0 0 ( 0 ‘^-5

Bonner Bodenkundliche Abhandlungen Band 37

Mahdi Osman

Rainfall and its Erosivity in Ethiopia with special consideration of the central highlands

Rheinische Friedrich-Wilhelms-Universitat Bonn 2001 Herausgeber: G.W. Briimmer, A. Skowronek Schriftleitung: Christine Klein Vertrieb: Institut fur Bodenkunde NuBallee 13, 53115 Bonn Tel. 0228-732780/81 Fax. 0 2 2 8 -732782

ISSN 0939 - 7809 (D 98) Bonner Bodenkundl. Abh. 37 249 S., 87 Abb., 27 Tab. Bonn 2001

Mahdi Osman

Rainfall and its Erosivity in Ethiopia with special consideration of the central highlands

Institut fur Bodenkunde Bonn 2001 ISSN 0939 - 7809 (D 98) Inaugural-Dissertation zur Erlangung des Grades Doktor der Agrarwissenschaften (Dr. agr.) der Hohen Landwirtschaftlichen Fakultat der Rheinischen Friedrich-Wilhelms-Universitat zu Bonn

vorgelegt von Mahdi Osman aus Addis Ababa, Athiopien

Referent: Prof. Dr. A. Skowronek Korreferentin: Frau Dr. P. Sauerborn Korreferent: Priv.-Doz. Dr.-Ing. A. Rieser Tag der miindlichen Priifung: 14.02.2001

Osman, M.

Rainfall and its Erosivity in Ethiopia with special consideration of the central highlands

Bonner Bodenkundliche Abhandlungen Bd. 37, 249 S. (2001)

Druck: Martin Roesberg, 53347 Witterschlick

iv

M a h d i O s m a n : R a in f a l l a n d its E r o s iv it y in E t h io p ia w it h s p e c ia l consideration o f t h e C e n t r a l H ig h l a n d s Bonn (2001) pp. 249, Faculty of Agriculture, Diss., accepted on 14.02.2001

ABSTRACT In this research project the characteristics and the erosive potential of precipitation in Ethiopia were examined, with particular emphasis on the central highlands of the country. Furthermore, the hydrologic implications of rainfall variability and socio-economic context of water erosion were studied. Long-term rainfall data spanning from 1898 to 1997 of 168 weather stations were statistically analysed. A detailed assessment of precipitation and its erosive potential in the central highlands was made using data of 44 selected weather stations. The erosivity of rainfall was estimated using the modified Fournier’s index. The consequence of rainfall variability on flow regime was examined based on hydrometric records ranging from 1982 to 1997 of 5 gauging stations. The socio-economic context of water erosion was assessed by examining previous project documents, supplemented by socio-economic and field survey. Generally, the summer is the major rainy season in most parts of Ethiopia. Two types of seasonal precipitation distribution patterns, namely single and double peaked were noted. The latter was further subdivided into two groups: One with a short gap between the two peaks and another with a long gap. The lowest rainfall variability in the country was noted in April and the highest in June. With regard to precipitation in the central highlands, the area has seen periods of extremely wet and drought events. On the whole, the region has experienced high positive deviations from the long-term average in the first half of the 20lh century, but dominantly negative departures have been evident since then. However, a general decline in precipitation was noted in the area. In the central highlands, highly erosive rainfalls were observed in the first half of the 20,h century, compared with the succeeding periods. Concerning prediction tools, a linear regression model was found to be the best predictor of regional water erosion. In the central highlands of Ethiopia, long-term rainfall erosivity tended to increase from east to west and from south to north. Generally, the summer precipitation was found to be the major agent of water erosion in the study area. The erosive potential of rainfall varies temporally and spatially. Hence, various prediction models were developed for the weather stations in the study area for different seasons. Six standard classes of annual and summer rainfall erosivity, namely very low, low, moderate, high, very high and extremely high were established for the region. Regarding the implications of rainfall variability on flow regime, it was noted that flow is affected by the variability in areal precipitation. However, further research is needed to determine unexplained factors influencing the flow regime. Concerning the socio-economic context of water erosion, Ethiopian farmers have long­ standing indigenous knowledge of land and water management techniques. These became less effective with intensification of land-use. The compensatory interventions were neither socially acceptable nor environmentally useful. Today, socio-economic and environmental sustainability are gaining attention in resource management policy design and implementation. Social participation in land and water management should be encouraged, especially through defined and clear property rights. The database established by this study is a useful resource for environmental management planning. The models developed are appropriate to forecast water erosion under conditions of data scarcity and financial constraints. It is recommended that the database and its management be regularly improved for appropriate land and water management planning. Future research should focus on testing and adaptation of modem water erosion prediction technologies as well as models to devise the best predictive tools with maximum efficiency. Watershed and river basin approach to curb rainfall erosion, sedimentation assessment, and promotion of indigenous knowledge should gain much focus in land and water management research in Ethiopia.

Rainfall and its erosivity in Ethiopia M a h d i O s m a n : N iederschlag und R egenerosivitat in A t h io p ie n u n t e r b e s o n d e r e r B erucksichtigung d es z e n t r a l e n H o c h l a n d e s Bonn (2001) 249 S., Landw. F„ Diss. v. 14.02.2001

KURZFASSUNG In der vorliegenden Arbeit wurden die Charakteristik und das Erosionspotential der Niederschliige in Athiopien untersucht unter besonderer Berucksichtigung des zentralen Hochlandes. Dariiber hinaus wurden die hydrologische Bedeutung von Niederschlagsvariabilitat sowie sozio-okonomische Zusammenhange von Wassererosion untersucht. Insgesamt wurden langjahrige Niederschlagsdaten von 168 Wetterstationen in Athiopien fur den Zeitraum 1898 bis 1997 statistisch analysiert. Fur eine detaillierte Untersuchung des zentralen Hochlandes wurden Daten von 44 ausgewahlten Wetterstationen herangezogen. Als KenngroBe der Regenerosivitat fand der modifizierte Foumier-Index Anwendung. Die Auswirkung der Niederschlagsvariabilitat auf das AbfluBregime wurde mittels hydrometrischer Daten von 5 MeBstationen zwischen 1982 und 1997 untersucht. Die sozio-okonomische Bedeutung der Wassererosion wurde an Hand von Projektunterlagen mit empirischen Untersuchungen und Gelandestudien erarbeitet. 1m iiberwiegenden Teil Athiopiens ist der Sommer die Hauptregensaison. Es lassen sich zwei Typen saisonaler Niederschlagsverteilungen feststellen: mit einem und mit zwei Maxima, wovon der zweite (Typus) sich in Untergruppen mit je einer langen bzw. kurzen Phase zwischen den beiden Hochstwerten untcrteilen laBt. Die Niederschlagsvariabilitat ist im April am niedrigsten und im Juni am hochsten. Im zentralen Hochland gibt es abwechselnd Phasen groBer Dtirre und extremer Feuchte. In Athiopien kann fur die erste Halfte des 20. Jahrhunderts eine starke positive Abweichung vom langjahrigen Niederschlagstrend festgehalten werden; danach jedoch zeigt sich eine negative. Im zentralen Hochland ergibt sich ein abnehmender Niederschlagstrend. Es konnte ein lineares Regressionsmodell zur Vorhersage der regionalen Niederschlagserosion festgehalten werden. Die langjahrige Regenerosivitat steigt von Ost nach West und von Stid nach Nord an. Generell bilden die Sommerwerte den wichtigsten Einflussfaktor der Wassererosion im Untersuchungszeitraum. Da die Variabilitat saisonal und raumlich schwankt, wurden fur die einzelnen Wetterstationen verschiedene Vorhersagemodelle fur unterschiedliche Jahreszeiten entwickelt. Sechs Gefalirenstufen des jahrlichen und sommerlichen Regenerosionspotentials konnen abgegrenzt werden: sehr niedrig, niedrig, mittel, hoch, sehr hoch und extrem hoch. Die Variabilitat der Niederschlage im Untersuchungsraum wirkt sich auch auf das AbfluBregime aus. Zur Feststellung der noch unerforschten Einflussfaktoren wie z. B. Bodennutzung sind weiterfuhrende Untersuchungen notig. Athiopische Farmer haben eine gute Kenntnis traditioneller Land- und Wasser- Managementtechniken, die mit der Intensivierung der Landnutzung jedoch ihre Wirkung verloren. Anfangliche AusgleichsmaBnahmen waren weder sozial akzeptabel noch umweltvertraglich. Heute jedoch wird Ressourcenmanagement der sozi-okonomischen und der okologischen Nachhaltigkeit eine groBe Bedeutung beigemessen. Eine Beteiligung der Betroffenen sollte hier gefordert werden, insbesondere durch eine klare Definition des Eigentumsrechts. Die in dieser Studie entwickelte Datenbasis stellt eine wichtige Ressource fur das Umweltmanagement dar. Die hier erarbeiteten Modelle ermoglichen eine vereinfachte Vorhersage der Wassererosion auf der Basis begrenzter Daten und geringer finanzieller Kapazitaten. Es kann eine Verbesserung der Datenbasis und ihres Managements fur die Planungen des Wasser- und Landmanagements empfohlen werden. Kunftig sollten zunachst die modemen Technologien zum Schutz vor Wassererosion sowie die Erosionsmodelle getestet und angepasst werden, um Vorhersageinstrumente mit maximaler Effizienz zu erhalten. In der kunftigen Forschung des Land- und Wassermanagements sollten Wassereinzugsgebiete und Flussbecken integriert werden. Die Forderung des einheimischen Wissens sollte in der Land- und Wassermanagementforschung Athiopiens stark beriicksichtigt werden.

Rainfall and its erosivity in Ethiopia vi

M a h d i O s m a n : P l u v io s it e e t E r o s io n - e t u d e p o r t \ n t en p a r t i c u l a r s u r l e s P l a t e a u x C e n t r a ux E t h io p e e n s Bonn (2001) pp. 249, Faculte d'agriculture, these de doctorat. acceptee le 14.02.2001

RESUME Ce present travail a examine le caractere et le potentiel erosif des precipitations en Ethiopie, en particulier ceux des hauts plateaux du pays. De plus, l’etude a porte sur les implications hydrologiques de la variability pluvieuse et sur l’erosion due aux eaux dans le contexte socio- economique ethiopien. La pluviosite a long terme a ete analysee statistiquement a partir de donnees recueillies entre 1898 et 1997 par 168 stations meteorologiques. L'evaluation detaillee des precipitations et de leur potentiel erosif dans les hauts plateaux centraux s’est appuyee sur les donnees foumies par 44 stations meteorologiques selectionnees. L’erosivite due aux precipitations a ete estimee selon 1’index modifie de Foumier. Des enregistrements hydrometriques effectues entre 1982 et 1997 ont permis d’estimer les consequences de la variabilite pluvieuse sur le regime d'ecoulement. L’impact socio-economique de l'erosion diluvienne a ete evalue sur la base d’etudes anterieures; un complement d’information a ete foumi par des etudes socio-economiques et des recherches sur le terrain. Dans la plupart des regions de l'Ethiopie l'ete est la principale saison des pluies. Deux modeles de distribution des precipitations saisonnieres ont ete distingues. Le premier se caracterisant par un seul et le second par deux points culminants. Ce dernier se subdivise en deux types se distinguant par la duree de la periode qui separe les deux points culminants. La variation des precipitations se revelent plus basse en avril et plus haute en juin. Concemant la pluviosite dans les hauts plateaux centraux. la region a connu des periodes d'humidite et de secheresse extremes. Dans 1’ensemble, la region a vecu des ecarts positifs eieves par rapport a la moyenne a long terme pendant la premiere moitie du 20imc siecle, mais depuis lors, des ecarts negatifs predominent. On constate une diminution generalisee des precipitations dans la region. Par comparaison aux periodes suivantes. c'est pendant la premiere moitie du 20imc siecle qu’on a observe, dans les hauts plateaux centraux, des chutes de pluies hautement erosives. Le meilleur instrument de prevision d’erosion regionale due aux pluies s’est avere etre un modele lineaire regressif. Dans les hauts plateaux de l'Ethiopie. l'erosion diluvienne a long terme, a tendance a augmenter d’est en ouest et du sud au nord. Generaiement, les precipitations estivales se sont averees etre la cause principale de l'erosion dans la region examinee. Le potentiel erosif des pluies est variable dans le temps et dans 1’espace. On a done developpe, selon les saisons. des modeles variables de prevision pour les diverses stations meteorologiques dans la region. Six echelons standardises d’erosion due aux precipitations annuelles et estivales allant de tres bas, bas, modere, eleve, tres eleve a extremement eleve ont ete definis pour la region. Ceci a permis de mieux evaluer le potentiel de Terosion diluvienne. On a pu constater que le regime d’ecoulement est influence par la variabilite regionale des precipitations. Toutefois ces facteurs restent jusqu'alors inexpliques. Le contexte socio-economique: en matiere de gestion des eaux et des sols, les paysans ethiopiens possedent une experience indigene de longe date; cependant a cause d’une exploitation de plus en plus accrue des terres. ces techniques se sont revelees inefficaces car les interventions compensatoires n'etaient pas conciliables avec les conditions sociales et environnementales. Aujourd’hui. 1’elaboration et l’appiication d’une politique de gestion des ressources tient davantage compte d'un developpement durable du point de vue socio-economique et environnemental. La base de donnees etablies par cette etude constitue une ressource utile pour la planification de la gestion de l’environnement. Les modeles developpes conviennent a la prevision de l’erosion diluvienne dans des circonstances des donnees et des moyens financiers restrictives. Une gestion et une amelioration regulieres de la base des donnees sont recommandees en vue d'une planification appropriee de la gestion des sols et des eaux. Dans le future Taccent devra porter sur Foptimisation des technologies modemes preventives. En outre la prise en compte des lignes de partage des eaux et des bassins fluviaux ainsi que du savoir-faire indigene ethiopien en matiere de gestion des terres et des eaux seront egalement necessaire pour refrener l’erosion.

Rainfall and its erosivity in Ethiopia TABLE OF CONTENTS PAGE

LIST OF FIGURES...... xi LIST OF TABLES...... xvii ABBREVIATIONS...... xix

1 INTRODUCTION...... 1 1. 1 Background and statement of the problem...... 1 1.2 Objectives of the study...... 6

2 MATERIALS AND METHODS...... 7 2. 1 The Ethiopian and the central highlands...... 7 2. 2 Rainfall data set and data analysis...... 9 2. 2. 1 Concepts of rainfall erosivity...... 9 2.2.2 The data set...... 11 2. 2. 3 Data entry and statistical analysis...... 15 2. 2. 4 Regression analysis and isoerodent contours...... 16

3 PRESSURE PATTERNS, PRECIPITATION AND RAINFALL EROSIVITY IN ETHIOPIA...... 17 3. 1 Pressure patterns over Ethiopia...... 17 3. 2 Characteristics of long-term rainfall in Ethiopia...... 23 3. 3 Estimated values of long-term rainfall erosivity for Ethiopia...... 29 3. 4 Rainfall erosivity factors for the rainfall pattern regions of Ethiopia...... 30

4 LONG-TERM RAINFALL VARIABILITY IN THE CENTRAL HIGHLANDS OF ETHIOPIA...... 38 4. 1 Representativeness analysis...... 41 4. 2 Time series analysis of rainfall...... 42 4. 2. 1 Graphical analysis...... 43

Rainfall and its erosivity in Ethiopia 4. 2. 2 Systematic trend analysis...... 47 4. 2. 3 Autocorrelation and persistence...... 49

5 RAINFALL EROSIVITY IN THE WHOLE CENTRAL HIGHLANDS OF ETHIOPIA...... 53 5. 1 Statistical characteristics of long-term areal annual and summer rainfall erosivity...... 53 5. 2 Representativeness analysis...... 55 5. 3 Time series analysis...... 57 5. 3. I Graphical time series analysis...... 57 5. 3. 2 Temporal variability of long-term areal mean annual and summer rainfall erosivity...... 60 5. 3. 3 Systematic trend analysis, autocorrelation and persistence...... 62 5. 4 Relationship between areal rainfall and rainfall erosivity in the whole central highlands of Ethiopia...... 65 5. 5 The isoerodent contours for the central highlands of Ethiopia...... 68

6 RAINFALL EROSIVITY AT SELECTED WEATHER STATIONS IN THE CENTRAL HIGHLANDS OF ETHIOPIA...... 71 6. 1 Long-term annual rainfall erosivity...... 72 6. I. 1 Statistical parameters of annual rainfall erosivity...... 72 6. 1.2 Relationship between long-term annual rainfall erosivity and annual rainfall...... 74 6. 2 Long-term summer rainfall erosivity...... 77 6. 2. I Statistical parameters of summer rainfall erosivity...... 77 6. 2. 2 Relationship between long-term summer rainfall erosivity and summer rainfall...... 79 6. 3 Long-term rainfall erosivity during short rainy seasons...... 82 6.3.1 Autumn...... 82 6. 3. 2 Spring...... 83

Rainfall and its erosivity in Ethiopia 6. 4 Temporal and spatial characteristics of long-term rainfall erosivity...... 85 6. 4. 1 Temporal and seasonal variability of rainfall rosivity...... 85 6. 4. 1. 1 Graphical trend analysis...... 85 6. 4. 1.2 Departures of annual and summer rainfall erosivity from long-term mean...... 100 6. 4. 1.3 Systematic trend analysis...... 117 6. 4. 1. 4 Autocorrelation and persistence analysis of long-term rainfall rosivity...... 120 6. 4. 2 Spatial characteristics of rainfall erosivity...... 123 6. 4. 2. 1 Vertical characteristics of rainfall erosivity...... 123 6. 4. 2. 2 Horizontal characteristics of rainfall erosivity...... 126 6. 5 Rainfall erosivity classes for the central highlands of Ethiopia...... 132

7 IMPLICATIONS OF RAINFALL VARIABILITY AND RAINFALL EROSIVITY ON SURFACE FLOW AND LAND AND WATER MANAGEMENT...... 135 7. 1 The influence of rainfall variability on surface flow ...... 135 7. 1. 1 Materials and Methods...... 136 7.1.2 Results and discussion...... 138 7.1.3 Conclusion and recommendations...... 148 7. 2 Land and water management in a context of sustainability...... 149 7. 2. 1 The context...... 149 7. 2. 2 Indigenous Land and Water Management Technologies - ILWMTs...... 150 7. 2. 3 Modern Land and Water Management Technologies - MLWMTs...... 153 7. 2. 4 Socio-economic context of water erosion in the central highlands of Ethiopia...... 157 7. 2. 5 Land and water management policies and strategies of Ethiopia...... 159

Rainfall and its erosivity in Ethiopia 7. 2. 6 Retrospect and prospect...... 163

8 SUMMARY AND CONCLUSIONS, ZUSAMMENFASSUNG UND AUSBLICK, RESUME ET CONCLUSIONS...... 165 8.1 Summary...... 165 8.2 Conclusions...... 168 8.3 Zusammenfassung...... 170 8 .4 Ausblick...... 174 8 .5 Resume...... 175 8 .6 Conclusions...... 179

9 REFERENCES...... 181

10 ANNEXES...... 203

Rainfall and its erosivity in Ethiopia Fig. 30: Relationship between long-term areal annual rainfall and its erosivity in the central highlands of Ethiopia...... 67 Fig. 31: Relationship between long-term areal summer rainfall and its erosivity in the central highlands of Ethiopia...... 67 Fig. 32: Annual isoerodent contours for the central highlands of Ethiopia...... 69 Fig. 33: Summer isoerodent contours for the central highlands of Ethiopia...... 70 Fig. 34 a (I): Time series of annual rainfall erosivity at the Ejaji weather station...... 86 Fig. 34 a (II): Time series of annual rainfall erosivity at the Shola Gebeya weather station...... 87 Fig. 34 a (III): Time series of annual rainfall erosivity at the Tulu Bolo weather station...... 87 Fig. 34 b (I): Time series of summer rainfall erosivity at the Ejaji weather station...... 88 Fig. 34 b (II): Time series of summer rainfall erosivity at the Shola Gebeya weather station...... 88 Fig. 34 b (III): Time series of summer rainfall erosivity at the Tulu Bolo weather station...... 89 Fig. 35 a (I): Time series of annual rainfall erosivity at the Addis Ababa weather station...... 91 Fig. 35 a (II): Time series of annual rainfall erosivity at the Debre Markos weather station...... 91 Fig. 35 a (III): Time series of annual rainfall erosivity at the Debre Zeit weather station...... 92 Fig. 35 a (IV): Time series of annual rainfall erosivity at the Fiche weather station...... 92 Fig. 35 a (V): Time series of annual rainfall erosivity at the Kombolcha weather station...... 93 Fig. 35 a (VI): Time series of annual rainfall erosivity at the Majete weather station...... 93

Rainfall and its erosivity in Ethiopia x iv

Fig. 35 a (VII): Time series of annual rainfall erosivity at the Sheno weather station...... 94 Fig. 35 a (VIII): Time series of annual rainfall erosivity at the Wonji weather station...... 94 Fig. 35 b (I): Time series of summer rainfall erosivity at the Addis Ababa weather station...... 96 Fig. 35 b (II): Time series of summer rainfall erosivity at the Debre Markos weather station...... 96 Fig. 35 b (III): Time series of summer rainfall erosivity at the Debre Zeit weather station...... 97 Fig. 35 b (IV): Time series of summer rainfall erosivity at the Fiche weather station...... 97 Fig. 35 b (V): Time series of summer rainfall erosivity at the Kombolcha weather station...... 98 Fig. 35 b (VI): Time series of summer rainfall erosivity at the Majete weather station...... 98 Fig. 35 b (VII): Time series of summer rainfall erosivity at the Sheno weather station...... 99 Fig. 35 b (VIII): Time series of summer rainfall erosivity at the Wonji weather station...... 99 Fig. 36: Departures of annual and summer rainfall erosivity from long-term average at the Addis Ababa weather station...... 101 Fig. 37: Departures of annual and summer rainfall erosivity from long-term average at the Debre Markos weather station...... 103 Fig. 38: Departures of annual and summer rainfall erosivity from long-term average at the Debre Zeit weather station...... 105 Fig. 39: Departures of annual and summer rainfall erosivity from long-term average at Ejaji weather station...... 106

Rainfall and its erosivity in Ethiopia XV

Fig. 40: Departures of annual and summer rainfall erosivity from long-term average at the Fiche weather station...... 108 Fig. 41 a: Departures of annual and summer rainfall erosivity from long-term average at the Majete weather station...... 111 Fig. 41 b: Departures of annual and summer rainfall erosivity from long-term average at the Sheno weather station...... 111 Fig. 41 c: Departures of annual and summer rainfall erosivity from long-term average at the Shola Gebeya weather station...... 112 Fig. 41 d: Departures of annual and summer rainfall erosivity from long-term average at the Tulu Bolo weather station...... 112 Fig. 42: Departures of annual and summer rainfall erosivity from long-term average at the Kombolcha weather station...... 115 Fig. 43: Departures of annual and summer rainfall erosivity from long-term average at the Wonji weather station...... 117 Fig. 44: Relationship between long-term mean rainfall erosivity and altitude for selected weather station in the central highlands of Ethiopia...... 125 Fig. 45: Relationship between latitude and variability in annual rainfall erosivity - regional diminution...... 127 Fig. 46: Relationship between longitude and variability in annual rainfall erosivity - regional dimension...... 128 Fig. 47: Relationship between latitude and variability in summer rainfall erosivity - regional dimension...... 129 Fig. 48: Relationship between longitude and variability in summer rainfall erosivity - regional dimension...... 130

Rainfall and its erosivity in Ethiopia x v i

Fig. 49: Percentage of weather stations in the central Ethiopian highlands assigned to rainfall erosion risk classes...... 133 Fig. 50 a: Comparison of mean annual rainfall and surface runoff at the Hombolegauging station...... 140 Fig. 50 b: Comparison of mean summer rainfall and surface runoff at the Hombole gauging station...... 140 Fig. 51 a: Comparison of mean annual rainfall and surface runoff at the Kessem gauging station...... 141 Fig. 51 b: Comparison of mean summer rainfall and surface runoff at the Kessem gauging station...... 141 Fig. 52 a: Comparison of mean annual rainfall and surface runoff at the Meleka Kunture gauging station...... 142 Fig. 52 b: Comparison of mean summer rainfall and surface runoff at the Melka Kunture gauging station...... 142 Fig. 53 a: Comparison of mean annual rainfall and surface runoff at the Modjo gauging station...... 143 Fig. 53 b: Comparison of mean summer rainfall and surface runoff at the Modjo gauging station...... 143 Fig. 54 a: Comparison of mean annual rainfall and surface runoff at the Teji gauging station...... 144 Fig. 54 b: Comparison of mean summer rainfall and surface runoff at the Teji gauging station...... 144 Fig. 55: Traditional hillside terraces for soil and water management in the central highlands of Ethiopia (North Shoa)...... 152 Fig. 56: Integrated traditional agricultural land-use practices for land and water management in the central highlands of Ethiopia (North Shoa)...... 153 Fig. 57: Conventional check dams for land and water management in the central highlands of Ethiopia (North Shoa)...... 154

Rainfall and its erosivity in Ethiopia LIST OF TABLES PAGE

Tab. 1: Selected weather stations in the central highlands of Ethiopia and their general Information...... 14 Tab. 2: Statistical parameters of long-term rainfall and its erosivity for the FAO (1984) rainfall pattern regions in Ethiopia...... 33 Tab. 3: Statistical parameters of long-term annual rainfall erosivity...... 73 Tab. 4: Regression models of long-term annual rainfall erosivity for selected weather stations in the central highlands of Ethiopia...... 76 Tab. 5: Long-term summer (kiremt) rainfall erosivity for selected weather stations in the central highlands of Ethiopia...... 78 Tab. 6: Regression models for long-term summer (kiremt) rainfall erosivity...... 80 Tab. 7: Statistical parameters of autumn (Tseday) and spring (Belg) rainfall erosivity (mm) for selected weather stations in the central highlands of Ethiopia...... 84 Tab. 8: Trend correlation coefficients of rainfall erosivity for selected weather stations in the central highlands of Ethiopia...... 118 Tab. 9: Autocorrelation of long-term annual rainfall erosivity...... 121 Tab. 10: Autocorrelation of long-term summer rainfall erosivity...... 121 Tab. 11: Classes of annual rainfall erosion risk for the central highlands of Ethiopia...... 132 Tab. 12: Classes of summer rainfall erosion risk for the central highlands of Ethiopia...... 132 Tab. 13 : Selected gauging stations in the central highlands of Ethiopia...... 137

Rainfall and its erosivity in Ethiopia XVIXI

Tab. 14: Statistical characteristics of hydrometric records at the selected gauging stations...... 138 Tab. 15: Cross correlation coefficients between hydrometric records and areal rainfall...... 145 Tab. 16: Non-parametric trend correlation coefficients of surface runoff and rainfall...... 147

Rainfall and its erosivity in Ethiopia LIST OF FIGURES PAGE

Fig. 1: Natural division of Ethiopia showing the highlands and the central highlands...... 8 Fig. 2: Network and distribution of selected meteorological stations in Ethiopia...... 13 Fig. 3: Air stream over Ethiopia during the major dry season (October - May)...... 18 Fig. 4: Air stream over Ethiopia during the major rainy season (late June - early September)...... 19 Fig. 5a: Major disturbances in air stream over Ethiopia during March - May, penetrating south...... 21 Fig. 5b: Major disturbances in air stream over Ethiopia during March - May, penetrating south...... 22 Fig. 6: Isohyetes of long-term mean annual rainfall over Ethiopia...... 24 Fig. 7: Seasonal rainfall distribution [±5%] for selected weather stations in Ethiopia (1898-1997)...... 25 Fig. 8: Long-term seasonal rainfall distribution data series for Ethiopia (1898-1997)...... 28 Fig. 9: Modified rainfall pattern regions of Ethiopia, FAO (1984)...... 32 Fig. 10: Relationship between precipitation and rainfall erosivity in rainfall pattern region A ...... 35 Fig. 11: Relationship between precipitation and rainfall erosivity in rainfall pattern region B ...... 35 Fig. 12: Relationship between precipitation and rainfall erosivity in rainfall pattern region C ...... 36 Fig. 13: Relationship between precipitation and rainfall erosivity in rainfall pattern region D ...... 36 Fig. 14: Relationship between precipitation and rainfall erosivity in rainfall pattern region E...... 37

Rainfall and its erosivity in Ethiopia xix

LIST OF ABBREVIATIONS

AC Autocorrelation ACF Autocorrelation Function ACC Autocorrelaion Coefficient ADMM Annual Departures from the Multi-year Mean AFm Annual Fni AGBAG Allgemeine Bodenabtragsgleichung CEC Commission for the European Communities CFSCDD Community Forestry and Soil Conservation and Development Department CSE Conservation Strategy of Ethiopia CV Coefficient of Variation DA Development Agents E.T. Equatorial Trough EHRS Ethiopian Highlands Reclamation Study EPA Environmental Protection Authority ERCS Ethiopian Red Cross Society FAO Food and Agriculture Organisation FaWCDA Forestry and Wildlife Conservation and Development Authority FDRE Federal Democratic Republic of Ethiopia FINNIDA Finnish International Development Authority Fm Modified Fournier's index GTZ Deutsche Gesellschaft fur Technische Zusammenarbeit I.T.C.Z. Inter Tropical Convergence Zone ICRAF International Centre for Research in Agroforestry ILWMT Indigenous Land and Water Management Technology IUCN International Union for the Conservation of Nature LCL Lower Confidence Limit LPB Lower Prediction Band MEDC Ministry of Economic Development and Co-operation MLWMT Modem Land and Water Management Technologies MoA Ministry of Agriculture

Rainfall and ist erosivity in Ethiopia XX

MoACE Ministry of Agriculture Conservation and Environment MoWR Ministry of Water Resources N Number of observations NDVI Normalised Difference Vegetation Index NMSA National Meteorological Service Agency NO A A National Oceanic and Atmospheric Administration NRCDMD Natural Resources Conservation and Development Main Department NRHS Natural Resources and Human Settlement OAU Organisation of African Unity ONCCP Office of the National Committee for Central Planning PA Peasant Association Prob. Probability R2 Coefficient of Determination RRC Relief and Rehabilitation Commission RUSLE Revised Universal Soil Loss Equation SCLTRD Soil Conservation and Land-use Technology and Regulatory Department SFCDD State Forest Conservation and Development Department SIDA Swedish International Development Authority SUDMM Summer Departures from the Multi-year Mean SUDSLTA Summer Departures from the Summer Long-term Average SUFm Summer Fm SWC Soil and Water Conservation SWCD Soil and Water Conservation Department UCL Upper Confidence Limit UIA Union of International Associations UNDHA United Nations Department of Humanitarian Affairs UNDP United Nations Development Programme UNEP United Nations Environment Programme UNESCO United Nations Educational, Scientifc, and Cultural Organisation UNICEF United Nation Children’s Fund UNSO United Nations Sahelian Office

Rainfall and ist erosivity in Ethiopia xxi

UPB Upper Prediction Band USLE Universal Soil Loss Equation VITA Volunteers in Technical Assistance WFP World Food Programme

Rainfall and ist erosivity’ in Ethiopia 1 INTRODUCTION

1. 1 Background and statement of the problem

In the human time scale, soil is a non-renewable natural resource that is difficult to replace once lost. The word “erosion” is derived from the Latin word “ erosio” meaning “ to gnaw away” (L a l 1990). The term ’’soil erosion” refers to the naturally occurring phenomenon of soil deterioration caused by eroding agents (most important of which are water, wind, snow and gravity), but vastly accelerated process of soil removal and redeposition brought about by human interference for economic activities with the normal dynamic equilibrium state between soil formation and depletion (B e n n e tt 1939; UIA 1994-95). Conceptual definitions of soil erosion are proposed and discussed in the literature (see inter alia R ic h te r 1976; S c h u ltz e 1976; B o rk 1982; B o rk 1987; B e c k e d a h l 1998).

Manmade soil erosion is as old as the introduction of land use by man for settled as well as intensive agriculture and causes major environmental problems world­ wide (P imentel et al. 1995). It is considered to be responsible for the vanishing

of many ancient civilisations (Eckholm 1976; O lson 1981). Warnings of the danger of soil erosion came from Plato more than 2 000 years ago and from

other writers more recently (e.g., M uckenhausen 1976). The current extent of human-induced soil erosion and its potential danger are recorded for different

parts of the world (e.g., M iddleton and Thomas 1997). Globally, 1 094.6 million hectares of land, of which 440.6 are found in Asia, 227.4 in Africa, 123.2 in South America, 46.3 in Central America, 59.8 in North America, 114.5 in Europe and 82.8 in Australia, are degraded due to water erosion in a form of

loss of top soil accompanied by terrain deformation (e.g., O ldeman et al. 1991; Steiner 1996). According to the study conducted by Pimentel et al. (1995), about 80% of the world's agricultural land suffer moderate to severe erosion and

10% slight to moderate erosion. Buringii (1989) reported that agricultural land degradation alone can be expected to depress world food production between 15 to 30% by the year 2020.

Rainfall and its erosivity in Ethiopia 2

Water and wind erosion, usually associated with deforestation, overgrazing and shifting cultivation, have left their mark on large areas of the African continent

(Lal 1988; Dregne 1990; M iddleton and Thomas 1997; N ill et al. 1996). There is presumptive evidence that the granary of Rome experienced some erosion in the waning days of the empire, but the first real threat seems to have appeared in the seventh Century, after the Arab conquest of present day Tunisia,

Algeria and Morocco (M ikesell 1960). Warnings about soil erosion threats in Africa appeared from time to time after about 1910. By 1930, a torrent of reports appeared on the harm that erosion had done and was doing in North Africa, the

British colonies and South Africa (D regne 1990). Presently, severe accelerated soil erosion occurs in all major ecological regions of tropical Africa (see Lal

1988). For instance, according to Richter (1998), 2.27 million km2 of land area in Africa are already degraded through water erosion; of which 25% are slightly, 30% moderately, 43% severely and 2% extremely damaged. Frequent drought occurrence with alternating extreme rainfall and flood events has exacerbated not only water but also wind erosion (N icholson 1998)

A b a te (1997) wrote that the Ethiopian highlands had been inhabited by agricultural Hamites long before the Axumite Civilisation around 3 000 B. C. By the first century A. D., the heartland of ancient Abyssinia was already an area with farmland degradation, thus significantly contributing to the collapse of the great Ethiopian civilisation (see Brown 1973; Bard et al. 2000). This supports the hypothesis that agriculture in the country has been progressively intensified since the beginning (A statke and Kelemu 1993; A b a te 1997). The evidence for the early use of soil and water conservation practices, as reported by H u ffn a g e l (1961), W e stp h a l (1975) and other authors (e.g., H allpike 1970; A lem ay eh u 1996; A srat et al. 1996; Gebre-Michael 1998), indicates that soil degradation has been the major environmental problem in Ethiopia since early historical times. In their efforts to secure food and other basic demands, Ethiopians over centuries have taken little account of future generations’ means of existence, to the extent that soil degradation went beyond the mitigating capacity of traditional conservation measures. For instance, Kuru (1986) mentioned that even though traditional soil and water conservation has been part

Rainfall and its erosivity in Ethiopia 3 of subsistence agriculture in north Shoa, the heart of the central highlands, nearly non of the top soil is left on the agricultural land. Subsequently, even trees and various plants which could otherwise have survived on poor soils are dwindling. Hoben (1997) narrates that increase in population and consequent activities such as intensive cultivation, overgrazing by livestock, deforestation and inappropriate land use practices aggravated the problem of soil erosion in the country. The author compares the situation in Ethiopia with the Neo- Malthusian environmental degradation scenario. Progressive deforestation and lack of watershed management in the highlands have resulted in high water yield due to increased runoff caused by reduced water retention capacity of the soil. Only some hundred years ago, more than 40% of land area in Ethiopia was covered with natural forests and woodlands (B reitenbach 1963). According to the United Nations Environment Programme (UNEP), as recently as two decades ago, 16% of land surface in Ethiopia was under forest cover, but by

1982 this area had diminished to 3.1% (UNEP 1983 cited in K uru 1986). This figure on current forest cover in Ethiopia at large and the central highlands of Ethiopia in particular is already old but repeatedly reported. There are controversies in literature whether the reported figures on forest cover are true or not (e.g., Woien 1995; Hoben 1997).

Besides the onsite impacts, the problem of soil erosion has expanded to low lying areas in a form of reservoir, lake and marsh sedimentation, damaging of agricultural land, settlement areas and roads (FAO 1993). For example, G ebre-

M ariam (1998) reported that water chemical and biological quality deterioration due to soil erosion and sedimentation in Ethiopia is critically advancing. The Ethiopian agriculture has developed in such a way as to threaten its future capacity to sustain development and support life by ignoring conservation of resources; the consequent impact on soil and land resource is immense. However, the comparative degree of soil erosion, land degradation and its severity varies widely within and among the major agro-ecological zones, land units, altitudinal belts, drainage and river basins (e.g., V irgo and M unro 1978; Kuru 1986). At present, Ethiopia is believed to be facing an environmental crisis of greater magnitude than ever before. The degradation of

Rainfall and its erosivity in Ethiopia 4 soil resulting from erosion and consequent deterioration of soil quality - physical, chemical and biological degradation - in Ethiopia is much faster than the rate of soil formation (HURNI 1993). The Ethiopian Highland Reclamation Study (EHRS) (FAO 1986) indicated that, of the 53.5 million ha of the Ethiopian highlands, 28% are very severely affected by accelerated water erosion and 24% are moderately affected but still to a serious degree. This leaves only 48% of the area free from erosion problems of which more than half (ca. 58%) is at risk owing to its high to very high susceptibility to accelerated erosion if conservation agriculture is not practised. Every year, 1.9 billion tonnes of soil are eroded from the highlands of Ethiopia. H urni (1987) reported that, at the current rate of soil erosion, some 38 000 km" of agricultural land will disappear by the year 2 010 and some 60 000 km" will have less than 10 cm soil depth. The current rate of soil erosion reduces food production in Ethiopia by at least 2% per annum (K a ppel 1996). A rough estimation of FAO (1986) indicated that agricultural soil degradation will cost Ethiopia about U.S. $ 7 246.4 million over the next 25 years (1986 average exchange rate U.S. $ 1 = 2.07 Ethiopian Birr). This is an annual average of U.S. $ 290 million. Nearly 80% of the losses are attributed to reduced crop production and the rest to reduced livestock production. This will mean a substantial decrease in per capita income (ca. 30%) and, hence, increased mass poverty.

Water erosion, accentuated with its variability in time and space, strongly influences decision making on land-use and land/water management practices. The temporal and spatial variability of rainfall has tremendous consequence on surface and subsurface hydrologic regime (e.g., M oreda and B auwens 1998) and, hence, on rainfall erosive potential (W ischmeier 1962). For instance, FAO (1965) reported a mean specific degradation of 20.66 t/km2/annum for the Awash River basin which makes a large part of the central Ethiopian highlands. However, these figures are subject to fluctuation depending on rainfall variability and the consequent hydrologic regime in the area. While frequent flooding in Ethiopia at large and in the central highlands of Ethiopia in particular has resulted in huge socio-economic and environmental damage (VITA 1996a, 1996b, 1997a, 1997b), its impacts extended beyond the country’s

Rainfall and its erosivity in Ethiopia 5 boundary into the neighbouring regions; mainly Somalia, Kenya and the Sudan (VITA 1998).

An appropriate and sustainable environmental management planning and implementation requires thorough understanding of those variables responsible for environmental degradation. Although factors such as rainfall, wind and land- use/land and water management practices play an important role in soil degradation, water erosion is the major but highly neglected degrading agent in agricultural and environmental research in Ethiopia in general and in the central highlands of Ethiopia in particular. Rainfall, as opposed to other eroding agents, is not directly influenced by human measures and, hence, the most uncertain element in water erosion prediction. Agricultural development and land management plans and implementation are highly sensitive to this uncertain element of water erosion prediction - the rainfall and runoff variable. Therefore, the most viable preventive approach to water erosion is the understanding of the erosive potential of water and integration of protective measures into sustainable land-use and water management planning. Moreover, thorough knowledge of rainfall erosivity potential and its temporal as well as spatial characteristics is a fundamental prerequisite to develop empirical predictive tools of water erosion risk and devise control measures.

Several authors quantified potential rainfall erosivity for different parts of the world using various methods (e.g., Roose 1977; Ram Babu et al. 1978; W ischmeier and Smith 1978; Sauerborn 1994; Oduro-A friyie 1996; Favis-

M ortlock and G uerra 1999; Sauerborn et al. 1999). Hardly any research on water erosion potential risk assessment in a medium to long-term perspective was carried out in Ethiopia. Attempts to quantify these were limited in scope, spatial and temporal coverage (e.g., A begaz 1989; Bosshart 1998). Thus, little is understood about water erosion potential in the Ethiopian highlands as a whole and in the central highlands of the country in particular, especially concerning its long-term temporal and spatial characteristics. The viability of land development and environmental management plan based on investigations with limited scope is rather questionable and rarely realistic. A detailed

Rainfall and its erosivity’ in Ethiopia 6 investigation of sufficiently long-term climatic data covering large areas of the country and various climatic pattern regions (see FAO 1984) should be made to get a thorough knowledge of rainfall erosivity in Ethiopia and the central highlands of the country.

The database established on the basis of such detailed investigation, and the know-how gained thereof, enable to develop reliable quantitative prediction tools readily applicable to land and water resource management planning. Moreover, these will facilitate the design of realistic environmental conservation policies and strategies as well as advise land-users on appropriate land management or land-use practices. In addition, the results of such study would provide a functional database for scientific research on the Ethiopian climate and ecosystem and, consequently, contribute to the better understanding of local and global environmental change.

1. 2 Objectives of the study

This study aims at: i. Assessing the structure of rainfall and its short/long-term temporal and spatial variability in Ethiopia with major focus on the central highlands; ii. Estimating rainfall erosivity for the whole of Ethiopia, with special consideration of the central Ethiopian highlands; iii. Carrying out elaborate quantitative analysis of rainfall erosivity in wide ranging temporal and spatial framework/coverage and developing prediction tools of rainfall erosivity which are readily applicable in Ethiopia in situations of limited data availability; iv. The study will further investigate the implications of rainfall and rainfall erosivity in the central highlands of Ethiopia in relation to surface hydrology and socio-economic links; v. Furthermore, it is the objective of this study to develop precipitation and rainfall erosivity database for meso to macro scale land and water resources management planning in Ethiopia which is useful for future research, policy and strategy making.

Rainfall and its erosivity in Ethiopia 7

2 MATERIALS AND METHODS

2. 1 The Ethiopian and the central highlands

Nearly 44% of Ethiopia’s land area are highlands 1 500 m above sea level. Some authors reported that more than 50% of the area in the Ethiopian highlands have slope exceeding 16% (e.g., G etahun 1978; E l W akeel and

Astatke 1996; A bate 1997). The Ethiopian highlands are generally divided into western and eastern highlands. Further sub-divisions are suggested by several authors (e.g., M ohr 1971; W estphal 1975; M ayer-L eixner 1999).

Agro-ecologically, W right and Adamseged (1984) divided the Ethiopian highlands into three broad zones: (1) the High Potential Cereal/Livestock Zone in high and intermediate altitudes with a good natural resource base, but under severe human pressure; (2) the Low Potential Cereal/Livestock Zone, which is quite similar to the High Potential Cereal/Livestock Zone, but with advanced ecological degradation; (3) the Low Potential Cereal/Livestock Highlands Escarpment which is characterised by severe seasonal water shortage for both human and livestock uses, except in the river valleys. The Ethiopian highlands are home for ca. 85% of the country’s human and about 75% of livestock population. They account for 95% of Ethiopia’s cultivated land area and are origin of more than 90% of the country’s agricultural products, thus playing a major role in the national economy.

The central highlands were chosen as area of focus for the investigation of rainfall variability and rainfall erosivity in Ethiopia because they are believed to represent the highlands of the country with regard to socio-economic, demographic as well as ecological aspects. The ecological disasters faced by the central highlands are believed to be representative of those ecological problems faced by the whole highlands. The area hereby designated as the central highlands was arbitrarily defined for the purpose of this study (Fig. 1).

Rainfall and its erosivity in Ethiopia Fig. 1: Natural division of Ethiopia showing the highlands and the central highlands

The adoption of such definition was necessary because there is no distinct and geographically delimited topographic unit as such called the central highlands of Ethiopia yet. The region covers various regional states and provinces, thus making impossible to deal with rainfall erosivity at a single province or state level. As per the arbitrary definition, the study area lies between latitudes 7° 02' to 11° 46'N and longitudes 36° 27' to 40° 12' E. The climate of the area is mainly tropical highland climate with dry winter season. According to Ethiopian climate classification (Tessema et al. 1993), the climate of the area ranges from dry Kolia to moist Dega. The topography of the area is dominated by high mountains dissected by the Great Ethiopian Rift System. The Shoa Plateau, the Arsi and Bale Mountains constitute the outstanding physiographic features of the region. The soils of the central Ethiopian highlands are highly degraded due to long history of intensive agricultural land-use and progressively increasing pressure on land (FAO 1986). The topographic characteristics of the area, poor

Rainfall and its erosivity in Ethiopia 9 land-cover, inappropriate land-use and management have enhanced the process of water erosion. The major populous central Ethiopian towns; namely, Addis Ababa, Asela and Debre Zeit are concentrated in this geographic unit. Adaba, Kombolcha, Goba, Yirga Alem, Debre Markos and Albuco are also important towns located near the frontier of the region. The central highlands of Ethiopia are the most densely populated region of the country and, hence, the most intensively cultivated area.

2. 2 Rainfall data set and data analysis

2. 2. 1 Concepts of rainfall erosivity

Rainfall erosivity is defined as the potential ability of rainfall to cause soil erosion (H udson 1971). While studying the impact of rainfall on soil, Ellison ( 1952) described soil erosion as a work process and the resulting forms of erosion; namely gullies, sheet-erosion scars and related erosional damage, as the product of work done by rainfall on soil surface. The blasts of raindrops beating on bare soil surface erode and damage the soil by breaking down its crumbs and splashing them into parts. Investigations by several authors have led to a wide acceptance that rainfall drop impact is the major eroding agent (e.g., FREE 1952;

Hudson 1961; W ischmeier and Smith 1978). Studies on tropical rains have indicated that rainfalls in the tropics are more erosive than temperate rains. This is attributed to the high rainfall intensity, the relatively large drop size and the accompanying high wind velocity. El Swaify and Dangler (1982) reported a maximum rainfall intensity of 800 mm/h for Jamaica. Kowal and Kassam (1977) registered peak intensities of 120-160 mm/h for Nigeria and extreme intensity of 190 mm/h for western part of Nigeria. Lal et al. (1980) reported a median drop size of 3 mm for Nigeria and an association of highly erosive rains with high wind velocities. In his investigation of the physical characteristics of subtropical rainfall in southern Rhodesia (Zimbabwe), Hudson (1961) reported that an intensity of 152 mm/h was always expected and a value of 254 mm/h recorded several times. The author reported that extreme values were also observed during the study. The largest median drop size D5q coincided with peak

Rainfall and its erosivity in Ethiopia 10

intensities. De Ploey (1972), in his comparative study on erosive capacity of tropical and middle-latitude rains, found that even an intermittent, discontinuous runoff and splash erosion were very active on gentle crest slopes in Kinshasa. This is comparable with what would occur only on down medium slopes of about 10° in Antwerp. El Hassanin et al. (1994) reported that rainfall in Burundi contributed to the major proportion of soil erosion.

Various water erosion estimation and prediction techniques used in different parts of the world; e. g., the Universal Soil Loss Equation (USLE) (W ischmeier and Smith 1978), the Allgemeine Bodenabtragsgleichung (ABAG)

(Schwertmann 1981), the Revised Universal Soil Loss Equation (RUSLE)

(Renard et al. 1997) are inappropriate to apply in Ethiopia because of data limitations. Monthly and annual rainfall data are the only readily available information in Ethiopia from which such estimation could be made. This study applied the Modified Fournier’s Index (A rnoldus 1977) which has already found wide range of acceptance after it had been tested under various climatic regions in different countries. Based on his work in Morocco, Arnoldus (1977) proved that Fournier’s climatic erosivity index (c) (F ournier 1960) was weakly correlated to that of W ischmeier and Smith (1978). Therefore, he modified the Fournier’s climatic erosivity index to use in countries with weak database to calculate either of the indices developed in different parts of the world. The Modified Fournier’s Index (Fm) of rainfall erosivity is given by the following mathematical expression:

12 2 i = 1 F m P

Whereby: Pi = monthly precipitation and P = annual precipitation.

Rainfall and its erosivity in Ethiopia 11

Based on numerous regression models he developed for various climatic zones,

A r n o ld us (1977) recommended to use the Modified Fournier’s Index only in regions with homogenous climate. He partitioned Africa, north of the equator, and the Middle East into homogeneous climatic zones and developed an isoerodent map of the region in metric units (A rn old us 1980). O d u r o -A friy ie (1996) developed an isoerodent map of Ghana using the Fournier’s index.

G r iffith s and Ric h a r d s (1989) used the index for the Awash River Basin in

Ethiopia as recommended by FAO (B o erw in k el and P aris 1984) and suggested its use in other parts of Ethiopia where sufficient data are available.

2. 2. 2 The data set

Precipitation data were obtained from the Ethiopian National Meteorological Service Agency (NMSA), the Ministry of Water Resources (MoWR), the climate CD-ROM of the FAO (FAO 1995) and the database of the National Oceanic and Atmospheric Administration (NOAA). The original data were made up of total monthly rainfall (mm), latitude/longitude in degrees and minutes, altitude (m) and local names of the weather stations. The weather stations are named either after the old provincial, Awraja or Woreda names. Errors in topographic information and station location were corrected where mistakes were found in the original documents. Currently, some of the weather stations do not operate since they were shutdown due to personnel and operating budget constraints while others function occasionally (personal communication with authorised NMSA staff members 1998). The rainfall data records obtained from various sources were meticulously scrutinised for their quality and quantity. Weather stations with a data series shorter than seven years were excluded from the study, and those longer than that were further checked for data quality. For this study, stations with irregular data characteristics were eliminated from further investigation at the outset. In spite of their series length, several stations had incomplete data set whose gaps were difficult to estimate due to heterogeneity of the Ethiopian topography and missing basic information which would otherwise have been a springboard for data estimation. Such stations were eliminated from the study as well. Many weather stations had a

Rainfall and its erosivity in Ethiopia 12 data set with frequent discontinuity and breaks of various length and features. Such is a special characteristic of the weather stations in northern Ethiopia, for instance Tigray. Data for these stations were corrected, where possible, and respective stations included in the study or cancelled from the investigation otherwise. The final data set covers the period extending from 1898-1998. Addis Ababa has the longest rainfall data series (1898-1996); Dembi Dolo (1978- 1993), Elias (1981-1987), Merawe (1981-1987) and Gumuro (1978-1985) have the shortest, whereby all have data sufficiently long for this study. Weather stations used for the assessment of rainfall characteristics and rainfall erosivity in the central highlands of Ethiopia were chosen from the final data set. Weather stations with data series longer than thirty years were selected for elaborate investigation of the variability of rainfall and rainfall erosivity in the central highlands of Ethiopia from the data set established for the central highlands of Ethiopia. Fig. 2 and Tab. 1 depict the weather stations finally chosen for the assessment of rainfall erosivity in Ethiopia and the central highlands, respectively.

Rainfall and its erosivity’ in Ethiopia 13

Fig. 2: Network and distribution of selected meteorological stations in Ethiopia

Rainfall and its erosivity in Ethiopia 14

Tab. 1: Selected weather stations in the central highlands of Ethiopia and their general Information Serial Reference Weather Latitude Longitude Altitude Years of Nr. Nr. in all stations (North) (East) (m) observation stations list 1 2 Adaba 7°02' 39°40' 2410 1968-1987 2 3 Addis Ababa 9°02' 38°42' 2408 1898-1996 3 7 A ear fa 7°16' 39049' 2520 1968-1989 4 11 Albuco 11°46' 39°34' 2100 1962-1983 5 13 Alem 10°02' 39°02' 2280 1973-1997 6 18 Ankober 9°00' 39°40' 2600 1962-1983 7 22 Artuma 10°35' 4 0 °0 r 1880 1975-1987 8 23 Asasa 7°12' 39019' 2370 1971-1987 9 24 Asela 7°57' 39°08' 2350 1966-1988 10 28 Awassa 7°05' 38°29' 1700 1972-1987 11 39 Bora 10°41' 40°03' 1500 1975-1987 12 40 Boru Meda 11°22' 39°28' 2720 1962-1983 13 50 Debre 10°20' 37°40' 2515 1954-1991 14 52 Debre Work 10°44' 38°08' 2740 1975-1987 15 53 Debre Zeit 8°44' 38°57' 1900 1951-1984 16 57 Dessie 11°06' 39°38' 2500 1962-1987 17 59 Dhera 8°20' 390, 9 ' 1770 1976-1996 18 65 Eiaii 9°00' 37°19' 1900 1966-1997 19 70 Fiche 9°48' 38°42' 2750 1954-1997 20 76 Gessera 7°22' 4 0 ° ir 2320 1968-1986 21 79 Goba 7°02' 40°00' 2710 1961-1984 22 80 Gobessa 7°48' 37°30' 2300 1969-1985 23 82 Gohatsion 10°02' 38°14' 2560 1972-1997 24 94 Hosana 7°33' 37°52' 2200 1953-1968 25 95 Huruta 8°09' 39°22' 2200 1969-1988 26 96 Idoto 7°3 r 40°03' 2480 1969-1985 27 100 Kachise 9°35' 37°20' 2520 1986-1996 28 103 Keleta 8°19' 39025' 1500 1962-1986 29 108 Kofole 7°04' 38°47' 2620 1961-1987 30 110 Kombolcha 11°07' 39044' 1916 1952-1991 31 112 Kore 7° 13' 38°54' 2500 1969-1985 32 117 Mai ete 10°27' 39°51' 2000 1962-1996 33 121 Mararo 7°27' 39°22' 2940 1968-1988 34 133 Munessa 7°35' 38°54' 2510 1958-1975 35 135 Nekemte 9°05' 36°27' 2080 1971-1981 36 138 Robi 7° 15' 39037' 2400 1982-1990 37 139 Saeure 7°46' 39°09' 2480 1973-1983 38 140 Sedika 7°43' 39°40' 2545 1968-1985 39 141 Seru 7°40' 40°12' 2480 1968-1985 40 145 Sheno 9°20' 390 is ' 2655 1962-1995 41 146 Shola 9° 13' 39027' 2650 1962-1996 42 148 Sire 7° 13' 38°53' 2390 1953-1975 43 152 Tulu Bolo 8°40' 38°13' 2100 1962-1996 44 158 Wonii 8°29' 39°15' 1540 1951-1984

Rainfall and its erosivity in Ethiopia 15

2. 2. 3 Data entry and statistical analysis

Selected weather stations across Ethiopia were plotted on the country’s map

(Fig. 2) using Arc View 3.1 (Liebig 1997). The original data, which were available only as hard copy, were entered using Microsoft Excel 97, preliminary preparations and calculations were made on the same programme. Detailed statistical analyses were carried out using SPSS 8.0/SPSS 9.0 (F elix 1998). Basic statistics on measures of location (mean, minimum, maximum), dispersion (standard deviation) and distribution coefficients (skewness, kurtosis and variability coefficient) were calculated for rainfall and rainfall erosivity in the central Ethiopian highlands and for the stations throughout the whole country.

To assess temporal characteristics of rainfall and rainfall erosivity, time series analyses were carried out on the data of the central Ethiopian highlands based on weather stations with data series longer than thirty years. The choice of a thirty year period as the minimum time span for time series analysis was based on the fact that rainfall erosivity is a climatic element (and, hence, a function of climate) and experience of studies made on climate changes (M itchell et al. 1966). For the time series analysis, short-term missing data ranging from 1 to 3 consecutive years were substituted by series mean (e.g., Sneyers 1990). Long­ term annual rainfall erosivity series for selected weather stations were smoothed using 5-years and 11-years simple running means (Essenwanger 1986). Linear trend lines were fit to the smoothing curves to investigate the general direction of change the series underwent in the course of observation period. A systematic trend analysis was performed using Spearman’s non-parametric statistics

(M itchell, et al. 1966; Kothyari et al. 1997). The data series were tested for normality using One-Sample Kolmogorov-Smimov test. The two dimensional linear Product-Moment correlation coefficient - Pearson’s correlation coefficient - was calculated for the series for further comparison with the results of Spearman’s coefficient and test of representativity (Rapp and SCHONWIESE 1996). Auto-correlation analysis for rainfall and rainfall erosivity was made to investigate the persistence of the erosivity data in the course of observation period (B rockwell and Davis 1996).

Rainfall and its erosivity in Ethiopia 16

In order to analyse the temporal variability of the rainfall erosivity in the central highlands of Ethiopia, departures of annual and seasonal rainfall erosivities from multi-year/long-term seasonal average were calculated for those stations with more than or equal to thirty years data availability. The seasons of the year in Ethiopia used for the calculation of seasonal rainfall erosivity were adopted from Ek lu n d h and P ilesjo (1990).

2. 2. 4 Regression analysis and isoerodent contours

In order to develop the isoerodent contours for the central highlands of Ethiopia and the whole country, simple regression analyses were performed, whereby annual and seasonal rainfall erosivities were regressed to the long-term annual and summer rainfall (S a u erb o rn and G r u n er t 1995). For this purpose, a series of linear and non-linear (logarithmic, quadratic, cubic, power and exponential) regression models were fit and the best functional form chosen based on quantitative, theoretical and practical criteria (e.g., A n sc o m b e 1973; G o m ez

1984; ESSENW a n ger 1986). The behaviour of rainfall erosivity, as evidenced by research experience in tropical countries (e.g., Pa u w ely n et al. 1988), was taken into account while choosing the best regression models.

Rainfall and its erosivity’ in Ethiopia 3 PRESSURE PATTERNS, PRECIPITATION AND RAINFALL EROSIVITY IN ETHIOPIA

Before detailed assessment is made on rainfall erosivity, it is important to discuss the atmospheric circulation influencing rainfall and its formation over Ethiopia. The understanding of precipitation generating processes enhances to view water erosion from climate perspective and, hence, make a reasonable comparison of precipitation and its erosive potential, especially concerning its spatial distribution and variability. The rain generating circulation and rainfall in Ethiopia are discussed in the succeeding section.

3. 1 Pressure patterns over Ethiopia

According to the investigations of various authors (e.g., T a to 1964; SUZUKI

1967; G r iffiths 1972), two great anticyclones dominate Ethiopia during the major dry season which spans from October to May. As illustrated by Fig. 3, there is an east-west oriented anticyclone centred over the Sudan and another, oriented north-south, overlies Arabia and projects south into Ethiopia. The northerly current from the Sudan high pressure is very dry and stable because of its land trajectory across the Sahara. The second anticyclonic current with north­ easterly orientation and originating from the Arabian high pressure has followed an anticyclonic path around the eastern and southern side of the Arabian high and, as a result, has a sea trajectory of only small length. The air may be moist, but the air mass is subsident and stable aloft. The dry subsident air streams moving southward result in dry weather during the period October to May. The thermal low to the south-east of the Ethiopian highlands remains stationary during this time, but moves north ward when it is associated with Inter Tropical

Convergence Zone (I.T.C.Z.) (G riffith s 1972).

Rainfall and its erosivity in Ethiopia 22

—* Isobaric flow -♦ Warm dry air stream Base map: TATO (1964) j Central highlands Draft: M. Osman — International boundary

L Low I. T. C. Z. Inter Tropical Convergence Zone

| Ethiopian capital

Fig. 5b: Major disturbances in air stream over Ethiopia during March - May, penetrating south

Seasonal migration of I.T.C.Z. over Ethiopia

The climate of Ethiopia is highly influenced by seasonal migration of the I.T.C.Z., which is in turn governed by annual variation in the declination of the

Rainfall and its erosivity in Ethiopia 23

sun and the associated shift of the main pressure system (H urni 1982; K uhnel 1983). Investigations of the Ethiopian National Meteorological Service Agency (NMSA 1996) indicate that, from January to June, the I.T.C.Z. moves in a northward direction. One of the largest excursions of the I.T.C.Z. occurs along longitude 35° East. The I.T.C.Z. is at its most northerly position in July and August at about 16 to 20° North (see Fig. 4). At this time, most parts of Ethiopia receive rainfall. From September to November, the I.T.C.Z. migrates southward at a faster rate than to the north. This marks the end of the rainy season. During December, the I.T.C.Z. shifts southwards outside Ethiopia, over the equator to the tropic of Capricorn, reaching a mean low position of 10° South in January,

and back towards the sphere of influence over the country (T ato 1964). The migration pattern of the I.T.C.Z. is temporarily interrupted or even reversed and

exhibits an irregular feature from one year to another. T ato (1964) reported that great fluctuation in the pattern of movement was observed not only periodically but also diumally. During its north and south excursions, the I.T.C.Z. is distorted over the Ethiopian highlands due to the effect of terrain and topographic barrier. It has no uniform speed and occasionally seems to be stationary over an area. In addition, some active parts appear to jump rather than move steadily from one position to another.

3. 2 Characteristics of long-term rainfall in Ethiopia

The lowest multi-year mean rainfall in Ethiopia was noted in Elidar (N = 11), with multi-year mean of 117.30 mm, standard deviation of 86.00 mm and a variability coefficient of 73% (see annex, Tab. 19). The highest multi-year mean rainfall in Ethiopia during the years of investigation was recorded for Nekemte (N = 11), with long-term mean rainfall of 2128.73 mm, standard deviation of 210.90 mm and a variability coefficient of 10%. The data show that the least long-term rainfall variability coefficient was noted for Nekemte (10%) and the highest for Bahir Dar (N = 36, coefficient of variation 172%) weather station. The maximum annual rainfall record received by the Elidar weather station (235.00 mm), marks the least value observed throughout Ethiopia for the period

Rainfall and its erosivity in Ethiopia 24 of observation. The highest annual rainfall in Ethiopia was recorded for Deke Stefanos weather station, with an annual value of 4821.60 mm.

Fig. 6: Isohyets of long-term mean annual rainfall (mm) over Ethiopia

Annual rainfall in Ethiopia tends to decrease from west to east and from south to north (Fig. 6). The western parts of Ethiopia and the central highlands generally receive a high amount of rainfall, as compared to other areas of the country. The seasonal distribution of rainfall in Ethiopia, as described on the basis of synoptic weather stations, varies strongly in different parts of the country. Among other things, the seasonal migration of the I.T.C.Z. and local topographic characteristics are the dominant factors influencing rainfall distribution over Ethiopia ( T a t o 1964; GRIFFITHS 1972). Fig. 7 shows the long-term monthly distribution of precipitation for selected weather stations in Ethiopia.

Rainfall and its erosivity in Ethiopia 25

350 j_ _ 225 -j -S' 300 -: Addis Ababa rFj y 200 -=■ ' Awassal 1 250 o 175 -=■ 200 150-1- | * 125 • | 150 .a i e 100 -=■ £ 75 -s- $ 100 * • V I 50 .L...... { • 50 -r 0 . = p l i , t , 1. .L. ^aqpc 25 -i 0 - L M I t | | % I* I* s s - Months Months

500 250 T * 200 Debre Zeit - 1- Bahir Darj i 150 *- 100 50 4 - _i_ XX 0 2? +■* > l ! i J S I ^ £ .9 14 h w Z 3a a # I Months Months

300 150 Dire Dawal & 250 - r.... Gambelaj" 125 £ f t

350 300 =3 250 t?1 200 S s 150 100 9 50 S 0 & ■ s ■ GO& Months Months

Fig. 7: Seasonal rainfall distribution [±5%] for selected weather stations in Ethiopia (1898-1997) Rainfall and its erosivity in Ethiopia 28

Fig. 8: Long-term seasonal rainfall distribution data series for Ethiopia (1898-1997)

As can be revealed from Fig. 8, the major proportion of precipitation (more than 65%) in Ethiopia occurs during summer and autumn, of which nearly 80% fall only in the months July and August, with relatively low variability coefficient (ca. 50%). Rainfall in June is marked by the highest relative variability (84%) and in April by the lowest (40%). The general tendency is that, the drier the months, the higher their relative rainfall variability and vice versa.

The long-term areal precipitation time series data are normally distributed at a significance level of 5%. The assessment of the relationship between long-term mean precipitation of synoptic weather stations and the areal mean rainfall data indicated that there is a good relationship between them. The weather station Gode, with a correlation coefficient of -0.24, was found to be an exception to the

Rainfall and its erosivity in Ethiopia 29 general tendency. The correlation coefficients for the other weather stations were noted to be statistically highly significant at 1% level, with correlation coefficients ranging from 0.72 to 0.97. The coefficient for Gode (r = -0.24) indicates that the Gode precipitation time series is negatively correlated with that for the whole of Ethiopia. This can be explained by the variation between the rainy seasons of most of the selected weather stations and that of Gode.

According to T a to (1964) and K o u sk y et al. (1998), this variability in the rainy season is explained by geographic location of the weather stations and the general atmospheric circulation pattern over the area. As can be revealed from Fig. 7, the Gode weather station remains dry during the major rainy period of the other areas in Ethiopia. It can be inferred from the results of correlation analysis between precipitation data of the selected weather stations and the generalised rainfall time series data for the whole of Ethiopia that the areal precipitation time series highly represents these weather stations. Hence, the applicability of its analytical results and interpretations regarding characteristics of rainfall (see

Ra pp and Sc h On w iese 1996). However, inference from investigation results of one to many time series should be made cautiously, since topographic situations and terrain characteristics, which are not further considered within the scope of this study, can play an important role in influencing seasonal distribution patterns of precipitation. It can be concluded that, since the areal precipitation time series (1898-1997) developed from the synoptic weather stations throughout Ethiopia well represents the data for selected weather stations, it can be meticulously used for regional strategic land and water resources management planning. Furthermore, this is applicable to agricultural development programming and advance preparedness against disaster (e.g., extreme drought and flood events) and its prevention in the country.

3. 3 Estimated values of long-term rainfall erosivity for Ethiopia

The least long-term mean rainfall erosivity in Ethiopia was found for Elidar weather station (see annex, Tab. 20). Accordingly, a long-term average rainfall erosivity value of 30.00 mm was calculated for the Elidar weather station (N = 11), with standard deviation of 17.22 mm and a variability coefficient of 57%. A

Rainfall and its erosivity in Ethiopia 32

L egend

Central highlands - Boundary of rainfall pattern regions

Base map: FAO (1984) - International boundary Draft: M. Osman A-E Rainfall pattern regions Ethiopian capital______

Fig. 9: Modified rainfall pattern regions of Ethiopia, FAO (1984)

Description of the FAO (1984) rainfall patter regions:

A One short rainy season in summer B One long rainy season with dry winter C One long rainy season with rainfall peaks in spring and summer to autumn, separated by a season with less rainfall D Two short rainy seasons, main rains in spring, minor rains in autumn E Minor rains in spring, major rains in summer

Rainfall and its erosivity in Ethiopia 33

Tab. 2: Statistical parameters of long-term rainfall and its erosivity for the FAO (1984) rainfall pattern regions in Ethiopia

Rainfall Rainfall erosivity Statistical AB C DEA B C DE parameters N 34 15 39 20 60 34 15 39 20 60 Mean 1145 1246 1036 792 911 249 205 160 144 163 Median 1120 1120 913 830 939 228 197 150 144 160 Standard deviation 393 341 342 304 300 82 52 42 42 49 Minimum 143 783 646 284 117 116 149 110 60 30 Maximum 1867 2129 1946 1276 1604 498 324 265 231 362 Variability coefficient (%) 34 27 33 38 33 33 26 26 29 30 Kurtosis 0.24 2.01 0.71 -0.79 0.84 1.25 0.26 -0.03 -0.14 0.33 Skewness -0.16 1.27 1.14 -0.41 -0.24 0.99 1.01 0.87 0.00 -0.23

The long-term mean rainfall in various rainfall pattern regions (FAO 1984) exhibits large differences from one pattern region to another (Tab. 2). The highest mean precipitation (1246 mm) was recorded for the rainfall pattern region B (N = 15), with standard deviation of 341 mm and a variability coefficient of 27%. The minimum long-term mean rainfall (792 mm) was observed in region D (N = 20), with standard deviation of 304 mm and a variability coefficient of 38%. Of all the rainfall pattern regions in Ethiopia, long-term rainfall in region B is characterised by the least variability (27%) and in region D by the highest (38%). It is evident from the statistical parameters of rainfall in Tab. 2 that lower precipitation is characterised by higher relative variability in rainfall and vice versa.

The long-term mean rainfall erosivity, like precipitation, varies considerably in different rainfall pattern regions. Accordingly, the highest long-term areal average rainfall erosivity for the rainfall pattern regions (249 mm) was determined for region A (N = 34), with standard deviation of 82 mm and a variability coefficient of 33%. The lowest long-term areal average rainfall

Rainfall and its erosivity in Ethiopia 36

Long-term mean annual rainfall (mm)

Fig. 12: Relationship between precipitation and rainfall erosivity in rainfall pattern region C

Y =258.69533-0.95228 X + 0.00156 X2-6.87326E-7 XJ

Long-term mean annual rainfall (mm)

■ Mean annual Fm data point ------Polynomial fit

Fig. 13: Relationship between precipitation and rainfall Erosivity in rainfall pattern region D

Rainfall and its erosivity in Ethiopia 37

Fig. 14: Relationship between precipitation and rainfall erosivity in rainfall pattern region E

It was observed that 50% of rainfall erosivity in rainfall pattern region A, 67% in region B, 84% in region C, 80% in regions D and E, respectively, are explained by long-term areal mean precipitation of the respective area (see Figs. 10 to 14). It can be confirmed from the results of regression analysis that there exists fair to very good relationship between long-term areal mean rainfall and its erosivity for the rainfall pattern regions in Ethiopia. Hence, these regression models are considered to be good prediction tools of rainfall erosivity in each region because they are selected based on best model criteria (see A nsco m be

1973; Ja n ssen and La a tz 1997; V o elk el and G erber 1999). Therefore, these regression models are recommended for use in Ethiopia to forecast the potential risk of water erosion towards the soils of these rainfall pattern regions and, especially, to devise land and water resources management plans and strategies at zonal scale. Moreover, these models are readily applicable to meso-scale and regional land and water management policy design and the making of pertinent decisions. In order to apply these models to local level soil and water conservation efforts and to advise land-users, the respective local climatic situations and land-use practices should be carefully taken into account.

Rainfall and its erosivity in Ethiopia 38

4 LONG-TERM RAINFALL VARIABILITY IN THE CENTRAL HIGHLANDS OF ETHIOPIA

The whole region of east and north east Africa, including Ethiopia, has had a long history of drought in the last 1 100 years (e.g., V er sc h u re n et al. 2000). The consequent calamities, thanks to the Ethiopian Synaxarium or Book of Saints, can be traced back to early medieval times. These phenomena could be identified for many centuries through the Ethiopian royal chronicles, lives of Saints and foreign travel literature. Detailed studies on recent agricultural production shortfalls due to rainfall variability in the central highlands of

Ethiopia were carried out by several researchers (e.g., W o l d e -M ariam 1984;

D egefu 1987). H urni (1993) describes climate change and variability as one of the dominant natural elements causing vulnerability to famine in Ethiopia. Shortage of precipitation and its space-time variability in Ethiopia has already led to recurrent and substantial shortfall in agricultural production, which took thousands of human and animal lives. A serious reduction of rainfall over large parts of Ethiopia in 1984/1985 affected almost ten million people and caused enormous damage to the socio-economy and ecology of the country (W o l d e-

M ariam 1984). During these years, Ethiopia suffered significant production shortfall of about 20% in the agricultural sector; resulting in a decrease of total annual production by about one million tonnes, mainly involving cereals and pulses (H urni 1993). The central highlands of Ethiopia, which play a great role in the country’s agricultural economy, have suffered the most, compared with other parts of the Ethiopian highlands. In spite of these experiences, ecological impacts of the past climatic variability in Ethiopia and its consequences, especially change in land-use system and land management practices, were not sufficiently examined. Development strategies focusing on resource management as a means to attain sustainable self-reliance failed to take into account the investigation of the effect of long-term climatic change on this macro objective. The primary reasons for this are the lack of database as well as preliminary studies on long-term rainfall variability. Therefore, the filling of this information gap is essential for future activities in agricultural development planning and ecosystems management. Figs. 15 and 16 illustrate histograms of

Rainfall and its erosivity in Ethiopia 39 long-term areal annual and summer rainfall data distribution (1898-1997) used in this study to investigate precipitation variability in the central highlands of Ethiopia during the last decade of the 19lh century and during the whole period of the 20Ih century.

G 3 O U

Std. Dev = 240.30 Mean =1148 N = 99

Annual rainfall (mm)

Fig. 15: Long-term annual rainfall distribution in the central highlands of Ethiopia

Rainfall and its erosivity in Ethiopia 40

c o3 U

Std. Dev = 162.07 Mean = 666 N = 99

Summer rainfall (mm)

Fig. 16: Long-term summer rainfall distribution in the central highlands of Ethiopia

The central highlands of Ethiopia received a long-term average annual (1 148.00 mm) and summer (666.00 mm) rainfall in the period of observation, with standard deviations of 240.30 mm and 162.07 mm, respectively (Figs. 15 and 16). A variability coefficient of 21% was calculated for long-term annual rainfall data, and 24% for that of summer. The minimum recorded annual rainfall amount was 690.00 mm and the maximum 2025.00 mm, thus resulting in an extreme average of 668.00 mm (see P a n o fs k y and B r ie r 1958; Schonwiese 1985). Whereas, a minimum amount of 354.70 mm and a maximum of 1593.48 mm were recorded for the summer period, with an extreme average of 619.39 mm. A skewness coefficient of 1.34 and a kurtosis coefficient of 2.60 were determined for long-term annual rainfall. For the long-term summer rainfall, a skewness coefficient of 2.30 and a kurtosis coefficient of 11.00 were calculated. In addition, the result of normality test for the distribution type of the long-term rainfall data indicated that both long-term annual and summer rainfall data of the

Rainfall and its erosivity in Ethiopia 41 central Ethiopian highlands are approximately normally distributed at a significance level of 5%. These results justify the applicability of parametric statistical methods to representativeness analysis of long-term areal rainfall data for the central highlands of Ethiopia. The average annual and summer rainfall in the region is relatively higher, compared with most of the values recorded for the synoptic weather stations throughout the country, but the coefficients of variation are comparatively lower (see Fig. 7).

4.1 Representativeness analysis

The representativeness of a regional precipitation data series shows the extent to which the areal data represents records and trends at individual weather stations (R app and SchOnwiese 1996). Following the results from the analysis of distribution types for annual and summer rainfall data in the central highlands of Ethiopia, the use of a parametric statistical method can be justified to investigate the representativeness of areal rainfall time series data for the whole region (see R app and Schonwiese 1996; M ik e ls e n et al. 1998). The following results were obtained according to the correlation analysis made between the long-term rainfall data series of each selected weather station and the long-term areal rainfall time series of the central Ethiopian highlands for both annual and summer duration. For the annual period, the best correlation coefficient was obtained for Addis Ababa weather station, with a correlation coefficient of 0.85. A good correlation was obtained for Majete and Tulu Bolo weather stations, each with a correlation coefficient of 0.60, and fair relation was obtained for Fiche and Kombolcha weather stations, each with a correlation coefficient of 0.55. The correlation coefficients for the remaining weather stations lie between 0.20 and 0.41. It can be inferred from these results that the Addis Ababa, Majete and Tulu Bolo weather stations are significantly represented at 5% level. All other weather stations are weakly to fairly represented.

The highest coefficients obtained from similar analysis of summer rainfall time series data were noted for Addis Ababa weather station, with a correlation coefficient of 0.86. Good correlation was determined for Debre Zeit, Fiche and

Rainfall and its erosivity in Ethiopia 42

Kombolcha weather stations (each with a value of 0.60) and fair correlation was obtained for Sheno, Tulu Bolo and Wonji weather stations (each with a value of 0.55). The coefficients for the remaining weather stations lie between 0.20 and 0.45. It was confirmed from these results that significant relationship at 5% level was obtained for Addis Ababa, Debre Zeit, Fiche and Kombolcha weather stations. The remaining weather stations are weakly to fairly represented.

It can be deduced from the results of representativeness analysis that, the longer a series is, the stronger it is represented by the regionalised rainfall time series data. This implies that, with increasing time series length and improved database development and management, the degree of representativeness can be substantially enhanced. Despite the weak representativeness of the areal rainfall time series for some of the selected weather stations in the central highlands of Ethiopia, it is the only available and useful data to investigate rainfall variability and its erosivity at regional scale. Moreover, these data can be used in water resource management and development planning (e.g., M ik e ls e n et al. 1998) to improve agricultural production and ecological conservation at regional scale.

4. 2 Time series analysis of rainfall

Climatic elements, for instance precipitation, are normally subject to temporal and spatial variability and fluctuation ( M itc h e l l et al. 1966; G o o s s e n s and B e r g e r 1986). The variability in rainfall can range from relatively short-term (2 to 3.5 years cycle) to long-term (10 or more years cycle). In addition, a variability of 100 years cycle, referred to as secular or instrumental climate change, can occur with significant impact on the global climate affecting the whole ecosystem, especially the hydrosphere and the biosphere (R app and Schonwiese 1996). Temporal scales of global as well as local climate change and their corresponding nomenclature are thoroughly discussed by some authors (e.g., M i tc h e l l et al. 1966; R isb ey et al. 1999). The investigation of rainfall variability and its fluctuation is essentially as well as practically a problem of time series analysis (e.g., M a h e r a s 1990; C e b u la k 1997; Pavlopoulos and

Rainfall and its erosivity in Ethiopia 43

G ritsis 1999). A time series approach to the investigation of rainfall variability was applied to the precipitation data series of the central highlands of Ethiopia.

4. 2. I Graphical analysis

Rainfall in the study area has followed clearly decreasing observed trend in the period of investigation which spans from 1898 to 1997 (Figs. 17 and 18). The declining trend commenced in the year 1917 and continued with apparently progressive downward trend. However, phenomenal fluctuation was observed in the course of time, while the declining tendency of both annual and summer rainfall persisted unabated. The individual data points are relatively clustered along the smoothing curves, while still maintaining a general declining trend in the successive years. Long-term annual precipitation reached its utmost bottom in the years 1951 and 1984, where as the summer records reached their lowest point in the year 1951 and 1987, when the central highlands of Ethiopia suffered extremely serious rainfall deficit (e.g., S e le s h i and D e m a re e 1993; H u rn i 1993; M o r e d a and B a u w e n s 1998). As can be confirmed from the data, the year 1984 coincided with the largest drought disaster Ethiopia has seen in the 20!h century. Contrary to the general declining trend in rainfall noted from the 5- years and 11-years moving averages and their respective trend lines, of which only the 11-years trend lines are statistically significant for annual and summer period at 95% level of confidence, the central Ethiopian highlands have also seen years of extremely high rainfall amount. Instances for this are the years 1915 (annual) and 1922 (summer); a period with annual and summer precipitation surplus in the central highlands of Ethiopia. In addition, extreme highs in summer rainfall were also detected in 1916, 1922 and 1946. Despite the overall declining trend in rainfall amount, it was frequently reported that a series of flooding had inflicted huge ecological as well as socio-economic damage to the central highlands of Ethiopia (see UNDHA 1995).

Rainfall and its erosivity in Ethiopia 44

2000

| 1750 B

Ic 1500 c5 I 1250 B

I 1000 c00 o j 750

500

1898 1913 1928 1943 1958 1973 1988

Years

® Annual rainfall data point ------5-years running mean

------11-years running mean - - - Linear (5-years running mean)

“ “ ‘Linear (11-years running mean)

Fig. 17: Long-term annual rainfall time series data for the central highlands of Ethiopia (1898-1997)

Rainfall and its erosivity in Ethiopia 45

Fig. 18: Long-term summer rainfall time series data for the central highlands of Ethiopia (1898-1997)

The observed decline in precipitation in the central Ethiopian highlands during the study period was characterised by precipitation anomalies which might even have had long-term historical antecedents (e.g., V erschuren et al. 2000). Such anomalies would best be indicated by the deviation of annual and seasonal rainfall records from long-term average precipitation. Fig. 19 shows the departures of long-term annual and summer rainfall from their respective long­ term average.

Rainfall and its erosivity in Ethiopia 46

Fig. 19: Departures of long-term annual and summer rainfall from their respective long-term average in the central Ethiopian highlands (1898-1997)

It is apparent that the departures of long-term areal annual and summer precipitation from the corresponding long-term average were very low at the initial phase of the series records (Fig. 19). While deviations in this period were dominantly positive, pronounced departures of both long-term annual and summer precipitation were noted in the period between 1915 and 1917, with deviations from long-term average ranging from 45-66%. The summer rainfall variability in the year 1922 was situated far above the annual figure, with a deviation of 78% from the long-term average. The comparatively lower annual rainfall variability in this period might be due to the cancelling effect of rainfall deficit during the short rainy seasons of the respective years. This period was succeeded by years of consecutively diminishing rainfall variability which, however, did not persist over long duration. The years 1925 and 1926 were characterised by rainfall shortage in summer, which is the major rainy season in

Rainfall and its erosivity in Ethiopia 47 the central Ethiopian highlands. Accentuated precipitation extreme marked by high positive departure was noted in 1947, with annual and summer values higher than the corresponding long-term average by 70% and 46%, respectively. However, the year 1951 has seen the highest negative deviation from the long­ term average ever recorded in the central Ethiopian highlands. The occurrence of such high discrepancy between the positive and negative departures within a short interval of time has a negative implication on the dependability of expected variability. However, the data show that persistent recurrence of precipitation shortfalls in the central highlands of Ethiopia has prevailed since 1951, with major positive departures observed only in 1958, 1985, 1993 and 1996. Generally, the central Ethiopian highlands were predominantly characterised by positive rainfall departures from the long-term mean in the first and negative deviations in the second half of the 20th century. The positive departures observed in the first half of the 20th century are highly pronounced in the first three decades. It was noted that the second half of the century has predominantly suffered negative rainfall deviations, with both annual and summer values frequently lower than the long-term average. This statement agrees with the results found by S e le s h i and D e m a re e (1993, 1995) for north Ethiopian and Eritrean highlands. The authors also reported that there was high concentration of meteorological drought in Addis Ababa in the period from 1948 to 1973. High frequency of drought occurrence in the second half of this century in Ethiopia was also reported (W o o d 1977; T e s fa y e 1990). The authors recorded meteorological, hydrological as well as agricultural drought in Ethiopia in the second half of the 20th century. It was revealed that only since the early 1990s have the central Ethiopian highlands experienced successive years with rainfall higher than long-term average. This might have had considerably influenced the surface flow regime of major rivers springing from and flowing through the central highlands of Ethiopia.

4. 2. 2 Systematic trend analysis

The systematic trend analysis of rainfall time series data in the central highlands of Ethiopia enhances the knowledge of temporal precipitation behaviour in the

Rainfall and its erosivity in Ethiopia 48 region by means of statistical evidence. Moreover, the observed trend of rainfall time series is considered bona fide only when it is corroborated by statistical tests ( M itc h e l l et al. 1966). The result of statistical time series analysis of long-term areal annual and summer rainfall time series data in the central Ethiopian highlands is consistent with that of graphical analysis presented by trend diagrams (Figs. 17 and 18). Whilst a trend correlation coefficient of -0.50 was obtained for the annual rainfall time series data, a value of -0.41 was determined for the summer series, thus showing a decreasing trend. It was noted that the declining trends in both long-term annual and summer rainfall in the central highlands of Ethiopia are statistically significant at 1% level, each with a probability value of 0.00. The trend correlation coefficients calculated for rainfall time series data of most of the selected weather stations in the central highlands of Ethiopia also show a decreasing trend in rainfall in respective areas during the period of investigation. However, the strength and degree of decline in the trend for these weather stations vary seasonally and spatially. The trend correlation coefficients of the selected weather stations used to compare with the statistical trend of the areal rainfall time series data are given in Fig. 20.

Fig. 20: Correlation coefficients for precipitation time series data of selected weather stations in the central highlands of Ethiopia

Rainfall and its erosivity in Ethiopia 49

Annual and summer rainfall time series data at Debre Zeit and Ejaji weather stations, contrary to the series of the other stations and the areal rainfall time series, exhibit a positive trend; thus indicating a concurrent increase in annual and summer rainfall records in these areas (Fig. 20). However, the increasing trend in rainfall in the two areas is not statistically significant. As is evident from Fig. 20, a decreasing trend in both annual and summer rainfall was noted at the weather stations Addis Ababa, Debre Markos, Kombolcha, Sheno, Shola and Wonji. It was confirmed that, whilst the declining trend in the annual rainfall time series data was significant at a level of 5% only for Debre Markos and Kombolcha, located in north-western and north-eastern part of the central Ethiopian highlands, respectively; the summer falling trend was significant at 5% level only for Addis Ababa and Kombolcha. Hence, it is apparent that precipitation in Kombolcha area and its neighbourhood is obviously diminishing. Long-term annual and summer rainfall at the weather stations Fiche, Majete and Tulu Bolo show contrasting feature, with summer rainfall experiencing a decreasing trend in the former two and an increasing trend in the later. However, these trends are found to be statistically not significant.

Generally, it is evident from the analytical results of the areal precipitation time series data and the time series rainfall records of the selected weather stations that rainfall has decreased throughout the central highlands of Ethiopia during the 20th century. Therefore, water resource management measures and supporting policies should be thoroughly designed and strictly implemented in the central highlands of Ethiopia to tackle the challenges of rainfall deficit and meet the ever increasing demand of progressively growing population for water for survival.

4. 2. 3 Autocorrelation and persistence

The correlation of a set of observations within themselves is referred to as autocorrelation (Brockwell and D a v is 1996). With regard to this study, the autocorrelation coefficient provides a measure of temporal correlation between rainfall data points with different time lags; namely 1, 2, 3, ..., n years. Thus,

Rainfall and its erosivity in Ethiopia 50 autocorrelation provides initial information relevant to the internal organisation of each time series data (Essenwanger 1986; M eek et al. 1999). The prevalence of autocorrelation in a data series is also an indication of persistence in the series of observations. The auto-correlation coefficients provide an essential hint whether forecasting models can be developed based on the given data (J a n s s e n and L a a tz 1997). For a purely incidemal event, all autocorrelation coefficients are zero, apart from r (0) which is equal to 1. The results of autocorrelation analysis of the annual and summer rainfall time series data in the central Ethiopian highlands are presented in Figs. 21 and 22.

Cl, U <

Confidence Limits

^Coefficient 1 6 11 16 21 26

L ag N u m b er

Fig. 21: Autocorrelation of long-term areal annual rainfall in the central highlands of Ethiopia

Rainfall and its erosivity in Ethiopia 51

u« U <

Confidence Limits

^Coefficient

L ag N u m b er

Fig. 22: Autocorrelation of long-term areal summer rainfall in the central highlands of Ethiopia

The autocorrelation coefficients of annual rainfall in the central highlands of Ethiopia lie within the range of -0.09 to 0.42, with standard error of 0.08 and 0.09, respectively (Fig. 21). The autocorrelation coefficients of summer rainfall lie between -0.14 and 0.33, each with standard error of 0.09 (Fig. 22). As shown by the autocorrelogram (Fig. 21); the 1st, 2nd, 6th, 8th, 10th, 11*, 12th, 13lh, 19th, and 21st order lag autocorrelation coefficients of long-term areal annual rainfall exceeded the upper 0.25 confidence limit and, hence, are statistically significant (J a n s s e n and L a a t z 1997). For summer areal rainfall data series, the 1st, 6th and 11th order lag autocorrelation coefficients were found to be statistically significant at 0.25 confidence limits (Fig. 22). The fact that the ls! order lag autocorrelation coefficients for both annual and summer rainfall time series data significantly deviate from zero confirms the existence of persistence in the series at the respective order lag. At the same time, for both annual and summer rainfall data, the values of second and third order lag autocorrelation coefficients are greater than the square and the cubic values of the first order lag

Rainfall and its erosivity in Ethiopia 52 autocorrelation coefficients, respectively, i.e. r2 > r|2 and r3 > r|3. Hence, there is a clear indication of the persistence of the first order linear Markov process ( M itc h e l l et al. 1966).

In the absence of any reliable physical basis for predicting seasonal and/or annual climatic conditions, as the case most likely holds true for the Ethiopian climate data set, any assumptions about future climatic conditions and, especially, water resources, have to rest on the experiences of past occurrences (e.g., Meigh et al. 1999). Mostly, the characteristics of rainfall in the past are the only accessible guide for advance decision making regarding water resources development and management, particularly for an agrarian economy like that of Ethiopia. Similar procedures were recommended for Scotland (S m ith 1995). The author stressed the use of best normal periods with the least extrapolation variances for statements about precipitation conditions and, hence, water resources management. In spite of the fact that constant persistence in the observed precipitation series data was not found for the central highlands of Ethiopia, the above trend analysis would help the practitioners, specifically water managers and agricultural development planners, with their decision making process.

Rainfall and its erosivity in Ethiopia 53

5 RAINFALL EROSIVITY IN THE WHOLE CENTRAL HIGHLANDS OF ETHIOPIA

In order to make regional land and water management plans and policies, the assessment of rainfall erosivity and development of prediction tools for the corresponding area are needed. The following section deals with regional analysis of rainfall erosivity.

5. 1 Statistical characteristics of long-term areal annual and summer rainfall erosivity'

Long-term areal mean annual and summer rainfall erosivity figures of 207.60 mm (N = 100) and 259.35 mm (N = 100), with standard deviations (SD) of 33.58 mm and 47.04 mm, respectively, were determined for the whole central highlands of Ethiopia (Figs. 23 and 24). A minimum long-term areal mean rainfall erosivity value of 110.56 mm and a maximum of 370.55 mm were observed for the annual period, resulting in an extreme average value of 130 mm. For the summer period, a minimum value of 140.57 mm and a maximum of 439.10 mm, and hence an extreme average of 491.40 mm were determined. A variability coefficient of 16.20% was determined for long-term areal annual rainfall erosivity and a value of 18.14% was calculated for long-term areal summer rainfall erosivity. Figs. 23 and 24 illustrate the histograms of areal annual and summer rainfall erosivity in the central highlands of Ethiopia.

Rainfall and its erosivity in Ethiopia 54

Fig. 23: Histogram of long-term areal annual Fm (mm) for the central Ethiopian highlands

Fig. 24: Histogram of long-term areal summer Fm (mm) for the central Ethiopian highlands

Rainfall and its erosivity in Ethiopia 55

It is apparent from the histograms (Figs. 23 and 24) that long-term areal mean annual and summer rainfall erosivity in the central highlands of Ethiopia show similar statistical characteristics. This can be confirmed from the positive coefficients of skewness and kurtosis calculated for the data. Accordingly, skewness coefficients of 1.06 and 0.61 were determined for the long-term areal annual and summer rainfall erosivity, respectively. Kurtosis coefficients of 2.84 and 1.65 were calculated for the annual and summer erosivity, respectively. It can be inferred from these results that long-term areal annual and summer rainfall erosivity are positively skewed and are mesokurtic (H a w k in s and W e b b e r 1980). The results of normality test indicate that long-term areal rainfall erosivity in the study area approximate normal distribution. Subsequently, the use of parametric statistical methods to analyse representativeness of the areal rainfall erosivity time series data for selected weather stations is justified (R app and Schonwiese 1996). As noted from the results of descriptive statistical analysis, summer rainfall in the central highlands of Ethiopia is comparatively more erosive than annual precipitation, but subject to higher relative dispersion, as is evident from the corresponding variability coefficients. The fact that annual rainfall erosivity values are lower than summer values is due to the cancelling effect of less erosive precipitation during the dry season.

5. 2 Representativeness analysis

A representativeness analysis of rainfall erosivity data series of the whole central Ethiopian highlands was carried out for each of the selected weather stations in the region. It was revealed that the degree of representativeness of the areal long-term rainfall erosivity data series varies spatially as well as seasonally. For the long-term annual rainfall erosivity, the correlation coefficients range from 0.25 to 0.81. Wonji (r = 0.25), Debre Markos (r = 0.27), Kombolcha (r = 0.30), Shola (r = 0.32), Tulu Bolo (r = 0.35) and Ejaji (r = 0.42) weather stations are weakly represented by the long-term areal annual rainfall erosivity. In contrast, the Fiche weather station, with r = 0.60, is well represented. Very good representation coefficients were determined for the

Rainfall and its erosivity in Ethiopia 56

weather stations Sheno (r = 0.70), Debre Zeit (r = 0.70), Majete (r = 0.73) and Addis Ababa (r = 0.81).

For the long-term summer rainfall erosivity, the representativeness correlation coefficients range from 0.33 to 0.83. The long-term summer rainfall erosivity at weather stations of Debre Markos and Wonji, with correlation coefficients of 0.33 and 0.45, respectively, are weakly represented by the areal long-term summer rainfall erosivity data. In contrast to this, the weather stations Ejaji (r = 0.47), Tulu Bolo (r = 0.47), Majete (r = 0.50), Shola (r = 0.53) and Kombolcha (r = 0.55) are fairly represented. A good representation coefficient was calculated for Sheno and Debre Zeit weather stations, with r = 0.58 and 0.60, respectively. Very good representation was observed for the Fiche (r = 0.70) and Addis Ababa (r = 0.83) weather stations.

As can be confirmed from the results of correlation analysis, annual and summer rainfall erosivity in Wonji and Debre Markos areas are weakly represented by the corresponding regional long-term rainfall erosivity. The Addis Ababa weather station is strongly represented by both long-term annual and summer rainfall erosivity. The remaining weather stations are relatively fairly to well represented. It can be implied from the results of representativeness analysis that the tendency of representativeness of the annual and summer rainfall show clear similarity. Generally, it can be deduced from the results that areal summer rainfall erosivity has major implication, as compared with that of annual period in the region. This is evident from the relatively better representativeness of the long-term areal summer rainfall erosivity for most of the weather stations. Moreover, it was noted that the degree of representativeness of areal time series data varies from one locality to another. Subsequently, while the regionalised rainfall data can be used for land and water resources management planning and development policy making at regional scale, it should be supported by results of investigations in specific localities for meso-scale application. This practice would enhance the validation and practicability of regional data for micro-level resource management. Hence, in-depth regional and micro-scale rainfall erosivity analysis for the central highlands of Ethiopia is essential in this regard.

Rainfall and its erosivity in Ethiopia 57

5. 3 Time series analysis

5. 3. 1 Graphical time series analysis

Graphical time series analysis enables visual inspection of trend directions (see P a n o fs k y and B r ie r 1958; C e b u la k 1997). As illustrated by the trend diagrams (Figs. 25 and 26), long-term areal rainfall erosivity in the central Ethiopian highlands shows a very weakly decreasing trend for the annual and summer rainfall erosivity data series in the observation period which spans from 1898-1997. The 5-years and 11-years running mean curves reveal that, except for the year 1955, the long-term areal mean annual rainfall erosivity in the central highlands of Ethiopia has shown relatively similar characteristics in the course of the observation period (Fig. 25). The linear trend lines, with slopes of - 0.06 and -0.11 for 5-years and 11-years running means, respectively, show that long-term areal annual rainfall erosivity exhibited a very weakly declining trend. Long-term areal annual rainfall erosivity attained its highest in 1922, with a value of 370.56 mm, and its lowest in 1951, with a value of 110.56 mm. High annual rainfall erosivity values were also noted in the years 1915 (342.5 mm), 1947 (328.05 mm) and 1965 (338.75 mm).

As can be revealed from the smoothing curves of 5-years and 11-years running means (Fig. 26), long-term areal summer rainfall erosivity in the central Ethiopian highlands shows quite similar pattern to long-term areal annual erosivity series. The 5-years and 11-years linear trend lines, with slopes o f-0.05 and -0.11, respectively, confirm that the areal long-term summer rainfall erosivity in the central Ethiopian highlands followed a weakly declining trend during the period of observation (1898-1997). Long-term areal summer rainfall erosivity attained its peak in 1922, with a value of 439.09 mm and its lowest in 1951, with a value of 140.57 mm. Very high long-term areal summer rainfall erosivity values were also recorded in the years 1946 and 1947, with values of 367.81 mm and 409.93 mm, respectively.

Rainfall and its erosivity in Ethiopia 58

Generally, the central highlands of Ethiopia have seen both extreme highs and lows of annual and summer rainfall erosivity in the late 1910s, early 1920s and mid 1940s. These years were characterised by the occurrence of catastrophic water erosion events as well as extremely low water erosion within a period of only few years. The 1930s and 1970s were characterised by high concentration of relatively lower areal annual and summer rainfall erosivity, whereas the 1920s and 1960s were characterised by high concentration of relatively higher rainfall erosivity values. Currently, rainfall erosivity shows an increasing tendency, especially since the mid 1980s. The period of high concentration of highly erosive rainfalls might have led to enormous soil loss in the central highlands of Ethiopia, depending on land use practices and land cover of the area during these years. However, the period of low water erosion, which coincide with years of rainfall deficit in the region, might have been characterised by high wind erosion, as dry unprotected soil aggregate particles are susceptible to wind erosion (see D e x te r and Kroesbergen 1985; Le Bissonnais 1996; Fox and Le Bissonnais 1998).

Rainfall and its erosivity in Ethiopia 59

360

E 310 E uf s 260

210

160

110 1898 1908 1918 1928 1938 1948 1958 1968 1978 1988

Y ears

s Areal annual Fm data point ------5-years running mean

1 11 -years running mean - - * Linear (5-years running mean)

— - • Linear (11 -years running mean)

Fig. 25: Long-term areal annual time series of Fm in the central highlands of Ethiopia

Rainfall and its erosivity in Ethiopia 60

440

390 ? w 340 uF

| 290 3 V3 I 240 <

190

140 1898 1908 1918 1928 1938 1948 1958 1968 1978 1988 1998 Years

■ Areal summer Fm data point ------5-years running mean ------11 -y e a rs ru n n in g m e a n — - L in e a r (11 - y e a rs ru n n in g m e an ) ■ ■ • Linear (5-years running mean)

Fig. 26: Long-term areal summer time series of Fm in the central highlands of Ethiopia

5. 3. 2 Temporal variability of long-term areal mean annual and summer rainfall erosivity

The course of rainfall erosivity was compared with long-term areal average in order to investigate its temporal variability. Accordingly, areal annual and summer rainfall erosivity were found to be below the long-term average in the years 1890-1910. The 1920s-1950s predominantly experienced negative deviation of areal annual and summer rainfall erosivity from the corresponding long-term areal average (Fig. 27). The years 1922 and 1947 were exceptions, where annual and summer rainfall erosivity figures were higher than their respective long-term mean. Thus, annual rainfall erosivity was 67% higher than the long-term areal mean in 1922 and 48% higher in 1947. Summer rainfall

Rainfall and its erosivity in Ethiopia 61 erosivity was 61% and 50% higher than the long-term average in the years 1922 and 1947, respectively. Compared with the long-term average, the least erosive annual and summer rainfall in the central highlands of Ethiopia in the 20th century was noted in 1951, when annual and summer rainfall erosivity were 50% and 48% lower than their corresponding long-term averages. Very accentuated erosive annual rainfall was noted in the central highlands of Ethiopia in the years 1915, 1922, 1947 and 1965. Rainfall erosivity in the years 1970-1997 was generally lower than the long-term areal average, except for the year 1985 where both annual (+20%) and summer (+26%) rainfall erosivity were notably higher than their respective long-term areal averages. It can be concluded from the data that highly erosive rainfall events have more frequently occurred in the first half of the 20tn century than in the second half. However, the effect of erosive rainfall on the soils of the central highlands of Ethiopia in the second half of the 20th century was even more aggravated due to high pressure on land for cultivation and deforestation (W q ien 1995; L u p w ay i et al. 2000).

Rainfall and its erosivity in Ethiopia 62

Fig. 27: Comparison of temporal variability of long-term annual and summer rainfall erosivity in the central highlands of Ethiopia (1898-1997)

5. 3. 3 Systematic trend analysis, autocorrelation and persistence

The results of the systematic trend analysis of long-term areal annual and summer rainfall erosivity in the central highlands of Ethiopia agree considerably with those of graphical analysis (see Figs. 25 and 26). Both parametric and non- parametric statistical methods revealed that the erosive potential of rainfall in the central highlands of Ethiopia has followed slightly decreasing course during the period of investigation (1898-1997). Parametric trend correlation coefficients of -0.17 (P value 0.095) and -0.15 (P value 0.128); non-parametric trend correlation coefficients of -0.18 (P value 0.072) and -0.15 (P value 0.125) were determined for long-term areal annual and summer rainfall erosivity, respectively. It can be confirmed from the values of correlation coefficients that parametric and non-parametric statistical trend tests led to similar inference regarding the trend of rainfall erosivity in the central highlands of Ethiopia.

Rainfall and its erosivity in Ethiopia 63

Nevertheless, neither the parametric nor the non-parametric trend correlation coefficients of long-term areal annual and summer rainfall erosivity were found to be statistically significant at 5% level. It can be inferred from this statistical proof that the erosive potential of water in the central Ethiopian highlands has not been significantly decreasing even though precipitation has shown a significant decline in the observation period (see Figs. 17 and 18). In addition, it is worth noting that highly erosive rainfalls in the study area concentrated only in few years of the investigation period, and thereby caused ecologically and agriculturally catastrophic denudation in the area during the respective years.

Autocorrelation coefficients of long-term areal annual rainfall erosivity in the central highlands of Ethiopia range from -0.182 to 0.31, with standard errors of 0.091 and 0.099, respectively (Fig. 28). As is evident from the results, the first order lag autocorrelation, coefficient used to check whether the time series is purely incidental or not (M itch ell et al. 1966) significantly deviates from zero (P value 0.002) and, hence, is statistically significant at 1% level. It follows that there exists persistence (M itchell et al. 1966; E ssenw a n g er 1986;

B ro c k w e ll and D avis 1996) in long-term annual rainfall erosivity figures of the study period. However, no occurrence of first-order linear Markov process was detected.

The autocorrelation coefficients of long-term areal summer rainfall erosivity in the central highlands of Ethiopia range from -0.21 to 0.22 (Fig. 29), with standard errors of 0.083 and 0.093, respectively. The first order lag autocorrelation coefficient does not exceed the 95% confidence limits and, hence, is statistically not significant. The 95% confidence limits of autocorrelation coefficients were exceeded only at 11,h (r = 0.222, standard error = 0.093) and 29th (r = -0.21, standard error = 0.083) order lags.

The results of autocorrelation analysis suggest that the erosive potential of areal annual rainfall estimated for various years is significantly affected by that observed in the preceding years. Hence, significant deterministic relationship exists between the values of consecutive years (e.g., M eek et al. 1999). For areal

Rainfall and its erosivity in Ethiopia 64 summer rainfall erosivity, a value of a specific year is likely to be followed by a higher, lower or similar figure, thus indicating the existence of stochastic relationship between values of subsequent years (Bahrenberg et al. 1992). It is worth noting that, from practical point of view, the response of soil surface to the erosive energy of water is considerably influenced by the foregoing moisture condition of soil ( P o t r a t z et al. 1991; Fox and Le Bissonnais 1998; F o h r e r et al. 1999; T o r r i et al. 1999). These authors reported that the initial soil moisture and soil surface structure affect the erodibility of soil. Subsequently, dynamic soil loss prediction models should take this into account. Even though the erosive potential of summer rainfall in the study area does not show persistence, developing a water erosion prediction regression model for the summer period is essential for practical application to soil conservation planning and implementation in the region.

Confidence Limits

^Coefficient

L ag N u m b e r

Fig. 28: Autocorrelation of long-term areal annual rainfall erosivity in the central highlands of Ethiopia

Rainfall and its erosivity in Ethiopia 65

u- U <

Confidence Limits

^Coefficient

Lag Number

Fig. 29: Autocorrelation of long-term areal summer rainfall erosivity in the central Ethiopian highlands

5.4 Relationship between areal rainfall and rainfall erosivity in the whole central highlands of Ethiopia

The relationship between long-term areal rainfall and its erosive potential in the central highlands of Ethiopia varies for annual and summer period. The best regression models for the two periods were found to be linear relationships (Figs. 30 and 31). Accordingly, 50% of long-term annual and 80% of summer rainfall erosivity potential in the study area are explained by the corresponding long-term areal rainfall. As can be visually detected from Fig. 30, most of the areal annual data points are situated beyond the 95% confidence limits. However, almost all of the observations lie within the 95% prediction limits of the regression model. Hence, this model can be used for the estimation of regional annual rainfall erosivity. It can be noted from Fig. 31 that most of the summer rainfall erosivity data points are well clustered along the 95% confidence limits. Furthermore, almost all of the observations lie within the 95%

Rainfall and its erosivity in Ethiopia 66 prediction limits. Subsequently, the linear regression model of rainfall erosivity during summer season is a good predictor of water erosion risk in the study area. Figs. 30 and 31 show the relationship between long-term areal annual and summer rainfall erosivity in the region.

Rainfall and its erosivity in Ethiopia 67

Long-term annual rainfall (mm)

Fig. 30: Relationship between long-term areal annual rainfall and its erosivity in the central highlands of Ethiopia

Fig. 31: Relationship between long-term areal summer rainfall and its erosivity in the central highlands of Ethiopia

Rainfall and its erosivity in Ethiopia 68

According to best regression model criteria, the models fit for long-term areal annual and summer rainfall erosivity factors fulfil statistical conditions of a good prediction model (see A n sc o m b e 1 9 7 3 ; J a n ssen and L a a tz 1 9 9 7 ). Moreover, the models agree to a large extent with theoretical and practical concept of rainfall - rainfall erosivity relationship. According to W isch m eier and S m ith ( 1 9 7 8 ), rainfall erosivity is, among other variables such as wind, positively associated with rainfall amount which, in turn, influences the kinetic energy of rainfall. Positive relationship between long-term areal rainfall and rainfall erosivity in the central highlands of Ethiopia can be observed from the slopes of the regression curves (annual 0 .1 3 , sum m er 0 .3 3 ). Therefore, the models are appropriate to predict regional rainfall erosivity in the study area, under conditions of constrained data availability. While applying these models, due consideration should be given to the factors resulting in unexplained component of variation. As is evident from the determination coefficients of rainfall erosivity regression models and their respective adjusted coefficients of determination, annual 5 0 % and summer 8 0 % , the long-term areal annual rainfall erosivity regression model is a comparatively weaker predictor than the summer regression model. Regarding the application of the regional models to land and water resource management planning and policy decision making in specific localities, the models should be carefully adapted to the micro-climate and land- use systems of the respective area before they are fully applied. The primary reason for this is that, in contrast to the regional models, the local-specific erosivity models are expected to be empirically and spatially more representative for the area. Consequently, local condition based models are more reliable for area-specific land and water resources management planning as well practical implementation.

5. 5 The isoerodent contours for the central highlands of Ethiopia

The isoerodent contours for the central highlands of Ethiopia were developed using the rainfall erosivity regression models and the rainfall distribution map of Ethiopia. The purpose was to present the distribution of rainfall erosivity potential spatially for the whole central highlands of Ethiopia. It was found that

Rainfall and its erosivity in Ethiopia 69 the spatial presentation of rainfall erosivity is highly influenced by spatial rainfall distribution and seasonal rainfall variability. It was noted that the iso­ erodent contours of rainfall erosivity in the central highlands of Ethiopia are underestimated due to the normalised long-term rainfall isohyet contours which also underrated rainfall at each weather station in the area. Figs. 32 and 33 illustrate the annual and summer isoerodent contours of the area.

Fig. 32: Annual isoerodent contours for the central highlands of Ethiopia

Rainfall and its erosivity in Ethiopia 70

Fig. 33: Summer isoerodent contours for the central highlands of Ethiopia

It can be noted from Figs. 32 and 33 that both long-term annual and summer rainfall erosivity in the central highlands of Ethiopia tend to increase from east to west and from south to north. Compared to long-term annual rainfall erosivity, long-term summer rainfall erosivity depicts vivid irregular features. This is a reflection of the spatial precipitation pattern in the area which is, in turn, governed by topographic characteristics and circulation patterns.

Rainfall and its erosivity in Ethiopia 71

6 RAINFALL EROSIVITY AT SELECTED WEATHER STATIONS IN THE CENTRAL HIGHLANDS OF ETHIOPIA

The investigation results of rainfall erosivity for the whole central highlands of Ethiopia, as dealt with in the preceding part, enable to develop land and water management policies, strategies and implementation plans, mainly at regional scale; namely, the central highlands of Ethiopia. However, when the primary concern is land and water resources management at micro scale; for example, meso-scale soil and water conservation or watershed development at community level, it is important to establish the plans on the basis of local-specific assessment of rainfall and its erosivity. Important justifications for this are: 1. Local climatic conditions influence not only rainfall and its erosivity, but also land and water management practices pursued as a response to water erosion. 2. Regional models are less representative of concrete realities at local level. Problems related to the use of point and regional data for water resource management are discussed in the literature (e.g., E i n f a l t et al. 1998; M ik e ls o n et al. 1998). 3. Results of local-specific studies are directly applicable at on-farm scale and practically useful, especially to work with resource-users on issues of sustainable environmental management. 4. Point approach enables not only the exploitation of local knowledge and experience accumulated over many years, but also enhances the validation, calibration and adaptation of regional as well as point specific rainfall

erosivity models. For instance, F a v is-M o r tlo c k and G u e r r a (1999),

Sa u erbo r n et al. (1999), when attempting to stress the importance of point approach to water erosion research, have downscaled the General Circulation Model (GCM) for the purpose of investigating local rainfall erosivity. In light of these facts, this investigation was carried out for selected weather stations throughout the central Ethiopian highlands for various seasons of the year.

Rainfall and its erosivity in Ethiopia 72

6. 1 Long-term annual rainfall erosivity

6. I. 1 Statistical parameters o f annual rainfall erosivity

Long-term mean annual rainfall erosivity in the central highlands of Ethiopia varies considerably from one locality to another (Tab. 3). The least erosive annual rainfall in the region was determined for Asasa weather station (N = 17). A long-term average rainfall erosivity value of 106.00 mm (mean rainfall 571.00 mm), with standard deviation of 34.00 mm and variability coefficient of 32% was estimated for the Asasa area. The highest multi-year mean rainfall erosivity in the study area was found at Nekemte weather station (N = 11). A mean rainfall erosivity value of 323.73 mm (mean rainfall 2128.73 mm), with standard deviation of 31.20 mm and variability coefficient of 10%, was estimated for the Nekemte locality. The skewness coefficients of long-term annual rainfall erosivity data of the investigated weather stations range from -2.00, determined for Kore, to 3.96, calculated for Awassa weather station, thus showing a wide range of variation in skewness coefficients in various areas. The long-term annual rainfall erosivity data in Adaba, Alem Ketema, Asela, Ejaji, Fiche, Hosana, Idoto, Kore, Munessa and Sagure weather stations are negatively and in all other selected weather stations positively skewed (see B a h ren ber g et al. 1999). The kurtosis coefficients of long-term annual rainfall erosivity data of the selected weather stations in the central Ethiopian highlands range from -1.21, calculated for Gobessa, to 15.83, determined for Awassa weather station. The long-term annual rainfall erosivity data in Albuco, Awassa, Huruta, Kore, Seru, Shola Gebeya and Tulu Bolo weather stations is extremely high-peaked (leptokurtic); in Addis Ababa, Artuma and Asasa weather stations normally peaked (mesokurtic) and in all the remaining areas extremely low-peaked

(platykurtic) (compare H aw k in s and W eber 1980). As can be noted from the descriptive statistics of the data (Tab. 3), long-term annual rainfall erosivity in most of the localities in the central highlands of Ethiopia is approximately normally distributed. Tab. 3 depicts details of basic statistical parameters of long-term annual rainfall erosivity in the central highlands of Ethiopia (see also annex, Tab. 20).

Rainfall and its erosivity in Ethiopia 73

Tab. 3: Statistical parameters of long-term annual rainfall erosivity Coefficients Mean annual W eath er N M ean S tandard Variability Kurtosis Skewness rainfall S tations (m m ) Deviation (mm) (%) (m m ) A d ab a 20 132.02 27.35 21.00 -0.10 -0.16 725.01 Addis Ababa 99 207.64 40.92 20.00 2 .9 4 1.19 1208.95 A g arfa 20 149.63 97.80 65.00 3.33 1.98 1082.00 Albuco 18 223.52 160.45 72.00 12.62 3.30 1048.35 Alem Ketema 25 243.28 80.92 33.00 1.00 -0.77 1010.72 A n k o b er 21 252.47 58.84 23.00 1.13 0.54 1603.65 Artuma 10 252.61 80.10 32.00 2.91 1.26 1584.70 A sasa 17 106.00 34.00 32.00 3 .00 1.00 571.00 Asela 23 176.60 32.03 18.00 -0.63 -0.01 1102.20 A w assa 16 175.08 217.76 124.00 15.83 3.96 1015.10 B ora 13 149.07 39.40 2 6.00 2.31 1.21 815.60 B oru M eda 20 206.00 59.00 2 9.00 1.82 0.74 1117.75 Debre Markos 38 224.67 32.45 14.00 0 .05 0.60 1334.50 Debre Work 13 178.48 54.70 31.00 1.44 1.23 981.20 Debre Zeit 34 171.66 52.82 31.00 0.33 0.41 845.51 Dessie 26 203.28 47.08 2 3.00 -0.2 0 0.04 1107.00 D hera 21 118.08 31.60 27.00 0.30 0.31 616.80 E jaji 32 171.45 36.74 21.00 0.63 -0 .7 0 962.99 F iche 44 256.98 106.13 41.00 -0.38 -0.36 979.97 G essera 19 182.06 71.09 39.00 -0.15 0.56 1225.57 G oba 23 115.55 19.79 17.00 -0.28 0.20 916.51 G o b essa 17 149.12 28.00 19.00 - 1.21 0 .26 1105.95 Gohastion 26 220.46 42.22 19.00 -0.80 0.43 1118.60 H osana 16 153.60 38.20 25.00 0.0 2 -0.60 1048.06 Huruta 14 134.71 43.10 3 2.00 3.52 1.40 617.96 Idoto 17 112.40 18.91 17.00 -0 .1 8 -0.24 810.80 K achise 11 2 7 7 .8 0 52.80 19.00 -0 .5 0 0.50 1470.01 Keleta 25 118.00 24.00 20.00 0.00 1.00 692.00 K ofole 27 138.00 18.00 13.00 0.00 0.00 1016.00 K o m b o lch a 40 187.03 50.10 27.00 -0.10 0.50 1015.00 K ore 17 137.00 32.00 23.00 5.0 0 - 2.00 1095.00 M ajete 35 196.55 54.83 28.00 -0 .8 6 0.13 1092.58 Mararo 21 162.00 60.00 37.00 - 1.00 0.00 1107.00 M u n essa 10 176.13 22.23 13.00 1.38 -0.72 1307.80 N ek em te 11 323.73 31.20 10.00 1.80 1.35 2128.73 R obi 9 121.00 26.00 21.00 0.00 1.00 865.00 S agure 11 116.00 35.00 30.00 1.00 -1.00 723.00 S ed ik a 18 151.70 37.50 25.00 -0 .6 2 0.70 9 7 7.00 Seru 17 210.00 79.00 38.00 4 .0 0 2.00 1434.00 Sheno 34 208.18 66.16 32.00 1.85 0 .90 936.23 Shola Gebeya 35 206.28 69.30 34.00 15.13 3 .34 946.12 Sire 17 136.00 45 .0 0 33.00 1.00 0.00 652.00 Tulu Bolo 35 2 0 9 .2 0 74.40 36.00 3 .23 1.54 1063.42 W onji 34 147.57 48.06 33.00 -0 .7 3 0.5 4 762.62

Rainfall and its erosivity in Ethiopia 74

6. 1. 2 Relationship between long-term annual rainfall erosivity and annual rainfall

Local climatic as well as topographic conditions have great influence on the erosive potential of rainfall. With this in mind, various regression models were fit for long-term annual rainfall erosivity of selected weather stations in the central highlands of Ethiopia. The regression models thus developed and chosen as best are given in Tab. 4. The best regression models of long-term annual rainfall erosivity for different localities in the study area vary considerably. This shows that the erosive potential of rainfall varies from one area to another in the central highlands of Ethiopia. Models with highest coefficients of determination were found for Agarfa (N = 20, R2 = 0.93), Albuco (N = 18, R2 = 0.91), Awassa (N = 16, R2 = 0.99) and Fiche (N = 44, R2 = 0.90) weather stations. It can be inferred from the values of determination coefficients of the respective weather stations that 90% and more of water erosion potential in these localities are explained by long-term mean annual precipitation (mm). Very low coefficients of determination were calculated for Ankober (N = 21, R2 = 0.23), Asasa (N = 17, R2 = 0.28), Bora (N = 13, R2 = 0.34), Debre Markos (N = 38, R2 = 0.33), Gobessa (N = 17, R2 = 0.34), Gohatsion (N = 26, R2 = 0.30) and Munessa (N = 10, R2 = 0.24) weather stations. It can be concluded from the values of determination coefficients that the erosive potential of water in these areas is weakly explained by long-term annual rainfall (mm) recorded for the respective weather stations. The coefficients of determination for the remaining weather stations range from 50-90%, thus indicating relatively good to very good explanation of water erosion potential by long-term annual rainfall (mm) in these areas. As can be implied from the heterogeneity of best models for the various selected weather stations, local-specific soil and water management planning and conservation practices should be based on rainfall characteristics of the respective locality. To achieve efficiency o f the models, especially to the advantage of land resource beneficiaries, conservation technicians and conservation planners, other variables affecting water erosion potential (e.g., land-use, terrain characteristics and micro-climate) should also be taken into consideration when these models are applied to predict the erosive potential of

Rainfall and its erosivity in Ethiopia 75 water. Tab. 4 shows a list of rainfall erosivity regression models for selected weather stations in the central Ethiopian highlands.

Rainfall and its erosivity in Ethiopia 76

Tab. 4: Regression models of long-term annual rainfall erosivity for selected weather stations in the central highlands of Ethiopia

Weatherstations ,v . _ „ R^ssion models Detonation (Y = rainfall erosivity, X = mean annual rainfall) coefficient (R ) A daba Y = 453.1 - 2X + 0.0032X2 - 1.6 * 10'GXJ 0.43 Addis Ababa Y = 33.4 + 0.14X 0.51 A g arfa Y = 0.06X11 0.93 A lbuco Y = -137.63 + 0.7X - 0.0004X2 + 9.07 * 10‘8X3 0.91 Alem Ketema Y = 2 .6 X 07 0 .84 A n k o b er Y = 156.1 + 0.06X 0.23 A rtu m a Y = 66.11 + 0.32X - 0.0002X2 + 3.73 * 10'8X3 0.84 A sasa Y = 222.7 - 1.08X + 0.002X2 - 1.3 * 10'6X3 0.28 A sela Y = 6 .7 X CS 0.50 A w assa Y = 243.7 - 0.4X + 0.0003X2 0 .99 B ora Y = 66 .0 4 + 0. IX 0.34 B oru M eda Y = 404.76 - 0.61X + 0.0004X2 0.63 Debre Markos Y = 64.13 + 0.12X 0.33 D ebre W ork Y = - 176.35 + 0.4X 0.89 D ebre Z eit Y = 0 .4 X 09 0.80 D essie Y = -180 + 0.6X - 0.0002X2 0.46 D hera . Y = 5X0'5 0 .44 E jaji Y = 10X ° 42 0.71 F iche Y = 0 .2 6 X 099 0 .90 G essera Y = - 943.76+159.86 In (X) 0.75 G oba Y = 41.8 + 0.08X 0 .64 G o b essa Y = 3 .2 5 X 054 0.34 G o h atsio n Y = 72 + 0.13X 0.30 H osan a Y = 52.4 + 0.9X 0.63 H u n ita Y = 631.2 - 2.8X + 0.005X2 - 2.3 * 10‘6X3 0 .44 Idoto Y = 50 + 0.07X 0.60 K ach ise Y = 175.8e00003X 0.43 K eleta Y = 120.14 - 0.3 IX + 0.001X2 - 3.6*10’07X3 0.50 K o fo le Y = 140- 0.08X + 6.94*10‘5X3 0.51 K o m b o lch a Y = 192.7 - 0.0003X2 + 2.4 * 10'7X3 0.54 K ore Y = 1.8X 06 0.88 M ajete Y = -1018.51 + 174 in (X) 0.50 M araro Y = 0.38X0'86 0.70 M u n essa Y = 224.2 - 0.13X + 6.5 * 10'5X3 0.24 N ek em te Y = 1623.62- 1.3X + 0.0003X2 0.56 R obi Y = 57 4 e0 0008x 0.63 S agure Y = 3 6 .9 3 e0002x 0 .89 S edika Y = 44.8+ 0.1 IX 0 .50 Seru Y = 2.53 + 0.14X 0 .87 S heno Y = 6 7 .5 1 e0 00IX 0.50 Shola Gebeya Y = 191.15 - 0.14X + 0.0002X3 0.82 S ire Y = 307.71 - 0.5X + 0.0004X2 - 6.4 * 10'8X3 0.89 T u lu B olo Y = 179.98 + 0.0002X2- 0.14 0.71 W onji Y = 59.3e°'oolx 0.71

Rainfall and its erosivity in Ethiopia 77

6. 2 Long-term summer rainfall erosivity

6. 2. 1 Statistical parameters o f summer rainfall erosivity

Long-term mean summer (in amharic Kiremt) rainfall erosivity in the central Ethiopian highlands lies within the range of 102.37 mm (determined for Goba, N = 23) to 400.47 mm (determined for Nekemte, N = 11) (Tab. 5). The least long-term mean summer rainfall erosivity in the region was determined for Goba weather station, with a mean value of 102.37 mm, standard deviation of 27.86 mm and variability coefficient of 27%. The highest long-term mean summer rainfall erosivity was determined for Nekemte weather station, with a mean figure of 400.47 mm, standard deviation of 41.51 mm and variability coefficient of 10.20%. Even though the multi-year summer minimum and maximum rainfall erosivity averages in the central highlands of Ethiopia were observed for Goba and Nekemte weather stations, respectively, the minimum inter summer value for the area was observed for Bora (2.00 mm) and the maximum for Awassa weather station (1282.00 mm) (see annex, Tab. 21). The long-term summer mean rainfall erosivity data of Awassa weather station are characterised by the highest standard deviation (291.20 mm) and the highest variability coefficient (151%) in the whole central Ethiopian highlands, followed by Adaba (standard deviation 201.77 mm, variability coefficient 86%). The skewness coefficient of long-term summer rainfall erosivity in the central highlands of Ethiopia varies from -1.37 (Bora) to 3.96 (Awassa). The kurtosis coefficient of long-term summer rainfall erosivity in the region varies from -1.42 (Kore) to 15.75 (Awassa). The frequency distribution of long-term summer rainfall erosivity at Ankober, Artuma, Bora, Dessie, Ejaji, Fiche, Huruta, Idoto, Kachise, Kore, Robi, Sagure and Sire weather stations is negatively skewed. The frequency distribution of long-term summer rainfall erosivity in all the remaining weather stations is positively skewed. The frequency distribution of long-term summer rainfall erosivity for ca. 80% of the investigated weather stations are platykurtic and the rest leptokurtic (see H a w k in s and W e b e r 1980). Tab. 5 illustrates details of descriptive statistics on summer rainfall erosivity for selected weather stations in the central highlands of Ethiopia.

Rainfall and its erosivity in Ethiopia 78

Tab. 5: Long-term summer (kiremt) rainfall erosivity for selected weather stations in the central highlands of Ethiopia

W eath er N M ean Coefficients Mean S tandard sta tio n s (m m ) Variability Kurtosis Skewness rain fall deviation (m m ) A d ab a 20 234.05 2 0 1.77 86.00 3.29 1.85 378.46 Addis Ababa 99 2 5 7.90 49.55 19.00 1.41 0.76 688.44 A g arfa 20 131.01 70.17 54.00 2.42 1.54 333.49 A lbuco 18 287.43 186.11 65.00 10.32 2.88 563.65 Alem Ketema 23 325.82 82.54 25.00 -0.60 0.18 674.41 A n k o b er 20 316.22 76.59 24.00 -0.87 -0.21 716.83 A rtum a 9 319.37 54.11 17.00 0.18 -0 .5 9 660.22 A sasa 17 123.34 40.64 33.00 2.43 0 .24 309.00 A sela 23 206.20 47.51 23.00 -0.73 0.12 499.19 A w assa 16 193.48 291.20 151.00 15.75 3.96 436.64 B ora 13 173.84 76.08 44.00 1.17 -1.37 364.52 B orn M eda 18 263.12 92.78 35.00 3.17 0.66 522.36 Debre Markos 38 282.83 41.10 15.00 0.60 0.50 769.48 D ebre W ork 13 232.70 66.54 29.00 1.00 1.11 584.13 Debre Zeit 34 217.20 66.40 31.00 0.38 0.36 530.33 D essie 24 276.07 65.94 24.00 1.16 -0.53 568.31 D hera 19 144.43 39.77 28.00 -0.35 0.40 345.81 E jaji 23 2 2 0 .7 9 42.52 19.00 1.77 - 1.01 583.37 F iche 36 292.83 128.68 44.00 -0.11 -0.56 609.94 G essera 19 3 7 2 .6 6 141.21 38.00 0.15 0.78 372.66 G oba 23 102.37 27.86 27.00 1.84 1.11 269.35 G o b essa 17 151.48 36.75 24.00 2.37 1.48 382.49 G o h tsio n 26 2 7 9 .6 7 53.06 19.00 -0.20 0.42 711.89 H o sana 15 173.40 27.37 16.00 -0.25 0.58 450.27 H uruta 11 138.46 23.50 17.00 0.78 -0 .2 4 2 3 7.36 Idoto 17 120.46 32.18 27.00 0 .06 - 1.10 3 0 0.60 K achise 11 347.72 70.64 20.00 0 .70 -0.13 826.71 K eleta 21 139.43 35.76 26.00 0.27 0.24 338.00 K ofole 21 148.43 36.48 25.00 -0.24 0.82 3 7 6 .0 0 K o m b o lch a 40 251.17 69.57 28.00 0.31 0.15 536.56 K ore 16 171.78 27.03 16.00 -1.42 -0.09 460.00 M ajete 32 2 6 4.27 85.43 32.00 -0.43 0.28 538.57 M araro 21 2 0 1 .0 6 71.82 36.00 -1.24 0.44 528.00 M u n essa 10 177.81 41.15 23.00 -0.95 0.12 478.00 N ek em te 11 4 0 0 .4 7 41.51 10.20 2 .36 1.24 1182.00 R obi 9 123.25 28.54 23.00 0.89 -0.87 324.00 S agure 10 157.39 36.63 23.00 -1.34 -0.17 416.00 S ed ik a 17 175.73 50.48 29.00 -0.36 0.68 4 2 2 .0 0 S eru 16 211.33 108.17 51.00 7.09 2.28 520.00 S heno 34 264.66 77.29 29.00 2.18 0.38 6 1 4.37 Shola Gebeya 35 268.65 87.23 32.00 14.65 3.25 595.10 S ire 17 143.84 46.71 32.00 -0.63 -0 .4 7 319.00 T ulu B olo 35 258.45 89.73 35.00 1.50 1.08 7 1 8.60 Wonji 33 184.93 63.06 34.00 -0.97 0.46 4 3 3 .3 7

Rainfall and its erosivity in Ethiopia 79

6. 2. 2 Relationship between long-term summer rainfall erosivity and summer rainfall

The summer period is the major rainy season in the whole central highlands of Ethiopia and, hence, it is the main period when erosive rainfall occurs in the region. Therefore, regression models were fit for summer rainfall erosivity for selected weather stations in the region. The choice of the best models was based on statistical, theoretical and practical criteria. The best models thus fit are presented in Tab. 6, along with the respective weather station.

Rainfall and its erosivity in Ethiopia 80

Tab. 6: Regression models for long-term summer ([kiremt) rainfall erosivity

W eather stations Regression models Determination (Y = rainfall erosivity, X = mean summer rainfall) coefficient (R2)

A d ab a Y = 952.82 - 3.1X + 0.003X2 0.80 Addis Ababa Y = 5.8 + 0.4X 0.86 A garfa Y = 0 .3 X 1'06 0.97 A lbuco Y= -89.46 +0.96X - 0.0006X2 + 2.25 * 10‘7X3 0.98 Alem Ketema Y = 125.1 e° 0,2X 0.62 A n k o b er Y = 131 e000,2X 0.70 A rtum a Y = 395.2 - 0.6X + 0.0007X2 0.37 Asasa Y = 30.5 + 0.3X 0.80 A sela Y = 156.1 - 0.08X + 0.0003X2 0.70 Awassa Y = 70 + 0.0005X2 - 7 * 10'8X3 0.99 B ora Y = 1.08X 087 0.97 B oru M eda Y = - 122.73 + 1.7X - 0.003X2 + 1.8 * 10'6X3 0.84 Debre Markos Y = - 1533 + 273.62 In (X) 0.84 Debre Work Y = 80e°'002X 0.81 D ebre Z eit Y = 75.3e° 02X 0.84 D essie Y = 1.4X 083 0.74 Dhera Y = 40.34 + 0.3X 0.77 Ejaji Y = 8.33 + 1.2X - 0.003X2 + 2 * 10'6X 3 0.87 F iche Y = 102.30 + 0.31X 0.68 G essera Y = - 74.33 + 1.3X - 0.003X2 + 2.56 * 10‘6X3 0 .96 G oba Y = 3 6 .9 2 e0 004X 0.91 G o b essa Y = 0.4X101 0.98 Gohatsion Y = -35.3 + 0.44X 0.82 H o sana Y = 103.2 + 0.2X 0.20 H uruta Y = 8 + 1.4X - 0.0 IX2 + 6.6 * 10'6X3 0.93 Idoto Y = 40.5eOOO4X 0.78 Kachise Y = 226.28e00005X 0.54 K eleta Y = 31.3 + 0.31X 0.85 K ofole Y = 113 - 0.13X + 0.001X2 0.72 K o m b o lch a Y = 1.2X084 0.75 K ore Y = 6 9 .1 e0002x 0.80 Majete Y = 0.80X092 0.80 M araro Y = 0.56X0 94 0 .96 M u n essa Y = - 935.5 + 179.3 In (X) 0.98 N ek em te Y = - 23.44 + 0.36X 0.98 R obi Y = 41.5e°'003X 0.88 S agure Y = 375.02 - 2.4X +0.007X2 - 5.2 * 10'6X3 0.91 S edia Y = 47.2e0003X 0.83 Seru Y = - 202.95 + 1,6X - 0.002X2 + 1.2 * 10'6X3 0.96 S heno Y = 3X 0'7 0 .84 Shola Gebeya Y = 188.2 - 0.1X + 0.0003X2 0.90 S ire Y = 7.73X0'51 0.83 T ulu B olo Y = 0.94X0 95 0.94 W onii Y = 141.05 - 0.58X + 0.002X2 - 1.5 * 10'6X3 0.89

Rainfall and its erosivity in Ethiopia

L 81

As can be revealed from the rainfall erosivity regression models (Tab. 6), the coefficients of determination range from 0.20 (determined for Hosana, N = 15) to 0.99 (calculated for Awassa, N = 16). The best coefficients were obtained for the weather stations with R2 greater than or equal to 0.90. These are Agarfa, Albuco, Awassa, Bora, Gessera, Goba, Gobessa, Huruta, Mararo, Munessa, Nekemte, Sagure, Seru, Shola Gebeya and Tulu Bolo. More than or equal to 90% of rainfall erosivity at these weather stations are explained by the long-term summer precipitation (mm) of the respective weather station. Areas with R2 in a range of 0.70 to 0.89 are considered to have good to very good regression models. This category of weather stations comprises more than half of the selected areas. Long-term summer precipitation (mm) in these areas explains 70%-89% of rainfall erosivity in the corresponding locality. Fair coefficients of determination were found for Alem Ketema (N = 23, R2 = 0.62), Fiche (N = 36, R2 = 0.68) and Kachise (N = 11, R2 = 0.54). Very low coefficients of determination were calculated for Artuma (N = 9, R2 = 0.37) and Hosana (N = 15, R2 = 0.20). The coefficients for the remaining weather stations lie within a range of 70%-90%.

Compared with the long-term annual rainfall erosivity models, the prediction capacity of summer models is higher. This is substantiated by the comparatively better coefficients of determination of summer rainfall erosivity obtained for most of the selected weather stations and the statistical significance of the regression coefficients. The erosive potential of precipitation in Alem Ketema, Artuma, Debre Work, Fiche, Hosana, Kore and Sire areas can be better predicted by the long-term annual rainfall (mm) of the corresponding weather station. In the Awassa area, the erosivity of rainfall can be equally well predicted using both annual and summer long-term precipitation (mm). Therefore, either of the available data can be used for soil and water management planning as well as prediction. Water erosion in all the remaining areas of the central highlands of Ethiopia can be best predicted using long-term summer rainfall (mm) recorded for the respective weather station. It can be noted that, for most of the weather stations, better coefficient of determination was obtained by using local specific data than aggregate time series data of the

Rainfall and its erosivity in Ethiopia 82 whole central highlands of Ethiopia (Figs. 30 and 31). However, it can be recommended that the aggregate rainfall erosivity time series data and the corresponding prediction model be used for local land and water management practice when data are not available for that specific locality.

6. 3 Long-term rainfall erosivity during short rainy seasons

Autumn (in Amharic Tseday) and spring (in Amaharic Belg) periods are short rainy seasons in most parts of the central Ethiopian highlands. Precipitation during this time, though agriculturally not as important as summer rains, contributes to the causes of soil erosion in the central highlands of Ethiopia. Especially since land cover during this period is sparse and the soil is relatively dry, short rains cause significant soil degradation in the area ( W r ig h t and A d a m se g e d 1984). Due to this fact, rainfall erosivity during short rains in the central highlands of Ethiopia was investigated and the descriptive statistical results presented in Tab. 7.

6. 3. 1 Autumn

Long-term mean autumn rainfall erosivity (mm) for selected weather stations in the central highlands of Ethiopia ranges from 51.64 mm, determined for Asasa (N = 16), to 257.16 mm, estimated for Nekemte (N =11) (Tab. 7). The standard deviations of rainfall erosivity (mm) for the Asasa and Nekemte weather stations are 39.42 and 106.74 mm, respectively. As presented in detail (see annex, Tab. 22), the least inter autumn rainfall erosivity in the region was observed for Sire weather station, with erosivity value of 0.00 mm. A long-term average autumn rainfall erosivity of 62.21 mm, with standard deviation of 61.37 mm (variability coefficient = 99%) was determined for Sire. This implies that the lower the rainfall erosivity is, the higher the relative dispersion of long-term autumn rainfall erosivity in the central highlands of Ethiopia. From this tendency of variability in rainfall erosivity, it can be concluded that extremely erosive rainfalls concentrated within short duration and inflicted intolerable damage to the soil. The highest inter autumn rainfall erosivity was noted for Albuco

Rainfall and its erosivity in Ethiopia 83 weather station, with a value of 697.24 mm. A long-term mean rainfall erosivity value of 128.34 mm, with standard deviation of 150.59 mm and a variability coefficient of 117%, was noted for Albuco (see annex, Tab. 22).

6. 3. 2 Spring

The highest long-term mean spring rainfall erosivity in the central highlands of Ethiopia was noted at Artuma weather station (N = 10), with a multi-year mean spring rainfall erosivity value of 231.64 mm, standard deviation of 163.84 mm and a variability coefficient of 70.73% (Tab. 7; annex, Tab. 23). The least long­ term mean spring rainfall erosivity in the region was obtained for Wonji weather station (N = 32), with a value of 64.09 mm, standard deviation of 28.16 mm and a variability coefficient of 43.93% (Tab. 7; annex, Tab. 23). The highest inter spring rainfall erosivity in the region was observed at the Artuma, with a figure of 660.15 mm, and the lowest the Fiche weather station, with a value of 2.58 mm for that specific spring. A multi-year mean spring rainfall erosivity of 69.15 mm, with standard deviation of 42.92 mm and a variability coefficient of 62.07%, was determined for the Fiche weather station.

Autumn rainfall is comparatively more erosive than spring rainfall in some of the selected weather stations (Tab. 7). Contrary to this generalisation, long-term autumn and spring rainfall in some of the areas investigated are approximately equally erosive. Examples are Awassa, Dhera, Goba, Hosana, Keleta, Kore and Tulu Bolo weather stations. The highest seasonal long-term rainfall erosivity for the weather stations Agarfa, Goba, Gobessa, Robi and Seru was noted for the spring period. However, it can be concluded that summer rainfall in the central highlands of Ethiopia causes most of the damage to soil in the area. The summer precipitation at the majority of the investigated weather stations (88.64%) was found to be more erosive than rainfall during the other seasons.

Rainfall and its erosivity in Ethiopia 85

As shown in the annex (Tab. 25 and 26), a good relationship between long-term autumn and spring rainfall erosivity with the corresponding long-term rainfall was noted for the weather stations studied. The models developed here are applicable for long-term forecasting of rainfall erosivity during the short rainy seasons in the respective areas.

6. 4 Temporal and spatial characteristics of long-term rainfall erosivity

In spite of their close location to each other in the central Ethiopian highlands, the weather stations selected for this study experienced quite different spatial and temporal characteristics of rainfall erosivity. In order to better understand this, a detailed time series analysis was conducted on rainfall erosivity using various methods.

6. 4. 1 Temporal and seasonal variability o f rainfall erosivity

6. 4. 1. 1 Graphical trend analysis

Long-term annual and summer rainfall erosivity at Ejaji, Shola Gebeya and Tulu Bolo weather stations followed an increasing trend (Figs. 34a and b). The highest figures during both periods at Ejaji were recorded in 1994, with annual and summer values of 236.48 mm and 289.52 mm, respectively. The lowest annual and summer rainfall erosivity at Ejaji was noted in 1976, with annual and summer values of 77.77 mm and 100.20 mm, respectively (Figs. 34 a [I] and b [I]). The Shola Gebeya weather station has seen both the highest annual and summer rainfall erosivity in 1985, with annual and summer values of 536.11 mm and 681.13 mm, respectively (Figs. 34 a [II] and b [II]). The Tulu Bolo weather station experienced its highest annual (451.90 mm) and summer (523.50 mm) rainfall erosivity in the years 1992 and 1993, respectively. The lowest annual (90.71 mm) and summer (106.60 mm) rainfall erosivity at the Tulu Bolo weather station were noted in 1983 (Figs. 34 a [III] and b [III]). As illustrated by the trend lines and their corresponding equations (Fig. 34), rainfall erosivity at Tulu Bolo has followed the steepest trend, as compared with Ejaji and Shola Gebeya weather stations. The weakest increasing trend in annual rainfall

Rainfall and its erosivity in Ethiopia 86 erosivity was observed at the Ejaji weather station. For the summer period, the weakest increasing trend was noted for the Shola Gebeya (5-years running mean) and Ejaji (11-years moving average) weather stations. Figs. 34 a and b illustrate trends in rainfall erosivity at Ejaji, Shola Gebeya and Tulu Bolo.

250 230 Y5 = 0.50t- 823.16 ...... c ...... 1...... A ~ ? R2 = 0.06 N = 32 210 6 ° ° : O y .iQo..... 190 .....o.„ ...... ' :...... A 0 / 0 ...... / \ O 170 : _ a - H r i Y,, =0.531 - 889.38 150 R2 = 0.13 N = 32 130 ...... rv...... °\ 110 O o 90 M aximum 70 l “ ■ 1 I^ 7,741, j, ■, ■ 1,,,, ■ M inim um 1965 1970 1975 1980 1985 1990 1995 2000

Years

o AFm data point — 5-years running mean ------11-years running mean - Linear (11 -years running mean) -----Linear (5-years running mean)

Fig. 34 a (I): Time series of annual rainfall erosivity at the Ejaji weather station

Rainfall and its erosivity in Ethiopia 87

1000 Y 5 = 0.7612t - 1299 R2= 0.0385 N = 35

Y„ =2.01091-3776.4

R2 = 0.4821 N = 35

M aximum ■ M inim um

1:17.28 100 4- 1960 1965 1970 1975 1980 1985 1990 1995

Years o A Fm data point ------5-years running rrean

1 11 -years running mean ------Linear (5-years running mean)

------Linear (11-years running mean)

Fig. 34 a (II): Time series of annual rainfall erosivity at the Shola Gebeya weather station

Y5 = 6.935t - 13441.00 R2 = 0.13 N = 35

Y ,, = 11.3 5 4 t- 22217.00 R2 = 0.56 N = 35

M axim um ■ M inimum

1962 1967 1972 1977 1982 1987 1992

Years

o AFm data point ------5-years running mean

------11-years running mean ------Linear (5-years running mean)

------Linear (11 -years running mean)

Fig. 34 a (III): Time series of annual rainfall erosivity at the Tulu Bolo weather station

Rainfall and its erosivity in Ethiopia Ys = 1 -85t - 3455.90 R2 = 0.20 N = 23

Y,, = 2.47131 -4698.90

R2 = 0.67 N = 23

M axim um CtZ ■ M inim um

1966 1976 1986 1996

Y ears

o SUFm data point ------5-years running mean ------11 -years running mean ------Linear (5-years running mean) — ■ Linear ( 11 -years running mean)

Fig. 34 b (I): Time series of summer rainfall erosivity at the Ejaji weather station

Ys = 1.03t - 1765.60 R2 = 0.045 N = 35

Y„ = 2.67t- 5012.00 R2 = 0.50 N = 35

M axim um ■ M inim um

Y ears

5 SUFm data point — 5-years running mean

— 11 -years running mean - Linear (5-years running mean)

" • Linear (11 -years running mean)

Fig. 34 b (II): Time series of summer rainfall erosivity at the Shola Gebeya weather station

Rainfall and its erosivity in Ethiopia 84

Tab. 7: Statistical parameters of autumn (Tseday \ and spring (Belg) rainfall erosivity (mm) for selected weather stations in the central highlands of Ethiopia

A utum n Spring Weather stations N Mean Standard Mean rainfall N Mean Standard Mean rainfall (m m ) deviation (m m ) (m m ) deviation (m m )

A daba 18 72.06 35.58 106.02 19 91.03 34.01 190.11 Addis Ababa 99 172.54 65.78 209.22 99 103.95 46.16 236.74 A garfa 20 123.05 80.67 217.78 19 182.80 145.48 420.66 Al buco 18 128.34 150.59 198.35 16 98.75 47.78 182.04 Alem Ketema 22 129.20 58.28 142.40 22 77.08 35.88 149.47 A nkober 20 177.44 95.78 272.98 20 204.37 95.71 421.88 A rtum a 9 166.96 68.14 248.57 10 231.64 163.84 488.05 A sasa 16 51.64 39.42 78.00 16 68.23 33.67 139.00 A sela 21 163.13 56.68 235.20 23 139.50 56.67 298.21 A w assa 16 120.44 34.49 198.35 15 125.25 33.91 303.86 Bora 12 87.63 46.64 129.96 12 114.94 66.35 244.98 B oru M eda 19 102.74 59.77 136.46 19 147.54 63.34 291.31 Debre Markos 38 186.60 52.00 285.21 38 89.00 36.00 202.87 D ebre W ork 13 116.07 55.28 182.42 13 70.92 26.80 160.82 D ebre Z eit 34 97.86 48.14 121.99 34 73.87 35.26 152.63 D essie 25 135.16 50.76 196.19 24 108.58 47.53 259.51 D hera 18 80.04 36.50 99.70 18 76.01 37.04 142.67 Ejaji 20 140.84 39.66 202.94 24 110.40 48.20 236.67 Fiche 33 110.51 61.86 124.01 35 69.15 42.92 146.97 G essera 19 179.43 101.75 302.69 18 206.52 108.36 477.71 G oba 23 125.23 29.56 226.24 22 126.63 33.63 300.67 G obessa 17 118.19 33.26 202.08 17 162.42 48.83 365.62 Gohatsion 25 113.39 53.72 157.69 26 94.14 43.71 204.90 H osana 16 135.91 67.93 204.06 16 139.52 64.05 333.63 H uruta 11 120.51 45.40 97.01 13 94.24 45.92 154.67 Idoto 17 88.43 44.82 156.20 17 115.31 24.77 272.08 K achise 10 237.79 79.25 315.28 11 122.19 53.05 299.57 K eleta 21 91.36 37.49 124.00 20 89.36 39.93 178.00 K ofole 19 128.03 34.33 224.00 20 140.01 32.65 362.00 Kombolcha 40 161.77 53.83 171.77 40 97.86 39.67 219.46 K ore 16 125.71 38.76 219.00 17 127.21 34.09 317.00 M ajete 33 104.41 42.21 150.86 33 122.30 48.35 256.30 M a raro 20 95.59 49.27 151.43 21 131.84 68.62 305.91 M unessa 10 161.14 30.76 266.33 10 170.45 53.14 379.45 N ekem te 11 257.16 106.74 440.27 11 207.39 76.59 386.91 Robi 9 85.29 27.41 137.78 9 139.01 47.08 315.71 Sagure 10 157.39 36.63 108.70 11 84.13 32.76 191.05 Sedia 17 105.93 49.11 177.71 17 131.48 58.83 271.86 Seru 16 176.42 75.49 295.34 15 226.69 129.08 488.85 Sheno 34 94.87 65.02 117.56 33 70.18 26.04 159.89 Shola Gebeya 33 98.79 45.13 123.86 35 80.07 35.38 50.91 Sire 16 62.21 61.37 251.78 15 87.55 36.46 180.23 T ulu Bolo 30 87.32 43.84 113.52 35 88.16 34.41 192.24 W onji 33 96.10 58.35 127.00 32 64.09 28.16 144.74

Rainfall and its erosivity in Ethiopia 89

550 '■ 5 2 3 .5 0 a 500 Y5 = 7.04t- 13589.00 ! ? 4 5 0 R2 = 0.13 N = 35 Q § 40 0 Y„ = 11.57t- 22593 3 ' . i 350 o C/5 o o o o R2 = 0.54 N = 35 J 300 : / / 0 | 25 0 - ° \ M axim um I 200 ■ M inim um (S 0 ° ° ? 0 ° 0 0 0 150 106.60 100 ■ , . .

1962 1972 1982 1992

Y ears

o SUFm data point ------5-years running mean

------11-years running mean ------Linear (5-years running mean)

“ * Linear (11-years running mean)

Fig. 34 b (III): Time series of summer rainfall erosivity at the Tulu Bolo weather station

Annual and summer erosivity of rainfall at Addis Ababa, Debre Markos, Debre Zeit, Fiche, Kombolcha, Majete, Sheno and Wonji weather stations followed decreasing course during the period of investigation. The lowest annual rainfall erosivity in Addis Ababa (132.74 mm) was observed in 1902, Debre Markos (159.72 mm) in 1987, Debre Zeit (69.38 mm) in 1951, Fiche (5.60 mm) in 1984, Kombolcha (99.93 mm) in 1984, Majete (103.74 mm) in 1984, Sheno (69.08 mm) in 1967 and Wonji (67.97 mm) in 1978. The highest annual rainfall erosivity in Addis Ababa (370.56 mm) was observed in 1922, Debre Markos (292.72 mm) in 1970, Debre Zeit (304.46 mm) in 1964, Fiche (453.92 mm) in 1969, Kombolcha (321.16 mm) in 1964, Majete (300.90 mm) in 1963, Sheno (392.64 mm) in 1966 and Wonji (256.40 mm) in 1975. The 5-years and 11-years running mean curves fitted to the annual and summer rainfall erosivity data points of these weather stations confirm that rainfall erosivity has been declining

Rainfall and its erosivity in Ethiopia 92

330 Y, = -0.14t + 444.26 A 304.46 290 O R2 = 0.0014 N = 34 250 ^ \ i ^ 210 0 ;------Y „ = -0.6731+ 1478.50 z J J " 0 170 R‘ = 0.05 N = 34 / o O: 0

130 0 0 • o 0 • M aximum 90 ■ M inim um « 69.) 8 50 1950 1955 I960 1965 1970 1975 1980 1985 1990 Years

o AFm data point ------5-years running mean

------11 -years running mean ------Linear (11 -years running mean)

------Linear (5-years running mean)

Fig. 35 a (III): Time series of annual rainfall erosivity at the Debre Zeit weather station

Y5 =-3.441 + 7050.10

R2 =0.Z7N=44

Yn =-5.20t + 10557.00

R2 = 0.60N =44

-*• Maximum ■ M inimum

1954 1964 1974 1984 1994

Years

o AFm data point ------5 -y a rs running m a n

------11-yars running rrcan ------Linear (5-yeais running nran)

------Linear (11-years running rrean)

Fig. 35 a (IV): Time series of annual rainfall erosivity at the Fiche weather station

Rainfall and its erosivity in Ethiopia 93

: i 321.16: Ys - -0.76t +1673.40 R2 = 0.14 N = 40 280 _-.„o...... i 255 . : O .L.O.A...... -....IQ... o 230 o : o Y ,,=-0.64t + 1440.20 •? O ° j ....o...... TB T-S ^ R2 = 0.31 N = 40

-*■ Maximum ■ Minimum 80 : ------] ------| ------[ ? ?9;93 1952 1962 1972 1982 1992 Years

o AFm data point ■ 5-years running mean ------11 -years running mean ------Linear (5-years running mean) ------Linear (11-years running mean)

Fig. 35 a (V): Time series of annual rainfall erosivity at the Kombolcha weather station

300.90

Ys = -5.5t + 110910 R2 = 0.32 N = 35

Ylt =-3,69331 + 7530.5 R2 = 0.41 N = 35

^ Maximum ■ Minimum

1960 1965 1970 1975 1980 1985 1990 1995 2000 Years o AFm data point ------5-years running mean ------11-years running mean ------Linear (5-years running mean)

L— - Linear (11-years running mean)

Fig. 35 a (VI): Time series o f annual rainfall erosivity at the Majete weather station

Rainfall and its erosivity in Ethiopia 96

430 Y5= -0.401+ 1041.10

390 R2 = 0.18 N 98

350 Y„ =-0.52t+ 1258.60 310 R: = 0.52 N = 98

270 M a x im u m 230 ■ M in im u m 190

150 1898 1913 1928 1943 1958 1973 1988

Y ears o SUFm data point ------5-years running mean ------11 -years running mean ------Linear (5-years running mean) — ■ Linear (11 -years running mean)

Fig. 35 b (I): Time series of summer rainfall erosivity at the Addis Ababa weather station

4 00 Ys = -1.23t + 2700.80 375 R2 = 0.30 N = 38 3 50

325 Y„ = -l.lt + 2397.30

300 R2 = 0.34 N = 38

275 M a x im u m 250 ■ M in im u m 225

200 1954 1964 1974

Y ears

o SUFm data point ------5-years running mean

------11-years running mean ------Linear (11 -years running mean)

------Linear (5-years running mean)

Fig. 35 b (II): Time series of summer rainfall erosivity at the Debre Markos weather station

Rainfall and its erosivity in Ethiopia 97

Y ,= 0 .1 3 t-18.23 R2 = 0.0007 N = 34

Yn = -0.68t + 1577.80 'I* Rz = 0.031 N = 34

-*• Maximum ■ Minimum

1951 1956 1961 1966 1971 1976 1981 Years

o SUFm data poini ------5-years running mean “ 11-years running mean ------Linear (5-years running mean) ~ » Linear (11-years running mean)

Fig. 35 b (III): Time series of summer rainfall erosivity at the Debre Zeit weather station

500 ■ * 516.26 Ys = -7.20t + 14556.00 0 ° 0 0 R2 = 0.10 N = 36 1 400 ■ / \ o A. ° 0 S’ — Y ,,= -9. lOt+ 18355.00 s 3oe- ° u 0? ° .r * 3 R2 = 0.23N = 36 6 a 200- = °\1 o -*• Maximum 1 ™- o o ° ° o ■ Minimum

• , . .,0 ,. 1954 1964 1974 1984 1994 Years

o SUFm data point ------5-years running mean ------11-years tunning mean ------Linear (5-years running mean) — - 'Linear (11-years running mean)

Fig. 35 b (IV): Time series of summer rainfall erosivity at the Fiche weather station

Rainfall and its erosivity in Ethiopia 100

6. 4. 1. 2 Departures o f annual and summer rainfall erosivity from long-term mean

Addis Ababa

The largest departure of annual rainfall erosivity from the multi-year mean at the Addis Ababa weather station was noted in 1922, when the annual value was 78% greater than the long-term mean (Fig. 36). In the subsequent years, the departures were much lower and showed considerable fluctuations in magnitude. The least erosive annual rainfall at the Addis Ababa weather station was observed in 1902, where the annual rainfall erosivity was lower than the multi­ year mean by 36%. Highly erosive annual rainfall occurred in Addis Ababa in the years 1915 and 1947, where the annual values exceeded the long-term mean by 65% and 58%, respectively.

As can be observed from Fig. 36, the departure of summer rainfall erosivity from the multi-year mean at the Addis Ababa weather station attained its highest in 1922, where it exceeded the multi-year average by 111%. Relatively highly erosive summer rainfall also occurred in the years 1947 and 1970, where the summer values were higher than the multi-year mean by 97% and 66%, respectively. Compared with multi-year average, the least erosive summer rainfall at the Addis Ababa weather station was noted in the year 1987, where summer rainfall erosivity value was 21% less than the multi-year average. The Addis Ababa weather station has experienced summer rainfall with nearly equally erosive potential in the years 1906, 1916, 1936, 1958 and 1993, as compared with multi-year average.

The highest deviation of summer rainfall erosivity from the corresponding long­ term average was noted in 1922, with the value exceeding the long-term mean by 70%. The minimum deviation was noted in the year 1987, where the summer value was 38% lower than the respective long-term average (Fig. 36). These results agree with those of the comparison between summer rainfall erosivity and multi-year mean, which signifies the occurrence of highly erosive summer

Rainfall and its erosivity in Ethiopia 101 rainfall at the Addis Ababa weather station during 1922. Generally, as can be confirmed by the high positive deviation, summer rainfall erosivity at the Addis Ababa weather station is far higher than the multi-year average for most of the investigation years. It is obvious from these results that summer rainfall in Addis Ababa is the major causative agent of soil erosion and, hence, leads to severe land degradation. The summer rainfall erosivity departures from the respective long-term average were found to be generally lower than the summer departures from multi-year value. Therefore, intra seasonal variability is apparently lower than seasonal variability in rainfall erosivity at Addis Ababa weather station. Fig. 36 shows departures of rainfall erosivity from long-term mean in Addis Ababa.

Fig. 36: Departures of annual and summer rainfall erosivity from long-term average at Addis Ababa weather station

Rainfall and its erosivity in Ethiopia 102

Debre Markos

The highest positive annual rainfall erosivity departure from multi-year mean at the Debre Markos weather station was identified in 1970, where the annual rainfall erosivity was 30% higher than the corresponding multi-year mean (Fig. 37). Highly erosive annual rainfall at the Debre Markos weather station was observed in 1958, 1960 and 1985, where the rainfall erosivity values were 29% higher than the multi-year mean in each of these years. The least erosive annual rainfall at the Debre Markos weather station was noted in the year 1988, where the rainfall erosivity value was lower than the multi-year mean by 29%.

As illustrated by Fig. 37, compared with the multi-year mean, the highest summer rainfall erosivity at the Debre Markos weather station was observed in the year 1958, where summer rainfall erosivity was higher than the multi-year mean by 70%. Pronounced positive departures of summer rainfall erosivity from the multi-year average were also noted in 1960 and 1985, with summer rainfall erosivity values exceeding the multi-year average by 68% and 63%, respectively. The deviation of summer rainfall erosivity from the multi-year mean was at its lowest point in 1987, when it was 12% lower than the multi-year mean. It can be inferred from this analysis that summer rainfall dominantly causes water erosion, as compared with annual rainfall, in Debre Markos area.

As compared with the long-term summer average rainfall erosivity (Fig. 37), the highest departure of summer rainfall erosivity from the corresponding long-term average was found in 1958, where the value was higher than the long-term average by 35%. Highly erosive summer rainfall at Debre Markos weather station was also noted in the years 1960 and 1985, where the summer values were higher than the long-term average by 34% and 30%, respectively. The departure of summer rainfall erosivity from its long-term mean was at its lowest point in 1987, where summer value was 30% less than the long-term mean. Negative extreme value of summer rainfall erosivity was also detected in 1978, with summer value being 25% less than the long-term summer average. As can be confirmed from the deviation lines in Fig. 37, the departure of summer

Rainfall and its erosivity in Ethiopia 103 rainfall erosivity from the long-term summer average follows a pattern similar to the departure of summer rainfall erosivity from multi-year mean. This suggests that, in Debre Markos area, inter-seasonal and intra-seasonal variability show nearly similar characteristics. Fig. 37 depicts departures of rainfall erosivity from long-term averages at Debre Markos weather station.

Fig. 37: Departures of annual and summer rainfall erosivity from long-term average at the Debre Markos weather station

Debre Zeit

The maximum departure of annual rainfall erosivity from multi-year mean at the Debre Zeit weather station was found in 1964, where the annual value exceeded the multi-year mean by 77% (Fig. 38). The deviations in the subsequent years fell continuously for nearly a decade but rose abruptly afterwards; however, without continuation of dramatic fluctuation. The Debre Zeit weather station has seen consecutive years of highly erosive annual rainfall during 1963-1966,

Rainfall and its erosivity in Ethiopia 104 where the departures were higher than the multi-year mean by 34%-77%. Annual rainfall erosivity at Debre Zeit was at its lowest in the year 1951, where a record 61% lower than the multi-year average rainfall erosivity was noted. Compared with the remaining years of investigation, annual rainfall erosivity at Debre Zeit weather station was relatively very low during 1951-1956, when least erosive rainfall was highly concentrated, with departures lower than the multi-year mean by 3 1%-61 %.

Compared with the multi-year mean, the deviation of summer rainfall erosivity at the Debre Zeit weather station (Fig. 38) reached its highest positive in 1964, where it was 121% greater than the multi-year mean; then continued falling till 1972, but rose again and attained a secondary high in 1975 (+65%) which was followed by a steady drop in summer rainfall erosivity. Summer rainfall erosivity at the Debre Zeit weather station attained other minor highs in the years 1981 and 1983, with the values exceeding the multi-year mean by 34% and 47%, respectively. The least erosive summer rainfall at Debre Zeit was observed in 1951, where it was 47% lower than the multi-year mean. Summer rainfall erosivity, approaching the multi-year mean, continued to increase to the extent that the value reached a pronounced secondary high in 1955 (+58%), but sharply fell to a secondary minimum in 1956 (-39%). Debre Zeit experienced major concentration of low summer rainfall erosivity during the period 1951- 1956. This period coincided with the occurrence of high concentration of less erosive annual rainfall in the area. It can be noted from the results of comparative investigation that the summer rainfall erosivity at the Debre Zeit weather station mostly lies above the multi-year mean. Subsequently, summer rainfall is the main causing agent of water erosion in Debre Zeit.

The departure of summer rainfall erosivity from long-term summer average at the Debre Zeit weather station reached its highest in 1964, with summer value exceeding the long-term summer mean by 75%. Compared with long-term summer rainfall erosivity, the least erosive summer rainfall was recorded in 1956, where the summer value was 52% lower than the long-term summer average. Parallel tendency in the deviations of summer rainfall erosivity from

Rainfall and its erosivity in Ethiopia 105 multi-year mean and long-term summer average was noted (Fig. 38). It can be observed from the results that the most erosive summer and annual rainfall at the Debre Zeit weather station coincided in the year 1964. Fig. 38 shows the departures of rainfall erosivity from long-term averages.

Fig. 38: Departures of annual and summer rainfall erosivity from long-term average at the Debre Zeit weather station

Ejaji

The Ejaji weather station is characterised by high irregularity in rainfall data and, hence, limited analysis was made on seasonal variability of rainfall erosivity in the area. The rainfall erosivity calculated from the available data indicates that maximum departure of annual rainfall erosivity from long-term mean occurred in 1994, with annual value 38% greater than the multi-year mean (Fig. 39). The Ejaji weather station also experienced high positive deviations of annual rainfall erosivity in 1975 and 1997, where the annual values were higher than the multi-year mean by 24% and 36%, respectively. The departure of annual rainfall erosivity reached its extreme low in 1976, where the annual value

Rainfall and its erosivity in Ethiopia 106 was 55% less than the multi-year mean. Secondary lows were noted in 1981 and 1990, where the annual values were 43% and 42% less than the multi-year mean, respectively. The data indicate that rainfall erosivity during 1980-1993 was relatively very low at Ejaji weather station.

The deviation of summer rainfall erosivity from multi-year mean attained its highest in 1994, where the summer rainfall erosivity was 69% higher than the multi-year mean. As can be confirmed from Fig. 39, the deviation curve of summer rainfall erosivity from long-term summer mean runs parallel to the deviation line of summer rainfall erosivity from multi-year mean and both their highest positive departures coincided in 1994. The deviation of summer rainfall erosivity from the corresponding long-term summer mean in 1994 was 31%. Compared with long-term summer average, high summer rainfall erosivity was also noted in 1985, where the summer value was 24% higher than the multi-year mean. Fig. 39 indicates the departures of rainfall erosivity from the long-term average at Ejaji.

Fig. 39: Departures of annual and summer rainfall erosivity from long-term average at Ejaji weather station

Rainfall and its erosivity in Ethiopia 107

Fiche

Long-term annual rainfall erosivity at Fiche is characterised by obvious irregularities because of frequently missing rainfall records. As depicted by Fig. 40, the estimated long-term annual rainfall erosivity data of Fiche weather station show that the highest positive departure from multi-year average occurred in 1969, with annual value exceeding the long-term mean rainfall erosivity by 77%. Major positive deviations were also observed in 1955, 1968 and 1970; where annual values were higher than the multi-year mean by 61%, 59% and 58%, respectively. This signifies the occurrence of highly erosive annual rainfall during these years. Except for the years 1959 and 1961, where annual rainfall was relatively far less erosive, the period from 1954-1971 was clearly dominated by highly erosive precipitation. The departure of annual rainfall erosivity from the multi-year mean in Fiche reached its negative extreme in 1984, where the annual value was lower than the long-term mean by 89%. It can be noted from Fig. 40 that less erosive annual rainfall also occurred in the years 1959 (-78%), 1961 (-61%), 1972 (-44%), 1978 (-60%), 1982 (-65%) and 1983 (-70%).

The departure of summer rainfall erosivity from the multi-year mean at the Fiche weather station shows relatively similar characteristics to annual rainfall erosivity (Fig 40). Summer rainfall erosivity runs almost parallel to the annual rainfall erosivity for most of the observation period. Compared with the multi­ year average, summer rainfall erosivity reached its highest positive deviation in 1955, where it was higher than the multi-year mean by 101%. Secondary positive highs were also noted in 1969 and 1977, where summer values exceeded the multi-year mean by 90% and 82%, respectively. The least erosive summer rainfall occurred in the year 1984, where the summer value was 98% less than the multi-year mean. Summer rainfall with low erosivity potential at Fiche weather station was also detected in 1959, 1961, 1978 and 1980, with summer rainfall erosivity values lower than the multi-year mean by 65%, 92%, 57% and 47%, respectively.

Rainfall and its erosivity in Ethiopia 108

The deviation of summer rainfall erosivity from the corresponding long-term average in Fiche area shows similar pattern to that of summer departure from multi-year mean (Fig. 40). Both attained their extreme departures during the same year. In addition, minor peaks and lows occurred in the same period. It can be confirmed from these results that, in Fiche area, summer rainfall plays the major role in water erosion. It was noted from this similarity between the pattern of summer rainfall erosivity deviations from the multi-year average and long­ term summer mean that intra and inter seasonal variability of rainfall erosivity clearly share common features. Subsequently, both can be applied in water erosion mitigation planning in the area.

125%

-25%

-75%

-125% 1954 1959 1964 1969 1974 1979 1984 1989 1994

Years

IADMM — b— SUDMM SUDSLTA |

Fig. 40: Departures of annual and summer rainfall erosivity from long-term average at the Fiche weather station

Majete, Sheno, Shola Gebeya, Tulo Bolo

The Majete, Sheno, Shola Gebeya and Tulu Bolo weather stations have rainfall erosivity data series of equal length (1962-1996). Subsequently, they were grouped for the characterisation of the departures of rainfall erosivity from long­

Rainfall and its erosivity in Ethiopia 109 term average in each of these areas. As can be revealed from Figs. 41 a through d, the departure of annual rainfall erosivity from the corresponding multi-year mean at the Majete weather station reached its historical highest in 1963 (+53%), Sheno in 1966 (+89%), Shola Gebeya in 1985 (+160%) and Tulu Bolo in 1993 (+116%). Extremely low annual rainfall erosivity at Majete was noted in 1984 (-47%), Sheno in 1967 (-67%), Shola Gebeya in 1963 (-43%) and Tulu Bolo in 1983 (-57%). Except for the years 1962-1964, where annual rainfall erosivity persistently deviated far above the multi-year mean, annual rainfall erosivity at Majete was characterised by clearly alternating highs and lows throughout the period of investigation. Both the most and the least erosive annual rainfall at the Sheno weather station were noted during 1963-1967, which did not recur in the subsequent years with similar magnitude. As compared with the remaining years of investigation, this period was characterised by contrasting highs and lows in annual rainfall erosivity in the area only within few years. In spite of the very high erosive annual rainfall occurrence in 1985 (+160%), departures of annual rainfall erosivity from the multi-year mean at the Shola Gebeya weather station showed relatively uniform behaviour with hardly any extreme lows and highs. Less erosive annual rainfall at the Tulu Bolo weather station was observed in the years 1974 (-36%), 1987 (-35%) and 1995 (-32%).

Maximum positive departures of summer rainfall erosivity from the multi-year mean at the Majete, Sheno, Shola Gebeya and Tulu Bolo weather stations were noted in the years 1964 (+121%), 1966 (+131%), 1985 (+230%) and 1993 (+150%), respectively (Figs. 41 a through d). This suggests that extremely erosive summer rainfall occurred in these areas during these years. As compared with annual rainfall erosivity, the least erosive summer rainfall at the Majete, Sheno, Shola Gebeya and Tulu Bolo weather stations was noted in the years 1984 (-56%), 1967 (-68%), 1972 (-21%) and 1983 (-49%), respectively. Secondary high positive departures of summer rainfall erosivity from the multi­ year mean at the Majete weather station were observed in the years 1975 (+83%), 1985 (+66%), 1988 (+105%) and 1995 (+70%); Sheno in the years 1963 (+71%), 1981 (+62%) and 1985 (+55%); Shola Gebeya in the years 1964

Rainfall and its erosivity in Ethiopia 110 and 1981, (each +60%), 1983 (+81%) and 1986 (84%); Tulu Bolo in the years 1969 (+55%), 1977 (+66%), 1980 (+62%), 1982 (+99%) and 1992 (+126%). Compared with the multi-year mean rainfall erosivity, low summer rainfall erosivity was also observed at Majete weather station in 1972 (-20%) and 1982 (-15%), Sheno in 1990 (-30%) and 1972 (-20%), Shola Gebeya in 1963 (-11%), Tulu Bolo in 1974 (-24%) and 1987 (-21%).

As illustrated by Figs. 41 a, b, c and d, the departure of summer rainfall erosivity from the corresponding long-term average at the Majete weather station attained its peak value in the year 1964 (+65%), Sheno in 1966 (+82%), Shola Gebeya in 1985 (+154%) and Tulu Bolo in 1993 (+103%). Secondary positive highs in the departure of summer rainfall erosivity from the respective long-term mean, the consequence of highly erosive rainfall occurrence, were detected for the Majete weather station in 1988 (+53%) and 1975 (+36%), Sheno in 1965 (+81%) and 1981 (+28%), Shola Gebeya in 1983 (+39%) and 1964 and 1981 (each +23%), Tulu Bolo in 1982 (+61%) and 1992 (+83%). The least erosive summer rainfall at the Majete weather station was recorded in the year 1984 (-67%), Sheno in 1967 (-75%), Shola Gebeya in 1972 (-39%) and Tulu Bolo in 1983 (-59%). Compared with long-term summer average, relatively low rainfall erosivity was detected for the Majete weather station in 1972 (-40%) and 1982 (-37%), Sheno in 1990 (-45%) and 1972 (-37%), Shola Gebeya in 1962 (-33%) and 1984 (-28%), Tulu Bolo in 1974 (-39%) and 1995 (-35%).

The summer rainfall erosivity at the Majete, Sheno, Shola Gebeya and the Tulu Bolo weather stations persistently exceeded the course of multi-year rainfall erosivity during the investigation years (Figs. 41 a-d). This suggests that summer rainfall is the major causative force of water erosion in these areas. As can be revealed from their mostly parallel deviation curves, the departures of summer rainfall erosivity from multi-year mean and the corresponding long­ term summer average have shown similar characteristics in the course of observation period. In addition, simultaneous occurrence of major and minor peaks as well as major and minor lows were detected in these areas in several years.

Rainfall and its erosivity in Ethiopia Ill

Fig. 41 a: Departures of annual and summer rainfall erosivity from long-term average at the Majete weather station

Fig. 41 b: Departures of annual and summer rainfall erosivity from long-term average at the Sheno weather station

Rainfall and its erosivity in Ethiopia 112

2 5 0 %

200%

1 5 0 %

100% ia. 5 0 %

0%

- 5 0 %

1962 1977 1982

Y e a rs

EZZ3ADMM —e—SUDMM - ^ ~ S U D S L T a |

Fig. 41 c: Departures of annual and summer rainfall erosivity from long-term average at the Shola Gebeya weather station

120% -

7 0 % -

20%

- 3 0 %

- 8 0 % 1962 1967 1972 1977 1982 1 9 8 7 1 9 9 2

Y ears

C Z H A D M M —Q— S U D M M —a -^SUDSLTA |

Fig. 41 d: Departures of annual and summer rainfall erosivity from long-term average at the Tulu Bolo weather station

Rainfall and its erosivity in Ethiopia 113

Kombolcha

The departure of annual rainfall erosivity from the multi-year mean at Kombolcha reached its highest in the year 1964, where the annual value was higher than the multi-year mean by 72% (Fig. 42). It was noted that highly erosive annual rainfall at Kombolcha concentrated in the years 1952, 1953 and 1955, with the records exceeding the multi-year mean by 47%, 45% and 38%; respectively. The least erosive annual rainfall at the Kombolcha weather station was noted in the year 1984, where the annual value was less than the multi-year mean by 47%. Annual rainfall with low erosive potential was also observed at the Kombolcha weather station in the years 1962, 1965, 1972 and 1976, where the annual figures were less than the multi-year mean rainfall erosivity by 44%, 28% and 31%, respectively.

The comparison of the course of summer rainfall erosivity with the multi-year average in Kombolcha, as illustrated by Fig. 42, revealed that the maximum departure of summer rainfall erosivity from the multi-year mean occurred in the year 1964, where the value was higher than the multi-year mean by 130%. As compared with annual rainfall, the least erosive summer rainfall at Kombolcha occurred in the year 1984, with summer value 53% lower than the multi-year mean. Less erosive summer rainfall was noted in 1962, with the record lower than the multi-year mean by 32%. As was the case for annual rainfall erosivity, the Kombolcha weather station has experienced high concentration of erosive summer rainfall in the years 1952, 1954, 1953 and 1955, where the values were higher than the multi-year mean by 95%, 71% and 89% in the latter two years, respectively. Similar comparison indicated that the Kombolcha weather station has seen relatively highly erosive rainfall during the years 1969, 1970 and 1988, with summer values exceeding the annual records by 81% in the former and 74% in the latter two years, respectively. The comparison of summer rainfall erosivity with the multi-year mean revealed that summer rainfall is highly responsible for water erosion in Kombolcha area.

The comparison of the course of summer rainfall erosivity with the respective

Rainfall and its erosivity in Ethiopia 116

erosivity at the Wonji weather station, as illustrated by Fig. 43, indicates that most extreme departures of from the multi-year average occurred in the years 1963 and 1975, where summer rainfall erosivity was higher than the multi-year mean by 100% in each of these years. The same comparison shows that the least erosive summer rainfall at Wonji was registered in 1978, where the summer rainfall erosivity value was 50% less than the multi-year mean. It was noted that summer rainfall erosivity dominated the annual rainfall erosivity in the years 1966, 1958 and 1961, with summer values exceeding the multi-year average by 96%, 95% and 86%, respectively. It can be concluded from these results that, compared with annual rainfall, summer rainfall in Wonji area plays the primary role in causing water erosion and, hence, land degradation .

The comparison of the course of summer rainfall erosivity with the respective long-term mean (Fig. 43) showed that the highest positive departures of summer rainfall erosivity at the Wonji weather station occurred in 1963 and 1975. Rainfall erosivity in each of these years exceeded the long-term summer average by 60%. Very high erosive summer rainfall was also recorded in the years 1966 (+56%), 1958 (+55%), 1961 (+48%), 1952 and 1969 (each +41%). The least erosive summer rainfall at Wonji occurred in the year 1978, where the annual value was less than the long-term average by 60%. Similar comparison indicated that summer rainfall in 1953, 1957, 1979 and 1980 was characterised by low erosive potential, where the summer erosivity values were lower than the long- ^ term average by 36% in 1953, 42% in 1957, 38% in 1979 and 36% in 1980. It can be concluded from these results that highly erosive rainfall occurs during summer season in the Wonji area, as compared with the annual period. It can be confirmed from Fig. 43 that rainfall erosivity at the Wonji weather station was characterised by alternating highs and lows throughout the study period. This contrasting behaviour of rainfall erosivity at Wonji is exhibited by both annual and summer rainfall. Fig. 43 depicts departures of rainfall erosivity from long­ term average at Wonji weather station.

Rainfall and its erosivity in Ethiopia 117

120% 100% 80% 60% «3 40% c 20% I 0% Q -2 0 % -40% -60% -80% 1951 1956 1961 1966 1971 1976 1981 Years

IZ Z 1A D M M - b -SUDMM -^r-SUDSLTA|

Fig. 43: Departures of annual and summer rainfall erosivity from long-term average at the Wonji weather station

6. 4. 1. 3 Systematic trend analysis

Rainfall erosivity, like other climatic elements of soil eroding agents (for instance, wind and snow) is subject to variability. Such variability in rainfall erosivity is described as a trend followed by rainfall erosivity in a definite period and can assume different forms lasting from few to many years (R a pp and

S ch o n w iese 1996). Graphical investigation of trends in rainfall erosivity in the central highlands of Ethiopia enables only a broad qualitative description of temporal behaviour of the data series. Such analysis, though comprehensive in nature, does not lead to a conclusive statement regarding the direction of rainfall erosivity time series in a strict sense. Therefore, a systematic statistical trend analysis based on significance test is required to make a quantitatively corroborated and concrete statement about their direction and their degree of strength. The Spearman’s rank statistic is one of the most widely used and robust non-parametric statistical tests against randomness and for the existence of a trend in time series data (M itch ell et al. 1966; Sn ey er s 1990). In addition,

Rainfall and its erosivity in Ethiopia 118 a parametric trend correlation coefficient (Pearson’s correlation coefficient), which is widely used in trend analysis (e. g., R a p p and S c h o n w ie s e 1996), was calculated for the rainfall erosivity data to support the results of non-parametric trend test. Tab. 8 presents the trend correlation coefficients of rainfall erosivity for selected weather stations in the central highlands of Ethiopia.

Tab. 8: Trend correlation coefficients of rainfall erosivity for selected weather stations in the central highlands of Ethiopia

Annual Sum m er W eather Spearm an's Pearson’s Spearm an’s Pearson’s JTp Jrp rD rD Stations rank correlation rank correlation value value value value statistic coefficient statistic coefficient Addis Ababa -0.25 0.02 -0.23 0.03 -0.23 0.02 -0.22 0.03 Debre M arkos -0.21 0.22 -0.22 0.18 -0.20 0.28 -0.22 0.20 Debre Zeit 0.13 0.46 0.12 0.50 0.11 0.53 0.16 0.37 Fiche -0.50 0.002 -0.40 0.01 -1.50 0.32 -0.13 0.42 Ejaji 0.20 0.38 0.14 0.44 0.24 0.27 0.26 0.24 Majete -0.14 0.43 -0.20 0.25 -0.13 0.45 -0.21 0.23 Kom bolcha -0.26 0.10 -0.26 0.07 -0.28 0.08 -0.32 0.046 Sheno -0.23 0.20 -0.32 0.07 -0.27 0.13 -0.26 0.13 Shola Gebeya 0.03 0.87 0.13 0.48 0.73 0.68 0.13 0.45 Tulu Bolo 0.11 0.51 0.25 0.14 0.12 0.50 0.22 0.21 Wonji -0.03 0.85 -0.08 0.66 -0.06 0.72 -0.11 0.52

The correlation coefficients in Tab. 8 reveal that long-term annual rainfall erosivity at the weather stations Addis Ababa, Debre Markos, Fiche, Majete, Kombolcha, Sheno and Wonji followed a decreasing course during the period of investigation. Both parametric and non-parametric trend coefficients, with P values of 0.03 and 0.02, respectively, confirm that long-term annual rainfall erosivity at the Addis Ababa weather station has followed significantly declining course, at 5% level of significance. The non-parametric trend correlation coefficient determined for the long-term annual rainfall erosivity data series of the Fiche weather station shows that rainfall erosivity underwent significantly declining course, at 1% level of significance (P value 0.002), and the parametric trend correlation coefficient indicates a significant decline in annual rainfall

Rainfall and its erosivity in Ethiopia 119 erosivity, at 5% significance level (P value 0.011). Long-term annual rainfall erosivity at all the other weather stations (Debre Zeit, Ejaji, Shola Gebeya and Tulu Bolo) has followed increasing course during the period of investigation. However, the increasing trend was found to be statistically not significant.

Long-term summer rainfall erosivity at the Addis Ababa, Debre Markos, Fiche, Majete, Kombolcha, Sheno and Wonji weather stations pursued a decreasing course during the study period and increasing course at all the remaining weather stations. The parametric and non-parametric trend correlation coefficients (P values of 0.03 and 0.02, respectively) indicate that rainfall erosivity at the Addis Ababa weather station has significantly declined at 5% level. It was confirmed from the parametric correlation coefficient that long­ term summer rainfall erosivity at Kombolcha weather station has significantly decreased, at 5% level of significance (P value 0.046), during the period of study. The declining trend in long-term summer rainfall erosivity at all the remaining weather stations was found to be statistically insignificant. This can be explained by the increasingly erosive precipitation in these areas. Long-term summer rainfall erosivity at Debre Zeit, Ejaji, Shola Gebeya and Tulu Bolo weather stations followed increasing trend which was, however, found to be statistically not significant at all of these weather stations.

It can be concluded that a concurrent and statistically significant decline in long­ term annual as well as summer rainfall erosivity in the study area was detected only for the Addis Ababa area. Statistically significant declining trend in long­ term annual and insignificant declining trend in long-term summer rainfall erosivity at the Fiche weather station can be explained by a remarkable decline in precipitation during the short rainy seasons. These led to the diminishing in erosive effect of precipitation.

Rainfall and its erosivity in Ethiopia 120

6. 4. 1.4 Autocorrelation and persistence analysis of long-term rainfall erosivity

Many meteorological observations for a given area are samples of the state of nature over time and, thus, are not from a designed, controlled experiment

(M eek et al. 1999; Pa v lo po u lo s and G ritsis 1999). The same applies to rainfall erosivity since it is a derivative of climate variable, that is precipitation. Autocorrelation is one of several problems that may be at hand in the analysis of such data. This was examined for the long-term annual and summer rainfall erosivity data of selected weather stations in the central highlands of Ethiopia. Tabs. 9 and 10 summarise the results.

Rainfall and its erosivity in Ethiopia 121

Tab. 9: Autocorrelation of long-term annual rainfall erosivity

W eath er L ag M inim um S tan d ar L ag M ax im u m S tan d ard Prob. Prob. statio n s N r. AC e rro r N r. AC erro r A ddis A b ab a 15 -0 .106 0.0 9 2 0.080 1 0.301 0.099 0.002 Debre Markos 7 -0 .244 0.143 0.216 9 0.108 0.138 0.364 Debre Zeit 12 -0.342 0.134 0.094 1 0.388 0.1 6 4 0.018 Ejaji 16 -0.247 0.121 0.835 9 0.185 0.145 0.969 F iche 16 -0.281 0.118 0.003 2 0.491 0.144 0.000 Kombolcha 10 -0.240 0 .1 3 4 0.452 16 0 .0 7 6 0 .1 2 0 0.265 M ajete 3 -0.311 0.157 0.052 1 0.312 0.162 0.055 S heno 9 -0.209 0.143 0.491 3 0.333 0 .159 0.170 S hola 13 -0.161 0.131 0.966 2 0.136 0.160 0.679 Tulu Bolo 6 -0.232 0.150 0.402 12 0 .2 5 4 0.133 0.299 Wonji 1 -0.401 0.164 0.015 9 0.201 0.143 0.059

Tab. 10: Autocorrelation of long-term summer rainfall erosivity

W eath er L ag M in im u m S tan d ard L ag M ax im u m S tan d ard P rob. Prob. statio n s N r. AC erro r N r. AC erro r Addis Ababa 9 -0.093 0 .095 0.514 11 0.229 0 .094 0.221 Debre Markos 16 -0.255 0.120 0.278 1 0.238 0 .156 0.126 D ebre Z eit 13 -0 .3 6 7 0.131 0.003 1 0.465 0 .1 6 4 0.005 Ejaji 10 -0.236 0.132 0.677 4 0.239 0.150 0.574 F iche 10 -0.237 0.130 0.186 2 0.248 0 .144 0.069 K o m b o lch a 8 -0.258 0 .1 3 8 0.691 11 0.248 0.131 0.506 Majete 3 -0.276 0.157 0.050 1 0 .3 3 7 0.162 0.038 Sheno 2 -0.254 0.162 0.264 3 0.225 0 .159 0.199 S hola 13 -0.189 0.130 0.978 2 0.109 0.1 6 0 0.770 Tulu Bolo 6 -0.252 0.150 0.330 12 0.311 0.133 0.122 Wonji 1 -0.326 0.1 6 4 0.047 6 0.191 0.151 0.330

The Autocorrelation Coefficients (AC) of long-term annual rainfall erosivity for selected weather stations in the central highlands of Ethiopia vary from -0.401, determined for the Wonji weather station, to 0.491, calculated for the Fiche weather station, with standard errors (S. error) of 0.164 and 0.144; probability value (Prob.) of 0.015 and 0.00, respectively (Tab. 9). These values are found to be statistically significant at 5% and 1% significance level, respectively. It was noted that the ACs of long-term annual rainfall erosivity for most of the weather stations investigated are not statistically significant. As shown in the annex (Fig.

Rainfall and its erosivity in Ethiopia 124 weather stations at higher altitude were exposed to relatively less mean annual and summer rainfall erosivity, compared with those stations situated at lower altitude. For instance, Mararo, situated at 2 940 m.a.s.l, is characterised by less erosive long-term annual and summer rainfall, with erosivity values of 162.00 mm (variability coefficient 37%) and 201.06 mm (variability coefficient 36%), respectively. In contrast, Fiche, located at 2 750 m.a.s.l., has annual and summer erosivity values of 256.98 mm (variability coefficient 41%) and 292.83 mm (variability coefficient 44%), respectively; Debre Work, found at 2 740 m.a.s.l., has long-term mean annual and summer rainfall erosivity of 178.48 mm (variability coefficient 30%) and 232.70 mm (variability coefficient 29%), respectively; Alem Ketema, located at 2 280 m.a.s.l., has annual and summer rainfall erosivity values of 243.28 mm (variability coefficient 33%) and 325.82 mm (variability coefficient 25%), respectively. The Goba weather station, located at an altitude of 2 710 m.a.s.l, was found to have erosivity value less than those found at the lowest altitudinal boundary of the central Ethiopian highlands (1 500 m.a.s.l), namely Bora, Keleta and Wonji weather stations. Weather stations located at equal height were noted to have different long-term mean annual and summer rainfall erosivity values. An example for this are, the Bora and Keleta weather stations which have different long-term mean annual and summer rainfall erosivity values, with different variability coefficients (see annex, Tabs. 17, 20 and 21). Fig. 44 shows the relationship between long-term annual and summer rainfall erosivity and altitude (m.a.s.l.).

Rainfall and its erosivity in Ethiopia 125

450 - r ...... , = 0.0155Altitude+ 143.59

400 ...... a R2 = 0.0123 N = 44

350 + ...... a YSUFm = 0.0137Altitude+ 189.42 a • * 300 ...... R- = 0.0044 N = 44

250 ...... • a A...... i A —A- ♦»- 200 - : _ : ' A -B IS© j " t • 100 ------I f l

50 -j__i__‘ i 1 * 1 I 1 1 1 I 1500 1700 1900 2100 2300 2500 2700 2900

Altitude (m.a.sl.)

Mean AFm a Mean S U F m ----- Linear (Mean SUFm)------Linear (Mean AFm) |

Fig. 44: Relationship between long-term mean rainfall erosivity and altitude for selected weather station in the central highlands of Ethiopia

Compared with elevation, no clear pattern of distribution of long-term rainfall erosivity was observed in the central highlands of Ethiopia (Fig. 44). The rainfall erosivity data points are distributed along the linear trend lines without showing a clearly defined pattern. The slopes of trend lines (annual [0.0155] and summer [0.0137]) show that for the weather stations studied, altitude change has hardly any influence on the erosivity of rainfall. The same can be confirmed from the determination coefficients of the trend equations. It was noted that only 1.23% of long-term annual and 0.44% of summer rainfall erosivity are explained by elevation. The relationship between rainfall erosivity and elevation detected in this study for the central highlands of Ethiopia is similar to that found by

H e l l d e n and E k l u n d h (1988), E k l u n d h and P il e s j o (1990) for mean annual rainfall vs. elevation for the whole Ethiopian highlands.

Rainfall and its erosivity in Ethiopia 126

6. 4. 2. 2 Horizontal characteristics o f rainfall erosivity

Variability of rainfall erosivity across the landscape causes variability in soil degradation within the same landscape depending on terrain characteristics, land-use and land cover of an area. Subsequently, spatial behaviour of rainfall erosivity has a direct implication on land and water management, land-use and agricultural development. The knowledge of spatial behaviour of rainfall erosivity is especially useful in resources management decision making and environmental conservation policy design as well as implementation. Several methods are used to investigate spatial variability of agro-meteorological variables in relation to their effect on the soil environment. Coefficient of variation is the most widely used approach to investigate spatial characteristics of rainfall and other weather variables affecting the soil system (e.g., B e c k e r and S m ith 1990; H u b b a rd 1994; C a m a rg o and H u b b a rd 1999; R e n s c h le r et al. 1999). In this study, the method of variability coefficient is used to investigate the spatial behaviour of rainfall erosivity in the central highlands of Ethiopia. Figs. 45 to 48 illustrate the relationship between variability coefficient of annual and summer rainfall erosivity for selected weather stations in the central highlands of Ethiopia.

Rainfall and its erosivity in Ethiopia 127

Latitude (degrees and minutes)

Latitude (degrees and minutes)

• AFmCV vs. Latitude AFm Annual rainfall erosivity (mm) ------Linear fit: CV Coefficient of Variation AF CV = 0.34 + -0.0044Lat, r = 0.0011

Fig. 45: Relationship between latitude and variability in annual rainfall erosivity - regional diminution

Rainfall and its erosivity in Ethiopia 128

L o n g itu d e 0 2 4 6 8 10

Longitude (degrees and minutes)

® AFmCV vs. Longitude AFm Annual rainfall erosivity (mm) ------Linear fit: CV Coefficient of Variation AFmCV = -0.82 + 0.03Long, r = 0.14

Fig. 46: Relationship between longitude and variability in annual rainfall erosivity - regional dimension

Rainfall and its erosivity in Ethiopia 129

Latitude

10

4 .y sO ou U

Latitude (degrees and minutes)

SUFmCV vs. Latitude SUF Summer rainfall erosivity (mm) ■ Linear fit: CV Coefficient of Variation SUF CV = 0.57 -0.03Lat, r = -0.17

Fig. 47: Relationship between latitude and variability in summer rainfall erosivity - regional dimension

Rainfall and its erosivity’ in Ethiopia 130

L o n g itu d e 10

Longitude (degrees and minutes)

SUFmCV vs. Longitude SUFm Summer rainfall erosivity • Linear fit: CV Coefficient of Variation SUF CV = -1.42 + 0.045Long, r = 0.17

Fig. 48: Relationship between longitude and variability in summer rainfall erosivity - regional dimension

The variability in annual rainfall erosivity shows weekly decreasing tendency with increasing latitudinal location of observation points (Fig. 45). However, it was noted that their relationship, with a correlation coefficient of 0.07, is statistically not significant at a 5% level (P value = 0.64). The variability increases slightly eastwards (Fig. 46). The increasing tendency, with a value of correlation coefficient of 0.24 was, however, found to be statistically not significant at a level of 5% (P = 0.12). As illustrated by the scatter diagrams (Figs. 45 to 48), relatively slight difference in general spatial characteristics was noted in the variability of summer rainfall erosivity from that of annual values. The variability in summer rainfall erosivity shows slightly decreasing tendency with increasing latitudinal location of observation points, with a correlation

Rainfall and its erosivity in Ethiopia

I 131 coefficient of -0.11 (Fig. 47). The decreasing tendency was, nevertheless, found to be statistically not significant at 5% level (P = 0.56). The variability in summer rainfall erosivity shows increasing tendency towards the east with a correlation coefficient of 0.42 (Fig. 48). This tendency, unlike that of annual rainfall erosivity, was found to be statistically significant at a level of 1% (P = 0.005). Concerning the longitudinal variability, the spatial pattern of the variability of annual and summer rainfall erosivity in the central highlands of

Ethiopia contrasts with that of precipitation. Eklundh and Pilesjo (1990) reported that mean annual precipitation in Ethiopia decreases to the east direction. Similar results were reported by Hurni (1982) and Seleshi and

Demaree (1995). Contrasting relationship between the variability in annual rainfall and its erosivity was noted with increasing latitude. Contrary to that, the latitudinal variability of summer values agrees with that of rainfall.

The results of multiple regression analysis indicated that altitude, latitude and longitude together explain 65% of spatial variability of annual and 10% of summer rainfall erosivity in the central highlands of Ethiopia. The analysis of variance showed that these geographic variables do not significantly explain, at a level of 5%, the spatial characteristics of annual and summer rainfall erosivity, with P values of 0.43 and 0.22, respectively. This statement agrees with the conclusion reached by Eklundh and Pilesjo (1990) for precipitation in the Ethiopian highlands.

It can be generalised that rainfall erosivity varies less in the western part of the central Ethiopian highlands than in the east. Furthermore, a uniform latitudinal pattern in the variability of both annual and summer values was detected. The spatial pattern of the variability in rainfall erosivity in the region seems to be closely associated with conditions producing precipitation and its variability, as reported by T ato (1964). Therefore, additional factors behind the spatial variability of rainfall erosivity in the study area should be sought in order to achieve highly reliable and quality models for land and water management planning as well as practical application to environmental protection.

Rainfall and its erosivity in Ethiopia 132

6. 5 Rainfall erosivity classes for the central highlands of Ethiopia

A generalised and simplified presentation of rainfall erosion risk is a widely used method in the investigation of potential soil erosion risk and the resulting

land degradation hazard (F A O 1977; O l d e m a n et al. 1991). Such presentation is

based on classes of rainfall erosion index (e.g., G r if f it h s and R ic h a r d s 1989;

S a u e r b o r n 1994; O d u r o -A f r iy ie 1996). Rainfall erosivity classes developed for the central highlands of Ethiopia based on clusters of cases are given by Tabs. 11 and 12 as follows.

Tab. 11: Classes of annual rainfall erosion risk for the central highlands of Ethiopia

Class Annual rainfall Range of rainfall Number of cases number erosion risk class erosivity (mm) (N) % 1 Very low <100 0 0 2 Low 101-150 3 3 3 Moderate 151-200 36 36 4 High 201-250 51 51 5 Very high 251-300 5 5 6 Extremely high >301 5 5 Total 100 100

Tab. 12: Classes of summer rainfall erosion risk for the central highlands of Ethiopia

Summer rainfall Range of rainfall erosivity Number of Class number erosion risk class (mm) cases

(N) % 1 Very low <150 1 1 2 Low 151-200 8 8 3 Moderate 201-250 31 31 4 High 251-300 42 42 5 Very high 301-350 13 13 6 Extremely high >351 5 5 Total 100 100

Rainfall and its erosivity in Ethiopia 133

Six reference classes of rainfall erosion risk ranging from very low to extremely high were established for annual and summer rainfall erosivity, respectively for the central highlands of Ethiopia (Tabs. 11 and 12). These can be utilised as a standard reference to define the degree of water erosion in the region. It was noted that more than fifty percent of the observed cases of annual rainfall erosivity fell into the high class. The second largest class of annual rainfall erosion risk was found to be the moderate class, with 36% of the observed cases falling into this category. For the summer period, the largest category was found to be the high class (42%), followed by the moderate category (31%). It can be noted that the high class of water erosion risk constitutes the majority of the cases observed for both annual and summer period. The weather stations in the central highlands of Ethiopia are assigned to the various classes as illustrated by Fig. 49.

40%-

35%-

.2 30%-

25%-

% 20%-

15%-

10%-

Very low Low Moderate High Very high Extremely high Classes of rainfall erosivity

□ Annual: Percentage of weather station SI Summer: Percentage of weather station

Fig. 49: Percentage of weather stations in the central Ethiopian highlands assigned to rainfall erosion risk classes

Rainfall and its erosivity in Ethiopia 134

The majority of the weather stations in the central highlands of Ethiopia fell into the low to high annual rainfall erosivity classes (Fig. 49). Accordingly, 39% of the weather stations fell into the low and 50% into the moderate to high risk category. Only 2% of the weather stations fell in the extremely high class. For the summer period, most of the weather stations were found to be in the very low to high category. Thus, 23% of the investigated weather stations were found in the very low, 20% in the low, 16% in the moderate and 27% in the high class. Only 5% of the weather stations fell into the extremely high class. Of the weather stations studied, 9% were noted to be in the very high water erosion risk class for both annual and summer period. Details of class numbers assigned to different weather stations in the study area are given in the annex (Tab. 21). The fact that the majority of the weather stations in the region are categorised into the low to moderate rainfall erosion risk classes agrees with the results of

G r if f it h s and R ic h a r d s (1989). The authors concluded that precipitation in Ethiopia is not highly erosive by tropical standards, and that the extremely high rate of soil erosion is the consequence of land mismanagement. However, this statement undermines the acceleration of land degradation due to water erosion following even the slightest land mismanagement. The classification of water erosion risk in the central highlands of Ethiopia agrees, to a large extent, with the findings of O l d e m a n et al. (1991). According to the authors, most areas of the central highlands of Ethiopia fall into the medium to very high class, whereby the most the areas seem to be in the very high risk category. The

O l d e m a n et al. (1991) classification is very general. Therefore, it is recommended that the classes established by this study be adopted for Ethiopia, since they are relatively specific to the study region.

Rainfall and its erosivity in Ethiopia 135

7 IMPLICATIONS OF RAINFALL VARIABILITY AND RAINFALL EROSIVITY ON SURFACE FLOW AND LAND AND WATER MANAGEMENT

The long-term variability in precipitation and its erosive potential have significant geoecological implications. Particularly they influence the pedosphere, hydrosphere and the biosphere qualitatively as well as quantitatively. The following section deals with the influence on flow regime, which is part of the hydrosphere.

7. 1 The influence of rainfall variability on surface flow

The consequences of climate variability on global water resources is a subject of major concern among scientists and policy makers. The variability of river flows and lake water depth is the major characteristics of surface hydrology which is gaining special focus in this regard (M icklin 1996). The impact of climate variability on surface hydrology and water resources was investigated by several authors (e.g., K u n d zew ic z and So m ly od y 1997; M a jo r and Fred eric k 1997;

W o o d et al. 1997; K iely 1999). P fister et al. (2000) investigated the trend of the relationship between rainfall and hydrological time series in the Alzette

River Basin in Luxembourg. F a r qu ha rso n and SUTCLIFFE (1998) concluded that temporal variability as well as occasional periodicity of sub-saharan African river flows is primarily induced by temporal and inter-temporal rainfall variability. S ele sh i and D em aree (1995) found out that rainfall variability in the Ethiopian and Eritrean highlands significantly affected the flow regimes of the Blue Nile measured at different gauging stations. M o r ed a and Bau w ens (1998) reported that rainfall time series of the Addis Ababa weather station was significantly correlated with discharge measurement of the Awash River at its upper watershed.

The variability in surface hydrologic regime, expressed as overland flow, has a direct implication for rainfall erosivity. SCHMIDT et al. (1999) indicated that the erosive impact of overland flow and droplets is a function of the momentum flux

Rainfall and its erosivity in Ethiopia 136 exerted by overland flow and droplets, respectively. Rainfall variability, through its direct consequence on surface flow regime, influences the whole soil water system. Yoo et al. (1998) found out that the impact of rainfall on soil moisture regime is widespread during storm, but decreases with falling precipitation amount.

The impact of long-term mean rainfall variability with large area coverage on river flow regime and lake depth in the central highlands of Ethiopia has not been well investigated yet. In addition, little consideration was given to this subject in land and water management research context. This study will facilitate the understanding of the impact of climate change and variability on the water resources of the country. It will also enhance in the planning exercises of water conservation measures for agricultural development and other economic uses. Furthermore, it enables to devise soil and water resources development as well as management plans based on improved knowledge of the relationship between surface hydrologic regime, temporal and inter-temporal rainfall variability. The understanding of the relationship between rainfall variability and surface flow regime fosters watershed approach to ecological management and/or agricultural development. Furthermore, the application of this knowledge to land and water management enhances to make realistic land resources management strategies. The aims of this section are: a To analyse the relationship between seasonal as well as long-term areal mean rainfall and surface flow regime ® To assess the temporal variability of surface flow regime ® Compare and contrast the long-term trend of surface hydrologic regime and rainfall.

7.1.1 Materials and Methods

The gauging stations used to examine the relationship between long-term areal mean rainfall variability and surface flow regime in the central highlands of Ethiopia are located in the Awash River basin of Ethiopia. This river was chosen

Rainfall and its erosivity in Ethiopia 137

for this study because of its high economic and ecological importance for the country. It is considerably exploited for agricultural and industrial development. Moreover, it is the largest and most representative for the study area. The Awash basin gets 65-70% of its annual rainfall during the main rainy season of the region and the remaining during the short rainy seasons. The major proportion of surface runoff in this area drains to the Awash River and its tributaries. Subsequently, the Awash River is a destination of sediments eroded from large areas of the central highlands of Ethiopia. Tab. 13 shows the gauging stations used to assess the relationship between rainfall variability and surface flow regime in the central highlands of Ethiopia.

Tab. 13: Selected gauging stations in the central highlands of Ethiopia

Gauging station Latitude Longitude Elevation Drainage Y ears o f (m) area (km2) observation H om bole 8°23' 38°47' 2300 7722.50 1986-1997 K essem 9 ° 1 0 ' 3 9 ° 0 4 ' 2800 50.00 1986-1997 Melka Kunture 8 °4 2 ' 3 8 ° 3 6 ' 2332 4 4 5 6 .0 0 1986-1997 M odjo 8 °3 6 ' 3 9 ° 0 5 ' 2175 1264.40 1989-1998 T eje 8 °5 1 ' 3 8 °2 5 ' 2569 66 2.50 1982-1988

Data set and data analysis

Hydrometric records spanning from 1982-1997 were obtained from the Ministry of Water Resources (MoWR) as hard copies. The data were visually controlled for quality and quantity and gauging stations with frequently missing measurements were omitted from further investigation. The basic criteria to filter out the stations were length of record and regularity of keeping. The gauging stations thus selected for further analysis are shown in Tab. 13. The length of hydrometric measurements at the various stations were compared with areal precipitation time series and matching years which have records adopted for the study. Data entry and preliminary preparation were accomplished using various statistical packages.

Basic statistics of gauge measurements was calculated for each station.

Rainfall and its erosivity in Ethiopia 138

Graphical comparison and correlation analysis between long-term hydrometric data and areal mean rainfall records were made. The purpose was to visually investigate and quantify the relationship between surface flow records and long­ term areal mean rainfall in the study area. Time series analysis was conducted to assess changes in surface runoff in the course of investigation period. Accordingly, the trend in areal mean rainfall and hydrometric records visually perceived from graphical analysis was tested for significance using parametric and non-parametric statistical methods, as suggested in the literature (e.g.,

K o t h y a r i et al. 1997; P f i s t e r et al. 2000).

7.1.2 Results and discussion

The magnitude of surface runoff at the various gauging stations in the central highlands of Ethiopia varies considerably. This can be best explained by spatial variability of rainfall, terrain characteristics, soil physical properties, soil depth, soil water storage capacity and land-use practices in the area. Tab. 14 shows summary of statistical characteristics of hydrometric records for the selected gauging stations.

Tab. 14: Statistical characteristics of hydrometric records at the selected gauging stations

G auging Annual Sum m er stations Coefficients Coefficients N Mean Stand. M ean Stand. CV CV (mill, m 3) dev. Skewness Kurtosis (mill, m 3) dev. Skewness Kurtosis (%) (%) Hom bole 12 583.69 614.44 105 0.75 -1.39 908.10 371.18 41 1.74 1.23 Kessem 12 33.07 13.04 39 0.01 0.33 26.65 10.96 41 0.63 1.33 M elka K unture 12 499.58 159.50 32 1.79 3.44 726.17 200.52 28 0.16 -0.31 Modjo 10 150.94 121.56 81 1.89 3.28 111.90 106.92 96 1.96 3.48 Teji 7 113.85 41.81 37 o:76 -0.44 89.84 39.17 44 1.09 1.72

The highest mean annual and summer runoff records were noted for the Hombole gauging station (Tab. 14). A mean annual and summer value of 582.69 mill. m? (N = 12, standard deviation 614.44 mill, m3) and 908.10 mill, m3 (N =

Rainfall and its erosivity in Ethiopia 139

12, standard deviation 371.18 mill, m3), respectively were determined for Hombole. The lowest values were found for the Kessem gauging station, with mean annual and summer runoff records of 33.07 mill, nr (N = 12, standard deviation 13.04 mill, n r) and 26.65 mill, nr' (N = 12, standard deviation 10.96 mill, nr), respectively. The variability coefficient for the annual period varies from 32%, observed for Melka Kunture, to 105%, observed for Hombole gauging station. For the summer period, the value of CV varies from 28%, observed for Melka Kunture, to 96%, observed for Modjo. The skewness coefficients of annual and summer runoff vary from 0.01 to 1.89 and 0.16 to 1.96, respectively. The kurtosis coefficients for the annual and summer runoff records vary from -1.39 to 3.44 and -0.31 to 3.48, respectively. Generally, surface runoff is the highest at the Hombole and the lowest at the Kessem gauging station. Moreover, Hombole is characterised by the highest relative variability in annual runoff and Modjo by the highest summer relative variability. The annual and summer runoff records show the least variablity at Melka Kunture gauging station. The results of descriptive statistical analysis show that annual and summer runoff measurements for the studied gauging stations are approximately normally distributed.

Relationship between long-term areal mean rainfall and hydrometric records

Graphical analysis of the relationship enables to visually inspect how precipitation and runoff measurements are related. Figs. 50 through 54 show the results of graphical illustration of the relationship between long-term rainfall and the hydrometric records at the selected gauging stations.

Rainfall and its erosivity in Ethiopia 140

Years

Fig. 50 a: Comparison of mean annual rainfall and surface runoff at the Hombole gauging station

Years

Fig. 50 b: Comparison of mean summer rainfall and surface runoff at the Hombole gauging station

Rainfall and its erosivity in Ethiopia 141

Fig. 51 a: Comparison o f mean annual rainfall and surface runoff at the Kessem gauging station

Years

Fig. 51 b: Comparison of mean summer rainfall and surface runoff at the Kessem gauging station

Rainfall and its erosivity in Ethiopia 142

Fig. 52 a: Comparison of mean annual rainfall and surface runoff at the Meleka Kunture gauging station

Fig. 52 b: Comparison of mean summer rainfall and surface runoff at the Melka Kunture gauging station

Rainfall and its erosivity in Ethiopia 143

Fig. 53 a: Comparison of mean annual rainfall and surface runoff at the Modjo gauging station

Fig. 53 b: Comparison of mean summer rainfall and surface runoff at the Modjo gauging station

Rainfall and its erosivity in Ethiopia 144

Fig. 54 a: Comparison of mean annual rainfall and surface runoff at the Teji gauging station

Fig. 54 b: Comparison of mean summer rainfall and surface runoff at the Teji gauging station

Rainfall and its erosivity in Ethiopia 145

The implication of the variability in long-term areal mean annual and summer rainfall on the flow regime in the central highlands of Ethiopia can be noted from the results of the gauging stations examined (Figs. 50 to 54). The trend of precipitation was directly reflected on surface runoff at the Kessem (Figs. 51a and b), Melka Kunture (Figs. 52a and b) and Teji (Figs. 54a and b) gauging stations. Irregular characteristics were noted at the Modjo gauging station, where surface runoff only slightly fluctuated from 1991-1995, as compared with marked rise and fall in the long-term mean annual and summer rainfall data series (Figs. 53a and b). Annual runoff series at the Hombole gauging station exhibited a constant trend between 1991 and 1998, where surface runoff showed only a small change. However, it can be confirmed that the trend of summer rainfall series at Hombole was directly reflected on surface runoff (Figs. 50a and b). In order to support the results of visual inspection with quantitative facts, trend correlation analysis between hydrometric records and rainfall for the selected gauging stations was made. Tab. 15 shows the results.

Tab. 15: Cross correlation coefficients between hydrometric records and areal rainfall

Annual Summer Gauging stations N Correlation Correlation P values P v alues coefficients coefficients H o m b o le 12 -0.01 0.97 0.45 0.09 K essem 12 0.60 0.86 0 .1 0 0.74 Melka Kunture 12 0.70 0.01 0.53 0.07 M odjo 9 0.60 0.09 0.40 0.28 T eji 7 0.30 0.51 0.60 0.15

The cross correlation coefficients between long-term areal mean annual rainfall and the respective surface runoff for the selected gauging stations in the study area (Tab. 15) vary from -0.01 (P value 0.97) to 0.70 (P value 0.01). For the summer period, the cross correlation coefficients vary from 0.10 (P value 0.74) to 0.60 (P value 0.15). For the annual period, statistically significant relationship between long-term areal mean rainfall and long-term mean surface runoff was only noted at Melka Kunture, with a correlation coefficient of 0.70 (a = 5%, P

Rainfall and its erosivity in Ethiopia 146 value 0.011). The relationship between long-term mean annual and summer surface runoff and the corresponding rainfall in all the remaining gauging stations is found to be statistically not significant. It can be implied from these results that point precipitation records might have better correlation with surface runoff than do areal rainfall. However, as was clearly noted from visual analysis of the graphs (Figs. 50 to 54), long-term mean areal rainfall has clear consequence on long-term surface runoff in the central highlands of Ethiopia. This statement can be supported by the results obtained at a specific location.

For example, M o r e d a and B a u w e n s (1998) reported that rainfall variability had great impact on surface flow regime at the Hombole gauging station.

SELESHI and D e m a r e e (1995) indicated that rainfall variability in the Ethiopian and Eritrean highlands echoed on the annual flow regime of the Blue Nile at a gauging station in Khartoum. Conway et al. (1998) found that rainfall variability over the Ethiopian highlands was responsible for the high Nile flows during the 1890s.

Systematic trend analysis o f hydrometric records at selected gauging stations

The trend correlation coefficients of long-term areal mean annual and summer rainfall and the respective surface runoff during the same years (1982-1997) were compared. The purpose was to check if there is statistically significant concurrent trend in long-term areal mean rainfall and the corresponding surface runoff for the selected gauging stations in the study area. Tab. 16 illustrates the non-parametric trend correlation coefficients of precipitation and surface runoff for the period of investigation.

Rainfall and its erosivity in Ethiopia 147

Tab. 16: Non-parametric trend correlation coefficients of surface runoff and rainfall

Annual Summer Runoff Rainfall Gauging stations N Runoff Rainfall Cor. P Cor. P value Cor. P Cor. P coef. value coef. coef. value coef. value Hombole 12 -0.76 0.004 -0.09 0.78 0.350 0.27 0.48 0.12 Kessem 12 -0.03 0.930 -0.09 0.78 -0.007 0.98 0.48 0.12 Melka Kunture 12 -0.11 0.750 -0.09 0.78 0.040 0.89 0.48 0.12 Modjo 10 0.20 0.580 0.10 0.79 0.430 0.21 0.57 0.11 Teji 7 -0.07 0.880 0.54 0.22 -0.110 0.82 0.18 0.70

A concurrent decline in long-term areal mean annual rainfall and the corresponding mean surface runoff was noted at the Hombole, Kessem and Melka Kunture gauging stations (Tab. 16). However, the declining trend in surface runoff was statistically significant, at 1% level, only for the Hombole gauging station, with a P value of 0.004. The trends for all the other gauging stations are not statistically significant. The Teji gauging station has experienced contradicting trends in rainfall and surface runoff. The decline in surface runoff, as opposed to increase in rainfall, might be due to the diversion of runoff from the measuring point or retention in the watershed area. The trends are, however, statistically not significant. Further investigation is required to find out the reasons for this behaviour. Concurrently increasing trend in long-term areal mean annual precipitation and surface runoff was observed at the Modjo gauging station. Nevertheless, the trend was found to be statistically not significant. The precipitation data compared with runoff have not exhibited a significant declining trend at all the gauging stations.

For the summer period, a declining trend in long-term mean surface runoff was found only for the Kessem and Teji gauging stations. Nonetheless, the trend does not seem to be a result of decline in long-term areal mean summer rainfall. For all the remaining gauging stations, a concurrently increasing trend in long­ term areal average summer rainfall and the corresponding mean surface runoff was noted. The increasing trend was, however, found to be not statistically significant.

Rainfall and its erosivity in Ethiopia 148

Generally, it can be noted that the trend in long-term surface runoff for the selected gauging stations in the central highlands of Ethiopia can not be fully explained only by long-term areal mean rainfall variability. Subsequently, it is postulated that change in surface runoff at each gauging station can be more explained by local precipitation characteristics, soil physical properties; mainly soil water storage capacity, land-use and cover of the area than the areal precipitation. B u ll et al. (2000) arrived at similar conclusion for south-eastern Spain. According to them, lithology, morphology and land-use are factor combination which strongly influence the generation of floods. O po k u-

A n k o m a h and A m isigo (1998), based on their investigation in Ghana, concluded that change in surface runoff is affected by infiltration to ground water and evapotranspiration during hydrological process. They also noted that the same should be linked with climate change.

7.1.3 Conclusion and recommendations

Rainfall variability in the central highlands of Ethiopia has clear implication on surface flow regime. However, the consequence is not uniform at all the gauging stations investigated, and it varies for the annual and summer period. Even though statistically not significant, a decline in annual surface runoff was noted at most of the stations. The trends in summer surface runoff are not statistically significant. Subsequently, based on the annual trends, it can be recommended that water management be seriously practised to avoid water deficit. Areal precipitation used for this study does not give full explanation of the variability in flow regime in the study area. Increased gauging station density with reasonable spatial distribution can substantially improve the assessment results of the temporal and inter-temporal variability of surface flow regime. This will enhance to make more precise decision regarding sustainable land and water management. In addition, detailed assessment should be conducted on watershed hydrology and meteorology, soils, land-use and cover for careful land and water management planning.

Rainfall and its erosivity in Ethiopia 149

7. 2 Land and water management in a context of sustainability

7. 2. 1 The context

Progressive land degradation threatens the agro-ecology, most crop and pasture land in the highlands of Ethiopia. The major type of degradation is that caused by water erosion which is a result of erosive rainfall. This is aggravated by deforestation, overgrazing and land mismanagement (B r o w n 1973; J u t z i 1990). Water erosion has already caused adverse effects in the Ethiopian highlands. Soil depth is rapidly sinking, water-holding capacity of the soil is diminishing, crop water availability is dwindling, nutrient level as well as organic matter in the soil are vanishing and crop productivity is continuously dropping (FAO 1986). The degradation of crop land is of special concern, since agriculture is the mainstay of the country’s economy and food self sufficiency as well as poverty minimisation are the most high ranking national goals (FAO

1998; B l o c k 1999; W o r l d B a n k 2000). The Ethiopian farmers have long traditional experience in soil and water management practices (see K r u g e r et al. 1996; G e b r e -M ic h a e l 1998; O s m a n et. al. 2000). These, however, turned out to be less efficient with time. To compensate for the inefficiency of the traditional techniques, immediate actions were taken. The steps involved the development, adoption and implementation of wide ranging water erosion aversion technologies. These encompassed biomass mulching, no-tillage, ridge tillage, agroforestry systems, terracing, grass strips, crop rotation, and combinations of these when required. Nevertheless, the modem techniques were neither sustainable, nor socially acceptable or useful. Former natural resources policies disregarded the socio-economic priorities of the beneficiaries. In spite of this, a critical review of the Indigenous Land and Water Management Technologies (ILWMT) and the modem programmes regarding their achievements and drawbacks was not documented. The current status of the water erosion problems and future trends as well as alternative solutions are not satisfactorily addressed. The questions posed are: What went wrong and why? Are useful lessons drawn; which corrective measures were taken? V/hat is the future trend? This study takes the view that in order to deal with these questions,

Rainfall and its erosivity in Ethiopia 150 a critical review of the experiences is needed. The study enables to point out positive achievements gained through the use of the prevailing technologies and exploit their potential advantages for better future land and/or water management. Furthermore, it is an approach of this assessment that a critical review would facilitate the exploration of shortcomings of past water erosion mitigation programmes, projects, policies and strategies. This fosters the capitalisation on experiences and the attainment land and water resource management goals in a sustainable manner. The overall objective of this investigation is to conduct a post evaluation of land and water management activities. The specific aims are to:

1. Critically review experiences in soil and water conservation activities in the highlands of Ethiopia; 2. Compare and contrast traditional versus modem methods of land and water management. In this regard special attention was given to their sustainability, social acceptance, integration into agricultural practices and their possible future trends; 3. Assess land and water management policies as well as strategies and examine their failures and possible future directions.

This chapter focuses on the central highlands of Ethiopia, as described in the introductory section. The limited data were derived through intensive archive research and examination of previous studies, exploratory social survey, field investigations and interviews with authorities, farmers as well as experts. Familiarity with the area has also played a determining role to objectively investigate the facts.

7. 2. 2 Indigenous Land and Water Management Technologies - IL WMTs

Ethiopia has a wealth of experience in Land and Water Management Technologies (LWMT). In the face of growing population pressure and increasing demand for land, farmers in the central highlands of Ethiopia became increasingly reliant on ILWMT domestic to the environment. A series of these

Rainfall and its erosivity in Ethiopia 151

technologies is documented by several authors (e.g., A l e m a y e h u 1996;

K r Og e r et al. 1996; G e b r e -M ic h a e l 1998). ILWMTs are farming practices that have evolved over long time, without any known external institutional intervention (K r u g e r et al. 1996). Various engineering, biological and agronomic techniques which protect the soil against the erosive impact of rainfall as well as runoff are encompassed by this term. From personal experience, the central highlands of Ethiopia are one of the country’s regions where various ILWMTs are intensively practised. This is a reflection of high pressure on land for agricultural use and the consequence of severe ecological degradation in the area. Generally, biological, physical as well as integrated land and water management measures are widespread in most of the areas. It was noted that biological techniques are used where there is plant growth potential, physical methods where this is limited and on steep slopes. These measures are combined where topographic and soil moisture conditions permit plant growth, especially cultivated plants.

Today, farmers prefer the ILWMTs, since they have been part of agricultural practices for several years. Moreover, the techniques compete less for cultivated land and are easy to repair and maintain, as compared with the modem measures. Their disadvantage is that they are less effective where the pressure on land for agricultural use is high. Integrated land and water management practices are the most efficient ILWMTs. They enable to diversify the agricultural activities, including livestock production. The combination of various Indigenous Technologies significantly reduce the risk of soil loss in case of heavy rainfall. Repair and maintenance of the measures is combined with the routine farming operations. Figs. 55 and 56 illustrate typical traditional technologies commonly used in the central highlands of Ethiopia.

Rainfall and its erosivity in Ethiopia 152

Fig. 55: Traditional hillside terraces for soil and water management in the central highlands of Ethiopia (North Shoa)

Hillside terraces (Fig. 55) have been experienced for a long time in the central highlands of Ethiopia. They have double purposes in that they are not only used for soil conservation but also for water diversion from the field in case of excessive runoff as well as retention in the field. These measures need frequent repair and maintenance.

Rainfall and its erosivity in Ethiopia 153

Fig. 56: Integrated traditional agricultural land-use practices for land and water management in the central highlands of Ethiopia (North Shoa)

Integrated land and water management practices (Fig. 56) are very efficient. Households using such practices have the ability to diversify their farming systems. The combination of the various traditional techniques ensures less risk of excessive soil erosion during heavy rainstorms.

7. 2. 3 Modern Land and Water Management Technologies - MLWMTs

In the face of overwhelming land degradation due to water erosion, experts and policy makers were convinced that ILWMTs alone were not sufficient to conserve the ecosystem. Hence, state intervention with the traditional techniques through MLWMTs was firmly justified. While the introduction of modem methods dates back to the 1960s, their application in the central highlands of Ethiopia reached its climax in the 1980s. The Ethiopian government, with bi- and multi-lateral international support, launched a colossal environmental protection programme in 1980s (FAO 1986). The central highlands of Ethiopia

Rainfall and its erosivity in Ethiopia 154 were one of the project target areas (see FAO 1986; EPA and MEDC 1997). The activities predominantly focused on mechanical land and water management technologies, occasionally accompanied by afforestation and reforestation. The MLWMTs can be grouped under five major categories: i. Benches: terracing and bunding ii. Water diversion or drainage channels iii. Water ways or courses iv. Check dams v. Tillage practices

Most of the mechanical land and water management structures were built on hillsides so that tree planting could be carried out concurrently. Tillage practices were exclusively performed by the land-users. Fig. 57 shows an example of modem land and water management measures predominantly applied in the study area within the framework of the huge conservation programmes.

Fig. 57: Conventional check dams for land and water management in the central highlands of Ethiopia (North Shoa)

Rainfall and its erosivity in Ethiopia 155

Check dams (Fig. 57) are multi-functional structures. They are used to trap sediments, divert excess runoff, slow down the velocity of channel flow and form a micro-basin for the growth of conservation plants.

Even though progress was made through the intervention programmes of the 1980s, their environmental sustainability and social acceptability is quite frequently questioned today (compare C am pbell 1991; Woien 1995; EPA and

MEDC 1997; A d m a ssie 1998). This emanates primarily from the nature of the introduced MLWMTs, wrong planning approaches of the programmes and their implementation policies or strategies. First and foremost, the measures were highly biased towards mechanical methods which involved breaking and moving of the earth, with little care for the natural setting of the soil system (FAO 1993). They encompass activities which are inharmonious with the very concept of land and water management. Even though they enable to reduce water erosion (sheet wash, rill, inter-rill and gully erosion) and sedimentation in the field through safe drainage of excessive surface runoff, they facilitate the detachment as well as transportation of soil particles as a result of less energy expenditure required for the process. Furthermore, the drained turbid water is disposed out of the agricultural land and directed to the streams, rivers and lakes; thus causing offsite damage through sedimentation and pollution of recipient areas. Water harvesting and recycling for agricultural use is neither thought about nor practised. Encouraged by the erected structures, not accompanied by conservation based agricultural development and extension services, land users were tempted to cultivate on steep slopes without appropriate management practices. This has not only enormous negative consequences for agricultural land-use systems, it also permanently damages the whole ecosystem. Land and water management experts, road engineers as well as villagisation planners have given no consideration to the impacts of soil excavation on water erosion, neither have they integrated safety measures into the construction works. This exposed the soil to fatal damage due to water erosion (H e r w eg 1992). The mechanical structures used against water erosion were not cared for once the construction works had been completed. Repairing and maintenance was given low level priority, if any at all. Subsequently, most

Rainfall and its erosivity in Ethiopia 156 of the MLWMTs collapsed. Today, deformed landscape and disturbed ecology are the heritage in most parts of the central highlands of Ethiopia. To make the situation worse, land degradation and environmental devastation due to water erosion continue at an accelerated rate, the fastest ever since the 1950s.

Large reforestation and afforestation programmes were triggered by the extreme ecological disaster caused by water erosion, accelerated by human pressure on the ecosystem. The target area covered huge part of the country which included the central highlands. Accordingly, various projects with land and water management goals, community forest development as well as pre-urban plantations were launched in the area. These efforts were extremely biased towards planting of Eucalyptus trees (Eucalyptus globulus (Labillardiere)), which practically proved inefficient in other countries to mitigate the impact of water erosion (Poore and Fries 1985 cited in Turnbull 1999; Tekle 1999). Controversial ecological impacts of Eucalyptus globulus which including soil water, plant nutrients and soil erosion are also reported (e.g., Pozo et al. 1998;

T urnbull 1999).

No doubt that progresses were recorded by reforestation as well as afforestation within the framework of land and water management. Nevertheless, their achievements were reported to be unsatisfactory, compared with the extent of the problems. It is documented that the rate of deforestation was higher than the rate o f reforestation and afforestation together (Omiti et al. 1999). The process of accelerated deforestation in the central highlands of Ethiopia has greatly exacerbated the impact of water erosion. Consequently, soils in many areas, especially in North Shoa. are irrecoverably degraded and have lost their ecological function so that even trees which could otherwise have grown on very degraded soils, e.g., E. globulus, failed to survive. In spite of the reforestation and afforestation programmes and efforts to improve land cover in the area, water erosion continues to degrade the land at accelerating rate. It is the major present and future environmental and agricultural challenge.

Rainfall and its erosivity in Ethiopia 157

7. 2. 4 Socio-economic context of water erosion in the central highlands of Ethiopia

The socio-economic welfare of Ethiopia depends on land resource base of the country. However, the human and ecological interaction in the country is considered to be destructive. Increasing human and livestock pressure are frequently blamed for land degradation which, in turn, is made responsible for low performance of the agricultural sector (W0IEN et al. 1995: H o ben 1997). In the central highlands of Ethiopia, land degradation due to water erosion is an obvious phenomenon which persistently impoverishes not only the soil resources but also land-holders' subsistence economies. Hence, the causes of poverty in the area are interwoven with and deeply rooted in the consequences of water erosion (see H urni 1993; A bate 1997; H oben 1997; S hiferaw and

Ho lden 1999; Sh ifer w a and Ho lden 2000). Implications of water erosion and activities to alleviate its impact on the socio-economy of the region are quite heterogeneous throughout the area. Individual as well as collective reactions towards the problems and combating measures vary from one specific area to the other. This mainly depends on value judgements and the extent of problem perception. However, common features of individual and collective reactions towards top-down oriented water erosion mitigation programmes and strategies can be discerned (e.g., STAHL 1990; CAMPBELL 1991; T ekle 1999). The authors noted that little attention was given to the attitude of the farmers towards conservation programmes and to their priorities regarding land and water management. The farmers were neither contacted nor included in design and implementation of the conservation programmes. As a result, both the community as well as the individuals did not feel much sense of responsibility to manage the established land and water conservation infrastructure.

Conflicts between socio-economic interests of land-users and land and water management programmes as well as strategies were ubiquitous. Whereas land- users' economic objectives were to maximise the efficient use of land through cultivation and grazing, the aims of most land and water management programmes were only to conserve soil and water. The modern intervention

Rainfall and its erosivity in Ethiopia 158 programmes did not give due consideration to the land-users' socio-economic decision criteria. Thus, competition for land between these seemingly harmonious objectives have probably led to the collapse of the conservation efforts in the area. S t Ahl (1990) reported that most of the conservation works have considerably reduced the available land for cultivation and grazing.

A d m a ssie (1998) indicated that hill-side closure in southern Wello led to conflict on grazing land and, finally, conservation closures were destroyed.

Land and water management programmes in the central highlands of Ethiopia have created confusion and ambiguity in property and use rights of land resources, especially forests. EPA and MEDC (1997) noted that impositions which increasingly and cumulatively eroded the rights of land-users to manage their own resources might have led to a significant negative environmental impact. Conservation structures were laid across several plots of land whose ownership was diverse, ranging from individual to collective ownership. However, property rights and responsibilities to maintain the structures were not clearly defined. Consequently, the land-users, without whose participation any achievement in environmental management is unthinkable, were confused with respect to property rights and their feeling of ownership of land resource diminished. This situation led to the alienation of resource beneficiaries from their own traditional property.

The land tenure system has aggravated the ambiguity in land-use and property rights. This led to the lack of individual and public responsibility to care for land and water management infrastructure. The FAO (1993) reported that uncertainty about land tenure was a disincentive to investments in soil conservation in the highlands of Ethiopia. Ez r a and K a ssah un (1988) cited in T ek le (1999) noted that the uncertainty of ownership and use rights of tree plantations contributed to low performance of the community forestry programme in Ethiopia. Negative consequences of colliding social and political interests in environmental conservation were also reported from other African countries (P ile 1996; L ad o

1998; W an itzek and S ippel 1998; E llis-Jones and T en g berg 2000). For exam ple, W a n itzek and S ippel (1998) noted that conflicting interests arising

Rainfall and its erosivity in Ethiopia 159 from uncertain land-use and land property right in Tanzania have resulted not only in low sustainability of environmental conservation programmes but also in local social conflicts. The authors suggested a community based conservation approach as a possible solution. Land development with clear land-use and property rights, whereby the public was given a sense of responsibility, proved to be effective in the Niger Republic (Norman 1998). A broad based conservation approach, which takes the role of the local community in land and water management as well as agricultural development decision making into account, was suggested for the Ethiopian highlands (FAO 1993). However, due to limitations of financial resources needed for implementation, their realisation and achievements are far below expectations, especially in the highlands of Ethiopia.

7. 2. 5 Land and water management policies and strategies o f Ethiopia

The 1984/85 drought disaster triggered high environmental awareness, state interventionism with environmental management and political eco-fanatism in

Ethiopia (see H o b en 1995; A dm a ssie 1998). This conditioned the development of land and water management policies and strategies in the country which, thenceforth, stood high on the state natural resource management agenda. A national policy framework for land and water management within the context of environmental conservation was issued in 1987 (see St Ahl and W oo d 1989; IUCN 1993). The policy, empowering the government, spelt out that: ’’the state shall ensure that ecological balance is maintained through the conservation and development of natural resources; namely: land, water, forest and wildlife, and their utilisation for the benefit of the working people” (IUCN 1993).

Several institutional arrangements were made to implement the natural resource conservation policy of Ethiopia. The Office of the National Committee for Central Planning (ONCCP) was the highest state authority which co-ordinated issues related to resource management through its Department of Natural Resources and Human Settlement (NRHS). Relevant bodies were the Ministry of Agriculture, Conservation and Environment (MoACE), the Ministry of Mines

Rainfall and its erosivity in Ethiopia 160 and Energy and the Ministry of Education. The Natural Resources Conservation and Development Main Department (NRCDMD) of the Ethiopian Ministry of Agriculture (MoA) was the major organ responsible for all conservation matters. Two departments of the MoA played an important role in land and water management. First, the Soil and Water Conservation Department (SWCD), primarily responsible for activities pertaining to soil and water management and second, the Forestry and Wildlife Conservation and Development Authority (FaWCDA), which was mainly concerned with afforestation and reforestation schemes. FaWCDA implemented land and water management activities through the State Forest Conservation and Development Department (SFC'DD) and the Community Forestry and Soil Conservation and Development Department (CFSCDD). The SFCDD was mainly responsible for state forestry and CFSCDD for on-farm and off-farm soil and water management and community forestry. The SWCD discharged its responsibilities through its four divisions at the headquarter and the Conservation Teams at regional offices of the MoA. The duty of the headquarter was to develop indicative work targets and finally distribute these to the implementing regional offices. The work sites were selected by district soil and water management specialists. Development Agents (DAs) and leaders of Peasant Associations (PAs). Feedback on performance and problems encountered was channelled to the headquarter through the same bureaucratic hierarchy (FAO 1986: 1UCN 1993).

In addition to the government offices, several non-governmental institutions were involved in the implementation of the land and water management policies. Among those were the Ethiopian Relief and Rehabilitation Commission (RRC), the Ethiopian Red Cross Society (ERCS), the Swedish International Development Authority (S1DA), the Swiss Government/University of Berne, the Deutsche Gesellschaft fur Technische Zusammenarbeit (GTZ), the United Nations Sahelian Office (UNSO), the Finnish International Development Authority (F1NN1DA), and the United Nations World Food Programme (WFP)

(see FAO 1986: S i Am. and W o o d 1989).

The Government constitution of the Federal Democratic Republic of Ethiopia

Rainfall and its erosivity in Ethiopia 161

(FDRE) and the national economic policy underscore environmental sustainability as a key prerequisite for lasting success in economic development (EPA and MEDC 1997). It is stated in the objectives of the Federal policy on the Environment that: “ The overall policy goal is to improve and enhance the health and quality of life of all Ethiopians to promote sustainable social economic development through sound management and use of natural, human-made and cultural resources and the environment as a whole so as to meet the needs of the present generation without compromising the ability of future generations to meet their own needs” (EPA and MEDC 1997). Public, especially grassroot participation in planning as well as implementation, inter- and intra-sectoral co­ ordination as well as collaboration, adoption, adaptation and dissemination of appropriate technology are important features of the current environmental policy. These are expected to directly or indirectly foster land and water management. The reduction of tenure uncertainty is an other issue gaining attention.

As stated in the Conservation Strategy of Ethiopia (CSE) (EPA and MEDC 1997), sectoral environmental policies which are concerned with land and water management are mainly dealt with in the Soil Husbandry and Sustainable Agriculture, Forest, Woodland and Tree Resources sub-sectors of the economy.

The following are subsectoral policy components which directly address strategies to tackle the ecological impacts of water erosion:

« Land cover and soil organic matter management; • Improvement of biological, physical as well as chemical properties of soil; ® Water resource development for land care and bio-mass production; ® Reforestation and afforestation.

These are expected to enhance land and water management for sustainable socio-economic development of the country.

Cross-sectoral policies which implicitly or explicitly contribute to the mitigation

Rainfall and its erosivity in Ethiopia 162 of negative environmental consequences of water erosion are mainly considered in the Population and the Environment as well as the Community Participation and the Environment sub-sectors. These particularly aim at maintaining the environmental human carrying capacity, especially land; promoting public environmental awareness and fostering community participation as well as tenure security. It is expected that these policies and strategies will facilitate the implementation as well as the sustainability of land and water management measures proposed by the Soil husbandry and Sustainable Agriculture sub­ sector. However, EPA (1999) noted that the soil and water conservation draft policy, although it recognises community participation, appears to be too prescriptive in that it extremely relies on legal measures to deter land mismanagement.

The current institutional arrangements and frameworks provide for grassroot participation and sharing of responsibility by the government and the public to conserve as well as protect land and water resources of the country. The EPA, established in 1995 by virtue of Proclamation 9 of 1995, is the highest federal organ concerned with activities in the field of environmental management. The EPA is accountable to the Council of Ministers of the FDRE (EPA 1999). Specifically, it co-ordinates environmental policies and strategies, action plans and legislation making (EPA 1999). The MEDC co-ordinates the planning, programming as well as consolidating of the overall investment projects for environmental management and development. The MoA, acting through its Soil Conservation and Land-use Technology and Regulatory Department (SCLTRD), and the Ministry of Water Resources (MoWR) are responsible federal organs for land and water management policy and regulatory activities of Ethiopia and the highlands. Regional State Agricultural and Water Bureaux are responsible for implementing land and water management programmes at local level. In addition, a number of community-based organisations are involved in land and water management. Several international organisations; for example, ICRAF, UNESCO, UNICEF, OAU, are directly or indirectly involved in supporting as well as promoting sustainable land and water management in the highlands of Ethiopia (EPA 1999).

Rainfall and its erosivity in Ethiopia 163

7. 2. 6 Retrospect and prospect

Ethiopia has a wealth of traditional know-how to manage natural resources. However, this became less and less efficient with increasing human population pressure on land for economic activities. To compensate for the inefficiency, modem programmes were launched. Nevertheless, due to top-down policy and strategy failures, past efforts of the modem intervention programmes were ineffective. Today, farmers are not only strongly adhering to the ILWMTs, but also have become pessimist about any intervention programme. Notable lessons drawn from past experiences are: Increased awareness of land degradation problems, their extent and consequences for the environment as well as agricultural development, especially with regard to food production, poverty alleviation and economic development.

The current policies stress sustainable resource management through public participation. However, clear action plans at local level and transparent property rights are lacking. Land degradation problems are accentuated due to the increasing demand for resources because of population growth and degradation- promoting land exploitation. Hence, it is highly unlikely that any mitigation effort will solve the problems in the foreseeable future. Consequently, land degradation will remain a major environmental threat to Ethiopia. Therefore, a fundamentally new environmental thinking which aims at breaking the vicious circle of land degradation driven low agricultural production and poverty should be introduced and implemented.

It is recommended that the ILWMTs be promoted and improved. To ensure effective implementation, frequent follow-up and continuous evaluation of intervention measures should be made for immediate rectification of the failures. Clear property rights and responsibilities should be stated. Farmers should be encouraged to participate in defining the research problems and the findings should be tested on the farm in collaboration with the beneficiaries. Future research should focus on fostering the ILWMTs and adaptation of MLWMTs to the physical and socio-economic environment of the target people. Basic and

Rainfall and its erosivity in Ethiopia 164 applied research on water erosion potential and climate variability should be intensively carried out to develop prediction tools. Rainfall erosion should be assessed in detail and the findings integrated into design and development of conservation measures.

Rainfall and its erosivity in Ethiopia 165

8 SUMMARY AND CONCLUSIONS, ZUSAMMENFASSUNG UND AUSBLICK, RESUME ET CONCLUSIONS

8. 1 Summary

The subject of this study is rainfall and its erosivity in Ethiopia with high emphasis on the central highlands of the country. Land degradation caused by water erosion is the major ecological problem in Ethiopia. Particularly the central highlands of the country are extremely affected. Water erosion has already adversely affected large areas of the region to the extent that a huge proportion of agricultural land is irrecoverably degraded (see section 1.1). However, rainfall erosivity, especially in relation with long-term precipitation and its variability, is not yet documented in the country; this holds true for the central highlands. This study investigated rainfall as well as its erosivity, and assessed their spatial and temporal variability in Ethiopia with a special emphasis on the central highlands. Furthermore, it examined the influence of the variability in precipitation on flow regime and finally summarised the Ethiopian experience with land and water management. The principal aims were to develop a land and water resources management database as well as to establish rainfall erosion prediction tools applicable to the Ethiopian environmental conditions (consult section 1.2).

Long-term precipitation data (1898-1997) of 168 carefully chosen weather stations were statistically analysed. Rainfall erosivity was estimated using the widely accepted and applied Modified Fournier’s climatic erosivity index (Fm), as recommended by the FAO for countries where data are scarce. Forty-four stations were selected to examine rainfall erosivity in the central highlands of which finally eleven representative stations with 30 years or more record length were chosen for detailed time series analysis. Descriptive and inferential statistical analyses were conducted on precipitation and its long-term erosivity. The characteristics of precipitation erosion potential in time and space was thoroughly assessed. Regression models to predict rainfall erosivity at regional as well as specific weather station level were carefully developed. Annual and

Rainfall and its erosivity in Ethiopia 166 summer iso-erodent contours were drawn (see section 2). The relationship between rainfall variability and flow regime was examined based on hydrometric records of five gauging stations with years of data records ranging from 1982 to 1997 (refer to section 7.1). Documents of terminated as well as ongoing projects were meticulously examined and social survey was conducted to review the Ethiopian experience in land and water management (see section 2).

Generally, the summer period is the major rainy season in Ethiopia, especially in the central highlands. Two types of rainfall distribution patterns can be noted. First, one with a single maximum and, second, another with two peaks. The latter was subdivided into two groups, namely one with a small and another with a large gap between the two peaks. The seasonal variability gradient, as confirmed by the coefficient of variation, is highest in June and lowest in April. Generally, the variability decreases with increasing precipitation amount. This phenomenon is reflected on rainfall erosivity, resulting in a variability of up to 130% (see section 3). Rainfall erosivity estimated for the FAO climate pattern regions which range from A to E indicated that precipitation is most erosive region A and least region D. The prediction regression models developed for the pattern regions can be used to estimate rainfall erosivity at regional scale. Particularlly they are readily applicable to zonal land and water management planning as well as resources development policy design (refer to section 3.3).

In the central highlands of Ethiopia, a dramatic fluctuation in long-term precipitation involving extremely wet and drought events was noted. The area predominantly experienced positive departures of annual and summer rainfall from the respective long-term averages in the first half and negative deviations since the second half of the 20!h century. However, a declining tendency in rainfall during the whole period was observed (consult section 4). It was confirmed that highly erosive rainfall events have occurred more frequently in the first half of the 20th century than in the succeeding period. Even though precipitation has shown a general declining tendency, the effect of water erosion has accentuated since the 1950s due to the exploitation of land resources, which

Rainfall and its erosivity in Ethiopia 167 strongly promoted degradation. A linear regression model was found to be the best predictor of regional rainfall erosivity. Long-term annual and summer erosivity of precipitation in the region tend to increase from east to west and from south to north. This is a reflection of the spatial precipitation pattern in the area which is, in turn, governed by the topographic characteristics and atmospheric circulation. Since local land and water management practices based on regional models are inexact, a detailed analysis was conducted for the selected weather stations in the study area. It was found out that the erosivity of rainfall varies considerably at different locations. Nevertheless, it could be generalised that summer precipitation is the major causing agent of soil degradation, as compared to the precipitation in the other seasons. Predictive models of water erosion were meticulously developed for each of the selected areas. These are readily applicable to land and water management practices at the respective locality. Regarding temporal variability, a decline in rainfall erosion potential was noted for some, whereas an increasing trend was observed for the other weather stations. However, the decline was found to be statistically non-significant. Departures from the long-term mean vary seasonally as well as spatially. This justifies the viewpoint that local specific land and water management practices should include the adaptation of techniques appropriate for the respective area as well as season (consult section 6.4).

It was confirmed that the variability in rainfall erosivity is not significantly related to altitude. The variability shows a decreasing tendency towards the north and increasing one towards the east. However, a multiple regression analysis indicated that this relationship is not statistically significant. Hence, further investigation is required to be able to define these yet unexplained determining variables (see section 6.4.2).

Six standard classes of water erosion risk were established for the study area. These are: very low, low, moderate, high, very high and extremely high. The dominant class for annual and summer rainfall erosivity was found to be the High category. Succeeding this, individual weather stations were clustered to the predetermined classes. Accordingly, most of the weather stations examined fell

Rainfall and its erosivity in Ethiopia 168 into the Low class when annual rainfall erosivity was considered. This is due to the cancelling effect of dry seasons. Regarding summer rainfall erosivity, the majority of the weather stations assessed fell into the high class. This supplements the fact that most of the erosive precipitation events occur in summer (refer to section 6.5).

The decline in areal rainfall was reflected on flow regime at the gauging stations that were assessed. However, the effect varied at different measuring points. The probable causes of this phenomenon are topographic variations, land-use, and cover of the respective areas. A comprehensive geo-ecological investigation with a basin or a watershed approach is required to determine which of these factors play the major role and how this affects soil erosion as well as sedimentation (consult section 7.1.2).

The Ethiopian farmers have rich experience in Indigenous Land and Water Management Technologies, namely engineering, biological and integrated soil and water conservation techniques (refer to section 7.2.2). Nonetheless, these turned out to be less effective with the accelerating land degradation. The compensatory intervention actions taken by the Ethiopian Government were found to be neither sustainable nor socially acceptable. The reasons are inappropriate project planning and top-down implementation policy approach (see section 7.2.3). Subsequently, there is a tendency that public participation and sustainability are gaining more attention in policy design (consult section 7.2.5).

8. 2 Conclusions

Data limitation is the major bottleneck to scientific investigation of climate variability and water erosion in Ethiopia (see section 2.2.2). In order to alleviate the problem, there is an immediate need to improve the database as well as its management. To ensure the accuracy of the records, obsolete climate measuring as well as hydrometric equipment should be renewed or replaced with modem recording apparatus. This facilitates the introduction, verification and

Rainfall and its erosivity in Ethiopia 169 application of modem water erosion prediction technologies, which would also enhance appropriate land and water management.

The data set o f rainfall and its erosivity established by this study is a useful resource for further climate, land and water management research. It is recommended that this valuable information be continuously enriched as well as regularly updated (see annex).

The models developed by this study are applicable to forecast water erosion in the study area under conditions of data scarcity and financial constraints. Moreover, they are useful for land and water resource management planning and policy making. However, they need to be tested, verified and calibrated at field conditions for precise estimation of land degradation (refer to sections 3.4, 5, 6 and annex).

Future research should focus on testing and adaptation of modem water erosion prediction technologies as well as models to find out the best prediction tools with maximum efficiency. This should go hand in hand with research aiming at developing and adapting various water erosion mitigation techniques as well as agricultural development. Watershed and river basin approach to rainfall erosion and sedimentation assessment as well as promotion of indigenous knowledge of environmental conservation should gain much focus in land and water resources management research in Ethiopia, and particularly in the central highlands of the country.

Rainfall and its erosivity in Ethiopia 170

8.3 Zusammenfassung

Die vorliegende Untersuchung befaBt sich mit Regen und dessen Erosivitat in Athiopien; Schwerpunkt der Betrachtung ist das zentrale Hochland. Bodendegradation durch Wassererosion ist das groBte okologische Problem in Athiopien. Insbesondere das zentrale Hochland ist davon stark betroffen. Bis heute hat die Wassererosion derart groBe Regionen zerstort, daB ein iiberwiegender Teil des Acker- und Weidelandes unwiederbringlich geschadigt ist (vgl. Kap. 1.1). Trotzdem ist die Regenerosivitat insbesondere im Zusammenhang mit langjahriger Niederschlagsvariabilitat bis heute nicht eingehend untersucht worden; gleiches gilt fur das zentrale Hochland des Landes. Die vorliegende Arbeit beschaftigt sich mit Regen und seiner Erosivitat und analysiert deren raumliche und zeitliche Variability in Athiopien. Das zentrale Hochland bildet einen besonderen Schwerpunkt der Betrachtung. Dariiber hinaus wird der EinfluB der Flachenniederschlagsvariabilitat auf das AbfluGregime untersucht und abschlieBend mit athiopischen Ansatzen von Land- und Wassermanagement in Verbindung gebracht. Vorrangiges Ziel war es, eine Datenbasis fur das Land- und Wassermanagement zu entwickeln und Planungsinstrumente zum Schutz vor Regenerosivitat fur Athiopien bereitzustellen, die den athiopischen Umweltbedingungen angepaBt sind (vgl. Kap. 1.2).

Langjahrige Niederschlagsdaten (1898-1997) von 168 ausgewahlten Wetterstationen wurden statistisch untersucht. Die Regenerosivitat wurde an Hand des international anerkannten und angewendeten modifizierten Fournier Index (Fm) bestimmt, entsprechend dem Vorgehen der FAO fur Lander, in denen Datenmangel es nicht zulaBt, die ublicherweise angewandten Methoden wie z. B. die USLE und ABAG einzusetzen. Vierundvierzig representative Wetterstationen wurden fur die Untersuchung der Regenerosivitat im zentralen Hochland eingesetzt, davon kamen elf mit mindestens 30-jahrigen Datenreihen fur eine detaillierte Zeitreihenanalyse des Niederschlags und seiner Erosivitat zum Einsatz. Beschreibende und analysierende statistische Auswertungen wurden fur den Niederschlag und seine langjahrige Erosivitat durchgeflihrt. So

Rainfall and its erosivity in Ethiopia 171 ist das zeitliche und raumliche Verhalten der potentieilen Niederschiagserosion naher untersucht worden. Zudem wurden Regressionsmodelle entwickelt, die regionale und lokale Aussagen tiber das potentielle Wassererosionsrisiko ermoglichen konnen. Aus den regionalen Modellen wurden jahrliche und sommerliche Isoerodentkarten entwickelt (vgl. Kap. 2). Die Beziehung zwischen Niederschlagsvariabilitat und AbfluGregime wurde an Hand der hydrometrischen Daten von 5 MeBstationen im zentralen Hochland mit Aufzeichnungen zwischen 1982 und 1997 untersucht (vgl. Kap. 7.1). Dariiber hinaus wurden Dokumente vergangener, laufender und geplanter Projekte zusammengetragen und ausgewertet, um den Stand und die Perspektiven der Bodenerosion und athiopischer Ansatze im Land- und Wassermanagement zu bewerten (vgl. Kap. 2).

Generell ist das Sommerhalbjahr die Hauptregensaison in Athiopien, insbesondere im zentralen Hochland. Es konnten zwei Typen von Niederschlagsverteilungen festgestellt werden: mit einem Maximum und mit zwei Maxima. Der zweite Typ konnte wiederum differenziert werden: in eine Verteilung mit einer kurzen und eine mit einer langen Phase zwischen den beiden Hochstwerten. Der saisonale Variabilitatsgradient, der sich aus dem Variationskoeffizienten ergibt, ist im Juni am hochsten und im April am niedrigsten. Generell nimmt die Variabilitat mit zunehmender Regenmenge ab. Dieses Phanomen wirkt sich deutlich auf die Regenerosivitat aus und resultiert in einer Variabilitat von bis zu 130% (vgl. Kap. 3). Ubertragt man die Regenerosivitat auf die FAO-Klimaregionen, welche in Gruppen von A bis E untergliedert sind, so laBt sich feststellen, daB der Regen in Region A am erosivsten ist und am wenigsten erosiv in Region D. Regressionsmodelle, die fur die Regenerosion entwickelt wurden, konnen zur regionalen Vorhersage des Regenerosivitatspotentials verwendet werden. Insbesondere konnen sie der Planung des zonalen Land- und Wasserressorcenmanagements und dem Entwurf der entsprechenden politischen MaBnahmen dienen (vgl. Kap. 3.3).

Wie es von Daten des Flachenniederschlags bestimmt wurde, sind im zentralen Hochland Athiopiens dramatische Schwankungen der langjahrigen

Rainfall and its erosivity in Ethiopia 172

Niederschlage festzustellen, die sich in extremen Trockenzeiten aber auch in Oberschwemmungen auBem. In der ersten Halfte des 20. Jahrhunderts wies die Region iiberwiegend positive Abweichungen des Jahres- und Sommemiederschlages vom langjahrigen Mittel auf. In der zweiten Halfte iiberwogen negative Abweichungen. Insgesamt kann eine abnehmende Tendenz der Regensumme liber den gesamten Untersuchungszeitraum festgestellt werden (vgl. Kap. 4). Es wurde belegt, daB die starksten erosiven Ereignisse haufiger in der ersten Halfte des 20. Jahrhunderts stattfanden als in der Zeit danach. Trotz der abnehmenden Tendenz der Niederschlage hat sich ihre erosive Wirkung seit 1950 zugespitzt, bedingt durch die extreme Nutzung der Landressourcen, welche die Bodendegradation zusatzlich beschleunigt hat. Es wurde festgestellt, daB ein lineares Regressionsmodell das beste Prognoseinstrument der regionalen Regenerosivitat im Untersuchungsraum ist. Raumlich gesehen, nimmt die langjahrige Jahres- und Sommemiederschlagserosivitat im Untersuchungsgebiet von Ost nach West und von Siid nach Nord scheinbar zu. Diese Tendenz hangt eng mit der raumlichen Verteilung der Niederschlage zusammen, die wiederum durch topographische Bedingungen und atmospharische Zirkulation beeinfluBt sind. Da lokale Land- und Wassermanagement-Praktiken, die auf regionalen Modellen basieren, nicht exakt sein konnen, wurde eine detaillierte Analyse fur ausgewahlte Wetterstationen des Untersuchungsgebiets vorgenommen. Es wurde festgestellt, daB die Regenerosivitat an den verschiedenen Stationen betrachtlich schwankt. Trotzdem konnte festgestellt werden, daB im Vergleich mit den Niederschlagen in den anderen Jahreszeiten. die Sommemiederschlage die Hauptausloser fur die Bodenerosion sind. Der Vollstandigkeit halber wurden Prognosemodelle fur die Wassererosion fur jeden einzelnen der Untersuchungsraume erarbeitet. Diese ermoglichen die Planung von Land- und Wassermanagement am jeweiligen Ort. Bei der Untersuchung der zeitlichen Variabilitat des Regenerosionspotentials konnte fur einige Stationen ein abnehmender, fur andere ein ansteigender Trend verzeichnet worden. Dennoch ist die absteigende Tendenz statistisch nicht signifikant. Abweichungen vom langjahrigen Mittel variieren sowohl zeitlich als^auch raumlich. Dies belegt die Hypothese, daB das lokalspezifische^ Land- und Wassermanagement die Adaption der benutzten Techniken an die ortlichen und jahreszeitlichen

Rainfall and its erosivity in Ethiopia 173

Gegebenheiten einschlieBen soil (vgl. Kap. 6.4).

Es konnte festgestellt werden, daB die Beziehung zwischen der Variabilitat der Regenerosivitat und den Hohenstufen nicht signifikant ist. Die Variabilitat zeigt eine abnehmende Tendenz im Norden und eine ansteigende im Osten. Jedoch zeigte eine multiple Regressionsanalyse, daB diese Variabilitat nicht signifikant ist. Demzufolge sind weitere zusatzliche Untersuchungen notig, um die noch unbekannten bestimmenden Variablen feststellen zu konnen (vgl. Kap. 6.4.2).

Im Untersuchungsgebiet konnten sechs Gefahrdungsstufen des potentiellen Risikos von Regenerosivitat abgegrenzt werden: selir niedrig, niedrig, mittel, hoch, sehr hoch und extrem hoch. Die vorherrschende Klasse sowohl fur Sommer- als auch fur Jahresniederschlagserosivitat ist die hohe Kategorie. Als nachstes wurden die einzelnen untersuchten Wetterstationen durch Clusteranalyse den vorbestimmten Klassen zugeordnet. In Folge dessen fielen die meisten Untersuchungsstandorte in die niedrige Klasse, sofern die jahrliche Regenerosivitat betrachtet wurde. Dies ist auf den abschwachenden Effekt der Trockensaison bzw. Winter zuriickzuflihren. Bei der Analyse der sommerlichen Regenerosivitat ist ein iiberwiegender Teil der Stationen der hohen Klasse zuzuordnen. Dies unterstiitzt die Tatsache, daB die meisten erosiven Regenereignisse im Sommer stattfinden (vgl. Kap 6.5).

Die Abnahme des Gebietsniederschlags wirkt sich auch auf das AbfluBregime der untersuchten MeBstationen aus. Jedoch ist die Auswirkung an den verschiedenen MeBpunkten unterschiedlich groB. Mogliche Griinde fur dieses Phanomen sind topographische Unterschiede, sowie Landnutzung und Bodenbedeckung der jeweiligen Einzugsgebiete. Eine eingehende geookologische Untersuchung auf der Basis von Einzugsgebietsanalysen wurde empfohlen, um eine Gewichtung dieser Faktoren und ihren EinfluB auf Regenerosivitat und Sedimentation zu bestimmen (vgl. Kap. 7.1.2).

Die athiopischen Bauem haben viel Erfahrung mit einheimischen Land- und Wassermanagement-Praktiken insbesondere in Bezug auf mechanische,

Rainfall and its erosivity in Ethiopia 174 biologische und integrierte Wasser- und Bodenschutz-Techniken (vgl. Kap. 7.2.2). Diese Techniken sind jedoch bei steigender Degradation der Boden zunehmend ineffizient geworden. Die GegenmaBnahmen der athiopischen Regierung waren weder nachhaltig noch sozial akzeptabel. Griinde hierfur sind vor allem eine unangemessene Projektplanung und hierarchische Implementierungspolitik (vgl. Kap. 7.2.3). Die jetzige Tendenz ist, daB eine Beteiligung der Offentlichkeit bzw. der Betroffenen an Entwurf und Durchsetzung der Umweltpolitik an Wichtigkeit gewonnen hat (vgl. Kap. 7.2.5).

8.4 Ausblick

Die begrenzte Datenbasis ist der groBte EngpaB fur wissenschaftliche Untersuchungen der klimatischen Variabilitat bzw. der Regenerosion in Athiopien (vgl. Kap. 2.2.2). Um diesem Problem entgegen zu wirken, muB die Datenbasis und ihre Verwaltung verbessert werden. Um die Giite der Daten zu gewahrleisten, sollten vergleichende Messungen stattfinden; hierfur sollten die veralteten MeBgerate gegen moderne Ausstattungen ersetzt werden. Dies erleichtert die Einfuhrung, Verifikation und Anwendung der modemen Wassererosionsprognose-Technologien, die zudem ein geeignetes Land- und Wassermanagement unterstiitzen konnen.

Die Datenbasis fur die Analyse des Niederschlages bzw. seiner Erosivitat, die der vorliegenden Arbeit zugrunde liegt, stellt eine Grundlage fur kunftige Untersuchungen des Klimas sowie des Land- und Wassermanagements dar. Es bleibt festzuhalten, daB die verfugbaren Informationen standig erganzt und aktualisiert werden sollten (siehe Anhang).

Die in dieser Studie entwickelten Modelle sind zur schnellen Abschatzung bzw. Prognose des Wassererosionspotentials im Untersuchungsraum in Situationen mit Datenmangel und finanziellen Beschrankungen anwendbar. Dariiber hinaus sind sie fur die Planung des Land- und Wassermanagements sowie fur den Entwurf der Ressorcenmanagementpolitik nutzbar. Jedoch miissen sie getestet, verifiziert und an die Feldbedingungen angepaBt werden, um prazise

Rainfall and its erosivity in Ethiopia 175

Voraussagen fur die Landdegradation treffen zu konnen (siehe Kap. 3, 4, 5, 6 und Anhang).

Zukiinftige Untersuchungen sollten auf das Testen und Anpassen modemer Wasserosionsprognose-Techniken sowie Modelle an athiopische Bedingungen abzielen, um die besten Schatzinstrumente fur Wassererosion und Landdegradation mit maximaler Nutzleistung zu erhalten. Dies sollte mit begleitenden Forschungen verkniipft werden, deren Ziele die Entwicklung und Anpassung von verschiedenen Wassererosionsminderungs-Techniken sowie die landwirtschaftliche Entwicklung sind. Einzugs- und FluBbeckenansatz zur Regenerosion- bzw. Sedimentationsforschung ebenso wie die Forderung des einheimischen Wissens zum Umweltschutz sollten den Vorrang in Land- und Wassermanagementstudien in Athiopien erhalten. Eine besondere Betonung sollte hierbei auf dem Hochland liegen, da dort Umweltdegradation am starksten ist.

8.5 Resume

Cette etude a pour objet la pluie et l'erosion qui en resulte en Ethiopie et se concentre en particulier sur la region centrale des hauts plateaux du pays. La degradation du sol engendree par l'erosion pluviale constitue le probleme majeur en matiere d'ecologie en Ethiopie. C'est en effet la region centrale des haut plateaux qui s'en trouve la plus affectee. L'erosion a deja fortement touche de vastes parties de la region a tel point qu'une partie importante de la terre cultivable est devenue definitivement inexploitable (voir Chap. 1.1). Toutefois, on ne dispose, en Ethiopie, d'aucune documentation traitant l'erosion causee par les pluies, notamment en ce qui conceme les effets a long terme et leurs variations; ceci est valable pour la partie centrale des hauts plateaux du centre. Ce travail s'est done propose d'une part d’etudier la pluviosite et l'erosion qui en resulte et de l’autre d’evaluer les variations aussi bien dans l'espace que dans le temps en Ethiopie, principalement pour la region mentionnee ci-dessus. En plus, ce travail a examine 1'influence des variations de la pluviosite sur le regime fluvial. Pour Finir, ce travail tire les conclusions concernant l'experience

Rainfall and its erosivity in Ethiopia 176 ethiopienne acquise en matiere de gestion de l'eau et des sols. Les objectifs principaux etaient de mettre au point une base de donnees relative a la gestion des ressources en eaux et sols, ainsi que de mettre a disposition des outils de prediction de 1' erosion adaptes aux conditions environnementales ethiopiennes. (Chap. 1.2)

Les donnees pluviometriques s'appuyant sur le long terme (1898-1997) foumies par 168 stations meteorologiques soigneusement selectionnees ont fait l'objet d'une analyse statistique.

L'erosion pluviale a ete determinee avec l'aide de l'index d'erosivite climatique modifie de Fournier (Fm) qui est largement accepte et applique et recommande par la FAO pour les pays dont les donnees sont peu abondantes. 44 stations meteorologiques furent selectionnees afin d'etudier l'erosion due aux precipitations dans les hauts plateaux du centre parmi lesquelles 11 stations representatives possedant des donnees vieilles de 30 ans et plus furent retenues pour effectuer des analyses detailiees dans le temps. Des analyses statistiques descriptives et deductives furent conduites sur les precipitations et leur caractere erosif a long terme. Les caracteristiques du potentiel erosif des precipitations au niveau temporel et spatial ont ete etudies en detail. Des modeles de regression permettant de prevoir l'erosion pluviale au niveau regional aussi bien qu’au niveau d'une station meteorologique specifique ont ete soigneusement mis au point. Des courbes annuel les et estivales iso-erosives furent tracees (voir Chap. 2). La relation entre les variations des precipitations et le regime fluvial a ete etudie a l'aide des donnees hydrometriques foumies par cinq stations de mesures fluviales dont les donnees annuelles ont ete enregistre entre 1982 et 1997 (voir Chap. 7.1). Des documents presentant des projets termine et en cours d'execution ont ete examines avec soin et une enquete sociale a ete menee afin d'evaluer l'experience ethiopienne en matiere d'amenagement des eaux et des sols (voir Chap. 2).

La periode estivale est generalement la principale saison pluvieuse en Ethiopie, tout particulierement dans les hauts plateaux du centre. Deux types de repartition

Rainfall and its erosivity in Ethiopia 177 des pluies peuvent etre distingues. Le premier se caracterisant par un seul et le second par deux points maximales Ce dernier se laisse en outre subdiviser en deux categories qui se distinguent par la duree de la periode separant les deux points maximales. Le quotient de variation saisonniere, egalement confirme par le coefficient de variation, est le plus eleve en juin et le plus bas en avril. En general, la variation decroit dans la mesure oil les precipitations augmentent. Ce phenomene se reflete dans l'erosion pluviale, qui a comme effet une variation de plus de 130% ( voir Chap.3 ). Dans les regions classees selon la FAO et situees sur une echelle allant de A a E. l'erosion due aux precipitations est la plus forte en A et la moins forte D. Les modeles de regression des predictions developpes pour les regions modeles peuvent etre utilises pour estimer l'erosion dues aux pluies a l’echelle regionale. Ils sont prets a etre appliques au projet d'amenagement des eaux et des sols de la zone ainsi qu'en politique d’exploitation des ressources (voir Chap. 3.3).

Dans les hauts plateaux du centre ethiopien, on a note une fluctuation dramatique des precipitations a long terme, provoquant des periodes de tres grande humidite suivie de periodes de secheresse. La zone a surtout connu des ecarts positifs par rapport a la pluviosite annuelle et estivale moyenne de la premiere moitie du 20e siecle et des ecarts negatifs par rapport a la seconde moitie du 29eme siecle (voir Chap. 4). On note cependant une tendance a la baisse des precipitations sur toute la periode observee. II a ete confirme que les precipitations tres erosives sont apparues plus frequeinment pendant la premiere moitie du 20e siecle que dans la seconde moitie. Bien que les precipitations ont montre une tendance generale au declin, l'effet d'erosion s'est accentue depuis les annees 50 en raison de 1'exploitation des terres qui a aggrave fortement la degradation des sols. Un modele de regression lineaire s'est avere etre le ineilleur indicateur de prevision de l’erosion due aux pluies regionales. L'erosion estivale et annuelle a long-terme due aux precipitations relevee dans la region tend a augmenter d'est en ouest et du sud au nord. Ceci est le reflet de la repartition des precipitations dans la region, elle-meme tributaire de la topographies et de la circulation atmospherique. Comme les pratiques de gestion locale des eaux et des sols s'appuient sur des modeles regionaux inexactes, une

Rainfall and its erosivity in Ethiopia 178 analyse detaillee a ete menee par les stations meteorologiques selectionnees dans la zone d'etude. Cette analyse a revele que l'erosion due aux pluies varie considerablement selon les lieux. On peut cependant affirmer en generalisant que les precipitations estivales sont, par comparaison aux precipitations pendant d'autres saisons, l'agent principal qui provoque la degradation des sols. Des modeles de prevision concemant l'erosion par les eaux ont ete soigneusement mis au point pour chacune des regions selectionnees. Ils sont prets a l'emploi pour la gestion des eaux et des terres dans les localites respectives. Concemant les variations dans le temps, une diminution du potentiel de l'erosion pluviale a pu etre constatee pour certaines stations, tandis qu'il y a une augmentation pour d'autres stations meteorologiques. Cependant, la diminution est insignifiante du point de vue de la statistique. Des ecarts de la moyenne a long terme varient selon les saisons et selon les regions. Ceci justifie l'idee que des pratiques locales specifiques concemant la gestion des eaux devraient inclure l'adaptation de techniques valables aussi bien pour la region respective que pour la saison en question (voir Chap. 6.4). II s'est avere que les variations de la pluviosite ne sont pas reliees de maniere significative a l'altitude. Les variations tendent a decroitre en direction du nord et a augmenter en direction de Test. Cependant, une analyse multiple de regression a montre que cette relation n'est pas significative du point de vue de la statistique. II est done necessaire de mener une enquete plus poussee pour etre capable de definir ces variables inexpliquees jusqu'alors (voir Chap. 6.4.2).

Pour la zone d'etudes, six classes de risques concemant l'erosion par les eaux ont ete etablies. Le risque est defini comme tres bas, bas, modere, eleve, tres eleve et excessivement eleve. La classe predominante concemant la pluviosite annuelle et estivale se trouve etre celle au risque eleve. A la suite de cela, des stations meteorologiques ont ete reliees aux classes predetermines. La plupart des stations etudiees se sont retrouvees dans la classe " bas " quand on a considere l'erosion annuelle. Ceci est du a l'effet d'annulation des saison seches. En considerant l'erosion estivale, la plupart des stations meteorologiques considerees se sont retrouvees dans la classe "eleve". Ceci confirme le fait que la plupart des precipitations erosives tombent pendant l'ete (voir Chap. 6.5).

Rainfall and its erosivity in Ethiopia 179

La diminution de la pluviosite regionale a ete refletee par le regime fluvial des stations de mesures fluviales examinees. Cependant, 1'effet etait different selon les differentes stations de mesures. Les causes probables de ce phenomene sont des variations topographiques, 1'exploitation des terres et la couverture des differents cinq regions. Une etude geophysique globale portant sur un bassin ou un point de repartition des eaux est indispensable pour determiner lequel de ces facteurs joue le role majeur et pour savoir comment cela joue sur l'erosion des sols aussi bien que sur la sedimentation (voir Chap. 7.1.2).

Les agriculteurs ethiopiens ont une experience riche en matiere de technologies de gestion de la terre indigene et des eaux, a savoir le genie (engineering), les techniques biologiques et de sol integre ainsi que de la conservation des eaux (voir Chap.7.2.2). Cependant, ces techniques se sont averees etre moins efficaces face a la degradation acceleree des sols. Les mesures d'intervention compensatoire prises par le gouvernement ethiopien se trouvent etre ni durables ni socialement acceptables. Les raisons en sont un projet inapproprie et une approche politique cle mise en place et arbitree par les dirigeants. En consequence, une tendance montre que la participation du public et le caractere durable attirent davantage Pattention dans le concept de la politique (voir Chap. 7.2.5)

8. 6 Conclusions

Le caractere limite des donnees est Pobstacle majeur de P etude scientifique sur la variation climatique et Perosion en Ethiopie(voir Chap.2.2.2) Afin de resoudre le probleme, il est urgent d‘ameliorer les bases de donnees en disposition ,ainsi que de leur gestion. Afin d‘assurer la precision des donnees, Pequipement de mesures climatiques obsolete ainsi que Pequipement hydrometrique devront etre renouveles ou remplaces par un appareil d'enregistrement modeme. Ceci facilitera Pintroduction, la verification et Papplication des technologies modemes de prevision de Perosion due aux pluies, ce qui devrait egalement encourager a une gestion des eaux et des sols appropriee.

Rainfall and its erosivity in Ethiopia 180

L‘ensemble des donnees concemant les precipitations et leur caractere erosif etabli par cette etude est une ressource utile pour d‘autres climats, pour la recherche sur la gestion des eaux et des sols. II est recommande que ces informations precieuses soient continuellement enrichies et mises a jour.

Les modeles mis au point au cours de cette etude sont applicables afin de prevoir 1‘erosion dans la zone d‘etude dans des conditions de contraintes de donnees insuffisantes et de contraintes financieres. De plus, ils sont utiles a la mise en place d‘une politique et d‘un programme d‘amenagement des ressources en eau et des ressources terrestres. Les modeles ont besoin d‘etre soumis a des tests, a des verifications et a un calibrage selon des conditions de zones pour une estimation precise de la degradation des sols (voir Chap. 3, 4, 5, 6 et Annexe).

Les recherches devront a Pavenir se concentrer sur la mise en place de tests et Padaptation des technologies modemes de prevision de Perosion due aux precipitations ainsi que des modeles afin de definir les meilleures outils avec un maximum d‘efficacite. Ceci devrait aller de pair avec le but de la recherche: developper et adapter differentes techniques afin de reduire Perosion ainsi que le developpement agricole. Les cascades et les bassins des rivieres vus sous Pangle de Perosion due aux precipitations et Petude de la sedimentation devraient davantage encourager la recherche dans la gestion de Peau et du sol ethiopien.

Rainfall and its erosivity in Ethiopia 181

9 REFERENCES

Abate, K. (1997): Poverty, Desertification and the Impact of Drought in Ethiopia. Desert. Cont. Bull. 31, 60-66.

A b e g a z , G. (1989): Total annual precipitation as a predictor of the annual erosion index. Eth. J. Agric. Sci. 11, 71-83.

A d m a s s ie , Y. (1998): Twenty Years to Nowhere: Property Rights, Land Management and Conservation in Ethiopia. Diss., Red Sea Press, Lawrenceville, New Jersey, 537 pp.

A l e m a y e h u , M. (1996): Traditional Ditches in Northern Shewa, The Ethiopian Highlands. In: Reu, C., Scoones, I. and C. Toulmin (eds.), Sustaining the Soil: Indigenous Soil and Water Conservation in Africa. International Institute for Environment and Development, London, 163-169.

A n s c o m b e , F. J. (1973): Graphs in statistical analysis. A m . Stat. 27, 17-21.

A r n o l d u s , H. M. J. (1977): Methodology used to determine the maximum potential average annual soil loss due to sheet and rill erosion in Morocco. FAO Soils Bull. 34, 39-51.

A r n o l d u s , II. M. J. (1980): An approximation of the rainfall factor in the

USLE. In: D e B o o d t , M. and D . G a b r i e l s (eds.), Assessment of Erosion, John Wiley and Sons Ltd., Chichester, 127-132.

Asrat, K., Idris, K. and M. Semegn (1996): The ‘flexibility’ of indigenous SWC techniques: A case study of the Hararge Highlands, Ethiopia. In: Reu, C., Scoones, I. and C. Toulmin (eds.), Sustaining the Soil: Indigenous Soil and Water Conservation in Africa. International Institute for Environment and Development, London, 156-162.

Rainfall and its erosivity in Ethiopia 182

A sta t k e , A. and F. Kele m u (1993): Modifying the traditional plough - Merasha

for better management of vertisols. In: M a m o , T., A st a t k e , A., S r iv a st a v a ,

K. and A. D ibabe (eds.), Improved Management of Vertisols for Sustainable Crop-Livestock Production in the Ethiopian Highlands, International Livestock Centre for Africa, Addis Ababa, 85-100.

B a h r en ber g , G., G ie se , E. and J. N ipper (1992): Statistische Methoden in der Geographie. Bd. 2., Multivariate Statistik, 2. Aufl., B. G. Teubner, Stuttgart, 415 pp.

Ba h r en b er g , G., G ie se , E. and J. N ipper (1999): Statistische Methoden in der Geography. Bd. 1., Univariate und Bivariate Statistik, 4. iiberab. Aufl., B. G. Teubner, Stuttgart, 234 pp.

B a r d , K. A., C o l t o r t i, M., D iB la si, M. C., D r a m is, F. and R. Fa ttov ich (2000): The Environmental History of Tigray (Northern Ethiopia) in the Middle and Late Holocene: A Preliminary Outline. Afr. Arch. Rev. 17, 65-86.

B ec k ed a h l , H. R. (1998): Subsurface Soil Erosion Phenomena in South Africa. Peterm. geogr. Mitt., Erg.-H. 290, 128 pp.

B ec k e r , P. and A. P. S m ith (1990): Spatial autocorrelation of solar radiation in a tropical moist forest understorey. Agric. For. Meteorol. 52, 373-379.

B en n ett, H. H. (1939): Soil conservation. McGraw Hill Book Co., New York, 993 pp.

B iseteg n e, D., O g a l l o , L. and J. Inin d a (1986): Rainfall characteristics in Ethiopia. Technical Conference on Meteorological Research in Eastern and Southern Africa. 1986, Nairobi, Kenya, 35-41 pp.

B lo ck , S. A. (1999): Agriculture and Economic Growth in Ethiopia: Growth Multipliers from a Four-sector Simulation Model. Agric. Econ. 20, 241-252.

Rainfall and its erosivity in Ethiopia 183

B o e r w in k e l, F. and S. Paris (1984): Methodology used in the development of soil loss rate map of the Ethiopian Highlands. MoA, U N D P and FAO, Addis Ababa, 19 pp.

BORK, H.-R. (1982): Die holozane Relief- und Bodenentwicklung im Untereichsfeld und im siidvvestlichen Harzvorland. Diss., Braunschweig, 187 pp.

B o r k , H.-R. (1987): Bodenerosion und Umwelt - Verlauf, Ursachen und Folgen der mittelalterlichen und neuzeitlichen Bodenerosion. Bodenerosionsprozesse, Modelle und Simulationen. Landschaftsgenese u. Landschaftsokologie 13, 249 pp.

B o s sh a r t, U. (1998): Catchment discharge and suspended sediment transport as indicators of the performance of physical soil and water conservation in the Ethiopian highlands. Advances in GeoEcology 31, 403-413.

B r e it e n b a c h , F. (1963): The indigenous trees of Ethiopia. Second, revised and enlarged edition. Ethiopian Forestry Association, Addis Ababa, 306 pp.

B r o c k w e ll, P. J. and R. A. Davis (1996): Time Series: Theory and Methods. Second edition. Springer-Verlag, New York, 577 pp.

Br o w n , L. H. (1973): Conservation for survival: Ethiopia’s choice. Haile Selassie I University, Addis Ababa, 96 pp.

B u ll, L. J., K ir k b y , M. J., S h a n n o n , J. and J. M. H o o k e (2000): The impact of rainstorms on floods in ephemeral channels in southeast Spain. Catena 38, 191-209.

B u r in g h , P. (1989): Availability of agricultural land for crop and livestock

production. In: P im e n tel, D. and C. W. H all (eds.), Food and Natural Resources, San Diego, Academic Press, San Diego, 69-83.

Rainfall and its erosivity in Ethiopia 184

C a m a r g o , M. B. P. and K. G. H u b b a r d (1999): Spatial and temporal variability of daily weather variables in sub-humid and semi-arid areas of the United States high plains. Agric. For. Meteorol. 93, 141-148.

C a m p b e l l , J. (1991): Land or peasants?: The Dilemma confronting Ethiopian Resource Conservation. Afr. Affairs 90, 5-21.

C e b u l a k , E. (1997): Variability of precipitation in selected regions of the Carpathians in the years 1951-1995. Geographia Polonica 70, 64-76.

Conway, D., Brooks, N. and K. R. Briffa (1998): Historical climatology and dendroclimatology in the Blue Nile River Basin, Northern Ethiopia. In: Servat, E., Hughes, D., Fritsch, J.-M. and M. Hulme (eds.), Water Resources Variability in Africa during the XXlh Century, Proceedings of the Abidjan 1998 Conference, Cote d’Ivoire, November 1998, IAHS Publication No. 252, IAHS, Wallingford, 243-251.

De Ploey, J. (1972): A quantitative comparison between rainfall erosion capacity in a tropical and a middle-latitude region. Geographia Polonica 23, 141-150.

D e g e f u , W. (1987): Some aspects of meteorological drought in Ethiopia. In:

G l a n t z , M. (ed.), Drought and hunger in Africa: Denying famine a future. Cambridge University Press, UK, 23-36.

D e x t e r , A. R. and B. K r o e s b e r g e n (1985): Methodology for determination of tensile strength of soil aggregates. J. Agric. Eng. Res. 31, 139-147.

D r e g n e , H. E. (1990): Erosion and soil productivity in Africa. J. Soil Water Conserv. 5, 431-436.

E c k h o l m , E. P. (1976): Loosing ground. W. W. Norton and Co., New York, 233 pp.

Rainfall and its erosivity in Ethiopia 185

E in fa lt, T h ., J o h a n , G. and A. Pfister (1998): On the spatial validity of heavy point rainfall measurements. Water Sci. Technol. 37, 21-28.

E k l u n d h , L. and P. P il e s j o (1990): Regionalisation and spatial estimation of Ethiopian mean annual rainfall. Int. J. Clim. 10, 473-494.

El H a ssa n in , A. S., L a b ib , M. T. and E. I. G a b er (1994): Rainfall erosivity and soil erodibility in Burundi. Egypt. J. Soil Sci. 34, 237-252.

E llis-Jo n e s, J and A. T en gberg (2000): The impact of indigenous soil and water conservation practices on soil productivity: Examples form Kenya, Tanzania and Uganda. Land. Degrad. Deveolp. 11, 19-36

E l l is o n , W. D. (1952): Raindrop energy and soil erosion. Emp. J. of Exp. Agric. 20, 82-99.

E l S w a if y , S. A. and E. W. D a n g l e r (1982): Rainfall erosion in the tropics: A

state of the art. In: K a r l , D . M. (ed.), Soil erosion and conservation in the tropics. ASA and Soil Sci. Soc., Madison, Wisconsin. ASA Spec. Publ. 43, 1- 25.

E l W a k e e l , A. and A. A s t a t k e (1996): Intensification of agriculture on vertisols to minimise land degradation in parts of the Ethiopian Highlands. Land. Degrad. Develop. 7, 57-67.

EPA [Environmental Protection Authority] and MEDC [Ministry of Economic Development and Co-operation] (1997): The Conservation Strategy of Ethiopia: The Federal Policy on the Environment, Vol. II, EPA/MEDC, Addis Ababa, 114 pp.

EPA [Environmental Protection Authority] (1999): National Report on the Implementation of the United Nations Convention to Combat Desertification. EPA, Addis Ababa, 54 pp.

Rainfall and its erosivity in Ethiopia 186

E s s e n w ANGER, O. M. (1986): Elements of Statistical Analysis. In:

L a n d e s b e r g , H. E. (e<±), World Survey of Climatology, Vol. IB, General Climatology, Elsevier Publishing Company, Amsterdam, 424 pp.

E z r a , M. and B. K a s s a h u n (1988): A review of the community forestry programme and an evaluation of its achievements: A socio-eonomic survey.

SIDA, Addis Ababa, 42 pp. [Cited in T e k l e 1999]

FAO (1965): Survey of the Awash River basin. Climatology and hydrology, Vol. 3. Imperial Ethiopian Government, United Nations Special Fund Project, FAO, Rome, 182 pp.

FAO (1977): Assessing Soil Degradation. FAO Soil Bull. 34, Rome, 70 pp.

FAO (1984): Assistance to land-use planning, Ethiopia. Agroclimatic resources inventory for land-use planning, Vol. I and II, Technical Report, No. 2, FAO/UNDP, Rome, Various pages.

FAO (1986): Ethiopian Highlands Reclamation Study. Final Report, Vol. 1, MoA/FAO, Rome, 378 pp.

FAO (1993): The state of food and agriculture. FAO Agriculture Series No. 26, FAO, Rome, 306 pp.

FAO (1995): A CD-ROM with World-wide Agroclimatic Data: FAOCLOM 1.2. FAO, Rome.

FAO (1998): The state of food and agriculture. FAO Agriculture Series No. 30, FAO, Rome, 371 pp.

F a r q u h a r s o n , F. A. K . and J. V. S u t c l i f f e . (1998): Regional variation of

African river flows. In: S e r v a t , E v H u g h e s , D ., F r i t s c h , J.-M. and M.

H u l m e (eds.), Water Resources Variability in Africa during the XXIh

Rainfall and its erosivity in Ethiopia 187

Century, Proceedings of the Abidjan 1998 Conference, Cote d’Ivoire, November 1998, IAHS Publication No. 252, IAHS, Wallingford, 161-169.

Favis-M ortlock, D. T. and A. J. T. G u e r r a (1999): The implications of general circulation model estimates of rainfall for future erosion: A case study from Brazil. Catena 37, 329-354.

Felix, B. (1998): Datenanalyse mit SPSS. Springer-Verlag, Berlin, 256 pp.

Fohrer, N., Berkenhagen, J. H e c k e r , J.-M. and A . R u d o l f (1999): Changing soil surface conditions during rainfall Single rainstorm / Subsequent rainstorms. Catena 37, 355-375.

F o u r n i e r , F. (1960): Climate et erosion: la relation entre l’erosion du sol par l’eau et les precipitations atmospheriques, 1. ed. Presses Universitaires de France, Paris, 201 pp.

F o x , D. M. and Y. Le B isson na is (1998): Process-based analysis of aggregate stability effects on sealing, infiltration, and interrill erosion. Soil Sci. Soc. Am. J. 62, 717-724.

Free, G. R. (1952): Soil movement by raindrops. Agr. Eng. 33, 491-494, 496.

G e b r e - m a r ia m , Z . (1998): Human Interaction and Water Quality in the Horn of Africa. Symposium on Science in Africa: Emerging Water Issues. Philadelphia, Pennsylvania, 17 February 1998. AAAS-Sub-Saharan Africa Project. http://www.aaas.org/Intemational/ssa/ewmi/, (10. 06. 2000).

G e b r e -M ic h a e l , Y . (1998): Indigenous Soil and Water Conservation Practices in Ethiopia: New Avenues for Sustainable Land Use. Adv. GeoEcol. 31, 1359-1366.

Rainfall and its erosivity in Ethiopia 188

G e t a h u n , A. (1978): Zonation of the Highlands of Tropical Africa: The Ethiopian Highlands. Working Paper, International Livestock Centre for Africa, Addis Ababa, 178 pp.

G o m e z , K. A. (1984): Statistical procedures for agricultural research. Second edition, John Wiley and Sons, New York, 679 pp.

G o o ssen s, C h . and A. B erg er (1986): Annual and seasonal climatic variations over the northern hemisphere and Europe during the last century, Ann. Geoph. 4, B. 4, 385-400.

G r if f it h s , J. F. (1972): Eastern Africa. In: G r if f it h s , J. F. (ed.), Climate of

Africa. In: L a n d e s b e r g , H. E. (ed.): World Survey of Climatology, Elsevier Publishing Company. Amsterdam, vol. 10, 313-347.

G r if f it h s , J. S. and K. S. R ic h a r d s (1989): Application of a low-cost database to soil erosion and soil conservation studies in the Awash Basin, Ethiopia. Land Degrad. Rehab. 1, 241-262.

Ha l l pik e, C. R. (1970): Konso Agriculture. J. Eth. Stud. 3, 31-43.

H a w k i n s , C. A. and J. E. W e b e r (1980): Statistical Analysis: Application to Business and Economics. Clark A. Hawkins and Jean E. Webber, New York, 626 pp.

H e l l d e n , U . and L. E k l u n d h (1988): National drought impact monitoring - A NOAA NDVI and precipitation data study of Ethiopia. Lund Studies in Geography, Ser. C. General Mathematical and Regional Geography, No. 15, 55 pp.

H e r w e g , K . (1992): A Survey Method for Soil Erosion Assessment and

Conservation Control. In: H u r n i, % H . and K . T a t o (eds.), Erosion,

Rainfall and its erosivity in Ethiopia 189

Conservation, and Small-Scale Farming. International Geographica Bemesia, Missouri, 1-13.

Ho b e n , A. (1995): Paradigms and Politics: The Cultural Construction of Environmental Policy in Ethiopia. World Develop. 23, 1007-1021.

H o b e n , A. (1997): The Cultural Construction of Environmental Policy: Paradigms and Politics in Ethiopia. The Ecologist 27, 55-62.

H u b b a r d , K. G. (1994): Spatial variability of daily weather variables in the high plains of the USA. Agric. For. Meteorol. 68, 29-41.

Hudson, N. W. (1961): Raindrop size distribution, kinetic energy and intensity of sub-tropical rainfall. Inter-African Conference on Hydrology, Publication No. 66, 127-133.

H u d so n , N. W. (1971): Soil conservation. B T Batsford Ltd., London, 320 pp.

H u ffn a g e l , H. P. (1961): Agriculture in Ethiopia. Rome, 484 pp.

H u r n i, H. (1982): Klima und Dynamik der Hohenstufung von der letzten Kaltzeit bis zur Gegenwart. Hochgebirge von Semien-Athiopien. Vol. II. Jahrbuch der Geographischen Gesellschaft von Bern, Beiheft 7, G 13, 196 pp.

H u r n i, H. (1987): Erosion-productivity-conservation systems in Ethiopia. In:

P la S entis (ed.), Soil conservation and productivity, Proceedings of the 4lh International Conference on Soil Conservation, Maracay, Venezuela, 3.- 9.11.1985,654-764.

HURNI, H. (1993): Land degradation, famine, and land resources scenarios in Ethiopia. In: PIMENTEL, D. (ed.), World Soil Erosion and Conservation, Cambridge University Press, UK., 28-61.

Rainfall and its erosivity in Ethiopia 190

IUCN [International Union for the Conservation of Nature] (1993): Ethiopia: Environmental Synopsis. IUCN and CEC, 33 pp.

J a n ssen , J. and W. La a t z (1997): Statistische Datenanalyse mit SPSS fur Windows. 2. neubearbeitete Auflage, Springer-Verlag, Heidelberg, 636 pp.

J u tzi, S. C. (1990): Deep black soils (Vertisols) management options for

Ethiopian highlands. In: H u r n i, H. and B. M esserli (eds.), African mountains and highlands: Problems and perspectives. African Mountains Association, Missouri, 99-106.

Ka p p e l , R. (1996): Economic Analysis of Soil Conservation in Ethiopia: Issues and Research Perspectives. Soil Conservation Research Report No. 35, 29 pp.

KlELY, G. (1999): Climate change in Ireland from precipitation and streamflow observations. Adv. Water. Resour. 23, 141-151.

K o t h y a r i, U. C., S in g h , V. P. and V. A ra v a m u th a n (1997): An investigation of changes in rainfall and temperature regime of the Ganges Basin in India. Water Resour. Manage. 11, 17-34.

Kousky, J. V., Thiaw, W. M. and R. C. Levy (1998): Variation in satellite- derived precipitation estimates and atmospheric circulation over Africa. In: Servat, E., Hughes, D., Fritsch, J.-M. and M. Hulme (eds.), Water Resources Variability in Africa during the XXth Century, Proceedings of the Abidjan 1998 Conference, Cote d’Ivoire, November 1998, IAHS Publication No. 252, IAHS, Wallingford, 11-19.

Ko w a l , J. M. and A. H. K assam (1977): Energy load and instantaneous

intensity of rainstorms at Samaru, northern Nigeria. In: G r e e n l a n d , D. J. and

R. Lal (eds.), Soil conservation and management in the humid tropics. John Wiley and Sons Ltd., New York, 57-70.

Rainfall and its erosivity in Ethiopia 191

K r u g er , H.-J., F a n t a w , B., G e b r e-M ic h a el, Y. and K. Ka jel a (1996):

Creating inventory of indigenous SWC measures in Ethiopia. In: Re d , C.,

Sc o o n e s, I. and C. T o u lm in (eds.), Sustaining the Soil: Indigenous Soil and Water Conservation in Africa. International Institute for Environment and Development, London, 170-180.

KUH N EL, I. (1983): Bewolkung, Niederschlag und atmospharische Zirkulation in

Ostafrika. In: BUDM IGER, G. (ed.), Berner Beitrage zur Afrikaforschung. Jahrbuch der Geographischen Gesellschaft von Bern, Bd. 54, 139-155.

K u n d z ew ic z, Z. W. and L. So m ly o d y (1997): Climatic change impact on water resources in a systems perspective. Water Resour. Manage. 11, 407- 435.

K u r u , A. (1986): Soil Erosion and Strategic State Policy: The case of Ethiopia. Diss., Publications of the Department of Environmental Conservation at the University of Helsinki, No.7, Department of Environmental Conservation, University of Helsinki at Viikkii, Helsinki, 123 pp.

Lado, C. (1998): The transfer of agricultural technology and the development of small-scale farming in rural Africa: Case studies from Ghana, Sudan, Uganda, Zambia and South Africa. GeoJoumal 45, 165-176.

L a l, R. (1988): Soil degradation and the future of agriculture in Sub-Saharan Africa. J. Soil Water Conserv. 43, 444-451.

La l , R. (1990): Soil erosion in the tropics: Principles and management. McGraw-Hill, Inc., New York, 580 pp.

La l , R., La w so n , T. L. and A. H. A n a sta se (1980): Erosivity of tropical rains.

In: D e B o o d t , M. and D. G a b riels (eds.), Assessment of Erosion. John Wiley and Sons Ltd., Chichester, 143-151.

Rainfall and its erosivity in Ethiopia 192

L e B isso n n a is, Y. (1996): Aggregate stability and assessment of soil crustability and erosibility: I Theory and Methodology. Eur. J. Soil Sci. 47, 425-437.

L iebig , W. (1997): Desktop-GIS mit Arc View. Leitfaden fur Anwender. Wichmarm Vg., Heidelberg, 340 pp.

L u pw a y i, N. Z., G ir m a , M. and I. H a q ue (2000): Plant nutrient contents of cattle manures from small-scale farms and experimental stations in the Ethiopian highlands. Agric. Ecosyst. Environ. 78, 57-63.

M a h er a s, P. (1990): Temporal and spatial characterisation of annual precipitation over the Balkans in the twentieth century. Int. J. Clim. 10, 495- 504.

M a jo r , D. C. and K. D. F red er ic k (1997): Water resources planning and climate change assessment methods. Clim. Change 37, 25-40.

M a y er -L elxner, G. (1999): Probleme der Energieversorgung landlicher Gebiete in Afrika: Okologische und Soziookonomische Aspekte der Brennholzkrise in der Dritten Welt unter besonderer Berucksichtigung von Athiopien. Weltforum Verlag GmbH, Munchen, 356 pp.

M ee k , D. W ., P r u eg e r , J. H., S a u er , T. J., K u sta s, W. P., H ipps, L. E. and L.

J. H a tfield (1999): A note on recognizing autocorrelation and using autocoregression. Agric. For. Meteorol. 96, 9-17.

M eigh J. R., M c Ke n z ie , A. A. and K. J. Sene (1999): A Grid-based Approach to Water Scarcity Estimates for Eastern and Southern Africa. Water Resour. M anage. 13, 85-115.

M ic k l in , P. P. (1996): Man and the water cycle: Challenges for the 2P ‘ century. GeoJoumal 39, 285-298.

Rainfall and its erosivity in Ethiopia 193

M id d l e t o n , N. and D. T hom as [eds.] (1997): World Atlas of Desertification. Second edition, UNEP, John Wiley and Sons, Inc., New York, 182 pp.

M ik elsen , P. S., M a d se n , H., A rn b jer g -N ie lsen , K., Jo r g e n se n , H. K.,

R o sb je r g , D. and P. H a r rem o es (1998): A rationale for using local and regional point rainfall data for design and analysis of urban storm drainage systems. Water Sci. Technol. 37, 7-14.

Mikesell, M. W. (1960): Deforestation in northern Morocco. Science 132, 441 - 454.

M itc h ell, J. M., D zerzeevskii, B., F lo h n , H., H o f m e y r , W. L., L a m b , H. H.,

Ra o , K. N. and C. C. W allen (1966): Climate change. Technical Note 79, W M O , Geneva, 79 pp.

M o h r , P. A. (1971): The geology of Ethiopia. University college of Addis Ababa, Asmara, 268 pp.

Moreda, F. and W. Bauwens (1998): Influences of variability of rainfall on flow regimes in central Ethiopia. In: Servat, E., HUGHES, D., Fritsch, J.-M. and M. Hulme (eds.), Water Resources Variability in Africa during the XXth Century, Proceedings of the Abidjan 1998 Conference, Cote d’Ivoire, November 1998, IAHS Publication No. 252, IAHS, Wallingford, 297-305.

MUCKENHAUSEN, E. (1976): Bisherige Untersuchungen iiber Bodenabtrag in Deutschland und anderen europaischen und auBereuropaischen Landem. In:

R ic h t e r , G. and W. S perling (eds.), Bodenerosion in Mitteleuropa, Wege der Forschung. Wissenschaftliche Buchgesellschaft, Darmstadt, Bd. 430, 26- 33.

NICHOLSON, S. E. (1998): Interannual and interdecadal variability of rainfall

over the African continent during the last two centuries. In: S e r v a t , E.,

H u g h e s, D., Fr it sc h , J.-M. and M. H ulm e (eds.), Water Resources

Rainfall and its erosivity in Ethiopia 194

Variability in Africa during the XXth Century, Proceedings of the Abidjan 1998 Conference, Cote d ’Ivoire, November 1998, IAHS Publication No. 252, IAHS, Wallingford, 107-117.

N ill, D., Sc h w e r t m a n n , U., S a b el -K o sh ell a , U., Ber n h a r d , M. and J.

Breuer (1996): Soil erosion by water in Africa. Deutsche Gesellschaft fur Technische Zusammenarbeit (GTZ), Eschbom, 292 pp.

NMSA [National Meteorological Service Agency] (1996): Climatic and Agroclimatic Resources of Ethiopia. NMSA Meteorological Research Report Series (Ethiopia), V. 1.1, NMSA, Addis Ababa, 137 pp.

N o r m a n , W. R. (1998): Changes in land tenure and use during thirty years of small irrigation system operation: A case study from the Sahel. Irrig. Drain. Syst. 12, 161-183.

O d u ro -A fr iy ie , K. (1996): Rainfall erosivity map for Ghana. Geoderma 74, 161-166.

Oldeman, L. R., Hakkeling, R. T. A. and W. G. Sombroek (1991): World map of the status of human-induced soil degradation. An explanatory note. Global assessment of soil degradation (GLASOD). Second revised edition, International Soil Reference and Information Centre (ISRIC) and United Nations Environment Programme (UNEP), Wageningen, 34 pp. + 3 maps.

O lso n , G. W. (1981): Archaeology: Lessons on future soil use. J. Soil and Water Cons. 36, 261-264.

O m it i, J. M., Pa r to n , K. A., SlNDEN, J. A. and S. K. EHUE(1999): Monitoring changes in land-use practices following agrarian de-collectivisation in Ethiopia. Agric. Ecosyst. Environ. 72, 111-118,

Rainfall and its erosivity in Ethiopia 195

Opoku-Ankomah, Y. and B. A. Amisigo (1998): Rainfall and runoff variability in the southwestern river system of Ghana. In: Servat, E., Hughes, D., Fritsch, J.-M. and M. Hulme (eds.), Water Resources Variability in Africa during the XXth Century, Proceedings of the Abidjan 1998 Conference, Cote d'Ivoire, November 1998, IAHS Publication No. 252, IAHS, Wallingford, 307-314.

O sm a n , M., S k o w r o n ek , A. and P. S au erb or n (2000): Land and water resources management in Ethiopia: What did we leam, where do we go? In: Universitat Hohenheim (ed.), International agricultural research: A contribution to crisis prevention, Proc. Int. Agric. Res. Hohenheim, 11-12 October 2000, 138-139

Pa n o f sk y , H. A. and G. W. B rier (1958): Some Applications of Statistics to Meteorology. The Pennsylvania State University, Pennsylvania, 222 pp.

Pauwelyn, P. L. L., Lenvain, J. S. and W. K. Sakala (1988): Iso-erodent map of Zambia, Part I: The calculation of erosivity indices from a rainfall data bank. Soil Technology 1, 235-250.

Pa v lo po u lo s, H. and J. G ritsis (1999): Wet and dry epoch durations of spatially averaged rain rate, their probability distributions and scaling properties. Environ. Ecol. Stat. 6, 351-380.

Pfister, L., Humbert, J. and L. Hoffmann (2000): Recent trends in rainfall- runoff characteristics in the Alzette River Basin, Luxembourg. Clim. Change 45, 323-337.

Pile, K. (1996): Soil erosion, policy and people’s perceptions in a rural community in KwaZulu-Natal, South Africa. GeoJoumal 39, 59-64.

Rainfall and its erosivity in Ethiopia 196

P im e n tel, D., H a r v e y , C., Reso su d a r m o , P., S in c l a ir , K., K u r z , D.,

M cN a ir , M., C r ist, S., Sh pr it z , L., F it t o n , L., Sa f fo u r i, R. and R. B lair (1995): Environmental and Economic Costs of Soil Erosion and Conservation Benefits. Science 267, 1117-1123.

Po o r e, M. E. D. and C. F ries (1985): The Ecological Effects of Eucalyptus. Food and Agriculture Organization of the United Nations (FAO) Forestry

Paper No. 59, FAO, Rome, 87 pp. [Cited in T u r n b u ll 1999].

Potratz, K., Henk, U. and A. Skowronek (1991): Luftsprengung, Aggregatzerfall und Verschlammung als wichtige Prozesse der Erosionsdynamik - Ergebnisse von Starkregensimulationen an LoBboden. Z. Geomorph. N.F., Suppl.-Bd. 89, 21-33.

Pozo, J., Basaguren, A., E lo s e g u i, A., M o lin e r o , J., F a b r e , E. and E. C h a u v e t (1998): Afforestation with Eucalyptus globulus and leaf litter decomposition in streams of northern Spain. Hydrobiologia 373/374, 101- 109.

Ram b a b u , T e jw a in , K. G., A g a r a l, M. C. and L. S. B h u sh a n (1978): Distribution of erosion index and iso-erodent map of India. Indian J. Soil Conserv. 6, 1-12.

Ra p p , J. and C.-D. Sc hOn w iese (1996): Atlas der Niederschlags- und Temperaturtrends in Deutschland 1891 - 1990. Frankfurter geowiss. Arb., Serie B, 5, 253 pp.

R e n a r d , K. G., Fo s t e r , G. R. and G. A. W eesies (1997): Predicting soil erosion by water: A guide to conservation planning with the revised universal soil loss equation (RUSLE). Agricultural Handbook, 703. United States Department of Agriculture (USDA), Washington D. C., 384 pp.

Rainfall and its erosivity in Ethiopia 197

Re n sc h l e r , C. S., M a n n a er t s, C. and B. D iek k ru g er (1999): Evaluating spatial and temporal variability in soil erosion risk - rainfall erosivity and soil loss ratios in Andalusia, Spain. Catena 34, 209-225.

Ric h t e r , G. (1976): Was ist Bodenerosion? In: Ric h t e r , G. and W. S perlin g (eds.), Bodenerosion in Mitteleuropa, Wege der Forschung. Wissenschaftliche Buchgesellschaft, Darmstadt, Bd. 430, 75-105.

Ric h t e r , G. (1998): Bodenerosion als Weltproblem. In: R ic h t e r , G. (ed.), Bodenerosion: Analyse und Bilanz eines Umweltproblems. Wissenschaftliche Buchgesellschaft, Darmstadt, 231-242.

Risb e y , J., K a n d l ik a r , M., D o w la ta b a d i, I-I. and D. G r a etz (1999): Scale, context and decision making in agricultural adaptation to climate variability and change. Mitig. Adapt. Strat. Glob. Change. 4, 137-165.

R o o se , E. J. (1977): Use of the Universal Soil Loss Equation to predict Erosion in West Africa. Soil Conservation Society of America, Ankeny, Iowa. Spec. Publ. 21,60-74.

Sa u er b o r n , P. (1994): Die Erosivitat der Niederschlage in Deutschland - Ein Beitrag zur quantitativen Prognose der Bodenerosion durch Wasser in Mitteleuropa. Diss., Bonner Bodenkundl. Abh. 13, 189 pp.

Sa u e r b o r n , P. and J. G r u n er t (1995): Neueste Ergebnisse zur potentiellen Erosionsgefahrdung durch Niederschlage in Deutschland. Regensburger Geogr. Sehr. 25, 157-164.

Sa u e r b o r n , P., Kl e in , A., B o t sc h e c k , J. and A. Sk o w r o n ek (1999): Future rainfall erosivity derived from large-scale climate models - Methods and scenarios for a humid region. Geoderma 93, 269-276.

Rainfall and its erosivity in Ethiopia 198

Sc h m id t , J., W e r n er , M. v ., and A. M icha el (1999): Application of the EROSION 3D model to the CATSOP watershed, The Netherlands. Catena 37, 449-456.

Sc h o n w ie se , C.-D. (1985): Praktische Statistik fur Meteorologen und Geowissenschaftler. Gebriider Bomtraeger, Stuttgart, 231 pp.

Sc h u lt ze , J. H. (1976): Das Wesen der Bodenerosion und ihre Problematik in

Thiiringen. In: R ic h t er , G. and W. S perling (eds.), Bodenerosion in Mitteleuropa, Wege der Forschung. Wissenschaftliche Buchgesellschaft, Darmstadt, Bd. 430, 57-74.

Sc h w er tm a n n , U. (1981): Die Vorausschatzung des Bodenabtrags durch Wasser in Bayern. Lehrstuhl fur Bodenkunde d. TU., Miinchen/Freising- Weihenstephan, 126 pp.

S elesh i, Y. and G. R. D em a r ee (1993): The temporal distribution of Ethiopian

meteorological droughts in the XXlh century. In: S y m o n e s, J.-J., D ev o s, P.,

Ra m m e l o o , J. and C h . V erstra eten (eds.), Biological Indicators of Global Change. Symposium, 7-9 May 1992, Brussels, 77-91.

S ele sh i, Y. and G. R. D em aree (1995): Rainfall variability in the Ethiopian and Eritrean highlands and its links with the southern Oscillation Index. J. Biogeography 22, 945-952.

S h ifer a w , B. and S. T Ho ld en (1999): Soil Erosion and Smallholders’ Conservation Decisions in the Highlands of Ethiopia. World Develop. 27, 739-752.

S h ifer a w , B. and S. T. Ho ld en (2000): Policy instruments for sustainable land management: The case of highland smallholders in Ethiopia. Agric. Econ. 22, 217-232.

Rainfall and its erosivity in Ethiopia 199

S m it h , K. (1995): Precipitation over Scotland, 1757-1992: Some Aspects of Temporal Variability. Int. J. Clim. 5, 543-556.

S n e y er s, R. (1990): On the Statistical Analysis of Series of Observations. Technical Note, 143, WMO, Geneva, 192 pp.

S tAh l, M. (1990): Environmental degradation and political constraints in Ethiopia. J. Disast. Stud. Manage. 14, 140-150.

St Ah l, M. and A. W ood (1989): Ethiopia: National Conservation Strategy. Inception report, IUCN, Addis Ababa, 76 pp.

STEINER, K. G. ( 1 9 9 6 ): Causes of soil degradation and development approaches to sustainable soil management. Deutsche Gesellschaft fur Technische Zusammenarbeit (GTZ) GmbH, Eschborn, 50 pp.

S u z u k i, H. (1967): Some aspects o f Ethiopian climates. Eth. Geog. J. 4, 19-22.

T a t o , K. (1964): Rainfall in Ethiopia. Eth. Geog. J. 2, 28-36.

T e k le, K. (1999): Land degradation problems and their implications for food shortage in South Wello, Ethiopia. Environ. Manage. 23, 419-427.

T esfa y e, H. (1990): Drought in Ethiopia. In: WMO (ed.), Extended abstracts of papers presented at the 3rd WMO symposium on meteorological aspects of tropical droughts with emphasis on long-range forecasting. WMO/TD, No. 353, Geneva, 273-278.

T e sse m a , A. B., B ir n ie, A. and B. T engnAs (1993): Useful Trees and Shrubs for Ethiopia, Propagation and Management for Agricultural and Pastoral Communities. Regional Soil Conservation Unit (RSCU)/Swedish International Development Authority (SIDA). Technical Handbook Series, 5, Nairobi, 474 pp.

Rainfall and its erosivity in Ethiopia 200

To r r i, D., R eg u es, D., P e ll eg r in i, S. and P. B a zzoffi (1999): Within-storm soil surface dynamics and erosive effects of rainstorms. Catena 38, 131-150.

T u r n b u ll, J. W. (1999): Eucalypt plantations. New Forests 17, 37-52.

UIA [Union of International Associations] (1994-95): Encyclopaedia of World Problems and Human Potential. Fourth edition, 300 pp.

UNDHA [United Nations Department of Humanitarian Affairs] (1995): Ethiopia: Floods. DHA Geneva situation report Nr. 1, Geneva. Volunteers in Technical Assistance (VITA), Disaster Information Centre. http://www.vita.org/disaster/sitrep/95b/0068.html (5 November 1997).

UNEP [United Nations Environment Programme] (1983): Ecology and Environment. What do we know about desertification? Desert. Cont. Bull. 3, 2-9.

V e r sc h u r e n , D., La ir d , K. R. and B. F. C u m m in g (2000): Rainfall and drought in equatorial East Africa during the past 1 100 years. Nature 43, 410- 414.

V ir g o , K. J. and R. N. M u n ro (1978): Soil and erosion features of the central plateau region of Tigray, Ethiopia. Geoderma 20, 131-157.

VITA [Volunteers in Technical Assistance] (1996a): Ethiopia: Floods. DHA Geneva Information Report No. 2. VITA, Disaster Information Centre. http://www.vita.org/disaster/sitrep/96b/0000.html (22 August 1996).

VITA [Volunteers in Technical Assistance] (1996b): Ethiopia: Floods. DHA Geneva Information Report No. 3. VITA, Disaster Information Centre. http://www.vita.org/disaster/sitrep/96b/0005.html (26 August 1996).

Rainfall and its erosivity in Ethiopia 201

VITA [Volunteers in Technical Assistance] (1997a): Ethiopia: Floods. DHA Geneva Information Report No. 1. VITA, Disaster Information Centre. http://www.vita.org/disaster/sitrep/97b/0095.html (5 November 1997).

VITA [Volunteers in Technical Assistance] (1997b): Ethiopia: Floods. DHA Geneva Information Report No. 3. VITA, Disaster Information Centre. http://www.vita.org/disaster/sitrep/97b/0068.html (11 December 1997).

VITA [Volunteers in Technical Assistance] (1998): Somalia: Floods. DHA Geneva Situation Report No. 12. VITA, Disaster Information Centre. http://www.vita.org/disaster/sitrep/98a/0000.html (6 January 1998).

V o e l k e l , K. E. and S. B. G er ber (1999): Using SPSS for Windows: Data Analysis and Graphics. Springer-Verlag, New York, 228 pp.

W a n it z e k , U. and H. S ippel (1998): Land rights in conservation areas in Tanzania. GeoJoumal 46, 113-128.

W est ph a l , E. (1975): Agricultural systems in Ethiopia. Agricultural Research Report, Centre for Agricultural Publishing and Documentation, Wageningen, 278 pp.

W isc h m e ier, W. H. (1962): Rainfall erosion potential. Agric. Eng. 42, 212- 225.

W isc h m e ier, W. H. and D. D. S m ith (1978): Predicting Rainfall Erosion Losses - A Guide to Conservation Planning. Agricultural Handbook 537, United States Department of Agriculture (USDA), Sci. and Educ. Admin., Washington, D. C., 57 pp.

W o ie n , FI. (1995): Deforestation, Information and Citations: A Comment on Environmental Degradation in Highland Ethiopia. GeoJoumal 4, 501-511.

Rainfall and its erosivity in Ethiopia 202

W o l d e -M a r ia m , M. (1984): Rural vulnerability to famine in Ethiopia: 1958- 1977. Vikas, New Delhi, 191 pp.

Wood, A. W., L ett en m a ier , D. P. and R. N. P a l m e r (1997): Assessing climate change implications for water resource planning. Clim. Change 37, 203-228.

Wood, C. (1977): A preliminary chronology of Ethiopian droughts: Drought in Africa. International African Institute, London, 68-73.

W o rld Ban k (2000): World Bank atlas. 32nd edition, World Bank, Washington, DC., 63 pp.

Wright, C. and Y. A damseged (1984): An Assessment of the Causes, Severity, Extent and Probable Consequences of Degradation in the Highlands of Ethiopia: Ethiopian Highlands Reclamation Study (EHRS), Working Paper 3, Land Use and Regulatory Department of the Ministry of Agriculture/FAO, Addis Ababa, 59 pp.

Yoo, C., Valdes, J. B. and G. R. N orth (1998): Evaluation of the impact of rainfall on soil moisture variability. Adv. Water Resour. 21, 375-384.

Rainfall and its erosivity in Ethiopia 203

10 ANNEXES PAGE List of annex Tables and Figures

Tab. 17: Network o f selected meteorological stations o f Ethiopia...... 204 Tab. 18: Weather stations in FAO rainfall pattern regions of Ethiopia...... 208 Tab. 19: Statistical summary of long-term annual rainfall at selected meteorological stations in Ethiopia...... 211 Tab. 20: Statistical summary o f long-term annual rainfall erosivity factors for selected meteorological stations in Ethiopia...... 218 Tab. 21: Statistical characteristics of long-term summer rainfall erosivity in the central highlands of Ethiopia...... 225 Tab. 22: Statistical characteristics of long-term autumn rainfall erosivity in the central highlands of Ethiopia...... 227 Tab. 23: Statistical characteristics of long-term spring rainfall erosivity in the central highlands of Ethiopia...... 229 Tab. 24: Statistical characteristics of long-term winter rainfall erosivity in the central highlands o f Ethiopia...... 231 Tab. 25: Regression models of long-term autumn rainfall erosivity for the central highlands of Ethiopia...... 233 Tab. 26: Regression models of long-term spring rainfall erosivity for the central highlands of Ethiopia...... 235 Tab. 27: Regression models of long-term winter rainfall erosivity for the central highlands of Ethiopia...... 237 Fig. 58: Autocorrelation coefficients o f long-term rainfall erosivity for selected weather stations in the central highlands of Ethiopia...... 239

Rainfall and its erosivity in Ethiopia 204

Tab. 17: Network of selected meteorological stations of Ethiopia

Weather Latitude Longitude Elevation Years o f Data Nr. station (N) (E) (m) observation source 1 Abella 6°56' 38°28' 1900 1932-1962 NM SA/M oW R 2 Adaba 7°02' 39°40' 2410 1968-1987 NMSA/MoWR 3 Addis Ababa 9°02' 38°42' 2408 1898-1996 NOAA 4 Addis Zemen 12°07' 37°52' 1550 1974-1986 NM SA/M oW R 5 Adigrat 14°27' 39°45' 2457 1970-1991 NMSA/MoWR 6 Adi Gudom 13°27' 39°52' 2200 1971-1988 NMSA/MoWR 7 Agarfa 7°16' 39°49' 2520 1968-1989 NM SA/M oW R 8 Agaro 6°51' 35°23' 1360 1953-1970 NOAA 9 Alaba Kulito 7°19' 38°06' 1750 1953-1997 NMSA/MoWR 10 Alamata 12°31' 39°41' 2200 1957-1991 FAO 11 Albuco 11°46' 39°34' 2100 1962-1983 NMSA 12 Alemaya 9°26' 42°01' 2125 1957-1993 NMSA 13 Aiem Ketema 10°02' 39°02' 2280 1973-1997 NMSA 14 Aleta Wondo 6°33' 38°25' 1910 1954-1971 NMSA/MoWR 15 Amaro Kelo 6°09' 37°54' 1400 1974-1983 NM SA/M oW R 16 Amed Ber 11°53' 37°50' 2090 1977-1985 NMSA/MoWR 17 Ancharo 11°03' 39°38' 2000 1980-1987 NMSA 18 Ankober 9°00' 39°40' 2600 1962-1983 NMSA 19 Arba Minch 6°05' 37°38' 630 1970-1988 NMSA 20 Areka 7°40' 37042' 1750 1971-1982 NMSA/MoWR 21 Arjo 8°45' 36°30' 2560 1985-1996 NMSA 22 Artuma 10°35' 40o01' 1880 1975-1987 NMSA 23 Asasa 7°12' 39° , 9 ' 2370 1971-1987 NMSA/MoWR 24 Asela 7°57' 39°08' 2350 1966-1988 NM SA/M oW R 25 Askuna School 10°58' 36°54' 2060 1980-1987 NMSA/MoWR 26 Asosa 10°01' 34°31' 1560 1978-1996 NMSA 27 Assita 11°34' 4I°2 7 ' 430 1964-1985 NMSA 28 Awassa 7°05' 38°29' 1700 1972-1987 NMSA/MoWR 29 Aykel 12°52' 37°05' 2800 1960-1991 FAO 30 Bacco 5°50' 36°38' 2300 1954-1966 NMSA 31 Bahir Dar 12°00' 37°40' 1805 1961-1996 NOAA 32 Bati 11°22- 40°05' 1660 1960-1991 FAO 33 Bedeno 9°08' 41°38' 2050 1966-1994 NMSA 34 Bedessa 8°55' 40°46' 1820 1957-1997 NMSA 35 Bele 7°05' 37035. 1200 1974-1985 NM SA/M oW R 36 Bilate Farm 6°45' 38°04' 1300. 1969-1987 NM SA/M oW R 37 Birr Valley 10°38' 37°07' 1700 1979-1987 NM SA/M oW R 38 Bokoksa 11°22' 39°53' 1800 1975-1987 NMSA 39 Bora i0°41' 40°03' 1500 1975-1987 NMSA 40 Boru Meda 11°2 2 ' 39°28' 2720 1963-1983 NMSA 41 Burji 5°22' 37°54' 1720 1956-1975 NM SA/M oW R 42 Butajira 8°07' 38°22' 2000 1968-1997 NM SA/M oW R 43 ChefTa 10°35' 39°27' 1400 1963-1985 NMSA

Rainfall and its erosivity in Ethiopia 205

Tab. 17: Continued

Weather Latitude Longitude Elevation Years of Data Nr. station (N) (E) (m) observation source 44 Chelelektu 6°or 38°22' 1686 1973-1985 NM SA/M oW R 45 Chencha 6°15' 37°34' 2800 1953-1966 NMSA 46 Chiro 9°04' 40°52' 1900 1962-1997 NMSA 47 Dabat 12°59' 37°45' 2685 1970-1984 NMSA/MoWR 48 Dambacha 10°34' 37°28’ 2100 1966-1987 NMSA/MoWR 49 Dangla 11°07' 36°25' 1290 1955-1969 NMSA/MoWR 50 Debre Markos 10°20' 37°40' 2515 1954-1991 NOAA 51 Debre Tabor 11°53' 38°02' 2410 1960-1991 NOAA 52 Debre Work 10°44' 38°08' 2740 1975-1987 NM SA/M oW R 53 Debre Zeit 8°44' 38°57' 1900 1951-1984 NOAA 54 Deke Stefanos 1I°54' 37°16' 1795 1975-1987 NMSA/MoWR 55 Delo Sebro 7° 15' 40°28' 2200 1968-1985 NM SA/M oW R 56 Dembi Dolo 8°31' 34°48‘ 1850 1987-1993 NMSA 57 Dessie 11°06' 39038- 2500 1962-1987 NMSA 00 Dhagahbour 8°13' 43°33' 1070 1954-1997 NMSA 59 Dhera 8°20' 39°19' 1770 1976-1996 NMSA/MoWR 60Dila 6°25' 38°18' 1670 1954-1986 NMSA/MoWR 61 Dinsho 7°04' 39°46' 3450 1969-1987 NMSA/MoWR 62 Dire Dawa 9°36' 41°5 1' 1260 1952-1990 NMSA/NOAA 63 Dodola 6°59' 39°11' 3000 1954-1979 NM SA/M oW R 64 Ebnet 12°08' 38°03' 1800 1970-1984 NM SA/M oW R 65 Ejaji 9°00' 370, 9 ' 1900 1966-1997 NMSA 66 Elias 10°18' 37°28' 2140 1981-1987 NM SA/M oW R 67 Elidar 12°04' 4I°55' 660 1975-1985 NMSA 68 Enfranz 12° 11' 37°41' 1500 1977-1985 NMSA/MoWR 69 Felege Neway 6° 18' 36°53' 1380 1970-1985 NMSA 70 Fiche 9°48' 38°42' 2750 1954-1997 NMSA 71 Finoteselam 10°41' 37°16' 1900 1979-1987 NMSA/MoWR 72 Galamso 8°49' 40°31' 1820 1964-1991 NMSA 73 Gambela 8°15' 34°35' 0 1908-1983 NOAA 74 Gato 5°25' 37027' 1730 1973-1987 NMSA 75 Gerese 5°55' 37°18' 2300 1873-1987 NMSA 76 Gessera 7°22' 40°11' 2320 1968-1986 NM SA/M oW R 77 Gidole 5°39' 37°27' 2550 1957-1972 NMSA 78 Ginir 7°13' 40°42' 1750 1968-1987 NMSA/MoWR 79 Goba 7°02' 40°00' 2710 1961-1984 NMSA/MoWR 80 Gobessa 7°48' 37°30' 2300 1969-1985 NM SA/M oW R 81 Gode 5°54' 43°35' 295 1967-1981 NMSA 82 Gohatsion !0°02' 38°14' 2560 1972-1997 NMSA 83 Gonder 12°53' 37°40' 1966 1960-1996 NM SA/M oW R 84 Gorgora 12°15' 37°18' 1830 1972-1985 NM SA/M oW R 85 Goro 7°00' 40°28' 1780 1974-1988 NM SA/M oW R 86 Grawa 9°08' 41°50' 2100 1968-1997 NMSA

Rainfall and its erosivity in Ethiopia 206

Tab. 17: Continued

Weather Latitude Longitude Elevation Years of Data Nr. station (N) (E) (m) observation source 87 Gumadaye 5°35' 37°40' 1730 1976-1987 NMSA 88 Gumara 11°50' 37°38' 1880 1978-1985 NM SA/M oW R 89 Gursum 9°21' 42°23' 1900 1969-1987 NMSA 90 Hagere Mariam 5°36' 38°20' 1900 1956-1983 NM SA/M oW R 91 Hagere Selam 6°28' 38°31' 2840 1956-1985 NMSA/MoWR 92 Haik 11°19' 39°4' 1900 1962-1984 NMSA 93 Hebeno 6°44' 38°59' 2700 1968-1986 NM SA/M oW R 94 Hosana 7°33' 37°52' 2200 1953-1968 NOAA 95 Huruta 8°09' 39°22' 2200 1969-1988 NM SA/M oW R 96 Idoto 7°31' 40°03' 2480 1969-1985 NMSA/MoWR 97 Jijiga 9°20' 42°47' 1775 1952-1984 N MS A/NO A A 98 Jima 7°40” 36°50' 1665 1952-1990 NOAA 99 Jinka 5°47' 36°34' 1420 1970-1988 NMSA 100 Kachise 9°35' 37°20' 2520 1986-1996 NMSA 101 Kebri Dahar 6°44' 44°18' 545 1957-1996 NMSA 102 Kelem Meda 11°11' 39°31' 2500 1979-1987 NMSA 103 Keleta 8°19” 39°25' 1500 1962-1986 NMSA/MoWR 104 Kemba 6°03' 37°10' 1850 1974-1987 NMSA 105 Key Afer 5°31' 36°44' 1550 1979-1988 NMSA 106 Kibre Mengist 5°50' 38°58' 1680 1974-1985 NMSA/MoWR 107 Kobo 12°08' 39°38' 1420 1960-1991 FAO 108 Kofole 7°04' 38°47' 2620 1961-1987 NMSA/MoWR 109 Kolme 5°46' 3 7 ° ll' 1200 1976-1985 NMSA 1 10 Kombolcha 11°07' 39°44' 1916 1952-1991 NOAA/FAO 111 Konso 5°20' 37°27' 1550 1971-1987 NMSA 112 Kore 7°13' 38°54' 2500 1969-1985 NMSA/MoWR 113 Korem 11 °3 r 39°31' 3000 1976-1987 NMSA/MoWR 114 Kule Meskel 11°57' 39°14' 2300 1973-1987 NMSA/MoWR 115 Kunzela 11°51' 37°01' 1920 1980-1987 NMSA/MoWR 116 Lalibala 12°02' 39°03' 2500 1969-1985 NMSA/MoWR 117 Majete 10°27' 39°51' 2000 1962-1996 NMSA 1 18 Maksenget 12°22 ' 37°33' 1450 1970-1985 NMSA/MoWR 119 Mandura 1l°07' 36°25' 1290 1972-1987 NMSA/MoWR 120 Mankusa I0°40' 37°09' 2000 1972-1987 NMSA/MoWR 121 Mararo 7°27' 39°22' 2940 1968-1988 NMSA 122 Mechara 8°36' 40°19' 1790 1968-1976 NMSA 123 Mekele 13°50' 39°48' 2121 1960-1991 FAO 124 Merab Abay 6°18' 37°47' 1260 1972-1987 NMSA 125 Merawe 11°25' 37°09' 2110 1981-1987 NMSA/MoWR 126 Mersa 11°40' 39°39' 2100 1962-1985 NMSA/MoWR 127 Meta Hara 8°52' 39°54' 930 1952-1974 NMSA 128 Metekel 10°57' 36°30' 1650 1973-1987 NMSA/MoWR

Rainfall and its erosivity in Ethiopia 207

Tab. 17: Continued

Weather Latitude Longitude Elevation Years of Data Nr. station (N) (E) (m) observation source 129 Midre Genet 6°55' 38°25' 1750 1978-1985 NM SA/M oW R 130 Mieso 9°14' 40°45' 1400 1962-1997 NMSA 131 Minneh 8°19' 40°10' 0 1952-1961 NMSA/MoWR 132 Muja 12°00' 39°18' 2820 1973-1982 NMSA/MoWR 133 Munessa 7°35' 38°54' 2510 1966-1975 NM SA/M oW R 134 Nedjo 9°30’ 35°27' 1800 1974-1992 NMSA 135 Nekemte 9°05' 36°27' 2080 1971-1981 NOAA 136 Rike 10°46' 39°52' 1749 1962-1987 NM SA/M oW R 137 Robe 7°08' 40°00' 2480 1960-1991 FAO 138 Robi 7° 15' 39037- 2400 1982-1990 NM SA/M oW R 139 Sagure 7°46' 39°09' 2480 1973-1983 NMSA/MoWR 140 Sedika 7°43' 39°40' 2545 1968-1985 NM SA/M oW R 141 Seru 7°40' 40°12' 2480 1968-1985 NMSA/MoWR 142 Shakiso 5°46' 38°55' 1620 1975-1985 NMSA/MoWR 143 Shemena 6°48' 38°15' 1600 1974-1985 NM SA/M oW R 144 Sheneka 7° 15' 40°00’ 2400 1960-1991 FAO 145 Sheno 9°20' 39°18' 2655 1962-1995 NMSA 146 Shola Gebeya 9°13' 39°27' 2650 1962-1996 NMSA 147 Sinana 7°40' 40°13' 2500 1980-1988 NM SA/M oW R 148 Sire 7° 13' 38°53' 2390 1953-1975 NM SA/M oW R 149 Sodo 6°50' 37°43' 2002 1970-1987 NMSA/MoWR 150 Tendaho (Dubti) 11°42' 40°48' 400 1960-1991 FAO 151 Tibila 8°27' 39°35' 0 1965-1975 NMSA/MoWR 152 Tulu Bolo 8°40' 38°13' 2100 1962-1996 NMSA 153 Turmu 5°06' 36°30' 1200 1980-1988 NMSA 154 Urgessa 11 °33' 39°37' 2000 1962-1985 NM SA/M oW R 155 Wadera 5°53' 39°14' 2000 1972-1985 NM SA/M oW R 156 Wogel Tena n ° 3 6 ' 39°13' 3000 1978-1987 NMSA/MoWR 157 Woldia 1 l°50' 39°37' 2010 1962-1987 NMSA/MoWR 158 Wonji 8°29' 39°I5' 1540 1951-1984 NOAA 159 Worancha 6°58' 38°05' I960 1975-1985 NM SA/M oW R 160 Worota 11°55' 37°41' 1980 1971-1985 NM SA/M oW R 161 Wuchale 11°32* 39035' 2000 1962-1985 NM SA/M oW R 162 Yetmen 10°20' 38°08' 2060 1960-1991 FAO 163 Yirba Dimbicho 6°52' 38°20' 2000 1973-1985 NMSA 164 Yirba Muda 6° 12' 38°43' 2050 1972-1985 NMSA/MoWR 165 Yirga Alem 6°45' 39°10' 1835 1956-1977 NMSA/MoWR 166 Yirga Cheffe 6° 19' 38°14' 1925 1966-1985 NM SA/M oW R 167 Yitnora 10° 12' 38°08' 2540 1960-1991 FAO 168 Zege 11°41' 37°19' 1800 1975-1987 NMSA/MoWR

Rainfall and its erosivity in Ethiopia 208

Tab. 18: Weather stations in FAO rainfall pattern regions of Ethiopia

Region A Region B

Nr. Weather station Ref. Nr." Nr. Weather station Ref. Nr." 1 Addis Zemen 4 1 Agaro 8 2 Adi Gudom 6 2 Areka 20 3 Adigrat 5 3 Arjo 21 4 Amed Ber 16 4 Asosa 26 5 Askuna school 25 5 Bele 35 5 Aykel 29 6Dembi Dolo 56 7 Bahir Dar 31 7 Ejaji 65 8 Birr Valley 37 8 Gambela 73 9 Dabat 47 9 Gobessa 80 10 Dambacha 48 10 Hosana 94 11 Dangla 49 11Jima 98 12 Debre Markos 50 12 Kachise 100 13 Debre Tabor 51 13 Nedjo 134 14 Debre Work 52 14 Nekemte 135 15 Deke Stefanos 54 15 Tulu Bolo 152 16 Ebnet 64 17 Elias 66 18 Enfranz 68 19 Finoteselam 71 20 Gohatsion 82 21 G onder 83 22 Gorgora 84 23 Gum ara 88 24 Kunzela 115 25 M aksenget 118 26 M andura 119 27 M ankusa 120 28 M ekele 123 29 M erawe 125 30 Metekel 128 31 W orota 160 32 Yetmen 162 33 Yitnora 167 34 Zege 168

H Reference Number, see station numbering in Tab. 17.

Rainfall and its erosivity in Ethiopia 209

Tab. 18: Continued

Region C Region D

Nr. Weather station Ref. Nr." Nr. Weather station Ref. Nr." 1 Adaba 2 I Bacco 30 2 Agarfa 7 2 Burji 41 3 Alemaya 12 3 Delo Sebro 55 4 Aleta Wondo 14 4 Dhagahbour 58 5 Amaro Kelo 15 5 Gato 74 6 Arba Minch 19 6 Gidole 77 7 Bedeno 33 7 Ginir 78 8 Bedessa 34 8 Gode 81 9 Chelelektu 44 9 Goro 85 10 Chencha 45 10 Gumadaye 87 11 Dila 60 11 Hagere Mariam 90 12 Dinsho 61 12 Jinka 99 13 Dodola 63 13 Kebri Dahar 101 14 Felege Neway 69 14 Key Afer 105 15 Galamso 72 15 Kibre Mengist 106 16 Gerese 75 16 Kolme 109 17 Gessera 76 17 Konso 111 18 Goba 79 18 Shakiso 142 19 Gravva 86 19 Turm u 153 20 Gursum 89 20 Wadera 155 21 Hagere Selam 91 2 2 Hebeno 93 #Reference Number, see station numbering in Tab. 17. 23 Idoto 96 24 Jijiga 97 25 Kemba 104 26 M echara 122 27 Merab Abay 124 28 Minneh 131 29 Robe 137 30 Sedika 140 31 Seru 141 32 Sheneka 144 33 Sinana 147 34 Sire 148 35 Sodo 149 36 Yirba Dimbicho 163 37 Yirba Muda 164 38 Yirga Alem 165 39 Yirga Cheffe 166

Rainfall and its erosivity in Ethiopia 210

Tab. 18: Continued

Region E

Nr. Weather station Ref. Nr.* Nr. Weatherstation Ref. N r.# 1 Abella 1 31 Keleta 103 2 Addis Ababa 2 32 Kobo 107 3 Alaba Kulito 9 33 Kofole 108 4 Alam ata 10 34 Kombolcha 1 10 5 Albuco 11 35 Kore 1 12 6 Alem Ketema 13 36 Korem 1 13 7 A ncharo 17 37 Kule Meskel 1 14 8 Ankober 18 38 Lalibala 1 16 9 A rtum a 22 39 Majete 1 17 10 Asasa 23 40 Mararo 121 11 Asela 24 41 Mersa 126 12 Assita 27 42 M eta Hara 127 13 A w assa 28 43 Midre Genet 129 14 Bati 32 44 Mieso 130 15 Bilate Farm 28 45 Muja 132 16 Bokoksa 38 46 M unessa 133 17 Bora 39 47 Rike 136 18 Boru Meda 40 48 Robi 138 19 Butajira 42 49 Sagure 139 20 Cheffa 43 50 Shemena 143 21 Chiro 46 51 Sheno 145 22 Debre Zeit 53 52 Shola Gebeya 146 23 Dessie 57 53 Tendaho (Dubti) 150 24 Dhera 59 54 Tibila 151 25 Dire Dawa 62 55 Urgessa 154 26 Elidar 67 56 Wogel Tena 156 27 Fiche 70 57 W oldia 157 28 Haik 92 58 Wonji 158 29 Huruta 95 59 Worancha 159 30 Kelem Meda 102 60 W uchale 161

# Reference Number, see station numbering in Tab. 17.

Rainfall and its erosivity in Ethiopia 1.83 1.17 0.44 0.33 0.24 2.58 - 1.00 -1.34 -0.75 Skewness 1.40 -0.90 1.93 1.10 -0.31 3.00 1.00 0.90 4.00 6.94 2.89 -1.62 2.00 0.61 3.74 1.82 - 1.00-0.83 -0.90 -1.50 0.30 -0.02 10.86 3.01 Kurtosis 15.52 1.8516.67 -0.30 -0.51 -0.33 16.79 20.74 0.20 -0.65 28.37 26.34 -0.51 -0.20 59.51 31.29 68.72 29.84 5.62 49.08 49.56 35.84 33.56 -1.53 46.10 0.00 -0.31 33.24 47.74 -0.60 -0.40 24.50 26.23 0.40 49.05 -0.55 -0.10 Variability coefficient (%) 190.17 184.80 343.85 335.95 299.79 456.00 55.31 203.03 485.87 1152.50 Standard deviation (mm) 901.10 234.83 1679.00 290.29 1125.00 188.70 23.97 1066.501224.60 135.00 1936.70 2377.30 417.92 22.39 (mm) Maximum 0.00 85.60 1420.80 26.60 1664.30 521.08 57.00 1559.00 325.02 37.25 40.40 1951.90 546.70 123.00 926.00 162.00 192.00 2873.30 612.73 56.63 582.10 1353.80 285.80 321.60 3300.00 623.90 927.20 3226.20 478.60 576.50277.30 1744.90 1004.70 270.10 302.10 841.80 934.00 2049.00 241.28 (mm) Minimum 775.70 526.60 858.00 827.00 526.00 534.70 912.50 (mm) 1422.80 281.40 3358.30 785.40 1871.60 1047.10 1339.05 899.40 4464.20 1748.50 957.10 2182.60 1088.70 1091.00 Median 810.00 650.24 650.24 913.41 Mean (mm) 1657.91 1584.70 1 102.201 1115.40 563.60 1677.01 1555.00 1557.50 1355.58 1322.40 477.20 2496.50 8 824.50 1022.50 115.40 1240.00 17 571.00 592.00 18 18 1048.35 862.55 N 30 1102.47 98 1208.95 1176.20 904.00 Station Weather Asasa Askuna School 8 1866.80 Amed Ber 8 Areka 12 1061.60 1294.05 Arba Minch 19 787.10 Amaro Kelo 9 Addis Zemen 13 AgarfaAgaro 22 1082.00 963.05 1 Abella 5 Adigrat 13 550.58 7 8 3 Addis Ababa 6 Adi Gudom9 18 Alaba Kulito 478.73 462.15 25 872.53 37.60 2 Adaba 20 725.01 753.20 4 17 Ancharo 19 18 Ankober 22 1603.65 1507.25 12 Alemaya 36 13 Alem Ketema 2515 1010.72 14 Aleta 16 Wondo 18 1145.10 1144.20 10 Alamata11 Albuco 35 21 Arjo25 12 24 Asela 23 20 22 Artuma23 10 Nr. Tab. 19: Statistical summary oflong-term annual rainfall at selected meteorological stations in Ethiopia

Rainfall and its erosivity’ in Ethiopia 212 1.00 1.60 2.22 2.70 -0.21 -0.40 -0.80 2.80 -0.30 0.03 -0.60 0.54 0.90 0.70 -0.90 -0.09 -0.63 0.40 -0.75 -0.32 -0.53 -0.30 10.00 2.80 - 1.00 -0.40 -0.20 -0.40 -0.70 Kurtosis Skewness 19.75 39.8476.26 0.84 0.20 1.00 43.00 5.83 1.97 45.65 61.81 2.6020.57 1.72 39.17 31.89 0.83 - 1.00 32.81 0.60 -0.90 33.17 6.72 24.99 72.38 9.42 35.95 Variability coefficient (%) 86.70 57.53 -0.80 0.50 172.15 463.40 977.04 328.03 597.41 262.41 246.50440.29 171.97 0.30 0.30 334.64 427.97 Standard deviation (mm) 1873.50 330.50 31.61 4.00 1495.60 453.48 50.14 1847.70 470.20 48.11 1767.50 154.83 11.60 (mm) Maximum 65.50 1976.80 120.80 3175.40 680.04 367.60 2147.10 402.34 601.00 4778.80 697.80 1479.00 220.80 488.00 1320.00 232.00 24.71 (mm) 1057.30 Minimum 859.60 247.50 1257.30 866.40807.87 189.50 314.80 1273.70 201.73 (mm) 932.00 1432.05 894.60 2035.00 1042.20 702.20 1066.45 134.60 1470.80 1210.10 765.10 2545.20 1153.00 92.10 Median 143.34 935.66 905.70 783.40 586.50 807.14 Mean 939.00 (mm) 904.35 939.56 854.70 1013.80 1007.45 35.00 1764.50 556.47 54.89 1019.84 1117.75 1150.80 1290.34 1389.40 1372.20 864.70 2293.50 323.73 23.30 0.82 0.62 1334.50 1326.60 8 844.10 872.40 700.70 990.60 117.30 13.90 -2.00 N 19 15 977.40 15 36 32 837.00 849.80 310.60 1285.20 36 32 Station W eather Bahir Dar BedenoBedessa 20 1105.28 26 998.90 837.38 754.30 526.40 330.60 2219.50 1565.20 Bati Asosa 15 Birr Valley Bacco 12 Dambacha Dabat Dangla 15 1229.60 142.50 445.90 1794.20 442.00 Tab. Continued 19: Nr. 26 27 Assita28 Awassa 22 16 150.70 1015.10 133.30 935.9035 Bele 394.70 14.40 2594.10 310.60 12 30 31 33 34 36 Bilate Farm 18 609.70 646.50 57.50 1051.90 263.80 43.27 32 29 Aykel 32 1580.82 1167.90 37 38 Bokoksa39 Bora 11 1045.60 13 815.60 892.00 333.60 1045.50 227.40 27.88 0.10 -1.00 50 Debre Markos 38 51 Debre Tabor 41 Burji42 Butajira 20 30 822.90 40 Boru Meda 43 Cheffa44 Chelelektu45 Chencha46 Chiro 23 13 48 13 47 49

Rainfall and its erosivity in Ethiopia 1.00 1.80 0.12 0.15 -0.24 -0.66 -0.20 - 1.10 -0.88 -0.54 1.12 1.00 1.40 3.84 0.50 -0.60 2.30 1.14 0.85 3.720.31 1.73 3.91 4.72 1.80 0.06 -0.09 -0.78 -1.80 0.45 Kurtosis Skewness 14.65 0.80 0.93 13.25 25.54 0.14 -0.36 31.6234.08 1.42 0.54 27.51 -0.05 -0.23 44.83 45.46 -0.3120.57 33.19 -0.30 0.9573.32 -1.50 -0.90 39.53 2.40 -1.21 65.01 34.25 43.76 -0.32 0.80 Variability coefficient (%) 86.00 127.14 193.72 31.41 189.40 123.90 12.87 232.57 501.30 44.98319.65 9.40 281.65 22.26 451.32 937.87 1087.65 59.69 Standard deviation (mm) 764.70 1073.10 1308.001302.80 143.70 1460.40 201.10 18.17 1596.00 1722.00 1875.90 436.58 (mm) 2383.20 432.553201.80 35.38 4821.80 4120.90 2306.00 453.60 26.64 Maximum 3.70 156.00 1177.50 187.84 354.00 215.60 443.70 210.80 281.13 (mm) Minimum 603.00 930.10995.10 523.00 139.00 1102.70 179.17 920.50 831.20 1141.10 (mm) 1109.00 632.50 1453.40 1129.10 1748.80 1317.55 1115.85 1061.60 24.20 1581.50 353.06 36.04 1041.10 0.00 1990.00 354.94 Median 845.51 850.45 360.50 283.58 283.58 6.30 809.13 809.13 92.50 1433.90 367.81 594.02 871.12 682.00 677.00 184.90 1165.90 232.40 997.71 935.40 438.50 979.70 962.94 Mean (mm) 1222.57 1107.00 1110.03 1140.40 640.20 1499.50 283.50 1442.56 8 13 981.20 999.30 824.40 18 19 15 11 117.30 112.00 1.50 235.00 16 15 15 1702.51 1857.80 652.40 N 33 1114.48 1154.50 26 21 616.80 619.10 44 Station Weather Dinsho Delo Sebro Dessie Deke Stefanos 13 1822.10 1743.30 Ebnet Finoteselam Felege Neway Gato Gerese 58 Dhagahbour59 Dhera 44 52 Debre53 Work Debre Zeit55 56 Dembi 34 57 Dolo 7 54 61 Tab. 19: Continued Nr. 60 Dila 74 70 Fiche 71 72 Galamso 27 1036.38 63 Dodola 17 62 Dire Dawa64 65 Ejaji 39 66 Elias67 Elidar 32 962.99 71429.54 68 Enfranz69 73 8 1141.62 Gambela 1169.20 75 76 1265.30 1262.00 511.00 1908.00

Rainfall and its erosivity in Ethiopia Tab. 19: Continued

Weather Mean Median Minimum Maximum Standard Variability Nr. N Kurtosis Skewness Station (mm) (mm) (mm) (mm) deviation (mm) coefficient (%) 76 Gessera 19 1225.57 1 108.40 492.70 2322.20 467.02 38.11 0.13 0.71 77 Gidole 15 1014.55 1014.55 192.00 1723.50 331.90 32.71 3.00 -0.40 78 Ginir 20 1089.17 952.40 401.90 2534.80 537.90 49.39 1.45 1.20 79 Goba 23 916.51 923.30 564.60 1279.20 197.17 21.51 -0.37 -0.08 80 Gobessa 17 1105.95 1097.70 702.30 1436.20 212.70 19.23 -0.70 -0.20 81 Gode 15 293.67 254.00 0.00 755.00 221.31 75.36 0.15 0.76 82 Gohatsion 26 1118.60 1060.70 885.00 1474.80 163.06 14.58 -0.58 0.58 83 Gonder 37 1063.65 1016.00 688.80 1828.10 237.80 22.36 2.03 1.20 84 Gorgora 14 911.80 914.50 367.60 1500.80 276.52 30.33 0.97 0.16 85 Goro 14 772.44 793.40 242.10 1093.60 265.86 34.42 -0.24 -0.72 86 Grawa 30 782.30 782.30 116.00 2568.40 545.67 69.75 2.82 1.30 87 Gumadaye 12 837.00 877.65 0.00 1369.40 356.30 42.57 1.80 - 1.00 88 Gumara 7 1090.50 1114.10 272.70 1622.60 409.81 37.58 3.40 -1.30 89 Gursum 19 740.31 680.60 137.90 1353.50 370.34 50.02 -0.84 0.40 90 Hagere Mariam 28 674.42 674.42 94.00 1202.10 284.77 42.22 -0.30 -0.40 91 Hagere Selam 24 1061.15 1067.60 21.00 1987.40 512.64 48.31 -0.13 -0.34 92 Haik 23 1135.00 1157.40 53.00 1641.30 314.10 27.67 6.00 -1.70 93 Hebeno 19 1881.11 1537.90 115.60 4376.90 1013.29 53.87 0.92 0.94 94 Hosana 16 1048.06 1115.50 215.00 1469.00 313.54 29.92 2.34 -1.42 95 Huruta 14 617.96 625.60 237.20 1014.50 261.71 42.35 -1.14 -0.15 96 Idoto 17 810.80 857.70 455.30 1174.10 188.87 23.29 -0.20 -0.30 97 Jijiga 33 711.49 628.20 349.80 1825.00 327.01 45.96 2.87 1.50 98 Jima 39 1291.03 1377.00 0.00 2011.00 431.91 33.45 3.54 -1.84 99 Jinka 19 1275.91 1262.50 822.60 1799.20 296.44 23.23 -0.95 0.09 100 Kachise 11 1470.01 1577.70 764.30 1998.80 409.89 27.88 -0.23 -0.73 215 1.40 1.29 0.11 0.05 0.95 - 1.00 -0.70 -0.43 -0.30 Skewness 1.21 -1.09 1.60 1.20 -1.30 2.10 1.20 3.00 - 1.00 5.00 -2.00 2.70 - 1.10 -0.20 -0.63 -0.20 -0.42 -0.30 -0.02 -0.40 -0.50 -0.30 0.12 -0.08 -0.40 Kurtosis 17.44 12.67 41.77 29.27 -0.54 46.71 27.93 -0.76 27.9528.87 2.00 -1.04 40.43 2.00 -1.50 36.28 35.95 - 1.00 0.00 23.4839.31 0.00 1.60 -0.10 54.25 3.60 1.26 22.92 45.76 28.22 43.83 3.00 -1.30 20.97 -0.50 Variability coefficient (%) 178.78 59.35 175.80 40.26 139.23 15.94 177.00 320.20 501.63 220.00 31.79 386.20 296.50 310.70 309.00 233.63 398.00 235.81 350.65 250.44 Standard deviation (mm) 853.70 872.00 162.16 1198.10 1008.00 1809.70 1269.00 274.90 1115.30 1879.00 1277.40 133.71 1425.30 228.78 1123.10 222.22 1346.40 1612.60 1558.20 176.95 12.54 (mm) Maximum 15.40 847.10 72.30 1074.80 142.00 1540.00 784.00 3575.40 721.82 363.60 989.60 153.27 20.89 225.00 1329.00 284.00 658.10 1868.70 (mm) 1075.50 Minimum 751.40 243.60 800.00 445.90 882.00 563.00 276.35 0.00 857.20 968.60 423.30 857.70 398.30 997.10 513.70 1436.30 735.15 421.60 134.40 931.30 124.10 1421.00 (m m ) 1016.00 1074.20 866.20 1063.35 604.00 1297.00 Median 1124.50 697.10 1683.70 301.25 766.50 692.00 743.00 0.00 580.51 563.50 311.50 533.00 524.20 Mean 733.64 (mm) 1073.96 1157.40 0.00 1 107.001 1015.00 631.00 1004.39 9 1091.10 10 733.4112 939.34 826.80 N 16 646.33 617.30 1.80 Weather r‘ r‘ Station 102 Kelem Meda 9 104 Kemba 14 101 101 Kebri Dahar 40 109 Kolme110 Kombolcha 40 10 1015.00 769.82 103 103 Keleta107 Kobo 25 108 Kofole 32 27 873.30 1016.00 105 105 K eyA fer 106 Kibre Mengist 120 Mankusa122 Mechara 8 856.51 114 Kule Meskel 15 118 Maksenget119 Mandura 16 16 1836.04 11 1 1 11 Konso112 Kore113 Korem 17 436.62 17 11 1095.00 1134.00 844.00 121 121 Mararo 21 115 Kunzela 8 1055.04 123 Mekele124 Merab Abay 125 Merawe 32 7 1410.60 1482.40 116 Lalibala117 Majete 9 35 1092.57 1086.50 Tab. 19: Continued

Rainfall and its erosivity in Ethiopia 1.15 1.58 1.00 1.14 0.84 0.87 0.00 0.32 0.54 0.40 2.00 -0.30 -0.04 -0.40 -0.65 -0.40 -0.06 -2.00 - 1.20 Skewness 1.60 -0.20 4.94 5.00 3.00 4.00 2.00 -0.80 -0.80 0.03 -0.30 0.50 -0.21 -2.00 - 1.00 0.00 Kurtosis 9.91 11.18 2.32 12.17 28.85 35.30 -0.70 21.85 1.20 59.26 -0.80 39.02 1.50 28.29 -0.90 29.05 58.19 -0.58 22.06 -0.30 0.50 35.55 34.73 29.32 2.73 30.10 2.40 65.56 20.54 0.80 Variability coefficient (%) 94.60 180.00 61.43 0.90 188.00 21.73 176.44 19.35 199.24 310.00 47.55 1.00 696.90 47.59 0.97 204.52 272.95 528.85 531.00 723.73 218.00498.00 22.31 263.64265.75 268.68 34.76 335.60 210.90 Standard deviation (mm) 770.00 992.00 257.00 1035.10 1430.00 1479.20 1748.80 302.56 1401.001983.40 289.00 1384.90 1599.00 1693.50 1350.70 198.81 1608.00 2526.80 2942.00 2030.90 472.31 (mm) Maximum 0.00 2586.70 13.00 866.00 257.38 55.00 190.00 105.00 307.60 3085.00 585.40576.20 1870.00 899.70 385.30 698.00 613.70 1263.80 453.20 1944.50 (mm) 1331.20 Minimum 138.45 738.00 760.00 50.00 920.20 451.20 971.90 571.50 (mm) 1398.00 787.00 1447.95 M edian 2124.00 1759.00 2478.00 846.00 840.00 705.50 652.00 911.77936.23 893.45 758.44 752.30 2.10 939.50 936.20 437.00 1271.10 Mean (mm) 1210.34 1003.70 1637.54 1649.00 1307.80 1319.05 757.30 8 9 865.00 934.00 610.00 1096.00 17 12 1497.97 10 810.00 870.00 15 18 1243.81 1478.55 N 32 34 24 960.00 988.00 95.00 Station 150 Tendaho (Dubti) 32 293.00 277.00 148 Sire 149 Sodo 18 1464.37 147 Sinana 8 146 Shola Gebeya 35 946.12 923.10 144 Sheneka 145 Sheno 140 Sedika 141 Seru 18 977.00 18 955.00 1434.00 142 Shakiso II 1031.86 138 Robi 139 Sagure 11 723.00 143 Shemena 128 Metekel 129 Midre Genet 130 Mieso 131 Minneh 132 Muja133 Munessa 36 10 10 134 Nedjo 137 Robe 32 901.13 867.65 135 Nekemte136 Rike 11 2128.73 26 1155.10 14.60 11 127 Meta Hara 23 434.31 517.00 126 Mersa Tab. 19: Continued ^ Weather

Rainfall and its erosivity in Ethiopia 217 0.50 0.26 0.51 -0.03 -0.41 -0.54 -0.62 - 1.00 Skewness 1.34 0.48 6.20 -2.00 0.80 0.60 0.60 0.65 -0.68-0.50 -0.32 -0.33 - 1.00 Kurtosis 15.82 1.50 0.74 15.43 -0.29 33.49 37.21 2.73 -1.60 51.10 -0.92 -0.33 55.35 35.30 - 1.10 23.74 0.6031.76 0.60 30.04 46.97 4.20 63.64 47.28 1.00 Variability coefficient (%) 190.13 184.60180.61 15.37 341.82 38.63 -0.20 -0.60 601.20 331.72 316.71 328.00 237.64 31.16 1.03 356.00 311.02 29.25 243.00 426.00 Standard deviation (mm) 855.70 281.38 1410.00 1740.80 1303.901552.00 402.26 251.30 21.95 1426.40 1657.10 2066.40 450.93 3511.80 994.25 2389.30 (mm) Maximum 7.10 2407.80 3.30 140.50 143.00 181.00 1 103.00 1333.00 856.90 1478.20 907.00 1785.00 544.00 628.40 904.10 1725.30 468.30 512.90 1912.90 479.30 405.80 1657.60 (mm) Minimum 538.60 27.50 (mm) 1425.80 58.10 1183.20 189.00 1 1003.30 Median 897.26 1002.25 442.17 Mean (mm) 1346.42 1200.80 1145.00 1063.42 1027.50 9 13 1170.7614 1124.10 1945.74 2153.90 243.70 14 32 1202.10 1188.30 20 1615.60 1737.95 32 N 24 26 1023.60 1000.00 24 1418.14 1388.30 Station Weather 167 Yitnora 168 Zege 13 166 Yirga Cheffe 165 165 Yirga Alem 22 884.93 884.93 162 Yetmen 163 Yirba Dimbicho 164 Yirba Muda 161 161 Wuchale 156 Woldia 158 Wonji160 Worota 34 762.62 15 741.50 726.73 726.73 155 155 Wadera 159 Worancha 1 1 1044.30 157 Wogel Tena 10 698.30 691.00 57.10 153 Turrnu 151 151 Tibila152 Tulu Bolo 35 11 753.00 822.00 154 Urgessa Tab. Continued 19:

Rainfall and its erosivity in Ethiopia Tab. 20: Statistical summary of long-term annual rainfall erosivity factors for selected meteorological stations in Ethiopia

W eather Mean M edian M inim um Maximum Standard Variability Nr. N Kurtosis Skewness Station (mm) (mm) (mm) (mm) deviation (mm) coefficient (%) 1 Abella 30 145.89 135.20 96.50 220.94 35.26 24.17 -0.64 0.76 2 Adaba 20 132.02 132.05 74.96 181.57 27.35 20.72 -0.10 -0.16 3 Addis Ababa 98 207.64 205.43 132.74 370.56 40.92 19.71 2.94 1.19 4 Addis Zemen 13 335.86 309.20 208.00 545.22 97.20 28.94 0.30 0.70 5 Adigrat 13 116.25 103.95 44.60 189.90 47.50 40.86 -0.96 0.26 6 Adi Gudom 18 147.95 10.41 10.41 304.91 77.67 52.50 -0.25 -0.04 7 Agarfa 20 149.63 129.49 16.07 444.27 97.80 65.36 3.33 1.98 8 Agaro 18 206.37 209.88 151.33 281.40 33.78 16.37 0.43 0.23 9 Alaba Kulito 25 130.00 124.15 55.04 287.75 40.86 31.43 9.73 2.53 10 Alamata 35 140.80 136.15 80.36 221.16 33.00 23.44 -0.14 0.44 11 Albuco 18223.52 192.01 68.34 820.97 160.45 71.78 12.62 3.30 12 Alemaya 36 110.30 104.41 7.00 265.42 43.08 39.06 4.28 1.15 13 Alem Ketema 25 243.28 240.09 51.83 385.94 80.92 33.26 1.00 -0.77 14 A leta W ondo 18 158.62 172.64 0.00 241.84 63.33 39.93 1.30 - 1.20 16 A m aro Kelo 9 141.58 130.24 106.76 206.97 31.60 22.32 1.05 1.10 15 Amed Ber 8 364.25 275.04 230.00 952.58 241.00 66.16 7.35 2.68 17 Ancharo 8 184.23 217.01 44.00 257.70 80.00 43.42 -0.36 -1.03 18 Ankober 21 252.47 248.33 129.66 394.88 58.84 23.31 1.13 0.54 19 Arba Minch 19 114.90 110.30 74.40 150.55 23.40 20.37 -1.15 -0.02 20 Areka 11 161.94 177.93 26.60 204.87 49.24 30.41 6.50 -2.42 21 Arjo 12251.13 247.89 184.63 329.34 37.98 15.12 0.69 0.33 22 Artuma 10252.61 238.00 137.25 437.36 80.10 31.71 2.91 1.26 23 Asasa 17106.00 96.00 53.00 203.00 34.00 32.08 3.00 1.00 24 Asela 23 176.60 170.62 112.23 228.42 32.03 18.14 -0.63 -0.01 25 Askuna School 8 391.00 404.15 325.43 458.32 52.23 13.36 -1.70 -0.22 Tab. 20: Continued

W eather Mean Median M inimum M aximum Standard Variability Nr. N Kurtosis Skewness Station (mm) (mm) (mm) (mm) deviation (mm) coefficient (%) 26 Asosa 15 197.75 175.46 121.31 422.29 74.80 37.83 5.56 2.20 27 Assita 22 43.91 42.55 7.51 78.45 21.30 48.51 -1.14 0.20 28 Awassa 16 175.08 121.70 98.84 989.89 217.76 124.38 15.83 3.96 29 Aykel 32299.17 232.48 156.50 825.60 152.90 51.11 3.60 1.83 30 Bacco 12 176.75 180.02 35.00 355.23 81.22 45.95 1.33 0.53 31 Bahir Dar 36 317.13 305.10 187.72 425.67 64.41 20.31 -0.95 -0.01 32 Bati 32 148.42 143.20 60.93 280.87 40.07 27.00 3.44 1.00 33 Bedeno 20 191.00 180.76 108.53 311.31 51.34 26.88 -0.20 0.53 34 Bedessa 26 153.43 145.33 80.00 268.37 47.52 30.97 0.16 0.80 35 Bele 12 151.50 106.50 33.13 468.55 126.50 83.50 3.00 1.76 36 Bilate Farm 18 112.02 100.71 13.64 347.92 68.67 61.30 8.53 2.45 37 Birr Valley 8 149.35 150.88 111.12 176.05 21.45 14.36 0.17 -0.50 38 Bokoksa 11 184.83 166.30 90.88 425.33 87.01 47.08 6.90 2.40 39 Bora 13 149.07 147.65 96.00 247.43 39.40 26.43 2.31 1.21 40 Boru Meda 20 206.00 207.00 103.41 366.60 59.00 28.64 1.82 0.74 41 Burji 20 143.05 143.74 97.69 193.51 28.38 19.84 -0.96 0.11 42 Butajira 29 157.51 157.32 83.93 210.74 32.48 20.62 -0.23 -0.45 43 Cheffa 22 174.00 175.00 123.06 233.73 34.59 19.88 -1.04 0.33 44 Chelelektu 12 207.24 175.59 129.89 587.34 122.29 59.01 10.73 3.21 45 Chencha 13 174.80 175.88 99.23 288.67 46.88 26.82 2.22 0.88 46 Chiro 36 125.00 132.04 0.00 235.20 47.00 37.60 1.80 -0.70 47 Dabat 15 204.70 173.90 62.56 576.56 121.11 59.16 6.50 2.30 48 Dambacha 15 223.54 252.91 46.10 292.57 69.03 30.88 2.07 -1.57 49 Dangla 15 245.47 258.84 109.05 350.52 67.70 27.58 0.64 -0.70 50 Debre Markos 38 224.67 219.60 159.71 292.72 32.45 14.44 0.05 0.60 1.23 1.78 1.20 2.70 0.78 0.04 2.56 0.31 5.20 0.29 3.13 0.84 0.17 0.40 -0.20 -0.19 - 1.21 -0.70 -0.03 1.44 1.30 8.02 0.33 0.41 9.12 0.83 -0.58 -0.38 -0.20 - 1.10 -0.20 -0.52 -0.38 -0.36 -1.15 -0.60 0.62 11.62 27.73 Kurtosis Skewness 30.65 81.67 38.64 27.73 -0.91 23.16 66.07 26.76 0.30 22.39 21.4321.31 57.40 0.63 36.96 -0.10 1.50 0.01 -0.89 39.49 65.63 1.84 24.8639.79 -0.34 36.23 Variability 78.80 17.22 54.70 52.82 30.77 23.1860.78 13.07 31.60 38.01 32.77 2.78 29.16 24.92 -0.57 51.62 43.46 33.09 27.0549.04 14.80 139.66 106.13 41.30 246.49 129.34 deviation (mm) coefficient (%) 57.26 190.40 168.48 184.75 310.10 304.46 414.45 316.83 67.94 297.35 47.08 251.13 280.26 46.70 306.54 428.76453.92 225.55 108.44 278.98 63.32 1734.50 406.74 1459.27 (mm) Maximum Standard 1.50 3.28 353.00 85.94 69.38 59.06 35.43 77.80 236.48 36.74 61.43 99.79 556.21 79.02 94.13 190.00 143.52 211.58 108.00 186.10 316.25 115.51 106.00 437.60 66.00 Minimum 75.32 168.43 125.30 149.19 197.65 119.96 146.83 3.70 156.14 186.00 150.00 383.50 299.60 79.94 238.95 76.27 466.78 93.14 92.00 30.00 27.00 (mm) (mm) (mm) M ean M edian 171.66 173.72 175.81 177.35 180.88 118.08 197.23 197.07 116.00 110.70 498.03 208.60 216.20 242.25 215.45 256.98 262.71 5.60 13 178.48 13 N 16 131.35 142.83 18 117.03 118.25 77.14 15 11 16 212.81 15 257.07 34 32 284.13 285.00 136.14 21 29 190.57 39 32 171.45 171.45 44 Station W eather Deke Stefanos Dessie 26 203.28 DhagahbourDhera 37 Ebnet Fiche Gambela 76 Elias 7 Finoteselam 8 182.73 51 Debre Tabor 52 Debre 53Work Debre54 Zeit 55 Delo Sebro 17 56 Dembi Dolo 7 57 58 59 60 Dila 61 D insho 62 Dire Dawa 63 Dodola64 72 Galamso73 26 167.11 154.06 65 Ejaji 66 67 Elidar 70 71 74 Gato75 Gerese 15 159.12 144.74 68 Enfranz69 Felege Neway 8293.40 Nr. Tab. 20: Continued

Rainfall and its erosivity in Ethiopia 1.07 1.04 0.46 0.20 0.26 0.43 0.52 0.50 -1.42 -0.60 -0.24 1.44 1.30 0.60 5.92 2.15 4.20 1.20 -0.44 0.75 -0.28 - 1.21 -1.51-0.80 0.20 -0.50 Kurtosis Skewness 17.13 18.78 19.15 16.82 -0.18 18.62 0.32 16.5119.01 3.03 1.21 30.6058.38 1.50 39.05 -0.15 0.56 35.16 0.55 0.07 65.69 27.46 - 1.10 0.17 38.58 1.03 1.13 40.9345.05 3.11 18.91 19.79 71.09 38.20 24.87 0.02 52.2181.80 24.82 4.40 1.80 79.44 58.64 30.04 52.80 60.5063.87 63.68 60.64 43.67 105.87 107.32 48.80 -0.17 Standard Variability deviation (mm) coefficient (%) 158.01 182.15 142.88 379.97 366.54 225.42 41.05 28.51 -0.30 0.47 245.00 275.83 255.25 454.69 203.45 250.00 43.10 31.99 3.52 1.40 404.51 (m m) M axim um 0.00 51.17 242.62 45.92 32.52 56.74 73.77 81.15 349.56 95.5046.56 220.93 384.49 41.33 148.12 112.37 201.82 28.00 103.98 222.68 28.83 128.54 125.22 366.30 (mm) M inimum 141.67 130.94 78.26 143.19 157.98 124.90 102.72 55.45 155.44 196.00 (mm) 265.40 204.62 277.40 213.22 203.04 M edian Mean (mm) 112.40 857.70 140.11 151.00 134.60 116.15 135.46 132.98 0.00 394.30 182.06 153.60 155.94 67.00 154.82 150.50 220.46 214.05 154.92 297.64 42.22 219.94 7 258.63 17 18 14 208.70 14 141.19 133.40 11 19 14 134.71 127.26 66.00 14 95.00 83.41 10.00 N 23210.76 182.46 27.00 697.65 118.16 56.06 14.15 3.30 33 113.20 37 181.92 178.38 24 37 210.32 Station W eather Huruta Kachise 11 277.80 Gorgora Grawa 30 Gum ara Ginir 20230.50 167.60 77.93 737.78 151.42 Jinka 19 76 Gessera 78 84 85 Goro 86 89 Gursum 19 77 Gidole79 Goba80 Gobessa81 Gode 1382 144.00 Gohatsion83 Gonder 23 17 115.55 149.12 113.90 26 87 Gum adaye 81.3988 95 96 Idoto 97 Jijiga 98 Jima 99 90 Hagere91 Mariam Hagere 92Selam Haik 24 140.89 136.00 21.00 232.70 49.53 93 Hebeno 94 Hosana 16 Nr. 100 Tab. 20: Continued

Rainfall and its erosivity in Ethiopia 222 1.44 1.21 0.40 0.37 0.62 0.33 0.00 0.95 0.40 0.55 -0.31 -0.81 -1.30 -0.34 1.30 3.24 1.20 0.30 5.00 -2.00 0.14 0.00 0.60 1.20 0.00 1.00 0.72 -0.30 -1.94 - 1.00 0.00 -0.26 Kurtosis Skewness 13.04 13.91 3.53 53.73 39.30 34.30 3.00 20.34 27.9030.49 -0.86 0.12 25.90 23.36 37.04 55.71 58.33 35.47 0.31 -0.31 50.10 26.79 -0.10 0.50 39.37 70.58 67.04 42.61 35.00 22.55 3.23 45.31 31.22 2.00 44.13 Standard Variability deviation (mm) coefficient (%) 180.79 175.00 18.00 725.66 157.43 40.89 0.55 1.31 239.13 244.54 37.00 24.62 -0.15 362.60 379.00 205.65 222.18300.90223.64 64.04 54.83 44.67 39.60 22.64 1.80 -1.17 279.75 33.88 14.61 -0.50 249.91 47.07 285.68 (m m) M aximum 1.80 276.10 74.00 63.12 80.00 169.00 24.00 87.00 48.76 253.13 61.36 276.4345.27 290.8898.00 55.53 60.85 128.13 109.41 103.74 104.40 277.23 47.00 27.13 100.00 321.16 175.55 204.35 233.90 (mm) Minimum 50.55 32.70 105.45 22.72 37.96 155.32 118.00 156.32 143.10 182.03 95.22 (mm) 137.15 66.93 328.72 223.70 202.00 234.17 278.15 M edian Mean (mm) 150.26 164.43 153.90 112.30 112.20 42.75 143.37 140.57 15.40 161.91 150.07 155.24 157.04 181.75 162.00173.22 153.00 169.00 81.00 281.00 60.00 7 283.09 8 174.91 8 231.96 9 13 10134.36 124.55 80.51 202.78 50.07 37.27 -1.80 0.30 N 15 16 117.24 95.37 10 25 118.00 40 103.68 94.51 0.00 27 138.00 137.00 35 196.55 21 Station W eather Kemba LalibalaM ankusa 9 Kule Meskel Mekele 32 145.12 M araro Kobo 32 Kore 17 137.00 145.00 38.00 185.00 32.00 Kolme Kombolcha 40 187.03 182.62 M erab Abay M ajete Nr. 101 Kebri 102Dahar Kelem M eda 9 157.34 108 Kofole 109 104 105 Key Afer 114 117 123 103 Keleta 107 111 Konso112 113 Korem 13 59.86 11 122 M echara 110 124 118 M aksenget 16 231.50 121 106 Kibre Mengist 12 115 Kunzela 116 119 M andura120 16 385.00 125 M erawe Tab. 20: Continued

Rainfall and its erosivity in Ethiopia 1.15 1.09 1.10 0.97 3.34 0.64 -0.08 -0.72 1.38 1.10 1.80 1.35 1.00 - 1.00 2.40 0.10 0.13 0.67 0.13 -0.17 0.10 0.00 1.00 -0.10 0.45 -0.62 0.70 Kurtosis Skewness 9.64 16.06 42.67 0.41 38.85 -1.00 0.00 24.72 33.60 15.13 31.78 1.85 0.90 28.7523.28 0.55 5.17 1.88 13.03 10.00 2.66 1.58 38.80 24.49 76.00 32.70 -1.14 39.41 51.70 31.23 47.07 43.00 32.26 52.3031.20 19.59 22.23 12.62 60.0026.0035.00 29.02 79.00 21.49 30.1752.50 37.62 4.00 2.00 29.28 25.35 1.85 61.82 31.57 3.62 1.52 66.16 26.92 69.30 33.11 50.57 -1.10 0.33 45.00 33.09 1.00 0.00 125.15 57.85 6.04 2.26 Standard Variability deviation (mm) coefficient (%) 170.00 158.00 127.18 266.17 231.27 237.00210.86 61.00 209.77 295.72 392.64 240.00 216.49 612.00 (mm) M aximum 13.00 183.00 17.00 74.76 82.43 64.00 69.08 41.73 100.91 177.20 377.31 112.50 350.58 129.56 103.40118.00 225.16 445.00 37.50 111.80 349.48 117.45 158.04 117.29 536.11 286.50 394.49 (mm) M inim um 92.35 168.00 71.00 157.13 128.50 2.10 250.71 152.03 64.60 (mm) 106.75 113.00 94.00 198.00 125.00150.13 38.00 175.44 133.00 227.40 128.07 346.40 292.78 227.32 398.07 317.62 204.56 M edian 65.48 55.12 (mm) Mean 133.30 158.43 165.56 176.13 177.41 157.00 182.58 125.60115.65 107.20 125.60 195.81 130.31 126.12 232.43 267.00 257.86 206.28 193.40 208.18 9 121.00 N 11 116.00 12 17 136.00 24 23 92.35 32 115.51 34 32 Station W eather Nedjo 16 Shemena Sheneka 32 Seru 17 210.00 Sedika 18 151.70 Nr. 132 M uja 10 131 M inneh 10 128 Metekel129 Midre 130Genet Mieso 15 8 293.04 36 127 Meta Hara 126 M ersa 133 M unessa 10 134 135 Nekemte 11 323.73 136 Rike 25 206.76 209.17 139 Sagure 140 143 144 137 Robe 138 Robi 141 142 Shakiso145 Sheno 11 147 Sinana148 Sire 149 Sodo 8 18 216.32 187.04 146 Shola Gebeya 35 150 Tendaho (Dubti) Tab. 20: Continued

Rainfall and its erosivity’ in Ethiopia 224 1.54 1.34 0.00 0.54 0.42 -0.54 -0.04 -0.90 Skewness 1.21 0.95 4.46 -1.33 3.23 4.18 1.50 2.00 0.74 1.00 -0.90-0.34 -0.30 -0.45 -1.30 Kurtosis 18.06 30.79 37.47 2.60 -0.80 36.9025.92 -1.51 -0.08 24.79 - 1.00 32.5728.70 -0.73 -0.33 33.48 34.55 22.60 -0.02 0.60 61.24 -1.2040.07 0.35 1.80 21.71 3.30 40.67 81.10 74.37 35.55 43.77 22.30 33.32 65.17 27.00 97.86 66.29 48.06 131.56 65.56 Standard Variability deviation (mm) coefficient (%) 352.57 531.50 121.36 342.00 337.85 42.10 330.07 67.76 260.34 7.10 343.00 71.43 0.00 85.73 20.65 122.51 412.25 115.07 280.57 48.00 (mm) (mm) Minimum Maximum 187.53 127.30 180.92 (mm) 310.25 26.46 285.24 M edian Mean (mm) 193.88 196.26 191.00 127.96 193.65 194.00 150.00 155.00 44.00 283.00165.45 168.40 147.57 135.51 61.00 67.80 256.40 196.11 232.02 220.40 200.67 230.50 209.20 206.30242.20 223.72 90.71 119.20 451.90 442.30 9 106.42 87.60 14.52 203.40 13 323.92 11 10 13 265.17 N 35 32 34 20 24 Station W eather Yirba Dimbicho 13 149.52 158.60 102.70 192.65 Yirga Alem 20 141.20 143.55 67.25 201.00 36.60 Urgessa 24 Zege Turmu Nr. 168 167 Yitnora 164 Yirba M165 uda 166 Yirga ChefTe 162 Yetmen163 32 159 Worancha160 Worota 11 116.10 122.13 15 73.03 181.29 151 Tibila 152 Tulu Bolo153 154 157 Wogel Tena 155 W adera 13 161 W uchale 156 Woldia158 Wonji 26 170.76 163.53 108.58 260.12 38.60 Tab. 20: Continued

Rainfall and its erosivity in Ethiopia 225 1 1 1 3 4 2 5 5 4 6 4 4 4 1.45 1 1.85 3 0.78 0.76 0.50 0.18 5 0.24 0.12 3 -0.24 ■1.01 -0.57 -1.37 2 1.77 1.84 1.11 1 1.17 1.41 0.15 2.37 1.48 2 3.170.60 0.66 4 0.38 0.36 3 3.29 0.18 -0.59 -0.20-0.25 0.42 0.58 -0.11 -0.35 0.40 -0.60 -0.73 10.32 2.88 4 Kurtosis Skewness Class 16.97 0.78 15.78 18.97 16.94 86.21 43.94 24.26 25.33 23.04 43.76 24.22 -0.8732.95 -0.21 2.43 150.51 15.75 3.96 2 Variability coefficient (%) 27.37 76.08 42.52 19.26 66.4065.94 30.57 23.88 1.16 -0.53 70.17 53.56 2.42 76.59 66.54 28.59 1.00 1.11 3 40.64 201.77 Standard deviation (mm) 180.72 23.50 181.46 27.86 27.22 516.30 128.68 252.91385.97 36.75 53.06 519.72 92.78 35.26 699.70 141.21 37.89 389.63 380.19 227.54 39.77 27.53 388.64 54.11 401.00 428.31 439.09 49.55 19.21 (mm) M aximum 86.13 81.79 1281.9973.01 291.20 100.20 289.52 165.20 102.73 172.57 121.02 297.75 47.51 (mm) M inimum 140.98 92.60 169.85 136.22 231.98 135.02 112.31 118.90 118.84 28.25 219.77 169.71 9.46 818.39 (mm) 335.80 179.40 226.06 214.96 91.73 278.00 199.00 381.74 41.10 14.53 299.64 166.24 476.34 82.54 M edian (mm) Mean 102.37 103.30 63.72 151.48 144.43 146.76 131.01 110.86 17.00 323.22 193.48 173.84 191.58 2.00 251.03 279.67 271.61 179.98 319.37 315.21 216.22 220.79 217.20 276.07 272.58 257.90 254.61 164.62 325.82 263.12 258.30 287.43 254.59316.22 314.66 82.79 953.20 186.11 64.75 9 11 138.46 15 173.40 19 18 13 232.70 215.00 17 123.34 N 36 292.83 315.21 6.63 23 38 282.83 24 20 234.05 99 station W eather Huruta Goba GobessaGohatsion 17 26 Debre Zeit 34 Debre Work Asela 23 206.20 207.03 Artuma Asasa Adaba Albuco 18 1 7 8 3 Agarfa5 Alem Ketema 20 23 9 6 A nkober 20 2 Addis Ababa 4 19 Fiche 11 Bora12 Boru Meda 13 18 Ejaji 23 14 15 16 Dessie17 Dhera 10 Awassa13 Debre 16 Markos Nr. 25 21 22 23 24 Hosana 20 Gessera 19 372.66 Tab. 21: Statistical characteristics oflong-term summer rainfall erosivity in the central highlands ofEthiopia

Rainfall and its erosivity in Ethiopia 226 1 1 4 3 6 3 2 4 4 4 1.24 1.08 0.15 0.24 0.82 1 0.44 2.28 0.38 3.25 0.46 2 -0.17 -0.87 0.06 - 1.10 1 0.27 7.09 0.89 2.36 -1.42 -0.09 2 -0.24 -0.95 0.12 2 -0.36 0.68 2 -1.34 -0.63 -0.47 1 -0.97 Kurtosis Skewness Class 15.74 10.37 20.32 0.70 -0.13 5 24.57 25.64 27.70 0.31 23.14 23.16 23.27 28.73 34.7234.10 1.50 Variability coefficient (%) 32.1835.76 26.71 27.03 85.43 32.33 -0.43 0.28 4 28.54 36.63 77.29 29.20 2.18 63.06 41.51 Standard deviation (mm) 159.05 210.79 330.10 71.82 35.72 -1.24 214.43 557.26 108.17 51.18 224.53 36.48 246.66 41.15 278.86 50.48 211.97 523.50 89.73 295.57 (mm) M aximum 88.83 430.87 69.57 86.31 435.19 69.56 170.07 82.09 65.66 66.50 481.24 53.50 210.89 46.71 32.47 74.13 120.50 163.50 681.13 87.23 32.47 14.65 106.56 (mm) M inimum 117.19 (mm) 167.43 104.93 180.42 128.89 341.90 211.70 470.67 70.64 264.45 247.87 256.89 261.60 Median Mean (mm) 120.46 148.43 135.60 89.10 123.25 347.72 400.47 395.70 343.76 497.62 264.27 251.17 268.65 211.33 190.60 9 N 10 177.81 17 143.84 151.74 21 139.43 135.52 69.80 21 station W eather

IdotoKachise 17 Kombolcha 11 40 Kofole W onji 33 184.93 157.01 Shola Gebeya 35 Sire N ekem te 11

Nr. 33 M araro 21 201.06 181.35 106.23 31 Kore32 M ajete 16 32 171.78 172.53 133.02 26 30 Tab. 21: Continued 34 M35 unessa 36 Robi 38 Sedika39 Seru 17 175.73 158.10 16 102.12 27 28 Keleta 37 Sagure 10 157.39 29 40 Sheno 34 264.66 41 42 43 Tulu44 Bolo 35 258.45 248.70 Rainfall and its erosivity in Ethiopia I 227 1.43 1.60 1.20 1.80 1.54 1.60 3.51 0.06 0.48 0.45 -0.33 -0.30 1.50 0.97 5.53 0.58 1.01 -0.86-0.55 0.14 0.22 -0.63 -0.41-0.81 0.14 -0.11 11.46 3.20 27.00 Kurtosis Skewness 65.56 5.8053.98 2.10 49.38 4.50 45.11 -0.85 28.64 28.14 1.20 1.16 49.98 0.29 0.67 47.63 49.19 47.38 3.60 117.34 13.54 Variabilit coefficient (%) 58.28 55.28 50.76 37.55 -0.50 0.51 65.78 38.1295.78 7.0439.4234.49 76.34 68.14 40.82 39.66 28.16 61.86 55.98 -0.51 0.27 46.64 53.22 -0.64 29.5667.93 23.60 101.75 56.71 0.53 1.10 Standard deviation (mm) 173.19 338.72 52.00 27.87 0.60 308.41 204.09 218.82 250.28 56.68 34.74 226.67 277.08 59.77248.50 58.17 2.91 412.90 (mm) M aximum 8.00 146.80 36.50 45.60 19.59 189.88 15.00 179.87 35.58 17.00 390.95 80.67 74.26 76.51 200.70 33.26 27.30 58.78 53.30 20.78 80.49 53.80 176.71 45.40 37.67 -1.61 67.93 188.76 61.95 61.15 429.04 49.01 257.13 (m m) M inim um 88.87 35.57 69.79 94.44 95.52 28.77 697.24 150.59 169.03 50.70 522.07 162.37 105.35 3.00 203.00 48.14 116.54 107.87 118.97 152.90 189.00 100.40 128.96 138.18 64.85 (mm) Median 72.06 (mm) Mean 128.34 186.60 135.16 116.07 179.43 147.01 129.20 133.99 102.74 9 166.96 1 11 120.51 127.54 18 13 N 18 19 19 98 172.54 20 123.05 34 97.86 23 125.23 33 110.51 112.01 0.50 226.19 21 163.13 155.53 38 station W eather Huruta Dessie 25 Bora 12 87.63 87.22 Hosana 16 135.91 119.43 22.00 281.42 Debre Work Fiche Adaba Addis Ababa Asela Alem Ketema 22 Agarfa Debre Zeit Artum a Asasa 16 51.64 41.14 Goba Gohatsion 25 113.39 105.70 19.20 282.20 53.72 Albuco Ankober 20 177.44 1 5 7 8 2 3 6 9 4 l ab. l 22: Statistical characteristics oflong-term autumn rainfall erosivity in the central highlands ofEthiopia 18 Ejaji 20 140.84 15 12 Boru Meda 14 16 17 Dhera19 18 80.04 68.41 10 Awassa11 16 120.44 123.69 13 Debre Markos Nr. 25 21 22 Gobessa23 24 17 118.19 20 Gessera

Rainfall and its erosivity in Ethiopia 1.14 2.79 0.22 0.91 -0.11 -0.08 1.43 0.66 1.16 9.56 2.75 3.21 1.45 4.82 2.01 2.16 1.50 0.61 0.65 -0.45-0.06 -1.34 -0.25 -0.17 Kurtosis Skewness 19.09 -1.54 50.68 26.91 32.14 50.21 -0.52 -0.02 41.04 40.42 60.72 Variabilit 37.49 38.76 30.83 0.8175.49 0.46 42.7958.35 0.22 44.82 43.84 Standard 199.82 61.37 98.66 -0.12 0.78 131.47 27.41 275.34 49.27329.48 51.55 214.44 211.97256.25 36.63 49.11 23.27 46.36 422.67 65.02 68.54 20.41 3.98 M aximum 1.00 264.70 15.02 35.89 40.13 138.82 365.18 69.20 114.03 203.37 30.76 88.4579.77 3.10 172.45 83.57 93.91 98.72 39.25 254.90 45.13 45.68 151.36 74.03 167.43 104.93 234.56 121.43 359.76 79.25 33.33 -0.50 94.87 62.21 43.2096.10 0.00 98.79 Mean Median Minimum 161.14 161.97 104.41 109.40 32.65 235.50176.42 42.21 237.79 10 16 125.71 123.08 58.95 10 157.39 19 128.03 143.27 67.60 184.36 34.33 26.82 -0.80 -0.43 N 17 88.43 84.79 12.40 212.67 33 34 40 161.77 160.50 74.90 360.00 53.83 33.28 3.28 station (mm) (mm) (mm) (mm) deviation (mm) coefficient (%) W eather

Wonji 33 M unessa Robi 9 85.29 Majete KachiseKeleta 10 21 91.36Mararo 80.22 43.42 20 95.59 195.77 80.55 Kore Sheno Tulu Bolo 30 87.32 Nekemte 11 257.16 254.31

Nr. 31 34 35 36 37 Sagure38 Sedika39 Seru 17 105.93 16 82.61 55.83 30 Kombolcha 32 33 42 Sire43 44 16 27 28 29 Kofole 40 41 Shola Gebeya 33 Tab. 22: Continued 26 Idoto Rainfall and its erosivity in Ethiopia ■ l

229 1.19 1.20 2.28 0.37 0.67 2.28 2.05 0.13 0.42 0.84 -0.16 Skewness 1.48 1.14 1.29 0.94 1.60 0.97 1.56 6.05 2.03 0.24 0.90 0.19 0.05 1.08 -0.07 0.91 -0.16 0.11 -0.29-0.32 0.19 0.74 Kurtosis 37.36 2.38 70.73 6.14 57.73 5.71 48.38 - 1.01 46.8349.35 0.93 - 1.10 1.14 27.07 47.74 30.06 - 1.12 0.05 42.93 2.51 1.54 48.74 43.66 62.07 46.43 48.73 1.09 0.72 45.90 -1.09 0.07 Variability coefficient (%) 35.88 46.55 -0.18 47.78 56.6733.91 40.62 36.00 40.45 -0.94 95.71 33.67 46.16 44.40 66.35 26.8035.26 47.53 37.79 43.77 63.34 37.04 33.63 26.56 64.05 48.83 43.71 45.92 42.92 108.36 52.47 Standard deviation (mm) 181.83 34.01 116.91 184.97 137.15 158.38 189.62 192.15 270.42 657.95 145.48 79.59 329.89 442.39 660.15 163.84 297.55 237.53 242.04 226.80 48.20 442.59 2.58 202.38 15.63 30.43 57.67 89.33 197.45 25.58 31.91 21.60 150.98 41.40 298.98 76.03 235.88 35.47 30.00 83.51 28.65 200.15 (mm) (mm) Minimum Maximum 85.02 74.10 87.00 30.00 157.60 80.85 13.38 154.08 83.22 138.23 14.68 188.14 87.27 110.05 117.83 82.18 (mm) 104.66 140.83 102.36 205.47 75.19 M edian 91.03 73.87 71.67 89.00 76.01 94.14 94.24 Mean (mm) 103.95 94.23 21.60 114.94 108.58 162.42 154.58 110.40 112.40 31.30 126.63 122.79 67.39 204.37 16 98.75 86.01 33.99 N 12 16 68.23 72.00 18 17 22 77.08 20 38 24 26 station W eather Addis Ababa 98 A daba 19 Debre WorkDebre ZeitDessie 13 70.92 34 66.68 24 34.87 Albuco Asela 23 139.50 134.04 Alem Ketema Hosana 16 139.52 Dhera Artum aAsasa 10 231.64 Fiche 35 69.15 64.48 Awassa 15 125.25 122.87 Debre Markos Huruta 13 Gohatsion Goba 22 1 2 3 Agarfa5 19 182.80 7 8 4 9 6 Ankober Nr. 10 1112 Bora Boru Meda14 15 1916 147.54 17 13 18 Ejaji 19 Tab. 23: Statistical characteristics oflong-term spring rainfall erosivity in the central highlands ofEthiopia 20 Gessera 18 206.52 162.07 23 24 21 25 22 Gobessa

Rainfall and its erosivity in Ethiopia 230 1.13 1.44 0.42 0.84 0.80 0.44 0.81 0.40 0.46 0.82 0.63 -0.19 -0.19 -1.80 -0.42 Skewness 1.78 2.38 0.99 0.37 0.59 2.21 0.17 -0.69 -1.76 -0.99-0.36 -0.32 -0.64 Kurtosis 39.03 -0.52 43.93 0.28 52.05 -0.28 36.93 41.64 -1.32 0.25 33.86 44.19 -0.23 38.94 56.9437.11 -0.10 44.68 26.79 39.54 0.78 21.48 44.75 43.41 23.32 40.54 0.82 Variability coefficient (%) 28.16 76.59 36.46 34.41 35.38 58.83 53.05 32.65 39.67 34.09 53.14 31.18 3.83 26.04 68.62 24.77 129.08 Standard deviation (mm) 157.49 149.82 194.99193.50 39.93 130.47 32.76159.68 157.90 241.04 204.42 265.05 48.35 243.17 47.08 (mm) M axim um 4.00 127.07 87.67 80.25 284.93 30.38 40.52 23.41 135.93 38.45 94.48 42.70 (mm) M inimum 66.56 88.79 65.19 92.48 30.57 197.86 115.93 181.47 40.50 218.25 125.50 128.41 42.90 113.29 41.73 230.11 (mm) M edian 88.16 82.28 27.15 64.09 80.07 74.78 33.00 70.18 89.36 84.14 26.03 (mm) Mean 131.48 131.84 111.58 8.50 279.96 140.01 132.75 122.29 125.78 170.45 226.69 214.82 24.00 495.51 9 139.01 15 87.55 77.70 17 127.21 11 207.39 262.98 11 122.19 17 115.31 110.87 71.03 N 10 32 33 21 40 97.86 station W eather Wonji Mararo Tulu Bolo 35 Sagure 11 84.13 Robi Sedika 17 Idoto Kofole 20 Nekemte Nr. 39 Seru 15 43 44 36 42 Sire 38 33 35 37 41 Shola Gebeya 35 40 Sheno 3031 Kombolcha Kore 32 M ajete 33 34 M unessa 28 Keleta 20 27 Kachise 29 Tab. 23: Continued 26

Rainfall and its erosivity in Ethiopia I 1.51 1.78 1.16 1.85 1.47 0.97 0.36 0.62 3.16 0.15 0.50 0.50 0.33 0.100.91 0.88 5.17 2.68 0.74 4.43 2.08 0.47 0.57 2.95 1.49 -1.25-1.16 0.28 0.56 -0.35 -0.31-0.98 0.05 Kurtosis Skewness 73.30 75.17 2.62 73.35 74.07 0.67 69.39 64.15 79.5370.57 2.71 1.71 54.16 -0.10 50.71 85.97 12.97 62.62 -0.40 59.54 75.31 53.30 -1.33 0.30 65.51 5.72 1.86 68.33 108.81 2.99 1.71 133.11 3.02 112.80 121.91 3.20 1.80 Variability coefficient (%) 19.39 17.86 33.09 47.97 52.38 79.92 48.91 97.37 35.39 69.19 33.08 81.71 74.05 41.5664.44 22.00 66.18 0.00 58.64 24.50 42.37 23.72 54.79 Standard deviation (mm) 74.00 53.85 14.84 63.50 120.65 30.62 175.35 105.81 24.94 150.12 28.15 242.34 454.86 1.20 185.76 1.50 106.11 28.61 1.00 89.17 2.602.35 167.01 3.40 181.12 199.30 8.20 269.37 2.50 84.62 7.32 191.80 2.00 0.73 112.05 0.40 250.92 0.80 113.75 30.72 6.50 0.10 12.20 10.17 212.95 42.71 25.40 48.63 27.60 8.69 62.48 2.40 147.61 41.53 5.43 117.30 47.00 52.35 36.84 27.73 5.27 42.79 134.31 9.19 (mm) (mm) (mm) Median Minimum Maximum 75.50 65.81 44.09 66.72 51.14 62.80 51.59 36.50 48.31 40.48 32.32 0.20 137.48 (mm) Mean N 17 16 38.63 33.21 13 27.40 24.92 3.74 13 19 16 43.30 19 12 60.74 31.40 20 20 25 14.56 5.67 0.98 21 132.75 23 25 33.51 28.20 station W eather

Addis Ababa 97 44.02 37.02 Albuco Asela Dhera Awassa 16 38.50 Ejaji Fiche 32 Gohatsion 25 32.74 Gessera Gobessa 17 77.86 60.67

1 Adaba 2 5 Aleni Ketema 3 Agarfa 8 Asasa 4 67 A nkober Artuma9 9 124.59 93.90 Nr. 10 1213 Boru Meda Debre Markos 20 38 91.31 35.13 86.27 31.11 14 Debre Work 11 Bora 15 Debre16 Zeit Dessie 31 37.56 25 24.09 17 18 19 Tab. 24: Statistical characteristics oflong-term winter rainfall erosivity in the central highlands ofEthiopia 20 24 Hosana 2122 Goba 23 65.20 61.84 25 Huruta 23 Rainfall and its erosivity’ in Ethiopia i 232 1.50 1.01 1.33 1.02 3.58 2.97 0.59 2.55 0.87 0.43 0.77 0.75 0.47 -1.38 Skewness 2.29 7.03 5.24 2.00 3.44 1.45 2.78 1.32 0.140.68 0.39 0.40 -1.19 11.34 -0.35 0.55 -0.85 -0.46 Kurtosis

74.11 -1.00 85.3079.91 1.83 80.92 57.46 51.6481.69 2.8373.98 96.36 78.49 53.13 65.63 64.39 -0.38 62.89 -0.55 111.94 113.33 220.17 13.10 Variability coefficient (%)

17.84 75.21 31.69 28.60 59.92 29.8764.15 23.22 97.61 48.74 45.99 31.79 75.91 40.37 40.17 Standard deviation (mm) 77.81 81.66 95.55 27.35 126.50 131.60 146.79 35.37 199.65 141.24 39.01 173.24 248.00 207.14206.11 64.31 265.55 228.60 427.54 110.09 (mm) Maximum 1.03 1.67 137.86 32.75 5.00 3.00 2.00 9.24 211.55 7.11 58.00 4.37 95.67 19.40 (mm) Minimum 14.67 3.00 33.31 81.14 50.29 1.40 55.53 78.71 25.13 1.51 74.07 17.18 26.70 3.60 29.00 (mm) Median 31.33 50.00 30.60 21.97 1.08 37.16 30.4243.54 1.66 86.93 84.74 9.00 66.57 92.95 75.99 61.55 60.59 58.82 12.48 34.85 28.37 41.89 35.28 Mean (mm) 9 52.87 37.07 8.51 14 17 11 28 17 34 N station Weather

Sheno Mararo 21 56.30 55.13 Sedika Tulu Bolo 32 35.79 27.80 Kofole 21 Nekemte 11 Kombolcha 40 49.90 39.57 Kachise 11

Nr. 43 44 Wonji 42 Sire 35 3738 Sagurc 39 Seru 16 36 Robi 40 41 Shola Gebeya 33 33 34 Munessa 10 30 3132 Kore Majete 17 33 77.78 29 28 Keleta 21 26 Idoto 27 Tab. Tab. 24: Continued Rainfall and its erosivity in Ethiopia f 233

Tab. 25: Regression models of long-term autumn rainfall erosivity for the central highlands o f Ethiopia

Weather Regression models Coefficient of Nr. station Y = rainfall erosivity, X = mean rainfall determination (R2) 1 Adaba Y - 1.12X088 0.88 2 Addis Ababa Y = 27 + 0.7X 0.75 3 Agarfa Y = 0.5X1 03 0.94 4 Albuco Y = 26.85 + 0.5X 0.97 5 Alem Ketema Y = 1,5X° 9 0.90 6 Ankober Y = 53.2e0

Rainfall and its erosivity in Ethiopia 236

240

Fig. 58: Continued

Debre Markos ACCs of long-term annual rainfall erosivity

Auto- S ta n d . Lag C o rr. Err. -1 -.75 -. 5 -.25 0 .25 .5 .75 1 Box-Ljung Prob. 1 1 1 1 1 1 1 1 1 1 1 1 1 1 .016 . 156 .010 .920 2 - .016 . 154 .021 . 990 3 - .241 . 152 2 .550 .466 4 .243 . 150 5 .194 .268 5 . 129 . 147 * ★ * 5 . 955 .311 6 .016 . 145 5 . 967 .427 7 - .244 .143 8 . 879 . 261 8 - .083 . 140 * ★ 9.228 .323 9 . 108 . 138 * * 9.841 . 364 10 . 247 . 136 * ★ * ★ ★ 13 . 145 .216 11 .013 .133 13.155 .283 12 -.098 . 131 ** 13.711 . 320 13 - . 118 . 128 * * 14.552 .336 14 . 119 . 126 * * 15.456 .348 15 . 069 . 123 * 15.772 . 397 16 - . 114 . 120 * * 16.677 .407

Plot Symbols Autocorrelations * Two Standard Error Limits . Total cases: 38 Computable first lags: 37

Debre Markos: ACCs of long-term summer rainfall erosivity

Auto- S ta n d . Lag C o rr. E rr. - 1 -.75 -.5 -.25 .25 .5 .75 1 Box-Ljung Prob. 1 1 1 1 1 1 1 1 1 I I 1 till 1 .238 . 156 2.337 .126 2 . 016 . 154 2 . 348 . 309 3 - . 131 .152 * * * 3 .093 . 378 4 .201 . 150 * * * * 4 . 905 .297 5 .151 . 147 ★ * * 5.959 .310 6 .022 . 145 5 . 983 .425 7 - .216 . 143 * ★ * * 8 .278 .309 8 -.061 . 140 * 8 .466 .389 9 . 153 . 138 * * * 9. 694 . 376 10 .203 . 136 * * * * ( 11. 933 .290 11 - . 057 . 133 * 12.115 . 355 12 - . 150 .131 13.438 .338 13 - . 120 . 128 ** 14.314 . 352 14 - . 007 . 126 14.317 .426 15 - . 008 . 123 14.321 .501 16 - .255 . 120 18 . 818 .278

P lo t Symbols Autocorrelations * Two Standard Error L im its . Total cases: 38 Computable first lags: 37

Rainfall and its erosivity in Ethiopia 241

Fig. 58: Continued

Debre Zeit: ACCs of long-term annual rainfall erosivity

Auto- Stand. :-Ljung P ro b .

1 .388 . 164 5 . 592 . 018 2 .306 . 162 9. 180 .010 3 .139 . 159 9.945 .019 4 - .011 . 157 9.950 . 041 5 . 104 . 154 10 .403 .065 6 . 017 . 151 10 .416 . 108 7 - . 084 . 149 10.737 . 151 8 - . 006 . 146 10.739 .217 9 - .055 . 143 10.885 .284 10 - . 048 . 140 11.002 .357 11 - . 154 . 137 12.262 . 344 12 - . 342 . 134 18.761 . 094 13 - . 348 . 131 25.822 . 018 14 - . 335 . 128 32.678 .003 15 - . 185 . 125 34.891 . 003 16 - .073 . 121 35.258 .004

P lo t Symbols Autocorrelations * Two Standard Error Limits T o tal c a s e s : 34 Computable first lags: 33

Debre Zeit: ACCs of long-term summer rainfall erosivity

Auto- Stanc Lag C o rr. E rr. ;-Lj ung P ro b .

1 .465 . 164 8 . 008 .005 2 .338 .162 12.376 .002 3 . 149 . 159 13.253 .004 4 .034 . 157 13.300 .010 5 . 102 . 154 13.743 .017 6 . 026 . 151 13.772 .032 7 - . 075 .149 14.025 .051 8 - .043 . 146 1 4 .I l l .079 9 - . 108 . 143 14.679 . 100 10 - . 076 . 140 14.976 . 133 11 - .215 .137 17.441 .095 12 - .320 . 134 23.152 . 026 13 - . 367 . 131 31.003 .003 14 - .334 . 128 37.829 .001 15 - . 186 .125 40.054 .000 16 - . 044 . 121 40 . 187 .001 P lo t Symbols Autocorrelations * Two Standard Error Limits T otal c a s e s : 34 Computable first lags: 33

Rainfall and its erosivity in Ethiopia 242

Fig. 58: Continued

Ejaji: ACCS of long-term annual rainfall erosivity

Auto- Stand Lag C o rr. E rr. -1 - .75 - . 5 - .25 .25 .5 -Ljung Prob. 1 1 1 1 1 1 1 1 1 1 t 1 1 . 147 . 169 ★ * * . 759 .384 2 - .073 . 166 ★ . 953 .621 3 - .010 . 163 . 956 .812 4 - .034 .160 • 1. 002 .910 5 . 020 . 158 ★ 1.017 . 961 6 - . 004 . 155 1.018 .985 7 - . 063 .152 ★ 1.193 .991 8 - .039 . 149 * 1.262 .996 9 . 185 . 145 ★ * * * 2 . 874 . 969 10 .037 . 142 * 2 . 944 . 983 11 - . 107 .139 ★ * 3 .534 .982 12 - . 103 . 136 ★ * 4 .117 . 981 13 -.011 .132 4 . 124 . 990 14 .002 .129 4 . 124 . 995 15 - . 190 . 125 # * * * * 6.432 . 972 16 - .247 .121 10.580 .835

Plot Symbols Autocorrelations * Two Standard T otal c a s e s : 32 Computable first lags: 31

Ejaji: ACCs of long-term summer rainfall erosivity

AutO- Stand Lag C o rr. Err. -1 -.75 -.5 -.25 .25 .5 .75 1 30X -Ljung 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 .046 . 172 * .070 .791 2 .040 .167 * .127 .939 3 - . 072 .144 * .373 . 946 4 .239 .150 2 . 907 . 574 5 -.078 .144 ** 3 .198 .669 6 . 118 .150 * * 3.816 . 702 7 .062 .150 ★ 3 . 985 .782 8 .072 .150 * 4.212 .837 9 - . 042 . 144 ★ 4.296 .891 10 - .236 . 132 7.508 .677 11 . 044 .125 * 7.634 .746 12 - . 127 .132 ★ ★ ★ 8 . 560 .740 13 - . 015 . 125 ■ 8 . 575 . 804 14 - . 181 .118 • 10.922 .692 15 . 011 .110 * 10.933 .757 16 - .135 .110 12.431 .714 •

P lo t Symbols Autocorrelations * Two Standard Error Limits T otal c a s e s : 32 Computable first lags: 17

Rainfall and its erosivity in Ethiopia 243

Fig. 58: Continued

Fiche: ACCs of long-term annual rainfall erosivity

Auto- Stand. Lag Corr. Err. -1 -.75 -.5 -.25 .25 .5 .75 Box-Ljung Prob. H— b 1 .371 .146 *****' * 6.471 .011 2 .491 .144 *****.**** 18.110 .000 3 .205 .142 * * * * 20.183 .000 4 .229 .141 *****' 22.829 .000 5 .162 .139 *** 24.194 .000 6 .167 .137 *** 25.673 .000 7 .061 .135 * 25.880 .001 8 .099 .133 ** 26.427 .001 9 -.028 .132 * 26.473 .002 10 - .069 .130 * 26.755 .003 11 .023 .128 , 26.787 .005 12 -.083 .126 27.220 .007 13 -.033 .124 * 27.293 .011 14 -.129 .122 *** 28.419 .013 15 - . 179 .120 , **** 30.654 .010 16 -.281 .118 * m* * ** 36.356 .003

Plot Symbols: Autocorrelations * Two Standard Error Limits Total cases: 44 Computable first lags: 43

Fiche: ACCs of long-term summer rainfall erosivity

Auto- Stand. Lag C o r r . B o x -Lj ung P r o b .

1 .225 .146 2.385 .123 2 .248 .144 5.340 .069 3 -.069 .142 5.573 .134 4 .025 .141 5.605 .231 5 -.073 .139 5.883 .318 6 -.065 . 137 6.111 .411 7 -.174 .135 7.757 .354 8 - .076 .133 8.079 .426 9 -.200 .132 10.388 .320 10 - .237 .130 13 .724 .186 11 .000 .128 13 .724 .249 12 -.023 .126 13.759 .316 13 .040 .124 13.862 .384 14 - . 135 .122 15.096 .372 15 -.112 .120 15.975 .384 16 - .182 .118 18.372 .303

Plot Symbols: Autocorrelations * Two Standard Error Limits Total cases: 44 Computable first lags; 43

Rainfall and its erosivity in Ethiopia 244

Fig. 58: Continued

Kombolcha: ACCs of long-term annual rainfall erosivity

Auto- Stand Lag C o rr. Err. -1 -.75 -.5 -.25 .25 .5 .75 1 Box-Ljung Prob. 1 1 1 1 1 1 1 1 till till 1 . 094 .152 * * .385 .535 2 .040 . 150 * .456 .796 3 - . 027 .148 + .489 .921 4 .064 . 146 * .679 .954 5 . 053 . 144 * . 816 .976 6 - . 098 . 142 * * 1.294 .972 7 .050 . 140 * 1.420 .985 8 -.297 .138 6.066 .640 9 - . 155 .136 * * * 7.375 . 598 10 - . 092 . 134 * * 7 . 850 . 644 11 .242 . 131 11.245 .423 12 - . 050 . 129 * 11.395 .495 13 - . 075 . 127 * 11.741 .549 14 .030 . 124 * 11.798 .622 15 -.045 .122 * 11.932 .684 16 . 017 . 120 ■ 11.952 .747

P lo t Symbols Autocorrelations * Two Standard Error L im its . T otal c a s e s : 40 Computable first lags: 3 9

Kombolcha: ACCs of long-term summer rainfall erosivity

Auto- S tand. Lag Corr. Err. -1 -.75 -.5 -.25 .25 .5 .75 1 Box-Ljung Prob. 1 I 1 1 1 1 1 . 141 . 152 I'.' * * * 1 . 855 .355 2 .028 .150 * .891 .641 3 - .105 . 148 ♦ * 1. 392 . 708 4 - .038 . 146 * 1.458 .834 5 . 116 . 144 * ★ 2 .105 .834 6 .004 .142 2 .106 .910 7 - . 011 . 140 2 .113 .953 8 - .258 .138 5.613 .691 9 - .074 .136 * 5.913 .749 10 - . 119 . 134 * * 6 .701 .753 11 .248 .131 10.272 .506 12 . 050 .129 * 10.423 .579 13 - . 016 . 127 10.439 .658 14 - . 009 . 124 10.445 .729 15 -.0 9 0 . 122 * * 10.990 .753 16 . 118 .120 ★ * 11.963 .747

Plot Symbols Autocorrelations * Two Standard Error L im its . Total cases: 40 Computable first lags: 39

Rainfall and its erosivity in Ethiopia 245

Fig. 58: Continued

Majete: ACCs of long-term annual rainfall erosivity

Auto- Stand. :-Ljung P ro b .

1 .312 . 162 3 .696 .055 2 .059 . 160 3.833 . 147 3 -.3 1 1 . 157 7 . 747 .052 4 - . 178 . 155 9.078 .059 5 .034 .152 9. 127 . 104 6 . 106 . 150 9.632 . 141 7 .052 . 147 9.759 .203 8 - .159 . 144 10.967 .204 9 - .215 . 142 13.278 .150 10 - . 042 . 139 13.369 .204 11 . 180 . 136 15.109 . 178 12 .209 . 133 17.574 . 129 13 .243 .130 21.041 . 072 14 - .057 . 127 21.242 .096 15 - .079 . 124 21.642 .118 16 - .247 . 121 25.793 . 057

P lo t Symbols Autocorrelations * Two Standard Error Limits T o tal c a s e s : 35 Computable first lags: 34

Majete: ACCs of long-term summer rainfall erosivity

Auto- Stanc Lag C o rr. E rr. Ljung Prob.

1 .337 .162 4 .318 . 038 2 . 101 . 160 4 .721 .094 3 -.276 . 157 7 . 809 .050 4 - .243 . 155 10.269 .036 5 - . 058 . 152 10.412 .064 6 . 031 . 150 10.456 . 107 7 . 060 . 147 10.624 . 156 8 - . 090 . 144 11.011 .201 9 - .209 . 142 13.188 . 154 10 - . 042 . 139 13.281 .208 11 . 147 . 136 14.453 .209 12 .219 . 133 17.154 . 144 13 .230 . 130 20.270 .089 14 - . 002 . 127 20.270 . 122 15 - . 069 . 124 20.578 . 151 16 - .236 . 121 24.375 .082 P lo t Symbols Autocorrelations * Two Standard Error Limits T o tal c a s e s : 35 Computable first lags: 34

Rainfall and its erosivity in Ethiopia 246

Fig. 58: Continued

Sheno: ACCs of long-term annual rainfall erosivity

Auto- Stand. Lag C o rr. E rr. -1 -.7 5 -.5 -.25 0 .25 .5 .75 1 Box-Ljung P ro b . 1 1 1 1 1 1 1 1 1 1 t 1 1 1 1 1 1 1 - .015 . 164 .008 . 929 2 - . 128 . 162 ★ * * . 630 .730 3 .333 . 159 5.020 . 170 4 - .009 . 157 5.023 .285 5 .099 .154 * * 5.438 .365 6 -.118 .151 * * 6.046 .418 7 .045 . 149 ★ 6.139 . 524 8 . 058 . 146 * 6.296 .614 9 - .209 . 143 * i t i t i t 8.436 .491 10 .055 . 140 * 8 . 587 .572 11 .020 . 137 8.608 .658 12 .034 . 134 * 8.671 .731 13 -.112 . 131 i t * 9.409 .741 14 - . 129 . 128 + + + 10.425 .730 15 .208 .125 * * * * 13.222 .585 16 - . 001 .121 13.222 .656

Plot Symbols: Autocorrelations * Two Standard Error Limits . Total cases: 34 Computable first lags: 33

Sheno: ACCs of long-term summer rainfall erosivity

Auto- Stand.

1 - . 071 164 .187 .666 2 - .254 162 2 . 663 .264 3 .225 159 4 .656 .199 4 . 020 157 4 . 671 .323 5 . 124 154 5 .319 .378

6 - . 112 151 5 . 865 .439 7 . 028 14 9 5 . 900 .551 8 .010 146 5 . 905 .658 9 - . 198 143 7 . 819 .552 10 . 028 140 7 . 859 .643 11 . 043 137 7 . 956 .717 12 . 100 134 8 . 511 .744 13 - . 116 131 9.301 . 750 14 - . 147 128 10.621 .715 15 . 191 125 12.977 .604 16 . 094 121 13.583 .630

Plot Symbols: Autocorrelations * Two Standard Error Limits Total cases: 34 Computable first lags: 33

Rainfall and its erosivity in Ethiopia 247

Fig. 58: Continued

Shola Gebeya: ACCs of long-term annual rainfall erosivity

AutO- Stand. Lag Corr. Err. -1 -.75 -.5 -.25 .25 .5 .75 Box-Ljung Prob. - 1 1 1 1 1 1 1 .033 .162 .042 .837 2 .136 . 160 .773 .679 3 - .063 . 157 * .933 .817 4 .082 .155 1.212 .876 5 -.070 . 152 * 1.423 .922 6 - .128 . 150 * * * 2.159 .904 7 - .117 .147 . ** 2.797 .903 8 - .074 .144 * 3.062 .930 9 -.100 .142 ** 3.565 .938 10 .026 .139 3. 599 .964 11 -.068 .136 3.845 .974 12 .000 .133 3.845 .986 13 -.161 . 130 * * * 5.370 .966 14 .017 .127 1 5.388 .980 15 .063 .124 5.647 .985 16 .049 .121 5.809 .990

Plot SymbolB: Autocorrelations * Two Standard Error Limits Total cases; 35 Computable first lags: 34

Shola Gebeya: ACCs of long-term summer rainfall erosivity

Auto- Stand. Lag Corr. Err. -1 -.75 -.5 - .25 ) * 25 .5 .75 1 Box-Ljung Prob. j 1 1 1 1 1 1 1 1 1 i 1 1 1 .038 .162 * .056 .812 2 .109 .160 ** .523 .770 3 -.050 .157 * .625 .891 4 .069 .155 * .823 .935 5 - .084 .152 ** - 1.127 .952 6 - .092 .150 ** 1.504 .959 7 -.096 .147 ** 1.929 .964 8 -.020 .144 1.949 .983 9 -.081 .142 . ** 2.274 .986 10 .018 .139 * 2.290 .994 11 -.091 .136 . ** • 2.737 . 994 12 - .018 .133 ■ 2.755 .997 13 -.189 .130 ^ **** 4.853 .978 14 -.026 .127 * 4 .895 .987 15 .028 .124 * 4 .947 .993 16 .071 .121 * 5.290 .994

Plot Symbols: Autocorrelations * Two Standard Error Limits . Total caBes: 35 Computable first lags: 34

Rainfall and its erosivity in Ethiopia 248

Fig. 58: Continued

Tuiu Bolo: ACCs of long-terrr, annual rainfall erosivity

Auto- Stand. Lag Corr. Err. -1 -.75 -.5 -.2 5 .25 .5 .75 1 Box-Ljung Prob. i 1 1 1 1 I 1 1 1 I i ! | 1 1 1 1 . 177 . 162 * * * * 1.187 .276 2 -.137 .160 * * ★ 1. 925 .382 3 -.161 .157 * * * 2 . 974 .396 4 . 137 . 155 * * * 3.754 .440 5 - .026 . 152 * 3.784 .581 6 - .232 . 150 6 .190 .402 7 -.088 . 147 * * 6.546 .478 8 .017 . 144 6.559 .585 9 -.1 2 9 . 142 * * * 7.392 .596 10 -.033 . 139 * 7.449 .682 11 .234 . 136 10.410 .494 12 .254 . 133 14.030 .299 13 .013 . 130 14 . 040 .371 14 - . 112 . 127 ★ * 14.820 .391 15 -.028 . 124 * 14 . 871 .461 16 . 026 . 121 * 14.917 .531

Plot Symbols: Autocorrelations * Two Standard Error Limits Total cases: 35 Computable first lags: 34

Tulu Bolo: ACCs of long-term summer rainfall erosivity

Auto- Stand.

1 . 151 . 162 . 866 .352 2 - . 184 . 160 2 .193 .334 3 - . 162 . 157 .257 .354 4 . 133 . 155 , 993 .407 5 - . 037 . 152 , 051 . 542 6 - .252 . 150 , 898 .330 7 - .082 . 147 .212 .407 8 - . 013 . 144 .220 .513 9 - . 186 .142 8 . 947 .442 10 -.041 .139 9. 036 . 529 11 .248 .136 12.352 .338 12 .311 .133 17.781 . 122 13 .005 . 130 17.783 . 166 14 - . 104 . 127 18.446 . 187 15 - .018 . 124 18.468 .239 16 .085 . 121 18.959 .271

Plot Symbols: Autocorrelations * Two Standard Error Limits Total cases: 35 Computable first lags: 34

Rainfall and its erosivity in Ethiopia 249

Fig. 58: Continued

Wonji: ACCs of long-term annual rainfall erosivity

Auto- Stand Lag C o rr. E rr. :-Ljung Prob.

1 - .401 . 164 5 . 962 .015 2 . 117 . 162 6.489 . 039 3 .070 .159 6.680 . 083 4 - .223 . 157 8.713 .069 5 - .075 . 154 8 . 953 . I l l 6 .201 .151 10.715 .098 7 - .285 . 149 14.399 .045 8 . 010 . 146 14.404 . 072 9 .201 . 143 16.374 .059 10 - .201 . 140 18.438 .048 11 . 160 . 137 19.796 .048 12 . 121 . 134 20.604 .056 13 - .212 . 131 23.213 .039 14 . 182 . 128 25.232 .032 15 - . 062 . 125 25.482 . 044 16 - . 172 . 121 27.495 .036

Plot Symbols Autocorrelations * Two Standard Error Limits T otal. c a se s: 34 Computable first lags: 33

Wonji: ACCs of long-term summer rainfall erosivity

Auto- Stand Lag C o rr. E r r . :-Ljung Prob.

1 - .326 . 164 3 . 938 . 047 2 . 077 . 162 4 . 166 . 125 3 .011 . 159 4 .171 . 244 4 - . 155 .157 5.149 .272 5 . 055 . 154 5.278 .383 6 . 191 . 151 6.870 . 333 7 - . 148 . 14 9 7 . 861 . 345 8 - . 019 . 146 7 . 879 .445 9 . 063 .143 8 . 072 . 527 10 - . 199 . 140 10.086 .433 11 . 107 . 137 10.701 .469 12 .076 . 134 11.024 .527 13 - .165 . 131 12.614 .478 14 . 133 . 128 13.695 .473 15 - . 120 . 125 14.625 .479 16 - . 132 . 121 15.808 .466

P lo t Sym bols: Autocorrelations * Two Standard Error Limits T o tal c a s e s : 34 Computable first lags: 33

Rainfall and its erosivity in Ethiopia ACKNOWLEDGEMENTS

Writing this dissertation was a great challenge for me. If it were not for the help of many people, nevertheless, this challenging work may have not been put into completion. Hence, I am very grateful for all the support given to me in various forms.

My first thanks go to the Friedrich-Ebert Foundation for sponsoring my doctoral research project in Bonn. Dr. Emst-J. Kerbusch, Director of the Department for International Co-operation and Dr. Ursula Mehrlander, Director of the Department for Labour and Social Research of the Foundation deserve special mention here for suggesting me for a scholarship award.

Most of all, I would like to sincerely acknowledge my primary supervisor, Prof. Dr. Armin Skowronek, for kindly accepting me as his advisee and for giving me a placement in the Institute for Soil Science of the Rheinische-Friedrich- Wilhelms-Universitat Bonn. I highly appreciate his constructive scientific as well as technical comments and suggestions, and above all his confidence in my ability to conduct research with a great degree of independence and self reliance. Working with him, I got outstanding opportunities to freely exercise my academic interests, and Prof. Dr. Armin Skowronek supported and encouraged me whenever necessary. It was a great pleasure for me to work with him; particularly the fact that he created the most attractive and conducive working environment for the scientists and students has continuously motivated me to achieve even more. Most of all I would like to thank Prof. Dr. Armin Skowronek for his dedications to the accomplishment of my project by supporting me in various ways and by sharing a competent and dedicated ear whenever I felt the need to review and discuss my research project.

My sincerest gratitude goes to my co-advisor, Dr. Petra Sauerborn from the University of Cologne, for showing a great and untiring interest in my project from the outset, and for her willingness to assist me. Most of all, I am grateful to Dr. Petra Sauerborn for closely advising me on research methods during the

Rainfall and its erosivity in Ethiopia years of my Ph D. study, and for recognising my capacity to conduct research on my own. Her commitment to bring my attention to the most relevant points is especially noteworthy. The numerous discussions we had are invaluable. Despite her own immense workload, Dr. Petra Sauerborn showed readiness to support me at all times, and her dedication to my progress was unparalleled. I strongly wish her a very successful life.

Special thanks go to Priv.-Doz. Dr.-Ing. A. Rieser from the Department of Water Engineering of the Rheinische-Friedrich-Wilhelms-Universitat Bonn, for his interest in my research since the initial phase and his positive response to my request when I needed his assistance.

I thank the Ethiopian offices which provided me with data necessary for this research. First of all, the Ethiopian National Meteorological Service Agency, for allowing me access to the National Climate Data Base. The Ministry of Water Resources and the Ministry of Agriculture deserve sincere acknowledgements for allowing me to use their archive materials and their library. In addition, I would like to thank the International Livestock Research Institute headquarter at Addis Ababa for permitting me to use its library and office facilities.

The names of friends and people who supported me in Ethiopia with the gathering of the data required for this research deserve special mention: Mrs. Almaz Demissie and Mr. Dula Shanko at the Ethiopian National Meteorological Service Agency, for kindly facilitating for me the rainfall data acquisition procedures; my former colleagues and friends, especially Mr. Mekonnen Ashenafi, at the Planning and Project Design Division of the Ministry of Water Resources, for assisting me with getting the precipitation and hydrometric data; my long-time friends Mr. Selamyihun Kidanu at the Debre Zeit Agricultural Research Centre, for his encouragement and organisation of transport as well as logistic necessary for the field work, Mr. Kedir Adam at the Oromia Bureau of Agricultural Development and Mr. Mesfin Tilaye at the Ethiopian Environmental Protection Authority, for assisting me with data gathering; Prof. Dr. Mesfin Abebe, a senior soil scientist at the Debre Zeit Agricultural Research

Rainfall and its erosivity in Ethiopia Centre, for the series of valuable discussions; and Mr. Tamiru Habte, Director of the Department of Natural Resources Management and Regulatory Department of the Ministry of Agriculture, for helping me with official procedures to get access to documents.

I wish to express my appreciation to all the Ph. D. students in the Institute for Soil Sciences, for creating a friendly working atmosphere; to the scientists and technical staff of the Institute, for the attractive working environment. Particularly worth mentioning is Priv.-Doz. Dr. Johannes Botscheck who was always ready to extend a helping hand whenever I needed it, not to forget all that he helped me deal with rough routines in and outside the Institute. Mr. Harindranath Bambarandage-Perera deserves special mention for his remarkable kindness and readiness to support all the times when I needed his help.

I express my deepest gratitude, which is beyond the capacity of w ords to say, to those close to my heart for sharing with me the moments of joy and sadness. Of especial worth mentioning here is BA. Nette.

I would like to sincerely dedicate this piece of work to my parents for giving me love and tenderness. Thus, I wish my parents good health, and long as well as happy life.

Rainfall and its erosivity in Ethiopia CURRICULUM VITAE

PERSONAL DTATA Full name: Mahdi OSMAN Date and place of birth: July 20, 1966, Harar, Ethiopia Nationality: Ethiopian Family status: Single Mother Mumina IBRAHIM Father Osman OMAR

HIGHER EDUCATION 01/1997-02/2001 Ph.D. candidate in Agricultural Science, University of Bonn, Germany (Dr. Agr.) 10/1994-12/1996 Studies of Agricultural Science and Resource Management in the Tropics and Subtropics, University of Bonn, Germany, (M Agr.) 04/1994-09/1994 Intensive German Language Course, University of Bonn, Germany (PNdS) 06/1996 Postgraduate Course on Legumes in Cropping Systems of the Tropics and Subtropics, University of Hohenheim, Germany, (.Postgraduate Certificate) 12/1992 Intensive Course on Resources Economics, Ethiopian Valleys Development Studies Authority Institutional Support Project and Richard Woodroofe & Associates, Addis Ababa, Ethiopia, (Certificate in Resource economics) 09/1984-07/1988 Studies of Agriculture and Agricultural Economics, Alemaya University of Agriculture, Alemaya, Ethiopia (B. Sc.) SCHOOL EDUCATION 09/1980-07/1984 Senior Secondary School / Grammar school, Alemaya College of Agriculture Evening Senior Secondary School and Harar Senior Secondary School, Hararge, Ethiopia (Ethiopian School leaving certificate examination) 09/1977-07/1980 Elementary and Junior Secondary Schools, Laga Harre and Bate Junior Secondary Schools, Hararge, Ethiopia 09/1974-07/1977 Arab Language School, Dire Dawa. Ethiopia

PROFESSIONAL AND SCIENTIFIC EMPLOYMENT 09/1988-04/1994 Economist, Socio-economic Researcher, Ethiopian Valleys Development Studies Authority, Ministry of Natural Resources, Addis Ababa, Ethiopia 01/1998-04/2001 Researcher in the Institute for soil science of the University of Bonn, Bonn, Germany

APPRENTICESHIP 02/1998-03/1998 Cypher and Information management officer, Embassy of the Federal Democratic Republic of Sri Lanka, Bonn, Germany 10/1995-02/1996 Laboratory and field methods practice in soil science, Institute for soil science of the University of Bonn, Germany. 07/1987-08/1987 Junior agricultural extension expert, Ministry of Agriculture, Hararge, Ethiopia

MAJOR SOCIAL DEDICATION 03/1984-07/1984 Teaching in National Literacy Campaign, Harar, Ethiopia Bonner Bodenkundliche Abhandlungen

Herausgeber: G.W. Briimmer. A. SkowTonek Schriftleitung: C. Klein Vertrieb: Institut fur Bodenkunde Nussallee 13, D -53115 Bonn Tel.: 0228 - 73 2780, Fax.: 0228 - 73 2782 e-mail: [email protected]

Bislang in der Schriftenreihe erschienen:

Band 1: Schollmayer, G. (1991): Untersuchungen zur Nitratmobilitat in sandigen Boden im Einzugsbereich der Wasserwerke Bocholt/Westfalen. 212 S., 81 Abb., 21 Tab., 1 Farbt. (vergriffen) Band 2: Hornburg, V. (1991): Untersuchungen zur Mobilitat und Verfugbarkeit von Cadmium, Zink. Mangan, Blei und Kupfer in Boden. 228 S., 30 Abb., 82 Tab. (vergriffen)

Band 3: Braun, P. (1991): Gefiigeverfestigung in Auenboden des Mittelrheins. 354 S., 204 Abb., 60 Tab. (vergriffen)

Band 4: Hiller, D.A. (1991): Elektronenmikrostrahlanalysen zur Erfassung der Schwermetallbindungsformen in Bdden unterschiedlicher Schwermetallbelastung. 172 S., 48 Abb., 50 Tab., 1 Farbt., 25,- DM Band 5: StrauB, R. (1992): Mechanismen der Phosphatbindung durch Goethit. Phosphat- adsorption und -diffusion in Abhangigkeit von der Goethitkristallinitat. 284 S., 77 Abb., 50 Tab., 30,- DM Band 6: Hennig, A. (1992): Vergleich verschiedener Methoden zur Berechnung und Simula­ tion des Bodenwasserhaushaltes - dargestellt am Beispiel von Auenboden bei Hennef / Sieg. 135 S., 44 Abb., 21 Tab., 20,- DM Band 7: Schlinkert, A. (1992): Jahreszeitliche Dynamik der Inhaltsstoffe von Bodenlosungen aus A-Horizonten unterschiedlicher Boden in Abhangigkeit von Bewirtschaftungsweise und Standorteigenschaften. 271 S., 54 Abb., 66 Tab., 30,- DM

Band 8: Veerhoff, M. (1992): Silicatverwitterung und Veranderungen des Tonmineralbestan- des in Waldboden als Folge von Versauerungsvorgangen. 249 S., 89 Abb., 12 Tab. (vergriffen) Band 9: Schobel, Th. (1993): Kennzeichnung der Stoffzusammensetzung tiefreichend humoser Sandboden im Westmiinsterland und Ermittlung von Kriterien zur bodentypologischen Abgrenzung. 260 S., 15 Abb, 77 Tab., 30,- DM Band 10: Heyder, D. (1993): Nitratverlagerung, Wasserhaushalt und Denitrifikationspotential in machtigen LoBdecken und einem Tonboden bei unterschiedlicher Bewirtschaftung. 171 S., 39 Abb., 40 Tab., Preis: 25,- DM Band 11: P o tra tz , KL-U. (1993): Bedeutung von Feuchte und Struktur der B odenoberflache fur die Bodenerosion. 72 S., 23 Abb., 10 Tab., 15,- DM Band 12: Weyer, Th. (1993): Kalkungsversuche mit carbonatisch und silicatisch gebundenen Kalk- und Magnesiumdungem - Initialeffekte auf versauerte Waldboden Nordrhein- Westfalens. 299 S., 60 Abb, 24 Tab., 30,- DM Band 13: Sauerborn, P. (1994): Die Erosivitat der Niederschlage in Deutschland - Ein Beitrag zur quantitative!! Prognose der Bodenerosion durch Wasser in Mitteleuropa. 189 S., 16 Abb., 45 Tab. (vergriffen)

Band 14: Tebaay, R. (1994): Untersuchungen zu Gehalten, zur mikrobiellen Toxizitat und zur Adsorption und Loslichkeit von PAKs und PCBs in verschiedenen Boden Nordrhein- Westfalens. 262 S., 79 Abb., 194 Tab., 30,- DM

Band 15: Schmidt, R. (1994): Porenentwicklung von Auftragsboden aus LoB im Rheinischen Braunkohlentagebau. 102 S .19 Abb., 23 Tab., 20,- DM

Band 16: Denkler, M. (1994): Mikrobieile Toxizitat, Abbau und Verlagerung ausgewahiter Pflanzenschutzmittel in Ackerboden aus LOB: Labor- und Freilanduntersuchungen sowie Modellrechnungen. 229 S., 55 Abb., 60 Tab., 30,- DM Band 17: Zeien, H. (1995): Chemische Extraktionen zur Bestimmung der Bindungsformen von Schwermetallen in Boden. 284 S., 35 Abb., 89 Tab., 30,- DM Band 18: Klimsa, K. (1996): Sorption, Verlagerung und Abbau von ausgewahlten Pflanzen- schutzmittel-Wirkstoffen in Boden unterschiedlichen StotYbestandes mit und ohne Mulch. 226 S., 64 Abb., 66 Tab., 30,- DM

Band 19: Heusch, K- (1997): Auenboden im fluvialen Stoffsystem - dargestellt am Beispiel der Siegaue bei Hennef. 182 S., 46 Abb., 23 Tab., 25.- DM Band 20: Schluter, W. (1997): Kationen- und Anionen-Gehalte in der Bodenlosung und im oberflachennahen Grundwasser von organisch und konventionell bewirtschafteten Ackerflachen der Siegaue sowie Simulation der StickstofTdynamik und Nitrat-Verla- gerung. 203 S., 33 Abb., 74 Tab., 30,- DM Band 21: Kehl, M. (1997): Experimentelle Laboruntersuchungen zur Dynamik der Wasser­ erosion verschieden texturierter Ackerboden Nordrhein-Westfalens. 120 S., 38 Abb., 36 Tab., 20,- DM Band 22: Thiele, S. (1997): PAK in belasteten Boden: Untersuchungen zur Charakterisierung, Optimierung und Prognose des mikrobiellen PAK-Abbaus sowie zur Abschatzung der filr Pflanzen verfugbaren PAK-Fraktion. 318 S., 47 Abb., 105 Tab., 30,- DM Band 23: Patzold, S. (1998): Herbizidanwendung im Obstbau - Messung und Simulation des Abbau-, Sorptions- und Verlagerungsverhaltens von Simazin und Diuron in LoB- boden. 231 S., 41 Abb., 55 Tab., 30,- DM Band 24 Everding, C. (1998): Kennzeichnung des Erosionsverhaltens und der Erosionsan- falligkeit verschieden texturierter Ackerboden Nordrhein-Westfalens mit Hilfe von Feldberegnungen. 195 S., 24 Abb., 58 Tab., 25,- DM Band 25 GeiBen, V., A. Schoning, G.W. Briimmer (1999): Entwicklung eines Sanierungs- konzeptes fur stark versauerte Waldboden - dargestellt am Beispiel des Kottenforstes bei Bonn. 135 S., 43 Abb. 24 Tab., 20,- DM Band 26 Gunster, N. (1999): Palaopedologische Untersuchungen an pliozanen und pleisto- zanen Sedimenten im Becken von Granada/Siidspanien. 238 S., 34 Abb., 49 Tab., 30,- DM Band 27 Keppler, J, (1999): Modellversuche und Gelandeuntersuchungen zur Kinetik der Schwermetall-Festlegung sowie zur Loslichkeit, Mobilisierung und Verlagerung von Schwermetallen in belasteten Boden. 252 S.. 38 Abb.. 50 Tab., 30,- DM Band 28 Liebe, F. (1999): Spurenelemente in Boden und Pflanzen Nordrhein-Westfalens - Gehalte verschiedener chemischer Fraktionen in Boden und deren Beziehung zur Bodenreaktion und den Gehalten in Pflanzen. 375 S. 206 Abb., 154 Tab., 35,- DM Band 29 Botschek, J. (1999): Zum Bodenerosionspotential von Oberflachen- und Zwischen- abfluB. Habil.-Schrift, 174 S., 46 Abb.. 29 Tab., 40,- DM Band 30 Fischer, L. (2000): Untersuchungen zur Ad-/Desorption und Diffusion von Schwer- metallen und Aluminium mit Goethiten unterschiedlicher Kristallinitat - Experimen­ te d Ergebnisse und Prozessmodellierung. 205 S., 39 Abb., 17 Tab., 25,- DM Band 31 GeiBcn, V. (2000): Reaktionen ausgewahlter Tiergruppen (Lumbricidae. Annelidae; Collembola. Arthropoda) auf Veranderungen der Nahr- und Schadstoffgehalte von Waldboden nach Kalkung und Dungung. Habil.-Schrift, 382 S., 114 Tab., 40 Abb., 35,- DM Band 32 Welp, G. (2000): Schwermetalle und organische Schadstoffe in Boden: Gehalte - Sorption und Loslichkeit - Wirkung auf Mikroorganismen. Habil.-Schrift, 169 S., 34 Abb., 11 Tab., 20,- DM

Band 33 Spang, G. (2000): Messung und Modellierung von Sorptions- und Diffusions- prozessen in einem LoBboden - Batch- und Saulenperkolationsversuche mit Stoffen unterschiedlicher Matrixaffinitat. 272 S.. 59 Abb.. 82 Tab., 30.- DM Band 34 Schoning, A. (2000): EinfluB chemischer Eigenschaften von W'aldboden auf deren Besiedlung mit Lumbriciden - Konsequenzen fur die Bodenschutzkalkung. 343 S., 37 Abb., 38 Tab., 30,- DM Band 35 Hamer, M. (2000): Saureneutralisationskapazitat und Mechanismen der Protonen- pufferung definierter Minerale sowie ausgewahlter Gesteine und Boden Nordrhein- Westfalens. 282 S., 61 Abb., 54 Tab., 30,- DM Band 36 Wolf, A. (2001): Gewasserschonende Landbewirtschaftung in einem Wasserschutzgebiet nordwestlich von Koln unter Beriicksichtigung typischer Ackerboden und Feldfriichte. 161 S., 37 Abb., 44 Tab.