An analysis of tropical storm surge trends for the Atlantic coast of the United States

Abstract

Tropical cyclone generate storm surge is responsible for damage to lives and livelihoods on a global scale. In spite of this inherent danger, the scientific community currently lacks a global database of peak storm surge events. Thus, SURGEDAT was created to fill this void. Research began on the Gulf of Mexico, and the research presented here creates a database of peak surges for the Atlantic Coast of the United States (ACUS) between the years of 1898-2011. A total of 25 sources were utilized for creation of this database, with many more being consulted but not included in the final product. A database of 72 surge events ≥ 1.22 m was created, with the largest event being a 6.49 m surge in Fairhaven, MA in 1938. Spatial analysis reveals high levels of surge activity in south Florida and North Carolina, with diminished activity from Virginia to Maine. Statistical analysis tested surge frequency and magnitude versus two teleconnections: the Atlantic Multidecadal Oscillation (AMO), and the Southern Oscillation Index (SOI). Statistically significant links were found between AMO phase and surge frequencies; however, no links were found between AMO phase and surge height. No statistically significant links between SOI and magnitude or frequency were detected. Return periods associated with surge events were calculated through three quantile estimation methods – the Pareto distribution and the Huff-Angel and Southern Regional Climate Center (SRCC) linear regression methods. The SRCC method produced the best quantile estimates of surge heights along the ACUS, with a 100-yr event of 6.49 m and a 2-yr event of 1.55 m. A K-means cluster analysis was performed to split the ACUS region into 10 zones, and the SRCC method was employed to produce quantile estimates on a regional level. Zone 7, centered on Charleston, SC, produced the highest 100-yr return period (5.64 m), followed by Zone 1, centered on New Bedford, MA (5.61 m).