Error Assessments in the GNSS Radio Occultation Excess Phase/Bending Angle Calculation

Global Navigation Satellite System (GNSS) Radio Occultation (RO) is an active remote sensing technique that provides valuable information on the vertical variations of electron density in the ionosphere and temperature, pressure, and water vapor in the stratosphere and troposphere. The bending angle profiles derived from RO signals have been used in operational numerical weather prediction (NWP) system through data assimilation to correct temperature bias in the forecasts. Conversion from the raw GNSS measurements (phase and time delay) to bending angle to higher level temperature can introduce different errors. These errors can be from the Precise Orbit Determination (POD) procedure, the clock error correction algorithm, the cycle slip detection algorithm, the wave/geometric optical transformation from time/excess phase space to bending angle/ impact height space, and the temperature retrieval algorithm etc. Different mission processing centers use different algorithms (software) to do the conversion. And the operational POD can only be obtained using ultra rapid GNSS orbital products which are usually less accurate but with a few hours latency than the final orbit with two weeks latency. In this study, we will go through a basic processing procedure from raw GNSS observations (the low rate positioning and the high rate occultation observations) to high level bending angle calculation. We use the Bernese software for experiments with publically available dataset, such as KOMPSAT5 and COSMIC, to derive the LEO and GNSS POD orbits and velocities. The excess phase is then modeled based on the POD results and a single differencing method to remove the receiver clock errors. The ROPP package is then used to derive the bending angle. We compare the results with those processed from UCAR CDAAC (COSMIC Data Analysis and Archive Center) and discuss/assess causes of uncertainties for the calculation of excess phase/bending angle. The retrieval uncertainties caused by cycle slip detections, different types of GNSS satellites, coordinate systems conversion (ECEF/ECIs) and different observation stages (i.e., openloop/close loop phase lock, rising/setting and geolocation parameters) will be carefully examined. This study of the RO data conversion procedure supports the Calibration/Validation work on recently launched COSMIC-2 mission and Commercial Weather Data Pilot program for NOAA/STAR.