GPS metrology: Bringing traceable scale to a local crustal deformation GPS network

Abstract

A constant scale difference between GPS solutions and traceable electronic distance measurement (EDM) results was found during semi-annually repeated campaigns performed in Olkiluoto, Finland. Since EDM results are very accurate and uncertainties are well-defined, this leads to an assumption that the GPS solution is biased. At the Kyviškės test field in Lithuania, the true lengths with traceable uncertainties between observation pillars were measured using a Kern ME5000 Mekometer as a scale transfer standard. GPS observations were processed using individual and type calibrated antenna tables, a local and global ionosphere model, and three different cut-off elevation angles, and several linear combinations and were then compared with the EDM results. The results show that the ambiguity resolution strategy and antenna calibration model play a significant role compared to the cut-off elevation angle and ionosphere model. Individual antenna calibration is required for the best metrological accuracy by means of the best agreement with traceable EDM results. The best metrological agreement was obtained with an L1 solution and individually calibrated antennas. The rms and maximum difference to the true (EDM) values were 0.3 and 0.7 mm, respectively. However, a clear distance dependency of 0.5 ppm was also evident. In particular, linear combinations with type calibrated tables caused variations up to 4 mm from the true value, even when high quality choke ring antennas were used. With individually calibrated antennas, all solutions were within ±1 mm of the true value.