Simulating Agricultural Non-Point Source Pollution From A Typical Agricultural Small Watershed Of Three Gorges Reservoir Area Based On SWAT Model

As the effective control of point source pollution in china, the effect of Non-point source pollution on the ecological environment is more and more important. Three Georges Reservoir Area is not only an important water resource conservation zone but also a priority for all in the ecological environment management of middle and upper reaches of Yangtse River. And the Non-point source pollution of local agriculture is acute. Above all. the research on agricultural Non-point source pollution in Three Gorges Reservoir Area has a great significance on the protection of regional water quality. This thesis is going to choose Changling Country Chenjiagou watershed in Wanzhou District, Chongqing which locates on the central part of Three Gorges Reservoir Region as research object. Based on SWAT model, this thesis will set up a Non-point source pollution database, then Using the measured data of basin runoff, sediment, total nitrogen and total phosphorus, it will do sensitivity analysis, calibration and verification to the model parameter. Finally, this thesis will do quantitative analysis to the space-time distribution features of agricultural Non-point source pollution in Chenjiagou watershed and the influence of different fertilization situations on Non-point source pollution load.Here are the research results.(1) The establishment of SWAT model database. Under the circumstance of geographic information system environment, this thesis recodes ihe Land Use Type Graph and Soil Type Graph, unifies the projection and coordinates of DEM Graph, Land Use Type Graph and Soil Type Graph; according to the requirements of SWAT model to data, it processes the physical/chemical properties of soil, meteorological parameters, crop management practices and so on; the research confirms the catchment area threshold as 15ha. and divides the whole watershed into 25 sub-basins; through the overlay of DEM Graph, Land Use Type Graph and Soil Type Graph, it furtherly gets 281 hydro]ogical response units.(2) Through the application of LH-OAT sensitivity analysis, confirming parameters which are more sensitive to the model. Via a parameter sensitivity analysis to runoff. sediment, total nitrogen and total phosphorus respectively, it eventually gets 17 parameters which are relatively more sensitive to model.simulation, which consists of the SCS runoff curve number, base flow alpha factor, deep aquifer permeability than compensating coefficient, soil evaporation, canopy entrapment quantity biggest, soil saturated hydraulic conductivity, shallow eroundwater runoff coefficient, the effective hydraulic conductivity of main channel and main channel of the effective hydraulic conductivity, soil depth, soil available water, sediment was linear index, the main channel to carry manning coefficient, surface runoff coefficient of lag, USLE equation keep measures the power factor, power index of sediment was renewedly carried, and nitrogen permeability coefficient. This thesis will calibrate and verify the model basing on these 17 parameters.(3) With pervasive applicability, the SWAT model can effectively simulate Non-point source pollutions of the typical small watershed in The Three Gorges Reservoir Area.The results of calibration and verification show that R2 and Ens for calibration period and verification period of the runoff simulation values 0.90,0.84 and 0.82,0.75 respectively, R2 and Ens for calibration period and verification period of the sediment simulation values 0.78,0.60 and 0.88,0.73 respectively, R2 and Ens for calibration period and verification period of the total nitrogen simulation values 0.78,0.71 and 0.71,0.65 respectively, R2and Ens for calibration period and verification period of the total phosphorus simulation values 0.65,0.72 and 0.55,0.61 respectively, which mean all of the R2 and Ens values can meet the requirements of regulations, and it has applicability to simulate the monthly runoff, sediment, total nitrogen, total phosphorus load in Chengjiagou watershed.(4) In Chenjiagou watershed the annual average of sediment yield is 4763.41t, the total nitrogen output 8.27t and the total phosphorus output 0.79t. From the point of time distribution, it will analyze the inter-annual variation and intra-annual variation:because of the only four years of data, inter-annual varies randomly, but from the trend line we can get the pollution output decreases as the decrease of rainfall; from the point of intra-annual variation, non-point-resource pollution mainly happens from July to September, while in low-rainfall period, the outputs of sediment, total nitrogen and total phosphorus are relatively low. In a word, no matter from inter-annual and intra-annual variation, the sediment, total nitrogen and total phosphorus load show a consistent trend. From the point of spatial distribution, Non-point source pollution is relatively more serious in outlet of water, two catchment area in the central watershed and north high level area, which can be chosen as key areas of Non-point source pollution.(5) The types of land use which have higher Non-point source pollution load per unit are dry land and paddy field. There is a kind of positive correlation relationship between Non-point source pollution and fertilizer levels; if the original fertilizer level can be decreased 50%, the total nitrogen load can decrease 21.13% and the total phosphorus load can reduce 27.91%. In the future, we can pay more attention to the influence of fertilizer on crop output in order to make the most reasonable fertilizer level, which can ensure crop output and reduce Non-point source pollution effectively at the same time.