Hydromorphological conditions in eighteen restored floodplain channels of a large river: linking patterns to processes

Abstract

Summary Over the past few decades, numerous floodplain restoration projects have attempted to re‐establish complex and diverse river floodplains. They often aim to restore lateral connectivity (i.e. interactions between the main river channel and floodplain channels) and rejuvenate floodplain habitats which are no longer maintained or created by fluvial processes. Nonetheless, results of these experiences in terms of hydromorphological conditions and adjustments are rarely shared. The Rhône River is a large, highly regulated system where lateral connectivity has been greatly reduced. We investigated habitat dynamics (using sedimentological indicators as proxies) in 18 floodplain channels that were restored between 1999 and 2006. Environmental data (bathymetry and grain size of surficial fine sediments) were acquired on 3–5 surveys for each channel covering 6–12 years after restoration. In addition, a pre‐restoration survey was made in 12 of the 18 channels. Using pressure sensors in the floodplain channels and rating curves in the main channel, we quantified the upstream overflow frequency and magnitude (i.e. maximum shear stress) in the channels and tested how these variables explain observed sedimentological patterns. Between‐channel diversity accounted for 81% of the sedimentological variability observed after restoration. Time‐averaged sedimentological conditions were robust and well predicted from overflow frequency and magnitude. Similarly, an indirect index of lateral connectivity used by hydrobiologists was also predictable from overflow frequency and magnitude. The remaining 19% of the sedimentological variability was attributed to temporal variation within channels and was mainly related to changes in longitudinal grain size gradient. This emphasises that grain size patterns are periodically reworked as a result of the flooding regime (backflow versus overflow) without significantly affecting average grain sizes. However, trajectories of grain size changes were stochastic and not always related to the hydrological regime. Accordingly, the partial pre‐restoration data suggest that post‐restoration sedimentological conditions were often similar to those observed before restoration, except in a few channels where major restoration works were performed. Our results quantify how changes in upstream overflow frequency and magnitude can modify physical conditions in the floodplain channels. They can be used to design habitats that are infrequent or missing at the floodplain scale. These results also suggest that changes in upstream plug morphology are a primary habitat driver. Such changes could be more frequently implemented in the Rhône and elsewhere to maximise the diversity of physical conditions in floodplains.