Discussion of “Laboratory Study of Anchor Ice Growth” by J. C. Doering, L. E. Bekeris, M. P. Morris, K. E. Dow, and W. C. Girling

Abstract

As a student of anchor ice phenomena, I read the paper with great interest. Although anchor ice has long been recognized ~Barnes 1928!, the formation of anchor ice is still something of an enigma. Identification of relationships between easily measured flow characteristics and anchor ice phenomena would be a great boon to everyone who deals with anchor ice. The authors conducted a series of 12 laboratory flume experiments to search for such relationships. They conclude that their data suggests two such relationships: ~1! a relationship between the Froude number of the flow and anchor ice ‘‘growth,’’ and ~2! a relationship between the flow Reynolds number and anchor ice release. Based on the authors’ observations and data, I argue that both of these conclusions are flawed. The authors define the anchor ice growth rate as the increase of anchor ice thickness over time, expressed as millimeters per hour. They found that the most rapid anchor ice growth occurred at a Froude number of 0.27. Increasing anchor ice thickness over time would be a valid measure of growth if the anchor ice had the same density in every experiment. However, the authors note that ‘‘qualitative observations indicate that the density of anchor ice increases for increasing Froude number’’ ~p. 65!. Based on this observation, it would seem obvious that anchor ice growth cannot be defined by thickness alone. Instead, growth should be defined by the change in anchor ice mass, because the denser anchor ice seen in higher Froude number experiments must contain more ice ~i.e., greater masses of ice! than the ‘‘fluffy patches of anchor ice observed for a low Froude number run’’ ~p. 63!. The technique the authors used to determine anchor ice growth rates has a built-in inconsistency ~varying ice density! so any observed relationship is meaningless. The authors’ other major conclusion is that for Reynolds numbers less than approximately 42,000 anchor ice was released from the bed. However, it appears the authors made an error in calculating the Reynolds numbers. The authors’ data ~from their Table 1! on experiment number ~run!, water depth ~h!, current velocity (Ve), Reynolds number ~R! and anchor ice release are shown in Table 1. Also shown are my calculations for the Reynolds numbers (Rnew). I use h and Ve , from the authors, a density of 1,000 kg/m 3 , and a viscosity of 1.787 310 23 kg/~m/s! to calculate the Reynolds number. Based on a comparison of the authors’ and my calculated Reynolds numbers, it appears that the column containing the Reynolds number in the authors’ original paper somehow got