Reusing ruptures with different Vs30 GMPETable using --hc

[Farzaneh]

Hello OQ users,

I am running an event-based hazard calculation for Eastern Canada using the CanadaSHM6 model. My GMPEs are all GMPETable (pre-computed HDF5 files) calibrated at specific Vs30 values (e.g. vs760, vs450, vs580, vs2000). Since the tables are pre-computed at fixed Vs30, I believe OpenQuake cannot apply site amplification automatically from the site model file. the reason is that I can input only one file for gsim_logic_tree_file. For example when I input Vs30_450.xml, it refers to a specific table:

<logicTreeBranch branchID="b11">

<uncertaintyModel gmpe_table="../tables/CanadaSHM6-trial/stablecrust/CanadaSHM6-trial_NGA-East_Model_01_AA13_sigma.vs450.hdf5">GMPETable</uncertaintyModel>

<uncertaintyWeight>0.0757</uncertaintyWeight>

<uncertaintyWeight imt="PGA">0.0757</uncertaintyWeight>

<uncertaintyWeight imt="SA(0.1)">0.07725</uncertaintyWeight>

<uncertaintyWeight imt="SA(0.2)">0.06505</uncertaintyWeight>

<uncertaintyWeight imt="SA(0.3)">0.07755</uncertaintyWeight>...

My sites have varying Vs30 values across the region. 

What I tried: I ran a parent job at Vs30=760 which completed successfully. I then tried to run a child job with a different gsim_logic_tree_file (Vs30=450) and different site file using --hc <parent_job_id>. The child job runs but only outputs Full Report, Realizations, and Site Model - no GMFs, even though I have ground_motion_fields = True.

My questions:

Does --hc support reusing ruptures for a hazard-to-hazard calculation with a different GSIM logic tree?
If yes, what is the correct child job ini file configuration?
If not, what is the recommended approach for handling sites with different Vs30 values when using GMPETable HDF5 files?

Thank you for your help