Liquefaction Hazard Mapping—Statistical and Spatial Characterization of Susceptible Units

Abstract

In this paper, we present a three step method for characterizing geologic deposits for liquefaction potential using sample based liquefaction probability values. The steps include statistically characterizing the sample population, evaluating the spatial correlation of the population, and finally providing a local and/or global estimate of the distribution of high liquefaction probability values for the deposit. When spatial correlation is present, ordinary kriging can be used to evaluate spatial clustering of high liquefaction probability values within a geologic unit which in turn can be used in a regional liquefaction potential characterization. If spatial correlation is not present in the data, then a global estimate can be used to estimate the percentage of samples within the deposit which have a high liquefaction probability. By describing the liquefaction potential with a binomial distribution (high versus low), a global estimate can provide an estimate of the mean as well as uncertainty in the estimate. To demonstrate the method, we used a dense data set of subsurface borings to identify and characterize liquefiable deposits for hazard mapping in Cambridge, Mass.