The Great Miami Hurricane of 1926: 100 Years of Hurricane Forecasting and Research

Hello, I'm Tara Garwood, the Communications and Multimedia Lead for the NOAA Heritage Program, and I'd like to welcome you to this presentation, The Great Miami Hurricane of 1926, 100 Years of Hurricane Forecasting and Research. This is part of our NOAA Heritage Week 2026 Speaker Series, and we'll be hosting these presentations this year at noon and 2 p.m. Also, if you're in the Washington, D.C. area, on Saturday, September 26, we'll be hosting NOAA Centuries of Science, a pop-up experience from 9 a.m. to 4 p.m. at the NOAA campus in Silver Spring, Maryland. The pop-up is free and open to the public, but you do need to pre-register by September 23 to see the main exhibit, and you can do that at noaa.gov slash nhw2026. We would like to thank the NOAA Studio and the NOAA Library for their support of the NOAA Heritage Week 2026 Speaker Series. And with that, I would like to introduce our two speakers, and then I will let them take it away with their presentation. First up is Dr. Jack Bevin. Dr. Bevin is a Senior Hurricane Specialist at NOAA's National Hurricane Center, where he's served since 1988. He holds a Ph.D. in Meteorology from Florida State University and has advanced through various roles, becoming a Hurricane Specialist in 1999. His primary responsibilities include issuing track intensity and wind forecasts for tropical cyclones in the Atlantic and eastern North Pacific. Beyond forecasting, he is a leading trainer for the technique and an instructor for workshops benefiting both U.S. emergency managers and our international forecasters. Dr. Bevin also chairs the Supervisory Committee for the Hurricane Databases, or HURDATs, Hurricane Tracks Reanalysis Project, and he actively contributes to operationally oriented research. In 2009, he was recognized with the State of Florida Emergency Management Preparedness Program of the Year Award for his contributions to disaster preparedness. And if that isn't enough, we are now going to bring on Dr. Sim Aberson. Dr. Aberson is a meteorologist with the Hurricane Research Division of the Atlantic Oceanographic and Meteorologic Laboratory, or as we call it, AOML. His primary research interest is in using data to improve forecasts of tropical cyclones globally, with an emphasis on observations obtained during NOAA's annual Hurricane Field Program. Dr. Aberson's research has ranged from developmental and optimal use of dropwinds on technology to improving and verifying numerical models and investigating small-scale features in tropical cyclones that may impact future intensity. As part of his research, he has penetrated the eyes of hurricanes nearly 300 times on NOAA aircraft during his career, and that, I think, is utterly amazing. And again, if that wasn't enough, he is also a member of the American Meteorological Society and Out2Innovate, and is the recipient of the 2004 Presidential Early Career Award for scientists and engineers. Sim is also a 2005 Kavli Frontier of Science Fellow and was named OAR Employee of the Year in both 2003 and 2021. Now, you can find an interview with Dr. Aberson in which he speaks on his career in hurricane research at NOAA in our NOAA Heritage Oral History Project. He started as an intern in 1981, so there is a lot of hurricane history there. Thank you, gentlemen, for being here today, and you may start your engines. Okay, thank you, Tara. I'm going to go first because I'm going to give the narration of what happened as the 1926 hurricane approached the Miami area and the overall overview of the life cycle of the hurricane. And you have to understand that back in 1926, and Sim will touch on this in much greater detail later, the data available to the forecasters when these storms were out over the ocean, which was much less than it is now, the little image there on the left is an actual excerpt from the monthly weather review article written by Charles Mitchell in 1926 about the Miami hurricane, and that table is a sum total of all the observations that were available to the Weather Bureau before the storm reached the Florida peninsula. And that is not a very long list, especially when you consider that's four or five days worth of data right there. The observations near this hurricane were very sparse, and as part of the reanalysis project, we went back and tried to see if we could find more observations to help cover the life cycle of the hurricane. We failed to find anything prior to the 14th of September, which is when the storm was first detected, and we failed to find anything during the last 18 hours during its approach to Miami. So even now, with the benefit of looking backwards and being able to comb through the old data sources perhaps a little bit more thoroughly, we could not find much more on the storm than what was available to the forecasters in real time back in 1926, and that was a bit of a problem. Unfortunately, as Sim will point out, we've improved the technology since then. But here's what happened during the chronology of the hurricane's passage and life cycle. On September 14th, the synoptic weather data from the Lesser Antilles, particularly the island of St. Kitts, first indicated that there was a tropical cyclone to the northeast of the Lesser Antilles. Looking at barometric pressure readings and various wind shifts that were occurring, it was apparent that some sort of low pressure area was present to the northeast of St. Kitts, and that was the first known location of the 1926 hurricane. The next day, it was known to be north of the Virgin Islands, based on the wind shifts in the islands and a couple of ship reports that came in. However, it was not until the 16th of September, when the center of the hurricane passed near Turks Island, which is now the Turks and Caicos, near the southeastern Bahamas, where the storm proved to be very intense based on all the reports of wind and pressure and damage that came in from that island. And that was the big alarm bell that went off for the forecasters in the U.S. Weather Bureau, both in Washington and in South Florida, that this storm was very intense and was on moving on a west-northwest track in a general direction of South Florida. Now, on the 17th of September, during the morning hours, it was known to be moving through the Bahamas towards Nassau, and the Weather Bureau requested that Nassau send in extra special observations to help them track the passage of the hurricane through the central and northwestern Bahamas. Unfortunately, the last of those observations was at 1 p.m. that day, as the hurricane got too close and knocked out the communications with the weather station on Nassau, and that was the last piece of weather data that we actually had from near the core of the hurricane before it moved inland over Miami the next morning. So, later that day at noon on the 17th, storm warnings were issued for South Florida. The forecasters had enough confidence at that time that the storm would get close enough that it was going to bring adverse conditions to a lot of the Florida peninsula, particularly the southern part of it, so they started issuing warnings for the approaching storm. Now, by 8 p.m. that evening, the 17th, the barometers at Miami and Key West, and the other people who had barometers in South Florida, showed that the pressures were a little abnormally low. However, they were not falling very fast. A fast-falling barometer would have been an indication that the storm was closing in on some area rather quickly, but at this time, the barometer was not falling fast at any of these stations. However, the forecaster of that day, my predecessor, had a dilemma that, based on persistence of motion and extrapolation from the previous known location of the storm, the center could reach South Florida by 8 o'clock the next morning, or 12 hours from the point that they were trying to make this forecast. Finally, around 10 o'clock on the 17th, the barometric pressure in Miami starts falling rapidly, an indication that the center is moving in that general direction, is getting closer. An hour later, hurricane warnings were issued for South Florida in anticipation that the storm would make landfall there. Now, how were these warnings communicated in 1926? Well, the monthly weather review excerpt talks a lot about the various coordination by phone with various agencies that were in charge of getting the word out to people. So, telephone was the biggest one. Radio was another big one. Radio was not perhaps what it became later, but it was a technology that helped get the word out to people, and in this case, hurricane warnings. Locally, they would use signal flags and lanterns. During the day, they would fly signal flags from various locations around the town, including the weather bureau offices and postcard offices and light guard stations on the beach, and at night, they would use coated lanterns with patterns of lights that reflected certain kinds of warnings, and that's what you see there in the image on the right side, various types of signal flags and lantern patterns. We actually have a replica of one of these lanterns that's hanging in our office that we turn on when we have a hurricane out in the Atlantic Ocean and try to recreate the appropriate lantern signal patterns during this storm event. But this is how we got the word out, and you judge for yourself how effective it was given some of the things that happened the next morning when the hurricane actually arrived. Now, a little bit after the hurricane warning went out, the leading edge of the stronger winds and the heavy rain started to move into South Florida sometime around midnight. Within the next three hours, the conditions were deteriorating over South Florida, and at that point, the telephone lines started to blow down, and that resulted in loss of telephone contact between the Weather Bureau offices and the customers up and down the Florida coast who were depending on the information. A few hours after that, the hurricane winds reached the Miami area, and that started to cause power failures and finished off the rest of the communications and made it very difficult to get any updates out of the Now, the lull in the Miami metropolitan area at the time, which was current downtown Miami, Miami Beach, lasted about 35 minutes. People went outside in the eye, which they thought the storm was over, which did not match what the warnings were telling them, and a lot of people had a mistaken idea of what the storm was about. That comes back to how effective were these warnings, but they went outside, and several of them got killed when the second half of the hurricane came in. Now, based on what we found when we went through this hurricane as part of the reanalysis project, the Miami and Miami Beach areas were actually in the northern part of the eye. They did not have the full lull that probably occurred further south. As best as we can determine, the center of what was probably a 35 to 40 mile wide eye made landfall south of downtown Miami, somewhere near the Deering Estate. The minimum pressure in Miami proper itself was 935 millibars at the Weather Bureau office, and we best estimate based on that that the central pressure further south was probably closer to 930 millibars, although we don't have an absolute reading from the center, and based on everything that we can put together about the size and position and pressure of the hurricane, we estimate that the maximum winds that landfall were near 145 miles per hour, a solid Category 4 hurricane on the Saffir-Simpson hurricane wind scale. Now, later in the day, the hurricane passed Miami, moved further inland, and passed near the southern end of Lake Okeechobee and towns of Moorhaven, Belle Glade, and Clewiston out in that area, and then several hours after that it reached the Florida west coast just to the southwest of Fort Myers. It was over land eight to nine hours before it emerged out into the Gulf. On September 20th and 21st, the hurricane made multiple landfalls along the northern Gulf Coast from the Florida panhandle westward to southeastern Louisiana, and the first of these was a major hurricane, still Category 3, but it made landfall up in that area. I should point out as a forecaster, I would still have to be tracking the hurricane on its track towards the northern Gulf Coast, even though I'd be dealing with the impact of the hurricane locally in Miami, just as we had to do with Hurricane Andrew in 1992. Hurricane would eventually dissipate over Louisiana and the western Gulf Coast states on the 21st and 22nd of September. So, talk a little bit now about the toll and the impact, because the storm was very costly and very deadly. As best it can be determined, 372 people died in the United States due to the hurricane, and 162 of these deaths were in what is now Miami-Dade County, mainly Miami Beach in Miami, and that's based on work done by Rusty Fose, the former meteorologist in charge of the National Weather Service office in Miami. At least 150 people died near Lake Okeechobee when the hurricane pushed the storm surge down into the towns at the south end of the lake. Unfortunately, this was an ominous harbinger of what was to come two years later when the Palm Beach-Okeechobee hurricane swept through that area with a far greater death toll. In terms of the property damage, the total for the United States in 1926 dollars was about 100 million dollars, with about 75 million of that in South Florida. Tremendous destruction occurred in Miami and on Miami Beach. Significant damage to property and crops occurred elsewhere along the southeast Florida coast from Homestead northward to Palm Beach. A similar hurricane today with all the same meteorological characteristics would likely cause property damage in the 150 to 200 billion dollar range due to inflation and the sheer amount of property that is now located in the strike zone of what was the 1926 hurricane. It would be a far more costly event than what we actually went through back in 1926. So that's the look at what the 1926 hurricane did and how it affected South Florida as it moved along. I'm going to turn it over to Sim, who's going to talk about the history of the communications and the forecasting and how it has improved significantly since the actual time of 1926 hurricane. So take it away Sim. Thanks Jack. As Jack said, communications at the time were definitely an issue, not only in gathering the data to make the forecast, but in disseminating the warnings so that the people could prepare their property and remain safe. In these timelines that I have, the red line represents 1926 when the hurricane happened. So what's to the left of the line are things that were available at the time. In the 1860s, the telegraph lines became important and long-range communications could be made. But it wasn't until 1905, just 20 years before this hurricane, that there was wireless telegraph on ships so that ship reports could be sent to land to be used to make forecasts. Remember hurricanes mainly happen over the water and that is why they were so difficult to find at the time. In 1920, the first radio news casts were made, but that was only six years before the hurricane and they weren't widely listened to at the time. But it was one way to get the forecasts out and the warnings. But as Jack said, they did use lanterns and flags, but remember that the warnings went up at 11pm and the hurricane hit at 6.30am standard time. So there wasn't much time and it was at night when people couldn't necessarily see the flags or were asleep. So many people didn't even get the warning. So things that have happened since then, in 1948, TV news started, although it wasn't regularly watched until a few years later. In 1954, the BBC had the first television weathercast where you see here the forecaster making a forecast for the United Kingdom there. So it still wasn't regularly available. Weather warnings weren't regularly available. Some communities had sirens that they used, but you had to watch the news in order to get the warnings and if things changed very quickly, it was very difficult to get the messages out. In 1982, the Weather Channel came online. So instead of having to watch the news every six hours as it was on the broadcast television, we suddenly had 24-hour weather. And I remember that that was a time where we would all gather at the TV at 10 minutes before the hour to watch the tropical weather outlooks on the Weather Channel and see what was going on. So we did have 24-hour weather and 24-hour news at that point, but still communication was a little more difficult. Now we have the internet, we have smartphones, so wherever we are, we have the ability to get news and get warnings and find out what's going on. So communications have definitely changed from that time when there were very few observations to now we can have all the information we need right at our fingertips. In relation to observing, Jack mentioned the very small number of observations that were available to the forecasters. There were some ways to take observations. Centuries ago, we did have an anemometer to measure wind speed, hygrometer to measure moisture, thermometer and barometer. And we knew a little bit about the weather, but not a huge amount. But not much changed through the early 1920s until the first weather balloon went up. And we did know by physics, the equations of motion, that what was going on above the surface was very important, but we really didn't have a comprehensive way to measure that until the first radio sound network in 1937. But that was after the hurricane. So there were very few observations available that we could use. In 1926, a three-cup anemometer that was able to measure wind speed was used. I'm not sure if that was available in South Florida. There were some wind observations taken, but the anemometers did blow away and they didn't survive the hurricane. But since then, a lot has happened. In 1944, regular aircraft reconnaissance began in the Atlantic and the Pacific, where aircraft went out and actually searched for hurricanes. A lot of times we didn't know that they were out there. We only had observations of land and the very few ships that were in the tropical Atlantic. And these aircraft reconnaissance missions certainly helped in discovering hurricanes. And in 1955, we started having land-based radar, which gave us a little bit more idea of what was going on, and we could see some of the structure of hurricanes. And after that, we started to get satellites. So in 1961, we had the first visible satellite imagery and the first images of hurricanes came up. And by 1975, we had the GOES satellite, so we had a very good idea of what was happening over the entire hemisphere that the satellite could see. And then by that time, we couldn't miss any hurricanes. We knew what was out there because we could take pictures of them. But also very important is the change in aircraft reconnaissance over time. In 1956, the National Hurricane Research Project formed, and that was to actually work on improving forecasts of hurricanes, which is now morphed into where I work at HRD. In 1977, we obtained the NOAA P3 Orions, which are still being used. And in 1997, we got the NOAA G4, which is still being used. And those have been the workhorses. They give us a lot of information that we can't get from any of these satellites, even though we now have the ability with some of these new satellite systems to see through the clouds and see some of the structures below the clouds and see things that we can't see with our eyes, like temperature and moisture in the atmosphere. There's still a lot that we can't measure without going into the hurricanes. So the P3s and the G4 have been extremely important in getting that information, especially the dropwind songs. The new one that we're using now became available using GPS in 1997. And the airborne Doppler radar became available in 1982. That tells us the entire wind structure of the hurricane. And by 2008, we're able to make real-time analyses of these observations. And in 2006, we started flying drones into hurricanes. And now we're actually doing that on a regular basis. And those are giving us information where it's unsafe for the other aircraft to fly, but are very important, like where the sea and the air meet and exchange the energy that drives the hurricanes. So what we do here in the research end is we use a lot of these observations, and we develop a lot of these observations. So as I mentioned, the drop songs became available in 1997. We did have a rudimentary form of drop songs that I'm holding there in my right hand, but that couldn't give us any information in clouds. And as you know, hurricanes have clouds. We have UASs and USVs, both are drones. One is in the air, one is in the ocean. They give us a huge amount of information. And we need to know what's going on below the surface of the ocean. And that's where the floats and the gliders come in. They give us a lot of information well below the surface to tell us what's going on down there. And on the aircraft circled, there is two of the radars that we have. We have a lower fuselage radar and the Doppler radar in the back that give us a huge amount of information. We drop songs and other instruments from the plane and the UASs from the plane, and a lot of other instruments that we develop and use that are coming online in the near future. And finally, what we understood about the hurricane was fairly rudimentary at the time and how we made forecasts. So it wasn't until just about 100 years before the 1926 hurricane that William Redfield finally showed that hurricanes were closed circulations. It was thought before that they were just areas of straight line winds. And the first hurricane warning was only 50 years before the 26th hurricane in Cuba. Father Benito Vines issued a hurricane forecast based on information that he received from Eastern Cuba. He was able to tell that there was going to be the hurricane in Western Cuba and made that forecast. In just before the 26th hurricane, scientist Louis Fry Richardson, he attempted the first numerical forecast. And his idea was to fill a stadium with people and they would make all these calculations using abacuses and slide rules. And it ultimately failed for various reasons, one of which was it took longer to make the forecast than the forecast time that they were forecasting for, but it was a good first attempt. But after the 26th hurricane in 1950, Jewell Charney conducted the first successful forecast using the ENIAC computer, but they were not regularly made until much later. Charlie Newman, a colleague of both Jack and I at the Hurricane Center in 1966, 40 years after the hurricane, developed the first hurricane track forecast model. And the first primitive equation, full physics forecast models became available. The first specific hurricane model they used, physics, became available in 1974. And then there were some rapid improvements. The global models came online in the late 70s. We started using the observations that we got from the hurricane flight in the forecast models in 1982, but it wasn't until even 1998, just less than 30 years ago, that skillful intensity forecasts were made available. And in 2008, we had fast enough computers and developed the technology enough that all the different observations that we are able to take from the aircraft, including the high resolution Doppler data, was put in models, a process led by Fuqing Zhang, who was at Texas A&M in Florida State. So there's been just huge improvements since the 1926 hurricanes in our ability to communicate observations and warnings, to take the observations and to make the forecasts and disseminate them. And that has led to really impressive improvements in hurricane forecasts. These are the official verifications of the National Hurricane Center puts out. And one thing that's just so amazing to me is that the five-day forecasts that they make now are about as, they're actually more accurate than the two-day forecasts that they made back in 1970. So huge strides in track forecasts, a little bit of improvement in intensity forecasts, but we're starting to make a little bit more headway now, given some improvements that we've made to observations and improvements to the models. So with that, I will send it back to Jack and he can sum it up. Thank you, Sam. I should point out everything that Sam talked about, all those individual things, satellites, the radar, all the various types of observations and all the developments and forecasting, we can easily talk an hour or two on each of those topics individually. And everything that has improved those forecasts is just the tip of a iceberg of a very large amount of work that has been done to better observe, to better understand, and to better forecast the storms since the 1926 hurricane. Indeed, since the time that Sam and I have become professional meteorologists, the improvement just during that time has been tremendous as well. And I'll point out with new AI techniques that are being experimented with right now, we may be able to go further in successfully improving the forecast. How far we can go, we don't know. People propose that there are theoretical limits to how accurate our forecast can get, but we've blown through several sets of such limits in the course of improving the forecast. So we probably have a ways to go yet that we can make things better. The bottom line is the 1926 hurricane was very deadly, and a repeat of that hurricane today would be very costly. But the good news is we would have a much better idea it was coming, knowing many days in advance and having it being able to track the storm also minute by minute basis as it was going on its rampage through the Bahamas and then across South Florida and then into the Southern United States. So that's how things have changed since the actual occurrence of the 1926 hurricane. So with that, I'm going to turn it back over to Tara. I just wanted to say thank you for the opportunity to talk about this hurricane and the science of how we've improved the forecasting since that time. Well, thank you both, Jack and Sim, for a really interesting presentation. And thank you everyone for watching this presentation. If you enjoyed it, you can find the NOAA Heritage Week 2026 speaker series schedule and links to join the live online presentations at noaa.gov slash nhw 2026. And if you love NOAA's history and culture, or if this presentation piqued your interest, you can also visit noaa.gov slash heritage to learn more about our long history of scientific innovation and service to our nation.