Hazard Potential of Compound Flooding from Rainfall, Storm Surge, and Groundwater in Coastal New York and Connecticut
Abstract. Compound flood events, defined here as the co-occurrence of more than one flood type, can result in flood hazard potential that is higher than if the events occurred independently. To evaluate compound flooding in a semi-urbanized coastal area, historical records dating back to 1970 are used to study the co-occurrences of high precipitation, storm surge, and shallow groundwater conditions for Long Island and the Long Island Sound vicinity across coastal New York and Connecticut. Joint return periods for coincident precipitation-surge events were computed using fitted copulas and compared to the assumption of independence as a ratio of return periods, referred to here as a return period adjustment. Results indicate distinct seasonality where compound events in the area disproportionately occur in the cold season between October and April. Return period shifts range from 1 to almost 9, demonstrating the range in precipitation-storm surge dependence across the study area. Across all 24 station triad locations, groundwater levels were elevated during times of precipitation-storm surge co-occurrence, in areas where the average depth to water is shallow (less than 20 feet or 6 m below land surface). The result is a pseudo-trivariate compound flood potential map that integrates dependence between daily precipitation-surge events and overall monthly groundwater levels into a relative compound hazard score. The location with the highest compound flood hazard score is in the south shore of Long Island, as well as locations across coastal Connecticut where groundwater levels are already near-surface during events where both heavy rainfall and high coastal storm surge occur at the same time.
Glas et al have presented statistical analysis to demonstrate the probability of multi-flooding-events occurrence in the New York City region. Additionally they have proposed a flood score based on their statistical results to qualify the spatial flood risk in the region. This is a highly effective study with rigorous analysis and potential applications.
While the analysis in the manuscript appears to be sound, I would suggest further detailing and describing the methodologies and results for improved understanding across a wide audience. I have listed my comments below for the authors consideration: