Observations reveal changing coastal storms around the United States

Understanding extreme storm surge events that threaten low-lying coastal communities is key to effective coastal flood mitigation and adaptation measures. However, estimates of extremes storm surge events from observations are sparse and highly uncertain along most coastal regions owing to data limitations. There is also lack of observational evidence about their long-term (underlying) trends and how such trends contribute to overall extreme sea-level changes. Here, we analyze the U.S. tide-gauge observational network using a spatio-temporal Bayesian hierarchical framework, which provides robust empirical spatially-continuous estimates of extreme storm surge likelihoods and their underlying long-term trend since 1950 along U.S. coastlines. We find that observational estimates have underestimated likelihoods of storm surge extremes across over 80% of tide gauge locations nationwide. Additionally, and contrary to prevailing beliefs, storm surge extremes show significant trends within widespread regions along the U.S. coastlines, providing new evidence of changing coastal storm intensity during the historical monitoring period. Regional hotspots exist where storm surge trends are comparable to, or even exceed, trends in mean sea-level rise and its major individual components. These findings advance traditional coastal design/planning practices that rely on estimates obtained from discrete observations and assume stationarity in storm surge extremes.