HR: 09:40h
AN: NG21A-07    [Abstracts]
TI: 100 Years after the San Francisco Earthquake of 1906: Earthquake Forecasting and Forecast Verification - Status, Prospects and Promise
AU: * Rundle, J B
EM: jbrundle@ucdavis.edu
AF: University of California, Center for Computational Science, Davis, CA 95616` United States
AU: Turcotte, D L
EM: turcotte@geology.ucdavis.edu
AF: University of California, Dept of Geology, Davis, CA 95616 United States
AU: Holliday, J R
EM: holliday@cse.ucdavis.edu
AF: University of California, Center for Computational Science, Davis, CA 95616` United States
AU: Rundle, P B
EM: jbrundle@cse.ucdavis.edu
AF: University of California, Center for Computational Science, Davis, CA 95616` United States
AU: Tiampo, K F
EM: kftiampo@uwa.edu.ca
AF: University of Western Ontario, Dept of Earth Science, London, ON N6A5B7 United States
AU: Chen, C
EM: chenCC@cse.ucdavis.edu
AF: National Central University, Dept of Earth Science, Taipei, 320 Taiwan
AU: Nanjo`, K
EM: nanjo@cse.ucdavis.edu
AF: Institute for Theoretical Mathematics, Minato-ku, Tokyo, 106-8569 Japan
AU: Donnellan, A
EM: andrea.donnellan@jpl.nasa.gov
AF: Jet Propulsion Laboratory, Earth and Space Science Division, Pasadena, CA 91125 United States
AU: Klein, W
EM: klein@buphy.bu.edu
AF: Boston University, Dept of Physics, Boston, MA 02215 United States
AB: The Elastic Rebound Hypothesis was proposed by H.F. Reid in the Carnegie Commission summary report published in 1910. As we approach the 100th year anniversary of the 1906 disaster, progress not possible in that era is made possible by the use of advanced computer models and simulations, combined with new data sets and data mining techniques, together with ideas about complex nonlinear systems. Modern computational technology allows us to construct models such as Virtual California that include many of the physical processes known to be important in earthquake dynamics. These include elastic interactions among the faults in the model, driving at the correct plate tectonic rates, and frictional physics on the faults using the physics obtained from laboratory models with parameters consistent with the occurrence of historic earthquakes. Models such as Virtual California represent "numerical laboratories" in which the event statistics and precursory patterns can be determined directly from simulations, rather than by assumption. Using such models, we have found that failure on groups of fault segments is usually best described by Weibull statistics. Simulations also allow us to develop new types of forecast methods such as the Pattern Informatics (PI) method, which develops forecast maps based upon locating areas of highest change in seismic activity of small earthquakes. A contrasting approach, which serves as a null hypothesis, is to compute maps based upon locating the highest rates of small earthquakes (Relative Intensity, or RI). Development, testing, and evaluation of these forecast methods is possible only if reliable testing mechanisms are available. The Relative Operating Characteristic (ROC) method of forecast verification has been developed in the study of binary forecasts of tornadoes and other weather-related events. We have adapted this test to earthquake forecasting and find that it allows the best objective assessment of which methods show optimal performance. Using the ROC test, it is possible to develop forecasts of future earthquake occurrence that improve upon the original PI method as published in PNAS on February 19, 2002.
UR: http://quakesim.jpl.nasa.gov/
DE: 0550 Model verification and validation
DE: 3238 Prediction (3245, 4263)
DE: 4460 Pattern formation
SC: Nonlinear Geophysics [NG]
MN: Fall Meeting 2005