A study on influence factors of soil erosion in Guizhou based on optimal parameter geodetector model

As soil erosion accounts for a key culprit for land degradation across the globe, quantitatively identifying the causes of soil erosion is a crucial step in our effort to conserve soil and water and restore regional ecosystems. In the present work, the revised universal soil loss equation (RUSLE) model was employed to calculate the soil erosion modulus of Guizhou Province in southwestern China for the years 2000, 2010, and 2020. The Markov transition matrix was integrated to analyze spatiotemporal characteristics and the optimal parameter geographic detector (OPGD) model was used to assess the impact and spatial variability of a range of factors on the distribution pattern of soil erosion in Guizhou. The main findings are as follows: First, the erosion in Guizhou was primarily of mild and slight intensity, exhibiting a pattern of mixed point-line distribution, and the area without soil and water loss has increased by 4971.96km² over the years of study; Second, the main drivers for the variations in the soil erosion levels in Guizhou Province were the increasing C factor and the declining P factor; Third, the optimal parameter combinations for different variables varied, showing distinct responses across spatial scales; Fourth, the factor “slope” was found to have the most significant impact on the soil erosion distribution in the province, and considerable spatial variation was observed between the variables and regions. The OPGD model has demonstrated enhanced performance in terms of the spatial data discretization and scale effects, reaching higher accuracy in identifying the influencing factors for soil erosion, improving understanding of the distribution of soil erosion and its patterns of changes in Guizhou, and facilitating improved management of soil and water loss in this area.