RSMAS homepage MICOM homepage MICOM homepage High Resolution homepage
Run 1: COADS forcing Run 2: ECMWF Forcing Run 3: ECMWF Daily Forcing
Model Configuration

Domain

North and Equatorial Atlantic (28oS to 65oN) with a horizontal resolution of 1/12o (mesh size on the order of 6 km on average) and 16 layers in the vertical. This configuration provides a satisfactory result for the Gulf Stream separation, supporting the view that an inertial boundary layer (achieved by the fine grid resolution) is an important factor in the separation process.

Bottom topography is derived from the ETOPO 2.5 quality-controlled data set.

Relaxation to observed monthly T and S (Levitus, 1982, except for data from the TTO section in the deepest two layers, from ~65oN to 62oN and 40oW to 30oW) is carried out in 3o wide buffer zones with relaxation times of 5 days at the northern and southern wall boundaries (65oN, 28oS) increasing to 30 days at the inner edge of the buffer zones. The relaxation time scale for the Gulf of Cadiz buffer zone (representing the Mediterranean Sea outflow) is 30 days.

Forcing

  • COADS climatology from Da Silva (1994)
  • wind stresses, u-star3, heat flux [Qnet + dQ/dT (Tmodel - Tobser)], and observed E-P plus small relaxation to observed surface salinity

Mixing parameterization

  • thickness diffusion:  0.3 cm/s × mesh size
  • laplacian momentum mixing:  the greater of:
    • .75 cm/s × mesh size
    • 0.1 × (mesh size)2 × velocity deformation, plus small biharmonic momentum mixing
      (1 cm/s  × (mesh size)3)
  • temp./salin. mixing:  0.5 cm/s × mesh size
  • diapycnal diffusivity:  0.001 × (buoyancy frequency)(-1)  cm2/s2

Computational requirements

The computational requirements are extreme for basin-scale ocean modeling at the resolutions of interest (less than 10 km) (Bleck et al., Parallel Computing, 1995). Each time that the horizontal resolution is increased by a factor n, the computational load goes up by a factor n3, since the n-fold reduction in linear mesh size necessitates n times more time steps to integrate the model over a given time interval. For example, the initial 1/12o two-year simulation performed by computer scientists at the University of Minnesota required a month of round-the-clock computer time on 256 processors of the Cray T3D at the Pittsburgh Supercomputing Center.

The most recent MICOM high resolution simulation incorporates improved initial conditions, an increase in vertical resolution to better resolve the North Atlantic water masses, and lateral boundary conditions at the northern and southern boundaries to account for the existence of adjacent ocean basins. This simulation is presently being integrated for a total of 20 years, and will allow direct and detailed comparisons withdobservations from satellites (sea surface height, sea surface temperature), moorings, inverted echo sounders, and free-floating drogues. The simulation is performed at the NAVO MSRC.

Results

Publications: Paiva et al, 1999, Chassignet and Garraffo, 2001, Garraffo et al, 2003, Schmid et al, 2003, Garraffo et al, 2001a,b, Bracco et al, 2003.

Numerical drifters were deployed at the model mixed layer and at 400, 1000, 1500, and 3000m, initially in a 1ox1o grid. Data are available under request. This research was funded by NSF (Griffa, A. , E.P. Chassignet, Z. Garraffo; Publications: Garraffo et al, 2001, Bracco et al, 2003).

Link to Lagrangian Analysis and Prediction of Coastal and Ocean Dynamics (LAPCOD)research.

 

 

 

Last Modified on August 11, 2004
Webmaster