| << | Validation of Measurements and Models | >> |
Abstract
Meteorological fields are required as indispensable input for air quality models. But they can be a source of significant errors which contribute to uncertainties in simulations of the atmospheric distribution of chemical species and aerosols. Wind velocity and wind direction are parameters of fundamental relevance for atmospheric dispersion and thus for chemistry transport modelling. Within the COST728 action a project has been initiated, in which wind profiler measurements are used to assess whether several regional-scale meteorology models widely in use for air pollution studies are properly simulating the spatial and temporal features imbedded in the observations. Wind profiler observations compared to e.g. radiosonde data have the advantage, that they are available with much higher time resolution (at least hourly) and that they provide quasi-local profiles. They allow for studies within the atmospheric boundary layer as well as in the free troposphere. In this study hourly wind profiler data from three European stations of the CWINDE wind profiler network: Pendine/United Kingdom, Cabauw/The Netherlands and Lindenberg/Germany for a target period in 2003 characterized by complex meteorology is used. Among the models evaluated during this study are MM5, WRF, KNMI-HIRLAM, COSMO, just to name a few. Beside the direct comparison of profiles, time series and distributions, we will present an analysis based on a suite of statistical measures for wind velocity and direction. In addition spectral analysis is employed for several altitude layers in order to assess the models capabilities to adequately resolve intraday fluctuations as they are obvious in the wind profiler data. Finally, some suggestions will be made to improve the interaction between the observational and modelling community in order to foster more frequent model evaluations based on sophisticated profiler data.