Abstract:

Forest fire is one of the important factors that affect the forest ecosystem. The spread and development of forest fire are deeply affected by the moisture content of forest fuels, especially the occurrence of forest fire is directly affected by the moisture content of dead fuels on the surface. Therefore, accurate prediction of the forest surface dead fuels moisture content is the key to predict forest fire risk and fire behavior, and it is particularly important to strengthen the study of forest dead fuels moisture content prediction model. This article summarizes the research status of forest fuels moisture content in terms of the research methods, research models and model accuracy, and comparatively evaluates the existing models. In view of problems in the current research, five prospects for future research are proposed: 1) Strengthen research on the dynamic of fuel moisture content in key fire risk zone. The existing forest fire danger factor collection stations and forest fire danger monitoring stations are used to obtain the monitoring data of forest fuel moisture content and meteorological factors under different environmental condition. The prediction model of forest fuel moisture content based on meteorological parameters in key fire danger zone is constructed. 2) Strengthen basic data monitoring and collection of forest fuels. In order to build a comprehensive forest fire risk rating system, the basic data monitoring and collection of forest fuels should be strengthened, a solid data foundation should be laid, and an accurate forest fuel type classification system should be established. 3) Strengthen the study on the spatial heterogeneity of the fuel moisture content. In the future research, the dynamic changes of fuel moisture content under different impact factors should be considered, especially the spatial heterogeneity of fuel moisture content of small-scale forests, so that the prediction of forest fire danger can be made more accurately. 4) Improve the accuracy of the models combined with boosted regression tree(BRT). In the study of the influencing factors of the accuracy of the fuel moisture content model, the BRT method should be used to randomly extract a certain amount of data multiple times to analyze the degree of influence of different influencing factors on the accuracy of the model. 5) Conduct research on large-scale fire risk early alarm combined with GIS. Based on RS and GIS technology, the remote sensing inversion model of the fuel moisture content is established. On the basis of accurate simulation of the spatial distribution of forest fuel moisture content, the fuel moisture content prediction models of different fire risk classes is established.

Key words: forest fuel, surface dead fuel moisture content, forecast model, model accuracy, fire danger prediction