Home  /  Workshops  /  SCEC-NASA Workshop: Evaluating Ground-Based and Space-Based Methods of Earthquake Forecasting

SCEC-NASA Workshop: Evaluating Ground-Based and Space-Based Methods of Earthquake Forecasting

Conveners: Tom Jordan (SCEC/USC, chair), Tom Bleier (Quakefinder, co-chair), Andrew Michael (USGS/Menlo Park), and Friedemann Freund (NASA Ames)
Dates: July 25-27, 2011
Location: Davidson Conference Center, University of Southern California, Los Angeles, CA
Participants: 44

Sypnosis: Arrays of ground-based and space-based sensors observe a variety of earthquake-related phenomena, including seismic, geodetic, electromagnetic, and geochemical signals. This two-day workshop will focus on the use of these signals in earthquake forecasting. It will provide a forum for exchanging views among different research communities about how forecasts are specified (e.g., in terms of location, time, and magnitude), how hypotheses regarding precursory behavior can be tested, and how the reliability, skill, and net information gain of forecasting methods can be evaluated. The goal of the workshop is to chart a course for forecast development that begins with exploratory research on earthquake precursors and the casting of testable precursory hypotheses, proceeds through retrospective and prospective testing of forecasting methods, and leads to the incorporation of significant precursory information into operational earthquake forecasting.

Preparing for the Workshop: The workshop aims to provide a collaborative environment for researchers from different fields to explore common earthquake forecasting issues and compare a variety of research programs on forecast development. The objective is not to evaluate specific forecasting methods but to build a consensus about general strategies for forecast evaluation.

Prior to the workshop, each participant involved in developing earthquake forecasts is encouraged to address the following questions:

  1. What physical hypotheses about earthquake predictability have motivated your research?
  2. What evidence can be used to support or reject these hypotheses?
  3. What data are used? What is the spatial and temporal extent of the data? How often are there gaps in the data? What uncertainties exist in the data? Are there authoritative data sources and are they openly available?
  4. How is “noise” (ambient or sensor-generated) treated in the data collection and analysis process?
  5. Have earthquake-forecasting models that incorporate these hypotheses been formulated? Is the model under development or ready for retrospective or prospective testing? Are these models automated such that they could be submitted for independent evaluation? Are there parameters (e.g. time, space, and magnitude windows, thresholds when forecasts are made) that still need to be determined?
  6. Is there corroborating evidence within a forecast (e.g. Is this forecast made based on 1 indicator or multiple indicators)?
  7. Under which circumstances have the forecasting models been tested Retrospectively? Prospectively?
  8. What are the statistical results of formal testing (e.g., false-alarm and failure-to-predict error rates, skill scores, or relative information/probability gain)?
  9. Is currently significant information from this research ready for “operational” earthquake forecasting? What is the scale of the forecast elements (e.g. Time, Location, Magnitude, depth, probability) and what weight is placed on each?
  10. What are next steps for moving towards the use of this information in earthquake forecasting?
  11. What are the next steps for improving our understanding of the physical hypotheses?

Agenda: The agenda will comprise a mix of brief presentations and panel discussions. Each session will have a moderator and reporter, who will summarize the presentations and discussions from their session. Reporter summaries will contribute to the State of Knowledge and Next Steps Session on the final day of workshop.

During the open Sessions III and IV, researchers are invited to present the status of forecast methods from their perspectives. Each presentation should be no more than 15 minutes in duration. We hope these researchers will address the questions posed in their presentations. For those who want to provide more detailed information of their forecast methods, space will be available to display posters.

July 25, 2011

15:00 - 18:30 Non-Local Attendees Check into Hotel
18:30 - 21:00 Welcome Reception

July 26, 2011

07:30 - 08:30 Breakfast / Registration & Check-In
08:30 - 09:00 Workshop Goals and Objectives (T. Jordan)
09:00 - 12:00 Session I: Physical Processes That May Produce Precursor Signals
  Five invited presentations (15 min each) plus panel discussion
Moderator: T. Jordan
Reporter: B. Minster
Presentations: F. Freund, M. Johnston, S. Pulinets, D. Lockner, J. Vidale
12:00 - 13:30 Lunch
13:30 - 15:30 Session II: Methods for Hypothesis Testing and Forecast Evaluation
  Five invited presentations (15 min each) plus panel discussion
Moderator: T. Bleier
Reporter: M. Blanpied
Presentations: D. Agnew, V. Kossobokov, D. Jackson, J. Rundle, T. Jordan
15:30 - 15:45 Break
15:45 - 17:30 Session III: Status of Forecast Methods
  Open session of volunteered presentations (15 min each),
Addressing the "11 questions" for specific forecasting methods
Researchers may bring posters if they desire to present more detailed information of their forecast methods
Moderator: A. Donnellan
Reporter: B. Shaw
Presentations: T. Bleier, L. Jones, V. Korneev, M. Eneva, J. Love
17:30 Adjourn
18:30 - 21:00 Group Dinner

July 27, 2011

07:30 - 08:30 Breakfast
08:30 - 10:00 Session IV: Status of Forecast Methods Continued
  Open session of volunteered presentations (15 min each),
Addressing the "11 questions" for specific forecasting methods
Researchers may bring posters if they desire to present more detailed information of their forecast methods
Moderator: T. Tullis
Reporter: D. McPhee
Presentations: G. Papadopoulos, D. Ouzounov, J. Zechar, M. Kamogawa, S. Pulinets, T. Liu, F. Freund, D. Agnew
10:00 - 10:15 Break
10:15 - 12:00 Session V: Official Use of Forecasting Information
  Five presentations from organizations (10 min each)
Moderator: F. Freund
Reporter: D. Jackson
Presentations: T. Tullis, D. Agnew, M. Blanpied, OJ. Goltz, C. Dobson
12:00 - 13:30 Lunch
13:30 - 15:30 Session VI: State of Knowledge and Next Steps
  Include summaries from 5 reporters (10 min each)
Moderators: T. Jordan, T. Bleier
Report Summaries: B. Minster, M. Blanpied, A. Donnellan/B. Shaw, D. McPhee, D. Jackson
15:30 Adjourn

Posters: J. Zechar, V. Korneev, D. McPhee, C. Dunson


  • Duncan Agnew (UCSD, Professor of Geophysics)
  • Michael Blanpied (USGS/Reston, Associate Coordinator, Earthquake Hazards Program) - Dr. Michael Blanpied joined the USGS earthquake research group in Menlo Park, CA in 1989 after completing his doctorate at Brown University. His research has focused on the physics of earthquakes, including experimental investigations of the physics of fault slip and frictional properties of fault surfaces; applications of laboratory data to earthquake occurrence and the deformation of the continental crust; computer and laboratory modeling of earthquake interactions; and the development and application of probabilistic assessments of earthquake likelihood. He served as co-chair of the Working Group on California Earthquake Probabilities, which developed a new methodology for forecasting the likelihood of damaging earthquakes in the San Francisco Bay Region. In 2003 Dr. Blanpied relocated to USGS headquarters in Reston, Virginia to take his current position in the Earthquake Hazards Program office, where his duties include oversight of the Program’s earthquake hazards assessments and its research on earthquake physics and occurrence. He serves as executive secretary to the National Earthquake Prediction Evaluation Council (NEPEC), an expert group that advises USGS on earthquake predictions and forecasting methods. He also serves as programmatic lead for international research activities, and coordinates the Earthquake Disaster Assistance Team (EDAT) activity in cooperation with the USAID Office of Foreign Disaster Assistance (OFDA).
  • Tom Bleier (QuakeFinder, CEO and CTO) - [Ref 1]
  • Nevin Bryant (NASA/JPL)
  • Robert Dahlgren (CSU/SETI, Research Associate) - [Ref 1]
  • Bruce Davis (DHS/HQ)
  • Craig Dobson (NASA/HQ)
  • Andrea Donnellan (NASA/JPL, Doctoral Level Staff Scientist and Principle Research Scientist)
  • Shakia Dorsey (DHS/HQ)
  • Clark Dunson (QuakeFinder, Lead Analyst)
  • Gary Egbert (Oregon State U.)
  • Mariana Eneva (Imageair, Inc., Professor, Research Scientist, Doctoral Level Staff Scientist) - [Ref 1]
  • Friedemann Freund (NASA/Ames/SJSU, Adjunct Professor, Research Faculty) - I came late to the earthquake field and from the outside. Having a lifelong interest in defects in crystals, primarily in the context of ceramics and catalysis, I studied high purity oxide single crystals. I discovered a type of defect previously unknown in the science community: pairs of oxygen anions, which had changed their valence from the usual 2– valence to the quite unusual 1– state. These pairs of oxygen form peroxy bonds. As long as the peroxy bonds are intact, the defects are electrically inactive and near-impossible to detect. That’s why they hadn’t been discovered earlier. When peroxy bonds break, the electrical conductivity shoots up by orders of magnitude. These electron vacancy defects behave as mobile “positive holes” in the semiconductor sense, and are symbolized by h•. I showed that, during the peroxy break-up, near-perfect insulators turn into mild p-type semiconductors. While most of my early work was done on MgO, I found out that the same defects exist in minerals. Through the 1980s and early 1990s I used essentially every available physical technique to learn more about these positive holes – from single crystal electrical conductivity, precision thermal expansion, dielectric polarization and magnetic susceptibility measurements, infrared and Raman spectroscopy, x-ray photoelectron, electron spin resonance, and muon spin relaxation spectroscopy. In the late 1990s a heated debate was taking place, printed in Nature, pitting seismologists against scientists from outside the seismology field who were trying to understand a wide range of phenomena preceding earthquakes. As I followed this debate, I realized that most of these fiercely disputed and seemingly inexplicable pre-earthquake signals pointed to electrical processes taking place in the Earth’s crust. It quickly became apparent to me that positive hole charge carriers seemed to be involved. Today, 15+ years later, I am convinced that I am still on the right track. Even though the scientific community at large has not recognized the idea of positive hole charge carriers in rocks, there is some strong evidence that most disputed pre-earthquake signals are generated by processes linked to the stress-activation of these elusive electronic charge carriers in the Earth crust. This is the fundamental concept that I would like to bring to our SCEC-NASA Workshop. [Ref 1]
  • Mino Freund (NASA/Ames, Director of Research for: Advanced Materials and Devices, Nano & Biotechnology)
  • James Goltz (CalEMA, Branch Chief: Earthquake, Tsunami and Volcanic Hazards Program)
  • Egill Hauksson (Caltech, Senior Research Associate)
  • Jorge Heraud (INRAS/PUCP, Professor, Director Institute for Radioastronomy) - General research area is RADIO SCIENCE (Research in space physics, astrophysics, magnetospheric, ionospheric and earth sciences in general, using electromagnetic waves and specially radio techniques). At the Pontificia Universidad Catolica del Peru (PUCP) we are building the research infrastructure (a 20m radio-telescope) and involved in a small satellites development program. We are also doing research in both pre-seismic and co-seismic luminescence and operate two Quakefinder magnetometers in high seismicity areas in southern Peru. We are currently in preparation for installing another 6 magnetometers in Peru. Concurrently, we are developing the infrastructure to conduct research in alternate methods as pre-seismic luminescence, HF emissions, induced potential, geodetic GPS displacements, radon gas and others.
  • Sue Hough (USGS/Pasadena, Geophysicist, Doctoral Level Staff Scientist)
  • Kenneth Hudnut (USGS/Pasadena, Geophysicist, Doctoral Level Staff Scientist)
  • Tran Huynh (SCEC/USC, Special Projects and Events Coordinator)
  • Stephen Hwang (Sandia National Laboratories)
  • David Jackson (UCLA, Professor of Geophysics)
  • Malcolm Johnston (USGS/Menlo Park, Research Geophysicist) - The overall focus of my work has been to (1) monitor the mechanics of failure of active faults and volcanoes using a variety of instruments that have included magnetometers, tiltmeters, linear and tensor strainmeters, telluric current monitors, differential lake level, borehole tensor and dilational strain, volcano displacement monitors, etc., (2) develop physical models describing crustal failure from the analysis of data from these instruments and (3) determine implications for the physics of the earthquake source from these data. [Ref 1] [Ref 2]
  • Lucy Jones (USGS/Pasadena)
  • Tom Jordan (SCEC/USC, Director for the Southern California Earthquake Center, Professor of Earth Sciences)
  • Masashi Kamogawa (U. Tokyo, Japan)
  • Valeri Korneev (DOE/LBNL, Geological Staff Scientist, Lawrence Baker National Laboratory) - I believe that earthquake forecasts should be based on deterministic rather than statistical approaches. Understanding of earthquake physics supported by concentrated monitoring should lead to a decision making as a result of earthquake prediction. [Ref 1] [Ref 2] [Ref 3]
  • Vladimir Kossobokov (MITP, Russia, Chief Scientist, Professor and Vice President of IUGG "GeoRisk" Comission) - [Ref 1] [Ref 2]
  • John LaBrecque [canceled] (NASA/HQ)
  • Jann-Yeng Liu (National Central University, Taiwan) - [Ref 1]
  • Masha Liukis (SCEC/USC, Research Programmer)
  • David Lockner (USGS/Menlo Park, Doctoral Level Staff Scientist)
  • Jeffrey Love (USGS/Golden)
  • Darcy McPhee (USGS/Menlo Park, Geophysicist) - The U.S. Geological Survey, Stanford University, and the Berkeley Seismological Laboratory are collaborating on an effort to establish a network of local and remote-reference, high-precision ground-based electromagnetic (EM) monitoring stations in northern California to determine the validity of pre-seismic, co-seismic and post-seismic ULF (ultra-low frequency; 0.01-10 Hz) EM signals (see attached full Project Description). [Ref 1] [Ref 2] [Ref 3] [Ref 4] [Ref 5]
  • John McRaney (SCEC/USC, Associate Director for Administration)
  • Andy Michael [canceled] (USGS/Menlo Park, Geophysicist) - My primary interests in earthquake prediction research are the use of earthquake clustering to estimate time-dependent earthquake probabilities and statistical methods for testing proposed earthquake prediction methods. [Ref 1] [Ref 2]
  • Bernard Minster (UCSD)
  • Stacy Okutani (DHS)
  • Dimitar Ouzounov (Chapman/NASA/GSFC, Associate Professor, Department of Physics, SSAI Principal Investigator at NASA GSFC) - The purpose of our research is to utilize existing global remote-sensing satellite data (thermal infrared observations, Global Positioning System Total Electron Content (GPS/TEC), and space plasma parameters), simultaneously with ground observations to detect and understand atmospherics signals prior to major earthquakes. Our approach is integrated analysis, a Sensor Web approach based on using model estimates with data fusion of satellite and ground data. Our validation methodology is based on historical multi-year (9-20 years) time series data obtained from satellite and ground observations monitoring. We integrate several different physical-environmental parameters, which have been found to be associated with known earthquakes. We use a sensor web of existing satellite sensors (Terra, Aqua, POES, DEMETER and other international sensors) and ground observations GPS/TEC, radon, air temperature, relative humidity, lineaments change, electric and magnetic field measurements. Our rationale for using this complement set of observations are: (I) there are insufficient spatial and temporal coverage of any one of these pre-cursor signals; and (2) Our experience show that is very unlikely that a single existing method of earthquake precursors (seismic, magnetic field, ionospheric field, thermal infrared (TIR), and GPS/TEC) can provide a successful solution for detecting in advance the complex earthquake phenomena on the global scale. However, simultaneous satellite and ground measurements as an integrated web should provide the necessary information by combing the information provided by multiple sensing sources, both on the ground and from space. The advantage of our approach is to enable multiple and already validated physical measurements to be fused into one framework, united by a common physical theory and to provide feedback on data gaps that may then be acquired from other sources. We make use of Lithosphere-Atmosphere-Ionosphere Coupling (LAIC) concept, which systematically integrates various atmospheric phenomenon proceeding or accompanying earthquakes in one common theory. [Ref 1] [Ref 2] [Ref 3] [Ref 4]
  • Gerassimos Papadopoulos (Inst. Geodynamics, Greece, Research Director) - Research of G. A. Papadopoulos related to the forecasting of earthquakes is focusing on the identification of the space-time-size properties of short-term foreshocks. Recently a global database of research papers on foreshocks was organized.
  • Sergey Pulinets (Fiodorov Institute, Russia, Leading Scientist and Head of Laboratory) - Research Interests include: Physical model of Lithosphere-Atmosphere-Ionosphere Coupling, Short-term earthquake prediction, Short-term earthquake precursors, Ionospheric precursors of earthquakes, Space and ground based instrumentation for the earthquake precursors registration, Design of satellites for earthquake precursors registration, Algorithms development for automatic identification of earthquake precursors from experimental data. Probably the most important concept for me was the paper of Scholz et al., 1973 where the problem of earthquake prediction is considered from the physical point of view. One of the precursors considered in the paper was radon with which we were able to find correlation studying the ionospheric variations in Tashkent region before earthquake. The second step in our developments was consideration of atmospheric electric field modification by radon emanation, and the role of atmospheric electric field in coupling with the ionosphere. It was the first stage of creation of the coupling model to make foundation for explanation of ionospheric precursors of earthquakes. More or less at the same time the idea of satellite monitoring for detection of ionospheric precursors of earthquakes was put forward. [Ref 1] [Ref 2]
  • John Rundle (UC Davis, Interdisciplinary Professor of Physics, Civil Engineering and Geology) - Rundle and group are interested in forecast evaluation using mainstream, well accepted methods adopted from other fields, as well as developing new methods of forecasting in earthquakes, weather, and finance. [Ref 1] [Ref 2]
  • Bruce Shaw (Columbia/Lamont-Doherty Earth Observatory, Lamont Associate Research Professor, Columbia Faculty)
  • Terry Tullis (Brown, Emeritus and Research Professor, University Chair, National Earthquake Prediction Evaluation Council) - My Personal focus is on laboratory experimental studies of the frictional properties of rocks, rate and state friction, and on numerical earthquake simulations using this non-linear rheology which shows accelerating creep prior to instability. As the references below show, students and I have also investigated proposed methods of earthquake prediction proposed by others that initially appeared to have some promise. [Ref 1] [Ref 2]
  • John Vidale (U. Washington, Professor) - As Director of the Pacific Northwest Seismic Network, a member of the NEPEC, and a seismologist with an interest in earthquake mechanics and societal hazard reduction, I would like earthquake prediction to improve. Relevant research topics of mine include fault zone properties, seismicity patterns, relation of stressing to earthquake occurrence, and the development of earthquake rupture.
  • Jeremy Zechar (ETH Zurich, Switzerland, Lecturer and Senior Researcher)