Subantarctic Mode Water and Antarctic Intermediate Water: Circulation and formation processes

1
Subantarctic Mode Water and Antarctic
Intermediate Water Circulation and formation
processes

• Bernadette Sloyan
• CSIRO Marine and Atmospheric Research,
• Lynne Talley, Teresa Chereskin and James Holte,
• Scripps Institution of Oceanography

2
Southern Ocean Circulation
Schematic diagram of circulation of Southern
Ocean and interactions with subtropical gyre
circulations.
3
Southern Ocean Meridional overturing and
interbasin exchange
Schematic cross section of Southern Ocean
4
SAMW Formation Regions Southeast Indian and
Pacific Oceans
Proxy for winter mixed layer depth is 95 Oxygen
saturation. Deep mixed layers identify sites of
SAMW formation.
From Talley 1999
5
Global Distribution of SAMW and AAIW
Global distribution of low-salinity intermediate
water. Formation region marked with X, and
strong mixing with
LSW
NPIW
AAIW
From Talley 1999
6
SAMW and AAIW T-S properties
Progression of T-S across sectors of Southern
Ocean. DP, Falkland Current, S.Africa, S.Aust,
mid-Pac.
Atlantic Sector
Indian Sector
Pacific Sector
Difference in T-S between north and southward
flow of subtropical gyres.
Cape B
Natal B
mid-Atl. ridge
Tasman Sea
SW Pacific Chile B
Central Ind. Perth B
Argentine B.
7
Circumpolar distribution of Potential Vorticity
Drake Passage
Atlantic 30oW
South Africa 10oE
Mostly summer sections, apart from Pacific 135oW
which was occupied during late Austral spring
(Oct-Nov).
Indian 90oE
South Aust. 145oE
Pacific 170oW
Pacific 135oW
Pacific 102oW
Pacific 88oW
Low PV weak stratification
PV (x 10-14 (cm s)-1 )
8
Summary of Circumpolar properties of SAMW and AAIW

• SAMW identified by minimum in potential
  vorticity, maximum oxygen concentration and deep
  mixed layers in Indian and Pacific sectors of
  Southern Ocean.
• AAIW is identified by salinity minimum lying
  below SAMW
• SAMW freshen and cools in the eastward direction
  from mid-Atlantic (30oW) to Drake Passage.
• SAMW and AAIW exchange freshwater and heat
  between the Southern Ocean the subtropical gyres.
  They, also supply nutrient to the upper ocean of
  the adjacent subtropical gyres.
• Circulation time scale of the order of decades.

SAMW and AAIW important components of influence
and potential impact Southern Ocean on global
climate. How well are these water masses
simulated in Climate Models?
9
Climate model simulation Comparison with CSIRO
Atlas of Regional Seas (CARS2006)
CARS2006 depth, temperature and salinity for
Austral Spring (Sept, Oct, Nov) on isopycnal
surfaces that define SAMW and AAIW.
Subantarctic Front
10
Climate Models IPCC AR4
Define model mean climate of 20th century as
20-year average (1981-2000)
11
Climate Model Simulation of SAMW and AAIW
Isopycnal Depth
RMS ensemble isopycnal depth error (m) relative
to CARS2006 for Austral Spring. Ensemble is for 5
(26.7, 26.9 27.0 27.2 and 27.4) observed
isopycnal surfaces that define SAMW and AAIW.
Subantarctic Front
12
Observed and Model Simulated AAIW Salinity
minimum in Eastern Indian Ocean (100oE)
Salinity (psu). Observed depth and latitude
extent of AAIW salinity minimum is overlain on
model salinity distribution.
13
Summary of Climate Model Simulation of AAIW and
SAMW

• Climate models, except for HadCM, CSIRO and MRI,
  provide a reasonable simulation of SAMW and AAIW
  in the Southern Ocean. (Note Hadley Centre and
  CSIRO have further developed their climate
  models.)
• In the formation regions, models generally better
  simulate SAMW and AAIW in the eastern Pacific
  Ocean than in the eastern Indian Ocean.
• Models display limited equatorward extension of
  low potential vorticity layer and salinity
  minimum layer of SAMW and AAIW, respectively.
  Indicating models are too diffusive north of 40oS
• Large isopycnal depth errors north of SAF and
  subtropical oceans explained by incorrect density
  of model SAMW pycnostad and too warm upper ocean
  which both deepen model isopycnal layers.
• Error in simulation of SAMW/AAIW properties in
  formation regions may be due to biases in
  position and strength of wind stress, inadequate
  representation of sub-grid scale mixing and
  eddies. Errors in subtropical ocean possibly due
  to model representation of eastern and western
  boundary processes.

How can we improve model representation of SAMW
and AAIW?
14
SAMW and AAIW Formation Processes
Schematic of processes likely to be involved in
the formation of Subantarctic Mode Water (SAMW)
and Antarctic Intermediate Water (AAIW)
15
Austral Spring (2005) and Summer (2006)
Austral Spring 2005
Austral Summer 2006
16
Mixed Layer Depth
Austral Spring 2005- Deepest mixed layers found
in southeast direction from 103oW. Surface
salinity maximum associated with deep mixed layers
Austral Summer 2006 Shallow mixed layers
throughout region (20-50m). Region covered by
freshwater lens.
17
Combined CTD and XCTD surveys
Austral Spring 2005 Position of CTD and XCTD.
447 profiles.
Austral Summer 2006 Position of CTD and XCTD.
434 profiles.
Contour maps are AVSIO dynamic height field
average for duration of each survey (45
days). Profile spacing is about 8km or less in
southern part of survey
18
Quality Control of XCTD profiles

• XCTD profiles were quality controlled in the
  following manner
• Each temperature profile was quality controlled
  against CSIROs Atlas of Regional Seas (CARS)
  using CSIRO XBT quality control software (Quest).
  Temperature QC flags were applied to salinity
  profiles
• Each XCTD salinity and temperature profile, and
  T-S diagram were compared to surrounding CTD
  stations. QC flags were applied to both property
  fields.
• Profiles that failed to reach 500m were removed.
  Finally, a filter was applied to XCTD salinity
  profiles to remove high frequency noise.
• All good XCTD profiles were then combined with
  CTD stations for each survey (Austral winter 2005
  and Austal Summer 2006).

19
Austral Spring Property Distribution
MLD all data
MLD CTD only
20
Austral Summer Property Distribution
MLD all data
MLD CTD only
21
Property Structure at frontal crossings
Austral Spring 2005
Significant interleaving between potential
density 27.0 and 27.3 at all frontal crossings.
22
Property structure at frontal crossings
Austral Summer 2006
Significant interleaving between potential
density 27.0 and 27.4. Intrusions can be traced
across a number profiles. Red and blue profiles
are sites of intensive XBT surveys.
23
Thorpe Scale (LT) - density overturns
Thorpe Scale (LT) provide a length scale
associated with density overturns. It is defined
as the rms of vertical displacement of a
reordered density profile over a gravitationally
unstable patch.
24
Criteria imposed to determine overturns
Profiles were re-ordered to be gravitationally
stable following Ferron et al 1998, with average
XCTD and CTD noise determined to be dr 0.001.
Overturns were determined for density difference
of 2dr. Following Galbraith and Kelly (1996)
minimum thickness of resolvable overturn is given
as Lrmin 2 g/N2 dr/ro, where N is buoyancy
frequency and ro is mean density. Additionally a
minimum XCTD overturn length of 3m was applied to
these profiles based on smoothing applied to XCTD
salinity profiles. Quality Control procedures
applied to XCTD removed the need to apply the
water mass criteria.
25
Thorpe Scale Austral Spring and Summer
Austral Spring 2005
SAF
Austral Summer 2006
MLD CTD only and XCTD/CTD
26
Turbulent Kinetic Dissipation Rate (e)

• Ozmidov (1965) related the turbulent kinetic
  dissipation rate (e) to the stratification and a
  turbulent mixing length such that
• e Lo2N3
• Mircostructure studies which directly measure Lo
  have been used to defined a relationship between
  the Lo and LT as
• Lo aLT,
• where a varies between 0.65-0.98. Here we use a
  0.8. Therefore,
• e a2LT2 ltNgt3
• where ltNgt is buoyancy over averaged over LT
  length.

27
Layer mean Turbulent kinetic Dissipation Rates
Austral Spring 2005
SAF
Austral Summer 2006
Wm_den 26.9, 27.0, 27.05, 27.1 27.2, 27.3 , 27.4
Magnitude of e O(10-7 - 10-8)
28
Diapycnal Diffusivity (kr)
Following Osborn (1980) turbulent dissipation (e)
can be related to diapycnal eddy diffusivity (kr)
as kr G e N-2 Here N is the background
buoyancy as the expression for kr relates to the
turbulent diffusivity working on the background
density gradient. G is the mixing efficiency
(0.2). Mixing events result in the vertical
exchange buoyancy and dissipation of turbulent
kinetic energy (Osborn, 1980 and St Laurent et
al. 2001) . We can then define the water mass
buoyancy flux associated with mixing events by
Jb G?e, Where ?e, is the mean dissipation over
over the depth of a water mass.
29
SAMW and AAIW Diapycnal Diffusivity
Austral Spring 2005
Range of kr O(10-2 -10-4)
SAF
Austral Summer 2006
Wm_den 26.9, 27.0, 27.05, 27.1 27.2, 27.3 , 27.4
30
SAMW and AAIW Turbulent Mixing Buoyancy flux
Austral Spring 2005
SAMW Jb north of deep ML in spring and throughout
in summer
SAF
Austral Summer 2006
Largest Jb associated with AAIW densities in
spring.
Magnitude of Jb O(10-8)
Thermocline value usually 10-10
31
Conclusions
The Austral spring (2005) and summer (2006)
hydrographic surveys provide a rich data base
from which to explore processes involved in
formation of SAMW and AAIW.
Estimates of Turbulent mixing show Significant
mixing and associated buoyancy flux in AAIW
layers, so gt 27.1 kg m-3, near the Subantarctic
Front. Slight enhancement of mixing in the
Austral Spring relative to Austral
Summer. Enhanced mixing and associated buoyancy
flux in SAMW layers, 26.9 lt so lt 27.1 kg m-3,
north of deep mixed layer in Austral Spring and
between Subantarctic Front and 45oS in Austral
Summer. Elevated interior mixing support by
rapid decrease in CFC concentration northward of
the SAF.
32
Future Directions
Assess importance on subsurface turbulent mixing
on formation and property characteristics of AAIW
relative to other formation processes.
Importance of austral summer (and early autumn)
turbulent mixing in SAMW layer on
pre-conditioning water column for effective
winter/spring convective mixing when water mass
isolated from surface forcing by seasonal
thermocline. Combine turbulent mixing estimate
with surface convective mixing.
33
Thank you
Contact Email Bernadette.Sloyan_at_csiro.au