Skip to main content
Patches of Arctic permafrost with brown green grass growing on top layer surrounded by pools of cold blue and clear water from ice thawing.

Thawing permafrost. Credit: Anna Liljedahl, Woodwell Climate Research Center

Common near the North and South poles, permafrost covers almost a quarter of Earth. These frozen grounds serve as a massive carbon sink that stores methane and carbon dioxide, preventing their release into our atmosphere.

As the Earth warms at an accelerating rate, more and more permafrost thaws and releases greenhouse gases, exacerbating climate change and impacting all living things on this planet. Permafrost’s remote locations make it difficult to study, so understanding its changing nature and characteristics is a challenge.

That’s why Arctic researcher Anna Liljedahl from the Woodwell Climate Research Center collaborated with University of Connecticut’s Chandi Witharana, the Arctic Data Center’s Matt Jones and NCSA Software’s Kenton McHenry. Through the National Science Foundation’s “Navigating the New Arctic” program, they sought to make permafrost data more accessible.

New areas of scientific interest are emerging that need large derived-data products to answer key questions relevant to our society. A number of these derived datasets have only recently been enabled by advancements in machine learning combined with advanced computation. The research software engineers at NCSA work with scientists to help bring these elements together.

Kenton McHenry, Associate Director, NCSA Software

The team utilizes its expertise, access to remote sensing data, software development and big-data management to help researchers and scientists explore these distant and hard-to-reach areas in a way that hasn’t been done before. In this particular case, they leveraged NCSA developments with Clowder and Parsl to automate workflows and establish an interactive gateway where researchers can access and explore this data.

Together, they’ve identified and mapped over a billion ice-wedge polygons in Arctic permafrost using deep learning, advanced software, satellite imagery and supercomputers. This data helps establish a baseline to detect, monitor and track changes in these regions. 

“This is a perfect example of how previous investments in computing infrastructure, combined with new understanding of deep learning techniques, are building a resource to help with an important issue in the Arctic,” said NSF Program Director Kendra McLauchlan.

Read more about this project and collaboration in an in-depth feature by the Texas Advanced Computing Center or NSF’s research news story.

Back to top