Revised-Thesis_Final_Haimanote_formatted-22

This study focuses on characterizing subsurface water flow and ground water table fluctuations in response to rainfall that leads to saturation excess runoff, the basicprinciple of varia

MODELING RAINFALL-RUNOFF RELATIONSHIP AND ASSESSING IMPACTS

OF SOIL CONSERVATION RESEARCH PROGRAM INTERVENTION ON SOILPHYSICAL AND CHEMICAL PROPERTIES AT MAYBAR RESEARCH UNIT,

WOLLO, ETHIOPIA

A ThesisPresented to the Faculty of the Graduate School

of Cornell University

in Partial Fulfillment of the Requirements for the Degree of

Masters of Professional Studies

ByHaimanote Kebede Bayabil

August 2009

© 2009 Haimanote Kebede Bayabil

This study focuses on characterizing subsurface water flow and ground water table fluctuations in response to rainfall that leads to saturation excess runoff, the basicprinciple of variable source area hydrology In particular, this study concentrates to develop a model that efficiently simulates the location of saturated runoff areas and predict river discharge, which finally could help in realistic planning of watershed interventions Furthermore, the study assesses the impact of soil conservation research program intervention on selected physical and chemical soil properties of the study area Long-term discharge and rainfall data was available at the watershed outlet and for four test plots In addition, 29 piezometers were installed in 2008 and water table measurements were taken during the main rainy season Based on major runoff

mechanisms identified at the catchment-level, a conceptual rainfall-runoff model was developed to compute runoff The model incorporates saturated excess overland flow from both bottomlands and subsoil exposed areas and baseflow and interflow from thehillsides The model was tested on a daily, weekly, and monthly basis and fitted well the discharge data at the bottom of the watershed In addition, the distributed model output agreed well with the ground water table measurements The watershed was saturated (and produced runoff) in the flat areas near the river while the hillsides were unsaturated with a perched water table that responded rapidly to rainfall Data from test plots showed that flatter areas produced more runoff than test plots at steeper slope areas The model has potential to predict runoff in ungauged basins but should

be further tested to do so On the other hand, soil samples were tested for selected physical and chemical properties The result indicated that AP and % OC contents of the soil were found in lower amount than before/early project intervention period, while the Db value has shown an increase

BIOGRAPHICAL SKETCH

Haimanote Kebede was born and raised in East Gojjam, in 1981 After

successfully completing his secondary school study at Debre Markos C.S.S School,

he joined Alemaya Universiy in 1999 During his four-year stay at university, he studied plant science and received his Bachelor degree in June 2003

In September 2003, he was employed by Elfora Agro Industries P.L.C where

he served as Junior Agricultural Expert for eight months Then in May 2004, he moved to Finchaa Sugar Factory as a Plantation Section Manager and was assigned to manage a plantation section, including the 1300-hectare farm cultivated under

irrigation and all the working staff under this section

In April 2005, he had an opportunity to participate in a training entitled

‘Sugarcane Micro-Propagation Techniques’ conducted in Havana, Cuba for six

months

After his return to Ethiopia, he continued to work at the Finchaa Sugar Factory until he left the organization in November 2007 for his Masters degree

Dr Amy S Collick was so wonderful She helped me organize things, gave mevaluable ideas, edited my manuscript, and was always in the front line to help me during hard times Thank you so much.

Very special thanks go to my co-advisor Dr Ingr Sileshi Bekele, East Africa Director of International Water Management Institute (IWMI), for his constructive ideas, encouragement, and worthy comments

I am thankful to ARARI, for allowing me to work at the Maybar research site and providing long term hydrological and sediment data

The support and help I got from Mr Derese G / Wold, former SCRP staff, was more than I could express; he was always generous, cooperative, and friendly He is anamazing person with never changing smile

International Water Management Institute and Bahir Dar University are

acknowledged for their financial support and Dr Ayalew Wondie helped me in

facilitating the financial issues with Bahir Dar University finance division

I am also greatly indebted to the technicians at Maybar Research Station: Gash Ali Ahmed, Seid Hussien, and Seid Belay They were welcoming and allowed me to share everything they have The life experience I got from them was invaluable

I would also like to thank my family for the unconditional love and support they provided me throughout my life and in particular, I must acknowledge my

younger sister Nitsuh Kebede, who has always believed in my potentials, and she was the reason and my strength to join this program

Finally, I would like to express my gratitude for all my friends, who have been helping and encouraging me by telephone and e-mail during the study and thesis writing periods

TABLE OF CONTENTS

BIOGRAPHICAL SKETCH III ACKNOWLEDGMENTS V TABLE OF CONTENTS VII LIST OF FIGURES X LIST OF TABLES XII LIST OF ABBREVIATIONS XIII

1 CHAPTER ONE 1

RESEARCH BACKGROUND 1

STUDY AREA 3

Location and Topography 3

Soils 5

Agro Climate, Land Use, and Cropping Pattern 5

REFERENCES 9

2 CHAPTER TWO 10

INTRODUCTION 10

MODEL DEVELOPMENT 12

MATERIALS AND METHODS 16

Discharge from Runoff Plots 17

Saturated Area Delineation 20

Data Checking and Analysis 20

MODEL EFFICIENCY EVALUATION 20

Calibration and Validation of Rainfall – Runoff Model 20

RESULTS AND DISCUSSION 21

Rainfall Amount, Intensity and Infiltration Capacity 21

Rainfall Intensity and Soil Infiltration Rate 23

Runoff from Test Plots 25

Groundwater in the watershed 28

Ground water level at different slope range 30

Ground water level at different land use areas 31

Simulating Watershed Discharge 32

Calibration and simulation 33

Runoff Source Area in the watershed 43

CONCLUSION 45

REFERENCES 46

3 CHAPTER THREE 51

INTRODUCTION 51

RESEARCH METHODS 52

Soil Sampling Techniques, Site Selection, and Sample Preparation 52

Laboratory Analyses 53

Soil Chemical Property Analyses 53

Organic Matter Content (%OC) 54

Available Phosphorous (AP) 54

Soil Physical Property Analysis 55

Bulk Density Determination (Db) 55

Statistical Analysis 56

RESULTS AND DISCUSSION 56

Available Phosphorous (AP) 56

Percentage Organic Carbon (% OC) 57

Bulk Density (Db) 58

CONCLUSION 59

APPENDICES 63

LIST OF FIGURES

Figure 1-1: Digital terrain map of the Maybar Watershed Low elevation at the

southern end of the watershed, near the outlet, is indicated by blue while high

elevation is indicated by red 4

Figure 1-2: Soil map of Maybar watershed (Source: Weigel, 1986) 5

Figure 1-3: Mean annual rainfall, river discharge, and suspended sediment yield 7

Figure 1-4: Long term daily climate record 8

Figure 1-5: Land use map of Maybar watershed (2008 cropping calendar 2nd crop) 8

Figure 2-1: Structure of the conceptual water balance model by Steenhuis et al (2008) .13

Figure 2-2: Location of piezometer transects at different slope range in the watershed 19

Figure 2-3: Long-term rainfall amount and distribution 22

Figure 2-4: Long-term average annual hydrograph 23

Figure 2-5: Long-term rainfall amount and intensity values 24

Figure 2-6: Infiltration test results (Source: Derib, 2005) 25

Figure 2-7: Average annual plot runoff and rainfall values 26

Figure 2-9: Plot runoff coefficients at different slope gradients 27

Figure 2-10: Comparison of simulated runoff from saturated area and plot runoff (Plot-1) on daily basis 28

Figure 2-11: Comparison of average response of piezometers to rainfall from upper and lower watersheds 30

Figure 2-12: Water level at different slope ranges calculated above the impermeable layer 31

Figure 2-13: Water levels at different land use types 32

Figure 2-14: Comparison of daily model calibration simulated and measured discharge 35Figure 2-15: Scatter plot of daily model calibration simulated and measured discharge 36Figure 2-16: Comparison of daily model validation simulated and measured discharge against rainfall amount 36Figure 2-17: Scatter plot of daily model validation simulated and measured discharge result 37Figure 2-18: Comparison of weekly model calibration output 38Figure 2-19: Scatter plot of weekly model calibration simulated and measured

discharge results 39Figure 2-20: Comparison of weekly model validation simulated and measured

discharge against rainfall amount 39Figure 2-21: Scatter plot of weekly model validation simulated and measured 40Figure 2-22: Comparison of monthly model calibration simulated and measured discharge against rainfall amount 41Figure 2-23: Scatter plot of monthly model calibration simulated and measured

discharge values 41Figure 2-24: Comparison of monthly model validation simulated and measured

discharge results 42Figure 2-25: Scatter plot of monthly model validation simulated and measured

discharge results 42Figure 2-26: Map of runoff source area 44

LIST OF TABLES

Table 1-1: Watershed characterization based on slope (Source: Weigel, 1986) 4

Table 1-2: Soil labels and their descriptions (Source: Weigel, 1986) 6

Table 1-3: Soil types and their area share 6

Table 2-1: Slope range, runoff coefficient, and land use type of test plots 27

Table 2-2: Optimized values of model parameters 33

Table 2-3: Summary of data used during modeling and model efficiency results for three time setups 38

Table 3-1: Statistical analysis result for AP 57

Table 3-2: Average AP, %OC, and Db values for different land use areas 57

Table 3-3: Statistical analysis result for (%OC) 58

Table 3-4: Statistical analysis result for (Db) 59

LIST OF ABBREVIATIONS

ARARI: Amhara Regional Agricultural Research InstituteEIAR: Ethiopian Institute of Agricultural Research

SCRP: Soil Conservation Research Program

SDC: Swiss Agency for Development and CooperationTP: Test Plot

UNDP: United Nations Development Program

WDR: World Development Report

1 CHAPTER ONE RESEARCH BACKGROUND AND STUDY AREA

RESEARCH BACKGROUND

In the 21st century, agriculture continues to be fundamental to the overall economy, food security, and poverty reduction in Sub-Saharan Africa countries (WDR,2007) In Ethiopia, agriculture is mainly rain-fed, traditional and small scale with low inputs, which often leads to low crop productivity and yield Furthermore, Ethiopia’s low crop productivity is further aggravated by water shortage due to scarce rainfall and land degradation caused by excessive soil erosion

Worldwide awareness of water scarcity has put an emphasis on finding better approaches to meet water demand (Anonymous, 2000b quoted by Bastiaanssen et al., 2003) and reduce erosion (Nyssen et al., 2008) Soil erosion and water scarcity are themajor problems in the Ethiopian highlands, affecting the livelihoods of millions as the associated sedimentation and flooding or drought cause additional problems for downstream populations

The majority of the Ethiopian human and livestock population reside in the Ethiopian highlands where soils are degraded due to exacerbated soil erosion reaching

up to 400 tons/hectare/year (UNDP, 2002) Increasing population pressure coupled with declining land productivity has led to a demand for additional food production

To meet the demand, all land types, irrespective of their suitability, are intensively cultivated using poor management practices In the period between 1950 and 2000, thepopulation in the Ethiopian highlands was estimated to have increased nearly four times from about 16 million to about 65 million (Hurni et al., 2005) In addition to excessive population pressure, the rain-fed, low-input subsistence agriculture of the highlands is further worsened by erratic and unpredictable rainfall resulting in drought

or flood conditions The rainfall in the highlands ranges from very little rain creating extreme drought conditions to excessive rainfall producing floods Both extremities result in severe crop damage and sometimes complete crop failure As a result, the Ethiopian highlands have become very fragile, sensitive to slight environmental changes, and food insecure

In Maybar, located in the northeastern escarpment of the central highlands of Ethiopia with attributes similar to the other highland areas in the country, farming practices are suffering from severe land degradation and acute water scarcity

problems Taking these problems into consideration, the Soil Conservation Research Program (SCRP) was implemented in 1981 by the Ethiopian Ministry of Agriculture (MoA) in collaboration with the University of Berne, Switzerland and with the support

of the Swiss Agency for Development and Cooperation (SDC) Under this program which lasted from 1981-1987, a total of seven research sites were established with the Maybar research station being SCRP’s first research site (SCRP, 2000) The

underlying objective was to provide measures that could be implemented to alleviate the aggravated land degradation and water scarcity problems During the

implementation, soil and water conservation measures, such as physical structures, area closures and biological structures, were put in place through a “food for work” campaign

Since the establishment of the site, fine resolution data on climate, hydrology and suspended sediment, from both river and test plots, has been collected and an expansive database was established that serves as a data source to carry out

hydrological, soil erosion, and conservation research activities at regional, national, and international levels

The data collected in this watershed has been analyzed by the Amhara

Regional Agricultural Research Institute (ARARI) and the Ethiopian Institute of

Agricultural Research (EIAR) researchers, national and international students at Masters and PhD levels, and other researchers The research activities, under taken in the watershed, differ both spatially and temporally depending on the objectives and intended outcomes

This study analyzes the data of the Maybar watershed, but it bases the analysis

on specific hydrological processes, specifically from the perspective of variable sourcearea hydrology that relies on saturation excess runoff mechanism To aid with the analysis, 29 piezometers were installed and ground water levels of the area were measured during the main rainy season of 2008 Furthermore, this study includes the impact assessment of the soil conservation research program intervention on selected physical and chemical soil properties of the study area

This thesis has three chapters This chapter gives insight into the complete research project and provides detailed information about the research site Chapter Two focuses on the hydrological modeling that incorporates the explanation of the major hydrological processes, identification of the major runoff mechanisms, and determination of the runoff sources and recharge areas in the watershed This

information was further used to model the rainfall-runoff relationships in the area Finally, Chapter Three addresses the impact assessment of SCRP interventions on selected physical and chemical properties of soils of the Maybar Watershed

STUDY AREA

Location and Topography

The study area consists of the Kori Sheleko catchment, which is found in the Maybar Watershed It is the first of the SCRP research sites established and is located

in the northern eastern part of the central Ethiopian highlands situated in the Southern Wollo administrative region, approximately 20 km south-southeast of Dessie town

The gauging station lies at 39o39’E and 10o51’N The area is characterized by highly rugged topography with steep slopes ranging between 2530 and 2860 meters above sea level (masl), a 330 meter altitude difference within a 112.8 ha catchment area (Figure CHAPTER ONE-1) Steep and very steep slope areas (> 25% slope) cover about 74% of the watershed Table CHAPTER ONE-1 defines and describes the slopeclasses, their area coverage and percent share of the watershed

Figure CHAPTER ONE-1: Digital terrain map of the Maybar Watershed Low

elevation at the southern end of the watershed, near the outlet, is indicated by blue while high elevation is indicated by red

Table CHAPTER ONE-1: Watershed characterization based on slope (Source: Weigel,1986)

The soil types in Maybar research unit have developed from the alkali-olive basalts and tuffs of the Ashangi group, which are part of the tertiary volcanic trap series (Weigel, 1986) Figure CHAPTER ONE-2 includes the soils map of the Maybar Watershed, and Table CHAPTER ONE-2 defines the soil labels found in the map’s legend Table CHAPTER ONE-3 clearly indicates that the watershed area is

dominated by shallow depth soils classified as phaeozems and phaeozems associated with lithosols and covering more than 93% of the total area in the watershed

Figure CHAPTER ONE-2: Soil map of Maybar watershed (Source: Weigel, 1986)

Agro Climate, Land Use, and Cropping Pattern

The Maybar research watershed receives an average annual rainfall of

1370mm, of which only 1148 mm is effective rainfall (rainfall contributing directly to runoff and recharge), and has an average annual river discharge of 407 mm The mean

annual suspended sediment rate was estimated to be 951 tons/year, which is

approximately 8.4tons/ha/year Figure CHAPTER ONE-3 provides a graphical representation of the annual rates of rainfall, discharge, and sediment yield from 1989

to 2004

Table CHAPTER ONE-2: Soil labels and their descriptions (Source: Weigel, 1986)

Table CHAPTER ONE-3: Soil types and their area share

The Kori River, the main river in Kori Sheleko catchment in Maybar, is the main inlet to Lake Maybar, which is approximately 0.5 km below the gauging station

Soil type Label Soil mappingunits Descriptions

Phaeozems

associated

with

Lithosols

a Hh1ls Hapllic phaeozems very shallow (10-25 cm), very stony, (sandy) clay loams.

b Hh2ls Hapllic phaeozems shallow to very sahllow (10-50 cm) stony phase, (sandy)

clay loams

Phaeozems

c Hh2s Hapllic phaeozems shallow (25-50 cm)

stony phase, clay loams

d Hh3s Hapllic phaeozems moderately deep (50-100 cm) stony phase, (sandy) clay loams.

e Hh4s Hapllic phaeozems deep to very deep (> 100 cm) stony phase, (sandy) clay loams.Regosols g Re2s Eutric regosols ver shallow to deep (10-100 cm) stony phase, clay loams.

Gleysols

h Gm1w Mollic Gleysols water table during

growing periods with in < 20 cm of the surface, clay loams

Mollic Gleysols water table during growing periods with in 20-50 cm of the surface, clay loams

The whole of the Maybar Watershed drains to the Borkena River, ultimately flowing tothe Awash River basin, a subcatchment of the central Ethiopian Rift Valley

Figure CHAPTER ONE-3: Mean annual rainfall, river discharge, and suspended sediment yield

The area is typical for the “Dega” thermal zone with an average daily

temperature of 16 OC The rainfall pattern commonly follows a bi-modal distribution (Figure CHAPTER ONE-4): the first rainy season, the shorter of the seasons, around mid-March to April and the second often begins around June/July and ends usually in September The Maybar area is known to be a low agricultural potential, intensively cultivated, oxen-ploughed cereal belt of the north-eastern escarpment region of the central Ethiopian highlands (Boshart, 1997)

According to Hurni et al (2005), approximately 60% of the total catchment area is cultivated whereas 20% is woodland and the remaining 20% is grassland (Figure CHAPTER ONE-5) There exists two cropping seasons and the predominant crops are cereals and maize, hence there exists two cropping seasons: the first, “Belg”,

is the small rainy season in spring and the second, “Kremt”, the main rainy season during the summer and autumn During the “Belg” season cereals are predominantly planted while in the “kremt” season pulses are dominant (SCRP, 2000)

Figure CHAPTER ONE-4: Long term daily climate record

Figure CHAPTER ONE-5: Land use map of Maybar watershed (2008 cropping calendar 2nd crop)

Bastiaanssen, W.G.M and L Chandrapala, 2003 Water balance variability across

Sri Lanka for assessing agricultural and environmental water use,

Agricultural Water Management 58(2)171-192

Bosshart, U 1997 Measurement of River Discharge for the SCRP Research

Catchments: Gauging Station Profiles Soil Conservation Research

Programme, Research Report 31, University of Berne, Switzerland

SCRP 2000 Area of Maybar, Wello, Ethiopia: Long-term Monitoring of the

Agricultural Environment.1981-1994 Soil Conservation Research

Programme, University of Berne, Switzerland

Hurni H., Tato K., Zeleke G.2005 The Implications of Changes in Population, Land

Use, and Land Management for Surface Runoff in the Upper Nile Basin Area

of Ethiopia Mountain Research and Development 25(2)147-154

Nyssen J., Poesen J., Deckers J 2008 Land degradation and soil and water

conservation in tropical highlands Soil & Tillage Research,article in press Journal homepage: www.elsevier.com/locate/still

United Nations Development Program 2002 Human Development Report Oxford

University Press, Inc 198 Madison Avenue, New York, New York, 10016.Weigel,G 1986 The Soils of Maybar Area Soil Conservation Research Programme

(SCRP), Report no 7 Berne, Switzerland: University of Berne

World Development Report 2007 Agriculture for development 1818 H Street NW

Washington D.C 20433

2 CHAPTER TWO MODELING RAINFALL - RUNOFF RELATIONSHIPS ATMAYBAR

RESEARCH UNIT: WOLLO, ETHIOPIA

interactions between groundwater and surface water and the resulting exchange fluxes are often characterized by high temporal and spatial variability Commonly the type ofinteraction is classified by the direction of the exchange fluxes: influent (flowing in)

fluxes and effluent (flowing out) fluxes (Kalbus et al., 2006; and Zehe, 2007).

There is a real need for improved concepts to determine the source and timing

of flow by studying the drainage morphology: from such knowledge watersheds can

be evaluated as intermediaries of water flow, and future behavior under specific conditions may be predicted with greater precision (Hewllet and Hibbert, 1963) Efficient prediction of quantitative runoff and river flow occupies a central place in thetechnology of applied hydrology (Nash and Sutcliff, 1970; Calvo, 1986; Hosking and Clarke, 1990; and Cabus, 2008) since these values are useful to avoid risk for water resource planning, flood forecasting, pollution control and many other applications

The modeling of rainfall-runoff relationships is not a simple task It requires sufficient knowledge and good understanding of the hydrological processes, rainfall

characteristics, runoff mechanism, and the identification of runoff source areas within the watershed, which in turn are determined by the physical properties of the basin (Shakya and Chander, 1998; and Gomi et al., 2008) If the relationships between theseproperties and the hydrological behavior could be defined, the hydrological responses

of basins could be easily predicted (Acreman and Sinclair, 1986)

Total rainfall falling in a given area will not be directly converted to runoff because before runoff is generated rainfall has to pass different steps (Hewllet and Hibbert, 1966; Wang et al., 1992; and Huang et al., 2008) As a result, the rainfall – runoff relationship in watersheds is non-linear (Szilagyi, 2007; and Leh et al., 2008)

Generally, there are two types of runoff mechanisms: saturation excess runoff and infiltration excess (Hortonian) runoff (Kubota and Sivaplan, 1995; Sen et al., 2008; and Wickel et al., 2008) Saturation excess runoff volume is dependent on the aerial extent of saturation within a watershed and the rainfall depth, but it is

independent of rainfall intensity In contrast, infiltration excess runoff volume is directly dependent on rainfall intensity and will not occur at low intensities (Walter et al., 2000) Identification of runoff generation processes within the watershed requires close observations and detailed investigations, but characterization of dominant runoff processes is not an easy task, especially when such processes occur below the soil surface (Beven, 1989 quoted in Latron and Gallart, 2008)

Computer-based rainfall-runoff models at different resolutions have been developed for several decades (Jayakrishnan et al., 2005) with the objective of

elucidating the complex and dynamic hydrologic processes and simulating runoff and river discharge from watersheds throughout the world Most of the models attempt to simulate the complex hydrological processes that lead to the transformation of rainfall into runoff, with varying degree of abstraction from different physical processes (Jacquin and Shamseldin, 2006) The models differ not only in their level of

complexity, but also in their level of applicability, efficiency, and specific data

requirements The efficiency of all the models that simulate the amount of runoff from

a given rainfall depends on the ability of the model to simulate, all factors that affect the rainfall-runoff process in a given area (Jacquin and Shamseldin, 2006)

Although hill slopes are responsible for generating 95% of the water in the streams (Shakya and Chander, 1998), hill slope hydrologic response to rainfall is not well studied (Meerveld and Weiler, 2008) Most of early models describing rainfall runoff processes relied on Horton’s (1933) infiltration excess principle (Shakya and Chander, 1998), but the Horton concept failed to predict runoff on vegetated hill slopes (Meerveld and Weiler, 2008) To better simulate runoff from hills slopes, Hewlett (1961) introduced the variable source area concept, which is based on

saturation excess runoff mechanism

In Ethiopia, saturation excess overland flow has been identified as one of the mechanisms for generating storm flow (Lui et al., 2008) This study in the Ethiopian highlands focuses on characterizing subsurface water flow and ground water table fluctuations in response to rainfall that leads to saturation excess runoff In particular, based on these processes, the goal is to develop a model that efficiently simulates the location of saturated runoff areas and predict river discharge The results of this study will help in realistically planning watershed interventions

MODEL DEVELOPMENT

Conceptual Watershed model: Watersheds in the Ethiopian highlands are

characterized by relatively flat bottomlands and gentle to steep sloping uplands In ourconceptual watershed model, the watershed is divided into two areas, based on slope steepness, soil depth, and infiltration capacity of the soil: runoff source areas near the river and recharge source areas on the hills The runoff source area was further divided

in to two sub-groups based on relative difference in soil depth and amount of moisture required to initiate runoff.

Figure -6 illustrates the structure of the watershed model developed in this study The basic assumption made was that hill slope areas have very high infiltration capacities and all the rainfall above field capacity percolates downward due to gravity

On the other hand, the excess rainfall when the soil is saturated from runoff source areas (flatter areas) becomes overland flow In addition the flatter areas remain wet even during the extreme dry months of the year, only the top most soil layer will dry due to small amounts of water percolating downward from the hills And hence these areas need only a small amount of rainfall, to start generating surface runoff

Figure -6: Structure of the conceptual water balance model by Steenhuis et al (2008)

Model Description: A water balance model was modified from the model in Collick

et al (2008) for small watersheds in the upper Blue Nile basin and in Steenhuis et al (2008) for the whole Blue Nile basin The basic inputs to the model are daily

precipitation and potential evapotranspiration Model outputs include daily runoff, interflow, and base flow according to the type and proportion of area under

consideration within the watershed

The amount of water stored in the topmost layer (root zone) of the soil, S (mm), for hill slopes and the runoff source areas were estimated separately with a water balance equation of the form:

Where P is precipitation, (mm d-1); AET is the actual evapotranspiration, (mm d-1), S

t-Δt, previous time step storage, (mm), R saturation excess runoff (mm d-1), Perc ispercolation to the subsoil (mm d-1) and Δt is the time step

During wet periods when the rainfall exceeds potential evapotranspiration, PET (i.e., P>PET), the actual evaporation, AET, is equal to the potential evaporation, PET Conversely, when evaporation exceeds rainfall (i.e., P<PET), the Thornthwaite and Mather (1955) procedure is used to calculate actual evapotranspiration, AET (Steenhuis and van der Molen, 1986) In this method, AET decreases linearly with moisture content, e.g.:

Where St (mm) is the available water stored in the root zone per unit area and Smax

(mm) is the maximum available soil storage capacity defined as the difference

between the amount of water stored in the top soil layer at wilting point and the maximum moisture content, equal to either the field capacity for the hill slope soils or

SPET

saturation (e.g., soil porosity) in runoff contributing areas Smax varies according to soilcharacteristics (e.g., porosity, bulk density) and soil layer depth Based on Eq 2-2 the surface soil layer moisture storage can be written as:

In this simplified model, direct runoff occurs only from the runoff contributing area, when the soil moisture balance indicates that the soil is saturated Recharge and interflow originate from the remaining hill slopes It is assumed that the surface runofffrom these areas is minimal This will underestimate the runoff during major rainfall events and, to test its significance, the model was run on a daily, weekly, and monthly basis

In the overland flow contributing areas when rainfall exceeds

evapotranspiration and fully saturates the soil, any moisture above saturation becomes runoff, and the runoff, R, can be determined by adding the change in soil moisture from the previous time step to the difference between precipitation and actual

evapotranspiration, e.g.:

For high infiltration areas on hill slopes the water flows either as interflow or baseflow to the stream Rainfall in excess of field capacity becomes recharge and is routed to two reservoirs that produce baseflow or interflow We assumed that the baseflow reservoir is filled first and when full, the interflow reservoir starts filling The baseflow reservoir acts as a linear reservoir and its outflow, BF, and storage, BSt,

are calculated when the storage is less than the maximum storage, BSmax as:

] 3 2 [

S t t t

]42 [

AET P

Where α is the half-life of the aquifer, or the time it takes for half of the volume of the aquifer to flow out without the aquifer being recharged

When the maximum storage, BSmax, is reached then:

Interflow originates from the hill slopes with the slope of the landscape as the major driving force of the water Under these circumstances, the flow decreases linearly (i.e., a zero order reservoir) after a recharge event The total interflow, IFt at time t can be obtained by superimposing the fluxes for the individual events

Where τ* is the duration of the period after the rainstorm until the interflow ceases, IFt

is the interflow at a time t, Perc*t-τ is the percolation on t-τ days

MATERIALS AND METHODS

The study was carried out in the Maybar watershed, fully described in Chapter One Discharge was collected at the outlet of the watershed for the periods 1988 –

1989, 1992-2000, 2002, 2004 and 2008 and from the test plots from 1988-1994 In addition, 29 piezometers were installed during 2008 and the (perched) groundwater table was measured during the rainy season Infiltration measurements were carried

out in a previous study by Derib et al., (2005) Additional information on major runoff

mechanism was based on the information and informal discussions with farmers and technicians

1 exp( ) [2 5 ]

]52 [

t BS

BF

a t

BF Perc BS

BS

t

t

t t t-Δ-

BF

a BS

Discharge at watershed outlet

At Maybar research station watershed discharge was measured with a flume installed in the Kori River The water level height is measured in two ways: float-actuated recorder and manual recording The maximum stage height at the gauging station was set to be 1.8 meter Using the discharge rating equation all water level records (stage height) including during the year 2008 was converted to discharge volumes (m3)

The final discharge rating curve was set to be as follows:

Discharge from Runoff Plots

In the research station, there are four test plots from which runoff and sedimentdata is being monitored Each test plot covers 2 m x 15 m area and represents differentland use areas and different slope gradients Surface runoff water is collected in a tank.Water is removed if the rainfall is in excess of 12.5 mm in less than six hours or if the water depth in the collection tank is more than 25.5 cm

Long term runoff data (1988-1994) from test plots was analyzed and comparedwith the rainfall data The runoff amount and runoff coefficient of each test plot was calculated and the result was compared taking land use/land cover and slope gradient differences into consideration

)1997,:

(

]82[

*023.1)18049

(

*003.0)4921

(

*016.2)

21

(

862 1

356 3

311 1

Bosshart Source

Q

H H

Q

H H

Q

H Q

Climatic Data

Daily maximum and minimum air and soil temperatures, wind direction, wind strength, and evaporation (using piche tube evaporimeter) were collected twice per day at 8:00 A.M and 18:00 P.M Rainfall data was monitored via two procedures: (1) using automatic rain gauge which uses chart role (one chart role for one month) and the data obtained was to be used for further determination of rainfall characteristics such as intensity, duration, frequency, and erosivity values of individual storm events and (2) using two manual rain gauges at two different locations, one in the upper part

of the catchment and the other near the office (climatic station)

Ground Water Table Measurement

Ground water table levels were measured with 29 piezometers during the (2008) main rainy season These were installed at two different locations in eight

transects, 16 pierzometers in four transects in the upper watershed (Atarimesk) and 13

piezometers in four transects in the lower watershed (near the gauging station) (Figure -7)

The piezometers used were prepared from 5 cm diameter PVC pipes of varyinglengths The bottom 30 cm of the piezometers was perforated at four places (at 5 cm interval) in four columns The perforated part was covered with cloth that allows waterinflow towards the tubes but prevents inflow of sediment The bottom end (opening)

of the pipe was closed with a plastic cap and sealed with plastic bandage to block inflow and outflow of water, while the above ground opening of the piezometers was capped to protect against the entrance of rainfall

Figure -7: Location of piezometer transects at different slope range in the watershed

Since the topography of the watershed is highly undulated and very steep, piezometers were installed at two different locations in the watershed that are

relatively suitable for piezometer installation The two locations were selected based

on the presence of better subsurface water flow from the top of the hill slope down to

the saturated area near the river During the installation, an “Idle Man Auger” was

used to drill the boreholes and drilling was done until the impermeable layer, bedrock,

or the ground water was reached Piezometer installation depth from the earth surface ranged form 0.64 to 2.02 meters Detailed information on piezometer installation sites can be found in Appendix 5

T-1

T-6

Saturated Area Delineation

The saturated area in the watershed was delineated and mapped by combining information collected using geographic positioning system (GPS) instrument, field observation, and ground water level data (piezometer head readings) The result (size and extent of saturated area) was cross validated with the results obtained from the topographic index (TI) map of the area Topographic Index maps are grids derived from digital elevation models (DEMs) It computes topographic indices for each grid cell based on upslope contributing area per unit length of contour and topographic slope of the cell As a result, bottomlands, with large upslope contributing areas, have higher topographic index values and are prone to saturation

Data Checking and Analysis

Data checking was done mainly for dates with missing values, time sequence discontinuities, and negative values Finally, the long-term data set from SCRP databases and primary data collected during the major rainy season (2008) were incorporated and analyzed using Microsoft Excel spreadsheets

MODEL EFFICIENCY EVALUATION

Calibration and Validation of Rainfall – Runoff Model

Evaluation of the hydrologic model behavior and performance is commonly made and reported through comparisons of simulated and observed values For model testing, the thirteen years of hydrological data was to be used was divided into two sets The first set was used for calibration and was comprised of the years 1992 to

2000, 2002, 2004 and 2008 The second set, used for model verification, consisted of data from 1988 and 1989, which was collected under supervision of Bern University and generally believed to be of the best quality

Model calibration was done manually through randomly varying input

parameters in order that the best “closeness” or “goodness-of-fit” was achieved between simulated and observed river discharge The calibrated input parameters consisted of maximum storage Smax of the three regions and the reservoir parameters t*, α, and SBmax

Model efficiency was evaluated based on Nash and Sutcliffe (1970) efficiency index (E) and coefficient of determination (R2) values Nash-Sutcliffe (E) Index can

be expressed as:

……….……… [2-10]

Where Pi is the simulated discharge for each time step, Oi is the observed discharge value, O the average measured discharge, N is the total number of values within the period of analysis The value of E ranges from -∞ to 1, where a value of 1 indicates perfect fit between simulated and measured values, while 0 implies the model

efficiency in predicting discharge is equal to the mean of the observed data, but if E is less than zero the observed mean is better than the model in predicting

RESULTS AND DISCUSSION

Rainfall Amount, Intensity and Infiltration Capacity

Temporal and spatial rainfall characteristics are very important factors that affect runoff generation The two rain gauges installed in the 112.8 ha watershed give temporal effects of rainfall i.e., intensity and amount, that were obtained from

pluviograph readings of the automatic rain gauge The annual precipitation based on

O P

E

1

2

13 years of observation is 1374 mm and the coefficient of variation for annual

precipitation is 0.15 (Figure -8)

Figure -8: Long-term rainfall amount and distribution

Rainfall is distributed annually into a major and a minor rainy season The

minor rainy season (Belg) extends from March to April while the main rainy season (keremt) starts from June/July and ends in September The average precipitation is

295 mm during the belg and 780 mm during the keremt For determining the

hydrologic response of a watershed, the effective rainfall (defined as the precipitation minus the potential evaporation) is an important parameter (Figure -9) During both rainfall seasons, precipitation exceeds evaporation The excess leaves the watershed over time Long-term rainfall and river discharge data is summarized on a monthly and annual basis in Appendix 1 and Appendix 2, respectively

Figure -9: Long-term average annual hydrograph

Rainfall Intensity and Soil Infiltration Rate

Rainfall intensity is an important parameter to model rainfall-runoff

relationships, especially in areas where infiltration excess runoff is expected (Beven, 2004; Amore et al., 2004) Rainfall intensity was calculated from continuous

pluviograph recordings by dividing the amount of storm rainfall by its duration The average daily rainfall intensity for six years (1988-1995 except data from 1990) was

7 mm/hr, Figure -10 depicts six years event based rainfall amount and corresponding intensity values The highest intensity rainfall, recorded in the watershed, was

equivalent to 162 mm/hr on 29 July 1992 from a rainfall amount of 2.7 mm lasted onlyfor one minute

Figure -10: Long-term rainfall amount and intensity values

Soil infiltration tests were conducted in the Maybar watershed by Derib

(2005) He performed 16 infiltration tests, inside and outside of the watershed down

to Lake Maybar, varying from just over one day to ten days The final infiltration rates at the end of the experiment ranged from 19 mm/hr to 600 mm/hr (Figure -11 andAppendix 3) In this watershed, all infiltration rates were greater than the observed average rainfall intensity rate of 7 mm/hr, and most of the 16 locations had a final infiltration rate that was less than the maximum observed rainfall intensity in six years Since the final infiltration rates in almost all cases were greater than the rainfallintensity, using the infiltration excess concept common to most models developed in temperate climates would result in the majority of the rainfall events not producing runoff However, there was significant runoff reported Hence, the saturation excess concept, used in our simple water balance model and discussed in the Model

Description section in Materials and Methods, is a valid concept to simulate runoff in the Maybar and other watersheds in the Ethiopian highlands with a monsoonal climate(Engida et al., 2009; and Legesse et.al.,2009 )

Figure -11: Infiltration test results (Source: Derib, 2005)

Runoff from Test Plots

The rainfall-runoff data collected from runoff plots for the years (1988, 1989,

1992 and 1994) allow us to clarify the runoff processes in the watershed further The average annual runoff from the four runoff plots is depicted in Figure -12

Figure -12: Average annual plot runoff and rainfall values

The runoff from runoff plot 1 is the greatest followed by plot 4 and 3 Plot 2 has the least amount of runoff Dividing the total runoff over the total precipitation, the runoff coefficients are obtained The runoff coefficients range from 0.69 to 0.33 (Table 2-1) A distinct relationship exists between slope and runoff coefficient, but not

as one would expect; the steeper the slope the smaller the runoff The results from

Figure 2-8 clearly indicate that as slope increases the runoff coefficient decreases, and this implies that areas with the steepest slope in the watershed have the least runoff coefficient compared with mid-slope and gentle-slope areas The results further strengthen the concept of saturation excess runoff from gentle slope areas being the major runoff mechanism in the watershed.

Table -4: Slope range, runoff coefficient, and land use type of test plots

Test Plot No Slope (%) Runoff Coefficient Land Use Type

Figure -13: Plot runoff coefficients at different slope gradients