Ocean advection, ArcticAtlantic Connections, Climate P.B.Rhines, University of Washington Sirpa Hakk

1
Ocean advection, Arctic-Atlantic Connections,
ClimateP.B.Rhines, University of
WashingtonSirpa Hakkinen, NASA Goddard SPCwith
David Bailey, Wei Cheng, Jerome Cuny, Trisha
Sawatzky WUN Climate teleseminar, 26iii2003

• North Atlantic Oceanic Heat Advection is
  Important to the Wintertime Storm Track, Eurasian
  climate and weather and Global Climate

www.ocean.washington.edu /research/gfd/gfd.html
2

• TABLE OF
  CONTENTS
• Polar amplification of global warming
• Storms and weather.ocean to stratosphere
  continuity of the
• Atlantic storm track/Icelandic Low from
  subtropics to Arctic
• Atmospheric and oceanic transports, strongly
  coupled, supply continental and polar warmth,
  fresh water
• The oceanic heat source, strongly channeled,
  local and imported heat
• driving the 3 scales atmospheric circulation,
• Transports by ocean circulation connecting the
  Arctic and Atlantic
• Erika Dan, 600N winter in the N Atlantic deep,
  shallow and
• shelf water masses
• Atlantic Arctic exchange T-S-transport
  (heat,
• fresh-water transport) diagrams
• Rapid and slow modes of response of Atlantic
  overturning
• Labrador Sea vs. Nordic Sea overflows
• What matters to the global overturning
  circulation
• The freshwater cap and its movement liquid and
  solid
• Layering of deep-ocean water masses, and its
• representation in models
• Shallow continental shelves and communication
  with

3

• Global surface temperature has seen two major
  warmings in the 20th Century in the 1920s-30s,
  when it was very concentrated around Greenland,
  and since the 1980s, when it is much more global,
  yet still concentrated in high northern
  latitudes.
• Cod and herring fisheries responded to the much
  warmer ocean temperatures, which lasted for more
  than 25 years.
• Ironically we anticipate the ocean circulation
  slowing during the current warming, whereas it is
  possible that an accelerated oceanic meridional
  overturning circulation was a factor in driving
  the 1920s warming.

4
Surface Air Temperature
DelworthKnutson, Nature 1999
Uppernavik, West Greenland SAT anomaly (Imke
Durre)
5
The standing-wave structure with wintertime
Icelandic and Aleutian Lows gives the atmosphere
some of the east-west structure and gyre
circulation familiar to the ocean. The rapid
radiative cooling in fall sets up a cold dome
which slumps under gravity, tending to create a
surface polar vortex which is anticyclonic, the
convergence overhead strengthening the polar
cylonic vortexyet mountains and vertical
momentum transport can intervene, opposing the
surface high. In particular, the Atlantic
storm-track forms a continuous connection
between subtropics and the high Arctic.
6
Surface air temperature shows the imprint of both
warmocean and the orographic component of the
circulation. The ice free region of the Arctic
and subArctic is marked by warmsurface
air. NCEP reanal 2 Jan 1993
7

• Standing Rossby wave models of the circulation
  suggest a cyclonic trough in the lee of the
  Rockies, which transmits wave track to the
  atmosphere
• The hemispheric baroclinity (cold air dome)
  provides energy for both standing and transient
  eddies, while ocean heating augments both (e.g.,
  Held et al., J.Clim 2002).
• The surface low lies east of the 500 mb. low in
  this winter JFM averaged 1993 data, expressing
  both poleward heat flux and diabatic heating by
  the ocean (note particularly the Labrador Sea and
  subpolar Atlantic).

8
(No Transcript)
9

• The atmospheric circulation is increasingly zonal
  with height, and variance-based maps of storm
  tracks are quite zonal in the upper troposphere
  yet the low-level tracks of storms fills out an
  extensive meridional gyre centered on the
  Icelandic Low. (Hoskins and Hodges, JAS 2002).
• Storms amplify at several sites along the track,
  as over the
• warm water of the Nordic and Barents Seas.
• The low-pressure storm centers circulate round
  the Icelandic Low and spill heat and moisture
  onto the Eurasian continent over a broad range of
  latitudes.

10
Lagrangian storm track density HoskinsHodges
JAS 02based on sea-level pressure (ECMWF
1979-2000 period). NAO/AO positive index is
prevalent during this period.
11

• Storm activity propagates remarkable distances
  round the northern hemisphere, with one storm
  seeding the next, with group velocity exceeding
  the speed of individual storms

12

• 5-week Høvmüller plot following the principal
  storm tracks (Chang, Lee and Swanson J Clim
  2002). This is variance of meridional velocity
  at 300 mb.

Atlantic
Pacific
13

• Heat flux between ocean and atmosphere is
  inferred from many sources atmospheric
  reanalysis fields (plus top-of-atmosphere
  radiation in the brief ERBE period), ship-board
  observations, aircraft boundary layer
  observations, satellite altimetery and
  thermometry, water-column heat storage, and ocean
  circulation.
• NCEP and ECMWF net heat fluxes appear to be less
  accurate than re-analyzed products guided by the
  aircraft observations.

14
NCEP Aircraft (Xue et al JPO95) NCEP
Smith-deCosmo
latent
sensible
RenfrewMoore, 2002
15

• Maritime storms grow explosively, and move
  along the Gulf Stream front and the lower
  atmospheric temperature front lying near the
  coast
• In mesoscale model simulations (Kuo, Reed and
  Low-Nam MWR 1991) of 7 explosive storms ocean
  heating caused SLP deepening (13.5 mb in 48 hours
  of 30-40 mb total deepening) for long forecasts
  begun when the storms were weaker.
• APE generation is
• -favored by latent heat in the warm sector
  (warming warm air),
• -damped by sensible heat in the cold sector
  (warming cold-outbreak air Orlanski et al.
  J.Clim.2002_

16
Laus JAS 1979 winter diabatic heating of the
700mb-1000mb lower atmosphere. Peak values are
100-150 watts/m2 in the subtropical storm track
regions
17

• This net atmospheric heating by the ocean, plus
  or minus radiation, is less than the surface
  upward flux, Qnet, inferred by most methods (as
  the atmosphere cools to space). The net surface
  heat flux peaks with winter-average values as
  great as 300 watts/m2 near the Gulf Stream and
  Labrador Sea. The Barents Sea in the Arctic is
  not far behind with peaks exceeding 200 watts/m2
• In the map that follows the annual-average
  plotted values of Qnet are less than these
  wintertime values, roughly by a factor of 4.
  This may be explained by the radiative cooling of
  this air mass to space, although accuracy of the
  observations is also unknown.

18
Qnet, net atmosphere-ocean heat flux, watts/m2
(Keith Tellus 95)(annual average)
It should be noted that because the sun heats the
ocean, O, but does not cool the atmosphere, A,
the most useful maps of Qnet for A will differ
those for O by the short-wave insolation.
19

• Keiths analysis uses top-of-the-atmosphere
  radiation observations from the ERBE satellite
  together with atmospheric reanalaysis data to
  infer the surface heat flux as a residual.
  Errors can immediately be seen in the large
  values over land, which cannot support
  significant net heat flux.

20

• The Arctic in winter is ice-free in the Barents
  Sea and in regions of the Arctic rim warmed by
  circulation of tropical heat as part of the
  oceanic MOC. The relationship of ice,
  fresh-water and heat is an essential part of the
  climate system. Shrinkage of ice-cover in both
  winter and summer has been observed, yet there
  are dynamical as well as thermodynamic causes
  (Rigor and Wallace, 2003).

21
Cryosphere today (Feb. 2003)
22

• Oceanic mixed-layer depth mirrors the Atlantic
  storm-track (here, Levitus records the depth at
  which the density is greater than the surface
  density by an amount equivalent to 0.50 C, which
  is deeper than the actual depth of wintertime
  mixing). Deep regions of weak stratification
  lie beneath the storm track.

23
Mixed layer depth (March, in m) based on 0.5C
24

• The evidence so far suggests a relationship
  between the heating originating in the oceanic
  overturning circulation, energizing the
  atmospheric storm track. Energy released from the
  polar cold-air dome through baroclinic
  instability is perhaps the primary source for
  transient eddy energy (Chang et al., J Clim
  2002). Held et al. (J Clim 2002) however argue
  that the oceanic heat source produces significant
  forcing for the standing wave energy in the
  northern hemisphere circulation, comparable with
  the Himalayan plateau/Rocky Mountain orographic
  source. Decadal NAO variability is another issue,
  of course, with the oceanic high latitude heat
  source being less prominent (e.g. Kushnir et al.
  J Clim 2002).
• The meridional heat transport curves, averaged
  zonally and in time, show this, yet with the
  atmosphere doing the majority of the work.

25
Trenberth J Climate 2001 Global meridional heat
transport (based on TOA radiation, atmospheric
reanalysis)
ocean atmosphere
atmosphere ocean
26

• Bryden and Imawaki, 2001, emphasize that the
  atmospheric contribution to meridional heat flux
  should be separated into sensible-heat and
  latent-heat components. 1 Sverdrup of water
  vapor (the scale amplitude of the hydrologic
  cycle) moving poleward carries 2.5 petawatts (2.5
  x 1015 watts) of thermal energy.
• They plot the break-down using Keiths 1995
  data.

27
Global meridional heat transport
atmos (latent)(residual method, TOA radiation
85-89 atmos atmos (sensible)and ECMWF/NMC
atmos obs) ocean (sensible)

Oceanic heat flux is convergent north of here
Ice-cover north of here
Atmospheric flux is divergent south of here
data of Keith (Tellus 95)
Error est. 9 at mid-latitude Bryden est
2.0 0.42 pW at 24N
28

• The latent heat flux is intrinsically a coupled
  atmosphere-ocean mode, with the fresh-water
  carried north by the atmosphere returning in the
  ocean (Bryden et al. 2001).
• A part of the poleward moisture flux switches
  from atmosphere to land, eventually draining in
  the massive outflow of Russian rivers.

29
Arctic river outflows (cubic km per year)
AagaardCarmack JGR 1989
Delivery of moisture to continental interiors
sensitivity to storm-tracks, NAO/AO (e.g.
Tigris-Euphrates, Cullen et al. 2001) Arctic
river outflows provide 4200 km3 yr-10.12
Sverdrup of fresh-water P-E, Bering Strait
inflow (0.8 Sv at 32.5 ppt) and distillation by
freezing supply the rest of the exported
fresh-water. (Vuglinsky 1997 ACSYS-Orcas Isl)
30

• Traditionally, a dominant atmospheric moisture
  pathway takes evaporation from the subtropical
  Atlantic and sends it eastward across Eurasia
  however with high NAO/AO index, a sizeable part
  of this moisture takes a more northern route,
  some of it entering the Arctic.

31
Impact of oceanic heat/fresh-water transport

• Despite its apparent smallness at higher
  latitudes, oceanic meridional heat flux is
  important
• Its divergence in the subtropics is the dominant
  source of fresh-water for the higher latitudes
• Heat transport is channeled geograpically in both
  ocean and atmosphere, and is seasonally enhanced
  (sensible heat in winter exceeds that in summer
  by a factor of about 2.5 (Trenberth 2001), and
  latent heat/moisture flux is more complexly
  enhanced in winter (Peixoto and Oort, 1992).
• Observations of ocean circulation have
  definitively altered the ratio of oceanic and
  atmospheric heat transports (see Bryden and
  Imawaki 2001), and we may not have heard the last
  word.

32
Atlantic meridional heat flux Sato Rossby JPO
2000
33

• Seasonal cycle how is the wintertime upward heat
  flux supplied?
• - locally stored summertime heating
• or
• - heat imported from the south by ocean
  circulation?

34
Atlantic north of 250N air-sea heat flux
calculated from reanalyzed COADS data of da Silva
et al. 1994. Black monthly heat loss Red its
cumulative sum starting in April
Monthly flux Reaches 4.5 pW
Heating of ocean (black) integrated from 1 April
returns to zero by early Dec (orange)
RhinesHakkinen 2003
35

• In the Atlantic north of 25N, on average the
  previous summers heating is withdrawn from the
  ocean by early December.
• Subsequently, December through March, the heating
  of the atmosphere by the ocean is supplied almost
  entirely by laterally advective ocean
  circulation.
• Thus it is the geography and seasonality
  (concentrated arteries, steadier in the ocean,
  cold-season in the atmosphere) that raises the
  profile of oceanic heat transport to dominate in
  winter.

36
(No Transcript)
37

• We can plot the year-day when last summers local
  heating is exhausted by fall-winter cooling.
• For orange areas this day falls in early December
  or earlier, suggesting that in these regions, for
  most of the winter, the oceanic warming of the
  atmosphere is entirely due to advection by the
  circulation.
• For blue areas it is later, often suggesting
  local one-dimensional closure of the heat storage
  with no need for advective heat-flux convergence.
  
• In the Southern Hemisphere the year-days are
  shifted by 6 months so as to have the same
  connotation.

38
by early Dec. local, seasonal heat storage is
used up here
In blue regions local 1-d Mixed layer heat
storage dominates
Year-day when seasonal heat is used up ( 6
mo in SH)
39

• There are several other ways to reach the same
  conclusion, for example mapping the ratio of
  seasonal cycle to annual mean heat flux
  divergence. Data sources range from da
  Silva/Levitus (1994) reanalysis of COADS, with
  calibration by meridional heat transport inferred
  from oceanic hydrography, and with constraints on
  net global vertical flux.
• As an example, Seager et al. 2002 describe some
  0.8 petawatts of heat transported northward
  across 35N in the Atlantic Ocean. On average this
  produces 37 watts/m2 of heating of the atmosphere
  over the ocean surface to the north. Yet the
  wintertime average Qnet is 135 watts/m2
  substantially larger.
• However the fraction, say ?, of this
  ocean-circulation heat flux lying deeper than
  about 100m has to wait for winter with its deep
  mixed layers to escape. Its contribution to the
  wintertime upward heat flux is about 4 ? (taking
  winter to be 3 months long in this sense). Thus
  maps of ocean-atmosphere heat flux (as seen by
  the atmosphere) should compare wintertime heat
  flux Qnet with annual-average ocean heat-flux
  convergence maps multiplied by 4 ?. If ? 0.5
  we are comparing
• 2 x 37 74 watts/m2 from ocean circulation
  with 135 watts/m2 observed wintertime Qnet. In
  this case about 1/2 of the surface area of the
  northern Atlantic can be fully supplied with heat
  in winter by the lateral circulation, after
  locally stored mixed-layer heat is removed.

40

• Thus the conclusion is that warming and
  moistening of the atmosphere by the circulating
  ocean in middle latitudes as well as tropics is
  crucial to 3 scales of atmospheric motion and
  climate winter storms, the stationary winter
  gyres of sea-level pressure (Icelandic and
  Aleutian Lows), and the zonally averaged poleward
  transport so vital to maintain equable
  temperatures and rainfall at higher latitude.
• While this may seem obvious to many
  oceanographers, it has become a point of
  contention in the literature

41
North Atlantic heat balance vs. month and
latitude CartonWang jpo 2002ABSTRACTHere,
seasonal heat transport in the North Pacific and
North Atlantic Oceans is compared using a
49-year-long analysis based on data assimilation.
In midlatitudes surface heat flux is largely
balanced by seasonal storage.

• Surface flux Qnet
• (daSilva 94)
• Heat storage rate
• Ocean heat transport divergence

Small but it accumulates
for release in winter
Contour interval 2.5 x 108 w m-1
42

• Several recent papers all argue that oceanic
  heat storage is local, and requires only the
  thin upper mixed layer to describe it. The
  advocates often cite Gill and Niiler (JPO 1973),
  who do a scale analysis based on large ocean
  gyres with weak mid-ocean currents (and disclaim
  any conclusions about more active western
  boundary current regions). Wang and Carton (JPO
  2002), Seager, Battisti, Gordon, Naik, Clement
  and Cane (QJ Roy Met Soc 2002), Voison and Niiler
  ( JPO 1998) are among these.
• Seager et al. state in their abstract
• Further, analysis of the ocean surface heat
  budget shows that the majority of the heat
  released during winter from the ocean to the
  atmosphere is accounted for by the seasonal
  release of heat previously absorbed and not by
  ocean heat-flux convergence. Therefore, the
  existence of the winter temperature contrast
  between western Europe and eastern North America
  does not require a dynamical ocean.

43
Wintertime (DJF) average atmosphere-ocean heat
flux
From Seager et al. 2002 who use da Silva et
al. 1994 COADS derived Qnet data
Annual mean atmosphere-ocean heat flux
Smaller but it accumulates
44
Effect of ocean heat transport convergence on
GISS T42 atmosphere (Seager et al. 2002)
45

• The geographic pattern of wintertime
  ocean-to-atmosphere heating by itself suggests
  involvement of the ocean circulation it seems
  unlikely that a non-dynamical mixed-layer ocean
  could exhibit the large values coinciding with
  the northward track of the major currents.
• Northward heat transport by ocean circulation
  would be easy to identify if it were all well
  beneath the seasonal thermocline simply look for
  deep mixed layers in winter yet it is
  partitioned between deep and shallow levels.
• Deep mixed layers are a sufficient but not
  necessary symptom of dominant oceanic heat
  transport.
• Direct observations of ocean currents and
  hydrography complement local water-column heat
  storage and altimetric observations in
  constraining heat transport. In these studies,
  oceanic heat advection has been shown to be
  dominant in interannual variability of the heat
  budget in the Gulf Stream/Sargasso Sea (the
  Bunker-Budyko Bullet) (Dong and Kelly, JPO
  submitted ix2002).

46
Winter convective mixing has reached 110m (mild
winter so far)
Igor Yashayaev, BIO Canada
47
Bravo mooring heat content annual cycle by layer
Lilly et al. JPO 99
deep warming due by ocean heat transport
48
Heat content and fresh-water content of the
oceanic water column, Bravo site, central
Labrador Sea (1964-74 and 1990-98, Lilly et al.
JPO 1999) the annual heat storage range is 2.5
to 3 GJ/m2 offsets between 1960s and 1990s
reflect the NAO/AO period of intensely cold
winters and declining ocean salinity.
49
Heat content in central Labrador Sea, 1997-8
(ECMWF flux integral, mooring, acoustic
tomography) Fischer et al. IFM Kiel Germany
Annual cycle is close to 2.5 GJ of this, about 1
GJ is bound up In the advective, annual mean
Note time of maximum heat content lagged 3 months
relative to SAT and integrated air-sea heat flux
and 6 months relative to insolation
50

• The vertical structure of meridional heat
  transport is not unique because of the uncertain
  choice of reference temperature (though the heat
  transport divergence is unique) yet plotting the
  temperature transport, ?CPTv, integrated
  zonally across the Atlantic, is of interest. The
  following plot from the Hakkinen model suggests a
  substantial transport beneath the upper mixed
  layer (note the shallow southward Ekman driven
  transport between 35N and 50N).

51
Annual-mean meridional temperature transport
vertical structure (terawatts per m) Hakkinen
Atlantic/Arctic model
52

• Impact of the extensive meridional extent of the
  Atlantic storm-track on the warming and
  moistening of Eurasia, and its decadal
  variability Thompson and Wallace (J.Clim 2000)
  plot the 30-year trend in temperature advection,
  which closely resembles the pattern associated
  with the NAO/AO positive phase strong eastward
  advection implicating the N. Atlantic and Arctic
  ocean sources

53
30 year trend in advection of time-averaged
winter temperature (925-500 Hpa av.)..anomalous
velocity and advection, cooling the ocean while
warming the land in both Atlantic and Pacific
sectors. Could these oceanic advective heat
sources be the root cause of this contribution to
global warming over Eurasia?
(from ThompsonWallace J Clim 2000)
54

• Heat and the hydrologic cycle are coupled in the
  atmosphere by latent and sensible heat flux
• In the ocean the coupling is evident in the
  potential temperature/salinity relation that
  defines water masses and their density (Stommel
  and Csanady, JGR 1980). Many facets of the MOC,
  are involved
• Buoyant low-salinity water can mediate the deep
  convection process, and modulate the MOC. ..the
  Arctic as an estuary
• Evaporation/precipitation and freezing/thawing
  and run-off control these upper-ocean effects
  while also affecting atmospheric climate
• Sea-ice cover divides regions of weak and strong
  air-sea interaction a strong climate feedback.
• Polar amplification of global warming even
  though mid-latitude ocean circulation is expected
  to slow, high-latitude circulation is predicted
  to increase as ice-cover diminishes (Holland and
  Bitz, J Clim submitted).

55
Erika Dan temperature section, 600N
Labrador-Greenland-Rockall-IrelandWorthingtonWri
ght, 1970
warm, saline water moving north from the
subtropics
deep winter mixing, sensitive to upper ocean
low-salinity waters
shallow continental shelf circulation (unresolved
in CCMs)
deep overflows from Denmark Strait/I.S. Ridge
56
Stommels cooling spiral at OWS Juliette (52.5N,
20W) is a symptom of upward Qnet
Plot of (u,v) velocity components with depth as a
parameter from hydrographic data Stommel,
PNAS 1979
57
Hudson Strait salinity at 50m August 1965
(J.Lazier)This is part of the only 3-dimensional
hydrographic survey ever made of the
Labrador/Irminger seas.
Northward moving Greenland shelf water
Cuny et al. 2003
Southward moving low- salinity water from Baffin
Bay
Irminger Current water circumnavigating the
Labrador Sea
58

• Earlier we described the coupled heat- and
  fresh-water transport of the atmosphere in the
  ocean the potential-temperature/salinity relation
  is remarkable in its stability and simplicity
  round the Earth. If we could plot volume
  transport as a 3d variable on this plane it would
  express the coupled heat- and fresh-water (a.k.a.
  salinity anomaly) transport for the ocean.
• For a section across a strait (Davis-, Fram-)
  these plots sum up the movement of water masses.
  If the strait enters a cul-de-sac (or
  approximately does, as with the AR7W section from
  Labrador to Greenland, across the Labrador Sea)
  then the transport/T/S diagram sums up conversion
  by mixing and air-sea interaction as water masses
  enter at on region of T/S and leave at another.
• The heat-flux is expressed by taking this
  diagrams transport density and multiplying by
  ?CpT (the 1st-moment with respect to T, nearly).
  Fresh-water flux is expressed by multiplying the
  transport by (S-So). The effect of different
  choices for So is readily observed through shifts
  on the diagram.
• Finally, the familiar overturning diagram,
  transport vs. potential density, s, is found by
  summing the transport/T/S along curved s
  surfaces. This is done in model examples below.

59
The Cuny et al. 2003 transport through Davis
Strait in temperature-salinity space blue bars
are southward flux, peaking in shallow Arctic
low-salinity water red bars are northward flux
of Irminger (warm, saline) water from the
Labrador Sea and farther south.
60

• The shallow continental shelves are particularly
  active parts of the transport, mixing and
  water-mass modification of ice-laden, buoyant
  surface waters. Yet climate models do not resolve
  continental shelves!
• Transports of low-salinity water on the Labrador
  and Greenland shelves are not well-known, but
  estimates suggest for the Labrador Current, seen
  in the earlier satellite ice-image, at least 0.2
  Sverdrup fresh-water (relative to 34.8 ppt),
  with volume transports of roughly 3 Sverdrups
  (Loder et al., The Sea, 1998). This is a crucial
  part of the Arctic estuary.

61
(No Transcript)
62

• Observations are wanting in the important regions
  that connect Arctic and Atlantic. The ASOF
  program is seeking to instrument the Arctic Rim
  with sustained moorings, hydrography, tracers and
  supporting satellite observations. In the
  Canadian Archipelago, for example this is a
  challenging problem due to ice-cover, difficult
  access, and a vast complex of passages.

63
ASOF (Arctic-SubArctic Ocean Flux)
www.asof.npolar.nomoored array sites in Canadian
Arctic Archipelago
Melling, in FW Budget of the Arctic, E.L. Lewis
Ed. 2001
64

• We need better climate models and ocean
  circulation models. High-latitude processes are
  not well represented, or altogether absent. Here
  is a meridional section from a new climate model
  built by Wei Cheng and Rainer Bleck, based on the
  MICOM ocean and CCM3 atmosphere, with a
  thermodyanamic ice model. Note how the density
  surfaces lift radically up topographic slopes
  where there are deep boundary currents. This and
  the attendant sinking, passage flows and deep
  convection are represented differently than in
  oceanic level-models.

65
Cheng et al. Clim Dyn subm 2003
Atlantic subpolar gyre
66

• Entrainment and mixing occurs as dense waters
  sink to the deep ocean. In this experiment from
  our GFD lab, you see dense water sinking at the
  right the colors are injected into the sinking
  branch at intervalsthey are time-lines. The
  sinking branch entrains water, driving a
  subsurface overturning cell. This is realistic,
  in that the transport of the observed Atlantic
  high-latitude overflows more than doubles on its
  way round the subpolar gyre.
• In most numerical ocean models, mixing is too
  large and sinking too inefficient (unless one
  tries to fix it with a Beckmann-Doescher-type
  benthic layer). We need to preserve the
  articulate stratification of water masses that
  make up the huge North Atlantic Deep Water mass.

67
boundary mixing/ upwelling
entrainment
broad upwelling, yet much of it recirculates
below the mixed layer
68
The remainder of the talk describes analysis of
the Häkkinen ocean/ice model carried out by David
Bailey at UW 20 level s-coordinate 0.70 x
0.90 resolution Arctic Atlantic to 350S where
relaxed to Levitus NCEP forcing for
1950-2000 elastic-plastic ice model specified
Bering Strait flux smoothed topography (Bailey
et al. 2003)
mean barotropic transport
69
The meridional circulation penetrates farther
north in density space Than in the physical y-z
plane.

• MOC vs y,z
• MOC vs. y,s

70

• If the volume transport measured on an oceanic
  section is plotted on the T-S (potential-temperatu
  re/salinity) diagram, it shows individual
  water-mass contributions between sections it
  shows water-mass conversion

71
Red shades are transport into the Labrador Sea,
blue shades exit the Sea Water mass conversion
is represented by the movement toward fresher,
colder (red blue)
Labrador Sea cross-section AR7W line 15
Sverdrups of conversion (Bailey et al. 2003)
72
Labrador Sea inflow/outflow/conversion
Greenland-Scotland Ridge overflows
73
Here at 25N one has a wind-driven gyre with both
northward and southward transports at low
density. The deep southward flows are prominent
at the highest densities.
Sargasso Sea 25N
74
Transports at various sections in Hakkinen model.
The models Labrador Sea responds quickly to
atmospheric forcing and strongly influence the
MOC at 24N. The overflows from Iceland-Scotland
Ridge evolve more slowly, as has been observed.
75

• The headwaters of MOC formation involve dense
  overflows from the Greenland-Iceland-Norwegian
  seas, and the mid-depth waters from the Labrador
  Sea. These multiple sources produce a
  well-stratified North Atlantic Deep Water that
  occupies much of the Atlantic between 1200m and
  5500m depth.
• Models need to find the right balance between
  these sources, yet diapycnal mixing resulting
  from resolution contraints make this difficult to
  achieve.
• Cold winds from Arctic Canada drive deep
  convection in the Labrador Sea, but buoyant
  low-salinity waters mediate and weaken the
  convective deepening again a challenge to
  modelling.

76
These plots show how much buoyancy must be
removed from the top of the Labrador Sea to yield
convection to a given depth. The contributions
of temperature and salinity to this
resistance-to- convection are shown in red and
green low salinity near the surface dominates
the stratification.
Blue total buoyancy flux needed Green
salinity contribution to buoyancy barrier Red
thermal contribution to barrier
CSS Hudson cruise, May 1994
77
Buoyancy barrier to convection 0-500m difference
between thermal and haline components the
low-salinity cap is stronger in the observed
Labrador Sea yet weaker in the observed Greenland
Sea. Levitus November
Hakkinen model November
78
Conclusions

• Atmospheric storm-tracks, stationary winter
  low-pressure centers and hemispheric circulation
  all show the imprint of thermal ocean forcing.
  The Atlantic storm-track continuously connects
  the subtropics with the high Arctic, and its
  non-zonally oriented image extends vertically
  from beneath the oceanic mixed layer up to the
  stratosphere.
• Air-sea heat-flux observations, direct inference
  of oceanic meridional heat transport from
  hydrographic sections, and the residual method
  (TOA radiation minus atmospheric analysis) all
  give heat-flux convergence which, if released
  substantially in winter, accounts for most of the
  upward wintertime heat flux in the western
  subtropical Atlantic, the entire subpolar
  Atlantic and ice free regions of Labrador, Nordic
  and Barents Seas.
• Fresh-water and heat are transported
  interactively in both atmosphere and ocean at
  high latitude particularly the movement of
  buoyant low-salinity water and changing ice-cover
  exert strong climate feedbacks. A mixed-layer
  ocean is not sufficient for such climate studies.

79

• The vertical structure of the oceanic heat
  transport determines the seasonal delivery
  curve and its partition between local
  mixed-layer heat storage and storage of advected
  heat.
• The water-column heat-storage annual cycle ( 3
  GJoules/m2 in Labrador Sea, Gulf-Stream hot
  spots) has the potential to more accurately give
  Qnet (based on sustained time-series of in situ
  hydrography or satellite altimetry).
• We have ignored seasonal modulation of the ocean
  heat transport, for example in the wind-driven
  Ekman layer, believing that its divergence will
  be important mostly in frontal regions

80

• Coupled heat- and fresh-water transport occurs in
  both ocean and atmosphere analyzing ocean models
  one finds that in the northern Atlantic both
  shallow and deep outflows from Arctic and
  Labrador Sea shape the meridional overturning and
  its thermodynamic transports.
• Oceanic transport diagrams drawn on the T-S
  plane express the heat/fresh-water interaction,
  show water-mass movement across sections, and
  water-mass conversion between sections. They help
  to diagnose, in the example given here, the loss
  of identity of high-latitude (Greenland-Scotland-R
  idge) overflows and the strong role played by the
  Labrador Sea in the Häkkinen model. In
  particular, shallow low-salinity waters from the
  Arctic strongly resist deep convection and their
  model representation is an ongoing challenge.

81
The solution? Instrument the Arctic Rim. Do
ASOF. Deploy the Eriksen Seaglider
82
(No Transcript)
83

• THE END