Last updated: 2007
see status report
The LM_Z Priority Project was planned to end in August 2007 with delivering a report about the performance of the Z-coordinate version of the COSMO-Model. Due to problems, which were detected during running the tests and experiments, the report is not ready yet. Therefore, work in writing this report is going on and it will hopefully delivered soon. Nevertheless, we did not put up a new working plan and no more FTEs are allocated during 2007/2008 for LM_Z.
Z-coordinate numerical models of the atmosphere have the advantage of representing the atmosphere at rest properly and therefore justify the expectation of improved forecasts of orographically induced winds. An example for such a Z-coordinate model is the eta-model, which is based on step-orography. The step approach does not allow for a proper representation of the meso-scale flow over smooth and well resolved mountains. This problem was pointed out by Gallus and Klemp (2000). Steppeler et al. (2002a) showed that the problem can be solved by formulating lower boundaries using a representation of the mountains by linear splines rather than the step approach. On this basis a Z-coordinate version of the model LM (see Steppeler et al. (2002b)), LM_Z, was developed and tested extensively in idealised situations using bell shaped mountains and the orography of Scandinavia.
In the last years, the dynamical core of the LM_Z has been developed in a first version. The physical parameterization package is used from the terrain following model. For that, an interface between the z-levels and the terrain-following levels has been implemented, which interpolates the tendencies of the physics routines to the terrain-following grid by cubic splines to the z-representation. Besides artificial test cases it is now also possible to run real data simulations and forecasts.
In 2005/2006 several test cases have been performed in Poland, Greece, Italy and Switzerland, which showed advantages of the LM_Z especially for the (quantitative) precipitation forecast. But also some shortcomings and bugs have been detected, which have to be improved, if not eliminated.
In the COSMO-year 2006/2007 the LM_Z has to be further developed and consolidated. Besides doing more test cases, real test suites for several weeks / months including data assimilation have to be performed to assess the quality of this model. A report about these tests has to be written. This Priority Project will end with delivering this report.
The source code development can be grouped in 3 categories:
Test Cases with the new versions will be run at HNMS, USAM and MeteoSwiss. At DWD some experiments for longer periods will be started.
After running the tests and experiments, the results will be published in a COSMO Technical Report.