Risk assessment of wildfire disaster in Guilin based on XGBoost and combination weight method
The present paper conducts a risk assessment of wildfire disasters to quantitatively evaluate the risk of social and ecological losses caused by wildfire disasters,which is an indispensable part of its research objectives.Taking Guilin as the study area,and based on the historical wildfire data from 2017 to 2021.A grid of 30 m x 30 m size was used as the assessment unit to select 15 susceptibility assessment factors such as elevation,slope,rainfall,and temperature.XGBoost algorithm was used not only to analyze the wildfire susceptibility but also to analyze the main factors affecting the occurrence of wildfire.Secondly,8 disaster vulnerability factors such as old and young population,female population,GDP,and road density were selected for each district and county as assessment units.A combined weighting method based on an analytic hierarchy process(subjective weighting)and entropy weighting method(objective weighting)was designed to assess the vulnerability of wildfire.Finally,the risk of wildfire disasters in Guilin was assessed and analyzed based on the established susceptibility and vulnerability assessment model.The results indicate that:(1)The AUC value of the wildfire susceptibility model is 0.993,and the prediction accuracy is 0.969 8,indicating that the wildfire susceptibility prediction model has excellent prediction performance.Temperature,elevation,slope,and wind speed are the most important factors affecting the occurrence of wildfire disasters.(2)Vulnerability assessment result of wildfire disasters is consistent with objective fact and have good credibility.(3)The proportions of the areas with very low,low,medium,high,and very high risk to the region's total area are 28.969％,14.681％,11.642％,29.986％,and 14.722％,respectively.The assessment result is not only highly consistent with the spatial occurrence probability and social carrying capacity of wildfire disasters in Guilin but also validates the feasibility and effectiveness of the proposed model.In addition,it can provide important decision support for disaster prevention and mitigation by disaster management authorities in Guilin.
public safetywildfire disasterwildfire risk assessmentXGBoost algorithmsubjective-objective combination weight
万方数据
维普
