Civil Engineering Reference
In-Depth Information
TSMIP network: the CHY, ILA, TAP, and TCU arrays. The CHY array is
located on the Chianan Plain - an area of thick sediments, from central west
to southwestern Taiwan; most stations of the ILA and the TAP arrays are
installed within alluvium-fi lled basins with relatively rapid lateral changes
of depth of sediments; the stations of TCU array are located partly on the
northern part of Chianan Plain and hilly areas in Northwestern Taiwan. The
comparison clearly reveals the infl uence of peculiarities of local geological
conditions on within-earthquake correlation. It seems that the correlation
depends on spatial variability of geological characteristics of the area; the
correlation should decrease for the sites that do not share the same geology,
or for the sites that are not affected by similar site effects.
Note also that regression analysis, which is used for development of pre-
diction equations, does not usually incorporate the spatial correlation struc-
ture. The infl uence of spatial correlation in the analysis was studied by Hong
et al.
(2009) and Jayaram and Baker (2010). It has been shown that the
effects of incorporating spatial correlation on the estimated regression coef-
fi cients are insignifi cant; however, the variances of the between-earthquake
and of within-earthquake residuals may change signifi cantly (Jayaram and
Baker, 2010).
Bearing in mind that a key element of seismic hazard assessment and loss
estimation is the consideration of uncertainties, which are classifi ed as epis-
temic and aleatory, here it is necessary to note that the ground-motion
correlation, in principle, depends on the chosen ground-motion model,
because the correlation describes the behavior of residuals between
the observations and predictions. Thus, different models for between-
earthquake and within-earthquake correlation combined with the corre-
sponding ground-motion models would represent a source of epistemic
uncertainty. Different realizations of the random ground-motion fi eld in
terms of between-earthquake and within-earthquake residuals and ground-
motion correlation represent aleatory variability.
3.3
Ground-motion correlation and seismic
loss assessment
3.3.1 Example application for seismic loss estimation
To illustrate the infl uence of ground-motion correlation on estimations of
seismic loss, we consider here the extreme cases of ground-motion correla-
tion, i.e. all the variability between earthquakes (
ρ
η
=
1.0) and all the vari-
ability within earthquakes (
0 km). The loss
estimations were performed for a set of hypothetical buildings mimicking
a real building stock, which was constructed based on three types of build-
ings that are typical for Taiwan. The whole area of 22 km
ρ
η
=
0.0;
ρ
(
Δ
)
=
0.0 for
Δ
>
×
18 km was
Search WWH ::
Custom Search