Source: AGU Advances
Models of glacial flow and retreat rely on estimates of glacial ice viscosity, the measure of the ice’s resistance to flow.
Ice viscosity is dependent on the stress applied to the glacier. Most ice sheet models use a standard equation to model ice flow that includes the variable n, called the stress exponent. A larger value of n means ice viscosity is more sensitive to changes in stress. For decades, glaciologists have, almost exclusively, used an assumed n value of 3 in the models they use to predict ice flow.
However, through recent experiments and observations, researchers have found that an n value of 4 may actually better represent the conditions of Earth’s ice sheets and glaciers.
Martin et al. created a model representation of the fast-retreating Pine Island Glacier in West Antarctica. The ice sheet in their model had a true n value of 4, but they ran model projections using both n = 4 and n = 3. That allowed them to observe how their model would incorrectly predict glacial flow and resulting sea level change, given an incorrect n value.
The researchers modeled glacial retreat for 100 years under both equations with two different glacial melting scenarios. They then modeled glacial recovery for another 300 years. Under a moderate scenario, the n = 3 model underestimated glacial retreat by 18% and sea level change contributions by 21%. Under an extreme melting scenario, the model underestimated sea level contributions by 35%.
Notably, those disparities in glacial retreat and sea level change contribution predictions increased more than would be expected between the two scenarios, potentially increasing the level of uncertainty in current projections of sea level change. The researchers also suggest that incorrect n values may be mistakenly attributed to other physical processes in current ice sheet models.
The results could have far-reaching implications for predictions of future glacial melt and may prompt investigations into its effects on sea level, the authors say. (AGU Advances, https://doi.org/10.1029/2025AV001946, 2026)
—Madeline Reinsel, Science Writer
