It would be an interesting exercise to see what would happen with current lake levels throughout the system if we had a 1983 type water year again. I'm assuming that there's quite a bit more diversion activity going on upstream - especially in Colorado, and with lake levels as low as they are everywhere else, the system could absorb 1983 right now quite well. Even with Powell at 3588, it would probably be OK. It's an interesting thought.
Welcome to the forum, Ben!
It is an interesting thought exercise. First, consider that the collective capacity of Lake Powell, Mead and all the other reservoirs in the system is something on the order of 57 maf. That is a gigantic capacity. But if we turn the wayback machine 39 years to May 4, 1983, here was the remaining capacity in the three largest reservoirs on that date:
Mead - 1.3 maf
Powell - 2.0 maf
Flaming Gorge - 0.5 maf
On that day, Powell was 13 feet below full, and Mead was 9 feet below full.
On April 1, Powell was already at 3685. On May 4, it was up to 3687.
That's the kind of situation that presents a potentially unmanageable flood hazard if there's a giant spring runoff. And of course, that's exactly what happened. Focusing on Powell, inflow to the lake from May 1-July 31 that year was 12.3 maf!! An entire typical annual flow in three months! Peak inflow reached 122,000 cfs on July 1, which is the equivalent of 243,000 af--in one day!! Of course, you can see there was no capacity for that kind of inflow, and so the releases through the dam in that period had to also be gigantic, and even then they could almost not keep up. In those three months, BOR released 7.9 maf through Glen Canyon Dam. That means an average of about 44,000 cfs every day for 3 months. Even then, you can see the math didn't work in terms of flood control--net gain in those three months was 4.4 maf. And Powell only had the capacity on May 1 for 2.0 maf. So that excess either had to spill over the dam and create real downstream havoc, or they had to take extraordinary measures, and that's what they did. They raised the dam 10 feet with plywood, and somehow it held, with 2 feet to spare. As close a call as it gets. Because while the lake was seeing an inflow greater than 100,000 cfs from June 26-July 4, BOR was cranking the releases at full capacity just to try to stave off disaster. Daily releases through the dam during that time ranged from 133-183,000 cfs! But any less, and the dam was in jeopardy.
Meanwhile, downstream at Lake Mead, inflow was picking up, peaking at 92,000 cfs on July 1. Ultimately, Lake Mead topped out at just under 1225 on July 24, which is essentially 5 feet above full, and just below the lip of Hoover Dam.
That's as close as it gets.
And then came 1984.
That was essentially a carbon copy of 1983, but that year the BOR was ready for it. Powell was lower going into spring, for one thing. It stood at 3674 on April 1, or 11 feet lower than on that date in 1983. That provided the cushion needed to handle any sort of huge inflow. And inflow it did. The main difference this time is that the giant inflows occurred earlier--the end of May instead of the end of June. But they were even larger! Inflows exceeded 100,000 cfs from May 23 to June 5. Fortunately, BOR learned its lesson from 1983 and already anticipated that kind of flow, so they released an average of 45,000 cfs after May 9, compared to about half that outflow the year before in the same timeframe. That allowed for a more controlled situation than in 1983, and made all the difference. The lake peaked at 3702 on July 4.
So back to the original question. The dams work great for flood control, provided there's enough remaining unused capacity. And even in 1983 and 1984, when the reservoirs were already nearly full in April, there was just barely enough remaining capacity to handle the largest inflow events since the reservoirs were constructed.
So any sort of huge inflow when the lakes are as low as they are today is really no problem to handle for downstream flood control. Ideally, you'd have somewhere from 10-15 maf capacity remaining in the system for flood control, which would allow for any sort of sustained inflow to be handled. In spring 1983, the entire system only had about 4 maf before the big runoff. Not enough.
At 3588, Powell would have about 14 maf of remaining unused capacity to absorb a flood. I'd say that's a more-than-sufficent built-in safety factor.