The wind driven circulation

Superimposing the NAO windstress-curl anomalies onto the climatological windstress-curl reveals a ${\bf Z}$-like pattern delineated by the tilted zero windstress-curl line (diagonal) and the two zero windstress-curl line anomalies (Fig. 2a). The diagonal marks the climatological separatrix between the subpolar and the subtropical gyre. The anomalous mechanical forcing of the NAO then drives a cross-separatrix circulation from one gyre to the other. Marshall [2000a] call this anomalous circulation pattern the ``intergyre gyre'', represented by a transport streamfunction $\Psi_{ig}\vert _w$, evaluated just inside the western boundary current. A positive NAO phase drives an anticyclonic ( $\Psi_{ig\vert w} \geq 0$) ocean circulation.

Based on linear, flat bottom, Sverdrup dynamics, the time evolution of $\Psi_{ig}\vert _w$ may be expressed as (in non-dimensional form)

$$\Psi_{ig}\vert _w = - \int_{t - t_d^{ig}}^{t} \mbox{curl} \; \tau \, {\rm d} t$$ (2)

where $\tau$ denotes the spatial pattern of the NAO surface wind stress, and t_d^ig is the travel time of a first baroclinic Rossby wave mode. It is about ten years at $40^\circ$N in the Atlantic [1997]. The complete Sverdrup circulation (contoured every Sverdrup) is shown in Fig. 3, computed from the NCEP-NCAR reanalysis surface windstress. It shows the equilibrium response of the thermocline to an idealized, constant NAO mechanical forcing, which is established after t_d^ig years. In nature, the NAO is primarily stochastic and the intergyre gyre is thought to fluctuate on timescales of order $\simeq t_d^{ig}$ (cf. Frankignoul [1997], hereafter refered to as [FMZ97]), thereby introducing decadal timescale in the NAO - ocean sytem. Based on Fig. 3 and the observed mean SST, one can estimate the advective heat flux induced by the intergyre gyre (a $v' \bar{T}$ term). It is found to be of the order of $Q_{ig} \, = \, \pm$0.07 PW for an anomalous transport of $\Psi_{ig}\vert _w \, = \, \pm$3 Sv [2000a]. The intergyre gyre is thus conjectured to play a key role on decadal timescales, being a possible candidate for anomalous advection of heat by the ocean associated with NAO fluctuations. Because its net effect is to transport heat across the mean position of the separated Gulf Stream, it modulates the cross Gulf Stream SST gradient, and may thus impact the SST pattern locally forced by the NAO (Fig. 2c).