Microsoft develops machine learning tool for improved weather forecasts

Microsoft has unveiled a new machine learning tool aimed at improving the accuracy of weather forecasts for extended periods. Traditional physics-based forecasting models and even prior AI models often become less reliable beyond two weeks, a challenge Microsoft's new approach seeks to address.

The company's researchers report that their new method is cost-effective and allows for more precise predictions two to six weeks in advance. This enhanced certainty benefits sectors such as agriculture and emergency response, enabling better planning. The findings were published in the journal Nature, and the tool is now available as open-source software.

Microsoft's entry into AI-driven weather forecasting aligns with developments from other major technology firms like Google and Nvidia. Rather than creating a standalone forecasting system, Microsoft's tool is designed to augment existing models.

The development is expected to provide valuable long-range weather data for industries reliant on accurate predictions, potentially improving preparedness for weather-related events.