Research On The Potential Distribution Of Solenopsis Invicta In China

PurposeIn this research, we applied three modeling methods to strive for accurately forecasting thepotential distribution of invasive species-Solenopsis invicta in China. Furthermore, we also researchedthe relationship between potential geographic distribution of S. invicta and environmental factors indetail to understanding how environment factors restricted distribution of Solenopsis invicta innon-suitable areas of Chinese mainland. The achievement of research provides a theoretical basis, whichcould guild the next step for the prevention and control stratagem ofS. invicta, even the implementationof eradication efforts.Through comparisons among the results predicted bye the three models, we analyzed strengths andweaknesses of three different modeling methods and then found the suitable prediction method forinvasive pest risk analysis. Finally, we completed a set of distribution prediction software as a platform,which was provided with functions such as potential distribution risk analysis, species richness analysis,constraints analysis, comparative analysis of environmental factors. At the same time, we suggested thatthis platform would achieve a breakthrough in the field of domestic distribution prediction software.MethodsWe applied biological climate model (CLIMEX) S. invicta population dynanaics model (ColonyDynamic Model), nonparametric regression model (Generialized Additive Model GAM), three differentmethods of analyzing and forecasting, to explore ways forecast biogeographical distribution correctly.1) After setting biological parameters of S. invicta, the use of CLIMEX software provides anEco-climatic index (El) distribution map as an indicator of potential distribution risk of S. invicta inChina. we also built distribution maps of the growth index (GI), temperature index (TI), soil moistureindex (MI), cold stress index (CS), hot stress index (HS) and ecoclimatic index (IR-EI) under irrigationconditions to further define effects of different climatic conditions on S. invicta colonization.2) We employed population dynamic model of S. invicta with the effective soil temperature dataranging from 5-14 years of 533 meteorological sites and air temperature data of 704 sites in China.Subsequently, we reached an average female alates production every year as an indicator of potentialdistribution possibilities. Whereafter we compared distribution maps of key climatic factors with mapof colony dynamic model result to understand difference effects of climatic factors which limitedalates production in colony dynamic model.3) After collection actual occurrence in United States from the relevant documents and combinedforecasting result of colony dynamics model, we built estimated distribution data of S. invicta in theUnited States. Afterwards we run this occurrence data and other environment factors such as biologicalclimate data (BioClimatic Data), normalized differential vegetation index (NDVI), digital elevation model data (DEM), derivative data (slope, slope) of DEM and landuse types in the mainland of UnitedStates, on the analyssi system for GAM modeling. Through GAM modelling, the statistical relationshipbetween S. invicta colonization and distribution of environmental factors was researched. Then weinputted the same series of data in the area of Chinese mainland into GAM model and predicted thepotential geographic distribution of S. invicta in China.Results1) Build an analysis system of species distributionThis study has achieved a system with basic functions of operation, management on spatial data,which provides relationships modeling between pest distribution and environmental factors inlarge-scale analysis. The basic function of entire process of modelling can receive supports in thissoftware system. Main modules divided into five parts: basic function modules of geographicinformation systems, data analysis module, modeling module, R platform management module, modelmanagement module, help system module.2) CLIMEX predicted resultsThere are high probabilities invaded by S. invicta throughout Guangdong, Guangxi, Hainan,Taiwan, Fujian, Zhejiang and southern Yunnan, northwest part of Guizhou, the western region ofChongqing, southeast Sichuan, northern and southern Hunan, and the western region of Jiangxi fall intohigh-risk situation. In those areas, El values are more than 30, which mean these areas are very suitablefor invasion and colonization of S. invicta. Thus outward expansion of those areas above, let's say, theeastern part of Yunnan, central Guizhou, central Hunan, western Jiangxi, southern Anhui, southernJiangsu and southern Hubei also holds possibility of a high-risk, which EI value between 20-30. Thesefurther north areas, the northwest of Yunnan, from Hengduan mountain to plateau of western Sichuan,southern Shaanxi, northern Hubei, southern Henan, northern Anhui and Jiangsu, may suffer invasion,but by the more restrictive natural conditions, colonization and expanding populations would meetgreater difficulties in those region with EI value between 5-20. The vast northern region to probablearea is not suitable for colonization of S. invicta.3) Colony dynamics model prediction resultsS. invicta colonization is very likely in Guangdong, Guangxi, Hainan and Taiwan, throughout thefour provinces, Fujian, Hunan, Jiangxi and southern Yunnan, where annual reproduction of femalealates is 6000 above. In addition, eastern Sichuan, central Hubei, central and northern parts of Anhuiand southern Jiangsu also have a high colonization possibility. It should be noted that, even in southernareas to the isoline, which means average female alates production is equal to 2100, there are still existunsuitable areas where the elevations are more than 1000m and female alates productions are near to 0.Such areas as Zhaotong in the northern region of Yunnan-Guizhou Plateau, the Jinfoshan Mountain inthe junction of Guizhou and Chongqing, Lvcongpo in southwest Hubei, Hengshan Mountain,Xuefengshan Mountain, Lushan Mountain, Huangshan Mountain, Wuyishart Mountain, Guacangshan Mountain, Jiuxianshan Mountain and areas surrounding those mountains, remains unfit for colonizationof S. invicta. This result demonstrates that the high-altitude area of low-temperature conditionssignificantly limited colonization of S. invicta.4) GAM model predictive resultsBased the potential distribution of S. invicta predicted by GAM model, the northern limit linegenerally lie in the northeastern Tibet, northwest Sichuan, Gannan, Guanzhong area in southern Shanxi,the northern part of Henan and northern Shandong. Near the contact line, sites that were predicted noprobability for colonization (0) and sites for the colonization (1) intersect. Comparative analysis ofresults in China between colony dynamics model and GAM model shows a big similarity. Meanwhile,the results by GAM model of two sites in Leizhou Peninsula are inconsistent with the facts. In Tibetthere are 10 sites in where the outcome is in internal Tibet, is forecasted to be colonized but not in CDMmodel,.Conclusions1) Based the secondary development on Esri ArcGIS and R DCOM Server, we have built a set ofsoftware what contains functions such as spatial data processing, data exploration, model selection,model performance evaluation, and other functions of the biogeographical distribution analysis platform.The platform can be applied to the distribution analysis of harmful organislns, analysis of relationshipbetween species population dynamics and the environment factors, alien pest potential distributionanalysis, and so on.2) In the determination northern limit of S. invicta, CLIMEX model and the colony dynamicsmodel are quite similar. The limits lay generally in the northwest of Sichuan, southern Gansu,Guanzhong area, southern Shanxi, the northern part of Henan and southern Shandong. It is impossiblefor S. invicta to survival under the conditions of nature where are more northern than the limit line. onthe contrary, S. invicta can colonize in those area where lay in the south to the limit line. In addition,even a factor plays a key role in the survival and development of species, explaining thebiogeographieat distribution will be very difficult in the use of the single environmental factor. A widerange of environmental factors decides the existence of species.3) Through the process in statistical assessment of modeling parameters and the completed modeltest, the accuracy and stability of GAM model can be ensured. In dealing with theexistence/non-existence (Presence/Absence) types of data, nonlinear regression model, GAM can beapplied to the invasion analysis of harmful organisms (and potentially invasive species) (potential)distribution, and still maintain a relatively high accuracy in spite of space transplantation error.