The Following excerpt describes the geologic landscape of the Columbia Plateau. It is taken from pages 120-131 of William Dietrich’s Northwest Passage: The Great Columbia River (New York: Simon & Schuster, 1995). Courtesy of William Dietrich

The age of the Columbia River is not known, but the landscape through which it flows and which directs its passage is relatively young. The Colorado has cut Grand Canyon rock that dates back nearly two billion years. Some of the Columbia’s volcanic landscape can be measured in a few centuries or millennia, and the great floods that modified its landscape ended just twelve thousand and eight hundred years ago. The oldest rocks in the Columbia River Gorge date back just 50 million years₁ and far younger volcanic craters, some still barren and others only slowly being recolonized by trees, dot the river basin. Twenty-two dormant volcanoes and two hundred and twenty-five basalt vents have been counted around the gorge. Lava beds pocketed with craters, fissures, and caves pock the forest north of the river.

Not only is the Columbia in a young part of the world; it occupies a stage of geologic drama so extreme that for fifty years it defied belief. Here were mountains such as Oregon’s Mazama that blew up like hydrogen bombs, sending hot gases rolling across the landscape for hundreds of miles and leaving behind Crater Lake. Here were eruptions of molten rock that rolled up to four hundred miles across the landscape, pooling like hot pudding into vast basins that eventually covered an area almost as big as New England and New York State combined. Here were water floods so vast they contained ten times the flow of all today’s rivers taken together, or the equivalent of sixty Amazon Rivers roaring across the landscape at once. On the Snake River plain were shield volcanoes so numerous they overlapped each other like a pile of coins, and along the Columbia rock slides so big that they briefly dammed the river. This is geologic history as grand opera, with smoke, cymbals, and kettle drums. No-this is geologic history as MTV, all quick cuts and dizzying flicker. Compared to the Columbia, much of the rest of the continent is a stately nineteenth-century novel. When the early Appalachians were 600 million years old, the Pacific Ocean still covered the Columbia’s future course.

Coming to grips with the fluidity and violence of this geological landscape was not easy, even for scientists. Two twentieth-century geological theories that are central to the Columbia’s story took a half century to be fully accepted by the profession. The first was the idea of continental drift, proposed by the German Alfred Wegener in 1912. He theorized that the continents are moving, and that the collision of continental and oceanic plates could explain much of the world’s topography. But while Benjamin Franklin had speculated as early as 1792 that parts of the surface were floating on a liquid interior, Wegener proposed that the continents somehow plowed across solid basalt. This was too much for most geologists to accept. Ridiculed for years, the German explorer died of exposure on Greenland’s Ice Cap in 1930 long before his theory was fully accepted.

Continental drift explained too much, however, particularly the jigsaw puzzle fit between Africa and South America. Oceanographers found a mid-Atlantic ridge like a seam between the two hemispheres, and Wegener’s idea was slowly modified to include a mantle that was solid but plastic and mobile enough to account for the motion of continents. In the 1960s it was refined to the theory of plate tectonics. Two hundred million years ago, most of the Columbia Basin was underwater. The collision of western North America with Pacific Ocean plates shoved up the land that the Columbia would subsequently traverse.

A second heretical idea was proposed in 1923 by I. Harlen Bretz, a Seattle high school biology teacher turned amateur geologist during the summer. Bretz became so enamored of earth science that he returned to the University of Chicago to earn a doctorate in geology and eventually become a professor there. Subsequent summer exploration of Columbia Plateau scablands made him both famous and controversial. In the previous century geologists and much of the general public had come to embrace one revolutionary proposal, that the earth was not created a biblical six thousand years before (the vice-chancellor of Cambridge University had more precisely determined the moment to be 9:00 A.M. on October 26, 4004 B.C.) but was instead staggeringly, unimaginably old, with 4.6 billion years the current estimate. With this realization came the theory that much of the planet’s land was shaped slowly and uniformly by erosion and mountain building: time enough for the Grand Canyon to be carved, Mount Everest raised, and all the diversity of life produced by the patient tinkering of evolution. “Gradualism” replaced Genesis and Noah’s Flood, and any hint that sudden catastrophe had also shaped our world smacked of biblical literalism. Then came Bretz, hiking out of the hot coulees with his notebook full of curious observations. He proposed that much of the Columbia Plateau and the river’s downstream drainage had been carved and formed in days or weeks by titanic Ice Age floods greater than any the world has since seen.

Geologists scoffed. None had seen the ravaged ground Bretz explored, and Bretz himself could not explain where such volumes of water came from. As an alternative, many scientists posed gouging of the plateau by glaciers. Then, at the 1940 meeting of the American Association for the Advancement of Science, Joseph Thomas Pardee announced he had found the water Bretz needed. A lake with half the volume of Lake Michigan had formed over western Montana when a glacier had dammed Ice Age rivers. The collapse of the glacier, Pardee quietly noted, could have provided the flood to accomplish the destruction Bretz had found. In 1952, Bretz led a scientific expedition with several sympathetic geologists across the scablands, and they concluded that not only was there abundant geologic evidence for a single great flood, but for more than forty, possibly as many as one hundred. Acceptance still came slowly. One basic geology textbook was not revised to include mention of the Bretz or Spokane floods until 1971, and not until 1979-when he was ninety-six years old-did Bretz finally receive the nation’s highest geological award, the Penrose Medal of the Geologic Society of America. In this case, his profession progressed considerably more slowly than a Spokane flood.

Keep in mind, then, that the geologic story of the Columbia is new, recently evolved, and will no doubt continue to be refined. Both plate tectonics and Bretz floods, however, help to explain the geographic drama seen in the Columbia Basin, and thus its subsequent human history.

For example, the most obvious peculiarity about the Columbia is its contorted course. The river is beset with a basic topographical dilemma. It must reach the sea, but plate tectonics pushed up mountain ranges that run primarily north and south, forming a series of walls in its path. The Columbia threads its way up and down until it finds weak points in these ranges to pierce. This serpentine persistence makes the Columbia influence a huge area. It also dashed hopes that the Great River of the West would be a straightforward highway of commerce.

About 200 million years ago, when the great age of the dinosaurs was just getting under way, the earth’s land was joined into a single supercontinent geologists have named Pangea. South America and Africa were welded together along the lines still clearly visible today, and North America and Europe were tilted and linked. The western shore of North America was roughly where the Idaho border with Washington and Oregon is today, though the future site of Spokane in Washington was dry and the shore cut back to the east to leave southwestern Idaho under water. Farther north in British Columbia the coastline ran slightly west of today’s continental divide. Washington’s Steptoe Butte near Colfax, called Eomoshtoss by the natives, is a remnant of this ancient shoreline. One can drive to its top, look out across rolling wheat fields, and imagine the swells of the ancient ocean.

If we consider the Columbia the sum of a mountain basin and not just a channel of water, its story begins when Pangea began breaking apart. The giant landmass first fractured into two supercontinents geologists call Gondwanaland and Laurasia and then fragmented into roughly the present continents. The Atlantic Ocean began to grow at the stately rate of two to three inches per year that it continues to broaden by today. At the same time the western shore of North America rode up and over the Pacific plate. The thick mountain belt of western North America seems a reasonable result of this collision when we understand that the continent has overridden fifteen hundred miles of the Pacific Ocean floor since Pangea broke up.

Not all of the melted oceanic plate broke clear to the surface. Sometimes the magma crystallized into granite without erupting, creating huge “batholiths” that make up the mass of much of Idaho and western Montana. Other mountains formed when the top of the sea floor diving under the continental edge was scraped off in the collision and mounded at the coastal fringe. This formed part of today’s Wallowa and Blue mountains in eastern Washington and Oregon, plus the Klamath Mountains in southwestern Oregon.

In Canada ancient sea floor was being heaved up to make more sedimentary mountains, shoving Cambrian sea fossils and reefs half a billion years old some eight thousand feet into the air. Meanwhile, large offshore islands similar in size to the main islands of Japan docked with northeast Washington and southeast British Columbia, forming the Okanogan Highlands and the Columbia Mountains the river winds around. At least three more of these additions occurred, including one 50 million years ago to form the base of today’s North Cascade Mountains and British Columbia’s Coast Range. University of British Columbia geologist Bill Matthews jokes that a proper name for our continent might be “the United Plates of America.”

Behind the docking of the Okanogan microcontinent developed the Rocky Mountain Trench, believed to be the product of mountain fractures and glacial gouging. This is the longest valley of its kind in the world. It runs from Montana to the Yukon-Alaska border, a distance of twenty-one hundred miles, and is the birthplace of the Columbia, Fraser, and Yukon rivers.

As the new island microcontinents welded to North America’s western shore, the ocean trench where sea floor was shoved under the continental plate jumped westward: first to the site of today’s Okanogan Valley, and then to a point off the present-day coast. As the ocean bottom descended under the continent it melted, some of the magma floated upward, and volcanoes erupted to form the Cascade Mountains. This continued until about 25 million years ago, when the volcanoes mysteriously snuffed out.

Even at this relatively late date, much of eastern Washington and Oregon-and almost all the western half of those states-remained under water. A vast bay occupied what is much of the Columbia Basin today. As the Cascades quieted, volcanic activity shifted eastward two hundred miles to the Idaho border region again. Huge cracks split the crust and magma poured out in volumes possibly matched only by India’s Deccan Plain. The magma filled this bay with what is today called Columbia River basalt, creating the Columbia River Plateau.

The new lava flows ultimately covered one hundred thousand square miles to an average depth of nearly a mile. This outpouring consisted of flows spaced over 12 million years and separated by centuries or even millennia. The climate of the interior Northwest was much wetter then, giving sufficient time for lakes and rivers to form on the cooled and hardened basalt surface. Sometimes plants recolonized and animals browsed before the next magma flood came. This deceptive pause must have made the next eruption all the more awesome and catastrophic. The molten basalt had the plasticity of molasses and ran across the landscape at speeds that are currently in dispute: initial estimates pegged them as fast as thirty to fifty miles per hour while more recent studies have suggested a pace ranging from a walk to a slow crawl. In any event the advancing wave would carry a skin of cooling basalt like elephant hide. Some of this would slide forward and fall in front of the flood, and contemporary observation has suggested that unless the flow is at least fifty feet high, this “clinker dam” would eventually build to the point where it would stop the flow. The typical Columbia basalt flood, however, was one hundred feet high and some exceeded two hundred feet. One eruption, the Roza Basalt Member, covered twenty thousand square miles in an area that extends from the vicinity of Grand Coulee Dam to Pendleton, Oregon, a distance of one hundred and fifty miles. The molten basalt entombed everything in its path. Near Park Lake where Grand Coulee Dam enthusiasts held their 1931 rally, a rhinoceros-probably already dead and lying in the mud of a pond-was swallowed by the red-hot tide. The rhino left a cast of its body, later discovered by young boys playing in a cave.

Near Vantage, not far from where Interstate 90 crosses the Columbia River, the flows entombed a swampy hardwood forest of elm, maple, gingko, Douglas fir, walnut, and spruce. The trees, deprived of oxygen, did not rot quickly, giving time for silicon to invade their cells and replicate their structure. The result was petrified wood. Glaciers later scraped away enough rock to expose this time capsule and in the 1930s the Civilian Conservation Corps dug pits to expose various trunks and built a winding trail. There are iron bars over the pits so souvenir-hunting humans won’t peck apart in a few years what nature preserved for a million generations.

The bemusing thing about the site is not just the preservation of ancient trees but the topography of Gingko Petrified Forest itself. What was once a flat swamp is now desert hills. A plateau has risen like bread over the swamp and then been eroded into ridges rounded like melons. This is what time can do, and relatively brief geologic time at that. Towering Mount Rainier, nearly three miles high, may be less than a million years old: a blink, by geologic standards.

These basalt flows pushed the river west toward the eastern Cascade foothills. There it was confined, between new rock and a high place. The Columbia chewed down through the basalt even as new flows added to the plateau’s height. Between Wenatchee and Chelan the river seems almost hurled against the foothill bluffs, pinned there by a wall of lava that in places still seems to bear down on the Columbia.

About 12 million years ago the great basalt flows ceased and volcanic action once more shifted back to the Cascades. The Columbia did not so much carve a path through these rising mountains as maintain one. It had flowed south along the mountains until it found their lowest point and turned west, and as the Cascades rose, the river cut to maintain its own level. This process was uneven, and at today’s Wallula Gap southwest of the Tri-Cities the river ponded behind a barrier of basalt and formed a huge lake. Sediment fell out and when the river finally broke through at Wallula and the lake emptied, centuries of sediment from the dry lakebed blew back over eastern Washington, the loess creating rolling hills hundreds of feet deep in rich soil. Without the Columbia’s net of rivers to carry down glacial flour, pond it, settle it, and finally drain away to allow the wind to redeposit it, the rich Palouse country might be stony basalt.

North America continued its westward march. The Oregon coast range, Willapa Hills and Olympic Mountains began to rise above the sea as scrapings from the Pacific plate. The Willamette Valley, which began its existence as a vast saltwater bay, slowly began to fill with eroded sediment and lava flows from surrounding mountains. The new coast range, for reasons not entirely clear, migrated north about fifty miles. That pushed the Columbia River’s estuary with it, producing the northward bend of the river past Portland and Vancouver that is still seen today.

The Snake River had an equally troubled history. Rising in the Rockies, it initially may have flowed out of southern Idaho and across southeastern Oregon to empty into the huge interior bay west of the Cascades. The uplift of the Blue Mountains, however, barred this straightforward path. The Snake bent north to eventually link with the Columbia. As the land continued to rise, the river continued to cut. The Snake had chosen a difficult course through the new Seven Devils Mountains, a path that forced it to saw a canyon up to a mile and a half deep. “A lesser river might have given up,” noted Bates McKee in his geology text Cascadia, “but the Snake had no place to go. It had to cut the deepest canyon in North America, Hells Canyon, and not through sedimentary strata like those found in the Grand Canyon but through hard, massive greenstone:

Even while the lower Snake underwent this heroic excavation, the upper Snake was besieged by volcanoes breaking out like acne on the Snake River plain. As volcanoes rose the Snake twisted between them, skirting first this fresh lava flow, then that, so that one wall of its canyon is often distinct in geology from the rock on the other side, each having come from a different vent. The resulting mass of basalt from all this volcanism is so fractured that for two hundred and sixty-eight miles, from Henry’s Fork down to lower Salmon Falls, the north bank of the Snake receives no tributaries. The rivers that pour down from the snowy mountains of central Idaho, such as the Big Lost River, disappear into cracks in the earth to form one of the nation’s greatest aquifers₁ the water continuing to migrate until it reappears as springs on the canyon wall below Milner Dam. The springs pump 200 billion cubic feet of water a year into the river.

By the arrival of the Ice Ages of the last few million years, then, the general course for the great rivers of the Columbia Basin has been set. The most dramatic scene, the real show-stopper of this geologic extravaganza, was about to begin. Enter the glaciers.

What a stark and magnificently hostile landscape the great ice sheets must have produced! The Selkirks and Purcells and Monashees and Cariboos were buried by ice, their summits poking into the air like treading swimmers. At the glacial edge the ice wall ground and growled against the tundra-covered basalt of the Great Columbia Plain. Each rhythmic glaciation tended to obliterate the effects of the one before, but when the ice came down from Canada the last time, beginning about eighteen thousand years ago, it set the stage for the greatest known flood in geologic history.

The edge of the ice sheet was twenty-five hundred feet high when it bulldozed down the Okanogan Valley and crossed the canyon of the Columbia River at the site of Grand Coulee Dam. When it struck the Columbia Plateau on the other side, the glacier peeled off house-size chunks of basalt and carried them south toward Waterville, dropping the boulders into future wheat fields. There they sit squat and immovable to this day, forcing tractors to detour around monoliths the farmers call “haystacks.”

The Columbia was unable to quickly push past this glacial dam. It backed up to form Glacial Lake Columbia and then spilled southward to begin carving the Grand Coulee. Another lobe of the Wisconsin glaciation formed an ice dam in the middle of Idaho’s panhandle, backing water up over much of western Montana. Lake Missoula covered the site of the future city of Missoula with water a thousand feet deep. About sixteen thousand years ago, the ice dam failed when the water floated the barrier off its rock base. Five hundred cubic miles of water roared toward eastern Washington in a wall a thousand feet high, moving as fast as fifty-eight miles per hour. The flood contained ten times the volume of all the rivers in the world, and pushed before it a shock wave of air and sound that must have broken over the Columbia Plateau like a thunderclap. Icebergs carrying boulders rode the flood like bobbing container ships, some of them not stopping until they reached western Oregon, four hundred miles away. The Columbia Basin’s most spectacular sculpting had begun.

The water scoured the land like a sandblaster across old paint. When the velocity of a stream doubles, its ability to move material increases up to sixty-four times. Here was a flood with speeds as much as ten times an ordinary river, scraping new coulees with a slurry of rock, sand, and ice. In Grand Coulee it ripped apart basalt at the lava fractures and began eating backwards, the lip of Dry Falls marching back up the canyon. It created today’s Palouse River canyon and falls. It gouged out the Marmes Rock Shelter, where the earliest evidence of humans in the basin would later be found shortly before the site was drowned by Ice Harbor Dam. The flood fanned over the Columbia Plateau in a branching pattern that produced the scablands of the basin, ponded to a depth of twelve hundred feet at Wallula Gap, and then broke through again to crash toward and through the Columbia Gorge. At the future site of Bonneville Dam the flood ran more than eight hundred feet high. Its energy was so great that near John Day River it gouged out a pothole one hundred and sixty-four feet deep, or a few feet below sea level-the home, Native Americans later believed, of a monster called the Swallower who kept an armored sturgeon as a pet. The water’s fury pounded the gorge to create Celilo Falls, sheered off abrupt cliffs to create Oregon’s series of lovely waterfalls, and foamed onward to flood much of the Willamette Valley. As valley water receded, icebergs settled and left behind rounded granite boulders that would puzzle Oregon pioneers sixteen millennia later.

As unbelievably violent as this flood was, it did not just happen once. The glacier pushed a new ice lobe into Idaho’s panhandle, and a new Lake Missoula quickly formed. Then the dam broke again. And again. And again. Each time another five hundred cubic miles of water came roaring down, in cycles estimated to have averaged fifty-five years. Geologists John Elliot Allen and Sam C. Sargent have calculated that the combined energy of these forty to one hundred floods may have been double that of the asteroid impact of 65 million years ago suspected of wiping out the dinosaurs.

For more than three thousand years this cyclic flooding went on. Finally, 12,800 years ago, the great ice sheet began to retreat for the last time. First Lake Missoula failed to reform. Then Glacial Lake Columbia broke through the Okanogan ice lobe and the Columbia returned to its present-day channel, cutting back down through hundreds of feet of glacial till to its present bed. Dry Falls was silenced. Grand Coulee went dry. The weather warmed and got drier. Humans appeared about two thousand years after the last flood and crept into the Marmes Rock Shelter. As the glaciers receded the Columbia slowly cleared itself of glacial silt, washing its gravel beds clean and producing ideal habitat for spawning salmon. By four thousand years ago the runs were probably reaching the huge numbers that awed the explorers and pioneers.

What happened when this water and debris reached the sea? It was used to build the estuary and beaches and spits on either side of the river mouth. Gravel from Oregon’s Blue Mountains has been found in sediments near Astoria. The Columbia’s power also formed an underwater canyon that descends from the six-hundred-foot depth of the continental shelf to the abyssal plain two miles below. It is here that the remnants of mountains in Canada, Montana, Wyoming, and Idaho finally come.

A cubic foot of water with sediment weighs more than a cubic foot of clear water; so it tends to sink along an underwater slope, causing what oceanographers call a turbidity current. The pattern has been for Columbia River sediment to slowly migrate down the Pacific shelf, accumulate, and then be jarred loose by an earthquake every five or six centuries. As it slides into the abyss the turbidity current can reach speeds of sixty miles per hour. One underwater avalanche stretched four hundred miles across the ocean bottom, transporting 600 million cubic yards of sediment. These flows have filled the trench off the Pacific Northwest coast where the oceanic plate dives under the continent.

Is the story finished there? No. Eventually this dirt and sand will form sedimentary rock such as shale and be transported by the expanding sea floor back toward the land. There it will either be scraped off to form new coastal mountains or sucked under the floating continent to melt, mix, and perhaps rise again through new volcanoes. The cycle of the Columbia’s water is matched by the cycle of its land, ever-eroding, ever-rebuilding, on a scale and time span that mocks everything we have done to the river