Predicting The Potential Distribution Of Invasive Alien Weeds In China

Biological invasion cause large economic and ecological impacts on invaded areas,predicting the potential distribution of invasive alien species are not only important in risk analysis of invasions, but also provide theoretic reference on early detection and preventions.In this paper,3 species distribution predictive models:CLIMEX,GARP,and Maxent,combined with Geography Information System(GIS),were applied to predict the suitable distribution areas of several invasive weeds in China,and the predictive performance of 3 models were also evaluated.The modules "Location compare" of CLIMEX was used to predict the potential distribution of Parthenium hysterophorus L.in China under current and future climate scenarios.The results indicated that,under the current climate,the climatic suitable areas of P.hysterophorus was larger than current recorded area,including Shandong,Hunan,Hubei,Jiangxi,Zhejiang,Anhui,Jiangsu, Guangdong,Guangxi,Fujian,Jiangxi,Yunan,Guizhou,Chongqing,Shanxi,Shannxi,Hebei,and milled-east parts of Sichuan Province,and under the climatic change scenarios,the potential distribution area would increase 4.2%-9.3%and the highly suitable area would increase 4.0%-9.7%, the suitability of P.hysterophorus in China would increase 2.0-4.7 compared with current climate. However,the suitability of several tropic areas(such as Sanya)would decrease.The "Match climates" of CLIMEX was also used to analysis the potential distribution of Wedelia trilobata(L.)A.S.Hitche in China,Four distributed locations from native and other invasive regions were selected to perform climatic match to 749 climatic locations in China.The results showed that 3 locations in Taiwai Province were highly suitable,29 locations in Gongdong showed less suitability,and marginal points located in north part of Guangdong,Guangxi,southwest part of Yunnan and south part of Fujian.Ecological Niche model "GARP and Maxent" were applied to predict the potential distribution of Solanum rostratum Dunal and Ipomoea cairica(Linn.)Sweet in China,respectively.The results suggested that the suitable areas of S.rostratum is highly extensive,including most areas of China except Tibet,Qinghai,Hainan province,some part of Xijiang Uigur Autonomous Region,south part of both Guangdong and Guangxi Zhuang Autonomous Region and east part of Taiwan Province. North,East and Center China show high suitability especially.The potential distribution of I.cairica should be constricted within south China,Such as Hanan,Guangdong,Guangxi,Yunnan and middle-south part of Fujian Province.Finally,3 models mentioned above were combined to predicting the potential distribution of Mikania micrantha H.B.K.in China,and predicting results were compared by the method of threshold-dependent(omission rate)and threshold-independent(Receiver Operating Characteristic curve)evaluations.The potential distribution of M.micrantha suggested by 3 models include Yunnan,Guangxi,Guangdong,Hainan,and Taiwan Province,and the results of GARP and Maxent share large similarity,the area predicted by CLIMEX obviously larger than GARP or Maxent.The evaluations indicated that Maxent showed best performance in 3 models,followed by GARP and CLIMEX under current conditions.