How IBM is Using Geospatial AI to Fight Climate Change

Through its work in mapping urban heat islands, supporting reforestation efforts, and improving weather forecasting, IBM's partnerships with various esteemed institutions and governments globally demonstrate a dedicated use of technology for positive environmental impact. In this blog, we’ll be exploring some of IBM’s key projects and partnerships that exemplify their commitment to leveraging AI for a more sustainable future. 

IBM's Collaborative Efforts in Geospatial AI

IBM is at the forefront of combating climate change through its partnerships with entities like Mohamed Bin Zayed University of Artificial Intelligence, the Government of Kenya, the UK's STFC Hartree Centre, and NASA. These collaborations are paving the way for projects focused on urban heat island mapping, reforestation, and enhancing climate resiliency in aviation and weather forecasting.

IBM, a global leader in technology, is making significant strides in addressing climate change through various collaborative initiatives using geospatial AI. These efforts span across regions and sectors, demonstrating the versatile application of AI in environmental sustainability.

• Mapping Urban Heat Islands in Abu Dhabi: In partnership with the Mohamed Bin Zayed University of Artificial Intelligence (MBZUAI), IBM is applying its advanced geospatial AI models to map and reduce urban heat islands in Abu Dhabi. This collaboration has already shown promising results, reducing heat island effects and offering valuable insights for urban planning and climate adaptation strategies​​.

• Reforestation Efforts in Kenya: IBM supports Kenya's ambitious National Tree Growing and Restoration Campaign, aiming to plant 15 billion trees by 2032. Leveraging its geospatial AI technology, IBM has developed a digital platform to monitor tree planting in critical water tower regions, contributing to forest restoration and effective carbon sequestration management​​.

• Climate Risk Assessment Tools in the UK: Collaborating with the UK's Science and Technology Facilities Council (STFC) and Royal HaskoningDHV, IBM is developing AI-driven tools to assess climate risks, with an initial focus on the aviation industry. The project also includes the TreesAI research initiative, which aims to identify optimal locations for tree planting to mitigate surface water flooding, providing a valuable tool for urban planning and development​​.

Revolutionising Environmental Insights with AI

Moving beyond IBM's collaborative projects, let's focus on the core technology at play: IBM's AI models. These models, enriched with extensive geospatial data such as satellite imagery, are subtly altering our approach to environmental challenges.

The practical application of these AI models extends across a range of critical tasks. They prove particularly useful in interpreting complex environmental data, which is essential for initiatives like analysing flood patterns or understanding the impact of forest fires.

What distinguishes these models is their proficiency in handling complex data to yield valuable insights. In scenarios like tackling urban heat in the UAE or aiding reforestation in Kenya, IBM’s AI models contribute to data-driven decision-making, albeit in a less conspicuous yet effective manner. Their capacity to manage and interpret large volumes of geospatial data is what subtly sets apart IBM’s approach in the realm of environmental research and initiatives.

In exploring IBM's innovative efforts, it becomes evident that the application of AI in environmental analysis is steadily evolving. This technological evolution is not only refining current strategies for environmental management but also laying the groundwork for future innovations in climate response and sustainability.

Extending the Reach with Nasa

In a significant leap forward, IBM and NASA's Marshall Space Flight Center have embarked on a collaborative journey to harness the power of IBM's AI technology in deciphering the vast expanse of Earth and geospatial science data. This partnership is set to apply AI foundation model technology to NASA's Earth-observing satellite data for the first time, marking a groundbreaking step in the field of geospatial intelligence​​.

Foundation models, central to this collaboration, are trained on extensive sets of unlabelled data and can be utilised for a variety of tasks. These models have revolutionised the field of natural language processing (NLP) and are now being extended to Earth science applications. The goal is to expedite the process of extracting insights from massive datasets, thereby accelerating our scientific understanding of Earth and responses to climate-related issues​​.

One of the key projects involves training an IBM geospatial intelligence foundation model on NASA's Harmonised Landsat Sentinel-2 (HLS) dataset. This dataset encompasses a record of land cover and land use changes captured by Earth-orbiting satellites. By analysing petabytes of satellite data, this foundation model technology aims to provide critical analysis of environmental systems, including phenomena like natural disasters, crop yields, and wildlife habitats​​.

Additionally, this collaboration will generate an easily searchable corpus of Earth science literature, thanks to IBM's development of an NLP model trained on nearly 300,000 Earth science journal articles. This model, powered by IBM's PrimeQA, is expected to be a valuable resource for researchers and may also be integrated into NASA's scientific data management processes​​.

The collaboration between IBM and NASA emphasises the importance of teamwork across different organisations, bringing together diverse perspectives, resources, and skills to build foundation models that can be used for numerous downstream applications​​.

Raghu Ganti, principal researcher at IBM, highlights the potential of applying foundation models to various non-language factors within Earth science data. This approach can unlock valuable insights and information, benefiting a wide array of researchers, businesses, and citizens, and ultimately addressing some of the most pressing climate issues​​.

Furthermore, other potential joint projects in this agreement include constructing a foundation model for weather and climate prediction using the MERRA-2 dataset, a collection of atmospheric observations. This initiative is a part of NASA's Open-Source Science Initiative, which aims to foster an inclusive, transparent, and collaborative open science community​​.

To summarise, IBM and NASA's partnership is a monumental stride in the use of AI for weather and climate prediction. By harnessing the power of foundation models, this collaboration is poised to transform our ability to understand and respond to the challenges posed by climate change, making weather and climate modelling more accessible and efficient for the global research community.