Questions on the effects of roughness and its analysis in non-equilibrium flows

The prediction of turbulent flows over rough surfaces is important for many applications in engineering and in the natural sciences. Since the resolution of the roughness requires significant computational resources, most modelling approaches rely on the related concepts of 'equivalent sand-grain roughness' and 'wall similarity'. While the validity of these concepts is well established for zero-pressure-gradient boundary-layers and for channel flows, such is not the case for non-equilibrium conditions. This raises a number of important questions, some of which are discussed in this paper. We also suggest some possible paths to answering these questions.