ECMF AI forecasts

AI-powered weather forecasting has reached a global turning point in 2025, driven by new models, technologies, and collaborations among technology leaders and meteorological agencies. The main technological breakthrough is the replacement of traditional physics-based numerical models with deep learning systems capable of processing vast atmospheric data rapidly and producing high-accuracy forecasts with far lower computational costs.

Prominent models include Cambridge’s Aardvark, ECMWF’s AIFS, Google DeepMind’s GraphCast and GenCast, Microsoft’s Aurora, Huawei’s Pangu-Weather, and Climavision’s Horizon AI suite. Aardvark, for instance, can issue global and local forecasts in minutes on a desktop while using just 10% of input data and outperforming the U.S. GFS system. GenCast, published by Google DeepMind, outperformed ECMWF’s own ensemble system on over 97% of targets and demonstrated specific superiority in storm tracking and prediction speed. Microsoft’s Aurora is operational at top European centers and is notable for its ability to forecast not only weather but also other Earth systems like ocean dynamics and air quality.

The main industry players are Google (DeepMind), Microsoft, Huawei, ECMWF, and Climavision, together with key university research groups like Cambridge. Tech giants have made forecasts 1,000 times more energy efficient, requiring only a fraction of supercomputer resources, which democratizes access for developing regions without large computational facilities.

AI now assimilates real-time data from satellites, radars, and ground sensors to generate forecasts almost instantly—crucial for extreme weather warnings and emergency response. Ensemble forecasting using AI, such as that from GenCast and Horizon AI, allows for hundreds of scenario simulations, improving confidence in predicting rare, high-impact events.

The impact extends to agriculture, energy, logistics, and disaster management, leading to safer societies and greater resilience in the face of climate extremes. As research integrates more physics-based knowledge, AI models are also beginning to improve rare event prediction and long-range outlooks, with expanding commercial and humanitarian applications.

The recovery of Earth's ozone layer will contribute 40 percent more to global warming than scientists previously calculated, potentially undermining decades of climate protection efforts. The research, published this week in Atmospheric Chemistry and Physics, suggests that ozone's role as both atmospheric shield and greenhouse gas creates an unexpected climate challenge.

Ice can generate electricity in two ways: flexoelectricity, triggered when ice is bent or deformed, and ferroelectricity, present at the surface in extremely cold conditions. This explains how ice particles in thunderclouds become charged, revealing a likely mechanism behind lightning initiation. Ice’s electrical output, comparable to high-performance ceramics, enables potential uses in sensors and transducers, especially in harsh, cold environments where traditional electronics fail. These findings open up new technological possibilities and deepen our understanding of natural electrical phenomena in polar and stormy regions.

QubitCast represents a breakthrough in long-range weather forecasting, using quantum physics principles on conventional computers to detect hidden patterns in Earth's climate data. Unlike traditional weather models limited to 10 days of forecasting, the system will predict extreme weather events weeks to six months in advance.