Statistical inference methods for n‐dimensional hypervolumes: Applications to niches and functional diversity
Daniel C. Chen, Alex Laini, Benjamin Blonder
Abstract
Abstract The size and shape of niche spaces or trait spaces are often analysed using hypervolumes estimated from data. The hypervolume R package has previously supported such analyses via descriptive but not inferential statistics. This gap has limited the use of hypothesis testing and confidence intervals when comparing or analysing hypervolumes. We introduce a new version of this R package that provides nonparametric methods for resampling, building confidence intervals and testing hypotheses. These new methods can be used to reduce the bias and variance of analyses, and well as provide statistical significance for hypervolume analysis. We illustrate usage on real datasets for the climate niche of tree species and the functional diversity of penguin species. We analyse the size and overlap of the respective niche or trait spaces. These statistical inference methods improve the interpretability and robustness of hypervolume analyses.