Earthquake Science Center Seminars

Elizabeth Reddy, Colorado School of Mines

The Sistema de Alerta Sísmica Mexicano is the world's oldest public earthquake early warning system. This talk by the author of the new book, ¡Alerta! (forthcoming in 2023 MIT Press), draws on over a decade of ethnographic and archival research to analyze how SASMEX and related Mexican efforts have sought to change relationships between the environment, society, and technology and develop insights relevant for risk mitigation projects more broadly. This talk will specifically address earthquake early warning in the context of ongoing, everyday life with both earthquakes and warnings—a project which can help us understand what it means for such projects to succeed, and what it will take to make that happen.

Ross Stein, Temblor

Valere Lambert

Determining principles and conditions governing motion along faults is crucial for assessing how earthquake ruptures start and how large they may ultimately become. One of the most notable constraints for the stress levels along mature plate boundary faults is the abundance of evidence that the resistance to shear motion during substantial fault slip must be low (< 20 MPa). Two end-member hypotheses are used to explain such low-stress operation of mature faults: either (I) mature faults are strong at slow, quasi-static sliding rates but undergo considerable enhanced dynamic weakening during rapid earthquake-producing slip rates, as widely observed in laboratory experiments; or (II) they are chronically weak, such as due to pervasive fluid overpressure. I will present recent progress exploring how the characteristics of different physical models for “weak” mature faults are reflected in a range of geophysical observations - including heat flow constraints, seismologically determined properties of earthquakes and earthquake frequency-magnitude statistics – using numerical simulations of earthquake rupture sequences with laboratory-derived fault friction laws, shear heating and effects of pore fluids. Our simulation results suggest that the two classes of fault models can, in principle, be distinguished based on the amount of seismic energy radiated from large earthquake ruptures, as well as earthquake frequency-magnitude statistics. Our findings taken with existing geophysical inferences suggest that persistently weak fault models with moderate enhanced dynamic weakening may be plausible representations of megathrust faults in natural subduction zones, with potential differences for some mature crustal strike-slip faults, or that existing estimates of seismically radiated energy are substantially underestimated. Our modeling results also highlight the potential scale dependence of average measures of shear stress and the importance of accounting for finite-fault effects, such as differences in how earthquake ruptures start and propagate, when relating local friction behavior determined in the lab to the field scale.

Fred Pollitz, USGS

Four large earthquakes struck the Intermountain West and Basin and Range Province in 2020 and 2021: M5.7 March 18, 2020 Magna, Utah; M6.5 March 31, 2020, Stanley, Idaho; M6.5 May 15, 2020 Monte Cristo Range, Nevada; and M6.0 July 8, 2021 Antelope Valley, California, earthquake. As noted by Wesnousky (2020 SRL), each of the first three occurred in areas of relatively high background seismicity and geodetic strain rate; the same is true for the Antelope Valley earthquake. The events sample different tectonic environments with distinct fault geometries, leading to unique rupture characteristics of each event. We explore kinematic slip models of these earthquakes based on observations of geodetic static offsets and seismic waveforms. The Magna and Antelope Valley earthquakes are both normal faulting events that did not approach Earth’s surface (centroid depths ~8-10 km). However, their fault dips differ substantially — 30 deg. and 50 deg., respectively, reflecting likely differing tectonic stress fields. The Monte Cristo Range event involves predominantly left-lateral strike slip on steeply dipping faults and normal slip on a small near-surface shallowly dipping fault; it is the only event of the four to have ruptured the surface. The Stanley event is the deepest of the four, with significant slip concentrated at depth ~8 to 16 km. It involves both strike slip and normal slip, with a gradual transition between the two in the along-strike direction, reflecting a spatially variable tectonic stress field. These events highlight the significant seismic hazards of the Intermountain West area and the importance of the geological record for understanding regional differences.

Mohamad Hallal, UC Berkeley

When attempting to model seismic site effects, most studies have focused on the importance of vertically characterizing subsurface geomaterials. However, laterally characterizing the geomaterials’ variability has generally been either overlooked or oversimplified. As such, little is actually known about the spatial area that influences seismic site response. This presentation will share insights that have been gained regarding the spatial area that influences site response from performing two-dimensional (2D) ground response analyses (GRAs) for the Treasure Island Downhole Array (TIDA) site. These numerical analyses utilize a site-specific, large-scale, three-dimensional (3D) shear wave velocity (V s) model that has been developed to a depth of 150 m over the entirety of Treasure Island (an area of approximately 1.6 km × 1.0 km). The lateral extent influencing site response and its azimuthal variability will be discussed by investigating: (1) four cross-sections with increasing lateral extent along the same azimuth, and (2) 12 cross-sections with different azimuths that uniformly sample the full 3D V s model. Although the TIDA site is often assumed to be archetype of one-dimensional (1D) site conditions, the findings shared in this presentation will support the notion that there may not be many truly 1D sites due to the large spatial area that influences site response. While some sites can be reasonably modeled using 1D GRAs, moving our profession forward requires focusing our efforts on developing site-specific subsurface models over larger areas to better understand and accurately model seismic site response.

Yongfei Wang, SCEC

Coseismic fault displacements in large earthquakes have caused significant damage to structures and lifelines located on or near fault lines. For buildings or distributed infrastructure systems located near active faults, engineering displacement demands are defined using probabilistic fault-displacement hazard analysis (PFDHA) models. However, fault displacement models (FDMs) used in PFDHA are sparse and poorly constrained in part due to the scarcity of direct observations. The physics-based dynamic rupture simulation method is an attractive alternative to address this important issue. Because fault displacements can be simulated for various geologic conditions as constrained by current knowledge of earthquake processes, they can be used alone or combined with empirical datasets to support FDM and thereafter PFDHA model development.

Simulations must first be validated against data, then the underlying physics can justify their extrapolation to other plausible events. This study summarizes our calibrated dynamic rupture models and their validation against displacement observations from empirical scaling relationships for strike-slip earthquakes from M5 to M8. This combination of calibration and validation of the model is critical in informing the functional forms used in the FDMs. We also perform a first-order validation of the near-fault ground motion to confirm that essential modeling physical factors important to ground motions are also properly addressed. This work is an essential first step in paving the way for dynamic rupture modeling to support PFDHA development.

Tristan Buckreis, UCLA

Next Generation Attenuation (NGA) West2 ground motion models (GMMs) include regional path adjustments for broad geopolitical regions. We extend that framework to account for systematic variations in anelastic attenuation for nine physiographical subregions in California that are defined in consideration of geological conditions. Using a large database that is approximately doubled in size for California relative to NGA-West2, we find relatively fast attenuation in coast range areas (North Coast, Bay Area, Central Coast), relatively slow attenuation in eastern California (Sierra Nevada, Eastern California Shear Zone), and state-average attenuation elsewhere, including southern California. As part of these analyses, we find for the North Coast region relatively weak ground motions on average from induced events (from the Geysers), similar attenuation rates for induced and tectonic events, and higher levels of ground motion dispersion than other portions of the state. The proposed subregional path model appreciably reduces within-event and single-station variability relative to the Boore et al. (2014) GMM model for ground motions at large distance (RJB > 100 km). The approach presented here can readily be adapted for other GMMs and regions.

Christina Morency, Lawrence Livermore National Laboratory

Classical approaches for Earth subsurface imaging rely predominantly on seismic techniques, which alone do not directly capture fluid-specific properties. On the other hand, electromagnetic (EM) measurements add constraints on the fluid phase through, for example, electrical conductivity. However, EM signals alone do not offer direct information of solid properties. In the recent years, there have been efforts to combine seismic and EM data for exploration geophysics. The most popular approach relies on joint inversion of decoupled seismic and EM data. However, by analyzing fully coupled poroelastic seismic and EM wave equations, one can capture a pore scale behavior known as seismoelectric effects (SEE) and more accurately resolve both solid and fluid properties.

I will present the equations used to model the seismoelectric response, which corresponds to electrokinetically couple Biot's poroelastic seismic and Maxwell's electromagnetic wave equations. To solve these equations, the spectral-element method (SEM) is used. The SEM, in contrast to finite-element methods (FEM) uses high degree Lagrange polynomials. Not only does this allow the technique to handle complex geometries similarly to FEM, but it also retains exponential convergence and accuracy due to the use of high degree polynomials. Finally, I will discuss the potential of the SEE technique for carbon storage sequestration and geothermal resources monitoring.

This work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.

Mark Benthien, Earthquake Country Alliance

The Earthquake Country Alliance has created a variety of earthquake safety messaging documents in the top 16 languages spoken and read in California, at www.EarthquakeCountry.org/languages. Each document is also accessible for people who use screen reader technology. Funding for the ongoing project is provided to the California Governor’s Office of Emergency Services by FEMA (NEHRP) and subgranted to the Southern California Earthquake Center which administers ECA. This presentation will provide an overview of the process, lessons learned, and next steps.

Jiuxun Yin, Caltech Seismological Laboratory

Seismograms contain multiple sources of seismic waves, from distinct transient signals such as earthquakes to continuous ambient seismic vibrations such as microseism. Ambient vibrations contaminate the earthquake signals, while the earthquake signals pollute the ambient noise's statistical properties necessary for ambient-noise seismology analysis. Separating ambient noise from earthquake signals would thus benefit multiple seismological analyses. This work develops a multi-task encoder-decoder network named WaveDecompNet to separate transient signals from ambient signals directly in the time domain for 3-component seismograms. We choose the active-volcanic Big Island in Hawai'i as a natural laboratory given its richness in transients (tectonic and volcanic earthquakes) and diffuse ambient noise (strong microseism). The approach takes a noisy 3-component seismogram as input and independently predicts the 3-component earthquake and noise waveforms. The model is trained on earthquake and noise waveforms from the STandford EArthquake Dataset (STEAD) and on the local noise of seismic station IU.POHA. We estimate the network's performance by using the Explained Variance (EV) metric on both earthquake and noise waveforms. We explore different neural network designs for WaveDecompNet and find that the model with Long-Short-Term-Memory (LSTM) performs best over other structures. Overall, we find that WaveDecompNet provides satisfactory performance down to a Signal-to-Noise-Ratio (SNR) of 0.1. The potential of the method is 1) to improve broadband SNR of transient (earthquake) waveforms and 2) to improve local ambient noise to monitor the Earth's structure using ambient noise signals. To test this, we apply a Short-Time-Average to a Long-Time-Average (STA/LTA) filter and improve the number of detected events. We also measure single-station cross-correlation functions of the recovered ambient noise and establish their improved coherence through time and over different frequency bands. We conclude that WaveDecompNet is a promising tool for a broad range of seismological research.