On the Road with NOAA As They Face Uncertainty in Their Agency and the Weather

O
n a long stretch of County Road 106 East, somewhere near the border of Texas and Oklahoma, three conspicuously souped-up trucks pull off to the side of the road. It’s around 6 p.m. on a warm spring evening in late April, and it just started raining, the thick drops teasing that a bigger show might soon follow. I hop out of the back seat, slamming the truck door shut against the persistent wind, and join the line of seven men staring up at the clouds.

“That’s a beautiful structure,” says Sean Waugh, his eyes trained upward under the brim of his baseball cap. “See the left side break, right side break.”

In front of us, the sky splits, a wall of gray clouds thinning at the middle as if being pulled apart in a tug-of-war. We’d been following one storm; now, we have two.

That means Waugh and his colleagues have a decision to make.

Waugh is a research meteorologist with the National Oceanic and Atmospheric Administration’s (NOAA) National Severe Storms Laboratory and a team lead of the LIFT project, a multiyear field experiment to collect data on tornadoes. By the time this storm starts splitting, I’d been with Waugh, two other NOAA research scientists and LIFT leads named Michael Coniglio and Jeffrey Snyder, and a group of four up-and-coming researchers for about a day and a half. That included hours waiting for action in a Buc-ee’s gas-station parking lot, a failed storm chase that ended in a disappointing drizzle, a night at a hotel off a highway, a morning briefing with NOAA meteorologists, a few football tosses, more than 200 miles of driving, and a stop at Taco Bell, which is a pre-storm superstition for the team.

Now, they know we’re in the right place — they’d released a weather balloon to check on the atmospheric instability and had their interpretations validated by a tornado watch sent through by their colleagues at NOAA.

But so far, no tornado.

As it starts warming up each spring, spotting those storms becomes a fixation for these researchers. They live in increments of hours described in universal time to get around pesky time zones. Planning ahead more than a day or two is a challenge; their schedules are decided by a careful combination of weather forecasts, radar images, and the logistics of life on the road.

At the core of this challenge is the fact that even with those decades of data, we still don’t know exactly how or why tornadoes form. That’s why they remain difficult to predict and prepare for, resulting annually in the United States in dozens of deaths and billions of dollars in damages, including property and business losses. In the first half of 2025, more than 60 people were killed by tornadoes, and insurers estimated tornado damages topped $10 billion.

Research at the National Severe Storms Laboratory (NSSL) is bringing us closer to answering that question and has helped lessen the loss of life and property along the way. Over the past 40 years, innovations like Doppler radar have caused a drastic drop in fatalities — almost halved the deaths from tornadoes — while in more recent years, advancements in forecasting and modeling have helped sharpen predictions. That can mean lifesaving warnings reaching communities faster, such as on April 29, just west of Fort Worth, Texas, when residents received notice from forecasters 90 minutes before a tornado touched down. Back in the 1980s, the average lead time for these alerts was closer to five minutes.

These advancements are encouraging, but they still often fall short against the whims of the weather, particularly as climate change makes it hotter and wetter. After a spring of deadly tornadoes, the summer saw a flush of flash flooding across the country, including the July 4 storm in Texas that took more than 130 lives. These disasters raised questions about the capacities of NOAA and its National Weather Service, which have been dealing with existential budget cuts and staffing slashes under the Trump administration.

President Trump’s proposed 2026 budget for NOAA called for cutting more than $1 billion in funding for the agency, in part by scrapping the Office of Oceanic and Atmospheric Research, NOAA’s research division. This would mean shuttering laboratories across the country including NSSL. Congress has the ultimate say over the budget and has so far rejected the president’s push to defund all research, offering a lifeline to keep the National Weather Service staffed and labs open. But the call coming from the White House to dismiss this work — stopping ­studies at a moment when researchers like Waugh believe we are getting critically close to answering centuries-old questions — has many fearing for the future of weather research, as well as the immediate safety of those who rely on their work to stay safe from tornadoes, hurricanes, and all other sorts of extreme weather.

Getting to the Bottom of What Causes Tornadoes

We go for the storm that splits to the right.

The path to following the other side of the split was too challenging, and this one spitting out lightning seems promising. We head back into the trucks with their NOAA logos and research-gear-supporting metal roof racks. I get in with Waugh and Tyler Pardun, a Ph.D. student at the University of Oklahoma who is putting his defensive-driving training to work. Our truck is perhaps the most impressive of the pack, with a panel of LED lights on the back that Waugh tells me produces 30 percent more light than the sun shines on the surface of the Earth. They can illuminate even the darkest storms so that the high-definition, high-speed cameras mounted on the truck can record the hail that often precedes tornadoes. Once the truck captures the images, Waugh feeds them into an algorithm he also designed to surface any notably large hailstones.

“This is a truly one-of-a-kind instrument,” Waugh says. “And it was only created because we have funding at the [Severe Storms] Lab that we can use to try things.”

Waugh settles into the passenger seat, firing up the computer screen installed in front of the dashboard to check the storm models. He cranks up the soundtrack to last summer’s blockbuster film Twisters — it’s been running on a loop, another superstition — and we peel back onto the road.

I had first reached out to Waugh last summer after seeing him onstage in Hollywood, speaking about being a consultant for that very film. He helped build out some of the vehicles used in Twisters, picked radar images that were seen onscreen, and advised on how a real storm chaser would respond in various scenarios. So when actor Glen Powell wanted to know how to react to a wind-turbine blade speeding toward his truck in the movie, he asked Waugh. (Waugh, for the record, said his character would probably be excited.)

Last summer, I received approval from the NOAA press office under the Biden administration to shadow Waugh to learn more about his work, but his field research for tornadoes was wrapping up for the year. So, we made plans to reconnect this spring. In the interim, the Trump administration took office, and by the time we started speaking again, hundreds of NOAA staffers had lost their jobs due to cuts from the so-called Department of Government Efficiency as well as early retirement and resignation offers. As of July, nearly 2,000 of the agency’s staff of 12,000 were gone, with more than 3,000 open vacancies listed. This widespread shake-up has left some offices scrambling to cover shifts and some regions without reliable data, including in tornado alley and along the hurricane-threatened coasts. (NOAA declined to comment on staffing changes.)

“The Weather Service is in a position where they can’t afford cuts regardless, but the particular concern here is that cuts are not being made in a targeted or strategic way,” says Alan Gerard, a 20-year agency veteran who served in leadership roles across multiple departments, including NSSL.

That means there was no oversight to make sure offices were not gutted by these losses, leaving many of them significantly understaffed and unable to conduct basic research, such as releasing weather balloons to collect atmospheric data. In a June 2 broadcast, Florida meteorologist John Morales told his audience that there has been a 17 percent reduction in weather-balloon launches.

“What we are starting to see is the quality of the forecast becoming degraded,” Morales warned.

This also threatens the validity of computer models. James Franklin, a former branch chief of the Hurricane Specialist Unit at the National Hurricane Center, explained to members of Congress that without this consistent collection, it’s unclear what is missing and how important that information might have been, leaving “holes in the data” that’s then used to make predictions.

For Houston meteorologist Matt Lanza, part of the frustration is how proven many of these models and teams already are. He points to NOAA’s Hurricane Forecast Improvement Program, which was started after deadly hurricanes in 2004 and 2005, including Hurricane Katrina. Last year, the agency had its best predictive season on record. According to a National Bureau of Economic Research analysis, the program saves the federal government about $5 billion per hurricane, which came out to about $55 billion in 2024.

“Science is not something where you can just flip a switch and turn something on or off based on who’s occupying the White House,” Lanza says. “A lot of these projects are done over the course of years, and they require a set of data that is continuous, and then if you all of a sudden cut that out, you’re basically risking throwing that whole project away and having to start from zero again.”

Cutting off funding for tornado research would interrupt more than 30 years of tornado-centric field projects from NSSL. While tornadoes have been a human fascination for centuries — mythologized in stories by Native American tribes and a favored subject for founding father Benjamin Franklin — modern tornado research largely emerged in the 1990s. Advancements in Doppler radar unlocked the ability to see wind fields and storm rotation, and in the mid-1990s, NSSL launched the Vortex project, its first major study of how tornadoes form, which LIFT is an extension of today. There was also the blockbuster success of the first Twister movie, which inspired a new wave of meteorologists, researchers, and chasers (Snyder, one of the other LIFT leads, tells me he’s part of “the Twister generation”). And the decade ended with one of the costliest tornado outbreaks on modern record.

In early May 1999, a spate of thunderstorms produced 74 tornadoes in less than 24 hours across Oklahoma, Kansas, and Texas; 46 people were killed, hundreds were injured, and thousands of homes were damaged or destroyed. This disaster changed the way researchers and meteorologists worked within NOAA, according to Gerard.

“Now there’s much more collaboration,” Gerard says. “[There’s] an ongoing feedback loop of ‘OK, this is the problem we’re having with forecasts and warnings. Are there research results that would help us with this, or are there research projects that could be developed?’”

The “lift” of the NSSL LIFT project stands for “low-level internal flows in tornadoes.” The team’s goal is to get information on the bottom 20 meters of tornadoes, which is known as the “damage layer” since that’s the part of the funnel that will tear through a building and toss up a car. By understanding more about the components of this sliver of the storm, the team believes it can help sharpen how we predict and warn for tornadoes, providing insight into why they sometimes show up on the radar but fail to form at the ground level, while also better informing how we categorize them and potentially ­getting even closer to explaining why they happen in the first place.

Analyzing this layer requires getting within a few miles of an active tornado, which is why we’re racing the clouds across Texas. While we wait for a storm, Waugh paints me a picture of what it’ll look like if we see a tornado form. The trucks will fan out around it and fire up their measurement tools, which include mobile mesonets to track data like temperature, pressure, and wind speed; mobile Doppler LiDAR to perform a horizontal scan of the tornado at that low level; and Waugh’s blindingly bright hail-camera setup. Each would capture data that would feed into the whirring computers installed in each truck.

Waugh and his colleagues declined to comment on the federal cuts, both to their agency’s staff and its budget. The closest we come to it is when Waugh talks about how important his agency is in deepening our knowledge of how the weather works.

“If we improve our understanding of tornadoes and what’s causing tornadoes and how different radar signatures look when [they’re] getting ready to produce a tornado — those are things that are tangible that do come out of our research, and those are things that can then be incorporated into a model to increase its ability to produce a valuable forecast going forward,” he says.

A Whole New Tornado Alley to Reckon With

Waugh pulls up another map on his phone, this one covered in pinpricks of red dots. Those, he tells me, are other people following the same storm as us.

It’s a crowded chase, even by Oklahoma standards, where spotting a tornado is a popular pastime. Throughout my time in the trucks, a consistent stream of locals pull up alongside us or walk over to the car to inquire about the equipment, share their favorite tornado sighting, or ask if they should change their evening plans given the weather. More than one offers effusive praise for the agency’s work.

At this point, we’ve been driving in the rain for almost an hour, the drops sliding off the windshield thanks to a hydrophobic coating Waugh rubbed it down with earlier. Waugh instructs Pardun to pull over again once we’ve crossed the border in Oklahoma.

Something’s wrong.

“It’s missing something,” Waugh says from the front seat, window cracked to convene with Snyder. “I don’t know what it is, but I think this thing’s dead.”

“This is the roller coaster,” Snyder tells me.

This is the team’s fourth day of deployment. A few of the guys have sunburns creeping out from under T-shirt sleeves, and the trucks are growing collections of coffees, sodas, waters, and energy drinks in their cupholders. It’s after 7 p.m., and the final remnants of the day’s sun start to peek out through the clouds, a sign that the storm is over but the day soon will be too. The team decides to call it on this storm.

To do this job, Waugh tells me, you have to have the memory of a goldfish. Even after striking out, you need to be able to get up the next day anyway, stare back up at the sky that misled you yesterday, and do it again.

On the surface, the odds seem to be increasingly in their favor. There are approximately 1,400 confirmed tornadoes in the U.S. every year. In 2024, NOAA preliminarily confirmed 1,796, the second-highest number on record since the agency started tracking them 75 years ago. In the first seven months of 2025, NOAA has already received more than 1,350 tornado reports.

But researchers don’t necessarily think this means there are more tornadoes; instead, they think we’ve just gotten better at spotting them. From radar and research teams to the chasers Waugh pointed out on the map, there are also more people looking for them than ever before; there are more people living in more places that are touched by tornadoes than ever before. That means more buildings to be destroyed, and new types of construction, such as wind turbines, put to the tornado test.

While these societal changes have repercussions for multiple types of extreme-weather events, including hurricanes and wildfires, they impact the very way we understand tornadoes. Unlike hurricanes, which are categorized by their recorded wind speed, tornadoes are categorized by the damage they cause. In the wake of these storms, NOAA employees will survey the torn-off shingles, the crumpled grain silos, and the trees with their branches snapped and scattered to assign it a number of the Enhanced Fujita, or EF, scale. That means it’s a scale that has to evolve as we do; as we get better at building wind-resistant homes, for example, it needs to account for how that changes the baseline. It also means that there needs to be sufficient staff to do this post-disaster work — in the wake of an early April tornado in Louisville, Kentucky, low staffing levels at the local NWS office delayed this surveying by days, creating a lag that could make the assessment less accurate or delay clean up.

Then there are other recent tornado trends that continue to baffle researchers, including the fact that the so-called tornado alley — the area where tornadoes have traditionally been most frequently spotted — seems to be moving. A 2018 study found a “significant upward trend in tornado frequency” in the Southeast, Midwest, and Northeast. The Federal Emergency Management Agency has expressed concern over increasing tornadoes in the Southeast in particular, noting that this area is more densely populated than the traditional tornado alley. In addition to being home to more than 85 million people, the housing stock has a high concentration of mobile and manufactured houses and fewer homes with basements, leaving residents more vulnerable to tornado damage. Not only does this spark concern about the readiness of these areas, it also calls into question the very methods researchers like Waugh have used to study them, which were designed to work in the flat, open stretches of the Great Plains.

Up until now, these new challenges were being met with robust research, including a dedicated NSSL team working on tornado research in the Southeast. The agency also isn’t working alone: Multiple universities study these storms, including Texas Tech University and the University of Oklahoma, both of which are partners on LIFT. But federal partnerships and grants are a significant source of support for this work, falling under the same research office that is on the chopping block in Trump’s budget. Expensive and plagued with uncertainty, tornado research requires a level of experimentation that is challenging to fund, even given the potentially life- and cost-saving benefits on the other side.

One research scientist at a cooperative institute that works with NOAA worries these cuts threaten the future of their work, too, which is why they asked to stay anonymous for fear of retribution. But they wanted to convey the importance of these partnerships and how much stands to be lost without them. While the effects of losing forecasters might become clear quickly, the loss of research funding can be harder to spot from the outside but is just as consequential, they say.

“If you get wrong answers as to things like where the biggest risks are and where people need to evacuate, then people die — I mean, that’s the long and short of it,” they tell me.

Teams Scrambling to Do More With Less

After working on Twisters, Waugh fielded questions about whether the tornado-stopping fix the characters devised — dumping chemicals into the tornado that would suck the moisture, and therefore, the life force out of it — could work in real life. For multiple reasons, the answer is no, he tells me. To start, you would need “silos” of chemicals to make this work, which Waugh imagines would create an environmental hazard that would rival, if not surpass, the damage done by a tornado.

“You’re messing with a system that we don’t understand but is desperately needed to balance the energy of the Earth,” he says. “And if you disrupt that in any way, you can cause massive unintended consequences.”

It’s a serious response to a silly question, but it offers a window into the respect that guides Waugh’s approach to tornadoes.

Half an hour after Waugh, Coniglio, and Snyder decided to give up, they start messaging about another spinning storm they see on the radar. It looks promising, but Waugh muses that there might just be too many storms; sometimes, that leads to merges and can reduce the likelihood of tornadoes, though this is another area Waugh says needs more research. Still, he puts the Twisters soundtrack back on.

“This is a weird enough environment to say ‘Alright, we’re here, worth seeing how it plays out,’” he tells me.

It’s dark as we head west, and hail hits the metal cage on the truck’s roof with persistent pings. Pardun sprays water from the shoulder of the road as he drives to yet another parking lot. This time, the team is optimistic enough that the LiDAR-scanning truck turns around and aims its system at the storm.

The sky flashes with lightning as frantically as a toddler with a light switch. The LIFT leads look over the data and decide to keep moving.

The warnings start coming in, first shrill alarms from our cellphones alerting us of a tornado warning, and then the stirring call of tornado-warning sirens outside of the windows. We pull over again, this time on the side of the road, and stay in the car — it’s too rainy, too windy, too dark to go outside. Pardun and Waugh stare out the windshield; I try to decipher from their faces what they’re seeing.

Waugh is the first to cast doubt. In the driver’s seat, Pardun joins in, venturing that it might be too cold for tornadoes, which thrive off of warm, moist air.

It’s now after 9 p.m. We are officially out of chances.

The drive back to Norman, Oklahoma, where NSSL is located and where Pardun and Waugh live, is about two hours. It’s the ­quietest ride I’ve spent with the two. Waugh reminds Pardun to get acquainted with the models for tomorrow morning — there’s a chance they’ll head out once again. After they drop me off at my hotel, the sky un­­leashes peals of rain.

The next day, Waugh texts me an image of a bright teal sky through a wet window and video of hail ping-ponging off the grass. They got one.

There will be more to come as the season stretches on, including one tornado that lazily crosses less than half a mile as it spins in front of them in a field for more than an hour. This is the sort of data that gets Waugh excited — the combination of pre-storm forecasts, metrics measured, and the photos and videos the team captured allow for a comprehensive overview of the conditions.

“That data set is probably a goldmine as far as tornadoes go,” he says. “I think it’s really going to open the door, not only for future observations, but increasing our understanding of how these events are starting, maintaining themselves, and then ultimately dissipating.”

But experts like Franklin of the National Hurricane Center worry that the full potential of research like this won’t be recognized as teams scramble to do more with less — if they’re able to keep operating at all. Because sometimes, when you mess with things you don’t quite understand, you can cause massive unintended consequences.

“My fear,” Franklin said to members of Congress, “is that we’re going to look back 25 years from now and say, ‘This is when the progress stopped.’”