[CIG-SEISMO] Using SPECFEM to calculate the impulse response of anisotropic layers

[Dimitri Komatitsch]

Dear Derek,

Yes, I think SPECFEM2D should be able to compute that. Maybe not in less 
than 10 seconds though (?). Try turning off all the I/Os i.e. color 
pictures and vector plots created by the code. Also make sure you output 
seismograms only once, at the end (i.e. turn partial seismogram output off).

Note that SPECFEM2D is optimized for flexibility rather than speed
(we assume 2D runs are not too expensive and thus flexibility matters 
more than CPU time).

Best regards,
Dimitri.

On 09/27/2012 10:48 PM, Derek wrote:
> Dear colleagues,
>
> I sent this email to the list before, and several people kindly pointed
> out to me that my subject line was incorrect and that my error may
> reduce the response rate. So, I will try again, and apologize for
> sending this a 2nd time.
>
> I've got a rather naive question about SPECFEM.
>
> I'm looking for code to calculate the approximate 1-D impulse response
> of a horizontally layered anisotropic medium to a near (but not exactly)
> vertically incident waveform from below.
>
> More specifically:
>
>     • I am hoping to have 2 or 3 ~100km thick layers that have constant
>     velocity but differing types of anisotropy;
>     • I would like to input an SV impulse function (a 1s half-width
>     gaussian would work);
>     • I would like to be able to iteratively change the orientation of
>     the anisotropy to try new models (say by using a shell script to
>     change a parameter file)
>     • and I would like this to run in < 10 s (say in mpi mode on a 12
>     node computer).
>
> I only need one specific ray path of energy propagating from below the
> horizontally oriented layers to the "surface", along a specific incident
> ray parameter. This is part of a project to do a parameter sweep over
> all reasonable aspects of hexagonal anisotropy to match observed
> waveforms, so the goal would be to integrate the impulse response
> generation into a larger bit of code/shell script/matlab
>
> This strikes me as something rather different from what SPECFEM is
> intended to do, but I am wondering if you think what I am hoping to do
> may be feasible, probably in SPECFEM 2D?
>
> I have some old F77 code that does what I want, but it is non-mpi and
> does not always compile on modern computers.
>
> Thanks and best regards,
>
> Derek
>
> --
> Derek Schutt
> Assistant Professor
> Geosciences
> Colorado State University
> 970-491-5786
>
>
>
> _______________________________________________
> CIG-SEISMO mailing list
> CIG-SEISMO at geodynamics.org
> http://geodynamics.org/cgi-bin/mailman/listinfo/cig-seismo

-- 
Dimitri Komatitsch
CNRS Research Director (DR CNRS), Laboratory of Mechanics and Acoustics,
UPR 7051, Marseille, France    http://komatitsch.free.fr