GNSS Radio Occultation Excess Phase Data including Integrated Uncertainty Estimation and Intercomparison between Processing Centers

Satellite remote sensing of the Earth’s atmosphere employing the Global Navigation Satellite System (GNSS) radio occultation (RO) technique provides accurate and precise measurements in the troposphere and stratosphere regions with global coverage and long-term stability. The active limb-sounding technique records GNSS signals, which are refracted while propagating through the atmosphere, on receiver satellites in Low Earth Orbit (LEO). The new Reference Occultation Processing System (rOPS) developed at the WEGC aims to process raw RO measurements into key climate variables such as temperature, pressure, and tropospheric water vapor with an integrated uncertainty propagation. In a first step of the processing, atmospheric excess phases, which denote the integrated refractivity proportional to pressure/temperature and water vapor, are derived from the raw RO measurements and the GNSS and LEO satellite orbit data.

In preparation for the excess phase processing the position, velocity, and clocks of the GNSS transmitter and the LEO receiver satellites need to be determined with high accuracy. For this purpose we use two independent orbit determination software packages Bernese (v5.2) and Napeos (v3.3.1) to perform a mutual consistency check and include estimates of systematic uncertainty bounds and propagated random uncertainties. Furthermore, the obtained orbit and clock products are compared with orbits and clocks from the EUMETSAT and UCAR processing centers.

Resulting monthly statistics for the MetOp, GRACE, and CHAMP RO missions show orbit uncertainty estimates of about 5 cm in position and 0.05 mm/s in velocity. For the COSMIC mission less favorable attitude behavior and restrictions in processing observations from two antennas cause somewhat decreased accuracy estimates near 20 cm in position and 0.2 mm/s in velocity.

Based on the rigorous evaluation of the quality of the calculated orbits and clocks we use two differencing approaches to eliminate the transmitter and receiver clock biases in the excess phase processing: 1.) zero-differencing for satellite missions with a sufficiently stable clock (MetOp and GRACE); 2.) single-differencing (COSMIC and CHAMP) using additional observations from a non-occulting reference GPS satellite.

Monthly intercomparison results between the obtained excess phase profiles and profiles provided by EUMETSAT and UCAR show a high consistency, with differences of a few millimeters in the tenuous atmosphere region above about 50 km altitude (in the mesosphere) and differences at centimeter level in the dense atmosphere region below about 12 km altitude (in the troposphere). In the latter region, excess phase magnitudes are many 10 to 100 meters already. These results align with the expected differences and estimated uncertainties, and signal an encouraging quality for broad application of the data in atmospheric and climate science.