基于4种生态位模型的金钱松潜在适生区预测

金钱松（Pseudolarix amabilis）是我国特有孑遗植物，为国家II级保护植物。基于4种生态位模型（GARP、Bioclim、Domain和Maxent）预测金钱松潜在适生区，采用受试者工作特征曲线（Receiver Operating Characteristic，ROC）和Kappa统计量检验模型的预测效果。预测结果表明金钱松在浙江西北部、安徽南部、湖北南部、湖南北部以及江西北部表现为高度适生，并以这些地带为中心向外延伸至北纬24.43°-33.35°和东经106.41°-123.42°之间，4种模型预测结果的受试者工作特征曲线下面积（Area under recriver operating characteristic curve，AUC）平均值均大于0.9，Kappa平均值亦大于0.75，精度较高。通过"刀切法"分析得出年均温是预测金钱松潜在适生区的关键影响因子，可能为当前金钱松分布格局形成的决定因素。模拟金钱松在末次盛冰期和2070年气候条件下的分布，结果表明其分布格局随气候变化由"南扩北缩"变为"南缩北扩"，未来分布面积将大幅减小，气候变化是导致其"南缩北扩"的主要驱动因子。建议在当前金钱松高适分布区域内（江西铜鼓县、湖南张家界和衡阳）建立自然保护区或种子园，并在未来气候条件下高适分布区域内（如安徽北部、河南南部、湖北东南部等地）通过人工引种辅助金钱松的北向迁移。

Predicting Pseudolarix amabilis potential habitat based on four Niche models

Pseudolarix amabilis, which distributes narrowly in China, is a rare and endangered relic plant listed in National Key Protected Wild Plants Grade II list. It is of great significance to the artificial protection and cultivation of P. amabilis to predict the potential suitable habitat of P. amabilis in China using niche models. In this paper, 19 climate factors and 3 topographic factors derived from 44 specimen distribution records from herbaria of China and the WorldClim webset were applied to four Niche models, i.e. GARP, Bioclim, Domain, and Maxent to predict the areas of potential distribution of P. amabilis. The results of different models were analyzed and compared using two statistical criteria: the area under the receiver operating characteristic curve (AUC) and Kappa value. The predictions of the four models were basically identical, showing that the naional areas for potential distribution of P. amabilis were predominantly in southern Jiangsu, southern Anhui, southern Hubei, southern Jiangxi, and northwestern Zhejiang, with the core range situating from 24.43°N to 33.35°N and 106.41°E to 123.42°E. The AUC average values of the four models were all above 0.90 and the Kappa average values were above 0.75, justifying their applications for predicting the potential areas of P. amabilis. Furthermore, our study showed that the annual mean temperature was the key environmental factor affecting the distribution of P. amabilis. It may be the decisive factor of the current distribution pattern of P. amabilis and the basis of the migration direction of P. amabilis in the future. The results in the Last Glacial Maximum (LGM) and 2070 showed that the distribution pattern changed from "South expanding and North shrinking" to "South shrinking and North expanding" due to climate change, and the distribution area would be sharply reduced in the future. We suggested that the nature reserves or seed orchards of P. amabilis should be established in the most suitable distribution areas (the Tonggu County of Jiangxi Province, the Zhangjiajie County and the Hengyang City of Hunan Province). Besides, the northward migration of P. amabilis should be assisted by artificial introduction in the most suitable distribution areas (such as northern Anhui Province, southern Henan Province, southeastern Hubei Province, etc.) under the future climate conditions.