修正的灰色模型在森林生物灾害发生面积预测中的应用  

Application of improved grey model in forecasting occurrence area of forest bio-disaster

作　　者：苓建强[1] 周英梅[1] 温俊宝[1] 许志春[1] 骆有庆[1] 

机构地区：[1]北京林业大学省部共建森林培育与保护教育部重点实验室,北京100083

出　　处：《中国森林病虫》2011年第1期19-22,43,共5页Forest Pest and Disease

基　　金：长江学者与创新团队发展计划(PCSIRT0607)

摘　　要：以2000-2008年《中国林业统计年鉴》中生物灾害发生面积为依据，采用原序列和残差序列进行灰色动态预测模型aM（1，1）建模，分别建立林业病虫鼠害发生总面积、病害发生面积、虫害发生面积、鼠害发生面积、松毛虫发生面积、杨树食叶害虫发生面积和杨树蛀干害虫发生面积及各指标重度发生面积的灰色预测模型，对森林生物灾害发生面积进行预测。通过精度检验，应用残差序列所建的14个指标模型中，有3个模型非常准确（平均相对误差小于1％），5个模型很准确（平均相对误差在1％-5％之间），5个模型较准确（平均相对误差在5％～10％），只有1个模型准确率较低（平均相对误差大于10％）。所建立的模型适用于发生面积趋势预测，准确率非常高，预测结果有较高置信度，且给出了5a预测结果。该研究为我国森林生物灾害预测提供了新的途径。Based on the publication data from China Forestry Statistical Yearbook from 2000 to 2008, the grey dynamic forecast model GM（ 1,1 ） was established with raw data and residual error series. The models of total occurrence area of forest bio-disasters, the occurrence area of diseases, the occurrence area of insect pests, the occurrence area of rodent pests, the occurrence area of pine caterpillars, the occurrence area of poplar defoliators and the occurrence area of poplar stem borers were established respectively and the trend prediction of the occurrence area were made. The results showed that among the 14 models established, the average relative errors of 3 models were less than 1%, 5 were between 1% and 5% ,the other 5 were between 5% and 10% and only 1 was more than 10%. The verified grey model GM （ 1,1 ） was suitable for the trend prediction of the area. The improved models passed accuracy test. The results of prediction for the next five years were given. The study provided a new way of predicting occurrence area of forest bio-disaster.

关 键 词：发生面积 森林生物灾害 灰色模型 预测 

分 类 号：S763.305[农业科学—森林保护学]