NASA and IBM’s Prithvi Model: A 12x Boost in Climate Forecast Accuracy

May 23, 2024
2 mins read
Global MERRA-2 atmospheric temperature at 650 hPa (approximately 11,500 feet) on July 12, 2016. Red/dark red colors indicate higher atmospheric temperatures; blue/purple colors indicate lower temperatures. Credit: NASA GMAO.
Global MERRA-2 atmospheric temperature at 650 hPa (approximately 11,500 feet) on July 12, 2016. Red/dark red colors indicate higher atmospheric temperatures; blue/purple colors indicate lower temperatures. Credit: NASA GMAO.

In collaboration with IBM Research and Oak Ridge National Laboratory, NASA has developed the new AI foundation model, Prithvi-weather-climate (Prithvi is Sanskrit for Earth) to support the improvement of regional and local weather and climate models.. This Prithvi-weather-climate model uses a wide range of datasets and AI learning capabilities to apply patterns identified in the datasets to various weather and climate scenarios, NASA officials announced on Wednesday.

“Our ambition is to accelerate and advance the impact of NASA’s Earth science to meet this moment of changing climate for the benefit of all humankind,” said Karen St. Germain, director of NASA’s Earth Science Division, in an online statement. She stated that the new weather-climate model would better support people, communities, and organizations in preparing for, responding to, and mitigating the effects of weather.

The Prithvi-weather-climate model enables researchers to support various climate applications used to detect and predict severe weather patterns and natural disasters, according to NASA. The new model also helps create targeted forecasts by using localized weather observations, improved spatial resolution at the regional level, and enhanced representations of how physical processes affect weather and climate.

In line with NASA’s open science policies, the Prithvi-weather-climate model will be openly available. The model and the code are scheduled for release in late 2024 through Hugging Face, a public repository for open-source ML models. “These transformative AI models are reshaping data accessibility by significantly lowering the barrier of entry to using NASA’s scientific data,” said Kevin Murphy, chief science officer of NASA’s Science Mission Directorate.

MERRA-2 aerosol optical depth (AOD) for July 21, 2012, when a massive dust storm was moving off the northwest coast of Africa. White and yellow colors indicate lower AOD values and a cleaner atmosphere; purples and reds indicate higher AOD values and higher concentrations of atmospheric aerosols. Credit: NASA GMAO.

The Prithvi-weather-climate model applies the complex dynamics of atmospheric physics even when information is lacking, generating weather and climate models without degrading resolution. “Our open approach to sharing these models invites the global community to explore and harness the capabilities we’ve cultivated, ensuring that NASA’s investment enriches and benefits all,” Murphy said.

Prithvi-weather-climate foundation model is pre-trained on 40 years of climate data from NASA’s Modern-Era Retrospective analysis for Research and Applications, Version 2 (MERRA-2), addressing the need to incorporate AI and machine learning (ML) methods into weather and climate applications such as storm tracking, forecasting, and historical analysis.

Similar Posts

Foundation models (FMs) are the basis for enabling AI and ML systems to ingest large amounts of data and generate results based on associations among the data. They serve as a baseline from which scientists can develop a diverse set of applications that can result in powerful and efficient solutions.

The foundation of Prithvi-weather-climate is 40 years of MERRA-2 data. MERRA-2 is the first long-term global reanalysis to assimilate space-based observations of aerosols and represent their interactions with other physical processes in the climate system. These data are available through NASA’s Earthdata Search. “With the Prithvi-weather-climate FM, NASA and IBM have led the creation of a unique AI representation of all knowledge available in 40 years’ worth of MERRA-2 data,” said Dr. Juan Bernabé-Moreno, director of IBM Research Europe and IBM’s accelerated discovery lead for climate and sustainability.

From a scientific and research standpoint, the model has been fine-tuned to increase the resolution of long-term climate models by a factor of 12x, a process known as “downscaling.” Using an AI model in this context avoids the high costs associated with the conventional approach using high-performance computing (HPC). The FM also improves the use of AI for better representation of small-scale physical processes in numerical weather and climate models.

Govind Tekale

Embarking on a new journey post-retirement, Govind, once a dedicated teacher, has transformed his enduring passion for current affairs and general knowledge into a conduit for expression through writing. His historical love affair with reading, which borders on addiction, has evolved into a medium to articulate his thoughts and disseminate vital information. Govind pens down his insights on a myriad of crucial topics, including the environment, wildlife, energy, sustainability, and health, weaving through every aspect that is quintessential for both our existence and that of our planet. His writings not only mirror his profound understanding and curiosity but also serve as a valuable resource, offering a deep dive into issues that are critical to our collective future and well-being.

Leave a Reply

Your email address will not be published.

Student Debt Protest, Photo Credit: American Association of University Professors (CC BY-NC-SA 2.0 DEED)
Previous Story

Biden’s $7.7 Billion Student Debt Forgiveness to Aid 160,000 Borrowers: Latest Round Announced

Representative Image
Next Story

EU Sets Global Standard with First AI Regulation: What It Means for Tech Sovereignty

Latest from Artificial Intelligence

Don't Miss