基于生态位模型的松材线虫潜在生境预测方法研究  

Methods for Predicting Potential Habitat of Bursaphelenchus xylophilus based on Ecological Niche Model

作　　者：吴文浩[1] 李明阳[1] 

机构地区：[1]南京林业大学森林资源与环境学院,江苏南京210037

出　　处：《林业调查规划》2009年第5期33-38,共6页Forest Inventory and Planning

基　　金：国家林业局948项目"重要外来森林病虫害记载成图与预测技术引进"(2008-4-56)

摘　　要：以江苏省松材线虫(Bursaphelenchus xylophilus)定点发生数据和31个环境变量为主要信息源,采用分类与回归树模型(CART)、基于规则的遗传算法(GARP)、最大熵法(Maxent)、逻辑斯蒂回归(LR)4种模型建立松材线虫在江苏省的潜在生境预测模型,从接受者曲线下面积(AUC)、Pearson相关系数、Kappa值3个方面来检验模型预测精度,分析松材线虫的空间分布规律及其环境影响因素.研究结果表明:在3个评价指标中,CART模型的预测精度较低,其它3个模型的预测精度均达到优良水平,其中Maxent在物种现实生境模拟、主要生态环境因子筛选、环境因子对物种生境影响的定量描述方面都表现出优越的性能.GARP模型对松材线虫潜在生境的预测方面表现出优越的性能.海拔、年降水量、降水的季节性变化、温度的年变化范围是影响松材线虫空间分布的主要环境因子.The occurrence point and 31 environmental variables of Bursaphelenchus xylophilus were gath- ered, and 4 models namely Classification and Regression Trees （CART） , Genetic Algorithm for Rule-set Prediction （GARP）, Maximum Entropy Method （Maxent） , and Logistic Regression （LR） were intro- duced to generate the models for potential prediction of invasive species of Bursaphelenchus xylophilus in Jiangsu of China. Then 3 statistical criteria of the area under the Receiver Operating Characteristic Curve （AUC）, correlation （COR） and Kappa were calculated to evaluate the performance of the models, fol- lowed by analyses of major contribution variables and environmental influential factors. The results showed that in term of three statistical criteria, except prediction results of the CART, other ecological niche models were excellent and outstanding, in which Maxent presented supper property in three aspects of imitating actual habitat, selecting for major environmental factors, quantitative description of the influ- ence of environmental variables on habitat. GARP showed well in predicting potential habitat for Bursaph- elenchus xylophilus. Elevation, annual precipitation, precipitation seasonality and temperature annual range were the four major environmental factors, which influenced spatial distribution of Bursaphelenchus xylophilus.

关 键 词：松材线虫 潜在生境 生态位模型 预测模型 环境因子 

分 类 号：S718.5[农业科学—林学] S763.305