An inter-comparison exercise of mesoscale flow models applied to an ideal case simulation

An exercise is described aiming at the comparison of the results of seven mesoscale models used for the simulation of an ideal circulation case. The exercise foresees the simulation of the flow over an ideal sea–land interface including ideal topography in order to verify model deviations on a controlled case. All models involved use the same initial and boundary conditions, circulation and temperature forcings as well as grid resolution in the horizontal and simulate the circulation over a 24-h period of time. The model differences at start are reduced to the minimum by the case specification and consist mainly of the parameterisation and numerical formulation of the fundamental equations of the atmospheric flow. The exercise reveals that despite the reduction of the differences in the case configuration, the differences in model results are still remarkable. An ad hoc investigation using one model of the original seven identifies the treatment of the boundary conditions, the parameterisation of the horizontal diffusion and of the surface heat flux as the main cause for the model deviations. The analysis of ideal cases represents a revealing and interesting exercise to be performed after the validation of models against analytical solution but prior to the application to real cases.