Questions about the data quality graphs
...as displayed on the "site pages" reachable from the station list or clickable map.
|Where do the numbers plotted come from?|
The output of the teqc program from UNAVCO, given the input of daily RINEX files. Specifically, they are figures from the SUM line discussed in teqc's qc modedocumentation: number of complete observations, RMS MP1 (L1 Multipath) and MP2 (L2 Multipath), and observations per slip (inverted and multiplied by 1000).
teqc qc output is available in the s/ subdirectories of the GPS data area of the CDDIS archive. The teqc software can easily be ran on any file which is not stored at CDDIS.
|What are the error bars in the 45-day average plots?|
Standard deviations of the previous 45 days of data as plotted in the "Recent data quality plots."
|What are "good" numbers?|
A site recording data at 30 second intervals all day will result in 20000 or more observations. Numbers over 40000 almost surely indicate a site running at a sampling rate higher than every 30 seconds.
Cycle slips X1000/observations is less than 5 for more than half of IGS stations, and less than 10 for more than 2/3 of IGS stations.
Half of IGS stations have RMS MP1 under 0.4m, and 2/3 have less than 0.5m. Absolute MP RMS values, however, do not necessarily correlate with site performance. RMS MP2 is less than 0.6m for half of IGS stations and less than 0.75m for 2/3 of IGS stations.
Possibly more productive than aiming for a particular absolute value of any of the numbers is looking for higher standard deviations than are common in the IGS (why is the site behaving erratically day-to-day?), or sudden departures from the site's usual behavior (what happened?).
|What does "Most recent value deviates from 45-day mean" indicate?|
This message is printed above a plot if the most recent point is more than a threshold number of standard deviations away from the site's 45-day mean. For mp1 and mp2 the threshold is currently 2 standard deviations; for number of observations and cycle slips, the threshold is 3 standard deviations.
This does not necessarily indicate trouble with a site, but rather is a flag inviting a human to take a close look and decide if more investigation is warranted.
|What do vertical lines with either rec, ant, ecc, or ? printed above indicate?|
Vertical lines are plotted onto a graph whenever a site change in receiver (rec), antenna (ant), or eccentricity (ecc) is documented in the IGS SINEX template file, which is generated daily from the site logs on file at the Central Bureau. A vertical line with a '?' indicates a possible change in the site's behavior was detected at this point.