Ice mélange in Greenland fjords can collapse suddenly as submarine melt rises

Why it matters

Mélange pushes back on glacier fronts and so slows calving. If year-round mélange turns seasonal, glaciers would lose that support for part of each year. The authors say this could mean more calving, ice loss and sea level rise, and that models of future ice loss should include the collapse process.

What they did

The authors coupled an ocean model (MITgcm) to a continuum model of granular mélange flow. They ran an idealized fjord based on Sermilik Fjord, with tens of thousands of icebergs, and raised the summer subglacial discharge each year. They then ran the mélange model alone to find stable and unstable states across fjord widths and calving fluxes. They fitted simple linear rules for the minimum stable length and maximum melt rate, and compared these with satellite and melt observations at five fjords.

Key findings

• For seven years the simulated mélange survived each summer. In year 8 it collapsed over 2 weeks in autumn, even though ocean conditions looked almost the same as in year 7.
• Collapse came from a runaway rise in iceberg discharge at the mélange edge, not directly from more melt. Thinner mélange has less sidewall drag, so it flows faster and thins further.
• Collapse needs melt that increases toward the glacier. A spatially uniform melt rate gave no collapse in the rates tested, so coupling to the ocean matters.
• The predicted minimum lengths match satellite observations at Rink and Umiammakku just before their yearly collapse. Helheim and Sermeq Kujalleq are mostly longer than the minimum, but dip below it in summer.
• Even for the narrowest fjords and highest calving fluxes, the maximum melt rate for a viable mélange is near 1 m/d.

Limitations

• The model stops after the first collapse, so it cannot rebuild a mélange from zero length or run a full seasonal cycle. Sea ice is not included.
• Wind and ocean current stresses are left out, and glacier-mélange calving feedbacks are not modeled. The authors say these may matter for short, dispersed mélange such as at Kangerlussuup Sermia, which is an outlier.
• Observed melt rates are sparse and may miss very high melt close to the discharge plume, so how close real fjords are to collapse is uncertain.

Glossary

• Ice mélange: A tightly packed mix of icebergs, sea ice and small ice pieces at the front of a glacier.
• Subglacial discharge: Meltwater that flows out from under a glacier into the fjord and drives submarine melt.
• Saddle-node bifurcation: A tipping point where a stable state and an unstable state merge and vanish, so the system shifts abruptly.
• Buttressing: The backward push that mélange puts on a glacier front, which holds back calving.