An alternative to the traditional retrieval algorithms currently used in the retrieval of temperature and moisture profiles from GPS based radio occultation are retrieval methods based on optimal estimation principles. Such variational algorithms have become a quasi-standard in most fields of remote sensing data, since they have been proven to be comparatively robust against errors in both observational and a priori data. In addition, sensible error estimates of the derived geophysical quantities can be obtained. Theoretical investigations of their merits in the field of GPS based radio occultations have been presented by Healy and Eyre (2001) and Palmer et al (2001). There is, however, only limited practical experience with this type of retrieval algorithms in a quasi-operational environment for large volumes of data. We discuss the 1D variational retrieval developed for CHAMP radio occultation data. Results from the variational retrieval are compared with those from the traditional one, and comparisons with independent meteorological measurements (especially radiosondes) will be presented.

---

Marquardt, C.; Beyerle, G.; Healy, S. B.; Schmidt, T.; Wickert, J.; Neumayer, H.; König, R.; Reigber, C. (2002): Variational retrieval of CHAMP radio occultation data. 27th General Assembly European Geophysical Society (EGS) (Nice 2002).