With the Oroville Dam situation now making headline news, I thought I would present a hydrologic perspective. I, myself, lived through something like this in 2009 although not on this scale. First some terms:
River flow is measured in cubic feet per second. One cubic foot is 7.5 gallons. Reservoir volume is measured in acre-feet. It is exactly what it sounds like. It's the volume of water needed to cover an acre of land one foot deep. An acre is 43,560 square feet, so an acre-foot is 43,560 cubic feet. You could do the math, but it works out that a flow of 1 cubic feet per second equals 2 acre feet per day (it's actually 1.987). So if a river is flowing at 10 cfs, the reservoir it flows into will gain 20 acre feet that day.
Dams can be used for several purposes. Water storage for cities and farms is usually the most important. Flood control is the other biggie. Then there is power generation and recreation. Those are usually secondary benefits but systems could be designed specifically for those purposes. Then there are smaller dams with can be used to aid navigation and divert water into canals or pipelines.
The most important part of a dam is its outlet works. You need to be able to get the water from out behind the dam. Usually its just a large pipe/conduit. It should be designed to release the capacity of the channel below without flooding. Many dams also have a powerhouse below it. If a dam's major purpose is power, you want as much flow to go through the plant and not bypassed by a pipe. Water is money! Next there is the spillway. This is your next line of defense to preserve the dam. If inflows are very unpredictable, it may be necessary to spill water during major floods. Normally you don't want to use the spillway because you're wasting water. It can also cause flooding downstream because now you might be releasing flows that are above channel capacity. At this point the main concern is the dam, not downstream (which would have been flooded anyway even if the dam wasn't there). This is where flood control comes into play. If the dam has a flood control component, you have to have enough capacity in reserve to handle the peak of the flood. Lastly there is the emergency spillway. This is your last line of defense to save the dam from overtopping. One of our dams has a "fuse plug" which is a small berm of dirt just below the elevation of the top of the dam. If water ever got to that elevation, it would blow out the plug and act as an emergency spillway. Once again, if you get to that point you should be shitting many bricks.
Oroville isn't a federal dam. It is owned by California. When I heard that they were going to use the emergency spillway I knew that shit was serious. Having seen aerial pictures of the dam, however, I saw that it was designed quite well in that the spillways were away from the dam itself. Even if there was a shitload of erosion, the dam shouldn't be compromised. There will be a hell of a lot of cleanup, however, in rehabilitation of the emergency spillway channel and of course the main concrete spillway with the huge gaping hole in it.
I would be curious to read the post mortem of this event. I'd like to know what the hell the hydrologist on staff was thinking. You have record snow and rain and the lake is almost full. You left yourself with no margin for safety. I have a feeling that his/her thinking was similar to mine during the Great Fuckup of 2009. I, too, was just coming off of a major drought. There was a fear that I had to store as much as possible as early as possible in case something went wrong. I filled the reservoir too quickly right before a major heat wave rapidly melted the remaining snowpack. I was a day too slow in my increasing the release and before I knew it I had water lapping up against the spillway gates. The spillway was in such bad shape that any significant flow down it would have resulted in something similar to what happened at Oroville. The big difference is that the spillway is right up against the dam. If the spillway failed, it could take the dam with it. Fortunately I ended up releasing enough water through the outlet works that the water never reached the spillway. Disaster was averted. It nearly cost me my job, but I made it through. I then went on a massive data mining effort to fully understand the relationship between snow volumes and runoff. I now go into runoff season with eyes wide open and multiple scenarios ready to go. The silver lining of the whole thing is that I am now known as the savior of the river. My fuckup caused a major improvement in the ecosystem of the river. It worked so well we did a "controlled fuckup" last year at my suggestion. Instead of a reprimand, I got an award. Now I'm being asked to fuck up as much as possible. *lol*