After inspecting the vertical axis, it becomes clear that the size of the quadrupole term gets smaller when the nucleus becomes smaller. So I think this means that we are dealing with smaller corrections as l/d becomes smaller and thus the multipole expansion converges faster.