It was no normal mission.
Many of the Hurricane Hunters on the flight were seasoned, accustomed to taking on the world’s most powerful storms. Some had notched hundreds of hurricane eyewall penetrations – badges of honor coined eyewall “pennies.”
But even for the veterans, this one felt different.
It was after midnight in late September 2022 as the crew of 17 – engineers, scientists, and pilots – took off from a Houston runway for a flight into Hurricane Ian, a Category 3 storm on a collision course with the sandy shores of southwest Florida.
They were flying a 1970s-era turboprop wrapped in bulky scientific equipment. In the fuselage of the craft was a large canister containing an AI-powered autonomous unmanned aircraft that would be deployed to gather data from inside the eye of the hurricane.
If the mission succeeded, drones like this one might one day redefine hurricane exploration, further improving forecasts, which in turn could protect more lives – a crucial task in an era of runaway coastal populations and extreme hurricanes made worse by climate change. Drones could even eliminate the need for dangerous flights like this one, which send humans directly into the eye of raging windstorms.
Or so that was the plan.
The first Hurricane Hunters
Pilots have been intentionally flying into hurricanes since before the Second World War. In July 1943, in the most famous of these early flights, U.S. Col. Joe Duckworth successfully piloted a single-engine trainer aircraft into the eye of a hurricane off the Texas coast. On his second flight that day into the eye of the so-called “Surprise” Hurricane, he took with him a weather officer who jotted down the first temperature readings gathered while flying inside the eye of a hurricane.
Within a month of Duckworth’s flight, the first planned hurricane surveillance flight was made by the United States Army Air Forces – the predecessor of the U.S. Air Force – just east of the Bahamas. Ever since, crewed aircraft surveillance of hurricanes, conducted by the squadron known as the Hurricane Hunters, has continued to support forecasters.
But flying people directly into a hurricane is risky business that can sometimes prove deadly. In 1955 a Navy Hurricane Hunter and its entire crew were lost while flying into Hurricane Janet south of Jamaica.
Even missions that don’t end in disaster can be accompanied by scary moments and close calls. In 1989, a NOAA research team that included flight meteorologist Jeff Masters – now a Yale Climate Connections contributor – hit extreme turbulence while flying into a rapidly strengthening Hurricane Hugo, which temporarily sent the airplane and its flame-filled engine hurtling toward the mountainous, white-capped seas of a Category 5 hurricane. Mercifully the plane recovered from its dive – only 880 feet before slamming into the ocean – and the crew limped back to safety in Barbados.
It raises the question of why – decades after the first flights into hurricanes – weather forecasters still send piloted airplanes into the most extreme storms on the planet. What if there was another way to get the data forecasters need, without the risk to human life?
The promise of drones
The idea of flying drones rather than humans into hurricanes isn’t new.
Taiwanese scientists initiated an autonomous reconnaissance project in 1998 for typhoons, the hurricanes of the western Pacific. From 2000 to 2004, the Taiwan Aerosonde Team launched six small unmanned aerial vehicles into typhoons, but all six attempts failed, with each drone crashing before reaching its target eye.
It wasn’t until September 2005 when the first successful uncrewed mission into a tropical cyclone took flight from atop the roof rack of a Chevy Silverado at NASA’s Wallops Flight Facility in eastern Virginia. As part of a U.S. research endeavor between NASA and NOAA, the same drone platform used by Taiwanese researchers completed a trip through the outer rainbands of Tropical Storm Ophelia while the storm churned off the coast of North Carolina. Though the aircraft never got to within about 25 miles (40 km) of Ophelia’s center, it did prove its worth, gathering weather observations at very low altitudes – as low as 1,200 feet (366 m) – an important part of a hurricane where it’s too dangerous for crewed aircraft to fly.
The experiment was repeated by NOAA scientists in Hurricane Noel in November 2007, but this time the Aerosonde vehicle made it into the hurricane’s eye, a first, down to an unprecedented low altitude of just 300 feet (91 m). The tests into Noel and Ophelia were a triumph, especially for research scientist Joseph Cione, a meteorologist out of NOAA’s Hurricane Research Division in Miami, the principal investigator on both projects.
Cione had been getting unmanned aircraft into hurricanes for almost 20 years. His quest sprung from his obsession with energy exchanges at the air-sea interface – the lowest layer of the atmosphere where the spray of the ocean kisses the winds of the sky – a place no right-minded scientist would ever venture within the violent, Colosseum-like walls of a hurricane.
Computer model simulations suggest that better data about the transfer of heat and moisture in this hurricane danger zone – within about 2,000 feet (610 m) of the ocean surface, where tiny beads of sea spray fuse into the turbulent air above – could be critical for accurately forecasting the intensity of a hurricane.
Forecasters use maximum winds, found within the donut-shaped eyewall immediately surrounding the eye, to estimate a hurricane’s intensity. And getting a hurricane’s winds right matters. Doubling a hurricane’s intensity from 75 mph to 150 mph doesn’t simply double the damage potential; it increases it by over 250 times.
But for decades leading up to the first drone experiments, the performance of hurricane intensity forecasts had essentially flatlined despite huge improvements in hurricane track forecasts, which predict the storms’ movements. Cione theorized that data, gathered with drones flying at low altitudes, might finally crack the code.
Coyote and Kermit
In the early days of Cione’s airborne lab, success felt like a luxury.
“We’d have one success and two failures, and it was a good day,” he says. “It was a lot of trial and tribulation. Funding was hard to get without a lot of success.”
But after the two successful land-based drone launches into Ophelia and Noel, Cione finally felt a tail wind. In late 2007, after presenting preliminary findings at a scientific meeting, a U.S. military official approached Cione with a novel idea. Instead of launching a drone from land, researchers could instead deploy a small, expendable unmanned aircraft traditionally used in military surveillance and warfare from aboard NOAA’s crewed Hurricane Hunter aircraft as it flew inside a hurricane.
“You supply the drone,” Cione responded, “and we’ll supply the flight hours.”
The Trojan-horse-like approach of air-based drone deployments was a breakthrough, potentially solving several problems with earlier land-based deployments. Land launches required scrambling to send teams to an approved launch site close enough to a hurricane – typically within 100 miles (160 km) or so – for the limited range of communication with a small drone. Hurricanes often form hundreds of miles away from land over the open ocean, leaving scientists only a few cases and a narrow window to operate. Getting prior FAA clearance for drone activity was an arduous process and bad weather, ironically, often impeded launch attempts from land.
Launching lightweight drones from crewed Hurricane Hunter aircraft would sidestep these logistical challenges and expand the pool of test candidates. It would also allow scientists to compare observations from the drone with those from crewed aircraft simultaneously investigating the storm, helping to verify their accuracy and offering a means of calibration for the new technology. What’s more, researchers aboard the crewed aircraft would be able to steer the drone toward features of interest only detectable by conventional Hurricane Hunter instruments.
It was an ambitious idea, but if achieved, it could catapult hurricane exploration.
In the ensuing years, Cione and his research team tested these new military drones from NOAA’s planes in blue skies. But it wasn’t until he found funding for his own drones in the wake of 2012’s Superstorm Sandy and the passage of the landmark $50 billion Disaster Relief Appropriations Act of 2013 that he was able to turn the idea into a reality.
In September 2014, a small Coyote unmanned aircraft system was launched inside a canister underneath the belly of the NOAA P-3 Orion research plane nicknamed Kermit – after the Jim Henson signature Muppet – in Hurricane Edouard. The 13-pound Coyote was only a fraction of the size of the first-generation land-based Aerosonde drones Cione had used seven years earlier.
After launch, a parachute attached to the canister immediately deployed and, after falling for about 15 seconds, the Coyote shed the canister shell, unfurled its wings, and flew for nearly half an hour, to heights as low as 3,000 feet (914 m), collecting wind, humidity, and pressure readings in the most intense section of the Category 3 hurricane.
It was a roaring success for meteorologists reaching deeper into the hurricane danger zone than anyone before. Following Edouard in 2014, seven other Coyote missions were successfully flown into Hurricanes Maria in 2017 and Michael in 2018.
Despite the wins, researchers thought they could do even better. The fleet of Coyote drones so far hadn’t strayed more than 18 miles (29 km) from their crewed launch plane. Their limited and unreliable communication range made longer jaunts nearly impossible.
What forecasters needed was an expendable drone capable of relaying observations to scientists a hundred or more miles away, operating not for dozens of minutes but for hours to fully scan the lowest and most dangerous layer of hurricanes.
In other words, they were searching for the Holy Grail of hurricane hunting: a small, AI-powered, uncrewed aircraft system that would no longer require an escort plane, untethering humans from a hurricane’s dangerous core.
Cione had an idea of how to do it. But testing the idea would require taking scientists on one of the wildest flights ever recorded – straight into the belly of a Category 5 hurricane.
A drone fit for war
The COVID pandemic didn’t do Cione and his research team any favors. The safety protocols and clearance hurdles during the early years of the pandemic temporarily sidelined his now decadeslong pursuit.
But the period also ushered in the next generation of small drones – those based on new machine-learning technology leveraging recent advancements in artificial intelligence, or AI.
The ALTIUS – short for Air-Launched, Tube-Integrated, Unmanned System – is a drone designed for military warfare. Built by Anduril Industries, a Silicon Valley defense startup named for a sword from J.R.R. Tolkien’s “The Lord of the Rings,” the ALTIUS 600 was among the equipment included as part of the $50 billion U.S. security assistance package committed in June 2024 to Ukraine. In combat, the ALTIUS can be equipped with an explosive warhead to carry out kamikaze missions, but outside of war, the workhorse system could be repurposed for dangerous reconnaissance, like hunting hurricanes.
What makes the ALTIUS groundbreaking are the independent sensors that fly it. They allow the drone to process hundreds of decisions each second, learning and adjusting for whatever the mission demands. When tailored for hurricane hunting, this advanced autopilot is preprogrammed to steer toward the hurricane eye. It then determines how to stay within the strongest winds encircling the eye, all while spiraling down toward the ocean surface and collecting data in the hurricane danger zone.
The ALTIUS wasn’t the smallest drone in NOAA’s fleet. Weighing 27 pounds (12 kg) with an unfurled wingspan of eight feet (2.4 m) once launched – making it over twice as wide as it is long – it technically qualifies as a small drone according to FAA regulations, but not by much. Its larger frame ensures the drone can operate even in the harshest conditions and stay airborne even after sustaining considerable damage. More importantly, whereas the older Coyote drones could only operate within tens of miles from the crewed launch aircraft, the new ALTIUS drone could communicate up to 275 miles (440 km), which in theory would keep the Hurricane Hunters at a safe distance while the drone surveilled the hurricane’s treacherous inside.
But for the ALTIUS’s inaugural flight, a theoretical range of communication was too risky to count on. Big hurricanes don’t come around often, and the drone would require a close chaperone until it proved its range within a hurricane. That meant Cione and his team would need to experience the inaugural flight hand-in-hand with the ALTIUS.
Into the eye
Hurricane Ian had the entire Gulf coastline of Florida on edge.
After tearing across western Cuba as a Category 3 hurricane on Tuesday, September 27, 2022, Ian had emerged over the Florida Straits and appeared to be reloading on final approach to southwest Florida.
Forecast models showed the possibility of a dangerous major hurricane striking somewhere between the state’s capital in north Florida and the coastal communities at the end of its long-tailed peninsula. Smack in the middle of the forecast cone, just 24 hours before projected landfall, was Tampa Bay, home to over 3 million people.
On the ground in Houston, Texas, Cione gathered with 16 other NOAA crew members to prepare to fly into the storm. It’s normal for Hurricane Hunter missions into major hurricanes to activate the nerves. Even so, the team had the sense this was going to be a historic flight. Ian was a Category 3 now, but what would they find when they flew into the core?
Cione knew it would be his only shot to test the ALTIUS in Ian. There wouldn’t be enough time for another mission before the hurricane struck.
It was 3 a.m. local time on Wednesday when they took off east across Galveston Bay toward the shroud of darkness on the horizon. Behind them, the blanket of lights from the city disappeared into the midnight sea.
‘That’s not normal’
Ian met them with a light show.
Most of the seasoned crew had never seen so much lightning in a hurricane. The sterile cabin of the hollowed-out, generations-old research plane, resembling an aging computer server room peppered with bulky terminals and server racks, was illuminated by the nearly continuous flashes of lightning as it bounced about at 10,000 feet.
As the plane jolted, the scientists aboard considered what the nonstop lightning could portend. Intense core lightning had been observed in rapidly strengthening hurricanes before, like in the run-up to Hurricane Maria’s Category 5 rating in 2017, but it was faraway satellites and land-based detection networks, not Hurricane Hunters, that had observed those strikes.
By 6 a.m., the plane reached the edge of the most extreme winds and rain: Ian’s western eyewall. Though the sun wouldn’t rise for another hour, rapid flashes lit up the sky. Then a sudden, violent pull made the plane feel like it would split in two. The crew members were accustomed to extreme up-and-down turbulence – the kind of momentary weightlessness you might feel while on an amusement park drop tower – but most had never felt anything like this uncontrollable twisting, a side-to-side movement called yawing.
Unsecured instruments heaved through the cabin. Bunks careened into the galleys. Hot coffee washed down aisles and across equipment. The passengers’ four-point harnesses ripped as the plane lurched wildly and the pilots wrestled to maintain airspeed and altitude and prevent a catastrophic engine stall. Some flight members turned ghost-white. Others laughed nervously.
One of the meteorologists aboard was Josh Wadler, now an assistant professor of meteorology at Embry-Riddle Aeronautical University in Daytona Beach, Florida. Although Wadler had flown into other hurricanes, including Category 5 eyewalls, he wasn’t as seasoned as others on the flight. He was seated next to another scientist who’d been flying hurricanes for 20 years and was anxiously holding on tight as the plane contorted.
“That’s not normal,” Wadler thought.
After 10 hellish minutes, Ian loosened its grip. What had likely been a mesovortex along Ian’s eyewall – a finger of swirling winds resembling a tornado in the sky curling into the calm center – thrust the 50-ton patrol plane into Ian’s eye. They’d made it.
Read: Meteorologist turns frightening flight into research project
With little time to waste, the crew readied the ALTIUS for launch. At 6:30 a.m., they jettisoned the drone. The ALTIUS dropped a few thousand feet, then spiraled downward, arcing counterclockwise along the eyewall’s edge. Next, it descended on Ian’s southwest side, where it struck the same violent mesovortex that had nearly dismembered its crewed launch aircraft.
The plane circled the eye, banking along the cloud cliffs that surrounded them, scoping upward to heights of over 50,000 feet. Lightning lit the predawn, cloud-free sky above, filling the towering stands of nature’s grandest amphitheater.
The 27-pound ALTIUS performed admirably inside the mesovortex. Not only did it survive the whirlpool of turbulence, it took measurements throughout its time inside, documenting wind gusts as high as 216 mph (348 kmh), the strongest winds ever measured by a small uncrewed aircraft.
The ALTIUS flew for another 62 minutes around Ian’s eye, sending back a steady stream of temperature, pressure, humidity, and wind readings.
Meanwhile, Cione and the NOAA flight team cut their mission short and found the safest escape route out of Ian’s core, eager to have engineers check the integrity of the aircraft in the aftermath of their brutal entry.
By the time the ALTIUS had completed its work nearly two hours after it was launched – crashing into the Gulf of Mexico as planned – it had blown past the mission objectives. The scientists had continued to receive data from the drone even when they were 150 miles (240 km) away, a range that proved they could launch the drone from safely outside the hurricane’s core on future missions. Most importantly, the ALTIUS survived some of the strongest winds imaginable.
Cione was ecstatic.
“It was a knockout mission,” he said. “It did everything.”
Are crewed missions on the way out?
Cione and his research team would go on to receive a Gold Medal, the highest award from the U.S. Department of Commerce, which oversees NOAA and the National Weather Service, for their successful drone deployment and distinguished performance in Hurricane Ian. It was the culmination of a career-long crusade that transformed a scientific curiosity into a lifelong search.
But that Gold Medal wasn’t merely a cul-de-sac in Cione’s career. It was a bridge that took Cione and his team to the next stage of hurricane hunting, one that includes smaller, safer, and more cost-efficient tools to study hurricanes, aimed at improving forecast models and protecting the lives and property of tens of millions living at the head of Hurricane Alley.
The 27-pound ALTIUS drone will stay in NOAA’s hurricane-hunting arsenal for now, but the team has augmented its fleet with three-pound Black Swift S0 fully autonomous drones. These ultra-low-cost drones not only capture temperature, pressure, humidity, and wind readings like the ALTIUS but are equipped with lasers to measure sea surface temperature and the height of waves inside the hurricane. The first of these was successfully tested during Hurricane Tammy in October 2023, where it flew as low as 100 feet (30 m) above the North Atlantic.
In early August 2024, four more Black Swift S0 drones were launched from crewed Hurricane Hunter missions into Hurricane Ernesto as it churned toward Bermuda. But this time, the drones were launched from outside the hurricane’s core, transmitting measurements back to the crewed launch plane at distances up to 164 miles (264 km), breaking the range record from Ian.
The research team is planning to launch 14 additional expendable S0s in the 2024 Atlantic season, along with up to four ALTIUS 600s, including one with a 360-degree camera capable of capturing streaming video in the hurricane danger zone.
Crewed Hurricane Hunters are still needed right now. They routinely launch other expendable instrument packages used by hurricane forecasters that can’t be replicated by existing drones. And until heavier and more expensive equipment like Doppler radar can be safely mounted aboard drones and recovered, people will continue to fly into the eye of the storm.
But Cione does see an end to crewed hurricane hunting coming. Though he’s not willing to bet it’ll happen before the end of his time in active research, the long-term picture seems obvious: “We won’t have Hurricane Hunters on a plane anymore at some point in my lifetime. It doesn’t make sense.”
We help millions of people understand climate change and what to do about it. Help us reach even more people like you.