An Updated Catalog of Low-Frequency Earthquakes Along the San Andreas Fault Near Parkfield, California

SUMMARY

This Data Release contains an updated version of the San Andreas catalog of low-frequency earthquakes (LFEs) originally published here:
 
Shelly, D. R. (2017), A 15 year catalog of more than 1 million low-frequency earthquakes: Tracking tremor and slip along the deep San Andreas Fault, J. Geophys. Res. Solid Earth, 122, 3739–3753, doi:10.1002/2017JB014047.
 
This updated catalog contains 88 LFE families, with each family consisting of events detected by cross-correlation with the associated waveform template.  These templates were originally identified and located by Shelly and Hardebeck (2010):
 
Shelly, D. R., and J. L. Hardebeck (2010), Precise tremor source locations and amplitude variations along the lower-crustal central San Andreas Fault, Geophys. Res. Lett., 37, L14301, doi:10.1029/2010GL043672.
 
Data for event detection comes from the High Resolution Seismic Network (HRSN, 2014), operated by UC Berkeley Seismological Laboratory:
 
HRSN (2014), High Resolution Seismic Network. UC Berkeley Seismological Laboratory. Dataset. doi:10.7932/HRSN.
 
Data were retrieved from the Northern California Earthquake Data Center (NCEDC, 2014):
 
NCEDC (2014), Northern California Earthquake Data Center. UC Berkeley Seismological Laboratory. Dataset. doi:10.7932/NCEDC.
 
 
DOCUMENTATION
 
For completeness, the documentation below repeats the documentation provided in the supplement of Shelly (2017) below, with minor modifications:
 
Catalog Time-Period and Format:
The low-frequency earthquake catalog spans from 1 April 2001 (first detection 6 April 2001) to 30 April 2024 and contains 1,528,117 events.
Format: 
 
year, month, day, s_of_day, hr, min, sec, ccsum, meancc, med_cc, seqday, ID, latitude, longitude, depth, n_chan
 
Explanations:
year,month,day  - Event time (template start time in UTC - ~1s prior to first S-wave arrival time at an HRSN station)
s_of_day   - Event time (template start time in UTC - ~1s prior to first S-wave arrival time at an HRSN station), second of the day (i.e. 0-86400), provided for consistency with earlier versions
 hour, minute, second - Event time (template start time in UTC- ~1s prior to S-wave arrival time at first HRSN station)
ccsum   - correlation sum across all stations (must exceed 4.0)
meancc  - mean correlation among stations with data
med_cc  - median correlation
seqday - sequential day since 1 March 2001, provided for consistency with earlier versions
ID     - reference ID of family
latitude(deg) longitude(deg)  depth(km)  - estimated location for that family (Shelly and Hardebeck, 2010)
n_chan - number of data channels existing for event (some channels may exist, but not have good data)
 
 
 
Family IDs:
Each family has an associated identification code, which is a number followed by 1-4 ‘s’.  The family IDs are almost meaningless and are simply used as unique identifiers.  Originally the numeric code was taken from the second of the day at which the initial template for this family occurred.  The number of ‘s’ indicates the number of iterations of stacking and cross-correlation that were applied to derive the template waveforms (see Methods).  The lower amplitude and more distant sources typically benefitted from multiple iterations of stacking and cross correlation, before the final template stabilized in its detection set.
 
Data channels used (station.channels):
GHIB.13, EADB.123, JCSB.1, FROB.123, JCNB.123, VCAB.123, MMNB.123, CCRB.123, LCCB.123, SMNB.123, RMNB.123, SCYB.123
 
JCNB failed in 2008 and was replaced by a shallow sensor.  New sensor not used.
 
RMNB failed in 2011 and was not replaced.
 
GHIB.2 was never operational
 
JCSB.23 have poor signal to noise and are not used.
 
VARB was replaced with a new sensor at a new depth in 2003, and this station was not used in original template formation.
 
As of 2024, detection capabilities were substantially degraded with a maximum of 16 channels of data available for detection.  This is due to outages in GHIB (since 2020), FROB (since 2023), VCAB (since 2023), and CCRB (since 2022), in addition to the outages described above. It is unclear when/if these stations might be repaired in the future.
 
Channel swap on FROB, VCAB (after BP channel designations changed to SP), before 2011-7-14):
 
2011/4/21-2011/7/14:  Swap VCAB.2 and VCAB.3
 
2010/11/10-2011/7/14:  Swap FROB.2 and FROB.3
 
 
Disregard mean correlation, enforce network correlation sum only (because of poor but present data):  2012/2/13-2014/4/23
 
 
Polarity corrections during initial processing:
 
2001/6/1 to 2001/12/13: reverse CCRB.123.
 
2010/12/10-2011/4/7: reverse FROB.123
 
2010/12/10-2011/4/7: reverse MMNB.123
 
2010/4/8 to 2011-7-14: reverse FROB.23
 
 
Polarity and other corrections applied in post-processing (these are minor and done after initial detection stage):
 
2011/4/7- 2011/5/27:  do not use FROB.2 channel (wiring mistake, FROB.2 duplicates FROB.3)
 
2005/4/11-2005/5/13: reverse GHIB.13
 
2005/12/15-end: reverse GHIB.3
 
2002/11/22-2003/1/16: reverse EADB.2
 
2002/11/21-2003/1/17: reverse VCAB.3