Forecasting the Long-Term Spatial Distribution of Earthquakes for the 2023 U.S. National Seismic Hazard Model Using Gridded Seismicity

Authors:
Publisher website
Google Scholar
Abstract
Probabilistic seismic hazard analyses such as the U.S. National Seismic Hazard Model (NSHM) typically rely on declustering and spatially smoothing an earthquake catalog to estimate a long-term time-independent (background) seismicity rate to forecast future seismicity. In support of the U.S. Geological Survey’s (USGS) 2023 update to the NSHM, we update the methods used to develop this background or gridded seismicity model component of the NSHM. As in 2018, we use a combination of fixed and adaptive kernel Gaussian smoothing. However, we implement two additional declustering methods to account for the fact that declustering is a nonunique process. These new declustering methods result in different forecasts for the locations of future seismicity, as represented by spatial probability density functions that are later combined with a rate model to produce a full gridded earthquake rate forecast. The method updates, particularly in the separation of the spatial and rate models as well as revised regional boundaries, in some places cause substantive changes to the seismic hazard forecast compared to the previous 2018 NSHM. Additional updates to catalog processing and induced seismicity zones also contribute to changes in the gridded seismicity hazard since 2018. However, these changes are well understood and reflect improvements in our modeling of gridded seismicity hazard.

This publication has 87 references indexed in Scilit: