The Oroville Dam spillway looms as a catastrophic at biblical scale. In the short term bigger than Flynn. Bigger than Trump. Bigger than any other issue faced by the country today. The potential downside is huge. A wall of water 30 feet high, inundating Oroville to 100’. Flooding of the Central valley as far as Sacramento, with the river itself flooded all the way to the ocean. Displacements of a up to a million people—or more. Billions in damage. Disease—that’s a lot of area suitable for mosquitos, and a lot of people close to each other with no reliable source of water.
Already, the California Water Board owes a mass resignation: meetings for the last 9 years had agendas of every possible issue except infrastructure safety. If the worst happens, the Governor must resign as well. No politician can—or should—survive mismanagement of this magnitude.
Of course, it might not happen. But right now, we don’t know—and the water board doesn’t either.
The Oroville dam is over 700’ high. (Stories describe the top of the dam at 901'; that is the elevation, not the depth.) The structure itself is not currently threatened. That could change, but it is not a big threat. The big threats, of course are
- How fast the main spillway will back erode. It has stabilized to a degree, sort of. But it has eroded back about 300’ since the hole opened. This was expected to stop as the erosion reached its current position at a ridge of bedrock on the hillside.
- To what extent the hillside will be stable with that much missing mass
- The solidity—or lack thereof—of the bedrock under the emergency spillway
Note: I couldn't upload an image of the damage to the emergency spillway. Use this one for reference; it’s the best I've found.
www.metabunk.org/...
The water
The current flow of water over the spillway is 100,000 cubic feet per second. That’s a lot of water; it’s roughly equal to the median flow of the Mississippi at St Louis for this time of year. But the Mississippi is flowing on level ground. The spillway is sloped at an angle of roughly 1:3, or 25% grade, which is very steep indeed; much steeper than any road, and comparable to a black diamond ski trail. Water flowing at that angle is highly destructive; indeed it has formed a waterfall at the dam. This is actually a good thing: the water isn't doing much back eroding as it leaps out into the abyss. It's mostly eroding forward and deeper down the hillside. The water is running almost entirely on bedrock at this point, as the white, unstained color in current photos attest. Earlier photos showed streaks of brown, evidence of active fresh erosion of overburden. Watch for more of this in the future; it means the erosion has jumped to a new location.
The median flow of the Feather River is fortunately much lower than 100,000 cfcs: closer to 10,000cfs with bursts approaching 200,000. The water is dropping rapidly now, and will continue to do so until it reaches the level of the base of the spillway—desirable to get it as low as possible before a breach. This further means that the spillway will not see continuous use. The power plant can handle roughly 16,000cfs when it's running correctly. Obviously getting it running again, at full bore, is the first order of business, even if they have to sell it for nothing. That energy has negative value where it is now. The threat to the high tension line come into play here: if the lines fail, there’s no way to get rid of the energy. (The plant cannot run without a load.)
Best scenario: power plant operates at 100% until May, drawing most of the inflow. Occasional surges over the spillway handle the rest, with minimal further damage.
Note that the DPW does not want to run the spillway at low volumes because the water will “dribble” down the cliff under the chute, and do further back erosion. The DPW rep wasn't lying when he said that sometimes lower flows do more damage, because the higher flow shoots out further from the eroding face.
But for this to work, the spillway cannot erode much beyond its current endpoint at the ridge.
The ridge
The ridge is an outcrop of bedrock that defines the shape of the hillside. Above, the hill is more level; below it's steeper. This is a critical point of failure. If erosion continues past this to the relatively level area above, there’s no real hope for the erosion to stop by itself. Over the next few days, it should become clear as to whether this hope is vindicated. There’s really not much else to do but draw down the water to a minimum and get the power plant running again. Once the water is drawn down and the next storm passes, the DPW will be able to stop the flow and assess the damage. If the ridge fails to stop the back erosion, the DPW will be faced with a terrible choice: let the main spillway fail, or let the emergency spillway fail.
A dam (or spillway, or hillside) may fail in two ways: a blowout, where a large section of dam collapses and the water flows through in a tremendous wall; or a 'loss of crest control’, where the water overtops the structure and begins to erode a channel from above, at a rate of erosion faster than the head of water behind it drops. This is a runaway situation, where water runs faster and faster, doing more damage as the depth of water flowing through the channel increases. It only stops when either the reservoir runs dry, or the erosion reaches harder bedrock, and slows to a rate below the rate at which the head behind it is dropping. “Loss of crest control” is also catastrophic, with massive flooding, but the flood occurs over a sustained period rather than all at once. It does much less damage to downstream land forms (but any houses are still destroyed, if only by submersion.)
If the main spillway continues to erode, the DPW is forced with a terrible choice. Can they rely on the spillway structure enough to avoid a blowout? Then 'loss of crest control' at the main spillway is desirable, as it begins with significantly less water in the dam. If a blowout is likely, then 'loss of crest control' at the emergency spillway may be a better solution, so long as the crest is eventually stopped by bedrock. That's the million dollar question in all cases: can the bedrock hold?
The rock
The Oroville dam rests on ancient seabed, a mix of schist and other olivines. This rock is very hard and tough, but fractures easily, and weathers quickly when exposed to air or water. The weathered rock is red; the unweathered rock varies from light to greenish “olive” grey. In the photo below, the bedrock to the right of the emergency spillway (from above) is red: weathered, fractured, and generally rotten. There’s an upthrust bit of weathered bedrock that was quickly eroded through under the minimum force in the 'experiment’ of two days ago. This is the damage that caused the evacuation Saturday afternoon, and forced the DPW to start drawing the water down at 100,000cfs to reduce load on the emergency spillway (ES.)
The bedrock at the far end of the ES is in much better shape. The color is better, and the rock wasn't just mined through, but rather lost any broken-off chunks from above. Still, a rough surface like this can't hold up to a huge amount of force. Filling in the cavities with concrete is likely to work in such case. For the rotten bedrock, the only hopes are covering it entirely with smooth, good quality—reinforced!—concrete, or verifying that it doesn't go too deep, and allowing runway erosion until hard rock is reached. Any concrete they do pour, they must allow to set. The stuff they poured last week never had time to set, and is already badly eroded—pockmarked like a moonscape.
The choices
The first choice that must be made is get that power plant fixed. If it can handle even a significant fraction of the mean flow, it needs to run at as close to 100% as is possible. So keep at least some lines running from the plant operable, and offer it at whatever rate the market will accept. The best option is to keep the lake drawn down using the plant, with occasional surges on the spillway until spring. That's the hope, but hope is not a plan.
Additionally, disaster planning must be done now, before things get worse. This is a grave situation, and the DPW lost a lot of trust by downplaying it for the first several days. Find places to shelter, let people return to collect valuables, get food and medicine supplies lined up. The evacuation doesn't need to happen until another few hundred feet of spillway are lost. But it cannot wait til the last minute. I don't pretend to know this stuff, but the potential scope for disaster is almost inconceivable, which is probably why it's getting less play than it might.
Beyond that: mitigate the actual worst case catastrophe
- Decide which failure mode is least bad, just in case. This will take a while, as it depends on if and how the spillway fails.
- Try to keep the rotten end of the ES from getting any flow; the rock at the other end looks much, much tougher.
- Possibly dynamite the ‘safe' half of the ES, to force the flow to that end. (Yes, things are this bad.)
- If the weir on the ES fails, that would count as a small blowout, releasing much more head than otherwise would be necessary. That's one more reason to possibly dynamite it down near the level of bedrock, but I don't actually know if this is feasible. But it’s a lot more desirable than just letting the weir fail, inevitably at the weaker end, with the highest potential for loss of crest control.
Going forward
There has been a real political let us down, both in specific instances and particular. Over the last 10 years, there have been two points at which different decisions might have been made, but the things were let slide. The first occasion during relicensing in 2005, with serious input from three environmental groups—they come out looking good, but very few others do. The groups noted that the original plan for the emergency spillway was predicated on the existence of the Marysville dam, which was never built. Without that dam, the spillway was much more likely to be used, and therefore should be updated to modern requirements. California, FERC, and the local water agencies decided the plan was too expensive and unnecessary. The second was 2012-2013 when there was damage to the spillway that was only patched up, not fully replaced. In retrospect, it’s clear that this was a serious sign that wasn't treated as such, and was not sufficiently investigated. So two governors, one from each party, have failed to do a good job in this. Ratepayers didn't want to spend more to fund more bonds, the responsibilities go on and on, but it really goes most heavily on the governor and a political situation where water fees are considered anathema. It was known before this event that maintenance to infrastructure was underfunded and insufficient.
Any politician in charge of a disaster of such magnitude—if the worst occurs—will be politically crippled. There's little evidence of action by governor Brown prior to the evacuation order by the local sheriff on Saturday, and the state was late in asking for federal aid. The only people looking good in this are the sheriff, the three environmental groups, and the workers actually doing the job in difficult conditions.
Caveat
I don't pretend to be an expert in this stuff, but I have been following forums where actual experts—civil engineers and disaster planners—are discussing the issues. I've done the best I could to summarize them in a diary here, my first serious one ever.
Resources
www.metabunk.org/… expert forum
pixel-ca-dwr.photoshelter.com/… DWR damage images
cdec.water.ca.gov/… YTD water level, Oroville dam. It blew up when they slowed the spillway
www.water.ca.gov/… DWR news releases
www.swrcb.ca.gov/… DWR annual meeting agendas
www.metabunk.org/… Progression of main spillway damage
www.linkedin.com/...essay by expert. He has others.
Edits:
Fixed responsibility section. Fixed elevation data.
Wednesday, Feb 15, 2017 · 2:19:42 AM +00:00 · p mac
Emergency spillway weir likely not founded on hard rock.
Expert consensus for the condition of the ES is that the original construction did not consider the possibility of erosion of weathered bedrock. So the 'ogee’ (parabolic shaped) weir is resting on rock that may well erode from beneath , possibly leading to it tipping over catastrophically. As it stands, everything rides on the main spillway staying stable at its current location.
From the experts’ forum
I and a few others made educated guesses regarding the placement of the weir the other day. If the weir was constructed as shown on the plans, there would have been no place for that "key" or ridge on the bottom if this method of cleaning the bedrock had been used, unless that initial undercut (that's what we call any excavation deeper than the planned depth of bearing) would have been backfilled with concrete to establish a flat bearing surface. However, jut a simple graded strip would have allowed for that, and given the obvious lack of concern about emergency-spillway erosion that old documents (posted here) have illustrated, I tend to doubt that they went to great lengths to anchor the weir in the illustrated manner, especially since the lateral pressure exerted on the weir would not be all that great in the first place (lateral pressure at the base of the dam would be enormous, leading to a much greater need for solid connection to the underlying rock). You are correct though, that this process of cleaning the rock would have greatly reduced the likelihood for erosion to work its way headward beneath the weir, but since the whole reason for cleaning the rock was entirely different at the dam location (and again, given the overall attitude back then about emergency-spillway serviceablity), I'm not confident this was done there.
Wednesday, Feb 15, 2017 · 7:52:00 PM +00:00 · p mac
Second followup
In recent images, the end of the broken spillway has appeared stable since Feb 11 (courtesy of metabunk). This is a fortunate turn, and is doubtless why the evacuation was suspended. Teams are indeed breaking up new sandbars in the river to get the hydropower turned back on. The electric company is working on shoring up their vulnerable towers with an alternate route.
These are both good signs, suggesting that—for now—the two most critical components are under control. First, the main spillway is at least temporarily stable; second, that they should be able to shut it down except for further big storms. That said, Oroville dam remains badly compromised, and this situation could change at any moment if the spillway is undercut. Oroville town and downstream will remain on alert.
Look for the post mortem studies on this. The initial public response of DWR was at best perplexing, at worst, intentionally misleading or simply incompetent. The political fallout will be extreme, no matter the final engineering result.