Imaging Seismic Attenuation at the Brady Geothermal Field Using Interferometry

Abstract:

The Poroelastic Tomography experiment (PoroTomo) was conducted in March 2016 at Brady Hot Springs in Nevada. A key goal of the experiment is to understand how fluids travel from shallow aquifers, through faults and fractures, to deep geothermal reservoirs. As part of the effort, seismic, geodetic, and hydraulic technologies were tested and developed to fully characterize the rock mechanical properties. This abstract focuses on the application of seismic interferometry to image variations in seismic attenuation at the site. During the experiment, a large seismic array was deployed and recorded more than two weeks of continuous data, active vibroseis sweeps, local traffic noise, and the ambient seismic wavefield. In this study, we use several methods of seismic interferometry to investigate the site. We focus on two techniques: sweep interferometry uses the energy from the vibroseis sweeps as sources of high frequency energy; ambient noise correlation uses the energy of the ambient background field. In each case, the data recorded at one seismometer are correlated with the data recorded at another to obtain an estimate of the Green function between the two. The 238 geophones, concentrated over a 1.5 square-kilometer area, allow us to calculate nearly 30,000 paths, which we use to characterize the site and measure the localized wavefield. In collaboration with Ormat, pressures were changed during four stages of operation, including shutdown, followed by increased injection and pulsing. These changes caused measurable differences in the material properties beneath the site including the attenuation of seismic energy. We use two methods to study the changes in seismic attenuation at the site. The simplest is to measure the normalized amplitudes of the Green functions and to compare the values during each stage of operation. The second is to model the full waveform of the data, separating out contributions of Qs and Qp. Because P and S have different sensitivity to fluids in fractures and pores, the ratio of Qs/Qp is highly sensitive to the fluid saturation. We see anomalously high values of Qs/Qp at depth at the injection site and following fault boundaries. Over the course of the experiment, we observe large changes in attenuation across the site, bounded by structural features.

Poster:

WED.Monroe.1445.Matzel

Regional EEW Applications in Marmara Region (NW Turkey) for Distant Large Earthquakes

Abstract:

KOERI (Kandilli Observatory and Earthquake Research Institute) operates a seismic network in Marmara Sea region (NW Turkey) consisting of broadband and strong motion stations which has a reliable topology for regional EEW studies. In addition, a seismic network of 10 strong motion stations located close to the North Anatolian Fault crossing the Sea of Marmara is utilized in onsite threshold based studies. The Virtual Seismologist, PRESTo and ELARMS2 are regional EEW applications which are operational at KOERI data center. In addition, the onsite EEW system is running for more than a decade. The early warning signal is communicated to the appropriate servo shut-down systems of the recipient facilities, that automatically decide proper action based on defined alarm levels. Warning time is a critical parameters in EEW studies affected by the difference between the arrival time of the P-wave at the seismometer and the picking time. We have observed that the main reason for delays in detecting earthquake is packet size of the waveform data. We confirm that Seiscom3 platform requires some time to locate earthquakes and it needs to be optimized for EEW studies. Even with current settings Istanbul could have up to 20 seconds of warning times for a possible Marmara earthquake in case the rupture initiates in western part of the Sea of Marmara. The observations reveal that each algorithm has its own strengths and weakness. We give an example event (11-05-2015 04:16 GMT Gemlik bay event, M=3.9) that was detected successfully by the three algorithms shown. The first estimation time of the origin time by the three algortihms was between 14-16 second. This examples shows that despite the proximity of the event to Istanbul Metropoliatan area a couple of seconds of warning time remains. The second example shows the neccessty of EEW for distant earthquakes where we utilize the offhore 2014 North Aegean Sea Earthquake (Mw=6.9). Because of the sparse seismic network the first estimation of the epicenter was done in 35 seconds after the origin time of the earthquake prividing about 50 seconds leading time for Istanbul area located about 300 km away from the epicenter where the long period waves were causative for 10-15 minute shaking duration of high rise buildings.

Slidecast:

https://vimeo.com/277559865

Recent Findings and Recommendations for an Updated Hazard Characterization of the Eglington Fault in Las Vegas Valley, Nevada

Abstract:

The Las Vegas Valley fault system (LVVFS) is a complex set of north- to northeast- trending, intra-basin Quaternary fault scarps up to 30 m high that displace alluvial fan, fine-grained basin fill, and paleo-spring deposits in the densely populated Las Vegas metropolitan area. Characterizing the seismic hazard of the LVVFS is currently the focus of a multi-year collaborative study involving researchers from the Nevada Bureau of Mines and Geology, University of Nevada, Las Vegas, and the U.S. Geological Survey. The Eglington fault is the only LVVFS fault currently included on the National Seismic Hazard Map (NSHM), and is a priority focus in the early stages of the investigation. Substantial uncertainty remains regarding the seismogenic potential of the LVVFS. Two endmember hypotheses have been proposed regarding the mechanisms responsible for producing the scarps associated with the LVVFS, including the Eglington fault: 1) tectonic (e.g., coseismic surface rupture) and 2) non-tectonic (e.g., prehistoric differential sediment compaction). In this presentation, we will summarize existing geologic, geodetic, geophysical, and geochronologic data that provide insight into the mechanism(s) responsible for scarp formation within the LVVFS, and present unresolved problems with both endmember tectonic and non-tectonic scenarios. We will also discuss in-progress efforts to characterize the seismogenic potential of the Eglington fault including: planned paleoseismic trenching, geologic mapping using lidar and predevelopment topography derived from historical aerial photographs, Optically Stimulated Luminescence (OSL) dating of the Las Vegas basin stratigraphy, and evaluation of the potential for differential sediment compaction across the fault scarps. In addition, we will present the recommendations from the 2018 Working Group on Nevada Seismic Hazards, including the details of a logic tree framework to address uncertainty in the LVVFS hazard assessment.

Slidecast:

https://vimeo.com/277703108

Operational Experience with Next-Generation Automatic Association Software NET-VISA

Abstract:

The NET-VISA software produces an automatic combined seismic, hydroacoustic and infrasound bulletin resulting from the key step of assembling detections from multiple stations within the processing chain of the International Data Centre (IDC) of the Comprehensive Nuclear-Test-Ban Treaty (CTBTO). The IDC waveform analysts are systematically evaluating the results of using it as a complement to the current operational software Global Association (GA), which is nearing its 19th anniversary in continuous operation at the IDC. Events that otherwise have been missed by the standard processing are presented to the analysts from a processing pipeline running in parallel with the GA software. After just seven days of evaluation, the number of events added to the Reviewed Event Bulletin (REB) that originate with NET-VISA represent on average 12.3% of the total of the events in the REB. Out of this total, the number of events with a valid body wave (mb) or local magnitude (ML) larger than or equal to 4 is 3.5%, indicating that most added events fall below this threshold. This paper will present a more complete analysis based on multiple weeks of operational use.

Slidecast:

https://vimeo.com/277699087

Towards Structural Imaging Using Scattering Artifacts Detected in Ambient Field Correlations

Abstract:

Correlations of the ambient seismic field have been used successfully for tomographic imaging of the Earth on a wide range of scales. This is based on the theoretical and experimental observations that correlation functions computed between the signals recorded by two stations contain an approximation of the impulse response (Green’s function) between these stations. The waves that comprise the ambient field are subject to scattering due to the heterogeneous earth which can generate supplementary arrivals on the correlations functions. It is possible to use these effects of scattering that do not correspond to the propagation between the two stations considered, to locate potential external sources of signal. For this analysis, we use correlation functions computed from continuous signals recorded between 2001 and 2017 by seismic stations in Central California. We identify supplementary arrivals in the correlation functions and use array analysis to map the source of scattering effects linked to strong structural variations. We are particularly interested in imaging scatters that result in coupling between the P-SV and SH systems because they are a consequence of lateral heterogeneities in Earth structure. For that reason, we particularly focus on components of the correlation tensor, different from the Vertical-Vertical component, and we expect the Radial/Vertical to Transverse components to be particularly helpful.

Slidecast:

https://vimeo.com/278010585

Robustness of κ0 Measurement: Insight from a Site-Specific Study in the Low-to-Moderate Seismicity Context of Southeastern France

Abstract:

The site component of κ (κ0) is used in engineering seismology to describe the high-frequency attenuation at a site. It is an important input parameter for various applications (stochastic modeling, ground-motion prediction equations, host-to-target adjustments, etc.). Its evaluation faces, however, several issues as it is difficult to properly isolate κ0 from the source and path terms of κ, and because its measurement is subjected to the operator subjectivity and to large uncertainties. This is particularly true in low-to-moderate seismicity areas because the quantity and bandwidth of the usable data are generally limited. Therefore, κ0 measurements might have higher sensitivity to site amplification, frequency-dependent attenuation, and to the earthquake source properties. Here, the κDS (displacement spectrum) approach of Biasi and Smith (2001) is compared with the original κAS (acceleration spectrum) approach of Anderson and Hough (1984) for three sites in an industrial area in Provence (southeastern France). A semiautomatic procedure is developed to measure individual values of κr that reduces inter-operator variability and provides the associated uncertainty. A good agreement is found between κ0_AS and κ0_DS for the two hard-rock sites, which yields κ0 ∼30 ms. The comparisons between these approaches are also used to infer the reliability of κ measurements by addressing their sensitivity to site amplification, frequency-dependent attenuation, and the earthquake source properties. First, the impact of site amplification on κ0 estimates is shown to be very important and strongly frequency-dependent for stiff-soil sites, and non-negligible for hard-rock sites. Second, frequency-dependent attenuation cannot be ruled out for κ, as indicated by comparison with the literature quality factor (Q) for the Alps. Finally, a source component for κAS is questionable from the comparison of κr_AS evaluated for a cluster of events that shared the same path and site components.

Poster:

WED.Monroe.1700.Perron