Why aren't dams designed to be overtopped? Or are they?

[Quartz]

Inspired by this thread, why is overtopping a dam such a big deal? The net outflow from a dam must meet or exceed the inflow. And it all goes down the same river bed. At least after the dam is filled, that is. But surely dams are designed for flooding?

I'm ignorant on this; educate me.

[Colibri]

Note: Edited thread title to fix typo.

Colibri
General Questions Moderator

[dolphinboy]

Dams are designed to release water at specific levels ensuring that the river doesn't flood downstream. If the dam overfills to the point of releasing more water than the river can handle it will lead to flooding. Typically the capacity of the dam it far greater than what would be expected to inflow so that the dam can effectively buffer the downstream system. But if the dam can't release normally, or otherwise becomes completely full, and starts to spill over in an uncontrolled fashion, it can certainly lead to flooding.

[Exapno Mapcase]

Dams are no different from the overflow opening near the top of your bathroom sink or bathtub. A moderate amount of water continually flowing into the sink will spill through the overflow without soaking the floor. A large amount of water will be more than the overflow can handle and everything will get soaked.

[Yllaria]

Dam management can be tricky. One of the purposes of a dam is retaining water to use for hydroelectric power and/or as a source of water through the dry season. Another purpose is flood control. For the best flood control response, you want a near-empty dam.

Notice that best practices for the different purposes are at cross purposes with each other. If you're managing a multi-purpose dam, you're going to be making water release decisions based on weather forcasts. Most of the time that works pretty well. Sometimes you get caught with your retained volume up.

[FoieGrasIsEvil]

Don't most dams have sluice gates to prevent overtopping?

[OtakuLoki]

Related question: Isn't having uncontrolled overflow across the top of a dam (not in the designed overflow spillway) bad for the whole structure, causing accelerated erosion?

[John Carter of Mars]

Originally Posted by OtakuLoki

Related question: Isn't having uncontrolled overflow across the top of a dam (not in the designed overflow spillway) bad for the whole structure, causing accelerated erosion?

Yes, if the dam itself is overflowed, the water will probably cut channels down through the dam and ruin it.

I have a fairly large lake (12 acres) with an earthen dam on this farm. It was designed with a proper spillway. Trust me, if water ever came into that lake at a rate greater than the spillway could dispose of, there would be bigger things to consider.

[UncleRojelio]

Some are designed to have water run over the top. For instance, Inks dam in the Texas Highland lakes chain has water running over the top when it isn't generating.

[NinetyWt]

Originally Posted by Quartz

Inspired by this thread, why is overtopping a dam such a big deal? The net outflow from a dam must meet or exceed the inflow. And it all goes down the same river bed. At least after the dam is filled, that is. But surely dams are designed for flooding?
I'm ignorant on this; educate me.

Overtopping is a big deal for earthen dams because the water can quickly erode the embankment, causing the dam to breach. The flood wave resulting from a breach can be extremely destructive (see the historic Johnstown, Pennsylvania flood which killed about 2,200 people). Concrete dams may be more like weirs and it may be acceptable to have water flowing over the crest. The important difference is the armoring of the crest to prevent erosion.

The net outflow of the lake does not necessarily have to meet or exceed the inflow; that depends upon how much storage is available in the reservoir. The standard setup (for earthen dams) is to have two spillways (or outlets). The principal spillway is used to control the normal pool of the dam. This is the level that you want your usual lake level to be. The elevation of this pool will be at the crest of the principal spillway. During times of drought the pool may be lower due to evaporation.

Water flowing into the lake will fill up the lake as the principal spillway lets some of the water out. Here in Mississippi, the principal spillway must be of sufficient size that the lake can pass or store the 1% -chance storm (for high hazard dams).

Here is a good graphic illustrating the dam, impounded water (lake), and spillways. Some folks refer to the auxilliary spillway as the 'emergency' spillway.

The auxilliary spillway is used to pass any runoff which exceeds the lake's ability to store/pass the 1%-chance storm. It will have a crest at a higher elevation than that of the principal spillway. Here in Mississippi, this spillway must be sized to pass the Probable Maximum Precipitation (PMP) without overtopping the dam (for high hazard dams). The top of dam should be higher than the crest of the auxilliary (I like to use a minimum of 3 feet).

Here in central Mississippi, the PMP is approximately 45 inches of rainfall in 24 hours. This differs across the state; and is different for different areas of the U.S. For reference, the 1%-chance storm in central Mississippi is about 9.6 inches in 24 hours.

[NinetyWt]

Originally Posted by dolphinboy

Dams are designed to release water at specific levels ensuring that the river doesn't flood downstream.

That's true if it is a flood-control structure. Not all dams are. Many are not used for flood control, and merely pass on the larger storm runoff.

Typically the capacity of the dam it far greater than what would be expected to inflow

A combination of the capacity to store water and the capacity to pass water on through (via the spillway(s)) should be greater than what is expected to inflow.

Originally Posted by FoieGrasIsEvil

Don't most dams have sluice gates to prevent overtopping?

To me, a sluice gate is a sliding valve which is used to drain the lake for maintenance. I'm not sure that's what you mean, but sluice gates typically won't be large enough to pass high volumes of water. Instead, the water-passing capacity will be in the spillways; either the principal or a combination of the principal and auxilliary.

Originally Posted by OtakuLoki

Isn't having uncontrolled overflow across the top of a dam (not in the designed overflow spillway) bad for the whole structure, causing accelerated erosion?

Absolutely. Water passing over a broad-crested weir (which is what the dam will function as) passes through critical depth. The velocity will be critical velocity, which will be extremely erosive. Once the water behind the dam begins to flow through the breach, it's curtains for the dam.

[bluezooky]

Originally Posted by NinetyWt

The auxilliary spillway is used to pass any runoff which exceeds the lake's ability to store/pass the 1%-chance storm. It will have a crest at a higher elevation than that of the principal spillway. Here in Mississippi, this spillway must be sized to pass the Probable Maximum Precipitation (PMP) without overtopping the dam (for high hazard dams). The top of dam should be higher than the crest of the auxilliary (I like to use a minimum of 3 feet).

Here in central Mississippi, the PMP is approximately 45 inches of rainfall in 24 hours. This differs across the state; and is different for different areas of the U.S. For reference, the 1%-chance storm in central Mississippi is about 9.6 inches in 24 hours.

That’s really interesting as a 0.5 % storm is as considered a very real risk down in Australia, many dams where strengthened in the eighties and nineties as some doubt was cast on the accuracy as the flood estimates. Without river level data going back for centuries you really only guess what nature likely to throw at a dam

[Quartz]

Originally Posted by NinetyWt

Water passing over a broad-crested weir (which is what the dam will function as) passes through critical depth. The velocity will be critical velocity, which will be extremely erosive.

I'm sorry, but I don't understand this. Can you elaborate?

[NinetyWt]

Originally Posted by Quartz

I'm sorry, but I don't understand this. Can you elaborate?

Surely. A weir is like a wall built across a stream channel. It can be made of a number of different materials - some indigenous folks used branches to make weirs which would trap fish, for example.

Concrete weir.

A weir controlls water flow by presenting a long, flat opening for the water to pass over. Weirs are often used to measure flow rates because the equations for weir flow are well known. Article explaining weir flow equations here.

Critical depth and velocity:
In two-dimensional flow, there are two regimes or types of flow: sub-critical flow and super-critical flow. For practical purposes, naturally flowing streams are assumed to be in sub-critical flow (although this does not necessarily hold true in real-life for all reaches of all streams, it is true for the majority). Critical depth is the breakpoint between these two types of flow. It is the depth at the point of minimum energy. Depths greater than the critical depth indicate subcritical flow and depths less than the critical depth indicate supercritical flow. Any obstruction in the stream which the water has to flow across will cause the water to pass through critical depth as it goes over the obstruction.

Critical velocity is greater than sub-critical velocity, and therefore has greater erosive power. If we expect high velocities, say in a man-made channel, we will armor the channel with stone, concrete, etc. to prevent scouring and erosion. We can pave over the dam with concrete, for example, if we are going to allow water to flow over it.

For a discussion of energy and flow regimes check out this . I also found a good discussion of critical flow here (Word document, not very large).

On preview, that is certainly rambling. I hope it helped.

[NinetyWt]

Originally Posted by bluezooky

That’s really interesting as a 0.5 % storm is as considered a very real risk down in Australia, many dams where strengthened in the eighties and nineties as some doubt was cast on the accuracy as the flood estimates. Without river level data going back for centuries you really only guess what nature likely to throw at a dam

I can certainly see where that makes sense. It would be fascinating to work on hydrology research in Australia and New Zealand, with so much yet unknown.

[Santo Rugger]

Do 1% and 0.5% storms mean there's a 1% of the storm happening every year? Are these equivalent to 100 year and 50 year floods?

[CalMeacham]

As has been mentioned already, you DON'T want water pouring over the top of your earthen dam, because it will wash it away. That's how the Johnstown Flood happened

http://en.wikipedia.org/wiki/Johnstown_Flood

http://www.jaha.org/FloodMuseum/history.html

http://www.nps.gov/jofl/



You may wonder "Why would people build dam,s that way, if it can have such awful consequences?"

The answer is that, as long as water doesn't flow over the top of the dam, this kind of construction is perfectly safe. I was astonished, when I visited the region of that flood a few years ago, to find that about 1/2 of the original dam is still there. The central half or so was washed away, but the rest remained intact, and has continued to for over a century.

The reason the dam washed away in the first place was that:

1.) It was flooding, with rapidly rising waters in the heavy rain

2.) No one was using the relief pipes to draw off the excess. In fact, the relief pipes had been removed altogether when the dam passed from the railroad company to private hands.

3.) The weir through which the overflow would go if the level got too high was blocked by floating debris, which no one was clearing away. So weater couldn't flow safely away over the spillway, which would've been safe.

4.) So, finally, the water had no place to go other than over the top of the dam, washing it away and eventually stripping out the clay core, all in one catastrophic break.





There's a similar earthenware dam not far from where I live. When the water gets high, I keep an eye on it. Fortunately, this one is an axctive part of the local water control system, and does have its drainage system intact. Nevertheless, I've called up the system when I saw water lapping the top during flooding a few years back.

[BobLibDem]

NinetyWt, reading your post makes me thankful I only took one course in fluid mechanics. Are you a hydrologist or a civil engineer that specializes in hydraulics?

[Snarky_Kong]

Originally Posted by Santo Rugger

Do 1% and 0.5% storms mean there's a 1% of the storm happening every year? Are these equivalent to 100 year and 50 year floods?

Yes. They are equivalent to 100 year and 200 year storms. Keep in mind that there is a fairly large amount of uncertainty about what exactly those storms consist of. While records of rainfall go back a long ways, the sampling instruments are fairly crude. Basically you have a bucket of a given volume, and when it fills up it is emptied and a check mark is recorded. The opening for one of these devices is something like 6 inches in diameter and it might be the only device within miles. As you can see that's not a very representative sample. In one case I studied a dam was designed for a 250 year flood, but when more data was collected and analyzed it was determined that it was more like a 60 year event.

[NinetyWt]

Originally Posted by Santo Rugger

Do 1% and 0.5% storms mean there's a 1% of the storm happening every year? Are these equivalent to 100 year and 50 year floods?

The 1% chance flood is that which has 1% chance of being equalled or exceeded within any given year. Formerly referred to as a 100-year flood because we expect it to occur at least once within a given 100-year period of time. I don't like to use '100-year' because people tend to think that it can "only" happen once in a 100-year period; and that's untrue. For example, it could happen once a week every week from now till Kingdom Come (unlikely but possible).

Divide the number 100 by the % chance to get the recurrence interval. 100/1 is 100; that's the 100-year. 100/0.5 is 200; that's the 200-year. The 50-year is a 2% chance flood. Not as elegant as the Ideal Gas Law eh?

CalMeacham, also the dam had previously been repaired and the top raised; the materials used were not suitable fill (straw, debris, etc.).

History buffs may want to poke around the links provided; the Railroad interviewed many employees who were working at the time. Fascinating transcripts of those interviews are available on the 'net.

Originally Posted by BobLibDem

NinetyWt, reading your post makes me thankful I only took one course in fluid mechanics. Are you a hydrologist or a civil engineer that specializes in hydraulics?

Ha. Now you know how I feel when you wax on about modulii of elasticity and whatnot. I'm a civil engineer specializing in H&H; I consider myself a hydrologist of sorts. Flood studies and dam breach analyses make up a lot of our work. Over the last two years I have been subcontracting with another firm to do dam breach analyses; together we have done 40+, most of them in central Mississippi.

[NinetyWt]

Originally Posted by Snarky_Kong

Keep in mind that there is a fairly large amount of uncertainty about what exactly those storms consist of.

So true. We depend on measured data yet there is error inherent in the measuring. The accuracy of the data improves as the period of record lengthens and the number of gages increases.

Related interesting factoid: areas downwind of large urban centers frequently experience larger rainfalls than would be expected.

Also, the 1%-chance rainfall may not necessarily produce the 1%-chance flood. Many other factors come into play such as antecedent moisture condition, % of impervious areas in the watershed, changes to storage areas in the watershed, changes to channel capacity, etc. etc.

[Snarky_Kong]

Originally Posted by NinetyWt

Related interesting factoid: areas downwind of large urban centers frequently experience larger rainfalls than would be expected.

Is this because the warm air from the city pushes the moist air down and condenses the moisture?

[NinetyWt]

Originally Posted by Snarky_Kong

Is this because the warm air from the city pushes the moist air down and condenses the moisture?

Alas, I have not followed the literature on it, so I don't know what has been theorized about the mechanism causing it. I agree that the warm city air has something to do with it. I would be tickled if someone else had time to find out.

I'm also vaguely remembering some place in the Southeastern U.S. which has abnormally high rainfalls (nothing to do with an urban center). I'll ask around and see if I can come up with the location.

[Santo Rugger]

Originally Posted by NinetyWt

The 1% chance flood is that which has 1% chance of being equalled or exceeded within any given year. Formerly referred to as a 100-year flood because we expect it to occur at least once within a given 100-year period of time. I don't like to use '100-year' because people tend to think that it can "only" happen once in a 100-year period; and that's untrue. For example, it could happen once a week every week from now till Kingdom Come (unlikely but possible).

Divide the number 100 by the % chance to get the recurrence interval. 100/1 is 100; that's the 100-year. 100/0.5 is 200; that's the 200-year. The 50-year is a 2% chance flood. Not as elegant as the Ideal Gas Law eh?

Ugh, don't remind me! Although the MatLab code did end up coming out very pretty, if I must say so myself.
I was curious about the nomenclature, because I'm working with some airplane crash data right now where the expected frequencies are in the range of 10^-4 to 10^-7, depending on the type of plane (carrier/taxi/military, etc.). Our old tables described them as once in 100 operating years to once in 10,000, but IIRC we've gone away from that definition and towards the one you and Snarky have described.

Ha. Now you know how I feel when you wax on about modulii of elasticity and whatnot.<snip>

I, for one, dig 'em all.

[CalMeacham]

NinetyWt writes:

CalMeacham, also the dam had previously been repaired and the top raised; the materials used were not suitable fill (straw, debris, etc.).

I looked over the site to refresh my memory (It's been ages since I red McCullouch's book, or visited the actual dam), and was surprised to find the website say that the dam was actually lowered, rather than raised the dam to allow a road to go over it; this made it easier for water to flow over the top. The addition of the fish weir didn't help (it trapped debris rather than letting it go over the spillway), and other changes made the dam less stable.

[NinetyWt]

Originally Posted by CalMeacham

I looked over the site to refresh my memory (It's been ages since I red McCullouch's book, or visited the actual dam), and was surprised to find the website say that the dam was actually lowered, rather than raised the dam to allow a road to go over it; this made it easier for water to flow over the top. The addition of the fish weir didn't help (it trapped debris rather than letting it go over the spillway), and other changes made the dam less stable.

Oops, I got that backwards didn't I? I was remembering that the elevation of the auxillary was higher compared to the top of dam. That is, of course, because of said lowering. My apologies for the mistake.

[Magiver]

We have a large flood control dam system in my area and the dams are designed as a 2 stage system. Stage one is a fixed flow outlet at river level. It is designed to flow only enough water that the downstream levies can handle. The 2nd stage is an overflow that is designed to protect the earthen dam from damage and thus prevent a catastrophic dam rupture. Example