Effects of large river restoration on currently used bioindicators and alternative metrics

Data-rich restoration experiments offer opportunities to test the ability of bioassessment tools, such as those currently used to assess the ecological status' of waterbodies targeted by the European Water Framework Directive, to detect observed ecological changes. Minimum flow increases in four regulated reaches of the French Rhone River modified the invertebrate and fish communities in a predictable way, as detailed in other articles of this Special Issue. We tested the ability of several fish and macroinvertebrate metrics currently used in bioassessment to detect these changes. In addition, we considered changes in metrics that are expected to respond specifically to flow increase. These metrics were related to the habitat requirements of species, the ecological specialisation of communities and the abundance of macroinvertebrate functional groups (seen as surrogates for ecosystem attributes). For invertebrate communities, bioassessment metrics based on richness had equivocal responses to restoration and the Potamon-Type Index demonstrated no or contradictory responses to restoration. The French biotic index was not sensitive to restoration and instead depicted spatial differences in biological quality. For fish communities, the French fish index was marginally sensitive in the reach with the largest minimum flow increase and some of its metrics were sensitive in other reaches. Contrasting with commonly used bioassessment indices and metrics, several metrics related to habitat requirements appropriately indicated the observed changes in community structure. Large flow changes increased the proportion of fish and macroinvertebrate individuals with preferences for midstream habitats, fast currents, deep waters and/or coarse substrates. However, these changes did not translate into the expected increase in ecological specialisation. In addition, functional metrics indicated that restoration led to higher proportions of grazers and higher availability of suspended food for filtering collectors, suggesting a return to the ecological conditions of a large river. The mixed and potentially contradictory responses of the different metrics confirm the difficulty of establishing benchmarks for ecological indicators in large-regulated rivers and the need to design appropriate bioassessment metrics.