Modelling of Non-Point Source Pollution Around Lake Victoria Using SWAT Model: A Case of Simiyu Catchment Tanzania

Pollutant loading in Lake Victoria is resulting from a wide range of anthropogenic activities. This study fo- cused on Non-Point Sources (NPS) of pollution and it was carried out to identify and characterize land-use activities and to quantify the sediment and nutrient loads (nitrogen and phosphorus). The study was conducted in Simiyu catchment of Lake Victoria using land-use data of 1975 and 2006 and comparing the relative impact of land-use change on sediment and nutrient load (P and N) into the lake. Possible best management practices were also identified for those sub-basins with the highest pollution yield. Remote sensing using the package ILIWIS 3.0 was used to identify and characterize the land-use and the Soil and Water Assessment Tool (SWAT) model was used to quantify sediment and nutrient load from these two different land-use scenarios. Land use classification according to the SWAT model shows that Agricultural Land-Generic (AGRL) contributes about 73.43%, Range-brush (RNGB) contributes 24.42%, Pasture (PAST), 2.10% Sa- vanna (SAVA) 0.03% and Water (WATR) 0.02% of the total catchment area of Simiyu. It was also found out that there was an expansion of agricultural land from covering 19.33% of the catchment to 73.43% at an annual change rate of 2.9%. However, average Nitrate load was higher for 1975 than 2006. The P load of 1975 was less compared to that in 2006. Model simulation at the catchment outlet for N gave 77.2 kg/km 2 .yr while observed values were 146 kg/km 2 .yr, simulated P was 47 kg/km 2 .yr while observed was 164 kg/km 2 .yr. Hence, the model underestimated nutrient yield in the catchment. Therefore, the applicability of the SWAT modelling tool in studying NPS pollution yields poor model per- formance due to the scantiness of data used for model calibration. More rigorous data campaigns have to be carried out along the two rivers of Duma and Simiyu for purposes of gaining enough information for model calibration and valida- tion. With good model performance, developing management plans to control NPS pollution around Lake Victoria could be achieved using the SWAT model.