← Back to Sessions

Methods for Evaluating and Improving Rupture Directivity Modeling in Seismic Hazard Assessment

Description: 

Recent advancements in modeling near-source effects using more realistic features in simulations of earthquake rupture have led to an improved characterization of ground motions. Specifically, these contributions have led to a better understanding of how hypocenter changes for a specific fault geometry affect the rupture and resulting ground motions. Incorporating these source characteristics and near-fault ground motion behavior may contribute to improved modeling of azimuthally varying effects that include the effects of rupture directivity. One example is the application of machine learning methods to support integration of new predictor variables in rupture directivity model development and to introduce more evaluation approaches to assess residuals. Here, we utilize several techniques leveraging an extensive synthetic database. We showcase two examples of how rupture directivity models can be developed, expanded upon, or constrained using artificial neural network models (ANNs).

One approach uses a set of hypothetical earthquakes with corresponding synthetic ground motions from the Southern California Earthquake Center (SCEC) CyberShake study to develop a ground-motion model adapted to incorporate rupture directivity information using an ANN. This extensive database enables us to train the model to capture magnitude, period, and distance variations of the rupture directivity effects and how these parameters relate to hypocenter locations within the finite faults. In some cases, misfit is reduced by better-representing source features not included in ground motion models that neglect hypocenter location (e.g., azimuthal variation, source-to-site terms). Another ANN model uses a shallow-layered neural network model to fit a hypocenter-independent model better. This method adjusts the median and aleatory variability to account for the averaged impact of various hypocenter distributions to fit the underlying directivity adjustment model. This method is a template for applying to other directivity models, improving computational efficiency and enabling more readily integrated integration in hazard codes.

Session: Recent Advances in Modeling Near-source Ground Motions for Seismic Hazard Applications [Poster]

Type: Poster

Date: 4/16/2025

Presentation Time: 08:00 AM (local time)

Presenting Author: Jeff

Student Presenter: No

Invited Presentation: 

Poster Number: 58

Authors