XSC1
Slack Canyon Creepmeter

On March 6, 1997, the transducer for this instrument
was replaced. The transducer had started to fail in
early 1997. In addition, critical micrometer measurements
were lost. The tie between the measurements prior
to March 6, 1997 is tentative; users should treat
the data gap as a potential offset.

Around Jan 23, 2007, the DCP and electronics amplifier were
changed. Noted that the old electronics, for some reason
"saturated" or 'hit the rail' when output of electronic exceeded
about 3.7 volts (or the micrometer exceeded 20 mm or so).
Inspection of micrometer log reveals that during operation of
this instrument, that has never happened. With new DCP
and electronics, telemetry voltage range now 0-5v; scale factor 
is now 119 counts(mv) per mm fault slip (creep).




 
          General Notes:
    FAULT CREEP DATA FROM PARKFIELD

The following data represent real-time measurments of fault slip across
the San Andreas fault or its "south-west" trace near Parkfield, CA.
Creepmeters consists of 2 piers separated by about 30 meters and
connected by an invar wire.  The main fault-trace lies between the 2
piers.  The wire is at roughly a 30 degree angle from the local trace
of the fault. A displacement tranducer measures the change in length of
the wire (or the change in distance between the piers). These
measurements are made once every 10 minutes and are telemetred to Menlo
Park. In addition, micrometer measurements are periodically made by
"hand". The telemetred creep data are reconciled with the micrometer
measurements. The data have been reconciled with the
micrometer measurements and then decimated to daily measurements and
scaled to represent fault slip in mm. 

The data provided here has been processed such that telemetry
glitches have been removed, data edited for electronic and other
problems, offsets due to resets of the instrument have been removed,
and the telemetried data have been reconciled with the Micrometer
measurements.

Please note that when computing secular rates and other statistics that
the assumption of data independence (ie gaussian distributed data) is
NOT correct. For starters, the statistics of the data are a mixture of
random-walk and white (gaussian) noise.  You'll need to construct the
appropriate data covariance matrix to account for these noise sources.
Otherwise, if you let your standard least squares routine compute the
secular rate and its standard error, that error bar will most likely be
too small relative to the true estimate of standard error.

Because the pier are installed at shallow depths (2-3 meters), these
piers will tilt in responce to rainfall. ONE NEEDS TO EXAMINE RAINFALL
RECORDS WHEN INTERPRETING THESE CREEP MEASUREMENTS. 

Many people have contributed to this project including:
Robert Burford
Sandy Schultz
Kate Breckinridge
Rich Liechti

For questions, contact John Langbein (langbein@usgs.gov)
  or Andy Snyder (asnyder@usgs.gov)