基于MaxEnt模型的云南红豆杉潜在适宜分布预测

摘要

云南红豆杉(Taxus yunnanensis)是国家一级保护植物,具有重要的经济价值和药用价值,预测其潜在适宜分布区域具有重要的理论和实际意义.选择云南红豆杉的主要分布区云南省作为研究区域,基于有效的地理分布数据及筛选后的环境变量,运用最大熵模型(MaxEnt)与地理信息系统(GIS)对云南红豆杉的潜在适宜分布区域进行了预测,进而以刀切法(Jackknife)分析影响其分布的主要环境因子,并采用受试者工作特征曲线(ROC)对预测结果进行检验.结果表明,云南红豆杉在云南省有较广的适宜分布区域,主要集中在滇西北、滇西及滇北等地;海拔和年平均温度对云南红豆杉适宜分布的影响最大;AUC值为0.864,说明模型预测效果良好.%Taxus yunnanensis,a rare and endangered species listed in National Key Protected Wild Plants of Grade Ⅰ,has important economical and medicinal value,and it is of great theoretical and practical significance to predict its potential suitable distribution area.We selected Yunnan Province,the main distribution area of Taxus yunnanensis,as the research area.Based on the effective data of geographic distribution and environment variables,the potential suitable distribution area of Taxus yunnanensis was simulated by using the maximum entropy model (MaxEnt) and geographic information system (GIS),moreover,the primary environmental factors that affected its distribution were analyzed by method of jackknife.And we used receiver operating characteristic (ROC) curve to examine accuracy of the simulation results.The results showed that the value of AUC is 0.864,which indicated that the predicted results of MaxEnt model do well.There is a relatively broad potential suitable distribution area for Taxus yunnanensis which mainly concentrated in the northwest,west and north of Yunnan.Two environment variables of altitude and annual average temperature make great influence on the distribution of Taxus yunnanensis.