Principal Investigator Joerg-Olaf Wolff Project g22
Antarctic CRC Machine VP
University of Tasmania
Co-Investigators Nathan Bindoff
Antarctic CRC, University of Tasmania
Southern Ocean and Atmosphere Process Model
The parameterization of small scale processes in coarse resolution ocean and sea-ice components of coupled climate models is an important and currently very active field of research. Recently proposed theoretical arguments which link the effect of the meso-scale eddy-field to large scale properties of the mean flow have been shown to dramatically improve certain characteristics of the oceanic circulation in the Southern Ocean (SO), e.g. the depth of the mixed layer. These problems are best studied using high resolution models in limited regions that can actually resolve these small scale processes. Such studies will allow us to characterize and develop or verify more detailed parameterizations of these processes in order to improve the current ocean and sea-ice models. A recent coupled coarse resolution GCM with a transient increase in the atmospheric CO2 concentration showed a significantly delayed warming of the Southern Hemisphere due to rapid ventilation through changed water masses in the SO. These changes could have important implications on the Australian regional climate. However, one limitation of coarse resolution ocean models is their inherent inability to reproduce the correct dynamical balance of the SO Circulation, especially the Antarctic Circumpolar Current (ACC). This balance is intimately linked to the existence of mesoscale eddies, which are thought to be the main carrier of poleward heat transport in this region. Strong small scale mixing and deep overturning within the Southern Ocean probably have a profound influence on the pattern of greenhouse warming in response to transient global change forcing. These processes of water mass formation and transport are critical to the uptake and storage of carbon dioxide and other tracers in the deep ocean.
We are using a state of the art primitive equation ocean model (HOPE, Hamburg Ocean Primitive Equation Model) at eddy resolution in a sector of the Southern Ocean south of Australia. The model was developed at the Max-Planck-Institute for Meteorology in Hamburg, FRG. A modified Hibler-type sea-ice model is coupled to the HOPE model and allows prognostic calculation of sea-ice thickness, concentration and velocity. This is the first time a coupled ocean/sea-ice model will be used in the SO in eddy resolution. Despite restricting the area of interest to a specific sector of the Southern Ocean considerable computer resources will be needed to allow for sensitivity studies and extended integration times.
What are the basic questions addressed?
The eddy resolution will allow the investigation of their role in and around the sea-ice zone, and their effect on bottom water formation in a dynamically consistent framework. In addition, the role of the eddies on the mean circulation in the upwelling of heat and transport of salt at the Antarctic Divergence as well as across the ACC will be determined.
What are the results to date and future of the work?
This project is aimed at testing, tuning and running a state of the art primitive equation ocean model (HOPE-Model). The first aim of this project is to test the sea-ice model and its interaction with the ocean. Initial problems with a too strong ice build-up in certain areas around Antarctica have been solved in the last allocation period and we have started a 100 year control run where ocean thermohaline fields have been relaxed to climatology with a time scale of about 150 days. The sea-ice model has been improved through consideration of a more complex radiation balance and the annual cycle of sea ice area compares very well with satellite observations with the exception of a too strong heatflux during July/August (see Fig. 1). The effect of the heatflux mismatch can be seen in Fig. 2 where large areas north-east of the Wedell Sea and north-west of the Ross Sea are basically ice free in constrast to observations.
The second aim is to investigate the relationship
between the rates and areas of water-mass formation, sea-ice distribution,
deep water pathways to the north, heat- and freshwater fluxes
in the SO and the dynamical balance of the ACC.
Fig 2 Sea Ice Thickness Year 100 Month 8
What computational techniques are used and why is a supercomputer required?
HOPE is a primitive equation model of the global ocean circulation, but may also be used for regional oceanographic studies. Prognostic variables are the three-dimensional velocity fields, sea-surface elevation and the thermohaline variables. The vertical distribution of variables is on prescribed levels and in the horizontal an Arakawa-E-type grid is used. The time discretization uses only two time levels. A simplified sea-ice model (including thermodynamics and dynamics) allows prognostic calculation of sea-ice thickness, compactness and velocity. HOPE is especially useful for altimetry data assimilation purposes because of the prognostic sea-surface elevation.
The code has been specifically designed for vector processing computers. Run time tests at the German Climate Computer Center on a CRAY YMP have shown an average speed of 180 Mflops for this particular code with a resolution similar to that planned for these experiments on the VP2200. Vectorization testing software has indicated none of the code needs further optimization. Production runs on the VP2200 have a 91% (with sea-ice) and 97% (without sea-ice) vector unit usage.
Climatology and Variability in the ECHO Coupled
GCM, Latif, M, T Stockdale, J-O Wolff,
G Burgers, E Maier-Reimer, M M Junge, K Arpe and L Bengtsson,
Tellus, 46(A), 367-380, (1994)
Regional high-resolution coupled ocean/sea-ice
model, S Marsland, and J-O Wolff, 3rd
Natl AMOS Conf, Hobart, Australia, 136, (1996)
The dynamical and thermodynamical balance of the
Antarctic Circumpolar Current in an eddy-resolving primitive equation
model with reentrant channel geometry,
B Munro, and J-O Wolff, 3rd Natl AMOS Conf, Hobart, Australia,
Southern Ocean and Atmosphere Process Model,
J-O Wolff, and N L Bindoff, 3rd Natl AMOS Conf, Hobart, Australia,