OP-E4.2 Ocean Vertical Mixing
| Poster Title | Moving toward an improved modeling of vertical turbulent heat and momentum fluxes in the ocean. |
|---|---|
| Authors | Luke Van Roekel (Unlicensed), Qing Li (Unlicensed), Rachel Robey (Unlicensed) |
| First Author | Luke Van Roekel (Unlicensed) |
| Session Type | E3SM Session |
| Session ID | E4 |
| Submission Type | Presentation |
| Group | Water Cycle / Cryosphere |
| Experiment | Cross-cutting |
| Poster Link |
Abstract
More than 90% of the anthropogenic emissions from the past 50 years has been absorbed by the ocean. The rate of absorption is critically dependent on small-scale turbulent fluxes of heat and momentum in the surface boundary layer of the ocean (O(100s)m deep). An accurate representation of these critical fluxes across a wide range of forcing conditions has been pursued for more than 50 years. Currently, many ocean models utilize the K-Profile Parameterization (KPP) to represent turbulent vertical fluxes of heat and momentum. KPP has performed reasonably well for coarse resolution ocean models. However, in critical regions (e.g. near the coast and high latitudes), KPP can exhibit large biases.
Our testing has shown that KPP biases can result from incomplete or missing physics and/or structural deficiencies of the model itself. For example, turbulence driven by wave-mean flow interaction (Langmuir Turbulence) is neglected despite evidence that it has a large impact on simulated climate fidelity. Structurally, KPP’s critical dependence on a diagnosed boundary layer depth to determine the turbulent heat and momentum fluxes cannot be remedied easily.
We examine the influence of Langmuir Turbulence on the ocean climate simulated by the Energy Exascale Earth System Model (E3SM). A simple empirical wave model is included in the ocean component of E3SM to parameterize the effects of Langmuir Turbulence without coupling with a full wave model (e.g., WaveWatch 3) which can be computationally expensive. Using this simple model, we are able to reduce mixed layer biases with virtually no additional computational cost.
We also present initial results from a new closure for ocean models that casts a number of terms from the traditional Reynoldsaveraged Navier-Stokes (RANS) equations in a mass flux framework. This allows for accurate modeling of additional terms in the turbulent heat and momentum flux budgets at a fraction of the cost of RANS models. Notably, our new closure does not exhibit resolution dependent bias and is not sensitive to a diagnosed boundary layer depth, which can manifest large temporal and spatial oscillations, instead using it as a passive diagnostic quantity.
Approved for unlimited release, LA-UR 18-28985