ISOPYCNIC AND DIAPYCNAL CIRCULATION OF THE UPPER EQUATORIAL ATLANTIC-OCEAN IN 1983-1984

Isopycnal circulation and diapycnal processes in the equatorial Atlantic Ocean are investigated using eight cruises from November 1982 to July 1984. An eastward decrease of the transport of the equatorial undercurrent is observed averaging 18.2 10(6) m3 s-1 at 35-degrees-W to 10.2 10(6) m3 s-1 at 4-degrees-W. There is also a meridional convergence at the undercurrent level and a larger meridional divergence in the surface layer of 15 10(6) m3 s-1 between 35-degrees-W and 4-degrees-W. In the eastern equatorial Atlantic, the undercurrent core density as well as the salinity above the undercurrent experience a large seasonal cycle. During the season where the eastern Atlantic thermocline is closest to the surface and surface waters are coldest, the current peaks in denser waters and subsurface salinity maximum are weaker. An analysis of the salinity on isopycnal surfaces indicates that there is significant diapycnal mixing in the upper thermocline. In the eastern Atlantic, the upper thermocline vertical heat diffusivity coefficient scaled over 1-degrees of latitude varies between 3 cm2 s-1 during the upwelling season and nearly 0 early in the year. This seasonal mixing is an important element of undercurrent dynamics. The turbulent heat flux at the base of the surface layer averages 50 W m-2 between 1.5-degrees-N and 1.5-degrees-S. The surface layer heat budget implies an average oceanic gain of 60 W m-2 between 1.5-degrees-N and 1.5-degrees-S which could result from exchanges with the atmosphere. The heat flux and fresh water seasonal cycles needed to close the budgets are not realistic, however. The eastern equatorial Atlantic thermocline seasonal upwelling is associated with a large inflow into the surface layer. 'Over the two years, this flow averaging 11-12 X 1(6) m3 s-1 originates from the upper thermocline, and diapycnal transports near the core of the undercurrent or below are found to be small. Below the core of the undercurrent, we also find that mixing is not intense: for instance, at sigma(theta) = 26.5, in the upper part of the thermostad, vertical heat diffusivity is only 0.6 cm2 s-1 (assuming that mixing takes place over 1-degrees of latitude). The uncertainties on these budgets are however large, and assumptions on the dynamics that were used could not bc checked.