← Back to Sessions

How Many Earthquake-rupture Simulations Are Required to Quantify Epistemic Ground-Motion Variability?

Description: 

Modern probabilistic seismic hazard assessment increasingly relies on ground-motion simulations to quantify epistemic uncertainties. Seismic source characteristics — such as rupture length and slip distribution — strongly influence ground motion intensity, highlighting the need for comprehensive analyses of how rupture process variations affect the resulting shaking levels and their spatial distribution. Physics-based simulations (PBS) have become an essential tool for rupture scenario computations by solving the seismic wave equation, thereby also bypassing the ergodic assumption. However, PBS’s high computational cost, especially for 3D Earth structure and frequency bands of engineering interest (up to 20 Hz), leads to a critical question: How many simulations are needed to accurately characterize the epistemic ground-motion uncertainties?

This study addresses the challenge of determining the optimal number of simulations for seismic hazard assessment across various locations and scenarios. Using a kinematic rupture generator, we simulate strike-slip and dip-slip earthquakes at magnitude 6.5. The radiated seismic wave fields are computed via a wave-number integration method in a 1D layered Earth model. We then use the resulting database of ~1600 simulations to quantify the standard deviations of seismic ground motion intensities. Based on these calculations, we estimate the number of simulations required to capture ground-motion intensity distributions within a specified confidence interval. Results reveal significant variability in the required simulations, ranging from a few dozen to thousands, depending on earthquake characteristics, the relative location of the station to the fault, and the adopted error margin for intensity estimation. For example, strike-slip earthquakes for stations aligned with the fault show high variability in peak ground velocity (PGV), requiring over 600 simulations to estimate the PGV distribution accurately. These findings provide a framework for optimized resource allocation and improved seismic hazard assessment using PBS-based approaches.

Session: Accuracy and Variability of Physics-based Ground Motion Modeling - II

Type: Oral

Date: 4/15/2025

Presentation Time: 05:15 PM (local time)

Presenting Author: David

Student Presenter: No

Invited Presentation: 

Poster Number:

Authors