|An icy stalactite in Oregon Caves National|
Monument. Photo by Phil Lachman
can be found here, where it was
the Photo of the Day on Nov. 8, 2010.
Icy stalactites have been observed beneath sea ice, e.g., in the McMurdo Sound area of the Antarctic. The bottom of a sea ice sheet has numerous "disconnected ice platelets protruding downward" (R.A. Paige, Stalactite growth beneath sea ice, Science, v. 167, pp. 171-172, 1970). This is called the "skeleton layer", and ranges in thickness from a few centimeters up to 60 cm. Part of this layer consists of freshwater stalactites. These can extend up to a meter or more below the skeleton layer. The process by which briny water forms freshwater stalactites is fairly complicated (see discussion in the Paige article above).
Ice stalactites in caves are similar to icicles observed to form on houses and trees in cold climates. In common with CaCO3 stalactites You can see videos of icicles grown under laboratory conditions here. Steven Morris of the University of Toronto has studied the growth mechanisms of icicles in detail and in lab experiments. The basic process involves the slow downward flow of water either into the ocean or into air. As the water flows, it may cool to form ice. Latent heat of fusion is given up and must be transported out through the flowing film of water and into the external seawater or air. (In the case of CaCO3 stalactites, CO2 liberated in the process must be transported out through the water film and into the air.)
Two effects operate to produce the rippled texture of icicles. Latent heat is more efficiently transferred out of the system on the convex protrusions than from the convex indentations, which tends to make the protrusions grow faster than the indentations. This is the so-called "Laplace instability". It is countered by heat transfer down the icicle by the flowing water. Operating together, they produce a remarkable constant ripple spacing of about 1 cm, although the amplitude of the ripples can vary from one icicle to another.
Morris asked: Do the ripples move? One group of researchers (Ogawa and Furukawa, Physical Review E, October 2002) that developed a physical model for the ripple development predicted that the ripples should migrate down an icicle at about half the speed that the icicle grows. Another author (Ueno, Phys. Rev. E69 (5) 2004) predicted that the ripples would move up. Morris was able to use edge detection methods on videos of the icicle development to show that the ripples moved upward very slightly.