Cloud forecast verification with satellite observations

esoscale models offer an ideal framework for performing detailed and explicit simulations of cloud and precipitation. First, these models are able to follow the time evolution of different moments of the hydrometeor distributions in the context of real meteorological conditions. Second, the grid mesh of a mesoscale model is of the same size as a satellite pixel, facilitating a comparison without it being necessary to resort to additional assumptions on size. Here we adopt a model-to-satellite approach, in which satellite brightness temperature (BT) images are directly compared to BTs computed from predicted model fields. The approach is especially powerful in identifying discrepancies of cloud cover forecasts with BTs at 10.8 µm. The model-to-satellite approach associated with the BT difference (BTD) technique can also verify specific forecasts such as cirrus cover, dust occurrence, convective activity and overshoots. Applications of the approach to the French mesoscale model MESO-NH will be shown using MSG and AMSU observations. In particular, the BTD technique leads us to improve the cloud scheme of the MESO-NH model by tuning a critical parameter in a cirrus parameterization. The model-to-satellite approach is further combined with the calculation of meteorological scores for an objective evaluation of the model forecasts. The applications of the approach will be shown in the context of case studies and meteorological campaigns over the midlatitudes and the tropics.