外来入侵假臭草在中国分布区的预测  

Predicted Potential Distribution of Praxelis clematidea in China

作　　者：邱宠华[1] 王奇志[1] 余岩[2] 

机构地区：[1]华侨大学生物工程与技术系,厦门361021 [2]四川大学生命科学院,成都610064

出　　处：《应用与环境生物学报》2011年第6期774-781,共8页Chinese Journal of Applied and Environmental Biology

基　　金：福建省自然科学基金项目(No.2010J05076);华侨大学引进人才基金项目(No.09BS305)~~

摘　　要：假臭草原产南美洲,上世纪80年代首次记载传入中国,是潜在的恶性杂草.为提供其在中国分布区的预测资料,采用397个采集数据和已有的146个分布点数据,利用生态位模型(GARP)和Domain模型对分布区进行预测分析,并通过AUC与MaxKappa进行模型评价.结果显示:近十年,假臭草入侵记录增多,入侵扩散加剧;太阳辐射、水汽压、极端高温、霜日频率和水流方向对假臭草的分布限制较小,雨日频率、坡向、坡度、海拔、年降水、年均温、地形指数和极端低温等8个环境因子对其分布影响显著;入侵的主要植被类型是农田、灌木或草类覆盖嵌合区,常绿、落叶灌木地带,栽培植被及相关联地带等.模型评价表明两种模型预测能力为较好及以上,预测结果显示:假臭草在西藏(除东南小块区域)、新疆、青海、内蒙古、宁夏,陕西和山西北部地区,黑龙江、吉林、辽宁(除辽东半岛小块区域)是不宜生长区;最易入侵广东、广西、海南等地区,其中华中部分省份、华北平原为适生区,还未见有假臭草入侵的报道,相关部门应引起重视,严防入侵.Praxelis clematidea, native to South America, was first reported in 1980s and has already been a noxious weed in China. In order to provide early warning for P. clematidea invasion, GARP and Domain models were applied to predict the species’ distribution range based on the 397 recorded points and 146 known points, and the two models were evaluated by AUC and Maxkappa. The results indicated that the distribution records of P. clematidea had increased and it had further spread during the past 10 years. Solar radiation, water vapour pressure, maximum temperature, frost days, and flow direction had little impact, while 8 environmental factors such as wet days, aspect, slope, elevation, annual precipitation, mean annual temperature, topographic index and minimum temperature had great influence on P. clematidea potential range. The invaded vegetation types of P. clematidea were the mosaic of cropland/shrub or grass cover, the shrub cover, the artificial surfaces and associated areas, etc. Model evaluation showed that the results by two models were good. The results of prediction showed that Tibet （except a small area in southeast）, Xinjiang, Qinghai, Inner Mongolia, Ningxia, north Shanxi and Shanxi, Heilongjiang, Jilin and Liaoning （except a small area in Liaodong Peninsula） were predicted as less vulnerable to P. clematidea invasion; P. clematidea was most liable to invade Guangdong, Guangxi and Hainan, etc. The invasion of P. clematidea had not been reported in some provinces of central China and the centre of North China Plain, which are likely to be suitable distribution area. Therefore urgent action should be taken and management must be reinforced to prevent this species from further spreading. Fig 4, Tab 3, Ref 40

关 键 词：GARP模型 Domain模型 入侵植物 假臭草 分布区预测 模型评价 

分 类 号：S451[农业科学—植物保护]