Abstract When a forest fire occurs, timely and effective initial attacks can control the fire in its early stages, reducing the risk of spread and minimizing disaster losses. Enhancing the efficiency of forest fire suppression and the success rate of initial attacks has become a critical issue. Therefore, this study thoroughly investigated the key factors of the initial attack phase of forest fires, employing the weighted cross-entropy similarity and relative entropy models, combined with Geographic Information Systems (GIS) and machine learning technologies for precise simulation and analysis. The research introduced additional firefighting water reserve points in the study area to enhance the suppression capability and efficiency of unit firefighting resources, thereby increasing the success rate of the initial attack and reducing the losses caused by fires. An empirical study in Xichang City, Sichuan Province, China, demonstrated that this method significantly improves the success rate of initial attacks. Simulation results indicate that, under various weather conditions, especially extreme weather, the newly established firefighting water reserve points greatly enhance the success rate of initial attacks. This approach not only aids in the scientific planning of firefighting resources but also provides a practical foundation and theoretical guidance for future research to further improve the success rate of initial attacks.
Zhang et al. (Thu,) studied this question.