Application of the Optimal Probability Method to the Retrieval of Temperature, Water Vapor and Ozone Profiles

A code has been developed to perform the retrieval of atmospheric state parameters using the method of nonlinear least squares in conjunction with a program to compute the forward problem (Isaacs, 1988). In the present study, FASCODE (Clough et al., 1986) has been utilized for the forward problem other algorithms including rapid algorithms may readily be accommodated. The method has been applied to retrievals of temperature profiles, surface temperature and pressure, water vapor profiles and other constituent distributions using both real and simulated data. The method is applicable to both sequential and simultaneous retrievals. Either of two approaches may be selected for performing retrievals: the method of ridge regression or the maximum likelihood method. Our implementation of the latter approach is similar to that discussed by Rodgers (1976, 1987) and more recently by Eyre (1989).