What We Have Learned from ROMEX: Processing and Evaluation of ROMEX Data from NOAA/STAR

With more than 35K occultation profiles per day collected from multiple Global Navigation Satellite System (GNSS) Radio Occultation (RO) missions, the Radio Occultation Modeling Experiment (ROMEX) provided a unique opportunity for the RO community to examine the quality of RO data collected from multiple RO missions, which are with different sensor characteristics and orbits. The ROMEX RO missions include Metop B, C (GRAS) COSMIC-2, Spire, FY3-C, D, E (GNOS), GeoOptics, PlanetiQ, Yunyao, Tianmu, KOMPSAT-5, PAZ, TerraSAR-X, TanDEM-X, and Sentinel-6. These missions use different receivers with various antenna designs, operating in different orbits and having varying signal noise ratios (SNRs). Meanwhile, the GeoOptics and Spire data collect GNSS RO measurements from CuteSats, and PlanetiQ data from SmallSats. NOAA has included GNSS RO data as one of the crucial long-term observables for weather and climate applications, just as those from IR and MW measurements. Critical questions for ROMEX are

1. Can we use the RO data collected from commercial CubeSats/SmallSats with other RO missions in the neutral atmosphere for climate and weather prediction studies
2. What may be an optimal approach to characterizing the uncertainty of GNSS RO data, especially in the lower troposphere

NESDIS STAR has become a GNSS RO processing and science center. Using the independently developed inversion packages, NESDIS/STAR processed the primary GNSS RO missions and distributed the data freely to the ROMEX data assimilation teams for further numerical weather prediction experiments. This study demonstrated how we developed consistent quality control criteria for all RO missions. We also carefully investigated the penetration depth, precision, stability, and accuracy among different missions. To demonstrate if RO measurements from different receivers provide consistent RO bending angles, we collect all available co-planar RO pairs among various RO missions to quantify their stability. In addition, we also calculated the estimated dynamic bending angle observation errors (DBAOE) using the local spectral width (LSW) approach. The bending angle observation errors affect how the RO data may be optimally used for weather forecasts in the data assimilation system. We focused on characterizing the uncertainty of GNSS RO bending angle errors in the lower troposphere. The initial results demonstrate that not all RO data from multiple missions are of the same quality. The antenna design and viewing geometry do affect the SNRs' distribution, which may, in turn, affect the retrieval results. We also discuss how results from this study may be beneficial for NOAA’s optimal configurations for future RO architecture missions.

The manuscript contents are solely the opinions of the authors and do not constitute a statement of policy, decision, or position on behalf of NOAA or the U.S. government.