Continuous geodetic time-transfer analysis methods.

We address two issues that limit the quality of time and frequency transfer by carrier phase measurements from the Global Positioning System (GPS). The first issue is related to inconsistencies between code and phase observations. We describe and classify several types of events that can cause inconsistencies and observe that some of them are related to the internal clock of the GPS receiver. Strategies to detect and overcome time-code inconsistencies have been developed and implemented into the Bernese GPS Software package. For the moment, only inconsistencies larger than the 20 ns code measurement noise level can be detected automatically. The second issue is related to discontinuities at the day boundaries that stem from the processing of the data in daily batches. Two new methods are discussed: clock handover and ambiguity stacking. The two approaches are tested on data obtained from a network of stations, and the results are compared with an independent time-transfer method. Both methods improve the stability of the transfer for short averaging times, but there is no benefit for averaging times longer than 8 days. We show that continuous solutions are sufficiently robust against modeling and preprocessing errors to prevent the solution from accumulating a permanent bias.

Continuous time transfer using GPS carrier phase.

The Astronomical Institute of the University of Berne is hosting one of the Analysis Centers (AC) of the International GPS Service (IGS). A network of a few GPS stations in Europe and North America is routinely analyzed for time transfer purposes, using the carrier phase observations. This work is done in the framework of a joint project with the Swiss Federal Office of Metrology and Accreditation (METAS). The daily solutions are computed independently. The resulting time transfer series show jumps of up to 1 ns at the day boundaries. A method to concatenate the daily time transfer solutions to a continuous series was developed. A continuous time series is available for a time span of more than 4 mo. The results were compared with the time transfer results from other techniques such as two-way satellite time and frequency transfer. This concatenation improves the results obtained in a daily computing scheme because a continuous time series better reflects the characteristics of continuously working clocks.