Habitat connectivity drives regional differences in assemblage predictability and beta diversity of stream macroinvertebrate assemblages

Background/Question/Methods

Niche models imply that local taxon probabilities of occurrence are solely determined by habitat suitability and that local assemblage composition is therefore predictable from environmental characteristics. Multi-taxon niche models are frequently used to conduct assessments of local biodiversity loss in in freshwater ecosystems, but the precision of these models varies significantly among regions. Both ecological and statistical mechanisms may influence model precision, but the causes of variation in precision are not yet generally understood. Here, we consider two hypotheses that might explain variation in model precision: (H1) predictability in assemblage composition declines with decreasing habitat connectivity as a function of increasing stochasticity in dispersal, colonization, and establishment of specific taxa, and (H2) models increasingly fail to account for between-site differences in taxonomic composition as within-region environmental heterogeneity increases. We quantified relationships between model precision, habitat connectivity (measured as perennial drainage density), within-region environmental heterogeneity, and regional beta diversity with data from 22 regions within the USA.

Results/Conclusions

Niche model precision was positively associated with habitat connectivity and negatively associated with beta diversity. However, beta diversity was only associated with habitat connectivity (negative association) and not environmental heterogeneity. We interpret these relationships to imply that habitat connectivity influences both regional beta diversity and assemblage predictability via its effect on local community assembly. Regional differences in habitat connectivity may therefore impose fundamental constraints on the precision of niche models and hence influence the assessments of local biodiversity status derived from them.