摘要
Accurate estimation of clocks, for example for the Gravity Recovery And Climate Experiment(GRACE)twin-satellites, is a critical part of precise orbit determination(POD) that ensures temporal gravity inversion. Characterizing the periodic variations of the receiver clocks is critical for precise clock modeling and prediction. In this study, the receiver clock is estimated using two different POD procedure: kinematic and reduced-dynamic approaches. Choices and the number of orbital parameters estimated in POD process affect the clock estimates, e.g., there are 8895 and 34,560 total parameters in the reduced-dynamic and kinematic approaches, respectively. In the both cases, the periodic variations of GRACE receiver clock are mainly dominated by the GPS orbit period, as well as once-(1-pr) and twiceper-revolution(2-pr) effects. Here the 1-pr effect is coupled with the relativistic effect, resulting in a difficulty to separate both signals. The clock amplitudes caused by the GPS orbit period, 1-pr and 2-pr are about 0.1, 0.03 and 0.01 ns, respectively. The GPS orbit period is almost one order magnitude larger than the 1-and 2-pr effect. The 0.1-ns amplitude of the 12-h periodic variation is equivalent to a 3-cm error in range. Such a systematic error should be considered in the receiver clock modeling for both the improvement of positioning accuracy and the reduction of number of unknown parameters, if the precise point positioning(PPP) technique is used for the orbit determination of the GRACE.
Accurate estimation of clocks, for example for the Gravity Recovery And Climate Experiment(GRACE)twin-satellites, is a critical part of precise orbit determination(POD) that ensures temporal gravity inversion. Characterizing the periodic variations of the receiver clocks is critical for precise clock modeling and prediction. In this study, the receiver clock is estimated using two different POD procedure: kinematic and reduced-dynamic approaches. Choices and the number of orbital parameters estimated in POD process affect the clock estimates, e.g., there are 8895 and 34,560 total parameters in the reduced-dynamic and kinematic approaches, respectively. In the both cases, the periodic variations of GRACE receiver clock are mainly dominated by the GPS orbit period, as well as once-(1-pr) and twiceper-revolution(2-pr) effects. Here the 1-pr effect is coupled with the relativistic effect, resulting in a difficulty to separate both signals. The clock amplitudes caused by the GPS orbit period, 1-pr and 2-pr are about 0.1, 0.03 and 0.01 ns, respectively. The GPS orbit period is almost one order magnitude larger than the 1-and 2-pr effect. The 0.1-ns amplitude of the 12-h periodic variation is equivalent to a 3-cm error in range. Such a systematic error should be considered in the receiver clock modeling for both the improvement of positioning accuracy and the reduction of number of unknown parameters, if the precise point positioning(PPP) technique is used for the orbit determination of the GRACE.
基金
funded by the Ministry of Science and Technology of Taiwan, under grant numbers 103-2116-M-008-014
