Introduction


      Several studies beginning in the early 1990s (Johns et al., 1990; Didden and Schott, 1993; Richardson et al., 1994) have shown that the North Brazil Current (NBC) pinches off large anticyclonic eddies in the western tropical Atlantic at about 8° N during the time of year (summer to winter) when the NBC retroflects into the ocean interior to feed the North Equatorial Counter Current (NECC). These NBC rings play a role in the Atlantic meridional overturning cell (MOC) as they contribute to the transport of South Atlantic upper ocean water into the North Atlantic.

      Two separate studies to quantify that portion of the MOC transport carried northward by NBC rings reached very different conclusions. This is a result of their disparate estimates of the number of anticyclones transporting South Atlantic water from the equatorial gyre to the North Atlantic subtropical gyre along the northeast coast of South America. Fratantoni et al. (1995, 2000) used a free-running, but realistic 1/4° Atlantic Ocean model forced by monthly mean wind stresses (Hellerman and Rosenstein, 1983) and an imposed Atlantic MOC with an amplitude consistent with Schmitz and McCartney (1993). They concluded that two to four rings are shed each year from the NBC and the intergyre transport by these rings is 20-25% of the annual mean northward Atlantic MOC transport. Goni and Johns (2001) examined about six years of TOPEX/POSEIDON satellite altimeter data in the NBC region to arrive at a warm ring estimate of more than five per year possibly carrying greater than 1/3 of the northward Atlantic MOC transport. This includes eddies they identified in the region that may or may not have been shed by the NBC retroflection.

      The study presented here is an attempt to reconcile this significant difference in estimated MOC transport by NBC rings. The methodology is a combination of the previous two approaches in that it utilizes a realistic numerical ocean model that assimilates satellite altimeter data. The model is nearly global in coverage and even without data assimilation it simulates realistic currents, including the currents of the equatorial Atlantic as well as NBC rings. Using a data assimilating realistic ocean model may yield a better estimate of the number of NBC rings shed each year because the model acts as a dynamical interpolator of the assimilated data which facilitates determination of the origin of the warm rings observed in the NBC region.

     Click on any figure embedded in the text to see a larger version of that figure.