Geoscience Reference
In-Depth Information
maybe more commonthanhad beenpreviouslyappreciated.Our experiencewhile
deploying the SHEBA station in the Beaufort Gyre in September 1997, was that
when we drilled the late summer ice, we often encountered multiple layers of liq-
uidmeltwaterinterspersedbetweenthinice layers,suggestiveofseveralsuccessive
cyclesoffalse-bottomformationandmigration.Therewas alsoa significantdiffer-
encebetweenestablishinghydroholesduringfallversusearlyspring.Inthelatter,it
isoftenpossibletoextracticethatisdrytowithin10-15cmoftheicebottom,while
in the former,we encountereda “water table” relatively high in the ice column be-
lowwhichlateralwatermovementappearedtoberelativelyunrestricted(andmade
further ice excavation more difficult). The conceptof a porouswater table that mi-
gratesdownwardfromthesurfaceastheicecolumnwarmsfromaboveimpliesthat
any pre-existing concavities in the ice undersurface will be filled with fresh water
regardlessofa directverticalconnectionto thesurface.
Falsebottomsaffectthegeneralice-albedofeedbackissueintwoimportantways.
First, they may substantially delay the transfer of heat from the upper ocean to the
icepackbyreducing
St
∗
)
eff
,whichallowstheupperoceantomaintainitsheatcon-
tent well past the time when sun angles are high. Second, as fresh water begins to
collectinundericeconcavitiesearlyinthemeltseason,falsebottomformationpro-
tectsthethinnesticefromcontactwiththewarmingupperoceanthusdelayingexpo-
sureofopenseawater.Theice-albedofeedbackismosteffectivewhentheice/upper
ocean system can absorb solar radiation at times near the summer solstice. Both of
the false-bottom mechanisms described here tend to retard this timing, hence rep-
resent a perhaps important negative feedback in the system. A general thinning of
the perennial pack (Rothrock et al. 1999) will mean that summer warming and the
presumptive downward migrating “water table” will reach the ice base earlier in
the summer, hence reinforcing the mitigating impact of underice melt ponds and
falsebottoms.
(
6.7 Freezing—Is Double Diffusion Important?
Melloretal.(1986)andSteele etal. (1989)showedthatifdouble-diffusivetenden-
ciescarryovertofreezinginthesamewaythattheyapparentlyaffectmelting,then
thereoughttobesignificantproductionofsupercooledwater,becauseheatwouldbe
extracted from the upper ocean faster than salt would be injected. Presumably, the
supercooledwater would either nucleate in situ and form frazil ice crystals distrib-
uted in some way through the IOBL, or would nucleate more or less uniformly on
theiceundersurface,regardlessoficethickness.Steeleetal.(1989),usingexchange
parameters inferred from MIZEX measurements, estimated that supercooling and
subsequentfrazil productioncould accountfor as much as half of the ice accretion
forthin(20cm)iceto 30%for80-100cmthickice.
If only one thickness of ice is considered, over time it matters little whether
the ice forms from congelation at the immediate interface, or by accretion of
frazil crystals drifting up to the interface from below. However, with different ice
