Experiments in the Four-Dimensional Assimilation of Nimbus 4 SIRS Data

Abstract

Temperature data derived from Nimbus 4 SIRS radiance measurements have been used in four-dimensional assimilation experiments with a simple two-level primitive equation model. Data are inserted into a running forecast on an orbit-by-orbit basis. The method of insertion is held fixed but initial conditions and techniques for achieving geostrophic adjustment are varied. It is shown that the real temperature data can be readily assimilated and retained by the model, and that the assimilation is greatly enhanced if a simple wind law initialization is performed at the time of insertion. There is a suggestion that four-dimensional assimilation is a viable operational procedure and a somewhat milder suggestion that it is a more effective way of utilizing asynoptic data than updating procedures currently in practice.