SCEC Award Number 17213 View PDF
Proposal Category Individual Proposal (Data Gathering and Products)
Proposal Title Continued Development of OpenSHA/UCERF3 in Support of Operational Earthquake Forecasting, Hazard Assessment, and Loss Modeling
Investigator(s)
Name Organization
Thomas Jordan University of Southern California Edward Field United States Geological Survey Kevin Milner University of Southern California
Other Participants
SCEC Priorities 4d, 5a, 5d SCEC Groups WGCEP, EFP, CS
Report Due Date 06/15/2018 Date Report Submitted 11/14/2018
Project Abstract
 Major developments were made in support of Operational Earthquake Forecasting (OEF). Development was primarily focused on the UCERF3-ETAS model; this model bridges the gap between point process spatiotemporal clustering models and the UCERF3-TD finite fault models with elastic rebound. We prepared an exhaustive set of possible UCERF3-ETAS data products for the HayWired scenario earthquake and presented them to the user community at a USGS Powell Center meeting (April 2017) and published the findings in SRL (Field & Milner 2018). A generic Fast-ETAS model was also developed in OpenSHA to eventually replace the previously developed Reasenberg & Jones (1989) Operational Aftershock Model which will soon provide aftershock statistics on USGS event pages. CyberShake collaboration has also been strong in this report period, primarily in support of the Utilization of Ground Motion Simulations (UGMS) committee. The OpenSHA-based back-end for the UGMS MCER & Design Response Spectra tool was finalized in 2017 and published in April 2018 (https://data2.scec.org/ugms-mcerGM-tool_v18.4/).
Intellectual Merit The UCERF3-ETAS model implemented in OpenSHA was published in BSSA in February 2017. Typical short term clustering models do not consider the proximity to known faults when assessing the probability of a large aftershock, although CEPEC has recognized the importance of this. UCERF3-ETAS is the first model to couple ETAS statistics with a full fault-based long-term earthquake forecast, illustrating both the advantages and challenges therein. This allows for computation of full short-term time-dependent hazard and loss data products which are consistent with the long term hazard models adopted by the user community (exemplified in Field & Milner 2018).
Broader Impacts OpenSHA development in this reporting period culminated in the public release of the UGMS MCER & Design Response Spectra lookup tool in April 2018 (https://data2.scec.org/ugms-mcerGM-tool_v18.4/). This project was presented to the engineering community at the 11th National Conference on Earthquake Engineering in June 2018 and received media attention through a New York Times article. Awaiting formal adoption into the Los Angeles building code, this project may help engineers design better design structures informed by SCEC CyberShake 3-dimensional deterministic ground motion simulations. The UCERF3-ETAS model was presented to its potential user community at an April 2017 USGS Powell Center meeting, with participation from FEMA, public utilities, and the insurance/loss modeling community.
Exemplary Figure Figure 1

(a) Average number of triggered  M≥2.5 earthquakes nucleating in 0.02°×0.02°  cells over a 7‐day period following an M 7.1 earthquake on the Hayward fault (the “HayWired” scenario, shown with the white line) based on 200,000 U3ETAS simulations. (b) The gain of  M≥2.5 nucleation rates (the rates plotted in (a) plus long‐term UCERF3‐TI nucleation rates and then divided by the latter). (c) Same as (a), but for a pure (no‐faults) ETAS model that is otherwise equivalent to U3ETAS. (d) Ratio of rates shown in (a) versus (c) (after adding long‐term rates to each to avoid division by zero in more distant regions), revealing the influence of faults in U3ETAS. The dashed box is the San Francisco (SF) region utilized for several of the hazard metrics discussed in this article.

Credits: Ned Field (USGS) and Kevin Milner (USC), Figure 2 of Field & Milner 2018