An adaptive framework for symbiotic environmental fragility and geohazards risk: A case study of Gaizhou City, China

The ecological environment is an important context for geohazards formation. Traditional geohazards risk assessment mostly focuses on the physical characteristics and pays insufficient attention to long-term impact on the ecological environment. This paper considers environmental fragility from a 'current-sensitive-responsive' perspective, comparing the entropy power, projection pursuit, and random forest methods to identify the optimal model. It also proposes a 3D matrix to establish an adaptive framework for geohazards ecological risk, tailored to scenarios of frequent geohazards in environmental fragile mountainous regions. The results were validated using Gaizhou City: Random forest achieves better zoning performance and accuracy, making it the optimal model for environmental fragility. The area classified as very high risk zoning is only 28.35 km² using 3D matrix, with a high geohazard areal density of 0.35 point/km². With less than 0.3 % of the total area, it accommodates 10.6 % of the region’s geohazard inventory, fully demonstrating its effectiveness. The townships of Kuangdonggou and Shizijie were key control units, and it was recommended that a combination of engineering and ecological measures be prioritized. The improvement method may provide a new perspective on the synergistic management of the ecological environment and geohazards. • Considering environmental fragility from "current-sensitive-responsive" dimensions, random forest model is optimal. • A 3D matrix is proposed to realize the synergistic assessment of geology and ecology. • An adaptive framework is established to serve as a reference for mountainous regions.