New snow deposited on the Greenland ice sheet has a density of 50-70 kg/m3, just 5-7% of the density of water which is 1000 kg/m3. In the central parts of the ice sheet the temperature never gets above freezing so the snow does not melt away but gets buried under new layers of snow. The weight of the overlying snow gradually compresses the layers below it, and the density increases.

When a density of 830 kg/m3 is reached at a depth of approx. 80 m, all air passages between the crystals are sealed off so that air only exists in closed bubbles. This defines the transition from firn to ice. With increasing depth, the air in the bubbles is compressed and the density approaches 917 kg/m3 which is the density of glacier ice. It is impossible to compress the ice any further, and at greater depths, the thinning of the layers only happens through deformation of ice by ice flow.

Change of density

The change of density with depth measured at DYE-3 in southern Greenland. The narrow spikes of increased density marked with red asterisks are melt layers. At DYE-3 the summer temperatures sometimes get high enough to melt the surface snow. When the liquid melt water refreezes, it has a higher density than the surrounding firn.

In ice flow models the change in density with depth in the firn zone has to be accounted for. This is done by incorporating a special densification model into the flow model. However, the firn zone is quite small (70-100 m) compared to the total thickness of the ice sheet (approx. 3000 m), and in model studies focusing on deeper parts of the ice sheet an approximation is often used. In this approximation the ice equivalent thickness of the ice sheet - the thickness the ice sheet would have if all the firn was immediately compressed to ice - is used instead of the true thickness. For central Greenland the ice equivalent thickness is 25 m less than the true thickness. In other words: The firn contains 25 m of air.

Read more about how the porous firn zone influences the composition of the gas bubbles.

Relevant reading:

  • N. Reeh, D. A. Fisher, R. M. Koerner and H. B. Clausen
    An empirical firn-densification model comprising ice lenses
    Annals of Glaciology, Vol. 42, Number 1, pp. 101-106(6), 2005

  • T. Kameda, H. Shoji, K. Kawada, O. Watanabe and H. B. Clausen
    An empirical relation between overburden pressure and firn density

  • C. C. Langway, Jr., H. Shoji, A. Mitani and H. B. Clausen
    Transformation process observations of polar firn to ice