Here is the to-do list for Cyerbshake:
Current version (18.6):
...
Computational tasks:
- Run empirical IM and ratio plots for the remaining faults
- Code IM ratio statistical analysis and plot histogram of the ratio for further investigation; then hand it over to software engineers to add to the workflow
- Conduct manual verification of the existing IM plots
- Run IM aggregate and hazard calc for SA(5.0) as an example case
- faults, and investigate manually
- Re-run IM aggregation for 18p6, as we have corrected Mw for some faults. (Note: IM aggregation runs are computationally cheap)
- Append new subduction result
- Run hazard map calculations for SA5.0
- Verify the hazard curves for single locations manually by comparing with the old version of cybershake and empirical resultsVerify the hazard curves as the code has changed again
- Run hazard calcs for CS18.6 for all IMs existing empirical PSHAs
- Compare hazard curves and hazard maps for 18.6 and 18.5
- Fill in the draft paper with updated results
- Incorporate the simulation_epsilon code to the workflow Review and revise – and send to Brendon for review
- Conduct hazard calculations via python code and compare with the openSHA results
Writing/presentation tasks:
- Re-structure the current draft to include more of computational aspects
- Add new figures based on the updated results
...
- Finalized the Cybershake "Run Manual" wiki page
- Run workshop for all software developer to familiarize them with ALL components of cybershake simulations
Future Version:
- Identify source parameters that can be considered in the list of variable parameters for source generation (besides Mw)
- Investigate the correlation structure between the chosen parameters
- Set up the SRF generation code to include these parameters
- Generate SRF and VM
- Run verification test on them (both via verification code + manual checks)
- Run GM sim via new workflow
- Run IM verification code + do a manual verification
- Plot IMs
- run IM a_agg and Hazard calc
- Compare results from the previous steps with the empirical and previous versions of cybershake runs
...