Accurate forecasts, finally? — As a potentially historic hurricane season looms, can AI forecast models help? New model that predicts global weather can run on a single desktop computer.
Eric Berger – Jun 3, 2024 11:00 am UTC Enlarge / AI weather models are arriving just in time for the 2024 Atlantic hurricane season.Aurich Lawson | Getty Images reader comments 0
Much like the invigorating passage of a strong cold front, major changes are afoot in the weather forecasting community. And the end game is nothing short of revolutionary: an entirely new way to forecast weather based on artificial intelligence that can run on a desktop computer.
Today’s artificial intelligence systems require one resource more than any other to operatedata. For example, large language models such as ChatGPT voraciously consume data to improve answers to queries. The more and higher quality data, the better their training, and the sharper the results.
However, there is a finite limit to quality data, even on the Internet. These large language models have hoovered up so much data that they’re being sued widely for copyright infringement. And as they’re running out of data, the operators of these AI models are turning to ideas such as synthetic data to keep feeding the beast and produce ever more capable results for users.
If data is king, what about other applications for AI technology similar to large language models? Are there untapped pools of data? One of the most promising that has emerged in the last 18 months is weather forecasting, and recent advances have sent shockwaves through the field of meteorology.
That’s because there’s a secret weapon: an extremely rich data set. The European Centre for Medium-Range Weather Forecasts, the premiere organization in the world for numerical weather prediction, maintains a set of data about atmospheric, land, and oceanic weather data for every day, at points around the world, every few hours, going back to 1940. The last 50 years of data, after the advent of global satellite coverage, is especially rich. This dataset is known as ERA5, and it is publicly available. Advertisement
It was not created to fuel AI applications, but ERA5 has turned out to be incredibly useful for this purpose. Computer scientists only really got serious about using this data to train AI models to forecast the weather in 2022. Since then, the technology has made rapid strides. In some cases, the output of these models is already superior to global weather models that scientists have labored decades to design and build, and they require some of the most powerful supercomputers in the world to run.
“It is clear that machine learning is a significant part of the future of weather forecasting,” said Matthew Chantry, who leads AI forecasting efforts at the European weather center known as ECMWF, in an interview with Ars. Its moving fast
John Dean and Kai Marshland met as undergraduates at Stanford University in the late 2010s. Dean, an electrical engineer, interned at SpaceX during the summer of 2017. Marshland, a computer scientist, interned at the launch company the next summer. Both graduated in 2019 and were trying to figure out what to do with their lives.
“We decided we wanted to solve the problem of weather uncertainty,” Marshland said, so they co-founded a company called WindBorne Systems.
The premise of the company was simple: For about 85 percent of the Earth and its atmosphere, we have no good data about weather conditions there. A lack of quality data, which establishes initial conditions, represents a major handicap for global weather forecast models. The company’s proposed solution was in its namewind borne.
Dean and Marshland set about designing small weather balloons they could release into the atmosphere and which would fly around the world for up to 40 days, relaying useful atmospheric data that could be packaged and sold to large, government-funded weather models.
Weather balloons provide invaluable data about atmospheric conditionsreadings such as temperature, dewpoints, and pressuresthat cannot be captured by surface observations or satellites. Such atmospheric “profiles” are helpful in setting the initial conditions models start with. The problem is that traditional weather balloons are cumbersome and only operate for a few hours. Because of this, the National Weather Service only launches them twice daily from about 100 locations in the United States. Page: 1 2 3 4 Next → reader comments 0 Eric Berger Eric Berger is the senior space editor at Ars Technica, covering everything from astronomy to private space to wonky NASA policy, and author of the book Liftoff, about the rise of SpaceX. A certified meteorologist, Eric lives in Houston. Advertisement Channel Ars Technica ← Previous story Related Stories Today on Ars
Jesse joined ESPN Chicago in September 2009 and covers MLB for ESPN.com.
The last-place Washington Nationals fired president of baseball operations Mike Rizzo and manager Davey Martinez, the team announced Sunday.
Rizzo, 64, and Martinez, 60, won a World Series with the Nationals in 2019, but the team has floundered in recent years. This season, the Nationals are 37-53 and stuck at the bottom of the National League East after getting swept by the Boston Red Sox this weekend at home. Washington hasn’t finished higher than fourth in the division since winning the World Series.
“On behalf of our family and the Washington Nationals organization, I first and foremost want to thank Mike and Davey for their contributions to our franchise and our city,” principal owner Mark Lerner said in a statement. “Our family is eternally grateful for their years of dedication to the organization, including their roles in bringing a World Series trophy to Washington, D.C.
“While we are appreciative of their past successes, the on-field performance has not been where we or our fans expect it to be. This is a pivotal time for our club, and we believe a fresh approach and new energy is the best course of action for our team moving forward.”
Mike DeBartolo, the club’s senior vice president and assistant general manager, was named interim GM on Sunday night. DeBartolo will oversee all aspects of baseball operations, including the MLB draft. An announcement will be made on the interim manager Monday, a day before the club begins a series against the St. Louis Cardinals.
Rizzo has been the top decision-maker in Washington since 2013, and Martinez has been on board since 2018. Under Rizzo’s leadership, the team made the postseason four times: in 2014, 2016, 2017 and 2019. The latter season was Martinez’s lone playoff appearance.
“When our family assumed control of the team, nearly 20 years ago, Mike was the first hire we made,” Lerner said. “Over two decades, he was with us as we went from a fledging team in a new city to World Series champion. Mike helped make us who we are as an organization, and we’re so thankful to him for his hard work and dedication — not just on the field and in the front office, but in the community as well.”
The Nationals are in the midst of a rebuild that has moved slower than expected, though the team didn’t augment its young core much during the winter. Led by All-Stars James Wood and MacKenzie Gore, Washington has the second-youngest group of hitters in MLB and the sixth-youngest pitching staff.
The team lost 11 straight games in a forgettable stretch last month. And during a 2-10 run in June, Washington averaged just 2.5 runs. Since June 1, the Nationals have scored one run or been shut out seven times. In Sunday’s 6-4 loss to Boston, they left 15 runners on base.
There was industry speculation over the winter that the Nationals would spend money on free agents for the first time in several years, but that never materialized. Instead, the team made minor moves, signing free agents Josh Bell and Michael Soroka, trading for first baseman Nathaniel Lowe and re-signing closer Kyle Finnegan. Now, the hope is a new management team, both on and off the field, can help change the franchise’s fortunes.
The rosters for the 2025 MLB All-Star Game will feature 19 first-timers — and one legend — as the pitchers and reserves were announced Sunday for the July 15 contest at Truist Park in Atlanta.
Los Angeles Dodgers left-hander Clayton Kershaw, a three-time Cy Young Award winner who made his first All-Star team in 2011, was named to his 11th National League roster as a special commissioner’s selection.
Kershaw, who became only the fourth left-hander to amass 3,000 career strikeouts, is 4-0 with a 3.43 ERA in nine starts after beginning the season on the injured list. He joins Albert Pujols and Miguel Cabrera as a legend choice, after the pair of sluggers were selected in 2022.
Kershaw said he didn’t want to discuss the selection Sunday.
Overall, the 19 first-time All-Stars is a drop from the 32 first-time selections on the initial rosters in 2024.
Kershaw would be the sentimental choice to start for the National League, although Pittsburgh Pirates ace Paul Skenes, who leads NL pitchers in ERA and WAR, might be in line to start his second straight contest. Philadelphia Phillies right-hander Zack Wheeler, a three-time All-Star, is 9-3 with a 2.17 ERA after Sunday’s complete-game victory and also would be a strong candidate to start.
“I think it would be stupid to say no to that. It’s a pretty cool opportunity,” Skenes said about the possibility of being asked to start by Dodgers manager Dave Roberts. “I didn’t make plans over the All-Star break or anything. So, yeah, I’m super stoked.”
Kershaw has made one All-Star start in his career, in 2022 at Dodger Stadium.
Among standout players not selected were New York Mets outfielder Juan Soto, who signed a $765 million contract as a free agent in the offseason, and Dodgers shortstop Mookie Betts, who had made eight consecutive All-Star rosters since 2016.
Soto got off to a slow start but was the National League Player of the Month in June and entered Sunday ranked sixth in the NL in WAR among position players while ranking second in OBP, eighth in OPS and third in runs scored.
Earning his fifth career selection but first since 2021 is Texas Rangers righty Jacob deGrom, who is finally healthy after making only nine starts in his first two seasons with the Rangers and is 9-2 with a 2.13 ERA. He has never started an All-Star Game, although Skubal or Brown would be the favorite to start for the AL.
“Red carpet, that’s my thing,” Chisholm said. “I do have a ‘fit in mind.”
Rosters are expanded from 26 to 32 for the All-Star Game. They include starters elected by fans, 17 players (five starting pitchers, three relievers and a backup for each position) chosen in a player vote and six players (four pitchers and two position players) selected by league officials. Every club must be represented.
Acuna, Wood and Raleigh are the three All-Stars who have so far committed to participating in the Home Run Derby.
ESPN baseball reporter. Covered the Washington Wizards from 2014 to 2016 and the Washington Nationals from 2016 to 2018 for The Washington Post before covering the Los Angeles Dodgers and MLB for the Los Angeles Times from 2018 to 2024.
NEW YORK — The New York Yankees were seemingly in deep trouble Sunday when Juan Soto cracked a pitch to left field in the seventh inning.
The New York Mets, down two runs, were cooking up a rally with no outs. Francisco Lindor stood at first base, Pete Alonso loomed on deck, and Brandon Nimmo was in the hole. This was the heart of the Mets’ potent lineup. Given the Yankees’ recent woes, fumbling their two-run lead and suffering a Subway Series sweep at the hands of their neighbors — and a seventh straight loss — seemed almost fated.
Then Cody Bellinger charged Soto’s sinking 105 mph line drive, made a shoestring catch and fired a strike to first base for an improbable double play to secure a skid-snapping 6-4 win — and perhaps rescue the Yankees from another dreadful outcome.
“Considering the context of this week and everything,” Yankees manager Aaron Boone said, “that’s probably our play of the year so far.”
Soto’s line drive off Mark Leiter Jr. had a 10% catch probability, according to Statcast, but Bellinger, a plus defender at multiple positions who started at first base Saturday, was just able to snatch it before it touched the grass. Certain that he caught it clean, he made an 89.9 mph toss that reached first baseman Paul Goldschmidt on a line, over Lindor, who didn’t slide into the bag.
“I saw it in the air and had a really good beat on it,” said Bellinger, who went 2-for-3 with a double and a walk at the plate.
The Mets challenged the catch, but the call stood.
“That was incredible,” said Yankees right fielder Aaron Judge, who swatted his 33rd home run of the season in the fifth inning. “I’ve never seen something like that on the field.”
For the past week, a stretch Boone described as “terrible” for his ballclub, poor defense has been an issue for the Yankees. Physical errors. Mental lapses. Near disasters. The sloppiness helped sink a depleted pitching staff, more than offsetting the offense’s strong production.
That combination produced the team’s second six-game losing streak in three weeks and a three-game deficit in the American League East standings behind the first-place Toronto Blue Jays.
The surging Blue Jays won again Sunday to extend their winning streak to seven games and keep their division lead at three games, but Bellinger’s glove and arm ensured it didn’t grow to four.
“That was an unbelievable play,” Goldschmidt said. “Amazing catch and absolute cannon to me at first. To make that play was a game-changing play and potentially game-winning play for us today. And we needed it.”