Multi-rupture scenarios - can't get consistency with single scenario

[Malcolm Haylock]

Until now I have run the full ISC-GEM catalog in OQ using individual scenario runs. I'm moving over to using the new multi-rupture scenario feature (thanks for the addition!) but am having trouble reproducing the GMFs though. I've based my conversion on the OQ case study no. 13.

The initial runs were done using an XML rupture file. Here's the one for Tohoku (2011):

<?xml version="1.0" encoding="utf-8"?>

<nrml xmlns="http://openquake.org/xmlns/nrml/0.5" xmlns:gml="http://www.opengis.net/gml">

<singlePlaneRupture>

<magnitude>

9.09

</magnitude>

<rake>

90

</rake>

<hypocenter depth="20" lat="38.285" lon="142.546"/>

<planarSurface>

<topLeft depth="3.261293" lat="40.08489" lon="144.1628"/>

<topRight depth="3.261293" lat="36.09552" lon="142.7041"/>

<bottomLeft depth="36.73871" lat="40.47426" lon="142.3781"/>

<bottomRight depth="36.73871" lat="36.464" lon="141.0079"/>

</planarSurface>

</singlePlaneRupture>

</nrml>

I run this specifying explicitly the GMPE to use (using gsim).

 Converting this to a ruptures csv gives:

#,,,,,,,,,,"trts=['Stable', 'Volcanic', 'Subduction', 'Active']"

"serial","mag","rake","lon","lat","dep","multiplicity","trt","kind","mesh","extra"

16461282,9.09,90,142.546,38.285,20,1,"Subduction","BaseRupture PlanarSurface","[[[[144.1628,142.7041,142.3781,141.0079]],[[40.0849,36.0955,40.4743,36.464]],[[3.2613,3.2613,36.7387,36.7387]]]]","{""occurrence_rate"": 1e-03}"

And I run this with a single-branch GSIM logic tree to specify the GMPE as for case study 13.

The problem is that for the first implementation I could run a single simulation (number_of_ground_motion_fields=1) and get a smooth field. When running as a multi-rupture case with a single simulation I get a noisy field, which perhaps is sampling the uncertainty somewhere. Is there a way I can run a single simulation using the multi-rupture scenario implementation without any uncertainty sampling? 

[Michele Simionato]

Are you using version 3.11? If I remember correctly I changed something in current master, you should check if the current master works as you would expect. Otherwise, please send me the job.zip  archive of the calculation you are running.

                      Michele