Open eyalshimony opened 3 years ago
I have read how the source code deal with the adjoint simulation. The calculation of the adjoint simulation is just to use the adjoint source as the source time function to do a forward simulation. Theoretically the adjoint simulation and the calculation of sensitive kernel does not conflict with CPML.
Lines 898-899 in read_parameter_file.f90 cause the program to stop if PML conditions are used either in the forward part (SAVE_FORWARD=true) or the adjoint part (SIMULATION_TYPE=3) of FWI. However, from the discussion here: http://lists.geodynamics.org/pipermail/cig-seismo/2017-October.txt, it seems that one can actually use PML conditions with FWI. Therefore, it seems that these lines should be removed.