mahesh_paperBSSA by xiaoyounan


									TITLE PAGE

Station Delays, Their Standard Deviations, and Event Relocations in the

  Reno-Area Basin from a Dense USArray Flexible Array Deployment

               during the 2008 West Reno Earthquake Swarm

Mahesh S. Dhar, John N. Louie, Kenneth D. Smith, Mayo Thompson,

and Annie Kell-Hills


Nevada Seismological Laboratory

University of Nevada

MS 174

Reno, Nevada, 89557

Electronic Supplement:

This article has electronic supplement which includes tables and figures that are

supportive to discussions of the main text of the article. It can be accessed from

The seismic hazard of a sedimentary basin is high. The basin depth and its geometry

control the amplification of ground motion. So from the point of earthquake hazard, it is

important to know the basin geometry and its depth for the hazard evaluation of an area.

Knowing the basin geometry and depth and velocity of sediments will enhance our

understanding of ground amplification to improve seismic hazard evaluations. In this

context, to evaluate the seismic hazard of the Reno Basin, the Nevada Seismological

Laboratory (NSL) conducted a citizen volunteer Reno Basin multiple instrument

temporary deployment using USArray Flexible Array single-channel RefTek RT-125A

(Texan) recorders. Texans were deployed at 106 locations in the Reno Basin during the

May-July 2008 period of the 2008 west urban Reno Mogul earthquake swarm. The

Flexible Array deployment supplemented recordings from 46 NSL/ANSS regional

network, Reno area strong-motion, and 12 IRIS RAMP stations. This allowed relocation

of 97 events in the West Reno/Mogul sequence during May-July 2008, using phase

arrivals on all network and volunteer Texan stations. Hypoinverse runs showed no less

location error with added Texans, but visually, the events are drawn into tighter clusters.

An average station delay map, a mode-of-station-delay map, and a map of the standard

deviations of the average station delays confirm that the western part of the basin is

thickest. Negative delays and high standard deviations in delay times in the Verdi Basin

confirmed the basin to be narrow and complex.


The seismic hazard of western Nevada is high, with many faults capable of producing

magnitude 7 and greater earthquakes (dePolo et al., 1996). Over a 50-year period, the

probability of at least one magnitude ≥ 6.0 event ranges from 34% to 98%, the

probability of a magnitude ≥ 6.6 event ranges from 9% to 64%, and the probability of a

magnitude ≥ 7.0 event ranges from 4% to 50% (dePolo et al., 1997). The present study

area (Fig 1) was referred to as the Reno-area Basin since it encompasses the Reno-Sparks

urban area (Pancha et al., 2007). Abbott and Louie (2000) also referred to this area as the

Reno-area Basin, which is delineated by gravimetry. A sedimentary basin can trap

earthquake waves resulting in a longer duration of shaking and amplification of ground

motions. For example, in Mexico City seismic waves trapped in the low velocity basin

produced high ground motions during the great 1985 Michoacan earthquake (Campillo et

al., 1989; Sanchez-Sesma et al., 1989). The Seattle sedimentary basin, underlying Seattle

and other urban centers in the Puget Lowland, Washington, amplifies long-period (1–5

sec) weak ground motions by factors of 10 or more (Snelson et al., 2007). The

amplification of earthquake ground motion is mainly controlled by basin depth and

shallow shear-wave speeds (Lee et al., 2008). So, from the point of view of earthquake

hazard, it is important to know the basin geometry, velocities, and basin depth for a

robust seismic hazard evaluation. Knowing the basin geometry and its depth enhances the

understanding of ground amplification and improves seismic hazard evaluations.

Different parameters have been applied to demonstrate a basin structure. Aoi (2002)

proposed a waveform inversion scheme that can estimate three-dimensional basin

structure, in particular, the depth of the boundary between sediment and bedrock. Based

on available exploration data, Kagawa et al. (2004) used the spline function to construct a

3-D structural model of the Osaka basin which was further verified and refined by

waveform modeling using a 3D finite difference method (FDM). The Southern California

Earthquake Center (SCEC) has established a basin model in which basin depth is defined

by the depth to a particular shear-wave velocity isosurface, and the model has been

presented in three versions: Version 0 (Magistrale et al., 1996), Version 1 (Magistrale et

al., 1998), and Version 2 (Magistrale et al., 2000). Hearn (1984) used “station delays” to

reflect variation in crustal thickness and sedimentary basin thickness.

This study applied the parameter “station delay” used by Hearn (1984), the time delay of

seismic wave arrivals for each station from different earthquake events. The basic

concept for this parameter is, the higher the delay time, the thicker the basin. It is

expected that the delays are negative for rock sites and positive for the soil sites, as

seismic waves travel faster in rock than in soil. Besides separating rock sites from the soil

sites, there are other practical applications of delay mapping. It can be used to describe

basin amplification of the area. The area of higher delay values may possess higher basin

amplification. The delay mapping can be also used to confirm the basin thickness from

gravity results.

Residents of Reno and Sparks, Nevada experienced hundreds of earthquake shocks from

February through July 2008. The swarm and an M5.0 mainshock on April 25, 2008

shook not only the Reno urban basin, but also the public’s patience. On April 25, the

Nevada Seismological Laboratory (NSL) at the University of Nevada, Reno deployed

NSL-owned RefTek RT-125 (Texan) recorders at eight different locations in Reno.

Fortunately, the Texans recorded the M5.0 Mogul main shock that occurred on that night

at 23:40 PDT. This motivated the NSL to deploy a dense network of Texan recorders in

and around the Reno area. The NSL conducted the Reno Basin Volunteer Deployment

from May 2008 to July 2008 to capitalize on the public’s interest in earthquakes and the

Mogul earthquake sequence. For this, NSL borrowed 90 USArray Flexible Array single-

channel RefTek RT-125A recorders from Program for Array Seismic Studies of the

Continental Lithosphere (PASSCAL) and deployed them at a total of 106 volunteer

locations across Reno and Sparks from May 15 to July 15, 2008. This was a very dense

deployment as the basin only occupies about 150 square kilometers. Such coverage

would not have been possible without extensive public participation.

The objectives of this study were to relocate earthquake events in the West Reno/Mogul

sequence in 2008 and to prepare a station-delay map of the Reno-area basin. Building

delay histograms for each station provided the variance of station delays as well as a

delay-variance map. These time-delay maps should help in understanding the geometry

of the Reno Basin. So, this study assessed the usefulness of using station delays as a

parameter to interpret basin characteristics.

Tectonic Settings and Geology

The Reno-area basin is bordered by the Carson Range of the Sierra Nevada Mountains on

the west and Pah Rah Range and Virginia Range on the east. The Carson Range contains

Mesozoic granite and older metamorphic rocks (Bonham and Bringler, 1973) where as

the other ranges dominantly consist of Tertiary volcanic rocks including andesitic flows

of the Kate Peak Formation (Trexler et al., 2000) and volcanic flows of the Alta

Formation (Bonham and Bringler, 1973). The eastern and some southwestern portions of

the basin contain Kate Peak Formation deposits. The northern hills are composed of the

rocks of the Alta Formation. The western part of the basin contains pre-Lake Lahontan

deposits along with Pliocene sedimentary rocks (NBMG Washoe County Geological

Map, 2000). The basin fill consists of Quaternary deposits containing glacio-fluvial sands

and gravel washed out by drainage from the Sierra Nevada. It also contains outwash from

the most recent glacial epochs (Bell et al., 1989). A significant area of the basin is

underlain by low-density diatomaceous sediments (Abbott and Louie, 2000).

Prior Investigations of the Reno Basin

Abbott and Louie (2000) studied the depth of bedrock using gravimetry in Reno and

Carson City with 200 new gravity measurements along with existing data. The study also

included information from 26 boreholes among the updated database of 56 boreholes

from Garside and Schilling (1979) and Hess (1996). They found an unexpected sediment

depth of 1.2 km below the western side of Reno suggesting that west Reno has the

greatest seismic shaking hazard. Sediment depths below the rest of the Truckee Meadows

basin are less than 0.5 km. In Eagle Valley below Carson City, the sediment depth is 0.53

km. Preston and von Seggern (2007) have performed a high-resolution tomographic

inversion using ~200,000 P and S absolute travel times from nearly 10,000 local sources

recorded at the Nevada Seismological Laboratory (NSL) from 1990-2006. An additional

nearly 200,000 P- and S-wave high-quality cross-correlation differential travel times

were also used in the inversion for double-difference relocation of earthquakes. Low P

velocities and high Vp/Vs were found at shallow depths under the western Nevada Basin.

Results indicated prominent low velocities in the Reno area. Vp/Vs is high within the

basins, as expected in alluvial areas. The data from gravity modeling showed that the

Quaternary outwash (a principal aquifer in the Trukee Meadows) was thickest in the west

Reno area (Widmer et al., 2007). Using the refraction microtremor method of Louie

(2001), Scott et al. (2004) concluded that their transect across the Reno basin lies mostly

in National Earthquake Hazard Reduction Program (NEHRP) classes C/D (Wills et al.,

2000), which suggested that it was due to presence of low shear-wave velocities materials

within the basin.

Prior Investigations of Station Delays

Hearn (1984) used the concept of station delays in southern California to estimate

sediment thickness, and found that the Ventura and Los Angeles basins have delays

which indicate sediment thickness near 10 km. Station delays for Pn travel times have

been used to describe an anisotropic Pn tomography of the western United States (Hearn,

1996). The positive delays can be explained by thick sedimentary basins in the Los

Angeles Basin, the Imperial Valley, and the Ventura/San Fernando Valleys and the

negative delays in Sierra-Nevada Range, Santa Monica Mountains and Eastern

Transverse Ranges, and the western Peninsular Range (Teng and Zhu, 2002). Most of the

estimated delays were near zero, which implies a flat Moho discontinuity beneath the

southern Great Basin (Hearn et al., 1994). Station delays reflect variations in crustal

thickness (Hearn et al., 2004; Hearn, 1999; Serrano, 2005; Zhu and Kamamori, 2000; Pei

et al., 2002). The negative station delays in eastern China suggest a thinner crust and

large positive station delays suggest an approximately 70 km thick crust beneath the

Tibetan plateau (Hearn et al., 2004). In Xinjiang, China, higher station delays showed

that the crustal thicknesses for Tian Shan Mountain and the Qinghai-Xizang Plateaus

were greater than that of the Tarim and Junggar basins (Pei et al., 2002). The station

delays in the western part of China were positive, while those in the eastern part were

generally negative, with the fact that the crust is thicker in the west and thinner in the east

(Suyun et al., 2002). The stations on the thick sediment basins of SW Calabria, Italy

showed zero or positive station delays and for other stations the values were negative

(Raffaele et al., 2006).

Seismic Data Acquisition and Processing

The primary data for this study was collected during the Mogul Earthquake sequence

from April 2008 to July 2008. NSL owned RefTek 125s were deployed at eight different

locations in the Reno area on 25 April, 2008. The RefTek 125 has the capacity to record

continuous seismic data at 50 Hz sampling for two days. On these two days, the RefTek

125 recorded many earthquake events including the M5.0 main shock on 25 April, 23:40

PDT. Following the successful of recording of the main shock, NSL conducted the Reno

Basin Volunteer Deployment under which USArray Flexible Array single-channel

RefTek RT-125A (Texan) recorders were deployed at 106 locations in the basin during

the May-July 2008 earthquake swarm (Dhar et al., 2008; Kell-Hills et al., 2009). In

advance of the deployments, volunteer recorder hosts were invited from among the

public. About 200 people and institutions responded. The deployment sites were selected

according to the people’s response and their availability. Since the earthquake swarm was

in the Mogul and Somersett neighborhoods, the deployments were conducted mostly in

west Reno, where public interest was highest. The five deployments (E see Table S1 in

the electronic supplement to this article) resulted in a total of 106 station locations (E for

list of locations, see Table S2 in the electronic supplement to this article). Many of the

locations were repeated between the various deployments. Each deployment continuously

recorded shaking on a vertical 4.5-Hz geophone for a four-day period. This was a very

dense deployment as the basin only occupies about 150 square kilometers. The Flexible

Array deployment supplemented recordings from 46 NSL/ Advanced National Seismic

System (ANSS) regional networks, Reno area strong-motion (E see Table S3 in the

electronic supplement to this article) and 12 The Incorporated Research Institutions for

Seismology (IRIS) Rapid Array Mobilization Program (RAMP) stations (E see Table S4

in the electronic supplement to this article). ANSS strong-motion stations are distributed

throughout the Reno-area Basin and IRIS RAMP instruments (broadband and

accelerometers) were deployed in the near-source area.

The data stored in the RT125A datalogger was in Texan Raw Data (TRD) format. All

waveform data were collected, and initial earthquake event locations were determined in

the Antelope Database System (v 4.11), product of Boulder Real Time Technologies Inc.; The Antelope database was created following the instructions

provided by the IRIS/PASSCAL Instrument Center. The raw TRD file from the RT125A

datalogger was converted to mini SEED seismic data format (MSEED). Using clockcorr,

time quality and corrections were evaluated, and the headers were modified using fixhdr.

Default fields on the trace to MSEED format were modified. Then, to build the Antelope

database, a batch file was created. MSEED day volumes were created and added to the

database. Calibration values from the calibration table were assigned to the wfdisc.

Finally, a dataless SEED volume was created (PASSCAL, 2009). The data processed in

the Antelope database were stored in the IRIS archive with the ‘YD’ Network ID.

Antelope provides the facilities to users to view and arrange large seismic datasets and

analyze these data in a graphical user interface. It allows the users to apply different types

of filters, manually pick phase arrivals, locate earthquakes, estimate magnitudes and

adjust with many other parameters that are useful for earthquake studies.

The first P-arrivals were picked manually from the raw waveform data recorded on

RefTek 125As to develope earthquake locations and magnitudes. From 97 earthquake

events (E see Table S5 in the electronic supplement to this article), the first P-arrivals

were picked. Later, the first P- and S-arrivals were recovered for the same 97 events

recorded at network stations, which included 46 NSL/ANSS regional networks and 12

IRIS RAMP stations. The end product of this integration is a high-quality earthquake

catalog that can be used to relocate the earthquakes and future seismological studies to

understand the seismic processes in this area.

The Antelope locsat2 location algorithm was used to determine initial locations of the

events from manually picked phase arrivals applying the KDS velocity model (Table 1).

This velocity model is a modification of a model developed to locate earthquakes in the

1997 Fish Lake Valley Earthquake sequence (Mainshock Mw 5.3). A total of 97 events

were located using Antelope. The location and phase information were stored in a

Datascope database and can be easily accessed and used for further analysis. Datascope is

a relational database system in which tables are represented by fixed format files in

ASCII format.

Earthquake Relocation and Time Delays Estimates

Precise earthquake locations play an important role in understanding earthquake source

processes. The quality of relocated events improves due to the significant reduction of

travel-time residuals (Raffaele et al., 2006). In the present study, two events located

using Antelope from the two different data sets for a single time span. One event was

based on the data recorded on Texans and another based on the data recorded on network

stations. Both data were stored in different database systems. Now it is important to

merge both data sets and relocate with integrated phase time. The database contains the

phase information of first arrivals and the event locations. 3545 phases were picked from

the data recorded from 106 Texan stations. For the same events, 1346 analyst-picked first

arrivals from the 58 network stations were used to initially locate all the events. The

USGS Hypoinverse (Klein, 1978, 2000) program has been used to relocate earthquakes,

utilizing both P- and S- wave arrivals recorded by densely deployed RefTek 125A at 106

locations and 58 network stations associated with the Nevada Seismological Lab.

Hypoinverse is standard location program which is a flexible and interactive input. It is a

widely used program that allows regional variations in velocity models with either 1D

layered or 1D layered with linear velocity gradients to be used (Smith et al., 2004). S-

arrivals can be used when available. The solution to the location problem is solved by

series of iterations in linear steps to converge on the minimum mean-square travel time

residuals (RMS) to estimate the best event location. The Hypoinverse calculated RMS


                    RMS =2    (w r )
                                   i i

                                                 …………………………. (1),
                              (w ) i

where ri are the individual station residuals and wi are their final weights.

To run Hypoinverse, stations and phase arrival times were compiled in Hypoinverse

format along with the velocity model. Different weights were assigned for the stations,

which showed that there is not much difference in the time residual outputs. Hypoinverse

minimizes the travel-time residuals in a least square sense for the best event location. If

the velocity model is perfect, each station should have an average residual of zero; if

random errors are absent, each station would furthermore have zero standard deviation of

errors (Nelson and Vidale, 1990). This study used the KDS velocity model, regularly

used by the Nevada Seismological Lab for the Reno area. Playing with different other

models, KDS velocity model was the one that worked best for the Reno area as the

resulting residuals were reasonable. Note that the KDS model accounts for an average

basin velocity in its top layer, with a Vp of only 3.0 km/s.

Arrival times of seismic waves at sites located over sedimentary basins show travel time

delays due to the differences in seismic wave velocity between basement rocks and

overlying sediments and difference between the model and time velocities. This time

delays lead to an error of the hypocenter determination (Makoto et al., 2004). So, for the

accurate hypocenter locations, time delays were calculated.

For this, Hypoinverse calculated the model time (tcal) using our velocity model and

subtracted it from the observed travel time (tobs), such that;

                          Time delays (Δt) = tobs – tcal................. (2)

With the precise (high accuracy) hypocenter locations, we can delineate fault zones that

help to accurately quantify regional seismic hazards.

Calculation of Average Time-delays

Time-delay is one of the important factors in determining basin geometry and basin

thickness. It is also an important parameter in assessing the hypocenter location. By

running the Hypoinverse program, time-delays for each station for each of 97 relocated

events were calculated. Then the average time-delay was taken for each station. It should

be noted that time-delays were determined to those stations and events in which P-and S-

wave arrivals were picked.

The travel time residuals were inverted for station delays, event delays, and lateral Pn

velocity variations using a modified time-term equation (Hearn et al., 1984):

                          tij = ai + bj+  dijk Sk………………… (3)

where tij is travel time between source i and station j, ai is the static delay for station i, bj

is the static delay for event j, dijk is the distance travelled by ray ij in mantle cell k, and Sk

is the slowness perturbation of cell k.

Static station delays are related to both crustal velocity and crustal thickness by the

equation (Hearn et al., 1984):

                         ai = zi (1/vcrust2 – 1/vmantle2)1/2………… (4)

where zi is the crustal thickness, vcrust is the mean crustal velocity, and vmantle is the mean

mantle velocity.

In this study, for a simple two-layered model, the equation can be modified as:

                         ai = zi (1/vbasin2 – 1/vrock2)1/2………….. (5)

where zi is total refractor depth, vrock is refractor velocity, and vbasin is the velocity of the

material above the refractor.


Earthquake Relocation

In this study, earthquakes occurring during the 2008 Mogul Earthquake swarm have been

considered. During this period, event magnitudes varied from M0.19 to M3.05 for May-

July events, while the highest was M5.0 during February-April events. I relocated a total

of 97 earthquake events by running the Hypoinverse program (Klein, 1978 and 2000).

By adding more stations, root- mean-square travel time residuals (RMS) were expected to

be minimized. The results showed that there was no less location error by adding the

Texan stations. The mean RMS considering ANSS stations only was 0.09, while when

adding the Texan stations, it was 0.11. Similarly, the mean horizontal error and the mean

vertical error considering ANSS stations only are 1.93 and 2.16 respectively. After

adding the Texan stations, they became 2.11 and 2.14 respectively. So, RMS and

horizontal error were not improved, while vertical error was slightly improved by adding

the Texan stations. The reason for the increase in RMS value and horizontal error might

be due to adding of Texans in the basin where there were significant basin effects. Other

reason might be due to the fact that some of the Texans were deployed in distant

locations (approximately 35 km) away from the Mogul area. In this context, the velocity

model used for the Reno-are Basin might not be appropriate for such far locations.

However, when locations were visually interpreted, it was found that the event locations

that were scattered were now tighter (Fig 2). So, visually adding Texan stations gave

better locations. The event locations obtained from ANSS stations were mostly shifted to

NW directions (E see Figure S1 in the electronic supplement to this article) once

relocated after adding the Texan stations. Excluding the events for which the shift was

more than 10 km, RMS between differences in horizontal location was 1.19, average

horizontal shift was 1.7 km, and average azimuth was 300 o (E see Table S6 in the

electronic supplement to this article). The shifting of event locations to NW may be due

to uneven station geometry. In this study, the events were clustered in Mogul area and

many Texans were deployed in SE of Mogul. So, Hypoinverse moved the event locations

to NW. The shifting direction of the event locations also depends on the velocity model

used. If the velocity model is slow, then the location program assumes that the travel

paths are longer. Therefore, locations may be pushed away from the dense part of the

network to make the travel time match with the velocities.

The delays were more advanced at network stations which were close to the earthquake

source while adding the Texans (E see Table S7 in the electronic supplement to this

article). Texans were mostly in the basin, so the network stations had later arrival time

making the origin time later in the Hypoinverse solution. Since the origin time was later,

at the stations close to the earthquake source the delays became more negative. Similarly,

the RMS between difference in event depth before and after adding Texans was 1.71 and

the average of the absolute values in the difference in depth was 1.27 km (E see Table S8

in the electronic supplement to this article).

The relocation of earthquakes with added stations showed some clear trend of occurrence

of earthquakes. The February-April events showed the trend of NW-SE, indicating an

active fault zone in that trend. The May-July events concentrated NE of the February-

April trend. This is the same distribution shown by Smith et al. (2008) double-difference

location, using ANSS and RAMP stations.

Delay Mapping and Analysis

The station delays are primarily influenced by the crustal thickness and velocity but are

also sensitive to systematic timing and picking errors (Mele et al., 1998). The events

considered for this study were shallow depth events. So, the delays were influenced by

the basin depth and the velocity along basin and rock layer. The time delays of each

station for 97 relocated events were determined by running Hypoinverse (Klein, 1978 and

2000). The average of time delays obtained from the events was calculated for each

station. Histograms were plotted for every station. Histograms of delays at each station

showed that the delays were relatively insensitive to the details of the Hypoinverse

location process. The histograms depicted that the time-delays were mostly positive for

the stations at soil sites (Fig 3.a), whereas delays were mostly negative at the rock sites

(Fig 3.b). But for some rock sites and soil sites, there were exceptions. For example,

station NOAA is on a rock site, but showed a positive delay. The nearby stations SWSN

and RFNV on the rock site showed negative delays. Pancha et al., (2007) also pointed out

that NOAA was located on altered volcanic rock and had lower velocity values. The

rocks surrounding NOAA were visibly more generally weathered (Bonham and Bringler,

1973). Regarding RFNV, amplitudes of recorded seismic ground motions were much

smaller in comparison to other ANSS stations indicating that RFNV was located on more

competent rock (Pancha et al., 2008).

An average station delay map was prepared by contouring the average delays (E see

Table S9 in the electronic supplement to this article) at the stations. The positive contours

of the delays were concentrated on the west side of the Reno-area Basin (Fig 4.a). This

depicts that the basin is deepest in the west, despite the possible effects of low-density

diatomite on the gravity studies (Abbott and Louie, 2000). To check the validity of the

average station delay, a mode-station-delay map (Fig 4.b) was also prepared. The

contours of higher mode of delay (E see Table S9 in the electronic supplement to this

article) confirmed the distribution of higher station delays in the west Reno Basin. A map

of the standard deviations (E see Table S9 in the electronic supplement to this article) of

the average station delays (Fig 4.c) was also prepared. The low standard deviation values

in the west Reno Basin validated the average station delays. Standard deviations of

average delays were higher in the narrow region adjacent to the Verdi Basin. Volcanic

rocks underlie, and are locally interbeded with, the lower parts of the sedimentary strata

of the Verdi basin (Trexler et al., 2000). The narrow and small basin is generally

composed of a variation of lithology and characterized by complex structures within a

small area giving a range of values. The higher quantity obtained from standard deviation

normalized by average delays (E see Table S9 in the electronic supplement to this article)

in the Verdi Basin also confirmed the basin to be narrow and complex (E see Figure S2 in

the electronic supplement to this article). The dominance of negative delays in the Verdi

area may be due to the distribution of rock around the area and use of the velocity model

assigned for the Reno Basin.

Are Delays Linearly Proportional to Basin Depth?

There is a strong correlation between basin depth reported in the California USGS 3D

seismic velocity model (ver.2) (Jachens et al., 1997) and arrival-time delays (Dolenc et

al., 2005; Dolenc and Dreger, 2005). Tkalčić et al. (2008) showed the possibility to

establish a single trend between basin depth and travel-time delays for stations located in

the areas of the basin corresponding to shallower depths. To test the hypothesis, a graph

of depth of basin from Widmer et al. (2007) and average station delay underneath each

station was prepared (Fig 5). There was a scatter of points suggesting that the hypothesis

was negative, i.e. delays were not linearly proportional to the basin depth. However, a

minimum delay line (Fig 5) suggested that minimum delays were proportional to basin

depth. The scatter of points was due to a range of average propagation angles () to each

station, with most events at one end of the basin. Some delays were very high even

though the events were at shallow depths.

This may have been due to a long travel path, which was possible if the ray propagated

with large angle . This hypothesis was checked by computing the maximum angle of

propagation observed through a simple model. For the model, the first layer velocity was

taken as v0=3.0 km/sec to 1 km depth and the second layer velocity as v1=4.5 km/sec

below from the KDS velocity model. A linear minimum delay line was drawn to account

for the average delay with respect to depth (Fig 5). It was drawn from -0.3 sec delay at 0

m depth. This translation was done as the delays were calculated relative to an average

model that included some basin velocity. Here, the KDS velocity model was used which

is based on some basin averages.

Calculations based on Figure 5:

                                     v0          3
min = critical angle (ic) = sin-1      = sin-1      = 41.8o
                                     v1         4 .5

                               0.37 sec
Change in slowness (Δs) =               x Cos 41.8o = 0.34 sec/km

Total slowness (s) = Δs + slowness of the first layer = 0.34 + 0.33 = 0.67 sec/km

                           1    1
The velocity of basin =      =     = 1.48 km/sec
                           s 0.67

Hence the basin velocity was 1.48 km/sec. Taking some larger delay averages at L, M,

and N in Figure 5, the resulting propagation angles calculated were 83.3o, 73.1o, and

66.6o respectively. Since the main events were on the west side of the basin, propagation

angles were greater for the stations on east side of the basin resulting on longer travel

path and positive delays.

To clarify this hypothesis, a simple basin model was developed as shown in Figure 6. Z is

an earthquake location at a shallow depth from where the seismic wave transmits and

strikes the basin boundary at A. AB is a vertical ray path within the basin, which is also

the shortest ray path with the minimum time delay. But, due to refraction across the basin

floor, the wave follows a certain propagation angle (min) through the basin to station X.

The propagation angle (min) gives the measured delays. If the propagation angle

increased to , then the length of ray path also increased, with added delays.

To check the validity of minimum delay line, the station delay equation by Hearn et al.

(1984) was used. With a calculated basin velocity of 1.48 km/sec and rock layer velocity

of 3 km/sec from the velocity model. The relation between the delays and depths was

       ai = 0.56 zi ..............................................................................Line A (Fig 5)

where, ai = station delay and zi = refractor depth.

This Line A was transferred to -0.3 sec delay because the delays were calculated relative

to an average model that included some basin velocity paths. Then Line A became,

       ai = 0.56 zi - 0.3 ......................................................................Line A’ (Fig 5)

The Line A’ was nearly parallel to the projected minimum delay line (Fig 5). Assuming

the basin depth was greater than 1 km and assuming rock layer velocity of 4.5 km/sec

from the velocity model, the relation between delays and depths was,

       ai = 0.62 zi................................................................................................................... Line B (Fig 5)

where ai = station delay and zi = refractor depth.

This Line B was transferred to -0.3 sec delay because the delays were calculated relative

to an average model that included some basin velocity paths. Then Line B became,

       ai = 0.62 zi - 0.3 ......................................................................Line B’ (Fig 5)

The Line B’ was nearly parallel to the projected minimum delay line (Fig 5).

Both lines showed a parallel trend to the projected minimum delay line (Fig 5),

confirming that depth was proportional to minimum delay.

Do Delays Increase with Distances?

To test this hypothesis, average delays for each station and average distances from

earthquake locations to receiver stations were examined. A plot of average distances

versus average delays (Fig 7) did not show any distinct linear relationship between

distances and delays. So, the hypothesis was negative. Similarly, mode distance versus

mode delays also did not show any distinct linear relationship, which was contrary to the

hypothesis. The reason for this may be due to the concentration of stations within the

single and small basin.

Comparison of Results with Prior Investigations

Abbott and Louie (2000) studied the depth of bedrock using gravimetry in Reno and

Carson City, in which they found an unexpected basin depth of 1.2 km below the

western side of Reno. From the gravity modeling, Widmer et al. (2007) found the basin to

be thickest in west Reno. Both of these results agreed with the result from this study, in

which the observed station delays were larger in the west Reno Basin, confirming that the

basin is deeper there. Hearn (1984) used the concept of station delays to estimate the

sediment thickness. The station delay equation (Hearn, 1984) showed a relationship

between delays and basin depths. Using this equation in the present study, the hypothesis

stating that depth was linearly proportional to the minimum delays was confirmed. Prior

investigations in station delays (Hearn et al., 2004; Hearn, 1999; Serrano et al., 2005; Zhu

and Kamamori, 2000; Pei et al., 2002, Raffaele et al., 2006, Teng and Zhu, 2002, and

Suyun et al., 2002) were consistent with the relationship between station delays and the

basin thickness, which is the main outcome of the present work.


A dense deployment of FA Texans was performed from May to July 2008, during the

Mogul earthquake swarm. The five deployments recorded activity at a total of 106

volunteer host locations. The FA Texan deployment supplemented recordings from 46

NSL/ANSS regional networks and 12 IRIS RAMP stations. The recorded data were

processed in the Antelope database system and stored in an IRIS archive with ‘YD’

network ID.

One of the important products of this work was an average station delay map. The delay

measurements have confirmed the deep western sub-basin in agreement with gravity

results, despite suspicions that low density diatomite sediments in the basin

may have been giving false depths. This fact was also supported by the larger mode

values and low standard deviation in the western part. Negative delays and the high

standard deviations observed in the Verdi Basin were due to narrow and complex

structures within the basin and the distributions of volcanic and sedimentary rocks in the


In general, location error was less while adding the stations, but the study showed that

there was no less location error by adding FA Texan stations on ANSS and IRIS RAMP

stations. The RMS and horizontal error were not improved, while vertical error was

slightly improved by adding the stations. This might be due to adding stations in basin,

where there were significant basin effects. However, visually, adding the Texan stations

gave better locations. The scattered in event locations became less. The locations were

mostly shifted in the NW direction. The strong seismic lineation in NW-SE showed some

structures indicating an active fault in that trend.

For better results for event locations, the Texans must be deployed around the main event

locations. For better basin structure and depth there must be equal distribution of Texans

in rock and soil sites all around the basin. Then active sources can be applied at the

middle of the basin, such that the seismic waves transmit all over the basin.


Research supported by the U.S. Geological Survey (USGS), Department of the Interior,

under USGS (07HQAG0015). The views and conclusions contained in this document are

those of the authors and should not be interpreted as necessarily representing the official

policies, either expressed or implied, of the U.S. Government. Thanks to the Earthscope

Flexible Array and the PASSCAL Instrument Center at New Mexico Tech, who provided

ninety Flexible Array single-channel REF TEK RT-125A recorders for this experiment.

Thanks are given to all the personnel involved in collecting the field data. Thanks go to

all those who generously volunteered their properties to host for this project. Without

public participation, this experiment could not have been conducted.


Abbott, R. E., and J. N. Louie (2000). Depth to bedrock using gravimetry in the Reno and

Carson City, Nevada, area basins, Geophysics, 65(2), 340 – 350.

Aoi, S. (2002). Boundary shape waveform inversion for estimating the depth of three-
dimensional basin structure, Bull. Seism. Soc. Am. 92, 2410-2418.

Bell, J. W., R. J. Watters, P. A. Glancy, J. R. Keaton, and R. N. Morris, (1989).
Engineering geology of the Reno-Lake Tahoe area, Nevada: Field trips for the 28th
international geological congress, in Environmental, engineering, and urban geology in
the United States; Vol 2; Engineering geology of Western United States urban centers,
edited by P.M. Hanshaw, pp. 41-50, Am. Geophys. Union, Washington.

Bonham, H. F., Jr., and E. C. Bingler (1973). Geologic map, Reno quadrangle: Nevada
Bureau of Mines and Geol. Map 4Ag.

Campillo, M., J. C. Gariel, K. Aki, and F. J. Sanchez-Sesma (1989). Destructive strong
ground motion in Mexico City; source, path, and site effects during great 1985
Michoacan earthquake, Bull. Seism. Soc. Am. 79 (6), 1718-1735.

dePolo, C. M., J. G. Rigby, G. L. Johnson, S. L. Jacobson, J. G. Anderson, and T. J
Wythes (1996). Planning scenario for a major earthquake in western Nevada: Nevada
Bureau of Mines and Geol. Special Publication 20.

dePolo, C. M., J. G. Anderson, D. M. dePolo, and J. G. Price (1997). Earthquake
Occurrence in the Reno-Carson City Urban Corridor, Seism. Res. Let. 68, 401-412.

Dhar, M. S., M. Thompson, A. Kell-Hills, J. N Louie, K.D. Smith, J. Tirabassi , S. Tom,
and T. Irwin (2008). Educating a community impacted by an earthquake swarm: 106
volunteers host Earthscope Flexible Array recorders during the Mogul, Nevada sequence:
EOS Trans. AGU 89(53), PA13B-1342 (abstract).

Dolenc, D., D. Dreger, and S. Larsen (2005). Basin structure influences on the
propagation of teleseismic waves in the Santa Clara Valley, California, Bull. Seism. Soc.
Am. 95, 1120-1136.

Dolenc, D., and D. Dreger (2005). Microseisms observations in the Santa Clara Valley,
California, Bull. Seism. Soc. Am, 95, 1137-1149.

Garside, L. J., and J. H. Schilling (1979). Thermal Waters of Nevada, Nevada Bureau of
Mines and Geology Bulletin 91.

Hearn, T. M. (1984). Pn travel times in Southern California, J. Geophys. Res. 89, n B3,

Hearn, T. M., and A. C. Rosca (1994). Pn tomography beneath the Southern Great Basin,
J. Geophys. Res. 21, n 20, 2187-2190.

Hearn, T. M. (1996). Anisotropic Pn tomography in the western United States, J.
Geophys. Res. 101, n B4, 8403-8414.

Hearn, T. M. (1999). Uppermost mantle velocities and anisotropy beneath Europe, J.
Geophys. Res. 104, n B7, 15,123-15,139.

Hearn, T. M., S. Wang, J. F. Ni, Z. Xu, Y. Yu, and X. Zhang (2004). Uppermost mantle
velocities beneath China and surrounding regions, J. Geophys. Res. 109, B11301, doi:

Hess, R. H. (1996). Nevada Oil and Gas Well Catalog (NVOILWEL), in Nevada Bureau
of Mines and Geology Database 3.

Jachens, R. C., R. F. Sikora, E. E Brabb, C. M. Wentworth, T. M. Brocher, M. S.
Marlow, and C. W. Roberts (1997). The basement interface: San Francisco Bay area,
California, 3-D seismic velocity model, EOS Trans. AGU 78, F436 (abstract)

Kagawa, T., B. Zhao, K. Miyakoshi, and K. Irikura (2004). Modeling of 3D basin

structures for seismic wave simulations based on available information on the target area:
case study of the Osaka Basin, Japan, Bull. Seism. Soc. A., 94, 1353–1368.

Kell-Hills, A. M., M. S. Dhar, M. Thompson, J. N. Louie, and K. D. Smith (2009).
Community-outreach efforts in data collection and analysis for the 2008 Mogul
earthquake sequence: 2009 SCEC Annual Meeting poster 1-050, Palm Springs, Calif.,
Sept. 13-16; abstract in Proceedings and Abstracts, Volume XIX, p. 218-219.

Klein, F.W. (1978). Hypocenter location program-HYPOINVERSE: users guide to
versions 1, 2, 3, and 4, U. S. Geol. Surv. Open-File Rept. 78-694, 1-113.

Klein, F. W. (2000). HYPOINVERSE-2000, a FORTRAN program to solve for earth- 5
quake locations and magnitudes 4/2002, U.S. Geol. Surv. Open File Rept., 02-171.

Lee, S. J., H. Chen, Q. Liu, D. Komatitsch, B. Huang, and J. Tromp (2008). Three-
Dimensional Simulations of Seismic-Wave Propagation in the Taipei Basin with Realistic
Topography Based upon the Spectral-Element Method, Bull. Seism. Soc. Am. 98, no. 1,
253-264. doi: 10.1785/0120070033.

Louie, J. N. (2001). Faster, better: shear-wave velocity to 100 meters depth from
refraction microtremor arrays, Bull. Seism. Soc. Am. 91, no. 2, 347-364.

Magistrale, H., K. McLaughlini, and S. Day (1996). A geology based 3-D velocity
model of the Los Angeles basin sediments, Bull. Seism. Soc. Am. 86, 1161–1166.

Magistrale, H., R. Graves, and R. Clayton (1998). A standard three-dimensional seismic
velocity model for southern California: version 1, EOS Trans. AGU 79, F605.

Magistrale, H., S. Day, R. Clayton, and R. Graves (2000). The SCEC southern California
reference three-dimensional seismic velocity model version 2, Bull. Seism. Soc. Am. 90,

Makoto, N., S. Einoshin, N. Hiroo, and N. Koichi (2004). The travel time delay
correction of seismic wave at CEORKA strong ground motion observation stations in the
Osaka Basin, Journal of Geosciences, Osaka City University, 47, 141-147.

Mele, G., A. Rovelli, D. Seber, T. M. Hearn, and M. Barazangi (1998). Compressional
velocity structure and anisotropy in the uppermost mantle beneath Italy and surrounding
regions, J. Geophys. Res. 103, n B6, 12,529-12,543.

Nelson, G. D., and J. E Vidale (1990). Earthquake location by 3-D finite difference travel
times, Bull. Seism. Soc. Am. 80, 395–410.

NBMG- Nevada Bureau of Mines and Geology (2000). NBMG Open-File Report 97-1,
County Digital Geologic Maps 1:250,000 scale.

Pancha, A., J. G. Anderson, J. N. Louie (2007). Characterization of near-surface geology
at Strong-Motion stations in the vicinity of Reno, Nevada, Bull Seism. Soc. Am. 2096–

Pancha, A., J. G. Anderson, J. N. Louie, and S. K. Pullammanappallil (2008).
Measurement of Shallow Shear Wave Velocities at a Rock Site using the ReMi
Technique, Soil Dynamics and Earthquake Engineering 28, 522-535.

PASSCAL (2009).

Pei, S., Z. XU, and S. Wang (2002). Pn Velocity tomography        in Xinjiang, China and
adjcant regions, Chinese Journal of Geophysics 45, no. 2, 217-224.

Preston, L., and D. von Seggern (2007). High-Resolution 3-D Regional Double-
Difference Tomography of the Reno-Tahoe-Carson City Region, AGU T33A-1147

Raffale, R., H. Langer, S. Gresta, and F. Moia (2006). Tomographic inversion of local
earthquake data from the Gioia Tauro basin (South-western Calabria, Italy), Geophys. J.
Int. 165, 167-179.

Sanchez-Sesma, F. J., M. Camplillo, P. Y. Bard, J. C. Gariel, and K. Aki (1989). The
Great 1985 Michoacan Earthquake: A Unified Approach Considering Source, Path, and
Site Effects, in Cakmak, A.S., and Herrera, I., Eds., Engineering Seismology and Site
Response, edited by A. S. Cakmak, and I. Herrera, Computation Mechanics Publ, 53-57.

Scott, J. B., M. Clark, T. Rennie, A. Pancha, H. Park, and J. N. Louie (2004). A shallow
shear-wave velocity transect across the Reno, Nevada area basin, Bull. Seism. Soc. Am.
94, 2222–2228.

Serrano, I., T. M. Hearn, J. Morales, and F. Torcal (2005). Seismic anisotropy and
velocity structure beneath the southern half of the Iberian Peninsula, Physics of the Earth
and Planetary Interiors 150, 317–330

Smith, A. J., P. R. Cummins, T. Baba, S. Kodaira, Y. Kaneda, and H. Yamaguchi (2004).
Intra-plate seismicity in the subducting Philippine Sea Plate, southwest Japan:
magnitude–depth correlations, Physics of the Earth and Planetary Interiors 145, 179–

Smith, K., D. von Seggern, D. dePolo, J. G. Anderson, G. P. Biasi,                 and R.
Anooshehpoor(2008). Seismicity of the 2008 Mogul-Somersett West Reno,Nevada
Earthquake Sequence, EOS Trans. AGU 89(53), Fall Meet. Suppl., S53C-02 (abstract).

Snelson, C. M., T. M. Brocher, K.C. Miller, T. L. Pratt, and A. M. Tre´hu (2007).
Seismic Amplification within the Seattle Basin, Washington State: Insights from SHIPS
Seismic Tomography Experiments, Bull. Seism. Soc. Am. 97, No.5, 1432-1448. doi:

Suyun, W., T. M. Hearn, X. Zhonghuai, J. Ni, Y. Yanxiang, and Z. Xiaodong (2002).
Velocity structure of uppermost mantle beneath China continent from Pn tomography,

Science in China, 45, n 2, 143-150.

Teng, T., and L. Zhu (2002). Study of the moho depth and crustal Vp/Vs variation in

southern California from teleseismic waveforms, University of Southern California,

USGS award no: 00HQGR0007.

Tkalčić, H., A. J. Rodgers, N. Rawlinson, D. J. McEwan, and C. M. Snelson (2008).

Teleseismic Travel-Time Delays in the Las Vegas Basin, Bull. Seism. Soc. Am. 98, no. 4,

2047- 2060. doi: 10.1785/0120050239.

Trexler, J. H., P. H. Cashman, C. D. Henry, T. Muntean, K. Schwartz, A. TenBrink, J. E.
Faulds, M. Perkins, and T. Kelly (2000). Neogen basins in western Nevada document the
tectonic history of the Sierra Nevada-Basin and Range transition zone for the last 12 Ma,
in Great Basin and Sierra Nevada, D. R. Lageson, S. G. Peters and M. M. Lahren
(Editors), Boulder, Colorado, Geol. Soc. of Am. Field Guide 2, 97–116.

Widmer, M. C., P. Cashman, J. Trexler, and C. Benedict (2007). Neogene through
Quaternary stratigraphy and structure in a portion of the Truckee Meadows basin: a
record of recent tectonic history, Geol. Soc. Am. 39(4), 9 (abstract).

Wills, C. J., M. Petersen, W. A. Bryant, M. Reichle, G. J. Saucedo, S. Tan, G. Taylor,
and J. Treiman (2000). A site-conditions map for California based on geology and shear-
wave velocity, Bull. Seism. Soc. Am. 90, no. 6B, S187–S208.

Zhu, L., and H. Kanamori (2000). Moho depth variation in southern California from
teleseismic receiver functions, J. Geophys. Res. 105, no. B2, 2969-2980.

Authors’ affiliations, Address:

Nevada Seismological Laboratory

University of Nevada

MS 0174

Reno, Nevada, 89557

M. T. now at:

Zonge Geosciences

Denver, Colorado

FIGURE Captions

Figure 1:   Location map of Reno-area Basin showing stations and event

Figure 2:   Tighter event locations by adding FA Texan stations

Figure 3:   (a) The histograms of time delays at the stations on soil sites showed
            that the dominant delays were positive or late, but set had clear
            exceptions such as FENE and MOGL. (b) The histograms of time
            delays at the stations on rock sites showed that the dominant delays
            were negative or early, but set had clear exceptions such as PEST and

Figure 4:   (a) Average station delay map showing positive contours on the west
            side of the Reno Basin. Contour interval 0.1 sec. (b) Mode-station-
            delay map showing the higher positive mode delays concentrated in
            the western part of the Reno Basin. Contour interval 0.05 sec. (c)
            Standard deviation of average station delay map of Reno Basin.
            Contour interval 0.2 sec with a subsidiary contour at 0.05 sec.
            Contour 0.5 is removed to prevent showing false maxima.

Figure 5:   Delays versus depth from Widmer et al. (2007) with a minimum
            delay line showing linear relationship between depths and minimum
            delays. Solid line is projected minimum delay line. Dotted line is
            obtained taking rock layer velocity equal to 3 km/sec and dashed line
            is obtained taking rock layer velocity equal to 4.5 km/sec.

Figure 6:   Simple basin model and ray propagation diagram.

Figure 7:   No relationship between average delays for each station versus the
            average distance from earthquake location to the receiver station.


To top