AOML South Atlantic MOC related Observations and Plans

1
AOML South Atlantic MOC related Observations and
Plans

• Silvia L. Garzoli

• Molly O. Baringer (AOML)
• Christopher S. Meinen (AOML)
• Carlisle Thacker (AOML)
• Shenfu Dong (CIMAS)
• Zulema Garraffo (RSMAS)
• George Halliwell (RSMAS)
• Ricardo Matano (OSU)
• Alberto Piola and Ariel Troisi (SHN)
• Edmo Campos (USP)
• Mauricio Mata (URGDS)
• Sabrina Speich (LPO,UBO)

2
Outline

• Heat transport across 35S
• SAM Moorings at 35S
• Model data experiments

3
SAMOC Workshop, Estancia San Ceferino, Buenos
Aires Argentina, May 8, 9, and 10, 2007
Map of the South Atlantic and Southern Ocean,
including the two principal choke point regions,
the Drake Passage and south of South Africa, with
the current and proposed locations of instrument
deployments and the institutes leading the
corresponding associated projects.
4
AOML High density XBT AX18 line
Correlation of MOC and heat transport across 33S
(C.I.0.2)
High density XBT 20 lines conducted by AOML
since July 2002
(Barreiro, 2007)
5
Mean surface circulation in the region
(Lumpkin and Garzoli, 2008)
(Goni et al., 2008)
6
Heat Transport Methodology (Baringer and Garzoli,
2007)

• XBT data is collected using Sippican T-7 probes
  typically to depths of about 800 meters
• Data is extended to the ocean bottom using
  Levitus 0.25 data set.
• Salinity is estimated for each XBT profile by
  using S (T, P, Lat, Long) derived from Argo and
  CTD data (Thacker 2006).
• Meridional Ekman transports are computed as the
  Ekman mass (My) and Ekman heat (Hy) transport
  using NCEP daily reanalysis winds and
  interpolating the daily NCEP values to the time
  and location of the XBT observation.

7
Mean Heat transport (14 realizations CT-BA)
0.53 PW Std 0.11 PW (Garzoli and
Baringer, 2007) Mean Heat transport (3
transect CT-R) 0.54 PW Std 0.10 PW.
Mean Heat transport (20 realizations) 0.53
PW Std 0.12 PW
Total, Ekman and Geostrophic components of the
Heat transport across the AX18 lines.
8
Time series of the Ekman Heat transport
integrated across the basin as a function of
latitude. Dots indicate mean latitude of each
cruise.
9
Annual cycle of the Ekman component, geostrophic
component, and total heat transport across the
AX18 20 realizations. Results from the 3 lines
occupied from Cape Town to Rio are shown in a
different color (yellow).
10
Comparison with other results
11
Observing the DWBC
One of the largest uncertainties in the measured
heat transport is the lack of direct measurements
of the barotropic component of the flow, which is
largest to the west of 47W. This is particularly
important because at the western boundary the
Malvinas Current and the DWBC both flow in the
same direction, creating a strong barotropic flow
whose magnitude and variability are poorly known.
Mean model velocities at 1500 m depth
Model sections of the meridional velocity showing
the DWBC at 30S (left) and 34.5S (right).
Negative velocities indicate southward flow.
(From POCM model, Tokmakian and Challenor, 1999).

12
SAM
In conjunction with the CPIES deployed by Sabrina
Speich (Fr)
Positions of the first stations of the
BONUS-GOODHOPE transect. In green the two
stations where the two C-PIES moorings have been
deployed.
Proposed cruise track and tentative instrument
locations for the new IES line in the South
Atlantic. This program will be conducted in
collaboration with scientists from Argentina and
Brazil.
13
Can we do it? A combination of CTD with IES
(using the GEM technique) and pressure gauges
proved to be an adequate mean to monitor the
transport of the Deep Western Boundary current in
the North Atlantic across 26.5N .
Comparison of the DWBC transport at 26.5N
estimated from three different techniques. All
transports are integrated between 1200 and 4800
dbar. The same bottom pressure gauge data is
used to provide the bottom absolute velocity
reference for both the IES and the Dynamic height
moorings. Note that the magenta and green lines
switch to dotted in late January when the top of
one mooring broke off and the remaining segment
of the mooring slumped down due to the loss of
buoyancy current meter and dynamic height
mooring data after this point should be viewed as
having larger error bars.
(Meinen et al., 2004)
14
SAM first cruise October 08Deployed 4-5
yearsBiannual visits to recover the data

• Pop up system

15
AMOC variability in the South Atlantic Data
analysis and a numerical model simulation

• Characterize the mean and time varying pathways
  of the AMOC
• Evaluate the correlation between the AMOC
  strength and the meridional Heat Transport.
• Defining the importance of variations in
  inter-ocean and inter-basin exchange and the
  connectivity of the MOC.

AX22 1996 to present AX18 2002 to present AX25
2004 to present
Comparison transports from XBT and altimeter
16
Global 1/12º HYCOM Climatological
simulation. Produced at NRL (J.Metzger, J.
Shriver, A. Wallcraft, E. Chassignet, H.
Hurlburt). ECMWF ERA40 forcing plus 6 hourly wind
anomalies from a repeat year of NOGAPS winds.

• A South and Tropical Atlantic regional model will
  be nested inside the Global model. In this way
  better accuracy can be achieved for
• Trajectories of numerical drifters and floats
• Transports and diapycnal flux diagnostics.

daily boundary conditions from the global model
Global SSH in the Regional model domain
(From Garraffo,2008)
17
Particles will be launched in the regional model,
in a 1x1 regular grid at several depths,
following the model 3-D motion, plus at 0.25 deg
in transects near the inflow into the domain and
selected locations (yellow rectangles) e.g.,
DWBC, NADW.
Deep flow (mean of model year 15, 3000m)
AC
(From Garraffo, 2008)
18
"Observing System Simulation Experiments for the
AMOC Carlisle Thacker (NOAA/AOML), George
Halliwell (RSMAS/UM)

• Objective
• To estimate the effectiveness of strategies for
  monitoring the overturning circulation.
• Technique
• Simulate observations and then evaluate how
  uncertainties associated with the overturning
  circulation are reduced when they are assimilated
  into an ocean model.
• As accurate quantitative characterization of
  model errors and their correlations and of the
  noise characteristics of the simulated data
  strongly influence the results of an OSSE, these
  issues will receive considerable attention.
• Methodology
• The system will be calibrated by simulating data
  similar to those from the RAPID/MOCHA program and
  seeing whether they have an impact similar to the
  real data.
• The impact of similar data from other sections,
  e.g. near 30S, will be evaluated.
• Once the system is working, the value of other
  types of observations can also be considered,
  e.g., extending the depth range of Argo profiling
  floats.

19
Thank you