Utilizing geodetectors to identify conditioning factors for gully erosion risk in the black soil region of northeast China

• DOI: 10.1016/j.iswcr.2024.07.004
• Publication Date: 2024-07-31T22:28:56Z
• Authors: Donghao Huang; Xinrui Zhao; Zhe Yin; Wei Qin
• Keywords: Gully erosion; Black soil region; Geodetectors; Gully erosion risk; 0401 agriculture, forestry, and fisheries; 04 agricultural and veterinary sciences; TA1-2040; Engineering (General). Civil engineering (General); Random forest

In the black soil region of Northeast China, issue gully erosion persists as a significant threat, resulting in extensive damage to farmland, severe degradation soil, and decreased productivity. It is therefore utmost importance accurately identify areas that are susceptible effectively prevent control its negative impact. This study tried utilize geographical detectors (geodetectors) means factors contribute distribution gullies assess risk (GER) five catchments within region, with ranging from approximately 80 km2 200 km2. By employing geodetectors method, fourteen geo-environmental were analyzed, including topographic attributes (such aspect, catchment area, convergence index, elevation, plan curvature, profile slope length, slope, stream power wetness index), channel network distance, vegetation index (NDVI EVI), well land use/land cover (LULC). The modeling GER was conducted using random forest algorithm (RFA). Out examined factors, only subset, comprising less than or equal 50%, demonstrated (p < 0.05) influence on spatial gullies. These selected sufficient assessing GER, LULC (mean q-value=0.270) elevation q-value=0.113) identified two most important factors. Furthermore, RFA exhibited satisfactory performance across all catchments, achieving AUC values 0.712 0.933 = 0.863) predicting GER. Overall, classified into high, moderate, low, very low-risk levels, representing 9.67% 15.95%, 19.28% 26.08%, 24.59% 30.55%, 30.54% 39.08% total respectively. Importantly, positive linear relationship (r2 0.722, p observed between proportion cropland area occurrence high-level Although primary levels categorized low high-risk exceeded existing coverage (0.34% 3.69%). findings highlight substantial potential for underscore necessity intensified efforts prevention China.