IDENTIFICATION OF BEST MANAGEMENT PRACTICE OPTION ON SEDIMENT

YIELD USING SWAT

(CASE STUDY IN GUMARA WATERSHED, ETHIOPIA)

MSc THESIS

GEBEYEHU AYSHESHIM

OCTOBER, 2015

ARBA MINCH, ETHIOPIA

IDENTIFICATION OF BEST MANAGEMENT PRACTICE OPTION ON SEDIMENT
YIELD USING SWAT
(CASE STUDY IN GUMARA WATERSHED, ETHIOPIA)

GEBEYEHU AYSHESHIM

A THESIS SUBMITED TO THE

DEPARTMENT OF HYDRAULIC ENGINEERING
INSTITUTE OFTECHNOLOGY, SCHOOL OF GRADUATE STUDIES
ARBA MINCH UNIVERSITY IN PARTIAL FULFILLMENT OF THE
REQUIREMENTS FOR THE
DEGREE OF MASTER OF SCIENCE IN HYDRAULIC ENGINEERING

OCTOBER, 2015
ARBA MINCH
Certification
I, the undersigned, certify that I read a thesis entitled “Identification of Best Management
Practice Options on Sediment yield using SWAT” Case study in Gumara Watershed,
Ethiopia” and hereby recommended for acceptance by Arba Minch University in partial
fulfillment of the requirements for the degree of Master of Science in Hydraulic Engineering.

____________________________

Dr. Ing. AdaneAbebe (AMU)

(SUPERVISOR)

Date ___________________

i
Declaration and Copyright
I, Gebeyehu Aysheshim Mengaw, declare that this thesis is my own original work and that it has
not been presented and will not be presented by me to any other University for similar or any
other degree award.

Signature: ____________

Date: _______________

ii
Approval Page
This thesis entitled with “Identification of Best Management Practice Options on Sediment
yield using SWAT “Case study in Gumara Watershed, Ethiopia has been approved by advisor,
co-advisor, examiner, chairman, department head, coordinator, and school of Graduate studies in
partial fulfillment of the requirement for the degree of Master of Science in Hydraulic Engineering
Department.

Submitted By: Signature Date

Gebeyehu Aysheshim ______________ _____________
Student
Date of defense: October

Members of approval committee Signature Date

Dr.Ing. Adane Abebe ______________ _____________

Major Advisor

Elias Gebeyehu (MSc.) ______________ _____________

Co-Advisor

Abdella K. (PHD) _____________ _____________

Internal Examiner

Sisay D. (PHD) ______________ _____________

External Examiner

Ayalkibet M. (Msc) ______________ _____________

Chair Person

Elias Gebeyehu (MSc.) ______________ _____________

Dep’t head

Demelash Wendimagegn (Msc) ______________ _____________

Coordinator
______________ _____________ ______________

SGS Coordinator

iii
Acknowledgements
First and foremost, I want to thank the almighty God and His Mother, Saint Virgin Mary, for giving
me the chance, strength and courage to continue my study and for all things done in my entire life.
I can say it is with the mercy of God and his holy mother that I succeeded in finishing this paper.

I would like to express my wholehearted gratitude to my advisors Dr. Ing. Adane Abebe (AMU)
for his support in supervising my thesis work, providing me with important references, and
encouragements since early age of the thesis work. I greatly acknowledge his admirable patience
for giving me a chance to fulfill and successes this thesis research. And also my Co-Advisor Elias
Gebeyehu (MSc) without them, this work wouldn’t have been realized.

I am thankful to my family (my mother (Melkam Mesfin) and my father, my sister husband
(Abebe), My brothers and sisters) for their encouragements, Especially, my appreciation goes to
my bioncy Tigist and also my friends for their support at all the time and for giving me a word of
confidence every once in a while, without their support the research work would impossible.

I would also like to express my appreciation to my employer Amhara National Regional State
water, irrigation and energy bureau for providing the chance to have this study and also for Arba
Minch University, I also extend my special thank for National Metrological Agency and Ministry
of Water Resource for providing me the necessary data.

Last but not least, I would like to thank my friends for their ultimate support and encouragement.

iv
ABSTRACT

The major goal of the present study was to identify the critical erosion areas of an agricultural
Gumara watershed and recommend the best management practices using a physical process based
watershed scale model, soil water assessment tool (SWAT). The semi-automated Sequential
Uncertainty Fitting (SUFI2) calibration process built in SWAT calibration and uncertainty
program (SWAT-CUP) were used to calibrate the model parameters using time series of flow and
sediment load data of 1994 to 2001 and validated with the observed data from years 2002 to 2004.
The performance of the model was evaluated using graphical methods to assess the capability of
the model in simulating the runoff and sediment yield for the study area. The coefficient of
determination (R2) and Nash Sutcliff efficiency (NSE) values for the daily flow calibration using
SUFI2 are 0.78 and 0.77 respectively for validation and it was 0.73 and 0.72. For monthly sediment
yield by using SUFI2 calibration technique the model evaluation coefficients R 2 and NSE for
calibration was computed as 0.76 and 0.74 respectively, for validation it was 0.71 and 0.70
respectively. The sensitivity analysis runoff and sediment producing parameters was also carried
out and discussed. This paper presents sediment yield simulations in the Gumara watershed under
different Best Management Practice (BMP) scenarios. Scenarios applied in this paper are (i) Base
scenario (ii) Applying (terraces) (iii) contouring (iv) introduce strip cropping. The scenario results
showed that applying terraces, contouring and introducing Strip cropping reduced sediment yields
both at the sub watershed and the watershed outlets. Considering the critical sub watershed for the
existing conditions scenario (Base scenario), the model result indicates that simulated annual
average sediment yield was 31.514 t/ha/yr. Depend on this from base scenario terraces are saving
29.45 tons/ha/year of soil loss. Contouring was found to reduce soil erosion by 20.54 tons/ha/year.
Strip cropping for agricultural fields in the watershed reduces erosion also by 18.75 tons/ha/year.
These results indicate that applying BMPs could be effective in reducing sediment transport for
sustainable water resources management in the watershed.

Keywords: Gumara watershed, Best Management Practice, Sediment, Scenario, SWAT model.

v
Table of Contents

Approval Page........................................................................................................................... iii

ABSTRACT ...............................................................................................................................v

List of Figure ............................................................................................................................ xi

List of Tables .......................................................................................................................... xiii

Acronyms and Abbreviations ...................................................................................................xiv

CHAPTER ONE .........................................................................................................................1

1. Introduction.............................................................................................................................1

1.1 Background .......................................................................................................................1

1.2 Statement of the Problem ...................................................................................................3

1.3 Research Questions ............................................................................................................4

1.4 Objectives of the Study ......................................................................................................4

1.4.1 General Objective ........................................................................................................4

1.4.2 Specific Objectives ......................................................................................................4

1.5 Significance of the Study ...................................................................................................5

1.6 Scope of the Study .............................................................................................................5

1.7 Thesis Outline ....................................................................................................................6

CHAPTER TWO ........................................................................................................................7

2. Literature Review ....................................................................................................................7

2.1 Soil Erosion and its Extent .................................................................................................7

vi
2.2 Soil Erosion Types .............................................................................................................8

2.2.1 Rain Splash Erosion ....................................................................................................8

2.2.2 Sheet Erosion ..............................................................................................................8

2.2.3 Rill Erosion .................................................................................................................8

2.2.4 Gully Erosion ..............................................................................................................9

2.3 Factors Affecting Soil Erosion by Water ............................................................................9

2.3.1 Climatic Factors ..........................................................................................................9

2.3.2 Soil Factors ............................................................................................................... 10

2.3.3 Topographic Factors .................................................................................................. 10

2.3.4 Ground Cover Factors ............................................................................................... 11

2.3.5 Conservation Practice Factors.................................................................................... 11

2.4 Soil Erosion Models ......................................................................................................... 11

2.5 Hydrological Models ....................................................................................................... 13

2.5.1 Hydrological Model Selection Criteria ...................................................................... 14

2.6 Introduction to SWAT Model .......................................................................................... 16

2.6.1 Hydrological Component of SWAT........................................................................... 17

2.6.2 Sediment Component of SWAT ................................................................................ 22

2.6.3 SWAT Model Application Worldwide ...................................................................... 27

2.6.4 SWAT Model Application in Ethiopia ....................................................................... 28

2.6.5 SWAT-CUP .............................................................................................................. 29

2.7 Previous Study ................................................................................................................. 29

vii
CHAPTER THREE................................................................................................................... 30

3. Materials and Methods .......................................................................................................... 30

3.1 General ............................................................................................................................ 30

3.2 Study Area ....................................................................................................................... 30

3.2.1 Location .................................................................................................................... 30

3.2.2 Topography ............................................................................................................... 31

3.2.3 Climate...................................................................................................................... 31

3.2.4 Land Use and Land Cover ......................................................................................... 34

3.2.5 Soil............................................................................................................................ 34

3.2.6 Hydrology ................................................................................................................. 36

3.3 Materials and data used .................................................................................................... 37

3.4 General approach of Method ............................................................................................ 37

3.4.1 Scenario development ............................................................................................... 38

3.5 Data Availability and Analysis ......................................................................................... 40

3.5.1 Spatial Data ............................................................................................................... 40

3.5.2 Hydro-Meteorological Data ....................................................................................... 40

3.5.3 Filling Missing Weather Data .................................................................................... 41

3.5.4 Checking Data Consistency of the Stations ................................................................ 41

3.5.5 Thiessen Polygon Method ......................................................................................... 42

3.5.6 Stream Flow Availability, and Homogeneity Test ...................................................... 44

3.6 Model Input ..................................................................................................................... 46

viii
3.6.1 Digital Elevation Model ............................................................................................ 46

3.6.2 Land Use Map ........................................................................................................... 47

3.6.3 Soil Map and Data ..................................................................................................... 48

3.6.4 Weather Data ............................................................................................................ 48

3.6.5 River Discharge and Suspended sediment data .......................................................... 48

3.7 Model Setup .................................................................................................................... 49

3.7.1 Watershed Delineation .............................................................................................. 49

3.7.2 HRU Definition ......................................................................................................... 50

3.7.3 Weather Data Definition ............................................................................................ 52

3.7.4 Identification of Critical Sub Watersheds................................................................... 53

3.7.5 Calibration and Validation Setup and Analysis .......................................................... 53

3.7.6 Management Practices Scenario Development ........................................................... 57

CHAPTER FOUR ..................................................................................................................... 61

4. Result and Discussions .......................................................................................................... 61

4.1 Sensitivity Analysis ......................................................................................................... 61

4.2 Model Calibration and Uncertainty Measures ................................................................... 64

4.3 Spatial Variation of Soil Erosion ...................................................................................... 68

4.4 Analysis of Results .......................................................................................................... 68

4.4.1 Scenarios ................................................................................................................... 68

4.4.2 Identification and prioritization of Critical Sub Watersheds ....................................... 69

CHAPTER FIVE ...................................................................................................................... 71

ix
5. Conclusions and Recommendations....................................................................................... 71

5.1 Conclusion ....................................................................................................................... 71

5.2 Recommendations ............................................................................................................ 73

References ................................................................................................................................ 74

Appendix .................................................................................................................................. 80

Appendix I. Symbols and description of Weather Generator Parameters (WGEN) used by the
SWAT Model ........................................................................................................................ 80

Appendix II. Soils Parameters and Legend used in SWAT Model .......................................... 80

Appendix III. Soils Parameter Values used in SWAT Model ................................................. 81

Appendix IV: Parameters used for Weather Generator in SWAT Model ................................ 82

x
List of Figure

Figure 3.1: Location of study area ............................................................................................. 30

Figure 3.2: Topography of study area ........................................................................................ 31

Figure 3.3: Average measured annual rainfall from 1988-2010 of study area ............................. 32

Figure 3.4: Daily maximum temperature at four stations (1988-2010) ....................................... 33

Figure 3.5: Daily minimum temperature at four stations (1988-2010) ........................................ 33

Figure 3.6: Daily mean temperature at four stations (1988-2010) .............................................. 33

Figure 3.7: Land use and land cover of study area ..................................................................... 34

Figure 3.8: Soil type of study area ............................................................................................. 35

Figure 3. 9: Gumara river gauging station ................................................................................. 36

Figure 3.10: Conceptual framework of the study area ................................................................ 39

Figure 3.11: Double mass curve for average of three stations against Bahir- Dar station ............ 42

Figure 3.12: Spatial distribution of gauge stations ..................................................................... 43

Figure 3.13: Rescaled cumulative deviations from the mean for the total annual flow at Gumara
gauging station .......................................................................................................................... 45

Figure 3.14: Probability of rejecting homogeneity of annual flow at Gumara gauging station .... 45

Figure 3.15: Digital elevation model ......................................................................................... 47

Figure 3.16: Land use map ........................................................................................................ 47

Figure 3.17: Soil map and data .................................................................................................. 48

Figure 3.18: Sediment rating curve for Gumara River ............................................................... 49

Figure 3.19: Number of sub watersheds ..................................................................................... 50

Figure 3.20: Land use as reclassified by SWAT into the four letter land use code ...................... 51

xi
Figure 3.21: Soil as reclassified into four letters soil name in SWAT database .......................... 51

Figure 3.22: The reclassified classes of slopes for Gumara watershed........................................ 52

Figure 3.23: Ways of selecting management practice options .................................................... 57

Figure 3.24: Terracing management practice ............................................................................. 58

Figure 4.1: Sensitivity analysis of flow ...................................................................................... 61

Figure 4.2: Sensitivity analysis of sediment yield ...................................................................... 62

Figure 4.3: Hydrograph of the observed and simulated flow from the watershed for the
calibration period on a monthly basis ........................................................................................ 65

Figure 4.4: Hydrograph of the observed and simulated flow from the watershed for the validation
period on a monthly basis .......................................................................................................... 65

Figure 4.5: Comparison of observed and simulated sediment loss from the watershed for the
calibration period on a monthly basis. ....................................................................................... 66

Figure 4.6: Comparison of observed and simulated sediment loss from the watershed for the
validation period on a monthly basis. ........................................................................................ 67

Figure 4.7: Spatial distribution SWAT simulated annual sediment yield classes by sub basin
(t/ha/yr) ..................................................................................................................................... 68

Figure 4.8: Comparison of average monthly sediment yield in each scenario in sub watersheds 69

Figure 4.9: Comparison of average monthly sediment yield in sub watersheds .......................... 70

Figure 4.10: Comparison of average sediment from the critical sub watershed in each scenario . 70

xii
List of Tables

Table 2. 1: Management practice factor for different management practice options ................... 25

Table 3. 1: Soils types and their erodibility factor ...................................................................... 35

Table 3. 2: Selected meteorological stations .............................................................................. 44

Table 3. 3: Area under different classes of soil erosion by water in Ethiopia .............................. 53

Table 3. 4: The parameters used to simulate the effect management practices ............................ 60

Table 4. 1: The final ranges and fitting values of the SWAT model parameters resulted from the
final calibration procedure ......................................................................................................... 63

Table 4. 2: Performance and uncertainty for flows and sediment in calibration and validation
period. ....................................................................................................................................... 67

Table 4. 3: Area under different classes of soil erosion by water in the given sub watersheds .... 69

xiii
Acronyms and Abbreviations

A.M.S.L Above sea level

BMPs: Best Management Practices

DEM: Digital Elevation Model

EMA: Ethiopian Mapping Agency

ENMA: Ethiopia National Meteorological Agency

ENS: Nash Sutcliffe Efficiency Coefficient

GIS: Geographical Information System

HRU: Hydrologic Response Unit

LULC: Land use land covers data

MoWIE: Ethiopian Ministry of water, irrigation and energy

MUSLE: Modified Universal Soil Loss Equation

NRCS: Natural Resource Conservation Services

NPS: Non point sources

R2: Coefficient of determination

SWAT: Soil water assessment tool

SWAT-CUP: SWAT Calibration and Uncertainty Procedures

SW: Sub Watershed

SWC: Soil and Water Conservation

USLE: Universal Soil Loss Equation

WXGEN: Weather Generator

xiv
CHAPTER ONE
1. Introduction
1.1 Background

In the Ethiopian highlands, soil and water are the most critical. Nearly 85% of the population
depends on subsistence agriculture. One process that threatens the resource base is soil erosion.
Studies have shown that in Ethiopia billions of tons of soil are lost annually (USAID, 2000). Due
to greater population pressure and consequently more intensive cultivation, erosion losses have
been increasing to an annual areal average of 7 ton/ha equivalent to depth 0.5 mm ( Garzanti et al,
2006). Local erosion rates are highly spatially variable ranging from less than 1 to over 400
tons/ha/year ( Hurni, 1988, Mitiku et al.,2006 and Tebebu et al.,2010).

Watershed management, in its broader sense, can be considered an attempt to reduce soil, water
and nutrient losses from nonpoint sources (NPS) of the watershed and to ensure sustainable
agricultural production (Tripathi et al.,2004).

Of the many at risk resources in the Ethiopian highlands, soil and water are arguably the most
critical. The level of agricultural productivity in these areas is highly influenced by erratic and
unpredicted rainfall in addition to degradation of resources, such as soil. Resource degradation,
particularly soil degradation in the form of nutrient depletion, is an important factor in the decline
in the country’s agricultural production (Bekele and Holden, 1998).

The intensive study of individual watersheds is necessary to enable management plans to be

developed and also to apply the results of one watershed to another having similar characteristics.
Effective control of soil and nutrient losses requires implementation of best management practices
in critical erosion-prone areas of the watershed. The use of physically based distributed parameter
models, remote sensing techniques and geographic information systems can assist management
agencies for both identifying the most vulnerable erosion-prone areas of the watershed and
selecting appropriate management practices.

1
Erosion models are an important tool in reducing soil loss in the future by predicting the location
of vulnerable areas that need to be managed for reducing soils loss. Erosion models applied in the
Ethiopian highlands range from the empirical relationships [Universal Soil Loss Equation
(USLE)], to physical based models. Hurni (1985) adapted the empirical USLE for Ethiopian
conditions.

In order to formulate management options, soil erosion must be considered. Soil loss from a
watershed can be estimated based on an understanding of the underlying hydrological process in
a watershed, climatic conditions, landforms and soil factors. One option for formulating
management options is to use models to elucidate processes controlling the hydrologic and
sediment fluxes. Assessing and mitigating soil erosion at the watershed level is complex both
spatially and temporally. Soil type, depth, and location, land cover type and management, topology
and other factors make the watershed a complex system where hydrologic and erosive process may
differ greatly over a small spatial scale. Erosion rates depend on the rainfall intensity and the total
amount of precipitation after the onset of the rainy season thus adding a level of temporal
complexity to the system. Hence, watershed models that are capable of capturing these processes
in a dynamic manner can be used to provide an enhanced understanding of the relationship between
hydrologic processes, erosion/sedimentation, and management options. There are many models
that can continuously simulate stream flow, erosion/sedimentation, or nutrient loss from a
watershed. However, few models have been developed or tested in the monsoonal climates of
Africa. The Soil and Water Assessment Tool (SWAT) (Arnold et al., 1998) model has been used
successfully to predict runoff, erosion, sediment and nutrient transport from agricultural
watersheds under various management regimes.

2
1.2 Statement of the Problem

Through time the population will increase. This causes various effects on resource bases like
deforestation, expansion of residential area, and agricultural land expansion at the expense of
forests, lands under communal use rights (grazing and woody biomass resources), cultivation of
steep lands and overgrazing are widely practiced in the area. In addition, the ragged topography
greatly contributed to the loss of huge amounts of fertile and productive soil from farm lands.

In Ethiopian watersheds, erosion, sediment transport, and sedimentation are critical problems. The
current level of degradation leading to erosion, and sedimentation are causing considerable loss of
soil. As a consequence, the soils are becoming shallow, less fertile. In addition, water storage is
declining and droughts are becoming more frequent and intense.

Sheet and rill erosion are by far the most widespread kinds of accelerated water erosion and
principal cause of land degradation in the country and their combined effect significantly affect
agricultural production and productivity (Constable, 1984).The loss of nutrient-rich top soil by
water leads to loss of soil quality and hence reduced crop yield. Soil erosion by water and its
associated effects are therefore recognized to be severe threats to the national economy of Ethiopia.
Nearly 85% of the population depends on subsistence agriculture. One process that threatens the
resource base is soil erosion. Studies have shown that in Ethiopia billions of tons of soil are lost
annually. The average annual rate of soil loss in Ethiopia is estimated to be 12 tons/hectare/year
with losses as high as 300 tons/hectare/year (USAID,2000). In the Ethiopian highlands, in
particular, soil erosion is a major problem with an estimated loss of 16-50 ton/hectare/year (Abegaz
Gizachew, 1995).

One of the possible solutions to alleviate the problem of land degradation (soil erosion) is therefore,
to understand the processes and cause of erosion and by developing an effective management plan
for the identified critical sub watersheds to wisely utilize these precious natural resources of soil
and water while maintaining environmental quality.

3
1.3 Research Questions

The major research questions were as follows:

 How could be accessed the management practice option on sediment yield on a given
watershed?
 What will be the sediment yield/sub-basin sediment for different management practices
scenarios?
 What is the impact of identification of best management practices on sediment yield on a
given watershed?

1.4 Objectives of the Study

1.4.1 General Objective

The overall objective of this study is use SWAT to identify areas of a watershed that is highly
affected by soil erosion and targets these areas for soil erosion control measures, and also used to
select the best management options to minimize soil loss. In order to formulate management
options, the following objectives were identified:

1.4.2 Specific Objectives

The specific objectives of the studies are:

1. To apply SWAT model in simulating sediment yield in a given watershed, with current land
use land cover.

2. Identification and prioritization of the critical sub-watersheds on the basis of observed and
model simulated sediment yield.

3. Recommendation of best management practices (BMPs) for critical sub-watersheds.

4
1.5 Significance of the Study

Best Management Practices (BMPs) are a menu of options for which entities within a water use
sector can choose to implement in order to achieve benchmarks and goals through water
conservation. Best Management Practices has significantly impacts on natural resources,
socioeconomic and environmental systems. However, to assess the effects of Management
Practices on sediment yield for a given watershed understanding of the management practice
options either structural or nonstructural practices is essential indicator for resource base
analysis and development of effective and appropriate response strategies for sustainable
management of natural resources in the country in general and at the study area in particular.

This research has a profound significance role to identification of Best management practice
options on Sediment yield under different management scenarios of the watershed. Since
identifying the best management practice option on sediment yield is important. Therefore such
study has practical relevance for devising strategies and policies for sustainable water use and to
give some major action on management practices in Gumara watershed. In addition, this study will
provide scientific information on the future water resource development and fill the gaps of other
research works by incorporating the recommendations in other research works.

1.6 Scope of the Study

The scope of this study is using a physically-based SWAT model by developing different scenarios
in the study area for better understanding of the management practice effect on soil erosion and
also to understand to the capabilities of the model to identify the best management practices.

However, there has been a few research conducted about stream flow and sediment yield on
Gumara watershed, The application of models to identification of soil and water conservation
practices in the watershed is less explored compared to the application of models to estimate soil
losses. It is for this reason that I addressed this research gap in my MSc thesis by developing
different management practice scenarios in order to choose the best one. Since sound watershed
management strategies are critical to wisely utilize these precious natural resources of soil and
water while maintaining environmental quality.

5
1.7 Thesis Outline

The thesis work is organized in six chapters and associated appendixes. The first chapter is an
introduction that includes back ground, statement of the problem, research questions and
objectives, significance and overall thesis outline.

The second chapter covers the review of the main related facts from references used in the study
and review of earlier studies in the basin. The various studies conducted in specific parts of the
entire study area are well described; planned and implemented projects in the study area are also
discussed.

Chapter three gives a general description of the study area. This includes location, topography,
climate, drainage network, land use/land cover soil of the study area, and also covers materials and
methods of the study. Data availability and Processing, and methods of data analysis are presented
in detail. Relevant data of rainfall, runoff (discharge), climate data, and the like are presented; and
any additional information of each data set has been highlighted where possible. Data infilling and
extension has also been discussed in this chapter.

Chapter four brief in detail results and discussions of the study. Conclusions and
Recommendations of the thesis work are discussed in chapter five.

Finally, references, appendices in the form of tables and figures and serving as a supporting
document to this thesis are attached to make the work a complete one.

6
CHAPTER TWO
2. Literature Review
2.1 Soil Erosion and its Extent

Soil erosion was defined earlier by (Bennett, 1939) as the vastly accelerated process of soil removal
brought about by human interferences with the normal equilibrium between soil building and soil
removal. (Mitchell and Mitchell,1980) also defined soil erosion and soil loss in the following way:
“Soil erosion is the gross amount of soil moved by drop detachment or runoff and Soil loss is the
soil moved off a particular slope or field”.

Soil erosion is one of the most serious environmental problems in the world today, as it threatens
agricultural and natural environment (Vrieling,2006). As sources cited by (Deore,2005), study on
global soil loss has indicated that soil loss rate in the United States is 16 t/ha/yr; in Europe it ranges
between 10 – 20 t/ha/yr, while in Asia, Africa and South America between 20 and 40 tons/ha/yr.

As mentioned earlier on this document by (Abegaz Gizachew, 1995), the recorded annual soil
erosion in Ethiopia ranges from 16-50 tons/ha/yr depending mainly on the slope, land cover, and
rainfall intensities. According to the Ethiopian highland reclamation study (FAO, 1984), in mid-
1980’s 27 million ha or almost 50% of the highland area was significantly eroded, 14 million ha
seriously eroded and over 2 million ha beyond reclamation.

In general, Soil erosion is one of the major factors causing severe land degradation problem in
Ethiopia, which in turn is threatening the agricultural productivity and the very survival of the
overwhelming majority of the rural population. The rate of soil loss, depletion of soil organic
matter and nutrients as a result is so high and much faster than they can be replaced. The Ethiopian
Highland Reclamation Study (FAO, 1986) estimated that water erosion moves nearly 1.9 billion
tons of fertile soil from highlands annually.

Soil erosion and sedimentation by water involves the processes of detachment, transportation, and
deposition of sediment by raindrop impact and flowing water (Foster and Meyer, 1977;
Wischmeier and Smith, 1978; Julian, 1998).

7
2.2 Soil Erosion Types

Soil erosion is the removal of soil from the earth’s surface by erosive agents such as water and
wind. Water erosion involves the detachment, transport and deposition of soil particles by the
erosive forces ( Renard et al.,1997). Any soil condition which encourages surface run-off brings
with it the risk of soil wash and erosion. If the water flow is significant or concentrated, then
shallow channels (rills) may be formed which can become deeper gullies. Soil erosion by water
occurs in various forms (e.g., splash, sheet, rill, gullies) depending on the stage of progress in the
erosion cycle and the position in the landscape. Some of the types of erosion are discussed below:

2.2.1 Rain Splash Erosion

Rain splash Erosion is the result of water falling directly on to the ground during rainstorms or
when it is intercepted by the canopy and finds its way through the ground (Morgan, 1995). Some
of the water infiltrates into the soil, while some water stays on the surface saturating it and
weakening natural soil aggregates so that the impact of subsequent raindrops breaks them down.

2.2.2 Sheet Erosion

Sheet erosion occurs as a shallow 'sheet' of water flowing over the ground surface, resulting in the
removal of a uniform layer of soil from the soil surface. Although often difficult to recognize,
sheet erosion is responsible for extensive soil loss in both cultivated and non-cultivated
environments. Raindrops detach the soil particles, and the detached sediment can reduce the
infiltration rate by sealing the soil pores. The eroding and transporting ability of overland flow
depends on the rainfall intensity, infiltration rate, slope steepness, soil properties, and vegetative
cover ( Fangmeier et al.,2006).

2.2.3 Rill Erosion

Rill erosion is the detachment and transport of soil by concentrated flow of water. Rills are eroded
channels that are small enough to be removed by normal tillage operations. Rill erosion is the
predominant form of surface erosion. Rill initiated at a critical distance down slope where Surface
runoff concentrates and becomes channeled. These channels are called rills when they are small

8
enough to not interfere with field machinery operations. The water in a rill has sufficient depth for
turbulence to develop in it and therefore entrain large particles (Morgan, 1995).

2.2.4 Gully Erosion

Gully erosion produces channels larger than rills. These ephemeral channels carry water during
and immediately after rains. Gullies are distinguished from rills in that gullies cannot be obliterated
by tillage. The amount of sediment from gully erosion is usually less than from upland areas, but
the nuisance from having fields or developed areas divided by large gullies is often a greater
problem ( Fangmeier et al.,2006).The rate of gully erosion depends primarily on the runoff-
producing characteristics of the watershed, the drainage area, soil characteristics, the slope in the
channel, and the alignment, size, and shape of the gully (Bradford et al., 1973).

2.3 Factors Affecting Soil Erosion by Water

The major factors affecting soil erosion are climate, soil properties, vegetation characteristics, and
topography. These factors are not totally independent, as geology affects topography, which can
influence climate and the like. Human disturbances, such as tillage and construction, and natural
disturbances, such as severe weather or fire, dramatically increase erosion. Of these, the vegetation
and some disturbances, and, to a lesser extent, the soil and topography can be managed to reduce
erosion ( Fangmeier et al.,2006).

2.3.1 Climatic Factors

Climatic attributes affecting erosion are precipitation, temperature, wind, humidity, and solar
radiation. Soil erosion occurs when raindrops act upon the soil particles. Soil loss is closely related
to rainfall partly through the detaching power of raindrops striking the soil surface and partly
through the contribution of rain to runoff. Potential ability of rain to cause erosion is known as
erosive (R) factor ( Renard et al.,1997). Raindrops while falling acquire kinetic energy and on
impact, the kinetic energy is used up in detaching the soil particles. Energy is required to break the
soil aggregates, splashing them and subsequently carrying them with runoff (Saavedra,2005).

9
2.3.2 Soil Factors

Physical properties of soil affect the infiltration capacity and the extent to which particles can be
detached and transported. In general, soil detachability increases as the size of the soil particles or
aggregates increase, and soil transportability increases with a decrease in the particle or aggregate
size. That is, clay particles are more difficult to detach than sand, but clay is more easily
transported. The properties that influence erosion include soil structure, texture, organic matter,
water content, clay mineralogy, and density, as well as chemical and biological characteristics of
the soil ( Fangmeier et al.,2006).The susceptibility of soil to erosion agents is generally referred to
as soil erodibility ( Renard et al.,1997).Generally, soils with faster infiltration rates, higher levels
of organic matter and improved soil structure have a greater resistance to erosion. Soils high in
clay have low K values, because they are resistant to detachment. Coarse textured soils, such as
sandy soils also have low k values, because of low transportability even though these soils are
easily detachable. Medium textured soils, such as silt loam soils, have a moderate k values, because
they are moderately susceptible to detachment and they produce moderate runoff. Soils having
high silt content are the most erodible of all soils as they cause a decrease in infiltration
(Saavedra,2005).

2.3.3 Topographic Factors

Topographic features that influence erosion is slope length and steepness, shape (including
concave, uniform, or convex) and size and shape of the watershed. The slope factors (LS) refer to
topographic and/or relief factor. Erosion would normally be expected to increase with increase in
slope steepness and slope length as a result of respective increases in velocity and volume of
surface runoff (Deore,2005). Steeper terrain slopes cause higher runoff velocities, more splashes
downhill and faster flow and therefore contributes greater soil erosion. Erosion only doubled for a
steepness change from 2-20% and the erosion rate tends to level off (Remortel et al., 2001). A
slight increase in detachment rate probably occurs as the raindrops strike at a greater angle, but
this effect should not cause a major change in total splash detachment.

10
2.3.4 Ground Cover Factors

Vegetation cover is one of the most crucial factors in reducing soil erosion. Vegetation reduces
soil erosion by: protecting the soil against the action of falling raindrops, increasing the degree of
infiltration of water into the soil, reducing the speed of the surface runoff, binding the soil
mechanically, maintaining the roughness of the soil surface, and improving the physical; chemical
and biological properties of the soil (De Asis and Omasa,2007).

To account for the effect of vegetation in erosion assessments, a cover and management factor (C-
factor) has often been used. The C-factor is defined as the ratio of soil loss from land cropped
under specified conditions to the corresponding clean-tilled continuous fallow (Wischmeier and
Smith, 1978).

2.3.5 Conservation Practice Factors

Especially in agricultural areas, conservation practices such as contouring, strip cropping, or

terracing, reduce soil losses. For instance, in areas where there is terracing, runoff speed could be
reduced with increased infiltration, ultimately resulting in lower soil loss and sediment delivery.
The effectiveness of such practices is often analyzed with a support practice factor (P-factor) which
is defined as the ratio of soil loss with the practice applied and up- and down slope cultivation
(Wischmeier and Smith, 1978; ( Renard et al.,1997).

2.4 Soil Erosion Models

A wide variety of models are available for assessing soil erosion risk. Erosion models can be
classified in a number of ways. One may make a subdivision based on the time scale for which a
model can be used: some models are designed to predict long-term annual soil losses, while others
predict single storm losses (event-based). Field studies for prediction and assessment of soil
erosion are expensive, time-consuming and need to be collected over many years. Though
providing detailed understanding of the erosion processes, field studies have limitations because
of complexity of interactions and the difficulty of generalizing from the results. Soil erosion
models can simulate erosion processes in the watershed and may be able to take into account many

11
of the complex interactions that affect rates of erosion. The choice for a particular model largely
depends on the purpose for which it is intended and the available data, time and money (Lal,2001).

Erosion models allow users to ascertain temporal trends, examine spatial variations, identify
critical processes and explore the possible impacts of remedial measures and the relative
effectiveness of implementations strategies for erosion and sedimentation controls (Baigorria and
Romero,2007). Modeling in soil erosion is the process of mathematically describing soil particle
detachment, transport and deposition on land surfaces. The objective of soil erosion models is
either predictability or explanatory (Petter,1992). In general, the models fall into three main
categories: conceptual, empirical and physically based models.

Conceptual models play an intermediary role between empirical and physical based models.
Whilst they tend to be aggregated, they still reflect the hypotheses about the process governing
system behavior. This is the main feature that distinguishes conceptual model from empirical
models (Beck, 1987). According to (Renschler,1996), conceptual models tend to include a general
description of catchment processes, without including the specific details of process interactions,
which would require detailed catchment information. This allows these models to provide an
indication of the qualitative and quantitative effects of land use changes, without requiring large
amount of spatially and temporally distributed input data (Merritt et al. 2003).

Physically based models are models based on the knowledge of the fundamental erosion
processes; and incorporate the law of conservation of mass and energy (Bennett, 1974). In theory,
the parameters used in physical-based models are measurable and so are known (Merritt et al.,
2003). In practice, the large number of parameters involved and the heterogeneity of important
characteristics, particularly in catchments, means that these parameters must often be calibrated
against observed data (Beck, 1987). Examples of physically based models are the Water Erosion
Prediction Project (WEPP) and the Soil and Water Assessment Tool (SWAT).

Empirical models are generally the simplest of the three model types. They are based primarily
on defining important factors through field observation, measurement, experimentation and
statistical techniques relating erosion factors to soil loss (Petter,1992). Empirical models are
frequently used in preference to more complex models as they can be implemented in a situation

12
with limited data and parameters input (Merritt et al., 2003). They are particularly useful as a first
step in identifying the sources of sediments. The Universal Soil Loss Equation (USLE) and its
revised version RUSLE are two of the empirical models that have been most widely used and
generally accepted by the natural resources community because they are relatively easy to use
(Saavedra,2005).

2.5 Hydrological Models

Hydrological models are mathematical descriptions of components of the hydrologic cycle. They
have been developed for many different reasons and therefore have many different forms.
However, hydrological models are in general designed to meet one of the two primary objectives.
The one objective of the watershed hydrologic modelling is to get a better understanding of the
hydrologic processes in a watershed and of how changes in the watershed may these phenomena.
The other objective is for hydrologic prediction (Tadele, 2007).They are also providing valuable
information for identification of best management practices.

On the basis of process description, the hydrological models can be classified in to three main
categories (Cunderlik, 2003).

Lumped models Parameters of lumped hydrologic models do not vary spatially within the basin
and thus, basin response is evaluated only at the outlet, without explicitly accounting for the
response of individual sub-basins. The parameters often do not represent physical features of
hydrologic processes and usually involve certain degree of empiricism. These models are not
usually applicable to event-scale processes. If the interest is primarily in the discharge prediction
only, then these models can provide just as good simulations as complex physically based models.

Distributed models Parameters of distributed models are fully allowed to vary in space at a
resolution usually chosen by the user. Distributed modeling approach attempts to incorporate data
concerning the spatial distribution of parameter variations together with computational algorithms
to evaluate the influence of this distribution on simulated precipitation-runoff behavior.
Distributed models generally require large amount of (often unavailable) data. However, the
governing physical processes are modelled in detail, and if properly applied, they can provide the
highest degree of accuracy.

13
Semi-distributed models Parameters of semi-distributed (simplified distributed) models are
partially allowed to vary in space by dividing the basin in to a number of smaller sub-basins. The
main advantage of these models is that their structure is more physically-based than the structure
of lumped models, and they are less demanding on input data than fully distributed models. SWAT
(Arnold, et al., 1993), HEC-HMS (US-ACE, 2001), HBV (Bergström, 1995), are considered as
semi-distributed models.

Hydrologic models can be further divided into event-driven models, continuous process models,
or models capable of simulating both short-term and continuous events. Event-driven models are
designed to simulate individual precipitation-runoff-events. Their emphasis is placed on
infiltration and surface runoff. Typically, event models have no provision for moisture recovery
between storm events and, therefore, are not suited for the simulation of dry-weather flows. On
the other hand, continuous-process models simulate instead a longer period, predicting watershed
response both during and between precipitation events. They are suited for simulation of daily,
monthly or seasonal stream flow, usually for long-term runoff-volume forecasting and for
estimates of water yield (Cunderlik, 2003).

Generally for this study, semi-distributed models are selected because of their structure is more
physically-based than the structure of lumped model, and they are less demanding on input data
than fully distributed models.

2.5.1 Hydrological Model Selection Criteria

There are various criteria which can be used for choosing the right hydrological model for a
specific problem. These criteria are always project dependent, since every project has its own
specific requirements and needs. Further, some criteria are also user-depended (and therefore
subjective). Among the various project-dependent selection criteria, there are four common,
fundamental ones that must be always answered (Cunderlik, 2003):

 Required model outputs important to the project and therefore to be estimated by the
model (Does the model predict the variables required by the project such as long-term
sequence of flow and sediment yield?),

14
 Hydrologic processes that need to be modeled to estimate the desired outputs adequately
(Is the model capable of simulating single-event or continuous processes?),
 Availability of input data (Can all the inputs required by the model be provided within the
time and cost constraints of the project?),
 Price (Does the investment appear to be worthwhile for the objectives of the project?).

The reasons behind for selecting SWAT model for this study are;

 These models are often used for evaluating the effectiveness of various best management
practices (BMPs).
 For identification of best management practice options.
 A variety of management strategies can be modeled without excessive investment in time
or money.
 The model simulates the major hydrological process in the watersheds.
 It can work for data scarce area.
 Freely available and that it is readily applicable through the development of geographic
information system (GIS) based interfaces.
 Allowed topographical, land use and management differences.
 Could be calibrated through field testing.
 Evaluate the hydrologic effects of land use change.
 Represent variable land use throughout the watershed, and to produce a full hydrograph
response from each sub-area.

A major limitation to large area hydrologic modeling of SWAT is the spatial detail required to
correctly simulate environmental processes. Another limitation is data files can be difficult to
manipulate and can contain several missing records. The model simulations can only be as accurate
as the input data.

15
2.6 Introduction to SWAT Model

The SWAT (Soil and Water Assessment Tool) watershed model is one of the most recent models
developed at the USDA-ARS (Arnold et al., 1998) during the early 1970’s. SWAT model is
semi-distributed physically based can continuously simulate stream flow, erosion/sedimentation,
or nutrient loss in watersheds with varying soils, land use and management conditions over long
periods and primarily as a strategic planning tool (Neitsch, et al, 2005).

The interface of SWAT model is compatible with ArcGIS that can integrate numerous available
geospatial data to accurately represent the characteristics of the watershed. In SWAT model, the
impacts of spatial heterogeneity in topography, land use, soil and other watershed characteristics
on hydrology are described in subdivisions. There are two scale levels of subdivisions; the first is
that the watershed is divided into a number of sub-watersheds based upon drainage areas of the
attributes, and the other one is that each sub-watershed is further divided in to a number of
Hydrologic Response Units (HRUs) based on land use and land cover, soil and slope
characteristics.

The SWAT model simulates eight major components: hydrology, weather, sedimentation, soil
temperature, crop growth, nutrients, pesticides, and agricultural management (Neitsch, et al, 2005).
Major hydrologic processes that can be simulated by the this model include evapotranspiration,
surface runoff, infiltration, percolation, shallow aquifer and deep aquifer flow, and channel routing
(Arnold et al., 1998).

Among several physically based models, the Soil and Water Assessment Tool (SWAT) model has
been used successfully to predict runoff, erosion, sediment and nutrient transport from agricultural
watersheds under various management regimes. SWAT is a spatially distributed, continuous time
hydrological model developed by the United States Department of Agriculture, Agricultural
Research Service (USDA-ARS) (Arnold et al., 1998).

The SWAT model is mainly used to predict the long-term impacts of management measures on
water, sediment and agricultural chemical yielding for large complex watersheds with varying
soils, land use and management conditions (Arnold et al.,1996). It can also be used to simulate
water and soil loss in agriculturally dominated small watersheds (Tripathi et al, 2003).

16
Sediment yield is computed for each sub-basin with the modified universal soil loss equation
(MUSLE) (Williams and Berndt, 1997).It used to analyze the management practices option for the
critical sub watersheds. Arnold et al. (1996, 1998) suggested that the best management practices
(BMPs) could be evaluated with the SWAT model for critical erosion-prone areas, and accordingly
the practices could be recommended for reducing the soil erosion and nutrient losses.

2.6.1 Hydrological Component of SWAT

SWAT splits hydrological simulations of a watershed in to two major phases: the land phase and
the routing phase. The land phase of the hydrological cycle controls the amount of water, sediment,
nutrient, and pesticide loadings to the main channel in each sub watershed. While the routing phase
considers the movement of water, sediment and agricultural chemicals through the channel
network to the watershed outlet.

The land phase of the hydrologic cycle is modeled in SWAT based on the water balance equation
(Neitsch, et al, 2005):

SWt  SWo   Rday  Qsurf  E a  Wseep  Qgw                         2.1

t

t 1

Where SWt is the final soil water content (mm), SWo is the initial water content (mm), t is
the time (days), Pday is the amount of precipitation on day i (mm), Qsurf is the amount of
surface runoff on day i (mm), Ea is the amount of evapotranspiration on day i (mm). Wseepis
the amount of water entering the vadose zone from the soil profile on day i (mm), and Qgw
is the amount of return flow on day i (mm).

The model has eight major components: hydrology, weather, sedimentation, soil temperature, crop
growth, nutrients, pesticides, and agricultural management (Neitsch, et al, 2005). However, brief
description of some of the SWAT computation procedures which are considered in this study are
presented under the following subsections. For complete model description, one may refer to
SWAT Theoretical Documentation (Neitsch, et al, 2005).

17
2.6.1.1 Surface Runoff Volume

SWAT 2005 uses the concept that surface runoff occurs whenever the rate of water application to
the ground surface exceeds the rate of infiltration. Based on this assumption, SWAT uses two
methods for estimating surface runoff: the Soil Conservation Service Curve Number (SCS-CN)
technique (USDA Soil Conservation Service,1972) and the Green and Ampt infiltration method
(Green and Amp,1911). The SCS curve number method is less data intensive than Green & Ampt
infiltration method. Hence, the SCS curve number was used to calculate surface run off in the
watershed since available spatial data is limited. In the Soil Conservation Service (SCS) curve
number method often called the Curve-Number (CN) method, land use and soil characteristics are
lumped into a single parameter (White et al. 2008). The initial value for CN is assigned by the user
for each HRU then SWAT calculates the lower and upper limit. For this calculation, SWAT uses
a soil classification based on the Natural Resource Conservation Services (NRCS).This classifies
soil into four hydrologic groups (a soil group has similar runoff potential under similar storm and
cover condition (NRCS, 1996)) based on infiltration characteristics of the soil (Neitsch et al.,2005).
After this classification the model defines three antecedent moisture conditions to determine the
appropriate CN for each day using the CN-AMC (Curve Number –Antecedent Soil Moisture
Condition) (USDA – NRCS, 2004) distribution based on the moisture content of the soil calculated
by the model (Neitsch et al.,2005).This daily CN is then used to determine a theoretical capacity S
(retention parameter) that can be infiltrated.

 1000 
S= 25.4  10                                   2.2
 CN 

The empirical model used to estimate direct runoff from storm is the SCS runoff equation.

Pday  I a 2
Q
surf


Pday  I a  S 
                             2.3

18
Where Qsurf is the daily surface runoff in millimeters (mm), P day is the daily Precipitation (mm),
Ia is the initial abstraction which is commonly approximated as 0.2S and S is the retention
parameter. Equation above becomes.


Pday  0.2S2
                         2.4
Thus, Qsurf
P
day  0.8 S 

For the definition of hydrological groups, the model uses the U.S. Natural Resource Conservation
Service (NRCS) classification. The classification defines a hydrological group as a group of soils
having similar runoff potential under similar storm and land cover conditions. Thus, soils are
classified in to four hydrologic groups (A, B, C, and D) based on infiltration which represent high,
moderate, slow, and very slow infiltration rates, respectively.

2.6.1.2 Potential Evapotranspiration

Potential Evapotranspiration is a collective term that includes evaporation from the plant
(transpiration) and evaporation from the water bodies and soil. Evaporation is the primary
mechanism by which water is removed from a watershed. An accurate estimation of
evapotranspiration is critical in the assessment of water resources and the impact of land use
change on these resources.

There are many methods that are developed to estimate potential evapotranspiration (PET). SWAT
provides three options for PET calculation: Penman-Monteith (Monteith, 1965), Priestley-Taylor
(Priestley and Taylor, 1972), and Hargreaves (Hargreaves et al., 1985) methods. The methods have
various data needs of climate variables. Penman- Monteith method requires solar radiation, air
temperature, relative humidity and wind sped; Priestley-Taylor method requires solar radiation, air
temperature and relative humidity; whereas Hargreaves method requires air temperature only. For
this study, the Penman-Monteith method was selected as the method is widely used and all climatic
variables required by the model are available for the stations in and around the study watershed
area.

19
Penman- monteith method

The Penman-Monteith equation combines components that account for energy needed to sustain
Evaporation, the strength of the mechanism required to remove the water vapor and aerodynamic
and surface resistance terms. The penman-Monteith equation is

E 

H net  G    air * C p * e 0 z  e z  ra
                     2.5
 r 
   * 1  c 
 ra 

where λE is the latent heat flux density (MJ m-2 d-1), E is the depth rate evaporation (mm d-1), ∆
is the slope of the saturation vapour pressure - temperature curve, de/dt (kPa °C-1), Hnet is the net

radiation (MJ m-2 d-1), G is the heat flux density to the ground (MJ m-2 d-1), ρair is the air density

(kg m-3), Cp is the specific heat at constant pressure (MJ kg-1 °C-1), is the saturation vapor

pressure of air at height z ( kPa), e0zez is the water vapour pressure of air at height z (kPa), γ is the

psychometrics constant (kPa °C-1), rc is the plant canopy resistance (s m-1), and ra is the diffusion
resistance of the air layer (aerodynamic resistance) (s m-1).

 


H net  G    * k1 *  0.622 *  * air  * e 0 z  e z  ra
E t   P 
             2.6
 r 
   * 1  c 
 ra 

Where λ is the latent heat of vaporization (MJ kg-1), Et is the maximum transpiration rate (mm d-

1), K1 is a dimension coefficient needed to ensure the two terms in the numerator have the same

units (for uz in m s-1, K1= 8.64 x 104), and P is the atmospheric pressure (kPa).

20
2.6.1.3 Ground Water Flow

To simulate the ground water, SWAT partitions groundwater into two aquifer systems: a shallow,

unconfined aquifer which contributes return flow to streams within the watershed and a deep,

confined aquifer which contributes return flow to streams outside the watershed (Arnold et al.,

1993). In SWAT the water balance for a shallow aquifer is calculated with equation 4.10 see below.

aq  aq W  Q W W W                2.7
sh, i sh, i  1 rchrg gw revap deep pump, sh

Where, aqsh, i is the amount of water stored in the shallow aquifer on day i (mm), aqsh, i-1 is the

amount of water stored in the shallow aquifer on day i-1 (mm), Wrchrg is the amount of recharge

entering the aquifer on day i (mm), Qgw is the ground water flow, or base flow, or return flow, into

the main channel on day i (mm),Wrevap is the amount of water moving in to the soil zone in

response to water deficiencies on day i (mm),Wdeep is the amount of water percolating from the

shallow aquifer in to the deep aquifer on day i (mm), and Wpump,sh is the amount of water removed
from the shallow aquifer by pumping on day i (mm).

2.6.1.4 Flow Routing Phase

The second component of the simulation of the hydrology of a watershed is the routing phase of

the hydrologic cycle. It consists of the movement of water, sediment and other constituents (e.g.

nutrients, pesticides) in the stream network.

Two options are available to route the flow in the channel network: the variable storage and

Muskingum methods. The variable storage method uses a simple continuity equation inrouting the

storage volume, whereas the Muskingum routing method models the storage volume in a channel

length as a combination of wedge and prism storages. In the latter method, when a flood wave

advances into a reach segment, inflow exceeds outflow and a wedge of storage is produced. As the

21
flood wave recedes or retreat, outflow exceeds inflow in the reach segment and a negative wedge

is produced. In addition to the wedge storage, the reach segment contains a prism of storage formed

by a volume of constant cross-section along the reach length.

The variable storage method was used for this study. The method was developed by (Williams,

1969). The equation of the variable storage routing is given by:

Vstored  Vin  Vout                                    2.8

Where ΔVstored is the change in volume of storage during the time step (m3water), Vin is the volume

of inflow during the time step (m3 water), and Vout is the volume of outflow during the time step

(m3 water).

2.6.2 Sediment Component of SWAT

Erosion and sediment yield are estimated for each HRU with the modified universal soil loss

equation, MUSLE, (Williams, 1975)

2.6.2.1 Sediment Yield

SWAT uses the Modified Universal Soil Loss Equation (MUSLE) to estimate the soil loss from

each HRU. MUSLE is modified version of the Universal Soil Loss Equation (USLE) developed

by Wischmeier and smith (1965, 1978).

USLE predicts average annual gross erosion as a function of rainfall energy. The USLE model
estimates average annual soil loss by sheet and rill on those portions of landscape profiles where
erosion but not deposition is occurring. The model does neither predict single storm event nor does
it predict gully erosion (Foster, 1982; Keneth et al., 1991). In MUSLE, the rainfall energy factor
is replaced with a runoff. This improves the sediment yield prediction, eliminates the need for
delivery ratios, and allows the equation to be applied to individual storm events. Delivery ratios

22
are not needed with MUSLE because the runoff factor represents energy used in detaching and
transporting sediment (Neitsch et al.,2005).

S ed  11.8  Qsurf  q peak  area, hru 0.56  KUSLE  CUSLE  PUSLE  LSUSLE  CFRG 0.56
 2.9

Where Sed is the sediment yield on a given day in metric tons, Qsurf is the surface runoff from the
watershed in mm/ha, qpeak is the peak runoff rate in cubic meter per second, area, hru is the area
of HRU, KUSLE is the USLE soil erodability factor, CUSLE is the USLE land cover and management
factor, PUSLE is the USLE support practice factor, LSUSLE is the USLE topographic factor, and
CFRG is the coarse fragment factor.

The factors KUSLE, CUSLE, PUSLE, LSUSLE, and CFRG are taken and used based on previous studies
on the watershed and the definition and calculations of the parameters presented in the SWAT
documentation (Neitsch et al.,2005).

Soil Erodibility Factor

The erodibility of a soil is an expression of its inherent resistance to particle detachment and
transport by rainfall. It is determined by the cohesive force between the soil particles and may vary
depending on the presence or absence of plant cover the soil’s water content and the development
of its structure. It essentially depends on the amount of organic matter in the soil, the texture of the
soil, the structure of the surface horizon and permeability.

Wischmerier and Smith (1978) defined the soil erodibility factor as the soil rate per erosion index
unit for a specified soil as measured on a unit plot, and also noted that a soil type usually becomes
less erodible with decrease in silt fraction, regardless of whether the corresponding increase is the
sand fraction or clay fraction.

Land Use and Land Cover and Crop Factor

The attribute and spatial information on the present status of land use/land cover is the pr-requisite
to identify and prioritize areas for soil conservation measure and minimizing further land
degradation.

23
The C-value is a ratio comparing the soil loss from land under a specific crop and management
system to the corresponding loss from continuously fallow and tilled land. It represents the ratio
of soil loss under a given crop to that of the base soil. The cover management factor (C) measures
the combined effect of cropping and management practices in agricultural system and the effect of
ground cover, tree canopy and grass covers in reducing soil loss in non-agricultural condition. It
also reflects the effect of cropping and management practices on the soil erosion rate.

Management Practice Factor.

The conservation practices factor (p-values) reflects the effects of practices that will reduce the
amount and rate of the water runoff and thus reduce the amount of soil erosion. It depends on the
type of conservation measures implemented and requires.

Therefore, erosion control practice factor is based on the soil conservation practices operated in a
particular area. The support practices considered in this study for cultivated land include terrace,
contour and strip-cropping system that result in storage of moisture and reduction of runoff and
sediment yield. Even though, it is more suited for small-scale erosion hazard assessment mapping
than regional or basin-wide, Hurni gives parameters for different land management practices on
cultivated land. Studies conducted by (Hurni,1985) have found P values for various support
practices and land use cover. Hurni used P value range between zero and one. This means the P
value indicates reduced erosion potential, with a range between 0 to 1 because of farming practices
or soil and water conservation measures. With no erosion control practice, P is equal to one. The
farming practices increasing erosion instead of reducing are ploughing in the direction of up and
down slope with equivalent P value of one, which is the worst case scenario.

24
Table 2. 1: Management practice factor for different management practice options

Operations P factor
Ploughing up and down 1.00
Ploughing on Contour 0.90
Strip Cropping 0.80
Intercropping 0.80
Terracing 0.70
Applying mulch 0.60
Stone Cover80% 0.30
Stone cover 40% 0.80

Slope Length and Slope Steepness Factor

These BMPs were represented in the SWAT model by modifying SWAT parameters to reflect the
effect the practice has on the processes simulated within SWAT (Bracmort et al.,2006). However,
selection of BMPs and their parameters values is site specific and should reflect the study area
reality. For this study, we selected appropriate BMPs and their parameters values based on
documented local research experience in the Ethiopian highlands ( Hurni,1985; Herweg and
Ludi,1999; Gebremichael et al.,2005).

SWAT assigns the SLSUBBSN parameter value based on the slope classes. In this application the
assigned values by the SWAT were 61 m, 24 m and 9.1 m for slope class 0–10%, 10–20% and
greater than 20%, respectively. The modified parameters values were SLSUBBSN is equal to 10
m for slope class 0–10% 15and 10–20% classes

2.6.2.2 Surface Runoff and Sediment Lag

In large sub basins with a time of concentration greater than one day, only a portion of the surface
runoff is lagged as well. SWAT incorporates a surface runoff storage feature to lag part of the
surface runoff release to the main channel. Once surface runoff is calculated, the amount of surface
runoff released to the main channel is calculated using by equation (2.10), the amount of sediment
released to the main channel and the amount of sediment released to the main channel is calculated
using equation (2.11), by the model.

25
  surlag 
  Qstore,i 1 ) * (1  exp 
Qsurf  (Qsurf  )              2.10
 tconc 

  surlag 
sed  ( sed   sed store,i 1 ) * (1  exp   )                2.11
 t conc 

Where Qsurf is amount of surface runoff discharged to main channel in a day (mm), Q’ surf is amount

of surface runoff generated in a sub basin in a day (mm), Qstore i-1 surlag is the surface runoff stored

lagged from the previous day (mm), surlage is the surface runoff lag coefficient, tconc is the time
of concentration for the sub basin (hrs) and sed is the amount of sediment load generated in the
HRU on a given day (metric tons), Sedstor, i-1 is sediment stored or lagged from the previous day
(metric tons).

2.6.2.3 Sediment Routing

Sediment transport in the channel network is a function of two process, deposition and degradation.
SWAT compute both of them by using the same channel dimensions for the entire simulation. The
amount of sediment degradation in the channel can be calculated by the model by using equation
(2.12) and the net amount of sediment deposited in the reach segment is calculated equation (2.13)

sed deg  (Concsed ,ch,mx  Concsed ,ch,i ) *Vch * K ch * Cch                  2.12

sed dep  (Concsed ,ch,i  Concsed ,ch,mx ) * Vch                  2.13

Where Seddeg is the amount of sediment re- entrained in the reach segment (metric tons), Concsed,ch,i
is the amount of initial sediment concentration in the reach (kg/1 or ton/ m3), Concsed, ch, mx is the
maximum concentration of sediment that can be transported by the water (kg/l or ton/ m3), Kch is
the channel erodibility factor (cm/hr/pa), Cch is the channel cover factor, Vch is the volume of water
in the reach segment (m3), Seddep is the amount of sediment deposited in the reach (metric tons).

Once the amount of degradation and deposition has been calculated by the above equations (2.12
and 2.13 respectively), then the final amount of sediment in the reach is determined by equation

26
(2.14) and the amount of sediment transported out of the reach is calculated by equation (2.15) by
the model.

sedch  sedch,i  sed dep  seddeg                    2.14

Vout
sed out  sed ch *                      2.15
Vch

Where sedch is the amount of suspended sediment in the reach (metric tons), sedch,i is the amount
of suspended sediment in the reach at the beginning of the time period(metric tons), seddeg is the
amount of sediment re-entrained in the reach segment (metric tons), Sedout is the amount of
sediment transported out of the reach (metric tons), Vout is the volume of out flow during the time
step (m3) and Vch is the volume of water in the reach segment (m3).

2.6.3 SWAT Model Application Worldwide

The SWAT model has good reputation for best use in agricultural watersheds and its uses have
been successfully calibrated and validated in many areas of the USA and other continents
(Ndomba, 2002; Tripathi et al., 2003). The studies indicated that the SWAT Model is capable in
simulating hydrological process and erosion/sediment yield from complex and data poor
watersheds with reasonable model performance statistical values. (Ndomba, 2002) was applied the
SWAT model in modeling of Pangari River (Tanzania) to evaluate the applicability of the model
in complex and data poor watersheds. Tripathi et al., (2003) applied the SWAT model for Nagwan
watershed in India with the objective of identifying and prioritizing of critical sub watersheds to
develop an effective management plan and the model was verified for both surface runoff and
sediment yield. Accordingly, the study concluded that the SWAT model can be used in ungauged
watersheds to simulate the hydrological and sediment processes.

SWAT has gained international acceptance as a robust interdisciplinary watershed modeling tool
as evidenced by international SWAT conferences, hundreds of SWAT-related papers presented at
numerous other scientific meetings, and large number of articles published in peer-reviewed
journals (Gassman, 2007).

27
However, (Cibin et al. 2010) indicated that SWAT model parameters show varying sensitivity in
different years of simulation suggesting the requirement for dynamic updating of parameters
during the simulation. The same study also indicated that sensitivity of parameters during various
flow regimes (low, medium and high flow) is also found to be uneven, which suggests the
significance of a multi -criteria approach for the calibration of the model.

2.6.4 SWAT Model Application in Ethiopia

The SWAT model application was calibrated and validated in some parts of Ethiopia, frequently
in Blue Nile basin. A study conducted on modeling of the Lake Tana basin with SWAT model
also showed that the SWAT model was successfully calibrated and validated (Setegnet al., 2008).
This study reported that the model can produce reliable estimates of stream flow and sediment
yield from complex watersheds. (Gessese2008) used the SWAT model performed to predict the
Legedadi reservoir sedimentation. According to this study, the SWAT model performed well in
predicting sediment yield to the Legedadi reservoir. The study further put that the model proved
to be worthwhile in capturing the process of stream flow and sediment transport of the watersheds
of the Legedadi reservoir.

In addition to the above, the SWAT model as tested for prediction of sediment yield in Anjeni
gauged watershed by (Setegn et al., 2008). The study found that the observed values showed a
good agreement at Nash-Sutcliff efficiency (ENS) of 80 %. In light of this, the study suggested
that the SWAT model can be used for further analysis of different management scenarios that
could help different stakeholders to plan and implement appropriate soil and water conservation
strategies. (Tekle 2010) through modeling of Bilate watershed also indicated that SWAT Model
was able to simulate stream flow at reasonable accuracy.

The literature reviewed and presented above showed that SWAT is capable of simulating
hydrological and soil erosion process with reasonable accuracy and for different management
practice options and also can be applied to large and complex watersheds.

28
2.6.5 SWAT-CUP

SWAT-CUP (SWAT Calibration and Uncertainty Procedures) is designed to integrate various

calibration and uncertainty analysis programs for SWAT (Soil & Water Assessment Tool) using
different interface. Currently the program can run SUFI2 (Abbaspour,et al.,2007), GLUE (pp.
Beven,1992), and Parasol (van Griensven and Meixner, 2006), PSO, and MCMC procedures.
Currently the program links with Generalized Likelihood Uncertainty Estimation (GLUE),
Parameter Solution (Parasol) [6], Sequential Uncertainty Fitting (SUFI2), and Markov chain
Monte Carlo (MCMC) procedures. SWAT parameters related to discharge and sediment were
estimated using those described optimization techniques.

2.7 Previous Study

Assessment of management alternatives on a small agricultural watershed have been studied using
several types of models, which varied from strictly empirical to physical-based distributed models.
Few studies have examined ( Mostaghimi et al,1991). It relies on distributed parameter models to
evaluating the currently installed BMPs in the watershed; and to simulate the possible impact of
alternative BMPs for reducing NPS pollutant loadings. The recent work Identification and
Prioritizations of Critical Sub-watersheds for Soil Conservation Management using the SWAT
Model (Tripathi et al.,2003) using swat maodel identification and prioritisation of critical sub-
watersheds to develop an effective management plan.In Ethiopia Identify Watershed Management
Options (Biniam,2009)

Though several studies on best management practice for reducing runoff and sediment yield have
been discussed several times on long-term impact of currently installed BMPs, still there is no
much study overseeing on hypothetical alternatives BMPs.

29
CHAPTER THREE
3. Materials and Methods
3.1 General
For any research, identifying clear and efficient methodology is crucial for the effectiveness of the
study not only from time budget point of view, but also from the quality of the research result. The
methodology in this study includes office and field level investigation of the study area and
applying methods and techniques for data analysis.
3.2 Study Area
3.2.1 Location
The study area is found in north west part of Ethiopia in Amhara Regional State, South Gonder
Zone.Gumara watershed covers partly four weredas (Districts) namely.Farta,fogera Dera and Estia
It is located in the south east of Lake Tana as shown in Figure 3.1. Gumara watershed is one of the
sub watersheds in Abbay River Basin which is part of the Blue Nile Basin in Ethiopia. It is
geographically located between 37030’ and 38011’ east longitude between 110 34’and 110 57’ north
latitude.

Figure 3.1: Location of study area

30
3.2.2 Topography

The major land form of the watershed includes flat, gently sloping to undulating plains, hills and
mountains. The upper and middle parts of the catchment are characterized by mountainous, highly
rugged and dissected topography with steep slopes and the lower part is characterized by valley
floor with flat to gentle slopes. Elevation in the watershed varies from 1795 to 3703 meter above
sea level with a mean elevation of 2200 meter above sea level.

Figure 3.2: Topography of study area

3.2.3 Climate

The climate of Ethiopia can be classified in different ways including the Traditional, Koppen’s,
Throthwaite’s, Rainfall regimes, and Agro-ecological zone classification systems. The most
common used classification systems are the traditional and the agro-ecological zones. Traditional
classification system, is mainly relied on altitude and temperature. There are five climatic zones
namely: Wurch (cold climate at more than 3000 masl), Dega (temperate like climate-highlands
with 2500-3000 masl), Woina Dega (warm at 1500-2500 masl), Kola (hot and arid type, less than
1500m in altitude), and Berha (hot and hyper-arid type). Based on this classification the study area
is under Woinadega class.

31
Precipitation is an important hydrological variable that is notoriously variable spatially. In the
temperate regions, precipitation is equal to rainfall since there is no snowfall. Information required
in estimating the influence of precipitation ranges from general regional and seasonal variability
to the frequency and magnitude of individual storm events. Rainfall recording stations in the study
area has been in operation since 1975 and increased gradually. Four meteorological Stations were
identified for the Gumara watershed, Bahir Dar,Agere Genet,Debre Tabor,Woreta and actually
used in this study.

1650

1600
Rainfall (mm)

1550

1500

1450

1400

1350

1300
Bahir-Dar Debre-Tabor Wereta Agre Genet
Stations

Figure 3.3: Average measured annual rainfall from 1988-2010 of study area

Based on long term the average data temperature variation in stations,within the year is minor
9.185⁰C in November (Agre Genet) to 25.01⁰C to in May (wereta) with monthly average daily
maximum as shown in Figure 3.4 and minimum temperature of 33.81⁰C (wereta) and 0.81⁰C (Agre
Genet) in Figure 3.5, respectively. Humidity varies 40% in April and 80% in August. Wind speed
is low which limits potential evapotranspiration to value between 95 mm/month in December and
140 mm/month in April based on Bahirdar meteorological station.

32
Tempreture (⁰C) Tempreture (⁰C) Tempreture (⁰C)

12
14
16
18
20
22
24
26
10
15
20
10
20
25
30
35
40

10
15

8
0
0
5

5
Jan-88 Jan-88 Jan-88
Mar-89 Feb-89 Feb-89
May-90 Mar-90 Mar-90
Jul-91 Apr-91 Apr-91
Sep-92 May-92 May-92
Jun-93 Jun-93
Nov-93
Jul-94 Jul-94
Jan-95
Aug-95 Aug-95
Mar-96
Sep-96 Sep-96
May-97
Oct-97 Oct-97
Jul-98
Nov-98 Nov-98
Time (Month)

Sep-99 Dec-99 Dec-99

Nov-00 Jan-01 Jan-01

Time (Month)
Time (Month)

33
Jan-02 Feb-02 Feb-02
Mar-03 Mar-03 Mar-03
May-04 Apr-04 Apr-04
Jul-05 May-05 May-05

Figure 3.6: Daily mean temperature at four stations (1988-2010)

Sep-06 Jun-06 Jun-06
Jul-07 Jul-07

Figure 3.5: Daily minimum temperature at four stations (1988-2010)

Nov-07
Figure 3.4: Daily maximum temperature at four stations (1988-2010)

Aug-08 Aug-08
Jan-09
Sep-09 Sep-09
Mar-10
Oct-10 Oct-10

Wereta
Wereta
Wereta

Bahir-Dar
Bahir-Dar
Bahir-Dar

Agre Genet
Agre Genet
Agre Genet

Debre-Tabor
Debre-Tabor
Debre-Tabor
3.2.4 Land Use and Land Cover

The land use and land cover pattern of a watershed is an outcome of natural and socioeconomic
factors and their utilization by man in time and space. Land is becoming a scarce resource due to
immense agricultural and demographic pressure. Hence, information on land use and land cover
and possibilities for their optimal use is essential for the selection, planning and implementation
of land use schemes to meet the increasing demands for basic human needs and welfare. In this
study area more than three quarter of the watershed is agricultural land. Areas used for crop
cultivation, both annuals and perennials, and the scattered rural settlement that are closely
associated with the cultivated fields. While a few highlands of the study area are covered by bush
or shrub land, grazing and, forest land, and built up area are other land cover types in the watershed
(WWDSE, 2009).

1%
2%

17% Built Up Area

3% Cultivated Land
Forest Land
Grass Land
77% Shrub and Bush Land

Figure 3.7: Land use and land cover of study area

3.2.5 Soil

Generally, the soil type of this watershed area are characterized with shallow, moderate to
deep and very deep in depth and sandy, clay to clay texture types. The erodibility of these
soils also varies from medium to very erodible characteristics.

Soil texture refers to the relative proportion of sand, silt and clay in a mass of soil. Soil texture
influences soil infiltration, moisture and nutrient retention, drainage, workability and susceptibility
to erosion.

34
The soil that covered the catchment are classified by six different classes of soils based on technical
classifications following the classification of MoWIE. Six of the major soil classifications in this
study area are Luvisols, Fluvisol, Leptsols, vertisols, in combination with diagnostic horizon
modifiers: chromic, Eutric, Haplic, and also urban are common soil types in Gumara watershed
based classification system on Ethiopian ministry of water, irrigation and energy. Vertisols are
deep to very deep, moderately well to poorly drained, very dark grey to dark yellowish brown in
the topsoil, and clay textured throughout. The soils have large surface cracks in the dry season.
Run-off formation from Vertisols is high and hence it is susceptible to erosion. The recent soils
which are not developed are classified as Fluvisols found in small extent in the watershed area.
The shallow and very shallow soils are classified as Leptosols. Leptosols are found in relatively
small areas in the watershed area. These are stony and rocky. The texture of Leptosols varies from
sandy clay loam to clay and has excessive drainage characteristics.

Haplic Luvisols exist in bigger extent in the watershed area. These soils show textural
differentiation with moderate to high clay content. These soils are almost intensively cultivated.

Table 3. 1: Soils types and their erodibility factor

Major soil groups Soil Color K Value

Vertisols Black 0.15
Fluvisols Yellow 0.30
Leptosols Brown 0.20

(Source: Hurni, 1985)

4.5% 0.5%

14%
Eutric Fluvisols
40%
Chromic Luvisols
26% Eutric Leptosols
Haplic Luvisols
15% Eutric Vertisols
Urban

Figure 3.8: Soil type of study area

35
3.2.6 Hydrology

Gumara is one of the major rivers which contributes significant amount of flow to Lake Tana. For
this reason, the Ethiopian Minister of Water Resources installed one gauging station downstream
of the river near the small town called Gumara figure. It has long term daily observation since
1975. Based on the data from 1988-2009 records, it has an average annual daily flow of 36.4m3/s.
Rainfall over the Gumara watershed is concentrated in the Summer season (June to September)
and with virtual drought from November to April. The four wettest months cover 85 percent of the
total annual rainfall. The dry season from October to May has a total rainfall of about 15% of the
mean annual rainfall (WWDSE, 2007).

Figure 3. 9: Gumara river gauging station

36
3.3 Materials and data used

Materials that were used in this research were spatial data (land use map, soil map and digital
elevation model), meteorological data (rainfall, maximum and minimum temperature, relative
humidity, wind speed and solar radiation), hydrological data, sediment data and stationery
materials has been collected and used in this study.

Tools and materials used:

 Arc GIS 10.1
 Arc SWAT 2012
 SWAT CUP 5.1.6
 Rainbow for data homogeneity test
 DEM (digital elevation model)
 Land use map and soil map
 DEW02, pcp STAT
 Excel Spread sheet

3.4 General approach of Method

Primarily relevant, data such as DEM data, topographic map, land use land cover map, soil map,
meteorological variables (minimum, maximum temperature and precipitation), solar radiation,
wind speed, and relative humidity and hydrological data’s are collected. From the collected data
such as DEM was used to delineate the topographic characterization of Gumara watershed/sub
basins and determine the hydrological parameters of the watershed such as slope, flow
accumulation, flow direction, and stream network. Then, using the delineated watershed, soil map,
slope and land use/cover map data, developed Hydrologic Response Unit (HRU). HRU creates for
unit combination of land use slope, and soil type with in each sub basin. After created HRU and,
weather data, SWAT model was running, and sensitivity analysis was done with the input of
hydrological data. From the sensitivity, which parameters govern the watershed could obtain using
SWAT CUP SUFI-2 algorithm method by taking the P-value and t- stat to cheek their sensitivity
and significance respectively. SUFI-2 program was used for a combined optimization-uncertainty
analysis. SUFI-2 used amulti-site, semi-automated global search procedure. Then, model

37
calibration followed sensitivity analysis. Model calibration and validation also done using SUFI-
2 algorithm method until the observed and simulated fitted is an acceptable condition. Finally, all
the results of the model are going to interpret and then used to identify the critical sub watersheds
for sediment yield. After identification the critical or prone area by taking different agricultural
and structural control measures, the best management practice options were selected to minimize
soil losses in critical erosion-prone areas of the Gumara watershed.

3.4.1 Scenario development

Base Scenario: - processes simulation of hydrology and sediment yield without management
practice.

Scenarios: - processes simulation sediment yield with the interference of different management
practice options. And finally identify and select the best management options for the critical area
in a given watershed.

38
Data Collection

Collecting
DEM Information on land
Meteorological management
Data practicesmanagem
Slope LULC Soil ent

practices
Hydrological
Data

Hydrological Model
ARC SWAT

HRU

Sensitivity Analysis

Management
Model Calibration and practice Scenarios
Validation

Legend
Model Simulation under
different scenario
Input Data
condition

Model Process

Select management Result

option for critical sub-
watersheds

for critical sub-

watersheds
Figure 3.10: Conceptual framework of the study area
Where LULC is the land use land cover of Gumara watershed.

39
3.5 Data Availability and Analysis

The required data were collected from the different data source like organization, private sectors
key informants and persons during field works. The model build up processes involve the
preparation of the inputs data classified as spatial and hydro and meteorological data input data.

3.5.1 Spatial Data

Those data for the study area was obtained as:

 Digital Elevation Model /DEM /: Topography data to delineate a watershed into multiple
sub watersheds and also to calculate watershed/sub-watersheds parameters such as average
slope, slope length, and the accumulation of flow for the definition of stream networks
obtained from Ethiopian Mapping Agency (EMA).
 Land use land cover data /LULC/: it is spatial dataset in the model defines the densities
and types of land use found within a given area. The LULC is processed and prepared as
map during the image classification from Ethiopian Ministry of water, irrigation and energy
MoWIE GIS department
 Soil map/ Soil Data /,digital stream network data were collected from Ethiopian Ministry
of water, irrigation and energy MoWIE GIS department

3.5.2 Hydro-Meteorological Data

The Ministry of water, irrigation and energy (MoWIE) has the sole authority for collection and
processing of river (Hydrological) flow data in Ethiopia. The Ministry operates gauging stations
located on various locations along Gumara River. Those data such as daily discharge and sediment
load from the appropriate gagging site in the study area were taken from MoWIE.

The availability and quality of meteorological data such as rainfall, temperature, sun shine hours,
wind speed, and relative humidity is vital for any water resources study. In this study,
meteorological data: Rainfall, temperature (maximum and minimum air temperature), relative
humidity, solar radiation and potential evaporation, and wind speed has been collected from
National Meteorological Agency (NMA).

40
3.5.3 Filling Missing Weather Data

Data reconstruction was necessary due to the fact that continuous time series data was needed for
most of the analysis performed in this study. Almost all the records have chuck of missing values
and appropriate techniques were required to fill the gaps.

There are number of methods for estimating missing precipitation data. These are: - station
average, normal ratio, inverse distance weighting, quadrat method and regression methods. From
these method normal ratio and arithmetic method was used for this study.

In the normal ratio method, The rainfall P A at station A is estimated as a function of the normal
monthly or annual rainfall of the station under question and those of the neighboring stations for
the period of missing data at the station under question.

NRA

n
i 1
* Pi
NRi
PA                         3.1
N

Where PA = Missing value of precipitation to be computed, NRA = is the normal monthly, seasonal

rainfall at station A, NRi = is the normal monthly, or annual rainfall at station i, Pi = Rainfall of
neighboring station during missing period, N = Number of stations used in the computation.

Filling of missing temperature data was done with the same procedure and method as rainfall data.

3.5.4 Checking Data Consistency of the Stations

Numerous factors could affect the consistency of the recorded data at a given station. These are
damage and replacement of a measuring gage, change in the gage location or elevation, growth of
high vegetation or construction of a building, change in measurement procedure, or human,
mechanical, or electrical error in taking readings.

41
A significant change occurs in and around a particular rain gauge station will affect the rain gauge
data, being reported from a particular station. After a number of years, it may be felt that the data
of that station is not giving consistent rainfall (Garge, 2005). To detect any such inconsistency,
and to correct and adjust the reported rainfall value of the selected meteorological stations, the
double mass curve method is used. It is the most common method of checking for in-consistency
records of meteorological data.

The curve is a plot on arithmetic graph paper of cumulative rainfall collected at a gauge where
measurement conditions may have changed significant against average of the cumulative rainfall
for the same period of record collected at several gauges in the same region. The data is arranged
in the reverse order that is the latest record as the first entry and the oldest record as the last entry
in the list. A change in the proportionality between the measurement at the suspect station and
those in the region is reflected in a change in the scope of trend of the plotted points. The data
series, which is in consistent, adjusted to consistent values by proportionality.

40000
COMULATIVE ANNUAL FOR EACH

Bahir dar
R² = 0.9999
30000
STATIONS IN (MM)

Debre tabor
20000 R² = 0.9999

Wereta
10000 R² = 0.9999

Agere Genet
0 R² = 0.9998
0 10000 20000 30000 40000
FOUR STATIONS AVERAGE ANNUAL COMULATIVE RAINFALL (MM)

Figure 3.11: Double mass curve for average of three stations against Bahir- Dar station

3.5.5 Thiessen Polygon Method

The Thiessen Polygon approach is one the most common method used for determining average
precipitation over an area when there is more than one measurement. The basic concept is to divide
the watershed into several polygons, each one around a measurement point, and then take a

42
weighted average of the measurements based on the size of each one’s polygon, i.e., measurements
within large polygons are given more weight than measurements within small polygons. The
weighted average is calculated by


n
Pi Ai
P  i 1
                         3.2

n
A
i 1 i

Where P is the weighted average, P’s are measurements, and A’s are areas of each polygon.

Table 3.2 list and name of selected rainfall stations and Figure 3.12 shows the spatial distribution
of selected rainfall stations in the study area.

Figure 3.12: Spatial distribution of gauge stations

The rainfall is never uniform over the entire area of the basin or catchment, but varies in intensity
and duration from place to place. Thus the rainfall recorded by each rain gauge station should be
weighted according to the area, it represents. This method is more suitable under the following
conditions: for areas of moderate size, when rainfall stations are few compared to the size of the
basin.

43
Table 3. 2: Selected meteorological stations

Latitude/Longitude
Sr.No Station Name Altitude (m) Easting(0) Northing(0) Area Cover (%)
Bahir Dar 1828 37.42 11.60 6.8
Agere Genet 3010 38.18 11.92 4.1
Debre Tabor 2690 37.98 11.89 69.6
Woreta 1819 37.6 11.86 19.5

3.5.6 Stream Flow Availability, and Homogeneity Test

The main object of any hydrological study is to be able to simulate stream flow at all sensitive
locations within a catchment/basin. This is because stream flow is the key variable to determine
the potential water resources availability and supply. This objective of the hydrologic studies
includes the verification of the simulated stream flows against all available observed data records.

Rainbow software has been used to check the homogeneity of data. Frequency analysis of rainfall
data and flow data and their potential use in agro- meteorological decision–making processes
requires that the data be of long series.

Homogeneity test: Frequency analysis of data requires that the data be homogeneous and
independent. The restriction of homogeneity assures that the observations are from the same
population. RAINBOW offers a test of homogeneity, which is based on the cumulative deviations
from the mean.

 _

S k  i1  X i  X                         3.3
k

 

_
Where Xi are the records from the series X1, X2… Xn and X the mean. The initial value of Sk=0

and last value Sk=n are equal to zero (Figure 3.13). When plotting the Sk’s (also called a residual

mass curve) changes in the mean are easily detected. For a record Xi above normal the Sk=i

increases, while for a record below normal Sk=i decreases. For a homogenous record, one may

44
expect that the Sk’s fluctuate around zero since there is no systematic pattern in the deviations of
_
the Xi’s from their average value X .k=1... , n.

When the deviation crosses one of the horizontal lines the homogeneity of the data set is rejected
with respectively 90, 95 and 99% probability.

Figure 3.13: Rescaled cumulative deviations from the mean for the total annual flow at Gumara
gauging station

Figure 3.14: Probability of rejecting homogeneity of annual flow at Gumara gauging station

45
3.6 Model Input

SWAT is a comprehensive model that requires information provided by the user to simulate runoff
and soil erosion. The first step in initializing a watershed simulation is to partition the watershed
into sub basins. The user has the option of allowing SWAT to automatically delineate the
watershed and sub basins using the Digital Elevation Model (DEM) or the user can provide
predefined sub basins. The land area in a sub basin is divided into hydrologic response units
(HRUs). Hydrologic response units (HRUs) are portions of a sub basin and possess unique land
use, slope range, and soil attributes (Neistch et al.,2004).

SWAT has different components. Hydrologic components of the model work on the water balance
equation, which is based on surface runoff, precipitation, percolation, evapotranspiration, and
return flow data; Weather is one of the model component that needs data on precipitation, air
temperature, solar radiation, wind speed, and relative humidity data; Sedimentation is another
component of the model that needs information on surface runoff, peak rate flow, soil erodability,
crop management, erosion practices, slope length, and steepness; Soil temperature, crop growth,
nutrient pesticides and agricultural management are also components of SWAT. Thus, the data
required for the model are DEM, soil data, land use data, precipitation and other weather data. For
calibrating the model and also for validation purposes, river discharge and sediment yield are
required at the outlet of the watershed.

3.6.1 Digital Elevation Model

To delineate the watershed and sub basins and to determine drainage networks SWAT uses the
digital representation of the topographic surface. DEM is the digital representation of the
topographic surface. A 90m by 90m resolution DEM was used from the Ethiopian Ministry of
water, irrigation and energy, see Figure 3.15. Sub basin parameters such as slope gradient, slope
length of terrain and the stream network characteristics such as channel length, width and slope
were calculated and used by the model.

46
Figure 3.15: Digital elevation model

3.6.2 Land Use Map

Land cover classification is used for many applications like: conservation measure, biodiversity
assessment, water quality assessment, assessing land cover change and so on. Based on this
classification majority of the study area is covered by cultivation lands while a few highlands of
the study area are covered by forests, bush and shrubs as shown below.

Figure 3.16: Land use map

47
3.6.3 Soil Map and Data

Soils in the watershed should be categorized and prepared as a map in a shape file format and then
linked to a customized soil database designed by the user if the soils are not included in the existing
SWAT soil database.

Figure 3.17: Soil map and data

3.6.4 Weather Data

SWAT requires daily or sub-daily meteorological data. The model can either read these
meteorological data from previously measured data stored in tables or can be generate it using a
weather generator model. In this study, measured meteorological data were used. The
meteorological data used were daily precipitation, daily maximum and minimum air temperature,
daily solar radiation, wind speed, and relative humidity on a daily basis.

3.6.5 River Discharge and Suspended sediment data

Daily discharge measurement data of Gumara River were collected from MoWIE. Sediment rating
curve used in this research was taken from the study made by (Mequanint,2008) and the gap of
daily suspended sediment was estimated using this rating curve equation.

48
Sediment Rating Curve
160000
Mean suspended sediment (ton/day)
140000

120000 y = 23.35x1.5
R² = 0.931
100000

80000

60000

40000

20000

0
0 50 100 150 200 250 300
Mean daily discharge (m3/s

Figure 3.18: Sediment rating curve for Gumara River

(Source: Mequanint,2008)

3.7 Model Setup

3.7.1 Watershed Delineation

SWAT allows the user to delineate the watershed and sub basins using the Digital Elevation Model
(DEM). This tool uses and expands ArcGIS and Spatial Analyst extension functions to perform
watershed delineation (Neitsch et al.,2002). The DEM of the area is loaded into a GIS as grid
format. Stream network was defined for the whole DEM by SWAT using the concept of flow
direction and flow accumulation. Before defining the stream network, the model processes the
DEM map grid to remove all the non-draining zones (sinks). To define the origin of streams a
threshold area was defined. The threshold area defines the minimum drainage area required to
from the origin of a stream. The size and number of sub basins and details of stream network
depends on this threshold area (Winchell et al., 2007). The threshold area was taken to be
2958.27ha, suggested by Arc SWAT. The threshold area, or critical source area, defines the
minimum drainage area required to form the origin of a stream. The watershed outlet is manually
added and selected for finalizing the watershed delineation. With this information the model

49
automatically delineated a watershed of 126950.6914 ha and 25 sub basins were produced see
Figure 3.19 below.

Figure 3.19: Number of sub watersheds

3.7.2 HRU Definition

After watershed delineation, the watershed was partitioned in to hydrological response unit. The
Hydrologic Response Units (HRUs), which are combinations of land use (Figure 3.20), soil
(Figure 3.21) and slope (Figure 3.22) within the watershed. There are two options: the one can be
assigning only single HRUs for each sub watershed considering the dominant soil/land use
combination. The second way is by assigning multiple HRUs for each sub watershed considering
the sensitivity of the hydrologic process based on a certain threshold values of soil/land use
combinations. In this study multiple HRUs was selected. Analysis tool in Arc SWAT helps to load
land use and soil layers to the project. The delineated watershed by Arc SWAT and the prepared
land use overlapped 100%. The five classes land use map was named into five classes swat
database.

50
Figure 3.20: Land use as reclassified by SWAT into the four letter land use code

The delineated watershed and soil map have also overlap of 100%. The soil classes in the input
soil map were decoded using a lookup table (a table created by the user in .txt format according to
SWAT input file format).

Figure 3.21: Soil as reclassified into four letters soil name in SWAT database

51
HRU analysis in Arc SWAT includes divisions of HRUs by slope classes in addition to land use
and soils. The multiple slope option (an option for considering different slope classes for HRU
definition) was selected for this study. The slope discretization (0-2, 2-8, and >8%).

Figure 3.22: The reclassified classes of slopes for Gumara watershed

The HRU analysis section takes land use, soil and slope data, and divides each sub basin in to
HRU with specific combinations of the three layer respective characterization. The layer produced
by this process is crucial to the ultimate analysis performed by the SWAT model, because it
determines the land use, soil and slope category assigned to each HRU. This category determines
how land will respond to precipitation, run off, infiltration and other hydrologic processes during
the simulation. Each sub basin can then have one or more major HRUs defined within it. Finally
total of 177 HRUs for 25 sub basins were created.

3.7.3 Weather Data Definition

The WXGEN weather generator included in SWAT model was used to fill in gaps in measured
records. This weather generator was developed for U.S. Since Gumara watershed is located outside
U.S. the WXGEN weather generator was provided with all the necessary statistical information
from the meteorological records of the watershed. The parameters needed for the weather
generator are listed in Appendix (for definition of each parameters listed look at Neitsch et

52
al.,2005. These statistical values were calculated from the meteorological data available in the
Guamara watershed and Bahir Dar station. The number of years for calculating the statistical
values depends on the availability of data in the stations. Other meteorological data (daily
precipitation, daily minimum and maximum air temperature, daily relative humidity, daily solar
radiation and daily wind speed) including the corresponding location table were prepared
according to the SWAT format and integrated into the model using the weather data input wizard.

3.7.4 Identification of Critical Sub Watersheds

The critical sub watersheds were identified on the basis of average annual sediment yield simulated
using the SWAT model from the sub watersheds for the simulated period. Annual sediment yields
were simulated for each sub watershed of the Gumara watershed, and this helps to identify the
critical watershed based on the soil erosion class by comparing the result output in all sub basins
in the watershed.

Table 3. 3: Area under different classes of soil erosion by water in Ethiopia

Soil erosion
classes Slight Very Low Low Moderate High Very high Severe Very severe
Soil erosion
range (t/ha/yr) 0-1 1-5 5-8 8-10 10-15 15-25 25-31 >31

(Source: Hurni, 1985 and Israel, 2011)

3.7.5 Calibration and Validation Setup and Analysis

The SWAT model was calibrated and validated for stream flow and sediment yield (SY) in the
Gumara watershed. Monthly discharge and sediment records from 1994 to 2004 at the outlet of
the watershed (the Gumara hydrological station) were split into two segments, 1994–2001 and
2002–2004, in order to calibrate and subsequently validate water and sediment-relative parameters.

For this study SWAT Calibration and Uncertainty Procedures (SWAT-CUP) software was selected
to do the calibration because of its capability to perform calibration, validation, sensitivity analysis
and uncertainty analysis and also because its performance was better than the auto-calibration
modulus embedded in the SWAT interface (Zhou et al., 2014).

53
The SWAT-CUP program contains different algorithms, including: Sequential Uncertainty Fitting
(SUFI-2), Particle Swarm Optimization (PSO), Generalized Likelihood Uncertainty Estimation
(GLUE), Parameter Solution (ParaSol), and Markov Chain Monte Carlo (MCMC). Compared with
other algorithms, SUFI-2 has a higher efficiency in achieving a similarly accurate predication
result, which has been widely used to model stream flow, sediment load, and water quality in
recent years (Abbaspour et al., 2007, 2009; Azimiet al., 2013; Faramarzi et al., 2009; Schuol et al.,
2008a, b).So SUFI-2 was selected to calibrate the stream flow and Sediment yield in the Gumara
watershed.

3.7.5.1 Sensitivity Analysis

The aim of the sensitivity analysis is to estimate the rate of change in the output of a model with
respect to changes in watersheds that result in a clear difference in hydrologic sensitivity
(Reungsang et al.,2005). Sensitivity analyses were conducted for the Gumara watershed hydrology
to determine the parameters needed to improve simulation results and thus to better understand the
behavior of the hydrologic system and to evaluate the applicability of the model.

Sensitivity analysis from SUFI-2 provided partial information about the sensitivity of the objective
function to model parameters. Different water-related parameters (global parameters), along with
sediment-related parameters with absolute minimum and maximum ranges in the SWAT model
documents were selected to do sensitivity analysis separately. Then see the sensitivity ranking, and
checking there stat. A t stat provides a measure of sensitivity (larger absolute values are more
sensitive), and p values determine the significance of the sensitivity (a value close to zero has more
significance).

3.7.5.2 Model Evaluation and Uncertainty Analysis

Model evaluation is an essential measure to verify the robustness of the model. In this study, two
model evaluation methods were used. In order to evaluate the model’s performance relative to the
observed data; Correlation coefficient (R2) and Nash and Sutcliffe simulation efficiency (ENS).

The coefficient of determination (R2) suits model evaluation and is very sensitive to extreme values
(Harmel and Smith 2007). In this study, the coefficient of determination (R2) and the Nash-

54
Sutcliffe coefficient (ENS) were used. The coefficient of determination (R2) is the square of the
Pearson’s product-moment correlation coefficient and describes the proportion of the total
variance in the observed data that can be explained by the model. R 2 ranges from 0.0 to 1.0 with
higher values indicating better agreement (Legate and McCabe, 1999).

 X * X 
2
 n 
 X obs  X sim ,
mean mean
 obs,i sim , i 
R 
2 i
                3.4
  X
i
n
obs,i  X obs 
mean 2
*  X sim , i  X sim , 
mean 2


Where Xobs,i= Observed variable (flow in m3s−1or sediment in tons), Xsim,i=Simulated variable

(flow in m3s-1or sediment in tons), Xobs(sim)mean = mean of n values, n = number of observations.

Model calibration generally consists of statistical tests like Optimization of the Nash Sutcliffe
Coefficient (ENS) ( Santhi et al.,2001a; Cotter, 2002; Grizzetti et al.,2003.

The second objective function used in this study was the Nash-Sutcliffe coefficient of efficiency
ENS which has been widely used to evaluate the performance of hydrological models (Leavesley
et al.,1983,Wilcox et al., 1990,Arnold et al., 1999; Kirsch et al., 2002). This coefficient ranges
from minus infinity to 1.0. ENS value of 0.0, means that the model predictions are just as accurate
as using the measured data average to predict the measured data. E NS less than 0.0 indicate the
measured data average is better predictor of the measured data than the model predictions while
with higher values indicating better agreement (Legate and McCabe, 1999). Nash and Sutcliffe
(1970) defined the coefficient as:

 
E NS  1 
n

 i 1 X obs ,i  X sim ,i 2
 
                        3.5
2
 n X mean  
 i 1  obs,i
 X obs 
 

According to Moriasi et al. (2007) model simulation Judged as satisfactory if NSE > 0.5, for flow
and NSE > 0.5, for sediment.

55
Parameter uncertainty accounts for all sources of uncertainties such as uncertainty driving
variables (such as rainfall and temperature), conceptual model, parameters and measured data
(such as discharge and sediment yield).SUFI-2 aggregates all sources of uncertainty to the
parameter ranges. Two indices were used to quantify the strength of a calibration uncertainty
analysis, namely the p factor and the r-factor (Abbaspour et al., 2007). The p-factor is the
percentage of measured data bracket by the 95 % prediction uncertainty (95PPU), The goodness
of calibration and prediction uncertainty is judged on the basis of closeness of the P-factor to 100%
(i.e. all observations bracketed by the prediction uncertainty) . Another measure quantifying the
strength of a calibration or uncertainty analysis is the r-factor which is the average thickness of the
95PPU band divided by the standard deviation of the measured data. The ideal situation would be
to account for 100% of the observed data in the 95PPU while at the same time have an r-factor
close to zero. But this is seldom the case because of measurement errors, conceptual model
uncertainty, and non-uniqueness issues, also r-factor to 1 (i.e. achievement of rather small
uncertainty band). When acceptable values of r-factor and p-factor are reached, then the parameter
uncertainties are the desired parameter ranges.

56
3.7.6 Management Practices Scenario Development

Scenario analysis is the process of evaluating the possible future events by considering alternative
possible outcomes. This analysis is designed to facilitate decision-making and assessment through
a complete consideration of possible outcomes and their implications. The process of development
a scenario is done by using SWAT model apply to simulate the impact of BMPs on sediment
reduction and select this management practices that curb soil erosion. The simulations were carried
out for different scenarios.

3.7.6.1 Selecting Management Practices Options

Arc SWAT model has proven to be an effective tool for evaluating BMP implementation, and for
identification best management practices in the watershed. In this study, the results of simulation
were compiled for interpretation of impact of watershed management on sediment yield.it has been
considered three management practices options to examine the effects of sediment yield in the
watershed. The basic management practices are terracing, contouring and strip-cropping.

Figure 3.23: Ways of selecting management practice options

57
The BMPs were represented in the SWAT model by modifying SWAT parameters to reflect the
effect the practice has on the processes simulated (Bracmort et al., 2006). However, the type of
BMPs and their parameter value selection is site specific and ought to reflect the study area reality.
Thus, appropriate BMPs and their parameter value were cautiously selected based on documented
local research experience in the Ethiopian highlands (Hurni, 1985).

Scenario I (Base Scenario).

This scenario presents without any consideration of management practice condition observe in the
watershed.

Scenario II (Terracing).
As slope length and steepness increase, runoff and soil loss also increase. Slope length can be
changed by installing terraces. A terrace is an embankment within the field design to intercept
runoff and prevent erosion. It is constructed across slope on a contour with several regularly spaced
terraces. There are two basic types of terraces: storage terraces and gradient terraces. Storage
terraces collectwater and store it until it can infiltrate into the ground or be released through a
stable outlet. Gradient terraces are designed as a channel to slow runoff water and carry it to a
stable outlet, such as a grassed waterway. The USLE practice (TERR_P) factor, the slope length
(TERR_SL) and curve number (TERR_CN) are adjusted to simulate the effect of terracing.

Stone terracing soil and water conservation practices at Konso southern Ethiopia

Figure 3.24: Terracing management practice

58
Scenario III (Contouring)

Contour planting is the practice of tilling and planting crops following the contour of the field ass
opposed to straight rows. These contour lines create a water break which reduces the formation of
rills and gullies during times of heavy water rainfall; which is a major cause of soil erosion. The
water break also allows more time for the water to settle into the soil. These contours are oriented
at the right angle to the field slope at any point.small ridges resulting from field operations increase
surface storage and roughness reducing runoff and sediment losses. It is simulated in swat by
altering curve number (CONT_CN) to acount for increased surface storage and infiltration and the
USLE practice factore (CONT_P) to acount for decrease erosion.

Scenario IV (Strip-cropping)

Strip cropping or strip farming is defined as alternating crop rows between heavy-rooted plants
and loosely-rooted plants to minimize erosion. Strip cropping helps to stop soil erosion by creating
natural dams for water, helping to preserve the strength of the soil. it simulated in swat by altering
the manning coeficient value for overland flow (STRIP_N) to represent increased surface
roughness in the direction of runoff. Cureve Number (STRIP_CN) may be adjusted to account for
increased infiltration. USLE Cropping Factor (STRIP_C) may be adjusted to reflect the average
value for multiple crops within the field. The USLE Practice factor may also be updated to
represent strip cropping conditions. The variables which may be entered on the harvest and kill
line are listed and described below. The USLE practice (STRIP_P) factor may also be updated to
represent strip cropping conditions.

59
Table 3. 4: The parameters used to simulate the effect management practices

Best management Arc swat model parameter values

practices Parameters min max Default Used
TERR_P 0 1 0.5 0.5
Terracing TERR_SL 0 100 60 10
TERR_CN 0 100 20 59
COUNT_CN 20 60 60 59
Contouring
COUNT_P 0 1 0.6 0.9
STRIP_N 0.001 0.5 0.15 0.15
STRIP_CN 20 100 60 59
Strip Cropping
STRIP_C 0 1 0.4 0.17
STRIP_P 0 1 0.7 0.8

Hurni (1985) recommended the P factor be adjusted to 0.5 to represent terracing practice
throughout Ethiopia, USLE P value was obtained from documented fieldexperience by
Gebremichael et al. (2005).and C value

Whereas,(USLE N) and CN2 value were assigned SWAT user’s manual version 2005 for contoured
and terraced and strip cropping condition (Neitschet al., 2005).

The SLSUBBSN represents the spacing between successive terracing at field condition i.e 10m and
the modified value was used as reported by Hurni (1985) and Herweg and Ludi (1999).

However, BMPs parameter value selection is site specific and ought to reflect the study area
reality. Appropriate BMPs parameter value based on documented local research experience in the
Ethiopian highlands (Hurni, 1985; Herweg and Ludi, 1999; Gebremichael et al., 2005), because of
those parameter value were field trials of different soil and water conservation technologies to test
a host of parameter values in the Ethiopian highlands.

60
CHAPTER FOUR
4. Result and Discussions
4.1 Sensitivity Analysis

Sensitivity analysis was carried out to identify which model parameter is most important or
sensitive. As said in above sensitivity analysis from SUFI-2 provided partial information about the
sensitivity of the objective function to model parameters.

In our study, 12 water-related parameters (global parameters), and sediment-related parameters

with absolute minimum and maximum ranges in the SWAT model documents were selected to do
sensitivity analysis separately. The sensitivity ranking, a t test is used to identify the relative
significance of each parameter: a t stat provides a measure of sensitivity (larger absolute values
are more sensitive), and p values determine the significance of the sensitivity (a value close to zero
has more significance).

P-Value t-Stat
1:R__CN2.mgt
12:V__RCHRG_DP.…
3:V__GW_DELAY.gw
11:R__SOL_K(..).sol
Parameter Name

8:V__CH_K2.rte
6:V__ESCO.hru
2:V__ALPHA_BF.gw
7:V__CH_N2.rte
13:V__EPCO.bsn
4:V__GWQMN.gw
5:V__GW_REVAP.…
10:R__SOL_AWC(..…

0.00 0.50 1.00 -5.00 0.00 5.00 10.00 15.00

Figure 4.1: Sensitivity analysis of flow

Note: the t Stat provides a measure of sensitivity (larger absolute values are more sensitive); the
p value determines the significance of the sensitivity (a value close to zero has more

61
significance); “R_” and “V_” means relative change and a replacement to the initial parameter
values, respectively;

In terms of water process, the most sensitive parameter was found to be a curve number (CN2),
followed by deep aquifer percolation fraction (RCHRG_DP), GW_DELAY (ground water delay
time: lag between the time that water exits the soil profile and enters the shallow aquifer),SOL_K
(soil saturated hydraulic conductivity), channel effective hydraulic conductivity (CH_ K2),soil
evaporation compensation factor (ESCO), ALPHA_BF (the base flow alpha factor) ,the Manning’s
‘n’ value for tributary channels (CH_N2), EPCO (plant uptake compensation factor).Those
parameters were checked their sensitivity for further calibration and validation water process, those
are CN2, RCHRG_DP, GW_DELAY, SOL_K, CH_K2, ESKO, ALPHA_BF, CH_N2, EPCO etc.
And their rank of sensitivity from the most to least in the above Figure 4.1.

On the other hand ten parameters were found to be sensitive for sediment simulation, based on
sensitivity ranking a t test and its p values significance of the sensitivity shown Figure 4.2.

P-Value t-Stat
1:V__USLE_P.mgt
2:V__SLSUBBSN.hru
3:V__USLE_C{..}.pla…
Parameter Name

8:V__CH_N2.rte
6:V__SPEXP.bsn
9:V__CH_K2.rte
10:R__PSP.bsn
4:V__SPCON.bsn
5:V__CH_COV2.rte
7:V__CH_COV1.rte
-30.00-25.00-20.00-15.00-10.00 -5.00 0.00 5.00
0.00 0.40 0.80

Figure 4.2: Sensitivity analysis of sediment yield

Those three parameters are most sensitive parameters were found to be the USLE support practice
factor (USLE-P), SLSUBBSN (average slope length), the agricultural land cover factor in the
USLE equation (USLE-C_AGRC), followed by the Manning’s ‘n’ value for tributary channels

62
(CH_N), SPEXP (exponent parameter for calculating sediment re-entrained in channel sediment
routing) explain the channel erosion and sediment re-entrainment, channel effective hydraulic
conductivity (CH_ K2), sediment routing, factor in main channel (PSP), SPCON (linear parameter
for calculating the maximum amount of sediment that can be re-entrained during channel sediment
routing), CH_COV2 (channel cover factor), the channel erodibility factor (CH_COV1), and others.

Those flow and sediment parameters were adjusted from the SWAT initial estimates to fit the
model simulations with the observed flow and sediment data. These parameters and their calibrated
values are displayed in Table 4.1.

Table 4. 1: The final ranges and fitting values of the SWAT model parameters resulted from the
final calibration procedure

Fitting Final parameter range

parameter
Parameters Name Description value Min-value Max-value
R__CN2.mgt Curve number 0.146614 -0.2 0.2
V__ RCHRG_DP.gw Recharge to deep aquifer 0.129482 0 1
V__ GW_DELAY.gw Groundwater delay time 81.035858 30 450
Saturated hydraulic
R__SOL_K(...).sol conductivity -0.401594 -0.8 0.8
Channel effective hydraulic
V__CH_K2.rte conductivity 126.264946 5 130
Soil evaporation
V__ ESCO.hru compensation factor 0.825100 0.8 1
V__ ALPHA_BF.gw Baseflow alpha factor 0.754980 0 1
Universal soil loss equation
V__ USLE_P.mgt support practice factor 0.033865 0 1

V__ SLSUBBSN.hru Average slope length 16.533865 0 100

V__USLE_C
{AGRC}.plant .dat USLE land cover factor 0.383699 0.001 0.5
Manning’s “n” value for
V__CH_N2.rte main channel 0.219323 -0.01 0.3
Exponent of re-entrainment
parameter for channel
V__ SPEXP.bsn sediment routing 1.331673 1 1.5

63
4.2 Model Calibration and Uncertainty Measures

Model calibration followed sensitivity analysis by considering those parameters. Calibration

involves testing the model with known input and output data in order to adjust some parameters,
while validation involves comparison of the model results with an independent dataset during
calibration without any further adjustment of the calibration parameters. Model calibration and
validation using SUFI-2 Al- algorithm, flow predictions were calibrated using 1994 to 2001 and
validated using 2002 to 2004 monthly flow data. Likewise, sediment calibration was carried out
for the years 1994, 2001 and validated from 2002 to 2004.

After calibrating for flow simulation was executed and the hydrographs are well captured. The
agreement between the measurement and simulation is generally very good, which are verified by
NSE and R2 an acceptable result were obtained according to the model evaluation guideline
(Moriasi et al., 2007). The results of these tests illustrated that the monthly coefficient of
determination and Nash- Sutcliffe coefficient was 0.78 and 0.77 for calibration period, 0.73, and
0.72 for validation period respectively Table 4.2.

The uncertainty of the calibrated model in SUFI-2, 95PPUs, is the combination of the uncertainties
in the input data, model structure, model parameters, and the measured data (which was not
separately evaluated). The uncertainty was represented by the p-factor and the r-factor. In terms of
monthly stream flow, the p-factor and the r-factor was 69 % and 0.64, 66% and 0.60, respectively,
for calibration and validation. This indicated about 69 % (66 %) (Out of a perfect 100 %) of the
measured monthly stream flow could be bracketed by the 95PPU with a very narrow 95PPU band
of 0.64 (0.60) (close to a perfect 0) in the calibration and validation period respectively shown in
Table 4.2.

64
The calibration and validation period of the model was eleven years from 1994 to 2004 (Figure
4.3) and (Figure 4.4) respectively. The result of calibration and validation for monthly flow
hydrograph showed that there is a good agreement between the measured and simulated monthly
flows.

250

200
Monthly Flow(m3\s)

150 95PPU

100 observed

Best_Sim
50

0
1994/1 1995/1 1996/1 1997/1 1998/1 1999/1 2000/1 2001/1

Figure 4.3: Hydrograph of the observed and simulated flow from the watershed for the
calibration period on a monthly basis

140
120
Monthly Flow(m3\s)

100
95PPU
80
observed
60
Best_Sim
40
20
0
2002/1 2003/1 2004/1

Figure 4.4: Hydrograph of the observed and simulated flow from the watershed for the validation
period on a monthly basis

In terms of monthly Sediment simulation, the agreement between the measurement and simulation
is generally very good, which are verified by NSE and R2 was 0.74 and 0.76 for calibration period,
0.70, and 0.71 for validation period respectively Table 4.2

65
The p-factor and the r-factor were 65 % and 0.61, 67 % and 0.59, respectively, for calibration and
validation. This indicated about 65 % 67% (close to a perfect 100 %) of the measured monthly SY
could be bracketed by the 95PPU with a 95PPU band of 0.61 (0.59) ( close to a perfect 0) in the
calibration (validation) period. Usually a higher p-factor will cause a higher r-factor. Behavioral
solutions can reduce the p-factor and the r-factor, and get a smaller prediction uncertainty in the
SUFI-2.

The time series plot of measured and simulated at monthly sediment basis for calibration and
validation period has a good agreement. Even though, it slightly over predicted the sediment on
the rising limb and the receding limb in the calibration period (Figure 4.5) in 1994 and 1999.The
model the also under predicted the peak concentrations of sediment both on the rising and the
falling limbs of the sediment hydrograph in the validation period (Figure 4.6). Generally we can
say the model well simulated sediment concentrations even if there is a gap in the observed
suspended sediment.

3000000

2500000
Monthly SY(ton)

2000000
95PPU
1500000
observed
1000000
Best_Sim
500000

0
1994/1 1995/1 1996/1 1997/1 1998/1 1999/1 2000/1 2001/1

Figure 4.5: Comparison of observed and simulated sediment loss from the watershed for the
calibration period on a monthly basis.

66
1600000
1400000
1200000
Monthly SY(ton)

1000000
95PPU
800000
observed
600000
Best_Sim
400000
200000
0
2002/1 2003/1 2004/1

Figure 4.6: Comparison of observed and simulated sediment loss from the watershed for the
validation period on a monthly basis.
Table 4. 2: Performance and uncertainty for flows and sediment in calibration and validation
period.

Uncertainty value
Period simulation
R2 ENS
p-factor r-factor

Calibration (1994-2001) 69% 0.64 0.78 0.77

Stream flow
Validation (2002 – 2004) 66% 0.60 0.73 0.72

Calibration (1994-2001) 65% 0.61 0.76 0.74

Sediment

Validation (2002 – 2004) 67% 0.59 0.71 0.70

As can be indicated in the Table 4.2, the model performance values for calibration and validation
of the flow and sediment simulations are adequately good. This indicates that the physically
processes involved in the generation of stream flows as well as sediment in the watershed were
adequately captured by the model. The p-factor and r-factor calculated result have given good
agreement by bracketing >50% observed data for both calibration and validation on flow as well
as for sediment yield and the r-close to zero has shown the parameter uncertainties are the closed
to desired parameter ranges.

67
4.3 Spatial Variation of Soil Erosion

The spatial distribution of sediment generation for the Gumara watershed is presented in figure
below. The spatial distribution of sediment indicated that, out of the total 25 SWAT sub basins, 12
sub-basins produce average annual sediment yields ranging from 15-31.5 ton/ha/yr, while most of
the low land areas are in the range of 0-14 ton/ha/yr.

Figure 4.7: Spatial distribution SWAT simulated annual sediment yield classes by sub basin
(t/ha/yr)

4.4 Analysis of Results

4.4.1 Scenarios

Each BMP has a different effect on flow and sediment variables, and hence represented by distinct
parameter(s) in SWAT. Table 4.4 shows parameters used to represent BMPs in for this study.
Comparing to base simulation, those scenarios such as Scenario II (Terracing), Scenario III
(Contouring) and Strip-cropping (scenario IV) produce essentially different monthly sediment
yield amount see in Fig (5.8).Scenario II is observed to reduce the sediment from the watershed
and found to give the lowest sediment yield compared to the other scenarios.

68
35
30
Sediment yield t_ha

25 Base Scenario
20 Terracing
15 Contouring

10 Strip Croping

5
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

Sub Watershed (SW)

Figure 4.8: Comparison of average monthly sediment yield in each scenario in sub watersheds

4.4.2 Identification and prioritization of Critical Sub Watersheds

The different scenarios were compared and analyzed on a monthly basis of sediment yield from
the watershed were analyzed. As described earlier the scenarios were compared on the watershed
based on Sub Watershed (SW) basis for the base scenario simulated years.

Table 4. 3: Area under different classes of soil erosion by water in the given sub watersheds

Soil erosion
Classes Slight Very Low Low Moderate High Very high Severe Very severe
Soil erosion
range(t/ha/yr) 0-1 1-5 5-8 8-10 10-15 15-25 25-31 >31

Sub Watershed 7,17 4,10,17 9,16,20 23 22 1,2,5,11, 3,6,14 13,18

(SW) 24,25 12, 15,21,19

Around 83% of the mean annual sediment yield is from (SW) 22,1, 2,5,11,12,15,21,19,3,6,14,13
and 18. The sub watershed 7, 17 are under the slight soil erosion class from the watershed. Out of
the 25 sub watersheds, the WS13 (31.2 t/ha/yr) and WS18 (31.5 t/ha/yr) falls under the Very severe
soil loss group of soil erosion classes (>31 t/ha/yr).

In view of the annual soil loss, the sub watersheds SW18 were selected to adopt the management
measures in sequential order as it appear to reduce the sediment with in the watershed for
sustainable.

69
35
Sedment yield t_ha
30
25
20
15
10
5
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Sub Watershed (WS)

Figure 4.9: Comparison of average monthly sediment yield in sub watersheds

Management Practice for the Critical Sub Watersheds.

A sample SW18 was selected as described earlier to examine the effect of management practices.
Those three management practice were consider to examine the effects.

The effect of those three management practices on average annual sediment yield for sub
watershed (SW18) is below.

35
Sediment yield (ton/ha/year)

30
25
20
15
10
5
0
Base Scenario Terracing Contouring Strip Croping
Best Management Pratices (BMPs)

Figure 4.10: Comparison of average sediment from the critical sub watershed in each scenario

70
CHAPTER FIVE
5. Conclusions and Recommendations
5.1 Conclusion
The simulation of sediment yield with different BMPs in the watershed were performed on the
selected, SWAT model. SWAT CUP advance calibration and uncertainty analysis tool has been
used for Sensitivity analysis, automatic calibration and validation of stream-flow and sediment
yield on monthly basis for the period 1994–2004 using SUFI-2. Sensitivity analysis also done
using SUFI-2 provided partial information about the sensitivity of the objective function to model
parameters. A t and p test is used to identify the relative sensitivity as well as the significance of
each parameter. It provides an effective graphical interface for visualization of outputs, including
simulated data, observed data, best-fit model results and 95PPU for all variables used in model
calibration. The sensitivity analysis adopted for the stream-flow and sediment calibration shows
variations between the parameter ranges which had been initialized for the model calibration on
monthly basis. The p-factor and r-factor calculated using SUFI-2 procedures have given good
agreement by bracketing >50% observed data for both calibration and validation on flow as well
as for sediment yield. The Nash Sutcliff efficiency and coefficient of determination for monthly
flow was 0.77 and 0.78 and also for monthly sediment 0.74 and 0.76 for calibration period
respectively. For validation period and 0.72 and 0.73 on flow and sediment yield of 0.71 and 0.70
respectively. This indicates that the model is well performing the trend of flow and sediment in
both calibration and validation period and the model result during validation is comparable to those
of calibration period.

The SWAT model was applied to model spatially distributed soil erosion/sedimentation processes
at monthly time step and to assess the impact of three Best Management Practices (BMPs)
scenarios on sediment reductions in Gumara watershed. The model showed that the erosion prone
area at Sub watershed (WS) and Hydrological Response Units (HRUs) level, which is already
useful information for catchment management planning. For existing conditions (Base scenario),
a reasonable agreement was obtained between the model sediment yields at the watershed outlet.
And the result Sub-watershed 18 was found to be more critical pertaining to sediment yield. The
simulation results showed that from the model output of sediment yield comparing to the base

71
scenario Terracing (Scenario II) has 93%, Scenario III (Contouring) 65% and Strip-cropping
(scenario IV) 60% of sediment reduction from the base scenario. The rate of reduction in sediment
yield for different scenarios was used to identify appropriate best management practices for the
watershed. On the basis of the percent reduction results, it is concluded that terracing is one of the
best management practice option in order to minimize the sediment losses. These results indicate
that applying BMPs could be effective in reducing sediment transport for sustainable water
resources management in the watershed. However, any implementation of watershed management
measures to reduce sediment yields involves the use of resources and willingness of decision
makers. Thus, it is concluded that, a physically-based SWAT model can be used not only to
identify the critical areas but also to explore the capabilities of the model to identify the best
management practices. If discussed with the community leaders about implementation of the
structures, making them aware about the efficient working of the scenarios, they will certainly
participate in the process of watershed management to make their communities self-sustained in
water needs.

72
5.2 Recommendations

Generally from this specific study the following recommendations could improve similar research
for future work: The future research should have concerned to

 The model simulations only considered those three scenarios including the base scenario
in this study area by assuming all other thing constant, but change in climate and other
management practice options and also has to consider at a local scale level for sustainable
water resources management in the watershed, and other land use variables will contribute
great impact on rainfall runoff process and sediment yield of the watershed.
 In addition, the farmers may need to change their practices to adapt climate change. If
implementing the management options is planned, considering the impact of climate
change on these management options, formulated in this study, is very important.
 Management practice were treated using sediment rating curve because of limitation of
sediment load data, this may create the problem with the efficient working of the scenarios.
Generally in order to improve the model performance, the weather stations should be
improved in both quality and quantity. Hence, it is highly recommended to establish a good
network of both hydrometric and meteorological stations.

73
References

Abbaspour, K. C., Yang, J., Maximov, I., Siber, R., Bog-ner, K., & Mieleitner, J. (2007).
Spatially Distributed Modelling of Hydrology and Water Quality in the Prealpine/Alpine
Thur Watershed Using SWAT. Journal of Hydrology, Vol. 333, No. 2-4,, pp. 413-430.

Abegaz, G. (1995). Soil erosion assessment: Approaches, magnitude of the problem and issues
on policy and strategy development (Region 3). Paper presented at the Workshop on
Regional Natural Resources Management Potentials and Constraints, Bahir Dar,
Ethiopia.

Arnold, J. G., Srinivasan, R., Muttiah, R. S., & Williams, J. R. (1998). Large area hydrologic
modeling and assessment Part I. Model development. Journal of the American Water
Resource Association34(1), 73– 89.

Arnold, J. G., Williams, J. R., Srinivasan, R., & King, K. W. (1996). SWAT: Soil and Water
Assessment Tool. User’s Manual USDA Agriculture Research Service Grassland, Soil
and Water Research Laboratory,808 East Blackland Road Temple, TX 76502,, 190.

Baigorria, G. A., & Romero, C. C. (2007). Assessment of erosion hotspots in a watershed

Integrating the WEPP model and GIS in a case study in the Peruvian Andes.
Environmental Modelling & Software 22, 1175-1183.

BCEOM. (1998). bbay River Basin Integrated Development Master Plan Project Volume Xiii –
Environment. Ministry of Water Resource, Addis Ababa.

Bekele, S., & Holden, S. T. (1998). Resource degradation and adoption of land conservation
technologies in the Ethiopian Highlands. A case study in Andit Tid, North Shewa,
Agricultural Economics 18: 233-47.

Bennett, H. H. (1939). Soil Conservation. Inc., New York. USA: McGraw-Hill Book Company.

Beven, K. (1992). Changing ideas in hydrology—the case of physically-based models. Journal

of hydrology 105, 157–172.

Biniam. (2009). SWAT TO IDENTIFY WATERSHED MANAGEMENT OPTIONS:(ANJENI

WATERSHED, BLUE NILE BASIN, ETHIOPIA).

Bracmort, K., Arabi, M., Frankenberger, J., Engel, B., & Arnold, J. (2006). Modeling long-term
water quality impact of structural BMPs . T. ASABE, 49, 367–374.

74
Constable, M. (1984). Resource of Rural Development in Ethiopia. Ethiopian High lands
Reclamation Study. FAO/Ministry of Agriculture, Addis Ababa.

De Asis, A. M., & Omasa, K. (2007). Estimation of vegetation parameter for modeling soil
erosion using linear Spectral Mixture Analysis of Landsat ETM data. ISPRS Journal of
Photogrammetry & Remote Sensing 62, 309–324.

Decoursey, D. G., & Selly, E. H. (1988). Mathematical models for point sources water pollution
control. Journal of Soil and Water Conservation. 44(2): 568-576.

Deore, S. J. (2005). Prioritization of Micro-watersheds of Upper Bhama Basin on the Basis of

Soil Erosion Risk Using RemoteSensing and GIS Technology. PhD thesis, University of
Pune, Pune.

Fangmeier, D. D., Elliot, W. J., Workman, S. R., Huffman, R. L., & Schwab, G. O. (2006). In T.
D. Learning, Soil and Water Conservation Engineering, 5th Edition. Printed in USA.

Foster, J. R., & Meyer, L. D. (1977). “Soil erosion and sedimentation by water-an overview”.
Procs National Symposium on Soil Erosion and Sedimentation by Water, Am. Soc. Of
Agric.Eng., St. Joseph, Michigan, 1-13 .

Garzanti, E., Ando, S., Vezzoli, G., Ali Abdel Megid, A., & El Kammar, A. (2006). Petrology of
Nile River sands (Ethiopia and Sudan): Sediment budgets and erosion patterns. Earth and
Planetary Science Letters, 252 (3-4).

Gebremichael, D., Nyssen, J., Poesen, J., Deckers, J., Haile, M., Govers, G., & Moeyersons, J.
(2005). Effectiveness of stone bunds in controlling soil erosion on cropland in the Tigray
highlands, Northern Ethiopia , Soil Use. Manage., 21, 287–297.

Green, W. H., & Ampt, G. A. (1911). Studies on soil physics; The flow of air and water through
soils. Journal of Agricultural Sciences; 4, 11-24.

Grizzetti, B., Bouraoui, F., Granlund, K., Rekolainen, S., & Bidoglio, G. (2003). Modelling
Diffuse Emission and Retention of Nutrients in the Vantaanjoki Watershed (Finland)
Using the SWAT Model. Ecological Modelling , 169(1):25-38.

Harmel, R. D., & Smith, P. K. (2007). Consideration of Measurement Uncertainty in the

Evaluation of Goodness-of-fit in hydrologic and Water Quality Modeling. Journal of
Hydrology, 337, 326-336.

Hermance. (2003). Evaluating the impact of land use land cover change.

75
Herweg, K., & Ludi, E. (1999). The performance of selected soil and water conservation
measures-case studies from Ethiopia and Er-itrea. Catena, 36, 99–114.

Hurni, H. (1985). Erosion – productivity – conservation systems in Ethiopia, in: Proceedings of

the 4th International Conference on Soil Conservation. Maracay, Venezuela, 654–674.

Hurni, H. (1988). Degradation and Conservation of the Resources in the Ethiopian highlands.
Mountain Research and Development 8(2/3).

Israel, T. (2011). Soil Erosion Risk Assessment with RUSLE and GIS in Dire Dam . 44.

Julien, P. Y. (1998). “Erosion and sedimentation”. Cambridge University Press.

Lal, R. (2001). Soil degradation by erosion. Land Degradation & Development, 12, 519 – 539.

Leavesley, G. H., Lichty, R. W., Troutman, B. M., & Saindon, L. G. (1983). Precipitation-runoff
modeling system user’s manual. Sur. Water , (pp. Rep. 83-4238, 207 pp). U.S. Geol.

Mequanint, T. (2008). SWAT BASED RUNOFF AND SEDIMENT YIELD MODELING

(ACASE STUDY OF GUMARA WATERSHED LAKE TANA SUB
BASIN),ETHIOPIA.

Mitchell, J. K., & Bubnezer, G. D. (1980). Soil loss estimation. In I. M. Sons. Chichester,
England.

Mitiku, H., Herweg, K., & Stillhardt, B. (2006). Sustaiable land management –A new approach
to soil and water conservation in Ethiopia. Mekelle, Ethiopia. Land Resource
Management and Environmental Protection Department, Mekelle university ; Bern,
Switzerland: Center for Development and Environment(CDE), University of Bern and
Swiss National Center of Competence in Research (NCCR) North-South.

MoARD. (2004). Woody Biomass Inventory and Strategic Planning Project A national strategic
plan for Biomass energy sector. final report, Addis Ababa, Ethiopia.

Moriasi, D., Arnold, J., & & Van Liew, M. W. (2007). Model evalution guidline for systematic
quanification of accuracy in watershed simulation.Multi variable catchement,. 885-900.

Mostaghimi, S., Wang, Y., & McClellan, P. W. (1991). Application of Agricultural Non-Point-
Source (AGNPS) pollution model to the Owl Run watershed,Report No.O-S1-9109.
Virginia Department of Conservation and Historic Resources, Division of Soil and Water
Conserva- tion, Richmond Virginia. .

76
Neistch, J. R., Arnold, J. G., Kiniry, J. R., Srinivasan, R., & Williams, J. R. (2004). Input/Output
File Documentation Version 2005. Retrieved from Grassland, Soil and Water Research
Laboratory, Agricultural Research Service, Blackland Research Center, Texas
Agricultural Experiment Station, Temple,Texas 76502.

Neitsch, S. L., Arnold, J. G., Kiniry, J. R., & Williams, J. R. (2002). Soil and Water Assessment
Tool User’s Manual, Version 2000. Grassland, Soil and Water . Assessment Tool User’s
Manual, Version 2000. Grassland, Soil and Water Research and Extension Center,
Temple, Texas BRC Report 02-06. Texas Water Resources Institute, College Station,
Texas TWRI Report TR-192.

Neitsch, S. L., Arnold, J. G., Kiniry, J. R., & Williams, J. R. (2005). Soil and Water Assessment
Tool. Theoretical Documentation Version 2005. Temple, TX. USDA.

Neitsch, S. L., Arnold, J. G., Kiniry, J. R., & Williams, J. R. (Version 2005). Soil and Water
Assessment Tool Theoretical Documentation. Retrieved from Temple, TX. USDA
Agricultural Research Service and Texas A&M Blackland Research Center.

petter, p. (1992). Gis and Remote senccing for Soil Erosion in Semi-arid Environment.phD.
University of Lund.

Renard, K. G., Foster, G. R., Weesies, G. A., McCool, D. K., & Yoder, D. C. (1997). Predicting
soil erosion by water: a guide to conservation planning with the Revised Universal Soil
Loss Equation (RUSLE). In A. R. Service, Agriculture Handbook No. 703. Washington,
DC.

Renschler, C. (1996). Soil erosion hazard mapping by means of Geographical Information

systems (GIS) and Hydrological modelling. MSc,. Technical University of Braunschweig
Braunschweig.

Reungsang, P., Kanwar, R. S., Jha, M., Gassman, P. W., Ahmad, K., & and Saleh, A. (2005).
Calibration and validation of SWAT for the Upper Maquoketa River Watershed. Center
for Agricultural and Rural Development, Iowa State University, Ames, Iowa 50011-1070,
Working Paper 05-WP 396.

Saavedra, C. (2005). Estimating spatial patternsof soil erosion and deposition in the Andean
region using geo-information techniques: a case study in Cochabamba. Bolivia
Ph.D.dissertation, Wageningen University The Netherlands.

Santhi, C., Arnold, J. G., Williams, J. R., Hauck, L. M., & Dugas, W. A. (2001a). Application of
a Watershed Model to Evaluate Management Effects on Point and Nonpoint Source

77
Pollution. Transactions of the American Society of Agricultural Engineers, 44(6):1559-
1570.

Setegn, S. G., Srinivasan, R., & Dargahi, B. (2008). Hydrological Modeling in the Lake Tana
Basin, Ethiopia using SWAT model. The Open Hydrology Journal Vol 2, 49-62.

Tebebu.T. (2010). Surface and subsurface flow effect on permanent gully formation and upland
erosion near Lake Tana in the Northern Highlands of Ethiopia. Hydrol. Earth Syst. Sci.
Discuss., 7, 5235–5265, 2010 www.hydrol-earth-syst-sci.

Tripathi, M. P., Panda, R. K., & Raghuwanshi, N. S. (2003). Identification and Prioritisation of
Critical Sub-watersheds for Soil Conservation Management using the SWAT Model. Soil
and Water Engineering Department, Faculty of Agricultural Engineering, Indira Gandhi
Agricultural University.

Tripathi, M. P., Panda, R. K., & Raghuwanshi, N. S. (2004). Development of effective

management plan for critical subwatersheds using SWAT model. Department of Soil and
Water Engineering, Faculty of Agricultural Engineering, Indira Gandhi Agricultural
University, Raipur 492 006.

USAID. (2000). Amhara National Regional State food security research assessment report.
Available at:http://crsps.org/amhara/amhara_rpt.PDF.

Vrieling, A. (2006). In Satellite remote sensing for water erosion assessment (pp. 2 – 18). a
review. Catena65.

White, E. D., Easton, Z. M., Fuka, D. R., Collick, E. S., Adgo, E., & McCartney, M. (2008).
Adapting the soil and water assessment tool (SWAT) for the Nile Basin (unofficial data
describes that 70% of the cost of operation and maintenance in the Blue Nile part of
Sudan is spent on sediment related and canal maintenance).

White, K. L., & Chaubey, I. (2005). Sensitivity Analysis, Calibration and Validations for a
Multisite and Multivariable SWAT Model. Journal of the American Water Resources
Association (JAWRA), 41(5):1077-1089. .

Wilcox, B. P., Rawls, W. J., Brakensiek, D. L., & Wright, J. R. (1990). Predicting Runoff from
Rangeland Catchments: A comparison of two models. Water Resource , Research, (26),
2401 – 2410.

Williams, J. R., & Berndt, H. D. (1997). Sediment yields prediction based on watershed
hydrology. Trans. ASAE 20 (6), 1100–1104.

78
Wischmeier, W. H., & Smith, D. D. (1978). “Predicting rainfall erosion losses –A guide to
conservation planning.”. U.S. Department of Agriculture handbook, No. 537.

WWDSE. (2007). Catchment Development Plan, Gumara Irrigation Project. Ministry of water
resources. Addis Ababa.

79
Appendix

Appendix I. Symbols and description of Weather Generator Parameters

(WGEN) used by the SWAT Model

S. No Symbol Description
1 TMPMX Average or mean daily maximum air temperature for month (⁰C).
2 TMPMN Average or mean daily minimum air temperature for month (⁰C).
3 TMPSTDMX Standard deviation for daily maximum air temperature for month (⁰C).
4 TMPSTDMN Standard deviation for daily minimum air temperature for month (⁰C).
5 PCPMM Average or mean total monthly precipitation (mm H2O).
6 PCPSTD Standard deviation for daily precipitation for month (mm H2O/day).
7 PCPSKW Skew coefficient for daily precipitation in month.
8 PR_W1 Probability of a wet day following a dry day in the month.
9 PR_W2 Probability of a wet day following a wet day in the month.
10 PCPD Average number of days of precipitation in month.
11 SOLARAV Average daily solar radiation for month (MJ/m2/day)
12 DEWPT Average daily dew point temperature in month (⁰C).
13 WNDAV Average daily wind speed in month (m/s)
Appendix II. Soils Parameters and Legend used in SWAT Model

NLAYERS Number of layers in the soil (min 1 max 10)

HYDGRP Soil hydrographic group (A, B, C, D)
SOL_ZMX Maximum root depth of the soil profile
ANION_EXCL Fraction of porosity from which an ions are exchanged
SOL_CRK Crack volume potential of soil
TEXTURE Texture of the layer
SOIL_Z Minimum depth from soil surface to bottom of layer
SOL_BD Moist bulk density
SOL_AWC Available water capacity of soil surface to bottom of the layer
SOL_K Saturated hydraulic conductivity
SOL_CBN Organic carbon content
CLAY Clay content
SILT Silt content
SAND Sand content
ROCK Rock fragmented content
SOL_ALB Moist soil albedo
USLE_K Soil erodibility factor (K)

80
Appendix III. Soils Parameter Values used in SWAT Model

ANION_EXCL(M

SOL_BD(g\cm3)

SOL_K(MM\hr)
SOL_ZMX(mm)

SOL_CBN (%)
SOL_CRK (%)

SOL_Z(mm)

SOL_AWC

SAND (%)
TEXTURE
NLAYERS

SOL_ALB
CLAY (%)
HYDGRP

SILT (%)

USLE_K
ROCK
Code
Major Soil

M)
Type

1 SiL 200 1.45 0.11 7 0.5 25 31 44 0.01 0.13 0.25

2 CL 260 1.46 0.11 37.2 0.3 14 66 20 0 0.13 0.25
3 CL- 460 1.45 0.1 34.8 0.21 19 59 22 0 0.13 0.25
Chromic
LVx 4 B 1800 0.01 0.01 CL- 650 1.49 0.1 33.6 0.2 22 56 22 0 0.13 0.25
Luvisols
5 C 9500 1.48 0.1 36 0.2 17 57 26 0 0.13 0.25
6 C 1350 1.49 0.1 36 0.12 17 57 26 0 0.13 0.25
7 C 1800 1.47 0.1 36 0.1 16 59 25 0 0.13 0.25
1 LS 200 1.1 0.11 25 2 50 34 17 5 0.13 0.3
2 LS 500 1.04 0.11 25 2.3 50.8 22 27.1 0.01 0.13 0.3
EutricFluvisols Fle 3 B 1700 0.01 0.01 LS 900 1.05 0.12 25 2.5 38.6 40 21.1 0.01 0.13 0.3
4 LS 1300 1.3 0.95 25 0.2 38.8 34 29.1 0.01 0.13 0.3
5 LS 1700 1.04 0.1 60 0.42 58.8 30 11.1 0.01 0.13 0.3
1 C 200 1.1 0.11 25 2 50 34 7 5 0.13 0.22
Lpe C 650 0.01 0.03
EutricLeptosol 2 C 650 1.23 0.1 13 1.1 66 14 17 0.01 0.13 0.22
1 C 250 1.08 0.12 6.8 1.7 54 26 21 0 0.09 0.15
2 C 363 1.27 0.11 4.54 1.37 611 19 21 0 0.09 0.15
3 C 847 1.28 0.1 5.16 1.41 63 17 20 0 0.09 0.15
EutricVertisols Vre 4 D 2422 0.01 0.3 C 1029 1.22 0.1 4.24 0.88 63 8 29 0 0.09 0.15
5 C 1392 1.13 0.1 4.34 1.17 63 9 28 0 0.09 0.15
6 C 1635 1.1 0.11 4.24 1.24 60 13 27 0 0.09 0.15
7 C 2422 1.1 0.09 4.04 0.34 64 17 20 0 0.09 0.15
1 CL 200 1.45 0.11 30 0.5 25 31 44 0.01 0.13 0.25
HaplicLuvisols LVh B 600 0.01 0.01
2 C 600 1.37 0.09 5.52 0.5 23 33 44 0.01 0.13 0.25
Urban UR 1 1524 0.5 0.5 VAR 1524 2.5 0.01 180 0 15 30 55 20 0.23 0

`81
Appendix IV: Parameters used for Weather Generator in SWAT Model

TMPMX TMPMN TMPSTDMX TMPSTDMN PCPMM PCPSTD PCPSKW PR_W1 PR_W2 PCPD RAINHHMX SOLARAV DEWPT WNDAV
27.49 8.79 1.80 2.36 2.25 0.60 10.47 0.02 0.30 0.87 51.10 21.18 9.80 0.65
29.10 10.64 2.17 2.51 3.03 1.06 13.48 0.02 0.25 0.87 45.98 22.52 9.70 0.71
30.21 12.68 2.06 3.02 9.10 2.50 18.06 0.06 0.32 2.61 42.45 22.90 9.71 0.88
30.48 14.89 2.09 2.91 31.32 4.16 5.46 0.10 0.44 4.61 43.27 23.46 10.61 0.98
29.53 15.51 2.43 1.93 73.52 6.78 4.80 0.21 0.51 9.91 52.43 21.90 13.09 0.90
27.30 14.75 2.00 4.13 197.72 11.01 2.66 0.53 0.71 19.83 66.65 19.58 15.44 0.93
24.58 14.28 1.51 1.32 424.12 15.60 2.19 0.82 0.92 29.30 76.47 16.54 15.80 0.75
24.51 14.15 1.71 2.37 389.36 15.32 2.50 0.84 0.89 28.22 82.24 16.34 15.98 0.70
25.72 13.49 1.32 1.40 198.58 10.42 2.76 0.55 0.76 21.83 72.18 19.03 15.33 0.68
26.67 13.58 1.13 1.56 93.15 7.09 3.23 0.22 0.56 11.22 63.59 21.36 13.94 0.73
27.24 11.24 1.28 2.37 13.42 2.28 6.78 0.05 0.30 2.43 56.70 21.12 11.88 0.66
27.11 9.09 1.39 2.52 2.68 0.73 10.71 0.03 0.17 1.00 53.43 20.50 10.32 0.60

82