Abstract: Forest fires, due to their frequent occurrence and unpredictability, pose a severe threat to forest ecological security and regional sustainable development. Scientific and rational delineation of fire risk grade zones has become a core way to enhance disaster prevention and control capabilities. Taking the Qierqi Forest Farm in Heihe City, Heilongjiang Province as the research object, we collected vector data and remote sensing data of the forest farm. We classified all the 34 forest fire-influencing factors into four categories: topographic factors, vegetation factors, climatic factors, and human factors. The principal component analysis (PCA) function in ArcGIS 10.8 software was used to conduct analysis of the influencing factors. Based on the principal component contribution rate and the principal component factor loading matrix in the analysis results, 15 factors with low internal correlation but strong correlation with fire factors were selected. The Super Decision 3.0 software was used to construct a factor network model to determine the factor weights. On this basis, the weighted superposition evaluation method was used to conduct a comprehensive assessment of the forest fire risk level. According to the natural break point method, the forest fire risk level was divided into five grades: high, medium-high, medium, medium-low, and low, accounting for 15.1%, 29.3%, 24.5%, 18.9%, and 12.2% of the study area, respectively. The zoning results were verified by historical fire data. The results showed that the distribution of historical fire points was highly consistent with the fire risk levels. By superimposing and comparing with the traditional analytic hierarchy process, the reliability and effectiveness of PCA in superimposing after the selection of fire risk factors were confirmed. Our results can provide scientific basis for the early warning of forest fires, the planning and allocation of fire-fighting resources in the forest farm, which is of great significance for improving the forest fire prevention and control capabilities of the forest farm and ensuring the safety of forest resources.

Key words: forest fire risk level, principal component analysis, analytic network process, weighted superposition, Qierqi Forest Farm