A structure-function integrated ecological index for ecological quality assessment in urbanizing subtropical regions

Conventional remote sensing ecological assessments are limited when applied to subtropical high-biomass regions with dense vegetation cover because greenness indicators tend to saturate and cannot adequately capture ecosystem structure and function. Using Guangdong Province, China, as a case study, we developed an improved Remote Sensing Ecological Index (IRSEI) by integrating kernel NDVI (kNDVI) to mitigate vegetation saturation, the abundance index (AI) to characterize structural complexity, and net primary productivity (NPP) to represent ecosystem functional vitality. We then assessed ecological quality changes from 2001 to 2022 and used XGBoost-SHAP to identify key natural and anthropogenic drivers and their nonlinear threshold effects. Ecological quality in Guangdong improved overall, with 65.98% of the area showing restoration and 19.00% showing degradation. Ecologically high-quality areas were concentrated in the northern ecological barrier, whereas degraded areas clustered in the Pearl River Delta and the estuarine zones of eastern and western Guangdong. Future trend analysis indicated that 60.48% of the province is likely to sustain improvement, while 16.41% faces a risk of sustained degradation and 5.50% is vulnerable to shifting from improvement to degradation. Topographic factors dominated the spatial differentiation of ecological quality, whereas climatic and anthropogenic factors exhibited pronounced nonlinear and threshold effects. These findings suggest that IRSEI reduces saturation-induced bias and provides a more reliable assessment of ecological quality in subtropical high-biomass regions, offering a scientific basis for ecological monitoring, early warning, and differentiated spatial management.