation Late Quaternary weakening and strengthening of the mon-
between incision/aggradation and high/low lake status, soonal circulation (Kutzbach and Street Perrott 1985; Coet-
compared with streams which graded at the lake levels zee and van Zinderen Bakker 1989; Street Perrott and Perrott
(Fig. 17.11; Benvenuti et al. 2006). During moister periods, 1990; Gasse and Van Campo 1994) that caused alternating
the rivers grading at rising lake levels may have experienced drier (off-mode of monsoons) and moister (on-mode of
aggradation (Fig. 17.12a), while those disconnected from the monsoons) conditions over the Lakes Region. In this per-
lakes may have seen valley incision and dominant flushing spective, the Milankovitch orbital forcing was considered as
of transported sediments under high discharges the primary control on the long-term hydro-climatic
(Fig. 17.12b). During dryer periods, the rivers connected to dynamics of the Lakes Region and other regions of northern
Fig. 17.11 Example of fluvio-

deltaic response to
high-frequency fluctuations of
Lake Shala. a The southern shore
of Lake Shala with the stranded
shorelines formed during the
Separated Lakes Phase.
b A section on lacustrine and
fluvio-deltaic deposits exposed on
a meander bend of a short stream
(red star in A for location), person
for scale, the lake is in the
background. c Interpreted section
in B): L lacustrine mudstone;
CF fluvial channel fill sandstone:
early (yellow), late (blue);
DF delta front sandstone;
gray: talus scree (see text)
and eastern Africa (Fontes and Gasse 1989, 1991; Hillaire- Following this short-lived lacustrine episode, the Lakes
Marcel et al. 1986; Gasse 1989; Taieb et al. 1991). The Late Region was again affected by arid conditions and lake
Pleistocene Zw-Sh I–III stages (Street 1979; Megalake reduction as indicated by the topmost slope and alluvial
phase) occurred during warm phases (MIS5c–a, MIS3; deposits of synthem 2 (Benvenuti et al. 2005; Sagri et al.
Fig. 17.7b) punctuating the last glacial period. The missing 2008). This stage, predating the Macrolake phase, is con-
geomorphic, stratigraphic, and chronologic evidence of lake sidered to record the Younger Dryas cold event (Fig. 17.8b)
development between Zw-Sh III and Zw-Sh IV (Reduced that determined again arid condition all over the African
Lakes phase) points to a dramatic lacustrine shrinkage cen- tropics (Roberts et al. 1993). The return to lacustrine condi-
tered around 20,000–18,000 y BP and coinciding with the tions during the Early–Middle Holocene Zw-Sh IV–VI
LGM (MIS2; Fig. 17.6b), when aridity dominated in the (Macrolake phase; Figs. 17.7a and 17.8a) corresponds to the
northern tropics (Coetzee and van Zinderen Bakker 1989). post-glacial climatic optimum (MIS 1). Besides the different
The stratigraphic evidence of lake level standing at about interpretation of the forcing mechanisms, it is now widely
1,720 m a.s.l. at the end of the Reduced Lakes phase (Ben- acknowledged that most of the latest Pleistocene–Holocene
venuti et al. 2002; Sagri et al. 2008) attests to a huge re- hydro-climatic events recorded in the Lakes Region have a
flooding of the lakes basin. Despite the lack of direct dating, regional to global significance (Mayewski et al. 2004;
this high lake level may be reasonably constrained to the Weijers et al. 2007). Marked lake lowering during the Mac-
post-LGM latest Pleistocene (Benvenuti et al. 2002) when rolake and Separated Lakes phase (Figs. 17.7a and 17.8a)
warm–cold fluctuations globally characterized the transition occurred around global cooling events, such as the 8,200 cal
to the Holocene. A warm stage between 14,700 and y BP event (Fig. 17.8b), resulting in arid conditions all over
12,700 cal y BP (Bølling-Allerød event; Fig. 17.8b) had the tropics (Mayewski et al. 2004). The relationship between
significant hydrological consequences in the African tropics high lake levels and global warm–moist periods seems not to
(Weijers et al. 2007), such as the massive refilling of an be satisfied in the last 2,000 years as palaeolimnological data
almost desiccated Lake Magadi in Kenya (Taieb et al. 1991). from the Sahel and East Africa (Maley 1976; Verschuren
Fig. 17.12 Models of fluvial incision/aggradation in response to Holocene discontinuous ephemeral streams draining the lower slopes
moisture fluctuation regulating river discharge and lake levels (after of the Gademotta Ridge; after Benvenuti et al. 2005). Codes:
Benvenuti et al. 2006). a River connected to lake level (inspired by CF channel fill; DF delta front; vbf valley bank mass wasting; csf-fsf
Fig. 17.11b–c). b River disconnected from lake level (inspired by Late coarse–fine-grained sheetflow deposits; df debris flow deposits
Fig. 17.13 Palaeohydrography north of Lake Ziway. a The regional high just north the Lake Ziway (presently the Meki-Awash watershed,
drainage during the Late Reduced Lakes phase (latest Pleistocene) MAW) determined a smaller Macrolake compared to the Megalake
toward the desiccated lake basin, a condition that dominated also maximum extension. These structures, joined to further transverse lines
during the Megalake phase and ensured high water supply to the lakes. northward (dotted red line) and SE-ward tilting possibly due to the
The Abosa fault may have been active during these stages making the activity of the WFB (thick black arrow), determined the deviation of
central portion of the basin actively subsiding. WFB Wonji Fault Belt; the Awash and Mojo rivers toward ENE, subtracting a significant
SDZFZ Silti-Debre Zeyt Fault Zone. Present Lake Ziway and Koka volume of water from the hydrological budget of the Early Holocene
Reservoir for reference. b Around the Pleistocene–Holocene transition, Lakes Region (modified from Sagri et al. 2008)
the possibly concurrent activity of the Abosa fault and of the transverse
et al. 2000; Chalié and Gasse 2002; Legesse et al. 2002) apparent paradox may be accounted for by considering the
equate colder to moister periods, such as the Little Ice Age, activity of an orthogonal system of tectonic structures during
and warmer to drier periods such as the Medieval Warm the latest Pleistocene, a period characterized also by intense
Period (Fig. 17.8b). explosive volcanism. During the Megalake and Reduced
Lakes phase, the Abosa fault (Fig. 17.13) may have deter-
mined subsidence on its hanging wall, creating future
17.6.5 Tectonic Control on the Basin accommodation space for the Macrolake. The geometry of
Geometry and Regional the Early Holocene lake basin was also conditioned by the
Hydrography concurrent activity of the northern transverse high. Never-
theless, the concurrent action of rift-parallel and transverse
Though the global-scale Late Quaternary climate changes structures not only created a depocentre for the Early
provide a reasonably convincing explanation for the pala- Holocene lakes but also forced cutting off of such large
eohydrologic evolution of the Ziway–Shala Basin, such a fluvial systems as the Awash and Mojo rivers, which fed the
primary control may be, at least partially, questioned when Late Pleistocene lakes from the north (Fig. 17.13a; Sagri
the Megalake and Macrolake extensions are compared. The et al. 2008). Yet, during the Reduced Lakes phase, the fault-
global increase in temperature and moisture during the controlled northeastward tilting of the rift floor progressively
LGM–Holocene transition was greater in magnitude than deviated the Awash River which also captured the Mojo
that experienced during any pre-LGM stadial–interstadial River (Fig. 17.13b; Sagri et al. 2008). At the onset of the
transitions (Fig. 17.7b; Dansgaard et al. 1993). In the tropics, Holocene climatic optimum, the lakes were fed exclusively
such conditions should have resulted in water budgets larger by the Meki and Ketar rivers (Fig. 17.13b) with a significant
at the onset of Holocene than at the start of any previous reduction of water supply due to the latest Pleistocene tec-
interstadial (Thomas and Thorp 1995). In spite of that, the tonically controlled reorganization of the river network. The
maximum extension of the Holocene Macrolake, regulated intense fault activity of the Reduced Lakes arid phase was
by the overflow threshold at 1,670 m a.s.l., was markedly also coupled with intense volcanic activity, as observed in
smaller than that of the Late Pleistocene Megalake. Such an other rifted basins of eastern Africa (Abebe et al. 2007).
17.7 Conclusions Ethiopian Rift and its implication for the volcano-tectonic evolution
of the rift. Geol Soc Am Map Chart Ser, MCH094
Abebe B, Acocella V, Korme T, Ayalew D (2007) Quaternary faulting
In the variety of the geomorphological landscapes of Ethio- and volcanism in the Main Ethiopian Rift. J Afr Earth Sci 48:115–124
pia, the Lakes Region resulted from the interaction of tecto- AERMAP (1969) Meki river diversion scheme: topographic surveys,
nism, volcanism, and climate, which had different impacts on Imperial Ethiopian Government, Awash Valley Authority, Addis Ababa
the surface processes at different timescales. The tectonic and Agostini A, Bonini M, Corti G, Sani F, Manetti P (2011) Distribution
of quaternary deformation in the central Main Ethiopian Rift, East
volcanic activity, operating since the Cenozoic in a portion of Africa. Tectonics 30(4) TC4010
a wide continental rift system, shaped the region as a fault- Alessio M, Allegri L, Belluomini G, Benvenuti M, Cerasoli M, Improta
controlled subsiding depression, mostly filled up with large S, Manfra L, Sagri M, Ventra D (1996) Le oscillazioni tardo-
volumes of volcanic products, in which endorheic lacustrine quaternarie del Lago Shala (Rift Etiopico): Analisi dell’evoluzione
ambientale dall’integrazione di evidenze morfologiche, sedimenta-
basins developed. They characterize the region at least since rie e cronologiche. Il Quaternario 9:387–392
the Late Middle Pleistocene with relatively shallow, tecton- Alley RB (2000) The Younger Dryas cold interval as viewed from
ically controlled water bodies, particularly sensitive to hydro- central Greenland. Quat Sci Rev 19:213–226
climatic variations. During the Late Pleistocene–Holocene Bacci G (1940) Molluschi fossili dell’antico fondo del Lago Ziway.
Ann Mus St Nat Genova 60:454–458
interval, the lake hydrology was primarily controlled by the Benvenuti M, Carnicelli S, Belluomini G, Dainelli N, Di Grazia S,
global climate, oscillating at different frequencies between Ferrari GA, Iasio C, Sagri M, Ventra D, Atnafu Balemwald, Kebede
warmer and colder conditions which, in the tropical belt, Seifu (2002) The Ziway-Shala lake basin (Main Ethiopian Rift,
resulted in alternating moist and arid phases. For this reason, Ethiopia): a revision of basin evolution with special reference to the
Late Quaternary. J Afr Earth Sci 35:247–269
the Lakes Region represents a classic site for exploring the Benvenuti M, Carnicelli S, Ferrari G, Sagri M (2005) Depositional
geomorphic and stratigraphic record of these palaeohydro- processes in latest Pleistocene and Holocene ephemeral streams of
logic and palaeoclimatic fluctuations. The emphasis on a the Main Ethiopian Rift (Ethiopia). In: Blum MD, Marriott SB,
predominant climatic control over the Late Quaternary Leclair SF (eds) Fluvial Sedimentology VII, Spec Publ int Ass
Sediment. vol 35, pp 277–294
development could minimize the role of the active regional Benvenuti M, Carnicelli S, Ferrari G, Sagri M (2006) Coupled evidence
volcano-tectonic setting that, on the contrary, was particu- of Late Holocene rapid moisture fluctuations in lake and river
larly significant during the Pleistocene–Holocene transition. system proxies, Main Ethiopian Rift. Holivar 2006 Open Science
In the 18,000–11,000 y BP interval, concurrent fault activity Meeting, Poster Presentation, University College London
Benvenuti M, Bonini M, Tassi F, Corti G, Sani F, Agostini A, Manetti
and explosive volcanism modified the basin geometry, sup- P, Vaselli O (2013) Holocene lacustrine fluctuations and deep CO2
plied high amounts of volcaniclastic sediments, and forced degassing in the northeastern Lake Langano basin (Main Ethiopian
reorganization of hydrography. This critical period, therefore, Rift). J Afr Earth Sci 77:1–10
was characterized not only by arid conditions established in Bigazzi B, Bonadonna FP, Di Paola GM, Giuliani A (1993) K-Ar and
fission tracks ages of the last volcano-tectonic phase in the
the LGM, which dramatically reduced the lake extension, but Ethiopian Rift Valley (Tulu Moye area). In: Geology and Mineral
also by tectono-magmatic activity. The typical processes of Resources of Somalia and Surrounding Regions. Ist Agronom
an evolving continental rift were evidently able to leave a Oltremare, Firenze, Relaz Monograf. vol 113, pp 311–322
clear imprint over the geomorphic evolution, keeping pace Boccaletti M, Getaneh A, Tortorici L (1992) The Main Ethiopian Rift:
an example of oblique rifting. Annales Tectonicae 6:20–25
with high-frequency climate transitions. Bonini M, Souriot T, Boccaletti M, Brun JP (1997) Successive orthogonal
and oblique extension episodes in a rift zone: laboratory experiments
Acknowledgments The authors are indebted with Prof. Mario Sagri
with application to the Ethiopian Rift. Tectonics 16:347–362
and Giovanni Ferrari for their effort and impulse in the field researches
Bull WB (1997) Discontinuous ephemeral streams. Geomorphology
in Ethiopia and the constructive contribution in the discussion of the
19:227–276
data, for understanding the geomorphological evolution of the Lakes
Carnicelli S, Benvenuti M, Ferrari G, Sagri M (2009) Dynamics and
Region during the Late Quaternary. The researches summarized in this
driving factors of late Holocene gullying in the Main Ethiopian Rift
paper involved between 1994 and 2005 also a large number of Italian
(MER). Geomorphology 103:541–554
undergraduate and graduate students and Ethiopian colleagues which
Chalié F, Gasse F (2002) Late Glacial-Holocene diatom record of water
benefitted from funds from the European Commission (EC-STD3
chemistry and lake level change from the tropical East African Rift
“Land resource inventory, environmental changes analysis, and their
Lake Abijata (Ethiopia). Palaeogeogr Palaeoclimatol Palaeoecol
application to agriculture in the Lakes Region, Ethiopia”) and the
187:259–283
Italian Ministry of University and Research (PRIN 1999 and 2003).
Chernet T (1982) Hydrogeology of the Lakes Region, Ethiopia.
Prof. Paolo Billi and Mario Sagri are thanked for their critical reading
Ministry of Mines and Energy, Addis Ababa
and review of an early version of this paper.
Coetzee JA, van Zinderen Bakker EM (1989) Palaeoclimatology of
East Africa during the last glacial maximum: a review of changing
theories. In: Quaternary and environmental research on East African
References mountains. Balkema, pp 189–198
Coltorti M, Corbo L, Sacchi G (2002) New evidence for the Late
Pleistocene and Holocene climatic changes in the Lake Region. In:
Abebe T, Manetti P, Bonini M, Corti G, Innocenti F, Mazzarini F, Dramis F (ed) Proceedings of the symposium international
(2005) Geological map (scale 1: 200,000) of the northern Main Geomorph Ass, Addis Ababa
Corti G (2009) Continental rift evolution: from rift initiation to Mohr PA (1962) The Ethiopian Rift System. Bull Geophys Obs Addis
incipient break-up in the Main Ethiopian Rift, East Africa. Earth-Sci Ababa 5:33–62
Rev 96:1–53 Mohr PA, Mitchell JC, Raynolds RGH (1980) Quaternary volcanism
Dansgaard W, Johnsen SJ, Clausen HB, Dahl-Jensen D, Gundestrup and faulting at O’A Caldera, central Ethiopian Rift. Bull Volcan
NS, Hammer CU, Hvidberg CS, Steffensen JP, Sveinbjörnsdottir 43:173–189
AE, Jouzel J, Bond G (1993) Evidence for general instability of past Morrison RB (1977) Quaternary soil stratigraphy, concepts, methods
climate from a 250-kyr ice-core record. Nature 364:218–220 and problems. In: Mahaney WC (ed) Quaternary soils. Geo-
Di Paola GM (1972) The Ethiopian Rift Valley (between 7°00’ and Abstracts, Norwich, pp 77–108
8°40’ lat. North). Bull Volcan 36:517–560 Nilsson E (1940) Ancient changes of climate in British East Africa and
Fontes JC, Gasse F (1989) On the ages of humid Holocene and Late Abyssinia; a study on ancient lakes and glaciers. Stockholm Geogr
Pleistocene phases in North Africa—remarks on “Late Quaternary Ann 22:1–79
climatic reconstruction for the Maghreb (North Africa)” by North American Commission on Stratigraphic Nomenclature (1983)
P. Rognon. Palaeogeogr Palaeoclimatol Palaeoecol 70:393–398 North American stratigraphic code (NASC). AAPG Bull 67:841–875
Fontes, J-C, Gasse F (1991) PAHLYDAF (Palaeohydrology in Africa: Packard FA (1974) The hydraulic geometry of a discontinuous
objectives, methods and major results). Palaeogeogr Palaeoclimatol ephemeral stream on a Bajada near Tucson, Arizona. Unpublished
Palaeoecol 84:191–215 Ph.D. Thesis, Univ Arizona
Gasse F (1989) Tectonic and climatic controls on lake distribution and Porter SC (1989) Some geological implications of average quaternary
environments in Afar from Miocene to present. AAPG Mem 50:19–41 glacial conditions. Quat Res 32:245–261
Gasse F, Street FA (1978) Late Quaternary lake-level fluctuations and Roberts N, Taieb M, Barker P, Damnati B, Icole M, Williamson D
environments of the northern Rift Valley and Afar region (Ethiopia (1993) Timing of the Younger Dryas event in East Africa from lake-
and Djibuti). Palaeogeogr Palaeoclimatol Palaeoecol 24:279–325 level changes. Nature 366:146–148
Gasse F, Van Campo E (1994) Abrupt post-glacial climate events in Rosendahl BR, Reynolds DJ, Lober PM, Burgess CF, McGill J, Scott
West Asia and North Africa monsoon domains. Earth Planet Sci D, Lambiase JJ, Derksen SJ (1986) Structural expression of rifting:
Lett 126:435–456 lesson from Lake Tanganika, Africa. Geol Soc Spec Publ 25:29–43
Geze F (1975) New dates on ancient Galla Lake levels. Bull Geophys Sedimentation in the African Rifts
Obs Addis Ababa 15:119–124 Sagri M (ed) and the staff of EU Project (1998) Land resources
Gillespie R, Street-Perrot FA, Switsur R (1983) Post-glacial arid inventory, environmental change analysis and their application to
episodes in Ethiopia have implications for climate prediction. agriculture in the Lakes Region (Ethiopia) Final report. European
Nature 306:680–683 Commission, DG XII, Bruxelles
Graf WL (1988) Fluvial Processes in dryland rivers. Springer, Berlin Sagri M, Bartolini C, Billi P, Ferrari G, Benvenuti M, Carnicelli S,
Grove AT, Goudie AS (1971) Late quaternary Lake levels in the Rift Barbano F (2008) Latest Pleistocene and Holocene river network
Valley of Southern Ethiopia and elsewhere in tropical Africa. evolution in the Ethiopian Lakes region. Geomorphology 94:79–97
Nature 234:403–405 Schumm SA (1973) Geomorphic thresholds and complex response of
Grove AT, Street FA, Goudie AS (1975) Former lake levels and drainage systems. In: Morisawa M (ed) Fluvial Geomorphology.
climatic changes in the rift valley of Southern Ethiopia. Geogr J Geomorphology, Binghamton State University, Binghamton, pp 299–310
141:177–202 Street FA (1979) Late Quaternary lakes in the Ziway-Shala Basin,
Haines V, Haas H (1974) Southern Methodist University radiocarbon Southern Ethiopia. Unpublished Ph.D. Thesis, University of
date list. Radiocarbon 16:368–380 Cambridge
Hillaire-Marcel C, Carro O, Casanova J (1986) 14C and Th/U dating of Street-Perrott FA (1982) Twentieth century fluctuations in lake level in
Pleistocene and Holocene stromatolites from east african paleola- the Ziway-Shala basin, Ethiopia. Palaeoecol Afr 14:99–110
kes. Quat Res 25:312–329 Street-Perrott FA, Perrott RA (1990) Abrupt climate fluctuations in the
ISSC—International Subcommission On Stratigraphic Classification tropics: the influence of Atlantic Ocean circulation. Nature 343:607–612
(1994) A guide to stratigraphic classification, terminology, and Taieb M, Barker P, Bonnefille R, Damnati B, Gasse F, Goetz C,
procedure, 2nd edn. IUGS, Boulder Hillaire Marcel C, Icole M, Massault M, Roberts N, Vincens A,
Kutzbach JE, Street-Perrott FA (1985) Milankovitch forcing of Williamson D (1991) Histoire paléohydrologique du lac Magadi
fluctuations in the level of tropical lakes from 18 to 0 kyr (Kenia) au Pléistocène supérieur. C R Acad Sci Paris 313:339–346
BP. Nature 317:130–134 Thomas MF, Thorp MB (1995) Geomorphic response to rapid climatic
Laury RL, Albritton CC (1975) Geology of the middle stone age and hydrologic change during the Late Pleistocene and Early
archeological sites in the Main Ethiopian Rift valley. Geol Soc Am Holocene in the humid and sub-humid tropics. Quat Sci Rev
Bull 86:999–1011 14:193–207
Le Turdu C, Tiercelin JJ, Gibert E, Travi Y, Lezzar K, Richert J, Thomas MF, Thorp MB (2003) Palaeohydrological reconstructions for
Massault M, Gasse F, Bonnefille R, Decobert M, Gensous B, Jeudy tropical Africa- evidence and problems. In: Benito G, Gregory KJ
V, Tamrat E, Mohamed MU, Martens K, Balemwal A, Chernet T, (eds) Palaeohydrology: understanding global change. Wiley, Chich-
Williamson D, Taieb M (1999) The Ziway-Shala lake basin system, ester, pp 167–192
Main Ethiopian Rift: influence of volcanism, tectonics, and climatic Verschuren D, Laird KR, Cumming BF (2000) Rainfall and drought in
forcing on basin formation and sedimentation. Palaeogeogr Palae- equatorial East Africa during the past 1,100 years. Nature 403:410–414
oclimatol Palaeoecol 150:135–177 Waters MR, Haynes CV (2001) Late quaternary arroyo formation and
Legesse Dagnachew, Gasse F, Radakovitch O, Vallet-Coulomb C, climate change in the American Southwest. Geology 29:399–402
Bonnefille R, Verschuren D, Gibert E, Barker P (2002) Environmen- Weijers JWH, Schefuß E, Schouten S, Sinninghe Damsté JS (2007)
tal changes in a tropical lake (Lake Abijata, Ethiopia) during recent Coupled thermal and hydrological evolution of tropical Africa over
centuries. Palaeogeogr Palaeoclimatol Palaeoecol 187:233–258 the last deglaciation. Science 315:1701–1704
Maley J (1976) Les variations du lac Tchad depuis un millénaire: Woldegabriel G, Aronson J, Walter RC (1990) Geology, geochronol-
conséquences paleoclimatiques. Palaeoecol Afr 9:44–47 ogy and rift basin development in the central sector of the Main
Mayewski PA, Rohling EE, Stager JC, Karlen W, Maasch KA, Meeker Ethiopian Rift. Geol Soc Am Bull 102:439–458
LD, Meyerson EA, Gasse F, van Kreveld S, Holmgren K (2004)
Holocene climate variability. Quat Res 62:243–255
Water–Rock Interaction and Lake
Hydrochemistry in the Main Ethiopian Rift 18
Azeb Belete, Luigi Beccaluva, Gianluca Bianchini, Nicolò Colombani,
Massimiliano Fazzini, Chiara Marchina, Claudio Natali,
and Tewodros Rango
Abstract
This study investigates the halogen budget of the Main Ethiopian Rift (MER) lithologies and
water, and on the basis of new laboratory experiments gives insights on the water–rock
interaction processes which ultimately mobilize fluoride in the environment. The halogen
composition, and in particular, the chloride content of MER lakes is also taken into
consideration to evaluate compositional variation occurred during the last 80 years that have to
be mainly related evaporative effects. The evaporation trends are also investigated on the basis
of new analyses of stable isotopes of oxygen and hydrogen that are compared with those
available in the literature since the 1970s. In such complex scenario, although the average
annual temperature increased *1 °C in 30 years, we did not observe systematic trends valid
for all the investigated lakes. The record defined for the last 30 years by δ18O–δD denote
fluctuations of the climatic parameters with extreme evaporation preceding the year 2005, then
declining to more “normal” conditions. The relation between the observed climatic parameters
and the water isotopic composition suggests that the study lakes quickly respond to the
environmental changes, possibly within one (or two) year(s). We therefore suggest to continue
the data acquisition of climatic and hydrochemical parameters in order to implement the
existing hydro-archive that could be useful to point out possible environmental changes.

Keywords
Ethiopian lakes Hydro-archive Geochemistry Stable isotopes
18.1 Introduction
A. Belete L. Beccaluva G. Bianchini (&) M. Fazzini
C. Marchina C. Natali Ethiopia is characterized by a great variety of aquatic eco-
Department of Physics and Earth Sciences, University of Ferrara, systems, including several lakes that are of great scientific
Ferrara, Italy
e-mail: gianluca.bianchini@unife.it
interest and economic importance. The majority of these
lakes are confined within the Main Ethiopian Rift (MER)
A. Belete
Department of Earth Sciences, Addis Ababa University, Addis
valley, which is part of a tectonic depression extending from
Ababa, Ethiopia the Afar region in the north to the Kenyan border in the
G. Bianchini C. Natali
south (see Chaps. 16 and 17, this volume, for background
Institute of Geosciences and Earth Resources–National Research information about the origin, hydrology, and evolution of the
Council, Pisa, Italy lakes). The water geochemistry within this rifted belt is
N. Colombani extremely variable because meteoric water recharge from the
Department of Earth Sciences, “Sapienza” University of Rome, surrounding highlands is affected by different degrees of
Rome, Italy water–rock interaction and evaporation processes. The
T. Rango widespread uprising of thermal water adds to the complexity
Nicholas School of the Environment, Division of Earth and Ocean of the rift hydrological system. Moreover, anthropogenic
Sciences, Duke University, Durham, USA
308 A. Belete et al.
activities that altered the water balance during the last dec- in the local populations (Rango et al. 2009, 2010a, b, c,
ades have also to be taken into account (Deribe et al. 2011; 2012, 2013).
Yohannes et al. 2013; Chap. 16, this volume).
The hydrological cycle is regulated by tectonic activity
that could influence the drainage systems at various tem- 18.2 Physiographic and Hydrological
poral/spatial scales modifying the basin input/output, and Framework
also by volcanism and associated geothermal processes that
provide significant geochemical fluxes (Goerner et al. 2009). Remarkable rifting processes, originated by active faulting,
Hydrology is also influenced by changing climatic condi- formed various volcano-tectonic depressions in the rift floor,
tions that variously affect the lacustrine systems, some of in turn favoring the development of closed lacustrine basins
which are endhoreic. Accordingly, geochemical variations in fed by surface- and groundwater. Accordingly, the major
the different water components can be associated with global rivers draining the surrounding plateaus on both sides of the
(or local) climatic variations, implying changes of tempera- rift convey water to lakes and swamps located at different
ture and rainfall (Wagesho et al. 2013). These variations can elevations. The hydrological regime is obviously influenced
be monitored through the isotopic composition of oxygen by climatic parameters such as the monthly distribution of
and hydrogen in the meteoric water (generally expressed as rainfall and temperature (Ayenew 2003; Bewketu 2010).
δ18O and δD) which is site-specific and related to tempera- Within the rift valley, climatic conditions are semiarid/semi-
ture-dependent fractionation processes occurring during the humid, with rainfall ranging between 800 and 1,100 mm/yr
hydrological cycle (Clark and Fritz 1997). In the lakes, the (uncommonly as much as 1,500 mm/year). The precipitation
isotopic fingerprinting is remarkably affected by evaporation regime generally shows a first maximum in the summer
processes (Gat 1995; Rozanski et al. 2000; Froehlich et al. months. Only in the area of Arba Minch, the regime
2005; Kebede et al. 2009), because evaporation preferen- becomes definitely bimodal, with an absolute maximum in
tially taps the lightest (1H and 16O) with respect to the spring and a second peak in late summer. The winter is very
heavier (2H, 18O) isotopes, leading to progressively higher dry everywhere. The average annual temperature is about
2 1
H/ H and 18O/16O proportions in the residual lake water. 20 °C, and the maximum is above 25 °C, with a thermal
Within this framework, we present and discuss new original trend of the last 30 years showing an increase in average
δ18O–δD analyses of MER water (Fig. 18.1). Further detail temperatures of about 1 °C. The potential evaporation is
on the limnological processes that affect the chemical com- around 1,000 mm/year (Billi 1998; Chap. 4, this volume).
position of MER lakes is based on observed hydrogeo- Therefore, rivers and lakes recharge is mainly due to water
chemical changes in the lakes of Ziway, Langano, Shala, deriving from the highlands, where the rainfall is up to
Abijata, Awasa, Abaya, and Chamo which occur in response 1,800 mm/year and the mean annual temperature between 14
to ongoing environmental variations in the area (e.g., Sagri and 16 °C (Billi et al. 2003; Kebede et al. 2009), while
1998; Legesse et al. 2004; Dessie and Kleman 2007). The potential evaporation is less than 700 mm. The series of
dataset considered includes both the present authors’ original measurements of rainfall and temperature in the investigated
data and literature analyses (Omer-Cooper 1930; Loffredo sectors of the rift reported in Fig. 18.2 will be discussed in
and Maldura 1941; Chernet 1982; Von Damm and Edmond the final section in relationships with the measured geo-
1984; Wood and Talling 1988; Gizaw 1996; Zinabu et al. chemical data.
2002; Kebede et al. 2009; Kebede and Travi 2012), focusing For the sake of simplicity, the MER lakes are subdivided
on the concentration of halogens such as the chloride and in three groups: those located in the Awash basin (lakes
fluoride in the water system (Fig. 18.1). It must be stressed Koka, Beseka, Gemeri and Abe), those in the central part of
that chloride is a highly conservative tracer (Apelo and MER (lakes Ziway, Langano, Shala, Abijata, Awasa, Abaya,
Potma 2005), which is commonly used to monitor evapo- and Chamo), and those in the southern MER, close to the
ration processes in surface water (Manno et al. 2007; Kenyan border (e.g., Lake Chew Bahir that stretches into
Kirchner et al. 2010), groundwater (Gianbastiani et al. 2013) northern Kenya). As the first group of lakes is concerned,
and in the vadose zone (Scanlon et al. 2006), whereas Koka Lake is a hydropower reservoir, resulting from the
fluoride is an anion marking the water–rock interaction impoundment of the Awash River, whereas Beseka Lake is
processes (Saxena and Ahmed 2001; Tirumalesh et al. mainly fed by hot springs outpouring in an area crisscrossed
2007). This study also addresses the geochemical source(s) by active faults (Goerner et al. 2009); Abe and Gemeri lakes
and water–rock interaction(s) on the basis of new laboratory correspond to the terminal base level of the Awash River in
experiments that are focused to understand the anomalous the Afar region. The second group of lakes, located in the
concentration of fluoride which is typically high in the nat- central part of MER, were sampled and investigated by the
ural water of the rift valley, causing serious health problems authors and are the specific object of this paper:
18 Water–Rock Interaction and Lake Hydrochemistry … 309
Fig. 18.1 Maps of the studied

lakes located in the Main
Ethiopian Rift (MER), reporting
recent geochemical data on their F Cl
waters. The data (Table 18.1)
include samples collected in Ziway Ziway
January 2006 for lakes Ziway, Abijata 1.5 Abijata 10
Langano, Shala, and Abijata 384 5361
(Rango et al. 2009) and new Langano Langano
analyses carried out on samples 13 131
collected in January 2012 for Shala Shala
lakes Awasa, Abaya, and Chamo. 90 1326
Emphasis is given to the
following tracers: chloride Awasa Awasa
9.6 31
(a) which is extremely
conservative and describes
enrichment by evaporative
processes, fluoride (b) which is a
marker of water–rock interaction Abaya Abaya
processes and limits the use of 9.0 76
MER water for drinking
purposes, and δ18O–δD isotopic
values (c, d) that can help to Chamo Chamo
understand the origin of waters 9.3 156
and the ongoing hydrological
processes
δ18Ο δD
Ziway Ziway
Abijata +5.5 Abijata +38.7
+7.9 +45.7
Langano Langano
+7.0 +43.6
Shala Shala
+4.0 +26.2
Awasa Awasa
+7.8 +45.4
Abaya Abaya
+6.2 +44.2
Chamo Chamo
+6.3 +43.3
• Lakes Abijata, Langano, Shala, and Ziway form an conveying water into Lake Abijata, which is a shallow,
interconnected lakes system, within a closed basin cov- small, closed vlake, characterized by very saline water
ering an area of *13,000 km2. Ziway is characterized by (conductivity >20,000 μS cm−1). Langano Lake, fed
freshest water (conductivity *400 μS cm−1) and from the southeast by five small streams and by several
receives water supply from the Meki and the Ketar rivers, hot springs, is characterized by water with intermediate
which flow from the west and the east, respectively. salinity (conductivity *1,700 μS cm−1), which is par-
Ziway Lake outflows through the Bulbula River tially conveyed to Abijata Lake by the Horakelo River.
Fig. 18.2 Average annual temperature (°C) and precipitation (mm/yr) in areas close to the studied lakes. Lines represent mobile averages in
respect to a 3-year period
Lake Shala (Fig. 18.3, characterized by extremely saline by freshwater (conductivity *800 μS cm−1) although no
water (conductivity >20,000 μS cm−1), is the deepest visible outlet is observed. Thus, most probably subsurface
lake of Ethiopia (up to 266 m) and occupies a down- inflows and outflows occur in this lake.
faulted caldera as do Langano and Abijata lakes; it is fed • Abaya (Fig. 18.5) and Chamo lakes are closely associated
by the Dijo River that drains the western escarpment and within the same drainage basin. In the past, these lakes
by several hot springs. were interconnected, with Abaya Lake water spilling into
• Lake Awasa (Fig. 18.4) is located within a collapsed Chamo Lake during high-stand intervals (Von Damm and
caldera and receives surface contributions from the Edmond 1984). Presently, this is no longer so common,
intermittent Tikur Wuha stream, originating from the though during the rainy season water from Abaya Lake
Cheleleka swamp, which only a century ago was included spills through a small river into Chamo Lake. Lake Abaya
in a larger paleo-Awasa Lake; some other small ephem- is fed by the Bilate and Hamesa rivers and by several hot
eral streams contribute to the lake, which is characterized springs (Chernet 2011) and is characterized by relatively
Fig. 18.3 Shala Lake is the

deepest lake of Ethiopia (266 m)
and formed within a large caldera
Fig. 18.4 Awasa Lake is a

shallow (maximum depth 22 m)
body of freshwater in the middle
of a very large caldera
low salinity (conductivity *800 μS cm−1). The neigh- Shala, *440 km2 for Ziway, and 1,130 km2 for Abaya. The
boring Lake Chamo, fed by the Kulfo River and by other lake surface extension of Ziway, Langano, Shala, and Abaya
minor streams, has a more elevated salinity (conductiv- remained almost constant in the last 30 years, whereas Awasa
ity *1,600 μS cm−1). Generally, the Chamo Lake is expanded and Abijata and Chamo shrunk (Ayenew 2004).
closed, though occasionally overflowing into the Sagan The average water depth is less than 20 m for all lakes, with
River and, in turn, in the ephemeral terminal lake of Chew the exception of Lake Shala which is a crater lake having an
Bahir close to the border with Kenya. This overflow takes average water depth of *90 m (maximum depth about
place only when precipitation in the region is above long- 266 m; Brunelli et al. 1941; Chernet 1982; Le Tardu et al.
term average (Von Dam and Edmond 1984; Kebede et al. 1999; Alemayehu et al. 2006).
1994). With increasing diversion of water from tributary
rivers for irrigation, the overflow toward Chew Bahir and
the surface water connection between Abaya and Chamo 18.3 Geological and Geochemical
lakes are expected to definitely disappear. Processes Influencing the Water
The study lakes are roughly aligned N–NE/S–SW, at Chemistry
altitudes varying from 1,680 to 1,170 m above sea level.
According to the literature data (Table 18.1), their surface The study area is characterized by volcano-tectonic activity.
areas are *90 km2 for Awasa, *160 km2 for Abijata, The paroxysmal phase occurred in the Oligocene, with
*230km2 for Langano, *320 Km2 for Chamo, *330 for lithospheric bulging and uplift of the whole region and a
Fig. 18.5 Abaya Lake is the

second largest lake of Ethiopia
(1,140 km2). Its water is fresh and
very shallow (maximum depth
13 m)
Table 18.1 Physical details of the studied lakes

Altitude (m) Mean Depth Max Depth Surface area
(m) (m) (km2)
Lake Wood and Wood and Kebede Wood and MCE Kebede FAO Wood and MCE Kebede
name Tailing Tailing et al. Tailing et al. Tailing et al.
1988 1988 2009 1988 2001 2009 1982 1988 2001 2009
Ziway 1,639 2.5 4 8.95 8.9 9 434 442 440 442
Langano 1,582 17 20 47.9 47.9 47 230 241 230 241
Abijata 1,578 7.6 5 14.2 14.2 10 205 176 180 159
Shala 1,558 87 103 266 266 266 409 329 370 329
Awasa 1,680 10.7 10 21.6 22 23 130 129 92 92
Abaya 1,285 7.1 – 13.1 24.2 – 1,161 1,162 1,140 1,127
Chamo 1,233 – – 13 14.2 – 551 551 317 347
Data sources are as follows: Wood and Talling 1988; Kebede et al. (2009 and references therein), Awulachew et al. (2007 and references therein).
The label FAO and MCE stand for “Food and Agriculture Organization” and “Metaferia Consulting Engineers”
huge emission of continental flood basalts (Beccaluva et al. oldest one in the early Pleistocene when the existing lakes
2009). Then, differentiated rhyolite magmas appeared late in were united in singular, large paleo-lake (Geze 1975).
the sequence marking the onset of rifting and block tilting The above mentioned weathering processes contribute to
(Natali et al. 2011), which progressively continued leading, release dissolved ions in the surface- and groundwater
ultimately, to the origin of the East African Rift (Corti 2009). interaction, as well as sediment load to rivers and lakes. The
Within this framework, the rift floor is mainly paved by amount of the TDS (total dissolved salts) can be easily
volcanic products (felsic tuffs, ignimbrites, and subordinate assessed in the field by measuring the electrical conductivity
lavas). Locally, these volcanic rocks have been weathered and calculating the salinity by the use of algorithms, which
and the resulting particles have been displaced by surface were properly defined for MER water by Wood and Talling
runoff, leading to fluvial-lacustrine sedimentary formations. (1988) and Zinabu et al. (2002).
The study of these deposits has been useful to understand the Surface and groundwater supplied from the highlands is
evolution of the rivers/lakes network (Gasse and Street 1978; typically characterized by low conductivity, low TDS, and a
Le Tardu et al. 1999; Sagri et al. 2008; Benvenuti et al. calcium bicarbonate facies. Pristine water progressively
2013), giving insights into the climatic fluctuation during the migrating downhill reacts with the interacting lithologies
Quaternary (Ghinassi et al. 2012). Benevenuti et al. (2002 increasing the TDS. This evolution ultimately leads to more
and Chap. 14, this volume) have shown that these lacustrine “mature” water types characterized by a sodium bicarbonate
basins underwent four main lake stand phases, with the hydrochemical facies, which is typical of the rift valley
(Rango et al. 2009). This hydrochemical evolution, resulting volcanic rocks and the fluvio-lacustrine sediments. The
from progressive water–rock/sediment interaction, is leaching experiments carried out for duration of 15, 30, and
accompanied by an increase of pH and the appearance of 90 days with both deionized water and a calcium bicarbonate
dissolved fluoride (and arsenic), which are affecting water water having TDS *130 mg/l that simulate the recharge
quality of springs and wells located in the rift valley, causing water conveyed from the highlands, demonstrated that
severe health problems to the local populations (Rango et al. fluoride is efficiently removed from MER lithologies
2009, 2010a, 2012, 2013). The chloride/fluoride ratio indi- (Table 18.3). In particular, experiments show that among the
cates the water “maturity,” as the halogen content of mete- distinct MER lithologies, the fluvio-lacustrine sediments
oric components is mainly influenced by marine aerosol contain (and release) higher amount of fluoride which is
where chloride is distinctly predominant, while fluoride is plausibly trapped and concentrated by clay minerals (i.e.,
mainly leached from the local rocks that are particularly phyllosilicates in which fluoride can vicariate the oxydril
enriched in this element. Coherently, the chloride/fluoride radicals). This was further constrained by column displace-
ratio of MER rainwater (*12 and 0.05 mg/l of Cl and F, ment tests on volcanic ashes having different grain size, in
respectively, have been measured in a rainwater collected in which the fluoride concentration increased with decreasing
Arba Minch during January 2012) is extremely high with Cl/ the grain size (Rango et al. 2010c). In addition, an elevated
F up to 200, whereas Cl/F progressively diminishes to values mobility of fluoride was observed in saturated conditions,
of 1 in some hot springs, as result of pervasive water–rock/ with much of the available budget leached in a few column
sediment interactions. pore volumes, similar to chloride.
Better understanding of these processes necessarily The real MER waters, and especially those of the study
requires the geochemical study of the lithologies outcrop- lakes, are also subject to effective evaporation processes that
ping in the rift valley, which are mainly constituted by further concentrate the halogens.
volcanic glass (amorphous silicate phases) or their weath-
ering products developed during supergene processes. In
Fig. 18.6, we report a SEM picture showing the particles of a 18.4 Lake Water Composition
volcanoclastic deposit sampled close to the Ziway Lake and
the significant presence of a crystal of fluoro-apatite among The open or closed nature of each drainage system and the
the vesciculated glass shards. The high fluorine activity of effective in/out seepage obviously influence the salinity of
the original magmas is supported by the SEM chemical the study lakes that is generally expressed as milligrams per
analyses of both glass and fluoro-apatite (F up to 0.38 and liter of TDS. Accordingly, in the study lakes, TDS is
8.2 wt%, respectively) (Table 18.2). To constrain the reac- extremely heterogeneous, with values of 0.3–0.4 g/l in Zi-
tivity of MER lithologies to the atmospheric agents, we way, *1.4 g/l in Langano, *11.5 g/l in Shala, *52.7 g/l in
carried out water extraction tests in which 10 g of powdered Abijata, *0.7 g/l in Awasa, *0.8–0.9 mg/l in Abaya,
rock was leached by 50 ml of water, investigating both and *1.0–1.2 mg/l in Chamo; pH is generally higher than 8,
Fig. 18.6 Backscattered SEM

image of a MER volcanoclastic
rock. Vesciculated particles
characterized by light gray color
consist of glass, whereas the
euhedral white crystal is fluorine-
vesciculated
rich apatite glassy shards
100 μm apatite
Table 18.2 SEM analyses of glassy particles and apatite crystals recorded within the volcanoclastic deposit shown in Fig. 18.3
EDS analyses of glassy particles in TW39. Analytical conditions were adjusted for the analyses of halogens, and F was
detected by WDS
SiO2 71.65 71.88 72.30 72.24 70.15 69.90 69.07 68.96 71.42 70.95 69.51 69.36 69.69 69.42 72.85
TiO2 0.32 0.29 0.32 0.30 0.28 0.27 0.28 0.31 0.33 0.32 0.30 0.31 0.28 0.31 0.31
Al2O3 8.12 8.07 8.12 8.13 7.93 7.83 7.77 7.74 8.07 7.96 7.69 7.87 7.80 7.76 8.20
FeO 6.03 6.11 6.09 6.21 5.93 5.92 5.79 5.78 6.04 5.96 5.94 5.87 5.92 5.82 6.15
MnO 0.27 0.20 0.25 0.24 0.22 0.25 0.23 0.25 0.24 0.24 0.24 0.23 0.26 0.25 0.27
MgO n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.
CaO 0.00 0.25 0.28 0.24 0.27 n.d. n.d. n.d. n.d. n.d. n.d. 0.27 0.26 n.d. 0.24
Na2O 0.79 0.70 0.61 0.36 3.93 3.14 4.39 3.43 1.34 1.18 3.81 3.09 2.19 3.28 0.43
K2O 3.00 2.88 2.77 2.21 4.14 4.12 4.07 4.17 3.76 3.44 4.10 4.18 4.14 4.09 1.99
P2O5 n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.
F 0.37 0.43 0.30 0.29 0.30 0.31 0.38 0.36 0.34 0.30 0.35 0.31 0.26 0.30 0.30
Cl 0.19 0.19 0.22 0.21 0.20 0.20 0.19 0.17 0.20 0.19 0.19 0.22 0.18 0.20 0.20
EDS analyses of apatite crystals in TW39
SiO2 1.43 1.34 1.36 1.26 1.13 1.20 1.44 1.42 1.66
TiO2 n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.
Al2O3 n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.
FeO n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.
MnO n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.
MgO n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.
CaO 52.48 53.53 53.71 53.58 54.02 54.13 53.60 53.71 53.61
Na2O n.d. n.d. n.d. n.d. 0.37 0.48 0.34 0.37 n.d.
K2O n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.
P2O5 36.52 37.37 37.69 37.39 37.89 37.71 37.38 38.15 38.43
F 7.64 6.91 7.27 7.76 7.57 8.15 7.47 6.76 6.63
Cl n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.
Glass shows dacite and trachydacite compositions with F/Cl ratio always greater than one. Apatite is extremely enriched in F with negligible amount of Cl;
n.d. = not detected
Table 18.3 Results of leaching experiments simulating water-rock interaction

Deionized water leachates (mg/l) Calcium bicarbonate water leachates (mg/l)
Bulk analysis (mg/kg) 15 days 30 days 90 days 15 days 30 days 90 days
Ted 4 Pumice deposit 1,690 7.73 7.48 8.93 6.42 6.79 7.79
TW31 Ignimbrite deposit 985 3.16 3.66 3.66 2.94 3.22 3.53
TW39 Pyroclastic deposit 1,610 1.62 1.82 2.35 1.77 1.98 2.41
TW22 Sediment 2,510 8.00 8.01 8.52 7.80 7.76 8.24
Ted 38 Sediment 1,590 8.73 7.64 n.a. 5.97 6.68 8.26
The first column reports the fluorine content of MER lithologies (mg/kg), including volcanic rocks and sediments which results from the
weathering, reworking, and deposition of the volcanic rocks. The analysis has been carried out by the ion-selective electrode (ISE) NexSens WQ-
FL after alkaline fusion (0.4 g of sample + 0.3 of KNO3 + 2 g of Na2CO3 heated at 1,000 °C for 40 min) and dissolution in deionized water. Other
columns report the fluoride content of solutions (mg/l) obtained by the interaction of the above mentioned MER lithologies (10 g of powder) with
water (50 ml). The leaching experiments were performed with deionized water and with a Ca-HCO3 water simulating the water conveyed in the rift
from the highlands, analyzing the solutions by IC (Ion chromatography; Dionex ICS 1000) after 15, 30, and 90 days; n.a. = not analyzed
Table 18.4 Physico-chemical parameters of MER water (pH, TC, and EC) measured in situ, laboratory analyses of halogens (Cl, F Ion
chromatography; Dionex ICS 1000) and isotopic ratios of oxygen and hydrogen
Code Year Coordinates T (°C) pH EC μS/cm Cl (mg/l) F (mg/l) δ18O ‰ δD ‰
Meki River RI30 2006 38.822 8.153 23.3 8.66 530 16 1.5 1.7 6.1 Rango et al. (2010a, b)
Ketar River RI31 2006 39.003 8.046 21.5 7.83 180 4 1.6 −2.8 −11.3 Rango et al. (2010a, b)
Groundwater flow WL14 2006 38.966 8.120 24.6 7.48 350 145 3.1 −3.4 −13.7
toward Lake Ziway WL18 2006 38.822 8.159 25.2 7.82 490 4 0.9 −2.7 −11.0 Rango et al. (2010a, b)
WL20 2006 38.532 8.006 36.4 7.89 730 7 3.1 −3.5 −19.7
WL23 2006 38.812 8.145 25.1 7.81 2750 176 20 −2.1 −5.2 Rango et al. (2010a, b)
13 2011/12 −4.2 −22.8 Rango et al. (2013)
14 2011/12 −3.7 −18.3 Rango et al. (2013)
15 2011/12 −2.0 0.0 Rango et al. (2013)
16 2011/12 −3.1 −8.5 Rango et al. (2013)
17 2011/12 −2.7 −3.9 Rango et al. (2013)
18 2011/12 −4.2 −24.2 Rango et al. (2013)
19 2011/12 −4.7 −24.7 Rango et al. (2013)
20 2011/12 0.2 7.3 Rango et al. (2013)
21 2011/12 −5.1 −28.1 Rango et al. (2013)
22 2011/12 −3.4 −11.8 Rango et al. (2013)
23 2011/12 −3.2 −13.8 Rango et al. (2013)
Ziway Island spring HS2 2006 38.853 7.934 78.0 6.79 2,040 13 128 −3.0 −15.7 Rango et al. (2010a, b)
Lake Ziway LW29 2006 38.736 7.919 25.2 8.60 420 10 1.5 4.9 33.8 Rango et al. (2010a, b)
2011/12 6.1 43.6 Rango et al. (2013)
Spring at the HS1 2006 38.773 7.690 96.0 8.36 4,530 429 24 −0.4 −1.0 Rango et al. (2010a, b)
Langano shore HS9 2006 38.811 7.664 38.8 7.33 630 27 1.9 −3.4 −15.9 Rango et al. (2010a, b)
HS10 2006 38.773 7.712 62.0 6.96 3,900 435 23 −1.2 −3.4 Rango et al. (2010a, b)
Lake Langano LW25 2006 38.684 7.538 24.2 9.03 1,730 131 13 7.0 43.6 Rango et al. (2010a, b)
Spring at the HS3 2006 38.634 7.477 93.4 8.16 14,440 1,462 97 2.5 20.7 Rango et al. (2010a, b)
Shala Shore HS4 2006 38.634 7.477 57.3 7.81 10,570 1,324 64 3.7 24.2 Rango et al. (2010a, b)
HS5 2006 38.634 7.477 91.4 8.00 13190 1,148 55 2.3 20.1 Rango et al. (2010a, b)
HS6 2006 38.637 7.477 52.6 8.15 1,780 17 4.5 −0.4 2.8 Rango et al. (2010a, b)
HS7 2006 38.723 7.402 59.3 7.30 4,630 321 18 1.4 12.3 Rango et al. (2010a, b)
Lake Shala LW27 2006 38.435 7.423 25.8 9.60 >20,000 1,326 90 4.0 26.2 Rango et al. (2010a, b)
Lake Abijata LW28 2006 38.595 7.671 27.6 9.67 >20,000 5,361 384 7.9 45.7 Rango et al. (2010a, b)
Springs in the Lake
Awasa catchment AZ1 2012 38.522 6.993 60.9 6.2 1,200 36 n.d. −1.0 0.3 Original data
AZ2 2012 38.523 6.991 >80 6.9 2,330 73 14.6 −0.4 3.4 Original data
AZ17 2012 38.638 7.079 65.6 7.4 1,160 25 1.6 −3.0 −7.4 Original data
AZ18 2012 38.644 7.071 59 7.9 1,120 27 1.8 −2.9 −5.8 Original data
AZ19 2012 38.643 7.076 59 7.7 1,020 25 1.5 −3.0 −6.6 Original data
Lake Awasa AZ3 2012 38.458 7.056 24.6 9.9 810 31 9.6 7.8 45.4 Original data
Springs in the Lake AZ4 2012 38.129 6.935 36.6 7.2 1,030 20 16.0 −2.9 −14.7 Original data
Abaya catchment AZ5 2012 38.079 6.935 39 6.9 950 17 17.0 −2.5 −14.0 Original data
AZ6 2012 38.121 6.938 39.2 6.8 960 18 17.0 −2.6 −12.8 Original data
AZ11 2012 38.134 6.928 56.5 8.3 1,150 37 26.5 −2.5 −10.2 Original data
AZ12 2012 38.088 6.810 45.3 7.8 1,340 51 18.4 −1.9 −3.4 Original data
AZ13 2012 38.087 6.811 56 7.5 1,480 52 21.0 −1.6 −3.1 Original data
AZ14 2012 38.046 6.755 66.5 7.6 5,530 147 45.2 1.7 2.5 Original data
AZ15 2012 38.046 6.756 63 7.8 5,500 162 43.5 0.7 4.4 Original data
AZ16 2012 38.045 6.723 > 80 7.9 6,200 144 44.9 −2.3 0.4 Original data
AZ22 2012 37.909 6.636 65 6.6 2,700 65 18.1 −3.1 −11.1 Original data
AZ23 2012 37.910 6.636 60.8 7.3 2,560 70 18.3 −3.0 -9.7 Original data
AZ31 2012 37.577 6.243 16.8 6 60 9 0.6 −2.8 −6.0 Original data
AZ32 2012 37.577 6.241 18.3 5.3 10 1 0.1 −2.4 −3.9 Original data
Hamesa River AZ24 2012 37.825 6.568 35.5 6.4 420 4 2.3 −2.7 −7.5 Original data
(continued)
Table 18.4 (continued)

Code Year Coordinates T (°C) pH EC μS/cm Cl (mg/l) F (mg/l) δ18O ‰ δD ‰
Lake Abaya AZ33 2012 37.629 6.119 25.5 9.1 970 76 9.0 6.2 44.2 Original data
Springs in the Lake AZ25 2012 37.691 5.887 40.1 7.9 2,270 116 11.0 −2.0 −1.7 Original data
Chamo catchment AZ26 2012 37.691 5.884 45.5 8.0 2,400 115 11.1 −1.8 −1.5 Original data
AZ27 2012 37.692 5.886 45.0 7.5 2,450 118 11.4 −1.8 −2.1 Original data
AZ28 2012 37.692 5.886 55.6 7.7 2,600 115 10.9 −2.0 −1.5 Original data
Lake Chamo AZ34 2012 37.535 5.929 28.8 9.3 1,640 156 9.3 6.3 43.3 Original data
18 16 2 1
In particular, the new isotopic analyses ( O/ O and H/ H) were determined by laser absorption spectrometry using the CRDS LOS GATOS LWIA 24d isotopic analyzer,
reporting results as δ18 O and δD ‰ relative to the Standard Mean Ocean Water (SMOW)
increasing up to 10 in the terminal lakes such as Shala and

Abijata. These characteristics delineate a notable salinity–
alkalinity series, in which the waters are invariably charac-
terized by a sodium bicarbonate hydrochemical facies. The
water of Abijata Lake, which is subjected to high rates of
evaporation (Ayenew 2002; Legesse et al. 2004), represents
an exception as the hydrochemical facies is displaying a
peculiar sodium chloride pattern.
Some of these lakes do not appear chemically stratified as
demonstrated by chemical analysis of Awasa and Shala
water before and after detonation causing effective mixing
(Klemperer and Cash 2007). Others, like Lake Chamo,
present variation of the salinity along the water column
(Deriemaecker 2013).
18.5 Results and Discussion
In this paper, particular attention is given to the changes of

lake water composition over time, presenting new data
(Table 18.4) which are discussed together with those avail-
able in the literature. This can be done taking into account
the earliest hydrochemical data that are available for most of
the considered lakes since the year 1937 (Loffredo and Fig. 18.7 Temporal evolution of the chloride content in the studied
Maldura 1941), and for Lake Abijata since 1926 (Omer- lakes. The authors’ data (Table 18.4) include samples collected in
January 2006 for lakes Ziway, Langano, Shala, and Abijata (Rango
Cooper 1930). In this comparison, an emphasis is focused on
et al. 2009) and new analyses carried out on samples collected in
chloride that is a conservative element, the variation of January 2012 for lakes Awasa, Abaya, and Chamo, and are compared
which has to be related mainly to evaporation/dilution pro- with older literature data (Omer-Cooper 1930; Loffredo and Maldura
cesses. The comparison of the chloride content of the study 1941; Chernet 1982; Von Damm and Edmond 1984; Wood and Talling
1988; Gizaw 1996; Zinabu et al. 2002) to emphasize possible
lakes in different periods is synthesized in Fig. 18.7.
evolutionary trends
As the Ziway–Langano–Abijata–Shala system is con-
cerned, we observe that in the last 80 years, the chloride
content is constant in Ziway Lake, slightly decreasing in This demonstrates that there is not a unique—ubiquitous—
Langano and Shala lakes, and increasing only in Abijata trend valid for the whole region and indicates a more complex
Lake. It is worth noting that the latter has been impacted by framework. The variation of the chloride content through the
anthropogenic activities (Ayenew 2002) and exploited to investigated interval suggests that evaporative rate remained
obtain trona (Na2CO3·NaHCO3·2H2O) and finally soda relatively constant in the Ziway–Langano–Abijata–Shala and
(Na2CO3). The chloride content of Awasa and Abaya lakes Awasa systems, whereas evaporation slightly increased in the
has remained quite constant, whereas in Chamo Lake, TDS Abaya–Chamo sector. This evidence implies that microcli-
increased during the last 20 years, plausibly reflecting pro- matic conditions diversify the water balance in distinct MER
gressive isolation of its basin. sectors. However, other factors have to be considered since
80 80
Groundwater from the
D Highlands surrounding
D
Ziway
60 60
River inflow (Meki,
Ketar)
Hot springs close Lake
40 40 Langano
Lake Ziway
Lake Langano
20 20
Hot Spring within Lake
Hot Springs close Lake
Ziway
Shala
0 0
River ouflow (Bulbula) Lake Shala
-20 -20
18 O Lake Abiyata
18 O
-40 -40
-10 -5 0 5 10 15 -10 -5 0 5 10 15
80 80
D D
60 60 Springs in the
Abaya catchment
40 Springs within the 40 Inflow river (Bilate)
Awasa catchment
20 Inflow River 20 Lake Abaya
Awasa Lake Springs in the

0 0
Chamo catchment
Lake Chamo
-20 -20
18 O 18 O
-40 -40
-10 -5 0 5 10 15 -10 -5 0 5 10 15
Fig. 18.8 δ18O–δD isotopic compositions of water of the studied lakes samples collected in January 2012 for lakes Awasa, Abaya, and
and related inflow rivers/springs. The authors’ data (Table 18.1) include Chamo, and are supplemented with data provided by Kebede et al.
samples collected in January 2006 for lakes Ziway, Langano, Shala, (2009)
and Abijata (Rango et al. 2010b; 2013) and new analyses carried out on
tectonic process and human impact (mainly land use change) measured in the ultra-evaporated water of lake Abijata,
could have contributed to overprint the climatic effects on the showing δ18O 10.8 and δD 74.8.
hydrological budget (Vallet-Coulomb et al. 2001; Ayenew The effects of evaporation processes are also emphasized
2002; Gebreegziabher 2004; Legesse 2004). in the δ18O–δD diagrams (Fig. 18.9), in which the lake
To provide further indication on the possible relationships compositions are compared with the local meteoric lines
between climatic changes and water composition, we (i.e., the isotopic composition of the local rains represented
focused on the water isotopic ratios of hydrogen and oxygen by the black lines), specifically proposed for the study area
(Table 18.4). As mentioned in the introductive section, the by Rango et al. (2010b) and Kebede and Travi (2012). The
δ18O–δD of meteoric water is related to temperature- isotopic compositions of the lake waters systematically
dependent fractionation processes occurring during the deviate from these meteoric lines, showing a δ18O dis-
hydrological cycle, and each geographic site displays a placement toward the right quadrant of the diagram. This
specific precipitation composition, which in turn is trans- deviation led Rango et al. (2010b) to define a peculiar local
ferred to the local surface waters (Clark and Fritz 1997). In evaporation line (LEL; dashed line), specific for the MER
the study area, the inflow rivers and the springs that signif- region.
icantly contribute to the lake water budget have light isotopic It is worth to note the comparison between our isotopic
composition often characterized by negative δ18O and δD, data and those reported in the literature for the last 35 years.
whereas the lake water invariably have heavy isotopic In particular, our sampling campaigns were carried out in
composition characterized by positive δ18O and δD January 2006 for Ziway, Langano, Shala, and Abijata lakes
(Fig. 18.8). The most extreme isotopic composition has been (and related rivers and springs) and in January 2012 for
Fig. 18.9 Temporal evolution of 90 90

the δ18O–δD isotopic D D
80 80
compositions of water of the
studied lakes. The authors’ data 70 70
(Table 18.1) include samples
60 60
collected in January 2006 for
lakes Ziway, Langano, Shala, and 50 50
Abijata (Rango et al. 2010b; Lake Ziway Lake Langano
2013) and new analyses carried 40 40
Pre-2000 Pre-2000
out on samples collected in 30 2005 30 2005
January 2012 for lakes Awasa, 2006 2006
Abaya, and Chamo, and are 20 20
2011
compared with older literature
10 10
data provided by Kebede et al. 18 O 18 O
(2009 and references therein). 0 0
Composition of local rains is 0 2 4 6 8 10 12 0 2 4 6 8 10 12
delineated by the local meteoric 90 90
water lines (LMWL; black lines) D D
defined for the MER region by 80 80
Rango et al. (2010b) and Kebede 70 70

and Travi (2012). Evaporative
evolutionary trends are also 60 60
delineated by the local 50 Lake Shala 50
evaporation line (LEL) defined by
Lake Abiyata
Rango et al. (2010b; dashed line) 40 Pre-2000 40
Pre-2000
30 2005 30
2003
2006
20 20 2006
10 10
18 O 18 O
0 0
0 2 4 6 8 10 12 0 2 4 6 8 10 12
90 90
D D
80 80
70 70
60 60
50 50
Lake Awasa Lake Abaya
40 40
Pre-2000 Pre-2000
30 30
2005 2005
20 2012 20 2012
10 10
18 O 18 O
0 0
0 2 4 6 8 10 12 0 2 4 6 8 10 12
90
D
80
70
60
50
40
Lake Chamo
Pre-2000
30
2005
20 2012
10
18 O
0
0 2 4 6 8 10 12
Awasa, Abaya, Chamo lakes (and related rivers and springs). preceding the year 2005 and then declining to more “nor-
The isotopic data available in the literature for the same mal” conditions.
lakes are referred to periods before the year 2000 (Craig We therefore propose to continue the monitoring of the
et al. 1977; Gizaw 1989; Ayenew 1998) and to the year 2005 same climatic and hydrochemical parameters in the next
(Kebede et al. 2009). years in order to improve the existing dataset, which could
An univocal temporal progression of evaporation trends be useful to unravel possible environmental changes.
(i.e., an increase of the δ18O–δD values along the LEL line)
is not observed. Considering literature data, we observe an
increase of the evaporative effects in the year 2005 with References
respect to the last decade of the twentieth century. By con-
trast, after the year 2005, our 2006 and 2012 data suggest a Alemayehu T, Ayenew T, Kebede S (2006) Hydrogeochemical and
decline of the evaporative effects. These trends seem to be lake level changes in the Ethiopian Rift. J Hydrol 316:290–300
supported by the climatological parameters (Fig. 18.2), Appelo CAJ, Postma D (2005) Geochemistry, groundwater and
pollution, 2nd edn. Balkema, Rotterdam
which emphasize fluctuations and a significant climatic Awulachew SB, Yilma AD, Loulseged M, Loiskandl W, Ayana M,
forcing around the years 2004/2005. This, in turn, implies Alamirew T (2007) Water resources and irrigation development in
that the study lakes quickly respond to the environmental Ethiopia. International Water Management Institute, Colombo, 78 p.
changes, possibly within one (or two) year(s). For this rea- (Working Paper 123)
Ayenew T (1998) The hydrogeological system of the Lake District
son, they represent important hydro-archives (Zuppi and Basin, Central Main Ethiopian Rift. PhD Thesis, Free University of
Sacchi 2004) and their continuous isotopic monitoring will Amsterdam, The Netherlands
be useful to evaluate ongoing climatic changes. Ayenew T (2002) Recent changes in the level of Lake Abiyata, central
main Ethiopian Rift. Hydrolog Sci J 47:493–503
Ayenew T (2003) Environmental isotope-based integrated hydrogeo-
logical study of some Ethiopian rift lakes. J Radioanal Nucl Ch
18.6 Conclusions 257:11–16
Ayenew T (2004) Environmental implications of changes in the levels
Evidence of ongoing changes in lake physiography and of lakes in the Ethiopian Rift since 1970. Reg Environ Change
water chemistry is potentially related to concomitant chan- 4:192–204
Beccaluva L, Bianchini G, Natali C, Siena F (2009) Continental flood
ges in the climatic conditions. Unfortunately, the climatic basalts and mantle plumes: a case study of the northern Ethiopian
influence is disturbed by multiple natural and anthropogenic plateau. J Petrolol 50:1377–1403
factors. For example, active tectonic processes could favor Benvenuti M, Carnicelli S, Belluomini G, Dainelli N, Di Grazia S,
the formation of new hydrothermal springs that change the Ferrari GA, Iasio C, Sagri M, Ventra ., Atnafu B, Kebede S (2002)
The Ziway–Shala lake basin (main Ethiopian rift, Ethiopia): a
water balance in terms of volumes and/or chemistry and can revision of basin evolution with special reference to the Late
alter the water inflow/outflow. Human activity (such as Quaternary. J Afr Earth Sci 35: 247–269
deforestation) could increase surface runoff facilitating and Benvenuti M, Bonini M, Tassi F, Corti G, Sani F, Agostini A, Manetti
speeding water flow toward the lakes, but at the same time, P, Vaselli O (2013) Holocene lacustrine fluctuations and deep CO2
degassing in the northeastern Lake Langano Basin (Main Ethiopian
withdrawal of water from the lakes, mainly for agricultural Rift). J Afr Earth Sci 77:1–10
purposes, may occur. Bewketu K (2010) Hydrodynamics of selected ethiopian Rift Lakes.
In such a complex scenario, we do not observe systematic MSc thesis, Addis Ababa University
trends valid for all the investigated lakes. The only two lakes Billi P, Fazzini M, Fratianni S, Dramis F, Biancotti A, Bisc C (2003)
Etude de la distribution spatio-temporelle des parametres clima-
that show a coherent temporal decrease in the surface tiques dans le territoire de l’Ethiopie: Premiers resultats [Study of
extension coupled with an increase of salinization are Lake the spatiotemporal distribution of climatic parameters in the territory
Abijata and Lake Chamo. The behavior of Lake Awasa is of Ethiopia: initial results]. Dokumentacja Geograficzna 29:49–52
difficult to be deciphered, because Table 18.1 suggests that Billi P (1998) Climatic change. In: Sagri M (ed.) Land resources
inventory, environmental changes analysis and their application to
the surface is shrinking, but other observations reveal that agriculture in the Lake Region (Ethiopia), Final Report, STD3
the level has slightly risen (Ayenew 2004) and the water Project n. CT92-0076, European Commission, 16–18
chemistry (Fig. 18.7) does not show substantial variations. Brunelli G, Canicci G, Loffredo S, Maldura CM, Morandini G,
Other lakes show nearly constant areal extension and Parenzan P, Vatova A, Zolezzi C (1941) Esplorazione dei laghi
della Fossa Galla. Collezione Scientifica e Documentaria dell’Africa
steady hydrochemistry, with second-order fluctuations that Italiana III, Vol I, Ministero dell’Africa Italiana, Roma
are difficult to be decoded in terms of climatic changes. This Chernet T (1982) Hydrogeology of the Lakes Region, Ethiopia (Lakes
incongruence probably suggests that the physiographic data Ziway, Langano, Abiyata, Shalla and Awasa). Report of the
reported in the literature are scarcely constrained and have to Ethiopian Institute Geological Survey, 97 p
Chernet T (2011) Geology and hydrothermal resources in the northern
be taken with caution. Moreover, the record defined for the Lake Abaya area (Ethiopia). J Afr Earth Sci 61:129–141
last 30 years by stable isotopes also denotes fluctuations of Clark ID, Fritz P (1997) Environmental isotopes in hydrogeology. CRC
the climatic parameters with a more extreme evaporation just Press, Boca Raton
Corti G (2009) Continental rift evolution: from rift initiation to Legesse D, Vallet-Coulomb C, Gasse F (2004) Analysis of the
incipient break-up in the Main Ethiopian Rift, East Africa. Earth Sci hydrological response of a tropical terminal lake, Lake Abiyata
Rev 96:1–53 (Main Ethiopian Rift Valley) to changes in climate and human
Craig H, Lupton J E, Horowiff RM (1977) Isotope geochemistry and activities. Hydrol Processes 18(3):487–504
hydrology of geothermal waters in the Ethiopian Rift Valley. Le Turdu C, Tiercelin JJ, Gibert E, Travi Y, Lezzar KE, Richert JP,
Scripps Inst. of Oceanography, Report 77–14, 160 p Massault M, Gasse F, Bonnefille R, Decobert M, Gensous B, Jeudy
Deribe E, Rosseland BO, Borgstrøm R, Salbu B, Gebremariam Z, V, Tamrat E, Umer M, Martens K, Atnafu B, Chernet T,
Dadebo E, Norli HR, Eklo OM (2011) Bioaccumulation of Williamson D, Taieb M (1999) The Ziway-Shala lake basin system,
persistent organic pollutants (POPs) in fish species from Lake Main Ethiopian Rift: Influence of volcanism, tectonics, and climatic
Koka, Ethiopia: the influence of lipid content and trophic position. forcing on basin formation and sedimentation. Palaeogeogr Palaeocl
Sci Tot Env 410–411:136–145 150:135–177
Deriemaecker A (2013) A comparison regarding the physico-chemical Loffredo S, Maldura CM (1941) Risultati generali delle ricerche di
variables and zooplankton community characteristics of two Ethi- chimica limnologica sulle acque dei laghi dell’Africa orientale
opian Rift Valley Lakes: Lake Chamo and Lake Abaya. Unpublished italiana esplorati dalla Missione Ittiologica. Esplorazione dei laghi
Master thesis at the KU Leuven autonomous university, 81 p della Fossa Galla. Collezione Scientifica e Documentari dell’Africa
Dessie G, Kleman J (2007) Pattern and magnitude of deforestation in Italiana III, Vol. I, Ministero dell’Africa Italiana, Roma: 181–200
the South Central Rift Valley Region of Ethiopia. Mt Res Dev Manno E, Vassallo M, Varrica D, Dongarrà G, Hauser S (2007)
27:162–168 Hydrogeochemistry and water balance in the coastal wetland area of
Froehlich K, Gonfiantini R, Rozanski K (2005) Isotopes in lakes Biviere di Gela, Sicily, Italy. Water Air Soil Poll 178:179–193
studies: a historical perspective. In: Aggarwal PK, Gat JR, Froehlich Natali C, Beccaluva L, Bianchini G, Siena F (2011) Rhyolites
K (Eds) Isotopes in the water cycle: present and future of a associated to Ethiopian CFB: clues for initial rifting at the Afar
developing science, 139–150 plume axis. Earth Planet Sci Lett 312:59–68
Gasse F, Street FA (1978) Late quaternary lake-level fluctuations and Omer-Cooper J (1930) Dr. Hugh Scott’s expedition to Abyssinia. A
environments of the northern Rift Valley and Afar region (Ethiopia preliminary investigation of the freshwater fauna of Abyssinia. Proc
and Djibouti). Palaeogeogr Palaeocl 24:279–325 Zool Soc Lond 195–207
Gat JR (1995) Stable isotopes of fresh and saline lakes. In: Lerman A, Rango T, Bianchini G, Beccaluva L, Ayenew T, Colombani N (2009)
Imboden D, Gat J (eds) Physics and chemistry of lakes. Springer, Hydrogeochemical study in the main Ethiopian Rift: new insights to
Berlin, pp 139–162 source and enrichment mechanism of fluoride. Environ Geol
Gebreegziabher I (2004) Assessment of the water balance of Lake 58:109–118
Awassa catchment, Ethiopia. MSc thesis, International Institute for Rango T, Bianchini G, Beccaluva L, Tassinari R (2010a) Geochemistry
Geo-Information Science and Earth Observation, Enschede and water quality assessment of central main Ethiopian Rift natural
Geze F (1975) New dates on ancient Galla lakes levels. Bulletin waters with emphasis on source and occurrence of fluoride and
Geophysics Observatory University Addis Ababa 15, 119-124 arsenic. J Afr Earth Sci 57:479–491
Ghinassi M, D’Oriano F, Benvenuti M, Awramik S, Bartolini C, Fedi Rango T, Petrini R, Stenni B, Bianchini G, Slejko F, Beccaluva L,
M, Ferrari G, Papini M, Sagri M, Talbot M (2012) Shoreline Ayenew T (2010b) The dynamics of central main Ethiopian Rift
fluctuations of Lake Hayk (northern Ethiopia) during the last waters: evidence from δD, δ18O and 87Sr/86Sr ratios. Appl
3500 years: geomorphological, sedimentary, and isotope records. Geochem 25:1860–1871
Palaeogeogr Palaeocl 365–366:209–226 Rango T, Colombani N, Mastrocicco M, Bianchini G, Beccaluva L
Giambastiani BMS, Colombani N, Mastrocicco M (2013) Character- (2010c) Column elution experiments on volcanic ash: geochemical
ization of the lowland coastal aquifer of Comacchio (Ferrara, Italy): implications for the Main Ethiopian Rift waters. Water Air Soil Poll
hydrology, hydrogeochemistry and evolution of the system. 208:221–233
J Hydrol 501:35–44 Rango T, Kravchenko J, Atlaw B, Peter GM, Jeuland M, Merola B,
Gizaw B (1989) Geochemical investigation of the Aluto-Langano Vengosh A (2012) Groundwater quality and its health impact: an
Geothermal field, Ethiopian Rift Valley. M. Phil. Thesis, University assessment of dental fluorosis in rural inhabitants of the main
of Leeds, UK Ethiopian Rift. Environ Int 43:37–47
Gizaw B (1996) The origin of high bicarbonate and fluoride Rango T, Vengosh A, Dwyer G, Bianchini G (2013) Mobilization of
concentrations in waters of the Main Ethiopian Rift Valley, East arsenic and other naturally occurring contaminants in groundwater
African Rift system. J Afr Earth Sci 22:391–402 of the Main Ethiopian Rift aquifers. Water Res 47:5801–5818
Goerner A, Jolie E, Gloaguen R (2009) Non-climatic growth of the saline Rozanski K, Froehlich K, Mook WG (2000) Surface water. In: Mook
Lake Beseka, Main Ethiopian Rift. J Arid Environ 73:287–295 WG (Ed) Environmental isotopes in the hydrological cycle:
Kebede E, Zinabu GM, Ahlgren I (1994) The Ethiopian rift valley principles and applications. Technical documents in hydrology
lakes: chemical characteristics of a salinity-alkalinity series. Hyd- (IAEA)
robiologia 288:1–12 Sagri M (ed) (1998) Land resources inventory, environmental changes
Kebede S, Travi Y, Rozanski K (2009) The δ18O and δ2H enrichment analysis and their application to agriculture in the Lake Region
of Ethiopian lakes. J Hydrology 365:173–182 (Ethiopia), Final Report, STD3 Project n. CT92-0076, European
Kebede S, Travi Y (2012) Origin of the δ18O and δ2H composition of Commission, Bruxelles
meteoric waters in Ethiopia. Quatern Intern 257:4–12 Sagri M, Bartolini C, Billi P, Ferrari G, Benvenuti M, Carnicelli S,
Kirchner JW, Tetzlaff D, Soulsby C (2010) Comparing chloride and Barbano F (2008) Latest pleistocene and holocene river network
water isotopes as hydrological tracers in two Scottish catchments. evolution in the Ethiopian Lakes Region. Geomorphology 94:79–97
Hydrol Process 24:1631–1645 Saxena VK, Ahmed S (2001) Dissolution of fluoride in groundwaters: a
Klemper SL, Cash MD (2007) Temporal geochemical variation in water–rock interaction study. Environ Geol 40:1084–1087
Ethiopian Lakes Shala, Arenguade, Awasa, and Beseka: possible Scanlon BR, Keese KE, Flint AL, Flint LE, Gaye CB, Edmunds WM,
environmental impacts from underwater and borehole detonations. Simmers I (2006) Global synthesis of groundwater recharge in
J Afr Earth Sci 48:174–198 semiarid and arid regions. Hydrol Process 20:3335–3370
Tirumalesh K, Shivanna K, Jalihal AA (2007) Isotope hydrochem- Wood RB, Talling JF (1988) Chemical and algal relationships in a
ical approach to understand fluoride release into groundwaters of salinity series of Ethiopian inland waters. Hydrobiologia 158:29–67
Ilkal area, Bagalkot District, Karnataka, India. Hydrogeol J Yohannes YB, Ikenaka Y, Nakayama SMM, Saengtienchai A,
15:589–598 Watanabe K, Ishizuka M (2013) Organochlorine pesticides and
Vallet-Coulomb C, Legesse D, Gasse F, Travi Y, Chernet T (2001) heavy metals in fish from Lake Awassa, Ethiopia: Insights from
Lake evaporation estimates in tropical Africa (Lake Ziway, stable isotope analysis. Chemosphere 91:857–863
Ethiopia). J Hydrol 245:1–18 Zinabu GM, Chapman LJ, Chapman CA (2002) Conductivity as
Von Damm KL, Edmond JM (1984) Reverse weathering in the closed- predictor of a total cations and salinity in Ethiopian lakes and rivers:
basin lakes of the Ethiopian Rift. Am J Sci 284:835–862 revisiting earlier models. Limnologica 32:21–26
Wagesho N, Jain M, Goel N (2013) Effect of climate change on runoff Zuppi GM, Sacchi E (2004) Hydrogeology as a climate recorder:
generation: application to rift valley lakes basin of Ethiopia. Sahara-Sahel (North Africa) and the Po Plain (Northern Italy).
J Hydrol Eng 18:1048–1063 Global Planet Change 40:79–91
The Landscape and Landforms of the Ogaden,
Southeast Ethiopia 19
Daniel Mège, Peter Purcell, Stéphane Pochat, and Thomas Guidat
Dedication This chapter is dedicated to the memory of hydrogeologist

Dr. Costantino Faillace, who worked extensively on the basalt
aquifers of Somalia and India, among many other countries, in a life
dedicated to providing clean water supplies to remote communities.
He died in 2012.
Abstract
The Ogaden region is located on the Somali Plateau, in southeast Ethiopia. Originally a clan-
based term, the Ogaden is now commonly used for the entire region below about 1,500 m a.s.
l., an area of some 300,000 km2 that encompasses most of the Somali Regional State and
includes the southwest portion of Oromia. The climate is hot, arid to semiarid, corresponding
to the Ethiopian bereha and kolla climatic zones. Three basic physiographic provinces are
recognized: the Genale and Shebele drainage basins and the Eastern Slope and Plains. The two
drainage basins include spectacular upstream canyons that witness the vertical movements that
have accompanied the succession of rifting events in the Ethiopian Rift, Afar, and the Gulf of
Aden. In strong contrast, the Eastern Slopes and Plains is dipping less than 0.4° on average
over hundreds of kilometers to the southeast and is mantled by red sands. Several remarkable
Ogaden landforms are described and analyzed, including volcanic, fluvial, and gravitational
features, some having few equivalents in other areas on Earth. A variety of volcanic landforms
are present across the region, reflecting the complex Cenozoic history of the Ogaden’s
margins. For instance, meandering basalt hills provide a textbook example of inverted
topography by fossilizing paleodrainage networks of various ages. The modern drainage
network provides information on the genesis of the mega-geomorphology of the Ogaden and
documents its uplift history. In western Ogaden, the deep incision has exposed the Cretaceous
evaporites and triggered the development of one of the largest gravitational spreading domains
on Earth, the Audo Range.
D. Mège (&)
WROONA Research Group, Institute of Geological Sciences,
Research Centre in Wrocław, Polish Academy of Sciences, ul.
Podwale 75, 50-449 Wrocław, Poland
e-mail: daniel.mege@univ-nantes.fr
D. Mège S. Pochat T. Guidat
Université de Nantes, CNRS UMR6112, , Laboratoire
dePlanétologie et Géodynamique de Nantes, Nantes, France
P. Purcell
P&R Geological Consultants, 141 Hastings Street, Scarborough,
WA 6019, Australia
D. Mège S. Pochat
Observatoire des Sciences de l’Univers Nantes Atlantique
(OSUNA, CNRS UMS 3281), Nantes, France
324 D. Mège et al.

Keywords

Ogaden Wabe Shebele Audo Range Marda Range Basalt Topographic inversion
Rift flank uplift Incision rate
19.1 Introduction rocks occur on or near the surface, underlain by a deep

sedimentary basin.
The landscape and landforms of Ethiopia, and the peoples The modern use of the Ogaden name is opposed by some
and cultures that evolved there, have captured the imagina- people in the region because they feel it implies Ogadeni
tion since biblical times. This dramatic topography is the political leadership, both in general and specifically over
creation of earth forces uplifting and splitting apart the Afro- non-Ogadeni lands (Hagman 2007). However, there is no
Arabian continent for more than 30 Myr. A vast rift complex useful alternative term and it is now widely used and
now cleaves south through the uplifted dome of Ethiopia, accepted in many fields of study.
separating western (Ethiopian) and southeastern (Somali)
plateaus, and creating a landscape that has profoundly
influenced the history and culture of the region. The high 19.2 Landscapes
and dissected western plateau, long known as Abyssinia,
sheltered the Coptic Christian kingdom from Islamic and 19.2.1 Human Geography
European expansionism. Conversely, the southeastern pla-
teau’s easy access to the Islamic coast saw early conversion The Ogaden occupies the southern portion of the Somali
of Sidama, Oromo, and Somali clans and the establishment Regional State of Ethiopia, and the southeastern portion of
of Harar in the thirteenth century as the religious center of the Oromia State. Jijiga is now the regional capital of the
the Horn of Africa. Somali State and the administrative center for the region.
In this chapter, we describe the landscape and landforms Adama (or Nazret) is the capital of Oromia. Figure 19.1
of the lower regions of the southeastern plateau, the area shows the Ogaden region, as described in this paper, with the
known informally as the Ogaden (Fig. 19.1). This poorly main towns, rivers, and roads. Important towns include
accessible and seldom visited region is not as well known or Aware (Teferi Ber), Daghabur, Gode, Kebri Dehar, Fik, Filtu,
studied as the Ethiopian Plateau and is often dismissed as and Werder (Harar, the historical capital of the region, is now
‘desert.’ Certainly, the region has vast red soil plains, but it designated as a separate state, as is Dire Dawa). The Jijiga/
also has the grand canyons of the Wabe Shebele and its Ferfer road coincides with a major topographic and geolog-
tributary rivers, the spectacular Audo Range gravitational ical divide between eastern and western Ogaden regions.
spreading complex, and a variety of landforms, soils, and The Ogaden region encompasses about 300,000 km2 and
vegetation that reflect the changing elevation of the plateau. has a population of the order of 5 million, predominantly
We summarize the geographical and geological information Somali and Oromo clans, with each constituting near 90 %
and then focus on three major features of the Ogaden of the population in their state (Ethiopian Government Portal
landscape: the volcanic landforms, the drainage system and 2014). Historically, the people of the Ogaden have been
its evolution, and the very large-scale gravitational spreading primarily nomadic pastoralists, but the population is
structures of the Audo Range. These landforms are described increasingly gathered into the large and small towns across
and also discussed from a mega-geomorphology perspective. the region. Agriculture predominates at higher elevations
There is no formal definition of ‘the Ogaden’ but, in and along the lower Wabe Shebele.
popular usage, it corresponds approximately to the area of the Major roads provide access to the Ogaden from the
Somali Plateau from about 1,500 m elevation to the Somalia national capital Addis Ababa (Fig. 19.1). In the north, where
and Kenya borders. The Ogaden takes its name from the numerous major river canyons occur, secondary roads
Ogaden clan (Somali, Ogaadeen), a subclan of the Darod, extend south from the main road to Jijiga. Access to
who live in the region. The term originally referred only to southwestern areas is limited from the north by the Wabe
the Ogadeni lands but is now commonly used for the entire Shebele canyon and is primarily via roads linking to the Imi-
Somali region, even, albeit rather imprecisely, as an alter- Gode-Mustahil road. A main road from Jijiga southeast
native for the Somali Administrative Region (Hagmann and through Kebri Dehar to Ferfer on the Somalia border links
Khalif 2006; Temin 2006). Climatically, it corresponds to the with secondary roads in the central and eastern region and
hot arid region of the Ethiopian bereha and kolla climatic provides the road link to Somalia capital, Mogadishu. Sec-
zones. In geological terms, it is the area where sedimentary ondary roads in the Ogaden, as well as some of the main
19 The Landscape and Landforms of the Ogaden, Southeast Ethiopia 325
Fig. 19.1 Location map, Ogaden region, Ethiopia
roads, are unsealed and can be in poor condition, with river to a broad flood plain. Pronounced NW-trending features,
crossing very difficult, especially in the wet seasons. such as the Marda Range complex, are clearly seen.
By contrast, the eastern Ogaden is a very gentle slope, on
average dipping less than 0.4° to the southeast, with no
19.2.2 Physical Geography major rivers or high topographic features. Comprehensive
geomorphological studies remain to be conducted; never-
The digital elevation model of southeastern Ethiopia of theless, an analysis of the most prominent landforms, the
Fig. 19.2 shows the main features of the Ogaden landscape scattered low basaltic hills, has been undertaken and is dis-
and the clear subdivision into highly dissected western and cussed in Sect. 19.3.2. Rock outcrops are relatively rare and
subdued eastern regions. The boundary is relatively abrupt, the plain is covered with alluvial and eolian red sand,
along the eastern side of the uplifted Marda Range, a commonly a few meters thick but locally as deep as 13 m,
prominent NW/SE-trending structure thought to have based on the results of oil and water bores (Walsh 1976;
formed by Phanerozoic multiphase reactivation of a Pre- Faillace 1993).
cambrian shear zone (Purcell 1976). In the western Ogaden, The main physiographic provinces and features of the
the high plateau rises to the rift margin rim and is dissected Ogaden are shown in Fig. 19.2. The basic subdivision rec-
by the steep canyons of south- and east-draining rivers, ognizes the main provinces of the Wabe Shebele and Genale
creating a rugged topography. The steep river canyons and watersheds and the Eastern Slopes and Plains. The further
profiles are evidence of the recent relative uplift along the subdivision into regions is clearly defined in some areas,
plateau rim. These rivers join the Wabe Shebele, which such as the Audo Range, but in other cases, the boundaries
flows southeast across the central Ogaden into Somalia, its between regions are diffuse. Figure 19.3 illustrates typical
valley slowly diminishing from a steep-walled high canyon landscapes from several of the physiographic regions,
326 D. Mège et al.
Fig. 19.2 Digital elevation model, southeast Ethiopia, showing informal physiographic subdivisions
including the Mustahil Plateau, the Eastern Slopes and Ethiopian-Yemenite province, which peaked at 30–29 Ma
Plains, and the Wabe Shebele valley. (Hofmann et al. 1997), was accompanied by vertical dis-
placements in the region, though whether and where it
19.2.2.1 Elevation resulted in uplift or subsidence depends on an interplay of
The Somali Plateau has its highest elevation along the parameters (Olson 1994), details of which are limited by the
uplifted faulted margins of the Afar Depression and the Main scarcity of adequate geological observations (see, however,
Ethiopian Rift (MER), reaching over 4,000 m a.s.l. in the Juch 1975). On the northern side of the Ogaden, rifting of
Arsi and Bale mountains in the west and over 3,000 m a.s.l. the Gulf of Aden (Leroy et al. 2012) started as early as Late
in the Chercher and Ahmar mountains in the north/north- Eocene, while uplift peaked at 20–18 Ma and stopped
west. Except in the west, the areas above 3,000 m in ele- around 16 Ma, when oceanization started (Watchorn et al.
vation occur only relatively near the plateau edge and, in all 1998; Fournier et al. 2010). Northward propagation of the
areas, are associated with volcanoes or thick Tertiary basalt Main Ethiopian Rift and initiation of the southern Afar
flows. The elevation declines to the south and east and is started at 11–10 Ma (Wolfenden et al. 2004; Bonini et al.
about 300 m along the Somalia border (Fig. 19.4). 2005), and rifting has proceeded until the present. Rift-flank
The present elevation of the Somali Plateau is the result uplift at the western edge of the Ogaden is thought to have
of geodynamic events that produced vertical movements of occurred during this interval. The overall southeastward
the crust over the Cenozoic. The Ogaden is located at the slope of the Ogaden seems primarily a consequence of the
edge of a dynamic mantle plume, upwelling for more than topographic evolution of this tectonism along the north-
30 million years in eastern Africa and western Arabia western edge of the Somalia Plate, although it does appear
(Moucha and Forte 2011). Flood lava eruption in the that the region had a preexisting southward slope dating to
Fig. 19.3 Selected Ogaden landscapes. a Somali state capital Jijiga, of Wabe/Ramis junction; e Mustahil Plateau, south-central Ogaden,
looking southwest toward Marda Range; b Red sand plains in eastern looking across Wabe Shebele floodplain; and f Genale Plateau on
Ogaden, with village in foreground; c Wabe Shebele near Gode; Gabredarre limestone above Genale River (Photos by P. Purcell except
d Canyons of the Wabe Shebele incising Uarandab Shale plateau, south e which is courtesy of Hunt Oil Company)
the Late Jurassic, as revealed by the thinning and transition elevation, where annual rainfall is less than 400 mm, result-
to less marine facies of the Upper Jurassic sediments out- ing in sparse vegetation with extensive bare ground. The
cropping in the northern and western areas (Purcell 1981). kolla zone (warm to hot semiarid) covers the region between
500 and 1,500 m in elevation, where the average annual
19.2.2.2 Climate rainfall is generally around 600–800 mm. The higher regions
The climatic zones in the Ogaden and adjacent region are of the Somali Plateau are classified as weyna dega (1,500–
shown using traditional Amharic terminology on Fig. 19.5 2,500 m a.s.l.; warm to cool semihumid) and dega (>2,500 m
(Ethiopian Mapping Authority—EMA 1988; Lemma 1996). a.s.l.; cool to cool humid). Under the Köppen classification,
The bereha zone (hot arid) covers the region below 500 m the Ogaden is classified as hot arid (Bwh) and hot semiarid
328 D. Mège et al.
Fig. 19.4 Simplified elevation

map, southeast Ethiopia
Fig. 19.5 Climate zones,

southeast Ethiopia (after Lema
1996; EMA 1988)
(Bsh). In the high country above 2,000 m (beyond the Oga- blow across the region are relatively dry and carry little rain
den sensu stricto), the climate is warm temperate (Cwb). For and, secondly, that the Ogaden is in the rain shadow of the
all intents and purposes, Koppen zones Bwh and Bsh cor- Ethiopian highlands as regards Atlantic moisture carried east
respond to the bereha and kolla zones. by the unstable Congo airstream (Nicholson 1996; Sepulchre
Ogaden aridification dates back to Late Miocene, when, et al. 2006). The rainfall over most of the Ogaden falls mainly
after the Eocene optimum, the Sahara desert formed (Micheels during the transition period between the monsoons, giving a
et al. 2009) and the Messinian salinity crisis in the Mediter- bimodal rainfall distribution (EMA 1988), with rainy seasons
ranean region occurred (Feakins 2013). Aridity has been in March–June (main) and September–November (minor),
globally maintained until modern times. The main causes in known in Somali as Dehr and Gu, respectively.
the Ogaden are, firstly, that the NE and the SE monsoons that
Fig. 19.6 Drainage basins,

southeast Ethiopia (after EMA
1988)
Annual rainfall is between 400 and 800 mm over much of Gerer (or Yerer), flows southeast along the Marda Range and
the Ogaden (refer to bereha and kolla zones on Fig. 19.5) then southward, drying into the desert, except in heavy wet
but is very irregular and variable in the extreme east and seasons when it flows into the Wabe Shebele.
southern areas, often 200 mm or less. Annual variability can The Genale catchment area covers about 168,000 km2 and
be considerable within and between local areas, mainly contains three main rivers, the Genale, Weyb (or Webi Ges-
linked to variability in Gu rains, but recent analyses show no tro), and Dawa, all of which meet near Dolo on the Somalia
overall decline in annual rainfall in the period 1965–2002 border and continue south as the Juba River. The topographic
(Cheung and Senay 2008). divide between the Wabe Shebele and Genale watersheds
runs through the Audo Range, an alignment of gravitationally
19.2.2.3 Hydrology unstable mesas, further discussed in Sect. 19.3.4.
The Ogaden region contains two vast drainage systems, the In terms of landscape evolution, the rainfall regime is
Wabe Shebele basin in the northeast and the Genale basin in responsible for the development of temporary streams and
the southwest, as shown in Fig. 19.6 (EMA 1988). Both small river channels throughout the Ogaden lowlands. They
systems drain into the Indian Ocean. are dry most of the time, but are reactivated during the short
The Wabe Shebele (Somali, River of Leopards), known as and intense rainfalls, washing out the surface of the hills and
the ‘Second Nile’ to early Arabian geographers, is over accumulating the erosional products downstream in pan-type
1,300 km long in Ethiopia (about 2,000 km overall) and has a depressions. Such depositional areas are easily identified
catchment area of over 205,000 km2. The river rises in the downstream from basaltic hills, because the dark clay
Arsi highlands in the west and flows initially northeast in accumulations contrast with the surrounding red sands. In
spectacular deep canyons, reaching over 900 m deep near the the vast Wabe Shebele basin, intense rainfall in the upper
ancient town of Sheik Hussein, before abruptly swinging catchment produces flash floods that periodically cause loss
southeast and meandering across the Ogaden into Somalia. of life and destruction of buildings, crops, and livestock, by
Near Mogadishu, it is deflected southwestward by coastal the strength of the flow of the crocodile-infested waters. At
dunes and, in the wet season at least, joins the Juba River and the same time, these waters and the soil they carry play a
enters the ocean near Kisamayo. The main tributaries of the considerable role in the local agro-pastoral economy by
Wabe Shebele are the Galeti, Ramis, Erer, and Dakota rivers, making possible flood-recession agriculture in the riverbed
rising in the north in the Ahmar Mountains and cutting deep in the Gu and Dehr seasons (UNDP 1999). Other major
narrow gorges in the northern slopes of the plateau, as dis- rivers, such as the Fafan and Gerer, are subject to intense
cussed further in Sect. 19.3.3. The easternmost major river in floods lasting hours or days (Bauduin et al. 1973) but of
the region is the Fafan, which, with its main tributary, the much lower dramatic consequences than the Shebele floods.
330 D. Mège et al.
19.2.2.4 Vegetation and Soils from desert vegetation to steppe to grassland to woodland
The changing elevation of the Ogaden region and its rim of and savannah and, finally, in the high country to coniferous
highlands, and the rainfall pattern they influence have a forests. Xeromorphic thorny plants and grasses, often salt
marked impact on the pattern of soils and vegetation across resistant species, and low acacias dominate the desert and
the Ogaden landscape. semidesert scrubland in the southeast. In the steppes,
Fluviosols are present along the lower reaches of several between 200 and 1,400 m a.s.l., the vegetation is similar but
rivers, notably the Wabe Shebele, and can support large with larger and more dense growth. The grasslands in the
irrigated agricultural programs. In many valleys, however, Ogaden, covered by short to medium, fire-resistant grasses,
the alternating wet periods and long dry spells have created occur between about 400 and 2,000 m a.s.l. and interfinger
heavy clay-rich vertisols that are less useful for agriculture. with the savannahs and woodlands that are dominated by
Over most of the Ogaden, the predominant soils are xerosols acacias and juniper trees (EMA 1988).
and yermosols, the latter often gypsiferous, and generally not
suitable for agriculture. The soils are alluvial and vulnerable
to wind and water erosion (EMA 1988). 19.3 Remarkable Landforms
Botanical description of the Ogaden flora is still very
limited, as illustrated by the recent discovery of probably 19.3.1 Geological Background
hundreds of thousands of specimens of a newly described
acacia species, Acacia fumosa, on the Cretaceous Mustahil The landscape and landforms of the Ogaden are a product of
limestone hills in eastern Ogaden (Thulin 2007). The basic the stratigraphic and structural evolution of the region. The
vegetation pattern reflects the soils and moisture conditions. surface today consists of sediments, Mesozoic to Neogene in
While a complex interfingering of zones is determined by age, as well as Precambrian basement rocks and Cenozoic
local conditions, there is an overall change from SE to NW, volcanics (Fig. 19.7). Jurassic and Cretaceous sediments are
Fig. 19.7 Geology map, southeast Ethiopia, showing location of volcanic features discussed in text
the predominant outcrop in the western Ogaden, while basalts covers a time span of *8 million years, between 30
Cenozoic sediments occur in the east; the contact between and 22 Ma (Mège et al. 2012a), encompassing the onset of
the two provinces being roughly coincident with the Marda rifting in the Red Sea and Gulf of Aden. Excluding the
Fault Zone. The underlying Ogaden Basin contains about 2,000 m or more of accumulations at major volcanic centers
8 km of sediments deposited from Permian to Eocene time, along the plateau rim, basalt and associated volcanics on the
approximately 260–35 million years ago. plateau proper are 200–500 m thick in the north and west
The Ogaden Basin formed initially as part of the Karoo and more commonly 50 m or less in southern and eastern
rifting in the Permian (Purcell 1981). In the Late Triassic to areas. While not comparable volumetrically with the Ethio-
Early Jurassic, it developed into a sag basin and marine pian Plateau volcanics, the Ogaden basalt flows were
waters from the newly opening Indian Ocean flooded across extensive, reaching over 600 km into coastal Somalia, as
the region, extending to northern Ethiopia, and depositing revealed by the widely scattered remnants seen on the sur-
some 1,500 m or more of predominantly carbonates and face and identified in the subsurface by aeromagnetic sur-
evaporates in a vast shallow sea. In the Oxfordian, the seas veys and intersections in water bores and oil exploration
reached their maximum flooding level and the Uarandab wells (Purcell et al. 2011). Low viscosity lavas, that flowed
Shale developed. It was overlain, as the seas shallowed, by for 100–200 km or more, filled river canyons and are pre-
the carbonate Gabredarre Formation. The alternating cliff- served today as the inverted topography of basalt hills
forming limestones and softer shale and evaporite units meandering across the landscape. The volcanic features of
result in a cliff-and-terrace topography in the eroded valleys the Ogaden are discussed in detail in Sects. 19.3.2 and
in the northern and western Ogaden. 19.3.3. A major feature of the Ogaden magmatism was the
In the Early Cretaceous, silling of the basin, probably by Late Oligocene Ogaden Dyke Swarm, which can be traced
a barrier reef trend near the present coastline, established a across the Ogaden on outcrop and magnetic data from the
broad sabhka environment over the Ogaden, where the thick Afar margin to the Somalia border (Mège et al. in press).
evaporites (1,600 m) of the Gorrahei Formation (previously, The tectonic activity at the northern and western edge of
the Main Gypsum) were deposited in the basin center, the Somalia plate since the Miocene, the volcanic activity
thinning to the north and west. A marine transgression in the since the Oligocene, and contrasts in rheology and strength
Aptian reestablished open marine conditions and deposited of the Meso-Cenozoic cover are the main geological influ-
the dense limestones of the Mustahil Formation. Regional ences on the geomorphological evolution of the Ogaden. In
uplift in the Santonian exhumed the western Ogaden and particular, they are key to understanding the three prominent
reoriented the eastern basin to a NE-trend along the newly landforms of the Ogaden discussed below: the volcanic
formed continental margin (Purcell 1981). During the Ma- landforms, the exhumed canyon-filling lava flows and the
astrichtian–Paleocene, the continental Yesomma Formation gravitational spreading structures.
was deposited over much of the Ogaden, becoming more
marine in the east. The Mustahil and Yesomma formations 19.3.2 Volcanic Landforms
were to collapse later into the underlying Gorrahei evapor-
ites by block tectonics of gravitational origin, as discussed The geomorphology of the volcanic outcrops in southeast
further in Sects. 19.3.4 and 19.4. Ethiopia have been discussed by Purcell et al. (2011), based
An Eocene transgression deposited the carbonates and primarily on mapping of the entire region on Landsat Geo-
evaporities of the Auradu, Karkar, and Taleh formations in cover mosaics 1990 and 2000, high-resolution satellite ima-
the eastern Ogaden and, thereafter, sedimentation was lim- ges available on Google EarthTM, and helicopter-supported
ited to the coastal Somalia area. Further widespread erosion field work in the eastern and southern Ogaden. Volcanics in
occurred in the Oligocene, driven by uplift of the Somali southeast Ethiopia occur as broad volcanic plateaus, linear
Plateau and creating a landscape of eroded Tertiary, Creta- outcrops, isolated hill complexes, and meandering ribbons of
ceous, and Jurassic sediments, similar in general terms to the exhumed paleocanyon basalt fill. The locations of various
present surface. outcrops discussed in this chapter are shown in Fig. 19.7.
The early plume volcanism along what is now the Somali Selected examples of Ogaden volcanic landforms are shown
Plateau margin is relatively undocumented, but probably in Fig. 19.8. It is interesting to note that the volcanics have
commenced during the Oligocene around 30 Ma, as on the locally influenced settlement patterns in the region. The
Ethiopian Plateau (Hofmann et al. 1997), since the two rolling basalt hills at Fik (Fig. 19.8b), for example, provide a
regions were not then separated by the Afar and the Main natural shelter for the town. At nearby Daga Medo, the town
Ethiopian Rift (MER). In the eastern Ogaden, the 40Ar/39Ar has been built atop a recent flow, protecting residents from
age of volcanic rocks geochemically akin to the plume flash flooding during occasional heavy rains.
332 D. Mège et al.
Fig. 19.8 Selected volcanic landforms of the Ogaden. a Ilbah Hills, a Shebele and flanking volcanic hills near Gode. North to the upper right;
dyke-related volcanic complex in east central Ogaden; b volcanic hills e Arid land farming around volcanic cones, east of Jijiga near the
near Fik town; c part of a Landsat ETM+ image (Geocover 2000) of Somaliland border; and f view southeast down the Marda Range from
basalt hills formed by inversion of canyon-filling lava flows in the quarry in hillside above the Marda Pass (Photographs by P. Purcell)
ancestral Shebele river. Arrow shows direction of view in (d); d Wabe
19.3.2.1 The Main Ethiopian Rift Shoulder reevaluation. Dating (40Ar/39Ar) of a sample from the Gara
The most extensive basalt features occur in the west as Mulata area was attempted by the authors; a well-defined but
eroded plateaus flanking the upper Wabe Shebele depressed plateau indicates obvious argon loss, which
(Fig. 19.7). These plateaus are considered to be remnants of compromises the age determination; nevertheless, an
flows that originally extended continuously from the rift emplacement age for the first basalts >24 Ma can be inferred.
margin far into the Ogaden. The ages of the basalts are not The Pliocene (2.2 Ma) age obtained for the overlying Ginir
well constrained. The older Trap volcanics overlying Formation rhyolites (Merla et al. 1973) also needs reevalu-
Jurassic sediments were dated in the early 1970s using K–Ar ation. The columnar-jointed Ginir rhyolite, rising sharply
as Oligocene to Miocene (Megrue et al. 1972) and need above the eroded slope of the Garbaharre sandstones, forms
a prominent north-facing cliff known as the Bilka Ridge East of the Marda Range, several complexes of basaltic
(Kibrie and Yirga 2008). The multilayered flows are com- hills rise prominently above the red sand plains (Fig. 19.8a).
monly over 100 m thick and reach a maximum of 365 m and Volcanism has been dated using the 40Ar/39Ar method,
are a prominent landform of the plateau adjacent to the Wabe evidencing a 30–24 Ma age range (Mège et al. in press). The
Shebele canyon. hills are generally broad mounds covered with rounded
basalt cobbles, and actual outcrops are few and scattered.
19.3.2.2 The Genale and Dolo Basaltic Tablelands The topographic linearity of the hills suggest the presence of
Several large, undated basaltic tables occur in southwest emergent dykes, but the poor quality of the outcrops makes
Ogaden and adjacent Somalia. Lava flows originating close identification difficult, except at Ilbah Hills where a dyke has
to the Main Ethiopian Rift shoulder flowed down the Uelmal been confirmed from its chilled margins (Mège et al., in
River, a tributary of the Genale River, and accumulated press). Satellite images and SRTM topography reveal
downstream on the Genale valley floor east of Filtu. They widespread NW/SE-trending lineaments associated with
are now manifest as two basaltic tables, both 150 km2 in several volcanic hills and confirmed by field studies to be
surface area (Fig. 19.7). The northern table stands 100– shallow depressions, 100–300 m wide and a few meters
150 m above the Genale River and is dissected into mesas by deep, marked by contrasting vegetation density and occa-
numerous antecedent rivers. About 20 km downstream, the sional calcrete exposures, but without any surface evidence
southern table culminates 300 m above the Genale River. of associated volcanic activity. Many are coincident with
Nearly 100 km to the southeast, the Dolo basaltic table is high-frequency linear magnetic anomalies (Purcell et al.
100–150 m high and covers some 6,500 km2. Basaltic out- 2011, Mège et al. in press) and are considered to be the
liers, now separated by rivers, testify to an original much product of hydrothermal alteration and limestone or gypsum
broader extent. The Dolo basalt table forms a cliff above the karst development along dykes and dyke-parallel fissures
Weyb river on its western side, while on the eastern side its (Mège et al. 2012b).
edge is masked by extensive Quaternary deposits and its true
extent is not known; aeromagnetic data suggest that it could
extend more or less continuously below the Quaternary 19.3.3 Inverted Basalt-Filled Paleo River
mantle to the Bulo Burti basalt outcrop in Somalia, and then Channels
as far as the Indian Ocean (Bosellini 1989; Purcell et al.
2011). In the south, the geometry of the present outcrops A prominent feature of Ogaden and Somali volcanism is the
suggests that the Dolo basaltic table could have fed the presence of long (>100 km) basaltic flows in paleo river
meandering basalts that filled the Juba paleo river channel, channels. In the central western Ogaden, an elegant chain of
as discussed in Sect. 19.3.3. meandering hills extends south for 120 km from the Daga
Medo flow. No recent age dating has been done on this flow;
19.3.2.3 The Marda and Jijiga Volcanics however, Maxus Ethiopia (1993) obtained 27.4 ± 1.4 Ma
The best known volcanic outcrop in the region is the Marda using the K-Ar method (Table 19.1). Another flow, albeit
Range, a chain of uplifted and eroded hills of Jurassic more eroded, can be seen northwest of Mustahil. It is 90 km
limestone, capped by a linear basaltic layer, 150 km long and long, but with eroded ‘gaps’ between the basaltic remnants
up to 2 km wide, trending SSE from Afar to the central and originally flowed in a gently meandering succession of
Ogaden, parallel to the first order drainage pattern narrow valleys in the Cretaceous Mustahil limestones.
(Fig. 19.8f). In tectonic terms, it is referred to as the Marda The Gode lava flow (Fig. 19.8c, d) can be traced for more
Fault Zone and described as a Precambrian mylonite zone than 200 km and appears to have flowed into the large
that has been reactivated several times during the Phanero- meandering channel of the paleo-Wabe Shebele. The Gode
zoic (Purcell 1976; Boccaleti et al. 1991). In the north, near and Mustahil flows were sampled during the 2008 helicopter
Jijiga, the Marda basaltic layer is about 200 m thick and survey. Locally, this flow appears to overflow the main
consists of a number of thick, columnar basaltic flows channel to form overbank basaltic pods. The Juba River in
(Fig. 19.8f). There is no evidence of a feeder dyke within the Somalia has meandering basaltic outcrops, following a
limited outcrop area, but recently cut quarries nearby reveal paleo-Juba River channel (Abdirahim et al. 1993; Ali Kas-
numerous hypovolcanic basaltic intrusions, ranging to over sim et al. 1987, 1993), which are geomorphologically very
10 m in width and from which narrow dykes (<1 m) have similar to the Gode flow outcrops. Initially thought Oligo-
propagated upward. The rare earth element spectra for the cene, the geomorphology of both the Juba and Wabe
intrusives and the Marda extrusives are similar enough to Shebele basalt outcrops argues in favor of a much younger
suggest the same parental magma, in agreement with a age of emplacement. A 7.5 ± 0.4 Ma 40Ar-39Ar age has been
similar 40Ar/39Ar age of 23–25 Ma (Mège et al. in press). determined for the Wabe Shebele basalts (Table 19.1).
334 D. Mège et al.
Table 19.1 Location and age of the basalt samples used in this work
Volcanic site Sample ID Type Coordinates (WGS84) Age Method Material References
Latitude N Longitude E (Ma ± 2σ)
Marda (north) MA03 Flow 9°21′52.13″ 42°41′57.74″ 23.68 ± 0.54 40
Ar/39Ar Groundmass Mège et al. (in press)
Marda (north) MQR7 Dyke 9°21′48.51″ 42°42′7.68″ 25.04 ± 0.65 40
Ar/39Ar Groundmass Mège et al. (in press)
Marda (south) L-2-3 Flow 8°43′50″ 43°4′20″ 25.5 ± 1.3 K–Ar Whole rock Maxus Ethiopia (1993)
Daga Medo K-2-1 Flow 8°25′40″ 42°54′30″ 27.4 ± 1.4 K–Ar Whole rock Maxus Ethiopia (1993)
Mustahil 20 Flow 5°32′22.12″ 44°35′37.80″ 28.09 ± 0.81 40
Ar/39Ar Whole rock Mège et al. (in press)
Fik A-3-1 Flow 8°12′ 42°40″ 28.4 ± 1.4 K–Ar Whole rock Maxus Ethiopia (1993)
Kebri Dehar K3.2 Flow 6°49′38.00″ 44°56′49.45″ 27.64 ± 0.40 40 39
Ar/ Ar Whole rock Unpublished*
Kebri Dehar K1.6 Flow 6°45′18.34″ 44°26′3.39″ 27.61 ± 0.59 40
Ar/39Ar Whole rock Unpublished*
Gode WS Flow 6° 1′12.79″ 43°21′55.28″ 7.46 ± 0.47 40 39
Ar/ Ar Groundmass Mège et al. (in press)
*The ages yet unpublished have been obtained from this chapter’s team of contributors. Sample K3.2 has a 78 % plateau age with MSWD (mean
square of weighted deviates) 0.61, and sample K1.6 an inverse isochron with intercept 289.3 ± 8.6 and MSWD 1.15
Fig. 19.9 Simplified stratigraphy

of the northern Audo Range
(Mège et al. 2013)
The current outcrops of the Gode and Juba basalts sug- (Fig. 19.9). It overlies the Kimmeridgian–Portlandian Gab-
gest an original total lava flow length of several hundreds of redarre Formation inner shelf limestones, which provides a
kilometers, similar to the length of lava flows in other flood rigid basement on which the spreading has occurred: the
basalt provinces, for example, the Columbia River flood Mustahil Formation (limestone) and the Yesomma Sand-
basalts (Tolan et al. 1989). Surprisingly, however, and stone overlying the Gorrahei Formation have spread spec-
contrary to the other very long basaltic flows, which appear tacularly by block faulting and tilting (Figs. 19.10 and
to have formed as part of the main flood basalt emplacement 19.11), forming one of the world’s largest gravitational
event, the age of the Gode flow very significantly postdates spreading domains. Presently covering about 5,000 km2, the
the age of the main Ethiopian flood basalt events. The Juba Ogaden complex is 25 times larger than the Canyonlands
basalts, which have a similar morphology and similar rela- grabens area in Utah, another spectacular continental-
tionships with the Juba river, might also be of Upper Mio- spreading domain.
cene age, as might the Dolo volcanic table, given the good The spreading zone extends from the Weyb River to the
geomorphological evidence that the Juba basalts flowed Wabe Shebele (Fig. 19.10a) and is mainly manifest in the Audo
from it. Range, an element of the water divide between the Genale and
Wabe Shebele basins. The Audo Range is markedly asym-
metric, with the most dramatic topographic relief and best-
19.3.4 Gravitational Spreading preserved gravitational features along the Weyb River, and
Structures more eroded and attenuated landscapes on the Wabe Shebele
side. An unnamed ridge north of the Audo Range also displays
19.3.4.1 Manifestation and Geographical Extent evidence of normal faulting, producing tilting of post-Gorrahei
The Portlandian–Neocomian Gorrahei Formation outcrops sedimentary layers up to 25° (Figs. 19.10a and 19.11). East of
extensively in western Ogaden, where it is dominantly the Wabe Shebele, the Goja mesa displays tilted blocks boun-
evaporites alternating with shales and over 150 m thick ded by downhill- and uphill-facing normal fault scarps on the
336 D. Mège et al.
◂ Fig. 19.10 a Geomorphological map of the Audo Range, based on controlling stream or river. Dip angle is given with an accuracy of 1°,
field work at the Kebenawa Ridge and along the Imi–el Kere road except for dip angles in nearly horizontal areas when a higher precision
complemented by Landsat ETM+ images and high-resolution satellite was thought to be significant. b Geomorphological map of the Chebis
images available through Google Earth and Bing Maps. Block Ridge. In (a) and (b), error on dip angles is <0.1, except when given
displacement arrows have a color that reflects the color pattern of the
Fig. 19.10 (continued)
east and a huge rotational landslide on the west. Both the Goja 1963), the Yesomma Formation was correctly identified on
mesa and the unnamed ridge indicate that the initial extent of the the first edition of the Geological Map of Ethiopia (Kazmin
gravitational spreading zone may have been much larger than 1972) and a contemporary hydrogeological map of the
presently observed. Shebele drainage basin by ORSTOM (Bauduin et al. 1973).
However, these sediments were mapped subsequently as
19.3.4.2 Previous Work volcanics of unknown age by Merla et al. (1973) and,
The Audo Range and surrounding area have been little stud- reflecting this, Beicip’s (1985) Geological Map of Ogaden
ied, as illustrated by the confusion in the reported stratigraphy. reported them as either the Yesomma sandstone or ‘volcan-
The rocks overlying the Gorrahei Formation are the Aptian ics.’ This indecision was not reflected on the second edition of
Mustahil Formation limestones and the Maastrichtian– the Geological Map of Ethiopia (Tefera et al. 1996) where
Paleocene Yesomma Formation sandstones (Fig. 19.9), but they are shown as Eocene basalts (Ashangi Formation).
are shown in many documents as Tertiary basalts. Unrecog- Evidence of halokinetic deformation predating gravita-
nized on the earliest geological maps of Ethiopia, where it tional spreading was found at the Kebenawa Ridge in the
appears as Jurassic (‘Antalo’) limestone (Dainelli 1943; Mohr northern Audo Range (Mège et al. 2013). Gravitational
spreading—in the form of tilted blocks surrounding mesas— based on the same sedimentary criteria. Fracture lines in the
was documented at the Chebis Ridge (Lopez-Gonzalez SW corner of the Audo Range map (Fig. 19.10a) which are
2006), with faults dipping outward in most cases parallel to the Weyb River and some Genale River tributaries
(Fig. 19.10b). Mège et al. (2013) suggested that gravitational (see Fig. 19.6) may testify to recent moderate tectonic
spreading commenced after the Yesomma regression movement related to the opening of the Main Ethiopian Rift
(Bosellini 1989; Purcell 1981), when erosional incision of (Gani et al. 2009), possibly controlling the incision of the
the uplifted Ogaden surface reached the level of the Gorrahei Weyb River and triggering the more recent gravitational
evaporites. Topographic debuttressing and evaporite rheol- spreading on the western side of the Audo Range.
ogy then triggered fragmentation and spreading of the Dip angles were measured at 196 sites in the Audo Range
overlying formations, similar to the spreading of the Hona- area. The dip angle of the tilted layers is proportional to the
ker Trail Formation and the Cutler Group over the evaporite quantity of stretching (e.g., Angelier and Colletta 1983) and
Paradox Formation in the Needles district of the Canyon- provides additional information as to the spreading mecha-
lands Park in Utah (Mège et al. 2013). A significant differ- nisms. Dip angles were measured using the OrionTM struc-
ence, however, is that the Canyonlands topography was tural analysis software from Pangaea Scientific (Fueten et al.
incised by the Colorado River only on one side, leading to a 2005) applied to a co-registered SRTM90 digital elevation
single spreading direction, whereas the Audo Range is model and a Landsat ETM+ image (14.25–28.5 m/pixel).
wholly surrounded by rivers incising the evaporite forma- SRTM90 data (87–89 m/pixel) have an error of 5 m in
tion, leading to multidirectional spreading. Another differ- southeast Ethiopia (Farr et al. 2007). A best-fit plane was
ence is that spreading has promoted the development of half computed from multilinear regression from manually selected
graben tilted up to 25° in the Audo Range, whereas points of known horizontal coordinates and elevation, picked
spreading in the Canyonlands produced nearly symmetric out along a stratigraphic plane on the satellite image. The
grabens (Moore and Schultz 1999). validity and accuracy of the computed plane is controlled by a
series of statistics (see Fueten et al. 2005) and 3D visualiza-
19.3.4.3 Landform Development tion tools in OrionTM, and cross-checked with the high-res-
The direction of displacement of the tilted blocks in a olution (0.5–2.5 m/pixel) 2D and 3D views using Google
gravitational spreading complex provides important insights Earth and Bing Maps. Geologically speaking, each reported
into the development of the structure. Because gravitational dip angle corresponds to the mean dip angle of a given strata
spreading occurs in response to debuttressing of the evapo- of constant dip angle, averaged over a measurement site
rite layer, displacement is predictably perpendicular to the varying in length from a few to several hundred meters (7–30
orientation of the ‘free’ boundary, that is, to the side where measurements). Dip errors correspond to a 95 % confidence
the river has cut down through the gypsum layer. Compar- level and are reported only when >0.1°. Representative dip
ison between block displacement and the drainage system at angles are reported on Fig. 19.10a. Figure 19.10b gives all the
the Audo Range suggests that gravitational spreading has dip angles measured on the Chebis Ridge.
occurred in response to incision by tributaries of the Wabe Dip angles measured in the tilted layers of the Mustahil
Shebele and by the Weyb River, but not by the Wabe and Yesomma formations are in the range 0–26°. The
Shebele itself. A correlation between stream or river incision uppermost tilted blocks are connected to remnant Yesomma
and block displacement direction is proposed on Fig. 19.10a. plateau fragments, which are either horizontal (north of El
The gravitational spreading features are much better Kere) or more frequently, tilted by a few degrees (e.g.,
preserved on the western side of the Audo Range than on the Elewak Ridge, Chebis Ridge). Dip angles could not be
eastern side, where the tilted blocks are much more dissected measured in the Gorrahei Formation, due to the difficulty in
by the Wabe Shebele tributaries. This indicates that gravi- identifying continuous layers on the satellite imagery. Dip
tational spreading is older in the east and that topographic angles measured from the top of the Chebis Ride to its
debuttressing of the western side by the Weyb River western foot (Fig. 19.10b) increase from the top to the
occurred later. From the fluvial nature of the Cretaceous– bottom, suggesting that spreading proceeded inward.
Paleocene upper Yesomma Formation sandstones and the Dip angles in the underlying Gabredarre Formation are
paleodeltaic deposits downstream in southern Somalia, the usually close to zero and were never found to exceed 2–3°.
Shebele basin has been dated as old as the Cretaceous– In the southernmost area, dip angles were found to be con-
Paleocene in western Ogaden (Bosellini 1989). Hence, there sistently 0.4–0.7° to the east, accordant with the general
has been considerable time for erosion of the eastern side of 0.8°E dip angle calculated for the underlying Jurassic
the Audo Range. No evidence for such an age has been sequences from seismic reflection data (Line 93-GR-07E)
reported for the Weyb drainage basin, which could be a about 70 km south of Fig. 19.10a (Beicip-Franlab 1998).
recent component of the Juba Basin—though the Juba Basin These low dip angles show that gravity gliding along a
itself is thought to be of similar age to the Shebele Basin, regional structural slope (in the sense given by Schultz-Ela
338 D. Mège et al.
Fig. 19.11 Tilted block in the eastern Audo Range: a in the Yesomma sandstones; b in the Mustahil carbonates; c at the Chebis Ridge. Location
on Fig. 19.10
Fig. 19.12 Topographic profile

along the Audo range water
divide
2001) over the rigid pre-Cretaceous substrate did not play a surfaces throughout the Audo Range, plus 3 points on the
significant role in the development of the observed Goja mesa, shows they fit a single plane of strike
landscape. N003°E ± 016 and dip 0.4 ± 0.2° (at a 95 % confidence
It is significant to note that the various plateaus of sub- level), with a goodness of fit of 98.2 %. This remarkable fit
horizontal Yesomma sandstone that constitute the summit indicates that the many geodynamic events that occurred in
blocks of the Audo Range Yesomma are fragments of the the Cenozoic after Yesomma sandstone deposition, includ-
same dipping surface. A topographic profile along the water ing flood lava emplacement, and the rifting in the Gulf of
divide shows a linear decrease of plateau surface elevation Aden, Ethiopian Rift, and Afar, had very little influence on
from NNW to SSE, with an inclination of 0.27° (Fig. 19.12). deformation in western Ogaden, beyond the uplift of the
An analysis of thirty-one points on the summit plateau plateau itself.
19.4 Landform Evolution at Regional The general organization of the Ogaden basin shows
Scale large-scale structural and geodynamic control. The regional
geometry of the Genale and Wabe Shebele river systems
19.4.1 Adaptation of Rivers indicates two main directions of flow (Fig. 19.13). A dom-
to Topography inant N130E orientation is exemplified by the main orien-
tation of both the Genale and Shebele basins and by
Analysis of drainage systems provides valuable information tributaries such as the Gerer, Fafan, and Dawa rivers. A
regarding vertical motions of the surface at a regional scale second dominant orientation is N050-060E and is especially
(Snyder et al. 2000; Schumm et al. 2002; Duvall et al. 2004; exemplified by a 350-km-long segment of the upper Wabe
Whipple 2004; Whipple et al. 2013) and can help under- Shebele in the Arsi highlands. It is also observed in the
standing of the evolution of landforms in relation to geo- lower part of the Dakota river. The N130°E trend is the
dynamics. In this section, the development of the present orientation of the tilt of the Somali Plateau. The N050°–
drainage network and the current topography is analyzed. 060°E trend is the orientation of the Main Ethiopian Rift
Fig. 19.13 Drainage system

map, with the location of the
basalt outcrops in the Ogaden
(white surrounded with black
line). The rivers are in blue and
the watersheds in dark grey. The
basalt flows used in the
morphometric analysis are
named. The black arrows indicate
lava flow directions. Local mean
incision rates since basalt
outpouring are indicated in the
boxes
340 D. Mège et al.
Fig. 19.14 Longitudinal profiles

of rivers of the Wabe Shebele and
Genale systems discussed in the
text, and comparison with profiles
of rivers at equilibrium (Darling
river tributary, Australia); in
equilibrium except in localized
segments, where they are slightly
disturbed by tectonics or localized
abrasion (Rhine and Loire rivers,
respectively, Europe); and in
disequilibrium owing to active
tectonic uplift and resulting
strong fall of base level (Amazon
tributary, Andes). The profiles
were extracted from ASTER
GDEM with RiverTools®
(Peckham 1998). a Wabe Shebele
system; b Genale system. The
profiles are leveled at the
elevation value at the end of the
Wabe Shebele and Genale
axis. The faults guiding the rivers could originate from the the concavity index (Langbein 1964), given in Table 19.2 for
early stage of rift development, before strain concentrates on the studied rivers. This index is zero for a straight river
a limited number of fractures (e.g., Olson 1993). profile; −1 is for strongly unequilibrated (convex) profile; 1
The longitudinal profiles of nine riverbeds overlain by is for fully equilibrated (concave) profile. For the Wabe
basaltic flows were analyzed (Fig. 19.14). These rivers were Shebele, the concavity index cumulated over the whole
selected as representative of their location compared to the profile length (cumulative concavity index, CCI) is 0.83
main geologic features of the Ogaden. In the Wabe Shebele (Table 19.2). The Wabe Shebele is characterized over its
system (Fig. 19.14a), the Wabe Shebele does not show the upper half by a succession of vertical-step knick-points.
characteristic fully concave longitudinal profiles of rivers in Between elevations *2,600 and *2,300 m a.s.l. in the Arsi
equilibrium. Profile equilibrium is usefully described with highlands, it is characterized by an alternation of relatively
Table 19.2 Geomorphometric parameters for the rivers on Figs. 19.14 and 19.15 (graphical definition on Fig. 19.15)
River name hmax hmin l Concavity indexes h
(m) (m) (km) 25 % 50 % 75 % 100 %
Main Ethiopian rivers
Genale 2,825 50 1,315 D 2% 5% 23 % 100 %
CCI 0.04 0.27 0.48 0.87
ICI 0.04 0.3 −0.2 0.54
Wabe Shebele 2,685 165 1,505 D 8% 10 % 23 % 100 %
CCI −0.12 −0.49 0.42 0.83
ICI −0.12 −0.01 0.16 0.63
Tributaries
Uelmal 4,040 300 305 D 6% 13 % 10 % 100 %
CCI −0.42 −0.82 0.29 0.87
ICI −0.42 0.34 0.32 0.68
Dakota 2,735 740 250 D 0.8 % 4% 22 % 100 %
CCI 0.48 0.58 0.81 0.72
ICI 0.48 0.32 0.60 0.12
Dawa 2,825 415 500 D 8% 24 % 46 % 100 %
CCI −0.48 0.43 0.11 0.60
ICI7 −0.48 0.18 −0.71 0.07
Weyb 3,750 220 700 D 10 % 25 % 35 % 100 %
CCI −0.49 0.56 0.17 0.69
ICI −0.49 0.28 0.18 0.74
Erer 2,235 615 250 D 4% 8.5 % 44 % 100 %
CCI 0.24 −0.05 0.59 0.57
ICI 0.24 −0.13 0.1 0.2
Fafan 2,230 280 800 D 1% 12 % 45 % 100 %
CCI 0.5 0.65 0.6 0.53
ICI 0.5 0.47 0.2 0.26
Gerer 1,875 855 300 D 11 % 33 % 66 % 100 %
CCI 0.27 0.23 0.32 0.22
ICI 0.27 0.16 0.04 0.004
Other rivers
Darling (Australia) 1,245 105 1,225 D 0.4 % 2.5 % 7% 100 %
CCI 0.67 0.52 0.75 0.91
ICI 0.67 0.25 0.43 0.78
Rhine (Europe) 1,360 0 2,170 D 0.07 % 2% 7% 100 %
CCI 0.18 0.36 0.7 0.82
ICI 0.18 0.21 0.25 0.54
Loire (France) 1,200 0 1,055 D 2% 8% 20 % 100 %
CCI 0.27 0.33 0.53 0.81
ICI 0.27 0.53 0.53 0.44
Amazon tributary (Andes) 4,490 255 890 D 11 % 33 % 55 % 100 %
CCI 0.69 0.16 0.11 0.33
ICI 0.69 −0.19 −0.51 0.74
hmax = maximum river elevation a.s.l.; hmin = minimum river elevation a.s.l.; l = river length; h = river height decrease since its source; D = horizontal distance along
the river taken from its source; CCI = Cumulative concavity index; ICI: Incremental cumulative index. Cumulative index (Langbein 1964) is defined as 2A/H, with A
the vertical difference between the profile midterm and a straight line joining the two ends (or any point) of the longitudinal profile; H is the elevation difference
between the uppermost and lowermost points of the straight line. CCI measures concavity of a river profile from its source to a given D. ICI measures concavity
between two given D’s, as illustrated in the insert of Fig. X.15. The indexes are 0 for straight river profile; −1 is for strongly unequilibrated (convex) profile; 1 is for
fully equilibrated (concave) profile. Indexes are bold when convex. In each category, the rivers are ranked by cumulative concavity index, from higher to lower
342 D. Mège et al.
Fig. 19.15 Normalized river

profile plot for the rivers
displayed on Fig. 19.14. The inset
shows relationships between
geomorphometric parameters
used in Table 19.2
flat areas and slopes with a gradient of 5–7 m/km, giving an Shebele, which is attributed to the structural control of the
overall gradient of about 3 m/km. However, at about 110 Marda Range.
km, the slope increases abruptly to 20 m/km over several In the Genale system (Fig. 19.14b), the Genale River
tens of kilometers and only recovers its average slope of (1,300 km) has a linear profile upstream (20 m/km upstream,
3 m/km after 340–350 km, at 660 m of altitude. This huge CCI = 0.04, Table 19.2) but changes abruptly at km 70 to a
vertical-step break in slope creates the spectacular deep spectacular convex profile, with the slope decreasing sharply
canyons for which the river is famous (Fig. 19.3d). In its to 2 m/km and then increasing regularly to 5 m/km between
lower reaches, the Wabe Shebele is again marked by a km 350 and 400 (ICI = −0.2). Another abrupt change occurs
succession of vertical-step knick-points but of one order of downstream, with the slope decreasing to around 0.3 m/km
magnitude less prominent. Eventually, the Wabe Shebele and staying constant for the next 900 km.
ends its Ethiopian course in a wide floodplain (10–20 km) The Weyb and Dawa rivers are both marked by very
with a slope of only 0.2 m/km. Even if its upper part is irregular profiles, with a succession of major slope breaks in
highly disrupted by knick-points, its overall profile is close their upper reaches giving convex profile sections
to those of Rhine or Loire rivers, being smooth concave with (ICI ≤ −0.48) and very high average slopes: 10 m/km for the
irregularities: a near equilibrium profile (Fig. 19.15), with a upper 300 km for the Weyb (CCI = 0.69) and 12 m/km
CCI of 0.82 and 0.81, respectively (Table 19.2). during the upper 250 km for the Dawa (CCI = 0.6). In
The Fafan River has a much lower CCI (0.53), denoting a contrast, the Uelmal River is close to an equilibrium profile
significantly different evolution. Except in the uppermost (ICI = 0.87). Its slope decreases gradually from the very high
20 km, where a high slope (25 m/km) is manifest (incre- value of 100 m/km down to 1 m/km during the 300 km
mental cumulative index ICI = 0.5), the Fafan has a linear course, albeit marked by knick-points at km 50 and km 100
profile for approximately 700 km, with a slope of 1.5 m/km (CCI ≤ −0.42).
and without any remarkable knick-points (ICI = 0.2–0.26). The highland part of both Genale and Wabe Shebele is
The same evolution is observed for the Gerer (linear profile not in equilibrium with their base level. Comparison with
at 3 m/km along 300 km), Dakota (2.5 m/km along 200 km) profiles from the Andes, where uplift is active (CI = 0.33 but
and Erer rivers (3.5 m/km along 225 km). High-gradient negative over much of its profile ICI), and Australia, where
slopes of *35–40 m/km are present in these rivers only in the selected river is in equilibrium (Fig. 19.15) (CCI = 0.91),
the initial 25 km. None of those valleys show an equilibrium reveals that the Wabe Shebele, Genale, and Uelmal are close
profile. The Fafan and Gerer valleys are spectacularly linear to an equilibrium profile, but the presence of irregularities
and long compared to the other tributaries of the Wabe indicates that none of them is a perfect equilibrium profile.
Fig. 19.16 Profiles of modern

rivers (blue lines) and the base of
basalt flows (black dots, names in
red). Age details are on
Table 19.1. The water divides in
(c) are located on Fig. 19.13. The
two data clusters in (f) are for the
two Genale basalt tables located
on Fig. 19.7. The difference in
elevation reveals the depth of
incision
The Dawa and Weyb profiles are closer to the Andean incision rates. For example, the elevation of the base of the
profile, whereas the Erer, Dakota, and Fafan show exotic Marda volcanics can be compared with the elevation of the
bimodal linear profiles, and the Gerer has a unimodal linear beds of the Gerer and Fafan rivers, and the Gode flow, with
profile, the latter being very uncommon. This analysis of the the Wabe Shebele (Fig. 19.13).
profiles of the Genale, and Wabe Shebele and their tribu- This difference in elevation reveals the depth of incision
taries suggests a major base-level fall in the Ogaden, most since the basalt outpouring and that can be converted to a
likely due to tectonic uplift of one of its margins. In the next mean incision rate. The reliability of such an analysis
section, we discuss the use of age-dated basalts in paleo river requires that the horizontal distance between paleo- and
valleys to help constrain uplift amplitude and timing. present rivers is minimal and both sites are located in the
same geological domain, without intervening deformation
subsequent to the lava flow.
19.4.2 Insight into the Evolution The elevation of the longitudinal profiles of six basalt-
of Ogaden Topography Since filled paleo river channels in the Wabe Shebele drainage
the Upper Oligocene basin was analyzed and compared with the elevation of the
present river profiles. In addition to the Marda and Gode
The presence of age-dated basaltic flows, which filled paleo flows mentioned above, the Daga Medo volcanics were
river valleys and presently are only a short distance from the compared with the Fafan river, and the Fik volcanics with
modern river valley, allows the determination of average the Erer and Dakota rivers in the north and a Wabe Shebele
344 D. Mège et al.
tributary in the south. The elevation of the two Genale Marda, Daga Medo, Fik, and Kebri Dehar (Fig. 19.16a–d),
basaltic tables was also compared with the elevation of the and also for the younger Gode flow. Moreover, the Fafan has
Genale River (Figs. 19.13 and 19.16). In all cases, the area incised faster than the Gerer (Fig. 19.16a); the Dakota than the
between the volcanic flow and the modern river valley has Gerer (19.16b), and the Erer than the Dakota (19.16c),
been geologically stable since the time of eruption. The revealing that the mean incision rate has increased from east to
Marda volcanic flow, being located on the eastern side of the west. The incision rate has also increased from south to north,
prominent Marda Precambrian deformation zone, warrants mean incision rate for the middle and upper Dakota being 7–
some caution in that regard but the excellent and continuous 11 m/Myr. (Figure 19.16c) and 13–20 m/Myr, respectively
rock exposures through the Marda volcanics and across the (Fig. 19.16b). For the Fafan it has increased, south-to-north,
neighboring Fafan and Gerer rivers on the Harar-Jijiga road from 10–13 m/Myr. (Figure 19.16b) to 16–36 m/Myr.
show no significant postflow deformation. (Fig. 19.16c). The maximum mean incision rate of about
The elevation of the base of every basalt outcrop was 30 m/Myr (Fig. 19.16e) was obtained for the Wabe Shebele in
measured using elevation data from ASTER DEM, with its upstream course at the northernmost outcrop of the Gode
geographical control from Landsat ETM+ images and high- flow. Incision rates calculated for the different rivers and the
resolution images available on Google EarthTM. Uncertainty age of the associated lava flows are given on Fig. 19.17.
in elevation control is estimated to 10–15 m. Some of these The base of the Fik flow has a linear profile but crosses
flows (Marda, Mustahil, Kebri Dehar, Gode) have been two water divides (Fig. 19.16c), showing that the drainage
dated recently using argon radiochronology and their age is network has changed since the basalt outpouring, with the
considered reliable. The Marda, Daga Medo, and Fik flows Erer and Dakota rivers achieving their present geometry
were dated in the 1990s by the K-Ar method but have the more recently than the Fafan and Gerer. This is well illus-
same age span as most of the argon-dated samples, sug- trated on the comparative river profile plot (Fig. 19.15), in
gesting these ages are plausible. The Genale tables have not which the Erer and Dakota river profiles are close to recently
been dated. More details are given on Table 19.1. equilibrated profiles, in contrast to the linear Fafan and Gerer
Figure 19.16 presents the results of these comparisons and profiles, which indicate a relict transient profile (Kirby and
shows clearly that incision is intense at the river head and Whipple 2012 and references therein). The Erer and Dakota
decreases downstream. This is true for the older basalts at have their convergence point with the Wabe Shebele much
Fig. 19.17 Mean incision rate

calculated from the data presented
in Fig. 19.16. Incision after the
Marda (a), Daga Medo, Kebri
Dehar, and Mustahil (b), and Fik
flows (c), starts at 28–22 Myr,
i.e., as soon as the lavas are
erupted. For Gode (d) erosion
starts at 8–5 Myr
higher upstream than the Fafan and have been therefore and concave. In the east, Fafan-type rivers have preserved a
more sensitive to the geodynamic events occurring in the tilted slope in the uppermost part of the profile, but most of
upper Shebele region. the profile has not changed. The Gerer River is an extreme
Downstream, the Fik flow follows a Wabe Shebele trib- case in which the tilted slope is atrophied and much of the
utary with a constant vertical base-difference of 50 m. This remaining slope is constant, denoting geodynamic stability
tributary converges on the Wabe Shebele 100 km southward, and probably an increasingly dry climate and subsequent
cutting through the Gode basalt outcrops. This moderate lower erosion rate since the Miocene (e.g., Feakins 2013).
incision reflects the lower uplift rate of this area.
Similarly, in the easternmost Shebele basin, where the
Kebri Dehar flow is located, incision in the Fafan tributary that 19.4.3 Uplift Rates
parallels the Kebri Dehar flow is slow (2.5 m/Myr) and nearly
constant. This is consistent with the general incision pattern in Commonly, large-scale river incision is the result of base-
the basin, with the mean incision rate increasing from east to level fall, which can be caused by either tectonic uplift or
west and south to north, toward the upper Wabe Shebele sea-level (or lake-level) fall. In the Wabe Shebele valley, the
(Fig. 19.13). An extreme and remarkable case is the Mustahil geomorphological analyses indicate that the Wabe Shebele
flow (Fig. 19.16b) on the southeasternmost border of the valley profile has nearly reached its equilibrium profile
Ogaden (that is, as far as Ogaden can be from Afar), which, in downstream. All the perturbations in valley profiles are
spite of its 28 Ma age, ends at the Wabe Shebele–Fafan con- located in the very upstream part of the system. Conse-
fluence, at appreciably the same elevation as the present Wabe quently, the calculated mean incision rates (Fig. 19.17) can
Shebele, pointing to the vertical stability of this confluence. be taken to approximate the uplift rates, with only minor, if
This specific location along the Wabe Shebele may be close to any, contribution coming from sea-level variations.
the point in the river profile where incision stops and depo- Vertical movements of the surface (and underpinning
sition starts. Perhaps not coincidentally, with the exception of crust) in the Ogaden region are a result of the complex
the northeastern Ogaden (Mège et al. in press), the Cenozoic geodynamics in the Horn of African since the Oligocene.
volcanic rocks in the Ogaden are outcropping and being ero- The calculation of river incision rates helps in the under-
ded, whereas further east in Somalia the volcanic rocks are standing of the present topography at a regional scale, but
covered by Miocene to Quaternary sediments (e.g., Bosellini does not inform regarding the uplift rate variations. On the
1989; Faillace 1993). More generally, there is a marked Somali Plateau, such variations were probably not negligi-
absence of Neogene rocks throughout the Ogaden, except its ble, as discussed above in Sect. 19.2.2 with regard to the
northeastern part, while sediments of this age are widespread origin of the plateau elevation, and mean uplift rates need to
in Somalia (Abbate et al. 1994). be interpreted with caution when reconstructing paleotopo-
The Wabe Shebele floodplain begins near the latitude of graphic evolution stages. Nevertheless, for young ages,
Gode town. Its beginning is apparent on Fig. 19.16e, at the additional information is given by the slope break in the
convergent point between the upstream profile, with a steep paleo-Wabe Shebele profile (Fig. 19.16e), which shows that
longitudinal slope (2 m/km) and a high mean incision rate at least part of the uplift postdates 7 Ma. Since that time, the
(30 m/Myr), and the downstream profile, where the slope is mean incision rate has been *30 m/Myr upstream,
gentle (0.5 m/km) and the mean incision rate is between decreasing to zero downstream near the Somalia border.
almost 15 m/Myr and zero, the latter occurring at the con- This incision and the causative uplift (as argued above) are
vergence point with the Fafan River. related to the development of the Main Ethiopian Rift. These
Floodplains usually develop in subsidence areas (Syvitski measurements have been obtained at a minimum distance of
et al. 2012), in which detrital deposits accumulate, elevating 250 km from the rift margin, and the uplift rate of the
the slope profile. As a result, the Wabe Shebele incision may westernmost Ogaden along the rift shoulder must have been
be underestimated. The sharp transition in the slope profile much higher, given the usual rift-flank concave curvature.
of the base of the Gode basalts observed on Fig. 19.16e may Furthermore, if rift-flank uplift has been the dominant
be due to either a variation in the ancestral Wabe Shebele mechanism of topographic building, then visco-elastic
profile, or a difference in uplift rate, with an increasing mean models predict that this *30 m/Myr uplift rate could have
uplift gradient from south to north. been relatively constant since the beginning of rifting
The analysis of the paleotopography preserved by the (Sachau and Koehn 2010). The mean incision rates obtained
basalts indicates that the present-day Wabe Shebele catch- for the Wabe Shebele would correspond to a constant uplift
ment is divided into two main parts that had contrasting rate since that time.
responses to uplift. In the west, river profiles of Wabe The Fafan River is located on the western side of a major
Shebele-type have nearly completed a return to an equilib- crustal discontinuity, now manifest at the surface in the
rium state, even though the entire profile is not yet smooth Marda Range; the Gerer River is located on its eastern side,
346 D. Mège et al.
and both rivers are parallel to it. The Marda Range serves, Acknowledgments Part of this work was funded by the CNRS/INSU
therefore, as the boundary between the northern uplifted Marges programme. Field work in 2008 was conducted in cooperation
with Pexco Ethiopia Exploration (East Africa) N.V. 40Ar/39Ar dating
Ogaden, with its rivers responding according to their setting, for this project is performed by Dr Fred Jourdan at the Western Aus-
and the eastern stable Ogaden for which the evolution of the tralian Argon Isotope Facility, Curtin University, Perth, Western Aus-
Main African Rift appears, perhaps deceptively, to have tralia. ASTER GDEM is a product of METI and NASA.
been only a remote influence.
References
19.5 Concluding Remarks
Abbate E, Sagri M, Sassi FP (1994) Geological map of Somalia,
The landscape and landforms of the Ogaden reflect the 1:1,500,000. Somali National University, Mogadishu
Abdirahim MM, Ali Kassim M, Carmignani L, Coltorti M (1993) The
geological development of the region. This is clearly mani- geomorphological evolution of the upper Juba valley in southern
fest at a regional scale: the steep, nearly impenetrable can- Somalia. In: Abbate E, Sagri M, Sassi FP (eds) Geology and
yons in the north, for example, are a consequence of the mineral resources of Somalia and surrounding regions. Florence,
relative uplift of the Somali Plateau margin. Similarly, the Italy, vol 113 (Ist. Agron Oltremare Relaz E Monogr), pp 241–250
Ali Kassim M, Carmignani L, Fazzuoli M (1987) Geology of the Luuq-
contrast between the eastern and western Ogaden landscapes Mandera basin, geology of Somalia and surrounding regions,
owes much to the pattern of Jurassic and Cretaceous sedi- excursion A, GEOSOM 87. Somali National University, Depart-
mentation in these areas, and the impact of erosion during ment of Geology, Mogadishu, pp 1–43
Cretaceous, Tertiary and Recent times. This interplay of Ali Kassim M et al (1993) Flood basalts of the Gedo Region (Southern
Somalia): geology, petrology and isotope geochemistry. In: Abbate
Mesozoic sediments and Tertiary–Recent erosion is also E, Sagri M, Sassi FP (eds) Geology and mineral resources of
seen on a local scale in the Audo Range, where erosional Somalia and surrounding regions. Florence, Italy, (Ist. Agron.
debuttressing of the Gorrahei evaporites has triggered Oltremare, Relaz. E Monogr) vol 113, pp 311–334
gravitation collapse and spreading of blocks of the overlying Angelier J, Colletta B (1983) Tension fractures and extensional
tectonics. Nature 301:49–51
Mustahil and Yesomma formations. Bauduin D et al (1973) Projet du Wabi Shebele: étude hydrologique.
In spite of its diversity of landforms, somewhat amazingly, ORSTOM, fdi:06586, Paris
the Ogaden forms a continuous and coherent mega-geomor- Beicip (1985) Geological map of the Ogaden and surrounding area,
phological unit: a long eastward-dipping slope of dominantly 1:1,000,000. Geological Survey of Ethiopia, Addis Ababa
Beicip-Franlab (1998) Petroleum potential of Ethiopia. The Ministry of
thermal origin that docilely recorded the tectonic jolts from its Mines and Energy of Ethiopia, Addis Ababa
neighborhood, mainly the many volcanic and rifting events. Boccaleti M, Getaneh A et al (1991) The Marda fault: a remnant of an
These resulted in dramatic upwarping on the western and incipient aborted rift in the paleo-African Arabian plate. J Pet Geol
northern side of the region, while the eastern and southern 14:79–92
Bonini M et al (2005) Evolution of the main Ethiopian rift in the frame
regions underwent negligible deformation. Vertical motions of Afar and Kenya rifts propagation. Tectonics 24, TC1007. doi:10.
aside, the huge size of the area has allowed landforms to 1029/2004TC001680
develop over broad surfaces: many identified lava flows Bosellini A (1989) The continental margins of Somalia: their structural
are >100 km long and the Audo Range gravitational spreading evolution and sequence stratigraphy. Mem Sci Geol 41:373–458
Cheung WH, Senay GB (2008) Trends and spatial distribution of
complex could be the largest gravitational spreading domain annual and seasonal rainfall in Ethiopia. Int J Climatol 28
on Earth. Other landforms not described in this chapter (13):1723–1734. doi:10.1002/joc
include the karstic landscape of the Gabredarre limestones Dainelli G (1943) Geologia dell’Africa orientale. Centro Studi per
south of Ginir, where the splendid Sof Omar caves are located. l’Africa Orientale Italiana, Rome
Duvall A, Kirby E, Burbank D (2004) Tectonic and lithologic controls
A better understanding of the geomorphology of Ogaden on bedrock channel profiles and processes in coastal California.
will require considerably more exploration and study, per- J Geophys Res 109(F3). doi:10.1029/2003JF000086
haps especially the easternmost region. That the landscape of Ethiopia Government Portal (2014) www.ethiopia.gov.et/stateoromia &
the Ogaden and its flora and fauna have not been better /statesomali. Accessed Jan 2014
Ethiopian Mapping Authority (1988) National atlas of Ethiopia. Addis
studied to date is, in part at least, a product of the remoteness Ababa, Ethiopia
and limited infrastructure but also of the difficult security Farr T et al (2007) The shuttle radar topography mission. Rev Geophys
conditions that have prevailed over large areas for many 45:1–33
years. Current issues are the modern guise of conflicts cen- Faillace C (1993) Hydrogeological importance of the sub-surface
basalts in the Mudug-Galgadud plateau. In: Abbate E et al (eds)
turies old and have their origins in that landscape and the rift Geology and mineral resources of somalia and surrounding regions,
that developed between the cultures on the Somali and vol B, Mineral and water resources. (Ist Agro l’Oltremare, Rel Mon,
Ethiopian plateaus. Firenze, Italy) vol 113, pp 649–664
Feakins SJ (2013) Pollen-corrected leaf wax D/H reconstructions of Micheels A, Eronen J, Mosbrugger V (2009) The late Miocene climate
northeast African hydrological changes during the late Miocene. response to a modern Sahara desert. Glob Planet Change 67:193–
Palaeogeogr Palaeoclimatol Palaeoecol 374:62–71. doi:10.1016/j. 204. doi:10.1016/j.gloplacha.2009.02.005
palaeo.2013.01.004 Mohr P (1963) Geologic map of African Horn of Africa. Philip and
Fournier M et al (2010) Arabia-Somalia plate kinematics, evolution of Tacey, Fulham
the Aden-Owen-Carlsberg triple junction, and opening of the Gulf Moore JM, Schultz RA (1999) Processes of faulting in jointed rocks of
of Aden. J Geophys Res 115, B04102. doi:10.1029/2008JB006257 Canyonlands National Park, Utah. Geol Soc Am Bull 111:808–822
Fueten F, Stesky RM, MacKinnon P (2005) Structural attitudes of large Moucha R, Forte AM (2011) Changes in Africa topography driven by
scale layering in Valles Marineris calculated from mars orbiter laser mantle convection. Nat Geosci 4:707–712. doi:10.1038/NGEO1235
altimeter data and mars orbiter camera imagery. Icarus 75:68–77 Nicholson SE (1996) A review of climate dynamics and climate
Gani NDS, Abdelsalam MG et al (2009) Stratigraphic and structural variability in East Africa. In: Johnson TC, Odada E (eds) The
evolution of the Blue Nile Basin, Northwestern Ethiopian Plateau. limnology, climatology and paleoclimatology of the East African
Geol J 44:30–56 lakes. Gordon and Breach, Amsterdam, pp 25–56
Hagmann T (2007) The political roots of the current crisis in region 5. Olson JE (1993) Joint pattern development: effects of subcritical crack
Social Science Research Council, Web forum Crisis in the Horn of growth and mechanical crack interaction. J Geophys Res 98
Africa, New York (B7):12,251–12265
Hagmann T, Khalif MH (2006) State and politics in Ethiopia’s Somali Olson P (1994) Mechanics of flood basalt magmatism. In: Ryan MP
Region since 1991. Bildhaan. Int J Somali Stud 6:25–49 (ed) Magmatic systems. Academic Press, New York, pp 1–18
Hofmann C, Courtillot V et al (1997) Timing of the Ethiopian flood Peckham SD (1998) Efficient extraction of river networks and
basalt event and implication for plume birth and global change. hydrologic measurements from digital elevation data. In: Barn-
Nature 389:838–841 dorff-Nielsen OE et al (eds) Stochastic methods in hydrology: rain,
Juch D (1975) Geology of south-eastern escarpment of Ethiopia landforms and floods. World Scientific, Singapore, pp 173–203
between 39° and 42° long. East. In: Pilger A, Rosler A (eds) Purcell P (1976) The Marda Fault Zone, Ethiopia. Nature 261
Proceedings of an international symposium on the afar region and (5561):569–571. doi:10.1038/261569a0
related rift problems. Stuttgart, E. Schweizerbart’sche Purcell P (1981) Phanerozoic sedimentary history and petroleum
Verlagsbuchhandlung potential. In: Chewaka S, de Wit MJ (eds) Plate tectonics and
Kazmin V (1972) Geological map of Ethiopia, 1st edn, 1:2,000,000. metallogenesis: some guidelines to Ethiopian mineral deposits.
Geol Surv Ethiopia, Addis Ababa Ethiopian Inst Geol Surv Bull 2:97–114
Kibrie T, Yirga T (2008) The geology of Bedesa area (NC37-16) Geol Purcell P, Mège D, Jourdan F (2011) Volcanic geomorphology of
Surv Ethiopia Memoir 10, Addis Ababa Southeast Ethiopia. In: Asrat A et al (eds) Geomorphology for
Kirby E, Whipple KX (2012) Expression of active tectonics in human adaptation to changing tropical environments. IAG/IAG
erosional landscapes. J Struct Geol 44:54–75 regional conference, Addis Ababa, 127
Lemma G (1996) Climatic classification of Ethiopia. National Mapping Sachau T, Koehn D (2010) Faulting of the lithosphere during extension
Authority, Addis Ababa and related rift-flank uplift: a numerical study. Int J Earth Sci
Leroy S et al (2012) From rifting to oceanic spreading in the Gulf of 99:1616–1632. doi:10.1007/s00531-010-0513-6
Aden: a synthesis. J Arab Earth Sci 5:859–901. doi:10.1007/ Schultz-Ela DD (2001) Excursus on gravity gliding and gravity
s12517-011-0475-4 spreading. J Struct Geol 23:725–731
Langbein WB (1964) Profiles of rivers of uniform discharge. US Geol Schumm SA, Dumont JF, Holbrook JM (2002) Active tectonics and
Surv Prof Pap 501-B:119–122 alluvial rivers. Cambridge University Press, Cambridge
Lopez-Gonzalez T (2006) L’étalement gravitaire de la ride de Chebis Sepulchre P, Ramstein G, Fluteau F, Schuster M, Tiercelin J-J, Brunet
dans le bassin de l’Ogaden, Ethiopie. MSc. Thesis, Laboratoire de M (2006) Tectonic uplift and eastern Africa aridification. Science
Planétologie et Géodynamique, Nantes University 313:1419–1423
Maxus Ethiopia (1993) Ogaden concession final report. Addis Ababa, Snyder NP, Whipple KX, Tucker GE, Merritts DJ (2000) Landscape
Ethiopia response to tectonic forcing: digital elevation model analysis of
Mège D, Le Deit L, Rango T, Korme T (2013) Gravity tectonics of stream profiles in the Mendocino triple junction region, northern
topographic ridges: Halokinesis and gravitational spreading in the California. Geol Soc Am Bull 112:1250–1263
western Ogaden, Ethiopia. Geomorphology 193:1–13 Syvitski JPM, Overeem I, Brakenridge GR, Hannon M (2012) Floods,
Mège D, Purcell, PG, Jourdan F (2012a) Volcanism in southeast floodplains, delta plains—a satellite imaging approach. Sedim Geol
Ethiopia and the Ogaden Dyke Swarm. Magmatic rifting & active 267–268:1–14
volcanism conference, Addis Ababa, Abstract Tefera M, Chernet T, Haro W (1996) Geological map of Ethiopia, 2nd
Mège D, Purcell P, Jourdan F (2012b) Dykes and linear troughs: new edn, 1:2,000,000. Geol. Survey Ethiopia, Addis Ababa
observations on the Somali Plate. Lunar and planetary science Temin J (2006) Grassroots conflict assessment of the Somali Region,
conference 43, Lunar and Planetary Institute, Houston, Texas, Ethiopia. CHF International
Abstract 1317 Thulin M (2007) Acacia fumosa sp. nov. (Fabaceae) from eastern
Mège D, Purcell P et al A major dyke swarm in the Ogaden region Ethiopia. Nordic J Bot 25:272–274
south of afar and the early evolution of the afar triple junction. In: Tolan TL, Reidel SP et al (1989) Revisions to the estimates of the areal
Wright TJ, Ayele A et al (eds) Magmatic Rifting and Active extent and volume of the Columbia River Basalt Group. In: Reidel
Volcanism. Special Publication Geological Society of London, 420, SP, Hooper PR (eds) Volcanism and tectonism in the Columbia
in press River Flood-Basalt Province, US Geol Surv Sp Pap 239:1–20
Megrue GH, Norton E et al (1972) Tectonic history of the Ethiopian rift UNDP (1999) Drought and floods stress livelihoods and food security
as deduced by K-Ar ages and palaeomagnetic measurements of in the Ethiopian Somali Region. UNDP Emergencies Unit for
basaltic dykes. J Geophys Res 77(29):5744–54 Ethiopia report
Merla G, Abbate E et al (1973) Geological map of Ethiopia and Walsh J (1976) Oil and gas in Ethiopia. Geol Surv Ethiopia Min Circ 4
Somalia, 1:2,000,000. Consiglio Nazionale Della Ricerche, Watchorn F, Nichols GJ, Bosence DWJ (1998) Rift-related sedimen-
Florence tation and stratigraphy, southern Yemen (Gulf of Aden). In: Purser
348 D. Mège et al.
BH, Bosence DWJ (eds) Sedimentation and tectonics of rift basins: to document the evolution of the Ethiopian Large Igneous Province and the
red Sea-Gulf of Aden. Chapman and Hall, London, pp 165–189 Afar triple junction, and of various bodies of the solar system. Daniel is a
Whipple KX (2004) Bedrock rivers and the geomorphology of active member of the Geological Society of America, the American Geophysical
orogens. Ann Rev Earth Planet Sci 32:151–185 Union, the Geological Society of Africa, and IEEE.
Whipple KX, DiBiase RA, Crosby BT (2013) Bedrock rivers. In:
Peter Purcell Peter Purcell is an Australian geologist with an MSc from
Shroder J Jr, Wohl E (eds) Treatise on geomorphology, vol 9.,
the University of Sydney. He is General Manager of P&R Geological
Fluvial Geomorphology. Academic Press, San Diego, pp 550–573 Consultants Pty Ltd, with over 40 years experience working in Australia, SE
Wolfenden E, Ebinger C et al (2004) Evolution of the northern Main Asia, and East Africa. Peter first worked in Ethiopia in 1973–1977 and
Ethiopian rift: birth of a triple junction. Earth Planet Sci Lett continues to be involved with exploration and research there. He has edited
224:213–228 several books on Australian geology and is the author of numerous articles
on geology and petroleum exploration in Australia and East Africa, as well
as environmental and social issues. Peter is a fellow of the Geological
Society of London and a member of AAPG, PESA, and ASEG.
Author Biographies Stéphane Pochat Stéphane Pochat received the MSc degree in sedimen-
Daniel Mège Daniel Mège obtained a M.Sc. degree in geophysics and tology and paleoclimatology in 1999 and a PhD degree in sedimentology
internal geodynamics in 1991 in Paris at University of Paris-Sud, Orsay, and tectonics at University of Rennes 1 (Geosciences Rennes), France. He
Ecole Normale Supérieure-Ulm, and Pierre and Marie Curie University. In was appointed as assistant professor, then associate professor in 2007 at
1994, he obtained a Ph.D. degree in planetary sciences at University Paris- University of Nantes (LPG Nantes), France. He shares his time between
Sud, Orsay, France. In charge of the NASA Regional Planetary Image paleoclimatic and sedimentary basins dynamic during the Paleozoic, the
Facility at University of Paris-Sud in 1995–96, he was also a postdoc at study of marine terraces, fluvial geomorphology in Ethiopia and Somalia,
DLR and at University of Nevada, Reno. He was then appointed as assistant and fluvial and glacial geomorphology on Mars.
professor at Blaise-Pascal University, Clermont-Ferrand, France, before Thomas Guidat Thomas Guidat is a geomorphologist who obtained a
obtaining an associate professor position at Pierre and Marie Curie Uni- MSc degree in planetology in 2013 at University of Nantes (LPG Nantes),
versity in 1997. In 2001, he obtained his habilitation in comparative plan- France. He worked on geomorphometric analysis of basalt-filled paleorivers
etology and moved to University of Nantes, France, in 2004. Since 2011, he in Ogaden and Somalia, and glacial evolution of Isidis Planitia on Mars. He
is a Visiting Professor at the Institute of Geological Sciences at the Polish is currently involved in the PhD program at Trinity College, Dublin, where
Academy of Sciences, where he is in charge of developing a research group he is studying eolian dynamics on Mars. Thomas is a member of the British
working on planetary sciences. He conducts research on the geology of the Society of Geomorphology.
Gojjam lowlands in western Ethiopia and the Ogaden in order to contribute
Part III
Applied Aspects
Geo-hazard in Ethiopia
20
Giandomenico Fubelli and Francesco Dramis
Abstract
This paper presents an overview of the potentially dangerous geological–geomorphological
processes that characterize the territory of Ethiopia. Due to the active crustal mobility causing
the ongoing rifting between the Arabian and African plates, this country is particularly prone
to high magnitude earthquakes and volcanic eruptions. Moreover, the climate, characterized
by long-lasting and intense precipitation during the rainy seasons, and the recurrent earthquake
shocks are the triggering factors of a variety of landslides that are particularly favoured by the
rugged topography, marked by deep valleys and high steep slopes. All these processes may
induce different levels of risk in relation to the different hazards and vulnerabilities, especially
after the recent development and planning of road networks, dams, reservoirs, and the
establishment of new urban settlements in previously uninhabited areas. Finally, an overview
of the measures taken by the Ethiopian authorities so far to mitigate geological hazards is
presented along with some proposals for future mitigation initiatives.

Keywords
Geo-hazards Earthquakes Volcanoes Landslides Risk Ethiopia East Africa
20.1 Introduction Earthquakes and volcanic eruptions are related to crustal

diastrophism and, in particular, to different systems of frac-
Geo-hazards, including earthquakes, ground fissuring, vol- tures and faults, the modern activity of which is testified by
canic eruptions, and landslides, are crucial problems for the dislocation of Late Quaternary volcanics and alluvial
Ethiopia (Fig. 20.1), causing heavy limitation for urbaniza- deposits (Faure 1975; Abebe et al. 2007).
tion and infrastructural projects and, more in general, for the Seismic events with magnitudes greater than 5 are not
socioeconomic development of the country. uncommon along the margins of the Main Ethiopian Rift
Two main factors are responsible for the occurrence of (MER), the northernmost sector of the EARS which divides
geo-hazards in the Horn of Africa: (1) The regional geody- the Ethiopian plateau into two uplifted Blocks (the north-
namics causing the ongoing progressive separation of the western Plateau and the southeastern Plateau), and in the Afar
Arabian and African plates, i.e., the East African Rift System Triangle, where the MER joins with the Gulf of Aden and Red
(EARS), and (2) the climate characterized by long-lasting Sea rifts, thus splitting the African Plate into the Nubian Plate
and intense precipitation during the rainy seasons. to the north and the Somali Plate to the south (Fig. 20.2).
Widespread evidence of seismically induced landslides,
mostly first generated rapid movements, is provided by the
historical record of earthquakes and related surface effects
G. Fubelli (&) F. Dramis (Gouin 1979). However, the main triggering factor of slope
Department of Sciences, Roma Tre University, Rome, Italy failures in Ethiopia is heavy rainfall, mostly concentrated in
e-mail: giandomenico.fubelli@uniroma3.it
July and August (Gamachu 1977; Ethiopian Mapping
F. Dramis Authority 1988; Chap. 3, this volume).
e-mail: dramis@uniroma3.it
352 G. Fubelli and F. Dramis
reference to the case study of Dese, a major Ethiopian town

heavily struck by landslides. A short description of the risk
mitigation measures undertaken so far by the Ethiopian
government is also given.
20.2 Earthquakes
The first studies on the seismicity of the MER and the Afar
Triangle are those of Gutenberg and Richter (1954), Sykes
and Landisman (1964), Gouin and Mohr (1967), Mohr
(1967), Fairhead (1968), Gouin (1970, 1979). More recently,
seismic episodes have been described by Asfaw (1988,
1992), Ayele (1995, 1997), Kebede (1996), Ayele et al.
(2007). According to Gouin (1979), who published a com-
prehensive study of earthquakes in Ethiopia and the Horn of
Africa up to 1977, at least 15,000 tremors, strong enough to
be felt by humans, including a large number of highly
destructive events, have been recorded since the earliest
known earthquake, dating back to AD 1431, during the
kingdom of Emperor Zara Yaqob.
Kebede (1996) recognized a total of 16 earthquakes of
magnitude higher than 6.0 during the twentieth century in
Ethiopia. Significant among these were the events that
affected southern Shewa (1906), Awasa (1960), the Kara
Kore-Mejete area (1961), Serdo (1969), Langano (1985),
and the Dôbi graben in the Afar (1989). Strong earthquakes
also struck Massawa, Eritrea (1921—M = 5.8), and French
Somalia, Djibouti (1960—M = 6.0; 1973—M = 6.0; 1989—
M = 6.3). In this century, moderate to high magnitude
Fig. 20.1 Location map of the cited localities: 1 Awasa; 2 Kara Kore-
Mejete; 3 Serdo-Dôbi graben; 4 Shewa; 5 Langano; 6 Massawa; 7 earthquakes affected the Afar (2005—M = 5.6), Hosaina
Djibouti; 8 Hosaina; 9 Yirga Alem-Moto; 10 Metehara; 11 Ziway; 12 (2010—M = 5.3), Yirga Alem (2011—M = 5.0), and the
Butajira; 13 Welenchiti; 14 Melka; 15 Fentale Volcano; 16 Afdera southern Eritrean–Ethiopian border (2011—M = 5.7).
Volcano; 17 Dubbi Volcano; 18 Erta’ Ale Volcano; 19 Dabbahu In many cases, the earthquake shocks heavily damaged
Volcano; 20 Mand Hararo Volcano; 21 Nabro Volcano; 22 Ankober;
23 Dese (Tossa—Azwa Gedel); 24 Kombolcha; 25 Wendo Genet; 26 built-up areas and, in some cases, caused fatalities (Gouin
Mt. Bosetti; 27 Borkena River; 28 Muleti; 29 Kone volcano-Galibardi 1979). However, until the 1960s, the structural damage to
Pass. For river location, see Fig. 4.2 of Chap. 4 (this volume) buildings and infrastructure was generally low due to their
scarcity in the seismic-prone areas of the country. The recent
All geodynamic and geomorphic processes may interact development of urban settlements close to and within the
with people, buildings, infrastructures, and farmlands, MER and the Afar Triangle has definitely increased the level
inducing different levels of risk in relation to the different of seismic risk in Ethiopia.
hazards and vulnerability levels (IUGS WGL-CRA 1997). In the occasion of major events, coseismic ground effects
The recent development and planning of new road net- such as surface faults and fissures, as well as gravitational
works, dams, reservoirs, and the establishment of new urban movements were produced (Gouin 1979; Jacques et al.
settlements in previously uninhabited areas has definitely 2011). This is the case of the strong earthquake that struck
increased the geological–geomorphological risk in Ethiopia, the Kara Kore-Mejete area in 1961, with more than
a developing country where “social vulnerability” to natural 3,500 shocks from the end of May to the end of September
disasters is particularly augmented by economic, social, as recorded at the Geophysical Observatory in Addis Ababa
political, and cultural constraints (Alcántara-Ayala 2002). (Gouin 1979). Two shocks had a magnitude >6.4 and seven
This chapter presents an overview of the geo-hazards were stronger than >5.0. The area affected was estimated to
(including earthquakes, ground fissuring, volcanic eruptions, be about 300,000 km2 and relatively higher intensities were
flooding, and landsliding) that affect Ethiopia with particular observed in the southeastern sector of the zone. In the
20 Geo-hazard in Ethiopia 353
Fig. 20.2 Geological scheme of

the Horn of Africa (1 Quaternary
alluvial deposits; 2 Quaternary
volcanics; 3 Oligocene–Pliocene
volcanics; 4 Cretacic continental
and marine successions; 5 Late
Jurassic marine and continental
successions; 6 Precambrian-Late
Proterozoic intrusive rocks and
Late Paleozoic-Triassic
conglomerates and tillites; 7 Main
faults)
epicentral area, the village of Mejete was completely

destroyed and the village of Kara Kore was heavily damaged 20.3 Ground Fissures in the Main
with an estimated intensity of VIII–IX on the Mercalli- Ethiopian Rift
modified scale (Gouin 1979). Ground cracks, fissures, rock
falls, and subsidence of up to 1 m deep developed on the Ground fissures up to some kilometers long (Fig. 20.5)
Addis Ababa—Mekelle road near the town of Kara Kore, recurrently open in the MER floor consisting of up to more
heavy damaging culverts and retaining walls. Gravitational than 100 m thick, unconsolidated lacustrine sediments
movements were observed on steep escarpments and a (Gouin 1971; Asfaw 1982). These features consist of a series
15–20 km long fissure, in places 6–7 m deep, formed in of subsidence pits, up to 7 m wide and several tens of meters
unconsolidated soil along the eastern scarp of the Borkena deep, connected by cracks ranging in width from a few
River marginal graben (Fig. 20.3). A more than 2 m high centimeters to a few meters.
piedmont scarplet formed in unconsolidated materials along Taking the orientation of fissures into account, generally
the escarpment of the Borkena graben, reactivating a pre- parallel (NE–SW) or slightly oblique (NNE–SSW) to the
existing north–south trending normal/strike slip fault Rift axis, their occurrence may be related to the present NE–
(Fig. 20.4) over a distance of 12–15 km (Gouin 1979). SW extension of the MER. However, only in few cases, the
Considering that this displacement is very large in relation to genesis of fissures in the MER floor has been related to
the earthquake magnitude (M = 6.6), it can be explained by seismic events (Gouin and Mohr 1967; Asfaw 1982) and,
the contribution of a strong gravity stress related to the high even in these cases, their opening during earthquakes has
difference of relief between the Ethiopian Plateau and the never been directly observed (Asfaw 1998) nor confirmed.
Afar (Chorowicz et al. 1999). A much more frequent and immediate triggering factor of
Numerous coseismic surface breaks, including scarplets ground fissures is heavy, long-lasting rain (Asfaw 1998) that
with vertical throws up to 30 cm high and open fissures up to induces ground collapse, pit subsidence, and enlargement of
30 cm wide, were observed in central Afar during the first cracks on top of underground tunnels excavated by
two days of August 1989 earthquake of Dôbi graben groundwater running along tectonic fractures possibly
(M = 6.3) (Jacques et al. 2011). affecting the deep-seated rocky basement. In many cases, the
Fig. 20.3 Panoramic view of the Borkenna graben in the Kara Kore area. Coseismic deformation affected the road on the right of the image due to
the 1961 seismic sequences
Fig. 20.4 The Kara Kore fault:

a quarry in the Kara Kore area
where the contact between the
basalt (right) and slope/alluvial
deposits (left) is exposed due to
the fault activity; b detail of slope
deposits displaced by the fault
activity and tilted (the white
dashed line marks the altitude of
the deposits) toward the fault
plane (indicated by red arrows)
and c dragged (as indicated by
yellow dashed lines) along the
fault plane (indicated by red
arrows)
fissures have been associated with groundwater flow as near the town of Metehara (Gouin 1971) and in 1966, when
inferred from the noise of rushing water at depth. On the the same railroad was interrupted for a few days near the
other hand, fissures in hard rocks, associated to Quaternary town of Wolenchiti (Gouin and Mohr 1967). In the fol-
faults parallel to the Rift (Fig. 20.6), have been reported lowing years, other relevant fissure opening affected com-
from several places of the MER (Gibson 1967; Williams munication lines (such as the road from Ziway to Butajira)
et al. 2004). and urban settlements (such as the village of Moto, the
Quite different is the explanation provided by Ayalew Awasa Melka area, the surroundings of Meki, and the town
et al. (2004), according to whom the ground fissures form by of Muleti in the Awasa basin) causing destruction of houses
the combined action of groundwater flow, aseismically (Fig. 20.7) and loss of property (Asfaw 1982, 1998).
released elastic strain, piping, and hydrocompaction. There is no doubt that the occurrences of fissures in the
The first reports on damage caused by ground fissures in MER are actually many more than those reported so far
the MER were provided in 1956 by railway engineers who since most of these features occurred in desert areas. As far
observed them across the Addis Ababa—Djibouti railroad as settlement programs develop and communication lines
infringe on previously uninhabited areas, the reports will

likely become more and more frequent.
20.4 Volcanic Eruptions
Also, volcanic activity is associated with the rifting. This is

the case of the traps, an extensive pile of basalt lava flows,
covering an area of ca. 600,000 km2 and up to more than
2,000 m thick, emplaced during the Oligocene on a large
part of Ethiopia. Volcanic activity has followed during
Miocene, Pliocene, and Quaternary, with the eruption of
basalt and subordinate rhyolitic, trachytic, and phonolitic
materials, resulting in major topographic changes and the
formation of a number of lake basins. Linear basaltic fissure
vents dominate the rift axis while point source basaltic vents,
such as those located around Debre Zeit, are located as far as
7 km from the rift axis. Off axis central volcanic complexes
host the products of young eruptions that are as far as 12 km
from the center.
In the last centuries, several volcanic eruptions have been
recorded in the MER and the Afar Triangle (Mohr 1962).
According to a local tradition, a huge basalt flow originated
in 1810 from the southwestern flank of Mt. Fentale, a
presently dormant stratovolcano located in the northernmost
part of the MER (Fig. 20.8). Other volcanic products, rec-
ognized in the MER at Galibardi Pass and Mt. Bosetti, could
be related to the same period (Mohr 1962). In June 1907, a
relevant eruptive episode occurred in the Afdera Volcano, an
isolated stratovolcano in northeastern Ethiopia. The emitted
Fig. 20.5 Ground fissure at Muleti in the Awasa basin MER
Fig. 20.6 Fissure in volcanic

rocks in the MER near Metehara
(Photograph by V. Acocella)
reaching ash plume that disrupted the airline traffic in the

region. The eruption killed seven persons and caused the
displacement of thousands of people. A series of earthquake
shocks, including two with magnitude 5.7, struck the region
in the hours preceding the eruption.
20.5 Flooding
Floods frequently occur in Ethiopia during the main rainy

season (Kiremt season—See also Chap. 3, this volume)
which extends from mid‐June to mid‐September. Overflow
of rivers from their banks usually happen as a result of heavy
rains (Fig. 20.10). This type of flooding usually grows up
slowly with magnitude depending on many factors such as
watershed area, rainfall intensity and amount, topography,
sediment deposition in river beds, presence/absence of
vegetation, and human activity.
Where there is little settlement and the main economic
activity is pastoralism, the damage is low and floods are even
well accepted by the population since they improve grazing
and water supply for livestock and people. The only major
problems are related to road interruptions, sometimes cou-
pled with the collapse of bridges (Fig. 20.11). On the other
hand, flooding may be extremely dangerous in densely
inhabited areas as in the case of the Awash River basin
whose flood plain includes about 70 % of the country’s
large-scale irrigated agriculture. It is estimated that almost all
of the area delineated for irrigation development may be
Fig. 20.7 House damaged by a ground fissure at Meki (Photograph by subject to floods with a submergible surface of ca. 200,000–
V. Acocella) 250,000 ha (Alemayheu 2007). Uncontrolled deforestation
and expansion of farmlands have induced in this basin
lava flow was about 5 m thick and was accompanied by a widespread soil erosion (Bishaw 2001; Tadesse 2012)
sequence of seismic shocks. The most hazardous of these resulting in heavy sediment deposition in the stream beds.
events was the explosive eruption of the Dubbi Volcano For this reason, the Awash River in the Lower Plains has a
which occurred on May 7, 1981, in southern Eritrea, close to very unstable course tending to change its channel pattern
the Ethiopian border, killing more than one hundred people. and to branch out into the adjoining alluvial plains. Siltation
One volcanic center erupted in the Afar Triangle during has caused a reduction of more than 40 % (Kefyalew 2003)
the first decade of this century: the Erta’ Ale (Fig. 20.9), a of the capacity of Koka Dam reservoir, which is used to
basaltic shield volcano with a NNW–SSE elongated caldera supply water for irrigation, for hydropower generation, and
hosting a rising lava lake. A major lava overflow occurred also for flood lamination to protect the downstream plains.
on September 25, 2005, killing 250 heads of livestock and Other areas where significant river flooding occurs are the
forcing thousands of nearby residents to move away. A lower reaches of the Wabe Shebele, Genale, and Dawa rivers
further lava flow occurred in August 2007, forcing the near the Somali order, the lowlands areas bordering the
evacuation of hundreds of people and probably killing two Baro, Gilo, and Akobo rivers in western Ethiopia, the
persons. More recent eruptions were reported on November downstream areas of Omo River, the extensive fields sur-
4, 2008, and November 27–28, 2010. The Dabbahu and rounding Lake Tana and the Gumara and Rib rivers.
Mand Hararo volcanoes started to erupt basalt flows, ash, Flooding in urban settlements, such as Addis Ababa,
and smoke plumes along a more than 70 km long north– causes recurrent damages to built-up structures and properties
south fissure. The Nabro Volcano, a stratovolcano thought to along streams coming down from the adjacent mountains.
be extinct, started erupting on June 12, 2011 in southern Much more dangerous for human lives and properties,
Eritrea, not far from the Ethiopia borderline, with the because of their suddenness and high speed, are the flash
emission of a huge amount of sulfur dioxide and a high floods that recurrently occur in many parts of Ethiopia as the
Fig. 20.8 The Fentale volcano

Fig. 20.9 The Erta’ Ale volcano

result of localized intense rainfall capable to generate high occurred in the same town in April 1981, April 1994, and
peak discharges in river channels. These events are favored May 2005 (Alemayheu 2007, Billi et al. 2015).
by the rugged topography of the country characterized by
steep slopes and river channels, sparse vegetation, and poor
infiltration capacity of soil (Abebe 1997). 20.6 Landslides
In the town of Dire-Dawa, the flash flood event of August
5–6, 2006, killed more than 250 people and caused heavy Landslides are widespread in Ethiopia (Ethiopian Institute of
damage to public infrastructures including flood defenses, Geological Surveys 1994, 1995; Asrat et al. 1996; Gezahegn
housing, and livelihoods (FDPPA 2007). Other major floods 1998; Ayalew 1999, 2000; Ayalew and Vernier 1999;
Fig. 20.10 Flooding of a small

river in the Rift valley, south of
Langano Lake during the rainy
season (Photograph by P. Billi)
Fig. 20.11 Bridge collapsed

because of flooding (Golina
River)
Temesgen et al. 1999, 2001; Ayalew and Yamagishi 2002, The main predisposing factor of landslides is high relief
2004; Nyssen et al. 2002, 2003; Woldearegay et al. 2004; (ranging on average between 2,000 and 3,000 m a.s.l.)
Fubelli et al. 2008a; Moeyersons et al. 2008; Abebe et al. induced by the Pliocene–Quaternary uplift (Almond 1986;
2010; Broothaerts et al. 2012) where they cause significant Mohr 1986) and the resulting rugged morphology with deep
damage. From 1993 to 1998, more than 200 houses were valleys and gorges, where river incision causes further slope
destroyed, more than 500 km of roads were interrupted, and steepening and instability. Steep, unstable escarpments, up
about 300 people were killed (Ayalew 1999). to several hundred meters high, are also produced by active
faults on the rift margins (Chorowicz et al. 1999; Abebe in Ethiopia are reported by Gouin (1979), among which
et al. 2007; Fubelli et al. 2008a). were the huge slope failure which destroyed the town of
Triggering factors are heavy rainfalls and, to a lesser Ankober in 1842 and the fast-moving landslides that killed
extent, earthquakes. Intense, long-lasting rains commonly two people in Dese town during the 1977 earthquake.
occur in July and August (Gamachu 1977; Ethiopian Map- The nature of the outcropping bedrock, together with the
ping Authority 1988; Chap. 3, this volume) resulting in high intensity and duration of rainfall/earthquake shaking, con-
rates of water infiltration into the ground, widespread slope trols the typology and size of failure: (1) on steep slopes in
wash, overflooding, and riverbank erosion. Rain infiltration hard bedrock, such as basalt, welded ignimbrite, limestone
plays a primary role in triggering landslides by rising the and sandstone, fast-moving mass movements like rock falls
water table, thus increasing the total weight of the slope (Fig. 20.12), topplings, and rock slides/avalanches (Varnes
material and the pore pressure in fine-grained deposits 1978; Dikau et al. 1996; Ayalew 2000) are common; (2) on
(Brand et al. 1984; Ayalew 1999; Gabet et al. 2004). The steep slopes in weathered pyroclastics and volcanic
infiltration of rain water into fractured bedrock may be agglomerates first generation rapid landslides, such as debris
enhanced by the presence of deep desiccation cracks in the slides (Fig. 20.13) and avalanches, debris flows, earth flows
covering soil, which develop especially at the end of the dry and mudflows frequently occur mobilizing the soil cover
season (Billi and Dramis 2001, 2003). (Varnes 1978; Dikau et al. 1996; Temesgen et al. 1999,
Landslides are the most common surface effects of Fubelli et al. 2008a); (3) rapid failures, such as topples and
earthquakes, even in case of moderate magnitudes (Youd slumps (Varnes 1978; Dikau et al. 1996), typically affect the
1978; Harp et al. 1986; Ishihara and Hsu 1986; Cotecchia alluvial banks of deeply incised rivers and gullies; (4) on
1987). The main triggering agents of coseismic landsliding clayey materials, such as shaly–marly formations, alluvial–
are: oriented accelerations; earthquake induced changes in colluvial deposits and thick weathered layers on volcanic
the groundwater table; cyclic changes of pore pressures and rocks, simple and multiple retrogressive rotational slides,
liquefaction; seismo-tectonic surface fracturing and faulting. sometimes passing to earthflows or mudflows (Varnes 1978;
The latter agents seem also able to induce landsliding even Dikau et al. 1996) are commonly generated; (5) fast-moving
long after the seismic event, by favoring rain water infil- translational rock slides moving along predisposed yielding
tration (Chiodo et al. 1999; Dramis and Blumetti 2005). The surfaces (Varnes 1978; Dikau et al. 1996) can be expected
triggering of landslides, as well as other seismo-gravitational where soft clayey beds, dipping downslope, are overlain by
effects, is also favored by litho-structural contacts and hard rocky layers; (6) slow-moving translational slides with a
geomorphological settings (e.g., narrow ridges and valleys, step-like evolution, characterized by long quiescence periods
and peaks), which may locally amplify seismic shaking and short reactivation phases after heavy rainfalls, may
(Dramis and Blumetti 2005). Seismically induced landslides involve thick alluvial–colluvial deposits with intercalated
Fig. 20.12 Rockfall-derived

boulder from the rim of the
Wendo Genet caldera
water infiltration into the ground from leaking aqueducts,

sewerages, and irrigation ditches on cultivated slopes
(Temesgen et al. 1999; Ayalew 2000; Nyssen et al. 2002,
2003; Abebe et al. 2010).
20.6.1 The Case of the Dese Urban Area
One Ethiopian area heavily affected by landslides is the

urban territory of Dese (Welo) (Ethiopian Institute of Geo-
logical Surveys 1995; Ayalew and Vernier 1999; Terefe
2001; Ayenew and Barbieri 2005; Fubelli et al. 2008a), a
medium-sized urban settlement with about 200,000 inhab-
itants located at ca. 1,800 m a.s.l. in a small tectonic
depression “hanging” 800 m over the more eastern depres-
sion of Kombolcha (Figs. 20.14 and 20.15). This depression
is part of a basin and range landscape produced on the
eastern margin of the northern Ethiopian Plateau by the
Cenozoic extension of northeast Africa (Boccaletti et al.
1998; Ukstins et al. 2002). The outcropping bedrock consists
of a sequence of variably weathered Tertiary ignimbrites and
lava flows (Tefera et al. 1996), mantled by colluvial and
alluvial deposits mainly derived from the two north–south
trending fault escarpments that border the basin (Tossa and
Azwa Gedel escarpments). Besides the main border faults,
southwest–northeast trending transfer faults are also present
on the northern and southern edges of the basin (Fig. 20.16).
An up to 100 m thick sedimentary sequence of alluvial-
swampy-lacustrine deposits fills the lower part of the basin
floor (Ayenew and Barbieri 2005). These deposits are deeply
incised by the Borkena River, forming a wide terrace over
Fig. 20.13 Debris slide at Wendo Genet. Note people climbing the
which a wide sector of Dese town is located. Two alluvial
talus for scale
fans, formed by right-side tributaries of the Borkena River,
are also occupied by a part of the Dese settlement. Both the
finer horizons (Ayalew 2000); (7) first generation, fast- terrace and the alluvial fans are connected to the Tossa and
moving shallow landslides, such as soil slips or mud flows, Azwa Gedel escarpments by up to 15 m thick talus belt.
commonly mobilize eluvial–colluvial material (Abebe et al. A large number of landslide events have repeatedly struck
2010); (8) where thick layers of hard rock overlie clayey this area since long, with heavy impact on buildings and
formations, large-scale lateral spreading phenomena, moving infrastructure and loss of human lives (Fubelli et al. 2008a).
at an extremely slow rate, are present (Dramis and Sorriso- The main triggering factor is long-lasting intense rainfall
Valvo 1994; Dikau et al. 1996; Coltorti et al. 2009); (9) (Ayalew 1999; Ayenew and Barbieri 2005) and, to a lesser
clayey/marly formations overlain by fractured rocky aquifers extent, the earthquakes which recurrently strike the area
(such as basalt, limestone, or sandstone) that may undergo (Gouin 1979). The latter triggers rapid slope failures such as
sudden water table fluctuations during heavy rain are par- rock falls and debris flows, as those occurred during the
ticularly prone to failure (Ayalew 2000); (10) groundwater seismic event of July 8, 1988, causing the death of 2 persons
pressure and seepage may trigger mass movements in col- (Gouin 1979). Man-made activities, such as the construction
luvial deposits covering the hydrogeological contact between of houses, roads, bridges, or leakage of water from aque-
a hard rock cliff and an underlying clayey/marly aquiclude ducts, also induce slope instability by adding load to incip-
(Ayalew 2000; Nyssen et al. 2003). ient landslides, modifying slope profiles, or changing
Finally, an increasing impact of human activities, such as groundwater levels and flow.
intensive agriculture, quarrying, road construction, and A detailed GIS-supported geomorphological analysis of
urbanization, is also inducing slope instability by overload- the Dese area (Fubelli et al. 2008b) shows that landslide
ing the slope, modifying the slope profile, and increasing incidence and typology vary with the different landform

Dese basin (after Fubelli et al.
2008a, modified)
Fig. 20.15 An overview of the

Dese basin. The escarpment in the
back is ca. 400 m high
units making up the basin. Rapid mass movements such as

rock falls and topplings, rock slides (Fig. 20.17), debris 20.7 Hazard Assessment and Risk
flows, and debris slides triggered by heavy rain or earth- Mitigation Measures
quakes, characterize the major north–south and southwest–
northeast fault escarpments, the upper parts of the talus belts Notwithstanding the growing information on the occurrence
placed at the foot of the escarpments, and the upper catch- of catastrophic earthquakes, volcanic eruptions, and land-
ment slopes of the Borkena River tributaries. The escarp- slides in Ethiopia and the valuable papers published on these
ment talus belts are characterized by high hazard levels due topics, the response of public administrators and decision
to deep-reaching translational earth slides (Fig. 20.18). makers is still inadequate though some research projects and
Finally, rotational slides (Fig. 20.19) affect the erosional new regulations have been launched on the assessment of
escarpments of the terraced area and the incised alluvial fans geo-hazards and related risk at a regional scale (IUGS WGL-
where a large part of the town is located. CRA 1997).
1960s after the catastrophic earthquakes of Kara Kore-

Mejete (1961) and later in Serdo (1969). The first seismic
hazard maps were prepared by Gouin (1976) using a prob-
abilistic approach previously applied in North America
(Cornell 1968). Subsequently, maps of earthquake epicenters
distribution and computerized analyses of intensity and
acceleration as functions of time were produced (Asfaw
1988, 1990). The first seismic building code was introduced
in Ethiopia in 1980 (Engeda et al. 2011). It defines four
seismic zones with a return period of 100 years and 90 %
probability of non-exceedance. Following the works of
Kebede (1996), Kebede and Asfaw (1996), Panza et al.
(1996), and Kebede and Eck (1997), a new seismic hazard
zoning map of Ethiopia, taking into account the ground
motion attenuation, was produced.
Attempts of seismic zoning at a detailed scale and seis-
mic response at site level were performed for some impor-
tant areas such as the city of Addis Ababa (IDNDR
RADIUS Project 1999; Mammo 2005; Yoseph and Ramana
2008). Between 2001 and 2003, the Ethiopia Afar Geosci-
entific Lithospheric Experiment (EAGLE) international
research project was carried out to explore the kinematics
and dynamics of the lithosphere using a broadband seismic
array (Nyblade and Langston 2002; Bastow et al. 2011).
Another noteworthy research project involves the Afar Rift
Fig. 20.16 Fault pattern of the Dese basin (after Fubelli et al. 2008a, Consortium, funded by the UK Natural Environment
modified) Research Council with partners from UK, Ethiopia, France,
New Zealand, and the USA. Its aim was to conduct
experiments in the Afar area to better understand the
20.7.1 Earthquakes seismo-tectonic processes active there. More in general, an
accurate delineation of seismogenic belts associated with
The need for earthquake hazard assessment in Ethiopia was active faults, fissures, subsidence, and volcanism is an
initially recognized by the scientific community in the early indispensable basis for future hazard assessments.
Fig. 20.17 The crown of a rock

slide in weathered basalt near
Dese
Fig. 20.18 Trench of a deep-

reaching translational debris slide
at Dese
Fig. 20.19 Rotational shallow

earth slide in the Dese basin
20.7.2 Ground Fissures mapping of ground cracks and incipient fissures by air
photo analysis, detailed field survey, and geophysical
The random occurrence of ground fissures after heavy rains methods is an indispensable task to be performed before
and their destructive potential clearly characterizes them as establishing new settlements and transportation lines
natural hazards. In this context, the identification and (Asfaw 1998).
20.7.3 Volcanic Eruptions at the base or at the edge of high fault escarpments in
earthquake prone areas, are also exposed to risk.
Concerning volcanic hazard and risk mitigation, very little In the last years, detailed geomorphological–geotechnical
has been done so far in Ethiopia even if more than 9 million studies on landslide susceptibility/hazard of selected areas in
people live within 30 km of a potential eruptive center and at Ethiopia have been published (Temesgen et al. 2001; Aye-
least 5–10 of the 65 Holocene volcanoes recognized in the new and Barbieri 2005; Fubelli et al. 2008b; Abay and
country have been classified as “high risk” by the Global Barbieri 2012; Zvelebil et al. 2010) or performed as Master
Facility for Disaster Risk Reduction (GFDRR); a World and PhD thesis (Terefe 2001; Woldearegay 2005; Weldeg-
Bank funded pilot study that aims to establish science-based iorgis 2008; Asfaw 2010; Jbrahim 2011; Suyum 2011; Ha-
evidence for better integration of volcanic risks (Sparks et al. gos 2013).
2012). At least in some cases, explosive activity with A preliminary assessment of landslide risk in the Dese
pyroclastic flows and gas emission is expected. In particular, urban area has been recently published by Fubelli et al.
there are, so far, no systematic studies, monitoring, and (2013). These works are, however, just a small step in a large
mitigation plans for active volcanic centers and zones of the populated country such as Ethiopia, where the official geo-
MER and the Afar Triangle, notwithstanding the growing logical maps, published by the Geological Survey, cover the
development of urban settlements, infrastructure, and eco- whole territory only at the 1:2,000,000 scale (Kazmin 1972)
nomic activities in these areas. and only a few parts of the country are mapped at the
1:25,000 scale.
20.7.4 Flooding
20.8 Conclusions
Flood management and control in Ethiopia is not yet at an
advanced stage. The only major interventions have been The authors hope that this general overview of geo-hazards
carried out in the Awash River Basin with the implemen- in Ethiopia could help to make people more aware of their
tation of irrigation networks protected from floods by dykes impact on the environment, built-up areas, infrastructure,
and the establishment of a Project Control Center, equipped economy, and human lives. Further investigation on seis-
with adequate maintenance facilities. Moreover, the Gov- micity, volcanism, and slope stability, with particular refer-
ernment established a River Training Unit for the Lower ence to hazard zoning and risk mitigation at a local scale, are
Awash Plains, equipped with facilities to remove silt from mandatory. A very weak point in this context is the absence
the river beds. The Unit also constructs and maintains flood of medium-scale geological maps covering the entire Ethi-
protection dykes along the river. opian territory, notwithstanding the availability of good
Watershed management projects have been recently 1:50,000 topo-sheets.
started for different parts of Ethiopia. They include the Other imperative actions include making use of institutional
emplacement of check-dams and weirs for water conserva- monitoring facilities with state-of-the-art instruments; training
tion and retaining floods, the construction of terraces along specialized personnel (engineers; geologists, geophysicists;
mountain slopes to reduce runoff and favor ground water geomorphologists) at the MSc and PhD level to be employed in
recharge. public institutes and agencies; improving research activity also
Moreover, in the last years, an extensive project entitled in cooperation with international research groups; organizing
Flood Preparedness and Early Warning System has been scientific conferences; disseminating proper information by
launched in the Eastern Nile countries (Sudan, Ethiopia, and school teaching, village seminars, and TV broadcasts able to
Egypt). The planned activities, when accomplished, should reach even the most remote villages.
enable these countries to improve their capacity of mitigat-
ing flood risk.
References
20.7.5 Landslides
Abay A, Barbieri G (2012) Landslide susceptibility and causative
factors evaluation of the landslide area of Debresina, in the
Very high-risk levels, due to the possible occurrence of rapid southwestern Afar escarpment, Ethiopia. J Earth Sci Eng 2:133–144
mass movements related to heavy rainfall, are typical for Abebe B (1997) Flood vulnerability in Ethiopia and needs for
settlement areas located at the foot of steep slopes in preparedness. Paper for workshop vulnerability in Ethiopia from
disaster to development. Addis Abeba, June 23–25, 1997
weathered volcanic rocks or at the outlet of mountain
Abebe B, Acocella V, Korme T, Ayalew D (2007) Quaternary faulting
streams, from which concentrated flood flows and debris can and volcanism in the Main Ethiopian Rift. J Afr Earth Sci 48:
suddenly burst out. Settlements and infrastructures, located 115–124
Abebe B, Dramis F, Fubelli G, Umer M, Asrat A (2010) Landslides in Billi P, Dramis F (2003) Geomorphological investigation on gully
the Ethiopian highlands and the Rift margins. J Afr Earth Sci erosion in the Rift Valley and northern highlands of Ethiopia.
56:131–138 Catena 50:353–368
Alcántara-Ayala I (2002) Geomorphology, natural hazards, vulnerabil- Billi P, Alemu Y, Ciampalini R (2015) On the increased frequency of
ity and prevention of natural disasters in developing countries. flash floods in Dire Dawa, Ethiopia: climate change or human
Geomorphology 47:107–124 impact?, Natural Hazards, submitted
Alemayheu DD (2007) Assessment of flood risk in Dire Dawa town, Bishaw B. (2001) Deforestation and land degradation on the Ethiopian
eastern Ethiopia, using GIS. Unpublished Master thesis, Addis Highlands: a strategy for physical recovery. In: International
Ababa University Conference on African Development Archives, Paper 2. http://
Almond DC (1986) Geological evolution of the Afro-Arabian dome. scholarworks.wmich.edu/africancenter_icad_archive/2
Tectonophysics 131:301–332 Boccaletti M, Bovini M, Mazzuoli R, Abebe B, Piccardi L, Tortorici L
Asfaw EY (2010) Landslide assessment in Blue Nile Gorge, central (1998) Quaternary oblique extensional tectonics in the Ethiopian
Ethiopia. Master thesis, Universiteit Gent—Vrije Universiteit Rift (Horn of Africa). Tectonophysics 287:97–116
Brussel, Belgium Brand EW, Premchitt J, Phillipson HB (1984) Relationship between
Asfaw LM (1982) Development of earthquake induced fissures. Nature rainfall and landslides in Hong Kong. In: Proceedings of the 4th
286:551–553 international symposium on landslides, Toronto, September 17,
Asfaw LM (1988) Catalogue of Ethiopian earthquakes. In: Woldeyes G 1984, vol 1, p 377–384
(ed), SAREC-ESTC conference. Addis Abeba, pp 252–279 Broothaerts N, Kissi E, Poesen J, Van Rompaey A, Getahun A, Van
Asfaw LM (1990) Seismicity and earthquake risk in the Addis Abeba Ranst E, Diels J (2012) Spatial patterns, causes and consequences of
region. SINET -Eth J Sci 13(1):15–35 landslides in the Gilgel Gibe catchment, SW Ethiopia. Catena
Asfaw LM (1998) Environmental hazard from fissures in the Main 97:127–136
Ethiopian Rift. J Afr Earth Sci 27:481–490 Chiodo G, Dramis F, Gervasi A, Guerra I, Sorriso-Valvo M (1999)
Asrat A, Eshete G, Taddesse S, Getaneh A, Fekede, K (1996) Land Frane sismo-indotte e pericolosità di sito: primi risultati dello studio
mass movement of November 10, 1994 in Goffa District, Northern degli effetti di forti terremoti storici in Calabria centro-settentrio-
Omo Zone, Southern Ethiopia. Abstract 3rd Ethiopian Geoscience nale. Atti del XVIII Convegno Nazionale GNGTS, Roma, CD-Rom
and Mineral Engineering Congress, 15–17 November 1996, Addis Chorowicz J, Collet B, Bonavia F, Korme T (1999) Left-lateral strike-
Ababa slip tectonics and gravity induced individualisation of wide
Ayalew L (1999) The effect of seasonal rainfall on landslides in the continental blocks in the western Afar margin. Eclogae Geol Helv
highlands of Ethiopia. B Eng Geol Environ 58:9–19 92:149–158
Ayalew L (2000) Factors affecting slope stability in the Blue Nile basin. Coltorti M, Pieruccini P, Berakhi O, Dramis F, Asrat A (2009) The
In: Bromhead M, Dixon N, Ibsen M (eds) Landslides in research, geomorphological map of Mt. Amba Aradam southern slope
theory and practice. Thomas Telford, Cardiff, pp 101–106 (Tigray, Ethiopia). J Maps 2009:56–65
Ayalew L, Vernier A (1999) Causes and mechanisms of slope Cornell CA (1968) Engineering seismic risk analysis. B Seismol Soc
instability in Dessie town, Ethiopia. In: Yagi N, Yamagami T, Am 18:1583–1606
Jang J-C (eds) Slope stability engineering. A.A. Balkema, Rotter- Cotecchia V (1987) Earthquake-prone environments. In: Anderson
dam/Brookfield, pp 1181–1186 MG, Richards KS (eds) Slope stability. Wiley, Chichester, pp 287–
Ayalew L, Yamagishi H (2002) Landsliding and landscape develop- 330
ment; the case of northern Ethiopia. In: Proceedings of the Dikau R, Brunsden D, Schrott L, Ibsen ML (eds) (1996) Landslide
international congress intrapraevent 2002, Matsumoto, Japan, recognition. Identification movement and causes. Wiley, Chichester
pp 595–606 Dramis F, Blumetti AM (2005) Some considerations about seismic
Ayalew L, Yamagishi H (2004) Slope failure in the Blue Nile basin, as geomorphology and paleoseismology. Tectonophysics 408:177–
seen from landscape evolution perspective. Geomorphology 57:95– 191
116 Dramis F, Sorriso-Valvo M (1994) Deep-seated gravitational slope
Ayalew L, Yamagishi H, Reik G (2004) Ground cracks in Ethiopian deformations, related landslides, and tectonics. Eng Geol 38(3–4):
Rift Valley: facts and uncertainties. Eng Geol 75:309–324 231–243
Ayele A (1995) Earthquake catalogue of the Horn of Africa for the Engeda S, Kebede A, Tessema E (2011) Notes and proposed guidelines
period 1960–93. Seismology Department, Uppsala, Report 3–95 on updated seismic codes in Ethiopia. Implications for large-scale
Ayele A (1997) Spatial and temporal variations of seismicity in the infrastructures. J Eth Eng Arch 28:90–110
Horn of Africa from 1960–1963. Geophys J Int 130:805–810 Ethiopian Institute of Geological Surveys (1994) A report of
Ayele A, Jacques E, Kassim M, Kidane T, Omar A, Tait S, Nercessian engineering geological problems of studies of part of the Blue Nile
A, Chabalier J-B, King G (2007) The volcano–seismic crisis in Gorge (Goha Tsion-Dejen). EIGS, Addis Ababa
Afar, Ethiopia, starting September 2005. Earth Planet Sci Lett Ethiopian Institute of Geological Surveys (1995) A report on landslide
255:177–187 problems of Dessie town. EIGS, Addis Ababa
Ayenew T, Barbieri G (2005) Inventory of landslides and susceptibility Ethiopian Mapping Authority (1988) National atlas of Ethiopia. Addis
mapping in the Dessie area, Northern Ethiopia. Eng Geol 77:1–15 Ababa
Bastow ID, Keir D, Daly E (2011) The Ethiopia Afar Geoscientific Fairhead JD (1968) The seismicity of the East African rift system 1955
Lithospheric Experiment (EAGLE): probing the transition from to 1968. MSc Dissertation, University of Newcastle upon Tyne
continental rifting to incipient seafloor spreading. Geol Soc Am Faure H (1975) Neotectonics in the Afar (Ethiopia, T.F.A.I.). In:
Spec Paper 478:51–76 Suggate RP, Creswell MM (eds) Quaternary studies. Royal Society
Billi P, Dramis F (2001) Recent climatic changes and the spreading of of New Zealand, Wellington, pp 121–126
gully erosion in Ethiopia. In: Andah K (ed) Water resources FDPPA Federal Disaster Prevention and Preparedness Agency (2007)
management in a vulnerable environment for sustainable develop- Regional summary of multi-agency flood impact assessment of
ment. UNESCO-IHP-CNR, Grifo Publ, Perugia, pp 131–141 2006. Addis Ababa
Fubelli G, Abebe B, Dramis F, Vinci S (2008a) Geomorphological assessment and design of structures for earthquake resistance, Addis
evolution and present-day processes in the Dessie Graben (Wollo, Ababa, Ethiopia, February 21, 1996, pp 25–38
Ethiopia). Catena 75:28–37 Kebede F, Asfaw LM (1996) Seismic hazard assessment for Ethiopia
Fubelli G, Vinci S, Abebe B, Dramis F (2008b) Cartografia geomorf- and the neighboring countries. SINET Eth J Sci 19 (1):15–50
ologica di dettaglio e analisi statistica in ambiente GIS dei processi Kebede F, van Eck T (1997) Probabilistic seismic hazard assessment
di versante nel bacino di Dessie (Wollo, Etiopia). Mem Descrittive for the Horn of Africa based on seismotectonic regionalization.
Carta Geol It 78:92–99 Tectonophysics 270(3–4):137–331
Fubelli G, Guida D, Cestari A, Dramis F (2013) Landslide hazard and Kefyalew A (2003) Integrated flood management. WMO/GWP Asso-
risk in the Dessie Town area (Ethiopia). In: Margottini C, Canuti P, ciated Programme on Flood Management, Addis Ababa
Sassa K (eds) Landslide science and practice, vol 6. Springer, Mammo T (2005) Site-specific ground motion simulation and seismic
Berlin, pp 357–361 response analysis at the proposed bridge sites within the city of
Gabet EJ, Burbank DW, Putkonen JK, Pratt-Sitaula BA, Ojha T (2004) Addis Ababa, Ethiopia. Eng Geol 79(3–4):127–150
Rainfall thresholds for landsliding in the Himalayas of Nepal. Moeyersons J, Van Den Eeckhaut M, Nyssen J, Tesfamichael G, Van
Geomorphology 63:131–143 de Wauw J, Hofmeister J, Poesen J, Deckers J, Mitiku H (2008)
Gamachu D (1977) Aspects of climate and water budget in Ethiopia. Mass movement mapping for geomorphological understanding and
Addis Ababa University Press, Addis Ababa sustainable development: Tigray, Ethiopia. Catena 75:45–54
Gezahegn A (1998) Slope stability assessment in the Blue Nile Gorge, Mohr PA (1962) The geology of Ethiopia. Addis Ababa University
Ethiopia. In: Proceedings of the 8th international IAEG congress, College Press, Addis Ababa
September 21–25, 1998, Vancouver, Canada, pp 1437–1442 Mohr PA (1967) The Ethiopian rift system. B Geophys Observ Addis
Gibson IL (1967) Preliminary account of the volcanic geology of Ababa 11:1–65
Fentale, Shoa. B Geophys Observ, Addis Ababa 10:59–67 Mohr P (1986) Sequential aspects of the tectonic evolution of Ethiopia.
Gouin P (1971) Surface cracks and subsidence in Mojo. Report TR- Mem Soc Geol It 31:447–461
011. Ministry of the Interior, Addis Ababa, 5 p Nyblade AA, Langston CA (2002) Broadband seismic experiments
Gouin P (1970) Seismic and gravity data from Afar in relation to probe the East African Rift. EOS Trans AGU 83(405):408–409
surrounding areas. Philos T Roy Soc A 267:339–358 Nyssen J, Moeyersons J, Poesen J, Deckers J, Mitiku H (2002) The
Gouin P (1976) Seismic zoning in Ethiopia. B Geophys Observ, Addis environmental significance of the remobilisation of ancient mass
Ababa 7:1–46 movements in the Atbara-Tekeze headwaters, Northern Ethiopia.
Gouin P (1979) Earthquake history of Ethiopia and the Horn of Africa. Geomorphology 49:303–322
International Development Research Center (IDRC), Ottawa Nyssen J, Poesen J, Moeyersons J, Deckers J, Mitiku H, Lange A
Gouin P, Mohr PA (1967) Recent effects possibly due to tensional (2003) Human impact on the environment in the Ethiopian and
separation in the Ethiopian rift system. B Geophy Observ, Addis Eritrean highlands—a state of the art. Earth Sci Rev 64:273–320
Ababa 10:69–78 Panza GF, Vaccari F, Costa G, Suhadolc P, Fäh D (1996) Seismic input
Gutenberg B, Richter CF (1954) Seismicity of the Earth and associated modelling for zoning and microzoning. Earthq Spectra 12:529–566
phenomena. Princeton University Press, Princeton Sparks S, Aspinall W, Auker M, Crosweller S, Hincks T, Mahony S,
Hagos AH (2013) Remote sensing and GIS-based mapping on Nadim F, Pooley J, Syre E (2012) Mapping and characterizing
landslide phenomena and landslide susceptibility evaluation of volcanic risk. In: Abstracts of the magmatic rifting and active
Debresina area (Ethiopia) and Rio San Girolamo basin (Sardinia). volcanism conference, Addis Ababa, January 11–13, 2012
Doctoral thesis, University of Cagliari (Italy) Suyum GM (2011) Landslide mapping assessment using GIS tech-
Harp EL, Wilson RC, Keefer DK, Wieczorek GF (1986) Seismically niques in Dessie area, Northern Ethiopia. Unpublished Master
induced landslides: current research by the U.S. Geological Survey. thesis, Universiteit Gent-Vrije Universiteit Brussel, Belgium
In: Proceedings of the international symposium on engineering Sykes LR, Landisman M (1964) The seismicity of East Africa, the Gulf of
geology problems in seismic areas, Bari, Italy, vol 2, pp 159–173 Aden and the Arabian and Red Seas. B Seismol Soc Am 54:1927–1940
IDNDR RADIUS Project (1999) Addis Ababa case study, final report, Tadesse Y (2012) Socioeconomic impacts of flooding in Dire Dawa,
prepared by the Addis Ababa RADIUS Group Ethiopia: flood triggering factors, social, infrastructural and eco-
Ishihara K, Hsu H-L (1986) Landslides in natural slopes during nomic impacts of flooding. LAP—Lambert Academic Publishing,
earthquakes. In: Proceedings of the international symposium on Saarbrücken
engineering geology problems in seismic areas, Bari, Italy, vol 5, Tefera M, Chernet T, Haro W (1996) Explanation of the geological
p 273–298 map of Ethiopia, 2nd edn. EIGS, Addis Ababa
IUGSWGL-CRA (Working Group on Landslides—Committee on Risk Temesgen B, Umer M, Asrat A, Berakhi O, Ayele A, Dramis F (1999)
Assessment) (1997) Quantitative risk assessment for slope and Landslide hazard on the Debicho, Wendo Genet area: the case of
landslides—the state of the art. In: Cruden D, Fell R (eds) Landslide June 18, 1996 event. SINET Eth J Earth Sci 22:127–140
risk assessment, Balkema, Rotterdam/Brookfield, pp 3–12 Temesgen B, Umer M, Korme T (2001) Natural hazard assessment
Jacques E, Kidane T, Tapponnier P, Manighetti I, Gaudemer Y, Meyer using GIS and remote sensing methods, with particular reference to
B, Ruegg JC, Audin L, Armijo R (2011) Normal faulting during the the landslides in the Wondogenet Area, Ethiopia. Phys Chem Earth
August 1989 earthquakes in central Afar: sequential triggering and PT C 26(9):665–675
propagation of rupture along the Dôbi Graben. B Seismol Soc Am Terefe K (2001) Engineering geological mapping and landslide assessment
101(3):994–1023 of Dessie town. MSc thesis, Addis Ababa University, Addis Ababa
Jbrahim J (2011) Landslide assessment and hazard zonation in Mersa Ukstins IA, Renne PR, Wolfenden E, Baker J, Ayalew D, Menzies M
and Wurgessa, north Wollo, Ethiopia. MSc thesis, Addis Ababa (2002) Matching conjugate rifted margins: 40Ar/39Ar chrono-
University, Addis Ababa, Ethiopia stratigraphy of pre- and syn-rift bimodal flood volcanism in
Katzmin V (1972) Geological map of Ethiopia. EIGS, Addis Ababa, Ethiopia and Yemen. Earth Planet Sci Lett 198:289–306
Ethiopia Varnes DJ (1978) Slope movements types and processes. In: Schuster
Kebede F (1996) Seismic hazard assessment for the Horn of Africa. In: RL, Krizek RJ (eds) Landslides, analysis and control. Transporta-
Proceedings of the Ethiopian Association of Seismology and tion Research Board, Special Report, vol 176, National Academy of
Earthquake Engineering (EASEE)—workshop on seismic hazard Sciences, Washington DC, pp 11–33
Weldegiorgis H (2008) Landslide hazard zonation mapping in Blue Worku A (2011) Recent developments in the definition of design
Nile gorges. MSc thesis, Addis Ababa University, Addis Ababa earthquake ground motion. Calling for a revision of the current
Williams FM, Williams MAJ, Aumento F (2004) Tensional fissures Ethiopian seismic code. EBCS 8: 1995. J Eth Eng Arch 28:1–15
and crustal extension rates in the northern part of the Main Yoseph B, Ramana G (2008) Probabilistic seismic hazard analysis at
Ethiopian Rift. J Afr Earth Sci 38:183–197 selected sites in Addis Ababa, Ethiopia. Int J Geotech Eng 2
Woldearegay K (2005) Rainfall-triggered landslides in the northern (2008):129–141
highlands of Ethiopia: characterization, GIS-based prediction and Youd TL (1978) Major cause of earthquake damage is ground failure.
mitigation. PhD thesis, Graz University of Technology Civil Eng 48(4):47–51
Woldearegay K, Riedmueller G, Schubert W, Mogessie A (2004) Zvelebil J, Šíma J, Vilímek V (2010) Geo-risk management for
Controlling parameters and failure mechanisms of a large-scale developing countries—vulnerability to mass wasting in the Jemma
landslide in Ethiopia. Felsbau 22(3):46–55 River Basin, Ethiopia. Landslides 7(1):99–103
Land Degradation in the Ethiopian Highlands
21
Jan Nyssen, Jean Poesen, Sil Lanckriet, Miro Jacob, Jan Moeyersons,
Mitiku Haile, Nigussie Haregeweyn, R. Neil Munro,
Katrien Descheemaeker, Enyew Adgo, Amaury Frankl,
and Jozef Deckers
Abstract
The high soil erosion rates in the Ethiopian highlands find their causes in the combination of
erosive rains, steep slopes due to the rapid tectonic uplift during the Pliocene and Pleistocene,
and human impact by deforestation, overgrazing, agricultural systems where the open field
dominates, impoverishment of the farmers, and stagnation of agricultural techniques.
Travelling in the Ethiopian highlands, one can see many soil and water conservation
structures. Indigenous knowledge and farmers’ initiatives are integrated with these introduced
technologies at various degrees. This chapter addresses the status and drivers of land
degradation in northern Ethiopia, including changes over the last century.

Keywords
Desertification Soil erosion Slope processes Soil and water conservation
21.1 Introduction landsliding in the Ethiopian highlands find their causes in the
combination of erosive rainfall, steep slopes due to the rapid
The rugged landscapes of the Ethiopian highlands have been tectonic uplift during the Pliocene and Pleistocene, and
imprinted and partly degraded by agriculture since 3 mil- human impact by deforestation, overgrazing, agricultural
lennia at least (Nyssen et al. 2004b). This chapter particu- systems where the open field dominates, impoverishment of
larly addresses rainfall, runoff, and soil erosion processes. the farmers, and stagnation of agricultural techniques (Ståhl
The high soil erosion rates by water and tillage as well as by 1974; Girma and Jacob 1988; Ståhl 1990). In flat areas and
on stone-covered slopes (Nyssen et al. 2002b; Van de Wauw
et al. 2008), soil profiles have not yet been fully truncated by
J. Nyssen (&) S. Lanckriet M. Jacob A. Frankl soil erosion that is concomitant to tilled agriculture. Agri-
Department of Geography, Ghent University, Ghent, Belgium
e-mail: jan.nyssen@ugent.be
cultural practices are well adapted to the environment: the
mahrasha tillage tool (the traditional ‘ard’ plough) was
J. Poesen R.N. Munro J. Deckers
Department of Earth and Environmental Sciences, KU Leuven,
developed during the high-tech Axumite period; the crop-
Leuven, Belgium ping systems fit seamlessly to soil catenas (Nyssen et al.
J. Moeyersons
2008a); and the farming systems are well adapted to inter-
Royal Museum for Central Africa, Tervuren, Belgium annual variation in rainfall conditions (Pietsch and Machado
M. Haile N. Haregeweyn
2014). Whereas, technically, under the traditional circum-
Department of Land Resources Management and Environmental stances, agricultural adaptation to soil and climate variability
Protection, Mekelle University, Mekelle, Ethiopia is nearly optimal, land management has for long been
K. Descheemaeker hampered by unequal access to land and prevalent free
Department of Plant Sciences, Wageningen University, grazing. Most reports from the first half of the twentieth
Wageningen, The Netherlands century (e.g. Giglioli 1938a, b; Joyce 1943) recognised the
E. Adgo soil erosion problem but did not consider that it was a major
College of Agriculture and Environmental Science, Bahir Dar problem. Frankl et al. (2011) have shown that the gullies
University, Bahir Dar, Ethiopia
currently visible in the landscape started to develop in the installed in central Tigray during one year (1975 and 2001,
1960s. This chapter addresses the status and causes of land respectively) indicate that 30–70 % of all rain events had an
degradation in northern Ethiopia over the last century. intensity >25 mm h−1 (Hunting Technical Services 1976;
Nyssen et al. 2005). Krauer (1988) obtained from the rainfall
data of six Soil Conservation Research Programme (SCRP)
21.2 Rainfall and Runoff as Driving stations mean annual universal soil loss equation (USLE)
Forces for Soil Erosion rainfall erosivity indices R between 166.6 (Afdeyu, Eritrea)
Processes and 543.7 J cm m−2 h−1 year−1 (Anjeni, Gojam). Hurni
(1979), in an analysis of rainfall erosivity in the Simien
The climates of Ethiopia are complex: ‘Within short hori- Mountains, insisted on two other particularities of Ethiopian
zontal distances, climates from tropical to sub-humid and mountains: erosivity due to hail (2.5 times more important
sub-tropical to arctic can occur’ (Krauer 1988). Precipitation than erosivity due to rain) and the influence of hillslope
and air temperature vary mainly with elevation, but slope aspect. A soil surface unit exposed to wind receives a greater
aspect also plays an important role. Furthermore, precipita- quantity of water than a surface unit with an opposite
tion decreases and seasonality increases with latitude. exposure.
Given that rainfall characteristics in tropical highlands are
21.2.1 Precipitation Patterns different from those of more temperate climates, it is difficult
to apply erosivity equations, such as those proposed in (R)
From the end of June onwards, the Intertropical Convergence USLE (Wischmeier and Smith 1978; Renard et al. 1997),
Zone (ITCZ) is situated at its most northerly position (16°N– which have been developed for North America, to rainfall on
20°N). The southern air masses, limited to the lower layers of the Ethiopian highlands. Based on drop size measurements,
the atmosphere, bypass the highlands and reach them from Nyssen et al. (2005) showed that for the same rainfall
the south-west, giving way to the main rainy season (Goebel intensity, rainfall erosivity is significantly higher in the
and Odenyo 1984). Generally, clouds are formed at the end of Ethiopian highlands compared to elsewhere in the world
the morning, as a result of evaporation and convective cloud because of larger raindrop sizes, also during low intensity
formation due to daytime heating of the soil, and cause rains rain events. Moreover, in the absence of a network of
in the afternoon. In Afdeyu station, on the Eritrean highlands, automatic rain gauges, maximum hourly rainfall intensities
80 % of daily precipitation takes place between 12 and 16 h could be measured only in a small number of research sta-
(Krauer 1988). (All mentioned localities are indicated on tions in Ethiopia.
Fig. 21.1). This convective nature of rainfall also explains Rainfall erosivity is a function of the depths and inten-
why individual showers have a very local distribution. At the sities of the individual rainstorms, and these are not closely
end of the summer, the ITCZ returns quickly to the south, related to annual precipitation. However, the United States
preventing the arrival of monsoon rain. This is the end of the data indicate that for a given annual precipitation, the range
rainy season in the highlands. of likely erosivity values can be somewhat narrowed by
Abebe and Apparao (1989) calculated from 241 stations in knowledge of the general climatic conditions in the partic-
Ethiopia a mean annual precipitation of 938(±83) mm year−1. ular geographic area (Wischmeier and Smith 1978). In East
For the highlands, annual precipitation varies between Africa (i.e. Tanzania, Kenya, and Uganda), the relationship
450 mm year−1 in Tigray and more than 2,000 mm year−1 in between total precipitation and erosivity index improves if
the south-west of the country (Krauer 1988). The interaction rainfall stations are grouped by geographical area (Moore
of latitude and altitude controls total annual precipitation 1979). For Ethiopia, Hurni (1985) and Krauer (1988) elab-
(Troll 1970). At the regional scale, one should, however, also orated, from monthly data of 6 SCRP stations, correlations
take into account that during the rainy season winds come between USLE’s R-factor and mean annual rainfall and
essentially from the south-west, as well as orographic effects. Krauer (1988) presented an isoerodent (rain erosivity) map
Valleys are preferred flow paths for the penetration of humid of Ethiopia. More recently, several studies have reported rain
air masses into the highlands (Nyssen et al. 2005) and rainfall erosivity data for Ethiopian rain stations, as well as for other
distribution is highly erratic (Jacob et al. 2013). stations in Africa (e.g. Vrieling et al. 2010; Diodato et al.
2013).
21.2.2 Rainfall Erosivity in the Ethiopian

Highlands 21.2.3 Runoff and Infiltration
High rainfall erosivity is an important factor of soil erosion In Ethiopia, surface runoff production has been measured at
in the highlands. Data from two automatic rain gauges various temporal and spatial scales (from runoff plot to
21 Land Degradation in the Ethiopian Highlands 371
Fig. 21.1 Map of Ethiopia and Eritrea, with indication of localities Sina, 12 Dogu’a Tembien; summits are represented by triangles: a Ras
mentioned in this chapter. Minor localities indicated with numbers: 1 Dejen, b Ankober, c Amba Alage; open dots represent lakes: B Lake
Adama/Nazret + Debre Zeit, 2 Afdeyu, 3 Anjeni, 4 Kelafo, 5 Mustahil, Besaka, K Koka reservoir, L Lake Langano
6 May Makden, 7 Adwa, 8 Ambo, 9 Dizi, 10 Hunde Lafto, 11 Debre
catchment). Runoff has been monitored in the SCRP (<1,000 m2) runoff plots are very variable (0–50 %) (Nyssen
catchments and data series of up to 12 years are available et al. 2004b), which is attributed to the variable experimental
(SCRP 2000). Generally, runoff coefficients (RC) from small conditions. Besides different slope gradients, local
Fig. 21.2 Runoff coefficients 60

(RC) versus drainage area (A) for
catchments of the basins of the
Blue Nile, Tekeze, Wabe 50 Blue Nile
Shebele, Awash, and Rift Valley y = -4.1Ln(x) + 64.1
lakes (after Nyssen et al. 2004b) r2 = 0.34 n = 23 P < 0.01
40
RC (%)
30
20
Others
10 y = -3.83Ln(x) + 45.7
r2 = 0.75 n = 18 P < 0.001
0
100 1 000 10 000 100 000 1 000 000
A (km2)
Blue Nile Tekeze Wabe Shebele Awash and Rift Valley lakes
differences in soil texture, land use, vegetation cover, areas of the Blue Nile catchment along the border with
organic matter content, or rock fragment cover result in a Sudan, which reduces the overall runoff depth for the whole
wide range of infiltration rates obtained from runoff plots. catchment (Conway, personal communication 1999). In situ
Results on RC from runoff plots are therefore not repre- water harvesting and the construction of small reservoirs
sentative for RC of catchments. have both led to strongly decreasing RC at catchment scale,
For large catchments (A ≥ 100 km2), RC decreases with and to increased levels of the water tables (Nyssen et al.
increasing catchment area (Fig. 21.2). The already men- 2010; Berhane et al. 2013). However, significant differences
tioned conditions for high RC (presence of open field and in RC between the sub-catchments within the 5,000 km2
high rainfall) are mainly found in the Blue Nile basin. For Geba basin could not be demonstrated, most probably due to
this reason, two data series can be considered. The Tekeze, the overall implementation of soil and water conservation
Awash, and Wabe Shebele basins are mainly situated in dry (SWC) activities (Zenebe et al. 2013). Additional research
sub-humid to arid regions (Mersha 2000). In the Wabe on this topic is currently being conducted in Lake Tana basin
Shebele basin, Bauduin and Dubreuil (1973) explained a (Poppe et al. 2013; Dessie et al. 2014). Particularly, in cases
decreasing RC with an increasing catchment size by the fact of large-scale conversions of cropland and rangeland to
that small catchments are mostly situated in the headwaters forest, such as on the escarpment upslope from Alamata,
where nearly impervious, basalt-derived soils dominate, and effects are very clear, particularly in terms of decreased
also by a smaller mean annual basin precipitation in the downstream flooding and changes of hydrogeomorphology
larger catchments which include (semi)arid lowlands. The (i.e. river channel incision and narrowing) (Gebreyohannes
rainfall and runoff data for catchments in the Blue Nile basin et al. 2014).
suffer, according to its authors (USBR 1964), from the lack
of precision in delimiting drainage areas (A) for smaller
basins. Representative catchment precipitation data are also 21.3 Weathering and Soil Formation
difficult to obtain given poor station density and large spatial
variability of rainfall (Conway and Hulme 1993). Conway Few studies have been made on weathering of parent
(1997) pointed to short mean observation periods (i.e. material in the highlands. Hövermann (discussion in Bakker
1.5 years) and possible errors in rainfall data. Despite the 1967) studied the basal Precambrian granites in northern
wide scatter for the Blue Nile basin, it can be observed that Ethiopia where weathering mantles are up to 120 m
RC are larger than those for the other basins but that they deep. No studies exist for Mesozoic sedimentary rocks or for
still follow a parallel trend (Fig. 21.2). Decreasing RC values Tertiary volcanics, but the depth of weathering mantle is
with increasing A values in the Blue Nile basin are thought expected to be much less.
to be a result of (a) runoff transmission losses, due to Hurni (1983), through the study of soils developed on
evaporation and possibly lithological changes, and (b) less periglacial slope deposits, extrapolated soil formation rates
rainfall and larger potential evapotranspiration in the western for the different agroclimatological zones of Ethiopia.
Zonation in Ethiopia is based on altitude and more specifi-

cally on the corresponding local climate. These soil forma-
tion rates are mean rates, taking into account rainfall depth
and air temperature conditions, but not lithology. They are
intended to be compared with soil loss rates, but, to our
understanding, cannot be applied to the vast areas where the
soil mantle results from sediment deposition rather than from
pedogenesis.
21.4 Sheet and Rill Erosion
Most research on soil erosion in Ethiopia focused on sheet

and rill erosion (Fig. 21.3). Hurni (1975, 1978, 1979) studied
thoroughly the Jinbar valley (3,200–4,000 m a.s.l.) in the
Simien Mountains. Andosols occupy the whole valley,
which is partially under rangeland and degraded forest and
partially under barley. The depth of the A-horizon was
measured at some 300 sites in cropland and compared with
A-horizon depth in non-cultivated areas for similar slope
gradients. Mean total soil profile truncation depth from
cropland, occurring between the beginning of permanent
human occupation (500–200 years ago) and 1974, was
measured as 14.5 ± 2.1 cm, or 950 ± 200 t ha−1, or 2–
5 t ha−1 year−1. Due to elevation and to the proximity of the
climatic limit of barley cultivation, deforestation here has
started much later than in most other parts of the highlands
(Hurni 1982). The variability in soil loss depth is correlated
with slope aspect and probably with the age of deforestation
(Hurni 1975, 1978). Measurements of sheet and rill erosion Fig. 21.3 Rill erosion at a farmer’s field at Wonzima (Blue Nile
rates were conducted in the Ethiopian highlands (Hurni basin). Rills occur particularly on long and steep slopes without soil
1985, 1990; Kejela 1992; Herweg and Ludi 1999; SCRP conservation structures; here, the depth of the rill is controlled by the
2000; Nyssen et al. 2009c). tillage pan, on which plough marks of the ard are visible (Photograph
E. Monsieurs, August 2013)
Soil loss occurs mainly at the beginning of the main
summer rainy season (kiremt). In those regions where spring
rains (belg) are sufficient for cultivation, these crops have
been harvested and the land ploughed again before kiremt most rainfall infiltrates quickly in the dry, tilled farmlands
(Tesfaye 1988). In the northern highlands, spring rain is (Gebreegziabher et al. 2009; Zenebe et al. 2013). Further-
unreliable and the land is only sown at the beginning of the more, on Vertisols, which are well represented in Ethiopia
kiremt season, when rains are intensive and their onset more (Kanwar and Virmani 1986; Moeyersons et al. 2006), the
regular. The farmlands have then undergone at least two first rains are well absorbed, because of the deep shrinkage
tillage operations are bare and offer less resistance to splash cracks. After some time, with the closing of the cracks, these
and runoff erosion (Virgo and Munro 1978). Studies in soils become completely impervious and favour significant
southern Ethiopia, where deforestation is ongoing, show a runoff production (Bauduin and Dubreuil 1973; Ge-
tremendous increase in soil loss over the last few decades breegziabher et al. 2009; Oicha et al. 2010; Araya et al.
(Kassa et al. 2013). In the northern highlands, with the 2011). Moreda and Bauwens (1998) found the most signif-
advance of the rainy season, soil loss decreases as crop cover icant correlation between monthly precipitation and summer
increases (Tesfaye 1988). This pattern was also observed flow in the Awash headwaters to occur in August, at the
and similarly accounted for by Billi (2004) for the suspended beginning of the second half of the rainy season, when ‘there
sediment concentration in the Meki River, a main tributary is greater opportunity for flow generation (even for smaller
to Ziway Lake in the Rift Valley. However, substantial storms) since the catchment is already moist’. Sutcliffe and
runoff is produced more than one month after the beginning Parks (1999) estimated that ‘early rainfall is required to
of the kiremt rains. In the beginning of the rainy season, replenish the soil moisture storage after the dry season’.
Table 21.1 The Revised Universal Soil Loss Equation (RUSLE)—adapted for field assessments in Ethiopia (Nyssen et al. 2009c)
Equation: annual soil loss rate A = R * K * S * L * C * P (Mg ha−1 year−1)
1. R: annual rainfall erosivity (MJ mm ha−1 h−1 year−1)
R = 5.5 Pr − 47
Pr = annual precipitation (mm)
2. K: soil erodibility (Mg h MJ−1 mm−1), including effects of rock fragment cover
K = [2.1 M1.14 (10−4)(12 − a) + 3.25 (b − 2) + 2.5 (c − 3)] * e−0.04 (d−10)
* 0.001317
M = particle size parameter = (% silt and very fine sand) * (100 − % clay)
a = percentage of organic matter
b = soil structure code, ranging between 1 (very fine granular) and 4 (blocky, platy, or massive), with default value 2
c = permeability class, ranging between 1 (rapid) and 6 (very slow), with default value 3
d = stone (rock fragment) cover (in %)
3. S: slope steepness factor (dimensionless)
S = −1.5 + 17/(1 + e(2.3−6.1 sinθ)
)
θ = slope angle (°)
4. L: slope length factor (dimensionless)
L = 0.232 λ0.48 (5 m ≤ λ ≤ 320 m)
λ = slope length (horizontal projection, in m)
5. C: cover-management factor (dimensionless)
Dense forest 0.001 Degraded rangeland (<50 % vegetation cover) 0.42 Badlands hard 0.05
Dryland forest; exclosure 0.004 Badlands soft 0.40
Dense grass 0.01 Degraded grass 0.05
Sorghum, maize 0.10 Tef (in high rainfall areas) 0.25 Fallow hard 0.05
Cereals, pulses 0.15 Tef (in semi-arid areas) 0.07 Fallow ploughed 0.60
6. P: supporting practices (dimensionless)
P = PC· PN· PM (on cropland); P = PN (on other land)
Ploughing and cropping practices PC Conservation structures PN In situ conservation PM
practices
Ploughing up and down 1 No conservation structures 1 Stubble grazing; no 1
mulching
Ploughing along the contour 0.9 Stone bund (average condition; smaller value for new s.b. and larger for older 0.3 Applying mulch 0.6
s.b.)
Strip cropping 0.8 Grass strip (1 m wide; slope ≤ 0.1 mm−1) 0.4 Zero grazing 0.8
Intercropping 0.8 Grass strip (1 m wide; slope > 0.2 mm−1) 0.8
Dense intercropping 0.7
Source Renard et al. (1997). Adaptations: R correlation by Hurni (1985); K adjustment for rock fragment cover by Poesen et al. (1994); L correlation by Hurni (1985); C values by
Hurni (1985) and Nyssen et al. (2009c); P model by Nyssen et al. (2009c); P values by Hurni (1985), Nyssen (2001), Gebremichael et al. (2005), Nyssen et al. (2007a, b, 2008b).
Limitations as mentioned in Sect. 21.4
From their research, Hurni (1985) and later Nyssen et al. average drop sizes in the Ethiopian highlands, should be
(2009c) adapted the Revised USLE (Renard et al. 1997) to carried out (Nyssen et al. 2005). Calculations of the slope
Ethiopian conditions for use by development agents in the steepness factor (S) and the slope length (L) factor are shown
field of SWC (Table 21.1). The soil erodibility factor K can in Table 21.1 (Sects. 21.3 and 21.4). The use of equations for
be assessed from soil textural data, organic matter content, L requires caution, since ‘slope length is the factor that
and rock fragment cover (Table 21.1, Sect. 21.2). We rec- involves the most judgement, and length determinations
ommend including the rock fragment cover, which is a made by users vary greatly’ (Renard et al. 1997). In Ethio-
widespread feature in the Ethiopian highlands, as a correc- pian highland conditions, this runoff length is generally
tion factor for the K-value, rather than in the management longer than one single farm plot and shorter than the whole
factor P (Nyssen et al. 2002b). slope, from ridge to foot. Cover-management C values
For the R-factor (rainfall erosivity), the Ethiopia-specific (Table 21.1, Sect. 21.5) have been reported by Nyssen et al.
equation (Table 21.1, Sect. 21.1) may be used, bearing in (2009c). The P factor (dimensionless) relates to supporting
mind that additional studies, taking into account the above practices and indicates reduced soil erosion potential due to
farming practices and conservation measures. Sub-factors considered to be the major cause of rapid gullying in many
yield one composite P-value (Foster and Highfill 1983) for a areas (Frankl et al. 2011). One should also stress the
conservation system (Table 21.1, Sect. 21.6): importance of cropland abandonment for gully initiation,
especially if it is converted into grazing land. The over-
P ¼ PC PN PM ð21:1Þ grazed soil surface has a higher runoff coefficient than reg-
ularly ploughed farmlands; SWC structures are no longer
where maintained, and bank gullying often starts at places where
PC = Sub-factor for ploughing and cropping practices; these structures collapse.
PN = Sub-factor for conservation structures; Brancaccio et al. (1997) explained the present-day pro-
PM = Sub-factor for in situ conservation practices. cesses of channel incision in northern Ethiopia by an
increasing erosional power of concentrated runoff due to a
21.5 Gullying decreasing sediment load (clear water effect), associated with
the advanced phase of soil erosion on the hillslopes where
Gullying (Fig. 21.4) is not restricted to the highlands of bedrock is now outcropping. Since the late nineteenth cen-
Ethiopia but is widespread at sub-continental scale in Africa tury, gullies were present and though they had become sta-
(Moeyersons 2000). In Tigray, the increase of runoff bilised by 1935, a strong incision phase started in the 1960s
response on many hillslopes has been attributed to an overall due to the above-mentioned factors (Frankl et al. 2011).
lowering of the infiltration capacity of the soils due to Gullies in Ethiopia can often be considered as discon-
removal of natural vegetation (Virgo and Munro 1978; tinuous ephemeral streams (Bull 1997) comprising a hill-
Machado et al. 1998). Buried soils indicate advanced slope gully, an alluvial–colluvial cone at the foot of the hill,
deforestation, which in the Ethiopian highlands might have and renewed incision with gully head formation further
started around 5,000 14C years BP (Machado et al. 1998; downslope in the valley Vertisol. Pediments dissected by
Nyssen et al. 2004b; Pietsch and Machado 2014). Since the gullies are a common feature in many areas (Riché and
twentieth century, however, vegetation removal has also Ségalen 1973; Berakhi and Brancaccio 1993; Berakhi et al.
affected shrub and small tree cover, as well as grass strips in 1997). In the valley bottoms, initial gully heads often coin-
between the farmlands and on steep slopes. This removal of cide with sinking polygonal structures in Vertisols (Nyssen
vegetation has further lowered the infiltration capacity of the et al. 2000b), where piping erosion is very active (Frankl
soils, favoured the occurrence of flash floods, and is et al. 2012).
Fig. 21.4 Gullies, like this one

in Harena (Dogu’a Tembien), do
not only result in soil loss, but
also drain out the landscape
(lowering of the water table) and
are major obstacles to
communication
Active gullying induced by road building on pediments stone bund building. Possible effects of soil profile trunca-
was described by Berakhi and Brancaccio (1993). In a case tion on the values of these two parameters are outbalanced
study along the Mekele—Adwa road, built in 1994, Nyssen by increased infiltration rates, induced by stone bund
et al. (2002a), demonstrated how road building, through the building (Vancampenhout et al. 2006). The most common
enlargement of drainage areas and the concentration of soils in the Ethiopian highlands (i.e. Regosols, Vertic
runoff, induced an artificial exceedance of the critical Cambisols and Vertisols) have a quite homogenous com-
catchment area at which gully heads are formed for a given position with depth, which explains low soil fertility gradi-
slope gradient. ents in terraced lands.
21.6 Tillage Erosion 21.7 Wind Erosion
Soil translocation due to tillage by the ox-drawn ard plough In the Ethiopian highlands, wind erosion has not been
(Fig. 21.5) appears to be an important soil erosion process in measured and was rarely mentioned. Wind erosion mainly
the Ethiopian highlands. Assessments of tillage erosion rates occurs as ‘dust devils’ in areas with important trampling by
indicate that this process contributes on average to half of the humans or cattle, such as market places, footpaths, unme-
sediment deposited behind stone bunds (Nyssen et al. 2000c; talled roads, around cattle drinking places or on cropland
Gebremichael et al. 2005). Colluviation occurs in the lower where post-harvest grazing has taken place. On the numer-
part of the farmland and soil profiles are truncated in the ous isolated mountains or ‘inselbergs’, important wind ero-
upper part (Herweg and Ludi 1999; Nyssen et al. 2000c). sion, including the formation of dunes, occurs due to local
Soil sequences on progressive terraces overlying strongly aerodynamic situation (Uhlig and Uhlig 1989). More
weathered rock were analysed in the central highlands of research on wind erosion in the Ethiopian highlands seems
Ethiopia, in the Ankober area (Bono and Seiler 1986). At the necessary, as it may have been insufficiently studied.
upper part of the terrace, soils are shallow and water and Wind erosion is especially important in low-lying, dry
nutrient storage capacity low. However, in Dogu’a Tembien and hot regions, adjacent to the highlands, such as many
(Tigray), intra-parcel variability of soil fertility parameters is places in the Rift Valley. Desert pavements, created by wind
small. A larger content of soil moisture and of soil organic erosion, exist around Lake Turkana (Hemming and Trapnell
matter was even observed at the foot of the stone bunds, at 1957). Wind erosion and deposition contribute to the for-
the very place where the soil profile has been truncated after mation of dunes in the alluvial plains of the Wabe Shebele
Fig. 21.5 Soil tillage by

mahrasha ard plough, here in
Dogu’a Tembien, causes a
downslope movement of the
topsoil (tillage erosion)
(Photograph A Roelofs, April
2005)
and to overall deposition of aeolian sediments in that region deposited within the catchment (Nyssen et al. 2007b). Reuter
(Riché and Ségalen 1973). The Eritrean coastal plain is in (1991) and Descheemaeker et al. (2006b) stressed the
many places covered by stone mantles produced by deflation magnitude of sediment and organic carbon stored in collu-
as well as by loose sand occurring either as a mantle of vium on footslopes and reforested areas (exclosures sensu
variable depth or in the form of mobile dunes (Hemming Aerts et al. 2009). Sediment deposition in floodplains and
1961). Aeolian sediments in the coastal plains can be com- natural lakes is important, but the rates have not been studied
posed of eroded materials from nearby rocks or brought in systematically in Ethiopia.
by dust storms, which are quite common (Horowitz 1967).
21.10 Land Degradation

21.8 Mass Movements and Desertification
Due to steep topography, the presence of lithologies with a Although climatic conditions (0.05 < annual precipitation/
low shear strength, torrential rainfall, and in some cases the potential evapotranspiration < 0.65) in parts of the northern
occurrence of earthquakes, the Ethiopian highlands are also highlands and in the low-lying parts of the country would
affected by various types of mass movements (e.g. rock falls, justify the use of the term ‘desertification’ (UNEP 1994), the
debris flows, and slumps). Several studies have mapped term ‘land degradation’ will be used to indicate environ-
landslides in Ethiopia and have analysed their controlling mental degradation throughout the country. Two major
factors (e.g. Moeyersons et al. 2008; Van Den Eeckhaut factors inducing land degradation in the Ethiopian highlands
et al. 2009; Broothaerts et al. 2012). Although many mass are generally considered: drought and land use changes.
movements have been initiated by natural factors, human
activities (i.e. land use change, undercutting and overloading
during road construction, and improper slope drainage sys- 21.10.1 Rainfall Variations and Drought
tems) have often contributed to the reactivation of landslides.
In south Ethiopia, (Broothaerts et al. 2012) observed many Attention to famines in Ethiopia has created a popular view
recent landslides along river channels which were triggered of a drought-stricken country, with a tendency towards
by river channel incision due to increased peak flow dis- decreasing annual rainfall. The decline of rainfall in the
charges following deforestation in their catchments. Large Sahel observed since about 1965 was also seen to a lesser
landslides redistribute large volumes of sediments in the extent in the north-central Ethiopian highlands (Camberlin
highlands, hence affecting the spatial patterns of soil types 1994; Seleshi and Demarée 1995). However, unlike the
(Van de Wauw et al. 2008). Sahel, a comparison between two reference periods (1931–
1960 and 1961–1990) yields no significant changes in mean
precipitation over Ethiopia, but an increased inter-annual
21.9 Sediment Deposition variability (Hulme 1992). Analyses of long-term time series
of annual precipitation, both for Addis Ababa and the
On the back- and footslopes of cliffs, a ‘classic’ sorting of northern highlands, show that although the succession of dry
deposited sediment generally occurs, the coarse sediments years between the late 1970s and late 1980s produced the
(rock fragments) being deposited on the debris slope, and the driest decade of the last century in the Ethiopian highlands,
finer material on the footslope, as shown by Riché and there is no evidence for a long-term trend or change in the
Ségalen (1973) in the Wabe Shebele basin. Belay Tegene region’s annual rainfall regime (Conway 2000a; Conway
(1998) emphasised the importance of continuous deposition et al. 2004).
of colluvium on convergent footslopes which prevents the With respect to the inter-annual rainfall variability,
development of mature soil profiles. Hurni (1985) shows, for Conway (2000b) found a coefficient of variation below 20 %
a 116 ha catchment in Welo, that the rate of sediment for the wetter areas, but far above that for drier areas to the
accumulation (17 t ha−1 year−1) is more important than the north and at lower altitudes (see also Chap. 4, this volume).
rate of sediment export through the drainage system (Hoffmann 1987) also found annual rainfall variability
(7 t ha−1 year−1). In a well-vegetated catchment in south- strongly dependent on climatic region: <10 % in the area
western Ethiopia, sediment accumulation rates are around Jimma with a tropical rain climate and >45 % in
30 t ha−1 year−1 and sediment export rates through the river semi-desert areas. Dry years were observed in 1913–1914,
only 1.1 t ha−1 year−1. Here, most of sediment deposition 1937, 1941, 1953, 1957, 1965–1966, 1969, 1973–1974,
occurs in densely vegetated areas along riverbanks. A sedi- 1976, 1979, 1983–1984, and 1987 (Camberlin 1994). It is
ment budget for a 200 ha catchment in Tigray highlands evident that, in an already degraded environment, a dry year
indicates that 59 % of sediment produced by water erosion is has a very negative impact, not only on agricultural
production, but also on the environment (i.e. overgrazing, bulk density, decreased soil organic matter content, soil
cracking of Vertisols, and groundwater depletion). RC in structure decay, and decreased hydraulic conductivity. All
such a year are higher (Casenave and Valentin 1992; Val- these factors contribute to decreased infiltration rates and
entin et al. 2005) and result in increased soil erosion. increased runoff volumes. Mwendera et al. (1997) carried
Besides yearly precipitation, its seasonal distribution out experiments on grazing land with slope gradients
must be considered as well. Unlike West Africa, according <0.08 mm−1 in an area between Ambo and Addis Ababa.
to Hulme (1992), the seasonality of rainfall over Ethiopia Comparing ungrazed, moderately, and heavily grazed land,
slightly decreased between 1931–1960 and 1961–1990. they found significant differences in runoff volumes for slope
However, evenly distributed rains mean also that a larger gradients in the range of 0.05–0.08 mm−1. Steady-state
percentage of precipitation falls outside the crop growth infiltration rates decreased significantly, even under light
season, or that there is a shift from one rainy season to grazing intensity, and showed the effect of animal trampling
another, particularly decreased summer (kiremt) and on soil compaction (Mwendera and Saleem 1997). On
increased spring (belg) rains in the northern Ethiopian cropland, stubble grazing (a widespread practice) dramati-
highlands (Camberlin 1994; Seleshi and Demarée 1995; cally decreases the infiltration capacity. Field observations
Seleshi and Demarée 1998). Differences between the tem- also indicate that topsoil degradation by cattle trampling
poral pattern of spring and summer rains are expected to significantly contributes to soil erosion and sediment deliv-
reflect different levels of influence from the Indian and ery to water reservoirs.
Atlantic oceans (Conway 2000b). Repeat photography has also revealed that in the late
nineteenth century, the landscapes were at least as barren as
they are nowadays (Nyssen et al. 2009b). In recent years,
21.10.2 Human Settlement and Changes since 1975, the tree cover has improved in 90 % of the
in Land Use and Land Cover analysed landscapes (Munro et al. 2008; Nyssen et al.
2008b). Exclosures (Aerts et al. 2009) have been established
Human settlement with concomitant agricultural exploitation in former communal grazing land with the aim of forest
induces significant changes in land use and land cover, restoration and land conservation. The establishment of ex-
which in turn alter infiltration and runoff conditions, as well closures was made possible by an important land tenure
as soil erosion processes (Olson 1981; Bunney 1990). change in the 1980s, in which large feudal agricultural lands
Detailed studies show that settlement decisions were made in the valley bottoms and other level areas were shared
on a clear ecological basis, especially from the beginning of among the local farmers and this decreased the need of poor
the pre-Axumite era (700 BCE). Preferred locations were at farmers to establish marginal farmlands on hillslopes. In
the margin between Vertisol areas and narrow alluvial valley these locations, exclosures could then be established after
bottoms which could be irrigated (Michels 1988). Human land reform (Rahmato 1994). Although centrally imposed,
activity expanded from such preferential places to the pres- the implementation of exclosures is rather a bottom-up
ent-day occupation of steep slopes for agriculture through a process. Participation is enhanced by the implementation of
number of stages, including forest clearing and removal of remunerated SWC activities and plantation works at the
remnant trees and shrubs. establishment of the exclosure. Location, area, local by-laws
In the Ethiopian highlands, livestock grazes on vast related to restrictions and management, instalment and
deforested areas, commonly called rangeland, as well as payment of guards are most often decided at the local
various types of climax grasslands: i.e. at high elevations, on community level (Muys et al. 2014). The villagers are
Vertisols and on dry places (Klötzli 1977). Much in the same overall convincingly participating in reforestation and other
way as in forests and woodlands, vegetation cover decreases conservation activities (Kumasi and Asenso-Okyere 2011).
in grass- and rangeland. Most of the above-quoted studies of However, the encroachment by eucalypts remains a bottle-
land use changes show, besides decreasing tree and shrub neck for biodiversity. The benefits of planting these trees are
cover, an increase of the area occupied by ‘bare land’, ‘no largely for individual farmers, whereas the negative effects
vegetation’, ‘open areas’, and the like. Overgrazing of of this water-demanding tree are borne by the communities
rangeland is a particular problem in the cereal zones of the (Muys et al. 2014).
highlands, where current stocking rates are well in excess of
estimated optimum rates (Hurni 1993). Livestock plays a
key role in the agricultural system of the highlands, pro- 21.10.3 Social and Historical Impulses
viding energy (traction, manure used as fuel), food, fertiliser, of Land Use and Cover Changes
insurance, and status (Kassa et al. 2002). Consequences of
overgrazing on the environment are decreased surface It appears that rainfall variability, apart from the catastrophic
roughness, compaction of fine textured soils, increased soil impact of dry years on the degraded environment, cannot be
Fig. 21.6 Plough marks on large

rock fragments and pedestal-
supported boulders indicate that
this 100–200-year-old Juniperus
forest at Kuskuam near Debre
Tabor has grown on previously
degraded farmland. Forest re-
growth has taken place, as also
evidenced by the well-branched
older tree in the centre of the
photograph that used to grow in
an open area (Photograph
J. Nyssen, July 2011)
invoked to explain the current land degradation. Causes are erosion (Ståhl 1974; Girma and Jacob 1988; Ståhl 1990).
to be found in changing land use and land cover, which are In short, in situations of poverty and social insecurity,
expressions of human impact (Reid et al. 2000; Feoli et al. short-term survival prevailed over medium- and long-term
2002). Though deforestation and removal of other vegetation conservation issues.
cover over the last 2,000–3,000 years have probably been a
cyclic rather than a linear process (Fig. 21.6), studies on land
use and land cover change show that, at present, there is a
tendency of increasing removal of vegetation cover. 21.11 Human Reaction to Land
At this stage, it appears necessary to briefly outline the Degradation
social and historical causes of this human impact. Under
feudalism (until 1974), agricultural techniques stagnated for 21.11.1 Agricultural Intensification
centuries (Crummey 2000). Until the 1940s, the Agricul- and Land Rehabilitation
tural Department’s only effective activity was collecting the
agricultural tax (Joyce 1943). Investment in agriculture Faced with a deteriorating environment, society reacts in
started only in 1950s and was in the beginning mostly order to maintain/improve agricultural production, often
oriented towards export crops such as coffee (Coffea leading to changes in the production system (Boserup 1981),
arabica L.), grown in southern Ethiopia. Therefore, there an innovative process in which modern science needs to be
was limited agricultural investment in the highlands, where involved (Blaikie and Brookfield 1987; Ståhl 1990). The
subsistence production dominated (Ståhl 1990; Mulugetta present-day rise in food production in Ethiopia (Fig. 21.7)
1992). Until the late 1970s, sharecropping prevented the can, besides re-established climatic conditions, also be
farmers from investing in their farmlands. Impoverishment attributed to a variety of human interventions at different
led them to prefer immediate returns, even if it induced levels (Nyssen et al. 2004a). Extension of cropped area and
environmental degradation (Taddesse 1995). On the other increased grazing pressure is still possible. However, limited
hand, recent land redistributions in order to allocate land- space is left for this and productivity decreases. Giglioli
less households had a positive impact on land productivity (1938a, b) and Joyce (1943) already reported the widespread
(Benin and Pender 2001). To increase agricultural pro- use of indigenous SWC technology since a long time. Such
duction, most trees and shrubs between the farmlands and indigenous technologies can be used as a starting point, but
on steep slopes were cleared during the nineteenth and need improvement in order to increase their ecological effi-
twentieth centuries, thereby increasing runoff and soil ciency (Hurni 1998; Nyssen et al. 2000a).
250 25 gardening and minimum tillage (on self-mulching Vertisols)

Cereal yield
Total cereal production (10 6 t / y)

may be suggested (Astatke et al. 2002; Araya et al. 2012), as
Cereal production per inhabitant
Cereal yield (t / km²) and
200 20 well as an extension of the cropping period on Vertisols

production (kg / inhab.)
Total cereal production

(Tedla et al. 1999). Asnakew et al. (1994) obtained good
150 15 maize yields with rock fragment mulching and no-tillage.
Besides these conservation measures, Ethiopia strongly
100 10
invested in agricultural inputs, particularly fertilisers and
improved seeds. As a result, total food production is now
50 5
higher than ever; also food production per capita in 2005–
0 0
2010 was 160 % of that in 1985–1990 (Fig. 21.7).
1960 1970 1980 1990 2000 2010
Fig. 21.7 Agricultural intensification in Ethiopia is evidenced by

cereal production trends (data retrieved from http://faostat.fao.org) 21.11.2 Soil and Water Conservation
The main agricultural intensification observed in Ethiopia is

Nowadays, changes in the agricultural system appear certainly the now widespread catchment management
such as haymaking (‘cut and carry’) (Hurni 1986), partially activities (Fig. 21.8). Throughout the Ethiopian highlands, it
from exclosures (i.e. land under strict conservation man- is apparent that many SWC structures, established during the
agement, often controlled by the community), which are 1980s, remain in place and are often maintained. Their
increasingly being organised in the most affected northern destruction is not as widespread as stated by Rahmato
highlands (Tekle et al. 1997; Shitarek et al. 2001; Aerts et al. (1994), often the farmers accept and adopt these structures.
2004, 2009; Descheemaeker et al. 2006b) and which lead to Many, probably most of the soil bunds throughout Welo,
sediment trapping and enhanced soil fertility status (Elias have evolved into full-grown lynchets. Even in the high
and Scoones 1999; Descheemaeker et al. 2006a; Mekuria rainfall Ankober area, soil bunds have often been ‘opened’
et al. 2007). to allow drainage, but are still in place over most of their
Different pathways of agricultural intensification are length.
possible in Ethiopia. Mineral fertilising should not be the Local knowledge and farmers’ initiatives are integrated
overall option, given scarce capital resources. Due to with these introduced technologies at various degrees
decreased landholdings, a shift in the soil tillage system to (Gaspart et al. 1997; Nyssen et al. 2000a, 2004a, 2008b,
Fig. 21.8 Catchment

rehabilitation in the sub-humid
May Zeg-zeg catchment (Tigray);
trenches behind the stone bunds
enhance infiltration and decrease
catchment runoff response
(Photograph K. Herweg, May
2005)
2009a; Haile et al. 2006; Gebresamuel et al. 2009). The SWC activities also enhance rain infiltration rates during
efficiency of particular techniques cannot be discussed in the short but heavy storms and improve the situation with
depth here; the reader is referred to specialised publications regard to flooding, soil erosion, and groundwater recharge
(Herweg and Ludi 1999; SCRP 2000; Nyssen et al. 2004c, (Nyssen et al. 2009a, 2010). The current land tenure system
2007a, 2009a, 2010; Gebremichael et al. 2005; Haregeweyn in which an equality of land holdings is attempted, favours
et al. 2006; Vancampenhout et al. 2006; Wondumagegnehu solidarity among the farmers to undertake communal
et al. 2007; Alemayehu et al. 2009; Reubens et al. 2009; catchment management activities (Kumasi and Asenso-Ok-
Gebreegziabher et al. 2009; Araya et al. 2011; Lanckriet yere 2011; Taye et al. 2013). Besides the need for collecting
et al. 2012; Muys et al. 2014; Gebreyohannes et al. 2014). a wide set of original data, conceptually, in all related
research, a good comprehension of the hydrological balance
is needed. Further, for nutrient, sediment, and water-related
21.12 Conclusions processes, it is important to understand the occurrence of
sinks and to keep the scale concept in mind. These principles
Ethiopia is on the map for research on land resources and are at the base of the successful implementation of catch-
implementation of sustainable land management (SLM) ment management activities in northern Ethiopia (Fig. 21.9).
(Haile et al. 2006). Future research priorities are identified. In this regard, the impacts of the May Zeg-zeg catchment
Cornerstones of SLM include forest development in critical management could be monitored in detail (Nyssen et al.
places (Descheemaeker et al. 2006b, 2009), over sufficiently 2009a, 2010; Walraevens et al. 2009) and future develop-
large areas, as demonstrated through the dramatic changes ment scenarios could be elaborated (Lanckriet et al. 2012).
that occurred on the Rift Valley escarpment near Alamata
(Gebreyohannes et al. 2014).
References
Abebe Y, Apparao G (1989) Annual rain water potential and its

variability in drought years over Ethiopia. In: Conference on climate
and water, Suomen Akatemian, Helsinki, Finland, pp 219–235
Aerts R, Wagendorp T, November E, Behailu M, Deckers J, Muys B
(2004) Ecosystem thermal buffer capacity as an indicator of the
restoration status of protected areas in the northern Ethiopian
highlands. Rest Ecol 12:586–596
Aerts R, Nyssen J, Haile M (2009) On the difference between
“exclosures” and “enclosures” in ecology and the environment.
J Arid Environ 73:762–763
Alemayehu F, Taha N, Nyssen J, Girma A, Zenebe A, Behailu M,
Deckers J, Poesen J (2009) The impacts of watershed management
on land use and land cover dynamics in Eastern Tigray (Ethiopia).
Resour Conserv Recycl 53:192–198
Araya T, Cornelis WM, Nyssen J, Govaerts B, Gebregziabher T, Oicha
T, Getnet F, Raes D, Haile M, D J, Sayre KD (2011) Effects of
conservation agriculture on runoff, soil loss and crop yield under
rain fed conditions in Tigray, Northern Ethiopia. Soil Tillage Res
27:404–414
Araya T, Cornelis WM, Nyssen J, Govaerts B, Getnet F, Bauer H,
Amare K, Raes D, Haile M, Deckers J (2012) Medium-term effects
of conservation agriculture based cropping systems for sustainable
soil and water management and crop productivity in the Ethiopian
highlands. Field Crops Res 132:53–62
Asnakaw W, Assefa A, Yematawork A, Abera S, Van Straaten P,
Groenevelt P, Chesworth W (1994) Report on the results of the
Ethiopia-Canada agrogeology project—rock mulch. IDRC-Project
88-1032
Astatke A, Jabbar M, Mohamed Saleem MA, Erkossa T (2002)
Technical and economic performance of animal-drawn implements
for minimum tillage: experience on Vertisols in Ethiopia. Exp Agric
Fig. 21.9 Soil and water conservation activities in the Tsinkaniet plain 38:185–196. doi:10.1017/S0014479702000248
have led to a situation where overland flow strongly decreased in 2006 Bakker JP (1967) Weathering of granite in different climates, partic-
(b—photograph J. Nyssen) as compared to 1975 when evidence of ularly in Europe.In: Symposium international de géomorphologie,
flooding is clearly visible at the footslope (a—photograph R.N. 8–16 juin 1966. Les congrès et colloques de l’Université de Liège,
Munro). At the far end, in 2006, a reservoir mainly fed by groundwater Liège-Louvain, pp 51–68
is visible
Bauduin D, Dubreuil P (1973) L’inventaire des ressources en eau pour Descheemaeker K, Muys B, Nyssen J, Poesen J, Raes D, Haile M,
l’aménagement intégré du Bassin du Wabi Shebelle d’Ethiopie. Cah Deckers J (2006a) Litter production and organic matter accumula-
ORSTOM, sér Hydrol 10:307–348 tion in exclosures of the Tigray highlands, Ethiopia. For Ecol
Benin S, Pender J (2001) Impacts of land redistribution on land Manag 233:21–35
management and productivity in the Ethiopian highlands. Land Descheemaeker K, Nyssen J, Rossi J, Poesen J, Haile M, Moeyersons J,
Degrad Dev 12:555–568 Deckers J (2006b) Sediment deposition and pedogenesis in
Berakhi O, Brancaccio L (1993) Some reflections on the origin and exclosures in the Tigray highlands, Ethiopia. Geoderma 132:291–
land use of pediments on the Ethiopian highlands. Geogr Fis Dinam 314. doi:10.1016/j.geoderma.2005.04.027
Quat 16:101–106 Descheemaeker K, Raes D, Nyssen J, Poesen J, Haile M, Deckers J
Berakhi O, Brancaccio L, Calderoni G, Coltorti M, Dramis F, Tegene (2009) Changes in water flows and water productivity upon
B, Umer M (1997) Geomorphological and sedimentary records of vegetation regeneration on degraded hillslopes in northern Ethiopia:
Holocene climatic changes and human impact in the highlands of a water balance modelling exercise. Rangel J 31:237–249
Northern Ethiopia. Geogr Fis e Din Quat Supp 3, T1:77 Dessie M, Verhoest NEC, Admasu T, Pauwels VRN, Poesen J, Adgo
Berhane G, Martens K, Al Farrah N, Walraevens K (2013) Water E, Deckers J, Nyssen J (2014) Effects of the floodplain on river
leakage investigation of micro-dam reservoirs in Mesozoic sedi- discharge into Lake Tana (Ethiopia). J Hydrol 519, part A: 699–
mentary sequences in Northern Ethiopia. J Afr Earth Sci 79:98–110. 710. http://dx.doi.org/10.1016/j.jhydrol.2014.08.007
http://dx.doi.org/10.1016/j.jafrearsci.2012.10.004 Diodato N, Knight J, Belolocchi G (2013) Reduced complexity model
Billi P (2004) Sediment yield of a closed river system: the Meki River, for assessing patterns of rainfall erosivity in Africa. Global Planet
western margin of the Ethiopian Rift Valley. In: Proceedings of the Change 100:183–193
IAHS international symposium on sediment transfer through the Elias Eyasu, Scoones I (1999) Perspectives on soil fertility change: a
fluvial system, Moscow, 2–6 Aug 2004. pp 94–100 case study from southern Ethiopia. Land Degrad Dev 10:195–206
Blaikie P, Brookfield H (1987) Land degradation and society. Methuen, Feoli E, Vuerich L, Woldu Z (2002) Evaluation of environmental
New York degradation in northern Ethiopia using GIS to integrate vegetation,
Bono R, Seiler W (1986) Soils and soil erosion in the area of Debre geomorphological, erosion and socio-economic factors. Agric
Birhan, Eastern High Plateau of Shewa, Ethiopia. Regio Basiliensis Ecosyst Environ 91:313–325
27:39–53 Foster GR, Highfill RE (1983) Effect of terraces on soil loss—USLE P-
Boserup E (1981) Population and technology. Blackwell, Oxford factor values for terraces. J Soil Water Conserv 38:48–51
Brancaccio L, Calderoni G, Coltorti M, Dramis F (1997) Phases of soil Frankl A, Nyssen J, De Dapper M, Haile M, Billi P, Munro RN,
erosion during the Holocene in the highlands of Western Tigray Deckers J, Poesen J (2011) Linking long-term gully and river
(Northern Ethiopia): a preliminary report. In: Bard K (ed) The channel dynamics to environmental change using repeat photogra-
environmental history and human ecology of Northern Ethiopia in phy (North Ethiopia). Geomorphology 129:238–251. doi:10.1016/j.
the Late Holocene. Instituto Universitario Orientale, Napoli, pp 30– geomorph.2011.02.018
48 Frankl A, Poesen J, Deckers J, Haile M, Nyssen J (2012) Gully head
Broothaerts N, Kissi E, Poesen J, Van Rompaey A, Getahun K, Van retreat rates in the semi-arid highlands of Northern Ethiopia.
Ranst E, Diels J (2012) Spatial patterns, causes and consequences of Geomorphology 173–174:185–195. http://dx.doi.org/10.1016/j.
landslides in the Gilgel Gibe catchment, SW Ethiopia. Catena geomorph.2012.06.011
97:127–136. http://dx.doi.org/10.1016/j.catena.2012.05.011 Gaspart F, Jabbar M, Mélard C, Platteau J-P (1997) Participation in the
Bull WB (1997) Discontinuous ephemeral streams. Geomorphology construction of a local public good: a case study of watershed
19:227–276 management in the Ethiopian highlands. Cahiers de la Faculté des
Bunney S (1990) Prehistoric farming caused devastating soil erosion. Sciences Economiques, Sociales et de Gestion de Namur, Série
N Sci 125:20 Recherche 181
Camberlin P (1994) Les précipitations dans la Corne orientale de Gebreegziabher T, Nyssen J, Govaerts B, Getnet F, Behailu M, Haile
l’Afrique: climatologie, variabilité et connexions avec quelques M, Deckers J (2009) Contour furrows for in situ soil and water
indicateurs océano-atmosphériques. PhD dissertation. Université de conservation, Tigray, Northern Ethiopia. Soil Tillage Res 103:257–
Bourgogne 264. doi:10.1016/j.still.2008.05.021
Casenave A, Valentin C (1992) A runoff capability classification Gebremichael D, Nyssen J, Poesen J, Deckers J, Haile M, Govers G,
system based on surface features criteria in the arid and semi-arid Moeyersons J (2005) Effectiveness of stone bunds in controlling
areas of West Africa. J Hydrol 130:231–249 soil erosion on cropland in the Tigray highlands. North Ethiop Soil
Conway D (1997) A water balance model of the Upper Blue Nile in Use Manag 21:287–297. doi:10.1111/j.1475-2743.2005.tb00401.x
Ethiopia. Hydrol Sci J 42:265–286 Gebresamuel G, Haile M, Singh BR (2009) Agronomic and economic
Conway D (2000a) The climate and hydrology of the Upper Blue Nile performance of reservoir sediment for rehabilitating degraded soils
River. Geogr J 166:49–62 in Northern Ethiopia. Nutr Cycl Agroecosyst 84:23–38
Conway D (2000b) Some aspects of climate variability in the North Gebreyohannes T, Frankl A, Haile M, Nyssen J (2014) Catchment
East Ethiopian highlands—Wollo and Tigray. Sinet: Ethiop. J Sci rehabilitation and hydro-geomorphic characteristics of mountain
23:139–161 streams in the western Rift Valley escarpment of Northern Ethiopia.
Conway D, Hulme M (1993) Recent fluctuations in precipitation and Land Degrad Dev: Online Early View. doi:10.1002/ldr.2267
runoff over the Nile subbasins and their impact on main Nile Giglioli G (1938a) L’erosione del terreno agrario nei Tropici. Ist Agric
discharge. Clim Change 25:127–151 Colon Ital Rev 49
Conway D, Mould C, Bewket W (2004) Over one century of rainfall Giglioli G (1938b) Le sistemazioni idraulico agrarie dei terreni
and temperature observations in Addis Ababa, Ethiopia. Int J nell’Africa Italiana. Agric Colon 32:481–496
Climatol 24:77–91 Girma K, Jacob MJ (1988) Drought, famine and the political economy
Crummey D (2000) Land and society in the Christian Kingdom of of environmental degradation in Ethiopia. Geography 318:65–70
Ethiopia, from the Thirteenth to the Twentieth Century. University Goebel W, Odenyo V (1984) Ethiopia. Agroclimatic resources
of illinois Press, Addis Ababa inventory for land-use planning. Technical report DP/ETH/78/003,
vol I, 208 p; vol II, 95 p. Ministry of Agriculture, Land Use under Semi-Arid Conditions, Nairobi, Kenya, 1–6 Dec 1986.
Planning and Regulatory Department, UNDP, FAO, Addis Ababa IBSRAM, Bangkok, Thailand, pp 157–172
Haile M, Herweg K, Stillhardt B (2006) Sustainable land management Kassa H, Gibbon D, Singh B (2002) Livestock improve household food
—a new approach to soil and water conservation in Ethiopia. security and sustainability of Ethiopian small farms. J Sustain Agri
University of Berne, Berne, Switzerland, Mekelle University, 21:73–93
Mekelle, Ethiopia Kassa H, Nyssen J, Frankl A, Dondeyne S, Poesen J (2013) The impact
Haregeweyn N, Poesen J, Nyssen J, De Wit J, Haile M, Govers G, of human activities on natural resources in the southwestern
Deckers J (2006) Reservoirs in Tigray: characteristics and sediment highlands of Ethiopia—a state of art (in prep)
deposition problems. Land Degrad Dev 17:211–230 Tekle K, Backeus I, Skoglund J, Woldu Z (1997) Vegetation on
Hemming CF (1961) The ecology of the coastal area of Northern hillslopes in southern Wello, Ethiopia: degradation and regenera-
Eritrea. J Ecol 49:55–78 tion. Nordic J Bot 17:483–493
Hemming CF, Trapnell CG (1957) A reconnaissance classification of Kejela K (1992) The costs of soil erosion in Anjeni, Ethiopia. In: Tato
the soils of the south Turkana desert. J Soil Sci 8:167–183 K and H. Hurni (eds) Soil conservation for survival. Soil and Water
Herweg K, Ludi E (1999) The performance of selected soil and water Conservation Society, Ankeny, pp 219–230
conservation measures-case studies from Ethiopia and Eritrea. Klötzli F (1977) Wild und Vieh im Gebirgsland Aethiopiens. In:
Catena 36:99–114 Tuexen R (ed) Symp. der int Ver. für Veg.kunde. Cramer, Vaduz,
Hoffmann R (1987) Die Dürregefährdung äthiopischer Landschaften— Rinteln, pp 499–512
ein Beitrag zum Problem der Belastbarkeit tropischer Geoökosys- Krauer J (1988) Rainfall, erosivity and isoerdent map of Ethiopia. Soil
teme. Petermanns Geogr Mitt 131:217–224 conservation research project, research report 15, University of
Horowitz A (1967) The geology of Museri Island (Dahlak Archipelago, Berne, Bern
southern Red Sea). Isr J Earth-Sci 16:74–83 Kumasi T, Asenso-Okyere K (2011) Responding to land degradation in
Hulme M (1992) Rainfall changes in Africa: 1931–1960 to 1961–1990. the highlands of Tigray, Northern Ethiopia. IFPRI discussion paper,
Int J Clim 12:685–699 01142. International Food Policy Research Institute
Hunting Technical Services (1976) Tigrai rural development study, Lanckriet S, Araya Tesfay, Cornelis W, Verfaillie E, Poesen J, Govaerts B,
Annex 2: water resources, vol 1: hydrology and surface water. Bauer H, Deckers S, Haile M, Nyssen J (2012) Impact of conservation
Hunting Technical Services Ltd, Hemel Hempstead (G.B.) agriculture on catchment runoff and soil loss under changing climate
Hurni H (1975) Bodenerosion in Semien-Aethiopien (mit Kartenbei- conditions in May Zeg-zeg (Ethiopia). J Hydrol 475:336–349
lage 1:25 000). Geogr Helvetica 4:157–168 Machado MJ, Perez-Gonzalez A, Benito G (1998) Paleoenvironmental
Hurni H (1978) Soil erosion forms in the Simen mountains—Ethiopia changes during the last 4000 yr in the Tigray, northern Ethiopia.
(with map 1:25 000). Geogr Bernensia G8:93–100 Quatern Res 49:312–321
Hurni H (1979) Semien-Äthiopien: Methoden zur Erfassung der Mekuria W, Veldkamp E, Haile M, Nyssen J, Muys B, Gebrehiwot
Bodenerosion. Geomethodica 4:151–182 Kindeya (2007) Effectiveness of exclosures to restore degraded soils
Hurni H (1982) Klima und Dynamik der Höhenstufung von der letzten as a result of overgrazing in Tigray, Ethiopia. J Arid Environ
Kaltzeit bis zur Gegenwart. Hochgebirge von Semien-Äthiopien, 69:270–284. doi:10.1016/j.jaridenv.2006.10.009
Geogr Bernensia G13 Mersha E (2000) A desertification convention based classification of
Hurni H (1983) Soil erosion and soil formation in agricultural moisture zones of Ethiopia. Ethiop J Nat Resour 2:1–9
ecosystems Ethiopia and Northern Thailand. Mt Res Dev 3:131– Michels J, 1988. Regional political organization in the Axum-Yeha
142 area during the pre-axumite and axumite eras. In: Paper presented at
Hurni H (1985) Erosion—productivity—conservation systems in the Xth international conference on Ethiopian studies, Paris
Ethiopia. In: Proceedings 4th international conference on soil Moeyersons J (2000) Desertification and Man in Africa. Brussels. Bull
conservation, Maracay, Venezuela, pp 654–674 Séanc Acad Royal Sci Outre-Mer 46:151–170
Hurni H (1986) Guidelines for Development Agents on soil conser- Moeyersons J, Nyssen J, Poesen J, Deckers J, Haile M (2006) On the
vation in Ethiopia. Soil conservation research project. Community origin of rock fragment mulches on Vertisols: a case study from the
Forests and Soil Conservation Development Department, Ministry Ethiopian highlands. Geomorphology 76:411–429
of Agriculture, Addis Ababa Moeyersons J, Van Den Eeckhaut M, Nyssen J, Gebreyohannes T, Van
Hurni H (1990) Degradation and conservation of soil resources in the de Wauw J, Hofmeister J, Poesen J, Deckers J, Haile M (2008)
Ethiopian highlands. Mt Res Dev 8:123–130 Mass movement mapping for geomorphological understanding and
Hurni H (1993) Land degradation, famine, and land resource scenarios sustainable development, Tigray, Ethiopia. Catena 75:45–54. http://
in Ethiopia. In: Pimentel D (ed) World soil erosion and conserva- dx.doi.org/10.1016/j.catena.2008.04.004
tion. Cambridge University Press, Cambridge Moore TR (1979) Rainfall erosivity in East Africa. Geografiska
Hurni H (1998) Lessons of soil conservation research conducted in Annaler Series A 61A:147–156
Ethiopia. In: Proceedings, soil fertility management work- Moreda F, Bauwens W (1998) Influences of variability of rainfall on
shop. National Fertiliser Industry Agency, Addis Ababa, pp 31–35 flow regimes in central Ethiopia. IAHS Publication 252:297–306
Jacob M, Frankl A, Haile M, Zwertvaegher A, Nyssen J (2013) Mulugetta B (1992) Attempts in the transformation of Ethiopia’s agriculture.
Assessing spatio-temporal rainfall variability in a tropical mountain In: Doornbos M, Cliffe L, Abel Ghaffar M and Markakis J (eds) Beyond
area (Ethiopia) using NOAAs Rainfall Estimates. Int J Remote Sens conflict in the Horn. The Red Sea Press, Trenton, pp 143–153
34:8305–8321 Munro RN, Deckers J, Grove AT, Haile M, Poesen J, Nyssen J (2008)
Joyce F de V (1943) Notes on agriculture in Ethiopia, Part 3. East Afr Soil and erosion features of the Central Plateau region of Tigrai—
Agric J 9:35–38 learning from photo monitoring with 30 years interval. Catena
Kanwar JS, Virmani SM (1986) Management of Vertisols for improved 75:55–64. doi:10.1016/j.catena.2008.04.009
crop production in the semi-arid tropics: a plan for a technology Muys B, Nyssen J, du Toit B, Vidale E, Prokofieva I, Mavsar R, Palahi
transfer network in Africa. In: Proceedings of the first IBSRAM M (2014) Water-related ecosystem services of forests: learning from
(International Board on Soil Research and Management) Regional regional cases. In: Katila P, Galloway G, de Jong W, Pacheco P,
Networkshop in Africa on Improved Management of Vertisols Mery G (eds) Forests under pressure—local responses to global
issues. International Union of Forest Research Organizations, Deckers J (2009a) How soil conservation affects the catchment
Vienna, pp 423–440 sediment budget—a comprehensive study in the north Ethiopian
Mwendera EJ, Saleem MAM (1997) Infiltration rates, surface runoff, highlands. Earth Surf Proc Land 34:1216–1233. doi:10.1002/esp.
and soil loss as influenced by grazing pressure in the Ethiopian 1805
highlands. Soil Use Manag 13:29–35 Nyssen J, Haile M, Nauds J, Munro N, Poesen J, Moeyersons J, Frankl
Mwendera EJ, Saleem MAM, Dibabe A (1997) The effect of livestock A, Deckers J, Pankhurst R (2009b) Desertification? Northern
grazing on surface runoff and soil erosion from sloping pasture Ethiopia re-photographed after 140 years. Sci Total Environ
lands in the Ethiopian highlands. Aust J Exp Agric 37:421–430 407:2749–2755. doi:10.1016/j.scitotenv.2008.12.016
Nyssen J (2001) Erosion processes and soil conservation in a tropical Nyssen J, Poesen J, Haile M, Moeyersons J, Deckers J, Hurni H
mountain catchment under threat of anthropogenic desertification—a (2009c) Effects of land use and land cover on sheet and rill erosion
case study from Northern Ethiopia. PhD thesis. KU Leuven, Belgium rates in the Tigray highlands, Ethiopia. Zeitschrift für Geomorphol-
Nyssen J, Haile M, Moeyersons J, Poesen J, Deckers J (2000a) Soil and ogie 53:171–197
water conservation in Tigray (Northern Ethiopia): the traditional Nyssen J, Clymans W, Descheemaeker K, Poesen J, Vandecasteele I,
daget technique and its integration with introduced techniques. Vanmaercke M, Haile M, Haregeweyn N, Moeyersons J, Martens
Land Degrad Dev 11:199–208 K, Zenebe A, Van Camp M, Gebreyohannes T, Deckers J,
Nyssen J, Moeyersons J, Deckers J, Haile M, Poesen J (2000b) Vertic Walraevens K (2010) Impact of soil and water conservation on
movements and the development of stone covers and gullies, Tigray catchment hydrological response—a case in northern Ethiopia.
highlands, Ethiopia. Zeitschrift für Geomorphologie 44:145–164 Hydrol Process 24:1880–1895
Nyssen J, Poesen J, Haile M, Moeyersons J, Deckers J (2000c) Tillage Oicha T, Cornelis W, Verplancke H, Nyssen J, Deckers J, Behailu M,
erosion on slopes with soil conservation structures in the Ethiopian Haile M (2010) Short-term effects of conservation agriculture on
highlands. Soil Tillage Res 57:115–127. doi:10.1016/S0167-1987 Vertisols under tef (Eragrostis tef (Zucc.) Trotter) in the northern
(00)00138-0 Ethiopian highlands. Soil Tillage Res 106:294–302. doi:10.1016/j.
Nyssen J, Poesen J, Moeyersons J, Luyten E, Veyret-Picot M, Deckers still.2009.12.004
J, Haile M, Govers G (2002a) Impact of road building on gully Olson G (1981) Archaeology: lessons on future soil use. J Soil Water
erosion risk: a case study from the Northern Ethiopian highlands. Conserv 36:261–264
Earth Surf Process Landf 27:1267–1283 Pietsch D, Machado MJ (2014) Colluvial deposits—proxies for climate
Nyssen J, Poesen J, Moeyersons J, Lavrysen E, Haile M, Deckers J change and cultural chronology. A case study from Tigray, Ethiopia.
(2002b) Spatial distribution of rock fragments in cultivated soils in Zeitschrift für Geomorphologie, Supplementary Issues 58(1):119−136
northern Ethiopia as affected by lateral and vertical displacement Poesen JW, Torri D, Bunte K (1994) Effects of rock fragments on soil
processes. Geomorphology 43:1–16 erosion by water at different spatial scales—a review. Catena
Nyssen J, Haile M, Moeyersons J, Poesen J, Deckers J (2004a) 23:141–166
Environmental policy in Ethiopia: a rejoinder to Keeley and Poppe L, Frankl A, Poesen J, Admasu T, Dessie M, Adgo E, Deckers J,
Scoones. J Modern Afr St 42:137–147 Nyssen J (2013) Geomorphology of the Lake Tana basin, Ethiopia.
Nyssen J, Poesen J, Moeyersons J, Deckers J, Haile M, Lang A (2004b) J Maps 9:431–437. doi:10.1080/17445647.2013.801000
Human impact on the environment in the Ethiopian and Eritrean Rahmato D (1994) The unquiet countryside: the collapse of “socialism”
highlands—a state of the art. Earth Sci Rev 64:273–320. doi:10. and the rural agitation, 1990 and 1991. In: Zegeye A and
1016/S0012-8252(03)00078-3 Pausewang S (eds) Ethiopia in change, peasantry, nationalism and
Nyssen J, Veyret-Picot M, Poesen J, Moeyersons J, Haile M, Deckers J, democracy. British Academic Press, London
Govers G (2004c) The effectiveness of loose rock check dams for Reid RS, Kruska RL, Muthui N, Taye A, Wotton S, Wilson CJ, Mulatu
gully control in Tigray, Northern Ethiopia. Soil Use Manag 20:55– W (2000) Land-use and land-cover dynamics in response to
64. doi:10.1111/j.1475-2743.2004.tb00337.x changes in climatic, biological and socio-political forces: the case of
Nyssen J, Vandenreyken H, Poesen J, Moeyersons J, Deckers J, Haile southwestern Ethiopia. Landscape Ecol 15:339–355
M, Salles C, Govers G (2005) Rainfall erosivity and variability in Renard KG, Foster GR, Weesies GA, McCool DK, Yoder DC (1997)
the Northern Ethiopian highlands. J Hydrol 311:172–187. doi:10. Predicting soil erosion by water: a guide to conservation planning
1016/j.jhydrol.2004.12.016 with the Revised Universal Soil Loss Equation (RUSLE). Agricul-
Nyssen J, Poesen J, Gebremichael D, Vancampenhout K, D’Aes M, ture handbook 703. United States Department of Agriculture,
Yihdego G, Govers G, Leirs H, Moeyersons J, Naudts J, Washington, DC
Haregeweyn N, Haile M, Deckers J (2007a) Interdisciplinary on- Reubens B, Poesen J, Nyssen J, Leduc Y, Zenebe A, Tewoldeberhan S,
site evaluation of stone bunds to control soil erosion on cropland in Bauer H, Gebrehiwot K, Deckers J, Muys B (2009) Establishment
Northern Ethiopia. Soil Tillage Res 94:151–163 and management of woody seedlings in gullies in a semi-arid
Nyssen J, Poesen J, Moeyersons J, Haile M, Deckers J (2007b) environment (Tigray, Ethiopia). Plant Soil 324:131–156
Dynamics of soil erosion rates and controlling factors in the Reuter G (1991) Bodenkundliche Probleme im Irak und in Äthiopien.
Northern Ethiopian highlands—towards a sediment budget. Earth Mitteilungen der Deutschen Bodenkundlichen Gesellschaft 66
Surf Proc Land 33:695–711. doi:10.1002/esp.1569 (2):1181–1184
Nyssen J, Naudts J, De Geyndt K, Haile M, Poesen J, Moeyersons J, Riché G, Ségalen P (1973) Les sols et le modelé dans le Nord-Est du
Deckers J (2008a) Soils and land use in the Tigray highlands bassin du Wabi-Schebelle (Ethiopie). Cahiers ORSTOM, série
(Northern Ethiopia). Land Degrad Dev 19:257–274. doi:10.1002/ pédologie 11:237–247
ldr.840 SCRP (2000) Long-term monitoring of the agricultural environment in
Nyssen J, Poesen J, Descheemaeker K, Haregeweyn N, Haile M, six research stations in Ethiopia. Soil erosion and conservation
Moeyersons J, Frankl A, Govers G, Munro RN, Deckers J (2008b) database. 7 volumes. Soil conservation research programme, Berne
Effects of region-wide soil and water conservation in semi-arid and Addis Ababa
areas: the case of northern Ethiopia. Zeitschrift für Geomorphologie Seleshi Y, Demarée G (1995) Rainfall variability in the Ethiopian and
52:291–315. doi:10.1016/j.catena.2008.04.009 Eritrean highlands and its links with the Southern oscillation index.
Nyssen J, Clymans W, Poesen J, Vandecasteele I, De Baets S, J Biogeogr 22:945–952
Haregeweyn N, Naudts J, Hadera A, Moeyersons J, Haile M,
Seleshi Y, Demarée G (1998) Identifying the major cause of the UNEP (1994) United Nations convention to combat desertification.
prevailing summer rainfall deficit over the North-Central Ethiopian United Nations Environmental Programme, Nairobi
highlands since the mid-60s. In: Demarée G, Alexandre J and De USBR (1964) Land and water resources of the Blue Nile basin.
Dapper M (eds) Proceedings of internatonal conference tropical Appendix III. Hydrology. United States Department of Interior,
climate, meteorology, hydrology, Brussels 22–24 May 1996, Bureau of Reclamation
pp 516–531 Valentin C, Poesen J, Li Y (2005) Gully erosion: impacts, factors and
Shitarek T, Manaye Sintayehu, Abebe Berihun (2001) Stengthening control. Catena 63:132–153
user-rights over local resources in Wollo. IIED drylands pro- Van de Wauw J, Baert G, Moeyersons J, Nyssen J, De Geyndt K, Taha
gramme, issue paper, Ethiopia 103 N, Zenebe A, Poesen J, Deckers J (2008) Soil-landscape relation-
Ståhl M (1974) Ethiopia: political contradictions in agricultural ships in the basalt-dominated highlands of Tigay, Ethiopia. Catena
development. Stockholm, Rabén and Sjögren 75:117–127
Ståhl M (1990) Environmental degradation and political constraints in Van Den Eeckhaut M, Moeyersons J, Nyssen J, Amanuel Abaha, Poesen J,
Ethiopia. Disasters 14:140–150 Haile M, Deckers J (2009) Spatial patterns of old, deep-seated
Sutcliffe JV, Parks YP (1999) The hydrology of the Nile. IAHS Special landslides: a case-study in the northern Ethiopian highlands.
Publication 5. International Association of Hydrological Sciences, UK Geomorphology 105:239–252. http://dx.doi.org/10.1016/j.geomorph.
Taddesse B (1995) Deforestation and environmental degradation in 2008.09.027
Ethopia: the case of Jam Jam province. Northeast Arican Studies 2 Vancampenhout K, Nyssen J, Gebremichael D, Deckers J, Poesen J,
(2):139-156 Haile M, Moeyersons J (2006) Stone bunds for soil conservation in
Taye G, Poesen J, Van Wesemael B, Vanmaercke M, Teka D, Deckers the northern Ethiopian highlands: impacts on soil fertility and crop
J, Goosse T, Maetens W, Nyssen J, Hallet V, Haregeweyn Nigussie yield. Soil Tillage Res 90:1–15. doi:10.1016/j.still.2005.08.004
(2013) Effects of land use, slope gradient, and soil and water Virgo KJ, Munro RN (1978) Soil and erosion features of the Central
conservation structures on runoff and soil loss in semi-arid Northern Plateau region of Tigrai, Ethiopia. Geoderma 20:131–157
Ethiopia. Phys Geogr 34:236–259. doi:10.1080/02723646.2013. Vrieling A, Sterk G, de Jong SM (2010) Satellite-based estimation of
832098 rainfall erosivity for Africa. J Hydrol 395(3):235
Tedla A, Mamo T, Klaij M, Diedhiou M (1999) Effects of cropping Walraevens K, Vandecasteele I, Martens K, Nyssen J, Moeyersons J,
system, seed bed management and fertility interactions on biomass Gebreyohannes T, De Smedt F, Poesen J, Deckers J, Van Camp M
of crops grown on a Vertisol in the central highlands of Ethiopia. (2009) Groundwater recharge and flow in a small mountain
Zeitschrift Für Acker Und Pflanzenbau 183:205–211 catchment in Northern Ethiopia. Hydrol Sci J 54:739–753
Tegene B (1998) Pedogenesis and soil-geomorphic relationships on the Wischmeier WH, Smith DD (1978) Predicting rainfall erosion losses: a
piedmont slopes of Wurgo Valley, Southern Welo. Ethiop Sinet guide to conservation planning. Agriculture handbook, 537. United
Ethiop J Sci 21:91–111 States Department of Agriculture, Washington, DC
Tesfaye M (1988) Soil conservation experiments on cultivated land in Wondumagegnehu F, Tsegay A, Ashebir D, Tekie H, Gebre A, Kiros M,
the Maybar Area, Wello Region, Ethiopia. Soil conservation Geerts S, Raes D, Nyssen J, Deckers J (2007) Household water
research project, research report 16. University of Berne, Berne harvesting structures in Geba catchment. Tigray Livelihood Papers, vol
Troll C (1970) Die naturräumliche Gliederung Eritreas. Erdkunde 5. VLIR—Mekelle University IUC Programme, Mekelle, Ethiopia
24:249–268 Zenebe A, Vanmaercke M, Poesen J, Verstraeten G, Haregeweyn N,
Uhlig S, Uhlig K (1989) On the ecology and vegetation of the plateaus Haile M, Amare K, Deckers J, Nyssen J (2013) Spatial and temporal
on east Ethiopian isolated mountains. Archiv für Naturschutz und variability of river flows in the degraded semi-arid tropical
Landschaftsforschung 29:175–179 mountains of northern Ethiopia. Z Geomorph NF 57:143–169
Index
A D
Afro-montane, 8 Danakil depression, 70
Aggradation, 174 Danakil desert, 70
Alluvial fans, 168, 215 Debris flows, 361
Alluvial plain, 95 Debris slides, 359
Amba, 15, 34 Deflation, 377
Anabranching, 219 Deforestation, 9, 379
Andosols, 145 Degradation, 174
Antidunes, 221 Delta, 94
Ard plough, 157 Desertification, 377
Arenosols, 5 Desiccation cracks, 209
Aridity index, 85 Distributary systems, 95, 214
Arroyos, 174 Doline, 171
Artifacts, 50 Dome, 19, 47, 54
Drought, 68
Drylands, 213
B Dunes, 221
Bankfull, 218
Bankfull discharge, 102
Bedforms, 220 E
Bedload, 113 Earthflows, 359
Bedload sheets, 221 Endorheic, 90
Bega, 68 Endorheic basin, 289
Beheaded valleys, 129 Endorheic lakes, 55
Belg, 68 Ephemeral streams, 90
Bifurcation, 219 Erodibility, 6
Block fields, 150 Errace, 169
Boulder slopes, 150 Erratic boulders, 27
Boulder berm, 96 Etchplain, 150
Boulder talus, 150 Eucalyptus, 10
Eutric Vertisols, 5
Evapotranspiration, 68, 216
C Exclosures, 197, 378
Calcrete, 333
Caldera, 26, 48
Cambisols, 5 F
Canyon, 99, 324 Fan deltas, 243
Castellated tors, 150 Flash floods, 193
Check-dams, 193, 364 Floodouts, 220
Chilga beds, 54 Fluvial terraces, 243
Cinder cone, 25, 267 Fluvisols, 5
Cirques, 142
Climate change, 202
Concavity index, 340 G
Corestones, 151 Glacial deposits, 48
Crater lakes, 276 Gorge, 98, 215
Crystalline basement, 36 Gravitational spreading, 334
DOI 10.1007/978-94-017-8026-1, © Springer Science+Business Media Dordrecht 2015
388 Index
Ground fissures, 353 R

Gully, 191 Rainfall anomalies, 84
Gully erosion, 193 Rainfall erosivity, 370
Rainfall extremes, 68
Rainfall intensity, 75
H Rectangular drainage network, 94
Hohlkarren, 171 Regolith, 151
Hominid fossils, 36 Repeat photography, 195
Hominids, 53 Reservoir, 92
Hot springs, 284 Rhyolite plugs, 47
Humic Nitisols, 5 Riffle-pool, 96
Hydrothermal springs, 281 Rillenkarren, 171
Ripples, 221
River entrenchment, 25
I Roche mountonnées, 27
Inselbergs, 38 Rock cliffs, 100, 150
Intertrappean beds, 48 Rockfall, 179
Intertropical Convergence Zone, 68 Rotational slides, 361
Isotopic fingerprinting, 308 Runoff Concentration Index, 93
K S
Karren, 171 Saline water, 309
Kiremt, 68 Sapping, 154
Kopjes, 156 Saprolite, 153
Köppen classification, 81 Sediment concentration, 228
Sediment yield, 6, 94, 228
Sedimentary cycles, 118
L Seismic zoning, 362
Lacustrine terrace, 25, 243 Shear stress, 96
Land degradation, 377 Shield volcanoes, 25, 44, 48
Landslide, 360 Siltation, 235
Laterite, 54, 167 Slope deposits, 143
Lithic Leptosols, 5 Slope length, 374
Little Ice Age, 247 Slope steepness factor, 374
Slump, 359
Soil erosion, 112
M Soil piping, 193
Maar, 284 Solifluction, 143
Mesa, 100 Stelae, 147
Mid-slope benches, 150 Step-pool, 96
Modified Fournier Index, 75 Stone bunds, 376
Monsoon, 73 Stony bunds, 156
Moraines, 27 Strato-volcanoes, 25, 271
Mudflows, 359 Stream power, 113
Stromatolitic biostromes, 240
Suspended sediment, 111
O Swamps, 94
Obelisks, 19
T
P Tableland, 333
Palaeosoil, 119 Talus, 168
Palaeovalley, 130 Tarns, 27
Paraconformity, 121 Temperature increase, 84
Particle clusters, 96 Terminal fans, 95
Peneplain, 119 Terminal splay, 219
Peneplanation, 3 Terra rossa, 169
Physiography, 95 Thermal amplitude, 69
Planation surfaces, 118 Thermochronometry, 131
Plugs, 150 Tillage erosion, 376
Pothole, 202 Time series, 67
Precipitation trends, 84 Toppling, 359
Index 389
Tors, 172 W
Trachyte plugs, 12 Water balance, 279
Trade winds, 68 Waterfall, 98
Transmission losse, 372
Transverse ribs, 96
Tufa dams, 201 X
Xerosols, 5
U
Unconformities, 120 Y
Updoming, 3 Yermosols, 5
V Z
Variation coefficient, 74 Z score, 70
Vertic, 5
Vertisols, 154
Volcanic plugs, 28
Vulnerability, 180

Landscapes and Landforms of Ethiopia: Paolo Billi Editor

Uploaded by

Copyright:

Available Formats

Landscapes and Landforms of Ethiopia: Paolo Billi Editor

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Landscapes and Landforms of Ethiopia: Paolo Billi Editor

Uploaded by

Copyright:

Available Formats

What are some of the major landforms and landscapes discussed in the document?

What are some of the major landforms and landscapes discussed in the document?

What are some of the geomorphological processes mentioned that shape landscapes?

What are some of the geomorphological processes mentioned that shape landscapes?

World Geomorphological Landscapes

Paolo Billi Editor

For further volumes:

Landscapes and Landforms

ISSN 2213-2090 ISSN 2213-2104 (electronic)

Library of Congress Control Number: 2015931921

Springer Dordrecht Heidelberg New York London

Printed on acid-free paper

Springer Science+Business Media B.V. Dordrecht is part of Springer Science+Business Media

Part I General Introduction to the Geomorphology of Ethiopia

1 Geomorphological Landscapes of Ethiopia . . . . . . . . . . . . . . . . . . . . . . . . . 3

2 Geology of Ethiopia: A Review and Geomorphological Perspectives . . . . . . 33

3 The Climate of Ethiopia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

5 Planation Surfaces and the Long-term Geomorphological

Part II Local Studies

6 Paleoglaciated Landscapes in Simen and Other High-Mountain

7 Geomorphology of the Archaeological Area of Aksum . . . . . . . . . . . . . . . . 147

8 Geomorphology of the Adwa District . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163

9 The Amba Landscape of the Ethiopian Highlands, Shaped by Rockfall . . . . 179

10 Gully Development in the Tigray Highlands. . . . . . . . . . . . . . . . . . . . . . . . 191

11 Tufa Dams in Tigray (Northern Ethiopia) as Late Pleistocene—Holocene

12 Geomorphology of Ephemeral Streams in the Kobo Basin . . . . . . . . . . . . . 213

13 Sediment Yield Variability at Various Spatial Scales and Its

14 Climatic and Hydrologic Changes in Northern Ethiopia

15 Rift-Related Morphology of the Afar Depression . . . . . . . . . . . . . . . . . . . . 251

16 Morphometric Characteristics and Hydrology

17 The Geomorphology of the Lake Region (Main Ethiopian Rift):

18 Water–Rock Interaction and Lake Hydrochemistry

19 The Landscape and Landforms of the Ogaden, Southeast Ethiopia . . . . . . . 323

Part III Applied Aspects

20 Geo-hazard in Ethiopia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 351

21 Land Degradation in the Ethiopian Highlands . . . . . . . . . . . . . . . . . . . . . . 369

1.1 Introduction 1.2 Geography

P. Billi (ed.), Landscapes and Landforms of Ethiopia, World Geomorphological Landscapes, 3

1.2.2 Soil Erosion Table 1.2 Soil loss in Ethiopia

Fig. 1.2 Severe land degradation

Fig. 1.3 A deep gully incised

Fig. 1.4 Block sliding apart due

Fig. 1.5 A mushroom stone

Fig. 1.6 The Aba Samule

Fig. 1.7 Example of a natural

Fig. 1.8 A natural Acacia forest

Fig. 1.9 Example of a bushed

Fig. 1.10 Cereals cultivations on

growing capital city, Addis Ababa. In the 1970s, the plan-

Fig. 1.11 Pasture on the western

Fig. 1.12 Land cover

Fig. 1.13 A huge sycamore tree

Fig. 1.14 Eucalyptus forest on

Fig. 1.15 A small farm on the

Fig. 1.16 Granite inselbergs,

Fig. 1.17 Granite inselbergs

Fig. 1.18 Structural basins

Fig. 1.19 Recent (Quaternary?)

Fig. 1.20 Impressive fault