Thermodynamic climatology of the disturbed stratospheric polar vortex used for statistical optimization of radio occultation data

Statistical optimization (SO) is commonly applied for inversions of radio occultation (RO) data in the neutral atmosphe in order to reduce the effect of RO bending angle (BA) observational noise (ionospheric residuals, thermal noise and clock errors) at heights where it becomes comparable to the neutral atmospheric variations of BA (commonly above 30 km). The SO commonly uses a priori atmospheric climatological model which is linearly combined with the observed RO BA, by replacing larger observational errors by smaller errors of the model and thus reducing error propagation downward after the Abel inversion. However, errors of the model may significantly increase during extreme periods resulting in increased errors of RO retrievals. In this study, we focus on the disturbed polar vortex periods occurred in the Northern Hemisphere winter stratosphere, and develop a modification of the standard climatological model by using ERA-Interim Reanalysis data for the winter seasons of 1979-2014. We localize the polar vortex using 2D geometric moments from the ERA-Interim geopotential height at 10 hPa for each disturbance event, and apply the spatial filter to create an area average that represents the thermodynamic quantities of the air column inside the polar vortex. The associated area-averaged refractivities inside/outside the polar vortex are calculated and compared with those from the RO observations. The modified climatology which accounts for the disturbed stratospheric polar vortex will be implemented into the SO of RO inversions and the results will be evaluated against existing algorithms.