Reducing the NSHM23 California Rupture Set for Earthquake Risk Applications and Adapting Time-Dependent Recurrence Models

Heather Crume, Jessica Velasquez, Jochen Woessner, & Chris Milliner

Submitted August 30, 2026, SCEC Contribution #15305, 2026 SCEC Annual Meeting Poster #154

The 2023 U.S. National Seismic Hazard Model (NSHM23) introduces a substantially updated California crustal fault model, including revised slip rates and expanded multi-fault rupture connectivity. While these updates improve physical realism, the resulting rupture set contains more than 400,000 ruptures, creating challenges for operational hazard and earthquake risk applications. We present Moody's approach to developing a reduced NSHM23 California crustal fault rupture set that preserves key hazard characteristics while improving computational efficiency for stochastic event set generation and portfolio-scale risk analysis. The reduction is based on a sparse, exposure-weighted least-squares inversion that derives rupture-rate scaling factors relative to the full NSHM23 rupture set. The inversion drives many rupture rates to zero while increasing rates on retained ruptures to preserve hazard metrics, magnitude-frequency distributions, and spatial patterns of seismicity. We also discuss development of a corresponding time-dependent model for the reduced rupture set. UCERF3 provides a useful methodological foundation through its treatment of elastic rebound, fault-section recurrence, and time-dependent rupture probabilities. However, direct application to NSHM23 is complicated by the prevalence of complex multi-fault ruptures. In UCERF3, aperiodicity is parameterized primarily as a function of magnitude. For NSHM23, rupture complexity may provide a more meaningful proxy for recurrence variability, as the number of participating fault sections, fault-to-fault jumps, and overall rupture connectivity may better reflect recurrence irregularity than magnitude alone. We therefore explore formulations in which aperiodicity scales with rupture complexity. A related challenge is subsection recurrence, which can become unrealistically short when subsections participate in numerous alternative multi-fault ruptures spanning a broad magnitude range. We investigate magnitude-dependent subsection recurrence measures to provide a more stable basis for estimating time-dependent rupture probabilities. This work aims to develop a reduced rupture set and corresponding time-dependent framework that preserve the essential behavior of NSHM23 while remaining practical for operational earthquake risk applications.

Citation
Crume, H., Velasquez, J., Woessner, J., & Milliner, C. (2026, 08). Reducing the NSHM23 California Rupture Set for Earthquake Risk Applications and Adapting Time-Dependent Recurrence Models. Poster Presentation at 2026 SCEC Annual Meeting.

Related Projects & Working Groups
Earthquake Forecasting and Predictability (EFP)