On the effects of parameterized Mediterranean overflow on North Atlantic ocean circulation and climate

A parameterized Mediterranean overflow, based on the marginal sea boundary condition of [Price, J., Yang, J., 1998. Marginal sea overflows for climate simulations. In: Chassignet, E.P., Verron, J., (Eds.), Ocean Modeling and Parameterization, Kluwer Academic, pp. 155–170], has been implemented in the ocean component of the Community Climate System Model to represent exchanges through the Strait of Gibraltar, associated entrainment and intrusion of overflow product water into the Atlantic. Previously, in coarse resolution model versions with a closed Strait, this physics has been either missing in uncoupled configurations or only partially and unphysically treated as a surface salt exchange when fully coupled. Parameter choices are evaluated by comparing climatologically forced solutions to observations and process model results. The two major criteria satisfied by the implementation in a fully coupled climate model and an uncoupled global ocean model are stable solutions and projection of the overflow signal across the Atlantic basin at about 1000 m depth. Both of these configurations are low resolution, and in both the transports of inflow, source and entrainment water are all within the range of observed estimates, but there is too little product water. This bias is attributed to inadequate modeling of water masses in the Mediterranean source region. Nevertheless, the properties of the product water differ little from observed estimates and both the uncoupled and coupled models develop a Mediterranean salt tongue that spreads west and south from the Strait with a signature reminiscent of the observed hydrography. The improvements relative to either blocking the Strait, or excavating a too wide channel are presented. In the coupled solution, the impact of the improved overflow physics on the global climate is minimal, with North Atlantic sea surface temperatures and heat fluxes changing generally by less than 1 °C and 15 W m−2, respectively. However, there is interesting spatial variability in the coupling strength, which ranges between ±20 W m−2 °C−1 in the coupled case.