NorCal Broadband CyberShake Validation

This page documents our CyberShake validation work in Northern California.

Proposed Events

We plan to perform validation for the following events:

• Alum Rock (M5.45)
• San Simeon (M6.6)
• Loma Prieta (M6.94)

Validation Process

To perform validation, we calculate results with both the 1D Broadband Platform and with CyberShake, and then compare these to observations and the GMMs.

1D BBP

1. Run 64 realizations of the validation event, with the same hypocenter, magnitude, and fault surface, but different slip distributions.
2. Calculate broadband seismograms at the locations with observed seismograms, typically about 40 sites.
3. From the seismograms, calculate response spectra.
4. Calculate goodness-of-fit, comparing the BBP results to the observations, for each realization.
5. Combine the individual realization results into an overall GoF plot.
6. Calculate the bias for each of the 64 realizations.

3D CyberShake

1. Run same 64 realizations of the validation event at the same 40 sites to calculate seismograms. Use the same velocity model as is planned for use in the Northern California CyberShake study. (Realistically, most simulation volumes will be small enough that this will only include SFCVM).
2. Using the BBP tools, calculate response spectra.
3. Using the BBP tools, calculate goodness-of-fit comparing the CyberShake results to observations.
4. Combine the individual realization results into an overall GoF plot.
5. Calculate the bias for each of the 64 realizations.

Comparison Data Products

• Side-by-side comparisons of BBP and CyberShake aggregate (over all sites and all ruptures) GoF plots.
• Side-by-side comparisons of BBP and CyberShake for the realizations with the lowest (best) bias.
• Aggregate numerical comparisons over multiple realizations (to be developed)

Alum Rock

BBP

Overall GoF	GoF, best CS realization	GoF, best BBP realization (#60)

GMMs