Just How Unprecedented Was Hurricane Sandy?

Mean sea level pressure for 0600 UTC October 29, 2012 (contoured) with standard deviations from normal (shaded). Image courtesy Pennsylvania State Meteorology Department.

The first question we ask is: just how unprecedented was Hurricane Sandy? While the Perfect Storm of 1991 is a good analogue in terms of meteorological setup, it did not have nearly the kind of impacts Sandy produced over the Mid-Atlantic states and Long Island since it developed much further out to sea. The historical record shows that there have been other hurricanes to affect the northeastern United States. Notable examples include Hurricane Donna of 1960, Hurricane Carol of 1954, the Great New England Hurricane of 1938, the New York Hurricane of 1893, and the Norfolk and Long Island Hurricane of 1821. While these hurricanes have been classified as being fully tropical, it is difficult to know for sure about those that occurred during the pre-satellite era. Limited data we have suggests that the 1938 hurricane was perhaps an event similar to Sandy in terms of baroclinic enhancement of a tropical system. In either case, a “Sandy-like” event appears to occur no more than once every 50-100 years. Another unique and ultimately devastating attribute about Sandy is that she had the highest Integrated Kinetic Energy (IKE), a measure that combines size and strength of the wind field, of any tropical system on record with almost 4 times the IKE of Hurricane Katrina. The high IKE associated with Sandy can be primarily attributed to the storm’s incredible size. Lastly, the minimum sea level pressure field associated with Sandy was less than 9 standard deviations below normal while off the Mid-Atlantic coast, confirming that Sandy was perhaps a once-in-a-lifetime event.

NHC track forecast errors, courtesy NOAA.

While Sandy was both historic and disastrous for the Northeast, there was one other historic side of the story that is actually positive: the forecasts for the track of Sandy were spot-on. For perspective, 5-day National Hurricane Center (NHC) forecast track errors back in 1970 were 518 mi, which would have encompassed the entire shoreline from southern South Carolina through northeast Maine. Even in 1990, 5-day forecast errors were still 345 mi, which encompasses everything from southern North Carolina through northern Massachusetts. For Sandy, NHC forecasted a landfall near Atlantic City, NJ five days prior to actual landfall. Amazingly, the verifying landfall location was only 30 mi south of the 5-day forecast! Thanks to improved model resolution, superior data assimilation techniques, improved model physics parameterizations, and utilization of ensemble methods, NHC is now able to predict points of landfall with unprecedented skill. Additionally, because of recent advances in ensemble forecasting, it is now also possible to better convey the uncertainty in the forecast than ever before. Ensemble forecasts can now reliably depict whether the uncertainty is in the cross-track or the along-track direction, and implementation of this information into the forecast is currently underway. These advances are only possible because of improved computing power and the unwavering ambition of hard-working meteorologists and computer scientists who devote their lives to improving the models and advancing the science. While we cannot always expect a 5-day forecast to go as well as it did for Sandy, track forecast errors continue to fall year-by-year in the Atlantic. This means better advance warning for the public and emergency management, more time to make preparations and protect property, and, most importantly, fewer lives lost.

Will Komaromi
Ph.D. Student
Meteorology & Physical Oceanography
Follow the Rosenstiel School on Twitter: @UMiamiRSMAS
“Like” the Rosenstiel School on Facebook: www.Facebook.com/Rosenstiel School
Circle the Rosenstiel School on Google+ : Rosenstiel School