SAFOD—The San Andreas Fault Observatory at Depth and Its Relevance to Oil and Gas
by Bill Rizer
Introduction
A long-standing dream of many geoscientists was realized on August 2, 2005, when a drill hole at the SAFOD site near the town of Parkfield, California, penetrated a seismically active segment of the San Andreas Fault (the Fault) at a depth of approximately 2 miles (Figure 1).
The ambitious Earthscope project (van der Vink et al, 2005) is a major national research effort designed to further understanding of the properties, the structure, and the forces and deformation processes operative in the crust of North America. A part of Earthscope, the San Andreas Fault Observatory at Depth (SAFOD) is itself a major research effort of the USGS and the State of California that is funded by the National Science Foundation (NSF). The primary objective of SAFOD is to “study the physical and chemical processes that control deformation and earthquake generation within an active plate-bounding fault zone” (Zoback et al, 1998). SAFOD will establish an observatory within a segment of the Fault to study the basic mechanical, fluid, and seismic properties and processes operative along the active San Andreas system. Principal investigators for the SAFOD project are Mark Zoback of Stanford University and Steve Hickman and Bill Ellsworth of the USGS at Menlo Park, California.
Figure 1.
The location of the SAFOD site was chosen near Parkfield for a number of reasons.
• The area is accessible.
• It is just north of the section of the Fault that slipped in the M 6.0 Parkfield earthquake in 1966 (Figure 1).
• Parkfield was already the site of a major research effort by the USGS in earthquake prediction and, therefore, was very well documented geologically and geophysically.
In this region, the Fault was slipping through a combination of small-to-moderate magnitude earthquakes and aseismic creep (Hickman et al, 2004). The Fault at the surface was creeping at about 2 cm/year, with most of the displacement occurring in a zone that was at most only 10 m wide. Numerous microearthquakes (less than M 2.0) had been detected along the Fault near SAFOD at depths of 2.5 to 12 km. This area had been the focus of repeated magnitude M ~6.0 earthquakes over the past 150 years—in 1857, 1881, 1901, 1922, 1934 and 1966 (Bakun and McEvilly, 1979). The first, in 1857, was a foreshock to the great Fort Tejon (M 7.9) earthquake that ruptured the Fault from Parkfield to the southeast for over 180 miles. When drilling started in 2002, another M 6.0 earthquake was overdue. The idea was to locate SAFOD at a position along the Fault just northwest of the segmented expected to rupture next.
The Pilot Hole
Prior to drilling the main borehole, a 2.2-km-deep vertical pilot hole was drilled about 2 km southwest of the surface trace of the Fault (Figure 2). Drilling of the pilot was funded by the International Continental Drilling Program (ICDP), with NSF and USGS support (Hickman et al, 2004). The location was chosen to be close enough to the Fault to help
identify the most likely area of slip on the Fault and to guide the primary borehole to intersect that area. The pilot hole was logged for fractures, stress and temperature, and packer tests were run for stress, permeability and fluid sampling. A 40-level multicomponent seismic array was installed in the casing for monitoring microseisms and for serving as a part of 2-D and 3-D seismic surveys run to better define the structure of the site. The hole was instrumented for long-term monitoring of pore-pressure, strain, temperature and seismic activity. The well was completed in the summer of 2002.
Geologic data, microseismic monitoring and geophysical imaging from sensors in the pilot hole and on the surface were used to locate and guide drilling of the primary observation well at a sufficient accuracy to allow for drilling and coring deviated holes through the fault zone. The plan for the lateral core and wellbores (Figure 2) called for starting “kickoff” at a vertical depth of about 2.5 to 3 km and continuing through the fault zone into the “intact” rock on the other side.
Figure 2.
The main SAFOD observation well was spudded only 10 m from the surface location of the pilot. Drilling of the main observation well began in June 2004 on the Pacific Plate about 2 km west of the surface trace of the Fault (Figure 1) and continued to the middle of October. Drilling resumed in June 2005 with the wellbore penetrating the Fault in August 2005. The SAFOD observatory will be completed in 2007.
Somewhat ironically, the anticipated M 6.0 Parkfield earthquake occurred in September 2004, before drilling of the main observation well was completed. The quake ruptured roughly the same segment of the Fault that had ruptured in 1966, as predicted.
Fault Strength and Stress
While most of the research at SAFOD is geared toward fundamental questions related to earthquake prediction, many of the results could have very real impact on the oil and gas industry. One area being addressed has been the subject of considerable debate for many years: the strength of faults and the level of shear stress acting on them. Some of the data and analyses generated by SAFOD have already had an effect on this debate. Data from the pilot hole have provided important new information on the state of stress in the crust immediately adjacent to the Fault. Interpretation of well tests, image logs, shear wave logs and cores seem to bolster earlier arguments (e.g., Zoback et al, 1987) that suggest the San Andreas Fault may be very “weak,” that is, it may have little frictional resistance to slip or, equivalently, can support only limited levels of shear stress.
The strength of faults like the San Andreas has been a contentious issue for quite some time. The debate involves what has been called the stress/heat flow paradox (Lachenbruch and Sass, 1988; Zoback et al., 1998). A “weak” fault is one whose strength is on the order of the stress relieved by an earthquake on that fault (typically < 20 MPa), while a “strong” San Andreas would have a substantially greater strength, on the order of 50-100 MPa (e.g., Lachenbruch and McGarr, 1990; Fletcher and Mariagiovanna, 1999; Scholz, 2000). According to Zoback et al (1998) arguments in favo