Dynamics of an idealized Beaufort Gyre: 1. The effect of a small beta and lack of western boundaries

The Beaufort Gyre in the Arctic Ocean differs from a typical moderate-latitude gyre in some major aspects of its dynamics. First, it is located in a basin without a western boundary, which is essential for closing midlatitude circulations. Second, the gradient in Coriolis parameter, , is small and so the validity of the Sverdrup balance is uncertain. In this paper, we use an idealized two-layer model to examine several processes that are related to these two issues. In a circular basin with closed geostrophic contours in interior, the variability of vorticity in the upper layer is dominated by eddies. But in the time-mean circulation, the main dynamical balance in the basin's interior is between the curl of wind stress and the eddy vorticity fluxes. The torque of friction becomes important along the boundary where the rim current is strong. It is found that the smallness of has only a relatively small impact in a circular basin without a meridional boundary. The gyre is considerably more sensitive to the existence of a meridional boundary. The time-mean circulation weakens considerably when a peninsula is inserted between the model's center and the rim. (One side of the peninsula is dynamically equivalent to a midlatitude western boundary.) The gyre's sensitivity to has also increased significantly when a meridional boundary is present. Subsurface ridges have similar effects on the gyre as a boundary, indicating that such topographic features may substitute, to some extents, the dynamical role of a western boundary.