CESM Ocean Model Working Group (OMWG) Near-Term Plans
04 February 2013
The CESM OMWG Winter Meeting was held at NCAR on January 22-23, 2013. The meeting was scheduled for 1.5 days with the half-day session on January 23rd being joint with the annual meeting of the Climate Process Team on internal wave mixing. The meeting was very well-attended, with good talks and fruitful discussions. To further our collaborations with the CESM Biogeochemistry Working Group (BGCWG) in addressing ocean model biases, we also had representation and presentations from the ocean-focused members of the CESM BGCWG.
In our discussions, we critically evaluated progress towards our near-term (next 2-3 years) goals that were established during our December 2011 winter meeting (see OMWG December 2011 meeting summary). The planned POP2 algorithmic developments were i) elimination of virtual salt fluxes in favor of true freshwater surface fluxes, ii) implementation of a new vertical coordinate system, z*, and iii) introduction of a conservative Robert time filter to replace the time-averaging time step. While the latter allows efficient high-frequency coupling of the ocean model to resolve the diurnal cycle and inertial periods explicitly, the former developments are designed to better address sea level rise and land ice – ocean interaction science questions. The basic implementation of the conservative Robert time filter has been completed, but there is still work to be done in ensuring exact restartability, budget conservations, and implementation for passive tracers. These are expected to be completed by June 2013. The z* vertical coordinate implementation in POP2 is also progressing with the majority of work expected to be completed by June 2013. We need to assess the consequences of issues related to the lack of local conservation as well as determine whether pressure averaging is still feasible with our implementation of the z* vertical coordinate. The latter may have consequences for the efficiency of the model. During our discussions, we received a request from the community to make the partial bottom cell treatment fully operational in POP2. We need to assess the prioritization of this request in terms of resources, as it requires somewhat extensive modifications particularly in the mesoscale and submesoscale parameterization subroutines along with extensive testing.
Progress on Model for Prediction Across Scales Ocean Component (MPAS-O) has been rather impressive. The first paper describing the model formulation and some early results has been submitted to Ocean Modelling by the LANL model developers. In addition, MPAS-O has been brought into the CESM framework as a new model component. During 2013, our plan is to implement the KPP vertical mixing parameterization and the mesoscale mixing parameterization in MPAS-O. A primary goal is to entrain the involvement of the OMWG members in MPAS-O development and assessment of its results by first producing credible ocean – sea-ice simulations forced with the Coordinated Ocean-ice Reference Experiments inter-annually varying atmospheric datasets (CORE-II experiments). We note that the KPP development indicated above is part of our modularization of vertical mixing activities under the CVmix (Community Ocean Vertical mixing module) umbrella. This effort was initiated in December 2011 and involves NCAR, GFDL, and LANL collaboration. An additional task that we hope to complete during this year is improving computational efficiency of MPAS-O.
As in the past year, a significant amount of time during this year will be devoted to various subgridscale parameterization and model development activities, some as part of our CPT and NSF Earth System Model (EaSM) projects. These all involve work with our current ocean model POP2. They include a possible parameterization for topographic control of the Gulf Stream; estuary-shelf freshwater exchange parameterization; a parameterization for sea-ice heterogeneity and ocean mixing; a Langmuir mixing parameterization; completion of the anisotropic mesoscale eddy parameterization; and improvements to the near-inertial wave mixing, tidal mixing, and geothermal heat flux parameterizations.
During the past year, the high-resolution ocean model configuration (0.1° horizontal resolution) has been updated to have 62 vertical levels and the submesoscale parameterization. Our working group is committed to support this configuration as we anticipate increased usage of this version of the model with availability of the Yellowstone environment. At the opposite end, the OMWG offers a nominal 3° horizontal version of the ocean model, but the usage of the nominal 1° horizontal resolution version is encouraged if computer resources are available.
Finally, we plan to provide ocean model evaluation metrics along with updated diagnostics (including BGC variables) to the ocean model community. A caveat regarding our near-term plans is that the prioritization of these activities is largely dictated by available resources.