Assembling a Community Velocity Model
for the Western Great Basin
John N. Louie, Nevada Seismological Laboratory, University of Nevada, Reno
December 1, 2001
Project Objectives and Expected Results
This project will assemble a three-dimensional reference seismic velocity model for
the western Great Basin region of Nevada and eastern California (figure 1). This model will
be rule-based and distributed as software on the Internet, similar to the SCEC Community
Velocity Model (CVM) of Magistrale et al. (2000). These qualities will make the model
useful for multidisciplinary research activities including earthquake-hazard assessment,
strong-motion modeling, basin amplification modeling, high-precision earthquake location
and source-parameter estimation, and crustal-structure imaging.
A faculty geophysicist and a graduate student will assemble this model almost
entirely from existing results. The two project years will both include literature review and
examination of gray literature and data donated by the mining, petroleum, and geotechnical
industries. We will assemble especially detailed information on Nevada's urban basins, and
on the variability of shallow geotechnical measurements both in basins and in rock. During
the first project year we will also conduct an inexpensive, 1- to 3-day seismic refraction
survey, 500 km long, between mine blasts and crossing the northern Sierra Nevada and
Walker Lane. In the second project year we may need to develop additional
receiver-function analyses from recently installed stations. We will also develop in the 2nd
year the rule-based model and adapt SCEC's Internet software to the region.
A Sept. 2000 FEMA report shows that Nevada ranks fifth of all the states in terms of
annualized losses from earthquakes. Nevada is a very urban state, so FEMA's HAZUS
analysis picked out the state's two metropolitan areas, Las Vegas and Reno, as most at risk.
Among metropolitan areas outside California, Las Vegas ranks 7th in risk and Reno ranks
11th. Nevada's average annualized losses are high. The model finds an average loss of about
$55 million dollars per year.
Such hazard requires detailed study of seismic hazards, precise earthquake locations,
path and site effects, and simulated strong ground motions. Accurate velocity structure in
and near basins is essential for correct modeling of basin-induced seismic resonances and
amplifications, as shown by Kawase and Aki (1989), Field (1996), Joyner (2000), Olsen
(2000), Olsen et al. (2000), and many others. The widely used USGS seismic hazard maps
(Frankel et al., 1996) are based on a 30-meter geotechnical velocity of 760 m/s; strong
amplifications or de-amplifications of earthquake shaking are expected for any sites having
significantly different velocity. Phillips and Aki (1986) and Su and Aki (1995) observed
coda amplified by velocity heterogeneity. Field and Jacob (1995), Seekins et al. (1996),
Steidl et al. (1996), Bonilla et al. (1997), Satoh et al. (2001b), and others have found
geotechnical velocities helpful in resolving differences among site-amplification estimation
techniques. Shear velocity is a fundamental base parameter in understanding nonlinear site
effects as well, as in Ni et al. (1997).
Anderson et al. (1994) required crustal velocity estimates to compute the levels of
dynamic strain, generated by the 1992 Landers event, that triggered earthquakes in the
western Great Basin. Unknown path effects are a significant source of error for efforts to
predict ground-motion attenuation curves (Spudich et al., 1999). Savage and Sheehan (2000)
require detailed velocity structure as a first input to creating mantle anisotropy models, and
to explain shear-wave splits from sites near strong crustal heterogeneities. Scientists and
engineers require a realistic three-dimensional seismic velocity model to understand
earthquake hazards in the western Great Basin. The type of rule-based representations
developed by SCEC are very appropriate to defining velocity on the spatial scales
appropriate to each application, particularly in the western Great Basin. Crustal velocities
and Moho depths are known only at wide (100 km) spacing, but the structure of the urban
basins is known at some detail (0.2 km spacing).
We will compile velocity information from sources in the literature, results of
previous seismic experiments and earthquake-monitoring projects, and data donated from
mining and petroleum companies. We will also conduct one new crustal refraction profile
between mine blasts (figure 1). Harder and Keller (2000) demonstrated that a 150-km-long
deployment of 300 PASSCAL Texan instruments could be completed in a single day, with
the efforts of seven crews of 2 or 3 people. They observed crustal P, Pg, PcR, PmP, and SmS
phases from a single ripple-fired mine blast. The proposed profile is three times as long as
Harder and Keller's (2000), and will be reversed between mines in the western Sierra
Nevada and west of Carlin (figure 1). However, the profile is centered on Reno and entirely
accessible from Interstate 80. Thus, the proposed budget only needs include wages for three
undergraduate helpers in addition to the PI and graduate assistant, and funds for 5 vehicles
for the 3 days we anticipate for layout and pick-up. The profile will also use 300 Texans; the
PI has been trained in their use during two previous experiments.
Aside from the single inexpensive refraction survey, most of the proposed effort will
be to assemble velocity information of several types and at several scales to define the
model at different depths:
Upper mantle— The tomographic image of Humphreys and Dueker (1994)
provides a starting framework for mantle velocity in the western Great Basin; although their
coverage north and east of Reno (figure 1) is poor. Dueker and Sheehan (1997) tracked
upper-mantle discontinuities across the Snake River Plain (figure 1) with long-period
Pn velocities— Thompson et al. (1989) review regional constraints on Pn velocities.
Our proposed profile across the northern Sierra and Walker Lane will provide some of the
constraints available for the southern Sierra and Death Valley from Fliedner et al. (1996).
Moho depth— Mooney and Braile (1989) and Kaban and Mooney (2001) reviewed
all available constraints on Moho depth for the western Great Basin (figure 1). We will find
out if Dueker and Sheehan (1997) also made shorter-period receiver functions more suitable
for estimating mantle depths in the northern Great Basin. For the central Great Basin
constrained receiver-function analyses are available from Ozalaybey et al. (1997). We will
likely need to develop additional receiver-function analyses from recently installed stations
in the Nevada Seismic Network. G. Biasi of UNR has collected several years of teleseismic
data from many stations; their sensors are adequate for integration accuracy to more than 20
seconds period, so the receiver-function analysis will not be terribly difficult. The proposed
refraction survey between mine blasts will provide Moho depth information across the
northern Sierra and Walker Lane (figure 1), where constraints are poorer than to the south
near Death Valley.
Middle & lower crustal velocities— Mooney and Braile (1989), Thompson et al.
(1989), and Fliedner et al. (1996) provide reviews of crustal velocity information that will
form a basis for a 3-d crustal velocity model. This model will be parameterized as functional
profiles at the locations of control points, with interpolation extending the model laterally
between controls. In the northern and eastern Basin and Range control may be sparse
enough that we will need to employ the CRUST 5.1 global model of Mooney et al. (1998).
Ozalaybey et al. (1997) constrained crustal velocity profiles at several locations in the
central Great Basin, establishing low-velocity zones exist at very few. We will assemble as
well published and unpublished studies of joint aftershock relocation and velocity inversion
such as by Asad et al. (1999) for the Eureka Valley sequence north of Death Valley (figure
Upper crust— Louie and Qin (1991) and Louie et al. (1997) used surface waves
and COCORP reflection surveys to constrain upper-crustal velocities west of Death Valley
(figure 1). The optimization methods of Pullammanappallil and Louie (1993; 1997) have
proved effective in obtaining velocities to 5 km depth from reflection surveys. We will pick
and optimize first-arrival and reflection times where needed from available COCORP and
industry data. In addition to crustal thickness, much of the work reviewed by Kaban and
Mooney (2001) constrains P velocities well to 5-10 km depth. Additional constraints are
reviewed by Thompson et al. (1989) and Fliedner et al. (1996); many of them come from
long COCORP surveys extending from the northern Sierra to the Ruby Mountains, and in
the Death Valley region (figure 1). We will compile all available information from reflection
stacking velocity analyses, and seek to examine copies of commercial spec surveys,
abundant in the Carlin gold trend.
Basin depths and velocities— Honjas et al. (1997), Chavez-Perez et al. (1998) and
Abbott et al. (2001) estimated basin depths and velocities for Death Valley and Dixie Valley
(figure 1) from the first-arrival times recorded in reflection surveys. Jachens and Moring
(1990) summarize relations between density and depth in Nevada basins from oil-well logs,
mostly from Railroad Valley (figure 1). Langenheim et al. (2001) and Abbott and Louie
(2000) used these relations together with some borehole and seismic data to detail the depths
and density profiles of Nevada's urban basins in Las Vegas, Reno, and Carson (figure 1).
We have already developed rule-based velocity models for the Las Vegas and Reno
basins, based on the published depth and density data, and viewable at
http://www.seismo.unr.edu/ftp/pub/louie/reno/basinseis.jpg. These models are expressed in
Java code. One measurement of shear velocity to the basement has been done in Reno
(Louie, 2001), using the method of Horike (1985), Liu et al. (2000), and Satoh et al. (2001a).
COCORP stacking velocities from basins will provide a few P-velocity constraints for
basins between Reno and Carlin, and near Death Valley; spec seismic data we are able to
view will gain us some data for the central Great Basin.
Unlike how Magistrale et al. (2000) created the SCEC CVM with rules for formation
depths and velocities gained from oil wells, Nevada's urban basins have far too few deep
boreholes. However, SCEC's CVM must be accurate for basins under compression with
kilometers of thrust deformation. All of Nevada's urban basins are principally extensional,
and still receiving sediment. We propose to form the western Great Basin CVM using
instead rules for depth within a basin, and possibly the basin's proximity to Tertiary volcanic
centers, and its age and subsidence rate. Controlling for these factors, as far as is possible,
will enable us to predict velocities within the Las Vegas and Reno basins from the Railroad
Valley density profiles, and the Death Valley and Dixie Valley velocity optimizations.
Geotechnical— Louie (2000) published some shallow geotechnical velocity
information on the Reno basin. We have continued to apply the refraction microtremor
technique in and around this basin, resulting in measurements at a dozen rock sites around
the basin. Many of these measurements were made at borehole sites with assistance from
local engineering consultants. During both project years we will seek out geotechnical data
from consultants working in both Las Vegas and Reno. We will seek out as well shallow
geophysical data such as the study between Death Valley and Las Vegas (figure 1) by
Shields et al. (1998).
On-going is an effort to measure a shallow shear-velocity profile extending 20 km
across the entire Reno basin. This refraction microtremor (Louie, 2001) effort is inexpensive
enough to be university-funded, is employing 45 PASSCAL Texan recorders, and is about
70% complete. Figure 2 shows preliminary results from this profile at 50% completion.
Within a floodplain on the east side of the basin, the 30-meter average velocity jumps by
50%, although no geomorphic features suggest this (the spike in the 100-m velocity is not
well constrained). The 10-meter velocity shows a significant increase there as well. Actual
measurement along the profile shows far more lateral velocity variation than soil mapping
would suggest. More information is available at www.seismo.unr.edu/hazsurv.
The velocity model we develop will consist of rule-based representations of the
region's urbanized basins (Las Vegas, Reno, and Carson) embedded in a 3-d crust over a
variable-depth Moho, as done by SCEC for southern California (Magistrale et al., 2000).
Shallow velocities will be constrained by geotechnical data; receiver-function and refraction
analyses will constrain Moho depths. The model will be specified in a form compatible with
computer codes developed for the SCEC Community Velocity Model, for dissemination
through similar means. The model will be available on the Internet for cross-discipline
seismic and earthquake-hazard studies. Aside from presentation of the model at one national
meeting per year, the budget also includes travel funds for the PI and/or student to attend
workshops on the SCEC CVM in southern California. This will assure that development of
the western Basin and Range CVM proceeds compatibly with the SCEC CVM and its
Results From Related Previous NSF Support of John Louie:
Geophysical test of low-angle dip on the seismogenic Dixie Valley fault, Nevada,
NSF EAR-9706255, 9/97-8/99 for $91,313 between 3 PIs, Louie 2 weeks.
Research Activities: Our field collection of geophysical data in Dixie Valley used equipment and
techniques similar to surveys we propose here. Shallow seismic refraction, reflection, gravity, and
electromagnetic measurements examined a normal fault. Surface-wave data collected along this profile helped
to test the microtremor analysis methods. Time-domain electromagnetic data showed higher water content in
the footwall block of the fault than in the unsaturated shallow part of the basin.
This project conducted a seismic reflection and gravity experiment in March 1998 to test whether or
not part of the 16 December 1954 Dixie Valley Earthquake (Ms=6.8) produced slip on a low-angle normal
fault. Our geophysical field work included high- and medium-resolution seismic reflection profiles along
Cattle Road from the range-front scarp eastward. Gravity and magnetic transects were conducted across the
valley along Settlement and Cattle Roads and along the scarp from Willow Canyon to Brush Canyon. The
high-resolution survey was conducted within 130 m of the rupture with 100 Hz geophones and a
sledgehammer source. The medium-resolution line used near-surface explosive sources, 8 Hz geophones, and
extended from the rupture to 2.9 km into the basin. A dozen time-domain EM soundings profiled resistivity to
50 m depth across the fault. Fieldwork involved UNR graduate and undergraduate students, and a drilling and
blasting contractor. After data collection, we conducted seismic processing, seismic imaging, velocity
optimization, acoustic modeling, resistivity modeling, and gravity and magnetic modeling studies. Several of
these methods required additional development and adaptation. We presented our results at the 1998 Fall AGU
meeting and have published a paper in JGR (Abbott et al., 2001).
Contributions within Discipline: Project results show that slip along a section of the 16 December
1954 Dixie Valley earthquake rupture took place along a fault plane of unusually low dip (25-30 degrees). In
this regard, it is the first large historical earthquake for which slip on a low-angle normal fault has been
documented. This low-angle normal rupture may represent a relatively rare event, possibly triggered by the M7
Fairview Peak event that preceded it by 4 minutes. In the context of Great Basin faulting, however, with typical
event intervals in the thousands of years, low-angle normal faulting may be a common basin-building tectonic
Research Findings: Both high- and medium-resolution profiles define the 1954 rupture surface as
dipping at 25-30 degrees to 1.5 km depth. The high-resolution survey shows fault-surface reflections from
within 10 m of the surface to 50 m depth; the medium-resolution survey shows fault-surface reflections from
100 m to 1.5 km depth. Reflections subparallel to the fault-surface may be seen in the granite footwall, from 10
to 200 m depth; TEM results suggest these reflections arise in water-saturated fractures. Seen deeper in the
basin are rollover anticlines and buried rupture-graben structures in the sedimentary hanging wall. A vertically
coincident refraction off the fault-surface from a long-offset source is also consistent with the presence of a
smooth, shallowly dipping fault from 50 to 700 m depth. Gravity results support a shallow basin model, as
required by the shallowly dipping boundary normal fault. Our results indicate that slip along a section of the 16
December 1954 Dixie Valley earthquake rupture took place along a fault plane of unusually low dip (25-30
Publications and Presentations:
R. E. Abbott, J. N. Louie, S. J. Caskey, and S. Pullammanappallil, 2001, Geophysical confirmation of
low-angle normal slip on the historically active Dixie Valley fault, Nevada: Jour. Geophys. Res, 106,
R. E. Abbott, J. N. Louie, S. J. Caskey, and S. G. Wesnousky, 1998, Geophysical test of low-angle dip on
the seismogenic Dixie Valley Fault, Nevada: presented at Amer. Geophys. Union. Fall Mtg., Dec.
6-10, San Francisco.
John N. Louie and Abbott, Robert E., 2000, Geophysical confirmation of the 1954 Dixie Valley, Nevada,
rupture as a low-angle normal fault active since 25 Ma: presented at the Geol. Soc. Amer. Ann. Mtg.,
Nov. 16, Reno, Nevada; GSA Abstr. with Progr., 32, no. 7, A-507.
Internet Dissemination: Links to presentations, data, and publications are at:
Contributions to Education and Human Resources: This project provided practical and research
experience to undergraduate and graduate geology and geophysics majors. It helped several to gain
employment in Nevada geophysical industries. As well, it funded part of the Ph.D. thesis research of Robert
Abbott, a UNR graduate student.
Industrial Partnership: Optim LLC, a Nevada software startup owned by UNR Geophysics program
graduates, participated in the analysis of the project's seismic refraction data. Optim demonstrated that their
SeisOpt@3D product, used on the Beowulf 16 Gflop supercomputer they operate at UNR, could quickly
produce more reliable basin-depth and compressional velocity models than traditional tomography techniques.