Studies On Suitable Geographie Distribution Of Xanthium Italicum Moretti In Xinjiang

The biological invasion altered the biodiversity and the ecological balance of original ecological system.Therefore,in order to diminish the negative effect of invasive species to ecosystems and social economics,and it is necessary to understand the distribution of invasive species at different spatial scale.The potential distribution of invasive species predicted by its distribution model provides fundamental data for the eradication of invasive species.Although the potential distribution predicted by such method is accepted,the distribution with higher reliability is still possible if the human activity can be included as one of the predictors.The proposed research will firstly construct the spatial layer of the human activity index,and then integrate the layer to the data set of other environmental variables to predict the potential distribution of Xanthium italicum Moretti use ensemble modeling methods,The results showed that:(1)In order to contrast the prominent effect for more apparent,we first adopt the single model——MAXENT.Although the AUC value i1 s 0.989,but we can only according to the distribution of its MAXENT model which generated the distribution map to reflect the distribution probability of stalk,following the unified rules of the division,it can be divided into the parts as follows: the most suitable grass growing region(OV? 0.9),secondly suitable for growing region(0.7?OV<0.9),thirdly suitable for growing region(0.5?OV<0.7),and unsuitable for growing region(2)However,when we apply the integrated model method to predict it,we choose data according to sample model type,and the species and environmental data into below three models,those used for comparison and select the most suitable model.(3)AUC value is 0.9797,BIOCLIM model accurately predict areas are the north hill of Altay prefecture,southwestern plains of Tacheng prefecture,Bortala,Karamay,Urho,Huoshao shan,northern plains of Ili prefecture,etc.AUC value displayed as 0.8723 of DOMAIN model,accurate prediction areas for the Eerqisi river,Barkol county,northern Jungar basin,plains of western Jungar region,Bortala prefecture,the plains of Ili prefecture,and upstream mountains of Kizil river.One-Class SVM AUC value is 0.8896,accurately predict areas for the Ertis river,Barkol county,northern Junggar basin,western Jungar region,Bortala prefecture,the plains of Ili prefecture,and upstream mountains of Kizil river(4)We choose the highest AUC value and its forecast with the actual distribution area which close to BIOCLIM model,to integrate the human activity in intensity index and other environment variables set,again to the model test,the result showed that the AUC value is 0.9882.Predicting the Ertis river,north hill area in northern mountain,called Ruth too grasslands,western mountainous area of Jungar,Bortala,Karamy ur grain,fire mountains,colorful bay,northern plains Ili region.This study uses a single model to predict MAXENT for the first time,obtained Xanthium italicum Moretti's four suitable range for the Ebinur lake basin.Model results are ambiguous.And when we put the data of species and environment variables into a comprehensive model of three kinds of samples in the model,we got the three kinds of prediction results,from which we choose AUC value and highest predict BIOCLIM model of the region and the actually is most conforms to the human activity intensity index integrated into other environment variables set,again to the model test,results show that the AUC value is improved obviously the forecasting area are more precise and more actual.