Geoscience Reference
In-Depth Information
Fig. 18.9. Fromtop to bottom: (a) comparison between the spatial variation of
PGSa
obtained by numerical simulations (open circles) and by eq. (18.5) (solidtriangles); (b)
PGSa
vs.
PGV
pairs at ground surface simulated (open circles)and obtained by
eq. (18.5) (solidtriangles). The linecorresponding toeq. (18.3) isalso shown for
comparison
evaluation.InFigure18.8,thehorizontaldisplacementtimehistoriessimulatedatthesur-
faceoftheDuzcebasinareplotted,togetherwiththecorrespondinghorizontal(in-plane)
components of ground strain. It is clear that, corresponding to the lateral (South) edge
of the basin, surface waves are generated and that the largest horizontal (in-plane) strain
components are carried by such waves, while the strains induced by the nearly in-phase
S arrivals are much smaller. Furthermore, since
PGS
is carried in this case by Rayleigh
waves,while
PGV
byS-waves,astraightforwardcorrelationbetween
PGS
and
PGV
such
as eqs. (18.1) or eq. (18.3) should beconsidered withcare.
In Figure 18.9 the spatial variation of
PGSa
along the cross-section is shown, together
with the corresponding relationship with the horizontal
PGV
. Both numerical values and
the ones obtained by eq. (18.5) are shown, demonstrating a reasonable agreement. It
is worth noting that the values obtained by 2D analyses lie close to the 16
◦
percentile
of variability of the experimental results shown in the previous section (eq. 18.3). The
underestimation of ground strains with respect to the median value of (3) can likely be
attributed to an oversimplified soil model assumed in the numerical calculations, and
a consequent excess of spatial coherency of the simulated ground motion with respect
to reality. It should also be noticed that linear-elastic behavior of soil was considered
in these analyses. Non-linearity should act as a “filter” of strains larger than a certain
threshold, depending on soil strength: this effect should be considered in more detail in
future studies.
