Predicting river water quality in rural area using a watershed scale model

River water is a major source of agricultural water uses, but the quality of it is not always suitable for the agricultural purpose. Among others, agriculture is considered as a major contributor of nitrogen (N) pollution of river water and is expected to control N emission. However, the mechanism of water pollution by the agriculture is not sufficiently understood. In order to understand the mechanism of controlling N levels in river water, N loadings on watersheds were estimated using the modified pollutant loading unit method and municipal based statistical data by using a geographic information system (GIS). In the case of a less agricultural watershed, the estimated total N loading well reflected the water quality of the river and proved itself as a useful indicator for predicting water quality. However, the estimation was not enough to predict water quality in a more agricultural watershed. Then a simulation model was developed by incorporating both effects of denitrification in the rice paddy and the storage of N in the surface soil in order to explain the apparent loss of N loaded on the watershed. The present model sufficiently explained the river water quality, although uncertainties and possible modification were also discussed.