Author Topic: Extracting simulated IM fields from R2D  (Read 14599 times)

Anne Hulsey

  • Jr. Member
  • **
  • Posts: 70
    • View Profile
Extracting simulated IM fields from R2D
« on: September 16, 2021, 10:46:31 PM »
Hello,

I would like to use the R2D Hazard tool to simulate IM maps, rather than select ground motions. Additionally, I would like to be able to upload openquake files to identify the rupture scenario and include a gmpe logic tree.

I think the most important feature is being able to extract the predicted median and std dev at each location (perhaps a vector of them, representing the realizations of the logic tree). Is it possible to export this in the same format as EQHazard? Actually, even better than that, to export it as the .h5 files that are used in https://github.com/annehulsey/seaturtles/? (See the extract_eqhazard_data function in the main.py file.)

Beyond that, it could also be nice to run the whole simulation in R2D, then extract the IM maps.

Thanks,
Anne

steveg

  • Newbie
  • *
  • Posts: 22
    • View Profile
Re: Extracting simulated IM fields from R2D
« Reply #1 on: September 17, 2021, 01:15:12 AM »
Hi Anne,

In the current version of R2D you can avoid ground motion selection by selecting the None option in the Record Selection->Database dropdown. The resulting csv file will be an IM field. Openquake import functionality is coming soon in the next release of R2D. I will talk to Kuanshi about including a GMPE logic tree and export of the median/std. dev. at each location (in .h5 format alongside the csv) and get back to you. 

Best,
Steve

Anne Hulsey

  • Jr. Member
  • **
  • Posts: 70
    • View Profile
Re: Extracting simulated IM fields from R2D
« Reply #2 on: September 17, 2021, 05:31:08 PM »
Hi Steve,

Thanks for pointing out the current feature! The SiteIM.json just has a single string of lnSa values for each site - are these the predicted medians? If so, it would be great to also see the standard deviations in that file. (I also see the multiple IM values listed in each site.csv, just looking for clarity on what the SiteIM.json is.)

Also, I was surprised by the zTR values in the EventGrid.csv file that it produced (attached). Do you know what would explain the zero values?

Thanks,
Anne
« Last Edit: September 17, 2021, 05:38:51 PM by Anne Hulsey »

kuanshi

  • Jr. Member
  • **
  • Posts: 56
    • View Profile
Re: Extracting simulated IM fields from R2D
« Reply #3 on: September 17, 2021, 06:32:19 PM »
Hi Anne,

The SiteIM.json is designed for saving the final individual realizations of ground motion spectra - as median and standard deviations are actually not really used by later building response/damage/loss simulations.  But if you're interested, I could save them to an individual file under the same directory.

And the zTR is the depth to the rock - zero indicates a rock site, while it is exported for site-response simulation, and not actually used in "Earthquake Scenario Simulation" - sorry about the confusion!

Best,
Kuanshi

Anne Hulsey

  • Jr. Member
  • **
  • Posts: 70
    • View Profile
Re: Extracting simulated IM fields from R2D
« Reply #4 on: September 17, 2021, 09:12:16 PM »
Thanks, Kuanshi. I don't need the individual file at the moment but, if we are able to get R2D set up to run OQ input for rupture simulations, retrieving the medians and std dev would be a very helpful feature. (i.e., as described at the beginning of the thread)

Anne