Dams and road-crossings can prevent migratory fish from moving upstream into rivers and streams to breed. The Great Lakes Basin has over 100,000 of these barriers that block access to spawning grounds for fishes such as sturgeon, suckers, and trout. Our team has developed models to select the optimal set of barrier removals to maximize fish habitat given a specified level of investment. However, most stakeholders interested in barrier removal operate at smaller scales and with less data than was used in our analysis. In this study, our objective was to assess patterns in the location and traits of the highest priority barriers targeted for removal relative to the entire set of candidate barriers. We also wanted to evaluate how including watershed stress level, a surrogate for the kind of habitat quality data that is commonly available to local stakeholders, affects the optimal suite of barriers recommended for removal.
Across all budget levels, barriers recommended for removal tend to be in watersheds with few people where large stretches of habitat can be opened by fixing just one barrier. The optimization model avoided areas with dense road networks where multiple barriers in close proximity prevent large areas of habitat from being opened efficiently. Barriers recommended for removal at the lowest budgets are dams that are impassable to fish. Return on investment (ROI) for these selected barriers is higher than the averages for both the entire database and the set of barriers selected at larger budgets, indicating that spatially isolated dams are strong conservation targets. These high ROI barriers are mainly on the north shores of Lake Superior and Lake Huron, and sometimes included multiple barriers within the same watershed. Preliminary analysis suggests that accounting for watershed stress has only modest effects on the optimal portfolio of removals at large spatial scales.