February 19, 2017

More Threatening Weather at the Oroville Dam

11 AM Monday Precipitation Totals

Some of the terrain about the dam has received over 4 inches already.


The latest forecasts are worrying regarding the heavily damaged Oroville Dam in California.  And I am surprised there is to so little talk by CA state officials and the media about the danger.

During the past few days, with little rain and active drainage of water, the dam level has been reduced substantially (see image)

But central California and the Sierra Nevada mountains are about to be hit again...quite hard, with a strong atmospheric river (see moisture plot for 1 AM Monday).  The red and white colors shows high column integrated water vapor values.  Aimed right at the dam.
The forecasts show substantial precipitation with this atmospheric river.  Here is the accumulated precipitation for the 24h period ending 4 PM Monday from the excellent NCAR high-resolution ensemble system (the ensemble mean). 4-8 inches over the Sierra.

But that is the ensemble mean or average.  But there is a chance it could be more, so here is the ensemble maximum, the ensemble member with the highest amount of precipitation.  Much more (30-50% more in some locations)..and a serious problem for Oroville Dam.

The 48h precipitation total ending 4 AM Tuesday from the GFS model projects 5-10 inches in the mountains.

And the European Center model has the same idea, with 5-7 inches

Looking at the NCEP SREF ensemble system at Blue Canyon, about 15 miles away from the dam,  shows an ensemble mean of 7 inches, with a range from 5.5 to 13 inches.  Great confidence for a major rain event.

Hydrological forecasts by the National Weather Service, project a rapid rise of the Feather River, the main feed into the reservoir behind the dam, with the flow quadrupling from around 7.5 thousand cubic feet per second to around 32 thousand.

This is a dangerous situation and, as always, there is uncertainty with our forecasts.  Why is there so little discussion of the rapidly increasing precipitation?

And if you want to read an interesting blog about an unusually strong cold front that hit Hawaii, check this out:



  1. They have been doing quite a bit of repair work to the 'emergency spillway' (ESW) while letting out a lot of water via the main spillway.

    The ESW is basically a hill with a concrete overtopping area. The problem was that the downhill part of the ESW was just dirt, no rockfall to break up the water, so there was much dirt erosion. The worry was that the water overflow over the dam might cause a void to build back into (towards) the lake. If that happened, there would be quite a bit of water that would surge through the void (30 elevation feet?). That would not bode well for downstream. Basically, all of the highway 70 corridor (runs N/S) would be flooded. Lots of people there.

    The main spillway appears to be handling the increased rate, even though the lower part of the concrete main spillway has that big hole. But that, I think (I am not an engineer, nor do I play one on TV), is not a big issue compared to the ESW.

    They have been putting in large rocks/boulders, and concrete/grout to hold the rocks, below the ESW. If that is done correctly, that will reduce (no eliminate) the danger.

    I think that the DeptWaterResources staff/brass, though, is not telling all of the story. In particular, before the ESW overflowed, they were telling people not to worry. Until the emergency 'Get Out of Town' order came.

    I drove over Shasta Dam this weekend. It used to be very low. Now it is very high, about (estimated) 10-15 feet below the cleared area (no trees, just dirt) shoreline. And there is lots of water going downstream. We missed the flooding on I5 (at Williams) on Sat due to all of the rain on Friday.

    They are not out of the woods yet.

  2. Glad to see your information on the Oroville Dam situation. It's unfortunate that so little actual data is available in the news. I did find this site which is interesting.


    Title is at least premature, and hopefully wrong, but interesting information. Even if the spillway holds together, enough rain still appears to be a major problem.

  3. Again, not a flood engineer, but lived in the area for years. Flooding in Sac Valley is more of a problem than dam failure.

    I suspect some levee failures, especially in Sacramento River delta area (lots of low-lying farmland surrounded by high levees; farmland is lower than river levels). Central/Upper valley area (Stockton north) has lots of flood-prone farmland areas, although much of the farmland is 'built' to flood.

    Already some flooding north of Woodland (small towns in rural farm areas) with 1-2 feet of water in homes; I-5 flooding areas; rice farms flooded; irrigation ditches flooded. Area rivers (north of Woodland) are really high, flooding parkland (although parkland is 'built' in flood plains, so expected.

    This week will see some more flooding, I predict. I don't see dam failures, but some levee areas, especially in delta area, will fail. Levee maintenance on those rural areas usually happens during excess rain.

    Lots of residents haven't seen this problem; it's been at least a decade since heavy rains caused problems.

  4. ..also, don't expect mainstream media to cover issues very well. Regional papers (Sacramento, Woodland, Oroville, Red Bluff, Redding) will be better source for information

    (Ask the googles for local newspaper sites, don't rely on news aggregators like Google News or similar.)

  5. 32,000 cfs is actually not that much. Actually, inflows this morning were estimated at 40,000 cfs.

    They let out 100,000 cfs over the main spillway for the better part of a week after realizing they needed to stop allowing flow over the emergency spillway. While they hadn't intended to, the need to spill that much over the main spillway demonstrated that the damage did reach a relatively stable extent after a while, and the emergency spillway usage was unnecessary.

    Having brought the water level behind the dam back down to the normal flood control allowance for this time of year, they reduced flows to 60,000 cfs to reduce the chances of the main spillway erosion starting up again, and allow them to start clearing debris that is backing water up into the power plant.

    Once they start running the powerplant again, that will provide a near constant 14,000 cfs of flow, and that's enough that if the late winter and spring turn out to be typical this year, the main spillway might not be needed any more this season. As it has been a wet season so far and will probably continue to be so, they'll probably actually need to use it some, but it looks like they have the spare capacity in the reservoir to keep the main spillway use minimal.

  6. I was seeing quite of bit of local and regional news on the issue plus weather service storm and flood warnings and also a lot of local and regional radio. It's been there from the start.

  7. http://www.krcrtv.com/oroville-evacuation

    Plans today toincrease the level another 5K through the main spill way

    Just not making much noise around here.

  8. Cliff, would you care to comment on the use of weather data and predictions by key customers like water/dam authorities? That is, how could rainfall filling the streams catch them by surprise? Is the modeling of the hydrography far less sophisticated than the weather? Logically if one had an idea of the weather the volume of water on the ground (and thus into a reservoir) should be no surprise?

  9. Regarding the Oroville emergency spillway repair work, and the damage to the main spillway, take a look at this video (and story about full lakes feeding into Oroville Dam):


  10. Cliff, I am interested in your opinion of the following.

    1. Lake Oroville is 25 sq mi. The watershed is 3,611 sq miles, or 144.4x as large.

    2. 1 inch of rain throughout the watershed = 12 feet in the lake IF there is no soil absorption or transpiration.

    3. Soils are saturated, and tanspiration will be minimal in these conditions. In the last storm, 86% of the rain made it to the lake. Let's call it 1 inch of rain = 10 feet in the lake.

    4. The formula above does not take other inflow sources into account, such as the possibility that relatively warm rain accelerates snowmelt -- particularly given that snow in the watershed has already gotten heavy rain on top, which ought to make it more likely to start melting as it gets more rain.

    5. The CA Water Resources people have drawn down the lake by 50 feet. This suggests that 5 inches of rain over the watershed would refill the reservoir. Add a few inches to compensate for ongoing drawdowns, and it looks like anything more than about 8 inches of rain would (again) overtop the emergency spillway.


    I am very interested in a critique of this.

  11. In addition to my other comment:

    The erosion below the weir (lip) at the top of the emergency spillway showed not just soil erosion, but signficant rock erosion. The fear -- justifiable, from what I can tell -- is that this erosion, if it happens again (think 10 inches of rain throughout the watershed?) would undercut the weir and open up a hole in the emergency spillway at a level a lot lower than the current weir.

    If that happened, both spillways would see catastrophivc failure, followed by the dam itself. This is not likely, but I think it's a good deal more possible than any of the authorities want to admit.

  12. @Rick, I read the Sacramento Bee article that you linked. A couple things really jumped out at me:

    1. Lake Oroville reservoir is the biggest of a whole reservoir system, and that whole system is full. I wonder whether this will magnify the effect of a hard rain. Seems like it would, but I don't have the numbers.

    2. The Oroville managers seem to be unaware of the upper watershed reservoirs that feed Oroville, including one of them that is a) one-third the size of Oroville, and b) full.

    I sure hope there isn't a huge amount of rain, and that the Oroville managers aren't as clueless as they appear to be.

  13. The dam will be fine, Cliff. The managers lowered the level to 79% of capacity as the rain started, and can still release a lot of water as necessary. I doubt that I will eat these words. When the trouble started it was at 101% capacity.

  14. @ gregg daugherty:
    "That is, how could rainfall filling the streams catch them by surprise?"

    If you're referring to the original emergency, it didn't. Part of the main spillway collapsing into a sinkhole under the weight of 55,000 cfs of water flowing over it is what caught them by surprise. They initially stopped spilling water to assess the damage, then started spilling again at a slower pace that deliberately allowed the emergency spillway crest to overtop, because the erosion of the emergency spillway was not expected to threaten undermining the crest.

    If you're curious about hydrologic modelling, the URL Dave posted above details one of the most basic methods: empirical observations. That is, for a given amount of rain in a given period of time, how much has the river risen historically?

    @ Placeholder:
    "This suggests that 5 inches of rain over the watershed would refill the reservoir. Add a few inches to compensate for ongoing drawdowns, and it looks like anything more than about 8 inches of rain would (again) overtop the emergency spillway."

    The main thing to keep in mind is that your estimate does not account at all for reservoir outflows. 1 inch of rain over 3611 sq miles is ~193,000 acre-feet of water. They know they can spill at least 100,000 cfs over the main spillway in its current state, which is 198,000 acre-feet per day.

    Another big factor is rainfall is not even over that 3611 square miles. Maximum amounts of 10", which is an absolutely drenching downpour (Seattle's wettest day in recorded history was 5"), might average only half that.

    Lastly, even if the ground is saturated, the time to runoff into the dam is buffered by the time it takes water to flow from the furthest extents of the watershed.

    "If that happened, both spillways would see catastrophivc failure, followed by the dam itself."

    This is not correct. The main spillway would not be affected. The emergency spillway is not on the dam, but on the adjacent hillside. If the weir failed, the emergency spillway would likely erode quickly to the bed rock, and then slow, roughly 30 feet below the spillway crest.

    "The Oroville managers seem to be unaware of the upper watershed reservoirs that feed Oroville, including one of them that is a) one-third the size of Oroville, and b) full."

    Huh? Who do you think is telling the Bee about the other reservoirs in the Feather River watershed. The Oroville managers are not clueless about that factor.

  15. Most of the Oroville watershed (ie Feather River watershed) is in the Sierra above 5000ft. Snow level is more important than inches of precipitation in figuring out what will happen to the reservoir inflows. Usually love the blog but this post is outside the typical geographic scope here and seems to have missed the key point.

  16. @iamlucky, I won't argue it. I've spent more time today looking at this, and I agree that a dam collapse, or the equivalent if the weir above the emergency spillway were to collapse, is a much more remote possibility than I'd thought when I wrote my earlier messages.


Please make sure your comments are civil. Name calling and personal attacks are not appropriate.

Substantial Snow Returning to the Pacific Northwest

During even the strongest El Nino years there is often a period at the tail end of winter when the atmospheric circulation associated with E...