Impact of Climate Change on Groundwater Recharge and Base Flow in the Sub-

Catchment of Tekeze Basin, Ethiopia

Kiflom Degef Kahsay1, 2*, Santosh M. Pingale1, Samuel Dagalo Hatiye1

1
Department of Water Resources and Irrigation Engineering, Arba Minch Institute of
Technology, Arba Minch University, Arba Minch, Ethiopia, P.O. Box 21
2
Tigray Agricultural Research Institute, Alamata center, Ethiopia
Email: [email protected];
[email protected];
[email protected]
*Corresponding author

Abstract
The impacts of climate change are significant on both surface and groundwater resources.
However, little attention has been given to the effect of climate change on groundwater
resources. Therefore, the present study is concerned with the effect of climate change on
groundwater recharge and base flow in Tekeze sub-catchment in Ethiopia. The future climate
variables were obtained from Coordinated Regional Climate Downscaling Experiment
(CORDEX) Africa program for Representative Concentration Pathways (RCPs) of RCP 2.6
and RCP 4.5 scenarios. The Mann-Kendall test and Sen’s slope estimator were used for trend
detection using XLSTAT software package. Further, the downscaled and bias corrected
precipitation, temperature, and potential evapotranspiration were used as input to the WetSpa
model to simulate future water balance changes. The results indicated a decreasing trend in
annual rainfall and an increasing trend in average temperature and evapotranspiration for
selected scenarios. At the catchment level, precipitation decreases by 20% for both RCP 2.6
and RCP 4.5 scenarios, and actual evapotranspiration shows 0.4% and 8.1%increment for
RCP 2.6 and RCP 4.5, respectively. Consequently, the groundwater recharge decreases by
3.4% for RCP 2.6 and 1.3% for RCP 4.5. Base flow will also decrease by 1.5%and 0.55% for
RCP 2.6 and RCP 4.5, respectively. The results of this study would help policymakers,
scientists, government officials and local stakeholders in planning and management of the
surface and groundwater resources in the Ethiopian regions.

© 2017. This manuscript version is made available under the Elsevier user license
http://www.elsevier.com/open-access/userlicense/1.0/
Graphical abstract

Keywords: Tekeze catchment, Climate change, Recharge, Base flow, Trend analysis,
WetSpa

1 Introduction
Water is indispensable for life. However, its availability in sustainable quality and quantity is
threatened by many factors, of which climate plays a leading role (Woldeamlaket al., 2007;
Kumar, 2012). The evidence is mounting that we are in a period of climate change brought by
increasing atmospheric concentrations of greenhouse gasses (GHGs). The continuation of this
phenomenon may significantly alter global and local climate characteristics, including
2
temperature and precipitation (Kumar, 2012). Climate change can have profound effects on
the hydrologic cycle through precipitation, evapotranspiration and soil moisture with
increasing temperature. However, the extra precipitation will be unequally distributed around
the globe. Some parts of the world may see significant reductions in precipitation or major
alterations in the timing of wet and dry seasons (Kumar, 2012). Therefore, knowledge on the
impacts of climate change on hydrological processes and water resources is becoming more
critical.

Findings of the IPCC (2013) strongly suggest that climate change has the potential to
deteriorate the groundwater availability, water quality, and water supplies. Herrera-Pantoja
and Hiscock (2008) concluded that future climate may present a decrease in potential
groundwater recharge that will increase stress on local and regional groundwater resources.
In understanding the consequences of climate change on groundwater resources and water
balance system (especially, precipitation and evapotranspiration), an important component
that needs to have due attention is the influence of anthropogenic factors on groundwater
recharge and base flow. Current research has mainly focused on surface water resources and
very little is known about the potential impacts of climate change on groundwater. The
impacts on groundwater are far-reaching and need to be investigated, especially in African
regions, where most people rely on groundwater for different purposes (Taylor and Howard,
1996; Kulabako et al., 2007; Abiye 2016).

Ethiopia has abundant water resources potential of which groundwater has a great role. The
preliminary estimated amount of yearly groundwater recharge of the country is about 28,000
Mm3 (Obuobie, 2008; MoWIE, 2017). Awulachew et al., (2007) indicated that the potential is
much greater than this amount. Most of the developed groundwater resources are mainly used
for irrigation, domestic and industrial water supply. Groundwater is an important water
resource and needs more attention on its vulnerability to overexploitation (Dey et al. 2017),
pollution and, especially, climate change. Quantifying the impact of climate change on
groundwater resources requires not only reliable prediction of changes in the major climatic
variables but also the accurate estimation of groundwater recharge. The information on
groundwater recharge is necessary for sustainable management of water resources (Tilahun
and Merkel, 2009) and fluid and contaminant transport modeling within the subsurface
(Healy, 2010).

3
Groundwater recharge is a function of climatic factors, local geological formation,
topography and land use types of the area under consideration (Dragoni and Sukhija, 2013).
The regime has a direct relationship with precipitation and physiochemical properties of soils.
As precipitation varies with climate change as a result of temperature and evapotranspiration,
there is possible variation in groundwater recharge. Due to variations in the distribution of
rainfall, drought is a frequently recurring phenomenon in Ethiopia (Gebrehiwot and Veen,
2013). Earlier, Meze-Hausken (2004) also reported the perception of climate variability and
meteorological drought and further, its impact on decreased water availability in Northern
Ethiopia. There are several studies reported concerning drought phenomenon in Ethiopia in
recent past (e.g., Gedif et al., 2014; Legesse and Suryabhagavan, 2014; Suryabhagavan,
2017). Evidently, Ethiopia in general and the study area in particular, have suffered from a
shortage of food due to erratic rainfall (Lewis, 2017) and unsustainable use of water
resources (Gebrehiwot and Veen, 2013).

The conservation of the surface runoff and groundwater recharge, and the impact of climate
change need to be taken into consideration. Many researchers have conducted studies to
investigate the climate change impacts on groundwater and recharge in and around Tekeze
river basin (Adem, 2006; Walraevens et al., 2009; Beyene et al., 2011; Tesfagiorgis et al.,
2011; Gebremeskel, 2015; Haile, 2015). However, most of the earlier studies were focused
on the effect of land use/land cover change on hydrological processes including groundwater
and the rate of groundwater recharge in the basin (e.g., Gebremicael et al., 2013; Ashenafi,
2014; Haregeweyn et al., 2015; Welde and Gebremariam, 2017). Haile (2015) estimated
groundwater recharge and potentials under changing climate conditions in Werii watershed in
Tekeze river basin. However, the study is limited to IPCC A1B and B1 scenarios. The impact
of climate change on groundwater recharge and base flow were not considered. Very few
studies were done on groundwater recharge and base flow (e.g., Zomlot et., 2015). Therefore,
studying the basins groundwater characteristics in terms of groundwater recharge and base
flow and climate change impact is essential. The main objective of this study is to estimate
the effect of climate change on groundwater recharge and base flow using climate and
hydrological models in the Tekeze River basin. The trends of major climate variables such as
precipitation, evapotranspiration, and temperature affecting groundwater recharge in the
Tekeze sub-catchment were also assessed usingMann-Kendall (MK) test. The results from

4
this study can be used by local stakeholders and water managers for planning and solve water
resource management problems associated with water supply for domestic and agricultural
purposes.

2 Materials and Methods

2.1 Study Area
Tekeze river basin is situated in the north-western part of Ethiopia and forms the northern
most part of the Nile Basin within Ethiopia. Specifically, the Tekeze sub-catchment is located
southeast of the basin in the geographical location of 11o39’32.17” and 13o27’15.96’’ East
longitude, and 37o33’27.63’’ and 39o40’7.24’’ North latitude (Fig.1). The catchment lies in
both Amhara and Tigray regional states covering a total surface area of 29405 km2. The
elevation of the catchment varies from 1028 to 4536 metres above mean sea level (m.a.s.l.)
(Fig. 1). The slope of the watershed varied between 0 to 78.2% (Fig. 2a) The Tekeze River
contributes nearly 13% of the total annual flow of the Nile waters during the dry season and
22% during the flood season (Degefu, 2003).

The Tekeze basin has highly variable rainfall concentrated in one rainy season (June to
September) separated by relatively long dry seasons. The mean annual rainfall in the Tekeze
basin ranges from 600 mm to 1300 mm. The climate of the study area is classified as tropical
to alpine due to its great difference in altitude and topography. The average maximum and
minimum temperatures are around 24.8oC and 12.5oC, respectively. There is high temporal
variation of rainfall and high variation between daily maximum and minimum temperatures
in the study area. However, the seasonal variation of temperature is less than the diurnal
variation. The major land use/land cover (LULC) classes in the basin include intensively
cultivated land (7%), sparsely cultivated (58%), open woodland (12%), open grassland (5%),
sparsely vegetated (0.2%), complex land (15%), and others (2.8%) (Fig. 2b). According to
the USDA soil classification system, five soil types are identified in the catchment namely
sandy loam, silty clay, silt, silty loam, and clay. Silty loam, clay, and silt are the dominant
soil types of the catchment area (Fig. 2c).

2.2 Data used

The meteorological data such as precipitation, minimum and maximum temperature from the
years 1994 to 2013 were collected from the Ethiopian National Meteorological Agency

5
(NMA). Precipitation data were utilized from six meteorological stations (Gonder, Lalibela,
Mitsebri, Maichew, Mekelle, and Samre). The streamflow data of the catchment at the
Embamadre gauging station was collected from the Ethiopian Ministry of Water, Irrigation,
and Energy (MoWIE). The digital elevation model (DEM) of 30m resolution was obtained
from the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER)
Global Elevation Model (GDEM) version 2 databases(http://aster.usgs.gov) which is a
product of Ministry of Economy, Trade, and Industry (METI) of Japan and National
Aeronautics and Space Administration (NASA) agency of United States of America. Slope
map of the watershed was derived from the DEM (Fig. 2a). The soil and LULC map is taken
from the Ethiopian land cover data set, which is derived from the original raster based global
land cover of Africa archive (http://www.africover.org./index.htm) of the year 2008 (Fig. 2b
& c). Initially, rainfall data was subjected to data quality checks such as estimation of missing
rainfall, homogeneity and inconsistencies in the records by using statistical methods. For this
study, missed data (Table 1) was filled by using meteorological stations located in the nearby
catchment through Normal ratio and distance power methods following the approach
suggested by Subramanya (1998)) depending on the percentage of missed data available for
each station. The double mass curve method was applied for checking consistency in the data
among the selected meteorological stations and data were found to be consistent. Thiessen
polygon method was used to estimate the areal precipitation of the Tekeze sub-catchment.

The daily large scale RCM predictors from regional climate model REMO developed at the
Max-Planck Institute for Meteorology (MPIM) in Hamburg, Germany (Jacob, 2001) were
used to simulate the future climate changes. Downscaled rainfall and temperature (minimum
and maximum) projected climate data for the period 1951-2100 were obtained from
CORDEX-Africa database at a spatial resolution of 55km (0.44o×0.44o) for RCP 4.5 and RCP
2.6 scenarios. This data was used from Coupled Model Inter-Comparison Project Phase 5
(CMIP5) of IPCC fifth assessment report. The future climate change parameters on daily
basis for the period of 1951 to 2100 for rainfall and temperature (minimum and maximum)
data on a grid basis were collected from IWMI. The selected RCP data were subjected to bias
corrections. The selected study area was covered with six RCM grid cells. RCP 2.6 scenario
represents emissions of greenhouse gasses (GHG) and air pollutants and measured peak
emissions during 2010 to 2020 and declines substantially thereafter. The RCP 4.5 scenario
represents a range of technologies and strategies for reducing GHG emissions until 2100. For

6
daily rainfall and temperature, future scenarios were considered for two different periods of
near–term (2020–2049) and mid–term (2050–2079).

2.3 Methods
The conceptual framework of the study is shown in Fig. 3 and discussed in the following
sections.

2.3.1 Bias correction

The bias corrections are necessary since climate models often exhibit systematic error
(biases) which are caused due to the limited spatial resolution, simplified physics and
thermodynamic processes, numerical schemes or incomplete knowledge of climate system
processes (http://ccafs-climate.org/bias_correction/). Bias correction is therefore applied to
compensate for any tendency to overestimate or underestimate of the mean of downscaled
variables. In this study, both linear and non-linear bias correction methods were used for
correcting daily temperature and precipitation data. Leander and Buishand (2007) used a
power transformation function, which corrects the coefficient of variation (CV) and the mean.
In this nonlinear correction method, each daily precipitation amount, P, is being transformed
to a corrected P* using the equation:
P*  aP b (1)
where P* is the simulated data in the projection period; ‘a’ and ‘b’ are the parameters
obtained from calibration in the baseline period and subsequently applied to the projection
period. These parameters are determined by matching the mean and CV of the simulated data
with the observed data.
For temperature, monthly systematic biases were calculated for the baseline period (1994–
2005) by comparing RCP outputs with the observed temperature. The monthly mean bias
corrections were made according to equation 2 (Ho et al., 2012).
o
Tc  Tom  * ( Tr  Trm ) (2)
r
where Tc= bias corrected future RCP temperature
Tom= mean of observed temperature in base period
Trm= mean of RCP temperature in base period
Tr= RCP temperature of base period

7
  r and  o = the standard deviation of the daily RCP output and observations in the
reference period, respectively.

2.3.2 Estimation of potential evapotranspiration (PET)

After the bias correction for daily minimum and maximum temperature, potential
evapotranspiration was estimated for use in the WetSpa (Water and Energy Transfer between
Soil, Plants and Atmosphere) model. The PET time series data were calculated using the
Hargreaves equation (Allen et al., 1998).
PET  0.0023( Tmean  17.8 )( Tmax  Tmin )  0.5Ra (3)
where, PET is the potential evapotranspiration (mm/day); Tmean, Tmax and Tmin are average,
maximum and minimum temperature (°C) values, respectively; Ra is extraterrestrial radiation
(mm/day). The corrected average daily PET, precipitation, and observed stream discharge
were used as an input for the WetSpa model for simulation of the hydrological water balance
components. These climate data were generated based on the RCP scenarios for the future
climate change projections.

2.3.3 Trend Analysis

Detecting and assessing temporal and spatial trends is important for many environmental
studies and monitoring programs (ITRC 2013). For this study, trend analysis was done using
non-parametricMann-Kendall (MK) test. This is a statistical test widely used for the analysis
of the trend in climatological variables (e.g., Pingale et al., 2014; Seyoum et al., 2015;
Pingale et al., 2016) and in hydrologic time series analysis (Yue and Wang, 2004). There are
two advantages of MK test. First, it is a non-parametric test and does not require the data to
be normally distributed. Second, the test has low sensitivity to abrupt breaks due to
heterogeneous time series (Tabari et al., 2011). In fact, this test statistic is used to test the null
hypothesis (Ho) regarding the existence of trends in time series data. If |Zs| ≥Zα/2, where α
represents the significance level (e.g. 5% with Z0.025 = 1.96) then the null hypothesis is invalid
implying that the trend is significant (Motiee and McBean, 2009). If p-value less is than the
significance level (α=0.05), the null hypothesis (Ho) is rejected (statistically significant),
while for p-value greater than the significance level (α=0.05), Ho is accepted (statistically
non-significant) (Onoz and Bayazit, 2012). Failing to reject Ho does not mean that there is no
trend. Rather, it is a statement that the evidence available is not enough to conclude on the
existence of trends (Helsel and Hirsch, 2002). The positive and negative values of the S

8
statistic indicate upward and downward trends, respectively. The XLSTAT version 2015
software was used to perform the statistical MK test analysis. The null hypothesis is tested at
95% confidence level for precipitation, temperature and evapotranspiration data for the six
meteorological stations. Further details about the MK test can be found in Pingale et al.,
(2016).

2.3.4 Groundwater recharge and base flow estimation

2.3.4.1 Description of WetSpa Model
WetSpa is a physically based distributed hydrological model for predicting the Water and
Energy Transfer between Soil, Plants, and Atmosphere on regional or basin scale and daily
time step, developed in the Vrije Universiteit Brussels, Belgium (Batelaan et al., 1996; Wang
et al., 1996). It is one of the modeling packages used for simulating hydrologic processes
under changing climateconditions. Different researchers (e.g., Adem and Batelaan, 2006;
Bahrem and De Smedt, 2008; Nyenje and Batelaan, 2009; Jaroslaw and Batelaan, 2011; Haile
and Kassa, 2015; Yenehun et al., 2017) have studied hydrological processes and groundwater
rechargeand associated impacts of climate change using the WetSpa model.

Historical hydro-climatic and other data such as precipitation, minimum and maximum
temperature, potential evapotranspiration, streamflow, DEM, land use and soil type were
used as an input for this model. According to Liu and De Smedt (2004) and Nyenje and
Batelaan (2009), river flow hydrographs, soil moisture, infiltration rates, groundwater
recharge, surface water retention and runoff are the main outputs of the WetSpa model. The
water balance of the root zone is considered as a key factor controlling runoff, interflow and
groundwater recharge and is calculated for each grid cell as:

D  PI S EF R (4)
t
where D is the root zone depth [L];  is the change in soil moisture content [L/L]; t is the
time interval [T]; P is the precipitation [L]; I is the initial abstraction (interception and
depression losses) [L]; S is the surface runoff [L]; E is the actual evapotranspiration [L]; F is
the interflow [L] and R is the percolation out of the root zone [L]. The percolation out of the
root zone recharges the groundwater storage (SG)[L], which then contributes to groundwater
discharge forming the base flow of a stream (Liu and De Smedt, 2004). Base flow is a basic
element in groundwater studies as it explains the behavior of water movements in the
subsurface (Gu et al., 1998).

9
Recharge is estimated using the relationship between hydraulic conductivity and effective
saturation (Brooks and Corey, 1966):
   s ( 23 B ) B
R  Ks( ) (5)
s r
where R is the recharge or percolation (mm); Ks is the saturated soil hydraulic conductivity
(mm/h); θ is the soil moisture content(m3/m3); θs is the saturated soil moisture
content(m3/m3); θr is the residual moisture content (m3/m3); B is the pore size distribution
index, dimensionless. These values for recharge estimation were obtained based on the
identified soil textural class in the study area and provided to the model (Cosby et al., 1984;
Rawls et al., 1982 cited in Liu and De Smedt, 2004).
The base flow simulation is performed at the sub-catchment scale using a lumped linear
reservoir approach (Liu and De Smedt, 2004),
SG( t )
Qg ( t )  C g ( ) (6)
1000
where Qg(t) is the base flow (m3/s); SG(t) is the groundwater storage (mm) and Cg is the
groundwater flow recession constant (m2/s).

2.3.4.2 WetSpa model set up

Gridded model parameters were derived from topography, land use and soil maps of the
catchment automatically together with attribute lookup tables prepared in dbf format.
Hydrological features of the watershed such as surface slope, hydraulic radius, flow direction,
flow accumulation, stream network, order, as well as sub-catchments were delineated from
the DEM. Soil hydraulic conductivity, pore size distribution index, plant wilting point
porosity, field capacity, and residual moisture content for each grid cell were derived from
the soil map. Similarly, Manning’s roughness coefficient, interception storage capacity and
root depth parameters were derived from the land use map (Table 2 and 3).

A combination of elevation, soil and land use grid maps were used to provide grids of
potential runoff coefficient and depression storage capacity of the watershed by means of
attribute lookup tables. The instantaneous unit hydrograph (IUH), travel time to the basin
outlet, grids of flow velocity and standard deviation were generated at the final time step. The
parameter values that are taken as a threshold value and derived from other parameters and
constants are listed in Table 2.In order to run the WetSpa model, the input base maps must

10
have similar area and cell size. Hence, same area and cell size of the watershed was created
for the base maps of topography, land use and soil type. Accordingly, the catchments base
maps are made for 800m grid cell size with an average of 287 rows and 248 columns.

2.3.4.3 Calibration and validation

The main calibration global parameters and corresponding measurement units of WetSpa
model are given in Table 3. Interflow scaling factor (Ki) is a parameter for reflecting the
organic matter in the root zone associated with soil hydraulic conductivity. Groundwater flow
recession coefficient (Kg) is a global parameter for reflecting catchment’s groundwater
recession regime and relative soil moisture parameter (K_ss) is related to field capacity for
soil moisture content. Similarly, potential evapotranspiration is associated with a correction
factor K_ep and G0 is the depth of initial groundwater storage. The maximum groundwater
storage parameter (G_max) is dependent on groundwater depth and K_run is an exponent for
reflecting the effect of small rainfall intensity on surface runoff. P_max is a modeling time-
dependent threshold for rainfall intensity. Since snow melting and accumulation did not occur
in the catchment, temperature data were not taken as an input for the modeling process.
Hence, the global parameters such as base temperature (TO) and degree day coefficient
(K_snow) for estimating snowmelt, as well as the rainfall degree day coefficient (K_rain)
were set to a negative value (-1)so that they did not affect the model result.

To investigate the performance of the model, calibration and validation steps were
undertaken. The hydro-meteorological data such as daily precipitation, evapotranspiration,
and runoff were divided to use for independent calibration (1995-2001) and validation (2002-
2004) processes. Hence, data recorded within similar timescale for all the meteorological
parameters and spatial data derived from the base map, land use, and soil texture were used
for calibration as well as validation of the model (Table 2). Model calibration was carried out
by repetitive trial and error method and fine-tuning the global parameters within the range
(Table 3). The calibration and validation results were also compared with observed values
using model performance indicators. After the calibration of the WetSpa model using the
global model parameters, the water balance components were estimated based on the
measured input parameters to the model. The water balance components simulated for the
catchment include total interception, surface runoff, infiltration, percolation, actual
evapotranspiration, interflow, groundwater drainage, soil moisture storage and groundwater

11
storage. More details about these parameters and respective equations are found in WetSpa
user manual (Liu and Smedt, 2004).

2.3.4.4 Model performance indicators

Model bias
Model bias (MB) is a relative mean difference between predicted and observed stream flows,
reflecting the ability to reproduce water balance. Model bias is given by the equation:


N
(Qsi  Qoi )
MB  i 1
(7)
i1 Qoi
N

where MB is the model bias, Qsi and Qoi are the simulated and observed stream flows at time
step i (m3/s), and N is the number of time steps over the simulation period.

Model bias measures the systematic under or over prediction of a model for a set of
predictions (Stockdale et al., 2011). A lower MB value indicates a better fit, and the value 0.0
represents the perfect simulation of observed flow volume. Model bias values tend to vary
more during dry periods than during wet periods for streamflow (Gupta et al., 1999). Model
simulation values can be accepted if MB is between -0.25 and 0.25 for streamflow (Moriasi
et al., 2007).

Model confidence
Model confidence is expressed by the coefficient of determination, which is one of the
important criteria in the assessment of continuous model simulation. It is calculated using:


N
(Qsi  Qo )
R 2
 1 i 1
(8)

N
i 1
(Qoi  Qo ) 2

where R2 is the model determination coefficient and Qo is the mean observed stream flow. R2
values greater than 0.5 are considered as acceptable (Santhi et al., 2001). R2 is very sensitive
to outliers and less sensitive to additive and proportional difference values between simulated
and observed data (Legates and McCabe, 1999). Therefore, Nash-Sutcliffe efficiency is
employed besides the coefficient of determination.

12
Nash-Sutcliffe efficiency
Nash-Sutcliffe efficiency (NSE) indicator was used to describe how well discharges are
simulated by the model and can be described as (Nash and Sutcliffe, 1970):


N
( Qsi  Qoi )2
NSE  1  i 1
(9)

N
i 1
( Qoi  Qo )2

The NSE value can range from a -∞ to 1, with 1 indicating a perfect fit between the simulated
and observed hydrographs. NSE is used to calibrate highly variable flow regimes
characterized by extremely high and low flows. Hence, NSE was found to be the best
objective function for reflecting the overall fit of a hydrograph. Model simulation can be
judged as satisfactory if NSE > 0.50 (Moriasi et al., 2007).

3 Results and Discussion

3.1 Bias corrections
The bias corrections in monthly precipitation and temperature (maximum and minimum) of
the RCP 2.6 scenario for the base period 1994–2005 are presented for Mekelle station (Fig.
4a to c). Statistical parameters indicating the corresponding uncorrected RCP and bias-
corrected RCP scenarios from observation for monthly mean maximum and minimum
temperatures were depicted in Table 4. The Standard Error of Mean (Mean Abs Error),
Standard Deviation (Std-Error) and mean of the bias-corrected temperature data shows close
agreement to the observed value. Similarly, the magnitude of bias-corrected RCP
precipitation data are fitted with observed precipitation data for Michew station (Fig.5 and 6).
For other stations, the corrected RCP precipitation is found to be consistent with observed
precipitation. The bias-corrected temperature data also show close agreement with the
observed values.
3.2 Trend analysis
Trend analysis in this study helped to compare future groundwater recharge and base flow of
the catchment with climate and precipitation changes within the newly emerged RCP climate
change scenarios. In this study, the trend was analyzed for current and future projected
precipitation, temperature and evapotranspiration for the observed period of 1994–2013 and
the future period of 2020–2079.

13
3.2.1 Precipitation
The MK test statistic for the period 1994-2013 indicates that the stations (Mekelle, Mitsebri,
Miachew, Lalibela, Samre and Gonder) show anon-significant decreasing trend in
precipitation (Table 5). The MK test statistic (S) indicates a non-significant decreasing trend
for the observed and future projected precipitation for both the climate change scenarios
(RCP 2.6 and RCP 4.5) except Mekelle and Samre stations (Table 5). The Samre station
shows a statistically significant increasing trend in annual average precipitation (at the rate of
4.538 mm per annum, while Mekelle station showed decreasing trend at the rate of 1.917 mm
per annum at 5% level of significance. The Sen’s slope test also indicates the decreasing
trend of rainfall for the observed period, on an average of 8.762 mm per annum over the
study area. Future (2020–2079) projected rainfall again decreases by 1.724mm/annum and
1.135 mm/annum for RCP 2.6 and RCP 4.5 scenarios, respectively for the next 60 years. This
result agrees with the previous studies conducted at Gilgel Abay catchment which drains to
Lake Tana (Abdo et al., 2009), climate variability in the northern part of Ethiopia
(Gebrehiwot and Veen, 2013), hydro-climatic trends in the Abay/upper Blue Nile basin,
Ethiopia (Tekleab et al. 2013), trend and variability of rainfall in northern Ethiopia (Hadgu et
al., 2013) and IPCC (2013) using the Special Report on Emissions Scenarios (SRES) both for
the current and future periods. However, Gizaw et al. (2017) have reported that the projected
mean annual precipitation may increase by about 6%(9%) in the 2050s(2080s) which results
in increase in stream flow by about 3% (9%) in the major Ethiopian river basins including
Tekeze rivers under RCP 4.5 and RCP 8.5 scenarios. This may be due to ensembles of 10
CMIP5 GCMs average projections considered using Hydrologic Simulation Program-
FORTRAN (HSPF) hydrologic model and different sources of uncertainties caused because
of model uncertainty as well as Ethiopia’s complex topography (Berhanu et al., 2014). The
authors emphasized that more realistic climate projections could be provided using high
resolutions regional climate models.
3.2.2 Temperature
The MK test provided an interesting insight about annual average temperature for the Tekeze
catchment. The MK trend test for temperature shows a statistically significant increasing
trend for the current (1994-2013) and future (2020-2079) climates .There is a statistically
significant increasing temperature trend at Maichew and Mitsebri stations at 5% significance
level (Fig. 7). The annual average temperature was found to be increasing at the rate of
0.001oC at Michew and 0.013oC at Mitsebri stations, respectively. However, trend results

14
were found non-significant for Mekelle, Lalibella and Gonder stations during the observed
period (Table 5). Only a statistically significant increasing trend for future projected
temperature (2020-2079) was found in the case of RCP 4.5 scenarios at Gondar station (Table
5). It has also been revealed that the average temperature will increase at the rate of 0.011oC
at Gondar station. However, other stations showed non-significant increasing trends on the
average annual temperature in both RCP 2.6 and RCP 4.5 scenarios (Table 5). The average
temperature was found to be increasing at the rate of 0.023oC over the observed period in the
study area. Similarly, the future projected average temperature was found to be increasing at
the rate of 0.0012oC for RCP 2.6 and 0.004oC for RCP 4.5 scenarios in the study area (Table
5). This result is in line with studies conducted in and around the basin (e.g., Gebrehiwot and
Veen, 2013; Tekleab et al., 2013; Hadgu et al., 2013; Haile and Kassa, 2015; Gizaw et al.,
2017). Future temperature increases could have a significant impact on the sustainability of
surface water resource and groundwater recharge. This evidence is in good agreement with
Yu et al. (2015), an increase in temperature along with reduced precipitation is likely to result
in the loss of agricultural land due to decreased soil moisture, increased aridity, increased
salinity and groundwater depletion.

3.2.3 Potential evapotranspiration

The present and future potential evapotranspiration were estimated for all the six stations for
RCP 2.6 and RCP4.5 climate change scenarios (Table 5). Only statistically significant
increasing trend in average annual evapotranspiration was found at Michew (at the rate of
2.738 mm per annum), Mekelle (at the rate of 3.438 mm per annum) and Mitsebri stations (at
the rate of 1.765 mm per annum) in the study area. However, other remaining stations did not
show any statistically significant trends in annual average evapotranspiration at 5% level of
significance. For observed and future evapotranspiration, the MK test shows that there is an
increasing trend during the study period (1994-2013 and 2020–2079). The annual
evapotranspiration was found to increase by 1.676 mm/annum during the period 1994-2013.
Evapotranspiration in the study area may increase by 0.053 mm/annum and 0.180 mm/annum
for RCP 2.6 and RCP 4.5 scenarios, respectively for the coming 60 years. The MK trend test
showed that the change in average annual evapotranspiration for both RCP 2.6 and RCP 4.5
scenarios and observed period (1994-2013) is not statistically significant at α=0.05 over the
study area (Table 5 and Fig. 8).

15
3.3 Estimation of Groundwater Recharge and Base flow
3.3.1 Calibration and validation
The statistical model performance evaluation results for both calibration and validation
processes are presented in Table 6. The statistical comparisons of the observed and simulated
discharge hydrographs have confirmed that WetSpa model is calibrated well (Fig. 9). Table 6
reveals the presence of a good agreement between the observed and model simulated
discharges values of the catchment. The R2 and NSE values are found to be 0.82 and 0.89 and
0.79 and 0.76 during calibration and validation periods, respectively. The statistical model
performance indicators are very close to their optimum best-fit values during the calibration
and validation periods (Table 6). Although the model predicts the average conditions well as
depicted by the statistical parameters, it does not capture the peak flow values well. This
would be attributed to the inherent nature of the physical hydrological models which better
predict average conditions well when compared with the prediction results for extreme cases.
Extreme hydrological events are often victimized by model uncertainties which emanate from
assumptions during the development of such models. Particular to this study, average values
of soil moisture, flow parameters etc. were used which probably affected the model
predictive ability for the peak flow events. Indeed, it is out of scope to determine watershed
parameters which produce such extreme events. This model results are in agreement with
Yenehun et al. (2017) where groundwater recharge was estimated in Geba basin in Ethiopia
using WetSpa model.

Further, total runoff, actual evapotranspiration, groundwater recharge, interflow and soil
moisture have been simulated during the simulation process (Table 7). Base flow is obtained
as a summation of interflow and groundwater flow simulated from the WetSpa extension.
During the calibration, the base flow was produced from interflow (29.5 mm) and
groundwater flow (331.2 mm). The total runoff (327.89 mm) was simulated as 10% of
surface flow (32.6 mm) and 90% (360.7 mm) base flow. The base flow (313.5mm) was
generated from interflow (23.8mm) and groundwater flow during the validation period.
Therefore, the total runoff was found to be 285.8mm (Table 7). During dry season, base flow
mainly contribute to the total runoff. The recharge to the aquifers mainly occurs from
precipitation and losing streams (Kebede, 2013). These model results are in good agreement
with similar studies found in and around the study area (e.g., Nyenje and Batelaan, 2009;
Arefaine et al., 2012; Yenehun et al., 2017).

16
3.3.2 Simulation of groundwater recharge and base flow
After running the calibrated WetSpa model, the basic water balance components such as
actual evapotranspiration, groundwater recharge, surface runoff and interflow at the outlet
have been simulated on a current and future time basis (Fig. 10). The simulated water balance
changes for future (2020-2079) periods for both climate change scenarios (RCP 2.6 and RCP
4.5) and baseline period (1995–2001) are illustrated in Table 8. The water balance
components were analyzed and presented on an annual average basis (Fig. 10a). The study
result indicates that precipitation in the study area is expected to decrease by 20% for both
scenarios (RCP 2.6 and RCP 4.5) (Fig. 10 b & c). This result is consistent with the similar
studies done by Goitom et al., 2012 in the Geba basin in Northern Ethiopia. Further, the
actual evapotranspiration is projected to increase by 0.4% for RCP 2.6 and 8.1% for RCP 4.5.
Similarly, the results in the section 3.2.3 also confirmed an increasing trend in potential ET.

Consequently, the future groundwater recharge is expected to decrease by 3.4% and 1.3% for
RCP 2.6 and RCP 4.5scenarios, respectively (Fig. 10b & 10c). This may occur as a result of
decreasing precipitation. Moreover, the base flow value is decreased for both scenarios, with
a slightly greater decrease for RCP 2.6 compared to the RCP 4.5 scenario. The annual
average estimate of groundwater recharge in this study is found to be close to 50 to 100
mm/year (Kebede, 2013). Similarly, the surface runoff also shows decreasing trend for both
emission scenarios by 15.6% for RCP 2.6 and by 19% for RCP 4.5 (Fig. 10). Moreover,
Groundwater recharge and other hydrological components are found to be highly controlled
by geology (Yenehun et al., 2017). Hence, the risk of annual flooding is limited in the
catchment due to the decreased amount of surface runoff in the future. It can be concluded
that for the climate change scenarios considered in this study precipitation, groundwater
recharge, and base flow will decrease while evapotranspiration will increase.

4 Conclusions
Groundwater is an important water resource and needs more attention on its vulnerability to
over exploitation, pollution and, especially climate change. Therefore, this study was
conducted to investigate the effects of climate change on major climatic variables and
groundwater components (groundwater recharge and base flow) in Tekeze sub-catchment in
Ethiopia. Precipitation indicates a negative trend for the observed and future projected

17
scenarios. For the observed period, the rainfall decreases on average by 8.762 mm per annum
over the study area and future (2020–2079) projected rainfall decreases by 1.724mm/annum
and 1.135 mm/annum for RCP 2.6 and RCP 4.5 scenarios, respectively. The average
temperature shows an increasing trend at 5% significance level in both periods.
Consequently, there is an increasing trend for present and future projected annual
evapotranspiration.

In this study, the WetSpa model was used to simulate future water balance changes in the
catchment. For both RCP 2.6 and RCP 4.5 scenarios, precipitation is projected to decrease by
20-21%. The actual evapotranspiration will be expected to increase by 0.4% for RCP 2.6 and
8.1% for RCP 4.5 which is similar to the evapotranspiration trend analysis. The future
groundwater recharge will decrease by 3.4% and 1.3% for RCP 2.6 and RCP 4.5 scenarios,
respectively. Base flow is projected to decrease more under the RCP 2.6 scenario. Based on
the study results, it can be recommended that rain water harvesting structures and artificial
aquifer recharge schemes should be implemented to enrich groundwater potential as an
adaptation strategy in the study region and mitigate the possible impacts of climate change.

Acknowledgments
The authors would like to thank Tigray Agricultural Research Institute, Ministry of Water,
Irrigation and Energy, National Meteorological Agency, Ethiopia and International Water
Management Institute for providing the necessary data. Mr. Kidane Welde and Mr.
Gebremedhin Gebremeskel are also acknowledged for their support and providing data. We
also acknowledge USGS Earth Resources Observation portal for providing ASTER DEM
data. We also gratefully acknowledge the editor and three anonymous reviewers whose
comments greatly improved the paper.

References

Abdo, K.S., Fiseha, B.M., Rientjes, T.H.M., Gieske, A.S.M., Haile, A.T., 2009. Assessment
of climate change impacts on the hydrology of Gilgel Abay catchment in Lake Tana
basin, Ethiopia. Hydrol. Process. 23, 3661–3669. doi:10.1002/hyp.7363.

Abiye, T., 2016. Synthesis on groundwater recharge in Southern Africa: A supporting tool for
groundwater users. Groundw. Sustain. Dev. doi:10.1016/j.gsd.2016.10.002

18
Adem, G., Batelaan, O., 2006. Modeling groundwater-surface water interaction by coupling
MODFLOW with WetSpa. In geophysical research abstracts 8, 03181.

Allen, R.G., Pereira, L.S., Raes, D., Smith, M., Ab, W., 1998. Crop evapotranspiration-
Guidelines for computing crop water requirements-FAO Irrigation and drainage paper
56, Irrigation and Drainage. doi:10.1016/j.eja.2010.12.001.

Ashenafi A. A. 2014. Modeling hydrological responses to changes in land cover and climate
in Geba River Basin, Northern Ethiopia. A PhD dissertation Department of Earth
Sciences, Institute of Geographic Sciences, Physical Geography, Freie Universität
Berlin, Germany.

Arefaine, T., Nedaw, D., Gebreyohannes, T., 2012. Groundwater Recharge,

Evapotranspiration and Surface Runoff Estimation Using WetSpass Modeling Method
in Illala Catchment, Northern Ethiopia. Momona Ethiop. J. Sci. 4, 96–110.

Awulachew, S.B., Yilma, A.D., Loulseged, M., Loiskandl, W., Ayana, M., Tena, A., 2007.
Water Resources and Irrigation Development in Ethiopia, IWMI.
doi:10.1017/CBO9781107415324.004.

Batelaan, O., Wang, Z.M., De Smedt, F., 1996. An adaptive GIS toolbox for hydrological
modelling. IAHS Publ. Proc. Reports-Intern Assoc Hydrol. Sci. 235, 3–10.
doi:10.5772/1944.

Beyene, A.Y., Batelaan, H., Goitom., 2011. Spatial and temporal simulation of groundwater
recharge for Geba catchment, Northern Ethiopia using WetSpa, MSc. Thesis,
Universiteit Gent, VrijeUniversiteitBrussel, Belgium.

Berhanu, B., Seleshi, Y., Melesse, A.M., 2014. Surface water and groundwater resources
ofEthiopia: potentials and challenges of water resources development. In A. M. Melesse
et al. (eds.), Nile River Basin: Ecohydrological challenges, climate change and
hydropolitics. Doi:10.1007/978-3-319-02720-3_6.

Brooks, R.H., Corey, A.T., 1966. Properties of porous media affecting fluid flow. J. Irrig.
Drain. Div. 92, 2, 61-90.

Cosby, B.J., Hornberger, G.M., Clapp, R.B., Ginn, T.R., 1984. A statistical exploration of the
relationships of soil moisture characteristics to the physical properties of soils, Water
Resour. Res., 20, 682–690.

19
Degefu, G.T., 2003. The Nile: historical, legal and developmental perspectives. Trafford
Publishing.

Dey, N.C., Saha, R., Parvez, M., Bala, S.K., Islam, A.S., Paul, J.K., Hossain, M., 2017.
Sustainability of groundwater use for irrigation of dry-season crops in northwest
Bangladesh. Groundw. Sustain. Dev. 4, 66–77. doi:10.1016/j.gsd.2017.02.001

Dragoni, W., Sukhija, B.S., 2008. Climate change and groundwater: a short review. Geol.
Soc. London, Spec. Publ. 288, 1–12. doi:10.1144/SP288.1.

Gebrehiwot, T.G., Veen, A., 2013. Assessing the evidence of climate variability in the
northern part of Ethiopia. Journal of development and agricultural economics 5, 3, 104-
119.

Gebremeskel, G., 2015. Estimation of groundwater recharge and potentials under changing
climate in Werii Watershed, Tekeze River Basin. Doctoral dissertation, Haramaya
University, Ethiopia.

Gebremicael, T.G., Mohamed, Y.A., Betrie, G.D., van der Zaag, P., Teferi, E., 2013. Trend
analysis of runoff and sediment fluxes in the Upper Blue Nile basin: A combined
analysis of statistical tests, physically-based models and landuse maps. J. Hydrol. 482,
57–68. doi:10.1016/j.jhydrol.2012.12.023.

Gedif, B., Hadish, L., Addisu, S., Suryabhagavan, K. V, 2014. Drought risk assessment
using Remote Sensing and GIS: The Case of Southern Zone, Tigray Region, Ethiopia. J.
Nat. Sci. Res. 4, 2225–921.

Gizaw, MS., Biftu GF., Gan T.Y., Moges S.A., Koivusalo H. 2017. Potential impact of
climate change on streamflow of major Ethiopian rivers. Climatic Change 143(3-4),
371–383.

Gu, C., Hornberger, G.M., Herman, J.S., Mills, A.L., 2008. Influence of stream-groundwater
interactions in the streambed sediments on NO3- flux to a low-relief coastal stream.
Water Resour. Res. 44. doi:10.1029/2007WR006739.

Gupta, H.V., Sorooshian, S., Yapo, P.O., 1999. Status of automatic calibration for hydrologic
models: Comparison with multilevel expert calibration. J. Hydrol. Eng. 4, 135–143.
doi:10.1061/(ASCE)1084-0699(1999)4:2(135).

Goitom, H., Smedt, F.D., Yohannes, T.G., Walraevens, K., Gebrehiwot, K., Bauer, H.,

20
 Deckers J., 2012. Modeling Climate Change Impact in the Geba Basin,
Ethiopia.International Conference on Environmental, Biomedical and Biotechnology
2012 IPCBEE vol.41 IACSIT Press, Singapore 240-244.

Hadgu, G., Tesfaye, K., Mamo, G., Kassa, B., 2013. Trend and variability of rainfall in
Tigray, Northern Ethiopia  : Analysis of meteorological data and farmers’ perception.
Acad. J. Agric. Res. 1, 88–100. doi:10.15413/ajar.2013.0117.

Haile, G.G., 2015. Estimation of groundwater recharge and potentials under changing climate
in Werii Watershed, Tekeze River Basin. A MSc thesis,Haramaya University,
Haramaya, Ethiopia.

Haile, G.G., Kassa, A.K., 2015. Investigation of precipitation and temperature change
projections in Werii Watershed, Tekeze River Basin, Ethiopia; Application of climate
downscaling model. Journal of Earth Science & Climatic Change 6:300. doi:10.4172/
2157-7617.1000300.

Haregeweyn, N., Tesfaye, S., Tsunekawa, A., Tsubo, M., Meshesha, D.T. segaye, Adgo, E.,
Elias, A., 2015. Dynamics of land use and land cover and its effects on hydrologic
responses: case study of the Gilgel Tekeze catchment in the highlands of Northern
Ethiopia. Environ. Monit. Assess. 187, 4090. doi:10.1007/s10661-014-4090-1.

Healy, R.W., 2010. Estimating Groundwater Recharge, Cambridge University Press.

Helsel, DR., Hirsch, R.M., 2002. Statistical methods in water resourcestechniques of water
resources investigations. Book 4, Chapter A3. USGeological Survey (522 pages).

Herrera-Pantoja, M., Hiscock, K.M., 2008. The effects of climate change on potential
groundwater recharge in Great Britain. Hydrol. Process. 22, 73–86.
doi:10.1002/hyp.6620.

Ho, C.K., Stephenson, D.B., Collins, M., Ferro, C.A.T., Brown, S.J., 2012. Calibration
strategies: a source of additional uncertainty in climate change projections. Bull. Amer.
Met. Soc. 93, 21–26.

http://ccafs-climate.org/bias_correction/ Date of access September 16, 2017.

http://aster.usgs.govDate of access July 10, 2016

21
IPCC, 2013. Summary for Policymakers. In: Climate Change 2013: The Physical Science
Basis. Contribution of Working Group I to the Fifth Assessment Report of the
Intergovernmental Panel on Climate Change, Ipcc. doi:10.1017/CBO9781107415324.

ITRC, 2013. Groundwater statistics and monitoring compliance. ITRC Guid. Doc. GSMC-1,
370.

Jacob, D., 2001. A note to the simulation of the annual and inter-annual variability of the
water budget over the Baltic Sea drainage basin. Meteorol Atmos Phys 77, 61-73, doi:
10.1007/s007030170017.

Jaroslaw, C., Batelaan, O., 2011. Application of the WetSpa distributed hydrological model
for catchment with significant contribution of organic soil. Upper Biebrza case study,
Ann. Warsaw Univ. of Life Sci. –SGGW. Land Reclamation 43, 1, 25–35.

Kulabako, N.R., Nalubega, M., Thunvik, R., 2007. Study of the impact of land use and
hydrogeological settings on the shallow groundwater quality in a peri-urban area of
Kampala, Uganda. Sci. Total Environ. 381, 180–199.
doi:10.1016/j.scitotenv.2007.03.035.

Kebede S. 2013. Groundwater Occurrence in Regions and Basins. In: Groundwater in

Ethiopia. Springer Hydrogeology. Springer, Berlin, Heidelberg

Kumar, C.P., 2012. Climate change and its impact on groundwater resources. Int. J. Eng. Sci.
1, 43–60.

Leander, R., Buishand, T.A., 2007. Resampling of regional climate model output for the
simulation of extreme river flows. J. Hydrol. 332, 487–496.
doi:10.1016/j.jhydrol.2006.08.006.

Legates, D.R., McCabe, G.J., 1999. Evaluating the use of “goodness-of-fit” measures in
hydrologic and hydroclimatic model validation. Water Resour. Res. 35, 233–241.
doi:10.1029/1998WR900018.

Legesse, G., Suryabhagavan, K. V., 2014. Remote sensing and GIS based agricultural
drought assessment in East Shewa zone, Ethiopia. Trop. Ecol. 55, 349–363.

Lewis, K., 2017. Understanding climate as a driver of food insecurity in Ethiopia. Climatic
change, 144, 317-328.

Liu, Y.B., Smedt, F. De, 2004. WetSpa Extension , A GIS-based hydrologic model for flood

22
 prediction and watershed management documentation and user manual. Management 1–
126.

Meze-Hausken, E., 2004. Contrasting climate variability and meteorological drought with
perceived drought and climate change in northern Ethiopia. Clim. Res. 27, 19–31.
doi:10.3354/cr027019.

Moriasi, D.N., Arnold, J.G., Van Liew, M.W., Bingner, R.L., Harmel, R.D., Veith, T.L.,
2007. Model evaluation guidelines for systematic quantification of accuracy in
watershed simulations. Trans. ASABE 50, 885–900. doi:10.13031/2013.23153.

Motiee, H., McBean, E., 2009. An assessment of long-term trends in hydrologic components
and implications for water levels in Lake Superior. Hydrol. Res. 40, 564-579.
doi:10.2166/nh.2009.061.

Ministry of Water, Irrigation and Electricity (MoWIE), 2017. Groundwater resources.

http://www.mowr.gov.et/index.php?pagenum=2.2&pagehgt=880px accessed on
September 15, 2017.

Nash, J.E., Sutcliffe, J.V., 1970. River flow forecasting through conceptual models part I - A
discussion of principles. J. Hydrol. 10, 282–290. doi:10.1016/0022-1694(70)90255-6.

Nyenje, P.M., Batelaan, O., 2009. Estimating the effects of climate change on groundwater
recharge and base flow in the upper Ssezibwa catchment, Uganda. Hydrol. Sci. J. 54,
713–726. doi:10.1623/hysj.54.4.713.

Obuobie, E., 2008. Estimation of groundwater recharge in the context of future climate
change in the White Volta River basin, West Africa. Dissertation, Rheinischen
Friedrich-Wilhelms-Universität Bonn, Germany.

Onoz, B., Bayazit, M. 2012. The power of statistical tests for trend detection. Turkish
Journal of Engineering & Environmental Sciences 27, 247-251.

Pingale, S.M., Khare, D., Jat, M.K., Adamowski, J., 2014. Spatial and temporal trends of
mean and extreme rainfall and temperature for the 33 urban centers of the arid and semi-
arid state of Rajasthan, India. Atmos. Res. 138, 73–90.
doi:10.1016/j.atmosres.2013.10.024.

Pingale, S.M., Khare, D., Jat, M.K., Adamowski, J., 2016. Trend analysis of climatic
variables in an arid and semi-arid region of the Ajmer District, Rajasthan, India. J Lan

23
 Wat Dev 28, 3–18. doi:10.1515/jwld-2016-0001.

Rawls, W.J., Brakensiek, D.L., Saxton, K.E., 1982. Estimation of soil water properties,
Trans. Amer. Soc. of Agric. Engin., 25, 5, 1316-1320, 1328.

Santhi, C., Arnold, J.G., Williams, J.R., Dugas, W.A., Srinivasan, R., Hauck, L.M., 2001.
Validation of the SWATmodel on a largeRwer Basin with point and nonpoint sources. J.
Am. Water Resour. Assoc. 37, 1169–1188. doi:10.1111/j.1752-1688.2001.tb03630.x.

Seyoum, W.M., Milewski, A.M., Durham, M.C., 2015. Understanding the relative impacts of
natural processes and human activities on the hydrology of the Central Rift Valley lakes,
East Africa. Hydrol. Process. 29, 4312–4324. doi:10.1002/hyp.10490.

Subramanya, K., 1998. Engineering hydrology, 2nd edn. Tata McGraw-Hill.

Suryabhagavan, K.V., 2017. GIS-based climate variability and drought characterization in

Ethiopia over three decades. Weather and Climate Extremes 15, 11-23.

Stockdale, T.N., Anderson, D.L.T., Balmaseda, M.A., Doblas-Reyes, F., Ferranti, L.,
Mogensen, K., Palmer, T.N., Molteni, F., Vitart, F., 2011. ECMWF seasonal forecast
system 3 and its prediction of sea surface temperature. Clim. Dyn. 37, 455–471.
doi:10.1007/s00382-010-0947-3.

Tabari, H., Marofi, S., Aeini, A., Talaee, P.H., Mohammadi, K., 2011. Trend analysis of
reference evapotranspiration in the western half of Iran. Agric. For. Meteorol. 151, 128–
136. doi:10.1016/j.agrformet.2010.09.009.

Taylor, R.G., Howard, K.W.F., 1996. Groundwater recharge in the Victoria Nile basin of east
Africa: support for the soil moisture balance approach using stable isotope tracers and
flow modelling. J. Hydrol. 180, 31–53. doi:10.1016/0022-1694(95)02899-4.

Tekleab, S., Mohamed, Y., Uhlenbrook, S., 2013. Hydro-climatic trends in the Abay/Upper
Blue Nile basin, Ethiopia. Phys. Chem. Earth 61–62, 32–42.
doi:10.1016/j.pce.2013.04.017.

Tesfagiorgis, K., Gebreyohannes, T., De Smedt, F., Moeyersons, J., Hagos, M., Nyssen, J.,
Deckers, J., 2011. Evaluation of groundwater resources in the Geba basin, Ethiopia.
Bull. Eng. Geol. Environ. 70, 461–466. doi:10.1007/s10064-010-0338-3.

Tilahun, K., Merkel, B.J., 2009. Estimation of groundwater recharge using a GIS-based
distributed water balance model in Dire Dawa, Ethiopia. Hydrogeol. J. 17, 1443–1457.

24
 doi:10.1007/s10040-009-0455-x.

Walraevens, K., Vandecasteele, I., Martens, K., Nyssen, J., Moeyersons, J., Gebreyohannes,
T., De Smedt, F., Poesen, J., Deckers, J., Van Camp, M., 2009. Groundwater recharge
and flow in a small mountain catchment in northern Ethiopia. Hydrol. Sci. J. 54, 739–
753. doi:10.1623/hysj.54.4.739.

Wang, Z.M., Batelaan, O., De Smedt, F., 1996. A distributed model for water and energy
transfer between soils, plants and atmosphere (WetSpa). Phys. Chem. Earth. 21(3), 189–
193. doi:10.1016/S0079-1946(97)85583-8.

Welde, K., Gebremariam, B., 2017. Effect of land use land cover dynamics on hydrological
response of watershed: Case study of Tekeze Dam watershed, northern Ethiopia. Int.
Soil Water Conserv. Res. 5, 1–16. doi:10.1016/j.iswcr.2017.03.002.

Woldeamlak, S.T., Batelaan, O., De Smedt, F., 2007. Effects of climate change on the
groundwater system in the Grote-Nete catchment, Belgium. Hydrogeol. J. 15, 891–901.
doi:10.1007/s10040-006-0145-x.

Yue, S., Wang, C., 2004. The Mann-Kendall test modified by effective sample size to detect
trend in serially correlated hydrological series. Water Resources Management 18, 201–
218.

Yu, Y., Disse, M., Yu, R., Yu, G., Sun, L., Huttner, P., Rumbaur, C., 2015. Large-scale
hydrological modeling and decision-making for agricultural water consumption and
allocation in the main stem Tarim River, China. Water (Switzerland) 7, 2821–2839.
doi:10.3390/w7062821.

Yenehun, A., Walraevens K., Batelaan, O., 2017. Spatial and temporal variability of
groundwater recharge in Geba basin, Northern Ethiopia. Journal of African Earth
Sciences 134, 198-212. http://dx.doi.org/10.1016/j.jafrearsci.2017.06.006.

Zomlot, Z., Verbeiren, B., Hysman, M., Batelaan, O., 2015. Spatial distribution of
groundwater recharge and base flow: assessment of controlling factors. “in review.” J.
Hydrol. Reg. Stud. 4, 349–368. doi:10.1016/j.ejrh.2015.07.005

List of Figures
25
Figure 1 Location of the study area Tekeze Dam catchment in Ethiopia.
Figure 2 The spatial input database for WetSpa model such as a) slope, b) soil map and c)
land use/land cover (the year 2008) of the study area.
Figure 3 Conceptual framework of the study.
Figure 4 Figure 4 Monthly bias uncorrected, corrected and observed a) precipitation, b)
maximum and c) minimum temperature distribution for Mekelle station for the baseline
period.
Figure 5 Monthly bias uncorrected, corrected and observed rainfall distributions for Michew
station for the baseline period.
Figure 6 Comparison between observed and bias corrected for Michew station precipitation
data for the baseline period (1994-2005).
Figure 7 Observed average temperature (1994-2013) for Mitsebri and Lalibela stations.
Figure 8 Trend of average present and future evapotranspiration for the period 1994-2014 and
2020–2079.

Figure 9 Results of (a) model calibration and (b) validation hydrograph for Tekeze catchment

Figure 10 Average annual a) water balance components (mm) and their future percentage
changes (b&c) from baseline in the Tekeze catchment.

Table 1 Percentage of the missed data of the selected six stations

No. Station Name Data type Missed data (%)

1 Gonder Temperature, Rainfall 5.35, 10.64
2 Laibela Temperature, Rainfall 0.23, 1.80
3 Maichew Temperature, Rainfall 7.10, 15
4 Mekelle Temperature, Rainfall, 0.25, 11.20
5 Samre Temperature, Rainfall, 8.52, 18.36
6 Mitsebri Temperature, Rainfall 5.23,20.50

26
Table 2 Main calibration global parameters and it’s common threshold values of the WetSpa
model
Parameter Description Unit Parameters Unit Value
Interflow scaling Stream network 10
Ki - Cells
parameter delineating threshold
Groundwater Sub-catchments 1000
Kg recession - determining threshold Cells
coefficient value
Relative soil Upstream drained area km2 >0.1
K_ss -
moisture by a particular cell
Correction Total 33
K_ep - Sub catchments
coefficient for PET
Initial groundwater Average sub-catchment km2 909.1
GO Mm
storage area
Maximum Meter 0.005
Average hydraulic
G_max groundwater Mm
radius at upland cells
storage
Base temperature Meter 1.5
o
Average hydraulic
TO for estimating C
radius at outlets
snow melt
Degree day Hour 91
coefficient for
K_snow mm/mm/oC/day Time of concentration
calculating
snow melt
Rainfall degree day Mean travel time for Hour 23
K_rain mm/mm/oC/day
coefficient entire watershed
Surface runoff Manning’s coefficient m-1/3s 0.055
K_run -
coefficient for lowest order
Threshold rainfall Manning’s coefficient m-1/3s 0.025
P_max mm/day
intensity for highest order
Impervious area within % 30
an urban cell
(Source: Liu and De Smedt 2004)

27
 Table 3 Global model parameters and calibration result for the catchment (Liu and De
Smedt, 2004)

Parameter Value range Calibration result

Ki 0-12 3.21
Kg 0-0.06 0.045
K_ss 0-2 1.04
K_ep 0-2 0.5
GO 0-100 15
G_max 0-3000 17.41
TO 0-1 -1
K_snow 0-10 -1
K_rain 0-0.05 -1
K_run 0-5 3.5
P_max 0-500 250

Table 4 Performance of bias corrected method on monthly rainfall amount

Maximum Temperature Minimum Temperature

Std. Mean Abs Mean, Std. Mean Abs Mean,
Error, oC Error, oC o
C Error, oC Error, oC o
C
Observed 2.303 0.035 24.27 2.130 0.321 11.86
RCP
2.797 0.053 35.36 3.520 0.042 22.90
Uncorrected
RCP Corrected 2.303 0.035 24.27 2.130 0.322 11.86

Table 5 Result of trend analysis of annual average precipitation, temperature and

evapotranspiration using MK test
Observed (1994-2013)
Precipitation (mm) Temperature (oC) Evapotranspiration (mm)
Statio Sen’ Sen’
P- P- P-
ns s Kendall Sen’s Kendall s Kendall
S value S value S value
slop ’s Tau slope ’s Tau slop ’s Tau
(Zs) (Zs) (Zs)
e e

28
Miche - 0.005* - 2.73
-0.462 93 0.001* 0.034 0.544 59 0.041 0.345
w 79 * 18.7 8
Mekell - 0.164* - 0.581* - 3.43
-8.5 -0.24 -0.099 75 0.008 0.439
e 49 * 17 * 0.006 8
- 0.143* -
Samre -0.251 49 0.06** 0.031 0.521 19 0.528 0.75 0.112
43 * 7.96
Mitseb 0.783* - 1.76
-9 -0.053 71 0.013* 0.026 0.415 67 0.019 0.392
ri * 3.31 5
Lalibel - 0.489* - 0.058* 1.36
-0.123 55 0.024 0.322 45 0.125 0.263
a 21 * 2.94 * 6
-
Gonde - 0.211* 0.080*
11.1 -0.216 51 0.033 0.298 1 1 0 0.006
r 37 * *
6
Averag - 0.287* 1.67
0.023
e 8.76 * 6
RCP 2.6 (2020-2079)
Precipitation (mm) Temperature (oC) Evapotranspiration (mm)
Sen’ Sen’
Statio P- P- P-
s Kendall Sen’s Kendall s Kendall
ns S value S value S value
slop ’s Tau slope ’s Tau slop ’s Tau
(Zs) (Zs) (Zs)
e e
-
Miche 0.979* 0.865* 0.001 0.06
-5 0.03 -0.003 27 0.016 58 0.709 0.034
w * * * 3
2
- -
Mekell 0.875* 0.001 0.09
31 0.04* 1.91 -0.184 25 0.015 65 0.675 0.038
e * * 7
5 7
34 4.53 0.804* 0.001 0.09
Samre 0.024* 0.203 39 0.023 50 0.749 0.029
7 8 * * 7
-
Mitseb - 0.834* 0.619* 0.002 0.06
0.44 -0.019 77 0.045 56 0.719 0.033
ri 33 * * * 4
5
Lalibel 0.979* 0.709* 0.001 0.07
5 0.08 0.003 58 0.034 91 0.556 0.054
a * * * 7
-
Gonde - 0.666* 0.676* 0.001 - -
0.66 -0.039 65 0.038 0.591 -0.049
r 67 * * * 83 0.08
6
-
Averag 0.001 0.667* 0.05
1.72
e 2 * 3
4
RCP 4.5 (2020-2079)
Precipitation (mm) Temperature (oC) Evapotranspiration (mm)
Sen’ Sen’
Statio P- P- P-
s Kendall Sen’s Kendall s Kendall
ns S value S value S value
slop ’s Tau slope ’s Tau slop ’s Tau
(Zs) (Zs) (Zs)
e e
- -
Miche 0.284* 12 0.414* 0.004 13 0.14
16 1.52 -0.14 0.074 0.37 0.081
w * 6 * * 8 5
5 7
- -
Mekell 0.083* 13 0.374* 0.003 11 0.16
26 1.87 -0.02 0.8 0.476 0.065
e * 7 * * 0 5
6 5
- -
0.108* 18 0.239* 0.006 0.13
Samre 24 3.37 -0.14 0.106 92 0.552 0.054
* 1 * * 8
7 6
-
Mitseb - 0.774* 17 0.263* 0.004 20 0.33
0.47 -0.11 0.101 0.184 0.12
ri 45 * 2 * * 4 3
2
Lalibel - 0.685* - -0.11 21 0.166* 0.004 0.124 16 0.295 0.13 0.094

29
 a 63 * 0.66 3 * * 1 5
8
Gonde 11 1.10 18 0.003 14 0.16
0.440* 0.011 0.4 0.363 0.082
r 9 7 0 * 0 5
-
Averag 0.373*
1.13 0.004 0.18
e *
5
Note: S is an MK test statistic, Zs is MK trend test, slope (Sen’s slope) is the change
(mm)/annual, * and ** indicate significant and non-significant trend at 0.05 significant level,
respectively. A positive value of S indicates that there is an increasing trend and vice versa.

Table 6 Results of the WetSpa model calibration and validation

Run MB R2 NSE
Calibration 0 0.82 0.79
Validation 0.002 0.89 0.76
Optimum 0 1 1

Table 7 Summary of water balance components for the Tekeze catchment

Calibration (1995- Validation (2002-

Average measured data
Water balance 2001) 2004)
(1995-2001)
components
(mm) (mm) (mm)
Precipitation 1050 1125.4 893.3
Surface runoff 32.6 37
Evapotranspiration 1292.22 1192.22 1320
Interflow 29.5 23.8
Groundwater flow 331.2 289.7
Base flow 360.7 313.5
Total runoff 327.89 327.89 285.8

Table 8 Summary of annual water balance changes as compared to the baseline period
Baseline Future RCP 2.6 Future RCP 4.5
(1995-2001) (2020-2079) (2020-2079)
Water balance
Annual Annual Annual
components Change Change
average average average
(%) (%)
(mm) (mm) (mm)
Precipitation 1050 693.54 -20.44 690.61 -20.65
Evapotranspiration 1292.22 1302 0.378 1519.12 8.071
Groundwater
45.23 42.25 -3.41 44.06 -1.31
recharge
Surface runoff 32.5 23.75 -15.56 22.12 -19.00
Base flow 272.4 266 -1.15 269.4 -0.55
30
Figure 1 Location of the study area Tekeze Dam catchment in Ethiopia
 Legend Legend

(a) (b)

Legend

(c)
Figure 2 The spatial input database for WetSpa model such as a) Slope, b) soil map and c) land
use/land cover (the year 2008) of the study area
Figure 3 Conceptual framework of the study
 250

200

Precipitation (mm)
150

100

50

0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Obs Un-corr Corr.

(a)

Obs Un-corr. Corr. 30

45 Obs. Uncorr. Corr.
40 25
Temperature, oC

35
Temperature, oC

20
30
15
25
10
20
15 5

10 0

Nov
Mar
Apr

Jun

Aug
May

July

Oct

Dec
Feb

Sep
Jan
Apr
Mar

May

Aug

Nov
Jun
Jul

Oct
Feb

Sep

Dec
Jan

Month Month

(b) (c)
Figure 4 Monthly bias uncorrected, corrected and observed a) precipitation, b) maximum and
c) minimum temperature distribution for Mekelle station for the baseline period
 250
Jan

200 Feb

Precipitation (mm)
Mar
150 Apr
May
100 Jun
Jul
50
Aug

0 Sep
Obs Un-corr Corr.

Figure 5 Monthly bias uncorrected, corrected and observed rainfall distributions for Michew station
for the baseline period

250
Corrected precipitation (mm)

Corr. Linear (Corr.) R² = 0.9803

200

150

100

50

0
0 50 100 150 200 250
Observed precipitation (mm)

Figure 6 Comparison between observed and bias corrected for Michew station precipitation data for
the baseline period (1994-2005).
 Mitsebri Lalibela
20 20
y = 0.0277x - 36.557
Temperature, oC

Temperature, oC
19.5 19.5 R² = 0.2373

19 19

18.5 y = 0.0322x - 45.501 18.5

R² = 0.3458
18 18
1995 2005 2015 1995 2000 2005 2010 2015
Year Year

Figure 7 Observed average temperature (1994-2013) for Mitsebri and Lalibela stations.

Observed RCP 2.6

1350 1345
Evapotranspiration, mm
Evapotranspiration, mm

1330 1325

1310 1305

1290 1285
y = 1.6803x - 2061.9
1270 1265 y = 0.0365x + 1226.8
R² = 0.2039
R² = 0.0013
1250 1245
1995 2000 2005 2010 2015 2020 2040 2060 2080
Year Year

RCP 4.5 y = 0.2033x + 1102.3

1570 R² = 0.0311
Evapotranspiration, mm

1550

1530

1510

1490

1470

1450
2020 2030 2040 2050 2060 2070 2080
Year

Figure 8 Trend of average present and future evapotranspiration for the period 1994-2014
and 2020–2079
 Preci q-obs q-simu
3000 0

2500
Daily discharge (m3/s)

50

Daily rainfall (mm)

2000
100
1500
150
1000

500 200

0 250

Preci q-obs q-simu

1200 0

1000
50
Daily discharge (m3/s)

Daily rainfall (mm)

800
100
600
150
400

200
200

0 250
01/01/2002 01/01/2003 01/01/2004

Figure 9 Results of (a) model calibration and (b) validation hydrograph for Tekeze catchment
 Annual average components, mm
1600
1400 Baseline (1995-2001)
1200 Future RCP 2.6 (2020-2079)
1000 Future RCP 4.5 (2020-2079)
800
600
400
200
0

a) Water balance components (mm)

-1.15 -0.55

-15.56 -19 -20.65

-20.44

-3.41 8.071

-1.31
Precipitation
0.378
Evapotranspiration
Groundwater recharge
Surface runoff
Baseflow

b) RCP 2.6 (2020-2079) c) RCP 4.5 (2020-2079)

Figure 10 Average annual a) water balance components (mm) and their future percentage changes
(b & c) from baseline in the Tekeze catchment