Modeled ranges of California plant species and scripts to run the models for Spatial phylogenetics of the native California flora (Thornhill et al. BMC Biology)

The species distribution modeling algorithm Maxent was used to model the range of each Californian species, using the cleaned species-level spatial dataset from Baldwin et al. 2017. Models were fit at the species level ; each OTU’s range was then taken to be the union of the ranges of its included species. Models were fit using four predictor variables representing major energy- and water-related variables known to be important for California plant distributions: climatic water deficit, annual precipitation, mean summer maximum temperature (June–August), and mean winter minimum temperature (December– February) as described in Thornhill et al. 2017. In addition to the standard Maxent modeling approach, we also used a second “distance hybrid” approach to model the area for each species that was both climatically suitable and geographically close to observed occurrences. The distance constraint arguably increases realism by accounting for spatial processes such as dispersal limitation and ...

本研究采用物种分布建模算法最大熵模型（Maxent），结合鲍德温等人（Baldwin et al.）2017年发布的经清洗的物种级空间数据集，对加州各物种种群的分布范围开展建模。建模以物种为单位进行拟合；随后将每个操作分类单元（Operational Taxonomic Unit, OTU）的分布范围定义为其包含的所有物种种群分布范围的并集。本次建模共引入4类表征关键能量与水分相关的环境变量，上述因子已被证实对加州植物分布具有重要调控作用，具体包括：气候水分亏缺、年降水量、夏季平均最高气温（6月—8月）以及冬季平均最低气温（12月—2月），相关建模方案参考索恩希尔等人（Thornhill et al.）2017年的研究。除标准最大熵建模流程外，本研究同时采用第二种「距离混合（distance hybrid）」建模方法，对每个物种同时满足气候适宜性与地理邻近已知观测点的分布区域进行建模。该距离约束条件通过考量扩散限制等空间生态学过程，可有效提升模型的生态合理性，以及……