Abstract: Objective To predict the potential distribution of Euglandina rosea in China under current and future climatic conditions via an optimized maximum entropy (MaxEnt) model, and to provide a theoretical basis for the prevention and control of invasion of E. rosea in China. Methods The global occurrence records of E. rosea were collected and screened using ENMtool. Maxent 3.4.1 and GraphPad Prism 8 were used to screen environmental variables according to the contribution rate of environmental variables, the jackknife method, and the correlation analysis of variables. The "kuenm" package was run in R 4.0.4 software to calculate and adjust the model parameters by means of regularization multipliers and feature combinations. The optimized MaxEnt model was used to predict the potential distribution of E. rosea in China under current and future climate scenarios. ArcGIS 10.7 was used to process the results and map the images. Results A total of 780 E. rosea occurrence records were identified, and four environmental variables were used to construct the MaxEnt model. The area under the curve (AUC) of the model was 0.963. The most influential environmental factors on the distribution of E. rosea were precipitation of the driest quarter and the highest temperature in February. The MaxEnt model predicted that the current potential areas of E. rosea were distributed in the southeast of China, concentrated in Fujian province, Guangdong province, and central Guangxi Zhuang Autonomous Region. It would gradually expand northward to Hunan province, Jiangxi province, central Anhui province, eastern Hubei province, and some areas of Zhejiang province in the future. Conclusions E. rosea has potential distribution in China. With global warming, its potential habitat areas will gradually expand, while at lower latitudes in the southeast they will slightly shrink. As a result, the overall center of potential areas shifts northward.

Key words: Euglandina rosea, MaxEnt, Biological invasion, Climate change, Suitable habitat

摘要： 目的 通过优化后的最大熵模型（MaxEnt）预测在当前和未来气候条件下玫瑰蜗牛在我国的潜在分布，为我国防控玫瑰蜗牛的生物入侵提供理论依据。方法 收集玫瑰蜗牛的全球分布数据，使用ENMtool进行筛选；用maxent 3.4.1软件和GraphPad Prism 8根据环境变量贡献率、刀切法以及变量相关性分析对环境变量进行筛选；在R 4.0.4软件中运行“kuenm”程序包计算调整模型参数，包括正则化乘数和特征组合；利用优化MaxEnt模型预测当前和未来不同气候情景条件下玫瑰蜗牛在中国的潜在分布范围；使用ArcGIS 10.7软件进行结果处理及可视化。结果 共筛选出780个玫瑰蜗牛分布点，4个环境变量构建MaxEnt模型，模型的曲线下面积（AUC）值为0.963。对玫瑰蜗牛分布影响最大的环境因子为最干季降水量和2月最高温。MaxEnt模型预测当前玫瑰蜗牛潜在适生区分布在我国东南地区，集中于福建和广东省以及广西壮族自治区中部。未来将逐步向北扩大至湖南省、江西省、安徽省中部、湖北省东部以及浙江省的部分地区。结论 玫瑰蜗牛在我国存在潜在适生区，随着全球变暖，其潜在适生区将逐步扩大，东南部较低纬度适生区则略有收缩，使得玫瑰蜗牛潜在适生区的整体重心向北移动。

关键词: 玫瑰蜗牛, 最大熵模型, 生物入侵, 气候变化, 适生区