- May 18, 2026
- By Georgia Jiang
For years, scientists have warned that melting Antarctic ice could push sea levels dangerously higher by the end of this century. But a University of Maryland-led study suggests those warnings may underestimate the threat because they leave out a crucial factor: the ocean’s circulatory system.
When ice melts into the ocean, it doesn’t just raise sea levels—it also changes how the ocean circulates, which in turn changes how much ice melts, according to Department of Atmospheric and Oceanic Science Assistant Professor Madeleine Young and her research team. Its study, published Friday in Nature Geoscience, revealed that this self-reinforcing chain reaction may contribute as much to rising sea levels as the direct effects of a warming atmosphere.
“Most current climate models that inform international policy don’t consider this feedback loop at all. The Intergovernmental Panel on Climate Change (IPCC) treats melting as a fixed, rather than interactive input,” Youngs said. “We need to include ice shelf melt feedbacks when we’re estimating future ice shelf melt, the primary component of sea level rise, if we want the most accurate understanding of what’s going on.”
The mechanism comes down to water temperature and density. Cold, dense water naturally sinks and forms a barrier layer near the ocean floor that keeps warmer deep-ocean currents from reaching the base of ice shelves. When meltwater flows in, it dilutes and weakens that cold barrier—allowing warmer water to push through and melt the ice from below. More melting produces more freshwater, which further weakens the barrier and lets in more warm water. The cycle feeds itself.
“It’s a positive feedback loop where more melt leads to warmer water reaching the ice, which causes even more melt,” Youngs said. “If we [humans] continue to do business as usual, it’s a distinct possibility that we reach the climate tipping point sooner than later, especially as we consider this positive feedback loop.”
In regions of the Antarctic such as the Weddell Sea, that positive feedback loop amplifies dangerously. As upstream ice melts and freshwater pours in, the cold-water barrier erodes and warm water floods through, accelerating further melt.
But in other regions, such as the West Antarctic Peninsula and the Amundsen Sea, home to the Thwaites Glacier (informally called the “Doomsday Glacier”), the picture is more complicated. There, meltwater flowing westward from upstream forms a cold freshwater barrier that temporarily shields the ice from warmer ocean currents.
“Our study suggests that these regions—usually regarded as the most at-risk—are actually more protected than we thought, at least in the short term, because of this negative feedback loop,” Youngs said. “But this protection depends on massive upstream melting happening first, and that upstream melt has its own severe consequences on sea levels.”
The researchers believe their findings point to a significant blind spot in how scientists and policymakers currently model sea level rise. Rather than treating Antarctic ice-shelf melt as a fixed input, they argue that it should be treated as a dynamic process that continuously reshapes the ocean around it. The team suggests that understanding meltwater feedback loops across different regions of the Antarctic will be crucial for accurately tracking the speed and rate at which ice shelves are melting.
According to Youngs, underestimating the impacts of the feedback loops could be catastrophic. Over 680 million people worldwide live in low-lying coastal zones vulnerable to sea level rise. Even a modest increase beyond current projections—the IPCC estimates that Antarctic ice melt could contribute up to 28-34 centimeters of additional sea-level rise by 2100 under high-emissions scenarios—would significantly expand the reach of storm surges and permanent flooding in cities from Miami to Mumbai.
“This is really just a first investigation into this topic,” Youngs said. “What we’re showing is that the feedbacks in the Antarctic region are real, extremely impactful and vary depending on where they take place on the continent. We can’t just consider the direct impact of a warming atmosphere.”
In a new phase of work, Youngs’ team is developing higher-resolution simulations using meltwater feedback processes that will trace future trajectories from today through the year 2100, with a particular focus on identifying which ice shelves are closest to the point of no return.
“The next step is understanding exactly when and where things tip—and what that means for all of us,” Youngs said.