The next intercomparison workshop of working group 1 of GCSS will be on non-precipitating (shallow) cumulus convection. Several individual LES studies of cumulus convection have been done in the past. However, no attempt have been undertaken sofar to do a systematic intercomparison of this type of boundary layer clouds . The questions we would like to answer during the next workshop are threefold:

• 3d Large Eddy Simulation Models:

  What is the quality of the (thermo)dynamics produced by LES models for shallow cumulus convection? Special emphasis will be put on output relevant for cumulus parameterizations of GCM's. This includes both diagnostic output that can be used to evaluate the overall performance of the various cumulus parameterizations, as well as output (mass flux, interaction between clouds and environment) that can be used to evaluate the various parametric assumptions. Main question will be whether the LES output is reliable enough for this type of convection to test and improve cumulus parameterizations.

• 2d Cloud Resolving Models:

  2d CRM's are often used as an alternative for 3d LES models for the obvious reason that they can be used on a larger horizontal domain and with a higher resolution. In the previous GCCS workshop on stratocumulus (ASTEX) the 2d CRM's gave reasonable results. An interesting topic is to see how well such 2d model behave for a more intrinsic 3d situation, such as the present cumulus case. Special emphasis will be put on the interaction between cumulus clouds and the environment in this respect.

• One-Column Models (Cumulus Parameterizations):

  Quite different approaches for parameterizing cumulus convection coexist now for more than 10 years. Both the mass flux approach and the more traditional K-theory are still common practice. A systematic intercomparison combined with comparison with the LES (presupposed these give sufficiently coherent results) will hopefully shed some new light on this topic. More specifically, the LES output will also be used to evaluate parametric assumptions made in the various parameterizations.