Unintended consequences of restoration: loss of riffles and gravel substrates following weir installation.

We used pre- and post-restoration channel surveys of the Donner und Blitzen River, Oregon, to evaluate the effects of grade-control structures on channel morphology and baseflow habitat conditions for native redband trout and other aquatic biota. Six years after installation, we found that the channel had a smaller proportion of riffles and pools and less gravel substrate, combined with an increase in the proportion of flat waters and consolidated clay on the bed surface. Both local scour downstream from weirs and backwater effects upstream from weirs appear to have caused the general flattening and fining of the channel. A direct-step backwater calculation indicates that backwaters extended to the upstream weir at both low and high flows, creating long sections of flat water separated by short, steep drops. Despite backwater effects, a comparison of longitudinal profiles before and six years after weir installation showed bed erosion downstream of nearly all weirs, likely a consequence of the cohesive clay material that dominates the channel bed and banks. A deep inner channel reflects the cohesive nature of the clay and the mechanisms of abrasion, and indicates that sediment load is low relative to the transport capacity of the flow. Unfortunately, weirs were problematic in this system because of the cohesive clay substrate, limited sediment supply, and low channel gradient. Although deeper flows due to backwaters might be more favorable for resident trout, less gravel and fewer riffles are likely to negatively impact trout spawning habitat, macroinvertebrate communities, and biofilm productivity. Our results demonstrate the potential limitations of a single-feature approach to restoration that may be ineffective for a given geomorphic context and may overlook other aspects of the ecosystem. We highlight the need to incorporate geomorphic characteristics of a system into project design and predictions of system response.