Background

CS Scientific output

Story / Deliverable

a) For a given fault (median and sigma ratios) are plotted (histogram and map)

b) For a given realisation, a map plot of the simulation deviation from empirical is plotted (histogram and map)

c) When all realisations are complete for a fault, Empirical validation for a fault (a) is done automatically

 

a) For all realisations epsilon is calculated using Empirical median and sigma. -Done

b) For all realisations a mean is calculated over all stations in domain - Done

c) Plot 22 histograms to summarise cybershake - Done

d) When a Cybershake run is running, the first realisation of a fault is automatically verified/validated against empirical (b) before running other realisations

Tasks

1) Calculate empirical for each fault for 18p6 - Done

2) Calculate median and sigma for an 18p6 fault - Done

3) Plot median and sigma ratios on a histogram and map plot for AlpineF2K - Will not Do

4) Calculate realisation standard deviation from empirical for all stations for AlpineF2K-HYP01 - Modified to calculating epsilon

5) Plot this on a map and histogram - 0.5 day - modified to plotting histogram of means across all realizations

6) after researcher validation on the workflow, automate this process

a) Calculate rrup and empirical automatically for each fault - mgmt db - 1 day

b) Create new management step for realisation validation and automate it - 0.5 day

c) Create new management step type for fault validation and automate it - 1 day

d) Block running other realisations if the realisation validation didn't pass - 1-2 days

Notes

Workflow

This is an initial trial - no actual automation has been implemented yet.

If the results from this are promising this will be incorporated into the automated management of runs.

Rrup calculation for 1 realisation of each fault

  • Done
  • Computed for HYP01

Empirical calculation for 1 realisation of each fault

  • Issues with python modules on maui / mahuika
  • Needs work to parallelise this effectively. Currently starting 492 processes at once on 40 cores on Maui.
  • Calculations done
  • Aggregation is pending

Ratio calculation for all realisations vs the fault empirical data

  • For loop looping over every IM file and comparing that with the corresponding empirical fault
  • To-Do
  • Generates .xyz file for plotting over a map

Plotting / data analysis

  • No codes exist for data analysis Karim to examine .xyz file and provide guidance
  • Plots for all realisations are not required at this time
  • No labels