Google’s new AI weather model could make 15-day forecasts more accurate

Google’s new AI-powered weather model could redefine how we predict and prepare for changing conditions.

On December 4, Google DeepMind, the company’s AI research division focused on solving scientific problems, released a new AI weather model called GenCastwhich promises to give more accurate probabilities of various weather conditions up to 15 days in advance.

While weather forecasting typically involves physics-based models, which can take hours to compute on massive supercomputers, Google said it aims to achieve greater accuracy in just minutes. GenCast can generate a 15-day forecast scenario in just eight minutes using a Google Cloud Tensor processor chip, according to the company.

GenCast is part of Google’s expanded suite of AI-based weather models, which includes improving Google search and maps with improved predictions for precipitation, fires, floods and extreme heat.

The details of his latest testing were published in the magazine Priroda.

While Google’s AI model provided a single best estimate of future weather, GenCast uses probability-based forecasting with 50 or more predictions of how the weather might change and an estimate of the likelihood of those scenarios. The technology works on a diffusion model, similar to machine learning models used in generative artificial intelligence.

“We trained it on 40 years of historical data from (the European Center for Medium-Range Weather Forecasts), which included variables such as temperature, wind speed and pressure at different altitudes — allowing it to learn,” the company explained in tweet.

When predicting extreme heat, cold and high wind speeds, Google said GenCast outperformed current forecasting models. It was also noted to deliver “superior predictions” of tropical cyclone paths up to five days in advance.