Question about landslide tsunami modelling in NHWAVE and FUNWAVE

[Francisco Uribe]

Dear FUNWAVE group,
I'm using NHWAVE (I'm a novice in tsunami modeling) to model a landslide generated tsunami in a fjord, and I have big questions in order to carry out the simulation. I'm not sure, but based on what I have read on the internet (user manuals and forums) I have to do the simulation in NHWAVE for wave generation, and FUNWAVE for wave propagation. However, there are some things about the modeling that I'm not clear about, and I hope that with your help I can solve these questions.
- How large should the spatial discretization of the grid be? (the fjord is about 13 km long, and about 500 meters wide).
- Which flags should be activated in FUNWAVE (after having done the simulation in NHWAVE)?
- How to do the simulation in FUNWAVE from the results generated by NHWAVE?
Thank you!
Francisco

[fyshi...@gmail.com]

Francisco

The main reason for using the combined NHWAVE-FUNWAVE approach is to save computational costs. NHWAVE alone can do the work as long as your HPC can afford the job. Another reason is that FUNWAVE has better wetting-drying performance, which is good for modeling coastal inundation. Unfortunately, we don’t have detailed documentation for the coupling procedure. You may find some clues in the published work. Here are brief answers to your questions

- How large should the spatial discretization of the grid be? (the fjord is about 13 km long, and about 500 meters wide).

NHWAVE should be performed in a smaller region covering landslide and wave generation activities. The grid should resolve the slide and generated waves. I think ~10 m may be good for your case. The FUNWAVE grid should cover the entire domain with a similar or smaller grid size if you want to resolve the detailed geometry of the fjord. 

- Which flags should be activated in FUNWAVE (after having done the simulation in NHWAVE)?

If you use a Cartesian coordinate, set -DCARTESIAN, other flags are the same as a wind wave simulation

- How to do the simulation in FUNWAVE from the results generated by NHWAVE?

A snapshot result (eta ,u, v) from NHWAVE should be mapped into the FUNWAVE grid, and use the option of initial condition in FUNWAVE: INI_UVZ = T. An example can be found in the FUNWAVE-TVD package: /simple_cases/tohoku_tsunami/

Hope the answers help. 

-Fengyan

[Francisco Uribe]

Thank you very much Mr. Shi for the reply, you have been a great help. However, I still have quite a few doubts. If an event has 3 large landslides that generated a tsunami in a fjord, can I do the modeling separately (of those 3 landslides) or should I do the modeling of the 3 events at the same time (with the high computational cost involved)?
If I have a subaerial landslide, how do I get the NHWAVE software to consider the height of the centroid of the landslide?

As far as I see, the input files are .txt, which are not the format of the output files of NHWAVE , do I have to make a conversion for them to be read by FUNWAVE? (just change the format of the last output file?)

Thank you!

Francisco

[Francisco Uribe]

I think I'm making progress in NHWAVE-FUNWAVE modeling, as I obtained the output files in NHWAVE, then used those files as the initial condition for FUNWAVE. However, regardless of the inputs I use the simulation always stops at time = 117.387367, and I don't know why, because I thought that the inputs would generate numerical instability, but I modified the values to those recommended by the manual, but still the simulation is not carried out. I attach a screenshot as a sample.

I appreciate the help provided, and the time you have taken to respond to my answers.