Download 1 Apparent vertical Moho offsets under continental strike

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Plate tectonics wikipedia , lookup

Magnetotellurics wikipedia , lookup

Earthquake wikipedia , lookup

Shear wave splitting wikipedia , lookup

Surface wave inversion wikipedia , lookup

Earthscope wikipedia , lookup

Transcript
Apparent vertical Moho offsets under continental strike-slip faults from lithology contrasts in the
seismogenic crust
Vera Schulte-Pelkum1, Yehuda Ben-Zion2
1
Cooperative Institute for Research in Environmental Sciences and Department of Geological Sciences,
University of Colorado, Boulder, Colorado 80309-0399; [email protected]
2
Yehuda Ben-Zion
Department of Earth Sciences, University of Southern California, Los Angeles, CA 90089-0740
[email protected]
Bull. Seism, Soc. Am.
In press, June 2012
1
Abstract
We demonstrate that observed contrasts of upper crustal structures across large continental strikeslip faults are sufficient to produce artificial vertical Moho offsets of several km in studies using
laterally homogenous models. This can have significant impact on efforts to understand how the
observed surface displacement is accommodated in the deep crust. An existing hypothesis that
continental strike-slip faults extend as discrete narrow features through the entire crust rests partly on
inferred vertical Moho offsets across the faults based on teleseismic converted wave (receiver function)
studies. However, such studies typically do not account for the juxtaposition of different structures
across the faults, which can lead to significant (e.g. 10%) contrasts of seismic velocities in the upper
crust. Observed profiles of velocity contrasts across faults can bias images of the deep structure to the
extent where a large (> 5 km) Moho offset can be purely apparent rather than real. Constraining the
existence and size of true vertical Moho offsets below continental strike-slip faults thus require
calculations that account for the across-fault lithology contrasts. Accurate results on the geometry of
Moho across faults can have important implications for lithospheric deformation, crustal tomography,
strain localization and partitioning in the crust, continental rheology, and geodynamics.
1. Introduction
Continental transform faults can have tens to hundreds of km of displacement across very narrow
(sub m) fault zones in the shallow crust [e.g., Chester et al, 1993; Sibson 2003; Rockwell and Ben-Zion
2007]. The geometry and strain localization of several continental strike-slip faults are now reasonably
well resolved in the seismogenic zone, where local seismicity can be used for fault imaging [e.g.
Eberhart-Phillips and Michael, 1998; McGuire and Ben-Zion, 2005; Lin et al., 2007]. It is more
difficult to resolve fault properties below the brittle-ductile transition, where one has to rely mostly on
surface wave and teleseismic body wave data. Neither surface wave dispersion nor teleseismic travel
2
time tomography are sufficiently sensitive to velocity discontinuities to be used for tracing deep fault
structures.
Teleseismic converted waves (receiver functions) provide a popular method for imaging
discontinuities in shear wave velocity. The presence or absence of a vertical offset in the Moho across a
fault has been used as a proxy for inferring whether a discrete narrow zone of deformation continues
throughout the crust [e.g. Stern and McBride, 1998; Zhu, 2000; Wittlinger et al., 2004] or whether
distributed deformation occurs over a wider diffuse zone in the lower crust [e.g. Wilson et al., 2004;
Pedrera et al., 2010]. However, the observable given by typical receiver functions is the time difference
between a teleseismic P arrival and the converted phase from incident P to S wave at a deep shear
velocity contrast (e.g. Moho). Converting delay time to Moho depth requires knowledge of the P- and
S-velocities in the crust.
Large continental strike-slip faults are likely to separate different crustal blocks. In several places
sharp contrasts have been imaged as extending to the bottom of the seismogenic zone, by using head
waves that refract along fault bimaterial interfaces [e.g., Ben-Zion and Malin, 1991; McGuire and BenZion, 2005; Bulut et al. 2012]. Tomographic studies also reveal clear property contrasts across large
strike-slip faults [e.g., Thurber et al. 2006; Roecker et al. 2006; Tape et al., 2009]. Strong upper crustal
velocity contrasts can distort images of deep structure significantly when they are unknown or not
taken into account explicitly.
A standard receiver function technique is the common conversion point (CCP) stack, where
receiver function amplitudes are projected along ray paths in an assumed velocity model. While a depth
migration using 3-D Vp and Vs models is possible, typically fixed crustal Vp and Vp/Vs ratio are
assumed to calculate depth when making CCP and similar receiver function profiles [e.g. Zhu, 2000;
Wilson et al., 2004; Yan and Clayton, 2007; Porter et al., 2011]. Biases in Moho topography from
receiver functions, when used as an a priori input for tomographic inversions, introduce a compounded
3
bias in tomographic results [Tian et al., 2007]. Other methods that stack receiver functions Moho
reverberations to resolve the velocity-depth tradeoff [Zhu and Kanamori, 2000; Kumar and Bostock,
2008; Helffrich and Thompson, 2011] may suffer biases from lateral heterogeneity and unmodeled
sediment reverberations (Yeck, Y., A. F. Sheehan and V. Schulte-Pelkum, manuscript in preparation,
2012).
Some of the most-studied examples in the debate on the depth extent of strike-slip faults are in
California, in particular the San Andreas fault system [Zhu, 2000; Fuis et al. 2001; Ryberg et al., 2009,
Thornton and Zhou, 2008]. Fig. 1 shows variations of surface shear velocities in California as a proxy
for changes in the shallow sedimentary cover. Receiver function estimates of Moho depth are affected
by sedimentary cover as well as by upper crustal velocity variations, both of which show significant
contrasts across faults [e.g., Ben-Zion et al. 1992; Wills et al., 2006; Tape et al., 2009]. In the following
we use available high-resolution velocity models for sections of the San Andreas and San Jacinto faults
with clear across-fault variations. We use the shallow models to estimate possible biases in Moho depth
that result from ignoring the observed lithology contrasts over the seismogenic depth section.
The velocity-depth tradeoff is a well-known problem in receiver function interpretation. Given the
recent advances in imaging across-fault velocity contrasts, our intent is to quantify the depth bias
introduced by such known contrasts in order to determine whether they impact traditional imaging at
Moho depths significantly or not. We deliberately use a simple modeling approach (two 1-dimensional
models joined at the fault) to estimate first-order possible effects on imaged depth. Various additional
effects may be introduced by unmodeled processes such as attenuation in sediments and scattering by
the fault, but those will be dependent on details of parametrization, geometry, and modeling technique
and are of second order compared to the fundamental question of depth bias while migrating the Moho
conversion to depth.
4
2. Method
We use determined velocity models from two opposite sides of the fault locations indicated in Fig.
1 (yellow circles). The first location is on the San Andreas Fault (SAF) south of Hollister in central
California, and the second is on the San Jacinto Fault Zone (SJF) between Hemet and Anza in southern
California. We create synthetic CCP profiles perpendicular to the faults by calculating synthetic
seismograms for separate 1-D models on the opposite sides of each fault, produce receiver functions,
and combine those in CCP stacks.
In a depth profile produced by CCP stacking, rays are projected onto a vertical plane connecting the
stations. Similar to tomography, crossing rays are required for full illumination of a 2- or 3dimensional structure. The 1-D approximation for calculations of the synthetic seismograms is
representative for the SAF and SJF locations used here because of the orientation of the major faults
relative to the teleseismic event distribution (Fig. 1, inset). Most of the seismicity illuminates the faults
at near fault-parallel, with incidence angles from the NW and SE (circum-pacific and Central/South
American seismicity, respectively). The only fault-perpendicular illumination is from the SW Pacific
subduction zones and is steeply incident, with no available crossing rays from the NE. Nevertheless, to
assess the generality of our results to other locations, we also calculate a maximum width of cross-fault
illumination for an ideal case with complete ray coverage in azimuth and distance ranges.
For the SAF location, we take the Vp models published for the SE and NW sides of the fault based
on inversion of observed fault zone head and direct P waves [Fig. 10 of Lewis et al., 2007]. This
location has a several km thick sediment layer on the NE side of the fault that is not present on the SW
side (Fig. 1). The models have a maximum depth of 15 km. We conservatively continue them from 15
km down to an assumed Moho depth of 30 km with the same velocity on both sides of the fault (Fig.
2c). The Vs and density values are obtained from Vp using the scaling relations of Brocher [2005, eqs.
1 and 6]. We calculate synthetic seismograms for this example using the ray-based code of Frederiksen
5
and Bostock [2000], which offers the option of turning reverberations off (Fig. 2a) or on (Fig. 2b). We
use a pulse width of 0.75 s, comparable to commonly used frequencies in receiver function analysis.
The velocity profiles from Lewis et al. (Fig. 2c) utilize head waves that refract along the fault, so in
this sense are high-resolution images of the across-fault contrast, but they are also fairly simple in
having 4 constant velocity layers. For the SJF location (Fig. 3), we use travel-time tomographic images
of Allam and Ben-Zion (2012) for Vp and Vs to 28 km depth (Fig. 3c), and a reflectivity code
[Randall, 1994] to calculate synthetic seismograms, with the same pulse width as in the ray-based code
for the San Andreas Fault location. The velocity models of Allam and Ben-Zion [2012] include more
details on variations along strike and with depth than the simple layer model of Lewis et al. [2007], but
they probably represent smeared lower bound results on the velocity contrast across the fault. In
contrast to the SAF example, the SJF models lack a significant one-sided sediment layer.
For all three sets of synthetic receiver functions, CCP stacks are calculated assuming a constant
bulk crustal Vp of 6.4 km/s and Vp/Vs ratio of 1.75 everywhere; these values are standard in CCP
processing and were used for profiles across the SAF by Zhu [2000] and Porter et al. [2011]. The
resulting apparent Moho offsets are not very sensitive to the choice of the constant Vp and Vs used in
CCP stacking.
3. Results
The CCP stack for the San Andreas Fault example without reverberations (Fig. 2a) illustrates the
distortion on the 30 km deep Moho due to the differences in upper crustal velocity structure across the
fault. The Moho appears more than 7.5 km deeper on the NE side of the fault than on the SW side.
When reverberations are added (Fig. 2b), the Moho arrival on the NE side of the fault is nearly
canceled by a negative sediment reverberation (2 S and 1 P legs in the crust), and the dominant arrival
near the expected Moho depths is a preceding positive sediment reverberation (2 P and 1 S leg). This
6
may be picked as the apparent Moho, leading to an apparent opposite sense vertical Moho offset of 5
km. The broadening of the apparent Moho arrival can bias estimates of the sharpness of the velocity
transition across the Moho. Sediment reverberation amplitudes would be muted if attenuation and
lateral thickness variations were to be added to the modeling; however, this would also affect the Moho
conversion amplitude.
Fig. 2a and 2b also show converted shear wave ray path ranges for the ideal case of complete
teleseismic P distance range coverage and hypothetical fault-perpendicular incidence for the velocity
model on each side of the fault. The maximum horizontal range of overlapping ray paths across the
fault, where distortions from the fault interface may affect the image, is ~20 km at Moho depth and ~10
km for a contrast reaching down to 15 km depth given ideal station placement.
The SJF example demonstrates Moho distortion resulting from gradual velocity variations without
a significant one-sided sediment layer as in the San Andreas Fault example (Fig. 3). The CCP stack
from the full waveform synthetic receiver functions (including reverberations) shows an apparent
vertical offset of 3.5 km despite the uniform input Moho depth of 28 km.
Fig. 4a presents the apparent vertical Moho offset as a function of the thickness of an unmodeled
low Vp and high Vp/Vs ratio layer (sediment cover is used as an example). Fig. 4b shows the apparent
Moho offset due to a bias in the average crustal Vp or Vp/Vs. Realistic perturbations give rise to
significant apparent Moho offsets in all cases.
4. Discussion and Conclusions
The question of whether continental strike-slip faults cut through the Moho has wide-ranging
implications for lithospheric rheology and dynamics. Proposed models include that (i) such faults sole
out into a decollement near the brittle-ductile transition within the crust in thin-skinned tectonics
settings [e.g. Laubscher, 2010]; (ii) deformation continues into the lower crust but in a broad diffuse
7
zone of deformation [e.g. Wilson et al., 2004]; (iii) localized faults (a few km wide at most) cut through
the entire crust or even the entire lithosphere including the lithospheric mantle [e.g. Meyer et al., 1998].
Recent observations of tremor and low-frequency earthquakes near Parkfield on the San Andreas fault
show sources close to Moho depths that tend to line up subparallel to the fault, with lateral offsets of a
few to 14 km horizontal distance to the surface trace of the fault depending on the location technique
[Shelly et al., 2009; Ryberg et al., 2010]; the tremor locations suffer from the same bias due to shallow
velocity structure as discussed here.
Some surface exposures of deep crustal strike-slip zones show single narrow (~5 km) mylonite
zones with > 100 km displacement at midcrustal depths [Dumond et al., 2008]. In contrast, exhumed
lower crust shows in places broad (~ 50 km) shear zones consisting of multiple narrow (~5 km) shear
bands separating blocks with little internal deformation, suggesting wide zones of distributed shear
[Williams et al., 2000; Dumond et al., 2010, S2 fabric therein]. In both of these examples, the fault
geometry in the brittle crust above is unknown, although generally shear localization tends to increase
with decreasing temperature [e.g. Kaus and Podlachikov, 2006].
Geophysical imaging of present-day deep fault structure remains somewhat inconclusive to date,
not the least due to the effects illustrated quantitatively in this study. Gravity may be added to support
interpretations of Moho structure (e.g. Zhu, 2000; Salmon et al., 2007). Although the magnitude of a
Bouguer anomaly caused by a sedimentary basin is comparable to that of a thickened crust and its sign
is the same, its lateral gradient should be steeper for a shallower density contrast than for a deep
contrast. Sediment thickness contrasts above a possible Moho step (Zhu, 2000) may still complicate
gravity interpretation.
An alternative H-κ stacking method for estimating Moho depth (Zhu and Kanamori, 2000) uses
reverberations to reduce the velocity-depth tradeoff. It may be applicable to imaging Moho offsets in
cases where seismicity is distributed such that rays illuminate the fault from both sides and lateral
8
heterogeneity away from the fault is minor. In practice, tectonically active areas often show significant
scatter in H- κ results and unclear reverberations in moveout curves due to lateral heterogeneity. The
results are therefore typically smoothed laterally before interpretation (e.g. Zhu and Kanamori, 2000;
Rumpfhuber et al., 2009; Eagar et al., 2011), so inferring a local Moho step may be questionable unless
several stations on each side of the fault give consistent results.
Our modeling shows that the accuracy with which deep crustal structure and processes can be
imaged trades-off to a large degree with accurate knowledge of the upper crust. Fortunately, the latter
exists (or can be derived) at various locations of interest in California and other well-instrumented
areas. In places where high-resolution upper crustal velocities are not available, determination of upper
to lower crustal tradeoffs with synthetic calculations of the type done in this work can place error
bounds on deep imaging results.
Data and Resources
Waveform data to determine azimuthal and slowness coverage of teleseismic P receiver functions
shown in the inset in Fig. 1 were obtained from the Incorporated Research Institutions for
Seismology’s Data Management Center. The Vs30 model shown in Fig. 1 was downloaded from
http://opensha.org/apps-SiteData (Field et al., 2003).
Acknowledgments
We thank Kevin Mahan, Karl Mueller, Yaman Ozakin, and William Yeck for discussions. Fig. 1
was plotted with GMT [Wessel and Smith, 1991] with Vs30 data obtained using OpenSHA [Field et
al., 2003] with assistance by Kevin Milner. The paper benefited from useful comments by two
anonymous referees and Zhigang Peng. The study was supported by the National Science Foundation
(grant EAR-0908903).
9
References
Allam, A. A. and Y. Ben-Zion, Seismic velocity structures in the Southern California plate-boundary
environment from double-difference tomography, Geophys. J. Int., doi: 10.1111/j.1365246X.2012.05544.x, 2012.
Ben-Zion, Y. and P. Malin, San Andreas fault zone head waves near Parkfield, California, Science,
251, 1592-1594, 1991.
Ben-Zion, Y., S. Katz and P. Leary, Joint inversion of fault zone head waves and direct P arrivals for
crustal structure near major faults, J. Geophys. Res, 97, 1943-1951, 1992.
Brocher, T., Empirical relations between elastic wavespeeds and density in the Earth’s crust, Bull.
Seism. Soc. Am. 95, 2081-2092, 2005.
Bulut, F., Y. Ben-Zion and M. Bonhoff, Evidence for a bimaterial interface along the Mudurnu
segment of the North Anatolian Fault Zone from polarization analysis of P waves, Earth Planet.
Sci. Lett., 327-328, 17-22, doi:10.1016/j.epsl.2012.02.001, 2012.
Chester, F. M., Evans, J. P., and Biegel, R. L., Internal structure and weakening mechanisms of the San
Andreas fault, J. Geophys. Res, 98, 771-786, 1993.
Dumond, G., P. Goncalves, M. L. Williams, M. J. Jercinovic, Subhorizontal fabric in exhumed
continental lower crust and implications for lower crustal flow: Athabasca granulite terrane,
western Canadian Shield, Tectonics 29, TC2006, 2010.
Dumond, G., N. McLean, M. L. Williams, M. J. Jercinovic, S. A. Bowring, High-resolutions dating of
granite petrogenesis and deformation in a lower crustal shear zone: Athabasca granulite terrane,
western Canadian Shield, Chem. Geol. 254, 175-196, 2008.
Eagar, K. C., M. J. Fouch, D. E. James, and R. W. Carlson, Crustal structure beneath the High Lava
Plains of eastern Oregon and surrounding regions from receiver function analysis, Journ. Geophys
10
Res. 116, B02313, 2011.
Eberhart-Phillips, D., and A. J. Michael (1998), Seismotectonics of the Loma Prieta, California, region
determined from three-dimensional Vp , Vp / Vs , and seismicity, J. Geophys. Res., 103(B9), 21,099–
21,120, doi:10.1029/98JB01984.
Field, E.H., T.H. Jordan, and C.A. Cornell, OpenSHA: A Developing Community-Modeling
Environment for Seismic Hazard Analysis, Seis. Res. Lett. 74, 406-419, 2003.
Frederiksen, A. and M. Bostock, Modelling teleseismic waves in dipping anisotropic structures,
Geophys. J. Int. 141, 401-412, 2000.
Fuis, G.S., Ryberg, T., Godfrey, N., Okaya, D.A., and Murphy, J.M., Crustal structure and tectonics
from the Los Angeles basin to the Mojave Desert, southern CA. Geology, 29, 15-18, 2001.
Grand, S. P. and D. V. Helmberger, Upper mantle shear structure of North America, Geophys. J. Royal
Astr. Soc. 76, 399-438, 1984.
Helffrich, G. and D. Thompson, A stacking approach to estimate Vp/Vs from receiver functions,
Geophys. J. Int. 182, 899-902, 2010.
Kaus, B. and Y. Y. Podladchikov, Initiation of localized shear zones in viscoelastoplastic rocks, J.
Geophys. Res. 111, B04412, 2006.
Kumar, M. R. and M. G. Bostock, Extraction of absolute P velocity from receiver functions, Geophys.
J. Int. 175, 515-519, 2008.
Lewis, M. A., Y. Ben-Zion and J. J. McGuire, Imaging the deep structure of the San Andreas Fault
south of Hollister with joint analysis of fault zone head and direct P arrivals, Geophys. J. Int. 169,
1028-1042, 2007.
Lin, G., P. M. Shearer, E. Hauksson, and C. H. Thurber, A three-dimensional crustal seismic velocity
model for southern California from a composite event method, J. Geophys. Res., 112, B11306,
2007.
11
McGuire, J. and Y. Ben-Zion, High-resolution imaging of the Bear Valley section of the San Andreas
Fault at seismogenic depths with fault-zone head waves and relocated seismicity, Geophys. J. Int.,
163, 152-164, 2005.
Meyer, B., P. Tapponnier, L. Bourjot, F. Metivier, Y. Gaudemer, G. Peltzer, G. Shunmin, C. Zhitai,
Crustal thickening in Gansu-Qinghai, lithospheric mantle subduction, and oblique, strike-slip
controlled growth of the Tibet plateau, Geophys. J. Int., 153,1-47, 1998.
Pedrera, A., F. D. Mancilla, Ruiz-Constan, A., Galindo-Zaldivar, J. et al., Crustal-scale transcurrent
fault development in a weak-layered crust from an integrated geophysical research: Carboneras
Fault Zone, eastern Betic Cordillera, Spain, Geochem. Geophys. Geosyst. 11, Q12005, 2010.
Porter, R., G. Zandt and N. McQuarrie, Pervasive lower crustal seismic anisotropy in southern
California: Evidence for underplated schists and active tectonics, Lithosphere 3, 201-220, 2011.
Randall, G. E., Efficient calculation of complete differential seismograms for laterally homogeneous
Earth models, Geophys. J. Int. 118, 245-254, 1994.
Roecker, S., C. Thurber, K. Roberts, L. Powell, Refining the image of the San Andreas Fault near
Parkfield, California using a finite difference travel time computation technique, Tectonophysics
426, 189-205, 2006.
Rockwell, T. K. and Ben-Zion Y., High localization of primary slip zones in large earthquakes from
paleoseismic trenches: Observations and implications for earthquake physics, J. Geophys. Res.,
112, B10304, 2007.
Rumpfhuber, E. M., G. R. Keller, E. Sandvol, A. A. Velasco, and D. C. Wilson, Rocky Mountain
evolution: Tying continental dynamics of the Rocky Mountains and Deep Probe seismic
experiments with receiver functions, Journ. Geophys. Res. 114, B08301, 2009.
Ryberg, T., C. Haberland, G. S. Fuis, W. L. Ellsworth, D. R. Shelly, Locating non-volcanic tremor
along the San Andreas Fault using a multiple array source imaging technique, Geophys. J. Int. 183,
12
1485-1500, 2010.
Salmon, M. L., T. A. Stern, and M. K. Savage, A major step in the continental Moho and its
geodynamic consequences: the Taranaki-Ruapehu line, New Zealand, Geophys. J. Int. 186, 32-44,
2011.
Shelly, D. R., W. L. Ellsworth, T. Ryberg, C. Haberland, G. S. Fuis, J. Murphy, R. M. Nadeau, R.
Buergmann, Precise location of San Andreas Fault tremors near Cholame, California using
seismometer clusters: Slip on the deep extension of the fault?, Geophys. Res. Lett. 36, L01303,
2009.
Sibson, R. H., Thickness of the seismic slip zone, Bull. Seis. Soc. Am. 93, 1169-1178, 2003.
Stern, T. A. and J. H. McBride, Seismic exploration of continental strike-slip zones, Tectonophys. 286,
63-78, 1998.
Tape, C., Q. Liu, A. Maggi and J. Tromp, Adjoint tomography of the southern California crust, Science
325, 988-992, 2009.
Thurber, C., H. Zhang, F. Waldhauser, J. Hardebeck, A. Michael, D. Eberhardt-Phillips, Threedimensional compressional wavespeed model, earthquake relocations, and focal mechanisms for
the Parkfield, California, region, Bull. Seism. Soc. Am. 96, S38-S49, 2006.
Tian, Y., D. Zhao and J. Teng, Deep structure of southern California, Phys. Earth Planet. Int. 167, 93113, 2007.
Wessel, P. and W. H. F. Smith, Free software helps map and display data, EOS Trans. Am. Geophys.
Union, 72, 441, 445-446, 1991.
Williams, M. L., E. A. Melis, C. F. Kopf and S. Hanmer, Mictrostructural tectonometamorphic
processes and the development of gneissic layering: a mechanism for metamorphic segregation, J.
Metamorph. Geol. 18, 41-57, 2000.
Wills, C. J. and K. B. Clahan, Developing a map of geologically defined site-condition categories for
13
California, Bull. Seism. Soc. Am. 96, 1483-1501, 2006.
Wilson, C. K., C. H. Jones, P. Molnar, A. F. Sheehan, and O. Boyd, Distributed deformation in the
lower crust and upper mantle beneath a continental strike-slip fault zone: Marlborough fault
system, South Island, New Zealand, Geology 32, 837-840, 2004.
Wittlinger, G., J. Vergne, P. Tapponier, V. Farra, G. Poupinet, M. Jiang, H. Su, G. Herquel, A. Paul,
Teleseismic imaging of subducting lithosphere and Moho offsets beneath western Tibet, Earth
Planet. Science Lett. 221, 117-130, 2004.
Yan, Z. and R. W. Clayton, Regional mapping of the crustal structure in southern California from
receiver functions, J. Geophys. Res. 112, B05311, 2007.
Zhu, L., Crustal structure across the San Andreas Fault, southern California from teleseismic converted
waves, Earth Planet. Sci. Lett. 179, 183-190, 2000.
Zhu, L. and H. Kanamori, Moho depth variations in southern California from teleseismic receiver
functions, J. Geophys. Res. 105, 2969-2980, 2000.
Author affiliations and addresses
Vera Schulte-Pelkum
Cooperative Institute for Research in Environmental Sciences and Department of Geological Sciences,
University of Colorado, Boulder, Colorado 80309-0399
[email protected]
Yehuda Ben-Zion
Department of Earth Sciences, University of Southern California, Los Angeles, CA 90089-0740
[email protected]
14
Figure Captions
Fig. 1: Topography (shading), surface shear velocity (color; average shear velocity in top 30 m, from
Wills et al., 2006; shown as proxy for sediment cover) and faults (black lines) in California. Yellow
circles are locations of velocity models immediately SW and NE of faults used in Figs. 2-3. Green
circle is location of vertical Moho offset inferred in a previous study (Zhu, 2000). Yellow triangles
mark 10 seismic stations whose teleseismic P distribution is shown in inset. The inset has a rose
histogram of backazimuths for P receiver functions (number of usable quality receiver functions per 5
degree backazimuthal bin) for 5 years of data at the indicated stations. Green lines show range of faultparallel and fault-perpendicular orientations of San Jacinto Fault Zone and San Andreas Fault in the
fault segments of interest. Blue portions of histogram show number of receiver functions that have a
slowness less than 0.055 s/km (steep incidence), red portions of histogram indicate receiver functions
with greater slowness (shallow incidence). The rays from the SW are steeply incident and no seismicity
exists to the NE, so illumination of the faults with crossing rays is nearly nonexistent.
Fig. 2: (a) Synthetic CCP profile perpendicular to the fault at 0 km (thick black line, continued as
dotted line at depths without input velocity contrast), calculated with synthetic seismograms using a
ray-based 1-D method (Frederiksen and Bostock, 2000), with reverberations not modeled (cf. b). The
used velocity models are Vp immediately SW (left) and NE (right) of the San Andreas Fault (SAF)
south of Hollister (Lewis et al., 2007; northern yellow circle in Fig. 1), and Vs and density scaled from
Vp using Brocher (2005). CCP stacking with the common assumption of constant bulk crustal
velocities, without accounting for across-fault velocity variations, leads to an apparent 7 km offset in
the Moho (white bar, Moho depths picked shown as dashed white lines). Triangles show assumed
station locations. Color scale is ratio of radial to vertical amplitude in percent; example receiver
15
functions, corrected to vertical incidence, are shown as thin blue lines on each side of the fault. Pairs of
thin black lines on each side of the fault show incidence angle ranges of phases converted from P to S
at the true model Moho (steepest and shallowest possible ray within teleseismic P range). The
horizontal distance traveled by the ray from an assumed conversion depth can be used to estimate the
maximum range of disruption of the image by fault-crossing rays if ray paths perpendicular to the fault
trace were available. (b) Same as a), but with reverberations turned on in synthetics. The negative
(blue) multiple (reflected once off the free surface with one P and two S legs) interferes with the true
Moho arrival near ~35 km apparent depth. The preceding prominent 2P+1S sediment multiple may be
picked as the Moho, leading to an apparent Moho offset of ~5 km (arrival is peaked at contrast depth).
(c) Velocity models used: blue – SW, red – NE, solid line – Vp, dashed – Vs; (d) Vp/Vs for SW (blue),
NE (red). The calculations assume conservatively no contrast below 15 km.
Fig. 3: (a) Fault-perpendicular synthetic CCP profile (fault at 0 km) calculated for contrasting
tomographic Vp and Vs models with velocity gradients (Allam and Ben-Zion, 2012) across the San
Jacinto Fault (SJF) between Anza and Hemet (southern yellow circle in Fig. 1). Legend same as in Fig.
2. In comparison to Fig. 2 based on simple layered structure, the gradients in these models lead to a less
sharp contrast. Here both Vp and Vs are derived from local P and S arrivals rather than assuming a Vp
scaling. Synthetics including reverberations were calculated with a reflectivity code (Randall, 1994).
Input Moho was set to 28 km depth on both sides of the fault; standard CCP processing as for Figs. 2a
and 2b leads to an apparent 3.5 km Moho offset. Color scale is radial to vertical amplitude in percent.
(b) Velocity models used: blue – SW, red – NE; solid line – Vp, dashed – Vs; (c) Vp/Vs for SW
(blue), NE (red).
Fig. 4: (a) Apparent increase in Moho depth when assuming standard CCP stacking parameters (bulk
16
crustal Vp of 6.4 km/s and Vp/Vs = 1.73) in the presence of a sedimentary layer with Vp = 2.8 km/s
(e.g. Fig. 3b) and Vp/Vs of 1.73 (red line) or 2.0 (blue line, more realistic for sediments). Thicker
layers with smaller Vp and Vp/Vs deviations from the values assumed in CCP stacking lead to
comparable Moho biases, as the travel time delay scales with layer thickness and velocity bias. (b)
Apparent increase in Moho depth as a function of decreasing Vp or increasing Vp/Vs on average in the
crust.
17
Fig. 1
18
Fig. 2
19
Fig. 3
20
Fig. 4
21