Statistically average atmospheric bending angle model based on COSMIC experimental data

The retrieval of profiles of meteorological variables from radio occultation observations requires knowledge of bending angle profiles up to heights of no less than 60-70 km. Because of the residual error of the ionospheric correction, retrieved profiles become too noisy by a height of about 40 km. In order to invert the bending angle profiles, the statistical optimization is used. This makes it possible to construct an optimal linear combination of the a priori estimate of the average bending angle profile and a posteriori noisy estimate based on observations. The estimate of the average bending angle profile for the given coordinates and the time of year is usually based on the climatological atmospheric model. MSIS and CIRA models that have been used for this purpose are now obsolete and do not describe the global changes in the atmospheric state. The model of average bending angles BA-IAP (Bending Angle-Institute of Atmospheric Physics) is built based on the processing of the array of COSMIC radio occultation observations during 2006-2013. The proposed model is statistically validated based on the COSMIC database. It is shown that our model describes the average bending angle profiles more accurately than the MSIS model.