Abstract

Sediments resulting from soil erosion are the major non-point source pollutant of surface waters in agricultural watersheds. Annualized Agricultural Non-Point Source Pollution Model (AnnAGNPS) is a computer-based watershed model that predicts non-point source pollutants and runoff loadings within agricultural watersheds. AnnAGNPS v5.2 was used in conjunction with Arcview 3.2 GIS to predict streamflow and sediment discharges from a 250.89 ha agricultural watershed, Upper Ebonyi River watershed, located at Obollo-Etiti in Udenu Local Government Area in Enugu State, South-eastern Nigeria. AnnAGNPS predictions were compared with two months (September and October) of streamflow and sediment discharge field measurements from the study watershed. The September data were used to calibrate the model to a reasonable agreement with predicted data (R2 = 0.934 for streamflow and R2 = 0.707 for sediment yield). Statistical performance evaluation of the model was carried out on the validation results. The model performed very well in following the trends, peaks and volumes of the measured hydrograph and sediment graph, with R2 = 0.991 for hydrograph and R2 = 0.968 for sediment graph. The results show that the model performed better in predicting runoff than predicting sediment yield. Further, the model was applied to identify erosion hot spots in the study watershed. The model gave a sediment delivery ratio (SDR) of 0.419 for the watershed. The effects of different land uses (fallow, cropland, forest, pasture, rangeland and tillage) on runoff volume, soil erosion and sediment yield were modeled. The results show that fallow land use gave the highest runoff volume, soil erosion and sediment yield followed by tillage land use while forest land use gave the least runoff volume, soil erosion and sediment yield. Therefore, forest land use should be predominantly practiced to reduce sediment loss on the watershed. More conservation practices should be adopted at the landscape to reduce sediment loss for fallow and tillage land uses. The application of AnnAGNPS on this experimental watershed demonstrated that the model has considerable potential as a research and management tool for comparative assessment, long-term monthly and average annual estimation of runoff and sediment yields, identification of erosion hot spots, and estimation of sediment delivery ratio. This prediction will be useful to decision makers and farmers for effective watershed management.