L-03 Land Cover of Selected Areas in Somaliland and Southern Somalia

The designations employed and the presentation of material in this information product
do not imply the expression of any opinion whatsoever on the part of the Food and
Agriculture Organization of the United Nations concerning the legal status of any country,
territory, city or area or of its authorities, or concerning the delimitation of its frontiers or
boundaries.
This document should be cited as follows:
Monaci, L., Downie, M., Oduori, S. 2007. Land Cover of Selected Study Areas in
Somaliland and Southern Somalia. FAO-SWALIM. Project Report No L-03. , Nairobi, Kenya.
Photointerpreters: Simon Mumuli Oduori, Anthony Ndubi, Paola Codipietro and Ronald
Vargas
Table of Contents
TABLE OF CONTENTS
List of Acronyms
Acknowledgment
1 INTRODUCTION ............................................................................................................................................ 1
2 CONCEPTUAL FRAMEWORK ................................................................................................................... 4
3 STUDY AREAS ................................................................................................................................................ 6
3.1 Introduction.........................................................................................6
3.2 General features of the Northern Area of Interest (NAOI) ....................6
3.2.1 Climate................................................................................................................................................ 6
3.2.2 Geology/Lithology ............................................................................................................................. 9
3.2.3 Landform and Soils ............................................................................................................................ 9
3.2.4 Land cover ........................................................................................................................................ 10
3.2.5 Land Use........................................................................................................................................... 11
3.2.6 Population......................................................................................................................................... 11
3.3 Southern Area of Interest (SAOI) .......................................................13
3.3.1 Climate.............................................................................................................................................. 13
3.3.2 Geology............................................................................................................................................. 15
3.3.3 Landform/Soils ................................................................................................................................. 15
3.3.4 Land Cover ....................................................................................................................................... 17
3.3.5 Land Use........................................................................................................................................... 18
4 MATERIALS AND METHODS................................................................................................................... 19
4.1 Materials ............................................................................................19
4.1.1 Bibliographic data ............................................................................................................................ 19
4.1.2 Ancillary data ................................................................................................................................... 19
4.1.3 RS data.............................................................................................................................................. 19
4.1.4 Software............................................................................................................................................ 19
4.2 Methodology ......................................................................................20
4.2.1 Preparatory phase ............................................................................................................................. 20
4.2.2 Base map production ........................................................................................................................ 23
4.2.3 Field data collection ......................................................................................................................... 30
4.2.4 Legend updating and map revision .................................................................................................. 31
4.2.5 Data analysis and spatial modelling................................................................................................. 32
5 RESULTS ........................................................................................................................................................ 36
5.1 Northern AOI results from Survey Validation......................................36
5.2 Land Cover cartographic aggregations................................................37
5.3 Land Cover surface statistics ..............................................................41
5.3.1 Dur-Dur and Gebiley AOI northern AOI ..................................................................................... 41
5.3.2 Riverine area southern AOI .......................................................................................................... 45
6 CONCLUSIONS AND RECOMMENDATIONS....................................................................................... 51
7 THE VEGETATION MAP OF THE NAOI................................................................................................ 53
7.1 Introduction.......................................................................................53
7.2 Materials and Methods .......................................................................53
7.2.1 Trees and shrubs ............................................................................................................................... 53
7.2.2 Herbaceous Vegetation .................................................................................................................... 53
7.3 Results ...............................................................................................56
7.4 Conclusions and Recommendations ....................................................60
8 BIBLIOGRAPHY........................................................................................................................................... 61
ANNEXES................................................................................................................................................................ 63
Annex 1 - From Land Cover to Land Use map................................................................................................... 64
Annex 2 Unverified Land Cover map of the Juba and Shabelle riverine area at 1:100 000 scale produced
during SWALIM Phase I ....................................................................................................................................... 65
Annex 3 Boundaries and mapping scale of the two areas of interest ............................................................ 66
Annex 5..................................................................................................................................................................... 68
Annex 6..................................................................................................................................................................... 69
Annex 7..................................................................................................................................................................... 70
Annex 8..................................................................................................................................................................... 71
Table of Contents
Annex 9 Land Cover Pictures from the NAOI................................................................................................. 72

Annex 10 Somalia Land Cover legend .............................................................................................................. 92
List of Tables
Table 1: Regions, districts, and their populations (Somalia UNDP 2005, draft version) .. 11
Table 2: CCs images produced by SWALIM GIS unit ................................................ 21
Table 3: Land Cover codes................................................................................... 37
Table 4: Rangeland areas .................................................................................... 42
Table 5: Wooded vegetation areas ........................................................................ 43
Table 6: Agriculture areas.................................................................................... 44
Table 7: Areas of Wooded vegetation .................................................................... 47
Table 8: Rangeland areas .................................................................................... 48
Table 9: Agriculture areas.................................................................................... 50
List of Figures
Figure 1: Rainfall and potential Evapotranspiration .................................................... 8

Figure 2: Relative Humidity of the Study Area .......................................................... 8
Figure 3: Rainfall distribution and climate classification of the study area ..................... 9
Figure 4: Carta geobotanica dellAfrica Orientale by Rodolfo E.G. Pichi Sermolli (Map of
the vegetation of East Africa) ............................................................................... 12
Figure 5: Mean Monthly Rainfall Patterns in the study area (1963-2001) .................... 14
Figure 6: Mean annual rainfall distribution map and climate of study area................... 16
Figure 7: Summary of the legend by main Land Cover type ...................................... 24
Figure 8: Photointerpretation ............................................................................... 25
Figure 9: Flowchart............................................................................................. 29
Figure 70: Cartographic aggregations.................................................................... 35
List of Charts
Chart 1: Distribution of main aggregations of Land Cover types ................................ 41

Chart 2: Rangelands distribution........................................................................... 42
Chart 3: Wooded vegetation types distribution ....................................................... 43
Chart 4: Agriculture distribution ........................................................................... 44
Chart 5: Distribution of main aggregations of Land Cover types ................................ 45
Chart 6: Wooded vegetation types distribution ....................................................... 46
Chart 7: Rangelands distribution........................................................................... 48
Chart 8: Agriculture distribution ........................................................................... 49
Table of Contents
List of acronyms
FAO Food and Agriculture Organisation of the United Nations

SWALIM - Somalia Water and Land Information Management System
AOI Area of Interest
LCCS Land Cover Classification System
LMU - Land Mapping Unit
NASA National Aeronautics and Space Administration
SRTM Shuttle Radar Topography Mission
DEM Digital Elevation Model
UNDP United Nations Development Project
DIMU UNDP Data and Information Management Unit
GIS Geographical Information System
RS Remote Sensing
UNEP United Nations Environment Programme
MSS Multispectral Scanner
VNIR - Very Near Infrared Radiometer
SVNIR - Short Wave Infrared Radiometer
IAO - Istituto Agronomico DOltremare (Florence)
IUCN - The World Conservation Union
EOS - Earth Observation System
CCs - Colour Composites
FCCs - False Colour Composites
NDVI Normalized Differential Vegetation Index
ETM Enhanced Thematic Mapper
DB Database
GPS Global Positioning System
SDRN Environment and Natural Resources Service (Sustainable Development
Department)
PAT Polygon Attribute Table
FAT - Feature Attribute Table
asl above sea level
Acknowledgements
Acknowledgements
The authors wish to acknowledge the considerable support given by the Project Task
Force during preparation of this report.
Many thanks to the ecologists Abdulkadir Aden, Mohamed Ibrahim and Ahmed Jama
Sugulle, who worked tirelessly collecting field data. The survey team benefited
immensely from the support of many people whose collective effort contributed to the
success of the field survey; their help is sincerely acknowledged.
Special thanks are due to the FAO staff in the Hargeisa office: Mohamed Jama Gahayr,
Amran Ibrahim and Asha the FAO-Hargeisa office secretary who provided effective
administrative and logistic backup, obtaining daily mission security clearances, hiring
vehicles, payments, purchasing of materials, record-keeping and meeting arrangements.
Sincere acknowledgement also to the Somali drivers, and the Somali Organisations that
were contacted for information are also thanked, including the Somali Mine Action
Centre, Ministry of Agriculture and Cadastral Survey.
Introduction
1 INTRODUCTION
This report is one of the outputs of Activity 2.B.8 of the FAO Somalia Water And Land
Information Management Project (SWALIM Phase II).
The main focus is the description of Land Cover mapping undertaken by the SWALIM
Photointerpretation unit of two Areas of Interest (AOI) in Somaliland and southern
Somalia.
This report includes descriptions of the methodology and the results obtained for:
- the land cover of the northern and southern AOI, with databases verified by field
survey
Due to worsening security in the riverine areas of southern Somalia, the anticipated
collection of field data necessary for validating the Land Cover map was delayed for the
southern AOI. Therefore, the results for the northern AOI are based on field survey of
May-June 2006 whereas those of the southern AOI are based on field survey in March
2007.
As indicated in the Implementation Plan of the SWALIM project, activities for the
production of land baseline information include an Assessment of Land Suitability and
Agricultural Production Potential for the Riverine Areas in Southern Somalia and for the
Dur-Dur/Gebiley AOI in North Western Somalia (Activity 2B.8).
Land cover is defined as the observed (bio) physical cover on the earths surface (LCCS
classification concepts and user manual FAO, 2005).
In order to predict behaviour of land in response to specific patterns of utilisation, data

on landform, land cover, land use, soil and climate need to be collected and properly
integrated into the process of generating Land Mapping Units.
A Land Mapping Unit (LMU) is a mapped area of land, homogeneous according to the
above-mentioned environmental factors, and is described in terms of its qualities. The
LMU is the spatial unit of analysis for evaluation of land suitability.
A Land Cover map contains data about the distribution of vegetation, agriculture, bare
areas and built up surfaces. In addition to its use as a thematic layer during suitability
analyses, it is also used in the generation of land use maps.
Transformation from a Land Cover to a Land Use map takes place through a link between
Land Cover classes and available socio-economic data (Annex 1).
The project document indicated two areas for assessment of physical land suitability: the
Dur-Dur and Gebiley AOI in Somaliland, and riverine areas in southern Somalia. The
detailed boundaries of these two AOI were delineated considering existing available
Somali territory datasets and also taking into account areas mapped by the unverified
1:100 000 Land Cover map of the Juba and Shabelle riverine areas produced during
SWALIM Phase I (Annex 2).
1
Introduction
1.1. Delineation of the study area
1.1.1 The Northern AOI

The northern AOI was defined following evaluation of the NASA 90 meters SRTM-DEM.
The delineation took into account the following areas:
- sub-catchments extracted from the DEM;

- agricultural areas identified from available satellite imagery.
Following morphological criteria, the following areas were selected:
- the basin including the town of Baki (Dur-Dur is the main togga within this area);
- the basin including the towns of Hargeisa and Arabsiyo (the main toggas of this
area are Hog, Kadar and Biji);
- certain sub-catchments on the eastern side of interest due to presence of irrigated
cultivation;
- certain sub catchments on the southern-eastern side of interest due to rainfed
cultivation and tiger bush areas.
The limits of the northern AOI are:
- the coast in the north;

- the sub-catchment of Togga Maroodijeex, including Hargeisa town, to the east;
- the boundaries with Ethiopia in the south;
- the limit of the Dur-Dur watershed in the north-west.
1.1.2 The Southern AOI

In the southern AOI the following data were evaluated:
- NASA 90 meters SRTM-DEM;

- FAO Africover - Landform;
- UNDP/DIMU Flooded areas data set 1981 and 1987
- FAO SWALIM Phase I - Land Cover;
- available satellite imagery.
Delineation took into account the following areas:
- watershed limits extracted from the NASA SRTM DEM;

- alluvial plains (code L11ap) and floodplains (both L11wa and L11we) extracted
from the Africover Landform;
- agricultural areas extracted from the Land Cover prepared during SWALIM Phase I
image interpretation.
After the above layers were overlapped, morphological criteria guided selection of the
shape of the study area. The Juba and Shabelle rivers form a unique watershed that can
be subdivided into many sub-watersheds.
2
Introduction
Taking into consideration the new type of information product requested in the project
document (Activity 2B.8) and the land cover map prepared during SWALIM Phase I, the
following specifications were adopted for defining the area of interest:
- the eastern boundary of the AOI corresponds to the Shabelle main watershed
limit;
- the southern boundary corresponds to the coastline;
- the western boundary corresponds to the Juba main watershed limit, not including
a tributary entering from Kenya;
- the northern boundary corresponds to the Somalia/Ethiopia border;
- the central, empty, part of the AOI has been excluded following the eastern main
watershed limits of the Juba, and the limits of the alluvial plain defined by the
FAO-Africover layer.
- If an important agricultural area was found to be outside of watershed limits but
in proximity to it, then it was included (Annex 3).
Comparing the AOI delineated during Phase I with the revised one of Phase II, the
following were the major modifications:
- addition of some areas on the north-eastern and eastern boundary of the AOI in
order to fully include the limits of the Shabelle main watershed;
- exclusion of a tributary coming from Kenya on the western boundary that was
only partially included in the Land Cover map produced during SWALIM Phase I;
- addition of a substantial area on the northern boundary in order to fully include
the limits of the Juba main watershed;
- addition of the Lower Shabelle area along the coast, which was not interpreted
during Phase I due to unavailability of cloud free satellite imagery;
- elimination of areas partially included in Xudur and Tayeglow districts (Bakool
region) in the central region.
It is important to highlight that these changes were made only with regard to the extent
of the area where the suitability assessment was to be performed.
Land Cover map boundaries as shown in Annex 4 are the result of the addition of new
areas interpreted during SWALIM Phase II to the land cover map originally produced
during Phase I.
1.1.3 General
The scale of work adopted for production of the land cover map and all thematic layers
needed for the assessment of the physical land suitability in agricultural areas in western
Somaliland and in some pilot areas in southern Somalia (still to be selected according to
field accessibility), was 1:50 000. Land cover, landform, land use and soil maps related
to the two AOI were produced at a scale of 1:100 000 (see Annex 3).
3
Conceptual Framework
2 CONCEPTUAL FRAMEWORK
The main issues that had to be addressed to produce the land cover map for the two AOI
were:
- lack of available data on land cover at a scale compatible with the expected
results (1:100 000 and 1:50 000 in the selected agricultural areas);
- field survey constraints in terms of accessibility (road condition, seasonality and
insecurity);
- the vast area to be investigated (12 939 km2 for the northern AOI and 103 976
km2 for the southern AOI) at a semi-detailed working scale;
- available time and human resources;
- the need to set up an adequate survey campaign to inventory the land resources.
The solution adopted was to produce a land cover map through on-screen interpretation
of satellite imagery and validated through collection of field data, the latter being a
crucial phase in validating information collected during the interpretive activities. Some of
the advantages of this methodology were:
- consistent stratification of the land into homogenous areas, providing boundaries

that separate different land cover types through identification of useful
interpretative keys. In this phase it is more important to assign consistent codes
following the same rules throughout the AOI, than to assign well-chosen codes.
This will facilitate updating of the adopted legend following field survey/s.
- stratification of the land cover optimises the selection of samples for field survey.
When a sample is optimised it can provide more information, thereby maximising
results.
- having a standard methodology available for employment in possible new AOI or
for periodical monitoring of present ones.
- using the most recent remote sensing and GIS technologies, in particular very
high-resolution images and DEM. Information extracted from these technologies is
increasingly accurate and reliable, enhancing the above-mentioned advantages.
The production of the Land Cover map can be subdivided into five main phases:
1) Preparatory phase:
- Preliminary overview of the AOI
- Remote sensing data acquisition
- Image processing
- Setting up of the preliminary legend
2) Base map production:

- Photointerpretation activity
3) Field data collection:

- Field work
- Field data analysis
4) Legend updating and Map revision:
4
Conceptual Framework
- Legend updating
- Revision of the map
5) Data analysis and spatial modelling:
- Aggregations
- Areas subdivision (tabular) analysis
A detailed description of each phase is provided in Chapter 4 - Material and Methods.

Production of the land cover maps (1:50 000 and 1:100 000 scales) of the Dur-Dur and
Gebiley AOI in Somaliland involved all the above-mentioned phases.
A slightly different, partial sequence of phases was adopted for riverine areas. New areas
added on the basis of morphological criteria adopted for the delineation of the AOI were
photointerpreted, and then matched to the unverified map produced during Phase I. Two
cloud-free images of the lower Shabelle area were acquired, and the interpretation made
during Phase I was revised accordingly.
Some of the considerations made during the revision of the legend in the northern AOI
following the results of the field data collection were also applied to the south, where a
revision of the adopted codes was performed as is explained in detail in Chapter 4.
THE OBJECTIVE OF THE STUDY

The objective of this study was:
To produce the land cover map for the northern and southern AOI, with their
databases
5
Study Areas
2. STUDY AREAS
2.1 Introduction
The following descriptions are a brief introduction to the climate, vegetation and
agriculture of the area, as the main environmental components determining land cover.
The two AOIs will be introduced separately.
All reported information is based on the available bibliography (see Chapter 8). Among
the available ancillary data were two vegetation distribution maps: the Carta geobotanica
dellAfrica Orientale (Eritrea Ethiopia, Somalia) produced by R.E.G. Pichi-Sermolli at a
scale of 1:5 000 000 (Pichi-Sermolli, 1957), and the Generalized map of vegetation
zones of Somaliland produced by C. F. Hemming (Hemming, 1966).
Both maps provide good, detailed community descriptions and locations at a small scale.
The map produced by Pichi-Sermolli, in particular, constitutes a comprehensive
description at the national level and allows a ready comparison between the two AOI
(Figure 5).
The primary factor limiting development of vegetation is rainfall. As a consequence of the

extreme and uniform climate in both of the AOI, the zonal vegetation is monotonous
and the number of plant species in vegetation communities is not high. Plant species
have developed several features to adapt to the arid and semiarid climate. To a greater
or lesser extent, they consist of xerophytes and heliophytes and are tolerant of high
calcium carbonate concentrations and poor soil conditions. Most of them occur over a
vast variety of soil types. On the other hand, their competitiveness decreases in more
developed and complex phyto-cenosis in wetter environments. The vegetation is very
resilient, and is able to recover even after serious droughts. Climax formations and
species distributions (particularly herbaceous species) have been altered, mainly by
human activities. The above-mentioned characteristics complicate determination of
vegetation subdivisions.
Most important azonal communities are found in the coastal areas in brackish and
gypsum soils, and along rivers.
Two different types of agriculture in terms of water supply and productivity, i.e. rainfed
and irrigated cultivation, are present in both AOI. Irrigated areas are mainly located in
the southern AOI, along the Juba and Shabelle rivers where there is high potential for the
development of large irrigated farming systems, while in the northern AOI only a few
scattered irrigated areas, suitable only for small or household farms, occur scattered
along seasonal streams (toggas).
2.2 General features of the Northern Area of Interest (NAOI)
2.2.1 Climate
The region lies at the extremity of the sub-Saharan semi-arid zone commonly referred to
as the Sahel, which traverses the continent from Senegal to Somalia. The climate of the
study area is hot, dry desert in the coastal plain (Lughaya and northern part of Baki
districts) and arid in Borama and surroundings. Semi-arid conditions prevail at higher
altitudes of the Al Mountains and south of Gebiley. Mean annual rainfall ranges from 200
300 mm in the coastal areas of Lughaye, to 500 600 mm in the east of Borama and
surroundings. The rest of the study area has a mean annual rainfall of 300 500 mm
(see Figure 3).
6
Study Areas
The study area lies entirely between the two subtropical anticyclone belts. The main
weather pattern is controlled by the seasonal monsoon winds, and rainfall in the area is
consequently bimodal (see Figure 1). The south-west monsoon brings the primary Gu
rains between March and June. The Gu is followed by a hot dry period called Xagaa
(June/July). The onset of the north-east monsoon marks the beginning of the short Deyr
rains from August to October followed by the cooler long Jilaal dry period between
November and February.
Temperatures in the area are influenced by altitude and the strength and temperatures
of the seasonal winds. Figure 2 illustrates how temperature decreases with increasing
altitude. At the higher altitudes of the Al Mountains and plateau areas, temperatures vary
considerably through the seasons, with a mean annual temperature of 20-24C, while the
coastal region has mean annual temperatures of 28-32C.
Relative humidity of the highlands is mostly around 40%, except during rainy periods
when it increases to 80%. High temperatures in the coastal areas combine with a high
relative humidity of more than 70% to create an exceedingly hot, humid environment.
The major winds in the study area occur during the Xagaa dry season, particularly (June
to July) and in Jilaal (December to February) every year when the weather is hot. The
hot, calm weather occurs between the monsoons (part or whole of April and part or the
whole of September). Generally, in the northwest the winds are strongest during the
southwest monsoon. Weaker winds generally occur during April-May and October-
November. Average wind speeds vary from 8 - 10 m/s, but during a large part of the
year, strong winds of up to 17 m/s occur, causing frequent dust-devils all over the
coastal plains and plateaus.
The study area is subject to high potential evapotranspiration (PET), with an annual
average of < 3 000 mm. Annual rainfall is very far below the potential evapotranspiration
and a large water deficit exists during most of the year throughout the region. This low
rainfall is not always sufficient for successful crop production (Figure 1).
Rainfall Potential Evapotranspiration Graph at Borama Rainfall Potential Evapotranspiration Graph at Hargeisa
90 300
Rain PET
120 140 80
Rain PET 250
70
100 120
60 200
Rainfall(mm)
100
PET(mm)
80
Rainfall(mm)
50
PET(mm)
80 150
60 40
60
30 100
40
40 20
20 50
20 10
0 0 0 0
JUN
DEC
OCT
FEB
SEP
JUL
SEP
DEC
NOV
OCT
JAN
MAR
APR
FEB
AUG
MAR
APR
JUN
NOV
JUL
JAN
MAY
AUG
MAY
Rainfall Potential Evapotranspiration Graph at Gebilley
90 140
Rain PET
80
120
70
100
Rainfall(mm)
60
PET(mm)
50 80
40 60
30
40
20
20
10
0 0
JUN
DEC
OCT
FEB
SEP
JUL
NOV
JAN
MAR
APR
AUG
MAY
7
Study Areas
Figure 1: Rainfall and potential evapotranspiration
Figure 2: Relative humidity of the study area
8
Study Areas
Figure 3: Rainfall distribution and climate classification of the study area
2.2.2 Geology/Lithology
The study area is covered by rocks dating from Pre-Cambrian to Recent, comprising
sedimentary, igneous and metamorphic rocks. The tectonic arrangement of rock
outcroppings in the region is complex and severely affected by many different systems of
faults and fractures, mainly oriented parallel to the coast (i.e. WNW-ESE).
The basement complex covers an extensive area of the Al Mountains around Borama and
Baki districts. In other parts of the region it is covered by Jurassic limestone and Miocene
bio-limestone, Pleistocene basalts and Recent alluvial and aeolian deposits. Igneous rocks
consist mostly of basalts and rhyolites, while metamorphic rocks include a wide range of
schists, ortogneiss, quartzite, migmatites, marble, calcosilicate and paragneiss, intruded
by granite, diorite and gabbro.
Pleistocene basalt outcrops and other volcanic rock outcrops occur dispersed along the
northern escarpment and coastal plain.
2.2.3 Landform and Soils

From a geomorphological point of view, the study area can be divided into the following
landscapes: Plateau (both dissected and normal), Mountainous and Hilland, Piedmonts
and the Coastal Plain. There are three main ephemeral river systems (Togga Durdur,
Togga Biji and Togga Waheen) that drain from the plateau, traversing the mountain
range in the direction of the Red Sea and from the southern side of the same mountain
to the southern highlands respectively.
According to the bibliography, the northern AOI is characterised by the following land
features:
1. A narrow coastal plain, washed by the waters of the Gulf of Aden which is fairly
depressed in relation to the mountain range at its rear;
2. A mountain range, oriented almost E-W, parallel to the coast, with a very rugged
topography that rises up from the coastal plain to the countrys highest peak (2
407 m, outside of the AOI);
3. Highlands and plateaus, southward of the mountain range, consisting of gently
undulating or almost flat terrains dipping toward the south-east, cut by several
wadis and with wide alluvial plains.
With regard to the hydrography, in the northern AOI there are a number of ephemeral
rivers (locally called togga, tug or wadi - Faillace, 1986), which contain water only during
the rainy season. Drainage follows the general relief from the almost E-W mountain
range to the northern Gulf of Aden coast, and also from the same mountain range to the
inner south-facing highlands. They give rise to deep gorges in the inner part of the
mountain range.
According to the Sogreah soil survey report (Sogreah, 1983), soil distribution patterns
closely follow the geomorphology of the region. On the high plateau, soils were mapped
as predominantly deep and heavy textured Vertisols. The Mountainous and Hilland areas
were mapped as rocky or covered by shallow Entisols and some Aridisols. The soils in the
Piedmont areas were classed as Entisols and Aridisols. A big portion of the region is
covered by Rocky soils that were mapped as a separate non soil class.
9
Study Areas
2.2.4 Land cover

On the basis of vegetation types described by Pichi-Sermolli (1957) (Figure 4), the
northern AOI is characterised by:
- Grass, perennial and sub-shrubs steppe (Steppa graminosa, perennierbosa e

suffruticosa)
- Shrub steppe (Steppa arbustata)
- Xerophyllous open woodland (arid zones) (Boscaglia xerofila rada)
- Xerophyllous open woodland (semi-arid zones) (Boscaglia xerofila)
- Montane evergreen thicket and scrub (Boscaglia e fruticeto sempreverdi montani)
- Savanna
- Coastal formations (Formazioni costiere)
Grass, perennial and sub-shrubs steppe (or sub-desert shrub and grass) - Includes all
those cenosis consisting of grasses, graminoids and suffrutexes. Perennial plants are
rare. Degrees of cover and floristic composition can change according to edaphic
conditions. They are common on gypsum soils. A typical species is Chrysopogon auchieri
var. quinqueplumis.
Shrub steppe (subdesert bush and thicket) - Similar to the Sub-desert shrub and grass
category, but differs from it due to the presence of a shrub layer and scattered trees.
Shrubs can be obconical or hemispherical, sometimes clustered in groups of 2-4 plants
together with smaller shrubs and herbs, forming a small thicket. There are wide empty
spaces between individual plants and plant groups.
Xerophyllous open woodland (arid zones) - An open vegetation formed by woody plants.
Shrubs 3-5 meters in height and perennial graminoids are dominant, and scattered trees
are typical. Shrubs are scattered, sometimes grouped but never impenetrable.
Suffrutexes and perennial grasses have a discontinuous cover that is filled by annual
(therophytes) after rainfall. The most common Acacias are Acacia bussei, A. etbaica, A.
nilotica, A. tortilis and A. senegal.
Xerophyllous open woodland (semi-arid zones) Of greater density and higher species
richness of both high shrubs and perennial grass when compared to the previous
vegetation formation.
Montane evergreen thicket and scrub (Evergreen shrubs) this is characterised by dense
gathering of suffrutexes (2-3 meters in height) or evergreen shrubs (3-5 meters in
height). Normally they are sclerophyllous or aphyllous. Rushes, succulent plants, lianas
and deciduous plants are present. Trees are scattered. Euphorbia candelabrum is present
in some cenosis as an emergent. This vegetation develops on slopes from 1 000 to above
2 000m asl; at lower altitudes suffrutexes dominate, while higher up where there is
greater humidity, evergreen shrubs dominate.
Savanna - Composed of annual grasses (usually xerophytes). Cyperaceae and

Graminaceae are the most common families, forming a continuous layer 80 cm high.
Trees or shrubs are usually present, but the herbaceous layer is dominant. They can be
isolated or in small groups or thickets.
Coastal formations - These associations are distinguished by various dominant life forms
according to different edaphic conditions, but they are characterised by the direct
influence of the sea. They occupy the coral beaches, sandy shores, estuaries, deltas,
coastal dunes (stabilised or mobile) and the salty plains. Typical examples are Suaeda
fruticosa, Adenium sp. on rocky areas, and Salsola foetida on deep, dry soils.
10
Study Areas
2.2.5 Land Use

The main land use in the study area is extensive grazing (pastoralism). Other land uses
include rainfed agriculture, irrigated orchards along alluvial plains, and wood collection.
Rainfed agriculture is found in what is considered as the sorghum belt of Somaliland,

practiced in combination with pastoralism and wood collection. This class of land use is
the economic basis of households in the study area.
Cultivation of irrigated orchards is a cash-oriented activity in the area, involving the

growing of fruit trees such as citrus, guava, papaya and mango. Supplementary water for
irrigating the crops is obtained from wells, dams and other water bodies.
Wood collection for charcoal production is very frequent, occurring in all well-treed areas.
Preferred tree species are Acacia bussei, A. nilotica and A. etbaica. Interventions to help
introduce sustainable sources of cooking energy are important and urgent.
Most of the area is used for extensive grazing, or pastoralism. Goats and sheep are
grazed mostly on sloping areas, whereas cattle and camels are grazed in flatter areas.
Sedentary pastoralism around homesteads is a common practice. Hay harvesting from
enclosures supports this land use, as harvested hay can be used in the dry season.
However, hay harvesting may be a source of conflict as enclosures are not generally
welcomed. Hay production requires further research to establish its levels of
sustainability without being a cause of conflict in the study area.
Urban centres offer a good market for farm produce, but due to poor access roads are
inaccessible to most farmers. The urban centres are also points of high demand for
charcoal.
2.2.6 Population
The study area constitutes the districts of Dila, Gebiley, Faraweyne and Allaybaday and
parts of the districts of Hargeisa, Borama, Baki and Lughaya. The size of this study area
is effectively a little more than one third of the total area of Awdal and Waqooyi Galbeed
Regions. According to Somalia UNDP 2005 (Table 1), the estimated population for
Hargeisa by mid-2005 was 560.028, making it the second largest town in Somalia.
Borama had a population of 215.616 and Gebiley 79.564. These three are the main
towns in the study area.
Table 1: Regions, districts, and their populations (Somalia UNDP 2005, draft version)
District Estimated population

( * Regional
Region
Zone capital ) 2005 (Mid-year)
Non-
Total Urban urban
North-
west 1,828,739 819,989 1,008,750
Awdal 305,455 110,942 194,513
Borama * 215,616 82,921 132,695
Baki 25,500 8,577 16,923
Lughaye 36,104 14,010 22,094
Zeylac 28,235 5,434 22,801
Woqooyi
Galbeed 700,345 490,432 209,913
Hargeisa * 560,028 422,515 137,513
Berbera 60,753 42,070 18,683
Gebiley 79,564 25,847 53,717
11
Study Areas
Figure 4: Carta geobotanica dellAfrica Orientale by Rodolfo E.G. Pichi-Sermolli (Map of the
vegetation of East Africa)
12
Study Areas
2.3 Southern Area of Interest (SAOI)
2.3.1 Climate
The climate of the river basin areas of southern Somalia is tropical/arid to dry/sub humid
and is influenced by the north-easterly and south-easterly air flows of the Intertropical
Convergence Zone (ITCZ). The north-easterly and south-easterly air masses meet in the
Intertropical Front (ITF) and consequently raise air upwards, producing rain. The annual
movement of the ITCZ from north to south across Africa and back, gives rise to four
different seasons in Somalia, comprising two rainy seasons alternating with two marked
dry seasons as follows:
o Gu: April to June; the main rainy season for the entire country
o Xagaa: July to September; littoral showers, but dry and cool in the hinterland
o Deyr: October to December; the second rainy season for the entire country
o Jilaal: January to March; longer dry season for the entire country
The rainfall in the study area is erratic, showing a bimodal pattern except in the southern
riverine areas close to the coast, where some showers may occur even during the Xagaa
(see Figure 5). Total rainfall varies considerably over the study area, with the Gu
delivering about 60% of the total mean. Total mean annual rainfall ranges from 200400
mm in areas bordering Ethiopia in Hiiraan, Gedo and Bakool regions, and 400500 mm in
the central Bay, and northern part of Middle and Lower Shabelle Regions. High rainfall
areas receiving more than 600 mm occur in the Middle Juba region around Jilib in the
southern riverine areas. Rainfall is usually in the form of intense, short rainstorms, which
are often temporally and spatially erratic.
The study area has a high inter-annual rainfall variation and is thus subject to recurrent
droughts every 3-4 years, and more severe droughts every 7-9 years.
Air temperatures are generally influenced by altitude and by the strength of the seasonal
winds. In Xagaa (the first dry season) the days are often cool and cloudy all over the
region, with short showers in the areas closest to the coast. In Jilaal (the second dry
season) the days are hot to very hot, and dry. The hottest period coincides with the
months of March and April.
13
Study Areas
Belet Weyne Jowhar

100 120
80 100
Rainfall(mm)
Rainfall(mm)
80
60
60
40
40
20 20
0 0
1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12
Afgoi 100
Genale
140
120 80
Rainfall(mm)
100
Rainfall(mm
60
80
60 40
40 20
20
0 0
1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12
Balad MOGADISHU
120 100
100 80
Rainfall(mm)
Rainfall(mm)
80
60
60
40
40
20 20
0 0
1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12
Luuq Bardera
100 140
80 120
Rainfall(mm)
Rainfall(mm)
100
60 80
40 60
40
20 20
0
0
1 2 3 4 5 6 7 8 9 10 11 12
1 2 3 4 5 6 7 8 9 10 11 12
Figure 5: Mean monthly rainfall patterns in the study area (1963-2001)
Temperatures vary in the different seasons, with mean annual temperatures ranging
between 23-30C, a maximum temperature of 41C in March (Baardheere) and minimum
temperature of 24C in July. In areas near the major rivers the relative humidity is high,
ranging between 70-80%, but further inland away from the rivers the air is much drier.
Relative humidity is higher in the coastal areas, where it is normally greater than 87%.
The high relative humidity is often compounded by higher temperatures.
The major winds are in response to the north and south seasonal movements of the
Intertropical Convergence Zone, particularly the Intertropical front. In the study area the
winds blow persistently from the northeast during Jilaal (December to February) when
the weather is hot or very hot, and from the southwest during Xagaa, (June to August)
when the weather is cool and cloudy.
The weather is hot and calm between the monsoons (part or whole of April and part or
whole of September). In the Jilaal periods the prevailing winds are strong and blow in
heavy dust storms from the Arabian Peninsula. Weak winds generally occur during the
inter-monsoonal periods of April-May and October-November. Average wind speed varies
from 26 m per second.
14
Study Areas
Evapotranspiration is consistently high throughout the study area, the highest potential
evapotranspiration occurring in the northern areas of Gedo, Bakool and Hiraan regions
with more than 2 000 mm/yr, and between 2 000 mm/yr and 1 500 mm/yr. over the rest
of the area. Annual rainfall (P) is far below the potential evapotranspiration (PET) and
there is a large moisture deficit over most of the year.
The area experiences a tropical semi-arid to arid climate (Figure 6), with seasonal rainfall
patterns influenced by the monsoon winds. Rains occur between April-June and
September-November.
Three broad climatic zones may be recognised, characterised by differences in rainfall

patterns:
- The coastal zone - with significant amounts of rain occurring from July to August
(Hagi rains) that lengthen the Gu season.
- The semi-arid zone with two strongly-defined rainy seasons and an additional
light rainy season that may occur during July-August.
- The arid zone with a lower annual rainfall amount and a dry period in July-
August. The monsoon winds are the important factor affecting the climate and the
timing of the rainy periods. The south-west monsoon winds prevail during June,
July and August. The north-east monsoon winds prevail during December, January
and February.
Rainfall tends to be variable and unreliable. Failures of the rains are not uncommon,
particularly in arid areas. Rainfall can vary considerably from the average. This is an
important factor to consider in assessing the value of seasonal rains in term of
agriculture.
2.3.2 Geology
The southern AOI is characterised by outcropping of the metamorphic basement
complex, made up of migmatites and granites. Sedimentary rocks such as limestones,
sandstones, and gypsiferous limestones and sandstone are present, as is an extensive,
wide coastal system of sand dunes. Basaltic flows are present in the northwest part of
the AOI. From a tectonic point of view, this AOI is characterised by a fault system parallel
to the coast in the alluvial areas, and by a system of NW-SE oriented faults in the
metamorphic basement complex.
Some late Tertiary fluvio-lagunal deposits occur in the Lower Juba plain and part of the
Southern Shabelle, consisting of clay, sandy clay, sand, silt and gravel.
Recent fluvial deposits are common along the two major rivers, the Juba and the
Shabelle, consisting of sand, gravel, clay and sandy clay. Other Recent Alluvial Deposits
occur in small valleys in Gedo and Bakool regions and in the Buur area, and consist of
gravelly sand or red sandy loam materials. A wide dune system occurs along the coast.
2.3.3 Landform/Soils
According to the bibliography, the southern AOI is characterised by the following land
features:
1. the two main river valleys (Jubba and Shabelle) that incise the generally
smoothed and undulating morphology of the area;
2. hilly topography in the middle of the AOI, cut by wadis, and gently
undulating wide plains toward the coast;
15
Study Areas
3. coastal dune complex, known as the Merka red dunes, following the coast
and extending beyond the Kenyan border, which separate the narrow
coastal belt from the Uebi Shebeli alluvial plain (Carbone & Accordi, 2000).
Figure 6: Mean annual rainfall distribution map and climate of study area
The hydrography of the southern AOI is dominated by the presence of the distal portions
of the two main perennial rivers of the Horn of Africa flowing from the highlands of
16
Study Areas
Ethiopia towards the Indian Ocean: the Jubba (700 km within Somalia out of its 2 000
km total length) and the Shabelle (1 560 km within Somalia out of its almost 1 800 km
total length). The Jubba flows into the Indian Ocean almost at Kismaayo, while the
Shabelle impounds itself a few kilometres before reaching the lower tracts of the Jubba.
Because of the predominance of alluvium, many soils consist of layers of deposited

materials which, because of the semi-arid climate, have been little-affected by normal
soil-forming processes. Despite their variability, most soils share the characteristics of
heavy texture and low permeability, with a tendency to poor drainage.
2.3.4 Land Cover

On the basis of the vegetation types described by Pichi-Sermolli (1957) the southern AOI
is characterised by:
- Grass, perennial and sub-shrubs steppe (Steppa graminosa, perennierbosa e

suffruticosa)
- Subdesert scrub (Fruticeto subdesertico)
- Xerophyllous open woodland (arid zones) (Boscaglia xerofila rada)
- Xerophyllous open woodland (semi-arid zones) (Boscaglia xerofila)
- Savanna
- Coastal formations (Formazioni costiere)
- Riparian formations (Formazioni riparie)
- Palustrine formations (Formazioni palustri)
Grass, perennial and sub-shrubs steppe (or sub-desert shrub and grass) - Includes all
those cenosis constituted by grass, graminoids and suffrutexes. Perennial plants are rare.
Cover values and floristic composition can change according to edaphic conditions. They
are common on gypsum soils. A typical species is Chrysopogon auchieri var.
quinqueplumis.
Subdesert scrub - Formed by low trees (less than 3 m high), shrubs, suffrutexes and
spaced-out high succulent grasses. Tubers and bulbs occur among the herbs. Annual
grasses grow in the rainy season. Perennial graminoids are absent.
Xerophyllous open woodland (arid zones) - An open vegetation type, formed by woody
plants. In this formation, shrubs between 3-5 m in height and perennial graminoids are
dominant. Scattered trees are typical. Shrubs are scattered or sometimes grouped, but
never impenetrable. Suffrutexes and perennial grasses have a discontinuous cover
occupied by annuals (therophytes) after rainfall. The most common Acacias are:
A.bussei, A.etbaica, A.nilotica, A.tortilis and A.senegal.
Xerophyllous open woodland (semi-arid zones) - Of greater density and higher species
richness of both high shrubs and perennial grass when compared to the previous
vegetation formation.
Savanna - Composed of annual grasses (usually xerophytes). Cyperaceae and

Graminaceae are the most common families, forming a continuous layer 80 cm. in height
Trees or shrubs are usually present, but the herbaceous layer is dominant. They can be
isolated or in small, sometimes impenetrable thickets.
Coastal formations - These are vegetated by various dominant life forms, according to
edaphic conditions but are affected by the direct influence of the sea. Essentially they
occupy the coral beaches, sandy shores, estuaries, deltas, coastal dunes (stabilised or
17
Study Areas
mobile) and salty plains. Typical examples are Suaeda fruticosa, Adenium sp. on stony
areas and Salsola foetida on deep, dry soils.
Riparian formations - These formations develop along temporary or permanent streams

and rivers. Due to higher soil moisture content they stand out from surrounding, dryer
vegetation, from which they are also distinguished by their floristic composition.
Palustrine formations - These comprise all plant communities within swamps, i.e. areas
which are wet for most of the year. Vegetation typically occurs in shallower waters.
Central, deeper parts of swamps are characterised by rooted or free-floating aquatic
plants. The most common are palustrine graminoid plants in association with
Cyperaceae.
2.3.5 Land Use

Land use in the study area consists mainly of grazing and wood collection for fuel and
building. Rangelands in the Juba and Shabelle catchments support livestock such as goat,
sheep, cattle and camels. Livestock ownership is private, but grazing lands are
communal, making it very difficult to regulate range use. Rangelands are utilised by
herders using transhumance strategies (Shaie, 1977). Land cover associated with this
land use includes forest, bushlands and grasslands (GTZ, 1990).
Farmers in these two river valleys are sedentary, practicing animal husbandry in
conjunction with crop production. They tend to keep lactating cattle, a few sheep and
goats near their homes, while non-lactating animals are herded further away, in the
manner of herding nomadic stock. However, rainfed and irrigation farmers keep relatively
small numbers of livestock, mainly cattle and small ruminants.
Animal feed is primarily from natural vegetation and crop residues, while dry season
watering of animals is from rivers. Crop residues provide forage for non-browsers such as
cattle and sheep. Several wars provide water in the wet season and also serve as
alternative water sources to rivers. Groundwater is also an important source of water for
livestock, other sources including hand-dug wells, swamps, creeks and boreholes.
Other land uses include rainfed agriculture, which includes agriculture that is entirely
dependent upon rainfall. Crops under this category of land use include sorghum, millet,
maize, groundnuts, cowpeas, mung beans, cassava and other minor crops, and are
grown twice a year in the Gu and Deyr seasons.
Small-scale irrigated fields are also found along the Shabelle and Juba river valleys,
growing maize, sesame, fruit trees and vegetables while large-scale plantations include
sugar cane, bananas, guava, lemon, mango and papaya.
Flood recession cultivation in desheks (natural depressions) on the Juba River floodplain
is common, crops including sesame, maize and vegetables. Major crops in the desheks
are maize, sesame, tobacco, beans, peas and vegetables, watermelon and (rarely)
groundnuts. Cropping is either single or mixed.
18
Materials and Methods
3 3. MATERIALS AND METHODS
The following is a list of materials used and tasks performed during production of the
land cover map.
3.1 Materials
3.1.1 Bibliographic data
A list of documents consulted concerning aspects of vegetation in Somalia and in the two
AOI in particular, is provided in Chapter 7.
3.1.2 Ancillary data
- 1:100 000 topographic maps from the 1970s (revised in the early 1990s)
produced by the Defence Mapping Agency USA (Annex 5 and Annex 6).
- 1:200 000 land cover map of Somalia produced by FAO Africover through visual
interpretation of digitally enhanced LANDSAT TM images (Bands 4,3,2) acquired
mainly between 1995-1998. The land cover classes were developed using the
FAO/UNEP international standard LCCS classification system.
- 1:100 000 land cover map of the Juba and Shabelle riverine areas by SWALIM
Phase I (unverified), produced through visual interpretation of digitally enhanced
LANDSAT TM images (Bands 4,3,2) acquired mainly between 2000-2002. The land
cover classes were developed using the FAO/UNEP international standard LCCS
classification system.
3.1.3 RS data
- Landsat 7 Enhanced Thematic Mapper 2000-2001-2002 (Annex 6 and Annex 7)

- Landsat 5 Thematic Mapper 1985-1986
- MSS 1972-1973 e 1976
- IKONOS 2001 2003
- Aster VNIR and SVNIR 2002 2005
- Shuttle Radar Topography Mission Digital Elevation Model (SRTM-DEM)
Details on sensor type, image acquisition data, pixel resolution and wavelengths are
provided in Interim Report GIS Li-01.
3.1.4 Software
Digital interpretation was performed using ESRI GIS package and Geovis (Terranuova).
The module Imagework (PCI) was used to support the photointerpretation routine,
allowing some enhancement of satellite images.
The FAO software Land Cover Classification System (LCCS) was selected for creating the
Land Cover legend.
19
3.2 Methodology
Methodology adopted for the production of the Land Cover map can be subdivided into
five main phases:
3.2.1 Preparatory phase
Preliminary overview of the AOI

SWALIMs photointerpretation and GIS/RS coordinators initially reviewed available
data on the vegetation and agriculture of Somalia and the two areas of interest in
particular, in order to evaluate the complexity of the landscape and the potential
usefulness of the data.
Available data (included geological maps and data on geomorphology) were

collected in Italy from Istituto Agronomico DOltremare (IAO), the libraries of the
faculties of Agriculture and Botany of the University of Florence, and the Geology
Departments of the Universities of Florence and Pisa. Additional documents and
reports were collected in Nairobi, from the Internet, from the library of IUCN, and
in Somaliland by the Land Coordinator.
Remote sensing data acquisition

Several sets of satellite images were acquired during SWALIM Phase I:
- Landsat 7 Enhanced Thematic Mapper 2000-2001-2002

- Landsat 5 Thematic Mapper 1985-1986
- MSS 1972-1973 e 1976
- IKONOS frames (5x5 and 7x7km) 2001 2003
Additional remote sensing data were collected during SWALIM Phase II:
- Aster VNIR and SVNIR 2002 2005

- Shuttle Radar Topography Mission Digital Elevation Model (SRTMDEM)
Image processing
Image processing preceded pre-processing. Pre-processing is the basic cleaning,
correction and registering of raw primary data. Image processing operations
include all operations performed in order to assist visual interpretation and
facilitating extraction of information from original images.
The GIS unit digitally processed the images and was also responsible for some
specific pre-processing, due to available RS data being in different formats and
projections.
Aster datasets were delivered in EOS format and required fine registering before
being integrated into the GIS. The SRTM-DEM data was imported into ArcInfo
grid, re-projected from decimal degrees (DD) to Universal Transverse Mercator
(UTM) and pre-processed for basins and stream delineation. All datasets were
available in single bands as shown in Annex 08.
20
The following processes were performed:
Colour composite
Enhancement
Calculation
Colour composite - Single bands are employed for generating Colour Composites (CCs). A
composite is a coloured image where, for each of the three primary (additive) colours
(red, green and blue), a spectral band is used. The true colour composite uses the bands
1 red, 2 green and 3 blue, for the primary colours red, green and blue.
There are particular combinations of bands (thermal band included) that are significant
for the identification of the spectral signature of land cover types and were chosen for
generating False Colour Composites (FCCs).
The list of the CCs images produced for each sensor type by the GIS unit is shown in
Table 2.
Table 2: CCs images produced by SWALIM GIS unit
Landsat Aster Ikonos

Composite 432 Composite 321 Composite 432
Composite 743 Composite 654 Composite 321 (true colours)
Composite 456 NDVI index NDVI index
Composite 765
Panchromatic
Pan (increasing details) and/or

sharpen (enhancing details) added
images eventually required among the
previous composites.
NDVI index
Thermal Capacity index
T = thermal; 2, 3, 4, 5 and 7 are the other Landsat7 ETM bands implied
Enhancement - Single band images, used in black and white or combined in composites,
were enhanced by the following processing:
Adding panchromatic applied to Landsat (ETM bands only) and IKONOS - the
information of a colour composite image is combined with the more detailed
panchromatic. Panchromatic sensors have a greater geometric resolution than sensors
devoted to a restricted wavelength range. A higher resolution shadow is added to the
added colours in order to increase object shape details and to help in evaluating surface
roughness (e.g. homogeneous tree cover).
Stretching (applied to any single band) this process expands (proportionally
reallocates) tonal distribution from lower and higher values present in the original image,
to the full available grey scale display (usually subdivided into 255 grey tones). Contrast
enhancement is characterised by improved appearance of different bodies with similar
tones, emphasising patterns and, to a lesser extent, roughness.
21
Filtering (sharpen) (applied to any single band) a new value is applied to each pixel
of an image, based on elementary calculations employing the value of the pixel itself plus
surrounding pixels. This process is called convolution. These pixels are selected and their
values are weighted through the employment of a matrix or box centred on the target
pixel. These matrices are usually 3x3 or 5x5 pixels in size and can provide different
results. In our case they sharpened the edges of objects in images, improving the
capability of distinguishing and delineating them. In particular, it enhances edges and
boundaries between different land cover types, emphasising pattern and roughness.
Calculation sets of single bands available for an image frame can be employed as
variables of mathematical equations. These equations are derived from conceptual
relationships between physical phenomena and the light reflectance/emission response.
From mathematical combination of bands it is possible to create indices. The ones used
for the interpretation of the Land Cover were the Normalized Differential Vegetation
Index (NDVI) and the Thermal Capacity. Images obtained from equations do not
represent reflectance of the bodies but some of their properties
NDVI - (Normalized Differential Vegetation Index) is the expression of the intensity of

photosynthetic activity. It is strictly correlated with photosynthesis, and through
implication can also supply information about vegetative activity and biomass. These two
factors can help to identify vegetation types, or can be associated with vegetation types
in order to formulate hypotheses on their vegetative condition.
Thermal Capacity Index - this is an expression of the moisture level or mass capacity
level, and is correlated with the mass of water present in soils. It can be used in support
of the composites to identify depressions with an aquifer close to the surface, and wet
areas. Response is strongly affected by season.
Setting up of the preliminary legend

Classification of the Land Cover was performed using the FAO Land Cover Classification
System software (LCCS) developed by Antonio Di Gregorio and Louisa J.M. Jansen (see
LCCS classification concepts and user manual FAO, 2005). The advantages of this
system are that:
- It allows class description at any convenient categorical level of detail according to

available data. It is therefore possible to employ the same classification system
for AOI with different map scale representation and working conditions.
- It supplies the advantage of a standardised apriori classification system, which is
consistent in the choice of legend codes.
- It keeps the flexibility of a posteriori classification system (because it is not
composed by a predefined classes set, but by a predefined classifier set and
syntax to build well-fitting classes).
- It is adaptable to any environment, for any purpose, and is provided with a tool to
import external legends. This feature allows accommodation and correlation with
any other Land Cover map.
- All the features characterising legend codes are organised as a text string. This
structure allows querying, selection or grouping of codes according to desired
characteristic at different taxonomic levels.
- It facilitates the extensive use of outputs by a wide variety of end-users through
enhanced capabilities in rearranging land cover classes.
22
Land cover classes are defined through the use of a combination of sets of independent
diagnostic criteria - the so-called classifiers - that are hierarchically arranged to assure a
high degree of geographical accuracy. Due to the heterogeneity of land cover, the same
set of classifiers cannot be used to define all land cover types. The hierarchical structure
of the classifiers may differ from one land cover type to another. Therefore, the
classification has two main phases:
- an initial Dichotomous Phase, where eight major land cover types are
distinguished; and
- a subsequent Modular-Hierarchical Phase where the set of classifiers and their
hierarchical arrangement are tailored to the major land cover types.
All Primarily Vegetated land cover classes are derived from a consistent physiognomic-
structural conceptual approach that combines the classifiers Life Form, Cover and Height
(in (Semi) Natural Vegetation) and Life Form (in Cultivated Areas) with Spatial
Distribution. The Primarily Non-Vegetated classes have a similar approach, using
classifiers which deal with surface aspects, distribution/density and height/depth.
The preparation of the legend was preceded by the inventory and study of all available
data collected during the preparatory phase, such as reports, documents, maps and any
information on the distribution of vegetation and types of agriculture in the areas of
interest. Information contained in these datasets was translated into classes through the
selection of appropriated classifiers during the creation of the legend using LCCS.
The original list of classes was therefore improved through revision and/or creation of
new classes on the basis of the Land Cover types identified on the interpreted images.
The final version of the legend adopted for the Land Cover map is presented in Annex 9.
Figure 7 shows a summary of the classes, grouped by main Land Cover type.
3.2.2 Base map production
Photointerpretation activity
A general definition of interpretation is the identification of the objects seen in the image
and communication of this information to others (Lillesand & Kiefer, 1979).
Homogeneous surfaces are identified and enclosed within boundaries. An area
delimitated in this way is treated as an object and is called a polygon. Polygons are
classified with the assignment of a label. The code(s) composing the label are selected
from the adopted legend (Fig 02).
The two basic principles to follow and to strictly monitor during the production of the
map, are internal and external consistency. The same interpretation keys identified on
the image have to be adopted for different parts of the frame in order to maintain the
consistency inside the interpreted frame (internal consistency). All photo-interpreters in
the team meet and verify interpretations they have attributed to elements identified in
their respective images, in order to avoid the risk of using different labels for labelling
the same thing, or vice versa. Consistency is important, as with a consistent classification
land is subdivided in stratified homogeneous areas, allowing optimisation of field data
collection and easier upgrading and correcting. This facilitates revision of the preliminary
map.
SWALIMs land theme photointerpreters edited the Land Cover geo-feature class, through
visual interpretation of available RS images using the ESRI (ArcGis) platform and Geovis
software. Polygons represented on the map are in vector format and can be directly
managed as a spatial feature dataset and for performing analyses (see Chapter 3.2.5 on
Data analysis and spatial modelling).
23
Figure 7: Summary of the legend by main Land Cover type
24
Legend codes were assigned following LCSS rules. For a detailed explanation, see LCCS
classification concepts and user manual FAO edition, Rome, 2000.
HR47
HL57
HR57/4HCJFF
5U
Figure 8: Photointerpretation
25
Satellite images such as Landsat (ETM sensor: 30 m sized pixels) or Aster (VNIR sensor:
15 m sized pixels) were interpreted on the basis of:
- Tone and colour brightness;

- Spectral signature;
- Texture, which is the frequency of changes and arrangement of tones;
- Contrast, or the difference in tone between a body and the background;
- Pattern, the regular repetition of colours, tones and features such as lines or
shadows;
- Size, shape of the objects and their topological relationship/s.
As already explained in the previous paragraph on Image processing, FCCs images were
created by selecting and combining bands considered significant for the identification of
different land cover types. Combinations used for the interpretation were:
Composite 432 mainly for vegetation, Composite T43 mainly for wet areas,
soils, and settlements irrigated agriculture, and vegetation
Composite 457 mainly for vegetation and Composite T75 mainly for wet areas, soil
soil conditions and vegetation conditions, and irrigated
agriculture
Typical composite Landsat432 and corresponding Aster 321 were the more frequently
used.
26
The following paragraphs explain the conceptual framework used to extract information
using images with differing spatial, radiometric and spectral resolutions, or in temporal
sequences.
The main evaluations and comparisons performed on available datasets were:
Superimposing Ikonos samples on Landsat or Aster - Using IKONOS images for land
cover identification can be considered as an activity that partially replaces some aspects
of the field survey. It is possible to accomplish a level of sampling for certain basic
parameters, allowed for by the very high resolution of the sensor.
IKONOS images allow direct identification and recognition of small objects such as plants,
and structure of vegetation formations. Furthermore, some parameters can be more
accurately collected if compared with field samples. An example is vegetation canopy
coverage, which is one of the most important parameters used in many processes of
analysis. While field samples mostly supply information about vertical structures,
IKONOS can supply information about horizontal structures, improving correlations with
vegetation and environmental conditions.
The vertical perspective can be compared with results from images such as Landsat and
Aster, which have a lower resolution.
Locations of IKONOS imagery were accurately selected for the southern AOI during the
preliminary photo-interpretation. Parameters considered were:
- combination/s of trees and shrubs, coverage and medium diameter;

- presence and coverage of eventual emergent;
- total coverage (max 100%);
- herbaceous presence in particular seasonal conditions;
- crop patterns, life forms and water supply;
- geomorphological information and related processes (drainage, erosion, etc.);
- vegetation horizontal patterns (plant distributions: regular, irregular or grouped).
Superimposing different Landsat derived images - This method uses a type of

stratification or matrix. The basic principle is that every image can be interpreted in
different ways (formulating hypotheses), but that sometimes the only way is to combine
two or more different images.
Specific band combinations:
- composites Landsat 432 and 457 and Aster (VNIR) 321, used in combination with
NDVI for detecting vegetation;
- composite Landsat 432, T43, 457, Aster (VNIR) 321, Landsat panchromatic, NDVI
and Thermal Capacity for detecting agriculture;
- composite T43, Thermal Capacity and NDVI indices for assessing soil moisture.
When soils are moist the composite T43 is dark red or violet; the Thermal
Capacity is bright while the NDVI is dark, showing that moisture is not held by
vegetation.
Temporal series crossing among corresponding Landsat images - Different temporal

series of Landsat images from the years 1988, 1990 and 2000 were made available by
the FAO Africover project from their archive, which were used to identify agricultural
schemes and wet areas. Images acquired at different times and, when available, of
different seasons (for the same area) were compared. The main feature noted was the
27
dynamism in terms of cultivated surfaces over the years, which assisted in understanding
actual distributions.
Draping images to SRTM-DEM - This allowed a better understanding of features relating

to terrain shapes, such as lithology or shadows, that influence and alter vegetation
spectral signatures.
An example is the approach followed to determine the Land Cover in the Lower Juba
area. A similar approach was followed for Medium Shabelle.
The basic features to detect and extract were:
- Agriculture (presence or absence of any kind of irrigation system)

- Natural vegetation (wet or terrestrial)
Two Landsat images were available for the area:
1. P165r60 collected on 2nd April, 2000

2. P165r60 collected on 17th September, 2000
The two images are separated by seven months and were collected at the end (image 1)
and the beginning (image 2) of the Jilaal, the long dry season, respectively.
The parameters used for classification were:
Heat response - The first analysis was based on the heat response. The thermal infrared
has a small emission while water is stored within soils. Water in soils can be associated
with wet natural vegetation and irrigated agriculture.
Image 2 was collected during the last part of Deyr, the short rainy season and clearly
shows wet areas, mostly distributed in the central-upper portion of the river. The shape
of these areas is consistent with depression formations, and their position is close to
water bodies that are apparently in spate. Image 1, collected between the end of Jilaal
and the beginning of Gu, the main rainy season, shows water bodies in flood (with some
associated wet areas) in the south-west areas, which matches with stream flows from
drainage systems emanating from the Kenyan border areas.
The two images seem to show an irregular rain distribution when comparing the two
lower Juba areas, but are able to indicate flood water dynamics.
Image patterns - Image patterns are repeated textures, colours, shapes and objects on
the earths surface. Natural vegetation is characterised by gradual changes in colour and
by irregular clustering of shapes, with variable density and distribution.
In this approach, geometric patterns are considered as main indicators of human

activities, including agricultural schemes. This kind of pattern is characterised by abrupt
colour changes, straight boundaries, corners and geometric shapes.
Pattern typology - Three main geometric patterns may be distinguished in the area:
- Burned areas, not strictly related with agriculture.

- Patches and dots, principally related with household agriculture characterised by
different types of water supply: rainfed, post flooding (desheks), bank breaking
and uncontrolled riverine flooding.
28
- Big irrigation schemes usually occur in big farming projects. The main channels
are straight, long (up to several kilometres) and clearly visible. The pattern is
squared and very regular. It is not possible to directly detect amounts of silt and
consequently the effective functionality of the channels, but it is possible to make
hypotheses on soil moisture content.
In the case of dry soils and big irrigation schemes it may be deduced that channels are
present, but silted and not in use. Presence or absence of rain-fed agriculture has to then
be verified. Ex-irrigated schemes, when identifiable, are mapped for their importance in
defining Land Utilization Types.
A flowchart of considerations made during the interpretation is presented in Figure 03.
Thermal
LOWER HIGHER
response
Moisture in Dry soils

the soils
Big EX-big
Surface irrigation irrigation Surface
Absent pattern Present Present pattern Absent
schemes farming
farmlands schemes
Permanent Terrestrial
Seasonal Pattern Channels Channels Pattern natural
periodic Burned Burned vegetation
Type patterns patterns Type
natural
vegetation
Post
flooding Patches Patches Rainfed
and small agriculture
or dots or dots
irrigation
systems
mixed with
rain fed
agriculture
Figure 9: Flowchart
29
3.2.3 Field data collection
Field work
Planning standards and procedures for a field survey are a critical step in implementing
land cover analysis, as the reliability of the entire work is based on the analysis of data
collected (ground truthing).
A tailored methodology for field data collection was defined, taking into account the skill
levels of selected Somali surveyors, the language of the classification system adopted
and the field survey methodology developed by FAO Africover.
A simplified version of the legend structure was developed together with forms for data
collection. The methodology, structure and content of the forms are explained in detail in
Report L-01 Field Survey Manual.
The field survey was performed only in the north-western AOI and the southern riverine
areas. For the northern AOI the field survey was performed in May-June 2006 while for
the southern riverine areas it was performed in March 2007. The delay in the southern
AOI field survey was due to the problem of insecurity that was attributed to the war in
South Somalia.
The field survey boxes, in both the areas, (indicating the areas to be investigated) were
selected after evaluation of their representativeness, distribution, total number,
accessibility, possibility of creating clusters and available time. Priority was given to:
- homogeneous areas (labelled by a code) that cover large areas;

- the most frequent codes (labelling more polygons);
- very complex areas not directly understood through image analysis;
- locations with better accessibility;
- clustering of field sample sites in order to facilitate planning and route selection;
- areas of major economic and/or environmental importance.
The following tools were provided for the field survey:
- preliminary 1:130 000 maps printed on A0 paper in sets of two sheets, each
showing land cover polygons, boxes, and roads on Landsat images;
- a hardcopy list of the boxes with information on coordinates, sizes, preliminary
codes and an identifier derived from the field samples boxes geo-dataset;
- an adequate number of field forms to be used in data collection;
- laminated tables designed to assist in estimation of percentage of cover and
evaluation of dominant natural vegetation layers;
- measuring tools (compass, tape measure, etc.) and digital cameras;
- hardcopies of the Field Survey Manual;
- manuals for plant identification.
Field data analysis
The data collected in each box during the field survey are composed by:
- a set of forms describing land cover types identified;

- four digital pictures (one from each cardinal point);
30
- additional information and descriptions on vegetation physiognomy, floristic

composition, etc.
In order to make data usable for analysis and revision, the following tasks were
performed:
Data entry - an appropriate Microsoft Access template was designed to convert and store
data in digital format. Every field form has its corresponding tables linked together in a
relational database. Data entry is handled through appropriate masks developed with
essential validation rules, to guide operators and reduce mistakes.
Built a point geo-feature class for sites location (x and y coordinates) - GPS positions of
sampled sites were converted to a geo-feature class (shapefile) in order to register
locations of sites together with images and land cover geo-feature classes, allowing
comparative viewing, querying and analysis through GIS.
Validation of sites forms set - descriptions and parameters collected were checked for
completeness and consistency. Forms with missing/divergent information, coordinates,
codes or photographs were not considered during analysis.
Attribution of an LCCS code to each site an LCCS code (both single class and mixed
unit) was attributed to each site form and photo set on the basis of the evaluation of
collected data. Some new LCCS classes were created to better fit part of the site sample.
A subsequent selection of codes provided the final legend.
3.2.4 Legend updating and map revision
Parameters and descriptions collected in the field in both the northern and southern AOI
were compared with the preliminary map, and results used for revising the legend and
interpretation.
Legend classes were consequently confirmed, revised and/or updated with new ones. The
updating of natural vegetation classes was based on descriptions of life forms (coverage
and heights of tree, shrub and herbaceous layers), evaluation of photographs and
eventual additional information recorded.
Agricultural classes were revised after consideration of field form descriptions of crop life
form and species and evaluation of photographs, and field sizes and densities were
derived from satellite images.
Extensive comparative work was conducted on information collected on the ground in

order to identify relationships among different zones of the AOI. Similar areas/codes
identified on maps were then evaluated in order to consistently extend analyses to un-
surveyed areas of the map.
A list of all classes obtained from the preliminary legend, as well as revised and new
classes resulting from field data analysis, was checked in order to avoid overlap of
classes (some aggregations of cover thresholds were required), and the final legend was
created in LCCS.
Before starting final revision of the interpretation, a conversion table was created to
automatically transform, where possible, old codes into new ones. In some cases it was
possible to automatically dissolve bordering polygons. The majority of classes were
however replaced or confirmed through visual revision of the interpretation.
Codes assigned to similar areas were updated according to results of field data collected
in the north.
31
This homogenisation process was mainly concerned with natural vegetation and
rainfed/irrigated agriculture codes. The final bordering polygons generated were
automatically dissolved where possible and manually checked in the remaining cases.
The final Land Cover legend adopted for both AOI is shown in Annex 09.
3.2.5 Data analysis and spatial modelling
Spatial analyses are the most important functions of a GIS. Relationships between the
database and the geographic components (topological connections) enable GIS to create
models and representations of the real world. All spatial modelling was performed in line
with the FAO SDRN method and may be subdivided into two main types:
Aggregations
Aggregations involve spatial features and tabular data, and are mappable. They are an
abstraction mechanism, used for clustering objects at a superior level. Aggregations are
used to progressively obtain more complex and meaningful new objects from elementary
objects. Starting from a lower taxonomic level, higher levels are created. This process is
extremely important as it allows display, understanding and practical usage of geo-
datasets. Different aggregations can be designed and performed for different users,
according to their technical competences and needs.
The aggregation model is characterised by two components:
- identification of subsets of objects, which may be grouped into a more

complex object;
- rules applied during selection of each subset.
As already explained, LCCS is an extremely sophisticated classification system. It can

store voluminous information concerning vegetation, agriculture physiognomy,
physiology and other associated environmental features.
LCCS was designed considering the linkages between land cover and the actions of
people in their environment. Production of aggregation according to the GLCN method
leads to a basic subdivision of the land in raw land-use types. These aggregations supply
a base map for land-use classes that will be developed in further studies, providing the
spatial locations for them.
In order to obtain this result from LCCS codes, a set of rules was applied to obtain
several taxonomic ranks as follows:
1) The first level of aggregation (Aggregated Classes) was created based on:
- the most important LCCS class present in the code, where agriculture is
always dominant: a code with 2 classes, where the main class is that of
natural vegetation and the secondary one is that of isolated rainfed
agriculture (coverage = 10-20%) grouped as an (isolated) agricultural area.
- In the case of agriculture, they share the same irrigation system, coverage
rate (spatial distribution or macropattern), crop types, and field size. All these
features are closely related with major differences in management and
labour/mechanization.
32
2) The second level of aggregation (Main Aggregations) was created by grouping all
very similar classes.
- In the case of vegetation, all classes conceptually belonging to similar natural

formation types were grouped. As a general rule, classes sharing the same
dominant life form were grouped (for example, Shrub Open with Trees
Sparse or Shrub Open with Herbaceous Open). The main formation types
obtained were: Grassland (herbaceous can be of any height and cover
value, but shrubs and trees are absent), Savanna and other spaced woody
formations (general sparse shrubs and/or trees; herbaceous can be of any
height and cover value), Shrubland (general open shrubs) and Woodland
(general open trees). Dominance between layers, according to LCCS rules, is
maintained. For example, the class Open Trees with Closed to Open Shrubs
is included in the aggregation General Open Trees (woodland), due to the
influence of the larger life form despite similar coverage between the two
layers. Regarding coverage of dominant layers in natural vegetation, only the
broad LCCS ranges are considered (generally sparse 1-15%, generally open
16-65% and closed >65%).
- At this level, agricultural aggregations report irrigation system and spatial
distribution. The last feature can suggest the weight of agriculture versus
other activities, such as pastoralism.
3) The third and highest level of aggregations (Vocation) arranged Land Cover in the
main agricultural or natural subdivisions affecting land-use:
- Rainfed Agriculture
- Irrigated Agriculture
- Post-Flooding Agriculture
- Wooded Vegetation (suitable for charcoal, wood exploitation and grazing)
- Rangeland (suitable for grazing only)
- Other (including urban areas, settlements, water bodies and bare areas)
Agricultural aggregations can include activities such as charcoal, wood exploitation

and grazing. This is particularly true if agriculture is scattered or isolated.
All these rules are implicitly present in the Conversion Table (Figure 04 shows
cartographic aggregations generated according to the Conversion Table). The table is a
reference that defines a typical link from One (the aggregations) to Many (all the related
polygons in the PAT). This table also stores the area rating for each land cover code.
Areas subdivision (tabular) analysis
This is based on spatial feature attributes (tabular data only). In the present exercise the
aim was a statistical allocation of surfaces according to meaningful land cover typologies.
Transformations and handling of data required the employment of two functions:
1. Generation function information contained in each land cover code was

extracted and merged in a new categorical grouping regarding LCCS structure. In
this classification system, the most elementary classes are the so called single
classes composing the legend. The codes assigned to a land cover type can range
from 1 up to a combination of 3 single classes (mixed units) following the system
rules. Each class is referred to a particular vegetation or agricultural type or to
other kinds of land cover, such as urban, bare areas or water bodies. Each class in
33
a code has an assigned coverage rate (the sum of the classes composing a code
must be 100%). This percentage can be assigned according to the attributes
describing the class or the relative importance of the class in the mixed unit. For
example, in a mixed unit consisting of two natural vegetation classes, the main
class is assumed to cover 60% of the area, while the secondary class covers 40%.
Agricultural classes are determined based on the coverage rate classifiers. In the
case of rain fed agriculture up to three classes can be required, specifying
whether coverage is between 10-20% (isolated), 20-50% (clustered), or more
(continuous).
Such fusions are categorical only, since they are not mappable. In fact, single
classes were extracted from the codes (labels) and were merged according to any
interested feature but those features concerning coverage rate (spatial
distribution). For a particular kind of rainfed agriculture, each concerned class has
a different weight in surface attribution according to assigned percentage.
2. Calculation function tabular information is used as a variable in mathematical

equations. In GIS applications, homogenous spatial features are called
polygons. Each polygon is described by a linked record in a Feature Attribute
Table (FAT), which stores the land cover code (simple classes or mixed units), and
areas derived by projecting the geo-dataset in Albers Equal-Area projection (in
which all areas are proportional to the same areas on the earth). In this phase,
each class code of each record in the FAT is assigned an area. For example, if the
polygon area is 142 ha and comprises two natural vegetation classes, the main
class comprises 60% = 85 ha and the secondary one 40% = 57 ha. Calculated
areas are summarised on the basis of defined categorical groupings. One subset
of very similar natural vegetation classes can be summarised together, without
taking into account whether such classes are primary or secondary in the original
code.
In conclusion, calculated areas are not referred to polygons themselves, but to different
land cover typologies that they include. Results are arranged in categorical groupings
very similar to the cartographic aggregations. The main differences between the two are:
- Categorical groupings never need distinctions about cover percentage. They are
not mappable.
- Cartographic aggregations are basically based on the dominant class of the code.
Secondary classes are not considered, despite the fact that they can represent an
important part of the polygons.
34
ACTIVITY OF
MAIN_AGR AGGR_AGR
FOCUS
Irrigated - Continuous Large Fields of Herbaceous crops (single - Rice/Banana - or multiple herbaceous crops)
IRRIGATED AGRICULTURE Irrigated - Continuous Medium Fields of Herbaceous crops (single (Rice, Banana) or multiple herbaceous crops)
(crop fields >80% of the area) Irrigated - Continuous Small Fields of Herbaceous crops (single (Rice, Banana) or multiple herbaceous crops)
Irrigated - Continuous Small Fields of Tree crops (fruit trees) and Herbaceous crop (pulses and vegetables) (with
distinct patterns on the same field)
General Open Trees with Shrubs 1-65% OR Irrigated Continuous Small Fields of Tree crops (fruit trees) and Herbaceous
MIXED UNITS OF IRRIGATED AGRICULTURE (crop
crop (pulses and vegetables) (with distinct patterns on the same field)
fields 80-50% of the area)
IRRIGATED Irrigated - Mixed Continuous Medium Fields of Herbaceous crops (single (Rice, Banana) or multiple herbaceous crops)
Irrigated - Clustered Medium Fields of Herbaceous crops (single (Rice, Banana) or multiple herbaceous crops)
CLUSTERED IRRIGATED AGRICULTURE (crop fields General Open Trees with Shrubs 1-65% OR Irrigated Continuous Small Fields of Tree crops (fruit trees) and Herbaceous
50-20% of the area) crop (pulses and vegetables) (with distinct patterns on the same field)
Irrigated - Clustered Small Fields of Tree crops (fruit trees) and Herbaceous crop (pulses and vegetables) (with distinct
patterns on the same field)
Irrigated - Isolated Small Fields of Herbaceous crops (single (Rice, Banana) or multiple herbaceous crops)
ISOLATED IRRIGATED AGRICULTURE (crop fields
20-10% of the area) Irrigated - Isolated Small Fields of Tree crops (fruit trees) and Herbaceous crop (pulses and vegetables) (with distinct
patterns on the same field)
RAINFED AGRICULTURE Rainfed - Continuous Medium Fields of Herbaceous crops (single or multiple) - cereals (Sorghum, Maize)
(crop fields >80% of the area) Rainfed - Continuous Small Fields of Herbaceous crops (single or multiple) - cereals (Sorghum, Maize)
MIXED UNITS OF RAINFED AGRICULTURE (crop Rainfed - Mixed Continuous Medium Fields of Herbaceous crops (single or multiple) - cereals (Sorghum, Maize)
fields 80-50% of the area) Rainfed - Mixed Continuous Small Fields of Herbaceous crops (single or multiple) - cereals (Sorghum, Maize)
RAINFED
CLUSTERED RAINFED AGRICULTURE Rainfed - Clustered Medium Fields of Herbaceous crops (single or multiple) - cereals (Sorghum, Maize)
(crop fields 50-20% of the area) Rainfed - Clustered Small Fields of Herbaceous crops (single or multiple) - cereals (Sorghum, Maize)
ISOLATED RAINFED AGRICULTURE Rainfed - Isolated Medium Fields of Herbaceous crops (single or multiple) - cereals (Sorghum, Maize)
(crop fields 20-10% of the area) Rainfed - Isolated Small Fields of Herbaceous crops (single or multiple) - cereals (Sorghum, Maize)
POST FLOODING AGRICULTURE

Post Flooding - Continuous Small Fields of Multiple Herbaceous crops
(crop fields >80% of the area)
MIXED UNITS OF POST FLOODING AGRICULTURE Post Flooding - Mixed Continuous Medium Fields of Multiple Herbaceous crops
POST (crop fields 80-50% of the area) Post Flooding - Mixed Continuous Small Fields of Multiple Herbaceous crops
FLOODING
CLUSTERED POST FLOODING AGRICULTURE (crop
Post Flooding - Clustered Small Fields of Multiple Herbaceous crops
fields 50-20% of the area)
ISOLATED POST FLOODING AGRICULTURE (crop Post Flooding - Isolated Medium Fields of Multiple Herbaceous crops
fields 20-10% of the area) Post Flooding - Isolated Small Fields of Multiple Herbaceous crops
MIXED WATER SUPPLY AGRICULTURE

Rainfed - Continuous Medium Fields of Herbaceous crops (single or multiple) - cereals (Sorghum, Maize)
(crop fields >80% of the area)
MIXED WATER SUPPLY AGRICULTURE Rainfed - Continuous Small Fields of Multiple Herbaceous crops (Cereals) OR Post Flooding Continuous Small Fields of
(crop fields >50% of the area) Multiple Herbaceous crops
MIXED WATER ISOLATED MIXED WATER SUPPLY AGRICULTURE Rainfed - Isolated Small Fields of Multiple Herbaceous crops (Cereals) OR Post Flooding Isolated Small Fields of Multiple
SUPPLY (crop fields 20-10% of the area) Herbaceous crops
Irrigated - Clustered Small Fields of Herbaceous crops (single (Rice, Banana) or multiple herbaceous crops)
MIXED UNITS OF MIXED WATER SUPPLY Rainfed - Continuous Small Fields of Herbaceous crops (single or multiple) - cereals (Sorghum, Maize)
AGRICULTURE
(crop fields >50% of the area) Rainfed - Clustered Small Fields of Herbaceous crops (single or multiple) - cereals (Sorghum, Maize)
Post Flooding - Mixed Continuous Small Fields of Multiple Herbaceous crops
CLOSED TREES - (Forest) Closed Trees (> 65%) with Shrubs from Closed to Open
CLOSED SHRUBS
Closed Shrubs (>65%)
(crown cover > 65%) - (Thicket)
General Open Trees (15-65%) with Sparse Herbaceous
GENERAL OPEN TREES General Open Trees (15-65%) with Close to Open Shrubs
(crown cover 65-15%) - (Woodland) CELLULAR Very Open Trees (40-65%) with General Open Shrubs and Herbaceous (Tiger Bush)
STRIPED Very Open Trees (40-65%) with General Open Shrubs and Herbaceous (Tiger Bush)
WOODED
General Open Shrubs (15-65%)
VEGETATION
General Open Shrubs (15-65%) with Open Herbaceous
GENERAL OPEN SHRUBS
General Open Shrubs (15-65%) with Sparse Trees
(crown cover 65-15%) - (Shrubland)
General Open Shrubs (15-65%) with Herbaceous from Closed to Open (on ex-irrigated schemes)
Open Woody (40-65%)
CLOSED TO OPEN TREES OR WOODY VEGETATION Closed Trees (> 65%) on Temporarily Flooded Area
ON TEMPORARILY FLOODED AREAS (cover > 40% Open Trees (40-65%) on Temporarily Flooded Area
- < 4 months)
General Open Woody vegetation (15-65%) on Temporarily Flooded Area
Herbaceous (15-100%)
HERBACEOUS - (Grassland) General Open Herbaceous (15-65%) (on ex-irrigated schemes)
Sparse Herbaceous (1-15%)
HERBACEOUS ON TEMPORARILY FLOODED AREAS Closed Herbaceous (> 65%) on Temporarily Flooded Area
(cover > 40% - <4 months) General Open Herbaceous (15-65%) on Temporarily Flooded Area
HERBACEOUS ON FLOODED AREAS
RANGELAND Closed Herbaceous (> 65%) on Perennial Flooded Area
(cover > 40% - >4 months)
Herbaceous (1-100%) with Sparse Shrubs
Herbaceous (15-100%) with Sparse Trees and Shrubs
SAVANNA (and other spaced Woody Vegetation) Sparse Shrubs (1-15%)
Sparse Shrubs (1-15%) with Sparse Herbaceous
Sparse Trees (1-15%) with Sparse Shrubs
URBAN AND ASSOCIATED AREAS Airport

Urban Areas/Settlements
Loose and Shifting Sands
Dunes
BARE AREAS Salt Crust (Sandy area)
Bare Soil
Bare Soil with Scattered Vegetation
OTHER
Artificial Perennial Waterbodies
Natural Waterbody
Seasonal Rivers (Wadi)
WATERBODIES
Seasonal Rivers with Scattered Vegetation (Wadi)
River
Tidal Area
Figure 70: Cartographic aggregations
35
Results
4 RESULTS
4.1 Northern AOI results from survey validation
All available field data was evaluated as a whole and combined with edited land cover
and satellite imageries to formulate an appropriate usage of classes that was adopted in
the final legend. The indications given for the revision of the base map were also based
on partial correlations with geomorphological features that became evident during field
data analysis. Integration between the two layers will be completely developed during
preparation of the Land Mapping Units. A description of the codes followed is provided in
Table 3.
The revision mainly concerned:
- Natural Herbaceous layer This layer is strongly affected by rainfall variability and
livestock movements. It is related to deeper soils (sandy and clay). Considering the
erratic rainfall and periodic droughts that characterise the country, it was decided to
describe the herbaceous presence by adopting a class defined by a wide range of
coverage (from scattered to open/closed).
- Rainfed Herbaceous crops The presence of one or more herbaceous intercropped

crops is related to the amount of rain in each season, making it difficult to recognize the
different typology on the ground. It was therefore decided to adopt new classes without
this information.
- Identification of classes to be used in specific areas:
1 Sparse Shrub with Herbaceous (2SR6) was confirmed when used on the dissected
plateau and mountain slopes, but replaced with Sparse Shrubs Only (2SR) where soil was
shallow to discontinuous (very stony or rocky) in the sedimentary/limestone crests (all
the hills, mesas, buttes and other outcrops on the plateau).
2 The classes 2HR, 2SP, 2SP7and 2TP8 were adopted on mountain slopes at medium-
high altitudes and on the eastern side of the AOI. In these classes the herbaceous layer
was indicated as sparse, or omitted.
3 A soil (6S) used to label toggas (riverbeds).
4 Bare Soil with Scattered Vegetation (6SV) was used to label stony/rocky slopes and
badlands.
5 Seasonally Wet Waterbodies (8WFN and 8WFNV) were confirmed in the coastal area,
but were changed to Bare Soil (6S) in riverbeds of river valleys within the mountain
range, pediment and dissected plateau.
- Identification of mixed units to be used in specific areas:
1 River valleys several toggas are buffered by two particular land cover types: Irrigated
Agriculture (tree crop and cash crops) and General Open Trees and Shrubs. It is easy to
separate them from surrounding areas such as riverbeds, but not from each other. It was
therefore decided to label areas with the mixed unit 2TP8//TR3H57 eventually combined
with bare soil (6S).
2 Farming systems/schemes on the plateau the distribution of cultivated, grazed and
charcoal collection areas and consequently the status of the adjacent areas of natural
vegetation, is related to the proximity and density of settlements. These variations are
however not easily detectable at the adopted scale of work unless they cause abrupt
changes and discontinuities in the vegetation physiognomy. As a consequence, the land
cover is more homogeneous than it appears from the field survey. It was therefore
decided to select some sets of classes to be used in combination with classes of rainfed
36
Results
agriculture: (2TP6) in the south east area, close to tiger bush; (2HL78) in the more
populated westward area; and (2HL8 and 2HL) in the more degraded areas. All these
classes are present on the gentle rolling plateau where soil is continuous and fairly deep
(sandy-loamy soil). In these classes the range adopted for describing the coverage of the
herbaceous layer is extremely wide (from scattered to open) to compensate for its
annual variability.
Table 3: Land Cover codes
2SR6 Sparse Shrubs with Herbaceous 1-15%

2SR Sparse Shrubs
2HR Sparse Herbaceous
2SP General Open Shrubs with Open Herbaceous
2SP7 General Open Shrubs with Trees 1-15%
2TP8 General Open Trees with Shrubs 1-65%
6S Bare Soil
6SV Bare Soil with Scattered Vegetation
8WFN Seasonal Rivers
8WFNV Seasonal Rivers with Scattered Vegetation
General Open Trees with Shrubs 1-65% OR Irrigated Continuous
2TP8//TR3H57 Small Fields of Tree crops (fruit trees) and Herbaceous crop (pulses
and vegetables) (with distinct patterns on the same field)
2TP6 General Open Trees with Herbaceous 1-65%
2HL78 Closed to Open Herbaceous (>15%) with Sparse Trees and Shrubs
2HL8 Herbaceous (>1%) with Sparse Shrubs (1-15%)

2HL Closed to Open Herbaceous (>15%)
4.2 Land Cover cartographic aggregations
Cartographic aggregations of the northern AOI are shown in Map 1, while Map 2
corresponds to the southern AOI.
There are two main considerations guiding the grouping of natural vegetation classes.
The first involved an attempt to display the ability of vegetation to protect soils. The
main related feature is cover (coverage). In order to show different levels of protection,
grouping according to LCCS spatial distribution rules was assumed to be sufficient.
The second consideration is that separating savanna from woodland is subjective, since
the two terms have several definitions, their floristic composition is similar and detection
by photointerpretation is difficult. It was finally decided that savanna aggregation refers
to any kind of African savanna (shrub savanna and tree savanna) and any very open
wooded vegetation that can be partially matched with Xerophyllous open woodland and
Subdesert bush and thicket.
NB - At the taxonomic level of Main Aggregations it is assumed that, if agricultural

types are defined as clustered or isolated, the influence of activities depending on natural
vegetation should also be considered.
Annex 1 gives the vegetation of the AOI as established from the field assessment.
37
Results
38
Results
39
Results
40
Results
4.3 Land Cover surface statistics
4.3.1 Dur-Dur and Gebiley AOI northern AOI
Results presented for the northern AOI are derived from statistical analysis of aggregated
surfaces of the Land Cover map, produced through visual interpretation of satellite
images and validated by data collected in the field.
The results described here show the composition and distribution of land cover types of
the Dur-Dur and Gebiley AOI (Map 1).
The northern AOI has a total area of 1 293 899 ha, with 83% covered by Natural
Vegetation (Chart 1).
LANDCOVER TYPES - MAIN AGGREGATIONS
7%
10%
49%
34%
RANGELAND WOODED VEGET AT ION AGRICULT URE OT HER
Chart 1: Distribution of main aggregations of Land Cover types
The more representative vegetation type is Rangeland, covering 49% of the mapped area
(about 638 ha). The remaining area is distributed between Wooded Vegetation (441 ha)
and Agriculture (125 ha).
About 91 ha are occupied by Non Vegetated Areas, indicated as Other in the chart. These
include:
- Urban and Associated Areas (Settlement/Towns and Airport)

- Bare Areas (Bare Soils and Sandy areas)
- Natural Waterbodies
41
Results
RANGELAND
The Rangelands of the AOI are mainly formed of Savanna and only in small part by pure
grassland (Chart 2).
RANGELAND
8%
92%
SAVANNA HERBACEOUS - (Grassland)
Chart 2: Rangelands distribution
The total area covered by Savanna is 587 000 ha (Table 4). The most representative
vegetation type is composed of Sparse Shrubs occasionally in association with Sparse
Trees. From a physiognomic point of view it consists mainly of sparse wooded vegetation
with a layer of herbaceous vegetation influenced by rainfall and soil types, as well as
grazing pressure).
Sparse shrubs with a herbaceous layer are mainly located on the dissected plateau and
on mountain slopes, while sparse shrubs only are found where soil is shallow to
discontinuous (very stony or rocky) on sedimentary/limestone crests (all the hills, mesas,
buttes, and other outcrops on the plateau). Grasslands cover a low percentage of the
area and are mainly distributed on sandy and clay soils.
Table 4: Rangeland areas
%
SAVANNA 586 729
cover
Sparse Trees (1-15%) with Sparse Shrubs 199 993 31
Sparse Shrubs (1-15%) with Sparse Herbaceous 133 695 21
RANGELAND Sparse Shrubs (1-15%) 129 095 20
637 556 Herbaceous (1-100%) with Sparse Shrubs 81 379 13
Herbaceous (15-100%) with Sparse Trees and
42 568
Shrubs 7
HERBACEOUS (Grassland) 50 828
Herbaceous (15-100%) 25 521 4
Sparse Herbaceous (1-15%) 25 307 4
42
Results
WOODED VEGETATION
Wooded vegetation present in the AOI is composed almost equally of general open
formations of shrubland and woodland (Chart 3).
WOODED VEGETATION
50%
48%
1%
1%
GENERAL OPEN SHRUBS (Shrubland)
GENERAL OPEN T REES (Woodland)
CLOSED SHRUBS (T hic ket)
CLOSED T REES (Forest)
Chart 3: Wooded vegetation types distribution
A main layer of Shrubs, in some areas associated with a layer of sparse trees and/or
herbaceous, forms Shrublands with a total surface area of 220 709 ha (Table 5).
Woodlands consist of a primary tree layer, with or without shrubs and very open trees
with shrubs with a fragmented macropattern (tiger bush). Tiger bush can display an
interrupted cover of either striped or cellular fragmentation. Degraded types have a
cellular pattern.
Table 5: Wooded vegetation areas
%
GENERAL OPEN TREES (crown cover 65-15%) (Woodland) 211 942
cover
General Open Trees (15-65%) with Sparse Herbaceous 12 193 3
General Open Trees (15-65%) with Close to Open Shrubs 169 004 38
CELLULAR Very Open Trees (40-65%) with General Open Shrubs and
Herbaceous (Tiger Bush) 15 028 3
STRIPED Very Open Trees (40-65%) with General Open Shrubs and
Herbaceous (Tiger Bush) 15 717 4
WOODED GENERAL OPEN SHRUBS (crown cover 65-15%) (Shrubland) 220 709
VEGETATION General Open Shrubs (15-65%) 47 467 11
440 715 General Open Shrubs (15-65%) with Open Herbaceous 52 668 12
General Open Shrubs (15-65%) with Sparse Trees 120 574 27
CLOSED SHRUBS (crown cover > 65%) (Thicket) 5 752
Closed Shrubs (>65%) 5 752 1
CLOSED TREES (Forest) 2 313
Closed Trees (> 65%) with Shrubs from Closed to Open

2 313 1
43
Results
With the exception of tiger bush present in the southern parts of the AOI towards the
Ethiopian border, these formations are mainly located on mountain slopes at medium-
high altitude and on the eastern side of the AOI.
Only 2% of the AOI is covered by dense vegetation; less than 6 000 ha is occupied by
closed shrubs vegetation; while forests cover about 2 300 ha.
AGRICULTURE
Cultivated areas represent 10% of the total AOI. Their total surface is equal to about 124
800 ha. The most widespread agricultural practice is rainfed agriculture (>90%) where
crop establishment and development is completely dependent on rainfall. Irrigated
cultivation occupies less than 10% of the mapped area (Chart 4).
AGRICULTURE
9%
91%
RAINFED IRRIGAT ED
Chart 4: Agriculture distribution
The total area of rainfed agriculture is about 114 000 ha (Table 6). It is characterised by
fields of small dimensions (<2ha) growing cereals. Maize and sorghum are the most
important crops and are cultivated either as single crops or in combination. Depending on
rainfall, the same crop (maize maize) or two different crops (maize sorghum) can be
cultivated sequentially, i.e. in sequence on the same field within one year.
Rainfed fields of a shrub crop are also present (Catha edulis Qaat) but their size is too
small and distribution too scattered to be mapped at the scale of work adopted.
Table 6: Agriculture areas

%
RAINFED 113 998
cover
Medium Fields (2-5ha) 24 744 19.8
AGRICULTURE Small Fields (<2ha) 89 254 71.5
124 834
IRRIGATED 10 837
Small Fields (<2ha) 10 286 8.2
Mixed units 551 0.4
44
Results
Nearly 11 000 ha are cultivated, supplying water to crops through an irrigation system.
Irrigated fields are usually distributed along seasonal rivers. Water is pumped to the
crops through pipes. Fields are usually cover less than 2 ha, with fruit tree crops
(papaya, mango, and lemon) and vegetables (tomatoes, onions and watermelons).
Due to their location along the toggas where open or fragmented riparian formations of
natural trees are also present, identification of irrigated fields is quite difficult.
This uncertainty, encountered during interpretation work, required the creation of a

special class described in the legend as General Open Trees with Shrubs 1-65% OR
Irrigated Continuous Small Fields of Tree crops (fruit trees) and Herbaceous crop (pulses
and vegetables) (with distinct patterns on the same field). It was felt to be most
appropriate to aggregate the areas classified with this code as Mixed units of irrigated
agriculture.
4.3.2 Riverine area southern AOI
Results for the southern AOI are derived from statistical analysis of aggregated surfaces
of the Land Cover map produced during Phase I, integrated with additional areas
interpreted from satellite images during Phase II. The preliminary results were validated
in the month of March 2007 when the field work was conducted. Ultimately, the final
results were obtained and presented in this report.
The southern AOI has a total area of 10 399 916 ha, 74% of which is covered by Natural
Vegetation (Chart 5).
LAND COVER TYPES - MAIN AGGREGATIONS
9% 2%
24%
65%
Wooded Agriculture Rangeland Other
Chart 5: Distribution of main aggregations of Land Cover types
45
Results
The total area of 7 627 926 ha of natural vegetation are composed of 65% wooded
vegetation (6 671 021 ha) and 9% rangelands. Around 2 536 709 ha are occupied by
cultivated areas.
The remaining part of the investigated area, around 235 280 ha, is covered by Non
Vegetated Areas, indicated as Other in the chart. This aggregation includes the
following Land Cover types:
- Urban and Associated Areas (Settlement/Towns and Airport)

- Bare Areas (Bare Soils and Dunes/Sandy areas)
- Natural/Artificial Waterbodies.
WOODED VEGETATION
This aggregation includes natural and semi-natural vegetated areas of trees and shrubs
on terrestrial or temporarily flooded areas (Chart 6).
Shrubland is the dominant wooded vegetation type with a total surface of 7731282 ha
(80%) of the total area covered by the wooded vegetation.
WOODED VEGETATION
3% 5% 2%
10%
80%
GENERAL OPEN SHRUBS (crown cover 65-15%) (Shrubland)
GENERAL OPEN TREES (crown cover 65-15%) (Woodland)
CLOSED TREES - (Forest)
CLOSED SHRUBS (crown cover > 65%) (Thicket)
CLOSED TO OPEN TREES OR WOODY VEGETATION ON TEMPORARILY FLOODED
Chart 6: Wooded vegetation types distribution
The dominant layer consists of shrubs or woody vegetation (an intricate mixture of trees
and shrubs), with crown coverage ranging from 15-65%, in some cases associated with a
second layer of herbaceous vegetation. This type of vegetation is also present on
formerly irrigated schemes areas abandoned during the civil war period.
Woodland represents 10% of the natural vegetation of the AOI. Both vegetation forms,
trees and shrubs, are present with crown cover from very open to open, but in this case
trees are the dominant layer.
46
Results
Very open trees with shrubs, characterised by a fragmented macropattern (tiger bush)
are also included in this group. The total surface covered by woodland was equal to 999
544 ha as indicated in Table 7.
Only 8% of the AOI is covered by dense vegetation: around 323 274 ha are occupied by
forests, while closed shrubs vegetation amount to less than 497 657 ha.
In the southern AOI there are transitional areas between pure terrestrial and aquatic
systems (Lower Juba and Lower Shabelle regions), consisting of areas that are regularly
flooded but where water cover does not remain for a substantial period of time. The
vegetation of these temporary flooded areas consists of closed to open trees or woody,
and occupies a total surface of about 195257 ha.
Table 7: Areas of Wooded vegetation
WOODED
VEGETATION GENERAL OPEN SHRUBS (crown cover 65-15%) (Shrubland) 7731282
7 060 016
General Open Shrubs (15-65%) 253288
General Open Shrubs (15-65%) with Open Herbaceous 4700582
General Open Shrubs (15-65%) with Sparse Trees 2729906
General Open Shrubs (15-65%) with Herbaceous from Closed to Open (on
7136
ex-irrigated schemes)
Open Woody (40-65%) 40370
GENERAL OPEN TREES (crown cover 65-15%) (Woodland) 999 544

General Open Trees (15-65%) with Close to Open Shrubs 873 867
CELLULAR Very Open Trees (40-65%) with General Open Shrubs and
11 326
Herbaceous (Tiger Bush)
STRIPED Very Open Trees (40-65%) with General Open Shrubs and
114 351
CLOSED TREES - (Forest) 323 274

Closed Trees (> 65%) with Shrubs from Closed to Open 323 274
CLOSED SHRUBS (crown cover > 65%) (Thicket) 497 657
Closed Shrubs (>65%) 497 657
CLOSED TO OPEN TREES OR WOODY VEGETATION ON
195257
TEMPORARILY FLOODED AREAS (cover > 40% - < 4 months)
Closed Trees (> 65%) on Temporarily Flooded Area 149 964
Open Trees (40-65%) on Temporarily Flooded Area 25 385
General Open Woody vegetation (15-65%) on Temporarily Flooded Area 19 908
47
Results
RANGELAND
Rangelands of the AOI consist of 63% Savanna and 37% grassland on terrestrial or
flooded areas (Chart 7).
RANGELAND
71%
24% 4%1%
HERBACEOUS ON FLOODED AREAS HERBACEOUS ON TEMPORARILY FLOODED AREAS HERBACEOUS (Grassland) SAVANNAH
Chart 7: Rangelands distribution

Savannas are characterised by sparse shrubs and/or sparse trees with a dominant herbaceous layer. The total area
of this Land Cover type is about 905203 ha (Table 8). Formations of herbaceous vegetation only cover 255462
ha, of which 86% is terrestrial and 14% (90 825 35 ha) is represented by grasslands on
flooded areas.
Table 8: Rangeland areas

SAVANNA 649741
Herbaceous (15-100%) with Sparse Trees and Shrubs 190840
Herbaceous (1-100%) with Sparse Shrubs 353135
Sparse Shrubs (1-15%) 105766
HERBACEOUS (Grassland) 213683
RANGELAND Herbaceous (15-100%) 195806
1 757 873 General Open Herbaceous (15-65%) (on ex-irrigated schemes) 17876
HERBACEOUS ON FLOODED AREAS (Grassland) 9161
Closed Herbaceous (> 65%) on Perennial Flooded Area 9161
HERBACEOUS ON TEMPORARILY FLOODED AREAS 32618
Closed Herbaceous (> 65%) on Temporarily Flooded Area 23844
General Open Herbaceous on Temporarily Flooded Area 8774
48
Results
AGRICULTURE
Cultivated areas represent 13% of the total AOI mapped in southern Somalia, with a
total surface area of about 1 315 032 ha. Similarly to the northern AOI, the most
widespread practice is rainfed agriculture, occupying 84% of the total agricultural areas
(Chart 8). The remaining surface is cultivated using different water supply systems.
Around 6% is represented by irrigated crops and 4% by post-flooding cultivation.
AGRICULTURE
6% 2% 5%
87%
Irrigated Mixed Water Supply Post Flooding Rainfed
Chart 8: Agriculture distribution
The total area occupied by rainfed agriculture amounts to about 2570484 ha (Table 9). It
is mainly represented by fields of small to medium dimensions (up to 5 ha) cultivated
with cereals and sesame. Maize and sorghum are the most important crops and are
cultivated either as single crops or in combination. This type of agriculture is completely
determined by rainfall.
Irrigated systems are characterised by an artificial, regular supply of water to crops, in
addition to rain. The total area of irrigated crops is 129 754 ha, mainly distributed along
the Juba and Shabelle rivers. Crops are rice and other cereals, bananas, vegetables and
trees such as mango and papaya, cultivated as single or multiple crops. Field dimensions
are almost equally distributed between small and medium-sized fields. Large fields
(>5ha) represent only 5% of total irrigated cultivation and belong mainly to big farms
developed before the civil war.
49
Results
Table 9: Agriculture areas
RAINFED 2241999
Medium fields (2-5ha) of herbaceous crops 602 726
Small fields (< 2ha) of herbaceous crops 510 300
IRRIGATED 144078
Large fields (> 5ha) of herbaceous crops 6 451
AGRICULTURE Medium fields (2-5ha) of herbaceous crops 51 433
1 315 032 Small fields (< 2ha) of herbaceous or tree crops 71 870
POST FLOODING 127212
Medium fields (2-5ha) of herbaceous crops 2 082
Small fields (< 2ha) of herbaceous crops 45 035
MIXED WATER SUPPLY 57195
Rainfed OR Postflooding Small fields (< 2ha) of herbaceous crops 25 136
Post-flooding cultivation systems are also present in the AOI, with a total area of about
47 000 ha. Inundated areas are cultivated when water recedes, and water infiltrated into
the soil is used intentionally as a water reserve for crops. The dimensions of the
herbaceous fields cultivated with this system are usually small (less than 2ha).
For several areas of the AOI, satellite images were not available for different seasons,
making it difficult in some cases to distinguish between rainfed and post-flood cultivation.
This required the creation of a special class described in the legend as Rainfed
Continuous Small Fields of Multiple Herbaceous crops (Cereals) OR Post Flooding
Continuous Small Fields of Multiple Herbaceous crops. During the thematic grouping for
surface estimation it was felt to be appropriate to aggregate the areas classified with this
code under the vocation Mixed Water Supply.
50
Conclusions and Recommendations
5 5. CONCLUSIONS AND RECOMMENDATIONS
The northern AOI, Dur-Dur and Gebiley AOI (Somaliland), has a total area of 1 293 899
ha, comprising 83% Natural Vegetation and 10% Agriculture. The remaining 7% includes
Non Vegetated Areas. The most common natural vegetation type is Rangeland (49% of
natural vegetation surfaces), represented almost totally (92%) by Savanna with a small
area covered by pure Grassland (8%).
Wooded vegetation occupies 34% of the natural vegetated areas and consists almost
equally of general open formations of Shrubland and Woodland
The majority of agricultural areas are characterised by rainfed cultivation. Less than 10%
of the area is cultivated with irrigated crops.
The southern, riverine AOI has a total area of 10 397 600 ha, of which 84% is covered by
Natural Vegetation and only 3% by Agriculture. The remaining 13% includes Non
Vegetated Areas. The most common natural vegetation type is Wooded Vegetation (67%
of natural vegetation surfaces), dominated by open formations of Shrubland (67%) and
Woodland (24%).
Rangelands occupy 17% of the natural vegetated areas, and are mostly covered by
Savanna (63%). The remaining surfaces are covered by Grasslands.
The majority of agricultural areas are characterised by rainfed cultivation. Only 10% of
the area is irrigated.
The land cover geo-dataset presented in this report is an important output containing
information about cover types of the mapped areas and their distribution. It is not just a
static set of thematic maps, printable in different map scales, but a geo-referenced
dataset (in vector format) serving several purposes and with various possible uses.
All attributes are stored following LCCS structure that enhances the performance of the
database in facilitating and improving its spatial modelling and statistical analysis
capabilities. An example of this has been presented in Chapter 5.
In particular, the classifiers and syntax adopted by LCCS allows handling of the codes
according to various environmental features that can match with characteristics and
qualities adopted in FAOs framework for land evaluation (FAO, 1976). The study and
understanding of these classifiers by users is highly recommended in order to maximise
LCCS potential.
Spatial modelling and aggregations provide further improved results. The Land Cover can
also be used as a base map. Additional information, particularly from field surveys, can
be combined through a process of synthesis and linkage, promoting the land cover to
new thematic layers involving different disciplines such as geo-botany (vegetation
communities and formations), planning (land use) and other environmental studies.
By producing the new thematic layers in this way it is possible to obtain a more stable
and consistent model, as:
- topological errors can be controlled more efficiently.

- conceptual consistency can be easily maintained, avoiding incompatibility between
layers.
- simplifying the geometry of the various feature classes dataset can facilitate and
optimise geo-processing activities like intersection and update, where different
51
Conclusions and Recommendations
layers are combined, and polygons are increased using the original boundaries of
all entered layers.
Further possible uses of the Land Cover might include interventions in sub-catchment
and river basin catchment areas (FAO, 1993)..
Both map scales adopted (1:50 000 for agricultural areas of the northern AOI and 1:100
000 for the whole extent of the two AOI) are suitable for inventory use and for
performing supporting interventions such as:
- district-level multipurpose land use planning for (re)settlement schemes

(Subcatchment areas - scales of inventory and evaluation: 1:50 000 to 1:100
000);
- catchment-level assessment of water resources and their potential use, especially
downstream (conservation or reforestation projects in upstream parts, irrigation
and drainage projects in lower parts; fishery development); identification and
protection of high biodiversity values; national parks and indigenous delineation of
reserves (river basin catchments areas - scale of inventory and evaluation: 1:100
000 to 1:250 000).
This geo-dataset constitutes one of the baseline thematic layers for assessment of land
suitability, soil erosion and land degradation, through generation of Land Mapping Units.
This kind of spatial analysis requires combination of the Land Cover with other thematic
layers, in particular landform, climate classifications and soil datasets.
In order to maximise quality and accuracy of the results, the following activities were
incorporated in the study:
- field data collection in both the northern and southern AOI for validation of the
unverified results;
- the use of IKONOS (or Spot) images in addition to or as a partial substitution of
the survey (in the event that security issues disallow it);
- take advantage of IKONOS and extend coverage - some IKONOS frames were
acquired during SWALIM Phase I and used to improve interpretation as explained
in Chapter 3.2.2. Continuing this activity in the southern AOI and extending it to
the northern AOI could be very useful for further validation of the map;
- perform meaningful aggregations, starting from the ones presented in the results,
for developing the land use dataset;
- use the Land Cover geo-dataset extensively, exploring its potential for forming the
facets constituting the administrative units of the landscape in the Land Mapping
Units. New and better-fitting aggregations should be developed in a gradual,
interactive way, checking the results at each step, leading to more complex
aggregations. Particularly, a the more reliable way of linking LCCS codes to well-
defined vegetation and land use types, should be explored with the support of
survey results. In a second phase, the facets aggregation should be combined
with optimized landform aggregations through an intersection operation.
52
The Vegetation Map of Northern AOI
6 THE VEGETATION MAP OF THE NAOI
6.1 Introduction
The vegetation map of the Dur-Dur and Gebiley Area of Interest (AOI) was derived from
the Land Cover map and from quadrat sampling during the land cover field surveys. The
detailed land cover map was a generalised one, providing broad land cover classes of the
AOI. The vegetation in each land cover class was described using results of the field
vegetation sampling, due to the limited number of samples available for producing a
more detailed vegetation map of the AOI. This may be attributed to the fact that the
vegetation assessment was not very detailed, focusing only on the composition of
dominant species. The main foci were vegetation cover and frequency. Vegetation was
assessed at the same time as the land cover and other thematic surveys. However, the
results presented here give a realistic picture of the vegetation of the AOI.
Floristic composition is an important attribute of vegetation studies as it aids in

facilitating comprehensive classification of vegetation. Knowledge of vegetation allows an
understanding of environmental conditions affecting intensity of degradation and grazing.
Data collection for this study was undertaken between May and June, 2006. The main
objective was to produce a generalised vegetation map of the Northern AOI, to provide a
plant species inventory and to determine the main vegetation types.
6.2 Materials and Methods

The adopted methodology provided a rapid assessment of the abundance and
frequency of all species sampled. Trees and shrubs were assessed separately from
herbaceous vegetation, using different methodologies.
6.2.1 Trees and shrubs

Trees and shrubs were assessed using a tape measure along line transects of 100 m. In
this method, each whole number (1 m, 2 m, 3 m and so on up to 100 m) was considered
as a node. For each of the 100 nodes, woody crown interceptions with the line transect
were noted. At each interception, the plant was identified and the corresponding node
noted. The approach involved assigning the closest node to the trunk.
For each species thus selected, the following were required:
1. an identifier in the species list form.

2. the species name in the species list form.
3. to note an identifier in the box of the corresponding node in the floristic relieve
form.
4. to measure the length of tape intersecting the crown (under its shadow) and to
write the measurement (in cm) in the box of the corresponding node in the
floristic relieve form.
6.2.2 Herbaceous Vegetation

Every 20 m along the line transect, a 50x50 cm box, or quadrat was used to assess the
herbaceous vegetation. The tape unrolled several times to reach 100 meters, allowing
the definition of 5 segments of 0.5 m each. A 25 cm ground radius was examined on
either side of the tape (i.e. a diameter of 50 cm). The details of this procedure are as
follows:
1. Starting from the first quadrat (a box of 50x50 cm, from 0 to 0.5 m) identify all
species inside it and list them on the floristic list form. Take care to write the
correct segment number on the floristic form corresponding to the line transect
segment.
53
2. For each species, assess the coverage rate expressed in percentage or using the
Braun-Blanquet scale, refer to the segment area.
3. For each species measure the height in cm.
4. Pass to the next segment, from 0.5 up to 1 m. Repeat the operation. All the
species already found in previous segments are included in the new floristic list
form with the same identifier.
5. Go through all the segments until the last (the tenth), repeating the same
operations.
All unidentified species were collected and labelled with a piece of masking tape, writing
numbers of the Site ID and Species ID on it. For herbaceous plants it is important to
collect sufficient parts of the plant to assist in identification; in the case of woody plants,
a small branch was collected. Care was taken to keep leaves and/or buds and/or flowers
and/or fruits to aid identification. Collected material was preserved in a plant press.
Crown section
Node
1m 2m 3m 4m 5m
Crown
Tape measure
Node number 2
corresponding to the
selected plant
Figure 6. Showing how to use the ruler, select plants and take measurements.
9m 10 11m 12
Figure 7. Showing the second 0.5x0.5 m segment, located at 10 m.
Figure 8. The Floristic Relieve Form
54
FLORISTIC RELIEVE
Node Number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
ID 1
Species Identifier
ID
ID
ID
ID
ID
ID
ID
ID
ID
ID
ID
ID
ID
ID
ID
ID
ID
ID
ID
cm 220
cm
cm
cm
cm
cm
cm
cm
cm
cm
cm
cm
cm
cm
cm
cm
cm
cm
cm
cm
Section Cm
FLORISTIC LIST FORM
Species Identifier Species Name Tot Crown Section Cm
1 Acacia bussei
Here is an example of how to fill the floristic relieve and the floristic list form.
Herbaceous FLORISTIC RELIEVE BOX N._________________

FLORISTIC LIST FORM SITE N.:________________
Segment N. (from1 to5):_______
Species Identifier species Name Height Cover
The Braun-Blanquet scale:
R = rare
+ = coverage not measurable
1 = coverage < 5%
2 = coverage 5% 25%
3 = coverage 25% - 50%
4 = coverage 50% - 75%
5 = coverage > 75%
In the data analysis, the cover and frequency of occurrence for the different species
present in each sample were established. The samples were grouped according to the
various cover classes and used to characterize and eventually map the vegetation in the
AOI. In this classification, the species composition, in order of dominance was
established.
55
6.3 Results
The results of the vegetation assessment included a vegetation map with 25 classes as
shown in Figure 70 below. A summary of results are:
1. Closed Shrub this class covered about 5,525 ha and included the following:
Woody Species:
Acacia etbaica, A. nilotica, A. senegal, A. bussei, Cassia obovata, Grewia tenax,
Hypoestes hildebrandtii, Aloe spp., Solanum carense.
Herbaceous Species:
Chrysopogon auchieri, Sporobolus marginatus, Cenchrus ciliaris
2. Closed Shrubs/Isolated Rainfed Crops (267 ha)

Woody Species:
Hypoestes hildebrandtii, Aloe spp., Solanum carense
Herbaceous Species:
Crops:
Sorghum, Maize, Cowpeas, Simsim, Millet
3. Closed Trees (2 198 ha)

Woody Species:
Acacia tortilis, A. nilotica, A. bussei, Prosopsis juliflora, Sueda fruticosa
4. General Open Shrubs (157 211 ha)

Woody Species:
Herbaceous Species:
5. General Open Shrubs/Clustered Rainfed Crops (31 707 ha)

Woody Species:
Herbaceous Species:
Crops:
Sorghum, Maize, Cowpea, Simsim, Quat
6. General Open Shrubs/Isolated Rainfed Crops (59 577 ha)

Woody Species:
Herbaceous Species:
Crops:
Sorghum, Maize, Cowpea, Simsim, Quat
7. General Open Trees (211 638 ha)

Woody Species:
Acacia etbaica, A.nilotica, A. mellifera, A. bussei, Balanites glabra, Solanum incanum
Herbaceous Species:
Aloe somaliensis, Chrysopogon auchieri, Dactyloctenium aegyptium, Cynodon
dactylon, Sporobolus marginatus
56
57
The Vegetation of Northern AOI
8. General Open Trees/Clustered Rainfed Crops (777 ha)

Woody Species:
Herbaceous Species:
Crops:
Sorghum, Maize, Cowpea, Simsim, Millet
9. General Open Trees/Irrigated Orchards (13 345 ha)

Woody Species:
Acacia tortilis, Dobera glabra, Acacia bussei, Grewia tenax
Crops:
Papaya, Citrus, Guava, Watermelon, Tomato, Banana
10.General Open Trees/Isolated Rainfed Crops (3 568 ha)

Woody Species:
Herbaceous Species:
Crops:
Sorghum, Maize, Cowpea, Simsim, Millet
11.Herbaceous Vegetation (3 541 ha)

Dactyloctenium aegyptium, Panicum turgidum, Cleome brachycarpa, Aristida
adscensionis
12.Herbaceous Vegetation/Clustered Rainfed Crops (7 338 ha)

Herbaceous Species:
Dactyloctenium aegyptium, Panicum turgidum, Cleome brachycarpa, Heliotropium
undilatifolia
Crops:
Sorghum, Maize, Millet, Simsim, Cowpea, Quat
13.Herbaceous Vegetation/Irrigated Orchards (136 ha)

Herbaceous Species:
undulatifolium
Crops:
Papaya, Citrus, Guava, Watermelon, Tomato, Banana
14.Herbaceous Vegetation/Isolated Rainfed Crops (49 618 ha)

Herbaceous Species:
undulatifolium
Crops:
Sorghum, Maize, Millet, Simsim, Cowpea, Quat
15.Irrigated Orchards (172 ha)

Crops:
Papaya, Citrus, Watermelon, Tomato, Guava, Cucumber, Mango, Banana
16.Rainfed Crop Fields

Crops:
Sorghum, Maize, Millet, Cowpea, Quat, Simsim
58
17.Savannah (519 476 ha)

Woody Species:
Acacia nubica, A. tortilis, A. senegal, A. bussei, Aloe spp., Croton gillettii, Hypoestes
hildebrandtii, Acalypha fruticosa, Grewia tenax, Balanites aegyptiaca,
Herbaceous Species:
Cenchrus ciliaris, Cynodon dactylon, Sporobolus marginatus, Tragus racemosus,
Aristida adscensionis.
18.Savannah/Clustered Rainfed Crops (10 915 ha)

Woody Species:
hildebrandtii, Acalypha fruticosa, Grewia tenax, Balanites aegyptiaca,
Herbaceous Species:
Crops:
19.Savannah/Isolated Rainfed Crops (2 500 ha)

Woody Species:
hildebrandtii, Acalypha fruticosa, Grewia tenax, Balanites aegyptiaca
Herbaceous Species:
Crops:
20.Very Open Trees (50 428 ha)

Woody Species:
Acacia etbaica, A. nilotica, A. mellifera, A. bussei, Balanites glabra, Solanum incanum
Herbaceous Species:
Crops:
Sorghum, Maize, Cowpea, Simsim, Millet, Quat
21.Very Open Trees/Isolated Rainfed Crops (5 573 ha)

Woody Species:
Acacia etbaica, A. nilotica, A. mellifera, A. bussei, Balanites glabra, Solanum incanum
Herbaceous Species:
Crops:
Sorghum, Maize, Millet, Cowpea, Simsim, Quat
22.Urban Settlement (4625 ha)
23.Rural Settlement (2 281 ha)
24.Airport (42 ha)
25.Bare Areas (44 259 ha)
59
6.4 Conclusions and Recommendations
Samples used to characterise vegetation in the AOI were few and limited, resulting in a
vegetation map that is very general, and a more detailed assessment of vegetation is
desired. Many vegetation attributes, such as biomass and density, were not considered in
the assessment. Information on species composition was not exhaustively established, as
most unidentified samples were not subjected to identification procedures and remain
unknown.
Sampling was only conducted during the dry season. It is recommended that more
intensive field work be performed in both dry and wet seasons to produce a more
representative picture of the vegetation of the AOI.
Given that negative activities like charcoal burning are on the increase in the AOI and
that charcoal burning is selective with regard to tree species, it is important to monitor
trends in tree density over time. However, the results of this study form an important
basis for further investigations on vegetation of the AOI.
The results indicate that the AOI is predominantly covered by natural vegetation. This
explains why the economy of the AOI is basically pastoral, which in turn explains why
vegetation studies should be given more attention as they form the basis of livestock
management.
It is also desirable that a more comprehensive report on species composition is produced,

showing which plant species are most threatened by human activities. This information
would be an important input to pastoral land use policy formulation. Acacia bussei is the
tree most exploited for charcoal for example, and therefore needs protection. Density
trends of A. bussei are highly desirable and can only be achieved through exhaustive
vegetation studies.
60
Bibliography
7 7. BIBLIOGRAPHY
AA. VV. 2003. Flowers of Ethiopia and Eritrea, Aloes and other Lilies. Shama's Nature
Series 227 pp.
Barkhadle, M.I.A. 1993. The Somali traditional "Deegaan" Ecological zone classification
map and plant names. Riv.Agric.Subtrop. e Trop. in press.
Barkhadle, M.I.A., Ongaro L. & Pignatti S. 1994. Pastoralism and plant cover in the lower
Shabelle region, Southern Somalia. Landscape Ecology 9(2):79-88
Bird, N.M. & Sheperd, G. 1989. Charcoal in Somalia: A woodfuel inventory in the Bay
region of Somalia. Overseas Development Administration Library, UK.
Carbone, F. & Accordi, G. 2000. The Indian Ocean coast of Somalia. Marine Pollution
Bulletin 41:141-59. (Univ La Sapienza, CNR, Ctr Studio Quaternario & Evoluz
Ambientale, Dipartimento Sci Terra, P Aldo Moro, 5I-00185 Rome, Italy,
Dharani, N. 2002. Field guide to common trees and shrubs of East Africa. Struik.
Dharani, N. 2006. Field guide to Acacias of East Africa. Struik.
Di Gregorio, A. & Jansen, J.M.L. 2000. Land Cover Classification System (LCCS)
Classification Concepts and User Manual FAO. Rome, Italy.
Jansen, J.M.L. 2005. Harmonisation of land-use class sets to facilitate compatibility and
comparability of data across space and time. 12th CEReS International Symposium,
Chiba, Japan.
Di Tommaso, P. 1992. Geobotanica. CUSL. 286 pp.
FAO Africover Project. 1999. Land cover mapping based on satellite remote sensing.
Somalia. http://www.africover.org/index.htm
FAO. 1968. Agricultural and water surveys Somalia. Final Report, Vol. I, General.
Report prepared by Lockwood Survey Corporation Ltd.
FAO. 1968. Agricultural and water surveys Somalia. Final Report, Vol. III, Landform
and Soils. Report prepared by Lockwood Survey Corporation Ltd.
FAO. 1968. Agricultural and water surveys Somalia. Final Report, Vol. IV, Livestock
and Crop Production. Report prepared by Lockwood Survey Corporation Ltd.
FAO. 1976. A framework for land evaluation. FAO Soils Bulletin 32. Food and
Agriculture Organisation of the United Nations, Rome.
FAO. 1990. Remote Sensing Applications to land Resources. 299 pp.
FAO. 1993. Land and water integration and river basin management. Proceedings of
an FAO informal workshop, Rome, Italy. 31 January-2 February 1993
FAO-UNDP. 1969. Project for the water control and management of the Shebelli River,
Somalia. Volume I General Report. 227pp. Hunting Technical Services Ltd. in
association with Sir M. MacDonald & Partners.
FAO-UNDP. 1969. Project for the water control and management of the Shebelli River,
Somalia. Volume V Soils and Agriculture. 262pp. Hunting Technical Services Ltd
in association with Sir M Macdonald & Partners.
FAO. 2005. Land cover classification system. Software version 2. Rome, Italy.
Gillet, J.B. 1941. The plant formation of western British Somaliland and the Harar
Province of Abyssinia. Kew Bull.1941:37-199.
Gilliland, H.B. 1952. The vegetation of eastern British Somaliland. J. of Ecology 40:91-
124.
Giordano A. 1989. Il telerilevamento nella valutazione delle risorse naturali. Relaz. e
Monogr. Agr. Subtrop. Trop. N.106. IAO, Firenze, Italy.
Giordano A. 1999. Soil and Soil conservation with focus on remote sensing, IAO Relaz. e
Monogr. Agr. Subtrop. Trop n. 116, IAO, Firenze, Italy, 152 pp.
Gomarasca M. 1997. Introduzione a telerilevamento e GIS per la gestione delle risorse
agricole ambientali. Italy.
61
Bibliography
GTZ. 1990. Masterplan for Juba Valley development. Annex 1 - Land resources and
Land Use, Annex 4 Crop production, Annex 8 - Forestry and Range. Somali
Democratic Republic.
GTZ. 1990. Masterplan for Juba Valley development. Main Report. Somali Democratic
Republic.
Hemming, C.F. 1965. Vegetation arcs in Somaliland. J. of Ecology 53(1):57-67.
Hemming, C.F. 1966. The vegetation of the northern region of the Somali Republic.
Proc.Linn.Soc.Lond. 177(2):173-248.
Iannelli, P. 1988. I pascoli della Somalia. Relaz. e Monogr. Agr. Subtrop. Trop. n.103,
IAO, Firenze, Italy. 161 pp.
IIRR. 2002. Managing Dryland Resources - A manual for eastern and southern Africa.
Sommerlatte M. & Umar A. 2000. An ecological assessment of the coastal plains of
North Western Somalia (Somaliland). Somali Natural Resources Management
Programme, IUCN Eastern Africa Programme. 63 pp.
Lillesand, T.M. & Kiefer, R.W. 1979. Remote Sensing and Image Interpretation. John
Willey & Sons Inc, New York.
Kruckeberg, A.R. 2004. Geology and plant life. University of Washington Press. 361pp
MacFayden, W.A. 1950. Vegetation patterns in the semi-desert plains of British
Somaliland. Geogr. J. 116(4/6):199-211.
Ministry of Pastoral Development and Environment. 2004. Case study on economic
impact of charcoal on Socio Economy and Environment (DRAFT)
Ongaro, L. 1998. Land unit mapping for Land evaluation. Relaz. e Monogr.Agr.
Subtrop.Trop. N.115. IAO, Firenze, Italy.
Pallister, J.W. 1963. Notes on the geomorphology of the Northern Region, Somali
Republic. Geogr. J. 1929(2):184-187.
Pichi-Sermolli, R.E.G. 1957. Una carta geobotanica dellAfrica Orientale. Webbia 8, 129
pp
Rossiter, G.D. 1994. Land Evaluation Lecture Notes. Cornell University.
http://www.css.cornell.edu/landeval/le_notes/lecnot.htm
Sogreah. 1983. North-West region agricultural development project - B irrigated
agriculture - Main Report and Annexes. Somali Democratic Republic. Ministry of
Agriculture.
Tonelli, M.A. 1997. Complementi di telerilevamento. Luni ed. 118 pp.
62
Annexes
ANNEXES
63
Annexes
Annex 1 - From Land Cover to Land Use map
LAND
UTILIZATION SOCIO-ECONOMIC
LAND COVER SYNTHESIS TYPES
PHOTOINTERPRETATION & DATA
SELECTION
SPATIAL DATA LINKAGE FIELD SURVEY
TABULAR DATA
LAND USE
64
Annexes
Annex 2 Unverified Land Cover map of the Juba and Shabelle riverine area at 1:100 000 scale produced during
SWALIM Phase I
65
Annexes
Annex 3 Boundaries and mapping scale of the two areas of interest
66
Annexes
Annex 4 Boundaries of the LC map of the riverine area
67
Annexes
Annex 5
68
Annexes
Annex 6
69
Annexes
Annex 7
70
Annexes
Annex 8
71
Annexes
Annex 9 Land Cover Pictures from the NAOI
Picture 1: Wide River Bed with water (Photo by M. Shaie)
Picture 2: Dry riverbed (Photo by M. Shaie)
72
Annexes
Picture 3: Sparse Shrubs/Isolated Rainfed Crop Fields (Photo by M. Shaie)
Picture 4: Sparse Shrubs/Herbaceous Open/Clustered Rainfed Crop Fields (Photo by M. Shaie)
73
Annexes
Picture 5: Closed Shrubs (Photo by M. Shaie)
Picture 6: Open Shrubs (Photo by M. Shaie)
74
Annexes
Picture 7: Sparse Shrubs (Photo by M. Shaie)
Picture 8: Sparse Shrubs/Open Shrubs/Isolated Rainfed Crops (Photo by M. Shaie)
75
Annexes
Picture 9: Open Shrubs/Sparse Shrubs/Isolated Rainfed Crop Fields (Photo by M. Shaie)
Picture 10: Sparse Shrubs/Isolated Rainfed Crop Fields (Photo by M. Shaie)
76
Annexes
Picture 11: Sparse Shrubs (Photo by M. Shaie)
Picture 12: Riparian Open or Fragmented Forest with Irrigated Crop Fields (Photo by M. Downie)
77
Annexes
Picture 13: Acacia etbaica forest (photo by M. Downie)
Picture 14: Acacia bussei forest (photo by M. Downie)
78
Annexes
Picture 15: Closed Herbaceous (Photo by M. Downie)
Picture 16: Open Herbaceous (coastal plain with the sea in the background) (Photo by M. Shaie)
79
Annexes
Picture 17: Irrigated papaya fruit trees near Hargeisa (Photo by S. Oduori)
Picture 18: Urban area (Hargeisa) (Photo by R.Vargas)
80
Annexes
Picture 19: Bare Areas (notice the village in the middle of the photograph) (Photo by S. Oduori)
Picture20: Bare area/stony surface (Photo by R.Vargas)
81
Annexes
Annex 10 Land Cover Pictures from the SAOI
Picture 21: Grazing of Standing Crop Residue. Photo by Musse Shaie Alim FAO
SWALIM
Plate
Picture 22: Rainfed Agriculture. Photo by Musse Shaie Alim FAO SWALIM
82
Annexes
Picture 23: Irrigated Fields. Notice the fruit trees and the herbaceous crops. Photo by
Mohamed Farah FAO SWALIM.
83
Annexes
Picture 24: Rainfed Agriculture Sorghum, Harvested Sesame and Maize

Inmtercropped. Photo by Musse Shaie Alim FAO SWALIM.
Picture 25: Sand Dune. Photo by Mohamed Farah FAO SWALIM.
84
Annexes
Picture 26: Water Body. Photo by Mohamed Farah FAO SWALIM.
85
Annexes
Picture 27: Rural Settlement with Pond for Water (human and Livestock). Photo
by Musse Shaiei FAO SWALIM.
86
Annexes
Picture 29: Urban Area. Photo by Ibrahim Dagane Ali FAO SWALIM
Picture 30: Open Shrubs. Photo by Ibrahim Dagane Ali FAO SWALIM
87
Annexes
Picture 31: Dry River Bed. Photo by Ibrahim Dagane Ali FAO SWALIM
Picture 32: Closed Shrub. Photo by Ibrahim Dagane Ali FAO SWALIM
88
Annexes
Picture 33: Sparse Shrubs. Photo by Ibrahim Dagane Ali FAO SWALIM
89
Annexes
90
Annexes
91
Annexes
Annex 10 Somalia Land Cover legend
SOMALIA - FINAL LEGEND

LCC Own Description Map Code LCC Level
LCC Label
A12-NATURAL AND SEMINATURAL TERRESTRIAL VEGETATION

Herbaceous terrestrial
Closed to Open Herbaceous (>15%) Continuous Closed to Open
2HL A2 = Herbaceous Herbaceous Vegetation
A20 = Closed To Open 15% - 100%
B4 = 3 - 0.03 m
C1 = Continuous
Herbaceous (>1%) with Sparse Shrubs (1-15%) Closed to Open Herbaceous

2HL8 A2 = Herbaceous Vegetation with Shrubs
A10 = Closed >65%
B4 = 3 - 0.03 m
C1 = Continuous
F2 = 2nd layer
F6 = Shrubs
Z1 = Sparse 15-1%
G3 = 5 - 0.3 m
Closed to Open Herbaceous (>15%) with Sparse Trees and Shrubs Closed to Open Herbaceous
2HL78 A2 = Herbaceous Vegetation with Trees and Shrubs
A20 = Closed To Open 15% - 100%
B4 = 3 - 0.03 m
C1 = Continuous
F2 = 2nd layer
F5 = Trees
F10 = Sparse 15-5%
G2 = > 30 - 3 m
F2 = 3rd layer
F6 = Shrubs
F10 = Sparse 15-1%
G3 = 5 - 0.3 m
General Open Herbaceous (on ex-irrigated schemes) Continuous Open Herbaceous

2HP-x A2 = Herbaceous Vegetation
A11 = Open General 65-15%
B4 = 3 - 0.03 m
C1 = Continuous
Z6 = ex-irrigated schemes
Sparse Herbaceous Sparse Herbaceous Vegetation

2HR A2 = Herbaceous
A14 = Sparse General 15-1%
B4 = 3 - 0.03 m
Shrubs terrestrial
Closed Shrubs Continuous Closed Shrubland
2SC A4 = Shrubs (Thicket)
A10 = Closed >65%
B3 = 5 - 0.3 m
C1 = Continuous
General Open Shrubs Open Shrubs (Shrubland)

2SP A4 = Shrubs
B3 = 5 - 0.3 m
General Open Shrubs with Open Herbaceous Shrubland with Open Herbaceous
2SP6 A4 = Shrubs
B3 = 5 - 0.3 m
C1 = Continuous
F2 = 2nd layer
F4 = Herbaceous
F9 = Open General 65-15%
G4 = 3 - 0.03 m
General Open Shrubs with Herbaceous from Closed to Open (on ex- Medium To High Shrubland with
2SPJ6-x A4 = Shrubs Herbaceous
irrigated schemes)
B14 = 5 - 0.5 m
C1 = Continuous
F2 = 2nd layer
F4 = Herbaceous
F7 = Closed to Open
G4 = 3 - 0.03 m
Z6 = ex-irrigated schemes
General Open Shrubs with Trees 1-15% Shrubland with Emergents

2SP7 A4 = Shrubs
B3 = 5 - 0.3 m
C1 = Continuous
F2 = 2nd layer
92
Annexes
F5 = Trees
Z2 = Sparse 15-1%
G2 = > 30 - 3 m
Sparse Shrubs Sparse Shrubs

2SR A4 = Shrubs
B3 = 5 - 0.3 m
Sparse Shrubs with Herbaceous 1-15% Sparse Shrubs and Sparse

2SR6 A4 = Shrubs Herbaceous
B3 = 5 - 0.3 m
F2 = 2nd layer
F4 = Herbaceous
Z3 = Sparse 15-1%
G4 = 3 - 0.03 m
Woody terrestrial
Open Woody Continuous Open ((70-60) - 40%)
2WO A1 = Woody Woody Vegetation
A12 = Open 65-40%
B1 = 7 - 2 m
C1 = Continuous
Trees terrestrial
Closed Trees with Shrubs from Closed to Open Trees with Shrubs
2TC8 A3 = Trees
A10 = Closed >65%
B2 = >30 - 3 m
C1 = Continuous
F2 = 2nd layer
F6 = Shrubs
F7 = Closed to Open
G3 = 5 - 0.3 m
General Open Trees with Herbaceous 1-65% Woodland with Herbaceous Layer
2TP6 A3 = Trees
B2 = >30 - 3 m
F2 = 2nd layer
F4 = Herbaceous
Z4 = Closed to Open 1-65%
G4 = 3 - 0.03 m
General Open Trees with Shrubs 1-65% Woodland with Shrubs

2TP8 A3 = Trees
B2 = >30 - 3 m
F2 = 2nd layer
F6 = Shrubs
G3 = 5 - 0.3 m
STRIPED Very Open Trees With General Open Shrubs And Herbaceous Fragmented (Striped) Open (40 -
2TV86Zs A3 = Trees (20-10)%) Trees with Open Shrubs
(Tiger Bush) and Open Herbaceous
A13 = Very Open 40-15%
B2 = >30 - 3 m
C2 = Fragmented
C4 = Striped
F2 = 2nd layer
F6 = Shrubs
F9 = Open 65-15%
G3 = 5 - 0.3 m
F2 = 3rd layer
F4 = Herbaceous
F9 = Open 65-15%
G4 = 3 - 0.03 m
CELLULAR Very Open Trees With General Open Shrubs And Fragmented (Cellular) Open (40 -
2TV86Zc A3 = Trees (20-10)%) Trees with Open Shrubs
A13 = Very Open 40-15% and Open Herbaceous
B2 = >30 - 3 m
C2 = Fragmented
C4 = Cellular
F2 = 2nd layer
F6 = Shrubs
F9 = Open 65-15%
G3 = 5 - 0.3 m
F2 = 3rd layer
F4 = Herbaceous
F9 = Open 65-15%
G4 = 3 - 0.03 m
Sparse Trees with Shrubs 1-15% Sparse Trees With Sparse Shrubs
2TR8 A3 = Trees
A14 = Sparse 1- 15%
93
Annexes
B2 = >30 - 3 m
F2 = 2nd layer
F6 = Shrubs
Z1 = Sparse 15-1%
G3 = 5 - 0.3 m
A24-NATURAL AND SEMINATURAL AQUATIC VEGETATION
On Temporarily Flooded Areas (more than 2 but less than 4 months)
Closed Herbaceous On Temporarily Flooded Land Closed Medium To Tall Herbaceous
4HCJF A2 = Herbaceous Vegetation On Temporarily Flooded
A12 = Closed >65% Land
B15 = 3 - 0.3 m
C2 = < 4 months/y
General Open Herbaceous On Temporarily Flooded Land Open Medium To Tall Herbaceous
4HPJF A2 = Herbaceous Vegetation On Temporarily Flooded
A13 = Open General 65-15% Land
B15 = 3 - 0.3 m
C2 = < 4 months/y
General Open Woody On Temporarily Flooded Land Open Woody Vegetation On

4WPF A1 = Woody Temporarily Flooded Land
B1 = 7 - 2 m
C2 = < 4 months/y
Closed Trees with Shrubs from Closed to Open On Temporarily Flooded Trees With Shrubs On Temporarily
4TCF8 A3 = Trees Flooded Land
Land A12 = Closed >65%
B2 = >30 - 3 m
C2 = < 4 months/y
F2 = 2nd layer
F6 = Shrubs
F7 = Closed to Open
G3 = 5 - 0.3 m
Open Trees with Shrubs from Closed to Open On Temporarily Flooded Woodland With Shrubs On
4TOF8 A3 = Trees Temporarily Flooded Land
Land A14 = Open 65-40%
B2 = >30 - 3 m
C2 = < 4 months/y
F2 = 2nd layer
F6 = Shrubs
F7 = Closed to Open
G3 = 5 - 0.3 m
Herbaceous aquatic - On Permanently Flooded Areas

Closed Herbaceous On Permanently Flooded Land Closed Medium To Tall Herbaceous
4HCJFF A2 = Herbaceous Vegetation On Permanently Flooded
A12 = Closed >65% Land
B15 = 3 - 0.3 m
C1 = flooded > 4 months/y
A 11-CULTIVATED TERRESTRIAL AREAS AND MANAGED LANDS

Herbaceous Crop - Rainfed
Rainfed - Continuous Medium Fields of Herbaceous crops (single or Permanently Cropped Area With
HM47-C A3 = Herbaceous crop Herbaceous Crop(s)
multiple) - Main crop: cereals (Sorghum, Maize)
B4 = Medium (2-5 ha) Crop Type: Cereals
B5 = Continuous
D1 = Rainfed
A3B1B5XXD1D9-B4-S3 D9 = Permanent
S3 = Cereals
Rainfed - Clustered Medium Fields of Herbaceous crops (single or Scattered Clustered Field(s) Of
HM147-C A3 = Herbaceous crop Permanently Cropped Area With
multiple) - Main crop: cereals (Sorghum, Maize)
B4 = Medium (2-5 ha) Herbaceous Crop(s)
Crop Type: Cereals
B6 = Scattered - Clustered
D1 = Rainfed
D9 = Permanent
S3 = Cereals
Rainfed - Isolated Medium Fields of Herbaceous crops (single or Scattered Isolated Field(s) Of
HM247-C A3 = Herbaceous crop Permanently Cropped Area With
multiple) - Main crop: cereals (Sorghum, Maize) Herbaceous Crop(s)
B4 = Medium (2-5 ha)
Crop Type: Cereals
B7 = Scattered - Isolated
D1 = Rainfed
D9 = Permanent
S3 = Cereals
Rainfed - Continuous Small Fields of Herbaceous crops (single or Permanently Cropped Area With
HR47-C A3 = Herbaceous crop Small Sized Field(s) Of Herbaceous
multiple) - Main crop: cereals (Sorghum, Maize) Crop(s)
B2 = Small (less than 2ha)
Crop Type: Cereals
B5 = Continuous
D1 = Rainfed
D9 = Permanent
S3 = Cereals
Rainfed - Clustered Small Fields of Herbaceous crops (single or multiple) Scattered Clustered Permanently
HR147-C A3 = Herbaceous crop
94
Annexes
- Main crop: cereals (Sorghum, Maize) Cropped Area With Small Sized
B2 = Small (less than 2ha) Field(s) Of Herbaceous Crop(s)
B6 = Scattered - Clustered Crop Type: Cereals
D1 = Rainfed
D9 = Permanent
S3 = Cereals
Rainfed - Isolated Small Fields of Herbaceous crops (single or multiple) - Scattered Isolated Permanently
HR247-C A3 = Herbaceous crop Cropped Area With Small Sized
Main crop: cereals (Sorghum, Maize) Field(s) Of Herbaceous Crop(s)Crop
Type: Cereals
D1 = Rainfed
D9 = Permanent
S3 = Cereals
Herbaceous Crop - Irrigated

Irrigated Continuous Large Fields of Herbaceous crops (single - Permanently Cropped Area With
HL57 A3 = Herbaceous crop Irrigated Herbaceous Crop(s)
Rice/Banana - or multiple herbaceous crops)
B3 = Large (more than 5 ha)
B5 = Continuous
D3 = Irrigated
D9 = Permanent
Irrigated Continuous Medium Fields of Herbaceous crops (single - Permanently Cropped Area With
HM57 A3 = Herbaceous crop Irrigated Herbaceous Crop(s)
Rice/Banana - or multiple herbaceous crops)
B5 = Continuous
D3 = Irrigated
D9 = Permanent
Irrigated Clustered Medium Fields of Herbaceous crops (single - Scattered Clustered Field(s) Of
HM157 A3 = Herbaceous crop Permanently Cropped Area With
Rice/Banana - or multiple herbaceous crops) Irrigated Herbaceous Crop(s)
D3 = Irrigated
D9 = Permanent
Irrigated Continuous Small Fields of Herbaceous crops (single - Permanently Cropped Area With
HR57 A3 = Herbaceous crop Small Sized Field(s) Of Irrigated
Rice/Banana - or multiple herbaceous crops) Herbaceous Crop(s)
B5 = Continuous
D3 = Irrigated
D9 = Permanent
Irrigated Clustered Small Fields of Herbaceous crops (single - Scattered Clustered Permanently
HR157 A3 = Herbaceous crop Cropped Area With Small Sized
Rice/Banana - or multiple herbaceous crops) Field(s) Of Irrigated Herbaceous
Crop(s)
D3 = Irrigated
D9 = Permanent
Irrigated Isolated Small Fields of Herbaceous crops (single - Scattered Isolated Permanently
HR257 A3 = Herbaceous crop Cropped Area With Small Sized
Rice/Banana - or multiple herbaceous crops) Field(s) Of Irrigated Herbaceous
Crop(s)
D3 = Irrigated
D9 = Permanent
Trees Crop - Irrigated

Irrigated Continuous Small Fields of Tree crops (fruit trees) and Permanently Cropped Area With
TR3H57 A1 = Trees Small Sized Field(s) Of Irrigated
Herbaceous crop (pulses and vegetables) (with distinct patterns on the Tree Crop(s) (One Additional Crop) (
same field) B2 = Small (less than 2ha) Herbaceous Terrestrial Crop With
Simultaneous Period).
B5 = Continuous
C2 = Multiple crop
C3 = One additional crop
C7 = Herbaceous terrestrial
C17 = Simultaneously
D3 = Irrigated
D9 = Permanent
Irrigated Clustered Small Fields of Tree crops (fruit trees) and Permanently Cropped Area With
TR13H57 A1 = Trees Scattered Clustered Small Sized
Herbaceous crop (pulses and vegetables) (with distinct patterns on the Field(s) Of Irrigated Tree Crop(s)
same field) B2 = Small (less than 2ha) (One Additional Crop) (Herbaceous
Terrestrial Crop With Simultaneous
B6 = Scattered - Clustered Period) .
C2 = Multiple crop
D3 = Irrigated
D9 = Permanent
Irrigated Isolated Small Fields of Tree crops (fruit trees) and Herbaceous Permanently Cropped Area With
TR23H57 A1 = Trees Scattered Isolated Small Sized
crop (pulses and vegetables) (with distinct patterns on the same field)
B2 = Small (less than 2ha) Field(s) Of Irrigated Tree Crop(s)
(One Additional Crop) (Herbaceous
B7 = Scattered - Isolated Terrestrial Crop With Simultaneous
Period) .
C2 = Multiple crop
95
Annexes
D3 = Irrigated
D9 = Permanent
General Open Trees with Shrubs 1-65% OR Irrigated Continuous Small Woodland with Shrubs OR
2TP8//TR3H57 A3 = Trees Permanently Cropped Area With
Fields of Tree crops (fruit trees) and Herbaceous crop (pulses and
A11 = Open General 65-15% Small Sized Field(s) Of Irrigated
vegetables) (with distinct patterns on the same field) Tree Crop(s) (One Additional Crop) (
B2 = >30 - 3 m Herbaceous Terrestrial Crop With
Simultaneous Period).
F2 = 2nd layer
F6 = Shrubs
G3 = 5 - 0.3 m
OR
A1 = Trees
B5 = Continuous
C2 = Multiple crop
D3 = Irrigated
D9 = Permanent
Herbaceous Crop - Post Flooding
Post Flooding - Continuous Medium Fields of Multiple Herbaceous crops Post Flooding Cultivation Of
HM3HY A3 = Herbaceous crop Herbaceous Crop(s) (One Additional
B4 = Medium (2-5 ha) Crop) ( Herbaceous Terrestrial Crop
With Simultaneous Period) .
B5 = Continuous
C2 = Multiple Crop
C3 = 1 add. Crop
C7 = Herbaceous Terrestrial
D2 = Post flooding
Post Flooding - Isolated Medium Fields of Multiple Herbaceous crops Post Flooding Cultivation Of
HM23HY A3 = Herbaceous crop Scattered Isolated Field(s) Of
B4 = Medium (2-5 ha) Herbaceous Crop(s) (One Additional
Crop) ( Herbaceous Terrestrial Crop
B7 = Scattered - Isolated With Simultaneous Period) .
C2 = Multiple Crop
C3 = 1 add. Crop
D2 = Post flooding
Post Flooding - Continuous Small Fields of Multiple Herbaceous crops Post Flooding Cultivation Of Small
HR3HY A3 = Herbaceous crop Sized Field(s) Of Herbaceous
B2 = Small (less than 2ha) Crop(s) (One Additional Crop) (
Herbaceous Terrestrial Crop With
B5 = Continuous Simultaneous Period) .
C2 = Multiple Crop
C3 = 1 add. Crop
D2 = Post flooding
Post Flooding - Clustered Small Fields of Multiple Herbaceous crops Post Flooding Cultivation Of
HR13HY A3 = Herbaceous crop Scattered Clustered Small Sized
(One Additional Crop) ( Herbaceous
B6 = Scattered - Clustered Terrestrial Crop With Simultaneous
Period) .
C2 = Multiple Crop
C3 = 1 add. Crop
D2 = Post flooding
Post Flooding - Isolated Small Fields of Multiple Herbaceous crops Post Flooding Cultivation Of
HR23HY A3 = Herbaceous crop Scattered Isolated Small Sized
(One Additional Crop) ( Herbaceous
B7 = Scattered - Isolated Terrestrial Crop With Simultaneous
Period) .
C2 = Multiple Crop
C3 = 1 add. Crop
D2 = Post flooding
Rainfed - Continuous Medium Fields of Herbaceous crops (single or Permanently Cropped Area With
HR47-C//HR3HY A3 = Herbaceous crop Small Sized Field(s) Of Herbaceous
multiple) - Main crop: cereals (Sorghum, Maize) OR Post Flooding -
B2 = Small (less than 2ha) Crop(s)
Continuous Small Fields of Multiple Herbaceous crops Crop Type: Cereals OR Post
B5 = Continuous Flooding Cultivation Of Small Sized
Field(s) Of Herbaceous Crop(s)
D1 = Rainfed (One Additional Crop) ( Herbaceous
D9 = Permanent Terrestrial Crop With Simultaneous
Period)
S3 = Cereals
OR
A3 = Herbaceous crop
B5 = Continuous
C2 = Multiple Crop
96
Annexes
C3 = 1 add. Crop
D2 = Post flooding
Rainfed - Isolated Small Fields of Herbaceous crops (single or multiple) - Scattered Isolated Permanently
HR247-C//HR23HY A3 = Herbaceous crop Cropped Area With Small Sized
Main crop: cereals (Sorghum, Maize) OR Post Flooding - Isolated Small Field(s) Of Herbaceous Crop(s)
Fields of Multiple Herbaceous crops Crop Type: Cereals OR Post
B7 = Scattered - Isolated Flooding Cultivation Of Scattered
Isolated Small Sized Field(s) Of
D1 = Rainfed Herbaceous Crop(s) (One Additional
D9 = Permanent Crop) ( Herbaceous Terrestrial Crop
With Simultaneous Period)
S3 = Cereals
OR
A3 = Herbaceous crop
C2 = Multiple Crop
C3 = 1 add. Crop
D2 = Post flooding
B15-ARTIFICIAL SURFACES AND ASSOCIATED AREAS

Airport Industrial And/Or Other Area(s)Built-
5A A4 = Non linear Up Object: Airport
A12 = Industrial and other
A21 = Airport
Urban Areas/Settlements Urban Area(s)

5U A4 = Non linear
A13 = Urban area
B16-BARE AREAS
Loose and Shifting Sands Loose And Shifting Sands
6L A6 = Loose and shifting sands
Dunes 6LD A6 = Loose and shifting sands Shifting Sands / Dune(s)

B1 = Dunes
Salt Crust (Sandy area) Loose And Shifting Sands With Salt
6LZ A6 = Loose and shifting sands Flats
B13 = Salt Flats
Bare Soil 6S A5 = Bare soil a/o other unconsol. mat. Bare Soil And/Or Other Unconsolid.
Material(s)
Bare Soil with Scattered Vegetation 6SV A5 = Bare soil a/o other unconsol. mat. Bare Soil And/Or Other
Unconsolidated Material(s)
U1 = Scattered vegetation (> 4%) Scattered Vegetation Present
B27-ARTIFICIAL WATERBODIES
Artificial Waterbodies Standing Artificial Perennial Waterbodies
7WP A1 = Artificial Waterbodies (Standing)
A5 = Standing
B1 = Perennial
B28-INLAND WATERBODIES
Tidal Area A1 = Inland Water Tidal Area (Surface Aspect: Sand)
8WT1
B3 = Tidal Area
B6 = Sand
River Perennial Natural Waterbodies

8WFP A1 = Natural Waterbodies (Flowing)
A4 = Flowing
B1 = Perennial
Seasonal Rivers Non-Perennial Natural Waterbodies

8WFN A1 = Natural Waterbodies (Flowing)
A4 = Flowing
B2 = Non-Perennial
Seasonal Rivers with Scattered Vegetation Non-Perennial Natural Waterbodies

8WFNV A1 = Natural Waterbodies (Flowing) Scattered Vegetation
A4 = Flowing Present
B2 = Non-Perennial
U1 = Scattered vegetation
Natural Waterbody Natural Waterbodies (Standing)

8W A1 = Natural Waterbodies
A5 = Standing
97

L-03 Land Cover of Selected Areas in Somaliland and Southern Somalia

Uploaded by

Copyright:

Available Formats

L-03 Land Cover of Selected Areas in Somaliland and Southern Somalia

Uploaded by

Document Information

Original Description:

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

L-03 Land Cover of Selected Areas in Somaliland and Southern Somalia

Uploaded by

Copyright:

Available Formats

The designations employed and the presentation of material in this information product

This document should be cited as follows:

Annex 9 Land Cover Pictures from the NAOI................................................................................................. 72

Figure 1: Rainfall and potential Evapotranspiration .................................................... 8

Chart 1: Distribution of main aggregations of Land Cover types ................................ 41

FAO Food and Agriculture Organisation of the United Nations

In order to predict behaviour of land in response to specific patterns of utilisation, data

1.1. Delineation of the study area

1.1.1 The Northern AOI

- sub-catchments extracted from the DEM;

Following morphological criteria, the following areas were selected:

The limits of the northern AOI are:

- the coast in the north;

1.1.2 The Southern AOI

- NASA 90 meters SRTM-DEM;

Delineation took into account the following areas:

- watershed limits extracted from the NASA SRTM DEM;

- consistent stratification of the land into homogenous areas, providing boundaries

2) Base map production:

3) Field data collection:

4) Legend updating and Map revision:

A detailed description of each phase is provided in Chapter 4 - Material and Methods.

THE OBJECTIVE OF THE STUDY

The primary factor limiting development of vegetation is rainfall. As a consequence of the

2.2 General features of the Northern Area of Interest (NAOI)

Rainfall Potential Evapotranspiration Graph at Gebilley

Figure 1: Rainfall and potential evapotranspiration

Figure 2: Relative humidity of the study area

Figure 3: Rainfall distribution and climate classification of the study area

2.2.3 Landform and Soils

2.2.4 Land cover

- Grass, perennial and sub-shrubs steppe (Steppa graminosa, perennierbosa e

Savanna - Composed of annual grasses (usually xerophytes). Cyperaceae and

2.2.5 Land Use

Rainfed agriculture is found in what is considered as the sorghum belt of Somaliland,

Cultivation of irrigated orchards is a cash-oriented activity in the area, involving the

District Estimated population

2.3 Southern Area of Interest (SAOI)

Belet Weyne Jowhar

Figure 5: Mean monthly rainfall patterns in the study area (1963-2001)

Three broad climatic zones may be recognised, characterised by differences in rainfall

Because of the predominance of alluvium, many soils consist of layers of deposited

2.3.4 Land Cover

- Grass, perennial and sub-shrubs steppe (Steppa graminosa, perennierbosa e

Savanna - Composed of annual grasses (usually xerophytes). Cyperaceae and

Riparian formations - These formations develop along temporary or permanent streams

2.3.5 Land Use

3 3. MATERIALS AND METHODS

3.1.1 Bibliographic data

3.1.2 Ancillary data

- Landsat 7 Enhanced Thematic Mapper 2000-2001-2002 (Annex 6 and Annex 7)

3.2.1 Preparatory phase

Preliminary overview of the AOI

Available data (included geological maps and data on geomorphology) were

Remote sensing data acquisition

- Landsat 7 Enhanced Thematic Mapper 2000-2001-2002

- Aster VNIR and SVNIR 2002 2005

The following processes were performed:

Landsat Aster Ikonos