Simulation Of Non-point Source Nitrogen Pollution In The Upper Reaches Of Chishui River Basin Based On SWAT Model And Its Influencing Factor

The upper reaches of the Chishui River basin is a typical karst mountain basin.Due to the widely distributed agricultural land,the steep terrain and the abundant precipitation conditions in the southwest of China,this area has long faced severe non-point source nitrogen pollution problems.Exploring the influencing factors of non-point source nitrogen load in the upper reaches of the Chishui River Basin,and further clarifying the impact mechanism of each influencing factor on the non-point source nitrogen load in the basin,can provide effective decision basis for protection of water environment and rational development and utilization of water resources in the basin.Therefore,this paper uses TN load(the total amount of organic nitrogen and inorganic nitrogen in all forms)as the measurement index of non-point source nitrogen load in the study area.Based on the obtained spatial data of the study area(spatial distribution data of land use and soil types,DEM data,etc.)and attribute data(meteorological data,soil attribute data,hydrology and water quality data,etc.),a SWAT model was constructed which is suitable for the upper reaches of the Chishui River Basin.And the runoff and TN simulation accuracy of the model was calibrated and verified.Based on the simulation results of the SWAT model,the temporal and spatial distribution of precipitation and the influence of land use type,soil type and slope grade on the TN load in the watershed were systematically analyzed from the sub-basin scale.there are significant differences in the TN load intensity even for the same land use type.Therefore,based on the TN load intensity of each HRU(hydrological response unit),this paper analyzes the reasons for this phenomenon from three aspects: precipitation,slope grade and soil type.Finally,based on the research results,this study proposes a division plan for the key areas of returning farmland to forests considering slope and soil type.and evaluates the plan from two aspects: the spatial distribution characteristics of the key areas for returning farmland to forests and the reduction rate of TN load.The main results of this study are as follows:(1)The runoff and TN simulation results of the model were calibrated and verified by SWAT-CUP software.The result show that the R2 and Ens are 0.82 and 0.79 during the calibration period in the runoff simulation,respectively.And the R2 and Ens are 0.89 and 0.84 during the validation period,respectively.In the TN simulation results,the R2 and Ens of Qingshuipu Station reached 0.66 and 0.62 during the calibration period,respectively.And the R2 and Ens of the verification period reached 0.66 and 0.59,respectively.The R2 and Ens of Qingchi Station reached 0.72 and 0.53 during the calibration period,respectively.And the R2 and Ens of the verification period were 0.79 and 0.70,respectively.Because of all of the results are meet the requirements of R2>0.6 and Ens>0.5,the model accuracy meets the requirements of subsequent experiments.(2)In terms of time distribution,the temporal distribution characteristics of TN load in the study area were significantly affected by precipitation,mainly distributed in the summer flood season(May to September),and the intra-and inter-annual variability of TN loads is usually greater than that of precipitation.In terms of spatial distribution,the spatial heterogeneity of precipitation,total TN load and TN load intensity in the study area is significant.Among them,the total amount of TN load in a sub-basin is mainly affected by the area of the sub-basin,while the intensity of TN load is affected by the combined effects of precipitation and underlying surface conditions.The spatial distribution characteristics of TN load intensity at the sub-basin scale are relatively consistent with the spatial distribution characteristics of precipitation,and only differ in the south of the middle part of the upper reaches of the Chishui River Basin.(3)From the sub-basin scale,the TN load intensity differs significantly among different land use types.The TN load intensity of paddy field and dry land is the largest,which are31.93kg/ha and 34.65kg/ha respectively.And the TN load intensity of forest land is the smallest,which is 3.5kg/ha.The TN load intensity of paddy field and dry land is about 7.5times that of forest land.However,the changes of soil type and slope grade had relatively little effect on the TN load intensity.The soil type with the highest TN loading intensity has30% greater than the soil type with the lowest TN loading intensity;The slope grade with the highest TN loading intensity has 31% greater than the slope grade with the lowest TN loading intensity.(4)Even with the same land use type,there are significant differences in TN load intensity in the study area.From the spatial distribution of the annual average TN load intensity of the five main land types(paddy field,dry land,forest land,shrub forest and grassland),the spatial distribution of the annual average TN load intensity of paddy field,dry land,forest land and shrub forest is significantly different,and the distribution characteristics are basically consistent with the spatial distribution characteristics of precipitation in the study area.The spatial distribution difference of the annual average TN load intensity of grassland is relatively small,and there is no obvious consistency with the spatial distribution characteristics of precipitation.The correlation analysis showed that the TN load intensity of paddy field,dry land,forest land,shrub forest and construction land was significantly positively correlated with precipitation,while the TN load intensity of grassland was not correlated with precipitation.For the TN load intensity of each land use type under different slope levels,the change of precipitation does not affect the trend of TN load intensity.The TN load intensity of paddy field,dry land,woodland,shrub forest and construction land increases with the increase of slope grade,and the trend of increasing is more significant only when the slope grade transitions from grade 1 to grade 2.However,the TN load intensity of grassland and bare rock is basically not affected by the change of slope.For the TN load intensity of each species under different soil types,the change of precipitation does not affect the trend of TN load intensity.Except for forest land,the TN load intensity of other land use types is less affected by soil type.Compared with the forest land under other soil types,the TN load intensity of forest land under highly active strong acid soil,loose lithologic soil and embryonic soil is reduced by about 10 times..(5)Based on the above research results,four scenarios of returning farmland to forest were set up to explore the difference between the plan that considers both slope and soil type and the plan that only considers the slope.The experimental results show that the area of the key areas delineated by the plan that considers both slope and soil type is greatly reduced,and the distribution is significantly more concentrated,which effectively improves the actual implementation efficiency of the measures of returning farmland to forest.For the TN load reduction rate,the reduction rate of TN load of the scheme considering both slope and soil type is about 2-3 kg/ha higher than that of the scheme considering only the slope.That is,the same area of returning farmland to forest will be able to more effectively reduce the TN load in the watershed.