← Back to Sessions

WITHDRAWN Clustering in PSHA: A Study on Short Return Period Risk Assessments

Description: 

WITHDRAWN Probabilistic seismic hazard assessment (PSHA) typically employs a Poisson assumption, i.e., random and independent events. However, the Poisson assumption fails for highly clustered processes such as aftershock sequences and earthquake swarms. PSHA often handles these processes by declustering earthquake catalogs in an attempt to remove non-Poissonian behavior. An alternative approach used in the U.S. Geological Survey (USGS) National Seismic Hazard Model (NSHM) 2023 combines full catalog rates and a Poisson model with the spatial pattern from a declustered catalog. This method is less dependent on declustering methods and is an acceptable approximation for low probabilities of exceedance (e.g., 2% or 10% in 50 yr) used in the USGS NSHM 2023 hazard metrics (Field et al., 2021; Marzocchi & Taroni, 2014). However, for shorter forecast durations that are applicable for the insurance industry, the limitations of the Poisson assumption are apparent and have significant impacts on hazard (Field et al., 2021). In this study, we investigate the sensitivity of earthquake rate estimates at shorter forecast durations using full catalogs with both Poisson and non-Poisson (i.e., time-varying) models. The key questions are how hazard intensities and losses change when using full catalogs with a Poisson assumption versus non-Poisson models at short return periods, and whether non-Poisson models are sufficient when using full catalogs or if fully time-dependent models are required to capture the behavior of shorter forecast durations. Building on recent work evaluating the impact on short return period losses in Puerto Rico when using full catalogs (Farghal & Velasquez, 2024), we compare the performance of the Poisson model in the western US from the USGS NSHM 2023 with non-Poisson models, such as the ETAS model and the negative binomial distribution. The outcomes of this analysis aim to 1) evaluate the impact of the Poisson assumption at shorter forecast durations using full catalogs, and 2) help determine whether fully time-dependent models are warranted for shorter forecast durations in seismic hazard and risk assessments.

Session: Advancing Time-dependent PSHA and Seismic Risk Assessment: Accounting for Short- to Medium-term Clustering [Poster]

Type: Poster

Date: 4/15/2025

Presentation Time: 08:00 AM (local time)

Presenting Author: Heather

Student Presenter: No

Invited Presentation: 

Poster Number: 73

Authors