The Moon Landing Team Reunite: IBM & NASA's New Green AI

Share
NASA and IBM create AI to help global warming (picture credit: NASA)
IBM and Nasa’s new AI model offers versatile applications beyond forecasting, aiding scientists and businesses in climate data analysis

As extreme weather events become alarmingly more frequent across the globe, the shadow of climate change looms ever larger. The race to understand and predict these chaotic atmospheric shifts has never been more urgent, so the tools we rely on must evolve accordingly.

Enter AI. Artificial intelligence is now offering scientists the ability to sift through colossal amounts of data, spot emerging patterns and even forecast what the climate might throw at us next.

By harnessing the power of AI, policymakers have a vital new resource at their disposal, one that could help steer us away from catastrophe and towards a more sustainable future.

In recent months, IBM and NASA have been collaborating on such a project. Together, they are developing an AI foundation model, specifically engineered to tackle climate and weather-related challenges.

Delivering groundbreaking projects has been a part of the two organisations shared history for a long time. IBM computers were essential to putting humans on the moon with NASA's Apollo projects, so there is hope that their professional relationship can be just as fruitful when they put their minds to climate change.

Think - this could be a giant leap for mankind.

What can IBM and NASA's new AI model do?

Born out of collaboration with Oak Ridge National Laboratory, IBM and NASA's new AI model is a versatile programme, capable of anything from precise short-term weather predictions to sweeping climate change forecasts.

Karen St. Germain, director of the Earth Science Division of NASA's Science Mission Directorate

Karen St. Germain, Director of NASA’s Earth Science Division, sees it as a game-changing tool and one that can produce actionable insights.

She says: "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."

Unlike many existing weather AI models, this new system is designed for a wide range of applications
  • Creating Targeted Forecasts from Local Data: By integrating observations from specific regions, the model can deliver more precise, localised weather predictions, tailored to the unique conditions of each area.
  • Detecting and Predicting Severe Weather Patterns: By analysing historical data, the model can identify and forecast extreme weather events, including hurricanes, tornadoes, and heatwaves, providing earlier warnings and better preparedness.
  • Improving Spatial Resolution in Climate Simulations: The model enhances global climate simulations by offering more detailed information on a finer spatial scale, leading to greater accuracy in predicting long-term climate changes.
  • Enhancing the Representation of Physical Processes in Weather Models: With its ability to more accurately represent the complex physical processes driving weather and climate, the model improves the performance of numerical weather and climate models, contributing to more reliable forecasts.

Building on decades of climate data

IBM & NASA's AI foundation model was built on 40 years of Earth observation data from NASA's MERRA-2 dataset, giving it a robust understanding of the Earth's climate system.

Thanks to the specificity of NASA's data, the model is equally effective for global and regional applications.

The model is available for researchers and developers to download from Hugging Face, a popular platform for machine learning models.

Two fine-tuned versions have also been created to tackle specific scientific and industry challenges:

  • Climate and Weather Data Downscaling: This version can downscale climate and weather data to higher resolutions, allowing for more accurate, localised forecasts and climate projections.

  • Gravity Wave Parameterisation: Designed to improve the representation of gravity waves in numerical models, this version enhances the accuracy of weather and climate predictions by addressing a key gap in current modelling.

These specialised capabilities offer something truly different, meaning that this new model could alter how we approach both everyday weather forecasting and long-term climate science.

Youtube Placeholder

This particular AI model isn't the first that IBM and NASA have co-created, nor is it the first to analyse meteorological and environmental patterns. Another is the IBM watsonx.ai geospatial foundation model, which was also built using NASA’s satellite data.

This programme is capable of analysing global weather patterns, tracking land use changes and even predicting crop yields—tools critical for addressing environmental shifts and food security.

Another is a model powered by IBM's PrimeQA, an open-source multilingual question-answering system. This one uses data from over 300,000 Earth science journal articles, designed to organise and make vast amounts of scientific literature more accessible, helping to speed up climate research.

Juan Bernabe-Moreno, Director of IBM Research Europe (UK and Ireland) and IBM's Accelerated Discovery Lead for Climate and Sustainability

Juan Bernabe-Moreno, Director of IBM Research Europe (UK and Ireland) and IBM's Accelerated Discovery Lead for Climate and Sustainability, reflects on the progress the two organisations have made in driving practical changes to help protect the planet.

He says: "This space has seen the rise of large AI models that focus on a fixed dataset and single use case—primarily forecasting.

"We've designed our weather and climate foundation model to push beyond these boundaries, allowing it to be tuned for a wide variety of inputs and applications."

******

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

******

Sustainability Magazine is a BizClik brand​​​​​​​

Share

Featured Articles

Which 10 Nations Met the UN's Deadline for NDC Targets?

Only a handful of nations met the UN’s NDC deadline and with COP30 looming, slow progress is raising concerns over action on global sustainability targets

What Next as Baltic States Switch off Russian Energy Supply?

Baltic states cut ties with Russian energy infrastructure, joining Europe's grid in a major shift for energy security, independence and regional stability

This Week's Top 5 Stories in Sustainability

This week's top stories include the EU Omnibus, Trump's tariffs, Google's scrapped DEI goals, Equinor's reneged-upon commitments and UK green construction

Equinor Halves Green Spend: What Does it Mean for Net Zero?

Renewable Energy

Google Scraps DEI Goals - Why has it Followed Meta & Amazon?

Diversity & Inclusion (D&I)

The Six Ways BCG Links Sustainability and Value Creation

Sustainability