气象法预测盘古林场可燃物含水率的外推精度  

Meteorological elements regression method is used to predict Pangu forest farm extrapolation accuracy analysis of fuel moisture content

作　　者：张恒[1,2] 金森[1] 张运林[1] 于宏洲[1] 

机构地区：[1]东北林业大学,黑龙江哈尔滨150040 [2]内蒙古农业大学,内蒙古呼和浩特010019

出　　处：《中南林业科技大学学报》2016年第12期61-67,共7页Journal of Central South University of Forestry & Technology

基　　金：中央高校基本科研业务费专项资金项目资助(2572015BA03);林业公益性行业科研专项(201204508)

摘　　要：在春防期和秋防期对东北地区大兴安岭盘古林场的3种典型林分：樟子松林、兴安落叶松林、白桦林下的地表细小可燃物含水率进行长时间测定。得到其预测模型,使用模型对大兴安岭现有的4种可燃物含水率预测模型进行外推分析。结果表明：这些模型外推时得到的平均绝对误差（MAE）是自建模型的1.6~7.2倍。以1.5倍左右作为可替代的标准,则这4个模型与自建模型应具有一定的可替代性,即可用外来模型预测本地可燃物的含水率。春防期的模型预测精度高于秋防期。从外推精度的可用性来看,4个模型外推误差在绝对值上最小为3.6%,最大为17.5%。因此,现在的这4种可燃物含水率预测模型的外推能力都不理想。模型外推误差与建模地区和外推地区的微环境差异有关,与距离不是完全成正比。Fuel moisture dynamics of surface dead fine fuels under larch stand, Scotts pine stand and birch stand in Pangu forest farm,Greater Khingan Mountains, Heilongjiang province was observed in spring and autumn. A local weather-variable-regression model for fine fuel moisture prediction was established using the measured moisture data and meteorological data. Four models of the same type from literatures established from dataset collected from places other than the study area were used for predicting fuel moisture of the local forest fuels. Evaluated from the ratio of mean absolute error （MAE） of the four extrapolated models and the local model and taking 1.5 as a benchmark of the ratio, the four models with a ratio range of 1.6 to 7.2, were exchangeable with local models, accuracy of model prediction for fire protection period is higher in Spring than Autumn. Considering the availability of extrapolation accuracy, the minimal absolute value of extrapolated errors was 3.6% and the maximum was 17.5%. While assessed by absolute errors, the four models have a bit larger errors than required. The extrapolated errors were due to differences of local conditions between different data collection area, not proportional to the distance between these regions.

关 键 词：大兴安岭 可燃物含水率 预测模型 精度验证 

分 类 号：S762.2[农业科学—森林保护学]