The spatial boundary conditions are the standard assumptions of zero CR density at the halo boundaries. This is actually not very physical. Free escape at the boundaries (which is not the same thing) would be more appropriate. I am going to implement this soon.

The momentum boundary conditions are indeed not rigorous, and GALPROP relies on the assumption that the momentum range is chosen large enough for this not to be important. At low energies with energy losses, the flow is only from high to low, so it is no problem. For reacceleration it could be a problem. Same for momentum advection. These issues deserve attention.

_________________
Andy Strong, MPE