Hydroelectric Near-Miss - The Glen Canyon Dam

The Glen Canyon dam is a very large dam on the upper Colorado River, on the northern border of Arizona.  The Glen Canyon dam impounds the Colorado River in a man made reservoir named Lake Powell.

Below:  The vast reservoir of Lake Powell.  The Glen Canyon dam is at the bottom left.  Click any image to enlarge.

Below, a closer view of the Glen Canyon dam from above.

The Glen Canyon Dam is a gravity arch dam, very similar in design to the more famous Hoover Dam.  Although it's 30 years newer, it's still quite similar in appearance.  Glen Canyon Dam is upstream of Hoover Dam, and impounds water for states in the Upper Colorado River Basin (Western Colorado, Utah, and New Mexico)

In 1983, Glen Canyon dam was nearly overtopped.  The photo below shows a small dam with a tiny reservoir being overtopped.  Even this image of a very small dam being overtopped is unnerving.  This is the San Clemente dam on the Carmel River in California.  The dam has since been removed.

Overtopping is very dangerous, because the overtopping water can erode bedrock adjacent to and/or underneath the dam, leading to tunneling, which of course leads to complete failure...

...and failure brings us back to Glen Canyon Dam, because it nearly did fail in 1983. 

I will now quote Wikipedia (while Bolding the most interesting parts)

During the El Niño winter of 1982–1983, the Bureau of Reclamation predicted an average runoff for the Colorado River basin based on snowpack measurements in the Rocky Mountains. However, snowfall during April and May was exceptionally heavy; this combined with a sudden rise in temperatures and unusual rainstorms in June to produce major flooding across the western United States.^[83]

With Lake Powell nearly full, the USBR did not have enough time to draw down the reservoir to accommodate extra runoff. By mid-June, water was pouring into Lake Powell at over 120,000 cubic feet per second (3,400 m³/s). Even with the power plant and river outlet works running at full capacity, Lake Powell continued to rise to the point where the spillways had to be opened. Other than a brief test in 1980, this was the only time the spillways had ever been used.^[84]

At the beginning of June, dam operators opened the gates on the left spillway, sending 10,000 cubic feet per second (280 m³/s), less than one-tenth of capacity, down the tunnel into the river below. After a few days, the entire dam suddenly began to shake violently. The rumblings were so notable that a worker in the employee dining room, located near the power plant, was reported to say that it "sounded like the barrages that he had experienced in Vietnam". ^[16][17]  

The spillway was closed down for inspections and workers discovered that the flow of water was causing cavitation – the explosive collapse of vacuum pockets in water moving at high speed – which was damaging the concrete lining and eroding the rock spillway tunnels from the upper ends of the diversion tunnels, which connect to the bottom of the reservoir.^[85] This was rapidly being destroyed by the cavitation and it was feared that a connection would be made to the bottom of Lake Powell, compromising the dam's foundation and causing the dam to burst.^[85]

(Brief aside from the Wikipedia article - below is a profile of the Glen Canyon dam spillway.  There was concern that ongoing cavitation would erode the plugs in the diversion tunnels at the bottom left.  At that point the lake would uncontrollably drain Lake Powell from underneath the dam, destroying the dam in the process)

Back to the Wikipedia article:

Meanwhile, snow continued melting in the Rockies and Lake Powell continued to rise rapidly. To delay having to use the spillways, the USBR installed plywood flashboards (later replaced by steel) atop the gates to increase the lake level.^[86] Even this additional capacity was exhausted; discharges through the left spillway reached 32,000 cubic feet per second (910 m³/s), and the right spillway was opened to 15,000 cubic feet per second (420 m³/s). At Lee's Ferry, the Colorado River peaked at 97,300 cubic feet per second (2,760 m³/s), which was and still is the highest water flow recorded there since the dam was built.^[87] On July 14, Lake Powell reached 3,708.34 feet (1,130.30 m) elevation, a level that has not been exceeded since.^[88]Just as it seemed inevitable that the dam would fail, inflows fell and the dam was saved. Upon inspection, it was found that cavitation had caused massive gouging damage to both spillways, carrying away thousands of tons of concrete, steel rebar and huge chunks of rock.^[89]

After the flood, it was suggested that the flashboards atop the spillway gates should be replaced with stronger boards and kept permanently; this would allow an "insurance" against a 1983-reminiscent flood.^[15] An air slot was constructed in each spillway tunnel afterwards to prevent catastrophic cavitation events like those of 1983. Most surprising, however, is that the flood of 1983, although it nearly caused catastrophic disaster, it was a "relatively small flood".^[19] It was, in fact, only a 25-year flood, or a flood that has a four percent chance of occurring in any given year.^[11][17]

Below is an image of the cavitation damage inside the left spillway after the 1983 event.

Point of interest:  The reason that the small San Clemente dam shown earlier in this post was over-topped was due to silt build-up in the reservoir.  This caused a massive loss of reservoir capacity, from the initial 1425 acre feet capacity to just 70 acre feet, in its 87 year life-span.  Due to the very high silt levels in the Colorado River, about 100 million US tons of silt settle in the reservoir annually.  That's about 30,000 dump trucks per day that settle into the reservoir, so capacity is being lost.  More importantly, without dredging, the silt will eventually block the dam's outlets, reducing its ability to store and release water.

Below is a 3 part government video of this event.  What these videos never say is how close to failure the dam was, nor how much danger people downstream of Glen Canyon dam were in.  It's great video though, because it examines in detail how much damage occurred when the dam's spillways had to deal with a 25 year flood.  You would think that a dam should be able to release *a lot* of excess runoff without running the risk complete failure.

I'm beginning to suspect that one of hydro-engineering's dirty secrets is that many dams lack the ability to safely and adequately bypass large quantities of runoff. 

On the bright side, nobody lost their life.  We will look into a bit of that in future posts.  Some dam failures have caused appalling loss of life.

Part 1

Part 2

Part 3

If you live downstream from a dam, keep a raft handy!