query on logic tree weights and realizations

[era]

Hello!

Hopefully this is not a completely ridiculous set of queries - we have been experimenting a bit with the ways we weight our GMPE logic tree branches and haven't been able to find concrete answers to some of the results we're seeing.

1. We are running a series of sensitivity tests, and when we set branches to zero (out of laziness), as in the attached logic tree, all realizations are still computed. Why is this? 

2. According to realization csv files produced by hazard runs with this logic tree, the weight of GMPE combinations that include a branch with a weight of zero is zero. So, we can safely assume that, even though the realization was produced, it is not being included in the mean result (ie. hazard curve)? (We have done no additional hazard runs to test this, but are happy to some if they would be useful).

3. The primary calculation mode we have been running these sensitivity tests in has been disaggregations. It appears that it is not possible to get a 'mean' disaggregation hazard automatically using the engine. Is this the case? In our search through manuals, etc. to find the answer, we came across this post in the user group, but the conclusion is somewhat unclear to us. If we cannot output a mean disaggregation through the engine, can we combine the appropriate results manually?

Please also excuse some of the classes called in the logic tree - we are working with some GMPEs the we have modified to be more consistent with various New Zealand design standards. 

Cheers,

Elizabeth and Matt

[Michele Simionato]

Hello Elizabeth and Matt,

unfortunately I am on vacation right now, so I cannot look at the points you raised in detail, however I will notice that there not a single test in the engine code base for the case of zero weight of a realization. The behavior there is currently undefined, you should not assume that a zero weight will discard that realization. The numbers are probably correct, but the realization will still appear in the file realizations.csv. What we do, when we want to remove a realization, is just to edit the logic tree file and comment out the realization we do not want, adjusting the weights on the other realizations so that the sum is 1.
HTH,

                                Michele