The spatial–temporal changes and driving factors of desertification in the Wuliangsuhai watershed based on remote sensing

• Barren lands has the largest reduction. • The results of the DDI analysis showed a decrease from 2020 to 2022. • The FVC had the highest level of explained variance. • The desertification has developed in a positive direction. Wuliangsuhai watershed is a large freshwater lake in Hulunbuir City, Inner Mongolia Autonomous Region, with important ecological, cultural and economic values. Based on the Landsat 8 OLI remote sensing images, this study used the object-oriented CART decision tree to map the land cover, the pixel dichotomy model to estimate the FVC and Albedo-NDVI formula to estimate DDI in the Wuliangsuhai watershed in 2020, 2021 and 2022. Besides, Geo-informatic tupu and geographical detector model were also used to analysis the changing patterns and drivers of desertification. The results indicated a decline in desertification in the Wuliangsuhai watershed from 2020 to 2022. The area of barren lands exhibited a 17.41% reduction with a 37.78% decline in the FVC in the no vegetation class. The area of the DDI in the no desertification level increased, while that of the extremely severe desertification level decreased. Furthermore, the geo-informatic tupu indicated that the desertification improvement area showed consistent growth during the study period. However, a considerable proportion of the grasslands (32.73%) remained degraded, and the area of FVC in the moderately dense vegetation class exhibited a notable decline (21.32%). And the area affected by severe desertification level of DDI increased. This can be attributed to a combination of climatic, ecosystem conditions and human impacts. In particular, the reduction in precipitation, coupled with the steep slopes and negative human activities, has accelerated the desertification process. Consequently, this study provides recommendations for policy makers to prioritize negative human activities and soil erosion in the Wuliangsuhai watershed.