Design and validation of the GEOS ozone assimilation system

An ozone data assimilation system (DAS) that provides global 3-dimensional ozone fields has been developed at the Data Assimilation Office of NASA/Goddard. Ozone observations from Total Ozone Mapping Spectrometer (TOMS) and Solar Backscatter Ultraviolet (SBUV) instruments are assimilated into an ozone transport model. Ozone transport is driven by winds from the Goddard Earth Observing System (GEOS) core DAS. After every model time step (15 minutes) observations are combined with the model using a global physical space based statistical analysis scheme. An anisotropic forecast error correlation model motivated by the observing geometry of instruments on polar orbiting satellites is implemented. This model provides a better fit, in the maximum likelihood sense, to the realizations of observed-minus-forecast residuals than any isotropic model. The system has been validated for winter 1992. The analyzed ozone fields agree well with independent Halogen Occultation Experiment (HALOE) and ozone sonde measurements. The statistics of observed-minus-forecast residuals show good agreement of these residuals with forecast and observation error covariance models used in the ozone system.