Confronting multi-year idealized LES with measurements at a mid-latitude meteorological site

Abstract:

Long-term Large-Eddy Simulation (LES) at permanent meteorological sites is increasingly being used for the evaluation and improvement of parameterizations of fast boundary-layer physics for weather and climate models. Typically, idealized LES are generated that represent fine-scale downscalings of a large-scale model state at locations of interest, resolved at temporal and spatial resolutions at which turbulence and boundary layer clouds can be expected to be resolved. The setup relies on prescribed large-scale forings in combination with continuous nudging, at a time-scale large enough to allow the resolved fast physics to have enough freedom to establish their own, unique state. This study critically assesses the representativeness of such long-term LES, and asks to what degree the continuous nudging affects the budgets of thermodynamic state variables. To this purpose long-term, multi-year LES is confronted with relevant observations at a European midlatitude continental site. The large-scale advective forcings and surface properties used to drive the LES are derived from analyses of the Integrated Forecasting System (IFS) of the European Centre for Medium-Range Weather Forecasts (ECMWF). Evaluated aspects of the boundary layer include near-surface meteorology, vertical structure, and bulk properties including clouds. The evaluation focuses on the diurnal cycle, with the oservational datasets derived from state-of-the-art instrumentation at the Juelich Observatory for Cloud Evolution (JOYCE) in Germany. Conditional sampling is used to highlight results for regimes of interest, including the clear convective and shallow cumulus topped boundary layer. We find that the LES is able to reproduce the observed amplitude and time-variation of key boundary layer properties, including clouds. Budget studies of the thermodynamic state variables reveal that a rough balance exists between the prescribed larger-scale advection and the turbulence/convection as resolved by the LES, with the nudging term only playing a minor, secondary role. These results support the use of idealized, continuous LES for the evaluation of parameterizations.