Maximum Entropy Parameterization of Unresolved Scales in a Two-Layer Quasi- Geostrophic Model

J. Camphuijsen, Ministry of Infrastructure and Water Management, Royal Netherlands Meteorological Institute, Universiteit van Amsterdam
De Bilt : KNMI

Modeling turbulence in the atmosphere is essential for day to day weather prediction and for predicting climate change. An idealized but fairly accurate model of the atmosphere can be constructed by stacking multiple uid layers of uniform density on top of each other. The simplest of such models is the two-layer quasi-geostrophic model, that uses geostrophic balance to get a closed set of evolution equations. As it is computationally impossible to resolve all scales of motion, we need to truncate the model at a certain scale and describe the small scales by parameterization. A parameterization based on the principle of maximum entropy, that was developed by [Verkley et al., 2016], has been implemented and tested on a two-dimensional model. The new parameterization has no tunable parameters, as its dependence on the system is fully determined by the formalism. In this report, the two-layer quasi-geostrophic model and its energy budget are derived. The maximum entropy parameterization is adapted for this system using the constraint of quasi-geostrophic energy conservation in the unresolved scales. The parameterization is analyzed using multiple diagnostics for both short-term deterministic and lon-term statistical performance. The qualitative performance is compared with a conventional increased-viscosity parameterization. In the analysis, a simulation run of double resolution is used as reference.

38 p.
Fig., tab.
(Internal report ; 2018-02)
With a bibliogr.