Abstract: Soil erosion reduces soil fertility and crop yields, posing significant threats to agricultural productivity and ecological sustainability. Soil erosion in Changchun exhibits pronounced spatial heterogeneity, due to the combined effects of long-term human activities, climate change, and complex topography. The spatiotemporal variations and underlying mechanisms remain insufficiently understood. We quantitatively assessed soil erosion in Changchun from 1990 to 2020 using the Revised Universal Soil Loss Equation (RUSLE), analyzed the spatiotemporal driving forces of soil erosion with the help of the Geodetector model, and projected soil erosion intensity in 2030 and 2040 under multiple future climate scenarios. The results showed that: (1) Soil erosion in Changchun was predominantly mild and slight during 1990 to 2020, following a spatial pattern characterized by higher erosion in the southeast and lower in the northwest. Mild and slight erosion consistently dominated, accounting for 97.6% and 2.2% of the total area in 1990, and changing to 98.1% and 1.8% by 2020, respectively. (2) Geodetector results identified slope as the dominant factor influencing soil erosion, with a q value of 31.6%, underscoring its significant role in spatial differentiation. The interaction between slope and elevation, as well as rainfall erosivity, further increased the explanatory power, with q values of 47.0% and 40.4%, respectively. (3) Soil erosion intensity was projected to slightly increase by 2030 and 2040 under the natural growth and economic development scenarios, while a decline was expected under the ecological priority scenario. This study provides valuable insights for improving soil conservation efficiency and optimizing agricultural production models.

Key words: soil erosion, RUSLE model, remote sensing monitoring, driving mechanism, multi-scenario prediction, spatiotemporal variation