Culverts can increase stream velocities as a result of reduced waterway areas and prevent upstream passage of small non-salmonid fish such as the Native Utah Leatherside chub (Gila copei) and Longnose dace (Rhinichthyscataractae). To mitigate this problem, current culvert design standards for fish passage match sustained fish swim speeds with average cross sectional velocity through the culvert. Such policies dictate relatively large barrels and do not recognize the role of reduced velocity zones near culvert boundaries. Obstacles and streambed substrate create turbulent regions with lower velocity zones that can increase upstream fish passage. A comparison of upstream passage success using native Utah fish in an experimental flume was conducted with three different conditions: (1) a smooth boundary, (2) a smooth boundary with strategically placed cylinders, and (3) a boundary consisting of natural substrate. The refuge provided by the cylinders and substrate allowed fish to expend less energy as they swam upstream. Energy expenditure was compared between the conditions by mapping the velocity field near the boundary and tracing fish swim paths. Substrate provided sufficient refuge for the fish to behave in a manner similar to their behavior in a natural environment and with significantly reduced energy expenditure. Cylinders provided limited refuge that allowed fish to rest periodically as they navigated the flume. The smooth boundary case required the highest energy expenditure as there was no refuge provided. Fish swimming capabilities in the form of prolonged and burst velocities have been recorded for most species. Streamwise velocity near the boundary can be compared to the prolonged and burst swim speeds to predict passage rates. Further field testing is necessary to fully substantiate the effectiveness of utilizing reduced velocity zones in non-salmonid fish passage prediction. If such a design approach can be used instead of using the conservative but overly simplistic average velocity to evaluate the retrofit of existing culverts and to design new culverts it will help minimize costs and result in fewer culvert replacements and smaller and simpler new designs. Other implications such as downstream effects on stream bed stability and scour remain an issue.

Native fishes have become an increasingly important concern when designing fish passable culverts. Many operational culverts constrict waterways which increase velocities and prevent upstream passage of small fish species. The current method to ensure fish passage is to match the average cross sectional velocity to the sustained swim speed of the fish. This study investigates the passage rates of leatherside chub (Lepidomeda aliciae) and speckled dace (Rhinichthys osculus) at three sites (an arch culvert with substrate bottom, box culvert with bare bottom, and a stream section with no culvert) located on Salina Creek near Salina, UT. It was found that fish were able to pass through all of the sites. However, fish were able to take advantage of the habitat within the culvert that had a substrate bottom more effectively than within the culvert that had no substrate within the barrel. This was reflected in population density estimates at each of the three test sites for each species. It was also found that the substrate at the arch culvert and stream sites scaled with the fish measured in this study. The D50 and D84 were 44 and 205 mm at the arch culvert site and 26 and 126 mm at the stream site. The average fish length was 76 mm for the chub and 64 mm for the dace. It is recommended that (1) a culvert size that produces a velocity equal to the prolonged swim speed of target fish in the near boundary region (2 cm above the bed) be used in the future, and (2) substrate that scales with the target fish species be placed in the culvert barrel.