By Mark Zoback, Stephen Hickman, William Ellsworth, and the safod science Team doi: 10. 2204/iodp sd. 11. 02. 2011

Scientific Drilling, No. 11, March 2011  

17

Science Reports

most of the two actively deforming fault traces identified in 

the SAFOD crossing. The microearthquake locations shown 

in Fig. 2 were determined utilizing subsurface recordings of 

these earthquakes from various geophone deployments in 

the SAFOD borehole along with surface recordings from the 

dense Parkfield Area Seismic Observatory (PASO; Thurber 

et al., 2004). This said, although the accuracy of location of 

HI is good (being determined by a seismometer deployed in 

SAFOD directly above the events), the location of SF and LA 

with respect to HI is relatively uncertain.

SAFOD Pilot Hole

In preparation for SAFOD, a 2.2-km-deep, near-vertical 

Pilot Hole was drilled and instrumented at the SAFOD site in 

the summer of 2002. The Pilot Hole was rotary drilled with a 

22.2-cm bit, and cased with 17.8-cm outside diameter (OD) 

steel casing. The Pilot Hole is currently open to a depth of 

1.1 km (explained below) and available for instrument 

testing, cross borehole experiments, and other scientific 

studies. Hickman et al. (2004) present an overview of the 

Pilot Hole experiment. 

There were a number of important technical, operational, 

and scientific findings in the Pilot Hole. These include geo-

logic confirmation of the depth at which the Salinian gran-

ites and granodiorites would be encountered (Fig. 3), and 

calibration of geophysical models with direct measurements 

of seismic velocities (Boness and Zoback, 2004; Thurber et 

al., 2004), resistivity (Unsworth and Bedrosian, 2004), den-

sity, and magnetic susceptibility (McPhee et al., 2004). In 

tural complexities in the near surface. 

These studies included an extensive 

microearthquake survey, high-resolution 

seismic reflection/refraction profiling, 

magnetotelluric profiling, ground and 

closely-spaced aeromagnetic surveys, 

gravity surveys, and geologic mapping. 

The repeating microearthquakes pro-

vide targets on the fault plane at depth to 

guide the drilling trajectory (Fig. 2A) into 

the microearthquake zone at less than 

3 km depth. Another reason for choosing 

this site is that there are three sets of 

repeating M~2 earthquakes in the target 

area. Surrounding these patches, fault 

slip occurs through aseismic creep. In a 

view normal to the plane of the San 

Andreas Fault Zone at 2.65 km depth 

(Fig. 2B), we see the source zones asso-

ciated with these three patches (scaled for 

a ~10-MPa stress drop). The seismograms 

from each of these source zones are 

essentially identical (Nadeau et al., 2004), 

and cross-correlation demonstrates that 

within ±10 m uncertainty these events are 

located in exactly the same place on the faults (F. Waldhauser, 

pers. comm.). 

As shown in Fig. 2B, we refer to the shallower source zone 

in the direction of San Francisco as the SF events, and the 

adjacent source zone in the direction of Los Angeles as LA 

events. Note that the SF and LA patches are adjacent to each 

other; it is common for LA events to occur immediately after 

SF events as triggered events. As seen in Fig. 2B, the third 

cluster of events (in green) occurs on a fault plane to the 

southwest of that upon which the SF and LA events occur.  

As this cluster of events is to the southwest of the other two 

clusters, these are referred to as the Hawaii (HI) events. 

The time sequences of the three clusters of repeating 

earthquakes are shown in Fig. 2C. Note that prior to the M6 

Parkfield earthquake of September 2004, each of the three 

clusters produced an event every ~2.5–3.0 years. Following 

the Parkfield earthquake, the frequency of the events 

increased dramatically, apparently due to accelerated creep 

on this part of the fault resulting from stress transfer from 

the M~6 main-shock. Following this flurry of events the fre-

quency of the repeaters slowed down and is presently in the 

process of returning to the background rate exhibited prior 

to the main shock. Similar behavior has been seen elsewhere 

along the San Andreas Fault system in California (Schaff 

et al., 1998).

Note in Fig. 2B that the HI events occur about 100 m below 

the fault intersection at 3192 m (measured depth), indicating 

that the HI microearthquakes occur on the southwestern-

Figure 3.

 Simplified geologic cross-section parallel to the trajectory of the San Andreas Fault 

Observatory at Depth (SAFOD) borehole. The geologic units are constrained by surface 

mapping and the rock units encountered along both the main borehole and the pilot hole. The 

black circles represent repeating microearthquakes. The three notable fault traces associated 

with the San Andreas Fault damage zone (SDZ, CDZ, and NBF) are shown in red. The depth 

at which the SAFOD observatory is deployed is shown.

Salinian Granite and

Granodiorite

     

Tertiary

Etchegoin 

Cretaceous

Great Valley

Formation

  

 

     

Cretaceous

Franciscan

Formation

  

   

SAFOD

    

Tertiary

(Undiff.)

Arkosic Sandstone

and Conglomerate

?

?

?

?

?

End of Phase 1

Pilot Hole

SW

NE

QTp

QTp

Tertiary Santa Margarita

Qa

San Andreas 

Fault

3,000

2,000

1,000

0

Depth (m)

SAFOD Observatory

SEA

LEVEL

Gold Hill Fault

Buzzard Canyon Fault

?

?

?

?

Arkosic Sandstone

and Conglomerate 

SDZ

CDZ

NBF

End

Phase 1

End

Phase 2

Tertiary

Etchegoin 

---