Pierre Aumond

As acoustic propagation, forest fire (Pimont et al., 2006), tree stability (Dupont and Brunet, 2008b) or pollen dispersion (Chamecki et al., 2009), a lot of applications needs an accurate representation of the boundary layer to observe its influence on these phenomena. Then, in order to better represent the dynamics into the surface layer, Large-Eddy Simulations (LES) are carried on using thinner and thinner vertical resolutions until the order of the meter (Bohrer et al., 2009, Dupont and Brunet, 2008b). In these cases, influence of some objects (e.g. trees, buildings, ...) on dynamics should not be taken into account only through surface schemes as usually done in mesoscale meteorological model but directly in the dynamic equations of the model. In this way, instead of using the current roughness approach, a drag force approach has been implanted in some meteorological mesoscale models: MM5 (Otte et al., 2004), ARPS (Dupont and Brunet, 2008a), RAMS (Bohrer et al., 2009). This study presents its implementation in Meso-NH. Here, the objective is to compare LES results with a real case experiment over a large domain (until 10 km2). In this way, the originality of this study is to quantify the positive impact of drag force approach on LES boundary layer representation in real case situation. After a discussion in the Section 2 about the formulation and implementation of the scheme into the research mesoscale model (Meso-NH), section 3 consists to validate it thanks to numerical predictions issued from (Dupont and Brunet, 2008c) and (Shaw and Schumann, 1992). Large Eddy Simulations at very fine scales (until 2m) over real case are presented in section 4. Simulation results are confronted, with and without the canopy scheme, to observations from Lannemezan-2005 experience (Junker et al., 2006). At last, Section 5 provides conclusions and implications for future work. 2. CANOPY PARAMETERIZATION