Ice Age Floods National Geologic Trail
Ice Age Floods National Geologic Trail
Balbas, Barth, Clark, Clark, Caffee, O’Connor, Baker, Konrad, Bjornstad — “¹⁰Be dating of late Pleistocene megafloods and Cordilleran Ice Sheet retreat in the northwestern United States.”
The study that re-anchored the flood chronology — placing the largest Missoula Floods at 18.2 ± 1.5 ka, three thousand years earlier than Bretz’s original framing, and resolving the sequence in which the Cordilleran Ice Sheet’s southern lobes opened, dammed, and re-routed the floods over roughly 4,000 years.
For most of the 20th century, the timing of the Missoula Floods was bracketed by indirect evidence — volcanic ash layers, lake-sediment records, and stratigraphic relationships to the Mount St Helens “S” tephra. The result was a range that read like a shrug: “sometime between 22 ka and 16 ka.” A range a third as wide as the floods themselves.
Balbas et al. 2017 did what no one had done before: directly date the flood-deposited boulders themselves. Cosmogenic ¹⁰Be (a radioactive isotope produced when cosmic rays strike rock at the surface) accumulates predictably from the moment a boulder is exposed. Sample enough boulders from enough sites — ice-rafted erratics high on the Columbia Valley walls, granitic boulders strewn across flood bars, bedrock exposed by flood erosion — and you can pin specific flood events to specific dates.
The result is the chronology this entire trail now runs on.
Massive floods route down the NW Columbia River valley before the Okanogan lobe of the Cordilleran Ice Sheet advances and blocks it. Wallula Gap erratics 365 m above sea level fix the maximum flood depth.
The ice tongue that dammed Glacial Lake Missoula in the Clark Fork Valley starts withdrawing northward.
Erosion of the granite cataract complex at Northrup Canyon clears the upper Grand Coulee, making it the primary path for floods until the end of the Missoula sequence.
Withrow Moraine boulders constrain the onset of retreat. The lobe’s withdrawal opens the upper Columbia path again.
Mean ¹⁰Be age of mega-ripples near Spirit Lake and a boulder near Lake Pend Oreille — downstream of where the floods were released — dates the last major Missoula flood.
Smaller floods continuing down the NW Columbia, likely sourced from glacial Lake Columbia. Lake Columbia persists for centuries after the last Missoula flood.
The single most important geographic story in Balbas et al. is what happened at Wallula Gap. The high-elevation ¹⁰Be ages along the NW Columbia River valley (18.2 ka) are older than the maximum extent of the Okanogan lobe (15.4 ka). That means the largest floods preceded the lobe’s southward advance — they flowed through the upper Columbia, not through the Channeled Scabland.
Once the Okanogan lobe advanced and dammed the Columbia, floods were forced south through the Scabland complex via Grand Coulee. The Scabland features — the Cheney–Palouse, Telford–Crab Creek, and Quincy basins — record floods after the largest events, when the routing flipped.
This is the inversion of the older narrative. Bretz mapped the Scabland; he assumed the Scabland recorded the floods. Balbas et al. show the Scabland records the later, smaller floods. The largest floods left their fingerprints up the Columbia Valley.
Every interpretive sign, museum exhibit, and chapter pamphlet along the Ice Age Floods National Geologic Trail that says “15,000 years ago” is using 1920s science. The 18.2 ka anchor — with the full 4,000-year chronology Balbas et al. resolved — should be the trail’s default frame.
This is one of the gaps the trail’s public-facing infrastructure can close: a coherent, accessible explanation of when the floods happened, rooted in the peer-reviewed chronology rather than the inherited shorthand.
The Okanogan lobe begins retreat at 15.4 ka. The Puget lobe was already retreating at 17.0–17.2 ka. That ~1,600-year offset between two lobes of the same ice sheet is a real puzzle — it tells us the controls on ice-margin retreat were different on the coast versus inland.
Balbas et al. attribute it to dual controls: marine-ice dynamics on the Juan de Fuca lobe versus surface mass balance on the terrestrial Okanogan and Purcell Trench lobes. But the model isn’t airtight. The exact pace of the routing flip — how Grand Coulee opened, whether the upper Coulee was partially open during Okanogan-lobe maximum — remains open.
These are the questions the next decade of fieldwork along the trail will answer. The infrastructure of the trail itself — well-mapped, well-photographed, well-modeled in 3D — is what makes that fieldwork possible at scale.