COS 74-6 - Cross-scale interactions contextualize the effects of native species richness and disturbance on non-native species richness in stream fish assemblages of the United States

Abstract: Although many established hypotheses exist to explain the drivers of biological invasion, the evidence for universal patterns is inconsistent across study systems and regions. We shift our focus instead to understanding why universal patterns remain elusive. Using a broad extent and fine resolution stream fish community survey dataset compiled from state agencies across the United States, we examined whether the local scale effects of biotic and abiotic drivers of non-native species richness are explained by other related metrics that vary at the regional scale. We used a Bayesian hierarchical model with two levels to examine these cross-scale relationships. The first level modeled the effect of native species richness (biotic) and hydrologic alteration index (abiotic) on non-native species richness after controlling for watershed area and number of sampling events. The second level modeled the effect of regional gamma diversity on the first level effect of native species richness (biotic) and the effect of regional impervious surface cover on the first level effect of hydrologic alteration index (abiotic).

For both biotic and abiotic factors, regional differences in local scale relationships could be explained by regional variation. Native species richness had a positive effect of on non-native species richness in regions with low gamma diversity, but that the relationship dissolved when gamma diversity was high. We also found that hydrologic alteration index, a measure of disturbance, had a stronger positive effect in regions with the greatest amount of impervious surface. Together this suggests that diverse stream fish assemblages are resistant to the establishment of non-native species and that anthropogenic land change can exacerbate disturbance in aquatic systems in ways that promote the establishment of non-native species when they are introduced. As the availability of large, spatially explicit datasets increases, our results suggest that acknowledging and utilizing regionally-based, cross-scale interactions will be important for understanding the mechanisms behind invasion patterns and processes.