[CIG-SEISMO] Reflections off of C-PML Boundaries in Specfem3D Cartesian

[Ian Stone]

Hello,

I am attempting to simulate an earthquake using Specfem3D, but keep running
into issues with reflections off the absorbing boundaries, even with C-PML
layers along the edges. I was wondering if anyone else had experienced a
similar problem and had recommendations on how to fix it.

Some information about my simulation:

   - The x/y/z dimensions of my model area are 15X12X9.9km, respectively.
   The mesh has 2 triplication layers, transitioning from 300m to 100m, and
   from 100m to 33.3m near the surface. To accommodate the PML layers, I
   simply extruded the mesh 2400m in each direction (except for the free
   surface); so on each side, there are 8, 300m-thick PML layers.
   - I use a CMTSOLUTION for a M4 thrust earthquake as a source. In the
   Par_file, I set f0_FOR_PML=1.0Hz, as this seems to be the frequency
   range with the most energy.
   - The reflections off the absorbing boundaries are most prominent with
   the longer period S-waves, though other waves and periods do produce some
   smaller reflections as well. Reflections are generated regardless of
   where the source is located in the model.
   - I have checked to make sure all the min/max absorbing boundary files
   correspond to the actual outer surface of the model, and not the
   PML/non-PML boundary. Everything looks ok.

I have attached a few screenshots of my model setup. The first is a
small-scale version of my large mesh, showing 4 extruded layers on each of
the bounding surfaces. The last 3 images show the cpml damping profile in
the x, y, and z directions, respectively; these were generated using the
.vtk files output during the simulation.

Thank you,

Ian Stone
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.geodynamics.org/pipermail/cig-seismo/attachments/20180809/608d5603/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: small_mesh_pml.png
Type: image/png
Size: 120810 bytes
Desc: not available
URL: <http://lists.geodynamics.org/pipermail/cig-seismo/attachments/20180809/608d5603/attachment-0004.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: cpml_damping_x.png
Type: image/png
Size: 214340 bytes
Desc: not available
URL: <http://lists.geodynamics.org/pipermail/cig-seismo/attachments/20180809/608d5603/attachment-0005.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: cpml_damping_y.png
Type: image/png
Size: 235473 bytes
Desc: not available
URL: <http://lists.geodynamics.org/pipermail/cig-seismo/attachments/20180809/608d5603/attachment-0006.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: cpml_damping_z.png
Type: image/png
Size: 234536 bytes
Desc: not available
URL: <http://lists.geodynamics.org/pipermail/cig-seismo/attachments/20180809/608d5603/attachment-0007.png>