In early July, as Hurricane Beryl churned through the Caribbean, a top European weather agency predicted a range of final landfalls, warning that that Mexico was most likely. The alert was based on global observations by planes, buoys and spacecraft, which room-size supercomputers then turned into forecasts.

That same day, experts running artificial intelligence software on a much smaller computer predicted landfall in Texas. The forecast drew on nothing more than what the machine had previously learned about the planet’s atmosphere.

Four days later, on July 8, Hurricane Beryl slammed into Texas with deadly force, flooding roads, killing at least 36 people and knocking out power for millions of residents. In Houston, the violent winds sent trees slamming into homes, crushing at least two of the victims to death.

The Texas prediction offers a glimpse into the emerging world of A.I. weather forecasting, in which a growing number of smart machines are anticipating future global weather patterns with new speed and accuracy. In this case, the experimental program was GraphCast, created in London by DeepMind, a Google company. It does in minutes and seconds what once took hours.

“This is a really exciting step,” said Matthew Chantry, an A.I. specialist at the European Center for Medium-Range Weather Forecasts, the agency that got upstaged on its Beryl forecast. On average, he added, GraphCast and its smart cousins can outperform his agency in predicting hurricane paths.

In general, superfast A.I. can shine at spotting dangers to come, said Christopher S. Bretherton, an emeritus professor of atmospheric sciences at the University of Washington. For treacherous heats, winds and downpours, he said, the usual warnings will be “more up-to-date than right now,” saving untold lives.

Rapid A.I. weather forecasts will also aid scientific discovery, said Amy McGovern, a professor of meteorology and computer science at the University of Oklahoma who directs an A.I. weather institute. She said weather sleuths now use A.I. to create thousands of subtle forecast variations that let them find unexpected factors that can drive such extreme events as tornadoes.

“It’s letting us look for fundamental processes,” Dr. McGovern said. “It’s a valuable tool to discover new things.”

Importantly, the A.I. models can run on desktop computers, making the technology much easier to adopt than the room-size supercomputers that now rule the world of global forecasting.

“It’s a turning point,” said Maria Molina, a research meteorologist at the University of Maryland who studies A.I. programs for extreme-event prediction. “You don’t need a supercomputer to generate a forecast. You can do it on your laptop, which makes the science more accessible.”

People depend on accurate weather forecasts to make decisions about such things as how to dress, where to travel and whether to flee a violent storm.

Even so, reliable weather forecasts turn out to be extraordinarily hard to achieve. The trouble is complexity. Astronomers can predict the paths of the solar system’s planets for centuries to come because a single factor dominates their movements — the sun and its immense gravitational pull.

In contrast, the weather patterns on Earth arise from a riot of factors. The tilts, the spins, the wobbles and the day-night cycles of the planet turn the atmosphere into turbulent whorls of winds, rains, clouds, temperatures and air pressures. Worse, the atmosphere is inherently chaotic. On its own, with no external stimulus, a particular zone can go quickly from stable to capricious.

As a result, weather forecasts can fail after a few days, and sometimes after a few hours. The errors grow in step with the length of the prediction — which today can extend for 10 days, up from three days a few decades ago. The slow improvements stem from upgrades to the global observations as well as the supercomputers that make the predictions.

Not that supercomputing work has grown easy. The preparations take skill and toil. Modelers build a virtual planet crisscrossed by millions of data voids and fill the empty spaces with current weather observations.

Dr. Bretherton of the University of Washington called these inputs crucial and somewhat improvisational. “You have to blend data from many sources into a guess at what the atmosphere is doing right now,” he said.

The knotty equations of fluid mechanics then turn the blended observations into predictions. Despite the enormous power of supercomputers, the number crunching can take an hour or more. And of course, as the weather changes, the forecasts must be updated.

The A.I. approach is radically different. Instead of relying on current readings and millions of calculations, an A.I. agent draws on what it has learned about the cause-and-effect relationships that govern the planet’s weather.

In general, the advance derives from the ongoing revolution in machine learning — the branch of A.I. that mimics how humans learn. The method works with great success because A.I. excels at pattern recognition. It can rapidly sort through mountains of information and spot intricacies that humans cannot discern. Doing so has led to breakthroughs in speech recognition, drug discovery, computer vision and cancer detection.

In weather forecasting, A.I. learns about atmospheric forces by scanning repositories of real-world observations. It then identifies the subtle patterns and uses that knowledge to predict the weather, doing so with remarkable speed and accuracy.

Recently, the DeepMind team that built GraphCast won Britain’s top engineering prize, presented by the Royal Academy of Engineering. Sir Richard Friend, a physicist at Cambridge University who led the judging panel, praised the team for what he called “a revolutionary advance.”

In an interview, Rémi Lam, GraphCast’s lead scientist, said his team had trained the A.I. program on four decades of global weather observations compiled by the European forecasting center. “It learns directly from historical data,” he said. In seconds, he added, GraphCast can produce a 10-day forecast that would take a supercomputer more than an hour.

Dr. Lam said GraphCast ran best and fastest on computers designed for A.I., but could also work on desktops and even laptops, though more slowly.

In a series of tests, Dr. Lam reported, GraphCast outperformed the best forecasting model of the European Center for Medium-Range Weather Forecasts more than 90 percent of the time. “If you know where a cyclone is going, that’s quite important,” he added. “It’s important for saving lives.”

Replying to a question, Dr. Lam said he and his team were computer scientists, not cyclone experts, and had not evaluated how GraphCast’s predictions for Hurricane Beryl compared to other forecasts in precision.

But DeepMind, he added, did conduct a study of Hurricane Lee, an Atlantic storm that in September was seen as possibly threatening New England or, farther east, Canada. Dr. Lam said the study found that GraphCast locked in on landfall in Nova Scotia three days before the supercomputers reached the same conclusion.

Impressed by such accomplishments, the European center recently embraced GraphCast as well as A.I. forecasting programs made by Nvidia, Huawei and Fudan University in China. On its website, it now displays global maps of its A.I. testing, including the range of path forecasts that the smart machines made for Hurricane Beryl on July 4.

The track predicted by DeepMind’s GraphCast, labeled DMGC on the July 4 map, shows Beryl making landfall in the region of Corpus Christi, Texas, not far from where the hurricane actually hit.

Dr. Chantry of the European center said the institution saw the experimental technology as becoming a regular part of global weather forecasting, including for cyclones. A new team, he added, is now building on “the great work” of the experimentalists to create an operational A.I. system for the agency.

Its adoption, Dr. Chantry said, could happen soon. He added, however, that the A.I. technology as a regular tool might coexist with the center’s legacy forecasting system.

Dr. Bretherton, now a team leader at the Allen Institute for A.I. (established by Paul G. Allen, one of the founders of Microsoft), said the European center was considered the world’s top weather agency because comparative tests have regularly shown its forecasts to exceed all others in accuracy. As a result, he added, its interest in A.I. has the world of meteorologists “looking at this and saying, ‘Hey, we’ve got to match this.’”

Weather experts say the A.I. systems are likely to complement the supercomputer approach because each method has its own particular strengths.

“All models are wrong to some extent,” Dr. Molina of the University of Maryland said. The A.I. machines, she added, “might get the hurricane track right but what about rain, maximum winds and storm surge? There’re so many diverse impacts” that need to be forecast reliably and assessed carefully.

Even so, Dr. Molina noted that A.I. scientists were rushing to post papers that demonstrate new forecasting skills. “The revolution is continuing,” she said. “It’s wild.”

Jamie Rhome, deputy director of the National Hurricane Center in Miami, agreed on the need for multiple tools. He called A.I. “evolutionary rather than revolutionary” and predicted that humans and supercomputers would continue to play major roles.

“Having a human at the table to apply situational awareness is one of the reasons we have such good accuracy,” he said.

Mr. Rhome added that the hurricane center had used aspects of artificial intelligence in its forecasts for more than a decade, and that the agency would evaluate and possibly draw on the brainy new programs.

“With A.I. coming on so quickly, many people see the human role as diminishing,” Mr. Rhome added. “But our forecasters are making big contributions. There’s still very much a strong human role.”

Exposure to wildfire smoke increases the odds of being diagnosed with dementia even more than exposure to other forms of air pollution, according to a landmark study of more than 1.2 million Californians. The study — released Monday at the Alzheimer’s Assn. International Conference in Philadelphia — is the largest and most comprehensive review of the impact of wildfire smoke on brain health to date, according to its authors.

“I was expecting for us to see an association between wildfire smoke exposure and dementia,” said study author Dr. Holly Elser, an epidemiologist and resident physician in neurology at the University of Pennsylvania. “But the fact we see so much stronger of an association for wildfire as compared to non-wildfire smoke exposure was kind of surprising.”

The findings have big health implications, particularly in Western states, where air pollution produced by wildfires now accounts for up to half of all fine-particle pollution — a figure that’s been trending upward as wildfires grow larger and more intense due to climate change and legacies of fire suppression and industrial logging that have altered the composition of many Western forests.

The researchers looked at a type of particulate-matter pollution called PM2.5. These particles are 30 times smaller than the width of a human hair — tiny enough to penetrate deeply into the lungs and cross over into the bloodstream, where they can cause inflammation. Exposure has been shown to raise the risk of dementia and a host of other conditions, including heart disease, asthma and low birth weight.

“We increasingly see that PM2.5 is tied to virtually every health outcome we look at,” said study author Joan Casey, associate professor of public health at the University of Washington.

Elser, Casey and fellow researchers analyzed the health records of more than 1.2 million Kaiser Permanente Southern California members 60 or older between 2009 and 2019. None had been diagnosed with dementia at the beginning of the study.

They estimated each person’s exposure to PM2.5 based on their census tract of residence and then separated that into wildfire and non-wildfire pollution using air quality monitoring data, satellite imagery and machine learning techniques.

They then looked at how many participants were eventually diagnosed with dementia. Unlike past studies, the researchers were able to determine this using patients’ full electronic health records, rather than relying on hospitalizations as a proxy for such diagnoses.

Looking at participants’ average wildfire PM2.5 exposure over three years, the researchers found a 23% increase in the odds of a dementia diagnosis for each increase of 1 microgram of particulate matter per cubic meter of air. When it came to non-wildfire PM2.5 exposure, they documented a 3% increased risk of dementia diagnoses for each increase of 3 micrograms of particulate matter per cubic meter of air.

“That’s what it comes down to, is what’s so different about wildfire smoke?” Casey said.

More research is needed to learn exactly what that is. Possibilities include the fact that wildfire particles are produced at higher temperatures, contain a greater concentration of toxic chemicals and are, on average, smaller than PM2.5 from other sources.

These ultrafine particles can translocate from people’s noses into their brains via the olfactory bulb, Casey said.

“Normally the brain is protected by the blood-brain barrier, but here there’s actually a direct route for ultrafine particles to get into the brain and possibly cause some of the problems that we’re seeing in folks living with dementia,” she said.

The way in which people are exposed to wildfire smoke also differs from other types of fine-particle pollution, the researchers said. Background or ambient fine-particle pollution levels are usually relatively constant in a given place over time. But wildfire particulate matter tends to fluctuate wildly, resulting in more exposure over shorter periods of time, which may overwhelm the body’s defenses.

Of some 5,500 abstracts submitted to the Alzheimer’s Assn. International Conference, this one stood out due to its novelty, importance and impact, said Dr. Claire Sexton, senior director of scientific programs and outreach for the Alzheimer’s Assn.

“There have been other studies looking at different types of pollution, but this was unique in terms of the extent and the way in which it was able to do these analyses,” she said.

The researchers found the effects to be stronger on Asian, Black and Latino people, as well as those living in high-poverty areas. The most heavily impacted group was one that researchers classified as “other” because it didn’t contain enough people to differentiate further, Casey said. That group included Indigenous people, Pacific Islanders and people whose race was unknown.

“So these disparities are playing out again, as we unfortunately often see with environmental exposures,” she said. “But the level at which we observed it here was fairly striking.”

Casey believes those disparities are due to differential exposure based on where populations are located, noting that her previous research has shown that Indigenous people in California have by far the highest levels of wildfire particulate exposure. Other factors could include poorer housing quality, lack of access to air filtration devices, jobs that prevent people from staying indoors during wildfire events and disparate responses to the same amount of pollution due to preexisting hypertension or diabetes, she said.

“All those things are driven by social determinants of health,” she said. “The fact that we need to allocate additional resources to these people and places to protect health and to try to reduce health disparities going forward is really important.”

The researchers did not differentiate between dementia subtypes like Alzheimer’s, the most common form, because they relied on diagnostic codes rather than using brain imaging or postmortem studies. That’s important to know — and a key area for future study — because in order to best protect people, clinicians need to have an understanding of what’s underpinning the relationship between wildfire smoke and different dementia subtypes, Elser said.

Still, the study is notable for its massive sample size and careful approach, taking into account sociodemographics like comorbidities and census tract poverty, said Rachel Whitmer, the director of the UC Davis Alzheimer’s Disease Research Center who was not involved in the research.

The prevalence of dementia is on the rise as the baby boomer generation ages, but environmental factors may also be contributing to the increase, she said.

Research like this lays the groundwork for future studies, she said.

“With the increase in wildfires, this is a really important question and I think they did a really rigorous job in exploring it,” she said.

Levels of PM2.5 had been declining since the Clean Air Act took effect in 1970. But wildfires have reversed those trends in California, undercutting efforts to reduce emissions. In recent years, wildfire smoke has also affected the Midwest and East Coast. In 2023, smoke from Canadian wildfires blanketed the Atlantic seaboard, triggering air quality alerts and forcing the cancellation of outdoor events.

“It’s a big problem in places where wildfires are endemic,” Elser said. “And I worry that as we continue to experience increasingly frequent wildfire events, this could affect more people over a larger geographical distribution, more of the time.”

The most decorated gymnast ever sprints down the vault runway. She tumbles gracefully onto the springboard, flings herself backward onto the vault table and pops off the surface. Soaring through the air, she folds her body in half and grabs the back of her legs for two head-over-heels flips.

The crowd erupts when Simone Biles stomps her feet into the mat.

Another successful Yurchenko double pike.

“It just makes your mouth drop open every time,” said UCLA gymnastics coach Janelle McDonald, who sat in the front row next to the vault landing area at the U.S. Olympic trials, where Biles competed her signature vault.

Of the five skills in the International Gymnastics Federation (FIG) code of points named after Biles, her most recent vault — the Yurchenko double pike — has become the most iconic.

It’s the first major leap in vault innovation for women’s artistic gymnastics in two decades. When leveling up the sky-high event used to mean extra twists, Biles flipped the game upside down. She was the first woman to attempt her double-flipping skill in competition and completed it in international competition for the first time at the 2023 World Championships, earning its name as the Biles II.

“Simone made impossible an opinion with that vault,” NBC analyst John Roethlisberger said on the telecast during the U.S. Olympic trials.

In a sport that blends power and grace, Biles’ Yurchenko double pike is at the center of its own Venn diagram: athletic feat, scientific marvel and artistic genius all in six seconds.

The entry

The height, the rotation and the landing earn the most gasps from the fans who follow Biles’ every move on the competition floor. But for her peers who continue to marvel at the 27-year-old’s revolutionary talent, the most impressive part of the vault happens before Biles even contacts the table.

“Your Yurchenko entry has to be so technically perfect and so consistent,” said 2016 Olympic gold medalist Kyla Ross. “You have no doubt coming off the table that you’re going to hit the double pike.”

The Yurchenko entry — a roundoff onto the springboard and a back handspring onto the vault table — is named after former Russian gymnast Natalia Yurchenko, who debuted her eponymous skill in 1982. Since the FIG replaced the vault horse — which resembled a pommel horse without handles and was about 5 feet long and 1 foot wide at the top — for a tongue-shaped vault table in 2001, the Yurchenko vault has become more common for elite female gymnasts. Athletes can still harness the power generated from the unique entry while having a larger, safer surface area for their hands.

2024 Paris Summer Olympic Games

Advertisement

Biles begins with a sprint down the runway and reaches her hands toward the ground while cartwheeling her legs over her head. The roundoff turns her momentum forward to backward. Slamming her feet down on the springboard, Biles compresses the springs that then uncoil and transfer energy back into her body as she reaches up and backward for the vault table.

The key to vault liftoff is how Biles contacts the equipment to transfer her momentum.

“Pre-springboard, all of their motion is forwards,” said Emily Kuhn, a physics PhD student at Yale who was a level 10 gymnast. “After the springboard, some of their motion is upwards. And so [the board is] really helpful for converting the rotational energy from that roundoff into an upwards velocity that is used to get the height on the vault.”

In an instant, Biles arches backward toward the vault table. Her body whips back in a lightning-fast handspring that leaves even the best athletes in the world in the dust.

“She gets that power because of how quick-twitched her roundoff-back handspring is, technically speaking,” said 2008 Olympic silver medalist Samantha Peszek. “No one is as quick-twitch as Simone.”

The block

During the block, the moment Biles’ hands strike the table, she extends through her shoulders in a motion that’s barely detectable in real time. The micro-movement lasts a tenth of a second as Biles applies force to the vault that is then returned in equal and opposite measure.

(Kenzo Tribouillard / AFP via Getty Images)

Aly Raisman had a front-row seat to the world’s best vaulters. The two-time Olympian watched 2012 Olympic teammate McKayla Maroney nail a nearly perfect two-and-a-half twisting Yurchenko in the team final. Raisman was also the team captain in 2016 when Biles won Olympic gold on vault. Both standouts shared a key ability that helped them soar above the competition.

“When you look at her elbows on the table, they’re always very straight,” Raisman said. “Their body was so tight when their arms hit the table that it just helps them get so much air.”

The moment Biles’ hands strike the table, she extends through her shoulders in a motion that’s barely detectable in real time. The micro-movement lasts a tenth of a second as Biles applies force to the vault that is then returned in equal and opposite measure. Keeping every muscle contracted during the split second on the table is vital to transferring energy efficiently for maximum height.

“The force gets dispersed in bad form,” said Gina Pongetti, a physical therapist with more than 20 years of experience working with college, national team and elite gymnasts. “[The muscles] are all tight at one time so that nothing gives, nothing buckles. Because of that, all of that force, or as much as possible, goes into the vault [and] goes back to her to transfer to height and rotation.”

NBC estimated that Biles’ feet reach about 12 feet in the air at the peak of her vault when she is upside down.

Former UCLA gymnast Nia Dennis knows the feeling. The three-time U.S. national team member trained a Yurchenko double tuck — with her knees bent and legs pulled toward her chest — into the foam pit during her elite career, eventually stacking up soft mats to be equal to competition height. While she is best known for her viral, energetic floor routines, Dennis loved vault the most. She still recalls the intoxicating feeling of hitting the perfect block that fired her into the air.

“I just felt like a cannon,” Dennis said.

The flips

After Biles blocks off the vault, she is just a projectile.

(Matthias Hangst / Getty Images)

Searching for ways to upgrade her vault difficulty, Dennis wanted to buck the trend of adding additional twists to her Yurchenko. She was always more of a flipper than a twister, and Dennis sometimes landed on her neck from over-rotating her warm-up drills. One day, her coach encouraged her to pull all the way around onto her feet for an additional flip.

“It was just straight power,” Dennis said. “All I had to do was run and close my eyes, for real. Just block really hard, close my eyes really hard and pull really hard.”

For Dennis, the daring skill was fun. Her former UCLA teammate Ross did not share the sentiment.

“I used to like, cry because I was so scared of it,” laughed Ross, who practiced Yurchenko double tucks into the foam pit alongside her longtime club teammate Maroney.

The second flip is what makes the vault so frightening for athletes. Gymnasts can adjust a twisting vault midair by reducing the number of revolutions by halves if they feel something gone awry. There is no safe Plan B between one and two flips.

After Biles blocks off the vault, she is just a projectile, Kuhn emphasized. At that point, there is nothing she can do to change how high she is or her path through the air.

That’s when her proprioception takes over. Biles’ air awareness is “unbelievable,” Pongetti said. The physical therapist, who specializes in treatment, diagnoses and biomechanics in gymnastics, has watched Biles train for years.

While she cannot change her flight path, Biles is an expert in making split-second decisions to rearrange her body midair to change how she will move. If she is too low, she can pull her body into a tighter shape to flip faster. If she is too high, she knows a more open shape can slow down her flip and help avoid over-rotation.

Only that caliber of spatial awareness can make the death-defying skill “safer,” Pongetti said. She would never say it’s “safe.”

“That sets apart the level 10 [gymnast] from the elite,” Pongetti said, “from the Olympian from Simone.”

The landing

Simone Biles of the United States reacts after performing a vault during the 2023 gymnastics world championships in Antwerp, Belgium.

(Tim Clayton / Corbis via Getty Images)

With five skills named for her in the code of points, Biles is at the forefront of the sport’s progression. Peszek remembers when the double-twisting, double back tucks she and 2008 Olympic teammate Shawn Johnson competed were arguably the hardest tumbling passes in the world. Now Biles casually does that skill in combination with a full-twisting front layout.

“It’s really special to see the generations pass the torch and just how they’ve been able to take this sport by storm by creating all these new elements and really pushing the boundaries,” Peszek said. “Seeing her do it so effortlessly, it’s really a work of art to see.”

The Biles II was awarded the highest-difficulty score of any vault by the FIG at 6.4. Astronomical-difficulty scores, which are combined on each event with an execution score out of 10, allow Biles to win competitions by whole numbers as most of her peers fight for half-tenths.

On vault, most of the top medal contenders have difficulty values of 5.4 for the two-and-a-half twisting Yurchenko known as an Amanar or 5.6 for a Cheng, which begins with a roundoff onto the springboard, a half turn onto the vault table and a one-and-a-half twist off.

But Rebeca Andrade could challenge Biles for the crown. The Brazilian Olympic confederation published a YouTube video that featured the Olympic and world vault champion training a triple-twisting Yurchenko. If she lands it in international competition, it will bear her name.

Biles considered the skill as the next Yurchenko progression from an Amanar but has said going for the double flip was safer for her to land. The landing on a twisting flip presents additional challenges, Kuhn said, as gymnasts must absorb rotational forces to stop the twist while also controlling the landing vertically.

A 2013 study estimated that gymnasts absorb 11 times their body weight on landings during competition. The force increases to 18 times body weight if a landing is uneven, a common consequence of twisting elements. What Biles feels when her 4-foot-8 frame is falling from more than two and a half times her height to land the Yurchenko double pike might be even greater, Pongetti said.

“Her quads and her glutes and her hamstrings [and her calves], which otherwise would work to allow her to jump high, work in reverse to slow her down,” Pongetti said. “They are her brakes. … She is so good at not being stiff-legged when she lands.”

Fans seem to hold their breath as Biles floats and flips through the air. At the moment her feet punch into the mat, the crowd exhales with a roar of applause.

Thousands in Paris’ Bercy Arena will see the vault’s official Olympic debut. Millions more will watch internationally.

They’ll see Biles push the boundaries of sport and science in a gravity-defying six-second burst.