U.S.

Geological Survey Award Number G09AP00077

Final Technical Report

Near-Surface Geophysical Imaging of Paleochannels for

the Determination of Slip Rate at the Green Valley Fault

Project Period: June 1st, 2009 - May 31st, 2010

Submitted September 9th, 2010

Mitchell S. Craig
Department of Earth and Environmental Sciences
California State University, East Bay
25800 Carlos Bee Blvd.
Hayward, CA 94542
Phone: 510-885-3425, Fax: 510-885-3486
mitchell.craig@csueastbay.edu

Key Words: near surface geophysics, seismic reflection, paleoseismology

U.S. Geological Survey Award Number G09AP00077

Final Technical Report
Near-Surface Geophysical Imaging of Paleochannels for
the Determination of Slip Rate at the Green Valley Fault
Mitchell S. Craig
California State University, East Bay
mitchell.craig@csueastbay.edu
Abstract
Seismic reflection surveying was conducted at the Green Valley fault at Mason Road with the
goal of identifying subsurface stream channels for use in measuring fault offset and slip rate.
Three 180-m long, fault-parallel profiles were recorded north of Mason Road in August 2009
near the site of previous paleoseismologic investigations, including near-surface geophysical
surveys. A 2007 seismic refraction survey south of Mason Road identified a trough-shaped, lowvelocity body on the east side of the fault at depths of 5-12 m, which was interpreted as a stream
channel truncated by the Green Valley fault. The present work was undertaken with the goal of
finding the corresponding portion of the channel on the west side of the fault, presumably offset
to the north due to motion on the fault during approximately the past 10-20 ka. Seismic reflection
data were recorded in order to obtain higher-resolution and more extensive coverage. The new
survey provided excellent resolution of near-surface stratigraphy at depths of 80-375 m, but did
not provide sufficient resolution at shallower depths to permit correlation with data from the
2007 seismic refraction survey, which had a maximum depth penetration of only about 15 m.
The new seismic reflection data appear to image lateral variations in fluvial stratigraphy at
depths of approximately 100 m that could be used to measure fault offset in a future study.
Introduction
The Green Valley fault is located in a portion of the northeastern San Francisco Bay area
(Frizzell and Brown, 1976) that is currently undeveloped but likely to undergo commercial and
residential development in the coming decades. Based on trench studies of the fault at Mason
Road and Lopes Ranch, Lienkaemper et al. (2008) documented at least four earthquakes large
enough to cause surface rupture during the past 1000 years. In a study at Lopes Ranch, Baldwin
et al. (2004), determined a geologic slip rate for the Green Valley fault of approximately 2-4
mm/yr based on approximately 3060 m of displacement during the past 15 ka. These rates are
consistent with the results of McFarland et al. (2007), who conducted theodolite surveys of the
Green Valley fault and found the average creep rate at Mason Road since 2005 to be 2.8 mm/yr,
and the average rate at Red Top Hill since 1984 to be 3.7 mm/yr. However, geodetic modeling of
ongoing regional strain from GPS (Savage et al., 1999) suggests that the slip rate of the Green
Valley fault may be as high as 10 mm/yr. The goal of the present study was to measure the slip
rate of the fault during the past 10-20 ka by finding a buried stream channel that was deposited
within this time range and subsequently offset by fault displacement.
1

Previous Work
A grid of nine fault-parallel seismic refraction profiles was recorded in 2007 with the goal of
imaging paleochannels crossing the fault (Craig et al., 2008). The profiles were 92 m long and
spaced 5 m apart, providing a 92 x 40 m footprint centered on the fault, with its long dimension
oriented N-S, parallel to the fault. The line locations are indicated in yellow in Figure 1. Data
were recorded using a 24-channel spread with 4 m receiver spacing, 24 m shot spacing, and 9
shot locations per profile. A tomographic velocity model was determined for each profile. All
five profiles on the east side of the fault appeared to image a paleochannel, but the profiles on the
west side did not. The abrupt truncation of the channel at the fault location suggested that it had
been offset due to faulting, with the matching portion of the channel on the west side of the fault
presumably displaced to the north. The base of the channel coincided with an abrupt increase in
seismic velocity. Refraction profiles provided velocity information to 10-15 m depth. Velocities
varied from 600 m/s near the surface to 1700 m/s at depth.
Sediment samples from a 7 m deep pit and auger hole south of Mason Road were dated using
radiocarbon and OSL methods. Dates range from Holocene to late Pleistocene age. The pit was
excavated in Summer 2005 to a depth of 7 m (Lienkaemper et al., 2007), and sediment samples
were collected from sandy horizons and dated using the optically-stimulated luminescence
(OSL) method. The deepest and oldest sample was from a depth of 5.6 m and was dated at 14.3
0.43 ka. An auger hole located 135 m to the south, at CPT 13, encountered a coarse-grained layer
at 11 m depth that was radiocarbon dated at 11-12 ka (J. Lienkaemper, personal communication).
Cone-penetration testing (CPT) was conducted at the Mason Road site in 2004 by T. Noce of the
USGS and in 2007 by Gregg Drilling. CPT holes were generally 10-15 m deep, with the deepest
25 m.

Reprocessing of 2007 Seismic Refraction Survey

Data from the 2007 seismic refraction survey were reprocessed in order to verify the previous
results and interpretation, namely the presence of a truncated stream channel at 5-12 m depth on
the east side of the fault. During reprocessing, special care was taken to ensure that all lines were
processed consistently, including filters, picking first arrivals, parameters for traveltime
inversion, and color scales for velocity displays. Geometrics Seisimager software was used to
pick first arrivals and perform traveltime inversions to obtain velocity profiles (Figures 2 and 3).

New Seismic Reflection Survey: Acquisition and Processing

Seismic reflection data were acquired during August 2009. Lines were surveyed and staked out
prior to mobilization of the seismic equipment and crew. Seismic data were recorded over the
course of three days by a crew of three people. The survey consisted of three 180-m long profiles
with 15 m line spacing (Figure 1). All three profiles were located on the north side of Mason
2

Road and on the west side of the fault. Recording parameters are listed in Table 1. Data quality
was generally good to excellent, apart from some extended periods of severe noise from strong
afternoon winds and an automatic pump from a nearby water well. The seismic survey was
conducted in a privately-owned field that was normally under cultivation, but that was briefly
available for data acquisition after the harvest of one crop and before plowing for the next. The
recording surface was essentially flat, with gradually varying topography and less than 0.5 m of
relief along the 180 m length of the profiles. The ground was dry, hard silty clay with no
vegetation. The weather was sunny and hot.
Seismic reflection data were processed in the Windows environment using Parallel Geoscience
Corporation (PGC) Seismic Processing Workshop (SPW) software, using standard methods (e.g.,
Sheriff and Geldart, 1995; Yilmaz et al., 2001), with some modifications to enhance resolution in
the near-surface zone. Prestack gathers were filtered to eliminate low-frequency sourcegenerated noise and high-frequency random noise, and edited to remove noisy traces (Figure 4).
First breaks were eliminated at offsets greater than 6 m using a stovepipe mute (Figure 5). Near
offset traces, at offsets of 3-6 m, were unmuted and served as the sole source of approximately
the first 30 ms of data. Interactive semblance velocity analysis was performed using supergathers
(Figure 6), and the velocity field for each of the three lines reviewed for any sharp lateral
discontinuities in velocity and repicked as needed to provide a smooth field. Figure 7 shows the
velocity field from line MA09-02 after the second pass of velocity analysis. Stacking velocities
typically increased from about 1600 to 2200 m/s over the 100-300 ms range (Figure 8) and were
well constrained by good data quality within this time range. The final stacked sections are
shown in Figures 9-11. The data were processed only to 500 ms since the target of this study was
the near-surface zone. A window of data from 0-300 ms was extracted from each of the three
lines and the three windows were combined in Figure 12. The profiles in Figure 12 are labeled
with both travel time and approximate depths that were calculated based on stacking velocities
from line MA09-02. Common reflectors were correlated using colored horizons.

Results and Conclusions

The reprocessed velocity profiles from the 2007 seismic refraction survey (Figures 2 and 3)
verified the general results and conclusions of the original processing, namely the presence of a
truncated channel on the east side of the fault. All four lines on the east side of the fault (Figure
3) show a broad channel defined by the 1.7 km/s velocity contour at 10-12 m depth, while the
lines on the west side (Figure 2) do not. The matching portion of the channel on the west side of
the fault has presumably been displaced to the north due to right-lateral motion on the Green
Valley fault. The velocity contours for all profiles on the east side of the fault appear to be
displaced downward relative to those on the west side, indicating a component of vertical offset
along the fault and a thicker layer of low velocity sediments on the east side of the fault.

The original processing of the data showed a narrower channel on the east side of the fault than
what is indicated by the reprocessed data. The broader channel shown by the reprocessed data
would be difficult to uniquely identify on the other side of the fault without additional, longer
seismic lines on both sides of the fault.
The new seismic reflection survey clearly imaged stratigraphy over the time range 100-300 ms,
corresponding to a depth range of about 80-375 m, but did not provide sufficient resolution at
shallower depths to permit correlation with data from the 2007 seismic refraction survey.
Arrivals prior to 50 ms tend to be dominated by energy from direct and refracted waves rather
than reflections, and cannot be reliably used to interpret the near-surface stratigraphy.
The new data indicate sedimentary units that are relatively horizontal and planar, with some
gentle undulations (Figures 9-11). Several of the reflectors exhibit lateral changes in character
that likely correspond to lateral variations in bed thickness, including some pinchouts and
bifurcations. These types of features are common in fluvial and alluvial fan environments, and
could also indicate lithologic or facies changes. Lines MA09-02 and MA09-03 contain several
reflectors with lateral changes in reflector character that may be associated the margins of fluvial
channels, e.g., in the interval between horizons 3 and 5 in Figures 12b and 12c. The fact that
similar lateral transitions occur in nearly the same location in all three lines gives us confidence
that they are not artifacts due to transient noise during recording, but rather linear stratigraphic
features that persist across several fault-parallel lines, and that could potentially be used as
piercing points for measuring fault displacement.

Recommendations for further work

In order to measure fault offset at this location, two additional seismic profiles are needed, one
on each side of the fault, using at least a 60-channel seismographic system with shot and
geophone spacings of 3 meters or less. The profiles should be 600 m long, extending 200 m
north of Mason Road and 400 m south of it. The longer profiles would provide the level of
coverage needed to positively identify and match offset stratigraphic features on opposite sides
of the fault. A Betsy Seisgun or equivalent source should be used to provide higher bandwidth
and better resolution of shallow strata. Data acquisition should be suspended during periods of
high noise caused by strong afternoon winds. New data should be processed using both
refraction and reflection methods. With proper survey design and adequate quality control during
data acquisition, these two methods can be used together to provide continuous coverage through
the near-surface zone. The recommended work would be contingent upon the availability the
field on the southern side of the road, which is scheduled for development in the near future.
The primary objective of this work was to search for paleochannels in the uppermost portion of
the data, and this guided the data processing. As may be surmised from the stacked sections
shown in Figures 9-11, clear reflections continue well beyond 500 ms where the data were
4

truncated. The data could be reprocessed utilizing the full original record length of 2.0 s to
provide additional information on deeper stratigraphy well into the 500-1000 m depth range.

Education and student involvement

Four students from California State University, East Bay (CSUEB), including two
undergraduates and two graduate students, participated in this project. Two students helped in
the field with data acquisition, and three were involved with data processing. Data from this
project were used for classroom instruction in two courses at Cal State East Bay: GEOL 4010
(Applied Geophysics) and GEOL 6200 (Advanced Topics - Exploration Seismology).

Acknowledgements
This work was supported by the U.S. Geological Survey, Award No. G09AP0007. Jim
Lienkaemper (USGS) was instrumental in developing the study and provided ongoing
encouragement and support. Uyanga Ganbaatar (CSUEB), Jonathan Arakaki (CSUEB), and
Johnathan Brown (USGS) helped record the seismic surveys. Rufus Catchings and Mark
Goldman (USGS) provided equipment used for data acquisition. Karen and Mike Waller, and
Anthony Russo allowed access to their property for seismic surveying.

References
Baldwin, J. N.; Koehler, R. D.; Thompson, S. C., 2004, Preliminary Late Pleistocene Slip Rate of
the Green Valley Fault at Lopes Ranch Creek, Cordelia, California, Eos Trans. AGU, 85(47),
Fall Meet. Suppl., Abstract S31A-1027.
Craig, M. S., Kimball, M. A., Fay, R., and Lienkaemper, J. J., 2008, Near-Surface Geophysical
Investigations of the Hayward and Green Valley Faults Using Seismic and Ground-Penetrating
Radar Methods, Seismological Research Letters, 79(2), p. 316.
Craig, M., Lienkaemper, J., Crook, N., Dieter, B., Narvaez, R., and Nase, S., 2007, Near-surface
geophysical surveying of the Green Valley fault, Fairfield CA, Eos Trans. AGU, 88(52), Fall
Meet. Suppl., Abstract NS24A-05.
Frizzell, V.A., Jr., and R.D. Brown, Jr., 1976, Map showing recently active breaks along the
Green Valley fault, Napa and Solano Counties, California: United States Geological Survey,
Miscellaneous Field Studies Map 743, 1:24000.
Lienkaemper, J. J., Sickler, R. R., Brown, J., Baldwin, J. N., Turner, R., Jr., and Guilderson, T.
P., 2008, The Green Valley Fault: Geologic Record of Four Large Earthquakes in the Past
Millennium (abstract), Third Conference on Earthquake Hazards in the Eastern San Francisco
Bay Area.

McFarland, F. S., Lienkaemper, J. J., Caskey, S. J., and Grove, K., 2007, Data from Theodolite
Measurements of Creep Rates on San Francisco Bay Region Faults, California: 1979-2007,
USGS Open-File Report 2007-1367.
Savage, J. C., Svarc, J. L., Prescott, W. H., 1999, Geodetic estimates of fault slip rates in the San
Francisco Bay area Journal of Geophysical Research. 104 (B3), pp. 4995-5002.
Sheriff, R. E. and Geldart, L. P., 1995, Exploration Seismology, Second Edition, Cambridge
University Press, 592 pp.
Yilmaz, O., Doherty, S. M., & Yilmaz, O. (2001). Seismic data analysis: Processing, inversion,
and interpretation of seismic data. Investigations in geophysics, no. 10. Tulsa, OK: Society of
Exploration Geophysicists.

Recording parameters:
Shot spacing
Receiver spacing
CMP spacing
Source type
Hammer blows per shot
Geophone frequency
Number of channels
Spread length
Minimum offset
Maximum offset
Sample interval
Record length

3m
3m
1.5 m
sledgehammer
3
40 Hz
60
180 m
0m
180 m
0.5 ms
2.0 s

Processing Sequence:
Convert from SEG-Y to SPW format.
Geometry assignment
Trace sort

Truncate data at t = 600 ms

Bulk shift -100 ms.
Band-pass filter: 40-150 Hz
Kill noisy traces
Spherical divergence correction, 4 dB/sec
Mute first arrivals
Sort to CMP gathers
Bin supergathers (5 CMPS per gather)
Semblance velocity analysis (15 m spacing)
Normal moveout (NMO) correction
Brute stack
Velocity field display
Semblance velocity analysis (2nd pass)
Final stack
Bandpass filter 40-150 Hz
f-x deconvolution

Table 1. Recording parameters and processing sequence for 2009 seismic reflection survey.

1985800

1986000

563400
101

101

563400

3
160 MA09-0
2
-0
9
MA0
1
0
9
0
0
6
1
MA

1985600

563200
563000

LEGEND
2009 reflection
2007 refraction
2007 trench

SCALE
0

25 50
1985600

100

150

Meters
200
1985800

Fault, approximate
Fault, inferred
1986000

Figure 1. Basemap showing locations of 2009 seismic reflection lines (blue), 2007 seismic
refraction lines (yellow), 2007 trench (green), and Green Valley fault (red). Fault trace and
trench location courtesy J. Lienkaemper. Coordinates are NAD 1983 State Plane Coordinates in
units of meters.
8

562800

562800

GV07-11
GV07-13
GV07-15
GV07-17
GV07-19

563000

GV07-12
GV07-14
GV07-16
GV07-18

563200

MASON ROAD

a) GV07-11

b) GV07-12

c) GV07-13

d) GV07-14

Figure 2. Seismic refraction profiles from 2007 survey on the west side of the Green Valley fault. Vertical scale is
depth and horizontal scale is distance, both in units of meters. Contours are velocities in units of km/s, contour
interval is 0.1 km/s. a) GV07-11, b) GV07-12, c) GV07-13, d) GV07-14. See Figure 1 for location map.

a) GV07-16

b) GV07-17

c) GV07-18

d) GV07-19

Figure 3. Seismic refraction profiles from 2007 survey on the east side of the Green Valley fault. Vertical scale is
depth and horizontal scale is distance, both in units of meters. Contours are velocities in units of km/s, contour
interval is 0.1 km/s. a) GV07-16, b) GV07-17, c) GV07-18, d) GV07-19. Note the similarity of the four profiles in
this figure, and that the velocity contours in these profiles are displaced downward relative to those on the other side
of the fault (Figure 2).

10

a)

b)

Figure 4. Filtering. a) Shot gather before filtering. The first few traces on the left-hand side of the gather are
dominated by low-frequency noise, and some of the traces on right-hand side of the record contain high-frequency
noise. b) The same shot gather after band-pass filter and removal of noisy traces.

11

a)

b)

Figure 5. Muting. a) Shot gather with early mute superimposed. The portion of data shaded in gray will be muted.
b) Shot gather after muting.

12


Figure 6. Example of CMP supergather velocity analysis from Line MA09-01. The colored figure on the left shows
semblance as a function of velocity (horizontal axis) and time (vertical axis). The figure on the right is a CMP
supergather that has been corrected for NMO using the velocity function that has been interactively picked (black
line) on the left-hand figure. The circular bullseye in the semblance plot at 100-130 ms indicates a well-constrained
velocity that is defined by a coherent reflector over an adequate offset range.

Figure 7. Velocity field for line MA09-02. Each of the vertical black lines indicates the location of a CMP
supergather and semblance velocity analysis, as shown in Figure 6 above. The point symbols indicate velocity picks.

13

Velocity(m/s)
500

1000

1500

2000

2500

3000

50

100
170
175

150

180
185

Time(ms)

200

190
195

250

200
205

300

210
350

400

450

500

Figure 8. Stacking velocities from Line MA09-02. Each line is a velocity function from a different velocity analysis
location. Symbols are picks from semblance velocity analysis and correspond to the velocities shown in Figure 7
above.

14

Figure 9. Seismic reflection profile MA09-01. The red box indicates the extent of the data shown in Figure 12a.

15

Figure 10. Seismic reflection profile MA09-02. The red box indicates the extent of the data shown in Figure 12b.

16

Figure 11. Seismic reflection profile MA09-03. The red box indicates the extent of the data shown in Figure 12c.

17

110

MA09-01

120

Shotpoint
130

140

150
0

Time (ms)

100

80

200

200

300

375

MA09-02
110

120

130

140

150

Time (ms)

1
100

200

110

MA09-03

120

130

140

80

4
5
6

300

c)

200

Approximate depth (m)

375

150
0

1
100

80

5
6

200

300

200

375

Figure 12. Seismic profiles and interpretations for a portion of seismic lines a) MA09-01, b) MA09-02, and
c) MA09-03. Selected horizons common to all three lines are indicated in color and labeled 1-7. Times are
two-way travel times. Labeled depths are estimated based on stacking velocities from Line MA09-02.
Sections are unmigrated. The complete uninterpreted profiles are shown in Figures 9-11.

18

Approximate depth (m)

b)

Time (ms)

Approximate depth (m)

a)