I'm not really convinced this data falsifies the narrative of a slow erosion. It's just the Pareto principle in action. I bet if you graphed the erosion of a hill you'd see the same pattern.