Modeling spatial cross-correlation of multiple ground motion intensity measures (SAs, PGA, PGV, Ia, CAV, and significant durations) based on principal component and geostatistical analyses

Ground motion intensity measures (IMs) were observed to be spatially correlated during past earthquakes. In this article, a new spatial cross-correlation model for a vector-IM, which consists of spectral acceleration (SA) ordinates at 17 periods and six non-SA IMs (e.g. peak ground velocity, Arias intensity, cumulative absolute velocity, and significant durations), is proposed using principal component analysis (PCA) and geostatistical analysis. A total of 3797 ground motion records are selected from the NGA-West2 database for such analyses. PCA is used to transform the spatially correlated within-event residuals into uncorrelated principal components; a permissible function is then proposed to fit the empirical semivariograms calculated by the principal components. It is evident that the proposed model performs well in capturing the spatial variability characteristics of the multiple ground motion IMs. A simple example is presented to illustrate the use of the proposed model in realizing spatially correlated ground motion residuals of multiple IMs. The model developed enables one to simulate spatially cross-correlated IMs over a large area in a rapid way.

Automated summary

In this article, a new spatial cross-correlation model for ground motion intensity measures is proposed using PCA and geostatistical analysis. The model successfully captures the spatial variability characteristics of multiple IMs, enabling rapid simulation of spatially cross-correlated IMs over a large area.