From jray at bipm.org Tue Jun 15 05:08:06 2004 From: jray at bipm.org (Jim Ray) Date: Tue, 15 Jun 2004 14:08:06 +0200 Subject: [IGSMAIL-4953]: update of day-boundary clock jump analysis Message-ID: <006701c452d1$6a885aa0$e57f68c1@TAI20> ****************************************************************************** IGS Electronic Mail 15 Jun 08:41:38 PDT 2004 Message Number 4953 ****************************************************************************** Author: Jim Ray The discontinuity in estimated station clock values at the boundaries between 1-day arcs is a sensitive measure of the average pseudorange (code) accuracy averaged over 24-hr periods. Ken Senior and I previously described how this can be used to assess station performance and we reported results for 30 IGS stations over the ~460 days after Oct 2000. Our paper in Metrologia (40(3), S270-S288, 2003) is available electronically at: http://www.bipm.org/static/gpst/refs/raysenior-metrologia.pdf I have updated the day-boundary clock jump analysis using data available through the end of last week, up to ~1310 days. The RMS of the day jumps for each station equipped with an H-maser clock is given in the table below. (Other types of clocks are not sufficiently stable for this test.) Unfortunately, several stations no longer use masers (and log files are incorrect). The editing and processing criteria are the same as our earlier analysis. The maximum data gap at the day boundary is 30 min (typically 5 min), so the effect of any instability in the H-maser standard itself should be negligible. Since the RMS statistics are for the differences between pairs of independent days, each daily accuracy estimate should be smaller by the factor 1/sqrt(2). Of particular note is the very large dispersion in RMS performance, about one order of magnitude, which reflects the wide range in code performance among IGS stations. This in turn shows the vast variation in multipath environments, both external to the antenna and internal to the GPS instrumentation. In some cases the performance has varied markedly with time, sometimes for known reasons that can be correlated with changes reported in the site logs. Plots and data files are available upon request. RMS clk Site jump/ps Remarks ---- ------- ---------------------------------------------- ONSA 149.49 excellent BREW 151.89 excellent OPMT 158.39 new station, so very limited data BRUS 164.82 after changes in summer 2003 improved to 118 ps! MAD2 170.19 very limited data, so RMS is not reliable WTZR 189.13 GODE 204.59 USN1 225.03 WSRT 226.73 slight degradation since summer 2003 KHAJ 232.90 limited data CRO1 235.90 maser no longer used USUD 265.84 maser no longer used NPLD 267.68 TID* 269.00 appears improved since summer 2003 YEBE 270.70 GOL2 271.23 almost no data, so RMS is not reliable AMC2 282.58 improved after ant/rx changes in June 2002 SPT0 286.38 WES2 296.02 PIE1 305.25 improved since Oct 2002 (rx change) STJO 334.01 USNO 353.97 appears worse since spring 2003 IRKT 359.15 NYAL 363.11 much better than NYA1 recently NLIB 367.83 MATE 388.62 significant time variations; recently better KOKB 460.19 large degradation before recent ant/cable changes FAIR 477.94 somewhat improves since summer 2003 DRAO 521.78 YELL 563.76 large seasonal variations, much worse in winters ALBH 587.13 after Sept 2002 greatly improved to 97 ps! HOB2 631.46 variations correlated with station changes MEDI 703.24 maser no longer used ??? FORT 705.50 NYA1 749.95 large degradation since summer 2003 ALGO 877.47 large seasonal variations, much worse in winters NRC1 935.94 large seasonal variations, much worse in winters METS 1064.80 maser no longer used