It’s late spring. An especially big mountain snowfall is melting. The Columbia River is running at its strongest and coldest of the year. Dams are spilling water, their reservoirs virtually full. It’s a weekend, so riverside vacation homes are lively. Boats are putting in. People are running their dogs along shorelines.
Suddenly, a 7.0-magnitude earthquake, centered near Entiat, shakes Rocky Reach Dam off its bedrock anchors. Concrete crumbles. A wall of water swirling with debris plows downstream sweeping away docks, riverside homes, entire swaths of prime orchard. The wave swamps Wenatchee and heads downstream, gaining in destructive force.
But how likely would this really be?
A comprehensive, five-year study finished last year for the region’s three PUDs concluded that the ground-shaking potential from a big earthquake in North Central Washington is about three times greater than was thought possible when the Mid-Columbia River dams were built decades ago.
That’s like going from a mild quake with shaking similar to the effect of a passing, heavy truck to a quake capable of causing partial collapse of ordinary buildings.
Even so, the region’s hydro engineers say that their own modeling and case studies from all over the world point to the extreme unlikelihood of such a dam failure happening here.
“There would be no wall of water,” says Bill Christman, a hydro engineering manager at the Chelan County PUD. “If you just plucked that dam out of the water, there would be a big water surge, but that’s not how dams fail.”
In fact, big concrete hydroelectric dams — like Grand Coulee, Chief Joseph and the five PUD-owned dams that follow along the Columbia’s next 142 miles — don’t fail very often.
Anchored to the bedrock with stout steel bars and concrete “keyed” into specially made trenches, most of these massive structures were built to withstand the powerful flows of the then-seasonally raging Columbia River.
Rocky Reach Dam, for example, is built to withstand river flows of 1.2 million cubic feet per second (cfs). Today, peak seasonal flows at the dam rarely exceed 310,000 cfs.
Canada’s massive water-storage reservoirs, built in the late 1960s for down-river flood control and power generation, tamed the mighty river and transformed it into a world model for hydroelectric production with 14 mainstem dams and billions spent annually for fish survival.
These big dams could sustain damage in a big quake, local engineers say, but they’d generally keep their shape and contain their reservoirs, even when subjected to violent shaking from quakes measuring 8.0 or more on the Richter Scale.
Taiwan’s Shih Kang Dam offers a spectacular example.
An article in the September 2011 edition of International Water Power and Dam Construction magazine — the kind of stuff Christman reads — examines case studies from six of 19 dams worldwide that were shaken by earthquakes stronger than would likely occur here.
The Shih Kang, the first concrete dam reported to have failed in an earthquake, broke apart in a 7.6-magnitude quake in 1999. The dam was built over a unknown branch of a well-known nearby fault. The branch fault ruptured, causing the ground to displace some 36 feet.
Despite the severe damage, the dam didn’t completely lose its reservoir, the report says. Flow under the mangled spill gates was gradual, not catastrophic.
As far as anyone has yet detected, no faults are below any of Washington’s Columbia River dams, the engineers say.
And Canada’s giant water-storage dams and reservoirs, although made largely of earth, not concrete, were built to withstand earthquake forces currently considered possible only every 10,000 years.
“In general, the dams are very, very robust,” agrees Kevin Marshall, Christman’s counterpart at the Grant County PUD. “It’s not like a building. The building codes have you design a skyscraper so you can exit in a one-in-500-year event. We hold the dams to a higher standard. That ‘wall of water’ is a very, very low probability thing.”
More likely
In 2009 the Portland-based Bonneville Power Administration (BPA) coordinated a simulated failure of Canada’s Keenleyside Dam. All Columbia River dam operators participated. The U.S. Army Corps of Engineers built a computer model showing effects from the potential, downriver flooding.
“They set up a command center and did everything except actually release water from the dam,” Christman said. “When the first calls were made, the dominoes fell the way you’d want them to.”
A failure at Keenleyside, the model showed, would cause water levels to rise gradually over minutes and hours, but the floodwaters lose force as they head downstream, flow against rocks and vegetation, fan out along the width of the river, and become absorbed by the reservoirs of down-river dams, Christman said.
It doesn’t hurt that most of NCW’s Columbia River shoreline area is very sparsely populated.
Wenatchee, built above river level, would not be swamped, the model shows.
Nor, likely, would Crescent Bar, a Columbia River island about 25 miles southeast of Wenatchee, along the eastern shoreline of Wanapum Dam’s huge reservoir.
“We have a lot of room for water to rise without damaging homes,” Christman said. “A big (seismic) event would no doubt be a massive hassle, but not a tragic failure.”
New findings
Seismic analysis is nothing new to hydro engineers, nor is retrofitting dam components to make them more earthquake resistant as new information becomes available, Christman and Grant PUD’s Marshall say.
For the latest five-year study, each of the PUDs hired its own expert who studied a region stretching roughly from the Canadian border south to Oregon.
Unlike older types of analysis, which would focus on single elements, like faults, and their potential to cause earthquakes of certain magnitudes, the researchers took a broader approach thought to produce more precise results.
This kind of study, called a “probabilistic assessment,” takes into account all the possible earthquake sources in a defined area of study. It examines not only the potential for quakes and the ground movement they’d produce in the vicinity of the dams, but also likelihood those quakes would occur.
Martin McCann, a professor of structural and seismic engineering at Stanford University, an expert in the probabilistic study method, combined the results into a single report, considered the most comprehensive to date on North Central Washington seismicity.
The study shows that a 6.0-magnitude earthquake centered relatively close to Rocky Reach Dam is possible, Christman says, but it’s expected to happen only once every 2,500 years.
The region around Entiat, north of Wenatchee and very near Rocky Reach, has an important historical reference point — a 6.8- to 7.2-magnitude earthquake, thought to be epicentered around what is today the city of Chelan, that shook the region in 1872.
The Entiat area, including a landmark just north dubbed “Earthquake Point,” continues to be very active. Seventeen minor quakes measuring 0.4 to 1.9 on the Richter Scale have already been recorded in the area this year through February on both sides of the Columbia.
But an absence of mapped faults in this area have left researchers wondering what actually caused that big, historical quake and exactly where it was centered.
That uncertainty, combined with that known quake’s estimated magnitude, lead researchers to assign the area a slightly higher ground-shaking potential than farther south around Grant PUD’s Wanapum and Priest Rapids dams, where more fault lines have been recorded and more tremors occur, but no major quake has been recorded.
Assessing weaknesses
The latest seismic assessment is based on the best scientific information available, but “it doesn’t come with a warranty,” Christman says.
Federal dam regulators are now reviewing the results.
The PUDs will spend the next several years putting their new knowledge to use, identifying their utilities’ vulnerabilities and deciding how to address them.
At the Chelan PUD, that will include a look at an earthen section of Rocky Reach dam along the Douglas County shore to gauge its potential to breach and potentially flood an area below a nearby switchyard.
At Rock Island Dam, just down river, ground shaking could still damage a critical dam-access road that has already been retrofitted to make it more resistant. A breach in this area could threaten part of busy Highway 28, nearby.
Marshall says Grant PUD engineers will look at gates and embankments, along with a potentially vulnerable section of Wanapum Dam designed for future addition of more turbines and generators.
If the threat merits action and justifies the cost, the utilities could either make the structures more robust, or make sure they’re well equipped for “triage” — stocked with enough replacement parts to keep the system working.
Douglas PUD officials say they haven’t identified vulnerabilities at Wells Dam, but will wait for feedback from the feds before acting.
“An earthquake is a sudden thing, and once you have one you either discover that you’re OK or are going to have a long and difficult recovery,” Christman says. Chelan PUD’s attitude is that we want to make sure we’re safe and robust before it ever happens.”
Marshall agrees. “We’re constantly looking at the dams to make sure they’re safe. If we think we need to do something to make them safer, we’ll be doing it.”



