MS-GWaM - A three dimensional transient parameterization for internal gravity waves in atmospheric models

Internal gravity waves (IGWs) are important distributors of energy and momentum in a stratified atmosphere. While most IGWs are presumably excited at lower altitudes their effects are most important between the upper troposphere and the mesopause (~85km). During propagating - both in the vertical and the horizontal - nonlinear IGWs can exert a wave drag on the large-scale winds, interact with the large-scale potential temperature, and influence transport and mixing of atmospheric constituents such as aerosols or greenhouse gases.

In state-of-the art weather and climate prediction models subgrid-scale IGWs are typically parameterized neglecting both the horizontal wave propagation (single-column assumption), the transient wave behavior including its effect on wave-mean-flow interactions (steady-state assumption) as well as time dependent wave generation. While being computationally efficient the missing physics, however, may lead to model errors and inaccurate predictions under varying boundary conditions. The potential importance of the horizontal wave propagation and wave transience has been shown in various theoretical, numerical and experimental studies.

The transient Multi Scale Gravity Wave Model (MS-GWaM) - implemented in the high-top model UA-ICON - aims to improve these shortcomings by allowing for transient and three dimensional wave propagation. The parameterization is based on a multi scale WKBJ analysis of the compressible atmosphere and includes various non-orographic wave sources, non-dissipative wave-mean-flow interactions as well as wave breaking. Internally, the parameterized gravity waves are treated as Lagrangian volumes with the dynamics prescribed by the well known gravity wave modulation equations. A suitable projection method of wave properties onto the unstructured model grid facilitates the calculation of wind and temperature tendencies. What is more, an efficient parallelization of the ray-tracing scheme allows for simulations in reasonable computation times, being much faster than corresponding wave resolving runs.

While satisfactorily reproducing the observed zonal-mean wind and potential temperature climatology the model results reveal new insight into the detail of the role of IGWs in the atmosphere. In particular, probability density functions of wave momentum fluxes exhibit the typical observed long tails (i.e. wave intermittency) which cannot be reproduced with steady-state parameterizations. Moreover, the three dimensional distribution of wave momentum and wave action fluxes differ greatly when relaxing the single-column assumption. As an example the well known three dimensional refraction of IGWs into polar jets can be shown.