Advancing Processing Algorithms for Excess Phase and Bending Angle Extraction in Multi-RO Missions at NOAA/STAR

The importance of Global Navigation Satellite System (GNSS) Radio Occultation (RO) data continues to grow due to its vital contributions to weather forecasting and climate analysis. Extracting accurate excess phase and bending angle information from GNSS RO signals remains a formidable challenge, given the intricate atmospheric conditions and the complexity of multi-GNSS signals. The expanding landscape of RO missions, including small satellites like Spire, GeoOptics, and PlanetIQ alongside established ones such as Metop-B/C, COSMIC-2, and Sentinel-6, has led to a diverse array of RO receivers deployed in space, each with varying quality.

While processing algorithms share a common framework across missions, the unique observational characteristics of each sensor and mission-specific demands require tailored processing approaches and specialized interfacing software.

Currently, NOAA Center for Satellite Applications and Research (STAR) has developed capabilities as a GNSS RO science and data center (STAR RO DSC, see https://www.star.nesdis.noaa.gov/smcd/GNSSRO/RO/index.php). This paper presents the development of an independent algorithm for processing excess phase and bending angle data, explicitly designed for multi-GNSS RO missions at NOAA/STAR. The objective is to comprehend uncertainties introduced at each processing stage, from pseudo-range and carrier phase observations to excess phase extraction and from excess phase data to bending angle profiles. The Bernese version 5.4 software addresses challenges related to Low Earth orbit (LEO) satellite clock bias and precise orbit determination (POD). Our algorithm processes excess phases and bending angles, featuring the STAR full spectrum inversion (FSI) algorithm that converts excess phases into bending angles. This FSI algorithm has been fully integrated into the Radio Occultation Processing Package (ROPP) version 10.0, offering users an alternative to wave optics and geometry optics through configurable settings.

In this study, a comprehensive comparison of each processing stage will be presented, comparing our approach against the UCAR method for COSMIC-2 and Spire data. The study characterizes different RO receivers with Signal-to-Noise Ratio (SNR) values, clock stability, satellite orbits, and receiver open-loop models across multiple RO missions. Furthermore, the study quantifies the discrepancies and uncertainties inherent in POD, excess phase determination, and bending angle processing.