UCERF3 Extended Earthquake Rupture Forecast

Notes Creating UCERF3 EERF for CyberShake

KM has finished verifying a MeanUCERF3 implementation, and have attached a table which shows comparisons with the full branch sweep aggregated over all 60 sites and 4 IMTs that are noted on this map:

http://opensha.usc.edu/ftp/pmpowers/UCERF3.3/HazardCurves/Sites/

This allows you to weigh the trade-offs between fewer ruptures and more accuracy. At the maximum averaging level you'll have 373,661 ERF ruptures, as opposed to 4,905,378 without any averaging. Once I have gone over everything with NF, we should open a dialog with Ned/Rob/Tom and possibly others about which parameter settings are best for CyberShake.

Here are descriptions for the MeanUCERF3 parameters:

Sect Upper Depth Averaging Tolerance: Some fault sections have different aseismicity values across UCERF3 logic tree branches. These values change the upper depth of the fault section. If > 0, sections with upper depths within the given tolerance of the mean will be combined in order to reduce the overall section and rupture count.

Use Mean Upper Depth: If true and upper depth combine tolerance is > 0, mean upper depth will be used, else the shallowest upper depth will be used when averaging. Note that averaging does not incorporate participation rates, it is an unweighted mean and may not be representative.

Rup Mag Averaging Tolerance: Each rupture has a suite of magnitudes from the different scaling relationships. These magnitudes can be averaged (within a tolerance) in order to reduce the total rupture count. Magnitudes are averaged weighted by their rate. Set to '1' to average all mags for each rupture.

Rupture Rake To Use: Each deformation model supplies rake values for each fault section (and thus each rupture). Invididual rakes can be used, or the rupture count can be reduced by either using the rate-averaged rake or rakes from a specific deformation model.