Process-driven improvements to hurricane intensity and storm surge forecasts in the mid-atlantic bight: Lessons learned from hurricanes irene and sandy

摘要

The coastal northeast United States was heavily impacted by hurricanes Irene and Sandy. Track forecasts for both hurricanes were quite accurate days in advance. Intensity forecasts, however, were less accurate, with the intensity of Irene significantly over-predicted, and the rapid acceleration and intensification of Sandy just before landfall under-predicted. By operating a regional component of the Integrated Ocean Observing System (IOOS), we observed each hurricane's impact on the ocean in real-time, and we studied the impacted ocean's influence on each hurricane's intensity. Summertime conditions on the wide Mid-Atlantic continental shelf consist of a stratified water column with a thin (10m-20m) warm surface layer (24–26C) covering bottom Cold Pool water (8–10C). As the leading edge of Irene tracked along the coast, real-time temperature profiles from an underwater glider documented the mixing and broadening of the thermocline that rapidly cooled the surface by up to 8C, well before the eye passed over. Atmospheric forecast sensitivity studies indicate that the over prediction of intensity in Irene could be eliminated using the observed colder surface waters. In contrast, Hurricane Sandy arrived in the late Fall of 2012 after seasonal cooling had already deepened and decreased surface layer ocean temperatures by 8C. The thinner layer of cold bottom water still remaining before Sandy was forced offshore by downwelling favorable winds, resulting in little change in ocean surface temperature as Sandy crossed and mixed the shelf waters. Atmospheric sensitivity studies indicate that because there was little ocean cooling, there was little reduction in hurricane intensity as Sandy came ashore. Results from Irene and Sandy illustrate the important role of the U.S. IOOS in providing the best estimate of the rapidly evolving ocean conditions to atmospheric modelers forecasting the intensity of hurricanes. Data from IOOS may enable improved hurricane - orecasting in the future.