Key points are not available for this paper at this time.
The state of practice for assessing soil liquefaction triggering relies on semiempirical models that use the peak ground acceleration (PGA) at the ground surface and the earthquake moment magnitude (M) to represent the seismic demand leading to liquefaction or lack thereof. Analysts have used various methods to compute PGAs at liquefaction case history sites from previous earthquakes. PGA was often taken as equal to the value recorded at the nearest seismic station, sometimes adjusted based on the results of one-dimensional (1D) ground response analyses to account for differences in site effects. Ground motion models or judgment were often used when a site was not located near a seismic station. These approaches do not account for the differences in path effects between the liquefaction site and the nearest ground motion recording, may neglect differences in site conditions between a seismic station and a liquefaction site, and do not generally consider PGA spatial correlation. This paper’s objective is to compute PGAs at 569 case history sites in the Next Generation Liquefaction (NGL) database using a consistent approach that builds on previous studies. Three alternative intensity measures (IMs) are also computed at the liquefaction sites, namely peak ground velocity (PGV), Arias intensity (IA), and cumulative absolute velocity (CAV). A comparison of legacy and newly estimated PGAs indicates a mean difference of −0.034 in natural logarithmic units (legacy values larger) and 95% of our values lie between 55 and 180% (i.e., a factor of 1.8) of legacy PGAs. This shows that the previous and new PGA values are comparable as a whole, but individual cases have appreciable differences, the causes of which are explained in this paper.
Pretell et al. (Wed,) studied this question.