Potential impact of climate and land use changes on the water resources of the Upper Blue Nile Basin

Abstract

The upper Blue Nile basin is the major crop production area of Ethiopia with about 6 million hectares of cultivated land. It supplies surplus food to the national market. The basin contributes about 763 MW of hydroelectricity to the national grid at present. Additional 6000 MW of hydroelectricity is expected from the renaissance dam in the near future. The major effect of climate change in the basin is through its effect on crop production and hydropower generation. The droughts of the 1960s, 70s and 80s have led to crop failure in the basin and the country at large and led to famine and death of several people and livestock. The recent droughts in 2008 and 2009 reduced water levels at many dams and caused below normal power generation and frequent blackouts. Expansion of large scale irrigation, upland area afforestation and expansion of private large scale irrigation schemes through foreign direct investment are the major land use changes expected. The objective of this study was to study the potential impact of climate and land use changes on the discharge of the upper Blue Nile River using climate change and formulated future land use change scenarios. Statistical downscaling of climate change scenarios, bias correction of regionally downscaled climate change scenarios and hydrological modeling were used in this study. Statistical downscaling approach was used to downscale monthly rainfall at five stations in the basin using six downscaling domains. Predictors were synthesized based on correlation analysis between large scale climate predictors and observed station data (rainfall). Area averaged monthly predictors were used to establish a regression model between the predictors and observed rainfall from the five meteorological stations. The regression models were validated against observed station data and were used to generate downscaled future rainfall. Data for future scenario were extracted from the ECHAM5 A1B run. Regional climate change projections of ECHAM5-A1B downscaled by the REMO and CCLM models were obtained and compared to observed climate. Raw REMO data was bias corrected using a probabilistic approach called CDF-transform. Bias corrected CCLM data was obtained from the Potsdam Institute for Climate Impact Research. Difference and ratios were used to bias correct temperature and rainfall. A future time land use change scenarios was developed in consultation with soil and water experts, review of project documents, discussion with small holder farmers, considering foreign direct investment and past trends. The Soil and Water Assessment Tool (SWAT2009) was used to study the hydrological impact of climate and land use change. The model was calibrated and validated using measured stream flow measured at El Diem. The impacts for two future periods (i.e. 2041-2070 and 2071-2100) were obtained by running the calibrated model with climate and land use change scenarios. The results of statistical downscaling of rainfall data showed a decrease in rainfall by 6-12% during the short rainy season in the basin but mixed results in the main rainy season. Future research in downscaling rainfall should consider different IPCC emission scenarios and use of multiple GCMs. In addition the effect of domain size on model performance, uncertainties in the statistical downscaling and the data used should be evaluated. Average annual projected runoff changes for the basin from the CCLM were 9.2% and -10.0% relative to the historical flow for 2041-2070 and 2071-2100 respectively. It showed consistent increase by 102.8% and 141.4% during the same period for REMO. Although there is some agreement on the direction of future runoff change, the discrepancy in the magnitude of change between the two downscaling models highlights the need for better understanding the uncertainties in downscaling. There is a need for further studies to better understand the reasons for contrasting results in the regional models in simulating rainfall. Projected future land use change in the basin resulted in a decline in runoff by 1%. As the large size and complex hydrology of the basin might mask the effect on the runoff, it is recommended to carry out calibration of the hydrological model at sub basin level in order to capture the change in runoff as a result of land use change in the basin.

DER POTENTIELLE EINFLUSS VON KLIMA- UND LANDNUTZUNGSWANDEL AUF DIE WASSERRESSOURCEN IM OBEREN BECKEN DES BLAUEN NILS Das Obere Becken des Blauen Nils ist mit ca. sechs Millionen Hektar Anbaufläche Äthiopiens Hauptanbaugebiet für landwirtschaftliche Produkte. Der hier erzielte Produktionsüberschuss wird in den nationalen Markt eingebracht. Die Wasserkraftwerke des Beckens speisen bereits jetzt ca. 763 MW in das nationale Stromnetz ein; der im Bau befindliche Renaissance-Damm soll zukünftig weitere 6000 MW an hydroelektrischer Energie liefern. Der Klimawandel wirkt sich im Becken insbesondere auf die Landwirtschaft und die Wasserkraftgewinnung aus. Die Trockenperioden der 1960er, 70er und 80er haben zu Ernteausfällen, Hungersnöten und dem Tod zahlreicher Menschen und Tiere geführt. Die Dürren der Jahre 2008 und 2009 hatten niedrige Pegelstände vieler Reservoirs zur Folge und damit einhergehend eine unterdurchschnittliche Energiegewinnung bis hin zu Stromausfällen. Die zunehmende Ausbreitung von großflächigen staatlichen und privaten Bewässerungssystemen, finanziert durch direkte Auslandsinvestitionen, sowie die Aufforstung von Hochlandbereichen sind die wichtigsten zu erwartenden Landnutzungsänderungen. Das Ziel der vorliegenden Untersuchung ist es, ausgehend von modellierten Szenarien, den potentiellen Einfluss von Klima- und Landnutzungswandel auf den Abfluss des Oberlaufs des Blauen Nils zu untersuchen. In der vorliegenden Studie wurden Daten des globalen Klimawandelmodells ECHAM5 (Emissionsszenario A1B) zunächst durch downscaling auf regionale Maßstäbe heruntergerechnet und anschließend biaskorrigiert. Darauf basierend wurde eine hydrologische Modellierung durchgeführt. Der downscaling-Ansatz unter Verwendung von sechs downscaling domains wurde zur Berechnung der zukünftigen monatlichen Niederschläge von fünf bestehenden Messstationen im Oberen Becken des Blauen Nils herangezogen. Ein kombinierter Prädiktor wurde erstellt, resultierend aus einer Korrelationsanalyse zwischen den großmaßstäblichen Klimaprädiktoren des ECHAM5-Models und den regionalen Niederschlagsmesswerten der fünf meteorologischen Stationen. Flächengemittelte monatliche Prädiktoren fanden bei der Erstellung eines Regressionsmodells zwischen den Prädiktoren und dem an den fünf Messstationen gemessenen Niederschlag Anwendung. Die Regressionsmodelle wurden mit Messdaten validiert und durch downscaling für die Berechnung des zukünftigen Niederschlags verwendet.