Southen Ocean response to Climate Change

1
Southen Ocean response to Climate Change

• by
• Siobhan OFarrell
• CSIRO Atmospheric Research
• Aspendale, Vic. 3195
• Acknowledgments Tony Hirst, Jorge Sarmiento
• Helene Banks, Bill
  Budd.

2
Outline

• Ocean response in CSIRO Mark2 and Mark3 models
  to IPCC A2 scenario.
• Comparsion of Gille observations with model.
• Why Southern ocean is one of the best locations
  to detect climate change (Hadley Centre).
• Comparison of response in 6 models for OCMIP
  (MPI, CSIRO, GFDL, NCAR, HADCM3, IPSL)

3
CSIRO Mk 2 Model

• Atmosphere Grid R21 (5.6 x 3.2 deg.), 9
  levels Semi-Lagrangian moisture transport
  Moist convective adjustment scheme incl. mass
  flux Relative humidity cloud scheme
• Sea Ice Flato-Hibler rheology Semtner
  thermodynamics (3 level)
• Ocean Cox-Bryan model Grid 5.6 x
  3.2 deg., 21 levels Gent-McWilliams scheme
  (advective form)
• Reference Gordon and OFarrell (1997)

4
CSIRO Mk 3 Model

• Atmosphere Grid T63 (1.8 x 1.8 deg.), 18
  levels UKMO convection scheme Liquid water
  cloud scheme (Rotstayn, 2000)
• Sea Ice (Formulation as per Mk 2)
• Ocean GFDL Modular Ocean Model 2.2 Grid
  0.9 x 1.8 deg., 31 levels Integer Power Ri
  mixing scheme Bulk Ri mixing
  criterion Griffies et al. (1999) isoneutral
  mixing Griffies (1999) skew diffusion for
  GM90 Quick third-order advection
• No flux adjustment

5
Where to look for climate change in the ocean, by
Banks and Wood, J of Climate

• Study of signal to noise in model showed best
  indices for detection SST, SSS, SAMW at 32S.
• Poor indices were heat transport, strength of
  polar gyres, overflow water, with overturning
  streamfunction, Indonesian flow intermediate.
• Detection possible within 30 years by Argo
  observations at individual water masses in key
  locations which may currently be data sparse.
• Southern Ocean has particularly high signal to
  noise ratios in HADCM3 making it a good location
  for ARGO deployment.

6
Conclusions

• Reduction in overturning in Mk3 is less than in
  earlier Mk2 model but is closer to the response
  in some of the other models though some have poor
  representation of AABW formation.
• Zonal temperature, salinity and density sections
  have similar patterns of change.
• Higher rates of convection driven by ice
  processes.
• Both CSIRO models reproduce well the pattern
  observed by Gille at 900m.
• There is a range of SO response across 6 OCMIP
  models but some common changes in SST and SSS,
  sea ice, mixed layer depth and overturning.

7
Issues for Argo deployment and data analysis in
Southern Ocean CLIVAR

• High priority for seeding floats into SAMW and
  AAIW for climate change detection which
  integrate changes in surface fluxes which are
  more difficult to measure.
• Need to have floats in regions which are
  currently data sparse in Southern Ocean (ACC may
  concentrate floats).
• Need to measure changes in salinity and
  temperature over a range of time scales to
  understand variability of Southern ocean, ACC
  strength, ACW-like anomalies etc.
• Need to deploy floats that store data while
  operating under seasonal ice cover.
• Integrate ARGO data with satellite data by
  assimilating into Southern ocean model for
  seasonal-interannual changes
• Dont neglect deep hydrographic/tracer sections
  for measuring changes in AABW, shelf waters etc.