Validation methods to study the consistency and quality of radio occultation electron density profiles: application to COSMIC

Radio occultation (RO) is a relevant source of information from the atmosphere. Besides providing global coverage, due to the geometry of the data acquisition, RO provides measurements that can help to suppress gaps from other techniques. In this sense, RO data assimilation has potential to improve atmospheric products such as ionospheric models and numerical weather prediction. Constellation Observing System for Meteorology, Ionosphere and Climate (COSMIC) (2006-2020) has been one of the main RO missions, with significant number of atmospheric profiles available, especially considering the ionosphere. The ionosphere is of special relevance because it can influence the accuracy of global navigation satellite systems (GNSS) and related applications. This way, the assessment and filtering of RO data is crucial in order to identify profiles with questionable information. Many investigations have been developed aiming to provide methods of validation for RO profiles, however, no clear methodology for filtering the RO data can be easily found. In this context, in this work, seven RO filtering methods are applied including manual filtering of noisy data and discrepancies considering the first principles-based Chapman model in a normal distribution. The set of strategies using the normal distribution criteria leads to large rates of profiles exclusion (close to 90 % in some scenarios), while in most of the cases the foF2 differences do not show improvement. On the other hand, the strategy with manual filtering, in general, excludes 35 % of the profiles, leading to gain of about 7 % in the foF2 error.