Abstract
A coupled one-dimensional ice-ocean model is used for studies of Arctic phenomena. The ice-snow system is represented by the simplified thermodynamic ice model of Semtner and a dynamic approximation that neglects the internal stresses. The ocean is represented by the Mellor–Yamada level-2 turbulence mixed-layer model together with a prescribed geostrophic velocity.
The thermodynamic coupling considers an ice front and a salinity flux generated by the freezing or melting of ice. The dynamic coupling occurs via the turbulent stress that exists in the mixed layer beneath the ice. Various boundary conditions for ice-ocean coupling are examined including an analytical representation of the constant flux layer.
Two test cases are used for model validation and scientific studies. One is the standard climatological test used by Semtner and others. The other test case is with the AIDJEX data.
The ice-ocean model is compared to Semtner's ice model to determine the effects of a variable-depth mixed layer as opposed to an isothermal, fixed-depth mixed layer. In the variable-depth mixed layer model, a warm spot develops in the surface layers of the mean during open water periods. When the ocean refreezes some heat remains in the warm spot and gradually diminishes as the ice continues to grow. This heat is released from the upper ocean through the mixing process. Its release significantly affects the heat budget and the growth rate of ice. Open water occurs nearly every year, in climatology simulations, as opposed to once every six years in the case Semtner examined.
Simulations of the AIDJEX Experiment predicted the general trends of the temperature and salinity measurements. Specific discrepancies may be due primarily to the omission of advection.
