Formed by Megafloods, This Place Fooled Scientists for Decades

Geologists couldn't account for the strange landforms of eastern Washington State. Then a high school teacher dared to question the scientific dogma of his day.

In the middle of eastern Washington, in a desert that gets less than 20 centimetres of rain a year, stands what was once the largest waterfall in the world. It is nearly 5 kilometres wide and 121 metres high—ten times the size of Niagara Falls—with plunge pools at its base suggesting the erosive power of an immense flow of water. Today there is not so much as a trickle running over the cataract’s lip. It is completely dry.

Dry Falls is not the only curiosity in what geologists call the Columbia Plateau. Spread over 40,000 square km are hundreds of other dry waterfalls, canyons without rivers that might have carved them (called “coulees”), mounds of gravel as tall as skyscrapers, deep holes in the bedrock that would swallow entire city blocks, and countless oddly placed boulders. All across south-east Washington, fertile rolling hills border eroded tracts of volcanic basalt, as if Kansas farmland and Utah canyon land had been chopped up and sewed together into a topographic Frankenstein.

The first farmers in the region named the rocky parts “scablands” and dismissed them as useless as they planted their wheat on the silt-rich hills. But geologists were not so dismissive; to them, the scablands were an enigma. What could have caused this landscape? It was a question hotly debated for several decades, and the answer was as surprising and dramatic as Dry Falls itself.

Tall as a five-story building, this wall of volcanic basalt in Drumheller Channels took shape 10 million years ago as lava cooled, shrank, and cracked vertically. Massive floods later ripped away sections, creating this pillared landmark.

For that matter, so was the source of that answer: a high school science teacher named Harley Bretz. In 1909, the Seattle teacher visited the University of Washington to see the U.S. Geological Survey’s new topographic map of the Quincy Basin, a large area on the west side of the Columbia Plateau. He was 27, with no formal training in geology, but when he looked at the map, he noticed a striking feature: a huge cataract (much like Dry Falls) on the western edge of the basin, a place where water appeared to spill out of the basin and into the Columbia River, gouging a canyon several hundred metres deep. The falls would have been bigger than Niagara, but there was no apparent source of water for them—no signs whatsoever of a river leading to the cataract.

Bretz asked faculty in the department about the feature, called Potholes Coulee, but they had no answers for him. Nor could they explain many of the other unusual features of the region. That’s when, as legend has it, Bretz decided to become a geologist. He earned his Ph.D. in geology from the University of Chicago four years later, changed his professional name from Harley to “J Harlen” to sound more respectable, and in 1922 returned to eastern Washington to take a closer look at the plateau and its scablands. And after two seasons in the field, his conclusions shocked even himself: The only possible explanation for the all the region’s features was a massive flood, perhaps the largest in the Earth’s history—“a debacle which swept the Columbia Plateau,” ripping soil and rock from the landscape, carving canyons and cataracts in a matter of days. “All other hypotheses meet fatal objections,” he wrote in a 1923 paper.

Carved by repeated flooding, a horseshoe-shaped canyon called Potholes Coulee lies along the Columbia River. Raging water dropped 850 feet in less than three miles here, stripping away topsoil and eroding the underlying basalt.

It was geological heresy. For almost a century, ever since Charles Lyell’s 1830 text Principles of Geology set the standards for the field, it had been assumed that geological change was gradual and uniform—always the product of, as Lyell put it, “causes now in operation.” And floods of quasi-Biblical proportions certainly did not meet that standard. It didn’t matter how meticulous Bretz’s research was, or how sound his reasoning might be; he seemed to be advocating a return to geology’s dark ages, when “scientists” used catastrophic explanations for the Earth’s features to buttress theological presumptions about the age of a Creator’s divine handiwork. It was unacceptable. How did canyons and cataracts form? By rivers, of course, over millions of years. Not gigantic floods. Period.

So in 1927, after Bretz had published yet another paper about the “Spokane Flood” and the landscape it carved, the nation’s geological bigwigs invited him to Washington, D.C., to present his findings—and receive his beatdown. Bretz was game, and explained to the expert assemblage how a massive ice-age flood had carved three parallel tracts of flood channels south of the Cordilleran ice sheet (which covered Canada and the northern United States), pooled in a temporary lake twice the size of Rhode Island at the southern edge of the scablands, and then drained like an overflowing tub into the Columbia River Gorge. On the way, the floodwaters carved the famous Grand Coulee, a canyon up to three miles wide with walls up to a thousand feet high, cut hundreds of waterfalls, washed away entire hillsides, deposited gravel bars hundreds of feet high, carried rocks the size of cars and even small houses, and created a terrain of braided channels across eastern Washington.

Floodwaters scoured this chaotic labyrinth of channels into the bedrock of Babcock Bench, perched 600 feet above the Columbia River.

A wall of basalt at Frenchman Coulee lures rock climbers and shows the shrinkage cracks that formed in cooling lava millions of years ago.

Rivers and streams could not have done this, Bretz said. The landscape bears none of the marks of riverine systems, with smaller tributaries joining into larger ones, forming tree-like, branch-and-trunk patterns. Instead, you see a pattern of braided channels—the crisscrossing pattern that flowing water creates when it makes its way across fresh terrain. The difference between the channels we typically see—say, after a rainfall or on the margins of a flooding river—and the channels in the scablands, is simply scale. These are just much larger and were carved into rock instead of sand or silt.

The key to the rapid erosion, Bretz said, was the volcanic basalt that forms the bedrock of the Columbia Plateau. When basaltic lava cools into rock, it forms vertical hexagonal pillars that have weak bonds to each other. Compared to, say, granite, which erodes grain by grain, basalt can erode chunk by chunk as these pillars separate. So a massive, high-energy flood could pluck apart the bedrock so quickly that a canyon like the Grand Coulee might be formed virtually overnight.

“During the ice-age floods, this entire scene was submerged beneath hundreds of metres of water,” says geologist Bruce Bjornstad. The Palouse River, shown here, "was hijacked and forced to follow a new route to the Snake River.”

Bretz’s research was thorough, and his map of the channelled scablands was so accurate that it’s a virtual tracing of modern-day satellite images, creating the immediate impression of channelled floodwaters. But his audience—none of whom had visited, much less studied, the scablands—was having none of it. Bretz’s hypothesis was not just “wholly inadequate,” in the words of one critic, but “preposterous” and “incompetent.” Compounding the problem of his unlikely hypothesis was the question of where all this water would have come from, and Bretz had no convincing answer.


For more than a decade afterwards, Bretz was on the losing side of a pre-ordained conclusion, as the other geologists who began studying the area concocted one laboured hypothesis after another for how the scablands’ features might have been created by gradual erosion. Then, in the early 1940s, the other shoe dropped: Joseph Pardee, a geologist for the USGS, reported that he’d discovered strong evidence of a massive flow of water in western Montana: a swath of current ripples 9 to 15 metres high—like the sand ripples that might form in river or tidal water, but made of gravel and orders of magnitude larger. Their source? A giant ice-age lake—Glacial Lake Missoula—that formed when the Cordilleran ice sheet progressed south and blocked the Clark Fork river valley, forming a dam of ice 610 metres high.

Behind that dam, water from the Clark Fork gathered, forming a lake with as much water as Lake Erie and Lake Ontario combined, stretching for hundreds of kilometres in Montana’s mountainous river valleys. Then the dam broke, and a torrent of water with ten times the combined flow of all the world’s rivers barreled into eastern Washington, reaching speeds approaching 128 kilometres an hour, decimating the terrain and leaving giant current ripples and gravel bars in its wake.

Rich soil called Palouse loess covers the rolling fields of eastern Washington. “This is what the topography might have looked like before the floods removed the loess,” says Bjornstad.

A farmer rakes hay into windrows near Moses Lake. The rows follow the circular pattern of pivot irrigation.

It would take another two decades to win the establishment over, but for many geologists, this was convincing evidence that Bretz’s flood was real. The impossible had happened after all.


It takes practice to see the world as a geologist does. When I got my first glimpse of the Channelled Scablands more than 20 years ago on Interstate 90 west of Spokane, I was struck by their strange beauty, by the way rolling fields of wheat could suddenly yield to a landscape of rocky buttes. I had no explanation for the terrain, and I didn’t need one—I had that primitive eye that looks at rocks and just sees rocks. But when I returned to the scablands with Bretz’s story in mind, suddenly I was in an entirely different world.

Plummeting nearly 61 metres, Palouse Falls is a trickle compared to the megafloods that carved this canyon and shaped the surrounding landscape of eastern Washington State.

Standing in the middle of a broad swath of scablands extending from horizon to horizon, my mind’s eye could clearly see the floodwaters blasting through, like a raging inland sea, ripping up everything not strong enough to stay moored. Driving through what’s known as the Ephrata Fan, a broad open area where floodwaters left the confines of the Grand Coulee and spread out and slowed as they neared what would become Ancient (and very temporary) Lake Lewis, I easily understood why the landscape was riddled with boulders: As the water lost speed, it began dropping all the rocks it was carrying. And when I stood on the lip of the dry falls of Potholes Coulee, looking at this immense canyon with farmland on three sides and a precipitous drop on the other, I felt what Bretz was thinking when he looked at that map a century ago: If a river didn’t carve this, what did?

With the flood story in mind, it all seems so obvious—so obvious, in fact, that it’s almost impossible to see the terrain and not see the floodwaters that shaped it. Why, then, were the experts in Bretz’s day so blind to what now seems like a self-evident geological record? I posed that question to Vic Baker, a geologist with the University of Arizona who became the pre-eminent scablands expert in Bretz’s wake when we met to tour several of the region’s features. “It’s the mistake people have made most in the history of science,” he said. “They forgot that nature has the answers, not us.”

“Bretz was making arguments, and no one was going into the field to see anything,” Baker said. “They were just countering his arguments with theory.” And because scientists are first and foremost human beings, they loathe to change their theories or their minds because of mere data.

The basalt and granite boulders now littering the Ephrata Fan were carried there by torrents of water that gushed out of a canyon called the Grand Coulee. The largest piece of rock is more than 7.5 metres tall.

Baker told me a story as we looked out at Palouse Falls, another dramatic cataract at the head of a massive canyon, with a stream running through it that seems comically out of scale, like a toddler wearing a grown man’s boots. Sometime in the late 1950s or early ’60s, a geologist named Aaron Waters brought one of Bretz’s most vocal critics—James Gilluly, the one who’d called his ideas “preposterous” and “incompetent”—to the scablands for a first-hand look. As they took in the sight of the falls and the canyon, Gilluly was dumbfounded by their scale. “Gilluly was just quiet the whole time,” Baker said, “and as they were leaving, he broke out into this immense laugh and said, ‘How could anybody be so wrong?’” After resisting Bretz’s theory for decades, simply seeing the landscape with his own eyes had changed his mind.

Of course, for some of Bretz’s most stubborn critics, even eyewitness experience wasn’t enough. Bretz’s arch-adversary, Richard Foster Flint, a Yale geologist who remained a premier authority in the field until the 1970s, spent years studying the scablands and resisted Bretz’s theory until he was virtually the only one left who did. He finally acknowledged the scablands flooding (grudgingly, with a single sentence in a textbook in 1971), but as philosopher Thomas Kuhn observed, new scientific truths often win the day not so much because opponents change their minds, but because they die off. By the time the Geological Society of America finally recognised Bretz’s work with the Penrose Medal, the field’s highest honour, it was 1979 and Bretz was 96 years old. He joked to his son, “All my enemies are dead, so I have no one to gloat over.”

It is tempting to see this story as a simple morality tale, with “good guy” geologists lining up against “bad guy” geologists in a battle between open-minded inquiry and closed-minded dogmatism. But that might just compound the error because it neglects the fact that scientists almost always favour their own theories over others’, and rarely are those theories completely right. Enter Richard Waitt, a geologist with the USGS. In 1977 Waitt was exploring the Walla Walla valley in southern Washington when he noticed that one of the 40 sediment layers from the temporary flood lake contained ash from an eruption of Mt. St. Helens. It had been assumed that all those layers had been laid by one flood event—but if only one of them had the volcanic ash, it meant that each of those layers must have represented a separate flood.

“I knew right away that there couldn’t have been just one flood,” Waitt said. But when he published his findings in 1980, arguing that there had been at least 40 ice-age floods in the scablands, he faced such stiff resistance that he felt like Bretz himself. “Baker and his students were totally against it for years,” he said. And the irony for Waitt is that the lines seemed to be drawn just as they had been during the initial controversy. The authorities in the field were invested in a particular theory, and contrary evidence was dismissed without an adequate hearing.

A tower of basalt stands as a testament to the forces of nature that created the Scablands. “The region is unique: let the observer take wings of the morning to the uttermost parts of the earth: he will nowhere find its likeness,” wrote J Harlen Bretz, the geologist who first described the flooding that sculpted this terrain.

It turns out that Waitt was right. In fact, subsequent research indicates that 80 or more floods ravaged the scablands near the end of the last ice age. Repeatedly over a two- to three-thousand-year span ending roughly 13,000 years ago, the Cordilleran ice sheet advanced to block the Clark Fork river, a new iteration of Glacial Lake Missoula formed, and then the ice dam broke, each time unleashing such a torrent of water that if it were to happen today, most of Portland’s skyline would be submerged by the floodwaters. What’s more, something similar might have happened during previous ice ages—meaning that perhaps the most dramatic features of the scablands, like Grand Coulee and Dry Falls, didn’t form in the blink of a geological eye after all, but were shaped by catastrophic erosion over an extended period of time. Which would make both Bretz and his early critics right—Bretz about the flooding, and his critics in their sceptical assessment of his timetable.

This wouldn’t have come as a complete surprise to Bretz. By the early 1950s he’d noticed that some scabland features appeared to be more weathered than others, and in his last paper on the subject, in 1969, he argued that there had been at least seven scabland floods. But by then the controversy that had defined his professional life had already come and gone. When I asked Waitt about the irony of Bretz’s story, he said, “I think if Bretz could have made the argument in the 1920s for several floods, it would have muted the opposition a great deal.”

Perhaps it’s just as well that he didn’t. That sort of neat resolution might obscure what’s arguably the most important lesson of the scablands’ story—the caution that “nature has the answers, not us.” Just when we think we’ve got nature figured out, we find that among her many powers is the power to confound us, again and again, and again.

Header Image: Remnant of a lost landscape, this island of ancient soil—crowned by a crop of wheat—survived the ice-age floods that sculpted the region known today as the Channelled Scablands. PHOTOGRAPH BY MICHAEL MELFORD, NATIONAL GEOGRAPHIC

Discuss this article


Never miss a Nat Geo moment

Your email address