An AI developed by Google Deepmind manages to outperform current medium-term weather forecast models. Baptized GraphCastit requires only a fraction of the computing power that is currently required to deliver accurate results in record time.
This will also interest you
[EN VIDÉO] Have the weather forecasts for 2050 already come true? What could the weather report look like in 2050? In partnership with WMO, the Organization…
If you find that the forecast weather reportweather report are not always reliable, especially in the medium term, artificial intelligence (AI) should help to change your mind. According to an article recently published in the journal Sciencea team of GoogleGoogle Deepmind has developed a new 10-day climate forecast program called GraphCast.
Powered by AI, GraphCast almost always outperforms existing forecasting tools. Better yet, this AI only uses a fraction of the computing power of the current system (HRES). What makes the difference is that the AI of GraphCast has ingested decades of weather information as well as around forty years of data from satellites, weather stations and radars. Current processes work differently. With colossal computing power requiring hundreds of machines on a supercomputer, 10-day forecasts are established with gigantic databases. They contain data relating to the thermodynamicsthermodynamicsfluid dynamics and other criteria related to atmospheric conditions.
Example of a 10-day GraphCast forecast showing humidity, surface temperature and wind speed. © GraphCast, Deepmind
More precise and less energy-consuming than the current model
Only one machine of the same type of supercomputer is needed for GraphCast delivers a very precise medium-term forecast in less than a minute. Unlike the current model, by limiting the scope of the analysis to the lowest part of theatmosphereatmosphere where the most visible weather events are located, GraphCast was even proven to be more accurate in over 99.7% of tests. Since its beginnings in September with the European Center for weather forecastweather forecast medium term (ECMWF), GraphCast was already able to predict, nine days before, the precise trajectory of Hurricane Lee which impacted Nova Scotia, three days less than the HRES. GraphCast comes from open source and its experimentation will continue with the ECMWF alongside the current HRES.
rewrite this content and keep HTML tags