On September 23, 2024, IBM and NASA launched the Prithvi Weather-Climate (Prithvi WxC) model on Hugging Face, marking it as the largest open-source geospatial AI model to date.

This collaboration between NASA and IBM represents a significant step in leveraging AI to address critical global challenges related to climate change.

Prithvi WxC is designed to be versatile, capable of functioning on both global and localized scales, making it suitable for a wide range of applications.

The model, along with fine-tuned versions for specific applications, is available for download on Hugging Face, targeting the scientific, developer, and business communities.

Prithvi WxC has been pre-trained on 40 years of Earth observation data from NASA's Modern-Era Retrospective Analysis for Research and Applications, Version 2, ensuring a robust foundation for its predictions.

The model aims to enhance precision, speed, and cost-effectiveness in solving climate-related challenges, addressing both short-term weather forecasts and long-term climate projections.

Beyond weather forecasting, Prithvi WxC has applications in analyzing disaster patterns, biodiversity, and land use, showcasing its broad utility.

While AI forecast emulation has advanced significantly, existing models often rely on 'gridded data' rather than utilizing the natural state of weather data, which Prithvi WxC seeks to improve.

The technology enhances model resolution, aiding in the understanding of meteorological phenomena and assessing future climate risks.

Improved weather forecasting capabilities could enhance commodity pricing and reduce costs across multiple industries, including agriculture and shipping.

IBM is collaborating with Environment and Climate Change Canada to further test the model's flexibility in weather forecasting.

A detailed explanation of the model can be found in the paper titled 'Prithvi WxC: Foundation Model for Weather and Climate' published on arXiv.