exporting rupture data Event Based

[Céline Beauval]

Dear developers,

I have a question about event based PSHA outputs.

running openquake with event based PSHA, the results obtained are

gmf*.xml
hazard_curve*.xml
hazard_map*.xml

to get the rupture output, I have to use the commands
 oq engine --lo id
 oq engine --eo id .

and I get a rup_data_grp*.csv file

My questions are
- is there a readme for this csv file ?
- is it possible to obtain this rupture file in another format ?

I was used to an older version of openquake, where a file 'ses-smltp_b1-gsimltp_b1.xml' was generated as output,
but apparently this has changed.

Many thanks for your help,

Céline

[Michele Simionato]

The old ses output is now called 'ruptures'. It is generated only if in the job.ini you set save_ruptures= true.
You can export it in .xml or in .csv.
rup_data contains other information that you may find useful or not ;)

[Céline Beauval]

Michele,
thank you.
There is no where we can find the detail of whats is listed in the .csv file? We might be interested in these details.

Another question, when using a large logic tree with many optional GMPEs, it would be nice that the name of the output file reflects the branches explored (the gsimTreePath, e.g. "b11_b21_b31").

Is it possible to generate output files with such names?
just in case ...
thank you
Céline

[Michele Simionato]

Il giorno giovedì 9 febbraio 2017 17:42:29 UTC+1, Céline Beauval ha scritto:

Michele,
thank you.
There is no where we can find the detail of whats is listed in the .csv file? We might be interested in these details.

The rup_data output is left undocumented on purpose, since it may change and even disappear in future versions of the engine. The headers, however, should be telling. For instance, if I run our
event based demo I get a rup_data file like the following:
 

rupserial,multiplicity,numsites,occurrence_rate,mag,lon,lat,depth,strike,dip,rake,boundary
2,1,1,3.965800E-05,5.050000E+00,0.08456,0.15503,5.00000,4.500000E+01,9.000000E+01,0.000000E+00,((0.07295 0.14342,0.09617 0.16664,0.07295 0.14342,0.09617 0.16664,0.07295 0.14342))
244,1,1,3.965800E-05,5.050000E+00,-0.00537,0.11007,5.00000,1.350000E+02,9.000000E+01,0.000000E+00,((-0.01698 0.12168,0.00624 0.09846,-0.01698 0.12168,0.00624 0.09846,-0.01698 0.12168))
...

`rupserial` is an unique integer ID for the rupture. Multiplicity is how many times the rupture occurs in the investigation period (usually 1). `num_sites` is how many sites are affected by the rupture.
`occurrence_rate` is the occurrence rate, `mag` is the magnitude, `lon`, `lat` and `depth` is the position of the centroid of the rupture, `boundary` is the geometry of the rupture (as a MULTIPOLYGON, you same that you can find also in the ruptures.csv files). Then there are other parameters, in this case `strike`, `dip` and `rake` but they may vary depending on the tectonic region type and the GSIMs used.

Another question, when using a large logic tree with many optional GMPEs, it would be nice that the name of the output file reflects the branches explored (the gsimTreePath, e.g. "b11_b21_b31").

Is it possible to generate output files with such names?

 

This was the behavior of the engine a long time ago. We removed that feature because in the case of extra-large logic trees the name of the files could exceeds the Unix limit of 256 characters (and this happened), So now the filename contains only the realization number. However, there is also an output called 'realizations': when exported, you get a csv file telling you the gsimTreePath of each realization. For instance, in
the case of our logic tree demo, you get something like this:

$ echo realizations.csv
ordinal,uid,model,gsim,weight
0,b11_b21_b31_b41_b51~b11_b21,source_model.xml,BooreAtkinson2008() ToroEtAl2002(),0.00307409
1,b11_b21_b31_b41_b51~b11_b22,source_model.xml,BooreAtkinson2008() Campbell2003(),0.00307409
2,b11_b21_b31_b41_b51~b12_b21,source_model.xml,ChiouYoungs2008() ToroEtAl2002(),0.00307409
3,b11_b21_b31_b41_b51~b12_b22,source_model.xml,ChiouYoungs2008() Campbell2003(),0.00307409
...

I am not sure it the format of the realizations.csv file is documented; if not, please feel free to open a documentation bug ;-)
HTH,

              Michele

[Céline Beauval]

thanks a lot for your answer
Céline

[Michele Simionato]

Le vendredi 10 février 2017 07:44:24 UTC+1, Michele Simionato a écrit :

Il giorno giovedì 9 febbraio 2017 17:42:29 UTC+1, Céline Beauval ha scritto:

Michele,
thank you.
There is no where we can find the detail of whats is listed in the .csv file? We might be interested in these details.

The rup_data output is left undocumented on purpose, since it may change and even disappear in future versions of the engine.

Dear Céline,

I can confirm that the rup_data output will disappear in the next version of the engine. If you need that information for some purpose, please let us know.

The information will still be available in the datastore.

               Michele 

[Hyeuk Ryu]

Hi Michele,

I was trying to pick a scenario from output of even-based PSHA using v3.5, and noticed that MULTIPOLYGON contains only lat and lon while planarSurface requires lon, lat, and depth.

Would you be able to let me know where I can find the depth or any formula that I can use to calculate the depth?

Thank you very much.

Regards,

Hyeuk

[Michele Simionato]

On Wednesday, July 10, 2019 at 1:38:36 PM UTC+2, Hyeuk Ryu wrote:

Hi Michele,

I was trying to pick a scenario from output of even-based PSHA using v3.5, and noticed that MULTIPOLYGON contains only lat and lon while planarSurface requires lon, lat, and depth.

Would you be able to let me know where I can find the depth or any formula that I can use to calculate the depth?

Current master of the engine already exports the depth information, Just upgrade to master or wait for release 3.6 of the engine, due next week.

         Michele

[Hyeuk Ryu]

Thanks for your kind reply. I managed to get it by exporting the output in xml format though.

Regards,

Hyeuk

[Hermon Alfaro]

Hi Michele, How can I in latest version of the openquake have the ocurrence rates of ruptures? I cannot find them in the ruptures.csv

Thanks a lot,

Hermon

[Michele Simionato]

On Wednesday, January 29, 2020 at 4:06:16 PM UTC+1, Hermon Alfaro wrote:

Hi Michele, How can I in latest version of the openquake have the ocurrence rates of ruptures? I cannot find them in the ruptures.csv

You can get more information about the ruptures with the command

$ oq show ruptures CALCID

or

$ oq show ruptures CALCID > theruptures.csv

 

if you want to save the info into a .csv file.

HTH,

             Michele