← Back to Sessions

A Probabilistic Fault Displacement Analysis (PFDHA) Hazard Model for Rail Transportation Systems

Session: Fault Displacement Hazard: New Data and Modeling Advances [Poster]

Type: Poster

Date: 4/22/2021

Presentation Time: 03:45 PM Pacific

Description: 

We present a PFDHA model for performing assessments at rail alignment crossings, where displacements are needed for both the principal fault trace and total tectonic displacement across the fault zone based on the formulation of Youngs et al. (2003). Past assessments of the components of PFDHA have been primarily ergodic in which data from multiple earthquakes have been combined to develop the needed probability distributions that are treated as aleatory variability in the PFDHA, leading to broad distributions for the displacement that may occur at a single point on a fault, which are not consistent with the geologic observations of Hecker et al (2013) showing a COV of about 0.5. Therefore, we partition the overall uncertainty in PFDHA into aleatory and epistemic components, with an aleatory COV for displacement consistent with the Hecker et al (2013) assessments.

An initial database consisting of reported fault displacements along the length of mapped ruptures was developed from published sources for 25 strike slip and 15 reverse earthquakes. The data consists of displacements at mapped locations and components of motion. These distributions were used to assess average displacement for each rupture and to update empirical models of average displacement versus magnitude. Normalizing the displacements for each rupture by the average displacement effectively removed event-to-event correlations in D/Daverage. The model treats the distribution of D/Daverage as aleatory and the distribution of Daverage given the earthquake as epistemic. The distributions for the variability in D/Daverage along the length of a rupture were assessed accounting for spatial correlation in observations using an exponential semi-variance model.

We summarize the initial model development and example PFDHA results, as well as illustrating the effect of alternative partitioning between aleatory variability and epistemic uncertainty on hazard assessments.

Presenting Author: Robert Youngs

Student Presenter: No

Authors