HR: 08:45h
AN: G11A-04 INVITED [Abstracts]
TI: IGS Clock Products for Accurate Geodetic and Timing Applications
AU: * Senior, K L
EM: ken.senior@nrl.navy.mil
AF: Naval Research Laboratory, 4555 Overlook Ave SW, Washington, DC 20375, United States
AU: Ray, J R
EM: jimr@ngs.noaa.gov
AF: National Geodetic Survey, 1315 East-West Highway, Silver Spring, MD 20910, United
States
AB:
The performance of any GNSS is intimately related to the characteristics of the satellite clocks, so an
understanding of the clock behavior is vital. The accurate products of the IGS enable daily point positions to the
sub-cm level and continuous global clock comparisons to the sub-ns level. Time transfers are less accurate
than associated positioning because of: 1) difficult-to-measure hardware delays; 2) the limiting pseudorange
measurement errors. Both factors arise from characteristics of the pseudorange signals, which are easily
degraded by multipath and other effects.
The behavior of the satellite clocks are also be important. Over sub-daily intervals, IGS products show that
approximate power-law stochastic processes govern all GPS clocks. The Block IIA Rb and Cs clocks obey
random walk noise, with the Rb clocks up to nearly an order of magnitude more stable. Due to the high-frequency
noise of the onboard Time Keeping system in the newer Block IIR and IIR-M satellites, their Rb clocks are
dominantly white noise up to a few 1000 s with standard deviations of 90 to 180 ps. Superposed on this random
background, periodic signals are present at four harmonic frequencies, n × (2.0029 ± 0.0005) cycles
per day for n = 1, 2, 3, and 4. The equivalent fundamental period is 11.9826 hours, which surprisingly
differs from the reported mean GPS orbital period of 11.9659 hours by 60 ± 11 s. We cannot account for
this apparent discrepancy but note that a clear relationship between the periodic signals and the orbital dynamics
is evidenced for some satellites by modulations of the spectral amplitudes with eclipse season. The Cs clocks
are more strongly affected than the Rb clocks. All four harmonics are much smaller for the IIR/IIR-M satellites
than for the older blocks. The strong 12- and 6-hour periodics in most GPS clocks dictate that these variations
should be modeled in all high-accuracy applications, such as for timescale formation, interpolation
of IGS clock products for higher-rate GPS positioning, and predictions for real-time uses. Modeling will also
overcome the effect at 6 hours due to the conventional neglect of the Earth's oblateness in time transformations
and is more effective than modifying the relativistic modeling since other causes of 6-hour periodics are usually
greater.
After accounting for the 12- and 6-hour harmonics, and the usual quadratic variations, it is possible to interpolate
or extrapolate the residual stochastic clock time series with minimal degradation for the IIA Rb satellites up to
about 30 s or up to 5 minutes within the accuracy as the IGS products. The white noise of the IIR/IIR-M Rb clocks
limits their interpolation and prediction to about 135 ps over intervals shorter than 1000 s. The high level of
random walk variance for the IIA Cs clocks immediately overwhelms the IGS measurement noise for any interval.
In any future improved GNSS clock systems, it would be beneficial to reduce the white noise floor of the Time
Keeping System by at least a factor of five. For short-term performance comparable to the level expected for the
Galileo H-masers, further reduction of the GPS clock instabilities would be necessary. If methods can be found to
attenuate the GPS harmonics at the satellites, that would be useful, otherwise modeling of these effects is
probably effective and does not require any understanding of the driving mechanisms.
DE: 1241 Satellite geodesy: technical issues (6994, 7969)
SC: Geodesy [G]
MN: 2007 Fall Meeting