Planning and Operation of Hydropower Reservoir in

Tekeze Basin Considering Hydrological Variability

and Climate Change

PhD Dissertation
By
Fikru Fentaw Abera (MSc.)

24th September, 2018

Addis Ababa Institute of Technology

School of Civil and Environmental Engineering
 PLANNING AND OPERATION OF HYDROPOWER
RESERVOIR IN TEKEZE BASIN CONSIDERING
HYDROLOGICAL VARIABILITY AND CLIMATE CHANGE

By
Fikru Fentaw Abera

A Dissertation Submitted to Graduate School of Addis Ababa

University, Addis Ababa Institute of Technology, School of Civil and
Environmental Engineering in partial fulfillment of the requirement
for the degree of
DOCTOR OF PHILOSOPHY
In
CIVIL ENGINEERING

Under the supervision of:

Dr.-Ing Dereje Hailu (Assoc. Prof., Addis Ababa University)

Co-supervision of:
Professor Assefa M. Melesse (Prof., Florida International University)
Dr. Agizew Nigussie (Assis. Prof., Addis Ababa University)

Addis Ababa, Ethiopia

24th September, 2018

ii
 Dissertation Approval Sheet
The thesis committee of
Addis Ababa University (AAU), in Addis Ababa Institute of Technology (AAiT) certifies that
this is the approved version of the following thesis

PLANNING AND OPERATION OF HYDROPOWER RESERVOIR IN TEKEZE BASIN

CONSIDERING HYDROLOGICAL VARIABILITY AND CLIMATE CHANGE

BY:
___________________________ _______________________
Fikru Fentaw Abera Date

APPROVED BY
BOARD OF EXAMINERS:

_______________________________ _______________________
Advisor: - Dereje Hailu (Assoc. Prof.) Date

______________________________ _______________________
Co-advisor: - Assefa M. Melesse (Prof.) Date

______________________________ _______________________
Co-advisor: - Agizew Nigussie (Assis. Prof.) Date

______________________________ ______________________
External Examiner: - Paul Block (Assis. Prof.) Date

_____________________________ _______________________
Internal Examiner 1: - Seifu Tilahun (Assoc. Prof.) Date

_____________________________ ________________________
Internal Examiner 2: - Yilma Seleshi (Assoc. Prof.) Date

__________________________ _____________________
Chairperson: - Agizew Nigussie (Assis. Prof.) Date

iii
 DECLARATION

I at this moment declare that I have prepared this Doctoral dissertation independently and that

only those sources, aide, and advisors that are duly noted and cited herein have been used and

consulted. I had fully mentioned and referenced all materials and results that are not original to

this work.

Fikru Fentaw Abera …………………… ………………….

Name of author Signature Date

iv
 DEDICATION
This work dedicated to:

My father, Fentaw Abera, without you, this chapter of my life would never have been

written. We loved you so much.

The wise females that are always at my side: my wife (Woinshet), my little daughter

(Esetemariam) and the soul of my mother (Asrebeb)

v
 ABSTRACT OF THE DISSERTATION

Reservoirs are built to manage surface water resources used to balance between water resources

and demands. Water in the arid, semi-arid and other climate regions like Ethiopia has available

finite, but the demand increases. Hence reservoirs should be optimally operated in this semi-

arid area to use water most efficiently. This study was conducted to assess hydrological

variability and climate change impacts on the operation of Tekeze hydropower reservoir in

Tekeze River basin. In this basin annual and inter-annual climate variability of precipitation

and climate change uncertainties are present challenges for water resources planning and

management.

This study contributed to provide a scientific basis for the changing characteristics of

precipitation and streamflow of Tekeze basin. The research is carried out using Mann-Kendall

(MK) test and Indicators of Hydrologic Alteration (IHA) to determine the potential trends and

variability of rainfall and streamflow. The trend in mean monthly precipitation data shows

increasing trends in the South-eastern part of the Tekeze basin whereas decreasing trend in the

North-western part of the basin. The streamflow trend analysis also showed a significant

increasing trend during wet seasons, in contrast dry season showed a significant decreasing

trend. Indicators of Hydrologic Alteration (IHA) used to evaluate hydrologic variables and

trends generated from daily streamflow data at Embamadre due to Tekeze hydropower reservoir

operation. The significant results in the streams downstream of Tekeze hydropower reservoir

had increased minimum flow, decreased maximum flows, decreased rise and fall rates, reduced

wet season monthly flow. All these hydro-climatological variability influence Tekeze

hydropower reservoirs planning and operation. This would be valuable for the water managers

and decision makers to make better decision on integrated water resources management and

ecological environment assessment in the future.

vi
Assessment of Tekeze River basin water resources undertaken through the application of bias-

corrected ensemble COordinated Regional climate Downscaling Experiment in African domain

(CORDEX-Africa) Regional Climate Models (RCMs) under Representative Concentration

Pathway RCP4.5 and RCP8.5 climate scenarios in the 2020s, 2050s and 2080s used as an input

to Soil and Water Assessment Tool (SWAT).This study compares the performance of different

bias correction methods successfully, in particular, the result of both precipitation and

temperature were greatly improved by the distribution mapping (DM) bias correction. Results

showed that an increase in rainfall and temperature in the future would be critical to future

inflow in Tekeze hydropower reservoir, with rainfall variability having a more significant

impact than temperature variability. For all future time periods, both the scenarios produce

moderate increases in annual and seasonal streamflow. Due to the future changes of streamflow

which is an input for reservoir planning and operation, climate change scenarios should

incorporate into the operation of hydropower dams and reservoirs in Tekeze basin.

US Army Corps of Engineer’s Reservoir Evaluation System Perspective Reservoir Model

(HEC-ResPRM), a network flow based reservoir system operations optimization monthly

model was used to reproduce optimum hydropower reservoir storage, release and water level

on Tekeze hydropower reservoir. In current operation, HEC-ResPRM well optimized Tekeze

hydropower reservoir and showed an increase in power storage, pool level and release

compared to current actual hydropower reservoir operation status. Future optimized power

storage in the operational Tekeze hydropower reservoir is expected to increase up to 25% and

30% under RCP4.5 and RCP8.5 climate scenarios, respectively. This result indicates that

Tekeze hydropower reservoir operation affected by future climate change. Therefore, it is better

to improve this reservoir operation before investing into newly planned hydropower reservoirs

in the basin. This analysis shows that integrating hydrological variability and future climate

vii
scenarios into optimal hydropower reservoir operation and other water resources planning and

development are essential in Tekeze river basin and other river basins projects in Ethiopia.

Keywords: Hydrological variability, SWAT, CORDEX-Africa, RCP, Climate Change, HEC-

ResPRM, Hydropower, Reservoir operation, Optimization, Tekeze Basin

viii
 ACKNOWLEDGEMENTS

Above all and everything, I thank Almighty God the Father, God the Son and God the Holy

Spirit, whose love, grace and unity has been with me throughout my life and helped me to

achieve my dreams. I also thank saint merry, Angels, saints, and monks for their help

throughout my life and completion of this research.

I owe my gratitude to all those people who have made this dissertation possible and because of

whom my PhD experience has been one that I will cherish forever.

I gratefully acknowledge my supervisors Dr.-Ing Dereje Hailu, Professor Assefa M. Melesse

and Dr. Agizew Nigussie for their inspiration, competent guidance, patience, and

encouragement through all the different stages of this dissertation. Special thanks to Professor

Assefa, who consistently advised me to target publishing and financially support. I thank all

my advisors for guiding me to conduct my PhD research in a challenging and enjoyable topic

of hydropower reservoir operation under hydrological variability and climate change.

I want to acknowledge the assistance of the Ethiopian Meteorological Service Agency, the

Ethiopian Ministry of Water, Irrigation and Electricity and Ethiopian Electric Power

Corporation in providing data for the research. I also would like to give my thanks to Florida

International University assisting for two months’ research visit.

Thanks to all the friends who shared the PhD experience with me. My special gratitude is

extended to my friends Bahru Mekuria and Abebe Arega with whom we passed all the good

and bad times together during the research journey. I would like to give special thanks to the

staffs at the school of Civil and Environmental Engineering of AAiT, Addis Ababa University

for smoothly facilitating my PhD work throughout my study period. I am also thankful for all

the support I received from my colleagues at Florida International University, especially the

Earth and Environment Department who helped me at a research visit such as Anteneh and

Abraham. Thanks!

ix
I would also like to remember the encouragement provided by my parents without their support;

I would not have been able to come a long way. My heartfelt thanks are also due to my brothers,

sisters, and other family members and friends for their support and encouragement. Special

thanks to Fentaw, Setargie, Demssie, Mengistu, Ketema, Tesfaye and Eden for their help and

patience, for every period I was away.

Finally, I would like to thank my wife, Woinshet Tefera, and daughter, Esetemariam, for their

loving support, endless patience, encouragement and understanding during this period. They

had always supported my work even if I used to take ‘family time’ to research and used to be

absent at times when they needed me most. Words will never adequately express my gratitude

to you for this. Esete, you are my real sources of inspiration. My little son Mikiyas, you are so

lucky that you were born during the final phase of my PhD study that I have more time for you

during your development stage.

Fikru Fentaw Abera

24th September, 2018

Addis Ababa, Ethiopia

x
PREFACE

The present thesis, entitled “Planning and Operation of Hydropower Reservoir in Tekeze Basin

considering Hydrological Variability and Climate Change” has been submitted and defended

as one of the requirements for the PhD Degree at the School of Civil and Environmental

Engineering (Addis Ababa Institute of Technology) under the supervision of Dr.-Ing Dereje

Hailu, Professor Assefa M. Melesse and Dr. Agizew Nigussie. The PhD project run from

December 2013 to July 2018 and was funded by Ethiopian Ministry of Education.

The content of the PhD thesis is based on four papers published in scientific journals. In the

text, the papers are referred to by their appendix number written with Arabic numbers.

1. Fikru Fentaw, Dereje Hailu, Agizew Nigussie and Assefa M. Melesse, (2018), Rainfall and

Streamflow Variability in Tekeze River Basin, Ethiopia, a chapter in the book titled:

Extreme Hydrology and Climate Variability: Monitoring, Modeling, Adaptation and

Mitigation, Assefa M. Melesse et al. (Eds). Elsevier Publisher, (Accepted for publication)

2. Fikru Fentaw, Assefa M. Melesse, Dereje Hailu, Agizew Nigussie (2018), Impacts of

Climate Changes on the Water Resources of Tekeze River Basin part of Eastern Nile,

Ethiopia; Geophysical Research, Vol.20, EGU2018-3102 Conference Paper.

3. Fikru Fentaw, Dereje Hailu, Agizew Nigussie, Assefa M. Melesse (2018), Climate Change

Impacts on the Hydrology of Tekeze Basin, Ethiopia: a projection of Rainfall-Runoff for

Future Water Resources Planning, Water Conservation Science and Engineering journal,

Springer, doi:10.1007/s41101-018-0057-3

4. Fikru F. Abera, Dereje H. Asfaw, Agizew N. Engida and Assefa M. Melesse (2018),

Optimal Operation of Hydropower Reservoir under Climate Change: the case of Tekeze

Reservoir, Eastern Nile; Water2018, 10(3), 273; doi:10.3390/w10030273

xi
 TABLE OF CONTENTS

Content Page

ABSTRACT OF THE DISSERTATION .............................................................................. vi

ACKNOWLEDGEMENTS .................................................................................................... ix
PREFACE ................................................................................................................................ xi
TABLE OF CONTENTS .......................................................................................................xii
LIST OF TABLES ................................................................................................................. xiv
LIST OF FIGURES ................................................................................................................ xv
LIST OF ACRONYMS .......................................................................................................xviii

1. INTRODUCTION................................................................................................................ 1
1.1 Background ..................................................................................................................... 1
1.2 Statement of the Problem ................................................................................................ 9
1.3 Objectives and Questions of the Research ................................................................... 12
1.4 Dissertation Outline ..................................................................................................... 14

2. DESCRIPTION OF THE STUDY AREA ....................................................................... 16

2.1 Location and Topography ............................................................................................ 16
2.2 Climate and Hydrology ................................................................................................ 17
2.3 Soil and Land use ......................................................................................................... 18
2.4 Water Resources Potential ........................................................................................... 21

3. RAINFALL AND STREAMFLOW VARIABILITY IN TEKEZE RIVER BASIN,

ETHIOPIA1 ............................................................................................................................. 24
3.1 Introduction .................................................................................................................. 25
3.2 Material and Methods .................................................................................................. 30
3.2.1 Datasets used ........................................................................................................ 30
3.2.2 Trend analysis methods ........................................................................................ 33
3.2.3 Indicators of Hydrologic Alteration ..................................................................... 38
3.3 Result and Discussion ................................................................................................... 41
3.3.1 Preliminary Analysis ............................................................................................ 41
3.3.2 Trends of precipitation.......................................................................................... 42
3.3.3 Streamflow Trend and Variability ........................................................................ 48
3.3.4 Influence of Tekeze reservoir operation on the streamflow regime ..................... 49
3.4 Conclusive Remarks ..................................................................................................... 58

xii
4. CLIMATE CHANGE IMPACT ON THE WATER RESOURCES OF TEKEZE
BASIN2,3 ................................................................................................................................... 60
4.1 Introduction .................................................................................................................. 61
4.2 Study Area and Datasets............................................................................................... 66
4.2.1 Study Area ............................................................................................................ 66
4.2.2 Dataset Used ......................................................................................................... 69
4.3 Methodology ................................................................................................................. 75
4.3.1 Bias Correction Method........................................................................................ 76
4.3.2 Hydrological Modeling Using SWAT .................................................................. 84
4.4 Results and Discussions................................................................................................ 92
4.4.1 Evaluation of bias correction methods ................................................................. 92
4.4.2 Change of Projected Precipitation and Temperature ............................................ 97
4.4.3 SWAT Model Performance ................................................................................ 102
4.4.4 Streamflow Projection ........................................................................................ 103
4.4.5 Climate Change Impact on Water Resources Planning ...................................... 107
4.5 Conclusive Remarks ................................................................................................... 109

5. TEKEZE HYDROPOWER RESERVOIR OPERATION UNDER CLIMATE

CHANGE4 ............................................................................................................................. 111
5.1 Introduction ................................................................................................................ 112
5.2 Material and Methods ................................................................................................ 116
5.2.1 Study Area .......................................................................................................... 116
5.2.2 Datasets used ...................................................................................................... 117
5.3 Methods ...................................................................................................................... 118
5.3.1 HEC-ResPRM Optimization Model ................................................................... 118
5.3.2 HEC-ResPRM Model Setup ............................................................................... 122
5.4 Results and Discussions.............................................................................................. 125
5.4.1 Impacts of climate change on reservoir inflow .................................................. 125
5.4.2 Current Reservoir Operation .............................................................................. 125
5.4.3 Reservoir Operation under Climate Change ....................................................... 126
5.5 Conclusive Remarks ................................................................................................... 136
6 CONCLUSIONS AND RECOMMENDATIONS ......................................................... 138
6.1 Conclusions ................................................................................................................ 138
6.2 Recommendations ....................................................................................................... 141
REFERENCES ..................................................................................................................... 143
APPENDICES ....................................................................................................................... 161

xiii
 LIST OF TABLES
Table Page

Table 3.1 Location of weather stations and lengths of precipitation series used in this study . 31

Table3.2 Summary of hydrologic parametres used in the IHA and their charactrestics .......... 40

Table 3.3 Annual precipitation time series basic statistical properties of the study area ......... 42

Table 3.4 None parametric RVA scores at Embamadre station of Tekeze River .................... 53

Table 4.1 Order of sensitive parameters and their optimal value ............................................ 89

Table 4.2 : SWAT model performance evaluation criteria(Moriasi et al., 2007) .................... 91

Table 4.3 Frequency based statistics of daily observed, raw RCM simulated (raw) and bias

corrected precipitation at the selected weather stations of Tekeze basin ................................. 93

Table 4.4 SWAT hydrological model monthly performance under validation and calibration

periods in Tekeze Basin .......................................................................................................... 102

Table 4.5 Mean annual and seasonal streamflow change (%) in Tekeze basin at Embamadre

from the baseline period (1994-2008) .................................................................................... 105

Table 5.1Mean annual optimized power storage under climate change scenarios. ................ 129

Table 5.2 Mean annual optimized pool level variation under RCP4.5 and RCP8.5 climate

scenarios in three future time periods. .................................................................................... 132

xiv
 LIST OF FIGURES
Figure Page

Figure 2.1 Location of Tekeze basin ........................................................................................ 17

Figure 2.2Tekeze basin location, climate and streamflow stations including planned cascade

hydropower reservoir ................................................................................................................ 23

Figure 3.1 Results of Mann Kendall annual precipitation time series trend test at 95%

confidence interval ................................................................................................................... 43

Figure 3.2 Annual precipitation trend magnitudes at different stations of Tekeze basin......... 44

Figure 3.3 Percentage annual and seasonal trend test results a) Overall negative and positive

trends b) significant positive and negative trends at 95% confidence level ............................. 44

Figure 3.4 Mann Kendall Z values at different Tekeze basin stations ..................................... 46

Figure 3.5Results of seasonal precipitation trend magnitude at different weather stations ..... 47

Figure 3.6 Annual precipitation increasing and decreasing trends in the Upper parts of Tekeze

basin .......................................................................................................................................... 47

Figure 3.7 Comparison of median monthly streamflow before and after Tekeze hydropower

reservoir dam construction ....................................................................................................... 50

Figure 3.8 November and August monthly median flow change before and after Tekeze

hydropower reservoir construction in Tekeze river .................................................................. 51

Figure 3.9 Hydrological alteration of 3-, 7-, 30-, and 90-day minimum flow before and after

Tekeze hydropower reservoir construction in Tekeze river ..................................................... 56

Figure 3.10 Hydrological alteration of 3- 7-, 30-, and 90-day maximum flow before and after

Tekeze hydropower reservoir construction in Tekeze river ..................................................... 57

Figure 4.1 Location of upper Tekeze basin upstream of Embamadre gauging station ............ 68

Figure 4.2Sub-basins of Tekeze basin study area watershed delineated by SWAT ................. 68

xv
Figure 4.3Mean monthly precipitation (mm) and maximum temperature (Tmax) & minimum

temperature (Tmin) (◦C) for selected stations of Tekeze basin ................................................. 69

Figure 4.4Major land use/land cover map of Tekeze Basin study area re-classified by

SWAT ....................................................................................................................................... 71

Figure 4.5 Major soil map of Tekeze Basin study area re-classified by SWAT ...................... 71

Figure 4.6 Conceptual diagram of the SWAT modeling process with climate change

scenarios ................................................................................................................................... 75

Figure 4.7 Comparison of climate scenario data before and after bias correction with respect

to observed data at (a) Lalibela, (b) Debretabor, (c) Gonder, and (d) Mekele selected

meteorological stations of Tekeze basin ................................................................................... 95

Figure 4.8 Observed, raw RCP simulated and bias corrected values of precipitation at the four

selected stations of Tekeze basin .............................................................................................. 96

Figure 4.9 Mean annual and seasonal precipitation variation over Tekeze basin under (a)

RCP4.5, and (b) RCP8.5 climate scenarios .............................................................................. 98

Figure 4.10 Long terms mean annual precipitation trend at Tekeze basin under RCP4.5 and

RCP8.5 climate scenarios ......................................................................................................... 98

Figure 4.11 Change of monthly precipitation amount for the selected scenarios and projected

periods. ..................................................................................................................................... 99

Figure 4.12 Mean annual and seasonal temperature projection in Tekeze basin under (a)

minimum temperature change at RCP4.5 (b) minimum temperature change at RCP8.5 (c)

maximum temperature change at RCP4.5 (d) maximum temperature change at RCP8.5

climate scenarios..................................................................................................................... 101

Figure 4.13 Rates of change of monthly mean temperature for the selected scenarios and

projected periods..................................................................................................................... 101

xvi
Figure 4.14 Observed and simulated monthly streamflow hydrographs during (a) Calibration

and (b) validation periods. ..................................................................................................... 103

Figure 4.15 Mean monthly and annual streamflow changes under RCP4.5 and RCP8.5

climate scenarios. ................................................................................................................... 106

Figure 5.1 Location of Tekeze hydropower reservoir and weather stations .......................... 117

Figure 5.2 HEC-ResPRM schematization of Tekeze Hydropower reservoir operation......... 123

Figure 5.3 Annual Tekeze hydropower reservoir inflow trend for future time periods ......... 125

Figure 5.4 Mean monthly reservoir inflow and optimized outflow (release) for future time

periods under: (a) RCP4.5 climate scenario; (b) RCP8.5 climate scenario. .......................... 128

Figure 5.5 Monthly optimized reservoir storage variations under RCP4.5 and RCP8.5 climate

scenarios for: (a) 2020s; (b) 2050s; (c) 2080s. ....................................................................... 130

Figure 5.6 Monthly optimum pool level variations under RCP4.5 and RCP8.5 climate

scenarios in three time periods: (a) 2020s; (b) 2050s; (c) 2080s. .......................................... 135

xvii
 LIST OF ACRONYMS
AR5 IPCC Fifth Assessment Report

AMMA-MIP African Monsoon Multidisciplinary Analysis-Model

Intercomparison Project

CD Coefficient of Dispersion

CDF Cumulative Distribution Function

CMIP3 Coupled Model Intercomparison Project phase 3

CMIP5 Coupled Model Intercomparison Project phase 5

CMIP6 Coupled Model Intercomparison Project phase 6

CN Curve Number

CORDEX Coordinated Regional Climate Downscaling Experiment

DEM Digital Elevation Model

DM Distribution Mapping

ENMSA Ethiopian National Metrological Service Agency

FAO Food and Agriculture Organization

GIS Geographical Information System

GCM Global Climate Model

GUI Graphical User Interface

HEC-DSS HEC’s Data Storage System

HEC-PRM Hydrologic Engineering Center's Prescriptive Reservoir Model

HEC-ResPRM Hydrologic Engineering Center's Reservoir Evaluation System

Prescriptive Reservoir Model

HEC-ResSim Hydrologic Engineering Center Reservoir Simulation

HRU Hydrologic Response Unit

IHA Indicators of Hydrologic Alteration

xviii
IPCC Intergovernmental Panel on Climate Change

ITCZ Inter-Tropical Convergent Zone

LH-OAT Latin Hypercube One-factor-AT-a- Time

masl Meter above sea level

MCM Million Cubic Meters

MK Mann-Kendall

MW Mega Watt

PBIAS Percent Bias

RVA Range Variability Approach

RCM Regional Climate Model

RCPs Representative Concentration Pathways

REMO Regional Model

SCS Soil Conservation Service

SRES Special Report on Emissions Scenarios

SRTM Shuttle Radar Topographic Mapping

SWAT Soil and Water Assessment Tool

USACE US Army Corps of Engineer’s

WAPCOS Water and Power Consultancy Service (India) limited

WCRP World Climate Research Program

WMA World Meteorological Agency

xix
1. INTRODUCTION

1.1 Background

Water is a unique natural resource essential to meeting human needs and vital to the existence of

life on earth. However, water resource distribution is variable in time and space that affects the

overall development of water resources infrastructure around the world. Nowadays, effective

planning and management of water resources become more critical and challenging due to

conflicting demands by various stakeholders, rapid population growth, industrialization,

hydropower development, irrigation, urbanization and land degradation (Larson et al., 2013).The

impact of climate change on the hydrological cycle may lead to further stress on the management

of water resources in the future. As the water resources are inevitably linked with climate, the

global climate change has severe implications for them (Bates et al., 2008) and therefore, has led

to the vulnerable state of the water resources worldwide.

Therefore, it is essential to meet society's energy and water requirements by using our water
resources more efficiently (Ehteram et al., 2017). In this regard, reservoirs are one of the most
crucial efficient infrastructure components for integrating water resources development and
management (Guo et al., 2011; Li et al., 2010; Lu et al., 2013). Reservoirs alter the spatial and
temporal distribution of runoff. So Reservoir serves many purposes, such as municipal, industrial
and irrigation needs, flood control, hydropower production, navigation, recreation and ecological
requirements (Azizipour et al., 2016; Birhanu et al., 2014; Cheng et al., 2008; Yeh, 1985).
Management of reservoir systems from planning to operation is very challenging since the problem
deals with many complex variables like inflows, storages, power production, irrigation, industrial
and municipal water supply demands. These types of systems are involved because of the nonlinear
storage-inflow relationship, conflicting objectives, and nonlinear constraints. The optimal
operation of reservoirs is therefore vital to the efficient development of water resources in a river

1
basin. Reservoir inflows and water levels are the critical variables for decision making on reservoir
operations, particularly during an extreme flood or drought events.
No general algorithm covers all type of reservoir operation problems (Labadie John W., 2004).
The choice for techniques usually depends on the reservoir specific system characteristics, data
availability, the objectives specified and the constraints imposed. Water resources reservoirs can
be operated according to non-adaptive and adaptive operation approaches under climate change
conditions in the future periods. In the non-adaptive operation approach the developed operational
rules in the historical (base) period are used for operating the reservoir in future periods, and
climate change impacts on reservoir operation are investigated. In the adaptive approach dynamic
real-time operation rules are modified in future period for adapting to climate change. Different
reservoir operation models have been developed and applied for planning studies to formulate and
evaluate alternative plans for solving water management problems; for feasibility studies of
proposed construction projects as well as for re-operation of existing reservoir systems.
Optimization and simulation are the two primary modelling approaches in system analysis of

reservoir operations (Labadie John W., 2004; Wurbs Ralph A., 1993), including hydropower

generation problems. Compared with simulation models that describe reservoir system

performance under a given set of conditions, optimization models apply mathematical

programming to determine a set of decisions that maximize reservoir system performance (Wurbs

Ralph A., 1993). Optimization model expresses reservoir operation problems in standard

mathematical programming forms and solves them through optimization algorithms. Operation

decisions are then determined as outputs of optimization models (Labadie John W., 2004; Liu et

al., 2011). This analysis typically involves optimization and simulation models which can provide

the quantitative information to improve operational water management. The optimization model

is used to minimize or maximize objective function under given constraints, and the simulation

model is used to examine how a water system behaves under a given set of control actions. In the

past, optimization problems have solved by using Linear Programming, Dynamic Programming,

Quadratic Programming, Non-Linear Programming and simulation models (Labadie John W.,

2
2004; Lu et al., 2013; Wurbs, 2005). Nowadays combinations of simulation and optimization

models are used.

Hence, for this study US Army Corps of Engineer’s Hydrologic Engineering Centre Reservoir

Evaluation System Perspective Reservoir Model (HEC-ResPRM) Optimization model is an

appropriate choice. HEC-ResPRM is a network flow programming prescriptive, optimization, or

operating purpose-driven model for reservoir system operations. Prescriptive implies that the

model automatically determines the best plan, as contrasted with descriptive models that

demonstrate what will happen if a specified plan is adopted. The model suggests promising

reservoir operation decisions, driven by quantitative descriptions of operating purpose values,

called penalty functions. This represents a new, but well-tested technology, which (together with

other optimization and simulation models) has become an increasingly popular tool for reservoir

system studies.

Dams/reservoirs have facilitated human utilization and control of rivers by enabling water

managers to convert natural flow variability in to water releases governed by human needs that

have had a significant impact on freshwater ecosystem and species. Dams/ reservoirs can alter the

downstream flow regime by affecting total flow quantity, water quality, magnitude, seasonal

timing, duration and rate of change of specific flow events. Several merits and indicators have

been proposed to assess the river hydrologic regime alterations after dam construction. The range

of variability approach (RVA) based on indicators of hydrologic alteration (IHA) in terms of

streamflow magnitude, timing, frequency duration and rate of change are widely used to assess

river regime alteration especially the impacts of dam construction (Richter et al., 1996; Yu et al.,

2016; Q. Zhang et al., 2014). The approach for assessing hydrologic alteration is based on the

differences in streamflow regime characteristics between two defined time periods at a given

3
stream gauge, for this study at Embamadre stream gauging station in Tekeze River. RVA uses the

pre-impact natural variations of IHA parameter values as a reference for defining the extent to

which natural flow regime have been altered. The advantage of this method is that it rigorously

evaluates changes in Tekeze river hydrological conditions before and after dam construction. By

doing so, it quantifies the magnitude of impacts and yields management relevant information that

can be used to modify reservoir operations for the benefits of stream biota and to guide strategies

aimed at restoration of river’s ecosystem.

Water resources planning based on the concept of a stationary climate is increasingly considered

inadequate for sustainable water resources management (Milly et al., 2008). Hydrological cycle

and water resources system are influenced by temperature and precipitation variability’s and

human activities. Mainly, climate variability and change will have various impacts on water-

resource infrastructure (Stocker et al., 2013). Hence, it is essential to study the effects of climate

variability and change on past (historical) and future water resources under different climate

change scenarios and time periods. Changing patterns of rainfall and its impact on surface water

resources sustainability which includes the availability, management and utilization is an

important climatic problem for the societies today. Thus, understanding of the regional level of

inter-annual and spatial rainfall variability from past data is of important pre-requisites for

enhancing the management of water to mitigate the negative effects of floods and droughts.

Rainfall is the primary atmospheric factor that directly affects streamflow patterns, so detection of

past trend, change and variability in the time series of hydro-climatic variables is very important

for understanding the potential impact of future changes of the Nile basin (Tesemma et al., 2010;

Tekleab et al., 2013; Zhao et al., 2015). Therefore, assessing rainfall and streamflow variability

have been an integral part of water resources planning and management. There are various

4
parametric and non-parametric tests which were used for identifying trends and variability in

hydro-climatologically time series. However, recent studies showed that non-parametric tests are

mostly used for non-normally distributed and censored data, including missing values, which are

frequently, encountered in hydro-climatologically time series (Huang et al., 2014; Shifteh Some’e

et al., 2012; Gocic and Trajkovic, 2013; Sen, 1968)). In this Tekeze basin study, the non-parametric

Mann-Kendall (MK) test and Theil-Sen’s approach was applied for hydro-climatic series over time

on seasonal and annual scale to detect the nature and magnitude of change in trends and other

fluctuations. Therefore, the detection of rainfall and streamflow trends and variability can be used

for hydrological based decision making and further hydrological and climatic modeling.

The hydrologic conditions of the basin need to be adequately understood through solid hydrologic

modeling to investigate the long-term availability of water resources under climate change and

varying land use patterns (Kim and Kaluarachchi, 2009). The use of hydrological models in

planning and management of water resources (Schuol et al., 2008) and analyses of the impact of

climate change on water components (Uniyal et al., 2015) has become a norm recently. Climate

change effect on water resources analyzed using a combination of hydrologic models and climate

change scenarios (Fowler et al., 2007; Xu, 2000; Xu et al., 2005). These hydrologic models provide

a link between climate changes and water yields through simulation of hydrologic processes within

watersheds. Hydrologic models then allow various simulations to be performed based on user

needs. Confidence in the results varies greatly and largely depends on the methods and structure

of the climate scenario and the hydrologic model. Climate change impact studies usually consist

of forcing hydrologic models with Global Climate Model (GCM) projections. However, there exist

many critical challenges in the application of GCMs and hydrological models (Chen et al., 2012;

Fowler et al., 2007).

5
The impact of climate change has been widely studied using GCMs, the most effective tools for

exploring the physical processes of the earth's surface-atmosphere system which can provide very

credible information regarding historical and future climate. The World Climate Research

Programmer's (WCRP) Coupled Model Intercomparison Project phase 3 (CMIP3) and Coupled

Model Intercomparison Project phase 5 (CMIP5) datasets have many GCM outputs to advance

understanding of past, present, and future climate variability and change. Both datasets use

different scenarios describing the amount of greenhouse gas in the atmosphere in the future.

CMIP3 uses climate scenarios from the Intergovernmental Panel on Climate Change's (IPCC)

Special Report on Emissions Scenarios (SRES). Whereas CMIP5 having higher spatial resolution

compared than CMIP3 which built up on CMIP3 and included more idealized process and fed-

back oriented experiments and output to facilitate understanding of the climate system and has

higher spatial resolution uses new established Representative Concentration Pathways (RCPs)

(Meinshausen et al., 2011; Taylor et al., 2011), which can depict a wide variety of possible future

climate scenarios recommended in the IPCC Fifth Assessment Report (AR5).Thus, RCP scenarios

represent an attractive potential approach for further research and assessment, including emissions

mitigation and impact analysis (Riahi et al., 2011; Thomson et al., 2011; Vuuren et al., 2011).

Another benefit of RCPs’ is its better resolution that helps in performing regional and local

comparative studies.

Climate change impacts estimated from the use of a single GCM need to be interpreted cautiously

because GCM introduced uncertainties. Hence, multi-model ensemble climate simulations have

been used in many of the recent studies (Gbobaniyi et al., 2014; Nikulin et al., 2012). The multi-

model mean has been found to be giving a better simulation of the climate variables compared to

the individual models. However, GCMs are limited to regional studies because their spatial

6
resolution is too coarse (approximately 100–250 km) to be compatible with hydrological models

(Fowler et al., 2007) and require downscaling to an appropriate scale to represent the catchment

characteristics.

Progress in regional climate models (RCMs) has recently made to transfer the large-scale

information from GCMs to scales (25–50 km), which are closer to the catchment scale. RCMs

simulate hydrological components such as runoff in addition to climate variables. But these

simulations do not often agree with streamflow observations (Teutschbein and Seibert, 2010) and,

thus, might not be directly useful for assessing hydrological impacts at the catchment scale

(Graham et al., 2007; Teutschbein and Seibert, 2012). Hence, hydrological variables rarely used

directly from RCMs for impact assessment studies.

However, also RCM simulations of temperature and precipitation must be handled with caution as

they often show significant biases (Christensen et al., 2008; Teutschbein and Seibert, 2010). These

preferences need to be considered carefully before applying these data for future projections.

Otherwise, the projections could be misleading. The reasons for such biases include systematic

model errors caused by imperfect conceptualization, discretization and spatial averaging within

grid cells. The bias makes the use of RCM simulations as direct input data for hydrological impact

studies more complicated. One recommendation is to use an ensemble of RCM simulations

(Teutschbein and Seibert, 2010) together with bias correction methods. Bias correction methods

are applied to help remedy the various problems with biased RCM output. Typical biases are the

occurrence of too many wet days with low-intensity rain or incorrect estimation of extreme

temperatures (Ines and Hansen, 2006), but also include under or overestimation and inaccurate

seasonal variations of precipitation (Christensen et al., 2008; Teutschbein and Seibert, 2010).

Several bias correction methods have been developed to downscale climate variables from climate

7
models (Chen et al., 2011). These plans range from simple scaling approaches to rather

sophisticated methods employing probability mapping or weather generators. They were initially

designed to downscale GCM data, but can also be applied to adjust RCM-simulated temperature

and precipitation (Teutschbein and Seibert, 2012).

Several researchers in the Eastern Nile tried to investigate climate change impacts on water

resources by the trend analysis of historical data (Abtew et al., 2009; Conway and Hulme, 1993;

Tabari et al., 2015) and future climate projections (Kim and Kaluarachchi, 2009; Setegn et al.,

2011). Most of the studies focused on the influence of climate variability, and change in the Eastern

Nile has been based on a limited number of GCMs downscaled (dynamic, statistical) RCMs. These

studies used different SRES relied on CMIP3 models for all time windows. However, nowadays

those scenarios have become outdated and needs to be updated and expanded in scope. Recently

the new CMIP5 models typically have more excellent resolution processes, incorporation of

additional physics, and better-developed or well-integrated earth system components used. CMIP5

downscaled as ensemble means at COordinated Regional Climate Downscaling Experiment-

Africa domain (CORDEX-Africa) with new emission scenario called Representative

Concentration Pathways (RCP) which overcomes the shortcoming of SRES scenarios for climate

change impact studies in Africa. A Semi distributed continuous hydrological model Soil and Water

Assessment Tool (SWAT) is increasingly used for climate change assessment applications in the

Eastern Nile basins (Mengistu and Sorteberg, 2012; van Griensven et al., 2012). Hence, SWAT is

a suitable hydrologic model to assess impacts of climate change on the water resources of the

Eastern Nile basin.

Translating the impacts of climate change affected runoff on hydropower reservoir operation

requires coupling of hydrological models and reservoir optimization models. This study integrates

8
these types of models to provide insight into the impacts of climate change on hydropower

reservoir operation. So, this study aims to present a regional overview of Ethiopia’s hydropower

reservoir operation at Tekeze Hydropower dam project under future climate change by using the

projections from SWAT model under ensembles of CORDEX-Africa RCMs outputs for future

climate scenarios of RCP4.5 and [Link], the US Army Corps of Engineer's Reservoir

Evaluation System Perspective Reservoir Model (HEC-ResPRM) used future inflow generated

using calibrated SWAT model to optimally operate Tekeze hydropower reservoir in the face of

climate change.

1.2 Statement of the Problem

Water resources problems worldwide in the future are found to become more complex due to

population growth, climate variability, regulatory requirements, project planning horizons,

temporal and spatial scales, social and environmental considerations, trans-boundary

considerations, etc. Water resources planning and management is going to face difficult challenges

mainly due to the growing water demand from increasingly populated and complex societies which

requires a shift towards an integrated approach; and climate variability and change from a

hydrological point of view assumes of stationarity in long term weather conditions not valid

anymore; and requires seeking for new approaches, strategies and tools. It has been established

that climate variability and change will have impacts on the availability of water resources as well

as on the operating policies for water resources systems.

Today, water resources planning under the paradigm of climate change includes a wide range of

projected climate scenarios, great variability in future supply projections, temperature-driven

increases in water demand, and many other sources of uncertainty. Additionally, the projections

are coarse in nature (i.e., produced at a low resolution) compared with the scale of information

9
utilities directly feed into the water resources planning models. It is then challenging for utilities

to incorporate highly uncertain, low-resolution climate change information into water resources

planning and management decisions. The uncertainties of climate projections depend on the

uncertainties of the scenarios of future emissions, GCMs, downscaling methods and hydrologic

modeling used to study the impacts of climate change on watersheds. Therefore, uncertainties of

climate change and the future non-stationarity of hydrological regimes require the use of the multi-

model ensemble of climate projections in water resources system optimization. A typical method

of evaluating effects of climate change on flow regime is to use an ensemble of GCMs, scenarios

and statistical downscaling/regional climate models to provide inputs to a hydrological model, and

examine the range of effects on statistics of the modeled flows.

Ethiopian government is pursuing plans and programs to use the water resource potential of the 12

river basins of the country for sustainable economic growth and poverty reduction. To alleviate

electricity problem, poverty and food insecurity, it is widely recognized to utilize water resources

such as Tekeze basin and improve water management practices which was one of the least planned

and managed sub-basins of the Nile. Owing to such rapidly developing water resource projects in

the basin, there is an increasing need for the management of the available water resources to boost

agricultural production and to meet the demand for hydro electrical power production. It needs to

be understood when, where and how much water is available and how an intervention plan will be

suitable based on historical and future climate variables with the expected changing climate.

Despite the importance of precipitation and streamflow to insure sustainable water resource

planning and management, long term trends and variability of flow regime and their association

with climate change in Tekeze basin are not yet well understood. Therefore, it is necessary to

identify the presence of hydro-climatic trends and variability in Tekeze basin as it is a prerequisite

10
step to understand the cause and effect of trends and the links to water resources and use in the

basin.

Climate change will cause changes in the patterns of water cycle and geographical distribution of

water resources in the future where impacts will see in climatic factors such as precipitation and

temperature. The impact of interest here is on streamflow which is the sources for reservoir

operation for hydropower reservoir planning and operations. The resulting flow effects could be

in the form of changes in average flows (amount), variability of flow or seasonal variability. Even

though there is marked progress in climate research in recent years, the climate of many parts of

Nile basin particularly Tekeze basin is still not fully understood. Further derived climate scenarios

are very coarse and do not usually adequately capture important regional variations that exist in

Tekeze basin’s climate.

Regarding research gaps and priorities, there is very little detailed information on the impacts and

vulnerabilities of the hydropower reservoir operation in Blue Nile particularly in Tekeze basin

specific to climate change and variability by applying ensemble scenarios and outputs of GCMs

and RCMs. The greatest challenge is that hydropower reservoir operation is depending on

magnitude and timing of streamflow, which in turn is dependent on precipitation. Precipitation

makes hydropower reservoir operation one of the most vulnerable to changing climate. Currently,

Tekeze hydropower reservoir has not been optimally operated and sometimes not fully functional

during dry periods of the year due to fluctuations of the inflow to the reservoirs. To overcome this

problem, the reservoir can be operated optimally to get optimum level of release and optimal

volumes of storage considering future inflows and needs in the face of climate change. In recent

years, combinations of simulation and optimization models are applied in reservoir operation. In

11
this research, US Army Corps of Engineer’s Reservoir Evaluation System Perspective Reservoir

Models (HEC-ResPRM), a combination of simulation and optimization model was used.

1.3 Objectives and Questions of the Research

1.3.1 Objectives of the Research

The main objective of this PhD research is to predict and assess the potential impact of climate

variability and change on the hydrology and hydropower reservoir operation in the Tekeze River

basin, Northwest of Ethiopia. This research focuses on the following specific objectives:

 Investigate the change and variability of long term historical records of precipitation, and

streamflow in the Tekeze river basin

 Quantifying Tekeze river hydrological alterations associated with Tekeze hydropower

reservoir dam operation by comparing the hydrological regimes from pre- and post-impact

period

 Investigate the possible effect of climate change on hydro-climatological variables and water

resources availability under bias corrected new ensembles of CORDEX-Africa RCMs outputs

for future climate scenarios under RCP4.5 and RCP8.5.

 Apply HEC-ResPRM optimization model to get Tekeze hydropower operation optimal

reservoir level and storage for optimal power production including in the face of climate

change.

12
1.3.2 Questions of the Research

To asses and address the above-mentioned problems and objectives, a number of hypothetical

questions can be formulated. Among those:

 What are the nature of the historical trends and variability in precipitation and streamflow in

the Tekeze river basin?

 How Tekeze hydropower reservoir dam operation can affect downstream river flow?

 What is the likely trend in precipitation and temperature for the future period under recent RCP

climate scenarios compared to the historical periods?

 How will climate change impact water resources in the Tekeze River Basin?

 How optimal reservoir operation could be achieved in Tekeze reservoir system in the face of

climate change?

13
1.4 Dissertation Outline

This dissertation includes six chapters as described below. The general flow follows the research

process from the introduction and main body of the thesis through to conclusions and

recommendations. Each chapter that contains a methodology and analysis also gives the results at

the end of that chapter. In conclusion, which is the final chapter a summary made for all the results

obtained. Each chapter is self-sustained and includes its conclusive remarks.

Chapter 1 Introduction: gives a general introduction to climate change and hydrological

variability that affect the water resources of the watershed and reservoir operation. This chapter

also includes statement of the problem, research objectives, research questions and structure of the

dissertation.

Chapter 2 Description of the Study Area: In this chapter, the description of the study area,

topography, climate, hydrology, land use, soil and water resources potential of Tekeze basin was

presented.

Chapter 3 Rainfall & Stream Flow Variability in Tekeze River Basin, Ethiopia: This chapter

presents the assessment of long-term changes in the critical hydro-climatic parameters

(precipitation and streamflow). The chapter aims to establish if there have been significant trends

in hydro-meteorological data. The investigation implemented in Embamadre streamflow and

different precipitation gauging stations of the basin. Rainfall data are used to detect patterns in

climatic variables. Trends and change points were investigated and summarized in this chapter.

Also, the streamflow variability in the watershed and the impact of Tekeze Arch dam hydropower

dam/reservoir operation on streamflow regime using indicators of hydrological alteration were

analyzed.

14
Chapter 4 Climate Change Impacts on the Water Resources of Tekeze basin: describes

hydrological modeling in the context of climate change and description of the hydrological model

and hydrological processes in the model elaborated. This chapter deals with the description of

hydrological and meteorological data used for this study, the methodology that has been employed

for bias correction of ensemble CORDEX-Africa RCMs under RCP4.5 and RCP8.5 climate

scenarios meteorological data sets, their results and discussion elaborated in detail. This chapter

also focuses on setup and calibration of SWAT hydrological model at Tekeze hydropower

reservoir watershed, mainly to investigate the effects of climate change on Tekeze hydropower

reservoir inflows for the coming 90 years (2011-2100). This chapter focuses on different future

climate bias correction methods and modeling of the basin under bias-corrected climate change

scenarios and analysis of future time flow carried out.

Chapter 5 Tekeze Hydropower Reservoir Operation under Climate Change: This chapter

describes hydropower reservoir optimal operation using HEC-ResPRM optimization network flow

monthly model. This chapter highlights the results of current hydropower reservoir optimal

operation and the impacts of climate change on reservoir storage, release and pool level for the

future time periods.

Finally, Chapter 6 summarizes the findings of this dissertation, concluding remarks,

recommendation and directions for further studies in the basin.

15
2. DESCRIPTION OF THE STUDY AREA
2.1 Location and Topography

The Tekeze-Setit-Atbara sub-basin (covers about 227,128 km2) is one of the three major sub-

basins in the Eastern Nile Basin originate on the Ethiopian plateau (Blue Nile, the Tekeze-Setit-

Atbara, and the Baro-Akobo (Sobat) Rivers), and is located at most north-eastern portion of the

Eastern Nile Basin. The Tekeze-Setit-Atbara sub-basin consists of the Tekeze river (known as the

Setit in Sudan), and its tributaries, the Goang (Atbara in Sudan) and Angereb, all of which

originates from the central north and north western highland plateaus of Ethiopia at an altitude of

above 3,000 masl and descends to Humera less than an altitude of 800 masl at the Ethio-Sudan

border, which also goes down to less than 500 masl at its mouth in the Sudan. Water availability

in this sub basin is erratic but it contributes 13% of the Nile flow (Arsano and Tamrat, 2005;

Mengistu and Sorteberg, 2012). This study focuses on the Ethiopian portion of Tekeze basin

(Figure 2.1) up to Ethio-Sudan border with geographic location 11°40' to 15°12' N, and 36°30' to

39°50'E covers an area of 86,510 square kilometers of which 82,350 Km2 (95.19%) in Ethiopia

and relatively small part of the basin 4,160 km2 (4.81%) is situated in Eritrea. From its source, a

spring around Lalibela in North Wollo, the Tekeze River traverses about 750 km to the Ethio-

Sudan border. The topography of Tekeze basin is complex, characterized by rugged topography

consisting of mountainous and terrains of gentle slopes, about 70% of the area lies in the highland

(above 1,500 masl) and some 40% of its watershed has an altitude of above 2,000 masl. The

Tekeze River has the Ras Dashen mountain chain (4,620 masl) in its watershed and Metema low

lands at Sudan Border (589masl). Tekeze basin is bordered by the Mereb River basin and Eritrea

in the north, Atbara River plains in Sudan in the west, Abay River basin in the south and Danakil

basin in the east.

16
 Figure 2.1 Location of Tekeze basin

2.2 Climate and Hydrology

The climate of Tekeze basin is a tropical type with semi-arid in the East and North, and partly

humid in the South. There are three seasons in Tekeze basin; Kiremt (wet) season, Belg (minor

rainy) season and Bega (dry) season. The basin’s rainy seasons are, the main rainy season locally

known as Kiremt, main rainfall season for almost all parts of the basin where more than 70% of

the total annual rainfall falling within three-month periods beginning in the late June and lasting

until September, and a minor rainy season locally known as Belg between February to May in

which most parts of the basin receive considerable amount of rainfall. Bega season from October

to January is mostly a dry season for most parts of the basin. Wide variation in rainfall from year

to year is characteristics of the basin. The variations of rainfall over the basin are mainly associated

with the seasonal migration of the inter-tropical convergent zone (ITCZ) (Seleshi and Zanke, 2004)

17
and complex topography. The mean annual rainfall is 600mm in the lowlands and 1,300mm in the

Semein Mountains which is increased from North to South parts of the basin.

The spatial distribution of temperature value in Tekeze basin is strongly related to altitude. The

area located in the highlands of the basin is characterized by lowest mean monthly temperatures

that occur between December and February, in contrary, low land areas near Sudanese border

characterized hottest temperature. The temperature characteristics in this basin have large

variations in a limited elevation range. The mean monthly temperatures in the basin vary from

100C in the Semein Mountains of Ethiopian highlands to 220C in the highlands and to 260C in the

lowlands. Also, the minimum and maximum temperature ranges from 3-210C in the Semein

Mountains and 19-430C in the lowlands areas.

The river flow pattern of Tekeze basin typically follows that of rainfall. Maximum discharge

occurs in August while it ceases completely during the dry season from October to [Link]

flows in this basin are highly variable (compared to the Blue Nile and Baro-Akobo-Sobat basins)

especially in the crucial low flow months. Seasonal distribution is highly erratic and variable

affecting agricultural production and water resources projects of the basin significantly. At

Embamadre stream gauging station between 1994 and 2008, the mean annual flow was 190m3/s,

and the monthly low and high flow periods were the months of May (15 m3/s) and August (2187

m3/s), respectively.

2.3 Soil and Land use

Soils in the Tekeze basin are classified by the FAO soil classification system (FAO, 1995). There

are four major soil types in the Tekeze basin: Vertisols, Cambisols, Leptosols and Alisols. The

major soils locations on the basin are Eutric Vertisols on the level lands; Eutric Leptosols, Eutric

Vertisols, Eutric Cambisols and Haplic Alisol on the sloping lands; Eutric Leptosols on the steep

18
lands; and Leptosols on composite landforms. Eutric Vertisols is difficult to cultivate. Eutric

Vertisols with soil depths of more than 50 cm are dominant on the level lands while Leptosols are

the most common soils on the sloping lands. Soils in the mountains and hilly-land areas are very

shallow whereas normally deep in the valleys and lowland areas. The Tekeze river basin highlands

support most of the population as they offer a favorable environment for human settlement.

Moreover, the volcanic parent material supplies a rich diversity of nutrients that makes the soils

more suitable for agriculture (Sonneveld and Keyzer, 2003) but in current agricultural production

techniques, this largely exceeds the lands carrying capacity. This is due to loss of topsoil by sheet,

rill and gully erosion as well as on land sliding as a result of continued intense land degradation

and deforestation (Nyssen et al., 2004).

Tekeze basin dominant land use includes cultivated land, open grass land, sparsely grown wood

land, bushes, shrubs and exposed rocks. Recently trees such as Eucalyptus globules and acacia

species have been planted in the protected areas. Apart from forests and bushes, other land use

types include widely grown rain fed annual crops in the basin area were Tef (Eragrostis tef), wheat

(Tritium species), Barely (Hordeum vulagaris), sorghum (Sorghum bicolor), beans (Phaseolus

vulagaris) and Pulses (e.g., Vacia faba). From this annual crops Eragrostis tef covered large

proportion of the cultivated land. Most people in the Tekeze basin lead to a subsistence agrarian

life with agricultural crop production has the oldest history in this area. Environmental

deterioration caused a declining in production which together with the population increase that

created a shortage of land. These processes further led to an expansion of agricultural and grazing

activities in to marginal and steep slopes, which accelerated environmental degradation.

This basin characterized by severs land degradation due to population growth, deforestation,

repeated drought, over grazing and cultivation on the rugged topography. This region has virtually

19
lost its forest cover, and left with only remnant vegetation. The present natural vegetation cover

comprises of sparse woodland of thorny acacia bushes and scrubs interspersed between cultivated

areas. The combination of rugged terrain which is sensitive to erosion as well as difficult for

utilization and management, poor surface cover and the prominent gullies have led the region to

be considered one of the most degraded and degrading.

Cultivated land is the main dominant land use over the basin, where there is agricultural land

expansion due to human economic activities and population pressure. In response to the increasing

demands for food production, agricultural lands are expanding at the expense of natural vegetation

and grass lands. The natural forest resource of the basin is over exploited and most of the climax

vegetation of the basin has disappeared and only little of the original vegetation is evident while

only little of the lowland woodlands and bush lands in the Western and Northern parts of the basin

are nearer to climax. However, the Afro-alpine and sub-afro-alpine heath vegetation lies above

3,700 to 3,900 masl around Semein Mountains. Rehabilitation activities are under way through

area closures, afforestation, plantation programmers and community mobilization only on the

North-eastern part of the basin (Gebremedhin et al., 2003). In this basin, the common agricultural

land use system is a mixed crop-livestock smallholder farming system with agricultural production

in small parcels for subsistence purpose with no external inputs. These changes in land use/land

cover system have great impact among others like agro-biodiversity, soil degradation and

sustainability of agricultural production.

20
2.4 Water Resources Potential

Water resources availability is often the most vital factor controlling the economic growth in

developing countries like Ethiopia, which depend on agriculture. Ethiopia is endowed with a

substantial amount of water resources. The country has adequate average annual rainfall, several

major rivers and lakes, and significant hydropower potential sites and groundwater resources. The

total renewable surface water resources are estimated at 122 billion cubic meters per year from 12

major river basins and 22 lakes. Renewable groundwater resources are estimated to be about 2.6

billion cubic meters while gross hydro-electric potential about 45,000 MW (160,000GWh/year)

with in 299 potential sites and the potentially irrigable land in the country has been estimated at

3.7 million hectares. From these adequate water resources of the country Tekeze basin has a

capacity of surface water potential of 8.2 billion cubic meters per year, ground water potential

about 0.20 billion cubic meters, gross hydro-electric potential 5980Gwh/year with in 15 potential

sites and potential irrigable land from three large scale irrigation projects 83,368 hectares

(Awulachew et al., 2007). The surface water resource potential as runoff is impressive all of which

disappears without being used and the basin is little developed.

Tekeze river basin study not yet well developed even if preliminary water resources assessment

carried out by different water resources researchers and agencies which show different hydropower

and irrigation potential of the basin. The Ethiopian valley development studies authority

commissioned Water and Power Consultancy Service (India) limited (WAPCOS, 1990) undertake

the preliminary water resources development master plan for 14 basins of Ethiopia covering all

aspects of water resources development including domestic, agricultural and industrial use,

hydropower navigation, flood control, environmental aspects and fisheries that identifies ten

potential hydropower sites of Tekeze river basin with technical energy potential of 5588 Gwh/year.

21
According to NEDECO investigation hydropower potential of the basin is quite large, the river is

quite steep and some sites have deep gorges which make ideal dam construction sites easier.

However, the draw backs are steep drops and high flows for the short rainy season and the presence

of high variability over the year. Tekeze basin has a technical hydropower potential of

5980Gwh/year with in 15 potential sites. This study also estimates the potential for irrigable land

to be 45,000hectare for small scale and 207,781 hectare for medium and large scale with most

irrigable land found in the lower part of the basin (NEDECO, 1998). Such irrigation scheme would

get its water from reservoir dams to be built for hydropower generation. According to Tekeze river

basin master plan studies by NEDECO, from 15 potential hydropower sites 10 were selected and

made ground checking, and feasibility study carried out for the three cascade TK04, TK05 and

TK07 hydropower dam sites (NEDECO, 1998). From these potential sites three planned and

operational dam cascades found upstream of Embamadre stream gauged station which includes

from upstream to downstream, TK04 (in planning; 133 MW), Tekeze hydropower reservoir or

TK05 (completed in 2009; 300 MW), and TK07 (in planning; 321 MW) (Figure 2.2). This study

focused only on the operational (TK05) Tekeze arch dam hydropower reservoir. Tekeze

hydropower reservoir (TK05) detail characteristics have shown in chapter 5 and in the annex.

The high surface water potential of Tekeze basin also leads to development of small scale irrigation

systems. The possibility of using about fifty percent of this potential could irrigate half a million

hectares, which could feed three times the population of the region. In order to exploit this

potential an ambitious plan of the regional government of Tigray in 1994 established a

Commission for Sustainable Agriculture and Environmental Rehabilitation for the Tigray region

(Co-SAERT) targeted to bring food self-sufficiency in the region through irrigation agriculture by

planning, design and construction of up to 500 small dams to irrigate 50,000 hectares, but up to

22
2003, around 54 small dams have been built (Haregeweyn et al., 2008), which increase the area

of potentially irrigable land by about 2000 hectares in Tigray region (Behailu, 2002) but only

small number of small dams built in the portion of Amhara region. A socio-economic impact

assessment study conducted for some built small reservoirs/dams indicated that farmers were able

to increase crop yields 3 to 7 fold by using irrigation from the water harvesting schemes (Behailu,

2002). This is due to people and livestock have been able to get enough water easily even during

drought periods. The reservoirs constructed in the Tekeze basin also contributed greatly to the

reduction in soil losses and the off-site effects such as rapid siltation of downstream planned and

operational dams and river. This construction of dams for irrigation and hydropower purpose

resulted in various economic, hydrologic and ecologic benefits.

Figure 2.2Tekeze basin location, climate and streamflow stations including planned cascade
hydropower reservoir

23
3. RAINFALL AND STREAMFLOW VARIABILITY IN TEKEZE RIVER
BASIN, ETHIOPIA1

Abstract: This research was carried out using non-parametric Mann-Kendall test and Indicators

of Hydrologic Alteration (IHA) to determine the potential trend and variability of rainfall and

streamflow variability in Tekeze River Basin. In this study monthly, seasonal and annual

precipitation and streamflow has been analyzed to determine the trends and variability.

Precipitation trend analysis result showed many significant trends, increasing trends for Annual,

Kiremt and Belg seasons from 1953-2013 throughout the basin. Streamflow trends analysis also

showed a decreasing trend in dry season (October to January), in contrary increasing trend

observed in the annual, small and high rainy season months over Tekeze basin. The indicators of

hydrologic alteration and range variability approach (RVA) methods were used to evaluate the

pre- and post-impact hydrologic regimes of Tekeze River due to dam construction. The result

showed that Tekeze hydropower reservoir significantly changed the hydrological regime

downstream of the dam i.e. the 1-, 3-, 7-, 30-, and 90-day intervals minimum streamflow increased,

the 1-, 3-, 7-, 30-, and 90-day intervals maximum streamflow reduced, high pulse count decreased,

fall and rise rates decreased, increased in number of annual hydrograph reversals, the number and

duration of high and low pulses increased. Hence, investigation of trends in the hydro-climatic

variables of Tekeze River basin revealed many significant trends, both increasing and decreasing.

The findings may assist water managers in better planning and management of water resources

under climate variability and change.

Keywords: Hydro-climatic Variability, Hydrologic alteration, Trend Analysis, Rainfall,

Streamflow, Tekeze Basin

24
 3.1 Introduction
Water resources problems are found to become more challenging and complex worldwide for

future. The complexity of water resources planning, and management is due to the contribution of

climate variability, social and environmental considerations, trans-boundary nature of the rivers

and population growth. The stress on water resources diversely increased due to the rapid increase

in population (Wu et al., 2013). Further, water managers have been concerned about the anticipated

impacts of climate variability and change on water resources (Dawadi and Ahmad, 2013; Vedwan

et al., 2008). Changes in climate variability enhance the uncertainties in the availability of fresh

water for the future (Middelkoop et al., 2001). As a result water resources managers face

challenges to meet future water demands on existing water infrastructure that may be inadequate

in the future and stress is increasing to meet environmental flow requirements and provide water

for the energy needs (McCartney and Menker Girma, 2012; Qaiser et al., 2011). Changes in the

hydrological cycle can result from both climate variability and anthropogenic interference. These

changes may be gradual (trend) or abrupt (shift). Changing hydrology may lead to under-designed

or over-designed projects (Mosquera-Machado and Ahmad, 2007), which may not meet long-term

needs; thus, the traditional assumption of stationarity for hydraulic designs requires review (Milly

et al., 2008).

Rainfall and river flow in Africa display high levels of variability across a range of spatial and

temporal scales, with significant consequences for the management of water resource systems

(Conway et al., 2009). Throughout Africa, this variability brings significant implications for

society and causes widespread acute human suffering and economic damage. Examples of

variability include prolonged periods of high flows for rivers draining large parts of East and

central Africa (Conway, 2002). Most parts of East Africa covered with Nile river basin watersheds

are sensitive to climate variations as shown in previous studies (Beyene et al., 2010; Kim and

25
Kaluarachchi, 2009; Setegn et al., 2011) implying that climate change will have a considerable

impact on the resource. About 70% of the Nile flow is from the three major sub-basins in the

Eastern Nile Basin (the Blue Nile, the Tekeze-Setit-Atbara, and the Baro-Akobo (Sobat) Rivers),

and located at the North-eastern portion of the Eastern Nile Basin (Arsano and Tamrat, 2005;

Swain, 1997). As the vast majority of agriculture is rain-fed in the Eastern Nile, precipitation plays

a pivotal role (Block et al., 2008). Precipitation also plays another equally important role in the

Ethiopian highlands, feeding the headwaters of the Blue Nile and Atbara Rivers, which eventually

supply the mighty Nile River. Policy and planning tools, including water management, economic,

hydropower, and irrigation models for Ethiopia and other downstream countries, rely heavily on

precipitation and streamflow as critical parameters. Any rise or fall in the annual and seasonal

rainfall in this region leads to stress on average yearly runoff flow in the area with consequent

implication on the planning of water resource, water allocation and overall integrated development

of the watershed. Changes in precipitation directly affect water resources management, hydrology,

agriculture, natural ecosystem, food security, water quality and quantity, and human health. For

this reason, it is vital to understanding and predicting the trends and variability of precipitation and

streamflow to improve water resources management strategies (Block et al., 2008; Cannarozzo et

al., 2006).

Many studies conducted to characterize trends and variability’s of rainfall and streamflow time

series at various locations across the globe (Casanueva et al., 2013; Melesse et al., 2011; Moges et

al., 2014). These studies have adopted several statistical techniques to quantify increasing or

decreasing trends in annual and monthly rainfall.

Most studies used trend detection techniques like Mann-Kendal, Spearman rho, and linear

regression tests to understand rainfall and streamflow trend and variability in the Eastern Nile

26
Basin. Many studies on Blue Nile basin have been conducted to detect changing pattern and

amounts of rainfall and streamflow in the last decades (Conway and Hulme, 1993; Mengistu et al.,

2014; Tabari et al., 2015; Taye and Willems, 2012). All these studies on rainfall, on the other

hand, do not show any consistent pattern or trends. Conway and Hulme (1993) reported declining

annual rainfall over the Blue Nile, and Tekeze-Atbara basins resulting in a reduction of river flows

between the years of 1945 and 1984. In contrast, recent investigations by Tabari et al. (2015)

agreed that rainfall over the Upper Blue Nile basin did not show a statistically significant trend.

Whereas streamflow showed both statistically significant increasing and decreasing trends in

annual and seasonal time scale at the different Blue Nile gauging stations for the last 40 years

(1964- 2005). Mengistu et al. (2014) showed that annual and seasonal rainfall for 1981-2010 in

the Upper Blue Nile River basin statistically non-significant increasing trends except spring season

that shows declining trend in the Northeastern part of the basin.

In Tekeze river basin part of Eastern Nile, there are little local level studies about rainfall and

streamflow trends. The long-term trend analysis of rainfall in some stations covering a small part

of Tekeze basin was studied various researchers (Gidey et al., 2013; Kiros et al., 2017; Seleshi and

Zanke, 2004; Tesfaye et al., 2017). Seleshi and Zanke (2004) attempted to investigate the pattern

of rainfall over the upper part of Tekeze River basin by considering only one climatic station. Their

output demonstrated that the amount of rainfall remained constant for the past 40 years (1962–

2002). Kiros et al. (2017) investigated a decline of rainfall at Geba catchment a tributary of Tekeze

River using Mann Kendall trend for the last 40 years. Gidey et al. (2013) indicate that there was

a slight decrease in rainfall of upper Tekeze basin considering only part of Tigray region in the

year of 1954-2008. However, all the above studies are specific to catchment level not included the

27
rainfall stations founding in the sources of Tekeze river in the South and Southwest parts of Tekeze

basin.

In pursuit of detecting the trend and the shift of trend in hydro-meteorological variables, various

statistical methods have been developed and used over the years in Ethiopia (Cheung et al., 2008;

David et al., 1999; Tabari et al., 2015; Tesemma et al., 2010).There are various parametric and

non-parametric tests which used for identifying trends in hydro-meteorological time series.

However, from recent studies, it is found that non-parametric tests mostly used for non-normally

distributed and censored data, including missing values, which frequently encountered in hydro-

climatologically time series. The non-parametric Mann–Kendall statistical test (Kendal, 1975;

Mann, 1945) has commonly been used to quantify the significance of monotonic trends in hydro-

meteorological time series (Huang et al., 2014; Shifteh Some’e et al., 2012). The Mann-Kendall

test does not provide an estimate of the magnitude of the trend itself. For this purpose, another

nonparametric method referred to as the Sen’s slope estimate or approach is very popular by the

researchers to quantify slope of the pattern (magnitude) (Gocic and Trajkovic, 2013; Sen, 1968).

Sen's slope estimator used more robust slope estimate than least-square method because it is

insensitive to outliers or extreme values and competes well against simple least squares even for

normally distributed data in the time series. However, these cited trend detection methods are not

designed to detect the duration or timing of changes. The timing of a shift and change-point tests

have been performed in association with a trend test, using the various methods, such as the

nonparametric Pettitt test (Pettitt, 1979). The Pettitt test able to detect potential change points in

the mean of time series; it has been widely used with precipitation and streamflow data (Mu et al.,

2007; Villarini et al., 2011). Change-point tests are also performed separately from trend tests to

provide information on both the significance and timing of the change. However, there is no means

28
for those studies to differentiate between abrupt (a step change) and gradual (a trend) changes

rigorously and consistently unless the nature of the shift is visually obvious.

The natural streamflow’s of rivers also worldwide have significantly altered due to construction

of reservoirs, weirs and other hydraulic structures for the purpose of irrigation, hydropower,

industry and/or domestic uses (Wang et al., 2016). Reservoir operation has significant impacts on

river hydrology, primarily through changing the magnitude, frequency, duration, and timing of

flow regime. Indicators of hydrologic alteration (IHA) widely used to assess river regime

streamflow alteration due to dam construction (Richter et al., 1997). IHA is statistical software

developed by US Nature of Conservancy to assess the change in hydrologic conditions (e.g., River

flows, lake levels) over time including changes caused by water management activities (dams,

diversions, ground water pumping), climate shifts and land use change due to human activities

(Maingi and Marsh, 2002). The IHA method embodies the range of variability approach (RVA)

proposed by Richter et al. (1997), which has successfully used to evaluate hydrological disturbance

in many regulated rivers around the world (Jiang et al., 2014). The RVA assess the extent to which

flow conditions after dam construction (disturbance) lies within a user defined target range of flow

conditions that typified the hydrological regime in pre-impact periods. The IHA assess

hydrological alterations based on 33 parameters characterized by streamflow magnitude, timing,

frequency, duration and rate of change which is essential for understanding and predicting the

impact of altered flow regimes in the river.

The objective of this study was to investigate the change and variability of long term historical

records of precipitation and hydrological data in the Tekeze basin and to evaluate Tekeze river

streamflow regime change caused by Tekeze hydropower reservoir dam construction.

29
 3.2 Material and Methods

3.2.1 Datasets used

The basic datasets that are required by the trend analysis test and indicators of hydrologic alteration

for trend, change point detection and streamflow regime change at Tekeze basin are precipitation

and streamflow data’s.

[Link] Rainfall/precipitation data

Examination of climate trends and variability needs long and high-quality records of climatic

variables. There are now many precipitations recording stations located in Tekeze Basin. However,

only a few stations have continuous records which are found in the South and North-Eastern part

of the basin essential to hydro-climatic studies. In this study, long periods of daily, monthly and

annual precipitation recorded data corresponding to eleven selected gauging station sites out of the

more gauging stations available for Tekeze River Basin collected from the Ethiopian National

Metrological Service Agency (ENMSA) used. The data were analyzed to determine whether there

is evidence of specific trends in the characteristics of inter-annually and annually rainfall events

in the upper parts of the basin. The period of record was from 1953 to 2013 with varying record

length. The length of recording varies due to differences in site establishment and data gaps. To

check on the spatial coherence of the variability results across the study area. Some potential data

problems, for instance missing values, data entry errors, outliers, etc., were solved by careful

inspection. Due to lack of continuous available data this study focused on the middle and higher

latitude of half part of Tekeze basin. Location of the study area and spatial distributions of the 11

stations are shown in Figure 2.2, and their characteristics and data availability are presented in

Table 3.1.

30
Table 3.1 Location of weather stations and lengths of precipitation series used in this study
Long. Altitude Period of
Station
Lat. (o) (o) (masl) Pmean (mm) record
Axum 14.12 38.73 2105 723.94 1992-2012
Debre Tabor 11.53 38.02 2690 1439.04 1988-2013
Gonder 12.33 38.02 1967 1175.18 1953-2004
Hager Selam 13.39 39.09 2000 692.49 1994-2012
Hawzen 13.58 39.26 2242 531.90 1971-2012
Korem 12.31 39.31 3000 980.50 1985-2012
Lalibela 12.31 39.03 2500 799.07 1976-2004
Maichew 12.48 39.32 2400 733.03 1971-2012
Mekele 13.3 39.29 2070 603.68 1980-2012
Nefas Mewucha 11.44 38.27 3000 1103.41 1986-2004
Wukro 13.46 39.36 2070 581.29 1992-2012
Pmean is mean annual precipitation

[Link] River flow record at Tekeze Basin

There are more than 20 streamflow gauging stations in Tekeze Basin covering small tributaries

obtained from the Department of Hydrology – Ethiopian Ministry of Water, Irrigation and

Electricity. But all stations except Embamadre station cannot be used for the analysis because of

massive data gaps during civil war particularly at the end of 1970s up to mid-1980s and after that

most stations were not in operation for long periods of time. The remaining station which was

operational after 1990 has unreliable discharge data mainly due to the short period of record and

high missing data. The minimum length of streamflow record is 21 years (1994-2014) at

Embamadre which is impacted by hydropower reservoir operation since 2009 and less than 15

years (1998-2014) for the remaining stations. A longer period of recorded historical flow data at

Embamadre gauging stations increases the visibility of trends and the reliability of results from

trend analysis than shorter period records of flow in other stations found in the small tributary of

Tekeze basin even if they enhance spatial coverage on different parts of the basin. Based on the

quality of the data, time series length and influence of infrastructure (Tekeze hydropower

31
reservoir) Embamadre station (Figure 2.2) was selected for detailed analysis. The temporal

resolution of data available is daily.

[Link] Data Quality Check and Analysis

The hydro-climatic data of Tekeze basin is scanty and have many data gaps. Therefore, it is

essential to devote careful screening and quality check for all data before use in any hydro-

climatological analysis. The quality of collected data sets was checked in several ways, mainly

depending on the type of the data set and perceived uncertainties. The three steps visualization,

comparison to the nearest station with in the same zone and regression were taken for verifying

the hydro-climatic data used in this research. The climatic data was sorted and arranged in excel

and checked the homogeneity of it using the standard normal homogeneity test and Von Neumann

ratio test. The data series from these eleven weather stations that was used in this study were found

to be homogeneous. Visual inspection and regression relations between neighboring stations have

been used to detect outliers and fill in the missing gaps in the data series. For precipitation and

temperature data the regression coefficient (r2) ranges between values of 0.77 to 0.89. Large data

gaps of each weather station for the duration of one year and above, where excluded from the

analysis. Also, according to World Meteorological Agency (WMA) standards, it is not

recommended to fill more than 10% of missing data. Missing values less than 8% for daily stream

flow at Embamadre gauging stations and less than 10% of the eleven selected stations precipitation

and temperature data have been used over the considered period. According to the result all the

selected eleven stations were consistent and have been used for climate variability and change

studies and simulation purposes as discussed in the next chapter.

32
3.2.2 Trend analysis methods

There are various parametric and non-parametric tests available such as moving average, Bayesian

procedures, Mann-Kendall test, filtering technology, Pearson correlation coefficient, etc. which

were used to detect trends and variability in hydro-meteorological time series. However, from

recent studies it is found that non-parametric Mann-Kendall test are most frequently used for non-

normally distributed and censored data, including missing values, which are frequently

encountered in hydro-climatic time series. Non-parametric trend analysis is preferred by various

researchers recently due to several advantages over parametric method. (i) Non-parametric tests

do not require the assumption of normality or homogeneity of variance which are frequently

violated by hydro-climatic data or are difficult to verify; (ii) It compares medians rather than means

and, as a result, if the data have one or two outliers, their influence is negated; (iii) Prior

transformations are not required, even when approximate normally could be achieved; (iv) greater

powers is achieved for the sketching distribution and (v) data below the detection limit can be

incorporated without fabrication of values or biases. The statistical significance trend detected

using a non-parametric model such as Mann-Kendal test can be complemented with Theil–Sen’s

slope estimation to determine the magnitude of the trend and the Pettit test to determine the change

point detection. Brief explanations of these methods are discussed as follows:

[Link] Mann–Kendall test

The non-parametric Mann–Kendall test (Kendal, 1975; Mann, 1945) is widely used to evaluate

statistically significant trends in hydro-meteorological time series (Gocic and Trajkovic, 2013;

Shifteh Some’e et al., 2012). Like many other trend methods, the Mann–Kendall test assumes that

the time series under research are stable, independent and random with equal probability

distributions. It has the advantage of being a simple calculation that assumes no special data

33
distribution. The Mann-Kendall test, used to detect a trend in a time series, yields an incorrect (too

large) rejection rate when applied to an auto correlated series with no trend. Pre-whitening corrects

this situation, but reduces the power of the test when a trend exists. Pre-whitening is the most

commonly used procedure to eliminate the effect of serial correlation in trend analysis. It

efficiently removes the possibility of finding a significant trend in the Mann-Kendall test when

actually there is no trend. But pre-whitening has the disadvantage of accepting the hypothesis of

no trend with a high probability when a trend exists.

In the Mann–Kendall test, the null hypothesis H0 states that the data x1; x2; …; xn is a sample of n

random variable independent and distributed identically without considering seasonal changes.

The alternative hypothesis H1 of a two-sided test is that the distributions of xk and xj are not

identical for all k, j ≤ n with k j. The test statistic S is given by:

n 1

 sgn x  xk 
n
S  j (3.1)
k 1 j  k 1

in which n is number of data points, xi and xj are the data values in time series i and j (j>i),

respectively and sgn(xj-xi) is the sign function as

 1 if x j  xk  0

sgn( x j  xk )   0 if x j  xk  0 (3.2)
 1 if x  x  0
 j k

The variance is computed as

 m

 n n  12 n  5    ti ti  12ti  5b
VarS    i 1  (3.3)
18

Where n is number of data points; m is the number of tied groups which has a set of sample data

with same value and ti is the number of ties for the i value. When the sample size n > 10, the

standard normal variable, Z, is computed from the following equation(Douglas et al., 2000):

34
 S−1
, ifS > 0
√Var(S)
Z = 0, ifS = 0 (3.4)
S+1
, ifS < 0
{ √Var(S)

A hypothesis test based on normalized Kendall’s statistics for a significance level of  is used to

analyze all variables in the Mann– Kendall test. The null hypothesis, H0, is accepted at the

significance level of  in the two-sided trend test if -Z/2< Z < Z/2, where  Z/2 are standard

normal deviates. Alternatively, H0 is rejected or Z is statistically significant if Z > Z/2 or if Z < -

Z/2. Moreover, positive values of Z indicate an increasing trend while a negative Z reflects a

decreasing trend. In most of the researches around the globe, changing trends tested at 0.01, 0.05

and 0.1 significance levels. The null hypothesis of no trend is rejected if [Z] > 1.65, [Z]> 1.96 and

[Z] > 2.57 at the10%, 5% and 1% significance levels, respectively.

In this study, significance level α=0.05 which usually used in the Nile Basin hydro-climatic trend

analysis were used (Tabari et al., 2015).The Z values are approximately normally distributed, and

a positive Z value larger than 1.96 (based on normal probability tables) denotes a significant

increasing trend at the significance level of 0.05, whereas a negative Z value lower than -1.96

shows a significant decreasing trend.

[Link] Sen's slope estimator

The Sen’s slope method is a nonparametric, linear slope estimator that works most effectively on

monotonic data. Sen’s slope method (Sen, 1968) used to determine the magnitude of the trend line.

Sen’s slope proceeds by calculating the slope as a change in measurement per change in time. The

slope of trend in the sample of N pairs of data estimated as:

x x
j k (3.5)
Q  for i  1,..., N
i j k

35
Where Qi is slope between data points, xj and xk are the data values at times j and k (j > k),

respectively.

If there is only one datum in each time period, then N nn  1 ; where n is the number of time

2

periods. If there are multiple observations in one or more-time periods, then N < nn  1 ; where n
2

is the total number of observations.

The N values of Qi are ranked from smallest to largest, and the median of slope or Sen's slope

estimator calculated as:

Q (N+1) if N is odd
2
Qmed = { Q N +Q N+2 (3.6)
( ) ( )
2 2
if N is even
2

The Qmed sign reflects data trend reflection, while its value indicates the steepness of the trend. To

determine whether the median slope is statistically different than zero, one should obtain the

confidence interval of Qmed at specific probability.

The confidence interval about the time slope can be computed as follows:

Cα = Z1−α/2 √Var(S) (3.7)

Where Var(S) defined in Equation (3.3) and Z1−α/2 is obtained from the standard normal distribution

table. In this study, the confidence interval was computed at significance level α=0.05.

Then M 1  N  C and M 2  N  C are computed. The lower and upper limits of the confidence
2 2

interval, Qmin, and Qmax are the M1th largest and the (M2+1) th largest of the N ordered slope

estimates. The slope Qmed is statistically different than zero if the two limits (Qmin and Qmax) have

a similar sign. When hypothesis of no trend is rejected by Mann Kendall test, the Sen’s slope is

36
used to quantify the trend. Sen's slope estimator widely used in hydro-meteorological time series

(Huang et al., 2014; Shifteh Some’e et al., 2012).

[Link] Pettit’s test for change-point detection

This test, developed by Pettitt (Pettitt, 1979) is a nonparametric test, which is useful for evaluating

the occurrence of abrupt changes in hydrological and climatic records with continuous data. One

of the reasons for using this test is that it is more sensitive to breaks in the middle of the time series

and mostly used change point detection in climatic records (Smadi et al., 2006). This method

detects a significant change in the mean of a time series when the exact time of the change is

unknown. The test uses a version of the Mann-Whitney statistic Ut, N, that tests whether two sample

sets X1,. . ., Xt and Xt+1, ... XN are from the same population. The test statistic Ut, N is given by:

N
U t , N  U t 1, N   sgn( X t  X j ) (3.8)
J 1

Where t = 2, 3, …, N and

if ( X t  X j )  0, sgn( X t  X j )  1
if ( X t  X j )  0, sgn( X t  X j )  0 (3.9)
if ( X t  X j )  0, sgn( X t  X j )  1

The test statistic counts the number of times a member of the first sample exceeds a member of

the second sample. The test statistic KN and the associated probability (P) used in the test given as:

K N  max 1t  N U t , N (3.10)

  6K 2  (3.11)
P  2 exp 3 N 2 
 N N  
The significance probability of KN approximated for p≤ 0.05.

37
3.2.3 Indicators of Hydrologic Alteration

The Nature Conservancy developed Indicators of hydrologic alteration (IHA) method to enable

rapid processing of daily hydrologic records to characterise natural flow conditions and facilitate

evaluations of human induced changes to flow regimes (Richter et al., 1996; Yu et al., 2016; Q.

Zhang et al., 2014). IHA is one of the most common and widely accepted methods for assessing

the extent to which human disturbance has altered the hydrological regime of river [Link]

program was designed to calculate the value of thirty three hydrologic parameters that characterize

the daily, monthly and annual flow regime by five hydrologic features: (i) magnitude of monthly

discharge, (ii) magnitude and duration of annual extreme flows, (iii) timing of annual extreme

discharge (iv) frequency and duration of high and low pulses, and (v) rate and frequency of

discharge change (Table 3.2). Range variability approach (RVA) method used to assess the

hydrological regime alterations for regulated rivers based on IHA (Yu et al., 2016). In the RVA

analysis, hydrological parameters were calculated using parametric (mean and standard deviation)

or non-parametric (median and percentile) statistics. For most situations, non-parametric statistics

are a better choice, because of the skewed nature of many hydrologic datasets. But parametric

statistics may be preferable for certain situations such as flood frequency or average monthly flow

volumes. IHA detail description can be found in Richter et al. (Richter et al., 1998, 1997, 1996).

It is recommended to use long years of daily record streamflow data to obtain reliable pre- vs. post

impact comparison, as well as for trend analysis (Richter et al., 1997), while twenty years should

be considered as good baseline requirement for the amount of data needed. In this study, a twenty-

one years daily streamflow data at Embamadre guage records were analyzed using the IHA

methods to determine hydrologic shifts of Tekeze river basin streamflow in response to Tekeze

hydropower reservoir dam construction and operation. Analysis were conducted on mean daily

38
discharges for the water year (June-May) for the period of record prior to Tekeze hydropower dam

construction (reference) and then again after dam construction completed (disturbance). For the

case of Tekeze river the historical (1994-2008) and post-dam construction (2009-2014) hydrologic

conditions were evaluated. As the analysis focused on two time intervals of dams pre-impact and

post-impact periods, the range variability approach (RVA) was used to evaluate hydrological

change (Gao et al., 2013; Richter et al., 1997; Wang et al., 2016). The RVA uses the pre-impact

variation of IHA parameter values as references for defining the extent to which the flow regime

has been altered by dam construction (Richter et al., 1998). Based on these references, RVA

analysis generates a series of hydrologic alteration factors, which quantify the degree of alteration

of 33 IHA parametres.

The hydrological alterations in the RVA analysis are evaluated by comparing the frequency with

which pre-impact and post-impact variables (usually the IHA) fall within the three categories.

Rechiter et al. (1998) divide the range of hydrologic alterations in to three classes of equal ranges

with a distinict pattern as no alteration (0-33%), moderate alteration (34-67%), and high degree

of alteration (68-100%). Hydrologic alteration (HA) assumed to occur if the number of post-impact

values falling in the central interval (34th to 67th percentile) differ from the expected ones i.e. the

number of the pre-impact values. The RVA target range for each parameters is barcketed by 25th

and 75th percentail values of the pre-impact daily flow. The deviation of the post-impact flow

regime from the pre-impact period quantified using hydrologic alteration of Tekeze river. The

degree of hydrologic alteration, HA, is calculated for each variable as:

𝑁𝑂 −𝑁𝑒
𝐻𝐴 = [ ] 𝑥100 (3.12)
𝑁𝑒

𝑁𝑒 = 𝑝𝑥𝑁𝑇 (3.13)

39
Where NO is observed number of post-impact years for which the values of hydrologic parametres

falls with in the RVA target range, Ne is expected number of post-impact years for which the

values of hydrologic parametres falls with in the RVA target range, and p is percentage of post-

impact years for which the values of hydrologic parametres falls with in the RVA target range,

and NT is total number of post-impact years. When the observed frequency of post-impact annual

values falling with in the RVA target range equals the expected frequency, HA is equal to zero. A

positive HA factor means that the frequency of values in the category has increased from the pre-

impact to post impact period (maximum: 2), while a negative value means that the frequency of

values has decreased (minimum: -1). The coefficient of dispersion was a commonly used indicator

used to evaluate the variability of daily streamflow. It is calculated as: the coefficient of dispersion

(CD) = (75th percentile-25th percentile)/50th percentile.

Table 3.2 Summary of hydrologic parametres used in the IHA and their charactrestics
IHA statistics group Hydrologic Parameters used in Tekeze basin Total No. of
characteristics streamflow parameters
1. Magnitude of monthly discharge Magnitude, Median discharge for each 12 parameters
condition timing calendar month
2. Magnitude and duration of annual Magnitude, Annual maxima1-, 3-, 7-, 30-,
extremes discharge condition duration and 90-day means
Annual minima1-, 3-, 7-, 30-, 12 parameters
and 90-day means
Number of zero days, base flow
index
3. Timing of annual extremes Timing Julian date of each annual 1-day 2 parameters
discharge condition maximum
Julian date of each annual 1-day
minimum
4. Frequency and duration of Magnitude, Number of high pulses each
high and low pulses frequency year
Number of low pulses each year
Duration Median duration of high pulses 4 parameters
each year
Median duration of low pulses
each year

5. Rate and frequency of hydrograph Frequency, Number of rises and falls 3 parameters
changes rate of change Number of reversals

40
3.3 Result and Discussion

3.3.1 Preliminary Analysis

The rainfall data screened and comparisons between stations were made using the statistical

metrics mean, standard deviation (STD), the coefficient of variation (CV), skewness (Cs), and

actual excess kurtosis (Ku). The mean annual rainfall varied between 581.29mm in the Northern

part of the Tekeze River basin (Wukro station) and 1439.04mm in the Southwest part (Nefas

Mewucha station) of this basin. The skewness (Cs), which is a measure of asymmetry in a

frequency distribution around the mean, varied between 1.47 and 2.41, positive skewness

indicating that annual precipitation during the period is asymmetric and it lies to the right of the

mean over all the stations. Kurtosis (Ku) is a statistic describing the peakedness of a symmetrical

frequency distribution, for Tekeze basin it varied from 0.65 to 5.44 (Table 3.3). For time series

data to be considered normally distributed, the coefficient of skewness and kurtosis must be equal

to 0 and 3, respectively. Table 3.3 indicates, therefore, that the data in Tekeze basins precipitation

stations are positively skewed and not normally distributed. The coefficient of variation (CV), a

statistical measure of the dispersion of data points in a data series around the mean, was computed

for all stations to investigate spatial pattern of inter-annual variability of annual precipitation over

the study area. The coefficient of variation varied between 1.18 (Maichew station) and 1.86

(Wukro station). Table 3.3 shows that stations found in the Northern part of Tekeze basin (Wukro,

Mekele, Hawzen, and Axum stations) have a more inter-annual variability of precipitation than the

stations observed in the South and Southwest part of the basin. It can be concluded from the results

that the areas of usually heavy precipitation are the zone of least variability and areas of lowest

precipitation are the zone of highest variability.

41
The highest mean annual precipitation recorded at Debre Tabor station (1439.04 mm) and the

lowest registered in Hawzen station (531.90mm) according to the analysis of the present datasets.

These two rainfall stations also recorded the maximum (148.15 mm) and minimum (66.91 mm)

standard deviations respectively. However, 1998.40mm is seen for Debre Tabor station while

minimum (757.81 mm) indicated for Wukro station compared to other stations. The annual rainfall

series are positively skewed for all the eleven stations and the coefficient of variation shows no

significant differences among the stations. Mean annual rainfall at the stations varied from 531.90

mm /year to 1439.04 mm/year.

Table 3.3 Annual precipitation time series basic statistical properties of the study area
Station Pmean (mm) Pmax (mm) STD CV Cs Ku
Axum 723.94 1067.20 88.24 1.46 1.91 3.47
Debre Tabor 1439.04 1998.40 148.15 1.24 1.32 0.65
Gonder 1175.18 1772.80 120.62 1.23 1.41 1.26
Hager Selam 692.49 900.00 82.41 1.43 1.90 3.50
Hawzen 531.90 768.50 66.91 1.51 2.07 3.90
Korem 980.50 1272.20 103.00 1.26 1.68 2.14
Lalibela 799.07 1100.10 96.12 1.44 1.93 3.47
Maichew 733.03 1051.00 72.05 1.18 1.47 1.58
Mekele 603.68 917.90 84.18 1.67 2.21 4.47
Nefas Mewucha 1103.41 1105.70 113.82 1.24 1.90 4.13
Wukro 581.29 757.81 90.07 1.86 2.41 5.44
Pmean: mean annual precipitation; Pmax: maximum annual precipitation

3.3.2 Trends of precipitation

[Link] Annual trends of precipitation

Analysis of the Tekeze basin annual precipitation time-series using non-parametric Mann-Kendall

test found that 64% of the stations with a positive trend and the rest with a negative trend (Figure

3.3). Annual precipitation trend magnitude, direction and significance in the Tekeze river basin

are shown in Figure 3.1 through Figure 3.3. In the Tekeze basin selected stations, the level of

significance using Z value identified as positive and negative non-significant trends.

42
The Theil Sen’s Slope estimator (Q mm/year) summarizes the results of change per unit time of

the trends detected in the basin. The highest decreasing trend detected in the Northern part (Axum,

Hawzen, and Hagere Selam stations) whereas an increasing trend identified in the other stations

found in Southern (Lalibela, Nefas Mewucha, Debre Tabor and Korem stations) and Eastern

(Maichew and Mekele stations) part of Tekeze basin (Figure 3.6). Sen’s slope indicates that the

magnitudes of the non-significant positive trends at 95% confident level varies in the range of 8.12

mm/year at Lalibela station to 2.32 mm/year at Nefas Mewucha station. A negative non-significant

trend also predicted in the basin and varies from –1.26 mm/year at Gonder station to –6.22

mm/year at Axum station. Positive trends mostly happened in the Eastern part of Tekeze basin

whereas the negative trends occurred in the Northern and Southwestern parts of the basin. Mann-

Kendal test results in all selected stations of Tekeze basin shows a non-significant increasing and

decreasing annual precipitation trend observed at the 5% significance level which is similar to

other studies done in the Eastern Nile (Bewket and Conway, 2007; Conway, 2000; Gebremicael

et al., 2013; Mengistu et al., 2014; Tabari et al., 2015).

Figure 3.1 Results of Mann Kendall annual precipitation time series trend test at 95% confidence

interval

43
 Figure 3.2 Annual precipitation trend magnitudes at different stations of Tekeze basin

Figure 3.3 Percentage annual and seasonal trend test results a) Overall negative and positive

trends b) significant positive and negative trends at 95% confidence level

44
[Link] Seasonal precipitation Trend

For all selected stations of Tekeze basin, Mann Kendall test and Sen’s slope estimator method

were also applied to detect the temporal trends of seasonal precipitation time series during 1953-

2013. Pre-whitening did not change the results of the seasonal MK test, implying that the trend

was strong enough not to be missed, even after pre-whitening. Like the annual precipitation series,

the seasonal time series in the Tekeze basin showed a mix of positive and negative trends but

significant trends (Figure 3.3 and Figure 3.4).

In Kiremt main rainy season (Wet season: June-September) precipitation, 82% of the stations and

18% of the stations show positive and negative trends, respectively (Figure 3.3a). Only 55% (6

stations) found with significant positive trends and none with significant negative trends (Figure

3.3b). These significant positive trends are mostly observed to exist in the sources of Tekeze basin

at stations Debre Tabor, Lalibela, Nefas Mewucha in the southern mountainous areas and stations

like Maichew and Korem in the eastern part of the basin. Kiremt precipitation trend magnitude

varied between −1.19 and 6.4 mm/year. In general, it can be stated that the Kiremt season has been

experiencing mild increasing precipitation trends over the past 50 years. Kiremt season trend was

like the annual precipitation trend, which indicates that Kiremt precipitation has high contribution

for annual precipitation in Tekeze basin.

The Belg (Small rainy season: February-March) precipitation trends showed a similar trend to the

Kiremt season time series. Most of the trends in the Belg precipitation time series were positive

accounting for about 91% of the stations except Korem station (Figure 3.4 and Figure 3.3a).

Nevertheless, the significant positive trends in Belg precipitation were higher compared with those

in the other seasonal series. Nine significant positive trends (82%) detected in the Belg time series

(Figure 3.3b). This is due to the South-Easterly winds from the Indian Ocean, and the Gulf of Aden

45
produce the Belg rains to the East-central part of the Northwestern highlands of Ethiopia (Seleshi

and Zanke, 2004). Hence these stations record rainfall during this rainy period. The magnitude of

the significant increasing trend varies from 0.47mm/year at Northern part (Wukro station) to

2.41mm/year at Southern part (Debre Tabor station) of Tekeze basin. According to these results,

the significant increasing trend in Belg precipitation series experiences increasing trend throughout

Tekeze basin for the last 50 years. The result of this study in higher variability and increasing trend

Belg precipitation in Tekeze basin experiences similar result done by Cheung et al.(2008).

Figure 3.4 Mann Kendall Z values at different Tekeze basin stations

46
Figure 3.5 Results of seasonal precipitation trend magnitude at different weather stations

Figure 3.6 Annual precipitation increasing and decreasing trends in the Upper parts of Tekeze
basin

47
[Link] Change Point Results

Since the Mann-Kendall tests showed significant trend in seasonal precipitation at the significance

level of 0.05, the Pettit test was further used to detect the change points of transitional years. For

annual precipitation, there is no change point year could be detected at 95% confidence level as

all stations annual precipitation shows non-significant trend. Most of the station's precipitation in

both Kiremt (main rainy) and Belg (small rainy) seasons shift occurred around the 1970s or 1980s

either positive or negative. This is mainly due to the drought period of Tekeze basin that lasted

from 1978 to 1986. This result is confirmed the study of Conway (2000) and Seleshi and Zanke

(2004), during the late 1970s to mid-1980s near-minimum precipitation, recorded confirming the

unusually low rainfall received by Ethiopia during those years and the significant shift occurred.

3.3.3 Streamflow Trend and Variability

Streamflow is a very useful indicator of long-term hydro-climatic changes. From a water resources

management perspective, the identification of trend and variability in streamflow are critical for

planning of water resources purposes. Trend analysis is useful for understanding dynamics and

behaviors of hydrological and climatic variables over a long-term period. The non-parametric

Mann–Kendall test was applied to the annual and seasonal streamflow data at Embamadre station

over the period 1994 to 2008 with no substantial influence by water withdrawals for hydropower

or other water use purposes.

The Z statistic of streamflow was 0.62 and showed a non-significant decreasing trend at the 5%

confidence level. The annual streamflow was increased at a rate of 1.867m3/year. The observed

increases in yearly streamflow come primarily from rainy season runoff.

To better analyze of the trend and variabilities on seasonal streamflow, according to the

precipitation characteristics, one year divided into wet season (Kiremt: June-September), dry

48
season (Bega: October–January) and small rainy season (Belg: February-March). Kiremt (wet)

and Belg seasons streamflow shows a significantly increasing trend with Z= 4.43 and Z = 8.14

respectively whereas Bega (dry) season showed a significant decreasing trend with Z value -10.35.

The magnitude of the increasing trend in Kiremt varies up to 4.69m3/year and Belg season up to

0.14m3/year, the decline in Bega season up to 0.59m3/year.

This study of streamflow variability at Embamadre gauging station is significant as all

streamflow’s generated from the high altitude and mountainous regions usually reaches its

maximum value at this station and hydrological regime of the Tekeze river mainstream is strongly

affected by human activities like Tekeze hydropower reservoir and planned irrigation, hydropower

and water conservation projects. Although annual precipitation exhibited an increasing trend for

the past five decades, which in theory should result in more runoff, streamflow and water

availability in the Tekeze mainstream at Embamadre has decreased and the environmental

situation has been severely impaired because of limited water resources and may dried-up of the

main stream during dry periods. Therefore, land degradation and landscape change with human

activities, as well as the change and variability of climate all contributed to the trends of streamflow

detected in this study. This result is in agreement with the study of Tesfaye et al., (2017) in the

Tekeze basin and Tesemma et al., (2010) in the neighboring Upper Blue Nile basins who found

main and small rainy season’s significant increasing trends and decreasing trends in the dry

seasons. Overall, this study provides an elaborate view of past precipitation and streamflow trends

in the upper half part of Tekeze basin which should be use full for further research.

3.3.4 Influence of Tekeze reservoir operation on the streamflow regime

In the Tekeze River at Embamadre gauged stations all the 33 hydrologic parameters of medians,

coefficient of dispersion (CD), and measure of HA were calculated with the IHA software. The

49
25th and 75th percentile values were calculated based on the available pre-impact streamflow

records with considering low and high boundaries of the RVA target range. The RVA analysis

showed that the natural flow regime in the middle Tekeze River at Embamadre station significantly

changed after the operation of Tekeze hydropower reservoir (Table 3.4).

[Link] Magnitude of monthly streamflow

The result in Figure 3.7 indicated that the river flow become more smoothness in the post-impact

period by two major changes, a decrease in high flow and an increase in low flow. Flow regime

alterations were closely related to Tekeze hydropower reservoir operation that stores more water

in rainy seasons (July through September) and release water downstream for low flow season

power production. Tekeze hydropower reservoir operation altered the orginal hydrologic process,

smoothend the peak swelling, and increase dry season discharge of Tekeze river.

Figure 3.7 Comparison of median monthly streamflow before and after Tekeze hydropower

reservoir dam construction

The magnitude of monthly flow from February to June, the normal low flow period, increased

after 2009 when the reservoir behind the dam stored water for power production. Median discharge

for all months after Tekeze hydropower dam construction differ significantly from values for these

50
months in the pre-dam construction period. Thus, during the low flow period of the year, operation

of the dam increases median discharges while decreases high flow month’s median flow. Figure

3.8 showed that an increase in November median flow and a decrease in August median flow.

Figure 3.8 November and August monthly median flow change before and after Tekeze
hydropower reservoir construction in Tekeze river

51
[Link] Timing of annual extreme streamflow change

The median Julian dates of each annual 1-day minimum move backward from the 130th day in pre-

impact period to the 337th day in the post-impact period, with the low alteration of -14%. The

median Julian dates of each annual 1-day maximum also move backward from the 224th day in

pre-impact period to the 239th day in the post-impact period, with the low alteration of 29%.

[Link] Rate and frequency of streamflow changes

Table 3.4 showed that there were a change in the medians of rise rate, fall rate and number of

reversals in the pre-impact and post-impact periods. The median of rise rate decreased from

39.3m3/s/day in the pre-impact period to 33.34m3/s/day in post-impact period, with high

hydrological alteration of 114%. The median of fall rate also decreased by -3.4 m3/s/day in pre-

impact period to -18.83 m3/s/day in post-impact periods with hydrological alteration of -100%.

Together with these changes indicate that the dam significantly decreased the rate of rise of

hydrographs, presumably owing to storage effects of the reservoir, and led to many more reversals

between rising and falling stages of flow in the river. The medians of number of reversals also has

been significantly altered and increased from 87 in pre-impact to 138 post impact, with high

hydrological alteration of -100%.

52
Table 3.4 None parametric RVA scores at Embamadre station of Tekeze River
Pre-impact period (1994- Post-impact period (2009-
Hydrologic 2008) 2014) RVA targets
parameters IHA
Medians CD Medians CD Low High
(%)
Parameter Group #1
July 477.70 0.64 488.60 0.84 296.30 526.20 29
August 1322.00 0.53 849.00 1.08 897.80 1400.00 -57
September 320.30 0.52 443.10 0.46 297.30 375.50 -57
October 110.80 0.79 264.90 0.65 86.87 139.40 -57
November 42.90 1.07 239.80 0.86 35.82 58.96 -100
December 32.50 0.84 211.30 0.67 25.77 37.46 -57
January 23.10 1.23 201.50 0.94 9.72 30.16 -100
February 21.30 1.02 304.10 0.48 9.10 24.90 -100
March 20.20 0.91 273.70 0.42 16.60 21.90 -100
April 16.10 1.50 293.20 0.76 7.95 29.15 -100
May 19.10 0.85 320.00 0.75 13.87 26.44 -100
June 58.55 1.13 312.00 0.69 31.40 74.95 -100
Parameter Group #2
1-day minimum 3.70 1.92 64.70 2.17 2.15 8.06 -100
3-day minimum 9.00 1.42 154.90 0.97 2.27 9.37 -100
7-day minimum 9.81 1.33 184.40 0.82 2.47 13.28 -100
30-day minimum 14.21 1.01 201.50 0.80 4.71 17.22 -100
90-day minimum 17.12 1.13 201.50 0.88 9.97 22.70 -100
1-day maximum 3033.00 0.50 1719.00 0.65 2213.00 3130.00 -64
3-day maximum 2249.00 0.50 1451.00 0.43 1724.00 2470.00 -14
7-day maximum 1951.00 0.48 1240.00 0.46 1447.00 1995.00 -57
30-day maximum 1344.00 0.68 1015.00 0.50 1088.00 1553.00 -57
90-day maximum 748.40 0.91 723.50 0.22 665.50 958.00 71
Base flow index 0.03 2.08 0.44 0.93 0.02 0.07 -57
Parameter Group #3
Date of minimum 130.00 0.21 337.00 0.46 108.10 154.20 -14
Date of maximum 224.00 0.05 239.00 0.05 220.60 233.90 29
Parameter Group #4
Low pulse count 4.00 1.00 0.00 0.00 3.00 5.72 -64
Low pulse duration 4.50 3.11 3.50 1.43 3.00 12.16 -69
High pulse count 4.00 1.25 3.00 3.33 3.00 4.72 -64
High pulse duration 4.50 2.44 34.75 3.26 2.00 9.00 -20
Parameter Group #5
Rise rate 39.30 1.20 33.34 0.35 23.48 45.74 114
Fall rate -3.40 -0.82 -18.83 -0.52 -4.67 -2.20 -100
Number of reversals 87.00 0.30 138.00 0.25 80.84 96.16 -100

Note: the unit for monthly flow, 1-, 3-, 7-, 30-, and 90-day maximum and minimum flow is in m3/s, unit for duration of high and
low pulse is days, and the rest of other IHA parameters are non-dimensional.

[Link] Magnitude and duration of annual extreme

Tekeze river time series of 1-day, 3-day, 7-day, 30-day, and 90-day maxima and minima medians

flow in the pre-impact and post impact periods together with the median value (dashed line) and

the boundaries of the middle category (i.e. from 25th and 75th percentile), which is computed with

53
reference to the pre-impact period are showed in Figure 3.9 and 3.10. In the RVA analaysis,

significance differences were obsreved in the annual maximum and minimum flows in the post

impact periods. The medians of annual 1-day, 3-day, 7-day, 30-day and 90-day minimum flow for

the post-impact period increased due to the reservoir attempt to capture high flood season flow for

later dry season for hydropower production. By contrast, as the medians of annual 1-day, 3-day,

7-day, 30-day and 90-day maximum flow for the post-impact period decreased greatly due to the

elimination of high magnitude flooding by storing water in the reservoir.

Except for low alteration in the 90-day annual maxima, the others were rather high. The

hydrological alteration of annual 1-day, 3-day, 7-day, and 30-day parameter values fell within the

RVA target value whereas the minima 1-day, 3-day, 7-day, 30-day and 90-day maxima medians

reached -100%, which means most values of these five parameters fell out of the RVA target value

(Table 3.4). The dispersion coefficients (CD) of annual minima and maxima flows in the post-

impact period ranging from 0.22 to 0.97 are generally lower than those in the pre-impact period

ranging from 0.48 to 1.92. The base flow index is larger in post-impact period because of low flow

season water released from reservoir for hydropower production when natural flow is at its

minimum. This is shown by a higher persistence of annual base flow index HA = 1.57 for the

upper category and accordingly lower persistence in lower and middle categories by negative HA

index of -1.0 and -0.57, respectively. Therefore, the result showed that daily, weekly, monthly and

annual maximal/minimal flow cycles were positively influenced by Tekeze hydropower reservoir

operation.

54
55
Figure 3.9 Hydrological alteration of 3-, 7-, 30-, and 90-day minimum flow before and after
Tekeze hydropower reservoir construction in Tekeze river

56
Figure 3.10 Hydrological alteration of 3- 7-, 30-, and 90-day maximum flow before and after
Tekeze hydropower reservoir construction in Tekeze river

[Link] High and low pulses

Low pulse count, low pulse duration, high pulse count and high pulse duration have been changed,
with hydrologic alteration of -64%, -69%, -64% and -20% respectively. Except high pulse
duration the median of low pulse count, low pulse duration and high pulse count in the post-impact
period were lower than those in the pre-impact period (Table 3.4). The coefficient of disperion in
the low pulse count and low pulse duration were higher in the pre-impact period, in contrast, the
high pulse count and high pulse duration value were higher in the post-impact period. This
indicates that the frequency and duration of low and high flow pulses in the Tekeze river had been
influenced by Tekeze hydropower reservoir construction.

57
3.4 Conclusive Remarks

In this study, the non-parametric Mann–Kendall trend test and Pettit test were used to investigate

the spatiotemporal trends and variability of precipitation data from different stations in Tekeze

basin on the annual and seasonal timescales for the period 1953–2013 and streamflow at

Embamadre for the period [Link] non-parametric Mann–Kendall test showed that annual

precipitation has an increasing trend in southern part of Tekeze basin varies from 0.98mm/year to

8.12mm/year whereas the decreasing trend prevails in the northern part of the basin which ranges

from -6.22mm/year to -2.35mm/year. There was no significant positive, or negative trends were

detected by the trend tests in annual precipitation in this basin. The analysis of the seasonal

precipitation time series showed a mix of positive and negative trends. In Belg and Kiremt season

more than 80% of the precipitation stations showed positive trend whereas Bega showed a similar

percentage but decreasing trend. All the three seasons show statically significant increasing and

decreasing precipitation trends with abrupt change detected in the late 1970s and mid-1980s. In

Belg and Kiremt seasons significant positive trends were found in 82% and 55% of the stations

respectively, whereas in Bega 73% of the station show significant negative trends. The strongest

positive trend of 2.14mm/year and 6.41mm/year was detected in Belg and Kiremt season at Debre

Tabor station respectively and negative trend of -1.33mm/year detected in Bega season at Nefas

Mewucha station.

Annual streamflow in Tekeze basin showed a non-significant increasing trend at a rate of

1.867m3/year. The seasonal streamflow showed similar projections like precipitation both

increasing and decreasing trend. Streamflow trend increases in the Kiremt and Belg seasons

whereas decreases in Bega (dry) periods. Statistical analysis performed using IHA at Embamadre

streamflow station shows an increased in minimum flow duration and also decreased in maximum

58
flow duration, fall and rise rate. After the inauguration of Tekeze Dam, the hydrology altered as

significant decline of high flows and increase of low flows, which were mainly attributed to stored

water during rainy season and releasing it during dry season.

The findings of this research can provide some information to the government and community on

the variability of rainfall and streamflow for current and future planned dams and irrigation

projects. Such information also used for policy makers and managers within the context of water

resource management, hydrology, agriculture, and ecosystem in Tekeze river basin.

59
4. CLIMATE CHANGE IMPACT ON THE WATER RESOURCES OF
TEKEZE BASIN2,3
Abstract: Climate change impact on Tekeze Basin hydrology studied using bias corrected

ensembles of CORDEX-Africa RCP4.5 and RCP8.5 climate scenarios projections of precipitation

and temperature. Soil and Water Assessment Tool (SWAT) used to simulate streamflow and

distribution mapping bias correction improves precipitation and streamflow simulations. Tekeze

basin showed an increase in mean temperature up to 1.07 OC for RCP4.5 and 2.21 OC for RCP8.5

climate scenarios in all time periods. Annual, long rain (Kiremt) and dry (Bega) seasons

precipitation also showed an increasing trend up to 48% whereas short rain (Belg) season showed

a decreasing trend up to 52% under both RCP climate scenarios for all future time periods. For

this study, SWAT performs well with values of Nash-Sutcliffe efficiency (ENS) and R2 greater than

0.7 for simulating monthly streamflow with a reasonable accuracy. Projected streamflow in the

basin will increase and its magnitude varies up to 58-66% in dry and 21-55% in rainy seasons for

both RCP4.5 and RCP8.5 scenarios in 2020s, 2050s and 2080s projected time periods. Results

from this study indicated that climate changes will alter quantity and timing of water that affect

the basin hydrology and water resources. This increasing streamflow for future time horizon is

advantageous for planning and operation of large water resources infrastructures in Tekeze basin

in particular, and in the downstream Nile Basin countries; Ethiopia, Sudan and Egypt in general.

Keywords: Tekeze River Basin; Water Resources; SWAT; CORDEX-Africa; RCP Scenario;

Climate Change

60
4.1 Introduction
Water resources issues faced around the world are complex, costly and challenging. Together with

activities of humans like land cover and land use change, deforestation, land degradation, soil

erosion, urbanization, population increase, changing social value and others (Liu et al., 2009),

climate change impacts water resources planning and management. Climate change may alter

quantity, quality, distribution and timing of water due to resulting changes in the hydrological

cycle. These changes affect different water resources operations and managements like

hydropower generation, irrigation practices, flood risk reduction, water quality, navigation,

ecosystem and a combination of these future water resources planning and managements. In order

to solve these water resources operation and management problems it is necessary to study the

impact of climate change on the hydrology of the watershed (Adem et al., 2014; Dessu and

Melesse, 2013; Grey et al., 2014; Melesse et al., 2011, 2009, Setegn et al., 2014, 2011).

It has been predicted that climate change and the resulting changes in precipitation and temperature

regimes will affect the availability of water resources in different regions of the world (Piao et al.,

2010; Setegn et al., 2011). African river basins are highly vulnerable to these changes. Many

studies conducted to assess climate change impacts on water resources of Africa (Beyene et al.,

2010; Gbobaniyi et al., 2014; Gizaw et al., 2017; Kim et al., 2014) especially in the Nile River

basin which have competent applications in ten riparian countries (Taye et al., 2011). From a total

flow of Nile River, about 86% of the flow contributes from Eastern Nile basin covering a total area

of 60% which originates from Ethiopian highlands (Arsano and Tamrat, 2005; Mengistu and

Sorteberg, 2012; Swain, 1997). This is the reason why several studies focus on Eastern Nile basin.

Many studies done in the Eastern Nile basin especially Upper Blue Nile basin (Beyene et al., 2010;

Enyew et al., 2014; Melesse et al., 2009; Mengistu and Sorteberg, 2012; Taye et al., 2011) to assess

61
the impact of climate change on water availability but only very little studies done on Tekeze river

basin part of Eastern Nile in the Northwest of Ethiopia (Gizaw et al., 2017).

The primary tools for projecting climate are GCMs that are typically run at a horizontal resolution

of 100km-250km due to their massive computational and data storage requirements. Since typical

impact assessment models require inputs at much finer spatial resolutions, GCMs data are often

downscaled using RCMs. RCMs are a widely used tool for producing regional climate data that

use boundary conditions from forcing coarse scale global data sets such as a GCMs or reanalysis

to simulate the climate of a region. To systematically explore uncertainties associated with

downscaling, coordinated RCM simulations are necessary. Now a day’s such coordinated efforts

are available in different regions of the world. The CORDEX is a new initiative by World Climate

Research programme (WCRP) (Giorgi et al., 2009) aims to foster international collaboration to

generate an ensemble of high resolution historical and future climate projections at regional scale,

by downscaling the GCMs participating in the CMIP5 (Taylor et al., 2011). Africa is a prime focus

of this initiative due to its heightened vulnerability to climate change, and poorly developed

adaptation structure. More recently, analyses in relation to CORDEX simulations over Africa can

be found in the work of different researchers (Hernández-Díaz et al., 2013; Jacob et al., 2012;

Nikulin et al., 2012). Nikulin et al. (2012) evaluate the ability of ten RCMs over Africa and

conclude that all RCMs simulate the seasonal mean and annual cycle quite accurately. Likewise,

it is verified that the mean of multi model outputs do better than individual simulation. Hernández-

Díaz et al. (2013) strengthen the achievement of Nikulin et al. (2012). They successfully reproduce

the overall features of geographical and seasonal distribution over most Africa. In their report,

CORDEX simulations succeed in reproducing the average distribution of precipitation and its large

geographical differences. Jacob et al. (2012) have integrated Regional MOdel (REMO) over six

62
CORDEX continents and found that REMO is well suited to examine projected future changes in

all these domains despite wet and dry biases appear over the mountainous regions and East Africa,

respectively. On the other hand, the African Monsoon Multidisciplinary Analysis-Model

Intercomparison Project (AMMA-MIP) provides contribution relevant to CORDEX and achieves

good correlation between the accumulated rainfall over the Sahel and the latitude of the African

Easterly (Ruti et al., 2011). Ensemble outputs from CORDEX experiment are not only a

compulsory for climate studies, and a roadmap to adaptation and mitigation strategies but also

good source of data for us to apply them to our region, Ethiopia, where climate change is already

observed (Conway, 2005; Elshamy et al., 2009; Mengistu and Sorteberg, 2012).

Nowadays there is available most recent Representative Concentration Pathways (RCPs) scenarios

(Meinshausen et al., 2011) of the fifth phase of CMIP5 developed for fifth IPCC assessment report

(AR5) (Taylor et al., 2011). CMIP5 incorporated new paradigms for developing future emission

scenarios, introduced experiments to explore carbon-climate interactions and have finer resolution

process and well-integrated earth system components than the earlier phase of CMIP model an

ongoing process without specific end date, example third phase of [Link] CMIP5 scientific

gaps like poor quantification and understanding of radiative forcing (more coordinated

representation of the atmospheric aerosol and land surface processes) and model biases will be

considered in the next new sixth phase of Coupled Model Intercomparison Project (CMIP6).

CORDEX after applying on an ensemble RCMs for multiple GCMs produced high resolution

downscaled historical and future climate data based on CMIP5 simulations used for impact and

adaptation studies (Gbobaniyi et al., 2014; Kim et al., 2014; Nikulin et al., 2012; Panitz et al.,

2014). However, these high resolutions downscaled RCM simulations may not be directly used in

hydrological model for hydrological impact assessment in watershed scale due to biases

63
(Christensen et al., 2008; Graham et al., 2007; Kim et al., 2015; Teutschbein and Seibert, 2010).

These biases should be corrected if realistic future hydrology requires (Rojas et al., 2011) using

different bias correction methods (Teutschbein and Seibert, 2012).The climate projection

framework within CORDEX is based on the set of GCM simulations in support of the IPCC fifth

assessment data referred to as CMIP5 climate projections. CORDEX focus on the GCM

experiments using new set of climates forcing scenarios known as RCPs which represents RCP4.5

(mid-level) and RCP8.5 (high-level) emission scenarios. The definition of the RCPs allows for a

parallel development of new socioeconomic, technical and policy scenarios that provide insights

into the impact of policy decisions on the future climate (Vuuren et al., 2011). The RCPs focus on

the ‘concentrations’ of greenhouse gases that lead directly to a changed climate and include a

‘pathway’ the trajectory of greenhouse gas concentrations over time to reach a particular radiative

forcing at 2100.

Hydrological models that allow for a description of the hydrology of the region play an important

role in predicting river discharges from gauged and ungauged catchments and understanding the

rainfall–runoff processes in the catchments to enhance hydrological and water resources analysis.

As such, a number of models have been developed and applied to study the water balance, soil

erosion, climate and environmental changes in the Blue Nile Basin (Conway, 1997; Johnson and

Curtis, 1994; Kebede et al., 2006; Mekonnen et al., 2009; Tekleab et al., 2011; Tilahun et al.,

2013). But some of these models have limitations which include inappropriate scale, inability to

perform continuous-time simulations, inadequate maximum number of sub watersheds, and the

inability to characterize the watershed in enough spatial detail. Soil and Water Assessment Tool

(SWAT) (Arnold et al., 1998; Neitsch et al., 2005), a watershed scale physically-based simulation

model, was developed to overcome these limitations. The SWAT model offers continuous-time

64
simulation, high level of spatial detail, unlimited number of watershed subdivisions, efficient

computation, and capability to simulate changes in land-management. Besides its obvious

advantage as a hydrological modeling tool that includes modularity, ability to predict long-term

impacts as a continuous model, and ability to use readily available global datasets, availability of

a reliable user and developer support has contributed to its acceptance as one of the most widely

adopted and applied hydrological models worldwide (Gassman et al., 2014). SWAT has been

tested for a wide range of regions, conditions, practices and time scales (Gassman et al., 2007).

SWAT is increasingly used for climate change assessment applications all around the world

(Gassman et al., 2007; Schuol et al., 2008) and in the Upper Blue Nile Basin (Mengistu and

Sorteberg, 2012; Setegn et al., 2010b; van Griensven et al., 2012).These studies showed that the

SWAT model could describe the study areas with a quality that makes it suitable for water resource

planning and management use. Hence, SWAT is a suitable hydrologic model to assess impacts of

climate change on the water resources of Tekeze River basin part of the Eastern Nile Basin.

For the Tekeze basin, we focus on water as the key resource for development and flood security

as well as economic development and livelihood and especially for hydropower production.

Moreover, the most up-to-date knowledge the model outputs from CMIP5 of CORDEX-Africa for

the representative concentration pathways, RCPs, to investigate climate impacts in the basin water

resources. In this study, a climate change impact assessment was carried out for the Tekeze basin,

which contributes a significant flow to the main Nile. The Ethiopian government has planned to

increase reservoir water storage in this basin to increase hydropower generation and irrigation to

facilitate growth and stabilize the national economy. In this regional specific Tekeze basin study,

CORDEX-Africa that focused on bias corrected ensembles of GCMs/RCMs experiments using

mid (RCP4.5) and high (RCP8.5) level emission scenarios were used.

65
The objectives of this study are to (1) test SWAT hydrological model in the Tekeze Basin (2)

investigate the changes and trends in future precipitation and temperature across Tekeze river

basin, and (3) assess the climate change effect on the future stream flow magnitude (water

availability) under bias corrected ensembles of CORDEX-Africa RCM outputs of mid-range

(RCP4.5) and high-level (RCP8.5) climate scenarios in this basin.

4.2 Study Area and Datasets

4.2.1 Study Area

This study focuses on the Ethiopian portion of Tekeze basin (Figure4.1) with topographic location

11O40’ to 14O32’N and 36O30’ to 39O50’E with catchment area 43,000km2 upstream of

Embamadre gauged station. The river begins at the spring near Lalibela in the central Ethiopian

highlands. The topography varies from 4455meter above sea level (masl) in the Semein Mountains

to 589masl in the Metema lowland areas. This topographic variation has a high potential for

hydropower production in the mountainous areas and irrigation lands in the lowlands.

Climate of Tekeze basin varies depending on altitude. Rainfall increases as altitude increases but

temperature decreases as altitude increases. Minimum and maximum temperature varies from 3-

21OC in the highlands of Semein Mountains and 19-43OC in the lowland areas. The rainfall varies

from 600mm in the lowlands to 1300mm in the highland areas. During the months from July to

August 70% rainfall falls in the area. There are three seasons controlling the climate in Tekeze

basin. These are the dry (Bega) season runs from October to January, no rain and dry in all parts

of the basin. Small rain (Belg) season runs from February to May and most part of the basin receive

considerable amount of rainfall. Wet (Kiremt) season runs from June to September all parts of the

basin receive rain.

66
The main soils of Tekeze basin as per FAO classification are Eutric Cambisols, Calcric Cambisols

and Eutric Vertisols mainly found in the highland and lowland areas of the watershed and Eutric

Leptosols in the central part of the basin. Most of the basin areas are covered with rainy season

agriculture and cultivated land in the form of intensively cultivated, seasonally cultivated lands

and cultivated land with scattered trees and shrubs. Generally, the basin land use and land cover

includes more than 70% of cultivated land and the remaining are shrub land (Range grasses),

mixed forests, pasture/ grazing lands and exposed rocks (Tesfaye et al., 2017) . Major crops grow

in this basin are Teff, Wheat, Barley, Maize and other Cereals.

67
Figure 4.1 Location of upper Tekeze basin upstream of Embamadre gauging station

Figure 4.2Sub-basins of Tekeze basin study area watershed delineated by SWAT

68
Figure 4.3Mean monthly precipitation (mm) and maximum temperature (Tmax) & minimum

temperature (Tmin) (◦C) for selected stations of Tekeze basin

4.2.2 Dataset Used

For climate change impact assessment studies using hydrological model different climatic and

spatial data were used. These were digital elevation model, streamflow, temperature; precipitation

and land use and land cover data.

69
[Link] Digital Elevation Model (DEM)

DEM used to delineate the watershed and the drainage patterns of the surface area analysis in the

SWAT model. Sub-basin parameters such as slope gradient, slope length of the terrain, and the

stream network characteristics such as channel slope, length, and width were derived from DEM.

This study used a DEM that was a processed Shuttle Radar Topographic Mapping (SRTM)

30mx30m resolution topographic map obtained from Ethiopian Ministry of Water, Irrigation and

Electricity, GIS and Remote Sensing department.

[Link] Soil and Land use data

Soil, land use/land cover data were used as an input into SWAT hydrological model to delineate

sub-watersheds further into hydrologic response units (HRUs). The Soil data used includes the

information to describe the physical and chemical properties of the soil like soil texture, hydraulic

conductivity, bulk density, water content, organic carbon content and percentage of sand, silt, and

clay content for each soil horizon. Land use/ land cover influences the hydrological properties of

the watershed and used as an input for SWAT model. Tekeze basin land use/ land cover changes

time to time due to several factors mainly changing agricultural practice, over increase in

population density, urbanization, new hydropower and irrigation development. This basin is

covered with agricultural lands, forests, grasses, bushes and shrubs. Soil, Land use/ land cover data

of Tekeze basin were obtained from Ethiopian Ministry of Water, Irrigation and Electricity in

shape file format. The land use and land cover of Tekeze basin after reclassification by SWAT

includes 64% of agricultural land generic (AGRL), 14% of shrub land (range grasses, RNGE),

13% of mixed forests (FRST), and 9% of pasture/grazing lands (PAST).

70
Figure 4.4Major land use/land cover map of Tekeze Basin study area re-classified by SWAT

Figure 4.5 Major soil map of Tekeze Basin study area re-classified by SWAT

71
[Link] Climate Data

The SWAT model requires daily climate data of precipitation, maximum and minimum

temperature to simulate the hydrologic response. There are numerous weather stations in Tekeze

basin. For this study, stations recording precipitation and temperature data that have long period

of records with small data gaps were used. The climate data for eleven stations (Figure 2.2) found

within and near by the Tekeze basins from 1976-2013 collected from Ethiopian National

Meteorological Service Agency. However, there are missing data on average less than 8% for both

precipitation and temperature. The three steps visual inspection, comparison to the nearest station

with in the same zone and regression relations between neighbouring stations and SWAT built in

weather generator were taken to detect outliers and fill in the missing gaps in the data series. The

SWAT built in weather generator used a first-order Markov chain model. For each sub basin input

to the weather generator was observed precipitation data for the weather station that was nearest

the centroid of the sub-basin and having a record length from 1994–2008. Given the observed wet

and dry days’ frequencies, the model determines stochastically if precipitation occurs or not. When

a precipitation event occurs, the amount is determined by generating values from a skewed normal

daily precipitation distribution or a modified exponential distribution which is calculated based on

the observed data.

[Link] Streamflow/River Discharge Data

The hydrological discharge (streamflow) daily data was required for performing sensitivity

analysis, calibration and validation of the SWAT model. In the Tekeze River basin most of flow

gauging stations located on upper part of the basin relatively small tributaries and/or near the head

waters of the main river covering small catchment area with short periods of record, large data

gaps, and high amount of missing data which was discussed in previous chapter, but Embamadre

72
station has relatively long record periods and cover almost half of the basin area. Daily stream

flow data of the Tekeze basin at Embamadre gauged station (Figure 2.2) which has 15 years of

record period continuous data from 1994-2008 was collected from Ethiopian Ministry of Water,

Irrigation and Electricity.

[Link] Climate Scenario Data

In this research, climate change scenarios data from the newly available CMIP5 (Gbobaniyi et al.,

2014; Kim et al., 2014; Nikulin et al., 2012; Vuuren et al., 2011) RCM ensemble output of

CORDEX-Africa for African domain projections under Representative Concentration Pathways

(RCP4.5 and RCP8.5) were used as input to hydrological model. Currently, CORDEX-Africa

initiated by WCRP provides an opportunity for the generation of high resolution regional climate

projections over Africa that is used to assess future impacts of climate change at regional and local

scales. In this study, results of CORDEX-Africa ensemble RCMs simulations for the past (1951–

2005) and future (2006–2100) climate projections downscaled from different GCMs under RCP4.5

and RCP8.5 with spatial resolution of 0.44° is used. CORDEX-Africa RCMs generate an ensemble

of high resolution baseline (past) and future climate projections at regional scale by downscaling

different GCMs forced by RCPs based on the CMIP5 (Nikulin et al., 2012; Vuuren et al., 2011).

CORDEX-Africa climate projections use RCP4.5 and RCP8.5 climate scenarios.

RCPs are new climate change scenarios established by CMIP5 (Meinshausen et al., 2011; Taylor

et al., 2011a), which can depict a wide variety of possible future climate scenarios. RCP scenarios

have a better resolution that helps in performing regional and local comparative studies compared

to previous climate scenarios, and RCP scenarios also represents an attractive potential approach

for further research and assessment, including emissions mitigation and impact analysis. The fifth

Assessment Report (AR5) scientific literature selects one mitigation scenario (RCP2.6), two

73
medium stabilization scenarios (RCP4.5 and RCP6.0) and one high emission scenario (RCP8.5).

RCP2.6 scenario sees emissions peak early, then fall shown to be technically feasible. But one of

RCP2.6 scenario key assumptions is the full participation of all developed and developing

countries in the world in the short run to reduce all the main emitters, which is not possible in

actual cases. Due to this, it is decided to choose one medium scenario (RCP4.5) and high scenario

(RCP8.5) covering entire range of radiative forcing. RCPs represent pathways of radiative forcing,

not linked with exclusive socio-economic assumption in contrary to Special Report on Emission

Scenarios (SRES). Any single radiative forcing pathway can result from a diverse range of socio-

economic and technological development scenarios. RCP4.5 is a mid-range scenario that stabilizes

radiative forcing at 4.5 W/m2 (approximately 650 parts per million (ppm) CO2-equivalent) in the

year 2100 without exceeding this value, but this does not imply the climate system are stable

(Thomson et al., 2011; Vuuren et al., 2011). Whereas RCP8.5 is upper bound of all RCP scenarios

that stabilizes radiative forcing at 8.5 W/m2 (greater than 1370 ppm CO2-equivalent) in the year

2100 (Riahi et al., 2011; Vuuren et al., 2011).The works of (Meinshausen et al., 2011; Riahi et al.,

2011; Thomson et al., 2011; Vuuren et al., 2011) briefly describe each RCPs (RCP4.5 and RCP8.5)

in detail and discussed the modeling system that have created each RCP, the main socio-economic

assumptions, the underline trends in energy use and detail on emissions and land use.

Precipitation and temperature data from these scenarios are available from an ensemble of

CORDEX-Africa regional climate model for the domain of Ethiopia at 0.44O resolution from 1951-

2005 for baseline (past) used to calibrate and validate with observed data and 2006-2100 for future

periods which is expected to capture a reasonable range in climatic and hydrological projections

for Tekeze Basin.

74
4.3 Methodology

The most common approach for hydrological impact studies of climate changes is to run

hydrological models with climate scenarios, usually provided by the outputs of precipitation and

temperature from climate models bias corrected to the catchment of interest. In this study,

ensemble of CORDEX-Africa RCMs under RCP4.5 and RCP8.5 climate scenarios with bias

correction methods were used as an input for SWAT to assess future water resources of Tekeze

Basin.

Figure 4.6 Conceptual diagram of the SWAT modeling process with climate change scenarios

75
4.3.1 Bias Correction Method

Numerous studies have assessed the impacts of climate change on water resources using climate

variables from global climate models (GCMs) and water resources models. However, because of

the relatively low spatial resolution (100–250 km) of GCMs, regional climate models (RCMs)

are widely used for regional impact studies at catchment scales (25–50 km) climate variables.

Although RCMs are able to simulate local climate at finer resolutions, outputs from RCMs cannot

be used as direct input data for hydrological models due to systematic errors and require post-

processing of the model outputs to remove biases (Christensen et al., 2008). Typical systematic

model errors of RCMs include errors in estimation (over or under) of climate variables, incorrect

estimates of seasonal variations of precipitation (Christensen et al., 2008), and simulations of

more wet days of low intensity rainfall (drizzle effect) than they actually observed (Ines and

Hansen, 2006). The physical characteristics of precipitation make it more difficult to correct and

most studies tend to concentrate on precipitation correction. Methods of different complexity

have been put forward, aimed at correcting various moments of the rainfall distribution.

Bias correction procedures employ a transformation algorithm for adjusting RCM output. The

underlying idea is the identification of possible biases between observed and simulated variables,

which is the bias for correcting both control and scenario RCM runs. A crucial assumption of

bias correction is stationary of the bias, i.e., the correction algorithm and its parameterization for

current climate conditions are also valid for future conditions (Christensen et al., 2008;

Teutschbein and Seibert, 2012). Therefore, biases between the models are in general relatively

stable, such that bias correction on average considerably improves future scenarios for many

regions and all seasons. Bias adjustments methods usually do not account for the origins of biases

in climate models and instead perform empirical adjustments. Bias correction methods of varying

76
complexity are applied to precipitation and temperature variables. This research limited to these

two variables as they are the main atmospheric drivers for most impact and hydrological models,

and due to the availability of high quality observational data records.

Various methodologies have been developed over the last decades to perform bias correction,

from very simplistic methods, such as the so-called delta method only correcting the statistical

mean of the simulations, to more sophisticated ones for example based on distribution functions

(Piani et al., 2010; Teutschbein and Seibert, 2012). In this research, five methods, i) the delta

change approach ii) Linear scaling, iii) local intensity scaling, iv) variance scaling, and v)

distribution mapping were explored to adjust raw ensemble of CORDEX-Africa RCMs RCP

scenarios output simulation data.

i. Delta Change Method

The delta change method consists of altering an observed (reference) climate series with change

factors to obtain a new series representative of future changes. For the flux variables like

precipitation relative change factors applied whereas for state variables like temperature absolute

change is applied. Monthly change factors are derived and perturbed as follows for day and

month (i, j), where i = 1, 2, 3, …., 31 and j = 1, 2, …., 12:

Pobs  Pref (4.1)

avg
Pfut
P(i , j )   p ( j ) * Pobs (i , j ) ;  p ( j )  ( j)
(4.2)
Prefavg( j )

Tobs  Tref (4.3)

avg avg
T(i , j )  T ( j )  Tobs (i , j ) ; T ( j )  T fut( j )  Tref ( j ) (4.4)

Where P (i, j) and T (i, j) are delta change perturbed daily climate change variables, P obs(i, j) and

Tobs(i, j) are observed precipitation and temperature climate variables in the reference period, ΔP(j)

77
and ΔT(j) are the changes in climate as simulated by RCM-RCP scenarios and Pavg(j) and Tavg(j)

are daily precipitation and climate means by month, the index ref indicates the reference (control)

and fut indicates a future period.

The delta change method differs from other bias correction methods because it uses observations

and only the RCM change signal but does not adjust the RCM simulations. Therefore, the delta

change method cannot be evaluated for the control (baseline period) run as it gives perfect

simulation by definition (equation 4.2 and 4.3 not valid in actual case). Due to this limitation the

delta change method could not be assessed in this study.

ii. Linear scaling

The linear scaling approach operates with monthly correction values based on the differences

between observed and present day simulated precipitation and temperature raw data (raw

simulated ensemble of CORDEX-Africa RCMs RCP scenarios data in this case). Linear scaling

method aims to perfectly much the monthly mean of corrected RCM simulations with the

observations. Precipitation is typically corrected with a multiplier and temperature with an

additive term on a monthly scale. The applied correction factors and addends are assumed to

remain unvaried even for future conditions.

Precipitation is corrected with a factor based on the ratio of long term monthly mean observed

and control run data:

mpobs
, m ( d )  Praw , m ( d ) x
*
Pconcor (4.5)
mpcon

mpobs
, m ( d )  Pscen , m ( d ) x
*
Pscencor (4.6)
mpcon

Where P*concor, m (d) is corrected precipitation of control period on the dth day of mth month, Praw,

m (d) is the present raw precipitation on the dth day of mth month from 1976-2005, Pscen, m (d) is the

78
future raw precipitation on the dth day of mth month from 2011-2100, P*scencor, m (d) is corrected

precipitation of future scenario on the dth day of mth month, µmpobs the mean value of observed

precipitation at a given month m and µmpcon the mean value of control period raw simulated

precipitation at a given month m.

Raw simulated temperature data is corrected with the help of an additive term based on the

difference of long-term monthly mean observed and control run data:

, m ( d )  Traw , m ( d )   mtobs   mtcon

*
Tconcor (4.7)

, m ( d )  Tscen , m ( d )   mtobs   mtcon

*
Tscencor (4.8)

Where T*concor, m (d) is corrected temperature of control period on the dth day of mth month, Traw, m

(d) is the present raw temperature on the dth day of mth month from 1976-2005, Tscen, m (d) is the

future raw temperature on the dth day of mth month from 2011-2100, T*scencor, m (d) is corrected

temperature of future scenario on the dth day of mth month, µmtobs the mean value of observed

temperature at a given month m.

The temperature at a given month m and µmtcon the mean value of control period raw simulated

linear scaling approach adjusts monthly mean values and offers corrected data with a variability

more consistent with the raw RCM-RCP scenarios data. Similar to delta change approach, the

downside is that all events are adjusted with the same correction factor. It is furthermore not able

to correct frequencies.

iii. Local intensity scaling of precipitation

The local intensity method corrects the wet-day frequencies and intensities which can effectively

improve the raw precipitation data which have too many drizzle days (days with little

precipitation). It normally involves two separate steps:

79
1. The wet-day threshold for the mth month Pthres, m is determined from the raw precipitation

series to ensure that the threshold exceedance matches the wet-day frequency of observations.

The number of precipitation events for both control (1976-2005) and scenario (2011-2100)

run are corrected by applying the calibrated RCM precipitation threshold (Pthres, con) so that all

days with precipitation less than Pthres, conare redefined to dry days with 0mm precipitation:

0, if , Pcon( d )  Pthres , con

a
Pcon (d )  
Pcon( d ) , otherwise
(4.9)

0, if , Pscen ( d )  Pthres ,con

a
Pscen (d )   (4.10)
Pscen ( d ) , otherwise

Where Pcon (d) is the present raw precipitation on the dth day from 1976-2005, Pthres, con is

the present raw precipitation whose value greater than the threshold value, P acon(d) is the

present adjusted dry and wet-day precipitation of control period on the dth day, Pscen (d) is

the future scenarios raw precipitation on the dth day from 2011-2100 and Pascen(d) is the

future adjusted dry and wet-day precipitation of future period on the dth day. This

procedure allows the scenario run to have a different wet-day frequency than the control

and scenarios run.

2. A linear scaling factor (s) is estimated based on the long term monthly mean wet-day

intensities. This intensity factor is calculated by considering only wet-days (observed days

with precipitation larger than 0mm) in to account and the RCM-RCP scenarios simulated

days with precipitation larger than the adjusted precipitation threshold (Pthres, con).

 pobs , m ( d )
s (4.11)
 praw, m ( d )

Where s is linear scaling intensity factor, µpobs, m(d) the mean value of observed wet-day

precipitation (Pobs (d)> 0) at a given dth day and µpraw, m(d) the mean value of control period

80
 raw simulated precipitation greater than the controlled threshold value (Praw, m(d) > Pthres,

con) at a given dth day.

This factor used to ensure that the mean of the corrected precipitation is equal to that of

the observed precipitation. Finally, the RCM-RCP scenarios of present and future

precipitation are corrected as:

( d )  s * Pcon ( d )
* a
Pcon (4.12)

( d )  s * Pscen ( d )
* a
Pscen (4.13)

Where P*con(d) is corrected precipitation of control period on the dth day of mth month,

P*scen(d) is corrected precipitation of future scenario on the dth day of mth month.

Local intensity scaling is an improvement of the linear scaling approach, because it combines the

linear scaling advantages with a correction of the wet-day frequencies (precipitation threshold).

iv. Variance scaling of temperature

The variance scaling method was developed to correct both the mean and variance of normally

distributed variables such as temperature (Teutschbein and Seibert, 2012). Temperature is

normally corrected using variance scaling method with the following equation:

 Tobs,m
, m ( d )  Traw , m ( d )  (  Traw, m * )  Tobs,m
*
Tconcor (4.14)
 Traw,m

 Tobs, m
, m ( d )  Tscen , m ( d )  (  scen , m * )  Tobs, m
*
Tscen (4.15)
 Traw, m

Where T*concor, m (d) is corrected temperature of control period on the dth day of mth month, Traw, m

(d) is the present raw temperature on the dth day of mth month, Tscen, m (d) is the future raw

temperature on the dth day of mth month, T*scencor, m (d) is corrected temperature of future scenario

on the dth day of mth month, µTobs,m the mean value of observed temperature at a given month m,

81
µTraw,m the mean value of control period raw simulated temperature at a given month m, µ scen,m

the mean value of future period raw simulated temperature at a given month m, σTobs,m is the

standard deviation of observed temperature at a given month m, and σ Traw,m is the standard

deviation of control period raw simulated temperature at a given month m.

The variance scaling approach guarantees that the adjusted RCM control run has the same mean

and standard deviation (i.e. variance) as the observed time series. Like the other bias corrected

methods, the correction factors assumed to remain the same for future conditions, but allow for

changes in response between control and future scenarios run. Variance scaling adjusts both the

variance and the mean of raw RCM data. It performs much better than the linear scaling approach

in terms of correcting several statistical characteristics and in terms of the variability range. It

should be noted that variance scaling is not advisable as it is based on the invalid assumption that

all local variability is related to larger-scale variability and, furthermore, tends to augment the

mean square errors of uncorrected data.

v. Distribution Mapping

Distribution Mapping (DM) bias correction method (Piani et al., 2010) which can better transfer

the observed precipitation and temperature statistics to the raw ensemble GCM/RCM CORDEX-

Africa RCP4.5 and RCP8.5 climate scenarios were used. Distribution mapping was found to be

the best correction method for both climate projections and hydrological impact qualifications;

it performed especially well in terms of the simulation of hydrological extremes. To date, this is

probably one of the most thorough studies interms of inclusion of the most bias correction

approaches for hydrological impact studies.

Precipitation is more challenging to correct its bias due to its physical characteristics, but it has

more significant influence than temperature on streamflow simulation of Tekeze River basin.

82
Therefore, in this research, the precipitation bias corrected using DM gamma distribution

function with maximum-likelihood parameter estimation method. The gamma distribution

recommended function for distributions of precipitation with shape parameter  and scale

parameter  , and it has been proven to be effective (Fang et al., 2015; Piani et al., 2010). The

gamma distribution-based correction method assumes that the probability distributions of both

observed and simulated daily precipitation datasets can be approximated using a gamma

distribution.

The probability density function for gamma distribution is given by:

x
1 
f ( x,  ,  )   x  1e  (4.16)
 ( )

Cumulative distribution function

x
F ( x)   f ( x)dx (4.17)
0

Where f is the distribution function, e is Euler’s number, x is an independent (random) variable

and Г(α) indicates the gamma function evaluated at α. α is the shape parameter and β is the scale

parameter of the gamma distribution, estimated by the maximum likelihood estimation method.

The gamma function is defined by


( )   x  1e  x dx For α>0 (4.18)
0

This method was used to adjust the raw RCP4.5 and RCP8.5 climate scenarios precipitation data

mean, standard deviation, quantile and it preserves the extremes (Themeßl et al., 2012).

Ensemble RCM of CORDEX-Africa RCP4.5 and RCP8.5 climate scenarios in Tekeze basin

simulate a low number of dry days which are compensated by too much drizzle. In this research

for the local intensity scaling and DM methods small precipitation magnitude, less than 0.20mm

83
were excluded from this bias correction simulation to reduce much drizzle effects. This method

corrects all ranges of the systematic errors (i.e., biases) concerning each meteorological station

which is inherited in raw CORDEX-Africa RCP4.5 and RCP8.5 climate scenarios of

precipitation data.

For temperature time series, the Normal (Gaussian) distribution with location parameter μ and

scale parameter σ is usually assumed to fit best:

  x   2


fN X /  ,  2
 x  1 1
.e 2 2
(4.19)
 . 2

The scale parameter σ determines the standard deviation, i.e., how much the range of the

Gaussian distribution is stretched or compressed. A smaller value for σ results in a more

compressed distribution with lower probabilities of extreme values. By contrast, a larger value

for σ indicates a stretched shape with higher probabilities of extreme values. The location

parameter μ directly controls the mean and, therefore, the location of the distribution.

4.3.2 Hydrological Modeling Using SWAT

In this study, Soil and Water Assessment Tool (SWAT) (Arnold et al., 2013), one of the

extensively used hydrological model around the world (Gassman et al., 2007) and in Ethiopia

(Dessie et al., 2014; Mengistu and Sorteberg, 2012; van Griensven et al., 2012) was applied for

the simulation of stream flows in the Tekeze Basin.

SWAT is widely used to simulate hydrological processes under the scenario of changes in land

use, land management as well as climate change. Various studies used SWAT for watershed scale

hydrological modeling in the USA (Wang et al., 2008a, 2008b, 2006; Wang and Melesse, 2006),

Italy (Fiseha et al., 2014, 2013), Kenya (Dessu et al., 2014; Dessu and Melesse, 2013, 2012),

Ethiopia ( Mengistu and Sorteberg, 2012; Setegn et al., 2009, 2010b, 2010a, 2011; Tibebe and

84
Bewket, 2011; van Griensven et al., 2012; Yesuf et al., 2015) and Jamaica (Grey et al., 2014;

Setegn et al., 2014).

SWAT model is a physically based, continuous time watershed model which simulates

hydrological processes in the watershed. The details of SWAT were shown in Neitsch et

al.(2005).The SWAT used together with ArcSWAT in ArcGIS Geographical Information System

interface to process the datasets and construct the required input for the initial modeling setup.

In SWAT model application, the study basin is first sub divided in to sub-basins based on digital

elevation model and channel network, and further delineated into hydrological response units

(HRUs) considering dominant soil/land use category in each sub-basin where each sub-basin was

assumed to be constituted with a homogeneous soil, land use and climate. Routing of water is

simulated from the HRUs to the sub-basin level, and then through the stream network to the basin

outlet. The model predicts the hydrology ultimately streamflow (Neitsch et al., 2005) at each

HRU using water balance equation, which contains precipitation, surface runoff, evapo-

transpiration, infiltration and subsurface inflow. The water balance equation of the hydrologic

cycle is:

SWt = SWO + ∑ti=1(R day − Qsurf − Ea − Wseep − Qgw ) (4.20)

in which SWt is the final soil water content (mm), SWO is the initial soil water content on day i

(mm), t is the time in days, R day is the amount of precipitation on day i (mm), Qsurf is amount of

surface runoff on day i (mm), Ea is amount of evapotranspiration on day i (mm), Wseep is amount

of water entering to vadose zone from the soil profile on day i (mm), and Qgw is amount of return

flow on day i (mm).

When the rate of precipitation exceeds the rate of infiltration, surface runoff occurs. SWAT offer

the methods of the Soil Conservation Service (SCS) curve number procedure (United States Soil

85
Conservation Service., 1972), and the Green &Ampt infiltration method (Green and Ampt, 1911)

to estimate surface runoff. SCS curve number method was selected for this basin SWAT model

simulation as the SCS curve number procedure uses daily precipitation and the Green &Ampt

infiltration method requires precipitation input in sub-daily scale. In the SCS method retention

parameter defined by Curve Number (CN) is significant, and it is a sensitive function of the land

use, soil’s permeability and antecedent soil water conditions. Surface runoff was estimated using

Soil Conservation Service Curve Number (SCS-Curve number) method:

2
(Rday −Ia )
Qsurf = (P (4.21)
day −Ia −S)

Where Ia is the initial abstraction which includes surface storage, interception and infiltration

prior to runoff and S is the retention parameter (mm)

Retention parameter defined by

1000
S = 25.4 ( − 10) (4.22)
CN

CN is curve number for the day varies from 0 to 100 depending on soil permeability, land use

and the antecedent soil water condition.

Initial parameter approximated as 0.2S, Equation (4.21) becomes:

2
(Pday −0.2S)
R sur = (Pday +0.8S)
(4.23)

Where R surf is amount of surface runoff on day i (mm)

86
[Link] SWAT Model setup

The SWAT model utilizes the DEM to create stream network, sub-basin and delineate the

watershed boundary of Tekeze River basin using the elevation or topographic data and also

calculates the sub-basin parameters for example the longest path distance, the centroid and the

slope of sub-basin. Tekeze stream network and sub-watersheds were delineated using

ArcSWAT2012 integrated in ArcGIS version 10.2.2 suggested drainage area required to form

the origin of the stream i.e. 23 sub-watersheds; and 277 HRUs of Tekeze river were delineated

up to the point of outlet of a drainage area of 43,000 km2 at Embamadre gauging station.

Hydrological Response Unit (HRU) which is the smallest unit of the model to route streamflow

was established after reclassification of land use and soil maps and then overlaid one upon each

other. The threshold value was assigned to subdivide the watershed area into HRU which an area

is having unique land use, soil and slope combinations. Runoff is predicted separately for each

HRU and routed to obtain the total runoff for the watershed. This improves the accuracy of runoff

predictions of the model and provides a much better physical description of the water balance.

All model input data were projected to the same projection of Transverse Mercator projection.

[Link] Sensitivity analysis, Calibration and Validation of SWAT

It is necessary to identify most sensitive water flow parameters and the parameter precision for a

given watershed or sub-watershed required for calibration and validation process of SWAT

model. Generally local and global sensitivity analysis performed, and this analysis may yield

different results. Local sensitivity analysis is performed by changing values at a time whereas

global sensitivity analysis by allowing all parameter values to change at a time. Sensitivity of

one parameter often depends on the value of other related parameters; hence, the problem with

one-at-a-time analysis is that the correct values of other parameters that are fixed are never

87
known. The disadvantage of the global sensitivity analysis is that it needs many simulations.

Both procedures, however, provide insight into the sensitivity of the parameters and are necessary

steps in model calibration. After pre-processing of the Tekeze basin data and ArcSWAT2012

model set up, simulation was done. The built-in SWAT sensitivity analysis tool that uses the

Latin Hypercube One-factor-AT-a- Time (LH-OAT) was used to identify sensitive flow

parameters in Tekeze River Basin. Parameters identified from the sensitivity analysis were

varied in sequence of their relative sensitivity within their ranges (Table 4.1) until the volume is

adjusted to the required quantity. Out of 26 SWAT sensitive to water flow parameters the curve

number (CN2), available water capacity (Sol-AWC), Surface runoff lag coefficient (Surlag),

saturated hydraulic conductivity (SOL_K), soil evaporation compensation factor (ESCO),

Groundwater revap coefficient (GW_REVAP), Ground water delay time (GW_DELAY),

threshold water depth in the shallow aquifer for flow (GWQMN), Manning coefficient for main

channel (CH_N2), the shallow aquifer for “revap” (RVPMN) and base flow alpha factor (

ALPHA_BF) were identified as being parameters to which the flow has medium, high or very

high sensitivity. The ranking of the parameters is presented in Table 4.1. The curve number

(CN2) was the main sensitivity parameter. This is so because the curve number depends on

several factors including soil types, soil textures, soil permeability, land use properties etc. In

addition, the relative sensitivity of the available water capacity (Sol-AWC), the soil evaporation

compensation factor (ESCO) and the saturated hydraulic conductivity (Sol-K) were also highly

sensitive.

88
 Table 4.1 Order of sensitive parameters and their optimal value

Parameter
Initial Optimal
S. No Parameter Description Unit Range value value
1 CN2 Curve Number, Condition AMCII - 35-98 83 54
2 SOL_AWC Soil available water capacity mm/mm 0-1 0.1 0.63
3 ESCO Soil evaporation compensation factor - 0-1 0.95 0.02
4 SOL_K Hydraulic conductivity of the saturated soil mm/hr 0-2000 45 1.56
5 GW_REVAP Ground water revap coefficient - 0.02-0.2 0.02 0.18
6 GW_DELAY Ground water delay time Day 0-50 31 9.5
Threshold water level in shallow aquifer for
7 GWQMN base flow - 0-5000 1000 1290
8 EPCO Plant uptake compensation factor - 0.01-1 0.4 0.11
9 ALPHA_BF Base flow alpha factor Days 0-1 0.048 0.54
10 Surlag Surface run off lag coefficient - 0-10 0.5 0.3
11 CH_N2 Manning coefficient for main channel - (-)0.01-0.3 0.14 0.1
Threshold water depth in
12 RVPMN the shallow aquifer for “revap” - 0-500 450 45
Effective hydraulic conductivity of main
13 CH_K2 channel mm/hr (-)0.01-500 0.01 75

Proper model calibration is important in hydrologic modeling studies to reduce uncertainty in

model simulation. Model calibration is performed by carefully selecting values for model input

parameters (within their respective uncertainty ranges) by comparing model predictions (output)

for a given set of assumed conditions with observed data for the same conditions. The SWAT

model was calibrated and validated for Tekeze river streamflow using the measured data at

Embamadre gauge station. The available data divided into two-time periods, for calibration

(1994-2002) and for validation (2003-2008) processes. During the calibration process, the

model's input parameters guided by the sensitivity analysis adjusted to match the observed and

simulated streamflow’s. In general, SWAT accurately tracked the measured streamflow’s of

Embamadre for the time, even though some peak flow months were over predicted. In the

validation process, the model was run with input parameters set during the calibration process

without any change. After manual and automatic calibration monthly streamflow were compared

89
against the observed data. There are several ways in which SWAT can be calibrated and

validated.

In this study, the Nash-Sutcliffe coefficient (ENS), coefficient of determination (R2 ) and Percent

bias (PBIAS) were used to evaluate performance of SWAT model simulation in the calibration

and validation process (Moriasi et al., 2007).

Nash-Sutcliffe efficiency (ENS): ENS is a normalized statistic that determines the relative

magnitude of the residual variance (“noise”) compared to the measured data variance

(“information”) (Nash and Sutcliffe, 1970). ENS indicates how well the plot of observed versus

simulated data fits the 1:1 line. The optimal values to get best model performance is at ENS =1.

ENS is computed as shown in equation 4.8:

2
∑n Obs
i=1(Qi −QSim
i )
ENS = 1 − [ Obs −Qmean ) 2 ] (4.24)
∑n
i=1(Qi Obs

Where 𝑄𝑖𝑂𝑏𝑠 is the ith observed streamflow of day i, 𝑄𝑖𝑆𝑖𝑚 is the ith simulated streamflow of day
𝑚𝑒𝑎𝑛 𝑚𝑒𝑎𝑛
i, 𝑄𝑂𝑏𝑠 the mean ofobserved streamflow,𝑄𝑆𝑖𝑚 the mean of simulated streamflow

Coefficient of determination (R2): describe the degree of co-linearity between simulated and

measured data. The R2 describes the proportion of the variance in measured data explained by

the model. The ranges of R2 varies from 0 to 1, with higher values indicating less error variance,

and typically values greater than 0.5 are considered acceptable. Although R2 has been widely

used for model evaluation, this statisticis over sensitive to high extreme values (outliers), and

insensitive to additive and proportional differences between model predictions and measured

data.

Obs 2
2 [∑n
i=1(Qi −Qmean Sim
Obs )(Qi −Qmean
Sim )]
R = Obs 2 2 (4.25)
∑n
i=1(Qi −Qmean n Sim
Obs ) ∑i=1(Qi −Qmean
Sim )

90
Where 𝑄𝑖𝑂𝑏𝑠 is the ith observed streamflow of day i, 𝑄𝑖𝑆𝑖𝑚 is the ith simulated streamflow of day
𝑚𝑒𝑎𝑛 𝑚𝑒𝑎𝑛
i, 𝑄𝑂𝑏𝑠 the mean ofobserved streamflow,𝑄𝑆𝑖𝑚 the mean of simulated streamflow

Percent bias (PBIAS): PBIAS measures the average tendency of the simulated data to be larger

or smaller than their observed counterparts. The optimal value of PBIAS is 0.0, with low-

magnitude values indicating accurate model simulation. Positive values indicate model

underestimation bias, and negative values indicate model overestimation bias. PBIAS is

calculated by Equation 4.26:

∑n Obs
i=1(Qi −QSim
i )∗(100)
PBIAS = [ ∑n Obs ) ] (4.26)
i=1(Qi

Where PBIAS is the deviation of data being evaluated, expressed as a percentage

Table 4.2 : SWAT model performance evaluation criteria(Moriasi et al., 2007)

Model Evaluation ENS R2 PBIAS

Excellent 0.75 < ENS ≤ 1.00 0.70 < R2 ≤ 1.00 PBIAS ≤ 10
Good 0.65 < ENS ≤ 0.75 0.60 < R2 ≤ 0.70 10 < PBIAS ≤ 15
Satisfactory 0.50 < ENS ≤ 0.65 0.50 < R2 ≤ 0.60 15 < PBIAS ≤ 25
Unsatisfactory 0.00 < ENS ≤ 0.50 0.00 < R2 ≤ 0.50 PBIAS > 25

The final step is validation for the component of interest (streamflow). Model validation is the

process of demonstrating that a given site-specific model can make sufficiently accurate

simulations. Validation involves running a SWAT2012 model using parameters that were

determined during the calibration process, and comparing the predictions to observed data not

used in the calibration.

91
4.4 Results and Discussions

4.4.1 Evaluation of bias correction methods

The outputs of an ensemble of high resolution regional climate models (RCMs) from the

Coordinated Regional-climate Downscaling Experiment for Africa (CORDEX-Africa) under

RCP4.5and RCP8.5 climate scenarios have been bias adjusted. The ensemble of CORDEX-

Africa RCMs RCPs scenario simulations is divided into past (1951-2005) and future (2011-2100)

periods, from which a reference period (1976-2005) to coincide with the observations.

All the four bias correction methods correct the biases in the raw RCM simulations and improve

the raw RCM data to some extent, but there are considerable differences in the quality of adjusted

RCM precipitation and temperature. Linear scaling a good estimation of the mean but does not

well adjust the standard deviations, wet-day intensities and probability of wet days which over

or under estimates depending on the stations, while the local intensity scaling and distribution

mapping have a good estimation of all frequency based statistics. These results confirm the study

by Teutschbein and Seibert (2012); i.e., the linear scaling method does not adjust the standard

deviation and the percentiles while the distribution mapping method does.

The local intensity scaling method provides a good estimation in the mean, median, wet day

probability and wet-day intensity, however there is a slight over estimation of the mean and under

estimation in the standard deviation. The distribution mapping method of precipitation bias

correction provides a good estimation in the mean, median, standard devastation, wet-day

probability and wet-day intensity than Local intensity scaling methods in all the four selected

stations of the basin. To analyze the performance of each precipitation bias correction methods,

several statistical measures compared (Table 4.3).

92
Table 4.3 Frequency based statistics of daily observed, raw RCM simulated (raw) and bias

corrected precipitation at the selected weather stations of Tekeze basin

Mean Median Standard Probability of Intensity of

Bias correction method
(mm) (mm) Deviation (mm) Wet-day (%) wet-day (mm)

Debre Tabor
Observed 4.14 0 8.64 40 10.32
Raw 3.4 0 7.05 32 9.28
Linear scaling 4.13 0 8.71 33 12.74
Local intensity scaling 3.45 0 7.26 33 10.69
Distribution mapping 4.14 0 8.57 37 10.46
Gondar
Observed 3.56 0 8.32 36 9.86
Raw 2.6 0 5.36 32 8.17
Linear scaling 3.56 0 7.35 33 11.12
Local intensity scaling 3.13 0 6.46 32 9.85
Distribution mapping 3.62 0 7.92 34 9.9
Lalibela
Observed 2.32 0 6.15 30 7.81
Raw 6.6 0 11.75 45 15.78
Linear scaling 5.78 0 10.3 42 13.93
Local intensity scaling 3.24 0 5.77 40 7.94
Distribution mapping 2.53 0 6.53 34 7.85
Mekele
Observed 1.58 0 5.23 21 7.38
Raw 1.43 0 3.72 32 5.11
Linear scaling 1.92 0 4.05 29 5.56
Local intensity scaling 1.54 0 4.98 27 6.73
Distribution mapping 1.6 0 5.17 23 6.89
These results in agreement with previous studies by different researchers stated that both

parametric and non-parametric distribution based bias correction methods give the best

performance in terms of reproducing the observed climate, whereas means based methods, in

particular, linear scaling is almost always ranked as the least skilled bias correction method

(Teutschbein and Seibert, 2012; Chen et al, 2013) next to delta change approach. In conclusion,

all the above mentioned bias correction methods are significantly and consistently improved the

93
biases of RCM climate models outputs of precipitation and temperature. Among these different

bias correction methods that have been suggested, distribution mapping have been found to

provide particular good results. Therefore, in this study the distribution mapping (DM) bias

correction method has been identified as the most efficient approach to bias correct climate model

ensemble of CORDEX-Africa RCP4.5 and RCP8.5 climate scenarios of precipitation and

temperature, while reproducing its statistics at spatial and temporal resolutions suitable to run

hydrologic models.

[Link] Precipitation Bias Correction results using distribution mapping

Distribution mapping bias correction method using gamma distribution function was used on

ensemble results of CORDEX-Africa RCP4.5 and RCP8.5 climate scenarios simulated

precipitation with observed meteorological stations: Lalibela, Gonder, Debre Tabor, Mekele,

Nefas Mewucha, Korem, Maichew, Wukro, Hagere Selam, Hawzen and Axum found in Tekeze

basin. The cumulative distribution function graphs (Figure 4.7) showed the effects of statistical

correction (distribution mapping) methodology in the distribution of corrected RCP scenarios

simulations at selected stations. These graphs can be used as a check how bias correction method

simulates high precipitation events. This process removes the poor ability of CORDEX-Africa

RCP scenarios in simulating the high precipitation spectrum. In CORDEX-Africa RCP

scenarios, daily simulated Tekeze basin precipitation has many dry days in combination with too

many drizzles which are the main sources of the bias in the mean and inability to reproduce

extreme events. This method efficiently improves simulated precipitation data, which improves

streamflow simulations which cannot be correct by calibration of SWAT hydrological model.

94
Figure 4.7 Comparison of climate scenario data before and after bias correction with respect to
observed data at (a) Lalibela, (b) Debretabor, (c) Gonder, and (d) Mekele selected meteorological
stations of Tekeze basin
The whole procedure applied separately for each month to correct for errors in the seasonal cycle.

By correcting the full distribution, corrections on days with little rainfall will be different than for

the days with extreme rainfall, thus accounting for the model’s poorer ability to simulate extremes.

This correction procedure corrects both the mean intensities and extremes. As a result, bias free

precipitation values from CORDEX Africa RCPs scenario obtained by this procedure from all-

weather observation stations for current and future climate condition.

95
Figure 4.8 Observed, raw RCP simulated and bias corrected values of precipitation at the four
selected stations of Tekeze basin

96
The monthly time series of observed precipitation for the base line period were in close agreement

with those from the bias corrected base line RCP scenario values (Figure 4.8). These results

indicated that the bias correction techniques significantly improved the quality of the CORDEX-

Africa RCP scenarios data. Overall, the results showed that bias corrected data from distribution

mapping bias correction methods could be reliable used for analyzing water availability in the

future. The stationarity assumption, i.e., that the same correction algorithm applies to both current

and future climate conditions is considered as the main drawbacks of distribution mapping method.

In this study, climate scenarios of both RCP4.5 and RCP8.5 evaluated under three future periods

of 2020s, 2050s and 2080s with reference to present day baseline period 1980s (1976-2005).

4.4.2 Change of Projected Precipitation and Temperature

[Link] Projected Precipitation

The result of the study after bias corrections of raw precipitation sample showed that the mean

annual precipitation in the future increases over Tekeze basin (Figure 4.10). The change in mean

annual precipitations varies from +32% to +40% for mid rang RCP4.5 and high-level RCP8.5

climate scenarios over all future time periods. In all future time periods, Kiremt (wet) season

(June–September) precipitation showed an increasing trend and varied from +21% to +48%. While

in Belg (minor rainy) season (February-May) precipitation amount decreased and changed from -

15% to -52% under both RCP4.5 and RCP8.5 climate scenarios. The Bega (dry) season (October-

January) showed a mixed trend where precipitation decreased up to -23% in the 2050s under

RCP4.5 and -39.5% in 2020s for RCP8.5 climate scenarios. On the other hand, precipitation

increases from +5.7% to +50% under both RCP4.5 and RCP8.5 climate scenarios. In the Tekeze

basin, the projected seasonal precipitation has not shown any systematic decreasing or increasing

trend (Figure 4.9) contrasting to a temperature which showed a rising trend under RCP4.5 and

RCP8.5 climate scenarios of future time periods.

97
Figure 4.9 Mean annual and seasonal precipitation variation over Tekeze basin under (a) RCP4.5,
and (b) RCP8.5 climate scenarios

Figure 4.10 Long terms mean annual precipitation trend at Tekeze basin under RCP4.5 and RCP8.5
climate scenarios

98
Figure 4.11 shows future percentage changes of monthly precipitation amounts for different
projected periods under RCP4.5 and RCP8.5 climate change scenarios. For RCP8.5 scenarios, the
months of March, April and May would exhibit a decrease in precipitation amount compared to
the baseline period whereas RCP4.5 scenario presented an increasing trend. The months of October
through February would show an increase in precipitation compared to reference period for both
scenarios and projected periods considered.

Figure 4.11 Change of monthly precipitation amount for the selected scenarios and
projected periods.

[Link] Projected Temperature

Tekeze basin exhibits an increase in projected minimum and maximum temperature under RCP4.5

and RCP8.5 climate scenarios (Figure 4.12). Mean annual minimum temperature varies from

+0.61OC to +1.87OC under RCP4.5 and from +0.74OC +4.12OC under RCP8.5 climate scenarios

for the three-time periods. Similarly, mean maximum annual temperature showed increasing trend

and varied from +0.80OC to +1.00OC under RCP4.5 and from +1.40OC +2.60OC under RCP8.5

climate scenarios of future time periods. It also showed that the increase in daily minimum

temperature higher than maximum temperature over the next 90years under both climate scenarios

but maximum temperature decreased up to -0.20OC under RCP8.5 in the 2020s. Generally, the

99
mean annual temperature increases up to 1.07OC under RCP4.5 and 2.21OC under RCP8.5 climate

scenarios for the next 90 years.

The projected seasonal maximum and minimum temperature in the Tekeze basin (Figure 4.12)

indicated that a consistent increase in Kiremt (June-September), Belg (February-May) and Bega

(October-January) in both scenarios for all time periods; except decrease in maximum temperature

in Bega season in 2020s. Moreover, under CORDEX-Africa climate scenarios Tekeze basin

temperature projection RCP8.5 predicted higher temperature increase than that of RCP4.5.

Projected mean annual minimum temperature may increase up to 1.1 °C and 3.38 °C under RCP4.5

and RCP8.5 scenarios, respectively in all future time periods. Similarly, mean annual precipitation

may increase up to 45% under both scenarios for all future time periods. Figure 4.13 shows future

change rates of monthly mean temperature in both scenarios for all future time periods. Mean

monthly temperature will increase under both scenarios in all time periods except the months of

January and February which showed a slightly decreasing trend in 2020s.

100
Figure 4.12 Mean annual and seasonal temperature projection in Tekeze basin under (a)

minimum temperature change at RCP4.5 (b) minimum temperature change at RCP8.5 (c)

maximum temperature change at RCP4.5 (d) maximum temperature change at RCP8.5 climate

scenarios

Figure 4.13 Rates of change of monthly mean temperature for the selected scenarios and
projected periods

101
In general, bias corrected projected annual precipitation and temperature showed an increasing

trend in 2020s, 2050s and 2080s over Tekeze basin under RCP4.5 and RCP8.5 climate scenarios.

This study projected annual precipitation and temprature results confirmed the same trend with

the study of Gizaw et al. (2017) conducted on Tekeze basin under both RCP4.5 and RCP8.5

climate scenarios using multiple CMIP5 GCMs not CORDEX data which showed increasing trend

in 2050s and 2080s.

4.4.3 SWAT Model Performance

In this study, SWAT hydrological model was calibrated and validated the streamflow of Tekeze

River basin at Embamadre gauging station for a calibration period of 1994-2002, with the

remaining 2003-2008 for validation. Figure 4.14 shows similar distribution of the observed and

simulated streamflow hydrograph for both validation and calibration periods. The model

performed well against the monthly streamflow. The calibration and validation results showed that

SWAT model could simulate the monthly streamflow well with coefficient of determination (R2 )

and Nash-Sutcliffe efficiency (ENS) greater than 0.70. Also, Percent bias (PBIAS) is positive with

reasonable underestimates with a value less than 10% (Table 4.4). SWAT well simulated the

hydrology of Tekeze basin and forced to generate future streamflow under both bias corrected

RCP climate scenarios.

Table 4.4 SWAT hydrological model monthly performance under validation and calibration
periods in Tekeze Basin

Parameter Calibration Validation

Nash-Sutcliffe Efficiency (ENS) 0.71 0.79
Coefficient of determination (R2) 0.73 0.80
Percent bias (PBIAS) 0.53 0.45

102
In this research, SWAT model results of streamflow of Tekeze basin was considered for

comparative analysis of observed streamflow with projection periods 2020s (2011-2040), 2050s

(2041-2070) and 2080s (2071-2100) under both climate scenarios.

Figure4.14 Observed and simulated monthly streamflow hydrographs during (a)

Calibration and (b) validation periods.

4.4.4 Streamflow Projection

Impact of climate change on the streamflow at Embamadre station downstream of the Tekeze

hydropower reservoir was analyzed. Results in Figure 4.7 show that SWAT successfully simulated

annual and monthly streamflow with a reasonable accuracy. Hence, the calibrated and validated

103
SWAT model forced to run for historical and future climate scenarios to generate future

streamflow for both RCP4.5 and RCP8.5 climate scenarios. The effect of climate change on annual

and monthly streamflow was also investigated as a percentage change with respect to the baseline

period (1994–2008) under the two scenarios in three time periods 2020s (2011–2040), 2050s

(2041–2070) and 2080s (2071–2100).

For all future time periods, both climate scenarios produced a moderate increase in mean annual

streamflow which is due to projected increase in precipitation. However, RCP8.5 climate scenario

in the 2080s showed a decrease in streamflow up to 23% due to higher increase in projected

temperature that leads to rising in evapotranspiration than increased in precipitation. Figure 4.15

showed the percentage change of annual streamflow for both climate scenarios and the three-time

periods. Under RCP4.5, the mean annual percentage change of streamflow will increase by 49%,

39% and 47% in the 2020s, 2050s and 2080s, respectively. Similarly, for RCP8.5, the mean annual

percentage change of streamflow increases to 22%, 19% and 2% in the 2020s, 2050s and 2080s,

respectively.

Seasonal projection streamflow showed a mixed increasing and decreasing trend (Table 4.5). In

Kiremt (main rainy) season (June-September) and Bega (dry) season (October-January)

streamflow will be increased under both RCP climate scenarios of future periods except Kiremt

RCP8.5 in the 2080s which will decrease. In the Belg season (February-May), streamflow will be

increased in RCP4.5 and will be reduced under RCP8.5 in all projected time periods.

104
Table 4.5 Mean annual and seasonal streamflow change (%) in Tekeze basin at Embamadre from
the baseline period (1994-2008)

Period Belg Kiremt Bega Annual

Streamflow change (%) RCP4.5
2020s 55.06 31.70 61.51 49.43
2050s 25.03 26.85 65.68 39.19
2080s 39.88 33.70 65.96 46.51
Streamflow change (%) RCP8.5
2020s -13.82 21.33 59.07 22.19
2050s -47.36 21.14 58.81 10.68
2080s -87.72 -39.12 58.49 -22.78

The monthly streamflow change shows a mix of positive and negative trends. Mean monthly

streamflow showed an increasing trend in RCP4.5 and significant mixed trend under RCP8.5

climate scenarios (Figure 4.15). Under RCP4.5 climate scenario, change in average monthly

streamflow ranges from 12 to 69% in 2020s and 13 to 67% in 2080s but in 2050s streamflow

change shows mixed trend that decreases in March to May by up to 9% and increases on other

months up to 39%. Mean monthly percentage change of streamflow under RCP8.5 climate

scenario showed mixed trends in all time periods. Under RCP8.5, the mean monthly streamflow

changes from −37 to 64%, −29 to 68% and −49 to 64% in 2020s, 2050s and 2080s, respectively.

Individual month’s trend showed that there was an increasing trend from August to February and

a decreasing trend from March to July. Therefore, climate change will have a clear impact on the

future streamflow an input of reservoir power production in Tekeze basin.

105
 Figure 4.15 Mean monthly and annual streamflow changes under RCP4.5 and RCP8.5

climate scenarios.
The changes and variability of monthly (inter-annual) streamflow will be much greater than the

annual streamflow changes in both scenarios in all time periods. This result showed that it is

important for the hydropower reservoir planners and managers to consider, the monthly

streamflow variability and changes for future planning and operation of reservoirs.

These mean annual projected results were in good agreement with the work of Gizaw et al. (2017)

which showed that an increase in streamflow at Embamadre station in 2050s in both RCP4.5 and

RCP8.5 but decreased in RCP4.5 2080s in the Tekeze basin and also the result were agreed with

the works done in the nieghboring basin using previous SRES scenarios by Adem et al. (2014)

simulated a consistent increase in future sterean flow at Gilgel Abay and Setegn et al. (2011)

projected a declining in streamflow of Tana basin in 2080-2100.

106
4.4.5 Climate Change Impact on Water Resources Planning

Climate change calls into question the reliability of water resources and the quality of the available

water, and challenges traditional water utility planning technique where water resource planning are

based on recorded hydrology and weather information assumed to be stationary. This stationarity

implies that hydrology and weather statistics and variability will not significantly deviate beyond the

observed past conditions and are good representations of future conditions. However, Water resources

planners, managers and policy makers are facing considerable uncertainties on future demand and

variability of water. Climate change and its potential hydrological effects are increasingly contributing

to this uncertainty. Streamflow is a critical hydrological parameter used for water resources planning

and management in Tekeze basin. However, streamflow was affected by climate change. In this study,

climate change scenarios of precipitation and temperature projected to transform the behavior of

Tekeze basin affecting timing and volume of streamflow used for hydropower and irrigation

production. The projected climate scenario after bias correction showed that an increase in temperature

and change in annual and seasonal precipitation (Figure 4.10 and Figure 4.11) that affects water

resources system of Tekeze basin. Projected streamflow showed a slight rise in streamflow in the basin

(Figure 4.15) which will be used as an input for hydropower and irrigation projects reservoir planning

and operation. Therefore, climate change will affect the planned and operational hydropower and

irrigation reservoir projects in the Tekeze basin.

Annual, seasonal and monthly Tekeze reservoirs inflow changed due to climate change may alter

frequency and duration of droughts that affect hydropower production of Tekeze reservoir. Tekeze

reservoir watershed is projected to show an increase in precipitation and streamflow that may increase

hydropower generation. Other implications based on this simulation includes,

107
 The rise of projected temperature in both RCP climate scenarios will increase evapotranspiration in

the reservoir surface and upstream watershed that will reduce hydropower production.

 Kiremt (wet) season streamflow timing will be altered or delayed due to climate change may impact

reservoir operations as well as release patterns of the downstream flow.

 The projected increase of extreme precipitation will lead to increased flooding events may be over

top the spillway, and other structures will be affected dam safety and operation rule curves designed

to prevent overtopping.

 Projected higher intensity precipitation will be eroded upstream farmland watersheds and increased

sediment load to the reservoir that increases reservoir dead storage level which reduces reservoir

capacity (lifespan).

As water utilities grapple with preparing for the large range of possible climate change impacts, many

are searching for new planning techniques to help them better prepare for a different, more uncertain,

future. The range of potential climate change impacts on rainfall and river flows should be included in

feasibility assessment of new hydropower and irrigation reservoir planning and should be carried out

within the context of basin-scale water resources management planning. Where a new hydropower or

irrigation reservoir projects are assessed to be the best choice, policy-makers and planners should place

more emphasis on investing in hydropower/irrigation schemes that maximize flexibility and that

embrace adaptive management. Therefore, water resources planners and managers should incorporate

climate change scenarios into the planning and design of Tekeze hydropower and irrigation projects in

particular and in all Ethiopian river basin projects in general.

108
4.5 Conclusive Remarks

This research evaluated the effects of climate change on Tekeze basin hydrology resulting from bias

corrected ensembles of CORDEX-Africa mid-range and high-level RCP climate scenarios (RCP4.5

and RCP8.5). Calibrated SWAT hydrological model was then used to transform these future climate

scenarios to projected streamflow used as an input for reservoirs planning and management. Key

findings of this chapter include:

1. Projected precipitation and temperature from ensemble CORDEX-Africa RCP scenarios have

systematic errors (biases) which may lead to biased simulated streamflow which is not corrected by

calibration of the hydrological model. However, distribution mapping method corrected the biases

and improved precipitation and streamflow simulations.

2. Mid-range RCP4.5 and high-level RCP8.5 climate scenario showed that projected temperature

consistently increases across Tekeze basin and precipitation projection may also increase annually

and in the Kiremt and Bega except for Belg season which shows a mixed trend.

3. The SWAT model could reproduce the current hydrological condition of Tekeze basin at monthly

time scale. The calibration statistical results were ENS= 0.71 and R2 = 0.73, also the validation result

was ENS= 0.79 and R2 = 0.80. These results are indicative of the SWAT model’s good performance

in this basin.

4. Projected higher temperature and precipitation increase under RCP4.5 and RCP8.5 climate

scenarios expected to increase projected streamflow of Tekeze basin. According to the latest inflow

projections total inflow to Tekeze reservoirs slightly increase in all time periods.

5. This study result showed that climate change would affect the future planning of Tekeze basin

hydropower reservoirs and irrigation projects. Therefore, the effects of projected precipitation and

streamflow should be included in the feasibility assessment of Tekeze water resources planning and

109
management. Where a new hydropower reservoir assessed to be the best choice in the watershed,

more emphasis needs to be placed on investing in hydropower schemes that maximize flexibility

with climate change impact. Water resources planners and managers should incorporate climate

change scenarios into the planning and design of Tekeze irrigation and hydropower

Dams/reservoirs.

110
5. TEKEZE HYDROPOWER RESERVOIR OPERATION UNDER CLIMATE

CHANGE4

Abstract: Optimal operation of reservoirs is very essential for water resource planning and

management, but it is very challenging and complicated when dealing with climate change impacts.

The objective of this chapter was to assess existing and future hydropower operation at the Tekeze

reservoir in the face of climate change. In this study, a calibrated and validated Soil and Water

Assessment Tool (SWAT) was used to model runoff inflow into the Tekeze hydropower reservoir

under present and future climate scenarios. Inflow to the reservoir was simulated using hydro-climatic

data from an ensemble of downscaled climate data based on the Coordinated Regional Climate

Downscaling Experiment over African domain (CORDEX-Africa) with Coupled Model

Intercomparison project Phase 5 (CMIP5) simulations under Representative Concentration Pathway

(RCP)4.5 and RCP8.5 climate scenarios. Simulated and projected inflows to Tekeze hydropower

reservoir were used as input to the US Army Corps of Engineer’s Reservoir Evaluation System

Perspective Reservoir Model (HEC-ResPRM), a reservoir operation model, to optimize hydropower

reservoir release, storage and pool level. Results indicated that climate change has a clear impact on

reservoir inflow and showed increase in annual and monthly inflow into the reservoir except in dry

months from May to June under RCP4.5 and RCP8.5 climate scenarios. HEC-ResPRM optimal

operation results showed an increase in Tekeze reservoir power storage potential up to 25% and 30%

under RCP4.5 and RCP8.5 climate scenarios, respectively. This implies that Tekeze hydropower

production will be affected by climate change. This analysis can be used by water resources planners

and mangers to develop reservoir operation techniques considering climate change impact to increase

power production.

111
 Keywords: Reservoir Operation; Optimization; SWAT; HEC-ResPRM; Climate Change; CORDEX-

Africa; Tekeze Hydropower Reservoir

5.1 Introduction

Water resources reservoirs are important tools for integrated water resources development and

management (Yazdi and Moridi, 2017; Zhou et al., 2015), but nowadays their operation and

management is challenging due to various factors (Birhanu et al., 2014; Rani and Moreira, 2010). The

reservoir operates to supply water for municipal consumption, hydropower production, irrigation and

industrial needs, flood control, recreation, navigation or ecological requirements. Currently, due to

water crisis the global freshwater supply to meet the needs of the different sectors is falling short (David

et al., 1999; Jury and Henry, 2007; Rijsberman, 2006). Factors that contribute to this include population

growth, urbanization, climate change, land use change, land degradation and poor water resources

management (Ghashghaie et al., 2014; Larson et al., 2013). Hence, to alleviate these problems and meet

the freshwater and energy demand of communities, it will necessitate optimal operation of water

resources reservoirs (Guo et al., 2011; Zhou et al., 2014).

Various researchers studied the reservoir operation throughout the world to get optimum level of release

and optimal volumes of storage considering inflows and needs (Azizipour et al., 2016; Cheng et al.,

2008; He et al., 2014; Lu et al., 2013). Most research conducted in the reservoir operations have specific

objectives like hydropower (Azizipour et al., 2016; Cheng et al., 2008; Lu et al., 2013), flood control

(He et al., 2014), irrigation (Birhanu et al., 2014) and environmental (Yin and Yang, 2011).

Water resources infrastructures have been designed and managed historically but these designs gave

little attention to the effect of climate change and non-stationarity in hydrologic variables (Milly et al.,

2008). Evidence suggests that these hydrologic variables used for water resources planning and

management previously assumed stationarity in time have changed by anthropogenic activities (Stocker

112
et al., 2013). The increase in temperature, changes in precipitation and evapotranspiration rates due to

climate change alters global hydrologic cycle (Huntington, 2010). The streamflow affected by intensity

and frequency of precipitation leads to increase the intensity of floods and droughts. These changes

affect water resources at local and regional levels (Vicuña et al., 2011). The hydrological processes and

water availability affected by a change in the patterns of precipitation and temperature impacts

agriculture, industry, communities, hydropower and aquatic life (Alazzy et al., 2014). Climate change

impact on fresh water resources may change the mean annual streamflow, shift seasonal flows,

increases floods and droughts and changes in sediment fluxes which affect reservoir operation (Lee et

al., 2016; Lumbroso et al., 2015; C. Zhang et al., 2014).

Many researchers in different parts of the world have studied the impacts of climate variability and

change on shifts in hydrological regimes and water resources (e.g., (Ehsani et al., 2017; Haile et al.,

2017; Setegn et al., 2014; Zhu et al., 2018)). These studies assessed the current and future water

resources availability and rainfall variability across the globe to support appropriate water resources

planning and management. Different studies showed that Africa is highly vulnerable to climate change

(Serdeczny et al., 2017). Climate change studies showed temperature increased and precipitation

pattern changed throughout arid and semi-arid regions of Africa (Conway et al., 2009; Hales, 2007;

Nikulin et al., 2012) and affected the hydrological processes that impacts reservoir operation. Most

studies showed the impacts of climate change on African hydropower reservoirs (Beyene et al., 2010;

Hamududu and Killingtveit, 2016; Kim and Kaluarachchi, 2009; Lumbroso et al., 2015; Yamba et al.,

2011). Kim and Kaluarachchi (2009) and Beyene et al. (2010) projected that precipitation and

temperature will be increased in the Nile River basin and have positive effect on hydropower

production, but Yamba et al. (2011) and Hamududu and Killingtveit (2016) investigated that in the next

60 years hydropower production show a gradual reduction with large variability in the Zambezi River

113
basin. In the Nile River basin, the rapidly growing hydropower based energy need, population growth,

food insecurity and finite water resources will lead to competitions for water in the riparian countries

and this will be aggravated by the climate change. Several studies have been conducted on the

variability of precipitation and streamflow in the Nile River basin (Abtew et al., 2009; Melesse et al.,

2009; Mengistu et al., 2014; Setegn et al., 2011; Tarekegn and Tadege, 2006) that affects reservoir

planning and management (Birhanu et al., 2014).

Most hydropower reservoir operators concern is existing hydrological variability without foreseeing

climate change as a particular serious threat (Lumbroso et al., 2015). Hence reservoir operation need

to incorporate plans to address hydrologic non-stationarity and uncertainty caused by climate change

(Ehsani et al., 2017; Jamali Saeed et al., 2013; Vicuña et al., 2011; Vonk et al., 2014). Due to this,

ensembles of Global Circulation Models (GCMs), scenarios and regional climate models (RCMs) used

as input to hydrological model to generate future streamflow (Raje and Mujumdar, 2010; Wilby and

Harris, 2006) that can be used as an input for reservoir operations.

Sedimentation may cause serious impacts on reservoir operation and functionality by reducing reservoir

storage capacity and shortening reservoir useful life for human benefits. Studies showed that Northern

part of the Tekeze basin watersheds are vulnerable to sedimentation and/or soil erosion problems for

the sustainable use of small reservoirs developed for irrigated agriculture and Tekeze reservoir

(Haregeweyn et al., 2008; Wolde, 2016). This reservoir sedimentation problem may lead serious

reduction in reservoir storage capacity, causing future hydropower generation problems. However, rate

of sedimentation of Tekeze reservoir remains unpredicted. More and wide knowledge is still needed to

better understand and solve the sediment problem, and hence may improve future reservoir operation.

But the focus of this research is to study potential climate change impact on hydropower reservoir

operation and management by not varying sedimentation level.

114
Nowadays, reservoir operation techniques become increasingly important and researchers still

searching the best technique. Many authors proposed and reviewed various reservoir operation models

and methods (Labadie John W., 2004; Oliveira and Loucks, 1997; Rani and Moreira, 2010; Wurbs

Ralph A., 1993; Yeh, 1985). Labadie (2004) extensively reviewed and evaluated various optimization

methods and reported that no universally approved algorithm for all reservoir operations. Rani and

Moreira (2010) investigated that optimization models usually require simulation models for verifying

and testing planned operating policies. Dam managers use simulation models more relaxed than

optimization models as simulation models are easier to interpret, apply and present to non-professionals

(Labadie John W., 2004; Oliveira and Loucks, 1997). But Optimization models give reliable results. In

recent years, to overcome these problems, a combination of simulation and optimization models applied

in reservoir operation. In this research, US Army Corps of Engineer’s Reservoir Evaluation System

Perspective Reservoir Models (HEC-ResPRM), a combination of simulation and optimization model

is used.

In this study, Tekeze hydropower reservoir was chosen due to: (1) Tekeze basin shows high rainfall

variability (Abtew et al., 2009; Ayalew, 1999) which affects reservoir inflow; (2) Tekeze hydropower

reservoir not designed by considering hydrological non-stationarity and climate change; (3) the

reservoir has not been optimally operated and sometimes not fully functional during dry periods.

Therefore, the objective of this chapter are to (1) assess impact of climate change on reservoir inflow

using Soil and Water Assessment Tool (SWAT) and recent Coordinated Regional climate Downscaling

Experiment over African domain (CORDEX-Africa) RCMs under Representative Concentration

Pathway (RCP)4.5 and RCP8.5 climate scenarios, and (2) apply HEC-ResPRM optimization model to

get optimal release, reservoir level and storage for optimal power production including in the face of

climate change.

115
5.2 Material and Methods

5.2.1 Study Area

Tekeze hydropower reservoir is located in the tributary of Tekeze-Setit-Atbera river part of Eastern

Nile upstream of Embamadre gauged stations approximately 80km west of the town of Mekele (Figure

5.1). The surface area of the Tekeze reservoir watershed is 29,404 km 2. This watershed has high

mountainous areas in its sources in the Northern Ethiopian highlands up to 4517 masl and low land

areas as low as 800 masl with varying climate depending on altitude change. The rainfall increases with

altitude from 600 mm to 1200 mm but it is a reverse for temperature which decreases from 26 °C to 10

°C. This watershed has a mean annual inflow of 4.4 Billion cubic meters at Embamadre gauging station

and annual potential evapotranspiration of 1778 mm. This watershed also has a large elevation drop

from its sources to low land areas near Ethio-Sudan border and offers significant hydropower potentials

of 5960 GWh/year as discussed in chapter 2.

Tekeze single purpose hydropower reservoir located at 13°21′ N and 38°45′ E (Figure 2.2 and Figure

5.1) is the second tallest double concrete arch dam in Africa next to Katse arch dam in Lesotho. The

purpose of this reservoir is for hydropower production with total installed capacity of 300 MW in four

75 MW Francis turbines at underground power house. The reservoir has a total storage capacity of 9293

million cubic meters (MCM) of which 5293 MCM live storage at 1140 m above sea level (masl) and

4000 MCM below dead storage level (1096 masl). The reservoir also has 147 km2 surface areas at full

supply level.

116
Figure 5.1 Location of Tekeze hydropower reservoir and weather stations

5.2.2 Datasets used

[Link] Historical and Future Hydrology

In this research, the four hydrological data periods analyzed at the outlet of Tekeze hydropower

reservoir using the calibrated and validated SWAT model as discussed in chapter 4. These were the

reservoir inflow data of: (1) observed and RCP scenarios historical records (1994–2008); (2) the near

future period (2011–2040), middle future period (2041–2070) and the far future periods (2071–2100).

SWAT simulates historical (past) and all future reservoir inflows using precipitation and temperature

projections for eleven stations shown in Figure 5.1 from an ensemble outputs of CORDEX-Africa

117
RCMs downscaled from different GCMs from Coupled Model Intercomparison project Phase 5

(CMIP5) simulations available in 0.44° resolution for Ethiopian domain under two recent

representative concentration pathways (RCP4.5 and RCP8.5) climate scenarios.

[Link] Reservoir data

HEC-ResPRM, a reservoir operation model, requires data like back ground map of the watershed,

reservoir outlet capacities, elevation-area-storage curve, historical reservoir storage and water surface

level, power production and flow time series to perform optimal operations. Water surface elevation

values include the minimum operation level, the maximum operating level and historical maximum

and minimum elevations needed for power production in each months of the year. These physical data

were used to develop model constraints and allow the model to calculate penalties. All these data except

power production were collected from Ethiopian Electric Power Corporation and Ministry of Water,

Irrigation and Electricity. Tekeze hydropower reservoir characteristics, historical observed reservoir

storage and pool level data are shown in appendix B.

5.3 Methods
5.3.1 HEC-ResPRM Optimization Model

Several computer models have been developed to design reservoir storage capacity and establishing

operational policies during preconstruction planning of new projects, to reassess the existing operation

policies of reservoir systems and to support release decisions during real time operation. Operational

models used for reservoir operation broadly classified as descriptive simulation, prescriptive

optimization and hybrid models. Descriptive models simulates decisions of reservoir releases in pre-

defined logical rules while the descriptive nature of simulation models allow for "what if' studies, their

prescriptive capabilities are limited, prescriptive optimization models uses mathematical programming

techniques to solve decision variables and the hybrid models are mainly describe simulation models

118
with piecewise optimization (McMahon and Farmer, 2009). Optimization models do not require

explicit statements of operating rules, since operations are suggested (or prescribed) by the model.

Instead, the objectives for reservoir operations must be explicitly stated in the form of penalty functions.

In addition to the specification of quantitative operating objectives (penalty functions), optimization

models of reservoir systems also require mathematical constraints to represent physical, engineering,

or legal constraints to the system and a representation of hydrologic inputs to the system. Physical and

engineering constraints on the system would include reservoir capacities and turbine or outlet

capacities. In some cases, minimum in stream flows or other operational or legal constraints might be

added, although it is usually preferable to represent such "soft" constraints with steep penalty functions.

An optimization model, using often complex numerical solution algorithms, then prescribes desirable

operating decisions which yield the minimum total penalty and satisfy all constraints defined for the

system.

In this study, the US Army Corps of Engineer’s Reservoir Evaluation System Perspective Reservoir

Model (HEC-ResPRM) (Connell and Harou, 2011; USACE, 1992), a hybrid reservoir system

operations optimizations software package developed to assist planners, operators and managers with

reservoir operation plan and decision making, was used. HEC-ResPRM is a reservoir system operations

optimization model developed to assist reservoir operators with decision making by demonstrating the

optimal possibilities for reservoir management in the system. As an optimization model, HEC-ResPRM

offers an idea of the best outcome that can be expected from the system or any operating strategy. The

HEC-ResPRM is designed to further the efficiency and use of HEC-PRM (Perspective Reservoir

Model) for a data intensive single or multi-objective reservoir system operation studies. HEC-PRM, a

computational core of user friendly HEC-ResPRM software, is a generalized computer program that

performs deterministic network-flow optimization of reservoir system operations. HEC-PRM

119
prescribes optimal value of flow and storage over time by minimizing penalty functions located

throughout the water resources network. It also addresses a reservoir system operation problem of

optimal long-term allocation of available water. HEC-ResPRM is a combination of simulation and

optimization model when Perspective Reservoir Model (PRM) is integrated in to HEC-RES modeling

platform. HEC-ResPRM can be used in conjunction with ResSim or alone to improve and analyze

reservoir operation. HEC-ResPRM is an implementation of HEC-PRM shared with HEC-ResSim a

sister reservoir system simulation tool in a graphical user interface for creating, running, sorting and

analyzing optimization runs. The integration of PRM in to the HEC-Res modeling platform was made

to facilitate the joint development and use of simulation and optimization models. The HEC-Res

modeling system allows different network configurations and model runs to be managed and visualized

within a single interface, thus forming a robust platform for complex data-intensive modeling studies.

The Res implementation also allows users to produce graphical results directly from the Graphical User

Interface (GUI). HEC-ResPRM uses HEC’s data storage system (HEC-DSS) to store and retrieve of

input and output time series data.

HEC-ResPRM is a monthly network flow programming model and gives optimal values of release and
storage by minimizing penalty functions (Faber Beth A. and Harou Julien J., 2006; Ostadrahimi et al.,
2012). Network flow programming is computationally efficient form of linear programming. A network
solver finds optimal flow for the entire network simultaneously based on the unit cost associated with
flow along each arc. In HEC-ResPRM, networks are constructed to represent a physical system, where
nodes are junctions and reservoirs, and arcs are river reaches. To build a single model network that
represents the whole network flow problem the network must encompass both space and time. The
larger network is made up of the duplicate networks, connected by storage arcs, which represent Tekeze
hydropower reservoir storage (flow in time) from one-time period to the next. Each arc is a possible
path of flow, and a unit cost is associated with each arc. The slope of the penalty function is the unit
cost. In this model the goal of optimization solver is to find minimum cost path for each unit of flow in
the network over the time window. Finding the minimum cost path is accomplished through a network

120
flow linear programming technique called primal network simplex method. The solution to this
problem is the set of minimum cost (optimal) releases and flows for every time step. These results
provide insights about the system’s operations and objectives under varying conditions and can be used
to develop or improve operating rules.
Penalty functions associate a penalty or reward (negative penalty) with Tekeze hydropower reservoir
levels of flow or storage (flow in time). It is the penalty functions that derive the solution to the
optimization problem, which is why it is so essential to provide HEC-ResPRM with meaningful penalty
functions. Therefore, a penalty function a representation of unit cost in terms of relationship between
penalty and flow or storage. Hydropower generation is a nonlinear function of both net head and release
(flow). In order to reasonably represent this relationship, HEC-ResPRM allows users to define multiple
penalty functions, each based on a different storage level. It then uses an iterative process to select the
release and corresponding storage levels of Tekeze Hydropower Reservoir for minimum penalty. This
approach is, of course, an approximation of the actual complex conditions of power generation. All
reservoir outlets of Tekeze hydropower reservoir in HEC-ResPRM are represented as one outlet
because it cannot divide reservoir releases in to those that goes in through various gates. Therefore, at
any time with added hydropower penalties, water being released from the reservoir which is assumed
to be generating power. Multiple penalty curves can be used, each based on performance at a different
reservoir storage. For Tekeze hydropower reservoir operation it is difficult to add hydropower penalties
to generate optimum power due to lack of observed power production data and turbine efficiency curves
as only one turbine is functional from a total of four turbines. Because in this reservoir all turbines are
not functional and only one turbine was operational in the historical periods. Due to this reasons for
this study, Tekeze hydropower reservoir operation under climate change were discussed based on the
changes in reservoir storage, water level and release not in the power production.
Optimization problem represented by the network with cost associated with flow as follows:

𝑀𝑖𝑛𝑖𝑚𝑖𝑧𝑒: ∑𝑛𝑡 𝐶𝑡 𝑄𝑡 (For all nodes) (5.2)

𝑆𝑢𝑏𝑗𝑒𝑐𝑡 𝑡𝑜: ∑ 𝑄𝑡 − ∑ 𝑎𝑡 𝑄𝑡 = 0 (For all nodes) (5.3)

𝐿𝑡 ≤ 𝑄𝑡 ≤ 𝑈𝑡 (For all arcs) (5.4)

121
in which n is total number of network arcs; 𝐶𝑡 is unit cost, weighting factor for flow along arc t; 𝑄𝑡 is

flow along arc t; 𝑎𝑡 is multiplier (gain) for arc t; 𝐿𝑡 is lower bound on flow along arc t; and 𝑈𝑡 is upper

bound on flow along arc t. In this case, node represents a reservoir and river or channel junctions. Arcs

represent inflow and outflow links in the reservoir system. Each arc has a minimum and maximum flow

that it must carry in the reservoir system. The arcs (inflow and outflow links) may transfer water

between two points in space (transferring water in channels) or in time (changing pool elevations in the

reservoir). Also, flow is conserved in the reservoir (node). Equations (5.2) through (5.4) are special

forms of linear programming problems solved using primal simplex method. An off the shelf solver

(using a modified Simplex Algorithm) is used to determine the optimal allocation of water within the

system. The results of the solver are processed to report and display reservoir release, storage volume,

channel flow and other pertinent variables.

5.3.2 HEC-ResPRM Model Setup

HEC-ResPRM uses a map-based schematic to provide a meaningful representation of the optimization

network as it relates to the physical river/reservoir system. The layout of the Tekeze river system is
represented with a "stream alignment" which can be drawn by the user or imported from geo-referenced
maps. It can also be imported from HEC-ResSim. HEC-ResPRM modeling system allows different
network configurations and model runs to be managed and visualized within a single interface, thus
forming a robust platform for complex, data intensive modeling studies. HEC’s Data Storage System
(HEC-DSS) is used for efficient storage and retrieval of input and output time series data.
In HEC-ResPRM, a watershed is a collection of data associated with a model of a particular water
resources system, Tekeze hydropower reservoirs watershed in this case. Data for HEC-ResPRM
includes the physical layout and properties of the system, time series input and output, and information
associated with the optimization such as penalty functions, parameters and settings. Tekeze hydropower
watershed data are viewed and edited in HEC-ResPRM through three separate modules: Watershed
setup (Includes configuration of projects), Networks (including building run alternatives and
Optimizations (where the model alternatives are run and results viewed). Each module provides access
to specific types and directories of data within the watershed data tree as shown in Figure 5.2.

122
Each module requires the setup and entry of specific data. Changes to any module require updates to
other module, so it is important recognize how each module interacts with the other. After performing
an optimization one or more alternatives results can view and print in both tabular and graphical forms.

Figure 5.2 HEC-ResPRM schematization of Tekeze Hydropower reservoir operation

Watershed setup module used for laying out the physical extents of the model in which the basic shapes
and branches of the Tekeze river system setting up one or more sets of projects (configurations) wish
to model. The map region of the window contains a background map of Tekeze river, Tekeze
hydropower reservoir, a stream alignment and the reservoir element. Back ground maps are optional
but helpful when tracing the stream alignment. Project elements such as stream reaches, reservoir and
diversion are added on to the stream alignment. The stream alignment forms the back bone onto which
all the different elements will be added. All configurations of various project elements to be analyzed
for this watershed share a common stream alignment. Stream alignment can be created in HEC-
ResPRM or imported from various spatial data formats.
Network module used for imputing and editing element data, including penalty functions and placing
additional elements on to the stream alignment. Network of reservoirs, stream reaches and diversions
can be created. A network is created based on a selected configuration of projects on the stream

123
alignment and includes Tekeze river reaches that connects the project. An alternative in HEC-ResPRM
is a group of data including climate change scenario effect that includes a network, a selection of
composite penalties and settings, and a variety of compute settings. In HEC-ResPRM penalty functions
are grouped in to penalty sets and composite penalties. Each penalty set is intended to represent one
particular interest and consists of 12 individual penalty functions one for each month. The penalty
functions may vary based on the season selected for each month. If a penalty applies consisting all year,
a single “all year” season can be applied to every month. Monthly seasons can be automatically
generated or created and applied to the months. A composite penalty also consists of 12 individual
monthly penalty functions. The composites are the sum of the individual monthly penalty sets and are
defined based on which penalty sets are selected. HEC-ResPRM is set up to accommodate the need to
model and compare these types of variations in alternatives. An alternative in HEC-ResPRM is a group
of data that includes a network, a selection of composite penalties and settings, and a variety of compute
settings. Optimization runs are performed on these alternatives.
Optimization module designed to organize alternative runs, run optimization modules based on the
alternatives settings and to visualize and analyze results. The optimization takes place through the HEC-
PRM engine and solver. The time span (in this case the historical and projected periods) over which to
perform the model run and the alternatives to be analyzed must specify for each optimization. Graphical
results can be produced for one or more alternatives include in the optimization.
Sensitivity analysis in this study was done by adjusting the demands, changing the shape or magnitude

of penalty curves and by changing initial and ending reservoir levels. After the HEC-ResPRM model

fine-tuned, tests can be run on the performance under various inflow conditions for the historical

periods from 1994-2008 and 2009-2011 after the reservoir constructed. A series of wet year streamflow

constructed from historical data were run to see how the optimal results differ from average conditions.

Calibration can be consisting of these repeated runs, which provide successive improvement of the

model. Finally, the HEC-ResPRM model were run using the historical 2012-2017 reservoir inflow data

and SWAT model simulated future reservoir inflow for the near (2020s), middle (2050s) and far

(2080s) future RCP4.5 and RCP8.5 climate scenario data to get optimized Tekeze hydropower reservoir

release, pool level and storage.

124
5.4 Results and Discussions

5.4.1 Impacts of climate change on reservoir inflow

The volumes of reservoir inflows over different time periods are unknown in real time operation and
their volumes can be estimated using inflow forecasting models. The application of inflow forecasting
models may be affected by large errors to forecast long period reservoir inflows. Due to this reason
Climate projection models combine with hydrological models were used to get the projected Tekeze
hydropower reservoir inflow volumes. Impact of climate change on the streamflow at Embamadre
station downstream of the reservoir was analyzed. Observed streamflow data from a period 1994–2002
was used for model calibration and from 2003–2008 was used for validation. As shown in chapter 4
SWAT model successfully simulated annual and monthly streamflow with a reasonable accuracy.
Hence, the calibrated and validated SWAT model forced to run for historical and future climate
scenarios to generate future reservoir inflow for both RCP4.5 and RCP8.5 climate scenarios in all time
periods. The total mean annual baseline (past) and future Tekeze hydropower reservoir inflow (m3/s)
trends under RCP4.5 and RCP8.5 climate scenarios for all time periods are shown in Figure 5.3.
Projected monthly reservoir inflow for 2050s time period under both RCP4.5 and RCP8.5 climate
scenarios are shown in the appendix B.

Figure 5.3 Annual Tekeze hydropower reservoir inflow trend for future time periods
5.4.2 Current Reservoir Operation

HEC-ResPRM optimization model run under current baseline condition (2009–2017). This model

optimized the current Tekeze hydropower reservoir operation. The current optimized value (Table 5.1) of

125
HEC-ResPRM optimization model showed an increase in reservoir storage compared to current actual

hydropower reservoir operation status. It is also indicated that the mean annual reservoir pool level

increased up to 7.87 m (Table 5.2) that will store more water to produce power throughout the year. It

contradicts the current actual Tekeze hydropower reservoir operation which produces insufficient

power even very little or no power production during dry months. This implies that Tekeze hydropower

reservoir was not optimally operated till now. The reservoir storage dropped to the minimum operating

level and sometimes dries in the non-rainy months. Therefore, the current actual reservoir operation is

not effective and should consider different well tested reservoir operation techniques under a changing

climate. Due to lack of recorded power production data of Tekeze hydropower reservoir; this study

uses reservoir release, storage and pool level for current and future reservoir operation under climate

change for comparison.

5.4.3 Reservoir Operation under Climate Change

The future reservoir inflows generated by SWAT under RCP4.5 and RCP8.5 climate scenarios in three-

time periods 2020s (2011–2040), 2050s (2041–2070) and 2080s (2071–2100) with other reservoir data

were used to run optimization model to get projected optimal reservoir outflow (release), storage and

pool level results.

[Link] Projected Reservoir Inflow and Outflow

Climate change impacted inflow and outflow (release) hydrograph of Tekeze hydropower reservoir

considered in this study are shown in Figures 5.4a, b. According to the inflow projections based on

ensembles of CORDEX-Africa RCM climate model simulations, total inflows to Tekeze hydropower

reservoir expected to increase under RCP4.5 and RCP8.5 climate scenarios for all future time periods.

Figures 5.4 a, b show that, total monthly inflow under RCP4.5 is greater than the total monthly inflow

projected under RCP8.5 climate scenarios. However, under RCP4.5 climate scenario, the reservoir

126
inflow projections exhibit high fluctuations inter-annually as compared to RCP8.5 climate scenario and

observed historical values. The highest inflow volumes under RCP4.5 were concentrated in the rainy

months that spilled easily and affect the dry period reservoir storage level and or release.

There would be an increase in excess reservoir inflow during the rainy months of August through

October under both RCP4.5 and RCP8.5 climate scenarios in all time periods. This increased spillage

of available water inflow occurs because of the effect of climate change that increased the hydropower

reservoir inflow under RCP4.5 and RCP8.5 future climate scenarios. According to the latest climate

simulations, the overall inflow volume is predicted to be higher during rainy months and provided that

the reservoir lacks sufficient storage capacity to accommodate these high flows. As a result, Tekeze

hydropower reservoir forced to spill water without generating hydropower. This indicates that the

increased in overall reservoir inflow volume does not necessarily be advantageous to produce more

power. Therefore, decision need to be taken on the amount of water to be released and or stored now

and retained for future considering the variations in inflow and demands.

127
 Figure 5.4 Mean monthly reservoir inflow and optimized outflow (release) for future time

periods under: (a) RCP4.5 climate scenario; (b) RCP8.5 climate scenario.

In this study, the reservoir outflow (release) was obtained by HEC-ResPRM optimization model under

RCP4.5 and RCP8.5 climate scenarios for 2020s, 2050s and 2080s-time periods. In all future time

periods (Figures 5.4 a, b) under the two climate scenarios, the reservoir release will be increased to

produce more power due to an increased future reservoir inflow and optimum water storage using

optimization model. Under RCP4.5 climate scenario average monthly reservoir outflow varies from

353 to 2590 m3/s in 2020s, from 435 to 2757 m3/s in 2050s and from 442 to 3090 m3/s in 2080s.

Similarly, average monthly reservoir outflow varies from 538 to 1445 m3/s in 2020s, from 514 to 1412

m3/s in 2050s and from 577 to 1396 m3/s in 2080s under RCP8.5 scenarios. In both scenarios, the

minimum and maximum outflow value occurred during dry and wet periods, respectively. In all time

periods, the optimum reservoir outflows (releases) under RCP8.5 climate scenario for the dry months

of November through February were greater than the optimum releases under RCP4.5 climate scenario.

128
These changes show that under RCP8.5, the optimized reservoir stored more water in wet months for

dry period release and projected higher storage level compared to RCP4.5 climate scenario.

[Link] Optimum reservoir power storage under climate change

HEC-ResPRM optimized result showed an increase in projected mean annual Tekeze hydropower

reservoirs storage under RCP4.5 and RCP8.5 climate scenarios. This increase was projected for three

future time periods (Table 5.2) and the projected optimum stored water varies from 24 to 25% (RCP4.5)

and 28.5 to 30% (RCP8.5).

Table 5.1Mean annual optimized power storage under climate change scenarios.

Optimized Reservoir Change in Optimized

Periods
Storage (Mm3) Reservoir Storage (%)
RCP4.5 RCP8.5 RCP4.5 RCP8.5
Current optimized 6639 24.0
2020s 6688 6880 25.0 28.5
2050s 6669 6903 24.6 29.0
2080s 6665 6958 24.5 30.0

HEC-ResPRM model result under both scenarios in current and all future time periods showed

minimum and maximum reservoir storage periods (Figure 5.5 a, b, c). Tekeze reservoir reached at

maximum storage (reservoir filled) in September and stayed somewhat constant optimum storage up

to November. During August to September, main rainy months, the reservoir is filled, and optimization

model keeps the maximum storage up to November. The reservoir storage tends to slightly be decreased

starting from end of November until the beginning of February. After February, the reservoir storage

decreased down to the optimization model capacity to store energy at a minimum flow and reached a

minimum storage level in June to prepare and capture inflows in the wet main rainy months. In all

129
future months, there will be a stored water to produce power which is always greater than the current

optimized value.

Figure 5.5 Monthly optimized reservoir storage variations under RCP4.5 and RCP8.5 climate

scenarios for: (a) 2020s; (b) 2050s; (c) 2080s.

130
Figure 5.5 shows that more power is stored in both scenarios for the next 90 years as compared to

current actual and current optimized value. The change in maximum optimum storage increased will

be occurred in January and varies from 1693.4 to 1800.8 Mm3 under RCP4.5 scenarios and from 1731.9

to 1851.1 Mm3 under RCP8.5 scenarios in all time periods. The minimum optimal storage change

increased will occur in July and varies from 392.2 to 424.7 Mm3 under RCP4.5 scenarios and 803.6 to

956.6 Mm3 under RCP8.5 scenarios in all time periods. This is due to climate change impact on the

reservoir inflow and the capacity of the optimization model to operate the reservoir optimally. HEC-

ResPRM optimization of future projections tends to make much greater seasonal use of reservoir

storage than the current actual operations.

The monthly optimum stored water increases in all months for future time periods under both RCP

scenarios as compared to the baseline period (base line varies from 4400 to 6500 Mm3). Optimized

monthly reservoir storage variations are shown in Figure 5.4. The mean monthly optimum reservoir

storage in the future time periods varies for RCP4.5 from 5100 to 8300 Mm3 in 2020s, 4700 to 8050

Mm3 in 2050s and 5000 to 8100 Mm3 in 2080s. It also varies for RCP8.5 from 4900 to 8100 Mm3 in

2020s, 4850 to 8020 Mm3 in 2080s and 750 to 7900 Mm3 in 2080s.

This storage versus monthly graphs shown in Figure 5.5 a, b, c can be used as a rule curve for a given

period and climate scenario. These Operational guidelines (rule curves) can be inferred using the results

of runs for different climate scenarios and time periods. Then these rules can be tested in ResSIM or

HEC-ResPRM can test rules by making them constraints rather than penalty curves to encourage the

desired behavior. These rules are particularly applicable to the refill season of the reservoir, where

inflows are in excess of hydropower water supply demands. During the draw-down season, where

reservoir inflows are less than demands, the system should be drawn down in order of the reservoir to

131
provide storage for the refill season. The storage versus mean monthly graphs (rule curve) of Tekeze

hydropower reservoir for s specific time and climate scenarios was shown in Appendix F13.

[Link] Optimum Reservoir Pool level (Elevation) under Climate Change

HEC-ResPRM optimization result indicates that Tekeze hydropower reservoir pool level will be

increased under RCP4.5 and RCP8.5 scenarios in all projected time periods. This comparison made

with the current mean annual reservoir operation pool level of 1112.61 masl from 2009–2017. In the three

projected time periods, optimal pool level (Table 5.2) change varies from 8.26 to 8.45 m under RCP4.5

and 10.66–11.24 m under RCP8.5 climate scenarios. This is large elevation difference that will store

more water in the rainy months for the dry season power production. The optimized pool levels under

both scenarios in all time periods are larger than the current optimized pool level. This is due to the

impact of climate change and hydrological non-stationarity on reservoir operation. The reservoir

storage pool level change in RCP4.5 scenarios is lower than RCP8.5 scenarios due to increase in each

year individual month’s fluctuations in RCP4.5 scenarios because of future inflow variability that

reduced the mean annual reservoir water storage level.

Table 5.2 Mean annual optimized pool level variation under RCP4.5 and RCP8.5 climate scenarios in

three future time periods.

Optimized Pool Level

Pool Level Change (m)
Time Periods (masl)
RCP4.5 RCP8.5 RCP4.5 RCP8.5
Current
1120.48 7.87
optimized
2020s 1121.06 1123.27 8.45 10.66
2050s 1120.87 1123.37 8.26 10.76
2080s 1120.89 1123.85 8.28 11.24

The Tekeze hydropower reservoir operational level is changing continually due to inflows occurred

and releases are made to produce power. The start of dead storage level at 1096 masl (minimum live

132
storage level) has been assumed for power production. Figure 5.6 a, b, c shows the optimal reservoir

pool level of Tekeze hydropower reservoir generated by HEC-ResPRM optimization model. These

optimal pool level results have a similar pattern with the optimal reservoir storage variations and may

be considered as rule curves for optimal operation of Tekeze hydropower reservoir under a given

scenario and time period. The reservoir pool level stayed at high level every year from August to

November when reservoirs filled during the rainy months of August through September. The drop of

pool level in June caused due to optimization model constraint reservoir not emptied and a transition

zone when the drawdown ends, and reservoir refill start.

The result in this study indicates the comparison of Tekeze hydropower reservoir optimal pool level

and storage changes under climate change impacts. This will have a positive outcome in terms of

showing the impacts of HEC-ResPRM optimization model and climate change impacts on the

hydropower reservoir operation. However, it is sound to discuss the results in terms of how much

hydropower produced in the future periods based on optimization models. This is difficult due to lack

off observed historical power production data for comparison and turbine performance/ efficiency

curves to prepare hydropower penalty functions as only one turbine is operational out four turbines.

In general, there is no doubt that the hydropower reservoir system of Tekeze basin will be affected by

climate change. With over all predicted increases in precipitation and streamflow, inflow to the

reservoir anticipated to increase. Therefore, even though both RCP4.5 and RCP8.5 climate scenarios

exhibited higher water inflow volume to Tekeze hydropower reservoir, this did not necessarily result

in significantly more hydropower generation. Optimal operation of the reservoir using HEC-ResPRM

considerably increases the power production by storing the higher inflow volume to inflow deficiency

periods even if hydropower production has a non-linear relationship with storage and water level

(head).

133
Based on this research, the potential for hydropower production in the Tekeze hydropower reservoir is

predicted to increase if appropriate reservoir operation techniques are used by operators and water

managers to store and use the wet month’s flow to the dry months. The optimal operation model

developed in this study can be helpful to dam operators in building generation plans in the future. It

also enables a flexible response when inflow occurs during actual operation that is different from the

initial planning expectations, as it can generate new operational alternatives appropriate for the

changing situation. Even if many optimization models and techniques have been developed in several

fields of water resources system analysis such as hydropower reservoir operation around the world, the

adaptation of such techniques and tools by water managers is slow. Researchers and scientists must

accept the fact that the gap still exists between research studies and applications in practice. There need

to be research on how to translate science to improve management operations of reservoirs for optimal

results.

134
Figure 5.6 Monthly optimum pool level variations under RCP4.5 and RCP8.5 climate

scenarios in three time periods: (a) 2020s; (b) 2050s; (c) 2080s

135
5.5 Conclusive Remarks
This study used a semi-distributed hydrological model (SWAT) and a reservoir optimization model

(HEC-ResPRM) to evaluate the hydrological impacts of climate change on Tekeze hydropower

reservoir operation in Tekeze basin part of Eastern Nile. In this research, it was evaluated climatic data

(past and future periods) from the ensemble outputs of CORDEX-Africa RCMs under RCP4.5 and

RCP8.5 climate scenarios for the periods of 2020s, 2050s, and 2080s. Calibrated SWAT model was

used to generate climate change induced streamflow that was used as an input for optimal reservoir

operation modeling. Analysis conducted on Tekeze hydropower reservoir inflows and outflow,

reservoir storage volume and reservoir pool levels revealed the following:

1. This study found that the impact of climate change would increase in precipitation, temperature

and streamflow in Tekeze basin under RCP4.5 and RCP8.5 climate scenarios over future

periods which have an impact on current and future Tekeze hydropower reservoir operation.

2. Projected annual and inter-annual reservoir inflow showed increasing trend under both RCP4.5

and RCP8.5 climate scenarios.

3. HEC-ResPRM incorporates water storage, water surface elevation, release and power

generation would provide better understanding of current and future conditions of Tekeze

hydropower reservoir operation.

4. Current optimized power storage and pool level show more optimal results than the current

actual operation, so it is recommended to change the current operating policy to produce more

power throughout the year.

5. The projected increase of reservoir inflow under an ensemble of CORDEX-Africa RCP4.5 and

RCP8.5 future climate scenarios lead to optimized reservoir power storage, pool level (head)

and release that greatly exceed those historically observed, indicating a shift in current water

system behavior.

136
6. The study showed that climate change clearly affects future reservoir planning and

management in Tekeze basin. Therefore, water resources planners, managers and operators

should consider climate change impacts in the design, planning and management of reservoir

systems.

7. In practice, many reservoir system operators and water managers feel more comfortable to use

pre-defined rule curves and simulation results which are easy to understand and operate as

most optimal operating rules developed by scientists using sophisticated optimization models

and algorithms are mathematically more complex. The use of a combination of simulation and

optimization models may solve this problem.

137
6 CONCLUSIONS AND RECOMMENDATIONS

6.1 Conclusions

This study has been carried out in the Tekeze River basin to understand how changes in hydrological

regime in the future caused by climate variability and climate change affects the performance of

hydropower dam/ reservoir. The result of these studies can help to inform decision-makers about

possible ways of planning and operation of reservoir/dam in the future to optimize power generation

and to ensure that the demands of other users met.

The seasonal and annual trends of precipitation at eleven gauging stations of Tekeze basin and

streamflow at Embamadre station which have good quality data sets with reliable data and adequate

record length were analyzed by the non-parametric Mann-Kendall and Sen’s method. The result

demonstrated good agreement of performance in detection of the trends for precipitation at different

parts of the basin. Based on the results of statistical methods, the stations found in the South and Eastern

parts of the basin showed non-significant positive trends whereas non-significant negative trends

observed at the Northern part of the basin at 5% significant level for the annual precipitation series.

However, most of the precipitation stations showed significant increasing trends in Kiremt and Belg

seasons. The trend and variability analysis of streamflow also showed that a non-significant increasing

annual trend but Kiremt and Belg seasons streamflow trend showed a significant positive and Bega

season with significant negative trends. The findings presented here on the spatio-temporal trends and

variability’s of Tekeze basin precipitation and streamflow can be implemented to improve the water

resources strategies in the basin.

The hydrologic regimes before and after Tekeze hydropower reservoir operation were quantitatively

analyzed using the range of variability approach based on the indicators of hydrologic alteration. The

results for many of the parameters, the magnitude, duration and direction showed a significant alteration

138
like a significant increase in 1- through 90-day minimum streamflow, decrease in 1- through 90-day

maximum streamflow, rise and fall rates decreased and increase in annual hydrograph reversals.

Generally, the hydrologic regime of the lower Tekeze River changed significantly after the hydropower

reservoir began operation. The result of this study will be beneficial to the future regional water

resources management and effects of this alteration on river ecosystems with in the lower Tekeze basin.

In this studya semi-distributed hydrological model (SWAT) and a reservoir operation optimization

model (HEC-ResPRM) used to evaluate the hydrological impacts of climate change on hydropower

reservoir operation in the Tekeze River basin, Ethiopia. Meteorological products assessed during the

past periods (1976-2005) and the ensemble of CORDEX-Africa RCMs under RCP4.5 and RCP8.5

bias-corrected climate change scenarios during a future period (2011-2100) in concert with the

proposed hydropower water infrastructure plans. Past simulations driver with ground-based forcing

were calibrated and tested with stream gauging data at [Link] conducted on reservoir

inflows, pool elevations and power storage volume for meteorological and infrastructure scenarios

revealed the following:

1. Tekeze River basin has extreme hydrological variability and seasonality (Highly variable rainfall)

that results in runoff highly variable and seasonal across the basin, and from year to year which

has an impact on current operations and future planned hydropower reservoirs operation.

2. This study found that the impact of climate change will increase in precipitation, temperature, and

streamflow in the Tekeze river basin in both biases corrected CORDEX-Africa RCMs under

RCP4.5 and RCP8.5 climate scenarios over future periods.

3. According to the latest inflow projections based on RCP4.5 and RCP8.5, climate scenarios total

inflows to Tekeze hydropower reservoirs slightly increases in all time horizons, stored water in the

139
 reservoir increases which has many implications for future hydropower dams/reservoirs planning

and operation.

4. The new reservoir optimization modeling approach HEC-ResPRM incorporates water storage,

water surface elevation (i.e., head), and release would provide the better understanding of current

and future conditions of optimum hydropower dams/reservoirs operation in the Tekeze basin.

5. The future period contained a significantly higher amount of precipitation relative to the historical

simulation with the bias-corrected CORDEX-Africa RCP scenarios approach. This trend led to the

projection of higher optimized reservoir inflow, pool elevation, and storage from the Tekeze

hydropower reservoir. As a result, the future scenarios lead to optimized reservoir power storage

and pool elevations in the operational reservoir that significantly exceeds those observed

historically, indicating a shift in current water system behavior. Therefore, it is better first to

improve existing hydropower reservoir (TK05) capacity rather than investing in new planned

hydropower projects.

6. Incorporate climate change scenarios into dam design: The significant implication of climate

change in Tekeze River basin for dams and reservoirs is that the future is uncertain, and can no

longer be assumed to mirror the past. Hydropower reservoir/dam projects to design and operation

should approach with extreme caution until reliable data series available in the basin. Climatic

uncertainty should be incorporate into dam design, to avoid the hazards of over- or under-designed

infrastructure and financial risk.

140
6.2 Recommendations

The modeling exercise was relatively successful in the planning and operation of Tekeze hydropower

reservoir operation under hydrological variability and climate change. However, this study still has

some limitations which could address in future research. Since it is the first study in Tekeze basin

hydropower reservoir operation under climate change, further studies are recommended using multiple

hydrologic models, climate models and hydropower reservoir simulation and optimization models

which will provide more precise results. Therefore, to improve the outcome and the use of the entire

model capabilities based on future insights of Tekeze hydropower projects, the following is proposed

for future work:

i. Statistical precipitation trend analysis of this study investigates only the magnitudes and changes

of historical data without being able to quantify more precipitation characteristics like frequencies

of high precipitations or low precipitations, increase or decrease in the number of rainy days,

rainfall concentration index, standardized anomalies of annual rainfall and changes in dry-spell

length. Therefore, further investigations are needed by considering all these characteristics and

regionalize the results for the whole basin. Further investigation also needed for streamflow trend

detection considering land degradation and landscape change with human activities.

ii. Further climate change modeling at the Tekeze basin is needed to more clearly define the trends

and ranges of climate change and extreme events that need to incorporate into hydropower

development plans.

iii. This study has not considered the changes in land use/land cover due to socio-economic

development in the future. Land cover /land use properties were considered constant throughout

the simulation period; such an assumption can affect the projections of streamflow in the basin.

Land use changes including projected changes in associated with climate change impacts and effect

141
 erosion and sediment yield which are necessary to evaluate predicted changes in runoff associated

with future Tekeze River basin development. Hence, further studies are recommended to couple

climate models and various bias correction methods with land use change to quantify the projected

change in streamflow and sedimentation load in Tekeze hydropower reservoirs as well as its

implication on future hydropower generation from Tekeze river basin.

iv. Lack of sufficient observed data highlighted some of the challenges Tekeze basin face. The climate

model analysis and hydrological modeling were affected by the absence of the data. Future work

is required to collect more data through means that are sustainable. It is highly recommended to

improve both climate and hydrological monitoring system in the whole Tekeze basin. Furthermore,

existing monitoring network and data are insufficient to calibrate and verify hydrological models

of Tekeze River basin. A long-term commitment to monitoring critical hydrological variables

associated with climate change needed at the sub-basin level. That mainly includes maintaining

and expanding monitoring networks for temperature, rainfall, river stage, and discharge. Specific

focus should give to improving knowledge and confidence concerning the relationship between

climate change and river discharge as discharge directly affects hydropower production.

v. Further studies recommended using hydro-economic model that allowed for scenario calculation

of how reservoir water levels, storage and power production changes with climate change scenarios

and with a change in operating scheme of the reservoir (increase in turbine capacity). Also Further

model development will enable the consideration of a variety of additional parameters, such as

water withdrawal for planned irrigation, drinking water supply, or altered energy policies.

142
REFERENCES

Abtew, W., Melesse, A.M., Dessalegne, T., 2009. Spatial, inter and intra-annual variability of the Upper
Blue Nile Basin rainfall. Hydrol. Process. 23, 3075–3082. [Link]

Adem, A.A., Melesse, A.M., Tilahun, S.A., Setegn, S.G., Ayana, E.K., Wale, A., Assefa, T.T., 2014.
Climate Change Projections in the Upper Gilgel Abay River Catchment, Blue Nile Basin Ethiopia, in:
Nile River Basin. Springr, Cham, pp. 363–388. [Link]

Alazzy, A., Lü, H., Zhu, Y., 2014. Impact of climate change on evaluation of future water demand in
the Euphrates and Aleppo basin, Syria, in: Proceedings of the International Association of Hydrological
Sciences. pp. 307–312. [Link]

Arnold, J.., Srinivasan, R., Muttiah, R.., Williams, J.., 1998. Large area hydrologic modeling and
assessment part I: model development. J. Am. Water Resour. Assoc. 34, 73–89.

Arnold, J.G., Kiniry, J.R., Srinivasan, R., Williams, J.R., Haney, E.B., Neitsch, S.L., 2013. SWAT
2012 Input/output Documentation.

Arsano, Y., Tamrat, I., 2005. Ethiopia and the Eastern Nile Basin. Aquat. Sci. 67, 15–27.
[Link]

Awulachew, S.., Yilma, D.., Loulseged, M., Loiskandl, W., Ayana, M., Alamirew, T., 2007. Water
Resources and irrigation development in Ethiopia. International Water Management Institute,
Colombo, Sri Lanka.

Ayalew, L., 1999. The effect of seasonal rainfall on landslides in the highlands of Ethiopia. Bull. Eng.
Geol. Environ. 58, 9–19. [Link]

Azizipour, M., Ghalenoei, V., Afshar, M.H., Solis, S.S., 2016. Optimal Operation of Hydropower
Reservoir Systems Using Weed Optimization Algorithm. Water Resour. Manag. 30, 3995–4009.
[Link]

Bates, Kundzewicz, Z. W., Wu, S., Palutikof, J. P, 2008. Climate change and water. Technical paper
of the Intergovernmental Panel on Climate Change. Geneva.

Behailu, M., 2002. Assessment and optimizations of traditional irrigation of Vertisols in northern
Ethiopia: a case study at Gumselalsa micro-dam using maize as an indicator crop. Unpublished PhD
Thesis. Gent, Belgium.

143
Bewket, W., Conway, D., 2007. A note on the temporal and spatial variability of rainfall in the drought-
prone Amhara region of Ethiopia. Int. J. Climatol. 27, 1467–1477. [Link]

Beyene, T., Lettenmaier, D.P., Kabat, P., 2010. Hydrologic impacts of climate change on the Nile River
Basin: implications of the 2007 IPCC scenarios. Clim. Change 100, 433–461.
[Link]

Birhanu, K., Alamirew, T., Dinka, M.O., Ayalew, S., Aklog, D., 2014. Optimizing Reservoir Operation
Policy Using Chance Constraint Nonlinear Programming for Koga Irrigation Dam, Ethiopia. Water
Resour. Manag. 28, 4957–4970. [Link]

Block, P.J., Strzepek, K., Rosegrant, M.W., Diao, X., 2008. Impacts of considering climate variability
on investment decisions in Ethiopia. Agric. Econ. 39, 171–181. [Link]
0862.2008.00322.x

Cannarozzo, M., Noto, L.V., Viola, F., 2006. Spatial distribution of rainfall trends in Sicily (1921–
2000). Phys. Chem. Earth Parts ABC, Time Series Analysis in Hydrology 31, 1201–1211.
[Link]

Casanueva, A., Rodriguez-Puebla, C., D. Frías, M., González-Reviriego, N., 2013. Variability of
extreme precipitation over Europe and its relationships with teleconnection patterns.
[Link]

Chen, H., Xu, C.-Y., Guo, S., 2012. Comparison and evaluation of multiple GCMs, statistical
downscaling and hydrological models in the study of climate change impacts on runoff. J. Hydrol. 434–
435, 36–45. [Link]

Chen, J., Brissette, F.P., Poulin, A., Leconte, R., 2011. Overall uncertainty study of the hydrological
impacts of climate change for a Canadian watershed. Water Resour. Res. 47, W12509.
[Link]

Cheng, C.-T., Wang, W.-C., Xu, D.-M., Chau, K.W., 2008. Optimizing Hydropower Reservoir
Operation Using Hybrid Genetic Algorithm and Chaos. Water Resour. Manag. 22, 895–909.
[Link]

Cheung, W.H., Senay, G.B., Singh, A., 2008. Trends and spatial distribution of annual and seasonal
rainfall in Ethiopia. Int. J. Climatol. 28, 1723–1734. [Link]

144
Christensen, J.H., Boberg, F., Christensen, O.B., Lucas-Picher, P., 2008. On the need for bias correction
of regional climate change projections of temperature and precipitation. Geophys. Res. Lett. 35,
L20709. [Link]

Connell, S.O., Harou, J., 2011. HEC-ResPRM, Perspective Reservoir Model quick start guide Version
1.0, CPD-95a. US Army Corps of Engineers, Institute for Water Resources, Hydrologic Engineering
Center (HEC), Davis, USA.

Conway, D., 2005. From headwater tributaries to international river: observing and adapting to climate
variability and change in the Nile Basin. Glob. Environ. Change 15, 99–114.

Conway, D., 2002. Extreme Rainfall Events and Lake Level Changes in East Africa: Recent Events
and Historical Precedents, in: The East African Great Lakes: Limnology, Palaeolimnology and
Biodiversity, Advances in Global Change Research. Springer, Dordrecht, pp. 63–92.
[Link]

Conway, D., 2000. Some aspects of climate variability in the North East Ethiopian Highlands – Wollo
and Tigray. [Link]

Conway, D., 1997. A water balance model of the Upper Blue Nile in Ethiopia. Hydrol. Sci. J. 42, 265–
286. [Link] 10.1080/02626669709492024

Conway, D., Hulme, M., 1993. Recent fluctuations in precipitation and runoff over the Nile sub-basins
and their impact on main Nile discharge. Clim. Change 25, 127–151.

Conway, D., Persechino, A., Ardoin-Bardin, S., Hamandawana, H., Dieulin, C., Mahé, G., 2009.
Rainfall and Water Resources Variability in Sub-Saharan Africa during the Twentieth Century. J.
Hydrometeorol. 10, 41–59. [Link]

David, S., Randolph, B., Upali, A., 1999. Water scarcity in the Twenty-first century. International
journal of Water resources Development 15, 29–42. [Link]

Dawadi, S., Ahmad, S., 2013. Evaluating the Impact of Demand-Side Management on Water Resources
under Changing Climatic Conditions and Increasing Population. J. Environ. Manage. 261–275.
[Link]

Dessie, M., Verhoest, N.E.C., Admasu, T., Pauwels, V.R.N., Poesen, J., Adgo, E., Deckers, J., Nyssen,
J., 2014. Effects of the floodplain on river discharge into Lake Tana (Ethiopia). J. Hydrol. 519, 699–

145
710. [Link]

Dessu, S.B., Melesse, A.M., 2013. Impact and uncertainties of climate change on the hydrology of the
Mara River basin, Kenya/Tanzania. Hydrol. Process. 27, 2973–2986. [Link]

Dessu, S.B., Melesse, A.M., 2012. Modelling the rainfall–runoff process of the Mara River basin using
the Soil and Water Assessment Tool. Hydrol. Process. 26, 4038–4049.
[Link]

Dessu, S.B., Melesse, A.M., Bhat, M.G., McClain, M.E., 2014. Assessment of water resources
availability and demand in the Mara River Basin. CATENA 115, 104–114.
[Link]

Douglas, E.M., Vogel, R.M., Kroll, C.N., 2000. Trends in floods and low flows in the United States:
impact of spatial correlation. J. Hydrol. 240, 90–105.

Ehsani, N., J. Vörösmarty, C., Fekete, B., Z. Stakhiv, E., 2017. Reservoir Operations Under Climate
Change: Storage Capacity Options to Mitigate Risk. J. Hydrol. 555.
[Link]

Ehteram, M., Allawi, M.F., Karami, H., Mousavi, S.-F., Emami, M., EL-Shafie, A., Farzin, S., 2017.
Optimization of Chain-Reservoirs’ Operation with a New Approach in Artificial Intelligence. Water
Resour. Manag. 31, 2085–2104. [Link]

Elshamy, M.., Seierstad, I.., Sorteberg, A., 2009. Impacts of climate change on Blue Nile flows using
bias-corrected GCM scenarios. Hydrol Earth Syst Sci 13, 551–565.

Enyew, B., Van Lanen, H., Van Loon, A., 2014. Assessment of the Impact of Climate Change on
Hydrological Drought in Lake Tana Catchment, Blue Nile Basin, Ethiopia. J. Geol. Geophys. 3, 1–17.
[Link]

Faber Beth A., Harou Julien J., 2006. Multiobjective Optimization with HEC Res-PRM ? Application
to the Upper Mississippi Reservoir System. Oper. Reserv. Chang. Cond., Proceedings.
[Link]

Fang, G., Yang, J., Chen, Y., Zammit, C., 2015. Comparing bias correction methods in downscaling
meteorological variables for hydrologic impact study in an arid area in China.
[Link]

146
FAO, 1995. Soils of EAST Africa, SEA,. Food and Agriculture Organization of the United Nations,
ACD-Rom Data, Rome.

Fiseha, B.M., Setegn, S.G., Melesse, A.M., Volpi, E., Fiori, A., 2014. Impact of Climate Change on
the Hydrology of Upper Tiber River Basin Using Bias Corrected Regional Climate Model. Water
Resour. Manag. 28, 1327–1343. [Link]

Fiseha, B.M., Setegn, S.G., Melesse, A.M., Volpi, E., Fiori, A., 2013. Hydrological analysis of the
Upper Tiber River Basin, Central Italy: a watershed modelling approach. Hydrol. Process. 27, 2339–
2351. [Link]

Fowler, H.J., Blenkinsop, S., Tebaldi, C., 2007. Linking climate change modelling to impacts studies:
recent advances in downscaling techniques for hydrological modelling. Int. J. Climatol. 27, 1547–1578.
[Link]

Gao, B., Yang, D., Yang, H., 2013. Impact of the Three Gorges Dam on flow regime in the middle and
lower Yangtze River. QuatInt 304, 43–50.

Gassman, P., Reyes, M., Green, C., Arnold, J., 2007. The Soil and Water Assessment Tool: Historical
Development, Applications, and Future Research Directions. CARD Work. Pap.

Gassman, P.., Sadeghi, A.., Srinivasan, R., 2014. Applications of the SWAT model special section:
overview and insights. J. Environ. Qual. 43, 1–8. [Link]

Gbobaniyi, E., Sarr, A., Sylla, M.B., Diallo, I., Lennard, C., Dosio, A., Dhiédiou, A., Kamga, A.,
Klutse, N.A.B., Hewitson, B., Nikulin, G., Lamptey, B., 2014. Climatology, annual cycle and
interannual variability of precipitation and temperature in CORDEX simulations over West Africa. Int.
J. Climatol. 34, 2241–2257. [Link]

Gebremedhin, B., Pender, J., Ehui, S., 2003. Land Tenure and Land Management in the Highlands of
Northern Ethiopia. Ethiop. J. Econ. 8, 46–63.

Gebremicael, T.., Mohamed, G.., Betrie, P., van der Zaag, 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 land use maps. J. Hydrol. 482, 57–68.

Ghashghaie, M., Marofi, S., Marofi, H., 2014. Using System Dynamics Method to Determine the Effect
of Water Demand Priorities on Downstream Flow. Water Resour. Manag. 28.

147
[Link]

Gidey, T.., Gebrehiwot, T.., Veen, A. van der, 2013. Assessing the evidence of climate variability in
the northern part of Ethiopia. J Dev Agric Econ 5, 104–119. [Link]

Giorgi, F., Jones, C., Asrar, G., 2009. Addressing climate information needs at the regional level: the
CORDEX framework. World Meteorol Organ Bull 58, 175–183. http://
[Link]/RCD_Projects/CORDEX/CORDEX_giorgi_ WMO

Gizaw, M.S., Biftu, G.F., Gan, T.Y., Moges, S.A., Koivusalo, H., 2017. Potential impact of climate
change on streamflow of major Ethiopian rivers. Clim. Change 143, 371–383.
[Link]

Gocic, M., Trajkovic, S., 2013. Analysis of changes in meteorological variables using Mann-Kendall
and Sen’s slope estimator statistical tests in Serbia. Glob. Planet. Change 100, 172–182.

Graham, L.P., Andréasson, J., Carlsson, B., 2007. Assessing climate change impacts on hydrology from
an ensemble of regional climate models, model scales and linking methods – a case study on the Lule
River basin. Clim. Change 81, 293–307. [Link]

Green, W.H., Ampt, G.A., 1911. Studies on soil physics: the flow of air and water through soils. J.
Agric. Sci. 4, 1–24. [Link]

Grey, O., St F G Webber, D., Setegn, S., Melesse, A., 2014. Application of the Soil and Water
Assessment Tool (SWAT Model) on a small tropical island (Great River Watershed, Jamaica) as a tool
in Integrated Watershed and Coastal Zone Management. Int. J. Trop. Biol. Conserv. 62, 293–305.

Guo, S., Chen, J., Li, Y., Liu, P., Li, T., 2011. Joint Operation of the Multi-Reservoir System of the
Three Gorges and the Qingjiang Cascade Reservoirs. Energies 4, 1036–1050.
[Link]

Haile, A.T., Akawka, A.L., Berhanu, B., Rientjes, T., 2017. Changes in water availability in the Upper
Blue Nile basin under the representative concentration pathways scenario. Hydrol. Sci. J. 62, 2139–
2149. [Link]

Hales, S., 2007. Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution of
Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change,
in: Journal of Environmental Quality - J ENVIRON QUAL. p. 976.

148
Hamududu, B.H., Killingtveit, Å., 2016. Hydropower Production in Future Climate Scenarios; the Case
for the Zambezi River. Energies 9, 1–18.

Haregeweyn, N., Poesen, J., Nyssen, J., Gerared, G., Verstraeten, G., de Vente, J., Deckers, L.,
Moeversons, J., Haile, M., 2008. Sediment yield variability in Northern Ethiopia: A quantitative
analysis of its controlling factors. Catena 75, 65–76.

He, Y., Xu, Q., Yang, S., Liao, L., 2014. Reservoir flood control operation based on chaotic particle
swarm optimization algorithm. Appl. Math. Model. 38, 4480–4492.
[Link]

Hernández-Díaz, L., Laprise, R., Sushama, L., Martynov, A., Winger, K., Dugas, B., 2013. Climate
simulation over CORDEX Africa domain using the fifth-generation Canadian Regional Climate Model
(CRCM5). ClimDyn 40, 1415–1433. [Link] 10.1007/s00382-012-1387-z.

Huang, J., Sun, S., Xue, Y., Li, J., Zhang, J., 2014. Spatial and Temporal Variability of Precipitation
and Dryness/Wetness During 1961–2008 in Sichuan Province, West China. Water Resour. Manag. 28,
1655–1670. [Link]

Huntington, T.G., 2010. Chapter One - Climate Warming-Induced Intensification of the Hydrologic
Cycle: An Assessment of the Published Record and Potential Impacts on Agriculture, in: Sparks, D.L.
(Ed.), Advances in Agronomy. Academic Press, Cambrige, MA, US, pp. 1–53.
[Link]

Ines, A.V.M., Hansen, J.W., 2006. Bias correction of daily GCM rainfall for crop simulation studies.
Agric. For. Meteorol. 138, 44–53. [Link]

Jacob, D., Elizalde, A., Haensler, A., Teichmann, C., Wilhelm, C., 2012. Assessing the transferability
of the Regional Climate Model REMO to different COordinated Regional Climate Downscaling
EXperiment (CORDEX) regions. Atmosphere 3, 181–199.

Jamali Saeed, Abrishamchi Ahmad, Madani Kaveh, 2013. Climate Change and Hydropower Planning
in the Middle East: Implications for Iran’s Karkheh Hydropower Systems. J. Energy Eng. 139, 153–
160. [Link]

Jiang, L., Ban, X., Wang, X., Cai, X., 2014. Assessment of Hydrologic Alterations Caused by the Three
Gorges Dam in the Middle and Lower Reaches of Yangtze River, China. Water 6, 1419–1434.
[Link]

149
Johnson, P., Curtis, P., 1994. Water balance of Blue Nile River Basin in Ethiopia. J. Irrig. Drain. Eng.
120, 573–590.

Jury, W.A., Henry, J.V.J., 2007. The emerging global water crisis: managing scarcity and conflict
between water users. Adv. Agron. 95, 1–76.

Kebede, S., Travi, Y., Alemayehu, T., Marc, V., 2006. Water balance of Lake Tana and its sensitivity
to fluctuations in rainfall, Blue Nile basin, Ethiopia. J. Hydrol. 316, 233–247.
[Link]

Kendal, M.., 1975. Rank Correlation Methods. 4th Edition, Charles Griffin, London. [WWW
Document]. URL [Link] (accessed
1.5.18).

Kim, J., Waliser, D.E., Mattmann, C.A., Goodale, C.E., Hart, A.F., Zimdars, P.A., Crichton, D.J.,
Jones, C., Nikulin, G., Hewitson, B., Jack, C., Lennard, C., Favre, A., 2014. Evaluation of the
CORDEX-Africa multi-RCM hindcast: systematic model errors. Clim. Dyn. 42, 1189–1202.
[Link]

Kim, K.B., Kwon, H.-H., Han, D., 2015. Bias correction methods for regional climate model
simulations considering the distributional parametric uncertainty underlying the observations. J.
Hydrol. 530, 568–579. [Link]

Kim, U., Kaluarachchi, J.J., 2009. Climate Change Impacts on Water Resources in the Upper Blue Nile
River Basin, Ethiopia1. JAWRA J. Am. Water Resour. Assoc. 45, 1361–1378.
[Link]

Kiros, G., Shetty, A., Nandagiri, L., 2017. Extreme rainfall signatures under changing climate in semi-
arid northern highlands of Ethiopia. Cogent Geosci, 1353719 3.
[Link]

Labadie John W., 2004. Optimal Operation of Multireservoir Systems: State-of-the-Art Review. J.
Water Resour. Plan. Manag. 130, 93–111. [Link]

Larson, K.L., Polsky, C., Gober, P., Chang, H., Shandas, V., 2013. Vulnerability of Water Systems to
the Effects of Climate Change and Urbanization: A Comparison of Phoenix, Arizona and Portland,
Oregon (USA). Environ. Manage. 52, 179–195. [Link]

150
Lee, S.-Y., Hamlet, A.F., Grossman, E.E., 2016. Impacts of Climate Change on Regulated Streamflow,
Hydrologic Extremes, Hydropower Production, and Sediment Discharge in the Skagit River Basin.
Northwest Sci. 90, 23–43. [Link]

Li, X., Guo, S., Liu, P., Chen, G., 2010. Dynamic control of flood limited water level for reservoir
operation by considering inflow uncertainty. J. Hydrol. 391, 124–132.
[Link]

Liu, P., Guo, S., Xu, X., Chen, J., 2011. Derivation of Aggregation-Based Joint Operating Rule Curves
for Cascade Hydropower Reservoirs. Water Resour. Manag. 25, 3177–3200.
[Link]

Liu, X., Ren, L., Yuan, F., Singh, V.P., Fang, X., Yu, Z., Zhang, W., 2009. Quantifying the effect of
land use and land cover changes on green water and blue water in northern part of China. Hydrol Earth
Syst Sci 13, 735–747. [Link]

Lu, B., Li, K., Zhang, H., Wang, W., Gu, H., 2013. Study on the optimal hydropower generation of
Zhelin reservoir. J. Hydro-Environ. Res. 7, 270–278. [Link]

Lumbroso, D.M., Woolhouse, G., Jones, L., 2015. A review of the consideration of climate change in
the planning of hydropower schemes in sub-Saharan Africa. Clim. Change 133, 621–633.
[Link]

Maingi, J.., Marsh, S.., 2002. Quantifying hydrologic impacts following dam construction along the
Tana River, Kenya. Arid Env. 50, 53–79.

Mango, L., Melesse, A., McClain, M., Gann, D., Setegn, S., 2011. Hydro-meteorology and water
budget of Mara River basin, Kenya: A land use change scenarios analysis, In: Nile River Basin:
Hydrology, Climate and Water Use, (Ed., Melesse, A.M).

Mann, H.B., 1945. Nonparametric tests against trend. Econometrica, 13, 245-259.

McCartney, M.P., MenkerGirma, M., 2012. Evaluating the downstream implications of planned water
resource development in the Ethiopian portion of the Blue Nile River. Water Int. 37, 362–379.
[Link]

McMahon, G.., Farmer, M.., 2009. Rule-based storage accounting for multipurpose reservoir systems.
J Water ResourPlann Manage 135, 286–297.

151
Meinshausen, M., Smith, S.J., Calvin, K., Daniel, J.S., Kainuma, M.L.T., Lamarque, J.-F., Matsumoto,
K., Montzka, S.A., Raper, S.C.B., Riahi, K., Thomson, A., Velders, G.J.M., Vuuren, D.P.P. van, 2011.
The RCP greenhouse gas concentrations and their extensions from 1765 to 2300. Clim. Change 109,
213. [Link]

Mekonnen, M.., Worman, A., Dargahi, B., Gebeyehu, A., 2009. Hydrological modelling of Ethiopian
catchments using limited data. Hydrol Process 23, 3401–3408.

Melesse, A., Bekele, S., McCornick, P., 2011. Hydrology of the Niles in the face of land-use and
climate dynamics. In Nile River Basin: Hydrology, Climate and Water Use, (Editor Melesse, A.M).

Melesse, A.M., Loukas, A.G., Senay, G., Yitayew, M., 2009. Climate change, land-cover dynamics
and ecohydrology of the Nile River Basin. Hydrol. Process. 23, 3651–3652.
[Link]

Mengistu, D., Bewket, W., Lal, R., 2014. Recent spatiotemporal temperature and rainfall variability
and trends over the Upper Blue Nile River Basin, Ethiopia. Int. J. Climatol. 34, 2278–2292.
[Link]

Mengistu, D.T., Sorteberg, A., 2012. Sensitivity of SWAT simulated streamflow to climatic changes
within the Eastern Nile River basin. Hydrol Earth Syst Sci 16, 391–407. [Link]
16-391-2012

Middelkoop, H., Daamen, K., Gellens, D., Grabs, W., Kwadijk, J.C.J., Lang, H., Parmet, B.W. a. H.,
Schädler, B., Schulla, J., Wilke, K., 2001. Impact of Climate Change on Hydrological Regimes and
Water Resources Management in the Rhine Basin. Clim. Change 49, 105–128.
[Link]

Milly, P., Betancourt, J., Julio, Falkenmark, M., Malin, Hirsch, R., M, R., Kundzewicz, W, Z.,
Lettenmaier, P, D., Ronald, S., J, R., 2008. Stationarity Is Dead: Whither Water Management? Science
319, 573–574.

Moges, S.A., Taye, M.T., Willems, P., Gebremichael, M., 2014. Exceptional pattern of extreme rainfall
variability at urban centre of Addis Ababa, Ethiopia. Urban Water J. 11, 596–604.
[Link]/10.1080/1573062X.2013.831914

Moriasi, D., Arnold, J., Van Liew, M., Bingner, R., Harmel, R.D., Veith, T., 2007. Model Evaluation
Guidelines for Systematic Quantification of Accuracy in Watershed Simulations. Trans. ASABE 50.

152
[Link]

Mosquera-Machado, S., Ahmad, S., 2007. Flood hazard assessment of Atrato river in Colombia. Water
Resour. Manag. 591–609. [Link]

Mu, X., Zhang, L., McVicar, T.R., Chille, B., Gau, P., 2007. Analysis of the impact of conservation
measures on stream flow regime in catchments of the Loess Plateau, China. Hydrol. Process. 21, 2124–
2134. [Link]

Nash, J., Sutcliffe, J., 1970. River flow forecasting through conceptual models part 1 – a discussion of
principles. J. Hydrol. 10, 282–290.

NEDECO, 1998. Tekeze river basin integrated development master plan project, Sectorial reports,
volume X; Water Resource, Dams, Reservoirs and Hydropower Development. The Netherlands
Engineering consultant in association with DHV consultant, MoWR, Addis Ababa, Ethiopia.

Neitsch, S., Arnold, J., Kiniry, J.R., Williams, J.R., King, K., 2005. SWAT theoretical documentation.
Grassland 494, 234–235.

Nikulin, G., Jones, C., Giorgi, F., Asrar, G., Büchner, M., Cerezo-Mota, R., Christensen, O.B., Déqué,
M., Fernandez, J., Hänsler, A., van Meijgaard, E., Samuelsson, P., Sylla, M.B., Sushama, L., 2012.
Precipitation Climatology in an Ensemble of CORDEX-Africa Regional Climate Simulations. J. Clim.
25, 6057–6078. [Link]

Nyssen, J., Poesen, J., Moeyersons, J., Dackers, J., Haile, M., Lang, A., 2004. Human impact on the
environment in the Ethiopian and Eritrean Highlands-a State of the art. Earth Sci. Rev. 64, 273–300.

Oliveira, R., Loucks, D.P., 1997. Operating rules for multireservoir systems. Water Resour. Res. 33,
839–852. [Link]

Ostadrahimi, L., Mariño, M.A., Afshar, A., 2012. Multi-reservoir Operation Rules: Multi-swarm PSO-
based Optimization Approach. Water Resour. Manag. 26, 407–427. [Link]
011-9924-9

Panitz, H.-J., Dosio, A., Büchner, M., Lüthi, D., Keuler, K., 2014. COSMO-CLM (CCLM) climate
simulations over CORDEX-Africa domain: analysis of the ERA-Interim driven simulations at 0.44°
and 0.22° resolution. Clim. Dyn. 42, 3015–3038. [Link]

Pettitt, A.N., 1979. A Non-Parametric Approach to the Change-Point Problem. J. R. Stat. Soc. Ser. C

153
Appl. Stat. 28, 126–135. [Link]

Piani, C., Weedon, G.P., Best, M., Gomes, S.M., Viterbo, P., Hagemann, S., Haerter, J.O., 2010.
Statistical bias correction of global simulated daily precipitation and temperature for the application of
hydrological models. J. Hydrol. 395, 199–215. [Link]

Piao, S., Ciais, P., Huang, Y., Shen, Z., Peng, S., Li, J., Zhou, L., Liu, H., Ma, Y., Ding, Y.,
Friedlingstein, P., Liu, C., Tan, K., Yu, Y., Zhang, T., Fang, J., 2010. The impacts of climate change
on water resources and agriculture in China. Nature 467, 43–51. [Link]

Qaiser, K., Ahmad, S., Johnson, W., Batista, J., 2011. Evaluating the impact of water conservation on
fate of outdoor water use: A study in an arid region. J. Environ. Manage. 2061–2068.
[Link]

Raje, D., Mujumdar, P.P., 2010. Reservoir performance under uncertainty in hydrologic impacts of
climate change. Adv. Water Resour. 33, 312–326. [Link]

Rani, D., Moreira, M.M., 2010. Simulation–Optimization Modeling: A Survey and Potential
Application in Reservoir Systems Operation. Water Resour. Manag. 24, 1107–1138.
[Link]

Riahi, K., Rao, S., Krey, V., Cho, C., Chirkov, V., Fischer, G., Kindermann, G., Nakicenovic, N., Rafaj,
P., 2011. RCP 8.5—A scenario of comparatively high greenhouse gas emissions. Clim. Change 109,
33. [Link]

Richter, B.., Baumgartner, J.., Braun, D.., Powell, J., 1998. A spatial assessment of hydrologic
alteration within a river network. Regulated Rivers. Res. Manag. 14, 329–340.

Richter, B.., Baumgartner, J.., Powell, J., Braun, D.., 1996. A method for assessing hydrologic
alteration within ecosystems. Conserv. Biol. 10, 1163–1174.

Richter, B.., Baumgartner, J.., Wigington, R., Braun, D.., 1997. How much water does a river need?
Freshw. Biol. 37, 231–249.

Rijsberman, G., 2006. Water scarcity: fact or fiction? Agric. Water Manag. 80, 5–22.

Rojas, R., Feyen, L., Dosio, A., Bavera, D., 2011. Improving Pan-European hydrological simulation of
extreme events through statistical bias correction of RCM-driven climate simulations. Hydrol Earth
Syst Sci 15, 2599–2620. [Link]

154
Ruti, P.., Williams, J.., Yang, X., Mari, C., Traore, A.., 2011. The West African climate system: a
review of the AMMA model inter-comparison initiatives. Atmos Sci Lett 12, 116–122.

Schuol, J., Abbaspour, K.C., Srinivasan, R., Yang, H., 2008. Estimation of freshwater availability in
the West African sub-continent using the SWAT hydrologic model. J. Hydrol. 352, 30–49.
[Link]

Seleshi, Y., Zanke, U., 2004. Recent Changes in Rainfall and Rainy Days in Ethiopia.
[Link]

Sen, P.K., 1968. Estimates of the Regression Coefficient Based on Kendall’s Tau. J. Am. Stat. Assoc.
63, 1379–1389. [Link]

Serdeczny, O., Adams, S., Baarsch, F., Coumou, D., Robinson, A., Hare, W., Schaeffer, M., Perrette,
M., Reinhardt, J., 2017. Climate change impacts in Sub-Saharan Africa: from physical changes to their
social repercussions. Reg. Environ. Change 17, 1585–1600. [Link]
2

Setegn, S., Srinivasan, R., Dargahi, B., Melesse, A., 2009. Spatial delineation of soil erosion
vulnerability in the Lake Tana Basin, Ethiopia. Hydrol. Process. 23, 3738–3750.
[Link]

Setegn, S.G., Dargahi, B., Srinivasan, R., Melesse, A.M., 2010a. Modeling of Sediment Yield
FromAnjeni-Gauged Watershed, Ethiopia Using SWAT Model1. JAWRA J. Am. Water Resour.
Assoc. 46, 514–526. [Link]

Setegn, S.G., Melesse, A.M., Rayner, D., Dargahi, B., 2014. Climate Change Impact on Water
Resources and Adaptation Strategies in the Blue Nile River Basin, in: Nile River Basin. Springer,
Cham, pp. 389–404. [Link]

Setegn, S.G., Rayner, D., Melesse, A.M., Dargahi, B., Srinivasan, R., 2011. Impact of climate change
on the hydro-climatology of Lake Tana Basin, Ethiopia. Water Resour. Res. 47, W04511.
[Link]

Setegn, S.G., Srinivasan, R., Melesse, A.M., Dargahi, B., 2010b. SWAT model application and
prediction uncertainty analysis in the Lake Tana Basin, Ethiopia. Hydrol. Process. 24, 357–367.
[Link]

155
Shifteh Some’e, B., Ezani, A., Tabari, H., 2012. Spatiotemporal trends and change point of precipitation
in Iran. Atmospheric Res. 113, 1–12. [Link]

Smadi, M.M., Zghoul, A., Smadi, M.M., Zghoul, A., 2006. A Sudden Change In Rainfall
Characteristics In Amman, Jordan During The Mid-1950s. Am. J. Environ. Sci. 2, 84–91.
[Link]

Sonneveld, B.., Keyzer, M.., 2003. Land under Pressure: soil conservation concerns and opportunities
for Ethiopia. Land Degrad. Dev. 14, 5–23.

Stocker, D. Qin, G.-K, Plattner, L.V., S.K. Allen, 2013. Technical summary. 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. Cambridge University Press, Cambridge, England, pp.
31–116. [Link]

Swain, A., 1997. “The Nile River Dispute: Ethiopia, the Sudan, and Egypt”, The Journal of Modern
African Studies (Cambridge), vol. 35, no. 4, December 1997, pp. 675-694. J. Mod. Afr. Stud. 35, 675–
694.

Tabari, H., Taye, M.T., Willems, P., 2015. Statistical assessment of precipitation trends in the upper
Blue Nile River basin. Stoch. Environ. Res. Risk Assess. 29, 1751–1761.
[Link]

Tarekegn, D., Tadege, A., 2006. Assessing the Impact of Climate Change on the Water Resources of
the Lake Tana Basin Using the Watbal Model. Discuss Pap 30.

Taye, M.T., Ntegeka, V., Ogiramoi, N.P., Willems, P., 2011. Assessment of climate change impact on
hydrological extremes in two source regions of the Nile River Basin. Hydrol Earth Syst Sci 15, 209–
222. [Link]

Taye, M.T., Willems, P., 2012. Temporal variability of hydroclimatic extremes in the Blue Nile basin.
Water Resour. Res. 48, W03513. [Link]

Taylor, K.E., Stouffer, R.J., Meehl, G.A., 2011a. An Overview of CMIP5 and the Experiment Design.
Bull. Am. Meteorol. Soc. 93, 485–498. [Link]

Taylor, K.E., Stouffer, R.J., Meehl, G.A., 2011b. An Overview of CMIP5 and the Experiment Design.
Bull. Am. Meteorol. Soc. 93, 485–498. [Link]

156
Tekleab, S., Uhlenbrook, S., Mohamed, Y., avenije, H.H.., Temesgen, M., Wenninger, J., 2011. Water
balance modeling of Upper Blue Nile catchments using a top-down approach. Hydrol Earth Syst Sci
15, 2179–2193. [Link] 2011

Tesemma, Z.K., Mohamed, Y.A., Steenhuis, T.S., 2010. Trends in rainfall and runoff in the Blue Nile
Basin: 1964–2003. Hydrol. Process. 24, 3747–3758. [Link]

Tesfaye, G., Mohamed, Y., van der Zaag, P., Yazew, E., 2017. Temporal and spatial changes of rainfall
and streamflow in the Upper Tekezē–Atbara river basin, Ethiopia. [Link]
2127-2017

Teutschbein, C., Seibert, J., 2012. Bias correction of regional climate model simulations for
hydrological climate-change impact studies: Review and evaluation of different methods. J. Hydrol.
456–457, 12–29. [Link]

Teutschbein, C., Seibert, J., 2010. Regional Climate Models for Hydrological Impact Studies at the
Catchment Scale: A Review of Recent Modeling Strategies. Geogr. Compass 4, 834–860.
[Link]

Themeßl, M.J., Gobiet, A., Heinrich, G., 2012. Empirical-statistical downscaling and error correction
of regional climate models and its impact on the climate change signal. Clim. Change 112, 449–468.
[Link]

Thomson, A., V. Calvin, K., J. Smith, S., Kyle, P., Volke, A., Patel, P., Delgado Arias, S., Bond-
Lamberty, B., A. Wise, M., E. Clarke, L., A. Edmonds, J., 2011. RCP4.5: A pathway for stabilization
of radiative forcing by 2100. Clim. Change 109, 77–94. [Link]

Tibebe, D., Bewket, W., 2011. Surface runoff and soil erosion estimation using the SWAT model in
the Keleta Watershed, Ethiopia. Land Degrad. Dev. 22, 551–564. [Link]

Tilahun, S.., Guzman, C.., Zegeye, A.., Engda, T.., Collick, A.., ., Rimmer, A., Steenhuis, T.., 2013.
An efficient semi-distributed hillslope erosion model for the subhumid Ethiopian Highlands. Hydrol.
Earth Syst. Sci. 17, 1051–1063.

United States. Soil Conservation Service., 1972. SCS national engineering handbook. U.S. Dept. of
Agriculture, Soil Conservation Service, Washington, D.C.:

Uniyal, B., Jha, M.K., Verma, A.K., 2015. Assessing Climate Change Impact on Water Balance

157
Components of a River Basin Using SWAT Model. Water Resour. Manag. 29, 4767–4785.
[Link]

USACE, 1992. Developing Operation Plans from HEC Perspective Reservoir Model results for the
Missouri River System: Preliminary Results. Hydrologic Engineering Center, Davis, USA.

van Griensven, A., Ndomba, P., Yalew, S., Kilonzo, F., 2012. Critical review of SWAT applications in
the upper Nile basin countries. Hydrol Earth Syst Sci 16, 3371–3381. [Link]
3371-2012

Vedwan, N., Ahmad, S., Miralles-Wilhelm, F., Broad, K., Letson, D., Podesta, G., 2008. Institutional
Evolution in Lake Okeechobee Management in Florida: Characteristics, Impacts, and Limitations.
Water Resour. Manag. 22, 699–718. [Link]

Vicuña, S., Garreaud, R.D., McPhee, J., 2011. Climate change impacts on the hydrology of a snowmelt
driven basin in semiarid Chile. Clim. Change 105, 469–488. [Link]
4

Villarini, G., Smith, J.A., Ntelekos, A.A., Schwarz, U., 2011. Annual maximum and peaks-over-
threshold analyses of daily rainfall accumulations for Austria. J. Geophys. Res. Atmospheres 116,
D05103. [Link]

Vonk, E., xu, Y., Booij, M., Zhang, X., Augustijn, D.C.M., 2014. Adapting Multireservoir Operation
to Shifting Patterns of Water Supply and Demand. Water Resour. Manag. 28, 625–643.
[Link]

Vuuren, D.P. van, Edmonds, J., Kainuma, M., Riahi, K., Thomson, A., Hibbard, K., Hurtt, G.C., Kram,
T., Krey, V., Lamarque, J.-F., Masui, T., Meinshausen, M., Nakicenovic, N., Smith, S.J., Rose, S.K.,
2011. The representative concentration pathways: an overview. Clim. Change 109, 5.
[Link]

Wang, S. Shang, W. Yang, A. M. Melesse, 2008a. Simulation of an Agricultural Watershed Using an

Improved Curve Number Method in SWAT. Trans. ASABE 51, 1323–1339.
[Link]

Wang, W. Yang, A. M. Melesse, 2008b. Using Hydrologic Equivalent Wetland Concept Within SWAT
to Estimate Streamflow in Watersheds with Numerous Wetlands. Trans. ASABE 51, 55–72.
[Link]

158
Wang, X., Melesse, A., Yang, W., 2006. Influences of Potential Evapotranspiration Estimation
Methods on SWAT’s Hydrologic Simulation in a Northwestern Minnesota Watershed. Trans. ASABE
Am. Soc. Agric. Biol. Eng. 49, 1755–1771. [Link]

Wang, X., Melesse, A.M., 2006. Effects of Statsgo and Ssurgo as Inputs on Swat Model’s Snowmelt
Simulation1. JAWRA J. Am. Water Resour. Assoc. 42, 1217–1236. [Link]
1688.2006.tb05296.x

Wang, Y., Rhoads, B.L., Wang, D., 2016. Assessment of the flows regime alterations in the middle
reach of the Yangtze River associated with dam construction: potential ecological implication. Hydrol
Process 30, 3949–3966. [Link] 10.1002/hyp.10921

WAPCOS, 1990. Preliminary Resources Development Master plan for Ethiopia, MoWR, Addis Ababa,
ethiopia. MoWR, Addis Ababa, Ethiopia.

Wilby, R.L., Harris, I., 2006. A framework for assessing uncertainties in climate change impacts: Low-
flow scenarios for the River Thames, UK. Water Resour. Res. 42, W02419.
[Link]

Wolde, E., 2016. Identification and prioritization of sub watersheds for land and water management in
Tekeze dam watershed, Northern Ethiopia. Int. Soil Water Conserv. Res. 4, 30–38.

Wu, G., Li, L., Ahmad, S., Chen, X., Pan, X., 2013. A Dynamic Model for Vulnerability Assessment
of Regional Water Resources in Arid Areas: A Case Study of Bayingolin, China. Water Resour. Manag.
3085–3101. [Link]

Wurbs, R.A., 2005. Comparative Evaluation of Generalized River/Reservoir System Models

(Technical Report). Texas Water Resources Institute.

Wurbs Ralph A., 1993. Reservoir‐System Simulation and Optimization Models. J. Water Resour. Plan.
Manag. 119, 455–472. [Link]

Xu, C., 2000. Modelling the Effects of Climate Change on Water Resources in Central Sweden. Water
Resour. Manag. 14, 177–189. [Link]

Xu, C., Widén, E., Halldin, S., 2005. Modelling hydrological consequences of climate change—
Progress and challenges. Adv. Atmospheric Sci. 22, 789–797. [Link]

Yamba, F.D., Walimwipi, H., Jain, S., Zhou, P., Cuamba, B., Mzezewa, C., 2011. Climate

159
change/variability implications on hydroelectricity generation in the Zambezi River Basin. Mitig.
Adapt. Strateg. Glob. Change 16, 617–628. [Link]

Yazdi, J., Moridi, A., 2017. Interactive Reservoir-Watershed Modeling Framework for Integrated
Water Quality Management. Water Resour. Manag. 31, 2105–2125. [Link]
017-1627-4

Yeh, W.W.-G., 1985. Reservoir Management and Operations Models: A State-of-the-Art Review.
Water Resour. Res. 21, 1797–1818. [Link]

Yesuf, H.M., Assen, M., Alamirew, T., Melesse, A.M., 2015. Modeling of sediment yield in Maybar
gauged watershed using SWAT, northeast Ethiopia. CATENA 127, 191–205.
[Link]

Yin, X.A., Yang, Z.F., 2011. Development of a coupled reservoir operation and water diversion model:
Balancing human and environmental flow requirements. Ecol. Model. 222, 224–231.
[Link]

Yu, C., Yin, X., Yang, Z., 2016. A revised range of variability approach for the comprehensive
assessment of the alteration of flow regime. Ecol. Eng. 96, 200–207.

Zhang, C., Zhu, X., Fu, G., Zhou, H., Wang, H., 2014. The impacts of climate change on water
diversion strategies for a water deficit reservoir. J. Hydroinformatics 16, 872–889.
[Link]

Zhang, Q., Xiao, M., Liu, C.-L., Singh, V.P., 2014. Reservoir induced hydrological alterations and
environmental flow variation in the East River, the Pearl River basin, China. StochEnv. Res Risk Assess
28, 2119–2131. [Link] 10.1007/s00477-014-0893-4

Zhou, Y., Guo, S., Duan, W., Chen, H., Liu, P., 2015. Dynamic control of flood limited water level for
cascade reservoirs. ShuiliFadianXuebaoJournalHydroelectr. Eng. 34, 23–30.

Zhou, Y., Guo, S., Liu, P., Xu, C., 2014. Joint operation and dynamic control of flood limiting water
levels for mixed cascade reservoir systems. J. Hydrol. 519, 248–257.
[Link]

Zhu, X., Zhang, C., Qi, W., Cai, W., Zhao, X., Wang, X., 2018. Multiple Climate Change Scenarios
and Runoff Response in Biliu River. Water 10.

160
 APPENDICES
Appendix A: Meteorological data of Tekeze River Basin
Appendix A1: Monthly rainfall data (mm) in some Tekeze Basin weather stations after missing data
filled
Year/Month Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
a) Lalibela Station
1994 0.00 3.10 8.40 46.50 3.40 41.60 446.00 311.00 65.00 3.70 1.80 0.90
1995 0.00 28.60 10.20 83.20 68.00 9.30 323.40 205.70 46.80 0.70 0.00 20.70
1996 31.90 1.80 118.70 47.00 56.20 161.60 296.50 288.40 20.50 0.70 35.00 0.80
1997 15.70 8.40 95.60 58.70 24.30 104.40 304.20 162.00 25.80 100.30 100.90 2.60
1998 5.90 7.70 37.40 10.80 44.00 17.60 340.80 258.60 58.10 21.50 0.00 0.00
1999 19.30 0.00 0.00 22.30 0.70 34.20 319.70 319.60 50.40 35.50 3.30 1.80
2000 0.00 0.00 25.90 79.90 13.50 16.20 213.10 206.30 71.10 80.70 36.60 9.30
2001 0.00 9.80 82.10 31.40 1.90 106.80 340.40 377.30 7.00 0.00 0.00 7.10
2002 34.40 19.40 45.60 34.70 6.80 51.80 214.30 245.10 55.70 1.50 12.80 0.00
2003 2.00 19.70 45.50 56.80 1.80 59.30 225.30 404.10 52.70 0.00 1.60 0.00
2004 2.80 11.60 12.00 20.30 0.00 45.50 242.40 188.60 19.00 10.20 0.80 0.00
2005 6.20 21.40 47.90 27.10 57.00 36.40 366.20 155.70 40.20 0.50 0.00 0.00
2006 0.00 1.00 54.40 43.80 20.30 23.90 301.90 324.30 43.60 25.60 23.60 15.30
2007 2.00 19.70 45.50 56.80 1.80 59.30 225.30 404.10 52.70 0.00 1.60 0.00
2008 2.80 11.60 12.00 20.30 0.00 1.52 242.40 188.60 19.00 10.20 0.80 0.00
b) Gondar Station
1994 0.00 0.00 0.00 5.80 84.50 156.00 289.10 272.90 125.00 38.00 20.00 2.80
1995 0.00 0.00 34.50 36.70 86.50 107.80 284.50 308.60 86.90 12.40 0.40 19.80
1996 0.00 4.40 22.20 83.60 183.80 194.70 249.30 290.00 75.80 67.70 23.20 0.40
1997 0.00 1.80 28.20 42.80 124.20 184.80 239.70 230.40 33.10 200.30 40.20 13.70
1998 0.00 0.00 13.70 3.70 88.50 284.60 383.00 487.90 125.70 126.40 4.80 0.00
1999 35.50 0.00 0.00 42.00 127.10 158.70 432.80 424.40 187.40 337.80 11.30 52.60
2000 0.00 1.40 3.90 73.20 60.70 364.30 451.40 368.60 166.60 268.70 1.90 0.00
2001 0.00 0.90 3.30 29.40 88.90 415.40 568.90 491.30 118.20 144.80 16.00 0.00
2002 0.00 0.00 0.00 16.60 87.10 197.40 312.70 247.60 76.80 45.20 5.80 4.20
2003 0.00 23.50 9.70 0.60 37.30 258.40 311.70 291.60 116.10 21.30 0.00 5.70
2004 1.60 3.70 5.90 37.60 1.40 181.40 378.30 312.30 112.40 67.60 65.70 0.00
2005 0.00 11.20 60.80 12.10 24.20 137.50 289.00 276.20 169.30 42.80 17.40 0.00
2006 0.00 0.00 10.80 27.80 152.60 98.70 291.50 339.10 192.50 86.70 4.20 35.00
2007 22.90 0.50 22.20 87.80 65.60 285.60 422.80 457.80 172.50 8.10 5.00 0.00
2008 81.40 0.00 1.50 81.90 211.50 6.98 376.40 331.80 228.60 51.80 2.50 18.50
c) Debretabor Station
1994 1.80 9.10 0.00 14.40 89.70 237.70 495.80 633.50 248.00 10.00 24.80 31.70
1995 0.00 0.00 27.60 31.40 98.70 75.70 411.50 405.90 168.10 20.70 7.70 25.30
1996 4.30 1.20 47.80 92.10 138.70 287.00 343.10 374.00 155.40 30.60 76.20 4.40
1997 3.40 0.00 73.70 43.10 197.30 225.10 449.70 359.00 197.00 314.00 12.30 137.60
1998 13.60 0.00 28.10 6.90 203.60 126.20 400.60 410.80 214.80 75.90 0.20 0.00
1999 34.50 0.00 0.00 16.70 44.60 200.10 453.70 345.70 245.00 250.40 11.20 19.50
2000 0.00 0.30 6.30 116.10 61.10 168.10 423.40 466.00 232.30 137.80 34.80 0.50
2001 0.00 1.30 17.20 24.00 95.40 197.50 496.70 410.00 184.80 60.10 4.50 7.20
2002 0.40 0.00 61.20 45.10 46.80 203.30 256.60 313.40 132.80 2.90 16.00 18.80
2003 0.00 13.90 33.70 19.20 9.80 97.00 438.40 404.30 200.70 16.70 33.30 14.80
2004 0.50 37.60 33.70 75.50 19.10 141.00 333.70 295.20 120.80 85.80 42.50 12.70
2005 1.30 0.00 34.10 10.30 56.30 224.40 473.60 436.00 216.20 5.00 29.70 0.00
2006 0.00 1.40 6.80 63.20 147.30 170.00 482.20 452.50 255.00 47.50 0.00 13.60
2007 22.90 0.50 22.20 87.80 65.60 295.80 422.80 449.60 172.50 8.10 5.00 0.00
2008 81.40 0.00 1.50 81.90 211.50 6.98 376.40 331.80 228.60 51.80 2.50 18.50
d) Mekele Station
1994 0.00 5.30 0.40 43.80 0.80 67.60 147.90 317.80 70.10 0.00 1.80 2.00
1995 0.00 5.90 31.20 29.20 27.10 8.20 267.50 249.90 38.50 3.00 0.00 2.70
1996 1.40 0.00 59.50 12.50 92.20 47.90 109.20 224.00 7.10 0.00 31.40 1.10
1997 0.00 0.00 20.40 32.60 29.80 32.40 243.10 100.50 16.30 59.90 15.70 0.00
1998 10.00 1.20 10.60 10.60 22.00 48.00 289.00 318.80 31.70 22.00 0.00 0.00
1999 22.00 0.30 10.90 0.00 0.00 8.20 292.80 359.20 22.80 0.90 0.00 0.00
2000 0.00 0.00 0.00 10.40 24.60 5.40 201.40 182.00 15.80 2.20 10.30 3.50
2001 0.00 0.00 38.10 18.70 8.70 65.50 267.90 226.30 9.20 2.90 0.00 0.00
2002 12.90 0.00 35.50 4.20 23.00 60.80 95.50 208.60 28.00 0.00 0.00 0.30
2003 0.00 25.90 18.20 8.40 35.20 101.10 127.80 186.00 23.40 0.70 0.00 0.10
2004 10.90 0.20 35.20 20.50 7.10 25.40 64.30 221.10 1.40 3.10 0.80 0.00
2005 0.00 1.40 15.60 48.90 55.10 18.20 110.50 314.00 34.30 0.00 1.30 0.00
2006 0.00 0.00 31.30 117.60 46.30 38.10 187.10 298.90 23.60 12.00 0.00 0.30
2007 1.10 2.30 11.20 34.50 22.20 58.00 278.70 135.10 76.20 0.00 0.00 0.00
2008 7.50 0.00 0.00 23.90 5.90 0.43 94.10 103.30 27.40 7.70 4.10 0.00

161
Appendix B: Tekeze Hydropower Reservoir data

Appendix B1: Observed mean annual Tekeze hydropower reservoir pool level

Appendix B2: Observed mean monthlyTekeze hydropower reservoir pool level (2009-2017)

162
Appendix B3: Observed mean monthlyTekeze hydropower reservoir storage (2009-2017)

Appendix B4: Characteristics of Tekeze hydropower reservoir/Arch Dam (TK05)

Type Characteristics Unit Value

Catchment Area Km2 29,692
Annual inflow 109m3 3.75
Catchment Annual precipitation Mm 850
Annual discharge at dam site m3/s 120
Sedimentation 106m3/year 30
Dam crest level Masl 1145
Dam height M 188
Dam features Dam type Type Arch dam
Commissioning year Year 2009
Dam crest length M 420
Maximum retention level Masl 1140
Minimum operation level Masl 1096
Reservoir Total storage 109m3 9.293
Live storage 109m3 5.293
Dead Storage 109m3 4
Surface area at MRL km2 147
Powerhouse type Under ground
Turbine number and type No 4-Francis
Power plant Total Installed capacity MW 300
Maximum net head M 162.8
Minimum net head M 120
Maximum Discharge m3/s 220

163
Appendix B4: Projected monthlyTekeze hydropower reservoir inflow under RCP4.5 scenario

164
Appendix B5: Projected monthlyTekeze hydropower reservoir inflow under RCP8.5 scenario

165
Appendix C: Climate and hydrological variability

Appendix C1: Hydrological alteration analysis by RVA-Non parametric test of Tekeze river flow
regime due to dam construction

166
Appendix C2: Minimum and maximum 1-, 3-, 7-, 30- and 90- days flow at Embamadre using IHA

167
Appendix D: SWAT Hydrological Model parameters

Appendix D1: Sensitive parameters for calibration of the SWAT model at Tekeze basin

168
Appendix D2: Land use and soil distribution of Tekeze basin re-classified by SWAT

Appendix E: Climate change in Tekeze basin

Appendix E1: annual and seasonal changes of precipitation at different stations of Tekeze basin in
different scenarios

Change in precipitation (mm)

Station Season RCP45 RCP85
2020s 2050s 2080s 2020s 2050s 2080s
Annual 1185.08 1250.80 1420.56 1408.78 1393.87 1119.64
Belg 9.36 98.39 63.22 66.61 35.10 34.36
Lalibela
Kiremt 612.26 1100.35 1240.98 1292.70 1242.75 978.91
Bega 14.81 52.06 116.36 49.47 116.02 106.37
Annual 430.37 449.71 451.69 456.88 501.68 524.63
Belg -13.87 41.86 -2.37 0.68 -1.59 37.10
Gonder
Kiremt 259.14 432.13 465.68 478.42 516.67 449.02
Bega 46.47 -24.28 -11.61 -22.23 -13.40 38.51
Annual 192.07 214.30 250.06 255.13 265.76 261.55
Belg -37.07 7.17 -23.90 -26.24 -28.53 5.47
Debre Tabor
Kiremt 91.87 235.33 300.71 318.27 321.60 233.39
Bega 137.27 -28.20 -26.75 -36.90 -27.31 22.69
Annual 1303.84 1369.56 1539.32 1527.54 1512.63 1198.40
Belg -72.83 16.20 -18.97 -15.58 -47.08 -87.82
Maichew
Kiremt 858.23 1346.32 1486.95 1538.68 1488.72 1224.89
Bega 18.53 7.03 71.34 4.44 70.99 61.34
Annual -144.09 -162.18 -119.03 -85.41 -103.60 -189.87
Belg -74.19 -64.19 -73.04 -71.18 -73.35 -79.95
Mekele
Kiremt -186.21 -103.14 -61.11 -17.02 -41.09 -120.26
Bega 116.31 5.15 15.11 2.79 10.83 10.33

169
Appendix E2: annual and seasonal changes of temperature at different stations of Tekeze basin in
different scenarios

Change in Temprature (oC)

Change in Minimum Temprature Change in Maximum Temprature
Station Season
RCP45 RCP85 RCP45 RCP85
2020s 2050s 2080s 2020s 2050s 2080s 2020s 2050s 2080s 2020s 2050s 2080s
Annual 0.04 0.90 1.26 0.23 1.78 3.97 0.49 1.42 1.64 0.75 2.14 4.43
Belg -0.51 0.28 0.60 -0.38 1.27 3.48 1.08 2.24 2.42 1.54 3.06 5.25
Lalibela
Kiremt 0.24 1.11 1.51 0.49 1.97 4.10 0.17 1.20 1.37 0.29 1.83 4.59
Bega 0.40 1.30 1.67 0.58 2.10 4.32 0.21 0.82 1.14 0.42 1.54 3.44
Annual -1.08 -0.18 0.22 -0.89 0.75 3.01 -1.84 -0.93 -0.65 -1.61 -0.18 -0.65
Belg -1.64 -0.66 -0.45 -1.39 0.24 2.47 -2.31 -1.22 -0.95 -1.87 -0.36 -0.95
Gonder
Kiremt -0.32 0.59 1.03 -0.05 1.46 3.71 0.07 1.07 1.29 0.12 1.64 1.29
Bega -1.28 -0.47 0.08 -1.25 0.54 2.85 -3.28 -2.63 -2.29 -3.09 -1.81 -2.29
Annual 2.74 3.65 4.00 2.86 4.58 6.85 1.34 2.24 2.49 1.54 3.00 5.27
Belg 2.49 3.47 3.69 2.66 4.46 6.70 1.20 2.29 2.55 1.62 3.15 5.33
DebreTabor
Kiremt 3.17 4.11 4.41 3.38 4.93 7.10 2.11 3.12 3.22 2.12 3.70 6.45
Bega 2.55 3.37 3.90 2.53 4.37 6.76 0.70 1.31 1.70 0.88 2.15 4.04
Annual 1.06 1.92 2.28 1.06 1.92 2.28 -1.95 -0.74 -0.80 -1.95 -0.74 -0.80
Belg 0.76 1.55 1.87 0.76 1.55 1.87 -1.21 -0.05 0.12 -1.21 -0.05 0.12
Maichew
Kiremt 0.53 1.40 1.80 0.53 1.40 1.80 -2.87 -1.84 -1.67 -2.87 -1.84 -1.67
Bega 1.90 2.80 3.16 1.90 2.80 3.16 -1.78 -0.31 -0.85 -1.78 -0.31 -0.85
Annual 0.30 1.58 1.58 0.44 2.11 4.47 1.05 2.00 2.21 1.21 2.66 4.98
Belg -0.74 0.41 0.41 -0.65 1.07 3.35 1.03 2.04 2.28 1.38 2.85 4.96
Mekele
Kiremt 1.41 2.79 2.79 1.65 3.35 5.71 1.91 3.01 3.15 1.91 3.50 6.41
Bega 0.23 1.54 1.54 0.33 1.92 4.35 0.21 0.95 1.20 0.34 1.62 3.57

170
Appendix E3: Comparison of mean annual precipitation on selected stations of Tekeze basin under
RCP 4.5 and RCP8.5 scenarios for 2020`s,2050`s and 2080`s as compared to the base period 1975-
2005.

Appendix E4: Comparison of mean annual maximum and minimum temperature at selected stations of
Tekeze basin under RCP 4.5 and RCP8.5 for 2020`s,2050`s and 2080`s as compared to the base period
1975-2005.

171
172
Appendix F: HEC-ResPRM Analysis Results

Appendix F1: HEC-ResPRM configuration map at optimization and network modules

173
Appendix F2: Monthly optimized storage and release at Tekeze hydropower reservoir at baseline period

Appendix F3: Reservoir pool level and release at Tekeze hydropower reservoir at baseline period

174
Appendix F4: Optimized storage versus release at Tekeze hydropower reservoir under RCP4.5 in 2020s

Appendix F5: Optimized pool level versus release at Tekeze hydropower reservoir under RCP4.5 in
2020s

175
Appendix F6: Optimized storage versus release at Tekeze hydropower reservoir under RCP8.5 in
2050s

Appendix F7: Optimized pool level versus release at Tekeze hydropower reservoir under RCP8.5 in
2050s

176
Appendix F8: Optimized storage versus release at Tekeze hydropower reservoir under RCP4.5 in
2050s

Appendix F9: Optimized pool level versus release at Tekeze hydropower reservoir under RCP4.5 in
2050s

177
Appendix F10: Optimized storage versus release at Tekeze hydropower reservoir under RCP4.5 in
2080s

Appendix F11: Optimized pool level versus release at Tekeze hydropower reservoir under RCP4.5 in
2080s

Appendix F12: Optimized Tekeze hydropower reservoir (TK05) pool level (m.a.s.l) using HEC-
ResPRM optimization model under climate change a) RCP4.5 climate scenario b) RCP8.5 climate
scenario

178
 a) RCP4.5 climate scenario (2050s)
Year/Month Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
2041 1096.60 1095.90 1095.90 1104.60 1101.80 1098.10 1102.60 1121.70 1131.00 1134.00 1132.90 1131.30
2042 1128.10 1123.60 1118.50 1113.10 1107.40 1104.00 1107.60 1121.30 1132.30 1133.00 1131.30 1127.80
2043 1123.20 1118.00 1112.50 1106.60 1101.10 1102.80 1115.60 1117.10 1131.30 1133.60 1132.50 1129.40
2044 1124.90 1120.00 1114.40 1110.80 1105.90 1114.80 1115.60 1128.20 1136.10 1136.20 1134.20 1130.70
2045 1126.10 1121.10 1115.60 1115.40 1110.70 1108.20 1110.50 1124.70 1133.50 1134.00 1131.90 1128.40
2046 1123.70 1119.30 1115.60 1111.60 1106.90 1102.30 1106.50 1124.00 1131.30 1133.50 1132.90 1130.20
2047 1125.90 1121.20 1115.60 1109.90 1103.30 1107.30 1115.60 1115.60 1131.30 1131.70 1131.30 1128.40
2048 1124.10 1120.90 1118.40 1115.60 1111.60 1106.50 1115.60 1115.60 1131.30 1131.60 1131.90 1129.10
2049 1125.10 1120.80 1115.60 1115.60 1111.50 1112.00 1117.50 1126.00 1133.30 1133.70 1131.80 1130.20
2050 1126.00 1121.10 1115.60 1112.60 1111.60 1107.20 1109.30 1115.60 1131.30 1134.10 1133.60 1132.20
2051 1128.70 1124.10 1119.20 1117.10 1115.60 1113.10 1112.90 1123.40 1130.80 1132.30 1130.80 1127.40
2052 1122.80 1117.60 1111.90 1112.90 1108.70 1108.00 1115.60 1115.60 1131.30 1131.80 1132.00 1129.80
2053 1125.80 1121.10 1115.60 1112.30 1110.20 1109.90 1115.60 1119.60 1131.00 1133.30 1132.80 1129.90
2054 1125.40 1120.80 1115.60 1111.10 1105.30 1104.80 1115.60 1115.60 1131.30 1126.10 1127.20 1125.20
2055 1121.50 1120.10 1116.20 1115.60 1125.60 1123.60 1115.60 1115.60 1131.30 1129.40 1132.60 1131.30
2056 1127.90 1123.50 1118.40 1114.30 1109.10 1108.70 1115.60 1120.00 1131.30 1134.40 1133.20 1130.20
2057 1125.70 1120.90 1115.60 1110.20 1103.80 1097.80 1115.60 1122.30 1131.60 1132.90 1131.30 1127.90
2058 1123.30 1118.20 1112.80 1109.00 1104.90 1107.70 1115.60 1124.20 1130.90 1132.60 1131.40 1128.40
2059 1124.80 1120.30 1116.20 1114.40 1109.70 1106.20 1115.60 1131.30 1133.60 1133.20 1131.10 1127.40
2060 1122.80 1117.70 1112.00 1105.70 1099.40 1104.50 1115.60 1126.70 1133.90 1133.60 1131.30 1127.40
2061 1122.60 1117.40 1114.00 1108.80 1115.30 1112.10 1115.60 1126.90 1136.70 1138.60 1137.10 1135.00
2062 1131.30 1126.60 1124.60 1123.50 1119.90 1115.50 1113.40 1122.00 1126.80 1126.30 1124.80 1122.10
2063 1117.90 1115.60 1113.20 1111.60 1110.40 1111.50 1114.20 1123.90 1132.40 1134.50 1132.90 1129.40
2064 1124.70 1120.20 1115.60 1110.20 1103.70 1096.90 1101.90 1115.70 1131.30 1133.50 1132.60 1129.90
2065 1125.70 1120.90 1115.60 1110.10 1103.50 1107.00 1115.60 1115.60 1135.90 1138.70 1137.90 1135.10
2066 1131.30 1127.80 1122.90 1118.90 1113.20 1107.50 1113.50 1125.40 1131.90 1131.30 1130.00 1131.20
2067 1130.00 1128.20 1125.20 1124.00 1121.20 1118.80 1115.60 1129.00 1127.90 1126.50 1124.50 1121.80
2068 1118.60 1112.90 1107.70 1104.10 1097.80 1095.90 1115.60 1115.60 1131.30 1132.80 1131.60 1128.60
2069 1124.10 1119.10 1115.60 1112.90 1110.10 1105.50 1115.60 1115.60 1131.30 1131.30 1131.70 1129.60
2070 1125.60 1120.90 1116.30 1111.20 1105.30 1100.10 1115.60 1130.50 1138.10 1140.00 1140.00 1134.70

179
 b) RCP8.5 climate scenario (2050s)
Year/Month Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
2041 1097.10 1097.10 1097.10 1095.90 1095.90 1115.60 1115.60 1131.30 1131.40 1132.10 1129.90
2042 1121.10 1115.60 1110.60 1104.80 1105.60 1115.60 1115.60 1127.90 1131.30 1130.30 1127.30
2043 1118.80 1117.60 1118.60 1116.60 1112.80 1115.60 1121.90 1131.90 1132.60 1131.30 1127.90
2044 1118.10 1112.40 1113.30 1108.80 1107.50 1115.60 1115.60 1131.30 1132.00 1131.30 1129.20
2045 1120.40 1116.70 1114.10 1110.70 1106.80 1115.60 1115.60 1131.30 1132.10 1132.30 1130.00
2046 1121.10 1115.60 1110.20 1104.80 1103.80 1115.60 1122.40 1131.30 1132.80 1131.40 1128.00
2047 1118.20 1112.50 1115.10 1111.50 1111.50 1115.60 1127.60 1131.30 1134.00 1132.50 1129.30
2048 1120.70 1115.60 1110.10 1103.60 1108.80 1115.60 1121.60 1132.10 1133.00 1131.30 1127.80
2049 1118.10 1112.40 1107.00 1101.30 1096.50 1115.60 1125.50 1133.40 1134.50 1133.20 1130.40
2050 1121.20 1115.60 1110.00 1103.50 1101.10 1114.60 1120.70 1131.40 1133.60 1133.10 1130.30
2051 1121.20 1115.60 1114.20 1109.80 1104.40 1103.30 1119.70 1132.80 1134.10 1132.50 1130.20
2052 1121.10 1115.60 1113.60 1108.60 1104.50 1115.00 1119.30 1131.30 1134.30 1134.90 1132.90
2053 1125.80 1121.60 1116.40 1110.80 1107.80 1110.60 1125.10 1131.90 1132.20 1130.30 1128.60
2054 1119.80 1114.20 1109.30 1105.80 1105.50 1114.70 1118.60 1128.90 1133.40 1132.10 1129.50
2055 1121.00 1115.60 1113.80 1110.00 1106.20 1115.60 1115.60 1131.30 1131.30 1131.30 1130.40
2056 1123.00 1117.90 1112.40 1107.30 1101.70 1115.60 1115.60 1131.30 1132.60 1133.20 1131.30
2057 1122.80 1118.90 1114.40 1110.50 1106.60 1113.30 1121.50 1132.40 1134.00 1132.30 1129.00
2058 1119.80 1117.10 1113.80 1109.00 1107.20 1111.40 1120.60 1131.30 1134.30 1133.60 1130.90
2059 1121.70 1116.40 1111.80 1105.80 1100.60 1108.60 1115.60 1131.30 1131.30 1132.70 1131.30
2060 1123.60 1118.50 1113.20 1108.10 1110.20 1115.60 1119.40 1133.30 1134.60 1133.20 1130.30
2061 1121.10 1115.60 1112.40 1107.00 1101.70 1101.50 1115.60 1124.70 1133.40 1133.40 1131.30
2062 1122.80 1117.60 1112.90 1107.70 1102.60 1109.30 1118.20 1131.30 1132.60 1136.20 1134.50
2063 1127.20 1122.40 1119.20 1116.30 1114.70 1110.40 1116.20 1123.60 1123.30 1129.10 1125.80
2064 1116.20 1112.50 1107.10 1101.00 1095.90 1113.50 1126.00 1137.20 1139.10 1138.00 1135.10
2065 1126.60 1121.60 1117.40 1111.60 1105.70 1110.50 1127.30 1131.90 1131.10 1128.30 1124.20
2066 1115.10 1110.30 1105.00 1099.00 1095.90 1115.20 1120.60 1131.30 1133.40 1131.90 1128.90
2067 1119.80 1114.10 1108.20 1102.10 1104.50 1115.60 1115.60 1131.30 1133.50 1132.60 1129.80
2068 1120.60 1115.60 1118.00 1114.60 1110.10 1108.20 1121.60 1131.30 1131.30 1130.10 1126.80
2069 1117.30 1111.70 1116.70 1115.60 1116.50 1115.60 1115.60 1131.30 1131.30 1134.40 1136.90
2070 1127.00 1122.10 1117.00 1111.40 1119.80 1126.10 1138.80 1140.00 1140.00 1137.20 1137.05

180
Appendix F13: Mean monthly reservoir power storage for future time periods (rule curve) under: (a)
RCP4.5 climate scenario; and (b) RCP8.5 climate scenario.

181