Open ocean deep convection

1
Open ocean deep convection

• Modelling issues
• Effects on meridional transports

2
Major sites
Labrador Sea - only site in direct contact
with NADW Greenland Sea Mediterranean Sea
- deep waters remain local to the marginal
seas (no direct link to the outflows)
OMIP-forced global ocean-ice model
(OPA, 0.5 deg resolution)
3
Labrador Sea convection in models
1983
1983
MICOM 1/12 deg (Paiva et al., 1999)
ATL6 1/6 deg (CLIPPER project)
1989
1989
POP 1/10 deg (Smith et al.,2000)
ATL6 1/6 deg (CLIPPER project)
Figures Anne-Marie Treguier, 2003
4
Lavender et al. (2000)
5
Phases of open ocean deep convection
(Marshall and Schott 99)
a preconditioning b deep convection c
lateral exchange and spreading
Figure 3. Schematic diagram of the three phases
of open-ocean deep convection (a)
preconditioning, (b) deep convection, and (c)
lateral exchange and spreading. Buoyancy flux
through the sea surface is represented by curly
arrows, and the underlying stratification/outcrops
is shown by continuous lines. The volume of
fluid mixed by convection is shaded.
6
Labrador Sea (schematic)cyclonic
circulation and preconditioning ofthe
convection regime(Marshall and Schott 99)
7
Convective plumes

• Observations (e.g., Schott et al. 1994)
• Width of O (1 km) or less
• Send and Marshall (1995)
• Dominant effect is vertical mixing
• Klinger, Send and Marshall (1996)
• Effect can be represented by a vertical mixing
  scheme
• Simple convective adjustment is perfectly
  adequate

however Canuto et al. (Ocean Modelling.,
2004) strong differences between global
model behaviors using different schemes
for turbulent mixing
?
8
Restratification cylinder collapse
experiments, e.g., Jones and Marshall (1997)
rim current eddies efficient
restratification mechanism ( for convective
patches of 20-100 km with a strong rim current)
Figure 43. Numerical illustration of the
baroclinic instability of a cylinder of dense
fluid,of depth 1500 m and radius50km in an
ambient fluid in which N/f 5. The plan view'
panels on the left chart the development of a
passive tracer at the base of the cylinder at a
depth of 1400 m at (from top to bottom) 5, 10,
35, and 50 days.
9
Observed eddy fields in the Lab Sea
e.g., Heywood et al. (1994), Fratantoni (2001),
Cuny et al. (1997), Lilly et al. (2003), Brandt
et al. (2004)
Brandt et al. (JPO,2004)
Mean EKE (altimetry)

• Enhanced EKE off Greenland
• associated with warm-core
• eddies shed from the WGC
• propagation into convection
• region
• seasonal and interannual
• variation

1997 1998 1999 2000 2001
10
Eddy fields in high-resolution models

• Example
• 1/12 deg FLAME
• Atlantic model
• Eden and Böning (2002)
• Eddies spawned from
• WGC instability area
• Temporal modulation
• associated with strength
• of the boundary current

11
Catsman, Spall and Pickart (JPO, submitted
2003) Boundary current eddies and their role
in the restratification of the Labrador Sea
Spin-down of cone-shaped convection regions
rim current eddies not effective
boundary current eddies dominate able to
balance a significant portion of the heat
loss
12
Spall (JPO, 2004) Boundary currents and water
masstransformation in marginal seas
Spatially uniform cooling of a circular basin
(750 km) connected to an open ocean (providing
a heat source)
Equilibrium solution Heat loss in the
interior balanced by lateral eddy fluxes
originating in the boundary current
13
Snapshots at the end of the cooling
periodupper level temperature
mixed layer depth
14
Relative role of boundary vs. rim current
eddies?
Experiments with FLAME 1/12 suggest two sources
of EKE

• (Lars Czeschel, Ph.D. study)
• Response to enhanced winter
• heat fluxes
• stronger convection,
• higher interior EKE
• Response to increased
• wind forcing
• stronger WGC and EKE,
• reduced convection

15
EKE increase due to changes in heat flux and wind
stress(corresponding to NAO 3)
WIND
HEAT
EKE (averaged over LabSea
(0-400m) Reference case NAO3 (HEAT) NAO3 (WIND)
(Lars Czeschel, Kiel)
16
Changes in mixed layer depth
HEAT (NAO3)
Reference case
6.2
WIND (NAO3)
Mixed layer volume (105 km3)
4.7
3.8
(Lars Czeschel, Kiel)
17
Salinity section(through an eddy in FLAME
1/12)surface water from fresh WGCinterior
water from Irminger C.
Boundary current eddies important for the
preconditioning!
18
Issues

• Preconditioning
• Intricacies of source water budgets
• saline North Atlantic (Irminger C.) vs.
• fresh Arctic (WGC) waters
• Lateral exchange with boundary current
• instability process very localized
• Restratification and export of newly formed LSW
• What eddies?
• Role of mean boundary current and
  recirculations?

19
Rapid drainage of newly formed LSWby the
boundary current/recirculation
Thickness of LSW layer in April (FLAME 1/12)
20
Tiefenwasser-Export in den subtropischen
Nordatlantik
CFC-11 concentrations in the LSW layer (1997)
Observations
Model
Rhein et al. (2002)
Böning et al. (2003)
A4, A6, A7
?
21
Many details where models might get wrongbut
does it matter?

• What are the net effects
• i.e., the corresponding changes in the
• large-scale transports of mass and heat?
• (a) of model-model differences in convection?
• (b) of interannual variability in convection?

22
(a) Effect of model-model differences in the
representation of convection

• Model intercomparisons, e.g., Chassignet et al.
  (WOCE North Atlantic Workshop, 1999)
• Model sensitivity studies, e.g., different
  lateral
• mixing parameterizations
• No correspondence between convection and MOC

23
Example Two model versions (FLAME 1/3) using
the same forcing, with different lateral
mixing paramererizations
horizontal mixing
isopycnalGM
MOC 14 Sv
17 Sv
(FLAME group, Kiel)
24
(b) Effects of convection variability

• NCEP/NCAR reanalysis-forced experiments
• e.g., Eden and Willebrand (2002), Bentsen et al.
  (2004),.
• convection in Lab Sea linked to NAO
• - lagged response of the MOC

25
Variability of LSW formation and MOC
response to NCEP/NCAR heat flux variability
(FLAME 1/3 deg)
Convection intensity and transport
changes Response to NCEP heat fluxes

Labrador Sea Convection
Annual LSW production rate
Sv
1/12
43N
Mixed layer depth in March
26
Overturning anomalies in response to HEAT flux
variability
Years of intense LSW production
MOC anomalies In HEAT
-1
1
C.I. 0.2 Sv
27
MOC response to HEAT and WIND forcing anomalies
HEAT
HEAT and WIND
28
MOC (45 N)
MOC variability at about 45N (I) NCEP-forcing
29
(No Transcript)
30
Uptake of anthropogenic trace gases
Biastoch et al. (in prep.)
31
Overflow representation and effect on MOC
Denmark Strait overflow and MOC
Max. density along the outflow plume
1
Obs.
27.9
2
Denmark Straits
Cape Farewellr
3
28.0
MOC at 45N
28.1
3
Sv
62
64
66 N
2

-- Model experiments with different mixing
parameterizations
1
Obs Girton, Sanford, Käse (2001)
Max. density at 64N
32
Eddy effects in realistic configurations?
Example illustrating the temporal evolution of ml
depth (FLAME 1/12)
33
(No Transcript)
34
Overturning and Heat Transport
27N
45N
35
Simulated variability of the Atlantic MOC
(Bentsen et al. 2004) MICOM, 60-80 km resol.
4 realizations, forced by NCEP/NCAR 1948-1999
Winter ml
MOC max