← Back to Sessions

Incorporating Results From Numerical Simulations Into Ground-motion Models

Description: 

Large data sets of simulated ground motions are being generated. These simulation data sets supplement the sparse sampling of ground motions from large-magnitude events in empirical data sets. Often, the simulations are conducted for just a few key scenarios in a region. An issue is how to use these two data sets to develop improved ground-motion models (GMMs). One approach is to combine the empirical data and the simulation data into one large data set and conduct a regression analysis in the traditional manner. The simulation data set will typically have many ground motions at many more locations than empirical data sets, leading to the question of how to weight these two data sets in the regression. An alternative approach is to compute the residuals of the simulated data relative to an ergodic empirical GMM and develop adjustments to the GMM to reflect the behaviors in the simulated data set. This approach allows inclusion of the features of the simulations that are considered reliable and exclusion of features that are less reliable. In particular, the simulations may include systematic path effects in the due to the 3-D velocity structure that are more reliable than the source effects which are based on limited validation exercises for the same sparse set of large-magnitude earthquakes used in the empirical GMMs. This represents an extension of the approach of using the scaling in the simulations rather than the absolute ground-motion amplitudes. The simulated data set can be used to adjust the ergodic scaling of median path effects (average path effects for a region) and to constrain the median non-ergodic path effects for a specific source-site pair. The variability of the ground motion from simulations has not been adequately validated and is currently not directly used in practice, but the relative scaling of the variability for different sites can be used to adjust the regional average standard deviation to site-specific standard deviations. Currently, simulations have not been adequately validated to replace GMMs, but they can be used to adjust GMMs is an interim step.

Session: Challenges and Opportunities in Constraining Ground-motion Models from Physics-based Ground-motion Simulations - I

Type: Oral

Date: 4/17/2025

Presentation Time: 08:00 AM (local time)

Presenting Author: Norman

Student Presenter: No

Invited Presentation: Yes

Poster Number:

Authors