Community-Supported Ground Motion Processing Software

Date: 4/25/2019

Time: 06:00 PM

Room: Grand Ballroom

We describe ongoing efforts at the U.S. Geological Survey (USGS) to develop automated open-source ground motion processing software. Within the USGS, numerous projects have highlighted the need for standardized and automated ground-motion processing algorithms, including basic seismological research, analysis by the USGS National Seismic Hazard Model Program, and operational systems such as ShakeMap. We expect that the software will be of value to the broader engineering seismology community, and we seek to include ongoing developments in processing algorithms by groups outside of the USGS. The software is written in Python and is built on the open source ObsPy library and incorporates methods developed for the USGS Processing and Review Interface for Strong Motion Data (PRISM). To facilitate data collection, the software reads a range of diverse file formats including any format supported by ObsPy and many formats available from the Center for Engineering Strong Motion Data (CESMD). Additionally, the software can directly query webservices such as: Federation of Digital Seismograph Networks (FDSN) servers and the Istituto Nazionale di Geofisica e Vulcanologia (INGV) Strong Motions Web Service. Processing steps are modular to aid future expansion. These options currently include methods for selecting filter corner frequencies and computing rotation-independent ground motion metrics as well as a wide variety of common engineering intensity measures (e.g., response spectra, duration, Arias Intensity). Our long-term goal is to integrate new algorithms developed in PRISM and by researchers outside of the USGS into a software repository to which others may easily contribute, and which may be used by the broader scientific and engineering community. In this presentation, we describe the ground-motion processing software, demonstrate its capabilities, and present validations of our processing by comparisons with results from existing datasets processed via other methods.

Presenting Author: Eric M. Thompson

Authors