ice stress – A study finds that using the common n=3 stress exponent instead of n=4 can significantly misestimate glacial retreat and sea-level contributions—especially under stronger warming.
Glaciers don’t just “melt”—they can also speed up in how they flow, and new modeling work is sharpening why that matters for future sea levels.
At the heart of the change is a parameter used to represent how ice moves under stress: the stress exponent. often labeled n.. In many ice-sheet and glacier models. scientists have leaned on an assumed n value of 3 for decades. treating it as a broadly reliable stand-in for how Earth’s ice behaves.. But an emerging body of experiments and observations suggests that an n value of 4 may better match real conditions in at least some parts of the cryosphere.
The difference between n=3 and n=4 is not cosmetic.. A larger n means ice viscosity is more sensitive to stress—so changes in how forces build up in a glacier can translate into different flow speeds.. When models use the wrong sensitivity. they can misread how quickly ice responds to warming. how fast retreat proceeds. and how much melt ultimately contributes to the ocean.
In the new work, researchers focused on the Pine Island Glacier in West Antarctica, a region known for fast retreat.. They built a model representation in which the “true” glacier behavior corresponded to n=4. then ran projections using both n=4 and the long-standard n=3.. That design—keeping the glacier’s underlying behavior fixed while swapping the equation used to simulate it—made it possible to isolate the damage done by an incorrect n value.
Over a 100-year window. the researchers simulated glacial retreat under two melting conditions. then continued the simulations for 300 years to explore how the system might recover.. Under a moderate melting scenario, the n=3 approach underestimated glacial retreat by 18% and underestimated sea-level change contributions by 21%.. Under an extreme melting scenario, the gap widened: sea-level contributions were underestimated by 35%.. Put simply. the models weren’t just a little off—they could increasingly diverge from what the “correct” sensitivity would have predicted as the warming conditions became more intense.
What makes the result especially important is the pattern.. The disparities grew more than might be expected just from the difference between n=3 and n=4 alone.. That suggests the wrong exponent doesn’t merely scale the outcome—it can steer the model’s dynamics in a way that amplifies uncertainty across long simulations.. Even more concerning. the researchers argue that using an incorrect n might cause effects that should be blamed on ice physics to be misattributed to other processes already represented in current ice-sheet models.
Why should readers outside the modeling community care?. Because sea level projections influence everything from coastal planning timelines to infrastructure risk assessments.. When the physical “wiring” inside a model is slightly wrong. the forecast can shift in ways that are difficult to detect after the fact—especially when models are later used to inform scenario planning under different warming futures.
There’s also a broader lesson for cryosphere science.. Sea-level research is crowded with moving parts—ocean heat, ice-bed interactions, surface melt, and changing flow geometry all matter.. The stress exponent is one ingredient in that larger recipe.. But if that ingredient is mismatched to real ice behavior. it can skew which processes look important and which look secondary.. In that light. refining n isn’t just a technical tweak; it’s part of calibrating the model’s causal story.
Going forward, the findings point toward more systematic checks of which n values best describe different ice regions and conditions.. They may also motivate updates in how modelers choose parameters rather than treating n=3 as the default.. Misryoum will be watching how the next generation of glacier simulations incorporates these insights—because when the ice flows differently than we expect. the coastline feels it first.
Fake photos aren’t new: Rijksmuseum shows how image tricks began
Orion helium leak: NASA says no threat to Artemis II, but Artemis IV needs redesign
Tylenol in Pregnancy Linked to No Autism Risk, Danish Study Finds