SWAT Model Based Analysis On N/P Loss Loadings From A Mountainous Watershed

SWAT (Soil and Water Assessment Tool) is a physically based model and a combination of hydrological model, soil erosion model and pollutant migration model. Using weather, soil properties, land use and other specific information, SWAT can simulate water movement, sediment movement, crop growth, nutrient cycling, etc. In this research, we studied the characteristics of non-point source pollution in East Tiao river watershed using SWAT model. The dominant land use types are forest and paddy land in the study area, accounted for58%and27%of the total area, respectively. Loam and silty clay loam are domain soil texture types. In the process of modeling, the position of Pingyao hydrologic station was set as watershed outlet and there was a reservoir located at Linan, Hangzhou. At last,24sub-basin and438HRUs was defined. Measured data of hydrology and water quality of2008-2012was used to calibration and verification.36parameters were modified. The result showed that the calibrated model had a good effect on flow, TP flux, TN flux and NH4+-N flux simulation. In calibration period, the coefficients of determination (R2) of the four indexes were0.96,0.84,0.84and0.73, respectively. The Nash-Sutcliffe coefficients (ENs) were0.83,0.83,0.81and0.60, respectively. In verification period, R2were0.98,0.89,0.81and0.85. ENS were0.82,0.86,0.69and0.80, respectively.SWAT output showed that there were significant special and temporal characteristics of N\P loss in study area. The regression analysis of monthly precipitation and TN, TP loss in sub-basin scale showed that nutrient loss were significant positive correlation with the monthly rainfall (all p-values were smaller than0.01). The coefficients of determination (R2) of precipitation-TN loss regression equations were ranged from0.59to0.85, while that of precipitation-TP loss were0.55-0.75.The results of multiple regression analysis showed the difference of nutrient loss between forest and paddy land. TN loss per unit area of paddy land was bigger than that of forest under same rainfall. This law was also real for TP loss characteristics when rainfall over89mm. The results indicated that non-point source pollution management of paddy land was more urgent.