Robustness of Taylor's law under spatial hierarchical groupings of forest tree samples

Abstract

AbstractTesting how well Taylor's law (TL) describes spatial variation of the population density of a species requires grouping sampling areas (patches of habitat) into blocks so that a mean and a variance of the population density can be calculated over the patches in each block. The relationship between specific groupings and TL remains largely unknown. Here, using tree counts from a deciduous forest, we studied the effect of four biological methods of grouping sampling areas into blocks on the form and parameters of TL. Regardless of the method of grouping, the species‐specific basal area densities obeyed TL, and the estimated slopes were not significantly different from one grouping method to another. Surprisingly, TL remained valid when four kinds of randomizations were performed to the biological groupings and tree census. These randomizations randomly assigned sampling areas to blocks, and/or randomized the species composition within or across sampling areas. We found that the form of TL was robust to different grouping methods and species randomizations, but its parameter values depended significantly on species compositions at sampling areas.