Integrating Seasonal Variation and Spatial Heterogeneity into Wind Erosion Driving Force Analysis in a Typical Steppe in China

Soil wind erosion (SWE) remains a significant challenge to improving ecological environmental quality and achieving sustainable socioeconomic development in drylands of northern China. An in-depth understanding of the spatio-temporal variations and underlying mechanisms of regional SWE is a prerequisite for the scientific prevention and mitigation of erosion-related hazards. However, in regions with high variability in intra-annual climate, quantitative studies on the spatial heterogeneity and intra-annual variability of drivers of SWE are scarce. This knowledge gap poses challenges for policymakers in developing effective landscape management strategies that are spatially and temporally specific. Here, the dynamics of SWE in the Xilingol typical steppe of China were simulated using the Revised Wind Erosion Equation (RWEQ) at seasonal and annual scales during 2000–2020. Stepwise regression and geographically weighted regression (GWR) were employed to examine the spatial heterogeneity in the relationships between SWE and environmental variables. The results revealed that RWEQ simulations were significantly correlated with the frequency of dust storm events at the seasonal scale (R 2 = 0.807, p < 0.01). SWE in spring accounted for approximately two-thirds of the annual total, indicating that spring was the critical period for SWE control. High SWE intensity was concentrated in sandy soil regions, with the Otindag Sandy Land and Gahai Elesu Sandy Land being identified as priority areas for desertification prevention and control. Over the study period, SWE exhibited an overall decreasing trend at both seasonal and annual scales, suggesting an enhancement in the ecosystem’s capacity for windbreak and sand stabilization. The stepwise regression results indicated that climatic factors generally had greater explanatory power than topographic and landscape pattern variables. Wind speed showed the strongest association with SWE across different time scales, whereas the relationships of normalized difference vegetation index (NDVI) and precipitation with SWE exhibited clear seasonal dependence. The GWR results further revealed pronounced spatial heterogeneity and seasonal variability in both the direction and magnitude of the associations between SWE and climatic and landscape pattern variables. These findings provide scientific support for identifying priority areas for desertification prevention and for developing spatio-temporally targeted landscape management strategies in dryland sandy regions.