Comparison of ocean vertical mixing schemes in the Max Plank Institute Earth System Model (MPI-ESM1.2)

We compare the effects of four different ocean vertical mixing schemes on the ocean mean state simulated by the Max Planck Institute Earth System Model (MPI-ESM1.2) in the framework of the Community Vertical Mixing (CVMix) library. Besides the PP and KPP scheme, we implemented the TKE scheme and a r...

Full description

Bibliographic Details
Main Authors: Gutjahr, Oliver, Brüggemann, Nils, Haak, Helmuth, Jungclaus, Johann H., Putrasahan, Dian A., Lohmann, Katja, Storch, Jin-Song
Format: Text
Language:English
Published: 2020
Subjects:
Online Access:https://doi.org/10.5194/gmd-2020-202
https://gmd.copernicus.org/preprints/gmd-2020-202/
Description
Summary:We compare the effects of four different ocean vertical mixing schemes on the ocean mean state simulated by the Max Planck Institute Earth System Model (MPI-ESM1.2) in the framework of the Community Vertical Mixing (CVMix) library. Besides the PP and KPP scheme, we implemented the TKE scheme and a recently developed prognostic scheme for internal wave energy and its dissipation (IDEMIX) to replace the often assumed constant background diffusivity in the ocean interior. We analyse in particular the effects of IDEMIX on the ocean mean state, when combined with TKE (TKE+IDEMIX). In general, we find little sensitivity of the ocean surface, but considerable effects for the interior ocean. Overall, we cannot classify any scheme as superior, because they modify biases that vary by region or variable, but produce a similar pattern on the global scale. However, using a more realistic and energetically consistent scheme (TKE+IDEMIX) produces a more heterogeneous pattern of vertical diffusion, with lower diffusivity in deep and flat-bottom basins and elevated turbulence over rough topography. In addition, TKE+IDEMIX improves the circulation in the Nordic Seas and Fram Strait, thus reducing the warm bias of the Atlantic water (AW) layer in the Arctic Ocean to a similar extent as has been demonstrated with eddy-resolving ocean models. We conclude that although shortcomings due to model resolution determine the global-scale bias pattern, the choice of the vertical mixing scheme may play an important role for regional biases.