← Back to Sessions

Capturing the Uncertainty of Seismicity Observations in Earthquake Rate Logic Tree Branches

Description: 

Earthquake rate models based on seismicity are a key element in probabilistic seismic hazard analysis. For the 2023 update of the USGS National Seismic Hazard Model, we split these seismicity rate models into two elements: a spatial kernel that gives the relative distribution of seismicity in a region and an absolute rate model. The absolute rate model gives the distribution of the Poisson rate of earthquakes with magnitudes above the minimum level used to calculate hazard (Mmin), along with a distribution of Gutenberg-Richter b-values that controls the extrapolation to other magnitudes. Poisson is a valid assumption for declustered catalogs or full catalogs when determining low probabilities of a single exceedance. To include uncertainty in the hazard logic tree we use the mean and +-2s (s=1 standard deviation) rate branches. The b-value distribution is determined with the b-positive method (van der Elst, JGR, 2021). The distribution of earthquake rates at the magnitude of completeness (Mc) is determined by considering the likelihood of rates given the number of observed earthquakes with M>=Mc. To get the distribution of rates for M>=Mmin, and other thresholds, we project the distribution of rates from Mc to Mmin using the full distribution of b-values. The mean and +-2s rates of earthquakes with M>=Mmin gives us the first part of the information needed for the three branches. These three rates are each associated with a range of b-values. But we wish to choose a single b-value for each branch. A simple approach is to use the mean b-value with the mean rate, the -2s b-value with the high-rate branch, and the +2s b-value with the low-rate branch. However, that approach can overestimate the uncertainty in earthquake rates for magnitude thresholds higher than Mmin. A second choice is to choose b-values that connect the +-2s rate quantiles at Mmin and the highest magnitude of interest (Mmax). We are exploring the implications of these options.

Session: Opportunities and Challenges in Source Modeling for Seismic Hazard Analysis [Poster]

Type: Poster

Date: 4/20/2023

Presentation Time: 08:00 AM (local time)

Presenting Author: Andrea L. Llenos

Student Presenter: No

Invited Presentation: 

Authors