The requirements of Section 21.1 shall be satisfied where site
response analysis is performed or required by Section 11.4.7.
The analysis shall be documented in a report.

A site response model based
on low-strain shear wave velocities, nonlinear or equivalent
linear shear stress-strain relationships, and unit weights shall be
developed. Low-strain shear wave velocities shall be determined from field measurements at the site or from measurements from similar soils in the site vicinity. Nonlinear or equivalent linear shear stress-strain relationships and unit weights
shall be selected on the basis of laboratory tests or published
relationships for similar soils. The uncertainties in soil properties shall be estimated. Where very deep soil profiles make the
development of a soil model to bedrock impractical, the model
is permitted to be terminated where the soil stiffness is at least
as great as the values used to define Site Class D in Chapter 20.
In such cases, the MCE

_{R}response spectrum and acceleration time histories of the base motion developed in Section 21.1.1 shall be adjusted upward using site coefficients in Section 11.4.3 consistent with the classification of the soils at the profile base._{R}response spectrum and acceleration time histories of the base motion developed in Section 21.1.1 shall be adjusted upward using site coefficients in Section 11.4.3 consistent with the classification of the soils at the profile base.

Base ground motion time histories shall be input to the soil profile
as outcropping motions. Using appropriate computational techniques that treat nonlinear soil properties in a nonlinear or equivalent-linear manner, the response of the soil profile shall be determined
and surface ground motion time histories shall be calculated.
Ratios of 5% damped response spectra of surface ground motions
to input base ground motions shall be calculated. The recommended surface MCE

_{R}ground motion response spectrum shall not be lower than the MCE_{R}response spectrum of the base motion multiplied by the average surface-to-base response spectral ratios (calculated period by period) obtained from the site response analyses. The recommended surface ground motions that result from the analysis shall reflect consideration of sensitivity of response to uncertainty in soil properties, depth of soil model, and input motions.The requirements of Section 21.2 shall be satisfied where a
ground motion hazard analysis is performed or required by
Section 11.4.7. The ground motion hazard analysis shall account
for the regional tectonic setting, geology, and seismicity; the
expected recurrence rates and maximum magnitudes of earthquakes on known faults and source zones; the characteristics of
ground motion attenuation near source effects, if any, on ground
motions; and the effects of subsurface site conditions on ground
motions. The characteristics of subsurface site conditions shall
be considered either using attenuation relations that represent
regional and local geology or in accordance with Section 21.1.
The analysis shall incorporate current seismic interpretations,
including uncertainties for models and parameter values for
seismic sources and ground motions. If the spectral response
accelerations predicted by the attenuation relations do not
represent the maximum response in the horizontal plane, then
the response spectral accelerations computed from the hazard
analysis shall be scaled by factors to increase the motions to
the maximum response. If the attenuation relations predict the
geometric mean or similar metric of the two horizontal components, then the scale factors shall be 1.1 for periods less than
or equal to 0.2 sec 1.3 for a period of 1.0 sec., and 1.5 for periods
greater than or equal to 5.0 sec., unless it can be shown that other
scale factors more closely represent the maximum response, in
the horizontal plane, to the geometric mean of
the horizontal components. Scale factors between these periods
shall be obtained by linear interpolation. The analysis shall be
documented in a report.

The probabilistic spectral response accelerations shall be taken as the spectral response accelerations in the direction of maximum horizontal response represented by a 5

*%*damped acceleration response spectrum that is expected to achieve a 1*%*probability of collapse within a 50-year period. For the purpose of this standard, ordinates of the probabilistic ground motion response spectrum shall be determined by either Method 1 of Section 21.2.1.1 or Method 2 of Section 21.2.1.2.At each spectral response period for which
the acceleration is computed, ordinates of the probabilistic
ground motion response spectrum shall be determined as the
product of the risk coefficient,

*C*, and the spectral response acceleration from a 5_{R}*%*damped acceleration response spectrum having a 2*%*probability of exceedance within a 50-year period. The value of the risk coefficient,*C*, shall be determined using values of_{R}*C*and_{RS}*C*from Figs. 22-17 and 22-18, respectively. At spectral response periods less than or equal to 0.2 s,_{R1}*C*shall be taken as equal to_{R}*C*. At spectral response periods greater than or equal to 1.0 s,_{RS}*C*shall be taken as equal to_{R}*C*. At response spectral periods greater than 0.2 s and less than 1.0 s,_{R1}*C*shall be based on linear interpolation of_{R}*C*and_{RS}*C*._{R1}At each spectral response period for which
the acceleration is computed, ordinates of the probabilistic
ground motion response spectrum shall be determined from
iterative integration of a site-specific hazard curve with a log-normal probability density function representing the collapse
fragility (i.e., probability of collapse as a function of spectral
response acceleration). The ordinate of the probabilistic
ground motion response spectrum at each period shall achieve a
1

*%*probability of collapse within a 50-year period for a collapse fragility having (1) a 10*%*probability of collapse at said ordinate of the probabilistic ground motion response spectrum and (2) a logarithmic standard deviation value of 0.6.The deterministic spectral response acceleration at each period shall be
calculated as an 84th-percentile 5

*%*damped spectral response acceleration in the direction of maximum horizontal response computed at that period. The largest such acceleration calculated for the characteristic earthquakes on all known active faults within the region shall be used. For the purposes of this standard, the ordinates of the deterministic ground motion response spectrum shall not be taken as lower than the corresponding ordinates of the response spectrum determined in accordance with Fig. 21.2-1, where*F*and_{a}*F*are determined using Tables 11.4-1 and 11.4-2, respectively, with the value of_{v}*S*taken as 1.5 and the value of_{S}*S*_{1}taken as 0.6.The site-specific MCE

_{R}spectral response acceleration at any period,*S*, shall be taken as the lesser of the spectral response accelerations from the probabilistic ground motions of Section 21.2.1 and the deterministic ground motions of Section 21.2.2._{aM}The design spectral response acceleration at any period shall be
determined from Eq. 21.3-1:

(21.3-1)

where *S*is the MCE_{aM}_{R}spectral response acceleration obtained from Section 21.1 or 21.2. The design spectral response acceleration at any period shall not be taken as less than 80*%*of*S*determined in accordance with Section 11.4.5. For sites classified as Site Class F requiring site response analysis in accordance with Section 11.4.7, the design spectral response acceleration at any period shall not be taken as less than 80_{a}*%*of*S*determined for Site Class E in accordance with Section 11.4.2._{a}Where the site-specific procedure is used to determine the design
ground motion in accordance with Section 21.3, the parameter

For use with the equivalent lateral force procedure, the sitespecific spectral acceleration,

*S*shall be taken as the spectral acceleration,_{DS}*S*, obtained from the site-specific spectra at a period of 0.2 s, except that it shall not be taken as less than 90_{a}*%*of the peak spectral acceleration,*S*, at any period larger than 0.2 s. The parameter_{a}*S*shall be taken as the greater of the spectral acceleration,_{D1}*S*, at a period of 1 s or two times the spectral acceleration,_{a}*S*, at a period of 2 s. The parameters_{a}*S*and_{MS}*S*shall be taken as 1.5 times_{M1}*S*and_{DS}*S*_{D1}, respectively. The values so obtained shall not be less than 80*%*of the values determined in accordance with Section 11.4.3 for*S*and_{MS}*S*_{M1}and Section 11.4.4 for*S*and_{DS}*S*_{D1}.For use with the equivalent lateral force procedure, the sitespecific spectral acceleration,

*S*, at_{a}*T*shall be permitted to replace*S*_{D1}/*T*in Eq. 12.8-3 and*S*_{D1}*T*/_{L}*T*in Eq. 12.8-4. The parameter^{2}*S*calculated per this section shall be permitted to be used in Eqs. 12.8-2, 12.8-5, 15.4-1, and 15.4-3. The mapped value of_{DS}*S*_{1}shall be used in Eqs. 12.8-6, 15.4-2, and 15.4-4.The probabilistic geometric mean peak ground acceleration shall be
taken as the geometric mean peak ground acceleration with a 2

*%*probability of exceedance within a 50-year period.The deterministic geometric mean peak ground acceleration shall be
calculated as the largest 84

^{th}-percentile geometric mean peak ground acceleration for characteristic earthquakes on all known active faults within the site region. The deterministic geometric mean peak ground acceleration shall not be taken as lower than 0.5*F*_{PGA}, where*F*_{PGA}is determined using Table 11.8-1 with the value of PGA taken as 0.5 g.The site-specific MCE

_{G}peak ground acceleration, PGA_{M}, shall be taken as the lesser of the probabilistic geometric mean peak ground acceleration of Section 21.5.1 and the deterministic geometric mean peak ground acceleration of Section 21.5.2. The site-specific MCE_{G}peak ground acceleration shall not be taken as less than 80*%*of PGA_{M}determined from Eq. 11.8-1.