Abstract: The Peninsula section of the San Andreas Fault is an obvious and significant hazard to San Francisco and peninsula cities, however the history past events on the SAF on the peninsula is poorly known. We opened a paleoseismic site at Lake Merced, within the city of San Francisco. The lake may be an asymmetrical pull-apart basin with the NSAF system. We conducted a coring and geophysical investigation to test the site for its potential for development of a long-term earthquake history. We collected four main long cores of 5-7 meters, and 4 overlapping cores. The lake has a record of turbidites that dominate the 3-4 m lacustrine gyttja section, overlying a sharp contact with shelly-sandy estuarine material below. The lake record above the contact includes 15 turbidites spanning ~ 2000 years. The turbidites for the most part have sharp bases, some with load features, and fining upward sequences that can be correlated around the lake. Their individual thicknesses show little relation to stream input. Breaching of the spit at the north end of the lake has occurred in historic times, but appears to have drained the lake without open access to the sea, suggesting the event beds are not externally sourced. Radiocarbon dating thus far shows the uppermost 2 events to include bomb carbon, and occurred in ~1955-57 and 1976-1996. These two beds may relate to a local earthquake near the lake in 1957, and the Loma Prieta earthquake of 1989. The third event is thick, and has a model age of ~ 1890 (1860-1930), likely the 1906 event, supported by anthropogenic Pb content. Events dated at ~ 1720, and ~ 1580 underlie the 1906 bed, similar in age to those reported for the penultimate events on the North Coast and Peninsular NSAF segments respectively. The remaining event bed ages appear mostly compatible with turbidites at Noyo Canyon, and the Vedanta Marsh, suggesting that likely 6 and possibly 11 North Coast segment ruptures extended to the Peninsula segment.
Slidecast:
https://vimeo.com/278230545
Abstract: In September of 2017 Puerto Rico experienced the devastation of two hurricanes, Irma and Maria. On January 10, 2018 a 7.6 Mw earthquake near Honduras promoted the activation of the tsunami advisory for Puerto Rico. These three recent events have changed the life of Puerto Ricans and have raised an urgency to be better prepared in case of natural disasters that may be affecting our welfare at any time. The Puerto Rico Seismic Network Education and Outreach department (E&O) serves the entire Island community for earthquakes and tsunamis education and preparedness. The E&O department, impacts thousands of people in our community not including larger annual exercise (ShakeOut and Caribe Wave) through talks, conferences, workshops and fairs. From September 20th until December the requests for talks and conferences decreased to almost zero due to the devastation of the Island. The lack of power and water supply, inaccessible roads, collapsed bridges, floods, and lack of communications affected our activities. This, plus the long recovery process changed the priorities of our citizens. By January 2018, the raised awareness of the fragility of our wellbeing and partial recovery from the hurricanes, caused a significant increase in demand for E&O services with many communities indicating a raised urgency to be prepared and take the right actions now, before another disaster strikes. A second peak in requests occurred after the tsunami advisory level message issued for Puerto Rico. Despite outreach efforts to present tsunami and earthquake information to the community through talks, web page, educational material, the media, among others, the community became aware that they have not taken the time to learn and understand the tsunami protocols. Here we present the effects of natural disasters on community demand for E&O services and methods employed to take advantage of the receptiveness of our citizens in order to understand these natural phenomena.
Slidecast:
https://vimeo.com/278230002
Abstract: The value of geodetic data towards rapid non-saturating magnitude and source estimation for medium-to-large earthquakes has been thoroughly quantified and studied over the past decade. These efforts have led to the development of several GNSS-based algorithms for earthquake and tsunami early warning. There are still however questions as to how valuable the improved magnitude and source estimates will be to actual operational earthquake early warning; that is, will additional communities receive a warning with a geodetic source estimate that would not with a seismic source estimate and will such warnings be timely and effective? To explore these questions, we utilize both point-source geodetic magnitude estimates (i.e. peak ground displacement scaling) and finite-fault models derived from rapidly computed coseismic displacements. We first look at several large earthquake case studies around the world and how their ground motion prediction, and ultimately, warning times would be modulated by geodetic source information. Secondly, we leverage synthetic earthquake examples in Cascadia to quantify where geodetic information is most valuable with respect to population centers and the seismic hazards, identify the strengths and weaknesses in the current network design, and finally move towards making cost-benefit assessments based upon the Gutenberg-Richter frequency-magnitude distribution of earthquakes in different regions.
Slidecast:
https://vimeo.com/278569444
Abstract: This presentation compares measured and predicted seismic site response at the Garner Valley Downhole Array (GVDA) using a wide range of shear wave velocity (Vs) profiles developed from both borehole methods and inversion of surface wave data. Only low amplitude ground motions, resulting in approximately linear-viscoelastic site response between the downhole accelerometer and the surface accelerometers, were considered in this study. Thus, uncertainties associated with the small-strain Vs profiles used for site response predictions play a considerable role in attempting to match the recorded site response and its associated variability. Prior to our study, two borehole Vs profiles extending into rock were available for the site. However, their predicted/theoretical transfer functions (TTFs) were quite different and in poor agreement with the measured/empirical transfer functions (ETFs). These differences provided motivation to collect and interpret an extensive set of active-source and passive-wavefield surface wave measurements in an attempt to develop deep Vs profiles for the site that might be used to more accurately match the measured site response and its associated variability. Suites of non-unique Vs profiles developed from inversion of the surface wave data, and in conjunction with experimental horizontal-to-vertical spectral noise ratios (HVSR), visually exhibited considerable differences. However, their predicted TTFs matched the measured ETFs very well. Accordingly, we propose that surface wave dispersion and HVSR noise data represent an experimental “site signature” that can be used as a quantitative means of assessing whether candidate Vs profiles are appropriate for use in site response analyses.
Slidecast:
https://vimeo.com/278564632
Abstract: Evidence of segmented ruptures along the Cascadia margin comes in the form of correlated turbidites offshore, lacustrine turbidites, and coastal paleoseismic records. Alternatives to segmentation of earthquake frequency have been suggested based on slope stability, sediment supply, and variation in local ground motions. Steeper slopes in S. Cascadia have been suggested as a mechanism to increase turbidite frequency on the southern margin. However, upper canyon slopes are steeper in Washington canyons (15-24º), and cores at the base of the steep reaches contain the same lower frequency records found on low gradient reaches (.5-2º), even those with locally expanded sedimentary sections. Sediment supply could also pertain, given the high sediment load delivered by the Eel River in southern Cascadia. However, core records show that the well-documented high-frequency turbidite record from the Eel system is largely contained between the Mendocino Ridge and the Trinidad Canyon plunge pool and fan. This leaves the turbidite frequency largely unchanged between the Rogue and Trinidad systems, with a slight southward increase. Grain size and proxies show a southward coarsening source, but no evidence of systematic southward increase in bed thickness, with some beds actually thinning southward. Sediment supply and stronger shaking in S. Cascadia are thus not likely causes of higher event counts. Vancouver Island records also show no increase in event frequency despite the higher Q (low attenuation) setting similar to that of S. Cascadia. While remote turbidity triggering from distant events has been suggested in Cascadia, the close match with the onshore paleoseismic record over the last 3500 years precludes this mechanism. In addition, lacustrine paleoseismic records closely parallel the southward increase in event bed frequency found offshore, thus increasing event bed frequency remains best explained by southward increase in earthquake frequency in S. Cascadia.
Slidecast:
https://vimeo.com/278230744
Abstract: Phase I of the Source Physics Experiment was conducted at Nevada NNSS site in granite. The goal of this experiment was to study the ground motion generated by underground chemical explosion sources. Of particular interest is the understanding of mechanisms of shear wave generation observed during underground explosions, which may look like natural seismicity in the far field. To understand observed ground motion, we have conducted numerical modeling of the SPE. This talk reviews the main results discovered during the last 5 years of the modeling efforts. It was found that the rock mass mechanical response is not isotropic due to the presence of joints. Transverse ground motion is generated due to mechanical anisotropy which can be characterized if joint orientation is known. Azimuthal variations of both magnitude of this shear motion and its polarity seem to correlate with the joint set orientations. Numerical modeling of the wave propagation through such medium is computationally expensive if these joints are represented explicitly, because of the joint spacing is typically two orders of magnitude smaller than the wave length. We have undertaken such modeling only in the near-field region. Since exact orientation of every joint in the rock mass never can be found, we have used stochastic approaches to build various equally probably realizations of joint distributions for our modeling efforts. With this approach we were able to predict the extend of the tangential motion observed in the near field during the experiment. This work performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.
Slidecast:
