Ground motion selection for seismic risk analysis of civil infrastructure - Seismic Risk Analysis and Management of Civil Infrastructure Systems - page 99

1

8.0

7.5

0.8

7.0

0.6

6.5

0.4

6.0

0.2

50% in 50 yrs
10% in 50 yrs
1% in 50 yrs

50% in 50 yrs
10% in 50 yrs
1% in 50 yrs

5.5

0

10 1 10 2

0 1 2 3 4

Distance, R rup (km)

Amplitude scale factor, SF

(a) (b)

4.8 Comparison of: (a) amplitude scale factors and (b) magnitude-
distance distribution of the ground motions selected for three different
exceedance probabilities (after Bradley, 2012b).

0.083 and 0.17 g , respectively. That is, the SA (3.0) amplitude changes by a
factor of approximately 2, as the exceedance probability level increases
each time. In Fig. 4.7a and Table 4.3, it can be seen that the largest differ-
ence between the SA ordinates of the selected ground motions for the three
exceedance probabilities is for SA (3.0) and decreases as the vibration
period of SA tends to be away from T 3.0 s. Figure 4.7b illustrates that

the median PGV value of the selected ground motions at the 50% and 1%
exceedance probabilities have ratios of 0.48 and 1.7 with the median PGV
value for 10% exceedance probability. Similarly, Figure 4.7c illustrates that
the ratio of the median CAV values for the 50% and 10% exceedance prob-
abilities is 0.49, while the same ratio of the 1% and 10% exceedance prob-
abilities is 1.2 (notably less than the equivalent ratio for PGV , for example).
Figure 4.7d illustrates that, in contrast to the other aforementioned IM i s,
the median values of Ds 595 are similar for the three different exceedance
probabilities. A thorough explanation for the IM i properties of the selected
ground motions for the 50%, 10%, and 1% exceedance probabilities
observed in Fig. 4.7 is given in Bradley (2012b).
Similar to Figs 4.6, Fig. 4.8 compares the amplitude scale factors and M w

=

− R rup distributions of the selected ground motion sets for the three different

exceedance probabilities. For the 50%, 10% and 1% exceedance probabili-
ties, it can be seen that the median scale factors of the selected ground
motion sets were 1.14, 1.10 and 1.60, respectively. It can also be seen that
the EDFs of scale factors for the 50% and 10% exceedance probabilities
are similar. The fact that the EDF of scale factors for the 1% exceedance
probability is slightly larger than that for the 50% and 10% exceedance
probabilities is likely a refl ection of the reduced number of as-recorded
ground motions (in the prospective database) with large SA (3.0) values.