This is a special OCP Seminar, celebrating the 75th anniversary of Lamont Doherty Earth Observatory.
Title: Why do tropical cyclone disasters surprise us in the mid-latitudes? Insights from probabilistic assessment, simulation, and flood modeling
Abstract:
A fundamental challenge with coastal storm hazard assessment for many mid-latitude areas is that tropical cyclones (TCs) are responsible for the largest events but occur infrequently relative to extratropical cyclones (ETCs). Therefore, historical TC hazard distributions are typically under-sampled. As a result, in spite of large differences in TC and ETC maximum intensities, observation-based assessments of surge, wind and rain hazards typically merge data from TCs with far more numerous data from ETCs.
First, I will demonstrate the challenge using Monte Carlo analyses of synthetic storm tide hazard data for a range of mid-latitude storm climates. Extreme value analysis using merged TC and ETC data results in low-biased estimates of the 100-year storm tide event, and the bias is not mitigated by adding longer durations of historical data. Separating storm types before performing EVA helps avoid low bias in hazard estimates, but also suffers from greater random error. Seeking and applying the longest possible datasets or model-based synthetic storm data can be used to help reduce this error.
Second, I will show how the threat of these TC “outlier” events can be accentuated by compounding hazards. The analysis uses New York City as a midlatitude demonstration site, applying bivariate statistical analyses including copula modeling with historical hourly rainfall (1948-present) and storm surge, separating storm types into TC, ETC and “Neither” events using storm track data. Results again show that TCs have markedly different driver characteristics from other storm types and dominate the joint probabilities of the most extreme rain-surge compound events, even though they occur much less frequently.
I will conclude by showing how these outlier events overwhelm defenses in outsized proportion, surprising society. Recent flood modeling research on Hurricane Sandy and post-tropical cyclone Ida reveals how infrastructure built for the bulk of the storm distribution (the ETCs or other frequent events) is easily overwhelmed by the outlier TC events.
