What Are Storm Chasers and How Do They Work?

Storm chasers are people who deliberately drive toward severe thunderstorms, tornadoes, and other dangerous weather systems to observe, document, or collect scientific data from them. They range from university researchers deploying mobile radar trucks to hobbyists with dashboard cameras and a fascination with extreme weather. How they work depends on who they are and what they’re after, but the basic process involves forecasting where severe weather will develop, positioning themselves in its path, and then making real-time decisions as storms form and evolve. The tools and motivations have changed dramatically over the past few decades, and the line between professional scientist and self-taught enthusiast has blurred considerably.

Who Storm Chasers Actually Are

The popular image of a storm chaser is someone in a beat-up truck racing down a dirt road toward a tornado. That image isn’t exactly wrong, but it captures only a sliver of the community. Storm chasers fall into several overlapping categories, and the mix has shifted over time.

Research chasers are scientists or graduate students attached to universities or government labs. They participate in organized field campaigns with names like VORTEX2 or RELAMPAGO, coordinating fleets of vehicles and instrument platforms to gather measurements that fixed weather stations simply can’t provide. Their goal is data: wind speeds, pressure drops, temperature gradients, humidity profiles. Everything they do feeds back into models that help forecasters predict severe weather more accurately.

Broadcast chasers work for television stations or freelance for media outlets. They supply the dramatic footage you see during tornado coverage. Their goal is visual documentation, and they tend to get closer to storms than most researchers would prefer, because compelling video is the product. A growing subset streams live on social media, which has created its own economy around real-time storm footage.

Then there are hobbyist or recreational chasers, who do it because severe weather fascinates them. Some are deeply knowledgeable about meteorology and contribute useful storm reports to the National Weather Service. Others are casual tourists who follow a guided storm-chasing tour during a vacation to the Great Plains. The community is larger than most people imagine, and on a high-profile tornado outbreak day, roads in central Oklahoma or the Texas Panhandle can get genuinely congested with chase vehicles.

How a Chase Day Works

Storm chasing isn’t a spontaneous activity. A typical chase begins hours or even days before anyone starts driving. The process has a predictable rhythm.

The first step is forecasting. Chasers study weather models, surface observations, upper-air soundings, and satellite imagery to identify where conditions favor severe thunderstorm development. They’re looking for a combination of instability (warm, moist air near the surface with cooler air above), wind shear (winds changing speed or direction with altitude), and a triggering mechanism like a front or dryline that can initiate storm development. The Storm Prediction Center, run by the National Oceanic and Atmospheric Administration, issues convective outlooks that most chasers use as a starting point, but experienced chasers dig much deeper into raw model data.

Based on that analysis, a chaser picks a target area, sometimes hundreds of miles from where they woke up that morning. Driving three to six hours to reach a target is routine. Once in the target area, they wait. This can mean sitting in a gas station parking lot for hours, refreshing radar on a laptop, watching the sky, and re-evaluating whether the forecast is coming together. Many chase days produce no tornadoes and no dramatic storms. The atmosphere is fickle, and even the best forecasters get blanked regularly.

When storms do fire, chasers position themselves relative to the storm’s structure. Most try to stay on the southeast flank of a supercell thunderstorm, which is typically where the best view of the rotating updraft and any tornadoes is, while remaining out of the path of the storm’s heaviest rain and largest hail. Real-time navigation requires constant decision-making: which road to take, whether to let the storm pass or reposition ahead of it, whether conditions are becoming too dangerous to stay close. GPS, mobile internet, and portable radar apps have made this easier than it was in the 1990s, but it still comes down to judgment.

Mobile Radar and Close-Range Observation

The most significant scientific tool in storm chasing is mobile radar. The Doppler on Wheels (DOW) program, which has been operating for decades, uses truck-mounted radar units that can be driven to within a few miles of a storm and scan it at much finer spatial and temporal resolution than fixed National Weather Service radars, which might be a hundred miles away. This close-range advantage matters enormously for understanding the small-scale processes inside severe storms.

During the RELAMPAGO field campaign in Argentina in 2018, three DOW radars were deployed simultaneously to observe a supercell thunderstorm. The dual-Doppler scans allowed researchers to reconstruct three-dimensional wind fields inside the storm, revealing a tight correspondence between vertical air motion and rotation at mid-levels, which confirmed the storm’s supercell structure.¹ That kind of detail is invisible to distant fixed radars.

DOW units have also been used in less obvious settings. During the winter of 2010–2011, a DOW radar was positioned near Lake Ontario to study lake-effect snowstorms, cataloguing over 1,100 small-scale vortices within just two storm events.² Storm chasing, in the scientific sense, isn’t limited to tornadoes on the Plains. Any severe or poorly understood weather phenomenon can become a target.

Probes, Drones, and Getting Inside the Storm

One of the persistent challenges in tornado research is that the most interesting part of the storm, the tornado itself, is also the most dangerous place to put an instrument. Researchers have spent decades trying to solve this problem, and the solutions have gotten progressively more creative.

In-situ probes are rugged, weighted instrument packages designed to be placed in the path of an approaching tornado and survive the experience. The Pressure Acoustics Recordings Inside Tornadoes Experiment deployed probes between 2016 and 2019 that successfully sampled three EF2-rated tornadoes, recording temperature, humidity, and the pressure deficit near or inside the vortex.³ Getting a probe into the right spot requires a chaser to predict where a tornado will travel, drive ahead of it, deploy the instrument on the road, and then get out of the way. The margin for error is measured in minutes and sometimes seconds.

Unmanned aircraft have opened a new approach. The Tempest Unmanned Aircraft System was the first drone specifically designed to fly into supercell thunderstorms. During the VORTEX2 campaign, it became the first unmanned aircraft to sample the rear flank downdraft of a supercell, a region of descending air that plays a critical role in tornado formation but is extremely difficult to observe from the ground.⁴

By 2019, the technology had advanced to the point where multiple autonomous fixed-wing drones could be deployed simultaneously into different regions of the same supercell, each collecting atmospheric data independently.⁵ This is a meaningful shift. Instead of one data point from one instrument in one part of the storm, researchers can now build a picture of conditions across the entire storm environment at the same time. The drones are expendable compared to a manned aircraft, which means they can be flown into environments where no pilot would go.

Infrasound and Newer Sensing Methods

Not all storm-chasing instruments are as dramatic as a probe sitting in a tornado’s path. Some of the newer tools are subtle enough to ride along in a chase vehicle without much fanfare. One example is GLINDA, a ground-based infrasound data acquisition system that records very low-frequency sound waves generated by tornadoes and severe thunderstorms. Tornadoes produce infrasound that travels long distances and carries information about the vortex’s intensity, but until recently there was no practical way to collect these signals in real time from the field. GLINDA, which has been deployed with storm chasers since May 2020, uses an infrasound microphone, GPS receiver, and inertial measurement unit to automatically monitor the acoustic spectrum and detect peaks that may correspond to tornadic activity.⁶

The appeal of a system like this is that it doesn’t require visual contact with the tornado. A chaser could be miles away, in the rain, at night, and the infrasound signature might still register. Night tornadoes are disproportionately deadly because they’re hard to see, so any detection method that doesn’t rely on eyesight is potentially valuable for both science and public safety.

Machine Learning and the Future of Targeting

Forecasting where severe weather will develop is the foundation of storm chasing, and machine learning is starting to reshape that process. Traditional forecasting relies on a meteorologist interpreting atmospheric models, radar imagery, and satellite data. That works well enough for identifying broad areas of risk, but pinpointing which specific storms will produce tornadoes or large hail has always been difficult.

Recent work has shown that combining multiple data streams through machine learning algorithms can substantially improve prediction accuracy. A study using ten years of data from Poland fused radar reflectivity, lightning detection data, and atmospheric parameters from a weather reanalysis dataset, then ran them through machine learning models. The result was a large-hail prediction system that significantly outperformed any single product used in standard operational forecasting.⁷ The approach isn’t limited to hail. The same data-fusion philosophy applies to tornado prediction, flash-flood forecasting, and identifying which storms are most likely to produce damaging straight-line winds.

For storm chasers, these advances could eventually mean better real-time guidance about which storms to target. Right now, a chaser might be watching three separate supercells on radar and making a gut call about which one to approach. If machine learning tools can flag the storm most likely to produce a tornado based on its radar signature, lightning characteristics, and the surrounding atmospheric environment, chasers waste less time on duds and collect more useful data from significant events.

The Risks of Getting Close

Storm chasing is dangerous, and not only in the obvious way. Tornadoes themselves account for some chaser fatalities, including several experienced researchers and chasers who were killed by the El Reno, Oklahoma tornado in 2013, an event that was unusually wide, erratic in its path, and embedded in rain. But the more common risks are mundane: car accidents on wet or hail-covered roads, hydroplaning, poor visibility, lightning, and collisions with other vehicles on narrow rural roads. Driving hundreds of miles under time pressure, often while monitoring multiple screens, is inherently risky even without a tornado at the destination.

Hail is an underappreciated hazard. Supercell thunderstorms can produce hailstones larger than softballs, and a chaser who misjudges the storm’s motion can find themselves under a barrage that shatters windshields, dents hoods, and can injure anyone caught outside. Many experienced chasers have driven through hail that totaled their vehicles.

Flash flooding kills more storm chasers’ cars than tornadoes do. Roads in the Great Plains can go from dry to impassable in minutes during heavy rain, and rural drainage is often poor. A chaser retreating from a storm may find their escape route underwater.

How Storm Chasers Feed Back Into Public Safety

Storm chasers play a direct role in the National Weather Service warning process. The NWS issues tornado warnings based on radar signatures, but radar can’t always distinguish a tornado from a non-tornadic rotation. Trained spotters and chasers who report what they’re seeing on the ground, whether a tornado is actually on the ground, how large it is, what direction it’s moving, provide confirmation that can make the difference between a warning that sounds urgent and one that sounds tentative. Many chasers are registered as storm spotters through the NWS Skywarn program.

Beyond individual reports, the scientific data collected by research chasers feeds into long-term improvements to weather models and warning systems. Every mobile radar deployment, every drone flight into a supercell, every probe that captures a pressure trace from inside a tornado adds to the dataset that modelers use to improve how computer simulations represent severe storms. The improvements are slow and incremental, but the cumulative effect over decades has been a steady increase in tornado warning lead times. In the 1980s, the average lead time for a tornado warning was roughly five minutes. Today it’s closer to fifteen, and the science behind that improvement rests partly on data that storm chasers gathered in the field.

Storm Chasing Outside Tornado Alley

The stereotype places storm chasers exclusively on the flat farmland between Texas and Nebraska, but the practice extends well beyond that geography. The RELAMPAGO campaign mentioned earlier took mobile radars and chasers to Argentina, where the terrain and storm environments are different from the U.S. Plains but equally capable of producing violent supercells.¹ Chasers also operate in the southeastern United States, where storms are harder to see because of tree cover and often move faster than their Plains counterparts. The Southeast has a higher rate of nocturnal tornadoes, which makes chasing there both more difficult and arguably more important for safety.

In Europe, severe convective storms produce large hail, damaging winds, and occasional tornadoes across a belt from France through Germany, Poland, and into the Balkans. European storm chasers face different challenges: road networks are denser but terrain is more varied, radar coverage has historically been less comprehensive than in the United States, and the storms tend to be somewhat smaller and shorter-lived. Still, the same basic approach applies. Forecast, target, position, observe.

Tropical cyclones represent another branch of storm chasing entirely. Hurricane hunters fly aircraft into the eyes of hurricanes to collect data, and ground-based chasers position themselves near landfall to record wind speeds and storm surge. The instruments are different, the scale is different, and the risks skew more toward storm surge and widespread flooding than toward a single vortex, but the fundamental motivation is the same: get close enough to collect data that remote sensing alone can’t provide.

What Storm Chasing Cannot Do

For all its value, storm chasing has real limitations. Chasers are constrained by road networks. A tornado in a remote area with no paved roads is effectively unreachable. This means that the data collected by chasers is biased toward storms that happen near accessible roads, which in the United States means the Great Plains get disproportionate coverage while storms in the Appalachians or Deep South are undersampled.

Chaser data is also inherently opportunistic. A research team might spend an entire spring field campaign and never intercept a significant tornado. The atmosphere doesn’t cooperate on schedule, and even when it does, being in exactly the right place at the right time is partly luck. This is why the shift toward autonomous instruments, drones, and sensor networks matters: it reduces the dependence on having a human in a truck at the precise spot where the tornado touches down.

There’s also a tension between the growing number of chasers on the road and the safety of the public and the chasers themselves. On major tornado outbreak days, dozens or even hundreds of chase vehicles converge on the same rural roads. Emergency responders have raised concerns about chasers blocking evacuation routes or interfering with rescue operations. Most organized research teams coordinate with local law enforcement, but freelance and recreational chasers operate independently, and there’s no licensing requirement or formal regulation governing who can chase storms. The community has debated self-regulation for years without reaching consensus on what that would look like.