Kessler syndrome: vertical diffusion rate?
If (when?) a Kessler avalanche occurs, it will presumably begin in LEO: highly occupied and with missions in many different orbital inclinations.
Simulations often show a rapid spread of secondary-collision debris following initial collision, especially if both impactors are large. Assuming random or equipartition of momentum, later-generation collisions might rapidly create new projectiles in almost every direction in the orbital shell - perhaps with higher inclinations giving faster spreading due to increased intersection frequency with the more-populated near-equatorial orbits? Comments welcome.
My main question, though, is how rapidly the fragment swarm will be carried to higher orbits. (Lower ones, again, should clear quickly due to atmospheric drag.) Would daughter collisions tend to keep momentum in the same shell? When impact angles happen to confer a lot of vertical momentum, would that just create fragments with highly elliptical orbits? If so, perhaps drag at perigee would clear out those as well. Seems like vertical diffusion would require unlikely rear-end collisions to loft debris into a stable, near-circular orbit.
Or is there some other mechanism to add orbital delta-V instead of just elongation?