Hurricane Andrew: A 20-Year Retrospective

Dade County, Florida
24 August 1992

On 24 August 1992 (20 years ago this morning), category-5 Hurricane Andrew buzz-sawed into southern Dade County FL, taking just a few hours to become the most destructive natural disaster in American history to that point. I was there for the storm and its aftermath, both personally as a resident, and professionally as a National Hurricane Center meteorologist.

Hurricane Andrew wasn’t merely memorable, compelling or life-changing. Those terms just don’t do it justice. No, Andrew stands alongside the 3 May 1999 tornado outbreak as the most deeply moving and powerful weather experience in the life of this connoisseur of atmospheric violence. Perhaps the most part was that I knew this would be the case, for better or worse, well before the eyewall made landfall, the only uncertainty being specifics of moments to come.

In August of 1992, I was a young meteorologist just a couple of years removed from graduate school, surely wet behind the ears but soaking up tropical meteorology like the sponges sometimes snagged in my Florida Keys fishing tackle. The chance to work at NHC, my first full-time job in atmospheric science, had descended from the clear blue after four years of working part-time at the National Severe Storms Lab. My first two hurricane seasons there–1990 and 1991–offered plenty of learning opportunities and practice on faraway “fish storms”, and the tantalizing teasers of some weak tropical systems nearby. Those only intensified an already ravenous hunger to experience my first hurricane–not wishing destruction upon the area I had come to love as an inhabitant, but instead, wishing to be there should it happen anyway.
I didn’t have this in mind!

The old cliche goes something like this: “Be careful what you wish for, lest you get it.” I wanted a hurricane, and soon was handed a category-5 storm of ages, one that would unleash the energy of multiple thermonuclear bombs directly overhead and for miles around.


The area of thunderstorms and clouds that would evolve into Hurricane Andrew moved westward off the African coast on 14 August 1992, as part of an easterly wave. Soon thereafter, as a chart analyst and satellite meteorologist at the National Hurricane Center, I had the privilege of making the satellite-based classification of this system that led to its initial designation as a tropical depression. Max Mayfield–highly respected satellite meteorologist and hurricane forecaster who eventually would become center director–concurred with the analysis, and issued the first advisory. What became Andrew, ironically, was my first solo storm classification after several months of training.

The system translated west-northwestward for several days and slowly intensified, becoming Tropical Storm Andrew on the 17th–the first named tropical cyclone of 1992. Andrew then turned northwestward, slowed and weakened under the influence of an upper-level low at higher latitudes. It always was a small cyclone–and those are more prone to rapid outside influences from air, land and sea. In fact, Andrew got sheared and disorganized so much by late on the 19th and early on the 20th that RECON aircraft couldn’t find a coherent center anymore. The storm was on the precarious precipice between getting shredded in that area and turning west away from the shearing to intensify.

Although most models turned the system westward into an area of more favorable atmospheric and oceanic conditions for strengthening, some didn’t. The storm looked quite ill at this point, and nearly fell apart. Miles Lawrence, the most experienced of the forecasters there, made the ultimately gutsy and accurate call for a westward turn. His intensity forecast was conservative, which I thought was justified, given its wretched state at the time, and given very real possibilities that it might not even survive the shearing. This was the time of greatest peril to Andrew’s survival and to forecasters’ predictions.

The storm that almost wasn’t anymore turned west to become the most devastating hurricane in Florida history, and the second most intense (behind the similarly sized and even stronger Labor Day storm of 1935).


Once Andrew escaped the hostile environment, cruising steadily westward as the upper low pulled away and weakened, it got over warmer waters. Free of the shackles of shear, Andrew underwent a rapid-deepening cycle, its central pressure plummeting 47 mb in 24 hours. Andrew’s maximum sustained winds later were analyzed to have reached 150 kt (173 mph) mph with a central pressure of 922 mb east of Eleuthera Island, Bahamas (from Landsea et al. 2004). This was the first of two category-5 peaks. Eleuthera was slammed hard.

Watches, then warnings, went up for South Florida, and we were smack in the projected hit zone off most of the model and human forecasts. Operational and some support-staff shifts were changed from 8 to 12 hours until further notice to keep the center at max capacity while allowing time for staff to prepare their homes. On Saturday the 22nd, I went to Home Depot on US-1 to grab and cut one sheet of heavy plywood to fit the sliding glass door of the apartment. Our window was unreachable; so I had to seal and barricade it from within with other materials. The hurricane watch update was broadcast on TV while I was in-store; within 15 minutes, the number of people in the store swelled dramatically. After I left, the place was mobbed. It helped to have some “insider knowledge” of the impending forecasts!

My residence was a fairly sturdy old apartment building that I very deliberately had selected for maximizing hurricane survivability: concrete-pier roof, concrete outer walls with coral-rock facade, second floor unit (of three floors, to stay above freshwater flooding from below and to provide a layer of flooring above against leaks), inner-hall entrance, only two glass exposures to the outside, windowless bathroom, and the glass door and window facing inward to a courtyard instead of outward to full wind. Through a combination of planning and luck, it worked.

Emotions really didn’t hit much, because I didn’t have time for them. What I did experience in the few quiet moments remaining until Sunday night were a mix of weather-weenie anticipation and excitement with dread (for co-workers living in the area and the community at large), and the sobering realization that my first classified system would be my first in-person hurricane. Obviously, within 36 hours, people would die, entire neighborhoods would be leveled, and this place wouldn’t be the same again in my lifetime. Indelible memories were about to be made, whatever happens.

Sunday the 23rd, I was working a noon-midnight shift when Max issued the 5 p.m. forecast package, including the discussion and public bulletin foretelling the area’s impending meteorological fate, and urging preparations to be rushed to final completion. I still had routine map analyses to perform and FAX out, and tropical weather discussions to issue, both on this and ensuing days.

Needless to say, this was no ordinary forecast shift. Assorted media swarmed the operations area, with standing people and wires everywhere, though the media-pool system instituted many years earlier kept the work environment manageable as a form of controlled chaos. Most of the media stayed with Bob Sheets (director) and the senior hurricane specialists; though they did pass off one stand-up live interview to me. I don’t remember which local station it was, but I do recall telling the audience, “This area is going to get absolutely hammered,” and that anyone not in sturdy shelter should get in one before midnight.

Martin Nelson, who lived well to the north in Broward County, was my shift relief. Concerned about Dade County traffic, he arrived there very early at 11 p.m., and sent me home. My then-wife and I used the extra hour to take a long and eerie midnight walk around the area near our apartment.

Strolling through streets utterly silent, we marveled at the absence of car or foot traffic. It was as if aliens had arrived and snatched up the populace. Never before or since have I seen the Miami area so bereft of visible humanity. People indeed were taking shelter, as ready as can be for a maelstrom the likes of which none ever had experienced. Yet our winds were light northerly, with flashes of lightning in the east from outer bands over the Gulf Stream. I pointed out which buildings would suffer what kind of damage from easterly winds of over 100 mph and which way we would find trees fallen and signs bent.

We headed back by 1 a.m., right as the band was moving ashore just a couple miles to our east, winds still no more than 15-20 mph. Winds steadily increased to tropical storm force through about 3 a.m., then the inner bands hit. Still with power as late as 4 a.m., I remained outside on a stairwell’s south-facing balcony near my unit, filming wind and rain with a high-lux, Super-VHS video camera, not getting wet at all thanks to surrounding concrete walls. [The video ended up predictably low in quality and unimportant scientifically, and I never have digitized it.] I stayed out as gusts blasted well above hurricane force, and the eyewall approached.


Greenish blue power flashes lit up here and there along the visible horizon, then quit, as electricity blinked out in neighborhood after neighborhood to the south. Power finally went out at my location, and a very few emergency lights were all I could see in the distance, through heavy rain. The eyewall was moving ashore and nearly upon me. The pressure gradient by now, between the airport and where the eye was coming inland to my south, was 70 mb over the span of 35 miles–insanely intense even for a hurricane. This meant wind–astounding wind–and when the north side of the eyewall moved overhead shortly after 4 a.m., it was obvious. That wind hit, hit suddenly and didn’t let up for a couple more hours.

Within the ambient wind roar, deeper whooshing pitches signaled another approaching downburst streak embedded in the eyewall. As each of these bursts neared, I heard the collective breaking and shattering sounds of glass, trees branches and construction materials to the east. Large objects flew horizontally through the air at well over 100 mph, taking less than the blink of an eye to race past. Based on silhouetted snapshots taken by the eyes, the large flying debris looked like tree limbs, sections of roofing, sheet metal and all manner of other stuff that was unidentifiable in the near-darkness. The concrete wall protected me from direct wind, but swirling eddies were dousing me with rainwater.

One vivid and inextirpable auditory recollection was the car alarms–hundreds and hundreds of them going off at once, most when the first big eyewall burst struck. A few had been sounding off and on since some inner-band gusts about an hour and a half before, but the suddenly massive, anarchic cacophony of them wailed quite audibly above the din of the wind, and in every direction. I thought then and still do today, “Why did all these nitwits leave their car alarms activated, knowing damn well there was a hurricane coming, cars would get throttled for hours, and continual re-activation of the alarms could drain their batteries?” It was mildly amusing later to learn of many folks who couldn’t start their cars as a result of their own lack of foresight. Oh well…it was that many fewer vehicles out clogging the roads anyway.

Still foolishly by that stairwell opening, I noted a few flashes of lightning…not power flashes, but obvious, silvery lightning in the rainy sky aloft. Lightning actually is an uncommon phenomenon in the core regions of most hurricanes due to their deep warmth, and lack of ice particles and upright convection to separate charge. I knew this had to be an unusually and intensely convective eyewall to be producing both lightning and embedded downbursts.

I briefly stuck a hand out flat to the wind just to feel the force–my non-writing left hand, in case a flying object took aim at that spot–and had it slapped right back at me. I’m a strong guy, and still could not hold it out there without bracing my forearm with my other hand. It was, and remains, the most intense wind force I have experienced–by far. Several seconds of direct eyewall winds and painful, stinging raindrops on bare skin were enough; and it was time to go inside.

Back at the NHC and WFO, satellite dishes outside the building swayed and disintegrated, leaving NHC dependent on land-line data feeds. The sixth floor (home of NHC) was shuttered, and one window still broke. Andrew’s road could be heard and felt. Wind-channeling effects around the large office building accelerated the already extreme flow, flipping cars bizarrely in the uncovered portion of the parking area. One of those belonged to an NHC forecaster, Hugh Cobb, who had gone to Washington, DC as part of a backup forecast crew in the event the center was incapacitated. Hugh turned on CNN without realizing what had happened, and shockingly viewed footage of his car in that now-famous predicament.

As the eyewall edge neared, a roof-level gust over 160 mph shook the building, destroyed the rooftop anemometer and shredded the heretofore weatherproof dome covering the 1957-vintage Miami radar. Now a heavy, fully exposed metal air catcher, the big dish and its base toppled onto the roof with a deep, frightening boom. The sound reverberated throughout the structure of the building, clearly audible to all the stunned forecasters and media gathered six stories below. Later, I was surprised it didn’t fall through the roof!

This loop from AOML/HRD has the same color enhancement as the radar image above, and shows the approach of the eyewall right up to the point the dish fell. Another HRD loop shows WSR-57 reflectivity for the entire storm as it approached South Florida. The only NEXRAD Doppler radar to sample the storm was the unit at Melbourne, 177 statute miles to the north. Please check out my new scan-by-scan javascript loop of Andrew’s crossing of South Florida, using that Melbourne data.

Having passed across very warm gulf-stream waters, with no atmospheric impediments to intensification, the eyewall actually was strengthening and the cloud tops cooling around 9 Z (5 a.m. EDT) as Andrew ‘s eye crossed the shoreline! Between 0914-0927 Z (5:14-5:27 a.m. EDT), pilots of a USAF C-13 Hurricane Hunter plane made a daring and very rare overland penetration of a major hurricane, north-south across Dade County at an altitude of about 10,000 feet over my residence. Ted Fujita, pioneering tornado scientist who also had a keen interest in tropical cyclones, included a plot and analysis of that aircraft’s wind data in his subsequent survey results. In the later reanalysis of Andrew’s winds, HRD researchers determined a maximum surface wind on the Biscayne Bay shore of about 145 kt or 167 mph, within a range of analysis methods yielding up to
153 kt or 176 mph (image from Landsea et al. 2004). The 145-kt figure stands as the maximum wind of record, with a central pressure at Florida landfall of 922 mb.

Several miles SW of NHC, we retreated to the interior restroom with our two cats and waited the rest of the eyewall out, listening to a radio simulcast of TV meteorologist Bryan Norcross. We could hear the muffled roar and assorted banging noises through the air-conditioning vents and feeling low-frequency vibrations in the walls and floor.

We had one funny moment in there, and it was how I suffered my one hurricane-related injury. One of the kitties was an exceedingly friendly and adorably stupid male tabby named Shadow, who seldom lacked clumsy and awkward experiences uncharacteristic of most cats. He is the only cat I’ve ever known who sat down on a lit candle; and he did it at least three times in his life. Shadow could leap and miss a shelf or table, then try again and again, seemingly unaware of each previous failure. The little guy got mildly panicked at a loud noise, and leaped right past a shower curtain into the filled bathtub. I saved the now fully-panicked, screaming kitty from drowning, but not before he sliced up both the shower curtain and part of my arm with his claws. Shadow did not repeat that maneuver! Ever since, the event has been known as “Hurricane Shadow and His Little Storm Surge”.

Winds began subsiding shortly before any daylight was visible, but still were hurricane force from the southeast at the crack of dawn. I trudged out into the inner hallway, now wet with in-blown rain spray, and over to a north-facing balcony shielded from wind. Overhead, even as the surface winds continued to subside, scud clouds silhouetted the grim gray-blue pall of a deeply overcast dawn, racing past at astonishing speeds that I haven’t seen before or since. Power poles lay bent or snapped. Broken trees and tree branches littered the area, along with the usual assortment of roofing and wall material. In the distance, I could see my rain gauge, long before wrapped and strapped to the top of a fencepost with chicken wire and reinforced the day before. [We had just 2.64 inches total.] Only a small fraction of vehicles had busted windows, and from a distance, mine didn’t appear to be one of them. Just a few lingering car alarms meekly cried themselves to sleep, headlights barely blinking amidst dying batteries.

I grabbed my video camera and headed for the car in between the thin inner bands, as soon as shingles and other small objects stopped flying. Before the storm, I couldn’t park the Meatwagon in sturdy shelter and still have it readily accessible to get to work at noon for my next shift. On Sunday night, knowing the direction of the strongest winds to come, I faced it westward, so that the front windshield would be on the leeward side. I fully was prepared to find it dented to hell, back glass broken by flying objects.

Luckily, the only damage was a dent on the top panel from a 6-inch diameter tree limb that had bounced there and come to rest on the front fender. It was drivable, and I had a well-inflated spare tire and several cans of Fix-A-Flat for upcoming days’ treks hither and yon. I threw the big limb aside, started her up, and drove into the immediate neighborhood. As expected, signs were bent and trees lay aground and westward; dark traffic signals oscillated back and forth in the gusts and rain; power lines lay across roads and sidewalks; and assorted boards and other small debris in the roads compelled slow, careful motoring to avoid nails and glass. Mine was one of the first few cars out and about; so I had to go slowly and dodge a great deal of obstacles. On a couple of occasions, I was able to get out and drag tree branches off the roadway in order to proceed.

Radio reports told of southern Dade being incommunicado, with the few reports coming from the area being frightening accounts of significant, tornado-like devastation across miles and miles. Of course, that’s what happened; the eyewall effectively acted like a 15-mile wide EF2-EF3 tornado (EF4 in spots!) translating at 20 mph. Every road leading farther south was blocked by stuff too big to move–whole trees, roofs, and in a couple of cases, overturned vehicles with no occupants inside. Presumably those had been parked, but blew over. A few people had begun wandering streets, some aimlessly, with dazed and confused expressions. Soon it was time to report back home and get ready for work.


Although still in an extended set of 12-hour shifts with overtime, I was able to snag a few hours here and there, over succeeding days, to drive the Meatwagon into accessible areas and photograph some of the damage, carefully securing permission from residents and owners before stepping onto any property. Most folks actually were quite eager to share their experiences and allow me to shoot, as soon as it became evident I was not a looter. Some asked where I worked, and I was honest; but I didn’t use NHC credentials for access. Far more time was spent talking to folks than shooting pictures. This was a necessary and sobering experience, imparting a lifelong appreciation for the hardships of disaster victims, and their under-appreciated resilience in the face of destruction.

Driving and stopping across mainly the Krome Avenue, Silver Palm Drive and US-1 corridors of southern Dade County, with side trips to the bayside areas of Coral Gables, I saw far more than I shot. One morning, while driving NE on US-1 near Perrine, I heard a few pops of small-arms fire about a block away and some hollering to the effect of “Stop! Come back here!” Somebody was shooting at looters, and I wanted no part of the crossfire…time to get the hell out of Dodge.

Back then, I had 20 years less photographic experience, but understood the unforgiving nature of Kodachrome slide film with respect to dynamic range. I also was worried about camera-battery longevity–the one thing I hadn’t remembered to address before the storm–and had just one old spare of questionable vigor. These factors meant I probably was too conservative in my shooting. My only regret is not doing more aftermath photography, despite the time and film limitations and 12-hour shifts at an air-conditioning deprived NHC. Nonetheless, two decades later, the shot below [and here, bigger] remains my best-selling in terms of licensing; and just like photos I make today, I didn’t have a dime in mind when I took it.

Shift work at NHC and the Miami forecast office was steamy and uncomfortable, as expected. Few complained, knowing of the much greater suffering going on to our south and southwest. I grew up in Dallas with no air conditioning, and re-acclimated to the conditions reasonably well. Bottled water was around, and somehow the taps still ran. As long as we didn’t drip sweat into the keyboards or each other’s food stashes, everything was fine. The A/C-supporting generator had failed, but the other one was just enough to run some of the lights, computers and COMMS equipment. To conserve power and temper the heat, many computer consoles and monitors (back then, running on hot cathode-ray tubes) were switched off. Only a handful of fans ran, also to save power. We even opened some windows to the Miami summer in order to disperse heat and allow a minuscule modicum of ventilation. All of us were surprised at how well and how long the remaining equipment ran in the sweltering conditions!

Come what may, NHC still had the full set of bulletins and routine products to issue, including those covering the swirling destroyer that had left us for a date with Louisiana. We managed to get all forecasts out without backup help. Eventually, air conditioning was restored to the NHC facility, debris was cleared in fits and spurts from southern Dade County, power came back slowly to devastated areas, and the populace set about recovery. Bob Sheets and his management crew were compassionate enough to allow extra time off as needed for those who lived in South Dade to deal with their own family and home situations; fortunately none of them were injured, and the rest of us made sure all shifts got covered. Bob himself, as well as Max, the late Stan Wright, and several others, had homes in the heavier damage areas. Those at NHC overnight did their jobs without interruption even as the eyewall was pounding their residences mercilessly.

Ultimately, we each got a certificate and a $2,000 cash award for our service. Unlike many governmental bonuses, no one dared begrudge us that. I don’t work for recognition or awards, but instead for the intrinsic merits of excellence. Still, that was a much-appreciated gesture on the part of the Department of Commerce, and the extra pay helped everyone along in recovery.

Post-event visitors included Tim Marshall, whom I knew from our Great Plains storm-chasing experiences, and Ted Fujita, each of whom came to perform detailed damage surveys. Fujita attributed a lot of the heaviest damage in Naranja Lakes and nearby areas to “mini-swirls”–horizontal-shear eddies on the inner eyewall. He specifically and deliberately declined to refer to the mini-swirls in the eyewall as “tornadoes”. I spoke with him about this while he was there–one of only two times I ever got to have a full conversation with him, despite our many common interests–and his reasoning resonates clearly to this day. He could not establish physical vertical continuity of the eddies entirely upward into Andrew’s intensely convective eyewall plume, precluding their classification as tornadoes, by definition. He also expressed some doubt to me that eyewall tornadoes are real, despite all the reports, in absence of direct evidence (photo, video, or something none of us knew about then–very high-resolution mobile radar). I’ll be discussing this issue more in an upcoming EJSSM manuscript that is in final production stages.

By most measures, Andrew still ranks as the second most expensive hurricane in U.S. history, behind Katrina, doing $26.5 billion worth of damage in 1992 dollars. Most of that was in South Florida. Despite the relatively thin population and suburban to rural nature of southern Dade County, Andrew laid waste to at least 28,000 homes. Mobile home parks and poorly built subdivisions, such as Country Walk, resembled the aftermath of nuclear war. Longstanding tourist attractions–Miami Metrozoo, Monkey Jungle, Parrot Jungle (since relocated northward and renamed Jungle Island), the Deering estate, Orchid Jungle, and others, suffered debilitating damage. Andrew destroyed Homestead Air Force Base, from which all but the disabled planes had been scrambled in the few days prior to landfall. Although the Turkey Point nuclear power plant survived mostly in good shape, one of its concrete stacks cracked severely, and had to be brought down with explosives.

Ecologically, Andrew had impacts in South Florida and beyond that last to this day, some destined to persist for generations to come. The storm busted open public, academic and private enclosures containing exotic flora and fauna, releasing them into the surrounding air, land and waters. These included the sacred ibis, some research monkeys since gone feral, and the ongoing scourge of Burmese pythons in the Everglades. The storm surge, 4-6 feet over most of Biscayne Bay but measured at 16.9 feet near the Burger King headquarters, ripped lose vegetation of all kinds and shoved it inland–ironically, creating a “debris dam” that kept saltwater penetration from going extreme distances. Along with the devastation of mangrove wetlands and assorted low-lying timber habitats by waves and winds, and the smothering of oyster reefs by sedimentation, a total of nearly 200 million fish were killed in Florida and Louisiana, according to USGS.

Andrew’s impacts really do last a lifetime and beyond. Even today, broken tree trunks, sidewalks to nowhere, and empty lots liberally speckle the southern parts of what’s now Miami-Dade County–fading pixels in the fuzzy picture of destruction’s footprint. As long as civilization’s digital footprint persists, photos, low-definition video and most of these stories will pass through the generations.

South Florida learned a lot from that experience, but what of that knowledge has it retained? Will ignorance and denial of hurricane safety, and of sound building practices, set the stage for even greater devastation and carnage in the inevitable next “big one”–perhaps even larger and stronger and more directly aimed into the city? As the scars fade and the young survivors still living there grow elderly, will the newer generations heed their words of caution? Or will it take another 1926, another Betsy, another Andrew, to reset the complacency meter in the metro populace? According to the Claims Journal, Andrew would be three times as expensive today, and that only gets worse as more and costlier stuff is placed in the next Andrew’s path.

My kids have four quite deliberately given hurricane names between them; and my son (born a little over a year later) got Andrew as a middle name. My children’s grandchildren likely will be told of this by their parents and pass the story on to their offspring sometime in the 2100s.

There was no need to chase Andrew, it hunted me down. I’m forever grateful for that experience, while empathetic toward those who lost far more. When I travel back to NHC for conferences, which has been nearly every year since the late 1990s, any of the current and former staff who were there for Andrew reminisce and swap some stories. Funny how, unlike fish tales, Andrew never seems to get bigger, badder, more intense, more fantastic with time. It was big, bad, intense, and fantastic enough.


More Hurricane Andrew information is available through these links, which will open in a separate panel. Links work as of this writing, but may change at any time!

HTML: NHC Web Summary of Andrew

HTML: NOAA Interactive Path Map

HTML: Max Mayfield remembers lives lost to Andrew

HTML: NHC Preliminary Report–storm summary and track-intensity table

PDF: A Reanalysis of Hurricane Andrew’s Intensity by Landsea, et al. 2004 (Bull. Amer. Meteor. Soc.)

HTML: SkyPix Hurricane Andrew damage photography

WEBINAR: Hurricane Andrew: A Look Back by Robert Molleda, WFO Miami

VIDEO: Assorted clips of Bryan Norcross’ TV coverage (WTVJ-4)

VIDEO: AP 20-year Retrospective

HTML: 10-year NOAA reanalysis announcement for category-5 upgrade

VIDEO: Miami Herald 20-year Retrospective

PDF: Monthly Weather Review 1992 season summary by Mayfield, Avila and Rappaport

VIDEO: Extended IR Satellite Loop from NOAA

HTML: Rainfall Summary by David Roth, HPC

HTML: SPC Tornado Map (Andrew produced no tornadoes in Florida.)

HTML: Effects of Hurricane Andrew (1992) on Wetlands in Southern Florida and Louisiana by USGS

HTML: National Hurricane Center main site


Leave a Reply

You must be logged in to post a comment.