Over the past several decades, billions of dollars have been spent trying to restore degraded streams and rivers in North America. Recent critiques of these restoration efforts have argued that much funding has been wasted on projects that have failed to restore the structure and function of lotic ecosystems. This has triggered a cascade of interest in detailing the limitations to successful restoration. Here we present the results of a study designed to explore factors that might limit the recovery of fish populations in a restored reach of the Merced River in the central valley of California. In 2001, California led a multi-agency effort to restore a 2.7-km reach of river habitat that had been degraded by gold mining operations, damming, and sedimentation from agricultural inputs. The stream was restored by engineering the channel to have substrate sizes considered optimal for salmon spawning, which were sufficiently sorted to allow the bed to periodically move during floods, thus reducing impacts of sedimentation. Counter to goals of the original project, surveys have shown that population sizes of numerous fish species have been reduced by 10 to 90% in the restored channel (relative to an upstream, reference site). We hypothesized that reduced fish population sizes might be related to well-documented changes in abundance and composition of invertebrate food items, which have been altered by reductions in heterogeneity of substrates and increases in mobility. To address this hypothesis, we sampled two representative species of fish with contrasting life histories. Pike minnow (Ptychocheilus grandis) live and feed in the water column, whereas Prickly sculpin (Cottus asper) live and feed in benthic habitats.

Results/Conclusions

After analyzing the gut contents of 154 pike minnow, and 78 sculpin, we detected shifts in the species composition of pike minnow diets whereas sculpin diets were similar among reaches. Despite differences in composition, we were unable to find any clear shifts in the volume of material consumed (measured as total carbon) for either species. Length-weight ratios also suggest there is little difference in the condition of fish between the restored and reference reaches. When taken alongside our other studies performed in this site, our study suggests that reduced fish populations in the restored river are not driven by food-web shifts that alter the availability or species composition of the prey assemblage. Fish populations may instead be limited by factors representing physical habitat quality, such as the availability of low flow rearing habitat.