Simulating drying and human impacts on river networks to evaluate biological quality indices performance through the lens of metacommunity theory

Abstract Incorporating metacommunity perspectives into bioassessment represents a major challenge in the management of drying river networks, where drying-induced fragmentation compromises the performance of biological indices to assess their ecological status. Indeed, current indices focus on local community responses to stressors and neglect the effect of regional processes, such as spatiotemporal connectivity and dispersal, on metacommunity assembly. In this work, we explored the effect of drying on the performance of a widely used biological index using metacommunity simulations on a synthetic drying river network. We assessed how different gradients of drying-driven fragmentation and human impact extent determine local richness and the biological index scores by combining simulations with biomonitoring information. We used a coalescent metacommunity model to simulate the exchange of individuals between local communities along synthetic drying river networks. These networks were subjected to different drying extent, drying intensity and human impact extent scenarios. Additionally, we considered two major characteristics for each simulated taxon: (i) tolerance to human impacts and (ii) dispersal strategy (flying, swimming, or drifting). For each simulation, we obtained local richness and the biological index value. Then, we calculated biological index performance, which we defined as the capacity to distinguish between impacted and non-impacted sites. Finally, we tested our approach in six non-impacted European drying river networks, from which drying information was available. Our results showed that low spatiotemporal connectivity consistently led to decreased local richness and low scores of the biological index (reflecting poor biological quality). As drying extent and intensity increased, drying-induced fragmentation significantly reduced the biological index performance. For example, with a 50% increase in drying extent, index performance fell over 70% and at high drying levels, this performance dropped more than 90%. This decay followed a convex pattern, with a marked drop in performance as soon as drying appeared in the catchment and leveling off at higher drying extents. Synthesis and applications : We show how simulations can be used to incorporate network fragmentation and metacommunity dynamics into the design and validation of biomonitoring tools, thereby increasing their efficiency.