Source time function for plane wave

[tianzeliu]

Hi There,
I was trying to customize the source time function for an incident plane wave in SPECFEM2D, but found that the option time_function_type=8 does not seem to work. No matter what source time function I specify in the file, the source time function is always a Ricker wavelet. Is that a bug or I was not using the correct parameter? Thanks.

[komatits]

Hi Tianzeliu, Thanks! Let me cc Alexis and Eric here, who successfully use it, as well as Bence. It could be that time_function_type=8 is not implemented for a plane wave, but it could likely be added (you would need to do an FFT of your external source and use it in routines ./src/specfem2D/paco_beyond_critical.f90 and ./src/specfem2D/paco_convolve_fft.f90, which implement the plane wave (they were written by Paco Sanchez-Sesma from UNAM in Mexico, and I think he uses the analytical Fourier transform of a Ricker, you could easily change that to the numerical FFT of your external source). If you add that please send it to us and we will make it part of the official source code and list you as an author / contributor in the SPECFEM2D manual. Thank you, Best regards, Dimitri.

…

On 07/11/2017 06:59 AM, tianzeliu wrote: Hi There, I was trying to customize the source time function for an incident plane wave in SPECFEM2D, but found that the option |time_function_type=8| does not seem to work. No matter what source time function I specify in the file, the source time function is always a Ricker wavelet. Is that a bug or I was not using the correct parameter? Thanks. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#803>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AFjDKeEKb-Pgajz4YxzX9qcMqOkolk1lks5sMwEqgaJpZM4OTur9>.

-- Dimitri Komatitsch, CNRS Research Director (DR CNRS) Laboratory of Mechanics and Acoustics, Marseille, France http://komatitsch.free.fr

[komatits]

Hi Tianzeliu, I have checked again, the solution is simpler than I thought, you can just enter your external source time function in the time domain in src/specfem2D/paco_beyond_critical.f90 and src/specfem2D/paco_convolve_fft.f90 instead of the current Ricker (which is indeed hardwired, and thus you are right, option time_function_type=8 is ignored in that case). To see where to do this, type this: grep -n --color=always -i ricker ./src/specfem2D/paco_* (please first type "git pull" because some old comments were in French, I just translated them to English). If you implement that in a flexible way then please send it to us and we will put it in the official version of the code. To see how to implement time_function_type == 8 in a flexible way you can go to line 335 of ./src/specfem2D/prepare_source_time_function.f90 (of the "devel" branch of SPECFEM2D) Thank you, Best regards, Dimitri.

…

On 07/11/2017 04:19 PM, Dimitri Komatitsch wrote: Hi Tianzeliu, Thanks! Let me cc Alexis and Eric here, who successfully use it, as well as Bence. It could be that time_function_type=8 is not implemented for a plane wave, but it could likely be added (you would need to do an FFT of your external source and use it in routines ./src/specfem2D/paco_beyond_critical.f90 and ./src/specfem2D/paco_convolve_fft.f90, which implement the plane wave (they were written by Paco Sanchez-Sesma from UNAM in Mexico, and I think he uses the analytical Fourier transform of a Ricker, you could easily change that to the numerical FFT of your external source). If you add that please send it to us and we will make it part of the official source code and list you as an author / contributor in the SPECFEM2D manual. Thank you, Best regards, Dimitri. On 07/11/2017 06:59 AM, tianzeliu wrote: > Hi There, > I was trying to customize the source time function for an incident > plane wave in SPECFEM2D, but found that the option > |time_function_type=8| does not seem to work. No matter what source > time function I specify in the file, the source time function is > always a Ricker wavelet. Is that a bug or I was not using the correct > parameter? Thanks. > > — > You are receiving this because you are subscribed to this thread. > Reply to this email directly, view it on GitHub > <#803>, or mute the > thread > <https://github.com/notifications/unsubscribe-auth/AFjDKeEKb-Pgajz4YxzX9qcMqOkolk1lks5sMwEqgaJpZM4OTur9>. > >

-- Dimitri Komatitsch, CNRS Research Director (DR CNRS) Laboratory of Mechanics and Acoustics, Marseille, France http://komatitsch.free.fr

[loycbt]

Dear Dimitri,

According to your suggestion, I changed the ricker wavelet in../paco_convolve_fft.f90 (Line 165 to 251) to the external time_function, but the source function is still rickerwavelet.

[bottero]

Hi Loycbt, For generating a plane wave a convenient hack is to use a line of (hundreds of) punctual sources whose amplitude is apodized near the edges. Check it out, Best regards, Alexis Bottero Marine Physical Laboratory Scripps Institution of Oceanography (+1) 619 677 9887 (+33) 695 170 097 Le lun. 4 mai 2020 à 08:15, loycbt <notifications@github.com> a écrit :

…

Dear Dimitri, According to your suggestion, I changed the ricker wavelet in../paco_convolve_fft.f90 (Line 165 to 251) to the external time_function, but the source function is still rickerwavelet. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#803 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABVY5YX7PUXX5O6LKXELAR3RP3LYJANCNFSM4DSO5L6Q> .

[loycbt]

Dear Bottero,

Thank you very much.

Best regards.