Ecological fire risk assessment based on remote sensing ecological index and machine learning methods

Forests, key to terrestrial ecosystems and human resources, suffer severe fire damage, with global risks elevated by extreme climates. Limited fire-fighting resources require targeted allocation to high-risk, high-ecological-value areas. This study conducts ecological fire risk assessment in China’s fire-prone Greater Khingan Mountains using machine learning and ecological indices. It constructs a fire risk map via 12 fire factors and logistic regression, and an ecological quality map via RSEI. Overlaying these maps identifies high-risk areas in the north and southwest, with 12.7% of the region at relatively high risk. The result of the spatial autocorrelation analysis shows that the Moran’s I value of ecological fire risk in the Greater Khingan Mountains Region is 0.2, indicating a strong clustering effect in the ecological fire risk areas. The research provides scientific basis for resource allocation, fire management and ecological protection, supporting sustainable forest development under resource constraints.