Verification/Validation of scientific outputs of Cybershake:
There are three levels of verification/validation applied to the outputs from Cybershake.
I) Rupture realization Level
1- calculate and plot simulation epsilon for the first run of a fault. Simulation epsilon is (one realization im) – (empirical median) / (empirical sigma). We need two plots for this calculation:
a) a map plot of ratio
b) histogram
note: we will decide on the pass-or-fail threshold to stop the simulation if they are too off.
II) Fault level
1- Calculated and plot these when all realizations of a given fault are finished running:
a) (Simulation median) / (Empirical median) plotted on a map for a given fault; and its histogram
b) (Simulation sigma) / (Empirical sigma) plotted on a map for a given fault; and its histogram
III) Hazard
- Generate hazard map from cybershake and empirical at N_haz hazard levels for N_im IM levels,
e.g., 10% and 2% in 50 years hazard for PGA, PGV, SA(0.5,1.0,5.0,7.5). - Generate map plot of (sim/emp) hazard ratio
- Histogram of ratios (which shows at what proportion of the non-uniform sites (which implicitly represent the spatial applicability of the CS results) we have over- or under-prediction)
- Generate hazard curves for the default list of sites (e.g., 12 major populated regions) automatically, using an extendable code that can do this for N_{site} user-specified site.
- Generate deagg for the N_{site} selected sites above, at N_haz hazard levels
- Plot the contribution of tectonic types of sources based on map. This requires conducting de-agg for all sites at a user-specified hazard level
Note: The minimal list of IMs considered for CS18.6 version is: PGA, PGV, SA(0.5, 1.0,3.0,5.0), Ds595, Ds575, CAV, AI
Here is the planning sketch from the discussion with the software developer group: