[CIG-SEISMO] SPECFEM1D: Bug in Local to Global numbering of the source

[Martin van Driel]

Dear all,

I think the source implementation in SPECFEM1D is buggy (it is commented 
out anyway, but it might save others some time to have it correct 
nonetheless):

the variables iglob_source and stf are not defined and the indices of 
the local to global numbering are swapped. This way it seems to work:

! add source at global level
    accel(ibool(i_source,ispec_source)) = &
	accel(ibool(i_source,ispec_source)) + &
         source_time_function(dble(it - 1) * DT - hdur,hdur) * source_amp


All the best,
Marti