Climatic and Topographic Factors Driving Large Wildfires: Decision Tree-Based Risk Mapping in South Korea

This study identifies the key climatic and topographic factors influencing large wildfires in South Korea and develops a spatial methodology for risk mapping. Using wildfire records (2017-2023) and meteorological datasets, the Classification and Regression Tree (CART) model was applied to extract the significant variables. Maximum surface sensible heat flux (sshf_max), maximum shortwave radiation (ssrd_max), average slope, and maximum evapotranspiration were the most influential variables, with sshf_max being the primary split variable. These variables were normalized and weighted using a WLC approach to produce a wildfire risk map for Gyeongsangbuk-do. This province has been most affected by wildfires in the past decade. The southeastern region was classified as having the highest risk (level 1), whereas the northwestern regions had lower risk (levels 2-3). Despite the limited number of large-fire cases, this study offers a baseline framework to prioritize fire prevention and resource allocation. Future improvements are expected through the integration of global datasets and advanced machine learning.