Fine Resolution Antarctic Model (FRAM)

Click figure for moving version (438Kb)
Geographic Coverage: 24S-79S, 1/4 lat x 1/2 long
Depths: 32 levels (6 in top 200m.) 21m thick at top, 233m at bottom
Time: 16 year simulation
Topography: DBDB5

Parameter: Potential temperature, salinity and velocity
Parameter: snow and ice thickness and temperature
File size: 90 Mbytes-all parameters at end of each model month
Format: 32 bit IEEE floating point (SUN internal format)
PI: D J Webb, SOC/UK
Contact: Beverly de Cuevas, Southampton Oceanography Centre, Southampton, U.K.
E-mail: bac@soc.soton.ac.uk
Media: Preferred medium is 2.3 Gbyte Exabyte tape
Notes: The archive datasets are available to the community together with the Fortran 77 software needed to read the file and extract horizontal and vertical (N-S and E-W) sections of the model variables. Users will need to write their own programs to extract sections at other angles or which cover only part of the model domain. Plotting programs in the 'C' language are also available for use on SUN 4 computers.

Model Description

FRAM is a primitive equation numerical model of the Southern Ocean between latitudes 24S and 79S. The code is based on that of Cox. The model starts from rest with initial temperature of -2C and salinity 36.69 everywhere, and relaxes to the Levitus annual mean temperature and salinity over 6 years. For the first 2 years 160 days, the relaxation timescale is 180 days for depths above 160 m and 540 days for the deeper levels. From then until 6 years the timescale is 360 days throughout. This is equivalent to the normal robust diagnostic scheme, except that the relaxation is so weak that the eddy field can develop. Surface forcing (Hellerman annual mean winds), is introduced after 2 years 6 months - increasing linearly from zero to the mean values over 6 months. After 6 years seasonal winds are introduced and surface forcing of temperature and salinity is operated by relaxing the surface layer only to the annual mean Levitus values with a time scale of one year. Towards the end of 9 years (day 3256), the viscosity and diffusion change from harmonic to a mixture of harmonic and biharmonic and the bottom friction changes from linear to quadratic. The model runs on to the end of 16 years.

Model Behaviour

After one model day the stream function shows large amounts of energy generated around topography. By day 10 the model has settled down and produced a recognisable Circumpolar Current with steering south of New Zealand, round Kerguelen and through the fracture zone of the mid-ocean ridge of the South Pacific. The amplitude of the Circumpolar Current, measured by the transport through Drake Passage, grows roughly linearly with time. When the winds are added, the model responds in about 10 days with a 23 Sv increase in transport. The transport, total kinetic energy of the model and other diagnostic fields settle down between years 5 and 6, indicating that the momentum budget of the model is near its asymptotic state. The transport through Drake Passage is about 200 Sv. The main regions of eddy formation are in the Agulhas Current and along the path of the Circumpolar Current. The stream function after six years is illustrated in Figure 1. Following the introduction of seasonal forcing after 6 years, the transport through Drake Passage has been oscillating between 195 and 200 Sv. There has been an increase in eddy energy near the ACC and eddies produced by the Agulhas Current (see the animation of temperature at 120m) are continuing to radiate into the South Atlantic. Eddies are also formed along the East Australia Current (see the second animation of temperature at 120m).

Further details can be found in FRAM publications.