Schedule Overview

All locations at the Salt Palace except where noted

Tuesday 16 April 2013
9:30–6:00 PMBoard of Directors Meeting
Room 260 (Upper Mezzanine)
1:00–5:00 PMPre-Meeting Career
Workshop

Cottonwood 1 & 2,
Radisson Hotel Downtown
3:00–8:00 PMRegistration
Lower Mezzanine
6:00–8:00 PMOpening Reception
Upper Mezzanine
Wednesday 17 April 2013
7:00 AM–6:00 PMRegistration
Lower Mezzanine
7:30–8:30 AMLight Breakfast
Hall 1
7:30 AM–NoonPoster Sessions
Hall 1
8:30 AM–NoonOral Sessions
Rooms 155A, B, C, & D
Noon–2:00 PMAnnual Luncheon
Hall 2
2:15–5:45 PMOral & Poster Sessions
Rooms 155A, B, C, & D
Hall 1
5:45–7:30 PMStudent Reception
Salt Palace West,
Above the West Lobby
7:30–9:00 PMTown Hall Meeting
Wasatch Ballroom,
Radisson Hotel Downtown
Thursday 18 April 2013
7:00 AM–6:00 PMRegistration
Lower Mezzanine
7:30–8:30 AMLight Breakfast
Hall 1
7:30 AM–NoonPoster Sessions
Hall 1
8:30 AM–NoonOral Sessions
Rooms 155A, B, C, & D
Noon–1:00 PMPublic Policy Luncheon
Hall 2
1:30–5:00 PMOral & Poster Sessions
Rooms 155A, B, C, & D
Hall 1
5:15–6:15 PMJoyner Lecture
Room 155E/F
6:15–7:15 PMMeet & Munch
Hall 1
Friday 19 April 2013
7:00 AM–6:00 PMRegistration
Lower Mezzanine
7:30–8:30 AMLight Breakfast
Hall 1
7:30 AM–NoonPoster Sessions
Hall 1
8:30 AM–NoonOral Sessions
Rooms 155A, B, C, & D
Noon–1:00 PMLuncheon
Hall 2
1:30–5:00 PMOral & Poster Sessions
Rooms 155A, B, C, & D
Saturday 20 April 2013
7:30–8:00 AM
Returning ~5:30 PM
Board bus for Field Trip
Radisson Main Lobby

Pre-Meeting Workshop

Career Game-Changers:
Strategic Avenues to
Landing the Right Job and
Finding Success in Science

Tuesday, April 16, 1:00–5:00 PM


with Alaina G. Levine, President, Quantum Success Solutions

Alaina asks the big questions early-mid career scientists and engineers often have the most difficulty answering.

The answers are not always straightforward and this workshop is here to give early-mid career scientists and engineers the tools needed to create their own path. Participants will leave the workshop and their one-on-one career consulting session with the skills necessary to understand and utilize their own value, to maximize their potential, and to achieve their career goals both in the short and long term.

Specifically targeted towards graduate students, postdocs, and early career professionals, this workshop will focus on the current and expanding crisis in the job and career market for scientists, and how early-career scientists can best prepare for this challenge. We will specifically address career planning strategies for international scholars, and touch on opportunities both in and outside of academia.

Topics Include

  • How to orchestrate a personal career plan and develop a Plan B and Plan C for contingencies;
  • What early-career scientists should do now to enhance their CVs/resumes and research reputations;
  • What traditional and non-traditional career opportunities are available;
  • How to access and assess “hidden” job opportunities;
  • The biggest mistakes early career scientists can make and how to avoid them (this can be removed or edited)
  • How to develop a networking strategy that delivers real results;
  • How to find unique sources of funding.

One-on-One
Career Consulting Sessions

After the workshop, Alaina will meet with workshop participants about career issues. As a neutral, external, third-party, she can be a completely confidential and objective resource who can provide advice specifically for your situation. Consultations will be approximately 20 minutes long and will be offered to the first 36 who sign up for the course. Participants meet to discuss any career-related issues they have including: career planning and searching, CV/resume writing; cover letter writing, professional development, professional etiquette, networking, negotiation, etc. Participants are encouraged to bring their CV/resume, cover letter, or any other item they wish to have evaluated.

Sponsored by the SSA Kanamori Fund

SSA 2013
Icebreaker
Opening Reception

Tuesday, April 16, 6–8 PM
Upper Mezzanine
Salt Palace Convention Center

Join your colleagues and peers for an evening of pre-meeting conversation.

Enjoy the scenic views
and tasty food and beverages.

Mingle, mix, and reconnect!

Icebreaker attendees must be
registered to receive drink tickets.

SSA Annual Luncheon

Featuring the President’s Address
by Christa von Hillebrandt-Andrade

Wednesday Noon–2:00 PM
Hall 2

Presentation of Awards

Harry Fielding Reid Medal
James Rice

Frank Press Public Service Award
Randall White

Charles Richter Early Career Award
Katsuichiro Goda
 

2013 Joyner Lecture
Thursday, April 18th 5:15 PM
Room 155E/F

Kelvin Berryman
Manager, NZ Natural Hazards Research Platform

The Context and Impacts of the Canterbury Earthquake Sequence of 2010-2011

Kelvin manages the newly-formed research platform that integrates all of New Zealand’s government-funded research in natural hazards. The portfolio ranges from geological and weather-related hazards integrated natural hazard risk, resilient engineering and infrastructure research, and societal and land-use planning aspects of natural hazard mitigation. Kelvin is a Principal Scientist at GNS Science and one of New Zealand’s leading earthquake geologists, with particular expertise in the behavior of active faults and seismic hazard in the Pacific region. He has carried out probabilistic earthquake and tsunami hazard studies from site specific to the national scale. Recent work has focused on hazard assessments for nuclear facilities in Australia and Japan. Kelvin is also a Principal Investigator in the GEM Faulted Earth - Global Active Fault and Seismic Source Database project. In June 2011 Kelvin was recognized in the Queen’s Birthday honors list as a Companion of the Order of the Queen’s Service Order for services to science and Canterbury earthquake recovery.


William B. Joyner
Memorial Lectures

The William B. Joyner Memorial Lectures were established by SSA in cooperation with the Earthquake Engineering Research Institute (EERI) to honor Bill Joyner's distinguished career at USGS and his abiding commitment to the exchange of information at the interface of earthquake science and earthquake engineering, so to keep society safer from earthquakes. Joyner Lecturers are chosen on the basis of their work at this interface, whether they are contributions from earthquake science to earthquake engineering or from earthquake engineering to earthquake science.

Town Hall Meeting
Utah Earthquakes and You – It’s Personal!

Join Us
For a Free Night of
Education, Community Discussion
and Hazard Awareness!

Wednesday April 17th
7:30 – 9:00 PM
Radisson Hotel Downtown
Wasatch Ballroom / 2nd Level
215 West South Temple
Salt Lake City, Utah

The public is invited to a Town Hall Meeting sponsored by the Seismological Society of America as part of their 2013 Annual Meeting, and held in conjunction with the Great Utah Shakeout. Much has been learned about the earthquake threat and vulnerability in Utah. We know where earthquakes are likely to occur and what they can do. We know how to reduce losses from large earthquakes but we have not done enough to prepare for the next large Utah event. The purpose of this Town Hall Meeting is to help you prepare yourself, your family, your friends, and your community to survive the Big One.

Speakers/Panel Members

Ivan Wong, Principal Seismologist, URS Corporation
Ralph Becket, Mayor, Salt Lake City
Walter Arabasz, Research Professor Emeritus, University of Utah
Bill Lund, Senior Scientist, Utah Geological Survey
Christopher DuRoss, Senior Geologist, Utah Geological Survey
James Pechmann, Research Associate Professor, University of Utah
Barry Welliver, Structural Engineer, BHW Engineers LLC
Bob Carey, Operations Chief & Earthquake Program Manager, Utah Division of Emergency Management

SSA 2013
Student Reception


Wednesday, April 17, 5:45–7:30 PM
Salt Palace West
Above the West Lobby

Attend the Student Reception with your peers and fellow SSA Student members, while enjoying yummy food and good spirits.

Don’t miss your opportunity to meet, discuss, and network!

Students must bring their ‘Student Reception invitation’ for admittance – Student invitation will be found in your badge holder.

Drink tickets are provided upon entrance check in to the Student Reception.

Wednesday Luncheon

President’s Address

Christa von Hillebrandt-Andrade,
SSA President 2011-2013
 

Thursday Luncheon

Public Policy Speakers

Walter Arabasz
Reasearch Professor
Emeritus,
University of Utah
Ivan Wong
Principal Seismologist,
URS Corporation,
Oakland CA

“The Big One Will Hit During this Luncheon (or Not)—
Utah Earthquake Probabilities
and Public Policy Making”
 

A Working Group on Utah Earthquake Probabilities will soon complete a report quantifying earthquake probabilities for the Wasatch Front area and the 380-km-long Wasatch fault zone, one of the best studied intraplate faults in the world. Utah has “lots of eggs in one basket” in the Wasatch Front area. Nearly 80 percent of Utah’s population and 75 percent of its economy are literally astride the five central and most active segments of the Wasatch fault. Nearly 130 years ago, G. K. Gilbert warned of the danger of a large surface-rupturing earthquake on the Wasatch fault. The challenge for public policy making is how to motivate a constant, long-term level of preparedness for the next Big One.

Field Trip

Sleeping Giant:
The Earthquake Threat Facing
Utah’s Wasatch Front

Leaders:Bill Lund and Chris duRoss
Utah Geological Survey
Departs:8 AM Sharp Saturday 4/20
(Meet in Radisson main lobby @ 7:30 AM to board bus)
Returning:Approx. 5:30 PM

The Wasatch Front in northern Utah is home to the striking Wasatch Range, numerous cities and communities that house about 80% of Utah’s population, and the most continuous, active normal fault in the conterminous United States — the Wasatch fault. Although no large earthquakes have ruptured the Wasatch fault historically, the fault has a well-documented history of numerous surface-faulting earthquakes in the recent geologic past.

On this trip, we will visit prominent fault scarps on the Salt Lake City segment of the Wasatch fault, review the Holocene surface-faulting history of the fault, and discuss important issues, such as the potential for partial- and multiple-segment ruptures. We will also consider the fault in the context of Pleistocene Lake Bonneville, whose well-defined shorelines and expansive delta deposits help constrain the timing of fault movement. At Little Cottonwood Canyon near the south end of the Salt Lake City segment, we will visit classic normal-slip fault scarps displacing glacial deposits that were first recognized and described by G.K. Gilbert in 1877, causing him to issue Utah’s first earthquake-hazard warning. In the Avenues neighborhood near the University of Utah, we will visit the site of a recent urban paleoseismic investigation on the Salt Lake City segment, present an updated earthquake chronology for the segment, and consider the seismogenic relation between the Wasatch fault and the antithetic West Valley fault zone about 10 kilometers to the west in the Salt Lake Valley.

The second half of the field trip will focus on earthquake monitoring and research, and risk-reduction measures that have been applied in Utah. At the University of Utah Seismograph Stations, we will discuss the regional seismograph network and historical earthquake catalog, the threat of both moderate and large earthquakes, and ongoing seismological research. Following lunch, we will tour the seismic retrofit of the Utah State Capitol building. On the Capitol tour, we will discuss expected ground motions at the site, the details of the base-isolation seismic design, and Utah’s approach to earthquake education and outreach.  Next, we will visit the Warm Springs fault, a trace of the Wasatch fault that enters downtown Salt Lake City and is the controlling geologic structure for the Capitol retrofit. Finally, we will tour the seismic retrofit of the Salt Lake City-County building, one of Utah’s finest historic buildings, and discuss the consequences of a Salt Lake City segment earthquake on the large number of unreinforced masonry buildings in the Salt Lake Valley.

Plans are to leave the Radisson at 8:00 am sharp and return around 5:30 PM on Saturday. This full-day trip includes lunch. There is a small amount of walking and some time outdoors, but it is a very non-strenuous field trip.

NOTE: Clothing Requirements

We plan to tour the new Salt Lake City Public Safety Building, which is currently under construction, to see the seismic-safety design features being incorporated into a state-of-the-art building. However, because the building is still under construction, tour participants are required to have hard hats, safety vests, safety glasses, long pants, and boots. Hard hats, vests, and glasses will be provided for you, but to attend the tour, you will need to have long pants and boots (boots do not have to be steel-toed, but do need to cover your ankle). Also, April weather in northern Utah can be very unpredictable - hope for the best, but come prepared for cool or wet conditions.

SSA 2013
April 17–19
Salt Lake City, Utah

SSA Annual Meeting Mobile Web App
Version 3.0.1

Copyright © 2011–2013
The Seismological Society of America

Questions and/or Comments?
Email webmaster@seismosoc.org

Developed using iUI

Help SSA Identify
Outstanding
Student Presentations!

Please help SSA identify student presentations deserving of the “Best Student Presentation” award by nominating excellent presentations. Student presentations are marked STUDENT in this app (and in the printed program). Student posters will be marked with lime green “Student” tags.

Any attendee (including you) can submit nominations. Please limit your nominations to a few truly outstanding student presentations. To ensure fairness, you may not nominate your own student or advisee, a student from your own institution, or a student on a project in which you are involved.

Nomination forms are available at the registration booth and in the technical presentation rooms. Or you can use your computer, tablet, or smart phone to submit your nomination online here.

Drop off completed paper forms at the registration booth or to the session chair no later than 6:00 pm on Friday, April 19.

GEM Advanced Magnetometers
135 Spy Court
Markham, ON L3R 5H6
Canada
(905) 752-2202
www.gemsys.ca

GeoSIG Ltd
Ahornweg 5A
5504 - Othmarsingen
Switzerland
+41-44-8102150
www.geosig.com

Geotech Instruments, LLC
10755 Sanden Dr
Dallas, TX 75238
214-221-0000
www.geoinstr.com

Guralp Systems Ltd.
3 Midas House
Calleva Park
Aldermaston RG7 8#A
United Kingdom
+44-118-9819056
www.guralp.com

Kinemetrics, Inc.
222 Vista Avenue
Pasadena, CA 91107
626-795-2220
www.kmi.com

Lettis Consultants International, Inc.
1981 N. Broadway, Ste 330
Walnut Creek, CA 9459
925-482-0360
www.lettisci.com

Micro-g LaCoste
222 Snidercroft Rd
Concord, ON L4K2K1
Canada
905-668-2280
www.microglacoste.com

Nanometrics Inc
250 Herzberg Rd.
Kanata, ON K2K 2A1
Canada
613-592-6776
www.nanometrics.ca

REF TEK A Division of Trimble
1600 Tenth St, Ste A
Plano, TX 75074
214-440-1265
www.reftek.com

Save the Dates!
Upcoming
SSA Annual Meetings

SSA 2014
Anchorage, Alaska
April 30–May 2, 2014
Egan Center & Hilton Hotel

SSA 2015
Pasadena, California
April 21–23, 2015
Pasadena Convention Center

SSA
Especially Wishes to Thank

Program Committee Chairs

Keith Koper (Univ of Utah, Salt Lake City)
Ivan Wong (URS, Oakland, CA)

Program Committee Members

Bob Carey (Utah Division of Emergency Management)
Chris duRoss (Utah Geological Survey)
Christine Gammans (Univ of Utah)
Bill Lund (Utah Geological Survey)
Jim Pechmann (Univ of Utah)
Barry Welliver (BHW Engineers)
Dave Wald (Board of Director's rep, USGS Denver)

Advances in High-Frequency Physics-Based Earthquake Systems Simulation
Conveners: Ricardo Taborda, Carnegie Mellon University, Kim Olsen, San Diego State University, and Jacobo Bielak, Carnegie Mellon University, and Tom Jordan, University of Southern California
Recent advances in high-performance computing and efficient wave propagation methods render it computationally tractable to simulate earthquakes using physics-based, deterministic modeling approaches at frequencies up to 10 Hz within the next decade. There are, however, a series of challenges that will need to be overcome in order to generate realistic synthetic seismograms that resemble records from past earthquakes, so that they can be used with confidence in engineering design. These challenges include the realistic characterization of both the earthquake source and the surrounding media at unprecedented fine scales, and the accurate and efficient solution of wave propagation problems at high frequencies, as well as the uncertainty associated with these and other aspects involved in earthquake simulations. This session solicits contributions in the broad range of earthquake systems simulation, oriented to advancing the field of physics-based, deterministic earthquake simulations at frequencies valid for engineering applications: source generation including random parameters, dynamic rupture with small-scale complexity in fault geometry and time variations, construction and improvement of seismic velocity models, numerical techniques in anelastic wave propagation, efficient grid and mesh generation, material attenuation models, integration of geotechnical data in structural velocity models, stochastic representation of near-surface soft-soil deposits, off-fault and near-surface plasticity, incorporation of surface topography, efficient simulation algorithms, and other related topics. This session will also serve as a forum for scientists interested in participating in the High-F project of the Southern California Earthquake Center.
Broadband Seismic Observations on the Seafloor
Conveners: Anne Trehu, Oregon State University and Gabi Laske, University of California San Diego
Technical and engineering advances in the past several years have resulted in an expansion in a variety of instruments available to the community to make broadband seismic observations on the seafloor. This includes the development of instruments designed to mitigate effects of currents and trawling on OBSs deployed in waters depths <1000 m on continental margins, which allowed the community to go forward with the Cascadia Initiative, a community-planned deployment of a 4-year US-Array-like deployment covering the Juan de Fuca plate and its boundaries. There have also been a number of PI-driven experiments around the globe. The objective of this session is to bring together researchers working with broadband OBS data, in general and those of the Cascadia deployments, to present results from recent experiments and compare notes on the particularly challenging aspects of recording broadband data on the seafloor.
Characterizing Active Faults for Seismic Hazard Assessments
Conveners: Shaun Finn, Boise State University, Thomas L. Pratt, University of Washington, and Pier Paolo Bruno, University of Utah
Fault identification and characterization are important for estimating the magnitudes and recurrence intervals of earthquakes on individual faults as part of seismic hazard assessments. Fault location, geometry and displacement are important for assessing potential levels of ground shaking, and deducing a history of past events is crucial for estimating recurrence intervals. These parameters can be investigated through geophysical imaging, modeling, and geologic, geodetic and paleoseismic studies. The goal of this session is to highlight and illustrate methods that can be used for characterizing faults. We invite papers that describe novel techniques for fault characterization and case studies, especially those that integrate several methods or succeed in difficult urban settings.
Data Products as Research Resources
Conveners: Chad Trabant, IRIS, Aaron Velasco , University of Texas El Paso, and Gavin Hayes , USGS
Data products derived from raw data or synthesized from models are becoming increasingly important with the advancement of processing techniques and growing volumes of raw data. These data products often serve as stepping-stones to further research. Data product offerings from organizations such as the IRIS DMC, the USGS NEIC, UNAVCO and others in the scientific community have grown significantly recently. Examples include simple record sections and other even-oriented waveform plots, standardized receiver-functions (EARS), Earth model repositories, integrations of cross-disciplinary research (e.g., the USGS Slab1.0 model), and repositories of synthetic seismograms. We welcome abstracts highlighting new or potential products of interest to the SSA community including, but not limited to, products relating to earthquake, global, regional and strong-motion seismology, geodesy and infrasound. We hope this session will broaden the awareness of available products and spur the development and sharing of new products.
Earthquake Source Physics
Conveners: Paul Segall and Eric Dunham, Stanford University
The session welcomes studies of the mechanics of earthquakes, from interseismic loading to earthquake nucleation, dynamic rupture propagation and arrest, and post-seismic relaxation. Of particular interest are studies that integrate fault constitutive laws consistent with laboratory experiments into models that relate to the full range of seismological, geodetic, and geologic observations.
Earthquake Source Studies
(Poster Only Session)
This session includes a wide range of papers focused on inferring seismic source properties. Topics include earthquake detection and location, magnitude scales, and moment tensor estimation.
Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Models
Conveners: Luis Dalguer, Institute of Geophysics ETH-Zurich, Sinan Akkar, Middle East Technical University, and Cecile Cornou, LGIT, and Aysegul Askan, Middle East Technical University, and Eser Cakti, , and Mathieu Causse, , and Andre Herrero, INGV
This session is organized jointly by SSA and the European Seismological Commission (ESC). To reduce the impact of earthquakes on human lives and our environments it is essential to connect earthquake science with engineering practice. To reach this goal it is necessary to promote the continued integration between earthquake science and engineering, which is the purpose of this session. Recent developments of physics-based earthquakes models have contributed to substantial advances in our understanding of the earthquake mechanism and spatial distribution of near source ground motion. A further collaboration with the engineering community certainly would improve our ability to merge these physics-based findings into engineering solution models. We welcome novel studies that promote the implementation of seismology research advancements into engineering practice. These advances range from increased understanding in the basic physics of earthquake ground motions and fault rupture mechanics to fundamental inputs for hazard maps, risk assessment, ground motion prediction equations (GMPE), ground motion spatial variability and coherency models, seismic designs, building codes and construction practices.
Including Ground Failure in Scenario Events, Rapid Response, and Loss Estimation Models
Conveners: Eric Thompson, Laurie Baise, Tufts University, and Keith Knudsen, USGS
Liquefaction, lateral spread, and landslides are all ground failures that can result from earthquake shaking. The inclusion of these secondary effects in scenario events, rapid response, and loss estimation is not common practice; yet, these effects can cause significant losses. This session will highlight developments in ground failure models as well as efforts to integrate ground failure in regional event specific hazard and loss estimates.
Induced Seismicity
Conveners: Ivan Wong, URS Corp and Wesley Brown, Stephen F. Austin State University
In the past two years, there has been a renewed interest in induced seismicity largely because of the apparent increase in seismicity and felt events in parts of the central and eastern U.S. The causes of induced earthquakes and how such earthquakes may be controlled, including seismicity from hydraulic fracturing and the injection of its waste fluid, have become important research topics. In this session, papers on induced seismicity associated with waste water-injection, geothermal exploration, hydraulic fracturing, and mining will be presented. These papers include several of the most recent and controversial cases of fluid-induced seismicity.
Infrasound and Seismoacoustics
Conveners: Stephen Arrowsmith, Los Alamos National Lab and Relu Burlacu, University of Utah
The combination of infrasound and seismology is yielding new insights into the coupling of mechanical waves in the solid earth and atmosphere as well as a more complete characterization of the source physics for manmade and natural events near the solid-earth atmosphere boundary. These advances are being driven by new datasets, such as those from the Transportable Array network of seismometers and microbarometers, and by improved atmospheric specifications and advances in infrasound modeling. This session invites contributions from all aspects of infrasound and seismoacoustics including instrumentation, data analysis methods, and source and propagation studies.
Intermountain West Earthquake and Volcano Characterization and Related Hazards
Conveners: Bob Smith, University of Utah, David P. Hill, USGS, and Keith Koper, University of Utah, and Bill Hammond, University of Nevada Reno, and Ivan Wong, URS Corp, and Bill Lund, Utah Geological Survey
This session will address the seismo-volcano processes and related hazards of the Basin-Range Province including Utah, Nevada, Idaho, Montana, and surrounding areas. Specific topics include characterization of physical processes of intraplate lithospheric extension, normal faulting active faulting and seismicity, and volcanism including the large calderas of Long Valley, CA, and Yellowstone National Park. Data from modernized earthquake and GPS monitoring networks and new compilations of Quaternary fault properties provide key information for probabilistic earthquake assessments including block-tectonic models for the next U.S. National Hazard Map. In addition, topics in Basin-Range implementation of dynamic and kinematic models for assessing strong ground motion in fault-adjacent valleys, stress-triggering by earthquakes and volcanoes, liquefaction, and importantly, how seismologists can interact more effectively and provide timely information for emergency management agencies and the public during times of increased activity.
Joyner Lecture - The Context and Impacts of the Canterbury Earthquake Sequence of 2010-2011
Joyner Lecture
Networks and Instrumentation
(Poster Only Session)
This session consists of papers on new developments in seismic networks and instrumentation. These include reports on new types of instrumentation, upgrades of existing seismic networks, and innovations in rapid deployments aimed at recording aftershocks of major events.
New Developments in Earthquake Forecasting and Predictability Research
Conveners: Danijel Schorlemmer, Thomas Jordan, University of Southern California, and Matthew Gerstenberger, GNS Science
The Tohoku (Japan) and Christchurch (New Zealand) events of last year have stimulated research on time-dependent earthquake forecasting models that can improve seismic hazard assessments. Much of this research is being carried out under the auspices of the international Collaboratory for the Study of Earthquake Predictability (CSEP), which provides an infrastructure for the blind prospective testing of earthquake forecasting models using standardized, community-accepted statistical methods. Researchers are expanding CSEP capabilities into the testing of ground-motion predictions and hazard assessments as well as the testing of earthquake early warning procedures, and they are exploring the assimilation of new types of data into physics-based forecasting models. We solicit submissions addressing one or more of the following issues related to earthquake predictability, with special focus on global and regional CSEP experiments: - status of completed or ongoing regional and global forecasting experiments, and plans for future experiments; - retrospective and prospective testing of new models and multi-model combinations; - new and more effective methods for testing and evaluation of forecasting models; - assimilation of new data types into forecasting models; and - prospective testing of ground-motion predictions and hazard assessments.
New Frontiers in Seismic Data Analysis
Conveners: Daniel Bowman, Jonathan M. Lees, University of North Carolina, and Richard C. Aster, New Mexico Inst of Mining & Technology
New analysis techniques can reveal hitherto hidden structures and can enhance the study of critical data features that are not optimally revealed by traditional methods. For example, wavelet analysis and the Hilbert Huang Transform can greatly improve upon the time/frequency resolution of traditional Fourier methods, allowing earthquake spectra to be examined in more detail, and curvelet, gradiometric, or other novel wavefield decomposition methods can reveal structures in array data that may not be apparent using F-K analysis. This session will focus on innovative ways of presenting and analyzing seismic data, with an emphasis on new developments. We invite contributors who have developed or are exploring new techniques and/or who wish to offer relevant evaluations of widely used methodologies.
Next Generation of Ground Motion Prediction Models
Conveners: Ivan Wong, URS and Jennifer Donahue, Geosyntec Consultants
There are currently several ongoing efforts in the U.S. to develop ground motion prediction models. PEER is sponsoring two major efforts: the Next Generation of Attenuation (NGA) West 2 and NGA-East. The purpose of NGA West 2 is to update the 2008 NGA models for tectonically active regions. NGA East is an effort to develop a suite of new ground motion prediction models for the central and eastern U.S. using the community approach taken in the first NGA effort. EPRI is sponsoring an update of the 2004/2006 EPRI ground motion models for the central and eastern U.S. for use in re-assessing nuclear power plant safety. Outside the U.S., ground motion prediction models are being developed in many regions due to the availability of new strong motion data and improved techniques. In this session, presentations will be made to describe the ongoing efforts in the U.S. as well as internationally to develop the next generation of ground motion prediction models. Papers are welcome that describe the databases, methodologies, and results of these efforts.
Oceanographic and Atmospheric Signals in Seismology
Conveners: Keith Koper, University of Utah, Rick Aster, New Mexico Inst of Mining & Technology, and Dan McNamara, USGS
With the recent success of ambient noise surface wave tomography, interest in understanding sources of ambient seismic energy has grown tremendously. This session welcomes studies that detect, locate, characterize, interpret, and model seismic signals from a broad range of environmental and climate related sources. Topics can include characterizing the modal structure of microseismic energy, quantifying temporal variations in background noise properties, and interpreting seismic observations in the context of cryospheric, oceanographic and atmospheric source processes. We also welcome theoretical and numerical simulation studies.
Realtime Monitoring and Early Warning
Conveners: Peggy Hellweg, Berkeley Seismological Laboratory, UC Berkeley
A major effort to develop earthquake early warning systems is underway in the western U.S. These approaches rely on real-time estimates of earthquake size and location that can be transmitted to relevant agencies and the general public before the arrival of damaging S and surface waves. This session highlights recent progress of these efforts, as well as related research in real-time seismic monitoring.
Seismic Hazards and Ground Motions
Conveners: Alan Yong, USGS and Nilesh Shome, Risk Management Solutions
In this session, a wide range of topics on seismic hazards and ground motions will be presented. Recently there has been considerable debate on the validity of the probabilistic seismic hazard analysis approach and its implementation, in part due to the perceived underestimation of seismic hazard along the northern coast of Japan prior to the 2011 Tohoku-Oki earthquake. As a result, there has also been a renewed focus on seismic hazards to nuclear power plants in the central and eastern U.S. Papers on probabilistic seismic hazard analysis, seismic and tsunami hazards worldwide, and earthquake ground motions will be presented.
ShakeMap-Related Research, Development, Operations, and Applications
Conveners: Kristine Pankow, University of Utah, Bruce Worden, , and Kuo-Wan Lin, USGS
The recent upgrade of ShakeMap (to Version 3.5) represents a major improvement in the way ground motions are geospatially interpolated. The weighted-average interpolation approach now accommodates intensity as well as ground-motion data, along with shaking estimates from prediction equations or modeling. While these new capabilities add important new functionality, they require more sophisticated validation of ground motion and intensity relations, and a fuller understanding of uncertainties and spatial correlation. This session explores current research in the modeling and validation of ground motion and intensity relations as well as new or emerging efforts to improve ground motion prediction, site correction and geospatial analyses (along with uncertainties) pertinent to ShakeMap. We also encourage presentations on research, operations, and applications related to ShakeMap. Such applications include, but are not limited to, earthquake scenarios, risk assessment, loss estimation, earthquake response tools, and engineering analyses that utilize ShakeMap ground motion estimates and associated uncertainty information.
The Magnitude X.X Earthquake on the YY of ZZZZ: Major Earthquakes of 2012/13
Conveners: Gavin Hayes, Paul Earle, USGS, and Harley Benz, USGS
Major and mega-earthquakes have arguably become more prevalent in the 21 st century than in preceding decades. The ability of the global community to rapidly and accurately respond to such events has also significantly improved over the same period of time, and seismologists are now regularly served with detailed source characteristics of a recent earthquake within several hours of the event occurring. Seismological meetings over the past few years – both AGU and SSA -- have involved special, late-breaking sessions incorporating such studies for earthquakes that have occurred between session proposal and abstract deadlines. This session solicits abstracts that address such studies for recent earthquakes of import – those that have already occurred in 2012, such as the extraordinary April Sumatra sequence and M7+ events in late August and early September in El Salvador and Costa Rica – and also earthquakes which occur in the intervening period between now and April 2013. We welcome all aspects of analyses of these significant earthquakes, from teleseismic to regional studies using seismic data, to co- and post-seismic GPS and InSAR analyses, and in particular research which tackles such event characterization from a multi-disciplinary approach.
Towards an Integrated Understanding of Slow Earthquakes: What We Know, What We Don’t Know, and How to Move Forward
Conveners: Justin Brown, California Institute of Technology and Harmony Colella, Miami University of Ohio
Investigations of fault slip behavior in the last decade revealed a new class of slow earthquakes. Although rapid progress continues to be made in understanding these new phenomena, the mechanisms and relationships between the variety of slow fault slip behaviors remains enigmatic. Despite numerous studies that utilize observations, theoretical models, and/or laboratory measurements, a unified understanding of these fault slip behaviors does not appear to exist. The purpose of this session is to showcase the current state of knowledge of slow earthquakes, highlight the differences from one tectonic setting to the next, and propose an integrated perspective of the possible relationship between behaviors such as slow slip, tectonic tremor, slow earthquakes, and, ultimately, great earthquakes. Abstracts that combine models, laboratory measurements, and observations are strongly encouraged. We also encourage abstracts that highlight the similarities and differences of cross-disciplinary observations, and comparative studies that examine and explain the differences between geographic regions.
Triggering of Seismic and Volcanic Events
Conveners: Kris Pankow, University of Utah and Chunquan Wu, Los Alamos National Lab
Increases in the number of seismic events, both at close and remote distances, following large earthquakes are well documented and suggestive of some type of triggering mechanism. There are also studies suggesting a correlation between large earthquakes and volcanic events. However, the physics behind the triggering mechanism(s) is poorly understood. In addition, it is not always clear how to distinguish seismic and volcanic events that were triggered from those that would have occurred as a result of background stresses. In this session, we invite abstracts that address all aspects of triggering of seismic and volcanic events by large earthquakes. Example topics include interaction among large earthquakes and implications for seismic hazard, statistical analyses used to define triggered events, the role of static and dynamic stress in the near-field, possible triggering thresholds, and hypotheses and proposals for alternative triggering mechanisms such as afterslip, pore fluids, viscoelastic relaxations, or time-dependent frictional properties.
Velocity Models and Modeling
Conveners: Michael Pasyanos, Lawrence Livermore National Lab and Elizabeth Paulson, USC
This session highlights recently developed velocity and attenuation models derived at local, regional, and global scales. It includes presentations on various techniques of inferring Earth structure, such as ambient noise tomography, as well as tectonic and geodynamical implications of various Earth models.
What are the Limits of Explosion Source Model Predictions?
Conveners: Howard J. Patton, Los Alamos National Laboratory and William R. Walter, Lawrence Livermore National Laboratory
Near-field and far-field observations for a number of chemical explosions point to inadequacies of P-wave source models to predict corner frequency and isotropic explosion moment for small and/or over-buried shots. These include the 1997 Kazakh Depth of Burial Experiments in Kazakhstan and recent Source Physics Experiments in Climax Stock. Some first-principle explosion source calculations point to issues surrounding the depth scaling of cavity radius, but the scope of the problem is unresolved and could involve issues related to chemical explosions as surrogates for nuclear tests and scaling to small yields. This problem highlights a source of uncertainty in nuclear test monitoring which arises from test emplacements that do not conform to standard containment practices used by the U.S. and former Soviet Union test programs. Available regional seismic data from past nuclear tests and models built from that data were influenced by those practices. We invite papers that shed light on this problem for both nuclear and chemical tests conducted under varying emplacement scenarios, particularly depth of burial and source medium. We encourage presentations documenting support for model predictions and/or variance from the predictions in order to assess the scope of the problem and identify future research directions.
When and Why do Earthquake Ruptures Stop? Evaluating Competing Mechanisms of Rupture Termination
Conveners: Austin Elliott, University of California, Davis and Julian Lozos, University of California, Riverside
Cessation of coseismic fault rupture has been suggested to result from a variety of mechanisms, ranging from fault-specific static properties to transient, rupture-history-driven dynamic effects. Field and modeling evidence alike implicate static or quasi-static properties such as fault geometry, frictional asperities or regions of creep, and time-dependent poro-elasticity as strong controls on rupture endpoints. However, static, dynamic, and quasi-dynamic numerical models, as well as mounting instrumental and field evidence demonstrate that, as stress evolves over multiple seismic cycles, transient effects may periodically overcome established static barriers, allowing rupture to continue. While much work has been done to investigate the effects of individual mechanisms on rupture cessation, the next step is to merge disparate studies of competing mechanisms in order to understand their relative roles within a given fault system. We invite presentations that summarize findings from numerical models, laboratory tests, observational analyses, and field and paleoseismic investigations that address various mechanisms that inhibit earthquake ruptures. We encourage comparison of these effects with one another, as well as discussion of how to evaluate which properties may dominate rupture through a given fault system, and of how to determine which effects are persistent over multiple earthquake cycles.
Session: Characterizing Active Faults for Seismic Hazard Assessments
Presenter   Stirling, Mark
Schedule   Thu 10:45 AM / Oral
Room   155B
Appropriate Selection of Magnitude Scaling Relationships for Seismic Hazard Analysis
STIRLING, M. W., GNS Science, Lower Hutt, New Zealand, m.stirling@gns.cri.nz; GODED, T., GNS Science, Lower Hutt, New Zealand, t.goded@gns.cri.nz; BERRYMAN, K. R., GNS Science, Lower Hutt, New Zealand, k.berryman@gns.cri.nz; LITCHFIELD, N. J., GNS Science, Lower Hutt, New Zealand, n.litchfield@gns.cri.nz
A fundamentally important, but typically abbreviated component of seismic hazard analysis involves the selection of magnitude scaling relationships. These are typically regressions of historical earthquake datasets, in which magnitude is scaled to parameters such as fault rupture length and area. The mix of historical data from different tectonic environments, and the different forms of the regression equations can result in large differences in magnitude estimates for a given fault rupture length or area. Regressions such as the extensively-used “Wells and Coppersmith” and “Hanks and Bakun” equations are liberally applied the world over, with limited consideration as to their applicability to a particular environment. Underestimation of the M7.1 4 September 2010 Darfield, New Zealand earthquake by the latter regression of about 0.3 magnitude units is evident. As part of a Global Earthquake Model (GEM) initiative, we have compiled a worldwide set of regressions, and have recommended the most suitable regressions for use in the range of tectonic regimes and fault slip types in existence around the world. Selection of appropriate regressions is largely based on the geographical distribution, age, and quality of earthquake data used to develop them. Our compilation is limited to regressions of magnitude (or seismic moment) on source area or length, and our priority selection of regressions show a large magnitude range (up to a full magnitude unit) for a given rupture length or area. These large differences in magnitude estimates underline the importance of choosing regressions carefully for seismic hazard application in different tectonic environments.
Session: Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Models
Presenter   Van Houtte, Chris
Schedule   Fri 4:00 PM / Oral
Room   155C
STUDENT
Near-Source Kappa Estimates from the Canterbury Earthquake Sequence, New Zealand
VAN HOUTTE, C. J., University of Auckland, Auckland, New Zealand, cvan071@aucklanduni.ac.nz; KTENIDOU, O. J., Institut des Sciences de la Terre, Grenoble, France, olga.ktenidou@ujf-grenoble.fr; LARKIN, T. J., University of Auckland, Auckland, New Zealand, t.larkin@auckland.ac.nz; HOLDEN, C., GNS Science, Lower Hutt, New Zealand, c.holden@gns.cri.nz
The physical mechanism causing the spectral decay parameter (κ) is still unclear. Many researchers conclude that it is a site parameter, while others suggest that it is more likely caused by the source. Most applications of κ assume it characterises attenuation in the shallow crust, e.g. stochastic ground motion simulations (Boore, 2003) and the adjustment of GMPEs between regions using the host-to-target method (Cotton et al., 2006). Typically the variability in κ estimates is large, with previous research attributing the scatter to spatial variations of Q or near-source effects. This study examines near-source estimates of κ (Re < 30km) and assesses whether the variability arises from differences in azimuth, magnitude or focal mechanism. We also determine whether it is reasonable to average κ from the two horizontal components of motion for near-source earthquakes. The opportunity for this study arises from the wealth of near-source, strong motion data from the 2010-2011 Canterbury earthquake sequence in New Zealand, which allows the path-dependent effects of κ to be neglected. Quantifying the variability in κ estimates is important for the seismic design of structures in a probabilistic framework, when using the host-to-target method for site-specific GMPE adjustments.
Session: New Developments in Earthquake Forecasting and Predictability Research
Presenter   Ji, Ping
Schedule   WITHDRAWN
Room   155D
On the Short-Term Seismic Gap before Wenchuan Ms8.0 Earthquake
NIU, A. F., China Earthquake Networks Center, Beijing, China, nafcsb2004@seis.ac.cn; YAN, W., China Earthquake Networks Center, Beijing, China, ywpro@163.com; JI, P., China Earthquake Networks Center, Beijing, China, jiping@seis.ac.cn
The Wenchuan earthquake of M8.0 on May 12,2008,has been the strongest earthquake that caused the heaviest disaster since the founding of PRC in 1949. The Wenchuan Ms8.0 earthquake occurred on the Longmen Shan fault in NE direction and its fracture is the eastern boundary of the Bayan Har block. The results of the geological survey of the event show that the earthquake caused surface rupture of 240 km and it was the thrust movement with strike-slip, accompanied by vertical displacement of 4-6m and strike-slip displacement of 3-5m. Based on historical seismic data, Wen(2009) reported that the epicenter of Wenchuan earthquake was located in a seismic gap formed by earthquakes of M7.0 for hundreds of years in the North-South Seismic Belt. Instead of the long term gap, we proposed a short-term seismic gap surrounded by Ms5.5 earthquakes in 7 years before the Wenchuan earthquake, which was bounded by an ellipse with the major axis of 1000km in SE50° and the minor axis of 200km. Along the major axis, the Guangzhou-Luzhou GPS baseline length showed rapid compression from Jan. 1, 2005 to Dec. 1, 2007, with the maximum compression of up to 28mm, then it rebounded to the normal level on Dec. 20, 2008. The seismic creep slew down at the same time while the GPS baseline length was compressed. In addition, we have found that the some significant anomalies in ground tilt and cross-fault leveling appeared outside the seismic gap, however, there were a few deformation anomalies within the seismic gap. These observation results are similar to the deformation changes before Loma Prieta, Parkfiled and L’Aquila earthquakes(Linde et al.,1992;Bilham,2005;Amoruso et al.,2010) . These characteristics of short-term seismic gap demonstrate that the seismic activity outside the gap maybe an indication of energy transfer to the gap for stress accumulation of Wenchuan earthquake. We have tried to estimate the stress accumulation. This work supported by 2012BAK19B and 201108009 programs.
Session: Velocity Models and Modeling
Presenter   Pasyanos, Michael
Schedule   Wed 8:45 AM / Oral
Room   155A
A Lithospheric Attenuation Model of North America
PASYANOS, M. E., Lawrence Livermore National Laboratory, Livermore, CA, pasyanos1@llnl.gov
Recent moderate-sized, but strongly-felt, earthquakes in eastern and central North America have highlighted the important role of the earth’s attenuation structure in estimating and predicting local and regional ground motions. Over the past several years, we have been developing methods to use the amplitudes of regional phases Pn, Pg, Sn, and Lg to invert for the crust and upper mantle attenuation structure in Eurasia, and have recently starting transporting the methodology to North America. We now have path coverage for most of North America, including Canada, the United States, Mexico, and portions of the Caribbean, with the best coverage in the United States. After describing the development of the model, we will discuss the results in the context of the tectonics of the region, most notably the large differences between western North America and areas east of the Rockies. We will then demonstrate the use of the model in a number of applications including estimating reliable moment magnitudes for the Wells, NV earthquake sequence, the use of the models in strong ground motion prediction for the Mineral, VA mainshock, and in both discriminating and estimating explosion characteristics (depth, yield) of events at the Nevada Test Site.
Session: Next Generation of Ground Motion Prediction Models
Presenter   Goda, Katsu
Schedule   Wed 5:30 PM / Oral
Room   155C
Effects of Finite-Fault Plane Models on Source-to-Site Distance Measures and Ground Motion Models for Mega-thrust Subduction Earthquakes
GODA, K., University of Bristol, Bristol, United Kingdom, katsu.goda@bristol.ac.uk; ATKINSON, G., Western University, London, ON, Canada, gmatkinson@aol.com
Mega-thrust inter-plate subduction earthquakes cause significant damage to modern urban cities. To mitigate catastrophic consequences due to future large earthquakes, development of accurate and reliable prediction tools for possible ground motions and seismic hazard potential is essential. In this study, we focus on a particular aspect of ground motion modeling, dealing with the effects of using different finite-fault plane models on the calculation of rupture distances, and their influence on ground motion modeling. This issue has been recognized in the literature; however, it has not been examined extensively. We address the problem empirically, analyzing strong ground motion data from three recent well-recorded large inter-plate earthquakes. We develop event-specific ground motion models using alternative finite-fault plane models obtained from scientific publications. Then, the developed models as well as the statistics of the observed ground motion intensities are compared to quantify the effects of adopting different finite-fault plane models. When ground motion models are developed using different fault-plane definitions, there are significant variations amongst the considered models, typically leading to differences in predicted ground motion amplitudes by factors of two to three for very large events. This is an additional source of uncertainty and variability in the development of GMPEs, which should be considered in seismic hazard assessment. Importantly, the results suggest that comparison of observed ground motion data with existing ground motion prediction models is not a simple task. Moreover, although the effect of alternative definitions of the rupture plane is most pronounced for mega-thrust events, even for events as small as M7.0, the effects can be significant. Thus we conclude that the way in which the fault rupture plane is defined is important whenever the fault dimensions are in excess of about 50 to 100 km.
Session: Characterizing Active Faults for Seismic Hazard Assessments
Presenter   Amosu, Adewale
Schedule   Fri AM / Poster
Room   Hall 1
STUDENT
Modeling Earth's Crustal Deformation in the Lower Mississippi River Basin
AMOSU, A. M., Center for Earthquake Research and Information, Memphis, TN, walbytes@yahoo.com; SMALLEY, R. J., Center for Earthquake Research and Information, Memphis, TN; PUCHAKAYALA, J., Center for Earthquake Research and Information, Memphis, TN
The surface of the Earth deforms in response to changes in the loads imposed upon it by the atmosphere the cryosphere and the hydrosphere. Of particular interest is the load exerted on the Earth by ice and water masses present on its surface. Global Positioning System (GPS) instruments can measure the deformation signal of the Earth.We measured the 3-dimensional displacements from 1998 to 2011 using GPS stations located along the Mississippi river in the region of the New Madrid seismic zone . The vertical displacements are strongly anti-correlated with the stage of the river.We use topographic and river stage data to estimate the surface loading from the flood. We model the Earth as an elastic half-space to estimate the deformation from the flood loading for comparison with the deformation observed by GPS. We invert the observed deformation to obtain elastic properties of the crust. Once the elastic properties of the crust are known, we can use the observed deformation to weigh loads.Extracting elastic properties of the crust from geodetic measurement will prove to be a useful technique in characterizing elastic structure of regions where near surface elastic structure data is not available; more so as the availability of geodetic measurements around the world is increasing. This method is taking a step towards more complex and more detailed models of the near surface Earth structure, such as a layered model, and provides preliminary information about interesting features in the Earth for further investigation by other methods such as seismic refraction surveys and other seismic techniques.
Session: The Magnitude X.X Earthquake on the YY of ZZZZ: Major Earthquakes of 2012/13
Presenter   Okal, Emile
Schedule   Fri 8:30 AM / Oral
Room   155A
The Showa Sanriku Earthquake of 1933: A Remarkable Octogenarian
OKAL, E. A., Northwestern University, Evanston, IL, emile@earth.northwestern.edu; KIRBY, S. H., U.S. Geological Survey, Menlo Park, CA; REYMOND, D., LDG-CEA, Papeete, Tahiti, French Polynesia; UCHIDA, N., Tohoku University, Sendai, Japan
The earthquake of 02 March 1933 is the largest known outer slope earthquake, and the only one whose tsunami caused damage in the far field. On its 80th anniversary, we offer a comprehensive reassessment of this event, using modern algorithms. Its relocation, and that of more than 100 immediate aftershocks, places the mainshock at 39.22N, 144.45E, on the outer slope of the trench, a site shared by five well-constrained aftershocks, defining a 170-km long fault zone. By constrast, a large number of "post-shocks" relocate across the trench, indicating that they were activated by stress transfer on different fault systems. A classification of these events is presented, based on their first motion waveforms at Mizusawa. We use the PDFM method to invert the moment tensor of the mainshock, using 10 mantle Rayleigh and Love waves, yielding a moment of 7 10**28 dyn*cm and a mechanism (str.= 200; dip = 61; slip = 271 deg.) rotated only 25 degrees from Kanamori's [1971]. We use a number of remarkable short-period records, on prototype seismographs then being developed by Benioff, to estimate the radiated energy, and find a parameter THETA = -4.26, typical of a short or snappy source, with high stress drops. This is also supported by an analysis of felt intensities, and their decay with distance, and by the identification of a T wave signal at Pasadena. Based on recent seismicity, we favor a model of downdip rupture not extending past 30 km, suggesting a slip on the order of 18 m. This source is used to simulate the far field tsunami. In conclusion, the Showa Sanriku earthquake shares many properties with modern-day outer rise and other intraplate earthquakes, and its exceptional size explains its large tsunami. This emphasizes the need to consider this class of events when assessing tsunami risk, both in the near field (where 1933 runup reached 29m), and in the far field, where the first successful warning and evacuation took place in Hawaii during that event.
Session: ShakeMap-Related Research, Development, Operations, and Applications
Presenter   Foulser-Piggott, Roxane
Schedule   Thu PM / Poster
Room   Hall 1
Development of an International Macroseismic Scale Based on EMS-98
FOULSER-PIGGOTT, R., Cambridge Architectural Research Ltd, Cambridge, UK, roxane@carltd.com; SPENCE, R., Cambridge Architectural Research Ltd, Cambridge, UK, robin.spence@carltd.com
The aim of this project is to modify the current version of EMS-98, for Intensity levels greater than VI, to make the scale more internationally applicable. The expected benefits for global risk assessment are: Improvements in the robustness of intensity assignments outside Europe; Provide a standard scale for international use, reducing the need for inter-scale conversions; Enable better correlations to be made between macroseismic intensity and other measures of shaking intensity.A review of field experience using EMS-98 has been conducted to identify areas in which the scale should be developed to make it more suitable for international application. To date, the majority of this data has been collected using an online questionnaire, completed by EMS-98 users in different countries, both inside and outside Europe. The questions covered two main areas: Key features of EMS-98 study-structure types encountered, vulnerability, damage levels, range of intensities; Problems encountered- absence of structure types, insufficient range of vulnerabilities for a structure type, difficulties in making intensity assignments.This paper describes how the EMS-98 field experiences of the international community of seismologists and engineers have been used to identify aspects of the EMS-98 which should be developed to make it more specifically applicable to areas outside Europe with significantly different building stocks. The results of this work are a series of recommendations for modifications to the EMS-98 document which include:Revision of damage descriptions; Review and revision of vulnerability class definitions and ranges; Inclusion of building types not currently represented in EMS-98 and representation of additional building types and vulnerability classes; Investigation of the impact on intensity assignments of changes to EMS-98 vulnerability types and the addition of new building types.Finally, the paper details the methods of validation of the new scale.
Session: New Developments in Earthquake Forecasting and Predictability Research
Presenter   Rundle, John
Schedule   Thu 11:45 AM / Oral
Room   155D
Challenges in Web-Based Real Time Earthquake Forecasting (RTEF): Localizing Forecast Probabilities in Space and Time
RUNDLE, J. B., University of California, Davis, CA, jbrundle@ucdavis.edu; HOLLIDAY, J. R., University of California, Davis, CA, holliday@physics.ucdavis.edu; GRAVES, W. R., Open Hazards Group, Davis, CA, graveswr@gmail.com; DONNELLAN, A., Jet Propulsion Laboratory, Pasadena, CA, andrea.donnellan@jpl.nasa.gov; TURCOTTE, D. L., University of California, Davis, CA, dlturcotte@ucdavis.edu
Earthquake forecasts have been computed by a variety of groups and countries world-wide for over two decades. An example of such a group is the Working Group on California Earthquake Probabilities, which has been responsible for the official California earthquake forecast since 1988. For the most part, forecasts have been computed for insurance, reinsurance and underwriters of catastrophe bonds. However, recent events clearly demonstrate that mitigating personal risk is becoming the responsibility of individual members of the public. Open access to a variety of web-based forecasts, tools, utilities and information is therefore required. The basic forecast model we consider is the Natural Time Weibull (NTW) method (JBR et al., Phys. Rev. E, 86, 021106, 2012). This model uses small earthquakes (‘seismicity-based models”) to forecast the occurrence of large earthquakes, via data-mining algorithms combined with the ANSS earthquake catalog. This method computes large earthquake probabilities using the number of small earthquakes that have occurred in a region since the last large earthquake. Localizing these forecasts in space so that global forecasts can be computed in real time presents special algorithmic challenges, which we describe in this talk. Using 25 years of data from the ANSS California-Nevada catalog of earthquakes, we compute real-time global forecasts at a grid scale of 0.1o. We analyzed the performance of these models using the standard tests, which include the Reliability/Attributes and Receiver Operating Characteristic (ROC) tests. It is clear from much of the analysis that data quality is a major limitation on the accurate computation of earthquake probabilities. We discuss the challenges of serving up these datasets over the web, ongoing results of which can be seen on web-based platforms at www.quakesim.org and www.openhazards.com.
Session: New Frontiers in Seismic Data Analysis
Presenter   Upegui-Botero, Fabio M
Schedule   Thu AM / Poster
Room   Hall 1
STUDENT
Analysis of Time-Frequency and Empirical Mode Decomposition Applied to Seismic Signals
UPEGUI-BOTERO, F. M., Puerto Rico Strong Motion Program UPR-Mayaguez, Mayaguez, PR, fabio.upegui@upr.edu; HUERTA-LOPEZ, C. I., Puerto Rico Strong Motion Program UPR-Mayaguez, Mayaguez, PR, m-huerta@alumni.utexas.net; MARTINEZ-CRUZADO, J. A., Puerto Rico Strong Motion Program UPR-Mayaguez, Mayaguez, PR, jose.martinez44@upr.edu; SUAREZ-COLCHE, L. E., Puerto Rico Strong Motion Program UPR-Mayaguez, Mayaguez, PR, luis.suarez3@upr.edu; Department of Civil Engineering and Surveying, University of Puerto Rico at Mayaguez
Data processing is a necessary step before any analysis, or interpretation of experimental measurements. Seismic waves encountered in earthquake time series and transient signals observed in dynamics systems, clearly show non-stationary behavior, in the sense of amplitude and frequency content. Joint Time-Frequency spectral decomposition basically describes the capability of this signal processing technique to describe the true structure of the signal when the frequency content varies with time. On the other hand, the Empirical Mode Decomposition (EMD) is an adaptive method to decompose any data into a finite small numbers of Intrinsic Mode Function (IMF) components, which become the basis data representation. The EMD is ideally suited for analyzing data from nonlinear and non-stationary processes. An instrumented building may provide adequate information to calculate the response of the structure in detail and compare it with the estimated by numerical modeling. The goal of the system identification is to estimate the system properties, such as dynamic characteristics and modal parameters of structures by means of an input excitation or just from the system response. The objective of this study is to assess the capability of Join Time-Frequency analysis applied in seismic records, moreover the applicability of the Joint Time-Frequency analysis and the EMD to identify structural properties of instrumented building based in earthquake records and/or ambient vibrations measurements. In this work, preliminary results of the Time-Frequency analysis conducted on time series of seismic signals recorded in a building and free-field is presented.
Session: Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Models
Presenter   Baker, Jack
Schedule   Fri 8:30 AM / Oral
Room   155C
Building Code Specifications of Ground Motions for Structural Analysis, and the Potential Role of Physics-based Simulations
BAKER, J. W., Stanford University, Stanford, CA, bakerjw@stanford.edu; BURKS, L. S., Stanford University, Stanford, CA, lynne.burks@stanford.edu; ZIMMERMAN, R., Rutherford + Chekene, San Francisco, CA, rzimmerman@ruthchek.com
ASCE 7 is the Standard that specifies how ground motions must be developed for dynamic structural analysis of new building designs. This talk will present an overview of the requirements in that Standard, and typical procedures for satisfying those requirements using recorded or simulated ground motions. An example application is presented for a building at a near-fault site in Northern California, using suites of both recorded and simulated ground motions that satisfy ASCE 7 guidelines. The properties of the two suites are discussed, with regard to their ability to capture properties of expected future ground motions at the site, and to satisfy the requirements of ASCE 7. Structural analysis results from the two suites are also presented, and links between ground motion properties and resulting structural demands are made. Practical aspects of the two approaches, including the availability of data and software tools, are also discussed. A notable aspect of this use of ground motions is that the hazard analysis and resulting target response spectrum are developed separately, so choice of recorded or simulated ground motions can be made separately from judgments as to whether empirical or simulation based methods are preferable for seismic hazard analysis.
Session: Advances in High-Frequency Physics-Based Earthquake Systems Simulation
Presenter   Shahjouei, Alireza
Schedule   Wed PM / Poster
Room   Hall 1
STUDENT
A Hybrid Broadband Approach to Generate Artificial Seismograms for Central and Eastern US
SHAHJOUEI, A., Department of Civil Engineering, The University of Memphis, Memphis, TN, shhjouei@memphis.edu; PEZESHK, S., Department of Civil Engineering, The University of Memphis, Memphis, TN, spezeshk@memphis.edu
A method to generate strong and moderately strong ground motions is proposed. This modeling of ground motions is applicable for the engineering design purposes in the case of absence of appropriate, sufficient, and reliable recorded earthquakes. We have generated broadband synthetic seismograms for finite fault sources using the numerical techniques of Spudich and Archuleta and the stochastic model which is obtained from the method of Boore using the program SMSIM compatible with Central and Eastern US (CEUS). The broadband synthetics are constructed for earthquakes of magnitudes, 4.5, 5.5, 6.5, and 7.5. Using Green functions, the complete response of the earth structure and near field effects have been incorporated in the model. A spatial random field model has been used to characterize complexity in the slip distribution of the source. These deterministic long period synthetics are combined with the stochastic synthetics at high frequencies generated by the SMSIM program.The desired engineering spectral accelerations for 0.2, and 1.0 second periods are compared with the set of next generation attenuation (NGA) relations. Also, for the engineering application of the time history analysis of structures, we compare and validate the spectrum-compatibility of synthetics with the international building code (IBC-2012).
Session: ShakeMap-Related Research, Development, Operations, and Applications
Presenter   Worden, Bruce
Schedule   Thu PM / Poster
Room   Hall 1
ShakeMap: What’s New?
WORDEN, C. B., Synergetics Inc., Pasadena, CA, cbworden@usgs.gov; LIN, K. W., USGS, Golden, CO, klin@usgs.gov; WALD, D. J., USGS, Golden, CO, wald@usgs.gov; THOMPSON, E., Dept of Civil & Environmental Engineering, Tufts University, Medford, MA
Over the past decade the U. S. Geological Survey's (USGS) ShakeMap® has evolved from a simple but very useful means of visually depicting earthquake ground motions to an indispensable tool for emergency response, planning, loss-modeling, and public information. We have gradually broadened the range of products available to consumers and improved the quality of its ground motion interpolations. ShakeMap is now the standard post-earthquake information tool for all Advanced National Seismic System (ANSS) regions in the US, and in the remainder of the country via the global ShakeMap (GSM) system run at the USGS’s National Earthquake Information Center (NEIC). For global earthquakes GSM forms an important component of many governments’ and agencies' post-earthquake response protocols through the USGS’s PAGER (Prompt Assessment of Global Earthquakes for Response) and ShakeCast systems. Here we discuss the current status of ShakeMap and provide updates on new capabilities and directions. We release the fully-revised Version 2 of the ShakeMap Manual, which provides a comprehensive overview of ShakeMap operations and uses. We also describe a number of new ShakeMap products and features now delivered through USGS product delivery layer (PDL) protocol; these include ESRI Raster files for GIS and refined KMLs and enhanced web-page rendering. Oversight of ANSS ShakeMap systems has led to improved backup strategies using PDL and standardized AQMS database queries. We have also released a major update of the ShakeMap Atlas of historic earthquakes. The expansion of the ShakeMap platform through its installation at numerous international networks will be highlighted with examples from systems throughout Europe, and in Asia and Central and South America. Finally, we invite suggestions from meeting attendees through our highly innovative "suggestion tool", that is an empty text box on our poster, to which we will attach a magic marker.
Session: New Developments in Earthquake Forecasting and Predictability Research
Presenter   Hough, Susan
Schedule   Thu 11:30 AM / Oral
Room   155D
Revisiting the Historical Earthquake Catalog in Northern California: New Insights into the Seismic Cycle
HOUGH, S. E., U.S. Geological Survey, Pasadena, CA, hough@usgs.gov; PAGE, M., U.S. Geological Survey, Pasadena, CA, mpage@usgs.gov; MARTIN, S., Victoria University of Wellington, Wellington, New Zealand
The magnitudes of moderate historical earthquakes in the greater San Francisco Bay Area, California, have been estimated using modern methods, but the intensity assignments used in these analyses have not been reconsidered since the seminal work by Toppozada. Since that time there has been a growing recognition, in part due to intensity data collected by the USGS “Did You Feel It?” (DYFI) system, that many early intensity assignments were too high. We revisit archival accounts for 81 events with previously estimated magnitudes above 5.0. Using two well-calibrated methods to analyze the intensity values, revised magnitudes are generally lower than earlier estimates. Lower magnitudes are supported by direct comparisons with DYFI distributions from recent events such as the 2003 San Simeon earthquake. However, the results support the conclusion of Toppozada et al. (2002) that the June 1838 earthquake was a large (low- to mid-M7) event. We discuss other notable events, including the 31 March 1898 Mare Island earthquake. The revised magnitude of this event (5.8) is more consistent than the earlier estimate with the absence of surface rupture. Using the revised historical catalog and the instrumental catalog, we reconsider evidence for the so-called seismic cycle. The conceptual understanding of a seismic cycle has long been based on apparent evidence that regional seismic activity increases prior to a large event, although Hardebeck et al. (2008) conclude there is no evidence for clustering beyond that predicted by an ETAS (epistemic triggering of aftershock) model. We use the revised catalog to test the hypothesis that seismic activity in northern California is Poissonian, with modulation by stress shadows following (relatively infrequent) M≥6.5 earthquakes. As suggested by Hardebeck (2008), the largest earthquakes tend to occur during the more active periods when activity is higher, leading to the appearance of a seismic cycle with precursory patterns.
Session: Characterizing Active Faults for Seismic Hazard Assessments
Presenter   Federschmidt, Sara
Schedule   Fri AM / Poster
Room   Hall 1
STUDENT
Paleoseismic and Structural Characterization of the Hines Creek Fault at Denali National Park and Preserve, Alaska
FEDERSCHMIDT, S. E., University of Kentucky, Lexington, KY, sara.federschmidt@uky.edu; BEMIS, S. P., University of Kentucky, Lexington, KY, sean.bemis@uky.edu
The Hines Creek fault is a Quaternary-active fault in the central Alaska Range. It is defined as a previously right-lateral strike-slip fault that is currently a dip-slip fault with 6m of south-side down displacement (Wahrhaftig, 1975). With the aid of lidar digital elevation models and field observations, I observed a fault scarp and seismic fissures that cross Pleistocene and Holocene fluvial deposits near the entrance to Denali National Park and Preserve. Little information is known about this fault's recent activity and past researchers disagree on whether the observed seismic features should be attributed to the Hines Creek fault or to the Parks Road fault (Wahrhaftig, 1958; Sherwood and Craddock, 1979; Csejtey et al 1992). Associated seismic hazards are unknown, which is a major issue since the fault crosses an infrastructural corridor that is essential to Alaska. I am determining the local framework, the late Pleistocene and Holocene slip rate, and the seismic history of this fault with geologic mapping, paleoseismic investigations, and dating of Quaternary landforms. Geologic mapping revealed young morphology of scarps and fissures on four different river terraces that span the last 20ka. Therefore, this fault is active, has a history of recurrent earthquakes, and is an important structure in the modern uplift and deformation of the Alaska Range. Four paleoseismic trenches were investigated, with three providing evidence for earthquakes. At least four seismic events can be defined from the paleoseismic investigations and there is a possibility for more events to be defined as radiocarbon samples continue to be processed.
Session: Seismic Hazards and Ground Motions
Presenter   Veeraraghavan, Swetha
Schedule   Fri 2:15 PM / Oral
Room   155B
STUDENT
3-D Dynamic Analysis of Precariously Balanced Rocks Under Earthquake Excitation
VEERARAGHAVAN, S., California Institute of Technology, Pasadena, CA, sveerara@caltech.edu; KRISHNAN, S., California Institute of Technology, Pasadena, CA, krishnan@caltech.edu
Analyzing the toppling behavior of precariously balanced rocks (PBR) can provide limits on the largest ground motions (of the type that the rocks are sensitive to) that could have occurred at the rock sites in the time that they have been precarious. Until now, PBRs have been modeled as 2-D blocks with 2 points of contact. 2-D models with multi-point contact and 3-D models incorporating realistic geometry behave distinctly different from the idealized 2-D block with 2 point of contact. Therefore, we are creating detailed 3-D models of some of the PBRs that have been imaged using Terrestrial Laser Scanning (TLS) and photogrammetric techniques in order to analyze the toppling behavior accurately. To establish the proof of concept, we are modeling the Echo Cliff PBR which is located in the Western Santa Monica Mountains. Under the assumption that the rock and the pedestal behave like rigid bodies, we use rigid body dynamics to solve for the response of the rock to different ground motions. We consider the rock and the pedestal to be separate rigid bodies with the interaction between them modeled using normal contact forces and frictional forces. The response of the rock is determined using force and moment balance equations. We are validating this algorithm by conducting shake table tests on small rocks. We are also analyzing the 3-D response of the rock to idealized saw tooth ground velocity pulses with varying peak ground velocity (PGV), time period (T) and number of cycles (n) to characterize the nature of ground motion that can topple the rock. It then follows that such ground motion could not have occurred at the rock site from regional earthquakes. We are in the process of creating models and performing similar analyses on several PBRs in Southern California to help improve seismic hazard maps and also to validate ground motion simulations.
Session: Including Ground Failure in Scenario Events, Rapid Response, and Loss Estimation Models
Presenter   Rathje, Ellen
Schedule   Wed 5:15 PM / Oral
Room   155B
Regional Predictions of Earthquake-Induced Landslides: How Well Do They Work?
RATHJE, E., University of Texas, Austin, TX, e.rathje@mail.utexas.edu; WANG, Y., University of Texas, Austin, TX, wangyubing1987@hotmail.com; DREYFUS, D., University of Texas, Austin, TX, danieldreyfus@gmail.com
Earthquake-induced landslides are a significant seismic hazard that can generate significant economic losses. To gauge the potential for this hazard, regional seismic landslide maps can be developed based on estimates of rigid sliding block displacement from empirical predictive models. Previous earthquakes with well-documented landslide inventories provide opportunities to evaluate these regional predictions of earthquake-induced landslides. This research investigates the landslide inventories from the 2004 Niigata-ken Chuetsu earthquake (Mw=6.6) and the 1994 Northridge earthquake (Mw=6.7). For each earthquake, the observed landslide inventory is compared with predictions of sliding displacement. Regional predictions of sliding displacement are made using ground motion estimates from ShakeMap®, yield accelerations computed using best-estimate shear strengths assigned to geologic units, and an empirical displacement model using PGA and PGV. Comparisons with the observed landslides indicate that only 40% of the landslides are captured. Larger percentages of the landslides can be captured by modifying various parameters in the yield acceleration calculation (e.g., shear strength), but these changes result in a significant over-prediction in the total landslide area. These results indicate that improvements to regional seismic landslide mapping can only be made with improvements in mapping the spatial variation in shear strengths within geologic units.
Session: New Developments in Earthquake Forecasting and Predictability Research
Presenter   Yan, Wei
Schedule   Wed PM / Poster
Room   Hall 1
A Possibility for Earthquake Prediction Based on Ground Tilt Measurement
NIU, A. F., China Earthquake Networks Center, Beijing, China, nafcsb2004@seis.ac.cn; YAN, W., China Earthquake Networks Center, Beijing, China, ywpro@163.com; ZHANG, L. K., China Earthquake Networks Center, Beijing, China, Zhll1023@163.com
A network for ground deformation measurements had been constructed with a number of stations in China, and a lot of precursory data had been observed and collected. We have found that some short-term anomalies were usually far from the earthquake epicenters, and where were no significant changes for great earthquakes. In order to establish relationships between short-term deformation anomalies and earthquakes, we proposed a time-dependent criterion, what is criterion with minimum waiting time for earthquake after anomaly ended. Based on this criterion, we collected data with sudden changes in ground tilt before earthquakes, and matched them. Then by the least square method, we have given statistic formulas between anomaly durations, amplitudes and earthquake magnitudes and epicenter distances as below: Between anomaly Duration T (in days)and Magnitude M : M=1.36 log T+3.786 (1)Here the averaged square root error in magnitude is 0.38. This formula is close to the Rikitake(1975), but with difference. Between anomaly Amplitude A (in rad) and Magnitude M and epicenter Distance D (in km): M=0.57 log(D)+1.18 log(A)+11.87 (2) Here the averaged square root error in magnitude is 0.4. The formula (2) gives us a possibility to estimate the ranges of magnitude and epicenter distance, while ground tilt anomaly was found out and the anomalous amplitude A was got, the formula (2) can be simplified as follow: M=0.57 log(D) +C (3) There C is a constant, and the D is a power function of M, and there was not characterized point. However, we can estimate the ranges of M and D while the fractal dimension is unchanged. Based on the matched data, we have found there is a minimum distance change of 120km as M changed from M0-0.4 to M0+0.4. The model test had been done. This work supported by project 201108009.
Session: Next Generation of Ground Motion Prediction Models
Presenter   Ghofrani, Hadi
Schedule   Thu AM / Poster
Room   Hall 1
Site Condition Evaluation Using H/V Spectral Ratios of Earthquakes in the NGA-West 2 Database
GHOFRANI, H., University of Western Ontario, London, ON, Canada, hghofran@uwo.ca; ATKINSON, G. M., University of Western Ontario, London, ON, Canada, Gmatkinson@aol.com; CRAWFORD, J., University of Western Ontario, London, ON, Canada, jcrawf4@uwo.ca
We examine the correlation between VS30, the time-averaged shear-wave velocity over the top 30 m, and the horizontal to vertical component ratio (H/V) of earthquake ground motions using the NGA-West 2 database. This is useful because these parameters are alternative simple measures of site response that carry complementary information. Furthermore, such a relationship can be used to predict one of these metrics, if the other is known. Based on the peak log(H/V) and its peak frequency in different regions (Northern and Southern California, Italy, China, Japan, and Taiwan), two model classes were distinguished; each of the regions falls within one of these classes. Class 1 regions include Japan and China, while the other regions are Class 2. Class 1 (“high frequency”) regions have sites with peak frequencies greater than 5 Hz, and Class 2 (“typical”) regions have much lower peak frequencies. The results can be summarized by the following equations: (1) VS30 = (538±223) – (806±404)*log(H/V)peak + (58±10)*fpeak with a standard deviation of 84 m/s on VS30 (Class 1); and (2) VS30 = (542±167) – (723±254)*log(H/V)peak + (107±27)*fpeak with a standard deviation of 106 m/s on VS30 (Class 2). Using these models it is possible to estimate VS30 for seismographic recording sites with unknown site conditions.
Session: Velocity Models and Modeling
Presenter   Yang, Wei
Schedule   Thu AM / Poster
Room   Hall 1
Seismic Velocity Temporal Variation Obtained from ACROSS and Ambient Seismic Noise in the Longmenshan Range-front Fault
YANG, W., Institute of Geophysics, China Earthquake Administration, Beijing, China, weiyang05@163.com; GE, H. K., Unconventional Natural Gas Institute, China University of Petrol, Beijing, China, gehongkui@163.com; CHEN, H. C., Institute of Geophysics, China Earthquake Administration, Beijing, China, seismichc@gmail.com; WANG, B. S., Institute of Geophysics, China Earthquake Administration, Beijing, China, wangbs@seis.ac.cn; QIAO, S., Institute of Geophysics, China Earthquake Administration, Beijing, China, qiaosen@cea-igp.ac.cn
The stress change and its exact measurement of the fault in earthquake hazard zone has been a physical basis of understanding earthquake process. We carried out continuous monitoring across the Longmenshan range-front fault from June 2009 after the Wenchuan Ms8.0 earthquake, using ACROSS for generate highly repeatable seismic waves, and sweep frequency is from 2 to 10 Hz. The recording system consists of 9 Guralp-40T seismometers, its frequency response is from 2s to 100Hz, and RefTek-130B recorder and the 9 portable seismic stations deployed from near the source to 10km away. The time of the seismic source working and the recording system were accurately and continuously controlled by synchronization to the GPS clock.First, we carry out a comparative analysis with cross-correlation, frequency-domain sweep deconvolution, cross-coherence and short-window correlation. The overall effect of the coherent method and frequency-domain deconvolution method are slightly better than the cross-correlation and short-window correlation. We extracted the subtle relative change of P wave velocity processed by frequency-domain sweep deconvolution, filter, stack and interpolation in fixed travel time windows, showed that there is positive correlation between the P wave velocity change and barometric stress, and the relative decline of ~0.3% in Ms 5.6 coseismic effects.Compare the relative change with the result from ambient seismic noise, which both show obvious seasonal variation pattern at the magnitude about ~0.1%, with peak and trough at winter and summer respectively, and we also investigate the relationship between barometer pressure, temperature, precipitation and theoretical tidal strain, suggesting that the observed seasonality might be induced by medium crack density variation due to barometer pressure loading. Of course, the results presented here is rather preliminary, and long-term velocity variation and healing process of fault zone need to be analyzed later.
Session: Towards an Integrated Understanding of Slow Earthquakes: What We Know, What We Don’t Know, and How to Move Forward
Presenter   Bilek, Susan
Schedule   Fri 4:00 PM / Oral
Room   155A
What Can Earthquake Rupture Heterogeneity Tell Us About Subduction Zone Fault Heterogeneity?
BILEK, S. L., New Mexico Tech, Socorro, NM, sbilek@nmt.edu; DESHON, H. R., Southern Methodist University, Dallas, TX, hdeshon@mail.smu.edu; ENGDAHL, E. R., University of Colorado, Boulder, CO, bob.engdahl@gmail.com; EL HARIRI, M., BP, Houston, TX, mayaelhariri@gmail.com
The range of new observations of rupture timescales for subduction zone earthquakes has expanded greatly over the recent decade to include geodetically observed slow slip, non-volcanic tremor and low frequency earthquakes, in addition to tsunami earthquakes. Our understanding of what causes this wide spectrum of fault slip behaviors is still limited, but the specific frictional conditions along the megathrust likely contribute to this observed range of behavior. Earthquake source parameters can help probe these frictional conditions, allowing us to examine larger areas than possible through other means of sampling megathrust material. Here we present results of over 500 subduction zone earthquake source parameters, specifically the rupture duration and apparent stress, in order to investigate along-strike variation in material parameters at 11 subduction zones around the Pacific. Earthquakes range from magnitudes less than 4 along Central America to events with magnitudes of 5.5 to 8+ in other areas. We compare known slow-slip producing regions with those where we have not (yet) observed slow slip processes. We also incorporate observations about the varied geologic and tectonic environments in each region, including level of accretionary versus erosive processes, presence of subducting bathymetry, and sediment type and thickness. In many of these regions, we see heterogeneous along-strike patterns of rupture duration and apparent stress, but find very little or no correlation between the locations of long duration events and incoming sediment thickness or type. We find limited spatial correlation with regions of past tsunami earthquakes, regions of observed afterslip, and subducting bathymetric features. We find events with long duration characteristics in both shallow and deeper portions of the seismogenic zone, suggesting that heterogeneous patches of fault conditions can extend in both the along-strike and down-dip directions in the subduction zone.
Session: New Developments in Earthquake Forecasting and Predictability Research
Presenter   Segou, Margarita
Schedule   Thu 9:15 AM / Oral
Room   155D
Retrospective Physics-Based Earthquake Forecasting in the Corinth Gulf Continental Rift (Greece)
SEGOU, M., GeoAzur, Sophia-Antipolis, France, segou@geoazur.unice.fr; DESCHAMPS, A., GeoAzur, Sophia-Antipolis, France, deschamps@geoazur.fr
We perform a retrospective forecast experiment on the Aigio sequence (15/06/1995, M6.4) comparing the predictive power of physics-based and statistical models for short and long-term forecasting classes. We examine four predictive models, three based on the combination of Coulomb stress changes and rate-and-state theory (CRS) and one epidemic type aftershock sequence (ETAS) model. Even though Corinth gulf, the fastest continental rift in the world with extension rates 11-15 mm/yr, has been studied extensively in the past there is still today a debate whether seismicity is related with the existence of either a shallow dipping structure or typical steeply dipping normal faults. In the light of the above statement, two CRS realization are based on resolving Coulomb stress changes on specified receiver faults, expressing the aforementioned structural models, whereas the third CRS model uses optimally-oriented for failure planes. ETAS parameters are taken as the maximum likelihood estimates derived from stochastic declustering of the modern seismicity catalog (1964-1995) within our study area. We support a 0.043 bars/yr stressing rate for the Aigio fault derived from the aftershock duration of the 1995 event, suggesting that the fault’s activity expresses almost 50% of the geodetic extension rate. We implement likelihood tests to evaluate our forecasts for their spatial consistency and for the total amount of predicted versus observed events with M greater than 3.5 in 10-day time intervals. We find that (1) geology based CRS models are preferred over optimally oriented planes and (2) CRS models are consistent forecasters (60-70%) of transient seismicity, having in most cases comparable performance with ETAS models. We note the ambiguity, related with integrating geologic, geodetic and seismological observations into the rate-and-state dependent friction law, which is further addressed by introducing stochasticity in the physics-based models.
Session: Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Models
Presenter   Liao, Zonghu
Schedule   Wed AM / Poster
Room   Hall 1
STUDENT
Shear Rupture Along a Fault with an Experimentally-determined Frictional Strength
LIAO, Z., School of Geology and Geophysics, University of Oklahoma, Norman, OK, zonghuliao@ou.edu; RECHES, Z., School of Geology and Geophysics, University of Oklahoma, Norman, OK, reches@ou.edu
We simulate the dynamic shear rupture along a 2D vertical fault in an elastic half-space using a spectral boundary integral method. The fault has the frictional strength properties that were determined experimentally on granite samples at velocities approaching 1 m/s (Reches and Lockner, 2010). Recent experimental observations indicated that the steady-state frictional strength of silica-rich igneous rocks (granite, syenite, diorite) alternate between dynamic-weakening under low velocity (V < 0.03 m/s) and dynamic-strengthening under higher velocities (V > 0.03 m/s). This strength alternation was attributed to powder-lubrication (weakening), and powder dehydration (strengthening). The observed friction-distance-velocity relations of the experimental results for Sierra White granite were converted into an empirical friction model referred to as WEST (WEakening - STrengthening). The spectral element scheme is specially designed for anti-plane shear (mode III) fracture problems and allows for simulation of spontaneous propagation of faults. In the test problem, the WEST is used to express the strength of the material on the fault while keeping all other parameters (crust properties and stresses) the same as SCEC benchmark problem 3. This approach allow us the directly compare the WEST results with the slip-weakening model used by Rojas et al (2008). We found the following differences between the models: (1) WEST-based rupture occurs earlier at the observation point away from the nucleation zone; (2) WEST-based model has lower (~ 35%) peak velocity and shorter rise-time outside the nucleation zone; and (3) WEST-based rupture shows rich vibration of slip velocity related to dynamic-weakening and dynamic-strengthening throughout the entire fault. Consequently, a more dynamic procedure is simulated including stress drop, displacements, and friction recovery. We demonstrate the importance of incorporating experimental-based friction model to the rupture models.
Session: Seismic Hazards and Ground Motions
Presenter   Denolle, Marine
Schedule   Fri 3:45 PM / Oral
Room   155B
STUDENT
Seismic Amplification in Los Angeles for M7+ Scenarios on the San Andreas Fault Using the Virtual Earthquake Approach
DENOLLE, M. A. M., Stanford University, Stanford, CA, mdenolle@stanford.edu; DUNHAM, E. M., Stanford University, Stanford, CA, dunham@stanford.edu; PRIETO, G., Universidad de los Andes, Bogota, Columbia, gprieto@uniandes.edu.co; BEROZA, G. C., Stanford University, Stanford, CA, beroza@stanford.edu
Predicting accurately the ground motion for scenario earthquakes is critically important for urban areas that are subject to high seismic hazard. Emprical ground motion prediction equations estimate intensity of shaking based on observed measurements from past earthquakes, and are limited by a shortage of data in the very near field of large earthquakes. Physics-based approaches simulate the ground shaking assuming a source model and a 3D crustal structure to propagate seismic waves. They are limited by the uncertainties in modeling assumptions and our incomplete knowledge of crustal structure. We propose to use the ambient seismic field to validate the long-period ground shaking predicted from simulation.We compute the Earth impulse response from the ambient seismic field and coda waves using deconvolution on continuous data. In the virtual earthquake approach we correct the derived impulse responses from surface point source to that of a buried double-couple source. We sum the point sources of seismic stations deployed along the San Andreas Fault, according to the representation theorem, to model extended earthquake rupture for large M 7+ scenario earthquakes on the southern San Andreas Fault. This approach allows us to explore the coupling between source directivity and amplification due to sedimentary basin structure in Los Angeles area that has emerged from physics-based ground motion simulations.
Session: Earthquake Source Physics
Presenter   Zhang, Zhenguo
Schedule   Thu PM / Poster
Room   Hall 1
STUDENT
Dynamic Rupture Simulation by Curved Grid Finite-Difference Method
ZHANG, Z., University of Science and Technology of China, Hefei, Anhui, China, zgzhang7@mail.ustc.edu.cn; CHEN, X., University of Science and Technology of China, Hefei, Anhui, China, xfchen1@ustc.edu.cn
Numerical methods have been used widely to investigate the earthquake rupture dynamics. Among those methods, finite-difference method (FDM) has its own accuracy and efficiency. However, limitation of the FDM to simulate the spontaneous rupture dynamics is also obvious, i.e., it can neither deal with the arbitrarily orientated fault plane, nor the arbitrary non-planar fault if the split-node is used to deal with the displacement (or velocity) discontinuity across the fault. That is because of the use of Cartesian coordinate.To overcome those limitations of the FDM in the simulation of earthquake rupture dynamics, we adopt the Curved Grid FDM (Zhang & Chen, 2006; Zhang et al., 2012) to model the rupture dynamics of an arbitrarily orientated or arbitrarily non-planar fault in the arbitrary curvilinear coordinate system. Our CG-FDM has been proved its advantage in dealing with irregular free surface in the seismic wave simulation. Here, we treat the non-planar discontinuous fault as special non-planar interface which can be conformed by choosing the interface conforming coordinate (a distinct arbitrary curvilinear coordinate), i.e., constructing discrete grid along the fault plane. We also use the split-node method in dealing with the discontinuity of fault.With the CG-FDM, we can simulate most of the complex dynamic faulting models, including the non-planar dipping fault and striking fault, and the fault intersects with the surface, etc. Some standard benchmark problems are modeled and compared to validate the accuracy of our CG-FDM, as well as other nonplanar simulations to show the capability of this method.
Session: Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Models
Presenter   Mai, P. Martin
Schedule   Thu 4:45 PM / Oral
Room   155C
The Source Inversion Validation (SIV) Project: Uncertainty Quantification in Earthquake Source Inversions
MAI, P. M., King Abdullah University of Science and Technology, Thuwal, Saudi Arabia, martin.mai@kaust.edu.sa
Finite-fault source inversions estimate kinematic rupture parameters of earthquakes using a variety of available data sets and inversion approaches. Rupture models are obtained by solving an inherently ill-posed inverse problem, subject to numerous a priori assumptions and noisy observations. Despite these limitations, near real-time source inversions are becoming increasingly popular, while we still face the dilemma that uncertainties in source inversions are essentially unknown. Yet, the accurate estimation of earthquake rupture properties, including proper uncertainty quantification, is critically important for earthquake seismology and seismic hazard analysis, as they help to adequately characterize earthquake complexity across all scales.The “Source Inversion Validation” (SIV) project (http://equake-rc.info/sivdb/wiki), a collaborative international multi-institutional effort, attempts to quantify the intra-event variability in rupture models (see for example the SRCMOD database, http://equake-rc.info/srcmod), and to propose robust uncertainty metrics for earthquake source inversions. The SIV efforts include a rigorous testing platform to examine the current state-of-the-art in earthquake source inversion (http://equake-rc.info/sivdb), and to develop and test novel source inversion approaches. In this presentation, we will summarize initial SIV results related to previous benchmark exercises, discuss the latest findings for a test case of a complex rupture embedded in a 3D heterogeneous Earth model, and propose metrics to quantify rupture-model variability, quality of data fitting, and model robustness.
Session: Next Generation of Ground Motion Prediction Models
Presenter   Scherbaum, Frank
Schedule   Thu AM / Poster
Room   Hall 1
The Adjustment of GMPEs for Differences in Source,- Propagation- and Site Effects: Problems and Solutions
SCHERBAUM, F., University of Potsdam, Potsdam, Germany, fs@geo.uni-potsdam.de; SINGH BORA, S., University of Potsdam, Potsdam, Germany, Sanjay.Singh@geo.uni-potsdam.de; KUEHN, N., University of Potsdam, Potsdam, Germany, nico@geo.uni-potsdam.de; STAFFORD, P., Imperial College London, London, UK, p.stafford@imperial.ac.uk
One of the major challenges related with the current practice in seismic hazard studies is the adjustment of empirical ground motion prediction equations (GMPEs) to different seismological environments. The Hybrid Empirical approach proposed by Campbell (2003), in which ratios of stochastic or theoretical ground motion estimates are used to adjust empirical ground-motion prediction equations developed for one region for use in another region, has become a popular framework for this purpose. Although conceptually appealing and seemingly simple, its implementation in practice can lead to rather subtle technical challenges, in particular with respect to "transferring" the high frequency spectral signal content from one region to another. The key issue in this context is that, loosely speaking, Fourier spectra and Response spectra are different "beasts" and this difference needs to be taken into account when generating the adjustment factors for the Hybrid Empirical approach. For oscillator frequencies within the signal passband of the Fourier spectrum, response spectra behave similar to Fourier spectra when it comes to adjustment issues while for oscillator frequencies outside the signal passband (e. g. for PGA) this is no longer the case. In this presentation we demonstrate that this has considerable consequences for the adjustment of the high frequency response spectral shape but also for the magnitude- and distance dependence of the response spectral adjustment factors. We discuss the performance of Inverse Random Vibration Theory (IRVT) based solutions in which the the response spectral adjustment factors are calculated for a single magnitude-distance scenario and finally propose a general solution based on a combination of Fourier spectral and duration models.
Session: Next Generation of Ground Motion Prediction Models
Presenter   Singh Bora, Sanjay
Schedule   Thu AM / Poster
Room   Hall 1
STUDENT
Fourier Spectral- and Duration Models for the Generation of Response Spectra which are Adjustable to Different Source-, Propagation-, and Site Effects
SINGH BORA, S., University of Potsdam, Potsdam, Germany, Sanjay.Singh@geo.uni-potsdam.de; SCHERBAUM, F., University of Potsdam, Potsdam, Germany, fs@geo.uni-potsdam.de; KUEHN, N., University of Potsdam, Potsdam, Germany, nico@geo.uni-potsdam.de; STAFFORD, P., Imperial College London, London, UK, p.stafford@imperial.ac.uk
We explore a new approach for the development of response spectral Ground Motion Prediction Equations (GMPEs), which can easily be adjusted to different seismological environments. We believe that the key to accommodating differences in regional seismological attributes of a ground motion model lies in the Fourier spectrum. Our approach consists of developing empirical prediction equations for Fourier spectra and for a particular duration estimate of ground motion which is tuned to optimize the fit between spectra obtained through the random vibration theory (RVT) framework and the classical way. The presented analysis for the development of GMPEs is performed on the recently compiled RESORCE strong motion database across Europe, the Mediterranean, and the Middle East. Although the main motivation for the presented approach is the adjustability and the use of the corresponding model to generate data driven host-to-target conversions, even as a standalone response spectral model it compares reasonably well with the GMPEs of Ambraseys et al. (2005), Akkar and Bommer, (2010), and Akkar and Cagnan (2010).
Session: Next Generation of Ground Motion Prediction Models
Presenter   Scherbaum, Frank
Schedule   Thu AM / Poster
Room   Hall 1
A Study of the Sensitivity of Response Spectral Amplitudes on Seismological Parameters using Algorithmic Differentiation
MOLKENTHIN, C., University of Potsdam, Potsdam, Germany, molkenthin@geo.uni-potsdam.de; SCHERBAUM, F., University of Potsdam, Potsdam, Germany, fs@geo.uni-potsdam.de; KUEHN, N., University of Potsdam, Potsdam, Germany, nico@geo.uni-potsdam.de; OHRNBERGER, M., University of Potsdam, Potsdam, Germany, mao@geo.uni-potsdam.de
Response spectra are of fundamental importance in earthquake engineering and seismic design. The concept of response spectral shapes is often part of seismic building codes. The relationship of response spectral amplitudes to seismological source-, path-, and site characteristics, however, is, in contrast to Fourier spectral amplitudes, not immediately obvious and might even be considered counter-intuitive for high oscillator frequencies. This notwithstanding its understanding is crucial for many aspects of seismic hazard analysis. The purpose of the present study is the characterization of response spectral amplitude variations due to perturbations in seismological parameters. This is done by calculating the absolute parameter sensitivities, defined as the partial derivatives of the model output with respect to its input parameters. We perform a local first order sensitivity analysis of response spectra simulated by a stochastic ground-motion model to assess the effects and importance of input parameter perturbations on the shape of response spectra in a quantitative and systematic way. To derive sensitivities, we use algorithmic differentiation (AD). This powerful approach is extensively used for sensitivity analysis of complex models in meteorology or aerodynamics but to our knowledge not yet explored in the seismic hazard context. The main advantages of the proposed method are the exact quantitative estimates of the first order sensitivities with low computational costs. In addition, AD can be applied directly on existing computer code. Within the present study, AD was successfully implemented for the stochastic ground-motion model code SMSIM (Boore, 2003) using the Tapenade AD tool. In order to analyze the influence of individual seismological parameters on the response spectral amplitudes we have derived absolute and normalized sensitivity coefficients for different seismological parameters and for different regional models (ENA, WENA).
Session: Including Ground Failure in Scenario Events, Rapid Response, and Loss Estimation Models
Presenter   Wald, David
Schedule   Wed 5:30 PM / Oral
Room   155B
Incorporating Near-Real-Time Landslide and Liquefaction Estimates in ShakeCast and PAGER
WALD, D. J., U.S. Geological Survey, Golden, CO, wald@usgs.gov; HEARNE, M., U.S. Geological Survey, Golden, CO, mhearne@usgs.gov; THOMPSON, E. M., Tufts University, Medford, MA, eric.thompson@tufts.edu; NOWICKI, M. A., Indiana University, Bloomington, IN, mnowicki@usgs.gov; LIN, K., U.S. Geological Survey, Golden, CO, klin@usgs.gov; JOHNSON, K., U.S. Geological Survey, Golden, CO, kljohnson@usgs.gov; GODT, J., U.S. Geological Survey, Golden, CO, jgodt@usgs.gov; KNUDSEN, K., U.S. Geological Survey, Menlo Park, CA, kknudsen@usgs.gov; ZHU, J., Tufts University, Medford, MA, jing.zhu@tufts.edu; BAISE, L., Tufts University, Medford, MA, laurie.baise@tufts.edu
The human consequences of earthquakes are often a result of subsequent ground-failure– related hazards, e.g., landslides, liquefaction, and lateral spreading. Although the U. S. Geological Survey (USGS) ShakeCast, ShakeMap, and Prompt Assessment of Global Earthquakes for Response (PAGER) systems are capable of estimating losses in near- real time from the shaking-based hazard, they do not currently provide impact estimates due to these ground failure (or secondary) hazards. Colleagues in this session have proposed statistical models for estimating liquefaction and landslide distributions that make use of standardized ShakeMap outputs, and have calibrated the models against historical ground deformation case histories. The historical ShakeMaps used in developing these models are provided as part of an updated release of the ShakeMap Atlas. We have implemented and are testing the reliability of these preliminary landslide and liquefaction models for use in ShakeCast and PAGER. Globally available susceptibility variables include topographic slope, geology, Vs30, distance to rivers, and compound topographic index (CTI). Peak ground acceleration from ShakeMap is, to date, the chosen shaking-related predictor variable. We present an initial approach for visualizing secondary hazards and thresholds that indicate their overall severity (to be used in alerting). We also describe susceptibility maps, a useful byproduct of our secondary hazard modeling approach, which could be further used in risk analyses. An ongoing challenge is how to accommodate small- scale maps using global proxy data sets as predictor variables for PAGER maps with aggregate hazard and loss estimates, yet employ more accurate methods when site- specific geotechnical data are available (e.g., in ShakeCast, which treats losses at specific facilities). Making near–real-time aggregate loss estimates from these secondary hazard input maps will require considerable further research and development.
Session: ShakeMap-Related Research, Development, Operations, and Applications
Presenter   Mah, Russell
Schedule   Thu PM / Poster
Room   Hall 1
STUDENT
Using ShakeMap to Synthesize an Observation-Based Seismic Hazard Map of the United States
MAH, R. T., U.S. Geological Survey, Golden, CO, rmah@usgs.gov; GARCIA, D., U.S. Geological Survey, Golden, CO, garciajimenez.d@gmail.com; DEWEY, J. W., U.S. Geological Survey, Golden, CO, dewey@usgs.gov; WALD, D. J., U.S. Geological Survey, Golden, CO, wald@usgs.gov
ShakeMap has become a standard representation of the shaking caused by an earthquake. Its ability to quickly combine source parameters with instrumental measurements and intensity observations to create a map of the shaking level of an earthquake has proven invaluable to institutions around the world. We expand ShakeMap from a single-event mapping tool to an areal seismic-hazard mapping tool, by creating peak ground motion and intensity maps across the United States (US) for a period of over one hundred years. The maps are generated from source data, macroseismic intensity observations, and strong motion records from several US earthquake databases and collections, the presentation of which we standardize both temporally and spatially. We discuss earthquake selection, temporal and spatial distribution of events, depth and other approximations for old events lacking parameters, and selections of prediction and conversion equations used by ShakeMap. Products include (1) a database of reported felt earthquakes and their observed intensities across the US for 1900-2012 (2) a database of ShakeMaps for the same events, and (3) corresponding summary peak intensity and ground motion maps for the entire US. We expect that the summary peak intensity and ground-motion maps derived via the ShakeMap procedure will provide different perspectives on seismic hazard than those provided either by simple summaries of historically observed intensities or by methodologies of extrapolating the historical and geological record that are used in probabilistic seismic hazard mapping. For example, we will be able to address previously unanswerable questions like: “How many times has location X been shaken at intensity Y?” and related queries.
Session: Networks and Instrumentation
Presenter   Thomas, Gavin
Schedule   Wed AM / Poster
Room   Hall 1
STUDENT
A Model of Ambient Seismic Noise Recorded by the Utah Regional Network of Strong-Motion Seismometers
THOMAS, G. L., University of Wyoming, Laramie, WY, gthomas8@uwyo.edu; KOPER, K. D., University of Utah, Salt Lake City, UT, koper@seis.utah.edu; BURLACU, R., University of Utah, Salt Lake City, UT, burlacu@seis.utah.edu; DROBECK, D., University of Utah, Salt Lake City, UT, drobeck@seis.utah.edu
A comprehensive analysis was performed on continuous seismic data recorded for the entire year of 2011 by the permanent strong motion seismometers of the Utah Regional Seismic Network (URSN). As of December 2011, there were 126 such stations with 378 channels (ENE, ENN, and ENZ) in the URSN. Over 2.2 million power spectral densities (PSDs) were estimated for individual hour-long segments of instrument corrected data using a standard sub-windowing approach. The PSDs were log-averaged and combined into probability density functions (PDFs) to visualize the noise properties of individual channels, and groups of channels defined by site classification, instrument type, and component of motion. A single summary PDF was also constructed, and smoothed versions of the 2.5 and 97.5 percentiles were adopted as a reference model for strong motion noise in Utah. Evaluation of strong-motion instruments in Utah using the new reference model shows clear variations in noise level according to instrumentation (both sensor and recorder) and site characteristics (as defined by VS30 values). Although the University of Utah Seismograph Stations (UUSS) preferentially installs higher-quality instrumentation at quieter sites, instrument specific problems can be identified. Clear temporal variations in the background noise are also evident. The strongest variation occurs only at short seismic periods (~ 0.1 s) with a weekly periodicity and is related to cultural activities that generate more noise during the normal work-week (Monday-Friday) than on the weekends (Saturday and Sunday). The next strongest variation occurs at daily periodicities over a wider range of seismic periods (at least 0.1-10 s), likely reflecting natural activities (day/night variations in atmospheric pressure) as well as cultural activities (day/night variations in traffic, construction, etc.)
Session: Data Products as Research Resources
Presenter   Thingbaijam, Kiran Kumar Singh
Schedule   Fri AM / Poster
Room   Hall 1
The eQuake-RC Platform: Resources for Earthquake Source Studies
THINGBAIJAM, K. K. S., King Abdullah University of Science and Technology, Thuwal, Jeddah, Kingdom of Saudi Arabia, k.thingbaijam@kaust.edu.sa; MAI, P. M., King Abdullah University of Science and Technology, Thuwal, Jeddah, Kingdom of Saudi Arabia, martin.mai@kaust.edu.sa
We present an online platform – eQuake-RC (http://equake-rc.info) – that provides a common framework for collaborative research on earthquake source modeling. The three intertwined missions are (1) collect and disseminate earthquake SouRCe MODels (SRCMOD), (2) investigate uncertainty in earthquake Source Inversions through verification and Validation experiments (SIV), and (3) distribute software Codes for Earthquake Rupture and ground-motion Simulation (CERS). The SRCMOD database is a unified repository of rupture models for past earthquakes. The SIV Benchmarks provide a comprehensive datasets to facilitate verification and validation experiments to develop, test, and calibrate source inversion methods, and to develop strategies for rigorous uncertainty quantification in source inversions. An SIV Wiki serves as a depository for information on the SIV Benchmarks. The CERS-Software currently distributes three packages namely RupGen, Stress2Slip, and BB-Simulation. RupGen, in short for Rupture Generator, is a set of codes useful for generating realistic finite-fault earthquake rupture models. Slip2Stress computes the on-fault static stress changes corresponding to slip-distribution on the fault-plane. BB-Simulation can be used to compute and integrate high frequency synthetics to low frequency seismograms for generating hybrid broadband near-field waveforms. The e-Quake-RC platform is envisaged to be dynamic evolving to meet any new requirements while facilitating collaborative research and dissemination of resources pertinent to earthquake sources.
Session: Earthquake Source Physics
Presenter   Ma, Shuo
Schedule   Thu 11:00 AM / Oral
Room   155A
Dynamic Wedge Failure as a Unifying Mechanism for Large Tsunamigenesis and Anomalous Energy Radiation of Shallow Subduction Earthquakes
MA, S., San Diego State University, San Diego, CA, sma@mail.sdsu.edu; HIRAKAWA, E. T., San Diego State University, San Diego, CA, evan.hirakawa@gmail.com
The shallow reaches of subduction interfaces (upper 10 – 15 km) host earthquakes of anomalous energy release attributes. Earthquake ruptures there can be exceptionally tsunamigenic, but produce weak high-frequency ground motion. Numerous observations indicate that they are also associated with unusually long rupture duration, slow rupture velocity, small stress drop, as well as low moment-scaled radiated energy. What gives rise to these anomalous characteristics and how they relate to large tsunamigenesis are, however, still not well understood. Motivated by the critical taper theory for accretionary wedges (e.g., Davis et al., 1983; Dahlen, 1990), Ma (2012) showed that for a wedge on the verge of failure, pore pressure increase due to updip rupture causes extensive Coulomb failure within the wedge, which gives rise to slow rupture velocity and significant seafloor uplift landward from the trench. Here we extend this work and further show that the inelastic seafloor uplift strongly dilates the shallow-dipping basal fault behind the rupture front, which is greatly enhanced by the presence of free surface. The dilation reduces the effective normal stress and sliding friction on the fault, and increases the dynamic stress drop and slip velocity, such that slip-velocity time histories in the shallow section of the fault have a ‘snail-like’ shape, leading to long rise times, increased smoothness in the slip-velocity field, and depletion of high frequencies in the resultant source time function. The failure in the wedge also acts as a large energy sink (while contributing to seismic moment), giving rise to distributed heat generation, low moment-scaled radiated energy and small rupture directivity, which provides a unifying interpretation for nearly all anomalous observations documented for shallow subduction earthquakes. This wedge-failure mechanism presents an alternative explanation for the devastating Tohoku tsunami without requiring large slip at the trench.
Session: Earthquake Source Physics
Presenter   Ma, Shuo
Schedule   Thu PM / Poster
Room   Hall 1
Biases in the Coseismic Slip Models of Shallow Subduction Earthquakes Induced by Using Elastic Green’s Functions
YAO, Q., San Diego State University & University of California, San Diego, San Diego, CA, q1yao@ucsd.edu; MA, S., San Diego State University, San Diego, CA, sma@mail.sdsu.edu
Elastic dislocation theory has been widely used to infer earthquake rupture processes in the shallow subduction zone. However, the assumption that materials deform elastically during the rupture process is questionable. In the classic critical taper theory of accretionary wedges (Davis et al., 1983; Dahlen, 1990) the overriding wedge is on the verge of Coulomb failure everywhere. Due to the proximity to failure updip rupture on a shallow-dipping plate interface can significantly increase the pore pressure, lowering the wedge strength and leading to widespread Coulomb failure in the wedge, which is greatly enhanced by the shallow dip of the fault (Ma, 2012). The extensive failure in the wedge gives rise to slow rupture velocity and large seafloor uplift. It also leads to the deficiency of high-frequency seismic radiation and low moment-scaled radiated energy (Ma and Hirakawa, this meeting), which has been documented for shallow subduction zone earthquakes (especially tsunami earthquakes) for four decades.The displacement field calculated by incorporating this poroplastic mechanism is distinctly different from the one predicted by an elastic model – the slip near the trench is small and the largest seafloor uplift is landward from the trench. If this mechanism operates during the rupture process how biased can the inverse models by using elastic Green’s functions be? Could the large inferred slip near the trench for the 2011 Tohoku earthquake be manifestations of extensive failure process in the wedge? In this work, we will use elastic Green’s functions to invert the numerically simulated seafloor displacement field by incorporating this poroplastic mechanism and discuss these issues.
Session: ShakeMap-Related Research, Development, Operations, and Applications
Presenter   Madin, Ian
Schedule   Thu PM / Poster
Room   Hall 1
Cascadia M 9.0 Scenario Ground Motion and Ground Deformation Maps for Oregon
MADIN, I. P., Oregon Department of Geology and Mineral Industries, Portland, OR, ian.madin@dogami.state.or.us; BURNS, W. J., Oregon Department of Geology and Mineral Industries, Portland, OR, bill.burns@dogami.state.or.us
In support of a legislatively-mandated state Cascadia earthquake resilience plan for Oregon, we prepared a series of ground motion and ground deformation maps for a scenario M 9.0 earthquake. The maps were based on new NEHRP site class, liquefaction susceptibility and landslide susceptibility maps that we prepared using the statewide digital geologic database (OGDC5). The new site condition maps provide much greater resolution of important geologic features floodplains and existing landslides than any previous maps. Site condition data was gridded at 30 m cells and used to calculate new PGA, PGV, landslide probability and displacement maps, and liquefaction probability and displacement maps at the same resolution.We used bedrock shaking values from the model used for the USGS Cascadia M 9.0 Shakemap scenario (http://earthquake.usgs.gov/eqcenter/shakemap/global/shake/Casc9.0_expanded_se/) as input for the PGA and PGV calculations and calculated site amplification using equations from Boore and Atkinson (2008). The maps show broad patterns similar to the site condition maps provided in the USGS Cascadia Scenario, but with better resolution and representation of critical high hazard areas defined by site geology.We used the methods described in the HAZUS-MH MR4 manual to calculate the landslide and liquefaction maps. The maps depict high to extreme liquefaction probability and lateral spread displacements extending to the Oregon Cascade range. The landslide hazard maps show severe landslide probability and displacement for all of the Oregon Coast Range, with high hazard extending into the Cascade Range. Recent lidar based landslide inventory of two basins in the Coast Range shows that 22-35% of each basin’s area is an existing landslide, which suggests that the modeled hazard values are reasonable.
Session: Triggering of Seismic and Volcanic Events
Presenter   Parsons (Video), Tom
Schedule   Fri 10:45 AM / Oral
Room   155D
The Global Aftershock Zone
PARSONS, T., USGS, Menlo Park, CA, tparsons@usgs.gov
There is little doubt that the whole planet becomes the aftershock zone of large (M≥7) earthquakes. Surface waves distort fault zones and volcanic centers as they pass through the crust, leading to seismic failures. From a hazard perspective we are obviously concerned that this dynamic process might encourage high magnitude earthquakes. Great strides have been made in operational earthquake forecasting near mainshocks in time and space, but we are far less certain how to assess the hazard posed at global distances. Results from global compilations demonstrate significant rate increases during, and immediately after (~45 minutes) M>7.0 mainshocks in all tectonic settings and ranges. However, it is difficult to find strong evidence for M>5 rate increases during the passage of surface waves in combined global catalogs. On the other hand, studies of individual large mainshocks show convincing occurrences of M>5 triggering at global range that are often delayed by a few hours to days after surface wave arrivals. These examples tend to lie in the noise when the global catalogs are examined simultaneously, which implies that they are relatively rare events. However, if large triggered earthquakes can occur in the global aftershock zone, then we must be concerned about how to calculate and convey the hazard they pose. Results from physical modeling of stresses imposed by surface waves, and from comparative responses of individual regions to hundreds of M>7 mainshocks give us some preliminary insights into the likelihood of damaging global aftershocks. One recommendation would be to monitor local networks for rate increases during surface wave arrivals, with the assumption that a vigorous response enhances the probability of large aftershocks.
Session: Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Models
Presenter   Graizer, Vladimir
Schedule   WITHDRAWN
Room   155C
North Anna Nuclear Power Plant Strong Motion Records of the Mineral, Virginia Earthquake of August 23, 2011
GRAIZER, V., U.S. Nuclear Regulatory Commission, Washington, DC, Vladimir.Graizer@nrc.gov; MUNSON, C., U.S. Nuclear Regulatory Commission, Washington, DC; LI, Y., U.S. Nuclear Regulatory Commission, Washington, DC
The Mw 5.8 Mineral, Virginia earthquake was recorded at an epicentral distance of about 18 km at the North Anna Nuclear Power Plant (NPP) by the magnetic tape digital accelerographs installed inside the plant’s containment at the foundation and deck levels. The North Anna NPP is operated by the Virginia Electric and Power Company (VEPCO) commonly called Dominion Power Company and has two pressurized water reactors (PWR) units that began operation in 1978 and 1980, respectively. Following the earthquake, both units were safely shutdown. The strong-motion records were processed to get velocity, displacement, Fourier amplitude and 5% damped response spectra. The basemat record demonstrated relatively high amplitudes of acceleration of 0.26 g and velocity of 13.8 cm/sec with a relatively short duration of strong motion of 2-3 sec. The higher elevation containment deck record had peak acceleration of 0.38 g and peak velocity of 26.1 cm/sec. These accelerograms are the strongest ground motions ever recorded at a US nuclear power plant. Basemat Response Spectra exceed Design Basis Earthquake for the existing Units 1 and 2, while comprehensive plant inspections performed by Dominion and U.S. Nuclear Regulatory Commission have concluded that the damage to the plant was minimal and not affecting any structures and equipment significant to plant operation. This can be explained in part by short duration of the earthquake ground motion at the plant. The North Anna NPP did not have free-field strong motion instrumentation at the time of the earthquake. Since the containment is founded on rock there is a tendency to consider the basemat record as an approximation of the free-field recording. However, comparisons of deck and basemat records demonstrate that the basemat recording is also affected by structural resonances in the frequency range of 3-4 Hz and higher. Therefore future seismological interpretation of these recordings should take into account the effect of structure.
Session: Infrasound and Seismoacoustics
Presenter   Bonner [video], Jessie
Schedule   Wed 2:15 PM / Oral
Room   155D
Seismic and Acoustic Signal Generation and Propagation at Local Distances: New Datasets from Surface and Shallow Explosions
BONNER, J., Weston Geophysical Corp., Lufkin, TX, jes_bonner@westongeo.com; WAXLER, R., National Center for Physical Acoustics, University, MS, rwax@olemiss.edu; REINKE, R., DTRA, Albuquerque, NM, robert.reinke@dtra.mil; LENOX, E., DTRA, Albuquerque, NM, elizabeth.lenox@dtra.mil; COLE, P., DTRA, Ft. Belvoir, VA, phillip.cole@dtra.mil
To better understand uncertainties of combined seismo-acoustic yield estimates, we have collected new datasets for 26 surface/shallow explosions in Israel, Massachusetts, and New Mexico in 2011/2012. We have recorded 1434 overpressure and 2876 seismic observations in the 0-50 km distance range from these explosions. The datasets have been complemented with precise origin time monitoring, accurate GPS measurements for both source and receivers, and extensive atmospheric profiling (temperature, wind, etc) prior to or during each explosion. We will present the initial observations and conclusions from these datasets, which include the strong effects that weather, source directivity, and topography have on overpressure generation and propagation. For example, strong winds (>10 m/s) can alter overpressure signals as large as 3000 Pa at distances less than 1 km. For seismic signals, we observe significant differences in coupling between above ground explosions in different rock types. For example, we have observed that surface explosions in alluvium will produce seismic amplitudes that are only 3-5x smaller than a fully coupled explosion in the same alluvium; however, for limestone, the difference between surface and fully coupled is significantly larger (15-30x). We will also show how depth of burial/height of burst changes the seismo-acoustic signals generated from explosions.
Session: Triggering of Seismic and Volcanic Events
Presenter   Hardebeck (Video), Jeanne
Schedule   Fri 9:15 AM / Oral
Room   155D
Insight into Static and Dynamic Stress Triggering from Preshock and Aftershock Focal Mechanisms
HARDEBECK, J. L., U.S. Geological Survey, Menlo Park, CA, jhardebeck@usgs.gov
The focal mechanisms of earthquakes in southern California before and after four M≥6.7 mainshocks, 1992-2010, provide insight into how fault systems respond to changes in stress. Aftershocks in distance ranges experiencing a 0.02-5 MPa absolute static stress change have focal mechanisms significantly more consistent with the modeled static stress changes than the preshock mechanisms, as measured by the fit of the rake to the predicted rake from the stress change. Static stress changes ≥0.02 MPa therefore play a role in earthquake triggering. The mechanism sensitivity to the stress change decays with time, over 0.5 year following Hector Mine to ~20 years following Landers and Northridge. Aftershocks preferentially occur in locations where the mainshock static stress change and the background stress have similar orientation. In these areas, the distribution of aftershock focal mechanisms is indistinguishable from the distribution of preshock mechanisms. In regions with dissimilar background stress and stress change, most of the aftershock mechanisms also have a similar distribution as the preshock mechanisms. However, in some of these regions, there is an additional group of aftershock mechanisms that are not common in the preshocks. The new mechanisms correspond to fault planes with relatively high Coulomb stress change and with similar predicted rake directions with respect to the background stress and the stress change. Only very near the mainshock rupture does the aftershock mechanism distribution rotate substantially relative to the preshock distribution or become substantially more heterogeneous.I find little consistency between aftershock mechanisms and the orientations of dynamic stresses estimated from waveforms. This implies that if dynamic stress plays a role in delayed triggering, it functions primarily through changes in fluid pressure or through physical changes to the fault zone, rather than through dynamic Coulomb stress changes.
Session: Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Models
Presenter   Mohammadi, Kami
Schedule   Fri 1:30 PM / Oral
Room   155C
STUDENT
Site-Specific Topography Amplification Factors at Southern California Strong Motion Stations
MOHAMMADI, K., School of Civil and Env. Eng., Georgia Institute of Technology, Atlanta, GA, kamimohamadi@gatech.edu; ASSIMAKI, D., School of Civil and Env. Eng., Georgia Institute of Technology, Atlanta, GA, dominic@gatech.edu
We study the effects of surface topography on the aggravation of seismic motion in the vicinity of two strong motion station sites at Southern California (SC). Simulations are conducted for idealized 3D surface topography models and for both homogeneous and stratified soil formations, and results reveal the additive effects of topography and soil response and the role of ground motion characteristics in site amplification in the vicinity of irregular topographic features. Results are also compared with corresponding 2D representations of the features to highlight the role of 3D effects. The most distinguishing characteristic of the two sites is the underlying soil stratigraphy. For the homogeneous site with Class A shear wave velocity, simulated amplification is attributed to the surface geometry alone. For the layered site with 3 different shear wave velocities chara¬cterized by a stiffness contrast of approximately 10, the observed ampli¬fication cannot be explained by the effects of topography or stratigraphy alone: coupling of the two effects leads to amplification patterns that can be underestimated or overestimated by a homogeneous topography model assumption, depending on the site response characteristics and ground motion frequency content. A parametric study is conducted to investigate the role of surficial soil layers for a wide range of incident frequencies represented by isolated or trains of Ricker wavelets of various central frequencies. Synthetic seismograms and spatially distributed peak ground response data normalized by the corre¬sponding free-field response in time and frequency are presented to illustrate the relative contri¬bution of site and topography effects in the overall ground motion amplification. Future work efforts include the development of an amplification database for strong motion stations in SC, which includes topography and soil effects and can be implemented in validation exercises of broadband ground motion simulations.
Session: Characterizing Active Faults for Seismic Hazard Assessments
Presenter   Feenstra, Jessica
Schedule   Fri AM / Poster
Room   Hall 1
STUDENT
Microseismicity and Three-dimensional Tomography of the Alpine Fault, New Zealand
FEENSTRA, J. P., University of Wisconsin, Madison, WI, jpfeens@geology.wisc.edu; THURBER, C. H., University of Wisconsin, Madison, WI, clifft@geology.wisc.edu; ROECKER, S. W., Rensselaer Polytechnic Institute, Troy, NY, roecks@rpi.edu; RAWLES, C., University of Wisconsin, Madison, WI, rawles@wisc.edu; LORD, N., University of Wisconsin, Madison, WI, lord@geology.wisc.edu; TOWNEND, J., Victoria University, Wellington, New Zealand, John.Townend@vuw.ac.nz; BANNISTER, S. C., GNS Science, Te Pü Ao, Lower Hutt, New Zealand, S.Bannister@gns.cri.nz
The UW-Madison-RPI WIZARD array was deployed along the Alpine Fault, South Island, New Zealand, with the primary goal of monitoring background seismicity and imaging the crustal structure around the Deep Fault Drilling Project (DFDP) site. Seismicity recorded on the WIZARD array is combined with data from two neighboring arrays, SAMBA (Victoria University of Wellington) to the southwest and ALFA'12 (GNS Science) to the northeast, along with several GeoNet permanent stations, to provide broad coverage of a ~150 km-long section of the fault. This extensive dataset will help delineate the geometry of the Alpine Fault and other nearby active faults at depth. In addition, the station coverage allows for regional three-dimensional tomography of the area surrounding the Alpine Fault. Our preliminary results indicate that the majority of the seismicity occurs in the hanging wall, as previous studies have found, but a number of events have been identified that may be on the Alpine Fault. We will determine focal mechanisms for these events to see if they are consistent with the oblique left-lateral/reverse motion. Northeast of the drill site, the hanging wall rocks are systematically faster and there is a well-defined velocity contrast near the Alpine Fault. Southwest of the drill site, the footwall rocks are generally faster, and the fault is not as evident in the velocity structure.
Session: Next Generation of Ground Motion Prediction Models
Presenter   Gallegos, Andrea
Schedule   Wed 3:15 PM / Oral
Room   155C
STUDENT
Lg Attenuation in the Central and Eastern United States Revealed by the EarthScope Transportable Array
GALLEGOS, A. C., New Mexico State University, Las Cruces, NM, agall@nmsu.edu; RANASINGHE, N. R., New Mexico State University, Las Cruces, NM, nrana001@nmsu.edu; TRUJILLO, A. R., New Mexico State University, Las Cruces, NM, andreat@nmsu.edu; NI, J. F., New Mexico State University, Las Cruces, NM, jni@nmsu.edu; SANDVOL, E., University of Missouri-Columbia, Columbia, MO, sandvole@missouri.edu
Data from EarthScope's Transportable Array Network (TA) were used to generate a new and more comprehensive attenuation map of the regional Lg phase for the central and eastern United States (CEUS). The two-station method, which eliminates source effects, was used to measure interstation Q. Regional Lg phases generated by 39 events recorded from 2010 to 2012 resulted in 76,937 interstation Q measurements. Preliminary results show northeast trending high Q regions (low attenuation) through the majority of the CEUS. Regions of low Q (high attenuation) were seen along the Minnesota-Wisconsin border, the Mississippi embayment, and along the Oklahoma-Texas border. The northeast trending high Q regions may represent the northeast trending sutures of island arcs that were accreted to the core of the Laurentia craton in the Proterozoic. The area of low attenuation in the Minnesota-Wisconsin region does not seem to correspond to known geological structures and may be due to high heat flow. Notably, the large Mid-Continent Rift feature extending from Lake Superior to Kansas seems to have little effect on Lg attenuation. The large region of attenuation in the Mississippi embayment is likely due to a combination of high heat flow and the large amount of sediment in the region based on previous geological and heat flow studies. The high attenuation region along the Oklahoma-Texas border corresponds to the Anadarko Basin, one of the largest oil and gas reserves in the United States. Finally, the low attenuation region extending from Kansas to Missouri may correspond to the Penokean orogen. These results are the first step in creating a more detailed model of crustal attenuation in the CEUS. This model can improve ground motion predictions of future large earthquakes for more accurate hazard assessment and improve overall understanding of the structure and assemblage of the CEUS.
Session: Intermountain West Earthquake and Volcano Characterization and Related Hazards
Presenter   Ingate, Shane
Schedule   Wed PM / Poster
Room   Hall 1
Seismicity Patterns Linked to Deformation Associated with the Mid-Crustal Socorro Magma Body, Central New Mexico
BILEK, S. L., New Mexico Tech, Socorro, NM, sbilek@nmt.edu; ASTER, R. C., New Mexico Tech, Socorro, NM, aster@ees.nmt.edu; INGATE, S., New Mexico Tech, Socorro, NM, shane@passcal.nmt.edu
Within central New Mexico, the Socorro Seismic Anomaly (SSA) is responsible for 20-40% of magnitude ≥ 2.0 events in the 60-year instrumental record. This relatively high rate of seismicity is due to the presence of an inflating (~2 mm/year maximal surface uplift), predominantly thin, mid-crustal magma body (the Socorro Magma Body; SMB) with an upper surface at ~19 km depth. The SMB bears similarities to other interpreted magmatic seismic “bright spots”, including others in the Basin and Range province. Reflected phases on microearthquake seismograms recorded above the SSA and COCORP active seismic studies were initially used to define the depth and the extent of this feature. More recently, InSAR studies confirmed and refined the boundaries of the associated uplift. Over the last few decades many hundreds of magnitude ≥ 2.0 earthquakes have been recorded in the SSA. The largest events in the historical record occurred in 1906 as part of a protracted (multi-year) swarm, and have been estimated to about magnitude 5.8. Notable earthquake events in the SSA often occur as a part of a seismic swarm lasting few days to a few weeks, with few events reaching magnitude ≥ 2.0. As expected in a rifting environment, most of these events represent movement along normal faults, sometimes with a strike-slip component. Recent volcanism in the region as young as ~5,000 years indicates the possibility of basaltic volcanism associated with this feature, and the extraordinary flatness of the upper SMB surface may be consistent with such a low viscosity composition. Alternatively, the SMB may be an active, primarily sill-like mafic intrusion that will ultimately solidify in the mid-crust without erupting. The long-standing seismicity above the SMB, including events as shallow as ~2 km, suggests that amagmatic migration of hydrothermal fluids and uplift accommodation can be expected to generate earthquakes above such features.
Session: Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Models
Presenter   Zielke, Olaf
Schedule   Wed AM / Poster
Room   Hall 1
Kinematic Earthquake Source Inversion using MCMC for Uncertainty Quantification in Rupture Models and Their Predictions
ZIELKE, O., King Abdullah University of Science and Technology, Thuwal, Saudi Arabia, olaf.zielke@kaust.edu.sa; PRUDENCIO, E., University of Texas, Austin, TX, prudenci@ices.utexas.edu; MAI, P. M., King Abdullah University of Science and Technology, Thuwal, Saudi Arabia, martin.mai@kaust.edu.sa
Seismic and geodetic data are frequently used to invert for the spatio-temporal evolution of slip along a rupture plane. Such kinematic rupture models help to investigate the underlying physics of earthquakes, but are also needed for ground-motion prediction and tsunami simulation in case of large subduction-zone earthquakes. It was noted that images of along-fault slip evolution for a single event can vary distinctly, depending on the adopted inversion approach and rupture model parameterization. This fact leaves the question, which --if any-- of the provided kinematic source inversion solutions is “correct” and how accurate are fault parameterization and solution predictions? These issues are not included in “standard” source inversion approaches. We present a MCMC-based (Markov Chain Monte Carlo) approach to identify --for a given earthquake-- the optimal fault parameterization and resulting model predictions, while simultaneously quantifying the corresponding uncertainties. Assuming a prior joint probability density function (PDF) for rupture plane dimension and orientation, rupture plane position relative to hypocenter location, spatial fault discretization, mean slip direction, and source time function, as well as a likelihood PDF involving seismic data and the aforementioned model parameters, we use Bayes theorem and MCMC algorithms to generate and sample the posterior joint PDF of the model parameters. Any already computed posterior PDF can in turn be considered prior information as soon as new seismic data --or any other auxiliary information-- become available. The presented process thus allows us to continuously update our knowledge about the seismic event as new data is provided. It also enables us to quantitatively track the uncertainties of rupture model and its predictions as a function of newly arriving data. The approach is tested using synthetic test cases and recent well-recorded earthquakes.
Session: Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Models
Presenter   Razafindrakoto, Hoby
Schedule   Wed AM / Poster
Room   Hall 1
STUDENT
Role of Source-Time Function and Earth Structure on Uncertainty in Earthquake Source Imaging
RAZAFINDRAKOTO, H. N. T., King Abdullah University of Science and Technology, Thuwal, Saudi Arabia, hoby.razafindrakoto@kaust.edu.sa; MAI, P. M., King Abdullah University of Science and Technology, Thuwal, Saudi Arabia, martin.mai@kaust.edu.sa
Different factors cause the non-uniqueness of kinematic source imaging such as the non-linearity of the problem, the non-unique choice of physical assumptions, and the observational errors. In this study, we investigate the effect of variability in Earth structure and the assumed source-time function using Bayesian inference. This approach allows us to map kinematic source parameters in term of probability density function (PDF). Our reference model is given by a spontaneous dynamic rupture calculation that assumes random initial stress on a steeply dipping strike-slip fault, for which near-fault synthetics are computed at 40 well-distributed sites. We find that the use of an isosceles triangle as source-time function adds an artificial correlation between parameters that does not occur in case of a Yoffe source time function. This correlation appears to compensate the dynamic source effects that are not included for the triangular source-time function. Since Earth structure is in general incompletely known, we additionally consider its variability, which may affect the rupture speed and hence the extension and location of high-slip patches. Our result shows a particular change in the extension of the high-slip region along the dip toward the surface. In our analysis we extract at each point of the fault the mean, median and maximum from the posterior PDFs. The results show that the pattern of the source-parameter estimates share the same major slip and peak slip-rate patches around the hypocenter, consistent with the reference model. However, due to the skewed shape of PDFs, consisting mostly of one-sided tails, the median is the most appropriate estimates. This result is confirmed in our analysis, as the median model generates the best fitting synthetics, with a variance reduction of 94%.
Session: Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Models
Presenter   Prudencio, Ernesto
Schedule   Wed AM / Poster
Room   Hall 1
Assessing Earthquake Source Models Under Uncertainty with Bayesian Analysis and Parallel MCMC
CRUZ-JIMENEZ, H., King Abdullah University of Science and Technology, Thuwal, Kingdom of Saudi Arabia, hugo.cruzjimenez@kaust.edu.sa; MAI, P. M., King Abdullah University of Science and Technology, Thuwal, Kingdom of Saudi Arabia, martin.mai@kaust.edu.sa; PRUDENCIO, E. E., ICES, The University of Texas at Austin, Austin, TX, prudenci@ices.utexas.edu
Recent destructive earthquakes such as those in Haiti (2010), Chile (2010), and Japan (2011), highlight the importance of advanced ground-motion simulations for mitigating human and economical losses. Reliable ground motion predictions, with quantified uncertainty, are critical for designing response plans and large civil structures (e.g. bridges, dams, buildings). The advent of HPC allows comprehensive uncertainty quantification of the expected shaking levels when considering a large set of rupture models. In this study, we use Bayesian analysis to calibrate and quantitatively rank, under uncertainty, candidate earthquake source model hypotheses for Mw 6.5 strike-slip earthquakes that rupture at constant rupture velocity (2.7 km/s) on a vertical fault plane. The first, and simplest, hypothesis is characterized by homogeneous slip distribution, while the second hypothesis is characterized by an elliptical asperity at a fixed position. In both hypotheses the width, length and hypocentral depth are treated as random parameters. The last, and most complex, hypothesis also includes an elliptical asperity and we treat as random the following ellipse parameters: its area relative to the fault plane (from 10% to 50%), its position and the length of the semi-axes. We compute ground-motions at a dense virtual seismic network composed of 56 stations. The ranking returns the candidate hypothesis that provides simulations of peak ground velocities that best match the Boore and Atkinson (2008) empirical attenuation relation. Our ranking approach provides a pathway for statistically and rigorously assessing the source-model complexity required to achieve reliable ground-motion simulation for engineering purposes.
Session: Velocity Models and Modeling
Presenter   Hinzen, Klaus-G.
Schedule   Thu AM / Poster
Room   Hall 1
The Mycenaean Palace of Tiryns, Elements of a Comprehensive Archaeoseismic Study
HINZEN, K. G., Cologne University, Cologne, Germany, hinzen@uni-koeln.de; FLEISCHER, C., Cologne University, Cologne, Germany, claus.fleischer@uni-koeln.de; HINOJOSA, H., Cologne University, Cologne, Germany, hhinojos@uni-koeln.de; MARAN, J., University of Heidelberg, Heidelberg, Germany, joseph.maran@urz.uni-heidelberg.de; MEINHARDT, U., University of Heidelberg, Heidelberg, Germany, MeinhardtU@aol.com; REAMER, S. K., Cologne University, Cologne, Germany, s.reamer@uni-koeln.de; SCHWEPPE, G., Cologne University, Cologne, Germany, gregor.schweppe@uni-koeln.de; TZISLAKIS, J., Cologne University, Cologne, Germany, jana.tzislakis@uni-koeln.de
In the 1970s, the archaeologist K. Kilian first published the hypothesis that several destructive earthquakes contributed to the decline of Mycenaean palatial society, culminating in collapse around 1200 B.C.E. Damaged buildings and structures of the Tiryns citadel in the Argolid, Peloponnese, Greece, formed the nucleus of the hypothesis. Early work at the site of Tiryns was pioneered by the famous archeologist Heinrich Schliemann in the late 19th century. Those early excavations and several significant phases of restoration that followed pose an obstacle for determining prima facie damage as a part of contemporary archaeoseismic studies. The ruins of the Mycenaean center of Midea, situated 7 km east of Tiryns, also exhibit damage, which was attributed to the same earthquake proposed by Kilian as the cause of destruction of neighboring Tiryns. In an initial archaeoseismological field campaign, we gathered geophysical data for an engineering seismological model of the Tiryns site, including the excavated structures. Measurements included P- and S-wave refraction tomography on twelve profiles, ambient noise single station and array measurements and a gravity survey. In addition, we used 3D laser scanning to aid construction of a virtual model of the northwestern fortification walls supplemented by a systematic comparison between old and current photographs of the site. As both Tiryns and Midea were built on top of cone-shaped limestone hills, topographic amplification of seismic waves may have been a contributing factor to any structural earthquake damage that occurred and will be included in our models. We present the current status of the engineering geophysical model and an outline of the second phase of measurements, including architectural remains in Midea, and 3D modeling of earthquake response in the Argolid.
Session: Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Models
Presenter   Zengin, Esra
Schedule   Fri 2:30 PM / Oral
Room   155C
STUDENT
Ground Motion Estimations for the 2011 Van, Eastern Turkey Earthquakes Using Stochastic Finite Fault Approach
ZENGIN, E., Bogazici University, Department of Earthquake Engineering, Istanbul, Turkey, esra.zengin@boun.edu.tr; CAKTI, E., Bogazici University, Department of Earthquake Engineering, Istanbul, Turkey, eser.cakti@boun.edu.tr
The 23 October 2011 Van (Mw7.1) earthquake that occurred in eastern Turkey causing heavy damage particularly in the city of Van and town of Ercis. Another earthquake hit the region on 9 November 2011, close to town of Edremit, with a magnitude of Mw5.6, resulted in additional damage to the building stock which already affected by the first event. In this study, we present ground motion simulations of Van Earthquakes by using stochastic finite fault method (EXSIM, Motazedian and Atkinson, 2005; Boore, 2009). The source, path and site model parameters are calibrated based on recordings in time and frequency domain. After setting the optimum models for path and site effects, we try to obtain the best model by investigating several scenarios with varying source parameters such as rupture velocity, stress drop and fault depth. The performance of the model is evaluated on the basis of the misfit values for peak ground accelerations, peak ground velocities, 5% damped acceleration response spectra and Fourier amplitude spectra in the frequency range of 1.0Hz-25Hz. Validated model parameters are then used to perform regional simulations. Synthetic results are verified by ground motion prediction equations and bias adjusted ShakeMap data. Despite the limitation of the method for incorporating the directivity effect and inadequate representation of the soil conditions at the individual stations, a satisfactory match between simulated and observed values are obtained both in time and frequency domain. Regional synthetic results also show reasonable correlation with ground motion prediction equations.
Session: Next Generation of Ground Motion Prediction Models
Presenter   Hermkes, Marcel
Schedule   Thu AM / Poster
Room   Hall 1
STUDENT
Simultaneous Quantification of Epistemic and Aleatory Uncertainty in GMPEs using Gaussian Process Regression
HERMKES, M., University of Potsdam, Potsdam, Germany, hermkes@geo.uni-potsdam.de; KUEHN, N. M., University of Potsdam, Potsdam, Germany, nico@geo.uni-potsdam.de; RIGGELSEN, C., University of Potsdam, Potsdam, Germany, riggelsen@geo.uni-potsdam.de
We present a Bayesian non-parametric method based on Gaussian Process (GP) regression to derive ground-motion models for peak ground parameters and response spectral ordinates. Due to its non-parametric nature there is no need to specify any fixed functional form as in parametric regression models. A GP defines a distribution over functions, which implicitly expresses the uncertainty over the underlying data generating process. This has the advantage that it is possible to capture the whole uncertainty involved in ground-motion modeling, both in terms of aleatory variability as well as epistemic uncertainty associated with the underlying functional form and data coverage – in ranges where data is sparse the uncertainty increases. The distribution over functions is updated in a Bayesian way by computing the posterior distribution of the GP after observing ground-motion data, which in turn can be used to make predictions.Standard GP models are not capable to model correlated outputs and predicting pseudoacceleration ordinates at different periods jointly. The easiest way to handle this problem is to describe each output variable by an independent GP. The loss of information by ignoring the dependency structure of the outputs can lead to a negative influence on the quality of ground-motion prediction in seismic hazard analysis. To overcome this issue we also present an extension of the standard GP framework by convolution process formalism, in which the dependency between the different outputs is formulated by the convolution of a smoothing kernel and a latent function.The proposed GP regression models are evaluated on the Next Generation Attenuation (NGA) data set The experiments show that GP models have a better generalization error than a simple parametric regression model. A visual assessment of different scenarios demonstrates that the inferred GP models are physically plausible.
Session: Next Generation of Ground Motion Prediction Models
Presenter   Hermkes, Marcel
Schedule   Thu AM / Poster
Room   Hall 1
STUDENT
Transfer Lerning between Geographical Regions for GMPE Estimation
HERMKES, M., University of Potsdam, Potsdam, Germany, hermkes@geo.uni-potsdam.de; KUEHN, N. M., University of Potsdam, Potsdam, Germany, nico@geo.uni-potsdam.de; RIGGELSEN, C., University of Potsdam, Potsdam, Germany, riggelsen@geo.uni-potsdam.de
The accurate quantification of expected ground-motion is of key importance in probabilistic seismic hazard analysis (PSHA). In recent years, both quality and quantity of strong-motion data has been increasing, which in turn led to better ground-motion prediction equations (GMPEs). However, there are still many regions where strong-motion data is scarce. In this case, one needs to include data from other (background) regions to reliably estimate a GMPE for a region with limited data (the target region).The easiest way to combine data from different geographical regions is to pool all data together and learn a model on the whole data set. However, if there are differences in the physical properties of the individual regions, a negative information transfer might occur.In this work, we present a transfer learning approach, based on Gaussian process (GP) regression, in which information is shared in an intelligent way. Here, the relevant information from the background data is automatically selected, i.e. only those features from the other regions are utilized that are similar to the target region. In particular, for both target and background regions a single GP model is estimated, which are convolved with a kernel that describes the similarity of the background regions with the target region.The experiments are performed on the Next Generation Attenuation (NGA) and the RESORCE data set for Califonrian and European data, respecetively. For comparison we also consider a pooling approach as baseline method. The results show improved prediction performance of the transfer Learning approach compared to the simple baseline method.
Session: Earthquake Source Studies
Presenter   Stierle, Eva
Schedule   Wed PM / Poster
Room   Hall 1
STUDENT
Resolution of Non-Double-Couple Components in Seismic Moment Tensors of Aftershocks of the 1999 Mw7.4 Izmit Earthquake in NW Turkey
STIERLE, E., Helmholtz-Centre Potsdam, German Research Centre for Geosciences, Potsdam, Germany, stierle@gfz-potsdam.de; VAVRYČUK, V., Geophysical Institute, Academy of Sciences of the Czech Republic, Prague, Czech Republic; ŠÍLENÝ, J., Geophysical Institute, Academy of Sciences of the Czech Republic, Prague, Czech Republic; BOHNHOFF, M., Helmholtz-Centre Potsdam, German Research Centre for Geosciences, Potsdam, Germany, bohnhoff@gfz-potsdam.de
We investigate non-double-couple (NDC) components in aftershock moment tensors of the 1999 Mw=7.4 Izmit earthquake. Focal mechanisms of Izmit aftershocks clearly indicate a segmentation of the fault plane into several segments, one of which is the Akyazi Plain, a pull-apart structure, where significant NDC components might be observed.First, we performed theoretical tests to analyze the ability of a network of 35 stations along the entire Izmit rupture to detect source-related NDC components in seismic moment tensors of small earthquakes. We modeled synthetic P- and S-wave amplitudes adopting a shear-tensile source model with a varying NDC content. When inverting for the full moment tensor, we contaminated the data with random noise and systematically incorporated an inaccurate knowledge of the velocity structure and realistic errors in hypocenter precision. Varying coverage of the focal sphere and focal mechanisms representing the regional tectonic setting were considered. We found that the double-couple (DC) mechanisms can successfully be resolved for the given network. However, the resolution of the source-related NDC components remains limited due to the lack of stations with epicentral distances less than 15 km. Spurious NDC components caused by modeling errors and noise in data can be as high as 20% and are thus non-negligible.Applying our technique to waveform recordings of Izmit aftershocks using strict quality criteria allows us to determine a total of 29 stable moment tensor solutions. We find that pure shear faulting dominates strike-slip segments along the rupture reflecting the overall right-lateral strike-slip domain of the North Anatolian Fault Zone. In contrast, NDC components are significant (~30%) below the Akyazi Plain, a pull-apart structure hosting a substantial coseismic slip-deficit that resulted in EW-extensional normal faulting following the Izmit mainshock.
Session: Next Generation of Ground Motion Prediction Models
Presenter   Baltay, Annemarie
Schedule   Wed 11:45 AM / Oral
Room   155C
Magnitude Dependence of PGA and PGV in NGA-West2 Data
BALTAY, A. S., USGS, Menlo Park, CA, annemarie.baltay@gmail.com; HANKS, T. C., USGS, Menlo Park, CA, thanks@usgs.gov
This study explores the moment-magnitude (M) dependence of PGA (peak ground acceleration) and PGV (peak ground velocity) in the NGA-West2 database at close distances (R = 10 km) and site condition Vs30 = 620 m/sec with simple point-source, constant stress-drop earthquake source models coupled with random vibration theory (RVT). These models, developed over 30 years ago, provide a remarkably good fit to the NGA PGA and PGV data for 3 < M < 8 at close distances. This magnitude range spans four different functional dependences on M. For 6.5 < M < 8, the magnitude dependence is very weak, attributable to saturation of PGA and PGV recorded at close distances from a finite fault. For 4 < M < 6.5, the M dependence of PGA is similar to that of Hanks and McGuire [1981], using a constant acceleration spectra assumption and a simple approximation of RVT. When the earthquake corner frequency fc, approaches fmax [Hanks, 1982] at M ~ 4, more precise RVT is necessary to model the data; we used the SMSIM software of Boore [1983], using kappa, the near-site attenuation term. At M < 3, for which fc is greater than fmax, PGA and PGV are proportional to seismic moment, equivalently log (PGA, PGV) ~ 1.5M [Hanks and Boore, 1984]. The NGA-West2 empirical models are consonant with this simple point-source, constant stress-drop source model, and exhibit the same slope discontinuities at M ~ 4 and ~ 6.5. Due to the magnitude limitation of the NGA-West2 data at M < 3, however, the log (PGA, PGV) ~ 1.5M dependence is not clearly seen; small magnitude, close-in PGA and PGV data from Switzerland for 2 < M < 4 do show this 1.5M dependence. The overlap of the Swiss and NGA data for 3 < M < 4 and their agreement with the 1.5M dependence at M < 3 suggests that ground motions in Switzerland at close distances and larger M will conform to a constant stress drop model similar to that found for the NGA data.
Session: Next Generation of Ground Motion Prediction Models
Presenter   Edwards, Benjamin
Schedule   Wed 5:00 PM / Oral
Room   155C
Improving Referenced Ground-Motion Prediction through Physically-Based Site-Specific Adjustments
EDWARDS, B., Swiss Seismological Service, ETH Zürich, Zürich, Switzerland, edwards@sed.ethz.ch; POGGI, V., Swiss Seismological Service, ETH Zürich, Zürich, Switzerland, poggi@sed.ethz.ch; FÄH, D., Swiss Seismological Service, ETH Zürich, Zürich, Switzerland, faeh@sed.ethz.ch
We present the analysis of strong-motion data from European and Japanese strong-motion databases in order to investigate regional and site-to-site variability in recorded ground-motion and the suitability of stochastic models to predict near-source strong ground-motion in regions of limited seismicity. Site-specific attenuation (κ0) is known to have a significant impact on the seismic hazard at frequencies typically of interest when considering nuclear power plant sites. However, κ0 is not particularly well accommodated by current GMPEs, making host-to-target adjustments problematic. We estimate crustal attenuation (Q0) and κ0 using two methods: a broadband spectral modelling approach and a high-frequency linear fit. κ0 is found to vary strongly from site-to-site: from negligible to κ0≈0.09s, consistent with the wide variety of recording-site conditions in the databases. κ0 is shown to weakly correlate with Vs30, while we show that improved modelling is possible through consideration of quarter-wavelength velocity models. We then estimate site-class or rock-referenced site-specific elastic amplification, and, using this information and the attenuation models previously determined, compute seismic moments and ω2 stress-parameters for the events. Site amplification is shown to vary strongly within a single site-class: showing considerable scope for the reduction of uncertainty in GMPEs, given better site characterisation. We show that the seismic moments determined are consistent with database Mw values, and show that the resulting average stress-parameters are independent of magnitude or depth. Finally we apply the results of the spectral analyses to develop a stochastic ground-motion model for the prediction of near-field strong ground-motion. We show that such models, with appropriate parameterisation to take account of fault geometry can be useful for predicting site-specific rock-referenced ground motion using physically based host-to-target adjustments.
Session: Next Generation of Ground Motion Prediction Models
Presenter   Baltay, Annemarie
Schedule   Thu AM / Poster
Room   Hall 1
Using Tremor to Determine Strong Ground Motion Parameters
BALTAY, A. S., USGS, Menlo Park, CA, annemarie.baltay@gmail.com; ASAWACHAISUJJA, S., Chulalongkorn University, Bangkock, Thailand, sirada.ning@gmail.com; BEROZA, G. C., Stanford University, Stanford, CA, beroza@stanford.edu
Due to its widespread occurrence, frequency content, and location, tectonic tremor presents an exceptional opportunity to test and improve strong ground motion attenuation relations for subduction zones. We characterize the amplitude of thousands of individual tremor events to constrain the distance attenuation relationship of peak ground acceleration (PGA) and peak ground velocity (PGV) of tremor. Ground motion prediction equations (GMPE) relate ground motion amplitude to earthquake magnitude and distance, and are critically important for creating seismic hazard maps. In some areas of high earthquake hazard, such as Cascadia, the data set of recorded earthquakes used in GMPE development is extremely sparse. Using locations and timings from Wech and Creager [2008] of thousands of individual tremor events occurring during an ETS event, we can observe tremor in Cascadia recorded at borehole Plate Boundary Observatory stations to distances of 150 km, sufficient to place important constraints on ground motion distance decay. Because tremor bursts do not have a defined absolute magnitude, we use a difference inversion to determine relative magnitudes of the events. We also remove a station-specific residual, representing the site term. After normalization by the relative magnitude and site term, we can precisely estimate the anelastic attenuation term (c4). We find the same c4 term, to three significant digits, for tremor events from each 2010, 2011 and 2012 in Cascadia, for both PGA and PGV; both PGA and PGV values are similar to subduction-zone-specific GMPEs developed from both data and simulations [e.g. Atkinson and Boore, 1997]. We extend this analysis to include other subduction zones in which tremor is recorded. The massive amount of data present in the tremor observations allows us to improve the fit of the GMPEs, to refine distance-amplitude attenuation relationships, and to search for regional or intra-subduction zone variations in behavior.
Session: Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Models
Presenter   Cakti, Eser
Schedule   Fri 4:30 PM / Oral
Room   155C
Simulation of the 1999 Kocaeli, Turkey Earthquake
CAKTI, E., Bogazici University, Istanbul, Turkey, eser.cakti@boun.edu.tr; SESETYAN, K., Bogazici University, Istanbul, Turkey; MADARIAGA, R., Ecole Normale Superieur, Paris, France
We simulate the 1999 Kocaeli, Turkey earthquake, with the intention of developing a working model that is capable of simulating other or future earthquakes in the Marmara region. We make use of a velocity model created earlier, improve it by incorporating information and data that became recently available. We validate the improved velocity model by regional, small and medium size events. There are several extended source models available for the Kocaeli earthquake. Using them we simulate the strong motion records from the earthquake and look at the differences caused by them. We then embark on grid-based regional simulation of the same event with the aim of coming up with ground motion estimates that are comparable with field observations that in the case of the Kocaeli earthquake extended over a very large area. The effective frequency range of the simulation is 0-0.8Hz. The low frequency ground motion is simulated by code FD3D.
Session: Earthquake Source Physics
Presenter   Prieto, German
Schedule   Thu 9:30 AM / Oral
Room   155A
Radiation Efficiency and Fracture Energy of Intermediate-depth Earthquakes
PRIETO, G. A., EAPS, MIT, Crambridge, MA, gprieto@mit.edu; FLOREZ, M., Universidad de los Andes, Bogota, Colombia; BARRETT, S. A., Stanford University, Palo Alto, CA; BEROZA, G. C., Stanford University, Palo Alto, CA
The behavior of rupture of earthquakes at depths below 50 km is not well constrained. In contrast to conditions in the crust and shallow lithosphere, at temperatures and pressures corresponding to these depths one would expect rocks to yield by creep or flow and not by brittle failure, so there has to be a physical mechanism that allows for brittle or brittle- like failure for intermediate-depth earthquakes. To better understand their mechanisms it is fundamental to constrain both the static parameters as well as the dynamic parameters of these earthquakes. We perform spectral analysis of over 200 intermediate-depth earthquakes in the Bucaramanga Nest and determine seismic moment, corner frequencies and radiated seismic energies in the M3.5 – 5.2 magnitude range. We determine radiation efficiencies and using simple models determine fracture energies for all earthquakes. Fracture energies tend to be higher than shallow earthquakes suggesting that frictional melting or concentrated shear zones are developed during seismic ruptures at intermediate-depths.
Session: Triggering of Seismic and Volcanic Events
Presenter   Pollitz, Fred
Schedule   Fri 11:00 AM / Oral
Room   155D
The Profound Reach of the M8.6 11 April 2012 Indian Ocean Earthquake: Short-term Global Rate Increase Followed by a Long-term Global Rate Drop
POLLITZ, F. F., USGS, Menlo Park, CA, fpollitz@usgs.gov; BURGMANN, R., Dept. Earth and Planetary Sci., UC Berkeley, Berkeley, CA, burgmann@seismo.berkeley.edu; STEIN, R. S., USGS, Menlo Park, CA, rstein@usgs.gov; SEVILGEN, V., USGS, Menlo Park, CA, vsevilgen@usgs.gov
The M=8.6 11 April 2012 Indian Ocean earthquake was an unusually large intra-oceanic strike-slip event. For several days the global M ≥ 4.5 seismicity rates at remote distances (i.e. thousands of km from the mainshock) were elevated (Pollitz et al., Nature 2012). But the M ≥ 6.5 rate subsequently dropped to zero for the succeeding 95 days; global rates at 4.5 ≤ M ≤ 6.3 were nearly identical to background during this period. Such an extended period without a M ≥ 6.5 event has happened rarely over the past century. We interpret both the short-lived global seismicity rate increase followed by the longer quiet period as the product of dynamic stressing of a global system of close-to-failure faults. Transient dynamic stresses can encourage short-term triggering but, paradoxically, can also inhibit rupture temporarily until background tectonic loading restores the system to its pre-mainshock stress levels. We construct a statistical model of global seismicity involving tens of thousands of potential M ≥ 6.5 source patches governed by a single state variable (the shear strain) which is randomly distributed among all possible strain states between full strain drop and critically strained. When this system is subjected to a transient strain of εd = 0.2 μstrain, approximately the transient perturbation of the April 2012 event transmitted globally, we find that 6% of those patches within εd of failure were triggered (with a delay) by passage of the seismic waves; 88% of the remainder were inhibited from failure over the subsequent 95 days regardless of how close they were to failure before the April 2012 mainshock. This carries important implications for fault mechanics when faults are subjected to a transient stress.
Session: Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Models
Presenter   Edwards, Benjamin
Schedule   Fri 3:45 PM / Oral
Room   155C
A Kappa Model to Predict the Attenuation Characteristics of Gradient-Like Velocity Profiles
POGGI, V., Swiss Seismological Service, ETHZ, Zurich, Switzerland, poggi@sed.ethz.ch; EDWARDS, B., Swiss Seismological Service, ETHZ, Zurich, Switzerland, edwards@sed.ethz.ch; FÄH, D., Swiss Seismological Service, ETHZ, Zurich, Switzerland, donat.faeh@sed.ethz.ch
The definition of a common reference is a key issue in probabilistic seismic hazard analysis, microzonation studies, local site-response analysis and, more generally, when predicted or observed ground motion is compared for sites of different characteristics. Nowadays methods requiring the definition of a reference condition generally prescribe the characteristic of a rock reference, calibrated using indirect estimation methods based on geology or on surface proxies. In most cases, a unique average shear-wave velocity value is prescribed. Moreover, in spite of its relevance in affecting the high-frequency part of the spectrum, the definition of the associated reference attenuation is in most cases missing or, when present, still remains quite uncertain.In this study we propose an approach for the combined estimation of the attenuation characteristics and shear-wave velocity of a regional reference. Such approach is based on the comparison between empirical anelastic amplification functions from spectral modeling of earthquakes and average S-wave velocities computed using the quarter-wavelength approach. The method is an extension of the approach originally proposed by Poggi et al. (2011) for Switzerland, and is here applied to Japan. With respect to the previous study, we now analyze separately the elastic and anelastic contributions of the estimated empirical amplification by using the frequency independent (and site-dependent) attenuation operator kappa (κ). By comparing the dependency of κ with the quarter-wavelength velocity at selected sites, a predictive equation is established to model the attenuation characteristics of an arbitrary rock or stiff-soil velocity model, such as the target reference profile. As an additional output of the present study, we also propose a simplified method to estimate κ from the average velocity estimates over the first 30m (Vs30).
Session: Next Generation of Ground Motion Prediction Models
Presenter   Atkinson, Gail
Schedule   Wed 2:45 PM / Oral
Room   155C
Empirical Evidence for the Frequency-Dependence of Geometric Spreading in Eastern North America
ATKINSON, G. M., Western University, London, ON, Canada, gmatkinson@aol.com; BOORE, D. M., US Geological Survey, Menlo Park, CA, boore@usgs.gov
Analysis of ground-motion data from small-to-moderate earthquakes on rock sites in eastern North America (ENA) provides strong empirical evidence that the apparent geometric spreading coefficient for Fourier amplitudes depends on frequency at distances within 100 km of the source. An attenuation model that provides near-zero residuals at all distances from 10 to 500 km is a bilinear model with a slope (apparent geometric spreading) of R-1.3 to a transition distance Rt, with a geometric spreading of R-0.5 at greater distances. The associated anelastic attenuation is event specific, with an average value being given by a regional Quality factor of Q=680 f0.33. The transition distance (Rt) decreases with frequency from 70 km at ≤2 Hz to 30 km at ≥10 Hz. The frequency-dependence of Rt may be responsible for the second corner in the apparent source spectrum that is often observed for ENA events of M >4.5. It is important to stress that this apparent geometric spreading model is only an empirical description of how the amplitudes appear to decay, as opposed to a physical model. In reality, the apparent geometric spreading may be reflecting other physical effects; for example, the actual geometric spreading could be 1/R (say), but frequency-dependent radiation pattern and directivity effects could boost near-distance amplitudes, resulting in the observed apparent frequency-dependence in geometric spreading.
Session: Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Models
Presenter   Askan, Aysegul
Schedule   Fri 8:45 AM / Oral
Room   155C
Assessment of Synthetic Ground Motion Records from Alternative Simulation Methods in Dynamic Response Analyses of Building Structures
KARIMZADEH, S., Middle East Technical University, Ankara, Turkey, sh_naghshineh@yahoo.com; ASKAN, A., Middle East Technical University, Ankara, Turkey, aaskan@metu.edu.tr; YAKUT, A., Middle East Technical University, Ankara,Turkey, ayakut@metu.edu.tr; AMERI, G., GEOTER International, Auriol, France, gabriele.ameri@gmail.com
Nonlinear Time History Analysis (NLTHA) is a numerical technique commonly used in earthquake engineering for the evaluation of nonlinear dynamic response of a structure. NLTHA requires full time history of the ground motions. For regions with sparse ground motion data, ground motion simulations provide alternative acceleration time series for this purpose. Alternative simulation methods provide different levels of accuracy in terms of modeling any given ground motion record. It is thus critical to investigate the nonlinear response of structures to synthetic records of alternative levels of accuracy. For this purpose, in this study, we present nonlinear time history analyses of multi-storey frame buildings under real and corresponding synthetic ground motions. Synthetic records of 6 April 2009 L’Aquila (Italy) earthquake are simulated using the Hybrid Integral-Composite method and the stochastic finite-fault method. Results of nonlinear time history analyses from real and alternative synthetic records of this event are expressed in terms of maximum displacement of each storey levels and inter-storey drift ratios. We compare the results from synthetic and real records to investigate the ability of synthetic ground motions to predict the seismic responses of reinforced concrete frame structures.
Session: New Developments in Earthquake Forecasting and Predictability Research
Presenter   Wang, Zhenming
Schedule   Thu 2:00 PM / Oral
Room   155D
Earthquake Prediction and Forecast: From a Layperson’s Viewpoint
WANG, Z., University of Kentucky, Lexington, KY, zmwang@uky.edu
Although earth scientists have made significant efforts to study earthquake prediction and forecasting, as well as seismic hazard and risk assessment, recent earthquakes such as the 2008 Wenchuan (China), 2010 Haiti, and 2011 Christchurch (New Zealand) and Tohoku (Japan) earthquakes showed that earthquake prediction and forecasting are still great challenges for earth scientists. In Particular, the 2009 L’Aquila (Italy) earthquake brought prosecution and six-year prison sentences for six scientists for “failure to predict the earthquake” or “inadequate communication of seismic hazard and risk” before the M6.3 quake. The conviction of the six scientists not only shocked the earth science community, but also brought intense discussion about the role earth scientists should play in seismic hazard mitigation and risk reduction. The earth science community recognizes that earthquakes are difficult to predict or forecast because the probability of a large earthquake is generally quite small: less than 1 percent for any given day. Although the low probability of earthquake occurrence can be quantified by earth scientists, communicating this probability to the general public is difficult, even impossible. For example, a probability of 1 percent may not be significantly different from a probability of 0.1 percent from the point of view of a layperson. Earth scientists must find better ways to communicate with the general public.
Session: What are the Limits of Explosion Source Model Predictions?
Presenter   Stroujkova, Anastasia
Schedule   Wed 10:45 AM / Oral
Room   155D
Explosion Spectra from Near- to Far-Field: Source Effects, Structure or Dissipation?
STROUJKOVA, A., Weston Geophysical Corp., Lexington, MA, ana@westongeophysical.com; MOROZOV, I., University of Saskatchewan, Saskatoon, SK, Canada, igor.morozov@usask.ca
Monitoring of nuclear explosions is commonly performed at regional or teleseismic distances, thus the spectra of the recorded seismic waves are significantly affected by the propagation. Structural effects add significant complexities to the P-wave spectra that can distort the source signature. Viscoelastic attenuation as well as scattering may potentially change the high-frequency roll-off of the spectra. Nevertheless, the spectra recorded at local to regional distances are used to quantify the explosion source. Our recent seismic observations of chemical explosions at distances from several meters to several kilometers show that corner frequencies are altered by the filtering effect of the medium even at short distances. Range-dependent spectral ratios offer ways of quantifying the viscoelastic and scattering effects of the media. However, variations of amplitudes and fracturing in the near field may modify the dissipation characteristics in a complex manner, causing range-dependent corner frequencies and spectral roll-offs. Therefore, special care should be taken when interpreting the far-field explosion spectra, as they may not be representative of the source spectra. In this work we attempt to quantify the effects of the medium on the source characteristics measured from the near-source to local distances.
Session: Intermountain West Earthquake and Volcano Characterization and Related Hazards
Presenter   Pancha, Aasha
Schedule   Wed PM / Poster
Room   Hall 1
Determination of 3D Velocity Structure across the Deepest Portion of the Reno-Area Basin: Development of the Western Basin and Range Community Velocity Model and the Reno-Tahoe Urban Hazard Map
PANCHA, A., OptimSDS, Reno, NV, aashap@optimsds.com; PULLAMMANAPPALLIL, S., Optim Inc., Reno, NV, satish@optimsoftware.com; CASHMAN, P. H., Geological Sciences & Engineering,University of Nevada, Reno, NV, pcashman@mines.unr.edu; TREXLER, J. H., Geological Sciences & Engineering,University of Nevada, Reno, NV, trexler@unr.edu; WEST, L. T., OptimSDS, Reno, NV, travisw@optimsds.com; LOUIE, J. N., Nevada Seismological Laboratory, University of Nevada, Reno, NV, louie@seismo.unr.edu
We assessed shear-wave velocities across the deepest portion of the Reno-area basin as defined by gravity. We were able to delineate basin thicknesses of up to 900 m, and subsurface structure, including a buried river terrace. To do this we measured shear velocities to depths of 1000 m using refraction-microtremor (ReMi) arrays across this portion of the basin with 50 m depth resolution. This was achieved through the deployment of 30 stand alone wireless instruments, in arrays 2.9 to 5.8 km long, to record ambient urban noise. Data were obtained along two parallel east-west arrays and one shorter north-south array. The ReMi technique was employed to obtain 1D velocity profiles as a function of depth across each array from these noise records. To characterize and map lateral velocity heterogeneity beneath the arrays, data from overlapping subsets of instruments were used to obtain a series of 1D soundings. These 1D velocity-depth profiles were then interpolated to obtain 2D and 3D structural representations of shear-wave velocities, which delineate the deep velocity structure of the basin.Estimation of shallow and deep shear velocity is a key element in the assessment of sites for potential earthquake ground shaking, damage, and the calibration of recorded ground motions. To date, no permanent seismic instrumentation has been located over this thickest portion of the Reno-area basin and little structural and velocity data had been available. Existing velocity models for the region were limited in resolution to intervals of 1 km to 3 km. As a result, 3D basin details were insufficient for scenario modeling and ShakeZoning, essential components of seismic hazard evaluation. Our new efforts contribute these basin details towards development of the Western Basin and Range Community Velocity Model and the Reno-Tahoe urban hazard map.
Session: Advances in High-Frequency Physics-Based Earthquake Systems Simulation
Presenter   Tanimoto, Toshiro
Schedule   Thu 11:45 AM / Oral
Room   155C
High-Frequency Monochromatic Oscillations in the Los Angeles Basin and Numerical Simulation
TANIMOTO, T., UCSB, Santa Barbara, CA, toshiro@geol.ucsb.edu; OKAMOTO, T., Tokyo Institute of Technology, Meguro-ku, Tokyo, Japan, okamoto.t.ad@m.titech.ac.jp
What are the highest-frequency seismic waves that can be modelled deterministically? Currently, a conventional approach is to use deterministic waveform modeling below the frequency about 0.5-1.0 Hz and to use a stochastic method above it. Furthermore, for a long time, stochasticity in the source process was included but stochasticity in the wave propagation, i.e. from small heterogeneity in the medium, was often not included. This situation is now changing because it is now feasible to numerically compute waveforms in the 3D media up to 5-10 Hz. There still remain, however, some difficult problems that lie ahead in terms of constraining stochastic properties in the source process and those in the wave propagation process.We propose to get better constraints on the wave-propagation part by using unique oscillation data in the Los Angeles basin. The data are the results of monochromatic oscillations, generated by a building in Pasadena (at Caltech). Seismic waveforms were recorded by Southern California Seismograph Network (SCSN) and are available from the SCEC data center. We will present the shaking results of two experiments at 1.64 Hz and at 1.11 Hz, both of which lasted 4-6 hours. Because of their long durations, signals have higher spectral amplitudes than microseisms and can be identified at stations over 300 km. These data can be used to verify the 3D reference model in the area (SCEC Community Velocity Model) through comparison between data and simulated results. The source is basically a simple sinusoidal horizontal force at the surface and can be modeled easily. The sensitivity kernels of waveforms can be obtained for these oscillations in a straightforward, adjoint-operator based method. We will discuss how these data sample the medium and may be used to improve our understanding of seismic structure.
Session: Velocity Models and Modeling
Presenter   Simmons, Nathan
Schedule   Thu AM / Poster
Room   Hall 1
LLNL-G3Dv3: Global P-Wave Tomography Model for Improved Regional and Teleseismic Travel Time Prediction
SIMMONS, N. A., Lawrence Livermore National Laboratory, Livermore, CA, simmons27@LLNL.gov; MYERS, S. C., Lawrence Livermore National Laboratory, Livermore, CA; JOHANNESSON, G., Lawrence Livermore National Laboratory, Livermore, CA; MATZEL, E., Lawrence Livermore National Laboratory, Livermore, CA
We develop a global-scale P-wave velocity model (LLNL-G3Dv3) designed to accurately predict seismic travel times at regional and teleseismic distances simultaneously. The model provides a new image of Earth’s interior, but the underlying practical purpose of the model is to provide enhanced seismic event location capabilities. The LLNL-G3Dv3 model is based on ~2.8 million P and Pn arrivals that are re-processed using our global multiple-event locator called Bayesloc. We construct LLNL-G3Dv3 within a spherical tessellation based framework, allowing for explicit representation of undulating and discontinuous layers including the crust and transition zone layers. Using a multi-scale inversion technique, regional trends as well as fine details are captured where the data allow. LLNL-G3Dv3 exhibits large-scale structures including cratons and superplumes as well numerous complex details in the upper mantle including within the transition zone. Particularly, the model reveals new details of a vast network of subducted slabs trapped within the transition beneath much of Eurasia, including beneath the Tibetan Plateau. We demonstrate the impact of Bayesloc multiple-event location on the resulting tomographic images through comparison with images produced without the benefit of multiple-event constraints (single-event locations). We find that the multiple-event locations allow for better reconciliation of the large set of direct P phases recorded at 0-97° distance and yield a smoother and more continuous image relative to the single-event locations. Travel times predicted from a 3-D model are also found to be strongly influenced by the initial locations of the input data, even when an iterative inversion/relocation technique is employed. This work performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344. LLNL-ABS-559093
Session: New Frontiers in Seismic Data Analysis
Presenter   VanDeMark, Thomas
Schedule   Thu AM / Poster
Room   Hall 1
Template Based Association of Seismic Phases
JUNEK, W. N., US National Data Center, Air Force Technical Applications Center, Patrick Air Force Base, FL, william.junek@us.af.mil; POPE, B. M., US National Data Center, Air Force Technical Applications Center, Patrick Air Force Base, FL; VANDEMARK, T. F., US National Data Center, Air Force Technical Applications Center, Patrick Air Force Base, FL; SAULTS, T. R., US National Data Center, Air Force Technical Applications Center, Patrick Air Force Base, FL; WEHLEN, J. A., US National Data Center, Air Force Technical Applications Center, Patrick Air Force Base, FL; MORENCY, C. D., US National Data Center, Air Force Technical Applications Center, Patrick Air Force Base, FL; ZEILER, C., US National Data Center, Air Force Technical Applications Center, Patrick Air Force Base, FL
Aftershock sequences and earthquake swarms often produce numerous events that are close in space and time. These sequences significantly delay the interactive analysis of seismic data by human analysts due to the incorrect association of seismic phases by the Global Association (GA) application. We have developed an algorithm that uses alphanumeric phase association templates to assist the GA application with building events from large aftershock sequences and earthquake swarms. Template information is mined from recently reviewed event solutions archived in a relational database. Each template is derived from a collection of alphanumeric feature vectors that represent the event's space, time, and source characteristics. Feature vectors contain a series of event specific phase measurements and uncertainty estimates from each station in the seismic network which include; event to station distance, travel time, back azimuth, amplitude, and period. Positions in the feature vector with no corresponding measurements are populated with zeros to represent the absence of an observation at a particular station. Vectors are automatically grouped into families of similar events using an unsupervised clustering algorithm. Association templates from each event cluster are assembled using the mean and standard deviation of each alphanumeric feature for all stations in the seismic network. Templates are automatically regenerated during a developing earthquake sequence as the analyst reviewed bulletin grows. As a result, the quality of the automated bulletin will improve and analyst burden will decrease over time. The improvement to the automated catalog is measured using a bulletin quality metric that quantifies the similarity between the automated and analyst reviewed bulletins.
Session: New Frontiers in Seismic Data Analysis
Presenter   Langston, Charles
Schedule   WITHDRAWN
Room   Hall 1
Point Gradiometers for “Continuum” Seismic Recording
LANGSTON, C. A., University of Memphis, Memphis, TN, clangstn@memphis.edu; DUNN, R. W., Hendrix College, Conway, AR, Dunn@hendrix.edu
Aside from fundamental contributions made by Hugo Benioff in the 1930’s concerning strain seismographs, the principles of seismic recording have basically remained the same for nearly 150 years through the use of inertial seismographs that are primarily sensitive to ground velocity or acceleration. A seismic wave’s temporal vector ground motion is but shadow of all of the attributes that could be known before using the wave to understand the Earth. In particular, a wave’s spatial variability can be used to find additional information of wave velocity, wave type, wave propagation direction, and spatial amplitude changes. Such data would offer an order-of-magnitude greater number of constraints to studies of velocity heterogeneity (e.g., tomography, wave scattering, anisotropy), source complexity (e.g., rupture propagation, finiteness), and studies of media non-linearity for strong ground motions. This kind of seismic recording is being explored through the use of dense seismic arrays (gradiometers) that have apertures of less than 10% of a wavelength. However, a system consisting of a collocated, three-component seismometer with a three axis seismic rotation meter and a horizontal tensor strain meter can also sample the local continuum at the scale of a seismic wavelength by taking measurements of ground motion and its spatial gradients at a single point – effectively creating a “point seismic array” or point gradiometer. Such observatories are now possible and a practical configuration is being designed for testing in the central United States.
Session: ShakeMap-Related Research, Development, Operations, and Applications
Presenter   Huerfano, Victor
Schedule   Thu PM / Poster
Room   Hall 1
Improvements to the Puerto Rico SHAKEMAP System
HUERFANO, V. A., Puerto Rico Seismic Network, Mayaguez, PR, victor@prsn.uprm.edu; TORRES, M., Puerto Rico Seismic Network, Mayaguez, PR, maria@prsn.uprm.edu
The island of Puerto Rico is located between two major plates, Caribe plate (CP) and North American plate (NA). The CA is moving to the east, and the NA is moving to the west; the energy is being accumulated along the Puerto Rico Trench and other fault systems. Major earthquakes from off-shore sources have affected Puerto Rico in 1520, 1615, 1670, 1751, 1787, 1867, 1918 and 1946 (Mueller et al, 2003; PRSN Catalogue). Recent trenching has also yielded evidence of possible M7.0 events inland in the southwestern and south regions of the Island. The high seismic hazard, large population, high tsunami potential and relatively poor construction practice can result in a potentially devastating combination. Efficient emergency response in case of a large earthquake will be crucial to minimizing the loss of life and disruption of lifeline systems in Puerto Rico. The ShakeMap system (Wald et al, 2004) developed by the USGS to rapidly display and disseminate information about the geographical distribution of ground shaking (and hence potential damage) following a large earthquake has proven to be a vital tool for post-earthquake emergency response efforts, and is being adopted/emulated in various seismically active regions worldwide. Implementing a robust ShakeMap system is among the top priorities of the Puerto Rico Seismic Network (PRSN). However, the ultimate effectiveness of ShakeMap in post- earthquake response depends not only on its rapid availability, but also on the effective use of the information it provides. In this work we will present recent developments of the PRSN ShakeMap system, as the correction using the online DFI/CIIM reports and rapid instrumental corrections base on direct computation of ground motions from the PRSN real time system.
Session: Data Products as Research Resources
Presenter   Trabant, Chad
Schedule   Fri AM / Poster
Room   Hall 1
Near Real-Time Simulations of Global CMT Earthquakes
BOZDAG, E., Princeton University, Princeton, NJ, bozdag@princeton.edu; TROMP, J., Princeton University, Princeton, NJ, jtromp@princeton.edu; KOMATITSCH, D., CNRS/University of Aix-Marseille, Marseille, France, komatitsch@lma.cnrs-mrs.fr; HJORLEIFSDOTTIR, V., Universidad Nacional Autonoma de Mexico, Mexico City, Mexico, vala@geofisica.unam.mx; LIU, Q., University of Toronto, ON, Canada, liuqy@physics.utoronto.ca; ZHU, H., Princeton University, Princeton, NJ, hejunzhu@princeton.edu; PETER, D., Princeton University, Princeton, NJ, dpeter@princeton.edu; MCRITCHIE, D., Princeton University, Princeton, NJ, dmcr@princeton.edu; FRIBERG, P., Instrumental Software Technologies, Inc., New Paltz, NY, p.friberg@isti.com; TRABANT, C., IRIS DMC, Seattle, WA, chad@iris.washington.edu; HUTKO, A., IRIS DMC, Seattle, WA, alex@iris.washington.edu
We have developed a near real-time system for the simulation of global earthquakes. Prompted by a trigger from the Global Centroid Moment Tensor (CMT) Project, the system automatically calculates normal-mode synthetic seismograms for the Preliminary Reference Earth Model (PREM), and spectral element synthetic seismograms for 3D mantle model S362ANI in combination with crustal model Crust2.0. The 1D and 3D synthetics for more than 3000 seismographic stations operated by members of the international Federation of Digital Seismograph Networks are made available via the internet (global.shakemovie.princeton.edu) and the Incorporated Research Institutions for Seismology Data Management Center (IRIS; iris.edu). The record length of the synthetics is 100 min for CMT events with magnitudes less than 7.5, capturing R1 and G1 at all epicentral distances, and 200 min for CMT events with magnitudes equal to or greater than 7.5, capturing R2 and G2. The mode simulations are accurate at periods of 8 s and longer, whereas the spectral-element simulations are accurate between periods from 17 to 500 s. The spectral-element software incorporates a number of recent improvements, for example, the mesh honors the Moho as a first-order discontinuity underneath the oceans and continents, and the performance of the solver is enhanced by reducing processor cache misses and optimizing matrix–matrix multiplication. In addition to synthetic seismograms, the system produces a number of earthquake animations, as well as various record sections comparing simulated and observed seismograms. Together with near real-time simulations, synthetic seismograms for old earthquakes are also being computed, ultimately to have 1D and 3D synthetics for all earthquakes in the CMT catalogue.
Session: What are the Limits of Explosion Source Model Predictions?
Presenter   Taylor, Steven
Schedule   Wed 11:45 AM / Oral
Room   155D
Can Teleseismic mb be Affected by Rock Damage Around Explosions?
TAYLOR, S. R., Rocky Mountain Geophysics, Los Alamos, NM, srt-rmg@comcast.net; PATTON, H. J., Los Alamos National Laboratory, Los Alamos, NM, patton@lanl.gov
Effects of rock damage on teleseismic mb are investigated with P-wave synthetic seismograms using a moment dipole Mzz as the equivalent elastic model for damage around buried explosions. Two manifestations of late-time damage, cavity rebound and bulking from block rotations, are represented by model decompositions into compensated linear vector dipole and monopole sources, respectively. For high-velocity media, P waves from damage destructively interfere with those from the explosion. This interference reduces the rate at which mb yield scales for a pure monopole source, and provides a physical basis for observed scaling in hard rock, mb~0.75log[yield]. For over-buried explosions, such as the North Korean tests, P waves from damage are weaker, and higher scaling rates are expected than explosions conducted under standard containment conditions. Our results highlight a cautionary note of transporting the same mb-log[yield] relation between test sites to estimate yield when source phenomenology is likely to be very different.
Session: New Frontiers in Seismic Data Analysis
Presenter   Kijko, Andrzej
Schedule   Wed 5:30 PM / Oral
Room   155A
Flaw in the Bayesian Procedure for Estimation of the Maximum Regional Earthquake Magnitude
KIJKO, A., University of Pretoria, Hatfield, Pretoria, South Africa
The Bayesian procedure for the estimation of the regional maximum possible earthquake magnitude mmax will be the focus of this presentation. It will be shown that the currently used Bayesian procedure for mmax, as developed by Cornell (Cornell, 1994; EPRI NP-4726), has a mathematical flaw, and a statistically justifiable alternative approach is suggested. The fundamental problem in the application of the current Bayesian formalism for mmax estimation is that one of the components of the posterior distribution is the sample likelihood function, for which the range of observations (earthquake magnitudes) depends on the unknown parameter mmax. This dependence violates the property of regularity of the maximum likelihood function. The resulting likelihood function therefore reaches its maximum at the maximum observed earthquake magnitude and not at the required maximum possible magnitude mmax. Since the sample likelihood function is a key component of the posterior distribution, the posterior estimate of mmax is biased. The degree of bias and its sign depend on the applied Bayesian estimator, the quantity of information provided by the prior distribution and the sample likelihood function. It will be shown that if the maximum posterior estimate is used, the bias is negative and the resulting underestimation of mmax can be as big as 0.5 unit of magnitude. This study explores only the maximum posterior estimate of mmax, which is conceptionally close to the classic maximum likelihood estimation. However, conclusions regarding the shortfall of the current Bayesian procedure are applicable to all Bayesian estimators, e.g. posterior mean and posterior median. A simple, ad hoc solution of this non-regular maximum likelihood problem is also presented.
Session: ShakeMap-Related Research, Development, Operations, and Applications
Presenter   Wald, David
Schedule   Thu PM / Poster
Room   Hall 1
ShakeCast: What’s New?
LIN, K. W., USGS, Golden, CO, klin@usgs.gov; WALD, D. J., USGS, Golden, CO, wald@usgs.gov; LAWALL, T. E., Cherokee Services Group, LLC (CSG), Fort Collins, CO, lawallt@usgs.gov
The U.S. Geological Survey's (USGS) ShakeCast® system, an earthquake response tool of the Advanced National Seismic System (ANSS), is a derivative product of the ShakeMap program. ShakeCast automatically retrieves earthquake shaking data from ShakeMap, compares intensity measures against users’ facilities, sends notifications of potential damage to responsible parties, and generates facility damage assessment maps and products for emergency managers and responders. Since its inception in 2004, ShakeCast has consistently evolved to match both ShakeMap products and data interpretations. Primary users of the system now represent government, business, transportation, critical lifeline, utility, emergency management, and international community sectors. To facilitate the use of ShakeCast, many of the recommended improvements come from several prominent users, including Caltrans, the U. S. Nuclear Regulatory Commission (NRC) and the International Atomic Energy Agency (IAEA). We present the current status of ShakeCast and provide updates on new capabilities and directions. We release the new version 3 of the ShakeCast program and manual, which marks a major advancement regarding the overall system functionality and its applications. In addition to the full probabilistic interpretation feature for HAZUS-styled fragility functions, the new data analysis interface allows ShakeCast users to incorporate custom computations as part of their post-earthquake response protocols. Established modules include calculation of Generation 2 Fragility (g2F) for Caltrans bridges, regulatory level exceedance for nuclear power plants, and landslide and liquefaction likelihoods. We also describe the new ShakeCast user interface and its backend programming interface, which comes with full Representational State Transfer (REST) architecture support. It adapts common web standards for data exchange and permits transactions between ShakeCast and users’ in-house emergency management systems.
Session: Advances in High-Frequency Physics-Based Earthquake Systems Simulation
Presenter   Olsen, Kim
Schedule   Thu 9:00 AM / Oral
Room   155C
Ground Motion Prediction From Low-velocity Sediments Including Statistical Models of Inhomogeneities in Southern California Basins
OLSEN, K. B., San Diego State University, San Diego, CA, kbolsen@mail.sdsu.edu; SAVRAN, W., San Diego State University, San Diego, CA, wsavran@gmail.com; JACOBSEN, B. H., University of Aarhus, Aarhus, Denmark, bo@geo.au.dk
As ever-increasing computational resources allow earthquake scientists to push the frequency limits of deterministic ground motion estimates higher, understanding small-scale, near-surface heterogeneities becomes paramount. These small-scale heterogeneities may significantly affect ground motion in geologic basins, and are not included in state-of-the-art Community Velocity Models (CVMs). Toward characterizing the variability of shallow sediment amplification, we have collected readily available near-surface velocity data, including several datasets of direct and indirect Vs30 values (775 measurements) and downhole sonic logs from oil exploration surveys (687 measurements) in the Los Angeles area. We have prepared semi-variograms of the near-surface heterogeneities in an effort to quantify statistical parameters characterizing the existing shear-wave velocities in Los Angeles Basin. Our preliminary analysis suggests that the data are conforming to a fractal distribution with fractal dimensions of 1.5-1.8 and related Hurst exponents of 0.2-0.5. We then generate statistical models of seismic velocities and densities in agreement with the results of the data analysis and integrate these models into the most recent SCEC CVMs. The effects of the near-surface heterogeneities on ground motion and scattering are tested using simulations of the 2008 Mw 5.4 Chino Hills, CA earthquake in a subset of the SCEC CVM 4.0, with frequencies up to 5 Hz. We assume a minimum Vs of 200 m/s using a highly scalable finite-difference code (AWP-ODC). Preliminary results indicate a trade-off between the statistical model (in terms of scattering Q) with anelastic attenuation, and that the currently employed sediment Q models may need to be tuned (toward higher Q values) for models including the near-surface heterogeneities. The ratios of peak ground velocities between models with and without the near-surface heterogeneities (both including the Q model) vary between ~0.5 and 2.
Session: Intermountain West Earthquake and Volcano Characterization and Related Hazards
Presenter   Doser, Diane
Schedule   Wed PM / Poster
Room   Hall 1
Determining Historic Earthquake Locations in the Rio Grande Rift Region Using Intensity Information
DOSER, D. I., University of Texas, El Paso, TX, doser@utep.edu; GALVAN, P., University of Texas, El Paso, TX
We have used intensity information from well located, recent earthquakes occurring within the Rio Grande rift and surrounding regions to develop an intensity-distance attenuation model using calibration events from west Texas, New Mexico, eastern Arizona and southern Colorado. This attenuation model lies between those developed for the Basin and Range province by Bakun (2006) and eastern North America by Bakun and Hopper (2004) with MMI = 1.86±4.05 + 1.52±0.77 M - 0.0013±0.0025 D – 2.71±0.61 log D, where MMI is Modified Mercalli Intensity, M is magnitude, and D is epicentral distance. We then used this new model to relocate historic earthquakes based on their intensity observations. We obtained new intensity observations for a number of events by searching newspaper archives from the U.S. and Mexico. We tested our new attenuation model by relocating the 1931 Valentine, Texas earthquake. We found that there were two minima for the intensity location due to the lack of intensity information for points in Mexico. The epicenter minimum located in Texas was comparable to the instrumental epicenter and the estimated intensity magnitude of 6.4 at this location is comparable to its moment-magnitude of 6.3. Intensity centers for the July 12 and 16, 1906 central New Mexico earthquakes are located just north of the town of Magdalena with intensity magnitudes of 5.7 and 5.9, respectively. We are continuing our analysis of other ~15 other events located within the Rio Grande rift, easternmost Colorado Plateau and westernmost Great Plains.
Session: Next Generation of Ground Motion Prediction Models
Presenter   Ktenidou, Olga
Schedule   Wed 5:15 PM / Oral
Room   155C
Kappa Needs a Subscript!
KTENIDOU, O. J., Université Joseph Fourier, CNRS, ISTerre, Grenoble, France, olga.ktenidou@ujf-grenoble.fr; COTTON, F., Université Joseph Fourier, CNRS, ISTerre, Grenoble, France, fabrice.cotton@ujf-grenoble.fr; ABRAHAMSON, N., Pacific Gas & Electric, San Francisco, CA, abrahamson@berkeley.edu; ANDERSON, J. G., Nevada Seismological Laboratory, Reno, NV, jga.seismo@gmail.com
The high-frequency attenuation parameter κ (kappa), coined by Anderson and Hough (1984), is one of the most used yet least understood or agreed-upon parameters in Engineering Seismology. To date κ constitutes an important input parameter for several applications, yet it is measured, interpreted and used in many different ways. In terms of interpretation, the debate is still on as to whether κ is due to site, path, and/or source effects. But more importantly, even in applications where κ is used to represent site effects (such as ground motion simulations and GMPE adjustments from host to target regions), it is measured using a variety of approaches: on the high-frequency part of the acceleration Fourier spectrum, on the low-frequency part of the displacement spectrum, on the site transfer function, from fittings to the acceleration response spectrum, from broadband inversions, or using the inverse random vibration theory to derive GMPE-response-spectrum-compatible Fourier spectra. We suggest that in order to reduce the ambiguities, κ should always be given a notation consistent with the approach used to measure it. We propose such a notation system and suggest that certain families of measurement approaches may be more compatible with certain applications.
Session: Next Generation of Ground Motion Prediction Models
Presenter   Petersen, Mark
Schedule   Wed 11:15 AM / Oral
Room   155C
Evaluation of Ground-Motion Models for the 2014 USGS National Seismic Hazard Maps
PETERSEN, M., USGS, Golden, CO, mpetersen@usgs.gov; REZAEIAN, S., USGS, Golden, CO, srazaeian@usgs.gov; MOSCHETTI, M., USGS, Golden, CO, mmoschetti@usgs.gov; POWERS, P., USGS, Golden, CO, pmpowers@usgs.gov; MUELLER, C., USGS, Golden, CO, cmueller@usgs.gov; FRANKEL, A., USGS, Golden, CO, afrankel@usgs.gov; LUCO, N., USGS, Golden, CO, nluco@usgs.gov; HARMSEN, S., USGS, Golden, CO, harmsen@usgs.gov
The U. S. Geological Survey’s National Seismic Hazard Mapping Project (NSHMP) is updating the 2008 maps for consideration by the engineering community in developing future seismic design provisions. These maps need to be based on published information that is generally accepted by the hazard community as “best available science”. A public workshop was held December, 2012 in Berkeley, CA to discuss new ground motion prediction equations and their impact on seismic hazard across the U.S. New equations are available for earthquake sources in the Central and Eastern U.S. (CEUS), subduction and deep intraslab earthquakes, and shallow crustal earthquakes in the Western U.S. (WUS). Preliminary assessments indicate that these new equations produce significant differences in hazard compared to the 2008 maps. For the CEUS, new seismic data was interpreted as supporting alternative geometric spreading and high frequency site-attenuation (kappa) models. Uncertainties in these two parameters can cause significant differences in the ground motions for distances within 80 km compared to models applied in 2008. For great subduction-zone earthquakes, new equations result in higher ground motions over the coastal Pacific NW but similar or lower contributions elsewhere. Ground motion models for intraslab earthquakes are similar to previous models applied in 2008. New WUS equations are still preliminary and are based on a new ground motion database that includes significant additional shaking data, especially for moderate-size earthquakes. Several of the preliminary NGA models and other new equations result in generally higher ground motions over much of coastal CA and the Intermountain West regions. Implications to the hazard from applying these new equations are being evaluated and the NSHMP will release draft maps for public comment in mid 2013.
Session: Induced Seismicity
Presenter   Boltz, Meagan
Schedule   Fri AM / Poster
Room   Hall 1
STUDENT
Analysis of Mining-Induced Seismicity at Central Utah Coal Mines
BOLTZ, M. S., University of Utah Department of Mining Engineering, Salt Lake City, UT, m.boltz@live.com; KUBACKI, T. M., University of Utah Department of Mining Engineering, Salt Lake City, UT, texmk1@gmail.com; CHAMBERS, D. J. A., University of Utah Department of Mining Engineering, Salt Lake City, UT, derrick.ja.chambers@gmail.com; WHIDDEN, K. M., University of Utah Seismograph Stations, Salt Lake City, UT, whidden@seis.utah.edu; PANKOW, K. L., University of Utah Seismograph Stations, Salt Lake City, UT, pankow@seis.utah.edu; KOPER, K. D., University of Utah Seismograph Stations, Salt Lake City, UT, koper@seis.utah.edu; MCCARTER, M. K., University of Utah Department of Mining Engineering, Salt Lake City, UT, k.mccarter@utah.edu
The University of Utah Seismograph Stations and Mining Engineering departments are conducting research to improve knowledge of mining-induced seismicity (MIS) related to underground coal mines in central Utah. Ultimately, the goal of this research is to improve underground mine safety. Here we report on four active projects. (1) In relocating MIS generated at the Trail Mountain Mine during 2000 and 2001 using enhanced (master event and double difference) relocation techniques, we find the MIS locates in the mine roof. We also find an increase in the number of seismic events and a migration in location from the edges of the long wall panel across the entire panel width, corresponding with geology changes in the roof and floor of the mine. A finite difference model is being developed to examine the relationship between MIS and near-seam geology. (2) On 6 August 2007, the Crandall Canyon Mine experienced a catastrophic collapse (Mw 4.1) which trapped six workers. Cross-correlation detection techniques are used to uncover 524 small events that occurred on the day of the collapse, as well as 466 events in the preceding week. These detections are made at a completeness magnitude of -0.5 Mc and indicate a p-value of 0.71. This study offers a more complete understanding of mine-collapse aftershock sequences. (3) The existence of non-mining related (i.e., tectonic) seismicity within 15 km of active mining in the southern Wasatch Plateau is investigated after several large events (ML>3.0) were recorded. Events in the region are relocated using hypoDD. Waveforms, depths, and magnitudes are examined to discriminate events caused by mining from events that occur naturally. (4) Full moment tensors, including the isotropic component, are routinely calculated for earthquakes in Utah and surrounding areas for ML>3.5 seismic events. Efforts are underway to investigate methods for determining full moment tensors for smaller MIS events.
Session: Characterizing Active Faults for Seismic Hazard Assessments
Presenter   Cramer, Chris
Schedule   Thu 11:00 AM / Oral
Room   155B
Magnitude Estimates of M7.3–7.8 for the 1811–1812 New Madrid and M7.0 for the 1886 Charleston Earthquakes from a Monte Carlo Analysis of Mean MMIs
CRAMER, C. H., CERI, University of Memphis, Memphis, TN, ccramer@memphis.edu; BOYD, O. S., U.S. Geological Survey, Golden, CO, olboyd@usgs.gov
The earthquake-intensity observations, modified Mercalli intensities—MMI, and assigned magnitudes, M, of the 1811–1812 New Madrid events have been revised several times in the last decade. Evernden (1975) suggests that the close-in, higher intensity values are influenced by magnitude, source depth, and other factors, while the distant, lower intensity values are influenced mostly by the magnitude of the earthquake. We estimate magnitudes from mean MMIs of the 1811–1812 New Madrid and 1886 Charleston earthquakes for the distance range 800–1200 km using the relation ΔM = 0.63*ΔMMI, derived from the ground motion intensity relations of Dangkua and Cramer (2011) and the definition of magnitude from amplitude in Richter (1958). With two reference earthquakes of known magnitude—the 1929 M7.2 Grand Banks and 2001 M7.6 Bhuj, India earthquakes— we apply a Monte Carlo analysis using a logic tree having four branches representing alternatives of Reference Earthquake, Reference Earthquake Magnitude, Bhuj Mean MMI, and Historic Mean MMI. We estimate magnitudes of M7.5, M7.3, and M7.8 for the three New Madrid 1811–1812 mainshocks, M6.8 for the largest aftershock, and M7.0 for the 1886 Charleston, South Carolina earthquake. As a check on our method, we find that the magnitude of one reference earthquake is correctly estimated from the other reference earthquake. We demonstrate that the 1929 Grand Banks earthquake occurred in eastern North America (ENA) crust and had typical ENA MMIs. We also adjusted the Bhuj earthquake MMIs for differences in intensity attenuation between cratonic India and ENA. The error in the estimate of the mean MMIs for each earthquake’s alternative MMI datasets provide a logic tree estimate of magnitude uncertainty for the historic earthquakes of 0.2 to 0.4 magnitude units (95% confidence level). This variability comes primarily from the uncertainty in the mean MMI estimates, the reference earthquake used, and the alternative MMI datasets.
Session: Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Models
Presenter   Sleep, Norman
Schedule   Fri 11:00 AM / Oral
Room   155C
Inelastic Response Within Sedimentary Basins During the Passage of Strong Love Waves
SLEEP, N. H., Stanford University, Stanford, CA, norm@stanford.edu; ERICKSON, B. A., Stanford University, Stanford, CA, brittanya.erickson@gmail.com
Major earthquakes on strike-slip faults produce reverberating Love waves within sedimentary basins, e.g., the Los Angeles area. Sufficiently strong waves produce frictional failure in the upper few 100 meters, nonlinearly attenuating the wave. These waves attenuate nonlinearly along their full paths. In contrast, ascending body waves typically become nonlinear only just before they reach the surface. Simple scaling relations are available to relate peak ground velocity (PGV) and peak ground acceleration (PGA) to dynamic stress and nonlinear attenuation. These relationships can be appraised with numerical calculations and by comparison with seismograms. Dynamic strain and stress for Love waves scales with PGV and the net effect clips the wave on velocity at its dominant period (3-4 sec for Los Angeles) on a displacement seismogram. Conversely, shallow rock within basins is a fragile geological feature. Each seismic episode damages the rock reducing its shear modulus. Waves impose dynamic strain (particle velocity/phase velocity). Frictional yield stress is proportional to depth times the effective coefficient of friction. Each shaking episode cracks shallow rock so that the shear modulus decreases over repeated earthquakes cycles. Quasi-steady state is approached when typical imposed dynamic strains produce shear stresses that barely cause frictional failure. Then, the shear modulus then increases linearly with depth with the shallow region that becomes nonlinear. Calibration of rock strength is available from S-waves refracted into near vertical paths. The dynamic acceleration in g’s at failure is the effective coefficient of friction. The velocity seismogram is sawtoothed at its dominant period. The effective coefficient of friction at Cholame 2WA for the 2004 Parkfield mainshock is ~0.4, appropriate for a shallow water table. Attention needs to be given for the depth to the water table near Los Angeles in 1857 as well as for future events on the San Andreas Fault.
Session: Triggering of Seismic and Volcanic Events
Presenter   Hong, Tae-Kyung
Schedule   Fri 9:30 AM / Oral
Room   155D
Intraplate Seismicity Around the Korean Peninsula and its Response to the 2011 Tohoku-Oki Earthquake
HONG, T. K., Yonsei University, Department of Earth System Sciences, Seoul, South Korea, tkhong@yonsei.ac.kr; TAHIR, M., Yonsei University, Department of Earth System Sciences, Seoul, South Korea; HOUNG, S. E., Yonsei University, Department of Earth System Sciences, Seoul, South Korea
The Korean Peninsula is located in the far-eastern Eurasia plate, and belongs to intraplate seismicity regime with low seismicity. The instrumental seismicity is scattered over the peninsula, with relatively high seismicity at paleo-tectonic regions that are associated with paleo-rifting and paleo-collision. The paleo-rifting is responsible for the separation of Japanese islands from the Eurasian plate. The paleo-continental collisions formed a current shape of Korean Peninsula. It is known that dozens of devastating earthquakes with magnitudes of 7 occurred historically. The paleo-structures appear to be reactivated by the current ambient stress field. The Korean Peninsula is located at about 1200 km away from the 11 March 2011 M9.0 Tohoku-Oki earthquake. Nine triggered earthquakes within one hour after the Tohoku-Oki earthquake were identified in the Korean Peninsula. The sizes of triggered earthquakes are small, allowing observation only at stations in nearby distances. The short PS differential times suggests that the earthquakes appear to occur by brittle failure in shallow crust. The focal mechanism solution of the largest triggered earthquake presents normal-faulting sense of motion with a tensional axis in the southeast, suggesting that the event occurred due to the response to the displacement caused by the Tohoku-Oki earthquake. We present the spatio-temporal evolution of the Korean seismicity before and after the Tohoku-Oki earthquake. The response of intraplate regime to the rapidly sweeping dynamic stress field is discussed.
Session: Seismic Hazards and Ground Motions
Presenter   Ebel, John
Schedule   Fri 11:15 AM / Oral
Room   155B
The Hazard of Offshore Earthquakes and Tsunamis along the U.S. East Coast
EBEL, J. E., Weston Observatory, Boston College, Weston, MA, ebel@bc.edu
A swarm of earthquakes at the edge of the continental shelf east of New England in 2012 raised questions about the potential of strong offshore earthquakes and concomitant tsunamis along the U.S. east coast. The 2012 swarm was comprised of 15 earthquakes, 12 of which took place on April 12, centered about 280 km east of Boston, MA. The largest event in the swarm had MLg 4.0, and five of the events had magnitudes above MLg 3.0. The tectonic setting of this swarm at the edge of the continental shelf is very similar to that of the 1929 M 7.3 Grand Banks earthquake, which generated a 10-m tsunami along part of the coast of southern Newfoundland. Curiously, in 2012 several other earthquakes were detected at other locations along the edge of the Atlantic continental shelf of North America: MLg 3.5 on April 9 in the epicentral area of the 1929 earthquake; MLg 2.2 on April 30 about 200 km NE of the April 12 swarm area; and MLg 2.8 on June 12 about 250 km southeast of Cape Cod, MA. All of these areas have experienced earthquake activity during the past few decades prior to 2012. In contrast to this activity along the edge of the continental margin of the northeastern U.S. and eastern Canada, there have been no earthquakes detected since 1973 along the edge of the continental margin of the southeastern U.S. One MLg 3.7 earthquake was detected in 1990 about 150 km east of the continental slope off North Carolina, and one MLg 3.3 earthquake was detected in 2001 on the continental shelf east of the northernmost part of Florida. These observations suggest that the greatest threat of a strong earthquake and tsunami, approaching or exceeding that in 1929, probably exists from south of Long Island, NY northeast off the New England and Nova Scotia coasts. There appears to be less possibility of such an earthquake and tsunami along the southeast U.S. coast from southern New Jersey to Florida.
Session: Intermountain West Earthquake and Volcano Characterization and Related Hazards
Presenter   Shelly (video), David
Schedule   Wed 11:15 AM / Oral
Room   155B
Migrating Activity of the 2010 Madison Plateau, Yellowstone National Park, Earthquake Swarm: Evidence for Fluid Triggering?
SHELLY, D. R., U.S. Geological Survey, Menlo Park, CA, dshelly@usgs.gov; HILL, D. P., U.S. Geological Survey, Menlo Park, CA, hill@usgs.gov; MASSIN, F., BRGM-Guadeloupe Observatory, Guadeloupe, France, fred.massin@gmail.com; FARRELL, J., University of Utah, Salt Lake City, UT; SMITH, R. B., University of Utah, Salt Lake City, UT
Yellowstone National Park includes one of Earth’s largest volcanic-hydrothermal fields and a youthful caldera. The 2010 Madison Plateau earthquake swarm, near the northwest rim of Yellowstone caldera, was one of the three largest swarms recorded in the park since monitoring began in the 1970s. This swarm lasted from January 15 to February 6 and consisted of ~2200 cataloged events, including 17 M 3+ events, with the largest being M 3.9. To fully explore the activity and evolution of the swarm, we combine waveform-based event detection with precise double-difference relative relocation. Detection and location goals are accomplished in tandem, using cross-correlation with continuous seismic data and waveform templates constructed from cataloged events. Using this procedure, we detected ~8700 events that can be precisely located, a factor of ~4 more events than included in the Yellowstone Seismic Network standard catalog. These newly detected and relocated hypocenters reveal distinct migration of activity. Most events are located on a NNW-striking, ENE-dipping structure, with dimensions of approximately 3 by 3 km, between 8 and 11 km depth. Activity initiated abruptly at about 10 km depth, and expanded systematically outward (both shallower and deeper) along this structure over time. We hypothesize that the swarm activity may have been triggered by the rupture of a confined high-pressure fluid system into neighboring pre-existing crustal fractures. Double-couple-constrained fault solutions suggest strike-slip faulting on this dipping structure, which may be a reactivated Basin and Range normal fault. While strike-slip motion on a dipping fault is not mechanically optimal, it could occur if the fault is sufficiently weak relative to its surroundings, perhaps as a result of locally high fluid pressure.
Session: Next Generation of Ground Motion Prediction Models
Presenter   Ogweno, Luke Philip
Schedule   Thu AM / Poster
Room   Hall 1
STUDENT
Comparison of Current ENA GMPEs with the NGA East Ground Motion Database using Model Bias Residual Analysis
OGWENO, L. P., CERI-The University of Memphis, Memphis, TN, logweno@memphis.edu; CRAMER, C. H., CERI-The University of Memphis, Memphis, TN, ccramer@memphis.edu; ALNOMAN, M. N., CERI-The University of Memphis, Memphis, TN, malnoman@memphis.edu
The Next Generation Attenuation (NGA) East project has developed an updated database of eastern North America (ENA) ground motions (Cramer et al., 2009, 2010, 2012) containing over 11,000 records and covering distance and magnitude ranges of 1-3500 km and M 2.2-7.6, but mostly less than M6.0. The database was categorized into three classes depending on the soil condition at the site. Using model bias [log(Calc) – log(Obs)] residual analysis we compared the NGA East observations to current ENA ground motion predication equations (GPMEs). For each GPME, we binned the residuals into distance and magnitudes bins and then average the residuals in each bin and plot the averages for each soil site class respectively. We mainly compared the GMPEs used in the USGS national seismic hazard maps (Petersen et al., 2008), but also review more recent GMPEs and the EPRI (2004) GMPEs. Initial comparisons have been made at peak ground acceleration (PGA) and at 0.2 s and 1.0 s spectral accelerations (Sa).In general newer GMPEs tend to predict lower ground motion levels than older GMPEs. This appears to be associated with the geometrical spreading used in the GMPE (the newer R-1.3 vs. the older R-1 models). Some GMPEs only predict ground motions well out to 500 km while others do well out to 1000 km, mostly in keeping with their stated valid distance ranges. Model bias results are period dependent with newer GMPEs showing better predictions at short periods and older GMPEs showing better predictions at 1 s. We list the mean model bias for each current GMPE for distances less than 100 km at each period and show an average model bias across all periods.
Session: Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Models
Presenter   Assatourians, Karen
Schedule   Wed AM / Poster
Room   Hall 1
Validation Study of Stochastic Finite Fault Module (EXSIM) Implemented in SCEC Broadband Platform
ASSATOURIANS, K., Western University, London, ON, Canada, kassatou@uwo.ca; ATKINSON, G. M., Western University, London, ON, Canada, gatkins6@uwo.ca
The stochastic finite fault (SFF) algorithm EXSIM, as developed by Motazedian and Atkinson (2005) and Boore (2009), is among the alternative models implemented on the Southern California Earthquake Center (SCEC) broadband earthquake ground motion simulation platform. Proper implementation of this module requires a parameter validation study. The validation is carried out by comparing pseudo-response spectral (PSA) amplitudes of simulated motions to the corresponding values for four well-recorded earthquakes in North America. These earthquakes include the M5.89 1987 Whittier Narrows, M6.94 1989 Loma Prieta, M6.73 1994 Northridge earthquakes in the west, and the M5.81 1988 Saguenay in the east. Results show that the SFF is not strongly sensitive to the slip model; the main source parameter of interest for validation is the stress parameter. This conclusion is in accord with previous studies (Motazedian and Atkinson, 2005).In the past few years a number of modifications and conceptual changes have been applied to the EXSIM algorithm (e.g. Boore, 2009). Although these changes are improvements overall, they mean that current results obtained with EXSIM are not directly comparable to previous results, in terms of parameter values required to match simulations to observations. Thus we need to recalibrate EXSIM, for both eastern and western earthquakes, as part of the validation studies needed to implement EXSIM on the broadband platform. We report on that recalibration exercise, including conclusions regarding the following parameters: crustal amplification; geometrical and inelastic attenuation; high-frequency diminution (kappa); ground motion duration; and average stress parameter.
Session: Next Generation of Ground Motion Prediction Models
Presenter   Cramer, Chris
Schedule   Wed 2:30 PM / Oral
Room   155C
Investigating Q Transitions and Regional Q(f) Using NGA East and USARRAY Data
CRAMER, C. H., CERI, University of Memphis, Memphis, TN, ccramer@memphis.edu; AL NOMAN, M. N., CERI, University of Memphis, Memphis, TN, malnoman@memphis.edu
The NGA East ground motion database and EarthScope USArray data provide an excellent opportunity to improve the scientific understanding of crustal attenuation in the continental US. The purpose of this study is to explore the major seismic Q transition boundaries using NGA East and USArray observations. A clearer definition of the location and sharpness of these Q boundaries will improve seismic hazard evaluations such as the USGS national seismic hazard maps. We have performed an initial evaluation of the location and character of Q transitions for (1) the WUS-CEUS transition in the Rocky Mountains and Great Plains between 100°W and 115°W and (2) the CEUS mid-continent to Gulf Coast transition. We collected the regional waveform data and followed the standard procedure of NGA East ground motion database project for processing those observations (Cramer et al, 2009, 2010). We also used data from the NGA East ground motion database (Cramer et al., 2012). We then analyzed the transect data radially from the source along specific azimuths to define the transition-of-Q boundary and come up with an initial estimate of apparent Q on either side of each transition region. Our preliminary results suggest a strong, sharp transition between WUS and CEUS located near 113°W in Montana, near 110°W in Utah, and near 108°W near Texas. A similar transition between CEUS and the Gulf Coast in Oklahoma and Arkansas is near 35°N. The apparent Qo estimate for the WUS and Gulf Coast region is found to be ~ 227 and ~185 respectively, while the CEUS mid continent has a much lower attenuating Qo of ~ 670. These initial estimates use spectral acceleration values instead of narrow bandpass filtered amplitudes centered on a set of frequencies. Improved estimates of Q(f) using the approach of Benz et al. (1997) and Erickson et al. (2004) will be calculated in the near future.
Session: Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Models
Presenter   Kamai, Ronnie
Schedule   Fri 1:45 PM / Oral
Room   155C
Adding Fling Effects to Processed Ground Motion Time Histories
KAMAI, R., Pacific Earthquake Engineering Research Center, Berkeley, CA, rkamai@berkeley.edu; ABRAHAMSON, N. A., Pacific Gas & Electric Company, San Francisco, CA, naa2@pge.com; GRAVES, R. W., U.S. Geological Survey, Pasadena, CA, rwgraves@usgs.gov
Fling is the representation of the permanent tectonic offset in recorded ground motions, expressed by a one-sided pulse in ground velocity and a non-zero final displacement. Standard processing of earthquake time histories includes application of high-pass filters and baseline correction, which together remove the permanent displacement from the time history, but retain some of the fling effects at intermediate periods. Abrahamson (2001) developed a method to parameterize the fling acceleration time history by a single cycle of a sine wave. Design ground motions including fling effects were then developed by superimposing the expected fling acceleration time history onto the design time history. A short-coming of this method is it double counts some of the fling effects that are not removed by the filtering. An improved method for developing design ground motions with fling effects is presented. First, an updated parametric model for the fling step components is developed which is based on an extensive set of finite-fault simulations conducted on the SCEC broadband platform rather than on sparse empirical data. The simulations used the Graves and Pitarka (2010) hybrid simulation method and included strike-slip and reverse scenarios for magnitudes of 6.0 to 7.8, and dips of 30 through 90. Next, a procedure is developed to superimpose the fling time history onto the traditional time history without double counting the part that was not removed by the processing. In this procedure, the fling time history given by the parametric fling model is first processed with a matching filter before being superimposed on the design time history. The proposed procedure is validated against a set of 85 empirical ground motions containing fling.
Session: Advances in High-Frequency Physics-Based Earthquake Systems Simulation
Presenter   Galis, Martin
Schedule   Wed PM / Poster
Room   Hall 1
Verification of Seismic Wave-Propagation Codes in 3D Random Scattering Media
GALIS, M., KAUST - King Abdullah University of Science & Technology, Thuwal, Saudi Arabia, martin.galis@kaust.edu.sa; IMPERATORI, W., KAUST - King Abdullah University of Science & Technology, Thuwal, Saudi Arabia, walter.imperatori@kaust.edu.sa; MAI, P. M., KAUST - King Abdullah University of Science & Technology, Thuwal, Saudi Arabia, martin.mai@kaust.edu.sa
In recent years the fast progress in High-Performance Computing, including efficient optimization of numerical codes on petascale supercomputers, has permitted the simulation of 3D seismic wave propagation at frequencies of engineering interest (up to 10Hz) in highly heterogeneous media (e.g. Hartzell et al., 2010; Imperatori and Mai, 2013). Several verification exercises (e.g. Day et al., 2001; Bielak et al., 2010, Chaljub et al., 2010) have demonstrated the importance of assessing the accuracy of available numerical tools already at low frequency in presence of large-scale features (basins, topography, etc.). However, high frequency numerical simulations involving random scattering media, characterized by small-scale heterogeneities, are much more challenging for most numerical methods, and their verification may therefore be even more crucial than in the low-frequency case. Our goal is to investigate the behavior of three different numerical codes for seismic wave propagation in 3D random scattering media at high frequency. We use two 4-th order staggered-grid finite-difference codes (FD1, see Olsen et al., 2009 and FD2, see Moczo et al., 2007). The codes are characterized by slightly different medium representations; FD1 uses point values of material parameters in each cell, FD2 uses the effective material parameters at each grid point (Moczo et al., 2002). The third code, SO, is based on 2-nd order support-operator method (Ely et al., 2008). Random velocity perturbations are described by means of a van Karman correlation function with different correlation lengths and different standard deviations. Our results show significant variability in solutions obtained by different codes, suggesting that the numerical representation of material heterogeneity in different codes exerts non-negligible effects on ground-motion simulations. The detailed quantification of these effects may need to be included into the epistemic uncertainty of ground-motion prediction.
Session: Advances in High-Frequency Physics-Based Earthquake Systems Simulation
Presenter   Lee, En-Jui
Schedule   Wed PM / Poster
Room   Hall 1
STUDENT
Full-3D Waveform Tomography for Southern California
LEE, E., Department of Geology and Geophysics, University of Wyoming, Laramie, WY, elee8@uwyo.edu; CHEN, P., Department of Geology and Geophysics, University of Wyoming, Laramie, WY; JORDAN, T. H., Department of Earth Sciences, University of Southern California, Los Angles, CA; MAECHLING, P. J., Department of Earth Sciences, University of Southern California, Los Angles, CA; DENOLLE, M., Department of Geophysics, Stanford University, Stanford, CA; BEROZA, G. C., Department of Geophysics, Stanford University, Stanford, CA
Our full-3D tomography (F3DT) uses 3D SCEC Community Velocity Model Version 4.0 (CVM4) in Southern California as initial model, a staggered-grid finite-difference code to simulate seismic wave propagation and the sensitivity (Fréchet) kernels are calculated based on the scattering integral and adjoint methods to iteratively improve the model. We use both earthquake recordings and ambient noise Green’s function data, stacking of station-to-station correlations of ambient seismic noise, in our F3DT inversions. To reduce errors of earthquake sources, the epicenters and source parameters of earthquakes used in our F3DT are inverted based on full-wave method. An automatic waveform analysis algorithm that based on continuous wavelet transforms and a topological watershed method is used to pick waveforms full-wave inversions and make frequency dependent phase and amplitude measurements. Our current model shows many features that relate to the geological structures at shallow depth and contrasting velocity values across faults. The perturbations with respect to the initial model in some regions could up to 40% and relate to some structures do not exist in the initial model, such as southern Great Valley. The waveform fittings of earthquake waveforms and ambient noise Green’s function data are both improved after iterations. The earthquake waveform misfit between observed waveforms and updated synthetic waveforms reduced more than 60% and variance of ambient noise Green’s function group velocity delay time reduce more than 80%.
Session: Characterizing Active Faults for Seismic Hazard Assessments
Presenter   Zöller, Gert
Schedule   Fri AM / Poster
Room   Hall 1
Simulated Spatio-Temporal Patterns of Seismicity in the San Jacinto Fault Zone Assimilating the Available Instrumental and Paleoseismic Data
ZÖLLER, G., University of Potsdam, Potsdam, Germany, zoeller@uni-potsdam.de; BEN-ZION, Y., University of Southern California, Los Angeles, CA, benzion@usc.edu
We investigate spatio-temporal properties of earthquake patterns in the San Jacinto fault zone (SJFZ), California, between Cajon Pass and the Superstition Hill Fault, using long record of simulated seismicity constrained by available seismological and geological data. The model provides an effective realization of a large segmented strike-slip fault zone in 3D elastic half space, with heterogeneous distribution of static friction chosen to represent several clear stepovers at the surface. The simulated synthetic catalog reproduces well the basic statistical features of the instrumental seismicity recorded at the SJFZ area since 1981 including distributions of hypocenters, magnitudes, recurrence times, and stress drops. The model also produces events larger than those included in the short instrumental record, consistent with paleoearthquakes documented at sites along the SJFZ for the last 1400 years. The general agreement between the synthetic and observed data allows us to address with the long simulated seismicity questions related to large earthquakes and expected seismic hazard. In particular, we provide estimates for the current hazard associated with large earthquakes (m>7) inferred from the long record of simulated seismicity. We also discuss possible interactions of the different fault segments.
Session: Earthquake Source Physics
Presenter   Aso, Naofumi
Schedule   Thu PM / Poster
Room   Hall 1
STUDENT
Modeling Semi-Volcanic Deep Low-Frequency Earthquakes
ASO, N., The University of Tokyo, Bunkyo, Tokyo, Japan, aso@eps.s.u-tokyo.ac.jp; IDE, S., The University of Tokyo, Bunkyo, Tokyo, Japan, ide@eps.s.u-tokyo.ac.jp; TSAI, V. C., Seismological Laboratory, California Institute of Technology, Pasadena, CA, tsai@caltech.edu
While deep tectonic low-frequency earthquakes (LFEs) on plate boundaries are thought to be thrust events [Ide et al., 2007], the mechanisms of volcanic LFEs around the Moho beneath active volcanoes have not yet been firmly established. Recently, we examined a unique class of LFEs that occur far from active volcanoes but which are otherwise similar to volcanic LFEs [Aso et al., 2011; 2012 (in minor revision)]. Since these 'semi-volcanic' LFEs occur far from active volcanoes, they may provide clues to generally explaining why LFEs occur.Previous work has estimated various focal mechanisms of (semi-) volcanic LFEs [e.g., Ukawa and Ohtake, 1987], but most of these estimates used little information (e.g., polarities) from few events. To find a regionally dominant mechanism of (semi-) volcanic LFEs using waveform inversion for many events, we focused on the semi-volcanic LFEs in eastern Shimane in western Japan, where the second-most frequent (semi-) volcanic LFEs occur in a quiet region.We analyzed 38 events using five Hi-net stations. The focal mechanisms and moment rate functions were estimated by grid search and a linear inversion, respectively. The synthetic waveforms were calculated for a 1D structure and the local site amplification effect was corrected using body waves from deep earthquakes.The moment rate functions oscillate between positive and negative values. The focal mechanism is dominated by a CLVD component for many LFEs, and its symmetry axis is parallel to the lineation formed by the source distribution.Based on these observations, we are trying to build a physical model for these LFEs that involves three steps: stress accumulation, stress release, and oscillation excitation. We suggest that a solidification process of cooling magma accumulates stress, which then causes brittle deformation unlike fault slip. This brittle deformation then may excite an oscillating body, for which a magma or partially melted basalt would be a promising candidate.
Session: Seismic Hazards and Ground Motions
Presenter   Kuehn, Nicolas
Schedule   Fri AM / Poster
Room   Hall 1
Probabilistic Testing of Hazard Curves
KUEHN, N. M., University of Potsdam, Potsdam, Germany, nico@geo.uni-potsdam.de; SCHERBAUM, F., Unversity of Potsdam, Potsdam, Germany, fs@geo.uni-potsdam.de
Within the framework of probabilistic seismic hazard analysis (PSHA), the ground-motion values expected to be generated from future earthquakes at a given location are described in terms of a random variable. In this context, a seismic hazard curve, which is the most typical result of a PSHA, is merely one way to quantify the probability (density) distribution function of this random variable. In conjunction with the logic tree framework, which allows for capturing epistemic uncertainties on different models and/or parameters, modern PSHAs typically result in distributions of hazard curves. Recently, increasing efforts have been made to test the reliability of such hazard curve distributions, using ground-motion data which were observed subsequently to the generation of the hazard models. Here, we propose to test individual hazard curves by means of the likelihood of the data for a given hazard curve. In this context we treat each hazard curve as a probability forecast of future ground-motions and evaluate their accuracy using different scoring rules. This allows to assess which of the hazard curves is well calibrated against the observed data. In a similar vein, the likelihood of the observed data can be used to update the hazard curve distribution in a Bayesian way, resulting in the posterior distribution of hazard curves given observed data. The method is illustrated for simple synthetic examples as well as for the USGS 2008 hazard curves (Powers, Petersen, pers. communication) and for a site in California where a sufficient number of ground motion records is available.
Session: Earthquake Source Studies
Presenter   Tormann, Thessa
Schedule   Wed PM / Poster
Room   Hall 1
Generic Dependence of the Frequency-Size Distribution of Earthquakes on Depth and Its Relation to the Strength Profile of the Crust
SPADA, M., Paul Scherrer Institute, Villigen PSI, Switzerland, matteo.spada@psi.ch; TORMANN, T., ETH Zurich, Zurich, Switzerland, thessa.tormann@sed.eth.ch; WIEMER, S., ETH Zurich, Zurich, Switzerland, stefan.wiemer@sed.ethz.ch; ENESCU, B., University of Tsukuba, Tsukuba, Japan, benescu@geol.tsukuba.ac.jp
We explore the idea that the relative size distribution of earthquakes, quantified using the so-called b-value, is negatively correlated with differential stress. Because the maximum possible differential stress increases linearly in the brittle upper crust, we expect to find a decrease of b with depth. We test this expectation for seven continental areas around the world, each of which is described by a regional earthquake catalog. We find a monotonic decrease in b-value between 5-15 km depth. The decrease stops near the brittle-ductile transition. We specifically focus on the high-quality catalogs of earthquakes in California to perform a sensitivity test with respect to depth uncertainty; we also estimate the probability-depth gradient for the occurrence of a target magnitude event and study the behavior of b with depth in near-, and off-fault zones. We also translate the observed b-depth gradients into b-differential stress gradients. Our findings suggest that b-values are negatively correlated with differential stress and thus have the potential to act as stress-meters in the earth’s crust.
Session: Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Models
Presenter   Dalguer, Luis
Schedule   Thu 4:00 PM / Oral
Room   155C
Upper Bound Frequency of Deterministic Synthetic Ground Motion Compatible with GMPEs Resulted from Dynamic Rupture Models
DALGUER, L. A., Swiss Seismological Service, ETH-Zurich, Zurich, Switzerland, dalguer@sed.ethz.ch; BAUMANN, C., Swiss Seismological Service, ETH-Zurich, Zurich, Switzerland, cyrill.baumann@sed.ethz.ch
The increased interest of the engineering community to use synthetic ground motion for engineering applications, particularly near the source where recorded data are sparse, puts earthquake scientists with the responsibility to provide reliable ground motion data. The recent development of deterministic physics-based numerical simulations of earthquakes, that has contributed to substantial advances in our understanding of different aspects related to the earthquake mechanism and near source ground motion, appear as suitable approaches to merge earthquake ground motions models with engineering solutions. But it is important to evaluate the reliability of this models, particularly the upper frequency at which the simulated ground motion is reliable. Current engineering practice usually use ground motion quantities estimated from empirical Ground Motion Predicting Equations (GMPEs) such as peak ground acceleration (PGA), peak ground velocity (PGV), peak ground displacement (PGD), and spectral ordinates as input to assess building response. Therefore in this paper we propose that synthetic ground motion first has to verify its compatibility with current empirical GMPEs and provides the upper bound frequency in which the simulated data is useful for engineering purpose. As an example of this verification we evaluate a suite of deterministic ground motion simulation generated by earthquake dynamic rupture models. We use 360 rupture dynamic models with moment magnitudes in the range of 5.5-7. Initial stress distribution follows stochastic distribution compatible, in statistical sense, with past earthquakes. Grid size is 100m and minimum S wave velocity is 800m/s. We show quantitatively that the upper limit of our suite of simulations is 1Hz. Up to this frequency the synthetic data are compatible with the empirical model, which means that the residuals, which are defined as the differences between observed and predicted ground motions, do fall in the range of ±sigma.
Session: New Developments in Earthquake Forecasting and Predictability Research
Presenter   Zöller, Gert
Schedule   Thu 8:45 AM / Oral
Room   155D
Estimating Maximum Magnitude in a Time Horizon from Earthquake Catalogs with Varying Magnitude of Completeness: The M9 Tohoku Earthquake
ZÖLLER, G., University of Potsdam, Potsdam, Germany, zoeller@uni-potsdam.de; HOLSCHNEIDER, M., University of Potsdam, Potsdam, Germany, hols@math.uni-potsdam.de; HAINZL, S., GFZ German Research Centre for Geosciences, Potsdam, Germany
The technique to estimate the maximum magnitude in a time window of lengthTf as introduced by Zöller et al., BSSA 103(2a) (2013) is extendedfor earthquake catalogs with varying magnitude of completeness. In particular,we consider the case that two types of catalogs are available: a historiccatalog and an instrumental catalog. This leads to competing interests withrespect to the estimation of the two parameters from the Gutenberg-Richterlaw, namely the b-value and the event rate Λ (or the a-value): the b-value isestimated most precisely from the frequently occurring small earthquakes; thetendency of small events to cluster in aftershocks, swarms, etc. violates,however, the assumption of a Poisson process that is used for the estimationof Λ. We suggest to address this conflict by replacing the Bayesianestimation of b by a point estimate inferred from instrumental seismicity anduse large magnitude events for the remainder of the study. Applying themethod to Japan, we find that with a probability of about 10% the maximumexpected magnitude during any future time interval of length Tf = 50 years is m~9. The results indicate that in the light of the available earthquake catalogs, the size of the M9 Tohoku earthquake in 2011 is not atall a surprise. Finally, for typical scenario calculations related torequirements of nuclear safety, the maximum expected magnitude will be evenhigher.
Session: Seismic Hazards and Ground Motions
Presenter   Chandra, Johanes
Schedule   Fri 4:00 PM / Oral
Room   155B
STUDENT
Testing PGV/VS as a Proxy for Detecting the Nonlinear Seismic Soil Response using Centrifuge Test
CHANDRA, J., Institut des Sciences de la Terre, Grenoble, France, johanes.chandra@ujf-grenoble.fr; GUEGUEN, P., Institut des Sciences de la Terre, Grenoble, France, philippe.gueguen@ujf-grenoble.fr; BONILLA, L. F., IFSTTAR, Paris, France, luis-fabian.bonilla-hidalgo@ifsttar.fr
Nonlinear response of soil is a key issue for seismic risk analysis, including ground motion prediction, site effects, soil-structure interaction and seismic hazard. In this paper, soil dynamic seismic response, including shear wave velocity variation and stress-strain observation are analyzed through centrifuge test that is applied to a homogeneous soil column. The variation of the shear wave velocity(according to the level of shaking and the depth) is observed by computing the instantaneous shear wave velocity. The centrifuge test was done in IFSTTAR Nantes by considering different buildings resting on the soil column to recreate the real object behavior at a scaled-down model under seismic excitation. First, seismic interferometry by deconvolution method is applied instead of traditional cross-correlation to extract the shear velocity within the soil column, by using accelerometers located at the surface and in depth. This method is first validated by synthetic models created using CYCLIC 1D open source software, before proceeding to a more complex and delicate analysis using experimental data. In both cases, nonlinear behavior of the soil is observed using the PGA-PGV/Vs relationship and nonlinear curves are formed based on instantaneous shear wave velocity computed in depth and according to three levels of seismic shaking (low, moderate, strong). Finally, an effective and relevant proxy based on the PGV/Vs ratio is obtain, related with the stress-strain condition of the soil.
Session: Data Products as Research Resources
Presenter   Storchak, Dmitry
Schedule   Fri AM / Poster
Room   Hall 1
A New ISC Service: The Event Bibliography (1900-2012)
STORCHAK, D. A., ISC, Thatcham, Berkshire, UK, dmitry@isc.ac.uk; DI GIACOMO, D., ISC, Thatcham, Berkshire, UK, domenico@isc.ac.uk; OZGO, P., ISC, Thatcham, Berkshire, UK, przemek@isc.ac.uk; VERNEY, R., ISC, Thatcham, Berkshire, UK, rebecca@isc.ac.uk; SAFRONOVA, N., ISC, Thatcham, Berkshire, UK, natalia@isc.ac.uk; HARRIS, J., ISC, Thatcham, Berkshire, UK, james@isc.ac.uk; BONDAR, I., ISC, Thatcham, Berkshire, UK, istvan@isc.ac.uk
The International Seismological Centre (ISC) maintains the definitive worldwide summary of seismic events: earthquakes, mining induced events and explosions. A small yet not insignificant proportion of these events have been closely studied by individual researchers. These studies resulted in scientific publications in peer-reviewed journals, books, PhD thesis and individual reports. Some of the greatest earthquakes were described in several hundreds of articles published over a period of several decades.The ISC recently made a significant effort in linking bibliographical references (incl. doi) with event identifiers in the ISC database. This allows us to offer a new interactive map-based web service that makes it possible to select those scientific publications related to either specific events or events in the area of interest, covering the period over 100 years. Additional search parameters such as event type, date and magnitude, or publication author, journal or date are available. Results are offered in several standard output formats.The journals included in our database encompass a variety of fields in geosciences (e.g., engineering seismology, earthquake seismology, geodesy, tectonophysics, monitoring research, tsunami, geology, hydrogeology etc.), thus making this database useful in multidisciplinary studies.The ISC event bibliography now includes over 11,000 individual publications from over 400 titles related to events between 1900 and 2012.
Session: Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Models
Presenter   Lancieri, Maria
Schedule   Fri 9:00 AM / Oral
Room   155C
Study of the Damaging Nature of the Ground Shaking as a Function of the Structural Response Model and of the Selected Damage Criteria
LANCIERI, M., IRSN, Fontenay aux Roses, France, maria.lancieri@irsn.fr; RENAULT, M., IRSN, Fontenay aux Roses, France, marine.renault-manpower@irsn.fr; PERRAULT, M., UJF, Grenoble, France, perraulm@ujf-grenoble.fr; BAUMONT, D., IRSN, Fontenay aux Roses, France, david.baumont@irsn.fr; BERGE-THIERRY, C., CEA, Saclay, France, catherine.berge-thierry@cea.fr; BONILLA, L. F., IFSTTAR, Paris, France, luis-fabian.bonilla-hildago@ifsttar.fr; GELIS, C., IRSN, Fontenay aux Roses, France, celine.gelis@irsn.fr; GUEGUEN, P., UJF, Grenoble, France, pgueg@obs.ujf-grenoble.fr
Quantifying the damage level that could be produced by an earthquake on structures and components is a fundamental part of seismic risk assessment. It requires input signal that fully characterises the ground motion complexity — including its variability — and that, at the same time, is compatible with the structural calculation needs.The pseudo-acceleration spectrum, even if largely used, gives only partial information on the seismic shaking. Real records seem to be the better choice; however their selection for engineering purposes is not straightforward. The key point is to understand how representative and how much damaging the signals are.In the present work we propose to study this problem by investigating the relationships between a given set of accelerograms and the produced damage level. We selected 753 natural records from worldwide catalogues with peak ground acceleration greater than 0.1 g. The seismic ground motion is described by a set of shaking parameters. The structural behaviour has been modelled using a 1D Takeda hysteretic model. We investigated the damage level on high code and low code buildings having empirical frequencies corresponding to around 2, 5 and 10 storeys, for a total of 6 different structures. The damage is expressed in terms of relative drift between the roof and the floor, and as dissipated energy. Because of the great amount of parameters used herein, we performed a statistical study using the principal component analysis (PCA). We also perform a standard linear regression analysis by correlating each strong motion parameter with the damage criterion. We observe that all the parameters show a slope change in correspondence of the elasticity-plasticity limit, the integral quantities are better correlated with the damage level measured on the low frequency buildings and the peak quantities are better correlated with that measured on high frequency structures.
Session: Data Products as Research Resources
Presenter   Di Giacomo, Domenico
Schedule   Fri AM / Poster
Room   Hall 1
The ISC-GEM Global Instrumental Earthquake Catalogue (1900-2009)
DI GIACOMO, D., International Seismological Centre, Thatcham, Berkshire, UK, domenico@isc.ac.uk; STORCHAK, D. A., International Seismological Centre, Thatcham, Berkshire, UK, dmitry@isc.ac.uk; BONDAR, I., International Seismological Centre, Thatcham, Berkshire, UK, istvan@isc.ac.uk; ENGDAHL, E. R., University of Colorado, Boulder, CO, bob.engdahl@colorado.edu; VILLASEÑOR, A., Institute of Earth Sciences Jaume Almera, Barcelona, Spain, antonio@ictja.csic.es; LEE, W. H. K., 862 Richardson Court, Palo Alto, CA, willielee88@gmail.com; HARRIS, J., International Seismological Centre, Thatcham, Berkshire, UK, james@isc.ac.uk; BORMANN, P., GFZ Research Centre, Potsdam, Germany, pb65@gmx.net
The ISC-GEM global catalogue represents the final product of a two-year project sponsored by the Global Earthquake Model Foundation (GEM). The catalogue is available at the ISC website (www.isc.ac.uk) and consists of some 19,000 instrumentally recorded, moderate to large, earthquakes that occurred during the 110-year period between 1900 and 2009. Hypocenters for all events in the catalogue are re-determined using uniform and rigorous location and depth determination procedures. Because of limitations in resources, time and data availability, we introduced time-varying magnitude cut-offs for the earthquakes to be included in the ISC-GEM catalogue. These are 1900-1917: Ms ≥ 7.5 worldwide, as well as a selection of shallow events (Ms ≥ 6.5) in stable continental areas; 1918-1959: Ms ≥ 6.25; and 1960-2009: Ms ≥ 5.5. Owing to the ISC-GEM location procedures and to the substantial increase in the volume of observational data used in the relocations, the ISC-GEM catalogue offers an improved view of 110 years of global seismicity of the Earth. With respect to magnitude, surface and body wave magnitudes are recalculated using original amplitude-period measurements, and each earthquake in the catalogue is characterized by either a direct value of Mw, or an Mw proxy estimate based on newly derived non-linear regressions between Ms-Mw and mb-Mw. Finally, it is shown that the ISC-GEM catalogue represents a significant improvement in terms of magnitude homogeneity.
Session: Characterizing Active Faults for Seismic Hazard Assessments
Presenter   Bohnhoff, Marco
Schedule   Thu 4:15 PM / Oral
Room   155B
An Earthquake Gap South of Istanbul
BOHNHOFF, M., GFZ Helmholtz-Centre Potsdam, Potsdam, Germany, bohnhoff@gfz-potsdam.de; BULUT, F., GFZ Helmholtz-Centre Potsdam, Potsdam, Germany, bulut@gfz-potsdam.de; DRESEN, G., GFZ Helmholtz-Centre Potsdam, Potsdam, Germany, dre@gfz-potsdam.de; MALIN, P. E., University of Auckland, Institute of Earth Science and Engineeri, Auckland, New Zealand, p.malin@auckland.ac.nz; EKEN, T., GFZ Helmholtz-Centre Potsdam, Potsdam, Germany, ekent@gfz-potsdam.de; AKTAR, M., Bogazici University, Istanbul, Turkey, aktar@boun.edu.tr
The North Anatolian Fault Zone in Turkey has produced a remarkable sequence of westward propagating large (M>7) earthquakes throughout the last century, leaving the Sea of Marmara segment close to Istanbul as the only part of the entire fault zone that has not been activated since 1766. We investigate microseismicity along the eastern part of the Marmara segment at the transition to the 1999 Izmit rupture using microseismicity observed by a nearby seismic array on the Princes Islands offshore Istanbul and selected regional seismographs during a four-year period. We identify a 30 km long fault segment below the eastern Sea of Marmara in direct vicinity to the Istanbul population center that is entirely aseismic down to 10 km depth. This fault patch size is interpreted to be locked and seen as a potential nucleation point of the pending up to M7.4 Marmara earthquake.
Session: Advances in High-Frequency Physics-Based Earthquake Systems Simulation
Presenter   Mai, P. Martin
Schedule   Thu 10:45 AM / Oral
Room   155C
Broadband Near-Field Ground Motion Simulations in 3D Scattering Media
IMPERATORI, W., Kaust, Thuwal, Kingdom of Saudi Arabia, walter.imperatori@kaust.edu.sa; MAI, P. M., Kaust, Thuwal, Kingdom of Saudi Arabia, martin.mai@kaust.edu.sa
The heterogeneous nature of Earth's crust is manifested in the scattering of propagating seismic waves. In recent years, different techniques have been developed to include such phenomenon in broadband ground-motion calculations, either considering scattering as a semi-stochastic or pure stochastic process. In this study, we simulate broadband (0-10 Hz) ground motions using a 3D finite-difference wave propagation solver using several 3D media characterized by Von Karman correlation functions with different correlation lengths and standard deviation values. Our goal is to investigate scattering characteristics and its influence on the seismic wave-field at short and intermediate distances from the source in terms of ground motion parameters. We also examine other relevant scattering-related phenomena, such as the loss of radiation pattern and the directivity breakdown. We first simulate broadband ground motions for a point-source characterized by a classic omega-squared spectrum model. Fault finiteness is then introduced by means of a Haskell-type source model presenting both sub-shear and super-shear rupture speed. Results indicate that scattering plays an important role in ground motion even at short distances from the source, where source effects are thought to be dominating. In particular, peak ground motion parameters can be affected even at relatively low frequencies, implying that earthquake ground-motion simulations should include scattering also for PGV calculations. At the same time, we find a gradual loss of the source signature in the 2-5 Hz frequency range, together with a distortion of the Mach cones in case of super-shear rupture. Our simulations suggest that Von Karman correlation functions with correlation length between several hundred meters and few kilometers, Hurst exponent around 0.3 and standard deviation in the 5-10% range reproduce the available observations, although an unambiguous random-media characterization cannot be achieved.
Session: Earthquake Source Physics
Presenter   Bohnhoff, Marco
Schedule   Thu PM / Poster
Room   Hall 1
GONAF: A Borehole-Based Geophysical Observatory at the North Anatolian Fault Zone in NW Turkey
BOHNHOFF, M., GFZ Helmholtz-Centre Potsdam, Potsdam, Germany, bohnhoff@gfz-potsdam.de; DRESEN, G., GFZ Helmholtz-Centre Potsdam, Potsdam, Germany, dre@gfz-potsdam.de; BULUT, F., GFZ Helmholtz-Centre Potsdam, Potsdam, Germany, bulut@gfz-potsdam.de; RAUB, C., GFZ Helmholtz-Centre Potsdam, Potsdam, Germany, craub@gfz-potsdam.de; MALIN, P. E., Institute of Earth Science and Engineering, University of Auckla, Auckland, New Zealand, p.malin@auckland.ac.nz; ITO, H., Jamstec, Center for Deep Earth Exploration, Tokyo, Japan, hisaoito@jamstec.go.jp; KILIC, T., AFAD, Disaster and Emergency Management Presidency, Earthquake D, Ankara, Turkey, tugbay@deprem.gov.tr; NURLU, M., AFAD, Disaster and Emergency Management Presidency, Earthquake D, Ankara, Turkey, nurlu@deprem.gov.tr
The North Anatolian Fault Zone (NAFZ) below the Sea of Marmara represents a ‘seismic gap’ where a major earthquake is expected to occur in the near future. The Marmara segment of the NAFZ is located between the 1912 Ganos and 1999 Izmit ruptures and is the only segment that has not ruptured since 1766. The GONAF project (Geophysical Observatory at the North Anatolian Fault; www.gonaf.de) involves the installation of a high-resolution borehole seismic observatory at the NAFZ consisting of up to eight 300m deep vertical boreholes around the eastern Sea of Marmara to monitor the NAFZ segment at the transition to the recent 1999 Izmit rupture. GONAF is an international collaboration and co-funded by the International Continental Scientific Drilling Programme (ICDP), GFZ Potsdam and the Disaster and Emergency Management Presidency in Ankara/Turkey (AFAD). Further principal partners are MIT and UNAVCO/both US, IESE/New Zealand and JAMSTEC/Japan. The principal scientific objective of GONAF is to study physical processes acting before, during and after the expected M>7 earthquake along the Princes Islands segment of the NAFZ by monitoring microseismic activity at significantly reduced magnitude detection threshold and improved hypocentral resolution. In October 2012 the first GONAF borehole was successfully drilled in Istanbul on the Tuzla peninsula and an array of borehole seismometers was installed for permanent operation. In addition a surface station at the same site was installed consisting of short period, broadband and strong motions sensors.
Session: ShakeMap-Related Research, Development, Operations, and Applications
Presenter   Askan, Aysegul
Schedule   Thu PM / Poster
Room   Hall 1
Relationships Between Felt Intensity and Ground Motion Parameters for North Anatolian Fault Zone (Turkey)
BILAL, M., Disaster and Emergency Management Presidency, Ankara, Turkey, mustafa.bilal@afad.gov.tr; ASKAN, A., Middle East Technical University, Ankara,Turkey, aaskan@metu.edu.tr
Rapid response after major earthquakes is crucial for disaster mitigation purposes. ShakeMaps are used commonly for identifying the affected area after an earthquake in terms of either instrumental ground motion parameters or felt intensity measures. Currently, a local systematic relationship between instrumental parameters and felt intensity is not available for Turkey. Thus, the ShakeMaps are generated using relationships from other seismic zones. It is essential to study such a relationship for Turkey as the country lies on major active faults such as North Anatolian, North East Anatolian and East Anatolian fault zones that have caused catastrophic earthquakes. In this study, a relationship is obtained between felt intensity and instrumental ground motion parameters based on data from previous events in Turkey. The dataset consists of 92 records from 14 earthquakes of 5.7 < Mw < 7.6. Ground motion records are obtained from the online archive of the Earthquake Department of Disaster and Emergency Management Presidency (AFAD) in Turkey (www.daphne.deprem.gov.tr). The intensity values are gathered from the reports prepared by the experts of former General Directorate of Disaster Affairs. Using ordinary least squares regression and Deming regression techniques, two complementary relations are obtained relating MMI to PGA and PGV separately.
Session: New Frontiers in Seismic Data Analysis
Presenter   Boué, Pierre
Schedule   Wed 2:15 PM / Oral
Room   155A
STUDENT
Probing Deep Earth with Body Waves from Ambient Seismic Noise Correlation
BOUÉ, P. B., ISTerre, Grenoble, France, Pierre.Boue@ujf-grenoble.fr; POLI, P. P., ISTerre, Grenoble, France, Piero.Poli@ujf-grenoble.fr; BRIAND, X. B., ISTerre, Grenoble, France, Xavier.Briand@ujf-grenoble.fr; CAMPILLO, M. C., ISTerre, Grenoble, France, Michel.Campillo@ujf-grenoble.fr; PEDERSEN, H. P., ISTerre, Grenoble, France, Helle.Pedersen@ujf-grenoble.fr; ROUX, P. R., ISTerre, Grenoble, France, Philippe.Roux@ujf-grenoble.fr
Seismic noise correlation is now an established method to observe the propagation of surface waves between pairs of sensors, without involving transient sources. At low frequency, these observed surface waves are exploited to depict high-resolution image of the crust and upper mantle, or mapping the velocity changes associated with tectonic events. Recent works highlight more challenging observations such as attenuation measurement and body waves retrieval at regional scale. Here we focus on the detection of body waves at teleseismic distance using a global broadband dataset. We first show that body waves emerge from cross-correlations for every pairs of seismic stations whatever their geographical locations. Such worldwide body wave observations provide new and original information on the Earth structure and could now be used to imaging purposes, by applying classical methods for deep Earth study. In this work, we compare the global time-vs-offset representation of the cross-correlations computed in different period bands between 1 and 150s for which different phases are preferentially reconstructed. Then we analyze and discuss the deep Earth phases such as core phases which are usually extracted and studied from the earthquake seismograms.
Session: Data Products as Research Resources
Presenter   Mai, P. Martin
Schedule   Fri AM / Poster
Room   Hall 1
Comparing Earthquake Rupture Models in Space and Time
ZHANG, L., King Abdullah University of Science and Technology, Thuwal, Jeddah, Kingdom of Saudi Arabia, zhangling118911@gmail.com; MAI, P. M., King Abdullah University of Science and Technology, Thuwal, Jeddah, Kingdom of Saudi Arabia, martin.mai@kaust.edu.sa; ZIELKE, O., King Abdullah University of Science and Technology, Thuwal, Jeddah, Kingdom of Saudi Arabia, Olaf.Zielke@kaust.edu.sa; THINGBAIJAM, K. K. S., King Abdullah University of Science and Technology, Thuwal, Jeddah, Kingdom of Saudi Arabia, k.thingbaijam@kaust.edu.sa
Kinematic earthquake rupture models, as imaged by inverting seismic and/or geodetic data, are found to exhibit intra-event variability depending on the utilized inversion approaches and model parameterizations. Hence, there is need to quantify as well as visualize the observed variability in order to objectively identify features that are common in different models (for the same earthquake). We propose statistical measures according to parametric and nonparametric hypothesis testing for this purpose. The rigorous quantitative comparisons are implemented for the cases of (1) different models for the same earthquake as inferred by different research groups, and (2) different models from benchmark exercises for which a true solution exists. The emphasis is on complete models described by space-time data on the same space-time grid. The features of the approach include multiple metrics to handle the space- and time-dependent variability, interpolation to harmonize the different datasets, and fast computation for rapid application on large data volume. We attempt to explain the observed variability of the models for instance, in the slip distribution, the rupture-time and the slip duration. The deliverables also include resolution of the stable features of the models, and reconciliation of different inferred solutions for the "most likely model" from a set of proposed solutions.
Session: Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Models
Presenter   Imtiaz, A B Afifa
Schedule   Wed AM / Poster
Room   Hall 1
STUDENT
Magnitude and Distance Dependency of the Ground-Motion Variability from Numerical Simulations and Comparison with Real Data Analysis
IMTIAZ, A. B. A., Institut des Sciencec de la Terre (ISTerre), Grenoble, France, afifa.imtiaz@obs.ujf-grenoble.fr; CAUSSE, M., Institut des Sciencec de la Terre (ISTerre), Grenoble, France, mathieu.causse@obs.ujf-grenoble.fr; CHALJUB, E., Institut des Sciencec de la Terre (ISTerre), Grenoble, France, Emmanuel.Chaljub@obs.ujf-grenoble.fr; COTTON, F., Institut des Sciencec de la Terre (ISTerre), Grenoble, France, fabrice.cotton@obs.ujf-grenoble.fr
The ground-motion variability “sigma” is a fundamental component in Probabilistic Seismic Hazard Assessment since it controls the hazard level at very low probabilities of exceedance. So far, most of the analyses based on Empirical Ground-Motion Prediction Equations have not considered any distance dependeny of “sigma”. This study aims to capture some physical explanation on the origin of the ground motion variability, especially in the near field region, where the variability is poorly constrained due to the lack of available records .We, therefore, investigate the distance and magnitude dependency of “sigma” by performing ground-motion numerical simulations for some strike-slip events. Synthetic seismograms in terms of velocity time series (up to 3 Hz) have been generated from a suite of finite-source rupture models of past earthquakes. Green’s functions were calculated for a 1D velocity structure using a discrete wavenumber technique (Bouchon et al., 1981). Single station sigma was evaluated, from synthetic data, as a function of distance and magnitude. The simulations reveal that the within-event component of the ground-motion tends to increase with distance and decrease with magnitude. This tendency has also been observed by Rodriguez-Marek et al. (2011), based on accelerometric data analysis.
Session: ShakeMap-Related Research, Development, Operations, and Applications
Presenter   Riedel, Ismaël
Schedule   Thu PM / Poster
Room   Hall 1
STUDENT
Seismic Vulnerability of Urban Environment Using Datamining Machines
RIEDEL, I., ISTerre, Grenoble, France, ismael.riedel@ujf-grenoble.fr; GUEGUÉN, P., ISTerre, Grenoble, France, philippe.gueguen@ujf-grenoble.fr; COTTAZ, S., ISTerre, Grenoble, France
A complete seismic risk assessment requires not only the estimation of the seismic hazard (e.g. ground motion), but also a representation of the existent buildings quality and expected response through the definition of their vulnerability. The seismic vulnerability of an urban environment characterizes the capacity of buildings to stand the seismic ground motions. The difficulty in its estimation at the scale of a city is bound to the high amount of buildings to analyze, the construction type variability, the misunderstanding in the behaviour of an old structure due to the lack of information on its conception and material quality, and to the intrinsic variability in the response of structures to seismic loads. Over the last years, many methods for seismic vulnerability evaluation of existing buildings have been published; most of them calibrated using post-event damage information. In general, these methods characterize building stocks through structural parameter and establish empirical relationships between those parameters and the building typology, the vulnerability index or directly to the vulnerability class. However, these methodologies cannot be easily applied in moderate-to-low seismic prone regions, due to the large number of building parameters they require, which are hardly obtainable with a small budget. A way to handle these difficulties could be to start the estimation from the available data on buildings found in a national database (such as national census or remote sensing data), even if they cannot give a direct vulnerability classification following a vulnerability methodology. This approach was tested in France through two different datamining solutions: Machine rule learning and Support Vector Machine, calibrated on existing building-by-building Grenoble (France) dataset and compared to other methods of seismic vulnerability assessment performed in France or to historical earthquakes that produced moderate-to-strong motion in the territory.
Session: New Developments in Earthquake Forecasting and Predictability Research
Presenter   Schorlemmer, Danijel
Schedule   Wed PM / Poster
Room   Hall 1
On the Testability of Maximum Earthquake Magnitude
CLEMENTS, R. A., GFZ Potsdam, Potsdam, Germany, clements@gfz-potsdam.de; GONZALEZ, A., GFZ Potsdam, Potsdam, Germany, alvaro@gfz-potsdam.de; SCHORLEMMER, D., GFZ Potsdam, Potsdam, Germany, ds@gfz-potsdam.de
Disasters caused by unexpectedly large earthquakes illustrate the need for reliableestimates of the maximum possible magnitude Mmax at a given fault or in a particular zone. Such estimates are essential parameters in seismic hazard assessment, but their accuracy remains untested. In fact, whether Mmax can be tested or not is still uncertain. In this study, we discuss the testability of Mmax and the limitations that arise from testing such rare events. We use a simple extreme value theory approach to derive the sampling distribution for the maximum magnitude, i.e. the probability distribution for the maximum of a sample of earthquake magnitudes, and propose a straightforward hypothesis test for Mmax. The test is based on the largest magnitude observed in the sample. If it is larger than the estimate of Mmax, or it is too unlikely, given the assumed magnitude probability density function, the estimate of Mmax is rejected. We then perform a sensitivity analysis to identify which parameters have the most influence on this sampling distribution and conduct a power analysis for the test. Our results suggest that the sampling distribution is relatively insensitive to the overall Mmax, except when the b-value of the Gutenberg-Richter distribution is low and the size of the sample is high. Consequently, the power of the test is high only under optimal conditions, such as when the hypothesized value of Mmax is grossly different than the true Mmax, or when the seismicity rate is very high. Finally, we discuss that these limitations, in practice, may imply that a wrong maximum magnitude estimate can rarely be falsified, and express our concern about the use of these unfalsifiable estimates in seismic hazard assessment.
Session: What are the limits of explosion source model predictions?
Presenter   Matzel, Eric
Schedule   Thu AM / Poster
Room   Hall 1
Using Seismic Interferometry to Image the Site of the Source Physics Experiment
MATZEL, E. M., LLNL, Livermore, CA, matzel1@llnl.gov; MELLORS, R. J., LLNL, Livermore, CA; PITARKA, A., LLNL, Livermore, CA
The Source Physics Experiment (SPE) is a series of precisely designed explosions recorded by a dense network of seismometers. Its purpose is to obtain a physics-based understanding of how seismic waves are created at and scattered near the source. In order to separate source-specific effects from those due to geological heterogeneity, we need a precise picture of the subsurface. We are using several methods of seismic interferometry to obtain highly detailed images of the SPE site. Coda interferometry (CI) uses the diffuse coda from as a source of coherent energy. Ambient noise correlation (ANC) uses the energy of the ambient background field. In each technique, the data recorded at one seismometer are correlated with the data recorded at another to obtain an estimate of the Green's function (GF) between the two.More than 150 instruments were deployed around the site, predominantly along 5 lines extending radially outward from the shot point. We used the records of one of the SPE shots as an energy source for the CI technique and several weeks of high gain continuous data for ANC.Each technique has advantages over the other. CI is very fast and the GF obtained has the same frequency content as the original shot. The key disadvantages of CI are that we can only correlate energy propagating radially outward and the source point itself is hidden. ANC requires more data and processing time, but allows us to estimate the GF between any two of the seismometers. By combining the two techniques, we obtain a very sharp image of seismic velocity and attenuation beneath the site. Using CI, we recovered 860 GFs along the 5 main lines. ANC gave us several thousand more, including paths that cross the source point and between the lines. Among key observations, we are able to map a thin (20-50 m), very low velocity (150-250 m/s) layer at the surface in the SE of the site. Our end result is a fully 3D tomogram with lateral resolution as small as 100 m.
Session: Oceanographic and Atmospheric Signals in Seismology
Presenter   Hillers, Gregor
Schedule   Fri AM / Poster
Room   Hall 1
Noise-Based Observation of Wave Speed Variations Associated With Tidal Loading: In Situ Acoustoelastic Testing
HILLERS, G., ISTerre, Univ. Joseph Fourier, CNRS, Grenoble, France, gregor.hillers@ujf-grenoble.fr; RETAILLEAU, L., IPGP, Sorbonne Paris Cite, CNRS, Paris, France, retailleau@ipgp.fr; CAMPILLO, M., ISTerre, Univ. Joseph Fourier, CNRS, Grenoble, France, michel.campillo@ujf-grenoble.fr; INBAL, A., Caltech Seismolab, Pasadena, CA, ainbal@gps.caltech.edu; RIVERA, L., IPGS, Strasbourg University, EOST, Strasbourg, Paris, luis.rivera@unistra.fr; MA, K. F., Department of Earth Sciences, National Central University, Jhongli City, Taiwan, fong@eqkc.earth.ncu.edu.tw; AMPUERO, J. P., Caltech Seismolab, Pasadena, CA, ampuero@gps.caltech.edu; NISHIMURA, T., Department of Geophysics, Tohoku University, Sendai, Japan, nishi@zisin.gp.tohoku.ac.jp
We report on observations of systematic seismic velocity variations in response to tidal deformation. Linear---and potentially nonlinear elastic and/or anelastic---response types and associated material properties from a wide range of Earth materials can be inferred from laboratory acoustoelastic measurements. In this method a continuous low frequency load is applied to a sample and the resulting velocity and hence modulus changes are monitored with simultaneously applied high-frequency pulses. We use a conceptually similar technique to measure the linear response, i.e. the strain sensitivity, of a range of in-situ crustal and fault zone materials. Earth tides constitute the low frequency deformation; relative seismic velocity changes (dv/v) associated with different parts of the tidal strain protocol are resolved using standard ambient noise-based monitoring techniques. This “noise-based in-situ acoustoelastic testing” method is applied in an analysis of ambient seismic wave fields recorded by seismic arrays situated in various tectonic regions covering a range of spatial and temporal scales. We present results from multiple tests to estimate the resolution and temporal robustness of the dv/v signal associated with the deformation protocol, and we discuss the frequency dependence of the signal as a function of ambient wave field properties. We consider the potential to image spatially variable material properties associated with the damage state based on observations of different inter-station dv/v amplitudes in regions with strong lateral heterogeneities (e.g. across fault zones). The new method constitutes a complementary technique to active source experiments for imaging and monitoring rheologic properties and strain sensitivities, and we discuss implications for routine observation and monitoring of response fluctuations and their potential contribution to the detection of transients.
Session: Velocity Models and Modeling
Presenter   Kelemencky, Sara
Schedule   Thu AM / Poster
Room   Hall 1
STUDENT
New 1-D and 3-D Velocity Models and Hypocenter Locations for the Charlevoix Seismic Zone
KELEMENCKY, S., Center of Earthquakes Research and Information, Univ. of Memphis, Memphis, TN, sklmncky@memphis.edu; POWELL, C., Center of Earthquakes Research and Information, Univ. of Memphis, Memphis, TN, capowell@memphis.edu; LAMONTAGNE, M., Natural Resources Canada, Ottawa, ON, Canada, maurice.lamontagne@nrcan.gc.ca
We present our new 1-D and 3-D velocity models and interpret its rheological and compositional properties for the active Charlevoix seismic zone (CSZ), which is located ~100 km northeast of Quebec City, Canada. The CSZ has generated several damaging earthquakes but the causative reason for the seismic activity in this intraplate region is unknown. The region has a complex tectonic history including the Grenville Orogeny, opening and closing of the Iapetus Ocean, a Devonian meteor impact, and opening of the Atlantic Ocean; structure developed during any of these prior events could be influencing present day seismic activity. Determining the seismotectonic framework for the CSZ will require detailed 3-D velocity models. A previous local earthquake tomography study produced a 3-D P-wave velocity model but this analysis used an overly simplistic half-space starting model. We have developed improved 1-D P- and S-wave starting models using VELEST. These models contain several layers in the depth range 0 to 30 km and provide a better representation of the 1-D velocity structure than the half-space. The new 3-D P- and S-wave velocity models will provide insight into what factors, such as compositional variations or fracturing and elevated fluid pressures, most strongly influence earthquake occurrence in the CSZ. Relocated hypocenters may identify the spatial extent of seismogenic faults leading to improved seismic hazard assessment for the region.
Session: New Frontiers in Seismic Data Analysis
Presenter   Bowman, Daniel
Schedule   Wed 5:00 PM / Oral
Room   155A
STUDENT
The Hilbert-Huang Transform Applied to Seismo-acoustic Signals on Volcanoes
BOWMAN, D. C., University of North Carolina, Chapel Hill, NC, daniel.bowman@unc.edu; LEES, J. M., University of North Carolina, Chapel Hill, NC, jonathan.lees@unc.edu
We present an overview of the Hilbert-Huang transform (HHT) and discuss its application to three signals: a synthetic nonlinear wave, a transient signal recorded at Deception Island volcano, Antarctica, and quasi-harmonic volcanic tremor from Reventador volcano, Ecuador. The HHT is an empirical method designed to examine nonstationary and nonlinear time series with high resolution in both the time and frequency domain. Many spectral methods, such as Fourier analysis, assume that time series extend from positive to negative infinity (stationarity) and consist of a linear superposition of sinusoids (linearity). However, seismo-acoustic signals are never stationary and are not necessarily linear. The HHT largely avoids the linearity and stationarity constraints of Fourier-based spectral analysis tools by combining an adaptive signal decomposition method with the time/frequency precision of the Hilbert transform. We compare the HHT spectrogram with the Fourier spectrogram of the three examples above and discuss the differences and similarities between them. The synthetic signal highlights the different approaches of the HHT and the Fourier spectrogram. The transient signal demonstrates the high time/frequency resolution of the HHT, including a better preservation of causality. The volcanic tremor shows how HHT and Fourier differ in the analysis of a complex signal, including how the HHT eliminates high frequency harmonics. Finally, we discuss the advantages and disadvantages of the HHT with regards to other spectral methods.
Session: Seismic Hazards and Ground Motions
Presenter   Petersen, Mark
Schedule   Fri 11:00 AM / Oral
Room   155B
Updating the CEUS Model for the 2014 USGS National Seismic Hazard Maps
MUELLER, C. S., U. S. Geological Survey, Golden, CO, cmueller@usgs.gov; PETERSEN, M. D., U. S. Geological Survey, Golden, CO, mpetersen@usgs.gov; MOSCHETTI, M. P., U. S. Geological Survey, Golden, CO, mmoschetti@usgs.gov
The U. S. Geological Survey will update the national seismic hazard maps in 2014. Based on new research, including results from the recent Central and Eastern United States Seismic Source Characterization for Nuclear Facilities project (CEUS-SSC), as well as feedback from our workshops, we are considering several modifications to the source model for central and eastern North America: 1) use a new moment-magnitude-based seismicity catalog developed by CEUS-SSC; 2) implement new regional zones for analysis of catalog completeness levels and b values; 3) update methods for identifying and treating induced earthquakes; 4) implement adaptive smoothing for background seismicity; 5) update maximum-magnitude models; and 6) update models for specific faults and source areas, including New Madrid, Charleston, Meers, Cheraw, and Wabash. We will probably retain the four gridded and smoothed seismicity rate models from previous maps: magnitude 3+, 4+, 5+, and floor. We are also evaluating several new ground-motion prediction equations and alternative ways of grouping and weighting ground-motion relations. These assessments make use of the extensive new ground-motion dataset from the Next Generation Attenuation – East project as well as results from a recent ground-motion review sponsored by the Electric Power Research Institute, considering new ideas about near-source geometrical attenuation, site effects, and ground-motion regionalization. We will show the results of sensitivity tests and comparisons with the 2008 hazard maps.
Session: Triggering of Seismic and Volcanic Events
Presenter   Fereidoni, Azadeh
Schedule   Fri PM / Poster
Room   Hall 1
STUDENT
Statistical Features of Aftershock Temporal Behavior in Eastern North America: Application for Seismicity Declustering and Seismic Hazard Assessment
FEREIDONI, A., Western University, London, ON Canada, afereydo@uwo.ca; ATKINSON, G. M., Western University, London, ON, Canada, gatkins6@uwo.ca
In conventional seismic hazard assessment, the assumption of Poissonian distribution of earthquakes necessitates the removal of aftershock clusters from seismicity catalog (seismicity declustering). Over the years, various declustering techniques have been proposed for identifying aftershocks (dependent events) from mainshocks (independent events). Reasenberg (1985) introduced a simple declustering algorithm, based mostly on typical aftershock statistics in California (interplate seismicity). The algorithm has become quite popular and been applied widely in earthquake hazard assessment studies worldwide. However, users often apply this declustering technique blindly, without justifying the cluster criteria for the features in the study region. In this study, we calibrate the aftershock parameter values in Reasenberg (1985) algorithm for intraplate aftershock clusters in Eastern North America (ENA). For this purpose, we statistically analyze the aftershock activity of moderate to large earthquakes in ENA by fitting them to modified Omori formula (MOF). We incorporate residual analysis into our study to illustrate the goodness of fit and detect the ending time of aftershock sequences. The recommended values of parameters for ENA are obtained by averaging the values found for individual sequences. We apply the method of Reasenberg (1985) with standard and calibrated parameters to decluster the Canadian Composite Seismicity Catalog (Fereidoni et al, 2012). The resulting declustered seismicity is subsequently used to derive Gutenberg-Richter magnitude recurrence model in selected locations. This allows us to draw conclusions regarding the sensitivity of recurrence parameters to catalog clustering properties in ENA.
Session: Characterizing Active Faults for Seismic Hazard Assessments
Presenter   Gold, Ryan
Schedule   Thu 1:30 PM / Oral
Room   155B
Paleoseismic Investigation of the Mohawk Valley Fault System at the Sulphur Creek Sidehill Bench, Northern Walker Lane, California
GOLD, R. D., U.S. Geological Survey, Golden, CO, rgold@usgs.gov; BRIGGS, R. W., U.S. Geological Survey, Golden, CO, rbriggs@usgs.gov; ANGSTER, S. J., U.S. Geological Survey, Golden, CO, sangster@usgs.gov; PERSONIUS, S. F., U.S. Geological Survey, Golden, CO, personius@usgs.gov; CRONE, A. J., U.S. Geological Survey, Golden, CO, crone@usgs.gov
The dextral-slip Mohawk Valley fault system (MVFS) strikes northwestward along the eastern margin of the Sierra Nevada and is the westernmost member of a network of active dextral-slip faults in the northern Walker Lane. Geodetic data indicate that the MVFS may accommodate as much as 2-3 mm/yr of regional dextral strain; however, limited geologic data have been available to constrain the system’s slip rate and earthquake history. We mapped the complex and anastomosing MVFS using high-resolution airborne Light Distance and Ranging (LiDAR) data and field observations, and selected the Sulphur Creek Sidehill Bench site for paleoseismic investigation. At this site, oblique dextral-normal faulting has produced a closed depression on the steep valley margin that floods annually during spring snowmelt. In three fault-perpendicular trenches, we exposed fine-grained, pond sediment that interfingers with multiple wedges of colluvium derived from the bedrock scarp that bounds the eastern side of the pond. We interpret the wedges, and other stratigraphic and structural relationships, to record three or four surface-rupturing earthquakes. The faulted deposits appear to be latest Pleistocene and possibly Holocene in age based upon the degree of clast weathering and soil development in exposures. In progress radiocarbon and luminescence samples will bracket the times of major surface-rupturing earthquakes. Our investigation will improve our knowledge of the history of earthquakes on the MVFS, advance our understanding of strain accommodation in the northern Walker Lane, and refine regional seismic hazard assessments.
Session: Characterizing Active Faults for Seismic Hazard Assessments
Presenter   Woessner, Jochen
Schedule   Thu 11:15 AM / Oral
Room   155B
Modeling Earthquake Activity for Seismogenic Sources Across the Euro-Mediterranean Region
WOESSNER, J., Swiss Seismological Service, ETH Zurich, Zurich, Switzerland, j.woessner@sed.ethz.ch; BASILI, R., National Institute of Geophysics and Volcanology (INGV), Rome, Italy, roberto.basili@ingv.it; DANCIU, L., Swiss Seismological Service, ETH Zurich, Zurich, Switzerland, l.danciu@sed.ethz.ch; GIARDINI, D., ETH Zurich, Zurich, Switzerland, d.giardini@sed.ethz.ch
Probabilistic seismic hazard assessment (PSHA) aims at exploiting a variety of models for the earthquake source process and ground motion attenuation to cover epistemic and aleatory uncertainties. SHARE, an EC-FP7 funded project (www.share-eu.org), created a testable time-independent community-based hazard model for the Euro-Mediterranean region. The results will serve as reference model within Europe and will provide homogeneous input for seismic safety assessment of critical infrastructures.For the first time on a Euro-Mediterranean scale, earthquake rupture forecasts are created from a seismogenic source database compiled within the course of the project. The database provides access to all geometric and dominant earthquake parameters needed to model seismic activity. We divide the study area into background regions that contain entire fault systems and estimate the b-value within these regions from observed seismicity, assuming that b-value on faults is not different from that in the background. We then model the fault activity using several published methods.For all fault activity models, we find that the difference between modeled and observed seismic moment release increases with increasing slip rate. Given the current dataset, we also find that differences in the model parameterization have less influence than uncertainties in the input data.Hazard results based on this modeling approach tend to be higher in high slip rate regions and lower in others compared to previous results and other modeling approaches though with considerable uncertainty due to the parameters in the input data. This implies that strong efforts are required to improve the completeness and quality of active fault mapping especially the most critical parameters such as the slip rate.
Session: New Developments in Earthquake Forecasting and Predictability Research
Presenter   Field, Edward
Schedule   Thu 11:00 AM / Oral
Room   155D
The Uniform California Earthquake Rupture Forecast Version 3 – Key Assumptions and Testability
FIELD, E. H., USGS, Golden, CO, field@usgs.gov
The Working Group on California Earthquake Probabilities (WGCEP), a joint effort of the Southern California Earthquake Center, United States Geological Survey, and California Geological Survey, is in the process of developing the 3rd Uniform California Earthquake Rupture Forecast (UCERF3). Two primary objectives of this model are to relax segmentation and include multi-fault ruptures, the need for which was recently exemplified by the 2011 M9 Tohoku and 2012 M8.6 Sumatra earthquakes, respectively. Two other important objectives are to resolve self-consistency issues with elastic-rebound probability calculations in un-segmented models, and to include spatiotemporal clustering (triggered events and aftershocks). The 2011 M6.3 Christchurch earthquake exemplifies the need to represent the latter, and modeling such behavior will require deploying UCERF3 as an operational earthquake forecast (in terms of real-time access to network information). The purpose of this presentation is to summarize the key assumptions being made in UCERF3, prioritized by their practical implications, and to discuss the testability of these assumptions (in the opinion of this author, which may or may not agree with the views of other WGCEP participants).
Session: New Frontiers in Seismic Data Analysis
Presenter   Lees, Jonathan
Schedule   Wed 4:30 PM / Oral
Room   155A
R: New Seismic Analysis Packages
LEES, J. M., University of North Carolina, Chapel Hill, NC, jonathan.lees@unc.edu
A suite of R packages designed for a wide range of seismic analysis is currently available in the free software platform called R. R is a software platform based on the S-language developed at Bell Labs decades ago. Routines in R can be run as standalone function calls, or developed in object-oriented mode. The R approach is similar to Matlab, although it is open source and free to down load. R comes with a base set of routines, and thousands of user developed packages. The packages developed at UNC include subroutines and interactive codes for processing seismic data, analyzing geographic information (GIS) and inverting data involved in a variety of geophysical applications. On CRAN (acronym for the host website: Comprehensive R Archive Network, http://www.r-project.org/) currently available packages related to seismic analysis are RSEIS, GEOmap, RFOC, zoeppritz, RTOMO, and geophys. These include signal processing, data management, mapping, focal mechanisms, and deformation among other useful functionionality. All software in R packages is required to have documentation, making the exchange and modification of existing software easy. For example, we recently installed a new set of programs that allow users to plot focal moment tensor solutions on a Tape and Tape (2012) style, or ‘lune’, display. These will be illustrated with examples from volcanic regions in Ecuador.
Session: What are the Limits of Explosion Source Model Predictions?
Presenter   Yang, Xiaoning
Schedule   Thu AM / Poster
Room   Hall 1
Moment Tensor Inversion of Spatially and Temporally Distributed Sources Involved in an Underground Explosion
YANG, X., Los Alamos National Laboratory, Los Alamos, NM, xyang@lanl.gov; PATTON, H. J., Los Alamos National Laboratory, Los Alamos, NM, patton@lanl.gov
Recent developments in explosion-source model research illustrate that the secondary damage source during an underground explosion, which involves spallation of near-surface layers and deep-seated material damage under a tensile stress regime, has profound effects on surface-wave generation, Ms-mb discriminant, and yield estimation. Traditional moment-tensor representation of an explosion source in an inversion, however, has been restricted to a single point source, even though the damage source happens at different depths than the explosion and starts later in time. To better characterize a spatially and temporally distributed source, we developed a revised frequency-domain moment-tensor inversion method that accommodates a source with two centroids at different depths. Inversion tests using noise-free, near-source (< 2 km with a source depth of 55 m) synthetic data show that the new method recovers the input source perfectly. Adding realistic noise to the synthetic seismograms does not affect the result significantly. We investigate the resolution of the new inversion method and tradeoffs involved in different source parameterizations using synthetic data. Application of the new method to observed data from the first three of the U. S. Department of Energy’s Source Physics Experiment (SPE) explosions indicates that the new method holds promise to resolve and quantify the spatial extent and temporal variation of a complex explosion source.
Session: What are the Limits of Explosion Source Model Predictions?
Presenter   Pitarka, Arben
Schedule   Wed 11:00 AM / Oral
Room   155D
Three Dimensional Simulation of Far-Field Ground Motion from Source Physics Experiment Using a Hydrodynamic-to-Elastic Coupling Technique
PITARKA, A., Lawrence Livermore National Laboratory, Livermore, CA, pitarka1@llnl.gov; MELLORS, R. J., Lawrence Livermore National Laboratory, Livermore, CA, mellors1@llnl.gov; VOROBIEV, O., Lawrence Livermore National Laboratory, Livermore, CA; XU, H., San Diego State University, San Diego, CA; RODGERS, A. J., Lawrence Livermore National Laboratory, Livermore, CA; WALTER, W. R., Lawrence Livermore National Laboratory, Livermore, CA; ANTOUN, T., Lawrence Livermore National Laboratory, Livermore, CA; PETERSSON, A., Lawrence Livermore National Laboratory, Livermore, CA; SJOGREEN, B., Lawrence Livermore National Laboratory, Livermore, CA; MATZEL, E., Lawrence Livermore National Laboratory,Livermore, CA; WAGONER, J., Lawrence Livermore National Laboratory, Livermore, CA
Monitoring of nuclear explosions typically relies on distant recordings where the effects of non-linear near-source processes are combined with path-specific propagation effects, such as complex scattering, causing difficulties in extracting source properties from available waveform data. Therefore understanding the effects of wave scattering in the source region where the material response is non-linear as well as in the far-field where the material response is linear is crucial for studying the excitation of seismic energy during underground explosions as well as developing empirical source models. Ground motion recorded during the Source Physics Experiment (SPE) at near-field and far-field stations provided an excellent opportunity for understanding the excitation and propagation of seismic waves from underground explosions. We used three–dimensional hydrodynamic simulations, including non-linear rheology of the near-source material, coupled with elastic wave propagation to model SPE ground motion recorded at far-field stations. The simulations were used to extract seismic source characteristics and investigate near-source wave propagation effects. Near-source motions were computed with GEODYN-L, a Lagrangian hydrodynamic code for modeling the response of a well- characterized granite to explosion loading. Explosion generated ground motions were passed to WPP, an anelastic finite-difference code, through a coupling approach using various three-dimensional velocity models with specified material heterogeneities. Our physics based modeling results were analyzed to separate complex non-linear effects near the source from various along path wave propagation scattering effects due to topography and material heterogeneity in the source region. In particular, we focused on the contribution of these effects to S-wave generation.
Session: What are the Limits of Explosion Source Model Predictions?
Presenter   Yoo, Seung Hoon
Schedule   Thu AM / Poster
Room   Hall 1
Local Explosion Monitoring from the Source Physics Experiment (SPE): Detection, Yield, and Discrimination using Network Template Matching
YOO, S. H., Berkeley Seismological Laboratory / Weston Geophysical Corp., Berkeley, CA, syoo@seismo.berkeley.edu; MAYEDA, K., Weston Geophysical Corp., Oakland, CA, kevin.mayeda@gmail.com; BONNER, J., Weston Geophysical Corp., Lufkin, TX, jes_bonner@westongeo.com
We introduce a newly-developed explosion monitoring procedure using a locally-deployed, high-frequency seismic network. For event detection and location, we adopted a template matching method using empirically derived high-frequency coda envelope templates from local seismic records (peak energy arrival time and its decay rate). After detecting suspected explosions from continuous waveforms based on signal-to-noise ratios and cross-correlation coefficients, we mapped the averaged cross-correlation coefficients over the network on a 2D grid map to determine the event’s likely location. Once located, the program automatically measures its coda amplitudes, applies corrections, then forms source spectra and assuming either an earthquake or an explosions source, estimates source parameters such as seismic moment, yield, radiated seismic energy, and stress drop and forms high-frequency discriminants. Reference parameters and coefficients for coda amplitude measurements and calibrations for all stations (and all available channels) in the local network are pre-derived using previous local seismic records from a variety of local area networks. The method was successfully applied to recordings of the SPE2 and SPE3 experiments at Nevada National Security Site. This study will outline our method and provide new data from which to compare to other previous source physics studies such as the Source Phenomenology Experiment in Arizona, New England Damage Experiment, and Humble-Redwood I-III series in New Mexico, among others. For the future, we believe this new procedure can be directly applied in real-time explosion monitoring in local areas of interest with limited a priori calibration.
Session: Data Products as Research Resources
Presenter   Nakamura, Yosio
Schedule   Thu 2:30 PM / Oral
Room   155A
Lessons Learned in Preserving ALSEP (Apollo Lunar Surface Experiment Package) Data
NAKAMURA, Y., Institute for Geophysics, University of Texas at Austin, Austin, TX, yosio@ig.utexas.edu; FROHLICH, C., Institute for Geophysics, University of Texas at Austin, Austin, TX, cliff@ig.utexas.edu; NAGIHARA, S., Department of Geosciences, Texas Tech University, Lubbock, TX, seiichi.nagihara@ttu.edu
The geophysical data, including seismic data, acquired on the Moon from 1969 through 1977 during and following the Apollo lunar landing missions are attracting renewed interests recently mainly for two reasons: (1) Computers vastly more powerful than were available earlier and newly developed data analysis techniques are allowing us to extract new information; and (2) these data contain much valuable information the original investigators were unaware of and thus were not the subjects of earlier studies. However, in many cases, we are encountering obstacles because these old data were not preserved properly for use by future generations. Even though the original ALSEP data were acquired digitally, in some cases only selected analog plots of interest to the original investigators were archived; and when the original raw data were later reformatted to a standard format, such as SEED, which postdates the original data acquisition, important information, such as a clue to possible sources of data irregularities, which may be obvious in raw data streams, was often lost or incorrectly transferred. The only foolproof way that we have found to avoid these pitfalls will be to archive also the raw data as they were acquired with all the necessary metadata. Note that this applies not only to old data such as ALSEP but also to terrestrial data currently being acquired, for instance, active seismic data extracted and reformatted to SEG-Y from continuously recorded raw seismic data.
Session: New Frontiers in Seismic Data Analysis
Presenter   Abbott, Robert
Schedule   Thu AM / Poster
Room   Hall 1
Observations of Volcanic Activity at Kilauea Volcano, Hawaii, Using Rotational Seismometers
ABBOTT, R. E., Sandia National Laboratories, Albuquerque, NM, reabbot@sandia.gov; HART, D., Sandia National Laboratories, Albuquerque, NM, dhart@sandia.gov; THELEN, W. A., USGS Hawaiian Volcano Observatory, Hawaii National Park, HI, wthelen@usgs.gov
We present data and analysis from a multi-week deployment of two rotational seismometers at Uwekahuna, Hawaii, on Kilauea Volcano. The rotational seismometers are ATA models ARS-16 (three instruments in ZNE configuration), and ARS-24 (two instruments in vertical orientation). These instruments utilize magnetohydrodynamics to measure particle rotation rate with negligible sensitivity to translational motion. The instruments were paired with a Kinemetrics EpiSensor tri-axial accelerometer to create a six degree-of-freedom (6DOF) instrument (three orientations for translational motion and three for rotation rate). Controlled testing shows that instrument self noise for the ARS-24 and ARS-16 were 2.2 and 11.7 e-6 radians/s rms, respectively. Ambient noise at Uwekahuna appears to be 10-15 dB lower than the quieter ARS-24. Many high-amplitude events were recorded with significant signal-to-noise ratio, however. These include a magnitude 2 event within 2 km epicentral distance and a magnitude 3.3 event 20 km distant. 6DOF processing of these event data to determine back-azimuth had highly variable results, even though cross correlations between vertical-axis rotation rate and transverse horizontal acceleration were high. This is likely caused by the great degree of scattering (short mean free path) at Kilauea Volcano. 6DOF processing to determine in situ shear velocity is not sensitive to scattering, and yields shear wave velocities of 300-350 m/s at 10 Hz. This is in agreement with previous studies. Sandia National Laboratories is a multi-program laboratory managed and operated by Sandia Corporation, a wholly owned subsidiary of Lockheed Martin Corporation, for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-AC04-94AL85000.
Session: Earthquake Source Studies
Presenter   Bent, Allison
Schedule   Wed PM / Poster
Room   Hall 1
Evaluation of Lg Based Magnitude Scales for Eastern North America or Why is Your Magnitude Different from the USGS?
BENT, A. L., Geological Survey of Canada, Ottawa ON, Canada, bent@seismo.nrcan.gc.ca
While moment magnitude is generally the preferred magnitude for characterizing the size of an earthquake and the implementation of regional moment tensor inversions allows it to be calculated more rapidly and for smaller events than in the past, it is not a practical scale for rapid magnitude calculations of small earthquakes. In eastern Canada, mN remains the most commonly used magnitude scale. A conversion relation for mN to MW was previously established but ideally it should be applied to large data sets and not to individual earthquakes where the difference may vary from the mean. At the same time it has been noted that Lg based magnitudes calculated by other organizations using different formulae give magnitudes that are much closer to the expected or calculated MW values for eastern Canadian earthquakes than the mNs in the Canadian database. As part of a larger study to develop best magnitude practices for Canadian earthquakes, three Lg based magnitude scales are evaluated and compared: the Nuttli (1973) mN scale as applied by the Geological Survey of Canada (GSC) , the IASPEI (2011) mb_Lg scale adopted by the USGS and Herrmann and Kijko’s (1983) mLg(f) scale. The latter two give event magnitudes that are close to MW values obtained by moment tensor inversion and about 0.5 units lower than the GSC mN magnitudes. On comparing the differences between the mN and mLg(f) values on a point by point basis, however, a clear dependence on both distance and period is observed. Multivariate regression was used to develop a conversion relation between the two that removes these effects. An in-depth comparison of the mN and mb_Lg scales is ongoing. Ultimately, the goal is to implement a magnitude scale that allows rapid magnitude calculation from velocity records, is representative of the earthquake size and is independent of the data distribution.
Session: Infrasound and Seismoacoustics
Presenter   Jones, Kyle
Schedule   Wed 4:30 PM / Oral
Room   155D
Southwest U.S. Seismo-Acoustic Network: An Autonomous Data Aggregation, Detection, Localization and Ground-Truth Bulletin for the Infrasound Community
JONES, K. R., Sandia National Laboratory, Albuquerque, NM, krjones@sandia.gov; ARROWSMITH, S. J., Los Alamos National Laboratory, Los Alamos, NM, arrows@lanl.gov
The Southwest U.S. Seismo-Acoustic Network (SUSSAN) is a collaborative project designed to produce infrasound event detection bulletins for the infrasound community for research purposes. We are aggregating a large, unique, near real-time data set with available ground truth information from seismo-acoustic arrays across New Mexico, Utah, Nevada, California, Texas and Hawaii. The data are processed in near real-time (~ every 20 minutes) with detections being made on individual arrays and locations determined for networks of arrays. The detection and location data are then combined with any available ground truth information and compiled into a bulletin that is released to the general public directly and via the IRIS infrasound event bulletin.We use the open source Earthworm seismic data aggregation software to acquire waveform data either directly from the station operator or via the IRIS Data Management Center (if available). The data are processed using InfraMonitor, a powerful infrasound event detection and localization software program developed by Stephen Arrowsmith (LANL). Our goal with this program is to provide the infrasound community with an event database that can be used collaboratively to study various natural and man-made source. We encourage participation in this program directly or by making infrasound array data available through the IRIS DMC.
Session: Data Products as Research Resources
Presenter   Wald, David
Schedule   Fri AM / Poster
Room   Hall 1
Properties of the ‘Eastern, Central, and Mountain States of U.S., 1350 – 1986’ Database in the Online ‘Earthquake Catalog Search’ of the U.S. Geological Survey
DEWEY, J. W., U.S. Geological Survey, Denver, CO, jdewey@usgs.gov
This study revisits the ‘Eastern, Central, and Mountain States of U.S., 1350 – 1986’ database that is searchable from the U.S. Geological Survey’s Earthquake Catalog Search (http://earthquake.usgs.gov/earthquakes/eqarchives/epic/). The database has existed in its current form since the late 1980’s. Principal compilers were C.W. Stover, B.G. Reagor, and S.T. Algermissen. For U.S. states east of, and including, Idaho, Utah, and Arizona, the database contains origins of earthquakes that were also plotted in a series of state-specific USGS Miscellaneous Field Studies Maps published in 1977-1988: the database is most reliable for these states. The date of the earliest event in this part of the database is 1568. The database also includes many events from the Pacific states and Nevada, but the westernmost part of the database was still a work-in-progress at the time that compilation of the overall database was suspended; spot checking of the westernmost part of the database reveals data-entry errors and data-gaps. For the eastern, central, and mountain states for which compilation and review of the database were completed, the database is particularly valuable as a compilation of information on small, non-instrumentally-recorded, earthquakes that are omitted from many other earthquake listings. Knowledge of these earthquakes may shed light on seismogenic structures that have not been active in recent decades and on natural seismicity of regions where recent seismicity is hypothesized to be due to triggering by human activity. Origins in the database east of 105.5°W have been recently analyzed as part of the Central and Eastern United States Source Characterization for Nuclear Facilities (CEUS-SSC) Project (http://www.ceus-ssc.com). The present paper will analyze completeness thresholds for the Mountain States west of 105.5°W in light of thresholds proposed in the CEUS-SSC Project for the region east of 105.5°W.
Session: Characterizing Active Faults for Seismic Hazard Assessments
Presenter   Murray, Jessica
Schedule   Thu 4:45 PM / Oral
Room   155B
Spatially Variable Creep Rate on the Bartlett Springs and Maacama Faults, Northern California, Estimated via Bayesian Inversion of Global Positioning System Data
MURRAY, J. R., USGS, Menlo Park, CA, jrmurray@usgs.gov; MINSON, S. E., USGS, Seattle, WA, sminson@usgs.gov; SVARC, J., USGS, Menlo Park, CA, jsvarc@usgs.gov
Creep at seismogenic depths is thought to reduce the maximum magnitude of earthquakes a fault will produce, so quantifying the rates and spatial extent of creep is important for accurate hazard assessment. Surface creep is observed on several faults of the San Andreas system in northern California. Alinement arrays record creep rates of ~3 – 5 mm/yr at three locations on the Maacama (MF) and Bartlett Springs (BSF) faults. Data from recently expanded GPS networks in this region are generally consistent with these rates but provide broader spatial coverage. Our goal is to use the GPS data to infer slip rates below the seismogenic zone for the MF, BSF, and San Andreas Fault (SAF) and to image spatially variable creep rate on the MF and BSF. Inversions of geodetic data for distributed slip are underdetermined. Assumptions such as spatial smoothing can regularize the inversion but often lead to spurious results. Bayesian inversion provides posterior probability density functions that describe the ensemble of all solutions that fit the data, thus illuminating which model parameters are well-constrained and eliminating the need for regularization. Here we apply Bayesian inversion using a parallel Markov chain Monte Carlo simulation.In this analysis we subdivide the MF and BSF above 10 km depth into a mesh of subfaults. Initial results show that only a few subfaults above 5 km depth have creep rate estimates significantly greater than zero; most of these are ≤5 mm/yr. The depth extent of creep is poorly constrained. The SAF, MF, and BSF slip rate estimates below the assumed locking depths (11 km, 10 km, 15 km) are 18 – 19, 13 – 15, and 10 – 12 mm/yr at 95% confidence, respectively. The sum of these rates exceeds the 40 mm/yr expected from relative plate motion. Ongoing work is focused on examining this discrepancy; the trade-offs among model parameters; and the sensitivity of results to fault geometry, elastic structure, and details of the Bayesian approach.
Session: Including Ground Failure in Scenario Events, Rapid Response, and Loss Estimation Models
Presenter   Dhar, Mahesh
Schedule   Wed AM / Poster
Room   Hall 1
STUDENT
MAEHMP - Updating the Memphis Area Seismic and Liquefaction Hazard Maps
DHAR, M. S., CERI/ University of Memphis, Memphis, TN, msdhar@memphis.edu; CRAMER, C. H., CERI/ University of Memphis, Memphis, TN, ccramer@memphis.edu; VAN ARSDALE, R. B., DES/ University of Memphis, Memphis, TN, rvanrsdl@memphis.edu; PRYNE, D. E., DES/ University of Memphis, Memphis, TN, depryne@memphis.edu; PAUL, J. M., DES/ University of Memphis, Memphis, TN, jmpaul@memphis.edu; PATTERSON, G. L., CERI/ University of Memphis, Memphis, TN, glpttrsn@memphis.edu
In 2004, probabilistic and scenario urban hazard maps were generated for the six-quadrangle area centered on Memphis and southern Shelby County (Cramer et al., 2004, 2006, 2008) in SW Tennessee. These urban seismic hazard maps have been well received, particularly the liquefaction hazard maps, and used by consultants and public agencies to address seismic hazard mitigation in the Memphis area. We are updating and expanding the Memphis hazard maps to all of Shelby County (12+ quadrangles). The 3D geology model has been updated using the Stevens (2007) data and modeling approach plus additional well log control gathered during the summer of 2012. Included in the geologic model is improved modeling of sharp changes due to faults and/or erosion of strata. Seismic hazard maps are generated by folding the effects of local geology into the 2008 USGS national seismic hazard model and attenuation relations (Petersen et al., 2008). Improved hazard calculation approaches from the St. Louis, MO urban hazard mapping project are employed (Cramer 2009, 2011). The liquefaction hazard maps will be generated using the Memphis liquefaction hazard map approach of Cramer et al. (2008) and the liquefaction cumulative potential curves provided by Romaro-Hudock and Rix (2005).We will present seismic and liquefaction hazard maps for the whole of Shelby County and compare them with the older, smoothed model of Gomberg et al. (2003). The improved detail of the new geology model thins the low-velocity layers relative to the older model. This increases seismic hazard relative to the older hazard maps.A technical working group of interested local university, business, and government professionals has been formed to support and review this project. A professional outreach workshop is planned for March 13, 2013 to raise public, business, and professional awareness of Memphis area earthquake hazards, and to disseminate the urban hazard maps produced under this project.
Session: Next Generation of Ground Motion Prediction Models
Presenter   Graizer, Vladimir
Schedule   WITHDRAWN
Room   155C
Updated Graizer-Kalkan Ground Motion Prediction Equations for Western United States
GRAIZER, V., U.S. Nuclear Regulatory Commission, Washington, DC; KALKAN, E., U.S. Geological Survey, Menlo Park, CA
Ground motion prediction equations (GMPE) for peak-ground acceleration (PGA) and 5-percent damped pseudo spectral accelerations (SA) of horizontal component ground motions were developed by Graizer and Kalkan (2007, 2009) using the extended ground motion database of the Next Generation of Attenuation project. The main features of these GMPEs are: (1) minimum number of independent known or measureable parameters (Mw, Rrup, style of faulting and VS30) are used; (ii) predictive model for SA is a continuous function of spectral period (T), which eliminates the standard matrix of estimator coefficients, and allows for calculation of SA at any period of interest within the model range; (3) mathematical form of GMPEs constitutes a series of filters—each filter represents a certain physical phenomenon affecting the radiation of seismic waves from the source. This GMPE is shown to provide improved accuracy (expected median prediction without significant bias) and efficiency (relatively small standard error of predictions) (Graizer and Kalkan, 2011). Data from recent shallow-crustal earthquakes compare well with the predictions by the Graizer-Kalkan GMPEs for earthquakes with 5.0≤M≤7.9 at distances of up to 250 km (Celebi et al., 2010; Akkar et al., 2011; Segou and Kalkan, 2011; Graizer, 2011; Graizer et al., 2013). Here, the Graizer-Kalkan GMPEs are updated by both adding an anelastic attenuation filter as a function of Q and by improving the existing basin-effect filter, which is now a function of depth, distance and period T. Basin parameter is defined as the depth to 1.5 km/s shear-wave velocity isosurface. In our previous SA model, spectral shape decayed at long-periods with a slope of T-1.5, averaging basin and non-basin site effects. In the updated model, spectral shape decays at long-periods faster (T-2) for non-basin sites, and slower (T-1.4) for basin sites. The standard error of the updated GMPEs is on average 4% lower than that of our previous models.
Session: Earthquake Source Physics
Presenter   Martínez - Garzón, Patricia
Schedule   Thu PM / Poster
Room   Hall 1
STUDENT
Stress Tensor Changes Related to Fluid Injection at The Geysers Geothermal Field, California
MARTÍNEZ - GARZÓN, P., Helmholtz-Centre Potsdam German Research Centre for Geosciences, Potsdam, Germany, patricia@gfz-potsdam.de; BOHNHOFF, M., Helmholtz-Centre Potsdam German Research Centre for Geosciences, Potsdam, Germany, bohnhoff@gfz-potsdam.de; KWIATEK, G., Helmholtz-Centre Potsdam German Research Centre for Geosciences, Potsdam, Germany, kwiatek@gfz-potsdam.de
Studying spatiotemporal variations of the stress field caused by massive fluid injection is relevant towards an improved understanding of geomechanical processes in different types of reservoirs. However, a reliable determination of such stress changes based on inversion of focal mechanisms of induced seismicity requires dense local seismic networks with good azimuthal coverage and low magnitude-detection threshold as well as detailed hydraulic information.At The Geysers geothermal field (California), induced seismicity has been carefully monitored for more than 30 years. While it is evident that local seismicity is related to injection operations, it is not trivial to relate the hydraulic parameters from individual wells to the patterns of the crustal stress field and associated seismicity. Earlier attempts to determine the local stress field in the area indicated that the regional tectonic stress field dominates over the stresses induced by reservoir treatment.In this study we investigate spatiotemporal variations of the local stress field orientation at The Geysers geothermal field by using fault plane solutions of local events provided by the Northern California Earthquake Data Center. The inversion scheme is a linear approach with bootstrap resampling to analyze the uncertainties. First, we calculate stress field orientation at different depths over the whole reservoir using high quality focal mechanisms of induced seismicity. We verify the reliability of the stress inversions results by comparing them with those obtained with non-linear approaches. In addition, we search for potential temporal stress direction changes that could be related to periods with higher injection flow rates. For this part we use data from a prominent cluster of induced seismicity associated with an Enhanced Geothermal System project. This study will present the obtained results from described objectives and will discuss their interpretation and geomechanical implications.
Session: Induced Seismicity
Presenter   Wong, Ivan
Schedule   Fri 4:00 PM / Oral
Room   155D
A Probabilistic Framework for Evaluating Injection-Induced Seismicity
WONG, I., URS Corporation, Oakland, CA, ivan.wong@urs.com; NEMSER, E., URS Corporation, Oakland, CA; KULKARNI, R., URS Corporation, Oakland, CA
With the growth over the past decade in deep fluid injection related to oil and gas activities, risk-related issues pertaining to induced seismicity have become increasingly relevant and critical. The recent potential cases of fluid injection-induced seismicity near Guy, Arkansas; central Oklahoma; the Dallas-Fort Worth, Texas Airport; and Raton Basin, Colorado highlight these issues. However, it remains extremely difficult to make a definitive determination of whether an occurrence of seismicity is induced or not because of generally incomplete information and because of the large uncertainties in making such an assessment. For the same reasons it is also challenging to predict whether seismicity will be induced in the future at a site due to fluid injection. As is the state-of-the-practice in assessing the seismic hazards from earthquakes, it is appropriate to use a probabilistic approach to addressing the question of whether observed seismicity is induced or naturally-occurring; a probabilistic framework can also be used to predict the likelihood of future induced seismicity due to fluid-injection. We have developed probability evaluation matrices that can be used to evaluate whether 1) observed seismicity in the vicinity of active fluid injection wells has been induced by fluid injection; 2) active fluid injection wells without associated observed seismicity are likely to induce future seismicity; or 3) proposed (inactive) fluid injection wells are likely to induce future seismicity. Factors considered in the evaluation are the historical and contemporary seismicity, structural and hydrologic setting, injection history and parameters, induced pore pressure field, and the in situ tectonic stress field characteristics. To test the utility of these matrices, we provide examples using several well-documented examples of fluid injection-induced seismicity.
Session: Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Models
Presenter   Schmauder, Gretchen
Schedule   Fri 4:15 PM / Oral
Room   155C
STUDENT
Modeling Ground Motions in the Southern Lake Tahoe Basin, California and Nevada, Resulting from Rupture on the West Tahoe Fault
SCHMAUDER, G. C., University of Nevada, Reno, NV, gcschmauder@yahoo.com; LOUIE, J., University of Nevada, Reno, NV, louie@unr.edu; HALL, K. A., Whitmen College, Walla Walla, WA, hallka@whitmen.edu; GRAY, K. T., University of Nevada, Reno, NV, kylethomasgray@gmail.com; MCBEAN, A., University of Nevada, Reno, NV, alexamcbean@gmail.com; MCBEAN, K., University of Nevada, Reno, NV, kevinmcbean@gmail.com; KENT, T., University of Nevada, Reno, NV, tkent42@gmail.com
Recent studies in the Lake Tahoe basin (LTB) using both airborne remote sensing and shallow geophysical studies indicate that the West Tahoe fault (WTF), forming the western edge of the LTB, is responsible for large earthquakes within the past 11,000 years and that there is the potential for a future earthquake approaching a magnitude 7.0. An earthquake of this magnitude could have significant repercussions for the communities situated in the south Tahoe basin (STB), which is underlain primarily by quaternary sediments and, in the Tahoe Keys area, imported fill. Ground amplification resulting from an earthquake on the WTF may be significant in this region due, not only to the geotechnical properties of the basin fill, but also to the depth and focusing properties of the basin. This study begins to quantify the earthquake-shaking hazard in the STB using new geotechnical and geophysical data specific to the STB. We collected approximately 70 shallow shear-wave velocity profiles with Optim’s SeisOpt® ReMi™ method throughout the basin to determine the Vs30 values. We acquired 29 new gravity measurements and combined them with existing data to map a total of 270 gravity points. Geosoft’s Oasis Montaj helped us model the basin shape and thickness from the gravity data. The maps obtained from the Vs30 measurements and the gravity modeling are imported into the Nevada ShakeZoning (NSZ) community seismic modeling environment along with the location and moment of scenario earthquakes on the WTF. LLNL’s E3D finite-difference code propagates synthetic waves through the model STB assembled by NSZ. Using the specific geotechnical and geophysical information, as well as a reasonable rupture scenario, the model predicts Peak Ground Velocity (PGV) and shaking time histories for the southern portion of the Tahoe Basin. Areas identified as having high shaking are mapped and may be used by the Tahoe community for earthquake hazard mitigation.
Session: Characterizing Active Faults for Seismic Hazard Assessments
Presenter   Cramer, Chris
Schedule   Fri AM / Poster
Room   Hall 1
A New Macroseismic Intensity Prediction Equation and Magnitudes of the 1811-1812 New Madrid and 1886 Charleston, South Carolina, Earthquakes
BOYD, O. S., USGS, Memphis, TN, olboyd@usgs.gov; CRAMER, C. H., CERI, The University of Memphis, Memphis, TN, ccramer@memphis.edu
We develop a macroseismic intensity prediction equation (MIPE) for the central and eastern United States and estimate the moment magnitudes of the 1811–1812 New Madrid, MO and 1886 Charleston, SC earthquakes. We constrain the equation with North American census data, the National Oceanic and Atmospheric Administration modified Mercalli intensity (MMI) dataset having responses to questionnaires for earthquakes between 1924 and 1985, and the USGS ‘Did You Feel It?’ MMI dataset having web-based responses for earthquakes occurring between June, 2000 and August, 2012. The combined intensity dataset has more than 500,000 felt reports for 517 earthquakes with magnitude between 2.5 and 7.2.We assume that hypocentral distances for a given reported intensity are log-normally distributed. The coefficients of an MIPE predicting the mean of the log normal distribution of reported intensities as a function of MMI and moment magnitude are found by inversion. To find the magnitude for the older 1811–1812 New Madrid and 1886 Charleston intensity datasets, we simulate the distribution of the number of felt reports versus distance for each dataset and MMI to estimate the mean log-normal distance. The MIPE is then solved for magnitude, and the magnitudes determined for each MMI are combined in a weighted averaged where the weights are the inverse of the uncertainty in the magnitude estimate.We find that the new relation leads to magnitude estimates for the New Madrid earthquakes that are within the range of those determined previously. Depending on the MMI dataset used, of which we consider three for the New Madrid mainshocks, the new relation results in estimates for the moment magnitudes of the December 16th, 1811, January 23rd, 1812, and February 7th, 1812 mainshocks and December 16th dawn aftershock of 7.1–7.8, 6.8–7.6, 7.5–8.0, and 6.5–7.1, respectively, with a magnitude uncertainty of ±0.4–0.5. We estimate a magnitude of 7.1±0.4 for the 1886 Charleston, SC earthquake.
Session: Induced Seismicity
Presenter   Rubinstein, Justin
Schedule   Fri 1:30 PM / Oral
Room   155D
Addressing the Question of the Extent to which Earthquakes Induced by Fluid Injection at Depth Contribute to Seismic Hazard
MCGARR, A., USGS, Menlo Park, CA, mcgarr@usgs.gov
Activities involving the extraction of energy resources from low-permeability formations have resulted in many reports of earthquakes associated with these operations, which involve the injection of fluids at depth. These activities include (1) hydraulic fracturing of shales or coal seams to extract gas and oil (“fracking”), (2) disposal of the wastewater from these gas and oil activities by injection into deep aquifers, and (3) the development of Enhanced Geothermal Systems (EGS) by injecting water into hot, low-permeability rock. Of these three operations, wastewater disposal is observed to be associated with the largest, most damaging earthquakes, with maximum magnitudes sometimes exceeding 5. The largest earthquakes induced by EGS projects are somewhat smaller and, with a few exceptions, earthquakes due to fracking are much too small to be felt at the surface. Accordingly, the deep disposal of wastewater seems to be the principal cause of earthquakes that might contribute materially to the seismic hazard because of two factors: (1) wastewater disposal involves the largest volumes of injected liquid and (2) the number of disposal wells vastly exceeds the number of EGS injection wells. To estimate the maximum size earthquake that could be induced by a given fluid injection project, we assume that the rock mass is fully saturated, brittle, and responds to injection with a sequence of earthquakes having a typical magnitude-frequency distribution. If so, then it is easy to show that the maximum seismic moment, directly related to injection, is given by the volume of injected liquid times the modulus of rigidity. Evidence from numerous case histories involving wastewater disposal and EGS operations are consistent with this injection-only upper bound on seismic moment. Although we cannot rule out the possibility of stress triggering leading to earthquakes larger than this proposed upper bound, there is, so far, no solid evidence that supports such an outcome.
Session: The Magnitude X.X Earthquake on the YY of ZZZZ: Major Earthquakes of 2012/13
Presenter   Kwong, Kevin
Schedule   Fri PM / Poster
Room   Hall 1
STUDENT
Tracking the Short Period Energy Release of the 28 October 2012 Mw 7.7 Haida Gwaii Earthquake
KWONG, K. B., University of Utah, Salt Lake City, UT; KOPER, K. D., University of Utah, Salt Lake City, UT
The Mw 7.7 thrust earthquake that struck off the coast of Haida Gwaii on October 28, 2012 was the largest Canadian earthquake since the 1949 M 8.1 Queen Charlotte Islands event. Both earthquakes occurred along the Queen Charlotte Fault zone, a transpressive boundary between the Pacific and North American plates with about 50 mm/yr of right-lateral motion and 15-20 mm/yr of convergent motion. While most previous seismicity in the region was strike-slip, and no Benioff zone exists, receiver-function based imaging landward of the boundary showed evidence of a subducting slab, suggesting that large thrust earthquakes in this area were possible (Bustin et al., 2007). In this study, we investigate the short-period rupture properties of this unique earthquake with back-projection analyses of teleseismic P-waves. We use conventional back-projection (Xu et al., 2009), in a range of frequency bands between 0.1-2.0 Hz, with data from arrays of receivers in Europe, North America, Japan, and the Pacific Ocean, as well as a global configuration of mostly GSN stations. The most robust results come from regional arrays to the northeast (Europe) and southeast (eastern North America, Central American, and northern South America) and show bilateral rupture initially concentrated SE of the epicenter and later to the NW, lasting for a total of about 50-60 s. Interestingly, our results using the USGS hypocenter show peak short period energy release concentrated "down-dip" on the North American side of the boundary, while most of the early aftershock locations are concentrated, seaward on the Pacific Plate. We present robustness tests of these preliminary back-projection models using various hypocentral solutions and back-projections of synthetic data.
Session: Intermountain West Earthquake and Volcano Characterization and Related Hazards
Presenter   Ruhl, Christine
Schedule   Wed PM / Poster
Room   Hall 1
STUDENT
Stress Field Analysis of the Southern to Central Walker Lane Transition Zone, Western Nevada from Seismicity, Moment Tensors, and Short-Period Focal Mechanisms
RUHL, C. J., Nevada Seismological Laboratory, University of Nevada, Reno, NV, christineruhl@gmail.com; SEAMAN, T. C., Nevada Seismological Laboratory, University of Nevada, Reno, NV, tceeman@gmail.com; SMITH, K. D., Nevada Seismological Laboratory, University of Nevada, Reno, NV, ken@seismo.unr.edu
The dominantly NW-striking dextral strike-slip faults of the Central and Southern Walker Lane (CWL and SWL, respectively) of eastern CA and western NV are separated by a complex zone of NE-trending normal and primarily sinistral strike-slip faults, and recent volcanics, described as the Mina Deflection (MD) (Stewart, 1985; Faulds and Henry, 2008). The MD accommodates the transfer of right-lateral strike-slip motion from the SWL northward to the CWL east of the Wassuk Range over an ~50 km wide, ~80 km long right-step in a dramatic transition of the Walker Lane Belt strain fabric. Several slip transfer mechanisms have been proposed within the MD, from clockwise rotation of high-angle fault blocks (Wesnousky, 2005), to low-angle displacement within the Silver Peak-Lone Mountain complex (Oldow et al., 2001), and curved fault arrays associated with localized basins and tectonic depressions (Ferranti et al., 2009). The MD region has been a regular source of M4+ to moderate sized events, many of which occur in spatially distributed swarm-like sequences. Historical events include the 1934 M6.5 Excelsior Mountains event south of Mina, NV, and the 1932 M7.1 Cedar Mountains event east of the Pilot Mountains. Another feature of the seismicity is an ~40 km long arcuate distribution of activity extending from north of the White Mountains to Mono Lake that appears to reflect a southwestern boundary to NE-striking structures in the MD. We compile published and computed MT solutions for M3.5+ earthquakes as well as developed short-period focal mechanisms to evaluate the stress field in the MD-CWL-SWL transition. Inversions of fault plane solutions are computed where data is dense to better estimate the stress field. Based on the complex distribution of fault orientations, stress orientations are highly variable throughout the MD, however, in many cases, active fault planes can be isolated from high-precision locations to better constrain stress and slip orientations.
Session: Characterizing Active Faults for Seismic Hazard Assessments
Presenter   Li, Zefeng
Schedule   Fri AM / Poster
Room   Hall 1
STUDENT
Polarization Analysis of Fault Zone Head Waves Along the Parkfield Section of the San Andreas Fault
LI, Z., Georgia Institute of Technology, Atlanta, GA, zli354@gatech.edu; PENG, Z., Georgia Institute of Technology, Atlanta, GA, zpeng@gatech.edu; ZHAO, P., NORSAR, Norway, peng.zhao.gatech@gmail.com; BENNINGTON, N. L., University of Wisconsin-Madison, Madison, WI, ninfa@geology.wisc.edu; THURBER, C., University of Wisconsin-Madison, Madison, WI, clifft@geology.wisc.edu
Large mature faults like the San Andreas Fault generally have well-defined bimaterial interfaces that separate blocks with different elastic properties. Such a sharp material contrast is expected to generate fault zone head waves (FZHW) that propagate along the interface and are then recorded by stations on the slower velocity side. If the normal distance between a station and the fault is less than a critical distance, FZHW would arrive earlier than the direct P wave. Because the first-motion polarity of the FZHW is generally opposite to that of the direct P wave, most previous studies have utilized polarity of seismic waves in the vertical component to identify FZHW (Zhao et al., 2010). Recent studies (e.g., Bulut et al., 2012) show that horizontal polarizations of the FZHW and direct P wave are also quite different and could be used as another diagnostics to distinguish them. Here we examine polarizations of earthquakes that occurred along the Parkfield section of the San Andreas Fault and were recorded by the Parkfield Area Seismic Observatory (PASO) in 2001-2002. We plot evolving particle motions of the horizontal components and observe changes in the polarizations from a certain angle against the fault strike for FZHW to source-receiver directions for the direct P wave. We also measure the polarization direction and linearity of the FZHW and direct P wave, and the time delay between them. Our preliminary results show clear changes in the polarization direction between FZHW and direct P wave for many events, demonstrating its effectiveness in identifying FZHW and direct P wave phases. We are also in the process of applying the same procedure to identify possible S-type FZHW phases. Updated results will be presented at the meeting.
Session: Intermountain West Earthquake and Volcano Characterization and Related Hazards
Presenter   Kreemer, Corné
Schedule   Wed 8:30 AM / Oral
Room   155B
Improved Seismic Hazard Assessment for the Colorado Plateau Region from New Geodetic Data
KREEMER, C., University of Nevada, Reno, NV, kreemer@unr.edu; BLEWITT, G., University of Nevada, Reno, NV, gblewitt@unr.edu; HAMMOND, W. C., University of Nevada, Reno, NV, whammond@unr.edu; BROERMANN, J., University of Arizona, Tucson, AZ, rab@geo.arizona.edu; BENNETT, R. A., University of Arizona, Tucson, AZ, james@email.arizona.edu
Much of the Colorado Plateau region is geologically dramatic yet seismically enigmatic. Away from the Intermountain Seismic Belt, seismicity is sparse and recurrence times for large earthquakes in the Intermountain West are long. While there is evidence for large earthquakes from the Wasatch to the Sonoran Desert, only for the Wasatch Fault may there be enough evidence from microseismicity and paleoseismology to robustly constrain seismic hazard models. In light of these data deficiencies, very precise geodetic observations of strain accumulation may provide essential constraints on future seismic activity. Until recently, geodetic coverage of the area was very sparse. Much improvement was made with the installation of EarthScope's Plate Boundary Observatory, but much of the Colorado Plateau and surroundings were still left largely unmonitored. To improve constraints on crustal deformation rates, we installed 34 new continuous GPS stations in southern Utah, Nevada, and Arizona, and new semi-continuous stations across Las Vegas Valley and the Pahranagat Shear Zone. Moreover, GPS networks by various public and commercial organizations are now providing data, although station stability remains to be examined. We processed all data uniformly using the latest geodetic models. Time-series for all stations are now 2+ year long and we will present a significantly densified GPS velocity field. We will compare results against our recent hypotheses on Colorado Plateau motion/deformation, motion across the Pahranagat Shear Zone, and overall distribution of strain rates. We will also discuss the usefulness of data from non-tectonic networks and the evidence for non-tectonic deformation (e.g., hydrologic subsidence). We will also highlight that for all stations for which data is available the next day we process and make available 5-min solutions, useful for earthquake response purposes.
Session: Intermountain West Earthquake and Volcano Characterization and Related Hazards
Presenter   Farrell, Jamie
Schedule   Wed 11:30 AM / Oral
Room   155B
STUDENT
Crustal Velocity Structure and Seismicity of the Yellowstone Volcanic System from Automated Waveform Analysis of Body Waves, 1984-2011
FARRELL, J., University of Utah, Salt Lake City, UT, jamie.farrell@utah.edu; HUSEN, S., Swiss Seismological Service, Zurich, Switzerland, stephan.husen@sed.ethz.ch; SMITH, R. B., University of Utah, Salt Lake City, UT
Past tomographic studies of the Yellowstone magma system reveal a distinctive, low P-wave crustal anomaly beneath the 0.64 Ma caldera that has been interpreted to be the shallow crustal magma reservoir that provides the energy for Yellowstone’s youthful volcanic and hydrothermal systems. Until recently, limited seismometer coverage did not allow us to adequately resolve changes in seismic velocity north of the caldera, where recent gravity modeling reveals a shallow, low density body that extends ~20 km north of the caldera. Recent upgrades to the Yellowstone Seismic Network (YSN), including the addition of nine 3-component and broadband seismic stations provide much better ray coverage of the entire Yellowstone area with greater bandwidth data. This allows us to produce much-expanded and improved resolution images of the Yellowstone crustal velocity structure. We have compiled all the waveforms for the Yellowstone region earthquake catalog from 1984-2011 to analyze P-wave arrival times with an automatic picker based on an adaptive high-fidelity human mimicking algorithm. The automatic picker is calibrated using a reference dataset of 172 events that are manually picked based on strict criteria. There are approximately 33,000 earthquakes in the catalog and we will use roughly 115,000 P-wave arrival times to invert for the velocity structure. A complete dataset of earthquake first-arrival picks and consistent uncertainty estimations through time of the YSN allows us to better understand the velocity structure of the Yellowstone volcanic system and examine changes in the structure with time. In addition to inverting for velocity structure, this dataset is used to invert independently for S-wave velocity and attenuation variations associated with the Yellowstone magma reservoir and determine the complete crust and upper mantle structure.
Session: The Magnitude X.X Earthquake on the YY of ZZZZ: Major Earthquakes of 2012/13
Presenter   Mohammadebrahim, Ehsan
Schedule   Fri PM / Poster
Room   Hall 1
STUDENT
The Seismic Strong Motion Array Project (SSMAP) and the September 5, 2012 Mw=7.6 Nicoya, Costa Rica Earthquake: Preliminary Results
MOHAMMADEBRAHIM, E., California State University of Northridge, Northridge, CA, ehsan.mohammadebrahim.307@my.csun.edu; SIMILA, G., California Sate University of Northridge, Northridge, CA, gerry.simila@csun.edu
Since 2006, the seismic strong motion array project (SSMAP) for the Nicoya Peninsula in northwestern Costa Rica has been composed of 10 sites with Geotech A900/A800 accelerographs (three-component) and GPS timing. Our digital accelerograph array has been deployed as part of our ongoing research on large earthquakes in conjunction with the Earthquake and Volcano Observatory (OVSICORI) at the Universidad Nacional in Costa Rica. The country wide seismographic network has been operating continuously, and has been upgraded with broad-band seismometers and Episensors. The previous major earthquake in Nicoya was the 1950 M¬w 7.8 event. On September 5, 2012, a Mw=7.6 earthquake occurred in the seismic gap and appears to be the expected event based on recurrence interval, but Protti (2012) states that only 40% of the energy calculated to have accumulated since the 1950 quake was released by the September earthquake. The main shock focal mechanism is thrust faulting of the Cocos plate in the Middle America trench with strike N54W and dip 20 degrees NE. The preliminary result for the mainshock and associated aftershocks is a rupture zone of 79 km length and 45 km width. The maximum accelerations from two A900 stations perpendicular to the trench, Fortuna (distance 112km) and Pedernal (distance 128 km) are: 13.8% and 8.9 % g, respectively. The October 10 (MW 5.3) and 24 ( Mw 6.6) aftershocks recorded at Tamarindo (distances 40 km and 70 km, respectively) show accelerations of 2.4% and 8.2% g; respectively. The acceleration data are compared to existing attenuation relationships. Also, additional aftershock acceleration data and previous events for 2010-12 will be presented.
Session: Earthquake Source Physics
Presenter   Kwiatek, Grzegorz
Schedule   Thu PM / Poster
Room   Hall 1
The Spectral Ratio Method as a Tool to Investigate the Source Scaling Relations from km- to cm-scale (MW4 to -6) Earthquakes: Experiences from Induced seismicity, Volcanic-hybrid Seismic Events and Laboratory Experiments
KWIATEK, G., GFZ German Research Centre for Geosciences, Potsdam, Germany, kwiatek@gfz-potsdam.de; DRESEN, G., GFZ German Research Centre for Geosciences, Potsdam, Germany, dre@gfz-potsdam.de; BOHNHOFF, M., GFZ German Research Centre for Geosciences, Potsdam, Germany, bohnhoff@gfz-potsdam.de; HARRINGTON, R. M., Karlsruhe Institute of Technology, Karlsruhe, Germany, rebecca.harrington@kit.edu; BULUT, F., GFZ German Research Centre for Geosciences, Potsdam, Germany; GOEBEL, T., University of Southern California, Los Angeles, CA
A key question in seismology is whether large and small earthquakes are governed by the same physics and to what extent dynamic rupture and faulting observed in laboratory tests may provide insights into the nucleation and fracture processes of larger earthquakes. Studies of earthquake scaling relations involve comparison of static (stress drop) and dynamic (apparent stress) source parameters. Many authors have suggested that the stress drop is roughly independent of seismic moment suggesting a self-similar rupture process. It is still a matter of debate whether the apparent stress is constant over the whole magnitude range. A breakdown of self-similar scaling may occur due to the physical processes. However, it may likewise also be a result of poor data quality, selection and processing. For example, inaccurate knowledge on attenuation and scattering of seismic waves are found to be an important source of bias refraining from the reliable estimation of source characteristics.In this study we summarize results from recent scaling relations studies for seismicity covering a magnitude range between MW-6 and MW4 representing source radii in the range 1cm-1km. We considered seismic recordings from various environments including mining and fluid-induced seismicity, volcanic-hybrid earthquakes, and acoustic emissions from laboratory studies on rock samples. We focused on reliable assessment of source parameters in order to investigate the self-similarity of earthquake rupture process. We suppressed unknown medium effects introduced into the waveforms applying the spectral ratio technique to all datasets. The obtained results show that inappropriately corrected path effects as well as wrong selection of input data introduce an apparent breakdown in both static and dynamic scaling relations whereas spectral-ratio refined data confirm the self-similarity of earthquake rupture processes for small earthquakes down to at least MW-4.1.
Session: Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Models
Presenter   Wang, Feng
Schedule   Thu 4:15 PM / Oral
Room   155C
STUDENT
Comparison of Physics-Based Models and Ground Motion Prediction Equations in Seismic Hazard Analysis for Southern California
WANG, F., University of Southern California, Los Angeles, CA, fengw@usc.edu; JORDAN, T. H., University of Southern California, Los Angeles, CA, tjordan@usc.edu
CyberShake hazard model (CSHM) uses physics-based 3D earthquake simulations to explicitly incorporate deterministic source and wave propagation effects within seismic hazard calculations for Southern California. Seismic hazard models based on empirical ground motion prediction equations (GMPEs) account for the average source, propagation, and path effects by using functionals of model parameters. We have developed an averaging-based factorization (ABF) scheme that conforms to the implementation of ground motion predictions in seismic hazard analysis, and facilitates the geographically explicit comparisons between these models. Through a sequence of averaging operations over various model parameters, we uniquely factorize model residuals into several terms, which then characterize differences in various physical effects between models, such as basin excitation, distance attenuation, and source directivity. We compare CSHM with the Next Generation Attenuation (NGA) GMPEs using this technique. For spectral acceleration at long periods (>2s), NGA models underestimate the basin-excitation effects by up to a factor of 3 relative to CSHM; Moreover, basin excitations in the physics-based model are not a simple function of basin depth as parameterized in the NGA models. Using the directivity relations of Spudich and Chiou (2008), we quantify the extent to which the empirical directivity models capture the source directivity effects predicted by physics-based model. We found that the empirical directivity corrections for NGA models underestimate source directivity effects in CSHM, and they do not account for the directivity-basin coupling that substantially enhances the low-frequency seismic hazards in the sedimentary basins of the Los Angeles region. We also quantify the degree to which the use of physics-based CyberShake calculations can reduce uncertainties in seismic hazard estimations.
Session: Oceanographic and Atmospheric Signals in Seismology
Presenter   Anthony, Robert
Schedule   Thu 3:45 PM / Oral
Room   155D
STUDENT
Global Trends in Ocean Wave State and Extremal Storm Events through Microseism Analysis
ANTHONY, R. E., New Mexico Tech, Socorro, NM, ranthony@nmt.edu; ASTER, R. C., New Mexico Tech, Socorro, NM, aster@ees.nmt.edu; ROWE, C., Los Alamos National Laboratory, Los Alamos, NM, char@lanl.gov
The Earth’s seismic noise spectrum features two globally ubiquitous peaks, near 8 and 16 s period that arise when storm-generated oceanic waves are converted to seismic energy, predominantly as Rayleigh waves. Because of its regionally integrative nature, microseism intensity histories at long running sites can provide useful proxies for ocean sea state. Expanding on an earlier study of global microseism trends (Aster et al., 2010), we analyze multi-decadal seismic data from global stations associated with several networks to characterize the spatiotemporal evolution of wave climate through early 2013. Ground motion power spectral density (PSD) is calculated over 3-hour overlapping time series segments to produce a database of PSD statistics at each broadband station between 2 and 100 s. Isolating power in the primary and secondary microseism bands enables characterization of spatially-integrated trends in wave states and periods near the coastlines surrounding each station. In addition, extremal storm events are detected and are used to assess decadal changes in the location and frequency of oceanic storm activity, such as is hypothesized to occur under climate change scenarios. The results of these analyses are then correlated with recognized modes of atmospheric variability (e.g., El Nino-Southern Oscillation, Southern Annular Mode), which can impact storm statistics and thus regional wave states. Changes in ocean wave energy have the potential to drastically impact coastal environments and to impact cryospheric stability, particularly under rising global sea levels. Seismic characterization of global sea state through microseism analysis should usefully complement buoy, hindcast, satellite based, and other methods of characterizing ocean wave activity on hourly to decadal time scales.
Session: Induced Seismicity
Presenter   Rubinstein, Justin
Schedule   Fri 1:45 PM / Oral
Room   155D
The 2001 – Present Triggered Seismicity Sequence in the Raton Basin of Southern Colorado/Northern New Mexico
RUBINSTEIN, J. L., USGS, Menlo Park, CA, jrubinstein@usgs.gov; ELLSWORTH, W. L., USGS, Menlo Park, CA, ellsworth@usgs.gov; MCGARR, A., USGS, Menlo Park, CA, mcgarr@usgs.gov
The occurrence of an earthquake of magnitude 5.3 near Trinidad, CO, on 23 August 2011 renewed interest in the possibility that an earthquake sequence in this region that began in August 2001 is the result of industrial activities. Our investigation of this seismicity, in the Raton Basin of northern New Mexico and southern Colorado, led us to conclude that most of the earthquakes since August 2001 have been triggered by the deep injection of wastewater related to the production of natural gas from the coal-bed methane field here. The evidence that this earthquake sequence was triggered by wastewater injection is threefold. First, there was a marked increase in seismicity shortly after major fluid injection began in the Raton Basin. From 1970 through July of 2001, there were 5 earthquakes of M≥3 in the Raton Basin, and there were 100 from August of 2001 through the end of 2012. The statistical likelihood of this rate increase occurring naturally is 0.01%. Second, the vast majority of the seismicity is located close (within 5km) to active disposal wells in this region. This seismicity is primarily shallow, ranging in depth between 2 and 8 km, with the shallowest seismicity occurring within 500 m depth of the injection intervals. Finally, these wells have injected exceptionally high volumes of wastewater. Two high-volume disposal wells, located adjacent to the August 2011 M5.3 earthquake, injected about 4.9 million cubic meters of wastewater during the period leading up to the earthquake. This is more than 7 times as much as the disposal well at the Rocky Mountain Arsenal that caused damaging earthquakes in the Denver, CO region in the 1960s. The August 2011 M5.3 is the largest earthquake to date for which there is compelling evidence of triggering by fluid injection activities. Much of the seismicity since 2001 falls on a 15km-long, NE-trending lineation of seismicity dipping steeply to the SE, which is consistent with focal mechanisms and the regional tectonic regime.
Session: Advances in High-Frequency Physics-Based Earthquake Systems Simulation
Presenter   Shi, Zheqiang
Schedule   Thu 8:45 AM / Oral
Room   155C
Validation of Dynamic Rupture Simulations for High-Frequency Ground Motion
SHI, Z., San Diego State University, San Diego, CA, zshi@projects.sdsu.edu; DAY, S. M., San Diego State University, San Diego, CA, sday@mail.sdsu.edu
We generate high-frequency ground motion deterministically by performing 3-D dynamic rupture simulations on rough faults. The model fault geometry follows a self-similar roughness distribution with wavelength scales spanning three orders of magnitude, from ~102 m to ~105 m. In addition, our model accounts for strongly rate-weakening friction and off-fault inelastic yielding. The rupture simulations resolve wavefield spectral components up to greater than 10 Hz, permitting comparisons with empirical ground-motion intensity measures over much of the frequency range of engineering interest. Our simulation results show that dynamic rupture propagation on rough faults results in rupture irregularities that lead to ground motions with complex spatial patterns and extensive high-frequency content. The simulated ground accelerations have near-flat power spectra from a few tenths of a Hz to slightly less than 10 Hz. Our initial studies with a halfspace medium produced site-averaged synthetic response spectra having characteristics, including the distance and period dependence of the median values, absolute level and intra-event standard deviation, comparable to appropriate empirical estimates, throughout the period range 0.1-3.0 sec. We extend the preliminary validation study by incorporating more realistic media, different realizations of fault roughness, and a range of target event sizes.
Session: Seismic Hazards and Ground Motions
Presenter   Walling, Melanie
Schedule   Fri 10:45 AM / Oral
Room   155B
Seismic Hazard from Distant Sources in Eastern North America
WALLING, M. A., Lettis Consultants International, Inc., Boulder, CO, phillips@lettisci.com; VLASITY, J. A., Lettis Consultants International, Inc., Boulder, CO, phillips@lettisci.com; MCGUIRE, R. K., Lettis Consultants International, Inc., Boulder, CO, phillips@lettisci.com
In eastern North America, earthquake ground motions attenuate with distance at a slower rate than in plate margin regions, and rates of seismicity are generally low. These factors mean that seismicity beyond 320 km (which is an inclusion distance for background sources used in some guidelines) can contribute to seismic hazard at a site. Regions of higher-than-average historical seismicity (e.g. the eastern Tennessee seismic zone) and sources of repeated, large-magnitude earthquakes (e.g. the New Madrid seismic zone) are examples of sources that may contribute to seismic hazard at distances beyond 320 km. In hazard assessment of a site, a convenient rule is that sources of seismic hazard should be included so that at least 99% of the total mean hazard (mean annual frequency of exceedence) is calculated, for ground motion amplitudes of interest for seismic design and for all spectral frequencies of interest. These ground motion amplitudes typically correspond to mean return periods between 2,500 years and 10,000 years, depending on the application. Seismic hazard at low spectral frequencies (e.g. 1 Hz and lower) is generally more affected by the distance inclusion criterion than seismic hazard at high spectral frequencies. Example seismic hazard results are presented for sites in the southeastern US and in the central Atlantic coast, showing how distant sources can contribute to hazard. Following this 99% rule will mean that the total mean hazard will be accurately calculated for spectral frequencies of interest for seismic design, and the resulting seismic design spectra will be accurate.
Session: Broadband Seismic Observations on the Seafloor
Presenter   Shen, Yang
Schedule   Wed AM / Poster
Room   Hall 1
Offshore Structure of the Cascadia Subduction Zone from Full-Wave Ambient Noise Tomography
GAO, H., University of Rhode Island, Narragansett, RI, hgao@gso.uri.edu; SHEN, Y., University of Rhode Island, Narragansett, RI
We report a preliminary offshore model of the crust and uppermost mantle at the Cascadia subduction zone using a full-wave ambient noise tomographic method. In total, we use more than 150 seismic stations from 2011-2012, including 56 ocean bottom seismometers deployed by the Cascadia Initiative community experiment and Neptune Canada, and about 100 broadband stations on land. The vertical components of the continuous seismic records are normalized with a frequency-time waveform normalization method, and then cross-correlated between each station pair to extract empirical Green’s functions at periods of 7-50 s. We simulate wave propagation within a 3D Earth structure using a finite-difference method to generate a station Strain Greens Tensor database and synthetic waveforms. Rayleigh wave phase delays are obtained by cross-correlating the observed and synthetic waveforms. The sensitivity kernels of Rayleigh waves on the perturbations of Vp and Vs are calculated based on the Strain Greens Tensor database. We then invert for the velocity perturbation from the reference model and progressively improve the model resolution.
Session: Intermountain West Earthquake and Volcano Characterization and Related Hazards
Presenter   Gammans, Christine
Schedule   Wed PM / Poster
Room   Hall 1
STUDENT
Analysis of Aftershocks from the 3 January 2011 Mw 4.5 Tushar Mountains (Utah) Earthquake
GAMMANS, C. N. L., Univeristy of Utah, Salt Lake City, UT, christine.gammans@utah.edu; PANKOW, K. L., Univeristy of Utah, Salt Lake City, UT, pankow@seis.utah.edu; PECHMANN, J. C., Univeristy of Utah, Salt Lake City, UT, pechmann@seis.utah.edu; WHIDDEN, K. M., Univeristy of Utah, Salt Lake City, UT, whidden@seis.utah.edu; KOPER, K. D., Univeristy of Utah, Salt Lake City, UT, koper@seis.utah.edu
On January 3, 2011, an Mw 4.5 earthquake occurred in the Tushar Mountains near Circleville, Utah (38.246°, -112.325°). The Tushar Mountains are located in the transition zone between the stable Colorado Plateau (CP) to the east and the deforming Basin and Range (BR) province to the west. In this area, seismicity associated with the Intermountain Seismic Belt is relatively common. The University of Utah Seismograph Stations (UUSS) detected and located 97 aftershocks in the 33 weeks following the mainshock. On January 6, UUSS installed a portable station in the source region. Using three aftershocks recorded by the portable station as master events, including the largest (Mw 3.8), we relocated the mainshock/aftershock sequence. These refined locations were used as initial locations for the HypoDD method of Waldhauser and Ellsworth (2000) to produce a second, improved set of relocations. In addition to P- and S-arrival time picks, we used the lag-times from waveform cross-correlations as additional input into HypoDD for more accurate relocations. We also found the best empirical Green’s function event for the mainshock using cross-correlation analysis. Lastly, we determined focal mechanisms for several of the events using both first motion and waveform inversion methods. We will compare the results from these analyses to the local geology and past seismicity studies in order to improve understanding of deformation in the CP-BR transition zone.
Session: Oceanographic and Atmospheric Signals in Seismology
Presenter   Sufri, Oner
Schedule   Thu 4:00 PM / Oral
Room   155D
STUDENT
Microseisms from Superstorm Sandy
SUFRI, O., Dept. of Geology and Geophysics, University of Utah, Salt Lake City, UT, oner.sufri@gmail.com; KOPER, K. D., Dept. of Geology and Geophysics, University of Utah, Salt Lake City, UT, koper@seis.utah.edu
Superstorm Sandy was one of the most devastating and costly storms in recent history. It was the largest Atlantic hurricane on record with winds spanning 1,100 miles in diameter. We visualized and analyzed the microseisms generated by Sandy as recorded by the Earthscope Transportable Array (TA). We downloaded 19 days of continuous data from 428 broadband, three-component seismometers and performed continuous, frequency-dependent polarization analysis (Koper and Hawley, 2010). Instrument responses were removed from hour-long segments of data, which were then divided into ten overlapping sub-windows of length 819.2 s. Each sub-window was detrended, tapered with a 10% Hanning function and processed with a fast Fourier transform. For each sub-window, spectral covariance matrices were constructed from the three components at each station, and then linearly averaged to generate the overall covariance matrix for the original one-hour long time period. A log10 based averaging scheme was used to reduce the number of frequency bins from 32,768 to 301. Finally, eigen-analysis was performed on the spectral matrix, and polarization attributes were extracted from the dominant eigenvector. The results are visualized in a series of animations showing the spatial and temporal variation of various polarization attributes, at various periods, across the TA, as well as in individual spectrogram plots. We found that when Sandy sharply changed its direction due west-northwest (towards the East coast, specifically Long Island and New York), the maximum microseismic power and degree of polarization were observed across the TA; however, this energy level was comparable to the background microseismic field created by wave action in the North Atlantic and North Pacific oceans. We also observed finger-like structures in the amplitude field across the TA, that we interpret as a complex pseudo-radiation pattern, possibly created by near-surce bathymetry changes.
Session: Earthquake Source Physics
Presenter   McLaskey, Gregory
Schedule   Thu 9:15 AM / Oral
Room   155A
Seismic Source Spectra of Laboratory Earthquakes
MCLASKEY, G. C., USGS, Menlo Park, CA, gmclaskey@usgs.gov; KILGORE, B. D., USGS, Menlo Park, CA, bkilgore@usgs.gov; LOCKNER, D. A., USGS, Menlo Park, CA, dlockner@usgs.gov; BEELER, N. M., USGS, Menlo Park, CA, nbeeler@usgs.gov
Stick-slip instabilities observed in laboratory experiments are thought to be analogous to earthquakes, but natural earthquakes are typically characterized through analysis of their radiated seismic waves, while laboratory earthquakes (LabEQs) are described by fault slip and changes in stress on the samples—measurements cannot typically be made for natural earthquakes. To facilitate a more direct comparison between LabEQs and natural earthquakes, we study the spectra of seismic waves radiated from LabEQs (stick-slips) generated on a 2 m long fault cut through a granite block in a large-scale biaxial apparatus. With an array of 14 piezoelectric sensors, we first record the seismic waves radiated from ball impact, which has known source characteristics, and we use those results to characterize the effects of the wave propagation and instrument and apparatus response, so that the source spectra of LabEQs can be recovered from recorded signals in a frequency band from 200 Hz to 200 kHz. The laboratory sample is also instrumented with arrays of strain gages and slip sensors, so we supplement measurements of the spectra of the LabEQs (which average over space and time) with observations of the spatial and temporal variation of slip, stress changes, and seismic radiation during nucleation and dynamic rupture. Preliminary results indicate that the spectra of small events that are self-contained within the center of the simulated fault have corner frequencies ranging from 10 kHz to over 200 kHz while the spectra of large events produced from rupture of the entire fault surface have corner frequencies close to 1 kHz and are 2 to 4 orders of magnitude larger in amplitude. At high frequencies, displacement spectral amplitudes decay close to f-2, with some systematic variation. Finally, we compare our results to earthquake source parameters derived by previous researchers based on estimates of total fault slip, duration of slip, and apparatus stiffness.
Session: The Magnitude X.X Earthquake on the YY of ZZZZ: Major Earthquakes of 2012/13
Presenter   Rogers, Garry
Schedule   Fri 9:30 AM / Oral
Room   155A
The 2012 Haida Gwaii M7.8 Earthquake: Why a Megathrust Earthquake on the Queen Charlotte “Transform Fault Zone” Plate Boundary?
HYNDMAN, R. D., Pacific Geoscience Centre, Geological Survey of Canada, Sidney, BC, Canada, rhyndman@nrcan.gc.ca; ROGERS, G. C., Pacific Geoscience Centre, Geological Survey of Canada, Sidney, BC, Canada, grogers@nrcan.gc.ca
The Queen Charlotte fault zone forms the northern boundary between the Pacific and North America plates off northern Canada and the Alaska panhandle. It has exhibited large mainly strike-slip earthquakes including the M8.1 event of 1949, and M7.5 in 2013. Seafloor acoustic imaging shows a nearly linear fault trace near the coast and a microearthquake survey showed most events beneath this trace. Thus the 2012 M7.8 thrust event appeared remarkable. However, although off the Alaska Panhandle the margin is parallel to the relative plate motion, to the south in the region of the earthquake the margin is 20 degrees oblique with 20 mm/yr convergence, and there have been a number of previous smaller thrust events, including a M6.1 that caused a small tsunami. Uplift of the west coast of the islands has been explained by underthrusting that started about 5 m.y. ago. Although there are no Wadati-Benioff earthquakes, a “subduction” model is suggested by an offshore trench and accretionary sedimentary prism, the characteristic subduction gravity anomaly, and the down-bowing of the ocean crust and sediments horizons landward. Teleseismic receiver function analyses indicate a dipping layer beneath the islands interpreted to be underthrust oceanic crust. The heat flow across the margin decreases landward, in agreement with the predictions from subduction-type thermal models. The 2012 event has been shown by initial waveform analyses, aftershocks, GPS data, tsunami amplitudes, etc. to be a shallow thrust approximately orthogonal to the margin, mainly beneath the Q.C. Terrace. We therefore interpret the relative plate motion at the margin to be partitioned between the near-vertical strike-slip Queen Charlotte fault zone and subduction-like underthrusting of the Queen Charlotte Terrace which itself represents a “forearc sliver”. The 2012 event represents Pacific - Terrace relative motion.
Session: Advances in High-Frequency Physics-Based Earthquake Systems Simulation
Presenter   Withers, Kyle
Schedule   Thu 9:30 AM / Oral
Room   155C
STUDENT
Deterministic High-Frequency Ground Motions from Simulations of Dynamic Rupture along Rough Faults
WITHERS, K. B., SDSU/UCSD, San Diego, CA, quantumkylew@aol.com; OLSEN, K. B., SDSU, San Diego, CA, kbolsen@mail.sdsu.edu; SHI, S., SDSU, San Diego, CA, zshi@projects.sdsu.edu; DAY, S. M., SDSU, San Diego, CA, sday@mail.sdsu.edu; TAKEDATSU, R., SDSU, San Diego, CA, aaaum27@hotmail.com
The accuracy of earthquake source descriptions is a major limitation in high-frequency (~>1 Hz) deterministic ground motion prediction, which is critical for performance-based design by building engineers. We address this issue by an attempt to quantify the contributions to high-frequency ground motion from both small-scale fault geometry and media complexity and perform validation against recent Next Generation Attenuation (NGA) relations. Specifically, we compute the ground motion synthetics using dynamic rupture propagation along a rough fault imbedded in a velocity structure with heterogeneities described by a statistical model. First, simulations of dynamic rupture are carried out using a support operator method (SORD, Shi and Day [2013]), in which the assumed fault roughness follows a self-similar fractal distribution with wavelength scales spanning three orders of magnitude from ~10^2 m to ~10^5 m. The rupture irregularity caused by fault roughness generates high-frequency accelerations with near-flat power spectra up to almost 10 Hz. Next, we perform wave propagation simulations using the moment-rate time histories from the dynamic rupture simulation as a kinematic source to extend the ground motions out to farther distances from the fault with a highly scalable fourth-order staggered-grid finite difference method. The latter wave propagation simulations use a characteristic 1D rock model with and without small-scale heterogeneities. We then compare our ground motion results (e.g., distance and period dependence of peak ground accelerations and peak spectral accelerations) with the empirical curves given by recent NGA relations, for both the median and the standard deviation. Finally, the high-frequency ground motions from the rough-fault simulations are compared to those obtained by a hybrid deterministic–stochastic technique (Mai et al, 2010) in order to understand the limitations of high-frequency generation using the latter method.
Session: What are the Limits of Explosion Source Model Predictions?
Presenter   Vorobiev, Oleg
Schedule   Wed 9:00 AM / Oral
Room   155D
Analysis of Near-Source Shear Motions from Underground Explosions
VOROBIEV, O. V., LLNL, Livermore, CA, vorobiev1@llnl.gov; ANTOUN, T. A., LLNL, Livermore, CA; PITARKA, A. P., LLNL, Livermore, CA; HERBOLD, E. H., LLNL, Livermore, CA; GLENN, L. G., LLNL, Livermore, CA; LOMOV, I. L., LLNL, Livermore, CA
We have performed 2D and 3D simulations of underground explosions conducted recently in the Climax Stock granitic outcrop at the Nevada National Security Site (NNSS) as part of the Source Physics Experiment (SPE) campaign. The main goal of these simulations is to understand the nature of the shear motions recorded in the near field. Numerical simulations explicitly accounted for the joints present in the rock. The joints were well characterized in the experiment and had roughly a half meter spacing. Four main joint sets were identified at the site. Two of them were closely aligned with a high dip angle. These were approximated as a single vertical joint set in the calculations. The other two joint sets were represented discretely based on the joint characterization data. The joints in the top 15 m weathered layer were assumed to be more compliant and a higher friction angle than those in the deeper layer where the rock was fully saturated. The material model used to represent the deformational response of granite was calibrated to match the properties of the rock samples collected at different depths. We have conducted a sensitivity study by varying parameters characterizing the rock mass such as joint persistency, orientation and saturation. It was shown that significant shear motions can be generated by sliding on the joints caused by spherical wave propagation. Polarity of shear motion may change during unloading when the stress state may favor joint sliding on a different joint set. Though this presentation focuses on understanding shear wave generation in the near field, the overall goal of our research is to understand the far field seismic signatures associated with shear waves generated in the immediate vicinity of an underground explosion. This topic is the subject of a companion presentation (Pitarka et. al.)
Session: Characterizing Active Faults for Seismic Hazard Assessments
Presenter   Jiang, Junle
Schedule   Fri AM / Poster
Room   Hall 1
STUDENT
Detecting Penetration of Large Earthquake Ruptures Below the Seismogenic Zone Using Microseismicity
JIANG, J., California Institute of Technology, Pasadena, CA, jjle@gps.caltech.edu; LAPUSTA, N., California Institute of Technology, Pasadena, CA, lapusta@caltech.edu
It is typically assumed that earthquake ruptures are confined within the seismogenic zone - inferred from microseismicity - with regions below stably creeping. Rate-and-state fault models that reproduce such behavior represent the locked seismogenic zone (SZ) and the deeper creeping fault extension as velocity-weakening (VW) and velocity-strengthening (VS), respectively. The models predict that the creep of the VS regions would create stress concentration at the bottom of the VW zone, leading to concentrated microseismicity there. Such characteristic microseismicity is indeed present on some faults, e.g., the Parkfield segment of the San Andreas Fault (SAF) and the Calaveras Fault. However, microseismicity at the bottom of the SZ is not observed for several faults that hosted major historical earthquakes, e.g., the Carizzo segment of the SAF. We hypothesize that such absence of concentrated microseismicity results from the most recent large earthquake rupture penetrating several kilometers below the SZ, inhibiting the creep next to the VW/VS transition, and hence removing the stress concentration there, which in turn suppresses the occurrence of microseismicity. We confirm the proposed relation between the microseismicity and the depth extent of large seismic events by simulating earthquake sequences in 3D fault models governed by rate-and-state friction laws and co-seismic thermal pressurization (TP) of pore fluids. The depth extent of large earthquakes in the model is controlled by the depth extent of efficient TP, which is determined by the depth variation of the shear zone width and permeability, and need not coincide with the rate-and-state VW/VS boundary. Our work supports recent suggestions that earthquake rupture could penetrate below the SZ (e.g., Shaw & Wesnousky, 2008; Fagereng & Toy, 2011) and proposes that segments where such penetration has occurred in recent events can be detected from observations of microseismicity.
Session: New Frontiers in Seismic Data Analysis
Presenter   Punosevac, Predrag
Schedule   Thu AM / Poster
Room   Hall 1
Theoretical Developments in Three Dimensional Seismic Wave Gradiometry
PUNOSEVAC, P., Georgia Regents University, Augusta, GA, ppunosev@gru.edu; POPPELIERS, C., Georgia Regents University, Augusta, GA, cpoppeli@gru.edu
Wave gradiometry is a method that relates the spatial gradients of a wave field to its velocity andradiation patterns. Seismic gradiometry uses the spatial gradients of the wavefield, which can beestimated from an array of well-calibrated seismic instruments. Due to the small aperture requiredto estimate the wave field's spatial gradients, it becomes tractible to construct a three dimensionalseismic array via borehole instrumentation. We report on our recent progress inexpanding the theoretical foundations of wave gradiometry into three dimensions and summarizeour recent results for both scalar and polarized waves. For scalar wave gradiometry, we developrelationships to estimate wave speed and direction, as well as a set of partial differential equationsthat relate the spatial derivatives to the seismic radiation patterns. For polarized wave gradiometry,we derive a map between the Cartesian spatial derivatives and the spherical derivatives, as well as aset of partial differential equations that relate the spatial gradients and radiation patterns for all threebody wave types.
Session: ShakeMap-Related Research, Development, Operations, and Applications
Presenter   Johanson, Ingrid
Schedule   Thu PM / Poster
Room   Hall 1
Rapid Finite Fault Determination from Real-Time GPS for ShakeMap
JOHANSON, I. A., Berkeley Seismological Lab, Berkeley, CA, ingrid@seismo.berkeley.edu; DREGER, D. S., Berkeley Seismological Lab, Berkeley, CA, dreger@seismo.berkeley.edu; GRAPENTHIN, R., Berkeley Seismological Lab, Berkeley, CA, ronni@seismo.berkeley.edu; LOMBARD, P., Berkeley Seismological Lab, Berkeley, CA, lombard@seismo.berkeley.edu
Finite fault effects have been shown to be an important part of ShakeMap; they allow improved estimates of shaking near the fault and due to rupture directivity. Static offsets from GPS provide unique constraints on fault orientation and rupture extent and can be used to supply these to ShakeMap. We have developed a strategy to use real-time GPS data access with rapid post-processing (RPP) techniques to estimate static offsets from moderate to large earthquakes and use them to constrain a non-linear search for fault plane parameters. RPP requires waiting 1-2 minutes after the earthquake for data to accumulate, but displacement time series can then be generated within 5 minutes using the software Track, developed at MIT. From these, full fault plane determination can be performed within another 5 minutes. While real-time processing techniques are critical for Earthquake Early Warning, RPP provides higher precision in the static offset measurement. This allows GPS data to be used for smaller earthquakes and still finishes within a time frame appropriate for ShakeMap. We show results using data from the 2007 M5.6 Alum Rock and 2004 M6.0 Parkfield earthquakes. We find that rapid post-processing with Track allows more robust static offset determination from these smaller earthquakes than real-time processing (as it exists today). For both earthquakes, sub-cm offsets can be resolved and match post-event analyses well. We perform a search for the best-fitting rupture plane, using no a priori information. That is, we attempt to simulate a situation where the type and location of faulting for these earthquakes is not known. While low station density makes this difficult for the M5.6 Alum Rock earthquake, the fault plane for the M6.0 Parkfield earthquake can be quickly and robustly determined. We therefore expect this strategy to be successful for earthquakes of at least M6.0 and could be used for smaller events in areas with good station coverage.
Session: Advances in High-Frequency Physics-Based Earthquake Systems Simulation
Presenter   Taborda, Ricardo
Schedule   Thu 11:30 AM / Oral
Room   155C
Comparative Validation of a Set of High-Frequency Physics-Based Simulations Using Two Different Velocity Models
TABORDA, R., Carnegie Mellon University, Pittsburgh, PA, rtaborda@cmu.edu; BIELAK, J., Carnegie Mellon University, Pittsburgh, PA
We present a comparative validation of a set of simulations of the Mw 5.4 2008 Chino Hills earthquake that use two different velocity models available for Southern California. The simulations are tailored for a maximum frequency and a minimum shear wave velocity equal to 4 Hz and 200 m/s, respectively. We validate the ground motion synthetics from each simulation with data obtained from seismic networks on more than 300 recording stations, and compare the synthetic results from the two models with each other and with observations in order to draw insights about strengths of each model’s description of the crustal structure and the geotechnical layers. The simulations are done using Hercules, the parallel octree-based finite-element earthquake simulator developed by the Quake Group at Carnegie Mellon University. The source model corresponds to that of an independent inversion study, and the seismic velocity models used are the community velocity models developed by (i) the Southern California Earthquake Center (CVM-S) and by (ii) SCEC’s partners at Harvard University (CVM-H). Our results show strong sensitivity of the ground motion to the velocities of the shallow layers, especially at the higher frequencies. The areas where each model leads to better results, however, do not always coincide, suggesting that the models need to be cross-referenced in future studies. Although our results are particular to the event considered here, future similar simulations may help guide the improvement of both the seismic velocity and the source models. This type of simulations also suggests that extending the maximum frequency of deterministic earthquake simulations is an effort worth pursuing.
Session: Triggering of Seismic and Volcanic Events
Presenter   Sammis, Charles
Schedule   Fri 11:30 AM / Oral
Room   155D
Triggered Tremor, Phase-Locking, and the Global Clustering of Great Earthquakes
SAMMIS, C. G., University of Southen California, Los Angeles, CA, sammis@usc.edu; SMITH, S. W., University of Washington, Seattle, WA, swsmith@nwlink.com
A system of non-linear coupled relaxation oscillators is used to show how the communication between large earthquakes on a global scale can align their seismic cycles to produce a worldwide clustering of large events. Our model builds on recent observations that the seismic waves from a large earthquake can trigger an episode of non-volcanic tremor at the base of a distant fault. We assume that tremor is indicative of creep on the ductile extension of the fault zone that loads its overlying seismogenic layer thus advancing the fault toward failure. If this advance is larger toward the end of the seismic cycle, we show that two or more interacting faults will align their cycles, even if their recurrence intervals are not identical.
Session: Intermountain West Earthquake and Volcano Characterization and Related Hazards
Presenter   Stickney, Michael
Schedule   Wed 11:45 AM / Oral
Room   155B
The 2005 Dillon, Montana Earthquake; a Moderate-Sized Intermountain Seismic Belt Earthquake on a Blind Normal Fault and its Effects on Local Seismicity Rates
STICKNEY, M. C., Montana Bureau of Mines and Geology, Butte, MT, mstickney@mtech.edu
On July 26, 2005, a M 5.6 earthquake occurred 15 km north of Dillon, Montana near the northern Intermountain Seismic Belt’s western margin. This earthquake resulted from slip along a north-trending, 50° east-dipping normal fault at a depth of 11 km. No known late Quaternary faults lie near the seismically inferred fault plane’s up-dip projection. A vigorous aftershock sequence immediately followed the main shock and included 29 M≥3.0 events during the first 10 days, the largest of which–a M 4.4–occurred 36 hours after the main shock. Aftershocks during the next 120 days proceeded as a typical sequence with a b-value of -0.94 and a productivity (a-value) of -1.0. The aftershock epicentral distribution eventually expanded to approximately double the initial 24-hour, 3 by 1.5 km area. During the first two years after the main shock, aftershock rates decreased with time, but between two and five years after the main shock, rates remained virtually constant. Four additional seismic sequences occurred in the surrounding region, 5-38 km from the main shock, where predicted static coulomb stress changes induced by the main shock are <1 bar. Two earthquake sequences east of the main shock–a distinct cluster of events on the east edge of the aftershock zone and a M 4.7 earthquake in the previously aseismic Ruby Valley–at distances of 5 and 38 km respectively, occurred 0.48 and 1.78 years later, suggesting an outward initiation rate of 21 km/yr in an easterly direction. Two earthquake swarms, located 16 and 24 km WSW of the main shock, occurred 1.23 and 2.06 years later, suggesting that seismicity initiated outward at a rate of 11 km/yr in a westerly direction. The association of seismic sequences at distances of 16-38 km from the Dillon main shock is speculative, but, if true, suggests an as yet unrecognized mechanism for triggering seismicity in extensional tectonic environments at distances up to 10 times the main shock fault dimensions and as much as two years later.
Session: Triggering of Seismic and Volcanic Events
Presenter   Meng, Xiaofeng
Schedule   Fri PM / Poster
Room   Hall 1
STUDENT
Hurricane Irene’s Impacts on the Aftershock Sequence of the 2011 Mw5.8 Virginia Earthquake
MENG, X., Georgia Institute of Technology, Atlanta, GA, xmeng7@gatech.edu; ALLMAN, S., Randolph College, Lynchburg, VA, sjallman@randolphcollege.edu; PENG, Z., Georgia Institute of Technology, Atlanta, GA, zpeng@gatech.edu; GILSTRAP, T., Randolph College, Lynchburg, VA, tgilstrap@randolphcollege.edu
Recent studies have shown that typhoon could trigger shallow slow-slip events in Taiwan. However, it is unclear whether extreme weather events could affect the occurrence of regular earthquakes as well. A good opportunity to test this hypothesis occurred in 2011 when an Mw 5.8 earthquake struck Louisa County, Virginia. Roughly 5 days later, hurricane Irene struck the coast of Norfolk, Virginia near the aftershock region. Because aftershocks listed in the ANSS catalog could be incomplete immediately after the mainshock and during hurricane Irene, they did not provide any important clues about possible changes in the seismicity rate associated with hurricane Irene. Hence, we use a recently developed waveform-based matched filter technique to scan through the continuous seismic data to detect aftershocks that could be buried by the coda of the mainshock/large aftershocks and hurricane signals. A mixture of 7 permanent stations from the US National Seismic Network, Virginia Tech Seismic Network and 8 temporary stations were used. The temporary stations were set up between 24-72 hours following the mainshock around its epicenter. We use the 28 aftershocks listed in the ANSS catalog as template events and scan through the continuous data from 23 August 2011 through 10 September 2011. So far we have detected 791 events using a threshold of 12 times the median absolute deviation (MAD), which is about ~25 times more than listed in the ANSS catalog. The aftershock rate decayed with time following the modified Omori’s law. A statistically significant increase of seismicity rate is found when hurricane Irene passed nearby, which could be explained by a simple vertical unloading for thrust faults. We are currently in the process of evaluating the quality of the detected events, especially during the hurricane signals. Our observation, if true, would suggest that extreme weather events such as hurricanes could trigger not only slow-slip events, but also regular earthquakes.
Session: New Frontiers in Seismic Data Analysis
Presenter   Poppeliers, Christian
Schedule   Wed 4:45 PM / Oral
Room   155A
Practical Aspects of Two- and Three-Dimensional Seismic Gradiometry and its Applications
POPPELIERS, C., Georgia Regents University, Augusta, GA, cpoppeli@gru.edu; PUNOSEVAC, P., Georgia Regents University, Augusta, GA, ppunosev@gru.edu
Wave gradiometry is a method that relates the spatial gradients of a wave field to its velocity and radiation patterns. Seismic gradiometry uses the spatial gradients of the wavefield, which can be estimated from an array of well-calibrated seismic instruments. Due to the small aperture required to estimate the wave field's spatial gradients, it becomes tractible to construct a three dimensional seismic array via borehole instrumentation. Potential applications of three dimensional wave gradiometry are seismic source studies, seismic source detection, applications to hydraulic fracturing as used in the petroleum industry, and the observation of rotational body waves. In this talk, we outline our program of research in the development of three dimensional wave gradiometry. We focus primarily on the practical aspects such as correlated and uncorrelated noise in the data, experimental uncertainty, and practical limitations of the methods we are developing.
Session: Triggering of Seismic and Volcanic Events
Presenter   Meng, Xiaofeng
Schedule   Fri PM / Poster
Room   Hall 1
STUDENT
Systematic Search of Missing Earthquakes Near the Salton Sea Geothermal Field and San Jacinto Fault Around the 2010 Mw7.2 El Mayor-Cucapah Earthquake
MENG, X., Georgia Institute of Technology, Atlanta, GA, xmeng7@gatech.edu; PENG, Z., Georgia Institute of Technology, Atlanta, GA, zpeng@gatech.edu; WITHERS, K., San Diego State University, San Diego, CA, quantumkylew@aol.com; OLSEN, K., San Diego State University, San Diego, CA, kbolsen@mail.sdsu.edu; YU, X., Georgia Institute of Technology, Atlanta, GA, xyu40@gatech.edu; HONG, B., Georgia Institute of Technology, Atlanta, GA, bohong@gatech.edu
Whether static or dynamic stress changes are the dominant mechanism for triggering earthquakes in the near field is currently unclear. Earthquake triggering usually depend on seismicity rate changes from earthquake catalogs. However, such catalogs are often incomplete immediately after the mainshock, which may cause apparent seismicity rate changes unrelated to stress changes. The 2010 Mw7.2 El Mayor-Cucapah earthquake occurred in northern Baja California, and generated significant stress perturbations and seismicity rate changes in southern California. Here we focus on the Salton Sea geothermal field (SSGF) and San Jacinto fault (SJF) mainly because of the significant stress changes following the mainshock, dense borehole seismic network coverage and abundant background seismicity. To get a more complete catalog, we apply a waveform-based matched filter technique to systematically detect possibly missing events near SSGF and SJF from 03/01/2010 until 08/01/2010. We run the detection codes on multiple GPU clusters, significantly reducing the computation time. As a result, we detect ~20 times more events than listed in the Southern California Seismic Network (SCSN) catalog. Based on our detected events, the seismicity rate near SSGF and SJF both experienced significant increase immediately following the mainshock. However, the seismicity rate near SSGF, where static Coulomb Failure Stress (sCFS) decreased, dropped below the pre-mainshock level after ~50 days. On the other hand, the seismicity rate near SJF, where sCFS increased, remained high till the end of our detecting time window. Such pattern is consistent with dynamic stress changes being dominant in the short period, while sCFS changes may be more important in the longer term. Our next step is to apply the waveform detection technique to the entire southern California around the origin time of the El Mayor-Cucapah earthquake, and compare the seismicity rate changes with the peak dynamic and sCFS changes.
Session: Characterizing Active Faults for Seismic Hazard Assessments
Presenter   Cochran, Elizabeth
Schedule   Thu 8:45 AM / Oral
Room   155B
Shallow Crustal Anisotropy in the Yuha Desert, California from Aftershocks of the 2010 M7.2 El Mayor-Cucapah Earthquake
COCHRAN, E. S., U. S. Geological Survey, Pasadena, CA, ecochran@usgs.gov; KROLL, K. A., University of California, Riverside, CA, kkrol001@ucr.edu
Following the 4 April 2010 M7.2 El Mayor-Cucapah earthquake, eight temporary seismic stations were deployed in the Yuha Desert region, California to record aftershocks. Relocations of over 4,300 aftershocks (Kroll et al., 2012), coupled with geologic observations (Rymer et al., 2011) in the Yuha Desert region, show that these events occur on sets of geometrically complex structures consisting of a series of right- and left-lateral conjugate faults, and possibly, also on several steeply dipping normal faults. We employ this unique aftershock dataset to estimate the shallow crustal anisotropy (fast direction and delay time) using the automated shear-wave splitting analysis program MFAST outlined in Savage et al. (2010). Our preliminary analyses indicate that the station-averaged fast directions, between N30W and N18E, are relatively uniform across the region. Delay times, however, are more scattered, ranging between 0.103 s to 0.162 s. Additional analyses include a more comprehensive exploration of the spatial and temporal variation in the measured fast directions. Using the shear wave splitting tomography method developed by Johnson et al. (2011) we explore the detailed spatial variation in the fast directions and delay times. These results suggest there is a slight rotation from NNW to NS from north to south in the study area. We also see suggestions of perturbations in the local anisotropic field near some active faults. In addition, there appears to be temporal variation in the fast directions measured at some of the stations. However, there is no consistent temporal rotation in the measured fast directions that might indicate a post-seismic response; instead, we suggest the apparent temporal changes may reflect spatial variations due to migration of seismicity during the study period. Delay times are quite variable and do not appear to match any obvious geologic features.
Session: Data Products as Research Resources
Presenter   Convers, Jaime
Schedule   Thu 1:30 PM / Oral
Room   155A
STUDENT
Radiated Earthquake Energy and Rupture Duration as an IRIS Data Product
CONVERS, J. A., Georgia Institute of Technology, Atlanta, GA, jconvers@gatech.edu; NEWMAN, A. V., Georgia Institute of Technology, Atlanta, GA, anewman@gatech.edu
A single magnitude determination of an earthquake is a limited characterization of an event’s size and impact. While the near ubiquitous moment magnitude describes the static work performed during an earthquake, it contains no direct information about important dynamic rupture processes. As a compliment, radiated seismic energy is a valuable additional parameter that yields direct information about the event’s rupture speed and duration, parameters useful for identifying local strong shaking and tsunami potential. We use two non-saturating seismic energy parameters (magnitude and duration), and their scaling relationship, to characterize anomalous ruptures. Events that rupture more slowly include ‘tsunami earthquakes’, that slip through poorly consolidated near-trench section on or near the subduction interface. Such events are uniquely identifiable in real-time by their discrepancy in high-frequency radiated energy and excessive rupture duration, allowing for timely indicators of a likely tsunami. Earthquakes with high dynamic stress-drop radiate excessive high frequency energy that can be comparable to earthquakes an order of magnitude larger, creating considerably greater local strong ground shaking and larger local damage than expected given the event’s seismic moment alone.Here we report on the implemented tools for rapid determination of radiated seismic energy and rupture durations from globally observed earthquakes with magnitude ≥ 6.5, including small and strong earthquakes in New Zealand and Greece, slow tsunami earthquakes including the 2012 El Salvador earthquake, and giant total megathrust events including 2011 Tohoku-Oki. All event data, including individual-station energy time series, per-station source energy corrections, combined energy and duration solutions with error, and illustrations, are being developed as an IRIS data product and should be accessible at the time of this presentation.
Session: Oceanographic and Atmospheric Signals in Seismology
Presenter   Euler, Garrett
Schedule   Fri AM / Poster
Room   Hall 1
Global 26 s Noise Peak Produced by Primary Microseisms Generated in the Bight of Bonny, West Africa
EULER, G. G., Washington University in St Louis, Saint Louis, MO, ggeuler@seismo.wustl.edu; ADAMS, A. N., Washington University in St Louis, St Louis, MO, aadams@seismo.wustl.edu; WIENS, D. A., Washington University in St Louis, St Louis, MO, doug@wustl.edu; NYBLADE, A. A., Penn State University, State College, PA, andy@psu.edu
For more than 50 years seismologists have observed unusually intense global microseisms at a period of about 26s originating from near the Gulf of Guinea. Using observations from several temporary broadband seismometer arrays in Africa including one adjacent to the Gulf of Guinea in Cameroon, we present evidence that these microseisms are generated by ocean waves of the same period in the northern portion of the Gulf of Guinea, known as the Bight of Bonny, and thus should be classified as primary microseisms. In particular, we use spectrograms of individual seismic stations near the coast to substantiate our primary microseism interpretation and demonstrate that temporal peaks in microseism generation correspond to arrivals of ocean waves of the same period from storms in the South Atlantic. A new high-resolution frequency-slowness back-projection technique is used to demonstrate that the source location is near the Nigerian and Cameroon coast adjacent to the island of Bioko. Additionally, we observe heightened microseism generation at this location for several other periods: 28s, 21s, 19s, 16s & 14s. We attempt to establish a relationship between the significant wave heights of ocean waves in the Bight of Bonny and the power of microseisms from the same location but this effort is hampered by the lack of an ocean buoy in this region. These observations cast new light on primary microseisms which are thought to be generated by the pressure field of ocean waves near the shore. We hypothesize that the unusual coastal geometry and bathymetry in the Bight of Bonny focuses long period ocean waves arriving from storms in the South Atlantic and thereby enhances primary microseism generation.
Session: Characterizing Active Faults for Seismic Hazard Assessments
Presenter   Fitzenz, Delphine
Schedule   Fri AM / Poster
Room   Hall 1
From Hypothesis Testing to Bayesian Model Comparison: Rigorous Tools to Characterize Earthquake Recurrence
FITZENZ, D. D., RMS, Newark, CA, delphine.fitzenz@rms.com; JALOBEANU, A., UT Austin / ARL NPS / RSC, Monterey, CA, ajalobea@nps.edu
We review the definitions and limitations of both Null-Hypothesis Testing and Bayes Model Comparison (through Bayes factor), in the context of poorly understood physical phenomena, such as earthquake occurrence, in which unknown correlations between variables are likely to exist. Null-hypothesis testing can only reject or fail to reject a hypothesis, not select a hypothesis or reject alternative ones. Bayesian model comparison works only for 2 models or more and does not provide a definite answer about the truth of a given model, but their relative performance at explaining the data.Our work focuses on those cases where a hypothesis H0 can not be rejected on the bases of a given dataset, regardless of its number of points N. We show that an hypothesis H1 with more parameters than H0 might not need more data than N to be proven more performant, in the Bayesian sense, in particular when correlations between variables were overlooked in the original test of H0. We explore the implications of this conclusion in terms of the desirable complexity of earthquake occurrence models, either for faults considered independent, or for couples or groups of faults.
Session: New Developments in Earthquake Forecasting and Predictability Research
Presenter   Page, Morgan
Schedule   Thu 11:15 AM / Oral
Room   155D
The UCERF3 Inversion Methodology as a Tool for Hypothesis Testing
PAGE, M. T., U.S. Geological Survey, Pasadena, CA, pagem@caltech.edu; FIELD, E. H., U.S. Geological Survey, Golden, CO, field@usgs.gov; MILNER, K. R., University of Southern California, Los Angeles, CA, kmilner@usc.edu
Rather than the prescriptive methods used in the past to produce probabilistic seismic hazard maps, we have developed an inverse approach for the 3rd Uniform California Earthquake Rupture Forecast (UCERF3) that derives a range of rupture rates consistent with the data. Due to the global nature of the applied constraints, all aspects of the model are linked; hence, changing data for one particular fault can affect the solution elsewhere in the fault system. This can be problematic in that errors in the inputs can propagate spatially; however, it also provides a new tool to investigate the degree to which different data sets and hypotheses are consistent. For example, the inversion can be used to probe which logic tree branches are inconsistent with the hypothesis that earthquake magnitudes on individual faults follow a Gutenberg-Richter distribution. The connectivity of the fault system can also be investigated since an insufficient quantity of multi-fault ruptures can preclude matching the regional magnitude distribution. In addition, there are trade-offs between how well slip rates and paleoseismic data can be simultaneously fit. Finally, the inversion methodology used in UCERF3 allows us to easily test the influence or consistency of new model inputs and to find multiple solutions with similar data fits.
Session: Advances in High-Frequency Physics-Based Earthquake Systems Simulation
Presenter   Bydlon, Samuel
Schedule   Thu 9:15 AM / Oral
Room   155C
STUDENT
Rupture Dynamics and Ground Motion from Earthquakes in Heterogeneous Media
BYDLON, S. A., Stanford University, Stanford, CA, sbydlon@stanford.edu; KOZDON, J. E., Naval Postgraduate School, Monterrey, CA, jekozdon@nps.edu; DUNHAM, E. M., Stanford University, Stanford, CA, edunham@stanford.edu
Heterogeneities in the material properties of Earth’s crust scatter propagating seismic waves. The effects of scattered waves are reflected in the seismic coda and depend on the relative strength of the heterogeneities, spatial arrangement, and distance from source to receiver. In the vicinity of the fault, scattered waves influence the rupture process by introducing fluctuations in the stresses driving propagating ruptures. Further variability in the rupture process is introduced by naturally occurring geometric complexity of fault surfaces, and the stress changes that accompany slip on rough surfaces. Early results indicate that geometric complexity is the primary contributor to high frequency fluctuations in slip and slip velocity. We have begun a modeling effort to better understand the origin of complexity in the earthquake source process, and to quantify the relative importance of source complexity and scattering along the propagation path in causing incoherence of high frequency ground motion. To do this we extended our 2D high order finite difference rupture dynamics code to accommodate material heterogeneities. We generate synthetic heterogeneous media using Von Karman correlation functions and their associated power spectral density functions. We then nucleate ruptures on either flat or rough faults, which obey strongly rate-weakening friction laws. Preliminary results for flat faults indicate that off-fault material heterogeneity alone can lead to a mildly complex rupture process. Our simulations reveal the excitation of high frequency bursts of waves, which radiate energy away from the propagating rupture. The average rupture velocity is thus reduced relative to its value in simulations employing homogeneous material properties. By changing the statistical properties of the heterogeneous medium, we observe changes in ground motion and wave front shape. Increasing the level of heterogeneity results in amplified particle velocities and accelerations.
Session: Towards an Integrated Understanding of Slow Earthquakes: What We Know, What We Don’t Know, and How to Move Forward
Presenter   Sun, Wei-Fang
Schedule   Fri AM / Poster
Room   Hall 1
STUDENT
Using Seismic Arrays to Detect Non-volcanic Tremor in Taiwan
SUN, W., Georgia Institute of Technology, Atlanta, GA, ttsun.sun@gatech.edu; PENG, Z., Georgia Institute of Technology, Atlanta, GA, zpeng@gatech.edu; LIN, C., Institute of Earth Sciences, Academia Sinica, Taipei, Taiwan, lin@earth.sinica.edu.tw; CHAO, K., Earthquake Research Institute, University of Tokyo, Tokyo, Japan, kevinchao@gmail.com
Deep non-volcanic tremors (NVT) triggered by teleseismic surface waves have been recently observed in the southern Central Range in Taiwan, an arc-continental type collision environment. Aiming to capture more NVT events, we installed two dense 36-element, small-aperture seismic arrays in 2011 around the Liouguei and Lidao areas, which are located about 20 km in southwest and northeast to the tremor sources. In each array, the short-period, vertical-channel GS-11D sensors with 4.5Hz natural frequency seismometers were laid out on the relatively flat parts of the mountain areas with a spacing of approximately 100 by 80 meters. We had successfully recorded continuously for a total of 4,034 hours in 2011. As expected, the two arrays recorded tremor bursts triggered by the great Tohoku earthquake (Mw=9.0) on March 11, 2011. We apply the broadband frequency wavenumber (BBFK) beamforming method to measure the back-azimuth and incident angles for each tremor burst and both arrays. Our initial results show that obtained array parameters closely match those predicted from locations using tremor envelope cross-correlations. Our next step is to improve our array analysis technique, and use it to detect and locate ambient tremor occurred during our deployment. Our systematic analysis of deep tremor in Taiwan could help to better understand critical conditions related to tremor occurrence and fault mechanics at the bottom of the seismogenic layer.
Session: Infrasound and Seismoacoustics
Presenter   Park, Junghyun
Schedule   Thu AM / Poster
Room   Hall 1
STUDENT
A Comparative Study of Automated and Analyst Infrasound Detections Using PMCC and InfraMonitor
PARK, J., Southern Methodist University, Dallas, TX, junghyunp@smu.edu; HAYWARD, C. T., Southern Methodist University, Dallas, TX, hayward@smu.edu; ZEILER, C. P., Air Force Technical Application Center/TTRE, Patrick AFB, FL, cleat.zeiler@us.af.mil; ARROWSMITH, S. J., Los Alamos National Laboratory, Los Alamos, NM, arrows@lanl.gov; STUMP, B. W., Southern Methodist University, Dallas, TX, bstump@smu.edu
The automated detections calculated by PMCC (Cansi, 1995) and InfraMonitor (Arrowsmith et al., 2008) were compared to the signals identified by five independent analysts. PMCC uses a progressive cross-correlation technique to identify signals across an array; while InfraMonitor uses the modified F-statistic on a beam of the array. Each detector was applied to signals taken from a four-hour time sequence recorded by the Korean infrasound array CHNAR. The CHNAR array was used because of the small (<100 m) and large (~1000 m) aperture spacing of the elements. The four-hour time sequence contained a number of easily identified signals under a variety of noise conditions. The noise level varied in average RMS amplitude from 0.022 Pa to 0.058 Pa in the frequency band of 1 to 5 Hz, as estimated using progressive five-minute windows. The effectiveness, under the varied noise conditions, of both automatic detectors was determined for the small aperture, large aperture, small aperture combined with the large aperture, and the full array. The full and combined arrays performed the best for InfraMonitor for all noise conditions and the large aperture array had the poorest performance for both detectors. PMCC produced similar results to InfraMonitor during low noise conditions, but did not see dramatic improvement when using the full and combined arrays. The performance of the automated detectors and analysts showed a decrease in the rate of detection for high noise conditions. We compared the detection probabilities using Receiver Operating Characteristic (ROC) curves and found that the initial settings of consistency (PMCC, 0.1 and 0.5 s) and p-value (InfraMonitor, 0.01 and 0.05) did not impact the detection probability significantly. The detection probability was impacted the most by the noise level, with low noise having an average detection probability of ~30% and high noise average detection probability of ~10%.
Session: Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Models
Presenter   Miner, Krystle
Schedule   Fri 9:15 AM / Oral
Room   155C
STUDENT
Using the Charleston, SC Earthquake of 1886 to Develop New Models for Estimating Future Earthquake Damage
MINER, K. S., Department of Geology, College of Charleston, Charleston, SC, ksminer@g.cofc.edu; JAUMÉ, S. C., Department of Geology, College of Charleston, Charleston, SC, jaumes@cofc.edu; LEVINE, N. S., Department of Geology, College of Charleston, Charleston, SC, levinen@cofc.edu
In 1886, a magnitude 6.8 earthquake occurred near Charleston, SC, producing extensive damage in the city and surrounding region. Because earthquake damage on the Charleston peninsula is so well documented, there exists an opportunity to use this information to develop better models for predicting damage in future earthquakes. We are combining existing geological and geotechnical data and new seismological data to illuminate which factors best predict building damage due to earthquake shaking. Both GIS and HAZUS are used to test the usefulness of this data to predict observed building damage in 1886. We collected ambient seismic noise (microtremor) at 23 sites within a small section of downtown Charleston. We processed this data to produce horizontal-to-vertical spectral ratios (HVSR) to estimate the frequency and relative amplitude of seismic amplification. We find that site response within the study area shows the same correlation between surface geology and the 1 to 2 Hz peak amplitude that we see in previous work in Charleston; i.e., sites on artificial fill have higher peak amplitudes than sites on sturdier materials. Within GIS we created a continuous raster surface depicting site condition in the study area based upon these peak amplitudes. Using HAZUS software we created a catalog of damage predictions for a range of site conditions and event magnitudes. Damage estimates change significantly based upon the initial data map inputs. HAZUS analysis will be completed for a 1886 building inventory in order to establish which set of input parameters best matches the actual damage from the 1886 earthquake with the implication that the same combination of inputs will produce the most accurate future damage estimates for the current building inventory.
Session: Characterizing Active Faults for Seismic Hazard Assessments
Presenter   Dong, Shaopeng
Schedule   Fri AM / Poster
Room   Hall 1
STUDENT
Strike-Slip along Wassuk Range: A Piece of the Puzzle in Understanding Accommodation of Crustal Shear in the Northern Walker Lane
DONG, S. P., Institute of Geology, China Earthquake Administration, Beijing, China, dshaopeng@gmail.com; WESNOUSKY, S. G., Center for Neotectonics Studies, UNR, Reno, NV, wesnousky@unr.edu
The Wassuk Range fault is the easternmost of a set of normal fault-bounded ranges in the northern Walker Lane. Here geodesy shows that right-lateral shear is accumulating across the area at about 6 mm/yr (Wesnousky, Bormann, Kreemer, and Hammond, 2012). Reported slip rates across the ranges generally reflect normal fault displacements and have been too small and incorrectly oriented to explain the geodetic signal. The largest discrepancy is observed at the Wassuk Range where geodesy predicts right-lateral shear of ~0.7-1.6 mm/yr along the range front. Here we report evidence of ~13.9 m of right-lateral slip on a fault outboard and approximately parallel to the range front. The slip is recorded by displacement of several recessional beach ridges of pluvial Lake Lahontan. The offset is measured by displaced shadows on rectified low-sun-angle photography and differential backpack GPS surveys. The Lake Lahontan highstand has previously been reported to have occurred at about 15,500 cal yr BP(Briggs and Wesnousky, 2004). Dividing the offset by the age of the highstand is equivalent to a slip rate of ~>0.9 mm/yr, assuming the offset represents multiple repeated events and the Lahontan highstand is a maximum bound on the age of the displaced shorelines. The value accounts for a significant portion of the slip suggested by geodetic studies. Understanding how geodetic rates of shear accumulation compare to those recorded geologically is important to the analysis of seismic hazard. This work adds a piece to the puzzle in understanding the discrepancy between geodetic and geologic rates of slip and the geodynamics of the Walker Lane.
Session: Characterizing Active Faults for Seismic Hazard Assessments
Presenter   Allen, Trevor
Schedule   Fri AM / Poster
Room   Hall 1
Co-seismic Surface Deformation Relating to the March 23, 2012 MW 5.4 Ernabella (Pukatja) Earthquake, Central Australia: Implications for Cratonic Fault Scaling Relations
CLARK, D., Geoscience Australia, Canberra, ACT, Australia, dan.clark@ga.gov.au; MCPHERSON, A., Geoscience Australia, Canberra, ACT, Australia, andrew.mcpherson@ga.gov.au; ALLEN, T., Geoscience Australia, Canberra, ACT, Australia, treviallen@gmail.com; DE KOOL, M., Geoscience Australia, Canberra, ACT, Australia, marthijn.dekool@ga.gov.au
On 23 March 2012, at 09:25 GMT, a MW 5.4 earthquake occurred in the eastern Musgrave Ranges region of north-central South Australia, near the community of Ernabella (Pukatja). Several small communities in this remote part of central Australia reported the tremor, but there were no reports of injury or significant damage. This was the largest earthquake to be recorded on mainland Australia for the past 15 years and resulted in the formation of a 1.6 km-long surface deformation zone comprising reverse fault scarps with a maximum vertical displacement of over 0.5 m, extensive ground cracking, and numerous rock falls. The earthquake occurred in non-extended Stable Continental Region (SCR) cratonic crust, over 1900 km from the nearest plate boundary. The record of surface deformation relating to the Ernabella earthquake therefore provides an important constraint on models relating surface rupture length to earthquake magnitude. Such models may be employed to better interpret Australia’s rich prehistoric record of seismicity, and contribute to improved estimates of SCR seismic hazard worldwide. Based upon an analysis of new and existing surface rupture length data, faults in non-extended stable cratonic Australia appear to produce longer surface ruptures (for earthquakes larger than MW ~5.7) than those estimated using existing moment-to-rupture length scaling relations. The implication is that values of maximum magnitude earthquakes in such settings, as determined by analysis of surface rupture length data from paleo-fault scarps, may be overestimated.
Session: Intermountain West Earthquake and Volcano Characterization and Related Hazards
Presenter   Smith, Robert
Schedule   Wed 9:00 AM / Oral
Room   155B
Yellowstone Hotspot: Science and Hazards
SMITH, R. B., University of Utah, Salt Lake City, UT, robert.b.smith@utah.edu; FARRELL, J., University of Utah, Salt Lake City, UT, jamie.farrell@utah.edu; PUSKAS, C., UNAVCO, Boulder, CO, cmpuskas@gmail.com; MASSIN, F., Guadeloupe Observatory, Guadeloupe, French West Indies; CHANG, W. L., National Central University, Jhongli, Taiwan
The Yellowstone-Teton region is an active volcano-tectonic system has experienced the highest historic seismicity of the Intermountain West including the deadly Mw7.3 1959 Hebgen Lake, MT earthquake in an intraplate extensional regime with a Quat. history of caldera-forming silicic-basaltic volcanism. The Yellowstone hotspot results from interaction of a mantle plume producing a 400-km high topographic swell. Moreover a plume-fueled crustal magma reservoir 10 km beneath the Yellowstone caldera extends laterally~20 km NE of the caldera and is ~30% larger than mapped. Yellowstone seismicity is dominated by earthquake swarms and precisely located hyoocenters highlight an E-W zone of earthquakes that extends from the 1959 Hebgen Lake earthquake along the north side the Yellowstone caldera, then become less frequent, shallow beneath the caldera and occur along N-NW trending zones parallel to the post-caldera volcanic vents. Kinematically, GPS-derived deformation of Yellowstone is dominated by regional crustal extension at up to ~0.4 cm/yr but experienced a 2004-2009 accelerated uplift episode up to 7 cm/yr whose source is modeled by magmatic recharge of a sill at the top of the crustal magma reservoir. In contrast, the Teton fault is seismically quiescent with diffuse seismicity in the Gros Ventre Range. Paradoxically, Teton area reveals westward extension of the Jackson Hole valley compressing against the Teton fault as well as experiencing uplift, opposite to that envisioned for active normal fault loading. The earthquake, hydrothermal, and the volcano history are used to as input into a probabilistic seismic and volcano hazards analysis that reveals this region as a region of unusually high geologic hazard. In recognition of two highly visible national parks in an active volcano-tectonic setting with a ~7 million summer visitor population, we propose designating the Yellowstone-Teton system “As A Unique Volcano/Earthquake Hazard Area”.
Session: The Magnitude X.X Earthquake on the YY of ZZZZ: Major Earthquakes of 2012/13
Presenter   Bird, Alison
Schedule   Fri 11:00 AM / Oral
Room   155A
Shaking and Impacts from the October 2012 Magnitude 7.7 Earthquake near Haida Gwaii
BIRD, A. L., Geological Survey of Canada, Sidney, BC, Canada, Alison.Bird@NRCan-RNCan.gc.ca; HALCHUK, S., Geological Survey of Canada, Ottawa, ON, Canada, Stephen.Halchuck@NRCan-RNCan.gc.ca; ROSENBERGER, A., Geological Survey of Canada, Sidney, BC, Canada, Andreas.Rosenberger@NRCan-RNCan.gc.ca
Last October's magnitude 7.7 earthquake in the region of Haida Gwaii, Canada (formerly the Queen Charlotte Islands) is the second largest recorded in Canadian history. It was felt throughout British Columbia and as far away as the Yukon, Alberta and Montana, roughly 1500 km from the epicentre. In some locations (notably on Haida Gwaii), the perceivable shaking lasted 1.5 - 2 minutes, with very strong shaking for about 30 seconds. Strong motion seismometers recorded ground motions at three locations in the region, to a maximum horizontal acceleration of 0.2 g. Nevertheless, this earthquake resulted in very limited damage partly due to the population centres being located at least 80 km from the epicentre and 60 km from the fault rupture, but also due to the generally low, wood-frame construction on the islands. While relatively little visible impact and few, minor injuries resulted, many people were significantly traumatized by the experience and the numerous felt aftershocks.We will examine the various physical effects (e.g. landslides, building damage, loss of hot springs) from the shaking by this large earthquake, catalogued by NRCan field crews and by the inhabitants of Haida Gwaii. We also determine how the intensities gleaned from analysis of eye-witness accounts may have been affected by the rupture dynamics of the earthquake and the effects of surface materials as a guide toward the potential impact on the various Haida Gwaii communities from future large earthquakes. October's earthquake may also be used as a proxy for earthquakes in other, more populated areas of British Columbia, Canada and the world.
Session: The Magnitude X.X Earthquake on the YY of ZZZZ: Major Earthquakes of 2012/13
Presenter   Russell, David
Schedule   Fri PM / Poster
Room   Hall 1
Improved Ms Estimation Using Combined Rayleigh and Love Wave Energy: Application to the 17 May 2012 East Texas Event
BONNER, J. L., Weston Geophysical Corp., Lufkin, TX, jes_bonner@westongeo.com; RUSSELL, D., Weston Geophysical Corp., Satellite Beach, FL, dhrussell@westongeophysical.com
The 17 May 2012 Mw=4.83 East Texas earthquake provides an interesting test case for a new surface wave magnitude estimation procedure that combines Rayleigh- and Love-wave energies. The event was shallow (< 5 km) and had a dip-slip mechanism (Herrmann, 2012), which synthetic studies show is a scenario that will significantly increase variance in Ms estimates due to radiation pattern nulls. Rayleigh-waves from 700+ seismic stations were processed using the variable-period (T=8-25 s) Ms(VMAX) formula (Russell, 20006). The Ms Rayleigh=4.59+-0.20 with significant azimuthal variations that match the radiation patterns in the preferred focal mechanism (Herrmann, 2012). The Ms variance is increased when we use only Love waves to estimate Ms (e.g., Ms Love=4.74+-0.25) due to twice as many lobes/nulls in the radiation pattern. We propose a new surface wave magnitude calculated from the average energy contribution from both Rayleigh and Love waves at each station. Rayleigh and Love wave amplitudes at each station are corrected for spreading, attenuation, and source effects and are then averaged prior to forming the magnitude. The proposed method reduces the impact of nulls on Ms estimation by taking advantage of the orthogonality of Love and Rayleigh wave radiation patterns. The variance is reduced (Ms RL = 4.68+-0.15) and the resulting magnitude is a better indicator of the source size than a Rayleigh- or Love-wave only solution. There are advantages to using this method for both sparse and large network distributions. We have also investigated extending the period range to 40 secs for improved magnitude estimation of larger events, including the 2011 Virginia event.
Session: Intermountain West Earthquake and Volcano Characterization and Related Hazards
Presenter   Lund, William
Schedule   Wed 2:30 PM / Oral
Room   155B
Paleoseismic Trenching Investigation of the Northern Section of the Washington Fault Zone at the Dutchman Draw Site, Northwestern Arizona
LUND, W. R., Utah Geological Survey, Cedar City, UT, billlund@utah.gov; KNUDSEN, T. R., Utah Geological Survey, Cedar City, UT, tylerknudsen@utah.gov; DUROSS, C. B., Utah Geological Survey, Salt Lake City, UT, christopherduross@utah.gov; MCDONALD, G. N., Utah Geological Survey, Salt Lake City, UT, gregmcdonald@utah.gov
A Utah Geological Survey paleoseismic trenching investigation on the Northern section of the Washington fault zone provides evidence for two surface-faulting earthquakes at the Dutchman Draw site in Arizona. OxCal analysis of 14C and OSL ages constrains the earthquakes to the Holocene, at 7.7 ± 2.3 ka (P2) and 1.0 ± 0.6 ka (P1) (rounded to nearest 100 years and 2-σ uncertainty). The closed-seismic-cycle recurrence interval (also modeled using OxCal) for the two earthquakes is 6.6 ± 2.4 kyr. The trenches revealed indirect stratigraphic evidence for a possible older earthquake (P3) between 13.8 ± 1.2 and 17.1 ± 1.4 ka. The length of a possible P3–P2 recurrence interval (2.5 to 13.2 kyr) has a median value of about 7.9 kyr. The P2–P1 recurrence (6.6 ± 2.4 kyr) is within 2-σ uncertainty of the P3–P2 median value (7.9 kyr), suggesting that the P2–P1 recurrence value may be generally representative of average surface-faulting recurrence on the Northern section during latest Quaternary time.Net vertical displacement estimates came from scarp profiles, displaced stratigraphy in trenches, and scarp free-face heights extrapolated from colluvial-wedge thicknesses. Displacements ranged from about 1.0 m (P1) to 2.4 m (P2). The vertical slip rate for the P2–P1 recurrence interval (6.6 ± 2.4 kyr) and the P1 net vertical displacement (1.0–1.2 m) is 0.11-0.29 mm/yr. This slip-rate represents only the most recent closed seismic cycle, and should be treated with caution if used to extrapolate the long-term behavior of the Northern section.Using multiple regression relations to determine paleoearthquake magnitudes resulted in magnitudes from M 6.7 to 7.1, depending on the estimated surface-fault-rupture length. Available paleoseismic data are insufficient to fully characterize all possible Northern section rupture scenarios, but our analysis shows that the Northern section could produce future M > 7 earthquakes with average recurrence intervals of several thousands of years.
Session: Next Generation of Ground Motion Prediction Models
Presenter   Seyhan, Emel
Schedule   Wed 9:15 AM / Oral
Room   155C
STUDENT
Ground Motion Prediction Equations for RotD50 Horizontal Component for Active Crustal Regions
BOORE, D. M., U.S. Geological Survey, Menlo Park, CA, boore@usgs.gov; STEWART, J. P., University of California, Los Angeles, CA, jstewart@seas.ucla.edu; SEYHAN, E., University of California, Los Angeles, CA, eseyhan@ucla.edu; ATKINSON, G. M., University of Western Ontario, London, ON, Canada, gmatkinson@aol.com
We provide a relatively simple set of empirical equations for prediction of earthquake ground motions for the average horizontal component from shallow crustal earthquakes in active crustal regions. The equations were developed as part of the NGA-West2 project and are based on a combined data set that includes a large number of relatively small magnitude events from California for the approximate date range of 1998 to 2011 and international events from 1935 to 2011 spanning a wide magnitude range. The equations are formulated similarly to the Boore and Atkinson ground-motion model from the original PEER NGA project. Among the changes are modified site terms, based on the work of a site response task group in the NGA-West2 project, which include variable levels of Vs30 scaling above and below a corner velocity (Vc) along with region-specific Vs30 scaling for Vs30 < Vc; a modified magnitude scaling function, which now includes oversaturation of high frequency spectral ordinates at large magnitude at a reference distance of 1 km (although because of the magnitude-dependent apparent geometrical spreading, this does not result in oversaturation of predicted PSA); modified apparent anelastic attenuation terms based on analysis of the large small-magnitude data set from California with corrections for other regions; a magnitude dependent aftershock correction term; and optional correction factors for secondary effects such as top of rupture depth and basin depth. Coefficients for magnitude-scaling, distance-attenuation, and standard deviation terms are based on two-stage regressions using the much larger data set currently available. The resulting equations are applicable for events over a magnitude range of 3 to 8, distance range of null to 200 km, and for strike-slip, normal-slip, and reverse-slip focal mechanism for ground-motion prediction at spectral periods of 0-10 sec and PGV.
Session: Velocity Models and Modeling
Presenter   Patlan, Ezer
Schedule   Wed 11:00 AM / Oral
Room   155A
STUDENT
Microseismic, Ambient Noise Cross-correlation, and Two-Station Analysis Study of Menengai Caldera: Geothermal Prospect in the Central Kenya Dome
PATLAN, E., University of Texas, El Paso, TX, epatlan@miners.utep.edu; WAMALWA, A., Geothermal Development Company, Nairobi, Kenya, munikatony@yahoo.com; KAIP, G., University of Texas, El Paso, TX, gkaip@utep.edu; VELASCO, A. A., University of Texas, El Paso, TX, aavelasco@utep.edu
The Geothermal Development Company (GDC) and the University of Texas at El Paso (UTEP) have deployed fourteen seismic stations around the Menengai geothermal field along the Kenya rift system to monitor the seismicity around the host volcano to help identify active faults and fracture systems that may contain hydrothermal fluids and favorable drilling targets. We use waveform data from March 27 2011 to March 10 2012 to approximate seismic relocation and ambient noise tomography. For example, double difference relocation was used to locate events in order to identify faults within the margin of the caldera. Hence, we apply the ambient noise tomography approach and two-station analysis method in order to calculate phase velocities to be used for tomography. For the ambient noise, we cut the in the vertical component waveform data at a seismic station every 5 hours and preprocess the noise using 1-bit normalization to eliminate earthquake signals and instrumentation irregularities. We apply a spectral whitening to reduce the seismic amplitude. We cross-correlate the waveforms between two seismic stations to retrieve the Green’s function. For the two-station method, we measured between two stations approximately aligned with the teleseismic epicenter modern-to-large magnitude Mw > 5.7 on a great circle path, thus eliminating the common source phase. We used Frequency time analysis (FTAN) and match filter analysis (MTA) for both techniques in order to approximate the phase velocities. From the path dispersion curves, we will perform surface wave tomography that will show a subsurface shear wave velocity map, and with the micro-events, we can locate the hydrothermal activity for a future drill site.
Session: Earthquake Source Physics
Presenter   Ryan, Kenny
Schedule   Thu PM / Poster
Room   Hall 1
STUDENT
Modeling Rupture with Heterogeneous Prestress and Through Stable-Sliding Zones, and Implications for an Alaskan-Aleutian Megathrust Earthquake
RYAN, K. J., University of California, Riverside, CA, kryan003@ucr.edu; OGLESBY, D. D., University of California, Riverside, CA; GEIST, E. L., U.S. Geological Survey, Menlo Park, CA
Motivated by the 2011 M9 Tohoku-Oki event and potential earthquakes on the Alaskan-Aleutian (A-A) Megathrust, we investigate the effects of realistic fault dynamics on slip, free surface deformation, and resulting tsunami formation from an M9 megathrust earthquake. We model four scenarios: a spatially-homogenous prestress and frictional parameter model, two models with rate-strengthening-like friction (e.g., Dieterich, 1992), and one model with spatially-heterogeneous prestress. Firstly, we use the dynamic finite element code FaultMod (Barall, 2008) to show that simple slip-weakening friction (e.g., Ida, 1972) can serve as a very accurate proxy for rate-strengthening friction. Secondly, we use the dynamic finite element code EQDyna (Duan and Oglesby, 2006) to model 3-D ruptures, using time-weakening friction as a proxy for rate-strengthening friction, along a portion of the A-A subduction zone. Given geometric, material, and plate-coupling data along the A-A megathrust assembled from the Science Application for Risk Reduction (SAFRR) team (e.g., Bruns et al., 1987; Hayes et al., 2012; Johnson et al., 2004; Santini et al., 2003; Wells at al., 2003), we are able to dynamically model rupture. Adding frictional-strengthening to a region of the fault reduces both average slip and free surface displacement above the strengthening zone, with the magnitude of the reductions depending on the strengthening zone location. The heterogeneous prestress model has the same average slip as the homogeneous model, but with a heterogeneous slip distribution. Corresponding tsunami models, which use a finite difference method to solve linear long-wave equations (Shuto, 1991; Satake, 2002), match sea floor displacement, in time, to the free surface displacement from the rupture models (Tanioka and Satake, 1996). Tsunami models show changes in local peak amplitudes and beaming patterns for each slip distribution.
Session: Infrasound and Seismoacoustics
Presenter   Taylor, Steven
Schedule   Wed 4:45 PM / Oral
Room   155D
Multivariate Acoustic Detection of Small Explosions using Fisher’s Combined Probability Test
TAYLOR, S. R., Rocky Mountain Geophysics, Los Alamos, NM, srt-rmg@comcast.net; ARROWSMITH, S. J., Los Alamos National Laboratory, Los Alamos, NM, arrows@lanl.gov; ANDERSON, D. N., Los Alamos National Laboratory, Los Alamos, NM, dand@lanl.gov
A methodology for the combined acoustic detection and discrimination of explosions, which uses multiple detectors, is developed for the purpose of identifying weak explosion signals embedded in complex background noise. By utilizing physical models for simple explosions that are formulated as statistical hypothesis tests, the detection/discrimination approach does not require a model for the background noise, which can be highly complex and variable in practice. The detectors include an array F detector, a standard STA/LTA, a spectrogram detector and an N-wave detector. The N wave detector is a matched filter using a known signal model with parameters to be determined.Fisher’s Combined Probability Test is used to combine the p-values from all multivariate discriminants for the optimal detection of explosions that may be small or recorded remotely. This framework is applied to acoustic data from a 400g explosion conducted at Los Alamos National Laboratory and recorded at a distance of 5.1 km.
Session: Intermountain West Earthquake and Volcano Characterization and Related Hazards
Presenter   Lund, Bill
Schedule   Wed 11:00 AM / Oral
Room   155B
New Surficial Geologic Mapping Redefines the Northernmost Sections of the Quaternary-Active Washington Fault Zone, Washington County, Utah, and Mohave County, Arizona
KNUDSEN, T. R., Utah Geological Survey, Cedar City, UT
Detailed geologic mapping of the Washington fault zone (WAFZ) in northern Arizona and southwestern Utah has resulted in a new fault-section model for the WAFZ that includes the addition of the 22-km-long Washington Hollow section at the fault’s northern end. The 97-km-long WAFZ is one of several large normal faults in the structural transition between the Colorado Plateau and Basin and Range physiographic provinces in northern Arizona and southwestern Utah. In Utah, the WAFZ trends through the rapidly urbanizing St. George metropolitan area in Washington County. Scarps formed on unconsolidated basin-fill deposits and soft bedrock along the fault are evidence of late Quaternary surface faulting. Therefore, the WAFZ is considered active and represents a significant seismic hazard to the St. George area. Previous workers divided the WAFZ into three fault sections from south to north: Sullivan Draw, Mokaac, and Northern. New geologic mapping identified the Washington Hollow fault, which is along strike with and north of the WAFZ, as a fourth section of the WAFZ that is distinct from the Northern section to the south. Because the previously defined Northern section is no longer the northernmost section of the WAFZ, to avoid confusion I propose renaming the Northern section the Fort Pearce section. Additionally, the new geologic mapping shows that the Mokaac section and the Dutchman Draw fault, previously mapped as independent structures, are strands of the Fort Pearce section. The new mapping also identified two previously unknown fault scarps formed on late Pleistocene to Holocene alluvial deposits in Arizona. One scarp is on the main trace of the Fort Pearce section and the other is on a subsidiary fault of the Mokaac strand of the Fort Pearce section. The remaining known scarps on unconsolidated deposits along the Fort Pearce section in Utah and Arizona are mostly bedrock cored and not suitable for paleoseismic trenching.
Session: Earthquake Source Physics
Presenter   Lapusta, Nadia
Schedule   Thu 8:30 AM / Oral
Room   155A
Models of Earthquakes and Aseismic Slip Based on Laboratory-Derived Friction Laws: Accomplishments and Future Challenges
LAPUSTA, N., California Institute of Technology, Pasadena, CA, lapusta@caltech.edu
Fault processes involve both dynamic events (earthquakes) and complex patterns of quasi-staticslow slip. Understanding physics and mechanics of fault slip in its entirety is a fascinatingscientific problem. However, even for the more pragmatic goal of understanding only the behavior ofdestructive large dynamic events, it is still important to consider the entire slip history of the fault, since slow slip may determine where earthquakes would nucleate and arrest, modify stress and other initial conditions before dynamic rupture, and provide clues for the physics of the earthquake source.Fault models based on laboratory-derived rate-and-state friction formulations, including extensions to dilatancy and co-seismic weakening mechanisms, have been remarkably successful in reproducing, qualitatively and often quantitatively, the entire range of fault slip behaviors, including spontaneous earthquake nucleation, dynamic rupture, postseismic slip, repeating earthquakes, earthquake triggering, aftershock sequences, and slow slip transients. I will discuss the current rapidly evolving state-of-the-art in these models, fueled by the increasing stream of high-quality laboratory experiments, observational data, and computational resources, as well as significant future challenges. They include combining the richness of fault resistance laws with realistic fault geometries, including local non-planarity; studying interaction of multiple fault segments; and exploring scientifically justifiable ways of reducing the complexity of the resulting models. An important eventual goal is to harness the success of the laboratory-based models for predictive physics-based modeling, in which all available observations and experiments are used to inform the suitable ranges of physical and geometric parameters of a model, and then a range of potential fault behaviors is uncovered.
Session: Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Models
Presenter   Anderson, John
Schedule   Fri 11:45 AM / Oral
Room   155C
An Overview of the Largest Amplitudes in Recorded Ground Motions
ANDERSON, J. G., Nevada Seismological Laboratory, University of Nevada, Reno, NV, jga@unr.edu; KOKETSU, K., Earthquake Research Institute, University of Tokyo, Tokyo, Japan, koketsu@eri.u-tokyo.ac.jp; MIYAKE, H., Earthquake Research Institute, University of Tokyo, Tokyo, Japan, hiroe@eri.u-tokyo.ac.jp
A collection of 445 of the strongest openly-available observed ground motion accelerograms has been compiled. The extended compilation includes 71 with PGA over 1 g and 41 with PGV over 1 m/s. Response spectra have been found for all records in this extended compilation. We geneate distribution functions of the amplitudes of exceptional ground motions, and for response spectra we find spectral amplitudes of equal occurrence rate in additional to the current observational upper bounds. Perhaps it is just coincidental, but when the amplitude distributions are normalized by an estimate of the record time, they are similar to the hazard curves for a location of high seismic hazard (San Francisco based on the US National Hazard Map). Examples of exceptional records in different period ranges are shown, and the nature and cause of those motions is discussed. A strong source is an essential contribution in all cases. At long periods, exceptional ground motions are also generated by resonance in deep basins, sometimes very far from the sources. At intermediate and short periods, site resonance also frequently makes an important contribution. While non-linearity in soils generally tends to increase the site period, in rock and also at some soil sites, high frequency spikes on some of the records might also be generated by nonlinear processes near the stations.
Session: Characterizing Active Faults for Seismic Hazard Assessments
Presenter   Segou, Margarita
Schedule   Thu 4:30 PM / Oral
Room   155B
Multiple Methods for Analyzing GPS Velocity Fields Applied to the Active Tectonics of the Aegean Region
THATCHER, W., U S Geological Survey, Menlo Park, CA, thatcher@usgs.gov; SIMPSON, R., U S Geological Survey, Menlo Park, CA, simpson@usgs.gov; SEGOU, M., Géoazur, Valbonne, France, segou@geoazur.unice.fr
We apply several methods to highlight both common features of GPS velocity fields to identify both obvious and subtle features in the available data from the Aegean region of the eastern Mediterranean. We find that in searching for characteristic features of present-day interseismic velocity fields no one method is entirely satisfactory and so we use a suite of analysis techniques. Strain rate mapping using Delaunay triangulation is one crude but useful method of identifying regions with concentrated strain near faults, intra-fault straining, and regions of little or no deformation. Cluster analysis is a direct, objective method for finding localized velocity gradients, almost invariably near major faults, and determining fault slip rates. Gradient analysis applied to gridded versions of the velocity data can help locate both major and minor steps in velocity magnitude. When appropriate, geodetic block modeling may be applied using the methods described above to construct block geometry, often guided by locations of major active faults.Using these methods we clearly identify: the boundary between the North Anatolian fault-North Aegean Trough; the absence of tectonic deformation of the South Aegean block; the region of oblique strike-slip/normal extension in NW Anatolia; subduction off NW Greece (Corfu); high extensional strain rates across the Corinth Gulf, and lower but detectible extension in the Northern Gulf of Evia (Greece) and the Gulf of Datca (SW Turkey).
Session: Infrasound and Seismoacoustics
Presenter   Thelen (video), Weston
Schedule   Wed 5:00 PM / Oral
Room   155D
Enhanced Infrasound Monitoring Capabilities on the Big Island of Hawaii
THELEN, W. A., USGS Hawaiian Volcano Observatory, Hawai`i Nat'l Park, HI, wthelen@usgs.gov; GARCES, M., Infrasound Lab, HIGP, SOEST, University of Hawaii, Manoa, Kailua-Kona, HI; COOPER, J., USGS Hawaiian Volcano Observatory, Hawai`i Nat'l Park, HI; BADGER, N., Infrasound Lab, HIGP, SOEST, University of Hawaii, Manoa, Kailua-Kona, HI; PERTUU, A., Infrasound Lab, HIGP, SOEST, University of Hawaii, Manoa, Kailua-Kona, HI; WILLIAMS, B., Infrasound Lab, HIGP, SOEST, University of Hawaii, Manoa, Kailua-Kona, HI; LAHUSEN, R., USGS Cascade Volcano Observatory, Vancouver, WA
The Hawaiian Volcano Observatory (HVO) with the participation of the University of Hawaii Infrasound Lab (ISLA) installed three new permanent infrasound arrays on the south half of the Island of Hawaii. Together with three existing permanent arrays maintained by ISLA, the current infrasound network around Kīlauea volcano is one of the most advanced at any volcano in the world. Open-vent volcanoes like Kīlauea are particularly good infrasound emitters as lava spattering and turbulent gas release is common. The network was designed with two main goals in mind: 1) to monitor and study the infrasound sources associated with the ongoing Pu`u `Ō`ō eruption, and 2) to detect in near-real time new fissure eruptions at Mauna Loa or Kīlauea volcanoes. Each HVO array consists of 4 sensors, which form an equilateral triangle 100 m on a side surrounding a central sensor. Three other permanent arrays, operated by ISLA (I59US, MENE, KHLU), have been established since 2000, 2006, and 2009, respectively, and consist of a combination of Chaparral 25 and 50 sensors. Each infrasound instrument within the HVO arrays is built around an AllSensor MEMS sensor, which has higher noise characteristics than a Chaparral 25, but similar frequency response. Since the full network has been established, we have recorded several infrasound signals including infrasonic tremor from Halema`uma`u, rockfall from the craters of Halema`uma`u and Pu`u `Ō`ō, and other natural and anthropogenic infrasound from diverse sources on-island, offshore, and aloft. Future developments will include real-time detection, location, and identification of infrasonic signals for eruption notification. MENE data is presently available publically, and the HVO infrasound data should be available for research projects through the IRIS Data Management Center beginning in March 2013.
Session: Realtime Monitoring and Early Warning
Presenter   Hellweg, Margaret
Schedule   Fri 1:30 PM / Oral
Room   155A
Operating and Improving Earthquake Early Warning for Northern California and the US West Coast
HELLWEG, M., UC Berkeley Seismo Lab, Berkeley, CA, peggy@seismo.berkeley.edu; ALLEN, R. M., UC Berkeley Seismo Lab, Berkeley, CA, rallen@berkeley.edu; COLOMBELLI, S., Universita di Napoli, Napoli, Italy, scolombelli@unina.it; GRAPENTHIN, R., UC Berkeley Seismo Lab, Berkeley, CA, ronni@seismo.berkeley.edu; HENSON, I., UC Berkeley Seismo Lab, Berkeley, CA, henson@seismo.berkeley.edu; JOHANSON, I., UC Berkeley Seismo Lab, Berkeley, CA, ingrid@seismo.berkeley.edu; KUYUK, H. S., UC Berkeley Seismo Lab, Berkeley, CA, serdarkuyuk@gmail.com; NEUHAUSER, D., UC Berkeley Seismo Lab, Berkeley, CA, doug@seismo.berkeley.edu
A demonstration earthquake early warning (EEW) system, CISN ShakeAlert, is now operating on realtime seismic data in California. Over the past year, with support from the Moore Foundation and the USGS, we have made considerable improvements to the performance of ElarmS, the algorithm developed and operated by UC Berkeley, in terms of both reporting speed and accuracy. We are now operating ElarmS-2 or E2, which has two main modules. The first, continuous waveform processing of realtime seismic data, runs locally in the datacenters at UC Berkeley (UCB), Caltech, and USGS Menlo Park. The data are passed to the second module, for event detection, which runs only at UCB. It analyzes data from the three waveform processing streams, identifies earthquakes and sends alerts to the ShakeAlert Decision Module for publishing to test users from research, government agencies and some industrial partners. For example, BART is now slowing trains based on information from the EEW system.For E2, both waveform processing and event detection modules have been improved; communication software between modules is more flexible; and the method of associating triggers has been updated. In addition, we have added filters for split events and teleseisms and improved other aspects of event characterization to optimize performance. We are now moving to improve EEW performance for large earthquakes, with the funding from the Moore Foundation. We have developed and tested a method to use GPS data to rapidly estimate magnitude for very large earthquakes. In addition to continuing to maintain and improve E2, our activities are directed toward realtime GPS processing to that will rapidly pass such estimates to the EEW system for evaluation and incorporation into earthquake alerts.
Session: Earthquake Source Physics
Presenter   Peng, Zhigang
Schedule   Thu 11:45 AM / Oral
Room   155A
Immediate Foreshock Activity of the 2010 Mw6.9 Yushu, Qinghai Earthquake
PENG, Z., Georgia Institute of Technology, Atlanta, GA, zpeng@gatech.edu; WANG, B., Institute of Geophysics, China Earthquake Administration, Beijing, China, wangbs@cea-igp.ac.cn; TU, H., Earthquake Administration of Qinhai Province, Xining, Qinhai Province, China, tuhongwei33@sina.com
We conduct a detailed study of the foreshock sequence of the 2010 Mw6.9 Yushu, Qinghai earthquake in China, by examining continuous waveforms recorded at a seismic station near the mainshock rupture zone. We manually identify foreshocks with double peaks in the 5-Hz high-pass-filtered seismograms, and use waveforms of 12 earthquakes listed in a local catalog as templates to scan through the continuous waveforms for new detections. We have identified up to ~200 events in the last 9000 s before the Yushu mainshock. The waveforms of the newly detected events are very similar, suggesting that they occurred in a relatively compacted region. The entire foreshock sequence could be largely considered as aftershocks of an Ml4.6 earthquake that occurred about 2 hours before the mainshock, with statistical parameters (b and p values) not much different with aftershock sequences elsewhere. Our observation does not support an inference of accelerating foreshock activities that lead to the mainshock rupture, as was observed right before the Mw7.4 Izmit earthquake [Bouchon et al., 2011]. This suggests that the accelerated nucleation process of a large earthquake, as inferred by laboratory experiments and seismic studies, may not be reliably observed. This, together with other recent studies [e.g., Doi and Kawakata, 2012; Wu et al., 2013], indicates that foreshock sequences could be complex, and it is challenging to use them as possible warnings for upcoming large events.
Session: Velocity Models and Modeling
Presenter   Desser, Elizabeth
Schedule   Thu AM / Poster
Room   Hall 1
Anisotropy Beneath the Tibetan Plateau - A Survey of Shear Wave Splitting Analyses
DESSER, E. M., , Houston, TX, elizabeth.desser@gmail.com; AMMON, C. J., Penn State University, State College, PA, cja12@psu.edu
One of the most impressive surficial features on Earth is the Himalayan-Tibetan orogen, comprising the Himalaya Range and the associated Tibetan plateau to its north-northeast. Numerous and diverse seismic deployments have been conducted in Tibet to advance our understanding of the state of anisotropy in the crust and upper mantle beneath the plateau. Results from previous studies of shear wave splitting on the Tibetan Plateau suggest several models of mantle lithospheric flow, such as the rotational flow patterns observed around the Eastern Himalaya syntaxis. We present the results of a reanalysis of many of these observations. Specifically, we use the SplitLab processing environment to measure the shear wave splitting of teleseismic shear wave SKS and SKKS phases recorded on 386 broadband seismic stations within nine networks deployed across the Tibetan Plateau between 1992 and 2007. As had been noted by earlier researchers, we observe some strong variations in SKS splitting parameters across major terrane boundaries, but also detect substantial lateral variations in the SKS splitting in the west-to-east direction within both the Lhasa and Qiangtang terranes. In the Lhasa terrane, the apparent splitting decreases from west to east, in contrast with the Qiangtang terrane, which shows an eastward increase in SKS splitting. Patterns in much of the Himalaya and in southeastern Tibet are complex, with rapid apparent changes in the overall SKS splitting parameters. The results suggest a complex three-dimensional variation in mineral alignment, interacting with some of the boundaries obvious at the surface, but including another more subtle component most easily seen when all the observations are processed and analyzed in a uniform investigation.
Session: Earthquake Source Physics
Presenter   Wu, Chunquan
Schedule   Thu PM / Poster
Room   Hall 1
Non-Accelerating Foreshock Activity of the 1999 Mw7.1 Duzce, Turkey Earthquake
WU, C., Geophysics Group, Los Alamos National Laboratory, Los Alamos, NM, cwu@lanl.gov; PENG, Z., School of Earth and Atmospheric Sciences, Georgia Tech, Atlanta, GA, zpeng@gatech.edu; BEN-ZION, Y., Department of Earth Sciences, University of Southern California, Los Angeles, CA, benzion@usc.edu
Laboratory and theoretical studies have shown that accelerated local failure could occur at the transition from locked or slowly creeping fault to dynamic rupture. Recently, Bouchon et al. (2011) found tantalizing evidence for accelerated local failure shortly before the 1999 Mw7.4 Izmit, Turkey earthquake, by detecting in data of a near-fault instrument a family of small foreshocks that were buried in the noise. To clarify whether a similar accelerated foreshock activity exists for other large earthquakes, we use the recently developed match-filter technique to detect additional seismic events before the Mw7.1 Duzce earthquake. We employ extensive waveform data set recorded by a 10-station fault zone array deployed around the eastern end of the Izmit rupture zone a week after the Izmit mainshock. This deployment captured seismic activities before, during and after the Duzce earthquake that occurred ~3 months later. In details, we use ~4800 events within 20km to the Duzce epicenter listed in the catalog as templates to scan through waveforms recorded within ~65 hours before the Duzce earthquake. We detected in that time interval 690 events, which is ~11 times more than the 59 events listed in the original catalog. Most of the foreshocks occur to the west of the Duzce epicenter, while the Duzce mainshock initiated between the more active west and relatively quite east fault segments. We did not observe clear acceleration or migration of the foreshock activity during the 65-hour period before the Duzce earthquake. Instead, we found that the Duzce source region become less active during the 5 hours immediately before the main shock. Our results, together with other recent studies (e.g., Peng et al. 2013) suggest that progressive acceleration/localization of foreshock activities around the mainshock epicenters is not a general phenomenon.
Session: What are the Limits of Explosion Source Model Predictions?
Presenter   Walter, William
Schedule   Wed 9:30 AM / Oral
Room   155D
Insights from the Source Physics Experiments on P/S Amplitude Ratio Methods of Identifying Explosions in a Background of Earthquakes
WALTER, W. R., Lawrence Livermore National Laboratory, Livermore, CA, walter5@llnl.gov; FORD, S. R., Lawrence Livermore National Laboratory, Livermore, CA, ford17@llnl.gov; PYLE, M., Lawrence Livermore National Laboratory, Livermore, CA, pyle4@llnl.gov; PASYANOS, M. E., Lawrence Livermore National Laboratory, Livermore, CA, pasyanos1@llnl.gov; MELLORS, R. J., Lawrence Livermore National Laboratory, Livermore, CA, mellors1@llnl.gov; MATZEL, E., Lawrence Livermore National Laboratory, Livermore, CA, matzel1@llnl.gov; HAUK, T., Lawrence Livermore National Laboratory, Livermore, CA, hauk1@llnl.gov
Regional distance (200-1600 km) amplitude ratios of seismic P-to-S waves at sufficiently high frequencies (~>2 Hz) can identify explosions among a background of natural earthquakes. However the physical basis for the generation of explosion S-waves, and therefore the predictability of this P/S technique as a function of event properties such as size, depth, geology and path, remains incompletely understood. A goal of the Source Physics Experiments (SPE) at the Nevada National Security Site (NNSS, formerly the Nevada Test Site (NTS)) is to improve our physical understanding of the mechanisms of explosion S-wave generation and advance our ability to numerically model and predict them. Preliminary analysis of the first 3 SPE explosions reveal several interesting results: 1) The explosion P and S corner frequencies appear to be similar, in contrast to some model predictions. 2) The 1-ton chemical explosions (SPE-2 and SPE-3) appear to discriminate well from background earthquakes in the frequency band 4-8 Hz, where P and S signals are visible at the NVAR array located near Mina, NV about 200 km away, consistent with past nuclear explosions recorded at the site. 3) Local distance recordings (< 100 km) of P/S ratios do not appear to discriminate the SPE explosions from earthquakes at some stations across a broad range of frequencies from 4-64 Hz. Strong apparent structural focusing and defocusing effects on the waveforms appear to cause the poor discrimination and imply that local P/S explosion identification may require more complex path corrections than the simple 1 and 2-D corrections that are sufficient at regional distances. We are using the new SPE data to improve our explosion models and our ability to understand and predict where P/S methods of identifying explosions work and any circumstances where they may not.This work performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.
Session: Triggering of Seismic and Volcanic Events
Presenter   Peng, Zhigang
Schedule   Fri 9:00 AM / Oral
Room   155D
Dynamic and Delayed Triggering of Moderate-Size Earthquakes in East Asia
PENG, Z., School of Earth and Atmospheric Sciences, Georgia Tech, Atlanta, GA, zpeng@gatech.edu; WU, C., Geophysics Group, Los Alamos National Laboratory, Los Alamos, NM, cwu@lanl.gov; YAO, D., School of Earth and Atmospheric Sciences, Georgia Tech, Atlanta, GA, dyao30@gatech.edu; MENG, X., School of Earth and Atmospheric Sciences, Georgia Tech, Atlanta, GA, xmeng7@gatech.edu
Previous studies have shown that large earthquakes can dynamically trigger microearthquakes and deep tectonic tremor/slow-slip events at regional and teleseismic distances. In addition, a recent study has shown that the 2012 Mw8.6 earthquake in eastern Indian Ocean has triggered a transient increase of M>5.5 earthquakes globally for a few days [Pollitz et al., 2012]. However, it is still not clear whether large earthquakes could trigger moderate-size events at much later times.Here we conduct two lines of studies to investigate possible seismicity rate changes in East Asia following several large earthquakes in this region. In the first study, we examine changes in seismic activities from regional earthquake catalogs in Continental China and Taiwan. We select M>5 earthquakes within 6 months of each distant mainshock, and focus on the seismicity rate changes in the source regions of the selected M>5 events. In several cases, we observe clear increase of background seismicity following distant mainshock, followed by the occurrence of M>5 events ranging from a day to a few months afterwards. In the second study, we focus on the source regions of three moderate-sized (M>6) normal-faulting earthquakes occurred in 2004, 2005, 2008 in Zhongba, Tibet. We use 32 local earthquakes in the Zhongba region as templates, and scan through the continuous data recorded by the HiClimb array around the occurrence times of the 2004 Sumatra and 2005 Nias earthquakes. Based on newly detected events, we observe a transient increase of local seismic activity between the 2005 Nias earthquake and 2005 Ms6.5 Zhongba earthquake, suggesting a possible link between these two events. Through these systematic studies, we hope to better characterize the relationship among distant large mainshock, dynamically triggered microearthquakes and subsequent moderate (M>5) events in East Asia.
Session: Networks and Instrumentation
Presenter   Frassetto, Andy
Schedule   Wed AM / Poster
Room   Hall 1
Sensor Emplacement Techniques and Seismic Noise Analysis for USArray Transportable Array Seismic Stations
FRASSETTO, A., IRIS Consortium, Washington, DC, andyf@iris.edu; BUSBY, R., IRIS Consortium, Washington, DC, busby@iris.edu; HAFNER, K., IRIS Consortium, Washington, DC, hafner@iris.edu; WOODWARD, R., IRIS Consortium, Washington, DC, woodward@iris.edu; SAUTER, A., IRIS PASSCAL Instrument Center, Socorro, NM, allan.sauter@gmail.com
In preparation for the deployment of EarthScope’s USArray Transportable Array (TA) in Alaska beginning in 2014, the National Science Foundation (NSF) is supporting exploratory work on seismic station design, sensor emplacement, and communication concepts appropriate for this challenging high-latitude environment. IRIS has installed several experimental stations to evaluate different sensor emplacement schemes both in Alaska, the Yukon, and in the lower-48 of the U.S. The goal of these tests is to maintain or enhance a station’s noise performance while minimizing its footprint and the weight of the equipment, materials, and overall expense required for its construction. Motivating this approach are new developments in posthole broadband seismometer design and the unique conditions for operating in Alaska, where there are few roads, cellular communications are scarce, most areas are only accessible by helicopter, and permafrost underlies much of the northern tundra.In this study we review our methods used for directly emplacing broadband seismometers in comparison to the current standard for the TA. These new methods primarily focus on using a portable auger to drill three to five meters, beneath the active layer of the permafrost, or coring directly into surface bedrock to one meter depth with a portable drill. Subsequent station performance has been assessed using probability density functions summed from power spectral density estimates. These are calculated for the continuous time series of data recorded by each channel of the seismometer. There are five test stations currently operating in Alaska and the Yukon. One was deployed in August 2011 and the remaining four in October 2012. Our results show that the performance of seismometers in Alaska with auger-hole or core-hole installations equals or exceeds that of the quietest TA stations in the lower-48, particularly at long periods, and in exceptional cases approaches the performance of the GSN low noise model.
Session: When and Why do Earthquake Ruptures Stop? Evaluating Competing Mechanisms of Rupture Termination
Presenter   Bennett, Scott
Schedule   Fri PM / Poster
Room   Hall 1
Does Fault Segmentation Limit Earthquake Magnitude on the Wasatch Fault?
BENNETT, S. E. K., University of California, Davis CA , Golden, CO, sekbennett@ucdavis.edu; GOLD, R. D., U.S. Geological Survey, Golden, CO, rgold@usgs.gov; BRIGGS, R. W., U.S. Geological Survey, Golden, CO, rbriggs@usgs.gov; HAYES, G. P., U.S. Geological Survey, Golden, CO, ghayes@usgs.gov; POWERS, P. M., U.S. Geological Survey, Golden, CO, pmpowers@usgs.gov; FIELD, E. H., U.S. Geological Survey, Golden, CO, field@usgs.gov
Fundamental inputs for seismic hazard models include estimates of maximum earthquake magnitude (Mmax) and earthquake recurrence intervals on a given fault. However, it remains unclear if large earthquakes on segmented faults preferentially rupture multiple structural fault segments, or tend to be limited to individual segments, thereby limiting Mmax and influencing inter-event recurrence times. The Wasatch fault zone (WFZ) is a ~350 km-long active normal fault along the eastern edge of the Basin and Range province, composed of 10 structural segments. Segment boundaries are characterized by transfer faults or stepovers that vary in overlap (up to 12 km), underlap (up to 6 km), and stepover width (1.5 - 7 km). Past paleoseismic work suggests recurrence intervals of 1.3 - 2.5 kyr for large (M>7) earthquakes on the five most active central WFZ segments, adjacent to the largest cities in Utah, with an average collective recurrence interval of ~350 years. Much of this previous work focused on the centers of structural segments, making it difficult to assess if multi-segment fault ruptures have occurred. This study will test if recent surface ruptures on the WFZ have spanned fault segment boundaries and evaluate if the frequency of such large events are properly estimated in current seismic hazard models. To document the presence (or lack) of correlative fault ruptures near adjacent tips of the most active fault segments, we plan to acquire airborne LiDAR, conduct Quaternary mapping, and excavate paleoseismic trench ‘pairs’, using 14C and luminescence geochronology of faulted Quaternary deposits. Temporal correlation of correspondingly large single-event displacements near each fault tip in a stepover would imply that surface rupture spanned the stepover. These new data will allow reevaluation of rupture models along the WFZ, ensuring that future seismic hazard maps apply appropriate weights to multi-segment, single-segment, and partial-segment ruptures.
Session: The Magnitude X.X Earthquake on the YY of ZZZZ: Major Earthquakes of 2012/13
Presenter   Yue, Han
Schedule   Fri PM / Poster
Room   Hall 1
STUDENT
Source Rupture Process of the 5 September 2012 Costa Rica Mw=7.6 Thrust Event from Joint Inversion of High-Rate GPS and Teleseismic P Wave Data
YUE, H., University of California Santa Cruz, Santa Cruz, CA, hyue@ucsc.edu; RIVERA, L., Institut de Physique du Globe de Strasbourg, Paris, France, luis.rivera@unistra.fr; LAY, T., University of California Santa Cruz, Santa Cruz, CA, tlay@ucsc.edu; SCHWARTZ, S., University of California Santa Cruz, Santa Cruz, CA, sschwartz@pmc.ucsc.edu; PROTTI, M., Costa Rica Volcanological and Seismological Observatory, Costa Rica, marino.protti@gmail.com
On 5 September 2012, a large thrust event (Mw=7.6) ruptured a densely instrumented seismic gap on the shallow plate boundary beneath the Nicoya Peninsula, Costa Rica. Networks of strong motion accelerometers, broadband and short-period sensors, and high-rate (5-sps) GPS stations recorded ground motions directly above the rupture zone, providing a unique opportunity to study the detailed source process of a large shallow megathrust rupture using nearby land observations. An inland and relatively deep hypocenter (10.086°N, 85.305°W, 40 km) was estimated by the USGS, and teleseismic W-phase inversions also indicate a relatively large (30-40 km) centroid depth. Hypocenter relocation performed using the local seismic network data indicates that the event initiated with small emergent seismic waves from a hypocenter ~10 km offshore (9.80°N, 85.53°W) 15 km deep on the megathrust. The local origin time (14:42:05) is 3 s earlier than the USGS origin time, compatible with the shallower source depth. A joint finite-fault inversion of 0.2 Hz lowpass-filtered hr-GPS recordings and teleseismic P waves reveals that the primary slip zone is located beneath the Nicoya coastline up-dip from the USGS location. Complete ground motions are computed for the hr-GPS stations using a 1D regional velocity model and a wavenumber integration program from Robert Herrmann. The large-slip region extends ~50 km along strike and ~30 km along dip, with a centroid depth of ~25 km. The maximum slip is ~3.5 meters and Mw=7.6, consistent with teleseismic estimates. The inversion indicates that the rupture propagated down-dip from the hypocenter with a rupture velocity of ~2.5 km/s. The inversion has limited resolution of any offshore slip, but a small secondary arrival in the hr-GPS data appears to originate from a region about 30 km offshore along the northern half of the rupture zone. We consider the relationship between coseismic slip location, aftershocks and adjacent tremor and slow-slip regions.
Session: The Magnitude X.X Earthquake on the YY of ZZZZ: Major Earthquakes of 2012/13
Presenter   Ye, Lingling
Schedule   Fri 9:00 AM / Oral
Room   155A
STUDENT
Recent Large Earthquake Seismic Radiation and Rupture Process Variations on the Central America Megathrust
YE, L., UC Santa Cruz, Santa Cruz, CA, lye2@ucsc.edu; LAY, T., UC Santa Cruz, Santa Cruz, CA, tlay@ucsc.edu; KANAMORI, H., California Institute of Technology, Pasadena, CA, hiroo@gps.caltech.edu
The megathrust fault between the underthrusting Cocos plate and the overriding Caribbean plate recently experienced three large ruptures: the August 27, 2012 (Mw 7.3) El Salvador, September 5, 2012 (Mw 7.6) Costa Rica, and November 7, 2012 (Mw 7.4) Guatemala earthquakes. All three events involve shallow-dipping thrust faulting, but they ruptured the shallow, central, and deeper portion of the plate boundary in each region, respectively. The El Salvador earthquake ruptured from about 4 to 17 km depth (Domain A) along the megathrust with a relatively large centroid time of 19.4 s, low seismic moment-scaled energy release (Er/M0 ~ 0.78×10-4) and a depleted short-period source spectrum similar to that of the 1992 (Mw 7.6) Nicaragua tsunami earthquake that ruptured the adjacent shallow portion of the megathrust. The Costa Rica earthquake ruptured a depth range of about 15 to 23 km (Domain B), with a more typical source spectrum. Regional seismic recordings clearly display differences in P wave spectrum between the Costa Rica and El Salvador events consistent with teleseismic observations. The Guatemala event ruptured from about 22 to 30 km depth (Domain C), with a relatively enriched short-period source spectrum, high moment-scaled energy release, and a small centroid time (9.3 s). A waveform template analysis using regional broadband data for the 91 aftershocks of the El Salvador event, including 3 normal faulting and 3 thrusting events from the gCMT catalog, finds 58 likely thrust-faulting events and 21 likely normal-faulting events with 12 unidentified events. The activation of normal faulting by the El Salvador event is similar to other large tsunami earthquakes, possibly indicating relatively complete stress release on the megathrust. Modeling of spectral ratios of similar mechanism events separated by ≤ 30 km is used to examine spectral differences among the thrust and normal faulting populations.
Session: The Magnitude X.X Earthquake on the YY of ZZZZ: Major Earthquakes of 2012/13
Presenter   Yue, Han
Schedule   Fri 8:45 AM / Oral
Room   155A
STUDENT
En Echelon and Conjugate Fault Ruptures of the 11 April 2012 Great Indo-Australia Intraplate Earthquakes
YUE, H., University of California Santa Cruz, Santa Cruz, CA,, hyue@ucsc.edu; LAY, T., University of California Santa Cruz, Santa Cruz, CA, tlay@ucsc.edu; KOPER, K., The University of Utah, Salt Lake City, UT, koper@seis.utah.edu; HILL, E., Nanyang Technological University, Singapore, ehill@ntu.edu.sg
On 11 April 2012, the largest recorded strike-slip earthquakes (Mw 8.7) occurred ~100-200 km southwest of the Sumatra subduction zone. We jointly invert 3-component high-rate (1 sps) GPS records from the SuGAr geodetic network in Sumatra, and global teleseismic body and surface waves to constrain the rupture process of this event. Finite-fault inversions are constrained by back-projections off short-period seismic network data, but we test alternate fault geometries and rupture velocities motivated by some alternate interpretations now in the literature. The mainshock had an extraordinarily complex 4-fault rupture lasting ~160 s. The rupture initially expanded bilaterally with large slip (20-30 m) on a right lateral WNW-ESE trending strike-slip fault, then bilateral rupture was triggered on an conjugate left-lateral strike-slip fault trending N-S that crosses the first fault, followed by westward rupture on a second WNW-ESE strike-slip fault offset ~150 km toward the southwest from the first fault. Finally, rupture was triggered on another en echelon fault WNW-ESE fault ~330 km west of the epicenter crossing the Ninetyeast Ridge. We do not resolve any clear evidence for super-shear rupture, as has been proposed in one back-projection analysis. The broadband seismic and geodetic data provide limited resolution of fault geometries, but slightly favor the above geometry inferred from back-projection analysis. An Mw 8.2 aftershock at an epicenter 185 km SSW of the mainshock epicenter, appears to have ruptured bilaterally on a NNE-SSW fault based on back-projection of short-period seismic network data. The complex faulting limits our resolution of the slip distribution, however, both teleseismic and hr-GPS observations support rupture of the main right lateral strike-slip fault extending through the crust and into the upper mantle. This represents large lithospheric deformation that may eventually lead to a localized boundary between the Indian and Australian plates.
Session: Next Generation of Ground Motion Prediction Models
Presenter   Anderson, John
Schedule   Wed 11:30 AM / Oral
Room   155C
The Fukushima Hamadori, Japan, Earthquake: Extensive Ground Motion Data from a Surface-Rupturing Normal Fault
ANDERSON, J. G., Nevada Seismological Laboratory, University of Nevada, Reno, NV, jga@unr.edu; KAWASE, H., Disaster Prevention Research Institute (DPRI), Kyoto University, Uji, Kyoto, Japan, kawase@zeisei.dpri.kyoto-u.ac.jp; BIASI, G., Nevada Seismological Laboratory, University of Nevada, Reno, NV, glenn@seismo.unr.edu; BRUNE, J. N., Nevada Seismological Laboratory, University of Nevada, Reno, NV, brune@seismo.unr.edu; AOI, S., Natl Res Inst Earth Sci & Disaster Prevention (NIED), Tsukuba, Ibaraki, Japan, aoi@bosai.go.jp
A crustal normal-faulting earthquake (MJMA=7.0; MW=6.69) occurred in eastern Tohoku, Japan, on April 11, 2011. K-NET and KiK-net stations recorded 82 records from within 100 km of the fault. These data and data from associated foreshocks and aftershocks will make a critical contribution to future improvements of ground motion prediction for normal faulting earthquakes.Peak accelerations (PGA) and peak velocities (PGV) are compared with four ground motion prediction equations (GMPEs) that include the style of faulting as a predictor parameter. For distances under 100 km, and using a network average value of VS30, the average ratio of PGA to the selected GMPEs (the event term ) is high by factors of 2.3-3.7. Event terms for peak velocity (PGV) are high by factors of 1.4-1.8. Adjusting PGA and PGV with customized site terms (Kawase and Matsuo, 2004), the standard deviations of PGA and PGV residuals are reduced from 0.59 to 0.43, and from 0.53 to 0.35, respectively. The event terms decreased to relatively small factors of 1.1-1.8 for PGA and increased slightly to 1.5-2.0 for PGV. Thus site terms are very important, but positive event terms remain. The remaining positive event terms are not explained by high stress drop, which was typical of crustal events of all mechanisms globally or in Japan. Two subparallel faults rutpured, but source inversions we reviewed found that they ruptured sequentially, so simultaneous contributions from the two faults do not cause high motions. While these observations may tend to suggest that ground motions in large normal faulting events are larger predicted by the tested models, we are not aware of any observations from this event that contradict the precarious rock evidence of Brune (2000) that ground shaking is low on the footwall near the rupture.
Session: Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Models
Presenter   Neighbors, Corrie
Schedule   Wed AM / Poster
Room   Hall 1
Investigation of the High-Frequency Attenuation Parameter, κ (kappa), from Aftershocks of the 2010 Mw 7.1 Darfield, New Zealand Earthquake
NEIGHBORS, C., University of California, Riverside, Riverside, CA, cneig001@ucr.edu; LIAO, E. J., University of California, Riverside, Riverside, CA; COCHRAN, E. S., U.S. Geological Survey, Pasadena, CA; CHUNG, A. I., Stanford University, Stanford, CA; LAWRENCE, J. F., Stanford University, Stanford, CA; KAISER, A. E., GNS Science, Lower Hutt, New Zealand; FRY, B., GNS Science, Lower Hutt, New Zealand; CHRISTENSEN, C., Stanford University, Stanford, CA
Here we investigate the high-frequency spectral decay parameter, kappa (κ), for aftershocks of the 3 September 2010 Mw 7.1 Darfield earthquake on a network of over 180 low-cost, 14-bit Quake-Catcher Network accelerometers. Kappa is proposed to be largely influenced by the surface geology at a recording station and thus can potentially be used to characterize the local site response. The data show that aftershocks of magnitude >M4.0 within 40 km of the main network deployment in Christchurch were well recorded by QCN sensors. In addition to the QCN stations, we also consider records from the local New Zealand GeoNet network. We determine kappa values for 151 unique earthquakes recorded on 273 stations (181 QCN and 82 GeoNet), for a total of over 13,300 horizontal component records. We employ an automated method to estimate kappa from the direct S-wave acceleration spectra. Initial κ estimates range between 0.00017 and 0.067 sec. Preliminary results suggest κ increases as the source-receiver distance increases; however, some stations show a weak relationship due to a large amount of scatter in the kappa values. We find good agreement in the κ estimates between the low-cost QCN network and the GeoNet network; although the QCN stations recorded a more limited hypocentral distance range (<40 km) compared to the GeoNet stations (<700 km). Comparing kappa values to the mapped geology in Christchurch, we find lower kappa estimates at sites underlain by volcanics and higher kappa values at stations on fluvial sediments and dune deposits. Future work will include an investigation of κ with source magnitude and consideration of site attenuation in relation to the basin velocity model, the depth to basement, and the shallow water table, which has implications for liquefaction hazard.
Session: Triggering of Seismic and Volcanic Events
Presenter   Marrone, James
Schedule   Fri PM / Poster
Room   Hall 1
Curious Increase in the Occurrence Rate of Mainshock Seismicity in Southeastern US Following the 1886 Charleston Earthquake
MARRONE, J. E., Bechtel Corporation, San Francisco, CA, jmarrone@bechtel.com; TAVAKOLI, B., Bechtel Corporation, Frederick, MD, btavakol@bechtel.com
In some earthquake catalogs for the central and eastern US [CEUS], there appears to be a curious increase in the rate of mainshock earthquake activity throughout South Carolina, North Carolina, and Georgia, for ~1 year following the Mw 6.9 Sept. 1, 1886 Charleston, S.C. earthquake, particularly for the first 3 months. In the earthquake catalog prepared for the CEUS Seismic Source Characterization [CEUS SSC] Project (EPRI et al., 2012), the annual rate of mainshock earthquakes with magnitude Mw 2.2 and greater spikes at a rate 3 to 4 times that of near earlier and later time intervals. The CEUS SSC report details the catalog preparation, including 1) consideration of multiple source catalogs; 2) identification/removal of duplicate and non-tectonic events; 3) distinction between mainshock and cluster events [e.g., foreshocks/aftershocks]; and 4) assessment of the catalog completeness. These elements are critical when the catalog is used for the assessment of earthquake recurrence rates as input to a probabilistic seismic hazard assessment. An anomalous rate of seismic activity could impact local recurrence statistics where catalog analyses follow conventional assumptions. This increase in activity rate following the Charleston event can also be found in the USGS CEUS earthquake catalog used for development of the 2008 National Seismic Hazard Maps (Petersen et al., 2008), but, perhaps significantly, not in the EPRI-SOG (1988) CEUS catalog. This poster presents an investigation of these three earthquake catalogs, looking for the cause of the regional increased rate of cataloged seismicity – or lack thereof – following the 1886 Charleston earthquake. Dependence on minimum magnitude and time window is investigated. Initial thoughts of a cause: • Need of re-calibration of declustering parameters; • Regional triggered earthquakes; • Enhanced, temporary regional awareness to earthquakes by the public, resulting in a temporary increase in events cataloged.
Session: Towards an Integrated Understanding of Slow Earthquakes: What We Know, What We Don’t Know, and How to Move Forward
Presenter   Yabe, Suguru
Schedule   Fri 3:45 PM / Oral
Room   155A
STUDENT
Comparison of Tectonic Tremor Activities in Circum-Pacific Subduction Zones
YABE, S., The University of Tokyo, Tokyo, Japan, yabe@eps.s.u-tokyo.ac.jp; IDEHARA, K., The University of Tokyo, Tokyo, Japan, idehara@eps.s.u-tokyo.ac.jp; IDE, S., The University of Tokyo, Tokyo, Japan, ide@eps.s.u-tokyo.ac.jp
Tectonic tremor activities have been detected in many circum-Pacific subduction zones. Although these tectonic tremors share some characteristics (e.g., low frequency content, ambiguous P wave arrival, and successive occurrences), a large variety exists in their hypocenter distributions, migration behavior, durations, and recurrence intervals. In addition to a regional variety of tremor activity in each subduction zone, we also observe significant differences among different tremor zones. Here we investigate the variety of tremor activity based on updated tremor catalogs constructed using continuous seismic records, as an extended study of Ide (2012). The study regions are Nankai (including Kyushu), Taiwan, Cascadia, Mexico (Jalisco and Guerrero), South Chile, and New Zealand. We detect tectonic tremors using a modified version of envelope correlation method. After locating tremors, we measure the duration of each detected tremor, as the half value width of stacked envelope aligned at arrival times. We measure the sensitivity of tectonic tremor to tide (Ide, 2010), the average recurrence intervals, and the fractal dimension of temporal clustering (Idehara et al., submitted). Generally, tremor events with shorter pulse duration tend to be sensitive to tide, and less clustered with a shorter recurrence interval. Shallower tremors have longer recurrence intervals and deeper tremors have shorter recurrence intervals (Idehara et al., submitted), as previous studies suggested (e.g. Obara et al., 2010; Wech and Creager, 2011).
Session: New Developments in Earthquake Forecasting and Predictability Research
Presenter   Tsuruoka, Hiroshi
Schedule   Thu 8:30 AM / Oral
Room   155D
Predictability Study on the Aftershock Sequence of the 2011 Tohoku Earthquake using Preliminary JMA Catalogue
TSURUOKA, H., The Univ. Tokyo, Tokyo, Japan, tsuru@eri.u-tokyo.ac.jp; NANJO, K., The Univ. Tokyo, Tokyo, Japan, kazuyoshinanjo@gmail.com; SAYOKO, Y., The Univ. Tokyo, Tokyo, Japan, sayokoi@eri.u-tokyo.ac.jp; HIRATA, N., The Univ. Tokyo, Tokyo, Japan, hirata@eri.u-tokyo.ac.jp
We investigated a predictability of forecasting models for the aftershocks of the 2011 Tohoku earthquake. Our approach is based on testing five (ERS, ETES, ETAS, HIST-ETAS5pa, and HIST-ETAS7pa) different 1-day forecast models used in CSEP-Japan. The difference from the original CSEP experiment is in the use of the preliminary JMA catalogue to allow for immediate testing after the first weeks of the aftershock sequence that otherwise would be delayed due to the finalization of the regular JMA catalogue. All 1-day forecasts, except for ETES, seem to have been poor at forecasting the event numbers. The spatial accuracy of the HIST-ETAS models has been improved in a delayed fashion. These reveal a problem in the CSEP testing setup. The CSEP protocol for daily forecasts uses each forecast model to take into account the seismicity of the previous day. However, when seismicity is changing rapidly over time during the beginning of an aftershock sequence, earthquake forecast models tended to underestimate earthquake number. Therefore, we should be updated more frequently than 1-day intervals to improve accuracy of the forecasts. Lastly, we emphasize the need of further analysis on model improvement to assure the reliability of forecasts for the days immediately after the main shock.
Session: Infrasound and Seismoacoustics
Presenter   Johnson, Jeffrey
Schedule   Wed 2:30 PM / Oral
Room   155D
Probing Local Wind and Temperature Structure Using Infrasound from Volcan Villarrica (Chile)
JOHNSON, J. B., Boise State University, Boise, ID, jeffreybjohnson@boisestate.edu; ANDERSON, J. F., New Mexico Tech, Socorro, NM, ajakef@gmail.com; MARCILLO, O., Los Alamos National Labs, Los Alamos, NM, omarcillo@lanl.gov; ARROWSMITH, S., Los Alamos National Labs, Los Alamos, NM, arrows@lanl.gov
We use the continuous and intense (∼107 W) infrasound produced by Volcan Villarrica (Chile) to invert for the local dynamic wind and temperature structure of the atmosphere. Infrasound arrays deployed in March 2011 at the summit (2826 m) and on the NNW flank (∼8 km distant at 825 m) were used to track infrasound propagation times and signal power. We model an atmosphere with vertically varying temperature and horizontal winds and use propagation times (ranging from 23 to 24 s) to invert for horizontal slowness (2.75–2.94 s/km) and average effective sound speeds (328–346 m/s) for NNW propagating infrasound. The corresponding ratio of recorded acoustic power at proximal versus distal arrays was also variable (ranging between 0.15 to 1.5 for the peak 0.33–1 Hz infrasound band). Through application of geometrical ray theory in a uniform gradient atmosphere, these ‘amplification factors’ are modeled by effective sound speed lapse rates ranging from −15 to +4 m/s per km. NNW-projected wind speeds ranging from −20 m/s to +20 m/s at 2826 m and wind gradients ranging from −11 to +10 m/s per km are inferred from the difference between effective sound speed profiles and adiabatic sound speeds derived from local temperature observations. The sense of these winds is in general agreement with regional meteorological observations recorded with radiosondes. We suggest that infrasound probing can provide useful spatially averaged estimates of atmospheric wind structure that has application for both meteorological observation and volcanological plume dispersal modeling.
Session: New Frontiers in Seismic Data Analysis
Presenter   Asten, Michael
Schedule   Wed 5:15 PM / Oral
Room   155A
The Use of Wavenumber Normalization in Computing Spatially Averaged Coherencies (KRSPAC) of Microtremor Data from Asymmetric Arrays
ASTEN, M. W., Monash University, Melbourne, Australia, michael.asten@monash.edu; STEPHENSON, W. J., US Geological Survey, Golden CO, wstephens@usgs.gov; HARTZELL, S., US Geological Survey, Golden CO, shartzell@usgs.gov
The SPAC method of processing microtremor noise observations for estimation of Vs profiles has a limitation that the array has circular or triangular symmetry in order to allow spatial (azimuthal) averaging of inter-station coherencies over a constant station separation. Common processing methods allow for station separations to vary by typically +-10% in the azimuthal averaging before degradation of the SPAC spectrum is excessive. A limitation on use of high-wavenumbers in inversions of SPAC spectra to Vs profiles has been the requirement for exact array symmetry to avoid loss of information in the azimuthal averaging step. In this paper we develop a new wavenumber-normalised SPAC method (KRSPAC) where instead of performing averaging of sets of coherency versus frequency spectra and then fitting to a model SPAC spectrum, we interpolate each spectrum to coherency versus k.r, where k and r are wavenumber and station separation respectively, and r may be different for each pair of stations. For fundamental–mode Rayleigh-wave energy the model SPAC spectrum to be fitted reduces to Jo(kr). The normalization process changes with each iteration since k is a function of frequency and phase velocity and hence is updated each iteration. The method proves robust and is demonstrated on data acquired in the Santa Clara Valley, CA, (Site STGA) where an asymmetric array having station separations varying by a factor of 2 is compared with a conventional triangular array; a 300m-deep borehole with a downhole Vs log provides nearby ground truth. The method is also demonstrated on data from the Pleasanton array, CA, where station spacings are irregular and vary from 400m to 1200m. The KRSPAC method allows inversion of data using kr (unitless) values routinely up to 30, and occasionally up to 60. Thus despite the large and irregular station spacings, this array permits resolution of Vs as fine as 15m for the near-surface sediments, and down to a maximum depth of 2.5km.
Session: Advances in High-Frequency Physics-Based Earthquake Systems Simulation
Presenter   Thornley, John
Schedule   Wed PM / Poster
Room   Hall 1
Modeling the Site Response of a Downhole Array Site in Anchorage, Alaska
THORNLEY, J. D., Golder Associates Inc., Anchorage, AK, john_thornley@golder.com; DUTTA, U., University of Alaska, Anchorage, AK, udutta2@uaa.alaska.edu; YANG, Z., University of Alaska, Anchorage, AK, zyang2@uaa.alaska.edu
Alaska’s largest city, Anchorage, is located within a seismically active region including the Pacific Megathrust subduction zone and crustal faults on the North American Plate. A complex geology of glacial till, Bootlegger Cove Clays, and alluvium exists within the Anchorage area. A downhole array, with three component accelerometers placed at seven depths from the surface to 61 meters below ground, has been recording strong ground motions since 2004. Recorded ground motions at the borehole array have been analyzed as part of a systematic approach to evaluate the attenuation relationships and impedance characteristics of the soil column of the Anchorage basin. A one-dimensional equivalent linear model has been developed and calibrated using measured ground motions. Once the model was calibrated, input ground motions from other measured earthquakes at the borehole array were applied to the base of the model and response spectra at each of the seven depths were compared to measured response spectra at those respective depths. The shapes of the response spectra were similar for both cases. It was observed that two to three predominant frequencies, ranging from 1 to 10 Hz, generally occurred in the measured records. These same predominant frequencies were also observed in the response spectra at the same depths for the calibrated model. The comparison of the measured and modeled response spectra shows that predominant frequencies match well. Differences in the amplitudes of the measured and modeled predominant frequencies were reduced by applying minor adjustments to damping parameters in the model. A sensitivity study of model parameters was also performed to identify the model inputs that most affected the spectral matching results.
Session: Infrasound and Seismoacoustics
Presenter   Demonte, Philippa
Schedule   Thu AM / Poster
Room   Hall 1
STUDENT
Detection, Characterization, and Quantification of Geyser Eruptions: Insights from Infrasound Monitoring at Yellowstone National Park
DEMONTE, P. J., Boise State University, Boise, ID, philippademonte@u.boisestate.edu; JOHNSON, J. B., Boise State University, Boise, ID, jeffreybjohnson@boisestate.edu
Geysers are hot springs, which intermittently erupt hot water and steam. Their eruption phenomena includes bubble bursts and two-phase fluid jets, which are strong emitters of near-infrasound (1-20 Hz) and low frequency audible airwaves. This study investigates infrasound from geysers in Yellowstone to track eruption statistics, and to characterize and quantify eruption parameters.During October 2012, we deployed three 7-element arrays of low frequency microphones recording continuously in the Lower Geyser Basin. We use frequency slowness analysis to identify coherent waves and calculate signal back-azimuth. Intersecting back azimuths coincided with at least 6 distinct geysers at distances of up to 3 kilometers. We note improved detection capabilities of the 7-element electret condenser sensor arrays compared to a previous installation (in August 2011) of two 4-element MEMS sensor arrays. We also present infrasound waveform data from three distinctive geysers including Great Fountain, Sawmill Geyser, and Lone Star Geyser. The observed infrasound waveforms are compared spectrally and in terms of acoustic power and related to video observations of their corresponding eruption characteristics.Great Fountain Geyser typically produces episodic bi-modal pressure bursts, which we characterize as monopole, or volumetric, sources, with the largest of the events displacing more than 50 m3 of atmosphere. The periodic (0.7 s) pressure pulses produced by Sawmill contrast with Great Fountain in that they commence with an infrasound rarefaction, which we interpret as the implosion (or collapse) of steam bubbles. Observed signals of Lone Star Geyser are manifested as ~30-minute tremor episodes with higher frequency content than the other two fountain-type geysers. Its spectra and nature are indicative of jet signal, which transitions between relatively low frequencies (20-60 Hz; when dominated by water) to higher frequencies (40-85 Hz; when dominated by steam).
Session: Seismic Hazards and Ground Motions
Presenter   Gueguen, Philippe
Schedule   Fri 4:30 PM / Oral
Room   155B
Contribution of in Situ Measurements in the Evaluation of the Seismic Vulnerability of Existing Buildings
PERRAULT, M., ISTerre - CNRS/IFSTTAR - Université Joseph Fourier Grenoble I, Grenoble, France, matthieu.perrault@ujf-grenoble.fr; GUEGUEN, P., ISTerre - CNRS/IFSTTAR - Université Joseph Fourier Grenoble I, Grenoble, France, philippe.gueguen@ujf-grenoble.fr
This study is focused on the correlation between ground motion and the buildings response. In particular, we interested in reducing the uncertainties that occur in vulnerability methods in order to assess more precisely the seismic vulnerability of existing structures. In order to do this, this work is based on the use of data recorded within buildings. Records of ambient vibrations and low amplitude earthquakes within buildings are used to characterize the elastic behaviour of these structures. Linear models can then be defined in order to study the various components of the variability of the fragility curves, for the first damage level. Earthquake that have occurred since the 1970s have been recorded within Californian buildings. From these data, we focused on the relationship between the building response and parameters describing the noxiousness of ground motion. Depending on the parameter used to represent the noxiousness of earthquakes, the building response can be estimated more or less accurately. In particular, parameters involving the dynamic parameters of the buildings are more related with the response of structures, which is represented by its averaged inter-story drift. Grouping buildings by typologies (defined according to their main material of construction and their height) can improve the variability in the response of buildings. Indeed, by providing more information on the structure, we can reduce the epistemic component of variability. In addition, by combining noxiousness parameters, the accuracy in the prediction of the building response can be improved. A functional form is thus proposed to estimate the averaged inter-story drift within the structures, for several typologies of buildings. This functional form is then used to assess fragility curves and can also be used to get an estimate of damage after an earthquake, by comparing the values of inter-story drift from given by the functional form with reference values (FEMA, 2003).
Session: Oceanographic and Atmospheric Signals in Seismology
Presenter   Sheen, Dong-Hoon
Schedule   Fri AM / Poster
Room   Hall 1
Origin of Ocean Swell and Source Location of Microseism by a Huge Indian Ocean Storm 2007
SHEEN, D., Department of Geological Environment, Chonnam Nat'l Univ., Gwangju, Republic of Korea
It was reported that there were huge ocean waves in the Indian Ocean in the middle of May, 2007 which caused damage to the islands in the Indian Ocean and the surrounding areas. Satellite observation and ocean wave model show that the ocean waves lasted several days in the Indian Ocean. Correspondingly, microseismic peaks were generated and consistently observed at most of seismic stations around Asia and Australia, which shows there were two ocean swell systems. Comparison of microseismic peaks at island seismic stations in the Indian Ocean and inland seismic stations shows that the peaks at inland stations are strongly correlated each other, despite their distant locations, while those at island and near-coastal stations are not. This implies that microseismic energy generated at near-coastal areas efficiently propagates through the continental crust, while strong attenuation in the oceanic crust obstructs the observation of microseism even at near-coastal area. Dispersion of microseism is also observed in the single-frequency and the double frequency bands because the ocean waves have dispersed from their generation area and traveled long distances before generating microseismic peaks. The origins of ocean swells are estimated by measurements of slopes of the dispersion trends. And the generation areas of microseisms are located by the polarization analysis at 3 component seismic stations and the FK analysis of KSRS array stations at Korea.
Session: New Developments in Earthquake Forecasting and Predictability Research
Presenter   Yoder, Mark
Schedule   Wed PM / Poster
Room   Hall 1
1/f and the Earthquake Problem: Applying Scaling based Constraints to Seismicity Models
YODER, M. R., Dept. of Physics, UC Davis, Davis, CA, mark.yoder@gmail.com; RUNDLE, J. B., Dept. of Physics, UC Davis, Davis CA; TURCOTTE, D. L., Dept. of Geology, UC Davis, Davis, CA
The difficulty of forecasting earthquakes can fundamentally be attributed to the self-similar, or “1/f”, nature of seismic sequences. Specifically, the rate of occurrence of earthquakes is inversely proportional to their magnitude m, or more accurately to their scalar moment m. With respect to this “1/f problem,” it can be argued that catalog selection (or equivalently, determining catalog constraints) constitutes the most significant challenge to seismicity based earthquake forecasting. Here, we address and introduce a potential solution to this most daunting problem. Specifically, we introduce a framework to constrain, or partition, an earthquake catalog (a study region) in order to resolve local seismicity. In particular, we combine Gutenberg-Richter (GR), rupture length, and Omori scaling with various empirical measurements to relate the size (spatial and temporal extents) of a study area (or bins within a study area), in combination with a metric to quantify rate trends in local seismicity, to the local earthquake magnitude potential – the magnitudes of earthquakes the region is expected to experience. From this, we introduce a new type of time dependent hazard map for which the tuning parameter space is nearly fully constrained. In a similar fashion, by combining scaling relations and also by incorporating finite rupture length, area, and duration as constraints, we develop a method to estimate the Omori (temporal) and spatial aftershock decay parameters as a function of the parent earthquake's magnitude m and the catalog completeness threshold mc. From this formulation, we develop an ETAS type model. These models demonstrate promise with respect to earthquake forecasting applications. Moreover, the methods employed suggest a general framework whereby earthquake and other complex-system, 1/f type, problems can be constrained from scaling relations and finite extents.
Session: Infrasound and Seismoacoustics
Presenter   Steele, Alexander
Schedule   WITHDRAWN
Room   Hall 1
STUDENT
Seismo-Acoustic Analysis of Explosion Events at Tungurahua Volcano: Comparison of Eruptive Episodes in May-July 2010 and December 2012
STEELE, A. L., Boise State University, Boise, ID, alex.steele1837@gmail.com; RUIZ, M. C., Escuela Politecnica Nacional, Quito, Ecuador, mruiz@igepn.edu.ec; JOHNSON, J. B., Boise State University, Boise, ID, jeffreybjohnson@boisestate.edu
We analyze a continuous seismo-acoustic dataset of explosions at Volcan Tungurahua, Ecuador, from July 2006 to the present. Reduced amplitudes and energies are calculated for discrete events using four collocated broadband seismic and infrasound sensors. Of the 12 explosive episodes recorded during this time, we highlight eruptive activity in May-July 2010 and December 2012 because of their similar characteristics and temporal evolution. During both periods of unrest, the following sequence is observed: Stage 1 – Activity commences after more than 3 months of quiescence. Stage 2 – Inflation and increasing VTs lead to the initiation of one high-energy explosion. Stage 3 – A few high-energy explosions are followed by strong tremor, column generation and pyroclastic flows. Stage 4 - Hundreds of daily explosions are accompanied by chugging signals lasting about one week. Stage 5 - A sudden end to chugging coincides with a rapid decline in explosion numbers and seismo-acoustic energy radiated.We quantify the volcano acoustic-seismic ratio (VASR: acoustic energy/seismic energy) from discrete explosions during the early stages of activity in May-July 2010 and December 2012 and find relatively low values (VASR = 0.1-10) when compared with other episodes of activity at Tungurahua (where VASR = 10-1000). The periods of low VASR are related to explosions accompanied by the strong chugging of Stage 4. We suggest that the first high-energy explosions during Stages 2 & 3 were unable to completely clear the old viscous magma plugging the vent. Subsequent explosions through a blocked, or constricted conduit, are likely to exhibit low VASR due to strong seismic coupling with the surrounding plug and conduit walls. Following continuing explosive activity, we envision that the plug is eventually cleared, as chugging events rapidly disappear. An increase in explosion VASR during Stage 5 is suggestive of a period of more efficient acoustic radiation during explosive degassing.
Session: Seismic Hazards and Ground Motions
Presenter   Adnan, Azlan
Schedule   Fri 1:30 PM / Oral
Room   155B
The Effects of Considering Different Probability Density Functions, Maximum Magnitudes, and Attenuation Laws in Seismic Hazard Analysis of Peninsular Malaysia Due to Distant Sumatran Subduction Earthquakes
ADNAN, A. B., e-SEER, University of Technology Malaysia, Johor Bahru, Malaysia, azelan_fka_utm@yahoo.com, azlanadnan@utm.my; VAEZ SHOUSHTARI, A., e-SEER, University of Technology Malaysia, Johor Bahru, Malaysia, abdollahvaezshoushtari@yahoo.com
Even though Peninsular Malaysia is located in a low-seismicity region, the structures may be vulnerable to distant Sumatran earthquakes. Several studies have been conducted to assess the seismic hazard of this region which produced widely different Peak Ground Acceleration (PGA) values. The key factors that significantly influence the PGA obtained from the Probabilistic Seismic Hazard Analysis (PSHA) are mainly; the selected Probability Density Functions (PDF), the maximum magnitudes (M) of seismic sources, and the attenuation laws. This study attempts to consider the effects of the above factors on PGA results of Kuala Lumpur (KL), the main city of Peninsular Malaysia, based on distant Sumatran subduction zone earthquakes. PSHA studies done by Pan and Megawati (2002) using PDF by Dong et al. (1984), and Hendriyawan (2006) utilizing bounded Gutenberg-Richter are simulated and compared. Based on the two different PDFs and other factors, the PGA on rock sites of KL are 55.1 gal and 149 gal for 2 percents probability of exceedance in 50 years, respectively. The different is about 150%. In other studies, based on the methodology by Hendriyawan (2006), changing the Mmax of Sumatran subduction sources at Benioff from M=8.5 to 9.5 and Megathrust from M=9.2 to 9.5, result in widely different values of PGA (92.5 gal and 144.91 gal). Among the three factors, the attenuation laws influence the most in terms of PGA values. Using PSHA method by Pan and Megawati (2002), for 2 percents probability of exceedance in 50 years, based on Fukushima and Tanaka (1992) attenuation law, the PGA value is about 54.9 gal. By changing the attenuation law to the recently derived, Megawati and Pan (2010), the PGA value is calculated as about 12.11 gal which is 400 percent less than the previous one. Thus, the selection of PDF, maximum magnitudes and expecially the attenuation laws are critical in evaluation of any seismic hazard map of a region, particularly Peninsular Malaysia.
Session: New Developments in Earthquake Forecasting and Predictability Research
Presenter   Yan, Wei
Schedule   Wed PM / Poster
Room   Hall 1
STUDENT
The Mechanism Analysis of the Guza Station Borehole Strain's Wave Phenomenon before the 2008 MS8 Wenchuan Earthquake, China
YAN, W., China Earthquake Networks Center, Beijing, China, ywpro@163.com; NIU, A. F., China Earthquake Networks Center, Beijing, China, nafcsb2004@seis.ac.cn; LIU, Q., Institute of earthquake prediction, CEA, Beijing, China, liu-qi1985@hotmail.com
Guza station is located in the cross zone of Longmen Mountain fault and the Xianshuihe fault. The distance between Guza station and the MS8 Wenchuan earthquake in 2008 is 200km. Shimian station is located in the southern Guza station, the distance is 90km, which is in the Xianshuihe fault. Guza station and Shimian station are near to the Dadu River. Before the 2008 Wenchuan earthquake, the borehole strain of Guza station and water level observation of Shimian station appeared wave phenomenon, the period of these wave are about 3-5 days. This paper is base on the Guza borehole strain data and drilling pressure, temperature, rainfall and the Shimian water lever observation data, using the method of regression analysis and plane strain model to explain the wave phenomenon of the Guza borehole strain before the Wenchuan earthquake. The results showed that: 1st,the four component borehole strain self- checking results of Guzha station is good, showed that the observed wave before the earthquake anomaly is the reflection of the crustal strain change; 2nd, the regression correlation between the Guza station borehole strain and the rainfall is larger, the pressure's correlation coefficient is small, indicated that the years periodic variation we observed maybe more affected by rainfall influence, less affected by pressure; 3rd,using the borehole strain of Guza station and the water level of Shimian station, we built a planar load strain model. the inversion calculation result showed that, the wave phenomenon of Guza station borehole strain before the earthquake is mainly caused by Dadu River water level changes. The relationship between the 2008 MS8 Wenchuan earthquake and the wave phenomenon is not significant.Fund support: Chinese Science and Technology Support Foundation (2012BAK19B02), Chinese Seismic Special Foundation(201108009).
Session: New Frontiers in Seismic Data Analysis
Presenter   Askan, Aysegul
Schedule   Thu AM / Poster
Room   Hall 1
Determination of the Site Characterization Properties in Eastern Segment of the North Anatolian Fault Zone in Turkey based on the MMSPAC Method
SISMAN, F. N., Middle East Technical University, Ankara, Turkey, f.nurtensisman@gmail.com; ASKAN, A., Middle East Technical University, Ankara, Turkey, aaskan@metu.edu.tr; ASTEN, M. W., Monash University, Victoria, Australia, michael.asten@monash.edu
Erzincan (a small city in Eastern Turkey) is located in the conjunction of three active faults: North Anatolian, North East Anatolian and East Anatolian Fault Zones. Erzincan city center is in a pull-apart basin underlain by soft sediments which significantly amplify the ground motions. Combination of the tectonic and geological behaviour of the region have led to destructive earthquakes such as the 27 December 1939 (Ms=8.0) and the 13 March 1992 (Mw=6.6) events resulting in extensive losses. In this study, in order to perform site characterization in the region, we employ the passive microtremor survey technique at nine sites on three sides of the Erzincan city center in order to gain quantitative estimates of 1D shear-wave velocity profiles for key parts of the basin. The sites are aligned as two profiles in the North-South and East-West directions allowing construction of approximate 2D profiles. At each site we performed surface wave dispersion curve analysis using the Multi-Mode Spatial Autocorrelation (MMSPAC) technique on either single or nested four-station arrays together with horizontal to vertical spectral ratio measurements at array centers. The combination allows resolution of Vs for typically the upper 4 m from high-frrequency data, and as deep as 480 m using low-frequency SPAC and HVSR data in combination. The survey showed that a rate of production of two sites per day is achievable. We present our results in the form of one and two dimensional velocity structures and fundamental frequencies obtained from the HVSR spectra, a combination suitable for input to earthquake ground-motion prediction studies
Session: Infrasound and Seismoacoustics
Presenter   Terbush, Brian
Schedule   Thu AM / Poster
Room   Hall 1
STUDENT
Rock Falls Recorded at Volcan Santiaguito (Guatemala) with Seismic and Infrasound Arrays
TERBUSH, B., Boise State University, Boise, ID, brianterbush@u.boisestate.edu; JOHNSON, J. B., Boise State University, Boise, ID, jeffreybjohnson@boisestate.edu
Volcan Santiaguito is an active dome in Guatemala characterized by small explosive eruptions every 0.5 – 2 hours, daily pyroclastic flows, frequent block-and-ash flows from the active lava flow, and rock falls, including many from the neighboring Santa Maria volcano edifice. This high level of activity makes Santiaguito an excellent laboratory for study of gravity flow events using seismic sensors and acoustic arrays. In January and November 2012, two different types of acoustic arrays (using two different types of low-frequency microphones) were used to record rock fall, block-and-ash and pyroclastic flows. A three-element acoustic array of MEMS sensors were deployed in January. Although the MEMS transducers possess a flat amplitude response in the band of interest (~50 s to 50 Hz), their signal-to-noise is inferior to the 8-element electret condenser microphones (ECMs) array that was deployed in November. Unfortunately the low frequency roll-off (~2 Hz) and limited dynamic range (+/- 20 Pa) of the ECMs preclude extensive waveform analysis. We have focused on the MEMS infrasound array recordings of January 2012 and corroborate analyzed signals with observations from a continuous time lapse camera recording at 5 s intervals. We locate events by conducting slowness searches for candidate sources that might be located on the surface of the Santiaguito dome edifice. We find that the MEMS array is capable of effectively locating the moving sources of sound (i.e., block-and-ash and rock fall) for most visible rock falls. A co-located broadband seismic sensor also detected rock fall events, but the single sensor was incapable of effective localization.
Session: New Frontiers in Seismic Data Analysis
Presenter   Hrubcova, Pavla
Schedule   Wed 2:30 PM / Oral
Room   155A
Determination of Prominent Crustal Discontinuities from Waveforms of Local Earthquakes
HRUBCOVA, P., Institute of Geophysics, Prague, Czech Republic, pavla@ig.cas.cz; VAVRYCUK, V., Institute of Geophysics, Prague, Czech Republic, vv@ig.cas.cz; BOUSKOVA, A., Institute of Geophysics, Prague, Czech Republic, ab@ig.cas.cz; HORALEK, J., Institute of Geophysics, Prague, Czech Republic, jhr@ig.cas.cz
The West Bohemia/Vogtland area is known for increased geodynamic activity with reoccurrence of intraplate earthquake swarms. Previous geophysical studies, namely active and passive seismic investigations, revealed a high velocity lower crust in this area with increased reflectivity. To refine this result and retrieve a more detailed structure of the deep crust and the Moho discontinuity we analyzed waveforms of local microearthquakes that occurred in this area during the 2008 swarm.We developed a new multi-azimuthal approach in data processing to increase resolution of Moho phases in the waveforms. The waveforms typically display dominant direct P and S waves followed by converted and reflected waves secondarily generated at shallow and deep subsurface structure. Apart from the velocity structure and the source-receiver geometry, the waveforms are significantly affected by focal mechanisms of the earthquakes. Thus, the waveforms were grouped into clusters with similar focal mechanisms and clusters were processed separately. We applied the waveform cross-correlation, and rotated, aligned and stacked the seismograms to extract the Moho SmS, PmP, and PmS reflected/converted phases. These phases were inverted for laterally varying Moho depth by ray tracing and a grid search inversion algorithm. The model retrieved was verified using modeling of full waveforms computed by the discrete wave number method.The applied multi-azimuthal approach revealed details in the velocity structure of the crust/mantle transition at each station. Instead of a single interface with a sharp velocity contrast, the inversion indicates a reflective zone at Moho depths with one or two strongly reflective interfaces, which is in agreement with the zone interpreted by previous investigations. The thickness of the zone varies from 2-4 km within the depth range of 27-31.5 km and is delimited by reflections from its top and bottom boundaries, sometimes with strong reflectors within the zone.
Session: Characterizing Active Faults for Seismic Hazard Assessments
Presenter   Gulerce, Zeynep
Schedule   Thu 11:45 AM / Oral
Room   155B
Missing Fault Characteristics of North Anatolian Fault for Probabilistic Seismic Hazard Assessment
GULERCE, Z., Middle East Technical University, Ankara, Turkey, zyilmaz@metu.edu.tr; LEVENDOGLU, M., Middle East Technical University, Ankara, Turkey, mertlevendoglu@gmail.com; AGBAY, C. B., Middle East Technical University, Ankara, Turkey, canbarisagbay@gmail.com; SAROGLU, F., Kayen Kayı Energy Investments Inc., Ankara, Turkey, fsaroglu@gmail.com
North Anatolian Fault (NAF) system is one of the well-studied fault systems in Turkey, however fault characteristics and parameters that are crucial for a complete probabilistic seismic hazard assessment (PSHA) is still missing for many segments of this fault. The objective of this study is to build the seismic source characterization models for the rupture zones of 1942-1944 earthquake sequence to be used in the PSHA framework. NAF system ruptured progressively by eight large and destructive earthquakes in the last century; 1942 Erbaa-Niksar, 1943 Tosya, 1944 Bolu-Gerede earthquakes had broken approximately 500 kilometers of a uniform eastern trace. Linear fault segments are defined for each rupture zone, geometry of the sub-segments (length, width, and segmentation points) are determined and incorporated with the help of available fault maps. Composite magnitude distribution model (Youngs and Coppersmith, 1985) is used for all seismic sources in the region to properly represent the characteristic behavior of NAF without an additional background zone. Fault segments, rupture sources, rupture scenarios and fault rupture models are determined using the WG-2003 terminology and multi-segment rupture scenarios are considered. Events in the earthquake catalogue are attributed to the individual seismic sources and scenario weights are determined by balancing the accumulated seismic energy by the catalog seismicity on each source. Uncertainties involved in each parameter, especially the slip rate, annual creep rate, fault widths, b-value and scenario weights are considered. Effects of these uncertainties on the output hazard curve are presented to reveal the important missing parameters of NAF system that requires further research in the future.
Session: Induced Seismicity
Presenter   Vavrycuk, Vaclav
Schedule   Fri 4:15 PM / Oral
Room   155D
Triggering and Seismic Cycle of Earthquake Swarms in Geothermal Areas
VAVRYCUK, V., Institute of Geophysics, Prague, Czech Republic
The occurrence and specific properties of earthquake swarms in geothermal areas are usually attributed to a highly fractured rock and/or heterogeneous stress within the rock mass being triggered by magmatic or hydrothermal fluid intrusion. The increase of fluid pressure destabilizes the fractures and causes their opening and a subsequent shear-tensile rupture. The spreading and evolution of the seismic activity is controlled by fluid flow due to diffusion in a permeable rock and/or by the redistribution of Coulomb stress. The ‘fluid-injection model’, however, is not universal. We provide evidence that this model is inconsistent with observations of earthquake swarms in West Bohemia, Czech Republic. Full seismic moment tensors of micro-earthquakes in the 1997 and 2008 swarms in West Bohemia indicate that fracturing at the starting phase of the swarm was not associated with fault openings but rather with fault compactions. This can physically be explained by a ‘fluid-erosion model’, when the essential role in the swarm triggering is attributed to chemical and hydrothermal fluid-rock interactions in the focal zone. Since the rock is exposed to circulating hydrothermal, CO2-saturated fluids, the walls of fractures are weakened by dissolving and altering various minerals. If fault strength lowers to a critical value, the seismicity is triggered. The fractures are compacted during failure, the fault strength recovers and a new cycle begins.
Session: Characterizing Active Faults for Seismic Hazard Assessments
Presenter   Reicherter, Klaus
Schedule   Fri AM / Poster
Room   Hall 1
High Resolution t-LiDAR Scanning of an Active Bedrock Fault Scarp for Paleostress Analysis
REICHERTER, K. R., Neotectonics and Natural Hazards, RWTH Aachen University, Aachen, Germany, k.reicherter@nug.rwth-aachen.de; WIATR, T., Neotectonics and Natural Hazards, RWTH Aachen University, Aachen, Germany; PAPANIKOLAOU, I. D., Lab. of Min. & Geol., Dept. of Sci., Agricultural Univ. Athens, Athens, Greece; FERNÁNDEZ-STEEGER, T. M., Dept. of Engineering Geology and Hydrogeology, Aachen, Germany
Paleostress analysis of an active bedrock normal fault scarp based on kinematic indicators is carried out applying terrestrial laser scanning (t-LiDAR). For this purpose three key elements are necessary for a defined region on the fault plane: (i) the orientation of the fault plane, (ii) the orientation of the slickenside lineation or other kinematic indicators and (iii) the sense of motion of the hanging wall.We present a workflow to obtain palaeostress data from point cloud data using t-LiDAR. The entire case-study was performed on a continuous limestone bedrock normal fault scarp on Crete, Greece, at four different locations along the WNW-ESE striking Spili fault. At each location we collected data with a mobile t-LiDAR and validated the calculated 3-D paleostress results by comparison with the conventional palaeostress method with compass. Numerous kinematics indicators for normal faulting were discovered on the fault plane surface using t-LiDAR data and traditional methods, like Riedel shears, extensional break-outs, polished corrugations and many more. However, the kinematic indicators are more or less unidirectional and almost pure dip-slip. But, towards the tips of the fault, inclination of the striation tends to point towards the centre of the fault.When comparing all reconstructed paleostress data obtained from t-LiDAR to that obtained through manualcompass measurements, the degree of fault plane orientation divergence is around 005/03 (dip direction, dip). The degree of slickenside lineation variation is around 003/03 (trend and plunge). Therefore, the percentage threshold error of the individual vector angle at the different investigation site is lower than 3 % for the dip direction and dip for planes, and lower than 6 % for strike. The maximum mean variation of the complete calculate paleostress tensors is 005/03. t-LiDAR measurements are in the error range of conventional compass data. Remote paleostress analysis is possible with t-LiDAR.
Session: Advances in High-Frequency Physics-Based Earthquake Systems Simulation
Presenter   Olsen, Kim
Schedule   Thu 11:00 AM / Oral
Room   155C
Large-scale 3-D Simulations of Spontaneous Rupture and Wave Propagation in Complex, Nonlinear Media
ROTEN, D., ETH Zürich, Zürich, Switzerland, daniel.roten@sed.ethz.ch; OLSEN, K. B., San Diego State University, San Diego, CA, kbolsen@mail.sdsu.edu; DAY, S. M., San Diego State University, San Diego, CA, sday@mail.sdsu.edu; DALGUER, L. A., ETH Zürich, Zürich, Switzerland, dalguer@sed.ethz.ch; FÄH, D., ETH Zürich, Zürich, Switzerland, donat.faeh@sed.ethz.ch
A major challenge in seismic hazard assessment consists in the prediction of near-source ground motions resulting from large, rare earthquakes, which are not well represented in observed data. Realistic simulation of both the rupture dynamics and wave propagation during such events should be based on rheology models which are able to describe plastic yielding on and off the fault as well as nonlinear damping in soils near the surface. Towards this goal we have implemented the Drucker-Prager yield condition based on the return map algorithm in the highly scalable AWP-ODC finite difference code. We introduce a simple optimization method to reduce the number of stress interpolations needed for the staggered grid, which limits the extra computational cost of modeling plasticity to ~60% as compared to the corresponding elastic simulation. The implementation of Drucker-Prager plasticity is verified against four finite element codes in the framework of the SCEC/USGS spontaneous rupture code verification exercise (Harris et al., 2009). We apply the newly developed code to study how off-fault plasticity and nonlinear soil behavior affects ground motions during large scenario earthquakes. In particular, we use the Shakeout earthquake scenario using a kinematic source description (Graves et al., 2008) to study the effects on the resultant ground motion. We find that long-period ground motions in the downtown Los Angeles area, amplified by a wave guide of interconnected sedimentary basins, could be significantly reduced as compared to visco-elastic solutions, assuming that cohesions range from a few tens of kPa in soft soils near the surface to several MPa in deep bedrock. These results indicate that more research is warranted into the dynamics of rupture and wave propagation in nonlinear media. They also suggest a need to improve calibration of the additional parameters required for yielding models (e.g., friction angle and cohesion) from geotechnical and geophysical data.
Session: Velocity Models and Modeling
Presenter   Gao, Rui
Schedule   Thu AM / Poster
Room   Hall 1
Crustal Structure of the SW Margin of the Ordos Block from Wide Angle Seismic Data
GAO, R., Lithosphere Research Centre, Institute of Geology, CAGS, Beijing, China, gaorui@cags.ac.cn; KELLER, G. R., The University of Oklahoma, Norman, OK; LI, Q., Lithosphere Research Centre, Institute of Geology, CAGS, Beijing, China, liqiusheng@cags.ac.cn; LI, W. H., Lithosphere Research Centre, Institute of Geology, CAGS, Beijing, China; LI, Y. K., Geological Cores and Samples Center, Yanjiao, Hebei, China; WANG, H. Y., Lithosphere Research Centre, Institute of Geology, CAGS, Beijing, China; HOU, H. S., Lithosphere Research Centre, Institute of Geology, CAGS, Beijing, China; GUO, X. Y., Lithosphere Research Centre, Institute of Geology, CAGS, Beijing, China; XIAO, X., The University of Oklahoma, Norman OK
In 2012, a 410 km long seismic WAR/R profile was recorded jointly by the CAGS and the University of Oklahoma. Simultaneously, deep seismic reflection data were recorded along the same profile (165 km). The profile extended from the Ordos block, across the Liupanshan area to the Qilian orogen. The recording of seismic waves from 14 explosions was conducted in 2 deployments of 200 DFZ 3-components recorders with station spacing of 2-2.5 km. The P wave field on the sections provided good quality data for most of the profile. Arrivals from of refracted and reflected waves from sediments and basement (Pg), intracrustal phases (P1, P2, P3), the Moho (PmP), and reflector from mantel were typically observed.Hole’s 3D tomography program was used for first arrival upper crustal imaging. 2D P-wave velocity modeling was done layer by layer using the top to bottom approach. The velocity model was altered by trial and error, and the forward model was updated by damped least-squares inversion. In our modeling, calculated travel times fit observed arrivals for all 940 traces with average RMS of 0.117. Synthetic waveform modeling employing the Tramp program was used to constrain the velocity gradients within the model layers.The P wave velocity model shows: 1) The Qilian orogen and the Ordos show different characteristics in the first arrival topographic model, and indicate that the Haiyuan-Liupanshan fault is the boundary between them. 3) The Qilian orogen has relatively lower velocity crust, while the Ordos has higher velocity and relative simple crust. 3) The crustal thickness varies from ~50 km in the west to ~42 km in the east, and double Moho appears beneath the Liupanshan area. It suggest that the thicker Qilian crust pushed under the stable Ordos crust with far field effect of the collision between Indian plate and the Eurasian plate.Funded by NSF PIRE (0730154), Sinoprobe02 and China Geological Survey project (1212011220260)
Session: Characterizing Active Faults for Seismic Hazard Assessments
Presenter   Landgraf, Angela
Schedule   Fri AM / Poster
Room   Hall 1
Reverse-Fault Zone Topography Revealed by Combined Trenching and Terrestrial LiDAR Investigations
LANDGRAF, A., Inst. of Earth- a. Environmental Sciences, University of Potsdam, Potsdam, Germany, landgraf@geo.uni-potsdam.de; DJUMABAJEVA, A., Kyrgyz Institute of Seismology, Bishkek, Kyrgyzstan, arose3.60@mail.ru; ABDRAKHMATOV, K., Kyrgyz Institute of Seismology, Bishkek, Kyrgyzstan, kanab53@rambler.ru; ARROWSMITH, J. R., School of Earth and Space Exploration, ASU, Tempe, AZ, ramon.arrowsmith@asu.edu; STRECKER, M. R., Inst. of Earth- a. Environmental Sciences, University of Potsdam, Potsdam, Germany, strecker@geo.uni-potsdam.de
Fault scarps are the geomorphic manifestation of earthquake ruptures and might testify to the underlying near-surface processes. Especially thrust and reverse fault scarps are highly variable in their along-strike morphology, even if resulting from a single event. Complexities that affect morphology include hanging-wall collapse, backthrusting, folding or secondary far-field deformation. These impacts make any evaluation of fault-slip estimates and related earthquake parameters difficult. Here, we document high-resolution fault scarp topography, derived from a terrestrial LiDAR survey of a young reverse fault in the northern Tien Shan of Kyrgyzstan. This region was subjected to a series of major earthquakes during the late 19th and earliest 20th centuries: 1885 (Ms 6.9), 1887 (Ms 7.3), 1889 (Ms 8.3), and 1911 (Ms 8.1). The study area is located 35 km west of Belovodskoie, the epicenter of the 1885 earthquake, and about 75 km west of Bishkek, the Kyrgyz capital. Two fault scarps have formed in an alluvial fan, nested inside a late Pleistocene loess terrace. The northern, more prominent, E-W striking scarp, suggests a dominant reverse faulting mechanism, and can be followed for about 4 km across the alluvial fan. Distinct offsets between about 0.6 m and 3.8 m can be identified and the scarp location is aligned with a cumulative break in topography of about 13 m in the loess-covered surface to the west. Three trenches were excavated in the low-offset and intermediate-offset segments of this scarp. In all trenches, dip-slip motion was recognized as the dominant faulting process, resulting in hanging-wall collapse scarps with the deposition of colluvial wedges. The combination of our trenching results and the high-resolution surveys help to better constrain fault-slip parameters, but also define regional rates of erosional fault-scarp degradation. We use diffusion-age modeling to determine the age of the scarp and thus the responsible earthquake event.
Session: Earthquake Source Physics
Presenter   Perry, Stephen
Schedule   Thu PM / Poster
Room   Hall 1
STUDENT
Exploring Breakdown Energy in Simulations of Earthquake Sequences
PERRY, S., California Institute of Technology, Pasadena, CA, sperry@caltech.edu; LAPUSTA, N., California Institute of Technology, Pasadena, CA, lapusta@caltech.edu
The energy budget of earthquakes is a question of significant fundamental and practical interest. Using rate-and-state fault models that produce earthquake sequences, we focus on exploring the breakdown energy portion G of this budget and its relation to the characteristic slip distance L of rate-and-state friction, a parameter often selected based on numerical tractability. We explore values of L ranging two orders of magnitude and calculate fracture energy for events with various amounts of slip. We compare our results to calculations of G for natural earthquakes and find our computed values to be reasonable. However, we find nearly-constant values of G across a range of slips for a given L, as well as decreasing G with smaller values of L, as expected based on prior studies. Hence, simulations with laboratory-like values of L (0.01-0.1 mm), necessary for producing microseismicity, would result in breakdown energies too small for large events, compared with observations. We will also report on our current work, which is directed towards building models with variable L. Such models would aim to approximately match the observed increase of the breakdown work with slip and should allow, in a single model, for nucleation of small events while reproducing the large breakdown energies found in large events.
Session: Realtime Monitoring and Early Warning
Presenter   Guilhem, Aurelie
Schedule   Fri 2:15 PM / Oral
Room   155A
GRiDMT in Northern California: Preliminary Results of Continuous Detection and Characterization of Earthquakes Using Moment Tensors
GUILHEM, A., ETH Zurich, Zurich, Switzerland, aurelie.guilhem@sed.ethz.ch; DREGER, D. S., Berkeley Seismological Laboratory, Berkeley, CA, dreger@seismo.berkeley.edu; LOMBARD, P., Berkeley Seismological Laboratory, Berkeley, CA, lombard@seismo.berkeley.edu
The large earthquakes that occurred in the past decade have demonstrated that the rapid characterization of seismic events is critical especially when these events are capable of potentially generating damages and tsunamis. In the US, the western coast corresponds to the place, where damaging subduction zone earthquakes (i.e., up to M9) have occurred and where the seismic hazards are estimated to be high. In the southern end of the Cascadia Subduction Zone (CSZ) the Mendocino Triple Junction is the most seismically active region of Northern California with a large variety of anomalous seismic events such as low-frequency and repeating earthquakes, and tremors that occur in addition to the inter- and intra-plate seismic activity. Because it also corresponds to the section of the CSZ that most repeatedly ruptured according to geological surveys, an efficient monitoring of its activity is required.We implemented a continuous scanning of long-period realtime seismic records based on the method proposed by Kawakatsu (1998). This approach consists in the computation of moment tensors every two seconds over a 3D grid of points distributed over the MTJ. This method allows the automated detection, localization, and moment tensor characterization of M3.5+ earthquakes. Moreover, GRiDMT also offers the possibility to rapidly detect and correctly estimate the moment magnitude of megathrust earthquakes within a few minutes of their occurrences.We present the first results of the approach obtained in realtime, and we show that in addition of detecting regular earthquakes, GRiDMT also detects low-frequency earthquakes. This constitutes an important outcome as it helps highlight the unknown anomalous seismicity of the region. We also report on the implementation for the detection of M8-9 earthquakes along the CSZ, which would offer a faster and more reliable detection than the current in-cascade procedures and result in a particularly important step toward tsunami early-warning.
Session: Seismic Hazards and Ground Motions
Presenter   Sandikkaya, M. Abdullah
Schedule   Fri 2:30 PM / Oral
Room   155B
STUDENT
Evaluation of the Eurocode 8 and the NEHRP Site Amplification Factors
SANDIKKAYA, M. A., ISTerre, Universite de Grenoble, Grenoble, France, also at Dept. of Civil Engineering, Middle East Technical University, Ankara, Turkey, asandikkaya@hotmail.com; AKKAR, S., Dept. of Civil Engineering, Middle East Technical University, Ankara, Turkey, sakkar@metu.edu.tr; BARD, P. Y., ISTerre, Universite de Grenoble, Grenoble, France, pierre-yves.bard@ujf-grenoble.fr
Sandıkkaya et al. (2013) developed an empirical nonlinear site model that makes use of time based average of uppermost 30m shear-wave velocity profile (VS30) and peak ground acceleration at rock site (PGArock at VS30 = 750 m/s). The proposed site model resulted in compatible estimations with the site models of recent global predictive equations. The discussions on the NEHRP (BSSC, 2009) and the Eurocode 8 (CEN, 2004) site factors showed that the site model is not only valid in application to ground motion prediction equations but also could be used for developing site factors in seismic design codes. These discussions were based on the reference rock model that developed with a global database in the article. In order to decrease the possible regional variability in the reference rock model, the recently developed pan-European predictive equation (Akkar et al., 2013) is used to compute PGArock levels for a set of earthquake scenarios. Then the Sandıkkaya et al. (2013) site model is applied to compute the site amplification factors. Accordingly, both the NEHRP and the Eurocode 8 site factors are elaborately discussed with the findings of this study. The immediate observation is that the site amplification factors are strictly depended on period. Although the NEHRP considers this effect, the Eurocode 8 does not. The short- and long-period amplifications are distinct and become more apparent as sites get softer. When compared to the NEHRP site factors, the results are comparable. We find higher amplification factors, imposing less soil nonlinear behavior at short-period. On contrary to other site classes, the long-period amplification factors are found to be lower at stiff sites.
Session: Towards an Integrated Understanding of Slow Earthquakes: What We Know, What We Don’t Know, and How to Move Forward
Presenter   Koketsu, Kazuki
Schedule   Fri AM / Poster
Room   Hall 1
A Very-Long-Term Slow Slip Event Triggering the 2011 Tohoku Earthquake
YOKOTA, Y., Earthquake Research Institute, University of Tokyo, Tokyo, Japan, yusuke@eri.u-tokyo.ac.jp; KOKETSU, K., Earthquake Research Institute, University of Tokyo, Tokyo, Japan, koketsu@eri.u-tokyo.ac.jp
Several types of slow seismic events, such as episodic tremor, short-term slow slip event (SSE), and long-term SSE, have been discovered in various subduction zones, and they have long been believed to be key phenomena associated with megathrust earthquakes. The 2011 Mw 9.0 Tohoku earthquake, the latest megathrust earthquake, occurred in one of the most active subduction zones in the world, the Japan Trench subduction zone (JTSZ). However, except for slip deficits (drag of the overriding plate due to interplate coupling), no slow seismic event had been reported in the JTSZ before the Tohoku earthquake. Here we show that a large SSE with a much longer duration than any previously known type of slow event (very-long-term SSE) occurred in the JTSZ, triggering the Tohoku earthquake. We first assumed interseismic phenomena, recording on continuously observing GPS stations, to be composed only of slip deficits, and recovered the distributions of their annual rates through normal-faulting source inversions. A reverse-faulting source inversion was also performed for the coseismic slip distribution of the Tohoku earthquake. The results of these inversions show a weakening and migration of slip deficit rates in late 2002 and that their distributions then coincided with the coseismic slip distribution. The weakening and migration could have been caused by a very-long-term SSE lasting eight and a half years from late 2002. The coinciding distributions indicate that this very-long-term SSE triggered the Tohoku earthquake.
Session: New Developments in Earthquake Forecasting and Predictability Research
Presenter   Tormann, Thessa
Schedule   Thu 9:00 AM / Oral
Room   155D
Towards Global b-Value-Based Earthquake Forecasting: Tackling Subduction Zones and Their Megathrust Asperities
TORMANN, T., ETH Zurich, Zurich, Switzerland, thessa.tormann@sed.ethz.ch; WIEMER, S., ETH Zurich, Zurich, Switzerland, stefan.wiemer@sed.ethz.ch; ENESCU, B., University of Tsukuba, Tsukuba, Japan, benescu@geol.tsukuba.ac.jp; WOESSNER, J., ETH Zurich, Zurich, Switzerland, jochen.woessner@sed.ethz.ch
One promising candidate of mainshock forecasting models in the first completed 5-year test round in California was the Asperity-based-Likelihood Model (ALM), which uses local variation in the frequency-magnitude distribution (b-value) of earthquakes to forecast the rates of the target events. The underlying physical hypothesis is an observed relationship between the locally measured b-values and the relative stress levels: low b-values indicate currently highly stressed areas where large events are more likely to occur in the near future. The ALM hypothesis is currently also being tested in the testing regions Italy and Japan, and a modified version has been prepared for California (ALM2). Here we investigate how to move towards testing ALM at a global scale, which will allow drawing conclusions on the predictive skill of the hypothesis within a shorter period due to the higher rate of events, and also within different tectonic regimes. In a first step, we investigate how to apply the local b-value variation hypothesis and its analysis techniques to identify potential megathrust asperities in subduction zones. We analyze case studies with variable data quality and completeness from selected regions around the world, including Japan, Greece and Alaska. Preliminary results for Japan show that indeed asperities on subduction zones can be clearly imaged. We then investigate strategies to convert this information into a global seismicity forecasting model. Since in some places, e.g. around Loma Prieta in Central California, significant temporal variation of b-values is observed, we explore in a final step approaches to systematically account for reliably measured b-value variation through time to potentially improve the monitoring of time-dependent seismic hazard.
Session: Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Models
Presenter   Dutta, Utpal
Schedule   Fri 2:15 PM / Oral
Room   155C
Estimation of Spectral Acceleration Based Site Amplification in Anchorage Basin, Alaska
DUTTA, U., University of Alaska, Anchorage, AK, udutta2@uaa.alaska.edu
The amplification of the spectral acceleration (SA) is one of the important parameters for the assessment of seismic hazards. The spatial variations of amplification of spectral acceleration have been estimated for the Anchorage basin using the data recorded from 40 free-field station sites. These stations provide adequate seismic coverage for various lithological units of the basin and are equipped with Kinemetrics Altus K2 accelerographs operated with a data-sampling rate of 200 Hz. One site located on the metamorphic terrain of Chugach Mts. on the eastern side of the basin has been selected as the reference rock site. The recorded events provide good azimuthal coverage of the area and are of magnitude (ML) in the range of 2.6 to 5.5 located at depths from 10-150 km with the peak acceleration around 80 Gal. The 5% damped spectral acceleration of the recorded data from each site was determined from 0.04 to 15 sec period. Amplification of the SA values was computed by standard spectral ratio (SSR) as well as horizontal-to vertical ratio (HVR) method assuming unit response at the reference station site in the above period range of interest. The estimated amplifications values at each site were band averaged with the central period at 0.2 sec (Short Period Band) and 1.0 sec (Long Period Band). The band-averaged site amplification values were correlated with the available 30 m shear wave velocity values (β30) of the basin. A strong correlation exists between the long period site amplification and β30 values. However, for short period, the amplification is nearly constant with the β30 values. Two SA based site amplification maps at 0.2 sec and 1.0 seconds has been prepared for the area to capture the spatial variation of the ground motions characteristics that will be directly useful for the earthquake engineering analysis in future.
Session: Seismic Hazards and Ground Motions
Presenter   Kim, Kwang-Hee
Schedule   Fri AM / Poster
Room   Hall 1
Microseismicity in the Seoul Metropolitan Area, Korea, and Its Implications for the Seismic Hazards
KIM, K. H., Korea Institute of Ocean Sciences & Technology,, Ansan, Gyeonggi-do, Korea, kwanghee@kiost.ac.kr; KANG, S., Korea Institute of Ocean Sciences & Technology,, Ansan, Gyeonggi-do, Korea, sukang@kiost.ac; KIM, W. Y., Lamont Doherty Earth Observatory, Columbia University, Palisades, NY, wykim@ldeo.columbia.edu; RYOO, Y., Korea Meteorological Administration, Seoul, Korea, ryooyg@korea.kr; PARK, Y., Korea Polar Research Institute, Incheon, Korea, ypark@kopri.re.kr
An earthquake with magnitude 3.0 occurred in the Seoul Metropolitan Capital Area (SNCA), Korea, on 9 February 2010. The earthquake attracted much attention and raised concerns about seismic hazards and risks in the Korea Peninsula, in particular, to the SNCA. SNCA includes the Seoul and Incheon metropolitans and most of the Gyeonggi province. It has a population of 24.5 million (as of 2007) and is one of the largest metropolitan areas in the world. We applied waveform correlation detector to 2007-2011 continuously recorded seismic data to identify repeating earthquakes. We identify 9 micro-earthquakes during 2007-2010 periods which are not reported in the KNSN bulletin because their magnitudes are too small. Estimated magnitudes using amplitude ratios measured at the station SEO indicate the smallest event detected by the waveform cross correlation technique in the study is as low as 0.19. The number of events for our interpretation becomes 11 including 2 previously reported events and 9 newly identified ones. All of them occur in a very small area. While there are historic documents reporting earthquakes in the SNCA, repeating earthquakes or clustered seismicity from the instrumental earthquake record have not reported before. We have determined the focal mechanism solution for the representative events (9 Feb. 2010, Ml 3.0) using first motions. The preferred focal mechanism solution for the representative event is the WNW-ESE striking fault, which are consistent with the precisely determined earthquake hypocenter distribution. It is also consistent with the results in the previous studies of stress orientation in and around the Korean peninsula. The new list of earthquakes in this study is far from any complete, although we have adopted a well-established method to detect earthquakes. Considering the low seismicity, extensive efforts to monitor the micro-seismicity are definitely required to obtain comprehensive picture of the seismicity pattern in the area.
Session: What are the limits of explosion source model predictions?
Presenter   Oldham, Harrison
Schedule   Thu AM / Poster
Room   Hall 1
STUDENT
Near-Surface Velocity Modeling Using Body Wave Data, Surface Wave Dispersion Data and Short Period Ambient Noise Cross-Correlation Data
OLDHAM, H. R., Boston University, Boston, MA, remioldham@gmail.com; FERRIS, A., Weston Geophysical Corp., Lexington, MA; DALTON, C. A., Boston University, Boston, MA
A series of measurements were made at Rock of Ages Quarry, in Vermont, to determine the near-surface velocity profile in that region. Two lines of geophones and short-period seismometers were deployed at 90° to each other, extended out to approximately 35 km. Data from ballistic sources were analyzed to determine body-wave velocities, surface layer depth, and surface-wave dispersion curves. Ambient noise was processed to obtain, based on cross-correlations between all pairs of sensors, independent estimates of surface wave dispersion. Body wave data were used to make a simple single layer model. Surface wave ballistic data were then used to create a more refined model, based on Rayleigh wave group velocity measurements. Group velocity measurements derived from ambient noise were then used to make a further refined model. The inversion of group velocity measurements was done by means of constrained least squares as implemented by the Saint Louis University “Computer Programs in Seismology” software package. The initial model, obtained from body wave data, was a single 1.4 km layer with Vp = 6 km/s and Vs = 3.5 km/s. From the shot data group velocity we found a low VR velocity zone at 2 km depth. This result was confirmed by the noise-correlation analysis. The integration of multiple data types and techniques to develop a seismic model results in a better understanding of the local structure than would be obtained by any of these techniques independently.
Session: Earthquake Source Physics
Presenter   Kwiatek, Grzegorz
Schedule   Thu PM / Poster
Room   Hall 1
An Improved Method for Seismic Moment Tensor Inversion of Acoustic Emissions: The Effect of Incidence Angle, Sensor Coupling and Damage Evolution
KWIATEK, G., GFZ German Research Centre for Geosciences, Potsdam, Germany, kwiatek@gfz-potsdam.de; CHARALAMPIDOU, E. M., GFZ German Research Centre for Geosciences, Potsdam, Germany, elmachar@gfz-potsdam.de; DRESEN, G., GFZ German Research Centre for Geosciences, Potsdam, Germany; STANCHITS, S., TerraTek, a Schlumberger company, Salt Lake City, UT
Moment Tensor (MT) inversion is a standard tool allowing estimation of fault plane parameters as well as the amount of shear/volumetric change in the earthquake source. The inversion is sensitive to the quality of waveforms, sensor calibration and coupling. The application of MT inversion to laboratory deformed rocks is difficult since coupling of the Acoustic Emission (AE) sensors is not easily assessed at experimental conditions in a pressure vessel. In this study we present a procedure for the relative calibration of AE sensors and we analyze source processes of AE events recorded during two experiments on Bentheim and Vosges sandstones. We recorded AE activity and performed velocity measurements sending ultrasonic transmission (UT) pulses. The first P-wave amplitudes of UTs were used to estimate the coupling quality of the AE sensors and sensor sensitivity as a function of incident angle for a specific frequency band. We found that the quality of MTs improved significantly when data were corrected for coupling. We compared the developed procedure with the hybrid MT technique. Depending on sample and loading conditions, we observed significant compaction and tensile-shear-compaction components in the MTs of AE events recorded in tests performed on Bentheim and Vosges sandstones, respectively. The results shed light on the details of the rupture process in femtoscale.
Session: When and Why do Earthquake Ruptures Stop? Evaluating Competing Mechanisms of Rupture Termination
Presenter   Bonini, Lorenzo
Schedule   Fri PM / Poster
Room   Hall 1
Downdip Segmentation of Active Faults as a Key to Understand Hidden Earthquakes
BONINI, L., Universtà di Pavia, Pavia, Italy, lorenzo.bonini@unipv.it; DI BUCCI, D., 2Dipartimento della Protezione Civile, Rome, Italy, Daniela.DiBucci@protezionecivile.it; TOSCANI, G., Universtà di Pavia, Pavia, Italy, giovanni.toscani@unipv.it; SENO, S., Università di Pavia, Pavia, Italy, silvio.seno@unipv.it; VALENSISE, G., Istituto Nazionale di Geofisica e Vulcanologia, Rome, Italy, gianluca.valensise@ingv.it
Documenting seismogenic slip in relation with sub-surface and surface faults is one of the most challenging topic in earthquake studies. Best documented earthquakes worldwide showed a complex interaction between active faults with a tridimensional seismogenic fault segmentations. The well-documented 2009 L’Aquila earthquake affords the opportunity to observe the interaction of seismogenic ruptures with brittle crust heterogeneities. High-resolution geologic, geodetic and seismological available data allow to reconstruct with an unprecedented accuracy the geometry of the activated faults and to understand the relationships between sub-surface faults and surface deformations. We demonstrate that the earthquake was caused by a coseismic rupture strongly controlled at depth by pre-existing discontinuities and expressed at the surface by pseudo-primary breaks and by the reactivation of subsidiary faults resulting from coseismic and post-seismic phases. A comparison of this case with other earthquakes suggests that the structural complexity at depth is a common feature of hidden earthquakes, independently from the geodynamic contest. Failure to appreciate such complexity may result in severe over- or under-estimation of the local seismogenic potential.
Session: New Frontiers in Seismic Data Analysis
Presenter   Sebe, Olivier
Schedule   Wed 3:15 PM / Oral
Room   155A
Blind Deconvolution of the Seismic Source Time Function Based on Higher Order Statistics of Regional Coda Waves
SÈBE, O. G. R., CEA, DAM, DIF, Arpajon, France, olivier.sebe@cea.fr; GUILBERT, J., CEA, DAM, DIF, Arpajon, France, jocelyn.guilbert@cea.fr; BARD, P. Y., Institut des Sciences de la Terre, Université Joseph Fourier, Grenoble, France, pierre-yves.bard@ujf-grenoble.fr
At regional distance, the assessment of the source time function of moderate magnitude events is often a difficult task due to bad knowledge of the Source-station seismic Green’s function. Indeed the scattering effect on heterogeneities in the lithosphere generates a strong variability in space and time of regional wavefield. Several studies have already demonstrated the capability of the regional coda wavefield to provide valuable information on the seismic source. Nevertheless, all the current methods are based of the use of second order statistics providing source power spectral density without any information on the phase and thus the STF. By using Higher Order Statistics (Spectra), a modified version of our two step spectral factorization algorithm [Sèbe et al. 2005] of coda waves has been proposed allowing to recover the complete source time function. Belonging to the class of blind deconvolution methods, this new approach performs the estimation of STF without any prior knowledge of seismic Green’s function taking advantage of the stochastic nature of regional coda wavefield. Based on the strong assumption of a non-Gaussian probability density function, independently and identically distributed, of the coda excitation time series, the success of this algorithm provides, in addition to STF, interesting new insights on the statistical properties of the regional seismic wavefield. This algorithm has been applied on 2 events: the Rambervillers earthquake, France, 22/02/2003, and an artificial explosion in Kambara, Kyrgyzstan, 22/12/2009. Thank to the development of a multi-stations stacking procedure of 4th order statistics (tricorrelation), the proposed algorithm provides source time functions in good agreement with the source obtained by empirical Green function methods or with the known source phenomenology of the artificial explosion.
Session: ShakeMap-Related Research, Development, Operations, and Applications
Presenter   Mak, Sum
Schedule   Thu PM / Poster
Room   Hall 1
A GEM Testing Center for Intensity Prediction Equations: a Case Study for Italy
MAK, S., GFZ German Research Center for Geosciences, Potsdam, Germany, smak@gfz-potsdam.de; CLEMENTS, R., GFZ German Research Center for Geosciences, Potsdam, Germany, clements@gfz-potsdam.de; SCHORLEMMER, D., GFZ German Research Center for Geosciences, Potsdam, Germany, clements@gfz-potsdam.de
Macroseismic intensity has been, and will likely continue to be used for purposes such as seismic hazard assessment, loss estimation and communication to the public. Intensity prediction equations (IPEs) are being developed, although not at a rate as fast as its counterpart for physical ground motions. A Global Earthquake Model (GEM) testing center is being established to evaluate IPEs using truly prospective data (i.e., intensity reports recorded after the development of a model) and to periodically report their predictive skills. Here, we present the results from the prototype of this testing center on evaluating a collection of IPEs influential to seismic hazard assessment or model development for Italy. The mean absolute error between the model prediction and the observation is used as a measure of model performance. The observed intensity data are generated in two ways: the conventional way by expert investigations, or the modern way by internet questionnaires. The performance of the IPEs is found to be consistent on these two categories of data sets, verifying that the data generated by the two ways are compatible in quality, and justifying any subsequential use of intensity reports compiled by internet questionnaires. IPEs with a physically meaningful functional form are found to perform generally better. IPEs derived from a large amount of data are not found to perform better than those derived from a moderate amount of data. A recently developed global IPE is found to have comparable performance with the best indigenous model, suggesting that intensity attenuation characteristics local to Italy may not be significant. Only historical (retrospective) data contain sufficient strong earthquake data, and the performance of IPEs under these data are found to be less distinctive. The result of this study is not yet conclusive for strong earthquakes but IPEs with better predictive skills are seen for weak-to-moderate earthquake.
Session: Infrasound and Seismoacoustics
Presenter   Ford, Sean
Schedule   Wed 5:30 PM / Oral
Room   155D
Estimation of Yield and Height-of-Burst for Near-Surface Explosions from Seismoacoustic Data
RODGERS, A. J., LLNL, Livermore, CA, rodgers7@llnl.gov; FORD, S. R., LLNL, Livermore, CA, ford17@llnl.gov; RAMIREZ, A. L., LLNL, Livermore, CA, ramirez3@llnl.gov; XU, H., San Diego Supercomputer Center, La Jolla, CA, hxu6@hotmail.com; TEMPLETON, D. C., LLNL, Livermore, CA, templeton4@llnl.gov; DODGE, D. A., LLNL, Livermore, CA, dodge1@llnl.gov
Explosions near the Earth’s surface generate overpressure motions in the atmosphere and seismic motions in the solid earth. Amplitudes of these seismoacoustic signals depend on explosive yield and the observing range. Furthermore, the partitioning of energy into overpressure and seismic motions depends strongly on the height-of-burst (HOB) or depth-of-burial (DOB, where HOB for buried explosions is negative). Using experimental data from various near-surface explosions we developed models of these signal amplitudes at local distances (< 15 km). Signal amplitudes for a given yield and at a given distance are reduced when explosion emplacement is moved out of the observing medium. For example, seismic amplitudes are reduced as the emplacement moves from normal burial for containment toward the surface. Conversely, overpressure amplitudes are reduced as emplacement moves from above ground to below ground. We have quantified this behavior by developing empirical models to predict signal amplitudes for near-surface chemical explosions. These models can be used to simultaneously estimate explosive yield and HOB for a set of air-blast and seismic measurements. We find that signal amplitudes show strong trade-off between yield and HOB, which results in non-uniqueness in estimated parameters using a single measurement type (e.g. overpressure or seismic alone). However, by using both overpressure and seismic measurements, this non-uniqueness is greatly reduced. We find that yield can be estimated with errors less than 50% and that HOB can often be estimated with the correct sign (i.e. above or below ground). Additionally we have used hydrodynamic simulations to compute air-blast signals for a range of near-surface emplacements. We have validated modeling with experiment data for alluvium lithologies and extended model predictions to other geologic materials.
Session: Velocity Models and Modeling
Presenter   Morozov, Igor
Schedule   Wed 11:45 AM / Oral
Room   155A
Physics of the Correspondence Principle
MOROZOV, I. B., University of Saskatchewan, Saskatoon, SK, Canada, igor.morozov@usask.ca
Attenuation of Earth’s oscillations and seismic waves is among the most fundamental theoretical problems in seismology. This problem also has a very general solution, which is well-known as the correspondence principle (CP). This principle is thought to be applicable to all types of deformations from lab to planetary scales and environments, and all modern models of attenuation are based on it. Nevertheless, we note that the CP may not work in heterogeneous media, and it still needs to be tested by comparisons with full physical models. We present several examples of systems in which the CP leads to physically problematic results: 1) creep or oscillation of a thermoelastic bar; 2) reflection from a contrast in Q; 3) overestimation of dissipation within a Love wave. Other general and intriguing theoretical problems related to the CP are the relation between bulk and shear attenuation, analyticity of the “partial derivatives oh phase velocities,” and the existence of physical dispersion, i.e. dependence of material properties on frequency. Understanding of such problems elucidates the physical character of seismic attenuation and points to serious limitations of the existing, CP-based model.
Session: ShakeMap-Related Research, Development, Operations, and Applications
Presenter   Cauzzi, Carlo
Schedule   Thu PM / Poster
Room   Hall 1
The Next Generation Swiss ShakeMap: a Scientific and Technical Overview.
CAUZZI, C., Swiss Seismological Service, Zürich, Switzerland, carlo.cauzzi@sed.ethz.ch; CLINTON, J., Swiss Seismological Service, Zürich, Switzerland, john.clinton@sed.ethz.ch; CUA, G., , , georgia.cua@gmail.com; EDWARDS, B., Swiss Seismological Service, Zürich, Switzerland, edwards@sed.ethz.ch; FÄH, D., Swiss Seismological Service, Zürich, Switzerland, donat.faeh@sed.ethz.ch; KÄSTLI, P., Swiss Seismological Service, Zürich, Switzerland, kaestli@sed.ethz.ch; HEIMERS, S., Swiss Seismological Service, Zürich, Switzerland, stefan.heimers@sed.ethz.ch
Since 2007, the Swiss Seismological Service (SED) at ETH Zürich uses the ShakeMap (Wald et al., EQS 1999a-b) codes distributed by the USGS to automatically generate ground shaking scenarios for peak ground acceleration and velocity (PGA and PGV) within minutes of any felt event in Switzerland. The official Swiss ShakeMaps (see sedshakemap.ethz.ch) are based on a) ShakeMap v3.2; b) the ground motion prediction equation (GMPE) of Cua and Heaton (AGU 2008); c) the ground-motion to intensity conversion equation (GMICE) of Kästli and Fäh (ECEES 2006); d) the amplification factors derived for Switzerland by Cua et al. (AGU 2007) from macroseismic intensity data. Ground-motions as recorded at all realtime stations in Switzerland are processed through scwfparam (see www.seiscomp3.org/wiki/doc/applications/scwfparam and Cauzzi et al., ESC 2012), designed at the SED to provide rapid parameterisation of ground-motions and input to ShakeMap through the earthquake monitoring software SeisComP3 (www.seiscomp3.org and Hanka et al., NHESS 2010). In order to take advantage of the major improvements in the latest code, SED is presently transitioning to v3.5. We take this opportunity to concurrently undertake a careful revision of the scientific basis of Swiss ShakeMaps. Among the changes being implemented are: a) the parameterisation and implementation of the GMPE for hard rock sites by Edwards and Fäh (BSSA 2013), allowing the prediction of response spectra in addition to PGA and PGV; b) the use of the GMICE by Faenza and Michelini (GJI 2010); c) the verification of the amplification factors based on macroseismic intensity data. Decisions taken within the framework of the upgrade project are also critical to ensure coherency with other SED strong-motion products, from early warning to risk assessment studies. In this contribution, we discuss the main scientific and technical points related to the transition, with the aim of sharing our strategies and experience with the community.
Session: Advances in High-Frequency Physics-Based Earthquake Systems Simulation
Presenter   Thompson, Eric
Schedule   Wed PM / Poster
Room   Hall 1
Developing and Validating a Community Velocity Model for Boston
THOMPSON, E. M., Tufts University, Medford, MA, eric.thompson@tufts.edu; BERRY, B., Tufts University, Medford, MA, bradford.berry@tufts.edu; ZHU, J., Tufts University, Medford, MA, jing.zhu@tufts.edu; BAISE, L. G., Tufts University, Medford, MA, laurie.baise@tufts.edu; EBEL, J., Boston College, Chestnut Hill, MA, john.ebel@bc.edu; RODGERS, A., Lawrence Livermore National Laboratroy, Livermore, CA, rodgers7@llnl.gov
Developing and testing a three dimensional (3D) seismic velocity model for Boston provides the earthquake hazard community with an important case study for the eastern United States (US). Recent data, including updated geologic maps, compiled geotechnical data, microtremor surveys, and seismic velocity measurements have been assembled to create a 3D seismic velocity model for the greater Boston region. The City of Boston is located in a shallow sedimentary basin within the fault-bounded Boston Basin. The Boston Basin is defined by north-dipping faults that separate granitic and volcanic rocks from meta-sedimentary rocks. The bedrock is overlain by glacial deposits. In the downtown area and along the two rivers that feed Boston Harbor, the glacial deposits are overlain by marine clay, marsh deposits, and then artificial fill, creating a shallow sedimentary basin. The impedance contrast between the young sediments and the bedrock is on the order of three times stronger than the impedance contrast between sediment and bedrock in other well studied (in terms of earthquake hazard) urban sedimentary basins such as the San Francisco Bay. This strong impedance contrast at relatively shallow depths results in high frequency resonant peaks. This resonance is particularly important in the eastern US because ground motions tend to be rich in higher frequencies because they do not attenuate as rapidly as in the western US. Additionally, many eastern US sources have been observed to have large stress drops that result in more high frequency energy (e.g., 1988 Saguenay earthquake and 2011 Mineral earthquake). We perform 3D simulations to investigate the focusing/scattering effects of the bedrock structure in the Boston Basin and the shallower sedimentary basin.
Session: Next Generation of Ground Motion Prediction Models
Presenter   Bozorgnia, Yousef
Schedule   Wed 8:30 AM / Oral
Room   155C
NGA-West2: A Comprehensive Research Program to Update Ground Motion Prediction Equations for Shallow Crustal Earthquakes in Active Tectonic Regimes
BOZORGNIA, Y., PEER, University of California, Berkeley, CA, yousef@berkeley.edu; ABRAHAMSON, N., Pacific Gas & Electric Company, San Francisco, CA; CAMPBELL, K., EQECAT Inc., Beaverton, OR; ROWSHANDEL, B., California Earthquake Authority, Sacramento, CA; SHANTZ, T., California Department of Transportation, Sacramento, CA
In 2003, the Pacific Earthquake Engineering research Center (PEER) initiated a large research program to develop next generation attenuation relationships for shallow crustal earthquakes in active tectonic regions (now called NGA-West1). The project concluded in 2008 and provided several important deliverables, including a strong motion database of recorded ground motions and a set of peer-reviewed ground motion prediction equations (GMPEs). Many researchers, practitioners, and organizations throughout the world are now using the NGA-West1 models and the NGA database for research and engineering. As successful as the original NGA-West1 program was, there were some additional and complementary ground motion issues and supporting research projects that could not be addressed in NGA-West1 program due to time constraints. The goal of a follow-up study, NGA-West2, is to fill these gaps. The NGA-West2 project addresses the following technical issues: (1) directivity models for NGA GMPEs; (2) directionality of ground motions; (3) development of GMPEs for vertical ground motions; (4) updating NGA-West models for small magnitude data; (5) updating NGA-West models for moderate & large magnitude data: (6) damping scaling of GMPEs for response spectra; (7) development of epistemic uncertainty model; (8) further development of site response. For example, under the database Tasks (4) and (5), thousands of ground motions recorded worldwide since 2003 have been uniformly processed and added to the database, including the 2008 magnitude 7.9 Wenchuan (China), 2009 L'Aquila (Italy), 2010 Darfield (New Zealand), and 2011 Christchurch (New Zealand). The overall size of the NGA-West2 database is by over a factor of five larger than the NGA-West1 database. Incorporating the results and findings of such supporting research tasks, updated GMPEs for horizontal ground motion, as well as new GMPEs for vertical component, are being developed.
Session: Infrasound and Seismoacoustics
Presenter   Russell, David
Schedule   Wed 2:45 PM / Oral
Room   155D
Modeling Surface Waves from Above- and Below-Ground Explosions in Alluvium and Limestone
RUSSELL, D. R., Weston Geophysical Corp., Satellite Beach, FL, dhrussell@westongeophysical.com; BONNER, J. L., Weston Geophysical Corp., Lufkin, TX, jes_bonner@westongeo.com; REINKE, R., DTRA, Albuquerque, NM, robert.reinke@dtra.mil
The May 2012 HUMBLE REDWOOD III (HRIII) experiment in New Mexico provides a unique dataset to study surface wave generation from explosions conducted above and below ground for different rock types. Four 90.6 kg TNT-equivalent explosions were detonated either at 2 m height-of-burst (HOB) or 7 m depth-of-burial (DOB) at separate alluvium and limestone test sites. For the alluvium site, data from a temporary seismo-acoustic network show that fundamental-mode surface waves (Rg) were the largest seismic phase generated by shots both above and below ground. The 1-5 Hz Rg amplitudes from the 7 m DOB in-alluvium shot were 4-5x larger than the above-alluvium shot. When compared to data from the 7 m DOB limestone shot, the Rg amplitudes generated from in alluvium shot were 10x larger. The Rg amplitudes from the fully-confined limestone shot were 15x larger than Rg recorded from the co-located 2 m HOB shot. To model these different Rg characteristics, we generated 1-D velocity models for both test sites using observed surface wave dispersion. We considered two different methods for synthetic seismogram generation. For the above-ground shots, we have updated methods described by Murphy (1981) that couple near-field blast wave pressures and shapes with source medium properties to model seismic data at distance. For the below-ground shots, we estimate a moment using Denny and Johnson (1991) and generate explosion synthetics. For the alluvium test site, both methods provide synthetics that match observed 1-5 Hz Rg amplitudes and waveform shapes. For the limestone test site, the synthetic seismograms from both methods are slightly larger (factor or 2 or less) than the observed data, which may be due to the more complex 2D or 3D structure near this test site plus possible Rg radiation asymmetries. This modeling provides a viable technique to generate synthetic overpressures and ground motions that can be used to estimate combined seismo-acoustic yields for near-surface explosions
Session: Characterizing Active Faults for Seismic Hazard Assessments
Presenter   Li, Zefeng
Schedule   Fri AM / Poster
Room   Hall 1
STUDENT
Shear-wave Splitting Tomography along the Karadere-Duzce Branch of the North Anatolian Fault
LI, Z., Georgia Institute of Technology, Atlanta, GA, zli354@gatech.edu; ZHANG, H., University of Science and Technology of China, Hefei, Anhui, China, zhang11@ustc.edu.cn; PENG, Z., Georgia Institute of Technology, Atlanta, GA, zpeng@gatech.edu
Seismic shear waves propagating through anisotropic material are expected to separate into fast and slow components, known as shear wave splitting (SWS). Many previous studies on SWS have focused on spatio-temporal anisotropy distributions by simply averaging the splitting measurements along the entire ray paths (Peng and Ben-Zion, 2004, 2005). Zhang et al. (2007) developed a three-dimensional SWS tomography method to map the spatial distribution of anisotropy by back-projecting the splitting delay times along ray paths. Here we apply this method to the Karadere–Duzce branch of the north Anatolian fault (NAF) in western Turkey, where clear fault-zone induced anisotropy has been found following the 1999 Mw 7.4 Izmit and Mw 7.1 Duzce earthquakes. The input data consists of 7856 measurements of SWS delay time from aftershocks of both mainshocks, recorded by 31 temporary PASSCAL stations in a 6-month period.We calculate the strength of anisotropy using an anisotropic coefficient K. We also compute the derivative weighted sum to infer spatial resolution of the anisotropic image. The tomographic results show strong anisotropy and distinct spatial heterogeneity along the Karadere-Duzce branch of NAF. We find the existence of a continuous belt-like highly anisotropic zone along the fault strike, generally ~3 km wide and down to ~5 km. The strongest anisotropy in the study region appears at around a depth of 2.5 km, and anisotropy becomes weaker at larger depth. These results are generally consistent with the inference of shallow fault-zone induced anisotropy from previous studies (Peng and Ben-Zion, 2004). However, we also identify several highly anisotropic zones in isolated blocks outside the fault zones. These isolated blocks may provide possible explanations of abnormal fast directions observed in certain stations. Our results, together with previous studies, suggest that SWS tomography provides a useful tool to image high-resolution spatial distribution of anisotropy.
Session: New Developments in Earthquake Forecasting and Predictability Research
Presenter   Steacy, Sandy
Schedule   Thu 9:30 AM / Oral
Room   155D
A Hybrid STEP/Coulomb Model for Aftershock Rates
STEACY, S., University of Ulster, Coleraine, N. Ireland, s.steacy@ulster.ac.uk; GERSTENBERGER, M., GNS Science, Lower Hutt, New Zealand, m.gerstenberger@gns.cri.nz; WILLIAMS, C., GNS Science, Lower Hutt, New Zealand, c.williams@gns.cri.nz; RHOADES, D., GNS Science, Lower Hutt, New Zealand, d.rhoades@gns.cri.nz; CHRISTOPHERSEN, A., GNS Science, Lower Hutt, New Zealand, a.christophersen@gns.cri.nz
At present, there are two main methods for estimating aftershock rates – a physics based approach generally based on Coulomb combined with rate-state friction, or purely statistical techniques using ETAS or STEP models. The problem with the latter is that they leave out the stress interaction that appears to control much of the spatial distribution of aftershocks, whereas the former depend strongly on knowing the details of the slip distribution in order to accurately calculate the magnitude of the stress change.Here we develop a new model that combines the spatial constraints from Coulomb with the statistical approach of STEP. In essence, we use the Coulomb “map” as a Boolean filter and increase the likelihood of aftershocks in regions of positive Coulomb stress and decrease it in areas of negative stress. We do not consider the magnitude of the Coulomb stress change in this filter, simply whether the stress change is positive or negative.We apply this model (STEPC) to the Canterbury earthquake sequence and test it against the conventional STEP model and a Coulomb rate-state one. In all cases we find that Coulomb rate-state performs significantly worse than STEP and STEPC. Further, in most cases the new model performs better than standard STEP.
Session: Advances in High-Frequency Physics-Based Earthquake Systems Simulation
Presenter   Hollender, Fabrice
Schedule   Wed PM / Poster
Room   Hall 1
Taking Advantage of the Different Shear Wave Velocity Evaluation Methods for a Better Characterization of Sites
HOLLENDER, F., CEA Cadarache, Saint Paul lez Durance, France, fabrice.hollender@cea.fr; GUYONNET-BENAIZE, C., CEA Cadarache, Saint Paul lez Durance, France; DECHAMP, A., CEA Ile de France, Bruyères le Chatel, France; CORNOU, C., ISTerre, Université Joseph Fourier, Grenoble, France; FOTI, S., Politecnico di Torino, Torino, Italy; GAROFALO, F., Politecnico di Torino, Torino, Italy; COX, B. R., University of Texas, Austin, TX; OHRNBERGER, M., University of Potsdam, Golm, Germany; BARD, P.-Y., ISTerre, Université Joseph Fourier, Grenoble, France
Shear wave velocity is the most important property that controls ground motion amplification. Its determination is essential to characterize seismic site effects. Determining Vs features (derived in terms of soil classes, Vs30, or whole velocity profile) is a key issue in numerous engineering seismology fields, from GMPEs derivation (characterization of accelerometric station sites) to building code application (soil classes), including seismic hazard studies of industrial facilities (that often a thorough site characterization with assessment of geotechnical parameters and lateral variability). Different approaches can be used to assess Vs. While often considered as reference methods, the borehole techniques (e.g. cross- and down-hole tests) have however some drawbacks: they are costly and invasive, investigate a small volume of material compared to the wavelength of seismic waves, use a frequency range much higher than those involved during earthquakes. Conversely, non-invasive active or passive surface wave methods are less expensive, allow the characterization of a representative volume of material and involve directly the specific frequency range of interest in engineering seismology, even if their resolution capabilities are much smaller than those of invasive techniques. The implementation of surface wave tests, especially the inversion step, has not yet been framed within a standardized procedure, which may induce troublesome misinterpretations. We launched a project that aims 1/ to better assess the complementarity of these different methods, 2/ to write guidelines for standardized implementation of surface wave methods, and 3/ to propose a global methodology, taking benefits of different approaches, for a future site characterization of accelerometric networks stations. First intercomparisons and surveys will be presented, also taking into account the comparison with direct estimates of site effects from generalized inversion of French accelerometric data.
Session: ShakeMap-Related Research, Development, Operations, and Applications
Presenter   Cauzzi, Carlo
Schedule   Thu PM / Poster
Room   Hall 1
Scwfparam: a Tool for Rapid Parameterisation of Ground Motions and Input to ShakeMap in SeisComP3.
CAUZZI, C., Swiss Seismological Service, Zürich, Switzerland, carlo.cauzzi@sed.ethz.ch; CLINTON, J., Swiss Seismological Service, Zürich, Switzerland, john.clinton@sed.ethz.ch; BECKER, J., gempa GmbH, Potsdam, Germany, jabe@gempa.de; KÄSTLI, P., Swiss Seismological Service, Zürich, Switzerland, kaestli@sed.ethz.ch
A key input to an accurate ShakeMap is the near-real-time processing and parameterisation of all available recordings in the epicentral region. The key parameters for ShakeMap are peak ground acceleration, peak ground velocity and 5%-damped pseudo-spectral acceleration at vibration periods equal to 0.3, 1 and 3 s for all horizontal channels. These parameters must be written into ShakeMap compatible XML files, along with basic station information and event description. The Swiss Seismological Service (SED) at ETH Zürich uses the ShakeMap (Wald et al., EQS 1999a-b) codes distributed by the USGS to automatically generate ground shaking scenarios within minutes of any felt event in Switzerland. Spurred by the recent adoption of SeisComP3 (SC3, http://www.seiscomp3.org and Hanka et al., NHESS 2010) as the official earthquake monitoring system at the SED, we developed a SC3 module, scwfparam, that provides rapid parameterisation of ground motions and input to ShakeMap. The main features of this software are presented in this contribution. Scwfparam (http://www.seiscomp3.org/wiki/doc/applications/scwfparam) is open source and uses SeedLink and ArcLink as protocols for real-time and offline data retrieval. The module is highly flexible and customisable in terms of scheduling, processing options and parameterisation output. Part of the official SC3 release, scwfparam is compatible with any new module eventually developed and handled by SC3. It is expected that scwfparam will simplify the implementation of ShakeMap at institutions running SC3, that is already widely used across the world. The output from scwfparam is also written into an ad-hoc extended SC3 database. Beyond ShakeMap, scwfparam and the related extended datamodel constitute the framework for the Rapid Raw Strong-Motion Database, that will provide near real-time access to all on-scale waveforms from open datasets in Europe, presently under development within the EC-funded project NERA (http://www.nera-eu.org/).
Session: Infrasound and Seismoacoustics
Presenter   Hayward, Chris
Schedule   Wed 3:00 PM / Oral
Room   155D
Southern Methodist University's Experience with Infrasound Instrumentation
HAYWARD, C. T., Southern Methodist University, Dallas, TX, hayward@smu.edu; GOLDEN, P., Southern Methodist University, Dallas, TX, pgolden@smu.edu; STUMP, B. W., Southern Methodist University, Dallas, TX, bstump@smu.edu
In 1999, SMU began experiments to record infrasound at the Lajitas Seismic Array (TXAR) in the 0.5-5Hz band using a US Bureau of Mines instrument design due to its low cost and ruggedness relative to available sensors at the time.  The observations of seismo-acoustic signals prompted additional experiments including in-mine monitoring and near-source monitoring and required additional infrasound sensors, all based upon the simple pressure sensors.  Analysis of signals at TXAR indicated that instrument noise limited signal detection, and therefore the design evolved into a system that eventually included a random binary calibrator.  Again cost constraints limited the instrument choice.  With the success of the first Korean seismo acoustic arrays and the growing CTBT interest in infrasound, Chaparral M2 sensors became the instrument of choice for these applications. Obtaining and funding infrasound instrument purchase was still problematic for seismologists and vulcanologists, but there was a recognition that observations using inexpensive electret microphones could provide initial observations and validation.  While there was wide recognition that such systems were limited at low frequency, their low cost meant that any seismic recorder with an additional channel could become a seismo-acoustic recorder with little expense.  Various researchers experimented with methods to extend the usefulness of such electret systems, with one example finding use in snow avalanche detection.In SMU's seismo-acoustic and infrasound experience, installation effort and cost have been major controls in the scale and scope of observations.  Controlling wind noise, a major concern, has therefore been done in the least expensive manner using garden soaker hoses, still the viable option for field experiments.
Session: Characterizing Active Faults for Seismic Hazard Assessments
Presenter   Holland, Austin
Schedule   Fri AM / Poster
Room   Hall 1
Optimal Fault Orientations from Observed Focal Mechanisms During the Earthscope Transportable Array Deployment within Oklahoma
HOLLAND, A. A., Oklahoma Geological Survey, Norman, OK, austin.holland@ou.edu
154 focal mechanisms determined for earthquakes occurring in Oklahoma for about 30 months were used to determine the optimally oriented fault orientations within the region. A large number of these focal mechanisms were concentrated in central Oklahoma and can be associated with the ongoing Jones earthquake swarm and the November 5, 2011 Mw 5.7 Prague earthquake sequence. In addition there are focal mechanisms determined throughout Oklahoma while the Earthscope Transportable Array was located in the region. Focal mechanisms used in this study include Regional Moment Tensor solutions when available and first motion focal mechanisms. The probability density functions with a 10-degree bin were calculated for both possible nodal planes associated with the observed focal mechanisms. Focal mechanisms within the Jones swarm are clearly different than those from the rest of Oklahoma and show a greater range in variation than what would be expected from a single stress field. These orientations of focal mechanisms show a clear correlation to existing fracture orientations. The probability density functions (PDF) of focal mechanism orientations clearly demonstrate that strike-slip motion on steeply-dipping faults dominate the distribution. Fault strike was restricted to the range of 0 to 180 degrees. The optimal fault strike orientation ranges between 40°-60° and 130°-150° and represent fault orientations most likely to have an earthquake. Because damaging earthquakes have occurred in Oklahoma without surface rupture and only the Meers fault has been identified with Quaternary offset, identifying fault orientations likely to have earthquakes may help to better constrain future fault and seismic hazard studies within the region.
Session: Oceanographic and Atmospheric Signals in Seismology
Presenter   Wong, YeouHui
Schedule   Fri AM / Poster
Room   Hall 1
STUDENT
Survey of Broadband Seismic Noise Recorded at Permanent Stations of the Utah Regional Seismic Network
WONG, Y., Dept. of Geology and Geophysics, University of Utah, Salt Lake City, UT, spring_w2007@hotmail.com; KOPER, K. D., Dept. of Geology and Geophysics, University of Utah, Salt Lake City, UT, koper@seis.utah.edu; BURLACU, V., Dept. of Geology and Geophysics, University of Utah, Salt Lake City, UT, burlacu@seis.utah.edu
The University of Utah Seismograph Stations (UUSS) has maintained a network of seismometers in the Utah region since 1966. Continuous digital data from these stations are available at the IRIS DMC from June 2000 to the present. Currently, UUSS operates 25 broadband stations, corresponding to 75 channels (HHZ, HHE, HHN), which are sampled at 100 sps. We present here a comprehensive noise analysis of all available Utah broadband data (2000-2012) for periods of 0.04-500 s. At longer periods the noise is sensitive to natural phenomena such as wind and ocean wave interactions with the solid Earth, while at shorter periods the noise is sensitive to cultural activities like traffic, construction, and so on. The seismic noise is also indicative of overall station health and the accuracy of station metadata information. We estimate 3x3 spectral covariance matrices for each hour of data using 10 overlapping sub-windows, and then use a log10-averaging scheme to reduce the number of independent frequency bins to approximately 300. The diagonal elements of the covariance matrices represent power spectral density estimates for each of the three channels. We extract polarization information from the cross-spectra by performing eigen-decomposition of the matrix and examining the complex values of the dominant eigenvector. Results are visualized using a spectrogram format in which a particular scalar quantity is plotted as a function of time and period. At typical microseismic periods of about 3-20 s, Rayleigh wave energy arrives from both the Pacific and Atlantic oceans and gliding peaks of energy are observed within the single- and double-frequency bands.
Session: Characterizing Active Faults for Seismic Hazard Assessments
Presenter   Woolery, Edward
Schedule   Fri AM / Poster
Room   Hall 1
Toward a Geologic Boundary Characterization of the Recently Discovered Charleston Uplift—New Madrid Seismic Zone, Central United States
WOOLERY, E., University of Kentucky, Lexington, KY, woolery@uky.edu; VANARSDALE, R., University of Memphis, Memphis, TN, rvanrsdl@memphis.edu; PYRNE, D., Fugro GeoConsulting, Inc, Houston, TX, depryne@gmail.com; CSONTOS, R., Newmont Mining Corporation, Englewood, CO, Ryan.Csontos@Newmont.com
Exploratory seismic walkaway soundings were acquired across the northern boundary of a stratigraphic uplift in the northeastern vicinity of the New Madrid seismic zone for evidence of genesis (i.e., neotectonic or fluvial). The previously unknown uplift, called the Charleston uplift, was discovered using 520 electric logs from shallow (100 meter) lignite exploration wells and geospatial stratigraphic mapping. Although there are no known surface faults bounding this feature, the 30+ meters of structural amplitude exhibited in the well-log mapping of Quaternary and Tertiary horizons are hypothesized as having a tectonic origin. Two seismic soundings were performed north of and within the uplift to further test this hypothesis. Results indicated 47 m and 60 m of relief across the tops of the deeper Cretaceous and Paleozoic horizons, respectively. Although additional data are needed for confirmation, these soundings support a tectonic genesis. If uplift is proven to be fault controlled this is significant for the regional seismic hazard assessment, because faults of the northern segment of the New Madrid seismic zone are poorly constrained and contribute to uncertainty in the forecast of ground motions. Presently, the principal argument for a fault trending northeast from New Madrid, Missouri is a diffuse line of epicenters. Many of these earthquakes have focal solutions that exhibit reverse dextral movement along northeast trending structure, the most recent of these being the M 3.9 February 21, 2012 earthquake near Charleston, Missouri, also a suggested location for the M 6.7 1895 earthquake. Consequently, a fault-bounded uplift will define a major seismogenic structure and provide important physical constraint on the enigmatic seismotectonics of the New Madrid seismic zone.
Session: ShakeMap-Related Research, Development, Operations, and Applications
Presenter   Amato, Alessandro
Schedule   Thu PM / Poster
Room   Hall 1
ShakeMaps Improvement and Criticalities in the Epicentral Area of the 2012 Emilia Seismic Sequence: Implications for the Post-Seismic Actions
FAENZA, L., Istituto Nazionale Geofisica Vulcanologia, Rome, Italy, licia.faenza@ingv.it; CULTRERA, G., Istituto Nazionale Geofisica Vulcanologia, Rome, Italy, giovanna.cultrera@ingv.it; MICHELINI, A., Istituto Nazionale Geofisica Vulcanologia, Rome, Italy, alberto.michelini@ingv.it; MELETTI, C., Istituto Nazionale Geofisica Vulcanologia, Pisa, Italy, carlo.meletti@pi.ingv.it; AMATO, A., Istituto Nazionale Geofisica Vulcanologia, Rome, Italy, alessandro.amato@ingv.it; D'AMICO, V., Istituto Nazionale Geofisica Vulcanologia, Pisa, Italy, vera.damico@pi.ingv.it
On May 20, 2012 a Mw 5.9 earthquake struck the Emilia-Romagna region, causing casualties and severe damages to the historical buildings and to the economical activities of Northern Italy. The event was followed by a seismic sequence including six earthquakes with magnitudes >5.0, the biggest of them with Mw 5.7 on May 29 at about 12 km west of the first mainshock. During the Emilia sequence, the ShakeMap of peak ground velocity and acceleration (PGV and PGA) and acceleration response spectra (PSA) at 0.3s, 1.0s and 3.0s for earthquakes of Ml≥3.0 have been uploaded to the INGV web site and updated as more data became available. The on-line ShakeMaps have been determined only for rapid definition of the experienced strong ground shaking and thus provide preliminary and incomplete estimates of the experienced shaking. In spite of this, the maps were referred to in an Italian law (Legge 1 agosto 2012, n. 122) on urgent measures of post seismic action to secure the building of the area affected by the earthquakes occurred on May 20 and 29, 2012. This fact prompted us to verify the definition of the shaking suffered during the seismic sequence and the effect of different estimates in the application of the law. Moreover, we propose a possible, albeit expeditious, comparison between the maps of ground shaking and seismic hazard map, which is the basis of Italian seismic legislation.
Session: What are the Limits of Explosion Source Model Predictions?
Presenter   Larmat, Carene
Schedule   Thu AM / Poster
Room   Hall 1
Wave Speed and Attenuation of Seismic Waves Through Earth Structures On or Near Climax Stock Using Data from the Source Physics Experiment.
LARMAT, C. S., Los Alamos National Laboratory, Los Alamos, NM, carene@lanl.gov; ROWE, C. A., Los Alamos National Laboratory, Los Alamos, NM, char@lanl.gov; PATTON, H. J., Los Alamos National Laboratory, Los Alamos, NM, patton@lanl.gov
In order to gain a better understanding of the generation and propagation of seismic energy from underground explosions for hard rock media, the U. S. Department of Energy has funded a series of chemical tests on the National Nuclear Security Site (NNSS) as part of the Source Physics Experiment (SPE). To date, three such tests of yields 100, 1000, and ~900 kg have been conducted in the same borehole. Waveforms have been recorded on five linear geophone lines extending radially from ground zero and having offsets from 100 to 2000 m with station spacing of 100 m. Our work focuses on the characterizing of wave speeds and attenuation properties of seismic waves traveling through shallow Earth structures within 2 km of ground zero. 1D velocity models were derived from combined analyses of P-wave travel times and Rg dispersion curves. 2D raytracing to fit arrival times was performed when 1D models failed to match observations. Attenuation of Rg waves is measured with a spectral domain method, and the results are consistent from one test to the next along all five lines. We find that a near-surface, thin, weathered layer of varying thickness and low wave speeds plays a major role on the observed waveforms. Between 3 and 10 Hz, Rg Q values measured on granite paths are lower than those observed for granite path on the Balapan test site in central Asia. The interpretation of wave speeds and attenuation of one of these paths (line #2) is consistent with a simple 1D model. For the other path (line #1), travel time analyses point to the need for a 2D model, and our interpretation indicates a general thickening of low wave speed layers northward with higher topography. This basin-like structure is consistent with the observation of particularly weak Rg amplitude decay on this line. Full waveform modeling will be performed to validate this interpretation. Further analyses of observations for paths crossing major structural boundaries surrounding ground zero are ongoing.
Session: Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Models
Presenter   Herrero, Andre
Schedule   Fri 2:00 PM / Oral
Room   155C
Seismic Source and Displacement Response Spectrum at Large Periods
HERRERO, A., Istituto Nazionale di Geofisica e Vulcanologia, Rome, Italy, andre.herrero@ingv.it; AVALLONE, A., Istituto Nazionale di Geofisica e Vulcanologia, Rome, Italy, antonio.avallone@ingv.it; LATORRE, D., Istituto Nazionale di Geofisica e Vulcanologia, Rome, Italy, diana.latorre@ingv.it; ROVELLI, A., Istituto Nazionale di Geofisica e Vulcanologia, Rome, Italy, antonio.rovelli@ingv.it
The scope of this study is the variation in space and frequency of the displacement response spectra, especially for large periods (SDLP). Since the asymptotic value of the displacement spectrum is the peak ground displacement (PGD), we expect to observe directly the low frequency patterns of the seismic source such as the radiation pattern and the directivity.The first step is to check if we may consider the SDLP a reliable proxy of the PGD. We use comparison between displacement response spectra computed on data from co-located accelerometer and high-rate GPS recorded during the Emilia 2012 earthquake, to expose the critical points of this assumption. Our conclusion is that it is a simple and robust proxy if the period range is well chosen.We use the largest inland earthquakes of Japan, recorded on the accelerometric K-net network to study the SDLP spatial variation. The first feature is the geometrical spreading. The decay with the distance shows that it is linked mainly to the surface waves. Thus we present the data corrected by the square root of the distance. The plot of the corrected SDLP values on a map shows directly the nodes of the focal mechanism. When we reduce the period of the ordinates, the focal mechanism imprint vanishes around 3 s.The directivity pattern is more difficult to observe directly and required a more sophisticated analysis in order to quantify it. However, its characterization allows us to access to seismic source parameters like the mean rupture velocity vector and the rupture laterality (offset of the nucleation point on the fault).
Session: Intermountain West Earthquake and Volcano Characterization and Related Hazards
Presenter   Seaman, Tyler
Schedule   Wed PM / Poster
Room   Hall 1
STUDENT
Seismicity and Tectonics of the Lake Tahoe Basin through the Truckee, California Region
SEAMAN, T. C., Nevada Seismological Laboratory, University of Nevada, Reno, NV, tceeman@gmail.com; RUHL, C. J., Nevada Seismological Laboratory, University of Nevada, Reno, NV, christineruhl@gmail.com; SCHMAUDER, G. C., Nevada Seismological Laboratory, University of Nevada, Reno, NV, gcschmauder@yahoo.com; SMITH, K. D., Nevada Seismological Laboratory, University of Nevada, Reno, NV, ken@seismo.unr.edu
East dipping range bounding normal faults, NE-trending sinistral, and NW-trending strike-slip faults of the Tahoe-Truckee, California-Nevada area define the transtensional strain environment of the Northern Walker Lane. North of Long Valley, California, the Sierran front is comprised of a series of left-stepping normal faults that terminate in the northern Lake Tahoe basin. Directly north of Lake Tahoe, slip is transferred through the Truckee, California area and northward through a complex distribution of primarily conjugate strike-slip and minor normal fault systems. The recent identification of the Polaris fault zone extends Northern Walker Lake Belt dextral slip south of the Mohawk Valley fault zone to Truckee. The region accommodates as much as 10 mm/yr of dextral shear and has a history of M6+ earthquakes. In contrast to the general lack of background seismicity associated with major range bounding normal faults in the Tahoe region, there is abundant seismicity north of the Lake through this slip transfer zone (focal mechanisms of small magnitude earthquakes provide limited constraints on the dips of major range bounding normal faults). A NE-SW-trending zone of seismicity in the N. Tahoe area includes several M4+ strike-slip and normal faulting sequences, and numerous identifiable structures. An M4.5 (2004), M4.8 (2005), and associated general increase in seismicity in the N. Tahoe area may have been triggered by broader deformation resulting from a Moho-depth dike injection event in late 2003 (von Seggern et al., 2008). We incorporate moment tensor solutions, short-period focal mechanisms, and high-precision event locations to model slip transfer processes between the northern extent of primary range front faulting in the Lake Tahoe basin through the Truckee, California area and into the strike-slip faults of the Northern Walker Lane Belt.
Session: Including Ground Failure in Scenario Events, Rapid Response, and Loss Estimation Models
Presenter   Nowicki, M. Anna
Schedule   Wed AM / Poster
Room   Hall 1
STUDENT
Development of a Globally Applicable Model for Near Real-time Prediction of Seismically Induced Landslides
NOWICKI, M. A., Indiana University/USGS, Bloomington, IN/Golden, CO, anowicki28@gmail.com; WALD, D. J., US Geological Survey, Golden, CO, wald@usgs.gov; HAMBURGER, M. W., Indiana University, Bloomington, IN, hamburg@indiana.edu; HEARNE, M., US Geological Survey, Golden, CO, mhearne@usgs.gov; THOMPSON, E. M., Tufts University, Medford, MA, eric.thompson@tufts.edu
Substantial effort has been invested to understand where seismically induced landslides may occur in the future, as they are a costly and frequently fatal threat in mountainous regions. Though some regional efforts have succeeded, no uniformly agreed-upon method is available for predicting the likelihood and spatial extent of seismically induced landslides. We are developing a statistical model for estimating landslide distributions in near-real time around the globe for use in the U. S. Geological Survey (USGS) Prompt Assessment of Global Earthquakes for Response (PAGER) system. Here we use standardized outputs of ground shaking from the latest USGS ShakeMap Atlas to develop an empirical landslide probability model by combining these shaking estimates with broadly available landslide susceptibility proxies, such as topographic slope, surface geology, and climatic parameters. We include earthquakes for which digitally-mapped landslide inventories and well-constrained ShakeMaps are available. Using logistic regression, the resulting database is used to build a predictive model of the probability of landslide occurrence. The control landslide database includes the Guatemala (1976); Northridge, California (1994); ChiChi, Taiwan (1999); Chuetsu, Japan (2004); and Wenchuan, China (2008) earthquakes, as well as ShakeMaps for moderate-sized events without landslides. The performance of the regression model is assessed using statistical goodness-of-fit metrics (e.g., Aikaike Information Criterion) and a qualitative review to determine which combination of the proxies provides both the optimum prediction of landslide-affected areas and minimizes the ‘false alarms’ in non-landslide zones. Combined with near-real time ShakeMaps, we anticipate using our model to make generalized predictions of whether or not landslides are likely to occur (and if so, where) for earthquakes around the globe, and eventually to include these estimates into the loss forecasts of the PAGER system.
Session: Velocity Models and Modeling
Presenter   Arroucau, Pierre
Schedule   Thu AM / Poster
Room   Hall 1
A Shallow S-wave Velocity Model for the Eastern Tennessee Seismic Zone from Rayleigh Wave Ambient Noise Tomography
ARROUCAU, P., North Carolina Central University, Durham, NC, parroucau@nccu.edu; KUPONIYI, A. P., University of Victoria, Victoria, BC, Canada, ayodeji.kuponiyi@gmail.com; VLAHOVIC, G., North Carolina Central University, Durham, NC, gvlahovic@nccu.edu; POWELL, C. A., University of Memphis, Memphis, TN, capowell@memphis.edu
The eastern Tennessee seismic zone (ETSZ) is an intraplate seismic region characterized by frequent but low magnitude earthquakes and is the second most active seismic area in the United States east of the Rocky Mountains. One key question in the ETSZ is the actual relationship between earthquake distribution and geological structure at depth. Seismicity is mostly confined in the Precambrian basement, below the Paleozoic cover of the southern Appalachian foreland fold-and-thrust belt and shows little to no correlation with surface geological features. Since the middle of the seventies, the Center for Earthquake Research and Information (CERI) has installed and maintained several seismic networks in central and eastern United States. In this work, we use Rayleigh wave group and phase velocity dispersion information obtained from cross-correlation of seismic ambient noise at 24 short-period stations located in the vicinity of the ETSZ. The 3D velocity structure is estimated in four steps. First, dispersion curves are obtained for simultaneously recording station pairs for periods ranging from 2 to 20 s. Then, 2D group and phase velocity maps are determined for each period. Those maps are further used to reconstruct dispersion curves at fixed, regularly spaced locations. For each of these locations, a 1D shear-wave velocity profile is finally inverted for, that takes velocity information from previous studies into account. By providing new information about the upper crustal structure of this region, this work is a contribution to the understanding of the seismic activity of the ETSZ, and -to a broader extent- of the structure and evolution of the North American lithosphere.
Session: Velocity Models and Modeling
Presenter   Zhang, Jian
Schedule   Wed 9:00 AM / Oral
Room   155A
Ambient Noise Tomography of Surface-wave Attenuation using the USArray
ZHANG, J., Los Alamos National Laboratory, Los Alamos, NM, jjzhang@lanl.gov; YANG, X., Los Alamos National Laboratory, Los Alamos, NM, xyang@lanl.gov
Ambient noise tomography (ANT) based on the empirical Green’s function (EGF) from the cross-correlation (CC) has been successfully applied to seismic data for imaging high-resolution surface-wave velocity structures. Inverting attenuation models from seismic noise, on the other hand, has to first overcome the possible bias in the EGF amplitudes due to the uneven distribution of noise sources. In this work we conduct ANT of surface-wave attenuation using data from the USArray. We adopt a Bayesian approach similar to those used for earthquake-based inversion. We assume that the EGF amplitude depends mainly on the medium attenuation, site amplification, and noise-source intensity in the direction of the inter-station line, but little on noise-source intensities in other directions. We show that the correlation of the coda of correlation of noise (C3) can be used for EGF amplitude measurement and the resulting attenuation estimate is less biased. This is probably because of the fact that the coda of noise correlation contains more diffused noise energy and thus, effectively makes the noise-source distribution more homogeneous. Our results show a certain degree of correlation between the noise-based attenuation patterns and the regional geology.
Session: Intermountain West Earthquake and Volcano Characterization and Related Hazards
Presenter   Moschetti, Morgan
Schedule   Wed PM / Poster
Room   Hall 1
Variability in Earthquake Ground Motions from Kinematic Rupture Scenarios for the Salt Lake City Segment, Wasatch Fault
MOSCHETTI, M. P., US Geological Survey, Golden, CO, mmoschetti@usgs.gov; RAMIREZ-GUZMAN, L., UNAM, Mexico City, Mexico; HARTZELL, S., US Geological Survey, Golden, CO; ANGSTER, S., US Geological Survey, Golden, CO; PERSONIUS, S., US Geological Survey, Golden, CO; STEPHENSON, W., US Geological Survey, Golden, CO
We compute earthquake ground motions from kinematic rupture scenarios for the Salt Lake City segment of the Wasatch fault to investigate the variability of earthquake ground motions expected for a M7 earthquake. Kinematic rupture scenarios are computed from correlated random distributions using the method of Liu et al. (2006) and describe lateral distribution of slip amplitudes, rise times and rupture velocities on the fault. We constructed more than fifty rupture scenarios; the complete set includes ruptures sampling three hypocenters, three correlation lengths for the calculation of slip distributions, three corner frequencies and two sub-shear average rupture velocities. By using a kinematic source method, we are able to vary specific rupture parameters within ranges that are constrained by empirical relations and seismological observations; values selected for the scenarios are appropriate for a M7 normal-faulting earthquake and generally represent the mean and two-sigma values. Previous work demonstrates the importance of using a geometrically realistic fault model for the simulations. Our ground motion simulations utilize the fault model of Angster et al. (2013), which combines geological observations with geophysical measurements and is guided by analogous fault models. Here, we present preliminary results of our earthquake ground motion modeling through an analysis of the effect of correlation lengths on ground motions. We calculate the seismic wave propagation using the Hercules finite element tool-chain (Tu et al., 2006) and the Wasatch Community Velocity Model (Magistrale et al., 2006). Recent updates to this propagation code also permit us to model topographic effects within the Wasatch Front region.
Session: Networks and Instrumentation
Presenter   Spriggs, Neil
Schedule   Wed AM / Poster
Room   Hall 1
Comparison Study Between Vault Seismometers and a New Posthole Seismometer
SPRIGGS, N., Nanometrics Inc., Kanata, ON, Canada, neilspriggs@nanometrics.ca
Surface vault broadband seismometers have typically yielded good results on the vertical, but have been unreliable and noisy on the horizontal. There are several reasons for this issue, including inherent problems with surface tilt noise and thermal stability. A comparison study was undertaken between the highest performing vault seismometers and a new broadband Posthole seismometer in a downhole installation at different depths and in various environments at stations in North America, ranging from remote locations in Alaska to a noisy urban area. In the remote and urban studies, a spectral analysis was conducted and PSD plots were generated, the results of which will be provided in detail in the paper. This paper will discuss all of the results from these installations as well as the various installation techniques at these locations.
Session:Intermountain West Earthquake and Volcano Characterization and Related Hazards
Presenter   Angster, Stephen
Schedule   Wed PM / Poster
Room   Hall 1
A 3-D Structural Model of the Salt Lake City Segment of the Wasatch Fault: Resolving Normal Fault Geometries at Depth
ANGSTER, S. J., Geologic Hazards Science Center, U.S. Geological Survey, Golden, CO, sangster@usgs.gov; MOSCHETTI, M. P., Geologic Hazards Science Center, U.S. Geological Survey, Golden, CO, mmoschetti@usgs.gov; STEPHENSON, W. J., Geologic Hazards Science Center, U.S. Geological Survey, Golden, CO, wstephens@usgs.gov; PERSONIUS, S. F., Geologic Hazards Science Center, U.S. Geological Survey, Golden, CO, personius@usgs.gov
The Wasatch fault zone, one of the best-studied faults in the Intermountain West, poses a significant seismic threat to ~80% of Utah’s population. Recent studies show the need for detailed subsurface fault geometries to accurately model ground motions of future large earthquakes in the Salt Lake City (SLC) region. Here we present a new alternative 3-D structural model of the SLC segment to: 1) serve as an enhanced source model to perform deterministic ground motion modeling for improved hazard assessment of the SLC region; and 2) provide structural insight into normal fault geometries and kinematic relationships at seismogenic depth. Our fault model combines recent geological and geophysical data with insights from other fault models developed in similar geologic settings. Surface geometry of the SLC segment is constrained with previously mapped surface traces of the left-stepping Cottonwood Canyon, East Bench, and Warm Springs faults. Using analogous normal fault models we assume fault connectivity, where surficial fault strands merge at depth to form a single fault plane. We constrain fault geometry at depth with previously determined fault dips from geometric modeling, integrated with a community velocity model of the Salt Lake basin. A smooth gridding algorithm is used to extrapolate the fault plane at a constant dip of 50 degrees through the Precambrian basin floor to a depth of 20 km. Kinematic relationships at depths >5 km suggest slip partitioning between fault strands near the surface during large, complete segment-rupturing earthquakes. Apparent corrugations in the fault surface at depth also imply a dynamic slip distribution that affects slip direction and amount expressed at the surface.
Session: Induced Seismicity
Presenter   Ogwari, Paul
Schedule   Fri 3:45 PM / Oral
Room   155D
STUDENT
Using Induced Earthquakes to Estimate Hydraulic Properties of Subsurface Reservoirs in Central Arkansas
OGWARI, P. O., CERI - University of Memphis, Memphis, TN, opogwari@memphis.edu; HORTON, S. P., CERI - University of Memphis, Memphis, TN, shorton@memphis.edu
Disposal of hydrofracking-waste fluid by pressurized injection into subsurface aquifers triggered the Guy-Greenbrier earthquake swarm of 2010-2011 in central Arkansas (Horton, 2012). We estimate the hydraulic diffusivity of those aquifers by comparing the spatial expansion of triggered seismicity through time along the Guy-Greenbrier fault with pore pressure triggering fronts calculated for a range of diffusivities. The model calculations assume diffusion of fluids from the injection well into the surrounding formation increases pore pressure at a given distance over time. Preliminary results suggest the diffusivity is between 1.1 m2/s and 2 m2/s on the fault. We then estimate the permeability based on porosity of reservoir, viscosity and compressibility of the fluid. The diffusivity values imply a permeability value of 1.8× 10-15 m2 to 3.27× 10-14 m2. Using the Theis well equation, we can then model the pore pressure diffusion to infer storativity value varying between 5× 10-5 and 4.5× 10-4. The diffusivity values from the triggering front method are consistent with values computed from the ratio of transmissivity and storativity. Seismic activity begins at the center of the Guy-Greenbrier fault with activity propagating south faster than north along the fault during the next few weeks. This suggests a higher diffusion rate to the south.
Session: Earthquake Source Studies
Presenter   Escudero Ayala, Christian Rene
Schedule   Wed PM / Poster
Room   Hall 1
Seismic Studies at Ceboruco Volcano
ESCUDERO AYALA, C. R., Universidad de Guadalajara, Puerto Vallarta, Mexico, escudero.sisvoc@gmail.com; LOPEZ, J. I. P., Universidad de Guadalajara, Puerto Vallarta, Mexico, johndavid_927_9@hotmail.com; ANDRADE, H. E. P., Universidad de Guadalajara, Puerto Vallarta, Mexico, draco33_@hotmail.com; CORNU, F. J. N., Universidad de Guadalajara, Puerto Vallarta, Mexico, pacornu77@gmail.com
Many societies and their economies endure the disastrous consequences of destructive volcanic eruptions. The Ceboruco stratovolcano located at the west of the Mexican volcanic belt at 21.125o north, 76 km from the pacific coast and 2,280 meters above sea level has an eruptive recurrence of 200 years and its last activity was at 1875. This natural hazard could affect more than eight communities and important highways. Scientific knowledge constitutes the only way to avoid or at least to mitigate the negative effects of an eventual eruptive event. Accordingly we monitor and analyze the potential destructive effects of the Ceboruco volcano. Since March 2012 we deployed four seismic stations, equipped with a Nanometrics digital acquisition system TAURUS and a seismometer Lennartz 3D lite. We present seismic events classification, their average activity rate, and location; we also present a preliminary noise tomography; and an evaluation of the current seismic activity of the Ceboruco Volcano.
Session: Including Ground Failure in Scenario Events, Rapid Response, and Loss Estimation Models
Presenter   Zhu, Jing
Schedule   Wed 5:00 PM / Oral
Room   155B
STUDENT
A Geospatial Liquefaction Model for Rapid Response and Loss Estimation
ZHU, J., Tufts University, Medford, MA, jing.zhu@tufts.edu; BAISE, L. G., Tufts University, Medford, MA, laurie.baise@tufts.edu; THOMPSON, E. M., Tufts University, Medford, MA, eric.thompson@tufts.edu; WALD, D. J., U.S. Geological Survey, Golden, CO, wald@usgs.gov; KNUDSEN, K. L., U.S. Geological Survey, Menlo Park, CA, kknudsen@usgs.gov
Current regional liquefaction mapping techniques that heavily rely on surficial geologic maps are difficult to integrate into post-earthquake rapid response and loss estimation. We present a probabilistic approach to modeling spatial extent of liquefaction that uses globally available geospatial layers (e.g., derived from digital elevation models) and earthquake-specific parameters (e.g., ShakeMap produced by the United States Geologic Survey) and is relatively easy to implement for rapid response and loss estimation. In this approach, soil properties contributing to the likelihood that soils will liquefy, such as soil density and soil saturation, are modeled by explanatory variables that include shear-wave velocity, compound topographic index, distance to rivers, and a newly defined normalized distance parameter (distance to coast divided by the sum of distance to coast and distance to the basin inland edge). The magnitude and duration of earthquake load are modeled by peak ground acceleration weighted by a magnitude scaling factor. We develop a liquefaction database that is unbiased with respect to the areal extent of liquefaction from events where presence/absence of liquefaction has been documented. We develop two models: 1) the first, which is more accurate, is only applicable to coastal sedimentary basins, 2) the second is globally applicable. These models are developed with data from earthquakes in Christchurch, New Zealand and Kobe, Japan, and the model portability is then tested against earthquakes in additional regions such as Port-au-Prince, Haiti. We find that these models provide first-order approximations of the extent of liquefaction, appropriate for use in rapid response, loss estimation, and earthquake simulations.
Session: Velocity Models and Modeling
Presenter   Aleqabi, Ghassan
Schedule   Wed 10:45 AM / Oral
Room   155A
Imaging Shallow Midcontinent Rift Structure with Ambient Noise Tomography
ALEQABI, G., Washington University in St. Louis, St. Louis, MO, ghassan@seismo.wustl.edu; WIENS, D., Washington University in St. Louis, St. Louis, MO, doug@wustl.edu; WYSESSSION, M., Washington University in St. Louis, St. Louis, MO, michael@seismo.wustl.edu; SHORE, P., Washington University in St. Louis, St. Louis, MO, patrick@seismo.wustl.edu; VAN DER LEE, S., Northwestern University, Chicago, IL, suzan@earth.northwestern.edu; REVENAUGH, J., University of Minnesota, Minneapolis, MN, justinr@umn.edu; FREDERIKSEN, A., University of Manitoba, Winnipeg, MB, Canada, Andrew.Frederiksen@ad.umanitoba.ca; DARBYSHIRE, F., Université du Québec à Montréal, Montreal, QC, Canada, f.darbyshire@gmail.com; STEIN, S,, Northwestern University; JURDY, D., Northwestern University
Continental rifts are large-scale laboratories that can be used to study the initiation and cessation of spreading centers. The combination of seismometers from the SPREE deployment (Superior Province Rifting EarthScope Experiment) in Minnesota, Wisconsin, and Canada with those of the EarthScope USarray Transportable Array (TA) provides a unique spatial coverage of the mid-continent rift system (MCRS) as it impinges upon the Canadian Shield. The MCRS is roughly a 2000-km-long failed rift system, extending from Oklahoma to Canada, which formed about 1.1 billion years ago. Among the seismic methods we are using to investigate this region is a high-resolution ambient-noise surface-wave tomography study using empirical Green’s functions. Data available so far extend from May 2011 to October 2012. Ambient noise cross correlograms of each pair of stations are stacked and the group and phase velocities of the resultant empirical Green’s functions are measured at periods ranging from 10-70 s. The measured velocity values are inverted to produce two dimensional surface-wave velocity maps.The extension resulting from the formation of the MCRS was accompanied by the extrusion of gabbroic and granitic rocks at various locations along the rift axis that create a significant geophysical anomaly. This extrusive igneous activity was followed by the deposition of thick sediments. Preliminary ambient-noise tomography results show low shear-wave velocities at the shortest periods (corresponding to shallowest depths) along the rift valley, attributable to this clastic fill. At shallower depth, the low velocity anomaly is the widest over Lake Superior where the MCRS broadens and the anomaly changes orientation from NE-SW to NW-SE. At longer periods the velocity increases with the period along the MCRS, but remains lower than the surrounding terrain. This increase in velocity reflects the presence of mafic igneous rocks that fill much of the rift valley at mid-crustal levels.
Session: Induced Seismicity
Presenter   Brown, Wesley
Schedule   Fri 2:15 PM / Oral
Room   155D
Investigating the Cause of the 17 May 2012 M4.8 Earthquake near Timpson, East Texas
BROWN, W. A., Stephen F. Austin State University, Nacogdoches, TX, brownwa1@sfasu.edu; FROHLICH, C., Institute for Geophysics, University of Texas at Austin, Austin, TX, cliff@utig.ig.utexas.edu; BRUNT, M., Eagle Pass Junior High School, Eagle Pass, TX, mrbrunt59@hotmail.com; ELLSWORTH, W., Earthquake Science Center, U.S. Geological Survey, Menlo Park, CA, ellsworth@usgs.gov; LUETGERT, J. H., Earthquake Science Center, U.S. Geological Survey, Menlo Park, CA, luetgert@usgs.gov
The largest recorded earthquake in eastern Texas occurred in the early morning of 05/17/12 near Timpson, Texas. Following this earthquake, we installed three temporary seismographs and conducted felt report surveys in the epicentral area. Data from these instruments, from a permanent station in Nacogdoches (NATX), and from several temporary USArray stations allowed us to identify foreshocks and aftershocks, and to locate these events. The largest foreshock, a M3.9 event, occurred on 05/10/12 and was widely felt throughout east Texas. The best-recorded aftershock occurred on June 16th and its preferred location is about 3 km north of the highest-intensity area. S-P times are consistent with a depth of less than 4 km and indicate that all events originated from nearly the same focus. Published felt report surveys in local newspapers were followed up with telephone calls to respondents to confirm details and locations of their experience. To delineate the higher-intensity regions, we spent several days in the area interviewing local residents. We augmented these data with "did-you-feel-it" data provided by the USGS. The highest intensities of MMI VII occurred south of Timpson in a 10 sq km region where buildings suffered significant damage. Historical record indicates that in 1891 and 1981, long before injection began, M4.0 and M3.2 earthquakes occurred 80 km west, and 25 km southeast, of Timpson. These have been attributed to activity along the east-west trending Mt. Enterprise fault zone, situated within 25 km of the epicentral area. Thus we cannot rule out the possibility that the recent seismic activity has a natural origin. Three injection disposal wells are within 4 km of the epicentral region, one recording injection volumes of 42,750 cu m/mo since September 2006. Studies have shown, seismic activity is sometimes associated with wells having maximum monthly rates exceeding 24,000 cu m/mo, thus it is possible the Timpson earthquakes were induced by injection.
Session: Characterizing Active Faults for Seismic Hazard Assessments
Presenter   Rockwell, Thomas
Schedule   Thu 1:45 PM / Oral
Room   155B
Paleoseismology of the Aqua Tibia - Earthquake Valley Fault, Eastern Strand of the Elsinore Fault Zone
ROCKWELL, T. K., San Diego State University, San Diego, CA, trockwell@mail.sdsu.edu; AKCIZ, S., University of California, Irvine, Irvine, CA, sakciz@uci.edu; GORDON, E., San Diego State University, San Diego, CA
Information about the past behavior of large earthquake activity along the Elsinore Fault Zone (EFZ) (broadly including all of its parallel strands and cross-faults), arguably the least known type-A fault within the southern San Andreas Fault system, is critically needed in light of the recent 2010 El Mayor-Cucapah earthquake. The southernmost segment of the EFZ in the Coyote Mountains last ruptured in the past 300 years and was likely the first in the sequence that included the ca. 1750 (EFZ), 1892 (Laguna Salada Fault) and 2010 (Pescadores Fault, Borrego Fault, Paso Inferior and Superior faults). Could the next large EQ in southern California occur along the Agua Tibia - Earthquake Valley (ATEV) strand of the EFZ, as the 2010 earthquake has been inferred to have loaded the southern Elsinore Fault? In an attempt to better understand the past earthquake activity along the EFZ, we conducted a mapping and paleoseismic investigation along the central section of the ATEV fault, the eastern strand of the EFZ in southern California. Our preliminary mapping on high resolution Google Earth aerial imagery indicates that this strand is active and likely has a late Quaternary slip rate of 2-3 mm/yr, as the fault offsets basin deposits containing the Bishop Tuff (~750ka) by nearly 2 km. The ATEV fault, south of the rapidly growing Temecula Valley region, is characterized by geomorphic features typical of active strike-slip faults (e.g. sag ponds, pressure ridges, scarps, vegetation lineaments, linear drainages, offset geomorphic features). Our paleoseismic investigation, based on logging of a single 3 m deep, 25 m long trench and radiocarbon analysis of 13 detrital charcoal samples, all of which were in stratigraphic order within calibration uncertainty limits, indicates that the ATEV fault has ruptured three times in the past 2100 years, with the most recent rupture occurring in the past 300 years, similar in timing with the most recent rupture of the Coyote Mountain segment.
Session: Induced Seismicity
Presenter   Viegas, Gisela
Schedule   Fri 4:45 PM / Oral
Room   155D
Stress-Strain Conditions Associated with Induced Earthquakes
VIEGAS, G., Engineering Seismology Group Inc., Kingston, ON, Canada, gisela.fernandes@esgsolutions.com; URBANCIC, T., Engineering Seismology Group Inc., Kingston, ON, Canada, ted.urbancic@esgsolutions.com; BAIG, A., Engineering Seismology Group Inc., Kingston, ON, Canada, adam.baig@esgsolutions.com
Induced earthquakes are thought to have lower stress drops than natural-occurring tectonically driven earthquakes, although that relationship is difficult to establish because these earthquakes typically occur in different tectonic settings and under different conditions. To investigate differences in the source mechanisms of both types of events, we use seismicity recorded during a hydraulic fracturing completion in the Horn River Basin, BC, Canada, in which both type of events were observed. That is, micro-magnitude events characteristic of hydraulic fracturing activities which occur within the treatment formation and show tensile crack opening/closing components, and small-magnitude events, occurring below the treatment zone, associated with slip on pre-existing faults. The differences in source characteristics of these events can help identify the driving mechanism and local stress field conditions. To estimate source parameters, we use a spectral method that takes advantage of the commonality of the earthquake source and the availability of a very large number of recording sensors. The method simultaneously inverts the source spectra recorded at all sensors solving for corner frequency, seismic moment and whole-path attenuation, thus avoiding biasing the source and attenuation measurements with trade-off effects. We further improve the accuracy in the source measurements by imposing constraints derived from a statistical analysis of the attenuation results. Our results show that the events have an average low stress drop (2 MPa) when comparing to natural tectonic earthquakes (~17 MPa). We also observe a clear spatial separation of stress drops, with larger stress drop events occurring near and on the reactivated faults zones, indicating that strain is preferentially stored in these structures and that stress drop depends on the local stress conditions.
Session: Towards an Integrated Understanding of Slow Earthquakes: What We Know, What We Don’t Know, and How to Move Forward
Presenter   Walter, Jacob
Schedule   Fri 4:30 PM / Oral
Room   155A
Tremor, Slow Slip, and the Spatio-Temporal Evolution of Seismic Activity at the Fringes of a Megathrust Earthquake at the Nicoya Peninsula, Costa Rica
WALTER, J. I., Georgia Institute of Technology, Atlanta, GA, jakeiwalter@gmail.com; PENG, Z., Georgia Institute of Technology, Atlanta, GA, zpeng@gatech.edu; SCHWARTZ, S. Y., University of California, Santa Cruz, Santa Cruz, CA, susan@pmc.ucsc.edu; MENG, X., Georgia Institute of Technology, Atlanta, GA, xmeng7@gatech.edu; NEWMAN, A. V., Georgia Institute of Technology, Atlanta, GA, anewman@gatech.edu; PROTTI, M., Observatorio Vulcanológico y Sismológico de Costa Rica, Heredia, Costa Rica, marino.protti@gmail.com
The subduction megathrust interface, at the Nicoya Peninsula, exhibits a wide range of complex fault behavior, from recently discovered slow slip and tremor, numerous microearthquakes, to infrequent megathrust (> Mw 7) earthquakes. Over the last 10 years, an NSF-funded international and multi-University network of GPS stations and seismometers has detected at least 5 slow slip events accompanied with increased tremor activity. In contrast to other subduction zones, the tremor originates up-dip, down-dip, and within the seismogenic zone. On September 5, 2012, an Mw 7.6 earthquake occurred directly beneath this network, making the Nicoya Peninsula uniquely poised to investigate the wide range of fault behavior and spatio-temporal evolution of seismic activity around the mainshock. Preliminary matched-filter analysis using a template earthquake that precedes the mainshock by ~120 s indicates similar events occurring 20-40 min prior to the mainshock, as well as, immediately following the mainshock. So far we did not observe any accelerating foreshock activities. We are currently expanding this analysis with a broader catalogue of template events and utilizing matched-filter codes optimized for graphics processing units (GPUs). While detailed analysis of the foreshock/aftershock sequence is ongoing, the early aftershocks cluster in a distinct region that is immediately adjacent to regions that have undergone slow slip in past events. By systematically detecting all possible seismic events around the mainshock, we hope to gain better insight into the spatio-temporal transitions from stable sliding to stick-slip motion, and underlying physics of earthquake nucleation and interaction.
Session: Characterizing Active Faults for Seismic Hazard Assessments
Presenter   Sherrod, Brian
Schedule   Thu 2:00 PM / Oral
Room   155B
Characterizing Active Faults in a Glaciated Forested Landscape: Examples from the Puget Lowland, Washington
SHERROD, B. L., USGS, Seattle, WA, bsherrod@usgs.gov; BLAKELY, R. J., USGS, Menlo Park, CA, blakely@usgs.gov; KELSEY, H. M., Humboldt State University, Arcata, CA, Harvey.Kelsey@humboldt.edu; WEAVER, C. S., USGS, Seattle, WA, craig@ess.washington.edu
At least nine active fault zones cross portions of the Puget Lowland of Washington State, yet characterization of these faults is hampered by a series of vexing problems. First, marine embayments cover large parts of the region. Second, thick glacial deposits obscure all but the smallest vestiges of bedrock, making it difficult to pinpoint fault locations. Third, a thick carpet of vegetation conceals much of the landscape, preventing direct mapping of faults. To attack these problems, we employ a tripartite method: (1) high-resolution aeromagnetic surveys delineate magnetic anomalies potentially associated with active faults. Most Holocene surface-rupturing faults mapped thus far tend to closely follow linear aeromagnetic anomalies that reflect underlying basement structure; (2) LiDAR surveys evince potential fault scarps beneath the forest canopy along portions of the magnetic anomalies; (3) paleoseismic trenches across LiDAR scarps reveal active faults in the shallow subsurface and constrain the magnitude and timing of past earthquakes. Where anomalies cross shorelines, we quantify deformation associated with past earthquakes using stratigraphic techniques and microfossil-based paleoecology.A detailed study in the northern Puget Lowland illustrates our approach. Airborne magnetic data reveal a series of short-wavelength, low-amplitude, northwest-trending anomalies near Birch Bay, Washington. One anomaly coincides with an anticline in the shallow subsurface seen in seismic reflection data, suggesting that the anomaly has a tectonic origin. LiDAR surveys in the same area reveal evidence of raised Holocene coastal landforms. Subsequent paleoseismological studies at sites straddling the fault show as much as 2 m of uplift at one site and a lesser amount of submergence at a site nearby. The application of this tripartite approach helps constrain the location, length, and timing of past earthquakes on this and other active faults in the Cascadia fore arc.
Session: Seismic Hazards and Ground Motions
Presenter   Tasan, Hilal
Schedule   Fri 1:45 PM / Oral
Room   155B
STUDENT
Testing Probabilistic Seismic Hazard Assessments Against Accelerometric Data in France
TASAN, H., Institut des Sciences de la Terre, Grenoble, France, hilal.tasan@ujf-grenoble.fr; BEAUVAL, C., Institut des Sciences de la Terre, Grenoble, France, celine.beauval@obs.ujf-grenoble.fr; HELMSTETTER, A., Institut des Sciences de la Terre, Grenoble, France, agnes.helmstetter@ujf-grenoble.fr; GUEGUEN, P., Institut des Sciences de la Terre, Grenoble, France, pgueg@obs.ujf-grenoble.fr
Probabilistic seismic hazard (PSH) calculations rely on several models and assumptions. Understanding the impact of the uncertainties of PSH components on the final output of the PSHA is not straightforward. It is essential to look for independent data to constrain the PSHA results. In the present study, we test PSH models using accelerometric data at rock sites.Three different sets of PSHA maps established for the French territory (MEDD 2002, AFPS 2006, SIGMA (Alps) 2012) are tested. Two types of datasets are used for the tests: 1) a refined accelerometric dataset from the French Accelerometric Network (RAP) that provides 15 years of data; 2) a synthetic dataset inferred from the instrumental LDG earthquake catalogue, complete over 34 yrs for ML≥2.5 (combined with a GMPE).The distribution of the expected number of sites with exceedance (according to the PSHA model) is obtained by Monte Carlo sampling and compared to the observations. The testing is performed for a wide range of accelerations (and corresponding return periods).Considering a network of stations covering France, observations are on average coherent with PSHA estimates of AFPS 2006 (testing performed for 0.033-0.09g). Moreover, the use of the synthetic dataset shows that 15 years are long enough for testing these levels of accelerations. The results for MEDD 2002 are only available for classical return periods of interest in earthquake engineering (≥100 years). The testing at such long return periods does not permit to reject the model, but does not validate it either (0 exceedance is predicted). Moreover, when reducing the area sampled by considering only southeast France, the results suggest that the area is too restricted to permit testing, and the test of SIGMA2012 estimates are not conclusive. The methodology developed here will also be applied to historical intensity data, in order to test higher ground-motion levels and longer observation time windows.
Session: Towards an Integrated Understanding of Slow Earthquakes: What We Know, What We Don’t Know, and How to Move Forward
Presenter   Montgomery-Brown, Emily
Schedule   Fri AM / Poster
Room   Hall 1
Observations from a Small Aperture Array at Kilauea Volcano, HI
MONTGOMERY-BROWN, E. K., UW-Madison / USGS, Menlo Park, CA, emilymb1@gmail.com; THELEN, W., USGS - HVO, Hawaii Volcanoes National Park, HI, wthelen@usgs.gov; OKUBO, P., USGS - HVO,
We analyze data from the Hawaiian Volcano Observatory seismic network and a temporary small aperture array installed on Kilauea’s south flank from February 2012 to February 2013. The small aperture array consists of 16 three-component instruments (twelve Guralp L22, three CMG3-ESP and one 40T) and has a 500 m diameter footprint. During the May 28, 2012 slow slip event on Kilauea’s south flank, we observe increased regular volcano-tectonic seismicity on Kilauea’s south flank triggered by the slow slip event. We search for both volcanic and tectonic tremor with envelope cross correlations and by beamforming with the small aperture. We observe teleseismically triggered tremor at Kilauea in the two hours following the October 28, 2012 Masset, Canada earthquake and the January 5, 2013 Queen Charlotte Islands earthquake. The onset of the tremor is delayed several minutes after the teleseismic surface waves and therefore is not modulated by the teleseismic coda. There are several distinct tremor bursts that each last about 5-10 minutes.
Session: Data Products as Research Resources
Presenter   Herrmann, Robert
Schedule   Thu 3:45 PM / Oral
Room   155A
Evaluating Crustal Models
HERRMANN, R. B., Saint Louis University, St. Louis, MO, rbh@eas.slu.edu
Crustal models are used for earthquake and explosion location, modeling ground motion attenuation and the determination of source parameters through waveform modeling. The models are derived from travel times, the inversion of surface-wave dispersion and receiver functions and constrained by other geophysical parameters. Each use of a model is sensitive to just certain aspects of the model and typical models cannot be used for all purposes. The question to be addressed is whether tests exits that are independent of the data sets used to derive the model. This presentation will focus on testing whether models do or do not ahve sharp Moho's, whether models are adequate for source inversion, and hwo such models may be made available to the community.
Session: Broadband Seismic Observations on the Seafloor
Presenter   Dunham, Eric
Schedule   Wed AM / Poster
Room   Hall 1
Hydroacoustic Waves from Megathrust Ruptures
DUNHAM, E. M., Stanford University, Stanford, CA, edunham@stanford.edu; KOZDON, J. E., Naval Postgraduate School, Monterey, CA, jekozdon@nps.edu; LOTTO, G. C., Stanford University, Stanford, CA, glotto@stanford.edu
The occurrence of shallow coseismic slip in megathrust ruptures enhances seafloor uplift and tsunami excitation. Shallow slip also excites hydroacoustic waves in the ocean, which can be recorded on ocean bottom pressure sensors. We are exploring the possibility of using hydroacoustic waves, which arrive at near-shore sensors ~10 min prior to tsunami waves, for local tsunami early warning systems. To study this problem, we conduct two-dimensional dynamic rupture models of megathrust events. The simulations employ rate-and-state friction on the fault, realistic material structure (including fault and seafloor topography), and an ocean layer. The models presently capture seismic and hydroacoustic waves, and ongoing work will permit us to simultaneously model surface gravity waves (tsunamis). Preliminary results indicate pronounced excitation of ~10 s hydroacoustic reverberations in the ocean layer that appear as a dispersed wavetrain on seafloor pressure records. These modes involve motions of both the fluid and solid; at ~10 s period they have phase velocities close to the Rayleigh wave speed in the solid and group velocities close the sound wave speed in the ocean. Pressure signals at locations landward of the trench have amplitudes that correlate well with the amplitude of shallow slip and seafloor uplift (and presumably tsunami wave heights). Specific models of the 2011 Tohoku earthquake predict dynamic pressure changes ~0.1 to 1 MPa associated with these hydroacoustic waves at offshore stations; such waves were likely recorded on ocean bottom pressure sensors during that event.
Session: Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Models
Presenter   Jaume, Steven
Schedule   Wed AM / Poster
Room   Hall 1
Using Weak Motion Spectral Ratios to Predict Strong Motion Amplification in Charleston, South Carolina
JAUME, S. C., College of Charleston, Charleston, SC, jaumes@cofc.edu; MINER, K. S., College of Charleston, Charleston, SC, ksminer@g.cofc.edu
We have collected and processed ambient seismic noise for horizontal to vertical spectral ratios (HVSR) at over 90 sites in the greater Charleston, South Carolina region. The amplitude of the HVSRs in the 1-2 Hz frequency range correlates extremely well with both mapped surface geology and surface elevation of the Charleston peninsula. Modeling of HVSR using shear wave velocity profiles derived from geophysical and geotechnical data show that the HVSR amplitude in the 1-2 Hz range is controlled by a complex interplay between impedance contrasts at a range of depths beneath Charleston. Here we compare our weak motion amplification results to models of strong ground amplification conducted by Chapman et al. (2006). Our first point of comparison is between model results: a) both weak and strong ground motion model results suggest the greatest amplifications occur for Quaternary sediment thicknesses of 15 to 20 meters, with smaller amplifications for both thicker and thinner sediment, and b) both sets of models show the greatest amplifications for sites with VS30 ~240 m/sec, with the strongest correlation between weak and strong motion results occurring for strong motion models with larger (0.5 to 0.6 g) input accelerations. We also compared the observed HVSRs with the predicted strong motion amplification at 11 common sites. Here we find good to fair correlation, with weak motion HVSR over-predicting modeled strong motion by factors of 1.64 (0.1 g input) to 3.35 (0.6 g input). Contrary to the previous result, the best correlation between weak and strong motion amplifications occurs when the input acceleration to the strong ground motion model is small (0.1 to 0.2 g). We will explore other correlations between these data sets and explore their usefulness in predicting strong ground motion in Charleston.
Session: Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Models
Presenter   Seale, Sandra
Schedule   Wed AM / Poster
Room   Hall 1
Investigation of Seasonal Variations in the Response of the Soil-Foundation-Structure-Interaction Test Structure
SEALE, S. W. H., Earth Research Institute, UCSB, Santa Barbara, CA, sandy@eri.ucsb.edu; STINSON, E., Princeton University, Princeton, NJ, estinson@princeton.edu; STEIDL, J. H., Earth Research Institute, UCSB, Santa Barbara, CA, steidl@eri.ucsb.edu; HEGARTY, P., Earth Research Institute, UCSB, Santa Barbara, CA, hegarty@eri.ucsb.edu
The experimental soil-foundation-structure-interaction (SFSI) test structure installed at the NEES@UCSB Garner Valley field site is fully instrumented to record the response to ambient earthquakes. The SFSI experiment is unique in its ability to record in situ response of a structure in the field. A mobile shaker is installed on the underside of the roof slab of the SFSI. The shaker is run every night (60 sec sine sweep from 15 – 5Hz) and the response of the SFSI is recorded. Running this experiment daily has provided a valuable data set that documents the change in structural behavior with changes in the environment such as temperature and ground water level. Two years of nightly shaker test data were analyzed for behavior of the resonance of the structure. Fourier analysis shows a first mode (rocking) resonant frequency of ~5.6 Hz. That first resonance sometimes appears as two peaks in the amplitude spectrum. Comparing the resonance frequencies with the average daily temperature shows that the resonance frequencies are higher on warmer days. We expect that changing temperatures have the greatest impact on the steel members of the SFSI frame and that this, in turn, alters the resonance. It also appears that the “split” of the first resonance frequency is associated with higher temperatures. We also present a comparison of the rocking mode frequencies with water table depth.
Session: The Magnitude X.X Earthquake on the YY of ZZZZ: Major Earthquakes of 2012/13
Presenter   Lay, Thorne
Schedule   Fri 10:45 AM / Oral
Room   155A
The October 28, 2012 Mw 7.8 Haida Gwaii Underthrusting Earthquake and Tsunami: Slip Partitioning Along the Queen Charlotte Fault Transpressional Plate Boundary
LAY, T., University of California Santa Cruz, Santa Cruz, CA, tlay@ucsc.edu; YE, L., University of California Santa Cruz, Santa Cruz, CA, lye2@ucsc.edu; KANAMORI, H., California Institute of Technology, Pasadena, CA, hiroo@gps.caltech.edu; YAMAZAKI, Y., University of Hawai'i, Honolulu, HI, yoshikiy@hawaii.edu; CHEUNG, K. F., University of Hawai'i, Honolulu, HI, cheung@hawaii.edu; KOPER, K. D., University of Utah, Salt Lake City, UT, kkoper@gmail.com; KWONG, K. B., University of Utah, Salt Lake City, UT, kbkwong@gmail.com
The Pacific/North American plate boundary is undergoing predominantly right-lateral strike-slip motion along the Queen Charlotte and Fairweather transform faults, but there is about ~20 percent oblique convergence along the portion of the plate boundary from off the northwest coast of British Columbia to along the islands of Haida Gwaii. The Queen Charlotte Fault (QCF) hosted the largest historical earthquake in Canada; the 1949 Ms 8.1 strike-slip earthquake, with an epicenter off northern Haida Gwaii. There had been uncertainty in how the compressional component of plate motion along Haida Gwaii is being accommodated. The October 28, 2012 Mw 7.8 Haida Gwaii earthquake involved slightly oblique thrust faulting on a shallow (~20°) northeast-dipping fault plane with strike (~326°) parallel to the QCF, consistent with prior inferences of some Pacific Plate underthrusting beneath Haida Gwaii. The thrust motion had very large angle to the QCF, indicating significant interplate slip partitioning and possible decoupling of the underthrust segment of Pacific plate. The rupture extended to shallow depth offshore and generated a modest tsunami that spread across the northern Pacific and prompted a tsunami warning, coastal evacuation and beach closure in the Hawaiian Islands. The event was followed by a substantial aftershock sequence, in which almost all of the larger events involve normal faulting within the Pacific lithosphere extending about 30 km seaward of the QCF. The mainshock rupture appears to have spread with 2.5 km/s rupture velocity over a length of about 150 km, with slip averaging 3 m concentrated within 10 km from the surface. A purely strike-slip event occurred on January 5, 2013 (Mw 7.5) about 300 km to the northwest along the QCF, in a region where the plate motion parallels the plate boundary. How the accumulated boundary-parallel motion will be accommodated along Haida Gwaii is unclear, but future strike-slip faulting as in the 1949 event appears viable.
Session: ShakeMap-Related Research, Development, Operations, and Applications
Presenter   Bausch, Douglas
Schedule   Thu PM / Poster
Room   Hall 1
The Great Utah ShakeOut, ShakeMap and Hazus Applications
BAUSCH, D., FEMA Region VIII, Denver, CO, Douiglas.Bausch@dhs.gov; ROZELLE, J., FEMA Region VIII, Denver, CO, jesse.rozelle@fema.dhs.gov; PANKOW, K., University of Utah, Salt Lake City, UT, pankow@seis.utah.edu
The Federal Emergency Management Agency (FEMA) developed Hazus and released the first earthquake loss estimation version in 1997. Since then we have added flood loss and hurricane loss capabilities while continuing to improve the earthquake model. For more than a decade, a broad range of applications have emerged including the development of mitigation strategies, scenario driven catastrophic planning, exercise support, recovery and preparedness planning. This presentation will introduce several potentially valuable applications for earthquake scenarios, specifically our recent Great Utah ShakeOut www.shakeout.org\utah in April 2012 and 2013. The Utah application incorporated 20 ShakeMap scenario models http://www.shakeout.org/utah/scenarios/ from the University of Utah Seismograph Stations, liquefaction and landslide data from the Utah Geological Survey, as well as extensive updates to the Hazus inventory. The Hazus modeling analyses were updated with the 2010 census information, a site specific FEMA 154 vulnerable building inventory, a hospital inventory, Salt Lake County assessor data, and a detailed database of buildings on the University of Utah campus. In addition, we deployed the results in mapping interfaces that supported the development of a catastrophic plan, the functional exercise, and for the general public. The public outreach components included publication through media outlets, an interactive website and a new video “Utah: Preparedness Now”. Almost 1/3 of the State’s population participated in the drill and several hundred participated in the 72 hour exercise. Many were engaged as a result of successful communication concerning the potential impacts, including enhancing the realism for the exercise. The results also continue to support mitigation strategies, including addressing Utah’s large inventory of Unreinforced Masonry (URM) buildings.
Session: Triggering of Seismic and Volcanic Events
Presenter   Sevilgen, Volkan
Schedule   Fri 11:15 AM / Oral
Room   155D
Stress Imparted by the Great 2004 Sumatra Earthquake Shut Down Transform Faults and Activated Rifts up to 400 km Away in the Andaman Sea
SEVILGEN, V., Seismicity.net, San Carlos, CA, vsevilgen@usgs.gov; STEIN, R. S., US Geological Survey, Menlo Park, CA, rstein@usgs.gov; POLLITZ, F. F., US Geological Survey, Menlo Park, CA, fpollitz@usgs.gov
The cause and prevalence of triggered seismicity and remote aftershocks are under debate. As a result, they have been excluded from probabilistic seismic hazard assessment and aftershock hazard notices. The 2004 M=9.2 Sumatra subduction earthquake altered seismicity and hazard in the backarc rift-transform system in the Andaman Sea. Here we show that over a 300-km-long largely transform section of the backarc, M≥4.5 (the completeness magnitude) earthquakes stopped occurring for five years, and over a larger 750-km-long backarc section, the rate of transform events dropped by two-thirds, while the rate of rift events along this section increased eight-fold. We compute the propagating dynamic stress wavefield, and find the peak dynamic Coulomb stress is similar on the rifts and transforms. At long-period, dynamic stress amplitudes, which are thought necessary to promote dynamic failure, are higher on the transforms than on the rifts, opposite to the observed triggering. In contrast to the dynamic stress, we calculate that the mainshock brought the transform segments ~0.2 bar (0.02 MPa) farther from static Coulomb failure and the rift segments ~0.2 bar closer to static failure, consistent with the seismic observations. This accord suggests that changes in seismicity rate are sufficiently predictable to be included in post-mainshock hazard evaluations where the regional faulting system is known.
Session: Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Models
Presenter   Seale, Sandra
Schedule   Fri 11:30 AM / Oral
Room   155C
Modeling the Evolution of Pore Pressure Induced by Earthquakes: Results from the Brawley Swarm of August 2012
LAVALLEE, D., Earth Research Institute, UCSB, Santa Barbara, CA, daniel@eri.ucsb.edu; SEALE, S. W. H., Earth Research Institute, UCSB, Santa Barbara, CA, sandy@eri.ucsb.edu; ARCHULETA, R. J., Earth Research Institute, UCSB, Santa Barbara, CA, ralph@eri.ucsb.edu; STEIDL, J. H., Earth Research Institute, UCSB, Santa Barbara, CA, steidl@eri.ucsb.edu
Pore pressure built up during an earthquake and the hazard associated with soil liquefaction present a major challenge for our society. There is consensus among scientists that a better assessment of the liquefaction risk requires a better understanding of the coupling between pore pressure and ground motion time histories. An excellent data set was provided by the swarm events near Brawley, CA, in August 2012. Over a period of three days, hundreds of events were recorded at the NEES@UCSB Wildlife Liquefaction Array (WLA); 11 of those events generated pore pressure increases > 0.1 kPa. WLA is located at the southern end of the San Andreas Fault, on the banks of the Alamo River, at a site that liquefied in the Superstition Hills event of 1987. The instrumentation is composed of strong-motion accelerometers located at the surface and in boreholes. There is also an array of pore pressure transducers located in boreholes of different depths in the saturated, sandy layer. The correlation between the sudden onset of pore pressure increase in the data with the arrival of the large S-wave pulse led to the application of the heat diffusion equation to model the pore pressure increase. The basic assumption is that a pulse located in time and space (the arrival of the S-wave) can induce a sudden increase in pore pressure. Under this assumption, the time evolution of the pore pressure can be modeled by solving the heat diffusion equation with a forcing function where the forcing function is given by a pulse. Using data generated by a 4.9 event (Ru > 60%), pore pressure records were sampled at 5 sec (0.2 Hz) for estimation of the model parameters. The low-frequency response (“step” behavior and subsequent decrease) of the pore pressure data is well described by this model at different depths in the saturated layer.
Session: Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Models
Presenter   Dalguer, Luis
Schedule   Wed AM / Poster
Room   Hall 1
Propagation of 1-Point and 2-Point Statistics from Dynamic Source through Kinematic to Ground Motions
SONG, S. G., Swiss Seismological Service (SED), ETH Zurich, Zurich, Switzerland, song@sed.ethz.ch; DALGUER, L. A., Swiss Seismological Service (SED), ETH Zurich, Zurich, Switzerland, dalguer@sed.ethz.ch
The complexity of finite source processes control near-source ground motion characteristics significantly, especially for large events. Dynamic rupture modeling has been successfully adopted for physics-based source and ground motion simulations recently. However, the required input parameters of stress and frictional properties are generally not well constrained in the modeling. Following Song and Dalguer (2013), we systematically perturbed the heterogeneity of dynamic input parameters, especially stress drop, in the framework of 1-point and 2-point statistics, and performed dynamic modeling with high performance computing (HPC). After investigating kinematic source parameters and near-source ground motions in the same format of 1-point and 2-point statistics, we found that 1-point statistics of dynamic input parameters, which are often not explicitly considered in source modeling, plays a significant role in determining the characteristics of kinematic source and ground motion parameters. For example, larger standard deviation of input stress drop produces heavier upper tails in 1-point statistics of kinematic source parameters. Strong cross-correlations were also observed between kinematic source parameters and correlation patterns systematically change with increasing standard deviation of the stress drop. The increase in the standard deviation of stress drop also produces much stronger ground motions particularly in the forward rupture directivity region. We also found that the effect of the standard deviation is amplified or de-amplified, depending on 2-point statistics of stress drop. This study demonstrates that it is important to constrain input dynamic parameters for accurate simulation-based ground motion predictions, and we can efficiently quantify the characteristics of both earthquake source and ground motions in the framework of 1-point and 2-point statistics and investigate their relations in a consistent way.
Session: Earthquake Source Studies
Presenter   Bent, Allison
Schedule   Wed PM / Poster
Room   Hall 1
Source Properties and Effects of Two Recent, Moderate Earthquakes near Montreal
BENT, A. L., Geological Survey of Canada, Ottawa, ON Canada, bent@seismo.nrcan.gc.ca; PECI, V., Geological Survey of Canada, Ottawa, ON Canada, vpeci@nrcan.gc.ca; HALCHUK, S., Geological Survey of Canada, Ottawa, ON Canada, shalchuk@nrcan.gc.ca; HAYEK, S., Geological Survey of Canada, Ottawa, ON Canada, shayek@nrcan.gc.ca; HAYEK, K., Geological Survey of Canada, Ottawa, ON Canada, khayek@nrcan.gc.ca; MAJEWSKI, C., Geological Survey of Canada, Ottawa, ON Canada, cmajewsk@nrcan.gc.ca
On 10 October 2012 and again on 6 November 2012 the Montreal region was shaken by moderate earthquakes. Neither is known to have caused any damage but both received considerable attention. The earthquakes occurred within the West Quebec Seismic Zone, a region of moderate earthquake activity with occasional larger earthquakes. The first event (mN 4.5, MW 3.7) occurred near St.-Amable 35 km northeast of Montreal. More than 10000 people submitted felt reports to the Geological Survey of Canada’s (GSC) “did you feel it page”. The northwest striking thrust mechanism determined by regional CMT inversion and by first motions is typical of the region. The depth determined by CMT inversion and the regional depth phase method (RDPM) is 13-15 km, also typical of the region. The second event (mN 4.2, MW 3.9) occurred close to Hawkesbury, Ontario on the Quebec-Ontario border. It too was a northwest striking thrust event. The depth as determined by both the CMT and RDPM is 6-8 km, shallow but within the observed range for the region. Over 2500 online felt reports were received. The recent implementation of new methods of analysis have allowed us to study these moderate events in more detail than we could have in the past and may provide insight into what to expect from future earthquakes in the region as well as providing more detailed input for refining hazard assessment. Their proximity to large, urban areas proved a good test of changes made to improve the robustness of our website during times of heavy traffic.
Session: Velocity Models and Modeling
Presenter   Mostafanejad, Akram
Schedule   Thu AM / Poster
Room   Hall 1
STUDENT
The Teleseismic P-wave Palindrome for Average P and S Sediment Velocity: H/V ^ V\H Power Spectral Ratios
MOSTAFANEJAD, A., CERI, University of Memphis, Memphis,TN, mstfnjad@memphis.edu; LANGSTON, C., CERI, University of Memphis, Memphis, TN, clangstn@memphis.edu
Teleseismic P waves have nearly vertical ray paths in thick, unconsolidated sediments, such as those occurring in the Mississippi embayment of the central U.S., and create very large Ps conversions/reverberations on the radial component and large P reverberations on the vertical component due to the high velocity contrast basement-sediment interface. We use this dichotomy of wave propagation between the vertical and horizontal components of motion to examine S and P wave resonances in the sediment layer using power spectral ratios, similar to Nakamura's empirical H/V method, by calculating horizontal-to-vertical (H/V) and vertical-to-horizontal (V/H) power spectral ratios. Using Dart’s map of sediment thickness as a basis, a self-consistent model of average compressional and shear-wave velocity versus sediment thickness was developed for the embayment by utilizing the theoretical linear relationship between the frequency of the H/V and V/H peak and shear-wave and compressional-wave velocities, respectively. Teleseismic data used in this study come from the USArray transportable array, flex array stations of the Northern Embayment Lithosphere Experiment, and stations of the New Madrid Cooperative Seismic Network. The shear wave velocity model obtained from the teleseismic waveforms matches very well with the model obtained from a previous H/V ambient noise study in the region. The velocity model shows low average velocity near the edge of the embayment with average velocities increasing with increasing sediment thickness, consistent with increased sediment compaction. This velocity model is useful for geotechnical applications and for starting models in more detailed teleseismic transfer function studies.
Session: Earthquake Source Studies
Presenter   Ohrnberger, Matthias
Schedule   Wed PM / Poster
Room   Hall 1
Classifying Seismic Signals at Small-Aperture Arrays via Stochastic Modeling of f-k-Image Sequences
OHRNBERGER, M., University of Potsdam, Potsdam, Germany, mao@geo.uni-potsdam.de; HAMMER, C., University of Potsdam, Potsdam, Germany, conny.hammer@geo.uni-potsdam.de; GIANNOTIS, N., University of Potsdam, Potsdam, Germany, nikos.giannotis@geo.uni-potsdam.de; SCHWEITZER, J., NORSAR, Kjeller, Norway, johannes.schweitzer@norsar.no
Today's availability of huge amounts of continuous seismic data streams has influenced seismological observatory practice. Besides classic earthquake analysis, seismologists now also embrace the wealth of information exploitable from the continuous seismic wave field. Thus, there is strong need for robust automatic analysis systems that allow detecting and flagging waveform windows of interest from it. Standard detection algorithms are tuned to onset time detection of phase arrivals. Classification of some waveform of interest as a whole is less well established and often restricted to example pattern matching using cross-correlation techniques. Stochastic modeling techniques like Hidden Markov Modeling (HMM) are conceptually better suited to account for within-class variability of seismic waveforms usually requiring large training data sets that are difficult to obtain. We present a HMM based classification scheme to detect complex and diffuse waveform classes from small sample training sets. The approach is based on modeling the observed sequences of short-term f-k images computed from continuous array recordings as a seismologically meaningful wave field representation. For robust training of system parameters from this base representation we reduce the dimensionality using information preserving transformations (heuristics and/or learned from data). We test our system at array SPITS (Svalbard, Norway) targeting on the classification of high frequency glacier-related seismic signals that occur frequently in arctic environments. The purpose is two-fold: 1) monitor the seismic activity of glaciers for long term correlation with climate change; 2) clean the detection bulletins from glacier related events to allow analysts to focus on events of other origin. We find the overall classification rate of the system promising and comparable to expert judgement. However, quantitative assessment is difficult given the lack of ground truth data.
Session: Earthquake Source Studies
Presenter   Horton, Stephen
Schedule   Wed PM / Poster
Room   Hall 1
Seismological Observations Associated with the Development of a Sinkhole near the Napoleonville Salt Dome, Louisiana
HORTON, S. P., CERI, University of Memphis, Memphis, TN, shorton@memphis.edu; ELLSWORTH, W. L., U. S. Geological Survey, Menlo Park, CA, ellsworth@usgs.gov; RUBINSTEIN, J. L., U. S. Geological Survey, Menlo Park, CA, jrubinstein@usgs.gov; WITHERS, M., CERI, University of Memphis, Memphis, TN, mwithers@memphis.edu
On 06/08/12 and again on 07/03/12, strong shaking was reported by residents of the small community of Bayou Corne (adjacent to the Napoleonville salt dome), Louisiana. The U.S. Geological Survey and the University of Memphis deployed a narrow-aperture network of six broadband seismic stations to monitor activity. Real time data has been transmitted to USGS National Earthquake Information Center and archived at IRIS Data Management Center since the first stations became operational on 07/14/12. Initial seismograms showed that sharp tremors - similar to long period volcanic earthquakes with emergent body waves and well-developed surface waves lasting 10-15 seconds - were occurring in the northwest portion of the Napoleonville salt dome at a rate of 10-20 events per day. Preliminary depth estimates of the sharp tremors are less than 1km. The recent addition of a borehole sensor above a failed salt-cavern confirms the shallow depths of these LP events. The maximum energy is between 0.4 - 6Hz. On the morning of 07/24/12, the rate increased to hundreds of events per day. This intense tremor activity was accompanied by episodes of spasmodic bursts (perhaps overlapping sharp tremors) lasting 5-10 minutes. The intense tremor activity continued until 08/02/12, when the sharp tremors appeared to cease. The subsequent morning a sinkhole (~ 180m across and 128 m deep) was discovered on the edge of the salt dome adjacent to the seismic activity. After several months, sharp tremors are again being observed at a low rate. A very long period (VLP) signal (25-33s) - that is not directly related to the sharp tremors in time - began to occur within the network on 08/06/12. Polarization analysis indicates the VLP signals radiate from a point just northwest of the sinkhole. The VLP signals are low amplitude with duration about 5 minutes. The sinkhole and the tremors are within 1km of 3 natural gas pipelines, 3 gas storage caverns and 4 active brine-mining cavities in the salt dome.
Session: Intermountain West Earthquake and Volcano Characterization and Related Hazards
Presenter   Archuleta, Ralph
Schedule   Wed 2:45 PM / Oral
Room   155B
Broadband Ground Motion in the Salt Lake City Basin Inferred from a Combined Dynamic and Kinematic Simulation of a M 6.8 Earthquake on the Wasatch Fault
ARCHULETA, R. J., University of California, Santa Barbara, Santa Barbara, CA, ralph@eri.ucsb.edu; LIU, Q., University of California, Santa Barbara, Santa Barbara, CA, qliu@eri.ucsb.edu; SMITH, R. B., University of Utah, Salt Lake City, UT, robert.b.smith@utah.edu
The most significant seismic hazard for Salt Lake City is the segment of the Wasatch Fault that lies near the western edge of the city. This segment is capable of producing earthquakes M 7 and larger. To estimate the ground motion from a future earthquake we simulate earthquakes that might occur. Using the mapped fault trace for a reference we construct an irregular fault surface that dips 50 degrees. The fault is embedded in the 3D velocity structure for the area. We compute the slip history everywhere on the fault for a dynamically propagating rupture. We consider scenarios where the normal stress is constant on the fault or varying with depth as well as having different hypocentral locations. We also consider scenarios where the rupture can jump, or not, a major offset in the fault’s strike. The offset can have a major impact on size of the earthquake. The empirical prediction equations envelope the predicted peak ground velocities (lowpassed at 1.0 Hz) for all distances away from the fault. Because the dynamic models are limited to frequencies less than 1.0 Hz, we extend the frequency band of the ground motion by using a kinematic model. To estimate ground motions at higher frequencies we use the dynamic calculations to provide the information of slip and rupture time on the fault. We fit a functional form to the slip rate function at each point on the fault. With this slip rate function, timing of the rupture and approximating the basin structure we use the representation theorem to compute the ground motion throughout the basin using the method of Schmedes and others (JGR, 2010; GJI, 2013).
Session: Triggering of Seismic and Volcanic Events
Presenter   Milner, Kevin
Schedule   Fri 11:45 AM / Oral
Room   155D
Probabilities and Synchronization of Earthquake Sequences on Southern California Faults from a Million-Year RSQSim Catalog
MILNER, K. R., SCEC, University of Southern California, Los Angeles, CA, kmilner@usc.edu; JORDAN, T. H., SCEC, University of Southern California, Los Angeles, CA, tjordan@usc.edu; RICHARDS-DINGER, K. B., University of California Riverside, Riverside, CA, keithrd@ucr.edu; DIETERICH, , University of California Riverside, Riverside, CA, james.dieterich@ucr.edu
Southern California is a complex fault system comprising many fault segments with along-strike slip-rate variability, strike changes, and multiple branching. These heterogeneities contribute to complex interactions between neighboring segments. We use a one-million-year synthetic earthquake catalog generated by the RSQSim physics-based simulator (Dieterich and Richards-Dinger 2010) to explore these interactions in the Southern California fault system. An interesting example is the Mojave section of the SSAF, where, in the first week following large (M7+) events, the average rate of equally large earthquakes increases by two orders of magnitude, and almost 3 orders of magnitude for M7.5+ events. The rate gains conform to magnitude-frequency distributions that are characteristic rather than Gutenberg-Richter: the rates of M7.5+ aftershocks are significantly higher than M7.0+ aftershocks. We quantify the spatial distribution of the hazard increase in the 1st year following initial events by analyzing participation rate gains on neighboring faults. For initial events on the Mojave section, the rate gains are highest on neighboring sections (suggesting an “unzipping” of the SSAF). In particular, the participation rate of the Coachella section in magnitude 7+ events increases tenfold, from ~10^-2.3 to ~10^-1.3. We also observe synchronization between Mojave and other Southern California faults, most notably with the Coachella section which ruptures with or within 5 years of the Mojave section 54% of the time. Coachella and Carrizo appear to be intermittently locked out of phase, often rupturing with Mojave in an alternating sequence and only rupturing within 5 years of each other 11% of the time. Garlock and San Jacinto are not as strongly synchronized with the Mojave section, but do show some evidence of phase locking at 30% and 21%. These correlations represent potentially drastic changes in the short term earthquake hazard following large events in Southern California.
Session: Data Products as Research Resources
Presenter   Smith, Deborah
Schedule   Fri AM / Poster
Room   Hall 1
Developing an End-User Application for Generating 3D Stochastic Spatially Variable Stress
SMITH, D. E., Carnegie Institution of Washington, Washington, DC, desmith144@gmail.com; GOLDEN, S., Carnegie Institution of Washington, Washington, DC, sgolden@ciw.edu
Understanding the 3D crustal stress field and its spatial variation for differentregions in the Earth’s crust is a critical component for any earthquake hazardassessment or dynamic rupture modeling. Given that tectonically active areas can begeometrically complex, filled with fracture networks at different scales on whichdislocations occur, and the slip in earthquake dislocations can be quite variablespatially, it is reasonable to assume that stress is spatially variable orheterogeneous at some scale.Based on the model of Smith and Heaton [2011], but with expanded options, we havebegun building a downloadable software application that allows the end-user togenerate their own stochastic spatially-variable stress. We will show features andinitial results of this new application. The application is intended to be utilizedby scientists who wish to add a 2D stochastic stress overlay to their fault planesfor dynamic rupture simulations or add a 3D stochastic overlay to their regionalstress models. As in Smith and Heaton [2011], the model is controlled by two primarystochastic parameters: 1) alpha, which is the linear fall-off of the spectralamplitude as a function of spatial frequency along any 1D line and 2) HR, which isthe relative amplitude of the spatial variation to the spatial mean at some outerscale.The new application is written in C/C++, leading to runtime performance gains of anorder of magnitude or more compared to an original MATLAB prototype. Furtheroptimizations reduce its memory use, a limiting factor for stress model size. Theapplication also includes some new features, such as an option to generate adivergence-free 2D and 3D stochastic stress, which is particularly important fordynamic rupture simulations. Last, it may include code to generate synthetic focalmechanism catalogs similar to Smith and Heaton [2011] but with more options.
Session: Velocity Models and Modeling
Presenter   Zhao, Lian-Feng
Schedule   Thu AM / Poster
Room   Hall 1
Lg-Wave Attenuation Tomography in Middle East Using Single-Station, Dual-Station and Dual-Event Data Sets
ZHAO, L., Key Lab of the Earth's Deep Interior, IGGCAS, Beijing, China, zhaolf@mail.iggcas.ac.cn; XIE, X., IGPP, University of California at Santa Cruz, Santa Cruz, CA, xxie@ucsc.edu
We adopt an innovative Lg-wave Q tomography technique to simultaneously invert the Lg-wave attenuation distribution, source excitation function, and site response in Middle East. Broadband seismograms recorded at 144 stations form 318 crustal earthquakes are used to calculate the Lg-wave amplitude spectra at 58 discrete frequencies distributed log evenly between 0.05 and 10.0 Hz. At each frequency, the single-station, dual-station and dual-event datasets are formed for inversion. The single-station data has the highest ray coverage which benefits the inversion resolution. The dual-station data involve less tradeoff between the attenuation and source term. Similarly, the dual-event data eliminate the ambiguity between the station site response and the attenuation. However, the latter two data sets generate less useful data and make the high-resolution inversion difficult. By combining different data sets together, we obtain a high-resolution broadband Lg-wave attenuation model for the Middle East and its surrounding region. The maximum spatial resolution is approximately 1.0x1.0 degree in well-covered areas and for frequencies between 0.05 and 1.5 Hz (mostly for Tianshan mountains and Pamir plateau). The Lg Q images reveal the relations between attenuations and geological structures. From Turkish and Iran Plateau to Pamir Plateau, the continental-collision orogenic belt is characterized by a relatively low average Qo (1 Hz Lg Q) of 333 (253-438). Higher Qo values are found in stable regions: 578 (492-679) for Eurasian Plate, 549 (493-611) for Arabian Plate and 557 (521-595) for Indian plate, respectively. We investigate the frequency dependence of the Lg Q in different geology formations. The result illustrates that the Lg Q values generally increase with rising frequencies but show complex regional variations. The features of regional attenuation suggest that the strong crustal attenuation is mainly caused by the tectonic and thermal activities.
Session: The Magnitude X.X Earthquake on the YY of ZZZZ: Major Earthquakes of 2012/13
Presenter   Hayes, Gavin
Schedule   Fri PM / Poster
Room   Hall 1
Teleseismic Rupture Modeling of Recent Large Earthquakes
HAYES, G. P., USGS NEIC, Golden, CO, ghayes@usgs.gov; BENZ, H. M., USGS NEIC, Golden, CO, benz@usgs.gov; JI, C., University of California Santa Barbara, Santa Barbara, CA, ji@geol.ucsb.edu; MENDOZA, C., UNAM, Mexico City, Mexico, cmendoza@geociencias.unam.mx; HARTZELL, S., USGS GHSC, Golden, CO, shartzell@usgs.gov
Finite Fault Modeling forms a critical part of near real-time response to large earthquakes at the USGS National Earthquake Information Center (NEIC). After such events (generally M>7), we use high-quality, teleseismically-distributed broadband data to invert for the slip distribution of the earthquake following two approaches. For quick solutions that capture the first-order characteristics of the rupture, we use a P-wave inversion based on Mendoza et al. (2012). This method allows us to rapidly characterize key parameters such as approximate rupture length, slip amount and rupture duration, but is less sensitive to the details of slip distribution and rise times, and fixes other important parameters such as rupture direction and rupture velocity. To explore the details of the earthquake rupture, we follow the approach of Ji et al. (2002), using body-waves (P and SH) and long period surface waves (Rayleigh and Love) to simultaneously solve for slip amount, direction, duration, and initiation over an assumed fault plane geometry. This leads to a generally more accurate fault slip model, but requires a slightly longer time to generate than the preliminary solution (typically a few hours versus an hour or less). In the early stages of response, before or in parallel to our preliminary fault inversions, we also use an empirical Greens Function approach to derive the source time function of the earthquake; these are useful to constrain subsequent fault models.Here we compare results from each of these inversion approaches for a selection of recent large earthquakes in the 2012-2013 period: the August 2012 El Salvador earthquake (Mw 7.3); the September 2012 Philippine Islands earthquake (Mw 7.6); the September 2012 Costa Rica earthquake (Mw 7.6); the October 2012 Haida Gwai earthquake (Mw 7.8); the November 2012 Guatemala earthquake (Mw 7.4); and the January 2013 Haida Gwaii earthquake (Mw 7.5).
Session: Oceanographic and Atmospheric Signals in Seismology
Presenter   Moni, Aishwarya
Schedule   Fri AM / Poster
Room   Hall 1
Separation and Location of Microseisms
MONI, A., University College Dublin, Dublin, Ireland, aishwarya.moni@ucd.ie; CRAIG, D., University College Dublin, Dublin, Ireland, david.craig@ucd.ie; BEAN, C. J., University College Dublin, Dublin, Ireland, chris.bean@ucd.ie
Ocean gravity waves are driven by wind and atmospheric pressure systems and generate pressure changes at the sea bed. These pressure fluctuations generate continuous background seismic noise, called ‘microseisms’, which are associated with ocean wave activity and are generally stronger in coastal areas, although they are recorded on terrestrial seismic stations throughout the world. Background seismic noise levels increase during periods of increased ocean wave activity. Multiple spatially separate noise sources may be coincidentally active. This paper deals with the separation and location of microseisms. The algorithm used to separate data here is DUET (Degenerate Unmixing Estimation Technique). DUET is a method developed to separate simultaneous human voices from recordings from two sensors situated less than half the wavelength of the signal apart. Sources that are non-overlapping in the time-frequency domain are separated based on the time differences of arrival of the signals at the two sensors. This method has been successfully used to separate data in a volcano setting. The result from this algorithm is a 3-component seismogram at each station for each source.The method used to compute the direction of arrival of the sources of microseisms here is polarisation assuming Rayleigh wave arrivals (direction of arrival for which there is a 90 degree phase difference between the horizontal and vertical components). These methods were used to separate and locate multiple simultaneous sources of microseisms in a synthetic model, and then applied to field recordings in Ireland, in the North-East Atlantic. Results from the synthetic tests show that source separation prior to location improves the accuracy in location using single seismic stations. This will help improve transfer functions developed between land-based recordings of microseisms and wave parameters recorded at ocean buoys.
Session: Broadband Seismic Observations on the Seafloor
Presenter   Frassetto, Andy
Schedule   Wed AM / Poster
Room   Hall 1
Seismometer Orientations and Ambient Noise Observations for Year 1 of the Cascadia Initiative OBS Deployment
EVERS, B., Incorporated Research Institutions for Seismology, Washington, DC, brent.evers@iris.edu; FRASSETTO, A., Incorporated Research Institutions for Seismology, Washington, DC, andyf@iris.edu; ADINOLFI, A., Incorporated Research Institutions for Seismology, Washington, DC, andrew.adinolfi@iris.edu; WOODWARD, R., Incorporated Research Institutions for Seismology, Washington, DC, woodward@iris.edu; BABCOCK, J., Scripps Institution of Oceanography, San Diego, CA, jbabcock@ucsd.edu; BARCLAY, A., Lamont-Doherty Earth Observatory, Palisades, NY, barclay@ldeo.columbia.edu; COLLINS, J., Woods Hole Oceanographic Institution, Woods Hole, MA, jcollins@whoi.edu
The Cascadia Initiative ("Cascadia") is a National Science Foundation (NSF) American Recovery and Reinvestment Act (ARRA) funded project that encompasses a community designed and administered seismological and geodetic experiment addressing major geological, geophysical and geodetic questions specific to the Juan de Fuca plate system and the Cascadia subduction zone. A key element of the Cascadia Initiative is an amphibious array of seismometers deployed throughout the region, for which three Ocean Bottom Seismograph Instrument Pool (OBSIP) Institutional Instrument Contributors [IIC's: Woods Hole Oceanographic Institution (WHOI), Scripps Institution of Oceanography (SIO), Lamont-Doherty Earth Observatory (LDEO)] constructed 60 instruments for use starting in 2011. These instruments are in their second year of deployment to occupy a broad footprint spanning nearly the entire width of the Juan de Fuca plate and length of the Cascadia subduction zone from Vancouver Island to northern California.In an effort to make the Cascadia dataset available and useful to the widest possible number of investigators, the OBSIP Management Office at IRIS has calculated the horizontal orientations of the Cascadia instruments for the first year of deployment. In addition, we have produced probability density functions of power spectral density estimates of the continuous time series of recorded data at all stations to characterize the ambient noise of the ocean floor, which shows variations that vary seasonally and with water depth. Our preliminary estimates show that despite being an environment with a high level of ambient noise at long periods, reliable orientations have been obtained for most stations across the array. Furthermore, the overall data return for the OBS instruments is similar to that of many on-land portable deployments.
Session: When and Why do Earthquake Ruptures Stop? Evaluating Competing Mechanisms of Rupture Termination
Presenter   Galgana, Gerald
Schedule   Fri PM / Poster
Room   Hall 1
Locking and Present-day Slip Accumulation along the Nankai Trough, Japan: Analyses using Kinematic Models and Coseismic Ruptures
KLEIN, E. C., AIR Worldwide, Boston, MA, eklein@air-worldwide.com; GALGANA, G., AIR Worldwide, Boston, MA, ggalgana@air-worldwide.com; SHEN-TU, B., AIR Worldwide, Boston, MA, BShen-Tu@air-worldwide.com; MAHDYIAR, M., AIR Worldwide, Boston, MA, MMahdyiar@air-worldwide.com
Southwestern Japan is a rapidly deforming island arc due to the oblique convergence between the Philippine Sea plate and Eurasian plate. The potential for great megathrust earthquakes in this region is linked to the history of subduction along the Nankai Trough and has major implications for seismic hazard assessment. We investigate crustal deformation within the Japanese forearc and determine fault locking patterns along subduction zones using block models of the region. We use published geodetic data and detailed slab geometries of the subduction zones to predict block motions and coupling patterns along subduction zone interfaces. In addition, we use published coseismic slip distributions from past earthquakes and modeled rates of slip accumulation to estimate the distribution of present day slip accumulation for the Nankai trough. To quantify expected present day slip accumulation since the last great earthquake rupture, we analyze patterns of total slip accumulation using the block models and combine this with slip distributions from historic earthquake ruptures. Overall, we find that the Nankai Trough is highly coupled in at least two distinct regions (interpreted to be asperities) along the subduction zone interface, in agreement with previous studies. Our results suggest that portions of the Nankai Trough region may have accumulated as much as nine meters of slip since the last series of great earthquakes. We find that the total accumulated slip and tectonic moment along the southern segment of the Nankai trough is high, implying that the risk of a large event in the Nankai segment may be as high as that in the Tokai segment in the north.
Session: Triggering of Seismic and Volcanic Events
Presenter   Linville, Lisa
Schedule   Fri PM / Poster
Room   Hall 1
STUDENT
An Automated Algorithm to Detect Remotely Triggered Aftershocks Recorded by EarthScope's Transportable Array and Regional Seismic Networks: A Case Study of Four Large Earthquakes
LINVILLE, L. M., University of Utah, Salt Lake City, UT, linville@seis.utah.edu; PANKOW, K. L., University of Utah, Salt Lake City, UT, Pankow@seis.utah.edu; KILB, D. L., University of California, San Diego, CA, dkilb@ucsd.edu; VELASCO, A. A., University of Texas, El Paso, TX, aavelasco@utap.edu
The density of stations (~70 km station spacing) and broad coverage of the EarthScope Transportable Array (TA), supplemented with regional seismic network (RSN) data, enable the collection of an ideal dataset (albeit large and cumbersome for processing) for studying dynamic triggering. Digesting this large quantity of data in a reasonable manner requires an automatic processing and interpretation schema. In this study, for a target mainshock event, we examine ten hours of waveform data (±5 hours from the mainshock) from all TA and RSN stations. We high-pass filter these data at 5 Hz, and apply an optimized detection algorithm that uses Antelope’s short-term average (STA) to long-term-average (LTA) ratio to create a catalog of ‘detections’ at each station as a function of time. A detection is simply a signal above the noise level, which may, or may not, be an earthquake. The number of detections in the pre- and post-five hour windows are compared using a Poisson distribution. Stations with a significant increase (> 95% confidence level) in detections from the pre- to post-five hour window and having a minimum of five detections in the post-five hour window are flagged for use in a more detailed investigation. Following the automated processing, the flagged waveforms are individually analyzed, in both the time and frequency domains, to determine if the increase in detections correspond to local earthquakes (i.e., potentially remotely triggered aftershocks). We show results using this automated schema applied to data from four large, but characteristically different, earthquakes -- Chile (Mw 8.8 2010), Tokoku-Oki (Mw 9.0 2011), Baja California (Mw 7.2 2010) and Wells Nevada (Mw 6.0 2008). For each of our four mainshocks, the number of detections within the 10 hour time windows span a large range (1 to over 200) and statistically >20% of the waveforms show evidence of anomalous signals following the mainshock.
Session: Intermountain West Earthquake and Volcano Characterization and Related Hazards
Presenter   Zhang, Qiong
Schedule   Wed 10:45 AM / Oral
Room   155B
Seismic and Volcanic Characterization of the Coso Geothermal Field, California
LIN, G., University of Miami, Miami, FL, glin@rsmas.miami.edu; ZHANG, Q., University of Miami, Miami, FL, qzhang@rsmas.miami.edu
The Coso volcanic field is located at the west edge of the Basin and Range province and is well known as a geothermal area. Fumaroles are present along faults bounding the rhyolite-capped horst and locally within the rhyolite field. From the youthfulness of the rhyolite lavas and a zone of low seismic velocity crust roughly beneath the rhyolite, a magma body was proposed to be the source of thermal energy for the geothermal system. In this study, we combine the classic and state-of-the-art techniques to characterize the seismic and volcanic features of the Coso area. Seismic tomography inversions are carried out to map the variations in Vp, Vs and Vp/Vs ratios. The velocities in the top layers of our model are correlated with the surface geological features. The fault zones and basin areas, such as the Little Lake fault zone, the Airport Lake fault zone and the Indian Wells Valley show low-velocity anomalies, while the major mountain ranges, such as the south Sierra Nevada, the Coso Range, and the Argus Range show higher velocities. The resulting 3-D velocity model was used to improve absolute locations for all local events between 1981 and August 2011 in our study area. We then applied a similar event cluster analysis, waveform cross-correlation, and differential time relocation methods to improve relative event location accuracy. Over 66% of the seismicity falls into similar event clusters. A dramatic sharpening of seismicity patterns is obtained after using these methods. We also estimate in situ near-source Vp/Vs ratio within each event cluster using the differential times from waveform cross-correlation. The high-resolution Vp/Vs method confirms the trend of the velocity variations from the tomographic results. The high Vp/Vs ratios observed within a large event cluster centered at 3 km depth beneath the Coso geothermal field, together with the corresponding low Vp in this area, may indicate the existence of magma bodies.
Session: Velocity Models and Modeling
Presenter   Godoladze, Tea
Schedule   Wed / Oral
Room   155A
3D Wave Propagation Studies in the Caucasus
GODOLADZE, T. G., Ilia State University, Tbilisi, Georgia, tea_godoladze@iliauni.edu.ge; DREGER, D., BSL, UC Berkeley, Berkeley, CA, dreger@seismo.berkeley.edu; HUTCHINGS, L. J., Lawrence Berkeley Laboratory, Berkeley, CA, LJHutchings@lbl.gov; GOK, R. M., Lawrence Livermore Laboratory, Livermore, CA, Gok1@llnl.gov
The Caucasus is a tectonically and structurally complex region, Crustal and upper mantle velocities show great heterogeneity in this region and regional phases display significant variations in both amplitudes and travel time. Regional velocity and attenuation models fail to capture the complexity of the region, and the extent of the exact misfit is uncertain. The Caucasus belongs to the region of moderate seismicity. Large earthquakes that occurred in the Caucasus within the period of instrumental observation are: the 1991 Ms=7.0 Racha earthquake, the largest event ever recorded in the region; the 1992 M=6.5 Barisakho earthquake; Ms=6.9 Spitak, Armenia earthquake (100 km south of Tbilisi.In the past few years, a considerable number of stations, both broadband and short-period have become available in this region. Regional National Networks has developed a very strong collaboration in terms of online data exchange. Each country may access waveform data from the respective networks, and use it for scientific purposes.In order to improve the seismic velocity and crustal structure in the region we performed a study of 3D seismic wave propagation in the Caucasus. The data from national seismic networks of Georgia, Azerbaijan, Armenia, and Turkey are analyzed. An initial 3D velocity model is developed using local tomography studies. We performed local tomography studies based on data derived from temporal networks deployed throughout of the Caucasus during the strong events (Racha 2009. Javakheti 2010, Zakhatala 2012). We then perform waveform modeling to improve the understanding of regional phase propagation in complex areas, and use stacked receiver functions and surface wave dispersion curves to help constrain absolute shear wave velocity to a depth of 100 km at each station.
Session: Next Generation of Ground Motion Prediction Models
Presenter   Salomone, Lawrence
Schedule   Wed 11:00 AM / Oral
Room   155C
Updated EPRI (2004, 2006) Ground-Motion Model for the Central and Eastern United States
SALOMONE, L. A., EPRI Consultant, Palo Alto, CA, ceus_ssc@yahoo.com; TORO, G. R., Lettis Consultants International, Acton, MA; YOUNGS, R. R., AMEC, Oakland, CA; MCGUIRE, R. K., Lettis Consultants International, Boulder, CO; CHAPMAN, M., Virginia Tech, Dept of Geosciences, Blacksburg, VA; KASSAWARA, R