Advanced stratospheric data processing of radio occultation with a variational combination for multifrequency GNSS signals

The signal-to-noise ratio (SNR) is found to be a universal quality indicator, which concisely describes the uncertainty of RO measurements in diverse conditions. The measured SNR is used to parameterize a dynamic observation error, which is essential for the VAR to use the observation optimally. Tests with real data show that VAR significantly improves the accuracy of the RO retrieval even in the upper stratosphere, where the RO data were once considered to possess little observational value.

When compared with independent radiosonde observations, for instance, the VAR-produced data are more accurate than the analysis from the European Center for Medium-Range Weather Forecasts for which the radiosonde data have been assimilated. The VAR-produced data are also precise enough to reveal the systematic error of the radiosonde data.