As climate change intensifies, AI has emerged as a crucial tool in understanding and predicting environmental phenomena.

To tackle the climate crisis, IBM, a global technology firm and NASA have unveiled an AI foundation model designed to address weather and climate-related challenges.

This collaboration marks a significant step in applying AI to climate science, offering a versatile tool for analysing vast datasets, identifying trends and predicting future climate scenarios.

How can IBM and NASA’s AI model help the climate crisis?

The AI foundation model, developed with Oak Ridge National Laboratory, demonstrates flexibility in handling tasks related to short-term weather forecasting and long-term climate projection.

Karen St. Germain, director of the Earth Science Division of NASA's Science Mission Directorate, explains: "The NASA foundation model will help us produce a tool that people can use: weather, seasonal and climate projections to help inform decisions on how to prepare, respond and mitigate."

The model can create targeted forecasts based on local observations, enhancing prediction accuracy for specific regions.

Its capabilities also extend to detecting and predicting severe weather patterns by analysing historical data to identify trends indicating extreme events.

Additionally, the AI model can improve the spatial resolution of global climate simulations, providing more detailed information at a finer scale.

This enhancement could lead to more accurate representations of physical processes in numerical weather and climate models.

What are the technical specifications and accessibility assets of the AI model?

The foundation model was pre-trained on 40 years of Earth observation data from NASA's MERRA-2 dataset, providing a comprehensive understanding of the Earth's climate system.

IBM has made the model available for download on Hugging Face, a platform for sharing machine learning models.

This move allows researchers and developers to access and utilise the technology for various applications.

Two fine-tuned versions of the model have been also developed for specific use cases.

One version focuses on downscaling, a technique to increase the resolution of climate and weather data, enabling more localised forecasts.

The other aims to improve the representation of gravity waves in numerical models, potentially enhancing the accuracy of weather and climate predictions.

AI's expanding role in climate science

IBM and NASA’s approach marks a departure from existing large AI models in the climate space, which often focus on fixed datasets and single use cases, primarily forecasting.

Juan Bernabe-Moreno, Director of IBM Research Europe (UK and Ireland) and IBM's Accelerated Discovery Lead for Climate and Sustainability, highlights: "This space has seen the emergence of large AI models that focus on a fixed dataset and single use case — primarily forecasting.

“We have designed our weather and climate foundation model to go beyond such limitations so that it can be tuned to a variety of inputs and uses".

Yet the partnership has also produced an AI model trained on nearly 300,000 Earth science journal articles to organise and make scientific literature more accessible.

This project uses IBM's PrimeQA, an open-source multilingual question-answering system.

Another outcome is the IBM watsonx.ai geospatial foundation model, built using NASA's satellite data.

This model is designed to analyse global weather patterns, track land use changes and predict crop yields.

Juan concludes: “The model can run both on the entire earth as well as in a local context. With such flexibility on the technology side, this model is well-suited to help us understand meteorological phenomena such as hurricanes or atmospheric rivers, reason about future potential climate risks by increasing the resolution of climate models, and finally inform our understanding of imminent severe weather events."

*****

Make sure you check out the latest edition of AI Magazine and also sign up to our global conference series - Tech & AI LIVE 2024

******

AI Magazine is a BizClik brand