River restoration is prone to failure unless pre-optimized within a mechanistic ecological framework | Insights from a model-based case study.

River restoration with the use of in-stream structures has been widely implemented to maintain/improve physical habitats. However, the response of aquatic biota has often been too weak to justify the high costs of restoration projects. The ecological effectiveness of river restoration has thus been much debated over claims that large-scale environmental drivers often overshadow the potential positive ecological effects of locally placed in-stream structures. In this study, we used a two-dimensional hydrodynamic-habitat model to evaluate the ecological effectiveness of habitat restoration with the use of in-stream structures in various water discharges, ranging from near-dry to environmental flows. The habitat suitability of benthic macroinvertebrates and of three cyprinid fish species was simulated for six restoration schemes and at four discharge scenarios, and was compared with a reference model, without in-stream structures. We found that the ecological response to habitat restoration varied by species and life stages, it strongly depended on the reach-scale flow conditions, it was often negative at near-environmental flows, and when positive, mostly at near-dry flows, it was too low to justify the high costs of river restoration. Flow variation was the major environmental driver that our local habitat restoration schemes attempted -but mostly failed-to fine-tune. We conclude that traditional river restoration, based on trial and error, will likely fail and should be ecologically pre-optimized before field implementation. Widespread use of in-stream structures for ecological restoration is not recommended. However, at near-dry flows, the response of all biotic elements except for macroinvertebrates, was positive. In combination with the small habitat-suitability differences observed among structure types and densities, we suggest that sparse/moderate in-stream structure placement can be used for cost-effective river restoration, but it will only be ecologically effective -thus justifying the high implementation costs-when linked to very specific purposes: (i) to conserve endangered species and (ii) to increase/improve habitat availability/suitability during dry periods, thus proactively preventing/reducing the current and future ecological impacts of climate change.