Cavitation, air bubbles trapped in the water, striking the spillway surface chipped away over the years like tiny jackhammers.
Yes and considering how little use the main spillway at Oroville got used over the years; I was struck by how much damage from cavitation the close up images revealed in the floor of the structure.
One of the benefits of the current spillway Gate structure design is the "recombining" flows, from the concrete column separation, in that it creates a useful injection of larger air bubbles via the recombination "tail fins". This is effectively an "aeration" injection process. This pre-aeration of the water helped in reducing cavitation effects down the spillway. Thus, the criticality was to maintain a given smoothness to the concrete surfaces. Without aeration, the smoothness criticality becomes more acute.
A large example of aeration is the dissipation blocks at the end of the spillway. These were specifically designed to cause a huge "aeration" conversion of the dense & powerful waterflow while creating a "jump" upwards. Thus the plunge back into the Feather River was less destructive in erosive outcomes. All of the angles to these blocks were carefully designed to transition the water into an aerated form without causing cavitation erosion of the chute block. They did experience minor cavitation erosion over time on certain edges, but this was considered a manageable condition in resurfacing repairs.
The recently released BOC reports discuss inclusion of "aeration devices" at the higher velocity transition of the spillway flow - this is to reduce the issue of cavitation erosion down the lower spillway. The other issue, partially redacted, is their concerns of cavitation erosion on the rough surface characteristic of Roller Compacted Concrete (discussion then turned to investigating a "coating" of a smoother surface concrete).
However, with huge design flaws of the drain line thinning induced slab cracking, the patches to the cracking, the underflow erosion of slabs, the flexing of the slabs, the uplift of the slabs from hydraulic jacking, the "spalling" or "chipping" of pieces of the surface of the slabs - all were a target rich environment for cavitation erosion damage. But overall, the surface of the chutes likely would have been fine for this design if it weren't for these design flaws. Cavitation erosion likely accelerated a "spalled" chip off of one of the seams or at one of the drain patch cracked areas - thus this cavitation erosion accelerated a penetration of waterflow once a "hole" was made.
*how aeration works: formation of disruptive pressures, normally causing cavitation damage, is mitigated by expanding and reabsorbing into the existing aeration bubbles. Without the aeration, the expansion and collapse of a cavitation bubble is much greater in intensity in its collapse, focus, and energy. Thus the aeration act as a form of miniature "shock absorbers" to the cavitation forming (i.e sudden pressure differential from surface flow irregularity) and collapsing process.