[CIG-SEISMO] need some advices to learn the SPECFEM2D codes

[Carl Tape]

Dear Guoping:

My suggestion would be the following (if you haven't done them already):

1. Read the SPECFEM2D user manual. It is up-to-date and contains many
references.
2. Read some of the references listed in the manual.
3. Run one of the several examples that are provided with the code,
but note that only some of the examples contain the adjoint
capabilities.
4. Try stepping through the components of the code. The key point is
that the forward solver is used both to evaluate the misfit function
and to compute the sensitivity kernels. (I agree that it would be nice
if notation in published papers matched variable names in codes, but
this is generally not the case.)

SPECFEM2D can produce the gradient of the misfit function for a
tomographic inversion, but options for using the gradient within an
iterative inversion are left to the user (e.g., conjugate-gradient,
steepest descent). The plan is to include some examples in the future.

Hope this helps.

Carl

---------------------------------------------------------------
Carl Tape
Assistant Professor
Geophysical Institute
University of Alaska Fairbanks
Phone: 907-474-5456
Email: carltape at gi.alaska.edu
Web: http://www.giseis.alaska.edu/input/carl/
---------------------------------------------------------------

On Wed, Dec 14, 2011 at 7:39 AM, Guoping Chen, Ph.D.
<guoping.chen at duke.edu> wrote:
> Hi,
>  I am studying the SPECFEM2D codes, and very intersting in the adjoint algorithm. But then I finish read the "Finite-frequency tomography using adjoint methods—
> Methodology and examples using membrane surface waves" written by Carl Tape in 2007, and "Seismic tomography, adjoint methods, time reversal and banana-doughnut kernels" written by Jeroen Tromp in 2005, then tracing the SPECFEM2D codes, I am lost in the large program lines due to my poor Fortran skill and lack enough comments in the program codes.
>  Especially, I can not understand why the SPECFEM2D codes using like "rhop_ac_kl, rho_ac_kl" to define the phase-speed or density kernel, because them not dispear neither in Carl Tape paper published in 2007 nor in Jeroen Tromp paper published in 2005. That means in all your prior published papers, the kernel defined using K as the abbreviate for the kernel, why in program codes using "tho" as the pre-fix?
>
>  To sum up, would your can give me some advices to how to learn the algorithm introduced in the published papers and learn the implement through the SPECFEM2D codes?
>
>  Longer for your reply!
>
>
>
>
>
>
>  sincerely your,
>  ECE Duke, Guoping Chen
>  14th, Dec., 2011
> _______________________________________________
> CIG-SEISMO mailing list
> CIG-SEISMO at geodynamics.org
> http://geodynamics.org/cgi-bin/mailman/listinfo/cig-seismo