Using an eddy-resolving ocean circulation model endowed with active thermodynamics and a turbulence closure parameterization, a hierarchy of numerical experiments is carried out to investigate the relative contributions of the wind forcing, the surface thermohaline fluxes, the river runoff, and the Bosphorus inflow/outflow in driving the yearly mean circulation in the Black Sea. The model accommodates a topographic and boundary-fitted curvilinear coordinate system and resolves steep topographical changes around the periphery of the basin using O(10 km) grid spacing and 18 stretched vertical levels. Model experiments show that the topography, wind forcing, and buoyancy forcing are all first-order contributors to the primary circulation of the Black Sea. If any of these features are neglected, significant elements of the model circulation do not reproduce observations. Subbasin scale gyres are caused by both wind and thermohaline forcing. Annual mean wind stress is sufficient to produce the major interior cyclonic gyres. Stronger winds produce more defined interior flow than the weaker winds of the Hellermann and Rosenstein (1983) fields. Heat flux is an important contributor to subbasin scale cyclonic circulation. However, the annual mean heat flux with spatial structue given by the climatology produces unrealistic features in the interior circulation. The latter ones disappear when including the seasonal heat flux variability.

This result strongly suggests that the seasonal cycle of the wind stress is much less crucial than the heat flux seasonal cycle in producing a realistic basin circulation. The Rim Current is locked to the steep topographic shelf slope, regardless of the forcing mechanism. Without including the strong topography the Rim Current is absent for all forcings. Mesoscale variability arises from the dynamic evolution of the Rim Current. This variability is enhanced by the Danube inflow and the Bosphorus inflow/outflow, demonstrating the importance of these buoyancy sources in enhancing the mesoscale. The deep-layer circulation is controlled by the barotropic pressure gradient and is insensitive to the magnitude, seasonality, or strength of the surface forcing. A transition zone separates the surface and deep-layer circulation patterns. Its circulation is mainly driven by the slope of the pycnocline developed as a response to the surface forcing. ¿ American Geophysical Union 1995