source depth in netcdf file is the one from inparam source, not the one actually used

[martinvandriel]

which is the closest gll point. Maybe it makes more sense to store the "true" source depth, especially for long period simulation where this can easily be tens of kilometers different.