ar4
p_sv
I have conducted P-SV wave simulations using both Deepwave and PSVwave, but I found that the results from Deepwave are not as good as those from PSVwave. I'm not sure why this is the case, and it's possible that I haven't used Deepwave correctly. Could you provide me with a tutorial on how to perform P-SV wave simulations using Deepwave?
Hello and thank you for your message. I'm sorry to hear that you were not satisfied with the Deepwave results. I am glad that you took the time to contact me so that I can try to help you get better results. At the moment I only have this example of using the elastic propagator, but I would be happy to assist you myself. From the gathers that you provided, it looks as if there might be some numerical dispersion. This can be caused by the grid spacing being too large. Would it be possible for you to try running it again with a smaller grid spacing? If that doesn't resolve the problem, are you able to provide the code and perhaps also the models that you used to generate the data so that I can look to see if I notice any problems?
ok, I will provide the code and the models Uploading deepwave_forward_elastic_saveSAC.py…
It unfortunately does not seem to have uploaded correctly. If you want to try something else yourself, it looks to me like the sources and/or receivers might possibly be oriented in the other direction to those in your psvWave results. Do you think that might be the case? Do you have any questions about source/receiver orientation in Deepwave, or is the documentation already sufficiently clear?
Thank you for sending the code and models. As I suggested above, your grid spacing is too large and your source appears to be oriented in the wrong direction. The code you sent uses a grid spacing of 100m. For a frequency of 10Hz and a minimum velocity of 1000m/s, a spacing of 10m is more appropriate. In your code, the second dimension of your model is depth, however you used a source oriented in the y dimension, which in Deepwave is the first dimension. If you wish to use a source oriented downwards with this model, you should use source_locations_x and source_amplitudes_x. The velocity files that you provided also have regions where vs is negative. That will result in numerical instability. Getting reasonable results from a wave propagator requires great care, making sure that you understand all of the parameters of the code and how finite difference wave propagation works, carefully checking your models, and diligent investigation if the results are not as expected.
I'm going to close this issue now, but please feel free to reopen it or to create a new one if you have any other questions.