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

24  

Scientific Drilling, No. 11, March 2011

Science Reports

crustal volume. Their results refined earlier tomographic 

models for SAFOD to clearly image a deep low-velocity  

zone along the San Andreas Fault. This low-velocity zone 

supports the propagation of both P- and S-type fault zone 

guided waves. Observation of these waves on seismometers 

placed inside the fault zone places strong constraints on its 

geometry and continuity. Ellsworth and Malin (in press) 

determined that the low-velocity zone in which these waves 

propagate coin-cides with the zone of extensive rock damage 

seen in the downhole measurements (Fig. 4). The waveguide 

extends to the northwest and southeast of SAFOD for at least 

8 km. Wu et al. (2010) used the dispersion properties of the 

S-type guided waves recorded in the Main Hole to show that 

the low-velocity wave-guide extends downward to near the 

base of the seismogenic zone at 10–12 km depth.

The short hypocentral distances and high-Q environment 

of the SAFOD boreholes make it possible to study source 

parameters to smaller magnitude than with data from instru-

ments in shallow boreholes or on the surface. Only a small 

fraction (<1%) of the San Andreas Fault surface near SAFOD 

produces earthquakes, with the remainder of the fault 

moving through aseismic creep. The earthquakes that do 

occur are predominately located within clusters of repeating 

events. Static stress drops range from as low as 0.1 MPa to 

100 MPa (Imanishi and Ellsworth, 2006). The upper limit is 

comparable to the laboratory-derived frictional strength of 

the country rock from outside of the damage zone (Lockner 

et al., in press). McGarr and Fletcher (2010) determined the 

yield stress for a repeat of the SF target earthquake of 

64 MPa. These results suggest that the target events and 

other repeating earthquakes occur where the fault juxtapo-

ses normal crustal rocks patches embedded within an other-

wise weak, creeping fault. As a consequence, there is no con-

tradiction between such high stress drop events and an 

intrin-sically weak, creeping San Andreas Fault in a strong 

crust, as indicated by the in situ stress and heat-flow measu-

rements in the SAFOD Pilot Hole and Main Hole.

The twenty-seven experi-

mental deployments also 

guided the selection of sen-

sors for the observatory and 

revealed mechanical and 

environmental issues that 

dictated the design of the 

observatory. The ambient 

temperature of up to 120°C  

at the planned depth of the 

observatory controlled the 

choice of downhole electro-

nics and sensors. More 

seriously, the borehole fluid 

contains gases that penetrate 

past conventional O-rings 

and wireline insulation. 

Consequently, a design was 

To date, over 350 samples from the Phase 3 core have 

been distributed to investigators from around the world for 

laboratory analyses and testing; the latest results from these 

studies were discussed at two SAFOD special sessions of the 

2010 annual meeting of the American Geophysical Union. 

These include studies of the mineralogy and chemical evolu-

tion of the fault zone, the physical properties of fault zone 

materials, the frictional strength of fault and country rock 

under a wide variety of loading conditions, and the evolution 

of deformation mechanisms and fluid-rock interaction within 

the fault zone over time. Procedures for requesting samples 

or gaining access to the SAFOD thin-section collection are 

available online (Table 1). The GCR staff is responsible for 

maintaining records of core, cuttings and fluid sample 

requests filled; names of people to whom these samples were 

provided; and the final disposition of samples (date samples 

returned and condition of samples). The GCR staff is also 

responsible for entering data and results from SAFOD 

sample investigations into the EarthScope Data Portal, 

which is currently under construction (Table 1).

SAFOD Observatory

In preparation for the establishment of a geophysical 

observatory deep within the fault, a series of nineteen 

temporary deployments of seismometers, accelerometers, 

and tiltmeters in the Main Hole and an additional eight 

deployments in the Pilot Hole were conducted between 2002 

and 2008, leading up to the deployment of the SAFOD obser-

vatory in September 2008 (data available online, Table 1). 

Seismic data collected during the temporary deployments 

are yielding important new findings on the structure of the 

San Andreas Fault and properties of the earthquakes that  

it produces. By combining surface and borehole observa-

tions of surface explosions and local earthquakes with 

double-difference tomography, Zhang et al. (2009) deter-

mined a detailed Vp, Vs, and Vp/Vs model for the SAFOD 

H

ol

e 

G

 R

un

 2

 

Se

ct

io

n 

7

10487

10488

10489

Massive, Grayish-

Black Shale

Foliated Fault Gouge

Serpentinite block

 

TS

XRD

TS

XRD

TS

SEM

XRD

10490

10491

Note: Core section split in half

H

ol

e 

G

 R

un

 2

 S

ec

tio

n 8

XRD

Core section split in half

10492

Foliated Fault Gouge

H

ol

e 

G

 R

un

 2

 S

ec

tio

n 

9 

(c

or

e 

ca

tc

he

r)

Figure 6.

 Photographs of the section of core 2 that crosses the SDZ (see Figs. 4 and 5) as they appear in 

the Photographic Atlas of the SAFOD Phase 3 (Table 1). The colored lettering indicates where samples were 

used for TS, XRD, and SEM presented in the Phase 3 Core Atlas. Note that the center and bottom photos are 

of the core sections split in half. Measured depths (in the sidetrack) are shown in feet (1 ft=30.48 cm).

---