Europe’s new AI weather model is faster, easier and free, what you need to know here

The European Center for Medical Air Forecasters (ECMWF), the center’s most modern physics-based models have been introduced to the forecast for the EU, which approved 20%.

The model was named artificial intelligence forecasting system (AIFS). According to an ECMWF’s release, the new model works faster than physics-based models and receives about 1,000 times less energy to predict.

ECMWF produced Ens Enges, which is now one of the world’s leading average weather forecast models in the world. Medium-sized forecasting In advance of weather forecasts between three days and 15 days, ECMWF predicts the weather a year ahead. Weather forecast models are important to know what to do for daily needs, as well as for daily needs, but also for daily needs for daily needs.

Traditional weather forecasting models create forecasts by solving physics equations. One restriction of these models is the approximation of the atmospheric dynamics. An attractive side of the AI’s managed models is only to learn more complex relationships and dynamics, rather than relied on known and documented equations.

ECMWF announcement comes to Google Deepmind’s heels The Gencast model For AI-Powered Weather Forecast, for the next iteration of Google weather forecast program Neuralgcm and Stamp. Gencast has left the front Go downECMWF’s leading weather forecasting model, 97.2% of targets between different air variables. With more than 36-hour lead times, the Gencast, 99.8% of the targets was more accurate.

However, the European Center is innovative. The start of AIFS-Single is the first operating version of the system.

“It’s a great effort that makes models work hard and reliable,” said Florian Pappenberger, forecast and Service Director, Central Release of the Center in ECMWF. “At the moment, the solution of AIFS, using a physics-based approach to 9 km (5.6 miles), is less than our model (IFS).”

“We see each other and ifs of the Aifs and IFS meet each other and a number of products for a number of products for a number of products.

The team will examine the organization’s accuracy to improve the accuracy of the organization and explore the modeling based on physicist.

“Physical-based models are the key to the existing information-assimilation process,” said Matthew Chanthri, head of the strategic lead and innovation platform for car learning in ECMWF, gizmodo in an e-mail. “The same information-assimilation process is also important to allow machine learning models every day and to allow forecast.”

“One of the next borders for machine learning weather forecast, this is a data-assimilation step, which means that if it is resolved, the full weather forecast chain can be based on the learning of the machine,” Chantry said.

Chantry is a co-author of a research, which is controlled by a report that trusts physics-based reshangs, explains the final point system.

The system called the graphop, the system uses the amount observed as a bright temperature of Polar orbiters, “Creating a consistent secret representation of the location system,” The team wrote and the ability to predict the appropriate air parameters for five days. “

Connection of artificial integration methods with physics-controlled weather forecasting modeling is a place that promises to more accurate forecasting. To date, the test shows that the aims of the anI-power can produce historical models, but so far these models rely on reanalysis. The observations in the ground were important to teach models, and how much it will be impressive when the technology forecasting the forecasting skills are forced to leave the script.