Seminar: Implementing dynamic Smagorinsky closure in ICON

Abstract

Turbulence parameterization is essential for representing the effects of unresolved atmospheric motions in numerical weather and climate models. The Dynamic Smagorinsky Model (DSM), which was adapted for the spectral model KMCM, is being transferred to the ICON (ICOsahedral Non-hydrostatic) atmospheric model. A central challenge is the adaptation of the filtering procedure used in the spectral model to the unstructured triangular grid of ICON.

In this work, a coarse-graining-based filtering approach has been developed for the ICON grid and a diagnostic formulation of the DSM was implemented to dynamically determine the Smagorinsky coefficient. The dynamically computed coefficient was subsequently coupled to the horizontal diffusion scheme of the ICON dynamical core, allowing the turbulence parameterization to interact consistently with the resolved flow.

The implementation has been extended to a prognostic formulation and tested in global ICON simulations. A stable three-month simulation was performed to evaluate the spatial and temporal variability of the dynamically determined Smagorinsky coefficient and its influence on the resolved atmospheric flow. The ongoing analysis focuses on the impact of the dynamic closure on turbulence and resolved gravity-wave activity, with simulations using the conventional constant-coefficient Smagorinsky formulation providing a reference where feasible.