Assessment of land desertification in the Hexi Corridor of China in the last 20 years: Trends and drivers

• IRSDI effectively detects land desertification in arid and semi-arid regions. • SHAP-based mechanistic analysis reveals dual-factor coupling effects in desertification drivers. • Time-lag effects of climatic factors influence desertification trends. Located in the arid and semi-arid regions of northwest China, the Hexi Corridor is a crucial national ecological barrier with significant strategic importance for ecological security. To this end, the study constructed the Improved Remote Sensing Desertification Monitoring Index (IRSDI) to systematically monitor the desertification dynamics in the region and to deeply analyse the spatial and temporal characteristics of desertification during the period of 2000–2020. In addition, the study quantified the importance of different driving factors and their interactive effects, and further revealed the lagging and cumulative effects of climate factors on desertification. Key findings: (1) The IRSDI model achieved 90.65 % classification accuracy and a Kappa coefficient of 0.878, proving effective for large-scale monitoring. (2) Desertification in the Hexi Corridor exhibited significant regional heterogeneity, with a spatial pattern of “desertification in the northwest, greening in the southeast.” Over the past 20 years, 26.49 % of the area improved (less desertified), while 73.51 % worsened. Future predictions indicate 73.16 % of areas will continue to deteriorate, while 25.80 % will improve. (3) Soil total nitrogen (TN) was identified as the most influential driver (SHAP = 0.037). Fractional vegetation cover (FVC), total phosphorus (TP), slope, and digital elevation model (DEM) exerted inhibitory effects on desertification, whereas wind speed promoted it. Key coupling effects (FVC ∩ TN, DEM ∩ TN, and precipitation ∩ wind) further modulated desertification dynamics, revealing synergistic mitigation pathways. Concurrently, climate change and human activities jointly drove desertification in 50.8 % of the region, with distinct lag-cumulative effects observed for climatic factors: temperature (1-month cumulative effect with 1–3 month lag), precipitation (1–3 month lag), and wind speed (1–2 month lag).