Exciting news! We're transitioning to the Statewide California Earthquake Center. Our new website is under construction, but we'll continue using this website for SCEC business in the meantime. We're also archiving the Southern Center site to preserve its rich history. A new and improved platform is coming soon!

Group B, Poster #174, Ground Motions

A Bayesian approach for developing ground motion spatial correlation models

Renmin Pretell, Tristan Buckreis, Scott J. Brandenberg, & Jonathan P. Stewart
Poster Image: 

Poster Presentation

2023 SCEC Annual Meeting, Poster #174, SCEC Contribution #13186 VIEW PDF
Ground motion spatial correlation plays a crucial role in the assessment of seismic risk for distributed infrastructure, the development of ground motion models, and the validation of computer simulations and simplified engineering tools. The availability of ground motion spatial correlation models also allows for the informed estimation of ground motion intensity measures at case history sites, which in turn sheds light on the underlying factors leading to observed damage. This work introduces a Bayesian approach for developing ground motion spatial correlation models and an example application for the 2023 Türkiye Earthquake Sequence is presented.

Spatial correlation models...
have traditionally been estimated by fitting empirical semivariograms derived from available data to an exponential functional form conditioned only on separation distance using least-squares regression. This approach has two main limitations: (1) It solely relies on the available ground motion data, which might consist of a few recordings in the case of past events and poorly instrumented zones, and (2) it does not account for the uncertainty in the estimated spatial correlation model. This study addresses these limitations by introducing a Bayesian inference approach for estimating ground motion spatial correlation models. This approach allows for (1) the incorporation of prior knowledge about correlation models from past studies based on the broad body of available ground motion data, and (2) the incorporation of event-specific knowledge to an extent dependent on the amount of data. The Bayesian approach was applied to estimate spatial correlation models for peak ground acceleration, peak ground velocity, and 5%-damped pseudo-spectral acceleration at 1 second oscillator period for the mainshock and three aftershocks from the 2023 Türkiye Earthquake Sequence. The observed spatial trends and the inter- and intra-event variability are discussed.