The Xiamen Symposium on Marine Environmental Sciences (XMAS 2025)

Various ocean boundary layer vertical mixing parameterizations have been developed and a comparison among them is useful in revealing the defects of parameterizations and guiding future development. Such comparison is usually done by starting one-dimensional or three-dimensional simulations with different parameterizations from the same initial condition and running the simulations with the same forcing conditions. Differences in the simulated mean fields (e.g., velocity, temperature, salinity, etc.) and turbulent fluxes are then compared. While such approach is straightforward, errors in the simulated mean fields accumulate and are entangled with errors in the simulated turbulent fluxes. Sometimes they can partially cancel with each other and may not truly reflect the parameterization deficiencies. To address this, we directly compare the simulated vertical turbulent fluxes in single column simulations with different parameterizations and that in corresponding large eddy simulations (LES), while strongly nudging the mean fields to the LES throughout the simulations. By doing so, we have a better control that the mean fields in the single column simulations are always consistent with those of the LES during our simulations. Then, the deviations of turbulent fluxes from the LES can more fairly reflect parameterization defects. In addition, it can be regarded as an estimation of the nonlocal fluxes that a parameterization may not capture. We start with idealized open ocean cases, such as pure convection and pure shear scenarios, to demonstrate the usefulness of the method. Subsequently, we apply this method in an idealized merging boundary layer scenario, where the interaction between a surface boundary layer and a bottom boundary is important. Finally, we apply it to a more realistic case with initial conditions and forcings from the Ocean Station Papa for further method testing and parameterization comparison.
 

ocean boundary layer turbulence, mixing parameterization comparison