Dry Falls, Grand Coulee, Washington, United States

Channeled Scablands

About 17,000 to 13,000 years ago (last series of outburst floods)

A wide basalt cliff, with plunge pools at its foot and no river coming over the lip, under a clear sky. — Dry Falls, in the Grand Coulee. The lip is a cliff about 120 metres high and several kilometres wide, and no river runs over it now. The plunge pools are what a waterfall leaves behind. Ikiwaner, CC BY-SA 3.0. Shown unedited.

A cliff in central Washington, about as high as a forty-storey building and miles wide, has no river on top of it. Bretz said a flood did that, and it took a generation for the flood to become ordinary geology. The argument now is not whether the water came. It is how many times, and whether any one of those times was something other than an ice dam.

Why this is extraordinary

Eastern Washington is a scabland of bare basalt, dry coulees, and cataracts no river uses anymore. In the 1920s J Harlen Bretz argued that the Channeled Scablands were cut by floods of a size ordinary rivers never reach. The objection of the day was as much philosophical as geological. Landscapes, the feeling went, were supposed to be worn slowly, and a flood that could strip the Palouse loess off the basalt and leave a dry falls kilometres wide sounded like a story geology had outgrown.

The water had a source. Glacial Lake Missoula, in western Montana, was held by a lobe of the ice sheet, and when that dam failed the lake emptied west across Idaho and the Columbia plateau. J. T. Pardee's giant ripple marks on the old lake floor stand metres high and far apart. They are current dunes, not ordinary beach sand, and they are the lake's own receipt for water that left in a rush. The U.S. Geological Survey's review now counts dozens of Missoula floods in the last glaciation, likely more than a hundred, with volcanic ash, soils, and mud between them. Early fillings of the lake may have lasted on the order of a century. The last ones may have come a year or two apart. Exposure ages put this last series from about 17,000 to 13,000 years ago, and they put the cutting of upper Grand Coulee inside that window.

Lake Bonneville’s single spill is a different flood in the same review. About 18,000 years ago it came down the Snake. Its volume was larger than one Missoula flood and its peak was far smaller. It did not cut the Grand Coulee or Dry Falls. Older papers put the spill nearer 14,500 years. The two ages are both in print. They are not averaged.

Sometimes a single world flood, or one impact at the start of the Younger Dryas, is asked to do all of this work. The beds are the obstacle. Time passed between the floods, and many of them are older than 12,900 years ago. What the coulees record is a repeated ice-dam story. They are not a quiet river, and they are not one night. Bonneville is not that story either. It is one spill, down a different river, and it left the coulees to Missoula.

What this changed

The shift, in short. The weighing further down is the longer argument.

Older textbook view: Before Bretz, the coulees were read as the slow work of rivers and of ice, because a flood of this size was not a process geology allowed.
New evidence: Dry cataracts, stripped basalt, and gravel bars, then giant ripples on the floor of glacial Lake Missoula, then stacks of flood beds with ash and soils between them.
How it is dated: The landforms came first, and then radiocarbon, ash layers, and cosmogenic exposure ages on boulders, placing the last series at about 17,000 to 13,000 years ago.
What moved: Catastrophic flood became a normal geological process, and the uniformitarian veto did not survive the ripples and the beds.
What is still disputed: Not the existence of the floods. What stays open is the exact number, whether every release was a jökulhlaup under the ice or a more sudden collapse, and whether any late flood belongs to a different cause.

Why it matters

The scablands are the modern type-example of a megaflood on land. If they were one discharge, the ice-dam cycle is the wrong machine, and a single catastrophe, cosmic or otherwise, is back in the running. If they were a hundred discharges, the catastrophe reading has to shrink to a question about one bed among many. The count of the beds is what carries the weight.

A coulee is a steep-walled channel that now stands dry. A jökulhlaup is a flood from under a glacier, the usual way an ice-dammed lake is thought to empty itself. Loess is windblown silt. The Palouse hills are loess, and the floods cut through them down to the basalt. The Younger Dryas, a cold return about 12,900 to 11,700 years ago, is a different room, and it overlaps only the late end of this flood series.

Photo tour

2 photographs, unedited. On a phone, swipe the row. Choose one to look closely.

An aerial view of parallel earthen ridges covering a plain, with roads and fields and a scale bar. — Giant current ripples at Camas Prairie, Montana, on the floor of glacial Lake Missoula. The ridges are dunes built by deep, fast water, the labels and the scale are the maker's, and a kilometre bar sits in the corner.Paul H., CC BY 4.0

What is still unexplained

These sit beside the consensus account, with the same visual weight. None of them is marked solved.

Holds A hypothesis that was unpopular, and then confirmed Bretz published the flood reading from 1923 onward and met a long refusal. Pardee had pointed at Lake Missoula early, and the decisive landforms on the Montana floor, ripples about 15 metres high, were laid out in 1940. By then the channels, the bars, and the abandoned falls were already mapped. The delay was not a shortage of coulees. It was a rule that said a landscape this large must be slow, and the rule lost. Naming Bretz means naming the person who was doubted, not appointing a modern authority. The coulees are the evidence.

The survey’s count Many floods, not one Backwaters along the route keep flood beds stacked with signs that time passed between them, ash falls, soils, and, in lakes, mud. The Missoula basin itself shows many fillings and emptyings. The intervals seem to shorten, from as much as about a century toward a year or two at the end, which is how a self-dumping ice dam behaves as the ice thins. The survey says it is not yet clear that every flood drained the same way. Some may have emptied more suddenly, and "more suddenly" is still an ice-dammed lake, not a different planet. Exposure ages on boulders put the incision of upper Grand Coulee between about 17,000 and 14,000 years ago, with flooding through the coulee near 14,000. Older flood deposits, from earlier glaciations, are known, and they are too broken to count.

A poor fit to the stack of beds One catastrophe, read back onto the coulees The same landscape is sometimes folded into a single disaster, a world flood or the proposed impact at the start of the Younger Dryas. The cold interval begins near 12,900 years ago. The scabland floods of the last glaciation were already running thousands of years earlier, and they left pauses. One impact can be asked to explain one layer. It cannot cut Grand Coulee in a window that starts near 17,000 years ago and still leave ash and soil between dozens of beds. A reader can still ask whether the last flood was unusual. That question lives in the younger bed, not in the coulee as a whole.

A different machine Other famous landforms that are not this flood The Finger Lakes of New York are deep troughs the ice gouged and water later flooded, and the gorge at Niagara is the cut the river has made since the ice left, a falls that has walked upstream. Both are real, and both are Ice Age, yet they are not Missoula coulees. A tour that sets them beside Dry Falls as one megaflood is joining different tools. Pleistocene lakes in the Great Basin, such as Lake Manly in Death Valley, record a wetter climate, rain and runoff ponded in closed valleys. That is a pluvial lake, and it is not an ice-dam outburst.

Location and dating

Location: Dry Falls, in the Grand Coulee of Washington, is the pin on the map. The Channeled Scablands spread across the Columbia plateau of eastern Washington. Glacial Lake Missoula stood in western Montana, and the flood route also crossed northern Idaho and the Columbia River.
Period label: Current consensus — open to revision Dozens to more than a hundred floods fell in the last glaciation, about 17,000 to 13,000 years ago for the series dated by exposure ages. Older outbursts are known from earlier ice ages. The coulees themselves are late Pleistocene.
How this dating was done: Landforms came first, and then the stratigraphy of stacked flood beds, volcanic ash between some of them, radiocarbon where organic matter was buried, and cosmogenic exposure ages on the boulders the floods dropped. A ripple dates the current that built it only as "during the lake's last high water," unless someone takes a sample from it.
What could change it: What is still wanted is a bed-by-bed count that everyone accepts, and a test that separates a subglacial drain from a collapsing dam for the largest events. The existence of repeated floods does not wait on that distinction.

Published ranges, drawn as bars. A bar is a claim about a span, not a point.

Last glaciation’s floods

15000 BCE13000 BCE11000 BCE

Last glaciation’s floods. Exposure ages run about 17,000 to 13,000 years ago, and the incision of Grand Coulee falls inside the early part of that span.

What is actually there

Dry Falls is an abandoned cataract, its lip about 120 metres high and more than about 5 kilometres wide.
At the highest water the lip may have been only a step in a deeper flow, not a free fall for the whole flood.
Giant current ripples on the Missoula floor reach about 15 metres high.
Ash, soils, and mud sit between the flood beds.
The Finger Lakes, Niagara, and Lake Manly are Ice Age landscapes with other causes.

Open questions

Question	Status
Did every emptying go out under the ice, or did some of the dams fail more suddenly?	Open inside the ice-dam model. The survey says the jökulhlaup fits the shortening intervals, and that a more complete collapse is not ruled out for every bed.
Can one impact, or one world flood, stand in for the whole series?	Not for the coulees as a whole. The pauses and the early dates are the reason.

The current narrative

Where publication stands today. It is the best-supported reading, and it is open to revision. It is not a verdict.

The current narrative — open to revision

Repeated outbursts from an ice-dammed lake

The Channeled Scablands were cut by many floods from glacial Lake Missoula during the last glaciation, and by earlier floods in older ice ages. The dam was ice and the water was the lake. Time passed between floods.

What supports this

Coulees, cataracts, and stripped basalt of a kind ordinary rivers in this climate do not cut.
Giant ripples on the Montana floor of the lake that supplied the water.
Stacked beds with ash and soils between them, and a basin that filled and emptied many times.
Exposure ages from about 17,000 to 13,000 years ago for the last series.

What pushes back

The number is still "dozens, likely more than a hundred," not a final integer.
The survey leaves open whether every release was a jökulhlaup.
Older floods are too fragmentary to count, so "how many in the whole Pleistocene" is not known.

What would settle it

This is the centre of the room. What stays open is the count, and the style of each dam failure, not the existence of the water.

Emerging narratives

Other published readings. Each may be right. None is adopted here unless the evidence carries it.

An emerging narrative

One catastrophe did the coulees

A single megaflood, tied to a world disaster or to an impact at the Younger Dryas, cut the scablands, and the ice-dam series is an over-count of that one event.

What supports this

The landforms are outrageous for a normal river, which is why Bretz was doubted in the first place.
The Younger Dryas is a real and abrupt cold, and some flood beds are late enough to overlap its start.
Popular accounts prefer one story to a hundred that look alike.

What pushes back

Ash, soils, and mud keep the flood beds apart.
Exposure ages put the cutting of the canyons thousands of years before 12,900.
The source lake has its own fill-and-empty stratigraphy, over in Montana.

What would settle it

It does not survive the pauses. A question about one late bed can be asked in the Younger Dryas room, and it is not a theory of Dry Falls.

Should the current view be reevaluated?

Consensus holds well

Bretz's flood is no longer the heresy. Repeated Missoula outbursts are the working geology, dated for the last series and witnessed by ripples at the source. What stays open is the exact tally, and the mechanics of each dam failure. A single cosmic flood, or a single world flood, does not fit the stack.

What drives this

Abandoned cataracts, and basalt stripped bare.
Giant ripples on the floor of Lake Missoula.
Many flood beds, with time between them.
Ages that mostly fall older than the Younger Dryas.

This is the museum’s own weighing of the evidence. It is not a visitor vote, and it is separate from the slider.

Weigh it yourself

Anonymous. Nothing here is stored except on this browser. Your lean does not change the assessment beside it.

Repeated outbursts from an ice-dammed lake Unsure One catastrophe did the coulees

If the contested reading is right

What would follow if a reading that is not the consensus turned out to be true, and what could settle it. This explains the stakes. It does not adopt the reading.

Some of the largest floods were sudden dam collapses, not under-ice drains

The water would still be Lake Missoula, and the coulees would still be outburst floods. Peak discharge for those particular events would be even harder to model, and the count of floods would remain.

Why it is still worth a test

Sediment structures that only a near-instant collapse produces, found in identified beds, set beside beds that look like a slower jökulhlaup.

One late flood has a different cause

Only that one bed would leave the ice-dam series. The coulees, already cut, would not become an impact crater, and the claim would have to live in the deposits of that single flood.

Why it is still worth a test

A marker in one dated bed that is absent from the floods above it and from the floods below it.

Sources

The Missoula and Bonneville floods, a review U.S. Geological Survey
The timing of Missoula floods and the age of upper Grand Coulee Geology, preprint copy via the National Science Foundation
Introduction to the Ice Age floods Ice Age Floods Institute

Image credits

Dry Falls, in the Grand Coulee. The lip is a cliff about 120 metres high and several kilometres wide, and no river runs over it now. The plunge pools are what a waterfall leaves behind.. Ikiwaner, CC BY-SA 3.0. Shown unedited.
Giant current ripples at Camas Prairie, Montana, on the floor of glacial Lake Missoula. The ridges are dunes built by deep, fast water, the labels and the scale are the maker's, and a kilometre bar sits in the corner.. Paul H., CC BY 4.0. Shown unedited.

← United States World map