An Overview of Meteorological Data Assimilation

Abstract

The process of numerical weather prediction is classically viewed as an initial value problem whereby the governing equations of geophysical fluid dynamics are integrated forward from fully determined initial values of the meteorological fields at some initial time. Given the mathematical properties of the equations of motion applied to geophysical fluids and the complexity of energetic processes in the atmosphere, solving this “initial value problem” is by itself a tremendous task: many a numerical and physical “modeler” is currently engrossed in it. Still, providing the ultimate scheme for integrating the discretized equations of atmospheric dynamics would only be half the answer. For one must also attend to the problem of determining, from observations of the real atmosphere, the initial values of the many time-dependent quantities which define the state of a discrete, numerical, atmospheric model.KeywordsNormal ModeGravity WaveData AssimilationNumerical Weather PredictionSlow ManifoldThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.