Why are there differences in site-specific seismic hazard analyses for critical infrastructure and the National Seismic Hazard Model?
Ivan Wong
Lettis Consultants International
- Date & Time
- Location
- Hybrid In-Person and Online seminar via Microsoft Teams
- Host
- Justin Rubinstein
- Summary
It may come as a surprise to those who do not perform site-specific seismic hazard analyses, such as many in the engineering community, that there can be significant differences in the estimated hazard from site-specific studies and the National Seismic Hazard Model (NSHM). These differences are largely due to differences in methodology. The objective of the NSHM is to produce a national map where the hazard is uniformly consistent across the U.S. This objective requires making choices that are not considered in performing site-specific seismic hazard analyses. One such choice is with regards to addressing epistemic uncertainty. In the NSHM, only peer-reviewed publications can be used in model development. In a well-executed site-specific hazard analysis following the principles of SSHAC (Senior Seismic Hazard Advisory Committee), all available information and data whether published or peer-reviewed or not needs to be considered in model development to capture the center, body, and range of technically defensible interpretations.
Also there are computational constraints involved in calculating the hazard at thousands of grid points versus a single site that may preclude considering all possible sources of epistemic uncertainty. For example, the epistemic uncertainty in fault dip is not addressed in the NSHM. This uncertainty is particular important if the site is near an active fault. The epistemic uncertainties in rupture behavior and slip rates can be very large particularly for the less well-studied faults in the U.S. but the NSHM often has to make simplified choices on fault parameters that are input into hazard analyses. Addressing site effects is a significant issue now that the NSHM produces maps based on Vs30. Obviously at a national-scale, those effects need to be incorporated through the use of Vs30 in the western U.S. and generic amplification factors in the central and eastern U.S. For a particular site, it should not be surprising that the amplification can be very different performing a site response analysis versus using ground motion models and Vs30 or generic amplification factors.
We have compared the results of site-specific seismic hazard analyses for more than 100 sites spread across the U.S. with the 2023 NSHM and have noted frequency-dependent differences in hazard for both rock and soil sites more than 50%. Hence the engineering community needs to be aware of such differences when looking at the NSHM for estimating the hazard particularly for critical and important facilities/structures.