Implementation of Nowicki_Jessee_2018

[disotiri1]

Hello everyone!

I am trying to implement the Nowicki_Jessee_2018 model to predict the probability of landslide occurence and spatial extent through openquake v 3.23 (I already read the chat regarding the slope units). I understand that the implementation in openquake is based on the USGS implementation which has some differences with respect to the original paper. 

At the same time, I implement the same model in QGIS, through the raster calculation and by hand, as well. The probabilities computed from Openquake are quite low compared to both QGIS and hand calculations. I matched exactly the openquake result when the landcover coefficient is set to -1.08. However, my input landcover category was 100 and a cofficient of 1.5 should have been considered. 

Could there be a bug in the source code which alwats considers the landcover coefficient equal to -1.08?

[Anirudh Rao]

Dear Dimitris,

The source code for the Nowicki Jessee 2018 model is available here for engine v3.23: https://github.com/gem/oq-engine/blob/v3.23.4/openquake/sep/landslide/nowicki_jessee.py

It's been renamed in the current engine master, but the code is essentially the same: https://github.com/gem/oq-engine/blob/master/openquake/sep/landslide/probability.py

Best regards,

Anirudh

[Catarina Costa]

Hello,

We have looked into the issue and confirmed that there is indeed a bug in the current implementation affecting the handling of the landcover coefficient. As a result, the coefficient was always assigned the default value, as you correctly identified.

A fix has already been implemented in the following pull request:
https://github.com/gem/oq-engine/pull/11534

Thank you for spotting this issue. Best regards,

Catarina Costa

--
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 visit https://groups.google.com/d/msgid/openquake-users/395bb806-0506-41e7-88c6-47d2c22b4294n%40googlegroups.com.

[Dimitris Sotiriadis]

Dear Catarina,

Thank you for addressing this issue!

Best regards,

Dimitris Sotiriadis

[Ridho Alfi Mubarok]

Hi everyone,

Just wanted to add a quick note. I noticed in the lithology_values part of the source code, it seems like the parameter values for pi and vi are swapped compared to the Jessee (2018) paper.

Please correct me if I'm mistaken. Thanks!

[Lana Todorovic]

Dear User,

Thank you for raising an issue. Fixed here: https://github.com/gem/oq-engine/pull/11550

Kind regards,
Lana