Advancing Weather and Climate Forecasting: NASA and IBM Introduce Prithvi AI Model

  • NASA and IBM Research collaborate on the Prithvi-weather-climate AI model for weather and climate applications.
  • The model utilizes AI to enhance resolution, improving regional and local forecasts.
  • Trained on NASA’s MERRA-2 dataset, the model extrapolates patterns across various scenarios.
  • It supports diverse climate applications, including severe weather prediction and climate simulations.
  • NASA emphasizes the importance of actionable science for informed decision-making.
  • The model’s open approach to sharing invites global collaboration and innovation.
  • Developed in collaboration with IBM Research and Oak Ridge National Laboratory.
  • The Prithvi-weather-climate model scales globally and regionally, maintaining resolution.
  • NASA plans to release the model on the Hugging Face platform.
  • The initiative reflects NASA’s commitment to open science principles and global collaboration.

Main AI News:

In a groundbreaking collaboration, NASA and IBM Research unveil a pioneering artificial intelligence model designed to revolutionize various weather and climate applications. The innovative model, dubbed the Prithvi-weather-climate foundational model, harnesses the power of AI to potentially enhance the precision of regional and local weather and climate forecasts significantly.

The Prithvi-weather-climate model represents a pivotal advancement in leveraging artificial intelligence for meteorological and climatological purposes. By training on a comprehensive dataset, including NASA’s Modern-Era Retrospective analysis for Research and Applications (MERRA-2) data, this model employs AI learning capabilities to extrapolate patterns across diverse scenarios.

Karen St. Germain, director of NASA’s Earth Science Division, underscores the importance of delivering actionable science to address the evolving challenges facing our planet. She emphasizes the urgency of leveraging innovative strategies to provide decision-makers with reliable weather, seasonal, and climate projections for effective preparedness and mitigation efforts.

The versatility of the Prithvi-weather-climate model empowers researchers to support a myriad of climate applications across the scientific community. From identifying and forecasting severe weather events to refining global climate simulations at regional levels, this transformative model offers a multifaceted approach to enhancing weather and climate predictions.

Kevin Murphy, NASA’s chief science data officer, highlights the model’s role in democratizing access to scientific data, thereby fostering global collaboration and innovation. By adopting an open approach to sharing AI models, NASA aims to maximize the utilization of its resources for the benefit of all stakeholders.

Developed through a collaborative effort involving IBM Research, Oak Ridge National Laboratory, and NASA’s Interagency Implementation and Advanced Concepts Team (IMPACT), the Prithvi-weather-climate model epitomizes a synergistic approach to scientific innovation. Its adaptable architecture enables the model to capture intricate atmospheric dynamics, even in the presence of incomplete data, while maintaining scalability for both global and regional applications.

Rahul Ramachandran, leader of IMPACT at Marshall Space Flight Center, underscores the strategic significance of developing a comprehensive suite of AI foundation models to advance NASA’s scientific objectives. These models, he asserts, will augment the agency’s capacity to derive insights from vast repositories of Earth observations, facilitating informed decision-making processes.

As part of the broader Prithvi family of models, which includes variants trained on NASA’s Harmonized LandSat and Sentinel-2 data, the Prithvi-weather-climate model exemplifies NASA’s commitment to open science principles. Scheduled for release on Hugging Face later this year, this initiative reflects NASA’s dedication to fostering collaboration and innovation in the scientific community.

Tsendgar Lee, program manager for NASA’s Research and Analysis Weather Focus Area, emphasizes the agency’s ongoing efforts to pioneer new technologies for climate analysis and decision-making. By leveraging cutting-edge advancements in AI and data science, NASA endeavors to enhance its capabilities for addressing complex climate challenges.

The development of the Prithvi-weather-climate model represents a collaborative endeavor involving various stakeholders, including IMPACT, IBM Research, and a consortium of academic and research institutions. This collective effort underscores NASA’s commitment to advancing scientific knowledge and promoting global collaboration in addressing pressing environmental issues.


The collaboration between NASA and IBM Research to develop the Prithvi-weather-climate AI model signifies a transformative shift in weather and climate forecasting. By leveraging advanced AI technologies, this model not only enhances the accuracy of predictions but also promotes accessibility and collaboration within the scientific community. Its release on the Hugging Face platform indicates a broader trend towards democratizing scientific knowledge and fostering innovation, ultimately shaping the future of weather and climate research and its applications across various industries.


Your email address will not be published. Required fields are marked *