Spatial correlation of ground motions in estimating seismic hazards to civil infrastructure - Seismic Risk Analysis and Management of Civil Infrastructure Systems - page 71

Most of these studies considered the infl uence of between-earthquake and
within-earthquake components adopting a single model of the between-
earthquake correlation or analyzing, in addition to the model, two extreme
cases (no correlation of ground motion and perfect correlation). Different
models of within-earthquake correlation were used by Park et al. (2007),
Goda and Atkinson (2009), Molas et al. (2006), Crowley et al. (2008a), and
Sokolov and Wenzel (2011a,b).
Effects of variations in the between-earthquake correlation and in the
within-earthquake correlation on seismic loss and damage estimations for
extended objects (hypothetical portfolio) and critical elements (e.g. bridges)
of a network were analyzed by Sokolov and Wenzel (2011a) for a single
(scenario) earthquake and by Sokolov and Wenzel (2011b) for multiple
earthquakes. They showed that the impact depends on the hazard and prob-
ability level of interest (return period). Characteristics of the analyzed
portfolio or the set of critical elements of a network (e.g. seismic design and
location of elements) may be of considerable importance.
The general fi ndings obtained in the works mentioned above may be
summarized as follows. The higher the within-earthquake correlation and
between-earthquake correlation are, the larger the variation in losses to a
portfolio and the higher probability of extreme loss values become. For
the case of a single (scenario) earthquake, within-earthquake variability
increases the possibility of obtaining extreme motion at one of multiple
locations compared to the single-site probability. Between-earthquake vari-
ability increases the possibility of obtaining extreme motions at all consid-
ered locations simultaneously.
The variation in the shape of the within-earthquake correlation )

function may be also important (e.g. Sokolov and Wenzel, 2011a). The high
rate of decrease of within-earthquake correlation with separation distance
would increase infl uence of within-earthquake variability, which, in turn,
increases possibility of large losses at particular locations.
Some applications require assessing the probability that a specifi c event
will occur during a certain condition at least once or that several such events
will occur simultaneously (joint hazard/risk). For example, there may be
interest in knowing whether several vulnerable elements of a lifeline
network (e.g. bridges, local substations of the power supply network or
pipelines, etc.) are likely to be simultaneously affected by shaking strong
enough to disable them (joint damage) or at least one element will be
damaged, causing the failure of the system. Rhoades and McVerry (2001)
and McVerry et al. (2004) analyzed the infl uence of between-earthquake
and within-earthquake variation on such assessment; Sokolov and Wenzel
(2011a,b) considered the effect of within-earthquake correlation. The
increase of correlation distances, as well as the increase of between-
earthquake correlation, would increase the probability of joint damage.

ρ ε ( Δ