CIG-LLNL Computational Seismology Workshop

Table of Contents

1. Overview
2. Introduction
3. Part I: Setting up SPECFEM3D_GLOBE
4. Part II: Continental-scale Simulations
5. Part III: Visualization
   1. Visualize Seismograms
   2. Visualize Model Mesh
   3. Visualize Surface Movie
6. Part IV: Adjoint Simulations (Bonus)
7. Further Work
8. Resources

Part III: Visualization

In this third part of the tutorial, we will look at methods for visualizing the output simulation data. We will look at the following categories:

• Visualize Seismograms
• Visualize Model Mesh
• Visualize Surface Movie

Visualize Model Mesh

In order to visualize the model mesh, we need to modify the Par_file such that the mesh data are being saved. Therefore, open up DATA/Par_file and set the following parameter to .true.:

  ...

  # save mesh files to check the mesh
  SAVE_MESH_FILES                 = .true.

  ...

Once we have updated this parameter, we need to go through the following three steps in order to visualize the model mesh:

1. Rerun the mesh routine
2. Combine the output mesh data
3. Visualize using Paraview

• Step 1: Rerun the mesh routine

  This step is identical to what we did in Part II. We run the xmeshfem3D executable on the cluster by submitting the submit_mesher script:

      sbatch submit_mesher

And use squeue to monitor the job. Once the job finishes, we can check the OUTPUT_FILES/output_mesher.txt file to make sure that the mesher ran successfully, and then we can move on to the next step.

• Step 2: Combine the output mesh data

  Once we have rerun the mesh routine with the SAVE_MESH_FILES parameter set to true, the mesh files will be output to the DATABASES_MPI/ folder. The data is partitioned among the processees, and we will want to combine these data into single files.

  We can achieve this by running the combine_data.py script that is located in the root folder. This script will generate VTK files for the vp and vs model velocities which will be stored in the OUPUT_FILES/ folder.

  NOTE: Under the hood, the combine_data.py script is running the xcombine_vol_data_vtk executable that is located in the bin/ folder. For details on how to run this executable directly, inspect the combine_data.py script and consult the SPECFEM3D_GLOBE manual.

• Step 3: Visualize using Paraview

  In order to start Paraview, we first need to load the module

        module load paraview
  

  Then run the following command in the terminal to launch the program

        paraview
  

  Once Paraview is running, click the Open button in the top left corner, and navigate to the OUTPUT_FILES/ folder. In this folder you should see two file-bundles called reg_.._vp.vtk and reg_.._vs.vtk, which are the combined data files that we produced in the previous step. In order to visualize the vp velocity model, press and hold the ctrl key and select the three files reg_1_vp.vtk, reg_2_vp.vtk, reg_3_vp.vtk, and then click OK.

  Now that we have loaded the model, click Apply on the left-hand side in the Properties menu, and the model will appear.

  In order to see the model mesh, select the three model files in the Pipeline-Browser by holding down ctrl. Then go to the top menu and select Filter -> Common -> Group Datasets, and click Apply in the Properties menu. With GroupDatasets1 selected in the Pipeline-Browser go to the drop-down menu which says Surface and select Surface With Edges.

  The procedure is similar for visualizing the vs velocity model.

Visualization of the model mesh with the vp component.

In this section, we have looked at how to visualize the model mesh with different components (e.g. vp, vs).

In the next section, we will look at how to create a surface movie for the simulation data.

Previous section – Next section