Abstract: Objective To analyze and predict the suitable areas for Marmota himalayana in Golog Tibetan Autonomous Prefecture (Golog Prefecture), China using ecological niche models, so as to provide a theoretical basis for scientific and efficient plague host surveillance in the future. Methods According to field survey data on the longitude and latitude of 96 M. himalayana distribution points as well as data on elevation and climate (19 factors) from the WorldClim website, the suitable areas of M. himalayana in Golog Prefecture were modeled using a maximum entropy model (MaxEnt) and ArcGIS 10.6 software, and the results were evaluated through the receivers operating characteristic curve (ROC). Results The area under the curve (AUC) of the MaxEnt was 0.928, indicating that the prediction results of the model were reliable. The percentage contribution rate of the top five factors in descending order were precipitation of the driest month(35.7%), temperature seasonality(26.3%), precipitation seasonality(20.5%), monthly average diurnal temperature range(10.2%), and elevation(7.3%). When the monthly average diurnal temperature range was 14.7 ℃, the temperature seasonality was 800, the precipitation of the driest month was 2.4 mm, the precipitation seasonality was 95, and the elevation was 3 720 m, their respective response curves reached the highest. The highly suitable area for M. himalayana in Golog Prefecture was estimated at 2 445.24 km², and the moderately suitable area was estimated at 4 240.76 km², which were mainly located in Maqen County and Gade County. Conclusions The MaxEnt model can produce reliable prediction results that are consistent with the actual situation, providing a scientific and efficient theoretical basis for plague surveillance in the future. The highly and moderately suitable habitats of M. himalayana are largely located in Maqen and Gade counties, where surveillance should be strengthened.

Key words: Plague, Marmota himalayana, Suitable area, Maximum entropy model, Golog Tibetan Autonomous Prefecture

摘要： 目的 基于生态位模型分析并预测果洛藏族自治州（果洛州）喜马拉雅旱獭适生区，为今后科学高效的鼠疫宿主监测工作提供理论依据。方法 根据现场调查得到的96个喜马拉雅旱獭分布点经纬度及WorldClim网站获取的海拔及 19个气候变量，通过最大熵模型（MaxEnt）和 ArcGIS 10.6软件进行建模得出喜马拉雅旱獭在果洛州的适生区，并通过受试者工作特征曲线对模型预测精度进行评价。结果 根据 MaxEnt 得出该模型的曲线下面积（AUC）值=0.928。排名前 5 位的因素贡献度从高到低依次为最干燥月份降水量（35.7%）、温度季节性变化（26.3%）、季节性降水（20.5%）、昼夜温差月均值（10.2%）、海拔（7.3%）。各因素在模型中的响应曲线分析得出，在昼夜温差月均值为14.7 ℃、温度季节性变化为800、最干燥月份降水量为2.4 mm、季节性降水为95、海拔为3 720 m时各因素的响应曲线达到最高值。预测结果显示，果洛州喜马拉雅旱獭高适生区面积达2 445.24 km²，中适生区面积达4 240.76 km²，主要位于玛沁县、甘德县境内。结论 最大熵模型预测结果可靠，与实际情况相符，为今后鼠疫监测工作提供了科学高效的理论依据；玛沁、甘德县喜马拉雅旱獭高、中适生区分布区域占比较高，应加强对该区域的监测。

关键词: 鼠疫, 喜马拉雅旱獭, 适生区, 最大熵模型, 果洛藏族自治州