Abstract Addressing biodiversity loss requires knowing how different living beings are spatially distributed. For hyper‐diverse groups such as spiders, biogeographic dispersal‐related processes tend to be the main factor driving diversity patterns at large spatial scales, while the relevance of ecological filtering (species sorting) may increase at lower spatial scales. To determine how spider diversity and structure are shaped at local spatial scales, we sampled spider communities using a standardized optimized protocol and assessed habitat heterogeneity across 10 Mediterranean forest plots in the Northeastern Iberian Peninsula. We compared the spiders' composition and structure among forest types and across vegetation layers using generalized linear models and linear mixed models for assessing patterns of alpha‐diversity and constrained ordination analyses for understanding beta‐diversity patterns. Across forests, habitat heterogeneity was the only significant driver for structuring alpha‐ and beta‐diversity, and the geographic and climate distance explained a low variance of the models. Pine forests were separated in the redundancy analyses from holm oak and deciduous forests because a lower vegetation cover characterized the latter. Lineal models of alpha‐diversity at the microhabitat scale only showed significant negative tendencies for the rock cover. Beta‐diversity across vegetation layers within a plot and across forests appears to be influenced primarily by habitat heterogeneity, particularly by the presence of herbaceous and low shrub cover. Our results highlight the importance of species sorting over biogeography and climatic variables in shaping spider diversity patterns at the local scale. They also indicate that a mosaic of habitat structure within and between forest types is a significant driver of spider diversity. These findings have implications for assessing spider assemblages—and likely other arthropods—suggesting that a combination of methods capturing microhabitat descriptors is crucial for effective monitoring schemes.
Marquerie‐Córdoba et al. (Wed,) studied this question.