• Welcome to AlpineZone, the largest online community of skiers and snowboarders in the Northeast!

    You may have to REGISTER before you can post. Registering is FREE, gets rid of the majority of advertisements, and lets you participate in giveaways and other AlpineZone events!

New weather forecasting AI. Google DeepMind’s "GenCast" Sets a New Standard for Medium-Range Mountain Forecasting

MidnightJester

Active member
Joined
Oct 7, 2011
Messages
964
Points
43
GenCast combines a cutting-edge diffusion model with the full sweep of historical global weather data. Unlike some AI models that borrow techniques from image or music generation, GenCast is tuned to Earth’s spherical geometry and the subtle dance of the atmosphere. Trained on four decades of temperature, wind, and pressure readings, it has learned to tease out complex patterns that affect local snowfall from the largest weather systems. Impressively, it operates at a resolution fine enough—about a quarter-degree latitude and longitude—to broadly capture the shifting moods of the mountains. Even more remarkable is its speed: producing a 15-day ensemble forecast in around eight minutes, a fraction of the time traditional physics-based models require.


Looking forward to seeing this AI weather forecasting program put to the test : )~

This is Googles scientific paper on "Gencast" that is published for peer science review with open code provided.

Once the AI model was trained up, it was then asked to predict the weather for 15-day periods in 2019. When compared to the actual weather that occurred in that year, GenCast was 97.2% more accurate than the ECMWF predictions. When the window of the prediction time was narrowed to just 36 hours, GenCast did even better – it was 99.8% more accurate than ECMWF. Both metrics best Google's previous weather prediction program, known as GraphCast.

Google's weather tool gets it right 97% more times than best forecaster​

 
Last edited:

djd66

Well-known member
Joined
Sep 6, 2015
Messages
1,031
Points
113
GenCast combines a cutting-edge diffusion model with the full sweep of historical global weather data. Unlike some AI models that borrow techniques from image or music generation, GenCast is tuned to Earth’s spherical geometry and the subtle dance of the atmosphere. Trained on four decades of temperature, wind, and pressure readings, it has learned to tease out complex patterns that affect local snowfall from the largest weather systems. Impressively, it operates at a resolution fine enough—about a quarter-degree latitude and longitude—to broadly capture the shifting moods of the mountains. Even more remarkable is its speed: producing a 15-day ensemble forecast in around eight minutes, a fraction of the time traditional physics-based models require.


Looking forward to seeing this AI weather forecasting program put to the test : )~

This is Googles scientific paper on "Gencast" that is published for peer science review with open code provided.

Once the AI model was trained up, it was then asked to predict the weather for 15-day periods in 2019. When compared to the actual weather that occurred in that year, GenCast was 97.2% more accurate than the ECMWF predictions. When the window of the prediction time was narrowed to just 36 hours, GenCast did even better – it was 99.8% more accurate than ECMWF. Both metrics best Google's previous weather prediction program, known as GraphCast.

Google's weather tool gets it right 97% more times than best forecaster​

So what is the forecast for Warren VT?
 
Top