Ephemeral gully erosion (EGE) is a major driver of cropland soil degradation, exerting substantial impacts on soil properties and crop growth. These changes, in turn, profoundly alter soil erosion resistance (SER). Although SER is critical for soil conservation, its spatial patterns and driving mechanisms in long, gently sloping (LGS) cropland under the influence of EGE remain insufficiently understood. The aims of this study were to quantify the spatial variation in SER and to identify the influencing factors on a representative LGS cropland in the Mollisol region of Northeast China. Based on in-situ field sampling and controlled flume tests, the spatial heterogeneity of SER under EGE conditions was quantified. The dominant influencing factors and their interactive mechanisms were identified, and a stepwise regression model was developed to estimate rill erodibility (Kr). Results showed that SER varied significantly across slope positions but did not differ markedly between gully positions. Kr linearly increased with decreasing gully position, while critical shear stress (tic) first decreased and then increased, with variation ranges of-2.36 % to 50.00 %. Both Kr and tic showed quadratic relationships with slope position, with Kr peaking at the middle slope and tic at the upper slope. Dominant factors affecting Kr included clay content, sand content, soil cohesion (Coh), mean weight diameter (MWD), root mass density (RMD), and straw mass density (SMD), which collectively explained 79 % of the spatial variability. Notably, SMD had a significant regulatory effect on Coh, RMD, and MWD, and indirectly reduced Kr via these pathways (standardized path coefficient =-0.263). The developed Kr estimation model (Kr = 0.212MWD-2.405RMD-0.502) exhibited good predictive performance (R2 = 0.860; NSE = 0.863) but requires further validation under field conditions. The findings provide important theoretical support for site-specific erosion control strategies and contribute to the improvement of process-based soil erosion models at the hillslope scale in Mollisol regions of Northeast China.

The surface/subsurface "dual" structure in karst area complicates the regional soil erosion and sediment transport process. Knowledge of mechanism and processes of sediment transport in small watersheds may guide water-sediment management strategies. This study investigates runoff and sediment processes in three karst watersheds in southern China with different development degrees: Mahuangtian (MHT), Yangjichong (YJC), and Niulang (NL). The analysis is based on long-term, high-frequency hydrological data from 2014 to 2021. We qualitatively classified karst development based on rock exposure rate, vegetation coverage, and soil thickness, and applied hysteresis analysis to establish the relationship between sediment transport and karst development. The results indicated that YJC yielded the highest mean runoff depth of 36.6 mm, while MHT and NL were lower, at 12.5 and 22.4 mm, respectively. Correspondingly, the mean sediment modulus were 136, 30.5, and 119.6 t center dot km(-2)center dot a(-1)for MHT, YJC, and NL, respectively. In combination with the types of hysteresis, as well as the distribution of karst in the watershed and vegetation characteristics, we infer that the sediment in the MHT mainly originates from cultivated land, the riverbed near the downstream outlet (proximal source) and the subterranean karst conduit. Sediment in the YJC watershed originates mainly from upstream corn fields (distal source). Sediment in the NL watershed originates mainly from slopes on both side of the riverbed (proximal source). This study reveals the process mechanism of water and sediment transport in small watersheds with different karst development degrees, which can provide important support for formulating water and sediment management plans in areas with different karst development degrees.

Rocky desertification has formed a unique vegetation-soil-rock structure (VSRS) on the underlying surface of karst area, resulting in complex surface processes. Despite their significance, the response of runoff generation and erosion processes at binary interfaces to VSRS is still poorly understood. In this study, artificial simulated rainfall experiments were carried out on karst micro-plots with binary structure. Driven by rainfall intensity (R) and VSRS, the runoff and sediment output processes and its hydrodynamic mechanisms at the binary interface were studied. The results showed that the surface runoff rate (SRr) decreased with the increase of moss cover (MC) or rock exposure rate (RE), and increased with the increase of soil thickness (ST), with an amplitude more than or equal to 4.98%. With the increase of R, the surface runoff yield (SR) and runoff leakage (RL) showed an increasing trend, but the SR increased even more. This indicated that SR is more susceptible to R. Surface sediment yield (SS) showed complex changes with increasing test gradients, while sediment leakage (SL) showed a single increase (r >= 0.46, P < 0.05). Runoff and sediment leakage occurred at different VSRS, and the initial rate was greater than 0.09 Lmin(-1) or 0.02 gmin(-1). This suggested that soil and water leakage in karst areas may be widespread. Proportion of SR and SS increased with increasing R, and became the main loss pathway (>54.77 %) during heavy rainfall (120 mm/h). Prediction performance of hydrodynamic parameters for SRr was 12.25 % higher than that for surface sediment rate (SSr), and V was the best factor for SRr (R-2 > 0.84, P < 0.01). This study further enriches the theory of hydrological processes in karst rocky desertification areas, and provides important support for decision-making on soil and water conservation and ecological restoration.

Soil erosion is significantly altering the global carbon cycle, leading to considerable soil organic carbon (SOC) loss, which in turn affects carbon redistribution. However, the assessment of SOC loss in complex environments, such as karst regions, and its causal relationships with environmental factors remain underexplored. To fill it, this study proposes a novel approach for evaluating SOC loss in karst regions. This approach integrates rocky desertification factors to refine the Revised Universal Soil Loss Equation (RUSLE) for analyzing dynamic changes in soil erosion and SOC loss in karst regions. Additionally, causal inference theory is applied for the first time to construct a causal network for SOC loss, with the Tree-Augmented Naive Bayes method used for risk assessment. The findings indicate that the average SOC loss due to soil erosion over the study period was 2.16 x 1010 g C, with SOC loss closely mirroring the spatial patterns of soil erosion. The causal network revealed that rainfall exerted a significant direct causal effect on SOC loss, surpassing vegetation as the dominant factor, with its influence intensifying over time. Exposed bedrock, acting as a moderating factor, added complexity to the SOC loss process in karst regions. High SOC loss areas are mainly found in regions with abundant rainfall, dense vegetation, moderate slopes, high topographic wetness index, and high clay content. These results enhance our understanding of environmental interactions and carbon cycle responses in karst regions, providing a robust scientific basis for regional carbon sink strategies and effective management practices.

Intense land degradation had created a special vegetation-soil-rock complex structure (VCS) on karst slopes, which altered regional soil and water processes. In this study, we investigated the combined effects of heterogeneous VCS on soil erosion/leakage, rainwater transformation and hydrodynamic characteristics at the microplot scale by simulating the karst dichotomous structure slopes with steel tanks and indoor artificial rainfall. The analysis showed that the surface runoff rate decreased with the increase of VCS and the subsurface runoff rate decreased with the increase of VCS. When the rainfall intensity increased to 60-120 mm/h, there was obvious surface runoff yield on the VCS slope. When the rainfall intensity exceeded 60 mm/h, the VCS showed obvious surface sediment yield with an initial rate ranging from 0 to 4.03 gmin(-1). VCS showed obvious underground runoff and sediment yield under different rainfall intensities, and the initial rate was greater than 0.45 Lmin(-1) or 0.13 gmin(-1). This suggests that soil and water leakage from the karst rocky desertification slopes may be generalized. All the erosion flow regimes of VCS slopes were rapid laminar flow or slow laminar flow. The drag coefficient and flow shear increased with the increase of VCS, and the flow power showed a trend of increasing, then decreasing and then increasing. The water flow shear and water flow power showed a power function relationship with the sediment yield rate (R-2 >= 0.2293, P < 0.05). In terms of direct effects, hydrodynamic characteristics had the strongest influence on surface sediment yield (beta = 0.68, P < 0.05), and rock exposure rate had the strongest influence on subsurface sediment yield (beta = 0.56, P < 0.05). In terms of total effect, rainfall intensity was the dominant driver of surface/subsurface sediment yield (beta = 0.75/0.72, P < 0.05). This study provides insights into understanding the mechanism of hydraulic erosion on rocky decertified slopes and provides a theoretical basis for decision-making on soil erosion management in karst areas.

Extreme precipitation is a crucial trigger for soil erosion events in karst regions. However, the existence of a scale effect in suspended sediment characteristics of karst basins and which extreme precipitation variables control this effect remain unclear. To investigate this, we analyzed the scale effect on suspended sediment characteristics using monthly hydrological data from five karst basins of varying scales, consistently monitored from 2012 to 2019. We aimed to clarify the contribution of extreme precipitation to suspended sediment and identify the dominant influencing factors. The results showed a significant negative exponential correlation between basin scale and sediment transport rate (STR). The phenomenon occurs due to two main reasons, first, the increased sediment transit distances with growing basin dimensions, leading to higher sedimentation; second, the unique dual hydrological structure of karst landscapes, which enhances subterranean leakage and further sedimentation. Furthermore, extreme precipitation was found played a crucial role in explaining STR variability, explaining approximately 56 % to 78 % of the observed variance in the five karst basins. Simultaneously, the dominant extreme precipitation factors affecting suspended sediment variability varied with basin scale. Specifically, heavy precipitation days, consecutive wet days, the rainstorm amounts and rainstorm days were identified as the main determinants. These variations are mainly attributed to the watershed's sensitivity to extreme precipitation events and its intrinsic attributes, including dual pathways, land use patterns, and geomorphological types, etc. This study provides theoretical insights into the increasing soil erosion caused by extreme precipitation in karst basins with different scales.

The Revised Universal Soil Loss Equation (RUSLE) is the most widely used soil erosion modeling method worldwide. The karst regions, influenced by geological conditions and human activities, feature extensive exposure of carbonate rocks on the surface, which presents challenges for the application of the RUSLE model in these areas. This study introduces the rocky desertification factor (D) to characterize the influence of exposed surface rock on soil loss. The relationship between rock exposure rate and soil erosion was incorporated into the RUSLE model to develop a RUSLE-D model. We compared the performance of the RUSLE and RUSLE-D models using long-term high-frequency hydrological signals from two typical karst catchments to validate the applicability of the RUSLE-D model in karst areas. The results indicated that under natural rainfall conditions, soil erosion decreased as the rock exposure rate increased, showing a negative exponential relationship. The RUSLE-D model estimated the multi-year average soil erosion rates for the SBT and GC catchments to be 8.99 and 14.63 t ha−2·yr−1, respectively. The R2 values for the RUSLE and RUSLE-D models in the SBT catchment were 0.34 and 0.78, respectively, with NSE values of −0.03 and 0.55, and PBIAS values of −81.39 % and 13.87 %; for the GC catchment, the R2 values were 0.14 and 0.68, with NSE values of −13.82 and 0.43, and PBIAS values of −182.85 % and −24.27 %. The MCI indices for the SBT and GC catchments were 0.56 and 0.96, respectively. The RUSLE-D model significantly improved the accuracy of soil erosion simulation in typical karst watersheds. This study underscores the importance of incorporating the rocky desertification factor in soil erosion assessments within karst areas. The newly developed RUSLE-D model contributes to further developing the USLE/RUSLE series of models, enhancing their applicability in karst areas.

The source and transport processes of suspended sediment are issues that require further study. The complex relationship between the integrated surface characteristics and the mechanisms of sediment sources and transport in heterogeneous karst watersheds remains unclear. This study proposes a comprehensive assessment framework for sediment sources that combines the surface characteristics of karst watersheds, such as exposed rock and shallow topsoil in karst areas. Additionally, the Karst Surface Characteristic Index (KSCI) is developed to evaluate the potential supply capacity of sediment in karst watersheds. The hysteresis index (HImid) is calculated using a long series of hydrological data to study runoff and sediment relationships in eight watersheds in the karst region of southern China. Results indicated that plan curvature, terrain relief, NDVI, and rock exposure contributed most to KSCI, with these four factors contributing 65%. The Q -SSC hysteresis analysis results based on 445 hydrological events showed that a clockwise hysteresis pattern was the most frequent and efficient pattern of sediment transport in the studied watershed (clockwise hysteresis pattern accounted for 72.3% of 445 hydrological events) owing to the limited sediment supply in the karst watersheds. The quantitative analysis results of HImid and KSCI showed that HImid was closely related to KSCI. With an increase in KSCI, a clockwise hysteresis pattern was more likely to appear, indicating that watershed characteristics exerted a significant influence on sediment transport velocity and efficiency. In terms of spatial distribution, near the outlet of the watershed or the river channel, high KSCI values were distributed, and clockwise hysteresis was more likely to occur. In contrast, at the far end of the watershed outlet or the river channel, high values of KSCI were distributed, and anticlockwise hysteresis was more likely to occur. Our findings further enrich the theory of soil erosion and the runoff -sediment relationship at the watershed scale in karst areas, and serve as valuable decisionmaking references for integrated water and soil sources management.

Increasingly frequent extreme rainfall as a result of climate change is strongly damaging the global soil and water environment. However, few studies have focused on daily extreme sediment events (DESE) in heterogeneous karst watersheds based on long-term in -situ observations. This study quantitatively assessed the time effect of DESE on rainfall response, decoupled the impact of environmental factors on DESE by using structural equation modelling, and finally explored the modelling scheme of DESE based on the hybrid model. The results showed that DESE had the highest frequency of occurrence in May -July, with dispersed distribution in the value domain. Rainfall with a time lag of 1 day and a time accumulation of 2 or 3 days was an important contribution to DESE ( P < 0.01, R = 0.47 -0.68). Combined effects of environmental factors explained 53.6 % -64.1 % of the variation in DESE. Runoff and vegetation exerted the strongest direct and indirect effects on DESE, respectively (8 = 0.66/ -0.727). Vegetation was the dominant driver of DESE in Dabanghe and Yejihe (8 = -0.725/-0.758), while the dominant driver in Tongzhihe was climate (8 = 0.743). In the future, the risk of extreme sediments should be prevented and resolved through the comprehensive regulation of multiple paths, such as runoff and vegetation. Hybrid models significantly improved the modelling performance of machine learning models. Generalized additive model -Extreme gradient boost had the best performance, while Partial least squares regression -Extreme gradient boost was the most valuable when considering performance and input data cost. Two methods can be used as recommended solutions for DESE modelling. This study provides new and in-depth insights into DESE in

[Objective] The soil erosion issue is very prominent in the karst region of southwest China and significantly affects the sustainable development of the regional economy. However, it is currently not very clear how the characteristic environmental factors of karst basins affect the direct driving force of soil erosion and runoff erosivity. [Methods] Taking the Yeji River Basin in the karst region of southwest China as an example and basing on the long-term and high-frequency field measurement data at the outlet hydrological station of the basin, the spatiotemporal change characteristics of runoff erosivity from 2005 to 2020 were evaluated and the influence of main karst environmental factors on runoff erosivity was analyzed by the Soil and Water Assessment Tool (SWAT) and the Partial Least Squares Model (PLSR). [Results] The average annual runoff erosivity was 2 326.47 m4/(km2·s), with strong spatial heterogeneity. In 2005, the high runoff erosivity was mainly distributed in the northern and central parts of the basin. During 2010—2020, the high runoff erosivity area gradually shifted to the central and southern parts of the basin. In terms of time, benefiting from the implementation of ecological restoration projects such as comprehensive control of soil and water loss in sloping farmland and comprehensive control of rocky desertification since the 11th Five-Year Plan, the runoff erosivity of the Yeji River Basin had been generally decreasing year by year. Lithology, exposed bedrock, and slope were the dominant control factors, which jointly explained 57.7% of the spatiotemporal variation of runoff erosivity in the Yeji River Basin. [Conclusion] These results can provide a theoretical reference for the comprehensive control of soil erosion in karst basins and help to formulate more precise and effective policies and measures to improve the ecological environment of the karst area and promote the sustainable development of the region.

Soil erosion is a prominent environmental problem in karst regions.Exploring the spatiotemporal variability of soil erosion and the factors that influence soil erosion is of great significance for regional soil erosion prevention and control.However,the mechanisms influencing the characteristic features of the karst basins,such as bedrock exposure and lithology,still need to be further explored.This study used GIS technology,the Revised Universal Soil Loss Equation model,Getis-Ord Gi*,and partial least squares regression(PLSR)to identify the dominant factors influencing soil erosion and the spatiotemporal variability of soil erosion in 31 sub-basins of the Dabang River Basin(DRB),a typical karst area of Southwest China,from 2010 to 2020.The results indicated that soil erosion in the DRB from 2010 to 2020 was generally decreasing,the mean soil erosion in the DRB in 2010,2015 and 2020 was 18.46,16.51 and 15.29 t ha-1 a1,respectively.During the study period,the area of slight erosion increased by 26.39％(706.54 km2),while severe erosion enlarged by 26.36 km2.Spatially,the DRB was primarily affected by medium and slight soil erosion.The hot spot areas of soil erosion(key control areas)were mainly concentrated in the central and southern parts of the basin,decreasing each year,and the area of soil erosion hot pots has decreased from 43.22％to 20.60％.PLSR decoupling results show that elevation,bedrock exposure,land use type,interbedded limestone and clastic rock,and vegetation coverage were identified as the key variables affecting soil erosion,explaining 52.8％of soil erosion variability,with a high value of the Variable Importance on Projection(VIP)more than 1.These results can be used as a reference for comprehensive control of soil erosion and water loss in the basin.

Soil erosion is one of the most serious ecological threats in karst areas of Southwest China. The identification of priority areas for remediation and its driving factors is essential to improving the efficiency of prevention and control. The present study systematically considered natural and socio-economic factors not involved in the revised universal soil loss equation (RUSLE) model, and determined priority areas for soil erosion management based on minimum administrative units and karst landforms. Then, the driving factors were identified by using geographic detector. The results showed that the priority areas were mainly concentrated in the southwest, southeast and northeast, overlapping with the severely eroded areas (Erosion rate=45.79 t·ha −1 ·a −1 ). Gradient risk zones had geomorphological differences, but the most eroded zones were all controlled by bedrock exposure rates, elevation, or slope position. The spatial correlation and high erosion rate of priority areas provided opportunities to optimize the efficiency and cost of control. Driving factors were affected by karst landforms. The explanation power of slope position on soil erosion was higher in the peak cluster depressions and karst basins with small undulations ([Formula: see text]), while the karst gorges, trough valleys and plateaus with large undulations gradually decreased ([Formula: see text]). The interaction of driving factors will enhance the explanatory power for soil erosion. Among them, the repetition rate of elevation was 60%, and the repetition rate of lithology and development index was 40%. This study provides useful information for identifying and managing priority areas for soil erosion control, and enriches the theory of soil and water conservation in karst areas.

Study region: Guizhou Province, Southwest China Study focus: Karst slopes are highly heterogeneous, with a large number of rocks exposed on the surface. Empirical soil erosion models (such as the revised universal soil loss equation (RUSLE)) based on homogeneous soil areas have encountered significant challenges in their application to karst areas. Currently, a quantitative relationship between the exposed rock and soil erosion has not been systematically established, although the establishment of this relationship could provide opportunities for empirical soil erosion model optimisation in karst areas. In view of this, we combined a field natural rainfall test with an indoor simulated rainfall test and artificially simulated exposed rock to investigate rock exposure rate (RER) effects on soil loss and runoff processes on karst slopes and attempted to build a coupled RER and soil loss model. New hydrological insights for the region: As the RER increased, the runoff and sediment content decreased under both simulated and natural rainfall conditions. Soil loss decreased exponentially with an increase in RER under natural rainfall tests in the field. There was a relationship between the soil loss and the RER quadratic function under simulated rainfall tests indoors. The exponential relationship based on natural rainfall more accurately reflected karst slope soil erosion conditions. These results contribute to a deeper understanding of soil erosion mechanisms on slopes in karst areas and provide confidence in our subsequent introduction of soil erosion sensitivity-exposed rock relationships to optimise soil erosion models such as RUSLE.

Frequent rainstorms caused by climate change are causing significant stresses and impacts on karst zones and even global hydrological systems. However, few reports have focused on rainstorm sediment events (RSE) based on long series, high-frequency signals in karst small watersheds. Present study assessed the process characteristics of RSE and analyzed the response of specific sediment yield (SSY) to environmental variables using random forest and correlation coefficients. Management strategies are then provided based on revised index of sediment connectivity (RIC) visualizations, sediment dynamics and landscape patterns, and modeling solutions for SSY are explored through the innovative use of multiple models. The results showed that the sediment process showed high variability (CV > 0.36), and the same index had obvious watershed differences. Landscape pattern and RIC show highly significant correlation with mean or maximum suspended sediment concentration (p<0.01, |r|>0.235). Early rainfall depth was the dominant factor affecting SSY (Contribution = 48.15 %). The hysteresis loop and RIC infer that the sediment of Mahuangtian and Maolike mainly comes from downstream farmland and riverbeds, while Yangjichong comes from remote hillsides. The watershed landscape is centralized and simplified. In the future, patches of shrubs or herbaceous plants should be added around the cultivated land and at the bottom of the sparse forest to increase the sediment collection capacity. The backpropagation neural network (BPNN) is optimal for modeling SSY, particularly for running the variables preferred by the generalized additive model (GAM). This study provides insight into understanding RSE in karst small watersheds. It will help the region to cope with future extreme climate change and develop sediment management models that are consistent with regional realities.

Monitoring and evaluating the evolution of rocky desertification timely and studying the characteristics of soil erosion under different rainfall patterns are of great scientific significance for regional soil and water conservation, rocky desertification control and ecological environment construction. Four periods of remote sensing image data from 2005 to 2020 were selected to study the evolution characteristics of rocky desertification and its impact on soil erosion in the controlled boundary area of Shibantang hydrological station of Yeji River Watershed, Guizhou Province, China. According to the 408 erosive rainfall events, the soil erosion under different rainfall patterns in the watershed was analyzed. The results showed that: erosive rainfall events in the study area were mainly pattern A, accounting for 57.4% of the total rainfall events; the second was pattern B, accounting for 28.9% of the total rainfall events; the rainfall pattern of C occurred occasionally. Among them, pattern A was the main rainfall pattern leading to soil and water loss and had the largest contribution rate to soil erosion in the watershed. From 2005 to 2020, the area of rocky desertification showed a decreasing trend, accounting for 72.2% from 87.9%. Spatially, rocky desertification has mainly concentrated in the middle south of the watershed since 2010, while the rocky desertification mainly concentrated in the middle and north before 2010. The effects of different grades of rocky desertification on soil erosion were different, and the soil erosion modulus in areas with the medium, severe and extremely severe rocky desertification was generally small. The soil erosion modulus estimated by the RUSLE (Revised Universal Soil Loss Equation) model was still much higher than that calculated by the data measured by the hydrological monitoring station. Therefore, the application of the RUSLE model in karst area needs to be further modified. These results can provide reference for rocky desertification control, soil erosion control and fragile ecosystem restoration in karst area.

Soil erosion is the prominent ecological and environmental problem in karst area of southwest China, which seriously restricts the sustainable development of the region. Determining the priority areas of soil erosion governance and its driving factors can significantly improve the efficiency of prevention and control. However, at present, there are few researches on the priority areas of governance that comprehensively consider karst landform types and management requirements. Based on the minimum administrative unit and karst landform types, this study identified the priority areas of soil erosion control by comprehensive use of spatial clustering method and geographic detectors, and quantified the driving factors and their interactions. The results indicated that: (i) Priority areas within the smallest administrative unit are clustered in the southwest, southeast and northeast of the study area, overlapping with areas of intense erosion; Geomorphological differentiation of multi-factor gradient risk zones is obvious, but the areas with strong erosion are all controlled by bedrock exposure rate, altitude and slope aspect. (ii) The soil erosion in the priority area is concentrated and intense. By treating the priority area, which accounts for 12.77% of the total area, soil erosion can be reduced by 27.66%. (iii) The driving factors have a strong dependence on karst landforms, showing obvious differences in different karst landform areas; Interaction of factors, especially the interaction between human disturbance factors and natural influence factors, can significantly enhance the explanatory power of soil erosion. The research results have important theoretical significance for the planning and control of soil and water loss in karst areas.

The heterogeneity of the geographical environment determines the complexity of soil erosion patterns. Determining the effects of changing gradients and combinations of environmental factors on soil erosion is the key to combating soil erosion. In this study, based on the division of the karst development degree, the revised universal soil loss equation model was optimised based on the rocky desertification factor to determine the effect of the different gradients of environmental factors on soil erosion. The geographical detector method was used to identify the dominant and interactive factors of soil erosion under different karst development degrees. The gradient attribution shows that: soil erosion is influenced by the type of bedrock assemblage and the form of the assemblage. Soil erosion is strongest at 1400–1800 m above altitude. Grassland in karst areas is prone to soil erosion and the effect of vegetation coverage on soil erosion has a critical value of 0.6–0.7. Interaction analysis shows that the interaction between land use and slope plays a dominant role in the spatial differentiation of soil erosion, but has significantly higher explanatory power (q)in sub-karst and non-karst areas than in pure-karst areas. Combined with slope and rainfall, soil erosion is obviously intensified, and the pure-karst and sub-karst areas are higher than the non-karst areas. The significant interaction between lithology and slope is a typical feature of karst areas. The findings of this study have crucial theoretical significance for the site-specific control of soil erosion in areas with different karst development degrees.

Soil erosion is a process of migration and redistribution of soil substances in the landscape, which is regulated by topography, vegetation, human activities and their spatial pattern. At the watershed scale, changes in landscape pattern are important factors in determining the degree of soil erosion. Taking Dabang River Basin as the study area, based on the three phases of land use data, remote sensing image data and daily rainfall data from eight stations in the basin in 2010, 2015 and 2020, the rocky desertification factor ( D ) was introduced into the general soil loss equation RUSLE to calculate the soil erosion in Karst and non-Karst Areas in 2010, 2015 and 2020 respectively. The relationship between landscape pattern and soil erosion was analyzed from two aspects: type level index and landscape level index. The results showed that: 1) From 2010 to 2020, the average soil erosion modulus in The Dabang River Basin decreased first and then increased. The average soil erosion modulus in the non-karst region was about twice that in the karst region, and the average soil erosion modulus in the karst region decreased first and then increased. The mean soil erosion modulus in the non-karst area showed an increasing trend; 2) Under different slope grades, the erosion was mainly slight and mild, and the area of slight erosion was the largest, and the area of very strong and severe erosion increased as the slope increased. the area of strong, very strong and severe erosion increased in the slope zone below 15°, the area of light and moderate erosion decreased, and the area of slight, strong and very strong erosion increased in the slope zone from 15 to 25°, and the area of slight erosion increased in the slope zone above 25° area increased and light, moderate and strong erosion area decreased in the slope zone above 25°; 3) The landscape pattern of the Dadang River Basin changed significantly from 2010 to 2020. At the landscape level, the number of patches increased and the average patch area decreased. At the type level, the area of paddy field, woodland and shrubland decreases and the area of dry land, grassland, construction land and water body increased, and the dominant land type in the watershed changed from woodland to grassland; 4) The amount of soil erosion was positively correlated with patch type area, landscape percentage, maximum patch index and aggregation index, and positively correlated with edge density; 5) There was a linear relationship between soil erosion and Shannon diversity index (SHDI) and Shannon mean index (SHEI) at landscape level. The results can provide reference for land use planning and soil and water conservation measures.