Best software for seismic refraction tomography

[resour]

For multiscale tomography using Steepest Descent method for SAGEEP11 blind refraction data see

http://rayfract.com/tutorials/sageep11_smooth50.pdf

For multiscale tomography using Conjuate Gradient method for SAGEEP11 data see

http://rayfract.com/tutorials/sageep11_16.pdf

and final tomogram after 8 WET runs shown below :

[GeoALG]

ZondST2D is a very nice software for seismic refraction. I have used some old data with 7 shots. It gives excellent results. He Alex can you give me a temporary license to test it in full version?

[pitta]

I would like to point out a new software for seismic refraction tomography called SmartTomo. SmartTomo exploits an optimized implementation of the Moser method, ( “Shortest path calculation of seismic rays.” Geophysics 56.1 (1991): 59-67), and optimizes the results through the Simultaneous Iterative Reconstruction Technique.
More info can be found at the project homepage https://www.smarttomo.com
I tested the software with the "3layers.txt" dataset and with 20 iteration I got the following result:

The software was also tested using the SAGEEP11 dataset. The following image show the resulting tomographic profile.

It takes about 400 seconds to process the 101 sources recorded using 100 receivers of this dataset.

[resour]

Here our updated SAGEEP11 interpretation now with our new WDVS Wavelength-Dependent Velocity Smoothing :

Multiscale WET inversion with 1D-gradient starting model and WDVS at 60Hz took 22 minutes on Apple 2020 macMini.

For theory on WDVS see (Zelt and Chen 2016) at

https://www.researchgate.net/publicatio ... ction_data

We show more processing details and alternative interpretations in

https://rayfract.com/tutorials/sageep11_16.pdf

[esrasur]

I like SeisOpt -> requires no prior assumptions of subsurface structure, or any other subjective data, as input.

[kaminae]

This "subjective" data converts geophysical results to geology. So I can't say it is kind of advantage. What about equivalence, local minimums of solutions, undercoverage and data errors(we don't know BTW)? Actually first prior assumption made in inversion is model smoothness(model class).