Historical changes in overtopping probability of dams in the United States

Abstract

With concerns about aging dams and nonstationary changes in hydrologic extremes (e.g., flooding), questions arise about whether existing dams may be at risk of failure and pose threats to society. Here, we analyzed 33 dams across the United States to investigate temporal trends in dam overtopping probabilities of annual maximum dam water levels. These dams were selected because of the availability of public domain long-term time series of uncontrolled water levels (50 years or longer). We applied updated stationary frequency analyses using generalized extreme value distributions on 30-year rolling periods from 1973 to 2022. The results revealed an overall increasing trend in the number of dams exhibiting critical overtopping probabilities (i.e., low, moderate and high) alongside a decline in the number of non-critical overtopping probabilities (i.e., very low) over time. This approach uncovered overtopping probabilities that traditional analyses based solely on dam water levels could not reveal. We identified six dams having the greatest overtopping probability, with several being located near large population centers, posing potential risks to the downstream communities. All six dams are classified as large and high-hazard potential. This study provides insights into dam management and risk assessment, emphasizing the need for proactive measures to mitigate potential threats.