Modeling the effects of crop patterns and management scenarios on N and P loads to surface water and groundwater in a semi-humid catchment (West Africa)

Assessment studies of N and P loads to water systems have developed from simple descriptions based on monitoring and sampling into the analysis of multiple scenarios using simulation models. In this study, water, sediment, and nutrient delivery to the stream flow at the Donga-Pont river catchment outlet (586 km(2)) in the Republic of Benin, West Africa were simulated incorporating local management practices including detailed crop systems, fertilization and manure deposition, using the Soil and Water Assessment Tool (SWAT). At the study scale (586 km2), cropping systems are not mapped. Therefore, a land use map was refined for the study area, enabling the evaluation of four management scenarios: crop systems with fertilizer supplied only to cotton, rice and maize, as is common in Benin (Sc0); crop systems without the use of fertilizer (Sc0a); the original land use map with similar fertilizer inputs to all cropping systems (Sc1); and the original land use map without fertilizer inputs (Sc1a). Compared to the first scenario, the latter two scenarios, commonly used in regional scale modeling, exhibited distinct biases in plant growth parameters, crop yields, water yield, sediment yield and nitrogen load. Finally, it was apparent that at the catchment scale, decreases in water yield and nutrient loading were induced by reductions in rainfall (as the result of climate change scenarios), but the effects of the decline in precipitation were counteracted by the effects of changes in land use (land use scenarios). This indicates the strength of the relationship between agriculture and water quality (sediment and nutrient loads) within the Donga-Pont catchment. It was clear that management practices such as fertilizer inputs are among the principal factors controlling this dynamic. Moreover, high spatial variation in the groundwater nitrate concentration, reaching 42 mg/l, was observed. (c) 2012 Elsevier B.V. All rights reserved.