Abstract: The Parkfield segment of the San Andreas Fault (SAF) is one of the best instrumented seismic regions, due to the famous Parkfield Earthquake Prediction Experiment (Bakun et al., 2005). As a result, the 2004 Mw6.0 Parkfield earthquake generated a wealth of data to study seismic activity before and after the mainshock. Recently, Shelly (2017) released a 15-yr catalog of more than 1 million low-frequency tremor events along the Parkfield-Cholame section of the SAF, based on waveform matching with 88 tremor families. However, except a few previous studies that focused on microseismicity in a short time period (Peng and Zhao, 2009; Meng et al., 2013) or repeating earthquakes only (Lengline and Marsan, 2009), there is no systematic long-term detection of microseismicity in this region. Building upon previous studies of microearthquake detection in the region (e.g. Peng and Zhao, 2009), we perform a systematic detection of microearthquakes with a template matched filter detection technique within one year around the 2004 Mw6.0 Parkfield earthquake mainshock. We take advantage of additional seismic datasets that have not been fully used before and improve the detection of early aftershocks in the first tens seconds after the mainshock. With the analysis of a larger time span, we intend to further investigate and understand the migration patterns of the aftershock sequence and determine if any seismic signals are detected in the time period before the mainshock that could indicate a foreshock sequence. We also plan to relocate those newly detected microearthquakes and search for repeating events that occur at virtually the same location. Updated results will be presented at the meeting.
Media has not been submitted for this Presentation
Abstract: Seismic surface waves are valuable for investigating subsurface structures. However, in many other applications such as in seismic reflection imaging, it is desirable to separate the surface waves from the data. We propose a data-driven approach to predict and separate surface waves from the data based on the nonlinear dispersion measurement. In addition, we can also separate the surface waves into fundamental mode and overtones. The procedure has two steps. We first estimate high-resolution surface wave phase velocities from the recorded data using our nonlinear signal comparison (NLSC) approach. This enables us to predict the surface waves at each receiver location. We then subtract the predicted surface waves from the input seismic data. We applied our approach on two synthetic datasets and one field active-source seismic gather. From these examples, we can see that our new approach could effectively predict and separate surface waves with high fidelity.
Media has not been submitted for this Presentation
Abstract: The USArray has facilitated a revolution in the integration of seismic imaging and geodynamic modeling. We present a whole-mantle, variable resolution, shear-wave tomography model based on newly available and existing seismological datasets including regional body-wave delay times and multi-mode Rayleigh and Love wave phase delays. The dataset of previously published delays contributes ~600,00 S+ body wave delays, distributed globally, ~160,000 S wave delays used in the DNA13 regional tomographic model focused on the western and central US, and ~86,000 S and SKS delays measured on stations in western South America. Additionally, we derive ~4,100,000 S+ body wave delays through correlation of observed waveforms with 1-D PREM synthetics computed with Syngine for stations in the continuous US, Alaska, and the global seismic networks (IU, II). The surface wave dataset includes fundamental mode and overtone Rayleigh wave data from Schaeffer and Lebedev (2014), ambient-noise derived Rayleigh wave and Love wave measurements from Ekstrom (2013), newly computed fundamental mode ambient noise Rayleigh wave phase delays for the continuous US up to July 2017, and other, previously published, measurements. These datasets, along with a data-adaptive parameterization utilized for the SAVANI model (Auer et al., 2014), allow significantly finer-scale imaging than previous global models, rivaling that of regional-scale approaches, under the USArray footprint in the continuous US, while seamlessly integrating into a global model. We parameterize the model for both vertically (vSV) and horizontally (vSH) polarized shear velocities by accounting for the different sensitivities of the various phases and wave types. The resulting, radially anisotropic model allows for a range of new geodynamic analysis, including estimates of mantle flow or seismic anisotropy, without generating artifacts due to edge effects, or requiring assumptions about the structure of the region outside the well resolved model space. Our model images a number of geologically important features, including the Farallon slab and its constituent components, a complex undulating cratonic lithosphere, and a broad low velocity zone under the Cordillera.
Media has not been submitted for this Presentation
Abstract: It is not a question of if or why, but of where, how, and when the next Cascadia Subduction Zone (CSZ) megathrust earthquake will impact the Pacific Northwest. Synthetic ruptures from kinematic source models have attempted to account for future CSZ earthquake behavior and illustrate the importance of rupture directivity on ground motion amplification and duration in forearc basin sediments (e.g., Olsen et al., 2008; Delorey et al., 2014). However, these models do not incorporate source dynamics and are potentially missing finite fault effects that stem from stress heterogeneity or varying frictional properties. In particular, the “gap” region along the megathrust separating episodic tremor and slip (ETS) events from the up-dip seismogenic region deserves further scrutiny (Bruhat & Segall, 2016, 2017) because its stress state could strongly influence down-dip rupture propagation and ground motions in populated regions such as Seattle and Portland. A realistic CSZ rupture model must thus prescribe an appropriate gap rheology in concurrence with a low effective normal stress in the ETS region and also consider different hypotheses about the degree of slip-deficit accumulated since the last megathrust rupture. We approach this problem with 2D dynamic rupture simulations along the northern, central, and southern CSZ. Earthquake rupture is treated as both a dynamic crack governed by a linear slip-weakening friction law and as a pulse operating under a rate-and-state-friction framework. We explore a range of potential nucleation sites and stress profiles along the seismogenic, gap, and ETS regions and discuss the implications for strong ground motion. Because seismic risk to the Pacific Northwest is increased if rupture can penetrate into the ETS region, our models attempt to constrain what physical conditions along the gap are necessary for this to occur and aim to inform seismic hazard analyses by respecting the earthquake source physics.
Media has not been submitted for this Presentation
Abstract: We present the results of a surface wave group velocity tomography study for the Rio de la Plata craton (RPC). This craton represents the oldest Precambrian region of the end of southwest Gondwana in South America. The results were then inverted to estimate crustal and lithospheric thicknesses. Previous studies carried out in South America did not map some areas of the continent such as the RPC, clearly because of the insufficient number of crossing paths. To improve the resolution of the previously obtained crustal and upper mantle images, the number of group velocity measurements for the craton area was increased, achieving a better coverage of paths and a more uniform azimuthal distribution which enhances the tomographic images. Surface wave dispersion curves were obtained by a multiple filter technique with a phase-matched filter to better isolate the fundamental mode. A 2D tomographic inversion of the group velocity was applied using a conjugate gradient method. Our results include both Love- and Rayleigh-wave inversions for periods from 10 to 100 s. The obtained group velocity maps correspond well with the main tectonic structures along the studied area. Inversions of the group velocities were carried out to obtain the S-wave velocity distribution in the crustal region.
Media has not been submitted for this Presentation
Abstract: Earthquake recurrence interval is a fundamental concept of current earthquake models and a key parameter in earthquake hazard assessments. Whereas much effort has been devoted to estimate and refine the recurrence interval of large earthquakes on various faults, increasing evidence, especially from intracontinental faults, starts to paint a different temporal pattern of earthquakes: clusters of earthquakes within short periods, separated by long periods of quiescence. Such earthquake patterns can be mathematically described by the Cantor function, or the devil’s staircases. Devil’s staircase is a fractal property of complex dynamic systems in nature, including earthquakes. We show that seismicity at all scales, from global occurrence of large earthquakes to different tectonic regions to individual faults, demonstrates the devil’s staircase patterns. The average length of the quiescence periods seems inversely related to the rates of tectonic loading. Thus large earthquakes in stable continents have longer quiescent periods than those in tectonically active continents, and earthquakes at plate boundary faults have the shortest quiescent periods. The periods and frequency of clustered earthquakes also vary significantly. We have developed three-dimensional viscoelasto-plastic finite element models to investigate the underlying physics controlling the temporal patterns of seismicity. Our preliminary results confirm that tectonic loading rate is the key parameter for the length of the quiescent periods, viscous relaxation is a key process for earthquake clustering, and fault interactions a key factor for the devil’s staircases of seismicity.
Media has not been submitted for this Presentation
