AI Reveals New Insights into Antarctic Ice Flow, Challenging Existing Climate Models

AI reveals insights into the flow of Antarctic ice

This map of Antarctica shows the location of various Antarctic ice shelves in white; land is depicted in grey. | Agnieszka Gautier / NSIDC The researchers focused on five of Antarctica's ice shelves - floating platforms of ice that extend over the ocean from land-based glaciers and hold back the bulk of Antarctica's glacial ice. They found that the parts of the ice shelves closest to the continent are being compressed, and the constitutive models in these areas are fairly consistent with laboratory experiments. However, as ice gets farther from the continent, it starts to be pulled out to sea. The strain causes the ice in this area to have different physical properties in different directions - like how a log splits more easily along the grain than across it - a concept called anisotropy. "Our study uncovers that most of the ice shelf is anisotropic," said first study author Yongji Wang, who conducted the work as a postdoctoral researcher in Lai's lab. "The compression zone - the part near the grounded ice - only accounts for less than 5% of the ice shelf. The other 95% is the extension zone and doesn't follow the same law." Accurately understanding the ice sheet movements in Antarctica is only going to become more important as global temperatures increase - rising seas are already increasing flooding in low-lying areas and islands, accelerating coastal erosion, and worsening damage from hurricanes and other severe storms. Until now, most models have assumed that Antarctic ice has the same physical properties in all directions. Researchers knew this was an oversimplification - models of the real world never perfectly replicate natural conditions - but the work done by Lai, Wang, and their colleagues shows conclusively that current constitutive models are not accurately capturing the ice sheet movement seen by satellites. "People thought about this before, but it had never been validated," said Wang, who is now a postdoctoral researcher at New York University. "Now, based on this new method and the rigorous mathematical thinking behind it, we know that models predicting the future evolution of Antarctica should be anisotropic." The study authors don't yet know exactly what is causing the extension zone to be anisotropic, but they intend to continue to refine their analysis with additional data from the Antarctic continent as it becomes available. Researchers can also use these findings to better understand the stresses that may cause rifts or calving - when massive chunks of ice suddenly break away from the shelf - or as a starting point for incorporating more complexity into ice sheet models. This work is the first step toward building a model that more accurately simulates the conditions we may face in the future. Lai and her colleagues also believe that the techniques used here - combining observational data and established physical laws with deep learning - could be used to reveal the physics of other natural processes with extensive observational data. They hope their methods will assist with additional scientific discoveries and lead to new collaborations with the Earth science community. "We are trying to show that you can actually use AI to learn something new," Lai said. "It still needs to be bound by some physical laws, but this combined approach allowed us to uncover ice physics beyond what was previously known and could really drive new understanding of Earth and planetary processes in a natural setting."

Antarctica's hidden ice dynamics will redefine sea level projections - Earth.com

As the planet heats up and Antarctica's ice sheet melts at a faster pace, scientists warn of a steady rise in sea levels that could put coastal communities at risk. Home to enough frozen water to increase global sea levels by 190 feet if completely melted, Antarctica remains a focus of intense scientific scrutiny. Yet most climate models struggle to capture the complexities of its ice movement due to limited direct observations and the interplay between ocean currents, atmospheric conditions, and the frozen surface. In a paper published in the journal Science, a team from Stanford University utilized machine learning to analyze high-resolution remote-sensing data of Antarctic ice movements. By combining large volumes of satellite imagery with established physical principles, their approach identifies fundamental processes governing the large-scale flow of ice on the continent. The resulting insights could help refine how existing models estimate Antarctica's future and the associated threat of accelerating sea-level rise. Ching-Yao Lai is an assistant professor of geophysics at the Stanford Doerr School of Sustainability and senior author of the study. "A vast amount of observational data has become widely available in the satellite age," said Professor Lai. "We combined that extensive observational dataset with physics-informed deep learning to gain new insights about the behavior of ice in its natural environment." Antarctica's ice sheet is Earth's largest reservoir of frozen water, nearly twice the size of Australia. Acting like a vast sponge for the planet, it plays a pivotal role in regulating global sea levels. Understanding how this sheet moves and melts is essential as its rate of shrinkage increases every year. Historically, researchers have relied on laboratory experiments to approximate Antarctic ice's mechanical properties. However, simulating real Antarctic ice in a lab can be deceptively simplistic, explained Professor Lai. Factors such as variations between seawater-derived ice and snow-compacted ice, or the presence of cracks, air voids, and other irregularities, all contribute to ice movement in nature. "These differences influence the overall mechanical behavior, the so-called constitutive model, of the ice sheet in ways that are not captured in existing models or in a lab setting," Lai said. Rather than tackling every variable individually, the Stanford-led team used a machine learning model to assess large-scale patterns in the movement and thickness of ice. This model used satellite images and airplane radar data from 2007 to 2018, while adhering to basic physical laws that regulate ice flow. The researchers examined five Antarctic ice shelves - floating extensions of land-based glaciers. These shelves maintain stable levels of continental ice by keeping large amounts of it from sliding outright into the ocean. The experts discovered that areas of the shelves closer to land experience strong compression, a characteristic that aligns with typical lab-derived models for ice dynamics. However, ice located farther offshore experiences extension, causing it to respond differently in different directions - a phenomenon known as anisotropy. Study first author Yongji Wang completed the project as a postdoctoral researcher in Lai's lab and is now a postdoctoral researcher at New York University. "Our study uncovers that most of the ice shelf is anisotropic," said Wang. "The compression zone - the part near the grounded ice - only accounts for less than 5% of the ice shelf. The other 95% is the extension zone and doesn't follow the same law." This finding undermines a key assumption in many climate models that treat ice as the same in every direction. Real conditions, Wang's analysis indicates, are considerably more nuanced. "People thought about this before, but it had never been validated," Wang added. "Now, based on this new method and the rigorous mathematical thinking behind it, we know that models predicting the future evolution of Antarctica should be anisotropic." Understanding these complex ice dynamics is crucial, especially as polar regions warm rapidly. Rising seas already cause intensifying floods, coastal erosion, and severe storm damage. Models that overgeneralize the ice's properties may underestimate or misjudge how fast Antarctica's ice sheets can collapse or produce icebergs - a process called calving. Moreover, the scientists note that many existing models have assumed the Antarctic ice sheet's mechanical behavior is uniform. This oversimplification could translate into significant inaccuracies in projections about melt rates and sea-level rise. This project exemplifies an emerging strategy that merges "big data" with fundamental physics in Earth science research. Professor Lai emphasized that while machine learning can discover patterns from enormous sets of satellite images, scientists must still ensure the results adhere to the time-tested laws of physics that govern planetary processes. "We are trying to show that you can actually use AI to learn something new," Lai said. "It still needs to be bound by some physical laws, but this combined approach allowed us to uncover ice physics beyond what was previously known and could really drive new understanding of Earth and planetary processes in a natural setting." Since the data used here extend only to 2018, the team's next steps include refining their technique with newly available images and radar signals. The researchers also plan to apply the method to other parts of the Antarctic continent and possibly beyond, gleaning more clues about how these massive frozen landscapes might evolve under accelerating climate change. As researchers plug these results into broader climate simulations, the hope is that predictions about rising oceans will incorporate Antarctica's real-world complexities more faithfully. Policymakers and coastal communities around the globe stand to benefit from improved forecasts to guide adaptation strategies - from planning dikes to restoring coastal wetlands. Ultimately, the fusion of machine learning with physical knowledge marks a powerful advance in unraveling how Antarctica's ice moves and melts. By better capturing small-scale differences across the continent's ice shelves, scientists can glean deeper insights into the largest reservoir of freshwater on the planet - and how its future may shape the global coastline in coming decades. Image Credit: NASA's Goddard Space Flight Center Scientific Visualization Studio -- - Like what you read? Subscribe to our newsletter for engaging articles, exclusive content, and the latest updates.