Schedule Overview

Events will be held at the Egan Convention Center
or the Hilton Anchorage as noted.

Tuesday 29 April 2014
9:30–5:00 PM Board of Directors Meeting
Hilton Aspen/Spruce Room
10:00 PM–4:00 PM Pre-Meeting Media
Workshop

Hilton Dillingham Room
3:00–8:00 PM Registration
Egan Lobby
6:00–8:00 PM Icebreaker Reception
Egan La Perouse
Wednesday 30 April 2014
7:00 AM–6:00 PM Registration
Egan Lobby
7:15–8:30 AM Light Breakfast
Egan Cook & Arteaga Room
7:30–8:25 AM Government Relations
Workshop

Egan—Board Room A
7:30 AM–Noon Poster Sessions
Egan Cook & Arteaga Room
8:30 AM–Noon Oral Sessions
Egan Rooms 1, 2, 3, 4, & 7
Noon–2:00 PM Annual Luncheon
Egan La Perouse
2:15–5:45 PM Oral & Poster Sessions
Egan Rooms 1, 2, 3, 4, & 7
Egan Cook & Arteaga Room
5:45–7:30 PM Student Reception
Hilton 15th Floor—
Top of the World
5:45–7:30 PM Early Career Reception
Hilton 15th Floor—
Chart Room
Thursday 1 May 2014
7:00 AM–6:00 PM Registration
Egan Lobby
7:15–8:30 AM Light Breakfast
Egan Cook & Arteaga Room
7:30–8:25 AM Government Relations
Workshop

Egan—Board Room A
7:30 AM–Noon Poster Sessions
Egan Cook & Arteaga Room
8:30 AM–Noon Oral Sessions
Egan Rooms 1, 2, 3, 4, & 7
Noon–1:00 PM Public Policy Luncheon
Egan La Perouse
1:30–5:00 PM Oral & Poster Sessions
Egan Rooms 1, 2, 3, 4, & 7
Egan Cook & Arteaga Room
5:15–6:15 PM Joyner Lecture
Hilton Alaska Ballroom
6:15–7:15 PM Joyner Reception
Hilton Bristol Bay Ballroom
Friday 2 May 2014
7:00 AM–6:00 PM Registration
Egan Lobby
7:15–8:30 AM Light Breakfast
Egan Cook & Arteaga Room
7:30 AM–Noon Poster Sessions
Egan Cook & Arteaga Room
8:30 AM–Noon Oral Sessions
Egan Rooms 1, 2, 3, 4, & 7
Noon–1:00 PM Friday Luncheon
Egan La Perouse
1:30–5:00 PM Oral & Poster Sessions
Egan Rooms 1, 2, 3, 4, & 7
Egan Cook & Arteaga Room
Bus 1: 6:30 PM
 Returning ~9:30 PM

Bus 2: 7:30 PM
 Returning ~10:30 PM

Field Trip 2:
National Tsunami
Warning Center

Hilton Main Lobby*
Saturday 3 May 2014
Bus 1: 8:00 AM
  Returning ~6:30 PM

Bus 2: 9:00 AM
  Returning ~7:30 PM
Field Trip 1:
1964 Alaska Earthquake

Hilton Main Lobby*
Bus 1: 8:30 AM
 Returning ~11:30 AM
Field Trip 2:
National Tsunami
Warning Center

Hilton Main Lobby*

*Please Note: Field Trip Buses board 30 minutes prior to departure times and will leave promptly (please don’t be late). Return times are approximate.

Pre-Meeting Media Workshop

Communicating with Non-Technical Audiences

Tuesday, April 29, 10:00 AM–4:00 PM


with Nan Broadbent,
SSA Press Liaison

SSA is offering attendees an opportunity for practical, hands-on training in communicating technical information for a lay audience. Learn the foundation of working with the media — crafting a simple message that’s memorable. The majority of the workshop, scheduled for 10 AM–4 PM, will involve on-camera work and small group exercises. Each trainee will participate in at least one mock interview.

Nan Broadbent will facilitate the workshop. She is the press liaison for SSA, promoting the Annual Meeting and journals to the media. Prior to SSA, Nan worked in Washington, D.C., where she focused on research communications. She served as communications director for the journal Science and the American Association for the Advancement of Science. She conceived of and created EurekAlert!, which is a tool widely used by science reporters and public relations officers.

SSA 2014
Icebreaker
Opening Reception

Tuesday, April 29, 6–8 PM
La Perouse Hall
Egan Convention Center

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

Enjoy tasty food and beverages.

Meet, mingle, and reconnect!

Icebreaker attendees must be
registered to receive drink tickets.

Sponsored by the Alaska Earthquake Center

SSA Annual Luncheon

Featuring the President’s Address

“The Prediction Problems of
Earthquake System Science”

by Tom H. Jordan, SSA Board President

Wednesday Noon–2:00 PM
Egan La Perouse

Presentation of Awards

Harry Fielding Reid Medal
Kerry Edward Sieh

Charles Richter Early Career Award
Vedran Lekic
 

2014 Joyner Lecture
Thursday, May 1st 5:15–6:15 PM
Hilton Alaska Ballroom

David M. Boore
Geophysicist, USGS Menlo Park

Ground Motion Prediction Equations: Past, Present, and Future

Dr. Boore has published over 260 papers, most dealing with various aspects of the problem of estimating the ground shaking from large earthquakes. The topics covered in these publications range from the seismic source to site response, with stops in between.

Dr. Boore has served on a number of panels, boards, and committees including the Editorial Board of the Journal of Earthquake Engineering, U.S.-Japan Panel on Wind and Seismic Effects, the Department of Energy’s Tank Seismic Experts Panel, Nuclear Regulatory Commission’s Ground Motion Guidelines project, Senior Seismic Hazard Analysis Committee, SSA Board of Directors, Editor-in-Chief of BSSA, International Association of Seismology and Physics of the Earth’s Interior, and Commission on Strong Motion Seismology.

The U.S. Department of the Interior awarded Dr. Boore the Meritorious Service Award in 1993 and the Distinguished Service Award in 2005 in recognition of his research in many areas of engineering seismol- ogy. He is an AGU Fellow and an SSA honorary life member. In 2006 he was awarded the Indian Society of Earthquake Technology Trifunac Award for Significant Contributions in Strong Motion Earthquake Studies, and in 2010 he received the COSMOS/EERI/SSA Bruce A. Bolt medal. He is a coauthor on a paper given the Outstanding Paper of 2008 award by the Earthquake Engineering Research Institute (EERI).


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.

Government Relations Workshop

Wednesday & Thursday
7:30 – 8:25 AM
Top Floor of the Egan —
Board Room A

All attendees welcome!

[Elizabeth Duffy Photo] In this contentious legislative environment, with declining budgets and sequestration, every scientist needs to know how to make the best case for their science. Come for an informative session and find out about the inner workings of the US Congress, how laws are made, and how to become an effective advocate for science. Elizabeth Duffy, President of the Federal Affairs Office in Washington, D.C. and government affairs coordinator for SSA, will lead this session. Elizabeth will bring over 20 years of business, political and advocacy experience to show you how to best impact legislation in Congress, how to establish productive relationships with elected officials and their offices, and how to speak to representatives to get your message heard.

SSA 2014
Student and Early Career Receptions


Wednesday, April 30, 5:45–7:30 PM
Hilton 15th Floor
Students - Top of the World
Early Career - Chart Room

Attend the Student or Early Career Reception with your peers and fellow SSA members. Come and enjoy good food and good spirits.

Don’t miss this opportunity to make new connections
and network with your colleagues!

Students and Early Career attendees must bring their ‘invitations’ for admittance – Invitations will be found in your badge holder.

Drink tickets are provided upon entrance check in
at the reception.

Wednesday Luncheon

President’s Address

“The Prediction Problems of
Earthquake System Science”

Tom H. Jordan,
SSA President 2013-2014
 

Thursday Luncheon

Public Policy Speaker

John Davies
Senior Researcher, Energy Policy
at the Cold Climate Housing Research Center

“Lessons Learned by a Seismologist in Public Service”
 

John Davies received a Ph.D. in Geophysics from the University of Alaska Fairbanks in 1975. As a Research Associate at Lamont-Doherty Geological Observatory, he focused on the Aleutian Islands and Alaska Peninsula regions of Alaska. He served as State Seismologist for both the Alaska Division of Geological and Geophysical Surveys and at the Geophysical Institute at the University of Alaska where he was a prin- cipal in creating the Alaska Earthquake Information Center and the Alaska Volcano Observatory. John served on the Borough Assembly in Fairbanks before he was elected for five terms as Representative to the Alaska State Legislature (1993-2002). From there, John was hired as the Director of Research for the Cold Climate Housing Research Center (CCHRC). In 2010, John stepped down to focus on energy pol- icy issues as Senior Researcher – Energy Policy at CCHRC, his present position. He continues to serve on the Fairbanks North Star Borough Assembly, where he is Deputy Presiding Officer.

Field Trip 1

The 1964 Earthquake, Tsunamis, Paleoseismology, and Subduction Processes in Southcentral Alaska

[Turnagain Arm from Beluga_Point]
Leaders:Peter Haeussler, USGS Anchorage &
Ian Shennan, University of Durham
Departures:8 & 9 AM Saturday 5/3
(Meet in Hilton main lobby 30 minutes prior to your departure to board bus)
Returning: Approximately 6:30 & 7:30 PM

Field trip Group 1 will leave at 8 AM and Group 2 will leave at 9 AM on Saturday, 03 May 2014 (the day following the close of the regular meeting program) from the Anchorage Hilton lobby and return at approximately 6:30 – 7:30 PM. Please do not plan to catch a flight that evening as we cannot guarantee the bus will return on schedule.

This one-day field trip will travel along beautiful Turnagain Arm from Anchorage to Whittier and back – 120 miles, round trip. The field trip will feature stops that highlight effects and lessons learned from the 1964 M9.2 Great Alaska earthquake. The first stop will be at the famous Turnagain Heights landslide, which slid up to 610 m. The next stop will be in rocks of the Mesozoic accretionary complex, to show processes that occurred at seismogenic depths and discuss the long-term tectonic history of the region. The next stop will be at the Girdwood flats. This is a world-class paleoseismic site, which has excellent exposures of the stratigraphy of the 1964 earthquake and beautiful views of the ghost forest, which developed after the forest tree roots dropped into the intertidal zone. The next stop will be at Portage Lake, to discuss the history of the Little Ice Age in the region. The bus will then travel thru a unique 4 km long tunnel to the community of Whittier (population 150), at the western edge of Prince William Sound. Here, submarine landslide processes will be discussed that led to local tsunami generation and loss of 13 lives in the 1964 earthquake.

NOTE: Clothing Requirements

It is MANDATORY for people to have calf height rubber boots at the Girdwood stop. This is not a mere suggestion. They can be purchased at the Army Navy Store, which is in walking distance from the Egan Center. People may get muddy. Participants should bring rain gear as there is a reasonable chance of rain and a high chance of wind. Expect temperatures from the 40s to the 60s.

Field Trip 2

Tour of the
National Tsunami Warning Center

[National Tsunami Warning Center]
Leader:Paul Whitmore, NOAA/NWS National Tsunami Warning Center
Departures:6:30 & 7:30 PM Friday 5/2
8:30 PM Saturday 5/3
(Meet in Hilton main lobby 30 minutes prior to your departure to board bus)
Returning: Approximately 3 hours after departure

Come see one of two Tsunami Warning Centers in the USA. In this tour, you will see an active seismic observatory which monitors the entire planet and focuses on producing fast earthquake analysis and response. The Center also specializes in tsunami forecasting with forecasts based on the earthquake analysis and refined with tsunami observations. The tour will be indoors, though if time permits we will step outside to see a seismic vault.

The Center is in Palmer, Alaska, approximately 45 minutes from Anchorage.

SSA 2014
April 30 – May 2
Anchorage, Alaska

SSA Annual Meeting Mobile Web App
Version 4.0.0β

Copyright © 2011–2014
The Seismological Society of America

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

Developed using iUI

Help SSA Identify
Outstanding
Student Presentations!

Every year, SSA gives awards to outstanding student presentations. We need your help.

First, notice which presentations are given by Students. All Students will have yellow name badges. Look for these yellow student badges during oral and poster presentations (yellow tag will also be posted on the poster boards of Student poster presentations.)

What: Fill out short evaluation forms for only the student presentations you think are worthy.

Where: You can use your computer, tablet, or smart phone to fill out the form online here.

More Where: Forms are in the tech session rooms, on tables around the center, or at the registration booth.

Who: Everyone*

When: All forms must be completed by 6 PM on Friday, 2 May 2014. Drop off completed paper forms at the registration booth or give them to the session chair.

* 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 were involved.

Thanks for your help!

Geo Energy Monitoring Systems Inc.
167 Piedra Loop
Los Alamos, NM 87544
(928) 899-1875
www.geoems.com

GeoSIG Ltd
Wiesensrasse 39
8952 Schlieren
Switzerland
+41-44-8102150
www.geosig.com

Guralp Systems Ltd.
12 Southwood Dr.
Orinda, CA 94563
(925) 254-1357
www.guralp.com

IRIS (Incorporated Research Institutions for Seismology)
1200 New York Avenue NW, Suite 400
Washington, D.C. 20005
(202) 682-2220
www.iris.edu

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

Trimble Infrastructure
1600 Tenth St, Ste A
Plano, TX 75074
(214) 440-1265
www.trimble.com/infrastructure

Weir-Jones Engineering Consultants
2040 West 10th Avenue
Vancouver, BC V6J 2B3
Canada
(604) 732-8821
www.weir-jones.com

Alaska Earthquake
Center

Sponsor of the Icebreaker Rececption

The Alaska Earthquake Center is dedicated to reducing the impacts of earthquakes, tsunamis and volcanic eruptions in Alaska. We provide definitive earthquake information derived from an expansive network of seismic monitoring stations. We are scientists, engineers, data analysts, field technicians and educators with a shared passion for working in the real-time geology of Alaska. We are motivated by the reality that some of the great tragedies of our time are caused by earthquakes, as well as the knowledge that earthquakes are a powerful tool in understanding the science, hazards and resources of the landscape around us. The center is housed at the Geophysical Institute on the campus of the University of Alaska Fairbanks.

Bechtel

Sponsor of the 1964 Alaska Earthquake Field Trip

Bechtel is one of the world's largest engineering-construction firms. Founded in 1898, Bechtel provides premier technical, management, and directly related services to develop, manage, engineer, build, and operate installations for customers worldwide. Many of the objectives of the Seismological Society of America, such as promoting public safety and showing that it is reasonably practicable and economical to build for security from earthquake damage, are in close accord with Bechtel's vision and values.

Kinemetrics

Sponsor of the Thursday Morning Breakfast and Coffee Breaks

Kinemetrics is the world leader in innovative design, quality manufacturing and timely supply of earthquake instruments, systems and solutions, where reliability, flexibility and cost effectiveness are important. ISO 9001:2008 certified, KMI has 45 years experience providing seismologists and structural engineers with the highest standard implementations for field when, under extreme environmental conditions, secure continuous operation is critical, for research, where new vision must be achieved and for monitoring where new technological boundaries are tested.

Pacific Gas & Electric Company

Sponsor of the Annual Meeting Luncheon
and Internet Café

Pacific Gas and Electric Company, incorporated in California in 1905, is one of the largest investor-owned gas and electric utilities in the United States. Based in San Francisco, the company is a wholly owned subsidiary of PG&E Corporation. PG&E's infrastructure includes 131,000 miles of electric lines and 43,000 miles of gas pipelines. There are 18,400 employees who carry out Pacific Gas and Electric Company's primary business -- the transmission and delivery of energy. The company provides natural gas and electric service to approximately 13 million people throughout northern and central California, a region of moderate to high earthquake activity. PG&E provides strong leadership toward improving seismic safety throughout its service territory.

Trimble Infrastructure

Sponsor of Wednesday Morning Breakfast and Coffee Breaks

Trimble's portfolio of market-leading advanced positioning solutions (GNSS), application software and state-of-the-art REF TEK seismic recorders/sensors provide integrated proven tools for both geodetic, seismological, and meteorological scientific research and monitoring applications. Our robust technologies provide an unprecedented level of capability and reliability to improve safety decisions, ensure structural integrity, and future-proof your investment.

URS Corporation

Sponsor of the Public Policy Luncheon

URS Corporation is a leading provider of engineering, construction and technical services for public agencies and private sector companies worldwide. We offer a full range of program management; planning, design and engineering; systems engineering and technical assistance; information technology; construction and construction management; operations and maintenance; and decommissioning and closure services for federal, oil and gas, infrastructure, power and industrial projects. Established in 1951 with more than 50,000 employees in nearly 50 countries.

Sponsored by URS Corporporation

Weir-Jones Engineering Consultants

Level I Sponsor

Weir-Jones Engineering Consultants specializes in structural and geomechanical monitoring and testing services. Providing Shake Alarm and Shake Monitoring. The Company was founded in 1971 to provide specialized structural and geomechanical monitoring and testing services to the resource and transportation sectors. The Company’s capabilities subsequently expanded in the areas of data processing and testing system design, the application of this expertise has been extended considerably, notably in the fields of structural integrity monitoring for heavy structural, energy and offshore systems. The Company has its headquarters in Vancouver and has been active in projects in 55 countries.

Weston Geophysical Corp.

Sponsor of the Friday Meeting Luncheon

Weston Geophysical Corp. is a privately held consulting firm focused on seismic monitoring research. Our staff has expertise in all aspects of monitoring manmade and natural seismicity, at scales ranging from microseismic to global. Some specific areas of research focus include the following:

  • Advanced signal processing algorithms for highly accurate seismic event detection, location and characterization
  • Field deployments to examine source phenomenology and monitor induced/triggered seismicity
  • Coda-based studies of magnitude and other source parameters
  • Regional surface-wave magnitudes and discriminants
  • Three-dimensional seismic velocity models of the crust and upper mantle

Save the Dates!
Upcoming
SSA Annual Meetings

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

SSA 2016
April 20–22, 2016
Reno, Nevada
  The Peppermill Reno

Thank You!

SSA
Especially Wishes to Thank

Program Committee Chairs

Jeff Freymueller
  (University of Alaska, Fairbanks)
Peter Haeussler
  (USGS Anchorage)

Program Committee Members

Geoff Abers
  (Lamont-Doherty Earth Observatory, Columbia University)
Lloyd Cluff
  (Geosciences Director Emeritus, Pacific Gas &Electric)
Christine Goulet
  (UC Berkeley, UCLA)
Rich Koehler
  (State of Alaska, Div of Geological & Geophysical Surveys)
Thorne Lay
  (UC Santa Cruz)
Stephanie Prejean
  (USGS Alaska Volcano Observatory)
Julian Lozos
  (Stanford University & USGS Menlo Park)
Dave Wald
  (Board of Directors Liaison, USGS Denver)

… and all the other folks, too numerous too mention,
who’ve helped to make this year’s meeting a success!

75 Years of Frequency-Size-Distribution of Earthquakes: Observations, Models and Understanding
Conveners: Thessa Tormann, Swiss Seismology Service, Max Wyss, World Agency of Planetary Monitoring and Earthquake Risk Reduction, and Jeanne Hardebeck, US Geological Survey
Since the power-law scaling of the earthquake-size distribution has been documented for the first time by Ishimoto and Iida in 1939, 75 years ago, it has become most commonly characterized by the b-value in logN=a-bM, published in 1944 by Gutenberg and Richter and cited more than 1500 times since then. A multitude of studies have since been devoted to uncover the nature of b: From suggestions of being universal and unity to reports of significant spatial and temporal variation. From global catalogs to local microseismicity and acoustic emissions in laboratory experiments. From natural earthquakes to induced seismicity. From statistical robustness analysis to b-value imaging and physical interpretation.Beyond these aspects, the b-value is frequently used in seismic hazard analysis for linking the observed smaller to the infrequent larger event rates, and an improved understanding of its properties and meaning is crucial to advance future seismic hazard assessment. Recent years have offered exciting new insights, according to which the observed natural b-value variability can be related to physical properties of the crust, in particular its stress distribution, such as has been suggested from lab experiments for decades.We invite b-value related contributions from all scales – laboratory, induced and natural seismicity – addressing variability or stability in space and through time. We would like to discuss statistical requirements, appropriate techniques and uncertainty treatment; we are particularly interested in approaches of physical interpretation – (what) can we learn from b-value analysis about the physics of the crust and spatiotemporal earthquake potential? We also invite studies from different backgrounds that face similar analysis challenges or reveal related characteristics, such as strong along fault heterogeneity – do they offer new correlation prospects with b? Or what about b-values in the physical earthquake simulator world?We aim for a forum to discuss the state-of-the-art in b-value analysis and its impact and outline future challenges and opportunities.
A Decade of Great Subduction Earthquakes – What Have We Learned from Their Ground-Motions?
Conveners: Trevor Allen, National Resources Canada and Gail Atkinson, Western University
Of the nine largest subduction zone earthquakes documented since 1900, four have occurred in the last 10 years, including; 2004 Sumatra-Andaman Islands and 2005 Nias, Indonesia; 2010 Maule, Chile, and; 2011 Tohoku, Japan. Many of these earthquakes, together with their aftershocks and a host of “smaller” – but equally important – independent earthquakes have generated large databases of digital ground-motion recordings and macroseismic intensities. This session will discuss subduction zone ground-motion observations (both instrumental and macroseismic) for modeling hazard and risk globally. We seek contributions that address a range of questions, including: Can ground-motion models developed from one global database (e.g., Japan) be translated to another region (e.g., Alaska or Cascadia)? Are current magnitude scaling models appropriate for Great earthquakes? How do we model the ground-motion uncertainties? What macroseismic effects are commonly observed from these events? How can we leverage of modern and historical macroseismic data to inform ground-motion model development for hazard assessments? Are global monitoring seismic networks sufficient for monitoring large subduction events? How can we use this information to mitigate the potential impacts of future large subduction earthquakes?
Advances in Seismic Imaging and Monitoring of Time-Dependent Variations: Civil Structures, Near-Surface, and Shallow Crustal Scales
Conveners: Marco Pilz, GFZ-Potsdam and Nori Nakata, Stanford University
This session will cover new developments in seismic imaging and monitoring, from the sub-meter-scale in engineering applications for controlling civil structures to the meter-scale in near-surface seismology, up to the kilometer-scale of shallow crustal seismology. Contributions dealing with controlled-source continuous or repeated observations to detect temporal variations of seismic velocity or other properties are especially welcome. We invite submissions relevant to theoretical and practical aspects using both active and passive seismic data, methodological studies, and studies aimed at developing new instrumentation. We aim to highlight progress of imaging and surveillance studies and related research in real-time seismic hazard monitoring.
Advances in Understanding Earthquake Hazard in Central and Eastern North America
Conveners: Robert Williams, USGS, Mitch Withers, U Memphis, and William Stephenson, USGS, and John Ebel, Boston College
About twenty M4 to M5.8 earthquakes have occurred since 2010 in the intraplate region of Central and Eastern North America. Most of these events were well recorded by ANSS and USArray seismographs and provide data that can contribute to reducing uncertainties in seismic hazard assessment. From recordings of these events we seek contributions providing insights into seismic source characteristics, aftershock behavior, attenuation, site response, crustal structure relative to characterization of seismic and aseismic regions, and building damage. All aspects of studies related to the 2011 M5.8 Mineral, Virginia earthquake are of particular interest.
Alaska Update of the USGS National Seismic Hazard Maps
Conveners: Peter Haeussler, USGS, John Anderson, U Nevada Reno, and Mark Petersen, USGS
The United States Geological Survey (USGS) plans to update the Alaska portion of the U.S. National Seismic Hazard Maps to produce science-based products that will be considered for inclusion in future building codes, risk assessments, and other public policy applications. These seismic hazard maps are based on our assessment of the “best available science” at the time of the update, and incorporate a broad range of scientific input models and parameters. We invite papers discussing new geologic and geophysical information on Alaska seismic sources (faults and seismicity) and ground motion models. In particular we are interested in new paleoseismic fault studies, geodetic and geologic combined inversions of fault slip rates, and ground motion models for subduction and crustal sources.
Citizen Seismology: Citizens Helping Science Helping Citizens
Conveners: Paul Earle, Susan Hough, USGS, and Gail Atkinson, University of Western Ontario
There is a growing trend across scientific disciplines for citizens to aid scientific research and monitoring. In seismology, citizens now contribute both actively and passively to the collection and analysis of seismic data. These efforts include documenting macroseismic effects on a scale not imaginable before the advent of the Internet, contributing to early detection of significant earthquakes, and hosting of seismometers of various types. Citizen seismology clearly offers enormous potential to not only contribute to scientific research but also to push scientists to better understand and respond to public demands and expectations after natural disasters. In this session we invite contributions that focus on quantitative analysis of macroseismic data from both historical and instrumental earthquakes as well as papers that cover technical aspects of citizen science, ways to better engage with citizens,and other applications.
Cold Climate Installation Techniques and Instrumentation Developments for Temporary and Long-Term Networks
Conveners: Bruce Beaudoin, IRIS PASSCAL and Christopher Bruton, AEIC-UAF/GI
Temporary and long-term seismic networks reach all corners of the globe. Over the last decade, the increased interest in glacier and polar seismology concerning climate research, and more recently, the move of USArray to Alaska, have raised expectations that temporary seismic stations will operate continuously in these extreme environments. Polar climates pose particular challenges to temporary and long-term networks and the instrumentation that is deployed. Coldtemperatures, snow cover, freeze-thaw cycles and short daylight hours challenge the operational capabilities of these remote stations. Logistic constraints limit the size of deployable, off-the-grid power systems, and limited winter access dictates that equipment must operate for many months or years without maintenance visits, some blind even of state of health information.This session will focus on new developments in cold-climate seismic deployments, including enclosures, power systems, telemetry, instrumentation, and logistics.
Deciphering the Earthquake and Tsunami History of Subduction Zones
Conveners: Marco Cisternas, Universidad Católica de Valparaíso, Lisa Ely, Central Washington U, and Ed Garrett, Durham U, and Daniel Melnick, Universität Potsdam, and Alan Nelson, USGS, and Robert Wesson, USGS
Recent great earthquakes and their accompanying devastating tsunamis in Sumatra, Chile, and Japan highlight the need for subduction-zone histories that span the hundreds to thousands of years needed to sample the greatest events. Few instrumental measurements cover more than a fraction of the cycle of interplate strain accumulation and release at active subduction margins. Even on the few coasts with extensive historical archives, records are too short or incomplete to fully reconstruct the history of the greatest earthquakes and highest tsunamis. Without accurate reconstructions spanning multiple earthquake cycles, there is little hope of understanding the variables that control how megathrust ruptures nucleate and extend along subduction zones, how the highest tsunamis are generated and dispersed across ocean basins, and in assessing the hazards from these great events. The century to millennia-long recurrence times of these events dictates the use of an array of interdisciplinary techniques from, for example, the fields of stratigraphy, geomorphology, dating, paleontology, oceanography, and history, to determine the timing and characteristics of past events. This session will explore recent developments in reconstructing the histories of large earthquakes and tsunamis on active subduction margins worldwide. We especially encourage presentations comparing and contrasting earthquake and tsunami behavior at different subduction zones, presentations integrating multiple techniques to characterize earthquake and tsunami behavior, and descriptions of new and innovative methods of investigation.
Development of 2014 U.S. National Seismic Hazard Maps and Their Implementation in Engineering Applications
Conveners: Sanaz Rezaeian, USGS, Christine Goulet, U California Berkeley, and John Anderson, U Nevada Reno, and Mark Petersen, USGS
The 2014 National Seismic Hazard Maps (NSHMs) produced by the United States Geological Survey (USGS) incorporate significant changes in the earthquake catalog, source models, and ground motion models. These seismic hazard maps are based on the USGS’ assessment of the “best available science” at the time of the update, and incorporate a broad range of scientific input models contributed by the seismological research community. The NSHMs are used to produce science-based products that will be considered for inclusion in future building codes, risk assessments, and other public policy applications. We invite papers that discuss major changes to the maps or to the input data and models (e.g., catalog, source models, ground motion models) since 2008; or discuss improvements in data, methods, and models that could be used in future maps. Papers that discuss the impact of the 2014 NSHMs on design maps and future building codes are encouraged.
Development of Next Generation Field Methods for Portable Broadband Seismic Arrays
Conveners: Timothy Parker, IRIS PASSCAL and David E. Hawthorn, SEIS-UK
Researchers have started to deploy the next generation of directly buriable broadband seismic sensors after years of building vaults to protect arguably observatory-grade sensors that were neither rugged nor water resistant, using shallow simple vaults that limited their noise performance. The debate of whether to chase the final 2-10 dB of noise floor in shallow vaults has shifted to the question of what are the signal-to-noise ratios required to collect enough events that will allow more successful deployments with limited logistics and minimal budgets? This session will present techniques and developments to enable broadband seismology to increase the number of successfully deployed stations with decreasing logistical budget requirements including use of other types of shorter-period sensors in mixed-mode arrays.
Diverse Mechanisms of Subduction Zone Fault Slip: Exploring the Relationships among Seismic Rupture, Transient Slip, and Steady Creep
Conveners: Harmony V. Colella, Miami U Ohio, Emily C. Roland, USGS, and Aaron G. Wech, USGS
The release of elastic energy at subduction zone faults can result in intense ground shaking over broad regions and sea floor displacement that generates immense tsunamis. Tectonic tremor and/or slow slip, commonly observed between the seismogenic zone and down-dip stable sliding, represents an intermediate mechanism for faults to accommodate plate convergence. Yet the relative location of these different slip behaviors varies among subduction zones both along-strike and along-dip. Additionally, some subduction zones accommodate strain through aseismic stable sliding within the depth range where large earthquakes are expected to occur. Why such variability exists (e.g., frictional, rheological properties, material heterogeneity, variations in fluid pressure) and how it affects great earthquakes (e.g., up-dip or down-dip extent of seismic rupture, incomplete inter-seismic coupling) is poorly understood. It also remains unclear whether these diverse mechanisms are consistent over time or what role they play in the earthquake cycle. The objective of this session is to showcase the current understanding on these topics, and we seek to combine a wide range of abstracts that utilize observations, theory, or modeling with aspects from seismology, geodesy, lab experiments, and field observations.
Earthquake Physics and Interaction
Conveners: Jochen Braunmiller, University of South Florida
This session highlights recent advances in earthquake physics and the interrelationships between earthquakes in a particular region. Topics include earthquake clustering, time-dependent forecasting, scaling relationships, stress drop studies, and others.
Effects of Topography and Surface Loads on Earthquakes and Faulting
Conveners: Richard Styron and Eric Hetland, U Michigan
Topographic features can produce significant stress in the lithosphere. These topographic stresses can be on the order of inferred stress drops in earthquakes, raising the possibility that they may affect the earthquake processes. In the upper crust, the topographic stresses can be quite heterogeneous, resulting in stress heterogeneities of both fault shear and normal stresses even if the background tectonic stress is relatively smooth. Dynamic earthquake simulations have shown that stress heterogeneities on the order of several MPa or more have a strong effect on the earthquake rupture process. Additionally, flexure and isostatic rebound due to changes in water, ice or sediment loads on the Earth’s surface can affect fault behavior over many earthquake cycles, and in some cases may trigger earthquakes on relatively short timescales. We call for theoretical or empirical research on the interactions of topography and surface loads with earthquake behavior. We particularly welcome submissions addressing the following broad topics: Models of dynamic earthquake rupture with heterogeneous pre-earthquake fault normal and shear stress; Constraints on the stress evolution throughout an earthquake, including both stress drop and final post-earthquake stress; relationship of topographic stresses and rupture in an individual earthquake; Relationship between topography and location or style of faulting; Modulation of fault activity due to changes in surface loads; and Measurements and quantitative estimates of stresses near faults.
Emergence of Continuously Recording Very Large Array Capabilities in Seismology
Conveners: Rick Aster, IRIS PASSCAL & NMIMT and James Gridley, IRIS PASSCAL
Enhanced recovery, injection well monitoring, microseismic monitoring, energy exploration, and reservoir characterization/monitoring are driving dramatic new commercial-sector seismic instrument development and analysis techniques in the energy industry. The associated multichannel capabilities are up to orders of magnitude greater than have been available to non-industry-aligned academic seismology. This expansion of monitoring and imaging scope has also produced renewed emphasis on the use of broader-band instrumentation at very large network/array scales. This session will encompass both the diverse scientific/industry opportunities associated with the deployment and analysis of very large numbers of seismographs, and opportunities for associated novel collaborations within the greater seismological and Earth science community.
Explosive Source Characterization
Conveners: Catherine Snelson, National Security Technologies, Christopher Bradley, Los Alamos National Laboratory, and Sean Ford, Lawrence Livermore National Laboratory, and Kyle Jones, Sandia National Laboratories
In this session, we investigate the coupling of explosive energy to the radiated seismic wave field. Integration of detailed site setting, material characterization and modeling efforts for understanding explosive sources is critical for nuclear test monitoring and verification. In order to relate near field explosive source phenomenology with far-field observations in areas where characterization data may not exist, well-constrained input and validation from materials testing, structural analysis, field observations, overhead imagery, subsurface feature detection, first-principle calculations, and source-to-receiver modeling are required. These approaches must then be translated into useful, transportable methods for the verification community to characterize the source and quantify uncertainties. We invite contributions from all research focus areas and technologies that meld efforts described above.
Fault Structure, Heterogeneity, and Implications for Rupture Dynamics
Conveners: Thomas H. W. Goebel, California Institute of Technology, Thibault Candela, Pennsylvania State University, and Heather Savage, Lamont-Doherty Earth Observatory
Faults are complex zones of highly deformed rock containing hydro-mechanical heterogeneities from the grain to plate boundary scale. These heterogeneities influence frictional properties of faults and control the dynamics of slip instabilities from episodic tremor and slow slip to large earthquake ruptures. Rupture nucleation, propagation and arrest may be affected by variations in intrinsic fault characteristics, for example, fault hydrology, geometry and structure. While fault structure has been studied extensively on exhumed faults much remains unknown about fault structure at depth and its interplay with earthquake ruptures. Can fault heterogeneity limit the length of earthquake ruptures and control nucleation points of large earthquakes? Can fault heterogeneity explain the large diversity of slip behavior (from exotic slow events to large earthquakes) observed along faults? How do ruptures depend on faults characteristics, e.g., roughness, pore pressure and frictional properties? How does deformation localize within faults and is the degree of localization a function of fault lithology and age? How does fault heterogeneity evolve with successive ruptureevents? How are off-fault damage zones created? What is the contribution of dynamically vs. statically created damage? How is fault heterogeneity connected to the creation of fore and aftershocks? To encourage an exchange on some of these questions, we solicit contributions from observational, numerical and experimental studies of fault structure, heterogeneity and implications for earthquake ruptures.
Fifty Years of Tsunami Science: from the 1964 Earthquake and Tsunami to the SAFRR Tsunami Scenario - Advances in Tsunami Source Characterization, Numerical Analysis and Hazard Mitigation
Conveners: Elena Suleimani, U Alaska Fairbanks-GI, Stephanie Ross, USGS, and Hong Kie Thio, URS Corporation, and Dmitry Nicolsky, U Alaska Fairbanks-GI
In recent years, the study of tsunami hazard has evolved from the traditional numerical modeling environment to include loss estimates and mitigation measures, partly due to the large tsunami disasters of the last decade. For instance, the USGS SAFRR scenario for a large Alaskan tsunami's impact in California rolled out in 2013/2014 to coincide with the 50th anniversary of the 1964 Alaska earthquake and tsunami. This session brings together all the different facets of tsunami hazard analysis, from advances in source characterization, propagation and runup modeling to loss estimation and mitigation measures. The session will include several presentations on the SAFRR scenario, and we invite contributions in the broad range of tsunami research including tsunami source characterization in both deterministic and probabilistic framework, new techniques and technologies in numerical analysis and hazard mapping. Studies in the field of tsunami preparedness of the population, perception of risk, human behavior during the event and effectiveness of tsunami warnings are welcomed. We expect the contributed papers to help us better understand the missing pieces in the existing tsunami mitigation programs and improve resilience of coastal communities to tsunami hazards.
From the Earthquake Source to Damage of Buildings: Bridging the Gap between Seismology and Earthquake Engineering
Conveners: Luis A. Dalguer, Institute of Geophysics ETH-Zurich, Carlo Cauzzi, Swiss Seismological Service ETH-Zurich, and Aysegul Askan, Middle East Technical University, and Philippe Gueguen, Institut des Sciences de la Terre
Seismologists and engineers with the capability to understand the scientific and technical features of physical and numerical models including the physics of earthquake source, wave propagation effects, site specific ground motion predictions and seismic vulnerability of structures, are key professional figures within the framework of seismic risk assessment and mitigation studies. Building these professional skills requires seamless cooperative efforts between earthquake scientists and earthquake engineers. This expertise is routinely employed and optimized by a number of governmental, academic and industrial stakeholders. With the aim of gathering and promoting the discussion amongst the aforementioned scientists and engineers, in this session we welcome innovative studies on the topics mentioned above, with strong emphasis placed on practical applications, characterized by a successful integration of seismological and engineering advancements, tailored to the mitigation of seismic risk.
Geometric Complexities along Strike-Slip Systems: New Insights on Seismic Hazards, Earthquake Behavior, and Fault System Evolution
Conveners: Sean P. Bemis, U. Kentucky, Jeff Benowitz, U. Alaska Fairbanks, and Michele Cooke, UMass Amherst
Major, active strike-slip fault zones around the world often have discrete zones of deformation and elevated seismicity focused around restraining bends and other geometric complexities of the trace of the main strike-slip fault. Alaska contains two world-class examples of extreme deformation and complex seismicity patterns along major strike-slip fault systems – the Mount McKinley restraining bend of the Denali fault, and the Alaskan syntaxis associated with the Fairweather fault. With this meeting having these dramatic examples as a backdrop, we seek contributions that provide global insight into focused zones of deformation along strike-slip faults and their influence on earthquake rupture and recurrence behavior, fault system evolution and the differences in seismic hazards between the primary strike-slip fault and subsidiary faults. We welcome contributions from seismology, paleoseismology, thermochronology, structural geology, geodesy, and analogue/numerical modeling, exploring both theoretical studies and field-based examinations of restraining bends around the world, including (but not limited to) the well-studied examples from the San Andreas fault system, the major fault systems of central Asia and South/Central America, the Dead Sea fault, the North Anatolian fault, and the Alpine fault.
Great Earthquakes and Slip to the Trench (Seismological Society of Japan/Seismological Society of America Joint Session)
Conveners: Jim Mori, Kyoto U, David Wald, USGS, and Emily C. Roland, USGS
This session invites presentations about large and great earthquakes in convergent zones, especially those events that have significant slip on the shallow portion of the megathrust close to the trench. These types of seismic events, such as the 2011 Tohoku-oki earthquake, have the potential for producing very large tsunamis, as well as strong shaking, and the potential for triggering widespread landslides and liquefaction. We invite papers on to recent seismic and geodetic observations characterizing sources, constraints on the controlling structures and material properties, and modeling studies related to great subduction zone earthquakes.
Induced Seismicity
Conveners: Ivan Wong, URS, Justin Rubinstein, USGS, and Thomas Braun, INGV
Recent damaging earthquakes in Oklahoma, Colorado, and Arkansas have renewed interest in induced seismicity. Across the central and eastern United States, the seismicity rate has doubled over the past 11 years. This rate change can largely be attributed to earthquakes induced by fluid injection associated with oil and gas production. Given the occurrence of damaging earthquakes and the large increase in seismicity rate observed in the United States, it is of critical importance that induced earthquakes are better understood such that their hazards may be mitigated. Topics of interest to address these issues include: the causes of fluid-injection induced earthquakes; methods to distinguish natural and induced earthquakes; why some wells trigger earthquakes, but the vast majority of wells do not; how fluid-injection induced earthquakes may be controlled; and how to estimate the hazard associated with already developed and new fluid injection fields. We invite papers on all forms of induced seismicity but particularly those associated with waste water-injection, geothermal production, carbon sequestration, and hydraulic fracturing. Papers from industry are particularly welcome to better inform the research community on the state of knowledge and practices within the community. As part of this session, a roundtable discussion will be held to discuss (1) the factors influencing induced seismicity, (2) the uncertainties in characterizing induced seismicity, (3) strategies to control induced seismicity, and (4) methods to characterize the earthquake hazards associated with fluid injection.
Joyner Lecture
The William B. Joyner Memorial Lectures were established by the Seismological Society of America (SSA) in cooperation with the Earthquake Engineering Research Institute (EERI) to honor Bill Joyner's distinguished career at the U.S. Geological Survey 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. For more information on the Joyner Lectures see this web page.
Large and Damaging Earthquakes of 2013/2014
Conveners: Gavin Hayes, USGS, Thorne Lay, U California Santa Cruz, and Harley Benz, USGS, and Charles Ammon, Penn State U
Large earthquakes have continued to garner global attention in 2013. Significant events have struck Iran (Khash, April 2013, Mw7.7); China (Lushan, April 2013, Mw6.6); Russia (Okhotsk, May 2013, Mw8.3); New Zealand (Blenheim, August 2013, two Mw 6.5 earthquakes); Pakistan (Awaran, September 2013, Mw7.7); and the Philippines (Bohol, October 2013, Mw7.1), all in the past half-year. Further large and damaging events may also occur over the following six months. Studies of these earthquakes are important not only from a societal perspective – several of these events each caused hundreds of fatalities, and significant economic losses – but also seismologically – the Sea of Okhotsk earthquake was the largest deep earthquake ever recorded, and the Pakistan earthquake was another example of a large continental event involving complex slip on multiple fault structures. We welcome analyses of these earthquakes, from teleseismic to regional studies using seismic data, to co- and post-seismic GPS and InSAR analyses, to studies of the societal and economic impact of such events. Research that tackles event characterization from a multi-disciplinary approach is particularly encouraged.
Merging Paths: Earthquake Simulations and Engineering Applications
Conveners: Ricardo Taborda, U. Memphis, Aysegul Askan, Middle Eastern Technical U., and Tsuyoshi Ichimura, U. Tokyo
Recent and continuing efforts in broadband hybrid and deterministic modeling of earthquake processes indicate that we are approaching a junction where physics-based ground motion simulation techniques can be integrated with seismic hazard analysis, risk and loss assessment studies, engineering analysis and design. This session invites participants to contribute their work in the broader area of earthquake modeling and simulation (coupled) with engineering applications. Topics of interest include: high-frequency ground motion simulation; integrated earthquake simulation, end-to-end and/or rupture-to-rafters simulation; integration of nonlinear soil, surface topography and other site effects in simulation; local and regional simulations including soil-structure and site-city interaction effects; structural systems modeling, analysis and design using simulation synthetics as input; earthquake simulations coupled with GIS data; loss and damage assessment based on synthetic datasets; use of ground motion simulations for evacuation plans and emergency response.
Monitoring Dynamic Changes at Active Volcanoes and Fault Zones
Conveners: Ninfa Bennington, U Wisconsin, Matt Haney, USGS, and Silvio de Angelis, U Liverpool
Emerging time-lapse geophysical methods offer the opportunity to observe active geologic systems as they evolve. For example, examination of the coda of repeating earthquakes and ambient seismic wavefields have allowed for the detection of small temporal changes in velocity at active volcanoes and fault zones. Other recently discovered observables show promise for being even more sensitive measures of small changes. These dynamic methods of observation can provide insight into changing fault zone properties during the seismic cycle or the evolution of a volcanic eruption. In this session, we welcome studies that employ time-lapse geophysical methods to monitor active geologic systems. What can such methods tell us about these systems and how can they be efficiently transferred into tools for routine monitoring?
Near-Field Seismoacoustics of Natural and Man Made Explosions
Conveners: Daniel Bowman, Keehoon Kim, U North Carolina, and Jonathan Lees, U North Carolina
Multiparameter observations (e. g., seismic, acoustic, video, etc.) of natural and man made explosions have the potential to provide important details of source time functions and coupling of seismoacoustic waves. For example, a series of recent studies have observed signatures of rapid ground deformation, distortions due to topographic effects, production of nonlinear acoustic waves, and complex seismoacoustic partitioning during explosions. Understanding these relationships is important for characterizing geophysical phenomena in the atmosphere and the solid earth. Such detailed knowledge is also critical in volcano and nuclear explosion monitoring. We encourage contributions that describe extensions of our current seismoacoustic monitoring capabilities with innovative technologies, network configurations, and observational methods.
Network Operations and Data Centers
This session consists of papers on new developments in seismic and GPS networks and data centers. These include reports on new types of instrumentation, upgrades of existing seismic networks, and data processing. Great data is of no value unless it can get to the user. The session also includes new developments in serving data to the earthquake community.
New Directions in PSHA: Ins, Outs, and Uncertainty
Conveners: Peter Powers, Morgan Moschetti, USGS, and Ned Field, USGS
In light of recent advances in earthquake science, data, and computational resources, probabilistic seismic hazard analyses (PSHAs) have increased in complexity. Such complexity manifests as larger logic trees, more sophisticated data-integration techniques, better understanding of input data uncertainties, and the computational capacity to undertake high-resolution simulations (e.g., CyberShake) and inversions for earthquake rate models (e.g. UCERF3). We invite papers on topics that include uncertainty analysis in PSHA, new approaches to gridded, area, and finite-fault source representations, logic-tree analysis and trimming, deaggregation, site-specific methodologies, computational algorithms, simulation based methods, urban hazard maps, new source and ground motion models, and other research or approaches related to PSHA.
New Insights into Ground Failure as an Urban Earthquake Hazard
Conveners: Laurie Baise, Tufts U and Keith Knudsen, USGS
The 1964 Alaska Earthquake was significant for many reasons – including extensive observations of ground failure. As we celebrate the 50th anniversary, we will look at recent earthquakes, including the 2011 Tohoku earthquake and the 2010-2011 Christchurch Earthquake that have provided new datasets on ground failures. These new datasets facilitate research that improves ground failure models and provides insight into ground failure mechanisms. This session will highlight developments in ground failure models with special focus on urban earthquake hazard. Among the ground failure hazards to be addressed are surface rupture, liquefaction and landsliding.
Pillars of Simulation: Seismic Velocity and Material Models
Conveners: Ricardo Taborda, U Memphis, Po Chen, U Wyoming, and En-Jui Lee, U Wyoming
Various research and application areas in geophysics, seismology and engineering are strongly dependent on the seismic velocity and material models used in simulation today. While much progress has been done over the last decades in advancing such models, our knowledge about the geology and structure of the crust and its sedimentary basins and the near-surface geotechnical layers, as well as our modeling approaches to represent the mechanical properties of materials beneath the surface are somewhat limited. This session solicits contributions in the broad area of seismic velocity and material models and their application in earthquake (source and ground-motion) simulation. Topics of particular interest include but are not limited to: development of new and improvement of currently available community velocity models, description of velocity models software, discrete representation of velocity models used for simulation, verification and validation of velocity models, full waveform tomography and inversion, attenuation and/or quality factor rules, visco-elastic material models, off-fault and near-surface plasticity, initial stress conditions and elasto-visco-plastic models in simulations including material nonlinear behavior, integration of geotechnical data, and representation of heterogeneous media.
Recent Advances and Findings in Earthquake Geology and Paleoseismology
Conveners: Scott Bennett, USGS and Christopher DuRoss, Utah Geological Survey
Characterizations of the timing, recurrence, displacement, and rupture extent of recent surface-faulting earthquakes provide key inputs for earthquake-probability forecasts and seismic-hazard assessments. This session will highlight recent advances in the fields of earthquake geology and paleoseismology from investigations of Quaternary-active faults. We invite abstracts that present 1) paleoseismic earthquake histories and improved estimates of earthquake recurrence, fault slip rate, and rupture extent; 2) techniques for investigating surface ruptures, such as remote sensing of surface deformation and offset (e.g., InSAR, GPS, and pixel tracking); and 3) objective methods of synthesizing multiple paleoseismic datasets. The goal of this session is to critically evaluate how detailed fault and earthquake characterizations inform our understanding of seismic hazard.
Recent Advances in Ground Motions Simulation Methods and Their Validation
Conveners: Christine Goulet, University of California Berkeley and Kim B. Olsen, San Diego State University
This session is targeted to highlight recent advances in the development and validation of ground motion simulation methods. Topics of interest include: formulation of new or updated simulation methods with description of parameterizations of the source, path and/or (nonlinear) site effects; models including small-scale heterogeneity in the source or surrounding medium or incorporating 3D complexity of the medium and/or mountain topography; methods for better constraining input; sensitivity of methods to ranges of plausible input parameters; and quantification of uncertainty in simulated ground motions. We also invite case histories comparing ground motions from various simulation techniques, as well as calibration and validation of methods against recordings and/or ground motion prediction equations. The session is open to a wide range of simulation methodologies, including deterministic and stochastic models of wave propagation, with stochastic, kinematic or dynamic models of the source description.
Seismic Imaging as USArray Moves to Alaska (IRIS/Seismological Society of America Joint Session)
Conveners: Carene Larmat, Los Alamos National Laboratory, Frank Vernon, U California, San Diego, and Bob Woodward, IRIS, and Lindsay Worthington, U New Mexico
In 2014 the USArray Transportable Array will begin deploying in Alaska, after ten years successfully crossing the entire lower-48 states. The arrival of the Transportable Array in Alaska presents an unprecedented opportunity to image one of the most seismically and tectonically active regions of the world. We encourage contributions highlighting results achieved already with USArray data or which provide the context and framework for discoveries to come. The goals of the session are to identify opportunities for characterizing structures of the crust (including basins and volcanoes), slab, and mantle using complementary data sets. Example data sets include onshore and offshore data, active and passive seismic data, ambient noise and earthquake data (local or teleseismic), and potential field data. We encourage submissions featuring techniques that harness large emerging data sets and modern computational methods.
Seismic Location and Processing Techniques
Conveners: Natasha Ruppert, Geophysical Institute, University of Alaska
Modern earthquake location and monitoring techniques have revolutionized seismology, allowing better images of fault structure, stress, and time-dependent frictional properties. This session explores new techniques and results in this important seismological field.
Site Response: From Site-Specific Analyses to Predictive Models around the Globe
Conveners: Eric Thompson, San Diego State U and Dominic Assimaki, Georgia Institute of Technology
We invite contributions that fall within the broad category of site response analysis, including topics from advances in fundamental soil behavior modeling to simplified methods for mapping site response at the national and global scales. Site response has been a hotly debated topic between engineers and seismologists for decades. In the recent years, the two communities have reached consensus on the nonlinear nature and importance of site response as a constituent of ground motion predictions and seismic hazard assessment; faced with the ever increasing amount of recorded evidence in statistical studies, or pushing the envelope of high frequency ground motion simulations, many seismologists currently account for nonlinear soil effects in their studies. Thus, the classic debate has shifted towards more subtle and interconnected issues, such as the use of the average shear-wave velocity to 30 m (Vs30), empirical vs. theoretical modeling approaches, time-domain vs. spectral analyses, and the accuracy of fundamental modeling assumptions. Although different approaches can be cast as controversies, it is useful to see them as solutions to different problems rather than competing solutions to the same problem. For example, although a nonlinear time-domain analysis more accurately accounts for soil behavior than statistical correlations with Vs30, the latter approach may be the only option available for applications at sites where dynamic soil properties are unavailable. In this context, we encourage submissions to focus on the motivations behind their modeling choices, and to discuss the applications and scales where their approach is expected to be most beneficial.
Topics in Seismology: Hazards
A primary incentive to understanding earthquakes is in evaluating the hazard. This session includes presentations on a diverse array of hazard analysis, from precariously balanced rocks, to nuclear plants, to the L'Aquilla trial, and the utility of response spectra.
Topics in Seismology: Processes
Processes that lead to seismicity are diverse and varied from common slip of rocks along faults to ice quakes to fluid-induced seismicity. This session presents research on processes leading to seismicity.
Topics in Seismology: Regional Seismicity and Tectonics
A better evaluation and understanding of earthquake hazards can only come from regional studies of seismicity and active deformation. This session presents research on regional seismicity from around the globe.
Tracking Fluid Movement in Volcanic Systems
Conveners: John Lyons, USGS, Helena Buurman, U Alaska Fairbanks, and Diana Roman, Carnegie Institute, and David Fee, U Alaska Fairbanks
Seismology has long been the primary means through which to study and monitor the movement of magma and other fluids in active volcanic systems. However, despite decades of seismic monitoring at volcanoes, the ascent of magma and the circulation of fluids within volcanic systems, and how these phenomena are reflected in geophysical signals, remain poorly understood. In recent decades, improvements in instrumentation and processing techniques have led to the development of additional geophysical tools capable of tracking fluid movement, including infrasound, high-rate GPS, InSAR and gravity. These tools, often used in concert with seismological techniques, have brought forth many new insights that were previously unknown. We seek submissions highlighting interdisciplinary geophysical monitoring and study of fluid movement in active volcanoes and geothermal systems. We encourage contributions that emphasize advances in numerical modeling, feature new instrumentation or analytical methods, and/or provide new insights into the physical processes controlling fluid movement.
Session:Large and Damaging Earthquakes of 2013/2014
Presenter   Hobbs, Tiegan
Schedule   Thu 8:45 AM / Oral
Room   Room 2
STUDENT
Coulomb Stress Changes Resulting from the Mw 7.7 2012 Haida Gwaii Earthquake
HOBBS, T. E., University of Victoria, Victoria, BC, Canada, tiegan.hobbs@gmail.com; BRILLON, C., Geological Survey of Canada (Pacific), Sidney, BC, Canada, camille.brillon@nrcan-rncan.gc.ca; CASSIDY, J. F., Geological Survey of Canada (Pacific), Sidney, BC, Canada, John.Cassidy@nrcan-rncan.gc.ca; DRAGERT, H., Geological Survey of Canada (Pacific), Sidney, BC, Canada, herb.dragert@nrcan-rncan.gc.ca; DOSSO, S. E., University of Victoria, Victoria, BC, Canada, sdosso@uvic.ca
This study examines spatial changes to the local stress field resulting from the October 28, 2012, Mw 7.7 Haida Gwaii earthquake, which occurred west of Moresby Island off the coast of British Columbia. This event is thought to have occurred on a NE-dipping, blind thrust fault rather than on the sub-vertical Queen Charlotte Fault that represents the Pacific - North American plate boundary. This was the largest earthquake along the Canadian portion of this plate boundary since the 1949 Ms 8.1 Queen Charlotte earthquake. The USGS software `Coulomb' is used to quantitatively estimate the effect of the mainshock on the background stress field, the known aftershock nodal planes, and the nearby Queen Charlotte Fault. We use two different mainshock finite fault models, both of which are seismologically-derived (by Lay and Hayes, separately) and subsequently adapted by K. Wang to account for the motion detected at four nearby GPS stations. We also use the best-located set of aftershocks with information provided by a temporary array of ocean bottom seismometers. Preliminary results indicate an apparent clustering of aftershocks slightly seaward of the main thrust, which is consistent with the modeled zone of promoted normal failure likely related to extension in the footwall. Using existing models, we have found a high number of aftershocks to be consistent with triggering by the mainshock, suggesting that static stress is a dominant control in the months following a large earthquake in this area. Additionally, we find loading, greater than the threshold for triggering, on the Queen Charlotte Fault in an area believed to be within a seismic gap. This work provides an improved understanding of the evolving seismic hazard along the Queen Charlotte margin, while simultaneously testing the usefulness of Coulomb modeling in this environment.
Session:Recent Advances and Findings in Earthquake Geology and Paleoseismology
Presenter   Hobbs, Tiegan
Schedule   Fri / Poster
Room   Cook/Arteaga
STUDENT
Rupture Properties of the Mw 7.7 2012 Haida Gwaii Earthquake from an Empirical Green's Function Method
HOBBS, T. E., University of Victoria, Victoria, BC, Canada, tiegan.hobbs@gmail.com; CASSIDY, J. F., Geological Survey of Canada (Pacific), Sidney, BC, Canada, john.cassidy@nrcan-rncan.gc.ca; DOSSO, S. E., University of Victoria, Victoria, BC, Canada, sdosso@uvic.ca
This study examines rupture properties of the October 28, 2012, Mw7.7 Haida Gwaii earthquake off the coast of British Columbia using an empirical Green's function (EGF) technique. Surface waveforms from a 2001 Mw6.3 event, which ruptured only 15 km from the 2012 epicenter with an almost identical mechanism, are used as an EGF to be deconvolved from those of the 2012 mainshock. The resulting source-time function is free of path effects and instrument response, so the waveform displays only properties of the rupture itself. By examining azimuthal variations in these source-time functions we can constrain parameters such as average rupture velocity, rupture extent, and directivity. Additionally, information can be gathered about the possible existence of major sub-events and their locations. Preliminary results indicate that the overall rupture extends to the northwest for approximately 135-170 km along a direction similar to the strike (~320°). These findings are important, given that earthquakes with strong directivity, such as the 2002 Mw7.9 Denali earthquake, have been shown to be capable of triggering earthquakes thousands of kilometers away. One important question to be addressed is whether there is a link between the Haida Gwaii event and the Mw7.5 earthquake at Craig, Alaska (350 km northwest), just two months later. The Haida Gwaii earthquake was the largest event along the Canadian portion of the Pacific - North America plate boundary since the Ms8.1 Queen Charlotte earthquake of 1949. Focal mechanisms in this region are predominantly right-lateral strike-slip but with an element of oblique convergence off Moresby Island. This work aims to elucidate important information about the rupture process, which is thought to have involved a blind thrust fault dipping gently to the NE rather than the main, sub-vertical Queen Charlotte Fault.
Session:New Directions in PSHA: Ins, Outs, and Uncertainty
Presenter   Apel, Edwin (Trey)
Schedule   Fri 4:30 PM / Oral
Room   Room 2
Effects of Various Declustering and Smoothing Methods on Earthquake Catalogs and Consequences for Seismic Hazard and Seismic Risk in New Zealand
APEL, E. V., Risk Management Solutions, Newark, CA, USA, trey.apel@rms.com; NYST, M., ,
We apply different declustering methods, temporal completeness estimates, and smoothing techniques to the New Zealand earthquake catalog to estimate regional seismicity rate changes and the consequent impact on earthquake hazard and risk in the area. We explore the sensitivity of the declustering parameters (e.g. Reasenberg, 1985) as well as the variable magnitude and time windows (e.g. Gardner and Knopoff, 1974) on the magnitude-frequency relationships derived from declustered catalogs for the background zones in New Zealand. In addition, we test the effect of temporal completeness interval estimates and smoothing methods. We incorporate background rate estimates from the full catalog and compare them with the range of background rates from all of the declustering, completeness, and smoothing methods to our hazard model. We then compare a suite of metrics between the full catalog rates and the models with different rates to assess 1) the impact on hazard and risk and 2) the epistemic uncertainty associated with the range of catalog declustering and smoothing techniques. In areas where the seismic hazard is dominated by larger crustal faults (e.g. Wellington) the choice of declustering and smoothing methods has little to no impact on the hazard or risk. However, in areas like Auckland where the background zone contribution to the seismic hazard is greater, the choice of techniques is more significant (Christophersen et al., 2011). However, the impact of the seismic risk is less significant due to the lower estimated ground motions.
Session:Recent Advances and Findings in Earthquake Geology and Paleoseismology
Presenter   Allen, Trevor
Schedule   Fri 2:00 PM / Oral
Room   Room 3
A Tale of Two Seisms Down Under: 23 March 2012 Mw 5.4 Ernabella and 9 June 2013 Mw 5.6 Mulga Park
CLARK, D. J., Geoscience Australia, Canberra, ACT, Australia, dan.clark@ga.gov.au; MCPHERSON, A., Geoscience Australia, Canberra, ACT, Australia, andrew.mcpherson@ga.gov.au; ALLEN, T., Pacific Geoscience Centre, Geological Survey of Canada, Sidney, BC, Canada, Trevor.Allen@NRCan-RNCan.gc.ca
On 23 March 2012 a MW 5.4 intraplate earthquake occurred in the eastern Musgrave Ranges of north-central South Australia, near the community of Ernabella (Pukatja). This was the largest earthquake recorded on mainland Australia in the past 15 years and resulted in the formation of a 1.6 km-long surface deformation zone that included reverse fault scarps with a maximum vertical displacement of ~0.5 m (average ~0.1 m), extensive ground cracking, and numerous rock falls. Fifteen months later, on 09 June 2013 a MW 5.6 earthquake (the Mulga Park earthquake) occurred ~15-20 km northwest of the 2012 rupture. The P-axes of the focal mechanisms constructed for both events indicate northeast-oriented horizontal compressive stress. However, the focal mechanism for the Mulga Park earthquake suggests strike-slip failure, with a sub-vertical northerly-trending nodal plane favoured as the failure plane, in contrast to the thrust mechanism for the 2012 event. Despite being felt more widely than the 2012 event, ground cracking and minor dune settlement were the only surface expressions relating to the Mulga Park earthquake. No vertical displacements were evident, nor were patterns indicative of significant lateral displacement. An 18 km long north to north east trending arcuate band of moderate to high cracking density was mapped parallel to the surface trace of the Woodroffe Thrust, a major crust-penetrating fault system. A lobe of high-density cracking ~5km long, coincident with the calculated epicentral location, extended to the north from the centre of the main arc. We speculate that the rupture progressed to the south beneath the northern high-density lobe (consistent with the dimensions expected from new scaling relations), and that the larger arcuate band of cracking might relate to positive interference resulting from reflection of energy from the Woodroffe Thrust interface. Both events provide new insight into the rupture behaviour of faults in non-extended cratonic crust.
Session:Seismic Imaging as USArray Moves to Alaska (IRIS/Seismological Society of America Joint Session)
Presenter   Kim, YoungHee
Schedule   Thu 2:00 PM / Oral
Room   Room 4
Imaging the Megathrust Zone and Yakutat/Pacific Plate Interface in the Alaska Subduction Zone
KIM, Y., Seoul National University, Seoul, S. Korea, younghkim@snu.ac.kr; ABERS, G. A., LDEO, Columbia University, Palisades, NY, USA, abers@ldeo.columbia.edu; LI, J., LDEO, Columbia University, Palisades, NY, USA, jiyao.lee@gmail.com; CHRISTENSEN, D., University of Alaska Fairbanks, Fairbanks, AK, USA, doug@giseis.alaska.edu; CALKINS, J., LDEO, Columbia University, Palisades, NY, USA, jcalkins@ldeo.columbia.edu; RONDENAY, S., University of Bergen, Bergen, Norway, rondenay@geo.uib.no
We image the subducted slab underneath a 450 km long transect of the Alaska subduction zone. Densely spaced stations in southern Alaska are set up to investigate (1) the geometry and velocity structure of the downgoing plate and their relationship to slab seismicity, and (2) the interplate coupled zone where the great 1964 earthquake exhibited the largest amount of rupture. The joint teleseismic migration of two array datasets based on teleseismic receiver functions (RFs) reveals a prominent, shallow-dipping low-velocity layer at ~25-30 km depth in southern Alaska. Modeling of RF amplitudes suggests the existence of a thin low-velocity layer (Vs of ~2.1–2.6 km/s) that is ~20-40% slower than underlying oceanic crustal velocities, and is sandwiched between the subducted slab and the overriding North America plate. The observed low-velocity megathrust layer may be due to a thick sediment input from the trench in combination with elevated pore fluid pressure in the channel. Both velocities and thickness of the low-velocity channel increase downdip in central Alaska, in agreement with previously published results. Our image also includes an unusually thick low-velocity crust subducting with a ~20 degree dip down to 130 km depth at approximately 200 km inland beneath central Alaska. The unusual nature of this subducted segment results from the subduction of the Yakutat terrane crust. We also show a clear image of the Yakutat and Pacific crust subduction beneath the Kenai Peninsula, and the along-strike boundary between them at megathrust depths. Our imaged western edge of the Yakutat terrane, at 25-30 km depth in the central Kenai along the megathrust, aligns with the western end of a geodetically locked patch with high slip deficit, and coincides with the boundary of aftershock events from the 1964 earthquake. It appears that this sharp change in the nature of the downgoing plate could control the slip distribution of great earthquakes on this plate interface.
Session:Fifty Years of Tsunami Science: from the 1964 Earthquake and Tsunami to the SAFRR Tsunami Scenario - Advances in Tsunami Source Characterization, Numerical Analysis and Hazard Mitigation
Presenter   Melgar, Diego
Schedule   Fri 4:30 PM / Oral
Room   Room 4
Joint Earthquake Source Inversion of Land- and Ocean-based Geophysical Sensors for Tsunami Early Warning
MELGAR, D., Scripps Institution of Oceanography, La Jolla, CA, USA, dmelgarm@ucsd.edu; BOCK, Y., Scripps Institution of Oceanography, La Jolla, CA, USA, ybock@ucsd.edu
Computation of tsunami intensities in the regions adjacent to large earthquakes immediately after rupture remains a challenging problem. Seismological instrumentation in the near field cannot be objectively employed for real-time inversions and the non-unique source inversion results are a major concern for modelers. Employing near-field seismic, GPS and ocean wave gauge data from the M9 Tohoku-oki earthquake, we test the capacity of static finite fault models to produce reliable tsunami forecasts. We demonstrate the ability of seismogeodetic source models determined from combined land-based GPS and strong motion seismometers to forecast near-source tsunamis in ~3 minutes after origin time (OT). We show that these models, based on land-borne sensors tend to underestimate the tsunami but are good enough to provide a realistic first warning. We show that rapid ingestion of offshore shallow water (100 - 1000 m) wave gauge data significantly improves the model forecasts and possible warnings. We ingest data from ocean-bottom pressure sensors and GPS buoys into the inversion process. Tsunami Green functions are generated using the GeoClaw package, a benchmarked finite volume code with adaptive mesh refinement. These are used for a joint inversion with the land-based data and substantially improve the earthquake source and tsunami forecast. Model skill is assessed by detailed comparisons of the simulation output to 2000+ tsunami runup survey measurements collected after the event. We update the source model and tsunami forecast at 10 min intervals. By 20 min after OT the tsunami is well-predicted with a high variance reduction to the survey data and by ~30 minutes a final model is obtained with little changed observed afterwards. This approach minimizes operator interaction, it relies on a finite-extent CMT to distinguish among strike-slip, normal and thrust faulting events, all of which have been observed in subduction zones. and pose distinct hazards.
Session:Site Response: From Site-Specific Analyses to Predictive Models Around the Globe
Presenter   Cabas, Ashly
Schedule   Wed 8:45 AM / Oral
Room   Room 4
STUDENT
The Importance of the Elastic Half Space Assumption in Site Response Analysis
CABAS, A., Virginia Tech, Blacksburg, VA, USA, amcabas@vt.edu; RODRIGUEZ-MAREK, A., Virginia Tech, Blacksburg, VA, USA, adrianrm@vt.edu; GREEN, R. A., Virginia Tech, Blacksburg, VA, USA, rugreen@vt.edu
The influence of local soil conditions on the nature of earthquake damage has been recognized by seismologists and engineers for many years, especially after the Mexico City and Loma Prieta seismic events in 1985 and 1989, respectively. One of the fundamental decisions when conducting an equivalent linear site response analysis involves the determination of the depth and characteristics of the elastic half space (EHS). This boundary condition is usually defined at bedrock level and implies that the portion of the energy of downward-propagating seismic waves that are refracted into the EHS will be completely absorbed by the rock mass. Thus, the waves reflected from deeper material boundaries in the actual geologic profile are not considered in the numerical site response. For sites where a strong soil-rock impedance contrast is present, the EHS is typically set at this depth, with the EHS shear wave velocity often assumed to be approximately 800 m/s in the western US and approximately 3000 m/s in the eastern US. However, a case history of a site in the US that has significant shear wave velocity reversals (as a result of multiple sequences of basalt flows and sedimentary interbeds) will be presented to show that for less usual profiles, the determination of the EHS depth and characteristics is no longer intuitive nor trivial. In fact, results from a series of 1D equivalent linear site response analyses performed for the site, reveal that just the presence of a strong impedance contrast is not enough for determining the appropriate EHS depth in the numerical analyses. A parametric study is conducted to investigate which factors have the strongest influence on the EHS assumption.
Session:Seismic Imaging as USArray Moves to Alaska (IRIS/Seismological Society of America Joint Session)
Presenter   Yu, Chunquan
Schedule   Thu / Poster
Room   Cook/Arteaga
STUDENT
Crustal Thickness beneath the Western United States from Virtual Deep Seismic Sounding (VDSS)
YU, C. Q., EAPS, Massachusetts Institute of Technology, Cambridge, MA, USA, yucq@mit.edu; VAN DER HILST, R. D., EAPS, Massachusetts Institute of Technology, Cambridge, MA, USA, hilst@mit.edu; CHEN, W. P., Ocean College, Zhejiang University, Hangzhou, Zhejiang, China, wpchen@illinois.edu
The western United States has a complex tectonic history. Yet, the underlying geodynamic processes are still subjects to debate. In this study, we apply the newly developed virtual deep seismic sounding (VDSS) method to image the crust-mantle interface beneath the western United States, taking advantage of the wide-coverage USArray Transportable Array and much denser regional seismic arrays. VDSS has several inherent advantages over traditional receiver functions and active source seismic reflections and refractions. First, VDSS mainly focuses on the prominent, post-critical reflection, SsPmp, which can be as large as the direct S phase. To this end, one high-quality event is usually enough to image the crust-mantle boundary. Second, VDSS is less sensitive to near-surface structures, e.g. sedimentary cover, as the primary SsPmp phase is strong enough to overcome such signal-generated noise. Thirdly, when the crust-mantle boundary is gradational in nature, the SsPmp phase still provides a robust estimate of an average crustal thickness, because post-critical reflection is insensitive to details of the reflector. Results from this study reveal several distinct features of crustal structures in the western United States. We find significant crustal reworking along the Yellowstone - Snake River Plains system, the High Lava Plains, and the transition zone between the Great Basins and the Colorado Plateau. Interestingly, these regions are all characterized by active late Cenozoic magmatic volcanism. In contrast, The Wyoming craton is underlain by a rather thick crust, consistent with its longevity and high elevation. Although the Colorado Plateau is underlain by a thick crust within its interior, its edges, especially the western and southern parts, turn out to be encroached and thinned.
Session:Earthquake Physics and Interaction
Presenter   Wang, Jeng-Cheng
Schedule   Fri / Poster
Room   Cook/Arteaga
The Seismic Quiescence Phenomenon in Taiwan
WANG, J. C., Department of Applied Geoinformatics, Chia-Nan University, Tainan, Taiwan, ROC, jcwang@mail.chna.edu.tw; SHIEH, C. F., Institute of Seismology, National Chung-Cheng University, Chiayi, Taiwan, ROC, seifent@eq.ccu.edu.tw; WANG, J. H., Institute of Earth Sciences, Academia Sinica, Taipei, Taiwan, ROC, jhwang@earth.sinica.edu.tw
To detect the anomaly of seismic activities of large earthquakes in Taiwan, we first analyzed earthquake catalogue to obtain the minimum magnitude, Mc=2.4, for completeness of the catalogue. This means that earthquakes with magnitudes greater than Mc are complete and then can be used for the computation of seismicity rate changes, which are represented by the Z-values (with positive and negative values representing, respectively, active and inactive seismicity). To study the precursory phenomena, we examined pre-seismic seismicity prior to 65 large inland and offshore earthquakes with M>6.0 occurred during 1990 to 2013. After eliminating the obvious interactions between events, the Z-values are calculated for 7 inland and 10 offshore mainshocks. From the complete earthquake catalogue, the aftershocks of two very large earthquakes can influence the Z-values of related main shocks. Hence, 8 main shocks are removed. The variations of Z-values for 2 inland events show the appearance of anomalous states of seismic quiescence in 20 or more years prior to the events. Meanwhile, the length of time interval of seismic quiescence is positively related to the magnitude of the main shock. By contrast, precursory quiescence was absent or cannot be clearly recognized for offshore main shocks.
Session:Advances in Seismic Imaging and Monitoring of Time-Dependent Variations: Civil Structures, Near-Surface, and Shallow Crustal Scales
Presenter   Haney, Matthew
Schedule   Wed 10:45 AM / Oral
Room   Room 3
The Virtual Eruption of Okmok Volcano, Alaska
HANEY, M. M., U.S. Geological Survey/Alaska Volcano Observatory, Anchorage, AK, USA, mhaney@usgs.gov
When Okmok Volcano erupted violently in 2008, the new cone it formed swallowed a broadband seismometer within the caldera at site OKCD. The destruction of the seismometer constituted a short-term loss in terms of monitoring capability during the ensuing month-long eruption; however, in hindsight it has allowed the volcanic tremor source to be investigated using a novel technique in seismic interferometry known as the virtual earthquake approach (Denolle et al., 2013; JGR). In the case of Okmok, we refer to the technique as the virtual eruption. The virtual eruption is based on the ability to obtain accurate Green’s functions between the eruptive source and seismic stations using ambient noise cross-correlations. At Okmok, the local network consisted of 7 working stations during the 2008 eruption with a total of 11 components. Station OKCD existed for 4 years prior to the 2008 eruption. Over that time period, ambient noise derived Green’s functions can be computed between OKCD and the 7 operative stations from the 2008 eruption. In contrast to synthetic Green’s functions modeled computationally with an assumed subsurface velocity structure, these Green’s functions are exact so long as the noise sources are sufficiently well distributed in azimuth. Path effects due to unknown 3D heterogeneity are captured in the Green’s functions. The 2008 Okmok eruption lends itself nicely to the analysis since low frequency tremor, within the frequency band from 0.2-1 Hz, radiated continuously from beneath the new cone. Backprojection of the low frequency tremor demonstrates that it was predominantly composed of surface waves. Thus, the tremor and ocean noise share the same wave type and frequency range, further enabling the application of the virtual eruption technique. Green’s functions can be obtained for both a force and/or moment source and the eruptive source process estimated continuously over time with waveform inversion.
Session:Fifty Years of Tsunami Science: from the 1964 Earthquake and Tsunami to the SAFRR Tsunami Scenario - Advances in Tsunami Source Characterization, Numerical Analysis and Hazard Mitigation
Presenter   Porter, Keith
Schedule   Fri 2:30 PM / Oral
Room   Room 4
Advances in Tsunami Risk Assessment and Risk Mitigation
PORTER, K. A., University of Colorado at Boulder, Boulder, CO, USA, keith.porter@colorado.edu
Constructing a disaster scenario forces one to identify all potential damage from the peril in question, suppressing the streetlight effect in which one only looks for knowledge where it is easiest. The SAFRR Tsunami Scenario forced us to evaluate damage to several vulnerable asset classes for which tsunami risk models are sparse, especially coastal marinas, ports, bridge abutments, roadways, and to some extent coastal buildings. Scenario development brought to light several analysis and design needs. First, the American Society of Civil Engineers publishes design guidelines for small-craft harbors, which state that because "the frequency of occurrence of a tsunami at any location is very small, small craft harbors do not include design considerations for protection from their appearance." Consequently, pilings in California coastal marinas are too short to restrain floating docks in a reasonably foreseeable tsunami. California coastal ports are similarly exposed to several modes of foreseeable damage. For example, a port dispersal plan created by the US Coast Guard for dispersing the Ports of Los Angeles and Long Beach contains no guidance for decision-making in tsunamis, and there does not appear to be sufficient resources (especially pilots) to disperse vessels in the port within a reasonable tsunami warning time. A modest number of California highway bridges have abutments that intrude into the flow of a tsunami current and could be damaged by scour, and a few stretches of coastal highway are low enough to suffer scour damage in the event of a tsunami, some of them important routes and some of them lacking convenient alternatives during repairs. A large number of coastal buildings--the equivalent of 70,000 single-family dwellings--are close enough to the shore and at low enough elevation to suffer damage in a tsunami. Several modes of damage could be reduced with changes to design guidelines and zoning changes.
Session:Recent Advances and Findings in Earthquake Geology and Paleoseismology
Presenter   Kattenhorn, Simon
Schedule   Fri 4:00 PM / Oral
Room   Room 3
Revised Earthquake Hazard of the Hat Creek Fault, Northern California: Paleoseismology in Pleistocene Lava Flows
KATTENHORN, S. A., University of Idaho, Moscow, ID, USA, simkat@uidaho.edu; BLAKESLEE, M. W., University of Idaho, Moscow, ID, USA
Normal faults in basalt have distinctive surface-trace morphologies where they cut through near-surface lavas, providing information about slip history and earthquake hazards. Such faults may remain active during prolonged volcanic periods such that different ages of lava flows cut by the fault can be used as temporal markers of slip rates and slip history. The 47-km-long Hat Creek fault in northern California is a case example of this fault style. We address segment evolution, slip history, and earthquake potential based on surface morphologies and throw distribution in variable-aged lavas. In response to interaction with sporadically active volcanic systems, surface ruptures progressively migrated westward in the Late Pleistocene, with older scarps being successively abandoned. The most recent earthquake activity dissects 24 ± 6 ka basaltic lavas, forming a vertical scarp with a maximum throw of 56 m. Other lavas cut by the fault have ages of 38 ± 7 ka and 53.5 ± 2 ka. Prior to surface rupture, the fault formed a fault-trace monocline that attained a height of up to 33 m. Subsequent earthquake activity reduced the monocline to a rubble pile along the fault trace. Dislodged small columns thrown up onto the lava surface suggest >1 gn acceleration at the surface rupture location. We show that the length of the most recently active portion of the segmented system, which was suggested in past work to be 23.5 km, is actually comprised of at least 9 segments with a cumulative length of 30 km, thus increasing the potential regional seismic hazard. We estimate that a surface-breaking rupture could produce an earthquake of ~Mw 6.7 with a recurrence interval of 667 ± 167 yr in response to a rapid slip rate in the range 2.2–3.6 mm/yr. This implies a moderate risk given a lack of earthquake events in the past 200 yrs. Our methods for interpreting scarp geomorphology and slip history in faulted lava flows can be applied to any faulted basalt environment.
Session:Advances in Seismic Imaging and Monitoring of Time-Dependent Variations: Civil Structures, Near-Surface, and Shallow Crustal Scales
Presenter   Garcia, Victor
Schedule   Wed / Poster
Room   Cook/Arteaga
3D Shear Wave Model Development of the Texas Region Using an Optimization Scheme
THOMPSON, L. E., University of Texas at El Paso, El Paso, TX, USA, lethompson@miners.utep.edu; VELASCO, A. A., University of Texas at El Paso, El Paso, TX, USA, aavelasco@utep.edu; GARCIA, V. H., University of Texas at El Paso, El Paso, TX, USA, vhgarcia4@miners.utep.edu
Understanding Earth’s tectonic processes requires analyzing the Earth structure. The objective of this work is to expand on a constrained optimization approach for a joint inversion least-squares (LSQ) algorithm to characterize a one-dimensional Earth's structure of Texas with the use of multiple geophysical data sets. In particular, we apply the Multi-Objective Optimization Problem (MOP) Technique to the 1D inversion and use four geophysical data sets: receiver functions, surface wave dispersion curves, teleseismic body wave times, and gravity measurements to obtain a better shear wave model of Texas. For this research, we will use MOP formulation to generate quasi-3D shear wave models to determine regions of Texas at would have high and low velocities within the mantle. We used the MOP optimization technique to generate 3D imagery of Texas in order to help us to determine the petrophysical nature of the Lithosphere-Asthenosphere Boundary (LAB). Through the development of our quasi-3D models using multiple data sets, we will be able to evaluate and analysis the upper mantle seismic velocity structures in both the vertical and horizontal directions, which will be the key to our understanding of the overall tectonic structure of Texas. For this research, we collected and process seismic data (receiver function, surface wave, body wave travel times) from USArray, a National Science Foundation funded EarthScope project, plus the collection of gravity from other efforts. USArray, which is part of the EarthScope experiment, is a 15-year program to place a dense network of permanent and portable seismographs across the continental United States. The quasi 3D velocity models generated from the use of the USArray data that we acquired for the Texas region will provide insight into the tectonic history and physical properties of the Earth structure.
Session:Earthquake Physics and Interaction
Presenter   Gao, Bin
Schedule   Fri / Poster
Room   Cook/Arteaga
The Construction of the Fault Model in Ordos, China
GAO, B., Peking University, Beijing, China, gb310@126.com; WANGQL.BOX@263.NET, Q., Second crust Monitoring and Application Center, Xi'an, China; JIA, K., Peking University, Beijing, China, jk@pku.edu.cn; ZHOU, S., Peking University, Beijing, China, zsy@pku.edu.cn
Based on the main active faults of Ordos and the relocation results of the earthquakes of China from 2003 to 2012, we built a fault model of Ordos and the surrounding areas which included the longitude and latitude of the fault ends, the dip angle and the locking depth of the faults. Furthermore, based on this fault model and the GPS data of this area from 1997 to 2007, we used the least squared method to calculate the slip rate of faults of Ordos by considering the influence of the independent dislocation of different faults on the GPS data and the constraints between the adjacent faults. Thus, we initially built up a dynamic model of the faults of Ordos. And, we compared the results of slip rate of some of the faults with those obtained by other research groups and made a summary.
Session:Large and Damaging Earthquakes of 2013/2014
Presenter   Jia, Ke
Schedule   Thu / Poster
Room   Cook/Arteaga
Possibility of the Independence between the 2013 Lushan Earthquake and the 2008 Wenchuan Earthquake on Longmen Shan fault, Sichuan, China
JIA, K., Peking University, Beijing, China, jk@pku.edu.cn; ZHOU, S., Peking University, Beijing, China, zsy@pku.edu.cn; ZHUANG, J., The Institute of Statistical Mathematics, Tokyo, Japan, zhuangjc@ism.ac.jp; JIANG, C., Institute of Geophysics, China Earthquake Administration, Beijing, China, jiangcs@cea-igp.ac.cn
The relationship between the Lushan Ms 7.0 earthquake occurred on 20 April 2013 and the 2008 Wenchuan Ms 8.0 earthquake is still under debate. One view is that the Lushan earthquake is the strongest aftershock of the Wenchuan earthquake and the converse opinion believes that the Lushan earthquake is an independent event. To clean up this debate, we use the ETAS (Epidemic-type aftershock sequence) model and the stochastic declustering method to obtain the background/independent probabilities of the Lushan earthquake. The results show that, besides triggered aftershocks in its source fault, the Wenchuan earthquake also changed the background seismicity rate in the Lushan region, about 1.3 as the original rate. The proportion of contributions to the occurrence of the Lushan earthquake from the aftershock effect (in the sense of the ETAS model), the increment of the background seismicity, and the original background seismicity are, respectively, 12%, 50% and 38%. Considering the relaxation of the lower crust and upper mantle, the combined (coseismic plus postseismic viscoelastic) Coulomb failure stress (CFS) changes on the initial rupture point of the Lushan earthquake are brought 0.1~0.4 bar closer to failure. Thus, we propose the occurrence of the Lushan earthquake was dramatically affected by the Wenchuan earthquake.
Session:Geometric Complexities Along Strike-Slip Systems: New Insights on Seismic Hazards, Earthquake Behavior, and Fault System Evolution
Presenter   Ren, Junjie
Schedule   Wed 11:15 AM / Oral
Room   Room 2
Rupture Behavior on Two Parallel Strike–Slip Faults: A Case from the Central Longriba Fault Zone, Eastern Tibet
REN, J. J., Institute of Crustal Dynamics, Beijing, China, renjunjie@gmail.com; XU, X. W., Institute of Geology, Beijing, China, xiweixu@vip.sina.com; YEATS, R. S., Oregon State University, Corvallis, OR, USA, yeatsr@geo.oregonstate.edu; ZHANG, S. M., Institute of Crustal Dynamics, Beijing, China
Fault behavior on two parallel faults is worth studying for seismic hazard assessment because it is always related to a large earthquake. The 2008 Wenchuan earthquake that ruptured two faults on the Longmen Shan fault zone provides a case for co-rupture on two parallel thrust faults. Whether two parallel strike-slip faults can rupture simultaneously in a large earthquake remains unknown. The central Longriba fault zone, ~ 200 km northwest of and parallel to the Longmen Shan in eastern Tibet, is composed of two subparallel strands: Longriqu fault in the west and Maoergai fault in the east 20-30 km apart. Displaced landforms indicate that these two strands are dominantly right-lateral strike-slip faults in Late Quaternary. We conducted trenching combined with radiocarbon and Optically Stimulated Luminescence (OSL) dating to reconstruct paleoseismic history. Our results demonstrate that four surface-rupturing events occurred on the Longriqu fault at 5080 ± 90, 11,100 ± 380, 13,000 ± 260, and 17,830 ± 530 cal yr B.P. and three events on the Maoergai fault at 8510 ± 420, 7100 ± 70, and 5170 ± 80 cal yr B.P. If only a single fault is considered, each fault strand fits a cluster model with a moment magnitude (Mw) of ~7.2-7.4. If we consider these two fault strands as a whole, the last event has a good temporal correlation that implies a co-rupturing event equivalent to Mw ~7.6. Prior to the last event, the two strands have an alternative activity, and the central Longriba fault zone appears to undergo a regular recurrence with an interval of ~2000 years. These results suggest that these two fault strands probably ruptured simultaneously in the last larger event and may merge at the seismogenic depth, in accordance with available deep seismic data. This study suggests that two parallel faults linked at the seismogenic depth may rupture together and generate a superquake.
Session:Diverse Mechanisms of Subduction Zone Fault Slip: Exploring the Relationships Among Seismic Rupture, Transient Slip, and Steady Creep
Presenter   Watkins, William
Schedule   Wed / Poster
Room   Cook/Arteaga
STUDENT
Potential Causes for along-Strike Variability of Slow Slip Events in South-Central Alaska
WATKINS, W. D., Indiana University of Pennsylvania, Indiana, PA, USA, w.d.watkins@iup.edu; COLELLA, H. V., Miami University of Ohio, Oxford, OH, USA, hcolella@gmail.com; BRUDZINSKI, M. R., Miami University of Ohio, Oxford, OH, USA; DIETERICH, J. H., University of California-Riverside, Riverside, CA, USA; RICHARDS-DINGER, K. B., University of California-Riverside, Riverside, CA, USA
Slow slip events (SSEs) are observed in subduction zones around the world and exhibit a wide range of recurrence intervals, durations, and spatial extents. A ubiquitous feature of SSEs is the along-strike variability of these characteristics. However the cause and long-term effects of such variability is poorly understood. Additionally, it is unclear whether such variability and segmentation of SSEs persists beyond human time scales. Here we employ the earthquake simulator RSQSim to model a megathrust, which consists of seismogenic, slow slip, and continuous creep sections. The slow slip section is segmented to explore potential causes of along-strike variability in recurrence intervals, durations, and spatial extent, by varying parameters such as the effective normal stress, frictional properties, and slip rates. RSQSim enables simulations of long histories of SSEs over all orders of magnitude, which allows for robust characterization of the effects of variation in parameters. Preliminary results show even small variations in these parameters have a significant effect on observable characteristics of SSEs, which begins to illuminate the primary controls on along-strike variability. For example, a decrease in the effective normal stress from 9MPa to 3MPa results in a decrease in the mean recurrence interval and event duration from 35 to 15 months and 44 to 16 days, respectively, but increases the mean propagation speed from ~7 km/day to ~24 km/day. This research builds on the previous study and considers variations in fault geometry (i.e. tears in the fault) and fault roughness in addition to the parameters discussed above. Ultimately this work aims to reproduce observed SSEs in south-central Alaska.
Session:Topics in Seismology: Regional Seismicity and Tectonics
Presenter   Kim, YoungHee
Schedule   Thu / Poster
Room   Cook/Arteaga
Properties of the Subduction System in Mexico
KIM, Y., Seoul National University, Seoul, S. Korea, younghkim@snu.ac.kr; CLAYTON, R. W., California Institute of Technology, Pasadena, CA, USA, clay@gps.caltech.edu
We exploit dense seismic data (MASE, 2007; VEOX, 2010) to better define the subduction structure beneath Mexico. Our images provide insight into the process of subducting relatively young oceanic lithosphere and its complex geometry beneath continental North America. The converted and reverberated phase image for central Mexico shows complete horizontal tectonic underplating of the Cocos plate beneath the North American plate, with a clear image of a thin low-velocity oceanic crust which dips at 15-20 degrees at Acapulco then flattens to approximately 300 km from the trench. A low-pressure mineral phase such as talc is considered a prime candidate for anomalously low shear speeds at the subducted (top) interface of the Cocos plate. The evolution of this low-strength zone has important implications for the dynamics of the slab-flattening process as well as the geochemistry of the mantle wedge and arc in central Mexico. In southern Mexico, there are several unusual surficial features that cannot be explained by the subduction of the Cocos plate and low-relief relic fracture zone known as the Tehuantepec Ridge. The dip of the Cocos plate is ~25 degrees based on converted analysis, and the signature of the oceanic crust is absent beyond the depth of ~150 km. Our seismic images on the Isthmus of Tehuantepec show an enigmatic seismic structure dipping south from the Gulf of Mexico. The images argue that this structure represent a slab although standard models of Caribbean tectonics do not support this observation.
Session:Great Earthquakes and Slip to the Trench (Seismological Society of Japan/Seismological Society of America Joint Session)
Presenter   Okuwaki, Ryo
Schedule   Thu / Poster
Room   Cook/Arteaga
STUDENT
Relationship between High-frequency Radiation and Asperity Rupture, Revealed by Hybrid Back-Projection with Non-planer Fault
OKUWAKI, R., University of Tsukuba, Japan, rokuwaki@gmail.com; YAGI, Y., University of Tsukuba, Japan, yagi-y@geol.tsukuba.ac.jp; HIRANO, S., University of Tsukuba, Japan, hirano@kz.tsukuba.ac.jp
High-frequency (HF) radiation, in general, is generated by abrupt change of rupture front velocity and/or slip-rate. Therefore, analysis of HF waves is crucial to understand the dynamic rupture process of earthquakes. The hybrid back-projection (HBP) method (Yagi et al., 2012) enables us to obtain the detailed spatiotemporal distribution of HF radiation by stacking cross-correlations of observed waveforms with corresponding Green’s functions, By using the cross-correlation, the HBP method can mitigate the effects of subsurface structure. In this study, we developed a new HBP method by introducing a 3D non-planar fault model based on a subducting plate geometry (Tassara & Echaurren, 2012). We applied it to teleseismic P-waveform data (0.5 - 2.0 Hz) of the 2010 Maule, Chile earthquake (Mw = 8.8) and then compared our result with a coseismic slip distribution obtained by the new waveform inversion (Yagi & Fukahata, 2011). At the beginning, strong HF radiation is generated near the hypocenter, and then large slip event follows near the site of strong HF radiation. After 30 s, intermediate HF radiation area propagated from south to north, which concentrates along the rupture front obtained by the waveform inversion. Around 60 s, the second strong HF radiation occurs ahead of the rupture front obtained by the waveform inversion at 200 km north from the epicenter, and soon after that the second large coseismic slip event follows at about the same location. The strong HF radiation events before the large slip events may be understood in the framework of a cascading asperity model (Ide && Aochi, 2005). The precursory strong HF radiation may be generated by the rupture of a relatively small asperity, which contributes to triggering of rupture of a relatively large asperity. Intermediate HF radiation obtained in-between the first and second large slip events may play a key role in the interaction between the south and the north large slip events.
Session:Pillars of Simulation: Seismic Velocity and Material Models
Presenter   Rollins, John
Schedule   Wed / Poster
Room   Cook/Arteaga
STUDENT
Mechanisms of Postseismic Deformation Following the 2010 M=7.2 El Mayor-Cucapah Earthquake
ROLLINS, J. C., California Institute of Technology, Pasadena, CA, USA, jcrollin@caltech.edu; BARBOT, S. D., Earth Observatory of Singapore, Singapore, sbarbot@ntu.edu.sg; AVOUAC, J. P., California Institute of Technology, Pasadena, CA, USA, avouac@gps.caltech.edu
The Salton Trough is an ideal site in which to use postseismic deformation to study the rheology of the crust and upper mantle. In this region, lithospheric thinning has brought the low-viscosity asthenosphere up to within <50 km of the surface, high surface heat flow implies low viscosities throughout the thin lithosphere, and the transform component of relative plate motion produces large strike-slip earthquakes that impart large stress changes to this low-viscosity structure, likely inducing significant postseismic deformation. The 2010 Mw=7.2 El Mayor-Cucapah earthquake was the largest in the Salton Trough since at least 1892 and the postseismic deformation following this earthquake may provide insights into rheology at depth that are difficult to extract elsewhere. Three-year postseismic displacement timeseries extracted at UNAVCO GPS stations show four key attributes: 1) the Imperial Valley experienced ~2 cm of uplift in both coseismic and postseismic deformation; 2) significant westward displacement occurred in the Peninsular Ranges, suggesting a long-wavelength mechanism; 3) near-field timeseries show a rapid deformation signal in the first year after the mainshock; and 4) that signal is followed by a sustained signal that is still active at the time of writing. Forward modeling of stress-driven deformation reveals that the uplift in the Imperial Valley and the large displacements in the Peninsular Ranges are best fit to Newtonian viscoelastic relaxation in a low-viscosity asthenosphere with geometry matching the regional lithosphere-asthenosphere boundary inferred from receiver functions. The temporal signal in the near field can be fit to a variety of plausible mechanisms and rheologies but, among them, is well fit to a combination of deep afterslip and Newtonian viscoelastic relaxation in the lower crust, both plausible deformation mechanisms for a large strike-slip earthquake in a high-heat-flow region.
Session:Recent Advances in Ground Motions Simulation Methods and Their Validation
Presenter   Pasyanos, Michael
Schedule   Thu 1:30 PM / Oral
Room   Room 3
Using Attenuation in Ground Motion Prediction Equations: Validating Improvements to Strong Ground Motion Predictions
PASYANOS, M. E., Lawrence Livermore National Laboratory, Livermore, CA, USA, pasyanos1@llnl.gov
We have recently developed a lithospheric attenuation model of North America in which we estimate the attenuation quality factors (Qp and Qs) for the crust and upper mantle over a broad frequency band (0.5 – 10 Hz). The attenuation model can be used to improve strong ground motion predictions, as had been previously demonstrated for the 2011 Mineral, VA earthquake. In order to make them easier to test, the models are incorporated into standard 1D ground motion prediction equations (GMPEs), effectively making them 2D. Here, we test our ability to improve ground motion estimates for a much larger validation dataset of earthquakes in central and eastern North America compiled for the NGA-East. This dataset is a diverse set of recordings with a range of magnitudes, distances, tectonic settings, earthquake mechanisms, and ground motion levels. The use of attenuation models in GMPEs improves our ability to fit strong ground motions in the region, sometimes significantly, and should be incorporated in future national hazard maps.
Session:Seismic Location and Processing Techniques
Presenter   Horowitz, Frank
Schedule   Fri 8:30 AM / Oral
Room   Room 3
Hypocentral Relocations Aided by Virtual Receivers Constructed via Seismic Interferometry
HOROWITZ, F. G., Cornell University, Ithaca, NY, USA, frank.horowitz@cornell.edu
A 3D elastic wave propagation program[1] was used to investigate whether the technique of[2] can improve hypocentric relocations by employing virtual receivers (VRs) near a cloud of microearthquakes. The VR technique can be loosely described as a "dual" of the ambient noise technique from seismic interferometry - replacing noise sources on the boundary of a region of interest with physical receivers. Seismograms from events in the interior of the region of interest can be cross-correlated and integrated over all boundary receivers to estimate a seismogram from one of the interior events as if it were recorded at the location of another interior event. Unlike ambient noise interferometry, where raypaths from all directions impinge on the region of interest, VR raypath directions are constrained by the location of the physical receiver array. Hence, approximating the surface integral plays a large role in the practical success of the technique. Fortunately, stationary-phase arguments suggest that only a few physical receivers nearby the interior-source to virtual-receiver ray direction suffice to reconstruct the seismogram[2]. Arrival time error statistics and relocations of perturbed synthetic hypocenters using virtual receiver arrivals will be shown at the meeting. Improved performance in hypocc[3] - by replacing the embedded LSQR solver with LSMR[4] - will also be discussed. References: [1] Petersson, N. A., & Sjogreen, B. (2011). User's guide to WPP version 2.1.5. LL Nat'l Lab. [2] Curtis, A., Nicolson, H., Halliday, D., Trampert, J., & Baptie, B. (2009). Virtual seismometers in the subsurface of the Earth from seismic interferometry. Nature Geosci., 2 (10), 700-704. [3] Foulger, G. R., & Julian, B. R. (2013). Seismological software for geothermal monitoring. SGP-TR-198. Stanford University. [4] Fong, D. C., & Saunders, M. (2011). LSMR: An iterative algorithm for sparse Least-Squares problems. SIAM Journal on Scientific Computing, 33 (5), 2950-2971.
Session:Seismic Imaging as USArray Moves to Alaska (IRIS/Seismological Society of America Joint Session)
Presenter   Tape, Carl
Schedule   Thu 1:30 PM / Oral
Room   Room 4
Seismic Velocity Structure and Anisotropy of the Alaska Subduction Zone Derived from Surface Wave Tomography
WANG, Y., U. Alaska Fairbanks, Fairbanks, AK, USA, ywang@gi.alaska.edu; TAPE, C., U. Alaska Fairbanks, Fairbanks, AK USA, carltape@gi.alaska.edu
The Aleutian-Alaskan subduction zone is extremely active and driven by northwesterly subduction and collision of the Pacific/Yakutat plate. Better structural models of the crust and upper mantle are needed in order to constrain geodynamic models, which have examined the mantle flow at the eastern edge of the subduction zone. Tomographic studies have been performed based on short-period P and S body-wave traveltimes, either from local or global events, but Alaska structure remains unclear, partly because it has only been imaged from a small portion of the seismic wavefield. In this study, surface waves are used for the first time to investigate the structure of the Alaska-Yakutat-Wrangell subduction system at the eastern end of the Aleutian-Alaska subduction zone. Rayleigh waves are used within a two-plane wave representation with finite frequency sensitivity kernels to image the seismic velocity structure of the crust and upper mantle. The data set comprises 400 earthquakes Mw >= 6.0 at epicentral distances 30-120 degrees, and 168 broadband stations within Alaska (both permanent and temporary stations), for a total of 9865 source-station paths. Cross-correlation phase and amplitude measurements of Rayleigh waves over periods 22 s to 143 s provide maximal sensitivity to structures between in the depth range ~30-200 km. The 3D shear velocity model reveals a fast velocity anomaly associated with the subduction zone in the depth range of ~90-210 km and likely deeper. The onshore portion of the Yakutat block is associated with a slow anomaly at depths shallower than 70 km. Azimuthal anisotropy of phase velocity at 40 s shows an NW-SE fast direction in the Gulf of Alaska that continues into the backarc, which agrees with the convergence direction between the Pacific and North America plates. Anisotropy also shows abrupt fast direction change surrounding subsurface boundary of the Yakutat block, as identified by Eberhart-Phillips et al. (2006).
Session:Recent Advances and Findings in Earthquake Geology and Paleoseismology
Presenter   He, Zhongtai
Schedule   Fri / Poster
Room   Cook/Arteaga
STUDENT
Comprehensive Study on Holocene Paleoearthquakes in Daqingshan Piedmont Fault, Inner Mongolia of China
HE, Z. T., Institute of Crustal Dynamics, CEA, Beijing, China, hezhongtai@126.com; MA, B. Q., Institute of Crustal Dynamics, CEA, Beijing, China; HOU, J. J., School of Earth and Space Sciences, Peking University, Beijing, China, houjj@pku.edu.cn
Daqingshan piedmont fault is a normal dip-slip fault. It spreads along south piedmont of Daqingshan in Inner Mongolia of China. This fault happened many frequent great paleoearthquakes in Holocene. Paleoseismic events of this fault since Holocene were revealed through digging and trench techniques. This paper studies Holocene paleosols on the paleo-alluvial-pluvial fans of hanging wall and gullies geomorphology of footwall which record the fault activity history. On the hanging wall we study alluvial fans’ sections containing the ancient soil layers at intervals. The results show that Daqingshan piedmont alluvial fans in Holocene develop 3 periods ancient soil layers. We assume that in the same period of soil development, the development of paleosol formation was interrupted and formed in similar time multiple interbeds of paleosol and gravel layers, which are caused by the fault activities. That is, the gravel layer between two adjacent paleosol layers represents an ancient earthquake event. Thus we can date paleosol layers which are up and down the gravel layer to define paleoseismic events. On the footwall we extract gullies from high resolution IRS-P5 DEM data. 25 gullies across the fault on the footwall have been extracted from DEM. Knickpoints caused by fault activities in Holocene have been extracted from profiles of gullies for further step. Combining with the retreat distances and the retreat rates of knickpoints on the gullies, we get forming time of each rank knickpoints on the the fault. Thus, paleo-earthquake series in Holocene of the fault have been obtained. We get Holocene paleoearthquakes records from both walls of the faults. Then we contrastively analyse the results with trenches results together. The result demonstrates that paleosols ages on the Paleo-alluvial-pluvial fans and knick points sequence on the footwall have very close correspondence relations with paleoseismic events along the Daqingshan piedmont fault.
Session:Pillars of Simulation: Seismic Velocity and Material Models
Presenter   Jiang, Wenliang
Schedule   Wed / Poster
Room   Cook/Arteaga
Detailed Crustal Structure of the North China and Its Implication for Seismic Activity
JIANG, W. L., Institute of Crustal Dynamics, China Earthquake Administration, Beijing, China, jiang_wenliang@163.com; ZHANG, J. F., Institute of Crustal Dynamics, China Earthquake Administration, Beijing, China, zhangjingfa@hotmail.com; WANG, X., Institute of Crustal Dynamics, China Earthquake Administration, Beijing, China, wangxinjapan@163.com; TIAN, T., Institute of Crustal Dynamics, Beijing, China
Since the Mesozoic-Cenozoic era the North China Craton has experienced an important tectonic transition which gave rise to complicated crustal structure and strong earthquake activity. Based on the large-scale surface gravity data, we studied the detailed crustal structure and seismogenic mechanism of the North China. The results indicate that the North China presents typical characteristics of adjoining depression and uplift, alternating basins and hills, inhomogeneous density and also great differences in crustal structure and Moho topography. The upper and middle crustal structures are dominated by the NNE-striking tectonic units, with many faults cut down to the middle crust. The lower crust is characterized by the folding-structure, with high- and low-density placed alternately from west to east, presenting lateral heterogeneous feature. Adjusted by the gravity isostasy, Moho topography of the North China fluctuates greatly. Compared with the North China Basin, crustal thickness in the Western Taihang, northern Yanshan and Luzhong areas are much thicker while those densities are lower than the North China Basin. The dominating tectonic direction of the Moho topography strikes NE to NNE and undulates alternately from west to east. Gravity anomalies in the upper and middle crusts of the mountainous areas are related to the surface topography. The epicenters are mostly concentrated in the upper and middle crusts, especially the transitional areas between the high- and low-gravity anomalies. The Tancheng earthquake in 1668, Sanhe earthquake in 1673, Tangshan earthquake in 1976, and all other seismic tectonic zones of the North China are all distributed in areas where magma moves strongly beneath the crust, which is considered to be related to the movement of the high density, unstable and heat flows along the deep passage from the uppermost and asthenosphere due to the subduction of the Pacific slab towards the Eurasian plate.
Session:Earthquake Physics and Interaction
Presenter   Prieto, German
Schedule   Fri 4:15 PM / Oral
Room   Room 7/8
Seismic Evidence for Thermal Shear Runaway during Intermediate-Depth Earthquake Rupture
PRIETO, G. A., Earth, Atmospheric, and Planetary Sciences, MIT, Cambridge, MA, USA, gprieto@mit.edu
Intermediate-depth earthquakes occur at depths where temperatures and pressures exceed those at which brittle failure is expected. There are two leading candidates for the physical mechanism behind these earthquakes: dehydration embrittlement and self-localizing thermal shear runaway. We use seismic data recorded on a regional network around the Bucaramanga Nest, the highest concentration of intermediate-depth earthquakes in the world, to better constrain which mechanism is relevant during rupture. The combination of high stress drop and low radiation efficiency that we observe for Mw 4–5 earthquakes implies a temperature increase of 600–1000°C for a centimeter-scale layer during earthquake failure. This suggests that substantial shear heating, and possibly partial melting, occurs during intermediate-depth earthquake failure. Temporal and spatial clustering of these earthquakes shows aftershock sequences of repeating earthquakes, located within meters of each other and with similar focal mechanisms, in some cases within a few seconds of each other. This suggests that the same patch is rupturing over a short time interval, suggesting that the slow dissipation of the heat generated during previous rupture, allows for repeating slip along the same fault plane. Our observations support thermal shear runaway as the mechanism for intermediate-depth earthquakes, which would help explain differences in their behavior compared to shallow earthquakes.
Session:New Directions in PSHA: Ins, Outs, and Uncertainty
Presenter   Papoulia, Joanna
Schedule   Fri / Poster
Room   Cook/Arteaga
Seismic and Tsunami Hazard Assessment in the Southwestern Hellenic Arc Based on Multidisciplinary and Multiscale Geophysical Data - The EU SEAHELLARC Project
PAPOULIA, J. E., Hellenic Centre for Marine Research, Athens, Greece, nana@ath.hcmr.gr; MAKRIS, J. N., GeoPro GmbH, Hamburg, Germany, info@geopro.com; MASCLE, J., Geoazur, Villefrance sur Mer, France, mascle@geoazur.obs-vlfr.fr; SLEJKO, D., Osservatorio Geofisico Sperimentale, Trieste, Italy, dslejko@ogs.trieste.it; YALCINER, A., Technical University, Ankara, Turkey, yalciner@metu.edu.tr; PAPADOPOULOS, G., National Observatory of Athens, Athens, Greece, papadop@gein.noa.gr; NICOLICH, R., University of Trieste, Trieste, Italy, r.nicolich@univ.trieste.it; GULKAN, P., Cankaya University, Turkey, polatgulkan@cankaya.edu.tr; and the SEAHELLARC Working Group
Aim of the SEAHELLARC project was to develop a new approach in assessing seismic and tsunami hazard by exploiting new hardware and combining geological, geophysical and engineering research. The coastal zone of western Peloponnese, south western Hellenic arc, was used as a pilot area, as this margin is one of the most seismic and tectonic active regions of Europe. The most important achievement of the SEAHELLARC project was the integration of multidisciplinary and multiscale geophysical data (from seabed to deep crust and upper mantle) to better understand the structure, tectonic and geological hazards on an active margin. Seismic hazard was computed according to the most updated approach (the logic tree approach, that incorporates all uncertainties involved in the computation) and considering a new seismogenic zonation, established during the SEAHELLARC project on the basis of the results of passive and active experiments. Moreover, the seismic sources that have generated or could generate tsunami were identified and characterized from the geographical and seismic point of view.
Session:Site Response: From Site-Specific Analyses to Predictive Models Around the Globe
Presenter   Molina Villegas, Juan Camilo
Schedule   Wed / Poster
Room   Cook/Arteaga
Computation of H/V Spectral Ratios at Sites with Strong Lateral Heterogeneity Using Diffuse Field Theory and IBEM
MOLINA VILLEGAS, J. C., Instituto de Ingeniería, Universidad Nacional Autónoma de México, Coyoacan, D.F, México, jcmovi@gmail.com; PERTON, M., Instituto de Ingeniería, Universidad Nacional Autónoma de México, Coyoacan, D.F, México, mathieu.perton@gmail.com; SÁNCHEZ SESMA, F. J., Instituto de Ingeniería, Universidad Nacional Autónoma de México, Coyoacan, D.F, México, sesma@unam.mx
It is well known that horizontal-to-vertical H/V spectral ratios of microtremors are useful to identify the dominant shear frequency. In some cases H/V is considered to be either the S wave amplification or the Rayleigh wave ellipticity although there are little theoretical basis to support these assertions. It has been recently proposed a theory for microtremor H/V spectral ratios based on the diffuse field assumption (Sánchez-Sesma et al., 2011). In this theory H/V corresponds to the square root of the ratio of the sum of horizontal energy densities with respect to the vertical one. The directional energy densities are proportional to the imaginary parts of the corresponding components of Green’s tensor when both source and receiver are the same point. For horizontally layered medium, we can easily calculate the theoretical Green function. Thus, by observing microtremors we can assess the underground structure below by using the theoretical point source solution. On the other hand, for a laterally heterogeneous structure, the horizontal responses are different. In that case, to simulate microtremor H/V spectral ratios, a numerical approach is needed. In this communication the 3D indirect boundary element method (IBEM) is used to study laterally heterogeneous elastic layers over a half-space. In order to overcome the singularity within the IBEM the reference solutions correspond to the classical half-space problems by Lamb (1904) and Chao (1960), for normal and tangential loads, respectively. Alternatively, for irregularities and inclusions over an elastic half-space we can gather a set of incoming P, SV, SH and Rayleigh plane waves that insures the illumination that corresponds to a diffuse field. With these plane waves as excitation to 3D configurations and using IBEM to compute the response, the autocorrelations can then be easily obtained. A large number of cases are analyzed for which we compute the H/V in 3D settings.
Session:Induced Seismicity
Presenter   Leidig, Mark
Schedule   Thu 8:45 AM / Oral
Room   Room 1
The Curious Case of the Timpson, TX Earthquakes: Natural or Delayed Triggering?
LEIDIG, M., Weston Geophysical Corp., Houston, TX, USA, mleidig@westongeo.com; BONNER, J. L., Weston Geophysical Corp., Lufkin, TX, USA
In May 2012, residents of the small East Texas town of Timpson began feeling earthquakes. Homes and businesses have been damaged by several events with the largest being an M=4.8. Although a nearby E-W trending fault system is mapped, this region of the U.S. is generally considered aseismic. Weston Geophysical Corp., the USGS, and Stephen F. Austin State University installed local seismic networks to monitor and better define the earthquake activity. Approximately 100 earthquakes had been detected through 2013. Additional smaller earthquakes likely occurred prior to the installation of the seismic networks. We have located the events in two main clusters south of the town of Timpson. The event epicenters indicate a linear trend that likely images a NW-SE trending fault. As a by-product of the hydraulic fracturing and natural gas production that has increased significantly near Timpson since 2006, large volumes of brine water are produced and require disposal. Two wastewater disposal wells began operation near Timpson, TX in 2006 and a third in 2009. Total injection volumes have decreased steadily since a peak in early 2008. The epicenters for the Timpson events lie between the two primary wastewater injection wells. In verified cases of triggered seismicity from wastewater injection, the earthquakes often start within weeks or months of high-volume injection onset. If the Timpson earthquakes are being triggered by wastewater injection from nearby wells, an important question is why events were not detected or felt during the peak years of wastewater injection.
Session:Near-Field Seismoacoustics of Natural and Man Made Explosions
Presenter   Bowman, Daniel
Schedule   Fri / Poster
Room   Cook/Arteaga
STUDENT
The Acoustic Signatures of Ground Acceleration, Gas Expansion, and Spall Fallback in Experimental Volcanic Explosions
BOWMAN, D. C., University of North Carolina at Chapel Hill, Chapel Hill, NC, USA, daniel.bowman@unc.edu; TADDEUCCI, J., Istituto Nazionale di Geofisica e Vulcanologia, Rome, Italy, jacopo.taddeucci@ingv.it; KIM, K., University of North Carolina at Chapel Hill, Chapel Hill, NC, USA, keehoon.kim@unc.edu; LEES, J. M., University of North Carolina at Chapel Hill, Chapel Hill, NC, USA, jonathan.lees@unc.edu; GRAETTINGER, A. H., Center for Geohazards Studies, University at Buffalo, Buffalo, NY, USA, ahgraett@buffalo.edu; SONDER, I., Center for Geohazards Studies, University at Buffalo, Buffalo, NY, USA, ingomark@buffalo.edu; VALENTINE, G. A., Center for Geohazards Studies, University at Buffalo, Buffalo, NY, USA, gav4@buffalo.edu
Results from an infrasound array deployed during an experimental volcanology initiative shed new light on the sources of acoustic emissions during a buried explosion. Acoustograms of each explosion contain information on the magnitude of initial ground deformation, the contribution of gas breakout to the total amplitude of the signal, and timing of the fallback of displaced material. The presence and relative amplitudes of these three phases varied based on the size of the explosive charge, the burial depth, and whether or not the substrate had been disturbed by previous blasts. The amplitude of the acoustic pulse from ground deformation closely matched the observed ground acceleration of the blast region determined using particle image velocimetry. The arrival of the gas release phase corresponded with the timing of gas breakout at the ground surface as observed using high speed video. The ground fallback signal obeyed the relationship between time delay and explosion size as outlined in Lee and Walker (1980). Similar amplitudes and waveform characteristics were observed on the ground and on an airborne array held aloft by helium balloons, indicating that the radiation pattern was isotropic. We describe a conceptual model for the acoustic signals generated by buried explosions based on the results from this experiment and literature on subterranean nuclear explosions. Finally, we place recent observations of ground deformation induced acoustic signals at Santiaguito volcano, Guatemala and Sakura-jima volcano, Japan, in the context of this model.
Session:Development of 2014 U.S. National Seismic Hazard Maps and Their Implementation in Engineering Applications
Presenter   Borcherdt, Roger
Schedule   Thu 4:30 PM / Oral
Room   Room 7/8
On NGA and NEHRP Site Coefficients Inferred for Site Classes in U.S. Building Codes
BORCHERDT, R. D., US Geological Survey, Menlo Park, CA, USA, borcherdt@usgs.gov
Site coefficients, derived herein from Next Generation Ground Motion Prediction Equations (NGA GMPEs), agree well with NEHRP site coefficients at low levels of input motion (0.1g) and those observed from the Loma Prieta earthquake. For higher levels of input motion, the majority of the NEHRP values are within the 95 % epistemic-uncertainty limits implied by the NGA estimates with the exceptions being the mid-period site coefficient, Fv for site class D and the short-period coefficient, Fa for site class C, both of which are slightly less than the corresponding 95% limit. Proposals by the author to adjust the NEHRP coefficients to the mean of the NGA site coefficients inferred with respect to Vs30 midpoint for reference site class B yield site coefficients consistent with the definitions of the simplified site classes adopted in US building codes. The proposed mean NGA coefficients 1) yield unity amplification independent of ground motion level for the reference site class B, 2) are consistent with those used for the adopted Maximum Considered Earthquake Response maps (MCER), 3) account for the epistemic uncertainty implied by the four NGA GMPE models, and 4) are consistent with average surface spectral amplifications observed in past earthquakes with average Vs30 ~ 800 m/s and 5) can be easily updated with future improvements in NGA GMPEs. The NGA data base implies that the median Vs30 value for site class B (Vs30 ~ 900 m/s) is more typical than 760 m/s as a value to characterize firm to hard rock sites as a possible less conservative reference for future MCER maps.
Session:Monitoring Dynamic Changes at Active Volcanoes and Fault Zones
Presenter   Lai, Voon Hui
Schedule   Thu / Poster
Room   Cook/Arteaga
Ambient Noise Monitoring of Temporal Changes in Seismic Velocity at the Geysers Geothermal Field, California
LAI, V. H., Berkeley Seismological Laboratory, Berkeley, CA, voonhui.lai@berkeley.edu; TAIRA, T., Berkeley Seismological Laboratory, Berkeley, CA, taira@seismo.berkeley.edu; DREGER, D. S., Berkeley Seismological Laboratory, Berkeley, CA, dreger@seismo.berkeley.edu
To monitor the temporal stress change induced by tectonic events or fluid injections at the Geysers Geothermal Field, CA, we search for the small perturbations of the velocity structure by exploiting the correlations of ambient seismic noise. Vertical component of the 15-month-long continuous seismic data are obtained from an array of 30 seismic stations maintained by Lawrence Berkeley National Laboratory. The procedure of the study includes computing hourly cross-correlations of the seismic noise of each station pair and stacking these hourly data into 1-day and subsequently 30-day stacks to obtain a reference Green’s function (RGF) with high signal to noise ratio. We identify the relative seismic velocity change (dv/v) within the earth structure by measuring the relative travel time shift between the RGF and the 30-day stacked correlations. Our result identifies two types of changes in dv/v; (1) a significant decrease in dv/v of 0.3% due to three M3.0+ earthquakes occurred on the same day and (2) a long periodic change in dv/v due to fluid injection or possibly seasonality in ocean microseisms. To further explore the change associated with the local earthquakes, we reduce the stacking time window from 30 days to 1 day to increase temporal resolution. To determine the source of the long periodic change in dv/v, we correlate both the wave height amplitude during our study period, which is proportional to strength of ocean microseisms, and the fluid injection data to the measured dv/v. Additionally, using single value decomposition method, we estimate the spatiotemporal variations in dv/v near stations. An overall improvement in our processing method and inverse model increase our monitoring ability to differentiate the sources that induce change in stress at the Geysers.
Session:Cold Climate Installation Techniques and Instrumentation Developments for Temporary and Long-Term Networks
Presenter   Bernsen, Steven
Schedule   Wed / Poster
Room   Cook/Arteaga
STUDENT
Development of New Seismological Instrumentation for Ice Covered Environments
BERNSEN, S., New Mexico Institute of Mining and Technology, Socorro, NM, USA, stevenbernsen@gmail.com; WINBERRY, P., Central Washington University, Ellensburg, WA, USA, winberry@geology.cwu.edu; ASTER, R., Colorado State University, Fort Collins, CO, USA, rick.aster@colostate.edu; WOODWARD, R., IRIS Headquarters, Washington, DC, USA, woodward@iris.edu; CARPENTER, P., IRIS PASSCAL Instrument Center, Socorro, NM, USA, pcarpenter@passcal.nmt.edu; BEAUDOIN, B., IRIS PASSCAL Instrument Center, Socorro, NM, USA, bruce@passcal.nmt.edu; GRIDLEY, J., IRIS PASSCAL Instrument Center, Socorrow, NM, USA, james.gridley@iris.edu
Ice Covered regions comprise >10% of Earth’s continental area; and include regions with poorly understood ice dynamics, tectonic histories, and ongoing geological processes. Scientific investigations of these regions are challenged by extreme weather, limited and expensive logistics, and the physical conditions of the ice environment. We report on the ongoing development of a new NSF MRI-supported community seismic capability for studying ice-covered regions– the Geophysical Earth Observatory for Ice Covered Environments (GEOICE). This project is fundamentally motivated by the need to optimize the collection of high-quality data relevant to key solid Earth and cryosphere science questions. The instrument capability will included: 1) a hybrid seismograph pool of broadband and intermediate elements, for observation of both long-period (e.g., long-period surface waves and slow sources),, and intermediate-to-short-period (e.g., teleseismic body waves local seismicity, impulsive or extended glaciogenic signals). The GEOICE instrument, and its power and other ancillary systems, will be specifically designed to both withstand conditions associated with icy environments, including cold/wet conditions and high-latitude solar limitations, and to incur minimal installation time and logistical load (i.e., size and weight), while maximizing ease-of-use in the field, in data handling, and in telemetry compatibility. Key features will include an “all in one”, post-hole-compatible design that integrates the seismometer and data logger into a single environmentally and mechanically robust housing, very low power requirements (~0.2 W) for the intermediate-band seismographs, and advanced power/battery systems that optimize battery capacity and operational limits. The envisioned ~125 element GEOICE instruments will be maintained and supported at the PASSCAL Instrument Center to ensure full and flexible peer-reviewed community use.
Session:Great Earthquakes and Slip to the Trench (Seismological Society of Japan/Seismological Society of America Joint Session)
Presenter   Shirzaei, Manoochehr
Schedule   Thu 4:15 PM / Oral
Room   Room 2
Seismic Versus Aseismic Slip: Probing Mechanical Properties of the Northeast Japan Subduction Zone
SHIRZAEI, M., Arizona State University, Tempe, AZ, USA, shirzaei@asu.edu; BÜRGMANN, R., University of California, Berkeley, Berkeley, CA, USA, burgmann@seismo.berkeley.edu; UCHIDA, N., Tohoku University, Sendai, Japan, uchida@aob.gp.tohoku.ac.jp; HU, Y., University of California, Berkeley, Berkeley, CA, USA, yhu@seismo.berkeley.edu; POLLITZ, F., U.S. Geological Survey, Menlo Park, CA, USA, fpollitz@usgs.gov
Fault slip may involve slow aseismic creep and fast seismic rupture, radiating seismic waves manifested as earthquake. These two complementary behaviors accommodate the long-term plate convergence rate of major subduction zones and are attributed to the fault frictional properties. It is conventionally assumed that zones capable of seismic rupture on the subduction megathrust are confined to within about 10 to 50km depth; however, the actual spatiotemporal extent of fault mechanical parameters remains elusive for most subduction zones. The 2011 Tohoku Mw9.0 earthquake, challenging this conventional assumption, provides a unique opportunity to probe the mechanical properties of the Japan subduction zone. Drawing on the inferred distribution of coseismic and postseismic slip, it has recently been suggested that portions of the megathrust are capable of switching between seismic and aseismic behavior. Investigating the coseismic and 15-month postseismic deformation of this event, we find that the coseismic rupture triggered widespread frictional afterslip with equivalent moment magnitude of 8.41, in addition to viscoelastic relaxation in the underlying mantle. The identified linear relation between afterslip and the cumulative number of aftershocks at 35-55 km depth suggests that most aftershocks are a direct result of afterslip. Using a rate-state friction model we constrain the heterogeneous fault mechanical parameters of the subduction thrust. Our results indicate a variable pattern along dip and strike and that velocity-weakening properties persist in the shallow part of the megathrust. This finding indicates that earthquakes with the size of the Tohoku earthquake will likely repeat in the future, highlighting the importance of characterizing the spatiotemporal extent of the fault mechanical properties elsewhere.
Session:Explosive Source Characterization
Presenter   Zhao, Lian-Feng
Schedule   Wed 4:30 PM / Oral
Room   Room 1
Relocation, Discrimination, Magnitude Calculation and Yield Estimation of the 12 February 2013 North Korean Nuclear Test
ZHAO, L. F., KLEDI, IGG, Chinese Academy of Sciences, Beijing, China, zhaolf@mail.igcas.ac.cn; XIE, X. B., IGPP, University of California at Santa Cruz, Santa Cruz, CA, USA, xxie@ucsc.edu; FAN, N., KLEDI, IGG, Chinese Academy of Sciences, Beijing, China, fanna@mail.iggcas.ac.cn; WANG, W. M., KLCCPUI, ITPR, Chinese Academy of Sciences, Beijing, China, wangwm@itpcas.ac.cn; YAO, Z. X., KLEDI, IGG, Chinese Academy of Sciences, Beijing, China, yaozx@mail.iggcas.ac.cn
On February 12, 2013, North Korea conducted its third and the largest nuclear test to date in the China-North Korea border area. Due to its large magnitude, the seismic records from this event have better signal-to-noise ratios than those from two previous nuclear explosions (hereafter, these explosions are referred as NKT1, 2 and 3). We use regional waveforms recorded on China National Digital Seismic Network (CNDSN), Global Seismic Network (GSN), and Japan F-NET to investigate the NKT3 for its location, event discrimination, and yield estimation. Using 100 Pn arrival data and the relative location method, we obtain the location for NKT3 to be [41.2923±0.0004°N, 129.0727±0.0006°E], approximately 3.0 and 0.4 km away from the NKT1 and NKT2, respectively. The relocation precision is approximately 52 m. For the North Korean nuclear explosions and a group of nearby earthquakes, we calculate Fourier spectral ratios Pg/Lg, Pn/Lg, and Pn/Sn from vertical-displacement waveforms on stations with pure continental paths. After distance corrections and eliminating the data with signal-to-noise ratios below 1.8, these ratios can fully discriminate the explosion and earthquake populations at frequencies above 2.0 Hz, indicating an explosion within this magnitude range detonated in China-North Korea border area can be determined by a regional network without ambiguity. After calibrating the network with a Lg-wave dataset (685 broadband waveforms from 98 regional events between December 1995 and February 2013) and a broadband Lg-wave attenuation model, we obtain the Lg-wave magnitude mb(Lg)=4.91±0.22 for NKT3. By further using a fully coupled hard-rock site model, the yield for NKT3 is estimated to be 7.47 kt based on it is detonated at the scaled depth. However, if it is strongly over buried, this yield may be underestimated.
Session:Advances in Seismic Imaging and Monitoring of Time-Dependent Variations: Civil Structures, Near-Surface, and Shallow Crustal Scales
Presenter   Agrawal, Mohit
Schedule   Wed / Poster
Room   Cook/Arteaga
STUDENT
Lithospheric Earth Structure of Texas-Gulf of Mexico Margin from Surface Wave Dispersion and Migrated Ps Receiver Functions
AGRAWAL, M., Baylor University, Waco, TX, USA, Mohit_Agrawal@baylor.edu; PULLIAM, J., Baylor University, Waco, TX, USA, Jay_Pulliam@baylor.edu; SEN, M. K., University of Texas at Austin, Austin, TX, USA, msentx@yahoo.com
The seismic velocity structure beneath Texas Gulf Coastal Plain (GCP) is imaged by fitting of surface wave dispersion and velocity analysis of Ps receiver functions. The GCP is a portion of an ocean-continental transition zone, or "passive margin", where attempts to model lithospheric earth structure seismically have been rare. Seismic data from a linear array of 22 broadband stations, spaced 16-20 km apart along a ~330-km-long profile across the Texas-Gulf of Mexico passive margin, were employed to construct a coherent image of the crust and uppermost mantle. A Common Conversion Point (CCP) stacking technique is applied to teleseismic earthquake data to improve the signal-to-noise ratios of receiver functions. Using an incorrect velocity model for time-to-depth migration of a stacked CCP image may produce an inaccurate image of the subsurface, especially in the presence of substantial lateral velocity variations. To find accurate P- and S-velocity models, we first apply a nonlinear modeling technique to fit Rayleigh wave group velocity dispersion curves via Very Fast Simulated Annealing (VFSA), a global optimization method. We use ambient noise cross-correlation to compute Rayleigh wave group velocity dispersion curves. Statistical assessment tools, such as the parameter correlation matrix and the marginal Posterior Probability Density function, are estimated in order to evaluate the constraints that dispersion data place on model parameters.
Session:Deciphering the Earthquake and Tsunami History of Subduction Zones
Presenter   Cisternas, Marco
Schedule   Wed 2:15 PM / Oral
Room   Room 2
The 1730 and 1751 Chilean Earthquakes: Two Major Contributors to the Seismically Prolific 18th Century along the Pacific Coast of the Americas
CISTERNAS, M., Pontificia Universidad Catolica de Valparaiso, Valparaiso, Chile, marco.cisternas@ucv.cl; ELY, L., Central Washington University, Ellensburg, WA, USA, ely@geology.cwu.edu; WESSON, R., USGS, Boulder, CO, USA; PILARCKZIC, J., Rutgers University, New Brunswick, NJ, USA, jpilar@marine.rutgers.edu; GORIGOITIA, N., Pontificia Universidad Catolica de Valparaiso, Valparaiso, Chile, nga2043@gmail.com; DURA, T., University of Pennsylvania, Philadelphia, PA, USA, dura@sas.upenn.edu; MELNICK, D., University of Potsdam, Potsdam-Golm, Germany, Daniel.Melnick@geo.uni-potsdam.de; CARVAJAL, M., Pontificia Universidad Catolica de Valparaiso, Valparaiso, Chile, matias.carvajal.ramirez@gmail.com
Large underthrusting earthquakes along the Pacific coast of the Americas seem to have clustered during the first half of the 18th century. Fault ruptures that reportedly generated large tsunamis occurred in the northwestern United States in 1700, in south-central Mexico in 1732 and 1754, in central Peru in 1746 and in central and south-central Chile in 1730 and 1751. The latter two events damaged towns along 800 km and initiated destructive tsunamis that crossed the Pacific to Japan. Although these two Chilean earthquakes were only 21 years apart, there was significant overlap in the areas that sustained damages from each event. We studied the 1730 and 1751 earthquakes and tsunamis through their geological and historical records. We searched for primary historical sources in the archives of Spain, Peru, Argentina and Chile. We also scrutinized first-hand Japanese reports of the effects of both tsunamis on the eastern coast of Honshu. At two Chilean coastal sites, Campiche (32.7° S) and Tirúa (38.3° S), we described and dated sand sheets that were probably deposited by these tsunamis. Finally, we compared the effects of these historical events with those of two other modern and better-known ones: the Mw 9.5 1960 and the Mw 8.8 2010 earthquakes. Our results indicate that the effects of both the 1730 and 1751 events were greater than those of 2010 and less than those of 1960. If they were in fact larger than Mw 8.8, it would imply that the release of accumulated slip and seismic moment during the first half of the 1700s was exceptionally large along the Chilean coast, adding to the potential significance of this seismic period along the entire coast of the Americas.
Session:75 Years of Frequency-Size-Distribution of Earthquakes: Observations, Models and Understanding
Presenter   Nava, F. Alejandro
Schedule   Fri / Poster
Room   Cook/Arteaga
A Lower Bound for Gutenberg-Richter’s b-Value and Speculations on Fractal Dimensions
NAVA, F. A., CICESE, Seismology Department, Ensenada, Baja California, Mexico, fnava@cicese.mx
The Gutenberg-Richter distribution corresponds to an exponential probability density function for magnitudes, which for moment magnitudes MW is valid only for b>2/3. Smaller b values are thus indicative of either sampling errors or of the use of magnitudes which do not scale as MW. In 1981, K. Aki speculated that the fractal dimension of the areas of the seismic ruptures should be related to b as D=2b; the supposition that seismic moments are fractally distributed implies that their fractal dimension should be D=1.5b (so that for seismic moments D>1). This result validates Aki’s speculation for the case where the average slip on a fault scales as the cubic root of the rupture area.
Session:Topics in Seismology: Regional Seismicity and Tectonics
Presenter   Reeder, John
Schedule   Thu / Poster
Room   Cook/Arteaga
The Significance of the Yakutat Plate to the Alaska Orocline
REEDER, J. W., State of Alaska, Div. Geological & Geophysical Surveys, Retired, Fairbanks, AK, USA, jreeder@usgs.gov
The arcuate structures of major Alaskan faults such as the Denali are the bases of Carey’s (1955) Alaska Orocline, which has attracted numerous models such as rotation of blocks about a point (Coe, 1989; Plafker & Lahr, 1980; St. Amand, 1957), the nonrigid world of escape tectonics (Redfield et al., 2007) & finite element diffusion (Finzel et al., 2011), and the multiple bending of rigid layers like Grantz’s (1966) megakinks or multiple thrust bending like Glen’s (2004) conveyor belts. But, for all, the expected extensional and compressional structures on appropriate arc sides (Schultz & Aydin, 1990) and rotated and/or arched structures (Csejtey, 1990) are mainly lacking. None explain the existence of the Copper River, Nenana and Tanana basins. Another serious problem is the large offset disparities between E and W of the curved fault systems (Lanphere, 1978; Miller et al., 2002). Why all of these problems? It has been the lack of proper accounting for the subduction of the Yakutat Plate into the Katmai, Cook Inlet, and Mount McKinley parts of the Aleutian subduction zone. The Freymueller et al. (2008) “Southern Alaska Rotating Block” should be replaced with the Yakutat Plate. This Plate would have a boundary with the Bering Block at the Aleutian subduction zone. The Yakutat Plate is pulling away from the Wrangell volcanic arc and Copper River basin as a spreading ridge into the Cook Inlet part of the Aleutian subduction zone, with the Pacific Plate still causing a Wrangell subduction zone (Stephens et al., 1984). The Totschunda and Fairweather faults are the new westward developing Denali transform. The Nenana and Tanana basins formed earlier in a similar fashion to how the Copper River basin is now forming. The Yakutat Plate is forming at the expense of the Pacific Plate and it is transporting mass WSW along its own transforms into the Aleutian subduction; creating a more “steady state” Alaska Orocline!
Session:New Insights into Ground Failure as an Urban Earthquake Hazard
Presenter   Reeder, John
Schedule   Thu / Poster
Room   Cook/Arteaga
The Age of Landslides in Anchorage, Alaska
REEDER, J. W., State of Alaska, Div. Geological & Geophysical Surveys, Retired, Fairbanks, AK, USA, jreeder@usgs.gov
Numerous large Holocene landslides exist in Anchorage, Alaska, which are suspected to have been caused by megathrust earthquakes (Hansen, 1965; Schmoll & Dobrovolny, 1974; Reeder, 1974). Mean headscarp slopes of these pre-1964 landslides cluster into five assumed “degradation with time” groups. Based on subsidence events of the nearby Knik Arm Duck Flats, two types of megathrust earthquakes occur: the Pacific (P) Plate like the 1964 earthquake and a more continental Yakutat (Y) Plate (Reeder 2012 & 2013). Assuming these two types alternate, the following predicted Anchorage megathrust events result: 50P, 570Y, 950P, 1470Y, 1850P ybp etc. These ages agree roughly with actual paleoseismic records for Cook Inlet (Bartsch-Winkler & Schmoll, 1992; Combellick, 1991 & 1994) and Copper River Delta (Carver & Plafker, 2008), except the Delta is missing 570Y and 1850Y. In addition, the average inlet bluff erosion rate for Anchorage is 12 cm/yr (Schmoll & Dobrovolny, 1974 pers. com.). Given a width of up to 200 m for major landslides, it would require about 1700 yrs to remove the headscarp of a large landslide along inlet bluffs. Only the steeper three groups occur in inlet bluffs. Correlating megathrust histories with these slopes result with 36.6° for 570 ybp, 33.6° for 950 ybp and 31.6° for 1470 ybp, which is expected with the known erosion rate. Curve fitting yields 2370 ybp for 29.7° and a surprising 4550 ybp for 26.8°. The extensive headscarp for the Inlet View School at 12th Ave and N St is in this 4550 ybp category, which also correlates with the largest paleoseismic megathrust for the region (Plafker & Lienkaemper, 2012). Of note, this degradation curve does not follow the diffusion model as used for fault scarp degradations (Colman & Watson, 1983). Diffusion models require continuous processes. This is not the case for Anchorage where slopes appear to be a function of “soil” conditions during and just after earthquakes.
Session:Great Earthquakes and Slip to the Trench (Seismological Society of Japan/Seismological Society of America Joint Session)
Presenter   Reeder, John
Schedule   Thu / Poster
Room   Cook/Arteaga
The Occurrence of Two Types of Megathrust Earthquakes in South Central Alaska
REEDER, J. W., State of Alaska, Div. Geological & Geophysical Surveys, Retired, Fairbanks, AK, USA, jreeder@usgs.gov
Up to now, it has not been clarified whether the 1964 megathrust earthquake for the south central Alaska part of the thrust was between YAK (Yakutat Plate) & NA (North American Plate) and/or PAC (Pacific Plate) & YAK (Brocher et al., 1994; Freymueller et al., 2008). Berg (1965) recognized the P wave of the 1964 earthquake was divided into two first motions with the second striking N 16° W with a NE dip of 26°, which is the Wrangell subduction part of the earthquake. In 1964, the Patton Bay and Hanning Bay faults in the SW Prince William Sound had reverse uplift (Plafker, 1967). Based on 1964 horizontal displacements (Parkin, 1966), the YAP & NA moved S from the Aleutian subduction zone against the YAP & NA that moved W from the Wrangell subduction zone, causing these reverse uplifts. There is also a clear ENE displacement indicated by this data between YAK & NA for the Prince William Sound. But such movements are lacking to the W where the YAP appears stuck with the NA (Arke et al. 2013). The YAP needs to move NNW, which will happen when a more continental megathrust occurs, which would be principally between YAP & NA. The crust is now folding (Freymueller et al., 2008) as if the NA is indeed separating from the YAK at a wave length of about 200 kms. The crustal fold of the 1964 megathrust indicates the YAK was actually attached to the NA because its maximum fold wavelength was 475 km (Biot, 1961). Therefore, two types of megathrust earthquakes occur in south central Alaska: the 1964 Pacific type, principally between PAC & YAK; and the more continental type, principally between YAK & NA. Based on Knik Arm subsidence events (Reeder, 2012), these two types alternate at about 520 yrs for the Pacific and 380 yrs for the continental. The next megathrust earthquake would be expected in about 330 yrs. Such a continental megathrust would result in reverse uplift on the N side of the Castle Mountain fault.
Session:Earthquake Physics and Interaction
Presenter   Walter, Bill
Schedule   Fri 4:30 PM / Oral
Room   Room 7/8
A Global Comparison of Crustal Earthquake Scaling from Stable Event Ratio Levels
YOO, S. H., Weston Geophysical Corp., Lexington, MA, USA, hoonthhoonth@gmail.com; WALTER, W. R., LLNL, Livermore, CA, USA, bwalter@llnl.gov; MAYEDA, K., Weston Geophysical Corp., Lexington, MA, USA, kevin.mayeda@gmail.com
A challenge with using corner frequency to interpret stress parameter scaling is that stress drop and apparent stress are related to the cube of the corner frequency. In practice this leads to high levels of uncertainty in measured stress since the uncertainty in measuring the corner frequency is cubed to determine uncertainty in the stress parameters. We develop a new approach using the low and high frequency levels of spectral ratios taken between two closely located events recorded at the same stations. This approach has a number of advantages over more traditional corner frequency fitting, either in spectral ratios or individual spectra. First, if the bandwidth of the spectral ratio is sufficient, the levels can be measured at many individual frequency points and averaged, reducing the measurement error. Second the apparent stress (and stress drop) is related to the high frequency level to the 3/2 power so the measurement uncertainty is not as amplified as when using the corner frequency. Finally, if the bandwidth is sufficiently broad to determine both the spectral ratio low and high frequency levels, the apparent stress (or stress drop) ratio can be determined without the need to use any other measurements (e.g., moment, fault area), which of course have their own measurement uncertainties. For this study, we processed a wide variety of moderate-to-large crustal earthquake sequences using stable, high-resolution coda measurements and find overwhelming evidence that self-similar scaling is not supported by the data, regardless of the source model.
Session:From the Earthquake Source to Damage of Buildings: Bridging the Gap between Seismology and Earthquake Engineering
Presenter   Wyss, Max
Schedule   Fri 3:45 PM / Oral
Room   Room 1
Fatalities in Greek Earthquakes, Past and Future
WYSS, M., WAPMERR, Bougy-Villars, VD, Switzerland, wyss.adh@gmail.com; TOLIS, S., WAPMERR, Bougy-Villars, VD, Switzerland
There exist reports of fatalities (1 to several thousand) in approximately 100 Greek earthquakes since 464BC. We selected 17 of these for which a magnitude has been estimated, more than 100 fatalities were reported and there was no tsunami. Tsunamogenic events are excluded because we estimate the numbers of fatalities due to damaged buildings, but are unable to estimate flooding effects. Since 1881, there has not been a report of more than 500 fatalities due to an earthquake, thus it appears that Greece has been spared earthquake disasters during the last 140 years, whereas before some large human losses have been reported. We calculated how many fatalities would have to be expected for repeats of the 17 test earthquakes with the tool QLARM and its current model for Greek building properties. Events of M≥7 were modeled as line sources, smaller ones as point sources. Hypocentral depths were assigned in the top 5 to 20 km, depending on the magnitude and information found in the literature. For one event the depth is given as 90 km. The factors that introduce errors in the comparison include: Inaccurate source parameters, unreliable historical counts of fatalities, unknown deviations from average attenuation properties and local soil conditions, simplification of the QLARM models for population and building stock and unknown time of day for some of the older events. The fact that the QLARM loss calculation comes out within 6 persons of the number reported for the 1999 earthquake near Athens, gives some confidence that our theoretical estimates are meaningful. It turns out that today the numbers of fatalities to be expected are half (median) of those reported historically. Two events deviated from this trend; their estimates for the present were larger than the historically reported numbers. We interpret the overall result as indicating that the improvement of the building stock has reduced the fatality count more than the population growth has increased it.
Session:Induced Seismicity
Presenter   Atkinson, Gail
Schedule   Thu 2:15 PM / Oral
Room   Room 1
Seismic Hazard and Ground Motion Implications of Induced Seismicity
ATKINSON, G. M., Western University, London, ON, Canada, gmatkinson@aol.com
Induced seismicity poses a significant and as-yet-unquantified risk to the integrity of critical structures, such as major dams, as new unconventional energy activities (including hydraulic fracture treatments and wastewater disposal) are currently being planned and/or conducted in close proximity (within 10 km). The addition of a new seismicity source, such as a disposal well, fundamentally alters the seismic hazard at nearby sites. The additional hazard from the induced seismicity source may overwhelm the natural hazard, particularly at the low-probability levels of interest to critical structures. The impact is greatest in regions of low-to-moderate seismicity, for which existing facilities were designed with the implicit understanding that expected ground motions are modest. The impact is further exacerbated by the shallow depth of the induced events, which may result in large ground-motion amplitudes being generated from moderate events at very close distances. This presentation overviews the ground-motion and hazard issues associated with the addition of an induced seismicity source in close proximity to a site, in a low-seismicity region. I show that IF an induced seismicity sequence is initiated, then its contribution to hazard will greatly exceed that from the pre-existing natural hazard. This arises due to the very close proximity of the source to the site, and the fact that the background hazard is low. Therefore, it is crucial to assess the likelihood and rate parameters of the potential induced seismicity sequence; this would allow us to multiply the hazard parameters for the induced-seismicity source by the appropriate conditional likelihood parameters, and obtain a reasonable estimate of the added hazard. Characterizing the hazard contributions of induced seismicity sequences, for operations that have not yet been initiated, is a complex problem that will provide a grand challenge in engineering seismology over the coming decade.
Session:Fault Structure, Heterogeneity, and Implications for Rupture Dynamics
Presenter   Douilly, Roby
Schedule   Wed 2:30 PM / Oral
Room   Room 3
STUDENT
3D Dynamic Rupture Simulation across a Complex Fault System: The Mw7.0, 2010, Haiti Earthquake
DOUILLY, R., Purdue University, West Lafayette, IN, USA, rdouilly@purdue.edu; AOCHI, H., Bureau de Recherche Geologiques et Minieres, Orlean, France, h.aochi@brgm.fr; CALAIS, E., Ecole Normale Superieure, Department of Geosciences, Paris, France, ecalais@geologie.ens.fr; FREED, A. M., Purdue University, West Lafayette, IN, USA, freed@purdue.edu
Earthquakes ruptures often take places on faults adjacent to – and sometimes very close to – major plate boundary faults without activating them. The 1989 Loma Prieta earthquake is a classic case where rupture occurred on a blind reverse fault (with both dip-slip and strike-slip) while the adjacent San Andreas Fault was not triggered. In this study, we investigate the rupture dynamics of the Haiti earthquake in a complex fault system with multiple segments as identified by the aftershock relocations. We calculate a background stress regime that is consistent with the type of motion on the Léogâne rupture and with the regional tectonic regime. We initiate nucleation on the eastern segment of the Léogâne fault, at the location of the relocated earthquake centroid, by defining a circular region of 2-km radius where shear stress is slightly greater than the yield stress. By varying friction on faults and background stress, we find a range of scenarios that are consistent with the observed rupture. In the absence of near-field seismic records of the event, we score the models against the static surface deformation field derived from GPS, InSAR, and coral uplifts. All the plausible simulations show that the rupture propagates from the eastern to the western segment of the Léogâne fault, but not on the Enriquillo fault nor on the Trois Baies fault. The best-fit simulation shows a significant increase of shear stresses on the Trois Baies fault, consistent with aftershocks observed on this fault. This best-fit simulation also shows a significant increase of shear stresses mostly on the top and on the western part of the Enriquillo fault indicating that it has been brought closer to rupture and still constitutes a threat for southern Haiti. We also find that a shift of the western segment of the Léogâne fault by about 2 km to the north compared to recent studies provides a better fit to the coseismic InSAR and GPS displacements.
Session:Advances in Seismic Imaging and Monitoring of Time-Dependent Variations: Civil Structures, Near-Surface, and Shallow Crustal Scales
Presenter   Pilz, Marco
Schedule   Wed 9:00 AM / Oral
Room   Room 3
Shallow 3D Velocity Imaging by a Joint Rayleigh- and Love-Wave Noise Tomography
PILZ, M., Helmholtz Center Potsdam - GFZ, Potsdam, Germany, pilz@gfz-potsdam.de; PAROLAI, S., Helmholtz Center Potsdam - GFZ, Potsdam, Germany; BINDI, D., Helmholtz Center Potsdam - GFZ, Potsdam, Germany
Imaging shallow subsurface structures and monitoring related temporal variations are two of the main tasks for modern seismology. Although many observations have reported temporal velocity changes e.g. in fault zones, in volcanic areas and on landslides, precise measurements of velocity changes based on passive sources like ambient seismic noise still suffer from high computation times. Recently, we presented a method (Pilz et al. 2012, 2013) to overcome this problem for obtaining information about the local subsoil structure on the engineering scale. Here we present a fast one-step inversion procedure which accounts for the topographic relief and which is based on the computation of high-frequency correlation functions for both Rayleigh and Love waves between stations of a small-scale array deployed for recording ambient seismic noise. Constitutively, a tomographic inversion of the travel times estimated for each frequency is performed in due consideration of the different depth-sensitivity kernels of both Rayleigh and Love waves. We test our technique by using numerical simulations of seismic noise for a realistic site and by using real-world recordings from a small-scale array performed at Solfatara (Italy). The results imply that the method can provide a clear image of the subsurface structures and that, in turn, passive seismic interferometry allows local structural heterogeneities to be satisfactorily reproduced. In particular, the use of Love-wave phase velocities allows to further constrain the inversion and to improve the resolution of the model. Since reliable velocity estimates for the frequency range investigated can be obtained in almost real time, the results further imply the use of the proposed procedure as an exploration and monitoring tool.
Session:Advances in Understanding Earthquake Hazard in Central and Eastern North America
Presenter   Napoli, Vanessa
Schedule   Fri 11:30 AM / Oral
Room   Room 2
STUDENT
Relative Locations of Earthquakes and Potential Driving Mechanisms of Seismicity along the Northeast U.S. Atlantic Passive Margin
NAPOLI, V. J., Boston College, Chestnut Hill, MA, USA, napoliv@bc.edu; EBEL, J. E., Weston Observatory, Boston College, Weston, MA, USA, ebel@bc.edu
Understanding the possibility of a large magnitude, tsunamigenic earthquake along the continental shelf break of a passive margin is important as events similar to the 1929 M7.2 Grand Banks earthquake can unexpectedly devastate coastline towns. The tsunamigenic Grand Banks event raises the question, what is the potential for such an event to occur along other parts of the northeast Atlantic margin? Moderate seismicity has been observed near the continental shelf break east of the Gulf of Maine (~1000km southwest of the Grand Banks event). From 2008-2013 there were 16 earthquakes ranging in magnitude from 2.0-3.9 along the Gulf of Maine continental shelf break, with 10 of the earthquakes occurring on 12 April 2012. This research focuses on identifying the spatial and temporal migration patterns of these earthquakes calculated using a double-difference relative location analysis. In this procedure two waveforms are cross-correlated to determine the difference of arrival times of the P-waves and of the S-waves. The program then calculates the location of one event relative to another based on these differences in arrival times. The data show that generally as the earthquakes occurred further northwest they also happened deeper in the crust. Plotting the relative longitude versus relative latitude versus relative depth shows that the earthquakes map out a planar surface. If the events occurred on a single fault, then the preliminary interpretation of that fault is the strike is in the range from N60°E to N66°E, the dip is in the range from 23°-27.5° and the faults total extent is ~2.4km by ~2.7km. If this entire area ruptured in one earthquake, the event would be about a magnitude 5.0. The seismic hazard in the Gulf of Maine region will be better constrained with continued research on focal mechanisms of the 2012 events and by comparing the results of the locations and magnitudes of the 2012 events to those of the 1929 tsunamigenic Grand Banks event.
Session:Seismic Imaging as USArray Moves to Alaska (IRIS/Seismological Society of America Joint Session)
Presenter   Porritt, Robert
Schedule   Thu / Poster
Room   Cook/Arteaga
The Wyoming Province and the Llano Province: A Tale of Two Proto-continents
PORRITT, R. W., Univ. Southern California, Los Angeles, CA, USA, rporritt@usc.edu; ALLEN, R. M., Univ. California Berkeley, Berkeley, CA, USA, rallen@berkeley.edu; POLLITZ, F. F., USGS Menlo Park, Menlo Park, CA, USA, fpollitz@usgs.gov; MILLER, M. S., Univ. Southern California, Los Angeles, CA, USA, msmiller@usc.edu
USArray has facilitated significant advancement in tomographic models and methodologies. While there persists a fundamental tradeoff between horizontal and vertical resolution due to the components of the wave train analyzed, advances in joint inversions are continuing to refine the tomographic images generated with USArray. The DNA13 model incorporates teleseismic P observations, independent SH and SV observations, and surface-wave phase velocities from both teleseismic earthquakes and ambient noise to constrain the relative wave-speed from the crust down into the lower mantle. We address the validity of our models through forward prediction of observables and compare the predictive power of the DNA13 models to other models. In the shallow portion of DNA13, we image the Archean age Wyoming Province, which exhibits evidence of ocean closure at its northern and southern ends. The Llano Province in Central Texas is of Grenville age and still contains lithospheric evidence of subduction as the province accreted to North America. Comparison of these two provinces highlights the role of fossil slabs as part of the cratonic architecture. Analysis of the deep portion of the models highlights variations within the Farallon plate, including two distinct high wave-speed anomalies in the eastern U.S. and a shallow feature in the center of these two anomalies. We propose this is evidence of an oceanic plateau, which provides the necessary positive buoyancy to promote flat-slab subduction of the Farallon plate.
Session:75 Years of Frequency-Size-Distribution of Earthquakes: Observations, Models and Understanding
Presenter   Goebel, Thomas
Schedule   Fri 10:45 AM / Oral
Room   Room 4
b-Values, Stress and Fault Heterogeneity during Laboratory Stick-Slip Experiments
GOEBEL, T. H. W., California Institute of Technology, Pasadena, CA, USA, tgoebel@gps.caltech.edu; BECKER, T. W., University of Southern California, Los Angeles, CA, USA, twb@usc.edu; SAMMIS, C. G., University of Southern California, Los Angeles, CA, USA, sammis@usc.edu; SCHORLEMMER, D., German Research Centre for Geosciences, Potsdam, Germany, ds@gfz-potsdam.de; DRESEN, G., German Research Centre for Geosciences, Potsdam, Germany, dre@gfz-potsdam.de
Differences in crustal conditions and faulting mechanisms may lead to regional variations in the frequency-magnitude distributions (FMDs) of crustal seismicity. To better asses controlling parameters of variations in FMDs and b-values, we examine fault structure, stress and micro-seismic events during sequences of stick-slip events, a laboratory analog to earthquakes. The experiments were conducted on faults that developed from fresh fracture surfaces in Westerly granite. Our experiments revealed many similarities between laboratory and crustal faults, i.e., a hierarchical damage structure, fractal fault roughness that controls the spatial distribution of seismic events, Omori-Utsu aftershock decays and Gutenberg-Richter type frequency-magnitude distributions. The latter can be described by a powerlaw with an exponent (b-value) that varied between 0.7 to 2.2. We analyzed variations in b-value along fault strike and as function of differential stress. b-values are generally low in regions of relatively thin fault zone width, i.e. little to no gouge content between host-rock walls. These regions also show high seismic moment release and seismic event densities during the inter-slip periods and mark the nucleation sites of stick-slip events. Furthermore, b-values decreased systematically with increasing differential stresses during several stick-slip cycles so that seismic events are more likely to grow to larger sizes if ambient stresses are high. In addition, variations in b-values are sensitive to the structural complexity of faults observed in post-experimental CT-scans so that less complex faults and faults with larger displacements exhibit a stronger correlation between differential stress and b-values. Our results highlight the role of both fault heterogeneity and differential stress in controlling frequency-magnitude distributions during the faulting of rock samples in the laboratory.
Session:75 Years of Frequency-Size-Distribution of Earthquakes: Observations, Models and Understanding
Presenter   Ruhl, Christine
Schedule   Fri / Poster
Room   Cook/Arteaga
STUDENT
Evolution of the 2008 Mogul Earthquake Swarm, Reno, Nevada: Identifying Complex Structures in a Shallow Urban Seismic Sequence
RUHL, C. J., Nevada Seismological Laboratory, University of Nevada, Reno, Reno, NV, USA, cruhl@unr.edu; SMITH, K., Nevada Seismological Laboratory, University of Nevada, Reno, Reno, NV, USA, ken@seismo.unr.edu; ABERCROMBIE, R. E., Boston University, Boston, MA, USA, rea@bu.edu
Increased sensitivity of seismic networks in northern Nevada highlights low-magnitude swarm behavior common in the Reno-Tahoe-Carson City area, with many sequences including felt events and with b-values <1. Over a five-month period beginning 28 February 2008, an unusually shallow sequence produced 38 ML≥3.0 earthquakes in the Mogul-Somersett neighborhood of Reno, Nevada. The 26 April 2008 Mw5.0 mainshock was preceded by 18 of 38 ML≥3.0 earthquakes in two distinct foreshock phases that occur on closely spaced parallel structures. The second foreshock phase is characterized by increased activity with four ML≥3.0 events on 15 April 2008 approximately 1 km NE of the first structure. The second phase also included two ML≥4.0 foreshocks on 24 April 2008 and seven additional ML≥3.0 foreshocks prior to the “mainshock.” We relocate 1612 earthquakes occurring over 180 days with magnitudes greater than zero and a minimum of 20 defining phases using the double-difference algorithm of Waldhauser and Ellsworth (2000). The initial relocations indicate a complex fault zone with at least one auxiliary plane. We subsequently bin events using variable time windows and relocated each set independently to test if locating the entire swarm might be masking the temporal and structural resolution of the sequence. This relocation reveals several additional parallel and conjugate structures that have not been previously isolated. We use an empirical Green’s function method to estimate source dimension and stress drop for the larger earthquakes to isolate stress variations within the active structures that may correlate with the evolution of the stress field inverted from moment tensor and short-period focal mechanisms. The complexity and shallow nature of the sequence suggests potential involvement of fluids and provides an image of numerous localized faults that may be an overall characteristic of urban Reno area seismicity.
Session:Geometric Complexities Along Strike-Slip Systems: New Insights on Seismic Hazards, Earthquake Behavior, and Fault System Evolution
Presenter   Goebel, Thomas
Schedule   Wed 8:45 AM / Oral
Room   Room 2
Stress Drop Heterogeneity within Complex Tectonic Regions: A Case Study of the San Gorgonio Pass, Southern California
GOEBEL, T. H. W., California Institute of Technology, Pasadena, CA, USA, tgoebel@gps.caltech.edu; HAUKSSON, E., California Institute of Technology, Pasadena, CA, USA, hauksson@gps.caltech.edu; SHEARER, P. M., Univerisity of California, San Diego, San Diego, CA, USA, pshearer@ucsd.edu; AMPUERO, J. P., California Institute of Technology, Pasadena, CA, USA, ampuero@gps.caltech.edu
Slip along tectonic faults during earthquakes generally leads to a reduction in average shear-stress within the source region. The amount of stress decrease may vary, e.g., as function of fault geometry, depth, crustal thickness and faulting type. Here, we test a range of crustal parameters to advance our understanding of underlying mechanisms of spatial stress-drop variations. We analyze source spectra of small and intermediate magnitude seismic events within the San Gorgonio pass region. This region marks the location of a restraining bend within the San Andreas fault system resulting in complex, distributed deformation along systems of strike-slip and thrust faults. Stress drops are computed by fitting a Brune-type spectral model to individual event source spectra. These spectra are determined by iteratively stacking the observed amplitude spectra and correcting high frequency contributions by using a regional empirical Green's function. The results are subjected to a sensitivity analysis of input parameters to verify their stability. Stress drops show strong spatial variations ranging from 0.1 to 100 MPa with a median of about 5 MPa. Stress drops are higher for thrust events (about 8 MPa) whereas normal faulting events exhibit lower values of about 3 MPa. We observe an abrupt increase of stress-drops with depth at about 10 km. In addition, stress drops of events along the San Andreas fault correlate with geologic slip rates so that relatively low slip rates are associated with high stress-drops and vice versa. The spatial variations in stress-drops are likely sensitive to both crustal conditions and fault properties and can thus advance the assessment of regional differences in earthquake ground motions.
Session:Recent Advances in Ground Motions Simulation Methods and Their Validation
Presenter   Petersson, Anders
Schedule   Thu / Poster
Room   Cook/Arteaga
Two Summation-By-Parts Finite Difference Codes for Large Scale Simulations of Seismic Motion
PETERSSON, N. A., Lawrence Livermore National Lab, Livermore, CA, USA, petersson1@llnl.gov; SJOGREEN, B., Lawrence Livermore National Lab, Livermore, CA, USA, sjogreen2@llnl.gov; RODGERS, A. J., Lawrence Livermore National Lab, Livermore, CA, USA, rodgers7@llnl.gov
We report a methodology for finite difference simulations of anelastic seismic motions using the summation by parts (SBP) method. The SBP method provides provably accurate and energy stable solutions. We describe the underlying methodology and two parallelized open- source codes developed for a variety of applications, based on second (WPP) and fourth order (SW4) accurate space and time discretizations. Our method uses a node-centered discretization of the seismic wave equations in displacement formulation on a structured, or piecewise structured, grid. Our approach is different from the conventional staggered grid used in many seismic FD methods. The node-centered approach allows us to deform the mesh near the free surface to accurately enforce the free surface boundary condition along the actual topography. At depth, mesh coarsening can be used to allow the grid size to increase as wave speeds increase, saving memory and increasing the time step. Mesh generation is automatic allowing for great ease of use to change the domain size, orientation and grid spacing. Absorbing boundary conditions are implemented using a super-grid approach. Attenuation is modeled through a linear viscoelastic reological model using a small number of standard linear solid (SLS) elements coupled in parallel. Three-dimensional (3D) material models can be rendered on the grid in various ways, including hierarchical octree models such as the USGS San Francisco Bay Area model. The accuracy and convergence of our SBP FD codes is demonstrated by verification against canonical problems and semi-analytical solutions. Parallel scaling properties of our codes will also be reported. Furthermore, we demonstrate the ability of our method, together with a well- developed 3D material model, to predict recorded ground motions for earthquakes in the eastern San Francisco Bay Area. The effect of free surface topography is investigated by computing the response with or without topography.
Session:Near-Field Seismoacoustics of Natural and Man Made Explosions
Presenter   Petersson, Anders
Schedule   Fri / Poster
Room   Cook/Arteaga
A Seismo-Acoustic Infrasound Methodology for Accurate Large Scale Simulations
PETERSSON, N. A., Lawrence Livermore National Lab, Livermore, CA, USA, petersson1@llnl.gov; SJOGREEN, B., Lawrence Livermore National Lab, Livermore, CA, USA, sjogreen2@llnl.gov; RODGERS, A. J., Lawrence Livermore National Lab, Livermore, CA, USA, rodgers7@llnl.gov
We present recent progress towards a coupled seismo-acoustic simulation capability. This work aims to improve the modeling of energetic events that trigger mechanical waves that propagate through both the solid earth and the atmosphere. Acoustic wave propagation can be modeled by solving the linearized Euler equations for compressible fluid flow. The acoustic waves are modeled as small perturbations on top of a given background velocity, pressure, and density field, which may vary in space. Acoustic waves are advected by the wind, which can cause significant bias in sound propagation, and modify the refraction of energy compared to a calm atmosphere. Because of advection, the acoustic wave equations are naturally formulated as a first order hyperbolic system (i.e., containing up to first derivatives in space and time). This system is written on symmetric form, and then discretized by a high order accurate finite difference scheme that satisfies the principle of summation by parts (SBP). Using the symmetric form of the governing equations allows an energy estimate to be derived through integration by parts. By applying similar operations on the SBP difference approximation, a discrete energy estimate can be derived, implying stability of our finite difference discretization of the acoustic wave equation. This approach generalizes to curvilinear coordinates, allowing effects of realistic topography to be taken into account. The acoustic modeling capability will be coupled to our current parallel seismic modeling code SW4, which implements a 4th order accurate SBP scheme to simulate 3D seismic wave propagation in a heterogeneous viscoelastic material, satisfying the free surface boundary condition along realistic topographies. The acoustic and seismic domains will be coupled by imposing continuity of traction and normal velocity along the solid-air interface.
Session:Geometric Complexities Along Strike-Slip Systems: New Insights on Seismic Hazards, Earthquake Behavior, and Fault System Evolution
Presenter   Walton, Maureen
Schedule   Wed 10:45 AM / Oral
Room   Room 2
STUDENT
Basement Structure and Earthquake Hazards along the Queen Charlotte-Fairweather Fault System, Southeastern Alaska
WALTON, M. A. L., The University of Texas at Austin Institute for Geophysics, Austin, TX, USA, maureenlwalton@utexas.edu; GULICK, S. P. S., The University of Texas at Austin Institute for Geophysics, Austin, TX, USA, sean@ig.utexas.edu; HAEUSSLER, P. J., U.S. Geological Survey, Anchorage, AK, USA, pheuslr@usgs.gov; ROLAND, E., U.S. Geological Survey, Anchorage, AK, USA, eroland@usgs.gov
In southeastern Alaska, neotectonics are dominated by the Queen Charlotte-Fairweather dextral transform fault system, which accommodates motion between the Pacific and North American plates. The slip rate is around 4.5 cm/yr, with a vector of oblique convergence increasing southward up to 20 mm/yr. The Queen Charlotte Fault (QCF), bounded by the complex Explorer triple junction to the south and the Yakutat triple junction to the north, has caused several large (> Mw 7) earthquake events, and many questions still remain about the QCF’s surficial geometry, regional stresses, tectonics, and fault plane structure. Study of the QCF to this point has been confined to localized regions; we use a compilation of marine geophysical data along the fault trace to provide unique insight into the big-picture changes in fault geometry, structure, and plate dynamics. Two recent earthquakes – a Mw 7.8 oblique-thrust event on 28 October 2012 offshore of Haida Gwaii and a Mw 7.5 strike-slip event near Craig, AK on 05 January 2013 – emphasize the importance of understanding fault structure along this active margin, especially in the context of other historical events. Models for accommodation of oblique convergence along the southern QCF are debated, suggesting Pacific Plate underthrusting, distributed intraplate deformation, or some combination of the two. Using a compilation of USGS and academic seismic reflection surveys and GLORIA sidescan sonar data, we re-map the QCF trace and constrain its surficial geometry and splays. We also map basement structure throughout the region and note downward flexure of the Pacific Plate on the northern edge of Haida Gwaii, shallowing southward to the Explorer triple junction. Normal faults in the sediment atop the Pacific Plate in the flexure region suggest an element of extension, perhaps caused by flexure and/or transtension from a right-stepping right-lateral fault system.
Session:Monitoring Dynamic Changes at Active Volcanoes and Fault Zones
Presenter   Ortiz, Hugo
Schedule   Thu / Poster
Room   Cook/Arteaga
Study of the Site Effects at Cotopaxi Volcano: Construction of a Seismic Activity Index
ORTIZ, H. D., Instituto Geofisico - Escuela Politecnica Nacional, Quito, Pichincha, Ecuador, hortiz@igepn.edu.ec; PALACIOS, P. B., Instituto Geofisico - Escuela Politecnica Nacional, Quito, Pichincha, Ecuador, ppalacios@igepn.edu.ec; RUIZ, M. C., Instituto Geofisico - Escuela Politecnica Nacional, Quito, Pichincha, Ecuador, mruiz@igepn.edu.ec
Seismic wave amplitudes are mostly modified by the underlying ground and by the topography of the place where they are recorded. There are many techniques commonly used to quantify these effects. In the present work, through the inverse problem methodology and based on a series of measurements, we propose the determination of the parameters of seismic models which contain the information of the site effects. Once the first results about site effects have been obtained, a refinement of these values using IRLS iterative process based on L1 norm was performed. After this refinement, a quick inspection of the confidence levels obtained by the iterative process was executed, which used a Monte Carlo simulation of a set of measurements contaminated with ambient noise following a normal distribution. The final goal of the site effect calculations is to normalize the seismic recordings of various stations installed at Cotopaxi volcano, and thus to construct and implement a unified seismic activity index, which will enable visualizing and quantifying the internal dynamics of the volcano. Eventually, the seismic activity index can be used to supplement volcano monitoring, reducing volcanic risk of inhabitants living in danger zones around Cotopaxi.
Session:75 Years of Frequency-Size-Distribution of Earthquakes: Observations, Models and Understanding
Presenter   Bachmann, Corinne
Schedule   Fri 11:15 AM / Oral
Room   Room 4
Geomechanical Modeling of Frequency-Magnitude Distributions of Induced Seismicity
BACHMANN, C. E., Lawrence Berkeley National Lab, Berkeley, CA, USA, cebachmann@lbl.gov; WIEMER, S., ETH Zurich, Zurich, Switzerland, stefan.wiemer@sed.ethz.ch; GOERTZ-ALLMANN, B. P., NORSAR, Kjeller, Norway, bettina.goertz-allmann@norsar.no
Induced seismicity from different sources is becoming more and more a public concern. The largest events in the recent past came from wastewater injections and geothermal heat extraction. In the future, Carbon Capture and Storage is believed to be an additional source. To understand the ongoing processes better, seismicity has to be monitored with a dense local network, including borehole stations in the best case. The evolution of the seismic cloud indicates the location of the injected fluid in the ground. While this is crucial to know for geothermal heat extraction, as the water will be extracted through a second borehole, it is also important for the understanding and analysis of the other processes. To better predict size distributions of the induced seismic events, we introduce a geomechanical model, simulating the pore pressure diffusion in the media. We relate the induced event-sizes to the differential stress via an empirical inverse relationship, established from tectonic events and laboratory experiments. With this simplistic model, we can reproduce the frequency-magnitude distributions found from injection experiments, such as the Enhanced Geothermal System in Basel, Switzerland and a long-term brine injection in Paradox Valley, Colorado. We find significant temporal and spatial changes in the b-value of those injections; highest values are found closest to the injection points and earliest during the injection. Our model allows us to relate those changes to the pore pressure changes. We conclude that highest pore pressure changes lead to high b-values as those induce events with smaller sizes. We are therefore able to establish a link between the seismological observables and the geomechanical properties of the source region and thus a reservoir. Understanding the geomechanical properties is essential for estimating the probability of exceeding a certain magnitude value in the induced seismicity and hence the associated seismic hazard of the operation.
Session:Development of 2014 U.S. National Seismic Hazard Maps and Their Implementation in Engineering Applications
Presenter   Petersen, Mark
Schedule   Thu 1:30 PM / Oral
Room   Room 7/8
The 2014 Update of the United States National Seismic Hazard Models
PETERSEN, M., USGS, Denver, CO, USA, mpetersen@usgs.gov; MOSCHETTI, M., USGS, Denver, CO, USA; POWERS, P., USGS, Denver, CO, USA; MUELLER, C., USGS, Denver, CO, USA; HALLER, K., USGS, Denver, CO, USA; FRANKEL, A., USGS, Seattle, WA, USA; ZENG, Y., USGS, Denver, CO, USA; REZAEIAN, S., USGS, Denver, CO, USA; HARMSEN, S., USGS, BOYD, O., USGS, Field, N., USGS, CHEN, R., CGS, RUKSTALES, K., USGS, LUCO, N., USGS, WHEELER, R., USGS, WILLIAMS, R., USGS, and OLSEN, A., USGS
During the past three years, the USGS updated the U.S. National Seismic Hazard models for the lower 48 states. These products are considered for inclusion in future building codes, risk assessments, and public policy applications. The seismic hazard models are based on our assessment of the best available science at the time of the update, and incorporate a broad range of models and parameters. Several new datasets and models have been developed since the 2008 update. For the Central and Eastern U.S. we implemented a new moment magnitude (Mw) catalog and completeness estimates, updated the maximum magnitude distribution, updated and tested the smoothing algorithms for adaptive and fixed-radius methods for estimating earthquake rates away from faults, modified the sizes and rates of New Madrid Seismic Zone earthquakes, and considered induced earthquakes. In the Intermountain West we implemented new smoothing algorithms, fault geometries for normal faults, multiple Wasatch fault alternatives, and fault slip rates based on models obtained by inverting geodetic and geologic data together. For the Pacific NW we developed new Cascadia fault rupture models that incorporate additional earthquakes in the south, added the Tacoma fault and an eastern extension of the Whidbey Island fault, and implemented the new geodetic-geologic based slip rates. For California we implemented the new UCERF3 source model that allows for a more complete inventory of fault ruptures and is based on new deformation models, magnitude-area equations, and seismicity input parameters. We also applied several new published ground motion models for shallow crustal earthquakes, subduction interface, and deep earthquakes. The many improvements in input models resulted in small changes across most of the country, but have caused significant changes up to ± 25% at 1 Hz and 5 Hz spectral accelerations in localized places. Hazard information is available on-line: http://earthquake.usgs.gov/hazmaps/.
Session:Topics in Seismology: Regional Seismicity and Tectonics
Presenter   Castro, Raul
Schedule   Thu / Poster
Room   Cook/Arteaga
Seismicity in the Basin and Range Province of Sonora, México, near the Rupture of the 3 May 1887 Mw 7.5 Earthquake, between 2003 and 2011
CASTRO, R. R., CICESE, Dep. Sismología, Ensenada, Baja California, México, raul@cicese.mx
We studied the regional seismicity in the Basin and Range Province of Sonora, México near the epicentral region of the 3 May 1887 Mw 7.5 earthquake. We used body wave arrival times recorded by the local network RESNES (Red Sísmica del Noreste de Sonora) and two regional broadband stations of the RESBAN (Red Sismológica de Banda Ancha del Golfo de California) network to locate the seismic sources originating between 2008 and 2011. We first determined preliminary hypocenter coordinates with the program Hypoinverse (Klein, 2002) and then relocated the initial coordinates with the source-specific station term method (Lin and Shearer, 2005). We had used the same procedure in a previous study (Castro et al., 2010) to locate the earthquakes originating in the same region between 2003 and 2007. We also present in this paper a depurated catalog of events recorded from 2003 to 2011. Most relocated events cluster near the faults that ruptured during the 1887 earthquake. Additionally, we document seismicity along Basin and Range Province normal faults south of the 1887 rupture, such as the Villa Hidalgo and Granados faults and the faults bounding the Bacadéhuachi basin. The region of the largest population center, Hermosillo, appears to be seismically quiescent. This observation is supported by a lack of entries in the catalog of the International Seismological Centre (ISC) or documented historical seismicity in that area.
Session:Geometric Complexities Along Strike-Slip Systems: New Insights on Seismic Hazards, Earthquake Behavior, and Fault System Evolution
Presenter   Hatem, Alex
Schedule   Wed / Poster
Room   Cook/Arteaga
STUDENT
Analog Modeling of Restraining Bends: A Study of Strike-Slip Fault Evolution
HATEM, A. E., University of Massachusetts, Amherst, MA, USA, ahatem@cns.umass.edu; COOKE, M. L., University of Massachusetts, Amherst, MA, USA, cooke@geo.umass.edu; MADDEN, E. H., University of Massachusetts, Amherst, MA, USA, ehmadden@geo.umass.edu
Restraining bends migrate and evolve over time, by propagation of new faults and abandonment of faults. Such processes are poorly constrained in space and time, leaving open questions about how strike-slip faults interact over geologic time. Scaled analog modeling using wet kaolin allows our team to make qualitative observations and quantitative analyses about restraining bend evolution. We document fault initiation and abandonment to see how restraining bends overcome initial mechanical inefficiencies. Deformation is captured in photographs later analyzed using Particle Image Velocimetry, a particle tracking software. Restraining bends within the claybox grow oblique-slip faults outboard of the initial restraining segment on one or both sides of the restraining bend. These new faults often link with the pre-cut faults. We quantify the evolving mechanical efficiency of the fault systems as a ratio of measured fault slip per image to applied plate displacement (0.5mm/minute). At the start of fault slip, the efficiency correlates to restraining bend angle; shallower restraining bends are initially more efficient. Restraining bends with angles greater than 20 degrees have about the same initial efficiency. Restraining bends with higher angles begin inefficient, but efficiency increases as new faults propagate and link with the existing. Regardless of kink angle, all restraining bends have roughly equal efficiency (85%) after 55 mm applied displacement. We examine the Denali strike-slip system in the claybox with 13 and 17 degree kink angle restraining bends; such models resemble current relative topography distribution and surface fault traces within the Alaska Range. Both analog and crustal restraining bends may be inefficient structures along otherwise efficient strike-slip fault systems. Similarly, faults overcome such inefficiencies by abandoning high angle, moderately efficient structures in favor of propagating new oblique-slip faults.
Session:Seismic Location and Processing Techniques
Presenter   Gomez, Demian
Schedule   Fri / Poster
Room   Cook/Arteaga
STUDENT
A Linear Formulation for Earthquake Location in a Homogeneous Half-Space Based on the Bancroft Algorithm Developed for GPS Location
GOMEZ, D. D., Center for Earthquake Research and Information, Memphis, TN, USA, ddgomez@memphis.edu; LANGSTON, C. A., Center for Earthquake Research and Information, Memphis, TN, USA, clangstn@memphis.edu; SMALLEY, R., Center for Earthquake Research and Information, Memphis, TN, USA, rsmalley@memphis.edu
The traditional approach to both earthquake and GPS location problems in a homogeneous half-space produces a nonlinear relationship between a set of known positions, seismic stations or GPS satellites, and an unknown point, an earthquake hypocenter or GPS receiver. Linearization, followed by an iterative inversion, is typically used to solve both problems. Although sources and receivers are inverted in the earthquake and GPS location problems, the observation equation is the same for both due to the principle of reciprocity. Consequently, single step closed-form solutions for the GPS location problem, such as the Bancroft algorithm, can also be used to solve for earthquake hypocenters. We apply the Bancroft algorithm to synthetic data for the New Madrid seismic network, showing improvements in both precision and accuracy compared to traditional methods. We also show how tools commonly used by the GPS community can be used to better estimate the precision of locations obtained by a seismic network.
Session:Induced Seismicity
Presenter   Rubinstein, Justin
Schedule   Thu 2:30 PM / Oral
Room   Room 1
Quantifying the Seismic Hazard From Natural and Induced Earthquakes
RUBINSTEIN, J. L., USGS, Menlo Park, CA, USA, jrubinstein@usgs.gov; LLENOS, A., USGS, Menlo Park, CA, USA, allenos@usgs.gov; ELLSWORTH, W. L., USGS, Menlo Park, CA, USA, ellsworth@usgs.gov; MCGARR, A., USGS, Menlo Park, CA, USA, mcgarr@usgs.gov; MICHAEL, A., USGS, Menlo Park, CA, USA, michael@usgs.gov; MUELLER, C., USGS, Golden, CO, USA, cmueller@usgs.gov; PETERSEN, M., USGS, Golden, CO, USA, mpetersen@usgs.gov
Induced earthquakes are believed to be at least partially responsible for the dramatic increase in seismicity in the Central and Eastern US over the past 12 years, yet they are presently excluded from USGS estimates of earthquake hazard. Here we propose an approach to include potentially induced earthquakes into the USGS National Seismic Hazard Map that deemphasizes the need to evaluate whether the seismicity is natural or man-made. We first compile a list of areas of increased seismicity, including areas of known induced earthquakes. Using areas of increased seismicity (instead of just induced earthquakes) allows us to assess the hazard over a broader region, avoiding the often-difficult task of judging whether an earthquake sequence is induced. We then estimate the earthquake hazard for each zone using a four-branch logic tree: (1) The increased seismicity rate represents short-term variation within a longer-term background seismicity rate. Thus, these earthquakes would be included in the background seismicity rate catalog. (2) The increased seismicity rate is a new and permanent addition to the background seismicity. Thus, the background seismicity rate is computed beginning at the time of the earthquake rate change (3) Induced earthquakes are responsible for the increased earthquake rate and they should be accounted for separately. This branch would compute different magnitude-frequency distributions for both kinds of events. The hazard from both would then be summed (4) The increased seismicity rate will decay back to the background seismicity rate. This could be applied to natural earthquake swarms or areas where industrial activity has stopped. Based on the understanding of an individual earthquake sequence, one could weight these logic tree branches as appropriate. Although an imperfect solution, this approach is a first step in accounting for the contribution of both induced earthquakes and variations in natural earthquake rates to seismic hazard.
Session:Seismic Location and Processing Techniques
Presenter   Harrington, Rebecca
Schedule   Fri / Poster
Room   Cook/Arteaga
Low-Frequency-Earthquake Locations Determined Using Time-Reverse-Imaging
HORSTMANN, T., Geophysical Institute, Karlsruhe Institute of Technology, Karlsruhe, Germany, tobias.horstmann@kit.edu; HARRINGTON, R. M., Earth & Planetary Sciences Dept., McGill University, Montreal, QC, Canada, rebecca.harrington@mcgill.ca; COCHRAN, E. S., U. S. Geological Survey, Pasadena, Pasadena, CA, USA, ecochran@usgs.gov; SHELLY, D. R., U. S. Geological Survey, Menlo Park, Menlo Park, CA, USA, dshelly@usgs.gov
The low-frequency-earthquakes (LFEs) associated with non-volcanic tremor typically exhibit non-impulsive phase arrivals that inhibit the use of standard ray-tracing location methods. Here we use time-reverse-imaging techniques that do not require phase arrival identification to locate individual LFEs within tremor episodes on the San Andreas fault near Cholame, California. We propagate a time reversed seismic signal back through the subsurface using a staggered-grid finite-difference code. We then search for wave field coherence in time and space (e.g. LFE energy in 1.5 second time windows filtered between 1-5 Hz) to obtain the source location and origin time where the constructive interference of the curl field energy is maximum. The grid point and time window occupying the spatial median of cross-correlation values within 10% of the maximum value indicate the source location and origin time. Location errors are based on the spatial extent of all cross-correlation coefficient values exceeding 90% of the maximum value. Horizontal and vertical errors are on the order of 4 km and 3 km respectively. We check the locations determined here with an LFE catalog determined by stacking hundreds of LFEs [Shelly and Hardebeck, 2010]. The LFE catalog uses stacks of at least several hundred templates to identify phase arrivals to estimate the location. We find that the epicentral locations determined here using the time-reverse-imaging technique are within ~4 km of the catalog LFE locations [Shelly and Hardebeck, 2010]. LFEs locate at depths between 15-25km, similar focal depths to previously published locations of LFEs or tremor in the region. Overall, the method can provide robust locations of individual LFEs without identifying and stacking hundreds of LFE templates, and is more accurate than envelope location methods with errors on the order of tens of km [Horstmann et al., 2013]. Location errors may be further reduced with increased velocity model resolution.
Session:Seismic Imaging as USArray Moves to Alaska (IRIS/Seismological Society of America Joint Session)
Presenter   Jin, Ge
Schedule   Thu 4:00 PM / Oral
Room   Room 4
STUDENT
Automated Surface Wave Phase Velocity Measuring System and Its Application on USArray
JIN, G., Lamont-Doherty Earth Observatory, Columbia University, Palisades, NY, USA, ge.jin@ldeo.columbia.edu; GAHERTY, J., Lamont-Doherty Earth Observatory, Columbia University, Palisades, NY, USA, gaherty@ldeo.columbia.edu
Earthquake surface waves are recorded by dense seismic arrays as a strong and consistent signal among the stations. By taking advantage of the wavefield similarity of nearby stations, we have developed a new technique to estimate surface wave phase velocities precisely and automatically. First, a time window to isolate fundamental-mode surface wave energy is built based on the group delays of all frequency bands of interest. We then calculate multi-channel broadband cross-correlation functions of the isolated waveforms from nearby stations, and fit narrow-band filtered cross correlations with a five-parameter controlled wavelet to retrieve the optimized phase difference at a range of frequencies. The amplitude of this cross-correlation function can be used to estimate the coherence, which together with signal to noise ratio (SNR) are the two key factors to exclude unqualified measurements. The phase difference information between all the nearby station pairs for each event at each frequency is then used as the input to an Eikonal tomographic inversion for two-dimensional estimates of apparent phase velocity. We measure the amplitude of each station by adapting the same process on the auto-correlation function and perform Helmholtz tomography to estimate and remove interference (focusing and defocusing) effects. Finally, we stack the measurements of each individual event to get the final phase-velocity tomogram in each frequency band. The entire process requires no human interaction and is highly automated. This method can be applied on data from arrays of various apertures, ranging from continental scale such as USArray, to regional scales (few hundred km) typical of PASSCAL arrays, to local-scale high-frequency arrays employed in industry and hazard investigations. By combining this analysis with programmable data acquisition services such as IRIS DMC web services, we set up an automatic system providing up-to-date surface wave phase velocity maps of USArray.
Session:Advances in Understanding Earthquake Hazard in Central and Eastern North America
Presenter   Blakely, Richard
Schedule   Fri 11:15 AM / Oral
Room   Room 2
High-Resolution Aeromagnetic Data Illuminate Crustal Structure Associated with the New Madrid Seismic Zone, Missouri and Arkansas
BLAKELY, R. J., U.S. Geological Survey, Menlo Park, CA, USA, blakely@usgs.gov; WILLIAMS, R. A., U.S. Geological Survey, Golden, CO, USA, rawilliams@usgs.gov; SHAH, A. K., U.S. Geological Survey, Denver, CO, USA, ashah@usgs.gov; SHERROD, B. L., U.S. Geological Survey, Seattle, WA, USA, bsherrod@usgs.gov; WEAVER, C. S., U.S. Geological Survey, Seattle, WA, USA, craig@usgs.gov
In 2013, the U.S. Geological Survey completed a high-resolution aeromagnetic survey in the New Madrid seismic zone (NMSZ). Our goal was to map and characterize crustal structure and contribute to ongoing earthquake-hazard investigations. The survey was flown along NW-directed flight lines spaced 400 m apart at a height of 200 m above ground. The survey spans Crowley’s Ridge, the NW margin of the Reelfoot rift, and the southern arm of the NMSZ, and was designed to overlap a high-resolution LiDAR survey completed in 2012 to allow for cross-comparison of features found on both datasets. Power lines, pipelines, and town infrastructure dominate parts of the survey but their anomalies were attenuated with matched-filter analysis. The remaining crustal magnetic field is remarkably smooth, reflecting very low magnetizations of near-surface rocks and deposits. Nevertheless, aggressive data processing allowed us to illuminate detailed near-surface (upper 2 km) crustal features apparently at the top of Precambrian basement northwest of and mostly striking parallel to the Reelfoot rift NW margin. The most prominent of these anomalies corresponds with the Commerce geophysical lineament. Other shallow anomalies indicate basement structure and/or contacts. One of these strikes NE, crosses the entire survey (41 km), and corresponds with an alignment of three LiDAR topographic scarps interpreted as tectonic in origin but not associated with earthquakes in the modern record. The magnetic survey found no evidence of near-surface faults directly responsible for ongoing seismicity in the southern arm of the NMSZ. Various explanations are possible: (1) Near-surface lithologies in this part of the Reelfoot rift may be virtually nonmagnetic. (2) Faults in the southern arm of the NMSZ do not juxtapose lithologies with contrasting magnetic properties. (3) The southern arm of the NMSZ is short-lived, with insufficient time to accumulate discernible slip on individual faults.
Session:Citizen Seismology: Citizens Helping Science Helping Citizens
Presenter   Hough, Susan
Schedule   Fri 1:30 PM / Oral
Room   Room 7/8
Earthquake Intensity Distributions: A New View
HOUGH, S. E., US Geological Survey, Pasadena, CA, USA, hough@usgs.gov
Previous studies have demonstrated both the tremendous value of macroseismic data and the perils of their uncritical assessment. Traditionally, archival accounts of historical earthquakes are used to assign intensity values using the modified Mercalli intensity (MMI) or other scales. Traditional MMI values (MMI_T) for historical events can be analyzed using intensity-prediction equations developed from MMI_T values estimated for instrumentally recorded events. The U.S. Geological Survey “Did You Feel It?” (DYFI) system now collects and systematically interprets thousands of first-hand reports from felt earthquakes, using an algorithm to estimate Community Decimal Intensity (CDI) values within ZIP codes or geocoded cells. A comparison of the intensity-prediction equations developed independently for MMI_T and CDI data reveals a discordance in amplitude and distance-decay, with MMIT distributions suggesting more dramatic effects at greater distances, even when intensities are assigned according to more modern conservative practices. Using spatially rich CDI data for the 2011 Mineral, Virginia, earthquake, I show that CDI values within large cities are normally distributed; by design the DYFI system assigns mean values from multiple responses. I present evidence that MMI_T values tend to be controlled by extreme effects due to a fundamental tendency to report dramatic rather than representative effects. I introduce an empirical correction-factor approach to correct for this bias. This allows the growing wealth of well-calibrated DYFI data to be used as calibration in the analysis of historical earthquakes. As a case study I analyze the 1868 Hayward, California, earthquake, for which a particularly rich MMIT dataset is available, and estimate a magnitude of 6.4-6.6, a lower value than estimated by previous studies. I discuss the implications for probabilistic seismic hazard assessment of partially creeping faults.
Session:New Insights into Ground Failure as an Urban Earthquake Hazard
Presenter   Clahan, Kevin
Schedule   Thu / Poster
Room   Cook/Arteaga
Landslide Hazard Mapping and Quantitative Landslide Hazard Planning Scenario for Aizawl, India
CLAHAN, K. B., Lettis Consultants International, Inc., Walnut Creek, CA, USA, clahan@lettisci.com; DEE, S., Lettis Consultants International, Inc., Walnut Creek, CA, USA, dee@lettisci.com
Aizawl, the State Capital of Mizoram, in northeast India is a densely populated hill top city located in the Western Burmese ranges. Aizawl City is located along the topographic crest and steep limbs of an anticlinal fold in the Burmese fold and thrust belt. Steep slopes, often in concert with heavy monsoonal rainfall, are subject to devastating failure in the form of pervasive landsliding that has dominated the historical geologic hazard record in the region. This active and ongoing geologic process was reinforced most recently on May 12, 2013 when a catastrophic landslide occurred in the Laiputlang area resulting in the destruction of several properties and the deaths of 17 people. A comprehensive landslide hazard planning scenario was recently created for the City of Aizawl, India. The purpose of the landslide scenario was to estimate the percentage of the total area that will fail under the scenario earthquake. A landslide hazard map was created to estimate the percentage of failure as a function of hazard zone. This estimation is based on theoretical and empirical landslide inventory data as well as from site specific data collected from the study area and the results of our slope hazard calculations. We used both quantitative and qualitative methods to estimate the total landslide area (ALmax) that might fail during a Mw = 7.0 earthquake with PGA = 0.35 in close proximity to Aizawl. Malamud et al. (2004) presents a statistical distribution of landslide areas and landslide volumes by defining a probability density function using three complete landslide inventories. Based on the agreement between these three sets of probability densities a distribution equation is presented from which landslide areas and volumes can be predicted for scenario events (Malamud et al., 2004). The quantitative landslide model is combined with empirical relations of landslide density to calculate a percentage of area expected to fail during the seismic shaking.
Session:Advances in Understanding Earthquake Hazard in Central and Eastern North America
Presenter   Ogweno, Luke Philip
Schedule   Fri 9:30 AM / Oral
Room   Room 2
STUDENT
Comparing the CENA GMPEs Using NGA-East Ground Motion Database
OGWENO, L. P., The University of Memphis - CERI, Memphis, TN, USA, logweno@memphis.edu; CRAMER, C. H., The University of Memphis - CERI, Memphis, TN, USA, ccramer@memphis.edu
The Next Generation Attenuation (NGA) East project has an updated database for Central and Eastern North America (CENA) ground motions. The objective of this study is to analyze the performance of ground motion prediction equations (GMPEs) used in the United States Geological Survey (USGS) National Seismic Hazard Mapping Project (NSHMP) and some other potential GMPEs used in the CENA. Analyses include bias analysis and checks for model inadequacies using statistical tests such as cumulative probability plots, histograms and boxplots. Ranking of the GMPEs is accomplished using minimum residual, log likelihood (LLH), and Euclidean Distance Based Ranking (EDR) techniques. From the classical residual analysis, Atkinson and Boore (2011) (model A08p), Atkinson and Boore (2011) (model AB06p) and Atkinson and Boore (2006) (model AB06+) with 200 bar stress drop performed better than other GMPEs. These results were also analyzed using box plots to summarize the location, dispersion, and symmetry or skewness of the residual values. In general the NGA East database is positively skewed i.e., the right tail is more pronounced than the left tail. This implies that most of the observed ground motion values are concentrated to the left of the mean, with extreme values to the right. GMPE logarithmic residuals are leptokurtic and non-normally distributed, although not strongly so. EDR results show models Atkinson (2008) (model A08), AB06p and AB06+ as the best performing models for combined site classes. Models AB06p, EPRI (2004) cluster2 model (EPRI2), AB06+ and Silva et al., (2002) double corner model (SD02) matched the ground motion to the data well in rock sites. Generally, newer GMPEs tend to predict lower ground motion levels than older GMPEs. This is attributed to differences in the geometrical spreading used with the newer GMPEs using R-1.3 verses R-1.0 in the older GMPEs.
Session:Advances in Seismic Imaging and Monitoring of Time-Dependent Variations: Civil Structures, Near-Surface, and Shallow Crustal Scales
Presenter   Gueguen, Philippe
Schedule   Wed 11:00 AM / Oral
Room   Room 3
Nonlinear Elasticity and Slow Dynamics Observed at the Factor Building (UCLA-California) by Processing Continuous Recordings.
GUEGUEN, P., ISTerre/UJF/CNRS/IFSTTAR, Grenobe, France, philippe.gueguen@ujf-grenoble.fr; ROUX, P., ISTerre/UJF/CNRS/IFSTTAR, Grenoble, France, philippe.roux@ujf-grenoble.fr; JOHNSON, P. A., Geophysics Group, Los Alamos National Laboratory, Los Alamos, NM, USA, paj@lanl.gov
During earthquakes, a structure may suffer damage, ultimately changing its elastic properties. Earthquake engineers simplify building response by describing it as bi-linear behaviour. The appearance of cracks in the building materials during events causes a transient variation of Young’s modulus, a variation which becomes permanent if irreversible damage occurs. Frequency and damping are closely related to the system’s Young’s modulus and, assuming constant mass, control its dynamic response according to Newton’s second law. Monitoring these values therefore enables the assessment of rapid damage to the structure, or nonlinear response, in situations such as after an extreme event, or slow deterioration due to ageing (Structural Health Monitoring). However, real data also reveals a sometimes slow recovery of the structure’s elastic properties when stress returns to a very low level. The origin of this recovery is not clear. In this study, we will show that under slight internal deformation, the elastic parameters of the Factor Building (UCLA California) follow a nonlinear elasticity regime, typical of a slow dynamic. We found that using very precise measurement of frequency and damping under slight stress, these parameters are modified by around 0.1% for deformation levels of the order of 10-7. These observations tend to confirm the universal nonlinear elastic behaviour of actual systems, comparable to that of rock or soil, as observed in laboratory conditions. Our results show that the dynamics of a structure composed of composite material and linked to the ground, are similar to those of an Nonlinear Mesoscopic Elasticity material, which had not been considered previously. Our observation will enable better understanding of the dynamic response of structures; changes in design and constructi
Session:From the Earthquake Source to Damage of Buildings: Bridging the Gap between Seismology and Earthquake Engineering
Presenter   Gueguen, Philippe
Schedule   Fri 4:00 PM / Oral
Room   Room 1
Macroscale Vulnerability Assessment Using Association Rule Learning, a Data-mining Tool.
RIEDEL, I., ISTerre/UJF/CNRS/IFSTTAR, Grenoble, France, ismael.riedel@ujf-grenoble.fr; GUEGUEN, P., ISTerre/UJF/CNRS/IFSTTAR, Grenoble, France, philippe.gueguen@ujf-grenoble.fr
A complete seismic risk assessment requires not only the estimation of the seismic hazard, 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 analyse, the construction type variability, the misunderstanding in the behaviour of an old structure, and to the intrinsic variability in the response of structures to seismic loads. The well-known methodologies conceived using post-event damage information in earthquake prone countries cannot be easily applied in moderate-to-low seismic hazard regions like central Europe. We propose a way to perform a fast estimation with a comparable accuracy, using only few, easy-to-get and certain information on buildings. Using a dataset of existing buildings in Grenoble (France) with an EMS98 vulnerability classification and by means of two different datamining techniques – Association Rule Learning and Support Vector Machine - we developed a seismic vulnerability proxy. This proxy is applied to the entire country using basic information contained in national databases (census information), obtaining a vulnerability map nation-wide. We also show that including information on building obtained from remote sensing techniques (i.e. treatment of satellite images) we can improve the accuracy on the vulnerability estimation. The approach was validated in Nice (France) by comparison with the RiskUE method and finally, damage estimations were compared with historic earthquakes that produced moderate-to-strong damage in France.
Session:New Directions in PSHA: Ins, Outs, and Uncertainty
Presenter   Kijko, Andrzej
Schedule   Fri / Poster
Room   Cook/Arteaga
Straightforward Bayesian Procedure for Estimation of the Regional Characteristic, Maximum Possible Earthquake Magnitude mmax
KIJKO, A., Natural Hazard Centre, University of Pretoria, Pretoria, South Africa, andrzej.kijko@up.ac.za; SMIT, A., Natural Hazard Centre, University of Pretoria, Pretoria, South Africa, ansie.smit@up.ac.za
Upon investigations into the mathematical structure of the currently used Bayesian procedure for the estimation of the regional characteristic, maximum possible earthquake magnitude mmax (Cornell, 1994), a mathematical flaw was discovered which results in the biased assessment of mmax (Kijko, 2012). The degree of the 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 has been shown that if the maximum posterior estimate is used, the bias is negative and the underestimation of mmax can be as big as 0.5 magnitude units. The cause of the problem lies in the properties of 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 (LeCam, 1970) of the maximum likelihood function. The possibility to modify the current Bayesian procedure through the replacement of the ‘incorrect’ sample likelihood function with a seismicity-driven distribution of mmax is explored. The Monte-Carlo simulations show that the newly proposed approach is surprisingly very efficient and is capable of providing correct values of mmax even when a low number of seismic events are available References Cornell C.A. (1994). Statistical Analysis of Maximum Magnitudes”. EPRI, Palo Alto, Calif. 1994; NP-4726: 5-1-5-27. Kijko A. (2012). On Bayesian Procedure for Maximum Earthquake Magnitude Estimation. Research in Geophysics, 2:e7, Number 1. http://www.pagepress.org /journals/index.php/rg. LeCam L. (1970). On the Assumptions Used to Prove Asymptotic Normality of Maximum Likelihood Estimates. Ann. Statist., 41, 802-828
Session:Seismic Imaging as USArray Moves to Alaska (IRIS/Seismological Society of America Joint Session)
Presenter   Busby, Robert
Schedule   Thu / Poster
Room   Cook/Arteaga
EarthScope’s Transportable Array in Alaska
BUSBY, R. W., Inc Research Institutions for Seismology, Washington, DC, USA, busby@iris.edu; HAFNER, K., Inc Research Institutions for Seismology, Washington, DC, USA, hafner@iris.edu; WOODWARD, R. L., Inc Research Institutions for Seismology, Washington, DC, USA, woodward@iris.edu
Since 2003, EarthScope has been installing seismometers in a network, known as the Transportable Array, across the continental United States and southern Canada. The station deployments in the conterminous US were completed in the fall of 2013. In October 2013, and continuing for 5 years, EarthScope’s Transportable Array plans to create a grid of seismic sensors in approximately 265 locations In Alaska and Western Canada. The proposed station grid is 85 km, and target locations will supplement or enhance existing seismic stations operating in Alaska. Where possible, they will also be co-located with existing GPS stations constructed by the Plate Boundary Observatory. Stations will utilize the complete sensor suite as in the lower-48 (broadband seismic and atmospheric) with potentially some additional sensors for soil temperature, meteorological and/or strong motion. We describe the deployment plan and schedule for installation, and details of the engineering concept for the stations. Opportunities for collaborative measurements at TA stations will also be described. IRIS has installed several experimental stations to evaluate different sensor emplacement schemes both in Alaska and the lower-48 U.S. The goal of these tests is to maintain or enhance a station’s noise performance while minimizing its footprint and the equipment, materials, and overall expense required for construction. Motivating this approach are recent developments in posthole broadband seismometer design and the unique conditions for operating in Alaska, where most areas are only accessible by small plane or helicopter, and permafrost underlies much of the region. IRIS has tried different portable drills and drilling techniques to create shallow holes (1-5M) in permafrost and rock outcrops. The results indicate that seasonal changes can affect the performance of seismometers in different ways depending on the emplacement technique. Updates on these results will be described.
Session:Advances in Understanding Earthquake Hazard in Central and Eastern North America
Presenter   Hassani, Behzad
Schedule   Fri 9:15 AM / Oral
Room   Room 2
STUDENT
Referenced Empirical Ground-Motion Model for Eastern North America, Based on the NGA-West 2 GMPE
HASSANI, B., Western University, London, ON, Canada, bhassan7@uwo.ca; ATKINSON, G. M., Western University, London, ON, Canada, gatkins6@uwo.ca
Ground-motion prediction equations (GMPEs), giving expected ground-motion amplitudes as a function of magnitude and distance, are an important component of earthquake hazard analysis. A useful method for development of GMPEs in data-poor regions such as eastern North America (ENA) is the referenced empirical approach (Atkinson, 2008; Atkinson and Boore, 2011), in which sparse observational data are compared to an empirical GMPE from a data-rich region. We compare a recent GMPE for shallow events in active regions (Boore, Stewart, Seyhan and Atkinson, denoted BSSA14), with recorded ground motions in ENA (peak ground motions and response spectra), considering both Central U.S. data (CUS) and data from southeastern Canada and northeastern U.S. (east). We note that BSSA14 is well constrained at low magnitudes (M3.5+), making comparisons to ENA data robust. Residuals are defined as the ratio of the observed ENA values to the predicted values based on BSSA14. The residuals trends will be a function of frequency and distance, and reflect regional differences in source amplitude and attenuation. We observe that the residuals trends are similar in the CUS and east regions, indicating just slightly higher attenuation in the CUS. The Gulf Coast region, by contrast, shows significantly higher attenuation. In general, at close distances (<60 km), residuals tend to be insignificant except at very high frequencies (>10Hz); this suggests that western-based GMPE models might be used to predict eastern motions in high-hazard regions, for which sources at <60 km dominate (except for frequencies >10 Hz and PGA). As distance increases, the residuals gradually increase, presumably due to the slower attenuation rate at regional distances in ENA compared to that in active regions. The proposed referenced empirical GMPEs can be used for hazard analysis in ENA, in particular to provide empirically-constrained scaling of motions to magnitudes larger than those available in the ENA database.
Session:Seismic Imaging as USArray Moves to Alaska (IRIS/Seismological Society of America Joint Session)
Presenter   Pulliam, Jay
Schedule   Thu 4:30 PM / Oral
Room   Room 4
Lithospheric Modification at the Southern Edge of Laurentia Revealed by Broadband Seismology
PULLIAM, J., Baylor University, Waco, TX USA, jay_pulliam@baylor.edu; GRAND, S. P., University of Texas at Austin, Austin, TX USA, steveg@jsg.utexas.edu; GURROLA, H., Texas Tech University, Lubbock, TX USA, HAROLD.GURROLA@ttu.edu
The southern margin of Laurentia has been modified by a wide range of tectonic processes, including deformation due to repeated orogenies and continental collisions (Llano, Ancestral Rockies, Ouachita, Laramide) as well as ongoing rifting (at the Rio Grande Rift) and completed rifting (which created the Gulf of Mexico). Artifacts of these processes are likely to remain at lithospheric depths but, until recently, the tools needed to examine 3D structure at mantle depths beneath the region were not available. With the passage of the EarthScope’s USArray and two targeted broadband deployments, new images of the region’s lithosphere have emerged. These reveal lithospheric-scale anomalies that correlate strongly with surface features, such as a large fast anomaly that corresponds to the southern extent of the Laurentia (or “Great Plains”) craton and a large slow anomaly associated with the Southern Oklahoma Aulacogen. Other features were unanticipated, including slow shear velocities beneath the Texas-GoM margin, a layer that is bounded at its top and bottom by seismic discontinuities, and high levels of seismic anisotropy. Additionally a high velocity body underlying the Gulf Coast Plains may mark delaminating lower crust. If so, it provides indirect evidence that active rifting best describes the process that lead to the opening of the GoM. Lithospheric thinning associated with the Rio Grande Rift appears to be propagating eastward and eroding the Great Plains craton. A large, seismically fast anomaly imaged beneath west Texas and southeastern New Mexico has implications for the region’s structural evolution. The size, seismic velocity, and location of this anomaly suggests that it may be the result of a “step instability,” in which progressive deformation has locally removed mantle lithosphere beneath the eastern flank of the Rio Grande Rift, resulting in thickened lithosphere at the western edge of the Great Plains craton and uplift beneath the rift flank.
Session:Seismic Location and Processing Techniques
Presenter   Johnson, Courtney
Schedule   Fri / Poster
Room   Cook/Arteaga
STUDENT
Preconditioning Seismic Data Using the Hilbert Huang Transform Prior to Gradiometric Analysis
JOHNSON, C. E., Georgia Regents University, Augusta, GA, USA, cjohn102@gru.edu; POPPELIERS, C. J., Georgia Regents University, Augusta, GA, USA, cpoppeli@gru.edu
A recorded seismic signal is typically a superposition of several disparate wavefields. This poses a challenge for any seismic array data analysis method that assumes only a single wavefield is recorded by an array at any given time. For example, the new method of seismic gradiometry is extremely sensitive to the violations of the single wavefield assumption. As a result, gradiometry returns unreliable results when applied to a signal that contains several superposed wavefields. In such a case, preconditioning the data using some form of wavefield separation method prior to gradiometric analysis can be helpful if the data contains wavefields with distinct time or frequency components. However, most methods assume that the signal is stationary and linear. Given that seismic signals are neither stationary nor linear, it may be more helpful to employ a wavefield separation method that does not assume either. The Hilbert-Huang Transform (HHT) is an algorithm designed to decompose a signal which can contain nonstationary and nonlinear components. In this work, we explore the feasibility of applying the HHT to decompose the signal prior to gradiometric analysis.
Session:Alaska Update of the USGS National Seismic Hazard Maps
Presenter   Briggs, Richard
Schedule   Thu 10:45 AM / Oral
Room   Room 4
Implications of Recent Paleoseismic Observations for Models of Alaska-Aleutian Megathrust Rupture Patterns
BRIGGS, R. W., US Geological Survey, Golden, CO, USA, rbriggs@usgs.gov; WITTER, R. C., US Geological Survey, Anchorage, AK, USA, rwitter@usgs.gov; NELSON, A. R., US Geological Survey, Golden, CO, USA, anelsong@usgs.gov; KOEHLER, R. D., State of Alaska, Division of Geological Surveys, Fairbanks, AK, USA, richard.koehler@alaska.gov; HAEUSSLER, P. J., US Geological Survey, Anchorage, AK, USA, pheuslr@usgs.gov; ENGELHART, S. E., Department of Geosciences, University of Rhode Island, Kingston, RI, USA, engelhart@mail.uri.edu; GELFENBAUM, G., US Geological Survey, Santa Cruz, CA, USA, ggelfenbaum@usgs.gov; DURA, T., Department of Earth and Env. Science, University of Pennsylvania, Philadelphia, PA, USA, dura@sas.upenn.edu; Carver, G., Kodiak, AK, USA, cgeol@acsalaska.net
Here we summarize post-2007 paleoseismic observations along the Alaska-Aleutian megathrust and suggest that the next update of the USGS National Seismic Hazard Map (NSHM) for Alaska allow for a greater variety of multi-segment ruptures, revisit maximum magnitudes (Mmax), and relax the assumption that segment boundaries match historical rupture boundaries. The 2007 NSHM model allowed a characteristic M 9.2 earthquake with 650-year recurrence to rupture the Kodiak and Prince William Sound (PWS) segments. Recent work by Shennan et al. (2009) suggests that the Yakataga and PWS segments have ruptured together, and that the NSHM update should consider alternate multi-segment ruptures supported by paleoseismic data. The 2007 model fixed the 1964 rupture boundary between the Kodiak and Semidi segments. Our observations at Sitkinak Island of the 1964, 1788, and earlier ruptures suggest that this boundary may not persist at time scales relevant for hazard calculations. Our observations at Chirikof Island suggest that the Semidi segment has generated tsunamis every 250-400 years for the past ~3.4 ka, but it is not yet clear if the 1938 M 8.2 rupture released more or less moment than the 1788 rupture. This is important for time-dependent calculations and for re-evaluation of Mmax=8.5 for the Semidi segment. At Simeonof Island in the Shumagin segment, we find no geologic evidence for sudden coastal uplift or subsidence or for inundation from large tsunamis since ~3.4 ka. Although elastic dislocation models permit ruptures of up to M 8.5 trenchward of the island, an absence of high tsunami deposits on Simeonof implies such models are unlikely. Finally, tsunami deposits at Sedanka and Umnak Islands suggest 280–320 yr recurrence of heterogeneous ruptures in the area of the 1957 M 8.6 earthquake. Uncertainty in past rupture behavior in the western Aleutians segment supports continued use of truncated Gutenberg-Richter behavior until more specific rupture patterns are defined.
Session:Recent Advances and Findings in Earthquake Geology and Paleoseismology
Presenter   Tong, Xiaopeng
Schedule   Fri / Poster
Room   Cook/Arteaga
Is There a Discrepancy between Geological and Geodetic Slip Rates along the San Andreas Fault System?
TONG, X., SIO/UCSD, La Jolla, CA, USA, xitong@ucsd.edu; SMITH-KONTER, B., University of Hawaii, Honolulu, HI, USA, brkonter@hawaii.edu; SANDWELL, D. T., SIO/UCSD, La Jolla, CA, USA, dsandwell@ucsd.edu
Several previous inversions for slip rate along the San Andreas Fault System (SAFS), based on elastic half-space models, show a significant discrepancy between the geological and geodetic slip rates along a few major fault segments. We use a more realistic model of an elastic plate over a viscoelastic mantle to demonstrate that there is no significant discrepancy between long-term geologic and geodetic slip rates. The model includes steady slip along 40 major fault segments from the base of the locked zone to the base of the elastic plate along with episodic shallow slip based on known ruptures. The slip rates are constrained by 1989 velocity vectors from EarthScope GPS and ALOS InSAR data. In addition to slip rates, our inversion solves for the translation and rotation motion, as well as shallow creep along several creeping faults. The geodetic slip rates inferred from a model consisting of a 60 km thick plate and 10^19 Pa s viscosity agree to within the bounds of the geological slip rates, while the rates from a simpler half-space model disagree on the Mojave and the North Coast segments. In particular, along the Mojave segment, the recovered geodetic slip rate is 24.7 mm/yr for the half-space model but the result comes closer to the preferred geological rate of 34 mm/yr using a 60 km thick plate model (27.5 mm/yr) and a 30 km thin plate model (34.4 mm/yr). The plate models have generally higher slip rates than the half-space model because most of the faults along the SAFS are late in the earthquake cycle, so today they are moving slower than the long-term cycle-averaged velocity as governed by the viscoelastic relaxation process. We also explore the dependency of transient deformation on paleoseismology data (i.e., known ruptures extending back 1000 years). We found that the assumption of the characteristic earthquake model plays a minor role on the inferred slip rates, but the timing of the major events has a greater influence on slip rate result.
Session:Advances in Understanding Earthquake Hazard in Central and Eastern North America
Presenter   Bent, Allison
Schedule   Fri 10:45 AM / Oral
Room   Room 2
Time Dependence of Magnitude Conversion Relations for Eastern Canada: More Complex than We Thought
BENT, A. L., Geological Survey of Canada, Ottawa, ON, Canada, abent@nrcan.gc.ca
A change of 0.12 in the mN-MW relation for eastern Canada in the mid-1990s was previously established and found to be related primarily to changes in instrumentation during a seismograph network upgrade, which led to mN being calculated at higher frequencies than in the past. The national seismic monitoring network has undergone several previous large-scale upgrades raising the question whether there could have been additional such magnitude relation changes. There are not sufficient instrumental MW's, the preferred scale for hazard assessment, for past earthquakes for a thorough investigation. There are, however, amplitude-period data in the Canadian earthquake database back to 1940. For recent earthquakes MLg(f) is generally close to MW. MLg(f) was calculated and compared to mN for earthquakes in eastern Canada from 1940 through 2012. The difference between the two shows an increase from 0.0 to 0.6 with several steplike increases coinciding with network upgrades. Over the same period a decrease in the average period at which mN was calculated was noted. Data from recent earthquakes have shown that the difference between MLg(f) and mN is dependent on both frequency and distance. The same dependence is observed for the older earthquakes. That is, the time dependency of the magnitude relations appears to be related to changes in the input parameters and not directly to the calculations or instruments themselves. One must be careful not to assume that a change in the MLg(f)-mN relation is conclusive evidence for a change in the mN-MW relation. However, these results urge the use of caution in applying a single conversion constant for data amassed over several decades and raise the issue whether there may be similar time dependencies for other magnitude relations and in other regions.
Session:Development of Next Generation Field Methods for Portable Broadband Seismic Arrays
Presenter   Taylor, Steven
Schedule   Wed 11:45 AM / Oral
Room   Room 7/8
A Ground-Truth Mine Monitoring System
TAYLOR, S. R., Rocky Mountain Geophysics, Los Alamos, NM, USA, srt-rmg@comcast.net; JARPE, S. P., Jarpe Data Solutions, Prescott Valley, AZ, USA, jarpe@pobox.com; HARBEN, P. E., Rocky Mountain Geophysics, Los Alamos, NM, USA, philip.harben@gmail.com; HARRIS, D. B., Deschutes Signal Processing, Maupin, OR, USA, oregondsp@gmail.com
A Ground Truth Mine Monitoring System (GTMS) is currently being developed under a DOE Phase II SBIR entitled “Development of Mine Explosions Ground Truth Sensors”. The concept behind the SBIR is to develop an inexpensive, compact seismo/acoustic system that is simple to deploy and can be mailed to an institution for placement near mines that conduct large explosions. The ground-truth data provided by the GTMS can be used to calibrate regional nuclear-explosion monitoring networks. An on-board processor allows for single-station estimation of location, origin time and magnitude (yield) and signal diagnostic parameter data that is transmitted back to a Ground Truth Processing Center (GTPC) via two-way ORBCOMM satellite communications. Two-way satellite communication allows for deployment in remote regions and the possibility of limited waveform transmission that can be used to tune on-board processing parameters. Continuous or event-detected data is kept on two removable flash SD cards of up to 64+ Gbytes each. Low-power consumption allows for autonomous operation using only a 10 watt solar panel and a gel-cell battery. The GTMS has been successfully tested for long-term (> 6 months) remote operations over a wide range of environments including summer in Arizona to winter above 9000’ in the mountains of southern Colorado. On-board processing is being developed that includes detectors, P and acoustic wave phase picking, back azimuth estimation and magnitude (yield) estimates from coda waves and acoustic signals. A specialized leading edge picking algorithm has been developed to remove late bias from STA/LTA picks. A single-station Bayesian location algorithm is being developed that combines prior information on mine location with P and acoustic wave arrival time picks and P-wave back azimuth. A matched filter has been developed to obtain explosion yield estimates from automated impulse per unit area measurements.
Session:Pillars of Simulation: Seismic Velocity and Material Models
Presenter   Hearn, Thomas
Schedule   Wed / Poster
Room   Cook/Arteaga
STUDENT
Frequency Dependent Lg Attenuation in Northeast China
RANASINGHE, N. R., New Mexico State University, Las Cruces, NM, USA, nrana001@nmsu.edu; GALLEGOS, A. C., New Mexico State University, Las Cruces, NM, USA, agall@nmsu.edu; TRUJILLO, A. R., Schlumberger Information Solutions, Houston, TX, USA, andrea.trujillo87@gmail.com; SANDVOL, E. A., University of Missouri, Colombia, MO, USA, sandvole@missouri.edu; NI, J., New Mexico State University, Las Cruces, NM, USA, jni@nmsu.edu; HEARN, T. M., New Mexico State University, Las Cruces, NM, USA, thearn@nmsu.edu; TANG, Y., China University of Petroleum, Beijing, China., jacktangyc@gmail.com; GRAND, S. P., University of Texas at Austin, Austin, TX, USA, steveg@jsg.utexas.edu; Niu, F.L, Rice University, Houston, TX, USA, niu@rice.edu; Chen, Y.J., Peking University, Beijing, China, johnyc@pku.edu.cn; Ning, J.y., Peking University, Beijing, China, njy@pku.edu.cn; Kawakatsu, H., University of Tokyo, Tokyo, Japan, hitosi@eri.u-tokyo.ac.jp; Tanaka, S., Japan Agency for Marine-Earth Science and Technology, Yokosuka, Japan, stan@jamstec.go.jp; Obayashi, M., Japan Agency for Marine-Earth Science and Technology, Yokosuka, Japan, stan@jamstec.go.jp.
The 127 station Northeast China Extended Seismic Array (NECESSArray), previous PASSCAL experiments, GSN broadband stations, Chinese, Japanese and South Korean permanent stations provides unprecedented seismic data in Northeast China, allowing us to resolve lateral variations of Lg Q greater than or equal to 2.0°. We gathered more than six thousand Lg path-amplitudes from more than hundred crustal earthquakes. Using the reverse two-station/event method, we obtained a two-dimensional tomographic image of Lg Q with its values ranging from about 50 to 1500. A high degree of detail in the lateral variation of Lg attenuation is revealed in our tomographic image than the previous studies in the area. We also observe strong frequency dependent in attenuation. Low attenuation regions are found in the Great Xing’an, Lesser Xing’an and Songen-Zhangguangcai Ranges. Islands of low attenuation appear to be associated with granitic batholiths in the Great Xing’an Range. Attenuation is high in the Songliao, Sanjiang, Erlian Basins and Bohai Basins. The highest attenuation is found in the vicinity of the Changbaishan, Wudalianchi volcanic field, and Quaternary volcanic regions, the southern Songliao Basin, west of Erlian Basin, Bohai Basin and the Sanjiang Basin. Highly attenuating regions correlate well with regions of high heat flow and volcanism, suggesting intrinsic attenuation as the main cause of low crustal Q. Low Q regions also occur within basins with thick, unconsolidated sediments. Some of the high Lg attenuation regions, outside of the Great Xing’an Range and Songliao Basin, correlate well with the uppermost mantle low shear wave velocity anomalies, suggesting that hotter uppermost mantle also contribute to crustal attenuation.
Session:A Decade of Great Subduction Earthquakes – What Have We Learned From Their Ground-Motions?
Presenter   Ghofrani, Hadi
Schedule   Fri 8:30 AM / Oral
Room   Room 4
Ground-Motion Prediction Equations for Large Interface Earthquakes Based on Empirical Data from the 2011 M9.0 Tohoku, Earthquake
GHOFRANI, H. G., Western University, London, ON, Canada, hghofra@uwo.ca; ATKINSON, G. M., Western University, London, ON, Canada, Gmatkinson@aol.com
Ground motion prediction equations have a major impact on seismic hazard estimates because they control the predicted amplitudes of ground shaking. The modeling and understanding of ground-motion amplitudes due to mega-thrust earthquakes in subduction zones has been hampered by a paucity of empirical ground-motion data for the very large magnitudes of most interest to hazard analysis. Data from the Tohoku M9 2011 earthquake are important in this regard, as this is the largest well-recorded subduction event, and the only such event with sufficient data to enable a clear separation of the overall source, path and site effects. In this study, we use strong-ground-motion records from the M9 Tohoku event to derive an event-specific GMPE. We then extend this M9 GMPE to represent the shaking from other M>7 interface events in Japan by adjusting the source term. We focus on events in Japan to reduce ambiguity that results when combining data in different regions having different source, path and site effect attributes. Source terms (the average residuals of ground-motions with respect to the Tohoku GMPE) scale most steeply with magnitude at the lower frequencies; this is in accord with expectations based on overall source-scaling concepts. Interpolating source terms over the magnitude range of 7 to 9, we produce a GMPE for large interface events of M7 to M9, for NEHRP B/C boundary site conditions in both fore-arc and back-arc regions of Japan. We show how these equations may be adjusted to account for the deeper soil profiles (for the same value of VS30) in western North America, and demonstrate the importance of considering the variability in site conditions when translating ground-motions from one region to another. For the Cascadia subduction zone, the proposed GMPE predicts lower motions at very long periods, higher motions at short periods, and similar motions at intermediate periods, relative to the simulation-based GMPE model of Atkinson and Macias (2009).
Session:New Directions in PSHA: Ins, Outs, and Uncertainty
Presenter   Yagoda-Biran, Gony
Schedule   Fri / Poster
Room   Cook/Arteaga
Between – Event Uncertainty for “Repeating Earthquakes”
YAGODA-BIRAN, G., Nevada Seismological Laboratory, University of Nevada, Reno, NV, USA, birangony@gmail.com; ANDERSON, J. G., Nevada Seismological Laboratory, University of Nevada, Reno, NV, USA, jga.seismo@gmail.com; MIYAKE, H., Earthquake Research Institute, University of Tokyo, Tokyo, Japan, hiroe@eri.u-tokyo.ac.jp; KOKETSU, K., Earthquake Research Institute, University of Tokyo, Tokyo, Japan, koketsu@eri.u-tokyo.ac.jp
Recent studies of strong ground motions have placed increased importance on understanding the uncertainties in ground motion predictions. Ground motion prediction equations (GMPEs) often separate the between-event (tau) from the within-event variability, but recent studies have focused on more precise distinctions, such as single-station sigma. The concept of a fault that ruptures identically in repeated events was called a characteristic ground motion earthquake by Anderson and Brune (1999). An extension of this idea to observations suggests that perhaps the smallest uncertainty that can be expected in ground motion recordings is to be expected when the same station records motion from rupture of the same fault “segment” in repeated events. We have searched for records from repeated rupture of the same fault in similar, large-sized earthquakes (M>5), and found several cases, two of strike-slip faults from the US and five of thrust faults from Japan, which in our judgment are similar enough that the differences in ground motion are relevant to estimating the uncertainty due to the variability of multiple ruptures of a single source. We compare the 5% damped spectral accelerations for these events at the recording stations, and calculate the maximum-likelihood standard deviation tauep within a pair. For period ranges where records from both earthquakes are reliable, we find that tauep averages between 0.1 and 0.2 natural log units. Based on extensive numbers of synthetic seismograms with random variability, we find that the actual value of tauep is larger by a factor of 1.76 compared to the maximum likelihood estimates using event pairs. This suggests that the variability from repeating earthquakes is 0.18-0.36, which is comparable or lower than tau estimated in GMPE studies.
Session:From the Earthquake Source to Damage of Buildings: Bridging the Gap between Seismology and Earthquake Engineering
Presenter   Oth, Adrien
Schedule   Fri 8:45 AM / Oral
Room   Room 1
Stress Drop Variations and Their Relevance for Ground Motion Prediction
OTH, A., European Center for Geodynamics and Seismology, Grand Duchy of Luxembourg, adrien.oth@ecgs.lu; BINDI, D., GFZ German Research Centre for Geosciences, Potsdam, Germany, bindi@gfz-potsdam.de
Stress drop is one of the key parameters characterizing the earthquake source process and therefore also a fundamental input parameter for ground motion prediction. However, at the same time, stress drop is difficult to accurately determine. Stress drop estimates are commonly derived from corner frequencies of earthquake source spectra, and these estimates carry large uncertainties because stress drop depends on corner frequency cubed. Large earthquake populations usually show stress drop variations as large as three orders of magnitude. Assuming such a large range of stress drop variations leads to very large ground motion variability solely as a result of the source process, notwithstanding any variability due to propagation and site effects. On the other hand, it has recently been shown that the between-event variability obtained from ground motion prediction equations (GMPEs) implies significantly smaller variability in earthquake stress drop than source spectral studies generally do. In order to get new insights into the potential reasons for these contrasting findings, we present a dataset encompassing earthquakes of a wide magnitude range and spatial coverage in Japan for which corner-frequency based stress drops have been determined. These stress drop estimates show systematic lateral variations. Accounting for these systematic variations reduces the stress drop variability on local scale by a factor of two to three as compared with the full earthquake population, which is closer to the stress drop variability estimates obtained from GMPEs. The between-event terms obtained from regression analysis of the same dataset used for the source spectral study will be discussed in this context.
Session:Advances in Seismic Imaging and Monitoring of Time-Dependent Variations: Civil Structures, Near-Surface, and Shallow Crustal Scales
Presenter   Parolai, Stefano
Schedule   Wed / Poster
Room   Cook/Arteaga
Shear-Wave Quality Factor Qs Profiling Using Seismic Noise Data from Microarrays
PAROLAI, S., German Research Centre for Geosciences GFZ, Potsdam, Germany, parolai@gfz-potsdam.de; BOXBERGER, T., German Research Centre for Geosciences GFZ, Potsdam, Germany; PILZ, M., German Research Centre for Geosciences GFZ, Potsdam, Germany; BINDI, D., German Research Centre for Geosciences GFZ, Potsdam, Germany
The assessment of the shear-wave velocity (Vs) and the quality factor (Qs) profiles below a site is necessary to characterize its site response. Recently, methods based on the analysis of seismic noise have proved to be very efficient for providing a sufficiently accurate estimation of the Vs versus depth at reasonable costs for engineering seismology purposes. In this study, first it is investigated if the same methods can also provide, with just a few additional calculation steps, realistic Qs versus depth estimates. A data set of seismic noise collected at the Tito test site in southern Italy by a microarray of seismological stations (with maximum aperture of few tens of meters) was used, and the obtained Qs results are compared with those estimated by independent geophysical investigations. It is shown that the values are consistent and that the seismic noise analysis has the potential to also provide a more comprehensive (Vs and Qs from an engineering seismology point of view) description of the shallow geological structure below a site at low cost. The technique is further applied to data collected by other microrrays in Turkey and Kyrgyzstan. An empirical relationship between the estimated Qs and Vs values is investigated.
Session:Recent Advances and Findings in Earthquake Geology and Paleoseismology
Presenter   Grant Ludwig, Lisa
Schedule   Fri 11:45 AM / Oral
Room   Room 3
Multi-Method Measurement of Offset from the 1857 Fort Tejon Earthquake along the San Andreas Fault in the Carrizo Plain at Channel Sieh31
AKCIZ, S. O., Department of Earth and Space Sciences, UCLA, Los Angeles, CA, USA, sakciz@gmail.com; GRANT LUDWIG, L., Program in Public Health, UC Irvine, Irvine, CA, USA, lgrant@uci.edu; HALFORD, D., Stanford University, STanford, CA, USA, dhalford@stanford.edu; MALIYANI, G. I., School of Earth and Space Exploration, Arizona State University, Tempe AZ USA, gayatri.marliyani@asu.edu; SALISBURY, J. B., School of Earth and Space Exploration, Arizona State University, Tempe AZ USA, jbsalisb@asu.edu; KLEBER, E. J., School of Earth and Space Exploration, Arizona State University, Tempe AZ USA, ekleber@asu.edu; ARROWSMITH, J. R., School of Earth and Space Exploration, Arizona State University, Tempe AZ USA, ramon.arrowsmith@asu.edu; RHODES, E., Department of Earth and Space Sciences, UCLA, Los Angeles, CA USA, erhodes@ess.ucla.edu; Capaldi, T., Department of Earth and Space Sciences, UCLA, Los Angeles, CA, USA
Recent work using computer-based analytic tools and LiDAR topography suggested that offset during the 1857 Fort Tejon earthquake along the Carrizo section of the San Andreas fault was 5-6 m, significantly lower than the previously reported 8-10 m using traditional on-the-ground measurements. We are reevaluating surface displacement of the great 1857 Fort Tejon earthquake using 3D excavation of channel Sieh31, which has been interpreted to have been offset ~6m and ~9m by researchers using different methods. Increasing availability of high resolution topography and imagery make remote measurements more common. Continued validation of remote and field-based measurements is essential. We selected channel Sieh31 as the best candidate to measure 1857 slip and investigate the potential for superposed tectonically offset older and deflected younger sections of a single channel. In preliminary 2013 work, 3 trenches exposed well-defined channels incised into stratified alluvial fan deposits. The excavations show that the original thalweg of the Sieh31 does not coincide with either the present day thalweg nor the median of the present-day channel margins, emphasizing the importance of measuring the offset with a future 3D excavation. Preliminary optically stimulated luminescence (OSL) dates from an earlier channel fill stratum indicate that incision of Sieh31 channel predates ~1750+-30 A.D. While additional data are needed for confirmation, if the channel records a single offset, its age implies that the penultimate earthquake occurred between ~1650 and ~1750 A.D. Thus, the incision of channel Sieh31 predates the great Fort Tejon earthquake and the channel age is ideal to constrain displacement caused by the 1857 earthquake at this location. Preliminary results suggest different geomorphic and geologic measurements of tectonic displacement, and emphasize the importance of validation of surface offset reconstructions from remotely interpreted digital topographic data.
Session:Great Earthquakes and Slip to the Trench (Seismological Society of Japan/Seismological Society of America Joint Session)
Presenter   Hardebeck, Jeanne
Schedule   Thu 4:00 PM / Oral
Room   Room 2
Complete Stress Drop in Shallow Great Earthquakes
HARDEBECK, J. L., US Geological Survey, Menlo Park, CA, USA, jhardebeck@usgs.gov
Recent great earthquakes with significant shallow slip, including the 2011 M9.0 Tohoku, 2010 M8.8 Maule, and 2004 M9.2 Sumatra earthquakes, produce coseismic and postseismic rotations of the principal stress axes that imply complete or near-complete stress drop (Hasegawa et al., EPS, 2011; Hardebeck, GRL, 2012). A systematic analysis of coseismic stress rotations due to recent M≥8 subduction earthquakes shows little stress rotation below 25 km depth. This implies that the conditions that allow near-complete stress drop are limited to shallower depths, likely related to lower overburden pressure or to material properties in the near-trench region. I first invert together all moment tensors of inter- and upper-plate events for stress orientation. Prior to each shallow mainshock, the maximum compressive stress axis is near horizontal, while immediately after the mainshock, both the maximum and minimum compressive stress axes plunge at ~45°. Dipping faults can be oriented for either reverse or normal faulting in this post-mainshock stress field, depending on dip, explaining the normal-faulting aftershocks observed after some mainshocks. The stress axes rotate back rapidly in the months following the Tohoku and Maule mainshocks, and in the southern part of the Sumatra rupture. A rapid postseismic rotation is possible because the near-complete stress drop leaves very little stress at the beginning of the postseismic reloading. The spatial distribution of stress orientations before and after each mainshock is then mapped on a finer length scale to relate the complete stress drop regime to overburden pressure, material properties, and/or other fault behaviors. I bin events in the Tohoku, Maule, and Sumatra regions by both space and time, and invert their moment tensors. I will present preliminary results of the finer-scale spatial-temporal stress evolution, and relate the distribution of near-complete stress drop to depth and other observables.
Session:Great Earthquakes and Slip to the Trench (Seismological Society of Japan/Seismological Society of America Joint Session)
Presenter   Polet, Jascha
Schedule   Thu / Poster
Room   Cook/Arteaga
A Joint Seismic and Geodetic Rupture Model for the 2012 Haida Gwaii Earthquake and the Location of Slip on the Interface Relative to the Queen Charlotte Fault
POLET, J., Cal Poly Pomona, Pasadena, CA, USA, jpolet@csupomona.edu; SIMONS, M., Seismological Laboratory, Caltech, Pasadena, CA, USA, simons@gps.caltech.edu; FIELDING, E. J., Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA, USA, eric.j.fielding@jpl.nasa.gov; SAMSONOV, S. V., Canada Centre for Remote Sensing, Natural Resources Canada, Ottawa, ON, Canada, sergey.samsonov@nrcan-rncan.gc.ca
The October 28th, 2012, Mw7.8 Haida Gwaii earthquake occurred off the coast of northern British Columbia, in the transition region between subduction of the Juan de Fuca plate beneath the North American plate and a transform fault between the Pacific plate and North American plate: the Queen Charlotte Fault System. In the immediate vicinity of the Haida Gwaii earthquake rupture, relative motion between the North American and Pacific plates is more transpressional in nature than to the North, where the Queen Charlotte Fault System has produced several large (M8+) strike-slip earthquakes. The mechanism of the October 28th, 2012, earthquake is indicative of Pacific plate underthrusting beneath Haida Gwaii, showing slightly oblique thrust faulting on a shallowly dipping plane with a strike parallel to the Queen Charlotte Fault. We will present the results from a joint kinematic rupture inversion using teleseismic body and surface wave data, and InSAR data. We incorporate two RADARSAT-2 differential interferograms calculated from Wide Swath beam images from left and right-looking operations on two ascending orbital tracks with time intervals of 8-10 months. Preliminary results of our inversions indicate a single primary slip region with a width of 100-120 km along trench, from 52.75°N to the South, and a maximum of 5 m of slip. We will conduct Bayesian inversions of the interferogram displacements to further analyze the resolution of the depth of significant slip and in particular whether the rupture is constrained to the shallower part of the subduction interface to the West of the Queen Charlotte Fault.
Session:Earthquake Physics and Interaction
Presenter   Barrett, Sarah
Schedule   Fri 3:45 PM / Oral
Room   Room 7/8
STUDENT
Temporal and Spatial Clustering of Intermediate-Depth Earthquakes: Evidence for a Cascading Effect
BARRETT, S. A., Stanford, Stanford, CA, USA, sabarrett@stanford.edu; PRIETO, G. A., Massachusetts Institute of Technology, Cambridge, MA, USA, gprieto@mit.edu
Intermediate-depth earthquakes are often observed in dense concentrations of seismicity called nests. While these clusters occur globally, the most concentrated of this activity is observed in the Bucaramanga nest in northern Colombia. We detect events within the Bucaramanga nest using a regional seismic network. Each day, there are approximately 18 detectable events, many with similar waveforms, and some with reversed polarity waveforms. Cluster analysis strengthens the observation of the reverse polarity families. We use a pair-wise correlation matrix of all available stations, with a minimum of three observations, normalized for the total number of observations. An average-linkage algorithm, divides the events clearly into two distinct groups. We use another agglomerative hierarchical method (single-linkage) to investigate events within these established groups. We relocate these events using double-differences to illuminate two distinct features within the nest. Each of the two earthquake families corresponds, roughly, to one of these roughly linear features. We examine the recurrence interval of these two groups with respect to their new locations and observe clustering of the groups in time and space - in some cases showing repeating events within a few 10s of meters each another. We also investigate the focal mechanism of the repeating events to assess if the rupture might occur along the same fault patch, suggesting a cascading effect, possibly associated with thermal shear runaway.
Session:Seismic Location and Processing Techniques
Presenter   Mejia, Hannah
Schedule   Fri / Poster
Room   Cook/Arteaga
STUDENT
Accurate Depth Determination of Earthquakes along the Kuril Islands Outer Rise from Waveform Modeling
MEJIA, H., California State Polytechnic University Pomona, Pomona, CA, USA, hjpotter@csupomona.edu; POLET, J., California State Polytechnic University Pomona, Pomona, CA, USA; THIO, H. K., URS Corporation, Los Angeles, CA, USA; PITARKA, A., Lawrence Livermore National Laboratory, Livermore, CA, USA
We will present our results of waveform modeling of teleseismic P-waves, based on an application of a 3D finite difference method, for the purpose of determining improved locations of earthquakes located in the outer rise area near the Kuril Islands. A series of large outer rise events occurred in the region of the Kuril Islands between 2006-2009. The depth and timing of these outer rise compressional and tensional earthquakes may provide important insight into the coupling of the subduction zone. Prior to a large interface event, tensional earthquakes generally occur at shallower depths, while compressional events occur at greater depths. Tensional events can also occur after an interface event at greater depths; whereas compressional events typically become non-existent after large interface events, reflecting the change in stress in the outer rise area. However, the hypocentral depths of outer rise earthquakes are usually poorly constrained, and often held fixed at 10-15km depth in centroid moment tensor inversions. Outer rise events also tend to produce complex waveforms with large azimuthal variation due to the complicated velocity structure close to the trench and the interaction of the depth phases with this structure, the ocean bottom and water layer. The numerous earthquakes in this outer rise sequence, located in close proximity of each other, but diverse in terms of source mechanism, depth and horizontal location relative to the trench, offer a unique opportunity in the study of the importance of the incorporation of 3D subduction zone structure on depth phases. We will show the results of a comparison of 1D and 3D synthetics with teleseismic P-wave data for tens of tensional and compressional earthquakes.
Session:Explosive Source Characterization
Presenter   Rodgers, Arthur
Schedule   Wed 5:30 PM / Oral
Room   Room 1
Improving Yield Estimation for Near-Surface Explosions Using Seismic And Overpressure Data
RODGERS, A. J., LLNL, Livemore, CA, USA, rodgers7@llnl.gov; BONNER, J. L., Weston Geophysical Corporation, Lufkin, TX, USA, bonner@westongeophysical.com; FORD, S. R., LLNL, Livemore, CA, USA, ford17@llnl.gov; TEMPLETON, D. C., LLNL, Livermore, CA, USA, templeton4@llnl.gov; RAMIREZ, A. L., LLNL, Livermore, CA, USA; DODGE, D., LLNL, Livermore, CA, USA, dodge1@llnl.gov
We have developed a methodology to estimate the yield of near-surface explosions using a joint analysis of seismic ground motion and atmospheric overpressure. Because of the partitioning of energy into the ground and air for explosions near the Earth's surface, best results are obtained from combined analysis of seismic and overpressure data solving for both yield (W) and emplacement height-of-burst or depth-of-burial (HOB/DOB). Amplitudes of seismic P-waves and air-blast positive phase impulse are measured and used in grid search and Markov Chain Monte Carlo (MCMC, stochastic) inversions. The MCMC method allows us to propagate measurement and model errors without the need to assume underlying statistical distributions and has other advantages for this problem. We have applied this method to data from recent chemical high-explosive experiments and historical nuclear tests. The method has been successful for estimating W with errors of less than 50% and often inferring the correct sign of the emplacement (above or below ground). However, there are notable areas for improvement, particularly the empirical models for the seismic amplitude dependence on W, HOB/DOB and range. We show that the emplacement lithology (material strength) has a strong impact on seismic amplitudes. We report differences between seismic signal amplitudes for dry alluvium and hard rock, including W and HOB/DOB effects. We used recently digitized historical waveform data from chemical and nuclear explosions and will report differences in seismic and air-blast excitation from these source types.
Session:Pillars of Simulation: Seismic Velocity and Material Models
Presenter   Song, Xin
Schedule   Wed 5:00 PM / Oral
Room   Room 4
STUDENT
Anelastic Attenuation and Elastic Scattering Models of the Los Angeles Region for Use in Earthquake Simulations
SONG, X., University of Southern California, Los Angeles, CA, USA, xinsong@usc.edu; JORDAN, , University of Southern California, Los Angeles, CA, USA, tjordan@usc.edu
The accuracy of earthquake simulations needed for physics-based seismic hazard analysis depends on good information about crustal structure. As the frequencies increase, seismic wave attenuation becomes more important. We compare attenuation models that have been recently used in the CyberShake hazard model (Graves et al., 2011) and other simulation studies for the Los Angeles region (Olsen et al., 2009; Taborda & Bielak, 2013) with constraints from local earthquake data out to 10 Hz, which include those from Hauksson & Shearer’s (2006) attenuation tomography as well as our own measurements. We show that the velocity-attenuation scaling relationship for shear waves employed by CyberShake (QS = 50VS) provides a good approximation to the average crustal structure at f = 0.3 Hz, but it does not capture the lateral variations in QS at shallow depths. Moreover, this frequency-independent model is inconsistent with the high QS values observed throughout most of the crust at f >1 Hz. The data indicate a frequency-dependent attenuation of the form QS ~ f γ, where 0.5 ≤ γ ≤ 0.8. Anomalously low QS factors are observed at very shallow depths, which can be explained by a combination of anelastic attenuation and elastic scattering. The scattering parameters are roughly consistent with small-scale, near-surface heterogeneities observed in well-logs and seismic reflection surveys in the Los Angeles basin. Observation of small-scale heterogeneities from well-log shows strong non-Gaussian distribution (A. Plesch & J. Shaw), so scattering effect on attenuation must be described by using a Non-Gaussian model. We summarize the observations in a new attenuation and scattering model for the CyberShake region that is laterally heterogeneous and frequency dependent.
Session:From the Earthquake Source to Damage of Buildings: Bridging the Gap between Seismology and Earthquake Engineering
Presenter   Baltay, Annemarie
Schedule   Fri 9:00 AM / Oral
Room   Room 1
Stress Drop Variability and Its Relationship to Variability in GMPEs
BALTAY, A. S., USGS Menlo Park, Menlo Park, CA, USA, abaltay@usgs.gov; HANKS, T. C., USGS Menlo Park, Menlo Park, CA, USA, thanks@usgs.gov
Earthquake stress drop is an important predictive parameter in ground motion estimation and should correspond to the short-period event residual from a ground motion prediction equation (GMPE). The overall variability of seismologically estimated stress drops, however, is much larger than tau, the standard deviation of the event residuals from the GMPEs for peak ground acceleration (PGA). Using the NGA-West2 data base and KikNet data from Japan, we investigate the source of this discrepancy, employing several metrics for stress drop: the traditional Brune stress drop, dependent on moment and corner frequency cubed; the aRMS-stress drop of Hanks [1979], dependent on the moment and the square root of the corner frequency; and the stress parameter from Boore [1983]. The stress parameter is akin to a PGA-determined stress drop, estimated from the peak ground acceleration (PGA) of an event and the earthquake moment, which can also be modeled through stochastic ground motion simulations. We find that the reduced variability seen in the PGA event terms from GMPEs is attributable to several factors, including excellent data quality and processing in the NGA database; reduced natural variability of PGA measurements as opposed to spectral determination; and error propagation when calculating a Brune stress drop, due to the cube of the corner frequency. In addition, the NGA-West2 2014 GMPEs (e.g., Abrahamson, Silva and Kamai, 2014) consider mainshocks and aftershocks separately, finding a decrease in PGA amplitudes by 0.3 ln units for Class 2 events, on-fault aftershocks, as compared to the Class 1, stand alone events. Using the same classification, we find a similarly smaller stress parameter for the Class 2 aftershock events by ~30%. A lower stress drop for on-fault aftershocks may imply that they are re-rupturing damaged zones or weak asperities. Incorporating this difference when considering the variability of stress drop and of GMPEs is important.
Session:Merging Paths: Earthquake Simulations and Engineering Applications
Presenter   Quinay, Pher Errol
Schedule   Thu 3:45 PM / Oral
Room   Room 3
Simulation of Structure Seismic Response Based on Analyzing Fault-structure System in Large Computing Environment
QUINAY, P. E. B., Niigata University RINHDR, Niigata, Japan, pequinay@eng.niigata-u.ac.jp; ICHIMURA, T., University of Tokyo Earthquake Research Institute, Tokyo, Japan, ichimura@eri.u-tokyo.ac.jp; HORI, M., University of Tokyo Earthquake Research Institute, Tokyo, Japan, hori@eri.u-tokyo.ac.jp
Many physics-based earthquake simulation tools are increasingly being employed for site-specific ground motion estimation - owing to the continuous advancement in computation power, improvements in characterization of underground crust structures and availability of fault information. In the engineering field, such tools can be used to improve earthquake damage prediction to critical structures or buildings in urban area. In this study we conducted a fault-structure system analysis (combined earthquake simulation and analysis of building structure seismic response): the geologic-length scale processes are solved in the first step, and the second step analyzes engineering-length scale processes using the results in the first step as input boundary condition. In order to realize the computation in the first step within the range of frequencies relevant to engineering applications, we developed a parallel earthquake simulator suited to multicore computers. This simulator consists of parallel hybrid-grid mesh generator and a solver based on implicit time integration scheme. To utilize the architecture of large computers such as the K computer in Japan, the solver is implemented with MPI/OpenMP parallelization. For the structure response analysis, we consider a small region which includes building structures and local soil. As a preliminary test, we considered a Mw 6.4 scenario earthquake near the city of Niigata in Japan. First, we generated the crust model in K computer. Then we created the force data based on modeling the fault rupture using multiple point sources with varying starting times. For the soft soil response analysis, a fine-meshed model of a region under the city was generated. The results were inputted as base excitation to a structure seismic response analysis tool, Integrated Earthquake Simulator. We compared the response of several buildings based on case examples highlighting the advantage of three-dimensional ground motion modeling.
Session:Emergence of Continuously Recording Very Large Array Capabilities in Seismology
Presenter   Wang, Weitao
Schedule   Wed / Poster
Room   Cook/Arteaga
Retrieval of Core Phases from Ambient Noise Cross-correlation Using a Dense Large-aperture Broadband Array in Southwest China
WANG, W. T., Institute of Geophysics ,China Earthquake Administration, Beijing, China, wangwt@cea-igp.ac.cn; TSAI, C. V., Seismological Laboratory,California Institute of Technology, Pasadena, CA, USA, tsai@gps.caltech.edu; WANG, B. S., Institute of Geophysics ,China Earthquake Administration, Beijing, China, wangbs@cea-igp.ac.cn
With the deployment of very large broadband seismic arrays, there has been an explosion in the use of continuous seismic records as well as new techniques applied to dense arrays. From early 2011 through the end of 2013, more than 350 broadband portable seismic stations were deployed in southwest China. These stations, combined with the existing 88 broadband permanent stations, composed a dense large-aperture array covering a ten by ten degree area with average station spacing of less than 50 km. We calculated daily cross correlation functions (NCFs) up to 3600s using the vertical component of those 435 broadband stations using two years of continuous records. After stacking all daily NCFs of each path into one reference NCF, a further spatial stack, referred to as array interferometry, was conducted by stacking NCFs with similar inter-station distances into one trace. After applying array interferometry, we observe strong signals that arrive after the fundamental surface wave. By analyzing the travel times of these signals, we clearly identify the core reflected ScS around 900s and the twice refracted PKIKP2 phase around 2400s. Moreover, these also exist strong signals from 3000s to 3600s, which could be PcPPKPPKP and PcPPKPPKS. The temporal variability of the amplitude of ScS and PKIKP2 show positive correlation with the global seismicity, implying those core phases may originate from the long lasting scattered coda waves of significant earthquakes. Our results, combined with previous results using USArray (Lin et al., 2013), show that core phases can be extracted wherever high-density arrays exist. The geometry of sampling of the earth's deep interior can therefore be improved using array interferometry, overcoming the limitation of earthquake-station geometries. Further work on better understanding the detailed waveforms of the spatial stacked NCFs are still necessary in order to use those signals to investigate deep earth structure.
Session:Development of Next Generation Field Methods for Portable Broadband Seismic Arrays
Presenter   Pazos, Celia
Schedule   Wed / Poster
Room   Cook/Arteaga
STUDENT
Analysis of Micro-Earthquakes in the Sierra Madre-Cucamonga Fault Zone Transition as Recorded by a Temporary Seismic Deployment
PAZOS, C. S., California State Polytechnic University Pomona, Pomona, CA, USA, cspazos@csupomona.edu; LEVARIO, J. J., California State Polytechnic University Pomona, Pomona, CA, USA, jjlevario@csupomona.edu; NGET, D., California State Polytechnic University Pomona, Pomona, CA, USA, dnget@csupomona.edu; POLET, J., California State Polytechnic University Pomona, Pomona, CA, USA, jpolet@csupomona.edu
The Sierra Madre-Cucamonga fault zone is an active, North-dipping thrust fault system located at the base of the San Gabriel Mountains. Near the mouth of San Antonio Canyon, the fault bends and transitions from the Sierra Madre Fault to the Cucamonga Fault. Coinciding with this bend is the active San Antonio-San Jose fault system. Where the Sierra Madre fault transitions into the Cucamonga fault near the mouth of San Antonio Canyon, it is poorly exposed. Major population centers, including Upland, Rancho Cucamonga, and Claremont, are susceptible to seismic hazards presented by this fault system. During the summer of 2013, we conducted a micro-seismic survey of the Sierra Madre-Cucamonga fault zone. Nine Guralp CMG-6TD seismometers were deployed in the foothills of the San Gabriel Mountains. At the end of the two-month deployment period, our network had detected over 150 local earthquakes. Included in this catalog are two M3.7 earthquakes that occurred in the Pomona area. This strike-slip faulting doublet is part of a sequence of tens of events that may connect to the Upland earthquakes along the San Jose fault, which is thought to continue through Cal Poly Pomona campus. To better capture this sequence we moved one of our seismometers to a location directly above the doublet. The waveform data recorded by our network, together with data from the Southern California Seismic Network, will be processed in order to (better) locate these earthquakes. We plan to use a combination of the HYPOINVERSE, VELEST and, eventually, hypoDD software packages to determine absolute and relative earthquake locations. These locations will yield a better understanding of the fault zone geometry and clarify the three-dimensional relationships of the local fault systems.
Session:Seismic Location and Processing Techniques
Presenter   Kurzon, Ittai
Schedule   Fri 8:45 AM / Oral
Room   Room 3
Real-time Automatic Detectors of P and S Waves Using Singular Value Decomposition
KURZON, I., Geological Survey of Israel, Jerusalem, Israel, ittaik@gsi.gov.il; VERNON, F. L., University of California, San Diego, La Jolla, CA, USA, flvernon@ucsd.edu; ROSENBERGER, A., Pacific Geoscience Center, Geological Survey of Canada, Sidney, BC, Canada, andreas@arescon.com; BEN-ZION, Y., University of Southern California, Los Angeles, CA, USA, benzion@usc.edu
We implement a new method for automatic detection of P and S phases using Singular Value Decomposition (SVD) analysis. The method is based on a real-time iteration algorithm of Rosenberger (2010) for the SVD of three component seismograms. The algorithm identifies the incidence angle by applying SVD and separates the waveforms into their P and S components. We apply the algorithm on filtered waveforms and then either set detectors on the incidence angle and singular values, or apply Signal-to-Noise Ratio (SNR) detectors for P and S picking on the filtered and SVD-separated channels. A recent deployment in the San Jacinto Fault Zone area provides a very dense seismic network for testing the detection algorithm in diverse setting, including events with different source mechanisms, stations with different site characteristics, and ray paths that diverge from the approximation used in the SVD algorithm. A Butterworth band-pass filter of 2-30Hz gives the best performance in a large variety of events and stations. We tested the SVD detectors on many events and present results from one of the most complex and intense aftershock sequence of the MW 5.2 - June 2005 event. This sequence was thoroughly reviewed by several analysts, identifying 294 events in the first hour, located in a dense cluster around the main shock. We used this dataset to fine-tune the automatic SVD detection, association and location, reaching a 37% automatic identification and location of events. All detected events fall within the dense cluster and there are no false events. An ordinary SNR detector does not exceed the 11% success and has a wider spread of locations (not within the reviewed cluster). One of the main advantages of the SVD detectors is the pre-knowledge of the phases picked (P or S), reducing significantly the noise created by ordinary “blind” SNR detectors.
Session:Induced Seismicity
Presenter   Braun, Thomas
Schedule   Thu 9:00 AM / Oral
Room   Room 1
Trying to Discriminate Triggered from Natural Seismicity: Three Case Studies from Northern Italy
BRAUN, T., INGV - Seismological Observatory, Arezzo, Italy, thomas.braun@ingv.it; CESCA, S., GFZ German Research Centre For Geosciences, Potsdam, Germany, simone.cesca@gfz-potsdam.de; DAHM, T., GFZ German Research Centre For Geosciences, Potsdam, Germany, torsten.dahm@gfz-potsdam.de; *) Institute of Earth and Environmental Sciences, University of Potsdam, Golm, Germany
Human operations near seismogenic areas rise always the question, whether they are involved in triggering local seismicity. In Central and Northern Italy anthropogenic activity like oil and gas extraction, waterlevel changes of impoundment basins and production of geothermal energy are recent examples. We present the analysis of microseismicity observed after the rupture of the fill crest of a huge barrier lake and close to the production area of CO2. A third case study treats the two damaging earthquakes with magnitudes Mw 6.0 and 5.8 that struck the Emilia-Romagna region in the sedimentary Po Plain, on May 20th and 29th 2012. The shallow hypocentral depth (6-10 km) led to severe damage and substantial impact to the local economy, and 24 fatalities were reported. The foredeep basin in Northern Italy is bordered by two active thrust and fold belts, the Apennines and Southern Alps, and convergence between them is accommodated by buried outer thrust fronts in the Po Plain. The Emilia 2012 earthquakes, which occurred on these buried faults, were therefore considered to be caused by tectonic processes. In the closer area to the Emilia events, at least one oil/gas field (Cavone) was exploited since 1980. Besides, small amounts of wastewater are reinjected through one of the boreholes. This raised recently concerns whether these earthquakes could have been influenced by reservoir depletion from former exploitation. We apply a novel probabilistic discrimination method to assess the probability that the earthquakes were human triggered or possibly induced. The depletion-induced stress perturbation is modeled by means of boundary element method adapted for the nuclei of strain approach. Human- and tectonic-induced stress rates are compared using a rate and state frictional nucleation model to assess relative earthquake probabilities, a method previously tested with depletion-triggered earthquakes in the North Sea and the Northern German basin.
Session:Deciphering the Earthquake and Tsunami History of Subduction Zones
Presenter   Kempf, Philipp
Schedule   Wed 2:30 PM / Oral
Room   Room 2
STUDENT
A Comparison of Tsunami Deposits from the 1960 Great Chilean Earthquake in Two Coastal Lakes on Chiloé, Chile.
KEMPF, P., Renard Centre of Marine Geology, Ghent University, Ghent, Belgium, philipp.kempf@ugent.be; MOERNAUT, J., Geological Institute, ETH Zurich, Zurich, Switzerland, jasper.moernaut@erdw.ethz.ch; VANDOORNE, W., Renard Centre of Marine Geology, Ghent University, Ghent, Belgium, willem.vandoorne@ugent.be; VAN DAELE, M., Renard Centre of Marine Geology, Ghent University, Ghent, Belgium, maarten.vandaele@ugent.be; PINO, M., Instituto de Geociencias, Universidad Austral de Chile, Valdivia, Valdivia, Chile, mpino@uach.cl; URRUTIA, R., Centro EULA, Universidad de Concepción, Concepción, Chile, rurrutia@udec.cl; DE BATIST, M., Renard Centre of Marine Geology, Ghent University, Ghent, Belgium, marc.debatist@ugent.be
On the 22nd of May, 1960, the Valdivia segment of the Peru-Chile megathrust ruptured in the strongest earthquake (EQ) ever instrumentally recorded (MW 9.5). This EQ produced a tsunami which struck approx. 1000 km of the near-field coast of South-Central Chile. The Island of Chiloé is located in the middle of the Valdivia segment and its west coast was inundated by 3 to 4 strong waves with a maximum wave height of approximately 15 m (Sievers et al., 1963). For two coastal lakes on Chiloé, namely Lago Cucao and Lago Huelde, the land barrier between the ocean and the lakes with 1.2 km width and 5-6 m elevation did not prevent the tsunami from inundating them. The lakes are only 2 km apart, lie behind the same beach and their outlets meet before flowing into the ocean. This setting gives a unique opportunity to compare the influence of onshore topography and lake bathymetry on inundation characteristics and on how the event is archived in the geological record. Side scan sonar data and pinger sub-bottom profiles were obtained and 36 gravity cores were analysed for geophysical properties, grain size and by x-ray computer tomography. In both lakes the cores contain 1960 tsunami deposits in form of sandy to silty beds, often with mud rip-up clasts. These deposits stand out in contrast to the otherwise organic rich clayey silt background sediment. Spatial trends in physical properties, sedimentary features, grain size percentiles and deposit thickness are compared for the two lakes and possible causes are discussed. The two lakes record the tsunami event from 1960, but do so differently from each other, because of their different topographic and bathymetric setting. Both lakes are suited to reconstruct the tsunami recurrence for the west coast of Chiloé.
Session:Citizen Seismology: Citizens Helping Science Helping Citizens
Presenter   Kato, Mamoru
Schedule   Fri / Poster
Room   Cook/Arteaga
Frequency Distribution of JMA Seismic Intensity and Ishimoto-Iida Law
KATO, M., GS Human and Environmemtal Studies, Kyoto University, Kyoto, Japan, mkato@gaia.h.kyoto-u.ac.jp
Through our own experiences, we are aware that large seismic shakings are rare events. Past records show that large intensities are less frequently observed than small intensities, and frequency distribution of observed intensities obeys an exponential law, which is known as the Ishimoto-Iida law [Ishimoto & Iida, 1939]. We investigate statistics of JMA Intensity data by means of the Ishimoto-Iida law. While it is always difficult to represent complex ground motion with one index, JMA Seismic Intensity has been widely accepted in the Japanese society and it is often utilized when seismologists communicate with the public and discuss hazard assessment and risk management. Following a long history of JMA Intensity based on human sensory, JMA Instrumental Intensity was introduced in 1996, and the number of seismic intensity observation sites has substantially increased and the spatial coverage has improved vastly, and as a result JMA Intensity is going to become a unique quantitative dataset for strong motions in Japan. The Ishimoto-Iida law explains the statistics of the observation well for the recent instrumental as well as for pre-instrumental intensity data, At approximately 100 stations with long recordings, exponents of the Ishimoto-Iida law are almost identical between pre-instrumental and instrumental intensities, and continuity of observations is confirmed from this particular point of view. The observed numbers of large intensity is slightly less than extrapolated and predicted from those of small intensity assuming the exponential relation. Large number of intensity observations are annually reported in Kanto area.
Session:Large and Damaging Earthquakes of 2013/2014
Presenter   Di Giacomo, Domenico
Schedule   Thu / Poster
Room   Cook/Arteaga
The ISC Event Bibliography: A Useful Service for Studying Notable Seismic Events
DI GIACOMO, D., International Seismological Centre, Thatcham, United Kingdom, domenico@isc.ac.uk; STORCHAK, D. A., International Seismological Centre, Thatcham, United Kingdom, dmitry@isc.ac.uk; SAFRONOVA, N., International Seismological Centre, Thatcham, United Kingdom, natalia@isc.ac.uk; OZGO, P., International Seismological Centre, Thatcham, United Kingdom; HARRIS, , International Seismological Centre, Thatcham, United Kingdom
The International Seismological Centre (ISC) is a not-for-profit organization in charge of the definitive worldwide summary of seismic events (ISC Bulletin). The ISC Bulletin contains parametric data both for natural and man-made seismic events starting from beginning of last century. We linked the information available in the ISC Bulletin to bibliographic records (mostly peer-reviewed journals). This allows us to offer a map-based web service that supports users in selecting references to scientific publications related to either specific events or events in the area of interest. Notably, some of the greatest earthquakes were described in several hundreds of articles published over a period of few years. The journals included in our database encompass a variety of fields in geosciences (e.g., engineering seismology, earthquake seismology, geodesy and remote sensing, tectonophysics, monitoring research, tsunami, geology, geochemistry, hydrogeology, atmospheric sciences, etc.), thus making this service useful in multidisciplinary studies. Currently the ISC Event Bibliography includes over 15,000 individual publications from about 500 titles related to events occurred in last 100+ years. Although many publications for seismic events occurred in 2013 are yet to come, we have already links to references for some notable event like the 20 of April Lushan and the 24 May Sea of Okhotsk earthquakes. In this contribution we show how the ISC Event Bibliography helps in gathering information about specific seismic events and its usefulness during paper preparation for authors (hence for reviewers and editors in the revision process). We also suggest that the ISC Event Bibliography can be a useful resource to monitoring agencies that provide information on the seismicity of a region soon after a significant earthquake has occurred.
Session:Development of 2014 U.S. National Seismic Hazard Maps and Their Implementation in Engineering Applications
Presenter   Wang, Zhenming
Schedule   Thu / Poster
Room   Cook/Arteaga
Alternative Approaches for the National Seismic Hazard Mapping
WANG, Z., Kentucky Geological Survey, Lexington, KY, USA, zmwang@uky.edu; COBB, J., Kentucky Geological Survey, Lexington, KY, USA, cobb@uky.edu
The purpose of the national seismic hazard maps is to provide scientific bases for the development of engineering design and other policy considerations such as the International Building Code and Residential Code in the United States. Therefore, the maps should be scientifically sound. Although a scientific database including locations of active faults and earthquakes, earthquake recurrent intervals, and ground motion attenuations, has been generated through the national seismic hazard mapping project, the national seismic hazard maps produced are not scientifically sound, however, because of the methodology used to produce the maps: probabilistic seismic hazard analysis (PSHA). As a complicated model, PSHA is not consistent with modern earthquake science and contains a mathematical error: equating a dimensionless quantity (the exceedance probability in one year, or the annual probability of exceedance) to a dimensional quantity (the annual frequency or rate of exceedance with the unit of per year). This mathematical error leads to the so-called ergodic assumption: treating spatial uncertainty of ground motions as an uncertainty over time at a single point. PSHA is scientifically flawed and its results (i.e., hazard curves and maps) are artifacts and difficult to understand. Thus, uses of the national seismic hazard maps are problematic. Alternative approaches are need in order to utilize the scientific database and produce seismic hazard maps that can be readily used to develop engineering design and other policies. One of such alternative approaches is scenario/deterministic seismic hazard analysis. The scenario/deterministic maps are scientifically sound and easy to understand. Furthermore, the scenario/deterministic maps have been used in coastal California and Kentucky.
Session:From the Earthquake Source to Damage of Buildings: Bridging the Gap between Seismology and Earthquake Engineering
Presenter   Wang, Zhenming
Schedule   Fri / Poster
Room   Cook/Arteaga
Ground Motion Hazard Maps for Engineering Design in the New Madrid Seismic Zone of the Central United States
WANG, Z., Kentucky Geological Survey, Lexington, KY, USA, zmwang@uky.edu
The New Madrid Seismic Zone is one of the most active seismic zones in the central United States and poses a certain level of seismic hazard and risk to the communities in the region. It is well known that the best way to mitigate seismic hazard or reduce seismic risk is better design and construction of buildings and infrastructures. However, selecting an appropriate ground motion for engineering design is challenging because it not only involves seismologists, but also engineers and other stakeholders. It is critical that seismologists not only produce ground motion hazard maps based on the best available science, but also clearly communicate the science to engineers and other stakeholders. It is also critical that engineers and other stakeholders understand the hazard maps correctly. Otherwise, the selected ground motion might not be appropriate for engineering design. For example, the ground motion hazard maps with 2 percent probability of exceedance (PE) in 50 years were selected for engineering design in the central United States. The maps are not consistent with the best available science and are difficult to understand. The use of these maps has caused problems in engineering design of buildings and infrastructure in the New Madrid region. Thus alternative ground motion hazard maps are needed. One such alternative map has been produced using a scenario or deterministic seismic hazard analysis. The scenario or deterministic ground motion hazard maps are scientifically sound and easy to understand.
Session:Network Operations and Data Centers
Presenter   Storchak, Dmitry
Schedule   Thu / Poster
Room   Cook/Arteaga
Datasets for Seismology: ISC, EHB and GT Bulletins, Station Registry, ISC-GEM Catalogue and Event Bibliography
STORCHAK, D. A., International Seismological Centre (ISC), Thatcham, Berkshire, United Kingdom, dmitry@isc.ac.uk; DI GIACOMO, D., International Seismological Centre (ISC), Thatcham, Berkshire, United Kingdom, domenico@isc.ac.uk; RICHARDSON, W. P., International Seismological Centre (ISC), Thatcham, Berkshire, United Kingdom, wayne@isc.ac.uk; HARRIS, J., International Seismological Centre (ISC), Thatcham, Berkshire, United Kingdom, james@isc.ac.uk
The International Seismological Centre (ISC) is a non-governmental non-profit-making organization funded by 62 research and operational institutions around the world, including those in the United States: the NSF, USGS, IRIS and the University of Texas at Austin. The ISC is charged with production of the ISC Bulletin – the definitive summary of the global seismicity based on reports from over 130 seismic networks worldwide. We also distribute the EHB bulletin – a groomed subset of the ISC Bulletin widely used in studies of inner structure of the Earth. Jointly with NEIC, the ISC runs the International Seismic Station Registry (IR). The ISC station codes and event hypocenters are used by IRIS DMC in serving seismic waveform data to its users. The ISC provides a number of additional products including the IASPEI Reference Event list (GT). The ISC-GEM Catalogue is another dataset that was designed to be used for global and regional studies of seismic hazard and risk. The ISC Event Bibliography is an interactive facility to search for scientific articles devoted to both natural and anthropogenic seismic events that occurred within a region and time period of interest. The multitude of ISC data is widely used by the CTBTO and associated community, including that in the United States, charged with monitoring nuclear tests worldwide. The ISC data are distributed via the ISC and several mirror websites such as those at IRIS DMC, LLNL, University of Tokyo and CTBTO. The ISC has a substantial development program that ensures that its data remain useful for geophysical research.
Session:Induced Seismicity
Presenter   Cesca, Simone
Schedule   Thu 11:30 AM / Oral
Room   Room 1
The 2013 Seismic Sequence Close to the Gas Injection Platform of the Castor Project, Offshore Spain
CESCA, S., GFZ German Research Center for Geosciences, Potsdam. Germany, simone.cesca@gfz-potsdam.de; GRIGOLI, F., University of Potsdam, Potsdam, Germany, francesco.grigoli@geo.uni-potsdam.de; HEIMANN, S., GFZ German Research Center for Geosciences, Potsdam, Germany, sebastian.heimann@gfz-potsdam.de; GONZALEZ, A., University of Zaragoza, Zaragoza, Spain, alvaro.gonzalez@unizar.es; MAGHSOUDI, S., University of Potsdam, Potsdam, Germany, maghsoudi@geo.uni-potsdam.de; BUFORN, E., University Complutense of Madrid, Madrid, Spain, ebufornp@fis.ucm.es; BLANCH, E., Ebro Observatory, Roquetes, Spain, eblanch@obsebre.es; DAHM, T., GFZ German Research Center for Geosciences, Potsdam, Germany, torsten.dahm@gfz-potsdam.de
A spatially localized seismic sequence has originated few tens of kilometres offshore the Mediterranean coast of Spain, starting on September 5, 2013, and lasting at least until October 2013. The sequence culminated in a maximal moment magnitude Mw 4.3 earthquake, on October 1, 2013. The epicentral region is located near the offshore platform of the Castor project, where gas is conducted through a pipeline from mainland and where it was recently injected in a depleted oil reservoir, at about 2 km depth. We analyse the temporal evolution of the seismic sequence and use full waveform techniques to derive absolute and relative locations, estimate depths and focal mechanisms for the largest events in the sequence (with magnitude mbLg larger than 3), and compare them to a previous event (April 8, 2012, mbLg 3.3) taking place in the same region prior to the gas injection. Moment tensor inversion results show that the overall seismicity in this sequence is characterized by oblique mechanisms with a normal fault component, with a 30° low-dip angle plane oriented NNE-SSW and a sub- vertical plane oriented NW-SE. The combined analysis of hypocentral location and focal mechanisms could indicate that the seismic sequence corresponds to rupture processes along sub- horizontal shallow surfaces, which could have been triggered by the gas injection in the reservoir. An alternative scenario includes the iterated triggering of a system of steep faults oriented NW-SE, which were identified by prior marine seismics investigations. The most relevant seismogenic feature in the area is the Fosa de Amposta fault system, which includes different strands mapped at different distances to the coast, with a general NE-SW orientation, roughly parallel to the coastline. No significant known historical seismicity has involved this fault system in the past. Both of our scenarios exclude its activation, as its known orientation is inconsistent with the focal mechanism results.
Session:Monitoring Dynamic Changes at Active Volcanoes and Fault Zones
Presenter   Smith, Cassandra
Schedule   Thu / Poster
Room   Cook/Arteaga
STUDENT
Explosion Earthquakes during the 2007 Eruption of Pavlof Volcano, Alaska
SMITH, C. M., School of Geosciences, University of South Florida, Tampa, FL, USA, smithcm09@gmail.com; MCNUTT, S. R., School of Geosciences, University of South Florida, Tampa, FL, USA, smcnutt@usf.edu
Pavlof Volcano on the Alaska Peninsula began to erupt on August 15, 2007 after a 10.7 year repose. Precursor signals consisted of low-frequency earthquakes that began on August 14 and thermal anomalies that were likely coincident with the beginning of the eruption. The mainly strombolian eruptions occurred from a new vent high on the SE flank of the volcano, separate from the NNE vent that had been active over the previous several decades. Seismic activity, monitored by a network of 6 local instruments, consists of low-frequency events, explosion earthquakes, volcanic tremor, and lahar-generated signals. Here we focus on explosion earthquakes. The first such event occurred August 14 at 17:58:16 UT and the last on September 13 at 15:14:00 UT. Events were often embedded in continuous tremor, but could be distinguished by the air waves, which appeared as a high-frequency spikes superimposed on the lower-frequency ground waves. The time differences in the airwaves at different stations are consistent with the acoustic speed of ~330 m/s. Rates were as high as 14 explosion quakes per minute. Steam and ash plumes were generally below 15,000 ft, but rose as high as 20,000 ft on August 29 and 30. We are working to establish whether systematic changes in explosion rates, signal properties, and air-to-ground-wave amplitude ratios correlate with the height and ash content of the plumes. Measurements are corrected for changes in wind speed and direction. AVHRR remote sensing data showed an ash signal on August 29-30, consistent with pilot reports. On August 30 lightning was observed in the plume from Cold Bay, 59 km SW. Compared with recent eruptions of Pavlof in 1981, 1986, 1996, etc., the eruptive activity was of shorter duration (~5 weeks instead of ~3 months) and was somewhat less explosive with no episodes of ash columns to heights of 30,000 ft or more. Instead the 2007 eruption produced significant lahars that reached the Pacific Ocean on the south side of the mountain.
Session:Induced Seismicity
Presenter   Friberg, Paul
Schedule   Thu 9:30 AM / Oral
Room   Room 1
Characterization of an Earthquake Sequence Triggered by Hydraulic Fracturing in Harrison County, Ohio
FRIBERG, P. A., ISTI, New Paltz, NY, USA, p.friberg@isti.com; BESANA-OSTMAN, G. M., Ohio Dept. Nat. Resources, Div. of Oil and Gas Resources Mgmt., Columbus, OH, USA, Glenda.Besana-Ostman@dnr.state.oh.us; DRICKER, I., ISTI, New Paltz, NY, USA, i.dricker@isti.com
On Oct 2, 2013, OhioSeis and the IRIS Transportable Array (TA) stations in Ohio detected a sequence of 6 earthquakes, two as large as Mw 2.0. These events, located in Harrison County Ohio, near Clendening Lake, were followed by 4 positive (Mw 1.7-2.2) magnitude events from Oct 3 to Oct 19. A total of 10 earthquakes were located in a region where no prior seismicity was observed. HypoDD was used with cross correlation to invert for double difference locations resulting in a linear E-W trending cluster located ~800m to the West of 3 wells undergoing hydraulic fracturing at the time. To find lower magnitude events, a master event cross correlation technique was used to detect similar signals on the closest TA station (O53A ~4km away). Using one Mw 2.0 event on Oct 2 as the template event, 298 matching signals were found (with X-corr coeff >0.85) as far back as Sept 8. The same technique was applied back to Dec 2012 but revealed no other matches. During this sequence, hydraulic fracturing was performed from Sept 7 to Oct 6 on 3 wells ≤ 4.0 km SW of O53A. Using the same technique, an additional 190 earthquakes were observed between Oct 2 and Dec 13, 2013. Moment magnitudes for all signal detections were determined using scaling to the template event and ranged from -0.77 to 2.2 with a G-R b-value for the earthquakes of 0.88±0.08. To improve event location, ODNR-DOGRM deployed a seismic array with 4 short-period stations in late Oct. Using HypoDD on the array data, a cluster of 30 seismic events that occur between Nov 2 and Dec 13, 2013 were located below the 3 wells and defined a vertical plane striking N92. A composite focal mechanism using 5 of the best-located events resolved an East-West trending vertical focal plane (striking N97), matching the earthquake distribution. Thus, based on temporal and spatial occurrence of the above microseismicity, the concurrent hydraulic fracture operation above it most probably activated a preexisting fault in the area.
Session:From the Earthquake Source to Damage of Buildings: Bridging the Gap between Seismology and Earthquake Engineering
Presenter   Lancieri, Maria
Schedule   Fri 11:45 AM / Oral
Room   Room 1
A Naive Bayesian Approach to Strong Ground Motionn Selection for Non-Linear Seismic Analysis of Structures
LANCIERI, M., IRSN, Fontenay aux Roses, France, maria.lancieri@irsn.fr; REANULT, M., IRSN, Fontenay aux Roses, France, marine.renault@irsn.fr; LECOMTE, L., IRSN, Fontenay aux Roses, France, louis.lecomte@irsn.fr; PERRAULT, M., UJF, Grenoble, France, perraulm@ujf-grenoble.fr; BAUMONT, D., IRSN, Fontenay aux Roses, France, david.baumont@irsn.fr; GUEGUEN, P., IFSTTAR, Grenoble, France, philippe.gueguen@ujf-grenoble.fr; BERGE-THIERRY, C., CEA, Saclay, France
In this work, we investigate the possible role of ground motion intensity measures (IM) in constraining data selection. To this aim, we propose a naïve Bayesian inference scheme to predict the response of a single degree of freedom (SDOF) system in function of a set of IM. We modeled 6 non-linear SDOF systems with initial fundamental frequency of 1, 2 and 5 Hz and with different strength reduction factors. The Takeda hysteretic model describing the structural behavior and the structural demand is expressed in terms of relative drift. A database of 6373 strong motion records was built from worldwide catalogues and has been described by a set of “classical” IM. The SDOF response is reduced in a description of three possible statuses: elastic, if the induced drift is lower than the yield displacement, plastic if the drift ranges between the yield and the ultimate drift values and fragile if the drift reaches the ultimate drift. Our goal is to evaluate the conditional probability of observing a given status of the SDOF system in function of the IM array. The results of the Bayes classification on the training dataset are quite promising. Indeed, to validate the presented methodology and evaluate its prediction capability, we performed a blind test on a second dataset, completely disjointed from the training one, composed of 7000 Japanese waveforms. We classified Japanese data using the probability distribution functions issued on the first data set. The results on Japanese dataset are encouraging; however the critical point is the discrimination between the waveforms pushing the SDOF system in plastic state and those inducing a fragile behaviour. To better understand what are the seismic features related with the plastic and fragile behaviour, we will perform a time-frequency analysis using the Stockwell transform. This latter part of the work is devoted to interpret the Takeda behaviour making a link between the engineering and seismological features.
Session:Explosive Source Characterization
Presenter   Kendall, Lauren
Schedule   Wed / Poster
Room   Cook/Arteaga
STUDENT
Seismic-Wave Gradiometry Applied to a Small-Scale Exploration Dataset
KENDALL, L., University of Memphis, Memphis, TN, USA, lmkndall@memphis.edu; LANGSTON, C. A., University of Memphis, Memphis, TN, USA, clangstn@memphis.edu
Seismic-wave gradiometry is an array-processing technique that utilizes spatial gradients between receivers to give information about wave timing and amplitude that can be applied to many geophysical problems. The concepts of 2D seismic-wave gradiometry are explored using a huddle test and three small-scale experiments with 24 vertical geophones and sledge hammer sources. A huddle test is performed to calibrate the geophones. Results show that relative gains cannot be completely characterized because of geophone/ground interaction. Synthetic tests show that 3% gain errors introduces errors of up to 3% into gradiometry-derived wave attributes. A standard linear refraction array is constructed to investigate the slowness and frequency content of hammer sources in order to understand induced and ambient wavefields recorded by two gradiometer designs. One design consists of four, six-instrument gradiometers in a linear array to investigate the spatial/temporal variation of horizontal slowness, wave azimuth, radiation pattern, and geometrical spreading for sources close to the array as well as test the location abilities of the gradiometer array. The gradiometers were able to correctly determine the slownesses of the P-waves and surface waves identified in the refraction profile. A second gradiometer experiment involves four superimposed cells to explore precision in calculation of spatial gradients. This analysis was done for combinations of the eight center stations for refracted P-wave and surface wave arrivals. Refracted P-wave slowness precision increased by 70% and azimuth precision increased by 66% when two center stations were used instead of one. For a surface wave arrival, the slowness precision increased 57% and the azimuth precision increased 99% when two center stations were used instead of one. These small-scale experiments show that off-the-shelf geophones may be used to build useful gradiometer arrays.
Session:Explosive Source Characterization
Presenter   Yang, Xiaoning
Schedule   Wed / Poster
Room   Cook/Arteaga
Time-Dependent Moment Tensors of the First Three Source Physics Experiments Explosions
YANG, X., Los Alamos National Laboratory, Los Alamos, NM, USA, xyang@lanl.gov; PATTON, H. J., Los Alamos National Laboratory, Los Alamos, NM, USA, patton@lanl.gov; ROWE, C. A., Los Alamos National Laboratory, Los Alamos, NM, USA, char@lanl.gov; LARMAT, C., Los Alamos National Laboratory, Los Alamos, NM, USA, carene@lanl.gov
Using mainly vertical, high frequency geophone data recorded within 2 km from the Source Physics Experiments (SPE) explosions, we conduct frequency-domain moment-tensor inversions to retrieve time-dependent moment tensors of the first three SPE explosions. Building on a velocity model developed using P-wave travel times and Rg phase velocities for one of the radial arrays (Line 2), we constructed an attenuation model by fitting synthetic P-wave and Rg amplitude decay to those of observed seismograms. The one-dimensional velocity model along with attenuation is used in the moment-tensor inversion to calculate Green’s functions. We select data from all radial arrays that have P-wave travel time, Rg phase velocity and amplitude decay consistent with those of the Line-2 data. The inverted moment tensors are dominated by diagonal components. The source spectra show long-period levels consistent with predictions from a Mueller-Murphy nuclear-explosion source model for granite, the rock type at the SPE site, with modified Denny-Johnson cavity-radius scaling. Among off-diagonal moment-tensor components, Mxy (horizontal double-couple) have the largest peak amplitude. We also attempt to resolve the damage source that accompanies the explosions with a modified inversion scheme. Using the new scheme, we invert for two sources: an explosion and a damage source at a shallower depth. By systematically changing the depth of the damage source and comparing the predicted seismograms with input seismograms in a grid-search approach, we map the depth extent of the material damage during the explosions. It appears that most of the damage occurs within the weathered layer near the free surface for SPE explosions.
Session:Tracking Fluid Movement in Volcanic Systems
Presenter   Adams, Aubreya
Schedule   Fri / Poster
Room   Cook/Arteaga
Seismic Structure of the Tonga Arc and Lau Backarc Spreading Center from Local and Teleseismic Body Wave Tomography
ADAMS, A. N., Washington University in St. Louis, St. Louis, MO, USA, aadams@seismo.wustl.edu; WIENS, D., Washington University in St. Louis, St. Louis, MO, USA, doug@wustl.edu; BLACKMAN, D., Scripps Institution of Oceanography, La Jolla, CA, USA, dblackman@ucsd.edu; WEBB, S., Lamont-Doherty Earth Observatory, Palisades, NY, USA, scw@ldeo.columbia.edu; DUNN, R., University of Hawai'i, Honolulu, HI, USA, dunnr@hawaii.edu; CONDER, J., Southern Illinois University, Carbondale, IL, USA, conder@geo.siu.edu; ZHAO, D., Tohoku University, Sendai, Japan, zhao@aob.gp.tohoku.ac.jp
The Lau Backarc Spreading Center and Tonga Arc provide an excellent environment in which to apply mantle seismic imaging to the study the dynamics of magma production and migration. The Central Lau Spreading Center, Eastern Lau Spreading Center, and the Valu Fa Spreading Center show very different characteristics as a function of their distance to the arc and slab, which decreases from north to south. Geochemical studies also indicate a systematic change in magma composition along strike of the Lau Back-Arc Spreading Center with more water and fluid-mobile elements in the south. From 2009-2010, 16 broadband seismometers and 51 ocean bottom seismometers (OBS) were installed for one year across the Lau backarc spreading centers and in Fiji and Tonga to image the along-strike structural variation of the subduction zone and the dynamics of the melt production region. OBS performance was excellent and all OBSs except one were recovered, with a ~ 90% data return. We present preliminary results from the joint inversion of P and S wave arrivals from local and teleseismic events as recorded by this temporary network. Arrivals from local events are picked manually, while travel time residuals for teleseismic arrivals are determined via the adaptive stacking method developed by Rawlinson and Kennett (2004). A preliminary set of travel times for local events and residuals for teleseismic events are inverted using TOMOG3D (Zhao et al., 1994) to solve for three dimensional velocity structure along strike of the Lau Backarc Spreading Center and Tonga Arc.
Session:Diverse Mechanisms of Subduction Zone Fault Slip: Exploring the Relationships Among Seismic Rupture, Transient Slip, and Steady Creep
Presenter   Liu, Yajing
Schedule   Wed 9:15 AM / Oral
Room   Room 1
Scaling Relations and Along-Strike Segmentation of Episodic Slow Slip Events in Subduction Zones
LIU, Y., McGill University, Montreal, QC, Canada, yajing.liu@mcgill.ca
A wide spectrum of slow slip phenomena in subduction zones, ranging from low-frequency earthquakes to long-term slow slip events, demonstrates a linear relation between the equivalent event moment and duration; each property spanning 7 to 8 orders of magnitude. However, the linear relation is not clearly established for each type of phenomena, especially for episodic SSEs in individual subduction zones. Understanding their moment-duration relation is key to the study of SSE source physics and relation to megathrust earthquakes. In this study I set up a 3D planar Cascadia-like fault model, using gabbro rate-state friction data and near-lithostatic pore pressure at depths around the friction stability transition, to produce SSEs with spatiotemporal distributions similar to those observed in Cascadia. SSE source properties, including along-strike propagation distance, event duration, equivalent moment and stress drop, are quantified and their scaling relations are compared to the observation summarized by Gao et al. [BSSA, 2012]. Modeled SSEs have nearly constant stress drops of 0.001 to 0.01 MPa, due to the small effective normal stress around 1 MPa required in the SSE region. SSE moment scales approximately to the square of its duration, rather than following a generalized linear relation. This is due to the stress perturbation among multiple SSEs originated at different along-strike locations and their slip fronts approaching each other during propagation. The modeled scaling relations are independent of the individual SSE selection criteria using either the slip velocity at a particular depth or the moment rate in the SSE region. Synthetic surface displacement time series are also computed due to the slip evolution on the fault. Modeling results indicate that geodetically detected along-strike segmentation in Cascadia SSEs may result from slower fault slip velocities at these segmentation boundaries, which may cause displacement anomalies beneath GPS detection threshold.
Session:Deciphering the Earthquake and Tsunami History of Subduction Zones
Presenter   Melnick, Daniel
Schedule   Wed 2:45 PM / Oral
Room   Room 2
Linking Subduction-Earthquake Supercycles with Mountain Building in the Region of the Giant 1960 Chile Earthquake
MELNICK, D., U. of Potsdam, Potsdam, Germany, melnick@geo.uni-potsdam.de; CISTERNAS, M., U. Catolica de Valparaiso, Valparaiso, Chile, marco.cisternas@ucv.cl; WESSON, R. L., USGS, Golden, CO, USA, rwesson@usgs.gov; MORENO, M., GFZ-Potsdam, Potsdam, Germany, marcos@gfz-potsdam.de; NELSON, A., USGS, Golden, CO, USA, anelson@usgs.gov; JARA-MUNOZ, J., U. of Potsdam, Potsdam, Germany, julius.jara@geo.uni-potsdam.de; DURA, T., U. of Pennsylvania, Philadelphia, PA, USA, dura@sas.upenn.edu; ENGELHART, S., U. of Rhode Island, Kingstone, RI, USA, Simon Engelhart, engelhart@mail.uri.edu
Ranges in magnitude and recurrence of subduction-zones earthquakes allows them to be grouped into earthquake cycles within supercycles. Placing prehistoric as well as historical ruptures within such cycles of differing scale is necessary to evaluate the seismic hazard of plate boundaries. Linking these cycles with patterns of permanent forearc uplift will leas to improved models of plate-boundary segmentation. We use coastal stratigraphy and soils on raised intertidal deposits at Isla Guafo--60 km landward of the south Chile trench--to infer land-level changes associated with great subduction earthquakes, and numerical models to infer the relations of the changes with plate-boundary slip. Guafo was uplifted ~4 m during the 1960 earthquake (M9.5) and has been subsiding rapidly since (20 mm/yr from 3 yr continuous GPS). Five soils, each developed on coseismically uplifted intertidal deposits , were buried during the past 1.7 ka. The sequence of soils and deposits suggests land-level changes similar to those observed during and after 1960. Such repeated changes can be modeled as elastic strain accumulation and release in plate-boundary zones A and B (Lay’s zonation). Distinct changes in coastal morphology and stratigraphy suggest Gaufo was permanently uplifted during an earthquake in AD 681-887 (2sd). This age range overlaps the time when a sequence of subsided buried soils at Maullín (140 km from the trench) shows an abrupt change in lithology suggesting uplift. If the change records permanent uplift at Maullin during the same earthquake, slip in zones A, B and C is required. Emerged Holocene beach ridges and even higher Pleistocene marine terraces may be evidence of the permanent uplift resulting from earthquakes of this type. Such giant events rupture down to the continental mantle and may have millennial recurrence times instead of centuries as for 1960-type events.
Session:Recent Advances in Ground Motions Simulation Methods and Their Validation
Presenter   Cahayani, Ulrika
Schedule   Thu 2:00 PM / Oral
Room   Room 3
STUDENT
The Effect of Topography on the Seismic Wavefield
CAHAYANI, U., University of Alaska Fairbanks, Fairbanks, AK, USA, ucahayani@alaska.edu; TAPE, C., University of Alaska Fairbanks, Fairbanks, AK, USA, carltape@gi.alaska.edu
We perform a series of seismic wavefield simulations in 3D structural models in order to quantify the effect of topography on seismic waveforms. Our procedure involves three components: (1) constructing a finite-element, unstructured hexahedral mesh, (2) performing seismic wavefield simulations, and (3) quantifying differences between seismograms. We consider three different structural models: homogeneous, layered, and realistic 3D variations. For each structural model we produce a mesh with and without topography. For each of these meshes we produce a finer version in order to demonstrate the minimum resolvable period for the coarser meshes. The target regions are southern California and Alaska, both of which exhibit large topographic variations and also have abundant seismic stations that allow for comparisons with recorded data. Our simulations shows that the topography has the strongest effects on surface waves, whereby a phase shift of the main arrival occurs due to the different thickness of the uppermost layer (which includes the topography or not). Scattered waves are visible in the coda due to surface wave reflections that occur off the direct source-station path. Our analysis emphasizes seismograms with periods >= 1 s, with the motivation to understand the topographic signatures that may influence measurements used within adjoint-based tomographic inversions with earthquake data.
Session:Seismic Imaging as USArray Moves to Alaska (IRIS/Seismological Society of America Joint Session)
Presenter   Doser, Diane
Schedule   Thu / Poster
Room   Cook/Arteaga
Waveform Modeling of Large Earthquakes of the Southern Fairbanks Seismic Zone, Alaska (1929-1947)
DOSER, D. I., University of Texas at El Paso, El Paso, TX, USA, doser@utep.edu
A series of 5 to 6 magnitude >6 earthquakes occurred within the Northern Foothills Fold and Thrust Belt (NFFTB) and southern Fairbanks Seismic Zone (FSZ) between 1929 and 1947. Fletcher and Christensen (1996) studied the source properties of the October 16, 1947 (M~7.1) FSZ mainshock using teleseismic body wave deconvolution techniques but were unable to obtain a well-constrained focal mechanism for the event. First motion studies of the October 16 mainshock (Wickens and Hodgson, 1967) suggest a northwest striking thrust or reverse fault, a trend that does not match the strike of the FSZ or NFFTB or the strike of features revealed by regional aeromagnetic and gravity data. A preliminary comparison of body waveforms of the January 21, 1929, July 4, 1929, October 7, 1947, and October 20, 1947events indicate they appear to have similar focal mechanisms and depths as the October 16 event. The October 20 and January 21 earthquakes also appear to have similar magnitudes. An event on October 15 appears to be of smaller magnitude since clear body waves are not observed at distances of ~40º. Some earthquake catalogs indicate the July 4, 1929 is larger in magnitude than the January 21, 1929 event; however waveforms of the July 4 event are about 50% smaller than the January 21 event. Earthquake relocations suggest the entire southern FSZ extending from the NFFTB to ~64.5ºN ruptured between 1929 and 1947 and that the largest events may have nucleated near east-west trending magnetic highs. Using body waveform data from additional seismograph stations I hope to refine source models for the October 16 mainshock as well as other large events of this sequence.
Session:Recent Advances in Ground Motions Simulation Methods and Their Validation
Presenter   Crempien, Jorge
Schedule   Thu / Poster
Room   Cook/Arteaga
STUDENT
Separating Low- and High-Frequency Wave Propagation in Synthetic Broadband Ground Motion Modeling
CREMPIEN, J. G. F., UCSB, ERI, Santa Barbara, CA, USA; ARCHULETA, R. J., UCSB, ERI, Santa Barbara, CA, USA
Ground motion prediction equations for spectral acceleration in the western US have geometrical attenuation (GA) that is ~1/R (Earthquake Spectra, 2008). However earthquakes simulated in 1D earth models have a larger GA for epicentral distances up to ~70 km (e.g., Burger et al., 1987), This difference is most likely caused by the 1D velocity structure’s failure to represent the propagation of high frequency (HF) waves. Consider a finite fault representing the source embedded in a nearly homogeneous region of a 1D model. For epicentral distances greater than the source depth, upgoing rays will have a large angle of incidence when they impinge on the near-surface layers. These rays will then glance off the shallow structure and thus transmit insufficient energy to the surface. To overcome this effect in 1D models, we have modified the UCSB method for computing broadband synthetics (Liu et al, 2006; Schmedes et al., 2013). We use different velocity structures for HF and low-frequency (LF) propagation. The source remains the same. Liu et al. (2006) includes stitching of synthetics computed using a 1D velocity structure with those from a 3D velocity structure. Following this approach we compute ground motions up to a certain frequency using the full 1D model. We modify the full 1D model by replacing the medium above the Moho with a single elastic layer. Using the same source, we compute full waveform synthetics with the modified 1D model. We approximate the local amplification expected from the full 1D model by applying the quarter wavelength method (Boore & Joyner, 1997). We use wavelets to stitch the HF ground motion, computed with the modified 1D model and local amplification, to the LF ground motion, computed with the full 1D model. With these modifications we recomputed a set of validations used by SCEC. In particular we are interested in eliminating the bias (in the goodness of fit) that changed with distance.
Session:Site Response: From Site-Specific Analyses to Predictive Models Around the Globe
Presenter   Keshvardoost, Ruhollah
Schedule   Wed / Poster
Room   Cook/Arteaga
STUDENT
Site Response in Western Washington from Specral Ratios
KESHVARDOOST, R., Auburn University, Auburn, AL, USA, ru@auburn.edu; WOLF, L. W., Auburn University, Auburn, AL, USA, wolflor@auburn.edu
We compare site amplification and resonant frequencies from 3-component broadband ground motion data, strong motion data and ambient noise for locations encompassing the Seattle and Tacoma, Washington, basins. Records were acquired from the 2012 M 6.1 Vancouver Island earthquake and its M 4.5 aftershock and from the 2012 M 7.8 Queen Charlotte Islands earthquake for the comparison. Seismic stations were chosen to cover different geologic units close to Seattle and Tacoma basins and other areas in west part of Washington State. In order to determine the site amplification, we used Fourier spectral ratios of recorded horizontal ground motions, referenced to a rock site. SSR plots show a considerable amplification in in frequencies above 2Hz. As observed by others, a shift in frequency exists between the mainshock and the aftershock, which possibly is due to a non-linear response of soil sites. However this shift is not consistent in all frequencies and at sediment sites. H/V results show resonant frequencies ranging from 2Hz to 15Hz. A comparison of H/V values for noise and earthquake data suggests that the peak frequencies at each site are relatively stable, except for one of the stations. Peak frequencies likely correlate with different depths to basement. There is a good agreement between the results from this study and previous studies in the area, suggesting that ambient noise studies may provide useful information about ground motions in sedimentary basins.
Session:Recent Advances in Ground Motions Simulation Methods and Their Validation
Presenter   Withers, Kyle
Schedule   Thu 11:45 AM / Oral
Room   Room 3
STUDENT
High-Complexity Deterministic Q(f) Simulation of the 1994 Northridge Mw 6.7 Earthquake
WITHERS, K. B., San Diego State University/ University of California San Diego, San Diego, CA, USA, quantumkylew@aol.com; OLSEN, K. B., San Diego State University, San Diego, CA, USA, kbolsen@mail.sdsu.edu; SHI, Z., San Diego State University, San Diego, CA, USA, zshi@mail.sdsu.edu; DAY, S., San Diego State University, San Diego, CA, USA, sday@mail.sdsu.edu
With the recent addition of realistic fault topography in 3D simulations of earthquake source models, ground motion can be deterministically generated more accurately up to higher frequencies. The synthetic ground motions have been shown to match the characteristics of real data, having a flat power spectrum up to some cutoff frequency (Shi and Day, 2013). However, the earthquake source is not the only source of complexity in the high-frequency ground motion; there are also scattering effects caused by small-scale velocity and density heterogeneities in the medium that can affect the ground motion intensity. Here, we dynamically model the source of the 1994 Mw 6.7 Northridge earthquake using the Support Operator Rupture Dynamics (SORD) code up to 8 Hz. Our fault model was input into a layered velocity structure characteristic of the southern California region characterized by self-similar roughness from scales of 100 m up to the total length of the fault. We extend the ground motion to further distances by converting the output from SORD to a kinematic source for the finite difference anelastic wave propagation code AWP-ODC. This code incorporates frequency-dependent attenuation via a power law above a reference frequency, and we illustrate the importance of including Q(f) in ground motion estimation. For example, a power-law exponent of 0.6 can produce up to a factor-of-two increase in the Arias Intensity compared to a constant-Q 0-2.5Hz simulation. We model the region surrounding the fault with and without small-scale medium complexity, with varying statistical parameters. Furthermore, we analyze the effect of varying both the power-law exponent of the attenuation relation and the scale factor relating Q and media velocity, and compare our synthetic ground motions with observed accelerograms. Additionally, we calculate the intra-event variability, i.e. the single-station standard deviation, and compare with values from leading Ground Motion Prediction Equations.
Session:Alaska Update of the USGS National Seismic Hazard Maps
Presenter   Porto, Natanya
Schedule   Thu / Poster
Room   Cook/Arteaga
Earthquake Risk in Alaska
NYST, M., Risk Management Solutions, Inc., Newark, CA, USA, mnyst@rms.com; WILLIAMS, C., Risk Management Solutions, Inc., ; FITZENZ, D., Risk Management Solutions, Inc.,
In anticipation of the upcoming update of the Alaska seismic hazard map by the USGS we investigate the possible impact of hazard changes on seismic risk. First order standard components that build up a probabilistic seismic risk model are hazard, building response, and a financial loss model. The hazard component in the current RMS Alaska earthquake risk model is based on the 2007 probabilistic seismic hazard maps for Alaska (Wesson et al., 2007; Boyd et al., 2007). Statewide risk based on the 2007 hazard maps is controlled by the subduction zone mega-thrust earthquakes. We implement newly gained insights into generation (segmentation), maximum magnitude and impact (ground motions) of mega-thrust earthquakes and their time-dependent behavior. We expand the event set with earthquakes representing multiple-fault ruptures (following a modified UCERF3 approach). Instead of modeling the intraslab earthquakes at a uniform depth we implement these deep events in a step-wise manner, similarly to what is proposed for Cascadia by NSHMP, thus producing shallower events closer to the trench and deeper events under the mainland. We identify areas of potential extreme liquefaction and landslide. The resulting seismic risk model has a less uniform source distribution, with larger contributions from crustal fault and intraslab sources. Considered risk metrics include simple scenario calculations, where we model regional damage and loss due to one earthquake, average annual loss, an annualized expected loss level used by insurers to set their annual premium rates, and the loss exceedance probability curve used by insurers to address their solvency and manage their portfolio risk. We analyze risk profile changes in areas with large population density and for structures of economic and financial importance: the Trans-Alaska pipeline, industrial facilities in Valdez and typical residential wood buildings in Anchorage, Fairbanks and Juneau
Session:Deciphering the Earthquake and Tsunami History of Subduction Zones
Presenter   Liberty, Lee
Schedule   Wed 4:45 PM / Oral
Room   Room 2
Same Asperity, Different Rupture Pattern: Variations in Holocene Splay Fault Motion from Seward to Middleton Island, Alaska
LIBERTY, L. M., Boise State University, Boise, ID, USA, lliberty@boisestate.edu; HAEUSSLER, P. J., U.S. Geological Survey, Anchorage, AK, USA; FINN, S. P., Boise State University, Boise, ID, USA
Using tsunami run up, seismic reflection and bathymetric data, we identify tsunamigenic sea floor ruptures beneath the Gulf of Alaska from the 1964 M9.2 earthquake. Coseismic sea floor ruptures are rooted in faults that appear across the 500-km wide Gulf of Alaska continental shelf. Seismic and geochronology data suggest that this asperity, located at the trailing edge of the subducted Yakutat terrane, has remained in place for thousands of earthquake cycles. However, we show evidence that these faults produce a range of uplifts during consecutive earthquakes. Based on estimated tsunami travel times from 1964, we identify thrust faults that produced 5-10 m wave heights in the coastal town of Seward and along the Kenai Peninsula. We identify the Hanning Bay, Patton Bay, Cape Cleare, and Middleton faults as local tsunami sources. These faults moved vertically upwards of 12 m during the 1964 earthquake, but paleoseismic data provide no evidence that other late Holocene earthquakes produced such uplifts on the adjacent Montague Island. Additionally, new seismic evidence suggests offsets across a prominent late Holocene(?) unconformity show significant variations in fault slip for the past few earthquakes. The Cape Cleare and Patton Bay faults, with (post-glacial) sea floor scarps as large as 69 m and 25 m respectively, likely accommodates the greatest vertical uplift within the asperity region. Another highly active thrust fault, termed the Middleton fault, hosts a west-east 100 km long >20 m high sea floor scarp. Uplift along this fault is likely responsible for 3.5 m raised shoreline on Middleton Island from the 1964 earthquake and the total sea floor scarp height likely includes uplift from at least the prior 3 earthquakes that individually produced 6-9 m marine terraces on Middleton Island. Although rapid exhumation results from repeated coseismic slip on multiple faults, slip distribution between earthquakes is highly variable.
Session:Recent Advances in Ground Motions Simulation Methods and Their Validation
Presenter   Assatourians, Karen
Schedule   Thu / Poster
Room   Cook/Arteaga
Validation of the Stochastic Finite Fault Module (EXSIM) Implemented on the SCEC Broadband Platform
ASSATOURIANS, K., Western University, London, ON, Canada, karenassatourians@yahoo.com; ATKINSON, G., Western University, London, ON, Canada, gmatkinson@aol.com
The stochastic finite fault model of Motazedian and Atkinson (2005) and Boore (2009), as implemented in the EXSIM program (www.seismotoolbox.ca) is among the Southern California Earthquake Center (SCEC) broadband earthquake ground motion simulation modules. We calibrate model input elements of EXSIM by comparing pseudo response spectral acceleration (PSA) of ten well-recorded earthquakes in California, Eastern North America, and Japan with the corresponding simulated PSAs. In addition, simulated PSA amplitudes are compared to those of ground-motion prediction equations (GMPEs) from the Next-Generation-Attenuation (NGA) project. EXSIM simulations’ average spectral amplitudes are in very good agreement with both event-specific recordings, and with the predictions from GMPEs, over a broad distance and period range, even though we use basic region-specific parameters with no attempt to fit each earthquake individually. The important input parameters are the stress parameter and the attenuation model; an adjustment to the Brune point-source spectrum is made to introduce a “double-corner” effect (Atkinson and Silva, 1997) that reduces bias at longer periods, particularly for events in California. We confirm previous conclusions (Motazedian and Atkinson, 2005) that average PSAs from EXSIM simulations don’t vary much from one random slip realization to another. The mean bias is correctly reproduced using 5 to 10 realizations for large events (M>6.5), while smaller events may need more realizations. The average bias computed from an individual earthquake (comparing EXSIM PSA to recorded PSA), within the 0.1s to 1.0s period band, can be used to infer an event-specific stress parameter. The event-specific stress parameters vary from ~100 to 300bars for study events in California and Japan, and ~300 to 900bars for ENA; however note that the very high stress for ENA is associated with steep near-source geometric spreading.
Session:Citizen Seismology: Citizens Helping Science Helping Citizens
Presenter   Atkinson, Gail
Schedule   Fri 1:45 PM / Oral
Room   Room 7/8
Reconciling Intensity Prediction Equations with Ground Motion Prediction Equations
ATKINSON, G. M., Western University, London, ON, Canada, gmatkinson@aol.com; WORDEN, C. B., USGS, Golden, CO, USA, cbworden@usgs.gov; WALD, D. J., USGS, Golden, CO, USA, wald@usgs.gov
Equations that predict intensity as a function of magnitude and distance are a useful tool for quantitative shaking hazard and loss assessment, and in the interpretation of both contemporary and historical earthquakes. The intensity prediction equations (IPEs) of Atkinson and Wald (2007) (AW07) have been very successful in describing the level and intensity of motions reported under the “Did You Feel It?” (DYFI) program. Examination of the performance of AW07 for North American earthquakes, evaluated using an extensive compiled database of DYFI observations, suggests that there is little statistical basis for revising these equations. However, a problem with the AW07 equations is that they scale too uniformly (i.e. no magnitude or distance saturation effects) and thus predict unrealistically-large median intensities for large events (M>6) at close distances. Intensity data are insufficient to redress this issue, as they are too sparse in the magnitude-distance range of concern. The AW07 equations can be improved by ‘closing the loop’ between IPEs and ground-motion prediction equations (GMPEs) to reveal more suitable intensity scaling at large magnitudes and close distances. It has been shown (AW07) that DYFI intensity tracks instrumental measures of ground motion such as peak ground velocity very closely; moreover, statistical studies (Worden et al., 2012; W12) have provided robust correlations that relate DYFI intensity to ground-motion measures and vice versa. We use the equations of W12 to compare GMPEs (in particular Boore, Stewart, Seyhan and Atkinson, 2014) to AW07’s IPEs. Recent GMPEs are constrained over a wide range of magnitudes and distances, enabling a more robust calibration of MMI predicted by the GMPE (via W12) to DYFI observations, over a broad overlapping data range (M3.5 to M6.5 are plentiful to 100km in both datasets). Based on this reconciliation of intensity and ground-motion predictions, we propose a revised version of our IPEs.
Session:Recent Advances and Findings in Earthquake Geology and Paleoseismology
Presenter   Rockwell, Thomas
Schedule   Fri 11:30 AM / Oral
Room   Room 3
10 m Slips on the Pitas Point Thrust: M8 Earthquakes in the Western Transverse Ranges of Southern California
ROCKWELL, T. K., San Diego State University, San Diego, CA, USA, trockwell@mail.sdsu.edu; CLARK, K., GNS, Lower Hutt, New Zealand, K.Clark@gns.cri.nz; OSKIN, M., UCD, Davis, CA, USA, oskin@geology.ucdavis.edu; GAMBLE, L., UCSB, San Barbara, CA, USA, gamble@anth.ucsb.edu; HAAKER, G., SDSU, San Diego, CA, USA, echaaker@gmail.com; UCARKUS, G., UCSD, La Jolla, CA, USA, gucarkus@ucsd.edu; KENT, G., UNR, Reno, NV, USA, gkent@unr.edu; DRISCOLL, N., UCSD, La Jolla, CA, USA, ndriscoll@ucsd.edu
Slip on the Pitas Point thrust produces uplift and folding of the Ventura Avenue anticline (VAA) in the western Transverse Ranges (WTR) of southern California. Rapid convergence has resulted in as much as 2.7 km of uplift on the VAA over the past 200-300 ka, with as much as 320 m in the past 45-50 ka, indicating a long-term uplift rate of 6-7 mm/yr. Using vintage aerial photography form the 1920’s through 1950’s prior to construction of the 101 freeway, along with differential GPS and LiDAR, we mapped remnants of four emergent Holocene marine terraces between Carpinteria and Ventura, the youngest three of which are reasonably well-preserved and dated between Punta Gorda and Pitas Point. The most recent emergence event uplifted a Chumash Indian village at Pitas Point by 5-6 m about 900 years ago; this 1st terrace is nearly continuously preserved to the fold crest reaching an altitude of 7-8 m, which implies 9-10 m of slip on the causative fault at depth. Nearly 50 radiocarbon dates on marine shells and culturally derived charcoal indicate terrace emergence at about 0.9 ka, 1.9 ka, and 4.2-4.7 ka, with the 4th (highest and oldest) terrace inferred to be slightly older than 6.5 ka. The Holocene uplift rate indicated by the emergent terraces is 5-6 mm/yr, similar to the long-term rate. Scaling relationships between fault slip, earthquake magnitude , and rupture length suggest that paleo-earthquakes were close to M8, similar to the 1999 Chi Chi or 2008 Wenchan earthquakes, and that ruptures likely included both onshore and offshore faults. Large uplifts in the offshore may have also resulted in local tsunamis along the Santa Barbara and Ventura coasts. High-resolution CHIRP seismic reflection data acquired across the VAA offshore Rincon Point in summer 2013 show evidence for discrete folding/uplift events that are captured in the Holocene sediment record and complement the onshore studies.
Session:New Directions in PSHA: Ins, Outs, and Uncertainty
Presenter   Munguia, Luis
Schedule   Fri / Poster
Room   Cook/Arteaga
Results of a Preliminary Earthquake Hazard Study in North Baja California, México.
MUNGUÍA, L., CICESE, Ensenada, BC, México, lmunguia@cicese.mx
In the last decades several earthquakes of magnitude larger than 6 occurred in the Baja California-Southern California region. Most recently, the El Mayor-Cucapah earthquake of April 4, 2010 (M 7.2) produced severe damage along the eastern part of north Baja California. Its strong motion effects were our main motivation to start a seismic hazard assessment for the region. Here, we present the first preliminary results of a hazard analysis that is presently underway. For the hazard computations, a computer code that is based on Matlab programming language was developed. Validation of such computer program was done through satisfactory reproductions of results of a benchmark PSHA example (Kramer, 1996) and of other published results. The preliminary results of runs of this program are shown in maps with contour lines of peak ground acceleration (PGA) and spectral acceleration (PSA) in 2% and 10% exceedance probabilities in 50 years. To get the hazard results, all major faults of the area were considered as seismic sources, and each source was modeled as a Poisson earthquake process, for which the magnitude - frequency relation was a truncated Gutenberg–Richter curve. The a- and b-values for each source were estimated from regional catalogs containing the seismicity of the last 35 years. On the other side, the ground motion levels for given magnitudes and source-to-site distances were obtained from the NGA attenuation model of Boore and Atkinson (2008) or from a model by Boore et al. (1997). In addition to the above hazard maps, we also show maps with results of hazard disaggregation computations. Displays of the contributions to hazard in terms of the geographic coordinates provide a better insight about the most probable scenarios for specific sites.
Session:Pillars of Simulation: Seismic Velocity and Material Models
Presenter   Yan, Qiyun
Schedule   Wed / Poster
Room   Cook/Arteaga
Sn Attenuation in Northeast China
YAN, Q., University of Missouri-Columbia, Columbia, MO, USA, qy4f2@mail.missouri.edu; SANDVOL, E., University of Missouri-Columbia, Columbia, MO, USA, sandvole@missouri.edu; KU, W., University of Missouri-Columbia, Columbia, MO, USA, wkb57@mail.missouri.edu
Northeast China is a generally active continental region which is controlled by Pacific plate subduction and India-Eurasia collision. In this study of uppermost mantle seismic wave attenuation, we use the regional high frequency seismic phase Sn to measure mantle lid shear wave Q. We have used data from the three-component broadband seismograms recorded by the Northeast China Extended Seismic Array (NECESSArray). We collected waveform data recorded by 127 stations from 140 earthquakes within a rectangular region from 30°N to 60°N in latitude, 100°E to 145°E in longitude. Models of Sn Q are calculated by applying both two-station method (TSM) and reverse two-station method (RTM). The RTM is more accurate and eliminates the site responses and crustal contribution to Sn attenuation. The site effect is then estimated form comparing the results from these two methods. In general, the inversion results show high Q values lie in Songliao Basin, indicating relatively low temperature for the uppermost mantle. We can roughly divide the Q0 values into four regions: low (Q0 <400), medium low (400 <Q0 <800), medium high (800 <Q0 <1200), and high (Q0 >1400). We observed low Q0 in Xilinhot block, Songnen block, eastern Zhangguangcailing block and Yanji block which are part of Songnen unit, medium low Q0 in Okhotsk belt, medium high Q0 in northern Songliao basin, and high Q0 in southern Songliao basin and western Zhanggaungcailing block. The basins are characterized by high Sn Q0 values. The borders between high Q0 and low Q0 are mostly along the active faults. Resolution test using checkerboard shows good resolution at scale of 2°×2° within most of the study area.
Session:Fault Structure, Heterogeneity, and Implications for Rupture Dynamics
Presenter   Yang, Wei
Schedule   Wed / Poster
Room   Cook/Arteaga
Velocity Contrast along Ganzi-Yushu Fault from Analysis of Fault Zone Head Waves Associated with Aftershocks of 2010 Mw6.9 Yushu, Qinghai Earthquake
YANG, W., Institute of Geophysics, CEA, Bejing, China, weiyang05@163.com; PENG, Z. G., Georgia Institute of Technology, Atlanta, GA, USA, zpeng@gatech.edu; WANG, B. S., Institute of Geophysics, CEA, Bejing, China, wangbs@cea-igp.ac.cn; LI, Z. F., Georgia Institute of Technology, Atlanta, GA, USA, zli354@gatech.edu; MENG, X. F., Georgia Institute of Technology, Atlanta, GA, USA, xmeng.gatech@gmail.com; YUAN, S. Y., Institute of Geophysics, CEA, Bejing, China, ysy1@vip.sina.com; QIAO, S., Institute of Geophysics, CEA, Bejing, China, qiaosen@cea-igp.ac.cn
The 2010/04/13 Mw6.9 Yushu earthquake ruptured the Yushu segment of the Ganzi-Yushu Fault (GYF) in eastern Tibet, which separates Bayan Har block and Qiangtang block. After the mainshock, we deployed temporary seismic stations around the mainshock rupture zone to monitor subsequent aftershocks, and 9 stations were deployed within 5 km on the both sides of the GYF. During the 7-month deployment period, up to ~2000 aftershocks have been recorded, providing a rich data set for better imaging of detailed fault zone structures in this region. In this study, we conduct a systematic search for fault zone head waves (FZHWs) that refract along the bimaterial fault interface. Since FZHWs spend most of their propagation paths along the fault interface with velocities on the faster side, they provide a high-resolution tool for detecting and imaging of the velocity contrast across the major crustal faults. We identify FZHW phase as low-amplitude, long-period seismic signals with polarities opposite to the direct P arrivals. In addition, we also use polarization analysis to confirm the existence of FZHW phases. Our preliminary results have shown that several stations on the southwest side (Qiangtang block) of the GYF have recorded clear FZHWs, suggesting that the Qiangtang block has lower velocities than the opposite Bayan Har block. The corresponding velocity contrast is about 5-10%, and is coherent along the majority of the rupture zone. This observation is generally compatible with recent studies based on seismic tomography and active refraction survey in this region. The sharp velocity contrast is expected to produce a preferred rupture direction to the SE, which is consistent with the overall propagation direction of the Yushu mainshock, if the rupture speed is subshear. Our next step is to use phase arrivals of FZHWs and direct P waves into a tomographic imaging to better understand deep fault zone properties, and their relationship to earthquake rupture properties.
Session:Cold Climate Installation Techniques and Instrumentation Developments for Temporary and Long-Term Networks
Presenter   Bruton, Christopher
Schedule   Wed 9:30 AM / Oral
Room   Room 7/8
Tracking Telemetry State of Health with Nagios
BRUTON, C. P., Alaska Earthquake Center, Fairbanks, AK, USA, cpbruton@alaska.edu
The Alaska Earthquake Center maintains over 150 real-time seismic stations in Alaska, supported by a vast telemetry network spanning thousands of miles and including more than 300 individual IP-addressable devices such as radios, routers, and modems. Because of the cost and difficulty of accessing most field locations, it is important to closely monitor communications equipment remotely. This makes it easier to diagnose station outages if they occur, and even provide advanced warning of potential telemetry problems. In November 2013 we established a monitoring system using Nagios, an open-source network monitoring software package. Our Nagios setup checks ping response times to IP-addressable devices on our network. In addition, it uses SNMP queries to obtain extra information provided by certain devices. In particular, we monitor voltage, temperature, signal-noise, and reflected power on most of the radios on our network. Tracking this information over time is critical in evaluating radio link performance, and will allow us to better focus our maintenance efforts during the short Alaskan field season.
Session:Explosive Source Characterization
Presenter   Steedman, David
Schedule   Wed 2:15 PM / Oral
Room   Room 1
Hydrodynamic Modeling of the Near-Source Environment at a Jointed Site
BRADLEY, C. R., Los Alamos National Laboratory, Los Alamos, NM, USA, cbradley@lanl.gov; STEEDMAN, D. S., Los Alamos National Laboratory, Los Alamos, NM, USA, dwsteed@lanl.gov; ROUGIER, E., Los Alamos National Laboratory, Los Alamos, NM, USA, erougier@lanl.gov; KNIGHT, E. E., Los Alamos National Laboratory, Los Alamos, NM, USA, knighte@lanl.gov
We perform near source hydrodynamic modeling of explosive events in granite for the Source Physics Experiment (SPE). The effort includes constitutive material behavior of the rock, the optimal means for representing the granite joint response, and the best modeling approach for including both the high-deformation source region and the complex material response in the near field. Each of these factors contributes to better estimates of explosion to seismic phenomena to help the verification community. A newly developed material model, AZ-Frac, has been implemented in two separate hydrodynamic modeling codes. The continuum model includes: 1) visco-plasticity, 2) directional fracture, 3) initial and fracture related anisotropy, and 4) stochastic material heterogeneity. This model has been validated against observed acceleration data for both free field and surface measurements from SPE. We also demonstrate the effects of joints. We know that compliant joints reduce both amplitudes and wavespeeds of shock. The scale effects of joints (both micro- and macro-scale) are modeled as “smeared” joints. This is compared to explicit inclusion of macro scale joint sets. Finally, by analysis of the coupling phenomena, we observe aspects of the source that require explicit modeling, to contrast some approaches that minimize modeling complexity through the use of simplification. Our models are performed using both the fully coupled Euler-Lagrangian modeling available in the Abaqus code and LANL developed hydrodynamic codes. This approach allows for explicit modeling of source characteristics in our model and cross validation between algorithms.
Session:Explosive Source Characterization
Presenter   Jones, Kyle
Schedule   Wed / Poster
Room   Cook/Arteaga
Comparing Modeled and Observed Infrasound Signals from the Source Physics Experiment Tests 2 and 3
JONES, K. R., Sandia National Laboratories, Albuquerque, NM, USA, krjones@sandia.gov; WHITAKER, R. W., Los Alamos National Laboratory, Los Alamos, NM, USA, rww@lanl.gov; ARROWSMITH, S. J., Los Alamos National Laboratory, Los Alamos, NM, USA, arrows@lanl.gov
The overall mission of the Source Physics Experiment is to improve upon and develop new physics based models for underground nuclear explosions using scaled, underground chemical explosions as proxies. Jones et. al, (AGU 2013) previously presented results describing the use of the Rayleigh integral (RI) to model the source region of the SPE explosions. These results showed that the source region could be easily and accurately modeled using the RI. We follow-on to this work by analyzing and modeling infrasound data collected during SPE-3. This test was designed to be as similar as possible to SPE-2 and although the tests were similar in yield, we observed a reduction in peak acceleration and pressure amplitudes. We hypothesize that this could be due to the “damage zone” from the previous test. This work was done under award number DE-AC52-06NA25946. 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:Seismic Location and Processing Techniques
Presenter   Glowacka, Ewa
Schedule   Fri / Poster
Room   Cook/Arteaga
Looking for Non Volcanic Tremor in Mexicali Valley, Baja California, Mexico
MÁRQUEZ RAMÍREZ, V. H., CICESE/ UNAM Geociencias, Ensenada, BC, Mexico/ Juriquilla, Mexico, vmarquez@cicese.edu.mx; GLOWACKA, E., CICESE, Ensenada, BC, Mexico, glowacka@cicese.mx; VIDAL-VILLEGAS, J. A., CICESE, Ensenada, BC, Mexico, vidalv@cicese.mx; WONG, V., CICESE, Ensenada, BC, Mexico, vwong@cicese.mx; DÍAZ DE COSSÍO BATANI, G., CICESE, Ensenada, BC, Mexico, gbatani@cicese.mx; RAMIREZ RAMOS, E., CICESE, Ensenada, BC, Mexico, erik_ramirez_1@hotmail.com; ARREGUI OJEDA, S. M., CICESE, Ensenada, BC, Mexico, sergio.arregui@gmail.com; GARCIA ARTHUR, M. A., CICESE, Ensenada, BC, Mexico, marthur@cicese.mx; FARFAN, F.J., CICESE, Ensenada, Mexico, fjfarfan@cicese.mx
The presence of Non Volcanic Tremors (NVT) was identified for the first time in the Japan subduction zones in 2002 (Obara, 2002). Because of its close association with geodetically observed slow-slip events and low frequency earthquakes in subduction zones, it was named Episodic Tremor and Slip (ETS). Nadeau and Dolenc (2005) detected tremor in Central California under the San Andreas Fault near Parkfield. NVT triggered by surface waves has been observed many times in subduction zones (Beroza and Ide, 2011). NVT has also been observed in the San Andreas and San Jacinto faults triggered by the 2002 Denali earthquake, and in Taiwan triggered by the 2001 Kunlun earthquake (Peng et al., 2008, Peng and Chao, 2008, Chao et al., 2012). A seismic sequence and deformation were triggered by the Hector Mine, M 7.1, 1999, earthquake (HME), and signals similar to NVT were observed at a broadband instrument installed in Mexicali Valley, Baja California, Mexico, 250 km from the epicenter (Glowacka et al., 2002). A continuous seismic signal was recorded in Mexicali Valley in one instrument only and only during a few months in1999, and 2002 – 2005. We analyzed tremor-like signals triggered by the HME and Denali (M=7.2, 2002) earthquakes using the Kao (Kao et al., 2007) and Husker methods, (Husker et al., 2010). Signal characteristics obtained from the analysis are similar to NVT as noticed by Kao, but the fact that the analysis was made only for one instrument does not allow to conclude that what was recorded was in fact tremor, and where the source of the tremor is. Also the presence of noise related to the fluid extraction and injection in the Cerro Prieto geothermal field and very noisy agricultural activity in Mexicali Valley do not help to identify seismic signals. A more dense seismic network is necessary in order to carry out complete seismic studies in the tectonic and anthropogenic conditions of Mexicali Valley.
Session:Cold Climate Installation Techniques and Instrumentation Developments for Temporary and Long-Term Networks
Presenter   Anthony, Robert
Schedule   Wed / Poster
Room   Cook/Arteaga
STUDENT
The Seismic Noise Environment of the Antarctic and the Influence of Instrumentation Techniques
ANTHONY, R. E., Colorado State University, Fort Collins, CO, USA, Robert.Anthony@colostate.edu; ASTER, R. C., Colorado State University, Fort Collins, CO, USA, Rick.Aster@colostate.edu; WIENS, D. A., Washington University, St. Louis, MO, USA, doug@wustl.edu; NYBLADE, A., Pennsylvania State University, University Park, PA, USA, nyblade@psu.edu; ANANDAKRISHNAN, S., Pennsylvania State University, University Park, PA, USA, sxa17@psu.edu; WINBERRY, J. P., Central Washington University, Ellensburg, WA, USA, winberry@geology.cwu.edu; HUERTA, A. D., Central Washington University, Ellensburg, WA, USA, huerta@geology.cwu.edu; WILSON, T. J., Ohio State University, Columbus, OH, USA, wilson.43@osu.edu
Recent instrumentation development has opened up even the most remote Polar Regions to year-round, continuous seismographic deployments. Early multi-year networks deployed at high latitudes have illuminated novel seismic signals and noise environments. This noise environment is influenced by, in some cases, new oceanic sources and other processes. However, a handful of sites have experienced unwanted instrumentation and site-induced noise generated in these harsh environments. We analyze recently collected, broadband data from temporary and permanent Antarctic stations to quantitatively assess background seismic noise levels across the continent between 2007-2012, including substantial previously unsampled sections of the continental interior. We characterize three-component noise levels between 0.15-150 s using probability distribution function-derived metrics of power spectral densities and analyze seismic noise levels in multiple frequency bands to examine different noise sources. These metrics reveal patterns of significant seasonal and geographic noise variations across the continent, including the strong effects of seasonal sea ice variation on the microseism, at a new level of resolution. Additionally, we examine the influence of different siting techniques (e.g., ice vaults vs. rock sites) on the seismic noise environment. We include comparisons of the collocated ice cap borehole (146 m and 255 m depth) and surface seismometer installations at South Pole (QSPA) to quantitatively assess noise reductions attained through borehole emplacement. Thorough analysis of the seismic noise environment and its relation to instrumentation and siting techniques in the Polar Regions facilitates optimization of deployments for future seismological deployments in the Arctic and Antarctic regions as well as mountain glacier environments.
Session:New Directions in PSHA: Ins, Outs, and Uncertainty
Presenter   Cramer, Chris
Schedule   Fri / Poster
Room   Cook/Arteaga
Memphis Urban Seismic Hazard Mapping Update: New Geology Plus Shear-Wave Velocity and Earthquake Observations
CRAMER, C. H., CERI, University of Memphis, Memphis, TN, USA, ccramer@memphis.edu; DHAR, M. S., CERI, University of Memphis, Memphis, TN, USA, msdhar@memphis.edu
In 2012-2013 the Memphis urban seismic hazard maps were updated with an improved and expanded geology model covering all of Shelby County, Tennessee. The new geology-only model predicts fairly uniform ground motions on the Shelby County uplands (PGA ~0.3 g for M7.7) but lower ground motions on the western lowlands (Mississippi floodplain). Differences in ground motions are mostly related to alluvium/loess thickness variations, which have an assigned shear-wave velocity (Vs) less than 200 m/s. Vs-profile observations across Shelby County (Romero and Rix, 2001) show Vs <200 m/s consistently to depths of only ~10 m, unlike the geology-only model’s of ~ 20 m in the NW uplands and greater than 30 m in the lowlands. A new geology plus Vs model shows more uniform hazard across the uplands and lowlands. We check on the reality of the new geology plus Vs based hazard estimates using two recent earthquakes: the 10/29/2012 M3.6 Parkin, AR west of Memphis and the 02/21/2012 M3.9 East Prairie, MO to the north. Scenario ground motion models for these two earthquakes show the greatest difference between the geology-only and geology-plus-Vs models in the lowlands and mainly for long period (1.0 s). A residual analysis at 9 ANSS stations across Shelby County for these two earthquakes favors the geology-plus-Vs model. This comparison helps validate the shallow Vs model over the shallow geology model. The ultimate validation will come from future stronger motion observations that demonstrate a similarity or difference in ground motion between the lowlands and uplands in Shelby County.
Session:Seismic Imaging as USArray Moves to Alaska (IRIS/Seismological Society of America Joint Session)
Presenter   Maceira, Monica
Schedule   Thu / Poster
Room   Cook/Arteaga
Application of Advanced Multivariate Inversion Techniques to the Western U.S.
MACEIRA, M., Los Alamos National Laboratory, Los Alamos, NM, USA, mmaceira@lanl.gov; AMMON, C. J., The Pennsylvania State University, State College, PA, USA, cja12@psu.edu; CHAI, C., The Pennsylvania State University, State College, PA, USA, cxc754@psu.edu; HERRMANN, R. B., Saint Louis University, Saint Louis, MO, USA, rbh@eas.slu.edu
Western USA is an ideal location to test advanced multivariate imaging techniques due to excellent data density provided by the USArray. We simultaneously invert surface-wave dispersion, receiver functions, and gravity anomalies to determine the structure of the crust and upper mantle of the western US. Specifically, our target region spans latitudes from 27 to 54 degrees North and longitudes from 90 to 130 degrees West. Rayleigh and Love surface-wave dispersion data between 5 and 150 seconds come from multiple filter analysis of regional earthquakes. Receiver functions come from the EarthScope Automated Receiver system, and are stacked to produce an average model for each cell in our parameterized region. The gravity observations are extracted from the global EGM2008 model. Our starting model is comprised of an oceanic PREM model west of the Pacific coast, a western US model between that and the eastern front of the Rocky Mountains, and a continental PREM model east of the Rocky Mountain Front. The addition of receiver functions produces sharper boundaries for the observed crustal anomalies. The addition of gravity produces subtle changes to the final model. Our inversion offers a nice compromise between fitting the three different datasets and simultaneously reducing RMS residuals for all. Our final results are consistent with numerous previous studies in the region. In general, the craton exhibits higher velocities than the tectonically active regions to its west. We see high mid-crustal velocities under the Snake River Plain and the Colorado Plateau. In the lower crust we observe lowest velocities in the western Basin and Range and under the Colorado Mineral Belt. At 80km depth we see broad low velocities fanning out from the Snake River Plain likely associated with the mantle plume feeding Yellowstone Caldera. Additionally we see high and low velocity anomalies along the west coast that reflect ongoing subduction processes beneath the western US.
Session:New Directions in PSHA: Ins, Outs, and Uncertainty
Presenter   Grant Ludwig, Lisa
Schedule   Fri / Poster
Room   Cook/Arteaga
Preliminary Precariously Balanced Rock (PBR) Age Dates Based on Various Models of Erosion
BRUNE, R. J., USC, Los Angeles, CA, USA, richbrune@sbcglobal.net; BRUNE, J. N., University of Nevada Reno, Reno, NV, USA, brune@seismo.unr.edu; GRANT LUDWIG, L., University of California Irvine, Irvine, CA, USA, lgrant@uci.edu
The fragility and age of precariously balanced rocks (PBRs) are critical in constraining PSHA results . We present preliminary estimates of PBR ages based on preliminary 10Be concentrations provided by Dylan Rood (Rood et al., SRL 2010; EOS 2012). We consider increasingly complex models for erosion: (1) Instantaneous exhumation followed by a number of years of exposure (CRONUS), (2) Constant erosion rate of a flat surface from far above the rock to the time of exposure of the rock pedestal, (3) the 4 parameter model (3 erosion rates and age of the lowest sample on the pedestal) of Balco et al. (Quat. Geochron., 2011), and (4) a model similar to that of Balco et al., but with a tilted erosional surface. We have made corrections for paleo-intensity, improved shielding (5 deg by 5 deg sampling, and the effective attenuation length). Similar to Balco et al., we ran tests to fit 10Be concentration profiles for 5 PBRs, using 4 to 8 samples per PBR. In all cases the resulting pedestal age is at least a few thousand years. The pedestal age is a proxy for the elapsed time since the rock became precariously balanced. As pointed out by Balco et al (2011) various parameters in their model (and in our new models) are uncertain and could lead to changes in pedestal ages, but in no case could we get age dates less than a few thousand years. Results are similar to those found by Bell et al., (Geology 1998) using 36Cl, and by Purvance et al., (SCEC 2009) using varnish microlamination dating (VML).
Session:Fault Structure, Heterogeneity, and Implications for Rupture Dynamics
Presenter   McLaskey, Gregory
Schedule   Wed 3:00 PM / Oral
Room   Room 3
Stick-slip Instability in Granite Initiated as Acoustic Emission Event
MCLASKEY, G. C., USGS, Menlo Park, CA, USA, gmclaskey@usgs.gov; LOCKNER, D. A., USGS, Menlo Park, CA, USA, dlockner@usgs.gov
Recent modeling studies have explored whether large earthquakes begin with a detectable nucleation phase, or if such a phase can be accelerated or skipped by the rupture of a smaller fragile fault patch in a “cascade-up” process. To explore such a condition in the laboratory, we study the initiation of dynamic rupture on a fault in a 76 mm diameter cylindrical granite laboratory sample. The simulated fault is a smooth saw cut surface inclined at 30 degrees to the sample axis and was deformed at confining pressures of 40-120 MPa. In a sequence of 10 stick-slip instabilities, we measure radiated seismic waves using piezoelectric sensors glued directly on the granite sample and located inside the pressure vessel. For each stick-slip, the radiation of seismic waves (>20 kHz) appears to initiate abruptly, and we can locate the hypocenter. In many cases, we find that the beginning is relatively weak, and that the first few microseconds of the seismic records are nearly indistinguishable from those of many other very small magnitude discrete seismic events, commonly known as acoustic emissions (AE), that occur in the seconds prior to each large stick-slip. In this sense, it appears that stick-slip begins as an AE event that rapidly (~20 microseconds) grows about two orders of magnitude in linear dimension and ruptures the entire 150 mm length of the simulated fault. Yet, in many cases, the AE events that appear to initiate stick-slip are neither exceptionally large nor in unusual locations compared to other AEs that occurred in the previous seconds and did not dynamically grow into stick-slip instabilities. We also find indications of accelerating fault slip in the final seconds before stick-slip which could reduce fault strength and therefore allow an AE to grow into a larger rupture.
Session:Fifty Years of Tsunami Science: from the 1964 Earthquake and Tsunami to the SAFRR Tsunami Scenario - Advances in Tsunami Source Characterization, Numerical Analysis and Hazard Mitigation
Presenter   Suleimani, Elena
Schedule   Fri 2:00 PM / Oral
Room   Room 4
Inundation Mapping and Hazard Assessment of Tectonic and Landslide Tsunamis in Southeast Alaska
SULEIMANI, E., Geophysical Institute, University of Alaska Fairbanks, Fairbanks, AK, USA, elena@gi.alaska.edu; NICOLSKY, D. J., Geophysical Institute, University of Alaska Fairbanks, Fairbanks, AK, USA, djnicolsky@alaska.edu; KOEHLER, R. D., Alaska Division of Geological and Geophysical Surveys, Fairbanks, AK, USA, richard.koehler@alaska.gov
The Alaska Earthquake Center conducts tsunami inundation mapping for coastal communities in Alaska, and is currently focused on the southeastern region. This activity provides local emergency officials with tsunami hazard assessment, planning, and mitigation tools. At-risk communities are distributed along several segments of the Alaska coastline, each having a unique seismic history and potential tsunami hazard. Thus, a critical component of our project is accurate identification and characterization of potential tectonic and landslide tsunami sources. The primary tectonic element of Southeast Alaska is the Fairweather - Queen Charlotte fault system, which has ruptured in 5 large strike-slip earthquakes in the past 100 years. The 1958 “Lituya Bay" earthquake triggered a large landslide into Lituya Bay that generated a 540-m-high wave. The M7.7 Haida Gwaii earthquake of October 28, 2012 occurred along the same fault, but was associated with dominantly vertical motion, generating a local tsunami. Communities in Southeast Alaska are also vulnerable to hazards related to locally generated waves, due to proximity of communities to landslide-prone fjords and frequent earthquakes. The primary mechanisms for local tsunami generation are failure of steep rock slopes due to relaxation of internal stresses after deglaciation, and failure of thick unconsolidated sediments accumulated on underwater delta fronts at river mouths. We numerically model potential tsunami waves and inundation extent that may result from future hypothetical far- and near-field earthquakes and landslides. We perform simulations for each source scenario using AEIC’s numerical model of tsunami propagation and runup, which is validated through a set of analytical benchmarks and tested against laboratory and field data. Results of numerical modeling combined with historical observations are compiled on inundation maps and used for site-specific tsunami hazard assessment by local emergency planners.
Session:Tracking Fluid Movement in Volcanic Systems
Presenter   Prejean, Stephanie
Schedule   Fri 8:30 AM / Oral
Room   Room 7/8
Re-Examining the Tectonic/Magmatic Controls on Dynamic Triggering in the Brittle Crust
PREJEAN, S. G., USGS, Anchorage, AK, USA; HILL, D. P., USGS, Menlo Park, CA, USA
Theoretical considerations suggest that transtensional environments may be more likely to experience dynamic earthquake triggering than transpressional environments. This conclusion assumes optimally oriented faults, Anderson faulting theory, and that Coulomb failure is at play. However in some locales, including volcanic regimes, the factors controlling rupture initiation may be more complicated. It has been suggested that volcanoes are unusually susceptible to triggering due to high heat flow, high pore-fluid pressures, and dynamic fluid migration. To probe this issue, we assess dynamic triggering and lack thereof at 24 seismically monitored Alaskan volcanoes. Since these volcanoes have a wide range of seismicity and stressing rates, both spatially and temporally, as a group they provide a comprehensive perspective on the likelihood of dynamic triggering in active volcanic regimes. The time period of our study (2006-2013) includes six eruptions at five volcanoes and several non-eruptive magmatic intrusions. We find that clear evidence of immediate dynamic triggering is lacking after 48 M7+ events in all but two cases: Pavlof Volcano following the 2011 M9.0 Tohoku-Oki earthquake and Martin Volcano following the 2007 M8.1 Kurile earthquake. In the Pavlof case the long-period nature of triggered events suggests that disruption of the hydrothermal or magmatic system may be the culprit process. Delayed dynamic triggering is more common at several additional volcanoes and geothermal systems, also suggesting an aseismic root to the triggering process. Overall, however, dynamic earthquake triggering at Alaskan (and reportedly Japanese) volcanoes is rare compared to some areas such as the western contiguous United States, even during times of intense volcanic unrest.
Session:Induced Seismicity
Presenter   Ellsworth, William
Schedule   Thu 9:15 AM / Oral
Room   Room 1
The Azle, Texas, Earthquake Sequence of 2013-2014: Induced or Natural?
ELLSWORTH, W. L., U. S. Geological Survey, Menlo Park, CA, USA, ellsworth@usgs.gov; BENZ, H. M., U. S. Geological Survey, Golden, CO, USA; CAIN, C., Department of Earth Sciences, Southern Methodist University, Dallas, TX, USA; DESHON, H. R., Department of Earth Sciences, Southern Methodist University, Dallas, TX, USA; HAYWARD, C., Department of Earth Sciences, Southern Methodist University, Dallas, TX, USA; LUETGERT, J. H., U. S. Geological Survey, Menlo Park, CA, USA; QUITORIANO, V., U. S. Geological Survey, Golden, CO, USA; STUMP, B., Department of Earth Sciences, Southern Methodist University, Dallas, TX, USA; Wald, D. J., U. S. Geological Survey, Golden, CO, USA
In early November, 2013 a sequence of locally felt earthquakes began in north-central Texas near the city of Azle, approximately 30 km northwest of Fort Worth. The sequence included eight with magnitudes ≥ 3 by the end of 2013. These are the first earthquakes to be instrumentally detected in the area since at least 1973. Epicenters determined by the NEIC span an area 25 km in dimension, although waveform similarity suggests a more compact source region. Geocoded intensity values from area residents reported to the “Did-You-Feel-It?” website concentrated in a limited area north of Azle. Five NetQuakes seismographs were deployed in mid-December to more accurately determine the location of the earthquakes. The improved locations pinpoint the epicentral region to lie about 8 km north of Azle. Focal depths range between 4 and 6 km, placing the earthquakes in the shallow basement. Based upon these results, we were able to revise the locations obtained for the earthquakes reported by NEIC that occurred before the local network was installed. Using differential travel times at stations located 100-300 km from the activity all of the previously scattered epicenters now fall in the same limited area as the locally recorded events. The sudden onset of the seismicity and its persistence in an area without previous activity raises the question of whether they are natural or induced. There are many producing gas wells within a few km of the earthquakes in this portion of the Barnett Shale Play. Most were drilled years before the activity began and there were no hydraulic fracturing operations in progress at the time the earthquakes occurred. The Texas Rail Road Commission data base lists several active disposal wells within several km of the earthquakes, including one that had injected 6 million m3 since 2009, located within 2-3 km of the epicenters. As of this writing it is not known if there is a connection between the waste water disposal and the earthquakes.
Session:Citizen Seismology: Citizens Helping Science Helping Citizens
Presenter   Cramer, Chris
Schedule   Fri / Poster
Room   Cook/Arteaga
Developing Ground Motion Estimates from M>6.0 Earthquake Intensity Observations for Use in ENA Emperical GMPEs
CRAMER, C. H., CERI, University of Memphis, Memphis, TN, USA, ccramer@memphis.edu
Instrumental ground motion observations for eastern North America (ENA) earthquakes are limited to M<6.0. Intensity observations for M>6.0 ENA earthquakes are available. I convert intensities to ground motion estimates for use as constraints in empirical ground motion prediction equation (GMPE) development. My intensity to ground motion conversion uses the relations of Dangkua and Cramer (2011) for ENA. For a given intensity level, the ground motion estimate is assigned to the mean log distance for that intensity and for a given earthquake. This methodology is validated using the 1988 M5.9 Saguenay, QC and 2011 M5.7 Mineral, VA earthquakes, which have both ground motion and intensity observations. Magnitudes of historical events are taken from Cramer and Boyd (2014) for M7 ENA earthquakes. Magnitudes of M6 historical earthquakes are confirmed using the mean intensity Monte Carlo approach of Cramer and Boyd (2014) with the 1925 M6.2 Charlevoix, QC and 1988 M5.9 Saguenay, QC earthquakes as reference events. Mean magnitude estimates for the 1843 Marked Tree, AR and 1870 Charlevoix, QC earthquakes are M5.9–6.0 and M6.0–6.1, respectively. 95% confidence limits on the magnitude estimates are ±0.3-0.4. The 1870 earthquake magnitude estimate is consistent with Ebel (2013). The intensity data for the 1895 Charleston, MO earthquake is too contaminated by site effects (Bakun et al., 2003) to be useful at this time. For use in GMPE development, only the M7 1811–1812 New Madrid and 1886 Charleston, SC earthquakes plus the M6.2 1925 Charlevoix, QC earthquake provide useful additional constraints over a distance range of 50 to 1200 km. The 1929 M7.2 Grand Banks earthquake only provides constraints at large distances of 1200 to 1400 km due to the offshore location of the event. Other historical events are in the magnitude range of existing ground motion observations or the 1925 event, and hence not as useful.
Session:Advances in Seismic Imaging and Monitoring of Time-Dependent Variations: Civil Structures, Near-Surface, and Shallow Crustal Scales
Presenter   Kalkan, Erol
Schedule   Wed / Poster
Room   Cook/Arteaga
Extracting Site Characteristics at Delaney Park Geotechnical Array in Anchorage, Alaska Using Seismic Interferometry
KALKAN, E., Earthquake Science Center, United States Geological Survey, Menlo Park, CA, USA, ekalkan@usgs.gov; ULUSOY, H. S., , Menlo Park, CA, USA; WANG, F., Institute of Geophysics, China Earthquake Administration, Beijing, China; FLETCHER, J. P., Earthquake Science Center, United States Geological Survey, Menlo Park, CA, USA
Seismic interferometry, based on deconvolution of wave motions recorded at various borehole depths, is used to extract the site characteristics including wave attenuation, predominant frequencies, shear modulus and shear-wave velocity profile of a densely instrumented geotechnical array at Delaney Park in Anchorage, Alaska. Surface and six borehole (up to 61 m depth) three-component sensors recorded ten earthquakes having moment magnitude between 4.5 and 5.4 at epicentral distances of 11 to 162 km. Deconvolution of the waveforms recorded at various depths on horizontal sensors with respect to the corresponding waveform recorded at the surface provides upward (incident) and downward (reflected) travelling waves within the soil layers. The simplicity and similarity of the deconvolved waveforms from different earthquakes suggest that a one-dimensional shear velocity model is sufficiently accurate to represent the site characteristics. The travel times evaluated by deconvolution agree with the results obtained by the geophysical measurements of shear-wave velocity at the site. It is also shown that the deconvolution method is effective in computing amplification property of shallow soil layers. The estimated parameters are used for soil-structure interaction modeling of a densely instrumented 20-story steel framed Atwood building, located 166 m from the Delaney Park geotechnical array.
Session:Explosive Source Characterization
Presenter   Ezzedine, Souheil
Schedule   Wed 2:45 PM / Oral
Room   Room 1
Three Dimensional Simulation of Ground Motions Generated by Underground Explosions in Jointed Rock under Conditions of Uncertainty
EZZEDINE, S. M., Lawrence Livermore National Lab, Livremore, CA, USA, ezzedine1@llnl.gov; VOROBIEV, O. Y., Lawrence Livermore National Lab, Livremore, CA, USA, vorobiev1@llnl.gov; GLENN, L. A., Lawrence Livermore National Lab, Livremore, CA, USA, glenn5@llnl.gov; ANTOUN, T. H., Lawrence Livermore National Lab, Livremore, CA, USA, antoun1@llnl.gov
We have performed 3D high resolution simulations of underground explosions conducted recently in jointed rock outcrop as part of the Source Physics Experiment (SPE). The main goal is to understand the nature of the shear motions recorded in the near field under condition of uncertainties in a) the geological characterization of the joints, b) the spatial variability of the geomechanical material properties, and c) the spatial correlation of rock properties. The approach is stochastic. The joints are depicted using a Boolean stochastic representation conditional to observations and their probability density functions. Then, using a novel continuum approach, joints and faults are painted into the continuum host rock. Simulating wave propagation in discontinuous rock mass is highly non-linear problem and uncertainty propagation via intrusive methods is practically forbidden. Therefore, using nested Monte Carlo simulations, we have explored and propagated several uncertainty parameters. We have probabilistically shown that significant shear motions can be generated by sliding on the joints caused by spherical wave propagation. Polarity of the shear motion may change during unloading when the stress state may favor joint sliding on a different joint set. A sensitivity analysis of the response of the system to ranges of geomechanical input parameters has been conducted to evaluate the impact of the key geomechanical parameters on the generated ground motions. To understand the far field seismic signatures associated with near-field shear waves, we have abstracted the near field behavior into a probabilistic source-zone model that can be used in the far field wave propagation study. This work performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.
Session:Network Operations and Data Centers
Presenter   Yuan, Songyong
Schedule   Thu / Poster
Room   Cook/Arteaga
Field Observation and Quality Control of China Array Project
YUAN, S. Y., Institute of Geophysics, CEA, Beijing, China, ysy1@vip.sina.com; YANG, W., Institute of Geophysics, CEA, Beijing, China, weiyang05@163.com; SONG, L. L., Institute of Geophysics, CEA, Beijing, China, sllcea@163.com; WANG, B. S., Institute of Geophysics, CEA, Beijing, China, wangbs@cea-igp.ac.cn; XU, W. W., Institute of Geophysics, CEA, Beijing, China, wwxuww@sina.com; LIU, Z. Y., Institute of Geophysics, CEA, Beijing, China, liuzhengyi01@163.com; JIA, Y. H., Institute of Geophysics, CEA, Beijing, China; QIAO, S., Institute of Geophysics, CEA, Beijing, China, qiaosen@cea-igp.ac.cn
Temporally deployed dense seismic arrays have been proven to be powerful tools to explore deep structure. A lot of array observations have been carried out, among them US Array is the most successful. Recently, China launched its seismic observation project – China Array Project to exploring the deep structure of China mainland. This project is hopefully boosting seismic research and hazard mitigation in China. China Array is divided into 7 phases, and will be finish within 15 years. The first phase has finished in 2013. During the first phase 500 broadband seismic stations have been deployed in South-west China with average inter-station distance of 35 km. In the ongoing second phase of China Array, more than 600 stations will be deployed in North-west China. The observations were carried out in various areas as mountanious regions, Gobi, desert and grassland. Some stations suffer from serious temperature variation, thick sediments, frozen, and strong background noises. To meet the unified data standard, we designed various emplacement strategies according to different site status. In this abstract, we reported the field emplacement, data management, and quality control of China Array, one of the largest portable seismic array observations.
Session:New Directions in PSHA: Ins, Outs, and Uncertainty
Presenter   Cramer, Chris
Schedule   Fri / Poster
Room   Cook/Arteaga
Magnitude Dependent Site Amplification Seismic Hazard Calculation Outside the Hazard Integral for St. Louis, MO
CRAMER, C. H., CERI, University of Memphis, Memphis, TN, USA, ccramer@memphis.edu
Richard Lee (2000) developed an approach to site-specific seismic hazard calculation for magnitude-independent site-amplification that applies the site amplification distribution outside the hazard integral. Cramer (2009b) implemented this approach for urban seismic hazard mapping in St. Louis, MO-IL. However, the USGS National Seismic Hazard Mapping Project deaggregation for St. Louis shows that the predominant hazard is from M7s about 200 km from the study area and M5s and M6s within 50 km of the study area. Cramer (2009a) developed site amplification distributions corresponding to these three levels of magnitude and corresponding distance and showed that the M7 at 200 km case is different from the M5 and M6 <50 km cases. Hence site amplification is magnitude dependent in the St. Louis area. Richard Lee (2010, oral communication) suggested that the magnitude-dependent site-amplification problem could be divided into a series of magnitude-independent site-amplification calculations over limited magnitude ranges. The hazard curves from the series of outside the hazard integral calculations for different magnitude ranges could then be combined for the total hazard curve at a site or grid-point. This magnitude dependent site amplification approach has been implemented, tested, and applied to generating St. Louis urban seismic hazard maps. The outside the hazard integral approach for applying site amplification distributions is more computationally efficient (5-10 times faster) and separates the tasks of hazard and site amplification calculation so a change in one does not force the recalculation of the other as the inside-the-hazard-integral approach requires (Cramer, 2003, 2005).
Session:Development of 2014 U.S. National Seismic Hazard Maps and Their Implementation in Engineering Applications
Presenter   Murray, Jessica
Schedule   Thu / Poster
Room   Cook/Arteaga
Implications of Recent Findings from Bayesian Inversion of GPS Data for Seismic Hazard Assessment in California
MURRAY, J. R., US Geological Survey, Menlo Park, CA, USA, jrmurray@usgs.gov; MINSON, S. E., California Institute of Technology, Pasadena, CA, USA, minson@gps.caltech.edu
Geodetic data illuminate aspects of crustal deformation relevant for seismic hazard assessment including fault slip rates, the rate and spatial extent of creep, the relationship between regions of aseismic and coseismic slip, and the patterns of moment deficit recovery over multiple earthquake cycles. We investigate these issues through Bayesian inversion of Global Positioning System (GPS) data from northern and central California. Bayesian inversion allows incorporation of a variety of prior information (e.g., nonnegativity), provides the full probability density function for model parameters, and does not require subjective regularization such as spatial smoothing. It is well suited to characterizing uncertainties on parameters like slip rates and creep rates. Realistic uncertainties are required as direct input to hazard assessment through deformation models, for evaluating those models’ underlying assumptions, and in follow-on calculations used in time-dependent hazard forecasts. The following results from our analyses are directly relevant to future California seismic hazard assessments. 1) Using a newly expanded GPS dataset we estimate slip rates for the northern San Andreas, Maacama, and Bartlett Springs faults of 21.5 +/- 0.5, 13.1 +/- 0.8, and 7.5 +/- 0.7 mm/yr, respectively. 2) We infer that creep on the Bartlett Springs seismogenic zone keeps pace with slip on the fault’s deeper extension, while Maacama fault creep occurs primarily above 5 km resulting in a moment deficit on the seismogenic zone. 3) The depth extent of creep on strike slip faults, even for a high creep rate, well instrumented fault like the San Andreas at Parkfield, is poorly constrained by GPS data; lack of resolution must be considered in hazard assessment. 4) Spatial patterns of interseismic creep, coseismic slip, and afterslip at Parkfield, in the context of the San Andreas long-term slip rate, indicate persistent moment deficit even in areas that slip during M6 earthquakes.
Session:75 Years of Frequency-Size-Distribution of Earthquakes: Observations, Models and Understanding
Presenter   Page, Morgan
Schedule   Fri 11:45 AM / Oral
Room   Room 4
Local Magnitude Distributions in the 3rd Uniform California Earthquake Rupture Forecast (UCERF3)
PAGE, M. T., U.S. Geological Survey, Pasadena, CA, USA, pagem@caltech.edu; FIELD, E. H., U.S. Geological Survey, Golden, CO, USA, field@usgs.gov; MILNER, K. R., University of Southern California, Los Angeles, CA, USA, kmilner@usc.edu
Rather than the prescriptive approaches used to define rupture rates in hazard models of the past, the 3rd Uniform California Earthquake Rupture Forecast (UCERF3) uses an inversion method to derive rupture rates consistent with fault slip rates, paleoseismic data, and regional seismicity rates. The inversion method allows us to probe which hypotheses are consistent with the model parameterization and input data. We were unable to construct a model that imposed a Gutenberg-Richter magnitude distribution on the major faults in California. This is because such a model cannot put enough moment on the faults to match the target slip rates without a) assuming significantly more fault connectivity and therefore raising the maximum magnitude, b) over-predicting historically observed seismicity rates, and/or c) introducing significant b-value variation by lowering the b-value on major faults. We present evidence for some of these potential “fixes” which could make the Gutenberg-Richter branch tractable, but possibly not to the extent they would be needed. Interestingly, UCERF3 is significantly less “characteristic” than UCERF2, both statewide (where rates are constrained to match the Gutenberg-Richter distribution) and in smaller regions, due to the relaxation of segmentation and inclusion of multi-fault ruptures.
Session:Explosive Source Characterization
Presenter   Vorobiev, Oleg
Schedule   Wed / Poster
Room   Cook/Arteaga
Understanding Shear Motion from Underground Explosions in Granite
VOROBIEV, O. Y., LLNL, Livermore, CA, USA, vorobiev1@llnl.gov; STROUJKOVA, A., Weston Geophysical Corp., Lexington, MA, USA; EZZEDINE, S. M. E., LLNL, Livermore, CA, USA, ezzedine1@llnl.gov; HERBOLD, E. B., LLNL, Livermore, CA, USA, herbold1@llnl.gov; GLENN, L., LLNL, Livermore, CA, USA; ANTOUN, T., LLNL, Livermore, CA, USA
This work is focused on analysis of near-field measurements recorded during Source Physics Experiments as well as experiments conducted in central New Hampshire. One of the main goals of these experiments is to investigate the possible mechanisms of shear wave generation in the nonlinear source region. Experiments revealed significant tangential motion at many locations. Furthermore, azimuthal variations in radial velocities were also observed which cannot be generated by a spherical source in isotropic materials. Understanding the nature of this shear motion is important for discriminating between the natural seismicity and underground explosions signatures. Possible mechanisms leading to such motion include, but are not limited to, heterogeneities in the rock such as joints, faults and geologic layers as well as surface topography and vertical motion at the surface caused by material spall and gravity. We have performed 3D computational studies considering all these effects. We assume that the rock has preexisting fractures which can modify radial motion from the source based on local conditions such as friction, water saturation and fracture orientation. Also, in repeatable events these conditions may change due to the damage closer to the source and at the boundaries between various layers. We use history variables in the model which describe damage accumulation both at the preexisting fractures and in the continuum to model new fracture generation. Results of our studies in the near field can be used to build a better source model for seismic wave propagation from underground explosions which includes information from the site characterization. This work performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.
Session:Induced Seismicity
Presenter   Llenos, Andrea
Schedule   Thu / Poster
Room   Cook/Arteaga
Characterizing Potentially Induced Earthquake Rate Changes in the Brawley Seismic Zone, CA
LLENOS, A. L., US Geological Survey, Menlo Park, CA, USA, allenos@usgs.gov; MICHAEL, A. J., US Geological Survey, Menlo Park, CA, USA, michael@usgs.gov
The Brawley Seismic Zone (BSZ), located in the Salton Trough of southern California, has a long history of earthquake swarms and geothermal energy exploitation. Some of these earthquake rate changes may have been induced by fluid extraction and injection activity at local geothermal fields, particularly the North Brawley Geothermal Field (NBGF) and the Salton Sea Geothermal Field (SSGF) (e.g., Chen and Shearer, JGR, 2011; Brodsky and Lajoie, Science, 2013). We explore this issue by examining earthquake rate changes and interevent distance distributions in these fields. For example, in Oklahoma and Arkansas, where considerable wastewater injection occurs, increases in background seismicity rate and aftershock productivity and decreases in interevent distance were diagnostic of fluid-injection induced seismicity (Llenos and Michael, BSSA, 2013). Here we test if similar changes occur that may be associated with fluid injection and extraction in geothermal areas. We use the stochastic Epidemic-Type Aftershock Sequence (ETAS) model (Ogata, JASA, 1988) to detect changes in the underlying seismogenic processes, shown by statistically significant changes in the model parameters. Model changes in the SSGF roughly occur when large changes in net fluid production occur, but a similar correlation is not seen in the NBGF. Also, while increases in both the background seismicity rate and aftershock productivity parameters were associated with fluid-injection induced earthquake rate changes in Oklahoma and Arkansas, in the BSZ only the background rate increases significantly, which tends to correspond with net fluid production rate increases. Moreover, in both fields the interevent spacing does not change significantly with the start of field operations in 1982. This suggests that geothermal field activities may not significantly change the physics of earthquake interactions, but earthquake rates may still be driven by fluid injection or extraction rates, particularly in the SSGF.
Session:Recent Advances and Findings in Earthquake Geology and Paleoseismology
Presenter   Prentice, Carol
Schedule   Fri 3:45 PM / Oral
Room   Room 3
New Paleoseismic Study of the Septentrional Fault, Dominican Republic
PRENTICE, C. S., US Geological Survey, Menlo Park, CA, USA, cprentice@usgs.gov; SCHARER, K., US Geological Survey, Pasadena, CA, USA, kscharer@usgs.gov; GOLD, R., US Geological Survey, Golden, CO, USA, rgold@usgs.gov; DELONG, S., US Geological Survey, Menlo Park, CA, USA, sdelong@usgs.gov; PENA, L., Universidad Catolica Madre y Maestra, Santiago, Dominican Republic
The Septentrional Fault is one of two major strike-slip faults that accommodate motion between the Caribbean and North American plates at the longitude of Hispaniola. Earlier paleoseismic work (Prentice, et al., 1993; 2003) showed that east of Santiago, Dominican Republic, the most recent earthquake with surface rupture (MRE) on this fault occurred more than 800 years ago. However, recent analysis of the historical record by ten Brink et al. (2011) concluded that earthquakes in 1842 and 1562 both ruptured the Septentrional Fault across the island. We excavated two trenches at Santiago Viejo, the site of an early European settlement about 1 km northwest of Santiago. The Spanish established the settlement in 1508, and abandoned the site and moved the city after the 1562 earthquake. While the fault traverses the site, there is no evidence among the ruins for historical surface rupture; no colonial ruins are offset, and the scarp itself is degraded by erosion, with no evidence suggesting historical slip. We excavated two trenches across the fault zone to determine if either historical earthquake was associated with surface rupture at this site. Both trenches exposed the fault zone overlain by an unfaulted cultural horizon (defined by the presence of artifacts from the Spanish settlers), suggesting that the MRE occurred prior to European settlement. Radiocarbon dating of charcoal samples collected from post-faulting sedimentary horizons show that the MRE pre-dates the 1842 earthquake, and likely also pre-dates the 1562 earthquake. These results indicate that if 1842 surface rupture occurred on the Septentrional Fault, it occurred on the segment of the fault west of this site. Our preliminary results show no evidence for surface rupture in 1562, but do not entirely exclude it. Current data suggest an age older than 1562 for the MRE. We anticipate that forthcoming radiocarbon and pollen analyses will better constrain the age of the MRE at the Santiago Viejo site.
Session:Diverse Mechanisms of Subduction Zone Fault Slip: Exploring the Relationships Among Seismic Rupture, Transient Slip, and Steady Creep
Presenter   Erickson, Brittany
Schedule   Wed / Poster
Room   Cook/Arteaga
A Provably Stable Finite-Difference Method for Earthquake Cycle Simulations within Subduction Zones
ERICKSON, B. A., San Diego State University, San Diego, CA, USA, berickson@mail.sdsu.edu; DAY, S. M., San Diego State University, San Diego, CA, USA, sday@mail.sdsu.edu
We are developing a computational method for earthquake cycle simulations within complex geometries.  The method is developed for the classical plane-strain problem that can incorporate non-planar fault geometries and material heterogeneities.  The off-fault volume is discretized using finite-differences and coordinate transforms to handle the curvilinear grid that conforms to the fault. All boundary conditions are imposed weakly, yielding a provably stable method.  As a first step towards developing this method, we consider a planar, vertical strike slip fault governed by rate-and-state friction, and load the system at the remote boundaries at the slow, tectonic plate rate. During the long interseismic period we solve the equations for static-equilibrium.  Once an event nucleates we have the option of simulating quasi-dynamic events within the framework, or use the current numerical solution as initial input into the dynamic rupture code SORD that captures the fine details of wave propagation.  An important application for this computational method will be to fault geometries representative of subduction zones, where the earthquake cycle may be strongly influenced by both fault geometry and material property variations.
Session:Earthquake Physics and Interaction
Presenter   Martinez-Torres, Fernando
Schedule   Fri / Poster
Room   Cook/Arteaga
STUDENT
Applying the Time-Domain Moment Tensor Inversion Technique to Regional Earthquake Data in the Puerto Rico-Virgin Island Region
MARTINEZ-TORRES, F. A., University of Puerto Rico at Mayaguez, Mayaguez, PR, USA, fernan.martinez.2@gmail.com; LOPEZ-VENEGAS, A., University of Puerto Rico at Mayaguez, Mayaguez, PR, USA, alberto.lopez3@upr.edu
Although the Caribbean plate region is characterized by slow (~2 cm/yr) plate motions, it is still prone to earthquakes and tsunamis along its boundaries where most of the land is concentrated. Moreover, most of the population residing on these active tectonic regions is concentrated along the coastal areas, thus resulting in an increasing seismic hazard in the eventuality of tsunamigenic earthquakes. Being able to detect and alert coastal populations in a timely manner is critical to safeguard lives and infrastructure. This research project explores a thoroughly validated, well-formulated time-domain moment tensor inversion code in order to obtain in quasi real-time faulting parameters of significant regional earthquakes in the Puerto Rico – Virgin Islands region. The inversion code has been developed by researchers at the Berkeley Seismological Laboratory, whose main attractive is to decrease the time it takes to have an estimate calculation of a moment tensor for any major earthquake using regional data, approximately less than 7 minutes of an earthquake’s origin time. Three seismic events in the region have been used as testbed to the inversion code configured for this area. In order to compare our results, previously computed and published moment tensor inversions from the Global CMT and USGS for the same events were used to assess the deviations from results obtained in this study . Our results indicate the inversion method is capable of reproducing the regional and teleseismic solutions, and thus can be incorporated into daily earthquake location operations at the Puerto Rico Seismic Network (PRSN) for quick estimation of faulting mechanisms and tsunami warning purposes.
Session:Fifty Years of Tsunami Science: from the 1964 Earthquake and Tsunami to the SAFRR Tsunami Scenario - Advances in Tsunami Source Characterization, Numerical Analysis and Hazard Mitigation
Presenter   Nicolsky, Dmitry
Schedule   Fri 1:45 PM / Oral
Room   Room 4
Tsunami Modeling and Inundation Mapping in Southcentral Alaska
NICOLSKY, D. J., University of Alaska Fairbanks, Fairbanks, AK, USA, djnicolsky@alaska.edu; SULEIMANI, E. N., University of Alaska Fairbanks, Fairbanks, AK, USA, elena@gi.alaska.edu; KOEHLER, R. D., Alaska Division of Geological & Geophysical Surveys, Fairbanks, AK, USA, richard.koehler@alaska.gov; WEST, M. E., University of Alaska Fairbanks, Fairbanks, AK, USA, mewest@alaska.edu
The Alaska Earthquake Center (AEC) participates in the National Tsunami Hazard Mitigation Program by evaluating and mapping potential tsunami inundation of coastal Alaska. We evaluate potential tsunami hazards for several coastal communities near the epicenter of the 1964 Great Alaska Earthquake and numerically model the extent of their inundation due to tsunamis generated by earthquake and landslide sources. Tsunami scenarios include a repeat of the tsunami triggered by the 1964 Great Alaska Earthquake, as well as hypothetical tsunamis generated by 1964-type ruptures, a Cascadia megathrust earthquake, and a hypothetical Tohoku-type rupture in the Gulf of Alaska region. Local underwater landslide events in several communities are also considered as credible tsunamigenic scenarios. We perform simulations for each of the source scenarios using AEC’s recently developed and tested numerical model of tsunami wave propagation and runup. The tsunami scenarios are intended to provide guidance to local emergency management agencies in tsunami hazard assessment, evacuation planning, and public education for reducing future casualties and damage from tsunamis. During the 1964 earthquake, locally generated waves of unknown origin were identified at several communities, located in the western part of Prince William Sound. The waves appeared shortly after the shaking began and swept away most of the buildings while the shaking continued. We model the tectonic tsunami assuming different tsunami generation processes and claim the importance of including both vertical and horizontal displacement into the 1964 tsunami generation process.
Session:Site Response: From Site-Specific Analyses to Predictive Models Around the Globe
Presenter   Thompson, Eric
Schedule   Wed 9:00 AM / Oral
Room   Room 4
Site Response Mapping with Fewer Proxies
THOMPSON, E. M., San Diego State University, San Diego, CA, USA, ethompson@mail.sdsu.edu; WALD, D. J., USGS, Golden, CO, USA, wald@usgs.gov
Site response is frequently estimated through empirical correlations with the time-averaged shear-wave velocity to 30 m (Vs30), and empirical correlations have been developed to predict Vs30 from mapped quantities such as surface geology, topographic slope, or terrain. This creates an unnecessary "weak link" in the site response map, which we eliminate by correlating empirical amplification function (EAFs) directly to surface geology and topographic slope. By eliminating the unnecessary link of Vs30 between the mapped proxy and site amplification, we eliminate some uncertainty in the site response map. An additional benefit of focusing on EAFs is to achieve the best possible estimate of the site response at strong motion stations; site-specific EAFs at stations that have recorded numerous ground motions reflect the actual site response rather than approximations based on Vs30 or the correlations of EAFs. This will be particularly useful for ShakeMap when correcting recorded motions to reference rock motions, which are used to interpolate between stations. The use of EAFs for correcting recordings to rock conditions will improve the underlying rock reference layer of ShakeMap, which is currently estimated from recorded ground motions that are adjusted to rock conditions with Vs30-based corrections. Estimating nonlinearity is a key challenge to employing EAFs for larger ground motions. To address this, we select the representative Vs30, which is the Vs30 that minimizes the misfit between the EAF and the empirical site response equations. The representative Vs30 then determines the nonlinear factors that modify the EAFs for large-amplitude motions.
Session:Fault Structure, Heterogeneity, and Implications for Rupture Dynamics
Presenter   Hirano, Shiro
Schedule   Wed 2:15 PM / Oral
Room   Room 3
BIEM Simulation for Anti-Plane Dynamic Rupture Propagation across a Bimaterial Interface.
HIRANO, S., University of Tsukuba, Tsukuba, Ibaraki, Japan, hirano@kz.tsukuba.ac.jp; YAMASHITA, T., Earthquake Research Institute, University of Tokyo, Bunkyo, Tokyo, Japan, tyama@eri.u-tokyo.ac.jp
Boundary Integral Equation Method (BIEM) enables us to simulate dynamic rupture propagation on non-planar faults under some assigned stress conditions with high-precision. Although applicability of BIEM had been limited only for rupture in a homogeneous medium, we modify it for anti-plane rupture in a 2D bimaterial. For the modification, we treat a discretized fault element embedded in the bimaterial and derive an analytical solution that represents radiated waveforms of stress components due to a sudden slip on the element. Our method extends a simulation of dynamic rupture propagation to the case of a fault that intersects with a bimaterial interface. In our results, we find that temporal increase or decrease of slip velocity goes back from an intersection of the fault and the interface toward a nucleation point after a rupture front reaches the intersection. A sign of the temporal change is consistent with a sign of reflected wave that depends on material contrasts. Hence our results are qualitatively feasible and comprehensible in view of an elastic wave theory. Magnitude of the change is, however, not easy to predict because reflection coefficients have complicated dependence on a spatiotemporal relationship between a wavesource and a receiver. For example, in a case of a frictionless fault, we confirm that slip velocity temporally takes negative value when the temporal decrease of slip velocity passes. Temporal change of slip velocity can cause more complicated behaviour of faulting under a strongly rate-dependent friction law. Under the friction law, temporal increase/decrease of slip velocity leads weakening/strengthening of friction and vice versa, so that positive feedback works. We suggest a possibility of reactivation or deactivation of slip off a rupture front by considering our results and the friction law.
Session:Advances in Seismic Imaging and Monitoring of Time-Dependent Variations: Civil Structures, Near-Surface, and Shallow Crustal Scales
Presenter   Thompson, Lennox
Schedule   Wed 9:15 AM / Oral
Room   Room 3
STUDENT
Developing 3D Shear Wave Models Using a Multi-Objective Joing Inversion Scheme
THOMPSON, L. E., University of Texas at El Paso, El Paso, TX, USA, lethompson@miners.utep.edu; VELASCO, A. A., University of Texas at El Paso, El Paso, TX, USA, aavelasco@utep.edu; GARCIA, V., University of Texas at El Paso, El Paso, TX, USA, vhgarcia4@miners.utep.edu; ZAMORA, A., University of Texas at El Paso, El Paso, TX, USA, azamora3@miners.utep.edu; SOSA, U. A., University of Texas at El Paso, El Paso, TX, USA, usosaaguirre@miners.utep.edu
Understanding Earth’s tectonic processes requires determining the Earth structure, and we focus on determining Texas tectonic processes employing a joint inversion scheme for multiple geophysical datasets. In particular, we expand on a constrained optimization approach for a joint inversion least-squares (LSQ) algorithm to characterize a one-dimensional Earth's structure, We use seismic data from regional networks in Texas and the USArray, a dense network of permanent and portable seismographs placed systematically across the continental United States. We collect and process seismic data (receiver function, surface wave, body wave travel times), plus collect gravity data from other efforts. By jointly inverting these four geophysical data sets, we can avoid the inherent non-uniqueness from inversion.The Multi-Objective Optimization Problem (MOP) technique enables the capability to combine linear and non-linear problems. The technique enables the combination of inhomogeneous data sets and different statistical properties of error associated with each data set. We chose to use the MOP optimization approach because we want to find the best possible solution for our nonlinear geophysics inverse problem. We prove through numerical and experimental testing that our MOP scheme performs inversion in a more accurate, robust, and flexible matter than traditional inversion approaches. To develop quasi 3D models, we interpolate the 1D results using a kriging approach, only when we have excellent station coverage. Our quasi 3D velocity models provide insight into the tectonic history and physical properties of the Earth structure of Texas, and we will compare mantle structure beneath Texas to other ancient and active rift systems.
Session:Pillars of Simulation: Seismic Velocity and Material Models
Presenter   Tape, Carl
Schedule   Wed 2:15 PM / Oral
Room   Room 4
Seismic Wavefield Simulations of Earthquakes within a Complex Crustal Model for Alaska
TAPE, C., U. Alaska Fairbanks, Fairbanks, AK, USA, carltape@gi.alaska.edu
We explore the effects of two major sedimentary basins (Cook Inlet and Nenana) on the seismic wavefield in Alaska, in preparation for an iterative tomographic inversion using spectral-element and adjoint methods. Our 3D reference seismic velocity model contains three geometrical interfaces: the Moho surface, the basement surface of the sedimentary basins, and the topographic surface. The crustal and upper mantle tomographic model is from Eberhart-Phillips et al. (2006), but modified by removing the uppermost slow layer and then embedding models for the two major sedimentary basins. Synthetic seismograms are computed using the spectral-element method. We quantify the accuracy of the initial 3D reference model by comparing 3D synthetics with observed data for several earthquakes originating in the crust and underlying subducting slab. We explore the effect of complex crustal structure on the seismic wavefield of the 1964 Mw 9 earthquake source model of Ichinose et al. (2007), which was obtained using geodetic data, tsunami arrival times, and teleseismic P waves recorded from the earthquake. Cook Inlet basin has a prominent effect on the seismic wavefield in southern Alaska. This 8-km-thick sedimentary basin is situated between Anchorage and a major portion of the ~700-km-long rupture of the subduction interface. The relatively large simulation region can be used to test different scenario megathrust earthquakes as well as different models of the 1964 earthquake.
Session:Recent Advances in Ground Motions Simulation Methods and Their Validation
Presenter   Laurendeau, Aurore
Schedule   Thu / Poster
Room   Cook/Arteaga
Broadband Acceleration Time Histories Synthesis by Coupling Low Frequency Seismic Noise and High Frequency Stochastic Modelling
VIENS, L., IPGP, France now at ERI, Univ. Tokyo, Japan, viens@eri.u-tokyo.ac.jp; LAURENDEAU, A., CEA, DAM, DIF, F-91297, Arpajon, France; BONILLA, L. F., University Paris-Est, IFSTTAR, France; SHAPIRO, N., IPGP, France
Information carried by the ambient seismic noise is exploited to extract Green's functions between two seismic stations using one as a "virtual" source. The method developed by Prieto and Beroza (2008) is used and validated by comparing the ambient noise impulse response waveforms with earthquakes of moderate magnitudes from 4 to 6.1 in Japan. As the information is only available at low frequencies (less than 0.25 Hz), the seismic noise approach is coupled with a non-stationary stochastic model allowing to simulate time domain accelerograms from 0 to 50 Hz. This coupling allows the predicted ground motion to have both the deterministic part at low frequencies coming from the source and the crust structure and the high frequency random contribution from the seismic waves scattering. The resulting broadband seismograms show a good agreement with observed ground motions from real earthquakes in the northern Tokyo area in Japan.
Session:Seismic Location and Processing Techniques
Presenter   Chiu, Hung-Chie
Schedule   Fri / Poster
Room   Cook/Arteaga
Scaling Rotational Motion and Its Induces Effects
CHIU, H. C., Institute of Earth Sciences, Academia Sinica, Taipei, Taiwan, chiu@earth.sinica.edu.tw
A strong-motion accelerograph is designed mainly for measuring three-component translational accelerations. However, accelerometer can also detect rotation motions and related effects such as the centrifugal acceleration, gravity (ground tilt) effect and effects of rotation frame. Analyses of six-component ground motions (three-component translational motions and three-component rotational motions) can recover the translation motion and extract the rotational effects. Unlike the translational motion, the peak values of rotation motion and its effects are controlled by PGA rather than by the size and distance of earthquakes. We analyze a set of collocated and well recorded rotation rate-strong motion velocity data. Results show that the peak values of rotational effects vs. PGA are shown to be in quadratic form on a log-log scale which implies that these effects grow up very quickly in large ground motions. This relationship also implies that the rotational motions might have large variation at large translational motions. Both these features seem to fit the observations of near-field and extreme large ground motions.
Session:New Directions in PSHA: Ins, Outs, and Uncertainty
Presenter   McBean, Kevin
Schedule   Fri / Poster
Room   Cook/Arteaga
STUDENT
Event Terms for the Shallow 2011 Normal Faulting Earthquakes in the Ibaraki - Fukushima Prefectural Border Region, Japan
MCBEAN, K., Nevada Seismological Laboratory, University of Nevada, Reno, NV, USA, kevinmcbean@gmail.com; ANDERSON, J. G., Nevada Seismological Laboratory, University of Nevada, Reno, NV, USA, jga@unr.edu; KAWASE, H., DPRI, Kyoto University, Gokasho, Uji, Kyoto, Japan, kawase@zeisei.dpri.kyoto-u.ac.jp
In the global database of strong earthquake ground motions, there is considerably less data for large normal faulting events in the shallow crust than for events with strike-slip and reverse mechanisms. After the Mw9.0, 2011 Tohoku, Japan earthquake, a sequence of very well recorded normal faulting events occurred on land. The largest event, the Mw 6.6 Fukushima-Hamadori earthquake, caused surface faulting with up to 2 meters of vertical slip. These events greatly expanded the amount of strong-motion data for earthquakes with normal mechanisms, but occurred too late to be included in the 2013 updates of the NGA ground motion prediction equations. Anderson et al. (2013) found that the ground motions during the largest event were large compared with the 2008 NGA relations. This paper compares all 13 earthquakes in the sequence with Mw greater than 5.0 with the 2013 NGA relations. We applied the site-specific station corrections of Kawase and Matsuo (2004) to the raw data, as was done by Anderson et al. These corrections significantly reduced the variance compared to the raw data. For PGA and PGV we found event terms, i.e. the average of the residuals, ln(observed/NGA model), based on data within 100 km of the fault. While comparisons depend on the specific model, some general trends appear in our preliminary comparisons. For PGA, the event terms are near zero or positive at lower magnitudes (Mw under 6.0), and positive at higher magnitudes. For PGV, the event terms are positive at all magnitudes. For most comparisons, the residuals increase with distance; the slope of this distance dependence is greater for the lower magnitudes for both PGA and PGV. Our ongoing study aims to extend the results to more points on the response spectrum, and reconsider the site-response corrections.
Session:Seismic Location and Processing Techniques
Presenter   Baca, Austin
Schedule   Fri / Poster
Room   Cook/Arteaga
Depths of Indian Ocean Intraplate Earthquakes From Waveform Modeling
BACA, A. J., California State University Polytechnic of Pomona, Pomona, CA, USA, austinbaca1992@gmail.com; POLET, J., California State University Polytechnic of Pomona, Pomona, CA, USA, jpolet@csupomona.edu
The Indian Ocean is a region of complex tectonics and anomalous seismicity. The great magnitude 8+ strike-slip events that took place in the Indian Ocean on April 11th, 2012 are unique because their rupture extended to a depth within the plate where seismic activity was considered to be impossible. The oceanic lithosphere in this region possesses many bathymetric features, most notably the multiple inactive fracture zones within the Wharton Basin, and the Ninetyeast Ridge. Studies have described the high seismic activity within the oceanic lithosphere in this area as delineating a diffuse deformation region where the Indian sub-plate is separating from the Australian sub-plate. Other studies have found that the aforementioned fracture zones have been reactivated due to the current stress regime. We will present the results of our investigation of the focal depth of dozens of intraplate earthquakes that have occurred throughout the Indian Ocean in the past three decades. For these events, global moment tensor inversions are commonly carried out with fixed hypocentral depths, and a more detailed examination of the waveforms is required to determine accurate depths. A majority of the earthquakes we selected are between M5.5-6.5 with large tension axis plunge, chosen to provide optimal teleseismic P-waves for waveform analysis. 1D synthetic seismograms are computed for a range of depths based on existing lithospheric velocity models. Through a comparison of these synthetic waveforms with the recorded data we determine improved earthquake hypocentral depths
Session:Diverse Mechanisms of Subduction Zone Fault Slip: Exploring the Relationships Among Seismic Rupture, Transient Slip, and Steady Creep
Presenter   Yamaguchi, Tetsuo
Schedule   Wed 9:00 AM / Oral
Room   Room 1
Laboratory Experiments of Seismic Cycles: Effects of Normal Stress and Its Gradient
YAMAGUCHI, T., Kyushu University, Fukuoka, Japan, yamaguchi@mech.kyushu-u.ac.jp; HORI, T., JAMSTEC, Yokohama, Japan, horit@jamstec.go.jp; SAKAGUCHI, H., JAMSTEC, Yokohama, Japan, sakaguchih@jamstec.go.jp; AMPUERO, J. P., California Institute of Technology, Pasadena, CA, USA, ampuero@gps.caltech.edu
We report our studies on laboratory experiments inspired by seismic cycles in subduction zones. We conducted friction experiments between a Plexiglass block and a Silicone Gel plate driven at a constant plate velocity [1]. By changing the normal force and the inclination angle of the upper block, we controlled the normal stress level and its gradient. As a result, we successfully observed a wide variety of slip events and size statistics [2]. In small normal force and small inclination angle conditions, slip events were dominated by slow events which occur uniformly at the block-plate interface, and the size statistics obeyed the power law with cutoff. In intermediate force and angle conditions, coexistence of small slow events in “deep” (large normal stress) regions and giant events rupturing “shallower” regions were clearly seen and the size distributions obeyed power-law with a peak at large sizes. In larger force and angle conditions, however, the giant slip events disappeared and only small and intermediate slip events were observed. Interestingly, when we did experiments at the boundary between the latter two regimes, very complex stick-slip cycles wandering about the latter 2 dynamics modes were seen. Based on our visualization of 2D displacement fields at the frictional interface, we determined epicenter, rupture area, and the moment magnitude for each event, and we also estimated 2D stress distributions during stress accumulation processes. We found that rupture behavior for a giant event was significantly influenced by the stress distribution or stress history before the event. Reference: [1] T. Yamaguchi, M. Morishita, M. Doi, T. Hori, H. Sakaguchi, J. P. Ampuero, Gutenberg-Richter law in sliding friction of gels, J. Geophys. Res. Solid Earth, 116, B12306 (2011) [2] T. Yamaguchi, T. Hori, H. Sakaguchi, J. P. Ampuero, to be submitted.
Session:Recent Advances in Ground Motions Simulation Methods and Their Validation
Presenter   Moczo, Peter
Schedule   Thu / Poster
Room   Cook/Arteaga
A New Discrete Representation of Heterogeneous Medium for the Staggered-grid Finite-difference Modelling of Earthquake Motion
KRISTEK, J., Comenius University Bratislava, Bratislava, Slovakia, kristek@fmph.uniba.sk; MOCZO, P., Comenius University Bratislava, Bratislava, Slovakia, moczo@fmph.uniba.sk; CHALJUB, E., ISTerre - Institut des Sciences de la Terre, Grenoble, France, Emmanuel.Chaljub@ujf-grenoble.fr; DE MARTIN, F., BRGM - Bureau de Recherches Géologiques et Minières, Orleans, France, F.DeMartin@brgm.fr; KRISTEKOVA, M., Slovak Academy of Sciences, Bratislava, Slovakia, kristekova@savba.sk; GALIS, M., King Abdullah University of Science & Technology, Thuwal, Saudi Arabia, Martin.Galis@kaust.edu.sa
Sufficiently realistic models are necessary for numerical prediction of earthquake ground motion especially in local surface sedimentary structures capable to produce anomalous earthquake motion. It is also obvious that the realistic physical model has to be sufficiently accurately represented by discrete grid models in the (spatial) domain numerical methods such as finite-difference methods. Recent international comparative exercises for a typical deep Alpine Grenoble valley, France, (ESG 2006) and a shallow sedimentary Mygdonian basin, Greece, (E2VP - Euroseistest Verification and Validation Project 2008-2012) indicated that the existing finite-difference schemes do not reach sufficient accuracy if surface waves are locally induced and propagating along interface with strong velocity contrast. Recent finite-difference schemes represent a large variety of approaches with considerable differences in accuracy and computational efficiency in realistic models with large velocity contrast and complex geometry of material interfaces. This is mainly due to a level of inconsistency of the various discrete representations of the material interfaces with the boundary conditions at the interfaces. The E2VP experience led us to improve the existing discrete representation of strong material heterogeneity. We have developed a new discrete representation which is even more accurate than that presented by Moczo et al. (2002) while still consistent with the spatial distribution of particle-velocity and stress-tensor components in the finite-difference staggered-grid scheme. We briefly present theoretical principles of the new discrete representation and demonstrate its superior accuracy using extensive numerical tests against exact and reliable independent solutions.
Session:Advances in Seismic Imaging and Monitoring of Time-Dependent Variations: Civil Structures, Near-Surface, and Shallow Crustal Scales
Presenter   Gueguen, Philippe
Schedule   Wed / Poster
Room   Cook/Arteaga
Instantaneous Monitoring of Soil and Structure Dynamics Using Autocorrelation Functions Computed with the Stockwell Transform
BONILLA, L. F., IFSTTAR, Marne la Vallee, France, fabian.bonilla@ifsttar.fr; GUEGUEN, P., ISTerre/CNRS/UJF/IFSTTAR, Grenoble, France, philippe.gueguen@ujf-grenoble.fr
Monitoring the speed changes of a medium is very important in detecting variations related to the stress, temperature and fluid content. There are several techniques to do so, and one of them is the use of autocorrelation functions (ACF) that recover a zero-offset reflection seismogram. Traditionally, the autocorrelation functions are computed by stacking several time-windows of seismic data in order to enhance the signal-to-noise ratio. In this sense, the computed ACF shows an average of the dynamics of the physical process. To overcome this limitation, we propose to use the Stockwell transform, which is a time-frequency decomposition of a signal that preserves the phase information, to follow the evolution in time of the spectral properties of a seismic record. More specifically, at each time step, the Stockwell transform provides the amplitude and phase of the instantaneous wavelet that composes the signal, so that an instantaneous ACF can be computed. The analysis of these multiple ACFs allow to compute the time shift at a given ACF produced by the dynamics of the media. Changes of the time shift can easily be transformed into speed changes of the media. In this study, we present preliminary results on the analysis of K-net records that show liquefaction during the Tohoku-Oki 2011 earthquake. Furthermore, we show the effect of moderate to strong events of the response of the Factor building in UCLA, California.
Session:New Directions in PSHA: Ins, Outs, and Uncertainty
Presenter   Bora, Sanjay
Schedule   Fri / Poster
Room   Cook/Arteaga
STUDENT
A New Perspective towards the Generation of Response Spectral Ground Motion Prediction Equation for Seismic Hazard Analysis
BORA, S. 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. M., University of California, Berkeley, CA, USA, kuehn@berkeley.edu; STAFFORD, P. J., Imperial College London, UK, London, UK, p.stafford@imperial.ac.uk; EDWARDS, B., Swiss Seismological Service, ETH Zuerich, Zuerich, Switzerland, edwards@sed.ethz.ch
Adjustment of empirically derived ground motion prediction equations (GMPEs) to different seismological conditions is one of the major challenges in engineering seismology and seismic hazard studies. In the present study we propose a new and physically consistent framework for the development of a response spectral GMPEs which can be easily adjusted to different seismological conditions; and which does not suffer from the practical problems associated with adjustment in the response-spectral domain. The approach we present brings together two different paradigms of ground motion modeling. The first component, as described in its initial form by Bora et al. (2013), consists of an empirical FAS (Fourier Amplitude Spectrum) model and a duration model which are combined within the random vibration theory (RVT) framework (Boore, 2003) to obtain the full response spectral ordinates. In addition, we extrapolate the observed FAS beyond the frequencies which are supported by the observed acceleration records using stochastic FAS models, obtained by inversion as described by Edwards and Fäh (2013). This way we obtain response spectral models which are easily adjustable to different sets of seismological parameters such as stress parameter, kappa values, t* values, etc. We present this analysis for the recently compiled strong-motion database RESORCE-2012 obtained for data from Europe and the Mediterranean region.
Session:Fault Structure, Heterogeneity, and Implications for Rupture Dynamics
Presenter   Song, Seok Goo
Schedule   Wed / Poster
Room   Cook/Arteaga
Re-visiting Linear Source Inversion with the Full Complexity of Earthquake Rupture, Including Both Slip Re-activation and Supershear Rupture
SONG, S. G., ETH Zurich, Zurich, Switzerland, song@sed.ethz.ch; DALGUER, L. A., ETH Zurich, Zurich, Switzerland, dalguer@sed.ethz.ch
Kinematic source inversion is an important research tool in investigating earthquake source by inverting seismic and geodetic data. For the last two decades, nonlinear source inversion has become popular in the community. The fast growing computational power allows us to apply global searching algorithms in the nonlinear inversion. We can estimate temporal source parameters, such as rupture velocity and slip duration, more explicitly in the nonlinear inversion. It is also beneficial to implement dynamically compatible source time functions. However, dynamic rupture models show that earthquakes may rupture in a very complex fashion, exhibiting multiple rupture fronts propagating at subshear and supershear rupture speeds, back-propagating rupture and slip reactivation (Gabriel et al. 2012) even on a single planar fault. Lee et al. (2011) show that the 2011 Tohoku-oki earthquake may have multiple ruptures, involving slip reactivation, by kinematic source inversion. Classical nonlinear inversion methods assume single slipping with a single pre-defined source time function, which may not work well to resolve such complex rupture processes. The identification of the complex rupture requires inversion methods that allow such complexities in model space. We perform synthetic source inversion tests with dynamically generated complex source models, including both slip reactivation and supershear rupture. We adopt a linear source inversion method constructing a model space with multiple time windows, allowing slip from the nucleation to the termination of rupture for every point on the fault. We adopt the Bayesian inversion to regularize model space effectively. We find that both reactivated secondary rupture and supershear rupture front can be detected well as long as they produce a significant level of seismic moment. It would be beneficial to apply the linear inversion method to check the complexity of rupture before we assume single slipping in the nonlinear inversion.
Session:Advances in Seismic Imaging and Monitoring of Time-Dependent Variations: Civil Structures, Near-Surface, and Shallow Crustal Scales
Presenter   Campillo, Michel
Schedule   Wed 11:30 AM / Oral
Room   Room 3
Noise-based Crustal Seismic Velocity Changes Associated with the 2011 Tohoku-oki Earthquake
BRENGUIER, F., ISTerre/Univ. of Grenoble, Grenoble, France, florent.brenguier@ujf-grenoble.fr; CAMPILLO, M., ISTerre/Univ. of Grenoble, Grenoble, France, michel.campillo@ujf-grenoble.fr; BRIAND, X., ISTerre/Univ. of Grenoble, Grenoble, France, xavier.briand@ujf-grenoble.fr; TAKEDA, T., NIED, Tsukuba, Japan, ttakeda@bosai.go.jp; AOKI, Y., ERI, Tokyo, Japan, yaoki@eri.u-tokyo.ac.jp; SHAPIRO, N., IPGP, Paris, France, nshapiro@ipgp.fr; EMOTO, K., NIED, Tsukuba, Japan, emoto@bosai.go.jp
The occurrence of the giant 2011 Tohoku-oki earthquake and its record by the dense Japanese Hi-net seismic network provide a unique opportunity to study the response of the Earth's crust to strong shacking and deformation. Here, we use one year of continuous seismic data from the Hi-net network in Japan to estimate noise-based crustal seismic velocity changes associated with the occurrence of the 2011 Mw9 Tohoku-oki earthquake. We present results obtained from correlations in the 0.1-0.9 Hz frequency range and for lapse time up to 60 s. We observe a strong coseismic drop of seismic velocities (up to 0.2%) in volcanic regions throughout Japan, including Mt Fuji Volcano about 400 km away from the epicentral area of the Tohoku-oki earthquake. This is the prominent feature of the response during the first days after the earthquake. The velocity drop increased in the following days in the region of maximum deformation produced by the earthquake, followed by a phase of slow recovery. We compare the temporal response of the crust in Japan after Tohoku-oki earthquake with the one observed in Sichuan basin and Tibet after the 2008 M7.9 Wenchuan earthquake (Froment et al., 2013, Obermann, 2013). In both cases, the results indicate that, beyond a relaxation of the non-linear effect on the seismic speed, post seismic processes affect the seismic velocity at depth.
Session:Merging Paths: Earthquake Simulations and Engineering Applications
Presenter   KarimZadeh Naghshineh, Shaghayegh
Schedule   Thu 4:30 PM / Oral
Room   Room 3
STUDENT
Nonlinear Time History Analyses of Structures under Real and Synthetic Ground Motions
KARIMZADEH, S., Middle East Technical University, Ankara, Turkey, shaghayegh.karimzadehnaghshineh@metu.edu.tr; 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 Interntional, Auriol, France, gabriele.ameri@gmail.com
Full time series of ground acceleration is required for nonlinear time history analyses which are employed for evaluating the dynamic response of a structure under earthquake excitations. For regions with sparse ground motion data, alternative ground motion simulation techniques are used to generate acceleration time series with varying levels of accuracy. While using simulated records for engineering purposes, it is critical to investigate the efficiency of these records in predicting the real engineering demands. In the first part of this study, nonlinear time history analyses of 9 typical multi-story reinforced concrete frame structures are performed to compare structural responses to synthetic records with those to the real ground motions of a particular event. The 2009 L’Aquila (Italy) (Mw=6.3) earthquake is simulated using two alternative simulation methods: Hybrid Integral-Composite method and Stochastic Finite-fault method. Results of nonlinear time history analyses from the real and synthetic records of the mentioned event are compared in terms of maximum displacement of each story levels. We observe that the stochastic finite fault method is found to be more conservative in predicting the seismic responses of RC frame structures for this specific case. In the second part of the study, to further test the validation of the Stochastic Finite Fault method in prediction of real dynamic responses, 1992 Erzincan earthquake (Turkey) (Mw=6.6) is considered. For this second event, results show that for cases where near-field or basin effects are of concern, Stochastic Finite Fault method may not sufficiently predict real responses.
Session:From the Earthquake Source to Damage of Buildings: Bridging the Gap between Seismology and Earthquake Engineering
Presenter   Askan, Aysegul
Schedule   Fri 4:15 PM / Oral
Room   Room 1
Estimation of Seismic Losses in Urban Regions within an Interdisciplinary Framework: A Case Study
ASKAN, A., Middle East Technical University, Ankara, Turkey, aaskan@metu.edu.tr; ERBERIK, M. A., Middle East Technical University, Ankara, Turkey, altug@metu.edu.tr; YAKUT, A., Middle East Technical University, Ankara, Turkey, ayakut@metu.edu.tr; KILIC, N., Prime Ministry Disaster and Emergency Management Presidency, Ankara, Turkey, nazan.yilmaz@afad.gov.tr; KARIMZADEH, S., Middle East Technical University, Ankara, Turkey, sh_naghshineh@yahoo.com; SISMAN, F. N., Middle East Technical University, Ankara, Turkey, f.nurtensisman@gmail.com; ASTEN, M., Monash University, Melbourne, Australia, michael.asten@sci.monash.edu.au
Risk mitigation in urban regions starts with identification of potential seismic losses in future earthquakes. Estimation of seismic losses concerns a wide range of authorities varying from geophysical and earthquake engineers, physical and economic planners to insurance companies while the process naturally involves inputs from multiple disciplines. In this study we present a city model where potential seismic losses are expressed in terms of building vulnerabilities and regional seismic hazard. The main components of the study are probabilistic and deterministic seismic hazard assessment and estimation of potential ground motions, regional building vulnerability and fragility information, casualty and economic loss functions. As the study area, Erzincan, a city on the eastern part of the North Anatolian Fault zone is selected. Located within a triple conjunction of major fault systems within a basin structure, and experienced a major event (Ms=8.3) in 1939, this city has significant hazard potential. We present the initial results in terms of key components such as construction of a 2D velocity model, ground motion simulations of past earthquakes and scenario events, site-specific probabilistic seismic hazard analyses and fragility functions derived using regional building characteristics along with simulated regional ground motion data.
Session:Site Response: From Site-Specific Analyses to Predictive Models Around the Globe
Presenter   Montalva, Gonzalo
Schedule   Wed 9:30 AM / Oral
Room   Room 4
Site Effects in Concepcion Basin: Geometry and Amplification from Gravity and Microtremors
MONTALVA, G. A., University of Concepcion, Concepcion, Bio-Bio, Chile, gmontalva@udec.cl; CHAVEZ-GARCIA, F. J., Instituto de Ingenieria, UNAM, Ciudad Universitaria, Mexico DF, Mexico, paco@pumas.ii.unam.mx; TASSARA, A., University of Concepcion, Concepcion, Bio-Bio, Chile, andrestassara@udec.cl; JARA, D., University of Concepcion, Concepcion, Bio-Bio, Chile, dariojara@udec.cl
The city of Concepcion suffered significant damage during the February 27, 2010, great Maule earthquake. The magnitude of that event was evidently a major factor. However, observed damage distribution in the city was irregular and not related in a straightforward manner to construction quality. Site effects are suspected to be a major factor. However, during the main shock only two seismic records were obtained in the area. In this paper, we present a site effects study of Concepcion. We gathered available information related to site response and subsoil structure. This includes: geologic history and interpretation regarding the basin structure, gravity measurements, and previous seismic noise studies using both single measurements and array of stations. In addition, we performed microtremor measurements. The data were analyzed using the standard horizontal-to-vertical spectral ratios (HVSR) for single station records, and SPAC and seismic interferometry for data from arrays. The first results consist of a dominant period and maximum amplification maps derived from HVSR. We show that, in the case of Concepcion, amplification values derived from HVSR are significant and useful to understand site effects. Dominant period shows an excellent correlation with the basin structure derived from gravity measurements, suggesting that site effects are governed by the impedance contrast between the sediments and the underlying basement, and that sediments spatial variability is relatively low. Array microtremor measurements allowed estimating shear wave velocity profiles in the sediments. Preliminary simulations are in good agreement with amplitude estimates from HVSR. Our results will be used to explore the relation between damage distribution during the Maule earthquake with the variability of ground motion due to the irregular subsoil structure.
Session:Recent Advances in Ground Motions Simulation Methods and Their Validation
Presenter   Lees, Jonathan
Schedule   Thu / Poster
Room   Cook/Arteaga
GPU-based 3D Simulation of Acoustic Wave Propagation with Complex Topography
KIM, K., University of North Carolina, Chapel Hill, NC, USA, keehoon@live.unc.edu; LEES, J. M., University of North Carolina, Chapel Hill, NC, USA, jonathan.lees@unc.edu
By transferring parallel computations to a high end video card (GPU) we managed to reduce the time of finite difference wave propagation calculations from 11 hours to 27 minutes, an improvement of 95%. This allows us to pursue complex computations without incurring the high costs of large parallel computer clusters or supercomputers. As a first experiment we simulated acoustic wave propagation in an artificial urban environment to illustrate the power of the computational set up. Reverse-time migration is accomplished by propagating a wavefield after reversing the time axis. Using synthetic examples, reverse-time migration of resultant waveforms indicates that we can reasonably reduce the number of stations from 36 to 3 (a 91% reduction) and still achieve a source location solution, although the presence of noise is evident. Application of the reverse-time method using infrasound data in the complex topographic region of Sakurajima volcano, Japan, provides a high resolution localization of the source region. Acoustic pressures from explosive eruptions at Sakurajima were recorded by a temporary network deployed in July, 2013, <7 km from the active vent and amplitudes varied significantly with the crater-station azimuth. Results of numerical modeling are consistent with the observed azimuthal distribution of sound amplitudes, suggesting that scattering and diffraction along the source-to-receiver path geometry is significant. Forward modeling using the GPU-based approach further reveals acoustic nodes of the wave propagation that may be critical to deployment planning. Since topographic relief near volcanoes is often pronounced, scattering and diffraction from topography represents a major part of propagation effects for local infrasound propagation. The GPU-based approach opens new possibilities for exploring 3D wave propagation for testing hypotheses and predicting accurate arrival phases in the presence of wave field distortion.
Session:Network Operations and Data Centers
Presenter   Herak, Marijan
Schedule   Thu / Poster
Room   Cook/Arteaga
Croatian Seismicity Database
HERAK, M., Department of Geophysics, Faculty of Science, Zagreb, Croatia; HERAK, D., Department of Geophysics, Faculty of Science, Zagreb, Croatia; IVANČIĆ, I., Department of Geophysics, Faculty of Science, Zagreb, Croatia
We present the contents and the quality of data on seismicity of Croatia and the neighboring regions (including all of Bosnia and Herzegovina), collected at the Department of Geophyisics in Zagreb since the late 19th century, when systematic seismology research began in Croatia. The latest revision of the database comprises the Croatian Earthquake Catalog (CEC) which currently lists over 65000 earthquakes, the macroseismic digital archive, a collection of fault-plane solutions (FPS) for over 200 events, a large set of onset-times of local and regional phases from Croatian and nearby networks, and various supporting material (photos, reports, questionnaires, calibration records, etc.). Here we concentrate on the catalog and the FPS database only. The CEC is regularly updated twice a year. The rate of inclusion of new events rapidly increases, and now exceeds 9000 earthquakes/year. We also report on the recent magnitude (re)calibration and the new relation to convert ML to Mw. Analyses of catalog completeness in time and in space reveals rather large heterogeneity mostly caused by uneven process of densification of national seismological networks, but also by distribution of population and the degree of development of some subregions in the preinstrumental era. The thoroughly revised earthquake mechanisms database contains the first-polarity FPS that are since recently being routinely computed for most of Croatian earthquakes with magnitudes above about 3.0, as well as the best double-couple parameters of the CMT solutions from various sources. Although the inferred general orientation of P-axes is in agreement with the assumed SW–NE directed collision of the Adriatic microplate and the Dinarides as the main tectonic driving force, observed systematic deviation from this trend in some areas may indicate significant influence of local conditions in modifying the regional stress-field.
Session:Earthquake Physics and Interaction
Presenter   Barnes, Kyle
Schedule   Fri 4:45 PM / Oral
Room   Room 7/8
STUDENT
Stress Drop Studies of Earthquakes in the Southern North Island and Cook Strait Region of New Zealand
BARNES, K., University of Texas-El Paso, El Paso, TX, USA, kebarnes@miners.utep.edu; DOSER, D. I., University of Texas-El Paso, El Paso, TX, USA, doser@utep.edu; ABERCROMBIE, R. E., Boston University, Boston, MA, USA
We are using the empirical Green’s function technique to estimate stress drops of events in the southern North Island of New Zealand and Cook Strait. In this region the Pacific Plate subducts obliquely beneath the Australian Plate along the Hikurangi Trough. The plate interface is strongly coupled, with slow slip events occurring downdip of the strongly coupled zone. The goal of our study is to determine if earthquakes occurring at the edges of the strongly coupled zone have higher stress drops than those located in parts of the study area. Our preliminary efforts are focused on the 2004-2005 Upper Hutt sequence (M3 to M5.6) occurring within the Pacific Plate downdip of the locked zone and events of the 2013 Cook Strait sequence (2 magnitude ~ 6.5 events). The results of this study will provide insight into stress loading along the interface and is part of a larger study to determine if stress drops of New Zealand earthquakes vary with tectonic setting.
Session:75 Years of Frequency-Size-Distribution of Earthquakes: Observations, Models and Understanding
Presenter   Cesca, Simone
Schedule   Fri 11:00 AM / Oral
Room   Room 4
Growing Large Scale Fractures in a Salt Mine and their Effects on the Frequency-Magnitude Distribution
MAGHSOUDI, S., University of Potsdam, Potsdam, Germany, samira.maghsoudi@geo.uni-potsdam.de; CESCA, S., GFZ German Research Centre for Geosciences, Potsdam, Germany, simone.cesca@gfz-potsdam.de; HAINZL, S., GFZ German Research Centre for Geosciences, Potsdam, Germany, sebastian.hainzl@gfz-potsdam.de; DAHM, T., GFZ German Research Centre for Geosciences, Potsdam, Germany, torsten.dahm@gfz-potsdam.de; KAISER, D., BGR Hannover, Hannover, Germany
The spatiotemporal seismicity of a massive acoustic emission (AE) catalog recorded in the Morsleben salt mine is here investigated. We first summarize different approaches we have recently proposed to assess the catalogue completeness in strongly heterogeneous media and mining environments. The analysis of complete catalogue and the spatiotemporal distribution of the AE activity present characteristic signatures and the frequency-magnitude distribution locally deviates from the Gutenberg-Richter law. Almost a year after backfilling of the cavities, microevents are distributed with distinctive stripe shapes above cavities at different depth levels. The physical forces driving the creation of these stripes are still unknown. This study aims to find the active stripes and track fracture developments over time by combining two different temporal and spatial clustering techniques into a single methodological approach. Anomalous seismicity parameters values like sharp b-value changes for two active stripes are good indicators to explain possible stress accumulation at the stripe tips. We identify the formation of two new seismicity stripes and show that the AE activities in active clusters are migrated mostly unidirectional to eastward and upward. This indicates that the growth of underlying macrofractures is controlled by the gradient of extensional stress. Studying size distribution characteristic in terms of frequency–magnitude distribution and b-value in active phase and phase with constant seismicity rate show that deviations from the Gutenberg–Richter power law can be explained by the inclusion of different activity phases: (1) the inactive period before the formation of macrofractures, which is characterized by a deficit of larger events (higher b-values) and (2) the period of fracture growth characterized by the occurrence of larger events (smaller b-values).
Session:Advances in Seismic Imaging and Monitoring of Time-Dependent Variations: Civil Structures, Near-Surface, and Shallow Crustal Scales
Presenter   Bertrand, Etienne
Schedule   Wed 11:15 AM / Oral
Room   Room 3
Seismological Monitoring of a Tall RC Building in Nice, France
BERTRAND, E., CEREMA, DTer Med, Nice, France, etienne.bertrand@cerema.fr; DESCHAMPS, A., GEOAZUR, CNRS-UNS, Sophia-Antipolis, France, deschamps@geoazur.unice.fr; SANTISI D'AVILA, M. P., Université de Nice, Nice, France, msantisi@unice.fr; GUEGUEN, P., Isterre, Grenoble, France, philippe.gueguen@ujf-grenoble.fr; TAHMI, N., GEOAZUR, CNRS-UNS, Sophia-Antipolis, France; FERNANDEZ-LORENZO, G., GEOAZUR, CNRS-UNS, Sophia-Antipolis, France, guillermo.fernandez-lorenzo@unice.fr
The main goal of the National Building Array Program (NBAP), supported by both the French Ministery of Ecology, Sustainable Development and Energy and the French Ministery of Land Planing and Housing and managed by the French Accelerometric Network (RAP) is the seismic response analysis of typical French buildings for seismic vulnerability assessment purposes. In the frame of NBAP, GEOAZUR laboratory and CEREMA institute set up a seismological array in a tall reinforced concrete (RC) building raised in 1979 and located in a sedimentary basin in the city of Nice, southeastern France. The building is a 67-meters high tower composed of 20 levels and 2 basements. The structure is made of concrete slabs connected to two kernels that include the staircases and the elevators. Since july 2010, the building has been monitored by 24 accelerometric sensors distributed along the building elevation. This array is recording continuously, authorizing us to analyse both the ambiant vibrations and the ones due to local seismicity. We present here the first results obtained on the dynamic behavior of the building inferred by these data. About 10 earthquakes have been recorded in 2011-2012 with magnitude ranging between 4.0 and 6.1, all of them situated at epicentral distances greater than 40 km from the building. These earthquakes only produced weak motions in the building and didn’t cause any visible damage in the structure. In the longitudinal direction of the building, three modes of deformation are detected with frequencies around 1.2 Hz, 5.3 Hz and 9.3 Hz respectively. These eigenfrequencies show little variation with respect of the considered earthquake. We also investigate the frequency variation of the first mode throughout the year using the ambiant vibration recordings. Our first analyse indicates that this frequency may slightly vary during the year, but this variation is not larger than one standard deviation.
Session:Induced Seismicity
Presenter   Lajoie, Lia
Schedule   Thu 11:45 AM / Oral
Room   Room 1
Seismic Response to Power Production at the Coso and Salton Sea Geothermal Fields, South-Eastern CA: Using Operational Parameters and Relocated Events to Study Anthropogenic Seismicity Rates and Reservoir Scale Tectonic Structure
LAJOIE, L. J., Fugro Consultants, Inc., Lakewood, Colorado, USA, llajoie@fugro.com; O'CONNELL, D. R. H., Fugro Consultants, Inc., Lakewood, Colorado, USA, d.oconnell@fugro.com; BRODSKY, E. E., University of California, Santa Cruz, Santa Cruz, CA, USA, brodsky@pmc.ucsc.edu; CREED, R. J., Fugro Consultants, Inc., Lakewood, CO, USA, r.creed@fugro.com
With increasing demand for geothermal power, it is important to understand how geothermal operations interact with local (and regional) seismicity, and to determine if seismicity rates are predictable from operational parameters. Furthermore, geothermal injection and production strategies can be improved by identifying, locating and characterizing related earthquakes within the tectonic related background seismicity. We focus specifically on the Coso and Salton Sea geothermal fields in California. Local seismicity is statistically declustered into background and aftershock rates, and we demonstrate that the background rate (at both fields) can be approximated during many time intervals at the 90% + confidence level by a linear combination of injection volume and the net extracted volume. We also use relative relocations and focal mechanisms from Yang et al. (2012) to map fault planes within the Coso geothermal field. We use Bayesian S-wave picking of Coso borehole network microearthquake data and fully-nonlinear 3D hypocenter grid searches to obtain 18000+ well located hypocenters in the geothermal field. Coso geothermal seismicity is not diffuse; 87% of the hypocenters occur within 25 m of planes consistent with tectonic processes along the eastern California shear zone. 83% associate with vertical and steeply-dipping conjugate N-NW dextral and NE sinistral strike slip planes, 12% with on normal-slip planes with dips of 35-70 degrees, and 5% with reverse-oblique-slip on steeply-dipping planes. The non-diffuse nature of seismicity suggests that induced events occur preferentially on pre-existing structures and that flow is concentrated in fractures. With minimum horizontal stress oriented at 81 degrees, and no normal faults, in the western part of the field, and 106 degrees and pervasive normal faults in the eastern portion of the field, structures that are favorably oriented in the current stress field appear to be most commonly activated by geothermal operations.
Session:Topics in Seismology: Regional Seismicity and Tectonics
Presenter   Escudero, Christian R.
Schedule   Thu / Poster
Room   Cook/Arteaga
Seismic Studies of the Jalisco Block
ESCUDERO, C. R., Universidad de Guadalajara, Puerto Vallarta, Jalisco, Mexico, escudero.sisvoc@gmail.com; NUÑEZ-CORNU, F. J., Universidad de Guadalajara, Puerto Vallarta, Jalisco, Mexico; GOMEZ, A., Universidad de Guadalajara, Puerto Vallarta, Jalisco, Mexico; GUITIERREZ, Q. J., Universidad de Guadalajara, Puerto Vallarta, Jalisco, Mexico; PINZON, J. I., Universidad de Guadalajara, Puerto Vallarta, Jalisco, Mexico; PEREZ, O. G., Universidad de Guadalajara, Puerto Vallarta, Jalisco, Mexico; ROBLES, F. J., Universidad de Guadalajara, Puerto Vallarta, Jalisco, Mexico; CORDOBA, D., Universidad Complutense de Madrir, Madrid, España
Many societies and their economies endure the disastrous consequences of destructive earthquakes. The Jalisco region at Mexico is exposing to this natural hazard. Scientific knowledge constitutes the only way to avoid or at least to mitigate the negative effects of such events. Accordingly the study of geological and geophysical causes; structural, kinematics and dynamic characteristics; and destructive effects of such events are indispensable. Seismology constitutes a very important tool to study the structure as well as the earthquakes within the area; in this way we are analyzing different data sets. We use seismic data recorded from 2006 to 2008 by the Mapping the Riviera Subduction Zone (MARS) and the Colima Volcano Deep Seismic Experiment (CODEX) deployed in western Mexico. Additionally, we are analyzing data continually recorded by the Jalisco Seismic Network (RESAJ). We are also deploying another seismic experiment to characterize the seismic and tsunamic risk associated to the Rivera Plate and Jalisco Block structure (TSUJAL). We present a description of these seismic experiment as well as preliminary results that include seismicity maps, local and teleseismic tomography as well as an analysis of this result and the plan for future research.
Session:Seismic Imaging as USArray Moves to Alaska (IRIS/Seismological Society of America Joint Session)
Presenter   Abers, Geoffrey
Schedule   Thu 1:45 PM / Oral
Room   Room 4
Seismic Imaging along a Dense Broadband Transect through Central Alaska
ABERS, G. A., Lamont-Doherty Earth Observatory of Columbia University, Palisades, NY, USA, abers@ldeo.columbia.edu; CHRISTENSEN, D. H., Geophysical Institute, University of Alaska, Fairbanks, AK, USA, doug@giseis.alaska.edu
Alaska has North America’s longest and seismically active subduction zone. Over the last 15 years a series of broadband arrays have been deployed across its eastern end in central Alaska, including the Broadband Experiment Across the Alaska Range (BEAAR) and Multidisciplinary Observations Of Subduction (MOOS), together sampling a 450 km transect at 10-15 km spacing. The arrays provide scattered-wave migration images of the subduction zone. The rupture zone of the great (Mw9.2) 1964 earthquake shows a narrow low-velocity channel without thrust-faulting focal mechanisms, although abundant seismicity lies within the subducting plate. Where the subducting crust reaches mantle depths the thick low-velocity crust of the subducting Yakutat oceanic plateau is imaged. That seismically disappears at 130 km depth with eclogitization. Seismic attenuation shows 10-50x variations within the overlying mantle wedge. A cold high-Q forearc nose abruptly transitions where the plate interface reaches 80 km depth to low-Q, hot and presumably flowing asthenosphere, constraining thermal structure. Shear-wave splitting from arrays and others, from Pacific to Arctic coasts, show Alaska divided in two domains. Fabric rotates 90 degrees where the subducting plate is 80 km deep, i.e. where the mantle wedge transitions from cold to hot. The southern domain shows convergence-parallel fast directions for all plate depths, indicating sub-slab flow or fabric. The northern domain shows slab-parallel fast directions. Because delay times correlate strongly with thickness of the mantle wedge, a dominant component of the northern fabric comes from the mantle wedge, perhaps along-strike flow. However, that fabric continues far into North America where it parallels absolute plate motion. USArray will provide critical data to resolve the interaction between subduction- and plate-scale flow, to see the extent to which subduction leaves an imprint far inland from the trench.
Session:Seismic Imaging as USArray Moves to Alaska (IRIS/Seismological Society of America Joint Session)
Presenter   Obrebski, Mathias
Schedule   Thu / Poster
Room   Cook/Arteaga
Arc Structure around Mount Rainier From the Joint Inversion of Receiver Functions and Seismic Noise
OBREBSKI, M., Columbia University, LDEO, New York, NY, USA, obrebski@ldeo.columbia.edu; ABERS, G., Columbia University, LDEO, New York, NY, USA, abers@ldeo.columbia.edu; FOSTER, A., Columbia University, LDEO, New York, NY, USA, afoster@ldeo.columbia.edu
The amount and modes of magma transfer from the mantle to the crust along volcanic arcs is poorly understood. To better understand deep arc compositional variations, we analyze the shear wave velocity (Vs) distribution in the crust and uppermost mantle for a set of broad-band stations within 35 km of Mount Rainier, WA, in the Cascades arc. We resolve the depth of the main velocity contrasts using receiver function (RF) analysis. To alleviate the trade-off between depth and velocity intrinsic to RF analysis, we jointly invert them with dispersion constraints from earthquake surface waves (period 20-100 s) and ambient noise (5-20s). A transdimensional Bayesian scheme explores the model space (Vs in each layer, number of interfaces and their respective depths), minimizing the number of layers required to fit the observations given their noise level. Sites away from the forearc trend exhibit a middle-to-lower crust with Vs=3.6 km/s, presumably composed of continent-derived sedimentary rocks from the Puget Group. The Moho is sharp and the uppermost mantle has Vs=4.3 km/s, consistent with dry peridotite. In contrast, stations on the arc trend exhibit a rather fast crust over a slow mantle, yielding a weak Moho. We interpret these sites as magmatically modified. Their distribution correlates with the Southern Washington Cascades Conductor (SWCC) interpreted as partially melted mid-crust. The fast middle-to-lower crust (3.8-4.0 km/s) below the SWCC is consistent with mafic composition, presumably resulting from magma intrusion or chemical differentiation, either in the modern arc or its predecessor. A partially melted peridotite can account for the slow velocity in the uppermost mantle (4.0-4.2 km/s). This agrees with recent observation of large conductor connecting the SWCC to the top of the subducting slab below Mt Rainier, interpreted as flux melting fed by the slab dehydration.
Session:Advances in Understanding Earthquake Hazard in Central and Eastern North America
Presenter   Ebel, John
Schedule   Fri / Poster
Room   Cook/Arteaga
Earthquake Swarms Associated with the Saratoga-McGregor Fault System near Albany, NY
EBEL, J. E., Weston Observatory, Boston College, Weston, MA, USA, ebel@bc.edu; JACOBI, R. D., University of Buffalo, Buffalo, NY, USA, rdjacobi@buffalo.edu; O'HARA, A., University of Buffalo, Buffalo, NY, USA, aohara999@gmail.com; STARR, J. C., Weston Observatory, Boston College, Weston, MA, USA, starrjb@bc.edu
Swarms of small earthquakes took place about 25 km southwest and west of Albany, NY in 2007, 2009, 2010 and 2011, with the 2011 swarm starting about 1 day before the 2011 M5.8 Mineral, Virginia earthquake and continuing for about 110 hours after the Mineral event. Relative locations of each of these swarms reveal consistent spatial trends. The 2011 swarm consisted of 11 events ranging from M2.7 to M0.9. The relative locations trend about 350 m in a NE-SW orientation. The 2010 swarm was comprised of 13 events from M2.7 to M0.4, and it trended about 800 m in approximately an ESE-WNW orientation. A swarm of 8 events from M2.8 to M1.6 in early 2009 trended about 400 m in an NE-SW orientation, whereas 4 events with M3.0 to M1.6 took place in late 2009 with about a 5 km trend in an NE-SW orientation. The 2007 swarm involved 4 events with M2.7 to 1.2 and a possible trend of about 200 m in an NE-SW orientation. Each of these swarms was located in a different place, spanning about 16 km in a NE-SW orientation. The events in these swarms occurred on strike of and within a few kilometers of the extrapolated trace of the NNE-striking Saratoga McGregor (S-M) fault. Outcrops near the epicenters of the 2011 swarm exhibit NNE-striking fracture intensification domains with minor offset in the Ordovician Schenectady Formation. The 2010 swarm is also located on a WNW-trending lineament and has internal WNW-trends, consistent with WNW-trending faults similar to those in the Mohawk Valley. Possible focal depths for each of these swarms range from 8 to 18.5 km. These depths make correlation with specific surface structures tenuous, although seismic reflection data and fault traces through the Precambrian of the Adirondack Dome suggest that the S-M fault might be nearly vertical well down into the Precambrian. This conclusion is supported by the focal mechanism of an M2.6 earthquake on August 25, 2013 near the S-M fault about 60 km north of Albany.
Session:Site Response: From Site-Specific Analyses to Predictive Models Around the Globe
Presenter   Régnier, Julie
Schedule   Wed 10:45 AM / Oral
Room   Room 4
PRENOLIN Project: A Benchmark on Numerical Simulation of 1D Non-Linear Site Effect. 1 - Site Selection for the Validation Phase
REGNIER, J., CEREMA, DTer Méditerranée, Nice, France, julie.regnier@cerema.fr; KAWASE, H., DPRI, Kyoto, Japan, kawase.hiroshi.6x@kyoto-u.ac.jp; BONILLA, L. F., Université Paris-Est, IFSTTAR, Marne-La-Vallée, France, luis-fabian.bonilla-hidalgo@ifsttar.fr; BARD, P. Y., ISTerre, Grenoble, France, pierre-yves.bard@ujf-grenoble.fr; BERTRAND, E., CEREMA, DTer Méditerranée, Nice, France, etienne.bertrand@cerema.fr; HOLLENDER, F., CEA,ISTerre, Cadarache, France, fabrice.hollender@cea.fr; SICILIA, D., EDF, Aix-en-Provence, France, deborah.sicilia@edf.fr; NOZU, A., PARI, Yokosuka, Japan, nozu@pari.go.jp
The main objective of the PRENOLIN project is the assessment of the uncertainties associated to non-linear simulation of 1D site effects. An international benchmark is underway to test several numerical codes including various non-linear behavior models to compute the non-linear seismic site response. A preliminary verification phase (i.e. comparison between numerical codes on simple, idealistic cases) will be followed by a validation phase comparing the predictions of such numerical estimations with actual strong motion recordings obtained on well-known sites. The benchmark presently involves 23 different teams and 28 different non-linear computations. The site selection for the validation phase has been done within the KiK-net and the PARI (Port and AirPort Research Institute) Japanese networks. The purpose of this presentation is to detail and discuss the criteria that were used for this slection: (1) Sites must have recorded strong and weak events. (2) Sites must have a 1D geometrical configuration (i.e., horizontal layering): in the linear range the numerical and instrumental site response must agree. (3) The down-hole station must not be too far from the surface (depth <250 m). (4) the sites must have a large non-linear soil behavior and preferentially not associated to pore pressure effects (such as cyclic mobility or liquefaction) to allow the consideration of total stress numerical codes and (5) the local conditions should allow to drill complementary holes to measure the actual NL. Three sites were selected out of an initial selection of about 50 : one from KiK-net (KSRH10) and two from PARI (Onahama and Sendai). Specific in-situ and laboratory tests are being performed at the selected sites to have a complete characterization of the linear and non-linear site parameters. PRENOLIN is part of two larger projects: SINAPS@, funded by the ANR, the French national research agency, and SIGMA, funded by a consortium of nuclear operators (EDF, CEA, AREVA, ENL).
Session:Diverse Mechanisms of Subduction Zone Fault Slip: Exploring the Relationships Among Seismic Rupture, Transient Slip, and Steady Creep
Presenter   Abers, Geoffrey
Schedule   Wed 8:30 AM / Oral
Room   Room 1
What Do Seismic Imaging Constraints Really Tell Us about the Fault Zone in Which Great Earthquakes and Slow Slip Operate?
ABERS, G. A., Lamont-Doherty Earth Observatory of Columbia University, Palisades, NY, USA, abers@ldeo.columbia.edu; KIM, Y. H., Seoul National University, Seoul, Korea, younghkim@snu.ac.kr; SHILLINGTON, D. J., Lamont-Doherty Earth Observatory of Columbia University, Palisades, NY, USA, djs@ldeo.columbia.edu; SAFFER, D. M., Pennsylvania State University, University Park, PA, USA, dms45@psu.edu; LI, J., Lamont-Doherty Earth Observatory of Columbia University, Palisades, NY, USA, jiyao@ldeo.columbia.edu; JANISZEWSKI, H. A., Lamont-Doherty Earth Observatory of Columbia University, Palisades, NY, USA, helenj@ldeo.columbia.edu
Several lines of evidence suggest that subduction zone thrusts lie within weak, possibly overpressured channels. Seismic reflection data often shows a relatively thin, high-reflectivity surface with occasional bright spots, indicative of rapidly varying impedance contrasts over length scales of tens of meters. Scattered coda of teleseismic P waves, such as in receiver functions, often show a thin low-velocity layer corresponding to the top of the subducting plate. These have been best documented in Cascadia, where a 2-4 km thick very low velocity channel is seen above a moderately slow subducting crust; similar structure has been seen in Alaska. The channels are characterized by elevated Vp/Vs ratios (>2.0), and extend both throughout the locked, seismogenic fault zone into the region of episodic tremor and slip. Commonly, this combination of low velocities and high Vp/Vs is taken to indicate high pore pressures, and hence a fault zone that can withstand only very low shear stresses. However, models of these signals indicate static porosities of 2-5% throughout a 2-4 km thick layer, extending to depths of 40 km, a situation that seems difficult to sustain. Highly elevated pore fluid pressure may not be a unique explanation for the seismic observations. At both the Alaska and Cascadia margins, the low seismic wavespeeds, high Poisson’s ratios, and high anisotropies result in part from subduction of thick sediments. The presence of significant thickness of subducted and underplated sediment is consistent with observations of preserved subduction “channels” in exhumed rocks. Although some elevation of pore pressure may be still needed to explain observations, if subduction of 2-4 km sediment is a significant factor in generating the seismic signatures, then the geophysical observations would reflect a much stronger thrust zone than one sustained by high pore pressure alone.
Session:Merging Paths: Earthquake Simulations and Engineering Applications
Presenter   Lin, Ting
Schedule   Thu 4:45 PM / Oral
Room   Room 3
Rupture to Rafters to Response: Completing the Loop of Earthquake Science, Engineering and Policy
LIN, T., Marquette University, Milwaukee, WI, USA, ting.lin@marquette.edu
Ground motion definition is the link between seismic hazard and structural response, the first two elements of performance-based earthquake engineering. Site- and structure-specific ground motion selection is enabled by improved ground motion prediction and considerations of important seismic parameters for nonlinear dynamic analyses. Ground motion databases are growing, with denser instrumentation that provides more empirical recordings, high performance computing that facilitates large-scale simulations, and breakthroughs in geophysical understanding that push the frontiers to the high-frequency range. Structural models are also evolving with enhanced accuracy and complexity to capture important structural behaviors such as collapse, cumulative damage and “in-cycle” strength and stiffness degradation. In the past decade, significant progress has been made in the performance-based earthquake engineering framework to interpret structural performance results in terms of structural response, damage and loss. At the same time, Great ShakeOut Earthquake Drills are extending beyond California to improve earthquake preparedness, with over 24.7 million participants worldwide in 2013. The advancements in hazard characterization, structural modeling, performance interpretation, and emergency response offer a unique opportunity to integrate these elements to complete the loop of earthquake science, engineering and policy to reduce risk.
Session:Advances in Seismic Imaging and Monitoring of Time-Dependent Variations: Civil Structures, Near-Surface, and Shallow Crustal Scales
Presenter   Li, Zhiwei
Schedule   Wed / Poster
Room   Cook/Arteaga
Site Characterization for Marine Earthquake Engineering via Modeling of Ocean-Bottom-Seismometer Observations of Rayleigh Waves and Microtremors
LI, Z. W., Institute of Geodesy and Geophysics, Chinese Academy of Sciences, Wuhan, China, zwli@whigg.ac.cn; HAO, T. Y., Institute of Geology and Geophysics, Chinese Academy of Sciences, Beijing, China; NI, S. D., Institute of Geodesy and Geophysics, Chinese Academy of Sciences, Wuhan, China; YOU, Q. Y., Institute of Geology and Geophysics, Chinese Academy of Sciences, Beijing, China; LIU, L. H., Institute of Geology and Geophysics, Chinese Academy of Sciences, Beijing, China; BAO, F., Institute of Geodesy and Geophysics, Chinese Academy of Sciences, Wuhan, China; XU, Y., Institute of Geology and Geophysics, Chinese Academy of Sciences, Beijing, China
Shallow shear-wave velocity structure of the top tens to hundreds of meters is essential for site characterization. On land, both active and passive methods are used to determine subsurface velocity structures. With the increasing of ocean industry engineering (e.g., marine platform for petroleum exploitation), shallow shear-wave velocity structure in marine region need to be well investigated. However, it is not a trivial task to study the shallow shear-wave structure in the marine regions. With the ocean-bottom-seismometer (OBS) observations in the Bohai Sea, China, from the year of 2010 and 2011, high quality waveforms of short-period Rayleigh wave are obtained with air-gun source at close to large offset (from 200 m to >6000 m). The dense spacing of air-gun sources (~200 m) and high signal-to-noise Rayleigh waves provide a feasible way to investigate the shallow-shear wave structure beneath the OBS observation lines. Moreover, the horizontal-to-vertical (H/V) spectrums are calculated with the microtremors recorded by the OBS, which is also used to estimate the shear-wave velocity structure beneath the OBS. Shear-wave velocities of 120~220 m/s are obtained from both dispersion curves of Rayleigh waves within 1~5 Hz and H/V spectrum ratio. Our study suggests that HV method is also applicable in marine regions.
Session:Pillars of Simulation: Seismic Velocity and Material Models
Presenter   Allam, Amir
Schedule   Wed 5:30 PM / Oral
Room   Room 4
Finite-frequency Sensitivity of Seismic Waves to Fault Zone Structures
ALLAM, A. A., Scripps Institute of Oceanography, La Jolla, CA, USA, aallam@ucsd.edu; TAPE, C., University of Alaska Fairbanks, Fairbanks, AK, USA, carltape@gi.alaska.edu; BEN-ZION, Y., University of Southern California, Los Angeles, CA, USA, benzion@usc.edu
We analyze the volumetric sensitivity of fault zone seismic head and trapped waves by constructing finite-frequency Fréchet kernels for these phases based on idealized and realistic velocity models. We first validate numerical calculations by waveform comparisons with analytical results for a model with a bimaterial fault interface separating different elastic solids. After establishing numerical accuracy up to 10Hz, we construct sensitivity kernels for a ‘vertical sandwich’ model where a narrow low-velocity region is surrounded on either side by a higher velocity material, along with a suite of velocity models based on detailed tomographic results from the San Jacinto Fault Zone. In contrast to P body waves, which have little or no sensitivity to fault zone structure, the sensitivity kernels for head waves have sharp peaks with high values near the fault in the faster medium. Surface wave kernels show the broadest spatial distribution of sensitivity, while trapped wave kernels are extremely narrow with sensitivity focused entirely inside the low-velocity fault zone layer. Taken together, these phases contain complementary information about fault zone and regional velocity structure. These results indicate that adjoint tomography using high-frequency head and trapped waves kernels, combined with traditional body and surface wave kernels, can constrain fault zone structure at higher range of scales than has previously been possible.
Session:From the Earthquake Source to Damage of Buildings: Bridging the Gap between Seismology and Earthquake Engineering
Presenter   Ktenidou, Olga-Joan
Schedule   Fri 9:30 AM / Oral
Room   Room 1
Squeezing Kappa out of the Transportable Array: When the Going gets Tough
KTENIDOU, O. J., Université Joseph Fourier, CNRS, ISTerre and PEER, UC Berkeley, Grenoble, France, olga.ktenidou@ujf-grenoble.fr; SILVA, W., Pacific Engineering and Analysis, El Cerrito, CA, USA, pacificengineering@juno.com; DARRAGH, B., Pacific Engineering and Analysis, El Cerrito, CA, USA, pacificengineering@juno.com; ABRAHAMSON, N. A., Department of Civil and Environm, Berkeley, CA, USA, abrahamson@berkeley.edu; KISHIDA, T., Pacific Earthquake Engineering R, Berkeley, CA, USA, kishidapple@gmail.com; COTTON, F., Université Joseph Fourier, CNRS,, Grenoble, France, fabrice.cotton@ujf-grenoble.fr
The high-frequency attenuation parameter κ (Anderson and Hough, 1984), and, in particular, its site-specific component (κ0) is an important input parameter for ground motion prediction and simulations. The frequencies it controls (above 10-20 Hz) are important to the seismic response of safety-related equipment in nuclear facilities. We present a case where the seismicity and instrument characteristics render its measurement a challenge. Our site is located in a low seismicity region (S. Arizona). The available seismic records are few and often noisy. They are at distances between 10-300 km, which makes the trade-off between site and path attenuation significant. The event magnitudes (M1.2-M3.4) are rather low for the classic κ estimation method to be used. The stress drop values from different studies in the region vary greatly (1-50 bar), leading to large uncertainty in the source corner frequencies. Hence κ should be measured ideally above 10 Hz (above M3) or 20 Hz (below M2) to avoid trade-offs. Possible high-frequency resonances due to shallow soil layers may also interfere with the measurement. However, because our data comes from the Transportable Array, the low sampling rate limits the maximum usable frequency to 16 Hz. This allows us very little bandwidth to resolve source, path, and site effects and constrain κ0. We use three measurement approaches (above the corner frequency, below it, and across the entire frequency range) to define upper and lower bounds for κ0 in S. Arizona, as well as estimates of regional Q and stress drop. The TA has greatly increased the available dataset for North America and, in certain low-seismicity regions, represents the majority of available seismic records at short distances. The severe obstacles faced in this study will be relevant in future κ studies in other regions with such band-limited data. We believe that an increase in the sampling rate of the TA would help avoid them.
Session:Development of Next Generation Field Methods for Portable Broadband Seismic Arrays
Presenter   Bent, Allison
Schedule   Wed / Poster
Room   Cook/Arteaga
Crustal Structure of the St. Lawrence Corridor from Montreal to the Charlevoix Seismic Zone from a Temporary Broadband Array
BENT, A. L., Geological Survey of Canada, Ottawa, ON, Canada; KAO, H., Geological Survey of Canada, Sidney, BC, Canada
While many of the most seismically active zones of eastern Canada lie along the St. Lawrence River, the region is not uniformly active. The three primary seismic zones (Charlevoix, West Quebec and Lower St. Lawrence) are separated by regions of very low seismicity. The reasons for the different levels of activity are not well understood. To determine whether there are physical or structural differences which might explain the difference in activity rates temporary broadband seismometers were installed along the St. Lawrence River in the region from the Charlevoix Seismic Zone, northeast of Quebec City, to Montreal to fill in gaps in broadband coverage by the Canadian National Seismograph Network (CNSN). Stations were typically installed for about two years and then moved to new locations. Thirteen temporary sites as well as the CNSN stations in the region of interest have been evaluated. The crustal structure beneath the stations was modeled using teleseismic receiver function analysis. Preliminary models were obtained using a simple linear inversion. Once the station closed and no new data were forthcoming a Neighborhood Algorithm inversion was performed to provide a more exhaustive search of the model space and a quantitative ranking of the relative fits of the models tested. Regional scale features such as a crustal discontinuity at about 20 km depth and Moho at 35-40 km can be correlated from one station to another. The resulting structural models do not point to any significant difference between the seismic and aseismic regions, suggesting that perhaps the differences are temporal although small scale structural differences cannot be completely ruled out. As the US Transportable Array moves into the study area stations close to the St. Lawrence will be analyzed to further fill in gaps in coverage.
Session:Recent Advances and Findings in Earthquake Geology and Paleoseismology
Presenter   Bennett, Scott
Schedule   Fri / Poster
Room   Cook/Arteaga
Preliminary Paleoseismic Trenching Results from the Flat Canyon Site, Southern Provo Segment, Wasatch Fault Zone: Testing Holocene Fault-Segmentation at the Provo-Nephi Segment Boundary
BENNETT, S. E. K., U.S. Geological Survey, Golden, CO, USA, sekbennett@usgs.gov; DUROSS, C. B., Utah Geological Survey, Salt Lake City, UT, USA, christopherduross@utah.gov; GOLD, R. D., U.S. Geological Survey, Golden, CO, USA, rgold@usgs.gov; BRIGGS, R. W., U.S. Geological Survey, Golden, CO, USA, rbriggs@usgs.gov; PERSONIUS, S. F., U.S. Geological Survey, Golden, CO, USA, personius@usgs.gov; MAHAN, S. A., U.S. Geological Survey, Lakewood, CO, USA, smahan@usgs.gov
Paleoseismic data near fault segment boundaries provide direct information about fault rupture segmentation. The 350 km-long Wasatch fault zone (WFZ), the archetype of segmented normal faults, consists of 10 structural segments. Abundant paleoseismic data support a history of segmented Holocene surface ruptures, but recent findings document at least one case where an earthquake ruptured across a WFZ structural segment boundary. The extent and frequency of ruptures that span segment boundaries remains poorly known, in part because most paleoseismic studies are sited in segment interiors, adding uncertainty to seismic hazard models for this heavily populated region of Utah. To address these unknowns and reduce this uncertainty we have begun a paleoseismic trenching campaign targeting WFZ structural segment boundaries. We excavated a trench at Flat Canyon (Salem, UT), near the southern end of the Provo segment in the complex Provo-Nephi segment boundary, a 5–8 km-wide right stepover. Alluvial fan deposits at the site are displaced across a 13 m-high scarp. We document a minimum of four and maximum of seven earthquakes within a 17–20 m-wide fault zone, consisting of two graben systems within the scarp. The lower graben preserves evidence for four earthquakes and ≥5 m of vertical throw. The upper graben preserves evidence for two to three earthquakes with ~1 m of total throw. Ongoing optically stimulated luminescence and radiocarbon analyses will provide earthquake timing constraints and allow us to correlate earthquakes between the grabens. Our goal is to determine whether earthquakes at the Flat Canyon site correspond with earthquakes at several paleoseismic sites farther north on the Provo segment and/or with earthquakes at the northernmost sites on the adjacent Nephi segment to the south. Comparison of these earthquake chronologies will allow us to test whether Holocene earthquakes have ruptured across the Provo-Nephi segment boundary.
Session:Merging Paths: Earthquake Simulations and Engineering Applications
Presenter   Denolle, Marine
Schedule   Thu / Poster
Room   Cook/Arteaga
The Ambient Seismic Field Captures Complex Sedimentary Basin Effects
DENOLLE, M. A., Scripps Institute of Oceanography, La Jolla, CA, USA, marinedenolle@gmail.com; MIYAKE, H., Earthquake Research Institute - University of Tokyo, Tokyo, Japan, hiroe@eri.u-tokyo.ac.jp; NAKAGAWA, S., Earthquake Research Institute - University of Tokyo, Tokyo, Japan, nakagawa@eri.u-tokyo.ac.jp; HIRATA, N., Earthquake Research Institute - University of Tokyo, Toyko, Japan, hirata@eri.u-tokyo.ac.jp; BEROZA, G. C., Stanford University, Stanford, CA, USA, beroza@stanford.edu
Seismic hazard in urban environments can be strongly affected for cities that are situated in sedimentary basins. Metropolitan Tokyo is subject to high seismic hazard because it is located above the active junction of three major plates, and because the Kanto sedimentary basin underlies the area. The Kanto basin is fairly deep, and has poorly consolidated, low wave speed sediments, which yield a resonance frequency that approaches the natural period of very tall buildings. It is difficult to model basin effects with confidence because complexities of basin excitation and resonance depend on the details of three-dimensional basin structure, which are incompletely known. The ambient seismic field contains wave propagation information, and provides a new approach for characterizing basin effects in ground motion. We use the ambient field to construct the Earth’s response to an impulsive force by combining data from MeSO-net (the Metropolitan Seismic Observatory Network), which consists of 296 shallow borehole instruments in the Kanto basin to act as receivers, and 450 stations from the Hi-net (High Sensitivity Seismograph Network), which are deep borehole stations, to act as virtual sources. We find that at the natural resonance of 5-6 s, the amplification is proportional to sediment thickness. We also explore the impact of the basin shape on the wave field. We find, for example, strong Rayleigh-to-Love wave conversions at the basin edge, and evidence for basin-edge effects in propagation as well.
Session:A Decade of Great Subduction Earthquakes – What Have We Learned From Their Ground-Motions?
Presenter   Frankel, Arthur
Schedule   Fri 9:00 AM / Oral
Room   Room 4
Broadband Synthetic Seismograms for M9 Cascadia Earthquakes Using 3D Simulations
FRANKEL, A., U.S. Geological Survey, Seattle, WA, USA, afrankel@usgs.gov; PHILLIPS-ALONGE, K., U.S. Geological Survey, Pasadena, CA, USA, kphillips-alonge@usgs.gov; STEPHENSON, W., U.S. Geological Survey, Golden, CO, USA, wstephens@usgs.gov; DELOREY, A., Los Alamos National Laboratory, Los Alamos, NM, USA, andrew.delorey@lanl.gov
We are producing broadband (0-10 Hz) synthetic seismograms of M9 earthquakes on the Cascadia subduction zone (CSZ) for various rupture scenarios. At long periods (> 2 s) we are using synthetics from 3D finite-difference simulations for the region west of the Cascade Mountains and extending the 1100 km length of the CSZ. This 3D velocity model includes the dipping subducted plate and sedimentary basins, such as the Seattle, Tacoma, Portland, and Tualatin basins. We use earthquake slip models consisting of two parts: (1) random slip with a large correlation distance (about 100 km) and a specified wave number spectrum and (2) discrete asperities with high stress drops and short rise times of slip. For the first part of the model, an average slip velocity was specified to determine rise times of slip on the fault. The slip velocity was calibrated by comparing response spectral values of synthetics for the M8.8 Maule, Chile earthquake with the observed values. At short periods (< 2 s) we use a summation of stochastic seismograms from point sources. The short and long-period synthetics are combined using matched filters. We constrained the stress drop of the discrete asperities for the Cascadia simulations by modeling the observed spectral accelerations and waveforms of the M9.0 Tohoku, Japan earthquake and the Maule earthquake. For the Tohoku earthquake, the observed spectral accelerations from 10 s to 0.1 s can be explained by high stress drop sub-events located in the deeper portion of the rupture zone. The Maule earthquake consisted of at least three sub-events with high stress drops. We are comparing the spectral accelerations from the Cascadia synthetics with ground-motion prediction equations used in seismic hazard assessments. The synthetics exhibit large amplification of spectral accelerations at 3-5 s by the Seattle basin.
Session:Merging Paths: Earthquake Simulations and Engineering Applications
Presenter   Shahjouei, Alireza
Schedule   Thu / Poster
Room   Cook/Arteaga
STUDENT
Hybrid Low-High Frequency Ground Motion Time Histories for Central U.S.
SHAHJOUEI, A., Department of Civil Engineering, The University of Memphis, Memphis, TN, USA, shhjouei@memphis.edu; PEZESHK, S., Department of Civil Engineering, The University of Memphis, Memphis, TN, USA, spezeshk@memphis.edu
A hybrid broadband (HBB) simulation technique is applied to generate a suite of appropriate time histories in the central United States. Long period portion of the simulated ground motions are achieved through a kinematic model of the earthquake source and a deterministic wave propagation. A discrete wavenumber-finite element technique is used to obtain the wave displacement and the velocity time series in the zero to intermediate frequency range. Applying the kinematic source model, the near field effects and the forward directivity in the long period synthetics is captured for a set of station along the strike of the fault. The point source stochastic synthetic, using the computer program SMSIM (Boore, 2005), is used to compute the high frequency part of the synthetic ground motions. Finally, hybrid synthetics are obtained by implementing matched second order low-pass and high-pass Butterworth filters for the long period and the high frequency synthetics, respectively. The discussed HBB simulation technique is used to generate ground motions for earthquake magnitudes ranging from 4.5 to 7.5 for the central United States region. The HBB ground motion spectral accelerations at 0.2 and 1.0 seconds are compared with the existing ground motion prediction equations (GMPEs) for the central United States region.
Session:Recent Advances in Ground Motions Simulation Methods and Their Validation
Presenter   Graves, Robert
Schedule   Thu / Poster
Room   Cook/Arteaga
Refinements to the Graves and Pitarka (2010) Broadband Simulation Method
GRAVES, R. W., US Geological Survey, Pasadena, CA, USA, rwgraves@usgs.gov
Over the past several years, the performance of several ground motion simulation methods has been examined on the SCEC Broadband Simulation Platform (BBP). The analysis utilizes two complementary validation approaches: Part A compares simulated ground motions with recordings from past earthquakes and Part B compares median ground motion prediction equation results for hypothetical earthquakes with predictions from the BBP methods (Goulet et al., 2013). Here, I describe refinements to the Graves and Pitarka (2010) simulation method that have been guided by these studies. One refinement involves the addition of a deep weak zone to the rupture characterization, analogous to the shallow weak zone introduced by Graves and Pitarka (2010). The deep zone begins at 15 km and is characterized by a two-fold increase in the average rise time up to a fault depth of 18 km. Below 18 km, a constant factor of two increase applies. This is accompanied by a reduction in the width of slip tapering along the fault edges. These modifications reduce the radiation of strong motion energy from the deep fault while still allowing large fault displacement to occur in this zone. The result is to reduce the tendency of the simulations to over-predict observed ground motion levels in the period range of 1 to 5 sec, particularly for larger magnitude ruptures (Mw >6.5). A second refinement to the methodology is to add perturbations to the correlation structure for rise time and rupture speed parameterization. The original approach directly correlated variations of rise time and rupture speed with variations in local fault slip. Here, the correlation structure is relaxed such that the variations are randomly selected from a log-normal distribution having a median value that follows the specified correlation with local slip. The effect is to reduce the coherency of radiated motions around 1 to 2 sec period and provide a smoother transition into the stochastic approach used at shorter periods.
Session:Recent Advances in Ground Motions Simulation Methods and Their Validation
Presenter   Frankel, Arthur
Schedule   Thu 8:30 AM / Oral
Room   Room 3
Three-Dimensional Ground-Motion Simulations of Earthquakes in the Hanford, Washington Area
FRANKEL, A., U.S. Geological Survey, Seattle, WA, USA, afrankel@usgs.gov; THORNE, P., Pacific Northwest National Laboratory, Richland, WA, USA, paul.thorne@pnnl.gov; ROHAY, A., Pacific Northwest National Laboratory, Richland, WA, USA, alan.rohay@pnnl.gov
Observations and 3D finite-difference simulations for the Hanford area of eastern Washington state show that basin-edge generated surface waves at frequencies up to 2 Hz amplify and prolong the ground shaking compared to that expected from flat-layered velocity models. The Hanford site is a former nuclear production complex with large amounts of stored radioactive waste. The 3D model consists of a sedimentary basin with a relatively thin (130 m or less) layer of sediments above a thick sequence of basalts that overlie pre-Miocene sediments and basement. Recordings in the basin of a M3.7 earthquake at 20 km depth south of the basin show a prominent set of dispersed arrivals at 1-2 Hz following the S-wave. Synthetics derived from the 3D model match the observed later arrivals, whereas synthetics from flat-layered velocity models do not contain these arrivals. The simulations show that these later arrivals were high-frequency surface waves propagating in the sediments above the basalt, created by conversion of incident S-waves at the southern edge of the basin. 3D and 1D simulations were done for M6.7-6.8 earthquakes on the Rattlesnake Hills fault along the southern edge of the basin and M6.6 earthquakes on the Gable Mountain fault beneath the basin. The synthetics from the 3D model at some sites have spectral accelerations (SA) at 1-2 Hz that are 2 to 4 times higher than those from the 1D model, for earthquakes along the southern edge of the basin. This amplification is due to basin-edge generated surface waves. The 1-2 Hz SA of the synthetics from the 3D model for ruptures near the southern edge of the basin are significantly larger than those predicted from the NGA West 2 ground-motion prediction equations. We estimated the effects of the sedimentary basin on regional S-waves from earthquakes on the Cascadia subduction zone by using a dipping plane-wave source. The simulation for the 3D model produced SA values substantially larger than the flat-layered model.
Session:Development of 2014 U.S. National Seismic Hazard Maps and Their Implementation in Engineering Applications
Presenter   Shahjouei, Alireza
Schedule   Thu / Poster
Room   Cook/Arteaga
STUDENT
Ground Motion Attenuation Relations for Central U.S. Using Hybrid Broadband Synthetics
SHAHJOUEI, A., Department of Civil Engineering, The University of Memphis, Memphis, TN, USA, alireza.shahjouei@gmail.com; PEZESHK, S., Department of Civil Engineering, The University of Memphis, Memphis, TN, USA, spezeshk@memphis.edu
Pezeshk et al. (2011) proposed hybrid empirical ground motion prediction equations (GMPEs) for the central and eastern United States (CEUS) using the next generation attenuation (NGA-West) models and updated seismological parameters. In this study, we will produce a new comprehensive framework for using a hybrid broadband synthetic simulation technique for development of the ground motion attenuation relation in the central US. Synthetics in the low frequency and the high frequency parts of time histories are separately calculated through a kinematic and a stochastic approach, respectively. Hybrid synthetics are obtained by implementing matched second order low-pass and high-pass Butterworth filters for the long period and the high frequency synthetics, respectively. The directivity effects will be captured by applying a kinematic model of the earthquake source and a deterministic wave propagation in the generation of synthetics’ long period portion. GMPEs will be computed from simulations of different hypocenter locations and multiple slip models to consider the variability of the parameters. A new set of GMPEs will be generated for a magnitude range of 5.0 to 7.5 and a distance range of 2 km to 200 km. New questions will be compared with the GMPEs proposed by Pezeshk et al. (2011).
Session:Topics in Seismology: Processes
Presenter   Veitch, Stephen A.
Schedule   Thu / Poster
Room   Cook/Arteaga
STUDENT
Analysis of High-Frequency Icequakes at a Marine-Terminating Glacier in Greenland
VEITCH, S. A., Lamont-Doherty Earth Observatory, Palisades, NY, USA, veitch@ldeo.columbia.edu; NETTLES, M., Lamont-Doherty Earth Observatory, Palisades, NY, USA, nettles@ldeo.columbia.edu
Rapid deformation of fast-moving glaciers leads to the emission of seismic waves from sources internal to the glacier, at the interface between the glacier and the underlying rock or till, and at the glacier's terminus, over a wide-range of frequencies. Small, high-frequency "icequakes" are the most abundant glacial seismic sources, and provide information about glacier dynamics that complement flow data obtained at the glacier's surface using geodetic and remote-sensing methods. At Greenland's large outlet glaciers, long-period signals produced by "glacial earthquakes" during major calving events have been studied systematically. However, little is known about the high-frequency seismic sources within and just beneath these glaciers. We deployed a temporary array of six intermediate-period seismometers around Helheim Glacier in East Greenland with a sampling rate of 100 sps, complemented by a longer-operating broadband station sampling at 20 sps. Helheim Glacier is Greenland's third-largest outlet glacier, and our seven-week dataset contains tens of thousands of icequake signals. We analyze these signals to obtain timing and location information of discrete events, and to consider the relationship between glacier seismicity and other observable glacier-dynamic changes, and we apply our results to better understand the role of brittle deformation in the dynamics of these very large Greenland outlet glaciers. As fast-flowing glaciers such as Helheim Glacier are a primary means of mass-loss for the Greenland Ice Sheet, understanding their dynamics is an important part of the study of ongoing changes in the Arctic.
Session:From the Earthquake Source to Damage of Buildings: Bridging the Gap between Seismology and Earthquake Engineering
Presenter   Grapenthin, Ronni
Schedule   Fri 1:45 PM / Oral
Room   Room 1
Real-time GPS enhanced Earthquake Early Warning: The Northern California Setup
GRAPENTHIN, R., Berkeley Seismological Lab, University of California, Berkeley, Berkeley, CA, USA, ronni@seismo.berkeley.edu; JOHANSON, I. A., Berkeley Seismological Lab, University of California, Berkeley, Berkeley, CA, USA, ingrid@seismo.berkeley.edu; ALLEN, R. M., Berkeley Seismological Lab, University of California, Berkeley, Berkeley, CA, USA, rallen@berkeley.edu
In an effort to improve earthquake parameter estimation in earthquake early warning (EEW) for large earthquakes (e.g. Mw, finite fault geometry), the Berkeley Seismological Laboratory (BSL) works on integrating real-time GPS into EEW, and generates and archives real-time position estimates using data from 62 GPS stations in the greater San Francisco Bay Area. This includes 26 stations that are operated by the BSL as part of the Bay Area Regional Deformation (BARD) network, 8 operated by the USGS, and 29 stations operated by the Plate Boundary Observatory. We set up a fully triangulated network scheme in which neighboring station pairs are processed with the software trackRT. Positioning time series are produced operationally for about 170 station pairs; additional station pairs will be added as more real-time stations become available. G-larmS, the geodetic alarm system, sits on top of real-time GPS processors (e.g., trackRT) and analyzes real-time positioning time series, and determines and broadcasts static offsets and quality parameters from these. Following this, G-larmS derives fault and magnitude information from the static offsets and broadcasts these results as well. This Python implementation is integrated into seismic alarm systems (CISN ShakeAlert, ElarmS) as it uses their P-wave detection alarms to trigger its processing. The BSL now runs G-larmS in real-time test-mode for Northern California. Here, we present the setup and report results of the first months of operation. This includes analysis of system latencies, noise, and response to actual events and results produced by G-larmS. As the BSL generates differential positions over relatively short baselines, we also test how such baseline strains compare to absolute position changes in the case of a very large large earthquake. We perform this analysis on data from the 2011 Mw 9.0 Tohoku-oki earthquake, add randomly selected real-time noise, and invert for slip along the subduction zone interface.
Session:Site Response: From Site-Specific Analyses to Predictive Models Around the Globe
Presenter   Hayashi, Koichi
Schedule   Wed / Poster
Room   Cook/Arteaga
Detailed Near Surface Velocity Model and Site Response of the Hayward, East San Francisco Bay Area: Seismic Observation of a Building Implosion and Its Interpretation from Site Response Point of View
HAYASHI, K., Geometrics, Cupertino, CA, USA, KHayashi@geometrics.com; CRAIG, M., California State University, East Bay, Hayward, CA, USA, mitchell.craig@csueastbay.edu
A twelve-story building at California State University, East Bay, in Hayward, CA was demolished by implosion in August, 2013. The building was located ~300 m away from the Hayward Fault and seismic ground motion induced by the implosion was recorded to study velocity structure across the Hayward Fault. We deployed five sets of three-component broadband accelerometers across the Hayward Fault and recorded ground motion induced by the implosion. Accelerometers were placed along a line oriented perpendicular to the Hayward Fault with a spacing of approximately 800 m at distances ranging from 800 to 4000 m from the building. The implosion was clearly observed at all stations. The ground motion induced by the implosion lasted ~15 s and propagated with apparent velocity of ~400 m/s. Maximum acceleration was ~4.4 cm/s2 at a distance of 800 m from the building and ~0.3 cm/s2 at a distance of 4000 m. Spectra of all stations are basically flat from 0.5 Hz to 50 Hz, with no obvious peak. The spectrum from each station was divided by the spectrum of the station closest to the building to estimate site amplification. Resultant horizontal spectral ratios clearly show peaks at 1.5 Hz. A representative S-wave velocity cross section was constructed based on the active and passive surface wave methods and theoretical ground motion was calculated using a 2D viscoelastic finite-difference method. The theoretical ground motion shows a clear peak frequency at ~1 Hz and it is generally consistent with the observed implosion record. The S-wave velocity model constructed by the surface wave methods and the observation of the building implosion implies that the low frequency component of ground motion may be locally amplified on the west side of the Hayward Fault because of the effect of two-dimensional structure.
Session:Induced Seismicity
Presenter   Bradley, Chris
Schedule   Thu / Poster
Room   Cook/Arteaga
A Study of the Potential for Induced Seismicity Resulting from CO2 Injection at Kimberlina, Southern San Joaquin Valley, California
BRADLEY, C. R., Los Alamos National Laboratory, Los Alamos, NM, USA, cbradley@lanl.gov; LEE, R. C., Los Alamos National Laboratory, Los Alamos, NM, USA, rclee@lanl.gov; COBLENTZ, D. C., Los Alamos National Laboratory, Los Alamos, NM, USA, cobletz@lanl.gov; WILSON, J. E., Los Alamos National Laboratory, Los Alamos, NM, USA, jenwilson@lanl.gov; STONE, I. P., Los Alamos National Laboratory, Los Alamos, NM, USA, istone@lanl.gov
As part of the National Risk Assessment Program (NRAP) for geologic sequestration of CO2, a study of the potential for induced seismicity and consequent ground motion is being conducted for Kimberlina, a candidate location in the Southern San Joaquin Valley, California. The goal of this program is to assess the effect of the rates and locations of hypothetical CO2 injection in a predictive model to estimate the potential seismic impacts of induced stresses on the existing faulting regime. A probabilistic framework will be used to explore the production of both induced and tectonic earthquakes. This scheme will be used to guide characterization and monitoring programs prior to any injection program. Our study includes accumulating and estimating geologic and geophysical information of the basin structure and shallow velocity structure (e.g. Vs30), characterization of active and potentially active faults, and modeling the effect of the ambient stress field on the faults. Integrating this information, ground motion prediction is estimated for some of the most potentially hazardous faults by using several databases and techniques. First, a review of the historical seismicity provides baseline hazard, then ground motion prediction equations (GMPEs) derived from induced seismicity studies at hydrothermal injection sites are modified to include variations in Vs30. In addition, regional tectonic GMPE results are used to develop ShakeMap estimates for potential earthquakes in the Southern San Joaquin Valley. This research is a multi-laboratory effort in the Department of Energy Complex including Los Alamos, Lawrence Livermore, Lawrence Berkeley National Laboratories and the National Energy Technologies Laboratory.
Session:Advances in Understanding Earthquake Hazard in Central and Eastern North America
Presenter   Wallace, Katherine
Schedule   Fri 8:30 AM / Oral
Room   Room 2
STUDENT
Paleoseismic Evidence for Persistent Intraplate Seismicity Associated with Reactivation of Precambrian Crustal Structures in Central Canada
DOUGHTY, D., McMaster University, Hamilton, ON, Canada, doughty@utsc.utoronto.ca; EYLES, N., University of Toronto Scarborough, Toronto, ON, Canada, eyles@utsc.utoronto.ca; WALLACE, K. E., University of Toronto Scarborough, Toronto, ON, Canada, kwallace@utsc.utoronto.ca; EYLES, C. H., McMaster University, Hamilton, ON, Canada, eylesc@mcmaster.ca; BOYCE, J. I., McMaster University, Hamilton, ON, Canada, boycec@mcmaster.ca
Central Canada experiences recurring low to moderate intraplate earthquakes linked to reactivation of poorly- understood Precambrian basement structures. Unfortunately such structures are deeply buried below Paleozoic cover rocks, Pleistocene glacial and postglacial sediments and large lake bodies. Well defined laterally-extensive seismite horizons composed of ball and pillow structures, thrust faults and debrite horizons have been mapped in lengthy outcrops of early to mid-Silurian carbonate and siliclastic shallow marine strata. These structures record sediment disturbance by moderate paleo-earthquakes in the vicinity of the Grenville Front Tectonic Zone and other deeply buried crustal lineaments, and may be associated with phases of rapid subsidence of the Michigan Basin in response to far-field Appalachian orogenic events. Recent earthquake-related deformation in the last 10,000 years is recorded on high resolution seismic sub-bottom profiles that identify sediment faulting, diapiric deformation and slumping on many lake floors located above Precambrian terrane boundaries. Exceptionally thick mass flow successions and extensive faults in Lake Timiskaming located along the floor of the Ottawa-Timiskaming Graben within the seismically active Western Quebec Seismic Zone, suggest a higher frequency of earthquakes and slope failure during deglaciation and rapid glacio-isostatic rebound, but faulting and failure of lakefloor sediments is ongoing confirming the hypothesis that the graben is a weak zone within the North American plate. Results support the model that ongoing mid-plate earthquake activity is a consequence of brittle deformation of the fractured and faulted upper crust of the North American plate with significant implications for the assessment of seismic hazard in vulnerable intraplate regions with ageing infrastructure.
Session:Emergence of Continuously Recording Very Large Array Capabilities in Seismology
Presenter   Ross, Zachary
Schedule   Wed / Poster
Room   Cook/Arteaga
An Earthquake Detection Algorithm with Pseudo-Probabilities of Multiple Indicators
ROSS, Z. E., University of Southern California, Los Angeles, CA, USA, zross@usc.edu; BEN-ZION, Y., University of Southern California, Los Angeles, CA, USA, benzion@usc.edu
We develop an automatic earthquake detection algorithm combining information from numerous indicator variables in a non-parametric framework. The method is shown to perform well with multiple ratios of moving short- and long-time averages having ranges of time intervals and frequency bands. The results from each indicator are transformed to a pseudo-probability time-series (PPTS) in the range [0, 1]. The various PPTS of the different indicators are multiplied to form a single joint PPTS that is used for detections. Since all information is combined, redundancy among the different indicators produces robust peaks in the output. This allows the trigger threshold applied to the joint PPTS to be significantly lower than for any one detector, leading to substantially more detected earthquakes. Application of the algorithm to a small data set recorded during a 7-d window by 13 stations near the San Jacinto fault zone detects 3.13 times as many earthquakes as listed in the Southern California Seismic Network catalogue. The method provides a convenient statistical platform for including other indicators, and may utilize different sets of indicators to detect other information such as specific seismic phases or tremor.
Session:Explosive Source Characterization
Presenter   Jones, Kyle
Schedule   Wed / Poster
Room   Cook/Arteaga
Coupled Non-linear and Linear Acoustic Wave Propagation Codes for Explosion Monitoring
PRESTON, L., Sandia National Laboratories, Albuquerque, NM, USA, lpresto@sandia.gov; JONES, K. R., Sandia National Laboratories, Albuquerque, NM, USA, krjones@sandia.gov
Seismic and acoustic signals emanating from explosions are powerful phenomenologies that can be exploited for detection, identification and characterization. However, accurate and reliable identification of events requires the ability to model both the source physics and propagation of seismic and acoustic waves to global and regional monitoring stations. The ability to accurately simulate explosions is critical for nuclear monitoring, forensic evaluations and scenario modeling. Nonlinear algorithms such as Sandia’s CTH code can simulate the actual explosions, but they are inefficient and inaccurate at propagating these signals at great distances. A linearized wave propagation code, coupled to a nonlinear code, provides an efficient mechanism to both accurately simulate the explosion itself and also propagate these signals to distant receivers. We have successfully coupled radCTH and SNL’s linearized acoustic code, Acousti, for a simple test case. We will present results of this coupling simulation. The shape of the domain of coupling between linear and non-linear algorithms and the distance at which the coupling is implemented can affect far field wave amplitudes and shapes. This provides a means of investigating how the wavefield evolves toward linearity as a function of distance from an explosive source. The approximate distances where the wavefield is sufficiently close to linear will also be explored. Ultimately, the code can be validated with observations obtained from the Source Physics Experiment. This work was done under award number DE-AC52-06NA25946. 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:Induced Seismicity
Presenter   McGarr, Art
Schedule   Thu 8:30 AM / Oral
Room   Room 1
Factors That Enhance the Likelihood of Fluid Injection-Induced Earthquakes Large Enough to be Felt
MCGARR, A., USGS, Menlo Park, CA, USA, mcgarr@usgs.gov; RUBINSTEIN, J., USGS, Menlo Park, CA, USA, jrubinstein@usgs.gov
Felt earthquakes induced by fluid injection at depth typically are magnitude 2.5 and larger. Fluid injection operations that induce felt earthquakes include geologic disposal of wastewater, development of enhanced geothermal systems (EGS), and, on rare occasions, hydraulic fracturing. Of these three activities, wastewater disposal predominates both in terms of volumes of injected liquid and earthquake size, with magnitudes sometimes exceeding 5. Thus, total injected volume appears to be an important factor in the seismic response to injection, an observation consistent with case histories of induced earthquakes large enough to be of concern. Similarly, injection directly into crystalline basement seems to increase significantly the likelihood of felt earthquakes, as evidenced by case histories involving wastewater disposal and EGS. Another important factor is the presence of high-permeability fluid pathways that can channel the injection effects on pore pressure from the target aquifer into a fault zone, especially a fault that is well oriented for slip in the ambient stress field. Often, these faults are only revealed by the hypocenters of the induced earthquakes. The effect of injection rate is less straightforward. There is some evidence that during high-injection rate EGS stimulation earthquake rates are higher than those associated with lower-rate wastewater disposal. Assuming all other factors are equal (state of stress, appropriate fault orientation, and fluid pathways) this would suggest that the likelihood, per unit time, of a felt earthquake increases with injection rate. Given that high-rate fluid injection is typically very brief, the overall likelihood of felt earthquake being triggered by much longer operating lower-rate wells is likely higher than at high-rate wells. The available data corroborate the notion that the likelihood of induced earthquakes large enough to be felt is largely independent of injection rate.
Session:A Decade of Great Subduction Earthquakes – What Have We Learned From Their Ground-Motions?
Presenter   Skarlatoudis, Andreas
Schedule   Fri 9:15 AM / Oral
Room   Room 4
Broadband Strong Motion Simulations of Large Subduction Earthquakes
SKARLATOUDIS, A. A., URS Corp, Los Angeles, CA, USA, andreas.skarlatoudis@urs.com; SOMERVILLE, P. G., URS Corp, Los Angeles, CA, USA, paul.somerville@urs.com; THIO, H. K., URS Corp, Los Angeles, CA, USA, hong.kie.thio@urs.com; BAYLESS, J. R. B., URS Corp, Los Angeles, CA, USA, jeff.bayless@urs.com
The great subduction earthquakes that occurred recently in Peru, Chile and Japan have provided unparalleled information about the ground motions stemming from such earthquakes. Broadband ground motion simulations can enhance the usefulness of the recordings of these earthquakes by providing a means of interpolating and extrapolating the recorded data. In this study, we test our capability to simulate, with a hybrid method, broadband strong motion recordings of mega-thrust earthquakes by demonstrating that our simulations reproduce the amplitudes of the recorded ground motions without systematic bias. The broadband ground-motion simulation method that we used is based on the work of Somerville et al. (1991), Somerville (1993) and Graves and Pitarka (2004; 2010). It computes the long and short period ranges separately and then combines the two to produce a single time history using appropriate matching filters. At periods longer than 1s, the methodology is deterministic and contains a theoretically rigorous representation of fault rupture and wave propagation effects. At periods shorter than 1s, it uses an empirical representation of source radiation and scattering derived from the recordings of a smaller earthquake, which is combined with a simplified theoretical representation of wave propagation. The simulations have a model bias that is generally within 20% of zero across the full bandwidth. A systematic underprediction can be observed around the period of 0.5s. This underprediction could be driven by the characteristics of the empirical source function used in the high frequency part of the simulations. The standard deviation is comparable with the standard deviations of the most recent GMPE models (Zhao et al., 2006:Zea06 and Abrahamson et al, 2012:AGA12). The Atkinson and Boore (2003) (AB03) standard deviation is at the lower bound of the standard deviations from the simulations, while the Zea06 model has very comparable values across the examined period range.
Session:Large and Damaging Earthquakes of 2013/2014
Presenter   Ye, Lingling
Schedule   Thu 11:15 AM / Oral
Room   Room 2
STUDENT
Rupture Process of the November 17, 2013 Mw 7.8 Scotia Sea Earthquake
YE, L., Department of Earth and Planetary Sciences, UC Santa Cruz, Santa Cruz, CA, USA, lye2@ucsc.edu; LAY, T., Department of Earth and Planetary Sciences, UC Santa Cruz, Santa Cruz, CA, USA, tlay@ucsc.edu; KOPER, K. D., Department of Geology and Geophysics, University of Utah, Salt Lake City, UT, USA, kkoper@gmail.com; SMALLEY, R., CERI, The University of Memphis, Memphis, TN, USA, rsmalley@memphis.edu; BEVIS, M. G., School of Earth Sciences, Ohio State University, \Columbus, Ohio, USA, mbevis@osu.edu; ZAKRAJSEK, A. F., Ciencias de la Tierra, Dirección Nacional del Antártico - Instituto Antártico Argentino, Argentina, afz@dna.gob.ar; TEFERLE, F. N., Geophysics Laboratory, University of Luxembourg, Luxembourg, Germany, norman.teferle@uni.lu
On November 17, 2013, a very large (Mw 7.8) strike-slip earthquake occurred on the South Scotia Ridge (SSR), the boundary between the Scotia and Antarctica plates. The epicenter (60.27°S, 46.40°W) is northwest of the South Orkney micro-continent. The SSR is the most seismically active portion of the entire Antarctica plate boundary, having experienced about 10 M 7+ events since 1908 with additional events just east of the South Sandwich Island subduction zone. The earthquake sequence initiated with Mw 6.1 (Nov. 13) and 6.8 (Nov. 16) events located ~50 km west of the mainshock, and aftershocks extend ~250 km eastward along the SSR. On Nov. 24, 2013, a potentially triggered left-lateral Mw 7.0 event occurred ~900 km northwest on the North Scotia Ridge, the northern boundary of the Scotia plate with South America. The mainshock W-phase and GCMT point source inversions of long-period seismic waves indicate left-lateral strike-slip displacement on a fault striking 97°-102° and dipping southward at 45°-50°, with seismic moment of ~5.2-5.5 x 1020 N-m. Eastward rupture expansion for over 250 km at ~2.5 km/s is inferred from back-projections of teleseismic short-period P waves for networks of stations in the Americas and Australia, and from directivity in the effective source time functions for global long-period surface waves. Two intervals of large moment rate occurred from 20-50 s and 70-90 s after the origin, based on finite-fault inversion of broadband teleseismic P and SH waves. Iterative modeling of a 1 Hz GPS ground motion recording at station BORC on Laurie Island (South Orkney Islands) is used to constrain the source rupture parameters. The rupture extends to the vicinity of the rupture zone of an August 4, 2003 Mw 7.6 earthquake, which appears to have a similar southward dipping left lateral strike slip geometry. The 2013 event has relatively high short-period seismic wave energy release and high moment scaled radiated energy (ER/M0 ~ 3x10-5)
Session:Recent Advances in Ground Motions Simulation Methods and Their Validation
Presenter   Olsen, Kim
Schedule   Thu / Poster
Room   Cook/Arteaga
The SDSU Broadband Ground Motion Generation Module Version 1.5
OLSEN, K. B., San Diego State University, San Diego, CA, USA, kbolsen@mail.sdsu.edu; TAKEDATSU, R., San Diego State University, San Diego, CA, USA, rtakedatsu@mail.sdsu.edu
The Southern California Earthquake Center (SCEC) has completed Phase 1 of its Broadband Platform (BBP) ground motion simulation validation, evaluating the potential applications for engineering of the resulting PSAs generated by 5 different methods. The evaluation included part A, where the methods were evaluated based on the bias of simulation results to observations for 7 well-recorded historical earthquakes with source-station distances between 1 and 193 km, and part B, where simulation results for Mw 6.2 and Mw 6.6 strike-slip and reverse-slip scenarios were evaluated at 20 km and 50 km from the fault. The methods were assessed based on the bias of the median PSA for the 7 events in part A, and on a specified acceptance criterion compared to NGA-West2 GMPEs in part B. One of the 5 methods evaluated was BBtoolbox, a hybrid method combining deterministic low-frequency (LF) synthetics with high-frequency (HF) scatterograms (Mai et al., 2010; Mena et al., 2010; V1.4). In the validation exercise, the LFs are generated using 1D Green’s Functions and 50 source realizations from the kinematic source generator module by Graves and Pitarka (2010, ‘GP’). However, the results from BBtoolbox V1.4 did not appear to pass the validation exercise. In order to obtain more accurate BB synthetics, we generated BBtoolbox V1.5 which scales the HFs to a theoretical spectral level at the merging frequency (GP, Eq. 10) fixed at 1Hz, rather than the level of the LFs in V1.4. This modification has generated much improved spectral levels at higher frequencies, while we have seen little evidence of artifacts from this technique. In addition, V1.5 introduced a new source time function with rise-time scaled as a function of moment. With these modifications, BBtoolbox V.15 became one of three methods passing the SCEC validation exercise. Here, we describe the details of BBtoolbox V1.5, and show comparisons between BBtoolbox V1.5 synthetics and observations from the validation exercise.
Session:Explosive Source Characterization
Presenter   Saikia, Chandan
Schedule   Wed / Poster
Room   Cook/Arteaga
Yield and Depth of the SPE Chemical Explosions by Modeling of Local Waveforms Using a Source Consisting of Pure Explosion and Spall Sources
SAIKIA, C. K., Air Force Technical Applications Center, Patrick, FL, USA, chandanksaikia@gmail.com; WOODS, M., Air Force Technical Applications Center, Patrick, FL, USA, mwoods@aftac.gov
Regional waveforms generated by nuclear explosions are often used to invert for the moment tensor matrix to estimate the seismic moments of the isotropic, compensated linear vector dipole (CLVD) and double couple (DC) sources. In this paper, we modify this conventional approach to a grid-search technique assuming that the sources of the SPE chemical explosions consist of two mechanisms, isotropic and spall. In a separate analysis, we also include an additional DC source known to exist at a close proximity to the shot point. There is strong evidence of spall on the surficial accelerograms recorded very close to the shot point as well as that of a steeply dipping fault in the area surrounding the shots. The relative timing of the spall signal and dwell provide basic information regarding their delay time as well as their strengths relative to first P arrivals. We use this information in generating spall seismograms, assuming that source is equivalent to a vertically symmetric CLVD source or a vertical single force. In the single force formulation, we allow the force to act on the dislocation interface where the spall collapses at various depths. Synthetic spall seismograms are used in conjunction with the explosion seismograms to model recorded waveforms, determine the yield of the explosion, and seismic moment of the spall (or the CLVD) source. The source function of the explosion source is based on a time domain expression for a given yield and depth of burial (Saikia et al. ,2013). For the spall and DC mechanisms, source functions are assumed to be delta functions. A sensitivity study is being conducted to evaluate the dependence of model parameters on the depth of the spall sources. Green's functions used in this study are based on the inferred bore-hole data, but further refined to having the capability for entrapping high-frequency P and Rg waves observed from various shot points.
Session:A Decade of Great Subduction Earthquakes – What Have We Learned From Their Ground-Motions?
Presenter   Skarlatoudis, Andreas
Schedule   Fri / Poster
Room   Cook/Arteaga
Source Scaling Relations of Subduction Earthquakes for Strong Ground Motion and Tsunami Prediction
SKARLATOUDIS, A. A., URS Corp, Los Angeles, CA, USA, andreas.skarlatoudis@urs.com; SOMERVILLE, P. G., URS Corp, Los Angeles, CA, USA, paul.somerville@urs.com; THIO, H. K., URS Corp, Los Angeles, CA, USA, hong.kie.thio@urs.com; BAYLESS, J. R., URS Corp, Los Angeles, CA, USA, jeff.bayless@urs.com
At present, there is a wide range of uncertainty in the median rupture areas predicted for a given seismic moment M0 (Nm) by current relationships for subduction earthquakes. Our goal is to develop an updated set of earthquake source scaling relations that will reduce the large degree of epistemic uncertainty and improve the accuracy of seismic hazard analysis and the prediction of the strong motion characteristics and tsunamis of future subduction earthquakes. We compiled a database of interface earthquakes that occurred in the major subduction zones with Mw ranging from 6.75 to 9.1. We chose to fit the data using self-similar relationships between rupture area S (km2) and M0 and between the average slip D (m) and M0. The regression analysis was found to support the self-similarity of scaling relations up to Mw 8.6, as also suggested by previous studies (Murotani et al., 2013; 2008, Somerville et al., 2002). Above Mw 8.6, the rupture area increases more gradually with Mw and the average displacement increases more rapidly with Mw. We attribute this to the saturation of rupture width. For both S and D, the mean values computed in this study are within one standard deviation of those reported in Murotani et al. (2008; 2013). Comparison of the present study results with those of Somerville et al. (2002) indicates smaller areas and larger slip for the same M0 in the new relationships. These differences are less prominent for the average slip, where results lie within one standard deviation. The standard deviation is considerably smaller than the values estimated in previous studies as a result of the larger number of records used in the regressions. Another factor that might have contributed to smaller standard deviations is that we did not use multiple models of the same earthquake in the regression analyses. Instead, we used judgment to select the most representative one based on various criteria such as the number and type of data used in deriving the model.
Session:Topics in Seismology: Regional Seismicity and Tectonics
Presenter   VanDeMark, Thomas
Schedule   Thu / Poster
Room   Cook/Arteaga
Revisiting the Uljin Sequence of April 2006
VANDEMARK, T. F., Air Force Technical Applications Center, Patrick Air Force Base, FL, USA, thomas.vandemark.1@us.af.mil; KIM, T. S., Korea Institute of Geoscience and Mineral Resources, Daejeon City, ROK, tskim@kigam.re.kr; KANG, I. B., Korea Institute of Geoscience and Mineral Resources, Daejeon City, ROK, kang@kigam.re.kr
An earthquake sequence comprised of 10 reported events from the Korea Institute of Geoscience and Mineral Resources (KIGAM) catalog occurred in April of 2006. The sequence was located off of the eastern coast of the Southern Korean peninsula in the northwest corner of the Ulleung Basin. Due to the lack of a typical mainshock-aftershock sequence the seismicity is characterized as swarm behavior with magnitudes determined by KIGAM as ranging from ML 2.0 to 3.2 (Kang and Shin, 2006). Observations from regional seismic data show that these events are highly correlated. We employ a pseudo array correlation technique to detect an additional 4 events not previously associated to the sequence. Further, we perform relative relocations of the sequence’s events. This is accomplished by picking the phase arrivals manually by aligning waveforms of the events at common stations. The relative locations delineate two distinct parallel segments from SW to NE. This result supports assertions regarding the fault geometry in the northwest Ulleung Basin by previous researchers [Kang and Baag, 2004; Kang and Shin, 2006].
Session:Advances in Understanding Earthquake Hazard in Central and Eastern North America
Presenter   Vernon, Frank
Schedule   Fri 9:00 AM / Oral
Room   Room 2
USArray Transportable Array Evolving into the Central and Eastern United States Network (CEUSN)
VERNON, F., IGPP, UCSD, San Diego, CA, USA, flvernon@ucsd.edu; BUSBY, R., IRIS, Woods Hole, MA, USA, busby@iris.edu; WOODWARD, R., IRIS, Washington, DC, USA, woodward@iris.edu; ASTIZ, L., IGPP, UCSD, San Diego, CA, USA, lastiz@ucsd.edu
The Central and Eastern United States Network (CEUSN) is a long-term sub-array of ~160 seismometers selected from EarthScope’s USArray Transportable Array (TA) footprint. The mission of the CEUSN is to enable researchers and Federal agencies alike to better address the basic geologic questions, background earthquake rates and distribution, seismic hazard potential, and associated societal risks of this region. The network stems from collaboration between the National Science Foundation, United States Geological Survey, United States Nuclear Regulatory Commission, Department of Energy, and Incorporated Research Institutions for Seismology. Like the Transportable Array, the CEUSN stations are transmitting data in real time, allowing for continuous seismic monitoring. This long-term deployment of seismometers will produce a much larger set of observations than achievable during the originally planned 2-year duration of the TA deployment. The CEUSN, together with the existing coverage of permanently deployed seismometers in the central and eastern United States, will combine to form a network of ~300 broadband stations. The initial stations were deployed in November, and the rest will be operational by May 2012.
Session:Seismic Location and Processing Techniques
Presenter   Poppeliers, Christian
Schedule   Fri 9:00 AM / Oral
Room   Room 3
The Effects of Measurement Uncertainties on the Estimation of Seismic Gradients
POPPELIERS, C., Georgia Regents University, Augusta, GA, USA, cpoppeli@gru.edu; JOHNSON, C. E., Georgia Regents University, Augusta, GA, USA, cjohn102@gru.edu
Estimating the spatial gradients of a seismic wavefield has applications in seismic strain estimations, rotational motion estimates, and seismic gradiometry. Typically, seismic gradients are estimated using a small-scale seismic array of well-calibrated instruments to record the wavefield. The array stations occupy discrete coordinates in some sort of finite element grid and the spatial derivatives are estimated numerically. It's well known that the nominal aperture of the array must be smaller than 0.1 wavelengths to avoid numerical artifacts with estimating the spatial derivatives of the wavefield. However, there is little known regarding the minimum aperture of an array for reliable gradient estimation. Specifically, we can show that for given uncertainties in station coordinates and/or amplitude response, the uncertainties in the gradients are inversely proportional to array aperture. Our primary goal for the work shown here was to quantify this phenomenon. Our secondary goal was to devise a way to construct confidence intervals for the estimated spatial gradients. We perform a series of Monte Carlo simulations on synthetic array data to find an optimal array aperture that minimizes the uncertainty in the computed gradients (and attributes derived from them) for a given amount of measurement uncertainties. We then analyze actual seismic data from a high-frequency seismic array as well as a broadband seismic array. Both of these arrays were constructed specifically to estimate seismic gradients. By choosing specific stations within each array, we are able to vary the nominal aperture of the arrays between 0.01 and 0.1 wavelengths. We then compare our simulations with our actual data, and compute confidence intervals.
Session:Geometric Complexities Along Strike-Slip Systems: New Insights on Seismic Hazards, Earthquake Behavior, and Fault System Evolution
Presenter   Lay, Thorne
Schedule   Wed 11:00 AM / Oral
Room   Room 2
Implications of Recent Large Earthquakes on the Queen Charlotte Fault
LAY, T., Univ. California Santa Cruz, Santa Cruz, CA, USA, tlay@ucsc.edu
The Queen Charlotte Fault (QCF) accommodates predominantly right-lateral strike-slip motion between the Pacific and North American plates from the Explorer plate to the Fairweather Fault. The largest recorded earthquake along this boundary is the August 22, 1949 (Ms< 8.1) right-lateral strike-slip event. Southeastward from the 1949 rupture the QCF locates offshore of Haida Gwaii with a modest restraining bend resulting in compression and uplift of the Wrangelia terrane. The current obliquity of the relative plate motion is less than 20° from boundary parallel, prompting contention over the existence and extent of any underthrusting and associated seismic hazard. The October 28, 2012 (Mw 7.8) Haida Gwaii thrust earthquake ruptured offshore of Moresby Island, producing a substantial local tsunami and confirming the existence of strong slip partitioning into separate fault systems. Lay et al. (EPSL, 2013) modeled seismic and tsunami observations, finding an average slip of 3.3 m on a shallow-dipping plane below an offshore sedimentary terrace seaward of the QCF. Modeling of coseismic GPS deformation (Nykolaishen et al., AGU, 2013) and minor coastal submergence observed along Moresby Island (Haeussler et al., AGU, 2013) confirm the offshore slip. Lacking historical documentation of rupture of the QCF along this part of the margin, the potential for strike-slip failure of the QCF along Moresby Island remains unclear. With pre-event GPS observations in Haida Gwaii indicating plate-motion parallel deformations, there is significant boundary parallel shear stress that needs to be accounted for that was not released seismically in the 2012 event, nor in its afterslip. It is difficult to attribute this to shear-stress on the QCF, and more plausibly can be explained by viscous coupling of the deeper underthrust flange of Pacific plate in direct contact with North America beneath the islands. This implies low potential for large strike-slip faulting on this stretch of the QCF.
Session:Alaska Update of the USGS National Seismic Hazard Maps
Presenter   Wong, Ivan
Schedule   Thu 11:45 AM / Oral
Room   Room 4
Site-Specific Probabilistic Seismic Hazard Analyses of Anchorage, Alaska and Comparison to the 2007 National Seismic Hazard Maps
WONG, I., URS Corporation, Oakland, CA, USA, ivan.wong@urs.com; ZACHARIASEN, J., URS Corporation, Oakland, CA, USA, judy.zachariasen@urs.com; DOBER, M., URS Corporation, Oakland, CA, USA, mark.dober@urs.com
We have been performing site-specific probabilistic seismic hazard analyses for sites located in Anchorage, Alaska. The city is situated in one of the most seismically active regions in the U.S. The Alaskan subduction zone, which underlies the city, was the source of the 1964 M 9.2 Great Alaska earthquake. Wadati-Benioff earthquakes within the subducting Pacific plate and nearby crustal faults could also generate future strong ground shaking in the city. Unlike the National Seismic Hazard Maps (NSHM), we consider all active and potentially active fault sources that could affect Anchorage. For example in the Anchorage area, the 2007 NSHMs included only the Castle Mountain fault as a crustal fault source because of the lack of fault studies in the region. The nearby Cook Inlet faults are characterized in our analyses based on the studies of Haeussler et al. (2002). The uncertainties for all fault parameters are handled through the use of logic trees. We quantify the uncertainties in the seismogenic capability of all faults by assigning probabilities of activity. Our characterization of the Alaskan subduction zone megathrust uses many of the parameters used in the 2007 NSHMs with some revisions based on recent paleoseismic studies. Unlike the NSHM’s use of horizontal zones of gridded seismicity, we modeled the Wadati-Benioff zone as a dipping slab. We calculated recurrence for intraslab earthquakes in the historical catalog identified using the slab geometry. Both a uniform source zone and gridded seismicity were included in our probabilistic analyses. We used recent subduction zone ground motion prediction models (e.g., Abrahamson et al., 2013) and the NGA West-2 models. The intraslab zone dominates the PGA hazard. For long-period ground motions (e.g., 1.0 sec or more), the megathrust generally controls the hazard. We compare these site-specific hazard results in Anchorage to the 2007 NSHMs.
Session:Explosive Source Characterization
Presenter   Alvizuri, Celso
Schedule   Wed / Poster
Room   Cook/Arteaga
STUDENT
Full Moment Tensors for Small (Mw <3) Events at Uturuncu Volcano, Bolivia
ALVIZURI, C., Geophysical Institute, Fairbanks, AK, USA, alvizuri@gi.alaska.edu; TAPE, C., Geophysical Institute, Fairbanks, AK, USA, carltape@gi.alaska.edu
We use seismograms from a deployment of 24 broadband stations on Uturuncu volcano, Bolivia, to estimate full moment tensors for small (Mw ≤ 3) crustal earthquakes. We adapt the grid-search moment tensor inversion method of Zhu and Helmberger (1996), which considers first motion polarities, body waves, and surface waves; and use a parameterization of seismic moment tensor in terms of its eigenvalues (magnitude, isotropic, CLVD) and orientation (strike, dip, rake). Performing a full grid search on these parameters allows us to estimate uncertainties. The majority of events are in the range 0 ≤ Mw ≤ 1 and allow use of only P-wave polarities which results in larger parameter uncertainties, but our representation of misfit over parameter space still enables us to discriminate among different earthquake mechanisms. For the largest events (Mw >2) we are able to include full seismograms in the inversion, which reduces uncertainty estimates for the moment tensors. We also explore whether different subsets of waveforms (e.g., surface waves only or polarities only) will produce consistent moment tensors. The ability to resolve all six parameters, notably isotropic and CLVD, depends strongly on station coverage and signal-to-noise level in the seismogram. We present a comprehensive full moment tensor catalog for crustal events at Uturuncu volcano, with uncertainties in source mechanisms depicted visually on the fundamental lune of eigenvalue space. We use the lune to interpret the catalog, whereby crack tensors are at the boundary of the lune and double couple tensors are at the center. We find a predominance of source mechanisms with positive isotropic components, which we interpret as an indication of tensional crack processes beneath the volcano.
Session:Advances in Seismic Imaging and Monitoring of Time-Dependent Variations: Civil Structures, Near-Surface, and Shallow Crustal Scales
Presenter   Schwed, Martin
Schedule   Wed 9:30 AM / Oral
Room   Room 3
STUDENT
Joint Modeling of Complementary Data Functionals for Seismic Site Characterization
SCHWED, M., Baylor University, Waco, TX, USA, martin_schwed@baylor.edu; PULLIAM, J., Baylor University, Waco, TX, USA, jay_pulliam@baylor.edu; SEN, M. K., Institute for Geophysics, University of Texas at Austin, Austin, TX, USA, mrinal@ig.utexas.edu; WILLEMANN, R. J., IRIS Consortium, Washington, DC, USA, ray@iris.edu; HUERTA-LOPEZ, C., University of Puerto Rico, Mayaguez, Mayaguez, PR, USA, huerta@cicese.mx; MOSCHETTI, M. P., USGS, Golden, CO, USA, mmoschetti@usgs.gov; SCHMITZ, M., FUNVISIS, Caracas, Venezuela, Bolivarian Republic of., mschmitz@funvisis.org.ve; LOUIE, J. N., The University of Nevada, Reno, Reno, NV, USA, louie@seismo.unr.edu; POLANCO, E., Universidad Autonoma de Santo Domingo, Santo Domingo, Dominican Republic eugenio_polanco_rivera@msn.com; HUERFANO MORENO, V., University of Puerto Rico, Mayaguez, Mayaguez, PR, USA, victor@prsn.uprm.edu; PASYANOS, M., Lawrence Livermore National Laboratory, Livermore, CA, USA, pasyanos1@llnl.gov
New approaches suggest that it may be possible to determine shallow Earth structure through low-cost, non-invasive seismic surveys that make use of ambient noise and that the results can be used for “shake-casting” to produce scenarios for the purposes of urban planning, improving community resilience, and emergency response. We will present a strategy for determining seismic “site characterization” through joint modeling of horizontal to vertical spectral ratios (HVSR) and surface wave dispersion, determined via spatial autocorrelation (SPAC), refraction microtremor (ReMi), and/or multi-channel analysis of surface waves (MASW). Fitting of data functionals by synthetics is driven by global optimization and the models are assessed quantitatively. The products of this approach are shear wave velocity profiles for the shallow subsurface, accompanied by posterior probability distributions and parameter correlation matrices. Optimization strategies for solving nonlinear problems in geophysics have several advantages over linearized inversions. Jointly fitting dispersion curves and HVSR functionals via global optimization allows us to characterize the space of possible models, assess model reliability, identify parts of the “best-fit” model that are poorly constrained, and guide us toward new data that might improve constraints on the model. Tools such as the posterior probability distribution and the parameter correlation matrix allow us to assess the relative contribution of both types of data to model constraints and how to choose the optimal weights between data types. The joint modeling technique is applied to data acquired in an NSF-funded Pan-American Advanced Studies Institute in Santo Domingo, Dominican Republic, entitled “New Frontiers in Geophysical Research: Bringing New Tools and Techniques to Bear on Earthquake Hazard Analysis and Mitigation”, as a proof-of-concept survey in a highly built-up urban environment.
Session:Seismic Location and Processing Techniques
Presenter   VanDeMark, Thomas
Schedule   Fri / Poster
Room   Cook/Arteaga
Examination of the Storfjorden Aftershock Sequence
JUNEK, W. N., Air Force Technical Applications Center, Patrick Air Force Base, FL, USA, william.junek@us.af.mil; KVAERNA, T., NORSAR, Kjeller, Norway; PIRLI, M., NORSAR, Kjeller, Norway; SCHWEITZER, J., NORSAR, Kjeller, Norway; HARRIS, D. B., Deschutes Signal Processing, Maupin, OR, USA; DODGE, D. A., Lawrence Livermore National Laboratory, Livermore, CA, USA; WOODS, M. T., Air Force Technical Applications Center, Patrick Air Force Base, FL, USA; VANDEMARK, T. F., Air Force Technical Applications Center, Patrick Air Force Base, FL, USA, thomas.vandemark.1@us.af.mil
The Storfjorden aftershock sequence was triggered by a MW 6.2 earthquake that occurred off the Spitsbergen’s southeastern coast on February 21, 2008. Since its initiation, the sequence has generated thousands of aftershocks; several have exceeded MW 4.0. Recent seismotectonic investigations have suggested the source of the sequence is not related to the nearby Billefjorden fault zone. Instead, its source is most likely tied to a NE-SW trending Tertiary shear zone. This conclusion is supported by the orientation of the relative relocations and the orientations of focal mechanisms for numerous sequence events [Pirli et al., 2013; Junek et al., 2014]. The results shown in previous studies [Pirli et al. , 2013; Junek et al., 2014] were largely based on the analysis of manually reviewed events. However, the number of events in the NORSAR analyst reviewed bulletin represents a small fraction of the total number of events produced by the sequence. Here, an autonomous event detection and clustering framework is used to expand the available dataset. The expanded dataset is used to infer additional information about the tectonic structure within the fjord and to examine the sequence’s spatiotemporal properties. The relative relocation catalog orientation and mixture of strike-slip and normal focal mechanism suggests the fault has a SW-NE trending en echelon structure.
Session:Emergence of Continuously Recording Very Large Array Capabilities in Seismology
Presenter   Hole, John
Schedule   Wed 2:15 PM / Oral
Room   Room 7/8
Back-Projection Imaging of Aftershocks Recorded by the Dense AIDA Array after the 2011 Virginia Earthquake
HOLE, J. A., Virginia Tech, Blacksburg, VA, USA, hole@vt.edu; WANG, K., Virginia Tech, Blacksburg, VA, USA, wkai83@vt.edu; DAVENPORT, K. K., Virginia Tech, Blacksburg, VA, USA, kdavenport@vt.edu; CHAPMAN, M. C., Virginia Tech, Blacksburg, VA, USA, mcc@vt.edu; BESKARDES, G. D., Virginia Tech, Blacksburg, VA, USA, didem@vt.edu; QUIROS, D. A., Cornell University, Ithaca, NY, USA, daq7@cornell.edu; BROWN, L. D., Cornell University, Ithaca, NY, USA, ldb7@cornell.edu; MOONEY, W. D., U. S. Geological Survey, Menlo Park, CA, USA, mooney@usgs.gov
Aftershock Imaging with Dense Arrays (AIDA) recorded 12 days of seismic data following the 23 August 2011 magnitude 5.8 earthquake in central Virginia. AIDA utilized short-period, vertical-component seismographs at 201 locations at 200-400 m spacing to drastically reduce spatial aliasing as compared to traditional aftershock networks. Inter-station correlation enabled a visible detection threshold between magnitude –1.5 and –2. Traditional local earthquake tomography with ~1 km resolution revealed a homogeneous seismic velocity of Vp = 6.2 km/s and Vs = 3.6 km/s in the aftershock zone. This is consistent with the aftershocks occurring within a single crystalline-rock terrane. Hypocenters were located with an absolute accuracy of ~100 m without using double difference or cross-correlation methods. The hypocenters define a zone of seismicity that is >1 km wide, much larger than the hypocenter accuracy, suggesting a new or immature fault zone. The fault zone is weakly concave upwards in depth and along strike. Reverse time migration was applied to several of the aftershocks to back-project recorded seismic energy to the source. Events as tiny as magnitude –2 and with signal smaller than noise were successfully imaged as point sources with ~200 m resolution. The propagation of energy release as a function of time and space was observed for events from magnitude 2.5 to magnitude 3.7, the largest event recorded. Slip propagated at a high speed and ruptured a relatively long distance for the magnitude, suggesting that new fractures are being created in strong rock under high stress. Synthetic data tests show that resolution was primarily limited by the temporal sampling rate. Improved temporal and spatial sampling are feasible, and would produce images with sharper resolution.
Session:Citizen Seismology: Citizens Helping Science Helping Citizens
Presenter   Evans, John R
Schedule   Fri 2:15 PM / Oral
Room   Room 7/8
Performance of Several Low-Cost Accelerometers
EVANS, J. R., U.S. Geological Survey, Santa Cruz, CA, USA, jrevans@usgs.gov; ALLEN, R. M., Univ. of Calif. at Berkeley, Berkeley, CA, USA, rallen@berkeley.edu; CHUNG, A. I., Stanford Univ., Stanford, CA, USA, angelaichung@gmail.com; COCHRAN, E. S., U.S. Geological Survey, Pasadena, CA, USA, ecochran@usgs.gov; GUY, R., Calif. Inst. of Tech., Pasadena, CA, USA, rguy@gps.caltech.edu; HELLWEG, M., Univ. of Calif. at Berkeley, Berkeley, CA, USA, peggy@seismo.berkeley.edu; LAWRENCE, J. F., Stanford Univ., Stanford, CA, USA, jflawrence@stanford.edu
Several groups are creating low-cost host-operated accelerograph systems to support seismologists and engineers. These networks use very inexpensive sensors, host installation, computing, and telemetry to make networks economical to purchase and operate. The Advanced National Seismic System (ANSS) is exploring the potential of such systems for its networks. Examples of such networks include the Community Seismic Network (csn.caltech.edu) and the Quake-Catcher Network (Cochran et al., 2009; qcn.stanford.edu). The overarching goals of such efforts are to increase spatial density of existing networks but at very low per-site cost. Host operation brings new people into citizen science fostering interest in the issues of earthquakes and adding measurement that can mitigate damage. We summarize the performance of a wide range of low-cost sensors from testing at the Albuquerque Seismological Laboratory and elsewhere. They commonly have resolution from 12–16 bits over ±2 g, with a triaxial cost of about $100 to $200; they can have either USB or analog outputs; in our case USB outputs are recorded on software running on a laptop or within the instrument. We did “box flip” tests for 0 Hz for sensitivity, offset, and axis orientations, and we measured amplitude transfer functions. These are some of the tests commonly applied to ANSS accelerometers of any given precision. Two of the sensors we tested performed acceptably. Others, as tested, would not be viable in ANSS networks. The better devices have better resolution than the Kinemetrics SMA-1, which drove the early decades of earthquake building-code development. Axis orientation and sensitivity are less precise than in other ANSS accelerometers and might need to be corrected if applied for sensitive purposes (legal proceedings, design events, ground-motion models). However, in applications like ShakeMap where predictive uncertainties are nearly a factor of two, these low-cost devices are sufficient without correction.
Session:Recent Advances in Ground Motions Simulation Methods and Their Validation
Presenter   Assimaki, Dominic
Schedule   Thu 9:15 AM / Oral
Room   Room 3
Site-Specific Response in Validation Studies of Physics-based Earthquake Simulations
ASSIMAKI, D., Georgia Tech, Atlanta, GA, USA, dominic@gatech.edu; SHI, J., Georgia Tech, Atlanta, GA, USA, shijian@gatech.edu; TABORDA, R., University of Memphis, Memphis, TN, USA, ricardo.taborda@memphis.edu
Although site response has long been known to be an important factor in modifying seismic motion, its integration in physics-based earthquake simulations is constrained by challenges of computational cost, and by the scarcity of geotechnical information; in turn, site response is accounted for in terms of trends through Vs30-based amplification factors. However, as the computational capabilities of 3D physics-based earthquake simulations extend in the high frequency range (>1 Hz), the fine resolution of the geotechnical layers and their nonlinear response have become increasingly important for validation studies. Recent efforts by the United States Geological Survey (USGS) have produced geotechnical profiles at more than 100 stations in Southern California. In this study, we use a subset of these profiles to perform site-specific analyses at stations used for the validation of physics-based ground motion simulations of the 2008 Chino Hills earthquake; and compare our results to recorded and physics-based simulated ground motions without site-specific response. We quantify the differences of ground surface predictions and observations in terms of both seismological and engineering intensity measures. We show that the goodness-of-fit of high frequency predictions compared to the recorded motions improves significantly when we introduce site-specific geotechnical profiles. Our results thus promote the interdisciplinary efforts of seismologists and engineers to introduce broadband simulated ground motions in earthquake design practice.
Session:Recent Advances and Findings in Earthquake Geology and Paleoseismology
Presenter   Marliyani, Gayatri
Schedule   Fri 2:15 PM / Oral
Room   Room 3
STUDENT
Geomorphic and Geologic Evidence of Recently Active Faults and Folds in Java, Indonesia
MARLIYANI, G. I., Arizona State University, Tempe, AZ, USA, gayatri.marliyani@asu.edu; ARROWSMITH, J. R., Arizona State University, Tempe, AZ, USA, ramon.arrowsmith@asu.edu; HELMI, H., Arizona State University, Tempe, AZ, USA, hurien.helmi@asu.edu
Java is a highly populated island in Indonesia along the Sunda-Banda subduction system. There have been damaging earthquakes in Java; however, detailed neotectonic studies are limited. In this study, we present detailed maps of active faults derived from satellite and field-based neotectonic mapping, paleoseismic data, as well as new data on fault kinematics, tectonic geomorphology indices indicating relative rock uplift, and estimates of principal stress orientations from volcano morphology. The structures in East Java are dominated by NW-SE normal faults, Central Java by E-W folds and thrust faults, while NE-SW strike-slip faults dominate West Java. Our preliminary mapping indicates no large scale continuous structures in Java. Instead deformation is distributed over broad areas along small structures. We illustrate our results with examples from the many active structures identified in this study: the Baluran and Pasuruan normal faults in East Java, the Kendeng fold and thrust belt in Central Java, and the Cimandiri strike-slip fault system in West Java. In Pasuruan, two shallow trenches cut across the NW-SE scarp exposed faulted and folded alluvial and colluvial strata and provide evidence for at least three ground-rupturing earthquakes. Uplifted meanders and terraces indicate ongoing uplift along the Kendeng fold and thrust belt. In Cimandiri, the normalized channel steepness index delineates the active trace along the otherwise diffuse fault zone. The structure mapping and principal stress orientations from volcano morphology suggest that the distinctive patterns are likely controlled by inherited crustal structures and variable boundary stresses from subduction of bathymetric highs and angle of the subducting slab. Javan seismic hazards include great tsunamigenic earthquakes along the subduction interface to the south as well as M6-7, widely-distributed shallow earthquakes associated with upper plate structures.
Session:From the Earthquake Source to Damage of Buildings: Bridging the Gap between Seismology and Earthquake Engineering
Presenter   Crempien, Jorge
Schedule   Fri 8:30 AM / Oral
Room   Room 1
Does a 1D Velocity Structure Hurt or Help Ground Motion Predictions?
CREMPIEN, J. G. F., UCSB, ERI, Santa Barbara, CA, USA, crempien@eri.ucsb.edu; ARCHULETA, R. J., UCSB, ERI, Santa Barbara, CA, USA, ralph.archuleta@ucsb.edu
For linear response, ground motion is a convolution between the source and Green’s functions (GFs) for an anelastic structure. While 1D anelastic structures may provide a template to compute arrival times (high frequency), GFs (for these layered structures) convolved with the source can naturally underestimate the ground motion amplitude at high frequencies. With source being embedded in the mostly homogeneous structure, the up-going rays from the fault impinge with a large incidence angle the bottom of the near-surface layers and glance off. The low-frequency ground motion is also affected. The near-surface layering leads to natural waveguides for surface waves that tend to be larger than those observed for most earthquakes, thus overestimating response spectra at high periods. The fundamental erroneous assumption in using a 1D structure is that the uniform layers exist over a wide range of distances. Some boundaries, e.g., the Moho, obviously persist over large distances. But the assumption that near-surface uniform layers that are 10’s of m thick exist over 10’s of km, is certainly false. Imperatori & Mai (2013) incorporated randomness into a 3D velocity structure; the randomness leads to 50-60% attenuation in PGA and 20-30% attenuation in PGV even at distances less than 20 km. While this approach is a laudable, it is currently impractical (numerically expensive) for large distances. Recent results (Dunham et al., 2011; Fang & Dunham, 2013; Shi & Day, 2013) demonstrate that fault roughness leads to a random wavefield. The UCSB method (Liu et al., 2006; Schmedes et al., 2013) embeds randomness into the source by the choice of distribution functions for the kinematic parameters. Since ground motion is a convolution of source and path, we propose using randomness in the source convolved with a nearly homogeneous anelastic structure. We compare synthetics with response spectra based on ground motion prediction equations for the western US (Earthquake Spectra, 2008).
Session:Recent Advances in Ground Motions Simulation Methods and Their Validation
Presenter   Silva, Fabio
Schedule   Thu / Poster
Room   Cook/Arteaga
Using the SCEC Broadband Platform for Strong Ground Motion Simulation and Validation
SILVA, F., Southern California Earthquake Center, Los Angeles, CA, USA, fsilva@usc.edu; MAECHLING, P., Southern California Earthquake Center, Los Angeles, CA, USA, maechlin@usc.edu; GOULET, C., Pacific Earthquake Engineering Research Center, Berkeley, CA, USA, goulet@berkeley.edu; SOMERVILLE, P., URS Corporation, Los Angeles, CA, USA, paul.somerville@urs.com; JORDAN, T., Southern California Earthquake Center, Los Angeles, CA, USA, tjordan@usc.edu
The Southern California Earthquake Center (SCEC) Broadband Platform is open-source scientific software that can generate broadband (0-100Hz) ground motions for earthquakes, integrating complex scientific modules that implement rupture generation, low and high-frequency seismogram synthesis, non-linear site effects calculation, and visualization into a software system that supports easy on-demand computation of seismograms. The Platform has been developed by a scientific and engineering collaboration that involves geoscientists, civil engineers, graduate students, and scientific software developers. The Platform operates in two modes: validation simulations and scenario simulations. In validation mode, the Platform runs earthquake rupture and wave propagation modeling software to calculate seismograms of a historical earthquake for which observed strong ground motion data is available. In validation mode, the Platform calculates a number of goodness of fit measurements that quantify how well the model-based broadband seismograms match the observed seismograms. Based on these results, the Platform can be used to evaluate and validate different numerical ground motion modeling techniques. We have recently modified the software to enable the addition of a larger number of historical events, and we are now adding validation simulation inputs and observational data for 20+ historical events from the eastern and western United States, Japan, Taiwan, Turkey, and Italy. In scenario mode, the Platform runs simulations for hypothetical earthquakes. In this mode, users input an earthquake description, a list of station names and locations, and a 1D velocity model for the region of interest, and the Platform software then calculates ground motions for the specified stations. The Broadband Platform collaborative development process, involving earth scientists, engineers, and software developers, helps ensure that the Platform can be used in earthquake engineering applications.
Session:Recent Advances in Ground Motions Simulation Methods and Their Validation
Presenter   Rodgers, Arthur
Schedule   Thu 10:45 AM / Oral
Room   Room 3
Comparison of 3D and 1D Wave Propagation Modeling in the San Francisco Bay Area
PITARKA, A., Lawrence Livermore National Laboratory, Livermore, CA, USA, pitarka1@llnl.gov; RODGERS, A. J., Lawrence Livermore National Laboratory, Livermore, CA, USA, rodgers7@llnl.gov; PETERSSON, A., Lawrence Livermore National Laboratory, Livermore, CA, USA; SJOGREEN, B., Lawrence Livermore National Laboratory, Livermore, CA, USA
We performed 1D and 3D ground motion simulations in the San Francisco Bay area using kinematic rupture models for the M6.9 Loma Prieta earthquake, and M6.7 scenario earthquakes on the Hayward fault. The objective of our investigation is to evaluate three-dimensional wave propagation effects on ground motion from large earthquakes in an area with complex underground structure, and compare the efficiency of 3D and 1D velocity models, at capturing such effects in the period range of 0.5 to 20 s. We use a deterministic approach for modeling wave propagation, and a combination of deterministic and stochastic approaches for rupture and velocity model parameterizations. A suite of rupture scenarios with stochastic slip variations was used to analyze the sensitivity of 3D and 1D wave propagation effects to small-scale variations in the rupture kinematics. The rupture models for both earthquakes were generated on the SCEC Broad-Band Ground Motion Simulation Platform using the Graves and Pitarka (2010) method. The simulations were performed in the frequency range of 0.01 to 2 Hz using an anelastic 3D finite-difference method and the USGS San Francisco Bay area 3D velocity model, Version 8.3.0, with surface topography. We added small-scale random perturbations to the original USGS velocity model to account for seismic wave scattering caused by structural heterogeneities. Strong motion data from the USGS and the California Geological Survey strong motion arrays were used in the models validation. We will show results of our analysis using waveform comparison, and goodness of fit of spectral responses between simulated and recorded ground motion, as well as between 3D and 1D simulated ground motions, with respect to spectral period, fault distance, basin depth, and rupture scenario in the period range of 0.5 to 20 s. This work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.
Session:Geometric Complexities Along Strike-Slip Systems: New Insights on Seismic Hazards, Earthquake Behavior, and Fault System Evolution
Presenter   Scharer, Katherine
Schedule   Wed 9:00 AM / Oral
Room   Room 2
Comparison of Fault Behavior across the Big Bend of the San Andreas Fault, California
SCHARER, K., USGS, Pasadena, CA, USA, kscharer@usgs.gov; WELDON, R., UO, Eugene, OR, USA, ray@oregon.edu; STREIG, A., UO, Eugene, OR, USA, streig@uoregon.edu; YULE, D., CSUN, Northridge, CA, USA, Doug.Yule@csun.edu; WOLFF, L., CSUN, Northridge, CA, USA, lisa.humbert.346@my.csun.edu
The “Big Bend” of the San Andreas Fault (SAF) is a large restraining bend with localized areas of differing geometric and seismologic complexity. For example, compared to the dearth of seismicity along the simple, linear portions of the fault, the southern Big Bend in the San Gorgonio Pass Region is characterized by higher rates of recent seismicity and complex fault networks. In contrast, at the northern end of the Big Bend (where the Garlock Fault impinges) the SAF remains a single, strike-slip strand with low seismicity rates despite a relatively tight bend. We review paleoseismic event data from both ends of the Big Bend and place these results within the context of published regional geophysical studies and slip rates. At the Frazier Mountain paleoseismic site, the last ~600 years of large earthquakes shows remarkable consistency with the Carrizo section, yet differs from sites to the south on the Mojave section. Geologic slip rates taper somewhat along this 200 km span (geodetic rates vary), but geophysical studies indicate the fault transitions from a moderate SW dip to vertical just SE of the northern Big Bend, hinting that geometric changes could control earthquake behavior at smaller timescales but are not important at longer timescales. This pattern does not appear to hold in the San Gorgonio Pass region, which hosts dramatic changes in the fault orientation and complexity, high surface uplift rates (>2 mm/yr), and a marked drop in the horizontal slip rate. Here, the paleoseismic rates also change; the interval between ground-rupturing earthquakes at the Cabazon paleoseismic site is >7 times longer than on the linear strands of the fault outside of the Pass, suggesting that ruptures are rare and/or quite distributed through the Pass. Thus, greater fault complexity in the south versus greater continuity in the north appear to correlate with the scale of differences in short-term earthquake rates and the longer-term SAF behavior.
Session:Emergence of Continuously Recording Very Large Array Capabilities in Seismology
Presenter   Meng, Lingsen
Schedule   Wed 2:45 PM / Oral
Room   Room 7/8
Optimal Design of Large-Scale Seismic Array for Earthquake Source Imaging Constrained by Waveform Coherency of USArray
MENG, L., Earth, Planetary and Space Sciences, UCLA., Los Angeles, CA, USA, meng@ess.ucla.edu; AMPUERO, J. P., Seismo Lab, Caltech, Pasadena, CA, USA
Waveform coherence across a seismic array is crucial to back-projection (BP) earthquake source imaging. Previous work indicates the waveform coherence decays dramatically with inter-station distance and frequency. However, these studies use time windows with fixed duration, which naturally degrades the coherence at high frequency. Here, we measure the correlation coefficients of teleseismic waveforms recorded at USArray with window lengths proportional to 1/frequency and the same alignment procedure, filtering and window parameters adopted in teleseismic BP for source imaging. We find the coherency is high across the USArray over inter-station distances >10 wavelengths and up to 5 Hz. The coherence of large/shallow earthquakes decays faster with distance than small/deep earthquakes. For the same earthquake, coherence falls slower along the ray-path than across it. We explain these patterns by a finite source effect, including scattering near the source, supported by synthetic tests. Coherence also decreases with frequency, mainly due to the effect of signal to noise ratio. Our results suggest that coherence is not a limiting factor for array design. The optimal design can be thus inferred from classic array processing principles assuming perfect coherency. Aliasing is inversely proportional to station spacing and resolution is proportional to array aperture. BPs of the Tohoku earthquake based on different subsets of USArray show that spacing should be at most twice the actual USArray spacing (~150 km). As robust array processing requires windows much longer than the wave period, the effective temporal resolution is about 10 periods. Since the BP image averages the rupture process within this window, the spatial resolution of the array design does not need to be smaller than the rupture propagation distance over the window duration. This physical resolution limit provides a constraint on the optimal aperture of the array.
Session:Site Response: From Site-Specific Analyses to Predictive Models Around the Globe
Presenter   Assimaki, Dominic
Schedule   Wed 9:15 AM / Oral
Room   Room 4
Validation of Site-Specific Response Models Using KIK-Net Ground Motion Recordings
SHI, J., Georgia Tech, Atlanta, GA, USA, shijian@gatech.edu; ASSIMAKI, D., Georgia Tech, Atlanta, GA, USA, dominic@gatech.edu
Our ability to predict site response in regional ground motion simulations is constrained to a large extent by the scarcity of velocity profiles. To overcome this issue, site response is accounted for in terms of trends, parameterized by site stiffness proxies such as Vs,30. This approach is adequate for broadband ground motion models that capture high-frequency (>1Hz) components –that is, the frequencies mostly affected by site response– in terms of trends as well. Physics-based ground motion simulations, however, provide deterministic predictions of broadband time-series that cannot be corrected for site response by means of spectral amplification ratios. To validate the high‑frequency predictive capabilities of these models, it is important to account for the fine resolution of the geotechnical layers, and for the nonlinear effects of these layers in the time domain. In this talk, we present a validation study of four 1D site-specific response models that can be used to modify ground motion time-series for seismic wave amplification in layered media. We test both linear frequency-domain and nonlinear time-domain solutions on the site scale, by comparing surface predictions to observations at 30 KIK-Net stations that have recorded several strong events, including the 2011 Tohoku earthquake. Velocity profiles are used as the only input to all models, while additional parameters such as quality factor, density and nonlinear dynamic soil properties are estimated from empirical geotechnical engineering correlations. Thus, differences in site-specific response predictions are attributed to the model capabilities rather than to differences in input parameters. We use validation metrics that are of interest in both seismology and engineering, including bias ratios of peak ground response and visual comparisons of elastic spectra, on the one hand, and inelastic to elastic deformation ratio for multiple ductility ratios, on the other hand.
Session:Induced Seismicity
Presenter   Kwiatek, Grzegorz
Schedule   Thu 1:45 PM / Oral
Room   Room 1
Geomechanical Reservoir Characterization Using Induced Seismicity and State-of-The-Art Waveform Processing Techniques
KWIATEK, G., GFZ German Research Centre for Geosciences, Potsdam, Germany, kwiatek@gfz-potsdam.de; BOHNHOFF, M., GFZ German Research Centre for Geosciences, Potsdam, Germany, bohnhoff@gfz-potsdam.de; MARTÍNEZ-GARZÓN, P., GFZ German Research Centre for Geosciences, Potsdam, Germany, patricia@gfz-potsdam.de; BULUT, F., GFZ German Research Centre for Geosciences, Potsdam, Germany, bulut@gfz-potsdam.de; DRESEN, G., GFZ German Research Centre for Geosciences, Potsdam, Germany, dre@gfz-potsdam.de
In this study we present the application of three waveform processing techniques to investigate physical processes in reservoirs related to fluid injection operations performed at Berlín and The Geysers geothermal fields. An initial hypocenter catalog of 581 fluid-induced seismic events recorded at the Berlín geothermal field was significantly revised and refined using double-difference relocation and spectral ratio methods. The average resolution within the seismic cloud was improved by a factor of 15. The application of the spectral ratio technique effectively cancelled path effects resulting in improved source parameters for the seismic events. The refined catalog allowed for a detailed analysis of spatio-temporal evolution of the seismicity related to fluid injection initially not visible due to limited resolution: We find a clear clustering of events around the injection intervals and in areas of high conductivity, observe a migration of the seismicity away from the injection interval with increasing injection rates and along pre-existing faults, detect larger seismic events in undamaged parts of the faults and in areas with a high conductivity gradient. We also find that the static stress drop of seismic events is decreasing with distance from the injection point. The stress inversion technique was applied to a seismicity cluster in the NW part of The Geysers geothermal field in California in order to demonstrate how this technique can contribute to the monitoring of reservoir-geomechanical processes occurring in response to short-term fluid-injection. The analyzed cluster composed of 742 events with fault plane solutions was used to look for time-dependent stress changes. We observed significant rotations of the stress field orientation directly related to increasing flow rates during peak-fluid injection. Moreover, these stress rotations are reversible documenting that stress field orientation might serve as an additional parameter for reservoir monitoring.
Session:Geometric Complexities Along Strike-Slip Systems: New Insights on Seismic Hazards, Earthquake Behavior, and Fault System Evolution
Presenter   Onderdonk, Nate
Schedule   Wed / Poster
Room   Cook/Arteaga
Paleoseismology of the Northern San Jacinto Fault Zone and Implications for Rupture across Steps in the San Andreas Fault System, Southern California.
ONDERDONK, N. W., California State University, Long Beach, Long Beach, CA, USA, nate.onderdonk@csulb.edu; MCGILL, S. F., California State University, San Bernardino, San Bernardino, CA, USA, smcgill@csusb.edu; ROCKWELL, T. K., San Diego State University, San Dieogo, CA, USA, trockwell@mail.sdsu.edu
The San Jacinto fault is one of the major plate boundary faults in southern California, and recent geologic and geodetic data suggest it accommodates about one third of the total slip across the southern San Andreas fault system. To investigate the timing of past surface ruptures over the last 2000 years, we studied the prehistoric earthquake history on the northern SJF zone as recorded in shallow sediments at Mystic Lake, an ephemeral lake that formed at the north end of a 2.25 km-wide pull-apart basin between the Claremont and Clark segments of the San Jacinto fault zone. Between 2009 and 2013, we excavated nine trenches across the Claremont fault where it crosses through the north end of Mystic Lake. These trenches exposed evidence for 15 ground-rupturing earthquakes over the past 3700 years, with 12 of these occurring in the last 2000 years. 118 radiocarbon dates were used to determine the earthquake history recorded at the site. We compared the timing of events at Mystic Lake to the earthquake records at the Hog Lake site on the Clark fault to the south, and the Wrightwood site on the San Andreas fault to the north to evaluate the possibility and frequency of rupture across step-overs at both ends of the Claremont fault. At least six of the 12 earthquakes that occurred in the last 2000 years at the Mystic Lake site overlap in time with earthquakes at Hog Lake, suggesting that both the Claremont and Clark faults may rupture together in some large events. Between Mystic Lake and Wrightwood, four of nine earthquakes in the past 1500 years overlap in time, suggesting that some events may also have ruptured through the step-over between the northern end of the Claremont fault and the San Andreas fault. However, there are no events at Mystic Lake that overlap with both the Hog Lake and Wrightwood records, so we consider it unlikely that any earthquakes have ruptured across both ends of the Claremont fault in very large, multi-fault cascade events.
Session:New Insights into Ground Failure as an Urban Earthquake Hazard
Presenter   Moss, Robb
Schedule   Thu / Poster
Room   Cook/Arteaga
Influence of Near Surface Stiffness, Material Type, and Prior Rupture History on Reverse Fault Surface Rupture
STANTON, K. V., UNR, Reno, NV, USA, kevinstanton@unr.edu; MOSS, R. E. S., Cal Poly, San Luis Obispo, CA, USA, rmoss@calpoly.edu
The stiffness of the upper thirty meters of geologic material has been shown statistically to have a strong influence on the propagation of fault rupture to the ground surface for reverse events, based on empirical data from past earthquakes (Moss et al., 2013). To further investigate this, a series of fault box experiments that modeled reverse faulting were run on both sand and clay of varying stiffness. These lab experiments found that a relatively “soft” material accommodated on average 88% more normalized base displacement than a relatively “stiff” material, supporting the empirical field data. The material type (sand versus clay) had an even greater impact; with sand accommodating on average 175% more normalized base displacement than clay under similar conditions. When considering prior rupture, virgin material accommodated on average 411% more normalized fault displacement than previously ruptured material. These three variables; near surface stiffness, material type, and prior rupture history have been found experimentally to be strong indicators of the likelihood of surface fault rupture in reverse faulting events. Future research is investigating the numerical modeling of these effects to complement the completed statistical and experimental studies.
Session:Induced Seismicity
Presenter   Ackerley, Nick
Schedule   Thu / Poster
Room   Cook/Arteaga
Microseismic Network Performance Estimation: Comparing Predictions to an Earthquake Catalogue
GREIG, D. W., Nanometrics Inc., Kanata, ON, Canada, WesGreig@nanometrics.ca; ACKERLEY, N. J., Nanometrics Inc., Kanata, ON, Canada, NickAckerley@nanometrics.ca
The design of networks for monitoring induced seismicity is of critical importance as specific standards of performance are necessary. One of the difficulties involved in designing networks for monitoring induced seismicity is that it is difficult to determine whether or not the network meets these standards without first developing an earthquake catalog. We develop a tool that can assess two key measures of network performance without an earthquake catalog: location accuracy and magnitude of completeness. Site noise is measured either at existing seismic stations or as part of a noise survey. We interpolate measured values to determine a noise map for the entire region. This information is combined with instrument noise for each station to accurately assess total ambient noise at each station. Location accuracy is evaluated according to the approach of Peters and Crosson (1972). Magnitude of completeness is computed by assuming isotropic radiation and mandating a threshold signal to noise ratio (similar to Stabile et al. 2013). We apply this tool to a seismic network in the central United States. We predict the magnitude of completeness and the location accuracy and compare predicted values with observed values generated from the existing earthquake catalog for the network. We investigate the effects of hypothetical station additions and removals to a network simulating network expansions and station failures. We find that the addition of stations to areas of low noise results in significantly larger improvements in network performance than station additions to areas of elevated noise, particularly with respect to magnitude of completeness. Our results highlight the importance of site noise considerations in the design of a seismic network. The ability to predict hypothetical station performance allows for the optimization of seismic network design and enables the prediction of performance for a purely hypothetical seismic network.
Session:Diverse Mechanisms of Subduction Zone Fault Slip: Exploring the Relationships Among Seismic Rupture, Transient Slip, and Steady Creep
Presenter   Fry, Bill
Schedule   Wed 11:30 AM / Oral
Room   Room 1
Observations of SSE, Tremor, and Damaging Earthquakes on the Southern Hikurangi Margin, New Zealand
FRY, B. N., GNS Science, Lower Hutt, New Zealand, b.fry@gns.cri.nz; WALLACE, L., University of Texas, Austin, TX, USA, lwallace@utexas.edu
The Hikurangi margin was the site of 3 large slow slip events (SSEs) during 2013. These included a recurrence of the deep (30-50 km depth) long-term Kapiti SSE that began in January 2013 and is currently (January 2014) ongoing. Moment release in this event is equivalent to an Mw7.0 earthquake. Previous Kapiti SSEs occurred in 2003 and 2008. The 2013 event activated non-volcanic tremor in at least two discrete regions around the SSE. We use a template matching filter to search continuous data from 2013 for arrival times later used to locate the events. We determine our range of templates by manually inspecting waveforms from high quality permanent stations located around the tremor patches. We use multiple template durations to increase the variability in our recovered tremor events. The relative time of the maximum value of the correlation coefficient between the template and the continuous recordings are employed as ballistic arrival times to determine event locations with a non-linear location algorithm. With this approach, we have very little resolution of depth estimates for the tremor. However, if we assume the source is located on or near the plate interface, the events are occurring between about 20 and 35km depth. Interestingly, one of the tremor patches lies up-dip of the SSE, near a region of crustal earthquake activation later in the SSE cycle that included a damaging M6.6 event.
Session:Geometric Complexities Along Strike-Slip Systems: New Insights on Seismic Hazards, Earthquake Behavior, and Fault System Evolution
Presenter   Lozos, Julian
Schedule   Wed / Poster
Room   Cook/Arteaga
Rupture and Ground Motion Models on the Claremont-Casa Loma Stepover of the San Jacinto Fault, Incorporating Realistically Complex Initial Conditions
LOZOS, J. C., Pacific Earthquake Engineering Research Institute (UC Berkeley), Berkeley, CA, USA, jlozos@berkeley.edu; OLSEN, K. B., San Diego State University, San Diego, CA, USA, kbolsen@mail.sdsu.edu; OGLESBY, D. D., University of California, Riverside, Riverside, CA, USA, david.oglesby@ucr.edu; BRUNE, J. N., University of Nevada, Reno, Reno, NV, USA, brune@seismo.unr.edu
The stepover between the Claremont and Casa Loma strands of the San Jacinto Fault in southern California is a major discontinuity within a complex fault zone. This extensional stepover has an average of 4 km separation between the strands over an overlap length of 25 km. A primary question is whether rupture can jump between the Claremont and Casa Loma in one event. This region poses other questions about the dynamics of complex faults – in particular, whether small-scale geometrical complexity, contrasts in velocity structure, and heterogeneous initial stresses can control rupture behavior as much as the large stepover can, and how all of these factors contribute to ground motion. We use the 3D finite element method to conduct dynamic rupture models on a complex parameterization of this stepover. We implement fault geometry based on field mapping literature, material properties from the SCEC Community Velocity Model, a regional stress orientation taken from seismicity relocation literature, and several stochastic shear stress distributions. The stress field that arises from the combination of regional and stochastic stresses and fault geometry is the primary control on rupture extent. The contrast between high and low shear stress patches, rather than the magnitude of the stresses, determines the strength of a barrier. Greater contrast and smaller patches of high shear stress produce shorter ruptures. In most of our models, this stress pattern prevents rupture from reaching the end of either fault strand and having to negotiate the stepover. Smaller events that do not always reach the surface are consistent with historical behavior on the San Jacinto Fault. The complex velocity structure has more effect on ground motion distribution and intensity than on rupture extent. We find that the Claremont strand is more favorable for longer ruptures than the Casa Loma strand, and that the strongest ground motions are concentrated in the San Bernardino basin.
Session:Fifty Years of Tsunami Science: from the 1964 Earthquake and Tsunami to the SAFRR Tsunami Scenario - Advances in Tsunami Source Characterization, Numerical Analysis and Hazard Mitigation
Presenter   Wertman, Christina
Schedule   Fri 4:45 PM / Oral
Room   Room 4
STUDENT
Meteotsunamis Generated by Mesoscale Convective Systems along the U.S. East Coast on June 13th, 2013
WERTMAN, C. A., University of Rhode Island, Narragansett, RI, USA, caking@my.uri.edu; YABLONSKY, R. M., University of Rhode Island, ; SHEN, Y., University of Rhode Island, ; MERRILL, J., University of Rhode Island, ; KINCAID, C. R., University of Rhode Island, ; POCKALNY, R. A., University of Rhode Island,
Two high-frequency sea level oscillation events occurred on June 13th, 2013 along the east coast of the United States, causing injuries and property damages. No earthquakes or other known geologic triggers were detected near the coast at the time of the events. However, the events coincided with the passage of two eastward propagating mesoscale convective systems (MCSs) that produced high-frequency atmospheric surface pressure anomalies. While the MCSs were over land, the USArray Transportable Array recorded the associated pressure anomalies in detail; as the pressure anomalies moved offshore, they were recorded by NOAA tide gauge sensors. As the MCSs propagated over the ocean, the pressure anomalies generated shallow water waves, which were amplified by Proudman resonance under an approximately Froude number one regime. Analysis of the NOAA tidal data reveals that after the atmospheric forcing is negligible, these shallow water waves (i.e. meteotsunamis) propagated freely, reflected off the continental shelf break, and arrived at various coastal locations, where the local bathymetry and coastal geometry contributed to the meteotsunamis’ destructive impacts. This case study demonstrates that it is possible and perhaps advisable to quantify pressure anomalies associated with MCSs in the interior of the continental United States, which could lead to real-time prediction of MCS-generated meteotsunamis and set the stage for a meteotsunami warning system.
Session:Diverse Mechanisms of Subduction Zone Fault Slip: Exploring the Relationships Among Seismic Rupture, Transient Slip, and Steady Creep
Presenter   Hyndman, Roy
Schedule   Wed 11:45 AM / Oral
Room   Room 1
The Downdip Location of ETS and Dehydration Fluids Channelled to Forearc Mantle Corner
HYNDMAN, R. D., Pacific Geoscience Centre, Geological Survey of Canada, Sidney, BC, Canada, rhyndman@nrcan.gc.ca; MCCRORY, P. A., U.S. Geological Survey, Menlo Park, CA, USA, pmccrory@usgs.gov; WECH, A. G., Alaska Volcano Observatory, US. Geological Survey, Anchorage, AK, USA, awech@usgs.gov
We first show that for Cascadia the zone of Episodic Tremor and Slip (ETS) is not a downdip continuation of the seismogenic zone. Based on our recently published reviews, ETS occurs ~50 km downdip of the seismogenic zone which is concluded to be thermally limited for this hot subduction zone. We then address the question of what does control the location of ETS. In another review we have shown that the location of ETS coincides remarkably well with the forearc mantle corner for the whole Cascadia margin. This association can be explained by a focusing in this corner of the fluids that are released from deep subducting oceanic crust with downdip increasing temperature and pressure. It has been inferred from seismic velocity data that the overlying forearc mantle in Cascadia is strongly serpentinized from these fluids, and laboratory data show that such serpentinite has very low permeability. Once sufficient impermeable serpentinite is formed, deep-dehydration fluids should flow updip in the permeable underlying subducting crust to the forearc mantle corner where they are released upward into the more permeable forearc crust. As well as the inference from the coincidence of ETS and the forearc mantle corner, indications of this concentrated fluid flux come from seismic tomography. Exceptionally low Vp/Vs is found in the forearc continental crust above the corner that suggests a concentration of silica; quartz is the only common mineral that has adequately low Vp/Vs and silica deposition is expected from the large integrated volume of rising fluids over the ~40 m.y. of the current subduction regime. The fluids should be silica saturated and silica solubility decreases rapidly upward with decreasing temperature and pressure. In several studies elsewhere, concentrated quartz vein arrays in deeply eroded sections have been explained by episodic silica deposition from the large integrated fluid flux above subducted plates.
Session:Topics in Seismology: Processes
Presenter   Delorey, Andrew
Schedule   Thu / Poster
Room   Cook/Arteaga
The Response of Elastic Systems in the Earth to Dynamic Perturbations
DELOREY, A. A., Los Alamos National Laboratory, Los Alamos, NM, USA, andrew.a.delorey@gmail.com; CHAO, K., University of Tokyo, Tokyo, Japan, kevinchao@gmail.com; UCHIDA, N., Tohoku University, Sendai, Japan, uchida@aob.gp.tohoku.ac.jp; OBARA, K., University of Tokyo, Tokyo, Japan, obara@eri.u-tokyo.ac.jp; JOHNSON, P. A., Los Alamos National Laboratory, Los Alamos, NM, USA, paj@lanl.gov
Seismic waves have been observed to perturb the elastic properties of the crust at great distances from the hypocenter. These changes in elastic properties can be observed directly through seismic wave speeds or indirectly by observing changes in the rate and character of seismic emissions and aseismic processes like slow slip, changes in pore pressures, and changes in the packing of granular materials. Characterizing temporal variations in elastic properties is an emerging area of research with important implications for earthquake nucleation processes, earthquake forecasting, and seismic hazards. Here we show that by combining an earthquake catalog analysis with additional observations on temporal and spatial changes in microseismicity, seismic velocities, slip rates, and repeating earthquakes we can observe the signatures of dynamically induced changes in elastic properties in the southern part of the aftershock zone of the 2011 M9.0 Tohoku-Oki earthquake (TOE) following the 2012 M8.6 Indian Ocean earthquake (IOE). In the days following the IOE, several clusters of seismicity in the shallow accretionary wedge are initiated along a north to south trend over 150 km distance, which coincides with a change in behavior of repeating earthquakes on the plate interface and changes in seismic velocities in the shallow crust onshore. These observations are unique in the post-TOE period and are signatures of a dynamically perturbed elastic system.
Session:Recent Advances in Ground Motions Simulation Methods and Their Validation
Presenter   Shi, Zheqiang
Schedule   Thu 11:30 AM / Oral
Room   Room 3
Rupture Dynamics and Ground Motions from 3D Dynamic Rough-Fault Simulations of Dip-Slip Events
SHI, Z., Department of Geological Sciences, San Diego State University, San Diego, CA, USA, zshi@mail.sdsu.edu; DAY, S. M., Department of Geological Sciences, San Diego State University, San Diego, CA, USA, sday@mail.sdsu.edu
We perform 3D numerical simulations of dynamic rupture along rough dip-slip faults to study the properties of rupture dynamics and patterns of resultant ground motions in dip-slip events. Shi and Day (2013) deterministically generated high-frequency ground motions for up to ~10 Hz by numerically modeling rupture propagation along a rough strike-slip fault. 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 site-averaged synthetic response spectra have 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. In the current study we apply the methodology of Shi and Day (2013) to the deterministic simulation of high-frequency ground motions for dip-slip events by assuming that the surface topology of the dipping fault follows a self-similar roughness pattern. We explore the influence of geometrical and dynamic parameters of faulting on properties of rupture dynamics and ground motions. Parameters explored include fault dip, rake angle, depth to the top-of-rupture and stress distribution, with the goal of better understanding how ground motion phenomena such as directivity and hanging-wall amplifications are related to the dynamic and geometrical characteristics of dip-slip events.
Session:Fault Structure, Heterogeneity, and Implications for Rupture Dynamics
Presenter   Wu, Chunquan
Schedule   Wed 5:30 PM / Oral
Room   Room 3
Slip Characteristics of the Deep Portion of the San Andreas Fault Inferred from Low-Frequency Earthquakes
WU, C., Geophysics Group, Los Alamos National Laboratory, Los Alamos, NM, USA, cwu@lanl.gov; GUYER, R. A., Physics Department, University of Nevada, Reno, NV, USA, guyer@physics.umass.edu; TRUGMAN, D. T., Geophysics Group, Los Alamos National Laboratory, Los Alamos, NM, USA, dtrugman@lanl.gov; SHELLY, D. R., US Geological Survey, Menlo Park, Menlo Park, CA, USA, dshelly@usgs.gov; JOHNSON, P. A., Geophysics Group, Los Alamos National Laboratory, Los Alamos, NM, USA, paj@lanl.gov
The dynamics of fault slip in the lower crust are just beginning to be understood due to the recent discovery of non-volcanic tremor (NVT) and Low Frequency Earthquakes (LFEs). These phenomena occur in the deep crust where material properties are more ductile than in the brittle upper crustal seismogenic zone. Here we examine timing characteristics of LFE repeating sources of the deep crustal slip along the Parkfield segment of the San Andreas Fault (SAF). We apply an automatic LFE burst detection algorithm to all the LFE repeating sources to track the similarity and dis-similarity in source behaviors. We show that LFE repeating sources are comprised of bursts whose timing may be highly regular, highly irregular or intermediate in nature. We show further the coherence of burst activity among different sources varies along SAF, suggesting strong and weak coupling of different patches of fault. By comparing the observations to laboratory biaxial shear experiments, we infer possible complex slip regimes in the deep crust that vary along the fault, indicating that some portions exhibit stick-slip behavior, some are stably sliding, and others have a mixed slipping behavior.
Session:Explosive Source Characterization
Presenter   Napoli, Vanessa
Schedule   Wed 5:00 PM / Oral
Room   Room 1
STUDENT
Ms Unified: A New Magnitude for Rayleigh and Love Waves with Application in the Korean Peninsula and Yellow Sea Region
NAPOLI, V., Boston College, Chestnut Hill, MA, USA, napoliv@bc.edu; BONNER, J., Weston Geophysical Corp., Lufkin, TX, USA, jes_bonner@westongeo.com; RUSSELL, D., Weston Geophysical Corp., Satellite Beach, FL, USA, dhrussell@westongeophysical.com
In order to improve discrimination of small explosions and earthquakes, we need to change the focus of magnitude estimation from large scale network sampling to maximizing the available information from single stations and then combining into network estimates, designed to be effective with sparse network coverage. We have developed a new magnitude scale, based on the Ms(VMAX) technique of Russell (2006), that incorporates both Love and Rayleigh wave amplitudes for a unified surface wave magnitude. The Love and Rayleigh waves are narrow band filtered and corrected for propagation and source effects at periods between 8-25 seconds, in order to find filter bands of maximum energy propagation. The unified magnitudes (termed MsU) are calculated from average amplitudes of both Rayleigh and Love waves at each station, measured at each period, and then scaled to standard 20 sec Rayleigh magnitudes. This is essential to minimize the effect of earthquake radiation patterns for sparse networks. The data are also corrected for censoring effects at the station level given that either Rayleigh or Love waves may not be observed for an event. We have applied MsU to 39 earthquakes (3.21 <Mw <5.08) located in the Korean Peninsula and Yellow Sea region as recorded at near-regional distances (50-600 km). For comparison, we also estimated magnitudes based only on Rayleigh (MsR) and Love (MsL) waves. The MsU estimate consistently falls between the MsR and MsL estimates, with a significant reduction in outliers. The standard deviations of residuals for the different magnitudes are 0.15 (MsU), 0.19 (MsR), and 0.23 (MsL), with significantly improved stability of the MsU residuals as a function of distance. The reduction in variance for MsU will provide a magnitude that is a better indicator of source size for single-station and sparse network estimates, which has important implications in discriminating earthquakes and explosions at small magnitudes.
Session:Pillars of Simulation: Seismic Velocity and Material Models
Presenter   Plesch, Andreas
Schedule   Wed 3:00 PM / Oral
Room   Room 4
Stochastic Descriptions of Fine-Scale Basin Velocity Structure from Well Logs and the SCEC Community Velocity Model (CVMH)
PLESCH, A., Harvard University, Cambridge, MA, USA, andreas_plesch@harvard.edu; SHAW, J. H., Harvard University, Cambridge, MA, USA; SONG, X., University of Southern California, Los Angeles, CA, USA; JORDAN, T. H., University of Southern California, Los Angeles, CA, USA
The trend of numerical wave propagation studies to higher frequencies (> 2 Hz) has created a demand for higher resolution velocity models. Thus, we are developing a statistical description of fine-scale velocity structure, informed by local borehole observations and geological correlations, with the goal of enhancing community models so that they can support higher-frequency simulations. We examined the velocity (vp) structure in 108 wells across the Los Angeles basin and within a large oil field in the same region that provides more than 400k velocity samples from 70 deviated wells. These logs sample velocity to approximately 1m scales; the field is also constrained by 3D seismic reflection data. After trend removal and despiking, the data show symmetric distributions of velocity variability relative to the CVMH centered around a mean of 0% and with standard variations of 6.5% to 6.9%. The distributions are highly non-Gaussian: the kurtoses are high with values of ca. 13 in aggregate and up to 50 in individual wells. Using logarithmic frequencies, there is a linear drop-off from the mean to both sides indicating a power-law distribution of velocity variability. Second-order, but potentially important features include a small upward curvature on the log-log plot, indicating a fatter-tailed distribution than a simple power law, and an asymmetry in this curvature about the mean value. Vertical correlation lengths are on the order of 80 m. Further, the dense data sample suggests that horizontal correlation lengths are about 25 times that of vertical correlation lengths, defining a highly anisotropic nature of basin velocity structures. Finally, we show that lateral velocity structures are more highly correlated along stratigraphy than in the horizontal dimension in areas of dipping beds. This suggests that stratigraphic horizons may provide a means to guide the implementation of a stochastic description of fine-scale velocity structure in the basin models.
Session:Recent Advances in Ground Motions Simulation Methods and Their Validation
Presenter   Savran, William
Schedule   Thu / Poster
Room   Cook/Arteaga
STUDENT
Deterministic Simulation of the Mw 5.4 Chino Hills Event with Frequency-Dependent Attenuation, Heterogeneous Velocity Structure and Realistic Source Model
SAVRAN, W. H., San Diego State University, San Diego, CA, USA, wsavran@ucsd.edu; OLSEN, K. B., San Diego State University, San Diego, CA, USA, kbolsen@mail.sdsu.edu
We present simulations of the 2008 Mw 5.4 Chino Hills earthquake for frequencies 0.1 to 2.5Hz. Our model includes frequency-dependent attenuation with power law Q(f)=Q0fn (Withers et al., 2013), statistical models of small-scale heterogeneities in the medium, and a finite-fault source description (Shao et al., 2012). The background medium is the Southern California Earthquake Center (SCEC) community velocity model version 4 (CVM-4). We generate synthetic seismograms at 14 stations in the Los Angeles Area and perform sensitivity tests on the parameters of the small-scale heterogeneities, guided by analysis of sonic logs. Our results show that key ground motion parameters, such as peak ground velocity (PGV), peak ground acceleration (PGA), cumulative kinetic energy (CKE) and Arias intensity (AI) are affected by the statistical model parameters up to 60% at some stations. End-member tests show that correlation length (L) can change CKE, PGV, and AI by up to 30%. Varying the Hurst exponent (H) from 0.0 to 0.5 causes a slight increase in CKE and AI, but does not result in systematic changes of PGV or PGA. Spectral analysis shows the statistical model primarily affects frequencies greater than 1Hz. Based on comparisons with data, we recommend an attenuation relationship where Qs0=0.10Vs to Qs0=0.15Vs, and statistical models of small-scale heterogeneities with L=150m, H=0.0-0.1 and standard deviations of 5-10%. The statistical models include a horizontal-to-vertical anisotropy factor (~5), based on findings from borehole analysis (Plesch and Shaw, 2013). We find reasonable fits for ground motion parameters and frequency content; however, at some of the stations, particularly on hard-rock sites, the small-scale heterogeneities do not account for the misfit between observed and synthetic seismograms. As a result of this, we conclude the seismic velocities in the upper hundreds of meters are too large in the CVM-4 at these sites.
Session:Effects of Topography and Surface Loads on Earthquakes and Faulting
Presenter   Amosu, Adewale
Schedule   Wed / Poster
Room   Cook/Arteaga
STUDENT
Crustal Deformation from Surface Loading in the Great Salt Lake Region
AMOSU, A. M., CERI, Memphis, TN, USA, walbytes@yahoo.com; SMALLEY, R., CERI, Memphis, TN, USA, rsmalley@memphis.edu
We estimate the 3-component surface displacements associated with loading due to the Great Salt Lake volume variations between 1998 and 2013 using GPS stations located around the lake in North Utah. The vertical and horizontal displacements observed are inversely correlated with the lake stage levels. Our goal is to use the observed response of the Earth’s crust to estimate its elastic properties. To do this, we model the interaction of the Earth’s surface with the load. We use topographic, lake stage, groundwater, and atmospheric pressure data with an elastic half-space model of the earth to estimate the surface loading and response in the study region. We invert the observed deformation to estimate the Young’s modulus from which other seismic properties of the crust can be inferred.
Session:Pillars of Simulation: Seismic Velocity and Material Models
Presenter   Lee, En-Jui
Schedule   Wed 2:30 PM / Oral
Room   Room 4
Full-3D Waveform Tomography for Crustal Structure in Southern California Using Earthquake Recordings and Ambient-Noise Green’s Functions Based on the Adjoint and the Scattering-Integral Methods
LEE, E., University of Wyoming, Laramie, WY, USA, rickli92@gmail.com; CHEN, P., University of Wyoming, Laramie, WY, USA; JORDAN, T. H., University of Southern California, Los Angeles, CA, USA; MAECHLING, P. J., University of Southern California, Los Angeles, CA, USA; DENOLLE, M., Stanford University, Stanford, CA, USA; BEROZA, G. C., Stanford University, Stanford, CA, USA
We apply a unified methodology for seismic waveform analysis and inversions to Southern California. We developed a semi-automatic seismic waveform analysis algorithm for full-wave earthquake source parameters and tomographic inversions. The algorithm is based on continuous wavelet transforms, a topological watershed method, and a set of user-adjustable criteria to select usable waveform windows for full-wave inversions. The algorithm takes advantages of time–frequency representations of seismograms and is able to separate seismic phases in both time and frequency domains. The selected wave packet pairs between observed and synthetic waveforms are then used for extracting frequency-dependent measurements, which are used in our seismic source and structural inversions. Our full-wave waveform tomography uses the SCEC Community Velocity Model Version 4.0 (CVM-S4) as the initial model, a staggered-grid finite-difference code to simulate seismic wave propagations. 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 functions in our full-3D waveform tomographic inversions. To reduce errors of earthquake sources, the epicenters and source parameters of earthquakes used in our tomographic inversion are inverted by our full-wave CMT inversion method. After 26 iterations, our improved model (CVM-S4.26) shows many features that relate to the geological structures at shallow depth and contrasting velocity values across faults. The velocity perturbations could up to 45% with respect to the initial model in some regions and relate to some structures that do not exist in the initial model, such as southern Great Valley. The waveform misfits reduce over 70% when compared with that at the initial stage. The new tomography enables more accurate physics-based seismic hazard applications and the tectonic reconstruction of Southern California.
Session:Cold Climate Installation Techniques and Instrumentation Developments for Temporary and Long-Term Networks
Presenter   Hart, Darren
Schedule   Wed / Poster
Room   Cook/Arteaga
Development of Seismic Field System for Multi-Year Field Deployment at Poker Flat Research Range, Alaska
HART, D., Sandia National Laboratories, Albuquerque, NM, USA, dhart@sandia.gov; ABBOTT, R. E., Sandia National Laboratories, Albuquerque, NM,USA
The environment in polar regions poses a challenge for successful seismic deployments. Cold temperatures and lack of winter sunlight are obvious obstacles, but several other problems (such as remoteness and damage-causing animals) must be overcome, as well. We report on our design, installation, and maintenance of a small-aperture seismic array at Poker Flat Research Range, Alaska. Among the system requirements were: 1) Fully telemetered at sample rates exceeding 100 Hz per channel; 2) Controlled remotely for most operations; 3) Shallowly buried in permafrost; 4) Year round operation; and 5) Robust tilt tolerance. Implicit, as always, is that the system must be built within a finite budget. In our poster, we discuss the steps taken to select candidate components, testing of components that met our requirements, and then summarize the advantages and disadvantages of the components used in the final system. 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:75 Years of Frequency-Size-Distribution of Earthquakes: Observations, Models and Understanding
Presenter   Shcherbakov, Robert
Schedule   Fri / Poster
Room   Cook/Arteaga
Statistical Properties of Aftershock Sequences Generated by Large Subduction Earthquakes
SHCHERBAKOV, R., Department of Earth Sciences, Western University, London, ON, Canada, rshcherb@uwo.ca; GODA, K., Department of Civil Engineering, University of Bristol, Bristol, United Kingdom, Katsu.Goda@bristol.ac.uk; ATKINSON, G., Department of Earth Sciences, Western University, London, ON, Canada, Gmatkinson@aol.com
Large mega-thrust subduction earthquakes generate prolific aftershock sequences which last over an extended period of time and affect wide spatial areas. Among those aftershocks the largest ones can cause additional damage and pose significant risks to population and infrastructure. Therefore, modeling aftershock sequences of large subduction earthquakes is of considerable importance for seismic hazard assessment and earthquake risk mitigation. It can also play a prominent role in the ground shaking modeling of major mainshock-aftershock sequences. In this work, we analyze statistical properties of aftershock sequences of large subduction earthquakes world-wide which occurred from 1973 to present, including recent catastrophic events in Sumatra, Chile, and Japan. We use information provided in the NEIC catalog to extract 70 aftershock sequences generated by mainshocks of M7.0 and above. We construct their temporal decay rates and magnitude-frequency statistics. To model their temporal behavior, we estimate the parameters of the modified Omori law. In the magnitude domain, we model the frequency-magnitude statistics using the Gutenberg-Richter scaling relationship. We also analyze statistically the difference between the magnitude of the mainshock and the corresponding largest aftershock in the sequence and discuss this in terms of Båth’s law. One of the main goals of this work is to investigate the variation in parameter values of the above empirical laws with respect to the magnitude of the mainshock. Our main finding indicates that most parameters do not depend on the magnitude of the mainshock. However, they show some variation in values across different subduction settings.
Session:Pillars of Simulation: Seismic Velocity and Material Models
Presenter   Morozov, Igor
Schedule   Wed 4:45 PM / Oral
Room   Room 4
Physical Characterization of Seismic Attenuation in Earth Models
MOROZOV, I. B., University of Saskatchewan, Saskatoon, SK, Canada, igor.morozov@usask.ca
Seismic attenuation represents an important part of Earth’s models at all scales, from lab experiments with rock samples and engineering applications to mantle flows, free oscillations of the whole Earth, and tides. At present, all attenuation models are formulated in terms of the quality factor (Q), which is a specialized property attributed to the materials in order to explain the internal friction occurring during deformation. Special ‘principles’ for solving attenuating-wave problems (Boltzmann’s and correspondence principles) and rules for their implementation in numerical modeling (such as memory variables) are constructed to implement this model. However, these principles do not automatically agree with physics. The Q factor is a phenomenological property that may not always successfully represent all aspects of internal friction. Examples where the Q model can be problematic include: 1) distribution of frictional stresses in poroelastic media, 2) conservation of energy in surface waves, and 3) zero attenuation within the liquid outer core. To resolve such problems, we do not rely on the Q and the associated viscoelastic ‘principles’ and propose direct use of physical laws, such as the solid and fluid viscosity, thermoelasticity, poroelasticity, and other. This approach was used in early attenuation studies by Jeffreys, Knopoff, and Ricker, and it readily follows from classical continuum mechanics and thermodynamics. Time-integral relations for ‘material memory’ are replaced with instantaneous differential equations including frictional forces, and the ‘frequency dependences of Q’ are explained by the dependences of physical mechanisms (such as viscosity) on strains and strain rates. This approach guarantees causality and conservation of energy, and it is naturally implemented in finite-difference (FD) and finite-element algorithms. The approach is illustrated on FD modeling of wave propagation in a medium with nonlinear viscosity.
Session:Explosive Source Characterization
Presenter   Morozov, Igor
Schedule   Wed 3:15 PM / Oral
Room   Room 1
Modeling rel as the Nonlinear Inelastic-Elastic Transition near Explosion Source
MOROZOV, I. B., University of Saskatchewan, Saskatoon, SK, Canada, igor.morozov@usask.ca; STROUJKOVA, A., Weston Geophysical Corporation, Lexington, MA, USA, ana@westongeophysical.com
The ‘equivalent-cavity,’ or ‘elastic’ radius (denoted rel) separating the near- and far-field zones is the critical parameter of all explosion source models. While being measured from the far-field seismic data, the principal use of rel is in constraining the properties of the source. Conventional scaling laws for this radius developed from nuclear explosions do not always work for smaller and shallower chemical shots. Here, we investigate a physical definition of rel as the radius of the inelastic-elastic transition, i.e., the proximity to the source at which nonlinear effects become dominant. Three groups of mechanisms contribute to such nonlinearity: 1) dynamic strengthening of the material and increasing toughness of fractures under fast loading; 2) rock failure, damage, fragmentation, and fracturing, and 3) interactions with the free surface and preexisting near-surface structures. The first two of these groups can be described by the concept of effective medium with generalized, nonlinear viscoelastic rheology. Previous attempts for modeling such rheologies focused on constructing a strain-dependent quality factor (Q) for the medium. However, it is difficult to constrain or assume a meaningful frequency- dependent Q at near-source conditions. We propose a much broader, equivalent-medium approach using the concepts of nonlinear damage rheology and solid viscosity. These concepts are compatible with continuum mechanics, and thermodynamics, and they lead to straightforward finite-difference (FD) schemes. The waveforms and spectra of the resulting seismograms and the effective frequency-dependent Qs of the medium are explained by the dependences of elasticity and viscosity on strains and strain rates and by contributions of internal variables, such as temperature and the degree of damage. The model is illustrated by FD simulations of a radially-symmetric chemical source in an isotropic medium.
Session:From the Earthquake Source to Damage of Buildings: Bridging the Gap between Seismology and Earthquake Engineering
Presenter   Dalguer, Luis
Schedule   Fri / Poster
Room   Cook/Arteaga
STUDENT
3D Spontaneous Dynamic Rupture on Geometrically Complex Faults: The 2010 Mw 7.1 Darfield (New Zealand) Earthquake
ABOLFATHIAN, N., ETH Zurich, Zurich, Switzerland, nabolfat@student.ethz.ch; GALVEZ, P., ETH Zurich, Zurich, Switzerland, percy.galvez@sed.ethz.ch; DALGUER, L. A., Swiss Seismological Service, Institute of Geophysics, ETH Zurich, Zurich, Switzerland, dalguer@sed.ethz.ch
Field observations suggest that natural faults have a diverse degree of geometrical complexity such as branching and segmentations. Recent studies show that these non- planar geometrical complexities have dominant effects on the rupture dynamic process and near source ground motion. The 2010 Mw 7.1 Darfield (New Zealand) earthquake is one of the best recorded earthquake that exhibited geometrically complex faults, in which includes a vertical fault comprising bending and dipping branches. The rupture nucleates on a thrust branch fault and propagates on a strike-slip main fault with branches. We developed a 3D dynamic rupture model of this earthquakes considering its geometrical fault complexity. The unstructured 3D open source spectral element code SPECFEM3D is used for this investigation. This code allows modeling realistic non-planar fault geometries. Mesh generation software, CUBIT, is applied which employs hexahedral unstructured elements. Our simulations well describes the slip and rupture velocity from published kinematic source models obtained by source inversion. The role of the fault geometry on rupture dynamic process and stress changes and its effects on near-source ground motion is investigated.
Session:Induced Seismicity
Presenter   Weiser, Deborah
Schedule   Thu / Poster
Room   Cook/Arteaga
STUDENT
Geothermal Pumping and Induced Seismicity in California Geothermal Fields
WEISER, D. A., UCLA and USGS, Los Angeles and Pasadena, CA, USA, dweiser@ucla.edu; JACKSON, D. D., UCLA, Los Angeles, CA, USA, david.d.jackson@ucla.edu; JONES, L. M., USGS, Pasadena, CA, USA, jones@usgs.gov
Geothermal energy is an important source of power. However, an impediment to future growth is a concern over the potential that geothermal production and other pumping technologies can induce earthquakes, as observed in New Zealand, Switzerland, the US and elsewhere. Induced earthquakes in California could potentially trigger large earthquakes on the San Andreas and other major faults. Thus, it is critical to better understand the relationship between geothermal production and induced seismicity. We examine the relation between fluid volume change and seismicity criteria detailed in Davis and Frohlich (1993) at each geothermal field, and analyze data for other conditions or correlations that imply induced earthquakes have occurred. Initial results suggest that seismicity often increases when a new geothermal field begins pumping, and that there are temporal correlations between fluctuations in net fluid volume and seismicity. Monthly injection and extraction totals for geothermal fields in California are freely available from the California Department of Oil, Gas, and Geothermal Resources. Seismicity data are available from the Waveform Relocated Earthquake Catalog for Southern California (Hauksson et al., 2012), and the Double Difference Earthquake Catalog for Northern California (Waldhauser and Schaff, 2008). Southern California seismicity data are supplemented with Southern California Seismic Network (SCSN) data from 2011-2013 (Hutton et al., 2010).
Session:Near-Field Seismoacoustics of Natural and Man Made Explosions
Presenter   Sonder, Ingo
Schedule   Fri / Poster
Room   Cook/Arteaga
Large Scale Blast Experiments Integrate Field, Remote Sensing, and Laboratory Based Techniques
SONDER, I., University at Buffalo, Buffalo, NY, USA, ingomark@buffalo.edu; GRAETTINGER, A. H., University at Buffalo, Buffalo, NY, USA, ahgraett@buffalo.edu; VALENTINE, G. A., University at Buffalo, Buffalo, NY, USA, gav4@buffalo.edu; BOWMAN, D. C., University of North Carolina at Chapel Hill, Chapel Hill, NC, USA, daniel.bowman@unc.edu; LEES, J. M., University of North Carolina, Chapel Hill, NC, USA, jonathan.lees@unc.edu
Many problems in Geophysics arise from complex scaling behavior, and therefore bench top studies cannot be used to appropriately investigate all mechanisms critical for scaling to natural scenarios. Such inversion problems are addressed at the Geohazards Field Station (GFS), a user facility hosted by University at Buffalo, by setting up larger than bench top, free air experiments which can include much more natural complexities. The experimental scale also allows to monitor the process with remote sensing equipment and investigate using standard geologic techniques. For this purpose the GFS was used to conduct crater-forming experiments using chemical explosives (Pentex booster charges, 1-3 106 J) as driving energy source. The experiments were designed to study explosions which occur during the formation of maar volcanoes, and during the interaction of groundwater and magma, where multiple explosions take place at similar locations, close and far from the surface. However, they have been useful to highlight more universal behavior in subsurface explosions. The blasts created craters up to 2.7 m diameter and were monitored by a diverse set of sensors including high-speed, high-definition, and thermal cameras, microphones, infrasound microphones, electric field sensors, seismometers, geophones, high-frequency seismic sensor, and a pitot tube. The availability of data observing the dynamics of crater formation, and the static final products is a unique opportunity to better understand this process. Integrating the diverse sensor data and geologic records helps to interpret the sensor signals and map their individual parts to process phases, states and end members of a blast. The experiments serve as an excellent base for interdisciplinary meeting point, equipment testing, and student training.
Session:From the Earthquake Source to Damage of Buildings: Bridging the Gap between Seismology and Earthquake Engineering
Presenter   Moss, Robb
Schedule   Fri / Poster
Room   Cook/Arteaga
Quantifying the Spatial Variability of Ground Motion Residuals Using Site-to-Site Closeness Index as a Distance Metric
HOLLENBACK, J. C., UCB, Berkeley, CA, USA, jhollenback@gmail.com; MOSS, R. E. S., Cal Poly, San Luis Obispo, CA, USA, rmoss@calpoly.edu
There have been many investigations into the spatial variability of residuals from empirically based ground motion prediction equations (e.g., Wang and Takada 2005, Goda and Atkinson 2010, Loth and Baker 2013). In this study, and in similar studies, spatial variability refers to the similarity (or dissimilarity) of ground motion residuals at different locations for a given event. Most studies quantify spatial variability in terms of semi-variance, autocovariance or autocorrelation. Regardless of the statistic used, spatial variability is almost always described as a function of linear distance between locations of observations as a metric of separation. In this study we use site-to-site closeness index as the metric of spatial variability of ground motion residuals. The site-to-site closeness index is a modified form of the closeness index (Lin et al., 2012)and is a measure of the similarity between the wave propagation paths of ground motion observed at two different sites for a given event. Our investigation is performed using three different regional data sets, one from Japan (Rodriguez-Marek et al. 2011), one from Taiwan (Lin et al., 2011) and one from California (Chiou et al. 2010). For each data set we calculated the semi-variance for two different kinds of ground motion residuals: intra-event residuals and single-station intra-event residuals. These calculations were performed for peak ground acceleration and pseudo spectral acceleration at 0.3 and 1.0 seconds. Results showa reduction in scatter when calculating the semi-variance as a function of site-to-site closeness index when compared to semi-variance calculated as a function of linear distance. This is observed for all ground motion intensity measures for all three regional data sets.
Session:Site Response: From Site-Specific Analyses to Predictive Models Around the Globe
Presenter   Kaiser, Anna
Schedule   Wed 11:30 AM / Oral
Room   Room 4
Seismic Site Amplification in the Port Hills during the Canterbury Earthquake Sequence: Case Study of Critical Slopes
KAISER, A. E., GNS Science, Lower Hutt, New Zealand, a.kaiser@gns.cri.nz; HOLDEN, C., GNS Science, Lower Hutt, New Zealand; MASSEY, C., GNS Science, Lower Hutt, New Zealand
Significant ground failure and building damage occurred in the hillside suburbs of Christchurch during the Mw 6.2 February and Mw 6.0 June 2011 earthquakes. Damage patterns indicate that local amplification of ground motions likely contributed to the most severe effects. The Canterbury earthquake sequence provides an internationally significant case study to understand the influence of amplification effects in areas of steep topography. We present analysis of ground motion variability in the Port Hills focussing on a case study across a critical slope and deep-seated landslide above the suburb of Sumner. Our observations from small-scale temporary seismometer arrays show strong differences in ground motion polarization and amplification up to 3 times PGA over small distances (tens to hundreds of metres). Synthetic ground motion simulation of 2D amplification effects suggests site effects result from both topographic shape and material properties. Our results have engineering implications, given that all sites are classed as rock under current New Zealand design standards. At present in New Zealand such amplification effects are not routinely taken into account when designing earthworks or structures on slopes or in landslide risk assessments.
Session:Deciphering the Earthquake and Tsunami History of Subduction Zones
Presenter   Williams, Chesley
Schedule   Wed 3:15 PM / Oral
Room   Room 2
Understanding Mega-Earthquake Related Tsunami on Subduction Zones without Large Historical Events
WILLIAMS, C. R., Risk Management Solutions, Inc, Newark, CA, USA, Chesley.Williams@rms.com; MOHAMMED, F., Risk Management Solutions, Inc, Newark, CA, USA, Fahad.Mohammed@rms.com; LEE, R., Risk Management Solutions, Inc, Newark, CA, USA, Renee.Lee@rms.com
Following on the megathrust earthquake related tsunami in Chile in 2010 and in Japan in 2011, the insurance industry has become very aware of tsunami as a catastrophic peril. Though little exposure is currently covered for tsunami risk, insurers and their clients are interested in understanding the risk and having a tool to price this peril. To begin the process of understanding tsunami as a risk peril, Risk Management Solutions Inc. (RMS) is developing a global suite of tsunami inundation footprints. Many of these events represent well studied historical events including the events in 2010 and 2011. The dataset will also include a series of M9 events on subduction zones that have not historically ruptured in M9 events. These scenarios are to account for the potential risk posed by subduction zones that may be underestimated. This concern was raised following the 2011 Tohoku Japan M9 that far exceeded the expected maximum magnitude on the Japan Trench of M8.4 that was constrained based on the historical record. Subduction zones without M9 historical events being modeled include the Makran Trench, the Lesser Antilles Arc and the Hiku