Quantifying landscape change following catastrophic dam failures in Edenville and Sanford, Michigan, USA

Abstract

Dam failures due to changing hydroclimate and ageing infrastructure pose a significant threat to downstream river systems and communities. The detailed geomorphic effects of catastrophic dam failures are not well known because of a lack of high‐resolution topographic data before and after failures. On 19 May 2020, the 17‐m‐tall Edenville and 11‐m‐tall Sanford dams near Midland, Michigan, USA, failed as a result of significant rainfall over the preceding 2 days. We analysed the geomorphic impacts of these failures using a pre‐failure airborne lidar dataset and three uncrewed aerial vehicle (UAV)‐based lidar surveys collected 2 weeks, 3 months and 11 months after failure. Our survey following the dam failure revealed 47 100 ± 11 900 m3 of net floodplain erosion and 17 300 ± 4200 m3 of net deposition downstream of the Edenville and Sanford dam breaches, respectively. Over the year following failure, most geomorphic change was confined to new knickpoints migrating slowly through underlying glacial till substrate, with little change to riverbanks or surrounding floodplains. A lack of impounded reservoir sediment and antecedent downstream topography, including valley width and the location of the breaches relative to the river below the dams, contributed to relatively modest geomorphic changes despite the magnitude of water released. We provide insight into how landscapes are shaped by catastrophic floods, which are likely to become more common with ageing dams and a changing hydroclimate.

Four Years of Meander‐Bend Evolution Captured by Drone‐Based Lidar Reveals Lack of Width Maintenance on the White River, Indiana, USA

“…Short shrubs or dense tall grasses are the greatest impediment to bare-earth elevations, both because fewer bare-earth points are collected and because ground classification is more ambiguous. A full description of our data collection and processing workflow is provided in the Supporting Information, with additional details available in Martin, Edmonds, Yanites, and Niemi (2024).…”