Blonder Benjamin
Am Nat. 2016 Apr;187(4):E93-105. doi: 10.1086/685444. Epub 2016 Feb 15.
Hypervolumes are used widely to conceptualize niches and trait distributions for both species and communities. Some hypervolumes are expected to be convex, with boundaries defined by only upper and lower limits (e.g., fundamental niches), while others are expected to be maximal, with boundaries defined by the limits of available space (e.g., potential niches). However, observed hypervolumes (e.g., realized niches) could also have holes, defined as unoccupied hyperspace representing deviations from these expectations that may indicate unconsidered ecological or evolutionary processes. Detecting holes in more than two dimensions has to date not been possible. I develop a mathematical approach, implemented in the hypervolume R package, to infer holes in large and high-dimensional data sets. As a demonstration analysis, I assess evidence for vacant niches in a Galapagos finch community on Isabela Island. These mathematical concepts and software tools for detecting holes provide approaches for addressing contemporary research questions across ecology and evolutionary biology.
超体积被广泛用于概念化物种和群落的生态位及性状分布。一些超体积预计是凸形的，其边界仅由上限和下限定义（例如，基础生态位），而其他超体积预计是最大的，其边界由可用空间的限制定义（例如，潜在生态位）。然而，观察到的超体积（例如，实际生态位）也可能有漏洞，这些漏洞被定义为未被占据的超空间，代表了与这些预期的偏差，这可能表明存在未被考虑的生态或进化过程。迄今为止，在超过两个维度中检测漏洞是不可能的。我开发了一种数学方法，并在超体积R包中实现，用于推断大型高维数据集中的漏洞。作为一个示范分析，我评估了伊莎贝拉岛加拉帕戈斯雀类群落中空生态位的证据。这些用于检测漏洞的数学概念和软件工具为解决生态学和进化生物学中的当代研究问题提供了方法。
