S11A-1714
Predicted Liquefaction in the Greater Oakland and Northern Santa Clara Valley Areas for a Repeat of the 1868 Hayward Earthquake
Probabilities of surface manifestations of liquefaction due to a repeat of the 1868 (M6.7-7.0) earthquake on the southern segment of the Hayward Fault were calculated for two areas along the margin of San Francisco Bay, California: greater Oakland and the northern Santa Clara Valley. Liquefaction is predicted to be more common in the greater Oakland area than in the northern Santa Clara Valley owing to the presence of 57 km2 of susceptible sandy artificial fill. Most of the fills were placed into San Francisco Bay during the first half of the 20th century to build military bases, port facilities, and shoreline communities like Alameda and Bay Farm Island. Probabilities of liquefaction in the area underlain by this sandy artificial fill range from 0.2 to ~0.5 for a M7.0 earthquake, and decrease to 0.1 to ~0.4 for a M6.7 earthquake. In the greater Oakland area, liquefaction probabilities generally are less than 0.05 for Holocene alluvial fan deposits, which underlie most of the remaining flat-lying urban area. In the northern Santa Clara Valley for a M7.0 earthquake on the Hayward Fault and an assumed water-table depth of 1.5 m (the historically shallowest water level), liquefaction probabilities range from 0.1 to 0.2 along Coyote and Guadalupe Creeks, but are less than 0.05 elsewhere. For a M6.7 earthquake, probabilities are greater than 0.1 along Coyote Creek but decrease along Guadalupe Creek to less than 0.1. Areas with high probabilities in the Santa Clara Valley are underlain by latest Holocene alluvial fan levee deposits where liquefaction and lateral spreading occurred during large earthquakes in 1868 and 1906. The liquefaction scenario maps were created with ArcGIS ModelBuilder. Peak ground accelerations first were computed with the new Boore and Atkinson NGA attenuation relation (2008, Earthquake Spectra, 24:1, p. 99-138), using VS30 to account for local site response. Spatial liquefaction probabilities were then estimated using the predicted ground motions for the scenario earthquake magnitude and liquefaction probability curves that were developed with the methodology for probabilistic liquefaction hazard mapping proposed by Holzer (2008, ASCE Geotechnical Special Publication 181).
S11A-1715
3D Spontaneous Rupture Models of Large Earthquakes on the Hayward Fault, California
We are constructing 3D spontaneous rupture computer simulations of large earthquakes on the Hayward and central Calaveras faults. The Hayward fault has a geologic history of producing many large earthquakes (Lienkaemper and Williams, 2007), with its most recent large event a M6.8 earthquake in 1868. Future large earthquakes on the Hayward fault are not only possible, but probable (WGCEP, 2008). Our numerical simulation efforts use information about the complex 3D fault geometry of the Hayward and Calaveras faults and information about the geology and physical properties of the rocks that surround the Hayward and Calaveras faults (Graymer et al., 2005). Initial stresses on the fault surface are inferred from geodetic observations (Schmidt et al., 2005), seismological studies (Hardebeck and Aron, 2008), and from rate-and- state simulations of the interseismic interval (Stuart et al., 2008). In addition, friction properties on the fault surface are inferred from laboratory measurements of adjacent rock types (Morrow et al., 2008). We incorporate these details into forward 3D computer simulations of dynamic rupture propagation, using the FaultMod finite-element code (Barall, 2008). The 3D fault geometry is constructed using a mesh-morphing technique, which starts with a vertical planar fault and then distorts the entire mesh to produce the desired fault geometry. We also employ a grid-doubling technique to create a variable-resolution mesh, with the smallest elements located in a thin layer surrounding the fault surface, which provides the higher resolution needed to model the frictional behavior of the fault. Our goals are to constrain estimates of the lateral and depth extent of future large Hayward earthquakes, and to explore how the behavior of large earthquakes may be affected by interseismic stress accumulation and aseismic slip.
S11A-1716
Earthquake Stress Drops and Inferred Fault Strength on the Hayward Fault
The Hayward Fault and other faults of the East Bay provide an opportunity to study variations in earthquake stress drop with depth, faulting regime, creeping versus locked behavior, and the strength of the wall rocks on either side of the fault. We use the displacement spectra from borehole seismic recordings of 529 M1.0- 4.2 earthquakes in the East Bay to estimate stress drop using an empirical Green's function method (Shearer et al., 2006). The median stress drop is 8.7 MPa, and 50% of stress drops are between 3.2 MPa and 25 MPa. Several lines of evidence indicate that stress drop is controlled by the applied shear stress, even though the median stress drop values are significantly less than the theoretical shear stress assuming strong faults (Byerlee's law) and hydrostatic pore pressure. There is a trend of increasing stress drop with depth, with median stress drop of about 5 MPa for 1-7 km depth, about 10 MPa for 7-13 km depth, and about 50 MPa deeper than 13 km. Higher stress drops are observed for a deep cluster of thrust-faulting earthquakes near Livermore than for a deep cluster of strike-slip events on the Calaveras Fault. The changes in stress drops with depth and faulting regime imply that stress drop is related to the applied shear stress. We compare the spatial distribution of stress drops on the Hayward Fault to models of creeping versus locked behavior of the fault, and find that high stress drops are concentrated around the major locked patch near Oakland. This also suggests a connection between stress drop and applied shear stress, because the locked patch might be expected to experience higher shear stress as a result of either the difference in cumulative slip or the presence of higher-strength material. Comparison of stress drops with the wall-rock geology at depth does not show a correlation between stress drop and rock strength, suggesting that the fault strength is not directly related to the strength of the wall rock.
S11A-1717
Subsurface Creep and Geometry of the Hayward-Calaveras Stepover
The San Francisco Bay Area has not experienced a major earthquake beneath and urban center since 1906. The Hayward fault is the most populated fault in the area, and 140 years after its last rupture, also has the greatest risk associated with it. A contiguous subsurface stepover connecting the Hayward and the Calaveras faults appears to directly transfer slip between the two faults, and may affect earthquake rupture scenarios on both faults. Although the Hayward fault is partially locked for much of its trace, the southernmost fault creeps up to 9 mm/yr, equal to its long-term slip rate. At the proposed juncture region, the southern Calaveras fault also exhibits a surface creep rate close to total slip rate of 15mm/yr. Although, creep on the northern Calaveras is poorly constrained, its geologic slip rate is about 6 mm/yr, suggesting direct slip transfer from the southern Calaveras to the Hayward fault. Relocated seismicity outlines an eastward-dipping Hayward fault which appears to dip into and merge with the central Calaveras fault at depth. Additionally, characteristic repeating earthquakes through the stepover indicate that subsurface creep occurs between the two faults. Using both continuous GPS and updated GPS campaign data, we re-evaluate slip on Bay Area faults, and map slip through the Hayward-Calaveras stepover using a contiguous fault model. We use background seismicity and repeating earthquakes to constrain the geometry of the stepover, and invert GPS and InSAR data for slip. We interpret our map of creep transfer between the Hayward and Calaveras faults for its implications for seismic rupture through the stepover and seismic hazard in the Bay Area.
S11A-1718
Plio-Pleistocene Evolution of Concealed Basins Separated by a Bedrock Ridge West of the Rodgers Creek and Healdsburg Faults, Northern California
The Santa Rosa Plain west of the Rodgers Creek-Healdsburg Fault Zone conceals the Trenton Ridge, a long-lived NW-oriented structurally controlled drainage divide separating Windsor basin to the north, from Cotati basin to the south. Thick fills and geometries of these basins combine to amplify seismic shaking and influence ground water movement. The subsurface of Santa Rosa Plain was initially a single basin filled largely by deposition from a W-NW flowing, Miocene to late Pliocene fluvial to marine depositional system and interbedded volcanic rocks. Deposition in the initial basin was controlled by transtension from about 7 to 5 Ma, followed by transpression that formed the Trenton Ridge and separated the Windsor and Cotati basins. Gravity inversions and seismic-reflection data show the basins are more than 2 km deep and that basinward tilting of strata away from Trenton Ridge diminishes higher in the section. The modern Santa Rosa Plain, having no surface expression of the divide separating Windsor and Cotati basins, has evolved since 5 Ma from separate W-flowing interfingering fluvial systems, with transpressional uplift and erosion of the Trenton Ridge and dextral slip on the Rodgers Creek - Healdsburg Fault Zone. We use data from recently drilled water wells integrated with surface and subsurface geology, geophysical and geochemical data to constrain this evolution, focusing on non-marine deposits of the 3 to 1 Ma Glen Ellen Formation, which contain the most important and distinctive gravel aquifers of the region. Paleoflow data and the distribution of chemically fingerprinted obsidian clast suites show that Glen Ellen gravels in the Windsor basin were dominantly deposited from streams flowing westward across the Rodgers Creek - Maacama Fault system from NW Napa - Calistoga and Franz Valleys. Gravels in the Cotati basin are sourced from this area plus NW Sonoma Valley and Annadel, east of Santa Rosa. This suggests that drainage across the Rodgers Creek-Healdsburg Fault Zone from NW Sonoma Valley and Annadel into Windsor basin was blocked by the Trenton Ridge 3 to 1 Ma. Beneath Pleistocene fans of the Cotati basin, the Glen Ellen Formation is thin and irregularly distributed; in Windsor basin the formation is at least 165 m thick. From this relation we infer that uplift and erosion on the south side of Trenton Ridge continued until 1 Ma, after which younger alluvial deposits overlapped the Glen Ellen, forming the modern Santa Rosa Plain. This is consistent with thicker and more extensive gravel aquifers in Windsor basin, and suggests that parameters for modeling of seismic shaking in the two basins may differ significantly.
S11A-1719
Seismic Reflection Profiles Image the Rodgers Creek Fault and Cotati Basin Beneath Urban Santa Rosa, California
The USGS in collaboration with the Network for Earthquake Engineering Simulation (NEES) group at the University of Texas, Austin, the Sonoma County Water Agency, the city of Santa Rosa, and with support from NSF, collected 13-km of high-resolution seismic-reflection data in two profiles on the Santa Rosa Plain. The purpose of this survey was to image basin structure and stratigraphy in this seismically-active area and to provide constraints for earthquake hazard assessment. We acquired the data using a 9,990 kg minivib I truck in P-wave mode, which swept from 15 to 120 Hz, along city streets and creek-side roads. The common- midpoint spacing of these data is 2.5 m while nominal fold is 36 traces. The Rodgers Creek fault, a northward extension of the Hayward fault which passes through the city of Santa Rosa, has not been imaged previously by seismic reflection data. The east-west trending Santa Rosa Creek profile images several faults including the steeply dipping Rodgers Creek fault as it passes near Doyle Elementary School. In this vicinity the fault zone appears to consist of at least two strands with a set of arched reflectors between them. West of the Rodgers Creek fault, and in general agreement with preexisting gravity data and geologic mapping, we interpret a sedimentary basin more than 1 km deep that underlies downtown Santa Rosa, which was heavily damaged in the 1906 earthquake. This basin shallows to the west as the profile crosses the southeastern side of Trenton Ridge, a concealed basement high. Reflectors within the basin show a thickening sequence of layered strata and apparent dips of about 10 degrees east in the 400 to 800 m depth range that decrease to about 1 degree at 50 m depth. These new data will help to constrain existing seismic velocity models for this area which currently show only flat-lying basin fill.
S11A-1720
Assessment of Creep on the Bartlett Springs Fault using GPS and Alinement Array Measurements
The San Andreas Fault System (SAFS) north of Point Arena consists of three main strands in a 100 km wide zone: the San Andreas, Ma'acama, and Bartlett Springs faults. Geodetic coverage near the Bartlett Springs fault (BSF) has been sparse, but available data suggest the fault may slip at ~8 mm/yr at all depths (Freymueller et al., 1999). The BSF may be the northern extension of the Green Valley fault which is estimated to creep at 4.4 ± 0.1 mm/yr (Galehouse et al., 2003). Creeping behavior, if it extends to seismogenic depths, can reduce a fault's potential for damaging earthquakes. The rate and depth extent of creep may be estimated using geodetic data given adequate station coverage. To augment the 10 continuous Global Positioning System (GPS) sites within 50 km of Lake Pillsbury, which is located in the BSF zone, we established a dense network of 39 campaign GPS sites and an alinement array spanning the fault at the north end of the lake to measure near-surface creep. We collected data during three or more field surveys at 25 GPS sites; the rest have been surveyed twice. The BSF-parallel component of GPS station velocities as a function of position normal to the BSF shows a step of ~3 mm/yr at the fault, suggesting near-surface creep. Alinement array measurements have recorded 7.3 mm of episodic right lateral motion between November 2005 and August 2008, giving an average rate of 2.7 mm/yr. En eschelon cracks in pavement of a road crossing the BSF are another possible indicator of ongoing creep. Using BSF-parallel velocities for the 25 GPS sites with 3 or more surveys plus continuous sites spanning the SAFS at this latitude in a nonlinear optimization produced a preliminary creep rate estimate of 2 mm/yr (0.5 - 6.5 mm/yr at 95% confidence) from the surface to a depth of ~13 km. Further analysis to better constrain the depth-extent of creep, explore spatial variations, and interpret the data in the context of northern SAFS tectonics is underway and in the future will benefit from improved velocity estimates made possible by additional surveys.
S11A-1721
Earthquake Record of the Peninsula Segment of the San Andreas fault, Portola Valley, California
Previous paleoseismic studies on the Peninsula segment provide evidence of poorly constrained large magnitude earthquakes occurring in the late Holocene (Wright et al., 1999; Hall et al., 2001), and possibly in 1838 (Toppozada and Borchardt, 1998); whereas paleoseismic investigations on the Santa Cruz Mountains and North Coast segments provide moderately well constrained event chronology information over the last approximately 2,000 years (Fumal et al., 2004; Niemi et al., 2004; Kelson et al., 2006). The WGCEP (2003) report provides an estimate of mean recurrence values of 378 years for 1906-type events, and 230 years for ruptures occurring only on the Peninsula segment of the northern San Andreas fault. More recent paleoseismic investigations along the Santa Cruz Mountains segment (Fumal et al., 2004) and North Coast segment (Niemi et al., 2004; Zhang et al., 2003) suggest the occurrence of surface-fault rupture at these sites as frequently as every 105 to 250 years. It is unknown if these short recurrence intervals for the two segments represent local, smaller events, or if they could have been associated with through-going events that included the Peninsula segment located between the Santa Cruz Mountains and North Coast segments of the San Andreas fault. Paleoseismic trenching within the Portola Valley Town Center site, located on the Peninsula segment, provides preliminary late Holocene event chronology data for the Peninsula segment of the San Andreas fault. At the Town of Portola Valley site we interpret three, and possibly four, earthquake events within approximately the past 1,000 years. Based on stratigraphic and structural relations, available radiocarbon dating, and the presence of historical artifacts, it is permissible to interpret three and possibly four earthquakes that from oldest to youngest include: Event 1 (A.D. 1030 to 1490); Event 2 (A.D. 1260 to 1490); and Event 3 (1906 A.D.). It is permissible to interpret Event 2 for the site as two separate events: Event 2A (A.D. 1260 to 1490) and Event 2B (A.D. 1410 to 1640). The youngest stratigraphic units do not provide the resolution necessary to assess whether or not the postulated, historical 1838 event ruptured the Woodside trace at this site. Select, additional charcoal samples have been submitted for dating to further constrain our event chronology.
S11A-1722
New Criterion and Tool for Caltrans Seismic Hazard Characterization
Caltrans recently adopted new procedures for the development of response spectra for structure design. These procedures incorporate both deterministic and probabilistic criteria. The Next Generation Attenuation (NGA) models (2008) are used for deterministic assessment (using a revised late-Quaternary age fault database), and the USGS 2008 5% in 50-year hazard maps are used for probabilistic assessment. A minimum deterministic spectrum based on a M6.5 earthquake at 12 km is also included. These spectra are enveloped and the largest values used. A new publicly available web-based design tool for calculating the design spectrum will be used for calculations. The tool is built on a Windows-Apache-MySQL-PHP (WAMP) platform and integrates GoogleMaps for increased flexibility in the tool's use. Links to Caltrans data such as pre-construction logs of test borings assist in the estimation of Vs30 values used in the new procedures. Basin effects based on new models developed for the CFM, for the San Francisco Bay area by the USGS, and by Thurber (2008) are also incorporated. It is anticipated that additional layers such as CGS Seismic Hazard Zone maps will be added in the future. Application of the new criterion will result in expected higher levels of ground motion at many bridges west of the Coast Ranges. In eastern California, use of the NGA relationships for strike-slip faulting (the dominant sense of motion in California) will often result in slightly lower expected values for bridges. The expected result is a more realistic prediction of ground motions at bridges, in keeping with those motions developed for other large-scale and important structures. The tool is based on a simplified fault map of California, so it will not be used for more detailed evaluations such as surface rupture determination. Announcements regarding tool availability (expected to be in early 2009) are at http://www.dot.ca.gov/research/index.htm
S11A-1723
Development of Rapid Earthquake Shaking and Loss Assessment Methodologies
Main objective of this study, conducted under the JRA-3 component of the EU Project entitled "Network of research Infrastructures for European Seismology, NERIES", is to develop a methodology for real time estimation of losses after a major earthquake in the Euro-Mediterranean region. The earthquake shaking and loss information will be disseminated in a timely manner to related agencies for the planning and coordination of the post-earthquake emergency response. This multi-level methodology being developed together with researchers from Imperial College, NORSAR and ETH-Zurich is capable of incorporating regional variability and sources of uncertainty stemming from ground motion predictions, fault finiteness, site modifications, inventory of physical and social elements subjected to earthquake hazard and the associated vulnerability relationships The development encompasses the following general steps: 1. Finding of the most likely location of the source of the earthquake using regional seismotectonic data base, basic source parameters and if and when possible, by the estimation of fault rupture parameters from rapid inversion of data from on-line regional broadband stations. 2. Estimation of the spatial distribution of selected ground motion parameters at engineering bedrock through region specific ground motion attenuation relationships and/or actual physical simulation of ground motion. 3. Estimation of the spatial distribution of site-specific ground selected motion parameters using regional geology (or urban geotechnical information) data-base using appropriate amplification models. 4. Incorporation of strong ground motion and other empirical macroseismic data for the improvement and enchantment of the ground motion distribution (Shake Map) 5. Estimation of the losses and uncertainties at various orders of sophistication (Loss Map) A software called "ELER" based on this methodology is currently under development. Within the scope of this paper, results obtained from a pilot application of this methodology and the ELER software to the 1999 Kocaeli earthquake interms of ground shaking and losses are presented and comparisons with the observed losses are made.
S11A-1724
Real-Time and Off-Line Performance of the Virtual Seismologist Earthquake Early Warning Algorithm in California and Switzerland
The Virtual Seismologist (VS) method is a regional network-based approach to earthquake early warning that estimates earthquake magnitude and location based on the available envelopes of ground motion amplitudes from the seismic network monitoring a given region, predefined prior information, and appropriate attenuation relationships. Bayes' theorem allows for the introduction of prior information (possibilities include network topology or station health status, regional hazard maps, earthquake forecasts, the Gutenberg- Richter magnitude-frequency relationship) into the source estimation process. Peak ground motion amplitudes (PGA and PGV) are then predicted throughout the region of interest using the estimated magnitude and location and the appropriate attenuation relationships. Implementation of the VS algorithm in California and Switzerland is funded by the Seismic Early Warning for Europe (SAFER) project. The VS algorithm is one of three early warning algorithms whose real-time performance on California datasets is being evaluated as part of the California Integrated Seismic Network (CISN) early warning effort funded by the United States Geological Survey (USGS). Real-time operation of the VS codes at the Southern California Seismic Network (SCSN) began in July 2008, and will be extended to Northern California in the following months. In Switzerland, the VS codes have been run on offline waveform data from over 125 earthquakes recorded by the Swiss Digital Seismic Network (SDSN) and the Swiss Strong Motion Network (SSMN). We discuss the performance of the VS codes on these datasets in terms of available warning time and accuracy of magnitude and location estimates.
S11A-1725
Realtime Earthquake Detection and Hazard Assessment by ElarmS Across California
ElarmS is a network-based earthquake early warning methodology. As part of the ongoing project by the
California Integrated Seismic Network (CISN) to test early warning methodologies on the realtime geophysical
networks in the state, ElarmS is now running in realtime across California. At the time of writing 180 seismic
station are streaming their data into ElarmS. This includes all early-warning capable stations in northern
California and 15 stations from southern California. By the time of the meeting we expect to have included all
stations in southern California for a total of approximately 300 stations state wide. While the ElarmS
methodology and code has been tested using data from past earthquakes around the world, running the
algorithms in realtime has presented significant challenges both in the scientific/methodological and
engineering/technical domains. The primary methodological challenge encountered is triggering and
associating to identify events as rapidly as possible, while remaining robust. The large number of stations
continuously and inevitably generates spurious triggers and the variety and multiple generations of
instrument types make tuning of individual sites almost impossible. Instead we develop a space-time
association criteria that depends upon dense station coverage. The primary engineering challenge comes
from integrating the waveform data streams from the various contributing networks without using up all
available warning time. At the meeting we will present an overview of ElarmS performance during the last
year and highlight the most important methodological developments.
http://elarms.org/
S11A-1726
Analyzing the Capabilities of the ElarmS Methodology Using a Japanese Dataset
Earthquake early warning systems are algorithms that detect the initial P-waves from an earthquake, rapidly estimate the magnitude of the event, and predict subsequent ground shaking in the surrounding regions. Earthquake Alarm Systems, or ElarmS, is one early warning algorithm that uses a network of seismic stations to hone in on the size and location of the earthquake. Here we analyze the accuracy of ElarmS' magnitude, location, and ground motion estimates using earthquakes recorded by Japan's K-Net strong-motion seismic network. We consider the effects of station distribution, number of reporting stations, and number of seconds of P-wave arrival, and examine ElarmS' ability to process large (M6+) events. We verify the correlation between peak frequency and displacement in the P-wave and the final magnitude of the event. Our results show that utilizing data from several stations improves the overall magnitude estimate. We also find that using four seconds of P-wave data optimizes the accuracy and speed of the estimate.
S11A-1727
Real-time Testing of On-site Earthquake Early Warning within the California Integrated Seismic Network (CISN) Using Statewide Distributed and On-site Processing
Currently, the real-time performance of three algorithms for earthquake early warning is being tested within the California Integrated Seismic Network (CISN). We report on the implementation and performance of the τc-Pd on-site warning algorithm in California and describe recent improvements of the software. These include: (1) the development of a new τc-Pd based trigger criterion to reduce the number of false triggers and the scattering in magnitude estimates for small and moderate earthquakes; (2) the integration of additional broadband stations, including the ANZA network and stations with older dataloggers that provide waveform data with highest sampling rate of 80 sps; and (3) the implementation of leap-second capabilities in the real-time software used within CISN. At present, we are working on the implementation of the remote processing sites for the BK and NP, NC networks operated by UC Berkeley and USGS Menlo Park. These processing sites will analyze available local waveform data and provide τc-Pd values as well as Mw and PGV estimates. The new processing sites will provide more data for algorithm testing and improved data analysis for earthquakes located in northern California. We are also implementing the τc-Pd algorithm software on SLATE Field Processors to eliminate telemetry delays associated with waveform data. On-site the SLATE receives data from a Q330 datalogger and provides τc-Pd estimates from the first 3 seconds of P-waveforms. These τc-Pd values along with station-specific Mw and PGV values are transmitted to the central site as a short notification message. In the future, such station processors can also transmit warnings to local users directly. The τc-Pd algorithm software performed well during the recent July 29 2008 Mw5.4 Chino Hills earthquake. A total of 36 stations provided real-time estimates of τc-Pd values and derived Mw and PGV values. The current CISN network configuration would have provided a 6 second warning at Los Angles City Hall, which is located 50 km to the west-southwest of the mainshock epicenter.
S11A-1728
Extending the CISN Earthquake Early Warning (EEW) Web Site into the CISN EEW Testing Center
As a part of the California Integrated Seismic Network (CISN) earthquake early warning (EEW) algorithm
development, funded through USGS NEHRP, we have developed the CISN EEW web site to collect the results
of multiple EEW algorithms and to display these results in a comparative manner (www.scec.org/eew). During
the last year, the CISN EEW algorithm development group defined a set of EEW algorithm evaluation tests
(termed performance summaries). These compare EEW algorithm reports (generated by the real-time or
near real-time EEW algorithms) against seismicity data in the ANSS catalog and observed ground motion
information available through the SCEC Data Center (SCECDC) and the Northern California Earthquake
Data Center (NCEDC).
To automatically generate the performance summaries, a software development group at SCEC has
integrated elements of the SCEC Collaboratory for the Study of Earthquake Predictability (CSEP) Testing
Center into the CISN EEW web site. This has helped establish a CISN EEW Testing Center with capabilities
similar to the CSEP Testing Center. After the integration of the CSEP software, the CISN EEW testing center
now automatically creates EEW performance summaries and posts them on the CISN EEW web site each
day.
By leveraging the capabilities of the CSEP Testing Center, the CISN EEW Testing Center has been able to
implement several of the testing concepts originally developed on CSEP. These concepts include the
following: (a) earthquake, or ground motion, forecasts are reported in standardized data formats, (b)
commonly-agreed upon performance evaluation reports are used for all algorithms, (c) observed data is
retrieved from 'authorized' data sources and the same observed
data is used to evaluate all algorithms, (d) only forecasts and observed data for a specific testing region are
considered, and (e) the testing center saves information indicating how results were produced.
We present an overview of the CISN EEW Testing Center including the scientific design goals for the system,
and a description of the system's current capabilities. We describe the performance
summaries specified by the CISN EEW algorithm development group, and how the current CISN EEW Testing
Center produces those summaries using the automated testing capabilities from the CSEP software
framework.
http://www.scec.org/eew
S11A-1729
The California Post-Earthquake Information Clearinghouse: A Plan to Learn From the Next Large California Earthquake
In the rush to remove debris after a damaging earthquake, perishable data related to a wide range of impacts
on the physical, built and social environments can be lost. The California Post-Earthquake Information
Clearinghouse is intended to prevent this data loss by supporting the earth scientists, engineers, and social
and policy researchers who will conduct fieldwork in the affected areas in the hours and days following the
earthquake to study these effects. First called for by Governor Ronald Reagan following the destructive
M6.5 San Fernando earthquake in 1971, the concept of the Clearinghouse has since been incorporated into
the response plans of the National Earthquake Hazard Reduction Program (USGS Circular 1242). This
presentation is intended to acquaint scientists with the purpose, functions, and services of the
Clearinghouse.
Typically, the Clearinghouse is set up in the vicinity of the earthquake within 24 hours of the mainshock
and is maintained for several days to several weeks. It provides a location where field researchers can
assemble to share and discuss their observations, plan and coordinate subsequent field work, and
communicate significant findings directly to the emergency responders and to the public through press
conferences. As the immediate response effort winds down, the Clearinghouse will ensure that collected
data are archived and made available through "lessons learned" reports and publications that follow
significant earthquakes.
Participants in the quarterly meetings of the Clearinghouse include representatives from state and federal
agencies, universities, NGOs and other private groups. Overall management of the Clearinghouse is
delegated to the agencies represented by the authors above.
http://www.eqclearinghouse.org