Earthquake Science Center Seminars

An exposure of a creeping segment of the Bartlett Springs Fault (BSF), part of the San Andreas System in northern California, consists of a ~1.5 m-wide zone of serpentinite-bearing fault gouge sandwiched between late Pleistocene fluvial deposits. The fault gouge is a heterogeneous mixture of pale-colored antigorite serpentinite and dark bluish-gray zones rich in talc, chlorite, and tremolite-actinolite that are characteristic of metasomatic reaction zones that form between ultramafic and crustal rocks. The serpentinite entrained in the fault differs mineralogically and texturally from extensive nearby outcrops of low-temperature lizardite-chrysotile serpentinite associated with the Coast Range ophiolite. The antigorite and metasomatic minerals in the BSF gouge are stable at temperatures above 250°-300°C. They were tectonically incorporated into the fault and rose buoyantly to the surface from depths near the base of the seismogenic zone. Low-temperature alteration of the fault rocks accompanying near-surface shear has produced minor to moderate amounts of Mg-rich saponitic smectite clays. The mineral assemblages, textures, and mode of emplacement of the BSF serpentinitic gouge correspond to a younger, lower-offset equivalent of the creeping San Andreas Fault encountered in the SAFOD (San Andreas Fault Observatory at Depth) drillhole. This suggests a common origin for creep at both locations, and one that may be applicable to other creeping faults of the San Andreas system in northern California.

Interstory drift, the relative displacement between two adjacent floor levels of a building structure, is an important observable that is utilized in engineering design codes and standards to define building design limit states, to assess the earthquake demands on a building structure, and to evaluate a building condition after a major earthquake. The accurate measurement of interstory drift has historically been a technical challenge when using accelerometer-based measurement systems. In this presentation a new, laser-based sensor for measuring interstory drift is described and the results of combined experimental and numerical studies to assess the sensor performance are described. This recently developed sensor offers the potential for directly measuring interstory drift and providing building drift demands for decision making on building occupancy and continuity of operations immediately after a major earthquake.

Numerous constraints to the Cordillera crust and upper mantle temperatures indicate remarkably uniformly hot from Mexico to Alaska, 800-850C at the Moho. The high temperatures are a consequence of the Cordillera being a current or recent backarc with shallow asthenosphere convection fluxed by water driven off the downgoing oceanic plates. A global compilation has shown that all continental backarcs are all similarly hot.

Some of the consequences of the uniformly hot backarc temperatures are:

1. The Cordillera is a high elevation mountain belt in spite of having a uniformly thin crust, avg. 33 km. The elevation results from thermal expansion.

2. The Cordillera is a mobile belt, hot enough to be deformed by normal plate tectonic forces (and resulting strong seismicity). Cratons and other stable areas are too strong for significant deformation.

3. The Cordillera (and other backarcs) have widespread lower crust horizontal detachment and also channel flow, especially indicated by the constant thickness crust in spite of a history of widespread crustal extension and crustal shortening.

4. The Cordillera and other backarcs have widespread sporadic backarc igneous activity. Temperatures are close to the upper mantle H2O-saturated first melting.

5. Finally, there are implications for: (a) the involvement of hot backarcs in continental collision orogeny, and (b) backarcs are the origin of most regional (Barrovian) metamorphism, not orogeny-related deformation heating.

The Asteroid Threat Assessment Project (ATAP) at NASA Ames Research Center has developed an integrated approach, including experimental testing and computational simulations, to characterize the risk associated with Earth impacts. An overview of ATAP activities will be presented followed by a discussion about the use of numerical models to parametrically quantify the ground hazard associated with airburst and surface impacts. We'll specifically consider modeling the transfer of a meteor's kinetic energy to the atmosphere, including the propagation of the subsequent overpressure waves to the ground. Additionally, ocean impact simulations will be examined to characterize the initial cavity formation resulting from impact as well as implications on long-range wave propagation. The results from the hazard simulations enable the use of reduced order models in a probabilistic risk assessment. We will discuss how these models are used and cover ensemble risk estimates. Lastly, we will provide an example of how the integrated risk assessment was incorporated into the 2017 NASA Near Earth Object Science Definition Team report.

Previous fault-based earthquake forecast models have generally assumed segmentation, excluded multi-fault ruptures, and ignored spatiotemporal clustering (aftershocks and otherwise triggered earthquakes), each of which has been brought into question by recent earthquakes. The Working Group on California Earthquake Probabilities has addressed all these issues in their most recent model: the Third Uniform California Earthquake Rupture Forecast (UCERF3), which is the first model to provide self-consistent rupture probabilities over forecasting intervals from less than an hour to more than a century, and the first capable of evaluating the short-term hazards due to multi-event sequences of complex faulting (and earthquake triggering). We find that combining a fault-based forecast with statistical seismology clustering models (e.g., ETAS) requires both characteristic magnitude-frequency distributions on faults (non Gutenberg Richter), as well as the inclusion of elastic rebound, apparently settling the debate on whether either of these attributes is required. Potential usefulness with respect to operational earthquake forecasting (OEF) is presently being explored.

Geothermal energy extraction via Enhanced Geothermal Systems (EGS) is being explored at the pilot level, but still faces technical challenges related to hydraulic fracturing of rock, as well as placement of proppant into rough fractures to maintain a stable aperture for successful long-term operation. This presentation focuses on the role of micro-mechanical analyses in gaining a better understanding of the complex, coupled behavior associated with EGS development. The particular micro-mechanics tool used in this study is the Discrete Element Method (DEM). First, the Bonded Particle Model (BPM) in DEM is used to simulate granite behavior and hydraulic fracturing. The BPM was improved to investigate hydro-thermo-mechanical fracturing processes by implementing a novel convective-conductive heat transport model. Second, DEM coupled with Computational Fluid Dynamics (DEM-CFD) is used to study horizontal proppant flow and transport in narrow fracture zones and proppant settling in a narrow rough granite fracture. A new particle contact model was implemented into DEM-CFD to account for the effects of the fluid lubrication force on particle collisions and the dissipation of particle kinetic energy.

Novel contributions to the understanding of EGS formation using these micro-mechanical analyses in DEM will be presented, along with validation using recent laboratory results. First, a new understanding of hydraulic fracturing of rock will be presented, including the effects of fracturing fluid properties on fracture shape, branching, and secondary fracture formation, and the effects of a temperature difference between the fracturing fluid and surrounding rock on fracture initiation and propagation. Second, a new understanding of the effect of frequent particle collisions on flow and transport of granular slurry (mixtures of proppant and viscous fluid) will be presented. A particularly interesting finding is that in high viscosity fluid, particles remain in close vicinity form agglomerates. Fluid may flow around these agglomerates in rough fractures, which may lead to clogging or settling before reaching the desired location.

The Hazel Dell paleoseismic site on the Santa Cruz Mountains section of the San Andreas Fault provides the first definitive paleoseismic evidence of two pre-1906 19th century surface rupturing earthquakes. Historical records note a few large earthquakes that caused widespread damage in the region in this time period, however it was not previously known with any certainty on which Bay Area faults these earthquakes occurred. Trenches at the site revealed fine-grained interbedded sand and silt above a buried soil that is faulted in the most recent event, 1906, and in and two earlier earthquakes. We discovered hundreds of pieces of cut redwood chips coupled with numerous buried redwood tree stumps just below the ante-penultimate (E3) earthquake surface. This demonstrates that the redwood forest at the site was cut down right before earthquake E3, and thus E3 must be historic. Radiocarbon is frequently used to date pre-instrumental ruptures but can have combined contextual and analytical uncertainty on the order of decades to centuries, especially in the past 300 years. To better constrain age estimates for the onset of logging, we collected a slab from a buried tree stump and radiocarbon wiggle matched samples collected from growth rings to the INTCAL13 terrestrial 14C calibration curve. With these high-resolution age dates we place earthquakes E2 and E3 in the historic period. We use this high-resolution age data to evaluate the impact of C14 sample selection (including detrital charcoal, cones and needles), and use of non-native pollen analysis on our earlier earthquake age models for the Hazel Dell area.

This presentation will focus on the impact of unsaturated conditions on the dynamic response of unsaturated soil layers, with the goal of understanding the seismic compression of unsaturated soil layers during earthquakes. After an introduction to unsaturated soil mechanics, the first half of the presentation will focus on an experimental evaluation of the dynamic properties of unsaturated soils. This includes a summary of resonant column tests highlighting the roles of stress state, suction hardening, and hydraulic hysteresis on the small-strain shear modulus of unsaturated soils, and a summary of cyclic simple shear tests highlighting the modulus reduction and seismic compression of unsaturated soils. The second half of the presentation will focus on different effective stress-based methodologies that can be used to estimate the seismically-induced settlement of a free-field layer of unsaturated sand. This includes both an empirical methodology that synthesizes and extends available empirical methodologies for saturated and dry soils, and a semi-empirical methodology that extends the UBCSand constitutive model. The validation of both models as well as parametric evaluations will be presented to highlight model features and the path forward in future research.

The Mendocino Triple Junction (MTJ) lies at the transition point between the San Andreas Fault system, the Mendocino Transform Fault, and the Cascadia Subduction Zone. Many of the active faults of the triple junction are located offshore, making it especially difficult to characterize aseismic creep. In this work, we study aseismic creep rates near the MTJ using characteristically repeating earthquakes (CREs) as indicators of creep rate. Using seismic data from 2008-2017, we identify CREs as recorded by an array of eight 100-Hz borehole seismometers deployed in the Cape Mendocino area. We apply a repeating earthquake detection criterion to the waveforms, and identify several dozen sequences of repeating earthquakes. The CRE data implies a creep rate of ~3 cm/yr on the downgoing extension of the Mendocino Transform Fault, consistent with estimates of coupling on other oceanic transform faults. We also find repeating earthquakes on the southern margin of the North American accretionary wedge. The creep rates are mostly steady over the study period, although we investigate the potential for small transients due to stress interactions between moderate-sized earthquakes and the CRE sequences. This project demonstrates the ability of CREs to monitor aseismic creep and its time-dependent features in an important plate boundary region outside the coverage of current geodetic networks.