基于MAXENT模型和ArcGIS预测豚草在中国的潜在适生区  

Prediction of potential suitable area of Ambrosia artemisiifolia L.in China based on MAXENT and ArcGIS

作　　者：柳晓燕[1] 李俊生[1] 赵彩云[1] 全占军[1] 赵相健[1] 宫璐[1] 

机构地区：[1]中国环境科学研究院,北京100012

出　　处：《植物保护学报》2016年第6期1041-1048,共8页Journal of Plant Protection

基　　金：国家重点研发计划(2016YFC1201100)

摘　　要：为分析有害入侵杂草豚草Ambrosia artemisiifolia L.在中国的潜在适生区,基于其在中国的分布数据,结合气候、高程、土地利用类型等环境数据,利用最大熵模型(MAXENT)与ArcGIS软件相结合模拟豚草在中国的潜在适生区,运用受试者工作特征(ROC)曲线分析法进行验证,并运用刀切法分析影响豚草分布的主要环境变量。结果表明,中国东部大部分地区都属于豚草适生区,其适生面积为2.31×10~6km^2,占全国陆地总面积的24.10%。辽宁省东部、安徽省中部、江苏省南京市、浙江省杭州市、湖北省东部、湖南省东部、江西省及两广中部为豚草高适生区(适生值>0.5)。ROC曲线分析法得出下面积AUC值为0.955,表明预测结果可靠。刀切法分析显示,最冷季度平均湿度和最湿季度湿度对豚草在中国的分布状况影响最大,贡献率分别为30.89%和27.86%;最暖季度平均温度、温度变化方差、昼夜温差与年温差比值和土地利用类型有一定影响,贡献率依次为14.19%、8.05%、6.30%和5.93%;最冷月份最低温和坡向对豚草分布影响较小,贡献率分别为0.78%和0.20%。Based on the distribution of Ambrosia artemisiifolia L.and environmental data,such as climatic factors,DEM and landuse information,the potential suitable area of this species in China were predicted by using MAXENT modeling and ArcGIS.Moreover,the predictive data was tested and verified by receiver operating characteristic（ROC）,and the influence of the main environmental variables on distribution of A.artemisiifolia was analyzed with Jackknife test.The results showed that A.artemisiifolia may widespread in eastern China and its suitable area was 2.31 × 10~6 km^2,which accounts for 24.10%of the total land area of China.The highly suitable area（suitability index0.5） mainly located in eastern Liaoning Province,middle of Anhui Province,Nanjing of Jiangsu Province,Hangzhou of Zhejiang Province,eastern Hubei Province,eastern Hunan Province,Jiangxi Province,northern Guangdong Province,and middle of Guangxi Zhuang Autonomous Region.The AUC of ROC curve was 0.955,indicating that the predictive results with the MAXENT model were highly precise.Jackknife test showed that the mean precipitation of the coldest quarter and the wettest quarter,mean temperature of the warmest quarter,temperature seasonality,isothermality,and landuse dominantly determined the distribution of A.artemisiifolia,and the contribution rates were 30.89%,27.86%,14.19%,8.05%,6.30%and 5.93%,respectively,while the lowest temperature of the coldest month and slope aspect had less impact on the distribution of A.artemisiifolia,with a contribution rate of 0.78%and 0.20%,respectively.

关 键 词：豚草 MAXENT模型 外来入侵生物 潜在适生区 

分 类 号：S451[农业科学—植物保护]