Realisation weights in stochastic catalogues.

[Valerio Poggi]

Hi all,

I am interested in generating a stochastic catalogue of ruptures using the event based calculator.

Since my logic tree is complex, I am just using a sample of 1000 branches.

If I have correctly understood, in the events.csv output I get a separate stochastic catalogue for each logic-tree sample (aka realisation). The different events of a set are thus linked to the corresponding realisation through their rlz_ID.

As expected, each set has different number of events. However, it is not clear to me if the number of generated events (per set) is already scaled/proportional to the weight of the corresponding sampled logic tree branch or not.

I assume yes, given that in my case rlz_1 has 274247 events while rlz_999 has 921195 events (thus three times more).

I would be grateful if you could give me confirmation or any advice on a way to check this.

Thank you in advance.

Cheers,

Valerio

[Michele Simionato]

See this section of the manual (and the sections above): https://docs.openquake.org/oq-engine/advanced/rupture-sampling.html#the-difference-between-full-enumeration-and-sampling

Ciao Valerio,

          Michele

[Valerio Poggi]

Grazie Michele! It was super useful! :)

I only do not understand the last disclaimer at the end of the page, where it is indicated that "However, usually there are multiple tectonic region types, and a realization is associated to a tuple of GMPEs".

In my understanding, each realisation should be associated to a separate end branch of the logic tree (either GMM or source). Am I wrong?

Ciao,

Vale

[Michele Simionato]

On Wednesday, November 10, 2021 at 10:46:50 AM UTC+1 vpo...@inogs.it wrote:

Grazie Michele! It was super useful! :)

I only do not understand the last disclaimer at the end of the page, where it is indicated that "However, usually there are multiple tectonic region types, and a realization is associated to a tuple of GMPEs".

In my understanding, each realisation should be associated to a separate end branch of the logic tree (either GMM or source). Am I wrong?

No, we are saying the same thing. In case of a purely GMPE logic tree specifying an end branch of the logic tree means exactly specifying a tuple of GMPEs.

    Michele