Monday, 25 June 2007
Ballroom North (La Fonda on the Plaza)
Highly inertial dynamics is necessary but not sufficient to model the observed Gulf Stream separation at Cape Hatteras. Without a cold shelfslope "density current" flowing TOWARD Cape Hatteras along the bottom, the Gulf Stream tropical watermass flows along the shelfslope past Cape Hatteras due to western boundary current intensification involving Rossby waves propagation. Maintaining the thin, narrow shelfslope current during its long transit from its high latitude seas and Arctic Ocean sources to Cape Hatteras without excessive dilution requires accurate, low dissipation numerics. This is achieved by a six-grid, 4th-order-accurate coupled MEDiterranean Sea and North Atlantic (MEDiNA) model running on a Pentium 4 based PC. The MEDiNA model results are compared with observations based statistics. In a roughly wedge-shaped region between Gulf Stream and deeper shelfslope density current fronts having apex near Cape Hatteras, there is a clear tendency for isotherm (isopycnal) flattening -- a classic baroclinic instability signature. Gulf Stream cyclonic frontal eddies and warm-core eddies that pinch off northward Gulf Stream meanders are finite amplitude signatures of baroclinic instability processes that convert available potential energy into eddy kinetic energy. The Gulf Stream and density current maintain the available potential energy, thus playing roles like warm outer and cold inner wall of classic rotating annulus experiments. These results repudiate the widely held ocean modeling concept that z-level models cannot simulate density currents because of their SUPPOSEDLY inherent excessive dissipation.
Acknowledgement: The support from the Marcelino Botin Foundation, Spain, AcuSea, Inc. and National Science Council, Taiwan (grant number NSC952119M002048) is acknowledged.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner