Spatiotemporal evolution, driving factors, and policy impacts on ecological quality in the li river basin from 1995 to 2021

Ecological quality (EQ) protection constitutes a cornerstone of sustainable development. This study develops a Remote Sensing Ecological Index (RSEI) via the Google Earth Engine (GEE) platform, integrating Theil-Sen slope estimation, spatial autocorrelation, GeoDetector, geographically weighted regression (GWR), and text mining to investigate EQ dynamics in the Li River Basin (1995-2021). Key findings reveal: (1) The basin exhibited a "low-central, high-peripheral" RSEI pattern with a fluctuating upward EQ trend that peaked in 2003 (forest conservation and urban green policies) and 2009 (accelerated farmland-to-forest conversion and energy-saving initiatives). Mean RSEI values progressed from 0.59 to 0.60 and 0.62 across three phases, with 61.06% of areas showing improvement (predominantly construction/agricultural zones) versus 31.52% degradation (concentrated in water-source forests). (2) Land use intensity and elevation emerged as primary determinants of RSEI variability, while tourism activity intensity demonstrated escalating influence over time. (3) RSEI exhibits spatial autocorrelation, with the effects of slope, elevation, land use intensity, and tourism activity varying geographically and undergoing dynamic transitions across regions. (4) Policy evolution in the basin reflects a progressive shift toward sustainable landscape resource management and ecosystem conservation. This study provides crucial insights for preserving karst ecosystems and promoting landscape resource sustainability worldwide.