Civil Engineering Reference
In-Depth Information
be used, provided that the distribution of frequencies associated with high energy is relevant to the
fundamental period of the structure. This can be ensured by generating a record that conforms to an
approved spectral shape. From a structural engineer's viewpoint, it is instructive to note that the features
of strong motions that affect structural response are many and their inter-relationship is complex. It is
thus of importance to highlight the regional differences in strong-motion data and the criteria for the
selection of natural records. These two aspects are discussed hereafter.
3.5.1.1 Regional Differences
As a consequence of the proliferation of strong-motion databanks (e.g. Section 3.7 ), region - specifi c
ground-motion models (attenuation relationships) have been derived. These studies have generated
interest in regional differences in the characteristics of earthquake strong motion. A comprehensive
worldwide review can be found in Douglas (2001, 2006). In order to undertake valid comparisons of
strong-motion characteristics from different regions, records obtained under identical circumstances,
i.e. magnitude, depth, fault mechanism, travel path and site characteristics, are needed from each region.
This is clearly either very unlikely to be achieved or outright impossible.
The results of some studies suggest that regional differences in terms of strong- motion characteristics
in seismically active areas are quite small. The attenuation relationships for peak ground acceleration,
derived for western North America by Joyner and Boore (1981) and for Europe and adjacent areas by
Ambraseys
et al.
(1996), predict rather similar results. For a given magnitude and distance pair, the
difference between the two predictions is usually less than the standard deviation in the employed
attenuation relationships. A study by Spudich
et al.
( 1997 ) has examined strong - motion attenuation
in seismically active zones of tectonic extension. The study concluded that, in general, peak ground
acceleration and acceleration response spectral ordinates are lower in such regions than in other tectoni-
cally active areas. This supports the conclusion of McGarr (1984) who observed that PGAs in exten-
sional regimes to be about 2/3 of the values encountered in compressional regimes. However, it is
intuitively noted that regional differences in elastic response spectra may not carry over to inelastic
spectra, where the infl uence of hysteretic energy absorption and the continuous change in response
periods could possibly overtake regional differences on strong- motion records.
In the absence of a strong-motion databank for a specifi c region, it is necessary to select accel-
erograms from other regions that have produced signifi cant strong-motion recordings. An early deci-
sion is needed regarding whether inter- plate or intra -plate earthquake records are sought. It follows
that a defi nition of intra-plate and inter-plate earthquakes is needed. Dahle
et al.
( 1990 ) derived
attenuation equations for use in intra-plate regions by performing regressions on a data set from
earthquakes that they classifi ed as intra-plate. This data set includes records from regions such as
the eastern USA, Australia and Germany, but also regions such as Greece, Italy and Yugoslavia.
Dahle
et al.
(1990) classify areas as intra-plate on the basis of remoteness from active tectonic plate
boundaries. However, some of these aforementioned regions are areas of appreciable tectonic defor-
mation. Parts of mainland Japan, including the region of the 1995 Kobe earthquake, are also referred
to as intra-plate in some studies on the same basis (e.g. Wesnousky
et al.
, 1984), although, the latter
is also an area of active tectonic deformation. Johnston
et al.
( 1994 ) distinguish intra - plate regions
that are not being actively deformed and refer to these as stable continental regions. The latter defi ni-
tion is more convincing.
The differences between strong-motion characteristics in intra- plate and inter - plate regions are
usually attributed to source and path effects. In terms of the path effects, anelastic (engineering seis-
mology term for inelastic) attenuation is generally assumed to be greater in the more fragmented
inter-plate regions. It has been shown that earthquake ground motions attenuate less rapidly in the
eastern USA than in California (Atkinson and Boore, 1997). Abrahamson and Litehiser (1989 ) include
a term in their attenuation equation for peak acceleration, which implies that anelastic attenuation is
