Estimating empirical station timing biases using IGS clock products

We demonstrate a method to determine empirical station timing biases for geodetic receivers located at timing laboratories. Mutual observations of GPS time, as reported by the Bureau International des Poids et Mesures (BIPM) in its circular T and as observed by the International GPS service (IGS), are used together with the Circular T offsets of the lab timescales (UTC-UTC/sub i/). This method effectively extends the previously determined calibration bias of a lab's main time transfer system differentially to a colocated geodetic receiver system, without need of any additional local measurements. The empirical biases include both the internal delays within the geodetic GPS receiver/antenna hardware as well as the intra-laboratory offset to the source of UTC/sub i/. When applied to the subset of IGS stations located at timing labs, we find that useful results can be obtained provided that the geodetic system uses a frequency standard closely related to the local UTC/sub i/ realization and that the GPS receiver does not reset its internal clock too frequently. We find daily RMS repeatabilities in some cases of about 1 to 2 ns over periods up to about a year. Given sufficient available data, the mean station bias can be estimated from a single monthly circular T update to the 1 ns level, which is smaller than the uncertainty in current absolute calibrations. This method can be used to validate laboratory calibration measurements for the geodetic systems and to continuously monitor intra-laboratory calibration stability without disrupting normal operations, It can also be exploited to improve the near real-time steering of the IGS timescale to track UTC more closely.