Dynamic approach to meteorological data assimilation

The methods of four-dimensional assimilation of meteorological data are discussed on the basis of theoretical considerations as well as of numerical results obtained with two different models. The fundamental role of geostrophic adjustment and of the damping of inertia-gravity waves is particularly studied. A number of general dynamic concepts are defined, which lead to a possible understanding of the results of data assimilation experiments in terms of three parameters: the predictability of the reference flow by the model, the gravity wave damping time constant and the data acquisition rate. The equivalence between synoptic and asynoptic data is discussed, and proved in one particular case. Finally, “intermittent” and “continuous” assimilations are compared.