Technical Updates: 2013, Second Quarter

Data Sets (Task A)

  • The data reprocessing protocols are been finalized and a summary document was produced. After numerous tests, checks and improvements to the codes, the data re-processing started in mid-June. The process involves numerous participants including the Database Working Group members. Dr. Kishida, a visiting post-doctoral researcher at PEER is leading the software development and re-processing. He is working in close collaboration with Drs. Silva and Darragh (Pacific Engineering Analysis) and Dr. Goulet (PEER), TI Lead. The group benefited from multiple interactions with Dr. Cramer (Memphis University) and Dr. Boore (USGS).

Reference rock and site amplification models (Task B)

  • All the tasks below are being led by the Geotechnical Working Group (GWG). The group continues to be very active with their tasks and have monthly meetings attended by the TI Lead Dr. Goulet.
    • A paper on reference rock condition was submitted for journal review.
    • The group developed a Vs30 proxy method based on topographic slope conditioned by regional geology (glaciated and non-glaciated regions) in CENA.
    • The group’s report on reference rock remains under review by the TI team.
    • The group completed the work on Vs30 proxy at recording stations using P-wave seismograms.
    • The group continued their development of protocols of the large scale site response simulations that will be conducted using rock input ground motions and typical soil profiles. The GWG is finalizing the development of idealized soil profiles for the simulations of site response. This includes velocity profiles and dynamic soil properties. The output of these simulations will be used for development of nonlinear site amplification functions or models.
    • The group continues its work on establishing thresholds for when a nonlinear site response analysis is needed in addition to equivalent linear site response analysis. This will be used in defining the response analyses required for the large scale simulations described next.
    • The group continued their development of protocols of the large scale site response simulations that will be conducted using input ground motions and typical soil profiles. The group is finalizing the development of idealized soil profiles for the simulations of site response. The group has collected about 900 Vs profiles from various NPP license applications and the open literature for the CENA. The collected Vs profiles have been binned to 9 categories according to the geologic classes that were previously developed. The 9 categories are Old glacial, Old non-glacial, Young glacial marine, Young glacial non-marine, Young non-glacial marine, Young non-glacial non-marine, Residual from metamorphic rock, Residual from sedimentary rock, and Residual from soil. The group is accounting for variations in sediment depth in the profiles.

Regionalization & Source/Path Studies (Tasks C and D)

  • A lot of parallel tasks are progressing on this topic, which are truly critical for CENA and NGA-East. The updates below are grouped by task. Each research or team is working on a specific question in the broad scope of better understanding the regionalization of source and path issues in CENA.

Regionalization of Crustal Properties and Quality Factor
Dr. Mooney and his team (USGS)

    • The deep seismic velocity structure of the crust within Central and Eastern North America has been thoroughly investigated by this sub-Task during the past year. This investigation has centered around a detailed compilation of published interpretations numbering in excess of 2,000 measurements. A comprehensive database of the seismic crustal structure was delivered to the NRC in January, 2013. However, the seismic P-wave and S-wave velocity structure are not the only relevant physical parameters. It is also necessary to evaluate seismic attenuation within the crust, which is a function of the geometrical spreading factor and intrinsic attenuation.
    • Figure 1 shows the main geologic divisions of the North American crust. We are concerned with the central platform, the north-central craton, and the eastern regions, which inclose the Appalachians and coastal provinces . Figure 2 shows the geological history of the Appalachians and Coastal Provinces in cross-sections.

2013Q2_Figure_Mooney_RegionsFigure 1. Generalized geologic setting of North America. Key provinces for this study are: (1) the central Platform and cratonic regions; (2) the Gulf Coastal Province; (3) The Appalachians; and (4) the eastern Coastal Plains.

2013Q2_Figure_Mooney_GeolEvol Figure 2. Geological evolution of Eastern North America at the latitude of Washington, D.C. This complex evolution has resulted in regions of high seismic Q (cratonic crust) and low seismic Q (coastal plains).

    • The team investigated and compared published studies on the attenuation of seismic S-waves within the North American main continent. The material’s intrinsic damping of seismic waves throughout the earth’s crust is described by the Quality Factor (Q-Factor) which is defined to be inversely proportional to the attenuation. Furthermore the attenuation is frequency dependent, which is generally described as an exponential relation in the following form:

Q(f) = Q0 * fη

Where: Q(f) is the frequency depended Quality-Factor, Q0 the initial Q factor, f the frequency and η the frequency relation of Q on f.
About 24 published studies covering 50 study areas have been examined. Each study provided a Q(f) relation for a defined frequency band, usually reaching from 1 Hz to 20 Hz. Most of the studies investigated the attenuation of the crust as a whole by looking at the seismic Lg phase, which is a multiple reflected shear wave traveling within the continental crust. Fewer studies however investigated the attenuation of shallow sediments, by looking at the attenuation of P- and S-Waves. Figure 3 shows an example of comparison of Q factors for the Appalachian region.

2013Q2_Figure_Mooney_QAppalFigure 3. Comparison of Q-Factors within the Appalachian Highlands. The gray shaded area indicates ± 20% of the average Q of all plotted studies. The coherency in Q and trend is comparable between the different studies. Due to the nature of the rock material the frequency dependency is stronger in comparison to the older Interior Plains.

Geometrical Spreading and Attenuation for Eastern North America
Drs. Atkinson (University of Western Ontario) and Boore (USGS)

    • Drs. Atkinson and Boore have finalized their new model for attenuation in ENA which they documented in a paper submitted to BSSA entitled: “The Attenuation of Fourier Amplitudes for Rock Sites in Eastern North America.” The paper has been reviewed and is in the editorial pipeline. In this paper, the authors propose an update of previous empirical attenuation models they have developed for ENA. The new model provides updated values for the anelastic attenuation and a bilinear model to represent the geometrical spreading. The main new feature of the model resides in the development of a frequency-dependent hinge distance used to define the bilinear geometrical spreading part of the model. The authors also use event-specific anelastic attenuation as part of the model development. The model includes an adjustment factor to reduce some residuals seen at distances less than 50 km. This adjustment factor is frequency dependent, and can be combined with the geometric spreading to obtain an apparent geometrical spreading function. Figure 4 below shows this function for two frequencies, along with the unmodified geometrical spreading.

2013Q2_Figure_AtkinsonBoore_GSFigure 4. Illustration of standard geometric spreading and effect of frequency-dependent geometric spreading function for a hypocentral depth of 10 km for two frequencies.

    • The authors also presented a summary of their work at the SSA Annual Meeting in April:
      Atkinson and Boore: “Empirical Evidence for the Frequency-Dependence of Geometric Spreading in Eastern North America” (Abstract here)

Using EarthScope TA Data to Develop Regional Models for Geometrical Spreading and Q in the Central U.S.
Profs. Chapman (Virginia Tech) and Pezeshk (University of Memphis)

    • The objective of this task is to use broadband data from the EarthScope Transportable Array (TA), and ANSS stations to constrain the geometric spreading and the Q-factor for the different CENA regions. The TA has provided important new data to study path attenuation in the central and eastern United States. During 2010-2011, it recorded abundant data from swarms of small-to-moderate-sized earthquakes in Arkansas, Oklahoma, and Texas (Figure 5). The TA is now moving into the Appalachian region, and has recently recorded important data for that region as well.

2013Q2_Figure_Chapman_USARRAYEventsFigure 5. Earthquakes recorded by EarthScope Transportable Array in the south-central U.S.

    • The Arkansas February 28, 2011 magnitude 4.65 event was well-recorded by the EarthScope TA (Figure 6), and serves as a good example of the type of regional variation in Lg propagation. Stations to the north of the epicenter showed different Lg amplitude compared to stations to the south of the epicenter. Figure 6 shows two stations: one to the north and one to the south of the epicenter with approximately equal distance of 500 km. From Figure 6, one can observe difference in Lg amplitudes with the station near the Gulf coast exhibiting extreme attenuation of high frequency energy, compared to the station to the North in the cratonic platform.

2013Q2_Figure_Chapman_TA_MApRecordsFigure 6. Left: Geologic map showing the locations of EarthScope TA stations installed as of April, 2011, as well as other regional broadband stations. Geology from Garrity and Soller (2009). (Figure modified from Conn, 2013). Right: Acceleration waveforms from the Arkansas 2/28/2011 earthquake for two stations in Missouri (top) and Louisiana (bottom). Vertical lines represent the start and end of the Lg wave. (From Conn, 2013).

    • For this study, the team divided central United States into the following regions: (1) the Coastal Plain region which includes Cretaceous and younger marine sedimentary units that reach thicknesses in excess of 10km along the coast of Texas and Louisiana, as well as in the East Texas basin and in the Louisiana salt basin; (2) the Paleozoic Platform which represents the region of the mid-continent where medley deformed Paleozoic sedimentary rock overlies the crystalline Precambrian cratonic basement; and (3) The Great Plains where platform rocks are overlain by a veneer of Cretaceous sedimentary units. The boundary of these three regions was based on the map of surface geology by Garrity and Soller (2009) as shown in Figure 5 and Figure 6 .
    • In this study, special attention is focused on the determination of geometric spreading and Q factor for both geometric mean of the horizontal components as well as the vertical component.
    • Figure 7 shows a preliminary Q-tomography obtained by fixing the receiver terms and using a trilinear model similar to Atkinson (2004) for a central frequency of 8 Hz. Looking closely at Figure 7, one can observe that Q-tomography based on constraining the geometric spreading, give us values that closely follow the three regions defined by Conn (2013).

2013Q2_Figure_Chapman_QTomographyFigure 7. Q Tomography for 8 Hz.

    • Figure 8, from Conn (2013) shows 8Hz regression residuals from February 28, 2011 Arkansas earthquake. Note the large negative residuals in the Gulf Coastal region, which are in excellent agreement with the tomographic results, indicating that the Gulf Coastal plain is a region of high attenuation. Conn (2013) found that the negative residuals in the Gulf Coast could be reduced by including an additional anelastic attenuation term that depends on sediment thickness in her regression models. The Cretaceous and younger marine sediments in the Gulf Coastal Plain reach great thickness near the coastline, and clearly have a strong impact on ground motion attenuation.

2013Q2_Figure_Chapman_Conn_ResidualsFigure 8. Left: Regression residuals for 8 Hz Fourier amplitudes from the 2/28/2011 Arkansas earthquake, without inclusion of a term to account for attenuation in thick Coastal Plain sediments. Right: residuals with sediment term included. Figure is taken from Conn (2013).

    • The lead researchers presented an update on their work for this task at the SSA Annual Meeting in April:
      Hosseini et al: “Regional Investigation of Geometrical Spreading and Quality Factor in Central United States” (Abstract here)

Investigation of Focal Depth Effects on Attenuation
Dr. Frankel (USGS)

    • Dr. Frankel examined the hypocentral-depth dependence of the amplitude decay for a series of Charlevoix events. No clear dependence on the hypocentral depth could be resolved with the current set of earthquakes. Frankel determined coda envelopes for additional, recent earthquakes in the area, but found that they did not have sufficient signal at the longer lapse times in the coda used in the previous analysis. He is continuing the writing of a paper on the amplitude decay of the coda-normalized S-waves. He has also continued to run finite-fault stochastic simulations using 1/R geometrical spreading (up to 80 km distance), as the next step for producing broadband synthetics to develop new GMPE’s.

Finite Fault and Stochastic Simulations (Tasks E and F)

Source Models for Stochastic Method Simulations (Task F)

    • Dr. Boore (USGS) finalized his generalized two-corner source spectral models. The models have been fully implemented in his suite of stochastic method simulation programs (SMSIM). A modified version of the code was also developed by Drs. Abrahamson (University of California Berkeley) and Di Alessandro (Geopentech) to do a systematic parameter search to optimize the fit to validation scenarios for the collaborative simulation validation exercise described below.

CENA Parameterization for Finite Fault Simulations (Task E)

    • Dr. Frankel (USGS) has continued the process of making broadband synthetic seismograms for ENA earthquakes, for use in developing new GMPE’s. He is making a set of input files for M5.5, 6.5 and 7.5 sources for the low-frequency deterministic synthetics and for the high-frequency stochastic synthetics. Dr. Frankel has also been analyzing the broadband synthetics he made previously for the WUS to determine the relationship between apparent corner frequency and rupture properties over a range of distances.

Collaborative Simulation Validation Exercise (Task E)

    • The collaborative simulations group (PEER, SWUS and SCEC) has been very active in the last few months and continued to have their weekly phone calls. The first report on the evaluation of the validation for method currently implemented on the SCEC BroadBand Platform (BBP) is being prepared.
    • The first validation exercise includes two main parts:
      • Part A. Comparison of simulations to recorded events (five WUS and two Japanese events have been completed in this first round). WUS GMPEs are also compared to the recorded events as an independent way to evaluate frequency-dependent event terms.
      • Part B. Comparison of simulation to WUS GMPEs for generic scenarios empirically well-constrained.

Results from parts A and B provide complementary information on the methods’ performance and are used for the evaluation. Many different plots and summary tables are produced directly on the platform for parts A and B results. The BBP is constantly evolving as new products and features are implemented by Mrs. Silva and Maechling form the SCEC software development team.

    • Methods evaluated and currently implemented on the SCEC BBP are:
      • Graves and Pitarka (led by Dr. Graves, USGS and Mr. Bayless, URS)
      • University of California Santa Barbara method (led by Prof. Archuleta)
      • San Diego State University method (led by Prof. Olsen)
      • EXSIM (led by Dr. Assatourians and Prof. Atkinson, University of Western Ontario)
      • University of Nevada Reno method (led by Prof. Anderson)

Additional methods are also considered for future validations, including these so far:

      • Irikura Recipe (led by Dr. Somerville and Mr. Bayless, URS)
      • SMSIM (led by Dr. Di Alessandro, Geopentech and Dr. Boore, USGS)
    • The simulations group met in Pomona on June 11 and 12 to discuss preliminary results and to finalize the documentation required for the evaluation panel. The meeting was very productive and allowed the participants to discuss the science behind the various methods, the validation exercise process and the parameterization for forward scenarios.
    • The evaluation meeting was held in Menlo Park on June 26. The evaluation panel was formed under SCEC’s umbrella and consists of Prof. Dreger (University of California Berkeley, panel chair), Prof. Beroza (Stanford University), Prof. Day (San Diego State University), Dr. Goulet (PEER), Prof. Jordan (University of Southern California), Dr. Spudich (USGS) and Prof. Stewart (University of California, Los Angeles). At the meeting, each modeler presented a summary of their method and their self-assessment on its performance for WUS scenarios and GMPEs. The panel was presented with a large amount of documentation and is currently preparing the first validation report.
    • The CENA scenarios and site factors are being updated by Dr. Goulet and the second round of validation against CENA events is about to start. The first round was completed in late 2012. The NGA-East December 2012 working meeting on the topic led to improvements to the parameterization of events and to the validation process as a whole. Modelers are concurrently working on revising the parameters as needed for CENA events.
    • Several participants presented parts of the validation exercise at the SSA Annual Meeting in April in the special session “Implementation of Physics-Based Earthquake Source and Ground Motion Findings in Engineering Solution Model”:
        • Goulet et al.: “Development and Implementation of Validation Exercises Using the SCEC Broadband Strong Ground Motion Simulation Platform” (Abstract here)
        • Somerville et al.: “Validation of the Graves-Pitarka 2010 Strong Motion Simulation Procedure on the SCEC Broadband Platform” (Abstract here)
        • Bayless and Somerville. : “Behavior of Multiple Broadband Ground Motion Simulation Techniques for a Suite of Earthquake Scenarios Using Multiple Rupture Model Generators on the SCEC Broadband Platform” (Abstract here)
        • Crempien and Archuleta: “UCSB Broadband Ground Motion Method Validation for California Earthquakes” (Abstract here)
        • Assatourians et al.: “Validation Study of Stochastic Finite Fault Module (EXSIM) Implemented in SCEC Broadband Platform” (Abstract here)

Sigma (Standard Deviation) (Task J)

Development of models for standard deviation (within-event and single-station)

  • The Sigma Working Group had a conference call to review the latest progress on single-station sigma. They are analyzing the recently compiled NG-West2 dataset, especially for small magnitude data, to quantify single-station sigma. Dr. Al Atik (Al Atik Consulting) is leading and coordinating the efforts.
  • On a related topic, Dr. Goulet (PEER) collaborated with the USGS National Seismic Hazard Mapping Program (NSHMP) by providing residuals analyses for CENA. She used the preliminary NGA-East database and Vs30 assessments with existing SRCs GMPEs. The residuals were computed using a mixed-effects regression and also included the evaluation of single-station sigma through the computation of site terms.
  • The Sigma Working Group will repeat this type of exercise using both codes developed by Dr. Al Atik (Al Atik consulting) and Dr. Goulet with the NGA-East dataset once it has been completely reviewed for QA/QC.

References

Conn, Ariel (2013). Q Models for Lg Wave Attenuation in the Central United States, M.S. Thesis, Virginia Tech, URL http://vtechworks.lib.vt.edu/handle/10919/19317.

Garrity, C.P. and D.R. Soller (2009). Database of the geologic map of North America – adapted from the map by J.C. Reed, Jr. and others (2005). U.S. Geological Survey Data Series 424, URL http://pubs.usgs.gov/ds/424/.