Font Size: a A A

Dynamic And Prediction Model Development Of Forest Fire Occurrence In Jilin Province

Posted on:2024-04-19Degree:MasterType:Thesis
Country:ChinaCandidate:B GaoFull Text:PDF
GTID:2543307292495734Subject:Forest Protection
Abstract/Summary:PDF Full Text Request
Forest fire is a kind of natural disaster with strong suddenness,wide range of occurrence and great harm.Prevention is more important than fighting against this natural disaster.This study used historical fire data to analyze the occurrence law of forest fire in Jilin Province,and established a prediction model for the probability of forest fire occurrence in Jilin Province based on meteorological,topographical,vegetation,human and socio-economic factors,aiming at providing technical support for the prevention and control of forest fire in this area.The main research results were as follows:(1)From 1970 to 2019,forest fires in Jilin province exhibited significant temporal fluctuations and were spatially concentrated.There were 6919 forest fires in Jilin Province,which caused a total area of 137833.22 hm~2 and a burnt forest area of 71176.50 hm~2in the meantime.The 1990s marked a turning point for forest fires in Jilin province,as forest fires entered a stable period and the harm to the forest ecosystem decreased relatively.The occurrence period of forest fire in Jilin Province in spring was significantly longer than that in autumn.In recent years,the occurrence period of forest fire in spring and autumn had been shortened to varying degrees.In different years,forest fires had occurred from March 26 to May 25 and from September 28 to November 9,which should be focused on prevention.The number of forest fire points gradually decreased from the east to the west of Jilin Province,and most of them were distributed in the east and south.In the past 50 years,most of the forest fires that caused a large total area of fire were distributed in Yanbian Prefecture and Baicheng City,and most of the forest fires that caused a large burnt forest area were distributed in Yanbian Prefecture.(2)The fire sources that caused forest fires in Jilin Province were relatively complex,but the cycle period of forest fire was constantly extending.The number of forest fires with known fire sources was 5406 in Jilin Province,which caused a total area of 97450.19 hm~2and a burnt forest area of 46762.83 hm~2.At present,farming fire,sacrificial fire,outdoor smoking,arson,dementia fire,outdoor living fire and foreign burning were the main fire sources in Jilin Province.In the past 50 years,the cycle period of forest fire in Jilin Province had been continuously extended,from 1240 years in the 1970s to 101323 years in the 2010s.(3)There were certain regularities in the temporal and spatial dynamic changes of forest fires in Jilin province.On an annual scale,the forest fires in the 1970s were the most,with the largest total fire area and burnt forest area.On the monthly scale,forest fires were distributed in a bimodal pattern,mainly concentrated in March to May and September to November.The number of forest fires,total fire area and burnt forest area in spring was slightly higher than that in autumn.On the daily scale,the forest fire distribution was unimodal,and the forest fire occurrence frequency,total fire area and burnt forest area from 11:00-13:59 was higher.In space,compared with the 1970s,the forest fire spots in Jilin Province had shifted from east to west.In recent years,the central and southern part of Jilin Province had become another concentrated forest fire distribution area except the eastern part of Jilin Province.The total fire areas and burnt forest areas were relatively larger in the eastern and western of Jilin Province,which should be given special attention for prevention.(4)The established forest fire occurrence probability prediction models had good fitting effect and high prediction accuracy,which could provide technical support for developing scientific and effective forest fire management strategies.The prediction accuracy of the two machine learning models was higher than that of the three generalized linear regression models.RF had the highest prediction accuracy,followed by SVM model and GWLR model,and Lasso model and LR model had relatively low prediction accuracy.Meteorological factors were the main driving factors of forest fire in Jilin Province,while topographic factors,human and socio-economic factors had similar influence on the probability of forest fire occurrence.Level IV and V risk areas with high risk of forest fire were mainly distributed in the east and south of Jilin Province,and there was also sporadic distribution in Siping City and Changchun City.
Keywords/Search Tags:Jilin Province, Forest fire, Distribution pattern, Occurrence prediction, Forest fire driving factors
PDF Full Text Request
Related items