1 

CHAPTER C21 SITE-SPECIFIC GROUND MOTION PROCEDURES FOR SEISMIC

2  DESIGN

3  C21.0 GENERAL

4  Site-specific procedures for computing earthquake ground motions include dynamic site
5  response analyses and probabilistic and deterministic seismic hazard analyses (PSHA and
6  DSHA), which may include dynamic site response analysis as part of the calculation. Use of site-

LY
7  specific procedures may be required in lieu of the general procedure in Sections 11.4.2 through
8  11.4.7; Section C11.4.8 in ASCE 7-16 explains the conditions under which the use of these

N
9  procedures is required. Such studies must be comprehensive and must incorporate current

O
10  scientific interpretations. Because there is typically more than one scientifically credible
11  alternative for models and parameter values used to characterize seismic sources and ground
12  n
motions, it is important to formally incorporate these uncertainties in a site-specific analysis. For
io
13  example, uncertainties may exist in seismic source location, extent, and geometry; maximum
at
14  earthquake magnitude; earthquake recurrence rate; ground motion attenuation; local site
15  conditions, including soil layering and dynamic soil properties; and possible two- or three-
m

16  dimensional wave-propagation effects. The use of peer review for a site-specific ground motion
or

17  analysis is encouraged.

nf

18  Site-specific ground motion analysis can consist of one of the following approaches: (a) PSHA
and possibly DSHA if the site is near an active fault, (b) PSHA/DSHA followed by dynamic site
rI

19 
20  response analysis, and (c) dynamic site response analysis only. The first approach is used to
Fo

21  compute ground motions for bedrock or stiff soil conditions (not softer than Site Class D). In this
22  approach, if the site consists of stiff soil overlying bedrock, for example, the analyst has the
23  option of either (a) computing the bedrock motion from the PSHA/DSHA and then using the site

24  coefficient ( Fa and Fv ) tables in Section 11.4.3 to adjust for the stiff soil overburden or (b)

25  computing the response spectrum at the ground surface directly from the PSHA/DSHA. The
26  latter requires the use of attenuation equations for computing stiff soil-site response spectra
27  (instead of bedrock response spectra).

1 
 
 1  The second approach is used where softer soils overlie the bedrock or stiff soils. The third
2  approach assumes that a site-specific PSHA/DSHA is not necessary but that a dynamic site
3  response analysis should, or must, be performed. This analysis requires the definition of an
4  outcrop ground motion, which can be based on the 5% damped response spectrum computed
5  from the PSHA/DSHA, or obtained from the general procedure in Section 11.4. A representative
6  set of acceleration time histories is selected and scaled to be compatible with this outcrop
7  spectrum. Dynamic site response analyses, using these acceleration histories as input, are used to
8  compute motions at the ground surface. The response spectra of these surface motions are used

LY
9  to define a maximum considered earthquake (MCE) ground motion response spectrum.

10  The approaches described in the aforementioned have advantages and disadvantages. In many

N
11  cases, user preference governs the selection, but geotechnical conditions at the site may dictate

O
12  the use of one approach over the other. If bedrock is at a depth much greater than the extent of
13  the site geotechnical investigations, the direct approach of computing the ground surface motion
14  n
in the PSHA/DSHA may be more reasonable. On the other hand, if bedrock is shallow and a
io
15  large impedance contrast exists between it and the overlying soil (i.e., density times shear wave
at
16  velocity of bedrock is much greater than that of the soil), the two-step approach might be more
17  appropriate.
m

18  Use of peak ground acceleration as the anchor for a generalized site-dependent response
or

19  spectrum is discouraged because sufficiently robust ground motion attenuation relations are
nf

20  available for computing response spectra in western United States and eastern United States
tectonic environments.
rI

21 
Fo

22  C21.1 SITE RESPONSE ANALYSIS

23  C21.1.1 Base Ground Motions

24  Ground motion acceleration histories that are representative of horizontal rock motions at the site
25  are required as input to the soil model. Where a site-specific ground motion hazard analysis is
26  not performed, the MCE response spectrum for Site Class B (rock) is defined using the general
27  procedure described in Section 11.4.1. If the model is terminated in material of Site Class A, C,
28  or D, the input MCE response spectrum is adjusted in accordance with Section 11.4.3. The US

2 
 
 1  Geological Survey (USGS) national seismic hazard mapping project website
2  (www.earthquake.usgs.gov/hazards/products/conterminous) includes hazard deaggregation
3  options that can be used to evaluate the predominant types of earthquake sources, magnitudes,
4  and distances contributing to the probabilistic ground motion hazard. Sources of recorded
5  acceleration time histories include the databases of the Consortium of Organizations for Strong
6  Motion Observation Systems (COSMOS) Virtual Data Center website (www.cosmos-eq.org),
7  the Pacific Earthquake Engineering Research (PEER) Center Strong Motion Database website
8  (peer.berkeley.edu/products/strong_ground_motion_db.html), and the US National Center for

LY
9  Engineering Strong Motion Data (NCESMD) website (http://www.strongmotioncenter.org).
10  Ground motion acceleration histories at these sites generally were recorded at the ground surface

N
11  and hence apply for an outcropping condition and should be specified as such in the input to the
12  site response analysis code (Kwok et al. 2007 have additional details).

O
13  C21.1.2 Site Condition Modeling
n
io
14  Modeling criteria are established by site-specific geotechnical investigations that should include
at
15  (a) borings with sampling; (b) standard penetration tests (SPTs), cone penetrometer tests (CPTs),
16  and/or other subsurface investigative techniques; and (c) laboratory testing to establish the soil
m

17  types, properties, and layering. The depth to rock or stiff soil material should be established from
or

18  these investigations. Investigation should extend to bedrock or, for very deep soil profiles, to
19  material in which the model is terminated. Although it is preferable to measure shear wave
nf

20  velocities in all soil layers, it is also possible to estimate shear wave velocities based on
rI

21  measurements available for similar soils in the local area or through correlations with soil types
22  and properties. A number of such correlations are summarized by Kramer (1996).
Fo

23  Typically, a one-dimensional soil column extending from the ground surface to bedrock is
24  adequate to capture first-order site response characteristics. For very deep soils, the model of the
25  soil columns may extend to very stiff or very dense soils at depth in the column. Two- or three-
26  dimensional models should be considered for critical projects when two- or three-dimensional
27  wave propagation effects may be significant (e.g., sloping ground sites). The soil layers in a one-
28  dimensional model are characterized by their total unit weights and shear wave velocities from
29  which low-strain (maximum) shear moduli may be obtained and by relationships defining the

3 
 
 1  nonlinear shear stress–strain behavior of the soils. The required relationships for analysis are
2  often in the form of curves that describe the variation of soil shear modulus with shear strain
3  (modulus reduction curves) and by curves that describe the variation of soil damping with shear
4  strain (damping curves). In a two- or three-dimensional model, compression wave velocities or
5  moduli or Poisson ratios are also required. In an analysis to estimate the effects of liquefaction
6  on soil site response, the nonlinear soil model must also incorporate the buildup of soil pore
7  water pressures and the consequent reductions of soil stiffness and strength. Typically, modulus
8  reduction curves and damping curves are selected on the basis of published relationships for

LY
9  similar soils (e.g., Vucetic and Dobry 1991, Electric Power Research Institute 1993, Darendeli
10  2001, Menq 2003, Zhang et al. 2005). Site-specific laboratory dynamic tests on soil samples to

N
11  establish nonlinear soil characteristics can be considered where published relationships are
12  judged to be inadequate for the types of soils present at the site. Shear and compression wave

O
13  velocities and associated maximum moduli should be selected based on field tests to determine
14 
n
these parameters or, if such tests are not possible, on published relationships and experience for
io
15  similar soils in the local area. The uncertainty in the selected maximum shear moduli, modulus
16  reduction and damping curves, and other soil properties should be estimated (Darendeli 2001,
at

17  Zhang et al. 2008). Consideration of the ranges of stiffness prescribed in Section 12.13.3 is
m

18  recommended.
or

19  C21.1.3 Site Response Analysis and Computed Results

nf

20  Analytical methods may be equivalently linear or nonlinear. Frequently used computer programs
rI

21  for one-dimensional analysis include the equivalent linear program SHAKE (Schnabel et al.
22  1972, Idriss and Sun 1992) and the nonlinear programs FLAC (Itasca 2005), DESRA-2 (Lee and
Fo

23  Finn 1978), MARDES (Chang et al. 1991), SUMDES (Li et al. 1992), D-MOD_2 (Matasovic
24  2006), DEEPSOIL (Hashash and Park 2001), TESS (Pyke 2000), and OpenSees (Ragheb 1994,
25  Parra 1996, Yang 2000). If the soil response induces large strains in the soil (such as for high
26  acceleration levels and soft soils), nonlinear programs may be preferable to equivalent linear
27  programs. For analysis of liquefaction effects on site response, computer programs that
28  incorporate pore water pressure development (effective stress analyses) should be used (e.g.,
29  FLAC, DESRA-2, SUMDES, D-MOD_2, TESS, DEEPSOIL, OpenSees). Response spectra of
30  output motions at the ground surface are calculated as the ratios of response spectra of ground

4 
 
 1  surface motions to input outcropping rock motions. Typically, an average of the response
2  spectral ratio curves is obtained and multiplied by the input MCE response spectrum to obtain
3  the MCE ground surface response spectrum. Alternatively, the results of site response analyses
4  can be used as part of the PSHA using procedures described by Goulet et al. (2007) and
5  programmed for use in OpenSHA (www.opensha.org; Field et al. 2005). Sensitivity analyses to
6  evaluate effects of soil-property uncertainties should be conducted and considered in developing
7  the final MCE response spectrum.

( MCER )

LY
8  C21.2 RISK-TARGETED MAXIMUM CONSIDERED EARTHQUAKE

9  GROUND MOTION HAZARD ANALYSIS

N
10  Site-specific risk-targeted maximum considered earthquake ( MCER ) ground motions are based

O
11  on separate calculations of site-specific probabilistic and site-specific deterministic ground
12  motions.
n
io
13  Both the probabilistic and deterministic ground motions are defined in terms of 5% damped
at
14  spectral response in the maximum direction of horizontal response. The maximum direction in
the horizontal plane is considered the appropriate ground motion intensity parameter for seismic
m

15 
16  design, using the equivalent lateral force (ELF) procedure of Section 12.8, with the primary
or

17  intent of avoiding collapse of the structural system.

nf

18  Most ground motion relations are defined in terms of average (geometric mean) horizontal
rI

19  response. Maximum response in the horizontal plane is greater than average response by an
20  amount that varies with period. Maximum response may be reasonably estimated by factoring
Fo

21  average response by period-dependent factors, such as 1.1 at short periods and 1.3 at a period of
22  1.0 s (Huang et al. 2008). The maximum direction was adopted as the ground motion intensity
23  parameter for use in seismic design, in lieu of explicit consideration of directional effects.

24  When performing a ground motion hazard analysis, it is important to consider the regional
25  tectonic setting, geology, and seismology. An important geologic consideration is topography,
26  which has been shown to contribute significantly to the effects of local geology on ground
27  motions by increasing the amplitude of ground motions relative to flat ground conditions and
28  causing damage to structures near the crests of slopes (e.g., Assimaki et al. 2005). Since most

5 
 
 1  existing ground motion models do not account for the effects of site topography (Rai et al. 2017),
2  the ground motion hazard analysis should consider topographic effects where appropriate.
3  Topographic amplification is typically considered significant where all of the following are
4  present: (1) slope angles greater than about 17 degrees (Bouckovalas and Papadimitriou 2005),
5  (2) the horizontal or vertical dimensions of the topographic feature are comparable to the seismic
6  wavelengths of interest (e.g., the shear wave velocity of the feature divided by the frequencies
7  that impact structural response), and (3) the location of interest is within a horizontal distance
8  approximately equal to one wavelength from the slope crest. Topographic amplification can be

LY
9  estimated by semi-analytical solutions (e.g., Bouckovalas and Papadimitriou 2005), numerical
10  modeling (e.g., Jeong et al. 2019), laboratory experiments (e.g., Jeong et al. 2019), field

N
11  experiments (e.g., Wood and Cox 2016), and ground motion models (e.g., Rai et al. 2017).
12  Engineers should be cognizant of the large degree of spatial variability in the effects of surface

O
13  topography on ground motions (Wood and Cox 2016) and the limitations, applicability, and
14  uncertainty of each method of analysis.
n
io
15  C21.2.1 Probabilistic ( MCER ) Ground Motions
at
m

16  Probabilistic seismic hazard analysis (PSHA) methods and subsequent computations of risk-
17  targeted probabilistic ground motions, based on the output of PSHA, are sufficient to define
or

18  MCER ground motion at all locations except those near highly active faults. Descriptions of
nf

19  current PSHA methods can be found in McGuire (2004). The primary output of PSHA methods
rI

20  is a so-called hazard curve, which provides mean annual frequencies of exceeding various user-
21  specified ground motion amplitudes. Risk-targeted probabilistic ground motions are derived
Fo

22  from hazard curves using one (or both for comparison purposes) of the methods described in the
23  following two subsections.

24  C21.2.1.1 Method 1

25  The simpler, but more approximate, method of computing a risk-targeted probabilistic ground
26  motion for each spectral period in a response spectrum is to first interpolate from a site-specific
27  hazard curve the ground motion for a mean annual frequency corresponding to 2% probability of
28  exceedance in 50 years (namely 1/2,475 per year). Then, this “uniform-hazard” ground motion is
29  factored by a so-called risk coefficient for the site location that is based on those mapped in
6 
 
 1  Figures 22-18 and 22-19. Via the method explained in the next subsection, the mapped risk
2  coefficients have been computed from the USGS hazard curves for Site Class B and spectral
3  periods of 0.2 and 1.0 s.

4  C21.2.1.2 Method 2
5  The direct method of computing risk-targeted probabilistic ground motions uses the entire site-
6  specific hazard curve that results from PSHA. The computation is detailed in Luco et al. (2007).
7  In summary, the hazard curve is combined with a collapse fragility (or probability distribution of
8  the ground motion amplitude that causes collapse) that depends on the risk-targeted probabilistic

LY
9  ground motion itself. The combination quantifies the risk of collapse. Iteratively, the risk-
10  targeted probabilistic ground motion is modified until combination of the corresponding collapse

N
11  fragility with the hazard curve results in a risk of collapse of 1% in 50 years. This target is based

O
12  on the average collapse risk across the western United States that is expected to result from
13  design for the probabilistic MCE ground motions in ASCE 7.
n
io
14  C21.2.2 Deterministic ( MCER ) Ground Motions
at

15  Deterministic ground motions are to be based on characteristic earthquakes on all known active
m

16  faults in a region. The magnitude of a characteristic earthquake on a given fault should be a best
or

17  estimate of the maximum magnitude capable for that fault but not less than the largest magnitude
18  that has occurred historically on the fault. The maximum magnitude should be estimated
nf

19  considering all seismic-geologic evidence for the fault, including fault length and paleoseismic
rI

20  observations. For faults characterized as having more than a single segment, the potential for
21  rupture of multiple segments in a single earthquake should be considered in assessing the
Fo

22  characteristic maximum magnitude for the fault.

23  For consistency, the same attenuation equations and ground motion variability used in the PSHA
24  should be used in the deterministic seismic hazard analysis (DSHA). Adjustments for directivity
25  and/or directional effects should also be made, when appropriate. In some cases, ground motion
26  simulation methods may be appropriate for the estimation of long-period motions at sites in deep
27  sedimentary basins or from great ( M  8 ) or giant ( M  9 ) earthquakes, for which recorded
28  ground motion data are lacking.

7 
 
 1  When the maximum ordinate of the deterministic (MCER) ground motion response spectrum is
2  less than 1.5Fa, it is scaled up to 1.5Fa to put a lower limit or floor on the deterministic ground
3  motions. A single factor is used to maintain the shape of the response spectrum. The intent of
4  the exception defining site-specific MCER ground motions solely in terms of probabilistic MCER
5  ground motions (i.e., when peak MCER response spectral accelerations are less than 1.2Fa) is to
6  preclude unnecessary calculation of deterministic MCER ground motions.

7  C21.2.3 Site-Specific MCER

LY
8  Because of the deterministic lower limit on the MCER spectrum (Figure 21.2-1), the site-

N
9  specific MCER ground motion is equal to the corresponding risk-targeted probabilistic ground

O
10  motion wherever it is less than the deterministic limit (e.g., 1 .5 g and 0 .6 g for 0.2 and 1.0 s,
11  respectively, and Site Class B). Where the probabilistic ground motions are greater than the
12  n
lower limits, the deterministic ground motions sometimes govern, but only if they are less than
io
13  their probabilistic counterparts. On the MCER ground motion maps in ASCE 7-10, the
at

14  deterministic ground motions govern mainly near major faults in California (such as the San
m

15  Andreas) and Nevada. The deterministic ground motions that govern are as small as 40% of their
16  probabilistic counterparts.
or

17  The exception defining site-specific MCER ground motions solely in terms of probabilistic
nf

18  MCER ground motions (i.e., when peak MCER probabilistic ground motions are less than 1.2Fa)
rI

19  precludes unnecessary calculation of deterministic MCER ground motions. Probabilistic MCER
20  ground motions are presumed to govern at all periods where the peak probabilistic MCER
Fo

21  response spectral acceleration (i.e., < 1.2Fa) is less than 80% of peak deterministic (MCER)
22  response spectral acceleration (i.e., ≥ 1.5Fa).

23  The requirement that the site-specific MCER response spectrum not be less than 150% of
24  the site-specific design response spectrum of Section 21.3 effectively applies the 80% limits of
25  Section 21.3 to the site-specific MCER response spectrum (as well as the site-specific design
26  response spectrum).

27  C21.3 DESIGN RESPONSE SPECTRUM

8 
 
 1  Eighty percent of the design response spectrum, determined in accordance with Section 11.4.6,
2  was established as the lower limit to prevent the possibility of site-specific studies generating
3  unreasonably low ground motions from potential misapplication of site-specific procedures or
4  misinterpretation or mistakes in the quantification of the basic inputs to these procedures. Even if
5  site-specific studies were correctly performed and resulted in ground motion response spectra
6  less than the 80% lower limit, the uncertainty in the seismic potential and ground motion
7  attenuation across the United States was recognized in setting this limit. Under these
8  circumstances, the allowance of up to a 20% reduction in the design response spectrum based on

LY
9  site-specific studies was considered reasonable.

10  Values of the site coefficients (Fa and Fv) for setting the 80% lower limit are introduced to

N
11  incorporate both site amplification and spectrum shape adjustment, as described in the research

O
12  study, Investigation of an Identified Short-Coming in the Seismic Design Procedures of ASCE 7-
13  16 and Development of Recommended Improvements for ASCE 7-16 (Kircher 2015). This study
14  n
found that the shapes of the response spectra of ground motions were not accurately represented
io
15  by the shape of the design response spectrum of Figure 11.4-1 for the following site conditions
at
16  and ground-motion intensities: (1) Site Class D where values of S1 ≥ 0.2; and (2) Site Class E
17  where values of SS ≥1.0 and/or S1 ≥0.2. An adjustment of the corresponding values of Fa and Fv
m

18  was required to account for this difference in spectrum shape, which was causing the design
or

19  response spectrum to underestimate long period motions. Two options were considered to
20  address this short-coming. For the first option, the subject study developed values of new
nf

21  “spectrum shape adjustment” factors (Ca and Cv) that could be used with site factors (Fa and Fv)
rI

22  to develop appropriate values of design ground motions (SDS and SD1). The second option,
ultimately adopted by ASCE 7-16, circumvents the need for these new factors by requiring site-
Fo

23 
24  specific analysis for Site Class D site conditions where values of S1 ≥ 0.2, and for Site Class E
25  site conditions where values of SS ≥1.0 and/or S1 ≥0.2 (i.e., new requirements of Section 11.4.8
26  of ASCE 7-16). The spectrum shape adjustment factors developed by the subject study for
27  Option 1 provide the basis for the values of site coefficients (Fa and Fv) of Section 21.3 that
28  incorporate both site amplification and adjustment for spectrum shape. Specifically, the value of
29  Fv = 2.5 for Site Class D is based on the product of 1.7 (Site Class D amplification at S1 = 0.6,
30  without spectrum shape adjustment) and 1.5 (spectrum shape adjustment factor); the value of Fv
31  = 4.0 for Site Class E is based on the product of 2.0 (Site Class E amplification at S1 = 0.6

9 
 
 1  without spectrum shape adjustment) and 2.0 (spectrum shape adjustment factor), where values of
2  spectrum shape adjustment are taken from Section 6.2.2 (Table 11.4-4) of the subject study. The
3  value of Fa = 1.0 for Site Class E is based on the product of 0.8 (Site Class E amplification at SS
4  = 1.5 without spectrum shape adjustment) and 1.25 (spectrum shape adjustment factor), where
5  the value of the spectrum shape adjustment is taken from Section 6.2.2 (Table 11.4-3) of the
6  subject study. Site amplification adjusted for spectrum shape effects is approximately
7  independent of ground motion intensity and, for simplicity, the proposed values of site factors
8  adjusted for spectrum shape are assumed to be valid for all ground motion intensities.

LY
9  Although the 80% lower limit is reasonable for sites not classified as Site Class F, an exception
10  has been introduced at the end of this section to permit a site class other than E to be used in

N
11  establishing this limit when a site is classified as F. This revision eliminates the possibility of an

O
12  overly conservative design spectrum on sites that would normally be classified as Site Class C or
13  D.
n
io
14  C21.4 DESIGN ACCELERATION PARAMETERS
at

15  The SDS criteria of Section 21.4 are based on the premise that the value of the parameter SDS
m

16  should be taken as 90% of peak value of site-specific response spectral acceleration, regardless
or

17  of the period (greater than or equal to 0.2 s) at which the peak value of response spectral
18  acceleration occurs. Consideration of periods beyond 0.2 s recognizes that site-specific studies
nf

19  (e.g., softer site conditions) can produce response spectra with ordinates at periods greater than
rI

20  0.2 s that are significantly greater than those at 0.2 s. Periods less than 0.2 s are excluded for

21  consistency with the 0.2 s period definition of the short-period ground motion parameter, Ss ,
Fo

22  and recognizing that certain sites, such as Central and Eastern United States (CEUS) sites, could
23  have peak response at very short periods that would be inappropriate for defining the value of the

24  parameter SDS . The upper bound limit of 5 s precludes unnecessary checking of response at

25  periods that cannot govern the peak value of site-specific response spectral acceleration. Ninety
26  percent (rather than 100%) of the peak value of site-specific response spectral acceleration is

27  considered appropriate for defining the parameter SDS (and the domain of constant acceleration)

28  since most short-period structures have a design period that is not at, or near, the period of peak

10 
 
 1  response spectral acceleration. Away from the period of peak response, response spectral
2  accelerations are less, and the domain of constant acceleration is adequately described by 90% of
3  the peak value. For those short-period structures with a design period at, or near, the period of

4  peak response spectral acceleration, anticipated yielding of the structure during MCER ground

5  motions effectively lengthens the period and shifts dynamic response to longer periods at which
6  spectral demand is always less than that at the peak of the spectrum.

7  The SD1 criteria of Section 21.4 are based on the premise that the value of the parameter SD1

LY
8  should be taken as 100% of the peak value of site-specific response spectral acceleration for a

9  period range, 1 s  T  2 s , for stiffer sites vs,30 ft/ s  1,200 ft / s ( vs,30 m/ s  366 m/ s ) similar

N
10  to the previous requirements of Section 21.4 of ASCE 7-10 and for a period range, 1 s  T  5 s ,

O
11  for softer sites vs,30 ft / s  1,200 ft / s ( vs,30 m/ s  366 m/ s ), which are expected to have peak

12 
n
values of response spectral velocity at periods greater than 2 s. The criteria use the maximum
io
13  value of the product, TSa , over the period range of interest to effectively identify the period at
at
14  which the peak value of response spectral velocity occurs. Consideration of periods beyond 1 s
15  accounts for the possibility that the assumed 1/ T proportionality for the constant velocity
m

16  portion of the design response spectrum begins at periods greater than 1 s or is actually 1 / T n
or

17  (where n  1 ). Periods less than 1 s are excluded for consistency with the definition of the 1-s
nf

18  ground motion parameter, S1 . Peak velocity response is expected to occur at periods less than or

equal to 5 s, and periods beyond 5 s are excluded by the criteria to avoid potential misuse of very
rI

19 
20  long period ground motions that may not be reliable. One hundred percent (rather than a reduced
Fo

21  percentage) of the peak value of site-specific response spectral acceleration at the period of peak

22  velocity response is considered appropriate for defining the value of the parameter SD1 since

23  response spectral accelerations can be approximately proportional to the assumed 1/ T shape of
24  the domain of constant velocity for design periods of interest.

25  C21.5 MAXIMUM CONSIDERED EARTHQUAKE GEOMETRIC MEAN ( MCEG )

26  PEAK GROUND ACCELERATION

11 
 
 1  Site-specific requirements for determination of peak ground acceleration (PGA) are provided in a
2  new Section 21.5 that is parallel to the procedures for developing site-specific response spectra

3  in Section 21.2. The site-specific MCE peak ground acceleration, PGAM , is taken as the lesser

4  of the probabilistic geometric mean peak ground acceleration of Section 21.5.1 and the
5  deterministic geometric mean peak ground acceleration of Section 21.5.2. Similar to the

6  provisions for site-specific spectra, a deterministic lower limit is prescribed for PGAM , with the

7  intent to limit application of deterministic ground motions to the site regions containing active
faults where probabilistic ground motions are unreasonably high. However, the deterministic

LY
8 

9  lower limit for PGAM (in g ) is set at a lower value, 0.5 FPGA , than the value set for the zero-

N
10  period response spectral acceleration, 0.6 Fa . The rationale for the value of the lower

O
11  deterministic limit for spectra is based on the desire to limit minimum spectral values, for
12  structural design purposes, to the values given by the 1997 Uniform Building Code (UBC) for
13  n
Zone 4 (multiplied by a factor of 1.5 to adjust to the MCE level). This rationale is not applicable
io
14  to PGAM for geotechnical applications, and therefore a lower value of 0.5 FPGA was selected.
at

15  Section 21.5.3 of ASCE 7-10 states that the site-specific MCE peak ground acceleration cannot
m

16  be less than 80% of PGAM derived from the PGA maps. The 80% limit is a long-standing base
or

17  for site-specific analyses in recognition of the uncertainties and limitations associated with the
18  various components of a site-specific evaluation.
nf

19  REFERENCES
rI
Fo

20  Assimaki, D., Gazetas, G., and Kausel, E. (2005). “Effects of local soil conditions on the
21  topographic aggravation of seismic motion: parametric investigation and recorded field evidence
22  from the 1999 Athens earthquake”. Bulletin of the Seismological Society of America, 95(3),
23  1059-1089.

24  Bouckovalas, G. D., and Papadimitriou, A. G. (2005). “Numerical evaluation of slope

25  topography effects on seismic ground motion.” Soil dynamics and earthquake engineering, 25(7-
26  10), 547-558.

12 
 
 1  Chang, C.-Y., Mok, C. M., Power, M. S., and Tang, Y. K. (1991). “Analysis of ground response
2  at lotung large-scale soil-structure interaction experiment site.” Report NP-7306-SL. EPRI, Palo
3  Alto, CA.

4  Darendeli, M. (2001). “Development of a new family of normalized modulus reduction and

5  material damping curves.” Ph.D. Dissertation, Dept. of Civil Engineering, University of Texas,
6  Austin.

7  EPRI (Electric Power Research Institute). (1993). “Guidelines for determining design basis

LY
8  ground motions.” Report EPRI TR-102293. EPRI, Palo Alto, CA.

9  Field, E. H., Gupta, N., Gupta, V., Blanpied, M., Maechling, P., and Jordan, T. H. (2005).

N
10  “Hazard calculations for the WGCEP-2002 forecast using OpenSHA and distributed object

O
11  technologies.” Seismol. Res. Lett., 76, 161–167.

Goulet, C. A., Stewart, J. P., Bazzurro, P., and Field, E. H. (2007). “Integration of site-specific
12 
n
io
13  ground response analysis results into probabilistic seismic hazard analyses.” Paper 1486, In
14  Proc., 4th Intl. Conf. on Earthquake Geotechnical Engineering, Thessaloniki, Greece, CRC
at

15  Press, Boca Raton, FL.

m

16  Hashash, Y. M. A., and Park, D. (2001). “Non-linear one-dimensional seismic ground motion
or

17  propagation in the Mississippi embayment.” Engrg. Geol., 62(1-3), 185–206.

nf

18  Huang, Y.-N., Whittaker, A. S., and Luco, N. (2008). “Maximum spectral demands in the near-
19  fault region.” Earthquake Spectra, 24(1), 319–341.
rI

20  Idriss, I. M., and Sun, J. I. (1992). User’s manual for SHAKE91. Center for Geotechnical
Fo

21  Modeling, Department of Civil and Environmental Engineering, University of California, Davis.

22  Itasca Consulting Group. (2005). FLAC, Fast Langrangian Analysis of Continua, v. 5.0. Itasca
23  Consulting Group, Minneapolis, MN.

24  Jeong, S., Asimaki, D., Dafni, J., and Wartman, J. (2019). “How topography-dependent are
25  topographic effects?” Complementary numerical modeling of centrifuge experiments. Soil
26  Dynamics and Earthquake Engineering, 116, 654-667.

13 
 
 1  Kircher, C. A. (2015). Investigation of an identified short-coming in the seismic design
2  procedures of ASCE 7-16 and development of recommended improvements for ASCE 7-16,
3  prepared for Building Seismic Safety Council, National Institute of Building Sciences,
4  Washington, DC, prepared by Kircher & Associates, Consulting Engineers, Palo Alto, CA,
5  March 15, 2015.
6  https://c.ymcdn.com/sites/www.nibs.org/resource/resmgr/BSSC2/Seismic_Factor_Study.pdf

7  Kramer, S. L. (1996). Geotechnical earthquake engineering, Prentice-Hall, Englewood Cliffs,

8  NJ.

LY
9  Kwok, A. O. L., Stewart, J. P., Hashash, Y. M. A., Matasovic, N., Pyke, R., Wang, Z., et al.

N
10  (2007). “Use of exact solutions of wave propagation problems to guide implementation of
11  nonlinear seismic ground response analysis procedures.” J. Geotech. Geoenviron. Engrg.

O
12  133(11), 1385–1398.

13  n
Lee, M. K. W., and Finn, W. D. L. (1978). DESRA-2, Dynamic effective stress response analysis
io
14  of soil deposits with energy transmitting boundary including assessment of liquefaction potential,
at
15  Soil Mechanics Series 36, Department of Civil Engineering, University of British Columbia,
16  Vancouver.
m

17  Li, X. S., Wang, Z. L., and Shen, C. K. (1992). SUMDES, a nonlinear procedure for response
or

18  analysis of horizontally-layered sites subjected to multi-directional earthquake loading.

nf

19  Department of Civil Engineering, University of California, Davis.

rI

20  Luco, N., Ellingwood, B. R., Hamburger, R. O., Hooper, J. D., Kimball, J. K., and Kircher, C. A.
21  (2007). “Risk-targeted versus current seismic design maps for the conterminous United States.”
Fo

22  In Proc. SEAOC 76th Ann. Conv. Structural Engineers Association of California, Sacramento,
23  CA.

24  Matasovic, N. (2006). “D-MOD_2—A computer program for seismic response analysis of
25  horizontally layered soil deposits, earthfill dams, and solid waste landfills.” User’s manual,
26  GeoMotions, LLC, Lacey, WA, 20 (plus Appendices).

27  McGuire, R. K. (2004). Seismic hazard and risk analysis, Monograph, MNO-10. Earthquake
28  Engineering Research Institute, Oakland, CA, 221.

14 
 
 1  Menq, F. (2003). “Dynamic properties of sandy and gravely soils.” Ph.D. Dissertation,
2  Department of Civil Engineering, University of Texas, Austin.

3  Parra, E. (1996). “Numerical modeling of liquefaction and lateral ground deformation including
4  cyclic mobility and dilation response in soil systems.” Ph.D. Dissertation, Department of Civil
5  Engineering, Rensselaer Polytechnic Institute, Troy, NY.

6  Pyke, R. M. (2000). “TESS: A computer program for nonlinear ground response analyses.”
7  TAGA Engineering Systems & Software, Lafayette, CA.

LY
8  Ragheb, A. M. (1994). “Numerical analysis of seismically induced deformations in saturated
9  granular soil strata.” Ph.D. Dissertation, Department of Civil Engineering, Rensselaer

N
10  Polytechnic Institute, Troy, NY.

O
11  Rai, M., Rodriguez-Marek, A., and Chiou, B. S. (2017). “Empirical terrain-based topographic
modification factors for use in ground motion prediction.” Earthquake Spectra, 33(1), 157-177.
12 
n
io
13  Schnabel, P. B., Lysmer, J., and Seed, H. B. (1972). SHAKE: A Computer program for
at
14  earthquake response analysis of horizontally layered sites, Report EERC 72-12, Earthquake
15  Engineering Research Center, University of California, Berkeley.
m

16  Vucetic, M., and Dobry, R. (1991). “Effect of soil plasticity on cyclic response.” J. Geotech.
or

17  Engrg., 117(1), 89–107.

nf

18  Wood, C. M., and Cox, B. R. (2016). “Comparison of field data processing methods for the
rI

19  evaluation of topographic effects.” Earthquake Spectra, 32(4), 2127-2147.

Fo

20  Yang, Z. (2000). “Numerical modeling of earthquake site response including dilation and
21  liquefaction.” Ph.D. Dissertation, Dept. of Civil Engineering and Engineering Mechanics,
22  Columbia University, New York.

23  Zhang, J., Andrus, R. D., and Juang, C. H. (2005). “Normalized shear modulus and material
24  damping ratio relationships.” J. Geotech. Geoenviron. Engrg., 131(4), 453–464.

25  Zhang, J., Andrus, R. D., and Juang, C. H. (2008). “Model uncertainty in normalized shear
26  modulus and damping relationships.” J. Geotech. Geoenviron. Engrg., 134(1), 24–36.

15 
 
1  OTHER REFERENCES (NOT CITED)

2  Abrahamson, N. A. (2000). “Effects of rupture directivity on probabilistic seismic hazard

3  analysis.” In Proc., 6th Intl. Conf. on Seismic Zonation, Earthquake Engineering Research
4  Institute, Oakland, CA.

5  Somerville, P. G., Smith, N. F., Graves, R. W., and Abrahamson, N. A. (1997). “Modification of
6  empirical strong ground motion attenuation relations to include the amplitude and duration
7  effects of rupture directivity.” Seismol. Res. Lett., 68, 199–222.

LY
8 

N
O
n
io
at
m
or
nf
rI
Fo

16