Creepmeters on the San Andreas Fault System (click on the map for links to pre-1 Jan 2007 data, click here for data to 1 Jan 2008).
How does a creep-meter operate?
Access to calibrated creep-meter data 1994- 2006,
Access to calibrated creep data 2006-2008 UPDATED JANUARY 2008
Access to online telemetered data (user geo password hobo)
This week on the Hayward Fault and maps showing creepmeter monumentation.
Thirty page report on creep data 1993-2007 (pdf Final NEHRP Report 2003-7)
Nyland Ranch 2004-2006 (near San Jaun Bautista) Nail array data 1967-2002 &1999 photo
Creep-meters in the Coachella Valley, Southern California
Superstition Hills 2004-2006 Coachella Valley
Application note - Technical description of the low-power creep-meter ( pdf )
Application note - 1 sample/second, 3-m-range, 0.7 mm resolution, low power rupture meter (download pdf)
Thirty-four creepmeters operate in California, most of them operated by the USGS. Thirteen are maintained by the University of Colorado. Although most faults slip in earthquakes, a few slip continuously, or episodically too slowly to emit seismic radiation, several faults in California creep near the Earth's surface: the Hayward Fault (5-9 mm/yr), the Calaveras fault and the Concord Fault (0.5-5 mm/year), the San Andreas fault in central-California (0-30 mm/year), and in the Coachella Valley (0-3 mm/year), the Imperial fault (0-10 mm/year), and the Supersition Hills fault (<1.5 mm/year)
Creep is important in that it releases slip on a fault that would normally be available to drive a future earthquake. The region of the San Andreas fault between San Juan Bautista and Parkfield creeps sufficiently fast and to sufficient depth for it to have no significant likleyhood of a large earthquake. Usually, however, surface creep extends to depths of the order of 1-5 km. below which depth the fault is locked to the base of the brittle crust. This depth in California is of the order of 12 km. Hence creep typically releases much less than 1/3 of the slip that will drive future earthquakes.
Creep in California cracks road surfaces and walls, and offsets buried pipes and cables. In some places creep passes through structures requiring their continual repair.
The velocity of creep on a fault is proportional to the depth to which is locked and proportional to the shear stress applied to the fault. Thus if the depth of creep and the frictional properties of the fault do not change, the creep velocity provides a measure of the shear stress applied to the fault.
The friction on the creeping part of a fault is proportional to the stress applied at normal to the fault. Thus a slowing in creep rate could be caused either by a reduction in applied plate-boundary shear-stress, or by an increase in fault-normal stress. It could also be caused by a change in the friction of the fault surface caused, for example, by a change in water content.
Because surface creep is sensitive to the forces applied to a fault zone many of the faults in California are monitored by creepmeters. These instruments record or transmit information on relative movements of the edges of surface faults every 1-10 minutes (to an accuracy of 1-10 µm). Changes of creep rate have been seen to follow nearby earthquakes.
A pipe broke on the San Andreas near Parkfield the day before the 1966 Parkfield earthquake. For many years this fueled the notion that accelerated creep on the fault may have preceeded the earthquake, thereby offereing the possibility of short-term prediction of future earthquakes. The 2004 repeat of the Parkfield earthquake has dealt this a fatal blow. No creep occurred at Parkfield before the 2004 earthquake, and negligible strain at depth. Ironically, though, a pipe did indeed break before the 2004 Parkfield earthquake - but 135 km to the north at Nyland Ranch following slow creep measured in the preceding years by a local creep-meter.