Precision and accuracy of USNO GPS carrier phase time transfer: 2012 update

The United States Naval Observatory (USNO) produces GPS carrier-phase time-transfer (GPSCPTT) estimates for approximately 100 receiver clocks daily. All estimates are available with 16-h latency; a subset of approximately 34 are available every 6 h with 3-h latency plus 24 h of predictions. The once-per-day post-processed estimates are referred to as "rapids" (USR); the four-times-per/day estimates/predictions are referred to as "ultra-rapids" (USU). The ultra-rapids are suitable for real-time applications. We investigate the uncertainty of USR and the first 6 h of USU predictions by comparing them to four weeks of IGS Final Clock solutions along 4-6 timing links. No day-boundary discontinuities (DBDs) were removed. USR exhibited a 60-100 ps RMS difference and a few to tens of picoseconds time stability with respect to the IGS Finals, supporting an estimate of 125 ps USR standalone time stability. White FM characterized the USR-IGS difference for most averaging times observed, with USR-IGS frequency stabilities of 0.9-2.1·10^-15 at averaging time τ = 1 d. DBD removal would likely improve that value. USU clock predictions exhibited bi-modal performance with respect to IGS Finals, with links to site WES2 exhibiting noticeably worse behavior than links not utilizing it. USU-IGS differences were 139-220 and 470-584 ps RMS. USU-IGS time stabilities of 133 ps or better were observed for non-WES2 links, supporting an estimate of 153 ps time stability for USU clock predictions. Non-WES2 USU-IGS frequency stabilities exhibited 8·10^-14-8·10^-15 frequency stabilities for τ <; 6 h (the prediction length).