Water Quality Modelling of the Sosiani River Catchment in Western Kenya- applying SWAT model in a data poor African catchment

The study assessed and monitored pollution loading in the Sosiani River Catchment (369 km2) in Eldoret City which drains into Lake Victoria, the largest freshwater body in Africa. The study involved stream flow measurements, water quality tests and using SWAT (version 2012) model to simulate catchment hydrology and in-stream water quality. The model was setup, parametrized, calibrated, validated and simulations done. The existing soil map divided the study site into two areas, the Upper Plateau and Lower Plateau, both having two major soil types. This was a simplification of the catena in the area and an under-representation of slope positions such as shoulders, backslopes, footslopes, and toeslopes present in the study area which has an impact on the distribution of the available soil moisture contents in the catchment. The digital soil mapping (DSM) methods using k-means (KM) clustering and fuzzy logic approaches were utilized to obtain soil types with their slope positions taken into consideration. The enhanced soil maps were spatially detailed and better captured the soil-landscape relationships compared to existing soil maps. The novelty of the study was the determination of the available soil water content for the whole catchment based on the soil-landscape relationships. This was utilized in the modelling of runoff and water quality in the data-scarce Sosiani River Catchment with an aim of reverting trends of water quality degradation.