Deep structure beneath the Southern Rocky Mountains from
the Rocky Mountain Front Broadband Seismic Experiment
Arthur L. Lerner-Lam1, Anne Sheehan2, Steven Grand3, Eugene Humphreys4, Ken Dueker2,
Erin Hessler2, Hongsheng Guo5, Duk-Kee Lee6, and Martha Savage7
Lamont-Doherty Earth Observatory and Department of Earth and Environmental Sciences of Columbia University, Palisades, NY 10964, U.S.A.
2
CIRES and Department of Geological Sciences, University of Colorado at Boulder, Boulder, CO 80309, U.S.A.
3
Department of Geological Sciences, University of Texas at Austin, Austin, TX 78712, U.S.A.
4
Department of Geological Sciences, University of Oregon, Eugene, OR 97403, U.S.A.
5
formerly at: Lamont-Doherty Earth Observatory and Department of Earth and Environmental Sciences of
Columbia University, Palisades, NY 10964, U.S.A.
6
Polar Research Center, Korea Ocean Research and Development Institute, Seoul, South Korea
7
Institute of Geophysics, Victoria University, Wellington, New Zealand
1
ABSTRACT
The deployment of a two-dimensional array of broadband seismometers during the PASSCAL
Rocky Mountain Front experiment produced a teleseismic and regional event data set which
provides constraints on both the laterally averaged and laterally varying structure of the crust
and upper mantle beneath the Southern Rocky Mountains and Great Plains. Results from a
spectrum of seismological imaging and inversion techniques indicate that the western edge of
the North American tectosphere has been relocated from its position beneath the Paleozoic
passive margin to a position several hundred kilometers east of the Colorado Front Range. The
mantle and crustal structure presently beneath the Colorado Rockies suggest a laterally variable dynamic state, which, on average, must provide partial support for the high topography.
KEY WORDS: Seismology, structure of lithosphere and crust, Rocky Mountains.
INTRODUCTION
Large-scale seismic tomography images of the
North American plate show that the west to east tectonic transition between the Colorado Rockies and
western Great Plains is echoed by an equally dramatic transition in the velocity structure of the
upper mantle between the western United States and
the continental interior (e.g., Grand, 1994; van der
Lee and Nolet, 1997). While the crustal deformation within the Rockies is ascribed to the integrated
effect of a number of contractional events ending
with the Laramide orogeny, the relationship of this
deformation to the mantle transition, as well as the
involvement of the mantle in present-day regional
uplift and on-going extension are not well understood. Moreover, since the mantle transition marks
the western margin of a thick, cold mantle lithosphere keel beneath the North American craton, the
transition itself may provide insight into the evolution and stability of continental interiors.
The crustal deformation of the western United
States has been dramatic, long-lasting, and complicated (Karlstrom and Humphreys, this issue).
Throughout most of the Mesozoic, a volcanic arc
was active along the present Sierra Nevada, and
Antler, Sonoma, and Sevier fold-and-thrust belts
developed to the east. In the late Cretaceous, volcanism ceased in the Sierra Nevada, and the locus of
thrusting and magmatism shifted to the east of what
is now the Colorado Plateau. In the Rocky Mountains, the style of contraction was thick- rather than
thin-skinned and is termed the Laramide orogeny
(~75–50 Ma). As the Laramide orogeny waned in
the early Cenozoic, volcanic activity migrated west,
north, and south from the Rocky Mountains. Tec-
Rocky Mountain Geology, v. 33, no. 2, p. 199-216, 8 figs., October, 1998
199
LERNER-LAM ET AL.
tonic relief acquired during the Laramide orogeny
in the Rocky Mountains was subsequently beveled
down to a relatively gentle surface (Epis and Chapin,
1975; Dickinson et al., 1988; Pazzaglia and Kelley,
this issue). Regional uplift of 1–2 km in late Cenozoic time has been suggested (Epis and Chapin,
1975; Eaton, 1986, 1987), but the paleontological
corroboration for this inference (MacGinitie, 1953;
Axelrod and Bailey, 1976) has been challenged and
the uplift might be late Eocene (37–40 Ma; Gregory
and Chase, 1992, 1994).
Models for the evolution of the lithospheric
mantle that are compatible with this deformation
are only beginning to be formulated. While the foldand-thrust belts in and slightly east of the modern
Basin and Range province seem compatible with
“normal” contractile plate margins, the thickskinned Laramide orogeny, far from a plate boundary, has been difficult to understand. One popular
hypothesis is that the geometry of the subducting
slab changed in late Mesozoic time, becoming
subhorizontal and extending to the eastern edge of
the modern Rockies (e.g., Lipman et al., 1971; Bird,
1984). This in turn has led to the inference that there
has been wholesale transport of crust and upper
mantle from west to east (Bird, 1984, 1988). In this
particular model, the elevation of the High Plains
and Rocky Mountains is due to an increase in the
crustal thickness by lower-crustal flow during the
Laramide orogeny. Others have compared the shape
of the overall uplift to that of a mid-ocean ridge and
inferred that the elevations are thermally supported
from the mantle (e.g., Eaton, 1987), possibly reflecting the sinking of the subducted Farallon plate in
the Cenozoic (Bird, 1984). An entirely different
hypothesis is that there was not a horizontal slab
beneath North America (Molnar and Lyon-Caen,
1988), or at least not one transmitting forces into
the lithosphere (Livaccari, 1991), and that the
Laramide orogeny reflects transmission of stress
from the plate margin eastward through the lithosphere itself. Recent models offer a version of this
hypothesis and suggest that Laramide deformation
reflects collision at the west coast of north-translating arc terranes (Maxson and Tikoff, 1996).
A variety of geophysical data suggests that the
easternmost extent of western United States crustal
deformation may correlate with a lateral transition
in upper mantle properties. The recent uplift of the
North American mid-continent is centered on the
Rocky Mountain Front, which separates the uplifted
yet intact western Great Plains from the uplifted
and broken crust of the western United States. This
transition suggests a major strength contrast or con-
200
trast in effective flexural thickness somewhere beneath the western Great Plains (Bechtel et al. 1990).
Evidence for a seismic velocity contrast across this
transition was obtained by Cleary and Hales (1966),
who found that teleseismic P-waves recorded on the
craton arrived approximately 1 s earlier relative to
the surrounding crust. This study and the subsequent station-residual studies of Wickens and
Buchbinder (1980) (S-waves) and Dziewonski and
Anderson (1983) (P-waves) gave additional support
to the existence of the transition but lacked sufficient station density to resolve its sharpness. Grand
(1987, 1994) used a tomographic method to invert
several thousand hand-picked SH-wave residuals and
found a geographic pattern of upper-mantle shear
velocities that correlates with the patterns of station residuals and flexural thickness. This study,
though groundbreaking in the quality and amount
of data inverted, still suffers from the inherent
sparseness of the existing global network of seismographs. Recent work by van der Lee and Nolet
(1997) using surface-wave phase velocities, however,
provides additional confirmation of the transition.
Each of these studies places the largest lateral velocity gradients in the vicinity of the Rocky Mountain Front (RMF). If these gradients are taken as
proxies for the mechanical properties, thermal
structure, and chemical composition of the lithosphere and upper mantle, the implication is that a
major regional boundary is present in the mantle.
A second, though no less important, issue is the
correlation between localized or sub-regional variations in crustal deformation and the mantle structure below. Though regionally broad scale, the uplift
and extension centered on the Rockies is locally
heterogeneous, with variations associated with the
northern Rio Grande rift and other features. From
a crustal dynamics standpoint, it is important to
understand whether these variations are a consequence of heterogeneous reactivation of variable
preexisting crustal weaknesses, or are related to a
complex pattern of subcrustal mantle flow during
Laramide and/or Tertiary tectonism with relic structures observable at present.
Seismological imaging provides an effective
means for addressing these issues, but previous
imaging efforts have been limited by datasets with
insufficient ray-path density. Only a handful of stations from the old analog WWSSN (World Wide Standard Seismograph Network) were located close to
the RMF, and only one, station GOL near Golden,
Colorado, was actually on it. There were no stations
in eastern Colorado, western Kansas, or Nebraska.
The number of stations in the replacement digital
Rocky Mountain Geology, v. 33, no. 2, p. 199-216, 8 figs., October, 1998
DEEP STRUCTURE BENEATH THE SOUTHERN ROCKY MOUNTAINS
networks is even smaller. There were several activesource refraction experiments done across the Front
Range (e.g., as summarized by Prodehl and Lipman,
1989 and references therein), and these are useful
as constraints on crustal velocities and thicknesses.
It was in this context that the Rocky Mountain
Front Experiment was designed. The Program for
Array Seismological Studies of the Continental
Lithosphere (PASSCAL) of the Incorporated Research Institutions of Seismology (IRIS), funded by
the National Science Foundation, provides individual investigators use of a new class of portable,
autonomous, digital seismographs with sensitive
broadband seismometers. This facility is the ideal
tool for regionally dense imaging of the crust and
upper mantle, and is capable of recording both
teleseisms and regional earthquakes with very high
fidelity. Relatively large numbers of instruments
can be deployed in an array concentrated geographically over the structures of interest, and the array
dimensions, or aperture, and the interstation spacing can be designed according to a priori imaging
density and resolution criteria.
THE PASSCAL ROCKY MOUNTAIN FRONT
EXPERIMENT
The Rocky Mountain Front Experiment was
conducted in two phases. Ten seismographs were
deployed in a preliminary reconnaissance phase for
2.5 months during the summer of 1991. These were
oriented on an east–west line extending from eastern Utah to eastern Colorado at approximately 39°
20’ N latitude with a nominal inter-station spacing
of approximately 100 km. Subsequently, thirty-six
seismographs were deployed for nearly seven
months in 1992 for the main data acquisition phase
(Fig. 1). The instruments were arranged in a twodimensional pattern extending east–west from eastern Utah into western Kansas and north–south
between the state borders of Colorado, and included
reoccupation of most of the reconnaissance sites.
Three-component sensors capable of recording
teleseisms across a broad frequency band were installed at all sites, and the PASSCAL data loggers
were programmed to record both continuous broadband and triggered high-sample-rate data streams.
The latter programming is useful for high-fidelity
recording of regional earthquakes, which provides
average constraints on the structure of the uppermost mantle and crust. During the second phase of
the experiment, the array was deployed in a twodimensional rectangular pattern and 446 earthquakes with adequate body-wave phases were
recorded. Approximately 100 of these are classified
as regional events (located within 1000 km of the
array), and the rest are teleseisms.
There are several ways to process these data for
information on crust and upper-mantle structure.
Teleseismic body waves travel through the upper
mantle and crust beneath the receiver at steep
angles of incidence, and their traveltime anomalies are ascribable to the integrated velocity structure along the ray path. Tomographic inversion of
these observations for velocity structure yields good
lateral resolution of vertically integrated structure,
but vertical resolution is critically dependent on
the azimuthal distribution of events and their raypath crossing pattern beneath the array. Even if the
ray paths are well distributed in azimuth, the depth
above which the velocity structure can be resolved
is limited by simple geometry and the nominal
interstation spacing within the array.
Waveform analysis of the body waves is one way
of overcoming this limitation, at least in part. The
body of techniques known as “receiver-function” or
“boundary-interaction-phase” analysis relies on using portions of the three-component record as a reference signal against which the wave interactions
with near-receiver horizontal or slightly dipping
structure can be measured. Forward or inverse modeling of these body wave interactions can be used to
estimate the depth to major velocity discontinuities
and the interval velocities. These techniques are
particularly useful for estimating crustal thickness
and vertically averaged seismic velocities in the crust
and upper mantle. The available lateral resolution
depends once again on the nominal station spacing
within the array, while the vertical resolution largely
depends on the bandwidth over which the signal is
recorded and modeled successfully.
Neither of these techniques is particularly accurate in retrieving the velocity of the uppermost
mantle just beneath the Moho, a critical region for
discussing the role of the mantle in crustal deformation. The appropriate class of techniques for
examining this region involves the analysis of seismic waves propagating horizontally through it, such
as crustal phases from regional events and intermediate-period surface waves from teleseisms. The
analysis of these waves provides constraints on both
the absolute value of the seismic velocities as well
as the magnitude of velocity gradients, but the inferences are limited by the effect of lateral heterogeneities along the propagation path that cause
unmodeled lateral refraction. Thus these methods
are useful principally for constraining the regionally averaged velocity structure.
Rocky Mountain Geology, v. 33, no. 2, p. 199-216, 8 figs., October, 1998
201
LERNER-LAM ET AL.
FOR POSITION ONLY — 100%
Figure 1. Seismogr aph stations of Rocky Mountain Front PASSCAL experiment on gr ay-shaded map of t opographic
relief, 500 m cont our interval. Squares indicate station locations; stations labeled with st ation name ar e used in
receiver function analy sis (Sheehan et al., 1995). White dashed lines define st ate boundaries , whereas black dashed
lines define ph ysiographic provinces.
In this paper, we summarize the results of previously published analyses of the Rocky Mountain
data set. We will show how the combined analysis
provides constraints on the average regional structure and the smaller-scale variability superimposed
on it. Taken together, these studies provide the most
complete picture of crust and mantle velocity structure beneath the Southern Rocky Mountains.
Laterally averaged structure
One of the goals of this work is to establish where
and over what scale average mantle in the stable
continental keel gives way to average mantle beneath the tectonized western United States. To do
this, we first must show that the laterally averaged
structure beneath the east and west portions of the
RMF array can be characterized as “cratonic” or “tectonic” in comparison with other regional models.
Second, from the point of view of inversion theory,
202
we should find the location of the transition that
maximizes the difference between these averaged
structures. The station location density of the RMF
array is not well suited to a formal inversion for
this location, and so we proceed with a forwardmodelling analysis that seeks to demonstrate that
dispersion and traveltime measurements fall naturally into “cratonic” and “tectonic” populations. An
additional benefit of these regional models is that
they provide the background against which lateral
variation can be assessed. In simple terms, it is helpful to know whether observed lateral variation is
relative to an average tectonic mantle or an average cratonic mantle in addressing dynamic issues.
Localization of the transition
The dispersion of fundamental-mode Rayleigh
waves is a sensitive indicator of changes in average
upper mantle and crustal properties. Chen and
Rocky Mountain Geology, v. 33, no. 2, p. 199-216, 8 figs., October, 1998
DEEP STRUCTURE BENEATH THE SOUTHERN ROCKY MOUNTAINS
FOR POSITION ONLY — 56%
Lerner-Lam (1993) used a cross-correlation technique to measure the change of phase of Rayleigh
fundamental modes from station to station within
the RMF array, and obtained a map of interstation
phase velocities for periods greater than 10 s. Although there were significant lateral variations and
there is evidence for lateral refraction, Chen and
Lerner-Lam (1993) found a significant increase in
fundamental-mode phase velocities well to the east
of the Rocky Mountain Front. The dispersion measurements were grouped in two populations east and
west of the observed phase-velocity increase and
were averaged above 25 s periods to obtain regionalized dispersion curves. These curves were then
inverted for regionalized one-dimensional shearvelocity structures. The resulting models are shown
in Figure 2.
It is possible to test the resolution of small differences in velocity structure below a fixed depth
by inverting jointly the differences between regionalized dispersion for differences in velocity structure. The velocities are allowed to vary
independently above a fixed depth, but are required
to have the same variation below. As a series of inversion experiments are performed while varying
the fixed depth, and the change in variance reduction is monitored. The result is a bound on the minimum depth to which differences in regionalized
structure must persist. The bound is sensitive, however, to assumptions about the homogeneity of the
paths contributing to the regionalization. This
“squeezing” technique was developed by Lerner-Lam
and Jordan (1987) to test for differences between
continental and oceanic upper mantle.
Chen and Lerner-Lam (1993) applied this technique and found that the differences between the
east and west regionalizations could be confined to
depths above 180 km. The differences approach 9%
in the seismological lid (rheological lithosphere)
and decrease nearly monotonically with depth. A
quantitative examination of the interstation phase
velocities placed this transition 200–300 km east of
the Rocky Mountain Front.
The question of whether this transition is a formal lateral discontinuity, in the sense of a “step” in
the velocity model, or a gradual transition cannot
be addressed by this intermediate-period surfacewave analysis. However, the grouping of interstation
phase velocities into two populations suggests that
the transition must occur over only one or two
interstation lengths, that is, 100–200 km. Further
examination of refracted surface waves, or a search
for scattered phases, could resolve this issue.
Figure 2. Chen and Lerner-Lam’ s (1993) preferred “purepath” models f or the Southern Rocky Mountains (dotted
line) and eastern Color ado–western Kansas (solid line).
Laterally averaged structure of crust
and uppermost mantle
Higher-frequency surface waves and crustal
phases can be analyzed to constrain the laterally
averaged structure of the crust and uppermost
mantle. Guo (1995) analyzed the Pnl wavetrain and
short-period Rayleigh waves produced by the St.
George, Utah earthquake (September 2, 1992;
mb=5.7; Pechmann et al., 1994; Pearthree and
Wallace, 1992) and recorded on 15 RMF array stations stretching from western Colorado to western
Kansas. Pnl, described by Helmberger and Engen
(1980), is a combination of the mantle head wave
Pn and the crustal reverberation PL and is sensitive to laterally-averaged crust and upper mantle
structure (e.g.,Wallace, 1983; Shaw and Orcutt, 1984;
Clouser and Langston, 1990; Holt and Wallace,
1990). Figure 3 shows a typical Pnl waveform from
the St. George event recorded by the RMF array (station BTO).
The Pn portion of the waveform arrives first and
represents high-frequency seismic energy propagating in the uppermost mantle or the mantle-lid wave
guide. It has been modeled successfully both as a
“whispering gallery” phase (Menke and Richards,
Rocky Mountain Geology, v. 33, no. 2, p. 199-216, 8 figs., October, 1998
203
LERNER-LAM ET AL.
FOR POSITION ONLY — 74%
1980) and a mantle-lid refraction (Sereno and Orcutt,
1985). PL is the long-period wavetrain following Pn
(Fig. 3), characterized by normal dispersion, rapid
amplitude decay with range, and prograde elliptical
particle motion in the radial plane. Numerous observational and synthetic modeling studies (e.g.,
Helmberger and Engen, 1980; Wallace, 1983; Shaw
and Orcutt, 1984; Holt and Wallace, 1990; Clouser and
Langston, 1990) have demonstrated that PL can be
successfully modeled as the summation of multiply
reflected and converted P and SV energy trapped
within the crust with associated evanescent shear
energy loss to the mantle. Thus, the Pn portion of
Pnl provides constraints on uppermost mantle structure similar to the constraints provided by post-critical body waves in refraction profiles, whereas the
PL portion provides constraints on crustal structure
similar to a multimode surface wave. For example,
several studies (Hill, 1971; Clouser and Langston,
1990; Holt and Wallace, 1990) show that the gradient
in the upper mantle can profoundly affect the Pn
waveform, but has little or no effect on the PL por-
tion. A positive mantle gradient enhances Pn amplitudes, and vice versa, with respect to a constantvelocity lid. This single parameterization of Pnl has
been modeled, for example by Langston (1982), but
the ratio of Pn to PL amplitudes has been shown to
be a more robust measure of the mantle gradient
(Holt and Wallace 1990; Clouser and Langston, 1990,
Guo, 1995). The complexity of the waveform arises
from the constructive and destructive interference
of multiple reflections, refractions, and mode conversions, and its constraints on structure are best inferred through forward modeling with synthetic
seismograms.
Synthetics were calculated for one-dimensional
models using the vectorized wavenumber integration (reflectivity) scheme of Fuchs and Muller
(1971) and Muller (1985), which accounts automatically for all P-SV interactions. The synthetic seismograms were calculated in a slowness window of
0.04–0.35 sec/km and a frequency window of 0.01–
2.0 Hz. Guo (1995) assigned a representative attenuation of Qp = 1000 and 300, and Qs = 500 and 200,
Figure 3. Comparison betw een observ ed Pnl and Sn at RMF site B TO (bottom trace) and s ynthetic seismogr ams
computed f or differ ent models (see Guo [1995] f or complete e xplanation). Arrows point to features of the waveform
for which sensitivity tests were performed, and which ar e critical to the interpretation. Model m5 giv es the best fit,
and is the model fr om which RMF740 is deriv ed.
204
Rocky Mountain Geology, v. 33, no. 2, p. 199-216, 8 figs., October, 1998
DEEP STRUCTURE BENEATH THE SOUTHERN ROCKY MOUNTAINS
in the crust and mantle respectively. Data and synthetics were band-passed between 0.01 and 0.25 Hz.
Summary of Preferred Model results
Guo’s (1995) preferred model, RMF740, comprises a two-layer crust overlying a thin-lid/ thinLVZ upper mantle with a lid gradient of 0.0015 sec-1.
This model represents an average structure integrated along the path between the St. George earthquake and the RMF array. While the lateral
variability of these structures cannot be determined
with this analysis, model RMF740 represents an
adequate background against which smaller scale
variations can be assessed.
Figure 4 shows RMF740 plotted against other
significant one-dimensional structures obtained for
the western US and some cratonic regions. In Figure 4A, RMF740 is compared with two other P-wave
models obtained for the western United States:
model GCA (Walck, 1984) derived for the Gulf of
California spreading center and the more general
Basin and Range model T7 (Burdick and
Helmberger, 1978) derived from long-period-bodywave forward modeling. Figure 4B shows model
RMF740 plotted against the continental models S25
of Lefevre and Helmberger (1989), representing the
Canadian Shield, K8 of Given and Helmberger
(1980), representing northwest Eurasia, particularly
the Baltic Shield, and WCH of Beckers et al. (1994),
representing the mixed structures of western China.
Except for the case of GCA, which is derived from
mostly vertical ray paths, these models were derived
using seismological imaging techniques that, like
Guo (1995), average the lateral heterogeneity along
pseudo-horizontal propagation paths. WCH, K8 and
S25 typify the range of crust and upper mantle structures observed for stable continental interiors, while
T7 and GCA are representative of the upper mantle
structure seen beneath spreading crust.
A comparison of these models shows that the
upper mantle beneath the Southern Rocky Mountains is substantially different than the mantle beneath continental shields. Continental shield
models are characterized by very high Pn velocities (compressional velocity just beneath the Moho),
and, if present at all, relatively deep low-velocity
zones with small velocity decreases. RMF740, on the
other hand, is much closer to the models of
“tectonized” upper mantle (T7 and GCA), with similarly low values of the Pn velocity, thin lid, and shallow low velocity zone.
There are distinct differences among the models, however, which suggest that the average crust
and mantle beneath the Southern Rockies is distinct
from the average structure beneath the Basin and
Range. RMF740 has a faster Pn velocity (8.00 km/s
vs. 7.95 km/s) and a much smaller lid velocity gradient (0.15 vs. 0.5 km/s per 100 km) than T7. Moreover, the velocity difference between the lid and the
LVZ is smaller in RMF740 than T7, although the
difference in LVZ thickness is not well resolved by
the Pnl forward modeling procedure with these data.
Finally, the crust of RMF740 is thicker and has
higher velocity than the crust of the Basin and
Range.
Laterally varying structure
Summary of Receiver Function Analysis of Moho
converted phases
The coda of first-arriving, teleseismic P body
waves propagating upward through the upper
mantle and crust beneath the receiver contains
mode conversions between compressional and
shear energy caused by interactions with velocity
discontinuities. If these discontinuities are
subhorizontal, the mode conversions are large on
the radial component of the seismogram while the
vertical component remains relatively undisturbed.
Thus the vertical component provides a reference
for source mechanism, deep-mantle propagation
effects, and instrument filtering, against which the
perturbations to the radial component caused by
boundary interactions can be measured. “Receiver
function analysis” treats the radial component as
the linear convolution of the vertical reference signal with a time series, or “receiver function”, representing the near-receiver mode conversions.
Typically, the most significant mode conversions
are the conversion of P to S at the Moho, and the
subsequent first reverberation. Thus receiver function analysis is used widely to estimate crust thickness and structure (e.g., Phinney, 1964; Langston,
1977, 1989; and Owens et al. 1984, 1987).
With data from a two-dimensional deployment
such as the RMF experiment, receiver function
analysis provides a plan view of crustal thickness
variations, which can then be interpreted in terms
of the major structural trends exhibited in the geology. Sheehan et al. (1995) computed receiver functions at 20 Rocky Mountain Front deployment sites
for 12 teleseismic events with mb >5.5. The receiver
functions were interpreted in terms of depth to
Moho by computing synthetic receiver functions
(corrected for sediment cover) for a suite of crustal
thicknesses and determining the RMS misfit for
each event-receiver pair. The crustal thickness val-
Rocky Mountain Geology, v. 33, no. 2, p. 199-216, 8 figs., October, 1998
205
LERNER-LAM ET AL.
FOR POSITION ONLY — 62%
Figure 4. (a) Comparison of RMF740 with models deriv ed for active tectonic provinces: T7: Basin and R ange (Burdick
and Helmber ger, 1978); GCA: Gulf of Calif ornia spreading center (Walck, 1984). ( b) Comparison with models derived for more stable provinces: K8: northwest Eurasia (Given and Helmber ger, 1980); S25: Canadian Shield (Lefe vre
and Helmber ger, 1989); WCH: western China (Beck ers et al., 1994). Global a verage IASP9 1 (Kennet and Eng dahl,
1991) shown for comparison.
ues found at the RMS minimum for each event-receiver pair were then averaged over the event ensemble to produce a value for that receiver site.
Sheehan et al. (1995) performed a sensitivity analysis to show that for reasonable values of the compressional and shear velocities in the crust, the
estimates of crust thickness using this technique
are robust and have a standard deviation of about 2
km. The crustal thickness estimates thus obtained
are used to map the Moho topography beneath the
RMF array.
Figure 5 (Sheehan et al., 1995) shows an example
of the observed and synthesized receiver functions
computed for a single event, displayed as a pseudorecord section across the RMF array. It is apparent
immediately that the mode conversion at the Moho
does not vary significantly across the Rocky Mountain Front. To facilitate a geological interpretation,
Sheehan et al. (1995) mapped the receiver function
results onto four physiographic provinces with
the following results for regionally-averaged
mean crustal thickness: the Kansas Great Plains,
43.8 ± 0.4 km; the Colorado Great Plains, 49.9 ± 1.2
km; the Colorado Rocky Mountains, 50.1 ± 1.3 km;
and the northeast Colorado Plateau, 43.1 ± 0.9 km.
We will discuss the importance of these results
in the discussion section, but we point out that there
is essentially no variation in estimated crustal thick-
206
ness between the Colorado Great Plains and Rockies,
a result that is consistent with earlier refraction
studies (Pakiser, 1963; Jackson et al., 1963; Jackson
and Pakiser, 1965; Roller and Jackson, 1966; Prodehl,
1979; Prodehl and Pakiser, 1980).
Shear-wave tomography
Shear-wave tomography results from the Rocky
Mountain Front experiment are summarized in Lee
and Grand (1996). Traveltimes were obtained for 624
S-wave arrivals from teleseismic events from 30° to
120° epicentral distance. Traveltimes were measured
by cross-correlating the S-wave data with synthetic
seismograms computed using the WKBJ technique
(Chapman, 1978) and using the Harvard CMT solutions for the source mechanisms. The shear-wave
residuals are corrected for known crustal, sediment
and topographic variations (Sheehan et al., 1995)
and are thus sensitive to mantle shear-velocity variations beneath the Colorado array. Up to 5 s differences in shear-wave traveltimes are observed
between the central Rockies in Colorado and station PKS in western Kansas. These traveltime variations are nearly as large as those between the
Canadian Shield and the East Pacific Rise (Grand
and Helmberger, 1984). The shear-wave residuals are
inverted for shear-wave velocity variations from
Rocky Mountain Geology, v. 33, no. 2, p. 199-216, 8 figs., October, 1998
DEEP STRUCTURE BENEATH THE SOUTHERN ROCKY MOUNTAINS
FOR POSITION ONLY —100%
Figure 5. (a) Example of r eceiver functions from a single
event from Hokk aido, Japan, as determined f or eighteen
stations of the Rocky Mountain Front array. Stations ordered by longitude with westernmost st ation (HLD) at
top. Dominant arrival at 5–6 s is P conv erted to S at the
Moho (Ps). Sediment ary basin thicknesses ar e indicated
at the left. Station names and best fitting crust al thicknesses (measur ed fr om surf ace) are shown in the center
of the plot, and ar e italicized for stations with sedimentary basin thickness >1 km. ( b) Example of best-fitting
synthetic receiver functions for one-dimensional crust al
model (sediment ary basin layer is also included) f or
Hokkaido earthquak e. Station name and best-fitting
crustal thickness (measur ed from surf ace) shown to left.
Case where a local, rather than global, minima w as picked
identified b y *. Variations in sediment ary thickness preclude the simple use of hand-timing of Ps phase f or determination of crust al thickness, and synthetics which
include this layer (as shown here) must be used.
50- to 425-km depth using a back-projection
technique (Humphreys et al., 1984). The block
dimensions used in the inversion are 0.6 by 0.6
degrees in lateral dimension and 50 km thick in
the vertical direction. The tomographic models
show slow shear-wave velocities beneath the Rocky
Mountains relative to the Great Plains, extending
from the surface to 200 km depth (Fig. 6). The amplitude of the variation ranges from 9% at 50–100
km depth to 7% at 150–200 km depth. At greater
depths, lateral variations are smaller and not well
correlated with surface tectonics. The tomographic
images reveal three mantle structures beneath the
study area. Beneath the Rockies, the mantle is extremely slow. Intermediate velocities are found to
the west and east of the Rockies. At the Colorado–
Kansas border there is an abrupt increase in mantle
shear velocity, with faster velocities to the east.
Compressional-wave tomography
The data for the compressional-wave tomography
consist of approximately 2700 hand-picked P wave
arrival times from the Rocky Mountain Front experiment. The estimated error on the picks is 0.1 s, and
the range in traveltime residuals observed is 5 s. Relative residuals are created by subtracting the average
absolute residual for each event. Corrections for sedimentary basins, topography, and crustal thickness
variations are made to the traveltime residuals before inversion. The inversion consists of breaking the
study area into 50 x 50 x 30 km blocks, tracing rays
through this structure, and inverting using the SIRT
algorithm (Humphreys and Clayton, 1988).
The results of the compressional-wave inversion
reveal similar patterns to those found with the shearwave inversion (Fig. 7). The most prominent lowvelocity volume is found beneath the Rocky
Mountains in Central Colorado, with a transition
to higher velocities found beneath western Kansas.
The peak-to-peak velocity variation in the model at
150 km depth is 4%. The largest amount of heterogeneity is found in the upper 250 km of the upper
mantle. The tomographic model only explains
about 60% of the signal variance, which may indicate that unmodeled small-scale heterogeneity is
present.
Overview of tomographic results
If due solely to thermal effects, a 9% shear-wave
velocity variation implies an 800° v ariation in temperature (Nataf and Ricard, 1996). This is thought
to be unreasonably large, as this is nearly twice the
thermal variation predicted between oceans and
continents, and would also elevate the temperature
of the mantle rocks above their solidus and cause
massive melting. Given thermal expansion, a temperature variation of 800° would predict a density
anomaly more than double the estimate of Sheehan
et al. (1995) for the mantle density beneath the
Rockies needed to account for the topography and
gravity signature of the Rockies. The presence of
melt can have a profound effect on seismic velocity, and almost certainly is a contributor beneath
the Rockies. Assuming Faul et al.’s (1994) relation
between partial melt and shear velocity and density and Nataf and Ricard’s (1996) relation between
Rocky Mountain Geology, v. 33, no. 2, p. 199-216, 8 figs., October, 1998
207
LERNER-LAM ET AL.
FOR POSITION ONLY — 100% , COLOR
Figure 6. Four horizontal slices through a shear wave tomogr am, at depths 100, 150, 200, and 250 km (Lee and
Grand, 1996). Blue color s represent fast velocities, red represents slow velocities. Dotted lines ar e the Rocky Mountain Front and eastern boundary of the Color ado Plateau.
shear velocity and temperature, a model with
350 ° C difference in temperature between the
mantle beneath central Kansas and the Rockies as
well as a 1.5% partial melting of the shallow mantle
beneath the central Rockies predicts a shear velocity contrast of 8.8% and a density contrast of 1.1%
between the two regions, in close agreement with
our observations. Chemical heterogeneity may also
play a role and could lessen the amount of thermal
or melt heterogeneity required.
Using the scaling relations given in Humphreys
and Dueker (1994) a 4% P-wave velocity variation
implies a 640 ° C temperature anomaly. This is significantly less than the 800 ° C suggested by the
shear-wave tomography model. Some differences in
model amplitudes may be due to the parameteriza-
208
tion of the models, with larger amplitude velocity
variations needed if heterogeneity is confined to a
smaller depth interval, or smaller amplitudes if a
larger depth interval is used. Another possibility is
an anelastic contribution to the velocity variation,
which, for an attenuation parameter, Q, of 50, would
halve the thermal variation required (Karato, 1993).
For a given thermal variation, Q for shear waves
(Qs) is smaller than Q for compressional waves
(Qp). Therefore, following Karato (1993), the anelastic contribution to the variation in velocity for a
given thermal anomaly is expected to be larger for
shear waves than for compressional waves. This may
help explain the large shear-velocity anomalies relative to compressional-velocity anomalies beneath
the Rocky Mountains.
Rocky Mountain Geology, v. 33, no. 2, p. 199-216, 8 figs., October, 1998
DEEP STRUCTURE BENEATH THE SOUTHERN ROCKY MOUNTAINS
FOR POSITION ONLY — 100%, COLOR
Figure 7. Four horizontal slices through a compr essional-velocity tomogr am, at depths of 120, 150, 180, and 2 10 km.
Eastward increasing velocities belo w the Colorado–Kansas bor der are displaced some what east from the shear-v elocity results. The correlation with the shear-tomogr am suggests laterally varying P/S r atios is the mantle.
Regions of the study area with high heat flow
(Morgan and Gosnold, 1989; Blackwell and Steele,
1992) correlate strongly with regions where we observe low velocities. While crustal heat production
likely contributes to part of the heat-flow anomaly
observed, the heat-flow anomaly is of such significant magnitude (60 mWm-2) that a mantle contribution ascribable to conduction, convection, or
both, is very likely.
Summary of measurements of mantle anisotropy
It is well known that shear waves propagating
through a transversely anisotropic but otherwise homogeneous medium perpendicular to its axis of symmetry split into orthogonally polarized components
which travel at different speeds. Once these compo-
nents have passed through the region of anisotropy,
the time delay between them is proportional to the
magnitude of the anisotropy and thickness of the
anisotropic layer, while the azimuth of one of these
components is parallel to the symmetry axis of the
anisotropy. In the mantle, transverse anisotropy is
thought to arise from the lattice-preferred orientation of olivine resulting from finite strains
(McKenzie, 1979) or stress-mediated recrystallization.
In this circumstance, the axis of the fast polarization
lies along the symmetry axis of the anisotropy. As a
consequence, measurements of splitting of shear
body waves are used as proxies for the orientation of
olivine fabrics and hence as an indication of the orientation and strength of mantle deformation and
mantle flow. Obviously, only the present-day fabric,
Rocky Mountain Geology, v. 33, no. 2, p. 199-216, 8 figs., October, 1998
209
LERNER-LAM ET AL.
which is a result of the total strain, is observed. Circumstantial arguments and/or finite-strain flow
models must be developed to constrain the time history. See Silver (1996) for a recent review.
Shear-wave splitting parameters were determined for RMF array sites by Savage et al. (1996),
who examined core (SKS and SKKS) and body
phases (S) from 28 events recorded during both the
first and second phases of the experiment. The two
horizontal components of each shear-arrival complex were digitally combined and subsequently rotated and time-shifted in a two-dimensional
parameter grid until the waveform distortion caused
by splitting was minimized. The azimuth of the fast
polarization, φ, and the time delay between the fast
and slow polarizations, δt, were estimated along
with formal uncertainties. Savage et al. (1996) also
reported null measurements, that is, a deterministic observation of an unsplit arrival.
The interpretation of φ and δt in terms of mantle
anisotropy requires a choice of parameterization
for the medium. An effective parameterization,
assumed by Savage et al. (1996), localizes transverse
(hexagonal) anisotropy with a horizontal axis of
symmetry within a single homogeneous layer.
(Other parameterizations are possible: see Silver
and Savage, 1994; Levin et al., 1996.) The utility of
this parameterization lies in its role as a natural
proxy for the horizontal component of mantle flow,
as it is thought that the finite strain associated with
mantle flow results in the olivine constituent having a preferred orientation (McKenzie, 1979). Since
the incoming S, SKS, and SKKS phases are propagating through the upper mantle with near-vertical incidence, their splitting will be optimally
sensitive to the horizontal component of the olivine fabric. In this model, φ will be parallel to the
horizontal symmetry axis, and δt will be proportional to the thickness of the layer and the volume
fraction of oriented olivine.
Savage et al. (1996) found splitting parameters
varying among the array sites, as well as several null
measurements (Fig. 8). The interpretation of null
measurements of shear-wave splitting is particularly
interesting although somewhat subjective. A null
measurement could indicate isotropy, but it is also
consistent with transverse isotropy having a vertical symmetry axis that might arise from vertical
flow. In recent work, Saltzer et al. (1997) showed
that null measurements may also be consistent with
a stack of thin layers having randomized fabric,
orthorhombic fabric, or scattering heterogeneities.
While any of these scenarios would constitute
a useful null hypothesis for interpreting null split-
210
ting observations, the results of Savage et al. (1996)
suggest the presence of rapidly varying anisotropy
beneath the RMF array. If measurable finite strain
has thus occurred in the mantle beneath the RMF
array, it may be more reasonable to assume that
anisotropy is present rather than absent. For this
reason, Savage et al. (1996) interpret null measurements to indicate vertical fabric in this region.
The average δt found by Savage et al. (1996) is
0.7 s, which is consistent with an anisotropic path
length of 50 to 150 km using observed bounds on
the aligned volume fraction of olivine obtained
from kimberlite nodules (Mainprice and Silver,
1993). The lower bound is consistent with confinement of the anisotropy to the lithosphere or crust,
or consistent with a longer path length through a
more heterogeneous anisotropic structure.
DISCUSSION
The strength of an experiment configuration
such as the RMF array lies in the ability to make
hierarchical inferences based on the different spatial scales resolved by various observations and
modeling techniques. The RMF array provides data
constraining both the regional geophysical context
and the intra-regional variation needed for largescale geological interpretation. It is useful to examine these constraints beginning with the largest
spatial scales.
There are two persistent problems in understanding the large-scale tectonics of the western
United States: First, what is the role of the mantle
in the series of orogenies culminating in the
Laramide? Second, what is the source of the regional-scale uplift across most of the region (Eaton,
1987; Gregory and Chase, 1992, 1994). The results
of the analyses of the RMF data summarized in this
paper support the view that the mantle has participated in past deformation, and continues to contribute to the regional uplift. There are several lines
of argument.
The laterally-averaged shear velocities found
beneath most of the RMF array are much lower than
the shear velocities found beneath cratons. There
is a large body of geophysical and geochemical evidence indicating that stable continental cratons are
underlain by “keels” of depleted mantle (the
“tectosphere”; see Jordan et al. [1989] for a summary
discussion) which is stabilized against advective
disruption by compositional differences with the
surrounding, “normal” mantle. This cratonic mantle
is characterized by high average velocities which
persist to depths greater than 200 km and often
Rocky Mountain Geology, v. 33, no. 2, p. 199-216, 8 figs., October, 1998
DEEP STRUCTURE BENEATH THE SOUTHERN ROCKY MOUNTAINS
FOR POSITION ONLY — 89%
Figure 8. Shear wave splitting measurements fr om the Rocky Mountain Front experiment [ Savage et al., 1996]. A
positive measur ement is a single thick dark line oriented par allel to fast direction of polariz ation φ, with length
corresponding to delay time δt in seconds (scale on lo wer right). Null measur ements are plotted as two small crossed
lines with directions equal to the two allowed orientations of anisotr opy. All measur ements are plotted at the 220 km
depth projection of the split r ay. Small open cir cles denote seismogr aph station locations. Thin solid lines define
state boundaries , black dashed lines define ph ysiographic provinces.
deeper beneath the Archean cores (e.g., Grand, 1987,
Lerner-Lam and Jordan, 1987). The low velocities
found on average beneath Colorado are not consistent with cratonic mantle.
The surface-wave and tomographic studies summarized above place the transition between Rocky
Mountain mantle and cratonic mantle approximately 200 km east of the Rocky Mountain Front,
and suggest that this transition persists to at least a
depth of 180 km. A similar transition in mantle
structure beneath the eastern United States has been
detected some 400-450 km west of the North Atlantic passive margin (Fischer, personal communication,
1997; also Grand, 1994; and van der Lee and Nolet,
1997). Together, these may be interpreted as the
western and eastern edges of the North American
tectosphere. If a similar scaling operated in the Paleozoic beneath the western United States passive
margin, which extended from the Sierra Nevada
batholith on the west to the Colorado Plateau on the
east, the western edge of the North American
tectosphere would have extended to the vicinity of
the western edge of the Colorado Plateau. This is
supported by the observation that the sedimentary
hingeline of the Cordilleran miogeocline was at the
western edge of the Colorado Plateau, suggesting that
the plate was thinned at least this far east. The RMF
experiment results for present-day mantle structure
show that to first order, this presumed correlation
between the pre-orogenic Paleozoic passive western
margin and craton-like upper mantle structure has
been lost, and that the tectosphere boundary has
migrated eastwards since the Paleozoic. We suggest
that the lateral mantle transition observed beneath
western Kansas and eastern Colorado represents the
“new” western boundary of the North American
tectosphere, and that this new boundary has been
relocated on the order of several hundred kilometers eastward by the successive actions of the cordilleran orogenies and the subsequent uplift. This
boundary thus is a marker for the easternmost extent of mantle modification during Mesozoic–Cenozoic deformation in the western United States.
An outstanding question thus concerns the
mechanisms that relocated the western mantle margin of the tectosphere. A complete discussion is be-
Rocky Mountain Geology, v. 33, no. 2, p. 199-216, 8 figs., October, 1998
211
LERNER-LAM ET AL.
yond the scope of this work, but there are three somewhat related conceptual candidates: (1) advective
removal of previously resident continental
tectosphere through the action of subhorizontal subduction (Bird, 1984, 1988), (2) advective removal of
tectosphere through large-scale vertical transport,
lithosphere delamination, or the development of
Rayleigh–Taylor instabilities (Houseman and Molnar,
1997), and (3) convective disruption of the
tectosphere through the action of accelerated thermal or chemical diffusive mechanisms (Shapiro et
al., 1991). In the first and second cases, the location
of the mantle boundary 200 km east of the easternmost extent of Laramide crustal deformation suggests
a scale length for the distance between the decoupling
of the lithosphere from crustal deformation and its
interaction with the remaining cratonic lithosphere.
In the third case, the thickness of the lateral transition between Rocky Mountain mantle and cratonic
mantle should give a sense of the scale of the diffusive processes. Whatever the mechanism, the resulting forces were strong enough to overcome whatever
processes stabilized the tectosphere. It will be important to examine this boundary, and other tectosphere
margins, in detail. Nevertheless, the implication of
these findings is that the regional upper mantle beneath the Rockies has undergone substantial modification since the time of the Paleozoic passive margin,
presumably through mechanisms that produced the
successive contractional orogenies through the
Laramide and lithospheric thinning in the Cenozoic.
Whether this location represents a translation of the
pre-orogenic boundary east, or a marker for the easternmost extent of tectosphere removal is not directly
constrained by RMF observations. However, the
anisotropy observed at station CES, in western Kansas, is not consistent with finite compressional strain
in the mantle along an east–west azimuth. Thus we
do not favor the translational explanation.
Regardless of the details of the mechanism that
removed the Paleozoic tectosphere, the average residual mantle and crust structure observed at
present suggests continued mantle involvement in
western United States dynamics. It is useful to examine whether the average mantle structure argues
for any particular model. The surface-wave analysis yields models that are consistent with models of
mantle beneath other tectonically active regions,
but the Pnl regional phase analysis summarized
earlier provides more detail. The analysis of these
phases makes a case for an average structure with a
mantle lid with positive-downward velocity gradient, which is not consistent with models of midocean ridges or the Gulf of California (model GCA,
212
Walck, 1984), which predict little or no lid, and nil
or negative gradients. On the other hand, at a thickness of 30 km, the RMF mantle lid is less than half
the thickness of the mantle lid under stable continent, where it is found (e.g., Lefevre and
Helmberger, 1989) or beneath Phanerozoic platforms. Equivalent lid thicknesses can be found beneath oceanic plates that have cooled conductively
for approximately 10–20 m.y. (Leeds, 1975; Schlue
and Knopoff, 1977; Nishimura and Forsyth, 1988;
Jordan et al., 1989), but neither the conductive time
constant nor the implied subsidence is consistent
with late Eocene uplift. Seismological lids are also
found beneath currently active continental collisional orogenies, such as the southern Himalaya
(Molnar and Lyon-Caen, 1988) or in complicated
regions such as western China (Beckers et al., 1994).
Although in some cases these lids may be
orogenically thickened, these structures are in almost all cases much thicker than 30 km. It is difficult to draw robust conclusions from any single
path, but we suggest that the Pnl model, with its 30km lid and similarities to other regional tectonic
models, rules out coherent, shallow, unidirectional
spreading and subsidence. At the same time, it is
incompatible with excessively thickened lithosphere resulting from collision or contraction. Thus
this model argues indirectly for a mantle in a transitional dynamic state rather than steady-state in
the long-term. As we discuss below, however, some
elements have been quasi-static since the late
Eocene.
The evidence that this transitional structure
plays an active role in the current dynamics of the
Rockies and western Great Plains is less circumstantial. The crust thicknesses derived by Sheehan et
al. (1995) fall into two populations, with a thicker
crust of approximately 50 km underlying the
Rockies and Colorado Great Plains, and thinner
crust of 43–44 km beneath the northeast Colorado
Plateau and the Kansas Great Plains. The main
variation in crustal thickness is thus between western Kansas and eastern Colorado, and not between
the Colorado Great Plains and the Rocky Mountains.
Noting the observation that much of the western
United States, including the Rockies, is in isostatic
equilibrium, Sheehan et al. (1995) argued that to
first order, the crustal thickness results imply that
Rocky Mountain topography cannot be compensated by Airy-type crustal roots. An additional,
quantitative analysis by Sheehan et al. (1995) incorporating the gravity signature demonstrated this
more precisely, concluding that compensation of
Rocky Mountain topography can not be accom-
Rocky Mountain Geology, v. 33, no. 2, p. 199-216, 8 figs., October, 1998
DEEP STRUCTURE BENEATH THE SOUTHERN ROCKY MOUNTAINS
plished solely in the crust, and requires a mantle
component. Sheehan et al. (1995) calculated that
about half of the excess topography in the Rocky
Mountains is supported by the mantle.
There are several possible mechanisms for this
excess support, including thermally- and compositionally-driven buoyancy, and vertically oriented
dynamic forces. The results of the shear- and compressional-wave tomography show a large amount
of small-scale variability with the implication that
there may be several mechanisms operating at once.
However, both the shear and compressional tomography show a large region of relatively slow velocities centered on the Southern Rockies in the vicinity
of the Rio Grande rift near Leadville. The magnitude of the anomaly is too large to be consistent
with a compositional origin and is probably of thermal origin. Moreover, Savage et al.’s (1996) measurements of anisotropy are null in this region, but are
arrayed systematically around the locus of low velocities and are similar to measurements made elsewhere in the northern Rio Grande rift (Sandvol et
al., 1992). This region also correlates with high heat
flow, as noted above. Taken together, these observations are consistent with a spatially restricted zone
of vertical dynamic support beneath the southern
Rocky Mountains near the Rio Grande rift. It is possible that this localized vertical support could be
manifested through vertical flow, but this is not
demonstrated by our data, and does not preclude
other mechanisms operating elsewhere. As noted
by Savage et al. (1996), the near isostatic compensation of the Rockies argues for mantle flow driven
by shallow deformation rather than the reverse, at
least on a regional scale. This issue could be resolved
by dense lateral sampling and estimation of the
depth variation of anisotropy.
This excess buoyancy most likely has been a
quasi-steady state feature of the dynamics since the
late Eocene, if the conjecture by Gregory and Chase
(1992, 1994) on the timing of the topographic uplift is correct. This timing would also would argue
against a single thermal pulse of regional scale because, as suggested earlier, the subsequent topographic subsidence through conductive lithospheric
cooling is not observed. It is entirely possible, however, that smaller scale features could be the result
of successive thermal pulses and conductive cooling, although these may be below the resolving limit
of this experiment. Under these circumstances, it
is not clear what drives the passive flow, for large
amounts of post late-Eocene extension are not observed. The flow may be a residual effect of the
disruption of the tectosphere, or a “back-flow” asso-
ciated with continued Rayleigh-Taylor instabilities
of the sort proposed by Houseman and Molnar
(1997).
In circumstances such as those operating beneath the Rockies, it is legitimate to ask whether
the standard regionalized notions of lithosphere and
asthenosphere make sense. The removal of large
amounts of continental tectosphere by an unspecified mechanism may leave behind a mantle which
does not exhibit steady-state dynamics at the local
level. The motion is never coherent enough laterally to cause further deformation (extension or contraction) in the crust. However, the motion in the
aggregate provides enough vertical support so that
the uplift is regionalized and appears to be the consequence of steady-state flow. This implies that there
must be a balance between the net upward intrusion of asthenosphere and the net downward advection of lithosphere. Perhaps better spatial
resolution of the uplift history of the southern
Rockies will provide constraints on this balance.
CONCLUSIONS
The strength of an array deployment such as the
Rocky Mountain Front Experiment is its ability to
constrain features with scale lengths comparable to
the dimensions of the array as well as the nominal
station spacing. When analyzed by a variety of techniques, constraints on both the structure and the
dynamics can be developed.
Analysis of body-wave-traveltime and surfacewave-dispersion measurements show that a fast-velocity mantle tectosphere presumably present
beneath the western United States during the Paleozoic has been transformed in the Rocky Mountains
into a velocity structure having on average a 30-km
lid and slow velocities. The transition between this
structure and the cratonic upper mantle underlying stable North America is constrained to occur
beneath the Colorado-Kansas border and appears to
be sharp. The transition represents the western edge
of the North American tectosphere and is a marker
for the eastern limit of upper mantle deformation
associated with the western United States contractional orogenies and subsequent uplift. This eastern limit of mantle deformation appears to be
displaced from the eastern limit of Laramide crust
deformation by approximately 200 km.
The most significant feature appears to be a
restricted zone of passive upwelling beneath the
northern Rio Grande rift, in south-central Colorado.
This upwelling supplies a mantle mechanism for
partially supporting excess topography in the
Rocky Mountain Geology, v. 33, no. 2, p. 199-216, 8 figs., October, 1998
213
LERNER-LAM ET AL.
Rockies, which cannot be solely supported by
mechanisms confined to the crust. The magnitude
of the velocity anomalies in this upwelling zone
require the presence of partial melt. A significant
problem is bracketing the melt zone both laterally
and vertically.
Smaller scale features are observed in the tomographic models and the variation of anisotropy,
but it is not apparent whether these features are
correlated with smaller scale features in the crust.
An unresolved problem is whether these small-scale
variations represent dynamics operating in concert
to provide the source of the post–late-Eocene uplift.
ACKNOWLEDGMENTS
The authors thank K. Karlstrom, R. Johnson,
and G. R. Keller for reviews which greatly improved
the manuscript. Discussion with K. Fischer, M.
Anders, and K. Gregory were instrumental in formulating the conclusions. L. Gao assisted in the Pnl
analysis. The PASSCAL Instrument Center at
Lamont-Doherty provided technical support for the
RMF experiment deployments. We are especially
grateful to T. Boyd and P. Romig of the Colorado
School of Mines for providing laboratory and computer network support during the experiment. C.
Ebeling, J. Gaherty, S. Shapiro, and especially R.
and M. Palmer provided expert and cheerful assistance in the field. D. Jones, J. Lawrence, and C.
Bryant participated as undergraduate interns. This
work supported by NSF EAR90-18424 and EAR9305282. Lamont-Doherty contribution number 5795.
REFERENCES CITED
Axelrod, D. I., and Bailey, H. P., 1976, Tertiary vegetation, climate, and altitude of the Rio Grande depression, New
Mexico-Colorado: Paleobiology, v. 2, p. 235-254.
Bechtel, T. D., Forsyth, D. W., Sharpton, Y. L., and Grieve, R.A.F.,
1990, Variations in effective elastic thickness of the North
American lithosphere: Nature, v. 343, p. 636-638.
Beckers, J., Schwartz, S. Y., and Lay, T., 1994, The velocity structure of the crust and upper mantle under China from broadband P and PP waveform analysis: Geophysical Journal
International, v. 119, p. 574-594.
Bird, P., 1984, Laramide crustal thickening event in the Rocky
Mountain foreland and Great Plains: Tectonics, v. 3, p. 741758.
___ 1988, Formation of the Rocky Mountains, western United
States—A continuum computer model: Science, v. 239, p.
1501-1507.
Blackwell, D. D., and Steele, J. L., 1992, Geothermal map of North
America, continental-scale Map–006, The Decade of North
American Geology (DNAG): Boulder, Colorado, Geological
Society of America, scale 1:5,000,000, 1 sheet.
214
Burdick, L. J., and Helmberger, D. V., 1978, The upper mantle P
velocity structure of the western United States: Journal of
Geophysical Research, v. 83, p. 1769-1712.
Chapman, C. H., 1978, A new method for computing synthetic
seismograms: Geophysical Journal Royal Astronomical Society, v. 54, p. 481-518.
Chen, Q., and Lerner-Lam, A., 1993, Where is the western edge
of the North American craton?: EOS, Transactions of American Geophysical Union., v. 74, p. 423.
Cleary, J., and Hales, A. L., 1966, An analysis of the contractionals
of P waves to North American stations in the distance range
30° to 100° : Bulletin of the Seismological Society of America,
v. 56, p. 467-489.
Clouser, R. H., and Langston, C. A., 1990, Upper mantle structure of southern Africa from Pnl waves: Journal of Geophysical Research, v. 95, p. 17,403-17,415.
Dickinson, W. R., Klute, M. A., Hayes, M. J., Janecke, S. U., Lundin,
E. R., Mckittrick, M. A., and Olivares, M. D., 1988, Paleogeographic and paleotectonic setting of Laramide sedimentary
basins in the central Rocky-Mountain region: Geological
Society of America Bulletin, v. 100, p. 1023-1039.
Dziewonski, A. M., and Anderson, D. L., 1983, Travel times and
station corrections for P-waves at teleseismic distances: Journal of Geophysical Research, v. 88, p. 3295-3314.
Eaton, G. P., 1986, A tectonic redefinition of the Southern Rocky
Mountains: Tectonophysics, v. 132, p. 163-193.
___ 1987, Topography and origin of the southern Rocky Mountains and Alvarado Ridge, in Coward, M. P., Dewey, J. F., and
Hancock, P. L., eds., Continental extensional tectonics: Geological Society [London] Special Publication no. 28, p. 355369.
Epis, R. C., and Chapin, C. E., 1975, Geomorphic and tectonic
implications of the post-Laramide, late Eocene erosion surface in the Southern Rocky Mountains, in Curtis, B. F., ed.,
Cenozoic history of the Southern Rocky Mountains: Boulder, Colorado, Geological Society of America Memoir 144,
p. 45-74.
Faul, U. H., Toomey, D. R., and Waff, H. S., 1994, Intergranular
basaltic melt is distributed in thin, elongated inclusions:
Geophysical Research Letters, v. 21, p. 29-32.
Fuchs, K., and Muller, G., 1971, Computation of synthetic seismograms with the reflectivity method and comparison with
observations: Geophysical Journal Royal Astronomical Society, v. 23, p. 417-433.
Given, J. W., and Helmberger, D. V., 1980, Upper mantle structure of the northwestern Eurasia: Journal of Geophysical
Research, v. 85, p. 7183-7194.
Grand, S. P., 1987, Tomographic inversion for shear structure
beneath the North American Plate: Journal of Geophysical
Research, v. 92, p. 14,065-14,090.
___ 1994, Mantle shear structure beneath the Americas and surrounding oceans: Journal of Geophysical Research, v. 99, p.
11,591-11,621.
Grand, S. P. and Helmberger, D. V., 1984, Upper mantle shear
structure of North America: Geophysical Journal Royal Astronomical Society, v. 76, p. 399-438.
Gregory, K. M., and Chase, C. G., 1992, Tectonic significance of
paleobotanically estimated climate and altitude of the late
Eocene erosion surface, Colorado: Geology, v. 20, p. 581-585.
___ 1994, Tectonic and climatic significance of a late Eocene
low-relief, high-level geomorphic surface, Colorado: Journal of Geophysical Research, v. 99, p. 20,140-20,160.
Rocky Mountain Geology, v. 33, no. 2, p. 199-216, 8 figs., October, 1998
DEEP STRUCTURE BENEATH THE SOUTHERN ROCKY MOUNTAINS
Guo, H. S., 1995, Studies of earthquake dynamics and lithospheric
structure in the western United States [Ph.D. dissert.]: New
York, Columbia University, 192 p.
Lerner-Lam, A., and Jordan, T. H., 1987, How thick are the continents?: Journal of Geophysical Research, v. 92, p. 14,00714,026.
Helmberger, D. V., and Engen, G. R., 1980, Modelling the longperiod body waves from shallow earthquakes at regional
ranges: Bulletin of Seismological Society of America, v. 70,
p. 1699-1714.
Levin, V., Menke W., and Lerner-Lam, A., 1996, Seismic anisotropy in the northeastern US as a source of significant
teleseismic P traveltime anomalies: Geophysical Journal
International, v. 126, p. 593-603.
Hill, D. P., 1971, Velocity gradients and anelasticity for crustal
body wave amplitudes, Journal of Geophysical Research, v.
76, p. 3309-3325.
Lipman, P. W., Prostka, H. J., and Christiansen, R. L., 1971, Evolving subduction zones in the western United States, as interpreted from igneous rocks: Science, v. 174, p. 821-825.
Holt, W. E., and Wallace, T. C., 1990, Crustal thickness and upper
mantle velocities in the Tibetan Plateau region from the
inversion of regional Pnl waveforms: evidence for a thick
upper mantle lid beneath southern Tibet: Journal of Geophysical Research, v. 95, p. 12,499-12,525.
Livaccari, R. F., 1991, Role of crustal thickening and extensional
collapse in the tectonic evolution of the Sevier–Laramide
orogeny, western United States: Geology, v. 19, p. 1104-1107.
Houseman, G. A., and Molnar, P., 1997, Gravitational (RayleighTaylor) instability of a layer with non-linear viscosity and
convective thinning of continental lithosphere: Geophysical Journal International, v. 128, p. 125-150.
Humphreys, E. D. and Clayton, R. W., 1988, Adaptation of back
projection tomography to seismic travel time problems:
Journal of Geophysical Research, v. 93, p. 1073-1085.
Humphreys, E. D., and Dueker, K. G., 1994, Physical state of the
western United States upper mantle: Journal of Geophysical Research, v. 99, p. 9635-9650.
Humphreys, E. D., Clayton, R. W., and Hager, B. H., 1984, A tomographic image of mantle structure beneath southern
California: Geophysical Research Letters, v. 11, p. 625-627.
Jackson, W. H., Stewart, S. W., and Pakiser, L. C., 1963, Crustal
structure in eastern Colorado from seismic-refraction measurements: Journal of Geophysical Research, v. 68, p. 57675776.
Jackson, W. H., and Pakiser, L. C., 1965, Seismic study of crustal
structure in the southern Rocky Mountains: U. S. Geological Survey Professional Paper 525-D, D-85-D-92.
Jordan, T. H., Lerner-Lam, A., and Creager, K. C., 1989, Seismic
imaging of boundary layers and deep mantle convection,
in Peltier, W. R., ed., Mantle convection: Plate tectonics and
global dynamics, Gordon and Breach, Montreux, p. 97-201.
Karato, S., 1993, Importance of anelasticity in the interpretation of seismic tomography, Geophysical Research Letters,
v. 20, p. 1623-1626.
Kennet, B.L.N., and Engdahl, E. R., 1991, Traveltimes for global
earthquake location and phase identification: Geophysical
Journal International, v. 105, p. 429-465.
Langston, C. A., 1977, Corvalis, Oregon, crustal and upper mantle
receiver structure from teleseismic P and S waves: Bulletin
of Seismological Society of America, v. 67, p. 713-724.
___ 1982, Aspects of Pn and Pg propagation at regional distances: Bulletin of Seismological Society of America, v. 72,
p. 457-471.
___ 1989, Scattering of teleseismic body waves under Pasadena,
California: Journal of Geophysical Research, v. 94, p. 19351951.
Lee, D. K., and Grand, S. P., 1996, Upper mantle shear structure
beneath the Colorado Rocky Mountains: Journal of Geophysical Research, v. 101, p. 22,233-22,244.
MacGinitie, H. D., 1953, Fossil plants of the Florissant beds, Colorado: Carnegie Institution of Washington Publication 599,
198 p.
Mainprice, D., and Silver, P. G., 1993, Constraints on the interpretation of teleseismic SKS observations from kimberlite
nodules from the subcontinental mantle: Physics of the
Earth and Planetary Interiors, v. 78, p. 257-280.
Maxson, J., and Tikoff, B., 1996, Hit-and-run collision model for
the Laramide orogeny, western United States: Geology, v.
24, p. 968-972.
McKenzie, D., 1979, Finite deformation during fluid flow: Geophysical Journal Royal Astronomical Society, v. 58, p. 689715.
Menke, W. H., and Richards, P. G., 1980, Crust mantle whispering gallery phases: a deterministic model of teleseismic Pn
wave propagation: Journal of Geophysical Research, v. 85,
p. 5416-5422.
Molnar, P., and Lyon-Caen, H., 1988, Some simple physical aspects of the support, structure, and evolution of mountain
belts, in Clark, S. P., Burchfiel, B. C., and J. Suppe, J., eds.,
Processes in continental lithospheric deformation: Geological Society of America Special Paper 218, p. 179-207.
Morgan, P., and Gosnold, W. D., 1989, Heat flow and thermal
regimes in the continental United States, in Pakiser, L. C.,
and Mooney, W. D., eds., Geophysical framework of the continental United States: Boulder, Colorado, Geological Society of America Memoir 172, p. 493-522.
Muller, G., 1985, The reflectivity method: A tutorial: Journal of
Geophysics, v. 58, p. 153-174.
Nataf, H. C., and Ricard, Y., 1996, 3SMAC: An a priori tomographic model of the upper mantle based on geophysical
modeling: Physics of the Earth and Planetary Interiors, v.
95, p. 101-122.
Nishimura, C. E., and Forsyth, D. W., 1988, Rayleigh wave phase
velocities in the Pacific with implications for azimuthal
anisotropy and lateral heterogeneities: Geophysical Journal International, v. 94, p. 479-501.
Owens, T. J., Zandt, G., and S. R. Taylor, S. R., 1984, Seismic
evidence for an ancient rift beneath the Cumberland Plateau, Tennessee: A detailed analysis of broadband
teleseismic P waveforms: Journal of Geophysical Research,
v. 89, p. 7783-7795.
Leeds, A. R., 1975, Lithospheric thickness in the western Pacific:
Physics of the Earth and Planetary Interiors, v. 11, p. 61-64.
Owens, T. J., Taylor, S. R., and Zandt, G., 1987, Crustal structure
at regional seismic test network stations determined from
inversion of broadband teleseismic P waveforms: Bulletin
of Seismological Society of America, v. 77, p. 631-662.
Lefevre, L. V., and Helmberger, D. V., 1989, Upper mantle P velocity structure of the Canadian shield: Journal of Geophysical Research, v. 94, p. 17,749-17,765.
Pakiser, L. C., 1963, Structure of the crust and upper mantle in
the western United States: Journal of Geophysical Research,
v. 68, p. 5747-5756.
Rocky Mountain Geology, v. 33, no. 2, p. 199-216, 8 figs., October, 1998
215
LERNER-LAM ET AL.
Pearthree, P. A., and Wallace, T. C., 1992, The St. George earthquake of September 2, 1992: Arizona Geology, v. 22, p. 7-8.
Pechmann, J. C., Arabasz, W. J., and S. J. Nava, S. J., 1994, Refined analysis of the 1992 ML 5.8 St. George, Utah, earthquake and its aftershocks: Seismological Research Letters,
v. 65, p. 32.
Phinney, R. A., 1964, Structure of the Earth’s crust from spectral
behavior of long-period waves: Journal of Geophysical Research, v. 69, p. 2997-3017.
Prodehl, C. P., 1979, Crustal structure of the western United States:
U. S. Geological Survey Professional Paper 1034, 74 p.
Prodehl, C., and Lipman, P. W., 1989, Crustal structure of the
Rocky Mountain region, in Pakiser, L. C., and Mooney, W.
D., eds., Geophysical framework of the continental United
States: Boulder, Colorado: Geological Society of America
Memoir, v. 172, p. 249-284.
Prodehl., C. P., and Pakiser, L. C. , 1980, Crustal structure of the
southern Rocky Mountains from seismic measurements:
Geological Society of America Bulletin, v. 91, p. 147-155.
Roller, J. C., and Jackson, W. H., 1966, Seismic-wave propagation
in the upper mantle: Lake Superior, Wisconsin to Denver,
Colorado, in Steinhart, J. S., and Smith, T. J., eds., The Earth
beneath the continents: Washington, D.C., American Geophysical Union, Geophysical Monograph Series, v. 10, p. 270275.
Saltzer, R. L., Gaherty, J. B., and Jordan, T. H., 1997, What do
shear-wave splitting measurements mean when the anisotropy varies with depth? [abs.]: EOS, Transactions of the
American Geophysical Union, v. 78, F487.
Sandvol, E., Ni, J., Ozalaybey, S., and Schlue, J., 1992, Shear-wave
splitting in the Rio Grande rift: Geophysical Research Letters., v. 19, p. 2337-2340.
Savage, M. K., Sheehan, A. F., and Lerner-Lam, A., 1996, Shear
wave splitting across the Rocky Mountain Front: Geophysical Research Letters, v. 23, p. 2267-2270.
Schlue, J., and Knopoff, L., 1977, Shear-wave polarization anisotropy in the Pacific Basin: Geophysical Journal Royal Astronomical Society, v. 49, p. 145-165.
216
Sereno, T. J., and Orcutt, J., 1985, Synthesis of realistic oceanic
Pn wave trains: Journal of Geophysical Research, v. 90, p.
12,755-12,776.
Shapiro, S. S., Hager, B. H., and Jordan, T. H., 1991, Stability of
the continental tectosphere [abs.]: EOS, Transactions of the
American Geophysical Union, v. 72, p. 267.
Shaw, P., and Orcutt, J., 1984, Properties of PL and implications
for the structure of Tibet: Journal of Geophysical Research,
v. 89, p. 3135-3152.
Sheehan, A. F., Abers, G. A., Lerner-Lam, A., and Jones, C. H.,
1995, Crustal thickness variations across the Colorado Rocky
Mountains from teleseismic receiver functions: Journal of
Geophysical Research, v. 100, p. 20,391-20,404.
Silver, P. G., 1996, Seismic anisotropy beneath the continents:
probing the depths of geology: Annual Review of Earth and
Planetary Sciences, v. 24, p. 385-432.
Silver, P. G., and Savage, M. K., 1994, The interpretation of shearwave splitting parameters in the presence of two anisotropic layers: Geophysical Journal International, v. 119, p.
949-963.
van der Lee, S., and Nolet, G., 1997, Upper mantle S velocity
structure of North America: Journal of Geophysical Research, v. 102, p. 22,815-22,838.
Walck, M. C., 1984, The P-wave upper mantle structure beneath
an active spreading center: the Gulf of California: Geophysical Journal Royal Astronomical Society, v. 76, p. 697-723.
Wallace, T. C., 1983, Long period regional body waves [Ph.D.
dissert.]: Pasadena, California Institute of Technology, 185 p.
Wickens, A. J., and Buchbinder, G.G.R., 1980, S-wave residuals in
Canada: Bulletin of Seismological Society of America, v. 70,
p. 809-822.
MANUSCRIPT SUBMITTED DECEMBER 23, 1997
REVISED MANUSCRIPT SUBMITTED APRIL 15, 1998
MANUSCRIPT ACCEPTED JUNE 3, 1998
Rocky Mountain Geology, v. 33, no. 2, p. 199-216, 8 figs., October, 1998