Influence Of Slope And Flow Direction Algorithm On LS Factor

Soil erosion is an environmental problem that needs to be solved urgently in China and the world today.China has a wide range of soil erosion,and soil erosion in the Loess Plateau is very serious.In this paper,the gully in the loess hilly region is taken as the research object.The Moore algorithm is used to calculate the LS factor of the study area,simulate the soil erosion condition,explore the influence of different slope and flow direction algorithm models on the LS factor,analyze the scale effect of the LS factor,and explore the multi-scale.Under the DEM condition,the contribution rate of the unit catchment area SCA and sinβ to the LS factor is further studied.The soil erosion status in the loess hilly region is further understood,and the status of soil erosion in the loess hilly region is known,in order to provide a theoretical basis for controlling soil erosion and controlling soil erosion.The main conclusions are as follows:1.Compared the slope algorithm models of three third-order differential series,it is found that the difference in slope maximum value is small.Regardless of the influence of the single-flow D8 algorithm and the multi-flow MFD algorithm,the spatial distribution difference of the LS factor is relatively small.The LS factor values are mainly distributed in the range of 0 to 25.Among them,the average error of the LS factor calculated by the three-order finite difference model algorithm is closest to 0,and the medium error is the smallest.Compared with the area ratio curve of the LS factor obtained by the three-order finite difference weighted by reciprocal of squared distance slope algorithm model,they are the closest;when the two slope algorithms are benchmarks,the relative difference coefficient values are the closest.Therefore,when calculating the LS factor in this region,the slope algorithm model is recommended.2.The effects of single-flow D8 algorithm and multi-flow MFD algorithm on LS factor are compared.The results show that the spatial distribution of LS factors obtained by D8 and MFD algorithm is significantly different.When the LS factor obtained by the MFD algorithm is used as the reference algorithm,the relative difference coefficient is higher than that when the LS factor obtained by the D8 algorithm is used as the reference algorithm.The LS factor value is in the range of less than 40,and the cumulative frequency curve is largely different.The LS factor is in the range of more than 40,and the cumulative frequency curve is approximately coincident.Compared with the LS factor value obtained by the D8 algorithm,the proportion of the MFD algorithm in the small value interval is relatively small.3.The original DEM data of 6m resolution is resampled into DEM data of 10 m,15m,20 m and 25 m resolution for analyzing the scale effect of the LS factor.The results show that the slope value decreases with the decrease of grid resolution,and the SCA value shows a general fluctuation.The maximum value of the LS factor tends to decrease.When the accuracy of the grid is reduced to a certain extent,the maximum value of the LS factor will rise.With the decrease of grid accuracy,the LS factor value is mainly distributed in the range of 0~25,and the area percentage decreases gradually in the range of 0~5 and >30;in the range of 10~25,the area ratio gradually increases;In the range of 25~30,the proportion of the area showed a trend of increasing first and then decreasing.4.The multivariate regression analysis was used to analyze the effects of SCA and sinβ on LS factors under multi-scale conditions.The results show that when the resolution is reduced from 6m to 25 m,the correlation coefficient values generally increase,and there are slight fluctuations in the local.From the regression coefficient values,the SCA regression coefficient is larger than the sinβ regression coefficient,indicating that the effect of SCA on LS factor is greater than that of sinβ on LS factor.5.The geographically weighted regression model was used to compare the contribution rate of SCA and sinβ to LS factor under multi-scale DEM conditions.The results show that the values of SCA and sinβ regression coefficients were positive,and the spatial distribution of regression coefficient values varied greatly,indicating that there was a positive correlation between the independent variables SCA and sinβ and the dependent variable LS factor.As the accuracy of DEM decreases,the regression coefficient values of SCA and LS factors gradually decrease.The value of the sinβ regression coefficient and its range of variation are smaller than the SCA regression coefficient values.As the resolution of the grid decreases,the sinβ regression coefficient may increase,which may be due to the combined effects of DEM accuracy and topographical factors.