Aquatic macroinvertebrate response along a gradient of lateral connectivity in river floodplain channels

Large river floodplains potentially include the full range of freshwater ecosystems from permanently flowing channels to temporary pools and springs. Attempts to restore such complex systems require tools adapted to assess restoration success. In an analysis of invertebrate assemblages in the Rhoˆne River floodplain (France), taxonomic-based indices (rarefied richness and assemblage composition) were compared with functional metrics using trait-based ratios as surrogates of ecosystem processes. Their ability to respond to a gradient of hydrological connectivity was assessed in 7 cut-off channels. The sampling design included 2 sites/channel (upstream and downstream), 4 randomly chosen sampling points (0.530.5- m quadrats)/site, and 2 sampling seasons (spring and summer). Water physicochemical and habitat variables were recorded when invertebrates were sampled. Environmental variables, including water conductance, [NH3-N], submerged vegetation cover, diversity of sediment grain size, and organic matter content of the sediment, were used to construct a synthetic variable describing the hydrological connectivity of each site with the main river channel. A quadratic regression of rarefied taxonomic richness and the connectivity gradient was not quite significant, but assemblage composition was strongly related to the gradient. Four of 8 trait-based metrics were correlated with the connectivity gradient. Values of metrics that are surrogates for top-down control of assemblage structure and habitat stability (based on functional feeding groups) declined along the gradient from disconnected sites to more connected sites. Values of metrics that are surrogates for voltinism and food supply for water-column-feeding fish increased with connectivity. Top-down control and voltinism surrogates suggested a decline in predator–prey relationships and lower habitat stability, respectively, in the more connected sites. Assemblage composition and some of the trait-based metrics were sensitive to a flood that occurred before one of the sampling dates. Some of the trait-based metrics showed potential for explaining floodplain invertebrate assemblages and for monitoring postrestoration conditions in floodplain water bodies. However, the metrics were developed initially for studies of lotic systems and their use in heterogeneous floodplain water bodies will require further investigation, e.g., delineation of reference conditions for trait-based metrics