Google Updates AI Weather Model With More Satellite Data, Says Report

Google has updated its artificial intelligence-based weather forecasting model to use an expanded set of raw satellite data, according to a report published by Ars Technica on Wednesday.

The report states that the updated model benefits from the additional inputs in a manner similar to traditional, physics-based weather forecasting systems, which have long relied on incorporating more observational data to improve predictions.

Ars Technica's report indicates the change is intended to improve forecast accuracy, though specific performance metrics or details about the scale of the data expansion were not fully outlined in the available excerpt of the report.

AI-driven weather models have gained attention in recent years as an alternative or complement to conventional numerical weather prediction, which relies on solving physical equations governing atmospheric behavior. Machine learning approaches instead train on large volumes of historical and observational weather data to identify patterns and generate forecasts, often at a fraction of the computational cost of traditional methods.

Google has previously developed AI weather forecasting tools as part of its broader push into applying machine learning to scientific and environmental prediction problems. The company has not, according to the available report, released a detailed technical breakdown of the specific satellite data sources added in this update.

Further details about the model's architecture, the specific satellite instruments contributing data, and independent verification of accuracy improvements were not available in the source material reviewed for this report.