江西南昌活可燃物含水率动态变化和预测模型  

Dynamic change and prediction model of live fuel in Nangchang,Jiangxi province

作　　者：周勇 张贵[1] 张运林 丁波 满子源 ZHOU Yong;ZHANG Gui;ZHANG Yunlin;DING Bo;MAN Ziyuan(College of Forestry,Central South University of Forestry&Technology,Changsha 410004,Hunan,China;College of Mathematics and Statistics,Hengyang Normal University,Hengyang 421002,Hunan,China;School of Biological Sciences,Guizhou Education University,Guizhou 550018,Guiyang,China;Key Laboratory National Forestry Administration on Ecological Hydrology and Disaster Prevention in Arid Regions,Northwest Surveying,Planning and Designing Institute of National Forestry and Grassland Administration,Xi’an 710048,Shanxi,China)

机构地区：[1]中南林业科技大学林学院,湖南长沙410004 [2]衡阳师范学院数学与统计学院,湖南衡阳421002 [3]贵州师范学院生物科学学院,贵州贵阳550018 [4]国家林业和草原局西北调查规划设计院旱区生态水文与灾害防治国家林业局重点实验室,陕西西安710048

出　　处：《中南林业科技大学学报》2021年第8期28-35,共8页Journal of Central South University of Forestry & Technology

基　　金：国家自然科学基金项目(31370656);贵州省科技支撑项目(2017-2520-2)。

摘　　要：【目的】随着全球气候变暖,树冠火和极端火行为频繁发生,森林活可燃物含水率研究越来越受重视。然而,由于活可燃物含水率动态变化要比死可燃物含水率动态变化复杂,除气象要素外,还受本身理化性质的影响严重,因此目前关于活可燃物含水率的研究却很少。江西省森林资源丰富,森林火灾频发,但关于活可燃物含水率的研究几乎没有。【方法】以江西南昌3种典型活可燃物:毛竹、杨梅和油茶为研究对象,在防火期内以日为步长监测其含水率动态变化情况,并同步监测气象要素,分析活可燃物含水率与气象要素、湿度码的相关性,并基于逐步回归,分别建立气象要素预测模型和湿度码预测模型,比较了模型预测精度。【结果】1)监测期内,3种活可燃物被引燃的可能性由高到低依次为毛竹、油茶和杨梅;2)当日和前一日空气温度与杨梅和油茶活可燃物含水率呈极显著正相关,毛竹仅与前一日空气温度呈正相关。3种活可燃物含水率随着当日和前一日相对湿度、当日降水量的增加极显著增加。风速对所有活可燃物含水率动态变化没有显著影响;3)基于气象要素的活可燃物含水率预测模型中当日相对湿度均为预测因子,线性回归模型平均绝对误差和平均相对误差变化范围分别为10.590%~15.745%和7.171%~12.914%,非线性回归模型的平均绝对误差和平均相对误差变化范围分别为为9.844%~12.657%和6.830%~10.370%;对于湿度码预测模型,仅细小可燃物湿度码进入模型,线性回归模型的平均绝对误差和平均相对误差变化范围分别为10.259%~16.732%、6.960%~13.931%;非线性回归模型的平均绝对误差和平均相对误差变化范围分别为10.696%~31.285%、8.040%2~2.010%。【结论】不论预测是因为气象要素还是湿度码,都是线性回归模型更适于预测江西活可燃物含水率,结果可为江西活可燃物含水率研究提供基础数�【objective】With the global warming,canopy fires and extreme fire behaviors frequently occur,more and more attention is paid to the research on the live fuel moisture content.However,because the dynamic change of moisture content of live fuel is more complex than that of dead fuel,besides meteorological factors,it is also seriously affected by its physical and chemical properties,so there is little research on moisture content of live fuel at present.Jiangxi Province is rich in forest resources,but there are few studies on the moisture content of live fuels.【Method】Three typical live fuels in Nanchang,Jiangxi:bamboo,bayberry,and Camellia oleiferawere elected as the research objects.During the fire prevention period,the dynamic changes of moisture content are monitored with daily steps,and meteorological elements are simultaneously monitored.Based on the correlation between moisture content and meteorological elements and moisture codes,and based on stepwise regression,the meteorological element prediction models and moisture code prediction models were established respectively,and the prediction accuracy of the models was compared.【Result】1)During the monitoring period,the probability of ignition of the three live fuels in descending order was bamboo,camellia and bayberry;2)The air temperature of the day and the previous day had a extreme significant positive correlation with the moisture content of the live fuels of bayberry and Camellia oleifera,while the bamboo was only positively correlated with the air temperature of the previous day.The moisture content of the three kinds of live fuels increased significantly with the increase of the relative humidity of the day and the day before,and the rainfall of the day.Wind speed has no significant effect on the dynamic changes of moisture content of all live fuels;3)The relative humidity of the day in the meteorological element regression method is a predictive factor.The average absolute error and average relative error of linear regression model were 10.

关 键 词：活可燃物 含水率 气象要素 湿度码 预测模型 

分 类 号：S762.2[农业科学—森林保护学]