← Back to Sessions

Overcoming Factors That Limit the Predictive Power of Probabilistic Fault Displacement Hazard Models

Description: 

Fault displacement models provide surface slip probability distributions and are a key component of Probabilistic Fault Displacement Hazard Analyses (PFDHA). Relatively high residuals between predicted slip and slip observed in historical ruptures reflect large aleatory variability common across the models. We investigate variables that may modulate the amplitude of coseismic surface slip in order to refine slip predictions to match observations. Using new empirical models developed from the UCLA Fault Displacement Hazard Initiative global rupture database, we compare model residuals against variables thought to modulate surface slip, including depth to bedrock, topographic slope, relief, hypocentral and centroid distance, and number of fault strands.

Our initial findings show that the strongest source of discrepancy corresponds to portions of ruptures that violate the simple form of the predictive models, particularly where secondary faults are involved, reflecting the challenge of synthesizing slip observations in zones of complex geometry and multiple fault traces. This challenge is not easily overcome because models tend to have simple representations of slip distributions that do not capture real-world complexities. We propose that model predictions can be through careful selection of observational data that satisfy the premise of the model, that is, those which lie along the modelled fault trace(s). Conversely, quantifying fault complexity is a common objective warranting the inclusion of slip distributed off of a principal trace. For next generation fault displacement models, resolving accurate slip predictions for individual strands in areas of fault complexity may require careful filtering and selection of data on a case-by-case basis rather than through automated amalgamation approaches, though a hybrid method is possible. Robust interrogation of geological or geometric characteristics that modulate slip systematically at a global scale will first require refinement or filtering of the input observations to address cases that violate model assumptions.

Session: Assessing Seismic Hazard for Critical Facilities and Infrastructure – Insights and Challenges [Poster Session]

Type: Poster

Date: 5/3/2024

Presentation Time: 08:00 AM (local time)

Presenting Author: Austin

Student Presenter: No

Invited Presentation: 

Authors