Geoscience Reference
In-Depth Information
“frozen-field”hypothesisasdescribedinSection3.2.2providetheprimaryrationale
forestimatingverticalfluxesfromthecovariancestatistics.Toillustratethis,westart
witha1-htimeseriesofdata,adaptedfromMcPheeandStanton(1996),takendur-
ing the 1992 LeadEX project in the Canadian Basin. Although this example is not
typicalofwhatweusuallyobserveintheIOBL(othersarepresentedbelow),itwas
chosen because it accentuatesimportantaspects of the turbulentexchangeprocess.
At about the time of solar zenith on 7 April (year day 98), our TIC mast was posi-
tionednearthemiddleofthewellmixedlayeratthenorthedgeofakilometer-wide
lead, which had opened the previous day and was freezing fairly rapidly with air
temperatures in the
28 C range. The ice and lead were drifting south at
20 to
12ms 1 , so that in our moving reference frame, the current appeared to
comefromthesouth,acrosstheentireexpansethelead.Thedestabilizingbuoyancy
flux from salt rejected during freezing set up a boundary layer that appeared to be
dominatedbylargescale eddies(perhaps“rolls”)thatadvectedpastourinstrument
mast with some regularity. One of these “events” is indicated by shading of the
verticalvelocity(Fig.3.3a)showingdownwardmotionwith amaximumapproach-
ing 3cms 1 , andpersisting fornearly 5min. Duringthis time, temperature(b) and
salinity(c)bothdeviatein apositivesense fromtheir1-hmeanvalues.Forsalinity,
this is consistent with salt ejected at the interface enhancingturbulence by gravita-
tional acceleration. The large positive temperature anomaly indicates that, despite
rapid freezing and upward heat conduction through the ice, water near the surface
was being heated by incoming shortwave radiation that had penetrated the thin ice
cover,andcarrieddownwardinthe largeeddies.
In the shaded sample and several other events, there is obvious correlation be-
tween w and the deviatory scalars, quantified by forming the product series, as
shown for temperature and salinity in Fig. 3.3d and e, respectively. The product
series are characterized by a number of significant upward and downward excur-
sions, of roughly similar magnitudes, with peak values several times the 1-h mean
values(indicatedbythe dashedlines). In manycases a largeexcursionof the prod-
uct series is followed soon by an excursion of opposite sign, somewhat like one
mightexpectina wavefield wherethe verticalvelocityanddisplacementfieldsare
90 outofphase(theeventbetweenminutes28and31hassomewhatthisaspect).In
suchafield,themeanproductoveroneormoreperiodsiszero.Thisisperhapsone
way to visualize how the inertial instabilities we term eddies work. Somewhat like
a wave, most of the fluid carried up and down by the large-scale structures returns
to its former level, without too much changein its properties.Nevertheless, during
the large excursions there are smaller features (indicated by “fine structure” in the
timeseries) thatcontinuously“nibble”at thefluidwithinthelargeeddies,resulting
in a relatively small net exchange of propertiesacross the measurementlevel. This
thenshowsupinthecovariance(meanoftheproductseries)providedtheaveraging
timeisadequatetocaptureafairnumberofthelargeeddies.Averagedoverthe1h,
thecovariancessuggestadownwardheatfluxofabout70Wm 2 (roughly15%the
incomingshortwaveradiationattheuppericesurface),andsalinityfluxcomparable
tothat expectedmidwayin thewellmixedlayerifthe freezingratewasabout5cm
perday(McPheeandStanton1996).
about 0
.
Search WWH ::




Custom Search