PEER Ground Motion Database - Technical Report PDF

With special thanks to:
Technical Report for the

PEER Ground Motion Database Web Application
Beta Version – October 1, 2010
PEER Ground Motion Database – Beta Version

Technical Report i
TABLE OF CONTENTS
Page
1.0 INTRODUCTION.............................................................................................................1
1.1 OBJECTIVE, SPONSORSHIP, AND LIMITS ..................................................................1
1.2 EVOLUTION OF PEER GROUND MOTION DATABASE CONCEPT ..............................1
1.3 PRODUCTS ..............................................................................................................2
1.4 PROJECT TEAM .......................................................................................................3
2.0 DGML CAPABILITIES, FEATURES, AND OPERATION...........................................3
2.1 OVERVIEW ..............................................................................................................3
2.2 DATABASE ..............................................................................................................5
2.3 FORMING TIME SERIES SETS BASED ON RESPONSE SPECTRAL SHAPE AND
OTHER CRITERIA FOR HORIZONTAL COMPONENTS ................................................7
2.3.1 Step 1 – Developing the Target Spectrum.................................................7
2.3.2 Step 2 – Specifying Criteria and Limits for Searches for Time series
Records on the Basis of Spectral Shape ....................................................9
2.3.3 Step 3 – Search of Database, Selection of Records, and Saving of
Records, Plots and Supporting Information ............................................12
2.4 SUPPLEMENTAL INFORMATION ABOUT FORMING RECORDS SETS
CONTAINING RECORDS WITH PULSES ...................................................................13
2.4.1 Database for Records With Pulses ..........................................................13
2.4.2 Selecting Records with Pulses Within the PGMD ..................................15
3.0 REFERENCES ................................................................................................................16
TABLES
Table 1 Significant Duration of Shaking Based on Kempton and Stewart (2006)

Table 2a Fault-Normal Pulselike Ground Motion Records in PGMD
Table 2b Fault-Parallel Pulselike Ground Motion Records in PGMD
Table 3a Magnitude-Distance Distribution of Fault-Normal Pulselike Records
Table 3b Magnitude-Distance Distribution of Fault-Parallel Pulselike Records
FIGURES
Figure 1 Illustration and Calculation of a Conditional Mean Spectrum

Figure 2 Magnitude and Distance Distribution for PEER NGA Records in PGM Database
Figure 3 Code Response Spectrum
Figure 4 Example of Individual and Average NGA Response Spectra from Five Models
Constructed by PGMD Software Tool
Figure 5 Example of Conditional Mean NGA Response Spectrum Constructed by PGMD
Software Tool
Figure 6 Illustration of Construction of Conditional Mean Spectra for Different Periods for
Multiple Scenario Earthquakes (Based on Cornell 2006)
Figure 7 Example of Record Search Performed by PGMD Software Tool
Technical Report i
TABLE OF CONTENTS
(Continued)
Figure 8 Correlation of Pulse Period with Earthquake Magnitude by Baker 2007

Figure 9 Correlations of Pulse Period with Earthquake Magnitude by Different Researchers
APPENDICES
Appendix A PGMD Users Manual

Appendix B Summary of PEER-NGA Records Included in and Excluded from PGMD
Database
Table B-1 List of PEER-NGA Records Included in PGMD Database
Table B-1a List of PEER-NGA Records Included in PGMD Database (Taiwan)
Table B-1b List of PEER-NGA Records Included in PGMD Database (California)
Table B-1c List of PEER-NGA Records Included in PGMD Database (Other Regions)
Table B-2 List of PEER-NGA Records Excluded from PGMD Database

Technical Report ii
DESIGN GROUND MOTION LIBRARY
1.0 INTRODUCTION
1.1 OBJECTIVE, SPONSORSHIP, AND LIMITS

This document describes the development of the PEER Ground Motion Database (PGMD) –
Beta Version, which represents an improved web implementation of the Design Ground Motion
Library (DGML) Version 2.0 software. The objective of the current PGMD project is to create
an electronic library of recorded ground motion acceleration time series suitable for use by
engineering practitioners for time-history dynamic analyses of various facility types in
California and other parts of the western United States. The PEER Ground Motion Database
(PGMD) project is jointly sponsored by the California Geological Survey - Strong Motion
Instrumentation Program (CGS-SMIP) and the Pacific Earthquake Engineering Research
Center - Lifelines Program (PEER-LL). The PEER Ground Motion Database (PGMD) – Beta
Version is currently limited to recorded time series from shallow crustal earthquakes of the type
that occurs in the western United States. Time series from subduction zone earthquakes are not
part of the Library during this project. However, near future developments of the PGMD will
incorporate the extensive database collected within the PEER NGA-WEST 2 project, which
also includes events recorded after 2005. The library will be supplemented with elastic
response spectra generated for different damping ratios ranging from X to Z.
1.2 EVOLUTION OF PEER GROUND MOTION DATABASE CONCEPT

The original idea of creating a web-based, searchable database of ground motion records was
first embanked by PEER based on the effort to collect the most important ground motion
records from around the world. PEER took great effort in ensuring that all the data had been
processed consistently and that it was reliable in all regards. PEER also gathered related
metadata such as earthquake magnitude, various site-to-source distance measures, style of
faulting, local site conditions at the recording stations, and other relevant engineering
parameters. All the collected data were made available to the public by the form of web-based
ground motion databases: the first release was copyrighted in 2000 and the second release was
copyrighted in 2005. The 2005 database was used to develop improved attenuation relations
within the Next Generation Attenuation Models (NGA) project.
However, during the time of the web-based database creation, research on the relationship of
ground motion characteristics and structure response showed that selecting records to give
Technical Report 1
realistic (not overly conservative) estimates of structural inelastic response for different types
of structures required consideration of selection of time series dependent on ground motion
intensity and a wide range of structure characteristics. In fact, work by the PEER Ground
Motion Selection and Modification Working Group (GMSM), as presented in the 2006 and
2007 Technical Sessions of the Annual Meetings of the Consortium of Organizations for
Strong Motion Observation Systems (COSMOS 2006 and 2007) demonstrated the importance
to inelastic response of “conditioning” ground motion response spectra dependent on the
intensity of shaking and the characteristics of the structure. These considerations, as well as
considerations of a range of preferences by designers and analysts (such as the size of the time
series sets) indicated that a “dynamic” library was needed, i.e. a library permitting the tailoring
of the selection of time series records to specific project needs and designer preferences.
In this framework, it was decided to expand the existing ground motion database so to
incorporate an improved version of the Design Ground Motion Library (DGML) Version 2.0,
so that advance utilities has been developed to enable the rapid sorting and selection of the time
series from the PEER ground motion database based on appropriate criteria and user needs
rather than having pre-selected fixed record sets as it was used to be in the old practice.
1.3 PRODUCTS
Products of this project include:
(1) An internet web-based application that includes (a) the original unscaled PEER
ground motion database, inclusive of acceleration time series and corresponding
acceleration response spectra, and (b) a feature based on an improved DGML
(Version 2.0) software tool that allows the user to select, scale, and evaluate time
series for applications.
(2) This Technical Report describing the PEER Ground Motion Database – Beta Version
(PGMD) development.
(3) The Users Manual for the PGMD, which is included as Appendix A of this Technical
Report.
(4) Featured videos describing the product features and applications.

Technical Report 2
1.4 PROJECT TEAM
The PEER Ground Motion Database – Beta Version has been implemented in its on-line format
inside the PEER center: Professor Stephen Mahin and Dr. Bozorgnia led the project; James C.
Way was responsible for the web-based interface between the original DGML software and the
existing PEER Ground Motion Database; Camila Coria and Dr. Carola Di Alessandro provided
documentation and videos. The original DGML software was developed by a multi-disciplinary
project team of practitioners and researchers in structural engineering, geotechnical
engineering, and seismology, including the following organizations and individuals: Geomatrix
Consultants, Inc., prime contractor (Robert Youngs, Gang Wang, Maurice Power, Zhihua
(Lillian) Li, Faiz Makdisi, and Chih-Cheng Chin); Simpson Gumpertz & Heger, Inc. (Ronald
Hamburger and Ronald Mayes); Parsons Brinckerhoff (Roupen Donikian); Quest Structures
(Yusof Ghanaat); Pacific Engineering & Analysis (Walter Silva); URS Corporation (Paul
Somerville); Earth Mechanics (Ignatius Po Lam); Professors Allin Cornell and Jack Baker,
Stanford University; and Professor Stephen Mahin, University of California, Berkeley.
2.0 PGMD CAPABILITIES, FEATURES, AND OPERATION
2.1 OVERVIEW
The PGMD has the broad capability of searching for time series record sets in the library
database on the basis of (1) the characteristics of the recordings in terms of earthquake
magnitude and type of faulting, distance, and site characteristics, (2) the response spectral
shape of the records in comparison to design or target response spectra, and (3) other record
characteristics including duration and the presence of velocity pulses in near-fault time series.
Other criteria and limits can be specified by the user to constrain searches for time series. Also,
supplemental searches can be conducted for individual records or records from selected
earthquakes or stations and these records can be evaluated and incorporated in data sets of
search results.
Response spectral shape over a period range of significance to structural response has been
found to be closely correlated to inelastic structural response and behavior in a number of
studies (e.g. Shome et al. 1998; Cordova et al. 2001; Luco and Cornell 2006; Bazurro and
Jalayer 2003; Baker and Cornell 2004; Luco and Bazurro 2004; Baker and Cornell 2005, 2006,
2008; GMSM Working Group, 2009). The period range of significance may include periods
shorter than the fundamental structure period because of higher-mode effects and periods
longer than the fundamental structure period because of structure softening during inelastic

Technical Report 3
response. Although sets of time series are often formed in practice to provide an aggregate
match to a probabilistic response spectrum (equal or uniform hazard spectrum, UHS) for design
purposes, there may be conservatisms involved in doing so. As summarized by Cornell (2006),
the UHS is not the response spectrum of ground motion from a single earthquake and therefore
has an artificial shape. The spectral ordinates at different periods may be driven by two or more
different earthquakes and therefore the UHS may be overly broad and thus conservative for a
single earthquake. Hazard deaggregation can be carried out to identify the dominant earthquake
sources and narrower deterministic scenario earthquake design spectra scaled to the level of the
UHS can be defined as appropriate.
It has also been shown that a deterministic scenario earthquake spectrum may also be overly
broad if all spectrum ordinates are at a high “epsilon (ε)” value, where (ε) is defined as the
number of standard deviations above or below the median spectrum ordinate for a given
earthquake, distance, and site condition. In this case, it is technically justifiable and appropriate
to define “conditional mean spectra” that are more narrow-banded (Baker and Cornell 2006;
Baker 2006; Cornell, 2006). The term “conditional mean spectrum” refers to the mean of
spectra that are conditioned on a spectral value at a given period being at ε number of standard
deviations above or below the median ground motion for the particular earthquake, distance,
and site condition. The conditional mean spectrum is at ε number of standard deviations at the
period of conditioning, while the absolute value of the number of standard deviations at other
points on the conditional mean spectrum would be less. Figure 1(a) illustrates the construction
of a conditional mean spectrum given a target value of spectral acceleration at a particular
period on a spectrum constructed for a particular epsilon (in Figure 1(a), ε = 2 at a period of 1
second). It can be seen that at periods away from 1 second, the conditional mean spectrum is
below the ε=2 level. The conditional mean spectrum reflects the lack of perfect correlation
between spectral accelerations at different periods, so that if a rare high spectral acceleration
(e.g. ε=2 value in Figure 1(a)) is observed at one period, it is unlikely that it will be observed at
other periods. The steps involved in calculating a conditional mean spectrum are summarized in
Figure 1(b). The work of the PEER Ground Motion Selection and Modification Working
Group, as presented at the COSMOS 2006 and 2007 Annual Meeting Technical Sessions
(COSMOS 2006, 2007) and, at the GEESD IV 2008 Conference (Goulet et al., 2008 and
reported on by the GMSM Working Group (2009), illustrated that selecting time series having
response spectral shapes corresponding to the conditional mean spectrum was an effective
approach for obtaining a set of time series giving realistic inelastic structure response. As
described in Section 2.3, the PGMD – Beta Version web-based software tool enables the user
Technical Report 4
to specify different options, including the conditional mean spectrum option, for constructing a
design or target response spectrum and to search for time series having spectral shapes that are
most similar to the target spectrum over a user-defined period range of significance.
A number of studies have shown that strong velocity pulses in ground motion time series
records, such as often occur in near-source ground motions due to near-source fault rupture
directivity effects, can impose severe demands on structures (e.g. Bertero et al.; 1978;
Anderson and Bertero 1987; Hall et al. 1995; Iwan 1997; Krawinkler and Alavi 1998; Alavi
and Krawinkler 2001; Menun and Fu 2002; Makris and Black 2003; Mavroeidis et al. 2004;
Akkar et al. 2005; Luco and Cornell 2006; Baker and Cornell, 2008). The strongest pulses tend
to occur closer to the fault-strike-normal (FN) direction than the fault-strike-parallel (FP)
direction (Somerville et al. 1997). FN records having velocity pulses that may be associated
with directivity effects have been systematically identified in the PEER NGA database by
Baker (2007). (The PEER-NGA database is described in Section 2.2 below.) Records with
pulses have been identified by other researchers (e.g. Somerville 2003; Mavroeides and
Papageorgiou 2003; Bray and Rodriguez-Marek 2004; Fu and Menun 2004). Records with
velocity pulses are discussed in Section 2.4.1 and the presence of velocity pulses in records can
be a criterion in searches for records in the PGMD.
2.2 DATABASE
The source of the database for the PGMD is the PEER Next-Generation Attenuation (NGA)
project database of ground motion recordings and supporting information
(http://peer.berkeley.edu/nga/). This database was developed as the principal resource for the
development of updated attenuation relationships in the NGA research project coordinated by
PEER-Lifelines Program (PEER-LL), in partnership with the U.S. Geological Survey (USGS)
and the Southern California Earthquake Center (SCEC) (Chiou et al. 2006, 2008; Power et al.,
2008). The database represents a comprehensive update and expansion of the pre-existing
PEER database (Chiou et al., 2008). The ground motion records are originally from strong
motion networks and databases of CGS-CSMIP and USGS and other reliable sources,
including selected record sets from international sources. The PEER NGA database includes
3551 three-component recordings from 173 earthquakes and 1456 recording stations. 369
records from the PEER NGA database were not included in the current PEER Ground Motion
Database of 3182 records. The records were not included for various reasons including one or
more of the following: (a) records considered to be from tectonic environments other than
shallow crustal earthquakes in active tectonic regions, e.g. records from subduction zones; (b)
Technical Report 5
earthquakes poorly defined; (c) records obtained in recording stations not considered to be
sufficiently close to free-field ground surface conditions, e.g. records obtained in basements or
on the ground floors of tall buildings; (d) absence of information on soil/geologic conditions at
recording stations; (e) records had only one horizontal component; (f) records had not been
rotated to FN and FP directions because of absence of information on sensor orientations or
fault strike; (g) records of questionable quality; (h) proprietary data; (i) duplicate records; and
(j) other reasons. Records selected for the PGMD are tabulated in Table B-1 of Appendix B,
and records not included and reasons for exclusion are tabulated in Table B-2). Figure 2 shows
the magnitude and distance distribution of the included records.
Acceleration time series in the PGMD that can be searched for on the basis of record
characteristics and other criteria (see Section 2.3.2) are horizontal components that have been
rotated to FN and FP directions. The use of rotated time series in the PGMD does not imply
that they are for use in time series analyses in FN and FP directions only, and they can be used
in time series sets in the same manner as time series in the as-recorded orientations in other
databases. The rotation to FN and FP directions does, however, provide additional information
with respect to the seismological conditions under which the recordings were obtained, and, as
mentioned in Section 2.1, records in the FN direction have been found to often contain strong
velocity pulses that may be associated with rupture directivity effects.
Ground motion parameters quantified for time series in the DGML database are response
spectra, peak ground acceleration (PGA), peak ground velocity (PGV), peak ground
displacement (PGD), significant duration, assessments of the lowest usable frequency (longest
usable period) for response spectra, and presence and periods of strong velocity pulses.
Significant duration was calculated as the time required to build up from 5% to 95% of the
Arias Intensity (a measure of energy) of the acceleration time series (refer, for example, to
Kempton and Stewart (2006) for definitions of Arias Intensity and significant duration). The
recommended lowest usable frequency is related to filtering of a record by the record
processing organization to remove low-frequency (long-period) noise. Filtering results in
suppression of ground motion amplitudes and energy at frequencies lower than the lowest
usable frequency such that the motion is not representative of the real ground motion at those
frequencies. It is a user’s choice in PGMD on whether to select or reject a record on the basis of
the lowest usable frequency. Because of the suppression of ground motion at frequencies lower
than the lowest usable frequency, it is recommended that selected records have lowest usable
frequencies equal to or lower than the lowest frequency of interest.

Technical Report 6
A major effort was made in the PEER-NGA project to systematically evaluate and quantify
supporting information (metadata) about the ground motion records, including information
about the earthquake, travel path from the earthquake source to the recording station site, and
local site conditions. Metadata in the PEER-NGA database are described in the NGA flatfile
and documentation: http://peer.berkeley.edu/products/nga_flatfiles_dev.html. Every record in
the database was assigned a unique record number (NGA#) for identification purposes.
Metadata that have been included for records in the PEER Ground Motion Database are:
earthquake name, year, magnitude, and type of faulting; measures of closest distance from
earthquake source to recording station site (closest distance to fault rupture surface and Joyner-
Boore distance); recording station name; site average shear wave velocity in the upper 30
meters, VS30; and NGA#.
The PGMD also provides access to the vertical ground motion time series and their response
spectra if available. Vertical time series and response spectra are scaled by the same scale
factors developed for their horizontal components, and they can be visualized together with the
horizontal components. These features are provided as a convenience to users for developing
three-component sets of time series.
2.3 FORMING TIME SERIES SETS BASED ON RESPONSE SPECTRAL SHAPE AND OTHER
CRITERIA FOR HORIZONTAL COMPONENTS
The formation of data sets based on response spectral shape and other criteria is a three-step
process: (1) specification of the design or target response spectrum; (2) specification of criteria
and limits for conducting searches for time series records; and (3) search of database and
selection and evaluation of records.
2.3.1 Step 1 – Developing the Target Spectrum

Three options are provided within the PGMD for developing the target spectrum:
Option 1 – Specify Code Spectrum. For this option, the target spectrum is the design
earthquake spectrum or the maximum considered earthquake (MCE) spectrum as formulated in
the NEHRP Provisions, (BSSC 2003), ASCE Standard ASCE/SEI 7-05 (ASCE 2006), and the
International Building Code, (ICC 2006). As indicated in Figure 3, the Code design spectrum is
completely specified by three parameters which are obtained using the design ground motion
maps and other provisions in the Code document: site-class-adjusted 0.2 second spectral
acceleration, SDS; site-class-adjusted 1.0-second spectral acceleration, SD1; and the transition
Technical Report 7
period, TL, from constant spectral velocity (for which spectral accelerations are proportional to
1/T) to constant spectral displacement (for which spectral accelerations are proportional to
1/T2). The user enters the values for these three parameters and the DGML software tool
constructs and plots the response spectrum.
Option 2 – User-Defined Spectrum. The user may enter any response spectrum as a table of
periods and response spectral accelerations and the tool constructs and plots the spectrum. Such
response spectra may be either probabilistic (uniform hazard, UHS) or deterministic (scenario
earthquake) response spectra developed by the user.
Option 3 – Spectrum based on PEER Next Generation Attenuation (NGA) Relationships.

Developing a spectrum for Option 3 is a special case of Option 2. For Option 3, the tool
constructs a deterministic scenario earthquake spectrum using a user-selected set of ground
motion prediction equations (GMPEs) developed in the NGA project for shallow crustal
earthquakes in active tectonic regions. Five different GMPE models were developed in the
NGA project: (Abrahamson and Silva 2008; Boore and Atkinson 2008; Campbell and
Bozorgnia 2008; Chiou and Youngs 2008a; and Idriss 2008). The applicability of the Idriss
(2008) relation is currently restricted to Vs30 equal to or greater than 450 m/sec. The user enters
the earthquake parameters, travel path parameters, and site parameters (e.g., earthquake
moment magnitude, type of faulting, fault-to-site distance, site shear wave velocity in the upper
30 meters (VS30), and other parameters needed for the selected NGA models, and the tool
constructs the individual response spectra and an average of the spectra for the models. The
user also specifies whether the spectrum is a median spectrum or a spectrum at a selected
epsilon (ε) number of standard deviations. Figure 4 illustrates individual and averaged NGA
response spectra for a scenario earthquake constructed by the software tool. The user also has
the option of constructing the response spectrum as a conditional mean spectrum, such as
illustrated in Figure 1, using the correlation model of Baker and Jayaram (2008). The equations
and steps involved in calculating a conditional mean spectrum are also shown in Figure 1. The
construction of the conditional mean spectrum is done by the web-base software tool
conditional upon spectral acceleration at level ε for a user-specified T_eps, i.e., T_eps is the
period for which the spectrum is at the number of standard deviations, ε, specified for the
attenuation relationships. (Note that the symbol T1 in Figure 1 is equivalent to the symbol T_eps
as used herein and in the Users Manual, Appendix A.) Multiple conditional mean spectra can
be constructed for different periods T_eps. Figure 5 illustrates conceptually construction of a
conditional mean spectrum by the web-based software tool. Figure 6 illustrates conceptually

Technical Report 8
construction of conditional mean response spectra for different periods for two scenario
earthquakes in order to more closely match, in aggregate, a UHS design spectrum. As shown in
Figure 6 for scenario earthquakes A and B, a single conditional mean spectrum for each
earthquake could fall substantially below the UHS at periods distant from the period at which
the conditional mean spectrum is at the target epsilon. Therefore, as shown in the figure, two
(or more) conditional mean spectra could be required for each scenario earthquake to satisfy
Code requirements for an aggregate match to the design UHS.
In order to construct a conditional mean spectrum using the PGMD tool, the user must specify
the value of ε. There are a variety of approaches that may be used to select the appropriate
value of ε. Many modern PSHA software applications provide information on the values of ε
that are representative of the computed ground motion hazard (ε deaggregation). The user can
use this information to select an appropriate value of ε. Alternatively, the user may use a target
value of spectral acceleration at the spectral period of interest, T_eps, to define the appropriate
value of ε. The current version of the PGMD web implemented tool applies the same value of ε
to all of the selected NGA ground motion relationships. The user can adjust the entered value
of ε until the computed average conditional mean spectrum matches the target spectral
acceleration at T_eps. A third alternative is for the user to construct the conditional mean
spectrum outside of the PGMD tool and then enter this spectrum using Option 2 – User-
Defined Spectrum described above.
2.3.2 Step 2 – Specifying Criteria and Limits for Searches for Time series Records on
the Basis of Spectral Shape
A basic criterion used by the PGMD to select a representative acceleration time series is that
the spectrum of the time series provides a “good match” to the user’s target spectrum over the
spectral period range of interest. The user defines the period range of interest. The quantitative
measure used to evaluate how well a time series conforms to the target spectrum is the mean
squared error (MSE) of the difference between the spectral accelerations of the record and the
target spectrum, computed using the logarithms of spectral period and spectral acceleration.
The PGMD web-based tool searches the database for records that satisfy general acceptance
criteria provided by the user and then ranks the records in order of increasing MSE, with the
best-matching records having the lowest MSE.
The focus of the PGMD is on selecting “as recorded” strong ground motion acceleration time
series for use in seismic analyses. (In fact the records do include the effects of processing by

Technical Report 9
the supplying agency, such as filtering and baseline correction.) Therefore, the tool does not
provide the capability of altering the frequency content of the recordings to better match a
target spectrum. However, it does provide the ability to linearly scale recorded time series to
improve their match to the target spectrum and select time series that have the best spectral
match. The user has three options for scaling. One option is to apply a scale factor that
minimizes the MSE over the period range of interest. This approach results in selection of
records that have spectral shapes that are similar on average to the target over the period range
of interest, but whose spectra will oscillate about the target. The second option is to scale the
records so that the spectral acceleration at a specific period matches the target spectral
acceleration at that period. This provides a set of scaled time series whose spectral
accelerations are all equal to the target at the specified period. A third option of not scaling is
also available. The choice of scaling approach is up to the user. For all three options, the MSEs
of the records are calculated and ranked.
Calculation of MSE. The MSE between the target spectrum and the response spectrum of a
recorded time series is computed in terms of the difference in the natural logarithm of spectral
acceleration. The period range from 0.01 second to 10 seconds is subdivided into a large
number of points equally-spaced in ln (period, Ti) (100 points/log cycle, therefore 301 points
from 0.01 second to 10 seconds, end points included) and the target and record response spectra
are interpolated to provide spectral accelerations at each period, SAtarget(Ti), and SArecord(Ti),
respectively. The MSE is then computed using Equation (1) over periods in the user-specified
period range of interest:
 w (Ti ) {ln [SA target (Ti )]  ln [ f  SA record (Ti )]}2

MSE  i (1)
 w (Ti )
i
Parameter f in Equation (1) is a linear scale factor applied to the entire response spectrum of the
recording. Parameter w(Ti) is a weight function that allows the user to assign relative weights to
different parts of the period range of interest, providing greater flexibility in the selection of
records. The simplest case is to assign equal weight to all periods in the period range of interest
(i.e. w(Ti) = 1), but the user may wish to emphasize the match over a narrow period range while
maintaining a reasonable match over a broad period range. Arbitrary weight functions may be
specified, as described in the Users Manual.

Technical Report 10
The PGMD web-based tool allows the user to select recordings for which the geometric mean
of the two horizontal components provides a good match to the target spectrum. In this case the
MSE is computed over both components using Equation (1) with the same value of f applied to
both components. This process maintains the relative amplitude of the two horizontal
components.
Calculation of the Scale Factor. As discussed above, the user has three options for specifying
the scale factor f. The simplest is to use unscaled records, that is f = 1.0. The second approach is
to scale the records to match the target spectrum at a specific period, denoted Ts. In this case the
scale factor is given by:
SA target (Ts )
f  (2)
SA record (Ts )
The third option is to apply a scale factor that minimizes the MSE. This approach produces
scaled recordings that provide the best match to the spectral shape of the target spectrum over
the user-specified period range of interest. Minimization of the MSE as defined in Equation (1)
is achieved by a scale factor given by the mean weighted residual in natural logarithm space
between the target and the record spectra:
 SA target (Ti ) 
w(Ti ) ln 
 SA record

(Ti ) 
ln f  i (3)
 w(Ti )
i
When record selection is based on simultaneously considering both horizontal components, the
scale factor computed using Equation (3) minimizes the MSE between the target spectrum and
the geometric mean of the spectra for the two horizontal components. The geometric mean
(GM) of FN and FP horizontal accelerations is given by:
SA GM  SA FN  SA FP or ln SA GM  (ln SA FN  ln SA FP )/2 (4)
For all three scaling options, the MSE is computed using Equation (1). Note that for all options,
it is necessary for the user to specify the weight function because it is used to calculate the
MSE and order the results with respect to the degree of match between target spectrum and
spectra of recordings over the user-specified period range of significance.

Technical Report 11
Specification of Search Criteria for Records. The user specifies the ranges of parameters over
which searches are to be conducted and other limits and restrictions on the searches. These may
include: event name; NGA number; station name; earthquake magnitude range; type of
faulting; distance range; VS30 range; significant duration range; whether records are to exclude,
include, or be limited to pulse records; and limits on the scale factor f.
Table 1 summarizes median values of significant duration as defined in Section 2.2 as a

function of magnitude, closest source-to-site distance, and VS30 (m/sec) that can be used as a
guide in specifying duration ranges for searches (if desired) and/or in evaluating durations of
selected records. The values in Table 1 are based on Kempton and Stewart (2006). The
standard deviation of the natural logarithm of duration obtained by Kempton and Stewart
(2006) is 0.44, corresponding to a factor of 1.55 between median estimates and plus-or-minus
one standard deviation estimates. Pulse records are discussed in Section 2.4.
Other Criteria. Other criteria to be specified by the user are (1) total number of records for the
search that will be displayed in the “ground motion record display window” (see Users
Manual); and (2) total number of records for which the average spectrum will be calculated.
2.3.3 Step 3 – Search of Database, Selection of Records, and Saving of Records, Plots
and Supporting Information
The web-based software tool scans the database, selects all records meeting user-specified
criteria as summarized above, scales records to match the target spectrum, and ranks records in
order of increasing MSE. The implemented software tool also has the capability (termed
Supplementary Search in the Users Manual) to search for specific records according to
specified NGA record sequence number or by earthquake name or recording station name.
Selected records from a Supplementary Search are scaled and ranked by MSE and can be
incorporated into final data sets as desired by the user. This search capability was added so that
users can examine any record or group of records and further fine-tune the search results based
on user preferences.
Figure 7 (taken from Figure 43 of the Users Manual) illustrates the PGMD graphic interface
used to specify primary search criteria and list and plot search results including time series and
individual and average response spectra of scaled records sets compared to a specified design
or target spectrum. One-, two-, or three-component time series of a record can also be viewed
at an expanded time scale, if desired to examine details of the time series, using a feature called

Technical Report 12
“Zoom in Time”. The Users Manual describes in detail the procedures for specifying criteria
and obtaining and viewing results.
Search Report and Saving of Search Results. For a selected record set, a search report is
prepared as described in the Users Manual. The search report includes: search criteria;
summary of earthquake, distance, and station/site information; record scaling factors and
MSEs; scaled record characteristics including PGA, PGV, PGD, acceleration response spectra,
presence of pulses and pulse periods, significant durations, and recommended lowest usable
frequencies; and scaled average spectral accelerations for the selected record set along with the
target or design spectral accelerations. The search report can be saved as a Windows spread-
sheet file. Although the search results are based on horizontal records, the response spectra for
corresponding vertical records can also be saved together with their horizontal counterparts in
the search report. Spectra and acceleration, velocity, and displacement time series plots can be
saved as chart files in .png format. The horizontal and/or vertical components of the selected
acceleration time series can also be saved; the saved time series are the unscaled original data
from the PEER NGA database. Therefore, the user can further modify the time series if
required or desired for any purposes (e.g., fine-tune record scaling factors to meet building
code requirements for degree of match of an average spectrum of a selected records time series
set with a design spectrum; rotate time series; or adjust match of record spectra to a design
spectrum through frequency content altering methods.).
2.4 SUPPLEMENTAL INFORMATION ABOUT FORMING RECORDS SETS CONTAINING

RECORDS WITH PULSES
Within the NGA database, ground motion records have been identified as having strong
velocity pulses that may be associated with fault rupture directivity effects. Prior research
(Somerville et al., 1997) has indicated that the strongest pulses are generally more closely
aligned with the FN direction than the FP direction. As discussed previously, the time series
records in the PGMD have been rotated into FN and FP directions and their response spectra
calculated for these directions.
2.4.1 Database for Records with Pulses

The principal resource used in identifying and characterizing records with velocity pulses for
the PGMD has been the research by Baker (2007). Baker analyzed all records within the NGA
database and identified FN records having strong velocity pulses that may be associated with
rupture directivity effects. The basic approach followed by Baker was to use wavelet analysis

Technical Report 13
to identify the largest velocity pulses. General criteria that were used in defining records with
pulses were (1) the pulse is large relative to the residual features of the ground motion after the
pulse is extracted, (2) the pulse arrives early in the time series, as would be expected for pulses
associated with rupture directivity effects, and (3) the absolute velocity amplitudes are large
(PGV of record equal to or greater than 30 cm/sec). The detailed criteria and results for the FN
components are described by Baker (2007). The same criteria were applied by Baker for the FP
component and those results as well as more detailed results and documentation of analyses for
both components are contained on the website http://www.stanford.edu/~bakerjw/pulse-
classification.html. Note that there can be no assurance that velocity pulses of records in the
database are due to directivity effects without more detailed seismological study of individual
records. It is likely that other seismological factors may have caused or contributed to the
velocity pulses of some records. However, while the causative mechanisms for the pulses are
uncertain, it is expected that the pulses are similar to those caused by directivity and therefore
suitable for use in modeling effects of directivity pulses on structures.
Somerville (2003), Mavroeidis and Papageorgiou (2003), Bray and Rodriguez-Marek (2004),
and Fu and Menon (2004) also prepared lists of near-fault records considered to have strong
ground motion pulses. The focus of these researchers was on identifying pulses on the FN
components and only a few FP pulse records were identified. From examination of these data
sets, several additional records having FN pulses were identified. In determining the additional
records, we used the criteria that PGV for the records was equal to or greater than 30 cm/sec
(same as Baker’s criterion) and the records had been identified as pulse records in at least two
studies.
Pulse records have been identified in the PEER Ground Motion Database as follows: Sixty
records having pulses for the FN components only; nineteen records having pulses in the FP
components only; and thirty records having pulses in both FN and FP components. Several of
the records originally identified as having pulses are not part of the PEER Ground Motion
Database because they are among the records listed in Table B-2 of Appendix B as being
excluded from the PGMD for various reasons.
Tables 2a and 2b summarize information about individual FN and FP pulse records,

respectively, including NGA record no., earthquake name, year, magnitude, and type of
faulting (mechanism), recording station name, earthquake source-to-site distance, Vs30 of
geologic deposits at recording station sites, and estimates of pulse periods and significant

Technical Report 14
durations for the records. The distribution of the identified pulse records by earthquake
magnitude, type of faulting, and source-to-site distance is shown in Tables 3a and 3b for FN
and FP pulse records, respectively. Tables 3a and 3b indicate that the great majority of pulse
records are located within 20 km from the source, and, of these, most are within 10 km. There
are relatively very few pulse records beyond 30 km.
Estimates of pulse periods shown in Tables 2a and 2b were taken from Baker (2007) and the
website http://www.stanford.edu/~bakerjw/pulse-classification.html except for a few records
added to Baker’s compilation. We compared estimates of FN pulse periods for 28 records
where they had been identified in Baker (2007) and at least in two of other studies (Somerville,
2003, Mavroeidis and Papageorgiou, 2003, Bray and Rodriguez-Marek, 2004, Fu and Menun,
2004). For 23 records, the total range in estimated pulse periods among researchers was a factor
of 1.4 or less and for 15 of these records was within a factor of 1.2. For the remaining 5
records, estimated pulse periods had greater divergence and varied up to a factor of 5.
All of the researchers mentioned above found a trend for pulse period to increase with
magnitude, and this trend is expected based on the physics of fault rupture (Somerville, 2003).
Figure 8 shows the individual record estimates of FN pulse period and the mean correlation
between pulse period and magnitude of Baker (2007). Although the correlation for pulse period
to increase with magnitude is clear, considerable data scatter can also be noted. The standard
deviation of the natural logarithm of pulse period determined from Baker’s regression was
0.55, corresponding to a factor of about 1.7 between the median regression estimates and
median-plus-or-minus one standard deviation estimates. Figure 9 shows mean correlations of
pulse period with magnitude by different investigators. All the correlations show a similar trend
for pulse period to increase with magnitude.
2.4.2 Selecting Records with Pulses within the PGMD

If desired, a user of the PGMD can limit searches of records to those having pulses through
options available on the user interface. Searches can be made for records having FN pulses, FP
pulses, or both FN and FP pulses. Similar to other searches for records in the PGMD, a user can
specify criteria and limits described in Section 2.3.2 in searches for pulse records. Pulse records
can be scaled and ranked for spectral match as described in Section 2.3.
It is thought that the effects of type of faulting on pulse period may be significant for large
magnitude earthquakes, although the effect is not well defined. Therefore it is suggested that

Technical Report 15
for earthquakes greater than magnitude 6.5, records from strike-slip earthquakes not be used for
reverse-slip or normal-slip earthquakes and visa versa. Few pulse records from normal-slip
earthquakes are in the database, and records from reverse-slip earthquakes are suggested to be
used for normal slip earthquakes. In order to obtain a more detailed understanding of the nature
of pulses in time series records considered for analysis, it is suggested that the velocity time
series of candidate time series be displayed and examined. This can be readily done through the
PGMD graphic interface.
3.0 REFERENCES
Abrahamson, N.A., and Silva, W.J., 2008, Summary of the Abrahamson & Silva NGA ground-
motion relations: Earthquake Spectra, vol. 24, no. 1, pp. 67-97.
Akkar, S., Yazgan, U., and Gulkan, P., 2005, Drift estimates in frame buildings subjected to
near-fault ground motions: Journal of Structural Engineering, v. 131, no. 7, pp. 1014-
1024.
Alavi, B., and Krawinkler, H., 2001, Effects of near-fault ground motions on frame structures:
Blume Center Report #138, Stanford, California, 301 pp.
American Society of Civil Engineers (ASCE), 2006, Minimum Design Loads for Buildings and
Other Structures, ASCE/SEI 7-05.
Anderson, J.C., and Bertero, V.V., 1987, Uncertainties in establishing design earthquakes:
Journal of Structural Engineering, v. 113, no. 8, pp. 1709-1724.
Baker, J.W., 2006, Breaking the uniform hazard spectrum into component events: The effect of
epsilon on response spectra and structural response: Preprinted presentation from 2006
COSMOS Technical Session entitled “An Evaluation of Methods for the Selection and
Scaling of Ground Motion Time series for Building Code and Performance-Based
Earthquake Engineering Applications,” sponsored by Consortium of Organizations for
Strong-Motion Observation Systems and Pacific Earthquake Engineering Research
Center, Berkeley, California, November. http://www.cosmos-
eq.org/presentations/TS2006/6-Baker.pdf
Baker, J.W., 2007, Quantitative classification of near-fault ground motions using wavelet
analysis: Bulletin of the Seismological Society of America, v. 97, no. 5, pp. 1486-1501.
Baker, J.W., and Cornell, C.A., 2004, Choice of a vector of ground motion intensity measures
for seismic demand hazard analysis: Proceedings of the 13th World Conference on
Earthquake Engineering, Vancouver, Canada, August.

Technical Report 16
Baker, J.W., and Cornell C.A., 2005, A vector-valued ground motion intensity measure
consisting of spectral acceleration and epsilon: Earthquake Engineering and Structural
Dynamics, v. 34, no. 10, pp. 1193-1217.
Baker, J.W., and Cornell, C.A., 2006, Spectral shape, epsilon and record selection: Earthquake
Engineering and Structural Dynamics, v. 35, no. 9, pp. 1077-1095.
Baker, J.W., and Cornell, C.A., 2008, Vector-valued intensity measures incorporating spectral
shape for prediction of structural response: Journal of Earthquake Engineering, v. 12,
Iss 4, pp. 534-554.Baker, J.W., and Jayaram, N., 2008, Correlation of spectral
acceleration values from NGA ground motion models: Earthquake Spectra, vol. 24, no.
1, pp. 299-317.
Bazzurro, P., and Luco, N. 2003, Parameterization of non-stationary time series: Draft Report
for Task 1G00 to PEER-Lifelines Program.
Bertero, V., Mahin S., and Herrera R., 1978, Aseismic design implications of near-fault San
Fernando earthquake records: Earthquake Engineering and Structural Dynamics, v. 6,
no. 1, pp. 31-42.
Boore, D.M., and Atkinson, G.M., 2008, Ground-motion prediction equations for the average
horizontal component of PGA, PGV, and 5% damped PSA at spectral periods between
0.01s and 10.0s: Earthquake Spectra, vol. 24, no. 1, pp. 99-138.
Bray, J.D., and Rodriguez-Marek, A., 2004, Characterization of forward-directivity ground

motions in the near-fault region: Soil Dynamics and Earthquake Engineering, v. 24, no.
11, pp. 815-828.
Building Seismic Safety Council (BSSC), 2003, NEHRP Recommended Provisions for Seismic
Regulations for New Buildings and Other Structures (FEMA 450), 2003 Edition.
Campbell, K.W., and Bozorgnia, Y., 2008, NGA ground motion Model for the geometric mean
horizontal component of PGA, PGV, PGD, and 5% damped linear elastic response
spectra for periods ranging from 0.01s to 10.0s: Earthquake Spectra, vol. 24, no. 1, pp.
139-171.
Chiou, B., Power, M., Abrahamson, N., and Roblee, C., 2006, An overview of the project of
next generation of ground motion attenuation models for shallow crustal earthquakes in
active tectonic regions: Proceedings of the Fifth National Seismic Conference on
Bridges & Highways, San Mateo, California, September.
Chiou, B., Darragh, R., Gregor, N., and Silva, W., 2008, NGA project strong-motion database,
Earthquake Spectra, vol. 24, no. 1, pp. 23-44.
Chiou, B.S.J. and Youngs, R.R., 2008a, Chiou-Youngs NGA ground motion relations for the
geometric mean horizontal component of peak and spectral ground motion parameters,
Earthquake Spectra, vol. 24, no. 1, pp. 173-215.

Technical Report 17
Chiou, B.S.J., and Youngs, R.R., 2008b, NGS model for average horizontal component of peak
ground motion and response spectra, Report PEER 2008/09, Pacific Earthquake
Engineering Research Center, University of California, Berkeley.
Consortium of Organizations of Strong Motion Systems (COSMOS), 2006, Annual Meeting

Technical Session, An Evaluation of Methods for the Selection and Scaling of Ground
Motion Time series for Building Code and Performance-Based Earthquake Engineering
Applications, sponsored by COSMOS and Pacific Earthquake Engineering Research
Center (PEER) Berkeley, CA, November.
Consortium of Organizations of Strong Motion Observation Systems (COSMOS), 2007,

Annual Meeting Technical Session, Recommendations for Selection and Scaling of
Ground Motion for Building Code and Performance-Based Earthquake Engineering
(PBEE) Applications and on Evaluation of Selection and Scaling Procedures Proposed
for the Next Generation of PBEE, sponsored by COSMOS and Pacific Earthquake
Engineering Research Center (PEER) Berkeley, CA, November.
Cordova, P.P., Mehanny, S.S.F., Deierlein, G.G., and Cornell, C.A., 2000, Development of a
two-parameter seismic intensity measure and probabilistic assessment procedure:
Proceedings of the Second U.S.-Japan Workshop on Performance-Based Seismic
Earthquake Engineering Methodology for Reinforced Concrete Building Structures,
Sapporo, Japan, pp. 195-214.
Cornell, C.A., 2006, Should uniform hazard spectra be used for design? How should design
response spectra be determined considering uniform hazard spectra from design maps
and hazard deaggregation data?: Abstract and PowerPoint presentation in Preprints,
Third ATC-35/USGS National Earthquake Ground Motion Mapping Workshop, San
Mateo, California, December.
Fu, Q., and Menun, C., 2004, Seismic-environment-based simulation of near-fault ground
motions: Proceedings of the 13th World Conference on Earthquake Engineering,
Vancouver, Canada, 15 pp.
GMSM Working Group, 2009, C.B. Haselton (ed.), PEER Ground Motion Selection and
Modification Working Group, Evaluation of ground motion selection and modification
methods: predicting median interstory drift response of buildings: PEER Report
2009/01, Pacific Earthquake Engineering Research Center, University of California,
Berkeley.
Goulet, C.A., Watson-Lamprey, J., Baker, J., Haselton, C., and Luco, N., 2008, Assessment of
ground motion selection and modification (GMSM) methods for non-linear dynamic
analyses of structures, Proceedings of Geotechnical Earthquake Engineering and Soil
Dynamics IV Conference, Sacramento, CA, American Society of Civil Engineers,
Geotechnical Special Publication No. 181, 10 pp.
Hall, J.F., Heaton, T.H., Halling, M.W., and Wald, D.J., 1995, Near-source ground motion and
its effects on flexible buildings: Earthquake Spectra, v. 11, no. 4, pp. 569-605.
Technical Report 18
Idriss, I.M., 2008, An NGA empirical model for estimating the horizontal spectral values
generated by shallow crustal earthquakes: Earthquake Spectra, vol. 24, no. 1,
pp. 217-242.
International Code Council (ICC), 2006, International Building Code 2006.
Iwan, W.D., 1997, Drift spectrum: measure of demand for earthquake ground motions: Journal
of Structural Engineering, v. 123, no. 4, pp. 397-404.
Jalayer, F., 2003, Direct probabilistic seismic analysis: Implementing non-linear dynamic
assessments: Ph.D. thesis, Department of Civil and Environmental Engineering,
Stanford University.
Kempton, J.J., and Stewart, J.P., 2006, Prediction equations for significant duration of
earthquake ground motions considering site and near-source effects: Earthquake
Spectra, v. 22, no. 4, pp. 985-1013.
Krawinkler, H., and Alavi, B., 1998, Development of improved design procedures for near-
fault ground motions: Proceedings of the SMIP98 Seminar on Utilization of Strong-
Motion Data, Oakland, California Strong Motion Instrumentation Program, California
Geological Survey, pp. 21-41.
Luco, N., and Bazzurro, P., 2004, Effects of ground motion scaling on nonlinear structural
response: Presentation at PEER Annual Meeting, Palm Springs, California, February.
Luco, N., and Cornell, C.A., 2006, Structure-specific scalar intensity measures for near-source
and ordinary earthquake ground motions: Earthquake Spectra, v. 23, no. 2.
Makris, N., and Black, C., 2003, Dimensional analysis of inelastic structures subjected to near
fault ground motions: Earthquake Engineering Research Center, EERC 2003-05,
Berkeley, California, 96 pp.
Mavroeidis, G.P., Dong, G., and Papageorgiou, A.S., 2004, Near-fault ground motions, and the
response of elastic and inelastic single-degree-of-freedom (SDOF) systems: Earthquake
Engineering and Structural Dynamics, v. 33, no. 9, pp. 1023-1049.
Mavroeidis, G.P., and Papageorgiou, A.S., 2003, A mathematical representation of near-fault

ground motions: Bulletin of the Seismological Society of America, v. 93 no. 3, pp.
1099-1131.
Menun, C., and Fu, Q., 2002, An analytical model for near-fault ground motions and the
response of SDOF systems: Proceedings of the Seventh U.S. National Conference on
Earthquake Engineering, Boston, Massachusetts, 10 pp.
Power, M., Chiou, B., Abrahamson, N., Bozorgnia, Y., Schantz, T., and Roblee, C., 2008, An
overview of the NGA project: Earthquake Spectra, vol. 24, no. 1, pp. 3-21.

Technical Report 19
Shome, N., Cornell, C.A., Bazzurro, P., and Carballo, J.E., 1998, Earthquake records and
nonlinear response: Earthquake Spectra, v. 14, no. 3, pp. 469-500.
Somerville, P.G., 2003, Magnitude scaling of the near fault rupture directivity pulse: Physics of
the Earth and Planetary Interiors, v. 137, no. 1, p. 12.
Somerville, P.G., Smith, N.F., Graves, R.W., and Abrahamson, N.A., 1997, Modification of
empirical strong ground motion attenuation relations to include the amplitude and
duration effects of rupture directivity: Seismological Research Letters, v. 68, no. 1,
pp. 199-222.

Technical Report 20
TABLES
____________________________________________________________________________

Technical Report – Tables
TABLE 1
SIGNIFICANT DURATION OF SHAKING BASED ON KEMPTON AND STEWART (2006)
R Vs30 Median Duration (sec)*

(km) (m/sec) M=5 M = 5.5 M=6 M = 6.5 M=7 M = 7.5 M=8
760 1.6 2.6 4.1 6.4 10.0 15.5 24.1
0 520 2.4 3.3 4.8 7.1 10.7 16.2 24.8
250 3.2 4.2 5.6 7.9 11.5 17.0 25.6
760 3.2 4.3 6.0 8.7 12.9 19.3 29.3
10 520 4.1 5.2 6.9 9.6 13.7 20.1 30.1
250 5.0 6.1 7.8 10.5 14.7 21.1 31.1
760 5.2 6.5 8.5 11.6 16.4 23.9 35.5
20 520 6.2 7.5 9.5 12.6 17.4 24.9 36.5
250 7.3 8.6 10.6 13.7 18.5 26.0 37.6
760 9.7 11.0 13.0 16.1 20.9 28.4 40.0
50 520 10.7 12.0 14.0 17.1 21.9 29.4 41.0
250 11.8 13.1 15.1 18.2 23.0 30.5 42.1
760 17.2 18.5 20.5 23.6 28.4 35.9 47.5
100 520 18.2 19.5 21.5 24.6 29.4 36.9 48.5
250 19.3 20.6 22.6 25.7 30.5 38.0 49.6
760 32.2 33.5 35.5 38.6 43.4 50.9 62.5
200 520 33.2 34.5 36.5 39.6 44.4 51.9 63.5
250 34.3 35.6 37.6 40.7 45.5 53.0 64.6
* Significant duration is defined as the time required to build up from 5% to 95% of the Arias Intensity of an acceleration time history.

TABLE 2a
FAULT-NORMAL PULSELIKEGROUND MOTION RECORDS IN PGMD

Rjb* Rrup* Vs30** Duration
NGA# Comp. Event Year Station Tp Mag Mechanism
(km) (km) (m/s) (sec)
77 FN San Fernando 1971 Pacoima Dam (upper left abut) 1.6 6.6 Reverse 0.0 1.8 2016 7.1
150 FN Coyote Lake 1979 Gilroy Array #6 1.2 5.7 Strike-Slip 0.4 3.1 663 3.4
158 FN Imperial Valley-06 1979 Aeropuerto Mexicali 2.4 6.5 Strike-Slip 0.0 0.3 275 7.1
159 FN Imperial Valley-06 1979 Agrarias 2.3 6.5 Strike-Slip 0.0 0.7 275 11.5
161 FN Imperial Valley-06 1979 Brawley Airport 4.0 6.5 Strike-Slip 8.5 10.4 209 15.2
170 FN Imperial Valley-06 1979 EC County Center FF 4.5 6.5 Strike-Slip 7.3 7.3 192 14.9
171 FN Imperial Valley-06 1979 EC Meloland Overpass FF 3.3 6.5 Strike-Slip 0.1 0.1 186 6.2
173 FN Imperial Valley-06 1979 El Centro Array #10 4.5 6.5 Strike-Slip 6.2 6.2 203 13.0
184 FN Imperial Valley-06 1979 El Centro Differential Array 5.9 6.5 Strike-Slip 5.1 5.1 202 6.9
185 FN Imperial Valley-06 1979 Holtville Post Office 4.8 6.5 Strike-Slip 5.5 7.7 203 11.8
250 FN Mammoth Lakes-06 1980 Long Valley Dam (Upr L Abut) 1.1 5.9 Strike-Slip 9.3 16.2 345 7.2
292 FN Irpinia, Italy-01 1980 Sturno 3.1 6.9 Normal 6.8 10.8 1000 16.7
316 FN Westmorland 1981 Parachute Test Site 3.6 5.9 Strike-Slip 16.5 16.7 349 17.3
407 FN Coalinga-05 1983 Oil City 0.7 5.8 Reverse 2.4 8.5 376 2.8
415 FN Coalinga-05 1983 Transmitter Hill 0.9 5.8 Reverse 4.1 9.5 376 3.9
418 FN Coalinga-07 1983 Coalinga-14th & Elm (Old CHP) 0.4 5.2 Reverse 7.6 11.0 339 0.7
451 FN Morgan Hill 1984 Coyote Lake Dam (SW Abut) 1.0 6.2 Strike-Slip 0.2 0.5 597 3.1
459 FN Morgan Hill 1984 Gilroy Array #6 1.2 6.2 Strike-Slip 9.9 9.9 663 6.9
529 FN N. Palm Springs 1986 North Palm Springs 1.4 6.1 Reverse-Oblique 0.0 4.0 345 4.5
568 FN San Salvador 1986 Geotech Investig Center 0.9 5.8 Strike-Slip 2.1 6.3 545 3.8
615 FN Whittier Narrows-01 1987 Downey - Co Maint Bldg 0.8 6.0 Reverse-Oblique 15.0 20.8 272 8.3
645 FN Whittier Narrows-01 1987 LB - Orange Ave 1.0 6.0 Reverse-Oblique 19.8 24.5 270 8.3
721 FN Superstition Hills-02 1987 El Centro Imp. Co. Cent 2.4 6.5 Strike-Slip 18.2 18.2 192 18.8
723 FN Superstition Hills-02 1987 Parachute Test Site 2.3 6.5 Strike-Slip 1.0 1.0 349 10.5
738 FN Loma Prieta 1989 Alameda Naval Air Stn Hanger 2.0 6.9 Reverse-Oblique 70.9 71.0 190 6.0
763 FN Loma Prieta 1989 Gilroy - Gavilan Coll. 1.8 6.9 Reverse-Oblique 9.2 10.0 730 5.2
765 FN Loma Prieta 1989 Gilroy Array #1 1.2 6.9 Reverse-Oblique 8.8 9.6 1428 5.1
779 FN Loma Prieta 1989 LGPC 3.0 6.9 Reverse-Oblique 0.0 3.9 478 10.0
783 FN Loma Prieta 1989 Oakland - Outer Harbor Wharf 1.8 6.9 Reverse-Oblique 74.2 74.3 249 6.0
802 FN Loma Prieta 1989 Saratoga - Aloha Ave 4.5 6.9 Reverse-Oblique 7.6 8.5 371 8.4
803 FN Loma Prieta 1989 Saratoga - W Valley Coll. 1.9 6.9 Reverse-Oblique 8.5 9.3 371 11.0
821 FN Erzican, Turkey 1992 Erzincan 2.7 6.7 Strike-Slip 0.0 4.4 275 6.9
828 FN Cape Mendocino 1992 Petrolia 3.0 7.0 Reverse 0.0 8.2 713 16.2
838 FN Landers 1992 Barstow 8.9 7.3 Strike-Slip 34.9 34.9 371 17.4
879 FN Landers 1992 Lucerne 5.1 7.3 Strike-Slip 2.2 2.2 685 12.9
900 FN Landers 1992 Yermo Fire Station 7.5 7.3 Strike-Slip 23.6 23.6 354 17.2
1013 FN Northridge-01 1994 LA Dam 1.7 6.7 Reverse 0.0 5.9 629 6.5
1044 FN Northridge-01 1994 Newhall - Fire Sta 2.2 6.7 Reverse 3.2 5.9 269 5.5
1045 FN Northridge-01 1994 Newhall - W Pico Canyon Rd. 2.4 6.7 Reverse 2.1 5.5 286 7.1
1050 FN Northridge-01 1994 Pacoima Dam (downstr) 0.5 6.7 Reverse 4.9 7.0 2016 3.8
1051 FN Northridge-01 1994 Pacoima Dam (upper left) 0.9 6.7 Reverse 4.9 7.0 2016 6.0
1063 FN Northridge-01 1994 Rinaldi Receiving Sta 1.2 6.7 Reverse 0.0 6.5 282 7.1
1084 FN Northridge-01 1994 Sylmar - Converter Sta 3.5 6.7 Reverse 0.0 5.4 251 13.5
1085 FN Northridge-01 1994 Sylmar - Converter Sta East 3.5 6.7 Reverse 0.0 5.2 371 7.2
1086 FN Northridge-01 1994 Sylmar - Olive View Med FF 3.1 6.7 Reverse 1.7 5.3 441 5.8
1106 FN Kobe, Japan 1995 KJMA 1.0 6.9 Strike-Slip 0.9 1.0 312 9.6
1119 FN Kobe, Japan 1995 Takarazuka 1.4 6.9 Strike-Slip 0.0 0.3 312 5.1
1120 FN Kobe, Japan 1995 Takatori 1.6 6.9 Strike-Slip 1.5 1.5 256 10.8
1176 FN Kocaeli, Turkey 1999 Yarimca 4.5 7.5 Strike-Slip 1.4 4.8 297 15.4
1202 FN Chi-Chi, Taiwan 1999 CHY035 1.4 7.6 Reverse-Oblique 12.6 12.7 [555] 28.1
1244 FN Chi-Chi, Taiwan 1999 CHY101 4.8 7.6 Reverse-Oblique 10.0 10.0 259 29.1
1476 FN Chi-Chi, Taiwan 1999 TCU029 6.4 7.6 Reverse-Oblique 28.1 28.1 [426] 19.2
1477 FN Chi-Chi, Taiwan 1999 TCU031 6.2 7.6 Reverse-Oblique 30.2 30.2 489 24.1
1853 FN Yountville 2000 Napa Fire Station #3 0.7 5.0 Strike-Slip 8.4 11.5 271 3.3
2457 FN Chi-Chi, Taiwan-03 1999 CHY024 3.2 6.2 Reverse 18.5 19.7 428 8.6
2495 FN Chi-Chi, Taiwan-03 1999 CHY080 1.4 6.2 Reverse 21.3 22.4 [680] 2.9
2627 FN Chi-Chi, Taiwan-03 1999 TCU076 0.9 6.2 Reverse 13.0 14.7 615 3.0
3317 FN Chi-Chi, Taiwan-06 1999 CHY101 2.8 6.3 Reverse 34.6 36.0 259 18.4
Notes:
* Joyner-Boore distance (Rjb) and closest distance (Rrup) for earthquakes not having fault rupture models are shown in red;
distances were estimated using epicentral and hypocentral distances and simulations (Chiou and Youngs, 2008b).
** Updated preferred Vs30 values for CWB Taiwan sites are shown in brackets; values were estimated by B. Chiou (2009, personal
communication).
TABLE 2b
FAULT-PARALLEL PULSELIKEGROUND MOTION RECORDS IN PGMD

Rjb Rrup Vs30** Duration
NGA# Comp. Event Year Station Tp Mag Mechanism
(km) (km) (m/s) (sec)
173 FP Imperial Valley-06 1979 El Centro Array #10 2.0 6.5 Strike-Slip 6.2 6.2 203 11.9
184 FP Imperial Valley-06 1979 El Centro Differential Array 2.0 6.5 Strike-Slip 5.1 5.1 202 6.4
185 FP Imperial Valley-06 1979 Holtville Post Office 3.6 6.5 Strike-Slip 5.5 7.7 203 12.7
292 FP Irpinia, Italy-01 1980 Sturno 3.5 6.9 Normal 6.8 10.8 1000 12.1
316 FP Westmorland 1981 Parachute Test Site 4.2 5.9 Strike-Slip 16.5 16.7 349 15.5
319 FP Westmorland 1981 Westmorland Fire Sta 1.4 5.9 Strike-Slip 6.2 6.5 194 6.1
451 FP Morgan Hill 1984 Coyote Lake Dam (SW Abut) 1.1 6.2 Strike-Slip 0.2 0.5 597 5.4
496 FP Nahanni, Canada 1985 Site 2 0.8 6.8 Reverse 0.0 4.9 660 9.1
568 FP San Salvador 1986 Geotech Investig Center 1.8 5.8 Strike-Slip 2.1 6.3 545 4.2
569 FP San Salvador 1986 National Geografical Inst 1.0 5.8 Strike-Slip 3.7 7.0 350 4.6
722 FP Superstition Hills-02 1987 Kornbloom Road (temp) 2.1 6.5 Strike-Slip 18.5 18.5 207 13.4
738 FP Loma Prieta 1989 Alameda Naval Air Stn Hanger 2.3 6.9 Reverse-Oblique 70.9 71.0 190 4.6
764 FP Loma Prieta 1989 Gilroy - Historic Bldg. 1.8 6.9 Reverse-Oblique 10.3 11.0 339 9.8
767 FP Loma Prieta 1989 Gilroy Array #3 3.0 6.9 Reverse-Oblique 12.2 12.8 350 8.9
779 FP Loma Prieta 1989 LGPC 4.1 6.9 Reverse-Oblique 0.0 3.9 478 10.9
784 FP Loma Prieta 1989 Oakland - Title & Trust 1.7 6.9 Reverse-Oblique 72.1 72.2 306 13.3
803 FP Loma Prieta 1989 Saratoga - W Valley Coll. 5.0 6.9 Reverse-Oblique 8.5 9.3 371 12.4
821 FP Erzican, Turkey 1992 Erzincan 2.2 6.7 Strike-Slip 0.0 4.4 275 10.0
825 FP Cape Mendocino 1992 Cape Mendocino 4.9 7.0 Reverse 0.0 7.0 514 6.5
828 FP Cape Mendocino 1992 Petrolia 1.0 7.0 Reverse 0.0 8.2 713 17.3
1013 FP Northridge-01 1994 LA Dam 2.8 6.7 Reverse 0.0 5.9 629 6.5
1045 FP Northridge-01 1994 Newhall - W Pico Canyon Rd. 2.2 6.7 Reverse 2.1 5.5 286 9.0
1063 FP Northridge-01 1994 Rinaldi Receiving Sta 3.0 6.7 Reverse 0.0 6.5 282 10.1
1176 FP Kocaeli, Turkey 1999 Yarimca 4.6 7.5 Strike-Slip 1.4 4.8 297 14.9
1193 FP Chi-Chi, Taiwan 1999 CHY024 6.2 7.6 Reverse-Oblique 9.6 9.6 428 26.9
1463 FP Chi-Chi, Taiwan 1999 TCU003 11.0 7.6 Reverse-Oblique 86.6 86.6 517 41.3
1468 FP Chi-Chi, Taiwan 1999 TCU010 11.0 7.6 Reverse-Oblique 82.2 82.3 [484] 37.6
1605 FP Duzce, Turkey 1999 Duzce 5.6 7.1 Strike-Slip 0.0 6.6 276 10.7
3475 FP Chi-Chi, Taiwan-06 1999 TCU080 1.0 6.3 Reverse 0.0 10.2 [509] 6.8
Notes:
** Updated preferred Vs30 values for CWB Taiwan sites are shown in brackets; values were estimated by B. Chiou (2009, personal
communication).

TABLE 3
MAGNITUDE-DISTANCE DISTRIBUTION IN PULSE-LIKE RECORDS
(3a) FAULT-NORMAL
FN Rrup 0~10 km 10~20 km 20~30 km 30~50 km > 50 km

sum
Mag SS NOR REV SS NOR REV SS NOR REV SS NOR REV SS NOR REV
5~6 2 0 2 3 0 1 0 0 2 0 0 0 0 0 0 10
6~7 19 0 16 4 1 4 0 0 1 0 0 1 0 0 2 48
7~8 2 0 15 0 0 6 1 0 4 1 0 3 0 0 0 32
sum 56 19 8 5 2 90
(3b) FAULT-PARALLEL
FP Rrup 0~10 km 10~20 km 20~30 km 30~50 km > 50 km

sum
Mag SS NOR REV SS NOR REV SS NOR REV SS NOR REV SS NOR REV
5~6 3 0 0 1 0 0 0 0 0 0 0 0 0 0 0 4
6~7 7 0 6 2 1 3 0 0 0 0 0 0 0 0 2 21
7~8 2 0 7 0 0 6 0 0 2 0 0 2 0 0 5 24
sum 25 13 2 2 7 49

FIGURES
___________________________________________________________________________

Technical Report – Figures
a) Illustration of conditional mean spectrum (from Baker 2006)
b) Steps in calculating conditional mean spectrum (modified from Baker 2006)
Figure 1
Illustration and calculation of a conditional mean spectrum

Figure 2
Magnitude and distance distribution for PEER NGA records in DGML database

Figure 3
Code response spectrum

Note:
Response spectra are constructed at 1ε (one-standard-deviation) level for this example.
Figure 4
Example of individual and average NGA response spectra from five models constructed by PGMD software
tool.

Epsilon=0
Conditional Mean
Spectrum, Epsilon=1,
T_eps=0.5 sec
Epsilon=1
Note:
The lower and upper red curves are median and 1ε NGA response spectra (average of
five NGA models). Black curve is conditional mean response spectrum -- conditional on
spectral acceleration at 0.5 second period being at the 1ε level for this example.
Figure 5
Example of conditional mean NGA response spectrum constructed by PGMD software tool.

Figure 6
Illustration of construction of conditional mean spectra for different periods
for multiple scenario earthquakes (based on Cornell 2006)

Figure 7
Example of record search performed by PGMD software tool

Figure 8
Correlation of pulse period with earthquake magnitude by Baker 2007

Figure 9
Correlations of pulse period with earthquake magnitude by different researchers


PEER Ground Motion Database - Technical Report PDF

Uploaded by

Copyright:

Available Formats

PEER Ground Motion Database - Technical Report PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

PEER Ground Motion Database - Technical Report PDF

Uploaded by

Copyright:

Available Formats

With special thanks to:

Technical Report for the

PEER Ground Motion Database – Beta Version

Table 1 Significant Duration of Shaking Based on Kempton and Stewart (2006)

Figure 1 Illustration and Calculation of a Conditional Mean Spectrum

Figure 8 Correlation of Pulse Period with Earthquake Magnitude by Baker 2007

Appendix A PGMD Users Manual

PEER Ground Motion Database – Beta Version

1.1 OBJECTIVE, SPONSORSHIP, AND LIMITS

1.2 EVOLUTION OF PEER GROUND MOTION DATABASE CONCEPT

(4) Featured videos describing the product features and applications.

PEER Ground Motion Database – Beta Version

2.0 PGMD CAPABILITIES, FEATURES, AND OPERATION

PEER Ground Motion Database – Beta Version

PEER Ground Motion Database – Beta Version

2.3.1 Step 1 – Developing the Target Spectrum

Option 3 – Spectrum based on PEER Next Generation Attenuation (NGA) Relationships.

PEER Ground Motion Database – Beta Version

PEER Ground Motion Database – Beta Version

 w (Ti ) {ln [SA target (Ti )]  ln [ f  SA record (Ti )]}2

PEER Ground Motion Database – Beta Version

SA GM  SA FN  SA FP or ln SA GM  (ln SA FN  ln SA FP )/2 (4)

PEER Ground Motion Database – Beta Version

Table 1 summarizes median values of significant duration as defined in Section 2.2 as a

PEER Ground Motion Database – Beta Version

2.4 SUPPLEMENTAL INFORMATION ABOUT FORMING RECORDS SETS CONTAINING

2.4.1 Database for Records with Pulses

PEER Ground Motion Database – Beta Version

Tables 2a and 2b summarize information about individual FN and FP pulse records,

PEER Ground Motion Database – Beta Version

2.4.2 Selecting Records with Pulses within the PGMD

PEER Ground Motion Database – Beta Version

PEER Ground Motion Database – Beta Version

Bray, J.D., and Rodriguez-Marek, A., 2004, Characterization of forward-directivity ground

PEER Ground Motion Database – Beta Version

Consortium of Organizations of Strong Motion Systems (COSMOS), 2006, Annual Meeting

Consortium of Organizations of Strong Motion Observation Systems (COSMOS), 2007,

International Code Council (ICC), 2006, International Building Code 2006.

Mavroeidis, G.P., and Papageorgiou, A.S., 2003, A mathematical representation of near-fault

PEER Ground Motion Database – Beta Version

PEER Ground Motion Database – Beta Version

PEER Ground Motion Database – Beta Version

SIGNIFICANT DURATION OF SHAKING BASED ON KEMPTON AND STEWART (2006)

R Vs30 Median Duration (sec)*

PEER Ground Motion Database – Beta Version

FAULT-NORMAL PULSELIKEGROUND MOTION RECORDS IN PGMD

FAULT-PARALLEL PULSELIKEGROUND MOTION RECORDS IN PGMD

PEER Ground Motion Database – Beta Version

MAGNITUDE-DISTANCE DISTRIBUTION IN PULSE-LIKE RECORDS

FN Rrup 0~10 km 10~20 km 20~30 km 30~50 km > 50 km

FP Rrup 0~10 km 10~20 km 20~30 km 30~50 km > 50 km

PEER Ground Motion Database – Beta Version

PEER Ground Motion Database – Beta Version

b) Steps in calculating conditional mean spectrum (modified from Baker 2006)

PEER Ground Motion Database – Beta Version

PEER Ground Motion Database – Beta Version

PEER Ground Motion Database – Beta Version

PEER Ground Motion Database – Beta Version

PEER Ground Motion Database – Beta Version