Inter-gyre exchange: A process study

The North Atlantic Current (NAC) forms part of the boundary between the subtropical and subpolar gyres in the North Atlantic Ocean. The current has topographically controlled stationary meanders that appear to grow and decay. A region to the east of the current contains water of mixed subpolar/subtropical properties suggesting that there is exchange across the NAC. Using RAFOS floats we found that if a float is in the fast moving current for two days, it has a 3% chance of crossing from one side of the NAC to the other. The floats also showed that the current loses and re-entrains water parcels: 12% of the floats in the jet for two days are lost to the slower moving surrounding waters. The preferred locations of both inter-gyre exchange and losses from the jet were linked to the meander troughs and the recirculations associated with the meanders. A three layered, wind driven, two gyre numerical model was employed to examine further the importance of the inter-gyre exchange and the mechanisms involved. Although the numerical model does not attempt to precisely simulate the NAC, the meandering mid-latitude jet does exhibit similar values of exchange: 5% of the Lagrangian model floats in the jet for two days did cross from one gyre to another in the upper layer. Lagrangian particles were also lost from the jet to the slower moving surroundings. The inter-gyre exchange is important in understanding the vorticity budget; approximately 85% of the potential vorticity imparted to the southern basin by the wind is fluxed across the mid-latitude line. A passive tracer experiment showed that inter-gyre exchange was important for the property distribution. We differentiate between two types of inter-gyre exchange: permanent and short-term exchange. Since the latter could have an impact on property distribution, we term it property exchange; the former is both mass and property exchange. We examine different mechanisms: meander evolution, eddy formation, eddy-jet interactions, internal mode communication, and small scale turbulence. Three additional experiments were conducted: one with a tilted wind stress curl; one with a higher numerical diffusion and finally one with a biharmonic formulation of the numerical diffusion, rather than the Laplacian used in the previous experiments