Main content area

Watershed modeling for reducing future non-point source sediment and phosphorus load in the Lake Tana Basin, Ethiopia

Moges, MamaruA., Schmitter, Petra, Tilahun, SeifuA., Steenhuis, TammoS.
Journal of soils and sediments 2018 v.18 no.1 pp. 309-322
basins, best management practices, climate, developing countries, highlands, hydrograph, hydrologic models, hydrometeorology, intensive farming, lakes, model validation, monsoon season, phosphorus, pollutants, pollution load, prediction, runoff, sediment transport, sediment yield, sediments, statistical models, topography, water quality, watershed management, watersheds, Ethiopia
PURPOSE: Agricultural intensification to meet the food needs of the rapidly growing population in developing countries affects water quality. In regions such as the Lake Tana basin, knowledge is lacking on measures to reduce non-point source pollutants in humid tropical monsoon climates. The aim of this paper was, therefore, to develop a non-point model that can predict the placement of practices to reduce the transport of sediment and phosphorus (P) in a (sub) humid watershed. MATERIALS AND METHODS: In order to achieve the objective, hydrometeorological, sediment, and P data were collected in the watershed since 2014. The parameter efficient semi-distributed watershed model (PED-WM) was calibrated and validated in the Ethiopian highlands to simulate runoff and associated sediments generated through saturation excess. The P module added to PED-WM was used to predict dissolved (DP) and particulate P (PP) loads aside from discharge and sediment loads of the 700 ha of the Awramba watershed of Lake Tana basin. The PED-WM modules were evaluated using the statistical model performance measuring techniques. The model parameter based prediction of source areas for the non-point source sediment and P was also evaluated spatially and compared with the Topographic Wetness Index (TWI) of the watershed. RESULTS AND DISCUSSION: The water balance component of the non-point source model performed well in predicting discharge, sediment, DP, and PP with NSE of 0.7, 0.65, 0.65, and 0.63, respectively. In addition, the predicted discharge followed the hydrograph with insignificant deviation from its pattern due to seasonality. The model predicted a sediment yield of 28.2 t ha⁻¹ year⁻¹ and P yield of 9.2 kg ha⁻¹ year⁻¹ from Awrmaba. Furthermore, non-point source areas contributed to 2.7 kg ha⁻¹ year⁻¹ (29%) of DP at the outlet. The main runoff and sediment source areas identified using PED-WM were the periodically saturated runoff areas. These saturated areas were also the main source for DP and PP transport in the catchment. CONCLUSIONS: Using the PED-WM with the P module enables the identification of the source areas as well as the prediction of P and sediment loading which yields valuable information for watershed management and placement of best management practices.