Computer models that use artificial intelligence (AI) cannot forecast record-breaking weather as well as traditional climate models, according to a new study.
It is well established that AI climate models have surpassed traditional, physics-based climate models for some aspects of weather forecasting.
However, new research published in Science Advances finds that AI models still “underperform” in forecasting record-breaking extreme weather events.
The authors tested how well both AI and traditional weather models could simulate thousands of record-breaking hot, cold and windy events that were recorded in 2018 and 2020.
They find that AI models underestimate both the frequency and intensity of record-breaking events.
A study author tells Carbon Brief that the analysis is a “warning shot” against replacing traditional models with AI models for weather forecasting “too quickly”.
AI weather forecasts
Extreme weather events, such as floods, heatwaves and storms, drive hundreds of billions of dollars in damages every year through the destruction of cropland, impacts on infrastructure and the loss of human life.
Many governments have developed early warning systems to prepare the general public and mobilise disaster response teams for imminent extreme weather events. These systems have been shown to minimise damages and save lives.
For decades, scientists have used numerical weather prediction models to simulate the weather days, or weeks, in advance.
These models rely on a series of complex equations that reproduce processes in the atmosphere and ocean. The equations are rooted in fundamental laws of physics, based on decades of research by climate scientists. As a result, these models are referred to as “physics-based” models.
However, AI-based climate models are gaining popularity as an alternative for weather forecasting.
Instead of using physics, these models use a statistical approach. Scientists present AI models with a large batch of historical weather data, known as training data, which teaches the model to recognise patterns and make predictions.
To produce a new forecast, the AI model draws on this bank of knowledge and follows the patterns that it knows.
There are many advantages to AI weather forecasts. For example, they use less computing power than physics-based models, because they do not have to run thousands of mathematical equations.
Furthermore, many AI models have been found to perform better than traditional physics-based models at weather forecasts.
However, these models also have drawbacks.
Study author Prof Sebastian Engelke, a professor at the research institute for statistics and information science at the University of Geneva, tells Carbon Brief that AI models “depend strongly on the training data” and are “relatively constrained to the range of this dataset”.
In other words, AI models struggle to simulate brand new weather patterns, instead tending forecast events of a similar strength to those seen before. As a result, it is unclear whether AI models can simulate unprecedented, record-breaking extreme events that, by definition, have never been seen before.
Record-breaking extremes
Extreme weather events are becoming more intense and frequent as the climate warms. Record-shattering extremes – those that break existing records by large margins – are also becoming more regular.
For example, during a 2021 heatwave in north-western US and Canada, local temperature records were broken by up to 5C. According to one study, the heatwave would have been “impossible” without human-caused climate change.
The new study explores how accurately AI and physics-based models can forecast such record-breaking extremes.
First, the authors identified every heat, cold and wind event in 2018 and 2020 that broke a record previously set between 1979 and 2017. (They chose these years due to data availability.) The authors use ERA5 reanalysis data to identify these records.
This produced a large sample size of record-breaking events. For the year 2020, the authors identified around 160,000 heat, 33,000 cold and 53,000 wind records, spread across different seasons and world regions.
For their traditional, physics-based model, the authors selected the High RESolution forecast model from the Integrated Forecasting System of the European Centre for Medium-Range Weather Forecasts. This is “widely considered as the leading physics-based numerical weather prediction model”, according to the paper.
They also selected three “leading” AI weather models – the GraphCast model from Google Deepmind, Pangu-Weather developed by Huawei Cloud and the Fuxi model, developed by a team from Shanghai.
The authors then assessed how accurately each model could forecast the extremes observed in the year 2020.
Dr Zhongwei Zhang is the lead author on the study and a researcher at Karlsruhe Institute of Technology. He tells Carbon Brief that many AI weather forecast models were built for “general weather conditions”, as they use all historical weather data to train the models. Meanwhile, forecasting extremes is considered a “secondary task” by the models.
The authors explored a range of different “lead times” – in other words, how far into the future the model is forecasting. For example, a lead time of two days could mean the model uses the weather conditions at midnight on 1 January to simulate weather conditions at midnight on 3 January.
The plot below shows how accurately the models forecasted all extreme events (left) and heat extremes (right) under different lead times. This is measured using “root mean square error” – a metric of how accurate a model is, where a lower value indicates lower error and higher accuracy.
The chart on the left shows how two of the AI models (blue and green) performed better than the physics-based model (black) when forecasting all weather across the year 2020.
However, the chart on the right illustrates how the physics-based model (black) performed better than all three AI models (blue, red and green) when it came to forecasting heat extremes.
The authors note that the performance gap between AI and physics-based models is widest for lower lead times, indicating that AI models have greater difficulty making predictions in the near future.
They find similar results for cold and wind records.
In addition, the authors find that AI models generally “underpredict” temperature during heat records and “overpredict” during cold records.
The study finds that the larger the margin that the record is broken by, the less well the AI model predicts the intensity of the event.
‘Warning shot’
Study author Prof Erich Fischer is a climate scientist at ETH Zurich and a Carbon Brief contributing editor. He tells Carbon Brief that the result is “not unexpected”.
He adds that the analysis is a “warning shot” against replacing traditional models with AI models for weather forecasting “too quickly”.
The analysis, he continues, is a “warning shot” against replacing traditional models with AI models for weather forecasting “too quickly”.
AI models are likely to continue to improve, but scientists should “not yet” fully replace traditional forecasting models with AI ones, according to Fischer.
He explains that accurate forecasts are “most needed” in the runup to potential record-breaking extremes, because they are the trigger for early warning systems that help minimise damages caused by extreme weather.
Leonardo Olivetti is a PhD student at Uppsala University, who has published work on AI weather forecasting and was not involved in the study.
He tells Carbon Brief that “many other studies” have identified issues with using AI models for “extremes”, but this paper is novel for its specific focus on extremes.
Olivetti notes that AI models are already used alongside physics-based models at “some of the major weather forecasting centres around the world”. However, the study results suggest “caution against relying too heavily on these [AI] models”, he says.
Prof Martin Schultz, a professor in computational earth system science at the University of Cologne who was not involved in the study, tells Carbon Brief that the results of the analysis are “very interesting, but not too surprising”.
He adds that the study “justifies the continued use of classical numerical weather models in operational forecasts, in spite of their tremendous computational costs”.
Advances in forecasting
The field of AI weather forecasting is evolving rapidly.
Olivetti notes that the three AI models tested in the study are an “older generation” of AI models. In the last two years, newer “probabilistic” forecast models have emerged that “claim to better capture extremes”, he explains.
The three AI models used in the analysis are “deterministic”, meaning that they only simulate one possible future outcome.
In contrast, study author Engelke tells Carbon Brief that probabilistic models “create several possible future states of the weather” and are therefore more likely to capture record-breaking extremes.
Engelke says it is “important” to evaluate the newer generation of models for their ability to forecast weather extremes.
He adds that this paper has set out a “protocol” for testing the ability of AI models to predict unprecedented extreme events, which he hopes other researchers will go on to use.
The study says that another “promising direction” for future research is to develop models that combine aspects of traditional, physics-based weather forecasts with AI models.
Engelke says this approach would be “best of both worlds”, as it would combine the ability of physics-based models to simulate record-breaking weather with the computational efficiency of AI models.
Dr Kyle Hilburn, a research scientist at Colorado State University, notes that the study does not address extreme rainfall, which he says “presents challenges for both modelling and observing”. This, he says, is an “important” area for future research.
The post Traditional models still ‘outperform AI’ for extreme weather forecasts appeared first on Carbon Brief.
Traditional models still ‘outperform AI’ for extreme weather forecasts
The global roadmap on transitioning away from fossil fuels now being developed should be a “continuing conversation” which is part of UN climate talks, not just a one-off report, several governments told the Brazilian COP30 Presidency on Friday in Bonn.
During a 90-minute exchange of views at the annual mid-year climate talks in Germany, several European governments and the Marshall Islands said the roadmap that Brazil is due to finish by November should be incorporated into the official negotiations.
Any such push is likely to be resisted by nations whose economies are reliant on fossil fuel production. While Russia did not speak on Friday, it has said in earlier written submissions that the roadmap should not be referenced in any document approved by governments at UN climate talks.
At COP30 last year, Brazil tried to get governments to agree to produce a roadmap on how to transition away from fossil fuels but the proposal did not win consensus, with major nations like Saudi Arabia and Russia opposed.
Feedback in Bonn
To save the day, Brazil’s COP30 president André Aranha Corrêa do Lago promised at the closing plenary in Belem to draw up a voluntary roadmap in consultation with interested governments. Over 20 countries have officially submitted their opinions on this roadmap and, in Bonn on Friday, Corrêa do Lago sought their views – and those of civil society – in person after the presidency presented its findings so far.
The roadmap will also incorporate outcomes from the first global conference on transitioning away from fossil fuels held in Santa Marta, Colombia, in April and attended by around 60 countries.
A negotiator for the Marshall Islands told Friday’s meeting that at COP31 this year all governments should “welcome the collaborative effort behind the roadmap and the Santa Marta conference and for this work to be taken on to COP32 and beyond”.
-
Fewer journalists register for Bonn talks, as cuts to climate reporting bite
UN data shows a smaller number have signed up to cover the mid-year climate negotiations in Germany than in any year since the COVID-19 pandemic -
COP31 leaders unveil global targets, with spotlight on electrification
Co-hosts Türkiye and Australia want countries to sign on to global targets in three of their priority areas for this year’s COP climate negotiations
A spokesperson for Switzerland said on behalf of a group of nations which includes South Korea and Mexico that the roadmap must be a “sustained process, not a one-off report” and “we would welcome an ongoing platform for dialogue, for learning and cooperation including among fossil-fuel production countries”.
“We expect more than a document, rather a process whereby we come together to develop concrete steps, recommendations and tools to prepare for the transitions,” she said, calling on the COP31 co-presidents Australia and Turkiye and COP32 host Ethiopia to “take up the leadership” for implementing the roadmap”.
Global stocktake response
France’s negotiator said the roadmap “is a process and we will need continuing discussions” as “implementation needs time”, while the UK called for a “continuing conversation, including as we head towards the second [global stocktake]”.
The global stocktake (GST) is an official five-yearly report into how the world’s governments are doing on their Paris Agreement goal to limit global warming to 1.5C above pre-industrial temperatures.
The second stocktake will be published in 2028 and governments are likely to negotiate a response to it, which could include new commitments to reduce emissions, at COP33 that year. The response to the first global stocktake included the landmark COP28 commitment to transitioning away from fossil fuels in energy systems.
“Even though it’s not a formal part of the negotiation agenda, the roadmap can be a key input for the entire information-gathering phase of the second GST,” Enrique Maurtua Konstantinidis, an independent climate policy consultant, explained to Climate Home News.
“The key is for countries not to focus the discussion on defending the roadmap itself, but rather on its content, which is what truly matters,” he added.
At the Bonn event, civil society organisations also supported continuing the roadmap inside the formal climate process.
Natalie Jones, policy adviser for the International Institute for Sustainable Development, told Climate Home News the roadmap should be “an ongoing dialogue where countries can exchange their experiences, best practices and continue implementing the [transitioning away from fossil fuels] consensus”.
Russian resistance
But economies reliant on fossil fuel production are likely to oppose incorporating the roadmap into negotiations in Bonn and at COP summits. Russia’s written submission to Brazil’s consultation says the roadmap was not agreed by governments at COP30.
It says such work should therefore take place on the margins of the UNFCCC process, adding that “ the inclusion of any references to the “Roadmap” in the agenda or in official or informal documents” at Bonn or COP “would constitute a deviation from previously agreed consensus outcomes”.
Other major oil and gas producers like Saudi Arabia have not made written or spoken submissions and the US, as it has left the Paris Agreement, is not involved in discussions. But countries other than Russia are likely to resist incorporating the roadmap into official talks.
The submission by Japan, which is not a major producer of fossil fuels but consumes them from overseas, suggests nervousness about the roadmap. It asks Brazil for clarity on how the roadmap is “envisaged to be utilised” and argues that as many countries continue to rely on fossil fuels for electricity, a full and fast shift to “full decarbonisation” is “challenging.
After Friday’s event, Corrêa do Lago told Climate Home News that “the suggestions and the key milestones of the roadmap are not clear yet”. He added that the next step for the COP30 presidency will be to “sit down in July and August to really prepare” the content.
The veteran Brazilian diplomat added that the roadmap will have a section on the challenges of the transition and another section on solutions.
National fossil fuel roadmaps
Brazil, as COP30 president, is drawing up the global roadmap but its leader Lula da Silva has also ordered his officials to draw up a national roadmap.
In April, France became the first and so far only nation to produce a roadmap, which amalgamated different existing energy and decarbonisation plans and targets. Colombia is reportedly drawing up a roadmap too, based on a draft document by academics.
On Friday, a coalition of nearly 100 civil society organisations called on the COP31 co-presidents Australia and Türkiye to both come up with national roadmaps in order to “lead by example”. Türkiye produces about a third of its electricity from coal, while Australia is the world’s third-largest fossil fuel exporter, the NGOs said.
But in the Brazil-led consultation meeting, a Norwegian negotiator downplayed the importance of separate national roadmaps for transitioning away from fossil fuels.
While they can “have a supporting role”, the official said countries’ nationally determined contributions (NDCs) “must remain the primary vehicle for driving global climate transition.”
NDCs are climate plans, usually containing emissions reduction targets, which the Paris Agreement states governments must update with higher ambition every five years.
The post European, island states seek clear future for global roadmap to cut fossil fuels appeared first on Climate Home News.
https://www.climatechangenews.com/2026/06/12/european-island-states-seek-clear-future-for-global-roadmap-to-cut-fossil-fuels/
Big cuts in generating capacity are coming as the Colorado River struggles to meet demand.
Some day in the next 12 months—maybe in late August, maybe not until next spring— Lake Mead will drop below the critical threshold of 1,035 feet above sea level.
Climate Change
DeBriefed 12 June 2026: El Niño begins | COP31 hosts eye electrification | Atlantic current monitoring at risk
Welcome to Carbon Brief’s DeBriefed.
An essential guide to the week’s key developments relating to climate change.
This week
El Niño begins
‘DOMINO WEATHER’: The natural weather phenomenon El Niño, which can raise global heat and “bring domino weather effects across the planet”, is now underway, the US National Oceanic and Atmospheric Administration (NOAA) declared on Thursday, reported the Washington Post. The Japanese Meteorological Administration also identified the start of El Niño on Wednesday, said Bloomberg. According to the Japanese weather agency, the event is “expected to intensify in the coming months and become very strong later in the year, persisting into at least December”, reported the outlet.
‘SUPER EVENT’: BBC News reported that “many forecasts suggest this could end up as a so-called ‘super’ El Niño” and be “among the strongest ever recorded”. It added: “Coming on top of decades of human-caused warming, it could bring another record-hot year – most likely in 2027 – with disruption to weather, food supplies and economies running well into that year.”
COP31 hosts eye electrification
‘35 BY 35’: COP31 hosts Turkey and Australia have called for countries to support a target of electrifying 35% of global energy use by 2035, reported Politico. Speaking at climate talks in Bonn, Germany, Turkish minister Murat Kurum said that electrification would be a “flagship priority” at the COP31 summit, noted the publication. Kurum added that “electrifying daily life, from transport to buildings and industry” could “protect families and businesses from volatile energy markets”, said the outlet.
WASTE AND BUILDINGS: Climate Home News reported that electrification was one of three priorities unveiled by the COP31 hosts, with the other two being waste and buildings. On buildings, the COP31 hosts “quietly overhauled [their] goal”, Climate Home News said. It reported: “An initial press statement on Monday set out a target ‘to achieve at least a 25% increase in energy efficiency in buildings by 2035’. But…on Tuesday, that was replaced with a different goal to ‘reduce energy consumption intensity in the building sector by at least 25% by 2035’.”
‘HARDEST’ CHALLENGE: Elsewhere in Bonn, UN climate chief Simon Stiell said “governments must stop revisiting climate commitments and start delivering on them”, South Africa’s Mail and Guardian reported. It quoted Stiell as saying: “Tackling the global climate crisis is the hardest but most important thing humanity has ever tried to do together…We are not yet where we need to be. But we are somewhere we have never been before.”
Around the world
- ETS EXTRA: The EU has agreed “stronger” price controls on “ETS2”, its planned trading system for heating and transport emissions, according to Reuters.
- OCEAN STRESS: The rate of sea level rise has doubled in 10 years amid “severe and accelerating” pressures on oceans, said a UN report covered by Time.
- CLIMATE MIGRANTS: Donald Trump’s “immigration crackdown is largely targeting people from the countries most vulnerable to displacement from climate-driven disasters”, according to Guardian analysis.
- ULTRA-RICH: Investments by the world’s ultra-rich in 2022 are linked to nearly $1tn in climate damages, according to a Greenpeace Africa analysis covered by BusinessGreen.
Two
The number of bidders for Trump’s auction for drilling rights in an Arctic wildlife refuge, with big oil companies “sitting out the sale”, reported Bloomberg.
Latest climate research
- As the Arctic warms, increased iceberg activity could “reshape” deep-sea habitats and “elevate” navigational hazards as maritime traffic expands | Nature
- Around 11% of the population of the world’s “rarest great ape”, the Tapanuli orangutan, is estimated to have perished in an extreme rainfall event in Indonesia in 2025 | Current Biology
- Canada’s forests are shifting from a carbon sink to a carbon source, due to “wildfires disturbances” | Global Change Biology
(For more, see Carbon Brief’s in-depth daily summaries of the top climate news stories on Monday, Tuesday, Wednesday, Thursday and Friday.)
Captured
Solar power has overtaken gas in Asia to become the region’s third largest electricity source behind coal and hydropower, according to Carbon Brief analysis of data from the thinktank Ember. Solar became the third largest electricity source for Asia on an annual basis in April 2026, according to the analysis. In the year to April 2026, solar generated 1,727 terawatt hours (TWh), while gas generated 1,711TWh, it added.
Spotlight
Atlantic current monitoring at risk
This week, Carbon Brief reports on how Trump plans could disrupt efforts to track a major ocean current.
The Irminger Sea, a patch of frigid ocean east of Greenland, plays an outsized role in the Earth’s climate.
Here, surface water that has travelled thousands of kilometres from the tropics grows cold and dense enough to sink to the ocean’s depths – a transformation that must occur for the water to begin a long journey back to the southern hemisphere.
This makes the Irminger Sea an “action centre” for the mighty Atlantic Meridional Overturning Circulation (AMOC), the vast system of ocean currents that keeps temperatures in Europe mild.
Last week, the US government announced plans to dismantle ocean moorings installed in the Irminger Sea which, among other things, collect data on the health of the AMOC.
This came as part of a programme to “descope” the Ocean Observatories Initiative, a $368m network of ocean sensors installed in the Pacific and Atlantic oceans.
Two of the moorings earmarked for removal in the Irminger Sea form part of an internationally funded, trans-Atlantic AMOC monitoring array, known as OSNAP, that stretches from Canada to Scotland.
Experts told Carbon Brief the move by the Trump administration highlights the vulnerability of AMOC observation systems around the world. These deep-sea moorings – scattered across the Atlantic – collect real-time data on, among other things, ocean current, temperature, pressure and biochemistry.
Prof Penny Holliday, chief scientific officer of the UK National Oceanography Centre, told Carbon Brief that the OSNAP array, as well as the RAPID array at 26N, are “entirely dependent” on research grants that have to be “continually reapplied for”.
“Funding is perilous all the time,” she said.
A report prepared last month by scientists for Nordic ministers exploring the security of funding for AMOC observing systems warned that RAPID and OSNAP were in “critical condition” and faced “material exposure over an 18-month horizon”. Meanwhile, other key basin-wide and global components of the global AMOC observing system were rated as “at risk”.
It is not just US funding that is uncertain. The report notes, for example, that the five-yearly funding the UK provides to RAPID and OSNAP is “at risk from 2027 due to year-on-year budget reductions” at the Natural Environmental Research Council.
(RAPID is funded by the US and UK, whereas OSNAP is backed by five different countries, with the US contributing half of the total financial support.)
Report co-author Dr Femke de Jong from the Royal Netherlands Institute for Sea Research told Carbon Brief that “continued AMOC observations” are under pressure in “multiple countries”. She said:
“While the risk of a declining AMOC to society is starting to be recognised, there is not yet a system or institution in place to guarantee a way to monitor it.”
AMOC monitoring arrays are still in their infancy – RAPID, the oldest, was launched in 2004. Two decades of data captured so far shows that the AMOC is slowing down. However, scientists will need many more years of data to be able to confidently link the decline to climate change, rather than natural variability in the ocean.
NOC’s Holliday points to the disconnect between scientific and funder timelines:
“The timescale of observations needed in order to be able to detect a climate change signal from the very naturally variable ocean is around 40-60 years…. [And yet], in the Netherlands, they have to apply for a new grant for their ocean moorings every two years. They are going to have to do that for 40 years.
“This is a very inefficient way of getting funding for what should be critical infrastructure.”
This spotlight first appeared in Cited, Carbon Brief’s new fortnightly newsletter focused on climate research. Sign up for free.
Watch, read, listen
‘BEYOND GROWTH’: A group of economists set out a “roadmap for eradicating poverty beyond growth” in the Guardian.
OIL CAMPAIGN: Politico reported on how “oil industry allies” are campaigning against attribution science, including by working to discredit a US National Academies report that “will examine research into the ways corporate climate pollution is intensifying natural disasters”.
‘FIGHT BACK’: For the Apocalyptic Optimist podcast, Dr Dana Fisher spoke to historian and author Dr Naomi Oreskes about how to “fight back” against climate misinformation.
Coming up
- 8-18 June: Bonn climate talks, Bonn, Germany
- 16-18 June: 11th Our ocean conference, Mombasa, Kenya
- 18 June: International Energy Agency Global Hydrogen Review 2026 report launch
Pick of the jobs
- S-Curve Economics, head of road transport | Salary: £75,000-£80,000. Location: Remote (UK)
- UK Department for Energy Security and Net-Zero, speechwriter to the secretary of state | Salary: £62,595-£69,765. Location: London (hybrid)
- Basque Centre for Climate Change, postdoctoral researcher for JustBioSolar project | Salary: €27,040-€34,320. Location: Bilbao, Spain
- Boston Globe climate science and environment reporter | Salary: Unknown. Location: Boston, US
DeBriefed is edited by Daisy Dunne. Please send any tips or feedback to debriefed@carbonbrief.org.
This is an online version of Carbon Brief’s weekly DeBriefed email newsletter. Subscribe for free here.
The post DeBriefed 12 June 2026: El Niño begins | COP31 hosts eye electrification | Atlantic current monitoring at risk appeared first on Carbon Brief.
-
Climate Change10 months ago
Guest post: Why China is still building new coal – and when it might stop
-
Greenhouse Gases10 months ago
Guest post: Why China is still building new coal – and when it might stop
-
Greenhouse Gases2 years ago嘉宾来稿：满足中国增长的用电需求 光伏加储能“比新建煤电更实惠”
-
Climate Change2 years ago嘉宾来稿：满足中国增长的用电需求 光伏加储能“比新建煤电更实惠”
-
Climate Change2 years ago
Bill Discounting Climate Change in Florida’s Energy Policy Awaits DeSantis’ Approval
-
Renewable Energy8 months agoSending Progressive Philanthropist George Soros to Prison?
-
Carbon Footprint2 years agoUS SEC’s Climate Disclosure Rules Spur Renewed Interest in Carbon Credits
-
Greenhouse Gases11 months ago
嘉宾来稿：探究火山喷发如何影响气候预测