Overview of the GEWEX Atmospheric Boundary Layer Study

The aim of the GEWEX Atmospheric Boundary study (GABLS) is to improve the understanding and representation of the atmospheric boundary layer in weather-forecast and climate models. As such intercomparison studies have been organised for Single column models and Large Eddy Simulation models for an arctic site (GABLS1) and for a dry land surface case in CASES99 (GABLS2). The first two cases have been relatively simple in that the boundary layer was isolated from the surface by prescribing surface temperature and simplified external forcings. As such a quantitative comparison with observations was less meaningful and troublesome. For the third GABLS case an attempt is made to define a case with realistic forcings such that a direct comparison with observations is possible. The third GABLS inter-comparison case uses data gathered at Cabauw, NL. At this site on many nights a low level jet (LLJ) develops due to decoupling of the atmosphere from the surface and due to inertial oscillation. A relatively ideal, clear sky, baroclinic one day period (July 1, 2006) was selected from the multi-year data archive of Cabauw to explore the performance of the various models in comparison with the tower and remote sensing observations. Detailed dynamic forcings and surface conditions were prescribed on the basis of local observations and the outcome of 3D-atmospheric models. In total 12 institutes with 18 models ran their single column models with interactive land surface soil-vegetation schemes for the full day (including the difficult evening and morning transitions). In addition an intercomparison of LES models is made based by more than ten LES modeling groups on the nighttime and morning hours. In the presentation, we will describe the setup of the intercomparison case along with an overview of the results and conclusions. Emphasis will be given on the single column model performance and the open issues. Subsequent abstracts are submitted within this theme on the more detailed evaluation of the Single Column Models (Bosveld et al) and the LES models (Basu et al).