Simulation of Scenario-Earthquake Shaking in the Lake ...crack.seismo.unr.edu/.../Louie-TahoeShaking-preprint.docx · Web viewThe Tahoe basin also provides an important test bed for

Simulation of Scenario-Earthquake Shaking in the Lake Tahoe Basin–

A Comparison Between ShakeMap and Nevada ShakeZoning

John Louie1, Gretchen Schmauder1,2, Graham Kent1, Kenneth Smith1,Kevin McBean1, Alexa McBean1, Kyle Gray1, and Kelley Hall3

1Nevada Seismological Laboratory, University of Nevada, Reno, Mailstop 0174, Reno, Nevada 89557-0174, USA; [email protected], Inc., 2190 Fortune Drive, San Jose, California 951313Department of Earth and Space Sciences, University of Washington, Johnson Hall Rm-070, Box 351310, Seattle, Washington 98195-1310

Large normal faults, capable of producing earthquakes greater than magnitude 7.0, are responsible for the formation of the Lake Tahoe basin. Uplift and erosion over the past several million years have resulted in a structural geologic basin with thick sedimentary fill, significantly enhancing the potential for earthquake shaking. Using existing and newly collected geophysical data, we produced a 3-D model that serves as the basis for numerical simulations of earthquake shaking. Ground motions modeled using Nevada ShakeZoning (NSZ), our physics-based method incorporating 3-D geotechnical information and basin shape, produce peak ground velocity (PGV) maps that are considerably different than those obtained from ShakeMap, a standard USGS tool for 1-D ground motion estimation. Although the USGS tool conservatively over-predicts ground shaking outside the Lake Tahoe basin, it substantially and non-conservatively under-predicts potential ground motions within the basin. Our 3-D computations suggest “extreme,” sustained shaking in the basin, threatening several communities with “very heavy” earthquake damage. Comparing these new shaking models to very limited earthquake-monitoring station data for Tahoe suggests NSZ may perform well as new recordings become available. Given ten rupture scenarios on five major faults in and around the Tahoe basin, we developed a simplified probabilistic hazard map that includes the 3-D basin amplification effects. The simplified map sums the annual-rate-of-exceedance (ARE) of a key peak ground-motion level of 30 cm/sec computed by NSZ, on the threshold of “severe” shaking expected to produce “moderate” to “heavy” damage (a PGA of about 30%g). The map shows that areas at higher risk for dangerous shaking, more frequent than once in 1000 years, strongly correlate with basin shape.

INTRODUCTIONUntil recently, potential earthquake shaking within the Lake Tahoe basin

has been overlooked, in large part due to the water that covers the most active fault lines. We assess potential ground motions by comparing the one-dimensional USGS product ShakeMap (Wald et al., 1999; Worden et al., 2012), with a three-dimensional University of Nevada, Reno (UNR) product, Nevada ShakeZoning (NSZ). This updated ground motion information, along with new geologic and geotechnical information on basin material properties and the length/rupture

Louie et al. 1

history of the three main basin bounding faults, contribute to probabilistic maps showing the annual rate of exceedance of hazardous ground motion levels.

The NSZ process is based on simple and direct measurements of the local geology, specific to the region being studied (Louie, 2008; Louie et al., 2011; Flinchum et al., 2014). NSZ enables rapid, inexpensive measurements of shallow shear-wave velocity structure and basin geometry that feed into a community velocity model (CVM) for 3-D earthquake-scenario calculation. Unlike ShakeMap, which relies on one-dimensional ground-motion relations, NSZ produces a three-dimensional computation as a result of applying all the available information on local basin geometry. Since computational resources limit our NSZ calculations to shaking frequencies below 1.0 Hz, the 3-D shaking results cannot currently yield valid computation of ground accelerations or the peak ground acceleration “PGA.” NSZ directly produces shaking time-histories of ground velocity, and thus peak ground velocity or “PGV.” ShakeMap’s scale bar, calibrated by Worden et al. (2012), will allow a rough association of PGV levels computed here, to corresponding PGA levels.

The objectives of this study are to provide better ground motion estimates and more effective hazard maps to local communities in the Lake Tahoe basin. Using the NSZ methodology of incorporating all available geological information, we have developed a novel and simple probabilistic seismic hazard analysis (PSHA) for specific, high-population-density urban centers within the Tahoe basin. The resulting annual-rate-of-exceedance (ARE) maps give the annual probability of a “severe” level of shaking, 30%g, on the threshold of producing moderate to heavy damage to buildings.

This is a different type of hazard map than the USGS National Hazard Maps (Petersen et al., 2014) and the NEHRP Provisions of the International Building Code (BSSC, 1998). Those maps give a probabilistic ground-shaking level for each location on the map, expected within a certain period of time, such as 2% probability in 50 years (or a 2500-year interval). Our simplified ARE maps are, in a way, the inverse of the USGS maps. The ARE maps give the annual likelihood at each location on the map, of exceeding a fixed, damaging level of shaking- in our case a PGV of 30 cm/sec (30%g PGA). The USGS plans the next update of their hazard maps for 2016, and those maps are not likely to include the faults in the Tahoe basin that are published in this volume (e.g., Kent et al., 2016), or 3-D basin effects. Despite their simplicity and shortcomings, the approximate ARE maps will be much more realistic than the USGS hazard maps, and they will be available much sooner than realistic USGS maps. Our mapping of the annual likelihood of damaging shaking should give Tahoe communities the ability to make more informed land-use decisions. In the interim, before the USGS can update its hazard maps to include all the earthquake faults and all the 3-D basin effects in the Tahoe basin, we hope the simple ARE maps will provide guidance to Tahoe communities wanting to build earthquake resilience.

A long-term objective for this study would be to confirm that the NSZ process can provide accurate predictions of shaking and amplification, with a particular emphasis on the response of the Lake Tahoe basin. Developing an effective ground motion model requires some level of validation through comparison of a reasonable synthetic event co-located to a recorded seismic event. Full model validation provides credibility that the resulting hazard maps offer reliable data for city planners and engineers. Previous work in Southern California (Hartzell et al., 1999; Olsen and Mayhew, 2010), Northern California (Hartzell et

Louie et al. 2

al., 2006; Aagaard et al., 2008; Kim et al., 2010), southern Nevada (Flinchum et al., 2014), and the Seattle basin (Frankel et al., 2007; 2009) illustrate the methods used to validate such models; all show that the presence of geological basins and variations in geotechnical velocity elevate the shaking hazard well above the levels predicted by the USGS National Hazard Maps (Petersen et al., 2014). Such prior work, along with the very limited comparison data available for this study, are steps toward improving ground motion estimations for complex geological environments such as found in the Tahoe basin.

Unfortunately, observed ground-motion data are currently available for just a few events and very few monitoring stations near Tahoe. For comparison purposes, we simulate only six earthquakes of greater than M3.2 within and near the Tahoe basin. No recordings or events are available from the South Lake Tahoe sub-basins, where our CVM has the most geological and geotechnical detail. So the few comparisons between synthetics and recordings we can make do not meet the long-term objective of validation of our data or methods.

What we can accomplish at this point is our initial objective of informing the Tahoe community in a realistic way about the additional earthquake-shaking hazards posed by the geological basins and the geotechnical environment. There is no doubt that these features create hazard levels well above the level predicted by the USGS National Hazard Maps (Petersen et al., 2014). Validation of exactly how much additional hazard will be possible after new recording stations currently being established by the Nevada Seismological Lab (NSL) begin to return data. The community can join this effort at http://alerttahoe.seismo.unr.edu .

Tectonic SettingSeismologists’ recent recognition of earthquake hazards in the Tahoe basin

is partly due to its unique location (Figure 1). The Tahoe basin is a fault-controlled half-graben situated on the western edge of the Walker Lane, the youngest and most western basin in the Great Basin (Stewart, 1988; Unruh et al., 2003). Strain is partitioned between right-lateral shear and extension within the Walker Lane. Extension within the northern part of the Walker Lane shows evidence for a westward migration or encroachment trending from Walker Lake to Lake Tahoe since approximately 20 Ma (Surpless et al., 2002). Typical geomorphology within the northern Walker Lane consists of a series of fault-controlled, east- and west-tilted blocks that make up the north-trending mountain ranges and associated valleys. Together with several northwest-trending dextral fault systems both north and east of Lake Tahoe, the Walker Lane accommodates approximately 20-25 percent of the Pacific-North American plate motion (Faulds et al., 2005; Wesnousky, 2005).

Over the past decade, paleoseismic studies in the basin have identified at least three active normal faults with Holocene displacement (Figure 1). These three recognized faults; the West Tahoe—Dollar Point fault (WTDPF), the Stateline—North Tahoe fault (SNTF) and the Incline Village fault (IVF) have minimum vertical slip rates of 0.6, 0.45, and 0.2 mm/year, respectively (Kent et al., 2005; Dingler et al., 2009; Brothers et al., 2009). Additional work by Maloney et al. (2013) and Smith et al. (2013) have identified evidence to support at least three to four Holocene ruptures of the WTDPF with a recurrence interval of approximately 3,500 years. Evidence from seismic CHIRP studies, which have been validated through lake core samples (Maloney et al., 2013; Smith et al., 2013; Kent et al., 2016, this volume), shows that the last rupture of the WTDPF occurred at

Louie et al. 3

approximately 4,500 years ago. Based on the fault area and average event displacement, we suggest that the WTDPF is about 1000 years beyond its characteristic recurrence interval for a magnitude 7 earthquake (Maloney et al., 2013; Kent et al., 2016, this volume). The ground motions from such an event could be potentially catastrophic to the communities in the basin, especially if mitigation measures have not been undertaken.

Just outside the Tahoe basin, Ramelli and Bell (1998; 2012) assess the Genoa fault system to the east as capable of vertical slip as large as 5 m, with rupture lengths as long as 100 km. The most recent event dates to about 300 years ago, with the penultimate event at 1700 years, suggesting a recurrence interval of about 1500 years. Hunter et al. (2011) discovered the Polaris fault to the north, revealing a 35-km-long dextral strike-slip rupture zone with a Holocene slip rate of 0.4 mm/yr, which may be capable of an earthquake perhaps as large as M6.9.

Previous Work on Shaking in Basins Local observations immediately after the 1906 San Francisco earthquake indicated that certain areas in that city experienced more significant shaking than others (Lawson, 1908; Reid, 1910; McPhee et al., 2007). Lawson (1908) discusses the correlation between “made” land, which prior to the development of the city was below the water line, and higher ground shaking intensity. However, it was not until scientists noted the anomalously high, and very destructive, far-field ground motions during the 1985 Mexico City earthquake that researchers began developing 3-D modeling techniques to help explain basin effects. Salt Lake City (Olsen, 1994; Olsen and Schuster, 1994; Olsen and Schuster, 1995; Olsen et al., 1995b; Olsen et al., 1996), Southern California (Bielak et al., 1999; Graves, 1998; and Olsen and Archuleta, 1996; Olsen et al., 1995a; Olsen, 2000; Olsen et al., 2003), and the Seattle basin (Frankel et al., 2007; 2009) were among the first areas in the U.S. where basin analysis was included in the study of ground motion. Three-dimensional basin modeling differs from the USGS-sponsored ShakeMap procedure in that it requires detailed geological and geotechnical information specific to the area. TeraShake2, used by the Southern California Earthquake Center (SCEC), is one such model. It is the 3-D component of these models that allows for more reliable earthquake ground motion predictions (Olsen et al., 2008; SCEC, 2009).

Louie et al. 4

Figure 1 – Map of the Lake Tahoe basin showing the location of the major faults within the basin, the West Tahoe—Dollar Point fault (WTDPF), the Stateline—North Tahoe fault (SNTF), and the Incline Village fault (IVF); and two major faults outside the basin, the Polaris fault (PF), and the Genoa fault (GF). Seismic stations are shown as triangles and seismic events are shown as stars. Dashed box shows the scenario model area.

Louie et al. 5

METHODS

NSZ Model Development and Data SourcesThis project’s innovation is to create scenario-shaking models for the Tahoe

basin that make use of 3-D geological and geophysical data, and reflect the physics of wave propagation. Nevada ShakeZoning is an open-source numerical modeling environment (Louie, 2008; Louie et al., 2011; Flinchum et al., 2014; http://crack.seismo.unr.edu/NSZ) that can provide this capability to engineers and geologists, as well as seismologists. In this work NSZ interfaces to the E3D finite-difference seismic-wave propagation code developed by Shawn Larsen of Lawrence Livermore National Laboratory (Larsen et al., 2001). NSZ assembles specific geological and geophysical information for an area, creates 3-D grids, and then feeds the grids into the E3D code. More recent wave-modeling codes can also execute NSZ grids, such as the SW4 code maintained by the Computational Infrastructure for Geodynamics at http://geodynamics.org .

We used local geophysical datasets in the development of the Lake Tahoe basin model, including information on both basin structure and geotechnical shallow shear-wave velocity profiles. To evaluate basin shape, we use information from Saltus and Jachens (1995), as well as gravity points collected during this study, and points culled from the PACES Gravity Database at the University of Texas, El Paso (UTEP, 2013). A basement-gravity removal method derived from Saltus and Jachens (1995) and Abbott and Louie (2000) estimated sedimentary basin thicknesses within the southern portion of the Tahoe basin, under South Tahoe (Figure 1). Dense gravity measurements are not yet available for the northwestern part of the South Tahoe basin near Fallen Leaf Lake, or over Lake Tahoe. As a result, the poor data coverage in these areas leaves artificial boundaries between the gravity-derived South Tahoe thickness model and the rest of the lake.

Shawn Larsen of Lawrence Livermore National Laboratories developed the finite-difference elastic-wave modeling code E3D that the NSZ modeling environment drives (Larsen et al., 2001; Louie, 2008; Louie et al., 2011a,b; Flinchum et al., 2014). The parameters needed for E3D to produce specific wave behavior models include 3-D P- and S-velocity, density, and attenuation (QP and QS) grids. NSZ compiles the specific regional geophysical data, in this case the basin geometry and geotechnical velocities of the Tahoe basin, into 3-D grids holding these required parameters.

 South Tahoe Basin Model from Gravity– The basin model used to create this simulation is based on the results of Saltus and Jachens (1995), whose data yield a rough approximation of basin depth for the modeled region, east of longitude 120°W. In South Tahoe a UNR Applied Geophysics class collected 27 new gravity measurements in March 2012. We supplemented this information with approximately 250 existing gravity data points collected from the University of Texas, El Paso gravity database and from a study provided by the South Tahoe Public Utility District (STPUD). In certain areas, such as over the lake itself, there are gaps where gravity data are non-existent. In these areas, NSZ roughly estimates basin thickness based on proximity to bedrock outcrops. Joining the basins with depths estimated by very different procedures and data sets produces an artificial step in the basin floor, marked with “A” in Figure 2.

Louie et al. 6

Figure 2A through 2F – Maps of calculated 3-D peak ground velocity (PGV) values for each NSZ Tahoe intra-basin fault scenario, overlain on a shaded-relief basin-depth map. The color scale for ground shaking emphasizes, in orange and yellow, the areas of expected violent to extreme shaking for each scenario. The ShakeMap scale bar below allows association of NSZ’s computed PGVs with peak ground acceleration (PGA) levels, as developed by Worden et al. (2012). Maximum predicted PGV is 1.7 meter/sec (solid yellow, suggesting a PGA level of 130%g), with the strongest shaking confined to the deepest part of the basin. “A” marks the basin-depth-map artifact at the edge of the South Tahoe gravity model. The Meyers sub-basin “M”, Twin Peaks “T”, WT-DPF, Tahoe Keys “K”, Stateline “S”, Tahoe City “C”, and Incline Village “I” are discussed in the paper and are shown on their respective models.

Louie et al. 7

The NSZ process produces a 3-D grid file used for computing each scenario, occupying at least the area of Figure 1 outlined with a white dashed line. To evaluate the velocity structure, all of the available geotechnical velocities are interpolated into the surface of the 3-D grids. For the Tahoe basin model, we used a distance-weighted average with a specified search radius of one kilometer from measured values. Where geotechnical data are not available (such as within the Lake) we provide default values, which NSZ assigns to these areas. For the models shown in this paper, we assigned rock units a default Vs30 of 660 meter/sec, and soil units a default Vs30 of 250 meter/sec in basins where sediments are more than 10 m thick. Vs30 is the time-averaged shear-wave velocity from the surface to 30 m depth, as defined by BSSC (1998).

South Tahoe ReMi™ Geotechnical Measurements– For the southern portion of the basin, where the deepest terrestrial sediments are located, this project collected approximately 75 shallow shear-wave velocity measurements in June and July of 2012. Each measurement employed 88-meter-long, 12-channel array recordings of ambient microtremor, analyzed with the SeisOpt® ReMi™ software (© 2012, Optim) for a shear-velocity versus depth profile, to depths of 50 m or more. We distributed the measurements around South Tahoe’s suburban street network. The methods of Louie (2001) are the basis of this surface-wave sounding technology. The new geotechnical information provides more detailed modeling parameters at the surface and should allow for more realistic ground motion calculations.

Nevada ShakeZoning produces two attenuation grids that are centered about the peak frequency of the computed seismic source at 0.5 Hz. Estimated Q values for P and S waves are based on P and S velocities through the empirical relations described in Olsen et al. (2003) for the Los Angeles Basin. To estimate density, NSZ applies Gardner’s rule (Gardner et al., 1984) to estimate density from compressional velocity Vp. Shear velocity is estimated from a Vp-over-Vs ratio of the square root of three.

Within the basins below a geotechnical depth of 30 m, the Saltus and Jachens (1995) regionally averaged density versus depth profile gives an initial density estimate from which the velocity and Q parameters are derived. Outside the basins and within the bedrock below geotechnical depths, NSZ starts with a one-dimensional regional P-velocity model used for earthquake location from Smith and others (2001), and then estimates Vs, density, and Q values by applying the same relations used within the basins. Nevada ShakeZoning applies a minimum Vp criterion of 1.45 km/sec and a minimum Vs criterion of 0.36 km/sec to every grid-node cube, 200 m on a side.

According to Larsen et al. (2001), higher frequencies can be calculated for smaller grid areas using the same computational effort as for calculating lower frequencies over a larger area. In this case, the geography of the basin dictates the size of the grid. The area being modeled, approximately 50 km long and 24 km wide (dashed white box on Figure 1), encompasses the entire Tahoe basin. Despite the relatively large size of the model grid, we were able to calculate frequencies up to 0.5 Hz with acceptable accuracy using a grid node spacing of 200 meters, and frequencies up to 0.36 Hz without any grid dispersion artifacts. The gravity-derived basin-thickness model and ReMi™-measured geotechnical map for South Tahoe, and all other NSZ model inputs and parameters are available at http://crack.seismo.unr.edu/NSZ .

Louie et al. 8

Ground motion in the Lake Tahoe basin is modeled on geologically plausible event scenarios. Based on previous work (Brothers et al., 2009; Maloney et al., 2013; Smith et al., 2013; and Kent et al., 2016, this volume), normal faults in the basin are capable of producing up to a magnitude 7.2 earthquake with a vertical offset of approximately 3.5 meters, over a distance of ~50 km in length. Earthquakes of this magnitude should produce significant ground shaking throughout the basin.

This work models rupture scenarios for three faults within the Tahoe basin, and two faults outside the basin. Because information regarding the most likely hypocenter is not available for any of these faults, we evaluate several distinct scenarios for WTDPF, SNTF, PF, and GF ruptures. The IVF, because of its limited size and location within the basin, is only modeled with one hypocentral source. Each of these fault scenarios, except for the right-lateral PF, has a strictly normal sense of motion, and each is considered a reasonable event.

The WTDPF is modeled as a plane with a length of 44 kilometers, striking N8°W and dipping 70° to the east. The fault width is 15 km, with the top of the fault at approximately 2 km below the surface. The hypocenter is located at 9 kilometers below the surface. Three different scenarios are modeled for this fault, with the moment magnitude estimated to be approximately 7.1 for each event. Two of the scenarios for this fault include a rupture originating on the southern end of the fault and propagating northward, and a rupture originating on the northern end of the fault and propagating southward. Fault offset for these scenarios is estimated at a constant 3 meters, and is assumed to be evenly distributed over the entire fault plane.

Due to the resolution constraints of the available geophysical surveys, it is unknown whether the entire length of the WTDPF ruptures during an event, or instead, if discrete sections rupture at intervals of minutes to hundreds of years. For the third WTDPF scenario, we have modeled rupture for the north segment of the WTDPF. The modeled rupture length is 27 kilometers, striking N8°W and dipping 70° to the east. The fault width is 15 km, with the top of the fault at approximately 2 km below the surface, and with the hypocenter located at 9 kilometers below the surface. This model shows a source located on the north end of the segment and propagating southward. Fault offset for this scenario is estimated at 2 meters over the length of the fault.

The SNTF is modeled as a plane with a length of 21 kilometers, striking N11°E and dipping 70° to the east. The fault width is 15 km, with the top of the fault at approximately 2 km below the surface and the hypocenter located at 11 kilometers below the surface. Two different scenarios are modeled for this fault, with a moment magnitude of approximately 6.9 for each event. These include a rupture originating on the southern end of the fault and propagating northward, and a rupture originating on the northern end of the fault and propagating southward. Fault offset for these scenarios is estimated at 3 meters over the length of the fault.

The Incline Village fault is modeled as a plane with a length of 11 kilometers, striking N15°E and dipping 70° to the east. The hypocenter is located at 11 kilometers below the surface. Only one scenario is modeled for this fault, with an approximate moment magnitude of 6.6; this scenario depicts a rupture originating on the southern end of the fault and propagating northward. Fault offset for this scenario is estimated at 2 meters over the length of the fault.

Louie et al. 9

Outside the Tahoe basin, the Genoa fault system (GF) is modeled as a single plane with a length of 44 kilometers, striking N7°W and dipping 65° to the east. The fault width is 15 km, with the top of the fault at approximately 2 km below the surface with the hypocenter located 9.5 kilometers below the surface. Two different scenarios are modeled for this fault, with a moment magnitude of approximately 7.1 for each event. These include a rupture originating on the southern end of the fault and propagating northward, and a rupture originating on the northern end of the fault and propagating southward.

The Polaris fault (PF) is modeled as a plane with a length of 35 kilometers, striking N50°W and dipping 90°. The fault width is 15 km, with the top of the fault at approximately 2 km below the surface and the hypocenter located 12 kilometers below the surface. Both modeled scenarios have moment magnitudes of approximately 7.0, with one scenario involving a rupture originating on the northwestern end of the fault and propagating southeast toward the Tahoe basin, and the other with rupture propagation away from Tahoe toward the northwest.

An over-simplified rectangular fault plane with the dimensions and orientations given above represents each of the ten different rupture scenarios we modeled. Moment is constant across the kinematic rupture rectangle, as is the 2.8 km/sec rupture velocity. Each rupture element receives a smooth Gaussian slip-velocity versus time function as described by Larsen et al. (2001), with a rise time of 2 seconds, yielding synthetic seismograms with 0.5 Hz central frequency.

Simplified Probabilistic Hazard MappingFrankel et al. (2007; 2009) developed a modified probabilistic seismic

hazard map for Seattle that includes 3-D basin-amplification effects. The full probabilistic hazard mapping method requires evaluation of a large set of recorded strong-ground-motion data. Such data do not exist for Tahoe; nonetheless we seek to provide Tahoe communities with as much information as possible, in the interim before the USGS is able to more fully develop hazard mapping with basin effects, and data sets for Tahoe. The amplifications introduced by 3-D basin effects are substantial enough to dominate the expected-shaking results, overwhelming the uncertainties introduced by the lack of strong-motion recordings of earthquakes on Tahoe-area faults.

Following Frankel et al. (2007; 2009), we use a simplified method to estimate the annual rate of exceedance (ARE) of 0.30 meter/sec (30 cm/sec PGV) horizontal ground shaking velocity across a map of the Tahoe basin. Horizontal shaking at a velocity of 0.30 meter/sec is on the threshold of what ShakeMap (Wald et al., 1999; Worden et al., 2012) terms “severe” perceived shaking, with potential for moderate to heavy damage to buildings and infrastructure. On ShakeMap’s scale (Worden et al., 2012), this 0.30 meter/sec PGV level associates with a peak ground acceleration (PGA) of about 30% of the acceleration of gravity, or 30%g. We calculated the annual rates of exceedance (ARE), using the PGV maps computed for each fault rupture scenario, and annual return rates for each main fault within the basin, and the two faults outside the basin.

Each simulation contains the effects of both rupture directivity and the Tahoe and other basin amplification effects. The fundamental equation of PSHA determines the annual frequency λ of exceeding ground motion μ0 at a site from multiple faults or source locations by summing over source locations and magnitude as shown in Equation 1 below, from Frankel et al. (2007):

Louie et al. 10

λ (μ≥ μ0 )=∑M

∑source j

rate (M ,source j )P (μ≥ μ0|sitei , source j ,M ) , Eq. 1

where rate (M, sourcej) is the annual rate of occurrence for an earthquake with magnitude M at source location j. This annual rate can be determined from either a time independent or time dependent calculation; here, we use time-independent earthquake probabilities (Poissonian distribution of inter-event times), so the probability of exceeding ground motion μ0 in time t equals 1-e-λt .

The second factor (P) on the right-hand side of Equation 1 represents the probability of having ground motions μ greater than or equal to μ0 at site i, if an earthquake occurs at source location j with magnitude M. In a typical PSHA calculation, this factor is determined using a set of standard attenuation (ground-motion prediction) relations where the ground motion amplitudes depend only on the magnitude M and the distance D from the site to the fault. Here we modify this term so that the ground motion amplitudes μ are also dependent on the specific site and earthquake locations.

To follow Frankel et al. (2007), we should use the 3-D finite-difference simulations from E3D in the calculation of μ (1 Hz S.A.) for the basin and a set of ten scenario earthquakes. We could start with the ground motion predicted at each site μrock(M,D) from generic rock-site attenuation relations (for example, Abrahamson and Silva, 1997). Then we could calculate the site- and source-specific amplification using the 3-D simulations, so that:

μ=μrock (M ,D ) A3D (site i , source j) Eq. 2

However, we implement a much simpler scheme that provides a quick map of the Annual Rate of Exceedance (ARE) of the 0.30 meter/sec PGV level of shaking (30%g PGA), or any other level. Our quick estimate is based solely on the NSZ-computed PGV maps for each scenario, and the annual rate of each scenario. All of this information is really only roughly estimated. We do not consider any additional factors, such as attenuation relations, that may bring in considerable additional uncertainty for a region as poorly characterized as Tahoe. The simplistic ARE is determined with only:

ARE(μ≥μ0 | sitei) = Σsourcej ([(sitei, sourcej)/0]½ rate(sourcej)¿ , Eq. 3

with ARE(μ≥μ0 | sitei) the total annual rate of exceedance of ground motion 0

(0.30 meter/sec PGV) at site j on the map, (sitei, sourcej) the NSZ-computed PGV at site i for source scenario j, and the annual rate of scenario sourcej, summing rates from all sources j at each location i shown on the ARE map.

These ARE scenarios are calculated for shaking exceeding a threshold “severe” PGV value of 0.30 m (30%g PGA). Therefore, each point on the PGV map is divided by the threshold value, and the square root is taken to get an estimate of the relative probability of that level of shaking from that earthquake scenario. This normalized probability value is then multiplied by the annual return rate for each

Louie et al. 11

faulting scenario. Table 1 collects together the return-rate information we have for each of the five faults. Return rates are the inverse of return periods. Return periods for both the WTDPF and the SNTF are considered to be similar at 4,500 years, while the return period for the IVF is calculated to be 17,500 years. We assign the GF a return period of 1,500 years based on the interval between the latest and penultimate events dated by Ramelli and Bell (1998; 2012). The PF we give a period of 7,500 years based on the 0.4 mm/yr slip rate in Hunter et al. (2011) divided into the 3 m scenario slip. Where a fault such as the WTDPF has more than one computed rupture scenario, we assume each scenario to be equally likely, and the severe-shaking ARE values on each scenario’s map are divided by the number of scenarios for that fault.

Our quick severe-shaking ARE maps do conflate PGV levels exceeding our severe 0 of 0.30 meter/sec (30%g PGA) with higher annual rates of exceedance, with square-root dependence. PGV levels below 0 are conflated with lower ARE. While not strictly correct, the resulting ARE errors are less than the great uncertainties in the activity of the faults in question. Without this conflation, areas on the PGV maps experiencing less shaking than 0 would be given an ARE of exactly zero, which would be unrealistic, even dangerously non-conservative. Our simplistic severe-shaking ARE maps give the community a robust product they can use in economic models, assisting them in creating resilience.

Using information obtained from previous studies (Dingler et al., 2009; Brothers et al., 2009; Hunter et al., 2011; Ramelli and Bell, 1998 and 2012; Maloney et al., 2013; Smith et al., 2013; Kent et al., 2016, this volume), we estimated the average recurrence interval for the maximum expected event on each of the three major faults in the basin and the two outside the basin (Table 1). Since the GF and PF rates are taken from the work of others, we only present here their average rates. We used the PGV values from the ten scenario maps calculated for the Tahoe basin and then divided each value by threshold severe μ0PGV value of 0.3 meter/sec, take the square root, multiply by each scenario’s rate, and then sum the resulting rate maps over the ten scenarios. Summing annual rates is the operation called for by Eq. 1 to combine scenarios.

Fault Average Rate, /yr

MaximumRate, /yr

Minimum Rate, /yr

No. of Scenario

sProbability of Each Scenario

WTDPF 1/4,500 1/3,500 1/5,400 3 0.00007/year

SNTF 1/4,500 1/3,500 1/5,400 2 0.0001/year

IVF 1/17,500 1/15,000 1/20,000 1 0.00005/year

GF 1/1,500 - - 2 0.0003/year

PF 1/7,500 - - 2 0.00007/yearTable 1 – Recurrence Intervals for Earthquakes Affecting the Tahoe Basin

Comparisons Between Synthetics and DataChecking that our model parameters are representative of the basin, and

that the model has the potential to simulate realistic ground motion is an important

Louie et al. 12

step in model development. Although we cannot yet fully perform this step, we can still make a few comparisons of earthquake events culled from the Nevada Seismological Laboratory catalog (Univ. of Nevada, Reno, 1971), against synthetic events developed with rupture parameters similar to the recorded events. These details are discussed below.

Recorded Comparison Event Data– Multiple seismic stations are present within or near the basin (Figure 1), though only the Rubicon Station (RUB) and Willow (WIL) are high-quality digital stations with significant events recorded. Short-period, high-gain analog recordings for the events in Table 2 are available from the Emerald Bay (EMB) station in the basin, and the Genoa (GNO) station outside the basin to the east. No recording stations within the South Tahoe basin were yet available for this study. High-gain analog data are of limited use; here they provide only the travel time of the earthquake waves between the source and seismic station.

ModeledEvent Latitude Longitude Magnitude Moment

dyne-cmStrike

DegreesDip

DegreesRake

DegreesDepth

km146980 39.3097 -120.0602 4.69 1.2x1023 319 80 -177 1.0183059 39.2240 -120.0802 2.99 3.0 x1020 88 71 -90 1.0300426 39.3195 -119.9814 3.5 1.8x1021 69 71 -90 1.0323541 39.2320 -120.1101 3.24 6.0 x1020 337 71 -144 6.7380267 39.3178 -119.9904 4.21 1.32x1022 11 84 29 2.0972622 38.8325 -119.7924 3.55 3.0 x1021 330 70 -150 1.0

Table 2 – Source parameters of modeled comparison earthquakes

Recorded data were retrieved from the Nevada Seismological Laboratory (NSL) catalog of recent events. This study evaluated only a catalog of events greater than M2.5 and occurring within, or in close proximity to, the basin. The locations of these events are shown in Figure 1, labeled with their unique numerical IDs. Of this sample, only those events recorded digitally are used for comparison with the synthetic seismograms produced from NSZ models.

Though numerous earthquakes greater than a M2.5 have occurred within the basin since the year 2000, mostly in its northern portion, for this paper we examine six of the largest earthquakes that are more geographically distributed around the basin. Seismograms from these events are compared to synthetic seismograms created for similar model events utilizing NSZ. The information for these earthquakes is shown in Table 2.

Synthetic Comparison Data– For each recorded event, data collected by the NSL (Univ. of Nevada, Reno, 1971) are either resolved into a moment tensor solution (for the larger magnitude events) or a short period first motion solution (for the smaller magnitude events). Using this information, as well as available geological and seismological information (such as known fault strikes, aftershock distributions, etc.), the strike and dip of the fault plane can be reasonably ascertained. Using the cataloged local magnitude, we estimated the seismic moment for each event and used this information, as well as the depth and rake of the event.

For each NSZ model, the fault and earthquake rupture parameters are input to NSZ’s ModelAssembler community modeling environment. Due to the simplistic rupture model and large, 200-meter node spacing, the sources need not be located

Louie et al. 13

with high accuracy. The 3-D grid area for modeling had to be extended to include all sources and stations on Figure 1, and the faults outside the basin in Table 1, a larger area than that in the white-dashed box used for intra-basin scenario modeling.

RESULTSCalculating shaking for wave frequencies up to 0.5 Hz with the E3D code

took several hours for each of the ten scenario events from Table 1, and longer for each of the comparison events in Table 2, which required larger grids. For simplicity, computations were completed on single nodes of small clusters running a 32-bit Linux operating system. Thus the light computational effort matches the detail we have available for Tahoe basin geology and geotechnical measurements– many dozens of measurements in the South Tahoe portion of the basin, but very little and very general data from the other parts of the basin. We hope to inspire the collection of additional data and the development of more comprehensive and detailed models for Tahoe, which would then justify a greater computational effort, to reach wave frequencies as high as 1.0 or even 2.0 Hz.

Each E3D scenario computation resulted in multi-component time-histories of the particle velocity of shaking at key sites, including the recording stations yielding comparative data; map time slices animating wave propagation across the basin; and summary maps of the horizontal peak ground velocity (PGV) attained at all surface points of the grid. Figure 2 shows the summary PGV maps resulting from the six scenarios on the three faults within the Tahoe basin. The ShakeMap scale (from Worden et al., 2012) at the bottom of Figure 2 allows the reader to make an association between a color level on a PGV map to the potential shaking in terms of peak ground acceleration or PGA, in terms of the percentage of the standard acceleration of gravity, or “%g.” Space constraints prevent printing of any of the resulting seismograms or wave animations here. Figure 3 shows a few frames from the animations; however, all software, models, datasets, and computed results are available through http://crack.seismo.unr.edu/NSZ .

West Tahoe—Dollar Point Fault – Full Rupture from South HypocenterPeak ground velocity (PGV) results from this scenario indicate that strong

shaking occurs above the entire rupture length (western white line in Figure 2A), but is particularly strong in the Meyers sub-basin (Figure 2A, “M”). PGV within the confines of this small basin, almost 500 m deep and updip of the hypocenter, approaches an extreme 1.7 meters per second with very heavy damage potential. Here we adopt the shaking-severity terminology found on USGS ShakeMap products, after Wald et al. (1999), and Worden et al. (2012). A PGV of 1.7 meter/sec would be associated with an extreme PGA level of 130%g. Twin Peaks, a granite mountain situated between the main Lake Tahoe basin and the smaller sub-basin, still experiences severe ground motion at approximately 0.43 meters per second (associated with a PGA level of 40%g), with moderate to heavy damage potential (Figure 2A, “T”). An artificial step in the basin floor occurs at the boundary of the gravity-based South Lake Tahoe basin model, where basin thickness can only be guessed from bedrock proximity. It is worth noting that this artifact does not seem to have a prominent effect on the PGV results (Figures 2A-2F, “A”).

Figure 3A is a wave-propagation visualization of the basin for this scenario in map view, showing the calculated ground motion at 12.15 seconds after the

Louie et al. 14

rupture commences. At this point in time, the P-wave is just exiting the northern part of the basin near Tahoe City (Figure 3A, “C”). Long wavelength, high amplitude Rayleigh waves (alternating bright yellow/green and purple/blue) are propagating northward through the main portion of the basin. As expected, Rayleigh waves modeled in the competent Sierra granites on either side of the basin exhibit lower amplitudes (duller colors). As the Rayleigh waves propagate northward, shorter wavelength Rayleigh and Love waves are trapped in the southern portion of the basin and the Meyers sub-basin (Figure 3A, “M”). Radially propagating waves emanating from energy trapped in these southern basins are visible as lower amplitude waves exiting the sub-basin in all directions (Figure 3A).

At approximately 51 seconds after the initial rupture, the entire basin is filled with short wavelength Rayleigh waves (Figure 3B). The initial northward wave propagation is no longer apparent. Instead, the remaining energy from the rupture is trapped and reverberating within the basin. Energy is still trapped within the Meyers sub-basin (Figure 3B, “M”), though by this time the amplitude has diminished. Wave amplitudes are highest in the Tahoe City (Figure 3B, “C”) and Incline Village (Figure 3B, “I”) areas at this later time.

Figure 3A through 3C – Maps of wave propagation modeled for rupture scenarios on the WTDPF. For this figure, the color red represents shaking in the east and west direction (red arrow), green represents shaking in the north and south direction (green arrow), and blue represents up and down shaking (blue X). White represents shaking in all directions and shaking intensity is represented by the brightness of the color. Wave-motion colors are overlain on a shaded-relief map of basin-floor topography as assembled by NSZ. Figure 3A shows the wavefield at 12.15 seconds after the rupture. The P, S, Rayleigh and Love waves are identified. Long-wavelength Rayleigh waves are the most dominant feature in this figure. The Meyers sub-basin “M” is shown with waves reverberating within it, from which radially propagating waves emanate. Figure 3B is shown at approximately 51 seconds after the initiation of the rupture. This map shows the waves reverberating within both the main basin and the smaller Myers sub-basin. This reverberation is the source of the additional radial waves leaving the basin. Figure 3C shows a frame from

Louie et al. 15

a north-source rupture model of the West Tahoe fault at approximately 13.8 seconds after initiation. The frame shows the Rayleigh wave hitting the Tahoe Keys “K” but having less impact at Stateline “S;” without the 3-D effects of the basin edges, both locations would see more similar shaking. Refraction of surface waves on the map, around the basin edges, focuses the Rayleigh wave on the Keys. The large directivity pulses on the seismograms with the north-south Y component, larger than the east-west X component, and the Keys seismogram larger than the Stateline, allow the same conclusions.

West Tahoe—Dollar Point Fault – Full Rupture from North HypocenterPGV results from this alternate scenario for rupturing the WTDPF (Figure

2B) indicate that the strongest shaking occurs above the source of the rupture and along the strike of the rupturing fault (western white line on Figure 2B). Ground motions are still amplified within the Meyers sub-basin (Figure 2B, “M”) but not by as much as when the rupture initiates at the south end of the fault (compare to Figure 2A, “M”). The highest PGV value for this rupture scenario is approximately 1.4 meters per second (~110%g PGA), still extreme, and diminishes to a moderate <0.1 meter per second (<10%g PGA) away from the basin.

Figure 3C is a wave visualization frame captured at approximately 13.8 seconds after the initiation of a southward propagating rupture from the northern part of the WTDPF. With this scenario, the Tahoe Keys area (Figures 2B and 3C, “K”) exhibits much stronger shaking than observed at Stateline (Figures 2B and 3C, “S”). Without the 3-D effects of the basin edges, both locations would see more similar shaking, being similar distance and azimuth from the fault.

West Tahoe—Dollar Point Fault – Northern Segment RuptureResults from this scenario (Figure 2C) indicate that the strongest shaking

occurs near the source of the rupture and along the strike of the rupturing fault. The highest PGV value for this rupture scenario is approximately 0.88 meters per second (~75%g PGA), violent with heavy damage potential, and diminishes to a more moderate <0.1 meter per second (<10%g PGA) outside the basin. Though less intense than for the previous models, the Meyers sub-basin (Figure 2C, “M”) still exhibits increased shaking over the surrounding granite. The calculated maximum PGV for this basin is strong to severe, approximately 0.32 meters per second (~30%g PGA).

Stateline—North Tahoe Fault – Rupture from South HypocenterResults from this scenario (Figure 2D) indicate that the strongest shaking

occurs at the rupture epicenter, with PGV approaching 1.3 meters per second (~110%g PGA), extreme. This model predicts fairly strong PGV values near Tahoe City and Incline Village (Figure 2D, “C”, “I”). Estimated PGV values for these areas are approximately 0.8 meters per second (~70%g PGA), violent with heavy damage potential. Ground motion dies off quickly south of McKinney Bay (Figure 2D, “B”), with a strong PGV of approximately 0.1 meter per second. The Meyers sub-basin (Figure 2D, “M”) exhibits very little ground motion as a result of this event, though its light PGV value of 0.02 meter/sec (~3%g PGA), with no damage potential, remains higher than that of the surrounding bedrock.

Stateline—North Tahoe Fault – Rupture from North HypocenterResults from this scenario (Figure 2E) indicate that the strongest shaking

occurs at the rupture epicenter with PGV approaching an extreme 1.5 meters per second. This model predicts fairly strong PGV values near Tahoe City and Incline Village. Estimated PGV values for these areas are approximately 0.8 and 0.6

Louie et al. 16

meters per second (60-70%g PGA), respectively, still violent with heavy damage potential. Ground motion is more pronounced in the basin for this scenario, with the Meyers sub-basin exhibiting light PGV values around 0.2 meters per second (~20%g PGA). Again, ground motion outside the basin is weak.

Incline Village Fault – Rupture from South HypocenterOf the three faults modeled, the Incline Village fault (Figure 2F, “I”) is the

shortest fault and is expected to have the smallest magnitude earthquake. Results from this scenario (Figure 2F) indicate that the strongest shaking occurs at the rupture epicenter with PGV approaching a severe 0.5 meters per second (~50%g PGA), with moderate to heavy damage potential. The ground motions die off quickly away from the epicenter, though the highest PGV is located within Incline Village (Figure 2F, “I”). Ground motion is still observed in the basin for this scenario, though PGV values are light and expected to be around 0.02 meters per second (~3%g PGA). The Meyers sub-basin (Figure 2F, “M”) exhibits some light PGV values, while the ground motion outside the basin is weak, below 0.01 meters per second (<2%g PGA).

ShakeMap and No-Basin ShakeZoningIn addition to modeling each scenario with the corresponding basin and

geotechnical information, each scenario is modeled in Nevada ShakeZoning without the basin information; as well as in the USGS ShakeMap environment. The results from these models are very similar to each other, though substantially different from the full Nevada ShakeZoning model that includes the 3-D basin effects. The results for the full WTDPF rupture from the south show the maximum PGV prediction across the region obtained from ShakeMap (using the software of Wald et al., 1999, and Worden et al., 2012) is approximately 0.69 meters per second (~60%g PGA). The maximum PGV prediction obtained from Nevada ShakeZoning, without the basin information in the model, is 0.89 meters per second (~75%g PGA). These values are similar, as are the PGV maps (Figures 4A and 4B). This indicates that Nevada ShakeZoning procedures will yield similar predictions to the USGS’s ShakeMap procedures, given identical 1-D geological models.

Louie et al. 17

Figure 4A through 4C – PGV maps showing a comparison between ShakeMap predictions (Figure 4A), no-basin Nevada ShakeZoning (Figure 4B), and the shaking amplification due to the presence of the 3-D basin (Figure 4C). Figure 4A has a maximum PGV of 0.69 meters per second (~60%g PGA). Figure 4B has a maximum PGV of 0.89 meters per second (~75%g PGA). Figure 4C shows the NSZ PGV-ratio relationship where Figure 2A is the numerator and Figure 4B is the denominator. In figure 4C, the color red indicates where the PGV is greater than predicted by NSZ (yellow where >5x greater), with the 3-D basin, than without the basin; and blue indicates where the shaking is less than that predicted by NSZ with the basin. White indicates no difference between the two models, a PGV ratio of 1.0.

Annual Rates of Exceedance (ARE) of “Severe” ShakingFigures 5A through 5I show the simplistic ARE maps for a threshold severe

earthquake-shaking peak ground velocity (PGV) of 0.3 meter/sec (30%g PGA), estimated for each of the scenarios discussed above, with square-root conflation of higher or lower PGV than 0.3 meter/sec with the ARE according to Eq. 3. The highest severe-shaking ARE values approach 4 x 10-4 per year (1/2500 per year), observed in the Genoa fault (GF) rupture scenario from outside the basin (Figure 5G).

Many of the intra-basin scenarios surpass 1.6 x 10-4 per year (1/6250 per year). For each scenario, the high severe-shaking ARE values are near the source and follow the fault as expected. In Figure 5A, high severe-shaking ARE values are not only observed near the fault, but also in the southern part of the Lake Tahoe basin and within the Meyers sub-basin (“M”). This is not unexpected considering the proximity of these basins to the source fault; however, the severe-shaking ARE for these southern basins is anomalously frequent when compared to the surrounding areas.

Louie et al. 18

The highest severe-shaking ARE for the SNTF southern source is within Lake Tahoe proper, and therefore well away from developed communities. The highest severe-shaking ARE values for the SNTF north source are predominantly in the lake, though high values do extend beneath Tahoe City (Figure 5E, “C”) and towards Incline Village (Figure 5E, “I”).

The highest 0.3 meter/sec severe-shaking ARE for the IVF is beneath the lake just near Incline Village (Figure 5F, “I”). This is a fairly small zone, with the higher rates falling off quickly away from the fault. The severe-shaking ARE does follow the shape of the basin southward, though this value is also relatively low. For the four GF and PF scenarios, two of which are shown in Figures 5G and 5H, high annual return rates extend into the Tahoe basin.

Figure 5I is a summed map showing the ARE of 0.3 meter/sec PGV (roughly 30%g PGA according to Worden et al., 2012) for the entire basin and including the combined hazard associated with all ten scenarios on all five faults. As expected, the highest severe-shaking ARE is in the deepest part of the basin (“D”). High severe-shaking ARE values are also observed in the Meyers sub-basin (“M”), even though the highest hazard, a rate of exceedance of 1/700 per year, is in the northern half of the basin. The implication is that severe earthquake shaking with potential for moderate to heavy damage is expected in many parts of the Tahoe basin every 1400 years. Some parts of north Tahoe such as Incline Village (“I”) can expect severe shaking as often as every 1100 years.

Comparisons Between Synthetics and DataIn order to test the models used to predict the annual rates of exceedance,

we compared event seismograms recorded within Tahoe basin with synthetic seismograms for those events. Comparing the travel time between the source and the recording station between recorded and synthetic events is the main comparison we can make, and illuminates the effectiveness of the velocity model NSZ constructs. Because only two of the stations are digital, and event data are limited, much of our comparison depends on first arrival times. However, in the few instances where digital data are available, we compared the waveforms between the recorded and synthetic events. Prior to waveform comparison, some of the digital data were de-meaned and tapered prior to being band-pass filtered from 0.05 to 2.0 Hz.

Table 3 shows the first arrival times (FAT) for each of the recorded scenarios. Table 3 suggests ±10% FAT error. With the exception of the FAT for event 380267 to the Genoa station (GNO) and the FAT of event 323541 to the Rubicon station (RUB), all of the model velocities averaged over the path from event to station are within 10 percent error. Both of the outliers have short first arrival times, indicating that they are fairly close to the source. Small variations between our model and the actual basin structure may be enough to cause this discrepancy. Larger distances between a source and receiver allows for a greater averaging of travel times between the two. This would be especially noticeable for areas where the Vs30 values are not fully modeled and the variable subsurface material would more greatly influence the wave behaviors.

Louie et al. 19

Louie et al. 20

Figure 5A through 5I – These figures map the annual rate of exceedance (ARE) of a shaking level of 0.30 meter/second peak ground velocity (PGV), for each scenario over the Tahoe basin. (Equivalent to about 30%g PGA on the ShakeMap scale of Worden et al., 2012.) The WTDPF scenarios are shown in Figures 5A, 5B, and 5C. The SNTF scenarios are shown in figures 5D and 5E. The IVF scenarios is shown in Figure 5F. The selected GF and PF scenarios are shown in Figures 5G and 5H, respectively. Figure 5I shows the summed severe-shaking ARE for the entire basin and includes all ten scenarios on the five faults modeled. The highest severe-shaking ARE is in the deepest “D” part of the basin, with a return period of less than 1000 years.

Due to the limited digital station availability in the basin, amplitudes and waveforms of the synthetic seismograms were compared only for the Rubicon (RUB) and Willow (WIL) stations. The Rubicon station is located on the west side of the lake, just north of Meeks Bay near Quaternary glacial moraines. The upgraded station installed in 2011 has a broadband seismometer and strong motion instrument with six channels total, whereas the Willow station is located south of the basin on pre-Tertiary granodiorite bedrock, and only recorded Events 300426 and 183059. Waves from both events traversed the entire Tahoe basin north to south, on their way to WIL.

EventAct. FATEMB

Synth. FATEMB

Act. FATRUB

Synth. FATRUB

Act. FATGNO

Synth. FATGNO

Act. FATDON

Synth. FATDON

Act. FATWIL

Synth. FATWIL

146980 7.7 7.5 6.6 6.5 NA NA NA NA NA NA

183059 NA NA NA NA NA NA 5.5 5.9 10.1 9.2

300426 NA NA NA NA NA NA NA NA 10.1 10.4

323541 6.0 6.0 5.6 4.8 7.4 8.1 NA NA NA NA

380267 7.2 8.0 7.8 7.1 2.9 3.9 NA NA NA NA

972622 6.4 6.9 7.7 8.3 7.9 8.7 NA NA NA NA

Table 3 – Recorded (red) and synthetic (blue) first-arrival times (FAT). All values in seconds.

The largest earthquake, Event 146980, occurred on June 26, 2005 and has a ML of 4.69. It occurred on a northwest striking fault in the northern portion of the basin. At RUB, within Tahoe basin, the filtered seismograms show some recorded-to-synthetic amplitude correlation between the easting X and northing Y-axes. All of our waveform matches are poor, so we do not show any of the recorded or synthetic seismograms. Maximum recorded horizontal ground velocity (PGV) is 0.0029 cm/sec; the synthetic PGV is 0.0019 cm/sec, as close as Flinchum et al. (2014) could get matching recordings and NSZ synthetics in Las Vegas basin. The vertical component seismogram shows a first-arrival-time correlation between the recorded and synthetic data (Table 3). The duration of shaking for the synthetic event is approximately 55 seconds and the duration for the recorded event is considerably longer, approximately 90 seconds.

Event 183059 occurred on June 11, 2006 and has a ML of 2.99. It occurred on a northeast striking fault in Carnelian bay on the northwest side of the lake. The recording at WIL appears to represent the summation of two separate seismic events occurring closely in time to one another; this could partly account for the complete mismatch between the amplitudes of the recorded and synthetic data.

Louie et al. 21

Recorded PGV is 0.0016 cm/sec; the synthetic PGV is 0.70 cm/sec, two orders of magnitude larger. Despite the lack of amplitude correlation, the first arrival times are similar (Table 3). The Donner (DON) short-period station also recorded this event, and provided a similarly close FAT match.

Event 300426 occurred on December 23, 2009 and has a ML of 3.5. It occurred on a northeast striking fault in the northern part of the basin. Recorded at WIL, there is some correlation between both the first arrival times and the amplitudes of the seismograms with the synthetics. Recorded PGV is 0.0041 cm/sec; the synthetic PGV is 0.0016 cm/sec, less than half as large. Z-component recorded and synthetic PGVs were more in accord, at 0.0016 cm/sec and 0.0010 cm/sec respectively. The duration of shaking for the modeled event is approximately 55 seconds, while the duration of shaking for the recorded event is similar at approximately 65 seconds.

Event 972622 occurred on January 1, 2011 and has a ML of 3.55. It occurred on a northwest-striking fault in the southern part of the region near the Genoa fault in Carson Valley (Figure 1). The Rubicon station recorded this event, with the waves traversing the South Tahoe basin from southeast to northwest. Recorded PGV is 0.0009 cm/sec; the synthetic PGV is 0.0035 cm/sec. In this case the recording has smaller amplitude than the synthetic, about one-third. Table 3 shows more than a 10% FAT mismatch for this event. These discrepancies probably reflect our poor state of knowledge of the shape of the western side of the Tahoe basin near RUB and EMB. The FAT mismatch at GNO, very close to the 972622 event, similarly shows our poor knowledge of the Carson Valley basin. The seismograms between the recorded and modeled event are similar. Despite the amplitude discrepancy, the shaking duration is approximately 50 seconds for both recording and synthetic.

DISCUSSIONLevel and Distribution of Predicted Shaking– The 2014 USGS Hazard Map

for 2% probability of exceedance in 50 years (Petersen et al., 2014) predicts PGA levels at 40-70%g across most of the Tahoe Basin, reaching 80%g at Incline Village. For some of the basin, in some of the scenarios, our predictions are similar. However, the >1.0 meter/second PGV (associated with >90%g PGA) we predict in limited areas such as Meyers (Figure 2) from WTDPF scenarios exceeds the USGS shaking predictions by factors of two or more.

Even where our highest-impact scenarios (Figure 2) suggest similar levels of shaking as the USGS 2014 Hazard Map (Petersen et al., 2014), our severe-shaking ARE map (Figure 5I) suggests more frequent hazardous shaking. Much of the Tahoe basin, including Stateline and Incline Village, could see severe shaking more often than every 1400 years, half the 2500-year interval of the 2% in 50 years USGS Hazard Map. In South Tahoe where the expected shaking is comparable between the two maps, Figure 5I suggests a shorter interval of 1200 years for damaging shaking. The center of the Tahoe basin, and small sub-basins such as at Meyers, show much higher shaking at much shorter intervals than the USGS maps. Both maps include the Genoa fault as a prominent source, but Petersen et al. (2014) do not include the WTDPF in their model.

Results from this study indicate that 3-D, geology- and physics-based Nevada ShakeZoning scenario modeling predicts significantly stronger, extreme ground motions within the sedimentary basins of Lake Tahoe, compared to outside the basins (Figure 2). This finding differs from models developed using 1-D USGS

Louie et al. 22

ShakeMap procedures (Wald et al., 1999; Worden et al., 2012). ShakeMap produces a PGV map with fairly even ground motions across the entire basin, depending mainly on distance from the fault (Figure 4A). Comparing the 3-D NSZ-predicted PGV map of Figure 2A with the 1-D ShakeMap of Figure 4A and the 1-D no-basin NSZ map of Figure 4B suggests that both methods are predicting the same total wave energy. It is the distribution of shaking that is different between the 1-D and 3-D models; the PGV-ratio map of Figure 4C shows that the 3-D Tahoe sedimentary basin traps shaking within it. The trapping results in high levels of shaking within the basin, not predicted by ShakeMap, and lower levels of shaking than predicted by ShakeMap outside the basin.

3-D Basin Effects in Synthetics– The multicomponent synthetic seismograms created for each 3-D scenario evaluate detailed wave propagation properties within the basin. One would expect that motion on the easting X-axis would be the largest component of shaking for a basically north-striking, east dipping normal fault, in the absence of 3-D effects. The first model discussed is for the West Tahoe – Dollar Point fault (WTDPF), with a northern hypocenter rupturing to the south. For this scenario, one of the most striking results is the considerable shaking observed near Stateline in South Lake Tahoe. The 0.5-Hz synthetic seismogram (not shown) indicates shaking of long duration, over 30 seconds at levels close to a severe PGV value of 0.5 meter/sec (on Figure 2B, at “S”), roughly equivalent to 50%g PGA on the ShakeMap scale of Worden et al. (2012). The easting X-axis synthetic seismogram from this location is larger than the northing Y-axis seismogram. This indicates that the polarization of shaking in Stateline may not be closely correlated with 3-D basin effects. Instead, the geotechnical parameters of the soft soils, together with the overall size of the basin and its predominant thickness, may be controlling the large PGV and duration for this area.

Proximal to Echo Estates, by the Meyers sub-basin near Twin Peaks (Figure 2B, “H” and “T”, respectively) the easting X and northing Y components of the predicted shaking are similar, which indicates 3-D basin focusing and trapping effects. In fact, the north-directed Y component of shaking at Stonehenge (a landmark residence within Echo Estates) is unexpectedly large, suggesting that wave energy is trapped in the Meyers sub-basin (Figure 2B, “M”) adjacent to the site. Similarly, the Tahoe Keys (Figure 2B, “K”) and the Tahoe Airport (Figure 2B, “T”) are also subject to basin trapping and extreme shaking. Near the Tahoe Airport (“T”), the easting X component is much smaller than the northing Y component, indicating that 3-D basin effects are strongly affecting polarization in this area. For both of these sites, there are 30 seconds of very strong >0.2 meter/sec shaking (>22%g PGA).

The models also indicate the effects of rupture directivity in these areas, which could exacerbate any potential liquefaction hazards. Liquefaction occurs when the effective stress of a saturated soil is reduced to a value near zero. This corresponds to a complete loss of shear strength. During cyclic loading (such as that observed during earthquakes), the change in pore water pressure is greater than the strength of the soil, resulting in soil failure (Robertson and Fear, 1995; Liu et al., 2001). The Tahoe Center for Environmental Science at Incline Village (Figure 2B, “I”), would experience long-duration shaking, with several secondary arrivals, indicating wave propagation and trapping spanning the length and breadth of the basin (e.g., Figure 3).

A second model on the WTDPF for a north-rupturing, southern hypocenter indicates somewhat different results. For the easting X-axis seismogram near

Louie et al. 23

Angora Ridge (Figure 2A, “G”), the seismogram does not indicate rupture directivity nor does it indicate much shaking from surface waves trapped in the basin. Near Stateline (Figure 2A, “S”), being out of the path of the rupture’s directivity produces less shaking than for the northern-rupture scenario (compare to Figure 2B, “S”). However, the duration of shaking here is much longer than for any of the other sites. The longer duration results from waves reflecting off the basin sides, and propagating back and forth across the basin (as in Figure 3B). From an engineering standpoint, this could be problematic as anything damaged during the initial pulse could be further damaged or destroyed by the ongoing shaking (Bommer et al., 2009; Hancock & Bommer, 2005; Green & Terri, 2005; and Liu et al., 2001). Near the Tahoe Keys, the easting X-component of shaking is larger and more complex than the northing Y-component. For this area, liquefaction could be a significant threat if liquefiable layers are present.

Shaking Effects of Tahoe Basin and Small Sub-Basins– Based on studies by Olsen et al. (2006 and 2008), rupture simulations on the southern San Andreas fault using TeraShake suggested high PGV amplification values would be observed along a string of sedimentary basins extending from the San Bernardino basin through Whittier Narrows and into the Los Angeles Basin. The TeraShake model indicates that these motions exceed the 2% probability level predicted by the current empirical attenuation relationships, which do include a thickness-dependent basin amplification term (Campbell and Bozorgnia, 2006). The strongest shaking occurs in the smallest, most restricted basins, a “funneling” effect of wave-guide focusing.

This study makes similar observations. Small sub-basins amplify the PGV predictions to extreme levels; support for this can be seen when comparing areas such as Stonehenge to areas such as the Tahoe Science Center or the Rubicon station (RUB). Both the Rubicon station and the Tahoe Science Center are situated within the main Tahoe basin, not a smaller sub-basin. Near Stonehenge (Figure 3A, “N”), which is located on the edge of the small Meyers sub-basin (Figure 3A, “M”), shaking is significantly amplified and the northing Y-component of shaking is significantly increased, leaving no doubt that sub-basins have a strong effect on shaking amplitudes and polarizations.

Figure 4C shows the ratio between the PGV maps computed from two NSZ models. The ratio compares ground motions at each point on the map, with the 3-D basin-included full NSZ computation (Figure 2A) on the numerator, and the no-basin 1-D NSZ computation (Figure 4B, similar to the 1-D USGS ShakeMap output of Figure 4A) on the denominator. Blue indicates areas where the NSZ basin-included model has a lower PGV than either the NSZ basin-free or the ShakeMap models. Red indicates areas where the NSZ basin-included model (from Figure 2A) has a higher PGV than either the NSZ basin-free and the ShakeMap models; white is areas without much PGV difference between the two models. The yellow, shown within McKinney Bay, is where the highest PGV ratio values occur, a factor of five. This ratio map points to the fact that the yellow area has an NSZ PGV prediction (a violent 1.09 meter/sec, Figure 2A, suggesting nearly 100%g PGA) five times higher than the prediction derived from ShakeMap (only a very strong 0.22 meter/sec – 25%g PGA – for ShakeMap on Figure 4A; or 0.20 meter/sec – 22%g PGA – for the no-basin NSZ on Figure 4B).

Interestingly, the area of highest PGV ratio is within McKinney Bay, the source area for a large-volume (~6 km3) submarine landslide within the last 60,000 years (Kent et al., 2005). Based on our models, the PGV values in much of this

Louie et al. 24

region are four times those predicted by standard ShakeMap methods. There are at least two ways to interpret this prediction. One would be that the resulting shape of McKinney Bay after the landslide is trapping the seismic energy, resulting in a sub-basin effect within a larger basin. This reasoning suggests mutual exclusivity between the triggering of the landslide, and the high PGV values.

A second possibility is that the high PGV values predicted in this region indicate that earthquake shaking may have contributed to the occurrence of the McKinney Bay landslide. Landslide events are easily recognized within this portion of the basin (Smith et al., 2013; Maloney et al., 2013), suggesting that the till and lacustrine sediment here are less competent, weaker, and likely exhibit slower shear-wave velocity than the surrounding Sierran granites. The weaker rocks, together with the focusing provided by the shape of the basin overall, lead one to expect higher PGV values, which in turn could contribute to the landslide hazard. Careful examination of landslide hazard is beyond the scope of this study. Further work, such as deep Refraction Microtremor (ReMi) surveys (Pancha and Pullammanappallil, 2012; 2014) to evaluate the geotechnical properties of the material underlying the basin floor, might help quantify slide hazards.

In updating the SCEC TeraShake2 model, which is derived from a more complex source of rupture on the southern San Andreas fault than for the original TeraShake model, Olsen et al. (2008) describe that although the new model excites a less coherent wave field, showing less along-strike directivity, the TeraShake2 simulations still predict wave entrainment within basins, and a strong directivity pulse. Peak ground velocity values predicted in both the Los Angeles and San Gabriel basins are much higher than predicted by empirical, non-3-D studies. Olsen et al. (2008) goes on to make several points about basin effects on earthquake waves:1. Both the Los Angeles and San Gabriel basins have higher than anticipated PGV

values. These values are attributed to strong directivity pulses.2. Large parts of both of these basins have predicted PGV above the 2%

probability of exceedance (POE) level relative to current attenuation relationships.

3. Wave-guide focusing produces localized basin areas with PGV at roughly 0.1%–0.2% POE, a factor of 4.5 above the median.

4. Rock sites in the 0 to 100 km range of distance from the rupture show PGV values derived from TeraShake2 to be in close agreement to median empirical prediction and extremes nowhere reach the 2% POE level.

Our research in the Lake Tahoe basin predicts similar characteristics. Our severe-shaking ARE maps (Figure 5) suggest much of Tahoe basin is threatened by higher shaking levels at higher POE than predicted by the 2014 USGS Hazard Maps (Petersen et al., 2014). Directivity of the rupture plays a significant role in our PGV and shaking predictions; the strongest, most potentially damaging directivity pulse originates from the WTDPF northern rupture source (Figure 2B). Basin-driven directivity, as well as coincident high PGV values, result in sustained severe ground shaking beneath Tahoe Keys and the Tahoe Airport, such as shown in Figures 2A and 2B. Intense shaking, which also increases the threat of liquefaction, is of primary concern for these sites.

Basin-Edge Effects– Angora Ridge, a glacial moraine sitting within close proximity to the bedrock WTDPF trace (Figure 2A, “G”), exhibits high amplitude ground motion, but for a fairly short duration. This is seen for the Rubicon station (RUB) as well. Not surprisingly, the Stonehenge site (N), also located near the

Louie et al. 25

fault, is exhibiting extreme ground motions. What is surprising, however, is the high value of the 3-D NSZ PGV prediction. This site is situated on the edge of the smaller Meyers sub-basin within the larger Tahoe basin (Figure 2A, “M”). The amplitudes for this site are approximately 1.5 times larger than those observed at the Tahoe Airport, despite the fact that the geotechnical velocities at the Stonehenge site are far higher than those at the Tahoe Airport (T). This may be explained by basin edge effects, first examined by Olsen (1994) and Graves et al. (1998), as well as proximity to the fault rupture.

As expected, the annual rate of exceedance (ARE) maps for 0.3 meter/sec PGV (30%g PGA), a moderately to heavily damaging level of shaking (Worden et al., 2012), show larger, more frequent return values near the epicenters and within the basins. As shown in the PGV maps (Figures 2A-2F), high severe-shaking ARE values within the basins are to be expected, as the basin-floor structure in 3-D tends to trap, and the basin fill tends to amplify, the ground motions from any nearby earthquake. This prediction is in accord with point 3 above of Olsen et al. (2008). For Figures 5A through 5I, the Lake Tahoe basin channels the highest severe-shaking ARE values, while the rates become smaller within the surrounding bedrock. Reduced probabilities of intense shaking are predicted outside basins, in accord with point 4 of Olsen et al. (2008) above.

Prediction Accuracy– This study develops a model to predict the maximum PGV values for the basin. It also attempts to compare the NSZ model against recorded earthquake events. The purpose is to see how well the NSZ model performs when tested against recorded events. Given the very few good recordings we have from the basin, the travel time between the seismic source and a particular station is the indicator of model quality we can examine most carefully. Even high-gain analog stations can yield accurate travel times. For this study, we have modeled travel times for events within the basin to stations both within and outside of the basin. The difference between recorded and synthetic travel time for all of the events is less than 1.0 second. This is true even for stations located outside of the basin and further away from the source where error would be more prominently observed. This suggests that our NSZ model is a reasonable representation of the basin. Given those checks on the 3-D velocity model, together with our prior experience validating 3-D NSZ shaking computations for Las Vegas (Flinchum et al., 2014), the PGV values obtained here appear to be acceptable. The synthetics have some similarities to the recorded wave pulses with respect to peak ground velocity, timing, and phase. In the comparison between observed and synthetic seismograms at 0.5 Hz (not shown), it is clear that NSZ predicts a similar order of magnitude of times for the durations of shaking, though the method under-predicted shaking duration for one event and over-predicted shaking duration for another event. However, the ground motion shown on the event seismograms indicates that substantial shaking can occur for more than 40 seconds longer than the NSZ model predicts. Once the NSL records more seismogram data in the basin, we can undertake proper validation.

CONCLUSIONSThe Lake Tahoe basin is a major tourist destination and has some of the

highest home and real estate values in the USA (Antonucci, 2010). Throughout most times of the year, this basin has a relatively small population. However, during peak holiday times, the population can be quite large. A magnitude 7.0 earthquake would have an enormous impact on the local population, environment,

Louie et al. 26

lake clarity, economy, tourism, and home values. We have shown here that the current 1-D empirical method of predicting earthquake shaking, USGS ShakeMap (Wald et al., 1999; Worden et al., 2012), is non-conservative, ineffective, unrealistic and potentially dangerous within the Tahoe basin. In our 3-D predictions based on physics and detailed geological measurements, large earthquakes reasonably expected within the basin will cause “violent” to “extreme” levels of shaking, risking heavy to very heavy damage to buildings. The shaking will continue at dangerous levels for long durations of a minute or more. In limited areas of the basin, shaking levels will exceed those predicted by the USGS Hazard Maps (Petersen et al., 2014) by more than a factor of two. Ground particle motions could reach velocities of 1.7 meters per second, implying peak horizontal ground accelerations (PGA) exceeding 140% of standard gravity. Summing the effects of ten of the possible earthquake-rupture scenarios on the five major faults in and around the Tahoe basin, these dangerous ground motions can be expected to occur twice as frequently as predicted by the USGS Hazard Maps. More than 1 km outside the sedimentary basins, on Sierran bedrock, the USGS predictions remain conservative.

The Tahoe basin also provides an important test bed for methods that can subsequently be applied to other earthquake-prone areas. How many regions, underlain by basins, have the potential for higher than anticipated ground motions? How can an engineer design a structure capable of resisting collapse during an earthquake, if the severity of ground motions cannot be reasonably constrained? How can the economy be sustained, and building undertaken, without constraining the annual rate of earthquake damage?

Impacts of this Study– The responsibility of the seismologist is to develop predictive accuracy so that response spectra, shaking duration and peak ground motions are realistic and useful to engineers. This requires that models be updated and validated so that members of the engineering community have the best tools at their disposal for efficient design. By improving the models, which will in turn help improve communication between seismologists and engineers (Abrahamson and Gregor, 2006), the entire community benefits.

This study implements a new method for modeling earthquake-derived ground motions, yielding improved predictions in this complex region. Each model discussed here is a geologically probable scenario, and the geologic information has been carefully constrained through a large number of geophysical measurements. The results of this study indicate that dangerously severe earthquake shaking for the Tahoe region is more likely than previously thought, providing a strong indication of the urgency of improved seismic hazard studies. In the interim before the USGS can update its hazard maps to include all the earthquake faults and all the 3-D basin effects, we hope our simple severe-shaking ARE maps will provide guidance to Tahoe communities seeking improved earthquake resilience.

ACKNOWLEDGEMENTSOptim Inc. and the Nevada Applied Research Initiative funded the CCoG computing cluster. Dr. Satish Pullammanappallil of Optim kindly provided specialized equipment, trained personnel, data analysis, and access to proprietary software. Shawn Larsen of LLNL installed his E3D code at The Nevada Seismological Laboratory (NSL) at the University of Nevada, Reno (UNR). JNL carried out portions of the research at Victoria University of Wellington, New Zealand, partly sponsored by sabbatical leave from UNR. The NSL provided additional seed funding for Tahoe earthquake research. Nevada ShakeZoning

Louie et al. 27

software development, data collection, and research were partly supported by the U.S. Geological Survey (USGS), Department of the Interior; under USGS award numbers 07HQGR0029, 08HQGR0015, 08HQGR0046, G09AP00050, G09AP00051, and G10AP00002 to Louie and others, and awards G11AP20022, G12AP20026, G14AP00020, and G15AP00055 to Optim. The Northern Nevada Seismic Network is supported under USGS award G10AC0090, and the Western Great Basin Seismic Network Operations is supported under USGS award number G10AC0090. 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. Richard Saltus and Bob Jachens of the USGS kindly made their basin-thickness results available to us. The U.S. Forest Service Lake Tahoe Basin Management Unit through Joey Keely, the California Department of Transportation District 3 Maintenance Division, the Tahoe Keys Property Owners Association, and numerous homeowners very kindly granted us access to their lands for geophysical surveying.

REFERENCESAagaard, B.T., Brocher, T.M., Dolenc, D., Dreger, D., Graves, R., Harmsen, S., Hartzell, S.,

Larsen, S., and Zoback, M.L., 2008, Ground motion modeling of the 1906 San Francisco Earthquake, Part I: Validation using the 1989 Loma Prieta earthquake, Bulletin of the Seismological Society of America, 98(2), 989-1011.

Abbott, R.E., and Louie, J.N., 2000, Depth to bedrock using gravimetry in the Reno and Carson City, Nevada area basins, Geophysics, 65, 340-350.

Abrahamson, N., and Gregor, N., 2006, Crossing the seismology-engineering interface, Seismological Research Letters, 77(2), 272.

Abrahamson, N.A., and Silva, W.J., 1997, Empirical response spectral attenuation relations for shallow crustal earthquakes, Seismological Research Letters, 68(1), 94-127.

Antonucci, D.C., 2010, Tahoe FAQs, http://www.tahoefacts.com/#/lake-tahoe/, accessed April 22, 2015.

Bielak, J., Xu, J. and Ghattas, O., 1999, Earthquake ground motion and structural response in alluvial valleys, Journal of Geotechnical and Geoenvironmental Engineering, 125(5), 413–423.

Bommer, J.J., Stafford, P.J., Alarcón, J.E., 2009, Empirical equations for the prediction of the significant, bracketed, and uniform duration of earthquake ground motion, Bulletin of the Seismological Society of America, 99(6), 3217-3233.

Brothers, D.S., Kent, G.M., Driscoll, N.W., Smith, S.B., Karlin, R., Dingler, J.A., Harding, A. J., Seitz, G. G., and Babcock, J. M., 2009, New constraints on deformation, slip rate, and timing of the most recent earthquake on the West Tahoe—Dollar Point fault, Lake Tahoe Basin, California, Bulletin of the Seismological Society of America, 99(2A), 499-519.

Building Seismic Safety Council (BSSC), 1998, NEHRP Recommended Provisions for Seismic Regulation for New Buildings, FEMA 302/303, developed for the Federal Emergency Management Agency, Washington, D.C.

Campbell, K.W., Bozorgnia, Y., 2008, NGA ground motion model for the geometric mean horizontal component of PGA, PGV, PGD, and 5%-damped linear elastic response spectra for periods ranging from 0.01 to 10 s, Earthquake Spectra, 24(S1), 139–172.

Dingler, J., Kent, G., Driscoll, N., Babcock, J., Harding, A., Seitz, G., Karlin, B., and Goldman, C., 2009, A high-resolution seismic CHIRP investigation of active normal faulting across Lake Tahoe basin, California-Nevada, Geological Society of America Bulletin, 121(7-8), 1089-1107.

Faulds, J.E., Henry, C.D., Hinz, N.H., 2005, Kinematics of the northern Walker Lane: An incipient transform fault along the Pacific-North American plate boundary, Geology, 33(6), 505-508.

Louie et al. 28

Flinchum, B.A., Louie, J.N., Smith, K.D., Savran, W.H., Pullammanappallil, S.K., and Pancha, A., 2014, Validating Nevada ShakeZoning predictions of Las Vegas basin response against 1992 Little Skull Mtn. earthquake records, Bulletin of the Seismological Society of America, 104(1), 439-450, doi: 10.1785/0120130059. Preprint with color figures at http://crack.seismo.unr.edu/NSZ/validation/Flinchum-et-al-BSSA-preprint-color.pdf

Frankel, A.D., Stephenson, W.J., Carver, D.L., Williams, R.A., Odum, J.K, and Rhea, S., 2007, Seismic hazard maps for Seattle, Washington, incorporating 3-D sedimentary basin effects, nonlinear site response, and rupture directivity, U.S. Geological Survey Open-File Report 2007-1175, 77 pp., 3 plates.

Frankel, A., Stephenson, W., and Carver, D., 2009, Sedimentary basin effects in Seattle, Washington: Ground-motion observations and 3-D simulations, Bulletin of the Seismological Society of America, 99(3), 1579-1611.

Gardner, G.H.F., Gardner, L.W., and Gregory, A.R., 1984, Formation velocity and density - the diagnostic basics for stratigraphic traps, Geophysics, 39, 770-780.

Graves, R.W., Pitarka, A., and Somerville, P.G., 1998, Ground motion amplification in the Santa Monica area: Effects of shallow basin edge structure, Bulletin of the Seismological Society of America, 88, 1224–1242.

Green, R.A., and Terri, G.A., 2005, Number of equivalent cycles concept for liquefaction evaluations— revisited, Journal of Geotechnical and Geoenvironmental Engineering, 131(4), 477–488.

Hancock, J., and Bommer, J.J., 2005, The effective number of cycles of earthquake ground motion, Earthquake Engineering and Structural Dynamics, 34(6), 637–664.

Hartzell, S., Harmsen, S., Frankel, A., and Larsen, S., 1999, Calculation of broadband time histories of ground motion: comparison of methods and validation using strong-ground motion from the 1994 Northridge earthquake, Bulletin of the Seismological Society of America, 89(6), 1484–1504.

Hartzell, S., Harmsen, S., Williams, R.A., Carver, D., Frankel, A., Choy, G., Liu, P.C., Jachens, R.C., Brocher, T.M., and Wentworth, C.M., 2006, Modeling and validation of a 3D velocity structure for the Santa Clara Valley, California, for seismic-wave simulations, Bulletin of the Seismological Society of America, 96(5), 1851-1881, DOI: 10.1785/0120050243.

Hunter, L.E., Howle, J.F., Rose, R.S., and Bawden G.W., 2011, LiDAR-assisted identification of an active fault near Truckee, California, Bulletin of the Seismological Society of America, 101(3), 1162-1181.

Kent, G.M., Babcock, J.M., Driscoll, N.W., Harding, A.J., Dingler, J.A., Seitz, G.G., Gardner, J.V., Mayer, L.A., Goldman, C.R., Heyvaert, A.C., Richards, R.C., Karlin, R., Morgan, C.W., Gayes, P.T., and Owen, L.A., 2005, 60 k.y. record of extension across the western boundary of the Basin and Range province, Estimate of slip rates from offset shoreline terraces and a catastrophic slide beneath Lake Tahoe, Geology, 33(5), 365-368.

Kent, G.M., Schmauder, G.C., Smith, K., Driscoll, N., and Maloney, J., 2016, Reexamination of faulting in the Tahoe basin using airborne LiDAR and seismic CHIRP imaging, this volume.

Kim, A., Dreger, D.S., and Larsen, S., 2010, Moderate earthquake ground motion validation in the San Francisco Bay Area, Bulletin of the Seismological Society of America, 100(2), 819-825.

Larsen, S., Wiley, R., Roberts, P., and House, L., 2001, Next-generation numerical modeling: incorporating elasticity, anisotropy and attenuation, Society of Exploration Geophysicists Expanded Abstracts, 20, 1218-1221.

Lawson, A.C., 1908, The California earthquake of April 18, 1906, Carnegie Institution of Washington Publication 87, 1451 pp.

Liu, A., Stewart, J.P., Abrahamson, N., and Morikawi, Y., 2001, Equivalent number of uniform stress cycles for liquefaction analysis, Journal of Geotechnical and Geoenvironmental Engineering, 127(12), 1017–1026.

Louie et al. 29

Louie, J.N., 2001, Faster, better: shear-wave velocity to 100 meters depth from refraction microtremor arrays, Bulletin of the Seismological Society of America, 91(2), 347-364.

Louie, J.N., 2008, Assembling a Nevada 3-D velocity model: earthquake-wave propagation in the Basin & Range, and seismic shaking predictions for Las Vegas, Society of Exploration Geophysicists Expanded Abstracts, 27, 2166-2170.

Louie, J.N., Savran, W., Flinchum, B., Plank, G., Kent, G., Smith, K.D., Pullammanappallil, S.K., Pancha, A., and Hellmer, W., 2011, Next-Level ShakeZoning for earthquake hazard definition in the Intermountain West, presented at the Symposium on Engineering Geology and Geotechnical Engineering (EGGE), Las Vegas, NV, Mar. 25, 15 pp. Proceedings paper on line at http://crack.seismo.unr.edu/ma/scenarios/Louie-11EGGE.pdf.

Maloney, J.M., Noble, P.J., Driscoll, N.W., Kent G.M., Smith, S.B., Schmauder, G.C., Babcock, J.M., Baskin, R.L., Karlin, R., Kell, A.M., Seitz, G.G., and Kleppe, J.A., 2013, Paleoseismic history of the Fallen Leaf Segment of the West-Tahoe fault reconstructed from slide deposits in the Lake Tahoe Basin, California-Nevada, Geosphere, 9(4), 1065-1090, doi: 10.1130/GES00877.1.

McPhee, D.K., Langenheim, V.E., Hartzell, S., McLaughlin, R.J., Aagaard, B.T., Jachens, R.C., and McCabe, C., 2007, Basin Structure beneath the Santa Rosa Plain, Northern California: Implications for damage caused by the 1969 Santa Rosa and 1906 San Francisco Earthquakes, Bulletin of the Seismological Society of America, 97(5), 1449–1457.

Olsen, K.B., 1994, Simulation of three-dimensional wave propagation in the Salt Lake Basin, Ph.D. Thesis, University of Utah, Salt Lake City, Utah, 157 pp.

Olsen, K.B., and Schuster, G.T., 1994, Three-dimensional modeling of the site amplification in East Great Salt Lake Basin, U.S. Geological Survey Technical Report, 1434-93-G-2345, 102 pp.

Olsen, K.B., Archuleta, R.J., and Matares, J.R., 1995a, Three-dimensional simulation of a magnitude 7.75 earthquake on the San Andreas fault, Science, 270, p. 1628- 1632.

Olsen, K.B., Pechmann, J.C., and Schuster, G.T., 1995b, Simulation of 3-D elastic wave propagation in the Salt Lake Basin, Bulletin of the Seismological Society of America, 85, 1688-1710.

Olsen, K.B., and Schuster, G.T., 1995, Causes of low-frequency ground motion amplification in the Salt Lake Basin: The case of the vertically-incident P wave, Geophysical Journal International, 122, 1045-1061.

Olsen, K.B., and Archuleta, R.J., 1996, 3-D simulation of earthquakes on the Los Angeles fault system, Bulletin of the Seismological Society of America, 86, 575-596.

Olsen, K.B., Pechmann, J.C., and Schuster, G.T., 1996, An analysis of simulated and observed blast records in the Salt Lake Basin, Bulletin of the Seismological Society of America, 86, 1061-1076.

Olsen, K.B., 2000, Site amplification in the Los Angeles Basin from three- dimensional modeling of ground motion, Bulletin of the Seismological Society of America, 90, S77-S94.

Olsen, K.B., Day, S.M., and Bradley, C.R., 2003, Estimation of Q for long-period (>2 sec) waves in the Los Angeles Basin, Bulletin of the Seismological Society of America, 93, 627- 638.

Olsen K.B., Day, S.M., Minster, J.B., Cui, Y., Chourasia, A., Okaya, D., Maechling, P., and Jordan, T., 2008, TeraShake2: Spontaneous rupture simulations of Mw7.7 earthquakes on the Southern San Andreas fault, Bulletin of the Seismological Society of America, 98(3), 1162–1185, doi: 10.1785/0120070148.

Olsen K.B., Day, S.M., Minster, J.B., Cui, Y., Chourasia, A., Faerman, M., Moore, R., Maechling, P., and Jordan, T., 2006, Strong shaking in Los Angeles expected from southern San Andreas earthquake, Geophysical Research Letters, 33, art. no. L07305, doi 10.1029/2005GL025472.

Louie et al. 30

Olsen, K.B., and Mayhew, J.E., 2010, Goodness-of-fit criteria for broadband synthetic seismograms, with application to the 2008 Mw 5.4 Chino Hills, California, earthquake, Seismological Research Letters, 81(5), 705-723.

Pancha, A., and Pullammanappallil, S., 2012, Determination of 3D-velocity structure across the deepest portion of the Reno Area basin, Final Technical Report to the U.S. Geological Survey, External Grant Award No. G12AP20026, 54 pp. http://earthquake.usgs.gov/ research/external/ reports/G12AP20026.pdf.

Pancha, A., and Pullammanappallil, S., 2014, Determination of 3D-velocity structure across the northeastern portion of the Reno area basin, Final Technical Report to the U.S. Geological Survey, External Grant Award No. G14AP00020, 27 pp. http://earthquake.usgs.gov/research/external/ reports/G14AP00020.pdf.

Petersen, M.D., Moschetti, M.P., Powers, P.M., Mueller, C.S., Haller, K.M., Frankel, A.D., Zeng, Yuehua, Rezaeian, Sanaz, Harmsen, S.C., Boyd, O.S., Field, Ned, Chen, Rui, Rukstales, K.S., Luco, Nico, Wheeler, R.L., Williams, R.A., and Olsen, A.H., 2014, Documentation for the 2014 update of the United States national seismic hazard maps, U.S. Geological Survey Open-File Report 2014–1091, 243 pp., http://dx.doi.org/10.3133/ofr20141091.

Ramelli, A.R., and Bell, J.W., 1998, National Earthquake Hazards Reduction Program, Final Technical Report, 21 pp., scale 1:62,500.

Ramelli, A.R., and Bell, J.W., 2012, Spatial and temporal patterns of fault slip rates on the Genoa fault, Grant Award No. G09AP00020, 19 pp., 2 plts. http://earthquake.usgs.gov/research/external/ reports/G09AP00020.pdf

Reid, H., 1910, The Mechanics of the Earthquake, volume II of The California Earthquake of April 18, 1906, Report of the State Earthquake Investigation Commission, Carnegie Institution of Washington, Washington, D.C.

Robertson, P.K., and Fear, C.E., 1995, Liquefaction of Sands and Its Evaluation, Proceedings, 1st International Conference on Earthquake Geotechnical Engineering, Keynote Lecture, Tokyo, Japan.

Saltus, R.W., and Jachens, R.C., 1995, Gravity and basin-depth maps of the Basin and Range Province, Western United States, U.S. Geological Survey, Geophysical Investigations Map, Report: GP-1012, 1 sheet.

Smith, S.B., Karlin, R.E., Kent, G., Seitz, G.G., and Driscoll, N.W., 2013, Holocene subaqueous paleoseismology of Lake Tahoe, Geological Society of America Bulletin, 125(1-2), 691-708.

Southern California Earthquake Center (SCEC), 2009, Community Modeling Environment, from http://scec.usc.edu/research/cme/content/terashake-platform, accessed 30 April 2015.

Stewart, J.H., 1988, Tectonics of the Walker Lane belt, western Great Basin: Mesozoic and Cenozoic deformation in a zone of shear, in Ernst, W.G., ed., Metamorphism and crustal evolution of the western United States, Prentice Hall, Englewood Cliffs, New Jersey, 681-713.

Surpless, B.E., Stockli, D.F., Dumitru, T.A., and Miller, E.L., 2002, Two-phase westward encroachment of Basin and Range extension into the northern Sierra Nevada, Tectonics, 21(1), 2-1–2-10, doi: 10.1029/2000TC001257.

University of Nevada, Reno, 1971, Nevada Seismic Network, International Federation of Digital Seismograph Networks, Other/Seismic Network, doi:10.7914/SN/NN.

University of Texas, El Paso (UTEP), 2013, Pan American Center for Earth & Environmental Studies (PACES), http://research.utep.edu/Default.aspx?alias=research.utep.edu /paces, accessed 30 April 2015.

Unruh, J., Humphrey, J., and Barron, A., 2003, Transtensional model for the Sierra Nevada frontal fault system, eastern California, Geology, 31(4), 327–330.

Wald, D.J., Quitoriano, V., Heaton, T.H., Kanamori, H., Scrivner, C.W., and Worden, C.B., 1999, TriNet “ShakeMaps:” Rapid generation of instrumental ground motion and intensity maps for earthquakes in Southern California, Earthquake Spectra, 15, 537-556.

Louie et al. 31

Wesnousky, S.G., 2005, Active faulting in the Walker Lane, Tectonics, 24, TC3009, doi: 10.1029/2004TC001645.

Worden, C.B., Gerstenberger, M.C., Rhoades, D.A., and Wald, D.J., 2012, Probabilistic relationships between ground-motion parameters and Modified Mercalli intensity in California, Bulletin of the Seismological Society of America, 102, 204-221, doi: 10.1785/0120110156.

Louie et al. 32