Modeling of runoff pollution load in a data scarce situation using swat, Sondu watershed, Lake Victoria Basin

Abstract

Sondu watershed is one of the major rivers flowing to Winam Gulf on the Kenyan side of Lake Victoria. The lake is a source of fish and freshwater supply and also provides routes for transportation. Lake deterioration is being driven by excessive pollution load: sediments (TSS) and nutrients (TN & TP). Previous assessments of run-off load in the basin used simple unit load methods because of lack of data and there is still need for comprehensive information on lake pollution and deterioration. Monitoring of weather and water quality data in the watershed is still low. Incorporation of Geographical Information System (GIS) and remote sensing based tools are options to improve data assembly and pollution load assessment. This study used Soil Water Assessment Tool (SWAT) to simulate temporal-spatial distribution of surface water runoff (river flow), sediment and nutrient generation in Sondu watershed, and to identify soil erosion and nutrient source hot spots. Annual sediment generation to the lake is 80,000 t/yr composed of mainly silt while TN & TP are 3,388 t/yr & 312 t/yr, respectively. Sediment and nutrients are mainly generated from agricultural crop areas at downstream, central (Sondu) and upstream West (lower Kisii/Nyamira) area of the watershed while the high water runoff yielding areas are upstream (Kericho/Kisii/Kericho) areas. The nutrients peak periods lagged behind sediments’ by one to two months on average. Consideration of afforestation and adoption of best management practices such as filters on crop areas would be useful in reduction of pollution load to the lake and land conservation.

 Key Words: Lake Victoria, Sondu watershed, Pollution load, SWAT model