Geoscience Reference
In-Depth Information
Fig. 6.13
Ice temperature profiles on three days infast ice on VanMijen Fjord, Svalbard (see also
Colorplate on p. 207)
little snow), but show a gradient in the lower 10cm or so of the ice column that
would indicate upwardheatconductionthere of about21Wm
−
2
. Turbulencemea-
surements are summarized in Fig. 6.14, where the turbulence data have been bin
averagedaccordingtomeancurrentvelocity1mbelowtheinterfaceineach15-min
turbulencerealizations. A least-squares regression through the origin is quite close
tothelawofthewallforahydraulicallysmoothboundary(Hinze1975).
log
u
∗
0
u
1m
u
∗
0
=
1
κ
ν
+
4
.
9
The regression lines in Fig. 6.14a and b show a slight correlation between flux
magnitude and current speed, but are barely distinguishable from zero at the 90%
confidence level. Overall the conditions are not much different from the example
presentedabove,and the measuredheatflux is only slightly more than what would
be required to keep the well mixed layer near freezing as it became saltier. The
lack of much ocean heat flux in this controlled environment is probably the most
convincingevidencethatduringfreezingdoublediffusionisrelativelyunimportant,
in contrastto melting. By applyingnumericalmodelusing localturbulenceclosure
(described in Chapter 8) to longer term measurements from the VMF experiment,
we showed that the exchange coefficient had to be close to unity during the VMF
2001exercise(McPheeetal.2008,inpress).
