A new wrinkle in ice science: why a single exponent in Glen’s Law could tilt our predictions about ice sheets
Glen’s Law is the backbone of how we model glacier ice — a simple, stubborn relationship that says the rate at which ice deforms under stress scales with the stress raised to a power n, modulated by a temperature-dependent factor A. In the real world, that exponent isn’t a fixed, universal badge of certainty; it’s a material property that can vary across glaciers, terrains, and even within a single ice body. A recent study by Lilien, Ranganathan, and Shapero pushes us to confront a hard truth many climate models dodge: the precise value of Glen’s n matters, and it matters differently depending on where you look.
What I find most striking here is not just that n can tilt the balance of predicted mass loss, but that its effect is forked by glacier type. If you imagine two opposite scenarios, the same iceberg birth certificate can lead to divergent destinies:
- Dynamically controlled glaciers (where the flow is dominated by internal deformation and flow down ice streams): upping n makes ice flow faster into warmer, melt-prone zones, accelerating mass loss. It’s a story of leverage — small changes in the rule of deformation accelerate melting because the system’s energy pathways funnel more ice toward ablation.
- Surface mass balance-controlled glaciers (where gains and losses at the surface and in mass balance largely govern behavior): a higher n actually curbs mass loss by reducing flux at the equilibrium line, slowing the march of ice toward lower elevations where melt dominates.
From my perspective, this duality matters for policy and science alike. If climate models adopt a single, cavernous n for all ice, we’re painting with a brush that’s far too broad. The real world is spatially heterogeneous: temperature gradients, impurity content, grain size, and even bedrock interactions all modulate how ice yields to stress. The Lilien et al. work makes that point with force: we should allow the flow-law exponent to vary by region, glacier type, and perhaps even season, rather than pretend it’s one-size-fits-all.
What this reveals about modeling practices is unsettling but illuminating. A fixed n creates a fragility in forecasts — a kind of brittleness in the code that pretends complexity can be reduced to a single exponent. The more I reflect on it, the more evident it becomes that uncertainty budgets for ice-sheet projections must explicitly incorporate spatially varying flow laws. Otherwise, we risk overconfident projections that crumble when confronted with real-world heterogeneity.
Another layer worth unpacking is the nuanced interaction with A, the temperature-dependent elasticity factor. Lilien et al. show that changing A in tandem with n and varying sliding laws can reshape outcomes in nonadditive ways. This is a reminder that climate-ice systems are not linear machines where you can turn a knob and predictably dial up or down mass loss. Melt interacts with flow, flow interacts with geometry, and geometry feeds back into surface mass balance in a tightly coupled loop.
So what does this imply for forecasting and strategy?
- It argues for region-specific calibration: assign n values that reflect local ice physics, surface conditions, and sliding regimes. One global exponent is not just an approximation—it’s a structural flaw in long-range forecasts.
- It invites richer uncertainty analyses: instead of a single spread around a mean, use multi-dimensional ensembles that sample n, A, and sliding laws across subregions. This won’t be glamorous, but it will give us more honest projections with credible confidence intervals.
- It invites a shift in how we communicate risk: policymakers need to hear that predictions are contingent on how ice deforms, not just how much fuel we burn in the atmosphere. The message should be that the future of ice sheets depends on microphysics as much as macro trends.
What many people don’t realize is how tightly small model choices can bind large outcomes. The exponent n may seem like a technical detail tucked away in a numerical module, yet it sits at the intersection of physics, geography, and policy impact. If you take a step back and think about it, a single number embodies decades of lab work, field observations, and computational tradeoffs — and it still can be wrong in places where the ice behaves differently.
This raises a deeper question about how we model natural systems: how much heterogeneity should we encode, and where should we draw the line between tractable models and faithful representation? The more I ponder, the more I suspect the future of ice-sheet projection lies not in chasing a perfect single exponent but in embracing a framework that lets the flow law respond to the ice’s local story. A detail I find especially interesting is how these regional differences could coincide with changing climate regimes, potentially amplifying or dampening melt in ways we haven’t yet imagined.
In summary, Lilien and colleagues remind us that the march of ice toward the sea isn’t a uniform procession — it’s a mosaic of behaviors shaped by the very physics we’re still decoding. My takeaway: to forecast ice-sheet mass loss with credibility, we must abandon the comfort of a universal Glen’s Law exponent and start modeling with spatially aware, physically grounded variability. Only then can we hope to capture the true uncertainty of a warming world and translate it into meaningful guidance for adaptation and resilience.
