Varying congruence among spatial patterns of vascular plants and vertebrates based on habitat groups

: Proxies are adopted to represent biodiversity patterns due to inadequate information for all taxa. Despite the wide use of proxies, their efficacy remains unclear. Previous analyses focused on overall species richness for fewer groups, affecting the generality and depth of inference. Biological taxa often exhibit very different habitat preferences. Habitat groupings may be an appropriate approach to advancing the study of richness patterns. Diverse geographical patterns of species richness and their potential mechanisms were then examined for habitat groups. We used a database of the spatial distribution of 32,824 species of mammals, birds, reptiles, amphibians and plants from 2,376 counties across China, divided the five taxa into 30 habitat groups, calculated Spearman correlations of species richness among taxa and habitat groups, and tested five hypotheses about richness patterns using multivariate models. We identified one major group [i.e., forest- and shrub-dependent (FS) groups], and some minor groups such as grassland-dependent vertebrates and desert-dependent vertebrates. There were mostly high or moderate correlations among FS groups, but mostly low or moderate correlations among other habitat groups. The prominent variables differed among habitat groups of the same taxon, such as birds and reptiles. The sets of predictors were also different within the same habitat, such as forests, grasslands, and deserts. Average correlations among the same habitat groups of vertebrates and among habitat groups of a single taxon were low or moderate, except correlations among FS groups. The sets of prominent variables of species richness differed strongly among habitat groups, although elevation range was the most important variable for most FS groups. The ecological and evolutionary processes that underpin richness patterns might be disparate among different habitat groups. Appropriate groupings based on habitats could reveal important patterns of richness gradients and valuable biodiversity components.