Footprint model development and validation for homogeneous and inhomogeneous terrain using high resolution large eddy simulation
Zusammenfassung der Projektergebnisse
A new footprint evaluation algorithm that makes use of a Lagrangian stochastic model embedded in the parallelised large-eddy simulation model PALM has been developed in the framework of this project. For the first time an online coupling between the Lagrangian stochastic model and the large-eddy simulation model has been realised. The purpose of the development of the new footprint evaluation algorithm was the validation and advancement of conventional footprint models. The Lagrangian stochastic model driven by data from the large-eddy simulation model reproduces the results of former dispersion experiments in a convection tank and the results of numerical dispersion experiments. Moreover, the Lagrangian stochastic model can successfully be applied to the evaluation of footprints. The new model can be applied under a wide range of thermal stratifications and for measurements at arbitrary heights within the atmospheric boundary layer. The results of the footprint simulations of the large-eddy simulation model were compared to conventional Lagrangian forward and backward models. For this purpose a new system of 2D footprint comparison with a quality classification was developed. Both the size of the footprint and the location of the effect levels were taken into account. The agreement is generally better for intermediate measurement heights and for convective cases, whereas the two conventional flux footprint models agree best under near neutral conditions. The results of footprint calculations with the new developed Lagrangian stochastic model point out the existence of the phenomena of negative values of the flux footprint even in a horizontally homogeneous convective boundary layer. The negative values of the flux footprint are attributed to the typical dispersion pattern of a passive scalar that is released near to the surface into a highly convective boundary layer. Under neutral and stable stratification the veering of the wind with height that is often neglected in conventional footprint models explains the different orientation of the source areas for measurements at two different heights. As well, for the conventional Lagrangian footprint models, in addition to the extensive negative areas due to dominating downdraughts, negative contributions because of stochastic noise also have to be considered. As a large-eddy simulation model provides information on the turbulent flow field directly to the Lagrangian model, the new footprint evaluation algorithm can also be used to estimate footprints for measurements in horizontally thermally and / or aerodynamically heterogeneous terrain. Considerable differences compared to the results of conventional footprint models that assume a homogeneous flow field can be observed in a horizontally heterogeneous terrain. This possibility was applied for comparison with conventional Langrangian footprint models under heterogeneous situations. Therefore aerodynamical and thermal heterogeneous forcing was investigated for ideal conditions and for similar conditions during the LITFASS-2003 experiment. In the case of the absence of secondary circulations it is mainly the backward model that calculates comparable footprints in relation to the large-eddy simulation.
