Shortly after Hurricane Milton formed in the Gulf of Mexico in October 2024, forecasters watched in horror as the storm underwent explosive rapid intensification and made its way toward Florida, strengthening from a Category 1 to a Category 5 hurricane in a day.
Milton's wind speeds increased by 90 mph in 24 hours, which Jason Dunion, a University of Miami meteorologist specializing in hurricane research, said is the third highest rate of rapid intensification in recorded history.
"If you're a forecaster, you have to communicate that change really fast," Dunion said.
That kind of fast-paced strengthening, which in Milton's case famously brought one Florida meteorologist to tears on air, has long posed a problem for forecasters and hurricane-prone states like Louisiana, where residents are often left with little time to prepare.
While researchers understand the large-scale environmental factors that help fuel the phenomenon — warm sea-surface temperatures, low wind shear, moist air — much of the process remains a mystery.
"What's harder to predict," Dunion said, "is how fast the inner core, that heart of the storm, can change in a short period of time."
Hurricane Milton on October 7, 2024.
Part of the problem is that it's not easy to get to the core of the hurricane. But experimental tools that researchers think could help improve rapid intensification forecasting showed promise during the 2024 Atlantic hurricane season. And it was a good season for testing — it saw a boom in episodes of rapid intensification, nearly double the average of the last 10 years.
A few standouts in a vast field of ever-evolving technology are aimed at collecting data from within the hard-to-reach parts of a storm, where Dunion said researchers believe answers to the rapid intensification riddle could be hiding.
"That’s the part that we still need to understand better," Dunion said.
Tools to help hurricane prediction: 'No one lives at 10,000 feet'
Three pieces of new technology came up repeatedly in conversations with Dunion and other hurricane specialists: The tail Doppler radar, StreamSondes and drones.
A "flying laboratory" that allows researchers to fly into hurricanes to create a 3D X-ray of a storm, the tail Doppler radar consists of a dish inside a cone on the tail of the National Oceanic and Atmospheric Administration's Hurricane Hunter aircraft, Dunion said. It continuously measures wind speeds and rainfall as researchers fly through storms.
Hurricane researcher Jason Dunion on a Hurricane Hunter mission.
Hurricane Hunters can only safely fly through the tops of storms, generally at around 10,000 feet, but Dunion said the tail Doppler radar allows researchers to gather data from deep within a hurricane, as low as just above the ocean's surface.
With a 3D rendering of the entire storm, researchers can better see things like whether its inner core is tilting, which happens when the center of a storm's circulation in the lower and middle levels of the atmosphere are not vertically aligned. Stronger hurricanes tend to have aligned centers, while weaker storms tend to tilt.
Researchers think tilting, or the lack thereof, could have some correlation to rapid intensification, Dunion said. The tail Doppler radar could help them find out for sure.
Joseph Cione, the lead meteorologist for emerging technologies with NOAA's Hurricane Research Division, said his team is focused on closing that same data gap with the lower core of storms, which, really, are the parts that impact people and land.
"No one lives at 10,000 feet, right?" Cione said. "We all live down here."
Tiny weather stations called StreamSondes, which are about the size of a jar of tomato sauce and weigh just 14 grams, are dropped out of Hurricane Hunter airplanes directly into storms. StreamSondes measure temperature, humidity, pressure and wind speeds as they fall, transmitting the data back to researchers on the plane.
StreamSondes are tiny weather stations that are dropped into tropical storms and hurricanes to gather data.
Cione said StreamSondes, which were released in high volumes for the first time in 2024, are significantly smaller and lighter than previous generations of similar technology, allowing them to hover in the atmosphere for nearly twice as long. And longer hang time means more data.
Researchers also found last season that if they released multiple StreamSondes at once into a storm, a process known as "swarm deployment," they recorded varying wind speeds and other atmospheric measurements, even from within small sections of the same storm.
"So we learned that, wow, there's more variability than we thought," Cione said.
A graphic shows what StreamSondes look like while deployed in swarms.
Cione's specialty, though, is drone technology. His niche interest for the last 20 or so years, Cione said drones always seemed like an obvious way for researchers to get to the most dangerous parts of hurricanes.
In 2005, Cione led the first successful unmanned aircraft flight into the core of a mature tropical storm, Ophelia. Two years later, he flew a drone into Hurricane Noel.
Joseph Cione, NOAA's lead meteorologist for emerging technologies, with unmanned aircraft system Coyote.
Researchers have come a long way since then, Cione said: His team conducted 19 drone deployments into storms last season alone, breaking records for wind speeds, communication distance and durability. The air-deployed drones Cione is testing now weigh about 3 pounds, have some steering capabilities and can fly through a storm's most intense winds for over an hour.
Come June, Cione said he hopes to get that time up to two hours.
"So this hits the data void," he said. "This is an area we can't get information."
Data first
Despite strides in technology, the 2024 season proved to be a tough one for forecasters to nail down. Errors in the National Hurricane Center's storm intensity forecasts were up in 2024 compared to years prior, thanks in part to an explosion in episodes of rapid intensification.
John Cangialosi, a senior hurricane specialist at the NHC, said there were 34 episodes of rapid intensification — which the NHC defines as an increase in sustained wind speeds of 35 mph in 24 hours — recorded last year. That's compared to an average of 18 episodes per season over the last 10 years.
Still, NHC trends suggest hurricane intensity forecasting has become increasingly accurate since 1990. There's other evidence of movement in the right direction, too.
From NHC
When Cangialosi started his career at the NHC 19 years ago, he said forecasters might have guessed a storm would undergo rapid intensification, they might have discussed it with their colleagues, but they never would have said so in an official forecast. Now, they can forecast rapid intensification, and despite the phenomenon's unpredictable nature, those forecasts are getting better.
From 2010 to 2014, Cangialosi said NHC intensity forecasts were generally about 25 to 30 mph too low for storms that rapidly intensified. So if forecasters predicted a storm would peak with winds of 100 mph, it would end up with winds of 125 mph. Cangialosi called that a "big miss."
“Today that error has been cut in half," he said.
Cangialosi said that kind of progress in forecasting, which translates to lives and property saved, wouldn't have been possible without advancements in data gathering and storm hunting technology, which some experts worry could be threatened by the Trump Administration's cuts to NOAA.
“Any decrease in data at all would hurt," Cangialosi said. "From a perspective of data first, the more data you can get, the better your ability is to analyze the current state and make predictions. And the Hurricane Hunter data set is still the best for hurricane analysis and predictions.”
