Analysis of ecosystem services supply–demand relationship and influencing factors in Guangxi at different scales

Introduction Ecosystem service (ES) supply–demand relationship, directly affects human well-being and has become a critical aspect of ecological civilization construction. However, previous studies have focused more on ES supply–demand relationship at a single scale, neglecting the scale sensitivity of ecosystems. Therefore, this study employs Guangxi as a case study. Methods A multi-scale analysis framework was constructed, embedding municipal, county, and grid scales. The InVEST and RUSLE models were comprehensively applied to evaluate the supply–demand relationships and matching patterns of key ESs, including food production, carbon sequestration, soil conservation, habitat quality, and water yield. A geographically weighted regression model was utilized to analyse the scale dependency and spatial non-stationarity of influencing factors. Results (1) the matching patterns of ES supply and demand in Guangxi exhibit significant scale dependency. The city level presents an “overall surplus with structural deficits” characteristic; the county level reveals localised imbalances masked by municipal averages; while the grid scale further identifies micro-deficit patches within continuous spatial patterns. (2) Spatial correlation patterns between ES supply and demand undergo systematic evolution with increasing scale resolution. Municipal scale predominantly exhibit low-supply–low-demand agglomeration, while county scale show positive agglomeration for food and soil conservation services alongside negative correlations for water production services. Grid scales highlight localised mismatches between soil conservation and habitat quality. (3) Driving mechanisms demonstrate pronounced spatial heterogeneity and scale sensitivity. Factors such as forest cover and precipitation generally exert positive influences, yet their effect intensity and directionality vary spatially. As the scale becomes finer, the dominant factors shift from natural baseline conditions to human activity factors. At the county scale, a dual-track driving mechanism emerges where “natural factors dominate supply services while human factors dominate demand services.” At the grid scale, the influence of human activities further intensifies. Discussion This study reveals the transmission patterns of ES supply–demand relationships and the scale effects of driving mechanisms from a multi-scale perspective, providing scientific basis for differentiated and precise ecological restoration and spatial governance optimized for national territorial space.