基于MaxEnt模型的草原步甲属潜在地理分布研究

步甲(Carabid beetle)是草原生态系统重要的昆虫类群之一，对草原生态系统环境监测及维护生物多样性起着重要的作用。基于步甲属(雕步甲和长叶步甲)在宁夏草原131个有效调查记录和19个环境因子，利用最大熵模型(Maximum Entropy Model,简称MaxEnt)预测步甲属潜在空间地理分布，运用受试者工作特征曲线(ROC曲线)检测预测模型精度，刀切法(Jackknife)筛选影响步甲属潜在地理分布的主要生态因子。结果表明：(1) MaxEnt模型预测结果 ROC曲线AUC值分别为0.981、0.914,模型预测结果的可信度较高；(2) 19个环境因子中，对步甲属的潜在空间分布预测贡献率较高的环境因子为最冷月份最低气温、年均降水量、最干季均温、年平均气温；(3)雕步甲主要分布在温性荒漠草原及温性草甸草原的北部，长叶步甲主要分布在温性草甸草原。以上研究为步甲调查和保护提供参考依据。

Simulating potential distribution of Carabus beetle in the steppe based on MaxEnt model

Carabid beetle is an important taxon in grassland ecosystem, which plays an important role in environmental monitoring and biodiversity conservation in grassland ecosystem. Based on 131 investigation records and 19 environmental factor variables of carabid beetle in the natural steppe, the potentially geographical distribution of carabid beetles was predicted by MaxEnt model. The value of area under the receiving operator curve (AUC) of area under Receiver Operating Characteristic Curve (ROC) as the prediction accuracy standard of model, the jackknife test was used to evaluate the contribution of each environmental variable in deciding potential geographical distribution areas of Carabus glyptopterus and C. vladsimirskyi. The results show that:(1) the prediction accuracy of the MaxEnt model is relatively high, and the work curve area (AUC value) of the subject is 0.981 and 0.915, respectively. (2) Among 19 environmental variables, min temperature of coldest month, annual precipitation, annual temperature and mean temperature of driest quarter show a large amount of contribution to potential distribution of carabid beetles. (3) C. glyptopterus is mainly distributed in temperature desert steppe and the northern part of temperature meadow steppe, thus C. vladsimirskyi is mainly distributed in temperature meadow steppe. In brief, the conclusion provides reference basis for investigation and protection of carabid beetle.