Integration of Heterogeneous Data for Multiscale Regional Liquefaction Settlement Mapping

Earthquake-induced soil liquefaction is a geological hazard that can lead to infrastructure damages and economic losses during an earthquake. Accurate assessment and mapping of regional liquefaction require not only models for evaluating liquefaction and its spatial extent but also means to account for heterogeneous sources of information at different scales. In this work, a novel framework is presented to integrate geotechnical data (e.g., cone penetration test data) at the site-specific scale and geologic data (e.g., surficial geologic information) at the regional scale into regional liquefaction settlement mapping. The framework builds on a multiscale random field model to account for the spatial variability of quantities of interest across scales. The conditional sequential simulation algorithm is implemented to integrate geotechnical and geologic data into the random field model. The developed framework is applied to an earthquake-prone area for mapping of liquefaction-induced settlement. The generated liquefaction hazard maps are validated against existing knowledge and liquefaction observations of the region.