Only four days before hitting the Miami area as a Category 5 hurricane 20 years ago on Aug 23, 1992, Andrew seemed headed for a niche as a weather trivia question: “What happened to 1992‘s first named storm?”

Forecast model tracks for Andrew at 8 a.m. on Aug. 20, 1992. (Prepared by Bob Sheets for the book “Hurricane Watch”)

Few people paid much attention when Max Mayfield, the hurricane specialist on duty at the National Hurricane Center in Miami, issued a bulletin at 11 p.m. Sunday, Aug. 16 that a tropical depression had formed 3,500 miles east-southeast of Florida.

At 5 a.m. the next day, specialist Hal Gerrish reported satellite images indicated that the depression’s sustained winds were 40 mph, just strong enough to make it Tropical Storm Andrew.

The first U.S. Air Force Reserve WC-130 into the storm, flying from Antigua on Wednesday, found a surface air pressure of only 1006.1 millibars, which is high for a tropical storm. Was the new storm dying? When the airplane found a 73 mph wind, talk of Andrew dying was on the back burner, at least for the moment.

But at 5 a.m. Thursday morning, Aug. 20, an Air Force WC-130 couldn’t find a storm center at its 5,000-foot altitude. In fact, Mayfield says, the lowest pressure the airplane found was 1015 millibars; “there are high-pressure systems that aren’t that high. For a tropical storm you should have a pressure [no more than] 1005 millibars.”

With such a high “low” pressure and no sign of winds circling a center, Andrew technically wasn’t even a tropical depression.

On the other hand, even higher air pressures outside the storm created enough of a pressure difference to produce 62 mph winds in one area observed by the airplane. And, one of the eight computer models NHC forecasters used then showed Andrew would soon move into an area favorable for development.

Hurricane specialist Richard Pasch compared this model’s forecast with upper-air data charts and animated satellite images that show movements of water vapor in the air and concluded that the model’s forecast “looks realistic,” which wasn’t always the case in 1992.

After some discussion, NHC Director Bob Sheets decided that although Andrew technically might not be a tropical storm, it was better to continue calling it one for the next 24 hours. He didn’t want to downgrade the system and then upgrade it again a few hours later. Such announcements would be scientifically correct, but they could erode public perception of Andrew’s strength and possible future threat.

Bob Sheets, center, talks with forecasters at the National Hurricane Center. This area was in the background when Sheets, NHC Director from 1987 to 1995, gave his television briefings. (Credit: Jack Williams)

After deciding to keep Andrew listed as a tropical storm, Sheets and the hurricane specialists worked to make sense of models predicting a wide variety of future paths for Andrew and how fast it would move.

Thursday morning’s official forecast was that by 8 a.m. Sunday (Aug. 23) Andrew would still be a tropical storm with winds less than 74 mph approximately 550 miles east of Miami.

By Friday morning, however, forecasters analyzing time-lapse satellite images, data from the Bahamas and Caribbean islands, and from reconnaissance airplanes decided that Andrew is strengthening with winds at 60 mph. Mayfield predicted the storm would be a hurricane within 24 hours and about 390 miles east of Miami.

At 4 a.m. Saturday morning specialist Lixion Avila called Sheets at home to tell him that Andrew is gaining strength and moving faster toward Florida. An hour after this call Andrew became a hurricane. Sheets told emergency management directors they needed to speed up their preparations.

Sheets called in extra staff; men and women who now knew a hurricane was threatening their homes. He also activated the media pool agreement under which broadcasters provided a single camera and crew for the networks and a second camera and crew for local stations.

Inconsistent models continued to bedevil the forecasters. But at 5 p.m. Saturday the Hurricane Center posted a hurricane watch for the Florida Coast from Key West to Titusville, 212 miles north of Miami.

At 11 p.m. Saturday, after information from a hurricane hunter airplane showed Andrew strengthening, specialist Miles Lawrence, after talking with Sheets, predicted that Andrew would move ashore between Miami and Homestead on Monday as a Category 3 hurricane with speeds of 125 mph.

Tomorrow: At the Hurricane Center when Andrew hit

Jack Williams covered Hurricane Andrew from the National Hurricane Center in Miami for “USA TODAY” from early Sunday afternoon, Aug. 23, 1992 until after Andrew hit Louisiana the morning of Aug. 26. His account is based on the chapter about Andrew in “Hurricane Watch: Forecasting the Deadliest Storms on Earth” by Dr. Bob Sheets and Williams, published in 2001 by Vintage Books, and an interview this week with Max Mayfield.