Abstract
Due to the rapid urbanization and global climate change, arid and semi-arid regions are becoming more vulnerable to the growing disparity between the supply and demand of ecosystem services. Conventional ecological risk assessments have predominantly emphasized landscape pattern analysis, often overlooking considerations related to human well-being. This study focuses on the Xinjiang Uygur Autonomous Region (XUAR) and examines four key ecosystem services: water yield (WY), soil retention (SR), carbon sequestration (CS), and food production (FP). Using the InVEST model, geographic information system (GIS) spatial analysis, and statistical methods, we quantify the supply-demand dynamics of these ecosystem services and identify the risk classification of ecosystem service supply-demand (ESSD) using the Self-Organizing Feature Map (SOFM) method. The results show that: (1) From 2000 to 2020, the supply and demand of WY in XUAR increased from 6.02 × 10(10) m(3) and 8.6 × 10(10) m(3) to 6.17 × 10(10) m(3) and 9.17 × 10(10) m(3), respectively; SR supply and demand decreased from 3.64 × 10(9) t and 1.15 × 10(9) t to 3.38 × 10(9) t and 1.05 × 10(9) t, respectively; CS supply and demand rose from 0.44 × 10(8) t and 0.56 × 10(8) t to 0.71 × 10(8) t and 4.38 × 10(8) t, respectively; and FP supply and demand increased from 9.32 × 10(7) t and 0.69 × 10(7) t to 19.8 × 10(7) t and 0.97 × 10(7) t, respectively. A clear spatial differentiation in ESSD was observed, with higher supply areas mainly located along river valleys and waterways, while demand is concentrated in the central cities of oases. (2) The deficit areas for WY and SR are large and show a gradual expansion, while the deficit areas for CS and FP are small and shrinking. (3) The supply-demand ratio combined with the supply-demand trend index identifies four types of bundles in the study area: B1 (WY-SR-CS high-risk), B2 (WY-SR high-risk), B3 (integrated high-risk), and B4 (integrated low-risk), with B2 being dominant. Based on these findings, ecological management recommendations are proposed for different bundles based on ESSDR classifications. This study provides a comprehensive ecological risk assessment grounded in ESSD, offering important perspectives for ecological management in arid and semi-arid regions.