International Journal of Scientific & Engineering Research Volume 10, Issue 8, August-2019 1448

ISSN 2229-5518

Evaluation of Climate Change Impact on Stream Flow in Upper Awash River Basin,
Ethiopia

Kiya Kefeni1 * Tamirat Dessalegn1 Negese Roba1

1. Department of Water Resources Engineering , School of Civil Engineering and

Architecture, Adama Science and Technology University, Adama, Ethiopia

1*. Corresponding author: [email protected]

Abstract

Studies of climate change impact on water resources are very crucial for planning and
management to alleviate poverty; and for sustainable development specially in developing
countries. This study assessed the likely impacts of climate change on stream flow in the Upper

IJSER
Awash basin of Ethiopia using the Soil and Water Assessment Tool (SWAT). Change of projected
climate variables (temperature and precipitation) for the 2030s (2021-2040) and 2050s (2041-
2060) and estimated impacts of the projected variables on stream flow under RCP4.5 and
RCP8.5 scenarios were analyzed. Future climate data from CORDEX Africa was dynamically
downscaled using single Regional climate models and corrected for bias using delta change
approach. Performance of the SWAT model was evaluated using R2, NSE, and PBIAS. During
calibration (1999-2003) of the model, values 0.75, 0.74 and -7% were obtained for R2, NSE, and
PBIAS. Whereas during validation period (2004-2006) values 0.79, 0.78 and 3.1% were obtained
for R2, NSE, and PBIAS. The finding of this study indicates that streamflow will decrease by
6.51% and 12.33% at 2030s and 2050s under RCP4.5 scenario respectively. Likewise for
RCP8.5 scenario the stream flow will decrease by 10.76% and 26.74% at 2030s and 2050s
respectively.

Key Words: Climate change, CORDEX Africa, RCM, SWAT, Upper Awash

IJSER © 2019
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 10, Issue 8, August-2019 1449
ISSN 2229-5518

Introduction

One of the most significant potential consequences of climate change in the long-term
would be changes in regional hydrological cycle. According to Bates et al., 2008 the global
climate change has the potential to impose additional pressures on water availability. The main
contributors in the significant changes in global climatic patterns are the increasing
concentrations of atmospheric greenhouse gases (GHGs), which subsequently leads to global
warming (IPCC, 2014). Zhi and Jiming , 2017 also conclude that the increase in concentration
of greenhouse gases (GHG) seems to be one of the major driving forces behind the climate
change. Evapotranspiration and precipitation are the two vital hydrologic variables that can be
affected by changing temperature. For instance, rising temperature will have a major impact
on the magnitude and frequency of extreme precipitation events in some regions (Zhi and
Jiming , 2017). Based on different climate change scenarios the temperature is expected to

IJSER
rise with two or more degree centigrade, rainfall predicted to increase in some places and
to decrease in other parts (IPCC, 2014)

Developing countries, such as Ethiopia, will be more vulnerable to climate change mainly
because of the larger dependency of their economy on agriculture which is very sensitive to
climatic variations (Kassie et al., 2013). Therefore, assessing and evaluating the significance of
climatic change impacts on water availability of the basin at catchment and sub-basin scale is
crucial for water resource development.

Climate models are the main tools for developing projections of climate change in the
future. Global climate models (GCMs) are one suitable tool for the assessment of climate
variability and change (Claudia and Jan, 2010). Current GCMs have spatial resolution on the
order of 100– 250 km and have the potential to simulate the main characteristics of general
circulation at the range of this scale (Almseged and Rientjis, 2015; Claudia and Jan, 2010).
However, GCMs are not capable of capturing the detailed processes associated with regional
climate variability and changes (Taye et al., 2018). Hence, a Regional Climate Model (RCM) is
the best tool for dynamic downscaling of climate features in order to make predictions for a
particular region (Raneesh, 2011). Also finding of Taye et al., 2018 proves that regional climate
models (RCM) are better suited for regional impact studies.

IJSER © 2019
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 10, Issue 8, August-2019 1450
ISSN 2229-5518

Simulations of watershed hydrology are extensively used for water resources planning and
management. Generally, to utilize water in a sustainable manner, it is necessary to understand the
quantity and quality in space and time through studies. Therefore, the objective of this study is to
evaluate the impacts of climate change on stream flow of the upper Awash sub-basin using semi-
distributed SWAT model.

2. MATERIALS AND METHODS

2.1 Description of the Study area

The proposed sub-basin is located in the western highland part of the Awash river basin and
covers part of the basin above Hombole gauging station, including the capital city Finfinne
(Figure 2.1). The geographic location of the basin is between latitudes of 7°53’N and 12°N and
longitudes of 37°57’E and 43°25’E (Tajin et al., 2016). Awash River rises on the high plateau

IJSER
near Ginchi town west of Finfinne and flows along the rift valley into the Afar triangle, and
terminates in salty Lake Abbe on the border with Djibouti. The total length of the main course is
about 1200 km and it is the principal stream of an endorheic drainage basin covering
parts of the Oromia, Somali, Amhara, and Afar region (Yitea,2015)

IJSER © 2019
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 10, Issue 8, August-2019 1451
ISSN 2229-5518

IJSER
Figure 2-1 Location of upper Awash sub-basin.
The physical settings of the study area are characterized by the heterogeneity of the large natural
systems such as orographic groups, the high plains, mountains and plateaus (Yitea, 2015). Based
on physical and Socio-economic factors the Upper Awash Basin is topographic level of all lands
above 1500 m. The climate of the Awash basin is humid to sub-humid in the highlands and semi-
arid to arid in the rift valley. Annual rainfall ranges from 850mm to 1000mm in the plain area
and mountains of Upper Awash River basin respectively. Mean annual temperature is about
15°C in the highlands and around 21°C in the lowlands. The Sub-basin receives approximately
70% -75% of its annual rainfall during the wet season which covers the months June–September
(Daba, 2014).

IJSER © 2019
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 10, Issue 8, August-2019 1452
ISSN 2229-5518

2.2 Model Description

There are a number of integrated physically based distributed models. Among which, researchers
(Neitsch et al., 2011; Gassman et al., 2005) have identified SWAT as one of the most promising
and computationally efficient model. SWAT is a semi-distributed watershed model developed by
the Agricultural Research Service of the United States Department of Agriculture (USDA) to
simulate the impact of land-use and management practices on the quantity and quality of water
and to quantify sediment and agricultural chemical yields in large and complex watersheds with
changing soils, land use, and management conditions over long time periods (Neitsch et al.,
2011). SWAT is a physically based, basin-scale, spatially distributed, continuous daily time step
and computationally efficient hydrological model. SWAT is currently applied worldwide and
considered as a versatile model that can be used to integrate multiple environmental processes,
which support more effective watershed management and the development of better informed

IJSER
policy decision (Gassman et al., 2005).

The SWAT hydrological cycle simulation is based on the water balance equation (Neitsch et al.,
2011) shown in Equation (1).

SWt = SWo+∑ni=1(Rday, i − Qsurf, i − Eact, i − Wseep, i − Qlat, i) (1)

Where;
SWt is the final soil water content (mm); SWo is the initial soil water content on day i(mm); Rday
is the amount of precipitation on day i(mm); Qsurf is the amount of surface runoff on day
i(mm); Eact is the actual evapotranspiration on day i(mm; Wseep is the amount of water entering
the vadose zone from the soil profile on day i(mm water); Qlat is amount of return flow on day
i(mm) and t is the time (days).

Based on watershed physical characteristics either SCS CN method or Green and Ampt
infiltration method has been suggested for SWAT to generate runoff (Reza Kabiri, 2014).
Because of the unavailability of sub daily weather data for Green and Ampt method, SCS CN
method was used in this study to predict surface runoff from watershed.

IJSER © 2019
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 10, Issue 8, August-2019 1453
ISSN 2229-5518

2.2.1 Model Inputs

In order to run the SWAT model, it was necessary to prepare several sets of spatially distributed
data. For this study, spatial data include elevation, soil type, and land use/land cover were
collected from Ethiopian Ministry of Water, Irrigation and Energy (MOWIE).

The temporal input data for the model are meteorological and stream flow datasets. For this
study, the meteorological data elements such as daily precipitation, minimum and maximum air
temperature, relative humidity, wind speed and sunshine hours for thirty years (1985-2014) of
six representative weather were obtained from the National Meteorology Agency (NMA).
Station those have missing meteorological data values were filled using weather generator model
(WXGEN). Other parameters required in user weather generator were calculated using
pcpSTAT.exe and dew02.exe computer program. Stream flow data of the Hombole gauging

IJSER
station from (1995-2013) was collected from Ministry of Water, Irrigation and Electricity
(MoWIE) and were used for performing calibration and validation of the model.

2.2.2 Model Calibration, Validation and performance evaluation

In SWAT model calibration, the parameters are adjusted within their physical acceptable range in
such a way that practical agreement between simulated and observed stream flows is
accomplished. To perform parameter calibration and uncertainty analysis different programs are
introduced. SWAT-CUP is one of the programs which are currently used by different
researchers. It is a public domain and used for calibration, uncertainty or sensitivity analysis
procedures linked with SWAT. It enables sensitivity analysis, calibration, validation and
uncertainty analysis of SWAT models. Automatic calibration and uncertainty analysis
incorporated in SWAT 10.4 through SWAT-CUP software developed and tested by Abbaspour
(2007), with the semi-automated program SUFI2 was used for this study.
During calibration and validation of a hydrological model it is necessary to assess the
performance of the model. This is done by statistically comparing the model output and
observed values using various statistical measures. Model evaluation is normally based on its
ability to simulate major hydrological processes in a watershed. Model performance assessment
is normally carried out by comparing the model predictions at the basin outlet with the
corresponding observed records (Moriasi et al., 2007). Evaluation of model performance
involves assessing the ‘goodness-of-fit’ of the simulated and the observed hydrological variables

IJSER © 2019
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 10, Issue 8, August-2019 1454
ISSN 2229-5518

such as stream flow. There are a large number of model performance measures available to a
hydrologist. For this study, three model performance evaluation techniques such as; coefficient
of determination, Nash-Sutcliffe coefficient and PBIAS recommended by Moriasi et al. (2007)
were used.

3. Result and Discussions

3.1 Model calibration and Validation

3.1.1 Calibration

Once the sensitive parameters for the model are identified, the next step is to calibrate and
validate the model. Both manual and automatic calibration (using SUFI-2) calibration procedures
were used in this study. For model calibration five years Awash River data (1999-2003) at

IJSER
Hombole gauging station and two year data’s (1997-1998) was used for model warm up. The
performance of the model was evaluated using R2, ENS and PBIAS statistical measures.
Evaluations were performed at monthly time scales and the result of statistical parameters during
calibration obtained were 0.75 for R2, 0.74 for ENS and -7% for PBIAS. The values indicate that
there is good agreement between observed and simulated stream flow. Figure 3.1 below shows
hydrograph comparison during model calibration.

Simulated Observed
500.00
400.00
simulated flow (m3/s)

300.00
200.00
100.00
0.00
Jul-99

Jul-00

Jul-01

Jul-02

Jul-03
Apr-99

Apr-00

Apr-01

Apr-02

Apr-03
Jan-99

Oct-99
Jan-00

Oct-00
Jan-01

Oct-01
Jan-02

Oct-02
Jan-03

Oct-03

Month

Figure 3-1 Hydrograph comparison during model calibration

IJSER © 2019
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 10, Issue 8, August-2019 1455
ISSN 2229-5518

3.1.2 Validation
Validation of the model results is necessary to increase user confidence in model predictive
capabilities. Three year monthly basis data (2004-2006) were used for model validation without
any adjustment of fitted value during calibration and values of 0.79 for R2, 0.78 for ENS and 3.1%
for PBias were obtained. Values obtained for performance indices shows good agreement as they
are within recommended values by Moriasi et al., (2007). Figure 3.2 illustrates hydrograph
during validation.

simulated observed

500.00
simulated flow (m3/s)

400.00
300.00
200.00
100.00

IJSER
0.00

Month

Figure 3-2 Hydrograph comparison during model validation

Generally, according to Moriasi et al., (2007) model performance in terms of replicating the
observed hydrograph is acceptable during calibration and validation. With this performance the
model under estimated the observed stream flow compared to the simulated mean monthly
stream flow in 2000 during calibration and 2003and 2004 during validation period. One of the
factors that have contributed to uncertainty of the model might be the effect of the SWAT
parameters that are considered to have negligible influence on the stream flow but cumulative of
which would have affected the model performance. Unconsidered factors in modeling processes
that resulted in the model error are also other factors.

3.2 Projected changes in mean monthly, seasonally and annual stream flow

Future stream flow was simulated for two scenario periods 2030’s (2021-2040) and 2050’s (2041-
2060) considering river flow (1997-2006) as baseline flow. The stream flow projection takes
place by assuming the projected precipitation and temperature inputs to SWAT and also
considering land use/land cover, soil and other parameters similar to present in the future. The
mean monthly and annual percentage change in flow for both (RCP4.5 and RCP8.5) scenarios

IJSER © 2019
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 10, Issue 8, August-2019 1456
ISSN 2229-5518

for the period 2030s and 2050s are presented in Figure 3.3a and b respectively. The percentage
change of mean monthly flow increases for months January, February, March, April, July,
August, November and December and also decreases in months May, June, September and
October is observed for both 2030s and 2050s under RCP 4.5 scenario.

The mean monthly stream flow result shows that future monthly stream flow decreases in all
months and annual during 2030’s and 2050’s, under RCP 8.5 scenario. Dile et al., 2013 conclude
that decrease in flow volume observed in months which showed a decrease in monthly rainfall.
Generally results showed that the impact of climate change cause a decrease in mean monthly
flow volume between -0.89% to –18.01% during 2030s and between -0.44 to -32.98% during
2050s under RCP4.5 scenario.

a)Scenario RCP4.5
40.00

IJSER
30.00
Change of stream flow (%)

20.00

10.00

0.00

-10.00

-20.00

-30.00 2030 RCP4.5 2050 RCP4.5

-40.00

b) Scenario RCP8.5
5.00
Change of stream flow

-15.00
(%)

-35.00

-55.00
2030 RCP8.5 2050 RCP8.5

Figure 3-3 Mean monthly and annual percentage change of total flow pattern of Upper Awash
sub-basin

IJSER © 2019
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 10, Issue 8, August-2019 1457
ISSN 2229-5518

The projected mean monthly stream flow decreases in range between 10.75 % to 45.61% under
RCP4.5 scenarios in 2030s and 2050s respectively. Similarly for RCP8.5 scenarios the projected
mean monthly stream flow decreases in range between 15.95% to 52.78% in 2030s and 2050s
respectively.

At annual basis generally, decreasing trend of stream flow was observed in the study area at
2030s and 2050s under RCP 4.5 and RCP 8.5 scenarios. As can be noticed from figure (3 a and b)
the projected annual stream flow decreased by 6.51% and 12.33% under RCP 4.5 scenarios in
2030s and 2050s respectively. Similarly for RCP8.5 the projected annul stream flow decreased by
10.76% and 26.74% in 2030s and 2050s respectively. Overall, decreasing pattern of the average
total annual flow in future is mainly because of a decrease in average monthly flow in RCP 8.5
scenarios and RCP4.5. This is caused due to the fact that higher increment of projected maximum
and minimum temperature as well as significant increase in evapotranspiration in the future over

IJSER
Upper Awash sub-basin. Also the decreasing pattern of the average total annual flow in future is
higher in RCP8.5 than RCP4.5 due to high concentration of greenhouse gas, low technology
development and no police implementation for climate in RCP8.5. This cause’s high
concentration of heat occurred in this scenario in future.

The seasonal variation of streamflow of the projected climate from the baseline period was compu
ted for winter, spring, summer and autumn seasons (Table3.1). The seasonal projected stream
flow shows increasing in all seasons for the first time horizon (2030s) under RCP4.5 scenario.
Also for the second time horizon (2050s) the projected seasonal stream flow shows increasing
except for autumn (MAM) season which shows decrease by -7.66% under RCP4.5 scenario.
However, for RCP8.5 scenario the projected seasonal stream flow shows decreasing during both
time horizon (2030s and 2050s).

IJSER © 2019
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 10, Issue 8, August-2019 1458
ISSN 2229-5518

Table 3-1 Percentage change of projected seasonal stream flow pattern of Upper Awash sub-
basin
Time horizon
Scenario Season 2030s 2050s
Winter (DJF) 25.11 16.21
Spring (SON) 2.07 2.27
RCP4.5 Summer (JJA) 13.79 0.45
Autumn (MAM) 3.41 -7.66
Winter (DJF) -11.91 -17.29
Spring (SON) -25.98 -29.07
RCP8.5 Summer (JJA) -39.15 -44.04
Autumn (MAM) -30.79 -36.98

IJSER
When comparing the winter (dry) season with the summer (wet) season; the dry season average
projected stream flow increase by 25.11% and 16.21% in 2030s and 2050s respectively under
RCP4.5. Also the wet season average projected stream flow increase by 13.79% and 0.45% in
2030s and 2050s respectively. Similar findings were captured by Daba , 2014.

4. Conclusions

Projected hydrological variability is important for future resource and hazard management. This
study analyzed the response of stream flow to possible future climate change predicted using an
Ensemble of three GCMs driving models and single RCM with dynamically downscaled
approach. Physically based semi-distributed hydrological model (SWAT) was used to determine
the impact of climate change on stream flow. Dynamically downscaling method was used to
downscale the projected global rainfall and temperature for the 2030s and 2050s using data from
three Ensemble GCMs driving models. Simulation of the changes in stream flow (runoff) was
done using the soil and water assessment tool (SWAT) model. The impacts of the projected
climate variables on stream flow show that the stream flow will decrease by -6.51% and -12.33%
at 2030s and 2050s under RCP4.5 scenario respectively. Similarly for RCP8.5 scenario the
stream flow will decrease by –10.76% and –26.74% at 2030s and 2050s respectively.

IJSER © 2019
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 10, Issue 8, August-2019 1459
ISSN 2229-5518

References

Alemseged, T.H. and Tom, R., 2015. Evaluation of regional climate model simulations of
rainfall over the Upper Blue Nile basin. Atmospheric research, 161, pp.57-64.
Bates, B., Kundzewicz, Z., Wu, S., Palutikof, J., 2008. Climate Change and Water.
Daba, M., 2014. Evaluating Potential Impacts of Climate Change on Surface Water Resource
Availability of Upper Awash Sub-basin, Ethiopia rift valley basin.
IPCC (Intergovernmental Panel on Climate Change), 2014. Rainfall Projection for East Africa.
Fifth Assessment Report: Cambridge University Press, Cambridge, UK.
Dile, Y.T., Berndtsson, R. and Setegn, S.G., 2013. Hydrological response to climate change for
gilgel abay river, in the Lake Tana basin-upper Blue Nile basin of Ethiopia. PloS
one, 8(10), p.e79296
Gassman, W. P., Reyes, M. R., Green, C. H., and Arnold, J. G.: SWAT peer-reviewed literature:

IJSER
a review, Proceedings of the 3rd International SWAT conference, Zurich, 11–15 July, 1–
18, 2005.
Kassie, B.T., Rötter, R.P., Hengsdijk, H., Asseng, S., Van Ittersum, M.K., Kahiluoto, H. and
Van Keulen, H., 2014. Climate variability and change in the Central Rift Valley of
Ethiopia: challenges for rainfed crop production. The Journal of Agricultural
Science, 152(1), pp.58-74.
Taye, M.T.; Dyer, E.; Hirpa, F.A.; Charles, K. Climate Change Impact on Water Resources in
the Awash Basin, Ethiopia. Water 2018, 10, 1560
Moriasi, D. N., J. G. Arnold, M. W. Van Liew, R. L.Bingner, R. D. Harmel and T. L. Veith.
2007. Model Evaluation Guidelines for Systematic Quantification of Accuracy in
Watershed Simulations. Trans.ASABE 50(3):885-900.
Neitsch, S., Arnold, J., Kiniry, J.,Williams, J., 2011. Soil & Water Assessment Tool Theoretical
Documentation Version 2009. Tech. rep. Texas Water Resources Institute, College
Station, TX.
Raneesh, K. Y. & Santosh, G. T. (2011). A study on the impact of climate change on stream
flow at the watershed scale in the humid trophics. Hydrol. Sci. J. 56(6), 946–965.
Reza Kabiri, 2014. Simulation of Runoff using Modified SCS-CN method using GIS system,
case study: Klang Watershed in Malaysia. Research Journal of Environmental Sciences, 8:
178-192.

IJSER © 2019
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 10, Issue 8, August-2019 1460
ISSN 2229-5518

Tajin K., Adane A. & Behilu H. (2016). Study of Water Allocation for Existing and Future
Demands under Changing Climate Condition: Case of Upper Awash Sub River Basin.
Zhi Li and Jiming Jin , 2017. Evaluating climate change impacts on stream flow variability
based on a multisite multivariate GCM downscaling method in the Jing River of China.

IJSER

IJSER © 2019
http://www.ijser.org