(Near-)Real-Time Orbit Determination for GNSS Radio Occultation Processing

Kurzfassung

The processing of GPS radio occultation measurements for use in numerical weather predictions requires a precise orbit determination of the host satellite in near-real-time. Making use of data from the GRAS instrument on Metop-A, the performance of different GPS ephemeris products and processing concepts for near-real time and real-time precise orbit determination is compared. While previous analyses have focused on the achievable along-track velocity accuracy, the present study contributes a systematic comparison of the resulting estimated bending angles. This enables a more rigorous trade-off of different orbit determination methodologies in relation to the end-user needs for atmospheric science products. It is demonstrated that near-real-time GPS orbit and clock products have reached a sufficient quality to determine the Metop-A along-track velocity with an accuracy of better than 0.05 mm/s that was formerly only accessible in post-processing. The resulting bending angles are shown to exhibit standard deviation and bias differences of less than 0.3% compared to post-processed products up to altitudes of at least 40 km, which is notably better than 1% accuracy typically assumed for numerical weather predictions in this height regime. Complementary to the analysis of ground-based processing schemes the potential of autonomous on-board orbit determination is investigated for the first time. Using actual GRAS flight data it is shown that a 0.5 m 3D rms position accuracy and a 0.2 mm/s along-track velocity accuracy can in fact be obtained in real-time with the currently available GPS broadcast ephemeris quality. Bending angles derived from the simulated real-time processing exhibit a minor performance degradation above tangent point heights of 40 km but negligible differences with respect to ground based products below this altitude. Onboard orbit determination and, if desired, bending angle computation, can thus enable a further simplification of the ground segment in future radio occultation missions and contribute to reduced product latencies for radio occultation data assimilation in numerical weather predictions.