Improving Stream-Crossing Barrier Assessment With Multi-Scale Analyses of Fish Swimming Abilities and Structure Characteristics

Stream habitat fragmentation is driving critical declines in global freshwater fish diversity. A major factor causing this fragmentation is the installation of impassable road-related stream-crossing structures (e.g., culverts). Many fishes require movement within and among discrete habitat types to complete their life histories. These movements can be prevented when stream-crossing structures create conditions through which fish cannot pass. For example, inappropriately installed culverts can increase water flow velocity beyond fish swimming abilities or can create outlet drops beyond fish jumping abilities. Preservation of freshwater fish diversity requires remediation or removal of the most problematic of these barriers to allow natural movement patterns. Whether a structure acts as a barrier depends on the structure characteristics and the abilities of the species that negotiate it. Given the vast multitude of stream-crossing structures placed throughout stream networks, and the high diversity of fish species, our understanding of the characteristics and spatial distributions of passage barriers needs considerable improvement. Directly measuring fish passage through very large candidate sets of potential barriers is infeasible. Therefore, we require simplified models to reasonably predict fish passage probability for given structure characteristics, and the spatial distribution of structures with relevant characteristics. Here, I seek to improve these modeling processes in three broad ways: 1) improving understanding of swimming ability, and the traits that influence it, within and among various important species of southeastern U.S. fish; 2) improving understanding of the relationships between watershed characteristics and culvert severity metrics, and how local relationships between these variables are influenced by large-scale ecoregion context; and 3) demonstrating how species-specific swimming information can be integrated into rapid barrier assessment protocols, and how results of this integration depend both on fish abilities and watershed-specific culvert characteristics. Taken individually or in various combinations, results from this work will be useful to those interested in improving multiple aspects of small, in-stream barrier removal prioritization.