greater than on vertical surfaces in the February icing
mates. The derived brine volumes and air contents of
event, but there was no significant difference in March.
the accreted ice are calculated for the in-situ tempera-
Kultashev et al. (1972) found the density of ice on
tures measured at the time of sampling (Table 2, Fig. 9).
Soviet fishing trawlers to vary between 0.71 to 0.967
Total porosity and its constituent air and brine vol-
Mg m3. Tabata et al. (1963) found densities on four
umes varied considerably between ice on vertical and
ships, totaling 121 samples, to vary between 0.62 and
horizontal surfaces, and between the February and
0.94 Mg m3, though there was no systematic change
March icing events. In the February event, total
of density with location.
porosities were generally larger on vertical surfaces
(Table 2). Ice on all surfaces was cold in the February
event, and samples were taken hours after spraying had
The salinities of melted samples were measured with
ceased. With one exception, total ice porosity in the
a temperature compensating Beckman Solubridge. All
March event was much greater on horizontal surfaces
salinities were corrected to a reference temperature of
than on vertical surfaces (Table 2). During sampling,
25C. Measurement precision is estimated at 0.2‰.
most ice in the March event was very wet because of
Mean ice bulk salinity on horizontal surfaces was
continual splashing and warm weather, which may have
about 810‰ larger than on vertical surfaces in both
contributed to the generally higher porosity on the
icing events (Table 2). Ice samples were taken on the
ship whenever there was an opportunity, thus the length
A larger proportion of pores is filled with brine on
of time that brine had drained since the last splashing
horizontal surfaces than on vertical surfaces in both
by sea water is unknown. However, during several sam-
events (Table 2). Brine drains more readily from ice
pling excursions in the March spray event, spray was
pores on vertical surfaces, which then fill with air. A
lofted over the ice sampling team as ice was being
larger proportion of March ice pores are filled with brine
removed. In most cases, the ice was no more than a
than are February pores. This may be the result of the
few hours old since the last ice accretion had occurred.
higher temperatures in the March event (Fig. 10).
Panov (1972) found generally higher ice salinities from
similar sea water salinities on a Soviet trawler than were
MICROSTRUCTURAL STUDIES OF
measured on the USCGC Midgett. On both horizontal
and vertical surfaces, 10 to 12 hours into an icing event,
Panov's salinities ranged from 10.3 to 37.5‰.
Few studies have examined the crystalline structure
of ice created from bow spray. Ono (1968) sectioned
A significant portion of spray-accreted saline ice
ice removed from handrails of the patrol boat Chitose
often is composed of unfrozen water trapped within the
and photographed crystals through crossed polarizers.
ice matrix (Gates et al. 1986, Makkonen 1987). It is
He found tiny crystals, about 0.5 mm in diameter, with
called "spongy" ice; wind tunnel experiments have dem-
random orientations. The only other study, conducted
onstrated that unfrozen water contents can total up to
by Golubev (1972), was quite thorough, and examined
50% of the mass of the deposit.* If the unfrozen brine
the relationship between crystal structure and orienta-
eventually drains, as happens quickly during and
tion of the icing surface, substrate material, distance
immediately after accretion and more slowly later, the
within the ice from the substrate, and air temperature.
mass of the accretion will be considerably reduced
Golubev's work was done on a medium size Soviet fish-
below what thickness alone may suggest.
ing trawler in the Sea of Japan.
Air and brine volumes, collectively termed total
In this work, several ice structure characteristics were
porosity, of ice samples removed from decks and bulk-
examined, including ice crystal shape, size, and orien-
heads were computed using equations developed by Cox
tation, and inclusion size and shape. The February
and Weeks (1983) and Frankenstein and Garner (1967).
icing event is the least difficult to analyze because tem-
The volumes computed are estimates, as the equations
peratures were low during the entire period. The warm
were developed for floating sea ice, not spray ice, and
weather late in the March event complicates understand-
the answers rely upon density, salinity, and tempera-
ing of ice characteristics during that second event (Fig.
ture measurements that, with all potential errors con-
sidered, could cause uncertainty in the porosity esti-
Samples were prepared for microstructural analysis
by freezing portions of the superstructure ice onto glass
slides and then reducing them to the requisite thick-
ness on a microtome. Initially, these sections were
*Personal communication with Edward Lozowski, Univer-
thinned to 12 mm thickness and then photographed.
sity of Alberta, 1999.