Ground motion selection for seismic risk analysis of civil infrastructure - Seismic Risk Analysis and Management of Civil Infrastructure Systems - page 80

ground motions consistent with a desired exceedance probability from
PSHA non-trivial. The ground motion selection methodology discusses
herein is applied for the selection of PSHA-consistent ground motions.
However, because of its theoretical basis, minor modifi cations allow the
methodology to also be applied to ground motion selection for scenario
events (i.e. from DSHA).
4.2.3 An overview of contemporary ground
motion selection
Current ground motion selection methodologies are burdened by the
limited amount of information on the severity of a ground motion that is
conveyed (often using a single ground motion intensity measure ( IM ), e.g.
peak ground acceleration). For example, ground motion selection methods
utilized in both research and engineering practice commonly select motions
on the basis of their 'match' to some target response spectrum (e.g. see
Katsanos et al. , 2010, and references therein). Such target response spectra
maybe: (i) a design response spectra in seismic codes (ASCE, 2006; ATC-58,
2007; CEN, 2003); (ii) the response spectra of a single earthquake scenario
specifi ed in DSHA, or inferred as the governing scenario from PSHA
(McGuire, 1995); or (iii) the response spectra of a single earthquake sce-
nario conditioned on a particular spectral acceleration ordinate (Baker,
2011), among many others. Several studies have also recently proposed
methods to select ground motion suites which account for both the mean
and variability in response spectra for such target response spectra (Jayaram
et al. , 2011; Kottke and Rathje, 2008; Wang, 2011).
The primary limitation of the aforementioned ground motion selection
methods stems from the fact that only characteristics of the ground motion
represented in terms of (elastic) spectral accelerations are considered,
whereas it is well acknowledged that the severity of a ground motion, in
general, depends on its intensity, frequency content and duration. Spectral
accelerations, by defi nition represent the peak response of an elastic single-
degree-of-freedom oscillator of a specifi c period, and therefore do not
explicitly account for many other important features of ground motions,
such as the duration and energy (both total and its temporal accumulation)
of ground shaking, among others. It should be noted that while causal
parameters, such as earthquake magnitude, source-to-site distance, etc., are
often considered in addition to the target spectrum, they are only implicit
factors affecting a ground motion, but are not explicit measures of the
severity of the ground motion itself.
The ground motion selection methodology presented herein was devel-
oped bearing in mind the aforementioned limitations of conventional
methods. This chapter is intended to provide an overview of the methodol-