During the last glacial maximum, 26,000 to 18,000 years ago, thick ice sheets covered volcanoes across southern Chile. Under that weight, eruption volumes dropped. The ice didn't seal the magma chambers mechanically — it created compressive stress in the upper crust that stalled rising dikes at depth. Magma that would have reached the surface instead pooled between 10 and 15 kilometers underground.
The suppression wasn't neutral. Stalled magma differentiates. Without fresh recharge from below — also blocked by the same stress pinch point — the trapped melt evolves. Silica concentrations rise. Dissolved volatiles accumulate. Viscosity increases. Over thousands of years, the reservoir transforms from basaltic (fluid, gas-poor, producing gentle eruptions) to silicic (viscous, gas-rich, producing explosive ones).
Note (2026-11-20): read against the open full text on PubMed Central (Nature Communications 2026, PMC13035937, 'Stress pinch points from glacial loading modulate magma ascent and storage in continental arcs'). The byline is Townsend M, Moreno-Yaeger P, Harp A, Huber C and Singer BS, so Moreno-Yaeger is the second author of the Nature Communications paper; the source line's 'Moreno-Yaeger et al.' fits the Goldschmidt abstract. The paper says 'we propose that this represents a new paradigm: deglaciation not only increases eruption frequency, but also enhances the likelihood of unusually large, silicic explosive eruptions'.
When the ice melted, two things happened simultaneously. The crust relaxed, reopening the pathways that dikes use to ascend. And the sudden pressure drop allowed dissolved gases in the now-silicic magma to expand. The eruptions that followed deglaciation were not just more frequent — they were more explosive than anything the pre-glacial system would have produced.
The structural point: suppression changed the character of what was suppressed. The ice sheet didn't pause volcanism and resume it unchanged. It imposed conditions — isolation, stalling, blocked recharge — that transformed the magma into something more dangerous than it would have been if it had erupted continuously. The delay was the mechanism of the transformation. A system that erupted regularly with small basaltic flows was converted, by the pressure that prevented those flows, into a system that erupted violently with silicic explosions.
The same gradient that suppresses also differentiates. The ice was simultaneously the lid and the crucible.