Temperate forest heterogeneity decreases local and landscape-scale spider diversity through habitat filtering despite increasing species turnover.

Spiders are important arthropod predators in temperate forests. Their diversity depends on structurally heterogeneous habitats offering diverse microhabitats. Yet, modern silviculture has homogenized temperate forest structure at local and landscape scales. The consequences of this homogenization for landscape-level spider diversity, however, remain largely unknown. We sampled spiders using pitfall traps across 234 patches in a large-scale, replicated field experiment at 11 forest sites across Germany. At each site, one treatment district was experimentally heterogenized through canopy gap creation, thinning and deadwood enrichment, and a second homogeneous district remained untreated as a control. We applied a novel meta-analytic framework to compare α-, β- and γ-diversity of spiders between treatment and control districts, standardized for sample coverage along Hill numbers giving increasing weight to abundance and included taxonomic, functional and phylogenetic diversity facets. We also investigated spider community assembly in response to deadwood enrichment, canopy openness and heterogeneous forest structure. Based on 18,540 spider individuals from 206 species, treatment districts exhibited significantly lower γ- and α-diversity across all diversity facets and Hill numbers, particularly when focusing on rare species (q = 0). In contrast, β-diversity increased in treatment districts for phylogenetic and functional diversity across Hill numbers (q = 0, 1, 2). The simultaneous decrease in α- and γ-diversity despite higher β-diversity renders the increase in compositional turnover insufficient to compensate for local diversity losses. Although spiders were more abundant in treatment patches, habitat filtering, rather than niche competition, shaped the community. Our findings corroborate previous results of high spider abundances but lower taxonomic and functional diversity in canopy gaps due to strong habitat filtering effects. However, we demonstrate for the first time that this lower α-diversity is linked to a lower γ-diversity despite increases in β-diversity. Homogenous forests support higher γ-diversity through greater three-dimensional canopy habitat availability. Yet, failure to account for species frequencies using Hill numbers and coverage standardization may result in a substantial underestimation of arboreal spider diversity in pitfall traps. Nonetheless, higher abundances in heterogeneous forests point towards increased prey availability and predator pressure.