emsig/emg3d

Adjoint for magnetic receivers

prisae opened this issue · 1 comments

The adjoint-state gradient is currently not perfect. The magnetic response at receivers is obtained from taking the curl of the electric field (Faraday's law), and linearly interpolating the receiver response. In order to get the proper adjoint we need to:

  • distribute the residual with the adjoint of tri-linear interpolation to the magnetic field (adjust source fct for magn. grid);
  • get from the magnetic field to the electric field (curl transpose).

This was fixed in #251