CLIMATEWIRE | While artificial intelligence (AI) has made strides in various fields, it still fell short in predicting the recent floods in Texas.

During the floods of July 4, the most accurate weather predictions came from traditional models specifically engineered for local, high-resolution forecasting. In contrast, global models and AI-driven models did not perform as well, according to weather specialists.

“Those new, sophisticated AI models? They missed it as well,” remarked Daniel Swain, a climate scientist at the California Institute for Water Resources, during a live YouTube presentation on July 7.

---

Supporting Scientific Journalism

If you appreciate this article, please consider supporting our distinguished journalism by subscribing. Your subscription helps sustain the future of influential stories that are shaping our understanding of today’s world.

---

However, some weather experts believe AI’s role in meteorology could evolve. AI systems are beginning to better understand the complexities of atmospheric physics, which could eventually enable them to predict unique weather phenomena based on current atmospheric conditions.

“AI models are indeed capable of predicting events that are not in their previous datasets,” stated Corey Potvin, a scientist at NOAA’s National Severe Storms Laboratory in Norman, Oklahoma.

However, potential budget reductions at NOAA and a history of limited federal funding for AI in weather forecasting pose significant challenges to the advancement of federal AI weather models, experts argue. This issue underscores how the Trump administration’s broader intent to scale down government could undermine the nation’s ability to forecast weather, particularly as extreme weather events become more common.

Kim Doster, NOAA’s director of communications, assured in an email that the proposed budget cuts would not affect the agency’s research and forecasting priorities.

Commerce Secretary Howard Lutnick, who oversees NOAA, “is dedicated to incorporating cutting-edge technologies like AI to provide the quickest, most accurate data to the American public,” Doster explained. “This administration is actively collaborating with meteorologists and scientists to enhance research, reduce forecasting times, improve mapping, and streamline operations across the agency.”

Currently, the most sophisticated AI models for weather prediction are predominantly found in the private sector. Yet, many of these models did not accurately predict the Texas floods, similarly to their traditional counterparts.

One key issue is that many AI models continue to focus on predicting broad weather patterns at a global scale, noted Russ Schumacher, a meteorologist at Colorado State University and the state climatologist for Colorado.

“Predicting precipitation at a local scale is extremely challenging, and this has not been the main focus of most AI models currently in use,” he explained via email.

Despite suggestions that the Texas forecasts could have benefited from increased investment in AI at the National Weather Service, Tim Gallaudet, former acting NOAA administrator under Trump, recommended in a July 7 op-ed that the NWS should “incorporate more artificial intelligence” in its modeling systems to enhance accuracy during events like the Texas floods.

Nevertheless, some experts express concerns about AI’s capability to predict record-shattering weather events due to their rarity, which means there’s limited historical data for AI systems to learn from.

In a 2023 commentary published in Nature, weather experts Imme Ebert-Uphoff and Kyle Hilburn cautioned that AI systems are “often unpredictable under conditions they have never experienced before,” suggesting that extreme weather could lead to highly erratic predictions.

Potvin remains optimistic that new AI models will eventually be able to forecast rare events, albeit not as accurately as with more extensive training data. He also noted that while many AI models currently focus on large-scale patterns, more high-resolution models are likely forthcoming.

NOAA is developing several local weather modeling projects that incorporate AI. The National Severe Storms Laboratory’s experimental Warn-on Forecast system, or WoFS, is designed to integrate radar and satellite data into a high-resolution model rapidly, updating forecasts every 15 minutes to improve the accuracy of warnings for sudden extreme events like flash floods.

NSSL researchers are also refining an AI version of WoFS, dubbed WoFSCast. This model aims to match the performance of the original system but can potentially deliver forecasts much quicker and with significantly less computational power, making it a more cost-effective option for local NWS offices.

Additionally, NOAA’s High-Resolution Rapid Refresh model, known for its local storm forecasting prowess, was among the models that accurately predicted the Texas rainfall. An AI version of this model, named HRRRCast, is also in development.

“To my knowledge, WoFSCast and HRRRCast are the only AI models currently being developed for higher resolution forecasting,” Potvin added.

Investment Shortfalls

Despite the potential, NOAA continues to trail behind the private sector in investing in AI for weather prediction.

This poses a significant concern for NWS forecasts, as AI rapidly becomes a critical element in weather modeling that could greatly benefit NWS meteorologists—if there were more funds available for investment.

“The private sector has advanced much further than NOAA at this point, to such an extent that even in a normal budget environment, catching up would be a challenge,” commented Mary Glackin, former president of the American Meteorological Society, during an AMS panel last week.

Meanwhile, the White House has proposed approximately $2.2 billion in cuts to NOAA’s budget for fiscal year 2026.

One major proposal is to eliminate NOAA’s entire research arm, which includes a broad network of cooperative research institutes and laboratories, such as the NSSL, where researchers continue to enhance forecasting systems like WoFS and its AI counterpart.

Scientists have warned that these cuts could jeopardize NOAA’s weather forecasting abilities, potentially putting communities at risk during extreme weather events.

Relying solely on the private sector to compensate for the shortfall in federal investment in weather technology would be insufficient, according to scientists and meteorologists. Agencies like the NWS are committed to public service and providing free forecasts, ensuring that all U.S. communities—including those with limited resources—have access to accurate, lifesaving weather warnings.

More collaboration between the public and private sectors could help propel NOAA’s AI weather system development, Glackin suggested. Such partnerships would address the needs of the private sector, which seeks profit and competitive advantage, while staying true to the public service ethos and not neglecting the underserved.

However, these partnerships depend on the continued existence of NOAA’s research infrastructure, which may not survive if Congress approves Trump’s proposed budget cuts.

Moreover, while AI is a promising tool in weather forecasting, traditional weather models continue to improve as more data is collected and integrated. This has been evident in how hurricane predictions have significantly advanced over the past few decades.

“While I am a strong proponent of AI, it would be unwise to focus all our resources on AI at the expense of traditional weather modeling,” Potvin stated. “Ultimately, neglecting the traditional aspects would undermine future AI advancements.”

Meteorologists have cautioned that traditional weather forecasts could plateau—or even deteriorate—if Trump’s proposed cuts are enacted, crippling NOAA’s research capabilities.

“I am concerned about the loss of investments in science,” said Brad Colman, another former AMS president, during the July 10 panel. “These are our essential resources, and their impact will be long-lasting. I sincerely hope that better judgment prevails, and we maintain that capacity.”

Reprinted from E&E News with permission from POLITICO, LLC. Copyright 2025. E&E News provides crucial news for energy and environment professionals.