What exactly is up with the profound change in incision rates after the Deer Park lava shows up? Data artifact? Result of averaging over long periods of time? Or the effect of a gargantuan amount of sediment having to be delivered by the river following the demise of the Bogus Rim lava dam(s)?
I am leaning toward the latter as having an important effect. Recall that in my last field camp teaching episode (2010), I found Owyhee River terrace gravels underneath the Deer Park lava in the Hole in the Ground. Now speculating that these gravels related to terraces formed in this reach as the river was being forced to redistribute the huge sediment load liberated from the reservoir once held in by the Bogus lava dams. Can't imagine that this didn't matter. Also, the problem is potentially less significant in the case of the younger lava dams because of the very different geometries of the dammed lakes. Other factors include post-Bogus landsliding (tripartite effects of damming, armoring, and sediment loading), and regional climate change.
Thus, the roller coaster probably prevails in most cases over the long term.