Predictions Of Suitable Habitat For Wild Asian Elephant In China Based On Maximum Entropy Model(MaxEnt)in Climatic Background

Asian elephant(Elephas maximus)belongs to Mammalia,Proboscidea,Elephantidae,Elephas,which is the largest wild terrestrial animal in Asia.As IUCN Endangered species and species in Appendix I of CITES,the Asian elephant is listed as a National Class I key protected wild animal in China,which has extremely significant research,protection and culture value.To understanding the distribution of species is one of the basis of formulating strategies of biodiversity conservation,resource development and utilization.The Species Distribution Models(SDMs)is a method to analyze the spatial distribution and predict its potential distribution according to the known species distribution data and the corresponding data onto various environmental factors.The wild Asian elephant's distribution site in China was collected by the fieldwork(2013-2018),Investigation of wild Asian elephant distribution and activity data(2017-2018)and Wildlife Public Liability Insurance data of Yunnan province(2013-2017).Combining the 3S technology platform and MaxEnt model to predict the climate data as the condition of history and future of suitable habitat range and made a comparison of the same climate condition with the Domain model and Bioclim model.The main results of this research are as follows:(1)The original parameters of the MaxEnt model were used to carry out a retrospective study on the historical distribution of Asian elephants in China,and 23 variables contained in the 2.5 min historical climate factor data of Diva-GIS were used for simulation to select the better variables.Isothermality,Temperature Seasonality,Precipitation of Wettest Quarter,Annual Precipitation,Precipitation of Wettest Month,Mean Temperature of Driest Quarter,Mean Temperature of Coldest Quarter,Mean Monthly Temperature Range,Max_temp,Annual Mean Temperature,Precipitation of Wettest Quarter,Precipitation of Driest Quarter stand out,AUC values are 0.931,0.922,0.874,0.854,0.854,0.847,0.838,0.837,0.828,0.820,0.8739 and 0.811respectively,all of them are greater than 0.8,the result based on 23 climatic factors showed that the accuracy of the model was excellent.The result of historical average distribution of wild Asian elephant in China from 1968 to 2018 shows that in Xishuangbanna,Pu'er(Simao District,Lancang County,Jiangcheng County,Ning'er County,Jinggu County and Mojiang County),Lincang(Shuangjiang County,Gengma County and Yongde County)were likely as Asian elephants distribution,but the fieldwork survey shows that MoJiang,Shuangjiang,Gengma and Yongde had no Asian elephants,but Cangyuan County,The Nangunhe National Reserve has 13 individuals,theoretical traceability results differ from the actual distribution.(2)The or:iginal parameters of the MaxEnt model were used to carry out a retrospective study on the historical distribution of Asian elephants in China based on the former 7 kinds of better climatic variables factors screening:Isothermality,Temperature Seasonality,Precipitation of Wettest Quarter,Annual Precipitation,Precipitation of Wettest Month,Precipitation of Wettest Month,Precipitation of Wettest Month.The result showed Cangyuan County was one of the possible distribution areas of Asian elephants,and the investigation confirmed this historical trace.(3)The original parameters of the MaxEnt model were used to predict the feature distribution of Asian elephants in China,and 23 variables contained in the 2.5 min future climate factor data of diva-gis were used for simulation to select the better variables:Temperature Seasonality,Isothermality,Mean Temperature of Coldest Quarter,Precipitation of Wettest Month,Mean Temperature of Driest Quarter,Max_temp,Precipitation of Wettest Quarter,Precipitation Seasonality,Precipitation of Driest Month,Annual Mean Temperature,Annual Precipitation and Precipitation of Driest Month stand out.AUC values are 0.912.0.905,0.842,0.837,0.837.0.830.0.830,0.829,0.828,0.823,0.823 and 0.820 respectively,all of them are greater than 0.8.Meanwhile,AUC values of Temperature Seasonality and Isothermality both are greater than 0.9,which means those two factors in the feature distribution are key elements.The result based on 23 climatic factors showed that the accuracy of the model was excellent and credible.Lancang County,Mojiang County and Ximeng County are likely to be suitable habitats,and based on the fieldwork about the trace of wild elephant in Chian from 2013 to 2018,the result proved to be correct.Jiujing,Nuozhadu and Huiming in Lancang had recorded as Actual distribution areas for wild Asian elephants.(4)The original parameters of the MaxEnt model were used to predict the feature distribution of Asian elephants in China,based on the former 8 kinds of better climatic variables factors screening:Temperature Seasonality,Isothermality,Mean Temperature of Coldest Quarter,Precipitation of Wettest Month,Mean Temperature of Driest Quarter,Max_temp,Precipitation of Wettest Quarter and Precipitation of Wettest Quarter.The result showed that the prediction ability of the 8 climate factors after optimization was close to the ability of 23 climate factors before optimization.Through comparison of the actual distribution points,it is concluded that the effect of the optimized climate factors is better than that of the unoptimized ones.The prediction results show that Linyi City,Yongde County,Shuangjiang County and Huoma County in Linyi area have the potential to become potential distribution areas of Asian elephants;Mojiang County,Ximeng County,Menglian County and Zhenlai County of Pu'er City will become a distribution area of Asian elephants theoretically,and Ximeng County has clearly recorded the distribution of Asian elephants in history.(5)Under the same experimental conditions,the MaxEnt model was simulated with BC=15000 and RM=2 as the minimum model parameter to get the optimal result,followed by the Bioclim model and the Domain model with the worst simulation prediction result.The MaxEnt model has a high coincidence degree with the actual active area,and the result may be greatly affected by the density of the active record points(more than 80).In the results of the BioClim model,the most suitable distribution area is generally close to the actual distribution area,and the most suitable distribution area is also close to the distribution area of Asian elephants reported to relevant historical literature.(6)Through the simulation of the historical and future potential distribution of wild Asian elephants in China.In the MaxEnt model simulation results,except for Cangyuan County,there was no occurrence and distribution in Mojiang County(Mojiang County is adjacent to Ning'er County,and in 2018 there was a male adult individual crossing two counties).The stimulation effect of the model is in line with the actual distribution,and the BioClim model and the Domain model differ greatly from the actual distribution.Combined with field investigation and analysis,it is believed that the Asian elephant population first shrinks sharply,and then the cause of explosive spread is related to rubber planting and gun management.