Characteristic Fault uncertainty modelling

[Aasha Pancha]

Dear Openquake

I have a source model that includes background point sources and many Characteristic Fault sources.  Having run a disaggregation, I have identified the fault which dominates that hazard.

I would now like to run a new “classical” models in which I test the sensitivity of the hazard level to this fault.

One way of doing this is to run the model three times, in each case varying the values of minMag and occurRates according to the “minimum”, “best estimate”, and “maximum” values, where the geometry remains the same, and comparing the average Uniform Hazard Curves.

    <characteristicFaultSource id="21" name="Major_Fault" tectonicRegion="Active Shallow Crust">

      <incrementalMFD minMag="7.5" binWidth="0.1">

        <occurRates>0.0012</occurRates>

 

However, I was also wanting to add the uncertainty associated with this specific fault directly into the source model logic tree. How do I best do this? Do I make three copies of the source_modle_file.xml, and for each edit the details for this one fault specifying the “minimum”, “best estimate”, and “maximum” values and use a  source_model_logic_tree.xml as follows?

Or is there a more efficient way?

 

  <logicTree logicTreeID="lt1">

    <logicTreeBranchSet uncertaintyType="sourceModel"

                        branchSetID="bs1">

      <logicTreeBranch branchID="b1">

        <uncertaintyModel>source_model_min.xml</uncertaintyModel>

        <uncertaintyWeight>0.25</uncertaintyWeight>

      </logicTreeBranch>

      <logicTreeBranch branchID="b2">

        <uncertaintyModel>source_model_average.xml</uncertaintyModel>

        <uncertaintyWeight>0.5</uncertaintyWeight>

      </logicTreeBranch>

     <logicTreeBranch branchID="b3">

        <uncertaintyModel>source_model_max.xml</uncertaintyModel>

        <uncertaintyWeight>0.25</uncertaintyWeight>

      </logicTreeBranch>

    </logicTreeBranchSet>

  </logicTree>

</nrml>

 

 

[Marco Pagani]

Regarding the effectiveness of the LT you built, it depends on the epistemic uncertainties you model. What you did is general and fine.

--
You received this message because you are subscribed to the Google Groups "OpenQuake Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openquake-use...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openquake-users/a334f646-9af3-47c4-b013-1e642f90f8d9n%40googlegroups.com.

MARCO PAGANI | Seismic Hazard Team Lead | Skype mm.pagani | +39-0382-5169863

GLOBAL EARTHQUAKE MODEL | working together to assess risk

GEM - globalquakemodel.org | T - @GEMwrld | F - GEMwrld

[Aasha Pancha]

Thanks Marco,

As a sensitivity test, I ran  the model three times, in each case varying "occurRates "according to the “minimum”, “best estimate”, and “maximum” values.

As expected, the maximum value gave the lowest hazard, while the “minimum” gave the highest.

But I was wanting to weight these within the same computation of the hazard. I did create three copies of the entire source model, which contains many characteristic faults and point sources, changing the value "occurRates " in each for  just the fault of interest. However, regardless of the weighing applied, the UHS is always the same when just the  “best estimate” is set.

Is this to be expected? Is there a way of modelling the epistemic variation in a single un??? Or are the sensitivity test the most appropriate way to assess the effect of a single fault?

Aasha