Abstract

The basic principles and concepts of retrieval theory are straightforward. The problem is to find a profile (or to characterise a class of profiles) that is consistent with remote measurements of the atmosphere, within our present understanding of the physics of the measurement. The general approach is to construct a forward model of the measurement process, and then to design a numerical method (usually iterative) which finds a profile to more or less reproduce the measurements when it is inserted in the forward model.

© 1990 Optical Society of America