The primary, developed by Huawei, details how its recent AI model, Pangu-Weather, can predict weekly weather patterns all over the world far more quickly than traditional forecasting methods, but with comparable accuracy. 

The second demonstrates how a deep-learning algorithm was in a position to predict extreme rainfall more accurately and with more notice than other leading methods, rating first around 70% of the time in tests against similar existing systems.  

If adopted, these models may very well be used alongside conventional weather predicting methods to enhance authorities’ ability to organize for bad weather, says Lingxi Xie, a senior researcher at Huawei.  

To construct Pangu-Weather, researchers at Huawei built a deep neural network trained on 39 years of reanalysis data, which mixes historical weather observations with modern models. Unlike conventional methods that analyze weather variables one after the other, which could take hours, Pangu-Weather is in a position to research all of them at the identical time in mere seconds.

The researchers tested Pangu-Weather against one among the leading conventional weather prediction systems on this planet, the operational integrated forecasting system of the European Centre for Medium-Range Weather Forecasts (ECMWF), and located that it produced similar accuracy.

Pangu-Weather was also in a position to accurately track the trail of a tropical cyclone, despite not having been trained with data on tropical cyclones. This finding shows that machine-learning models are in a position to pick up on the physical processes of weather and generalize them to situations they haven’t seen before, says Oliver Fuhrer, the top of the numerical prediction department at MeteoSwiss, the Swiss Federal Office of Meteorology and Climatology. He was not involved within the research. 

Pangu-Weather is exciting because it could forecast weather much faster than scientists were in a position to before and forecast things that weren’t in its original training data, says Fuhrer.

Previously yr, multiple tech firms have unveiled AI models that aim to enhance weather forecasting. Pangu-Weather and similar models, akin to Nvidia’s FourcastNet and Google-DeepMind’s GraphCast, are making meteorologists “reconsider how we use machine learning and weather forecasts,” says Peter Dueben, head of Earth system modeling at ECMWF. He was not involved within the research but has tested Pangu-Weather.