Understanding the Physical Drivers of NYC's Extreme Rainfall from Hurricane Ida via Ensemble Sensitivity

The remnants of Hurricane Ida brought record-breaking rainfall rates to New York City on 1-2 September 2021. Central Park recorded 80+ mm of rain in just one hour (0100–0200 UTC), producing devastating urban flooding that resulted in 13 fatalities in NYC. To better predict—and prepare for—these types of events, it is crucial to analyze the synoptic and mesoscale mechanisms that facilitated such extreme rainfall. To this end, we used 6-h forecasts from a retrospective run of the National Severe Storms Laboratory's Warn-on-Forecast System (WoFS), a convection-allowing 18-member ensemble initialized hourly, to analyze storm attributes contributing to extreme precipitation and identify the sensitivity of forecasted precipitation to the surrounding environment and critical physical processes.

We evaluated the sensitivity of WoFS rainfall forecasts in Central Park and surrounding regions by investigating the spatial correlations between key atmospheric fields—such as total column precipitable water and mixed-layer convective available potential energy (CAPE)—and the hourly rainfall intensity in NYC at different lead times. This analysis provided insights into how specific environmental conditions correlated with the extreme rainfall observed in the wettest ensemble members. Additionally, we compared ensemble members that generated the heaviest rainfall in NYC during 0100–0200 UTC on 2 September with those that did not. The “wet” and “dry” ensemble groups were compiled by selecting the top 5 wettest and driest members based on NYC point rainfall from each initialization from 2000–2300 UTC, resulting in two groups of 20 ensemble members each. Composite means of the two ensemble groups were analyzed to identify statistically significant upstream differences at earlier lead times.

The results indicated that the WoFS ensemble members with the greatest 0100–0200 UTC rainfall rates in NYC were characterized by enhanced total column precipitable water along the New Jersey and Mid-Atlantic coasts at one- and two-hour lead times. Greater mixed-layer CAPE was also noted in a smaller region in Southern New Jersey for the wet ensemble members. The atmospheric circulation patterns, including an enhanced Atlantic subtropical ridge in the wet ensemble, were identified as potential contributors to steering heavy-rain-producing supercells into NYC. Other contributing mechanisms to the extreme rainfall are discussed. We conclude by integrating the various factors identified in our analysis into a comprehensive narrative highlighting some of the primary causes of the devastating Ida rainfall in NYC.