IBM, NASA using AI to answer climate change questions

A recent collaboration between tech giant IBM and NASA's Marshall Space Flight Center aims to use artificial intelligence (AI) to answer some of the most pertinent questions regarding climate change and its impact on the Earth, including environmental challenges facing Africa.

"Some of the questions that can be answered include how Africa is facing growing heatwaves and drought fueled by climate change? How is the smoke impacting air quality? With Africa reliant on subsistence farming to combat malnutrition, what impact is extreme weather conditions having on crop yield? By answering climate-related questions like these, we as a society and continent, are able to adapt to the changing environment around us," explained Kommy Weldemariam, distinguished engineer and chief architect for discovery science and applications at IBM Research.

"We live in a world where climate-related damages are real, [for example] the South African government declaring a state of emergency due to excessive rain [caused by] extreme weather patterns happening across the continent, there's no doubt that climate change is impacting Africa. Organizations, both private and public, are facing disruptions to supply chains and operations affecting the bottom line," Weldemariam told Connecting Africa.

By integrating AI with environmental, climate and weather data in industry-specific solutions, it is possible to adapt to and mitigate the effects of climate change with an accurate analysis of climate risk at scale and the [precise] accounting of carbon emissions, he explained.

"This will help identify changes in weather patterns that contribute to natural disasters or impact crop yields and wildlife habitats," he added.

Analyzing vast data sets

Weldemariam said that thanks to advancements in new satellite technologies, NASA's Earth observation satellite data is being gathered at unprecedented rates and volumes.

"This means that new and innovative approaches are needed to extract knowledge and insights from such vast data sets. We are working with NASA to create large-scale AI models that are capable of analyzing petabytes of NASA's Earth science satellite data and Earth science literature," he said.

This is with the objective to make it easier for data scientists and researchers to get the answers and insights they need by building custom AI models and applications tailored to specific climate-related questions and tasks.

"Our foundation model technology has the potential to increase the discovery and analysis of such data sets in order to quickly advance the scientific understanding of Earth and response to climate-related issues."

Partnering into the future

Collaborations between IBM and NASA are nothing new and the two groups have been working together for decades.

"It was IBMers that built and wrote many of the complex software programs that helped NASA with the Apollo Missions to put the first two astronauts on the moon in 1969," Weldemariam points out.

This most recent partnership with NASA's Marshall Space Flight Center will use IBM's AI technology to discover new insights in NASA's treasure trove of Earth and geospatial data.

This is a new application of AI foundational model technology to NASA's Earth observation satellite data.

Foundation models are types of AI models that are trained on a broad set of unlabeled data and can be used for different tasks. These models have rapidly advanced the field of natural language processing (NLP) technology over the last five years.

For this specific project, IBM is developing a net-new geospatial foundation model that could analyze petabytes upon petabytes of satellite data and which has been trained on NASA's Harmonized Landsat Sentinel-2 (HLS) dataset, which is the first time foundation models have been applied to NASA's satellite data.

IBM's foundation model technology has the potential to increase the discovery and analysis of large data sets as a way to quickly advance the scientific understanding of Earth and response to climate-related issues. (Source: NASA and IBM Research)

Additionally, geospatial foundational models will be able to help data scientists and researchers provide critical analysis of the planet's environmental systems by analyzing changes in the geographic footprint of phenomena such as natural disasters, change detection and cyclical crop yields.

"It is very important to understand this data. Having the capabilities to extract knowledge and gather insights will help scientists answer many of the pertinent questions we have regarding the environment. This will be achieved by making it easier for researchers to search the corpus of Earth science literature," Weldemariam said.

IBM's NLP model trained on nearly 300,000 Earth science journal articles to organize the literature and make it easier to discover new knowledge and insights. The model contains one of the largest AI workloads and it uses PrimeQA, an open-source multilingual question-answering system trained on Red Hat's OpenShift.

African scientific contributions

IBM's Research-Africa team is also actively using AI and hybrid cloud to develop technology that addresses climate and sustainability challenges.

One interesting example is their collaboration with global IBM Research colleagues to develop and operationalize nature-based carbon sequestration.

"By leveraging remote sensing data and geospatial foundation models, the team can quantify the amount of carbon being sequestered or stored in a region, such as the Kakamega forest [in Kenya]. This information can assist with the issuance of carbon credits for the voluntary or compliance carbon market," he explained.

Example of IBM's geospatial foundation models' ability to quantify the amount of carbon being sequestered in the Kakamega forest in Kenya. (Source: IBM Research)

Weldemariam said the African team is also working on developing early warning systems for natural disasters such as flooding or heatwaves. These systems improve disaster preparedness and reduce the impact of extreme events, contributing to overall resilience.

"For example, IBM is partnering with the African Risk Capacity (ARC) to develop early-warning systems for improve climate risk modeling and predictions. The team is also involved in the HEAT Center project, which aims to develop heat-health early warning systems for urban areas in Africa."

"These efforts demonstrate how technology can be leveraged to address climate and sustainability challenges and have the potential to make a significant impact," Weldemariam concluded.

*Top image is of Kommy Weldemariam, distinguished engineer and chief architect for discovery science and applications at IBM Research. (Source: IBM Research)

— Paula Gilbert, Editor, Connecting Africa