Hydrological processes and water security in a changing world

Proc. IAHS, 383, 69–78, 2020

https://doi.org/10.5194/piahs-383-69-2020 Open Access
© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

Extreme Rainfall and Temperature Changes Assessment

on late-twenty-first-century over the Mono River Basin,
Togo (West-Africa)
Batablinlè Lamboni1 , Agnidé Emmanuel Lawin2 , Celestin Manirakiza1 , and Dagbégnon Luc Olabodé1
1 Institute of Mathematics and Physical Sciences, University of Abomey-Calavi, BP 613 Porto-Novo, Benin
2 Applied Hydrology Laboratory, University of Abomey-Calavi, BP 4521 Cotonou, Benin
Correspondence: Batablinlè Lamboni (lbata1982@yahoo.fr)

Published: 16 September 2020

Abstract. This study assessed the extreme rainfall and temperature changes over Mono river basin by the end of
21st century under the highest greenhouse gas emission scenario RCP8.5. Simulations of eight regional climate
models (RCMs) provided by Africa-CORDEX program were considered. The maximum and minimum temper-
atures and rainfall were analyzed. Two groups of indices were analyzed. The first group consists of frequency
indices: maximum number of consecutive dry days and wet days. The second group is intensity indices: five-
day maximum rainfall and simple daily intensity index. These variables were calculated at annual and seasonal
scales. Changes from the baseline period 1971–2000 were computed for far future 2071–2100. As result, almost
all the RCMs considered predicted a decrease of the rainfall and increase of maximum and minimum temper-
atures over most parts of the Mono basin, particularly in the south. Declining mean monthly precipitation and
irregular rainy seasons at all scales were exhibited by most of the models. Simple daily intensity and five-day
maximum rainfall are projected to decrease by the majority of the used models. As for dry and wet sequences,
the RCMs showed an increase of the consecutive dry days and a decrease of wet days.

1 Introduction al., 2017.). Current and future trends suggested that regions:
West Africa, including the Sahara, the Horn of Africa, part
A global view of climate change impacts is important to of Central Africa and southern hemispheres, could undergo a
quantify the aggregated effects in a given region since large warming of order of 3 to 6 ◦ C by 2100. Pluviometric regimes
scale processes can influence driving forces at sub regional will also be affected and may decrease of more than 20 %
or mesoscale. Indeed, the detection of the planetary warming to 30 % compared to the reference 1961–1990 (Bigot, 2003).
trend, attributed to anthropogenic activities, have been pre- Other studies indicated an increase in extreme rainfall events,
sented with greater certainty in the 4th report of the IPCC. particularly in Sahel (Nicholson and Webster, 2007) but also
According to this report, the average global temperature will a decrease in Nigeria (Oguntundé et al., 2011) and (East-
increase and could reach 4.5 ◦ C by 2100, the precipitation erling et al., 2000), Guinea Conakry (Aguilar et al., 2009),
will change and the extreme events will be increasingly re- in eastern Niger (Ozer et al., 2009), in Ivory-Coast (Soro
curring. Several studies (Aguilar et al., 2013; IPCC, 2013; et al., 2014), in Benin (Yèkambèssoun et al., 2017) and in
Masson et al., 1999) showed that impacts of climate extremes South Africa (Mason et al., 1999). In the northern part of
are mostly depending on the spatial scale considered and Africa, there is a tendency towards wetter conditions; in con-
vary from a region to another, from country to country and trast with the eastern part, experienced with more drying
even from one local area to another within the same coun- trends, although, these trends are of low significance. The un-
try. For Africa, many previous studies have investigated ex- derstanding and characterizing climate variability at reduced
treme events and have showed that changes in extreme tem- scales are so fundamental to understanding the impacts of cli-
perature and precipitation are contrasting (Yèkambèssoun et

Published by Copernicus Publications on behalf of the International Association of Hydrological Sciences.

70 B. Lamboni et al.: Extreme Rainfall and Temperature Changes Assessment

mate change, local vulnerability and for future development Mono basin is located in the Gulf of Guinea region,
projects (Emmanuel et al., 2019). more precisely Between 06◦ 160 and 09◦ 200 N and 0◦ 420 and
For the specific case of Togo, according to the report of 20◦ 250 E. The Mono basin is almost entirely in Togo although
republic of Togo (MERF, 2009), floods, drought, late rains, its lower course is between Togo and Benin. Our study con-
high temperatures and high winds, are the major risks in the sidered the part in Togo (Fig. 1). It houses a dam of hydro-
country. The drought covers the entire territory except for the electric power plant called Nangbéto. Elongated, covering
coastal region which is however facing sea level rise risk. It 25 400 km2 , the Mono basin is the major river of Togo. In
should be noted that among the latest climatic events, flood- general, West Africa’s climate is controlled by the interaction
ing takes precedence over other risks with its share of mate- of two air masses; the inuence of which varies throughout the
rial damage and loss of life. It is becoming more and more year with the north-south movement of the Intertropical Con-
dangerous to the whole country. Activities which are most vergence Zone (ITCZ). Hot and dry continental air masses
exposed to these risks are farming, livestock farming, mar- originating from the high pressure system above the Sahara
keting of agricultural products and market gardening (MERF, desert give rise to dusty Harmattan winds over most of West
2009). Therefore, it compels scientists to conduct studies on Africa from November to February. In summer, moist equa-
climate change projection in Togo in order to take adapta- torial air masses originating over the Atlantic Ocean bring
tion measures to enhance these communities in their devel- annual monsoon rains. Within this West African context,
opment. However, studies of the future impact of climate rainfall in the study area is characterized by two types of
change on precipitation and temperature are very few. Nelson rainfall regimes. In southern basin there are two rainy sea-
et al. (2010) assessed impacts of precipitation and tempera- sons which extend from mid-March to mid-July and from
ture changes on agricultural productivity by 2050 in Togo. mid-August to October. In northern basin, there is one rainy
This study used global climate models (GCMs) with spatial season which extends from April to October. Indeed, accord-
resolution of 200 to 300 km that cannot give details at less ing to the report of PANA-Togo (MERF, 2009), the country is
scales. So these models are only important when it comes to inuenced by two large climate systems: The tropical Souda-
representing climate on a very large scale of space. But at nian system in the north with a wet season running from May
regional level, they present many limitations (Hulme et al., to October and a dry season from November to April. In this
2001). In order to fill this resolution issue, Regional Climate area pluviometry varies from 900 to 1100 mm per year with
Models (RCMs) from CORDEX (COordinated Regional Cli- an average temperature of 28 ◦ C; the tropical Guinean sys-
mate Downscaling EXperiment) are now proposed in our tem in the south is characterized by two dry seasons and two
study because of their high spatial resolution (10 to 100 km). wet seasons of unequal duration with pluviometry located be-
Furthermore Batablinlè et al. (2018) showed, at the Mono tween 1000 and 1600 mm yr−1 and an average temperature of
river basin scale, there was very few study focuses on future 27 ◦ C. According to the report of WAEMU in 2006, the pop-
evolution of the extreme temperatures, number of extremely ulation of the basin is approximately two million, with an
heavy rainfall days, consecutive dry days, consecutive wet annual increase of 2.9 %. This population also distributed at
days, five-day maximum rainfall, simple daily intensity in- high densities in the south of the basin has as main activities,
dex. We have therefore found it necessary to evaluate the agriculture (mainly rainfed); in the lower valleys, fishing and
future changes of climate indices using the World Meteo- salt-farming are the major activities (Ernest et al., 2015).
rological Organization (WMO) reference period 1971–2000
computed from Regional Climate Models (RCMs). In this 2.2 Data and Sources
study, rain and temperature simulated by 8 RCMs from the
Africa-CORDEX program have been analyzed through max- The observed daily rainfall and temperature data used in
imum and minimum temperatures, average rainfall and ex- this study, are provided by the Togo National Meteorology
treme precipitation events by 2100 on the Mono river basin Agency (Météo Togo) for the period 1970–2000 for eight
which is the major economical part of Togo in terms of agri- stations spatially located as shown in Fig. 1 and comple-
cultural and energy outputs. The objective of this work is mented by other sources such as the archives of ORTM Togo
therefore to provide knowledge and more details about ex- and the Nangbeto dam. The missing data values were re-
treme rainfall and temperature for the population of Mono placed statistically by similar day averages for the previous
river basin. years where there were existing data values. Table 1 shows
the Cordex models used. In order to develop future climate
change projections on Mono river basin, the most recent Re-
2 Data and the methods gional Climate Models (RCMs) provided by the COordi-
nated Regional Downscaling EXperiment (CORDEX) Africa
2.1 Study area were analyzed. These models were run using different emis-
sion scenarios, so-called Representative Concentration Path-
This study area has already been described by certain authors ways (RCPs) which provide different assumptions of green-
(Batablinlè et al., 2018; Lamboni et al., 2019b). house gas emission pathways. One of RCPs was selected

Proc. IAHS, 383, 69–78, 2020 https://doi.org/10.5194/piahs-383-69-2020

B. Lamboni et al.: Extreme Rainfall and Temperature Changes Assessment 71

of the following steps: (1) aggregate finer-scale observed data

to a RCM-specific scale; (2) correct the RCM’s systematic
biases through quantile mapping; and (3) disaggregate bias-
corrected outputs to a finer scale (Batablinlè et al., 2018). An
equiratio cumulative distribution function matching (equira-
tio CDFm): The statistical transformation can be written as:


h1 Xm−p
Xm−padjst = Xm−p
(1)
h2 Xm−p
and



Xm−padjst = Xm−p + h1 Xm−p − h2 Xm−p (2)

with (Eq. 1) for precipitation and (Eq. 2) for temperature,

where Fm−p is the cumulative distribution function (CDF) of
−1
the model for a future projection period, and Fo−c and
−1
Fm−c are quantile functions for observations and model in
the reference period, respectively. h1 and h2 were defined by:
−1
Figure 1. The geographical location (sources Batablinlè et al., h1 = Fo−c Fm−p (3)
2018).
and
−1
h2 = Fm−c Fm−p (4)
in this study to cover a wide high (RCP8.5) global anthro-
pogenic radiative forcing. We made the choice to download A complete description of this method is described by
the values of several climate variables of eight models avail- Gudmundsson et al. (2012).
able in a spatial resolution 0.44◦ or 50 km. It has already been Before this procedure, statistical measures such as the rel-
used over Mono basin by certain authors as (Batablinlè et al., ative error or mean absolute error were used to describe and
2018; Emmanuel et al., 2019; Lamboni et al., 2019a). compare the observed and simulated variables. The relative
error (EP and ET ), respectively for precipitations and tem-
2.3 Methods peratures of the estimated values of the rainfall may be cal-
culated from the following Eqs. (1) and (2):
The seasonal averages as well as the annual averages were
calculated using the Climate Data Operator Commands (Bat- (Xo − Xs )
EP = (5)
ablinlè et al., 2018; Emmanuel et al., 2019). Then, the free Xo
software R is used to compute all the statistical parameters ET = (Xo − Xs ) (6)
and all the plots presented in the following. In this study,
four climatic indices associated to the precipitations and tem- and finally, the changes from the reference period are as-
peratures were considered as potentially impacting on the sessed as shown by Eqs. (7) and (8) respectively for precipi-
availability of water resources, population and agricultural tations and indices (1XP,I ) and temperatures (1XT )
productivity in this zone (Emmanuel et al., 2019). These in- 
FF − HIST

dices, computed with Climate Data Operator Commands, are 1XP,I = × 100 (7)
HIST
defined in Table 2 as follow. To bias-correct the model fu-
ture projection, the cumulative distribution function match- or
ing (CDFm) of the model in the reference period is used to
1XT = FF − HIST (8)
identify the corresponding percentile values of the future pe-
riod (Batablinlè et al., 2018). Later the observation CDFm where FUT and HIST represent respectively for the future
is used to find the climate variable value for the same future (by 2100) and the historical or reference period (1971–2000).
cumulative probability, which represents the bias corrected
future value. The drawback of the above procedure of bias
3 Results and discussions
correction is the assumption of constant bias between the ref-
erence and future periods . However, we will accept this as-
3.1 Results
sumption as most of the recent studies did (Sarr and Camara,
2017) as mentioned earlier, bias correction and spatial disag- Before assessing the climate change scenario we evaluated
gregation relies on finer-scale observation data and consists the considered models’ performances to represent the climate

https://doi.org/10.5194/piahs-383-69-2020 Proc. IAHS, 383, 69–78, 2020

72 B. Lamboni et al.: Extreme Rainfall and Temperature Changes Assessment

Table 1. Used climatic models, global model under which they are run (column 1), their Institute of origin (column 2), their short name
(column 3) and RCM model Name (column 4).

Global Model Name Institute ID Model Short Name RCM Model Name
GFDL-ESM2M NOAA-GFDL NOAA SMHI-RCA4
NorESM1-M NCC NCC SMHI-RCA4
MPI-ESM-LR MPI-M MPI SMHI-RCA4
MIROC5 MIROC MIROC SMHI-RCA4
IPSL-CMA5-MR IPSL IPSL SMHI-RCA4
EC-EARTH ICHEC ICHEC KNMI-RCAMO22T
CNRM-CM5 CRNM-CERFACS CNRM SMHI-RCA4
CanESM2 CCCma CCCMA SMHI-RCA4

Table 2. Selected indices and their signification.

ID Name Definition Units

Frequency indices (adapted from WMO 2009)
CDD Consecutive dry days Maximum number of consecutive days d
with PRCP∗ <1 mm
CWD Consecutive wet days Maximum number of consecutive days d
with PRCP∗ >1 mm
Intensity indices (adapted from WMO 2009)
RX5day Max5-day precipitation Maximum consecutive 5 d precipitation Mm
amount Annual
SDII Simple daily intensity Index Annual total precipitation divided by mm d−1
the number of wet day (defined as
PRCP∗ >1 mm) in the year
∗ PRCP – daily total precipitation

of the study region. Table 3 showed that the spatial mean imum temperature, all the models predicted an increase for
rainfall MAE (Mean absolute error) of the selected RCMs are the whole basin except model canESM2 which showed a
weak and not greater than 3 mm d−1 in mean during the main slight decrease for a small part of the southern basin. The
part of the rainy season which is June to September (the agri- lowest increase is 0.42 ◦ C and is projected by the model
cultural period). NOAA model is the most accurate model CNRM-CM5. The highest increase is 3.3 ◦ C and is projected
while MIROC model showed the biggest MRE (Mean rela- by the rest of models. The north of the basin is forecasted
tive error). The higher bias between simulated and observed to experience the greatest increase according to almost all
data, may be caused by errors in measuring data and errors the models. The projections exhibited clearly an increase
associated with the model. The statistical analysis show that in days of high heat in the north as shown by the mod-
there is no enough an agreement between simulated and ob- els CNRM-CM5 (1), GFDL-ESM2M (6) , EC-EARTH (3),
served variables which such as mean temperatures (minimum IPSL-CMA5-MR (4), MPI-ESM-LR (6), NorESM1-M (7)
and maximum) and precipitations (Pm ). Obviously, current and Models-Mean (9). However, for models CCCMA, the
RCMs still have substantial deficiencies in simulating the re- maximum temperature will be relatively cooler on some area
gional climate over Mono. To produce the reliable projec- in the southward of the basin (mean value −0.29 ◦ C) and a
tions required for regional climate assessment, we need to strong increase on the rest of this area. In the north and the
bias-correct the raw RCM outputs. center, the maximum temperature will increase in average by
1 ◦ C for model CNRM-CM5 (1) and model MPI-ESM-LR
(5), 0.22 ◦ C for model GFDL-ESM2M (6), 1.45 ◦ C for mod-
3.1.1 Changes of Minimum and Maximum temperature
els EC- EARTH (3), IPSL-CMA5-MR (4) and NorESM1-
Predicted changes of maximum temperature and minimum M (7). Globally, the Models-Mean (9) predicted 1 ◦ C as rate
temperature between the far future (2071–2100) and the ref- of increase.
erence period (1971–2000) are shown by Fig. 2. For the max-

Proc. IAHS, 383, 69–78, 2020 https://doi.org/10.5194/piahs-383-69-2020

B. Lamboni et al.: Extreme Rainfall and Temperature Changes Assessment 73

Table 3. Spatial mean Rainfall MAE (in mm) and MRE (in %) between observations and bias corrected simulations over Mono river basin.

Main part of the rainy season

Models Biases J F M A M J J A S O N D
NOAA MAE 0 0 1.1 1.9 3.7 0.8 1.3 1.6 3.7 1.9 0.3 0
MRE 0 7 6 33 29 5 7 8 18 15 5 0
NCC MAE 1.6 1.4 1.1 3.7 3.4 2.3 2.6 2.2 3.9 0.8 0.9 1.3
MRE 6 9 6 51 34 14 14 10 15 12 37 6
MPI MAE 0.4 1.0 1.1 0.3 1.2 2.5 3.8 2.7 2.6 1.1 2.1 1.3
MRE 4 12 7 4 11 16 20 12 10 5 11 3
MIROC MAE 1.2 2.2 1.3 0.5 0.5 3.1 2.1 3.2 4.5 1.4 0 1.1
MRE 6 6 15 15 47 20 12 12 16 6 0 2
IPSL MAE 1.7 1.7 1.2 3.0 4.8 3.9 3.3 1.0 1.3 0.8 0.7 1.8
MRE 8 8 4 32 37 24 22 35 17 4 3 2
ICHEC MAE 1.8 0.8 1.3 2.7 3.8 4.0 0.5 1.3 1.8 2.2 0.9 3.1
MRE 9 3 4 28 31 26 31 27 19 11 15 4
CNRM MAE 2.1 0.2 1.4 2.0 2.9 2.2 1.3 2.0 3.3 3.2 0.4 0.5
MRE 2 26 5 33 28 13 18 22 15 13 11 6
CCCMA MAE 1.2 0.8 1.3 3.5 3.6 3.1 2.5 1.4 1.2 0.3 1.8 0.1
MRE 4 2 3 34 25 16 28 12 9 14 8 22
Models Mean MAE 0 0.1 0 1.9 1.3 2.8 3.1 2.5 2.3 2.7 0.2 0.3
MRE 0 9 0 11 13 8 9 18 10 7 2 3

The analysis of projected minimum temperature for far 3.1.2 Precipitation frequency indices
future revealed that there is a relative uniform evolution
over the entire study period for all models except models We have analyzed also the maximum duration of the number
CNRM-CM5 (1) and EC-EARTH (3). Indeed, in the north- of consecutive wet days (CWD) and the maximum number
ward, the minimum temperature were projected and will of consecutive dry days (CDD) in Fig. 3. The changes relat-
increase for (mean value 1.55 ◦ C), (mean value 1.45 ◦ C), ing to the maximum duration of the number of consecutive
(mean value 2 ◦ C), (mean value 1.4 ◦ C), (mean value 2 ◦ C), wet days (CWD) between the far future and the historical
(mean value 2.4 ◦ C) and (mean value 1.2 ◦ C), respectively, period are shown in Fig. 3. The lowest decrease is −40 %
for models MIROC5 (2), IPSL-CMA5-MR (4), MPI-ESM- and is projected by the model (M8) CanESM2. The high-
LR (5), GFDL-ESM2M (6), NorESM1-M (7), canESM2 est increase is 32 % and is projected by model (M5) MPI-
(8) and Models-Mean (9) . For the models CNRM-CM5 ESM- LR. Indeed, the projection models used showed a vari-
(1) and EC- EARTH (3), the results exhibit an increas- ability of the consecutive wet days on the northern part of
ing average for the minimum temperature in the north. the basin from 0 % to −14.5 %, −15 % to −21 %, −10 %
When considering the southward and the center, models to −20 %, −5 % to 15 %, −10 % to 11 %, −10 % to 17 %,
used predict an increasing average for, (mean value 2.5 ◦ C), −10 % to 5 %, −10 % to 20 % and −10 % to −15 % re-
(mean value 0.51 ◦ C), (mean value 1.82 ◦ C), (mean value spectively for models CNRM-CM5 (M1), MIROC5 (M2),
2 ◦ C), (mean value 1.22 ◦ C), (mean value 1.55 ◦ C) and EC-EARTH (M3), IPSL-CMA5-MR (M4), MPI-ESM- LR
(mean value 1.83 ◦ C), respectively, for models MIROC5 (2), (M5), GFDL-ESM2M (M6), NorESM1-M(M7), CanESM2
IPSL-CMA5-MR (4), MPI-ESM-LR (5), GFDL-ESM2M (M8) and Mean-models (M9). The southward and center (ex-
(6), NorESM1-M (7), canESM2 (8) and Models-Mean (9). cept models MPI-ESM-LR (M5) and GFDL-ESM2M (M6)),
While for models CNRM-CM5 (1) and (C) EC-EARTH (3), there is a decreasing average in general for the maximum du-
the results show a relative rise of the minimum temperature ration of the number of consecutive wet days (CWD), from
but more than 1 ◦ C. 0 % to −15 %, 0 % to −10 %, −5 % to −15 %, −15 % to 5 %,
0 % to −10 %, −10 % to −30 % and −10 % to −5 % respec-
tively for models CNRM-CM5 (M1), MIROC5 (M2), EC-
EARTH (M3), IPSL-CMA5-MR (M4), NorESM1-M (M7),
CanESM2 (M8) and Mean-models (M9). In summary, sev-

https://doi.org/10.5194/piahs-383-69-2020 Proc. IAHS, 383, 69–78, 2020

74 B. Lamboni et al.: Extreme Rainfall and Temperature Changes Assessment

Figure 2. Maximum and minimum temperature changes (◦ C) Figure 3. Mean consecutive wet days (CWD) and Mean consecu-
between far future (2071–2100) and historical period (1971– tive dry days (CDD) changes between far future (2071–2100) and
2000) during May to September for models: (1) CNRM-CM5, (2) historical period (1971–2000) during May to September for models:
MIROC5, (3) EC-EARTH, (4) IPSL-CMA5-MR, (5) MPI-ESM- (M1) CNRM-CM5, (M2) MIROC5, (M3) EC-EARTH, (M4) IPSL-
LR, (6) GFDL-ESM2M, (7) NorESM1-M, (8) CanESM2 and (9) CMA5-MR, (M5) MPI-ESM-LR, (M6) GFDL-ESM2M, (M7)
Models-Mean (ensemble mean of Model) in the North (N), Center NorESM1-M, (M8) CanESM2 and (M9) Models-Mean (ensemble
(C), South (S) and over mono basin (B). mean of Model).

eral models used have announced clearly a decrease the (M5), GFDL-ESM2M (M6), NorESM1-M (M7), CanESM2
northward of the maximum duration of the number of con- (M8) and Mean-models (M9). The RCMs and their mean
secutive wet days (CWD) and over the south and center of in general showed that the maximum number of consecutive
the Mono basin. To quantify the potential impact of climate rainy days (CWD) will decrease over the Mono basin while
change on future drought events, the change in the number of the maximum duration of dry sequences will go up.
dry periods was analyzed. A dry period is defined as at least Other important aspects of daily precipitation characteris-
five consecutive days for which the daily precipitation is less tics for evaluating climate variability and change impacts is
than 1 mm. The relative deviations of the maximum duration the Max 5 d precipitation and the simple daily intensity in-
of dry sequences between the far future and the historical dex (intensity indices) within the rainy season. The section
period are represented in the Fig. 3. Contrary to the case of following is reserved to the analysis of the intensity indices.
the consecutive wet days (in the same figure), the projection
models used show a variability of the consecutive dry days 3.1.3 Precipitation intensity indices
on the northern part of the basin from −5 % to 5 %, 20 %
to 50 %, 10 % to 40 %, 4 % to 18 %, 15 % to 25 %, −5 % to The relative differences in the number of rainy days during
10 %, −10 % to 20 %, −20 % to −10 % and 10 % to 25 %, the rainy season (May–June–July–August–September) be-
respectively, for models CNRM-CM5 (M1), MIROC5 (M2), tween the far future and historical period are shown in the
EC-EARTH (M3), IPSL-CMA5-MR (M4), MPI-ESM-LR Fig. 4. The model MPI-ESM-LR (M5) indicated lowest de-
(M5), GFDL-ESM2M (M6), NorESM1-M (M7), CanESM2 crease (mean value −9 %) against the model NorESM1-M
(M8) and Mean-models (M9). The southward and center, (M7) which indicates an increase of the max 5 d precipitation
models project an increase of the maximum duration of the the northward (mean value 20 %). In general, the projection
number of consecutive dry days (CDD), from 10 % to 25 %, models used showed a variability of the max 5 d precipitation
40 % to 50 %, 30 % to 50 %, 6 % to 12 %, 5 % to 10 %, −4 % in the north of the basin varying between, −7.5 % and −5 %,
to 5 %, 20 % to 40 %, 20 % to 30 % and −5 % to 10 %, re- −4 % and 4 %, 5 % and 15 %, 5 % and 20 %, −30 % and
spectively, for models CNRM-CM5 (M1), MIROC5 (M2), 10 %, −5 % and 5 %, 0 % and 20 %, 10 % to 15 % and −10 %
EC-EARTH (M3), IPSL-CMA5-MR (M4), MPI-ESM-LR and 20 %, respectively, for models CNRM-CM5 (M1),

Proc. IAHS, 383, 69–78, 2020 https://doi.org/10.5194/piahs-383-69-2020

B. Lamboni et al.: Extreme Rainfall and Temperature Changes Assessment 75

MIROC5 (M2), EC-EARTH (M3), IPSL-CMA5-MR (M4),

MPI-ESM-LR (M5), GFDL-ESM2M (M6), NorESM1-M
(M7), CanESM2 (M8) and Mean-models (M9). The south-
ward models projected also a fluctuation of the max 5 d,
between −8.5 % and −7.5 %), −6 % and −2 %, −5 % and
0 %, −5 % and 0 %, −20 % and 10 %, −15 % and −5 %, 0 %
and 20 %, 10 % and 15 %, and −10 % and 20 % , respec-
tively, for models CNRM-CM5 (M1), MIROC5 (M2), EC-
EARTH (M3), IPSL-CMA5-MR (M4), MPI-ESM-LR (M5),
GFDL-ESM2M (M6), NorESM1-M (M7), CanESM2 (M8)
and Mean-models (M9). In the center and south, models
NorESM1-M (M7) and CanESM2 (M8) show clearly a de-
crease of the max 5 d, respectively, for, −5 % and −10 %.
The RCMs and their mean showed in general that the max 5 d
will decrease over the south and center of the Mono basin.
The relative differences in the intensity of rainy days be-
tween the far future and the historical period are shown in the
Fig. 4. For the results of the variability of the simple daily in-
tensity index, the models have shown dipole structures, as
the frequency indices : The projection models used show a
variability of the simple daily intensity index in the north of Figure 4. Max 5 d precipitation andsimple daily intensity in-
the basin varying between, −6 % and −1 %, −4 % and 3 %, dex changes (in %) between far future (2071–2100) and histor-
−2 % and 8 %, −16 % to 8 %, −10 % and 10 %, −12 % and ical period (1971–2000) during May to September for models:
4 %, −15 % and 5 %, −15 % and 5 % and −5 % and 0 % , re- (M1) CNRM-CM5, (M2) MIROC5, (M3) EC-EARTH, (M4) IPSL-
spectively, for models CNRM-CM5 (M1), MIROC5 (M2), CMA5-MR, (M5) MPI-ESM-LR, (M6) GFDL-ESM2M, (M7)
EC-EARTH (M3), IPSL-CMA5-MR (M4), MPI-ESM-LR NorESM1-M, (M8) CanESM2 and (M9) Models-Mean (ensemble
mean of Model).
(M5), GFDL-ESM2M (M6), NorESM1-M (M7), CanESM2
(M8) and Mean-models (M9). The southward models project
an increase and decrease of the simple daily intensity index
(Fig. 4), from −3 % to −1 %, 0 % to 3 %, −6 % to −2 %, be associated with a palpable increase in temperatures in the
−16 % to −10 %, −2.5 % to 5 %, −12 % to −3 %, −10 % Mono river watershed by 2100.
to 5 %, −10 % to 5 % and −15 % to −5 %, respectively,
for models CNRM-CM5 (M1), MIROC5 (M2), EC-EARTH 3.2 Discussion
(M3), IPSL-CMA5-MR (M4), MPI-ESM-LR (M5), GFDL-
ESM2M (M6), NorESM1-M (M7), CanESM2 (M8) and The analysis of extreme weather indices shows that all cli-
Mean-models (M9). The RCMs and their mean in general mate projections indicate an overall increase in extreme tem-
showed that the simple daily intensity index will decrease peratures in the Mono catchment. The variations of the rain-
over a big part of the Mono basin during the rainy season fall and their indices will depend strongly on emission sce-
(May–June–July–August–September) particularly the south- narios and models. Indeed, some models suggested an in-
ward and in the center. crease in these indices while others indicated a decrease. To
have better comprehension of results of in this paper it is im-
portant to look at the results of other regions and countries.
3.1.4 Analysis of inter-annual irregularity The large disparity observed in the evolution of precipitation
by 2100 could be due to the type of forcing models or the
The analysis of the results presented in Table 4 shows that convection pattern used in West Africa (Sylla et al., 2016).
K2 tends to decrease in two stations (Tchamba and Lomé) This divergence of models on climate projections on pre-
from 1971–2000, indicating a tendency towards regularity. cipitation in West Africa is still uncertain. Yèkambèssoun et
On the other hand, in the six (6) others, K2 varies by increas- al. (2017) estimated that the spatially inconsistent trends of
ing, showing a tendency of increasing the irregularity of the climate indices noticed are likely linked to the high hetero-
precipitations. For the period 2071–2100, excepted stations geneousness of the rainfall. Indeed, the Mono basin is char-
Alédjo, Anié, Soutouboua and Lomé, which will be marked acterized by two types of rainfall regimes. In southern basin,
by increases but less significant, the other resorts will have there are two rainy seasons which extend from mid-March
to increase to reach 9.41 at Sokodé 7.12 Tabligbo, 5.18 at to mid-July and from mid-August to October. In northern
Atakpamé and 5.52 at Tchamba. Declining mean precipi- basin, there is one rainy season which extends from April to
tation heights at all scales and irregular rainy seasons may October. For Nelson et al. (2010), in the future (specifically

https://doi.org/10.5194/piahs-383-69-2020 Proc. IAHS, 383, 69–78, 2020

76 B. Lamboni et al.: Extreme Rainfall and Temperature Changes Assessment

Table 4. Results obtained by calculation of the K2 coefficient of inter-annual rainfall irregularity in the form of table of 1971–2000 and
2071–2100 for the Tabligbo, Atakpamé, Anié, Soutouboua, Sokodé, Tchamba , Alédjo and Lomé stations.

Station 1971/1980 1981/1990 1991/2000 2071/2080 2081/2090 2091/2100

Atakpamé 1.09 2.92 2.51 5.18 2.19 3.54
Tabligbo 1.53 1.22 2.04 7.12 4.12 2.01
Anié 1.86 2.38 2.17 1.51 1.38 2.01
Lomé 2.23 4.96 4.11 6.51 1.85 2.12
Sokodé 3.04 7.12 9.41 8.56 2.85 3.52
Tchamba 3.08 4.74 3.50 5.36 5.52 3.52
Alédjo 1.48 2.14 1.19 1.95 1.98 4.9
Soutouboua 2.54 1.13 7.38 1.56 3.80 8.02

by 2050), precipitation may rise or fall on average depending other rainfall indicators appear to remain stable for the pe-
on which model we look to. Despite everything, in the Mono riod 1960–2000. Likewise rainfall during the months of June
basin, several models used agree on an increase in tempera- to September appears to have witnessed declining trends over
ture (maximum and minimum) and a decrease of precipita- 1961–1993 in Nigeria (Oguntundé et al., 2011; Easterling, et
tion on a large part of the basin. Nelson et al. (2010) using al., 2000).
models GCM (general circulation models), showed that, in The contribution of our study in the area is the use of sev-
the future (specifically by 2050) the temperature is expected eral variables (analyzed with several models) based exclu-
to increase between 1 and 2.5 ◦ C in Togo. Diallo et al. (2013) sively on precipitation and other parameters (e.g. tempera-
pointed out that the decrease in mean rainfall in this area of ture) to analyze the rainfall season patterns and determine
West Africa could probably be induced in part by a weak- the factors of their temporal variability. The results of this
ening of the moisture from local sources, slowing down the study could enable policy makers to adapt strategies needed
hydrological cycle. Over the Mono basin, our work is also for better management of water and natural disasters and the
consistent with many works (Tall et al., 2016; Batablinlè et development of a resilient agricultural system in the face of
al, 2018; Lamboni et al., 2019a, b), which found a strong de- climate change. Since assessments of the impacts of climate
crease in precipitation by 2100 under the RCP8.5 scenario. It change in West-Africa are one of the issues most addressed
is also in line with the recent special IPCC report which by the IPCC, the use of multi-models for future extreme tem-
states that West Africa will likely experience longer and perature and precipitation assessment under the greenhouse
more intense droughts in the near future (IPCC, 2013). In gas emission scenario RCP4.5 and RCP8.5 over Mono basin
general, when considering the number of rainy days and the is a promising prospect that would contribute to a better un-
intensity of rainy days (mostly declining) were coherent with derstanding of the potential effects of climate change
that of rain simulated average on the Mono basin. By conse-
quence, Extreme rains in some areas could lead to some ex-
treme events such as floods and droughts. The analysis of the 4 Conclusions
maximum duration of wet sequences shows a sharp decline
in the Mono basin (Batablinlè et al., 2018). Ly et al. (2013) A new multi-source dataset is used in this paper to analyze
showed a gradual decrease in the maximum duration of the changes in extreme temperature and rainfall events over the
number of wet sequences in West Africa since 1960. Our Mono basin on the far future (2071–2100) and the histori-
work shows that such an evolution diagnosed on the present cal period (1971–2000). The analysis of these climatic data
climate is likely to persist in the future. This decrease of the has contributed to the knowledge of extreme phenomena and
maximum duration of wet sequences in Togo and particularly their evolution in the future over the Mono basin. These ex-
on the Mono basin appears to be consistent with the evolu- treme climatic phenomena are very random and highly vari-
tion of the maximum duration of dry sequences because a able in space and time. Their knowledge is therefore fun-
sharp decrease in the maximum duration of wet sequences damental to protect against natural disasters. The RCMs as
could favor an increase in the maximum length of the dry se- well as their average showed that rain will decrease over the
quences. This decrease (respectively increase) of the wet se- Mono basin while temperatures (Maximum and minimum)
quences (respectively dry sequences) in the Mono basin is a will increase. Moreover, others climatic indices have also
response to global warming (Sylla et al., 2016). In Benin, the been analyzed in order to characterize these events. In this
work on climate indices made. Yèkambèssoun et al., 2017 way, for the simple daily intensity index and five-day maxi-
showed that only the annual total precipitation, the annual mum rainfall, the models have presented variable structures
total of wet days and the annual maximum rainfall recorded but the average of the models showed a moderate diminution
during 30 d present a significant decreasing trend while the on almost whole Mono basin except on some parts. The anal-

Proc. IAHS, 383, 69–78, 2020 https://doi.org/10.5194/piahs-383-69-2020

B. Lamboni et al.: Extreme Rainfall and Temperature Changes Assessment 77

ysis of wet and dry sequences showed that the most of RCMs References
point out a fairly significant decrease in maximum duration
of the number of consecutive rainy days and an increase of
the maximum duration of the number of consecutive dry days Aguilar, E., Barry, A. A., Brunet, M., Ekang, L., Fernandes, A.,
Massoukina, M., and Zhang, X.: Changes in temperature and pre-
almost all over the Mono basin. Projections of the RCMs
cipitation extremes in western central Africa, Guinea Conakry,
showed that these extreme rain events do not evolve in the and Zimbabwe, 1955–2006, J. Geophys. Res.-Atmos., 114, 1–
same direction as precipitation and this could lead to flood- 11, https://doi.org/10.1029/2008JD011010, 2009.
ing intense in some parts of Mono basin particularly in the Batablinlè, L., Lawin, E., and Agnide, C. M.: Africa Cordex Sim-
north. Seasonal evolutions were investigated to provide more ulations Projection of Future Temperature, Precipitation, Fre-
detail about intra-annual behavior in extreme parameters for quency and Intensity Indices Over Mono Basin in West Africa,
the Mono river basin. The results show that declining mean J. Earth Sci. Clim. Change, 9, 490, https://doi.org/10.4172/2157-
precipitation heights at all scales and irregular rainy seasons 7617.1000490, 2018.
may be associated with a palpable increase in temperatures Bazyomo, S. D., Ouedraogo, A., and Lawin, E. A.: Seasonal trends
in the Mono river watershed by 2100. solar radiation available at the earth’s surface and implication of
future annual power outputs changes on the photovoltaic systems
with one and two tracking axes, J. Clim. Weather Forecast., 5, 1–
9, 2017.
Data availability. All data can be accessed via the references: The
Bigot, P.: Variability climatic, interaction and modification environ-
Table 1 presents the data used in this study. It is a set of eight cli-
nement, Variabilite climatique, Interaction et Modification, Env-
matic models outputs with daily time step and spatial resolution of
iron. J. Hydrol., 4, 8141, 2003.
0.44◦ and available at: https://esgf-data.dkrz.de/projects/esgf-dkrz/
Diallo, I. and Camara, M.: Comparaison du climat et de la vari-
(last access: 3 August 2017). The observed data used, are provided
abilite interannuelle de la pluie simulée au Sahel par les modèles
by the Togo National Meteorology Agency (Météo Togo). Please,
climatiques regionaux (Comparison of climate and interannual
contact the authors if you are interested in this research data.
variability of rainfall simulated in the Sahel by regional climate
models), 24, 96–106, 2013.
Easterling, D., Evans, J., Groisman, P., Karl T., Kunkel, K., and
Author contributions. BL and AEL conceived and designed the Ambenye, P.: Observed variability and trends in extreme climate
research; CM and DLO contributed in providing data; BL per- events: A brief review, B. Am. Meteorol. Soc., 81, 417–425, 200.
formed the research, analyzed the data and wrote the paper. Emmanuel, L. A., Lamboni, B., Célestin, M., and Hodabalo, K.: Fu-
ture ExtremesTemperature:Trends and Changes Assessment over
the Mono River Basin, Togo (West Africa), J. Water Resour. Pro-
Competing interests. The authors declare that they have no con- tect., 11, 82–98, 2019.
flict of interest. Ernest, A., Yves, T., Totin, H. S. V., Gil, M., and Pierre,
C.:Dynamique et modélisation des crues dans le bassin du
Mono à Nangbéto (Togo/Bénin), Hydrol. Sci. J., 59, 2060–2071,
Special issue statement. This article is part of the special issue https://doi.org/10.1080/02626667.2013.871015, 2015.
“Hydrological processes and water security in a changing world”. Gudmundsson, L., Bremnes, J. B., Haugen, J. E., and Engen-
It is a result of the 8th Global FRIEND–Water Conference: Hydro- Skaugen, T.: Technical Note: Downscaling RCM precipitation
logical Processes and Water Security in a Changing World, Beijing, to the station scale using statistical transformations – a com-
China, 6–9 November 2018. parison of methods, Hydrol. Earth Syst. Sci., 16, 3383–3390,
https://doi.org/10.5194/hess-16-3383-2012, 2012.
Hulme, M., Doherty, R. M., Ngara, T., New, M. G., and David,
Acknowledgements. I acknowledge the financial support from L.: African climate change: 1900–2100, Clim. Res, 17, 145–168,
the Deutscher Akademischer Austausch Dienst (DAAD). https://doi.org/10.3354/cr017145, 2001.
All my gratitude to UNESCO-IHP and local organizers who gave IPCC: Climate Change 2013: The Physical Science Basis, Contri-
me financial support and took charge of my participation in the 8th bution of WorkingGroup I to the Fifth Assessment Report of the
World Conference on Large River Basins in China. We are also Intergovernmental Panel on Climate Change, edited by: Stocker,
grateful to the anonymous reviewers whose thoughtful and con- T. F., Qin, D., Plattner, G.-K., Tignor, M., Allen, S. K., Boschung,
structive comments have improved the quality of the article. J., Nauels, A., Xia, Y., Bex, V., and Midgley, P. M., Cambridge
University Press, Cambridge, United Kingdomand New York,
NY, USA, 2013.
Financial support. This research has been supported by Lamboni, B., Emmanuel, L. A., and Manirakiza, C.: Future ex-
UNESCO-IHP. tremes Temperature and Rainfall: Trends and Changes As-
sessment over Mono Basin in WEST AFRICA, XXXII Col-
loque l’Association Internationale de Climatologie; Climatic
Change, Variability and Climate Risks, Thessaloniki Greece, 9–
14, 2019a.
Lamboni, B., Emmanuel, L. A., Manirakiza, C., and Djibib,
Z. M.: Variability ofFuture Rainfall over the Mono River

https://doi.org/10.5194/piahs-383-69-2020 Proc. IAHS, 383, 69–78, 2020

78 B. Lamboni et al.: Extreme Rainfall and Temperature Changes Assessment

Basin of West-Africa, Am. J. Clim. Change , 8, 137–155, Sarr, A. B. and Camara, M.: Projections Clima-
https://doi.org/10.4236/ajcc.2019.81008, 2019b. tiques Sur Le Sénégal, Europ. Sci. J., 13, 206–222,
Ly, M., Traoré, S. B., Alhassane, A., and Sarr, B.: Evolution of https://doi.org/10.19044/esj.2017.v13n17p206, 2017.
Some Observed Climate Extremes in the West African Sahel, Soro, G. E., Anouman, D., Goula Bi, T. A., and Srohorou, B.: Char-
Weather Clim. Ext., 1, 19–25, 2013. acterization of sequences of meteorological dryness at various
Mason, S., Waylen, P., Mimmack, G., Rajaratnam, B., and Harrison, time scales in Sudanese climate: case of the extreme northwest
J.: Changes in extreme rainfall events in South Africa, Climatic of the Ivory Coast, Larhyss J., 18, 107–124, 2014.
Change, 41, 249–257, 1999. Sylla, M. B., Elguindi, N., and Giorgi, F. W. D.: Projected ro-
MERF: Plan d’action national d’adaptation aux changements clima- bust shift of climate zones over West Africa in response to
tiques (PANA), (National Action Plan for Adaptation to Climate anthropogenic climate change for the late 21st century, Cli-
Change ), Lomé, Republic of Togo, 27–134, 2009. matic Change, 134, 241–253, https://doi.org/10.1007/s10584-
Nelson, G. C., Rosegrant, M., Palazzo, A., Gray, I., Ingersoll, C., 015-1522-z, 2016.
and Robertson, R.: Food Security, Farming, and Climate Change Tall, M., Sylla, M. B., Diallo, I., Pal, J. S., Mbaye, M. L., and
to 2050: Scenarios, Results, Policy Options, Washington, 2010. Amadou, T. G.: Projected impact of climate change in the hy-
Nicholson, S. E. and Webster, P. J.: Physical basis for the interan- droclimatology of Senegal with a focus over the Lake of Guiers
nual variability of rainfall in the Sahel, Q. J. R. Meteorol. Soc., for the 21st century, Theor. Appl. Climatol., 129, 655–665,
133, 2065–2084, 2007. https://doi.org/10.1007/s00704-016-1805-y, 2016.
Oguntundé, P., Abiodunb, B., and Lischeid, G.: Rainfall trends in Yèkambèssoun, N. O., Emmanuel, A. L., André, A., Benjamin, K.
Nigeria, 1901–2000, J. Hydrol, 411, 207–218, 2011. Y., Ganiyu, T. O., and Abel, A. A.: Decreasing past and mid-
Ozer, P., Hountondji, Y., and Laminou Manzo, O.: Evolution des century rainfall indices over the ouémé river basin, benin (West
caractéristiques pluviométriques dans l’est du Niger de 1940 à Africa), Climate, 5, 1–28, 2017
2007, Evolution of rainfall characteristics in eastern Niger from
1940 to 2007, GEO-ECO-Trop, 33, 11–30, 2009.

Proc. IAHS, 383, 69–78, 2020 https://doi.org/10.5194/piahs-383-69-2020