Balsmeier, Max H. (2017) Entwicklung eines barotropen Atmosphärenmodells und dessen Anwendung auf theoretische Szenarien und Reanalysedaten. (Bachelor thesis), Christian-Albrechts-Universität Kiel, Kiel, Germany, 47 pp.

Text Bsc. 2017 Balsmeier,M.pdf - Published Version Restricted to Registered users only Available under License German copyright act UrhG. Download (6MB) | Contact

Abstract

In this thesis a simple atmospheric model is developed in order to understand basic aspects of numerical weather prediction. The starting point is the füll vorticity equation, which is simplified such that a numerical solution is possible without much expense. The procedure to do this involves a system oflinear equations, which evolves out of the model equation by discretisation. Unfortunately, the unmodified model is unstable, so that the forecast diverges to infinity after a number of time steps that depend on initial and boundary conditions. As a consequence, the numerical procedure is extended by an additional diffusive term. The function and consequences of this modi:fication are discussed. In order to examine the model, short and long Rossby waves (wavenumbers two and eight) as well as ERA-Interim-reanalysis data (Dee et al., 2011) are used. Timedependent and constant boundary conditions are applied to the model to investigate the impact of the boundary conditions on the forecast using maps as well as quantitative calculations. The model is run under three different assumptions (f plane, ß plane and a linearised dynamic) to understand the importance ofthe different terms in the model equation, which represent different dynamical processes. Concerning the Rossby waves, the exact solution of the model equation is known, so that the phase velocities of the simulated waves can be compared with theoretical expectations. Here, profound deviations from theory are observed. The ERA-Interim data is used to run the different model versions with "real" data. The results show, that the model is approximately as reliable as the "persistence model" when constant boundary conditions are used. When time-dependent boundary conditions are used in the model, the prediction improves. Possibly, details in the grid structure and in the discretisation technique are responsible for the discrepancies between the model and theoretical expectations as well as for the poor performance concerning reanalysis data under constant boundary conditions. Investigating this further remains to be done.

Actions (login required)

View Item