Study On Forest Fire Risk Based On Temperature Vegetation Dryness Index

Forest fires are a natural disaster that occurs frequently around the world,causing a large amount of waste of vegetation resources every year,which not only has a huge impact on social and economic development,but also threatens the natural environment and destroys the ecological balance.Therefore,in order to protect the ecological environment,protect people’s lives and property safety,it is urgent to establish a scientific and effective fire risk monitoring and evaluation system.Hunan Province is very rich in forest resources,and there are many rare tree species,which play an important role in the development of China’s forestry economy.But at the same time,the frequency of forest fires in Hunan Province has also remained high,not only causing serious casualties,but also causing huge property losses.This paper uses the data of surface temperature,vegetation index,forest fire statistics,meteorology,humanities,vegetation distribution and other data of Hunan Province from 2008 to 2018,based on geographic information system,remote sensing and statistical analysis software Arc GIS,ENVI and SPSS.The forest fire risk probability model completes the division of forest fire risk levels in Hunan Province;and the forest fire risk zoning in Hunan Province is carried out according to the forest fire risk levels,aiming to provide science for the implementation of differentiated prevention and control measures for fire risk areas of different levels in the future Basis and strong reference.The main research contents and conclusions of this paper are as follows:(1)The spatial and temporal distribution of forest fires in Hunan Province varies greatly.From the perspective of spatial distribution,the forest fires in Hunan Province occurred in 14 cities and prefectures throughout the province,mainly distributed in central and southern Hunan and a small amount in other areas.Yongzhou and Chenzhou have more occurrences,which are the areas where forest fires occur intensively;Zhangjiajie and Xiangtan have fewer occurrences.From the perspective of time distribution,the frequency of forest fires was the highest in 2013,reaching 2055.After that,the number of forest fires decreased year by year and declined rapidly;the frequency of forest fires decreased significantly in summer,only 374,and the frequency in winter increased significantly.5589 cases;the month with the highest frequency is March,which is 3165;the month with the lowest frequency is June,only 11 times.(2)The correlation between the temperature vegetation drought index and forest fires is analyzed,and a multinomial regression forest fire risk probability model for Hunan Province is constructed.The temperature and vegetation drought index(TVDI)was calculated using surface temperature data and vegetation index data,and statistical analysis was conducted with forest fire monitoring data of Hunan Province.It was found that forest fire in Hunan Province mainly occurred in time and space when TVDI was 0.6-0.9 Within the scope of the region;using TVDI and forest fire monitoring data to establish a polynomial regression forest fire risk probability model.After testing,the determination coefficient R2 is 0.9182,indicating that the curve fitting effect is better.Using this model,the stratified map of forest fire risk in Hunan Province was obtained.(3)Nine forest fire risk factors were selected to construct the BLR forest fire risk probability model,and the correlation between risk factors and forest fire frequency was analyzed.Among them,the relationship between climate,vegetation distribution,topography,human factors and the frequency of forest fires is as follows:Forest fires are mainly concentrated in coniferous forests,shrubs and cultivated plantation forests with an average annual temperature of 16-20℃ and precipitation of 1400-2000mm.The area with an altitude of 0～800m and a slope of less than 25°.In addition,the frequency of forest fires decreases with the increase of average GDP and average population density,showing a significant negative correlation.The areas with the highest frequency are those with an average GDP of less than ￥3×107/km2 and an average population density of less than 600 people/km2.area.(4)Constructed the BLR "(Binary Logistic Regression)" forest fire risk probability model of Hunan Province.Based on the forest fire grassland fire prevention data of Hunan Province from 2008 to 2018,the BLR forest fire risk model was established based on the data of surface fire,vegetation index,meteorology,humanities and vegetation types.After testing,the AUC area of the model is as high as 0.779,far exceeding 0.5,and the model fitting effect is good,which can effectively identify fire points and non-fire points.Using this model,the probability of forest fire risk in Hunan Province is calculated,and the forest fire risk in Hunan Province is divided into five grades according to the equal interval method,and the forest fire risk grade zoning map of Hunan Province is obtained.Among them,the extremely low-risk areas account for 29.8%of the total area of the province;the low-risk areas account for 24.7%of the province;the medium-risk areas account for 15.3%of the province;the high-risk areas account for 17.7%of the province;Extremely high-risk areas account for 12.5%of the total area of the province.(5)The spatial distribution of forest fire risk in Hunan Province is significantly different.The areas with high fire risk grades and extremely high fire risk grades with a forest fire occurrence probability of>0.6 in Hunan Province are mainly distributed in central and southern Hunan,concentrated in Shaoyang,Hengyang and Yongzhou,southern Zhuzhou and parts of Chenzhou.Huaihua City,Loudi City,Xiangtan City and Changsha City are mainly moderate fire risk grade areas;Jishou City,Changde City,Zhangjiajie City Yiyang City and Yueyang City in northern Hunan have low and extremely low fire risk grades,and forest fires have occurred is less likely.