| In this paper,on the one hand,GIS is used to digitize rainfall,runoff,soil,topography,vegetation and other factors(soil erodibility)in different typical years in the study area,and the weight is analyzed by normalization and SPSS,so as to establish the grid calculation equation of water input coefficient of TN,TP,NH3N and COD load.Based on the statistical data of each typical year in the study area,the fitting load of agricultural non-point sources into water body of the improved output coefficient model of point sources was calculated.On the other hand,the daily flow data and monthly average LOAD data of wangjiaba,Bengbu Gate and Hengpaitou section in 2015 were used to fit the LOAD of agricultural non-point sources into water body based on monitoring data using LOADEST(LOAD ESTIMATOR)model.Based on the control section,the small watershed was divided by using the hydrological zoning function of SWAT to verify the scientific nature of the improved output coefficient model of non-point source at the small watershed level.Finally,the overall change trend of contribution rate of agricultural non-point source loads of each district and county in the study region along with time scale was calculated by the proportion of pollution load sources in the calculation results,and the spatial sensitivity of various loads in the study region to the study region was calculated based on the sources of pollution loads and intensity of various pollution loads.The following conclusions can be drawn from the above research contents:(1)According to GIS and SPSS,the average discontinuous values of TN and NH3N inflow coefficients in the study area are 0.419,0.411,0.423,0.463 and 0.394,0.396,0.429 and 0.447,respectively,through the grid calculation equation.Basically,it has a great correlation with the distribution of annual rainfall in the study area.The average discontinuous values of TP and COD into water body are 0.31,0.35,0.396,0.41 and 0.391,0.376,0.446,0.442,respectively.(2)Based on LOADEST model,the loading of TN,TP,NH3N and COD of WangJiaBa and BengBuZha into water body from non-point source is 4581.95 t,231.12 t,372.12 t,27291.26 t,1507.67 t,49.05 t,67.96 t and 11141.95t,respectively.Compared with the TN load in the small watershed,due to its relatively large cardinal number,the relative error range of the factor is small,which is 13.32%.The calculation error of NH3N load is 61.88%,the obvious error value of COD load is too large compared with other load types,and the relative error is 79.87%,while the TP error is relatively small,which is 21.3%.he annual TN,TP,NH3N and COD inflow into water from non-point sources are 262.15,16.28,46.89 and 3812.71t,respectively.The relative errors of TN and NH3N in small watershed are 64.16%and 13.01%,respectively.However,due to the relatively small cardinality,the error of TP estimation load is more than 110%,but the weight error among various factors is considered during the construction of the surface source model,so it meets the verification requirements.The relative error of COD load simulation is 94.45%,so it is considered to add conversion factor or reduction factor in the follow-up study to improve the COD non-point source model.(3)The grid graph of river inflow coefficient of the study area was extracted by county using GIS,and the fitting calculation of various loads was carried out by using the improved output coefficient method of surface source model.The results show that:1)The spatial distribution of water input coefficients of various pollution loads in the study area is similar to the distribution of rainfall in typical years.The TP load is added into the USLE((Universal Soil Loss Equation))model,so the weight proportion of rainfall factors is larger and the sensitivity is higher.2)The largest contribution ratio of COD load is livestock and poultry production,followed by rural life.Due to the implementation of the regulations on pollution prevention and control of livestock and poultry production in 2014,the overall COD load shows a trend of increasing first and then decreasing.3)The sensitivity of living sources in the study area in 2015 and 2018 increased significantly compared with that in 2008 and 2012,and the sensitivity levels of Huoqiu,Shouxian and Huoshan counties increased significantly.4)Rural life and livestock breeding were the main factors affecting the output of ammonia nitrogen fitting load,and the areas with developed livestock and poultry breeding industry contributed significantly to the increase of ammonia nitrogen fitting load output.5)The contribution rate of rural life to TP load is the largest.With the approach of time domain,the contribution rate of land use to TP load simulation increases gradually.According to the actual area and pollution load of each district and county in the study basin,the contribution of pollution load intensity of each pollution source is calculated,and the following conclusions are drawn:1)The sensitivity of living source load basically increases with the approach of time domain.By 2018,the sensitivity of living source load in Huoqiu and Funan county has been degraded,but Fuyang city and Huainan city are still in the highly sensitive areas of living source.2)The sensitivity of livestock and poultry sources in the middle and northern regions of the study area increased from 2008 to 2018.By 2018,the load intensity of livestock and poultry sources in Guzhen,Huaiyuan and Wuhe counties was 0.869,0.383 and 0.377 kg/(hm2·y),respectively,which were highly sensitive areas.3)The sensitivity of living sources in the study area in 2015 and 2018 increased significantly compared with that in 2008 and 2012,and the sensitivity levels of Huoqiu,Shouxian and Huoshan counties increased significantly.4)From 2008 to 2015,the comprehensive source load of each district and county in the study area showed an upward trend,and about 70%of the changes of comprehensive source load were related to the changes of livestock and poultry breeding sources.From 2015 to 2018,the sensitivity of comprehensive source load in the study area was obviously enriched.In 2018,guzhen,Huaiyuan and Wuhe counties in the northeast of the study area with developed livestock and poultry breeding were all highly sensitive areas of comprehensive source,and their pollution load intensity was 0.895,0.404 and 0.407 kg/(hm2·y),respectively.Figure[68]Table[32]Reference[116]... |
