Question about Time-Dependent Prescribed Boundary Velocities in Regional Models

ASPECT
1

Hello,

I am currently studying the tutorial on Prescribing Velocities from Global Convection Models as Boundary Conditions in Regional Models — ASPECT 3.1.0-pre. and I have a question regarding time-dependent boundary velocity input.

Actually, I have already successfully run the example model. The regional_velocity_files directory contains four files (e.g., chunk_3d_east.0.txt), which appear to provide velocity data only for the initial time step (t = 0).

subsection Boundary velocity model
  set Tangential velocity boundary indicators = outer
  set Prescribed velocity boundary indicators = south:ascii data, west:ascii data, east:ascii data, north:ascii data
  subsection Ascii data model
    set Data directory = $ASPECT_SOURCE_DIR/contrib/python/scripts/regional_velocity_files/
    set Data file name = chunk_3d_%s.%d.txt
  end
end

My question is: How can I extend this setup to provide time-varying boundary velocities over the full duration of a long simulation (e.g., end time = 200e3 years)?

Specifically:

  • Should I generate multiple data files with different time step indices (e.g., chunk_3d_east.0.txt, chunk_3d_east.1.txt, etc.)?

  • Or, could I instead use a single file (e.g., rename chunk_3d_east.0.txt to chunk_3d_east.txt)

Any help is appreciated.

Thanks,

Haolin

2

Hi @XHL-CUG,

Thank you for posting your question to the forum.

Here are a few links to documentation and examples on how to implement time-dependent velocity boundary conditions with the ascii data approach:

Please let us know if you have additional questions after reviewing this information.

Cheers,

John

3

Hi John,

Thank you very much for your fast reply!

Sorry for my delayed response. Thank you for providing the links. I think I have solved the problem, but a new one has arisen.
I’d like to use the boundary velocity (calculated by extract_local_velocity.py) in a box model. To do this, do I only need to convert the coordinates, or do I also need to transform the three velocity components? I’m a bit unsure how to handle the vector components correctly during this transformation.

Thanks,

Haolin

4

Hi @XHL-CUG,

Great to see someone is already using extract_local_velocity.py!

To your question, I will need to think a bit about what the best approach is to extract and project data from a spherical model onto the boundaries of a cartesian model. You could in theory do this all within ParaView or as an additional post processing (projection) step.

@daniel.douglas - Thoughts on this?

Cheers,

John