The Editor,

Journal of Glacialogy

Sir,

We recently observed a peculiar but fascinating melt pattern on fresh-water ice which we thought might be of interest to others. Basically, we had frozen a sheet of ice at −25°C in a 6 m by 3 m by 1 m pool inside one of our large cold chambers, and then had left the chamber controls set to idle at −2°c over a weekend. However, a malfunction in the refrigeration unit caused a loss of refrigeration, although the large evaporator fans continued to blow air across the ice surface. When we checked on the Monday, the chamber ambient temperature was +15°C and there was 2–3 cm of water on top of the ice. The top ice surface, however, was not planar but had large “craters” approximately 4 cm average diameter and 1–1.5 cm deep, quite equally spaced over the whole ice surface (Fig. 1). This figure shows the upper surface of the ice as viewed from an oblique direction. In general, each of the craters was surrounded by six others in a reasonably regular hexagonal pattern (Fig. 2). This figure is a plan view made with side lighting which clearly shows this hexagonal arrangement. To see whether this phenomenon was simply related to the crystalline grains in the ice, a thin section was made by simply melting a piece of the ice between crossed polaroids. This section indicated that the ice was columnar grained (type S2) with an average grain diameter of c. 0.5 cm (i.e. an order of magnitude smaller than the observed craters). The fine hair-like structure evident in Figure 2 is a result of the photographic technique, since it was not readily observable by the naked eye. These hair-like lines appear to be the result of a type of thermal etching (personal communication from N. K. Sinha) and, in some cases, the crystalline boundaries are clearly evident (as shown by the arrows in Figure 2). Finally, there was a series of thin (c. 0.2 cm diameter) channels which ran from many of the summits between these craters to the bottom of the sheet of ice. These can be partially seen in Figure 1 and are fully seen in Figure 3. This figure shows an oblique view of the bottom and side of an ice block which was cut from the sheet of ice and placed on wooden panels. Since the ice was transparent, both the underside of the top surface and the worm-like channels can be clearly seen in the photograph. The thickness of the ice block was c. 10 cm.

Fig. 1. Photograph of the upper surface of melting fresh-water ice showing the numerous cavities and worm-like channels through the ice. The cavities were c. 4 cm in diameter and 1-1.5 cm deep.

Fig. 2. Photograph of the upper surface of the ice using side lighting showing the regular hexagonal pattern of the cavities. The fine hair-like structures show internal defects in the ice and thermally etched grain boundaries on the surface (arrows).

Fig. 3. Photograph of an oblique view of the bottom and side of an ice block cut from ice and placed on two wooden panels. The underside of the upper surface of the ice and the thin worm-like channels through the ice are clearly visible. The thickness of the ice block was c. 10 cm.

In discussions with some of our colleagues on this matter, several interesting theories have been put forward to explain this curious melt pattern, although none has adequately explained the general regularity of the pattern nor the numerous thin channels through the ice. Can anyone, as a matter of general interest, propose a mechanism by which such a melt pattern may occur ?

We should like to thank Patricia Grichen for taking the excellent photographs shown here.