RICE UNIVERSITY
Imaging Lithospheric Structure Beneath the Colorado Plateau and its
Adjacent Regions Using Rayleigh Wave Tomography
by
Kaijian Liu
A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE
Master of Science
APPROVED, THESIS COMMITTEE:
Alan Levander, Carey Croneis Professor, Chair Earth Science
Fenglin Niu, Associate Professor Earth Science
David Alexander; Professor Physics and/Astronomy
HOUSTON, TEXAS
APRIL 2010
UMI Number: 1486042
All rights reserved
INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted.
In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if material had to be removed,
a note will indicate the deletion.
UMT Dissertation Publishing
UMI 1486042 Copyright 2010 by ProQuest LLC.
All rights reserved. This edition of the work is protected against unauthorized copying under Title 17, United States Code.
ProQuest LLC 789 East Eisenhower Parkway
P.O. Box 1346 Ann Arbor, Ml 48106-1346
ABSTRACT
Imaging Lithospheric Structure Beneath the Colorado Plateau and its Adjacent Regions using Rayleigh Wave Tomography
by Kaijian Liu
This thesis presents a new 3-D shear velocity (Vs) model of the lithospheric structure
beneath the Colorado Plateau and its surrounding regions to understand the complicated
lithospheric modifications caused by major Cenozoic tectonic and magmatic activities.
Prior to this work, lack of an overall velocity model had prevented a comprehensive
understanding of the current encroachment patterns near the margins, and of the
perplexing crust/mantle boundary beneath the plateau. Using the new USArray data, I
have inverted the isotropic Vs model from the Rayleigh wave dispersion curves obtained
by a modified two-plane wave method. The resulting Vs structures not only clearly image
the sublithospheric low-velocity channels, but also identify the high velocity anomalies
underlying the western plateau for the first time in surface wave tomography to support
the recent delamination hypothesis. This work also provides a high-resolution velocity
model for geodynamic modeling on lithosphere-asthenosphere interactions.
11
ACKNOWLEDGEMENTS
I am thankful and indebted to many people who offered me help and encouragement
during my thesis work. Much is owed to my advisor, Dr. Alan Levander, for his great
supervision, valuable suggestion, continuous support, helpful discussion and patient
thesis editing. I have benefited greatly from his profound knowledge of the western
United States and his enthusiasm in Earth science.
I would also like to thank Dr. Fenglin Niu for his invaluable help on the surface wave
tomography code, the data processing and many suggestions. I am grateful to Dr. David
Alexander for his constructive and insightful comments on my thesis.
I would like to thank the talented Rice seismology group for their suggestions and special
thanks go to Yongbo Zhai, Sally Thurner and Jennifer Mackenzie for help on my thesis
writing. Aimin Cao, Maximiliano Bezada, Yang He and Xin Cheng also gave useful
advice on my research. Particularly, I would acknowledge former post-doc Meghan
Miller for her collaboration on the Colorado Plateau project, and express my gratitude to
Dr. Yingjie Yang, Dr. Donald Forsyth and Dr. Aibing Li for their assistance on the
surface wave inversion code. I thank Tobias Hoink and Adam Ringler for discussion on
the results. Dr. Jan Hewitt, Dr. William Symes and Dr. Matthias Heinkenschloss in the
thesis writing class (CAAM 600) also offered plenty of great suggestions and comments.
Finally, I would thank my parents and brothers for their love and unconditional support
all the time. This research was supported by NSF EarthScope grant EAR-0844760.
i i i
TABLE OF CONTENTS
Abstract ii
Acknowledgements iii
Table of Contents iv
Chapter 1 Introduction 1
1.1 Geologic background 1
1.2 Previous seismological investigations 5
1.3 Motivation 6
1.4 Rayleigh wave tomography 9
Chapter 2 Data and Methodology 10
2.1 USArray stations and data processing 10
2.2 Methodology 15
2.2.1 Modified two-plane wave tomography 15
2.2.2 Shear velocity inversion 19
Chapter 3 Parameterization and Inversion 21
3.1 Phase velocity measurement 21
3.1.1 Parameterization 21
3.1.2 Inversion 23
3.1.3 Phase velocity maps 27
3.2 Shear velocity inversion 29
3.2.1 Parameterization and inversion 29
3.2.2 Average shear velocity structures 33
Chapter 4 Results and Discussion 36
iv
4.1 Crustal structure 36
4.1.1 Crustal heterogeneity 36
4.1.2 Interpretation 39
4.2 Upper mantle structure 42
4.2.1 Upper mantle seismic heterogeneity 43
4.2.2 Colorado Plateau mantle lithosphere 46
4.2.3 Basin and Range and CP-BR transition zones 48
4.2.4 Rio Grande Rift and Jemez lineament volcanism 49
4.2.5 Southern Rocky Mountains 50
4.3 Sublithospheric structure 50
4.4 Possible delamination 55
4.4.1 Comparison with receiver function imaging 55
4.4.2 Possible delamination mechanism 59
Chapter 5 Conclusions 63
List of References 65
v
CHAPTER 1
INTRODUCTION
This paper constructs a high-resolution shear velocity model of the lithospheric structure
beneath the Colorado Plateau and its surrounding regions. Such a seismic model is
essential for understanding how the major tectonic and magmatic activities altered the
continental lithosphere during the Cenozoic era, and how the plateau was uplifted to its
current high elevation while suffering little internal deformation. Using the dense
USArray data, I applied a modified Rayleigh wave method to build a better-resolved Vs
model to further examine the Cenozoic lithospheric modification and current complexity
of the crust/mantle boundary (Moho) beneath the plateau and to provide a high-resolution
velocity model for future geodynamic studies.
1.1 Geologic background
The "bowl-shaped" Colorado Plateau (CP) is a high standing and relatively stable block
across four states (Utah, Colorado, New Mexico and Arizona) in the southwestern United
States. As shown in Figure 1.1, this distinct province is bounded by the Uinta Mountains
and the Wasatch front in Utah in the north, the highly deformed Southern Rocky
Mountains (SRM) in the northeast, the Rio Grande Rift (RGR) valley in the east, and the
highly extended Basin and Range Province (BRP) in the south, west and southwest.
Between the Southern Basin and Range (SBR) Province and the plateau, there is a broad
transition zone of about 100 km with transitional geologic and geophysical
characteristics. The Colorado Mineral Belt (COMB) and the Jemez lineament (JL) are
1
two distinct geologic features lying at the transition regions between the CP and the SRM
and the RGR, respectively (Figure 1.1).
-118 -116" -114" -112" -110' -108" -106'' -104*
-118' -116" -114' -112" -110' -108 -106" -104'
Figure 1.1 Topography map of the major physiographic provinces of the western United States in this study, including the Colorado Plateau, the Southern and Northern Basin and Range, the Rio Grande Rift, the Southern Rocky Mountains and the Great Plains. Station distribution inside the study area (white box) from various USArray networks (TA: Transportable Array; CI: Caltech Regional Seismic; US: US National Seismic; IU: IRIS/USGS; UU: University of Utah Regional) are also shown in different symbols (see legend). The topography data is from the ETOPOl model (http://www.ngdc.noaa.gov/mgg/global/global.html; Amante and Eakins, 2009). Locations of the Colorado Mineral Belt (COMB) in Colorado and the NE trending Jemez lineament (JL) at the southeastern margin of the plateau are also indicated.
2
The modern lithospheric structures beneath the Colorado Plateau and its adjacent
provinces in the tectonically active southwestern United States are complicated mainly
due to significant Cenozoic modifications to a Paleozoic passive continental margin
outboard of a Proterozoic and Archean core (e.g. Lipman, 1992; Humphreys et al, 2003).
The northeast-striking continental lithosphere was assembled through progressive
southward accretion of island arc terranes during several Proterozoic tectonism, including
the Yavapai and Mazatzal orogeny (1.75-1.60 Ga), followed by anorogenic intrusion of
granitoids (1.48-1.32 Ga) and mafic intrusion (1.2-1.1 Ga) in the southwestern U.S.
(Karlstrom and Bo wring, 1988; Bowring and Karlstrom, 1990; Karlstrom and
Humphreys, 1998). The major lithospheric modification from the Paleozoic passive
North American continental margin to the cratonic edge (eastern limit of the North
American Cordillera) was associated with the major tectonism, magmatism and
deformation (shortening, extension etc.) occurring within the western U.S. as a
consequence of the Late Cretaceous-Early Cenozoic Farallon plate subduction beneath
the North American plate (e.g. Coney and Reynold, 1977, Humphreys et al, 2003). The
initially rapid subduction of the Farallon slab caused the basement-cored uplifts due to
crustal compression and thickening during the Laramide orogeny (~ 70-45 Ma), and also
initiated an eastward migration of the arc magmatism (Coney and Reynold, 1977;
Lipman et al, 1972; Snyder et al, 1976) in the foreland from Sierra Nevada (~ 74 Ma) to
as far as the NE trending Colorado Mineral Belt (Lipman, 1992; Miller et al, 1992)
several hundred kilometers inland from the trench. Due to the combined effects of the
oceanic plateau subduction and enhanced mantle wedge suction caused by rapid sinking,
the Farallon slab began to flatten and subsequently hydrate the base of the North
3
American mantle lithosphere (Humphreys et ah, 2003; English et ah, 2003; Humphreys,
2009) by introducing water content to the nominally anhydrous minerals and depressing
the solidus for subsequent melting (Lee, 2005; Li et ah, 2008).
Following the slab removal around 30 Ma, the Mid-Tertiary Basin and Range volcanism
("ignimbrite flare-up") and intensive extension were triggered by the combination of
post-subduction orogenic collapse and exposure of the fertile mantle lithosphere to the
hot ascending asthenosphere (Dickinson and Snyder, 1978; Humphreys, 1995). Small-
scale convection at scattered regions (Wilson et ah, 2005a, b; West et ah, 2004;
Hyndman et ah, 2005; Humphreys, 2009) and mantle asthenospheric upwelling from
deep source {e.g. Morgan et ah, 1986; Moucha et ah, 2008, 2009) further modified the
unstable southwestern U.S. lithosphere after the foundering of the slab. Magmatism and
deformation continue presently in the BRP and RGR provinces and are migrating onto
the CP margins at encroachment rates of ~ 3-6 km/Myr (Roy et ah, 2009). During the
past 30 Myr, the RGR at the southeastern edge of the plateau has undergone a two-stage
rifting process characterized by low- and high-angle normal faulting, consecutively, and
is separated by a ~ 10 Myr period of magmatic quiescence (West et ah, 2004; Gao et ah,
2004; Wilson et ah, 2005a, b). It still remains controversial in what manner these tectonic
and magmatic activities (such as plate subduction, slab flattening, post-Laramide
foundering, post-orogenic collapse, rock uplifts and incision) have altered the lithosphere
structure of the North America, especially beneath the CP and the transition zones (TZs)
between the plateau and the surrounding BRP, the RGR and the SRM (Humphreys, 1995;
Humphreys, 2009).
4
1.2 Previous seismological investigations
There have been numerous seismic studies on crustal and upper mantle structures
underlying the CP and the surrounding regions; however, they are limited to either large
continental scales with poorly sampled data, or focus on small regions or profiles. At
large scale, for example, P and S body wave traveltime tomography (e.g. Humphreys and
Dueker, 1994a, b; Burdick et al, 2008) and surface wave tomography (e.g. van der Lee
and No let, 1997) both characterized the western U.S. by pronounced low P and S velocity
anomalies in the upper mantle, from the Mendocino Triple Junction (MTJ) to the Rocky
Mountains front far in the east where the cratonic interior begins. Unfortunately, due to
uneven and sparse station coverage, these results provide little information of localized
heterogeneities on the plateau.
On the other hand, the regional studies in the past two decades investigated small scales,
though with higher lateral or vertical resolution to various degrees. For example, the
seismic reflection/refraction study from the Pacific to Arizona Crustal Experiment
(PACE) (e.g. Wolf and Cipar, 1993; Benz and McCarthy, 1994) and the Colorado
Plateau-Great Basin (CP-GB) PASSCAL experiment (e.g. Sheehan et al, 1995, 1997)
focus on the southern and northern Basin and Range-Colorado Plateau transition zones,
respectively. There are also four major seismic lines, one across the eastern Basin and
Range-western Colorado Plateau transition at 37°N by the University of Arizona (Zandt
et al, 1995), one along the N-S trending Deep Probe using active sources (Henstock et
al, 1998; Snelson et al, 1998; Gorman et al, 2002), one along the Continental Dynamics
of the Rocky Mountains line (CD-ROM) (e.g. Levander et al, 2005, Snelson et al,
5
2005), and the fourth and more recent line across the Rio Grande rifting zone between
western Texas to southeastern Utah by the Colorado Plateau/Rio Grande Rift Seismic
Transect Experiment (La RISTRA) group (West et al, 2004; Gao et al, 2004; Wilson et
al, 2005a, b). Although these active and passive seismic investigations on province
scales reveal parts of the complicated lithospheric architecture, many of which are
discordant with the tectonic features, a comprehensive understanding on how the plateau
interacts with the surrounding provinces requires an overall seismic velocity model that
will be provided in this study on this region.
1.3 Motivation
Such a full-scale crustal and upper mantle model can also be of general importance (1) to
understand the CP uplift mechanism, and (2) to reexamine the existing explanations of
the split and diminished signals of the converted phases in the receiver function studies.
First, a detailed 3D lithospheric model can provide an understanding of the timing and
uplift mechanism(s) of the CP, especially the post-Laramide uplift, which have been the
subject of long-standing debates. The present elevation of approximately 2 km was
mainly formed during the Cenozoic, since the plateau appears to have stayed at or close
to sea level until ~ 65 Ma, as evidenced by the widely distributed Late Cretaceous age
marine strata throughout the plateau {e.g. Oetking et al, 1967; Morgan and Swanberg,
1985). Numerous mechanisms have been suggested to explain the 2-km Cenozoic
rock/surface uplift of the plateau, which is roughly in Airy iso static equilibrium
(Thompson and Zoback, 1979). The buoyancy sources from the crust, lithosphere mantle
6
and asthenosphere that support the present elevation can be attributed roughly to four
major factors: thermal expansion (associated with phase change), mechanical lithospheric
thinning, crustal thickening and dynamic uplift. However, no consensus on the amount of
uplift or the timing contributed by each factor has been achieved yet. For example, based
on an early idea by Christiansen and Lipman (1972), Tompson and Zoback (1979)
suggested that the plateau uplift was caused mainly by gradual thermal warming
associated with eclogite to gabbros phase conversion, after cessation of shallow
subduction. Roy et al. (2009) further developed and quantified this idea based on a
simplified 3-D thermal conduction model, in which they assumed a transient mid-Tertiary
thermal perturbation to the thick heterogeneous and Fe-depleted lithosphere of the
plateau that triggered thermal re-equilibration, expansion and uplift, and the following
magmatism. Crustal thickening, as another significant uplift contributor, might have been
caused by Laramide-orogeny-related compression, magmatic intrusion or underplating
(Morgan and Swanberg, 1985; McKenzie, 1984). Another hypothesis is that intracrustal
transfer from an overthickened hinterland to the north (McQuarrie and Chase, 2000)
raised the CP while maintaining isostatic balance. Besides the thermal and crustal factors,
partial delamination of the mantle lithosphere has been suggested (Beghoul and
Barazangi, 1989; Spencer, 1996; Lastowka et al, 2001), as has the early complete
delamination hypothesis (Bird 1984, 1988). The latter does not appear consistent with
geophysical observations such as relatively high Pn and low heat flow beneath large parts
of the plateau (a thermal signal may not yet have reached the surface) could also produce
much of the plateau uplift while preserving petrologic and seismic signatures in the
uppermost mantle. In addition to all the passive sources to support the high elevation,
7
Moucha et al. (2008, 2009) suggested a regional dynamic topography (~ 500 m)
supported by a strong mantle flow coupled to the sinking Farallon slab. The uplift
contributions from various factors depend on two crucial depths: the Moho and the
lithosphere-asthenosphere boundary (LAB), which can be inferred from the Vs structures.
Second, the complexity and weak signals of converted phases in determination of the
crustal thickness beneath most of the plateau have been noted in many teleseismic
receiver function (RF) studies. From the common conversion point (CCP) stacked RF
profiles using several portable PASSCAL experiments and U.S. National Seismic
Network (USNSN) data, Gilbert and Sheehan (2004) suggested that the weak P-to-Sv
conversion energy in the northern Colorado Plateau might be introduced by gradational
transition from lowermost crust to uppermost mantle with small impedance contrast,
possibly caused by magmatic underplating in the lower crust of the plateau in the vicinity
of the Grand Canyon (Wolf and Cipar, 1993). Most of the converted amplitudes beneath
the northern plateau become separated positive events with weaker amplitude compared
to those of the surrounding tectonic regions. This feature was also identified by other RF
studies at different regions on the plateau using various datasets, e.g. La RISTRA data
(Wilson et al., 2005a), a combination of Source Phenomenology Experiment (SPE) and
La RISTRA array in southern plateau (Gilbert et al., 2007), the CP-GB in northwestern
plateau (Sheehan et al., 1997), the 37°N profile (Zandt et al., 1995), PACE across the
Arizona transition zone (McCarthy and Parsons, 1994; Wolf and Cipar, 1993; Benz and
McCarthy, 1994; Parsons et al., 1996) and the Deep Probe profile (Henstock et al, 1998;
Snelson et al, 1998; Gorman et al, 2002; Levander et al, 2005). The PdS and SdP RF
8
volumes of the western U.S. made by Levander and Miller (2010) from the dense
USArray/Transportable Array (TA) stations reveal a similar splitting of Moho depth
beneath the plateau with improved resolution. Instead, they interpreted this splitting
feature as a detachment of the lower crust and uppermost mantle at the western plateau,
rather than the low impedance contrast as suggested by Gilbert and Sheehan (2004).
Parsons et al. (1996) also claimed that the crust of the southwestern plateau possesses the
same crustal composition and appears as an unextended BRP to its southwest. The
surface wave tomographic result in this thesis will show why the delamination is a
preferred interpretation rather than the gradational Moho transition explanation.
1.4 Rayleigh wave tomography
In this study, to better understand the buoyancy sources for the high-elevated plateau, and
to examine the delamination hypothesis, I built a new high-resolution Vs model based on
a two-step inversion technique. In the next chapter, the details of data processing on the
records by the dense USArray/TA stations (-70 km) and the key ideas of the modified
two-plane wave inversion (Forsyth and Li, 2005; Yang and Forsyth, 2006b) and the
isotropic shear velocity inversion will be summarized. Further details on the
parameterization for both phase and shear velocity inversions and the phase velocity
maps will given in Chapter 3 before going to deep discussions on the Vs structures for the
crustal and upper mantle in Chapter 4. The final inverted Vs model based on the high-
quality data reveals detailed lithospheric structures deep to ~ 250 km, and successfully
images high velocity anomalies where the detached portion possibly occurred, and
therefore supports the delamination hypothesis by Levander and Miller (2010).
9
CHAPTER 2
DATA AND METHODOLOGY
This chapter first gives a brief summary of the selection and processing of the USArray
data to prepare the amplitude and phase data for the final 2-D phase velocity inversion.
Then I will describe the key idea of the two-step inversion method used in this study for
the final shear velocity model construction, especially the modified two-plane-wave
representation to improve the lateral resolution in the 2-D phase velocity inversion.
2.1 USArray stations and data processing
I used the phase and amplitude information of the fundamental mode Rayleigh
wavefields recorded by the EarthScope/USArray Transportable Array (TA) network and
several other seismic networks in the southwestern U.S. within the geographic boundaries
of 32° to 50° north latitude, and -116° to -105° west longitude. About 200 broadband
seismic stations in the TA network with an approximate station spacing of 70 km, and 18
stations from other networks were incorporated for analysis, including ten stations from
the Caltech Regional Seismic Network (CI), five stations from the US National Seismic
Network (US), two stations from the IRIS/USGS Network (IU), and one station from the
University of Utah Regional Network (UU), as shown in Figure 1.1.1 have used a total of
154 teleseismic earthquakes (30° < A s 120°, Figure 2.1a) with shallow focus < 70 km
and magnitude s 5.5 which occurred between June 2007 and September 2009, as the TA
network swept over the CP region. The good azimuthal distribution of events and well-
distributed stations provided extraordinarily dense raypath coverage (e.g. Figure 2.1b) to
10
(a)
m
112u -110"
Figure 2.1 (a) Azimuthal distribution of earthquakes (red dots) used as the Rayleigh wave sources centered on the Colorado Plateau (blue triangle). The neighboring concentric circles are in 30° distance, (b) Raypath (green lines) coverage (50 s) for the Colorado Plateau-Northern Basin and Range sub-region within the white box. Blue triangles represent the USArray stations.
11
better resolve the lateral heterogeneity of the phase velocities propagating across the
plateau region and its adjacent provinces.
I chose and processed the vertical components of the seismograms, which are free of
Love wave contamination in an isotropic medium, to extract the pure Rayleigh wave
signals. After resampling and correcting instrument responses to the same reference
transfer function, I applied 18 Butterworth bandpass filters (narrow bandpass width of 10
mHz, zero-phase shift and 4th order) with central periods at 20, 22, 25, 27, 30, 34, 40, 45,
50, 59, 67, 77, 87, 100, 111, 125, 143 and 167 s for the pre-selected traces (see example
in Figure 2.2). Filtered seismograms with unusual amplitudes or relatively low signal-to-
noise ratios (e.g. < 3) were excluded, and only those clean traces with high coherence
among stations ordered by epicentral distances were retained for further analysis (Figure
2.3). To isolate the fundamental mode from higher order modes and body wave phases, I
windowed the surface waves with variable-length windows with 50s half cosine tapers at
both ends (/. e. windows were several hundred seconds in length and were identical for all
receivers for the same earthquake). Fewer seismograms in short periods (e.g. less than 30
s) can be kept because of severe distortion and incoherent interference patterns among
stations due to strong, probably localized multipathing, focusing, or defocusing effects.
The effects of frequency-dependent anelastic attenuation and geometrical spreading
(Mitchell, 1995; Li et al, 2003) on the Earth's surface were taken into account to provide
a minor correction of the amplitudes after measuring the phases and amplitudes by
Fourier transform. Local site responses (Yang and Forsyth, 2006a) were also considered
to account for individual station corrections of the amplitude amplification and phase
12
20 s
. ' 22 s
A 25 s
f • """"
*l 27 s "I • tt 30s •i
j — —
V / p * l ^ ^ ^
•" • 'Hl i l l ' "
. .* i A i» l» ^ * -
• • j j i ' - ' I •
• ' , « * ;
.«• 34s
-r = • J 40 s tv
t .» 45 s
.» 50 s H '4 ' • f 59 s
67 s
i 77 s »
f 87 s
• - - • • • - - - • _ • • * v ? • - • - • • •
". *
A ' • > " ,
" V • I
—--A •'"',•' • / V / " * 1 , > • * > , — i — • —
> . / \ • ; • '• .' '", / ' " - — | 100 s
j . ,.
\ » * 125s
r
* \ - x
I 143 s >
» 167s "i .,
10 20 x tC-2
... i . . .
22 23 24 2S
. • * >
... . ., i . i .*".„ . . V i ,. . . , i y . . . . . . , . . . . . . . . .J . ~ -26 2? 23 2>> 30 31 32 33 34 35 36 37 3&
..I. . 'A 39 «0
Time (s)
Figure 2.2 An example of fundamental-mode Rayleigh waveform at different frequency bands recorded at the M23A station in the USArray/TA network for an earthquake (Magnitude 6.4, depth 35 km, epicentral distance 98.3 degree) on February 17, 2009. The vertical component of the raw data is shown at top and was filtered into 18 frequency bands with central periods ranging from 20 s to 167 s. The delay of arrival time at longer periods (> 50 s) indicates a low velocity zone in upper mantle.
13
02/17/2009 (0.015-0.025 Hz) Depth = 35 km
X1E+01
220
Time (sec) 240 260 280 300 320
oo 00 = _
o
J-..-.-. 220
i 240
...J 260 280 300
L 320
Figure 2.3 Filtered waveform (50 s) at the USArray stations within the study region ordered by epicentral distance for the same individual earthquake used in Figure 2.2.
14
shift that correlates with the localized tectonic features. Earthquake magnitudes were
normalized by scaling the observed amplitudes to the R.M.S. amplitude of all the stations
for each event.
2.2 Methodology
2.2.1 Modified two-plane wave tomography
I have used a two-stage inversion approach developed by Yang and Forsyth (2006b) to
determine the Vs structure beneath the study region from the Rayleigh phase and
amplitude data. In the first step to estimate the 2-D phase velocities, I adopted the
modified two-plane-wave tomography (TPWT) technique (Yang and Forsyth, 2006a, b;
Forsyth and Li, 2005) to effectively take account of interference of energy resulting from
relatively simple multipathing and scattering effects in the incoming Rayleigh wavefields
that will be poorly represented by the traditional one-plane wave approximation. This
TPWT method has been applied successfully to a large amount of datasets to estimate the
phase velocity maps (Forsyth et al, 1998; Forsyth and Li, 2005; Li et al, 2003, 2005;
Yang and Forsyth, 2006a; Yang and Ritzwoller, 2008; Weeraratne et al, 2003; Schutt et
al, 2008; Miller et al, 2009).
In this method, the complexity and distortion in the fundamental mode surface waves in
each individual frequency band are simply considered as the sum of two interfering plane
waves with different amplitudes, initial phases and propagation directions (deviation
angles from the great circle path). The vertical component at the k-th station for the i-th
15
incident event at each frequency co can thus be written in the form of (Forsyth and Li,
2005)
2
]uz{x) = 2 , V<°>>exp(-,"''M0'>) t2-1) ipw=\
with
%PA^ + "(rf " T?) + [r*cos(tf- ,0^) - *f }kiSavgco (ipw = 1, 2)
in which jAipw(co), ,.0/>w, and °0. are the amplitude, derivation angle from the great
circle path, and the phase at the reference station for the primary and secondary plane
waves, x\ and T° are the traveltime of the k-th station and the reference station along the
great circle path with respect to the edge of the study area, respectively, *5 is the
average slowness as a function of the weighted phase velocity coefficients, and r\, y\,
x\ are the local coordinates with respect to the reference station, based on a regional
flattening coordinate system.
In addition, I also followed Yang and Forsyth's (2006b) consideration on the finite-
frequency effect of the surface wave propagation caused by the off-azimuth structures by
introducing amplitude and phase sensitivity kernels during the inversion for the phase
velocity maps at different frequency bands. This approach is effective to further improve
the lateral resolution on the local small-scale structures, especially the heterogeneities at
wavelength scale. Simplified from the 3-D surface wave kernels (Zhou et ah, 2004) based
on the single-scattering approximation (Born approximation), the 2-D phase and
amplitude sensitivity kernels K^(r,co) and KcA(r,a>) for the incoming plane wave
propagation at each single frequency, which are defined as
16
( ** \ rr(KcAr,co)^ dlnA i-Jj \K'A{r,m)j
— dx2, c
can be expressed as (Yang and Forsyth, 2006b):
K;(r,co) = M k2R" • exp[-i(kx" -kAx + JC/4)])
/T(r,a>) = -Re
R^lnkx"
k2R"-exp[-i(kx"-kAx + Jt/4)]h
R-\l2jrkx" J
(2.2)
(2.3)
in which 8c Ic, 6(f) and dIn A are the phase velocity perturbation, phase delay and
amplitude perturbation, respectively; k is the wavenumber of the incident Rayleigh
wave; R" and R are the station polarization vectors for scattered and incoming waves,
respectively; x" denotes the scatter-receiver distance, and Ax denotes the propagation
distance difference between the incoming waves that arrives at the receiver and scatterer.
Figure 2.4 shows the computed sensitivity kernels at 20, 50 and 100 s assuming a single
plane Rayleigh wave incident from left and was recorded in the center of the region. The
broadening and weakening of the sensitivity peaks are well noted at the bottom of Figure
2.4, as the period increases from 20 to 100 s. The introduction of the 2-D finite frequency
kernels were demonstrated to be able to improve the lateral resolution of phase velocity
in regional surface wave tomography by synthetic tests (Yang and Forsyth, 2006b).
There are two main steps involved to invert the phase velocity anomalies as well as the
wave parameters: (1) I inverted the average phase velocity and wave parameters; (2) I
performed a 2-D inversion to obtain the wave parameters and the velocity coefficients
after placing the updated mean phase velocity as the initial velocity at each grid node. In
both steps, I first used the simulated annealing technique to find the best fitting wave
17
-1600
-1000
-500
0
500
1000
1500
-1500
-1000
-500
0
500
1000
1500
-1500
-1000
-500
0
500
1000
1500
Amplitude Kernel (20 s} j
i _ _ 1
I 1
X10"
-1000 0 km
1000
Amplitude Kernel (50 s}
-1000 0 km
1000
Amplitude Kernel (100 s)
-1000 0 km
1000
x 10'
4
2
0
-2
-4
-6
E
-1S00
-1000
-500
0
500
1000
1500
Phase Kerftel (20 s]
210"
1
0 i
-1-1
-2
x10"
-10
-1500-1000 1000 1500 -10 -1500-1000 1000 1500
Figure 2.4 Map views of 2-D amplitude (left) and phase (right) sensitivity kernels on the Earth's surface for left-incident Rayleigh plane waves at 20, 50 and 100 s. The bottom panel shows the cross-section profiles at x=-500 km for these three periods, showing that the sensitivity peaks become broad and weak as the period increases.
18
parameters based on a fixed velocity model for each event individually, and then used a
generalized linearized inversion technique (Tarantola and Valette, 1982) to invert the
wave parameters and phase velocity coefficients, simultaneously. The solution to the
generalized linear inversion technique shows that the change of model Am to the starting
model mo can be given by (Forsyth and Li, 2005; Tarantola and Valette, 1982)
Am = (GTC-;nG + C"l r 1 [GTC-;n Ad - C l (m - m0)] (2.4)
in which m is the current model vector, Ad is the difference vector of the predicted and
observed data, G is the partial derivative matrix, Cmm and Cnn are the a priori
covariance matrices for model and data vectors, respectively. Under this inversion
scheme, the perturbation to the current velocity model and wavefield parameters will be
iteratively updated, until the numerical criterion for the misfit was met or the maximum
number of iterations was reached. The standard deviations of the phase residuals for each
event were estimated after each iteration and were divided to down-weight the noisy
events that were poorly fit by the two-plane wave approximation. In practice, I need no
more than 10 iterations to obtain the final phase velocity variations as no further
significant changes occurred.
2.2.2 Shear velocity inversion
In the second step, I inverted the regionalized shear velocity structures from the localized
dispersion curves (20-167 s) using the DISPER80 package (Saito, 1988) which was
further modified by Forsyth et al.. The forward model shows that the dispersion curves
can be computed through analogy with the mathematically equivalent solutions to the
spheroidal-mode free oscillations at the short wavelength limit, for a 1-D model
consisting of a number of transversely isotropic layers with homogeneous density and
19
elasticity (Takeuchi and Saito, 1972). The DISPER80 package takes into account the
sphericity of the Earth and neglects the self-gravitating effect which is much smaller at
periods > 20 s (Takeuchi and Saito, 1972). A fourth-order Runge-Kutta-Gill (RKG)
scheme was implemented to integrate the motion-stress differential equations upward to
the free surface to satisfy the boundary condition of vanishing stress. Meanwhile, a
compound-matrix method was also introduced to overcome the numerical instability in
searching for the root of the characteristic equation (Saito, 1988; Aki and Richards,
2002). Knowing the forward modeling and starting from a 1-D shear velocity model, I
used a similar iterative and linearized solution given by Equation 2.4 to the nonlinear
least-square regional 1-D shear velocity inversion to construct a pseudo 3-D crustal and
upper mantle shear velocity model beneath the study area. The crustal thickness
constraints are also placed from the PdS receiver function estimates (Levander and
Miller, 2010).
20
CHAPTER 3
PARAMETERIZATION AND INVERSION
Based on the modified two-plane wave tomography technique summarized in the last
chapter, I implemented this method separately to the four subdivisions in my study focus.
For all the eighteen frequency bands, the phase velocities within each subdivision will be
jointed together by averaging overlaps to construct overall phase velocity maps and to
obtain the dispersion curves on each grid node. The following regional 1-D shear velocity
inversions carried on each node lead to the final 3-D Vs model with remarkable vertical
resolution deep to 250 km beneath the CP and the adjacent regions.
3.1 Phase velocity measurement
3.1.1 Parameterization
The phase velocity lateral variations at each frequency can be highly resolved from the
input phase and amplitude data using the modified two-plane-wave approximation. First,
to satisfy the basic assumption of the two-plane wave approximation that the study region
should be small enough so that the recorded surface waveforms can be simply
represented by interference of two plane waves, I divided the entire area by the
perpendicular bisectors at the longitude and latitude axes into four sub-regions (Figure
1.1): (1) Colorado Plateau-Northern Basin and Range Province (CP-NBR); (2) Colorado
Plateau-Southern Rocky Mountains (CP-SRM); (3) Colorado Plateau-Southern Basin and
Range Province (CP-SBR); and (4) Colorado Plateau-Rio Grande Rift (CP-RGR). This
approach also happened to take good advantage of the dynamic deployment status of the
21
USArray/TA station coverage (the majority of the transportable stations lies on the left
side before January 2008, and very few stations remain in the Northern Basin and Range
after December 2008). The phase velocities in the overlapping regions were averaged
after application of the two-plane wave method in each subdivision separately. Different
events were chosen inside each 'box' to ensure good station coverage and to maintain
similar azimuthal distribution to resolve lateral phase velocity variations (Figure 3.1a, b).
Second, as mentioned in last chapter, I incorporated the 2-D surface wave phase and
amplitude sensitivity kernels, instead of the 'fat ray' approximation with Gaussian-
shaped sensitivity kernels (Forsyth and Li, 2005; Schutt et al, 2008) using same width of
influence zone along the great circle ray path. For all the frequency bands, a Gaussian
averaging function with a smoothing length Lw of 65 km was used in the traveltime
calculation from the edge to the receivers, and it also acted as a 2-D filter with the same
characteristic length to smooth the Born-approximation sensitivity kernels. A checkboard
resolution test for this modified two-plane wave technique has also been done using the
same event distribution and station coverage.
The phase velocity variations were inverted separately at each sub-region. Since the
vertical components of the seismogram are considered as the sum of two independent
interfering plane waves (each with unknown amplitude, initial phase and propagation
angles off the great circle path), there will be six plane wave parameters to represent each
source event. Each 'box' is parameterized on an evenly distributed grid with a spacing of
0.5° in the receiver-covered region, and double-spacing nodes on the edges to absorb the
22
traveltime residuals (by assigning the boundary nodes 10 times of the a priori standard
deviations of the inner nodes) for the waves that are beyond the simple two-plane-wave
representation. Thus, there will be 6M1 + M2 model parameters, where Mi is the total
number of events and M2 is the number of phase velocity coefficients at each grid node.
There are 2N Rayleigh wave observations in which N stands for the total number of
retained traces and the factor 2 comes from the real and imaginary components of the
complex amplitudes in the frequency domain. The starting average phase velocities were
estimated from the synthetic dispersion curve based on the 1-D TNA model (Grand and
Helmberger, 1984). The two main steps, as described in the method section in the last
chapter, were then implemented to invert the 6M1 wave parameters and M2 phase
velocity coefficients simultaneously using both the simulated annealing technique and the
standard linearized iterative inversion technique.
3.1.2 Inversion
The phase velocity variations from 20 to 167 s in each geological 'box' were derived
from the TPWT method described above. The number of earthquakes and raypaths
(Figure 3.1a, b) used in the Rayleigh wave tomography is much smaller at short periods
(< 25 s) due to severe distortion and strong multipathing in the surface waves, and also
reduces slightly due to the decreasing energy of signals at longer periods (> 100 s). The
mean phase RMS misfits (Figure 3. Id) for all of the events stay below 3 s, indicative of a
good fitting except for those at the longest periods {e.g. 143 and 167 s) that have larger
uncertainties of phase measurements. To examine the validity of the two-plane wave
approximation used in the inversion, I checked the distribution of the plane wave
23
150
^00
50
0
6000
6000
4000
3000
2000
1000
0
ICP-NBR
(a) Number of Earthquakes
' .."."..] CP-SRM i i CP-SBR I ICP-RGRI
1 ;• I 20 22 25 2? 30 34 40 46 60 69 67 77 87 100 11<! 125 143 167
(b) Number of Raypaths
I CP-NBR ~~~~~! CP-SRM i ^ CP-SBR I JP-RGR
20 22 25 27 30 34 40 45 50 59 67 77 87 100 11 ' 125 143 167
75
70
65
60
55
60
-
(c) Rank/(number of model parameters) (%)
f CP-NBR CP-SRM CP-SBR CP-RGR| — —|
•j
20 22 2b 27 30 34 40 45 50 59 67 77 87 100 11* 125 143 167
(d) RMS phase misfit (s)
I CP-NBR CP-SRM CP-SBR CP-RGRi i
20 22 25 27 30 34 40 45 50 59 67 77 87 100 114! 125 143 167
Period (s)
Figure 3.1 (a) Number of Rayleigh wave sources, (b) total number of raypaths, (c) the ratio of rank to the total number of model parameters, and (d) the phase RMS misfit in second for the phase velocity inversion at the 18 frequency bands at the four separate sub-regions: CP-NBR, CP-SRM, CP-SBR and CP-RGR.
24
42" 42"
40°
1+ + + +
T + +
• + +
+ + , 38"
+ + +
-116' -114" -112'J
+ + +
+ + +
^ +
+ +
+ +
40"
4
+
+ + + *• +
+ +
+
m 1.C
0.8
38' •
-1K
42" •
-114'
+
+ +
-112"
m 3
Q-
0.6 J
0.4
40"
38"
X + * + + m lit- + + + + +^W| I
* + + + + * + '
l + + + + + I + + m. + * + +
1 0.8
1 0.4
1 rj 1 0.0 g
-0.4
• -0.8
-116" -114' -112"
a • o 3
Q. C
<
S O
X3
2 £
1.0 f
0.8 •
0.6 •
0.4 j
0.2 i
0.0 <•-• 0.0 0.2
n"
0.4 0.6
Observed Am
0.8
piitude
•
ji
,! *
1.0
I
"\ »; ; v • • '
• - '
•••
• /
J/
-0.8 -0.4 0.0
Observed
0,4
Phase
i;
0.8
0.8
0.4
0.0
-0.4
-0.8
US
.c 0.
"8 *** o T3 as
CL
Figure 3.2 Comparison of predicted and measured amplitude and phase data at 50 s in the CP-NBR sub-region. The map views of the amplitude (top) and phase (middle) show a good agreement between the observation and the prediction by the modified two-plane wave inversion technique, especially for the phase data. The bottom panel presents the same agreement in another way.
25
(a) Amplitude ratio
50 r
-°50
0.1 0.2 0.3 0.4 0.5 0.6 Amplitude ratio
(b) Primary wave
0.7 0.8 0.9
-15-12-9 - 6 - 3 0 3 6 9 12 15 Deviation angle from great circle path (degree)
(c) Secondary wave
-40 -30 -20 -10 0 10 20 30 40 Deviation angle from great circle path (degree)
Figure 3.3 Statistics of the inversion results of the two-plane wave approximation at 0.02 Hz (50 s). (a) Amplitude ratios of the smaller plane waves to the larger one. (b) Deviation angles of the primary plane waves from the great circle direction, (c) Deviation angles of the secondary plane waves from the great circle path. These angles are much larger than those of the primary waves.
26
amplitude ratio (secondary/primary amplitude), the range of deviation angles from the
great circle, and how well the predicted phase and amplitude fit with the observed data.
One example at 50 s of the inversion for one single earthquake (September 30, 2007) in
Figure 3.2 shows that the predicted phase and amplitude data fit pretty well with
observed data. Additionally, for most of events at the same period the amplitude ratio are
smaller than 0.4 (Figure 3.3a), and the deviation angles of the primary waves mainly fall
into ±10 degrees (Figure 3.3b, c), both of which suggest that the two-plane wave
representation is a reasonable approximation for the incoming wavefields at most of the
relevant frequency bands.
3.1.3 Phase velocity maps
The Rayleigh wave tomographic results of the phase velocity variations from 20 to 167 s
are plotted in Figure 3.4. From the phase velocity maps, several consistent velocity
anomalies associated with relevant geological features are identified: (1) a profound low
velocity area in the SRM at short periods (< 50 s) associated with a relatively low crustal
shear velocity beneath the Colorado Rockies; (2) at short periods less than 30 s, a large
low velocity area covering the COMB to the east of the CP, and relatively fast phase
velocities beneath the SBR which abruptly dropped once entering the CP-SBR transition
zone (TZ); (3) two long NE trending low velocity belts along the CP-NBR and the Jemez
lineament in the CP-RGR transition regions; (4) consistent relatively high phase
velocities in the CP interior compared to the surrounding tectonic margins; (5) when
periods are large than 45 s, the high velocity within the plateau is connected with the high
phase velocity anomaly to the south of the Wyoming Province, and this high velocity
27
34
32 K P L "t] iP9!3 L « i.2 125SJ ^ _1_67sJ 116 **4 112 MO 1CU *06 -116 -114 112 110 128 106 116 M 4 112 MO 108-106
6c/c(%) -6 0 2
Figure 3.4 Rayleigh wave phase velocity maps at 12 of the 18 periods. The phase velocity inversions were performed separately on the four 0.5° x 0.5° grid sub-regions, created by cutting along two geographic lines at 110.5°W and 37°N, and were then combined by averaging the overlaps.
28
anomaly feature exists consistently to longer periods, although it seems to be encroached
from the western, southwestern, southern and southeastern edges from the BRP and the
RGR; (6) the high anomaly beneath the SBR at short periods (e.g. 20, 25 and 30 s in
Figure 3.4) is associated with the relatively shallow Moho depths, since phase velocities
at the same period sense the deeper upper mantle velocity structures in the SBR while in
the other regions they mainly sense the crustal structures.
3.2 Shear velocity inversion
The regionalized isotropic crustal and upper mantle shear velocity structures beneath the
CP and the surrounding tectonic regions were inverted from the local Rayleigh wave
dispersion curves derived above from the two-plane-wave inversion at each grid node
(0.5° spacing). I performed the same standard iterative linearized inversion to solve the
nonlinear least-square problem by estimating the perturbation of the shear velocity
relative to the starting reference model. The Moho topography from teleseismic PdS
receiver functions (Levander and Miller, 2010) also put initial constraints on the Vs
inversion scheme to better resolve the lithospheric structure. A 3-D shear velocity model
was immediately constructed by unifying the grid-based 1-D Vs structures.
3.2.1 Parameterization and inversion
The Rayleigh wave phase velocities are primarily sensitive to the absolute shear velocity,
especially below the Moho, rather than the density and compressional velocity. The
sensitivity kernel profiles for S, P wave velocities and density at various periods (Figure
29
(a) Vs kernel
50
*oc
2QC
JZ 25"
I 30C
35E
4QS
"S
Vp kernel
'CO
203
I £ 253
! 303
353
- 43a | " _ 67a !
13Cs s — 1G?s ;
3.C- C C2
dc/dVs C.32
403
453
r t r - i *-
20s 40s G7s 103s 157s
Density kernel
! 3 0 '
1E0?
2MU
I ! £ 2C0J
I i 3C0>
350;
0.C 3.02 C.02
dc/dVp -C.C2
V
i ]
450 '•
20s
40s G7s
I 'OCs i " "" -07a
:C2
dc/dp
(b) fret- surface
t t : . ") , U
f t . j i . P
/()
IiiU'jirite Upward*
I
\ / /
\
\ ,
Figure 3.5 (a) Sensitivity kernel functions for S-wave, P-wave and density respectively at periods of 20, 40, 67, 100 and 167 s, based on the average shear velocity model, (b) The spherical Earth model with homogeneous Vp, Vs and density in each layer.
30
3.5a) were computed based on the reference inversion model, showing that the strongest
sensitivity comes from the Vs profile, while the P-wave velocity and density have strong
sensitivity mostly to structures within the crust. Thus, I kept density and Vp/Vs ratio fixed
in each layer (Figure 3.5b), and used a combined P and S-wave sensitivity to calculate the
perturbation to the previous velocity model. In practice, no significant change occurred
with fixed shear velocity when slightly varying the ratio of perturbation of Vp with
respect to that of Vs. The longest period (167 s) in the phase velocity measurements
indicates that the upper mantle velocity structure can be well resolved to approximately
250 km in depth, though with much larger vertical uncertainties.
I inverted iteratively for the regionalized shear velocity on a 0.5°x0.5° grid to best fit the
Rayleigh phase velocity observations, and thus built the final 3-D Vs model. The shear
velocity structures within the study area were parameterized as several 10km layers in the
crust and mantle lithosphere, and as layers of increasing thicknesses (20, 30 and 50 km)
in upper mantle. I used a modified TNA model (Grand and Helmberger, 1984), a well-
known shear velocity model for the tectonic North America, as the starting model for
shear velocity inversion. The average Vs structure was first inverted by taking the mean
crustal thickness of 42 km. Subsequently, the local S-velocity structures beneath each
grid point were inverted from the initial TNA model with crustal thickness constraints
from the RF study, in order to maintain typical crustal and uppermost mantle velocity
values above and below the Moho in the starting model. The linearized inversion scheme
used combined criteria of minimum length and smoothing regularization by introducing
nonzero diagonal and off-diagonal terms in the a priori model covariance matrix Cmm.
31
The choice of a damping factor of 0.2 suggested by the obvious turning point in the In
curve of the velocity model norm and residues of phase velocity (Figure 3.6a), also
agrees well with the choice of Schutt et al. (2008) by bounding the minimum shear
velocity in the upper mantle to a reasonable range.
(a) L curve (b) Rows of resolution matrix
0.035
50
100
150
- * 200
JZ 250 QL T3
300
350
400
450
-0.05 0 0.05 0.1 0.15 0.2 0,25
Residue
500 <— -0.2
15 km 55 km 105 km 155 km 205 km 310 km 425 km
0 0.2 0.4
resolution 0.6
Figure 3.6 (a) Damping parameter estimation from the L-curve of the Vs model norm and residues of the phase velocity prediction. The turning point with damping parameter of 0.2 is chosen for the Vs inversion, (b) Model resolution kernels for layers at depths of 15, 55, 105, 155, 205, 310, 425 km. Rows of the resolution matrix are calculated from the reference model.
32
3.2.2 Average shear velocity structures
The average lithospheric shear velocities beneath the entire study region as well as the
distinct tectonic regions were also inverted, as shown in Figure 3.7. I can observe
pronounced low velocity zones (LVZ), compared with the global compilation of the
continental and orogenic averages, with strong or moderate negative gradients underlying
the rigid lithospheric lid beneath the SBR, the NBR, the RGR and the CP. These
characteristics have also been noted in many continent-scale seismic studies (Humphreys
and Dueker, 1994a, b; Burdick et al, 2008; Grand, 1994; van der Lee and Nolet, 1997).
The inversion of Vs structures can be evaluated from the rank of the velocity parameters,
which represents the number of independent pieces of information in the model provided
by the dispersion data and the resolution kernels at different depths. The rank of the Vs
model has the largest value of 4.00 at the Colorado Rockies, the smallest of 3.24 in RGR.
The plateau has an intermediate rank of 3.84 between the SBR (3.62) and the NBR
(3.86). The rank for the rift area is lower than other regions due to larger uncertainties in
the phase velocities. In addition, in order to reveal the depth-dependent layer information
resolved by the observed dispersion relation, the resolution lengths for the reference
model were plotted in Figure 3.6 (b), with the resolution peak displayed around 20 km for
the row of the resolution matrix corresponding to 15 km. The number of adjacent layers
(10 km thick for each above 300 km) needed to recover one entire piece of independent
information of the model increases with the depth, because the resolution kernels become
broader in deeper structure, and more layers are needed to produce a summed resolution
component of 1.0. Thus, phase velocities at longer periods, which have deeper but
33
Dispersion curve Dispersion curve
(a)
•
average
50 100 Period (s)
150
TNA model j Average Vs I
3.4 3.6 3.8 4 4.2 4.4 4.6 4.8 Vs (km/s)
4.4
_ 4.2 45 E * 4 * o •55 3.8
| 3.6 0.
3,4
(c)
..*" ^^^*^ .
/ ' '
-
/ s<
Or | NBR | SBR i RGR i SRM |
E
C
D
0
20
40
60
80
100
120
140
160
160
200
220
240
260
280
300
50 <00 Period (s)
150
(d)
67.5 72.S
82.S 97.S
97.S
TNA modef CP NBR SBR RGR SRM
3.2 3.4 3.6 3.8 4 4.2 4.4 4.6 4.8 Vs (km/s)
Figure 3.7 Dispersion curves fitting and Vs inversion, (a) The predicted dispersion curves for the average (blue) phase velocities, (b) The average Vs (blue line) inverted from the starting modified TNA model (black) to fit the dispersion curve in (a), (c) The dispersion curves for five tectonic regions (CP, NBR, SBR, RGR and SRM). (d) The average Vs models beneath the corresponding geologic regions. The colors are the same as (c), and the black line is the initial TNA model. All error bars stands for the standard deviation of phase velocity measurements. Estimated LAB depths from the Rayleigh wave tomography at the center of the negative velocity gradients.
34
broader peaks of Vs sensitivity (Figure 3.5a), are less able to resolve deeper upper mantle
structures. The combined analysis from the Vs sensitivity and resolution kernels indicate
that the S velocity can be resolved well to a depth of- 250 km, above which vertical and
lateral seismic heterogeneities have been well imaged. The average depths of lithosphere-
asthenosphere boundary (LAB) beneath each tectonic province is inferred at the middle
depth of the low-velocity gradient, and the estimated depths show great consistency (see
Table 1) with the independent PdS and SdP receiver function studies (Levander and
Miller, 2010).
Table 1 Comparison of LAB depths (km) estimated from the PdS, SdP receiver function imaging and surface wave tomography (SWT)
NBR
SBR
RM
CP
RGR
Western U.S.
PRF
71.5±9.1
82.1±7.5
81.7 ±9.3
89.1±11.5
79.8±11.7
SRF
86.4±16.0
89.4±16.4
122.8±18.7
101.7±14.7
99.8±23.8
SWT
72.5
82.5
97.5 (SRM)
97.5
67.5
77.5 (average)
Note: The LAB depth in the last row for the SWT is averaged on the area in this study.
35
CHAPTER 4
RESULTS AND DISCUSSION
The final shear velocity model from the Rayleigh wave tomography reveals complex
lateral heterogeneity of the crustal and upper mantle structures related to the regional
tectonic features. After description and interpretation of the major low and high velocity
anomalies above 250 km, I will briefly discuss the geodynamic interactions between the
lithosphere and underlying asthenosphere based on the well-imaged low-velocity
channels. Last but most importantly, the comparison between this Rayleigh wave
tomographic result with the PdS receiver function volumes shows that the ongoing
delamination beneath the western portion of the CP could be a significant contributor to
the complexity of the crust/mantle boundaries and a cause of the recent CP uplift.
4.1 Crustal structure
4.1.1 Crustal heterogeneity
The crustal structure beneath the CP and the surrounding tectonic regions is characterized
by relatively short-wavelength high and low velocities (Figure 4.2 a, b). Due to lack of
short period surface wave data (< 20 s) and without consideration of the sedimentary
layer information, the upper crustal structure is relatively poorly resolved. However,
several distinct low velocity features can still be identified in the first several layers of the
Vs model with reasonable resolution, including three narrow transition zones (~ 50-100
km in width) along the southwestern, northwestern and southeastern margins of the
plateau, one belt (COMB) beneath the SRM (with the strongest anomalies under the San
36
(a) Surface heat flow -11S: -1*4 : -1*2 •10: -rt»8: -106:
4Z
3S"
32;
150
12b
1C0
E
E J,J
2b
L-J 0
-116: -1'4= -1*2" - '10 : -^08 : -106:
(b) Cenozoic magmattsm i8 : -1CKT
,—i 4500
i
3030
2500 "g"
2000 c
.£ 1500 a
1000 *3
509
0
-500
-1C00 -11G" -1*4" -V.'l -*1
Figure 4.1 (a) Surface heat flow maps (data from www.heatflow.und.edu) after interpolation. Several unusually high heat flow data are excluded, (b) Distribution of the Cenozoic volcanic rocks showing the magmatic pattern in this study area (data from NAVDAT: Western North American Volcanic and Intrusive Rock Database). Different colors indicate different rock ages (red: Tertiary, black: Quaternary), and different symbols indicate different types (circle: felsic; diamond: mafic; inverse triangle: intermediate). Topography from ETOPOl model is also plotted.
37
L JSBH 90kmJ
-'1G'-*'•»'-'" 2'-1'0'-''36'-T Co b / _ SfTOftto 90 kmj
r e _"4'-"!2' -"O'-'CE'-noc-' j^ff _Cru«tal_thlcknesirj ,„
.-•e _--i-_r2-*-0"-'cs-'C5'
33 4Z 50 SO d l n V s (%)
Figure 4.2 Maps of Vs anomalies (a-i) at depths of 15, 30, 60, 80, 100, 125, 150, 200 and 250 km, relative to TNA model, and comparison of Vs maps at 90 km from this study (j) and the S body wave tomography (k; Schmandt and Humphreys, 2010). The crustal thickness (1) is estimated from the PdS receiver functions (Levander and Miller, 2010). Two volcanic fields are marked: the San Juan volcanic field (SJVF) and the Navajo volcanic field (NVF). SMB in fig. a, b stands for the Socorro magma body. The red dot line shows the station distribution of the La RISTRA experiment.
38
Juan volcanic field and Aspen in central Colorado, see Figure 4.2 a, b), and one small
anomaly in the central RGR. The large-scale low crustal Vs anomalies beneath the CP-
BR Utah transition zone and the Colorado Rockies are broader than those in the CP-BR
Arizona transition zone and the NE trending Jemez lineament in the RGR-CP transition
region. The relatively high velocities under the SBR and the NBR are separated by a low
velocity zone located mostly in southern Nevada between 37°N and 40°N. The high NBR
velocities dropped abruptly to the east in north-central Utah near the Wasatch Front.
The plateau interior is relatively complex and has an intermediate average velocity
between the RGR and the stable Great Plains to the east of the RGR and SRM. The NE
trending high velocity anomaly in the southern plateau is bounded by the Jemez
lineament to the southeast and the Yavapai/Southern Yavapai accretionary boundary to
the northwest. It extends from the southeastern Navajo volcanic field (NVF) region to the
western edge of the SRM, and drops abruptly across the SRM-CP boundary with a very
sharp velocity gradient into the COMB. The relatively low velocity area to the west of the
northwestern NVF is consistent with the result of La RISTRA group (West et al, 2004;
Gao era/., 2004).
4.1.2 Interpretation
The seismic complexity within the relatively cold and stable plateau, especially the lower
crustal structure beneath the western plateau, is possibly associated with dynamic crustal
structure reset by a possible delamination beneath this region (see section 4.4). The high
velocities in the eastern part of the Navajo volcanic field could be related now cool mafic
39
intrusions (West et ah, 2004), while the relatively low velocities could result from the
heating from the ascending asthenosphere at the base of the crust under the western
plateau (west of the Navajo Volcanic Field).
The two remarkable low velocity anomalies under the COMB and the Jemez lineament
are consistent with previous tomographic images (e.g. Humphreys and Dueker, 1994b;
Dueker et ah, 2001; Humphreys et ah, 2003) and the refraction/wide-angle reflection
seismic results such as the Deep Probe line (Henstock et ah, 1998; Snelson et ah, 1998;
Gorman et ah, 2002) and the CD-ROM transect (e.g. Levander et ah, 2005, Snelson et
ah, 2005). The low crustal velocity feature under the SRM can be attributed to the results
of one of the most active Laramide-age magmatic activities (Mutschler et ah, 1987) in the
COMB region and the more recent (-37 Ma) large-volume San Juan volcanism. Both
volcanic events have caused compositional modification and produced partial melting of
the crust by heating the base of the COMB crust (Dueker et ah, 2001; Humphreys et ah,
2003). The COMB abutting the SRM-CP boundary is also the site of a large heat flow
contrast. The other crustal low velocity anomaly occurs along the Jemez volcanic
lineament (including the low anomaly under the Mount Taylor with complex sill and dike
structures) and corresponds to the suture between two Proterozoic island arc terranes: the
Yavapai-Mazatzal terrane to the north and the Mazatzal province to the south (Karlstrom
and Humphreys, 1998). The low velocities under the Jemez lineament match well with a
variety of geochemical and magmatic signatures that define the Jemez volcanic trend
(McMillan et ah, 2000). The low crustal velocities coincide with the high heat flow and
recent basaltic magma intrusion into the lower crust (Karlstrom and Humphreys, 1998;
40
Levander et al, 2005), as well as the low Bouguer gravity anomaly (Roy et al, 2005),
suggesting that in places the crust beneath the Jemez lineament is partially molten
(Hammond and Humphreys, 2000; Levander et al., 2005). The lack of a low velocity
signature directly beneath the RGR axis is consistent with recent magmatic quiescence
directly beneath the rift. The lithospheric thinning and rifting in the RGR as well as the
mechanical weakness caused by the plateau rotation could also contribute to reduce the
velocity beneath the Jemez lineament. One of the observed low velocity anomalies in the
central RGR is identified as the Socorro magma body (a mid-crustal sill structure) in New
Mexico (e.g. Balch et al, 1997), which is consistent with the observations by the
RISTRA group (West et al, 2004) and the tomographic result based on combined
ambient noise and surface wave data (Shen et al, 2009).
The Late Tertiary and Quaternary magmatism could have caused partial melt and reduced
the shear velocities along both the Utah and Arizona transition zones between the CP and
the Basin and Range (Wolf and Cipar, 1993; Benz and McCarthy, 1994; Parsons and
McCarthy, 1995; Parsons et al, 1996; Frassetto et al, 2006). Outside the transition
zones, the relatively high crustal velocities in the SBR and the NBR arise because of the
thinner crust in these regions (Figure 4.2 1, Levander and Miller, 2010; Frassetto et al,
2006)): The surface waves are sampling the lower crust and upper mantle compared to
middle crust of the CP interior at the same depth. In particular, in Figure 4.1b, the
strongest high velocity anomaly under the SBR at 30 km is larger than those elsewhere,
because the velocities at that depth represent both lower crust and upper mantle
41
lithologies. The low velocity zone between the SBR and the NBR, with distinct
nonbasaltic magmatism and extension to the north and to the south respectively,
coincides with the "amagmatic zone" close to the Utah-Arizona-Nevada tristate area (e.g.
Axen et al, 1993; Zandt et al, 1995). In a similar interpretation of the velocity difference
between the Basin and Range and the plateau interior, the seismic velocity reduction can
be attributed to thickened crust (-30-40 km, Levander and Miller, 2010; Zandt et al,
1995) in this region that have experienced more recent extension (<15 Ma, 60-100%)
than the Early and Late Tertiary intraplate extension (100-200%) in the BRR
4.2 Upper mantle structure
The Rayleigh wave tomography of the upper mantle is in very good agreement with the
independent P and S body-wave traveltime tomographic images from Schmandt and
Humphreys (2010) using USArray/TA and other data. For example, compare the shear
velocity anomaly map at 90 km shown Figure 4.2(j) with the surface wave tomography at
similar depths (Figure 4.2 k), the lateral heterogeneity patterns are similar, both showing
a large and continuous low-velocity anomaly surrounding the high velocity body in the
CP interior and its north. The slightly different magnitudes of low and high velocity
anomalies are largely due to the radial anisotropy of the S velocity structure, in which the
Rayleigh waves are sensing the vertical component (Sv) while the body-wave traveltime
tomography is inverting the SH component. Such radial anisotropy could reach more than
5% beneath the BRP while it stays low inside the plateau (Moschetti et al, 2010). In
general, the Rayleigh wave tomography model shows a high consistency with the
previous studies. The pattern of the lateral seismic heterogeneity strongly correlates with
42
the recent magmatic activities and the surface heat flow data (Figure 4.1a, b). The shear
velocity variations in the upper mantle are described first, followed by the interpretation
at each tectonic province.
4.2.1 Upper mantle seismic heterogeneity
The strong lateral seismic heterogeneity in the upper mantle beneath the CP and the
adjacent regions is clearly imaged by the surface wave tomography. A large high velocity
anomaly beneath much of the CP (1/2 to 2/3), extending from the Navajo volcanic field
in the south across the NE striking accretionary boundary (Cheyenne Belt) to the
southern edge of the Archean Wyoming Province. This high velocity body is consistently
observed from the uppermost mantle (60 km) to deeper upper mantle (-200 km) (Figure
4.2c-h). It is surrounded by a continuous, large-scale and strong low velocity anomaly
around the plateau rim beneath the southernmost Rocky Mountains, the rifted RGR, and
the extended SBR and NBR regions. At greater depths, the relatively high velocity
anomalies weaken, and a NE-trending narrow low velocity belt (e.g. Figure 4.2 f-h) is
observed to have penetrated into the high velocity body in the northwest of the Four
Corner region, which may indicate that the lithospheric root of the plateau are suffering
convective erosion from the underlying asthenosphere.
The tomographic images (Figure 4.2 c-h) also show one of the largest amplitude low-
velocity anomalies from depths of 60-150 km beneath the NBR to the west of the
Wasatch Front in Utah. The strong and sharp contrast between this low and high velocity
anomaly under the abutting CP interior is in good agreement with the extended traveltime
43
tomographic result from the new La RISTRA 1.5 data (Sine et al, 2008). The strong low
velocity zones, though bounded by the normal faults at the western edge of the plateau,
penetrate the margins of the less deformed plateau and become broader at greater depths.
The penetration distance from the SBR is about 100-150 km, greater than that from the
NBR (50-100 km). The inferred thickness of the mantle lid beneath the NBR from the
Rayleigh wave tomography model is relatively small (-20 km), consistent with the
observations from the inter-station surface wave technique using the CP-GB data
(Sheehan et al, 1997; Lastowka et al, 2001). The thin mantle lid is underlain by a more
steep negative velocity gradient than beneath the CP (Figure 3.7), with minimum Vs as
low as ~ 4.1-4.2 km/s at ~ 100 km (~ 7-8% slower than the global average). In the
southwestern margin of the plateau, a similar distinct LVZ with a thin mantle lid in the
transition region (identified by the PACE group, Benz and McCarthy, 1994; Parsons and
McCarthy, 1995) as well as the recent uplift, are suggested to be a delayed mechanical
response to the lithospheric thinning after the low-angle slab subduction and foundering
(Parsons and McCarthy, 1995). From the southwestern margin to the plateau interior, the
physical state of the upper mantle becomes more complicated, and an ongoing
delamination may provide an explanation for this complexity (see section 4.4).
In the southeastern plateau and the rift, a "lambda-shaped" (k) low-velocity pattern in
New Mexico is observed beneath the southern RGR and the Jemez volcanic fields
(Figure 4.2 c-e). These two low velocity trends merge at the Jemez Mountains close to
the New Mexico-Colorado boarder, and extend northwards into the SRM with slight
increase of Vs. The smallest shear velocities of ~ 8% lower than the global average are
44
centered directly beneath the rift axis and along the Jemez lineament (Figure 4.2 c-e),
extending from shallow uppermost mantle ~ 40 km to ~ 100 km deep, at which depth the
"X" shaped low velocity feature (Figure 4.2 f, g) blends with the large-scale SBR low
velocity anomaly. The velocities of the Colorado Rockies, though not as low as the Jemez
lineament, are still relatively low compared with those under the northern plateau and
north of the Cheyenne Belt. This "X" shaped low-velocity pattern is bounded by high
velocity anomalies beneath the plateau interior to the northwest of the Jemez lineament
and by the western edge of the Great Plains to the east of the rift. A large velocity
contrast, up to -10%, exists between the Jemez lineament and the high velocity anomaly
that separates the Jemez lineament and low velocities beneath the northwestern Navajo
volcanic field. This remarkable lateral velocity gradient from the rifting-related volcanic
fields to the stable plateau interior has also been clearly imaged by the RISTRA group
using both P and S traveltime tomography (Gao et ah, 2004; Sine et ah, 2008) and
surface wave measurements (West et ah, 2004), though their tomographic results are
limited to a single line (as indicated by the red dot lines in Figure 4.2 a-j). The higher
velocities to the east of the rift, determined by numerous seismic investigations {e.g. van
der Lee and Nolet, 1997; Dueker et ah, 2001), stop the rift low velocity zone at its eastern
edge which is associated with the westernmost edge of the cratonic lithosphere of the
North American Plate.
At a given depth above 200 km (Figure 4.2 c-g), the upper mantle shear velocity
variations can reach as high as ~ 12%, generally due to the contrast between high
velocities in the northern plateau and low ones along the CP-BR transition zones. This
45
strong lateral heterogeneity weakens with greater depths (Figure 4.2 h, i) due to relatively
weak sensitivity of the phase to shear velocities and lateral averaging effect at long
periods in the Rayleigh wave tomographic technique. The relatively high velocity relative
to the TNA model starts to dominant below 250 km (Figure 4.2 i), showing signature of
the base of sublithospheric low velocity zones (see Figure 4.3). This velocity increase
could be related to the change of grain size from ~ mm to ~ cm (Faul and Jackson, 2005).
4.2.2 Colorado Plateau mantle lithosphere
The CP interior has low heat flow (< 50 mW/m ) and relatively thick Proterozoic
lithosphere compared to the surrounding BRP and RGR, as inferred from the surface
wave model (see Table 1, Figure 3.7, Figure 4.1a) and previous studies {e.g. Thompson
and Zoback, 1979). Such cold and thick Proterozoic mantle lithosphere beneath the
plateau forms a thick thermal boundary layer, and thus shields the crust from deformation
and magmatism by limiting the heat transfer from the convecting mantle to the crust. This
thermal boundary layer is less dense and more ion depleted mantle lithosphere with
greater strength compared with the BRP, as proposed by Lee et al. (2001) based on the
peridotite xenoliths data. The slightly decreased Vs (~ 4.4 km/s) in the mantle lid of the
plateau compared to the global average (4.5 km/s), is consistent with the measurement of
relatively low Sn velocities (Nolet et al, 1998) in the same region. Even though a large
part of the plateau mantle lithosphere survived from the Laramide flat-slab subduction,
moderate hydration from the de-watering after the sinking of the Farallon slab could
cause minor reduction of the mantle lid shear velocity. This slightly low Vs and the
relatively high measured Pn of 8.0-8.1 km/s {e.g. Beghoul and Barazangi, 1989) yield a
46
high Vp/Vs ratio (-1.80-1.85), which has been observed by receiver function studies (e.g.
Levander and Miller, 2010; Bashir et al, 2009).
The crustal thickness of the CP used in this Rayleigh wave tomography from the PdS RF
estimates, is shallower than previous RF studies (e.g. Zandt et al, 1995; Gilbert and
Sheehan, 2004; Wilson et al., 2005a), and does not provide sufficient isostatic support for
the present 2-km elevation. The lithospheric thickness (-100 km) of the plateau inferred
from the surface wave model (see Table 1, Figure 3.7) rules out the almost complete
removal of the continental lithosphere beneath the plateau following flat-slab subduction,
as suggested by Bird (1988). In fact, the estimated LAB depth is largely consistent with
the partial delamination model using the mantle buoyancy to explain the anomalous
plateau elevations (Beghoul and Barazangi, 1989; Spencer, 1996; Lastowka et al, 2001).
It is also consistent with the fact that the cored mantle lithosphere imaged in the Vs model
preserves the Proterozoic signature including the relatively high seismic velocities and
strength, as observed in the xenoliths, despite small thermal modifications after removing
partial base of the lithosphere at - 30 Ma (Spencer, 1996). The relatively low surface heat
flow indicates that the recent heating at the base of the thick plateau lithosphere has not
yet reached the surface (Thompson and Zoback, 1979). In addition to the isostatic support
of the crust and dynamic uplift (Moucha et al, 2008, 2009), a mantle contribution to the
current uplift of the plateau is necessary in the form of asthenospheric support beneath
the low-density plateau lithosphere. Based on the surface wave tomography model in this
study, I favor the partial delamination model (Beghoul and Barazangi, 1989; Spencer,
1996; Lastowka et al, 2001) combined with dynamic uplift, isostatic response, and
47
asthenospheric buoyancy to support the current elevation of the plateau.
4.2.3 Basin and Range and CP-BR transition zones
These extremely low upper mantle velocities beneath the Great Basin cannot be merely
caused by the ~ 200-400°C temperature (Riter and Smith, 1996; Sine et al, 2008) or
compositional variations, but they also require the presence of partial melt and water
content in the nominally anhydrous minerals in the upper mantle lithosphere (Hammond
and Humphreys, 2000; Karato, 2006; Sine et al, 2008). The hydration-induced melting
is a consequence of the slab rollback (Humphreys et al., 2003) and the intensive Basin
and Range extension in Early and Late Tertiary. The differences between the magnitudes
of the low Vs in the northern and southern Basin and Range are probably related to
different intensity and history of the Cenozoic extension and magmatism separated by the
"magmatic gap" zone in the central Basin and Range, such as the chemical and thermal
variations in the upper mantle in the NBR and SBR (Armstrong and Ward, 1991; Axen et
al, 1993; Humphreys and Deuker, 1994b; Zandt etal, 1995).
The various penetrating distances agree well with the different magmatic encroachment
rates of 6.3 km Myr"1 in the SBR and only 4.0 km Myr"1 in the NBR (Roy et al, 2009).
These dominant low velocity features could be associated with the small-scale convection
(e.g. van Wijk et al, 2008) around the flanks of the plateau lithosphere with driving
forces from the warm BRP (characterized by 100-200% extension since Mid-Tertiary)
asthenosphere at relatively shallow depths (50-100 km), and is accompanied by increased
seismicity and a high surface heat flow regime (Figure 4.1a). This small-scale convection
48
could also provide dynamic uplift support from the northwestern and southwestern edges
of the plateau, as explanation for the excess marginal elevation. Additionally, the recent
Late Tertiary and Quaternary magmatism and the significantly metasomatic alternation at
the margins of the plateau (Roy et al, 2005, 2009; Smith and Griffin, 2005) could have
produced certain percentages of melt when the ascending asthenospheric materials heated
the base of the thin lithosphere between the extension regime and the relatively thick
Proterozoic lithosphere under the plateau.
4.2.4 Rio Grande Rift and Jemez lineament volcanism
This "X" shaped anomaly structures could have resulted from the Proterozoic structural
control on the mantle lithosphere similar to the crust (Karlstrom and Humphreys, 1998),
enhanced by the intrusion of hot asthenospheric materials to the base of the lithosphere
after removal of the Farallon slab. The temperature variations across the rift and the
mechanical thinning during the rifting could both contribute to such large velocity
gradients across the CP-RGR-GP zone. More recent magmatic and tectonic activities (in
the last 10 Myr) after the post-Laramide slab removal and Basin and Range extension is
likely caused by small-scale convection at the edge of the lithospheric root roughly
beneath the rift axis, which introduces a lateral temperature difference. Recent volcanic
fields along the Jemez lineament, rather than the RGR, coincide with a regional extension
induced by the small clockwise rotation of- 3° (Aldrich et al, 1986; Hamilton, 1988) of
the plateau that weakened the lithosphere abutting the rift and allowed magma ascent.
49
4.2.5 Southern Rocky Mountains
The upper mantle velocity structures beneath the SRM bear certain resemblance to the
Deep Probe, the CD-ROM and the RMF (Rocky Mountain Front Broadband Seismic
Experiment) projects (Snelson et ah, 1998; Henstock et ah, 1998; Gorman et ah, 2002;
Levander et ah, 2005; Dueker et ah, 2001; Li et ah, 2005), and exhibit more complicated
heterogeneous structures in 3-D perspective views of the shear velocity model (Figure 4.2
& 4.3). The less pronounced low velocity zones beneath the SRM could be caused by
about 1% partial melt in the upper mantle along the Colorado Mineral Belt (Dueker et ah,
2001; Humphreys et ah, 2003). The southern edge of the Rockies along the Jemez
lineament also correlates well with the Yavapai-Mazatzal boundary, further supporting
the structural control on tectonism by the ancient accretionary boundaries that weakened
the lithospheric zone (Karlstrom and Humphreys, 1998). The mantle velocity
perturbations beneath the SRM are not as strong as those in the upper and middle crustal
structures (Figure 4.2 c-g). However, the moderately low velocities compared to the
continental velocity still need a small amount of partial melt in the upper mantle
combined with a high temperature.
4.3 Sublithospheric structure
The surface wave tomography model has clearly imaged the sublithospheric channels of
low seismic velocity beneath the CP and the surrounding areas. Low velocity channels
are typically associated with low-viscosity zones, frequently attributed to the combined
weakening effects of water and partial melt. It stands to reason that I am imaging a sub
continental low-viscosity asthenosphere. The asthenosphere plays a key role in
50
-116 -114 -112 -110 -108 -106
E
I
-116 -114 -112 -110 -108 -106
Vs (km/s) 4.0 4.1 4.2 4.3 4.4 4.5 4.6
Figure 4.3 Vs cross-sections along 33.5°, 34.5°, 35.5°, 36.5°, 37.5°, 40.5°N showing location of low-velocity channels, with tectonic boundaries in red ticks and suggested delamination location in white arrows. The low-velocity channel beneath the RGR is relatively shallow compared with the Basin and Range Province.
51
understanding fundamental mantle convection and plate tectonics questions. Recent
numeric and analytic work revealed that flow in the asthenosphere is a Poiseuille-Couette
flow associated with the wavelength of convection and the relative strength of lithosphere
and asthenosphere (Hoink and Lenardic, 2008, 2009). The accurate depths of the
sublithospheric low-viscosity channels are difficult to be identified with Rayleigh waves
because these are relatively insensitive to the seismic boundaries of velocity changes. The
integrating nature of the measurements cannot distinguish sharp from gradual velocity
gradients beneath the lithosphere (Eaton et al, 2009). However, the period-dependent
sensitivity peaks to the absolute shear velocity over a certain depth range still provide
information on the relevant depths and thicknesses of the low-velocity channels in the
sublithospheric structures.
Six cross-sections along 33.5°, 34.5°, 35.5°, 36.5°, 37.5° and 40.5°N, as shown in Figure
4.3, clearly present the low-velocity channels beneath the southwestern US. The base of
the asthenosphere, which can hardly be inferred from the PdS or SdP receiver function
analysis, can be well estimated here from the surface wave tomographic images to study
the geodynamic interaction between the lithosphere and underlying asthenosphere at
various regional tectonic regions. The dominant strong low-velocity channels beneath
both southern and northern Basin and Range provinces are shallow (60-120 km) but
strong, and the magnitudes decrease to some extent when the region is far from the
plateau margins (Figure 4.3). The short-wavelength low-velocity asthenosphere beneath
the RGR is even shallower (to ~50 km) and a bit stronger compared to the Basin and
Range region, and this low-velocity channel is thinner (see the 34.5° cross-section in
52
Figure 4.3). The variety of the depths and thicknesses of the low-velocity channels
beneath the Basin and Range provinces, the RGR valley and the CP is essential to
understanding the current geodynamics after a complex superposition of multiple tectonic
and magmatic effects during the Cenozoic era.
The relative shallow but strong low-velocity zones were likely developed during Basin
and Range extension, as the lithosphere was thinned by a factor of up to 2. Basin and
Range asthenosphere might exert strong convective erosion to the relatively undeformed
plateau with increased depths. It correlates well with the migration of magmatism into the
margins around the Colorado Plateau from south, southwest and northwest. The
development of a strong low velocity zone was probably enhanced by hydration from the
Farallon slab, as well as the strongly coupled mantle convections rising to the shallow
base of the Basin and Range lithosphere. The slightly decreasing magnitudes of the low-
velocity (low-viscosity) channels away from the front of the magmatic encroachment
may indicate a cooling effect after the magmatic front moved away.
The consensus on classification of the Rio Grande Rift as a passive or active rift has yet
to be achieved. The observed emplacement of the hot asthenosphere (roughly to 150 km
at the base) at the replaced lithosphere below the rift in the Rayleigh wave model is
consistent with both surface and body wave traveltime tomographic results from the La
RISTRA group (Gao et al, 2004; West et al, 2004; Wilson et al, 2005b) and the
dynamic test from the numerical modeling (van Wijk et al, 2008), indicating that the Rio
Grande Rift may be a recent passive rift caused by the small-scale convection in response
53
to far-field-driven forcing of lithosphere. The consistency argues that it is relatively
unlikely that the RGR experienced an active rifting process controlled by the mantle
upwelling from deep source suggested by Moucha et al. (2008) based on the joint
seismic-geodynamic simulation.
The shear velocity model documents large lateral heterogeneity in depths thickness, and
minimum velocity in the low-velocity channels and maximum velocity in the overlying 1
lithosphere. The variations in the asthenosphere could be caused by the lateral
temperature gradients, which can generate strong pressure-driven flow in the
asthenosphere. In addition, a small amount could also be ascribed to the compositional
variations, and water content or melt. The asthenospheric channel beneath the Colorado
Plateau is moderately weak and shows less shear velocity reduction compared to the
surrounding BRP and RGR valley. The step in lithospheric thickness between these
distinct tectonic regions could cause strong lateral interaction between the ascending
asthenosphere and the lithosphere besides the vertical interaction (van Wijk et al, 2008).
The transitional regions of the lithospheric thicknesses strongly correlate with strongest
low-velocity zones that are at the margins between different tectonic regions. The well-
imaged low-velocity channels beneath the study area can help to understand the regional
lithosphere-asthenosphere interactions beneath the extended, rifted and undeformed
regions from various evolutions of mantle structures since Late Cretaceous.
54
4.4 Possible delamination
4.4.1 Comparison with receiver function imaging
The Rayleigh wave tomography model shows strong correlation with the PdS teleseismic
receiver function (RF) image volumes at the same area, despite the differences in
frequency bands (0.1-0.75 Hz in RF; 0.006-0.05 Hz in the Rayleigh wave tomography).
The receiver function imaging, which represents the converted teleseismic wave
responses beneath the stations, was obtained by applying a time-domain iterative
deconvolution approach to 106 teleseismic events recorded by the USArray/TA stations
(Levander and Miller, 2010). After converting the individual receiver functions from the
time domain to the depth domain, common conversion point (CCP) stacking was used to
migrate them laterally to the conversion point based on a 3-D reference velocity model.
The model was constructed from the Crust 2.0 model (Bassin et al, 2000), the University
of California at San Deigo upper mantle model (Buehler and Shearer, 2009), and the
University of Oregon P and S body wave tomography model (Schmandt and Humphreys,
2010). CCP stacking enhances the image by stacking redundant signals from the same
conversion point. The final stacked RF images reveal unusually complicated Moho
topography beneath the Colorado Plateau, especially beneath the northwestern portion of
the plateau, in which the Moho broadens and separates into two distinct positive-
amplitude events in the interior of the plateau. The top positive event is relatively shallow
compared with previous estimates of crustal thickness in the CP, while the deeper one
extends to depths (> 50 km) greater than the typical estimates and can hardly be
interpreted as the thickened Moho.
55
Figure 4.4 Map view of LAB depths picked from the PdS receiver function profiles with surface wave tomographic result as a guide.
The lithosphere-asthenosphere boundary (LAB) depths can also be better estimated from
the RF volumes with additional constraints from the surface wave tomography model in
this study. The LAB depth estimates of- 60-120 km from the PdS RF images, as shown
in white dash lines corresponding to the negative amplitude events, are well correlated
with the top of the low-velocity channel in the Vs images. This strong correlation
suggests that the Vs images can assist with determining LAB depths, particularly when
there are multiple negative events due to contamination from noise, reverberation, or
56
subsurface discontinuities within the upper or middle crustal structure, even though there
are larger uncertainties in the estimate of LAB depths directly from the Vs images than
from the RF due to the limitation on vertical resolution in the Rayleigh wave study.
Figure 4.4 shows the final estimated LAB depth map from the combined analysis of these
two studies, showing that the Basin and Range Province has relatively shallow LAB at
approximately 50-80 km, the CP has a deeper LAB of- 80-120 km, and the SRM LAB is
even deeper to -150 km.
Most importantly, the combined interpretation of these two studies suggests that a portion
of the lower crust and uppermost mantle are delaminating along the crust/mantle
boundary, and a new Moho is being formed at relatively shallow depths. Figure 4.5
shows a comparison of two profiles along 37°N and 111°W in these two independent
studies (except that crustal thickness constraint in the Rayleigh wave study comes from
the PdS receiver function estimates), with selected receiver function profiles (light red
means positive amplitude, while light green means negative) at every degree are also
displayed in the Vs images for direct comparison. As shown in Figure 4.5, a high velocity
anomaly is clearly imaged between 113°W and 111°W as hanging on to the western edge
of the plateau along the 37°N line, though more tilted than the second positive event as
imaged in the RF cross section. The key features in the RF images and the Vs images are
consistent, showing that a relatively high velocity anomaly appears to detach from the
plateau interior towards the western margins of the CP. Along the N-S profile at 111° W, I
also observed a high velocity anomaly with horizontal scale of ~ 1 ° in the Vs image, the
top of which we correlate with the lower part of the detachment as imaged in the RF
57
(a) 37 N Latitude
BRP CP
W
*
E W
\ '••
, It ;.
CP SRM .
'..: ^ * T .,»^;"-
KAJ aU^.^J MLu m i d rt , Ay , *j*W
•1C -*.J .'.2 .',.3 -V- -lb 4.6
1 3
(b)
SRM • H -
,'tf » 1' v**i
111W Longitude CP SBR SRM CP
S N
I » tl.
^ V ^ ^/ ̂ H- /*4
SBR
Amplitude
Figure 4.5 Comparison of cross-sections of shear velocity structure (left) and 1.25 Hz PdS receiver function images (right) along 37°N (a) and 111°W (b). Top of each image shows the elevation along the lines with arrows pointing at the Monument uplift (MU) where the delamination may start. I set the crustal velocity to 3.9 km/s for display, and plotted receiver function profiles (positive in light red and negative in light green) on the Vs cross-sections. Black solid lines and white dash ones indicate the Moho and LAB depths, respectively.
58
cross-section. This part of the high Vs anomaly is connected to a vertical high anomaly in
the northern part of the plateau, but cannot be imaged in the PdS images because the
receiver functions are not able to resolve vertical interfaces. The Vs model from the
Rayleigh wave tomography in this study supports the delamination hypothesis suggested
by the PdS RF study (Levander and Miller, 2010) to explain the two split positive events
in RF profiles beneath much of the plateau.
4.4.2 Possible delamination mechanism
One of the generally accepted descriptions of the delamination process is that the mantle
lithosphere peels off the overlying crust along the crust/mantle boundary and the
detached lithosphere sinks into the underlying asthenosphere (Morency and Doin, 2004).
The suggested ongoing delamination beneath the western and northwestern portions of
the Colorado Plateau develops from the observations of the high velocity anomaly in the
surface wave tomography and the second Moho-related positive event in the PdS receiver
function images. Although the mechanism of this regional delamination is not clearly
understood yet, the combined effects of regional surface uplift, magmatic underplating in
the lower part of the crust, and the high seismic attenuation might initiate the detachment
of the lower crust and uppermost mantle beneath the western plateau, probably starting
from the west of the Navajo volcanic field.
The newly re-dated Monument Uplift could be associated with the initiation of the
delamination in the western plateau. Along the 37°N profile in Figure 4.5a, the regional
59
elevation of the southern edge of the Monument Uplift (west of the Navajo volcanic
field) between 111°W and 110°W is relatively high and strongly correlates with the
delamination location where the deeper warm asthenospheric materials might rise up to
the base of the plateau crust, as indicated by the low velocity anomaly in the Vs cross-
section, or the negative event between the two separated positive events in the receiver
function image. Although the dominant rock/surface basement-cored uplifts and
monoclines in the CP are associated with the Sevier-Laramide orogenic deformation
(140-45 Ma) from Late Cretaceous to Early Tertiary, the rise of several small-scale
regions in the western and northwestern plateau such as the Circle Cliffs, the San Rafael,
the Monument, and part of the Kaibab uplifts were estimated to be formed more recently
(approximately 33 to 11 Ma after Laramide orogeny) by using the (U-Th)/He
thermochronology technique (Stockli et ah, 2002; Flowers et ah, 2008; Davis and Bump,
2009). This more recent regional uplift could respond to the initiation of delamination
that recently occurred.
The thickened crust caused by the regional crustal compression and magmatic
underplating could probably initiate the delamination, occurring along the localized and
weakened lower crust due to hydration and heating after the rollback of the Farallon slab.
The nearly west-east compression in the western part of the CP near 37°N (Saleeby,
2003) could cause a small part of the crustal thickening, and furthermore, the magmatic
intrusion or underplating at the base of the lower crust was suggested to contribute most
of the crustal thickening beneath the western plateau (Morgan and Swanberg, 1985; Wolf
and Cipar, 1993). The crustal thickening could stretch the geotherm under the southern
60
Monument Uplift and the magmatic underplating could also raise the Moho temperature,
which further thermally weakened the bottom of the lower crust. Additionally, due to the
suggested strong small-scale convection from the BRP around the flanks of the mantle
lithosphere beneath the plateau interior, the lithosphere of the western plateau was further
modified convectively and heated at shallower depths by the asthenosphere than the
preserved Proterozoic lithosphere beneath the plateau interior. The dehydration of
serpentinites in the cold Farallon subduction plate brought aqueous fluids to the
overriding continental lithosphere during subduction, further hydrated the plateau
lithosphere (e.g. Lee, 2005; Li et ah, 2008) and reduced the viscosity at the basal crust
(Humphreys et ah, 2003; Dixon et ah, 2004; Hyndman et ah, 2005). Furthermore, the
delamination started coincidently at the place with strong seismic attenuation of P and S
velocity (Ringler et ah, pers. comm., 2010). The thickening of the crust and the
weakening at the base of the crust could trigger the delamination that the lower part of the
crust and the upper mantle detached from the low-viscosity lower part of the crust and
fell into the warm and mobile asthenosphere driven by negative buoyancy (Meissner and
Mooney, 1998), and such delamination may be the cause of recent regional uplift of the
western plateau. This delamination mechanism described here is similar to the one
suggested by Morency and Doin (2004) based on 2-D numerical simulations using a
brittle-ductile rheological model. The possible ongoing delamination will reset the Moho
above its former depth and is therefore a dynamic crust/mantle boundary.
The initiation location of the delamination, such as under the southern Monument uplift
along the 37°N profile, is limited to the west of the Navajo volcanic field where no
61
significant magmatic underplating was suggested based on mantle xenolith evidence
(Mattie et al, 1997; Condie and Selverstone, 1999; Selverstone et al, 1999). Though the
garnetite and the associated eclogite xenoliths provide some evidence of mantle hydration
and metasomatism related to the Farallon subduction (Smith et al, 2004; Smith and
Griffin, 2005), the eastern side of the delaminating region has a relatively thick
lithosphere that could prevent the heating and weakening across the crust/mantle
boundary from the underlying asthenosphere (Figure 4.3). However, no significant
volcanic activity or thermal signals have been observed to be associated with the
suggested delamination. The moderately low velocity anomaly is indicative of the
replacement of the lower crust and upper mantle by the upward migration of the
asthenosphere though the velocity is not as low as under the BRP. An estimate of the
conductive time constant of about 13 Myr (using thermal diffusivity of 10"2 cm2/sec and
crustal thickness of ~ 40 km) may indicate that even if the base of crust was heated by the
ascending asthenosphere, the heat has not reached the surface yet, providing the
delamination occurred within 13 Myr. More geochemical and geophysical evidence is to
be investigated to support this delamination hypothesis.
62
CHAPTER 5
CONCLUSIONS
This study has developed a new high-resolution 3-D shear velocity model (to the depth of
~ 250 km) from the Rayleigh wave tomography technique using the latest USArray data
to reveal in detail strong seismic heterogeneity in the crust and upper mantle beneath the
CP and the adjacent tectonic provinces in the southwestern U.S.. Relatively high
velocities are observed in the Proterozoic mantle lithosphere beneath the CP interior, and
extend to the Archean province in the north. Continuous large-magnitude upper mantle
low velocity anomalies surround and underlie the edges of the plateau in the northwest,
south and southeast, including the CP-SBR and CP-NBR transition zones, the COMB,
the Jemez lineament and the RGR. These may be indicative of small-scale convection
involving shallow asthenosphere. In addition, the sublithospheric low-velocity (low-
viscosity) channel is observed to vary in thickness, depth, and minimum velocity. The
strong lateral heterogeneity of Vs can be attributed to the combined effects of temperature
variations, lithospheric thickness differences, compositional change, partial melt and
presence of water content in the mantle minerals. The complexity of the modern crustal,
upper mantle structures are strongly correlated with the Cenozoic lithospheric
modification related to the Farallon slab subduction, foundering, and the major the
extension, rifting and the associated magmatism.
The final shear velocity model agrees well with numerous previous regional or
continental-scale seismological investigations, such as the PACE, the CP-GB, the Deep
Probe, the CD-ROM, the RMF, as well as the more recent La RISTRA projects. It also
63
shows a good agreement with the new body-wave traveltime tomography models
(Schmandt and Humphreys, 2010) and the new PdS and SdP receiver function results
(Levander and Miller, 2010) using different frequency bands of the US Array data or
different imaging methods. The LAB depths inferred as the centers of the negative
velocity gradients from this Rayleigh wave tomography model have strong correlation
with the negative events in the PdS receiver function volumes.
The surface wave tomography model also successfully images, for the first time, high
velocity anomalies in the same location where lithospheric delamination has been
suggested based on receiver function imaging (Levander and Miller, 2010) to explain the
complexity of the crust/mantle boundary. This delamination event may be related to the
recent uplift in the Monument, where could be triggered by the magmatic underplating
and strong localized seismic attenuation to the west of the Navajo volcanic field.
The development of this high-resolution shear velocity model in this thesis can be used to
numerically study the geodynamic interactions between the lithosphere and
asthenosphere in regional scale. The successful implementation of the modified two-
plane wave technique makes it possible to apply the combined the surface waves and
ambient noise tomography to the new available Flexible Array Mendocino Experiment
(FAME) dataset to reveal the complicated upper mantle in the Mendocino Triple Junction
(MTJ) region. The dispersion curves in the Rayleigh wave tomography also provide the
surface data for the joint inversion with the receiver functions to examine the upper
mantle of the United States covered by the US Array stations.
64
LIST OF REFERENCES
Aki, K., and P.G. Richards (2002), Quantitative Seismology, second ed., University Science Books, Sausalito, CA, 700pp.
Aldrich, M.J., Jr., C.E. Chapin, and A.W. Laughlin (1986), Stress history and tectonic development of the Rio Grande Rift, New Mexico, J. Geophys. Res., 91, 6199-6211.
Amante, C , and B. W. Eakins, ETOPOl 1 Arc-Minute Global Relief Model: Procedures, Data Sources and Analysis. NOAA Technical Memorandum NESDIS NGDC-24, 19 pp, March 2009.
Armstrong, R.L. and P. Ward (1991), Evolving geographic patterns of Cenozoic magmatism in the North American Cordillera: The temporal and spatial association of magmatism and metamorphic core complexes, J. Geophys. Res., 96, 13,201-13,224.
Axen, G.J., Taylor, W.J., and Bartley, J.M. (1993), Space-time patterns of the onset of extension and magmatism, southern Great Basin, Nevada, Utah, and California, Geol. Soc. Am. Bull, 105, 56-76.
Balch, R.S., H.E. Hartse, A.R. Sanford, and K.-W. Lin (1997), A new map of the geographic extent of the Socorro mid-crustal magma body, Bull. Seismol. Soc. Am., 87, 174-182.
Bashir, L., K.H. Liu, and S.S. Gao (2009), Crustal thickness and Vp/Vs ratio variations beneath the Colorado Plateau, AGU Fall Meeting 2009.
Bassin, C, G. Laske, and G. Masters (2000), The Current Limits of Resolution for Surface Wave Tomography in North America, EOS Trans AGU, 81, F897.
Beghoul, N., M. Barazangi (1989), Mapping high Pn velocity beneath the Colorado Plateau constrains uplift models, J. Geophys. Res., 94, 7083-7104.
Benz, H. M., and J. McCarthy (1994), Evidence for an upper mantle low velocity zone beneath the southern Basin and Range-Colorado Plateau transition zone, Geophys. Res. Lett., 21, 509-512.
Bird, P. (1984), Laramide crustal thickening event in the Rocky Mountain foreland and Great Plains, Tectonics, 3,741-758.
Bird, P. (1988), Formation of the Rocky Mountains, western United States: A continuum computer model, Science, 239, 1501-1507.
Bowring, S.A., and K.E. Karlstrom (1990), Growth, stabilization, and reactivation of Proterozoic lithosphere in the southwestern United States, Geology, 18, no. 12, 1203-1206
Buehler, J.S., and P.M. Shearer (2009), Pn tomography of the western United States using USArray, J. Geophys. Res., submitted.
Burdick, S., C. Li, V. Martynov, T. Cox, J. Eakins, L. Astiz, F.L. Vernon, G.L. Pavlis, and R.D. van der Hilst (2008), Upper mantle heterogeneity beneath North America from travel time
65
tomography with global and USArray transportable array data, Seism. Res. Lett., 79:3, 384-392.
Christiansen, R.L., and P.W. Lipman (1972), Cenozoic volcanism and platetectonic evolution of the western United States, II. Late Cenozoic, Philos. Trans. R. Soc. London, Ser. A 271, 249-284.
Condie, K. C, and J. Selverstone (1999), The crust of the Colorado Plateau: New views of an old arc,/. Geol., 107,387-397.
Coney, P., and S. Reynolds (1977), Cordilleran benioff zones, Nature, 270, 403-406.
Davis, G.H., and A.P. Bump (2009), Structural geologic evolution of the Colorado Plateau, in Kay, S.M., Ramos, V.A., and Dickinson, W.R., eds., Backbone of the Americas: Shallow Subduction, Plateau Uplift, and Ridge and Terrane Collision: Geological Society of America Memoir 204, p. 99-124, doi: 10.1130/2009.1204(05).
Dickinson, W.R., and W.S. Snyder (1978), Plate tectonics of the Laramide Orogeny, in Laramide Folding Associated With Basement Block Faulting in the Western United States, edited by V. Matthews, Mem. Geol. Soc. Am., 151, 355-366.
Dixon, J.E., Dixon, T.H., Bell, D.R., and Malservisi, R. (2004), Lateral variation in upper mantle viscosity: Role of water, Earth Planet. Sci. Lett., 222, 451-467, doi: 10.1016/j.epsl.2004.03.022.
Dueker, K., H. Yuan, and B. Zurek (2001), Thick-structured Proterozoic lithosphere of the Rocky Mountain region, GSA Today, 11,4-9.
Eaton, D.W., F. Darbyshire, R.L. Evans, H. Griitter, A.G. Jones, and X. Yuan (2009), The elusive lithosphere-asthenosphere boundary (LAB) beneath cratons, Lithos, 109 (1-2), 1-22.
English, J.M., S.T. Johnston, and K.L. Wang (2003), Thermal modeling of the Laramide orogeny: testing the flat-slab subduction hypothesis, Earth Planet. Sci. Lett., 214, 619-632
Faul, U.H., and I. Jackson (2005), The seismological signature of temperature and grain size variations in the upper mantle, Earth Planet. Sci. Lett., 234, 119-134.
Forsyth, D.W., Webb, S., Dorman, L. Shen, Y. (1998), Phase velocities of Rayleigh waves in the MELT experiment on the East Pacific Rise, Science, 280, 1235-1238.
Forsyth, D.W., and A. Li (2005), Array analysis of two-dimensional variations in surface wave phase velocity and azimuthal anisotropy in the presence of multipathing interference, in Seismic Earth: Array Analysis of Broadband Seismograms, edited by A. Levander and G. Nolet, AGU, Washington, D. C.
Flowers, R.M., B.P. Wernicke, and K.A. Farley (2008), Unroofing, incision, and uplift history of the southwestern Colorado Plateau from apatite (U-Th)/He thermochronometry, Geol. Soc. Am. Bull, 120, 571-587, doi:10.1130/B26231.1.
Frassetto, A., H. Gilbert, G. Zandt, S. Beck, and M.J. Fouch (2006), Support of high elevation in the southern Basin and Range based on the composition and architecture of the crust in the Basin and Range and Colorado Plateau, Earth Planet. Sci. Lett., 249, 62-73.
66
Gao, W., S.P. Grand, W.S. Baldridge, D. Wilson, M. West, J.F. Ni, and R. Aster (2004), Upper mantle convection beneath the central Rio Grande rift imaged by P and S wave tomography, J. Geophys. Res, 109, B03305, doi: 10.1029/2003JB002743.
Gilbert, H. J., and A. F. Sheehan (2004), Images of crustal variations in the intermountain west, J. Geophys. Res., 109, B03306, doi:10.1029/2003JB002730.
Gilbert, H., A.A. Velasco, and G. Zandt (2007), Preservation of Proterozoic terrane boundaries within the Colorado Plateau and implications for its tectonic evolution, Earth Planet. Sci. Lett., 258 (1-2), 237-248. doi: 10.1016/j.epsl.2007.03.034
Gorman, A.R., Clowes, R.M., Ellis, R.M., Henstock, T.J., Spence, G.D., Keller, G.R., Levander, A.R., Snelson, CM., Burianyk, M.J.A., Kanasewich, E.R., Asudeh, I., Hajnal Z., and Miller, K.C. (2002), Deep Probe - Imaging the roots of western North America, Canadian Journal of Earth Sciences, v. 39, no. 3, 375-398.
Grand, S.P. (1994), Mantle shear structure beneath the Americas and surrounding oceans, J. Geophys. Res., 99 (B6), 11,591-11,621.
Grand, S.P., and D.V. Helmberger (1984), Upper mantle shear structure of North America, Geophys. J. R. Astron. Soc, 76, 399-438.
Hamilton, W.B. (1988), Laramide crustal shortening, in Interaction of the Rocky Mountain Foreland and the Cordilleran Thrust Belt, edited by C.J. Schmidt and W.J. Perry Jr., Mem. Geol. Soc. Am., 171,27-39.
Hammond, W.C., and E.D. Humphreys (2000), Upper mantle seismic wave velocity: Effects of realistic partial melt geometries, J. Geophys. Res., 105(B5), 10,975-10,986.
Henstock, T., and the Deep Probe Working Group (1998), Probing the Archean and Proterozoic lithosphere of western North America: GSA Today, 8, 1-17.
Hoink, T., and A. Lenardic (2008), Three-dimensional mantle convection simulations with a low-viscosity asthenosphere and the relationship between heat flow and the horizontal length scale of convection, Geophys. Res. Lett., 35(L10304), doi:10.1029/2008GL033854.
Hoink, T., and A. Lenardic (2009), Long wavelength convection, Poiseuille-Couette flow in the low-viscosity asthenosphere and the strength of plate margins, Geophys. J. Int., 180, 23-33, doi:10.1111/j.l365-246X.2009.04404.
Humphreys, E.D. (1995), Post-Laramide removal of the Farallon slab, western United States, Geology, 23, 987-990.
Humphreys, E.D., and Dueker, K.G. (1994a), Western U.S. upper mantle structure: J. Geophys. Res., 99,9615-9634.
Humphreys, E.D., and Dueker, K.G. (1994b), Physical state of the western U.S. upper mantle: J. Geophys. Res., 99, 9635-9650.
Humphreys E., E. Hessler, K. Dueker, E. Erslev, G.L. Farmer, and T. Atwater (2003), How Laramideage hydration of North America by the Farallon slab controlled subsequent activity in
67
the Western U.S, International Geology Review, 45, 575- 595.
Humphreys, E.D. (2009), Relation of flat subduction to magmatism and deformation in the western United States, Geological Society of America Memoirs, 204, 85-98.
Hyndman, R.D., C.A. Currie, and S.R Mazzotti (2005), Subduction zone backarcs, mobile belts, and orogenic heat: GSA Today, 15, 4-10.
Karato, S. (2006), Remote sensing of hydrogen in earth's mantle, in Water in Nominally Anhydrous Minerals, Reviews in Mineralogy and Geochemistry, vol. 62, edited by H. Kepler et ai, pp. 343-375, Mineral. Soc. Am., Chantilly, Va.
Karlstrom, K.E., and S.A. Bowring (1988), Early Proterozoic assembly of tectonostratigraphic terranes in southwestern North America, J. Geol., 96, 561-576.
Karlstrom, K.E., and G. Humphreys (1998), Influence of Proterozoic accretionary boundaries in the tectonic evolution of western North America: Interaction of cratonic grain and mantle modification events, Rocky Mountain Geology, 33, no. 2, 161-179.
Lastowka, L.A., A.F. Sheehan, and J.M. Schneider (2001), Seismic evidence for partial delamination model for Colorado Plateau uplift, Geophys. Res. Lett., 28, 1319-1322.
Lee, C.-T. A. (2005), Trace element evidence for hydrous metasomatism at the base of the North American Lithosphere and possible association with Laramide low-angle subduction, J. Geol, 113, 673-685, doi:l0.1086/449327.
Lee, C.-T., Q. Yin, R.L. Rudnick, and S.B. Jacobsen (2001), Preservation of ancient and fertile lithospheric mantle beneath the southwestern United States, Nature, 411, 69-73.
Levander, A., and M.S. Miller (2010), The lithosphere in the western U.S., Geology, submitted.
Levander, A., C.A. Zelt, and M.B. Magnani (2005), Crust and upper mantle velocity structure of the Southern Rocky Mountains from the Jemez Lineament to the Cheyenne Belt, in K. Karlstrom and G.R. Keller, eds., The Rocky Mountain Region: An Evolving Lithosphere, Geophysical Monograph Series 154, AGU, Washington, D.C., 293-308.
Li, A., D. W. Forsyth, and K.M. Fischer (2003), Shear velocity structure and azimuthal anisotropy beneath eastern North America from Rayleigh wave inversion, J. Geophys. Res., 108(B8), 2362.
Li, A., D.W. Forsyth, and K.M. Fischer (2005), Rayleigh wave constraints on shear-wave structure and azimuthal anisotropy beneath the Colorado Rocky Mountains, in The Rocky Mountain Region: An Evolving Lithosphere, eds Karlstrom, K.E. Keller, G.R., pp. 385^401, Geophys. Monogr., 154, AGU, Washington DC.
Li, Z.-X. A., C.-T. A. Lee, A. H. Peslier, A. Lenardic, and S. J. Mackwell (2008), Water contents in mantle xenoliths from the Colorado Plateau and vicinity: Implications for the mantle rheology and hydration-induced thinning of continental lithosphere, J. Geophys. Res., 113, B09210, doi: 10.1029/2007JB005540.
Lipman, P.W., HJ. Prostka, and R.L. Christiansen (1972), Cenozoic volcanism and plate-tectonic evolution of the Western United States. I. Early and Middle Cenozoic, Phil. Trans. R. Soc,
68
London Ser A, 271:117-148
Lipman, P.W. (1992), Magmatism in the Cordilleran United States: Progress and problems, in The Geology of North America, vol. G3, The Cordilleran Orogen, Conterminous U.S., edited by B. C. Burchfiel, P. W. Lipman, and M. L. Zoback, pp. 481-514, Geol. Soc. of Am., Boulder, Colo.
Mattie, P.D., K.C. Condie, and J. Selverstone (1997), Origin of the continental crust in the Colorado Plateau: Geochemical evidence from mafic xenoliths from the Navajo Volcanic Field, southwestern USA, Geochim. Cosmochim. Ac, 61, 2007-2021.
McCarthy J., and T. Parsons (1994), Insights into the kinematic and Cenozoic evolution of the Basin and Range-Colorado Plateau transition from coincident seismic refraction and reflection data, Geol. Soc. Am. Bull, 106, 747-759.
McKenzie, D. (1984), A possible mechanism for epeirogenic uplift, Nature, 307, 616-618.
McMillan, N.J., A.P. Dickin, and D. Haag (2000), Evolution of magma source regions in the Rio Grande Rift, southern New Mexico, Geol. Soc. Am. Bull, 112, 1582-1593.
McQuarrie, N., and C.G. Chase (2000), Raising the Colorado Plateau, Geology, 28, 91-94.
Meissner, R. and W.D. Mooney (1998), Weakness of the lower continental crust: a condition for delamination, uplift, and escape, Tectonophysics, 296, 47-60
Miller, D.M., T.H. Nilsen, and W.L. Bilodeau (1992), Late Cretaceous to Early Eocene geology of the United States Cordillera, in Burchfiel, B. C, Lipman, P. W., and Zoback, M. L., eds., The Cordillerian Orogen: Conterminous U.S.: Boulder, CO, Geological Society of America, Geology of North America, v. G-3, Plate 6.
Miller, M.S., Levander, A., Niu, F., and Li, A. (2009), Upper mantle structure beneath the Caribbean-South American plate boundary from surface wave tomography, J. Geophys. Res., 114, B01312.
Mitchell, B.J. (1995), Anelastic structure and evolution of the continental crust and upper mantle from seismic surface wave attenuation, Rev. Geophys., 33, 441-462.
Morency, C, and M.-P. Doin (2004), Numerical simulations of the mantle lithosphere delamination,/. Geophys. Res., 109, B03410, doi:10.1029/2003JB002414.
Morgan P., W.R. Seager, and M.P. Golombek (1986), Cenozoic thermal, mechanical and tectonic evolution of the Rio Grande rift, J. Geophys. Res., 91, 6263- 6276.
Morgan, P., and C. Swanberg (1985), On the Cenozoic uplift and tectonic stability of the Colorado Plateau, J. Geodyn., 3, 39-63.
Moucha, R., A.M. Forte, D.B. Rowley, J.X. Mitrovica, N.A. Simmons and S.P. Grand (2008), Mantle convection and the recent evolution of the Colorado Plateau and the Rio Grande Rift valley, Geology, 36, 439-442.
Moucha, R., A.M. Forte, D.B. Rowley, J.X. Mitrovica, N.A. Simmons, and S.P. Grand (2009),
69
Deep mantle forces and the uplift of the Colorado Plateau, Geophys. Res. Lett., 36, LI9310, doi:10.1029/2009GL039778.
Moschetti, M.P., M.H. Ritzwoller, F. Lin and Y. Yang (2010), Seismic evidence for widespread western-US deep-crustal deformation caused by extension, Nature, 464, 885-889.
Mutschler, F.E., E.E. Larson, and R. Bruce (1987), Laramide and younger magmatism in Colorado-New petrologic and tectonic variations on old themes, Colorado Sch. Mines Q. 82, 1-45.
Nolet, G., C. Coutlee, and R. Clouser (1998), Sn velocities in western and eastern North America, Geophys. Res. Lett., 25(10), 1557-1560.
Oetking, P., D.E. Feray and H.B. Renfro (1967), Geological highway map of the Southern Rocky Mountain region, U.S. Geol. Highway Map Ser., Map No. 2 1, American Association of Petroleum Geologists, Tulsa, Oklahoma, p. 1,900,800 approx..
Parsons, T., and J. McCarthy (1995), The active southwest margin of the Colorado Plateau: Uplift of mantle origin, Geol. Soc. Am. Bull., 107, 139-147.
Parsons, T., J. McCarthy, W.M. Kohler, C.J. Ammon, H.M. Benz, J.A. Hole, and E.E. Criley (1996), Crustal structure of the Colorado Plateau, Arizona: Application of new long-offset seismic data analysis techniques, J. Geophys. Res., 101, 11,173-11,194.
Riter, J., and D. Smith (1996), Xenolith constraints on the thermal history of the mantle below the Colorado Plateau, Geology, 24, 267-270.
Roy, M., J.K. MacCarthy, and J. Selverstone (2005), Upper mantle structure beneath the eastern Colorado Plateau and Rio Grande rift revealed by Bouguer gravity, seismic velocities, and xenolith data, Geochem. Geophys. Geosyst., 6, doi:10.1029/2005GC001008.
Roy, M., T.H. Jordan, and J.L. Pederson (2009), Colorado Plateau magmatism and uplift by warming of heterogeneous lithosphere, Nature, 459, 978-982, doi:10.1038/nature08052.
Saito, M. (1988), DISPER80: a subroutine package for the calculation of seismic normal-mode solutions, in: Seismological Algorithms, Computational Methods and Computer Programs, Academic Press, New York, 293-319.
Saleeby, J. (2003), Segmentation of the Laramide Slab - evidence from the southern Sierra Nevada region, Geol. Soc. Am. Bull., 115, 655-668.
Schmandt, B., and E. Humphreys (2010), Complex subduction and small-scale convection revealed by body-wave tomography of the western United States upper mantle, Earth Planet Sci. Lett., submitted.
Schutt, D. L., K. Dueker, and H. Yuan (2008), Crust and upper mantle velocity structure of the Yellowstone hot spot and surroundings, J. Geophys. Res., 113, B03310.
Selverstone, J., A. Pun and K.C. Condie (1999), Xenolithic evidence for Proterozoic crustal evolution beneath the Colorado Plateau. Geol. Soc. Am. Bull. I l l :590—606.
70
Sheehan, A.F., G.A. Abers, A.L. Lerner-Lam, and C.H. Jones (1995), Crustal thickness variations across the Colorado Rocky Mountains from teleseismic receiver functions, J. Geophys. Res., 100, 20,391-20,404.
Sheehan, A.F., C.H. Jones, M.K. Savage, S. Ozalabey, and J.M. Schneider (1997), Contrasting lithosphere structure between the Colorado Plateau and the Great Basin: Initial results from Colorado Plateau - Great Basin PASSCAL experiment, Geophys. Res. Lett., 242609-242612.
Shen, W., Y. Yang, M.H. Ritzwoller; J. Pulliam, and S.P. Grand (2009), The crust and upper mantle beneath the Rio Grande Rift and surrounding regions inferred from ambient noise and earthquake tomography, AGU Fall Meeting 2009.
Sine, C. R., D. Wilson, W. Gao, S.P. Grand, R. Aster, J. Ni, and W.S. Baldridge (2008), Mantle structure beneath the western edge of the Colorado Plateau, Geophys. Res. Lett., 35, L10303, doi:10.1029/2008GL033391.
Smith, D., J.N. Connelly, K. Manser, D.E. Moser, T.B. Housh, F.W. McDowell, and L.E. Mack (2004), Evolution of Navajo eclogites and hydration of the mantle wedge below the Colorado Plateau, southwestern United States, Geochem. Geophys. Geosyst., 5, 10.1029/2003GC000675.
Smith, D., and W.L. Griffin (2005), Garnetite xenoliths and mantle-water interactions below the Colorado Plateau, southwestern United States, J. Petrol, 46, 1901-1924.
Snelson, CM., T.J. Henstock, G.R. Keller, K.C. Miller, and A. Levander (1998), Crustal and uppermost mantle structure along the Deep Probe seismic profile, Rocky Mt. GeoL, 33, 181-198.
Snelson, CM., G.R. Keller, K.C Miller, H.-M. Rumpel, and C. Prodehl (2005), Regional crustal structure derived from the CD-ROM 99 seismic refraction/wide-angle reflection profile: the lower crust and upper mantle, in The Rocky Mountain Region—An Evolving Lithosphere: Tectonics, Geochemistry, and Geophysics, K. E. Karlstrom and G. R. Keller (Editors), American Geophysical Union, Washington, D.C, 271-291.
Snyder W.S., W.R. Dickinson, and M.L. Silberman (1976), Tectonic implications of space-time patterns of Cenozoic magmatism in the western United States, Earth Planet. Sci. Lett., 32, 9 1 -106.
Spencer, J.E. (1996), Uplift of the Colorado Plateau due to lithosphere attenuation during Laramide low-angle subduction, J. Geophys. Res., 101, 13,595-13,609.
Stockli, D.F., A.P. Bump, G.H. Davis, and K.A. Farley (2002), Preliminary (U-Th)/He thermochronological constraints on the timing and magnitude of early Tertiary deformation of the Colorado Plateau: Geological Society of America Abstracts with Programs, 34, no. 6, 321.
Takeuchi, H., and M. Saito (1972), Seismic surface waves, in: Methods in Computational Physics. Academic Press, New York, 11, 217-295.
Tarantola, A., and B. Valette (1982), Generalized non-linear problems solved using the least-squares criterion, Rev. Geophys. Sp. Phys., 20, 219-232.
Thompson, G.A., and M.L. Zoback (1979), Regional geophysics of the Colorado Plateau,
71
Tectonophysics, 61, 149-181.
Van der Lee, S., and G. Nolet (1997), Upper mantle S-velocity structure of North America, J. Geophys. Res., 102,22,815-22,838.
Van Wijk, J., Van Hunen, J. Goes, S. (2008), Small-scale convection during continental rifting: evidence from the Rio Grande rift, Geology, 36, 575-578.
Weeraratne, D.S., D.W. Forsyth, K.M. Fischer, and A.A. Nyblade (2003), Evidence for an upper mantle plume beneath the Tanzanian craton from Rayleigh wave tomography, J. Geophys. Res., 108,2427.
West, M., J. Ni, W. Baldridge, D. Wilson, R. Aster, W. Gao, and S. Grand (2004), Crust and upper mantle shear-wave structure of the southwest United States: Implications for rifting and support for high elevation, J. Geophys. Res., 109, B03309, doi:10.1029/2003JB002575.
Wilson, D., Aster, R., Ni, J., Grand, S., West, M, Gao, W., Baldridge, W.S., Semken, S. (2005a), Imaging the seismic structure of the crust and upper mantle beneath the Great Plains, Rio Grande Rift, and Colorado Plateau using receiver functions, J. Geophys. Res., 110, B05306, doi: 10.1029/2004JB003492.
Wilson, D., R. Aster, M. West, J. Ni, S. Grand, W. Gao, W. Baldridge, S. Semken, and P. Patel (2005b), Lithospheric structure of the Rio Grande rift, Nature, 433, 851-855.
Wolf, L. W., and J.J. Cipar (1993), Through thick and thin: a new model for the Colorado plateau from seismic refraction data from Pacific to Arizona crustal experiment, J. Geophys. Res., 98, 19881-19894.
Yang, Y., and D.W. Forsyth (2006a), Rayleigh wave phase velocities, small-scale convection, and azimuthal anisotropy beneath southern California, J. Geophys. Res., I l l , B07306, doi: 10.1029/2005JB004180.
Yang, Y., and D.W. Forsyth (2006b), Regional tomographic inversion of amplitude and phase of Rayleigh waves with 2-D sensitivity kernels, Geophys. J. Int., 166, 1148-1160.
Yang, Y., and M.H. Ritzwoller (2008), Teleseismic surface wave tomography in the western U.S. using the Transportable Array component of US Array, Geophys. Res. Lett., 35, L04308, doi:10.1029/2007GL032278.
Zandt, G., S.C. Myers, and T.C. Wallace (1995), Crust and mantle structure across the Basin and Range-Colorado Plateau boundary at 37°N latitude and implications for Cenozoic extensional mechanism,/. Geophys. Res., 100, 10,529-10,548, doi:10.1029/94JB03063.
Zhou, Y., F.A. Dahlen, and G. Nolet (2004), 3-D sensitivity kernels for surface-wave observables, Geophys. J. Int., 158, 142-168.
72