Latest news with #ECMWF
The Hindu
22-05-2025
- Science
- The Hindu
Microsoft AI weather forecast faster, cheaper, truer: study
Microsoft has developed an artificial intelligence (AI) model that beats current forecasting methods in tracking air quality, weather patterns, and climate-addled tropical storms, according to findings published Wednesday. Dubbed Aurora, the new system, which has not been commercialised, generated 10-day weather forecasts and predicted hurricane trajectories more accurately and faster than traditional forecasting, and at lower costs, researchers reported in the journal Nature. "For the first time, an AI system can outperform all operational centres for hurricane forecasting," said senior author Paris Perdikaris, an associate professor of mechanical engineering at the University of Pennsylvania. Trained only on historical data, Aurora was able to correctly forecast all hurricanes in 2023 more accurately than operational forecasting centres, such as the U.S. National Hurricane Center. Traditional weather predicting models are designed on first physical principles, such as conservation of mass, momentum and energy, and require massive computer power. The computational costs of Aurora were several hundred times lower, the study said. The experimental results follow on the heels of the Pangu-Weather AI model developed and unveiled by Chinese tech giant Huawei in 2023, and could herald a paradigm shift in how the world's major weather agencies forecast weather and potentially deadly extreme events exacerbated by global warming. "I believe that we're at the beginning of a transformation age in air system science," Perdikaris said in a video presentation distributed by Nature. "In the next five to 10 years the holy grail is how to build systems that can directly work with observations from remote sensing sources like satellites and weather stations to generate forecasts at high resolution anywhere we like." According to its designers, Aurora is the first AI model to consistently outperform seven forecasting centres in predicting the five-day trajectory of devastating cyclones. In its simulation, for example, Aurora correctly forecast four days in advance where and when Doksuri, the most costly typhoon ever recorded in the Pacific, would hit the Philippines. Official forecasts at the time, in 2023, had it heading north of Taiwan. Microsoft's AI model also outperformed the European Centre for Medium-Range Weather Forecasts (ECMWF) model in 92 percent of cases for 10-day global forecasts, on a scale of approximately 10 square kilometres (3.86 square miles). The ECMWF, which provides forecasts for 35 European countries, is considered the global benchmark for weather accuracy. In December, Google announced that its GenCast model had surpassed the European centre's accuracy in more than 97 percent of the 1,320 climate disasters recorded in 2019. These promising performances, all experimental and based on observed events, are being closely scrutinised by weather agencies. Several, including Meteo-France, are developing their own AI learning models alongside traditional digital models. "This is something we have taken very seriously," Florence Rabier, Director General of the ECMWF, told AFP. Their first "learning model", made available to member states in February, is "about 1,000 times less expensive in terms of computing time than the traditional physical model", she added. While operating as a lower resolution (30 sq km) than Aurora, the ECMWF model is already operational.
Irish Daily Mirror
16-05-2025
- Climate
- Irish Daily Mirror
Ireland weather: Expert highlights 'big flip' as forecasters are split
Scorching temperatures are set to continue over the weekend, but Met Éireann has warned rain could be on the horizon. Met Éireann has forecast scorching temperatures of up to 25C for the weekend, with the warmest areas being in the west and coolest along eastern fringes due to onshore winds. The national forecaster has warned that while Monday will see temperatures reach around 23C, there is a chance of some showers developing. For Tuesday, Wednesday and Thursday, Met Éireann has said that indications are the days will be largely dry with good sunny spells, though a few showers are possible, with temperatures widely reaching the mid-teens to low twenties. Alan O'Reilly from Carlow Weather shared that some weather models were showing rain for later next week, while others showed very little. He wrote: 'Now it's the ECMWF model showing rain for later next week when GFS shows very little. Big flip from this morning so don't put much faith in weather forecasts beyond early next week for now.' Here is Met Éireann's weather forecast for the week ahead: Friday Dry with plenty of sunshine today and just light easterly or variable breezes. Highest temperatures of 17C to 24C, warmest across the midlands, west and southwest. Dry and clear overnight apart from isolated mist patches. Lowest temperatures of 4C to 9C generally in light northeasterly or variable breezes. Saturday Dry tomorrow with plenty of sunshine. Highest temperatures of 17C along eastern fringes to 24C or 25C in the west, all in light to moderate easterly or variable breezes. Saturday night will be dry and mostly clear though a few mist and fog patches may develop. Lowest temperatures of 5C to 10C in light easterly or variable breezes. Sunday Mostly dry with good spells of sunshine. Highest temperatures of 17C to 24C or 25C in mostly light easterly breezes. Sunday night will be mostly dry and clear. Lowest temperatures of 5C to 10C in light easterly or variable breezes. Mostly dry with good sunny spells though with a chance of some showers developing. Highest temperatures of 16C to 23C in light to moderate easterly breezes. Rest of next week Indications are the days will be largely dry with good sunny spells though a few showers are possible. Temperatures widely reaching the mid-teens to low twenties.
Hindustan Times
30-04-2025
- Climate
- Hindustan Times
Pakistan nears 50 degree Celsius, may break its own world heat record
Pakistan temperatures are forecast to soar to a blistering 50 degrees Celsius this week, nearing the global April record. The sweltering heat has already pushed the temperatures in central and southern Pakistan to 48 degrees Celsius last weekend and is forecast to climb through Wednesday. Meanwhile, the Pakistan Meteorological Department predicted heatwave conditions from April 26 to 30, advising the public to take precautionary measures as Nawabshah in Pakistan is expected to surpass the world's highest recorded April temperature of 50 degrees, set in April 2018, The Washington Post reported Temperatures higher than 110 degrees Fahrenheit are forecast in 21 countries this week: Pakistan, Iran, Kuwait, Saudi Arabia, Mauritania, India, Iraq, Qatar, Sudan, United Arab Emirates, Oman, South Sudan, Bahrain, Mali, Senegal, Chad, Ethiopia, Niger, Eritrea, Nigeria and Burkina Faso. Current forecasts use the highly reliable ECMWF model, suggesting temperatures could reach 49 degrees on Wednesday and Thursday. Also Read | Doctor shares 'India's heatwave survival guide': 4 tips to stay safe and cool However, the same model underestimated temperatures last weekend by about 2 to 3 degrees -- raising the possibility that the actual peak could breach the record-setting 50 degrees mark. Late in the week, the unusually hot air mass will move eastward toward China, as a new heat wave causes stifling heat across Central Asia, where temperatures in Turkmenistan and Uzbekistan are forecast to soar past 38 degrees Celsius. Also Read | Una sizzles at 41.6 deg C, IMD sounds heatwave alert in parts of HP According to global temperature data, April 2025 has already been exceptionally warm, with 63% of the planet experiencing above-average temperatures. Warmer-than-average conditions have affected 116 countries; 39 have been cooler than average. These extremes are increasingly consistent with a rapidly warming climate. The first quarter of 2025 ranked as the second-warmest on record globally, only behind 2024, despite a transition from a warming El Niño to a typically cooler La Niña.
Yahoo
29-04-2025
- Climate
- Yahoo
Potentially record-breaking heat wave in Pakistan is testing human limits
An early heat wave stretching through the Middle East and South Asia could cause areas of Pakistan to reach the global April heat record, and is testing the limits of human survivability. Last weekend, the temperatures in central and southern Pakistan reached 118 degrees Fahrenheit and are forecast to rise through Wednesday, per The Washington Post. It is possible that the area will near the global April record of 122 degrees Fahrenheit. The record was set in 2018 when Nawabshah, a city in southern Pakistan, reached that extreme heat level. Weather historian Maximiliano Herrera maintains a list of monthly global temperature extremes that lists Nawabshah's 122 degrees as the record for April, at least in all of Asia. According to The Washington Post, there was a reading of 124 degrees in Santa Rosa, Mexico, in 2001 that may not be reliable. So far this month, temperatures have been more than 4 degrees above average in Pakistan, even before the arrival of the potentially record-breaking heat wave. Other countries such as Iraq, Iran, Niger, Turkmenistan and the United Arab Emirates have also seen above average temperatures this month. The wave of excessive warmth in the Middle East and South Asia is being caused by a 'sprawling dome of high pressure, like a heavy lid trapping heat in a pot,' according to The Washington Post. These areas typically experience heat waves in summer months, but this heat wave has come earlier than normal, per CNN. Heat wave conditions have been predicted by the Pakistan Meteorological Department from April 26-30. One of the most reliable weather models in the world, ECMWF, is predicting high temperatures around 120 degrees on Wednesday and Thursday in central Pakistan. The same model underestimated last weekend's high temperatures by around 2 to 3 degrees, meaning it's possible that Pakistan will reach temperatures in the low 120s this week, per The Washington Post. Temperatures over 110 degrees have been forecasted in 21 countries this week: Pakistan, Iran, Kuwait, Saudi Arabia, Mauritania, India, Iraq, Qatar, Sudan, United Arab Emirates, Oman, South Sudan, Bahrain, Mali, Senegal, Chad, Ethiopia, Niger, Eritrea, Nigeria and Burkina Faso. In the later part of the week, the heat is expected to move eastward toward China. According to CNN, experts say the rising temperatures coming earlier in the year are testing human limits. The arrival of early summer in India and Pakistan is testing the survivability limits of the people in these countries, putting huge strain on energy supplies, vital crops and livelihoods. Typically, both countries experience heat waves during the summer months of May and June, but this year's heat wave season has come earlier than normal and is expected to last longer as well. In recent decades, tens of thousands of people in India and Pakistan have been killed by extreme heat, per CNN. Climate experts say that by 2050, India will be one of the first places to go above human survivability limits. Pregnant women and their unborn children are particularly at risk under heat wave conditions, with many babies being born preterm in the summer and with respiratory issues. One major issue that comes with heat waves is persistent power outages, with areas losing power for up to 16 hours a day. 'This has intensified the impact of the heat, making it harder for people to cope,' said Ayoub Khosa, who lives in Pakistan, per CNN. The increased demand for electricity can also lead to coal shortages, causing more people to be without power for longer periods of time. It also causes problems for laborers and farmers, making it difficult to work; with a lack of drinking water, people fall sick, feel like vomiting or feel dizzy from the heat. It also causes issues for planting and harvesting crops, leading to lower yields. The effects of heat waves are felt more strongly by people living in communities without access to cooling measures or adequate housing, per CNN. So far this month, 63% of the world has experienced above-average temperatures and the other 37% has experienced below-average temperatures, per The Washington Post. Around 116 countries have been affected by warmer-than-average conditions. A strong El Niño caused 2024 to be a record-breaking warm year around the world, and this year's La Niña event hasn't done much to cool things down. This year's January to March period was the second-warmest on record for the world, behind only 2024, according to The Washington Post.
Forbes
28-04-2025
- Business
- Forbes
Increasing Weather Risks Demand A Longer View Of The Forecast
With the rapid advancement of AI-powered forecasting models like NVIDIA FourCastNet, which can generate a weeklong forecast in less than two seconds, decision-makers are now leaning on weather outlooks up to 14 days in advance. This shift isn't about trying to predict the unpredictable. It's about managing risk in a world where the cost of being unprepared has skyrocketed. From supply chain disruptions and staffing limitations to the growing intensity of weather events, the modern operational landscape has created a new dependency on long-range weather intelligence. But this change also introduces new challenges, as longer-range forecasts inherently carry more uncertainty. So how do stakeholders strike the right balance? One of the most immediate drivers is staffing and labor constraints. With widespread shortages across aviation, utilities, logistics and public works, many organizations are forced to finalize staffing schedules farther in advance. Waiting until a short-term forecast achieves higher confidence can be too late to mobilize a full response. Global supply chains and logistics dependencies add further complexity. Moving equipment, materials or personnel often requires significant lead time. For instance, utility companies may need to stage storm response crews across state lines or secure fuel for backup generators — actions that can't wait for last-minute confirmation. At the same time, extreme weather events are becoming both more frequent and more costly. In the United States, the National Oceanic and Atmospheric Administration reported 27 separate billion-dollar weather and climate disasters in 2024, marking the second-highest annual total on record. With rising volatility and intensity of these storms, the margin for error is shrinking. Even a low-confidence signal in a long-range forecast may now justify activating initial response plans. Finally, there's the human factor. Social media and public scrutiny have raised expectations for early, proactive communication. Municipal leaders and business owners face reputational risks if they appear unprepared, especially in the aftermath of a disruptive storm. The good news is that long-range forecasts are far more accurate and useful today than they were even a decade or two ago. A 7-day forecast today is roughly 25% more accurate than it was in 2000, and a 10-day forecast is about 45% more accurate, according to ECMWF data verification data. These gains mean that forecasts once considered speculative now offer enough reliability to guide real-world planning and risk mitigation. To meet these evolving demands, many organizations are turning to advanced decision support systems rather than relying on consumer-grade forecasts. These platforms integrate long-range model data, expert meteorologist interpretation and tailored insights specific to each operation's needs. In aviation, for example, some airports partner with specialized weather services that provide 10–15 day trend analysis for snow, wind or thunderstorms. This allows teams to preposition deicing equipment or adjust runway maintenance schedules well in advance. Similarly, utility companies rely on ensemble forecasts to assess the likelihood of high wind days, guiding mutual aid agreements and field crew rotations. Even with the increases in accuracy, the core challenge of utilizing long-range forecasting is the trade-off between early decisions and uncertain data. Yet many critical decisions must be made well before that window closes. This forces organizations to operate in a gray area, weighing probabilities and trends rather than waiting for definitive outcomes. Consider an outdoor event planner who sees a signal for heavy rain nine days out. Canceling too early could lead to unnecessary financial losses and disappointed attendees. Waiting too long may compromise safety. In these cases, the decision is less about exact rainfall totals and more about risk tolerance, contingency options and operational flexibility. Organizations that excel in this space don't wait for certainty. Instead, they treat long-range forecasts as planning tools to build flexible, tiered response plans. This might include staging resources, briefing internal teams, or outlining thresholds for escalation if the threat persists. The key is embracing uncertainty as part of the process, not a barrier to action. Long-range forecasting will only grow in importance as weather becomes more extreme and operational flexibility more constrained. But this evolution isn't only about advancing scientific capability, it's also about reshaping how organizations use forecasts. Meteorologists are increasingly taking on the role of risk communicators, helping decision-makers interpret long-range signals through the lens of operational impact. I have written about this topic several times in the past, including a recent Forbes article on how AI enhances the role of risk communicators. The key question is no longer around how many inches it will rain, but rather if the rain amount exceeds a threshold that requires action. Similarly, it's not just about predicting wind, but understanding whether gusts could damage infrastructure or tents, and how much lead time is needed to adjust staffing, relocate assets or modify logistics. Looking ahead, more organizations may adopt scenario-based planning, where forecast guidance informs a range of actions depending on confidence levels. This approach helps break free from the binary trap of 'go/no-go' decisions and supports a more flexible strategy — one that aligns with how modern weather risks are unfolding. In a world of growing complexity, organizations must plan further ahead, even if the view is a little fuzzier. By embracing probabilistic thinking, investing in customized weather intelligence and building operational agility, businesses and communities can stay one step ahead of what's coming. In high stakes situations with little time to spare, acting early even if imperfect is usually better than waiting too long.
