A Note on the Reanalysis of Hurricane Andrew’s Intensity
Aug 23, 2012
Editor's Note: AIR Principal Atmospheric Scientist Dr. Tim Doggett and Senior Writer Robert Zalisk recall NOAA's reclassification of Hurricane Andrew—more than a decade after the storm devastated South Florida.
Twenty years ago, on the morning of August 24, 1992, Hurricane Andrew struck the east coast of South Florida near the small town of Homestead (population about 30,000 at the time), roughly 40 miles southwest of Miami. According to the 5 AM EDT advisory from the National Hurricane Center (NHC), Andrew came ashore with winds of about 140 miles per hour—a Category 4 hurricane on the Saffir-Simpson Scale. Andrew's winds and a 17-foot storm surge caused massive damage. Parts of Homestead and nearby areas were literally flattened and when the losses were ultimately tallied, Andrew had cost the insurance industry $15.5 billion.
Ten years later, in the course of a multi-year National Oceanic and Atmospheric Administration review and evaluation of the entire Atlantic Basin hurricane database, Andrew was determined to actually have been a Category 5 hurricane at landfall.
The Atlantic Basin Hurricane Database Reanalysis Project
Beginning in about the year 2000, the National Oceanic and Atmospheric Administration began updating the official hurricane database for the Atlantic Basin (known as HURDAT). This is an ongoing effort to both extend the database back in time and to revisit and revise, as necessary, the official tracks and intensities of tropical storms and hurricanes from 1851 to the present. Following the project's sequential ordering of the database, Hurricane Andrew originally had been expected to be examined in 2005. In the summer of 2002, however, the NHC requested that Andrew's re-evaluation be addressed sooner since that year was the tenth anniversary of the hurricane's landfall.
The reanalysis concluded that Hurricane Andrew had actually manifested Category 5 conditions when it made landfall. It found that Andrew's peak winds in South Florida had not actually been measured—largely because the available measuring instruments had failed or had been destroyed. For example, the Coastal Marine Automated Network station at Fowey Rocks recorded—just before it failed—an 8-minute average wind of 142 mph with a peak gust of 169 mph in its last transmission. The reanalysis study noted that Andrew's winds at this station had increased dramatically in the last hour of reporting. Taking into consideration the location of the Fowey Rocks station, Hurricane Andrew's location and movement at the time, and an estimate of the surface radius of maximum winds for that portion of the storm, the study calculated that Fowey Rocks would likely have encountered the highest winds in Andrew's northern eyewall at about 08:20 UTC—about 20 minutes after its final report.
Assuming only a linear increase of winds during this time (a conservative assumption), the study estimated that the maximum one-minute surface wind at the Fowey Rocks station would have been approximately 170 mph.
The study went on to note that the observations of both the Fowey Rocks and another (similarly failed) recording instrument were taken in the northwest portion of the eyewall, but that Andrew's strongest winds were likely situated closer to the storm's center in the northern part of the eyewall. Thus, the study concluded, neither observation (each of which was in the Category 4 range of the Saffir-Simpson Scale) would seem to represent Andrew's maximum winds at landfall.
Hurricane Andrew's Official Reclassification and Closing Thoughts
In 2004, the NHC officially reclassified Hurricane Andrew as a Category 5 hurricane. It decided that it was "quite unlikely" that Andrew was a Category 4 hurricane at landfall in 1992, as had been thought. Rather, Hurricane Andrew's maximum one-minute surface winds were "very likely to be in the range" of 157-178 mph, with a single best estimate of 167 mph—a Category 5 hurricane.
It is important to note that the reclassification applied to the maximum sustained wind experienced at landfall. Furthermore, as noted in the reanalysis report, Category 5 conditions occurred only within a small area in the southern part of Miami-Dade County, close to the coast in Cutler Ridge. In reality, the damage and resulting insured loss from Andrew reflect the storm's wind field over a large area—both in Florida and in Louisiana, where the storm made a second landfall—and the distribution of the winds within that footprint. In modeling such events, there is always a range of uncertainty regarding what the true maximum sustained might have been, and the reclassified maximum for Andrew was within the range of values already considered.
Ultimately, the reclassification of Andrew did not materially change the modeled view of the frequency of intense storms making landfall in South Florida. Neither did it change the actual reported losses from this event—the billions of dollars of loss data that are used to validate model output. Thus, the act of reclassifying the landfall intensity did not have a significant impact on the modeled loss for this event.