Abstract
The optimum statistical inversion technique is applied to the evaluation of Umkehr observations from Arosa (46.5N, 9.4E), Aspendale (38.0S, 145.IE) and Tallahassee (30.2N, 84.2W). A Priori statistical information on the vertical ozone distribution is obtained from 511 balloon-sounding profiles at Boulder (40.0N, 105.1W). These latter data are also used to develop a regression method for obtaining a first guess at the O2 distribution, using total O2 as the independent variable. The ozone profiles inferred from the statistical method, when compared with concurrent direct measurements, display a significant improvement over profiles derived by the method in use at the World Ozone Data Center, Most of this improvement is attributable to the better first guess based on the total O2 regression. It is concluded that the utility of Umkehr observations is mainly restricted to estimation of the high-level O2 distribution above the main O2 maximum. For inferences about the ozone distribution at and below the maximum, effort should be directed toward regression techniques using total ozone as the independent variable.
