Earthquake Science Center Seminars

Distributed acoustic sensing (DAS) has been delivering us records of very high spatial and

temporal resolution. During the summer of 2017, about 22 days of passive recorded data were

acquired using a vertical downhole DAS array in the top 800 m of the San Andreas Fault

Observatory at Depth (SAFOD) main hole. In this talk, I will show two seismological applications

that take advantage of the revolutionary DAS data quality in a downhole array. First, certain

earthquakes may be used for P and S wave velocity model estimation, yielding results

comparable to (or better than) conventional downhole geophone surveys. This analysis may

unveil previously unknown geological structures. Ambient field records can also be used to

retrieve the P-wave velocity structure with reasonable quality. Then, I'll show how a seismic

velocity-based detection method can be applied for automatic earthquake detection. Despite

the high performance of the USGS network in the area, we discovered and analyzed a

previously uncataloged earthquake. This method is picking-free, easy to implement, and yields

some directional information during the detection process.

Tsunami earthquakes are highly destructive. The runup produced by their associated tsunami is much higher than expected for most megathrust events of the same magnitude. These rare events have comparatively modest magnitudes but rupture the shallowest portion of a subduction zone megathrust with exceptionally large seafloor displacements. Reliable and unsaturated estimates of the moment magnitude of these events are not generally available before their tsunamis inundate. Furthermore, the event magnitude alone is not an indication of the runup that these tsunami earthquakes produce. Thus it remains challenging to discriminate these events as tsunami earthquakes in time to issue local tsunami warning. Previous teleseismic observations found that tsunami earthquakes radiate seismic waves weakly, and should therefore not be strongly felt in the near field. To date no near‐source seismic recordings of these events exist that confirm this. I will present on an analysis of near‐field records of a dataset of subduction zone earthquakes, including the 2010 M7.8 Mentawai, Indonesia tsunami earthquake. The records of the Mentawai event show remarkably weak shaking, in comparison to both the other events as well as to estimates from ground-motion models. This is strong evidence that this earthquake does indeed have a weakly radiating or inefficient source process, in spite of its large slip. These findings imply that because other tsunami earthquakes are inferred to have similar source properties, our methodology should be applicable to other events.

Oceanic transform faults (OTFs) represent abundant, yet relatively unexplored pathways for fluid flow in the mantle lithosphere. Current models limit seawater circulation to 600 °C, the thermal limit of earthquakes recorded by teleseismic surveys. However, recent geologic and geophysical studies indicate that seismicity extends beyond 600 °C, creating a pathway for deep seawater circulation.

Peridotite mylonites dredged from OTFs on the Southwest Indian Ridge show evidence of brittle deformation and fluid flow concurrent with ductile deformation of olivine. Fluid-rock interactions resulted in fault weakening, grain size reduction, and growth of syn-deformational amphibole. The high chlorine content of amphibole and its association with brittle features suggest that seawater was the fluid source. Geothermometry and thermodynamic analyses of the mylonite mineral assemblage indicate that fluid flow and brittle-ductile deformation occurred at temperatures ~600-900 °C (~14-22 km depth). We interpret the shallow limit of the brittle-ductile transition zone as a fine-grained, hydrated shear zone, and the deeper limit as the transition to distributed, dry deformation of coarse-grained peridotite.

Our results are supported by ocean-bottom seismic studies of OTFs on the East Pacific Rise that have located earthquakes at depths corresponding to temperatures >1000 °C. The presence of earthquakes at the conditions of mylonite hydration and brittle-ductile deformation suggests that deep seawater circulation is controlled by the seismic cycle. We propose that feedbacks between seismicity, fluid flow, and mantle rheology determine how plate motion is accommodated on OTFs.

The detection of small, spatiotemporal changes in seismic velocity structure has shown promise in recent years as a tool for monitoring volcanoes. Much of this promise is due to the repeated observation that seismic velocity decreases prior to volcanic eruptions. These decreases are thought to occur in response to the opening and closing of cracks and are considered a proxy for volumetric strain. However, it is unclear which specific underlying processes create velocity changes. Kīlauea Volcano, Hawai'i, and its dense, modern geophysical monitoring network, is an ideal natural laboratory to test hypotheses about the genesis of seismic velocity changes. Using coda-wave interferometry of repeating earthquakes during the last two months of the summit eruption in 2018, we measured velocity changes with sub-hourly temporal resolution and compare these results with other timeseries of seismicity and deformation.

The Kīlauea summit caldera collapsed in a series of events with near-daily periodicity spanning mid-May through early August. Seismic velocity also showed a pattern of near-daily cyclicity in phase with seismicity levels and deformation, where velocity increased at the time of each collapse and slowly decreased between them. Multiple processes provide possible explanations for this cycle, including pressurization of the shallow magmatic system and stress-induced dilatation. Long-term trends in the seismic velocity timeseries can be divided into two distinct periods: a linear decrease in velocity from early to late June, followed by overall constant velocity in July through the end of the eruption in early August. We propose the first phase to be related to crack development and opening during the formation of the ring fault outlining the roof block above the shallow magmatic system, and the second phase to the down-dropping of the roof along that established margin.

Yellowstone hotspot volcanism started near the Oregon-Nevada border. But before propagating ENE, consistent with North America's plate motion, it erupted the Columbia River Basalts (CRB) from centers in eastern Oregon. There are several strange things about these eruptions: (1) they propagated rapidly north toward Washington, (2) volcanic vigor increased as the eruptions propagated "off track" and (3) an odd topographic bullseye, centered on the Wallowa Mts, uplifted in NE Oregon. Upper mantle tomography images a slab-like high-velocity structure hanging 100s of km beneath the bullseye. The story I'll present is this: The imaged high-velocity structure is Farallon lithosphere that was left flat beneath Oregon ~50 Ma, and then delaminated during the CRB event with a S-to-N rollback. Delamination was enabled by Yellowstone magma ponding between the Farallon slab and North America, and the northward rollback drew Yellowstone plume north with it. North America extension over this hot asthenosphere caused the increase in volcanic production. Delamination rollback passed beneath the large, pre-existing Wallowa batholith in NE Oregon, allowing its dense root to founder. This unloading caused uplift of the Wallowa Mts. Other supporting "details" will be presented during my presentation.

The current version of the USGS Bay Area Velocity Model is a 3D representation of the seismic velocity structure in the San Francisco Bay region and has been shown to successfully reproduce many aspects of ground motion and wave propagation from local earthquakes. Still, some users have noted that there are areas in the model that could be improved and should be investigated. Here, we present an exercise to evaluate and improve the model in a subregion centered around the East Bay. Special attention is paid to the area immediately east of the Hayward Fault in the Berkeley Hills. We use a finite-difference code (SW4) to simulate wave propagation from eight Mw3.5-4.2 earthquakes with epicenters in the East Bay. The synthetic seismograms obtained are compared with observed motions from each respective event. Similarity between synthetic and observed motions is judged by measures of arrival time difference and energy mismatch. First, some questionable features in the current version of the model are addressed and modified. Then, several variations of the model are generated by adjusting velocities assigned to particular geologic units. The evaluation procedure indicates which variation best represents the true velocity structure by the chosen metrics. The final adjusted model significantly improves certain aspects of the synthetic waveforms in comparison with the current version of the velocity model.

An accelerating cascade of ruptures in the days leading up the 2008 Mw4.9 Mogul, NV earthquake highlight complex, but organized, spatiotemporal patterns of seismicity and stress drop on the mainshock fault. Previous work using empirical Green's function (EGF) analysis indicated nearly two orders of magnitude of variation of stress drops over a small spatial area on the fault plane. The EGF-derived rupture areas encircle distinct seismicity voids, previously interpreted as "patches" of heterogeneous stress on a doubly-kinked fault plane. High stress drop foreshocks nucleate at the edges of these "patches" or near changes in the strike and/or dip of the fault, while low stress drop foreshocks exist primarily outside the main seismicity void and no aftershocks re-rupture it. New 3D rupture directivity analyses of 87 of the earthquakes, 36 occurring before the mainshock, further illustrate the propagation of events around the "patches" leading up to the mainshock. We measure azimuthal variation in duration of the EGF-derived source time functions for events as low as ML2.4 using a stretching- and correlation-based method. Then we perform a grid search to find the orientation and rupture velocity of the best-fitting line source. Combining robust directivity results with well-resolved seismicity locations and focal mechanisms enables a four-dimensional analysis of event triggering and interaction with respect to source physics. Our detailed analysis suggests that fault plane complexity, both geometry and proposed stress heterogeneity along the fault, plays a critical role in controlling the locations and migrations of seismicity, the stress drops of earthquakes, and the rupture direction of events within this energetic, shallow foreshock-mainshock-aftershock sequence. Of note, the Mw4.9 earthquake produced strong ground motions in excess of 1g, was preceded by over 60 days of foreshock activity, and was followed by aseismic slip equivalent to the moment release of the mainshock itself.

Liquefaction during earthquakes has been the subject of extensive study for over half a century and is now routinely addressed in engineering practice using a wide range of technical approaches that depend on the project size and importance. These past studies have produced major advances in our scientific understanding of liquefaction phenomena and the engineering practices used to address liquefaction hazards, but there remain numerous situations where knowledge gaps and engineering practice limitations hinder the efficient mitigation of earthquake-induced liquefaction damages to our infrastructure and communities. This presentation examines a number of lessons, challenges, and opportunities regarding the evaluation and mitigation of liquefaction hazards, including aspects of site characterization, engineering analysis methods, challenging soil types, remediation methods, performance-based engineering procedures, and risk management approaches.

Granular materials like sands, muds, soils, pebbles, boulders, and ground up rock are central to a wide range of earth science processes, and such natural shear flows are composed of particles exhibiting a range of irregular shapes, sizes, and material properties. For natural hazards involving granular media, like landslides and earthquakes, it is important to understand how the complexity of natural grains affects the transition between nearly static slow moving shear flow and fast moving shear flow. Laboratory experiments using naturalistic granular materials (e.g. angular sand instead of spherical beads) explore the effect of grain shape and material properties on the boundaries and transitions between these rheological regimes, looking to answer an enduring physics question of how energy entering a granular system is dissipated.