Icicles
water, bulk salinities can exceed 10‰. The shapes of
Sample M4. This was a lifeline sample--only bulk
grains composing the crystalline structure of accreted
ice ranged from rounded to polygonal. Our observa-
salinity (16.7‰) was obtained for this (Table 2).
tion that ice formed during the initial stages of accre-
Sample M11. This example of an icicle formed on
tion frequently displayed a polygonal crystalline struc-
a polypropylene lifeline was collected at a location for-
ture can probably be attributed to thermally driven
ward of the 5-in. gun mount immediately port of the
modification of the original microstructure, caused pos-
windlass (Fig. 28 and 29). The sample was about 6 cm
sibly by heat leaking from the interior of the vessel.
long and 8 cm wide, and was hanging down and away
This process is generally manifested by the straighten-
from the relative wind (winds were from the port bow).
ing of crystal boundaries, and by the formation of triple
Brine was dripping from the icicle, and it was very firm
junctions intersecting at equilibrium angles of approxi-
mately 120, as observed in thin sections.
and difficult to remove from the polypropylene life-
line. A vertical section, cut longitudinally through the
We noticed no trend towards a preferred orientation
length of the icicle, is shown here photographed be-
of crystallographic c-axes, either in freshly accreted ice
tween crossed polarizers (Fig. 28). The microstructure
or its thermally modified (recrystallized) variant. Mean
shown in Figure 29 is characterized by grains elongated
grain dimensions ranged from a minimum value of 0.56
longitudinally down the length of the icicle, occasion-
mm to a maximum of 1.15 mm. The only exceptions
ally consisting of crystals measuring several millime-
were one deck sample accreted in very warm weather
ters in length. An intermixing of these elongate grains
(M13) and icicle type ice, where the dimensions of the
with smaller rounded grains is also featured in this sec-
crystals often exceeded several millimeters and where
tion. A bulk salinity of 11.8‰ was measured. The thin
we commonly observed dimensional orientation (elon-
section gives the impression of grains substantially im-
gation of crystals in a preferred direction) in the micro-
mersed in brine.
structure. Excluding this one deck sample and the
icicles, we found that the mean dimension of crystals
Sample M12. This sample was retrieved from one
accreted on both horizontal and vertical surfaces of the
of many icicles dripping from the bottom of the 5-in.
USCGC Midgett were similar to those measured by
gun mount (Fig. 30). The icicle sample appears to be
Tabata et al. (1963), but generally much larger than those
two or more merged icicles. Brine was dripping from
derived from three-dimensional measurements of crys-
the icicles as they were removed. Crystals appear to
tals reported by Golubev (1972). Golubev also reported
radiate from two nodes, assumed to be drip tubes, and
significant levels of preferred orientation of the crys-
swirl counterclockwise around the left drip tube but not
tallographic c-axes. We did not, nor does this appear to
around the right tube. Crystals also appear to radiate
be the case of observations of c-axis orientation in
outward from each drip tube outside of the swirl zone.
accreted ice examined by Tabata et al. (1963).
A particular feature of the vertical section microstruc-
Estimates of brine volume and entrapped air con-
ture of this icicle is the incorporation of large
tent derived from measurements of the salinity, den-
millimeter-sized grains among much smaller grains, as
sity, and in-situ temperature of samples, in conjunction
is clearly demonstrated in Figure 30. These two con-
with the equations of Cox and Weeks (1983), indicated
trasting types of crystals were analyzed separately for
widely ranging values, as expected in view of the wide
grain size; results are presented in Table 3. A bulk
ranging values of density, salinity, and temperature.
salinity of 14.2‰ was measured on one of the icicles.
Though the icing rates experienced on the USCGC
Midgett were low, and ice thicknesses were small, use-
ful measurements were made. Ice thickness was greater
CONCLUSIONS
on decks than on bulkheads, with an approximately
Ice accreted on horizontal and vertical surfaces of
1.25:1 ratio between deck and bulkhead ice thickness.
the USCGC Midgett ranged from densely packed (0.92
If ice density on bulkheads and decks is also accounted
Mg m3) to loosely consolidated (0.69 Mg m3). Tex-
for, the ratio of mass per unit area increases the ratio to
turally, accreted ice clearly resembled frazil ice, formed
1.4:1 because the density of the ice on the decks was
from the consolidation of freely nucleated ice crystals
larger than that on the bulkheads. This ratio has impli-
in sea water. This resemblance is also reflected in the
cations for modeling superstructure icing, because
ratios and magnitudes of incorporation of brine. Bulk
higher levels on ships, where center of gravity is most
salinities ranged from to 7 to 25.4‰ compared to val-
affected by ice, are dominated by bulkheads. Ice on
ues of 67‰ measured in normal sea ice. However, in
horizontal surfaces (deck and hatch covers) also pro-
frazil ice formed on the surface of the ocean, which
duced fewer and smaller inclusions than did ice on ver-
frequently represents the initial mode of freezing of sea
tical surfaces (bulkheads).
30
to contents
Previous Page
