Did you see my comment about the length scale? If you are at 5+ times this length scale the inverse squared distance is a good scaling law. If you are below that length, it does not have to be a good approximation. It might be, but that depends on the beam then. In between, things are in between those two cases.