利用最大熵模型预测内蒙古高原长爪沙鼠鼠疫疫源地动物间疫情发生的风险

The potential risks of animal plague in natural foci of Meriones unguiculatus in the Inner Mongolia plateau predicted by Maximum Entropy model

目的研究气候环境因素与长爪沙鼠鼠疫疫源地动物间疫情的相关性,预测长爪沙鼠鼠疫疫源地动物间疫情风险分布,为动物间鼠疫防控提供科学依据。方法2005-2018年,内蒙古高原长爪沙鼠鼠疫疫源地30个疫源县的鼠疫细菌培养阳性数据,包括检菌时间、检菌数量、经纬度或详细地点、宿主种类,分别来自于中国疾病预防控制系统鼠疫防治管理信息系统以及相关鼠疫防治专业机构。采用logistic回归分析长爪沙鼠鼠疫疫源地动物间疫情与气候环境相关的危险因素,利用生态位模型的最大熵(Maxent)模型预测长爪沙鼠鼠疫疫源地动物间疫情的潜在分布,利用受试者工作特征曲线(ROC曲线)验证模型。结果经筛选,在气候环境相关因素中,有11个气候因子(年均温、昼夜温差和年温差比值、温度季节性变化、最干季平均温度、最热季平均温度、最冷季平均温度、年降水量、降水量的季节变异系数、土地覆盖物、归一化植被指数、坡向)与长爪沙鼠鼠疫动物间疫情发生相关(OR=1.302、0.455、0.957、0.930、4.864、0.179、0.986、1.126、0.992、0.981、0.721,P均<0.01),均被纳入模型中。其中年均温、最热季平均温度和降水量的季节变异系数升高会增加长爪沙鼠鼠疫疫源地动物鼠疫发生的风险;昼夜温差和年温差比值、温度季节性变化、最干季平均温度、最冷季平均温度、年降水量、土地覆盖物、归一化植被指数和坡向升高会降低长爪沙鼠鼠疫疫源地动物鼠疫发生的风险。验证预测模型显示,训练集和测试集的ROC曲线下面积(AUC)值分别为0.988和0.985,模型预测效果较好。长爪沙鼠鼠疫疫源地动物间疫情高风险区域主要集中于乌兰察布高原中北部、鄂尔多斯高原、河套平原东部。结论利用最大熵模型和气候环境数据预测长爪沙鼠鼠疫疫源地动物间疫情潜在风险空间分布,结果准确可靠。

Objective To forecast the risk distribution of inter-animal plague in Meriones unguiculatus effectively and provide scientific evidence for prevention and control of inter-animal plague,through studying the correlation between meteorological and environmental factors and inter-animal plague in Meriones unguiculatus.Methods Positive data of plague bacterial culture in 30 epidemic source areas of Meriones unguiculatus in the Inner Mongolia plateau from 2005 to 2018,including detecting time,number of bacteria,latitude and longitude or detailed location,host type,were from the Chinese Disease Prevention and Control System Plague Prevention Management Information System and related professional institutions for plague prevention and treatment.Logistic regression was used to explore the relationship between the inter-animal plague and climate-related risk factors.The Maximum Entropy(Maxent)model was used to predict the habitat distributions of inter-animal plague,and the receiver operating characteristic(ROC)curve was used to validate the model.Results There were 11 climatic factors including annual mean temperature,isothermality,temperature seasonality,mean temperature of driest quarter,mean temperature of warmest quarter,mean temperature of coldest quarter,annual precipitation,precipitation seasonality,globcover,normalized difference vegetation index and slope,were related to the outbreak of plague among Meriones unguiculatus and included in the model(OR=1.302,0.455,0.957,0.930,4.864,0.179,0.986,1.126,0.992,0.981,0.721,P<0.01).The increase of annual mean temperature,mean temperature of warmest quarter and precipitation seasonality will increase the risk of animal plague in the plague foci of Meriones unguiculatus;the increase of isothermality,temperature seasonality,mean temperature of driest quarter,mean temperature of coldest quarter,annual precipitation,globcover,normalized difference vegetation index,and slope will reduce the risk of animal plague in the plague foci of Meriones unguiculatus.The areas under the curve(AUCs)of the Maxent model training data and test data were 0.988 and 0.985,the prediction effect of the model was better.The habitat distribution of Meriones unguiculatus plague mainly concentrated in the central and northern Ulanqab plateau,Ordos plateau,and eastern Hetao plain.Conclusions The use of Maxent model and climate data can predict the potential risks and spatial distribution of animal plague in Meriones unguiculatus;the results are accurate and reliable.