Geoscience Reference
In-Depth Information
Another possible source of supercooling can arise from differential mixing of
salt and heat when there are large horizontal gradients in temperature and salinity.
Measurementsin 2007oftransientsupercoolingeventsinan energetictidalflowin
Fremansundet, Svalbard, were suggestive of this mechanism. The events occurred
at different times at two different levels when a front between slightly less saline
water from outside the fast ice was advected into the sound by the tide. The water
on both sides of the front was within millikelvins of freezing, so that the water
advected into the sound was slightly warmer. We interpreted the transient events
(each lasting about an hour) as the result of heat mixing more rapidly than salt, so
thattheincomingwatermassdippedbelowitsfreezingpointinthefrontalzone.Itis
perhapsworthnotingthatthesupercoolingnorthofSvalbardreportedbyLewisand
Perkin (1983) occurred in a region with strong horizontal gradients in temperature
andsalinity.
Given several possible sources of supercooling and subsequent frazil produc-
tion (but a lack of evidence that it occurs extensively under multiyear pack ice),
is it possible to examine in isolation the hypothesized mechanism of supercool-
ing associated with double diffusion at the interface during rapid ice growth? We
approached this problem as follows. Consider growth in thin ice in seawater at
freezing under with the following conditions:
u
∗
0
5mms
−
1
=
,
=
,
=
S
w
34psu
T
w
865
◦
C, with upwardheat conductionof 20Wm
−
2
in the ice column,
correspondingtoatemperaturegradientintheiceofabout
T
f
(
)=
−
.
S
w
1
10Km
−
1
.We assume
−
S
ice
=
7psu.Firstsolvetheinterfaceequationforsalinity(6.9),withnodoublediffu-
sion,i.e.,
α
h
=
α
S
=
0
.
0058(whichwasshownaboveto matchtheStantonnumber
constraint,
St
∗
=
067psu, the ice grows at a rate of
about 7mm per day, and under the assumption that ice salinity is 7psu, this pro-
duces a salinity flux
0
.
0057). In this case,
S
0
=
34
.
w
S
0
=
−
10
−
6
psums
−
1
, and an upward heat flux
1
.
96
×
4Wm
−
2
, which would be difficultto detectby covari-
ancemeasurement.It iseasy to confirmthat
from thewater column of 0
.
w
S
0
, i.e.,that heat is
extractedfromthewateratjusttheraterequiredtomaintainthewateratitsfreezing
temperature as salt is added at the surface. Note that none of these quantities are
extremebyanymeasure.
Next solve the problem with identical conditions except that we now let double
diffusion operate in the interface control volume at levels used in the false bottom
simulation(
w
T
0
=
−
m
859,con-
gelation growth is significantly reducedto about3.3mm per day, and now the heat
flux out of the water column is 10
α
h
=
0
.
0111;
α
S
=
α
h
/
50,seeFig.6.10).Inthiscase
S
0
=
34
.
7Wm
−
2
. This is easily measured, and thought
experiments like this convinced us that the best way to look for the supercooling
effect due to double diffusion at the freezing interface was by measuring upward
heatfluxinwaterjustbelowtheinterface.
From these considerations, we designed a field experiment in a relatively con-
trolled environment offered by fast ice and a gentle tidal flow in Van Mijen Fjord,
Svalbard. In March 2001, we occupied a site on smooth fast ice, and installed in-
strumentationtomeasureicecharacteristicsandturbulence1mbelowtheice/water
interface (McPhee et al. 2008, in press). Temperature profiles measured during the
fieldproject(Fig.6.13)showtheimpactofchangingsurfacetemperature(therewas
.
