Probabilistic seismic hazard analysis of civil infrastructure - Seismic Risk Analysis and Management of Civil Infrastructure Systems - page 15

quantifi cation of epistemic uncertainty. To improve the prediction accuracy
of site effects, the fundamental frequency of a site can be employed, which
renders supplementary information in addition to near surface average
shear wave velocity. Such an additional measure tends to reduce variability
associated with GMPEs (Cadet et al. , 2011), thus resulting in reduction of
overall uncertainty in PSHA. A challenging issue is the appropriate treat-
ment of epistemic uncertainty. Epistemic uncertainty, by defi nition, stems
from incomplete knowledge/understanding of the subject matter. There-
fore, it is not trivial to confi dently capture epistemic uncertainty, and events
or their consequences may not be foreseen (e.g. as in the tragic 11 March
2011 M w 9.0 Tohoku earthquake). Modelling of epistemic uncertainty
requires a synthesis of theories and available data from various fi elds, and
imagination to blend model components to evaluate possible scenarios.

1.3 Extension of probabilistic seismic hazard

analysis (PSHA) to advanced earthquake
engineering analyses

This section presents two illustrative applications regarding extension of
PSHA to other earthquake engineering analyses. Specifi cally, PLHA and
PSRA are discussed in Sections 1.3.1 and 1.3.2, respectively.

1.3.1 PLHA

Ground failure due to large earthquakes has disastrous consequences on
structures and infrastructure, particularly building foundations and buried
pipes. One of the major ground failure phenomena is soil liquefaction,
causing lateral spreading of slope, damage to pile-foundation, and failure
of retaining wall. Seismic loss due to liquefaction can be extremely high
(e.g. sliding of residential houses on slope and tilting of buildings result
in complete demolition of structures). A current trend in the popular
simplifi ed method, which was originally proposed by Seed and Idriss (1971)
as a deterministic procedure, is to carry out probabilistic assessment of
liquefaction triggering potential at a site of interest by taking uncertainties
of multiple earthquake scenarios and ground motions into account. This
is particularly motivated by the adoption of uniform hazard spectra as
a seismic design tool. Conventionally, two key input parameters for
liquefaction triggering analysis are PGA and moment magnitude. From
PSHA, PGA values are obtained as uniform hazard spectra ordinates. The
PGA value at a selected probability level is not governed by a single sce-
nario, rather by multiple scenarios. Therefore, a one-to-one relationship
between PGA from PSHA and moment magnitude is not possible to estab-
lish. Several methods have been developed to deal with this problem by