As our planet warms, the Antarctic ice sheet is melting faster than ever, playing a big role in rising sea levels around the world. With enough ice to potentially raise global sea levels by a staggering 190 feet, it’s crucial for us to get a handle on how this ice is moving and melting. Unfortunately, traditional climate models often miss the mark when it comes to predicting these movements because of the limited data and the complex dance between ocean, air, and ice.
But here’s where things get exciting. Researchers at Stanford University are using machine learning to dive into high-resolution remote-sensing data. This is giving us fresh insights into Antarctic ice movement for the first time. By tapping into AI, they’re uncovering the fundamental physics behind these massive ice shifts, aiming to boost our predictions about the continent’s future changes.
Despite being the world’s largest ice mass—about twice the size of Australia—the Antarctic ice sheet acts like a giant stabilizer, storing freshwater as ice and keeping sea levels in check. However, as it moves more quickly, it challenges the existing models that are based on lab assumptions. Unlike controlled lab conditions, Antarctic ice is far more complex, with differences in seawater ice, compacted snow, and even the presence of cracks and air pockets.
Ching-Yao Lai, an assistant professor of geophysics at the Stanford Doerr School of Sustainability, points out, “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.”
The team has crafted a machine learning model to examine large-scale ice movements and thickness using satellite images and airplane radar data from 2007 to 2018. This model sticks to the physical laws that govern ice movement, allowing them to develop new constitutive models that describe ice viscosity.
In their study of five Antarctic ice shelves, the researchers discovered that the sections closest to the continent experience compression, which aligns with lab models. But as the ice stretches further into the ocean, it undergoes strain, leading to anisotropic properties—meaning they vary in different directions.
Yongji Wang, the lead author of the study, shared, “Our study uncovers that most of the ice shelf is anisotropic. 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.”
Understanding how Antarctic ice moves is becoming increasingly important as global temperatures rise, worsening flooding, coastal erosion, and damage from severe weather. Although we don’t yet fully grasp why anisotropy occurs in the extension zone, ongoing research is set to refine these findings with new data from Antarctica.
This research is a big leap forward in developing a model that better simulates future conditions. By blending observational data with established physical laws, deep learning is revealing the physics of other natural processes too, potentially leading to more scientific breakthroughs and collaborations within the Earth science community.
Ching-Yao Lai concludes, “We are trying to show that you can actually use AI to learn something new. 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.”
