Non-point Source Pollution Study In The Longxi Rriver Watershed Based On SWAT Model

With the rapid development of society and economy,the pollution is becoming more and more serious,and the available water is decreasing.Water pollution has become a global problem.widely space-time distribution,large spatial differences,complex sources of pollution,difficult to capture processes,and hardly to control,are the characteristics of non-point source pollution.As the first grade tributary of the Yangtze River,Longxi River is an important source of water in Liangping County,Dianjiang county and Changshou county.It is also an important secondary river in the Three Gorges Reservoir area.It shows great significance to study.Therefore,this study selected the longxi River as the research area,the total nitrogen(TN)and total phosphorus(TP)pollution loads were taken as the research object.This paper use Soil and Water Assessment Tools(SWAT)of hydrological model to construct the non-point source pollution simulation of the Longxihe River,and analysis the character of the TN and TP pollution load from temporal and spatial variation dimensions.And then,based on the landscape pattern theories,this study use mathematical statistics,landscape pattern index analysis and redundancy analysis(RDA)to identify non point source pollution load process comprehensively.At last,this study use the Landscape heterogeneity to recognize the Non-point source pollution.The results are as follows:1)TN and TP of Longxi River showed significantly different temporal and spatial variation.The load in the flood period is significantly higher than low water period.The load of the main channel is larger than that of the tributary channel,and the downstream load is much larger than the upstream load.2)The response of the RICE-leading watershed,AGRL-leading watershed and FRST-leading watershed was different,and FRST-leading watershed has a certainreduction in the non-point source pollution,while AGRL-leading watershed has a higher contribution to non-point source pollution.3)LPI has high recognition ability for non-point source pollution load in RICE-leading watershed.ED,LSI and AI have better identification of non-point source pollution load in AGRL-leading watershed,and PD,LSI and AI are better for identifying pollution in the FRST-leading watershed.4)RDA shows that water land,industrial land,rural residential land and woodland are the strongest in the landscape type.The response of between water land and non-point source pollution load is the most sensitive and mutual expression,the larger the industrial land and rural residential land is,the lower non-point source pollution is.For the landscape index,the most obviously index response to the non-point source pollution load in the watershed is the Fragmentation/Aggregation,the Heterogeneity/Dominance,and different types of watershed has different response manners.