Geology Reference
In-Depth Information
spatial distribution is crucial. It is also required to initiate
thermodynamics and climate models over polar regions.
As a first approximation, some models adopt the crude
assumption of constant salinity. Other models assume
idealized time‐independent salinity profiles. The best
input, however, would be time‐varying salinity profiles
(as ice ages) that reflect also the spatial variation. This
information, however, is usually not available. Empirical
equations that relate the bulk salinity with the ice thick-
ness can be useful.
Surface salinity is the value found within the top few mil-
limeters of the ice. It affects the reflection of solar radiation,
emission of the thermal infrared radiation (TIR) and
microwave radiation, and the scattering of radar signal.
Bulk salinity is the average salinity over the entire thickness
of an ice sheet. Therefore, it is a function of ice thickness.
The bulk salinity of natural sea ice, subjected to unidirec-
tional growth, is significantly high at the initial stage of for-
mation. It decreases at a relatively fast rate during the early
ice growth [
Sinha and Nakawo,
1981], then at a rather slower
rate as ice grows and even after the growth stops (brine
drainage mechanisms are explained in section 2.3.3.2).
While surface salinity is directly influenced by atmos-
pheric temperature and precipitation, bulk salinity and
salinity profile are affected by the growth and melting
conditions of the ice. Therefore, they are indirectly influ-
enced by atmospheric temperature. Ice salinity is a mani-
festation of the temperature influence on sea ice. In
summer, surface melt water drains through sea ice above
freeboard, flushing away brine especially near the surface
[
Weeks and Ackley,
1982]. This decreases the salinity near
the surface significantly and to a lesser extent throughout
the entire ice thickness. Investigations on salinity profiles
and bulk salinity have been conducted in numerous field
experiments on different ice types and growth stages.
Unlike seasonal ice, salinity of MY ice is nearly steady
with values that vary between 0% and 2‰.
study concludes that thin ice (<5 cm thick) has bulk salinity
as high as 25‰, but the salinity decreases rapidly to about
6‰ when ice thickness reaches 50 cm. The decrease con-
tinues though at a reduced rate to about 4‰ for ice more
than 150 cm thick. In the beginning of the melt season
another tangible decrease in the ice bulk salinity to about
2‰ is observed. If the ice survives the summer to become
second‐year ice it will start with salinity less than 2‰ at
the surface.
Widely used equations that relate bulk sea ice salinity
to its thickness were developed by
Cox and Weeks
[1974],
based on data obtained from FY ice and MY ice in the
Beaufort Sea in winter. Two linear regression equations
are applied to two ranges of ice thickness:
S
14 24 19 39
.
.
h
for
h
04
.
m
(3.8)
B
i
i
S
788159
.
.
h
for
h
04
.
m
(3.9)
B
i
i
where
S
B
is in ‰. Equation (3.9) is a fit of FY and MY
ice data. The study found that the correlation coefficients
for the above two regression equations are −0.78 and
−0.94, respectively. It also suggested that the pronounced
break in the slope at 0.4 m is due to a change in the domi-
nant brine drainage mechanism from brine expulsion to
gravity drainage.
Cox and Weeks
[1974] regressed their
data also into a single exponential equation as many nat-
ural phenomena tend to follow the exponential trend.
Using salinity measurements from the Arctic and
Antarctic sea ice,
Kovacs
[1996] developed another set
of single monotonic equations relating bulk salinity to ice
thickness:
S
4 606 91 603
.
.
/
h
for Yice
F
(3.10)
B
i
2
S
180 99810 5
.
.
/
h
for MY ice (3.11)
B
i
Applications of these equations are limited to the polar
regions where cold winter conditions are nearly steady.
They cannot be used for sea ice in southern latitudes
where freeze‐thaw cycles are usually encountered in
winter. The equation for the thickness of undeformed
FY ice [Equation (3.10)] along with validation data from
Arctic and Antarctic measurements are plotted in
Figure 3.6. The correlation coefficient is 0.73 and the
standard deviation of the measurements from the empir-
ical relation is 1.47.
Kovacs
[1996] pointed out that the
variation in bulk salinity data from leveled FY ice grown
in the Arctic and Antarctic is similar. The errors in the
data (FY and MY ice) shown in Figure 3.6 are well
within the experimental error. This is in spite of the
difference in seawater surface salinity (within the top
50 m) between about 34‰ of the Southern Ocean and
about 32‰ of the Arctic Ocean The author stated that
3.2.1. Bulk Salinity
The three main factors that influence the bulk salinity
of sea ice
S
B
are: (1) the initial growth conditions of the
ice sheet (faster growth retains more salts in the ice),
(2) air temperature record during the lifetime of the ice
sheet (higher temperatures result in more brine drainage
and therefore less salinity), and (3) the porosity of the ice,
which depends on the crystallographic structure (brine
tends to accumulate along subgrain boundaries of ice
crystals as explained in section 2.3.3.2). As ice thickens,
its growth slows down and therefore less brine continues
to be added. On the other hand, the brine drainage con-
tinues. The net result is a reduction of the bulk salinity.
This is reported in
Kovacs
[1996] from measurements of
FY ice salinity data in the Arctic and the Antarctic. The
