The COSMO model (Consortium for Small-scale Modeling) is a non-hydrostatic LAM, designed for both operational Numerical Weather Prediction (NWP) and climate simulation. It is used and further developed for climate- related applications by the CLM-Community, an open international network of scientists (http://www.clm-community.eu), and it is referred to as the COSMO-CLM model (Rockel et al., 2008). There is an active collaboration between the NWP and climate communities. Since 2005, the COSMO-CLM model has been taking part in international coordinated regional climate modeling projects like PRUDENCE (Déqué et al., 2005). COSMO-CLM was present in the ENSEMBLES project and more recently CORDEX simulations have been performed with the COSMO-CLM model both at 12-km and 50-km resolutions.
A joint standard evaluation of the EURO-CORDEX RCM ensemble (Kotlarski et al., 2014) highlights the general ability of COSMO-CLM to represent the basic spatio-temporal patterns of the European climate and the good performance of the COSMO-CLM model compared to other ensemble members. Even with a mesh size of 12 km, as used in CORDEX, several processes are not resolved and have to be parameterized. This introduces model deficiencies and large uncertainties in the projections of future climate (e.g., Déqué et al. 2007). An important benefit increasing the mesh size to the kilometer-scale is that deep convection is (at least) partly resolved. Model simulations with COSMO-CLM have shown that the most important added value of these kilometer-scale integrations are found in the diurnal cycle with an improved timing of summer convective precipitation, the intensity of most extreme precipitation, and the size and shape of precipitation objects (Prein et al., 2013).
Recent simulations over the Belgian region have confirmed this added value (Brisson et al., 2016a,b). Experiments with COSMO-CLM also suggest that a more detailed representation of cloud physics (2-moment schemes) do not substantially improve most results at present, due to insufficient knowledge of the relevant parameters (van Weverberg et al., 2014), but appears valuable for the simulation of hail reaching the surface. In addition, kilometer-scale model integrations open the avenue to take into account land-use changes like urbanization into the climate projections for the future. The computationally efficient urban land-surface scheme TERRA_URB was recently implemented in COSMO-CLM (Wouters et al., 2016).