Reducing the Error Associated with Manning's Roughness in Culvert Design for Improved Fish Passage

There are an estimated 150,000 culverts in the United States that convey water from streams and creeks underneath our vast transportation network. These culverts were originally designed for hydraulic conveyance without consideration given to the varied aquatic ecology that was present, including fish. Each of these culverts represents a potential migration barrier for fish. The accepted design of culverts uses calculations based on Manning's roughness coefficients at full capacity flow. Manning's roughness in pipes has long been known to vary with depth (Chow 1959). There have been a handful of studies that have been conducted on the variation of Manning's roughness coefficient with depth inside pipes (Mills 1924, Camp 1946, Pomeroy 1967, and Fukuchi 2006). A culvert having partially full flow is the most common culvert condition and consequently represents the most likely flow that fish experience during migration. Unfortunately at these low flow depths the results of roughness experiments are contradictory. For instance, at flow depths less than 20% of the pipe diameter, Pomeroy 1967 found roughness coefficients less than full pipe roughness while Camp reports roughness coefficients greater than full pipe roughness. Over much of the flow depth there is 20% to 30% deviation in Manning's roughness which directly relates to differences in predicted velocity. Drag on a fish body is a function of velocity squared while the power required for swimming is a function of velocity to the third power. These relations indicate that the moderate errors in Manning's roughness that have been reported will have large effects on the predicted fish passage potential of a culvert. Further, existing literature uses data obtained from pipes having diameters 12 inches or less made from materials of either vitrified clay, concrete, or galvanized steel. According to the Ohio Department of Transportation, the most common diameters of culverts range from 12 inches to 10 feet and are comprised of corrugated steel, corrugated plastic or concrete. The purpose of this research is to examine pipe roughness in culverts at partially full conditions for the materials and diameters most commonly used in practice today in order to increase the efficacy of fish passage design. The paper will include a literature review and discussion of the previous work on this problem as it pertains to fish passage. Furthermore, in order to fill in gaps and clarify inconsistencies with the data, preliminary results from a culvert roughness study will be included.