Work package M1: Implementation and validation of turbulence parameterizations


Contact information

Principal investigators: Dr. Björn Maronga mail.gif, Prof. Dr. Siegfried Raasch mail.gif

Staff: M.Sc. Tobias Gronemeier mail.gif

Institution: Institute of Meteorology and Climatology (http://www.muk.uni-hannover.de/), Leibniz Universität Hannover (https://www.uni-hannover.de/)

Goals of WP-M1

The main goal of this subproject is to add a RANS (Reynolds-averaged Navier Stokes) mode to the model. A RANS mode is able to calculate the mean-flow condition of the atmosphere while parameterizing the atmospheric turbulence. This reduces the computational cost of a simulation drastically compared to a large-eddy simulation (LES) where the most energetic turbulent motions are directly calculated while only smaller turbulent motions are parameterized. The used parameterizations of the turbulence for the RANS mode need to be scale dependent as the model will operate under different spatial resolutions ranging from 1m up to 100m. For coarser resolutions building and vegetation effects must be parameterized as well because these structures might not be resolved at coarser resolutions. As the base model PALM is an LES model it already includes a subgrid-scale (SGS) model to parameterize the unresolved part of the turbulence in LES mode. However, this SGS model needs to be modified to account for a broader range of unresolved turbulence.


Work program

WP-M1.1: Selection of RANS turbulence schemes

The first work package includes the selection of a turbulence parameterization scheme for the RANS mode to be included to the model. Several turbulence schemes exist in other RANS models from which to choose. Also a proper parameterization scheme for buildings and vegetation needs to be chosen prior to the implementation.

WP-M1.2: Implementation and testing of parameterizations

The schemes selected in work package WP-M1.1 will be implemented to the model. The online analysis tools which already exist in the base model PALM will be updated to present all relevant information also for the RANS mode. Since within the RANS mode the mean-flow condition is simulated, the simulation should automatically stop after a stationary state is reached. Testing of the parameterization will be done by simulating simple test cases like a single cube or array of cubes where there are many results in the literature to compare with. Results of measurements e.g. from wind tunnel experiments might also be used for testing if data will be available.

WP-M1.3: Development and installation of an improved SGS model

The SGS model currently implemented in the base model PALM is used to parameterize the SGS turbulence in LES mode. This parameterization, however, only gives sufficient results if more than 90% of the turbulent fluxes are resolved by the grid. The new SGS model should be applicable for a wider range of unresolved turbulent fluxes as well as being computationally efficient and scalable (i.e. work efficient on a parallel-computing system).

Last modified 3 years ago Last modified on Sep 12, 2016 10:52:53 AM
                                                                                                                                                                                                                                                                                                                                                                               
  | Impressum | ©Leibniz Universität Hannover |