The impact of sedimentation and climate variability on the hydrological status

of Lake Hawassa, South Ethiopia

Dissertation

zur

Erlangung des Doktorgrades (Dr. rer. nat)

der

Mathematisch-Naturwissenschaftlichen Fakultät

der

Rheinischen Friedrich-Wilhelms-Universität Bonn

vorgelegt von

Mulugeta Dadi BELETE

aus

Dilla, Äthiopien

Bonn 2013
Angefertigt mit der Genehmigung der Mathematisch-Naturwissenschaftlichen Fakultät der
Rheinischen Friedrich-Wilhelms-Universität Bonn

1. Referent: Prof. Dr. Bernd Diekkrüger

2. Referent: Prof. Dr. Jürgen Herget

Tag der Promotion: 11.12.2013

Erscheinungsjahr: 2013

i
 Dedicated to:

My queen `Simegn Asmare´, my children `Fikir Mulugeta´ & `Nuhamin Mulugeta´, and my late
mother `Yeshumnesh Alemayehu´

ii
Acronyms and abbreviations
AMO Atlantic Multidecadal Oscillation
AnnAGNPS Annualized Agricultural Non-point Sources
AR Autoregression
CL Confidence limit
CSE Conservation Strategy of Ethiopia
DEM Digital Elevation Model
DPSIR Drivers-Pressures-State-Impact-Response
EIA Environmental impact assessment
ENSO El Niño-Southern Oscillation
ETo Potential evapotranspiration
GIS Geographic Information System
HSG Hydrologic Soil Group
IPPC Intergovernmental Panel on Climate Change
ITCZ Inter-tropical Convergence Zone
m.a.s.l meter above sea level
MER Main Ethiopian Rift
MJO Madden-Julian Oscillation
MK Mann-Kendall
MODIS MODerate-resolution Imaging Spectroradiometer
MoWR Ministry of Water Resources
MPSIAC Modified Pacific Southwest Inter Agency Committee
MUSLE Modified Universal Soil Loss Equation
NAO North Atlantic Oscillation
NASA National Aeronautics and Space Administration
NBS Net Basin Supply
NEP North-East Pacific
NNSMP National non-point source monitoring program
NOAA National Oceanic and Atmospheric Administration
NP North Pacific Index
NSE Nash-Sutcliffe efficiency
ONI Oceanic Niño Index
PDA Personal Digital Assistant
PDO Pacific Decadal Oscillation
PSIAC Pacific Southwest Inter Agency Committee
RSI Regime Shift Index
RUSLE Revised Universal Soil Loss Equation
SCS Soil Conservation Service
SPOT5 Satellite Pour I´Observation de la Terre 5
SRTM Shuttle Radar Topography Mission
SST Sea Surface Temperature
TLU Tropical Livestock Unit
UNEP United Nations Environmental Protection
USBR United States Bureau of Reclamation
USLE Universal Soil Loss Equation
VCF Vegetation Continuous Fields
WRDB Water resources development bureau
WWDSE Water works design and supervision enterprise

iii
 Contents

TABLE OF FIGURES ...................................................................................................................................................................... VII

LIST OF TABLES ........................................................................................................................................................................... IX
ABSTRACT ................................................................................................................................................................................. I
KURZFASSUNG .......................................................................................................................................................................... I
CHAPTER 1. INTRODUCTION ........................................................................................................................................................... 3
1.1. BACKGROUND INFORMATION .............................................................................................................................. 3
1.2. PROBLEM STATEMENT........................................................................................................................................ 5
1.3. OVERALL OBJECTIVES OF THE THESIS...................................................................................................................... 7
1.4. THESIS ARCHITECTURE AND GENERAL APPROACH ..................................................................................................... 7
CHAPTER 2. DESCRIPTION OF THE STUDY AREA ............................................................................................................................. 9
2.1. LOCATION........................................................................................................................................................ 9
2.2. CLIMATE AND AGRO-ECOLOGY .......................................................................................................................... 10
2.3. TOPOGRAPHY ................................................................................................................................................. 12
2.4. SOILS ............................................................................................................................................................ 14
2.5. LAND USE/COVER............................................................................................................................................ 15
2.6. GEOLOGY ...................................................................................................................................................... 16
2.7. MORPHOLOGY OF LAKE HAWASSA ..................................................................................................................... 17
CHAPTER 3. CHARACTERIZATION OF THE WATER LEVEL VARIABILITY OF THE MAIN ETHIOPIAN RIFT VALLEY (MER) LAKES .............. 20
3.1. INTRODUCTION ............................................................................................................................................... 20
3.2. OBJECTIVES OF THE CHAPTER ............................................................................................................................. 20
3.3. METHODOLOGY .............................................................................................................................................. 21
3.3.1. Description of the study area and characteristics of the lakes .......................................................... 21
3.3.2. Available data .................................................................................................................................... 23
3.3.3. Methods ............................................................................................................................................. 24
3.4. RESULTS AND DISCUSSION................................................................................................................................. 27
3.4.1. Classification of the lakes based on their long-term water balance .................................................. 27
3.4.2. Classification of the lakes based on their morphology ....................................................................... 28
3.4.3. Recent/actual situations of individual lake level regimes .................................................................. 29
3.5. CONCLUSIONS ................................................................................................................................................ 34
CHAPTER 4. THE IMPACT OF CLIMATE SHIFTS AND ENSO PHENOMENA ON THE HYDROLOGICAL STATUS OF LAKE HAWASSA ............ 35
4.1. INTRODUCTION ............................................................................................................................................... 35
4.2. HYPOTHESIS AND OBJECTIVES OF THE CHAPTER ..................................................................................................... 36
4.3. IMPACT OF EL NIÑO/LA NIÑA ON CLIMATE VARIABILITY OF EAST AFRICA.................................................................... 37
4.4. METHODS ..................................................................................................................................................... 38
4.4.1. Data availability ................................................................................................................................. 38
4.4.2. Estimation of coherence between ENSO index and lake level variability ........................................... 38
4.4.3. Significance limits of the spectral coherence estimation ................................................................... 40
4.4.4. Sequential regime shift detection using Regime Shift Index (RSI) ...................................................... 42
4.4.5. Detection of long-terms trends using Mann-Kendall test .................................................................. 44
4.5. RESULTS AND DISCUSSION................................................................................................................................. 47
4.5.1. Results of coherence analysis ............................................................................................................. 47
4.5.2. Variability in the lake level ................................................................................................................. 48
4.5.3. Rainfall variability in the watershed .................................................................................................. 51
4.5.4. Variability in the streamflow of Tikur Wuha River ............................................................................. 53

iv
 4.5.5. Shift detection in the runoff coefficient of Tikur Wuha sub-watershed ............................................. 54
4.5.6. Land use/cover changes: a potential anthropogenic factor .............................................................. 56
4.5.7. Detection of regime shift in lake-evaporation (1986-2007) ............................................................... 57
4.6. CONCLUSIONS ................................................................................................................................................ 59
CHAPTER 5. SIMULATION OF THE LONG-TERM WATER LEVEL VARIABILITY OF LAKE HAWASSA IN THE PRESENCE OF REGIME
SHIFTS IN WATER BALANCE COMPONENTS ..................................................................................................................................... 61

5.1. INTRODUCTION ............................................................................................................................................... 61

5.2. OBJECTIVES OF THE CHAPTER ............................................................................................................................. 61
5.3. PREVIOUS WATER BALANCE STUDIES OF LAKE HAWASSA ......................................................................................... 61
5.4. METHODS ..................................................................................................................................................... 62
5.4.1. Representing the water balance of Lake Hawassa ............................................................................ 62
5.4.2. Quantification of water balance parameters ..................................................................................... 63
5.4.3. Simulation procedure ......................................................................................................................... 65
5.4.4. Updating procedure using Autoregression (AR) ................................................................................. 66
5.4.5. Model efficiency test using Nash-Sutcliffe Efficiency (NSE)................................................................ 67
5.5. RESULTS AND DISCUSSION................................................................................................................................. 68
5.5.1. Results of water level simulation ....................................................................................................... 68
5.5.2. Result of residual error analysis ......................................................................................................... 68
5.5.3. Which components of the water balance show particularity at the year 1996? ............................... 70
5.5.4. Replacing the pan-evaporation by the Penman-Monteith model ...................................................... 70
5.5.5. Annual magnitudes of water balance components ........................................................................... 71
5.6. CONCLUSIONS ................................................................................................................................................ 72
CHAPTER 6. THE EFFECT OF RECENT SEDIMENTATION ON THE STORAGE CAPACITY OF LAKE HAWASSA ........................................... 74
6.1. INTRODUCTION ............................................................................................................................................... 74
6.2. HYPOTHESIS AND OBJECTIVES OF THE CHAPTER ..................................................................................................... 76
6.3. MATERIALS AND METHODS ............................................................................................................................... 77
6.3.1. The old bathymetric map (1999) ........................................................................................................ 77
6.3.2. Tying local bench marks with standard elevations ............................................................................ 77
6.3.3. The new bathymetric survey (2010) ................................................................................................... 77
6.3.4. Assessment of sediment by topographic differencing technique ....................................................... 78
6.4. RESULTS AND DISCUSSION................................................................................................................................. 80
6.4.1. Estimation of sediment thickness and distribution ............................................................................ 80
6.4.2. Sediment volume and its impact on the storage capacity ................................................................. 81
6.5. METHODOLOGICAL LIMITATIONS/ SHORTCOMING /CHALLENGES AND SOLUTIONS ........................................................ 81
6.6. THE LINKAGE OF IN-LAKE SEDIMENTATION TO THE SOURCES IN THE WATERSHED .......................................................... 82
6.6.1. Linkage to gully density in the watershed .......................................................................................... 83
6.6.2. Linkage to the disappearance of Lake Cheleleka ............................................................................... 83
6.7. CONCLUSIONS ................................................................................................................................................ 86
CHAPTER 7. THE LINKAGE OF SEDIMENTATION IN LAKE HAWASSA TO THE WATERSHED CHARACTERISTICS: AN APPLICATION OF
PSIAC MODEL ............................................................................................................................................................................ 87
7.1. INTRODUCTION ............................................................................................................................................... 87
7.2. RESEARCH QUESTIONS AND OBJECTIVE OF THE CHAPTER.......................................................................................... 89
7.3. PSIAC MODEL ................................................................................................................................................ 89
7.4. MATERIALS AND METHODS ............................................................................................................................... 92
7.4.1. Parameterization of individual factors ............................................................................................... 92
7.4.2. Arithmetic procedure for erosion/sediment risk assessment ............................................................. 95
7.5. RESULTS AND DISCUSSION................................................................................................................................. 96
7.5.1. Result of surface geology rating ........................................................................................................ 96

v
 7.5.2. Soil rating result ................................................................................................................................. 96
7.5.3. Climate rating result .......................................................................................................................... 97
7.5.4. Runoff rating results........................................................................................................................... 98
7.5.5. Topography rating result ................................................................................................................. 100
7.5.6. Land cover rating result ................................................................................................................... 100
7.5.7. Land use rating result ....................................................................................................................... 102
7.5.8. Upland slope erosion rating result ................................................................................................... 104
7.5.9. Channel erosion and sediment transport rating result .................................................................... 105
7.5.10. Identification of erosion/sediment source areas: the model output .............................................. 106
7.5.11. Comparison of sediment yield estimation to previous studies ....................................................... 108
CHAPTER 8. PARTICIPATORY ASSESSMENT OF ANTHROPOGENIC FACTORS AFFECTING THE HYDROLOGY OF LAKE HAWASSA: AN
APPLICATION OF DPSIR FRAMEWORK ........................................................................................................................................ 110

8.1. INTRODUCTION ............................................................................................................................................. 110

8.2. OBJECTIVES OF THE CHAPTER ........................................................................................................................... 112
8.3. METHODOLOGY ............................................................................................................................................ 113
8.3.1. General methodology....................................................................................................................... 113
8.3.2. DPSIR story sheet and participatory approach ................................................................................ 113
8.4. RESULTS AND DISCUSSION............................................................................................................................... 115
8.4.1. "Indirect drivers" (iD)........................................................................................................................ 115
8.4.2. "Direct drivers" (D) ........................................................................................................................... 118
8.4.3. "Pressures" (P).................................................................................................................................. 119
8.4.4. "State" (S) ......................................................................................................................................... 120
8.4.5. "Impact" (I) ....................................................................................................................................... 120
8.4.6."Responses" (R) ................................................................................................................................. 120
8.5. SUMMARY AND CONCLUSIONS ......................................................................................................................... 122
8.6. LIMITATION OF THE STUDY .............................................................................................................................. 123
CHAPTER 9. SYNTHESIS ............................................................................................................................................................ 125
9.1. GENERAL REMARKS ....................................................................................................................................... 125
9.2. SYNTHESIS OF CAUSAL LINKS ............................................................................................................................ 125
9.3. OVERALL CONCLUSIONS ................................................................................................................................. 126
9.4. PERSPECTIVES .............................................................................................................................................. 127
REFERENCES ............................................................................................................................................................................ 129
ACKNOWLEDGEMENT ................................................................................................................................................................ 146
APPENDICES ............................................................................................................................................................................. 147

vi
 Table of figures

Figure 1.1. Locations of the Ethiopian Rift Valley lakes ......................................................................... 4

Figure 2.1. Maps of the study area at different scales ............................................................................. 9
Figure 2.2. Distribution of monthly rainfall (a) and temperature (b) at Hawassa Station .................... 10
Figure 2.3. Isohytal map of Lake Hawassa watershed .......................................................................... 11
Figure 2.4. Partitioning of the watershed by Thiessen´s polygon .......................................................... 12
Figure 2.5. Hill shade view of the watershed landscape as processed from DEM ................................ 13
Figure 2.6. Three dimensional view of topographic diversity of the watershed .................................... 13
Figure 2.7. Elevation range of Lake Hawassa watershed...................................................................... 13
Figure 2.8. Soil types of the watershed .................................................................................................. 14
Figure 2.9. Geological map of Lake Hawassa watershed...................................................................... 17
Figure 2.10. Elevation diversity of some Rift Valley lakes..................................................................... 18
Figure 2.11. Elevation-area-volume curve for Lake Hawassa .............................................................. 19
Figure 3.1. Base map of the Ethiopian Rift Valley basin ....................................................................... 22
Figure 3.2. Classification of lakes by water balance criteria (aridity factor is not included) ............... 27
Figure 3.3. Classification of lakes by water balance criteria (aridity factor included)......................... 28
Figure 3.4. Long-term lake level plots of the eight Rift Valley lakes ..................................................... 30
Figure 3.5. Lake level rise of Lake Beseka in relation to climatic factors ............................................. 34
Figure 4.1. The procedure to determine sequential regime shifts .......................................................... 43
Figure 4.2. The procedure to account for the existence of autocorrelation ........................................... 44
Figure 4.3. Coherence between ENSO index and lake level variability in frequency domain .............. 47
Figure 4.4. Hydrograph of monthly maximum lake level ....................................................................... 48
Figure 4.5. Sequential regime-shifts in annual lake levels .................................................................... 50
Figure 4.6. Regime shift of over-lake rainfall ........................................................................................ 52
Figure 4.7. Sequential regime shifts in the annual rainfall on the watershed ....................................... 53
Figure 4.8. Regime shift in Tikur Wuha stream flow ............................................................................. 54
Figure 4.9. Tikur Wuha sub-watershed and Thiessen polygon .............................................................. 54
Figure 4.10. Shift detection in runoff coefficient of Tikur Wuha sub-watershed ................................... 56
Figure 4.11. Comparison of lake-evaporation estimates of pan vs. Penman-Monteith model .............. 58
Figure 4.12. Regime shifts in lake-evaporation as computed from pan records.................................... 59
Figure 4.13. Regime shifts in lake-evaporation as computed by the Penman-Monteith model ............. 59
Figure 5.1. Schematic representation of the water balance components for a closed lake ................... 62
Figure 5.2. Flow chart of the lake level simulation procedure .............................................................. 65
Figure 5.3. Schematic diagram of the linear Auto-Regressive (AR) updating model ............................ 68
Figure 5.4. Trend of residual error (predicted - observed lake level values) ........................................ 69
Figure 5.5. Simulated vs. observed lake level ........................................................................................ 69
Figure 5.6. Annual magnitudes of water balance components .............................................................. 70
Figure 5.7. Simulation result after replacing pan-evaporation by Penman-Monteith model ................ 71
Figure 6.1. Partial view of echosounders for the old and new bathymetry surveys .............................. 78
Figure 6.2. Basic steps in the topographic differencing between the two bathymetry maps ................. 79
Figure 6.3. Spatial distribution of sediment thicknesses during the period 1999-2010 ......................... 80
Figure 6.4. Map of gully density and pictures of active gullies in the watershed .................................. 83
Figure 6.5. The disappearance of Lake Cheleleka in the watershed ..................................................... 84
Figure 6.6. Time series of changes in the surface area of Lake Cheleleka ............................................ 85
Figure 6.7. Current view of the former Lake Cheleleka (2011) ............................................................. 85
Figure 6.8. Dropping down of Lake Haromaya till drying up ............................................................... 86
Figure 7.1. Field measurement of infiltration capacity using double-ring infiltrometer ....................... 93
Figure 7.2. Spatial distribution of geology factor values ....................................................................... 96

vii
Figure 7.3. Spatial distribution of soil factor values ............................................................................. 97
Figure 7.4. Location of meteorological stations and climate factor values ........................................... 98
Figure 7.5. Locations of infiltration measurement sites......................................................................... 98
Figure 7.6. Distribution of hydrologic soil groups in the watershed ..................................................... 99
Figure 7.7. Rating of topographic factor based on slope percentage .................................................. 100
Figure 7.8. Derivation of bare land percentage from Continuous Fields (VCF) ................................ 101
Figure 7.9. Rating of land cover factor ................................................................................................ 102
Figure 7.10. Derivation of tree percentage from Vegetation Continuous Fields (VCF) ..................... 102
Figure 7.11. Percentage of canopy cover computed from MODIS satellite data ................................ 103
Figure 7.12. Rating of land use factor ................................................................................................. 104
Figure 7.13. Upland land erosion rating ............................................................................................. 105
Figure 7.14. Sediment transport rating ................................................................................................ 105
Figure 7.15. Output diagram of model builder after running the raster input data by ArcGIS10 ...... 106
Figure 7.16. Spatial distribution of sediment yield potential classes in the watershed ....................... 107
Figure 8.1. Terminology of DPSIR model............................................................................................ 111
Figure 8.2. The DPSIR framework in a decisional context .................................................................. 112
Figure 8.3. Flowchart for causal chain assessment of anthropogenic factors .................................... 113
Figure 8.4. Template of DPSIR story sheet .......................................................................................... 114
Figure 8.5. Primary and secondary driving forces affecting the hydrology of the system .................. 115
Figure 8.6. Trends of agricultural land expansion in the watershed ................................................... 116
Figure 8.7. Link between direct drivers and pressures ........................................................................ 119
Figure 8.8. List of responses ................................................................................................................ 121
Figure 9.1. Summary of causal link and relationship among anthropogenic and natural factors ...... 126

viii
 List of tables

Table 2.1. Description of soil types in Lake Hawasssa watershed ........................................................ 15

Table 2.2. Types of land cover in Lake Hawassa watershed.................................................................. 16
Table 2.3. Summary of physical characteristics of Lake Hawassa ........................................................ 18
Table 3.1. Morphological characteristics of Rift Valley lakes ............................................................... 23
Table 3.2. Selected water quality parameters of the Rift Valley lakes ................................................... 23
Table 3.3. Water balance components of the eight Rift Valley lakes ..................................................... 24
Table 3.4. Results of characterization based on specific watershed ...................................................... 24
Table 3.5. Results of inflow, outflow, aridity, and the corresponding quadrants .................................. 27
Table 3.6. Results of characterization based on specific watershed ...................................................... 29
Table 3.7. Monotonic trends of level of individual lakes ....................................................................... 31
Table 3.8. Land use/cover changes in Lake Abaya watershed ............................................................... 33
Table 4.1. The core set of hydro-climatic data employed in the study ................................................... 38
Table 4.2. Annual total rainfall and runoff coefficients in the Tikur Wuha sub-watershed ................... 55
Table 4.3. Land use changes in Lake Hawassa watershed .................................................................... 57
Table 5.1. Annual volume of some components of water balance of Lake Hawassa ............................. 72
Table 6.1. Main technical parameters of the two echosounders ............................................................ 78
Table 7.1. Sediment source estimation techniques ................................................................................. 89
Table 7.2. PSIAC parameters and their diagnostic criteria................................................................... 90
Table 7.3. PSIAC factor ratings and degree of limitation ...................................................................... 91
Table 7.4. Rating of topographic factor ................................................................................................. 94
Table 7.5. Erosion class based on drainage density .............................................................................. 95
Table 7.6. PSIAC sediment classes ........................................................................................................ 95
Table 7.7. Rating of geological formations ............................................................................................ 96
Table 7.8. Soil rating of major soil types ............................................................................................... 97
Table 7.9. Spatial distribution of meteorological station within and near to the study area ................. 97
Table 7.10. Rating of runoff factor ......................................................................................................... 99
Table 7.11. Percentage of bare land in each land cover type (500m resolution) ................................ 101
Table 7.12. Percentage of canopy in each land use type (500m resolution) ........................................ 103
Table 7.13. Percentages of each potential classes ............................................................................... 107
Table 7.14. Summary of model outputs ................................................................................................ 108
Table 8.1. Population growth in seven administrative units in and around the watershed ................. 116

ix
 Abstract

ABSTRACT

Lake Hawassa is a topographically closed lake in the Central Main Ethiopian Rift Valley.
The water level of this lake has been rising significantly with an average rate of 4.9
cm/year over the study period (1970-2010). The cause of this rise is not yet sufficiently
investigated. The main target of this study is to investigate causal variables for lake level
variability in general, and its resultant rise in particular. The study is based on two main
hypotheses. The first is concerned with the effect of climate variability on the lake level
variability; and the second is related to the effect of sedimentation on the storage capacity
of the lake.

The first hypothesis (the effect of climate variability) was investigated through the
application of diverse statistical techniques. It comprises the coherence analysis to study
the linear relationship between the 3.4 ENSO index and lake level changes. A sequential
regime shift algorithm was employed to investigate the variations in the mean values of
some selected hydro-climatic variables. Trend test was also used to investigate the
variability of the hydro-climatic variables overtime. A simple water balance approach was
applied to simulate the lake level variability so as to examine how the model behaves
throughout the study period.

The second hypothesis (the effect of sedimentation) was approached by conducting a new
bathymetric survey. The result of the new survey was compared with the existing
bathymetric map of 1999. The Pacific-Southwest Inter-Agency Committee (PSIAC) model
was also employed to identify the "hot-spots" of sediment production in the watershed. In
this semi-quantitative model, nine factors affecting sediment yielding the watershed were
characterized, rated, and an overlay analysis was performed. Participatory assessment of
anthropogenic factors that affect the hydrological status of the lake was conducted through
the application of DPSIR (Drivers-Pressures-State-Impact-Response) analytical
framework.

The result of the coherence analysis between the monthly lake level changes and the
corresponding changes in the ENSO index reveals that the two variables have significant
linear relationship over frequencies ranging from 0.13 to 0.14 cycles/month or 1.56 to 1.68
cycles/year. This corresponds to a dominant average periodicity (coincident cycle) of about
7.4 months. Furthermore, the result of sequential regime shift detections show that most of
the significant change points coincide with the occurrences of ENSO events and climate
shifts. Generally, the lake level tends to be high during El Niño and low during La Niña
years. The typical example is the coincidence of extreme historical maximum lake level to
the strongest El Niño event of the century that occurred in 1997/98. The coincidence of
climate regime shift in the Pacific Ocean in 1976/77 with an equivalent regime shift in the
lake level and rainfall records of this period is considered as additional evidence. The study
further reveals the existence of sequential regime shifts in stream flow, runoff coefficient,
and lake evaporation which clearly coincide with the occurrences of ENSO phenomena.

i
 Abstract

Results of the Mann-Kendall trend analyses also reveal the significant increasing trend of
the lake level and streamflow. On the contrary, decreasing trend of evaporation was
observed while rainfall exhibits no trend over the study period.

The long-term increasing trend of streamflow from Tikur Wuha sub-watershed without the
corresponding increment in rainfall is found to explain the role of land use/cover changes
at least in modifying the impact of climate.

The application of simple spreadsheet water balance model estimates the long-term (1986-
2006) average annual magnitudes of the water balance components as follows: over-lake
precipitation (89 Mm³), evaporation from the lake surface (132 Mm³), streamflow from the
Tikur Wuha sub-watershed (94 Mm³), and streamflow from the un-gauged sub-watershed
(77 Mm³) and storage changes (3 Mm³).

Comparison of the two bathymetric maps shows that the average accumulated sediment
between the years 1999 and 2010 was estimated as 14 ± 5cm or 13.3 x106 m3. Assuming a
constant rate, the mean annual average rate of sedimentation in the lake is about 1.2
cm/year or 1.1 x106 m3. Accordingly, the mean annual reduction in storage capacity of the
lake due to siltation is 0.08 %.

The attempt to link sediment yield estimate of the bathymetric approach with the estimates
of the PSIAC model results in a considerable disagreement as the former estimates 967
m3/km2/year whereas the latter estimates the sediment yield to be in the range of 95-250
m3/km2/yr.

The result of participatory assessment of anthropogenic factors and review of previous

studies shows that anthropogenic factors show considerable impact on the hydrological
status of the lake. Sedimentation and increased runoff are perceived as pressures
(immediate causes) for the lake level rise (state). These pressures are perceived to arise
from drivers (land use changes, deforestation and misuse/mistreatment of land resources).
These drivers in turn had resulted from indirect drivers that comprised population growth
and density, agricultural development, the use of wood as fuel, socio-economic changes,
and the existing land tenure system. The interesting finding of this assessment of
anthropogenic factors is the presence of promising policy instruments (responses) that
support the integrated management of the lake and the watershed. The failure of
implementation of these policy instruments is the commonly complained issues among the
stakeholders.

ii
 Kurzfassung

KURZFASSUNG

Der Hawassa-See ist ein Endsee im afrikanischen Grabenbruch, dessen Wasserstand im

Zeitraum dieser Studie (1970-2010) jährlich im Durchschnitt um 4,9 cm gestiegen ist. Der
Grund für diesen Anstieg ist noch nicht ausreichend erforscht. Das Hauptziel dieser Studie
ist die Untersuchung der Ursachen für die Variabilität des Seewasserstandes im
Allgemeinen und für den beobachteten Anstieg insbesondere. Dieser Arbeit liegen zwei
Hypothesen zugrunde. Die erste bezieht sich auf die Auswirkungen der Klimavariabilität
und die zweite auf die Auswirkung der Sedimentation auf die Speicherkapazität des
Hawassa-Sees.

Für die Untersuchung der ersten Hypothese (Auswirkung der Klimavariabilität) wurden
verschiedene statistische Verfahren eingesetzt, darunter die Kohärenzanalyse, um die
lineare Beziehung zwischen dem 3.4 ENSO-Index und der Wasserstandsänderung zu
prüfen. Der sequential regime shift algorithm wurde verwendet, um zu untersuchen, ob die
Kipppunkte der Mittelwerte ausgewählter hydro-klimatischer Variablen mit dem Auftreten
bzw. der Intensität der ENSO-Ereignisse übereinstimmen. Weiterhin wurde eine
Trendanalyse durchgeführt, um die zeitliche Variabilität klimatischer Parameter zu
bestimmen. Mittels eines einfachen Wasserbilanzverfahrens wurden die
Wasserstandsänderungen simuliert, um das Modellverhalten im Untersuchungszeitraum zu
analysieren.

Für die Analyse der zweiten Hypothese (Sedimentationseffekt) wurde eine neue
bathymetrische Untersuchung durchgeführt und mit einer existierenden Bathymetrie aus
dem Jahr 1999 verglichen. Das Pacific-Southwest Inter-Agency-Committee-Modell
(PSIAC) wurde für die Bestimmung von „Hot-Spots“ der Sedimentproduktion eingesetzt.
In diesem Modell werden neun Faktoren der Erosion und Sedimentation im Einzugsgebiet
berücksichtigt, flächenhaft berechnet und überlagert. Abschließend wurde eine
partizipative Bewertung der beeinflussenden anthropogenen Faktoren im Rahmen der
DPSIR-Methode (Drivers-Pressures-State-Impact-Response) durchgeführt.

Das Ergebnis der Kohärenzanalyse zwischen monatlichen Wasserstandsänderungen und

den entsprechenden ENSO-Indices zeigt, dass die beiden Variablen eine signifikante
lineare Beziehung im Frequenzbereich von 0,13 bis 0,14 Zyklen/Monat bzw. 1,56 bis 1,68
Zyklen/Jahr aufweisen. Dies entspricht einer dominierenden mittleren Periodizität von ca.
7,4 Monaten. Darüber hinaus zeigen die Ergebnisse der sequential regime shift detection,
dass die überwiegenden Kipppunkte der ENSO-Ereignisse und der Klimaparameter
übereinstimmen. Der Seewasserstand tendiert in El Niño-Jahren zu höheren und in La
Niña-Jahren zu niedrigeren Werten. Ein typisches Beispiel ist die Übereinstimmung des
historisch höchsten Seewasserstandes mit dem stärksten El Niño-Ereignis des letzten
Jahrhunderts im Winter 1997/1998. Eine weitere Evidenz ist die Übereinstimmung der
Verschiebung des Klimaregimes im pazifischen Ozean 1976/1977 mit einer
entsprechenden Verschiebung des Seewasserstände und der Niederschläge im gleichen
Zeitraum. Die Untersuchung zeigt auch die Existenz von weiteren Regimeverschiebungen
in Abfluss, Abflussbeiwert und Evaporation in Übereinstimmung mit ENSO-Ereignissen.

i
 Kurzfassung

Die Ergebnisse der Mann-Kendall-Trendanalyse zeigen eine Übereinstimmung zwischen

Seewasserstand und gemessenem Zufluss, wohingegen die Evaporation abnimmt und der
Niederschlag keinen Trend zeigt.

Die langfristige Zunahme der beobachteten Zuflüsse am Pegel Tikur-Wuha ohne Änderung
des Niederschlags ist ein Hinweis auf die Bedeutung von Landnutzungs- und
Landbedeckungsänderungen im Einzugsgebiet.

Die Anwendung einer einfachen Tabellenkalkulation ergibt die langfristigen (1986-2006)

mittleren Jahresbilanzen: Niederschlag über dem See (89 Mm³), Evaporation des Sees (132
Mm³), Zufluss des Tikur-Wuha Einzugsgebietes (94 Mm³), und Zufluss des nicht
instrumentierten Einzugsgebietes (77 Mm³) sowie Speicheränderung (3 Mm³).

Der Vergleich der beiden Bathymetrien ergibt eine Sedimentakkumulation in der Zeit von
1999 bis 2010 in Höhe von 14 ± 5cm oder 13.3 x106 m3, was einem mittleren Wert von 1.2
cm/a oder 1.1x106 m3 entspricht. Dies bedeutet einen Verlust an Speichervolumen in Höhe
von 0.08% pro Jahr.

Beim Versuch, die Ergebnisse der Bathymetrie (967 m³/km²/a) mit denen des PSIAC
Modells (95-250 m³/km²/a) zu vergleichen, werden klare Unterschiede deutlich.

Die Analyse vorheriger Studien und die teilnehmende Bewertung der anthropogenen
Einflussfaktoren zeigen einen deutlichen Einfluss derselben auf die Hydrologie des Sees.
Sedimentation und zunehmender Gebietsabfluss werden als Belastung (pressure) für den
Seewasserstand (Status, state) angesehen. Diese Belastung ist eine Folge verschiedener
Treiber (drivers: Landnutzungsänderung, Abholzung, unangemessene Nutzung der
Landressourcen). Diese direkten Treiber werden von indirekten Treibern wie
Bevölkerungswachstum, landwirtschaftliche Entwicklungen, Feuerholznutzung, sozio-
ökonomische Änderungen sowie den existierenden Besitzverhältnissen beeinflusst.
Interessanterweise existieren vielversprechende politische Instrumente (response), die das
integrierte Management des Sees und seines Einzugsgebietes unterstützen. Das Versagen
der Implementierung dieser politischen Instrumente wird von den betroffenen Stakeholdern
beklagt.

ii
 Chapter 1: Introduction

Chapter 1. Introduction

1.1. Background information

A lake is generally defined as an inland body of fresh or saline water, appreciable in size (i.e.
larger than a pond), and too deep to permit vegetation (excluding submergent vegetation) to
take root completely across its expanse (Schertzer et al., 2012). They are subjected to multiple
interacting stressors (Christensen et al., 2006) such as atmospheric, meteorological, geological,
hydrological and astronomical influences (Altunkaynak, 2003). The human-induced changes
are also found to affect the hydrology of lakes in many parts of the world.

One of the most significant and broadly impacting effects of climate variability on lakes is the
changes in water level. Such changes reflect an alteration of the lake water balance, which can
result from changes in: precipitation, surface runoff, ground water flow, and evaporation from
the lake surface (Elsawwaf and Willems, 2012; Lenters et al., 2005). The water in a lake is
balanced by the basic hydrological relationship in which the change in water storage is
governed by the water input and output to the system (Limgis, 2001).

In the 1960s, lakes throughout East Africa were rising (Lamb, 1966) resulting from a series of
remarkably wet years (Flohn, 1987; Nicholson, 1995). The spatial extent and the magnitude of
fluctuations were considered as a signal to major global climate change (Lamb, 1966).
According to Arnell et al. (1996) and Bergonzini (1998), African lakes are known to be very
sensitive to climate variations with special sensitivity of closed lakes. The impact of non-
climatic factors on water level variability was also reported by different scholars in Ethiopia,
such as Görner et al. (2009) and Belay (2009).

In addition to climatic and non-climatic factors, the type of lake can also influence the water-
level fluctuation character of a lake (Deganovsky and Getahun, 2008). For instance, lakes
without outlets (called closed or terminal lakes) fluctuate in a greater degree as compared to
open lakes (Langbein, 1961). The Ethiopian Rift Valley Lake Basin contains such terminal
lakes that make the basin hydrologically sensitive. These terminal lakes are also sensitive to
pollution by constantly taking pollutants without chances of releasing them. Xu (2011), Zhao
et al. (2009), Milliman et al. (2008) and many other researchers concluded that climate change
and human activities are the main driving forces that affect the hydrological status of a given

3
 Chapter 1: Introduction

lake. However, the discrimination between these two causes is still one of the major
challenges in hydrology (Yang et al., 2012).

The Ethiopian Rift Valley (figure 1.1) is characterized by a chain of lakes varying in size,
hydrological and hydrogeological settings. The water levels of some of these lakes showed
dramatic changes in the last few decades (Alemayehu et al., 2006).

Figure 1.1. Locations of the Ethiopian Rift Valley lakes Source: Alemayehu et al. (2006)

The lakes in the Rift Valley are situated within three sub-basins: Awash basin (Lake Koka,
Beseka, Gemari, Abe), which is located in the Northern Main Ethiopian Rift (MER), the lakes
region (Lake Ziway, Langano, Abiyata, and Shalla) occupying a central part of the MER, and
the Southern basin (Lake Hawassa, Abaya, Chamo, and Chewbahir). Hydrologically, the
basins form separate units, but hydrogeologically they form a unique system within the rift due
to the underground interconnection by NE-SW aligned regional faults (Belay, 2009;
Alemayehu et al., 2006).

4
 Chapter 1: Introduction

Lake Hawassa has been experiencing a progressive rise in water level during the past two
decades (1981-1998) (Gebreegiziabher, 2004; WWDSE, 2001). The concern of this rise
achieved its peak in the aftermath of the extreme flooding of the surrounding area as a result of
extreme rise in 1998/99. It was because of this problem that the regional government funded
an extensive studies conducted by WWDSE (2001) and WRDB (2007). Regarding the lake
level rise, WWDSE (2001) explained that it was caused by deforestation which increased the
runoff and siltation of the lake. However, Ayenew and Gebreegiziabher (2006) argued that the
justifications are speculative rather than supported by scientific data. Another recent project
was undertaken between 2008-2010 by Ministry of Water Resources with the aim of
generating a development master plan for Ethiopian Rift Valley Basin in general (MoWR,
2008; 2009; and 2010). The governmental funding of the above three projects indicates the
level of concern of policy makers towards the management of water resources in the region.
These projects produced extensive information including the first bathymetry map of the lake,
land use dynamics, soil and geological classifications, gully networks, and supportive
information about the lake. They were development-oriented than dealing with scientific
arguments.

There were earlier researches to understand the hydrology of Lake Hawassa and many studies
associated the causes of the water-level rise of Lake Hawassa with climate changes (Lamb et
al., 2002; Ayenew, 2006; Deganovsky and Getahun, 2008; Gebreegziabher, 2004; WWDSE ,
2001; MoWR, 2008; Bewketu, 2010). Other researchers considered the problem as resulting
from land use changes that in turn affect the runoff generation mechanism (Lamb et al, 2002;
Gebreegziabher, 2004; Ayenew, 2004; MoWR, 2008; Bewketu, 2010). Less number of studies
reported the role of sedimentation process into the lake (Esayas, 2010; Gebreegziabher, 2004;
Geremew, 2000). The involvement of tectonic processes that affect the ground water flow
regime is also recognized by Ayenew (2006), WWDSE (2001) and others. Generally, the
underlying cause of the water level rise of Lake Hawassa is still a spot of confusion.

1.2. Problem statement

Over the past few years, several researchers have studied the long-term water balance of Lake
Hawassa, such as Gebreegiziabher (2004), Ayenew (2004), Deganovsky and Getahun (2008),
WWDSE (2001), Ayenew and Gebreegiziabher (2006), Gebremichael (2007), and
Shewangizaw (2010). Land use/cover changes have also been studied by Wagesho et al.

5
 Chapter 1: Introduction

(2012), Beetle (2009), and WWDSE (2001). Despite the number of studies and their
importance, the cause of Lake Hawassa‟s water level rise has not been concluded and not yet
explicitly investigated.

Limitations of previous studies:

(1) Previous studies focused primarily on long-term variations in the lake level which have
disadvantage of obscuring particular temporal responses of the lake to extreme events. A
better insight could have been grasped if the analyses had focused on both long-term
variations (trends) and temporal extreme events (regime shift) simultaneously. A trend is
likely to continue in the future but does not necessarily change the stationarity of the
system; but a regime shift is likely to persist until a new regime shift takes place (Villarini
et al., 2011);
(2) The implicit assumptions of the so-called stationarity (stability of mean values over time)
of hydro-climatic variables can be erroneous unless the presences of shifts in mean values
are statistically tested and the causes of those shifts assessed. Change points violate
stationarity and so their identification becomes an important issue (Breaker, 2007;
Wagesho et al., 2012, and Box and Jankins, 1970);
(3) No previous attempt was made to study the impact of lake-bed sedimentation on the
storage capacity of the lake; and
(4) Some of the previous studies analyzed part of the story and their results should be
synthesized in a logical way to show the cause-effect chain of the main environmental
problems by applying a suitable analytical model, such as DPSIR framework.

6
 Chapter 1: Introduction

1.3. Overall objectives of the thesis

The general aim of this research is to investigate the effect of natural and anthropogenic
factors on the temporal variability of Lake Hawassa water level. Even though the specific
objectives are within the respective chapters, the following list compiles the overall objectives:

 To test the coherence between Sea Surface Temperature (SST) anomalies and lake
level variability;
 To test the presence of significant variability over-time (trend) and variability across-
time (regime shift) of the hydro-climate variables (lake level, rainfall, stream flow, and
evaporation);
 To simulate the long-term variability of the lake level using spreadsheet water balance
model;
 To quantify the effect of sedimentation on the storage capacity of the lake by
conducting new bathymetry and comparing the results with the existing one;
 To assess the linkage of in-lake sedimentation to the watershed characteristics by
applying PSIAC model;
 To synthesize the preliminary cause-effect chain responsible for the lake level rise by
employing the Drivers-Pressures-State-Impact-Response (DPSIR) analytical
framework.

1.4. Thesis architecture and general approach

This thesis contains nine chapters where the first three cover "general introduction",
"description of the study area", and "literature review" consecutively. As shown in figure 1.2,
chapter four deals with an investigation into the impact of climate shifts and ENSO
phenomena on the hydrological status of Lake Hawassa; chapter five presents simulation of
the long-term lake level variability and computation of the magnitudes of water balance
components. Moreover, chapter six is about an investigation of the effect of recent
sedimentation on the storage capacity of the lake; whereas chapter seven is about tracing the
hot-spots of sediment sources in the watershed. Finally, chapter eight deals with an assessment
of the anthropogenic factors that affect the hydrology of the lake in a preliminary and
qualitative manner while chapter nine presents a synthesis of the causal-links by linking
natural and anthropogenic factors prevailing in the hydrosystem in a comprehensive way.

7
 Chapter 1: Introduction

The impact of

Chapter 4
climate shifts and
ENSO phenomena on the
hydrological status
of Lake Hawassa

Simulation of

y
ilit
long-term water

Chapter 5
b
ria
level variability of Lake

va
Hawassa in the presence of

te
ma
n regime shifts in water
tio

Cli
balance components
ula
What are sim
m
the causes of ste
Sy The effect of recent

Chapter 6

Chapter 9
variations in Lake Hawassa

Synthesis
sedimentation on the
water level over time (trend) Sedimentation
storage capacity of Lake
and across time So Hawassa
ur
(regime shift)? ce
so
fs
ed
im The linkage of
en
Hu

t sedimentation in

Chapter 7
ma

Lake Hawassa to the

nf

watershed characteristics: an
ac

application of
tor

PSIAC model
s

Participatory assessment

Chapter 8
of anthropogenic factors
affecting the hydrology of
Lake Hawassa: an appication
of DPSIR framework

Figure 1.2. Thesis architecture and general approach

8
 Chapter 2: Description of the study area

Chapter 2. Description of the study area

2.1. Location
Lake Hawassa watershed is located in the central North-East of the Ethiopian Rift Valley
Basin (figure 2.1) and covers an area of 143,651 ha. It contains five sub-watersheds:
Dorebafena-Shamena, Wedesa-Kerama, Tikur Wuha, Lalima-Wendo Kosha and Shashemene-
Toga. The geographical co-ordinates of the watershed are 60451 to 70151 North and 380151 to
380451 East latitude and longitude respectively. The city of Hawassa, named after the lake, is
located at 275 km south of the capital city-Addis Ababa and is established in the very eastern
shore of the lake (MoWR, 2010).
Main Ethiopian Rift Valley Basin (MER)

00 km 2
52,0

B
Lake Hawassa Catchment
3D v (Clipped SPOT5 image)
iew
of L
ak e Ha
was
sa

m 2
=96 k
ace area
Surf

D C

Figure 2.1. Maps of the study area at different scales

A: The 12 river basins of Ethiopia source: Vilalta (2010) ;
B: The Main Ethiopian Rift Valley basin source: MoWR (2010);
C: Lake Hawassa watershed as clipped from SPOT5 satellite image source: own study;
D: 3D view of Lake Hawassa as generated by ArcGIS10 from the 1999 bathymetry map source: own
study

9
 Chapter 2: Description of the study area

2.2. Climate and Agro-ecology

According to Legesse et al. (2003), the watershed is characterized by three main seasons. The
long rainy season in the summer from June-September is known locally as Kiremt and is
primarily controlled by the seasonal migration of the inter-tropical convergence zone (ITCZ),
which lies to the north of Ethiopia at this period. The wet period (locally named as Kiremt)
represents 50-70% of the mean annual total rainfall. The dry period (locally named as baga)
extends between October and February when the ITCZ lies to the south of Ethiopia (Legesse
et al., 2004). During March and May, the "small rain" season (locally named as belg) occurs
when about 20-30% of the annual rainfall falls. The climate in the area varies from dry to sub-
humid according to the Thornthwaite‟s system of defining climate or moisture regions (Dessie,
1995).

As computed from the long-term (1973-2010) rainfall record of Hawassa meteorological

station, the annual average magnitude is computed to be 961 mm and distributed as 50% for
Kiremt (June-September); 20% for baga (October-February) and 30% for belg season (March-
May). Figure 2.2 shows the long-term average monthly distribution of rainfall and
temperature at Hawassa meteorological station.
140
120 a
100
80
mm

60
40
20
0
Feb

Sep
Jan

Nov
Dec
Apr

Aug
Mar

May

Jul

Oct
Jun

Rainfall

30
25
b
20
oC

15
10
5
0
Feb

Sep
Jan

Nov
Dec
Apr

Aug
Mar

May

Jul

Oct
Jun

Minimum temperature Maximum temperature

Figure 2.2. Distribution of monthly rainfall (a) and temperature (b) at Hawassa Station

10
 Chapter 2: Description of the study area

The isohytal map of the watershed is shown in figure 2.3 and the rainfall time series of the five
meteorological stations in and around the watershed is shown in figure 2.4.

Figure 2.3. Isohytal map of Lake Hawassa watershed Source: Shamo (2008)

11
 Chapter 2: Description of the study area

Annual RF (mm)
1400,0 1300
1100

Annual RF (mm)
1200,0
900
1000,0 700

800,0 500
300

1984

2004
1974

1979

1989

1994

1999

2009
600,0
1973

1998
1978
1983
1988
1993

2003
2008
1400

Annual RF (mm)
1200

1000

800

600

1978

1983

1988

1993

1998

2003

2008
1400 1600
1300 1400
Annual RF (mm)

1200
Annual RF (mm)

1200
1100
1000 1000
900 800
800
600
700
2009
1984

1989

1994

1999

2004

400
1974

1979

1984

1989

1994

1999

2004

2009

Figure 2.4. Partitioning of the watershed by Thiessen´s polygon (the red lines show average
annual rainfall values. The green line for Shashemene station shows the presence of change in
mean values as tested by Pettit’s homogeneity statistics that detect single breaking point in a
series) (The raw data was obtained from the local Meteorology Agency)

2.3. Topography
Majority of the watershed is flat to gently undulating but bounded by steep escarpments. The
altitude ranges from 1,680m at Lake Hawassa to 2,700m on the Eastern escarpment: an
altitude range of 1,020m. Most slopes (56%) are flat to gentle (0-8%) with a further
33% moderately sloping (8-30%) and only 5% steep to very steep (>30%) (MoWR, 2010).
Figures 2.5, 2.6, and 2.7 demonstrate the topographical variations in the watershed.
12
 Chapter 2: Description of the study area

Figure 2.5. Hill shade view of the watershed landscape as processed from DEM
(The unprocessed SRTM DEM of 30 x 30 m resolution was obtained from Ministry of Water
Resources)

Figure 2.6. Three dimensional view of topographic diversity of the watershed (Elevations are
exaggerated to some extent) (The unprocessed SRTM DEM of 30 x 30 m resolution was
obtained from the Ministry of Water Resources)

Figure 2.7. Elevation range of Lake Hawassa watershed Source: Abraham (2007).
13
 Chapter 2: Description of the study area

2.4. Soils
Twelve soil types are identified in the watershed (MoWR, 2010) as shown in figure 2.8 and
described in table 2.1.

Figure 2.8. Soil types of the watershed Source: MoWR (2010). For details see table 2.1.

14
 Chapter 2: Description of the study area

Table 2.1. Description of soil types in Lake Hawasssa watershed

Soil type Soil type description
code
1 LVcr-Mde Well drained; deep to very deep; dark brown to dark reddish brown; fine and medium
textured; moderate, fine to coarse sub angular blocky structured chromic luvisols (eutric)
developed on medium to high gradient mountains with slope of 8-30%.
2 CMeu-Ede Excessively drained; moderately deep to deep (gravely and pumice below 60cm); very dark
grayish brown and very dark gray; coarse textured; weak, very fine crumb and massive
structured vitric Andosols developed on level plain land form with slope of 0-2%.
3 CMeu-Rcd Well drained; deep to very deep; medium textured; weak to moderate fine and medium sub
angular blocky structured; slightly to non-calcareous eutric cambisols developed on rolling
plain with a dominant slope range of 5-15%.
4 LVha-Rcd Well drained; deep to very deep; dark brown to dark reddish brown; fine and medium
textured; weak to moderate medium sub angular blocky structured; non-calcareous haplic
luvisols developed on rolling plain with a dominant slope range of 5-15%.
5 CMeu-PLab Well drained; very deep; dark brown over very dark grayish brown; medium textured;
weak, medium and coarse sub angular blocky and single grain structured eutric cambisols
developed on level plain land form with slope of 0-2%.
6 CMeu-PLbc Well drained; very deep; dark brown over very dark grayish brown; medium textured;
weak, medium and coarse sub angular blocky and single grain structured eutric cambisols
developed on level plain land form with slope of 2-5%.
7 ANvi-PLab Excessively drained; moderately deep to deep (pumice below 45cm); very dark grayish
brown; coarse textured; weak, medium crumb and massive structured vitric Andosols
developed on level plain with slope of 0-2%.
8 CMcr-Mde Well to excessively drained; moderately deep; dark reddish brown; fine and medium
textured; weak to moderate fine and medium sub angular blocky structured; non calcareous
chromic cambisols developed on a very steep topography with slope >8%.
9 ANvi-Rcd Well to excessively drained; moderately deep to very deep; dark brown to dark yellowish
brown; medium and coarse textured; weak fine and medium sub angular blocky structured
vitric Andosols developed on rolling plain (0-15% slope) with few to many fine pumice
gravels.
10 CMvr-PLab Very poorly drained; deep to very deep; very dark grey to black; fine and medium textured;
moderate, medium sub angular blocky structured non-calcareous vertic cambisols
developed on flat topography (0-2%) of alluvial plain landforms.
11 CMeu-Rbc Well drained; deep to very deep; medium textured; weak to moderate fine and medium sub
angular blocky structured; slightly to non-calcareous eutric cambisols developed on rolling
plain with dominant slope range 2-8%).
12 LP-Hde Excessively to well drained; very shallow; dark brown to very dark yellowish brown;
medium textured; weak to moderate medium sub angular blocky structured; friable moist;
slightly sticky and slightly plastic wet; slightly to non-calcareous leptosols developed on a
hill with slope >8%.
Source: MoWR (2010)

2.5. Land use/cover

According to MoWR (2010), land use in the watershed is dominated by cultivation which
occupies 61% of the total area (or 66% of the land area) with intensive cultivation. The major
land cover splits into smallholder cultivation (95%) of which 31% is cereals and perennials
(CI3) and 64% cereals only (CI4) and mechanized cultivation (5%) most of which is state
owned rather than private. Intensive cultivation with perennial crops occurs in the eastern hills
with cereal cultivation dominating the western, southern and northern areas.

15
 Chapter 2: Description of the study area

Other important land covers include disturbed and plantation forests in the Wendo Koshe hills
and around Wendo Genet comprising 3% of the area; dense and open shrubland in the Wendo
Koshe hills and west of Cheleleka comprising 6%, grassland (11%) comprising open grassland
in the Wendo Koshe hills (3%), in association with marshland at Cheleleka (4%), in
association with moderate smallholder cultivation in the Eastern hills (3%) and wooded
grassland (1%) in the eastern hills.

Table 2.2. Types of land cover in Lake Hawassa watershed

Land cover Land cover % of sub- % of sub-
[ha] basin basin land
area
Urban – U 2,531 1.76 1.88
Intensive Mechanized Cultivation (Private) CIMP 1,015 0.71 0.76
Intensive Mechanized Cultivation (State) CIMS 3,287 2.29 2.45
Intensive Smallholder Cultivation CI3 27,664 19.26 20.59
Intensive Smallholder Cultivation CI4 56,055 39.02 41.73
Total Intensive Smallholder Cultivation 88,021 61.27 65.53
Disturbed High Forest – FD 3,599 2.51 2.68
Plantation Forests – FP 328 0.23 0.24
Dense Shrubland – SD 2,104 1.46 1.57
Open Shrubland – SO 5,995 4.17 4.46
Open Grassland – GO 3,534 2.46 2.63
Open Grassland with moderate smallholder cultivation –
GO/CM3 4,520 3.15 3.37
Open Grassland and Marshland - GO/MA 5,333 3.71 3.97
Wooded Grassland- GW 1,559 1.09 1.16
Marshland – MA 2,335 1.63 1.74
Open Woodland - WO 422 0.29 0.31
Dense Woodland - WD 5,744 4.00 4.28
Bare Eroded Land with scattered vegetation – EES 7,913 5.51 5.89
Bare Rock – ER 388 0.27 0.29
Total land 134,328 93.51 100
Water 9,324 6.49
Watershed total 143,651 100
Source: MoWR (2010)

2.6. Geology
According to MoWR (2010), the watershed forms the Corbetti caldera with the steep western
and eastern escarpments of the caldera walls (figure 2.9). The majority of the watershed, the
flat caldera floor, is composed of lacustrine sediments of Pleistocene age, evidence of the
gradual desiccation and infilling of the former Lake Shallo. The Wendo Koshe hills to the
north-west of Lake Hawassa are composed of pumice, unwelded tuffs, obsidian and pitchstone
while other hills (Alge, Kike, Kuwe etc) and the steep escarpment immediately to the north of

16
 Chapter 2: Description of the study area

Lake Cheleleka are rhyolitic and trachrytic lava flows. The hills forming the eastern
escarpment are composed of Nazret silicicvolcanics comprising ignimbrites, unwelded tuffs,
ash flows, rhyolites and trachytes while the land to the east of the Wendo Koshe hills is
underlain by rocks of the Dino formation comprising ignimbrites, tuffs, water lain pyroclastics
and occasional lacustrine beds.

Figure 2.9. Geological map of Lake Hawassa watershed

[The red lines are the main roads crossing the watershed and the green dots are well points and
not relevant in our case] Source: WRDB (2007)

2.7. Morphology of Lake Hawassa

Lake Hawassa is the smallest and the highest in altitude among the Great Ethiopian Rift
Valley lakes (1680 m.a.s.l) and located at the geographic coordinates of Lake 71060 N and
381330 E between the Ziway-Shalla lakes to the north and Lakes Chamo and Abaya to the
south. The lake lies within a nested caldera complex and is predominantly underlain by highly
faulted ignimbrites and other silicic pyroclastic deposits (Kazmin, 1979 as cited in Lamb et al.,
2002).

When we compare the elevation (figure 2.10) of Lake Hawassa (1680 m) with lake Ziway
(1636 m), Langano (1585 m), Abiyata (1578 m), Shalla (1550 m) and Abaya and Chamo

17
 Chapter 2: Description of the study area

(~1180m) (Gebreegziabher, 2004), it is possible for ground water to flow from Lake Hawassa
to low lying lakes when hydrogeological condition permits.

Figure 2.10. Elevation diversity of some Rift Valley lakes Source: Gebreegziabher, 2004)

The bathymetry survey of this research, which was conducted on January 2011, revealed that
the maximum depth of the lake is 23.4 m and an average depth of 13.3 m. As extracted from
satellite imagery, the length of north-to-south axis is 16 km and the east-west axis is 8 km. The
water storage capacity of the lake is 1.36 km3 (Ayenew et al., 2007). The elevation-area-
volume curve of the lake is shown in figure 2.11.

Table 2.3. Summary of physical characteristics of Lake Hawassa

Parameters Size and location References
1 Watershed area (including 1436.5 km2 MoWR (2010)
the lake) @ Lat. 60451 to 70151North and
@ Long. 380151 to 380451East
2 Maximum lake depth 23.4 m (on Jan. 2011) Own study
@ Lat. 7.082019 deg. and
Long. 38.45225 deg.
3 Average lake depth 13.3 m (on Jan. 2011) Own study
4 Lake surface area 96 km2 Own study
5 Water storage volume 1.36 km3 Ayenew et al. (2007)
6 Residence time 1.3 year Ayenew et al. (2007)
7 Lake Surface area (m ) = 4*106 x d+ 9*106
2
Gebreegziabher
(rating curve) (where d is the actual depth of the lake m) (2004)
8 Lake volume (m3) = 2*106x d2+ 1*107d-5.95*107 Gebreegziabher
(rating curve) (2004)

18
 Chapter 2: Description of the study area
FIGURE 3.2 AREA - CAPACITY CURVES OF LAKE AWASA

120 1600

1400
100
1200

80

VOLUME(mcm)
AREA(Km^2)
1000

60 800

600
40 area
400

20
200
volume
0 0
1658.7 1660 1662 1664 1666 1668 1670 1672 1674 1676 1678 1680 1682
ELEVATION(m.a.s.l)

Figure 2.11. Elevation-area-volume curve for Lake Hawassa Source: WWDSE (2001)

19
 Chapter 3: Literature review
Chapter 3. Characterization of the water level variability of the Main Ethiopian Rift
Valley (MER) lakes

3.1. Introduction
Water level variability of a given lake results from water exchange characteristics within its
watershed (Vuglinskiy, 2009). Lake levels fluctuate naturally in response to climatic and
hydrological factors within natural amplitudes (Zohary and Ostrovsky, 2011) as far as they are
undisturbed by external forces such as climate anomalies or anthropogenic factors. Scheffer
and Carpenter (2003) also remarked that the usual state of affairs in nature is to fluctuate
around some stable average. The seasonal and annual water level fluctuation of lakes is a
common phenomenon in every lake. Such fluctuations are usually due to the differences
between precipitation and evaporation in that specific season (Kinshiro, 1974). These
dynamics are controlled by the balance between inputs and outputs of water, which are in turn
controlled by the hydrological processes (Hayashi and Kamp, 2007). These natural
fluctuations are an inherent feature of lake ecosystems and essential for the survival and well-
being of many species that have evolved to suit their life cycle to those fluctuations (Gasith
and Gafny, 1990).

In the Main Ethiopian Rift Valley region, there has been no increasing/declining precipitation
trend for the last 50 years except for the inter-annual and seasonal variations (Ayenew, 2004).
This kept the level of some lakes constant, with little or no change (Ayenew, 2007) but some
of the lakes in the region experienced either an increasing or decreasing trend (Belay, 2009;
Ayenew, 2004; Gebreegiziabher, 2004). These fluctuations are disturbing the stability of the
ecosystems, putting serious impacts on the lives of many animals and plants around the lakes
(Bewketu, 2010). Reviewing the characteristics of lake level variability in the region is
relevant to this study in providing an insight into the similarity or dissimilarity of such
variability among the lakes in the region. The hypotheses of this study arise from this review.

3.2. Objectives of the chapter

The aim of this chapter is to characterize the lake level variability of Rift Valley lakes in
general and Lake Hawassa in particular. The dominant processes controlling the lake level
variability are reviewed. Such characterization is expected to identify research gaps and
provide information while designing the hypotheses of the main thesis work. Diverse

20
 Chapter 3: Literature review
particularities of lake level regimes in the Rift Valley Basin are intended to answer the
question “what is common to these lakes?".

The lakes under consideration are: (1) Lake Ziway, (2) Lake Langano, (3) Lake Abiyata, (4)
Lake Shalla, (5) Lake Beseka, (6) Lake Hawassa, (7) Lake Abaya, and (8) Lake Chamo.

3.3. Methodology

3.3.1. Description of the study area and characteristics of the lakes

The Rift Valley Lakes Basin (RVLB) is one of the eleven major river basins in Ethiopia with a
total area of approximately 52,000 km2 (MoWR, 2010). The basin is characterized by a chain
of lakes varying in size, hydrological and hydrogeological settings (Alemayehu, et al., 2006).
It constitutes seven main lakes: Lake Ziway, Lake Langano, Lake Abiyata, Shalla, Lake, Lake
Abaya, and Lake Chamo (figure 3.1) where all are located south of the Ethiopian capital Addis
Ababa.

21
 Chapter 3: Literature review

Figure 3.1. Base map of the Ethiopian Rift Valley basin Source: MoWR, 2010)

22
 Chapter 3: Literature review
Table 3.1 and 3.4 depict the morphological characteristics of individual lakes in the Rift
Valley Basin as compiled from different sources. The water quality parameters are also
presented in table 3.2.
Table 3.1. Morphological characteristics of Rift Valley lakes
Altitude (m.a.s.l) Max. depth (m) Mean depth (m) Volume (km3)
1 Lake Ziway 1636 8.95 2.5 1.6
2 Lake Langano 1582 47.9 17 5.3
3 Lake Abiyata 1578 14.2 7.6 1.1
4 Lake Shalla 1558 266 87 36.7
5 Lake Hawassa 1680 22 11 1.34
6 Lake Abaya 1285 13.1 7.1 8.2
7 Lake Chamo 1233 13 6 3.3
8 Lake Beseka 1200
Sources: Wood and Talling (1988), Kebede et al. (1994), Chernet (1982), Ayenew (1998), Tessema
(1998), Halcrow and partners (1989), WWDSE (2001), Deganovsky et al. (2004), and Görner et al.
(2009)

Table 3.2. Selected water quality parameters of the Rift Valley lakes
Parameter Ziway Abiyata Shalla Langano Hawassa Abaya Chamo Beseka
pH 8.37 9.60 9.80 9.04 9.00 9.07 9.48
EC (µS/cm) 453 47,915 46,075 1,937 867 1,218 1,966 7,155
Na (mg/l) 61 7,520 6,475 390 165 234 428
F (mg/l) 1.6 220.0 188.0 9.1 8.7 8.1 9.1
SAR 3.0 653 267 41.5 10.2 15.7 27.0
Source: MoWR (2009), Ayenew (1998), Wood and Talling (1988), and Halcrow and partners (1989)

3.3.2. Available data

Table 3.3 presents the magnitudes of available water balance components for the eight Rift
Valley lakes of Ethiopia and table 3.4 shows the relative surface areas of the lakes and their
watershed.

23
 Chapter 3: Literature review
Table 3.3. Water balance components of the eight Rift Valley lakes (the units are as appeared
in their respective literatures, no conversion made)
Inflow Outflow
Name of the lake P E A References
Sin Run GWi Sout GWo
1 Ziway (in 106 m3) 323 656.5 48 80.5 890 184 28 14.6 Ayenew (2004)
(mm) 750 1530 1720 Deganovsky and Getahun (2004)
(mm) 753 0.692km3 0.05km3 100 1740 200 (net) Vallet-Coulomb et al. (2001)
2 Langano (in x106 m3) 186 212 135.4 463 46 18.9 Ayenew (2004)
3 Abiyata (in x106 m3) 113 230 15 26.8 372 0 13 1.2 Ayenew (2004)
(in x106 m3) 97.2 179.87 13.92 290.97 0 0 Ayalew (2003)
4 Shalla (in x106 m3) 232 245 18 40 781 0 Ayenew (2004)
5 Hawassa (in x106 m3) 106 83.1 132 0 58 Ayenew (2004)
(mm) 950 1440 1440 0 570 Deganovsky and Getahun (2004)
(in x106 m3) 80.6 74 90 164.6 0 71 WWDSE (2001)
(in x106 m3) 106 83 131 0 58 Ayenew and Gebreegiziabher
(2006)
(in x106 m3) 90 167 148 Gebremichael (2007)
(in x106 m3) 98.9 54.9 44.44 178.93 0 Shewangizaw (2010)
(in x106 m3) 90.72 88.29 91.57 3.2 166.66 71.5 WRDB (2007)
(in x106 m3) 106 83.7 - 132 0 58 Gebreegiziabher (2004)
(in x106 m3) 106 83.7 - 132 58 Ayenew et al. (2007)
(in x106 m3) 52.5 Ayenew and Tilahun (2008)
6 Abaya (in x106 m3) 556 1900 Ayenew (2004)
(in x106 m3) 980 750 691 2009 Belete (2009)
(mm) 730 1080 1700 Deganovsky and Getahun (2004)
7 Chamo (in x106 m3) 406 900.9 Ayenew (2004)
8 Beseka (in x106 m3) 22 30 52.8 98.8 Ayenew (2004)
(in x106 m3) 24.4 7.7 33.8 61.8 0.22 Belay (2009)
P=over lake precipitation; Sin= stream flow; Run= surface runoff from the watershed; E= evaporation from the
lake; Sout= stream outflow; A= abstraction; GWi= ground water inflow; GWo= ground water outflow

Table 3.4. Results of characterization based on specific watershed

Names of the Rift Valley lakes Surface area (km2) Watershed area (km2)
1 Lake Ziway 442 7025
2 Lake Langano 241 1600
3 Lake Abiyata 176 1630
4 Lake Shalla 329 3920
5 Lake Hawassa 90 1250
6 Lake Abaya 1162 17300
7 Lake Chamo 551 2210
8 Lake Beseka 43 505
Sources: Ayenew (2004), and Deganovsky and Getahun (2004)

3.3.3. Methods
This chapter intended to investigate the hydrology of Main Ethiopian Rift Valley lakes by
assessing their long-term water balances and their morphological characteristics. Assuming the
fundamental similarity of all lakes, the review adopted two different approaches to estimate
the natural responses of the lakes. These techniques of characterizing the lake level regime are
suggested by Szestzay (1974) based on long-term water balances and another suggestion by
Litinskaya (1973) based on morphological nature of lakes. The methods are meant to show the

24
 Chapter 3: Literature review
expected natural behavior of the lake hydrology and deviations from these are considered to be
shifts from the natural state. The following sections discuss the methods in detail.

[Link]. Water balance approach to characterize the lake level regimes

An earlier publication of Szesztay (1974) suggested the possibility of classifying lakes based
on their water balance as shown in figure 3.2 and 3.3. Inflow factor (i), outflow factor (o) and
aridity factor (a) are the basic criteria for characterization of the lakes. The basic equations of
these factors are presented below:

1. () .….……………... (3.1)

2. ( ) .………......... (3.2)

3. ( ) ……………..………………..….... (3.3)

A lake which belongs to one of the nine categories of figure 3.2 and 3.3 is considered as
having particular characteristics in terms of stability of the water balance and the factors
controlling water level fluctuation. For instance, the quadrant I-O represents those lakes
which are flow-dominated and equilibrium condition of their water balance are quickly
followed by corresponding changes in the height and regime of the water level. The quadrant
P-E comprises "atmosphere-controlled" lakes with self-regulating mechanism responsive to
climatic changes. The quadrants IP-E and I-E are expected to accumulate short term variations
of precipitation which in turn increase the imbalance during extreme dry and wet periods. The
other five quadrants of the scheme (I-OE, IP-OE, P-OE, P-O, and IP-O) are conceived as
representing intermediate situations between the "flow-controlled" and "climate-controlled"
lakes.

[Link]. Morphological approach to characterize the lake level regimes

This approach is based on the suggestion by Litinskaya (1973). In this approach, it is
recommended to use the term specific watershed (F) which is computed as:

Specific watershed (F) = lake basin area / lake surface area ……….…….. (3.4)

According to the approach, the lakes would be classified into three groups based on the
magnitude of specific watershed that is considered as a proxy to characterize the level-regime

25
 Chapter 3: Literature review
of the lakes. Those lakes having specific watersheds less than 10 are assumed to have stable
lake level regime with mean annual amplitude of fluctuation ranging from 30 to 65 cm. The
other category includes those lakes having specific watersheds ranging from 10 to 50 cm.
These lakes are expected to be less stable in terms of increased annual fluctuation (mean
annual amplitude of water-level fluctuations rises 50 to 130 cm). The third category of lakes
comprises those lakes with specific watershed exceeding 50 cm. The mean annual amplitude
of lake level variability in this case increases to 110 to 210 cm.

26
 Chapter 3: Literature review
3.4. Results and discussion

3.4.1. Classification of the lakes based on their long-term water balance

Table 3.5 presents the computational results of inflow factor, outflow factor and aridity using
equations 3.1, 3.2, and 3.3. The grouping of these lakes into their respective quadrants based
on their calculated particularities is also presented in table 3.5, figure 3.2, and 3.3.

Table 3.5. Results of inflow, outflow, aridity, and the corresponding quadrants
Inflow factor Outflow Aridity(a) Without With aridity
(i) factor (o) aridity factor factor
1 Lake Ziway 69.0 22.6 2.5 I-E* IP-E*
2 Lake Langano 65.1 12.3 2.5 IP-E* IP-E*
3 Lake Abiyata 68.6 3.7 3.1 I-E* I-E*
4 Lake Shalla 56.6 0.0 3.4 IP-E* I-E*
5 Lake Hawassa 53.3 23.2** 1.5 IP-E* IP-E*
6 Lake Abaya 59.5 0.0 2.6 IP-E* IP-E*
7 Lake Chamo incomplete incomplete 2.2
8 Lake Beseka 79.0 0.0 4.5 I-E* I-E*
* Interpretation: Climate controlled (with little role of inflow)
**The value represents the ground water outflow
100 I-E I-OE I-O
Inflow factor (i)= I: (I+PL) percent

80 Beseka
Ziway
Abiyata
IP-E IP-O
IP-OE
60 Abaya
Langano
Hawassa
40 Shalla

P-E P-OE P-O

20

0 20 40 60 80 100

Outflow factor (o)=O: (O+E) percent

Figure 3.2. Classification of lakes by water balance criteria (aridity factor is not included)

27
 Chapter 3: Literature review
10
I-E I-OE I-O

Beseka

Aridity factor (a)= E/P

3
Abiyata
Shalla

I=
I=2

5P
IP P Ziway
-
Abaya E
Langano
2
Hawassa
I=0

I=20P
.5 IP
P-E P - OE

1 P-O
E
IP-
O
P-O
0

0 20 40 60 80 100

Outflow factor (o)=O: (O+E) percent

Figure 3.3. Classification of lakes by water balance criteria (aridity factor included)

The water balance analyses show that most of the lakes have similar characteristics in terms of
their sensitivity to climate variability. This similarity is depicted in both cases of "with" and
"without" the use of aridity factors as classification criteria. All of the lakes are under I-E or
IP-E quadrant, and these two quadrants are known for their dominance in climate (with some
exceptions) during extreme dry and wet periods in which runoff from the watershed increases
the imbalance.

3.4.2. Classification of the lakes based on their morphology

Based on equation 3.4, the ratio of watershed area with lake surface area was computed and
results are presented in table 3.6 below.

28
 Chapter 3: Literature review
Table 3.6. Results of characterization based on specific watershed
Surface area Watershed Specific About level-regime Expected mean
(Km2) area (km2) watershed annual amplitude
(cm) **
1 Lake Ziway 442 7025 16 Moderately stable 50-130
2 Lake Langano 241 1600 6.6 stable 30-65
3 Lake Abiyata 176 1630 9.3 stable 30-65
4 Lake Shalla 329 3920 12 Moderately stable 50-130
5 Lake Hawassa 90 1250 14 Moderately stable 50-130
6 Lake Abaya 1162 17300 15 Moderately stable 50-130
7 Lake Chamo 551 2210 4 stable 30-65
8 Lake Beseka 43 505 11.7 Moderately stable 50-130
**The expected amplitudes are as suggested by Litinskaya (1973)

The result shows that lakes of mean stable level regime are dominant in the basin (Ziway,
Shalla, Hawassa, and Abaya) and the rest are in the range of stable level regime (Langano,
Abiyata, and Chamo) indicating the potential of the lakes to naturally regulate the surface
runoff flowing into them from their watershed. This technique appears to underestimate the
role of climate on Lake Hawassa as compared to the report of Tesfaye (1982) in which Lake
Hawassa is found to be sensitivity to slight climatic changes.

3.4.3. Recent/actual situations of individual lake level regimes

As shown in figure 3.4, the long-term water level records of individual lakes. Each lake has
experienced particular rise and/or drop in water levels which cannot be explained by
monotonic trends (defined as the slow move up or down from the mean value and keep on
moving in the same direction over time). Table 3.7 also shows the monotonic trend of each
lake under study based on raw data from literatures (for the first six lakes) and based on the
results of previous studies for Lake Beseka and Lake Shalla.

29
 Chapter 3: Literature review
Lake Ziway Lake Langano

Lake level above local

1,8
1,4

Lake level above local

bench mark (m)
1,6

bench mark (m)

1,2 1,4
1,2
1
1
0,8 0,8
0,6 0,6
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006
0,4

1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006
Lake Abiyata Lake Hawassa
3,5

Lake level above local

Lake level above local

bench mark (m)

bench mark (m)

2,5
5

3 1,5

1 0,5
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006

1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006
Lake Abaya Lake level above local Lake Chamo
Lake level above local

4 4
bench mark (m)
bench mark (m)

3 3

2 2

1
1

0
0
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006

Lake Shalla Lake Beseka

Figure 3.4. Long-term lake level plots of the eight Rift Valley lakes
N.B: Analysis of regime shifts were done for the first six lakes and the red lines represent mean lake
level (above local bench marks) before the regime shift (µ1) and the green line after the regime
(µ2).The blue bold lines for Lake Shalla and Lake Beseka represent the measured water level and the
pink color for Lake Beseka, though irrelevant to this review, represents modeled lake level as reported
by Belay (2009). (Raw data source for Lake Ziway, Langano, Abiyata, Hawassa, Abaya and Chamo is
Bewketu (2010); the graph for Lake Shalla is from Crétaux et al. (2011); and for Lake Beseka is from
Belay (2009).

30
 Chapter 3: Literature review
Table 3.7. Monotonic trends of level of individual lakes
MK ** Interpretation
1 Lake Ziway 0.324 Increasing
2 Lake Langano 0.037 No trend
3 Lake Abiyata -0.492 Decreasing
4 Lake Shalla - -
5 Lake Hawassa 0.531 Increasing
6 Lake Abaya 0.363 increasing
7 Lake Chamo 0.106 No trend
8 Lake Beseka - Increasing
** MK  is the Mann-Kendall coefficient as shown in section 4.4.5.

[Link]. Lake Ziway

The lowest level of Ziway was recorded in June 1975 (0.13 m) and the maximum in
September and October 1983 (2.17 m). However, for the last three years of the late 1970s and
early 1980s, the level was slightly lower due to the dry years of the 1970s. The lake shows a
slight reduction after the late 1980s due to the abstraction of water for irrigation (Legesse and
Ayenew, 2006; Vilalta, 2010). The existence of land degradation in the watershed that induced
large scale sedimentation rate was reported by Legesse and Ayenew (2006) and Billi and
Dramis (2003).

[Link]. Lake Langano

Lake Langano experienced only small seasonal water level variations of about 1 m, and lower
inter-annual water level variations compared to other lakes in the basin (Vilalta, 2010;
Ayenew, 2001). The absence of considerable water abstraction and large ground water flow
from springs are considered to be the factors against its relative stability of lake level
variability. Lake-bed sedimentation is also estimated to the magnitude of about 0.5 to 0.6
cm/yr, with 85-95% water content (Legesse and Ayenew, 2006).

[Link]. Lake Abiyata

Lake Abiyata is a saline-alkaline type (Wood and Talling, 1988) and in terms of lake level
variability, it has experienced a drop of about five meters over the last three decades
(Alemayehu et al., 2006) and also found to be heavily impacted by human activities
(Alemayehu et al., 2007; Vilalta, 2010). Its size, for instance, was decreased by 25% over the
last thirty years because the lake water is under pressure due to the production of Soda Ash
using solar evaporation of brines from the lake and the maximum drop coincides with the time

31
 Chapter 3: Literature review
of large scale water abstraction (Legesse and Ayenew, 2006). But the inter-annual fluctuations
are controlled by climate variability. According to Legesse et al. (2004), this lake also reacts
more rapidly to an abrupt shift to wetter conditions than to dry conditions. The production of
Soda Ash has not taken place for the last three years of the reporting time because of the
significant decline in the water level (MoWR, 2008). The fluctuation of Abiyata follows the
same trend as Lake Ziway, with an average time lag of about 20 days. Any abstraction of
water in the Ziway watershed results in a greater reduction in the level of Abiyata than in
Ziway (Legesse and Ayenew, 2006).

[Link]. Lake Hawassa

The monthly and annual Hawassa lake level and Tikur Wuha stream flow showed an
increasing overall trend (Wagesho et al., 2012). The possible causes of the water-level rise of
the lake is associated to climate changes (Lamb et al., 2002; Ayenew, 2006; Deganovsky et
al., 2008; Gebreegziabher, 2004; WWDSE, 2001; MoWR, 2008; and Bewketu, 2010); the
upset of hydrological variables (Lamb et al., 2002; Gebreegziabher, 2004; Ayenew, 2004;
MoWR, 2008; and Bewketu, 2010); sedimentation process (Esayas, 2010; Gebreegziabher,
2004; and Geremew, 2000) and geological tectonic processes that affect the ground water flow
towards the lake (Ayenew, 2006 and WWDSE, 2001).

[Link]. Lake Abaya

Lake Abaya experienced the rise of about 3.35m between 1987-1998 (12 years) followed by
continuous drop of 3.12m in the years 1998-2006 and then rose by 0.91m between 2006 and
2007. While discussing these variations, Belete (2009) stated that these fluctuations are mainly
caused by precipitation as input and evaporation as output and limited role of deforestation
and agricultural expansion in the watershed. Even though the role is limited, the watershed
experienced an expansion of agricultural lands by close to 200% in the year 2000, while bush
land increased by 17% during the same period, which can be explained by continuous
deforestation for agriculture and charcoal production for commercial and community use.
Table 3.8 below shows the land use/cover changes in the watershed.

32
 Chapter 3: Literature review
Table 3.8. Land use/cover changes in Lake Abaya watershed
Land use/Land cover In the year 1986 (ha) In the year 2000 (ha) Changes in percent
Bush land 50459.8 59442.4 17.8
Wet land 31512.7 20790.8 -34
Forest 180832 143195 -20.8
Agriculture 24506.7 72254.3 194.84
Water 137734 137320 -0.3
Grassland 17150.2 9192.48 -46.4
Source: Belete (2009)

[Link]. Lake Chamo

This lake rose in the years 1989, 2006 and 2007 only and the El Niño event in 1997/1998
which caused heavy rainfall and runoff in southern Ethiopia didn‟t cause substantial lake level
rise (Awulachew, 2006) in contrast to many other Rift Valley lakes.

[Link]. Lake Shalla

Regarding Lake Shalla, the available literature is very limited. That might be due to the little
interest on the lake water because of its alkaline nature (Vilalta, 2010) which discourages its
use for irrigation purpose.

[Link]. Lake Beseka

Despite small inter-annual variations, the water level of Lake Beseka has been rising for more
than three decades which is evidenced by the quadrupled expansion of its surface area from
11.1 km2 to 39.5 km2 between 1973 and 2002 with the corresponding rise in lake level (Görner
et al., 2009). The main cause for this expansion in surface area and rise in lake level is the
increased ground water flow from the western part of the watershed. The discharges to the lake
in the form of hot springs constitute the major water inflow to the lake (Görner et al., 2009;
Belay, 2009; Williams, 1981; and Ayenew, 2004). It is estimated to be 51% of the total inflow
to the lake (Belay, 2009). Some investigators relate the phenomena to neotectonism (Ayenew,
1998; Tessema, 1998). The average annual increment of the lake was 0.2m and the level of the
lake has risen by four meters between 1976 and 1997 (Zemedeagegneh and Egizabher, 2004).
Due to the expansion and flooding, the loss of 57 human lives, inundation of about 35 km2 of
grazing land, and displacement of 910 people was reported. The Methara sugar plantation has
also been inundated and the company lost income from 161.55 ha of land (WWDSE, 1999).
Damages on the nearby railway line and highway caused a loss amounting to 2.6 million US$
(Tessema, 1998; Ayenew, 2004).

33
 Chapter 3: Literature review

Figure 3.5. Lake level rise of Lake Beseka in relation to climatic factors (temperature,
precipitation, evaporation) Source: Görner et al., 2009

3.5. Conclusions
The results of this chapter suggest that the hydrological statuses of most Ethiopian Rift Valley
lakes are not stable in terms of their lake level variability. Few of them such as Lake Abiyata
tend to be at the verge of extinction as was observed from its drastic and continuous drop in
level.

In the previous section in which each lake in the Rift Valley was separately assessed, one can
observe the similarity among the lakes, for instance, Lake Abaya and Lake Hawassa
experienced lake level peaks in the year of 1998/99 and both of the peaks were caused by short
term climatic variability. The analyses and syntheses of this review showed that long-term
monotonic changes provide limited information in explaining the dynamics of lake levels. The
lake level changes seem to be explained better with the consideration of specific periods and
the corresponding events. In addition, the extent of the problem is not the same on each lake
and each lake suffers from diverse factors and deserves individual and separate analyses. In
terms of research gap identification, it was found that there existed nearly no attempt to
estimate the impact of sedimentation on the storage capacity of the lakes. The explicit attempt
to study the relationship between Lake Hydrology and climate anomalies is also absent. The
upcoming thesis work benefits from the above research gaps. This thesis is focused on
assessing the causal links of water level dynamics of Lake Hawassa, where there is a clear
research gap observed.

34
 Chapter 4: The impact of climate shifts and ENSO phenomena on the hydrological status of Lake Hawassa

Chapter 4. The impact of climate shifts and ENSO phenomena on the hydrological status
of Lake Hawassa

4.1. Introduction
The significant rising trend of Lake Hawassa water level (as shown in section [Link]) is one
of the main environmental threats for the city of Hawassa, which has been established at the
eastern shore of the lake. It is still the subject of concern and center of debate among the
stakeholders since the last few decades especially in the aftermath of the 1998 flood that
caused displacement of resident population, destruction of properties and infrastructure by
inundating vast areas along the lake shore. According to WRDB (2007) and WWDSE (2001),
the lake level rise and the associated surface expansion affected about 162 urban and 2244
farmers‟ households, 13 different organizations, water supply schemes, 10 ha of sand quarry,
roads, and forestland. In monetary terms, the total physical damage was estimated to be
43,490,524 Ethiopian birr (about € 5.4 million).

The hypothesis of "climate-hydrology link" was conceived in this study after the recognition
of coincidence between the lowest lake level record in the year 1975 with a strong La Niña
year and the maximum lake level in 1998 with the strongest El Niño year (please compare
figure 4.4 and appendix 1). La Niña and El Niño are anomalies in ocean surface water
temperature. They are commonly termed as "teleconnections" (Wallace and Gutzler, 1981).
There are reports of coherence between lake levels and teleconnection signals. For instance,
Namdar-Ghanbari and Bravo (2008) reported the significant coherence between Great Lake
water levels and some teleconnection signals.

One of the questions to be answered in this study is whether there is a quantifiable coherence
between ENSO signals and water level variability of Lake Hawassa. Appropriate pairs of
monthly step time series data were undergone through spectral analysis for the explicit
estimation of "coherency" between these series. The Niño-3.4 index (N3.4), which is the
average SST anomaly within the region 5°S-5°N, 170°-120°W is used as a representative
index for ENSO phenomena. This index is usually employed to predict rainfall in Ethiopia
(Korecha and Barnston, 2007; Babu, 2009). It is one of the most widely used ENSO index
(Barnston et al., 1997).

The use of spectral coherence analysis is quite recent in the area of hydrology. The coherence
analysis in this study was made following the idea of Jenkins and Watts (1968) and

35
 Chapter 4: The impact of climate shifts and ENSO phenomena on the hydrological status of Lake Hawassa

Bloomfield (1976). This technique is employed to quantify analyze the relationship between
Niño 3.4 index and lake level data series. The significance of coherence resulting from this
technique suggests that changes in one series is related to changes in the other.

In addition to coherence analysis, the coincidences of significant regime shifts in lake level,
streamflow, and lake-evaporation with climate shifts and timing/intensity of ENSO
phenomena were investigated to further strengthen the result of coherence analysis towards a
better understanding of climate-hydrology link.

The concept of "regime" in hydrology tells us the temporal pattern of the variable under
discussion over a period of time and "regime shift" was originally proposed in relation to
oceanic ecosystem (Steele, 1996; Hare and Mantua, 2000) to describe sudden drastic changes
in temporal characteristics of a variable (Yang et al., 2012). The definition of climatic regime
shifts can be viewed as "differing average climatic levels over a multi-annual duration"
(Overland et al., 2006). Shifts in the mean are the most common type of shifts considered in
literature (Rodionov, 2004; 2005). The main driving forces of variability in hydrological
variables are climate change and human activities (Zhao et al., 2009; Xu, 2011).

Improved understanding of the effects of climate variability on the water level of Lake
Hawassa can help in managing the hydrosystem in general. According to Lenters et al. (2005),
the changes in water level reflect alteration of water balance components. So, the explicit
analysis of hydro-climatic variables together with their linkage with climate anomalies would
provide a better insight into the inherent variability of hydrological status of the lake.

4.2. Hypothesis and objectives of the chapter

The sort of hypothesis to be proved in this chapter is stated as:

"The water level variability of Lake Hawassa is linked to Sea Surface

Temperature (SST) anomalies. It is further studied, whether regime
shifts occur in the hydro-climatic variables corresponding to the
occurrence of North Pacific climate shifts and El Niño/La Niña events".

36
 Chapter 4: The impact of climate shifts and ENSO phenomena on the hydrological status of Lake Hawassa

In line with the hypothesis, the objectives of this chapter are:

 To analyze the coherence between data series of Niño 3.4 Index (N3.4) and Lake
Hawassa water level;
 To analyze the long-term trends (variation over-time) and sequential regime shifts
(variation across-time) for lake level, rainfall, streamflow, and lake-evaporation data
series; and
 To compare significant change points of the above hydro-climatic variables with the
timing and intensity of North Pacific climate shifts/El Niño/La Niña occurences.

4.3. Impact of El Niño/La Niña on climate variability of East Africa

National Oceanic and Atmospheric Administration - NOAA's ([Link]) and many
other websites provide the detailed characteristics, impacts, intensities and answers to
frequently asked questions about El Niño/La Niña events. So any interested reader can refer
these sources. According to these sites, El Niño represents the warm phase of the El
Niño/Southern Oscillation (ENSO) cycle and La Niña represents the cool phase of the cycle,
and is sometimes referred to as a Pacific cold episode.

The El Niño-Southern Oscillation (ENSO) phenomena have a strong impact on the weather
and climate variability of Ethiopia (Haile, 1988). Farther to the north, Eastern Equatorial
Africa-a region that includes Kenya, Southern Ethiopia, Somalia, Uganda, and Tanzania -
generally experiences more rainfall during El Niño years. There, the deluge associated with
the 1997 El Niño was nearly unprecedented (Ropelewski, 1999). Similarly, Goddard and
Graham (1999) commented that the rainfall variability in Eastern and Southern Africa is the
conjunction of two competing effects of the Pacific and the Indian Oceans. Warming of the
Eastern Tropical Pacific, during an ENSO event, tends to alter the atmospheric circulation
dynamics above Eastern Africa and to reduce rainfall rate on this area. The effects of La Niña
are generally less pronounced in Eastern Equatorial Africa and tend to be the opposite of those
of El Niño (Nicholson and Selato, 2000). The interval between the two strongest El Niño
events occurred only 15 years apart and it should be typically 30 to 40 years and these changes
are unlikely to be due to natural variability alone (Trenberth and Hoar, 1997), and natural
atmospheric cycles such as the Pacific Decadal Oscillation (PDO), the Madden-Julian

37
 Chapter 4: The impact of climate shifts and ENSO phenomena on the hydrological status of Lake Hawassa

Oscillation (MJO) or the chaotic nature of the atmosphere might also have a role to play
(McPhaden, 1999).

The Oceanic Niño Index (ONI) has become the de-facto standard that NOAA uses for
identifying El Niño (warm) and La Niña (cool) events in the tropical Pacific. For the purpose
of reporting, for an event to be categorized as weak, moderate or strong it must have equaled
or exceeded the threshold for at least three months. The threshold is broken down into Weak
(with a 0.5 to 0.9 SST anomaly), Moderate (1.0 to 1.4) and Strong (≥ 1.5) events (Null, 2013).

4.4. Methods

4.4.1. Data availability

As shown in table 4.1, there exists fairly long sequence of hydro-climatic data for Hawassa
meteorological station which is the nearest station for the lake under consideration. Other
meteorological stations in the watershed (refer figure 2.4 in chapter two) have limited data.
Data gaps are filled by linear interpolation throughout the study.

Table 4.1. The core set of hydro-climatic data employed in the study
Data type Temporal scale Period Sources
Lake level records Daily 1970-2010 Ministry of Water Resources
Stream flow » 1980-2006 »
Rainfall for:
 Hawassa Daily 1972-2010 Meteorological Agency
 Wendo Genet Monthly 1974-2010 »
 Shashemene » 1974-2010 »
 Yirbaduwancho » 1974-2010 »
 Haisawita » 1974-2010 »
Pan-evaporation Daily 1986-2007 »
Wind speed » 1989-2010 »
Relative humidity » 1985-2010 »
Temperature » 1973-2010 »
Sun-shine hours » 1985-2010 »

4.4.2. Estimation of coherence between ENSO index and lake level variability
Time series data records of any two continuous variables suitable for computing a covariance,
if of sufficient length for computing a stable fast Fourier transform (fft), can be transformed
into the frequency domain for computation of a dimensionless squared spectral coherence
(Biltoft and Eric, 2009). Transforming from the time to the frequency domain and computing
the squared spectral coherence (CH) provides frequency-stratified results that can be tested for

38
 Chapter 4: The impact of climate shifts and ENSO phenomena on the hydrological status of Lake Hawassa

statistical significance using the F distribution (Biltoft and Eric, 2009). Frequency is defined as
the number of cycles per unit time. Coherence, also known as coherency spectrum, is a widely
used measure for characterizing linear dependence between two time series. Classical books
on time series analysis present coherence as “the frequency domain analogue of the
autocorrelation function” (Hernando and Bellegem, 2006). Further information on spectrum
analysis can be referred from books such as Koopmans (1974) and Bendant and Piersol
(1986).

The presence of trend in a time series data produces a spectral peak at zero frequency, and this
peak can dominate the spectrum in that other important features are obscured (GEOS, 2013).
Due to this, detrending should be part of the analysis. In this study, the time series were
detrended using linear regression (that means: the difference between the expected value
computed from a linear regression through the series and the data point is added to the mean of
the series). The autocorrelations in the time series were also removed by differencing
techniques with order 1.

According to the coherence analysis that is used in this study (Von Storch and Zwiers, 1999;
Jenkins and Watts, 1968; and Bloomfield,1976):

The cross-spectrum (coherence analysis) is defined from the covariance function Cxy:

( ) ∑ * + [ ⁄ ⁄ ]……………………… (4.1)

This is a complex function where the power is:

( ) ( ( )) ( ( )) …………………...…… (4.2)

and the phase is:

(( ( ))
( )= ( ) ………………………….……..…… (4.3)
( ( ))

A cross-spectrum for two similar processes, but with one shifted in time with respect to the
other (x(t) and x(t + )), gives the same power spectrum as for the same analysis applied to
two identical time series, x(t) but instead of a phase difference of zero, the phase is linear in
frequency with a slope proportional to the phase shift: Φxy(x) = 2  .

39
 Chapter 4: The impact of climate shifts and ENSO phenomena on the hydrological status of Lake Hawassa

The coherence spectrum is analogous to the conventional correlation coefficient and is defined
as:

( )
( ) ……………………………..…….…… (4.4)
( ) ( )

Namdar-Ghanbari et al. (2009) employed similar analysis to examine the relationships

between ice, local climate and the teleconnections, Southern Ocean Oscillation (SOI), Pacific
Decadal Oscillation (PDO), North Atlantic Oscillation (NAO), and Northern Pacific Index
(NP).

4.4.3. Significance limits of the spectral coherence estimation

As noted by Thomson and Emery (2001), the final step in any coherence analysis is to specify
the confidence limits (i.e. the level up to which the coherence-square values can occur by
chance) for the coherence-square estimates. This step places the spectral results in a complete
statistical context.

It is noted that each Fourier frequency is associated with only two degrees of freedom
(Thomson and Emery, 2001) regardless of length of the records. In this spectral analysis, there
exist 492 data points (pairs of monthly records) in the time series and the band width is
computed as 1/492=0.002 cycle/month (Band width is the width of the frequency interval
applicable to a spectral estimate (GEOS, 2013)). The 492 observations have 256 points in the
spectrum and each of these 256 (half of the total observation) spectral estimates would have
two degree of freedom. However, results based on two degrees of freedom are not statistically
reliable (Thomson and Emery, 2001) or unlikely to be reproducible (Hartmann, 2013). Hence,
some sort of ensemble averaging or smoothing of spectral estimates is required. As noted by
Engle (1976), the width of the window is an important parameter in the estimation. The wider
the window, the smaller is the variance of the resulting estimate. The wider the window, the
more serious may be the bias of smoothing over non-smooth portions of the spectrum. The
more smoothing we do, the narrower the confidence limits and the greater the reliability of any
observed spectral peaks. The trade-off is loss of spectral resolution and longer processing time.
The windowing approach, which partitions the time series into a series of shorter overlapping
segments, is one of the computation methods used to smooth (average) spectral estimates
(Thomson and Emery, 2001).

40
 Chapter 4: The impact of climate shifts and ENSO phenomena on the hydrological status of Lake Hawassa

According to Engle (1976), the spectral estimator resulting from smoothing the periodogram is
approximately proportional to another chi squared random variable, this time with more
degrees of freedom. The equivalent degrees of freedom (EDF) are equal to:

EDF=B * m …………………………………………..…………. (4.5)

where B is band width and m is the number of observations.

Considering the recommendation of Engle (1976) in that the sensible value of windows span
(the author used the term "range" in the paper) is the square root of the number of
observations. Span of 23 was used in this study. "Range" is defined as the number of spectral
points used in each moving average. It gives the separation between which two points are
known to be completely independent.

As a method to increase the degree of freedom, smoothing of the data series using Daniell´s
window with span of 23 was used and the degree of freedom was increased to 46 (which is
assumed to be against the resolution of the spectrum). "Resolution" is the ability of the
spectrum to represent the fine structure of the frequency properties of the series (GEOS,
2013). The new bandwidth of this spectrum now becomes 46/492=0.093 cycle/month.
Smoothing the spectrum means that we have fewer independent estimates but greater
statistical confidence in the estimate we retain. The number of degrees of freedom for each
spectral estimate is just twice the number of realizations of the spectrum that we average
together (Hartmann, 2013). Biltoft and Eric (2009) conformed that the best solution will likely
include equivalent degrees of freedom in the midrange between 10 and 100. The F test results
with degrees of freedom that fall within the middle of this range produce the most consistent
and reliable results. As presented by Ghanbari et al. (2009), the estimated coherencies are
considered significant at the 99% and 95% level of confidence when they are larger than the
critical value T derived from the upper 1% and 5% points of the F-distribution on (2, d-2)
degrees of freedom:

………………………………………………….……… (4.6)

41
 Chapter 4: The impact of climate shifts and ENSO phenomena on the hydrological status of Lake Hawassa

where d is the degrees of freedom associated with the univariate spectrum estimates.
Coherence peaks indicate frequencies at which the principal flux activity is occurring (Biltoft
and Eric, 2009).

4.4.4. Sequential regime shift detection using Regime Shift Index (RSI)
A jump in a series that is detected by a regime shift test can imply changes in either climatic
factors or watershed characteristics (Tu et al., 2004). According to Breaker (2007), change
points occur where the changes are relatively abrupt. Formally, a change point exists at a time
t0, if all of the observations up to t0 share a common statistical distribution, and those after t0,
share a different statistical distribution.

Rodionov (2004) introduced an algorithm for detecting sequential regime shifts in time series
data in seven steps. Figure 4.1 summarizes the seven steps of Rodionov (2004) in the form of
flow chart.

42
 Chapter 4: The impact of climate shifts and ENSO phenomena on the hydrological status of Lake Hawassa

Set the cut-off length l of the regime to be

determined for a given variable

Determine the difference diff between mean values of

subsequent regimes that would be statistically
significant according to the student´s t-test

Compute the mean value μR1 of the initial l

values as an estimate for regime R1

For each new value starting with i = l + 1, The current regime has not
If the new value < μR1 ± diff.
compute μR2 = μR1 ± diff. changed

If the new value > μR1 ± diff.

Recalculate the average μR1 to
include the new value xi and l-1
previous values of variable X and wait
Possible start point j of the new regime R2. for the next value to come

Use each new value of xi, where i > j, to confirm or

reject the null hypothesis of a regime shift at year j.

The test for a regime shift at year

Compute RSI value Negative value
j failed.

Positive value

Significant regime shift

Calculate the actual mean value of the new

regime μR2 (now, this value becomes base
one against which the test will continue)

Search for the next shift to regime R3 starts

from year i = j + 1

Figure 4.1. The procedure to determine sequential regime shifts N.B: the diagram is drawn
based on the seven procedural steps proposed by Rodionov, 2004

The above procedures are automated and freely downloadable in the form of an Excel Add-In
at [Link]. The latest version (Ver. 3.4) is used in this study which has
additional attribute of considering the presence of auto-correlation in the datasets using the
procedure as shown in figure 4.2 below.

43
 Chapter 4: The impact of climate shifts and ENSO phenomena on the hydrological status of Lake Hawassa

Figure 4.2. The procedure to account for the existence of autocorrelation

Source: Rodionov, 2004: version 3.4

4.4.5. Detection of long-terms trends using Mann-Kendall test

The statistical significance of long-term monotonic trend of Lake Hawassa water level had not
been computed before, at least to our knowledge, and need to be computed. Statistical trend
analysis is a hypothesis testing process. The null hypothesis (H0) is that there is no trend. Each
test has its own parameters for accepting or rejecting H0. Failure to reject H0 does not prove
that there is no a trend, but indicates that the evidence is not sufficient to conclude with a
specified level of confidence that a trend exists (NNSMP, 2011). Trend analysis enables to
detect significant variations over time. It is easily understood and communicated, and readily
accepted due to its widespread use (TSOA, 1995). In this study, the Mann-Kendall (MK)
statistical trend test (Mann, 1945; Kendall, 1975) was employed to investigate trends in time
series data. It is a kind of non-parametric test and compares the relative magnitudes of sample

44
 Chapter 4: The impact of climate shifts and ENSO phenomena on the hydrological status of Lake Hawassa

data rather than the data values themselves (Gilbert, 1987 as cited in Tabari et al., 2011;
Tabari and Marofi, 2011). It allows us to investigate long-term trends of data without assuming
any particular distribution. The other advantage is its low sensitivity to abrupt breaks due to
inhomogeneous time series (Jaagus, 2006 as cited in Tabari et al., 2011; Tabari and Marofi,
2011). In this study, the 5% level of significance was considered.

The test statistic S measures the monotonic dependence of X on t:

S = P –M …………………………………….………… (4.7)

where :

 P = # of (+), the # of times the X‟s increase with t, or the # of Xi < Xj for all ti< tj
(“concordant pairs”).

 M = # of (-), the # of times the X‟s decrease with t, or the number of Xi > Xj for all ti <
tj (“discordant pairs”).

 i = 1, 2, … (n-1); and j = (i+1), …, n.

There are n(n-1)/2 possible comparisons to be made among the n data pairs. If all y values
increased along the x values, S = n (n-1)/2. In this situation, τ= +1, and vice versa. Therefore,
dividing S by n(n-1)/2 will give a -1 < τ < +1.

 …………………………………….…………… (4.8)
( )

The null hypothesis in accordance with this test H0 states that the data (x1,…,xn) is a sample of
n independent and identically distributed random variables. The alternative hypothesis H1 of a
two-sided test is that the distributions of xk and xj are not identical for all k, j ≤ n with k ≠ j.
The test statistic S, which has mean zero and a variance computed by equation 4.11, is
calculated using equation 4.9 and 4.10, and is asymptotically normal:

∑ ∑ ( ) …………………….……….…… (4.9)

( )
( )={ ( ) } …………………..……..…….. (4.10)
( )

45
 Chapter 4: The impact of climate shifts and ENSO phenomena on the hydrological status of Lake Hawassa

[ ( )( ) ∑ ( )( )]
( ) ………………….….......…. (4.11)

where n is the number of data points, m is the number of tied groups (a tied group is a set of
sample data having the same value), and ti is the number of data points in the ith group. In
cases where the sample size n >10, the standard normal variable Z is computed by using
equation 4.12.

√ ( )
……………………………………..…….…… (4.12)

{√ ( ) }

Positive values of Z indicate increasing trends, while negative values of Z show decreasing
trends. When testing either increasing or decreasing monotonic trends at α significance level,
the null hypothesis is rejected for an absolute value of Z greater than Z1−α/2, obtained from the
standard normal cumulative distribution tables (Partal and Kahya, 2006; Modarres and Silva,
2007 as cited in Tabari and Marofi, 2011). This statistical analysis is performed
using xlstat2013 statistical software.

46
 Chapter 4: The impact of climate shifts and ENSO phenomena on the hydrological status of Lake Hawassa

4.5. Results and discussion

4.5.1. Results of coherence analysis

Figure 4.3 shows the result of coherence analysis. The values of coherency (y-axis) versus
frequency (x-axis) between the Niño 3.4 ENSO index and monthly mean lake level changes.

Summary statistics and options:

 Type of smoothing: Daniell´s window (a simple (equal weight) moving average spectral window)
 Window span (width)= 23 months (which must be an odd number)
 Degree of freedom = 2 x 23 = 46
 Preprocessing: Detrending (by linear regression) and removal of autocorrelation (by differencing)
 Total number of paired observations in the series: 492 monthly data (implying 256 in the spectrum)
 Bandwidth=46/492 = 0.093 cycle/month
 F2,44=5.12 (for 99% confidence limit) and F2,44 =3.21 (for 95% confidence limit)
 Critical coherence squared @99% confidence limit =0.196 and @95%= 0.133

Figure 4.3. Coherence between ENSO index and lake level variability in frequency domain

As evidenced by the result of coherence analysis (figure 4.3), the cyclic nature of Lake
Hawassa water level variability has significant linear relationship to the climate variability at
some frequencies. Here appear two significant peaks at 95% confidence limit. Further probe to
the prominent peak reveals that the peak occurred at a frequency between 0.13-0.14
cycle/month or 1.56-1.68 cycle/year. This corresponds to a period of about 7.14-to-7.69
months (=1/0.14 -to- 1/0.13) or a dominant average periodicity (coincident cycle) of about 7.4
months. A relevant finding was reported by Namdar-Ghanbari and Bravo (2008) in which the
levels of Great Lakes and Trans-Niño Index (TNI) show significant coherence in the
frequency range (3-7)-1 cycles/year.

The vital importance of the above analyses is the detection of significant coherence at some
specific frequency ranges and confirmed that significant portion of the lake level variability is
caused by factors operating on a scale larger than processes in the watershed. The upcoming

47
 Chapter 4: The impact of climate shifts and ENSO phenomena on the hydrological status of Lake Hawassa

sections attempt to further analyze the existence of regime shifts in some of hydro-climatic
variables and reconcile the timing and intensities of El Niño/La Niña events with regime shift
points in time domain.

4.5.2. Variability in the lake level

[Link]. Long-term trend (1970-2010)

The visual inspection of figure 4.4 (below) uncovers the underlying variability of the observed
lake level by suggesting that the overall oscillation tends to be chaotic than periodic. The
highest peak was observed in November 1998 (22.54 m) followed by October and December
of the same year (22.49 m each). The lowest level in this year (June) (21.8 m) was greater than
92.5 % of historical records. This particular year was known for its peak records in many parts
of the world. The cases of Lake Abaya (another Rift Valley lake in Ethiopia) (Belete, 2009);
Lake Nasser (Egypt), Lake Chad, Lake Turkana, Lake Tanganyika, Lake Victoria, and Lake
Mwero (Mercier et al., 2002) are among the few examples.

23,00
Monthly maximum lake

22,00
level (m)

21,00

20,00

19,00
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
Figure 4.4. Hydrograph of monthly maximum lake level

Despite the multiple rises and falls, the lake level experienced a significant resultant upward
trend with Mann-Kendall τ values of 0.558, 0.629, and 0.545 (at = 0.05 and p <0.01%) for
monthly maximum, average and minimum values respectively. The ultimate evolution of
increasing trend is not gradual and consistent in direction (monotonic) rather sharp rises and
falls have been frequently appearing and such variations are likely to bias the monotonic trend.
Similar comment was given by Hartmann and Wendler (2005) in that the use of trend analysis
in climate change research depends greatly upon the time period studied, and results can be
biased when an abrupt climate change is observed during the study period.

48
 Chapter 4: The impact of climate shifts and ENSO phenomena on the hydrological status of Lake Hawassa

The long-term annual average increment was also estimated to be 4.9 cm/yr (as computed by
regression equation) which is low as compared to Lake Beseka (another lake located in the
same basin), which has average annual increment of 20 cm (Zemedeagegneh and Egizabher,
2004).

Regarding the connection of ENSO events to the extreme values of observed lake levels, the
1998 record (historical maximum) can easily be justified for its connection to the worst El
Niño event of the twentieth century (Tereshchenko et al., 2002, Magaña et al., 1999, and Strub
and James, 2002) as measured by changes in the Pacific (Marucci, 2002). Globally, this El
Niño year caused loss of approximately 35-45 billion USD (Sponberg, 1999). On the contrary,
the lowest lake level was observed in 1975 which is likely linked to the two consecutive strong
La Niña events of 1973-74 (the strongest in the period 1950-2012) and 1975-76 (appendix 1).

[Link]. Sequential regime shifts in the lake water level

Figure 4.5 (a, b and c) and appendix 2 demonstrate that the observed annual average,
maximum, and minimum lake levels have undergone a couple of regime shifts reflecting the
instability of the hydro-system.

49
 Chapter 4: The impact of climate shifts and ENSO phenomena on the hydrological status of Lake Hawassa

22
a
Annual average water
21,5
21
depth (m) 20,5
20
19,5
19
1971

1990

2009
1970

1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989

1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008

2010
Lake Depth (m) Mean of the regime (m)

23,5
b
Annual maximum lake

22,5
depth (m)

21,5

20,5

19,5

1992
1993
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991

1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
Max lake depth Mean

22,5
Annual minimum lake depth

c
21,5

20,5
(m)

19,5

18,5
1984

2001
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983

1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000

2002
2003
2004
2005
2006
2007
2008
2009
2010
Minumum lake depth(m) Mean

Figure 4.5. Sequential regime-shifts in annual lake levels (a) average, (b) maximum and (c)
minimum

The important aspect of the prevailing regime shifts lies on their occurrence in the year 1976-
78 (figure 4.5: a,b,c) which was known for the climatic regime shift period of the North
Pacific (Miller et al., 1994, and Yletyinen et al., 2012). The year 1977 also experienced the
highest historical recorded annual total rainfall (1226 mm) (appendix 3). The maximum lake
level has undergone a regime shift in 1983 (figure 4.5: b) which is likely associated to the
devastating El Niño of 1983. The other smaller shift in mean value of lake level occurred in
1986 which was likely caused by moderate but prolonged El Niño of 1986-87. Another

50
 Chapter 4: The impact of climate shifts and ENSO phenomena on the hydrological status of Lake Hawassa

connection that is manifested by the overlap of regime shift of Lake Hawassa water level and
the North Pacific climate regime shift was observed in 1989 (figure 4.5: c) and Yletyinen et al.
(2012) reported that in 1989, a new regime shift (in the climate of the Pacific) had also
occurred but the changes were not as remarkable or pervasive as in the 1976-77.

The highest regime shift was observed in 1992 which showed an upward shift in mean value
of the lake level from 20.43 m to 21.2 m (appendix 2), implying a regime shift of 0.77 m. This
regime was extended up to 2002 and known for its frequent El Niño years of 1991-92 (strong),
1994-95 (moderate), and the 1997-98 El Niño (strong). Swanson and Tsonis (2009) also noted
that climate shifts occurred around 2001/2002 too and Lake Hawassa also experienced water
level regime shift in this year. The relatively sustained maximum lake level regime extended
from 1992 up to 1999 (figure 4.5: b) signifies the occurrences of three El Niños (strong,
moderate, strong consecutively) without the occurrence of La Niña in between (appendix 1).

The general upward shifts between 1978 and 1998 are in agreement with the work of Peterson
and Schwing (2003). They identified the PDO index to be negative for most years during
1948-1976 and positive during 1977-1998. In addition, Niebauer (1998) observed that before
the regime shift, the occurrence of El Niño and La Niña conditions was about even. Since the
regime shift, El Niño conditions are about 3 times more prevalent and this further signifies the
effect of climate.

4.5.3. Rainfall variability in the watershed

[Link]. Hawassa meteorological station as representing the over-lake precipitation

Figure 4.6 and appendix 3 show the sequential regime shifts in annual rainfall at Hawassa
meteorological station that represents the over-lake rainfall (refer figure 2.4). According to the
figure, the rainfall time series shows high variability with nine distinct regimes over the study
period. The relatively long and stable regime extended from 1986-to-1994 (upward shift)
followed by regimes of 1999-2004 (downward) and 2005-2010 (upward). The remaining
regimes are short lived and most of the breaking points coincided with the occurrences of
ENSO phenomena (1976, 1983, and 1994). The climate regime shift of North Pacific Ocean
that occurred in 1976/77 seems to manifest itself by causing an upward shift in both years. The
annual total rainfall record of 1977 was the highest of the records (1226 mm). The shift in

51
 Chapter 4: The impact of climate shifts and ENSO phenomena on the hydrological status of Lake Hawassa

1998 was also most likely linked to the transition from strong El Niño (97-98) to the two
consecutive strong La Niñas (98-99 and 99-00).

1300
1100
mm

900
700

1984

1991

1998
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983

1985
1986
1987
1988
1989
1990

1992
1993
1994
1995
1996
1997

1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
Over-lake precipitation (mm) Mean

Figure 4.6. Regime shift of over-lake rainfall

The monotonic trend analysis on this rainfall data shows no significant trend throughout the
recorded time span (both monthly and annual scale). The markedly uniform inter-annual
fluctuations of rainfall in East Africa were also reported by Nicholson (1996). The author also
showed the strong links between rainfall fluctuations and ENSO phenomena.

[Link]. Rainfall at the other stations

As shown in figure 4.7, the years 1986 and 1987 are the common breaking points for the
upward shift for Wendo Genet. The year 1982 (strong El Niño) likely caused the shift in the
rainfall of Wendo Genet and Shashemene. Because of the presences of wide data gaps, the
data from these stations are less reliable and may only serve as a support to the other analyses.
Rainfall data at Yirbaduwancho and Haisawita stations are not included in this analysis due to
similar reason.

52
 Chapter 4: The impact of climate shifts and ENSO phenomena on the hydrological status of Lake Hawassa

1600
1350

mm
1100
850
600

1984

1991

2006
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983

1985
1986
1987
1988
1989
1990

1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005

2007
2008
2009
2010
Wendo Genet Mean

1350
1100
mm

850
600
350
1984

1991

2006
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983

1985
1986
1987
1988
1989
1990

1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005

2007
2008
2009
2010
Shashemene Mean

Figure 4.7. Sequential regime shifts in the annual rainfall on the watershed

4.5.4. Variability in the streamflow of Tikur Wuha River

[Link]. Long-term trend

The trend analysis result of Tikur Wuha streamflow (the only perennial river flowing into the
lake) shows a significant increasing trend (MK =0.66 for the annual average and 0.385, 0.662
and 0.508 for the three local seasons of June-Sep (Kiremt), Oct-Feb (Baga) and March-May
(Belg) respectively. Monthly values also show similar trend (MK =0.440 at p <0.01%).

The increasing trend of the stream flow without the corresponding trend in rainfall indicates
the modification of the hydro-system (Chang, 2007). This argument is discussed in section
4.5.6.

[Link]. Detection of regime shifts

Figure 4.8 and appendix 4 demonstrate the variability of streamflow of Tikur Wuha across
time. The first breaking point occurred at 1986 which is known for its moderate El Niño. As
shown in the previous section, rainfall records at Hawassa and Wendo Genet experienced
similar upward shifts implying that the rainfall was the likely cause of the shift. The years
1994 and 1997 are also another change points corresponding to the timing of ENSO events.

53
 Chapter 4: The impact of climate shifts and ENSO phenomena on the hydrological status of Lake Hawassa

3,5

m3/s 2,5

1,5

0,5

1987

2004
1980
1981
1982
1983
1984
1985
1986

1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003

2005
2006
Mean annual stream flow of Tikur Wuha River Mean

Figure 4.8. Regime shift in Tikur Wuha stream flow

4.5.5. Shift detection in the runoff coefficient of Tikur Wuha sub-watershed

Tikur Wuha River is the only stream that has been gauged in the watershed and the time series
data of this river was used to analyze the regime shifts in runoff coefficient values. The runoff
coefficient is the ratio of total streamflow volume to the total precipitation over a certain area
and time (Kadioglu and Sen, 2001). Four meteorological stations are found in and around
Tikur Wuha sub-watershed as shown in figure 4.9.

Figure 4.9. Tikur Wuha sub-watershed and Thiessen polygon

The percentage contribution of each station to Tikur Wuha sub-watershed is: Hawassa (24%),
Wendo Genet (38%), Haisawita (27%) and Shashemene (11%). As presented in table 4.2, the
annual runoff coefficient (C) for Tikur Wuha River is computed by dividing the weighted

54
 Chapter 4: The impact of climate shifts and ENSO phenomena on the hydrological status of Lake Hawassa

rainfall depth to the resulting streamflow. The long-term average value is computed to be
C=0.14.

Table 4.2. Annual total rainfall and runoff coefficients in the Tikur Wuha sub-watershed
Year Haisawita Wendo Hawassa Shashemene Weighed rainfall* Stream flow C**
Genet (mm) (mm)
1981 1295 1314 1040 953 1110 106 0.10
1982 1403 821 992 1241 1132 108 0.10
1983 491 1158 1160 1108 947 136 0.14
1984 1348 1107 725 674 920 95 0.10
1985 1151 1229 902 892 991 87 0.09
1986 1079 1175 1194 565 1025 121 0.12
1987 1316 1241 955 1208 1123 110 0.10
1988 1035 931 957 887 954 126 0.13
1989 843 1191 1025 1073 987 126 0.13
1990 805 1277 751 962 847 146 0.17
1991 767 1124 889 594 809 131 0.16
1992 1287 1288 975 951 1078 129 0.12
1993 1133 1025 928 992 998 144 0.14
1994 1060 1250 861 745 921 132 0.14
1995 917 1403 1004 853 976 148 0.15
1996 1186 1305 1189 1207 1190 182 0.15
1997 1226 917 1055 1227 1114 143 0.13
1998 1265 943 1146 1175 1156 195 0.17
1999 817 1291 810 822 850 203 0.24
2000 1014 1161 822 986 931 135 0.14
2001 1298 1096 1022 662 1027 157 0.15
2002 1167 1031 920 338 878 173 0.20
2003 1037 1124 821 605 860 - -
2004 1030 1314 896 669 919 185 0.20
2005 1304 821 998 689 999 162 0.16
Average = 0.14
*computed as = value of (Haisawita x 0.285) + (Wendo Genet x 0.42) + (Hawassa x 0.195) + (Shashemene x
0.09)
**computed as the ratio of runoff from Tikur Wuha stream flow to the weighted rainfall for the sub-watershed
Figure 4.10 (below) and appendix 5 depict the sequential regime shift of the runoff coefficient
of Tikur Wuha. As depicted by the figure, the year 1997 (strong El Niño) is the breaking point.
In the same way, the maximum runoff coefficient in 1999 (table 4.2) is likely attributed to the
saturation of the soil as a result of high rainfall during 1997-98.

55
 Chapter 4: The impact of climate shifts and ENSO phenomena on the hydrological status of Lake Hawassa

Run-off coefficient (-)

0,25
0,2
0,15
0,1
0,05

1986

2001
1981
1982
1983
1984
1985

1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000

2002
2003
2004
2005
Runoff coefficient of Tikur Wuha sub-watershed (C) Mean

Figure 4.10. Shift detection in runoff coefficient of Tikur Wuha sub-watershed

4.5.6. Land use/cover changes: a potential anthropogenic factor

Often, the impact of climate change and human activities on hydrological variables cannot be
distinguished (Uhlenbrook, 2009) or is still a challenge in hydrology (Elfert and Bormann,
2010). Climate and land use are key factors controlling the hydrological behavior of a
watershed (Hörmann et al., 2005; Li et al., 2009). Technical details on the impact of land use
on watershed hydrology is given in Maidment (1993) and other hydrology books. In terms of
spatial scale, distinguishing the impact of land use changes on hydrology from the impact of
climatic variability is more difficult at the watershed scale than at the plot scale or small
watershed (Archer, 2003). Many studies have considered these factors separately. However,
these factors do not act in isolation, but rather interact to affect ecosystem structure and
function (Kulakowski et al., 2011). Their influence on the rainfall-runoff relationship is
usually investigated through the analysis of long hydro-meteorological time series or by
hydrological modeling (Tu et al., 2004).

Land use is a key factor controlling the hydrological behavior of watersheds and different
approaches are thinkable to identify possible impacts of land use change on watershed
hydrology. If long-term data series on the hydrological behavior as well as land use and other
influencing factors are available, statistics can reveal the contribution of land use change to
hydrological change in general (Elfert and Bormann, 2010).

A number of studies were conducted in Lake Hawassa watershed in relation to the impact of
land use/cover of the local water cycle. Abrha (2007) attempted to assess its impact on ground
water recharge. Gebreegziagher (2004) considered it as the most likely cause for the increasing
tendency of runoff over time in combination with the effect of climate.

56
 Chapter 4: The impact of climate shifts and ENSO phenomena on the hydrological status of Lake Hawassa

There could be different reasons for land use changes to occur and the social and physical
forces that drive those changes are explained in the DPSIR analysis of chapter eight. Table 4.3
presents some of the land use changes for the years: 1973, 1986, and 2000.

Table 4.3. Land use changes in Lake Hawassa watershed (units are in km2) (Abrha, 2007).
Shrubby
Agriculture Grass land Bush land wood land Urban Area
1973 323.3 15.5 165.9 704.7 6
1986 466.2 59.5 180.3 548.6 8
2000 565.9 68.7 145.6 448.2 13

On average, 9.5 km2 areas of shrub woodlands have been converted into other land uses types
mostly into agricultural lands and instead, 9 km2 new agricultural lands have been introduced.
The general trends in land use/cover changes at country level also show similar tendency. For
instance, forest cover in Ethiopia fell from 16% in the 1950s to 2.7% by the early 1990s, and
continues to decline by nearly 1% per year as woodlands are converted to fuel wood, farmland
and building materials (Shiferaw and Holden, 2001 as cited in Reynolds et al., 2010).

Generally, the long-term increasing trend of streamflow from Tikur Wuha sub-watershed
(figure 4.8) without a corresponding increment in rainfall (figure 4.6 and 4.7) is found to
justify the role of land use/cover changes at least in modifying the impact of climate. Chang
(2007) also argued similarly where such situation indicates the modification of the hydro-
system. This justification, which is based on statistical analysis of long hydro-meteorological
time series, is supported by Tu et al. (2004) and Elfert and Bormann (2010).

4.5.7. Detection of regime shift in lake-evaporation (1986-2007)

The variation in the rate of evaporation from the surface of the lake is considered as one of the
factors that affect the variations in water level. Monitoring of lake-evaporation has never been
done for Lake Hawassa. Due to this situation, indirect methods were employed in our case.
The first option was to use the pan-evaporation time series data and the second was to apply
the Penman-Monteith model (Monteith, 1965; Penman, 1948) (equation 4.15). Figure 4.11
shows the comparison between the estimates of lake-evaporation using both methods. A pan-
coefficient of 0.75 was used as recommended by Ayenew and Gebreegziagher (2006), Legesse
et al. (2003), and Ayenew (2002). The Penman-Monteith model uses five climate variables
(minimum and maximum temperature, relative humidity, wind speed, and sun-shine hours) to
compute the potential evapotranspiration (ETo) (equation 4.13), which is equivalent to

57
 Chapter 4: The impact of climate shifts and ENSO phenomena on the hydrological status of Lake Hawassa

evaporation from the surface of open water. In this study, the annual values of ETo were
computed from monthly values of input parameters using CROPWAT 8.0 software.

( ) ( )
( )
( ⁄ )

ETo = Water volume evapotranspired (mm s−1)

Lv = Volumetric latent heat of vaporization. Energy required per water volume vaporized.
(Lv = 2453 MJ m−3)
Δ = Rate of change of saturation specific humidity with air temperature. (Pa K−1)
Rn = Net irradiance (W m−2), the external source of energy flux
G = Ground heat flux (W m−2)
cp = Specific heat capacity of air (J kg−1 K−1)
ρa = dry air density (kg m−3)
δe = vapor pressure deficit, or specific humidity (Pa)
ga = Conductivity of air, atmospheric conductance (m s−1)
gs = Conductivity of stoma, surface conductance (m s−1)
γ = Psychrometric constant (γ ≈ 66 Pa K−1)

The long-term annual average estimates of lake-evaporation are to the magnitude of 1432 mm
(using the pan method) and 1406 mm (using the Penman-Monteith model) (figure 4.11).
Annual lake-evaporation

2000
1800
1600
(mm)

1400
1200
1000
1997
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996

1998
1999
2000
2001
2002
2003
2004
2005
2006
Pan evaporation Penman-Monteith model

Figure 4.11. Comparison of lake-evaporation estimates of pan vs. Penman-Monteith model

As observed from figure 4.11, pan-evaporation appears to have a striking drop in annual
magnitude between 1995 and 1996 and such drop is unusual and can be suspected of
artificially induced resulting from changes in the position of instrument or recording
technique. As witnessed by officials and experts of the meteorological agency (personal
communication), there were no changes in data recording. Even though a general decline of
pan-evaporation rate has been observed in many part of the world (Peterson et al., 1995;
Chattopadhyay and Hulme, 1997), such drastic drop needs special attention. It is likely that

58
 Chapter 4: The impact of climate shifts and ENSO phenomena on the hydrological status of Lake Hawassa

some noise entered in the measurement of pan-evaporation after the year 1995 and needs to be
further investigated.

In terms of variability across time, figures 4.12 and 4.13 present the sequential regime shifts
between the period 1986 and 2006.

2000
1800
1600
mm

1400
1200
1000
1990

1997

2004
1986
1987
1988
1989

1991
1992
1993
1994
1995
1996

1998
1999
2000
2001
2002
2003

2005
2006
Pan evaporation * 0.75 Mean

Figure 4.12. Regime shifts in lake-evaporation as computed from pan records (Data source:
Meteorological Agency)

1500
1450
mm

1400
1350
1300
1988

2001
1986
1987

1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000

2002
2003
2004
2005
2006
Penman-Monteith model Mean

Figure 4.13. Regime shifts in lake-evaporation as computed by the Penman-Monteith model

As depicted by figures 4.12 and 4.13, the lake-evaporation exhibited significant regime shift at
the year 1995 but this shift is pronounced in the estimates of pan-method. The effect of ENSO
phenomena seems better shown by the Penman-Monteith method in which the year 1999
(strong La Niña) is found to be the change point.

4.6. Conclusions
The diverse statistical analyses of this chapter provide a plausible explanation for the
interaction among the hydrology-climate-human components of the system. The hydrologic
component includes lake level and streamflow; the climate component comprises evaporation,
rainfall, and ENSO events; and the human component refers to the prevailing land use/cover
changes.

59
 Chapter 4: The impact of climate shifts and ENSO phenomena on the hydrological status of Lake Hawassa

More importantly, the evidences helped us to conclude about the effect of ENSO phenomena
and climate shifts on the local climate and hydrology of the lake. Generally, it is observed that
high lake level tends to follow moderate to strong El Niño and the reverse is true for La Niña
events.

The general suggestions of this study supported the idea of Szestzay (1974) in which water
level fluctuations of closed lakes are considered as meaningful indicators of climatic changes
which strengthens the results of Nicholson et al. (2000), Arnell et al. (1996), and Bergonzini
(1998) regarding the sensitivity of numerous lakes of East Africa.

The association of extreme lake level rises of Lake Hawassa to the occurrences of El Niño
events (as in the case of 1998 flood) could have two management dimensions. On one hand, it
would be difficult to mitigate the problem because of its dependence on macro-scale processes
and on the other hand, those large El Niño events which are notorious for their extreme floods
are acceptably predictable within period at lead times of up to two years (Chen et al., 2004).
Climate forecasts are also shown to be more accurate during El Niño and La Niña events and
furthermore, stronger ENSO events lead to greater predictability of the climate (Goddard and
Dilley, 2005). These are opportunities to get alarms against the urgency of flood occurrences
and it is recommended to mainstream the updated information regarding the probable
occurrences of ENSO events and climate shifts in a regular emergency and preparedness
actions to reduce the impact of potential flood risks.

60
 Chapter 5: Simulation of the long-term water level variability of Lake Hawassa

Chapter 5. Simulation of the long-term water level variability of Lake Hawassa in the
presence of regime shifts in water balance components

5.1. Introduction
In chapter four, it was shown that the lake level, rainfall, streamflow, and evaporation
experienced sequential regime shifts during the study period. In this chapter, the effect of
these regime shifts on the fitness of a lake level simulation model is hypothesized. As
reported by Ayenew et al. (2007), Ayenew and Gebreegziabher (2006), and Gebreegziabher
(2004), sharp rises in water level of Lake Hawassa had been occurred that could not be
explained in terms of the water balance components. They explained the existence of
divergence between the observed and simulated lake as it could be the effect of neotectonic
activities, which in turn possibly affect the ground water flow regime. They also suggested
the need for detailed investigation of hydro-climatic variables for better efficiency of the
water balance model. Water balance technique is a means of solving important theoretical
and practical hydrological problems (Chokolov and Chapma, 1974) and the idea of
simulating the lake level variability in this study was conceived after recognizing the
presence of hydro-climatic regime shifts that the simple spreadsheet water balance model
may not account for. In addition, this study extends the previous water balance studies by
about seven years.

5.2. Objectives of the chapter

The aim of this chapter is to simulate the long-term variability of Lake Hawassa water level
using a simple spreadsheet water balance model and to examine how the model behaves
throughout the study period.

5.3. Previous water balance studies of Lake Hawassa

Table 3.3 in chapter three presented the previous results of water balance modeling of Lake
Hawassa. Among these, the work of WWDSE (2001) was one of the earliest available
studies which computed water balance of Lake Hawassa for the period 1970-1998. This
study used the historical records of over-lake rainfall, stream flow, surface runoff (using a
runoff coefficient of 0.13 and 0.19) and evaporation (using pan coefficient of 0.8) together
with the observed lake storage (as computed from change in lake level) to estimate the
ground water flow component as the residual of the balance. The estimated magnitude of

61
 Chapter 5: Simulation of the long-term water level variability of Lake Hawassa

net ground out flow was 71 x106 m3. The works of Ayenew et al. (2007) and Ayenew and
Gebreegziabher (2006) followed similar approach. The assumed pan coefficient in these
studies was 0.75 and a runoff coefficient for the un-gauged part of the watershed was 0.14.
These studies estimated the constant annual ground water outflow of Lake Hawassa as 58
x106 m3. Using the above values together with historical records of hydroclimatic series, it
was shown that the observed and simulated lake level values were acceptably fitted for the
period from 1981 to 1999.

5.4. Methods

5.4.1. Representing the water balance of Lake Hawassa

Water balance for a lake is based on the law of conservation of mass that states any change
in water storage of a given lake during a specified period of time is equal to the difference
between the amount of water added to the lake and the amount of water withdrawn from it
and this balance can be constructed at any level of complexity (Lu et al., 2002). Figure 5.1
represents the water balance of closed lakes (a kind of lake with no surface outflow).

Rain over lake (t)

Evaporation (t)
Runoff (t)
Abstraction (t)


Ground water inflow (t) Storage (t) Ground water outflow (t)

Figure 5.1. Schematic representation of the water balance components for a closed lake

The use of water balance model to investigate the hydrology of a lake is a common
approach. For instance, Acreman et al. (1993) used this approach to explain the declining
level of Lake Toba in Indonesia and Bechtand Harper (2002) to understand the
anthropogenic impact upon the hydrology of Lake Naivasha in Kenya. Kebede et al. (2006)
also employed similar approach to study the hydrological sensitivity of Lake Tana
(Ethiopia) to variations in rainfall. The water balance of Lake Hawassa has been of wide

62
 Chapter 5: Simulation of the long-term water level variability of Lake Hawassa

interest for many years: initially because of scientific curiosity about the causes of the level
rise, but lately for its influence on the infrastructure of the rapidly growing Hawassa town
located in the eastern shore (Ayenew and Gebreegziabher, 2006).
According to the equation of continuity, the change in water level of a lake (H) is
controlled by the difference between the input and output of the water balance and the water
surface area (AL) (Szestzay, 1974) as shown in equation 5.1 below.

Total input – Total output

H = ………………………………..…….……… (5.1)
L

Whereby (AL) may considerably vary with water level H and the aggregate flux terms of the
numerator consist of a number of components largely differing from each other with regard
to the dimension and time pattern of controlling physical processes.

The elevation-area-volume equations for Lake Hawassa were derived by Gebreegziabher

(2004) as shown in equations 5.2 and 5.3. These equations were employed in this study.

Lake Surface area (m2) = 4*106d+ 9*106……………………………….…… (5.2)

Lake volume (m3) = 2*106d2+ 1*107d-5.95*107……………………….…… (5.3)
Where d is the actual depth of the lake in meter

The extended mathematical relationship among the components of water balance in figure
5.1 can be constructed based on conservation of mass as shown in equation 5.4 below for
hydrologically closed/terminal lakes like Hawassa in which the surface outflow component
is omitted because of their terminal nature.

Storage (t) = Rain over lake (t) + Runoff (t) - Evaporation (t)- Abstraction (t) + Gnet ground water flow (t)………(5.4)

5.4.2. Quantification of water balance parameters

[Link]. Over-lake precipitation

As shown in figure 2.4, time series rainfall records of Hawassa meteorological station
represents the direct over-lake rainfall.

63
 Chapter 5: Simulation of the long-term water level variability of Lake Hawassa

[Link]. Stream flow into the lake

Tikur Wuha River is the only flow that has been gauged and time series of this data was
used in this study.

[Link]. Streamflow from the un-gauged part of the watershed

This component of the water balance was computed by adopting the runoff coefficient of the
gauged sub-watershed as shown in table 4.2. The un-gauged portion of the watershed falls
under the span of Hawassa and Yirbaduwancho meteorological stations with proportional
share of 60 and 40% respectively (see figure 2.4). The two closed sub-watersheds: Muleti
and Wendo Kosha with areas of 91.6 km2 and 114 km2 respectively (MoWR, 2010) were
excluded from the computation. Refer figure 2.9 for the locations of these two sub-
watersheds.

[Link]. Evaporation from the lake (Elake)

Evaporation from a water surface is rarely measured directly (Jones, 1992) and the use of
standard pan with pan-coefficient is the most common method (Jensen 2010; Winter, 1981).
Similar attempts were made by Ayenew and Gebreegziagher (2006), Legesse et al. (2003),
and Ayenew (2002) in the Rift Valley of Ethiopia. This technique was adopted in this study
with a pan-coefficient (k) of 0.75 following the recommendation of Ayenew and
Gebreegziagher (2006), Legesse et al. (2003), and Ayenew (2002). As an alternative to the
pan method, the Penman-Monteith model (Monteith, 1965; Penman, 1948) was used as
described in section 4.5.7 of chapter four.

[Link]. Water abstraction from the lake (Elake)

As reported by Nidaw (1995) and Gebreegziagher (2004), there is no apparent evidence of
water abstractions from the lake, but in terms of water that could have entered the lake, if
not abstracted, are estimated to the magnitude of 22.56 x105m3/year which totally accounts
to about 1% of the total mean annual inflow. WWDSE (2001) and Ayenew et al. (2010) also
confirmed the absence of abstraction. In this study, this component is assumed to be nil.

[Link]. Consideration of ground water

The presence of considerable amount of net ground water outflow from the lake has been
reported by WWDSE (2001), Ayenew et al. (2007) and Ayenew and Gebreegziagher (2006).

64
 Chapter 5: Simulation of the long-term water level variability of Lake Hawassa

In terms of magnitude, WWDSE (2001) estimated the annual value to be 71 x106 m3, and
Ayenew et al. (2007) and Ayenew and Gebreegziagher (2006) estimated it as 58 x 106m3.
The average of the two estimates was used in this study.

5.4.3. Simulation procedure

The first step in simulating the lake level variability is the construction of the lake water
balance. In this study, the long-term (1986-2006) water balance of Lake Hawassa is
constructed on monthly basis according to equation 5.4. Figure 5.2 represents the simulation
procedure.

Monthly rainfall on Monthly flow Monthly rainfall on Constant ground

Monthly over-lake Monthly pan-
Tikur Wuha sub- (volume) of Tikur the un-gaged water outflow
rainfall (depth) evaporation depth
watershed Wuha River subwatershed (volume)

Run-off coefficient (C) Monthy run-off (Ungaged)

Observed lake depth

Depth-area-volume
at the beginning and
equation
end of each month

Computation of
lake volume at
the end of each
month

Computation of
lake depth at
the end of each
month

Comparison
of observed and
simulated lake
level series

Figure 5.2. Flow chart of the lake level simulation procedure

As represented in figure 5.2, the raw data of the water balance equation were sequentially
arranged on monthly basis in Microsoft Excel columns. Monthly rainfall and pan-
evaporation data series are available in depth terms (mm) and stream flow records from
Tikur Wuha River is available in volume term (m3). Time series of streamflow from the un-
gauged sub-watershed were generated by multiplying the runoff coefficients as computed
for Tikur Wuha River (table 4.2).

65
 Chapter 5: Simulation of the long-term water level variability of Lake Hawassa

The observed water depths at the beginning of each month serve as input for equations 5.2
and 5.3 to calculate the corresponding lake surface area and stored volume. The volume of
stored water at the end of a given month was calculated by adding the volume of on-lake
rainfall, streamflow from the watershed and subtracting volume of evaporation and net
ground water outflow from the amount of water computed at the beginning of the month.
Once the volume of stored water at the end of each month was known, the corresponding
lake depths were calculated by solving the positive roots of a quadratic equation as derived
from equation 5.3. The simulation process started on January 1986 and the corresponding
lake depth was 19.93m and its surface area of 88.72 km2, with a storage volume of
934,209,800 m3.

After the initial surface area and volume were set, the simulation process can continue in
two ways. The first approach is the simulation "with updating" as equation 5.7 and the
second is "without updating" as equation 5.8, normally known as "simulation mode" in
scientific literature. Model runs in updating mode normally can issue a reliable estimate at
one time step ahead of the current time step, but the ultimate performance of a model
depends on the model to give good estimates in the simulation mode. The simulation mode
involves the use of previously estimated values as the input function in the model in order to
issue a forecast (Kumambala and Ervine, 2010; Kachroo, 1992).

( ) ( )  ( ) ……………..………………………...… (5.7)

( ) ( )  ( ) ……………..……………….……….…. (5.8)

Where ( ) estimated lake level and ( ) is observed lake levelin a given period of
time. In this research, the second approach (equation 5.8) was employed.

5.4.4. Updating procedure using Autoregression (AR)

Since the simulation of the lake level is going to run in "simulation-mode", it is sensible to
expect the propagation of errors in the outputs of the model because of the reliance of a
given output on the accuracy of other preceding outputs and such dependency is termed as
autocorrelation. The autoregression (AR) model is one of the most favored updating
procedures, which is extensively used in applied hydrology (Serban and Askew, 1991). This

66
 Chapter 5: Simulation of the long-term water level variability of Lake Hawassa

procedure is based on the modification or partial correction of the un-updated output

variables of the model using their error estimates (Shamsedin and Connor, 1999).

According to Box and Jenkins (1976) and Kachroo (1992), an AR model of order p with
mean of zero or an error time series can be defined as:

∑ ………………………………………….…. (5.9)

where is the error at the ith time period and Φ is the AR model parameter set and (ideally
at least) is a pure white noise sequence having mean zero and constant variance . Once
the error series is determined, an AR model is separately calibrated on this series and
subsequently used for forecasting the output errors. In principle, the success of this
procedure depends on the degree of error persistence (Serban andAskew, 1991; Shamsedin
and Connor, 1999).

5.4.5. Model efficiency test using Nash-Sutcliffe Efficiency (NSE)

The Nash-Sutcliffe efficiency (NSE) (Nash and Sutcliffe, 1970) is a well-known normalized
statistic that indicates how well the plot of observed versus simulated data fit the 1:1 line. It
is computed as:

∑ ( )
∑
………………………………….…. (5.10)
( )

where Yi obs is the ith observed value in the series and Yisim is the corresponding simulated
value. Ymean is mean of the observed, and n is the total number of observations. NSE values >
0.5 are generally viewed as acceptable levels of performance, whereas values <0.0 indicates
unacceptable performance.

67
 Chapter 5: Simulation of the long-term water level variability of Lake Hawassa

5.5. Results and discussion

5.5.1. Results of water level simulation

As shown in figure 5.5 the model was unable to capture the full range of lake level
variability. The simulation up until 1996 was relatively acceptable but constant drift
occurred afterwards. Similar divergence was reported by Belete (2009) while simulating the
water level of Lake Abaya, another lake in Rift Valley Basin. Considering that the
simulation before 1996 is acceptable, the error analysis was done for the remaining years
after 1996 as presented in the upcoming section.

5.5.2. Result of residual error analysis

Figure 5.3 shows the procedure for error analysis. The persistent systematic error with
continuous over-estimation (figure 5.5) is considered to be attributed to the accumulation of
imbalances in the month-to-month error in model output. Such type of simulation is named
as "simulation-mode" (without updating) in literatures and the model error is usually
accumulated due to its auto-correlated elements (Kachroo, 1992; Kumambala and Ervine,
2010) and this may be particularly difficult to see unless analysis of the residual errors is
done (Kumambala and Ervine, 2010). The errors between the simulated and the observed
values can be compensated for through the use of implicit or explicit error updating
procedures (Shamsedin and O´Connor, 1999). The predicted lake level at a given time step
is a function of errors induced in the previously predicted values and its own. Such situation
can be analyzed by studying the trend of the error propagation and fitting to error models
like autoregressive (AR), so that persistent errors can be analyzed separately as shown in the
following section. Similar approach was employed by Kumambala and Ervine (2010).
Simulated lake
level series

Simulation Autoregression Updated lake

error updating model level series

Obsereved lake
level series

Figure 5.3. Schematic diagram of the linear Auto-Regressive (AR) updating model

Time series of model error was constructed by subtracting the observed lake level values
from estimated values. From the result of residual analysis (figure 5.4), a linear relationship

68
 Chapter 5: Simulation of the long-term water level variability of Lake Hawassa

was observed between series of model error and time steps t (chronological sequence of
months) with r2 = 0.97 (equation 5.11).

Error = 0.042 t (r2 = 0.97)…………………………………….….…. (5.11)

where t is the chronological sequence of months with t=1 for the first month (January 1996)
and t=132 for the last month in the sequence (December 2006).The model errors (residuals)
are assumed to be independent (uncorrelated) and normally distributed noise with mean
equal to zero and constant variance (Ajami et al., 2006).
6,00
Residual model error

Error = 0.042 t - 0.171

5,00
R² = 0.97
4,00
3,00
(m)

2,00
1,00
0,00
1 13 25 37 49 61 73 85 97 109 121
t = chronological sequence of months

Figure 5.4. Trend of residual error (predicted - observed lake level values)
(Jan 1996-Dec 2006)

As shown in figure 5.5, the updating procedure improves the performance of the model with
a Nash-Sutcliffe efficiency of 0.73.
28,00
26,00
Lake depth (m)

24,00
22,00
20,00
18,00
Jan 91
Jan 86
Jan 87
Jan 88
Jan 89
Jan 90

Jan 92
Jan 93
Jan 94
Jan 95
Jan 96
Jan 97
Jan 98
Jan 99
Jan 00
Jan 01
Jan 02
Jan 03
Jan 04
Jan 05
Jan 06

Observed lake level Modelled (before data assimilation)

Modelled(after data assimilation)

Figure 5.5. Simulated vs. observed lake level

69
 Chapter 5: Simulation of the long-term water level variability of Lake Hawassa

5.5.3. Which components of the water balance show particularity at the year 1996?
The divergence of simulated and observed lake levels (figure 5.5) is assumed to result from
the corresponding uncertainties in one or more input parameters of the model. Figure 5.6
plots volumetric magnitudes of the five components of the water balance against time.

180
160
Annual magnitude Mm3

140
120
100
80
60
40
20
0
-20
-40
-60
-80
1993

2004
1986
1987
1988
1989
1990
1991
1992

1994
1995
1996
1997
1998
1999
2000
2001
2002
2003

2005
2006
Over-lake rainfall Lake evaporation Streamflow (gaged)
Streamflow (ungaged) Storage changes

Figure 5.6. Annual magnitudes of water balance components

The visual interpretation of figure 5.6 reveals that the year 1996 was so particular in that
most of the water balance parameters experienced abrupt changes. The drastic drop in
annual evaporation and the sharp rise in storage changes between 1995 and 1996 are the
prominent particularities. The magnitude of annual rainfall also exhibited some increment
during this period with a corresponding increment in runoff. Another variable that worth
mentioning is the case of ground water flow in that the occurrence of particular imbalance
on the ground water flow was reported by Yirgu et al. (1997), Ayenew and Gebreegziabher
(2006), Gebreegziabher (2004), and WWDSE (2001). The analysis of ground water flow is
not included in this study because of the absence of monitoring data.

5.5.4. Replacing the pan-evaporation by the Penman-Monteith model

As shown in figure 4.11 (chapter four) and figure 5.6, evaporation shows a drastic drop in
annual magnitude and such occurrence induces suspect on the records of pan-evaporation to
be erroneous at some point. To account for such uncertainty, it was attempted to replace the
monthly values of pan based lake-evaporation estimates by an alternative time series based
on the Penman-Monteith model (Monteith, 1965; Penman, 1948) (which is based on
minimum and maximum temperature, relative humidity, wind speed, and sunshine hours to

70
 Chapter 5: Simulation of the long-term water level variability of Lake Hawassa

calculate evaporation from open water surface) as described in section 4.5.7 of the previous
chapter. The result of the new simulation is presented in figure 5.7 below.

25,0
24,0
Lake depth (m)

23,0
22,0
21,0
20,0
19,0
18,0

Jan 06
Jan 86
Jan 87
Jan 88
Jan 89
Jan 90
Jan 91
Jan 92
Jan 93
Jan 94
Jan 95
Jan 96
Jan 97
Jan 98
Jan 99
Jan 00
Jan 01
Jan 02
Jan 03
Jan 04
Jan 05
Observed lake level Simulated lake level

Figure 5.7. Simulation result after replacing pan-evaporation by Penman-Monteith model

As shown in figure 5.7, the simulation of the lake level is improved in the case of Penman-
Monteith model. However, the divergence is still in existence except that it begins around
the year of 1998. The situation partially shows the likely erroneous data of the pan-
evaporation and such uncertainties shall be investigated in the future.

5.5.5. Annual magnitudes of water balance components

Annual magnitudes of the water balance components were computed as the sum of monthly
values following equation 5.4. Table 5.1 shows the time series of annual magnitudes of Lake
Hawassa water balance components.

71
 Chapter 5: Simulation of the long-term water level variability of Lake Hawassa

Table 5.1. Annual volume of some components of water balance of Lake Hawassa (Mm3)
Over-lake Lake Stream flow Stream flow Storage
Year rainfall evaporation (gauged) (un-gauged) changes
1986 105 147 76 68 24
1987 86 150 69 53 15
1988 86 147 79 55 43
1989 94 141 79 63 -3
1990 69 145 91 66 -23
1991 80 140 82 69 -41
1992 87 142 81 64 53
1993 85 150 90 76 17
1994 79 164 82 72 -7
1995 91 146 92 73 -27
1996 109 116 113 94 98
1997 99 125 90 66 29
1998 111 115 122 103 49
1999 78 132 127 105 -50
2000 77 126 84 65 -54
2001 96 116 98 86 38
2002 87 121 108 56 -49
2003 76 121 100 89 -55
2004 82 113 115 99 -5
2005 91 116 101 90 9
2006 110 97 95 98
Mean 89 132 94 77 3
N.B
 Negative values show the removal of water from the lake while the positive values show the
addition of water.
 The gauged sub-watershed accounts for 625 km2, whereas the un-gauged sub-watershed =
512.66 km2. Area of the un-gauged sub-watershed was computed by subtracting the area of
the lake (93.24 km2), and the two closed sub-watersheds: Wendo Kosha = 114 km2; and
Muleti = 91.6 km2 from the total watershed area = 1436.5 km2

5.6. Conclusions
This chapter is not the first attempt to simulate the lake level variability of Lake Hawassa;
rather it is a sort of updating the existing water balance estimations. It also shows the
potential future inconveniences that would occur due to uncertainties in some of water
balance components.

Regarding the estimated magnitudes of long-term annual water balance components, the
over-lake precipitation (89 Mm3), evaporation (132 Mm3), and gauged stream flow from the
gauged sub-watershed (94 Mm3) are about the same magnitudes as compared to the previous
results (table 3.3 of chapter three). Streamflow from un-gauged sub-watershed (77 Mm3) is

72
 Chapter 5: Simulation of the long-term water level variability of Lake Hawassa

higher than the estimates of Shewangizaw (2010); and lower than the estimate of WRDB
(2007) and WWDSE (2001). Such disagreement seems to be associated with the
consideration of the two closed sub-watersheds: Wendo Kosha and Muleti. Gebremichael
(2007) computed magnitude of the combined (gauged and un-gauged) streamflow which is a
little lower than our estimate. The magnitude of storage changes (3 Mm3) could not be
compared to the previous results because of the absence of comparable estimates.

73
 Chapter 6: The effect of recent sedimentation on the storage capacity of Lake Hawassa

Chapter 6. The effect of recent sedimentation on the storage capacity of Lake Hawassa

6.1. Introduction
Most of the lowlands in the Central Rift Valley of Ethiopia are arid or semiarid and in
degradation, with frequent occurrence of droughts. Soil erosion by water during the rainy
season is a serious problem in the region, leading to declining agricultural production,
decreased food security, and a sedimentation risk for water bodies (Meshesha et al., 2012).
This situation has been accelerated by many human activities such as clearing forests and
woodlands, complete removal of crop residues, and overgrazing, exacerbated by poor soil
management and land use practices (Bekele, 2003).

Soil erosion causes accumulation of sediment in lakes and reservoirs, which results in the
degradation and impairment of use of these water bodies (Fitzpatrick et al., 1987).
Sedimentation has undesirable impacts on water quality, storage capacity, recreational value,
and natural lake bed habitat of natural and/or artificial lakes. For instance, John Redmond
Reservoir in US lost 37% of its storage capacity in 50 years (Martinko et al., 2011). LIA
(2010) also affirmed the impact of sedimentation in causing reductions in water depths,
smothering animals living in the bed, reducing light penetration (water clarity), and hence the
water depth in which aquatic plants with roots in the bed can grow. Drying-up of some lakes
in Ethiopia, such as, Lake Haromaya in the Eastern part (Alemayehu et al., 2007) and Lake
Cheleleka in the upstream of Lake Hawassa (see figures 6.5, 6.6 and 6.7) are among the live
examples for the impact of sedimentation on the reduction of storage capacities and then
disappearance of the entire lakes.

Recent siltation of Lake Hawassa has been perceived as one of the environmental dangers
threatening the lake that can lead to changes in its morphology, which may decrease the water
storage capacity that in turn contribute to the rise of the water level. There have been many
speculations on the impact of sedimentation process on the storage capacity and lake level rise
such as Esayas (2010), Gebreegziabher (2004), and Geremew (2000).

The study of sedimentation records has been widely used in palaeolimnological

reconstructions to evidence long-term trends of the climate change during the geological years
(Tiercelin et al., 1988) and these techniques often used only little number of cores (point data)
from the deepest part of a lake (Terasmaa, 2011). Such kind of study was conducted for Lake

74
 Chapter 6: The effect of recent sedimentation on the storage capacity of Lake Hawassa

Hawassa by Lamb et al. (2002). According to them, sediment accumulation rate in the lake
was estimated to vary between 1.2 and 2.0 mm/yr (mean rate 1.7 mm/yr) but this magnitude is
valid in time scale of centuries and less informative in explaining the contemporary situation.
So, it is timely task to study the recent in-lake sedimentation.

The most conventional technique and accurate determination of sediment load being carried to
a lake by streams is to measure the flow rate and sediment concentration of the inflowing
waters just upstream of the lake. The other conventional methods involve periodic bathymetric
surveys of the lake (limgis, 2001). Temporal comparison of bathymetry maps is an indicator
for environmental changes like lake or reservoir sedimentation. From this information, lake
ecosystem functioning, life time of reservoirs or erosion-sedimentation rates of watersheds can
be derived (Dost and Mannaerts, 2006).

The use of maps to study sediment accumulation in reservoirs is a common practice in many
parts of the world. For instance, prior to the mid-80‟s, the Natural Resource Conservation
Service (NRCS), formerly known as the U.S. Soil Conservation Service (US SCS),
hydrographically surveyed Triadelphia and Rocky Gorge Reservoirs approximately once every
10 years to determine the amount and rate of sediment accumulation. The approach in this
case is to calculate the temporal differences of reservoir capacities between long-term
consecutive mapping periods. For this, the range method, which utilized a number of transects
to determine the cross-sectional area of the reservoir at different locations and reservoir
volumes are calculated and from that, the deposited sediment volumes are deduced (Ortt et al.,
2008). Repeated bathymetric surveys can provide significant insight into the nature of
sedimentation within a reservoir. Changes in reservoir bottom topography can be monitored
over time to provide an overall estimate of the sediment accumulation rate and spatially
explicit representation of sediment accumulation (DeNoyelles and Jakubauskas, 2008).

Despite the importance, the in-lake sedimentation of Lake Hawassa has never been monitored
because of some perceived reasons such as cost, little awareness about the degree of the
problem, and probably due to the absence of an explicitly responsible organization for such
activities.

75
 Chapter 6: The effect of recent sedimentation on the storage capacity of Lake Hawassa

In response to the above situations, the assessment of in-lake sedimentation helps to recognize
its effect on the storage capacity of Lake Hawassa. At this moment, the bathymetric approach
appears to be an attractive and judicious choice because of the presence of previous
bathymetry maps to compare with the new one.

6.2. Hypothesis and objectives of the chapter

In accordance with the hypothesis on the effect of recent sedimentation on the storage capacity
of Lake Hawassa, the objectives of this chapter are:

 To produce a new bathymetry map (of 2010) from a new hydrographic survey and to
compare it with the old map (of 1999);
 To demonstrate the application of comparative analysis of the two bathymetric maps in
estimating the amount of sediment accumulation; and
 To compare the sediment volume with the storage capacity of the lake.

76
 Chapter 6: The effect of recent sedimentation on the storage capacity of Lake Hawassa

6.3. Materials and methods

6.3.1. The old bathymetric map (1999)

The first intensive bathymetry survey was done by Water Works Design and Supervision
Enterprise in 1999 using Bathy 1500 echo sounder (Version P02585). Echosounder is a device
for measuring depth of water by sending pressure waves down from the surface and recording
the time until the echo returns from the bottom (the free dictionary: accessed in 2012). The
technical specifications are shown in table 6.1. The output of this survey was obtained from
the archive of respective office in the form of hard copies as four A1 sized original blue prints
with a scale of 1:10000. To keep the originality of the information, these A1 sized papers were
scanned with a 300 dpi resolution scanner in University of Bonn (Germany) at the Department
of Geography and the scanned images were georeferenced (figure 6.2: a) to fit the lake using
the standard georeferencing technique available in ArcGIS10.

6.3.2. Tying local bench marks with standard elevations

Since the start of monitoring the lake level by Ministry of Water Resources, the data has been
recorded relative to the local bench marks. The elevation of the datum for water level
monitoring station of Lake Hawassa which is coded as 082004 and operated by MoWR was
set to be 1678.17 m.a.s.l. (WWDSE, 2001). The new survey also relied on this setup for the
sake of future comparison.

6.3.3. The new bathymetric survey (2010)

Before the commencement of the bathymetry survey, the SPOT5 satellite image of the lake
was gridded with 500 x 500m spacing and 337 points (figure 6.2: c) were generated to cover
the entire surface area of the lake. The contour lines of the old bathymetry map were
interpolated and converted into raster surface using ArcGIS 10 software and then with the use
of "extract values to points" option, the old elevation of each gridded locations were extracted
from the old bathymetry map as a kind of re-sampling.

The echosounder generates acoustic pulses for bottom recognition (water/sediment interface).
The data were collected as discrete x, y, z points where x and y are geographic coordinates of
the discrete points and z represents the depth of the lake bottom as measured from the surface
of the lake. A sonarmite echosounder, with technical specifications as shown in table 6.1, was

77
 Chapter 6: The effect of recent sedimentation on the storage capacity of Lake Hawassa

used, and a stadia rod was used to measure water depth shallower than 0.3m where large
grasses dominate. The basic components are: a recorder, the transmitting and receiving
transducer and a power supply. A total of 32 to-and-fro trips were made to cover the entire
lake surface during the survey which was made between December 2010 and January 2011.

Table 6.1. Main technical parameters of the two echosounders

Old (1999) New (2011)
Brand name Bathy 1500 echosounder (P02585) Sonarmite echosounder
Transducer frequency 33/200 kHz (Dual) 235 kHz
Minimum depth 0.5 m (transducer dependent) 0.3 m
Maximum depth 1000 m 75 m (software limit)
Beam pattern 8 to 24 Degrees 8 to 10 degrees
Accuracy + / - 0.025 m (for 0-40m depth) + / - 0.025 m

(a) Bathy 1500 echosounder (b) Sonarmite echosounder

Figure 6.1. Partial view of echosounders for the old and new bathymetry surveys

6.3.4. Assessment of sediment by topographic differencing technique

The bathymetric approach in this study was based on a direct comparison of lake bottom
elevation at two different time periods (topographic differencing technique), first in January
1999 (referred as old map) and second in December 2010 and January 2011 (referred as new
map) to detect the changes. The approach was based on the procedure used by Ortt et al.
(2008) that consisted of an assessment phase and of a historical comparison phase. The
distribution of sediment accumulation thicknesses and volume in the lake was determined by
comparing the new and old bathymetry maps. The sediment thickness map was generated by
subtracting the new bottom elevation from the old values.

78
 Chapter 6: The effect of recent sedimentation on the storage capacity of Lake Hawassa

(a) Original blueprint (1999) (b) Interpolated raster (1999) (c) New measurement points (2010)

Figure 6.2. Basic steps in the topographic differencing between the two bathymetry maps

79
 Chapter 6: The effect of recent sedimentation on the storage capacity of Lake Hawassa

6.4. Results and discussion

6.4.1. Estimation of sediment thickness and distribution

The result of topographic differencing between the two bathymetry maps is presented in figure
6.3. It is found that the average sediment thickness between 1999 and 2010 was to the
magnitude of 14 ± 5cm. The ± 5cm allowance is given to the estimation due to the accuracy
level of the echosounders (refer table 6.1). If a constant annual rate is assumed in the period,
the sedimentation rate would be 1.2 cm/year. Computation of sediment volume is discussed in
the next section. At the moment, we cannot compare this value with other results due to the
absence of recent sediment studies in the region. In terms of geological scale, which is less
informative regarding the recent situation, Lamb et al. (2002) estimated the average
sedimentation rate in Lake Hawassa to the magnitude of 0.17 cm/year placing the basal date at
14
6400 Cyr. Generally, as compared to the recently expanding gullies and the presence of
continuous land use changes as reported by Abraha (2007); Shewangizaw and Michael (2010);
Dessie and Christiansson (2008); WWDSE (2001); and Wagesho et al. (2012), the magnitude
of sedimentation rate seems to be underestimated or below the expectation of many
stakeholders. The distribution of sediment thicknesses over the entire lake bed is shown in
figure 6.3 below.

Figure 6.3. Spatial distribution of sediment thicknesses during the period 1999-2010

80
 Chapter 6: The effect of recent sedimentation on the storage capacity of Lake Hawassa

According to figure 6.3, sediment accumulation thicknesses range from 3 cm to 73cm. It is

higher at the western end of the lake and becomes lower while approaching the eastern end
where the city of Hawassa is located. The maximum deposition occurred at the western part of
the lake. Some of the northern parts and the entry of Tikur Wuha River also show
comparatively higher sediment accumulation than the rest of the areas. The linkages of such
sedimentation patterns to the catchment characteristics are discussed in section 6.6. Generally,
sedimentation rate is consistent with the degradation and pumice nature of the western sub-
watershed.

6.4.2. Sediment volume and its impact on the storage capacity

In terms of volume, the total accumulated sediment between 1999 and 2010 is estimated to be
in the magnitude of 0.0133 km3or 13.3x106m3 (taking the surface area of 94.85 km2 as
computed by ArcGIS based on depth measurement points) which is about 1.2% of the total
volume in twelve years. Assuming a constant rate over the period, the annual sedimentation
rate becomes 0.0011 km3 or 1.1x106 m3. As per this rate, the annual reduction in storage
capacity due to siltation is about 0.08%, which is a little higher than High Aswan Dam in
Egypt (0.05% between 1967 and 1991); and lower than Imagi reservoir in Kenya (0.8%);
Sennar (Makwar reservoir) in Sudan (0.6% between 1925 and 1986); and estimated global
average rate for annual loss of reservoir capacity of 1% (Douglas et al., 2001). The specific
sediment yield, which is the total annual sediment volume divided by the sediment
contributing area of the catchment, is estimated to be 967 m3/km2/year. One could imagine
that the effect of sedimentation is minimal in Lake Hawassa by looking at its physical
magnitude, but the reality is that this sediment carries chemicals from the watershed which
potentially affect the function of the watershed. So, this result should not be considered as
discounting the devastating effect of the sedimentation process on the ecosystem.

6.5. Methodological limitations/ shortcoming /challenges and solutions

The practical implication of our results is a more optimistic view towards the possibility of
estimating sediment depth and volume based on two bathymetric surveys. The success of such
estimation will be more strengthened after consecutive similar survey. Meanwhile, the
methodological problems encountered and their solutions while undertaking this study were
presented in the following paragraphs.

81
 Chapter 6: The effect of recent sedimentation on the storage capacity of Lake Hawassa

One of the methodological problems in this study was the difference in the formats of the two
maps which did not integrate directly. The old map was line data whereas the new map was
point data. To bring the two maps into a common format, the old line data was converted into
continuous surface in raster format using Kriging technique of ArcGIS 10. The Kriging
procedure generated an estimated surface from the scattered set of contour lines with z-values.
Hence, the "extract values to points" option of the spatial analyst tool was used to extract the
values of old bed elevation for each x,y coordinates of the new survey. By doing so, the x,y,z
values of the old bottom elevations were determined before the new survey and these values
were fed to the GPS to navigate into the point of interest. This technique is also assumed to be
more convenient for future comparison of changes in lake morphology.

The presence of large number of outliers (31 out of 337 pairs of values = 9 %) was another
problem in this approach. These outliers showed unrealistic elevation differences in that the
former elevation is greater than the new ones, which is unlikely and replaced by average
values of the eight points in the neighboring grids. The likely sources of the outliers might be
the interpolation technique and/or the accuracy of depth measurements. More accurate result
would come up if the two maps were generated by the same instrumentation and intensity.

In addition, most of the surveys were conducted between 11:00 a.m. - 4:00 p.m. because there
happened less wave occurrences during these hours. But, in few cases, there have been waves
with a potential to change the depth measurements and possibility to tilt the transducer away
from its vertical position. As a solution, repeated measurements were taken together with the
consideration of the wave heights. To avoid the potential effect of depth measurements at the
border of the lake where in-lake vegetation grow, a manually operated stadia rod was used to
directly measure the water depths.

6.6. The linkage of in-lake sedimentation to the sources in the watershed

The sedimentation process is dependent on a multitude of biophysical and anthropogenic
factors, such as the size of the lake, the size of watershed, soil type, climate, land cover, and
land use (Dost and Mannaerts, 2006). Such linkage is considered in chapter seven which
integrates nine factors affecting sediment production and transport processes. Those
characteristics of the watershed which are not included in chapter seven are discussed in the
upcoming sections.

82
 Chapter 6: The effect of recent sedimentation on the storage capacity of Lake Hawassa

6.6.1. Linkage to gully density in the watershed

Gully erosion is the dominant type of erosion in the watershed and it is considered as a
significant process for delivering sediment to the lake. Due to this, it was attempted to
compare the extent and pattern of gully erosion across the watershed with the pattern of
sediment accumulation in the lake. To this end, the spatial distribution of gully density is
assessed based on the data provided by MoWR (2010). Accordingly, the watershed contains
750 segments of gullies with a total linear length of about 668 km which are concentrated on
the Western side of the lake (figure 6.4). As shown in the figure, the western part of the
watershed is highly dissected by gully networks and it is in accordance with the result of in-
lake sedimentation figure 6.3.

Figure 6.4. Map of gully density and pictures of active gullies in the watershed Sources of the
raw data: MoWR (2010)

6.6.2. Linkage to the disappearance of Lake Cheleleka

In the nineteenth century, Lake Hawassa and Cheleleka had been a single lake (Grove et al.,
1975) and Lake Cheleleka was serving as a natural regulator of flow, sedimentation, and
biogeochemistry for Lake Hawassa. The progressive silting up of Lake Cheleleka over the last
35 years (figures 6.5; 6.6 and 6.7) is an example to the degree of sedimentation problem in the
watershed.

Unfortunately, the rate of siltation could not be estimated due to the absence of data but the
rate of shrinkage in lake surface area was monitored by a series of images. Topographic map

83
 Chapter 6: The effect of recent sedimentation on the storage capacity of Lake Hawassa

with a scale of 1:50,000 was used to delineate the surface area of Lake Cheleleka in the year of
1972; satellite imageries of thematic mapper (TM) was used for 1986 and 1995; ETM for 2000
and Spot5 for 2007.

1972 1986

1995 2000

2007

Figure 6.5. The disappearance of Lake Cheleleka in the watershed Source: Own study

84
 Chapter 6: The effect of recent sedimentation on the storage capacity of Lake Hawassa

14

Lake Cheleleka surface

12
10

area (km2)
8
6
4
2
0

1978

1991
1972
1973
1974
1975
1976
1977

1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990

1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
Figure 6.6. Time series of changes in the surface area of Lake Cheleleka

Figure 6.7. Current view of the former Lake Cheleleka (2011)

In 1972, the surface area of Lake Cheleleka was 12 km2. In 1986, it shrank to 5 km2. The
1995 image shows that the surface was reduced to 3 km2, and the 2000 image evidenced the
shrinkage of the surface area down to1 km2. The disappearance of Lake Cheleleka is shown in
the 2007 image. Currently, the area is serving as a grazing land as shown in figure 6.7. Similar
silting up of a large lake, Lake Haromaya in the Eastern part of Ethiopia, was reported by
Alemayehu et al. (2007) as shown in figure 6.8 below.

85
 Chapter 6: The effect of recent sedimentation on the storage capacity of Lake Hawassa

Figure 6.8. Dropping down of Lake Haromaya till drying up Source: Alemayehu et al., 2007

Generally, the susceptibility of Lake Hawassa for siltation seems high as observed from the
diminishing rate of Lake Cheleleka. Lake Cheleleka was serving as a natural silt trap, but
currently its damming effect is reduced exposing Lake Hawassa to an increased siltation.

6.7. Conclusions
This study is a first attempt to quantify the amount of contemporary sediment, to our
knowledge, in Lake Hawassa and found to provide important information about the effect of
sediment on the storage capacity of the lake. The case of sedimentation in a lake is different
from the water balance components in affecting the lake level changes. That is because once
this sediment joins the lake, it settles and remains there by creating permanent change in the
capacity of the lake. In terms of storage capacity, the annual reduction due to siltation is found
to be low (0.08% of the total volume) suggesting that the silting-up of Lake Hawassa is less
than global average rate of reservoirs‟ sedimentation.

Regarding the bathymetric approach employed in this study, the technique is found to give an
insight to the depth and volume of in-lake sedimentation and provides important baseline
information. The results suggested that the approach is promising and provide acceptable
evidence against the hypothesis despite its limitations. Furthermore, the requirement of small
commitment of time and no involvement of permanent monitoring make the approach more
applicable.

86
 Chapter 7: The linkage of sedimentation in Lake Hawassa to the watershed characteristics

Chapter 7. The linkage of sedimentation in Lake Hawassa to the watershed

characteristics: an application of PSIAC model

7.1. Introduction
Erosion/sediment risk maps are specialized form of land resource evaluations, as they classify
basins of similar erosion/sediment risk degree (Gournellos et al., 2004). Erosion risk
assessment methods can be used for various tasks such as: assessment of average pattern of
erosion risk, identification of high risk areas, identification of hot spots, location of
depositional and major concentrated flow areas, detailed erosion and deposition pattern and
effects of conservation measures, and detailed impact of erosion on roads (Blinkov and
Kostadinov, 2010). Land use and soil conservation planning also require erosion risk maps and
this mapping can be done using deterministic erosion models that describe processes and
quantitative outcomes. However, the common drawback of these models is that they are
developed for different regions than where they are applied (Vrieling et al., 2002). Vrieling et
al. (2002) noted that a qualitative approach can be more effective in erosion risk mapping than
the use of models that were not developed for the region to which they are applied. Morgan
(2005) also considered the attempts of using a model for conditions outside those specified as
bad practice and, at best, speculative.

Generally, three types of approaches exist to assess erosion risks: qualitative approach,
quantitative approach, and model approach. All these methods vary in their characteristics and
applicability (Eckelmann et al., 2006). The ideal erosion model considers all of the factors
controlling soil erosion as it constitutes: the erosivity of the eroding agent, the erodibility of
the soil, the slope of the land, and the nature of the plant cover (Morgan, 2005); but in practice
such consideration is not yet achieved. Depending on the scale, a single or combination of
indices has been used by different researchers to assess the erosion risks. For instance,
Stocking and Elwell (1976) used mean annual erosivity values to assess the generalized
erosion risk in Zimbabwe. Similar attempts were made by Wischmeier and Smith (1978) for
USA; Hudson (1981) for Bulawayo and Harare; and Rowntree (1983) for Kenya. Stocking and
Elwell (1973) devised a factorial scoring system for rating erosion risk in Zimbabwe to the
scale from 1 (low risk) to 5 (high risk) in respect of erosivity, erodibility, slope, ground cover,
and human occupation. Vrieling et al. (2006) also showed the use of information on the
steepness of slopes and vegetation cover to map erosion risk in a watershed. According to

87
 Chapter 7: The linkage of sedimentation in Lake Hawassa to the watershed characteristics

Grimm et al. (2002), a problem with most methods based on scoring is that the results are
affected by the way the scores are defined.

In the same way, assessment of spatial variations of sediment yield potential provides
information for prioritizing watershed management interventions. Watershed sediment yield is
the product of all sediment producing processes and sediment transport within a watershed
(Vente, 2009). The important factors affecting sediment yields are: size of drainage area,
topography, soils, cover conditions, and degree of channelization (Robinson, 1977).
Traditionally, the problems in predicting sediment yield at the basin scale are related to
model‟s high data requirements (Vente et al., 2005) and these data are predefined to fit to the
specific model. The pre-defined nature of those models does not allow the use of other data
which are even better or latest. Such nature of the existing numerical models hinders their
application in many regions. On the other hand, there exists local and global data base like
DEM, satellite images and some kind of major land use/cover classifications and the less strict
model that can accommodate available environmental information for its input can easily be
adopted in many areas as far as it can be validated to the region. The more appropriate
technique at this moment is assumed to adopt less strict model and modify its parameterization
and validate in an area.

Though they have received only limited attention in the international literature, there are some
models that purport to have holistic approach, at least to some extent. Often, these models are
a combination of descriptive and quantitative procedures to characterize a drainage basin and
result in a quantitative or sometimes qualitative estimate of sediment yield in a basin.
Therefore, these models can be classified in general as semi-quantitative (Vente et al., 2005).

The PSIAC (Pacific-Southwest Inter-Agency Committee) model is one of the semi-

quantitative models (PSIAC, 1968) and it was employed in this study. The idea of employing a
semi-quantitative model in Lake Hawassa watershed was conceived in response to the absence
of validated model to readily apply and compare the output for the estimation of lake
sedimentation.

88
 Chapter 7: The linkage of sedimentation in Lake Hawassa to the watershed characteristics

7.2. Research questions and objective of the chapter

Triggered by the sedimentation rate of the lake (that was estimated in chapter 6), other
research questions were raised as: "Where is the source of sediment?" and "How is it linked to
the watershed characteristics?".

In addressing the above questions, this chapter targets:

 To characterize the watershed in terms the nine sediment yield factors (geology, soils,
climate, topography, land use, ground cover, run off, upland erosion, and sediment
transport in the stream channel);
 To compare the output of PSIAC model with previous findings.

7.3. PSIAC model

Table 7.1 presents different techniques of sediment source estimation and the various
erosion/sediment processes to be estimated by the techniques. Further description can be
referred from Gee and MacArthur (1996).

Table 7.1. Sediment source estimation techniques

Method/model Sheet Gully Channel bed Mass Average Single
and rill erosion and bank movement annual event
erosion erosion yield yield
USLE √
MUSLE √
RUSLE √
PSIAC √
Aerial photography √ √
Topographic survey √ √
Thompson or SCS TR32 √
Dendy and Bolton √
Strand and Pemberten, USBR √
SCS yield rate map √
(Source: Gee and MacArthur, 1996)

According to PSIAC (1968), the model is probably the most known semi-quantitative model
which was developed by the Pacific Southwest Inter-Agency Committee (PSIAC) for
application in arid and semi-arid areas in the southwestern USA. The model was
recommended for use in broad planning purposes and for basins of at least 25 km2. The PSIAC
model consists of a rating technique that characterizes a drainage basin in terms of sensitivity

89
 Chapter 7: The linkage of sedimentation in Lake Hawassa to the watershed characteristics

to erosion, possibilities for sediment transport and flood plain storage, the protective role of
vegetation, and the influence of human land use practices.

Nine factors characterize a watershed in PSIAC model with a score to each factor (table 7.2
and 7.3). The first division includes watershed parameters related to geographic features,
namely: X1= geology; X2= soils; X5= topography; X7= land use; and X6= ground cover. The
aforementioned parameters are natural parameters related to the geographical features. These
parameters respond to other parameters, such as X3= climate (rainfall), which causes erosion
and the development of gullies and rivers. The response of the geographic parameters to the
rainfall is represented by the following parameters: X4= run off; X8= upland erosion; and X9=
sediment transport in the stream channel. In PSIAC, the most important parameters are X8 and
X9 (Seyed et al., 2008) which together can form one third to one-half of the total parameters
together.

Table 7.2. PSIAC parameters and their diagnostic criteria modified after PSIAC (1968)
PSIAC Description Included in this factor Diagnostic criteria Unit
parameters
X1 Surface Resistance of the surface rocks to Surface geology class
geology erosion and sediment yield types
X2 Soils Resistance of the soil against erosion soil texture, the resistance Soil texture class
of particles, lime stone,
clay disperse and primary
humidity of soil;
X3 Climate Aggressiveness of the rainfall to Rainfall erosivity Value
cause erosion (to be derived from of R
rainfall amount)
X4 Runoff Potential of runoff generation Hydrologic soil
group classes
X5 Topography Contribution of topography for runoff Slope class
generation and erosion processes
X6 Ground Availability of covering material on Vegetation, litter, and rock Number of trees per
cover or above the surface of the ground fragments hectare and
against the effect of precipitation abundance of coarse
fragments
X7 Land use Type and intensity of use of the land class
by human (degree of natural
vegetative cover removal) (degree of
natural balance)
X8 Upland Existence and extent of rill, sheet and Observed erosion class
erosion gully erosion
X9 Channel Transport expectancy of the streams Shape of the channel, flow
erosion and duration, channel cross
sediment section, drainage density,
transport channel gradient, and
width-depth ratio

90
 Chapter 7: The linkage of sedimentation in Lake Hawassa to the watershed characteristics

Table 7.3. PSIAC factor ratings and degree of limitation modified after PSIAC (1968)
Land quality Quantitative Qualitative Degree of Description of suitability classes
Ratings Ratings limitation
0 Low Nil (a) massive hard formations
Slight –to- (a) rocks of medium hardness, (b) moderately weathered, (c)
Surface geology 5 Moderate
moderate moderately fractured
(X1)
Severe-to- very
10 High
severe (a) marine shales and related mudstones and siltstone
(a) high percentage rock fragments, (b) aggregated clays, (c) high
0 Low Nil
in organic matter
Soils Slight –to- (a) medium texture, (b) occasional rock fragments, (c) caliche
5 Moderate
(X2) moderate layers
Severe-to- very (a) fine texture, easily dispersed, saline–alkaline, high shrink–
10 High
severe swell characteristics, (b) single grain silts and fine sands
(a) humid climate with rainfall of low intensity, (b) precipitation
0 Low Nil in form of snow, (c) arid climate with low-intensity storms, (d)
arid climate with rare convective storms
Climate Slight –to- (a) storms of moderate duration and intensity, (b) infrequent
5 Moderate
(X3) moderate convective storms
(a) storms of several days duration with short periods of intense
Severe-to- very
10 High rainfall, (b) frequent intense convective storms, (c) freeze–thaw
severe
occurrence
(a) low peak flows, (b) low volume of runoff per unit area, (c)
0 Low Nil
rare runoff events
Runoff Slight –to- (a) moderate peak flows, (b) moderate volume of flow per unit
5 Moderate
(X4) moderate area
Severe-to- very
10 High
severe (a) high peak flows, (b) large volume of flow per unit area
0 Low Nil (a) gentle upland slopes (<5%), (b) extensive alluvial planes
Topography
Slight –to- (a) Moderate upland slopes (<20%) (b) moderate floodplain
(X5) 10 Moderate
moderate development
Severe-to- very (a) steep upland slopes (>30%), high relief, little or no floodplain
20 High
severe development
(a) completely protected by vegetation, rock fragments, litter;
-10 Low Nil
Ground cover little opportunity for rainfall to reach erodible material
(X6) Slight –to- (a) cover <40%; noticeable litter, (b) if trees present understory
0 Moderate
moderate not well developed
Severe-to- very (a) ground cover <20%, vegetation sparse, little or no litter, (b)
10 High
severe no rock in surface soil
-10 Low Nil (a) no cultivation, (b) no recent logging, (c) low-intensity grazing
Land use
Slight –to- (a) <25% cultivated, (b) 50% or less recently logged, (c) <50%
(X7) 0 Moderate
moderate intensively grazed, (d) ordinary road and other construction
Severe-to- very (a) >50% cultivated, (b) almost all of the area intensively grazed,
10 High
severe (c) all of area recently burned
0 Low Nil (a) no apparent signs of erosion
(a) about 25% of the area characterized by rill and gully or
Slight –to-
Upland erosion 10 Moderate landslide erosion, (b) wind erosion with deposition in stream
moderate
(X8) channels
Severe-to- very (a) >50% of the area characterized by rill and gully or landslide
25 High
severe erosion
Channel erosion (a) wide shallow channels with flat gradients, short flow duration
and sediment 0 Low Nil (b) channels in massive rock, large boulders or well vegetated, (c)
transport artificially controlled channels
(X9)
Slight –to-
10 Moderate (a) moderate flow depths medium flow duration with
moderate
occasionally eroding banks or bed

91
 Chapter 7: The linkage of sedimentation in Lake Hawassa to the watershed characteristics

7.4. Materials and methods

7.4.1. Parameterization of individual factors

[Link]. Geology factor

There are nine geologic classes in the watershed as identified by MoWR (2010) with a scale of
1:250,000 and the factor ratings were done by professional judgment (Ayenew, 2011: personal
communication) at Addis Ababa University, department of Earth Science. The rating values
range from 0 to10 as recommended by the PSIAC model.

[Link]. Soil factor

The soil factor rating was done based on available erodibility (K) values of the four major soil
types in the watershed as conducted by MoWR (2008, 2009, and 2010). The factor rating
values were computed by multiplying the K values by 16.67 as recommended by MPSIAC
which is the modified version and intended to avoid subjectivity in scoring the sediment yield
factors.

[Link]. Climate factor

The most commonly used index of rainfall aggressiveness, which is shown to be significantly
correlated with sediment yields in rivers, is the ratio p2/P, where p is the highest mean monthly
precipitation and P is the mean annual precipitation (Fournier, 1960). Morgan (1976) obtained
significant correlation between p2/P and drainage texture (defined as the number of first-order
streams per unit area). In this study, the climatic factor rating was based on Fourier Index (FI)
which computes rainfall erosivity based on maximum monthly rainfall amount and annual
rainfall amount (equation 7.1).

⁄ ………………………………………………..…….. (7. 1)

Where p is the mean monthly rainfall of the wettest month and P is the mean annual rainfall.

[Link]. Runoff factor

The runoff factor rating values were assigned to each of the 12 soil types of the watershed
(table 7.10). As a procedure, the infiltration capacity of each soil types was measured in the
field using double-ring infiltrometer (figure 7.1). The double-ring infiltrometer is a simple and

92
 Chapter 7: The linkage of sedimentation in Lake Hawassa to the watershed characteristics

routinely used instrument which is used to determine the infiltration rate of water into the soil.
After presoaking of the test areas, the double-ring infiltrometer are established on a level
surface. The larger ring has 60cm and the smaller one has 30cm with both depths of 25cm.
The water drops within specified time limits are recorded. The cans are refilled after each
reading. The drops that occur in the inner ring during the final period or the average stabilized
rate, expressed as cm/hr represents the infiltration rate for that specific soil type.

Inner
Stop ring
watch
Outer
Graduated ring
float

Figure 7.1. Field measurement of infiltration capacity using double-ring infiltrometer

[Link]. Topographic factor

The topographic factor rating was computed from the SRTM Digital Elevation Model (DEM)
with 30 x 30 m resolution. Values were assigned according to the revised PSIAC as shown in
table 7.4 below.

93
 Chapter 7: The linkage of sedimentation in Lake Hawassa to the watershed characteristics

Table 7.4. Rating of topographic factor

(% Slope) = (Points)
> 30= 20 18-20= 10
28-30= 19 17-18= 9
27-28= 18 15-17= 8
26-27= 17 14-15= 7
25-26= 16 12-14= 6
24-25= 15 11-12= 5
23-24= 14 9-11= 4
22-23= 13 8-9= 3
21-22= 12 6-8= 2
20-21= 11 5-6= 1
<5=0

[Link]. Land cover factor

The rating of land cover factor was done as recommended by MPSIAC which utilizes the
percentage of bare land and computes the rating value by equation 7.2.
Y = 0.2X…………………………….…………….………… (7.2)

Where Y is vegetation cover factor value and X is bare soil (%)

The percentage of bare land is generated from VCF (Vegetation Continuous Fields) which is
the product of MODIS (MODerate-resolution Imaging Spectroradiometer)(Hansen et al.,
2002a, 2002b, 2003) and downloaded from NASA website. MODIS is an extensive program
using sensors on two satellites where each provides complete daily coverage of the earth. The
VCF collection contains proportional estimates for vegetation cover types: woody vegetation,
herbaceous vegetation, and bare ground. The product was derived from all seven bands of the
MODIS sensor onboard NASA's Terra satellite. This product is good for showing how much
of a land cover such as "forest" or "grassland" exists anywhere on a land surface. The VCF
product represents the total area of the watershed by 5432 cells with pixel resolution of 500m
(table 7.11).

[Link]. Land use factor

The land use factor rating was done in a similar manner with the land cover factor as equation
7.3.

Y7 = 20 - 0.2X7 …………………………………………………… (7.3)

Where Y7 is vegetation cover factor value and X7 is canopy cover (%)

94
 Chapter 7: The linkage of sedimentation in Lake Hawassa to the watershed characteristics

The percentage of canopy cover was derived by merging the percentage of trees and herbs
from VCF of MODIS (Hansen et al., 2002a; 2002b; 2003).

[Link]. Upland erosion factor

The rating of upland erosion was done based on the type and degree of upland erosion. MoWR
(2010) classified the entire watershed into three erosion zones as high, medium and low. The
same classes are used in this study and the class with high rating assigned a rating value of 25,
a value of 15 for the medium class and 5 points for the low potential parts of the watershed.

[Link]. Channel erosion and sediment transport rating

This rating was done based on the spatial distribution of the drainage density in the watershed.
This technique is adopted from Buoko and Mazurova, 1958 (in Stroosnijder and Eppink
(1993)) in which an erosion class is attached to an elementary watershed depending on its
drainage density (table 7.5).

Table 7.5. Erosion class based on drainage density (Buoko andMazurova, 1958 (in
Stroosnijder andEppink (1993))
Class Erosion Degree Drainage density (km/km2) Rating value
1 Slight < 0.1 2
2 Moderate 0.1 ≤0.5 4
3 High 0.5 ≤ 1.0 6
4 Severe 1.0 ≤ 2.0 8
5 Very severe Greater or equal to 2 10

7.4.2. Arithmetic procedure for erosion/sediment risk assessment

The sediment yield index is the sum of values for the appropriate characteristics of each of the
nine factors as shown in table 7.6. The final results are categorized into 5 classes as per the
recommendation in the PSIAC model.

Table 7.6. PSIAC sediment classes

Rating Sediment yield potential Qualitative classifications
classes
> 100 1 Very High potential
75 – 100 2 High Potential
50-75 3 Moderate potential
25-50 4 Low potential
0-25 5 Very low potential

95
 Chapter 7: The linkage of sedimentation in Lake Hawassa to the watershed characteristics

7.5. Results and discussion

7.5.1. Result of surface geology rating

The result of the geology factor rating is shown in table 7.7 and figure 7.2 below.

Table 7.7. Rating of geological formations (from least to most susceptible for erosion)
Coding Geological property PSIAC
in map Rating
1 Qwo Obsidian and pitch stone 2
2 NQs Nazreth group and dino formation, undifferentiated 3
3 Qwa Rhyolitic and trachytic lava flows 4
4 Qdi Ignimbrites, tuffs, water lain byroclatics, occasional lacustrine beds 5
5 N1_2n Stratoid silicics: ignimbrites, unwelded tuffs, ash flows, rhyolites and trachytes 6
6 Qvs Volcanic sedimentary rocks: lacustrine dominantly volcanoclastics sediments, tuffs 7
7 Qwpu Pumice and unwelded tuffs 8
8 QI Lacustrine sediments: sand, silt, pyroclastic sediments, diatomites 9
9 Qdp Coarse unwelded pumicious pyroclastics 10

Figure 7.2. Spatial distribution of geology factor values

7.5.2. Soil rating result

As shown in table 7.8 and figure 7.3, there are four major soil types in the watershed. These
are Andosols, Cambisols, Luvisols and Leptosols with Nitisols and Regosols (MoWR 2008,
2009, and 2010). By adopting the recommendation of the modified PSIAC, the soil erodibility
values (table 7.8) are multiplied by a factor of 16.67 to arrive at the soil rating values of each
major soil.

96
 Chapter 7: The linkage of sedimentation in Lake Hawassa to the watershed characteristics

Table 7.8. Soil rating of major soil types

Major soil type K (USLE erodibility) Rating value (=16.67 K)
Cambisols (CM) 0.13 2.17
Luvisols (LV) 0.11 1.8
Andosols (AN) 0.2 3.33
Leptosols (LP) 0.22 3.67
Source: MoWR (2010)

Figure 7.3. Spatial distribution of soil factor values

7.5.3. Climate rating result

Table 7.9 presents the coordinates, elevation, and long-term annual rainfall for the five
meteorological stations in and around the watershed. Figure 7.4 demonstrates how the
Thiessen´s polygon delineates the respective area coverage of each station and the
corresponding rating values of the climate factor.

Table 7.9. Spatial distribution of meteorological station within and near to the study area
Station Name X Y Elevation Mean Rainfall of Fourier Climate
coordin coordinate (m.a.s.l) Annual RF wettest index rating
ate(m) (m) (mm) month (mm) value
1 Wendo Genet 456960 778399 1800 1151 150,80 19.8 6
2 Yirbaduwancho 433423 765478 2000 1120 155,10 21.5 7
3 Haisawita 451228 763012 2240 999 137,70 19 5
4 Hawassa 442235 779921 1750 953 124,10 16.2 2
5 Shashemene 455869 795581 1950 918 128,80 18.1 4

97
 Chapter 7: The linkage of sedimentation in Lake Hawassa to the watershed characteristics

Figure 7.4. Location of meteorological stations and climate factor values

7.5.4. Runoff rating results

[Link]. Results of Infiltration test

With the support of handheld PDA, which is installed with ArcPAD 10, each soil type (table
2.1 in chapter two) was investigated for its infiltration capacity. Totally, 10 infiltration
measurements are made in this study (measurement sites are shown in figure 7.5).

Figure 7.5. Locations of infiltration measurement sites

98
 Chapter 7: The linkage of sedimentation in Lake Hawassa to the watershed characteristics

[Link]. Rating of runoff potential from infiltration rate data

Based on infiltration rate values, each soil type was grouped into the corresponding hydrologic
soil group (HSG) as shown in table 7.10 below. The field assessment on the two soil types
(CMeu-Rbc and LP-Hde) reveals that there exists water impermeable layer within 50 cm of
the soil surface. According to the limits on the diagnostic physical characteristics of the
hydrologic soil groups, both soil types are assigned to group D. The maximum value of 27
cm/hr was recorded on Andosols which is characterized by deep to very deep and mostly
pumice below 40 cm. It may be the underlying pumice responsible to the high infiltration
capacity in addition to its medium to coarse texture. The final rating is shown in figure 7.6.

Table 7.10. Rating of runoff factor

Soil type Minimum infiltration rate Hydrologic soil PSIAC rating
(cm/hr) group
LVcr-Mde 18 A 1
CMeu-Ede 12 B 4
CMeu-Rcd 7 B 4
LVha-Rcd 6 B 4
CMeu-PLab 5 B 4
CMeu-PLbc 9 B 4
ANvi-PLab 27 A 0
CMcr-Mde 6 B 4
ANvi-Rcd 15 A 1
CMvr-PLab 3 C 7
CMeu-Rbc Very shallow D 10
LP-Hde Very shallow D 10
Source: Findings of field measurement

Figure 7.6. Distribution of hydrologic soil groups in the watershed

99
 Chapter 7: The linkage of sedimentation in Lake Hawassa to the watershed characteristics

7.5.5. Topography rating result

The rating of topographic factor was done as shown in table 7.4. Figure 7.7 shows the final
rating.

Figure 7.7. Rating of topographic factor based on slope percentage

7.5.6. Land cover rating result

[Link]. Extracting percentage of bare land for each land cover type
The available land cover types of MoWR (2010) was used in combination with the Vegetation
Continuous Fields (VCF) of Hansen et al. (2002a; 2002b; 2003). The land cover types were
used as classification units and the average percentages of bare land (figure 7.8) were
extracted from VCF using ArcInfo tools as shown in table 7.11.

100
 Chapter 7: The linkage of sedimentation in Lake Hawassa to the watershed characteristics

Figure 7.8. Derivation of bare land percentage from Continuous Fields (VCF) (with 500m
grids)

Table 7.11. Percentage of bare land in each land cover type (500m resolution)
Land cover type (Code) Count Mean % Land cover rating value
of bare land ( = 0.2X6)
1 Open shrubland (SO) 355 1.5 0.3
2 CI2 45 0 0
3 Intensive Mechanized Cultivation (State) (CIMS) 142 2 0.4
4 Plantation forests (FP) 54 0.9 0.2
5 Open grassland (GO) 259 4.4 0.9
6 Open grassland /Open woodland (GO/WO) 135 0 0
7 Lake 392 - -
8 Dense Shrubland (SD) 92 5.9 1.2
9 Urban or Built-Up Areas (U) 177 3.3 0.7
10 Intensive smallholders cultivation (CI3) 2468 0.7 0.1
11 Marshland (MA) 295 33.5 6.7
12 Open grassland with moderate smallholder 141 0.2
cultivation (CM3/GO) 0.04
13 Intensive smallholders cultivation (CI4) 1134 1.6 0.3
14 Disturbed High Forest (FD) 119 0 0

[Link]. Land cover rating from percentage of bare land

The final rating of land cover factor was computed through equation 7.2 as shown in figure
7.9 below.

101
 Chapter 7: The linkage of sedimentation in Lake Hawassa to the watershed characteristics

Figure 7.9. Rating of land cover factor

7.5.7. Land use rating result

[Link]. Extracting percentage of canopy for each land use types

VCF was produced for percentage bare soil, percentage of herbs and percentage of trees which
constitute 100% of the entire area (figures 7.8 and 7.10).

Figure 7.10. Derivation of tree percentage from Vegetation Continuous Fields (VCF)

For the rating of land use factor, the percentage of herbs and trees were added up to give
canopy percentage (figure 7.11).

102
 Chapter 7: The linkage of sedimentation in Lake Hawassa to the watershed characteristics

Figure 7.11. Percentage of canopy cover computed from MODIS satellite data

The average percentage of canopy was extracted for each land use types MoWR (2010) as
shown in table 7.12.

Table 7.12. Percentage of canopy in each land use type (500m resolution)
Land use type (Code) Count Mean % of Mean % of Total canopy %) Land use rating value
tree (T) herbs (H) (X7= T+H) (= 20 - 0.2X7)
1 GB(FP) 507 17.3 82.4 99.7 0.06
2 Intensively cultivated land (IAC) 608 10 90 100 0
3 Intensively cultivated land (IAC) 54 7.8 91.9 99.7 0.06
4 Intensively cultivated land (IAC) 88 10 90 100 0
5 TMFP 54 11 88.1 99.1 0.18
6 Grassland (G) 33 9.3 86 95.3 0.94
7 Grassland (G) 204 10 86.3 96.3 0.74
8 Grassland (G) 22 4.9 84.6 89.5 2.1
9 Lake WSFR(NC) 392 - - - -
10 GB(FP) 75 30 66.2 96.2 0.76
11 Urban (RCI) 13.1
177 (bare=3.34) 87.2 96.66 0.67
12 IPAC 2468 25 75 100 0
13 Swamp area (DGF(NC)) 295 0 96.8 96.8 1
14 MPAC(L) 141 13.2 86.6 99.8 0.04
15 Intensively cultivated land (IAC) 571 14.7 84.7 99.4 0.12
16 TM(NC) 119 54.8 45.2 100 0

103
 Chapter 7: The linkage of sedimentation in Lake Hawassa to the watershed characteristics

[Link]. Land use factor rating from percentage of canopy cover

Based on equation 7.3, the final rating of land use factor was computed as shown in figure
7.12.

Figure 7.12. Rating of land use factor

7.5.8. Upland slope erosion rating result

MoWR (2008, 2009, and 2010) assessed upland erosion in the watershed based on types and
degree of erosion. This study categorized the entire watershed into three qualitative erosion
classes as "severe", "medium" and "slight". The same erosion classes are used in this study.
Following the recommendation of PSIAC model to assign the value of 25 for the worst case,
the areas which had been classed as "severe" were assigned to have a value 25. The remaining
"medium" and "slight" classes proportionally assigned values of 15 and 5 respectively as
shown in figure 7.13.

104
 Chapter 7: The linkage of sedimentation in Lake Hawassa to the watershed characteristics

Figure 7.13. Upland land erosion rating

7.5.9. Channel erosion and sediment transport rating result

This factor was rated based on the available drainage network as identified by MoWR (2008,
2009, and 2010) and the corresponding density classes were adopted from Buoko and
Mazurova (1958) (cited in Stroosnijder and Eppink, 1993) as shown in table 7.5. Based on the
line density calculation tool of ArcGIS10, the final rating was done as shown in figure 7.14.

Figure 7.14. Sediment transport rating

105
 Chapter 7: The linkage of sedimentation in Lake Hawassa to the watershed characteristics

7.5.10. Identification of erosion/sediment source areas: the model output

The final identification of the erosion/sediment source areas was made by overlaying the
ratings of the nine factors using the "model run" option of ArcGIS10 as shown in figure 7.15.

Figure 7.15. Output diagram of model builder after running the raster input data by ArcGIS10

The final output of the modeling process is presented in figure 7.16 and table 7.13 below.
According to the model result, 66.4% of the watershed area is found to be classed as "low
potential"; 22.7% under "very low potential"; 10.6% "moderate potential" and only 0.3% is
under "high potential" and none of the watershed area is categorized as "very high potential".

106
 Chapter 7: The linkage of sedimentation in Lake Hawassa to the watershed characteristics

Figure 7.16. Spatial distribution of sediment yield potential classes in the watershed

Table 7.13. Percentages of each potential classes

Sediment Qualitative PSIAC Qs (Sediment yield = total % of the
class categories value volume of sediment /sediment watershed
producing area) (m3/km2/yr) area
5 Very high Potential 100 1450 0
4 High Potential 75-100 450-1450 0.3
3 Moderate potential 50-75 250-450 10.6
2 Low potential 25-50 95-250 66.4
1 Very low potential 25  95 22.7

The overall result of sediment source identification shows that the distribution of sediment
sources in the watershed is disproportionate and two distinct patterns are distinctly identified
by the model in the Western and Eastern parts of the watershed. Regarding the western part,
the high erosion rate and the coinciding high drainage density (as shown in figure 7.14);
sensitive geological formation (as shown in figure 7.2); high gully density (as shown in figure
6.4); and higher percentage of bare land (as shown in figure 7.8) seem to make this part of
the watershed to be the principal sediment source. But, the Eastern part seems to be influenced
by its topography (very steep slope) (as shown in figure 7.7) and high drainage density (as

107
 Chapter 7: The linkage of sedimentation in Lake Hawassa to the watershed characteristics

shown in figure 7.14). This part is the likely source of sediment that silted up in Lake
Cheleleka.

7.5.11. Comparison of sediment yield estimation to previous studies

To validate the accuracy of prioritization, some measured data are required but such data are
not available. In the current situation of data availability, testing for the prediction capacity of
the PSIAC model cannot be undertaken, and also not intended. However, comparing the
outputs with other studies offers some clue on the general and crude performance of the
model. In this case, the output of PSIAC model is compared with the estimation in-lake
sedimentation of chapter 6 and the Annualized Agricultural Non-point Sources (AnnAGNPS)
model output as studied by Shamo (2008). Individual outputs are shown in table 7.14 below.

Table 7.14. Summary of model outputs

Estimated Sediment yield Annual values of erosion and/or sediment yield Sources
1 PSIAC 95-250 m3/km2/yr Own study
(66.4 % of watershed area) and (chapter 7)
95 m3/km2/yr
(22.7 % of watershed area)
2 Hydrographic survey 967 m3/km2/year Own study
(long-term average) (chapter 6)
3 Ann-AGNPS  1 mm/ha/yr Shamo
(most part of the watershed) or (2008)
1000 m3/km2/yr

The magnitude of specific sediment as computed in chapter 6 is 967 m3/km2/year. In this case,
sediment contributing area is considered by deducting the surface area of the lake (about 93.24
km2) and the two closed sub-watershed: Muleti = 91.6 km2, and Wondo Kosha= 114 km2)
from the total watershed area (1436.5 km2). Whereas the PSIAC model estimates this
parameter as to fall between 95-250 m3/km2/yr (66.4 % of the watershed area) and  95
m3/km2/yr (22.7 %). Shamo (2008), who employed the AnnAGNPS model reported that most
of the area of the watershed has sediment yield of less than 1 mm/ha/yr (1000 m3/km2/year).

As shown in the above comparisons, the results of the hydrographic technique and
AnnAGNPS can be considered as fairly similar, but the PSIAC estimation is considerably
lower than the former results. The accuracy of all the three methods can be evaluated only if
actual sediment flow from the watershed has been monitored. For the moment, it is thinkable

108
 Chapter 7: The linkage of sedimentation in Lake Hawassa to the watershed characteristics

that result of the hydrographic survey is better because of the fact that it is the result of direct
measurement and comparison of two bathymetric maps.

7.6. Conclusion

This study targets the identification of critical sediment source areas in Lake Hawassa
watershed using PSIAC model. The model screens the hot-spots and the results are in good
agreement with field verifications and justify its sufficiency in achieving the objective as a
screening model. The final product remains qualitative because of the absence of measured
data to validate the method. On the basis of our results, the watershed area is classified into
four sediment yield classes (high, moderate, low, and very low potentials).

The model is open in nature and can accommodate more number of factors as far as it
influences the erosion and sedimentation process in the watershed. Such opportunity paves a
way for future adoption of the model by including new inputs. The subjectivity of
parameterization and final result deserve a special care. Moreover, the model shall be
quantitatively validated for its maximum benefit. The result of this study can be used for
watershed prioritization that is an inevitable part of watershed management which embodies
reduction of sediment deliveries into the lake.

109
Chapter 8: Participatory assessment of anthropogenic factors affecting the hydrology of Lake Hawassa

Chapter 8. Participatory assessment of anthropogenic factors affecting the hydrology of

Lake Hawassa: an application of DPSIR framework

8.1. Introduction
The list of environmental issues has been growing and their inter-linkages with their complex
causes and consequences are getting complex. To tell an integrated story of these issues, the
need of structured process (framework) that can accommodate interdisciplinary knowledge is
of a paramount importance (UNEP, 2008). In recent years, most of environmental assessment
studies are based on the casual chain frameworks (e.g. Pressure-State-Response (PSR),
Driving force-State-Response (DSR), and Driving force-Pressure-State-Impact-Response
(DPSIR)). These frameworks have made an important contribution by emphasizing the
importance of causality (Niemeijer and Groot, 2008).

DPSIR (pronounced dipsir), as one of the conceptual tools, helps to identify and describe
processes and interactions in human-environmental systems (Burkhard and Mueller, 2008). It
has a potential to link the existing data, gathered from various previous studies, in causal
relationship (Sekovski et al., 2012). As different cause and effect chains are included in the
model and because it is intended as an iterative loop, the model is adaptive to arising changes
and developments.

The DPSIR framework is an extension of the Pressure-State-Response (PSR) model,

developed by Anthony Friend in the 1970s, and subsequently adopted by the Organization for
Economic Cooperation and Development (OECD). It was developed for reporting purpose and
structures the description of the environmental problems by formalizing the relationships
between various sectors of human activity and the environment as causal chains of links.
Figure 8.1 presents the meaning of the DPSIR elements.

110
Chapter 8: Participatory assessment of anthropogenic factors affecting the hydrology of Lake Hawassa

DRIVERS The underlying factors influencing a variety of relevant variables

PRESSURES The variables which directly cause environmental problems

STATE The current condition of the environment

IMPACT The ultimate effects of changes of state

RESPONSE The efforts of the stakeholders to solve the problems

Figure 8.1. Terminology of DPSIR model Source: modified after EEA (1999; 2000).

By definition, DPSIR model is considered as the best way to structure environmental

information providing environmental socioeconomic integrity - in order to build links between
natural and socio-economic sciences; science and management; qualitative and quantitative
analyses; measured and modeled data; and definition of environmental syndromes (Turner et
al., 1998; EEA, 2003). Since its conception, the framework has increasingly been applied in
research projects with the aim of supporting decision making. A number of attributes of the
framework regarding structuring and communication issues in research further strengthen its
original purpose of bridging the science-policy gap (Tscherning et al., 2011). The full-fledged
causal chain from driving forces to impacts and responses is a complex task, and tends to be
broken down into sub-tasks, e.g. by considering the pressure-state relationship (Kristensen,
2004).

The previous chapters furnished the bio-physical information of the hydro-system and we
know less about the associated socio-economic challenges that are often the causes of many
biophysical challenges (Gregersen et al., 2007). Chapter four and five were devoted to the
assessment of hydro-climatic factors and chapter six showed the rate and magnitude of
sedimentation in the lake. Some of these variabilities are the result of a complex interplay
between natural and anthropogenic factors.

This chapter is intended to create a platform on which the cause-effect chain of anthropogenic
factors can be viewed in an integrated manner. It considers the management of Lake Hawassa
as synonymous with the "response"; and the current hydrological status of the lake as "state".

111
Chapter 8: Participatory assessment of anthropogenic factors affecting the hydrology of Lake Hawassa

D and P are considered to cause the current status of the lake. Figure 8.2 shows how the
DPSIR framework used in a decisional context.

Figure 8.2. The DPSIR framework in a decisional context Source: Vázquez, 2003)

8.2. Objectives of the chapter

The objectives of this chapter are dual. On one hand, the perception of stakeholders about the
likely anthropogenic causes that affect the hydrology of Lake Hawassa is assessed. On the
other hand, the findings of previous studies are assembled in the DPSIR framework so as to
build an integrated story that tells "how the lake hydrologically operates".

112
Chapter 8: Participatory assessment of anthropogenic factors affecting the hydrology of Lake Hawassa

8.3. Methodology

8.3.1. General methodology

Figure 8.3 presents the flow chart of the general methodology in which the process starts with
identifying the stakeholders followed by contacting them as individual or group. Focus group
discussions and individual interviews were the two main techniques employed. After building
a common understanding on the existing cause-effect links, the participants were encouraged
to fill the DPSIR story sheets (figure 8.4). The results then cross-checked and supplemented
by secondary sources.

Stakeholders
identification

Contacting
stakeholders

Individuals Group

Individual Focus group

interviews discussion

Building common understanding on

DPSIR and cause-effect links

Filling DPSIR story sheets in

formal meetings

Reviewing secondary data for

more information

Causal chain

Figure 8.3. Flowchart for causal chain assessment of anthropogenic factors

8.3.2. DPSIR story sheet and participatory approach

The more detailed information was collected during the focus-group discussions (Cameron,
2000) which were conducted with participants (representing stakeholders) who have good
familiarity with the topics and the study area. To give everyone the opportunity to express
her/his opinion, six to eight people participated in each focus group. A total of five focus

113
Chapter 8: Participatory assessment of anthropogenic factors affecting the hydrology of Lake Hawassa

group discussions were conducted in this study. The participants were encouraged to fill the
DPSIR story sheet (figure 8.4). Other participatory methods which were used as community
truthing include (i) key informant interviews (Hay, 2004), and (ii) participants observations
(Cook, 1997). Those "experts" who were identified by local people as having special
knowledge (Warburton and Martin, 1999) were also interviewed.

Once the participants of the focus group discussions were, they were oriented about the
concept of DPSIR chain and provided with two DPSIR story sheets (figure 8.4) to write the
most likely cause-effect chains as per their own perception. The two main issues/topics in the
cause-effect chain were: lake level rise and lake sedimentation. The participants were
encouraged to talk openly and discuss the issues with other members and state their opinion
before filling the sheets. Due to the absence of incentives, the time of discussion was
deliberately reduced.

Drivers (D) Pressures (P) State (S) Impacts (I)

--------------- --------------- ---------------

--------------- --------------- ---------------
--------------
--------------- --------------- ---------------

Responses (R): policies - strategies - proclamations - programs - plans

Figure 8.4. Template of DPSIR story sheet

114
Chapter 8: Participatory assessment of anthropogenic factors affecting the hydrology of Lake Hawassa

8.4. Results and discussion

8.4.1. "Indirect drivers" (iD)

This study identified the need for one more component to the original DPSIR which helps to
explain causes of the drivers (D) and can be viewed as a variation to the framework. This
additional component is named as "indirect driver" and abbreviated as "iD" as shown in figure
8.5 below. Similar naming was used by Maxim et al.(2009).

The perceived principal indirect drivers/root causes underlying the causal-chain affecting the
hydrology of Lake Hawassa are found to include: population growth and density, agricultural
development, the use of wood as primary source of energy, socio-political changes, and the
existing land tenure system (figure 8.5). The upcoming sub-sections discuss the individual
issues in detail.

Indirect drivers (iD) Direct drivers (D)

1. Population growth and density

2. Agricultural development 1. Land use changes
3. The use of wood as fuel 2. Deforestation
4. Socio-political changes 3. Unsustainable use of lands
5. Existing land tenure system

Figure 8.5. Primary and secondary driving forces affecting the hydrology of the system

[Link]. Population growth and density

Demographics of the watershed communities, as one of the root cause in the causal chain was
also perceived by Esayas (2010), MoWR (2008), Dessie (2004), WWDSE (2001),
Shewangizaw and Michael (2010) and many others. Having 2005 as base year, the population
of the watershed is estimated to be 621,530 people. This is expected to double before 2025,
and grow to nearly 1.6 million by 2035 at an average growth rate of 3.15%, but with a higher
rate of 3.8% in the early years to 2010 (table 8.1). Population density of the watershed is
estimated as 624 people/km2 (MoWR, 2008) which is unusually high and about eight fold of
the average value at country level, that is about 77.72 people/km2 (World Bank, 2012).

The livestock population is also additional stress which constitutes 516,159 cattle, 95,035
sheep, 95,035 goats, 12,763 horses, 28,912 donkeys and 486 mules. The livestock density is

115
Chapter 8: Participatory assessment of anthropogenic factors affecting the hydrology of Lake Hawassa

calculated as 335 TLU/km2 (MoWR, 2008). TLU = tropical livestock unit; for example: 1
TLU = Camels 1.0, Cattle 0.7, Sheep/Goats: 0.1 (FAO, 2005). The impacts of population
growth are also being felt in the watershed with population pressure impacting many of the
other issues such as deforestation, land degradation, overgrazing, and increasing food
insecurity.

Table 8.1. Population growth in seven administrative units in and around the watershed
2005 2010 2015 2020 2025 2030 2035
Total Male Female Population Population Population Population Population Population
Place
ArbeGona 12,971 6,521 6,451 15,590 18,479 21,621 25,080 28,931 33,128
Hawassa 430,664 219,536 211,127 517,622 613,539 717,857 832,725 960,610 1,099,952
Shebedino 77,594 39,428 38,165 93,261 110,542 129,337 150,032 173,073 198,178
Siraro 1,811 888 924 2,241 2,695 3,142 3,614 4,167 4,790
Kofele 8,553 4,242 4,311 10,583 12,728 14,840 17,067 19,679 22,623
Shalla 15,748 7,717 8,031 19,485 23,433 27,320 31,421 36,229 41,647
Shashemene 74,190 36,836 37,354 91,802 110,410 128,725 148,048 170,708 196,238
Source: MoWR (2008)

[Link]. Expansion of agriculture

This perception of the stakeholders is also supported by Dessie (2004), Esayas (2010) and
others. According to these sources, expansion of agriculture in particular smallholder farming,
contributes to over 80% of the forest area loss. This expansion is characterized by two major
modes of change: 1) internal: clearings created by the intrusion of small farm plots, grazing
lands, and villages 2) external: expansion of agriculture from the exterior into the forests.
Generally, the changes from natural vegetation to cultivation as depicted by figure 8.6
indicates that agricultural expansion in the watershed is the most important proximate cause of
land use change and it takes place at the expense of other land uses.

Figure 8.6. Trends of agricultural land expansion in the watershed Source of raw data: Abrha
(2007)

116
Chapter 8: Participatory assessment of anthropogenic factors affecting the hydrology of Lake Hawassa

[Link]. The use of "wood" as energy source and absence of alternative energy
Fuel woods supplies 84% of total energy demands of which about 50% is from shrub-lands
and wood-lands (exceeding their mean annual increment of woody biomass) and only 5-10%
is from woodlots with the remainder from crop residues and dung (MoWR, 2008). Household
energy requirements are supplied largely by fuel wood collected from existing woodland and
shrub land, maize straw, and charcoal with cow dung also used in the western part of the
watershed. The use of biomass accelerates the rate of deforestation and erosion while the use
of crop residues and dung as fuel, rather than returning this organic matter to the soil, causes a
decline in soil fertility and deterioration in soil structure.

[Link]. Socio-political changes

Ethiopia has witnessed several dramatic political changes during the course of the last century.
These changes have been accompanied by transitional periods characterized by uncertainty
and insecurity. In the absence of firm political control, control of resources has also been
lacking. Many among the rural population have then taken the opportunity to usurp what has
been available in terms of, for instance, forest land and forest products, adding to the process
of forest decline (Dessie, 2004; Dessie and Christiansson, 2008).

[Link]. Land tenure system

All lands in Ethiopia are nationalized and redistributed in 1975. This policy has continued with
the present Government and the 1994 Constitution specifies that land cannot be subject to sale
or exchange (FAO, 2004). Issues of land tenure could include insecurity of tenure, ability to
use land as collateral and the transferability of property rights and the impacts these have on
land investment or factor (land, labor or capital) allocation. A major source of tenure
insecurity emanates from the periodic land redistribution to land-poor households (Mahmud
Joseph and Pender, 2005). This indirect driver is also identified by MoWR (2010). Currently,
the government is implementing a land certification programme that is expected to reduce the
effect of land insecurity among the land users but its effectiveness is not yet evaluated.

117
Chapter 8: Participatory assessment of anthropogenic factors affecting the hydrology of Lake Hawassa

8.4.2. "Direct drivers" (D)

[Link]. Deforestation, land use and land cover changes

As perceived by the stakeholders, expansion of agriculture and population growth are the two
main causes for the land use changes in the watershed. The magnitudes of land use changes
were attempted to be addressed by different studies, such as Wagesho et al. (2012) and Abrha
(2007) for the period of 1973, 1986, and 2000 at the watershed scale (table 4.3 in chapter
four). Ayenew et al. (2007) also stated that the land use changes in the watershed affected the
hydrology of the lake. While assessing the causes of these land use changes, Legesse et al.
(2003) noted that they are resulted from multiple forces such as: demographic trends, climate
variability, national policies, and macroeconomic activities.

The watershed has also undergone progressive deforestation (Lamb et al., 2002) and the
results of interview and individual/group discussions also reveal that "deforestation" is one of
the direct drivers which is perceived to be caused by the use of "wood" as source of energy for
cooking (its impact is also exaggerated by the absence of alternative energy); socio-political
changes especially during the period of political transitions; and agricultural expansions.
Similar causal chain was reported by Dessie (2004) which adds the economic activities and
local conflicts over resources to play an important role. According to Dessie (2004), the total
natural forest loss between 1972 and 2000 amounted to over 40,000 ha, which is over 80 % of
the forest cover that was present in 1972. This corresponds to an annual loss of over 1400 ha,
equivalent to 0.9% of the annual national loss. The decreasing trend of forest coverage in the
study area coincides with the general forest decline pattern in Ethiopia. The forest decline
during this period was not an isolated event, but rather a continuation of the past trend. Esayas
(2010) also mentioned that forest clearing exerted pressures on the natural resources of the
watershed.

During field visit, it was also observed that forests which were owned by the community were
depleting which might be explained by the lower regard for common property and common
access land than for individual holdings. Woodlands are also depleted which seems to be
driven by the demand for fuel wood and charcoal. There are few remnants of trees seen in the
Western side of the lake. These remnants are preserved against cutting due to few influential

118
Chapter 8: Participatory assessment of anthropogenic factors affecting the hydrology of Lake Hawassa

community members who created awareness among their neighbors, and now these trees are
taken as symbols of environmental protection.

[Link]. Unsustainable use of lands

Farmers seem to lack the commitment to implement and maintain the already implemented
measures. Previously constructed soil and water conservation works are also destroyed or
abandoned in many places in the watershed due to the lack of maintenance. Such reluctance is
perceived by the participants as to emanate from the existing land tenure system. Some
justified that it could be due to the lack of communal understanding on the importance of these
measures and lack of land management enforcement. Cultivation of steep slopes is also one of
the common misuses of the lands in the watershed.

8.4.3. "Pressures" (P)

The "pressure" component in this study constitutes the input of water and sediment to the lake
that potentially affect the hydrological status of the lake.

Direct drivers (D) Pressures (P)

1. Land use changes

1. Sedimentation
2. Deforestation
2. Increased runoff into the lake
3. Unsustainable use of lands

Figure 8.7. Link between direct drivers and pressures

[Link]. Sedimentation
In chapter 6, it was shown that the thickness of accumulated sediment in the lake between
1999 and 2010 was about 14 ± 5cm (figure 6.3). If a constant annual rate is assumed, the
sedimentation rate would be 1.2 cm/year and it can be considered as a pressure and part of the
causal chain. Soil erosion and sedimentation can be influenced by both climate and
anthropogenic factors and differentiation between them is recommended.

[Link]. Increased runoff from the watershed

As discussed in section 4.5.4 of chapter four, the runoff from Tikur Wuha River shows an
increasing trend whereas rainfall has no significant changes. The role of anthropogenic factors

119
Chapter 8: Participatory assessment of anthropogenic factors affecting the hydrology of Lake Hawassa

(through land use/cover changes) can be deduced from the interplay between the trending
runoff and non-trending rainfall.

8.4.4. "State" (S)

As a result of the factors that were described as pressures, the hydrology of Lake Hawassa has
been changed as indicated in chapter four that was expressed in terms of its significant
increasing trend (average annual rise of 4.9 cm/yr). As justified in section 4.5.6 (chapter
four), the anthropogenic factor has a stake atleast in modifying the impact of climate. This in
turn contributes for the resultant rise of the lake level. The sequential regime shifts in the lake
level series are less likely associated with human factors.

8.4.5. "Impact" (I)

The notable impact of Lake Hawassa water level rise is the historical recorded flood that had
occurred in 1998/99 (WRDB, 2007; WWDSE, 2001). In monetary terms, this destruction
accounted for about €5.4 million (WRDB, 2007). The corresponding impact on human well-
being and environment are not yet studied as to the author´s knowledge. However, it is
apparent that when the environmental factors change, for whatever reason, the individuals,
communities and even economic sectors that depend on these factors are also affected in
myriad ways (UNEP, 2008).

8.4.6. "Responses" (R)

Various policies are available in Ethiopia which can serve as legal ground towards sustainable
management of water resources in general and Lake Hawassa in particular. A "response" by
society or policy makers is the result of an undesired impact and can affect any part of the
chain between driving forces and impacts (Kristensen, 2004). Pursuing the enforcement of
current policy instruments is taken as an option in this study and the following paragraphs
discuss the existing policies, proclamations and programmes.

While grouping the main findings of the first four components of the DPSIR framework as
shown in the previous sections, the six major environmental issues as shown in figure 8.8 are
apparently identified to examine the existing "response" instruments. These issues are:
deforestation, population growth and density, the use of wood as fuel, land tenure system,
unsustainable utilization of land resources, and socio-political changes. Corresponding to the

120
Chapter 8: Participatory assessment of anthropogenic factors affecting the hydrology of Lake Hawassa

six key issues which belong to the D-P-S-I elements, there found about ten relevant and
overlapping response instruments (more number of responses can be identified if more time
was available than this study) that are directed towards managing them.

Population growth The use of wood as fuel Land tenure system

and density

io n
lat l
ra t
o pu Ethiopian Ru men gy
lP y Energy policy e p e
na lic Th elo trat
tio Po v S
N a De cy/
li
For Po
es
stra try res
tegi earc
c pl h l land
an ion on rura
Fores Proclamat n and use
io
action tr y administrat
Deforestation

progr

land resources
Unsustainable
amme Response

utilization of
Instruments MERET pr
oject
vation,
Conser lisation
Forest n Uti
d
pment a Prod
Develo oclamation uctive
S
Pr Prog afety Net
iopia r am m
yo f Eth e
ateg
n str ?
s er vatio
Con

Socio-political changes

Figure 8.8. List of responses (the red question mark at the bottom of the circle shows the gap
of response instrument)

As shown in the above figure, a couple of response instruments have been designed by the
government to mitigate the deforestation problem (the left most box). These include: Forestry
Research Strategic Plan, Forestry Action Programmes, Forest Conservation, Development, and
Utilization Proclamation, and Conservation Strategy of Ethiopia. These response instruments
provide legal framework for sustainable management of forest resources.

Regarding population growth, the National Population Policy (NPP) articulated the
Government‟s position on the relationship between demographic and economic growth, with
sustainable and equitable human development as its central theme. The policy stated that
without a reduction in population growth, the efforts to reduce poverty and the achievement of
national development goals would be jeopardized.

The Ethiopian Energy Policy encourages energy mix, and in the long-term, a replacement of
the traditional sources of fuel by modern technologies. Parallel to this, the policy promotes
country-wide afforestation programme to supplement traditional fuels.

121
Chapter 8: Participatory assessment of anthropogenic factors affecting the hydrology of Lake Hawassa

The rural development policy and strategy justifies the voluntary resettlement programmes
(can also be viewed as a response against population density). This response pays attention to
the land tenure issue and the proper use of land. Important changes such as the moratorium on
land re-distribution and the distribution of land certificates are given a legal basis in this
instrument.

The proclamation on rural land administration and use defines the individual land use and
disposal rights. It defines obligations of rural land users and land use restrictions. Thus,
protection of land becomes an obligation and failure to protect can lead to loss of title.

The last, but not the least, is the issue of socio-political changes that induce uncertainty and
insecurity during transition periods of political regimes. Dessie and Christiansson (2008)
identified that large areas of forest were cut down during periods of political transition as a
result of the political vacuum. Instruments to avoid such problems were not recognized by the
stakeholders (the red question mark at the bottom of figure 8.8) except that they recommend
creation of ownership feeling among the community so that socio-political changes have less
effect on the natural resources. Such occurrences are exemplified in some parts of Ethiopia.

Generally, as recognized from the causal chains, there is a loose link between response
instruments (R) and the rest of the DPSIR components which are probably attributed to the
inefficiency of institutional arrangements in implementing an established regulation. Such
situation reminds the policy makers to pay more attention to the appropriate implementation of
available management instruments together with designing new ones. Long-term education for
the implementing bodies and for the general community seems to work at this point.

8.5. Summary and conclusions

The overall result of this section shows that the lake level rise (State) of Lake Hawassa has
been accompanied by some anthropogenic factors. The immediate causes (Pressures) of this
level rise comprise sedimentation into the lake and increased runoff in which both pressures
are perceived to be influenced by land use changes, deforestation and unsustainable use of
land resources (Drivers). These drivers in turn impacted by indirect drivers (indirect Drivers)
that comprise population growth and density, expansion of agriculture, the use of wood as
fuel, socio-economic changes, and the existing land tenure system. The impact of lake level

122
Chapter 8: Participatory assessment of anthropogenic factors affecting the hydrology of Lake Hawassa

rise was assessed in view of its risk to produce flood (Impact) which usually impacts the
environment and human well-being in addition to economic losses.

This conceptual exercise about the causal paths is found to provide an aid to logically combine
information from different sources to tell an integrated story. It effectively highlights the
causal relationships and provides a useful first step towards the establishment of a full-fledged
causal network for the major environmental problems of Lake Hawassa hydro-system. The
results provide a better understanding of why and how people destroy their environment and
careful translation of this understanding into plans and actions enables the stakeholders to
prevent or reduce further destruction. The solutions are at the reach of stakeholders by acting
locally while thinking at watershed scale.

The structuring and integration of available and new environmental information using DPSIR
conceptual framework is found to illustrate the overall status of the Lake Hawassa
hydrosystem and paved a better understanding towards sustainable management of the
ecosystem as a whole.

With regard to the last component of the framework: R, a set of existing policies and legal
documents were assessed to explore the possible management options towards the other four
components of the model. It was then recognized that the available policies and legal
documents have a promising potential for mitigation, adaptation or curative actions against the
anthropogenic wings of the problems at hand.

8.6. Limitation of the study

Many of the relationships between the human system and environmental system are complex
and may not be well understood (Maxim et al., 2009). The underlying assumption of simple
causal relations cannot fully capture the complexity of interdependencies in the real world
(Spangenberg et al., 2002). Besides, it is sometimes difficult to provide conclusive evidence of
a cause-effect relationship as is required for the application of the DPSIR logic (EEA, 2005).

Generally, a full-fledged causal-link is not always necessary or, if any, it needs long and
intensive researches of integrated approach. In this study, the overall causal links were derived
from available information, researcher‟s and stakeholders‟ experience on the topic at hand. In

123
Chapter 8: Participatory assessment of anthropogenic factors affecting the hydrology of Lake Hawassa

this regard, it is strongly recommended to have future researches of similar framework, but
with filling gaps in that this study falls short of, if possible, by generating primary evidences.

124
 Chapter 9: Synthesis

Chapter 9. Synthesis

9.1. General remarks

The main objective of this study was to investigate the causes of Lake Hawassa water level
variability in general and its resultant rise in particular. A number of researchers attempted to
address the issue and this research can be viewed as one of those efforts. In this study, the
impact of climate variability and sedimentation was in a greater detail than has been done
before.

Particularity of this research is depicted by the diverse approaches employed to provide a

comprehensive insight into the hydrological status of Lake Hawassa. The logical combination
of coherence analysis, regime shift detection, trend analysis, lake level simulation, bathymetry
surveys, sediment yield modeling, and DPSIR analysis enables the thesis to provide an
integrated story about the characteristics of the hydrosystem. It can be considered as one of the
potential documents to guide the future management of Lake Hawassa. The following section
tries to summarize the over-all findings of this study in the form of DPSIR components.

9.2. Synthesis of causal links

The temporal variability of Lake Hawassa water level is shown to be influenced by the
combined effect of anthropogenic, natural and climate related factors. Figure 9.1 (below)
synthesized a simplified relationship among these factors in the DPSIR framework.

As depicted by figure 9.1, the causes of Lake Hawassa water level rise can be broadly
classified into two: hydro-climatic variability (a) and anthropogenics (b).The occurrences of
climate shifts and extreme events such as ENSO phenomena are among some of the
variabilities responsible for the resultant rise of the lake level. The neotectonic process is
believed to affect the lake level but not investigated in this study. Sedimentation in the lake
and the increasing trend of runoff from the watershed are among the immediate causes to
affect the lake level variability and these pressures can be related to both natural and human
inductions and the dominant factors are not yet fully analyzed. Further in the causal chain, the
sedimentation and increased runoff are manifested by land use changes, deforestation and
unsustainable use of land.

125
 Chapter 9: Synthesis

A ANTHROPOGENICS

Population
INDIRECT Expansion of The use of Socio-political Land tenure
growth and
DRIVERS agriculture wood as fuel changes system
density

Land use Unsustainabe

Deforestation
changes use of land
DIRECT
DRIVERS

Soil erosion

Increasing
PRESSURES Sedimentation trend of runoff
over time
(a) Regime shifts in
STATE lake level
(b) Lake level rise
(4.9 cm/yr)

Impact on the environment

Ecosystem
IMPACT b Flooding
disturbance
Impact on the society

B HYDRO-CLIMATIC VARIABILITY

On land surface On lake surface

Short-term
(a) Extreme and Short-term (a ) Extreme and
PRESSURES alteration of
(b) long-term runoff extreme (b) long-term decrease
ground water
changes rainfalls in evaporation
regime

a a b
Partially explained by:
DIRECT Climate shifts and
- Reduction of wind speed Neotectonism
DRIVERS ENSO phenomena - Increment of humidity

INDIRECT
Climatic puzzle Earthquake
DRIVERS

Figure 9.1. Summary of causal link and relationship among anthropogenic and natural factors

9.3. Overall Conclusions

The primary factor that influences the level rise of Lake Hawassa is found to be climate
variability. This factor is manifested by the extreme and simultaneous occurrences of high
rainfall and the corresponding runoff which are responsible for the entry of extreme amount of
water into the lake system.

126
 Chapter 9: Synthesis

The two prominent climate events which strongly influence the hydrology of Lake Hawassa
are: the climate shift of North Pacific Ocean that occurred in 1976/77 and the El Niño events
that occurred in 1972-73, 1982-83, 1997-98, and 2009-10. The neotectonic activities that
occurred in 1996, ´97 and ´98 consecutively, are also considered to involve in the interplay but
direct evidences are not available.

In the long-term perspective, "rainfall" has neither been significantly increasing nor decreasing
over time on both annual and monthly bases. Despite this, "runoff" has shown significant
increasing trend. The interplay between these different trends is considered to justify the role
of land use/cover changes at least in modifying the impact of climate.

The increased runoff has been perceived to be directly driven by land use changes,
deforestation, and unsustainable use of land. These direct drivers (D) have also been perceived
to result from population growth and density, agricultural development, the use of wood as
fuel, socio-economic changes, and the prevailing land tenure system and indirect drivers (iD).
The effect of sedimentation in reducing the storage capacity of the lake is found to be low
(0.08% per year) as compared to the total volume of water in the lake.

The historical maximum recorded flood which had occurred in 1998/99 appeared to be caused
by the simultaneous occurrence of the 1996 high rainfall followed by the worst El Niño (1997-
98) by which runoff, rainfall, and evaporation were affected. The tectonic activities that
occurred in the consecutive years of 1996, ´97 and ´98 were also assumed to affect the ground
water flow, but no concrete evidence to verify it.

It is generally concluded that the lake level variability of Lake Hawassa is more reactive to
extreme climate events than the long-term natural and anthropogenic factors.

9.4. Perspectives
This study provides a new dimension of articulating the causes of Lake Hawassa water level
rise by considering both anthropogenic and natural factors on one hand and long-term and
extreme temporal extents on the other. It was shown that almost the entire water balance
components have been interplayed in one way or another in affecting the temporal variability
of the lake level in general and its resultant increasing trend in particular.

127
 Chapter 9: Synthesis

The association of flood from Lake Hawassa to the climate anomalies resulting from natural
mechanisms could have two managerial implications; on one hand, it would be difficult to
mitigate the problem because of its dependence on macro-scale processes and on the other
hand, an optimistic view of those large El Niño events which are notorious for their extreme
floods are acceptably predictable within period at lead times of up to two years (Chen et al.,
2004). Such opportunities are useful to get an alarm against the urgency of flood occurrences,
even in the absence of local monitoring data and downscaling can be endorsed and
mainstreamed in a regular early warning and assessment activities to reduce the impact of
potential flood risks.

This causal-loop is by no means exhaustive, but believed to provide an orderly guidance for
future research and development interventions. As noted by Gregersen et al. (2007), integrated
watershed management can only be effective if they are grounded in the technical realities of
what is going on with the soil, water and biophysical resources and the interactions between
them. Thus fundamentally, an effective interaction or combination of institutional and
technical information is required for successful watershed management that results in lasting
benefits for the stakeholders living in the watershed.

128
 References

Abraha BA (2007) Assessment of downstream pollution profiles of Hawassa textile factory

effluent along Tikur Wuha River using physico-chemical and macro-invertebrate
indicators. Unpublished [Link] thesis, Addis Ababa University, Ethiopia
[[Link] .
Abraha L (2007) Assessing the impact of land use and land cover change on ground water
recharge using RS and GIS: a case of Awassa catchment, South Ethiopia. Unpublished
[Link] thesis, Addis Ababa University, Ethiopia
[Link]
Achamyeleh K (2003) Integrated flood management, flood management case study 1, the
associated program on flood management, Agricultural Science, 10, pp. 85-97
Acreman MC, Meigh JR, and Sene KJ (1993) Modeling the decline in water level of Lake
Toba, Indonesia. Advances in water resources, 16, pp. 207-222
Aklilu A (2011) Water Quality Assessment of Eastern shore of Lake Hawassa Using Physico-
chemical Parameters and Benthic Macro-invertebrates. Unpublished [Link] thesis,
Addis Ababa University. Ethiopia.
[[Link]
Alemayehu T (2007) Water use, management and industrial pollution in Awassa. Proceedings
of the public meetings on harnessing the water resources of Ethiopia for sustainable
development in the new Ethiopian millennium, forum for environment, Addis Ababa
[[Link]
Alemayehu T, Furi W, and Legesse D (2007) Impact of water over exploitation on highland
lakes of Eastern Ethiopia. Environmental Geology, 52, pp. 147-154
Alemayehu T, Ayenew T, and Kebede S (2006) Hydrogeochemical and lake level changes in
the Ethiopian [Link] of Hydrology, 316 (1-4), pp. 290-300
Altunkaynak, A, Özger M, and Sen Z (2003) Triple diagram model of level fluctuations in
Lake Van, Turkey. Hydrology and earth system sciences, 7(2), pp. 235-244
Arnell N, Bates B, Lang H, Magnusson JJ, and Mulholland P (1996) Climate change 1995:
impacts, adaptations and mitigations of climate change: scientific technical analyses.
Cambridge Univ. Press, Cambridge
[[Link]
Arnell NW, Liu C, Compagnucci R, Cunha LD, Hanaki K, Howe C, Mailu G, Shiklomanov I,
Stakhiv E, Döll P, Becker A, and Zhang J (2001) Hydrology and water resources. In:
McCarthy JJ, Canziani OF, Leary NA, David J. Dokken DJ, and White KS (eds).
Climate change 2001: impacts, adaptation, and vulnerability, Cambridge University
Press, pp. 191-233
Awulachew SB (2006) Characteristics Investigation of physical and bathymetric
characteristics of Lakes Abaya and Chamo and their management implications. Lakes
and reservoirs research and management, 11(3), pp. 133-140
Awulachew SB, Yilma AD, Loulseged M, Loiskandl W, Ayana M, and Alamirew T (2007)
Water resources and irrigation development in Ethiopia. International water
management institute, working paper 123, 78pp
Ayalew E (2003) Application of stable isotopes in the study of lake dynamics in Ziway-Shala
basin. Unpublished [Link] thesis, Addis Ababa University, Ethiopia
Ayenew T (1998) The hydrogeological system of the Lake District Basin. Central Main
Ethiopian Rift. Published PhD thesis (Book, ISBN. 9061641586), Free University of
Amsterdam, The Netherlands, 259pp

129
Ayenew T (2001) Numerical groundwater flow modeling of the Central Main Ethiopian Rift
lakes basin. SINET: Ethiopian J. Sci., 24 (2), pp. 167-184
Ayenew T (2004) Environmental implications of changes in the levels of lakes in the
Ethiopian Rift since 1970. Regional environmental changes, 4, pp. 192-204
Ayenew T (2006) Vulnerability of selected Ethiopian lakes to climatic variability, tectonism
and water use: global change system for analysis, research and training (START)
[[Link]
Ayenew T (2007) Water management problems in the Ethiopian rift: challenges for
development. Journal of African Earth Sciences, 48 (2-3), pp. 222-236
Ayenew T and Tilahun N (2008) Assessment of lake-groundwater interactions and
anthropogenic stresses, using numerical groundwater flow model for a Rift lake
catchment in central Ethiopia. Lakes and reservoirs: research and management, 13, pp.
325-343
Ayenew T and Gebreegziabher Y (2006) Application of a spreadsheet hydrological model for
computing the long-term water balance of Lake Awassa, Ethiopia, Hydrological
Sciences Journal, 51 (3), pp. 418-431
Ayenew T, Becht R, Lieshout A, Gebreegziabher Y, Legesse D, and Onyando J (2007)
Hydrodynamics of topographically closed lakes in the Ethio-Kenyan Rift: the case of
lakes Awassa and Naivasha. Journal of Spatial Hydrology 7 (1)
Babu A (2009) The impact of Pacific sea surface temperature on the Ethiopian rainfall. Paper
presented at workshop on high-impact weather predictability information system for
Africa and AMMA-THORPEX forecasters‟ handbook. Trieste, Italy
[[Link]
Barnston AG, Mason SJ, Goddard LD, DeWitt G, and Zebiak SE (2003) Increased automation
and use of multimodel ensembling in seasonal climate forecasting at the IRI. Bull.
Amer. Meteor. Soc., 84, pp. 1783-1796
Barnston AG, Chelliah M, and Goldenberg SB (1997) Documentation of a highly ENSO-
related SST region in the equatorial Pacific. Atmos.-Ocean, 35, pp. 367-383
Becht R, David D, and Harper M (2002) Towards an understanding of human impact upon the
hydrology of Lake Naivasha, Kenya. Hydrobiologia, 488, pp. 1-11
Bekele W (2003) Economics of Soil and Water Conservation. Theory and empirical
application to subsistence farming in the Eastern Ethiopian highlands. Unpublished
doctoral thesis (AGRERIA 404), Swedish University of Agricultural Sciences,
Department of Economics, Uppsala, Sweden
[[Link]
Belay EA (2009) Growing lake with growing problems: integrated hydrogeological
investigation on Lake Beseka, Ethiopia. Doctoral thesis. Rheinischen Friedrich-
Wilhelms-Universität Bonn, Germany [[Link]
Belete A (2009) Climate change impact on Lake Abaya water level. Unpublished [Link] thesis,
Addis Ababa University, Ethiopia
[[Link]
Bendant JS and Piersol AG (1986) Random data analysis and measurement procedures. 2nd ed.
Wiley, 566pp
Bergonzini L (1998) Bilans hydriques de lacs (Kivu, Tanganyika, Rukwa and Nyassa) du
rift est-africain. Mus. Roy. Afr. Centr., Tervuren, Ann. Sc. Geol. 103, 183pp
Bewketu K (2010) Hydrodynamics of selected Ethiopian Rift lakes. Unpublished [Link] thesis,
Addis Ababa University, Ethiopia
[[Link]

130
Billi P and Dramis F (2003) Geomorphological investigation on gully erosion in the Rift
Valley and northern highlands of Ethiopia. Catena, 50, pp. 353-368
Biltoft CA and Eric RP (2009) Spectral coherence and the statistical significance of turbulent
flux computations. Journal of atmospheric and oceanic technology, 26, pp. 403-409
Birkett C, Murtugudde R, and Allan T (1999) Indian Ocean Climate event brings floods to
East Africa's lakes and the Sudd Marsh, Geophys. Res. Lett., 26(8), pp. 1031-1034
Blinkov I and Kostadinov S (2010) Applicability of various erosion risk assessment methods
for engineering purposes. BALWOIS 2010 - Ohrid, Republic of Macedonia
[[Link]
ov%20&%20Kostadinov%202010_USLE_PESERA_KINEROS_WEP_WEPP_EPM.pdf.
Bloomfield P (1976) Fourier analysis of time series: An introduction. Wiley
Box GEP and Jenkins GM (1976) Time Series Analysis: Forecasting and Control. Holden-
Day. San Francisco, California, USA
Breaker LC (2007) A closer look at regime shifts based on coastal observations along the
Eastern boundary of the North Pacific. Continental Shelf Research 27, pp. 2250-2277
Brinkmann W (1983) Association between net basin supplies to Lake Superior and supplies to
the lower Great Lakes. J. Great Lakes Res. 9, pp. 32-39
Buishand TA (1982) Some methods for testing the homogeneity of rainfall records. J.
Hydrology, 58(1-2), pp. 11-27
Burkhard B and Mueller F (2008) Ecological indicators:Driver-Pressure-State-Impact-
Response. In: Jorgensen SE and Fath B (eds) Encyclopedia of Ecology. Elsevier
publishers, Amsterdam
Burn DH (1994) Identification of a data collection network for detecting climate change. Can.
Water Resour. J., 19, pp. 27-38
Cameron J (2000) Focusing on the focus group, In: Hay I(Ed.) Qualitative Research Methods
in Human Geography, Victoria, Australia, Oxford University Press
Chang H (2007) Comparative stream flow characteristics in urbanizing basins in the Portland
Metropolitan Area, Oregon, USA. Hydrol. Process, 21(2), pp. 211-222
Chen J and Gupta AK (2000) Parametric Statistical Change Point Analysis. Birkhauser,
Boston, Massachusetts, USA. 240pp
Chen D, Mark AC, Alexey K, Stephen EZ, and Daji H (2004) Predictability of Niño over the
past 148 years. Nature, 428, pp. 733-736
Cherinet T (1982) Hydrogeology of the Lakes Region, Ethiopia. Ministry of Mines and
Energy, Addis Ababa, Unpublished document, 97pp
Cheung WH, Gabriel BS, and Ashbindu S (2008) Trends and spatial distribution of annual
and seasonal rainfall in Ethiopia. International journal of climatology, 24, pp. 1345-
1358
Chokolov AA and Chapman TG (1974) Methods for water balance computations: An
international guide for research and practice. A contribution to the international
hydrological decade [[Link] 
Christensen MR, Graham MD, Vinebrooke RD, Findlay DL, Paterson MJ, and Turner MA
(2006) Multiple anthropogenic stressors cause ecological surprises in boreal
lakes. Glob. Change Biol.,12, pp. 2316-2322
Cohen SJ (1986) Impacts of CO2-induced climatic change on water resources in the Great
Lakes Basin. Climatic Change, 8, pp.135-53
Cohen S, Ianetz A, and Stanhill G (2002) Evaporative climate changes at Bet Dagan, Israel:
1964-1998. Agric. For. Meteor., 111, pp. 83-91

131
Cook KC (1997) Participant observation, In: Flowerden R and Martin D (eds) Methods in
Human Geography. A Guide for Students doing a Research Project. London, Sage
Publications
Craye M and Funtowicz S (2005) A reflexive approach to dealing with uncertainties in
environmental health
[[Link]
Crétaux JF, Jelinski W, Calmant S, Kouraev A, Vuglinski, Bergé NM, Gennero MC, Nino F,
R. Abarca R, Cazenave A, and Maisongrande P (2011) SOLS: A Lake database to
monitor in Near Real Time water level and storage variations from remote sensing
data, J. Adv. Space Res., 47(9), pp. 1497 - 1507
Darken PF (1999) Testing for changes in trend in water quality data. PhD thesis. Faculty of
Virginia Polytechnic Institute and State University
[[Link]
Darling WG, Gizaw B, and Arusei MK (1996) Lake-groundwater relationships and fluid-rock
interaction in the African Rift Valley: isotopic evidence. Journal of African Earth
Sciences, 22, pp. 423-431
Deganovsky AM and Getahun BA (2008) Water balance and level regime of Ethiopian lakes
as integral indicators of climate change. Proceedings of Taal2007: The 12th World
Lake Conference, pp. 1092-1102
DeNoyelles F and Jakubauskas M (2008) Current state, trend, and spatial variability of
sediment in Kansas reservoirs. In: Sedimentation in our reservoirs: causes and
solutions. Kansas State University agricultural experiment Stations and cooperative
extension service, 142pp
Dessie G and Christiansson C (2008) Forest decline and its causes in the South-Central Rift
Valley of Ethiopia: human impact over a one hundred year perspective. Ambio, 37 (4),
Royal Swedish Academy of Sciences
Dessie G (2007) Forest decline in South Central Ethiopia: extent, history and process. PhD
thesis, Department of Physical Geography and Quaternary Geology, Stockholm
University [[Link]
Dessie N (1995) Hydrogeological investigation of Lake Hawassa catchment. Unpublished
[Link] thesis, Addis Ababa University, Ethiopia
Desta Z (1997) Industrial Environmental Management: the case of Awassa Textile Factory,
Ethiopia. [Link]. thesis, Wageningen Agricultural University, The Netherlands, 95pp
[[Link]
Dost RJ and Mannaerts CM (2006) Generation of Lake Bathymetry using sonar, satellite
imagery and GIS. International Institute for geo‐information science and earth
observation (ITC). Enschede, The Netherlands
Douglas BC, Michael SK, and Stephen P (2001) Sea level rise: history and consequences.
International Geophysics, 75, pp. 1-232
Dunne T and Leopold LB (1978) Water in Environmental Planning. W.H. Freeman & Co.
New York. 818pp
Eckelmann W, Baritz R, Bialousz S, Bielek P, Carré, F, Houšková B, Jones J , Kibblewhite
RJA, Kozak M, Le BJ, Tóth CG, Tóth T, Várallyay G, Halla MY, and Zupan M
(2006) Common Criteria for Risk Area Identification according to Soil Threats,
Research report no. 20, Soil Information Working Group (SIWG), European Soil
Bureau Network (ESBN)
[[Link] 

132
EEA (1999) Environmental indicators: typology and overview. European Environment
Agency, Copenhagen [[Link]
EEA (2000) Down to earth: soil degradation and sustainable development in Europe.
Environmental Issues, Series, No. 16. European Environment Agency, Copenhagen,
32pp [[Link]
EE (2003) Europe‟s environment: the third assessment. Environment ssessment Report No.
10, EEA. Copenhagen
[[Link] 
EEA (2005) Sustainable use and management of natural resources, EEA Report No 9/2005,
72pp [[Link]
Elfert SS and Bormann H (2010) Simulated impact of past and possible future land use
changes on the hydrological response of the Northern German lowland „Hunte‟
catchment. Journal of Hydrology, 383 (3-4), pp. 245-255
Elsawwaf M and Willems P (2012) Analysis of the climate variability on Lake Nasser
evaporation based on the Bowen ratio energy budget method. J. Environ. Biol. 33, pp.
475-485
Engle RF (1976) Interpreting Spectral Analyses in Terms of Time-Domain Models. Annals of
Economic and Social Measurement, 5(1)
[ttp://[Link]/~boller/Econ.350/Engle_76.pdf
Environment Canada (2010) Global Effects of El Niño
[[Link]
Esayas Y (2010) Evaluating the impact of land use/land cover change on soil erosion and
runoff using SWAT model at Tikur Wuha Watershed. Unpublished [Link] Thesis,
Addis Ababa University, Ethiopia
[[Link] 
FAO (1977) Guidelines for soil profile description. Soil Resources Development and
Conservation Service. Land and Water Development and Conservation Division. 2nd
Edition, Rome
FAO (2004) Livestock information, sector analysis and polica branches-AGAL
[[Link] 
FAO (2005) Livestock Sector Briefs. Food and Agriculture Organization, Rome
[[Link] 
Fitzpatrick WP, William CB, and Nini GB (1987) Sedimentation and hydrologic processes in
Lake Decatur and its watershed. Illinois state water survey
[[Link]
Flohn H (1987) East African Rains of 1961/1962 and the Abrupt Change of the White Nile
Discharge, Palaeoecology of Africa,18, pp. 3-18
Fortin V and Gronewold AD (2012) Water balance of the Laurentian Great Lakes.
Encyclopedia of Earth Sciences Series, pp. 864-868
Fournier F (1960) Climat et érosion: la relation entre l'érosion du sol par l'eau et les
précipitations atmosphériques, Presses Universitaires de France, Paris, 203, pp. 281-
345
Gao P, Mu X-M, Wang F, and Li R (2011) Changes in stream flow and sediment discharge
and the response to human activities in the middle reaches of the Yellow River.
Hydrol. Earth Syst. Sci., 15, pp. 1-10

133
Gasith A and Gafny S (1990) Effects of water level fluctuation on the structure and function
of the littoral zone. In: Tilzer MM, Serruya C, (eds.) Large Lakes: Ecological Structure
and Function. Berlin (Germany), Springer-Verlag, pp. 156-171
Gebreegziabher Y (2004) Assessment of the Water Balance of Lake Awassa Catchment,
Ethiopia. [Link] Thesis, International Institute for Geo-information Science and Earth
Observation [[Link] 
Gebre-Mariam Z and Desta Z (2002) The chemical composition of the effluent from Awassa
Textile factory and its effects on aquatic biota. SINET: Ethiop. J. Sci., 25(2), pp. 263-
274
Gebremichael H (2007) Modeling and forcasting Hawassa Lake Level fluctuation,
Unpublished [Link]. Thesis, Addis Ababa University, Ethiopia
Gee DM and MacArthur RC (1996) Use of land surface erosion techniques with stream
channel sedimentation models. US Army Corps Engineers. Technical paper No. 152
[[Link]
GEOS (2013) Notes on time series analysis. University of Arizona
[[Link]
Geremew Z (2000) Engineering Geological Investigation and Lake Level Changes in the
Awassa basin. [Link]. thesis, Department of Geology and Geophysics, Addis Ababa
University, Addis Ababa, Ethiopia
Ghanbari RN, Hector RB, John JM, William GH, and Barbara JB (2009) Coherence between
lake ice cover, local climate and teleconnections (Lake Mendota, Wisconsin). Journal
of Hydrology, 374, pp. 282-293
Gilbert RO (1987) Statistical methods for environmental pollution monitoring. Van Nostrand
Reinhold, New York, 320pp
Giupponi C (2002) From the DPSIR reporting framework to a system for a dynamic and
integrated decision making process. MULINO Conference on “European policy and
tools for sustainable water management” 21-23 November, Venice (Italy)
[Link]
Goddard L and Graham NE (1999) The importance of the Indian Ocean for simulating
precipitation anomalies over Eastern and Southern Africa. J. Geophys. Res., 104 (16),
pp. 19, 099-19, 116
Goddard L and Dilley M (2005) El Niño: Catastrophe or Opportunity. J. Climate, 18, pp. 651-
665
Goerner A, Jolie E, and Gloaguen R (2009) Non-climatic growth of the saline Lake Beseka,
Main Ethiopian Rift. Journal of Arid Environments, 73, pp. 287-295
Goossens C and Berger A (1987) How to recognize an abrupt climatic change? Abrupt
Climatic Change: Evidence and Implications, In: Berger WH and Labeyrie LD (eds.),
Kluwer Academic Publishers, Dordrecht, pp. 31-46
Goswami M, O'Connor KM, Bhattarai KP, and Shamseldin AY (2005) Assessing the
performance of eight real-time updating models and procedures for the Brosna River,
Hydrol. Earth Syst. Sci., 9, pp. 394-411
Gournellos T, Evelpidou N, and Vassilopoulos A (2004) Developing an Erosion Risk Map
Using Soft Computing Methods (Case Study at Sifnos Island). Natural Hazards, 31, pp.
63-83
Gregersen HM, Peter FF, and Brooks KN (2007) Integrated Watershed Management:
Connecting people to their land and water. CAB International, 288pp

134
Grove AT, Street FA, and Goudie AS (1975) Former lake levels and climatic change in the
Rift Valley of southern Ethiopia, Geographic Journal, 142, pp. 177-194
Haile T (1988) Causes and characteristics of drought in Ethiopia. Ethiopian Journal of
Agricultural Sciences, X (1-2), pp. 85-97
Halcrow and partners (1979) Master plan for the development of surface water resources in ten
Awash basin, vol 6. Ministry of Water Resources, Addis Ababa, Ethiopia. Unpublished
document
Hansen M, DeFries R, Townshend JRG, Sohlberg R, Dimiceli C, and Carroll M (2002b)
Towards an operational MODIS continuous field of percent tree cover algorithm:
examples using AVHRR and MODIS data. Remote Sensing of the Environment 83, pp.
303-319 [[Link]
Hansen M, DeFries R, Townshend JR, Carroll M, Dimiceli C, and Sohlberg R (2003)
Vegetation Continuous Fields MOD44B, 2001 Percent Tree Cover, Collection 3,
University of Maryland, College Park, Maryland, 2001 [[Link]
Hansen M, DeFries, R, Townshend JRG, Carroll M, Dimiceli C, and Sohlberg R (2003)
Global percent tree cover at a spatial resolution of 500 meters: first results of the
MODIS vegetation continuous fields algorithm. Earth Interact., 7, pp. 1-15
Hare SR and Mantua NJ (2000) Empirical evidence for North Pacific regime shifts in 1977
and 1989. Progress in oceanography, 47, pp. 103-145
Hartmann DL (2013) ATM 552 Notes: Time Series Analysis - Section 6a
[[Link] 
Hay I (2004) Qualitative Research Methods in Human Geography. Oxford University Press,
Oxford
Hayashi M and Garth van der K (2007) Water level changes in ponds and lakes: the
hydrological processes. Plant disturbance ecology: the process and the response
[[Link] 
Henderson SA (1992) Continental cloudiness changes this century. Geo. Journal, 27, pp. 255-
262
Hernando O and Bellegem SV (2006) Coherence analysis of non-stationary time series: a
linear filtering point of view. Discussion paper - 618. UCL - “ EUEN/CORE - Center
for operations research and econometrics
[[Link]
Hirata FE, Jünior O, and Mata M (2010) Regime shifts, trends and inter-annual variations of
water level in Mirim Lagoon, southern Brazil. Pan-American Journal of Aquatic
Sciences, 5(2), pp. 254-266
Hirsch RM (1988) Statistical methods and sampling design for estimating step trends in
surface water quality. Water Resour. Res., 24, pp. 493-503
Hudson NW (1981). Soil conservation, 2nd edn. Batsford, London
Hurni H, Bantider A, Herweg K, Portner B, and Veit H (2007) Landscape transformation and
sustainable development in Ethiopia. Background information for a study tour through
Ethiopia, September 4-20, 2006, CDE (Center for Development and Environment )
University of Bern , Bern , 321pp
[[Link] 
Hurni H (1985) Erosion-Productivity-Conservation Systems in Ethiopia. Proceedings of the
4th International Conference on Soil Conservation, Maracay, Venezuela, pp. 654 - 674
Hutchinson GE (1957) A treatise on limnology. New York, John Wiley & Sons
Jenkins GM and Watts DG (1968) Spectral Analysis and its Applications, Holden Day

135
Jensen ME (2010) Estimating evaporation from water surfaces. Paper presented at the
CSU/ARS Evapotranspiration Workshop, Fort Collins, CO
[[Link] 
Jhonson CW and Gembhart AC (1982) Predicting sediment yield from Sagerbrush range
lands. USDA SEAARM Western Series, 26, pp. 145- 56
Ji ZG (2008) Hydrodynamics and Water Quality: Modeling Rivers, Lakes, and Estuaries. John
Wiley & Sons, Inc., Hoboken, New Jersey
Jöhnk KD, Dietmar S, and Wolfgang O (2004) Water level variability and trends in Lake
Constance in the light of the 1999 centennial flood. Limnologica, 34, pp.15-21
Jones FE (1992) Evaporation of Water: with Emphasis on Applications and Measurement.
Lewis Publ., Inc., Chelsea, MI, 200pp
Kachroo RK (1992) River flow forecasting. Part 1. A discussion of the principles. J.
Hydrology, 133, pp. 1 -15
Kachroo RK (1992) River flow forecasting. Part 5. Applications of a conceptual model. J.
Hydrol. 133, pp. 141-178
Kadioglu M and Sen Z (2001) Monthly precipitation-runoff polygons and mean runoff
coefficients. Hydrological Sciences, 46(1), pp. 3-11
Kahya E and Kalayci S (2004) Trend analysis of stream flow in Turkey. J. Hydrol., 289, pp.
128-144
Kane RP (1999) Some characteristics and precipitation effects of the El Niño of 1997-1998.
Journal of Atmospheric and Solar-Terrestrial Physics, 61(18), pp. 1325-1346
Kang HM and Yusof F (2012) Homogeneity tests on daily rainfall series in Peninsular
Malaysia. Int. J. Contemp. Math. Sciences, 7(1), pp. 9-22
Kazmin, V (1979) Stratigraphy and correlation of volcanic rocks in Ethiopia. Ethiopian
Institute of Geological Surveys Note 106, pp. 1-26
Kebede E, Zinabu GM, and Ahlgren I (1994) The Ethiopian rift-valley lakes: chemical
characters along a salinity-alkalinity gradient. Hydrobiologia, 288, pp. 1-12
Kebede S, Travi Y, Alemayehu T, and Marc V (2006) Water balance of Lake Tana and its
sensitivity to fluctuations in rainfall, Blue Nile basin, Ethiopia. Journal of Hydrology
316, pp. 233-247
Kendall MG (1975) Rank Correlation Measures. Charles Griffin, London, UK
Kinshiro N (1974) Hydrological Regime on the Regulation of Lake-level, A Contribution to
the Paleolimnological Study of Lake Biwa. Department of Geophysics, faculty of
Science, Hokkaido University, Hokkaido
[[Link]
Koopmans LH (1974) The Spectral Analysis of Time Series. Academic Press, 366pp
Korecha D and Barnston AG (2007) Predictability of June-September Rainfall in Ethiopia.
American Meteorological Society. Monthly weather review, 135(2), pp. 628-650
Kottegoda NT (1981) Some aspects of non-stationary behavior in hydrology. In: El-Shaarawi
AH and Esterby SR (eds). Time series methods in hydrosciences. Proceedings of an
international conference held at Canada Center for Inland waters, Burglington, Ontario,
Canada
Kristensen P (2004) The DPSIR Framework. Paper presented at the 27-29 September 2004
workshop on a comprehensive / detailed assessment of the vulnerability of water
resources to environmental change in Africa using river basin approach. UNEP
Headquarters, Nairobi, Kenya
[[Link] 

136
Kulakowski D, Bebi P, and Rixen C (2011) The interacting effects of land use change, climate
change and suppression of natural disturbances on landscape forest structure in the
Swiss Alps. Oikos, 120, pp. 216-225
Lamb AL, Leng MJ, Lamb HF, Telford RJ, and Mohammed MU (2002) Climatic and non-
climatic effects on the δ18O and δ13C compositions of Lake Awassa, Ethiopia, during
the last 6.5 ka . Quaternary Science Reviews 21(20-22), pp. 2199-2211
Lamb HH (1966) Climate in the 1960s. Geogr. J., 132, pp. 183-212
Langbein WB (1961) Salinity and Hydrology of Closed Lakes. A study of the long-term
balance between input and loss of salts in closed lakes. Geological survey professional
paper 412, Washington [[Link]
Legesse D and Ayenew T (2006) Effect of improper water and land resource utilization on the
central Main Ethiopian Rift lakes. Quaternary International, 148(1), pp. 8-18
Legesse D, Vallet-Coulomb C, and Gasse F (2004) Analysis of the hydrological response of a
tropical terminal lake, Lake Abiyata (Main Ethiopian Rift Valley) to changes in climate
and human activities. Hydrol. Process, 18, pp. 487-504
Legesse D, Vallet-Coulomb C, and Gasse F (2003) Hydrological response of a catchment to
climate and land use changes in Tropical Africa: case study South Central Ethiopia, J.
Hydrology, 275, pp. 67-85
Lenters JD, Kratz TK, and Bowser CJ (2005) Effects of climate variability on lake
evaporation: Results from a long-term energy budget study of Sparkling Lake, northern
Wisconsin (USA). Journal of Hydrology, 308, pp. 168-195
LIA (Lake Allawara Authority) (2010) [[Link]
Limgis (2001) Winter School on Essentials in Limnology and Geographic Information System
(GIS) [[Link]
Litinskaya KD (1973) Long-term water stage regime in natural lakes of Karelia and its
changes in view of conversion of lakes into storage reservoirs. Hydrology of Lakes
[Link]-AISH Publication No. 109
Lu Z, Walker GR and Dawes WR (2002) Water balance modeling: concepts and applications.
In: McVicar TR, Li R, Walker J, Fitzpatrick RW, and Changming L (eds), Regional
Water and Soil Assessment for Managing Sustainable Agriculture in China and
Australia, ACIAR Monograph No. 84, pp. 31-47
Magaña V, Pe´rez JL, Va´zquez JL, Carrisola E, and Pe´rez J (1999) El Niño y el clima.
Contribucio´n al libro Impacto dEl Niñoen Me´xico. In: Magaña, V., (Ed.), UNAM,
Mexico, pp. 23-66
Maidment DR (1993) Handbook of Hydrology. Mcgraw Hill. Newyork
Mann HB (1945) Non-parametric tests against trend. Econometrica, 13, pp. 245-259
Martinko EA, DeNoyelles J, and Jakubauskas M (2011) Bathymetric mapping and sediment
studies in Kansas Reservoirs: 2010-2011. Kansas Biological Survey
[[Link]
Marucci SD (2002) The impact of El Niño Southern Oscillation Phenomenon on the Panama
Canal and its Markets. Panama Canal Authority-Corporate planning and marketing
department: Marketing Division [[Link]
Maxim L, Spangenberg JH, and O‟Connor M (2009) n analysis of risks for biodiversity
under the DPSIR framework. Ecological Economics, 69, pp. 12-23
McBride JL and Nicholas N (1983) Seasonal relationship between Australian rainfall and the
Southern Oscillations. American Meteorological Society. Monthly weather review, 111
McPhaden MJ (1999) The child prodigy of 1997-1998. Nature, 398, pp. 559-562

137
Mercier F, Cazenave A, and Maheu C (2002) Inter-annual lake level fluctuations (1993-1999)
in Africa from Topex/Poseidon: connections with ocean-atmosphere interactions over
the Indian Ocean. Global and Planetary Change, 32 (2-3), pp. 141-163
Meshesha DT, Tsunekawa A, Tsubo M, and Haregeweyn N (2011) Spatial analysis and semi-
quantitative modeling of specific sediment yield in six catchments of the central Rift
Valley of Ethiopia. Journal of Food, Agriculture and Environment, 9 (3 and 4), pp. 784
-792
Meshesha DT, Tsunekawa A, Tsubo M, and Haregeweyn N (2012) Dynamics and hotspots of
soil erosion and management scenarios of the Central Rift Valley of Ethiopia.
International Journal of Sediment Research, 27, pp. 84-99
Miller AJ, Cayan DR, Barnett TP, Graham NE, and Oberhuber JM (1994) The 1976-77
climate shift of the Pacific Ocean. Oceanography, 7(1), pp. 21-26
Milliman J, Farnsworth K, Jones P, Xu K, and Smith L (2008) Climatic and anthropogenic
factors affecting river discharge to the global ocean, 1951-2000, Global Planet.
Change, 62, pp. 187-194
Modarres R and Silva VPR (2007) Rainfall trends in arid and semi-arid regions of Iran. J.
Arid Environ., 70, pp. 344-355
Mogaka H, Gichere S, Davis R, and Hirji R (2006) Climate variability and water resources
degradation in Kenya: improving water resources development and management.
World Bank working paper 69. Washington DC
[[Link]
[Link]?sequence=1
Monteith JL (1965) Evaporation and environment. pp. 205-234. In: Fogg GE (ed.) Symposium
of the Society for Experimental Biology, The State and Movement of Water in Living
Organisms, 19, Academic Press, Inc., NY
Morgan RPC (2005) Soil and erosion and conservation. Third edition. Blackwell publishing
Morgan RPC (1976) The role of climate in the denudation system: a case study from West
Malaysia. Chapter 11. In: Derbyshire E (ed) Geomorphology and Climate. John Wiley
and Sons, London
MoWR (2008) The Federal Democratic Republic of Ethiopia, Ministry of Water Resources:
Rift Valley Lakes Basin Integrated Resources Development Master Plan Study Project.
Phase 1 Report: Lake Hawassa Sub-basin Integrated Watershed Management
Prefeasibility Study. Part 1 and 2. Halcrow Group Limited and Generation Integrated
Rural Development (GIRD) Consultants. Addis Ababa. Unpublished document
MoWR (2009) The Federal Democratic Republic of Ethiopia-Ministry of Water Resources:
Rift Valley Lakes Basin Integrated Resources Development Master Plan Study Project.
Phase 2 Report: Lake Hawassa Sub-basin Integrated Watershed Management
Prefeasibility Study. Part 1 and 2. Halcrow Group Limited and Generation Integrated
Rural Development (GIRD) Consultants. Addis Ababa. Unpublished document
MoWR (2010) The Federal Democratic Republic of Ethiopia-Ministry of Water Resources:
Rift Valley Lakes Basin Integrated Resources Development Master Plan Study Project.
Phase 3 Report: Lake Hawassa Sub-basin Integrated Watershed Management
Feasibility Study. Part 1 and 2. Halcrow Group Limited and Generation Integrated
Rural Development (GIRD) Consultants. Addis Ababa. Unpublished document
Mu X, Zhang L, McVicar TR, Chille B, and Gau P (2007) Analysis of the impact of
conservation measures on stream flow regime in catchment of the Loess Plateau,
China. Hydrol. Process, 21, pp. 2124-2134
Muradian R (2001) Ecological thresholds: a survey. Ecological Economics, 38, pp. 7-24

138
Namdar-Ghanbari R and Bravo HR (2008) Coherence between atmospheric teleconnections,
Great Lakes water levels. and regional climate. Advances in Water Resources, 31(10),
pp. 1284-1298
Namdar-Ghanbari R, Bravo HR, Magnuson JJ, Hyzer WG, and Benson BJ (2009) Coherence
between lake ice cover, local climate and teleconnections (Lake Mendota, Wisconsin).
Journal of Hydrology, 374, pp. 282-293
Nash JE and Sutcliffe JV (1970) River flow forecasting through conceptual models. Part I- a
discussion of principles. Journal of hydrology,10, pp. 282-290
Ngaira JK (2006) Implications of climate change on the management of Rift Valley lakes in
Kenya. School of Environment and Earth Sciences, Maseno University, Kenya
[[Link]
management_of_Rift_Valley_lakes_in_Kenya._The_case_of_lake_Baringo
Nicholson SE (1996) A review of climate dynamics and climate variability in Eastern Africa.
In: Johnson TC and Odada E (eds). The Limnology, Climatology and Paleoclimatology
of the East African Lakes. Gordon and Breach, Toronto, pp. 25-56
Nicholson SE (1995) „Environmental Change within the Historical Period‟. In: Goudie AS,
Adams WM. and Orme A (eds.) The Physical Geography of Africa, Oxford University
Press, Oxford, pp. 60-75
Nicholson SE and Selato JC (2000) The influence of La Niña on African rainfall. International
journal of climatology, 20, pp. 1761-1776
Nicholson SE and Yin X, and Ba MB (2000) On the feasibility of using a lake water balance
model to infer rainfall: an example from Lake Victoria. Hydrological Sciences, 45(1),
pp. 75-95
Niebauer HJ (1998) Variability in Bering Sea ice cover as affected by a “regime shift” in the
North Pacific in the period 1947-96. J. Geophys. Res., 103(C12), pp. 27,717-27,737
Niemeijer D and Rudolf SG (2008) Framing environmental indicators: moving from causal
chains to causal networks. Environ Dev Sustain, 10, pp. 89-106
NNSMP-National non-point source monitoring program (2005) Monitoring data: Exploring
your data-the first step. Technical note 1
[Link]
exploring_data.pdf
NNSMP-National non-point source monitoring program (2011) Statistical analysis for
monotonic trends. Technical note 6
[Link]
_analysis.pdf
Null JJ (2013) El Niño and La Niña Years and Intensities: Based on Oceanic Niño Index
(ONI). Golden Gate Weather Services [[Link]
O‟connor TG (1985) A synthesis of Field Experiments Concerning the Grass Layer in the
Savanna Regions of Southern Africa. Report No 114, South African National Scientific
Programmes, Foundation for Research Development, Pretoria
Odada EO and Olago DO (eds.) (2002) The East African Great Lakes: Limnology,
Palaeolimnology and Biodiversity. Advances in Global Change Research, 12
Ortt RA, VanRyswick S, and Wells D (2008) Bathymetry and sediment accumulation of
Triadelphia and Rocky Gorge reservoirs. Watershed Services Department of Natural
Resources. US [[Link]
Overland JE, Percival DB, and Mofjeld HO (2006) Regime shifts and red noise in the North
Pacific. Deep Sea Research. Part I: Oceanographic Research Papers, 53 (4), pp. 582-
588

139
Partal T and Kahya E (2006) Trend analysis in Turkish precipitation data. Hydrol. Process,
20, pp. 2011-2026
Penman HL (1948) Natural evaporation from open water, bare soil, and grass. Proc. Roy. Soc.
London A193, pp. 120-146
Peterson WT and Schwing FB (2003) A new climate regime in Northeast pacific ecosystems.
Geophysical research letter, 30 (17), pp. 1-4
Peterson TC, Golubev VS, and Groisman PY (1995) Evaporation losing its
strength. Nature, 377, pp. 687-688
Pettitt AN (1979) A nonparametric approach to the change-point problem, Appl. Stat., 28, pp.
126-135
Piper BS, Plinston DT, and Sutcliffe JV (1986) The water balance of Lake Victoria.
Hydrological Sciences, 31(1), pp. 25-37
Pittock AB (1975) Climatic change and patterns of variation in Australian rainfall. Search, 6,
pp. 498-504
Potschin M (2009) Land use and the state of the natural environment. Land Use Policy 26S :
S170-S177
PSIAC (Pacific Southwest Inter-Agency Committee) (1968) Factors Affecting Sediment
Yield in the Pacific Southwest Area. Report of the Water Management Subcommittee
on Factors Affecting Sediment Yield in the Pacific Southwest Area and Selection and
Evaluation of Measures for Reduction of Erosion and Sediment Yield
[[Link]
Radziejewski M and Kundzewicz ZW (2004) Detectability of changes in hydrological records.
Hydrol. Sci. J., 49(1), pp. 39-51
Rao K, Mohan R, Abraha L, and Tibebu N (2011) Analyzing the impact of land useand land
cover change on ground water recharge using RS and GIS approach:
the case of Awassa Catchment, Ethiopia. The IUP Journal of Earth Sciences, 5 (2), pp.
46-56
Rasmusson EM and Wallace JM (1983) Meteorological aspects of the El Niño/Southern
Oscillation, Science, 222, pp. 1195-2002
Rayner DP (2007) Wind Run Changes: The Dominant Factor Affecting Pan Evaporation
Trends in Australia. J. Climate, 20, pp. 3379-3394
Republic of Kenya (2001) National Disaster Management Policy. Draft. Nairobi
[[Link]
Reynolds TW, Farley J, and Huber C (2010) Investing in human and natural capital: An
alternative paradigm for sustainable development in Awassa, Ethiopia. Ecological
Economics, pp. 2140-2150
Robinson AR (1977) Relationship between soil erosion and sediment delivery. IAHS Publ.,
122, pp. 159-167
Rodionov SN (2004) A sequential algorithm for testing climate regime shifts. Geophysical
Research Letters, 31
[[Link]
Rodionov SN (2005) A brief overview of the regime shift detection methods. In: Velikova V
and Chipev N (eds.) Large-Scale Disturbances (Regime Shifts) and Recovery in
Aquatic Ecosystems: Challenges for Management Toward Sustainability
[[Link] 

140
Ropelewski C (1999) The Great El Niño of 1997 and 1998: Impacts on Precipitation and
Temperature. A review assessment published in CONSEQUENCES-the nature and
implications of environmental changes, 5 (2), pp. 17-25
Rowntree KM (1983) Rainfall erosivity in Kenya: some preliminary considerations. In:
Thomas DB and SengaWM (eds) Soil and water conservation in Kenya. Institute of
Development Studies and Faculty of Agriculture, University of Nairobi Occasional
Paper 42, pp. 1-19
Salarijazi M, Akhond-Ali A, Adib A, and Daneshkhah A (2012) Trend and change-point
detection for the annual stream-flow series of the Karun River at the Ahvaz
hydrometric station. African Journal of Agricultural Research, 7(32), pp. 4540-4552
Scheffer M and Carpenter SR (2003) Catastrophic regime shifts in ecosystems: linking theory
to observation. Trends in Ecology and Evolution, 18 (12), pp. 648-656
Schertzer WM, Rouse WR, Lam D, Bonin D and Mortsch L (2012) Climate Variability and
Change - Lakes and Reservoirs. Environment Canada [[Link]
nwri/[Link]?lang=En&n=0CD66675-1&offset=17&toc=show
Schwing FB and Moore CS (2000) 1999-A year without summer for California or a harbinger
of a climate shift?. Eos Trans. AGU, 81, pp. 301, 304 - 305
Schwing FB, Murphre T, deWitt L, and Green PM (2002) The evolution of oceanic and
atmospheric anomalies in the Northeast Pacific during the El Niño and La Niña events
of 1995 - 2000, Prog. Oceanogr., 54, pp. 459 - 491
Schwing F, Moore C, Ralston S, and Sakuma KA (2000) Record coastal upwelling in the
California Current in 1999, Calif. Coop. Oceanic Fish. Invest. Rep., 41, pp. 148 - 160.
Sekovski I, Newton A, and Dennison WC (2012) Megacities in the coastal zone: Using a
driver-pressure-state-impact-response framework to address complex environmental
problems. Estuarine, Coastal and Shelf Science, 96, pp. 48-59
Sene KJ (1998) Effect of variations in net basin supply on lake levels and outflows.
Hydrological Processes, 12, pp. 559-573
Serban P and Askew AJ (1991) Hydrological forecasting and updating procedures. In: Van de
Ven FH, Gutknecht G, Loucks DP,and Salewicz KA (eds) Hydrology for the Water
Management of Large River Basins (Proc. Vienna Syrnp.), IAHS Publ. no. 201, pp.
357-369
Seyed H and Ali A (2008) Preparing the map of erosion hazard and sedimentation in Dez
watershed by remote sensing and GIS. Environmental Sciences, 6 (1), pp. 83-100
Shamo DA (2008) Assessment of non-point source pollution in Lake Hawassa watershed
using the annualized agricultural non-point sources (AnnAGNPS) model. [Link] thesis,
Addis Ababa University, Addis Ababa. Ethiopia
[[Link]
Shamsedin AY and O´Connor KM (1999) A real-time combination method for the outputs of
different rainfall-runoff models. Hydrological Sciences, 44 (6)
Shewangizaw D (2010) Assessing the effect of land use changes on the hydraulic regime of
Lake Hawassa. [Link] thesis, Addis Ababa University, Ethiopia
[[Link]
Shewangizaw D and Michael Y (2010) Assessing the Effect of Land Use Change on the
Hydraulic Regime of Lake Awassa. Nile Basin Water Science and Engineering
Journal, 3 (2)
Shiferaw B and Holden S (2001) Farm-level benefits to investments for mitigating land
degradation: empirical evidence from Ethiopia. Environment and Development
Economics, 6, pp. 335-358
141
Spangenberg JH, Pfahl S, and Deller K (2002) Towards indicators for institutional
sustainability: lessons from an analysis of Agenda 21. Ecological Indicators, 2, pp. 61-
77
Sponberg K (1999) Navigating the numbers of climatological impact. Compendium of
Climatological Impacts, University Corporation for Atmospheric Research,1, National
Oceanic and Atmospheric Administration, Office of Global Programs, 13pp
Steele JH (1996) Regime shifts in fisheries management. Fish Res. 25, pp. 19-23
Stocking MA and Elwell HA (1973) Soil erosion hazard in Rhodesia. Rhodesian Agricultural
Journal, 70, pp. 93-101
Stocking MA and Elwell HA (1976) Rainfall erosivity over Rhodesia. Transactions of the
Institute of British Geographers New Series 1, pp. 231-45
Strand RI and Penibertori EL (1982) Reservoir sedimentation: Technical Guideline for Bureau
of Reclamation
[[Link]
nTechGuide10_1982.pdf
Street FA (1979) Late Quaternary Lakes in the Ziway-Shala Basin, Southern Ethiopia,
unpublished PhD thesis, Cambridge University, Cambridge, UK, 457pp
Stroosnijder L and Eppink LA (1993) Principles of soil and water conservation. Lecture notes
of course K200- 500/510, WAU, Wageningen, The Netherlands
Strub PT and James C (2002) Altimeter-Derived Surface Circulation in the Large-Scale NE
Pacific Gyres. Part 2. 1997-1998 El Niño anomalies. Progress in Oceanography, 53,
pp. 185-214
Szesztay K (1974) Water balance and water level fluctuation of lakes. Hydrological sciences-
Bulletin, 19 (1), pp. 73-84
Tabari H and Marofi S (2011) Changes of Pan Evaporation in the West of Iran. Water
Resources Management, 25, pp. 97-111
Tabari H and Talaee PH (2011) Analysis of trends in temperature data in arid and semi-arid
regions of Iran. Global and Planetary Change, 79 (1-2), pp. 1-10
Tabari H, Somee BS, and Zadeh MR (1011) Testing for long-term trends in climatic variables
in Iran. Atmospheric Research, 100, pp. 132-140
Telford RJ (2000) Diatom stratigraphies of lakes Awassa and Tilo, Ethiopia: Holocene records
of groundwater variability and climate change. Unpublished PhD thesis, University of
Wales, UK
Telford RJ, Lamb HF, and Mohammed U (1999) Diatom-derived palaeoconductivity estimates
for Lake Awassa, Ethiopia: evidence for pulsed inflows of saline groundwater. Journal
of Paleolimnology, 21, pp. 409-422
Terasmaa J (2011) Lake basin development in the Holocene and its impact on the
sedimentation dynamics in a small lake (Southern Estonia). Estonian Journal of
Earth Sciences, 60 (3), pp 159-171
Tereshchenko I, Filonov A, Gallegos A, Monzo´n C, and Rodrı´guez R (2002) El Niño 1997-
98 and the hydrometeorological variability of Chapala, a shallow tropical lake in
Mexico. Journal of Hydrology, 264, pp. 133-146
Tesfaye C (1982) Hydrogeology of the Lakes Region, Ethiopia. Ministry of Mines and
Energy, Addis Ababa. Unpublished document
Tessema Z (1998) Hydrochemical and water balance approach in the study of high water level
rise of Lake Beseka. Unpublished [Link] thesis, University of Birmingham, UK, 90pp.
Thomson RE and Emery WJ (2001) Time-series Analysis Methods: Data Analysis Methods in
Physical Oceanography. Second and revised edition. Elsevier Science
142
Tiercelin JJ, Mondeguer A, Gasse F, Hillaire-Marcel C, Hoffert M, Larque P, Ledee V,
Marestang P, Ravenne C, Raynaud JF, Thouveny N, Vincens A, and Williamson D
(1988) 25,000 ans d‟histoire hydrologique et se´dimentaire du lac Tanganyika, Rift
Est-Africain. C. R. Acad. Sci. Paris, 307 (II), pp. 1375- 1382
Trenberth KE and Hoar TJ (1997) El Niño and climate change. Geophys. Res. Letters, 24, pp.
3057-3060
Tscherning K, Helming K, Krippner B, Sergio S, and Paloma G (2011) Does research
applying the DPSIR framework support decision making?. Land Use Policy, 29, pp.
102-110
TSOA (1995) Texas State Auditor's Office. Data Analysis: Analyzing Data - Trend Analysis-
2. Methodology Manual, rev. 5/95 [[Link]
Tu M, Hall MJ, De Laat PJ, and De Wit MJ (2004) Detection of long-term changes in
precipitation and discharge in the Meuse basin. GIS and Remote Sensing in Hydrology,
Water Resources and Environment (Proceedings of 1CGRHWE held at the Three
Gorges Dam. China, September 2003). IAHS Publ., 289 (169)
Turner RK, Lorenzoni I, Beaumont N, Bateman IJ, Langford IH, and McDonald AL (1998)
Coastal management for sustainable development: analyzing environmental and
socioeconomic change on the UK coast. Geogr J., 164, pp. 269-281
Uhlenbrook S (2009) Climate and man-made changes and their impacts on catchments, in:
Water Policy 2009, Water as Vulnerable and Exhaustible Resources, In: Kovar, P.,
Maca, P., and Redinova, J (eds). Proceedings of the Joint Conference of APLU and
ICA, Prague, Czech Republic, pp. 81-87
UNEP (2008) GEO Resource books: a training manuals on integrated environmental
assessment and reporting [[Link]
USDA (U.S. Department of Agriculture (2007) National Engineering Handbook, title 210-VI.
Part 630, Chapter 7. Washington, DC [Link]
Vallet-Coulomb C, Legesse D, Gasse G, Travi Y, and Chernet T (2001) Lake evaporation
estimates in tropical Africa (Lake Ziway, Ethiopia). Journal of Hydrology, 245, pp. 1-
17
Vandegrift TR and Stefan HG (2010) Annual Stream Runoff and Climate in Minnesota‟s
River Basins. Project Report No. 543. Minneapolis, Minnesota
[Link]
Vázquez JF (2003) A methodology for policy analysis in water resources management .
Fondazione Eni Enrico Mattei, Italy [Link]
Vente J and Poesen J (2005) Predicting soil erosion and sediment yield at the basin scale:
Scale issues and semi-quantitative models. Earth-Science Reviews, 71, pp. 95-125
Vente J (2009) Soil Erosion and Sediment Yield in Mediterranean Geoecosystems: Scale
issues, modelling and understanding. PhD thesis. Katholieke Universiteit Leuven
Faculteit Wetenschappen Departement Aard- en Omgevingswetenschappen
[[Link]
mediterranean-geoecosystems
Vilalta ER (2010) Water resources management in the central Rift Valley of Ethiopia.
Published MSc thesis in Civil Engineering. Barcelona, Spain
Villarini G, Smith JA, Serinaldi F, and Ntelekos AA (2011). Analyses of seasonal and annual
maximum daily discharge records for central Europe. J. Hydrol., 399, pp. 299-312
Von Storch H and Zwiers FW (1999) Statistical Analysis in Climate Research, ISBN 0 521
45071

143
Vrieling A (2006) Satellite remote sensing for water erosion assessment: A review. Catena, 65
(1), pp. 2-18
Vrieling A, Sterk G, and Beaulieu N (2002) Erosion risk mapping: A methodological case
study in the Colombian Eastern Plains. Journal of Soil and Water Conservation, 57 (3),
pp. 158-163
Vrugt JA, Diks CGH, Gupta HV, Bouten W, and Verstraten JM (2005) Improved treatment of
uncertainty in hydrologic modeling: Combining the strengths of global optimization
and data assimilation, Water Resour. Res., 41, W01017
Vuglinskiy V (2009) Water level in lakes and reservoirs, water storage. Global Terrestrial
Observing System, Rome [[Link]
Wagesho N, Goel NK, and Jain MK (2012) Investigation of non-stationarity in hydro-climatic
variables at Rift Valley lakes basin of [Link] of Hydrology, 444-445, pp.
113-133
Wagner G, Schröder HG, and Gurtz J (2002) A model approach for in- and outflow
calculation of Upper Lake Constance: an investigation of a 60 year time span and
observations about the flood of 1999. Limnologica 32, pp. 27-32
Wallace JM and Gutzler DS (1981). Teleconnections in the geopotential height field during
the Northern Hemisphere winter. Mon Weather Rev., 109, pp. 784-812
Warburton H and Martin A (1999). Local people‟s knowledge in natural resources research.
Socio-economic Methodologies for Natural Resources research. Chatham, UK: Natural
Resources Institute [[Link]
Westmacott JR and Burn DH (1997) Climate change effects on the hydrologic regime within
the Churchill-Nelson River Basin. Journal of Hydrology, 202, pp. 263-279
Wijngaard J, Klein T, and Können GP (2003) Homogeneity of 20th century European daily
temperature and precipitation series. Int. J. Climatol. 23 (6), pp. 679-692
Williams MA, Williams FM, and Bishop PM (1981) Late Quaternary history of Lake Beseka,
Ethiopia. Palaeoecology of Africa and the Surrounding Islands and Antarctica,13, pp.
93-104
Winter TC (1981) Uncertainties in estimating the water balance of lakes. Water resources
bulletin: American water resources association, 17 (1)
Wischmeier WH and Smith DD (1978) Predicting Rainfall Erosion Losses: A Guide to
Conservation Planning. Agriculture Handbook No. 537. USDA/Science and Education
Administration, US. Govt. Printing Office, Washington, DC. 58pp.
[[Link]
Wolde-Georgis T, Aweke D, and Hagos Y (2000) Reducing the impacts of environmental
emergencies through early warning and preparedness: The Case of the 1997-98 El
Niño [[Link] 
Wolff C, Haug GH, Timmermann A, Jaap S, Damsté S, Brauer A, Sigman DM, Cane AM, and
Verschuren D (2011) Reduced inter-annual rainfall variability in East Africa during the
last ice age. Science, 333 (6043), pp. 743-747
Wood RB and Talling JF (1988) Chemical and algal relationship in a salinity series of
Ethiopian inland waters. Hydrobiologia 158, pp. 29-67
World Bank (2012) World Bank indicators-Ethiopia-Density and Urbanization
[ttp://[Link]/ethiopia/rural-population-percent-of-total-population-wb-
[Link]

144
WRDB (Water Resources Development Bureau) (2007) Study of Pollution of Lakes and
rivers. AG consulting hydrogeologists and engineers Plc. Addis Ababa. Ethiopia.
Unpublished document
WWDSE (Water Works Design and Supervision Enterprise) (1999) Study of Lake Beseka,
Ministry of Water Resources, Addis Ababa, Ethiopia, Main Report, Vol 1, pp. 203.
Unpublished document
WWDSE (Water Works Design and Supervision Enterprise) (2001) The study of Lake
Awassa level rise. Southern Nations Nationalities and Peoples Regional State. Water,
Mines and Energy Resources Development Bureau. Addis Ababa, Ethiopia.
Unpublished document
Xu J (2011) Variation in annual runoff of the Wudinghe River as influenced by climate change
and human activity, Quatern. Int., 244, pp. 230-237
Yang Z, Zhou Y, Wenninger J, and Uhlenbrook S (2012) The causes of flow regime shifts in
the semi-arid Hailiutu River, Northwest [Link]. Earth Syst. Sci., 16, pp. 87-103
Yirgu G, Paola D, Kebede GM, Hagos F, Teferi MA, Tilahun G, Gemetesa N (1997) Study
report on ground fracturing in the Muleti area, Awasa Zuria Woreda. Addis Ababa
University and SNNPR Professional Paper, Awasa, 7pp. Unpublished document
Yletyinen J, Blenckner T, and Biggs R (2012) North Pacific Ocean. In: Regime Shifts
Database [Link]
Yu F, Chen Z, Ren X, and Yang G (2009) Analysis of historical floods on the Yangtze River,
China: Characteristics and explanations. Geomorphology, 113, pp. 210-216.
Ze'ev G and Smith DJ (2001) Interdecadal climate variability and regime-scale shifts in Pacific
North America. Geophysical research letters, 28 (8), pp. 1515-1518
Zemedagegnehu E and Egizabher R (2004) Determination for the cause of rising water levels
in Lake Beseka and design of remedial measure. Ethiopian Engineers Association
Journal, August issues. Unpublished document
Zhao FF, Xu ZX, Zhang L, and Zuo DP (2009) Streamflow response to climate variability and
human activities in the upper catchment of the Yellow River Basin, Science in China
Series E: Technological Sciences, 52, pp. 3249-3256
Zohary T and Ostrovsky I (2011) Ecological impacts of excessive water level fluctuations in
stratified freshwater lakes. Inland Waters, 1, pp. 47-59

145
 Acknowledgement

More people lend me their hands than I can possibly thank, but let me start with those who
have fully involved in the work over the last 3.5 years. I would like to thank ecbp (engineering
capacity building program) for funding me to pursue my PhD. The prime mover of the entire
work was Prof. Dr. Bernd Diekkrügger who showed me how to turn challenges into
opportunities. Prof. Dr.-ing. Jackson Roehrig is also an equivalent contributor of my work in
day-to-day supervision and purchasing of high tech equipments for my field work. Two of my
friends, Assegid Cherinet and Yonas Girma contributed much during the hydrographic survey
from which we benefited mutually. I would also like to express my appreciation to Mr.
Engidaw Zemedagegnehu from Water Works and Supervision Enterprise who provided me the
original bathymetry map. Ministry of water resources and meteorological agency also deserve
my thanks for their provision of time series of hydro-climatic data. My wife, Simegn Asmare,
played the biggest role by accomplishing dual assignments as a mother and a father for my
kids while I was abroad. I would like to wind up by praising the almighty God blessed me and
all around.

146
 Appendices

Appendix 1: (A) Oceanic Nino Index (ONI) Source: Null (2012) (accessed in September 2012)

(B) El Niño and La Niña Years and Intensities: Based on Oceanic Niño Index (ONI)
El Niño La Niña
Weak Moderate Strong Weak Moderate Strong
1969 1968 1972 1971 1970 1973
1976 1986 1982 1974 1998 1975
1977 1987 1991 1983 2007 1988
2004 1994 1997 1984 1999
2006 2002 1995 2010
2009 2000
2005
2011

147
Appendix 2: RSI result of annual average lake level
Av. Lake Depth (m) RSI Mean Weighed Length P Outliers
1970 19.64 0.00 19.70 19.68 8
1971 19.83 0.00 19.70 19.68 8
1972 20.26 0.00 19.70 19.68 8 0.76
1973 19.92 0.00 19.70 19.68 8
1974 19.53 0.00 19.70 19.68 8
1975 19.37 0.00 19.70 19.68 8
1976 19.36 0.00 19.70 19.68 8
1977 19.67 0.00 19.70 19.68 8
1978 20.40 0.63 20.12 20.10 9 0.02
1979 20.74 0.00 20.12 20.10 9 0.68
1980 20.33 0.00 20.12 20.10 9
1981 19.86 0.00 20.12 20.10 9
1982 19.72 0.00 20.12 20.10 9
1983 20.15 0.00 20.12 20.10 9
1984 20.22 0.00 20.12 20.10 9
1985 19.77 0.00 20.12 20.10 9
1986 19.90 0.00 20.12 20.10 9
1987 20.27 0.88 20.43 20.43 6 0.06
1988 20.32 0.00 20.43 20.43 6
1989 20.71 0.00 20.43 20.43 6
1990 20.78 0.00 20.43 20.43 6
1991 20.33 0.00 20.43 20.43 6
1992 20.17 0.00 20.43 20.43 6
1993 20.72 1.10 21.20 21.16 10 0.001
1994 20.70 0.00 21.20 21.16 10 0.95
1995 20.53 0.00 21.20 21.16 10 0.69
1996 20.92 0.00 21.20 21.16 10
1997 21.33 0.00 21.20 21.16 10
1998 21.98 0.00 21.20 21.16 10 0.53
1999 21.90 0.00 21.20 21.16 10 0.59
2000 21.25 0.00 21.20 21.16 10
2001 21.31 0.00 21.20 21.16 10
2002 21.37 0.00 21.20 21.16 10
2003 20.87 -0.29 21.00 20.99 8 0.34
2004 20.58 0.00 21.00 20.99 8
2005 20.56 0.00 21.00 20.99 8
2006 20.72 0.00 21.00 20.99 8
2007 21.46 0.00 21.00 20.99 8 0.93
2008 21.47 0.00 21.00 20.99 8 0.92
2009 21.19 0.00 21.00 20.99 8
2010 21.16 0.00 21.00 20.99 8
RSI: Regime Shift Index
Mean: Equal-weighed arithmetic means of the regimes
Weighed: Weighed means of the regimes using the Huber's weight function with the parameter = 1
Length: Length of the regimes
P: Significance level of the difference between the mean values of the neighboring regimes based on the Student's two-tailed
t-test with unequal variance (TTEST procedure in Excel)

148
Appendix 3: RSI result of annualtotal rainfall at Hawassa station
Annual rainfall at Hawassa Met Station (mm) RSI Mean Weighed Length P Outliers
1973 726 0 830 830 3
1974 937 0 830 830 3
1975 826 0 830 830 3
1976 954 0.18 954 954 1
1977 1226** 0.15 1130 1130 2
1978 1033 0 1130 1130 2
1979 968 -0.21 881 881 2
1980 794 0 881 881 2
1981 1040 0.047 1064 1064 3
1982 992 0 1064 1064 3
1983 1160 0 1064 1064 3
1984 725 -0.29 813 813 2
1985 902 0 813 813 2
1986 1194 0.13 948 942 9 0.55
1987 955 0 948 942 9
1988 957 0 948 942 9
1989 1025 0 948 942 9
1990 751 0 948 942 9 0.73
1991 889 0 948 942 9
1992 975 0 948 942 9
1993 928 0 948 942 9
1994 861 0 948 942 9
1995 1004 0.03 1099 1099 4
1996 1189 0 1099 1099 4
1997 1055 0 1099 1099 4
1998 1146 0 1099 1099 4
1999 810 -0.12 882 882 6
2000 822 0 882 882 6
2001 1022 0 882 882 6 1.00
2002 920 0 882 882 6
2003 821 0 882 882 6
2004 896 0 882 882 6
2005 998 0.43 1002 1025 6
2006 1198 0 1002 1025 6 0.81
2007 1157 0 1002 1025 6
2008 915 0 1002 1025 6
2009 704 0 1002 1025 6 0.44
2010 1039 0 1002 1025 6
**Highest annual rainfall in the period

149
Appendix 4: RSI result of Mean annual Stream flow at Tikur Wuha (m3/s)
Mean annual Stream flow (m3/s) RSI Mean Weighed Length P Outliers
1980 0.76 0 1.89 1.95 7 0.38
1981 2.09 0 1.89 1.95 7
1982 2.14 0 1.89 1.95 7
1983 2.69 0 1.89 1.95 7 0.61
1984 1.88 0 1.89 1.95 7
1985 1.71 0 1.89 1.95 7
1986 1.945 0 1.89 1.95 7
1987 2.18 1.12 2.39 2.39 3 0.08
1988 2.48 0 2.39 2.39 3
1989 2.5 0 2.39 2.39 3
1990 2.9 0.51 2.90 2.90 1
1991 2.6 -0.08 2.65 2.65 4
1992 2.55 0 2.65 2.65 4
1993 2.85 0 2.65 2.65 4
1994 2.6 0 2.65 2.65 4
1995 2.92 0.74 3.12 3.11 3
1996 3.6 0 3.12 3.11 3 0.91
1997 2.84 0 3.12 3.11 3
1998 3.86 0.04 3.39 3.39 9 0.96
1999 4.02 0 3.39 3.39 9 0.72
2000 2.66 0 3.39 3.39 9 0.61
2001 3.12 0 3.39 3.39 9
2002 3.42 0 3.39 3.39 9
2003 3.535 0 3.39 3.39 9
2004 3.65 0 3.39 3.39 9
2005 3.2 0 3.39 3.39 9
2006 3 0 3.39 3.39 9

Appendix 5: RSI result of annual runoff coefficient

Runoff coefficient of Tikur Wuha sub-watershed (C) RSI Mean Weighed Length P Outliers
1981 0.1 0 0.11 0.11 7
1982 0.1 0 0.11 0.11 7
1983 0.14 0 0.11 0.11 7 0.84
1984 0.1 0 0.11 0.11 7
1985 0.09 0 0.11 0.11 7
1986 0.12 0 0.11 0.11 7
1987 0.1 0 0.11 0.11 7
1988 0.13 0.85 0.13 0.13 2 0.01
1989 0.13 0 0.13 0.13 2
1990 0.17 0.31 0.17 0.17 2 0.09
1991 0.16 0 0.17 0.17 2
1992 0.12 -0.16 0.14 0.14 6 0.03
1993 0.14 0 0.14 0.14 6
1994 0.14 0 0.14 0.14 6
1995 0.15 0 0.14 0.14 6
1996 0.15 0 0.14 0.14 6
1997 0.13 0 0.14 0.14 6
1998 0.17 0.88 0.18 0.18 8 0.01
1999 0.24 0 0.18 0.18 8 0.48
2000 0.14 0 0.18 0.18 8 0.70
2001 0.15 0 0.18 0.18 8 0.93
2002 0.2 0 0.18 0.18 8
2003 0.2 0 0.18 0.18 8
2004 0.2 0 0.18 0.18 8
2005 0.16 0 0.18 0.18 8

150
Appendix 6: RSI result of annual total lake evaporation as computed from pan-evaporation (mm)
evaporation RSI Mean Weighed Length P Outliers
1986 1720 0 1717 1717 2
1987 1713 0 1717 1717 2
1988 1647 -0.29 1659 1659 8 0.08
1989 1554 0 1659 1659 8
1990 1625 0 1659 1659 8
1991 1606 0 1659 1659 8
1992 1682 0 1659 1659 8
1993 1696 0 1659 1659 8
1994 1818 0 1659 1659 8 0.97
1995 1645 0 1659 1659 8
1996 1298 -1.83 1329 1329 5 0.00001
1997 1343 0 1329 1329 5
1998 1236 0 1329 1329 5
1999 1385 0 1329 1329 5
2000 1382 0 1329 1329 5
2001 1258 -0.18 1295 1295 7 0.36
2002 1318 0 1295 1295 7
2003 1376 0 1295 1295 7
2004 1301 0 1295 1295 7
2005 1331 0 1295 1295 7
2006 1192 0 1295 1295 7
2007 1285 0 1295 1295 7

151