Earthquake Science Center Seminars: Recent Episodes
U.S. Geological Survey
Open dialogue about important issues in earthquake science presented by Center scientists, visitors, and invitees.
Chris Milliner, California Institute of Technology
Understanding the mechanics of faulting and accurately assessing seismic hazards are crucial for mitigating the impact of earthquakes. This seminar investigates the use of satellite pixel tracking data to unravel fundamental geomechanical properties of fault systems by providing novel estimates of friction, the absolute magnitude of stresses in the Earth's crust and evolution of inelastic strain as fault systems mature.
Here, I present a methodology based on Mohr-Coulomb failure theory with a quasi-static stress assumption to estimate fault friction parameters using satellite pixel tracking data. Applying this approach to the 2019 Ridgecrest earthquake sequence reveals statically strong and dynamically weakened friction, of 0.61 and 0.29, respectively, which are consistent with experimentally derived lab values. Additionally, I will argue that this approach can estimate the absolute magnitudes of stresses in the crust, shedding light on the driving forces behind earthquakes and their implications for future seismic activity.
In the second part of the seminar, I will present preliminary results of the friction analysis applied to the recent Mw 7.8 and Mw 7.6 Kahramanmaraş earthquakes in Türkiye. It is found that slip release along a specific fault segment deviates from the expected behavior under quasi-static stresses, suggesting an possible effect of strong dynamic stresses associated with rupture propagation.
Lastly, I will present estimates of surface deformation from radar and optical pixel tracking of 17 of the largest (6.0 ≤ Mw ≤ 7.9), historic, continental strike-slip surface ruptures to estimate the amount of inelastic coseismic off-fault strain. These results indicate progressively smaller amounts of off-fault inelastic strain with higher cumulative displacements, faster rupture velocities and geologic slip rates, supporting the notion that faults systems localize as they mature.
Karl Kappler, QuakeFinder
Short term (days to hours) earthquake forecasting is a hard scientific problem. Monitoring electromagnetic (EM) rather than mechanical/seismic activity may provide a breakthrough in this research. Anecdotal EM studies of single earthquakes lack of reproducible observations, and non-uniqueness of the anomalies when data are examined over the long term. However, in the early 2000s, a private humanitarian effort (QuakeFinder) was established to address the deficiency of data and to build out a dataset appropriate for statistical analysis. Three-component induction magnetometers were deployed at more than 100 stations running along the San Andreas and other major faults in California. More than 300,000 station days have been acquired with many stations nearby to significant earthquakes. The goal was to determine if there is a natural signal prior to moderate to large (>M4) earthquakes within range ( < 40 km) of the magnetometer instruments. Two major peer-reviewed statistical analyses (QuakeFinder and Google GAS) were done in 2019 and 2022. Both studies indicated rejection at the 99.7% confidence level of the null hypothesis (that there is not an electromagnetic precursor to earthquakes) at a baseline when crude noise compensation measures are applied. These statistical studies form the basis of further investigations, perhaps by USGS, to move the technology towards an operational, short term (days not seconds), earthquake warning system.
Rishav Mallick, California Institute of Technology
Viscoelastic processes in the upper mantle redistribute seismically generated stresses and modulate crustal deformation throughout the earthquake cycle. Geodetic observations of these motions at the surface of the crust-mantle system offer the possibility of constraining the rheology of the upper mantle. Parsimonious representations of viscoelastically modulated deformation through the aseismic phase of the earthquake cycle should simultaneously explain geodetic observations of (a) rapid postseismic deformation, (b) late in the earthquake cycle near-fault strain localization. To understand how rheological formulations affect kinematics, we solve a set of integral equations to simulate periodic earthquake cycles for a vertical strike-slip fault governed by rate-dependent friction in a homogeneous elastic crust underlain by a viscoelastic upper-mantle. We explore two popular viscoelastic rheological model classes in our simulations, linear Burgers and power-law rheologies; we show that these model classes are nearly kinematically indistinguishable with current geodetic observational techniques, though they may have different implications for stress transfer within the deforming medium. The degeneracy in predicted kinematics is broken when the system is perturbed by earthquake sequences i.e., aperiodic events with non-characteristic magnitudes. However, when the true rheology follows a power-law, estimating rheological parameters from earthquake sequence observations do not yield unique parameters. We discuss possible remedies and some issues related to non-steady state or transient rheologies. Distinguishing between linear, power-law and possibly transient rheologies and estimating associated parameters is a step towards building better regional seismic hazard models.
Itzhak Lior, The Hebrew University of Jerusalem
Distributed Acoustic Sensing (DAS) is revolutionizing observational seismology by allowing for seismic measurement every few meters along tens-of-kilometers long optical fibers. One application bearing immense scientific and societal implications is the use of DAS for Earthquake Early Warning (EEW). For optimal warning times, seismic sensors should be installed as close as possible to expected earthquake sources. However, while the most hazardous earthquakes on Earth occur underwater, most seismic stations are located on-land; precious seconds may go by before these earthquakes are detected. In this work we harness optical fiber infrastructure, ubiquitously deployed across the world both on-land and off-shore, for EEW. We devised methods for real-time magnitude estimation and ground motion prediction and validated them using earthquakes recorded in France, Greece and Chile. The results demonstrate the potential of DAS-based EEW and the significant time-gains that can be achieved compared to the use of standard sensors, in particular for offshore earthquakes.
Associated publication: https://www.nature.com/articles/s41598-023-27444-3
Te Yang Yeh, San Diego State University
We have simulated 0–3 Hz deterministic wave propagation in the Southern California Earthquake Center Community Velocity Model (CVM) version CVM‐S4.26‐M01 for the 2019 Mw
7.1 Ridgecrest earthquake. A data‐constrained high‐resolution fault zone model (Zhou et al., 2022) is incorporated into the CVM to investigate the effects of the near‐fault low‐velocity zone (LVZ) on the resulting ground motions, constrained by strong‐motion data recorded at 161 stations. The finite‐fault source used for the simulation of the Ridgecrest event was obtained from the Liu et al. (2019) kinematic inversion, enriched by noise following a von Karman correlation function above ∼1 Hz with a f−2 high‐frequency decay. Our results show that the heterogeneous near‐fault LVZ inherent to the fault zone structure significantly perturbs the predicted wave field in the near‐source region, in particular by more accurately generating Love waves at its boundaries. The fault zone decreases the 0.1–0.5 Hz mean absolute Fourier amplitude spectrum bias to seismic recordings for all sites in the model and in the Los Angeles basin area (∼200 km from the source) by 16% and 26%, respectively. The fault zone structure generally improves modeling of the long‐period features in the data and lengthens the coda‐wave trains, in better agreement with observations. We recommend that a data‐constrained fault zone velocity structure, where available, be included in ground‐motion modeling to obtain the least‐biased fit to observed seismic data.
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.
Joan Gomberg, USGS
Like many of its partners, the USGS has recognized the tremendous opportunity for advancing earthquake science that subduction zones (SZs) offer, because of 1) the extraordinary diversity of SZ earthquakes and other natural phenomena they initiate and respond to, 2) the new perspectives inspired by required multi-disciplinary and international work, and 3) technological advances enabling offshore exploration. Participation of USGS personnel in SZ-relevant partners’ programs helps to ensure that the USGS fulfills its mandate to assess earthquake hazards and provide products useful for reducing risk. This mandate distinguishes USGS’s SZ activities, but also complements the goals of its many partners. USGS SZ science has largely been motivated and guided by its scientists, but with growing engagement at higher authoritative levels to reprogram and request new SZ-targeted resources. Although not a formal USGS project or program, SZ issues figure prominently in those that are, such as Powell Center projects focused on Cascadia megathrust recurrence and another on tsunami sources, the ShakeAlert Earthquake Early Warning project, the USGS-USAID Disaster Assistance programs, and a Risk Research and Applications Community of Practice. We present a few examples of forefront USGS SZ research, conducted in partnership with numerous collaborators. These include field and modeling studies of paleo-earthquake shaking proxies such as ‘fragile geologic features’ and onshore and offshore landslides, implementation of new seafloor geodetic measurement capabilities for event to coupling studies, terrestrial-quality resolution of submarine subsurface and surficial imaging and coring to test paradigms about recent and paleo-earthquakes, and much more!
Harry Lisabeth, Lawrence Berkeley National Laboratory
The behavior of the upper crust is controlled in large part by fractures. Fractures are the conduits of fluid flow, facilitate reactions and transport of mass, and mediate deformation large and small. Fractures are not static features, but rather are sensitive to the hydraulic, chemical and mechanical environment in which they are set. I'll present the results of a multimodal experimental study of the physical properties of fluid-saturated, fractured rock in response to changes in fluid chemistry and stress. Complementary measurements were made of changes in the fracture's physical structure to investigate the origins of the modified physical properties. The role of stress is shown to dominate the role of fluid composition, particularly at low effective stress, where nonlinear behavior becomes prevalent. I'll discuss the results in the context of geological engineering efforts, induced seismicity and the potential for remotely mapping stress in the subsurface.
Ben Holtzman, Lamont-Doherty Earth Observatory
Geothermal heat mining has the potential to become a significant contributor to a clean energy transition. However, significant problems must be overcome to safely and effectively engineer fluid pathways through fracture networks. Multiple approaches to unsupervised machine learning in seismology have the potential to help us discover complex and subtle patterns in acoustic signals. Physics-constrained learning, though not discussed here in detail, has the potential to associate patterns in multiple data types in time, micro- and macroscopic. Our aim is to be able to associate changes in a reservoir's acoustic signals with changes in its thermal-mechanical state and fracture processes. Towards this, we are building sets of analyses on active reservoirs and laboratory experiments. I will show results from two sets of (unpublished, ongoing) studies- first on a local section in the northwestern corner of The Geysers geothermal field in Northern CA, and second on a set of deformation experiments on basalt.
Ethan F. Williams, Seismological Laboratory, California Institute of Technology
In recent decades, the volume of research seismographic data from broadband and nodal arrays has exponentially outpaced the growth of structural and geotechnical data. With this revolution, new frontiers in “large-N” and “large-T” seismology have emerged, from monitoring ground water in sedimentary basins to high-resolution microseismic event location. By contrast, without continuous recordings and dense networks, current engineering practice for seismic structural health monitoring and post-earthquake damage assessment remains comparatively primitive. In this talk, I will make the case for big data in engineering seismology, highlighting the multi-scale complexity of structural vibrations in time and space, as well as the value of new technologies like distributed acoustic sensing (DAS).
First, I will focus on a 20-year record from a single strong-motion station (“large-T”). Since 2001, the Southern California Seismic Network has archived continuous waveform data from CI.MIK in Caltech Hall (formerly Millikan Library), a nine-story reinforced concrete building in Pasadena, CA. Simple spectral analysis of ambient vibrations reveals that the building's fundamental frequencies have gradually increased by 5.1% (E-W) and 2.3% (N-S), with larger long-term variability up to 9.7% (E-W) and 4.4% (N-S). This finding is unexpected, as previous analysis of forced vibration tests and strong-motion records has shown that between 1968 and 2003 the fundamental frequencies decreased by 22% (E-W) and 12% (N-S), largely attributed to minor structural damage and soil-structure system changes from earthquakes. Today, the building's apparent stiffness is comparable to what it was in 1986, before the Whittier Narrows earthquake, indicating significant passive healing. Using data from earthquakes and forced vibrations, I also document the building's nonlinear dynamic elasticity, which is characterized by a rapid softening (decrease in apparent frequencies) at the onset of strong motion, followed by a slower, log-linear recovery trend over the scale of minutes. Importantly, nonlinearity persists down to the amplitude of ambient vibrations, and there is no linear elastic regime.
Second, I will present two examples of DAS arrays recording structural vibrations in the far-field (“large -N”). Through rocking, displacement, and shearing of the foundation, structural vibrations driven by cyclic loading efficiently generate seismic waves. (1) With DAS data from seafloor power cables in the Belgian North Sea, I show that Scholte waves generated by wind turbine structural vibrations dominate the high-frequency ambient seismic field between 1-5 Hz. Utilizing array processing, these vibrations can be localized to individual turbines, opening the door for remote operational and structural health monitoring of an entire offshore wind park with a single DAS system. (2) Returning to Caltech Hall, forced vibration tests are similarly effective at exciting Rayleigh waves at the building’s natural frequencies. The Pasadena Array, a 37.5-km fiber-optic loop instrumented with DAS since 2018, records the building-generated seismic waves at 10-m resolution throughout the city. This combination of a repeatable 1-10 Hz controlled source with an ultra-dense recording array permits high-resolution seismic microzonation on the scale of a city block and time-lapse near-surface velocity monitoring.
Sara McBride, USGS Earthquake Science Center
The collection of online videos and imagery to use in disaster reconnaissance is increasing in frequency, due to accessibility of platforms and the ubiquitous nature of smartphones and recording devices. This presentation explores the processes, goals, and utility of online footage and imagery of geohazards (earthquakes, volcanoes, tsunamis, and landslides) to better understand human behavior. Searching techniques and processes have grown increasingly sophisticated and organized. This talk focuses on three case studies: the 2018 M7.1 Anchorage, 2021 M7.2 Nippes, Haiti earthquakes, as well as the historic Hunga Tonga–Hunga Haʻapai volcanic eruption generated atmospheric shockwaves recorded around the globe, as well as a damaging Pacific-wide tsunami. These videos offer a significant source of data about physical and event-related human behavior, given that little is currently known about human reaction to earthquakes as well as volcanic eruptions followed by tsunami of these magnitudes, as well as the physical phenomena. For the Hunga Tonga–Hunga Haʻapai collected more than 400 videos for the eruption and this data set offers novel information about people's reactions to the eruption and attendant tsunami obtained from 11 different nations throughout the Pacific Ocean basin. This collection provides potential insights into human behavior across cultures and national boundaries related to tsunami impacts. Two findings were of interest: one is that when presented with multiple eruption-related hazards, people were more likely to do nothing than take self-protective action, indicating the need for multi-hazard drills. Further, surprisingly, in our dataset the presence of children seemingly reduced the likelihood that protective action was taken, particularly in relation to the tsunami. These findings can inform future education and outreach efforts to assist in strengthening standardized protective actions for the impacted regions.
Matt Herman, CSU Bakersfield
Since 2020, three large magnitude earthquakes have occurred in the vicinity of the Shumagin Islands in Alaska: the 22 July 2020 Mw 7.8 megathrust event, the 19 October 2020 Mw 7.6 intra-slab strike-slip event and the 29 July 2021 Mw 8.2 megathrust event. The 2020 events occurred in the transition from high plate interface coupling east of the Shumagin Islands to low coupling near the Shumagin Islands, and our finite element models of coupling demonstrate that the slip behavior and kinematics of these events reflect this coupling transition. The 2021 Mw 8.2 megathrust earthquake occurred further to the east, rupturing a large part of the 10 November 1938 Mw 8.2 rupture zone. Slip models and aftershock activity suggest that nearly all the slip deficit that could accumulate since 1938 was recovered in the mainshock. We find using our coupling models that an uncoupled Shumagin Gap is most compatible with observations from the July 2021 earthquake. Although the Shumagin Gap appears to be uncoupled based on these earthquake sequences, the coupling situation up-dip of the 2021 earthquake remains enigmatic. Using our modeling setup, we explore implications for scenarios with various degrees of coupling on the shallow plate interface.
Robert Goldman, University of Illinois at Urbana-Champaign
The 2018 eruption of Kīlauea Volcano was the first natural hazard crisis event in which USGS social media communications played a major role in providing reliable, around-the-clock eruption information to publics. These communications complemented traditional channels including HVO’s website, radio and television broadcasts, newspaper articles, automated text or email alerts, and in-person community meetings. As part of my National Science Foundation Graduate Research Internship, I conducted two complementary investigations of these communications. First, I analyzed semi-structured in-person interviews conducted with residents of Hawai‘i in January 2020 to identify themes describing residents’ preferred sources, messengers, and channels of 2018 eruption information. I then analyzed the public comment threads of the @USGSVolcanoes Facebook page to determine how well dialogues between @USGSVolcanoes scientists and non-USGS users answered users’ eruption-related questions, addressed occurrences of misinformation, and benefitted users’ overall emotional state.
From this work, my collaborators and I identified the role that messengers’ perceived trustworthiness and credibility played in Hawai‘i residents’ reliance on information delivered by those messengers. We also learned that among the most frequently viewed @USGSVolcanoes Facebook posts, 73% of users’ questions were directly answered while 54% of comments containing misinformation were directly corrected or called out. Moreover, most users’ comments on the @USGSVolcanoes Facebook page contained positive emotions that reflect appreciation of the information provided by @USGSVolcanoes scientists. I conclude this talk by discussing the implications of our findings for planning future hazard communications in Hawai‘i and elsewhere.
Dorian Golriz, Scripps Institution of Oceanography
In the first part of this talk, I will present a new method to determine the earthquake’s coseismic time window using a combination of seismic and GNSS data. Current practice is to use daily GNSS positioning for static offsets determination and finite slip models. When using high-rate (1-5 Hz) GNSS positioning for this purpose, it is a common practice to take a single time window across all the affected stations (typically a few minutes). Since the postseismic phase starts immediately after the earthquake, in a continuous manner, these static offsets include early postseismic deformation that occurs minutes to hours after the earthquake. Here I show the differences between these offset estimates, and how they can affect coseismic slip models. In the second part the talk, I will show how we can use this coseismic time window to estimate earthquake magnitude for tsunami warning purposes. The Pacific Tsunami Warning Center (PTWC) uses a variety of tools to issue a tsunami warning based on the size and location of the earthquake. However, current methods that rely on seismic data alone suffer from magnitude saturation or not timely enough for coastal communities located closest to the earthquake’s rupture. The combination of GNSS and strong-motion data using a Kalman filter yields both broadband velocity and displacement waveforms that do not clip and are sensitive to the entire spectrum of ground motions. We can use this combination to rapidly determine earthquake magnitude. Replaying seismogeodetic data for a number of tsunamigenic earthquakes around the Pacific basin, we show that useful and reliable estimates can be obtained before the end of rupture. Additionally, our method does not rely on empirical relationships derived from historical earthquakes, making it suitable for local tsunami warning systems and shows promise for earthquake early warning.
Verónica Rodríguez Tribaldos, Lawrence Berkeley National Laboratory
In recent years, there has been growing interest in the seismological community for regional characterization and monitoring of subsurface processes at high temporal and spatial resolution, for applications as varied as fault identification, near-surface characterization or hydrological process monitoring. However, it has become evident that the utilization of permanent local or regional seismic networks for these purposes is limited, mostly due to sensor sparsity. The cost of running large-N, long-term temporary experiments is almost prohibitive. Distributed Acoustic Sensing (DAS) re-purposes telecommunication optical fibers as dense arrays of seismic sensors. This developing technology enables recording ground motions for long periods of time across long distances (10’s of km) at high spatial (~1 m) and temporal resolution at frequencies ranging from the mHz to the kHz. Recently, the deployment of this novel sensing technique on existing, unused fiber-optic cable networks, known as dark fiber, has offered an attractive alternative to classical seismological studies, as it facilitates acquisition of high-resolution data at regional scale. However, some challenges still exist before we can fully exploit the capabilities of this technology, the main challenge being the large data volumes derived from these several km-long, high density arrays. In this talk, we will show some examples of using dark fiber DAS for basin-scale characterization and monitoring using ambient seismic noise at a variety of scales in the Sacramento Valley, CA, and present preliminary observations from an ongoing experiment aiming at characterizing geothermal systems in the Imperial Valley, CA. We will also discuss some of the challenges associated with regional, high-resolution characterization using this novel technology and our recent attempts at developing effective data management and processing strategies.
Will Steinhardt, UC Santa Cruz Seismo Lab
Many systems in geophysics, including faults, ice sheets, and hill slopes, are predominantly stable, but become unstable catastrophically, with severe societal consequences when they do. However, the behaviors of these systems are often difficult to predict because they involve extreme spatial and temporal scales, accumulating stresses over decades or centuries, but nucleating failure processes at the micron-scale in fractions of a second leading to kilometers of deformation. In this talk, I will discuss how I utilize applied physics techniques to build scaled-down experiments to explore these complex problems in systems where a wide range of system properties can be tuned to make otherwise impossible observations. I will present two examples: First, using a scaled, transparent laboratory fault where slip at the interface can be directly imaged, I will show that slow slip events in our system follow earthquake-like scaling, and demonstrate how finite fault effects alter stress drop. Second, I will discuss how material heterogeneity leads to brittle fracture roughness, and show that the resultant morphology of a crack is, surprisingly, not dependent upon the details of the medium, but is controlled entirely by a single parameter: the probability to perturb the fracture front above a critical size to produce a step-like instability.
Elnaz Seylabi, University of Nevada, Reno
We perform a series of large scale, nonlinear, earthquake ground motion simulations that account for the cyclic plastic behavior of sediments in the shallow crust. Our goal is to understand and quantitatively assess how idealized models of sediment nonlinearity influence the amplitude, frequency content, and duration of strong ground motion in broadband earthquake simulations. We use the Garner Valley region in southern California as a test case where near-surface nonlinearity has been reported for peak ground accelerations (PGAs) as small as 0.05-0.2g. We model the sediment cyclic response using a multi-axial constitutive model formulated within the framework of bounding surface plasticity in terms of total stress and implemented in a high-performance computing finite element code. We first describe a series of numerical experiments designed to verify our model implementation, and then present a series of idealized large-scale simulations where material properties were extracted from the Southern California Earthquake Center (SCEC) Community Velocity Model CVM-S4.26 (using its optional geotechnical layer). The modulus reduction curves and ultimate shear strength were selected empirically to constrain the nonlinear soil model parameters. Furthermore, we simulate the rupture of the 2010 Mw 5.4 Borrego Springs and a Mw 6.5 scenario earthquakes using a kinematic earthquake rupture model. Having the rupture simulations, we then compute synthetic ground motions with the nonlinear model (for a maximum frequency of 5 Hz) and discuss how modeling shallow crust nonlinearity affects the ground response intensity measures in the different cases considered.
Gabe Paris, University of Alaska Fairbanks AEC
Forecasting how many earthquakes will occur following a potentially damaging earthquake helps the public and emergency operators stay safe and make informed decisions. The USGS issues aftershock forecasts following potentially damaging mainshocks in the U.S. states and territories and updates these forecasts 75 times during the first year. Most of the forecasts are issued automatically, but some forecasts require manual intervention to maintain accuracy. To help identify the sequences whose forecasts will benefit from a modified approach, we built the oaftools R package. Oaftools includes functions that analyze and plot earthquake sequences and their forecasts, and the Operational Aftershock Forecast (OAF) Viewer, which incorporates the functions into an interactive web environment that can be used to explore aftershock sequences. The OAF Viewer displays two maps, one of the mainshocks and one of a sequence's aftershocks, and five analytical plots. The OAF Viewer will help seismologists understand complexities in the data, communicate with the public and emergency managers, and improve the OAF system by maintaining operational awareness. It will also help researchers study aftershock forecasting using their own methods. I used the OAF Viewer to study aftershock forecasting for sequences along the forearc of the Alaskan subduction zone. The USGS considers the forecast to be "successful" when the number of earthquakes observed within the forecasted duration period is within the 95% confidence interval. In the Alaskan forearc, the observed number of earthquakes consistently lie within this broad range of "success," however the forecasts systematically over-predict the number of aftershocks. I analyzed seventeen earthquake sequences in the Alaska forearc region and have developed a set of parameters that may improve early-sequence aftershock forecasting in the region. I stacked the 17 sequences and used maximum likelihood estimation to determine the new parameters and their uncertainty. Following the lessons of Page et al. (BSSA, 2016), I combined the inter-sequence variability and the uncertainty of the parameters from the stacked sequences to produce the total standard deviation for a new forecast model.
Max Schneider, USGS ESC
Public communication for a region’s seismic hazard and aftershock forecasts is often done using maps. But the public products released by scientists are not always built with best practices in visualization or cartography, nor are they empirically validated with rigorous user experiments. In this talk, I will present several projects that investigate how different visualization approaches can affect public understanding of seismic hazard and aftershock risk. First, I consider how seismic hazard can be mapped to comply with research-backed best practices in color selection, legend design and classification of the continuous hazard distribution onto a discrete color map. I apply these best practices to redesign the German seismic hazard map and evaluate it against the original map in a controlled user experiment. The redesigned map improves perception of key principles of hazard, including that it is not spatially concentrated only around areas with previous earthquakes. Next, I evaluate the effectiveness of three techniques for displaying the uncertainty in aftershock forecast maps. In an experiment, participants perform a comparative judgment task, which measures whether a visualization can succeed in reaching two key communication goals: indicating where an aftershock is either highly likely or highly unlikely (“sure bets”) and where the forecast is low but the uncertainty is high enough to imply potential risk (“potential surprises”). All visualizations perform equally well in the goal of communicating “sure bet’’ situations. But the visualization that shows uncertainty using lower and upper bounds significantly outperforms the others at communicating “potential surprises.” I conclude with an outlook on my postdoctoral research on creating and testing visual products for aftershock forecasting at the USGS.
Marius Isken and Sebastian Heimann, GFZ Potsdam and University of Potsdam, Germany
Pyrocko (https://pyrocko.org) is an open source seismology toolbox and library, written in the Python programming language. It can be utilized flexibly for a variety of geophysical tasks, like seismological data processing and analysis, modelling of InSAR, GNSS data and dynamic wave forms, or for seismic source characterization. The presentation introduces the framework and focuses on a flexible and computationally efficient f-k filtering technique for DAS data, which makes use of its dense spatial and temporal sampling, and can handle the large amount of data. The presented adaptive frequency-wavenumber filter suppresses the incoherent seismic noise while amplifying the coherent wave field.
Bio: Marius Paul Isken is a Geophysicist interested in observational seismology and earthquake source characterisation. He studied Geosciences (BSc) and Geophysics (MSc) at the University of Kiel and afterwards worked as a research seismologist at the USGS in Menlo Park and KAUST, Saudi Arabia. Before starting his PhD about distributed acoustic sensing (DAS) for earthquake modelling at the GFZ Potsdam with the section Physics of Earthquakes and Volcanoes in 2020, he worked as a software developer at the University of Kiel (2016 - 2020). In 2018 he co-founded QuakeSaver GmbH, pursuing the aim to develop smart seismometers and seismic networks.
Sabine Loos, Mendenhall Fellow, USGS, NHC
Many data sources that become available after a disaster—such as satellite imagery or reconnaissance surveys—have immense power to inform decisions that influence the trajectory of the affected country for years afterwards. However, often these data sources prioritize specific user groups, easy-to-measure metrics, and short-term understanding. In this talk, I will present on three main examples of designing earthquake information to be more actionable by centering user needs and more equitable by prioritizing vulnerable populations. The first is on developing actionable damage data, in which we identify how different sources of damage data can be applied to post-earthquake decisions to then inform a method to integrate multiple sources of building damage data to be more useful and accurate. The second is on developing post-earthquake data that acknowledges inequities in recovery, a relatively more complex phenomena than damage, looking at how we can integrate data on natural, social, and physical factors to estimate the spatial distribution of populations who will lag during recovery. The third is on current research to improve USGS’s near-real-time earthquake products to be more actionable and equitable, by taking a user-centered approach which combines focus groups with key stakeholders and web analytics to inform updates to current contend and the design of future content. Broadly, evaluating the what, why, and who of earthquake information supports the thoughtful design of future products that are more useful to local planners, reflect the multiple disciplines that study disaster, and, ultimately, inform decisions that lead to more effective and equitable outcomes.
Prof. Estéfan Garcia, Dept. of Civil and Environmental Engineering, Univ. of Michigan
Earthquake surface fault rupture has historically caused significant damage to structures, highways, lifelines, and other critical infrastructure. Field case histories from several large surface fault-rupturing earthquakes have provided valuable insight into the impacts of this hazard on the built environment, but their relative infrequency requires that mechanisms of these hazards be elucidated through supplementary means such as sandbox modeling. In the presented research, numerical models are developed using the discrete element method (DEM) to serve as “virtual experiments” that overcome the physical limitations of traditional sandbox modeling. The advantages of DEM in terms of accurately representing fundamentally discontinuous media and seamlessly capturing large-strain behavior are discussed. The results of DEM simulations of reverse and normal fault rupture are shown to be consistent with past physical studies in terms of the path of fault rupture propagation from bedrock to the ground surface and the manifestation of the resulting fault scarp. Pioneering work on the application of DEM with high-performance computing to surface fault rupture interaction with shallow foundations and in complex site conditions is presented. Finally, guidelines are provided for applying DEM in future studies of surface fault rupture, and potential application to larger scales of regional tectonic deformations over several kilometers are discussed.
Fred Pollitz, USGS ESC Moffett Field
The 15 January, 2022 eruption of Hunga Tonga Hunga Ha’apia is the largest since the 1991 Mount Pinatubo eruption. It produced a local tsunami with up to 15 m height and inundation of 500m; a plume that eventually reached 58 km height; globally recorded infrasound waves through Earth’s atmosphere; acoustic-gravity standing waves at two dominant resonant frequencies; worldwide sea waves driven in part by the atmospheric Lamb pulse. It produced globally observed seismic signals from coupling of the different atmospheric waves with the solid earth, as well as direct signals from the volcano due to the reaction force. We explore different seismological approaches to deriving the source time function of the reaction force, which is well characterized as a sequence of Impulsive vertical forcing that produced seismic wave energy in multiple packets for 5000 s after the initial Surtseyan eruption, with a late burst around 15000 s. The seismological results are consistent with the generation of the eruptive plume that expanded rapidly for the first 90 minutes, implying average forces of 10^12 N over this time but reaching magnitudes as high as 2x10^13 during the eruption subevents.
Kun Wang, Geophysics Group & Center for Nonlinear Studies, LANL
The previous 5 years demonstrate that machine learning is a powerful tool for studying laboratory faults using acoustic emissions (AE) that broadcast from the sliding interface. The AEs can inform us of the instantaneous fault frictional state in shear experiments and also contain information regarding the time remaining to the upcoming laboratory earthquake
(frictional failure). We present a deep learning approach using transfer learning for predicting the instantaneous status of the laboratory fault, with a unique emphasis on learning from extremely limited data. Although large training data sets are available from numerical simulations and
laboratory experiments, in Earth earthquake interevent times range from 10’s-100’s of years and continuous geophysical data typically exist for only a portion of an earthquake cycle. If the goal is predicting slip on seismogenic faults in Earth, then sparse data presents a serious challenge for applying machine learning. Hence, we developed a prototype transfer learning approach using fault-slip numerical simulations and laboratory data to address the sparse data situation. A convolutional encoder-decoder (CED) model learns a mapping between AE histories and fault friction from rich numerical simulations data. The model’s latent space is further trained using data from only a few laboratory-event earthquake-cycles, and even a small portion of pre- or post- failure data as an analog to a fault in Earth where data is sparse. Model predictions markedly improve by further training the model latent space and elucidate the potential of using machine learning models trained on numerical simulations and fine-tuned with small geophysical data sets for potential applications to faults in Earth. Furthermore, we address the open question
on whether the AE from laboratory experiments contains near-future frictional information. The approach applies a CED model and a self-attention transformer model that uses AE histories to predict the upcoming frictional behavior. Notably, information for predicting near future frictional failure and recovery are found to be contained in the AE signal, but when looking farther into the future the predictions are progressively worse. This first effort predicting future fault frictional behavior with machine learning will guide efforts for applications in Earth.
Alba Rodríguez Padilla, University of California, Davis
The past vicennial has witnessed four multi-fault, surface rupturing earthquakes in the Eastern California Shear Zone that left behind impressive footprints of widespread deformation in the desert. Inelastic processes from earthquakes contribute to the formation of fault damage zones that constitute a permanent sink of strain energy, modify the elastic properties of the shallow crust and amplify near-field ground shaking. Constraints on the extent of inelastic deformation differ depending on the dataset and methodology used. We combine fracture, strain, and aftershock maps from the Ridgecrest events, and fracture maps from the Landers, Hector Mine, and El Mayor Cucapah earthquakes to reconcile the properties of damage zones across different spatial scales and resolutions. Our observations reveal how macroscopic fracturing generates intense near-fault damage and that widespread damage accrues regionally over multiple earthquake cycles. To the east, the San Andreas and the San Jacinto faults come together in a releasing step-over at Cajon Pass, north of Los Angeles, where the record of multi-fault earthquakes through this junction has long been erased from the surface. Through paleoseismic trenching and finite-element modeling of secondary fault slip, we provide constraints on the frequency and mechanics of past multi-fault earthquakes at this location. Comparison of our record to independent chronologies shows that 20%–23% of earthquakes on the San Andreas and the San Jacinto faults are co-ruptures through Cajon Pass.
Kathryn Materna, Morgan Page, Lisa Schleicher, USGS ESC
We will have a selection of three talks from speakers at SSA:
Kathryn Materna - Determining strain rates and their epistemic uncertainties: Application to Southern California
Morgan Page - Finding the Next Layer of Seismicity Patterns in High-resolution Catalogs
Lisa Schleicher - Identifying Strong-motion Instrument Metadata Inconsistencies Before and After Earthquakes
Dara Goldberg, USGS Geologic Hazards Science Center
Following a significant earthquake, the USGS National Earthquake Information Center (NEIC) publishes a spatiotemporal estimate of the earthquake’s slip pattern, known as a kinematic finite fault model. These models are critical to informing downstream response products such as ShakeMap ground motion estimates and PAGER loss estimates. Because large earthquakes can involve slip over tens to hundreds of kilometers along a fault, it is vital to rapidly assess the amount and location of slip along the fault. Since its introduction to NEIC’s rapid response capabilities in late 2007, the finite fault product has been computed in the first several hours after a significant earthquake using only teleseismic data, for which it is generally possible to obtain a reliable model for earthquakes of magnitude 7 and larger. In this talk, I will highlight some recent and anticipated updates to the NEIC’s rapid finite fault modeling capabilities to include regional-distance seismic and geodetic observations. These new features allow rapid modeling of earthquakes as small as magnitude 6, and reduce the overall time required to produce a reliable model of slip.
Wenbo Wu, Woods Hole Oceanographic Institution
As the major buffer of Earth’s energy imbalance, the ocean plays a key role in regulating global climate and temperature changes. However, accurate estimation of global ocean temperature change remains a challenging sampling problem. To complement existing point measurements, we have developed a novel and low-cost method of using travel time changes of acoustic waves from repeating natural earthquakes to infer basin-scale average ocean temperature changes. Using the land-based seismometers and CTBTO hydrophones, we have detected not only seasonal signals, which are generally consistent with that in previous oceanographic datasets of ECCO and Argo, but also interesting signals missing in ECCO and Argo. Recently, we are working on the frequency and mode dependent travel time changes of these acoustic signals, which could reveal depth information of ocean property changes. I will also talk about the challenges of the multiple frequency measuring and mode tomography and the opportunities of Distributed Acoustic Sensing in T-wave study.
Shanna Chu, USGS ESC
The faults on which earthquakes occur sometimes form complex interconnected patterns. The level of this complexity may increase high-frequency ground motions from earthquakes occurring on such faults. In this talk, I describe ways of quantifying the complexity of groups of faults based on how they are aligned and how densely they are spaced. I found that high-frequency ground motions in Southern California tend to correlate with misaligned faults, suggesting that structural interactions between different parts of the fault system may play a role in generating the ground motions felt during earthquakes. I introduce a physical model to explain one possibility of how fault interactions could generate high-frequency ground motion and the settings where such interactions could be expected to occur. Finally, I explore some of the ways in which data-driven initiatives like the Ridgecrest stress drop validation project can help us gain understanding of the relation between fault complexity and high-frequency radiation.
Janet Watt, USGS Pacific Coastal and Marine Science Center
Distinguishing between seismic and aseismic fault slip in the geologic record is difficult, yet fundamental to estimating the seismic potential of faults and the likelihood of multi-fault ruptures. We integrated chirp sub-bottom imaging with targeted cross-fault coring and core analyses of sedimentary proxy data to characterize vertical deformation and slip behavior within an extensional fault bend along the Hayward-Rodgers Creek fault system in northern San Pablo Bay. We identified and traced four key seismic horizons (R1–R4), all younger than approximately 1400 CE, that cross the fault and extend throughout the basin. A stratigraphic age model was developed using detailed down-core radiocarbon and radioisotope dating combined with measurements of anthropogenic metal concentrations. The onset of hydraulic mining within the Sierra Nevada in 1852 CE left a clear geochemical and magnetic signature within core samples. This key time horizon was used to calculate a local reservoir correction and reduce uncertainty in radiocarbon age calibration and models. Vertical fault offset of strata younger than the most recent surface-rupturing earthquake on the Hayward fault in 1868 CE suggest near-surface vertical creep is occurring along the fault in northern San Pablo Bay at a rate of approximately 0.4 mm/yr. In addition, we present evidence of at least one and possibly two coseismic events associated with growth strata above horizons R1 and R2, with median event ages estimated to be 1400 CE and 1800 CE, respectively. The timing of both these events overlaps with paleoseismic events on adjacent fault sections, suggesting the possibility of multi-fault rupture.
Daniel Faulkner, University of Liverpool
Estimating the size and frequency of earthquakes during subsurface fluid injection is currently not possible. Understanding the fundamental controls on earthquake-size distributions is the first step to making these predictions. This study shows that the laboratory measured frictional stability parameter (a-b) relates closely to fault-related induced earthquake-size distributions (b-values) in a shale gas play. Fluid injection promoted seismicity along faults that cut lithologically distinct horizons, each with characteristic frictional stability parameters and earthquake-size distributions, are compared. Results indicate material properties play a key role in earthquake-size distributions. As such, it may be possible to constrain the expected seismicity from laboratory measurements prior to any fluid injection into the subsurface.
Joshua Crozier, USGS
Wavelet based seismic event detection yields a catalog of thousands of very-long-period seismic signals over the over the 2008-2018 eruption of Kīlauea Volcano, HI, USA. These represent impulsively triggered magma oscillations within the shallow conduit and lava lake, and thus provide a direct probe of the magma system. We invert for these events, along with geodetic and lava lake data, using a petrologically informed model of magma dynamics. The inversions reveal significant variation in magma temperature and highly disequilibirum volatile contents over days to years, and suggest an evolving magma system geometry. Such changes can modulate eruption style and hazards, making in situ inference of their temporal evolution vital for volcano monitoring.
Laura Pinzon-Rincon, Université Grenoble Alpes
Seismologists eagerly seek new and preferably low-cost ways to map and monitor the complex structure of the top few kilometers of the crust. Passive seismic imaging appears as a novel, low-cost, and environmentally-friendly approach for exploring the sub-surface. Usually, passive seismic interferometry relies on blind correlations within long time series of seismic noise or coda waves. Here, we propose a complementary approach: seismic interferometry using opportune sources, specifically moving sources that are not stationary in time. This new approach relies on an accurate understanding of the seismic source's mechanism, a careful signal time-window and station pairs selection, and seismic phase identification. For example, massive freight trains were only recently recognized as a persistent, powerful cultural (human activity-caused) seismic source. Thus, these train signals can be considered an opportune seismic source for passive seismic interferometry because they are readily available, detectable, repeatable, and generate high-frequency broadband energy. To illustrate this novel method's potential, we show a case study in a mineral exploration context at the Marathon site, Ontario, Canada, where we deployed a dense nodal array of 1020 sensors. We retrieve high-frequency energy using train signals only, and we use these arrivals to generate a 3D shear-wave velocity model. We discuss the pros and cons of the method compared to more standard approaches with the help of numerical modeling. We showed that by correlating train tremors, we retrieved high-frequency arrivals with higher quality than using the standard method while using fewer data. Far from being restrained to near-surface imaging, this new way of analyzing opportune seismic sources can be applied in various contexts and scales using natural or man-generated signals.
Eva Kostyu, American Geophysical Union
This presentation will introduce the audience to the AGU Bridge Program and the program's guiding principles, origin, and impact. The AGU Bridge Program is part of the multi-disciplinary coalition, IGEN (Inclusive Graduate Education Network), which aims to increase representation of historically marginalized populations in graduate STEM education through the promotion of equitable practices and student supports. The AGU Bridge Program has a dual focus working with both graduate geoscience departments and individuals interested in pursuing an advanced degree. Forty-six departments across the US partner with AGU in this effort, with new department cohorts selected each year through a competitive application process. Students who apply through the shared IGEN application and accept an offer of admission from a partner department enter the program as Bridge Fellows with a cohort of peers and access to AGU resources. With this two-part focus on departmental adaptations and direct student support, the AGU Bridge Program is pushing for graduate education practices rooted in equity resulting in improved educational experiences and greater diversity in the geosciences.
What can hematite textures and (U-Th)/He thermochronometry tell us about fault mechanics in the shallow crust?
Earthquake Swarms: Volcanic and Tectonic Lessons from the Earth's Crust, and the Mysteries that Remain
Asperity interactions in laboratory earthquake sequences illuminate delayed earthquake triggering