Ice Age FloodsIce Age Floods National Geologic Trail
Home·Erratics·Ice-Rafted Boulder Site 390 — 240 ft, Sherwood / Parrett Mountain area
Ice-rafted erratic · Willamette Valley

Ice-Rafted Boulder Site 390 — 240 ft, Sherwood / Parrett Mountain area

This ice-rafted boulder is physical evidence of the cataclysmic Missoula floods that swept through this valley roughly 15,000 years ago. The specific rock type is unrecorded. Many erratics in the valley are so deeply weathered that their original minerals are masked beneath a...

Lithology
Unknown
Elevation
240 ftabove sea level
Cataloged
Allison (Fig)
Coordinates
45.3361, -122.8122WGS84

This ice-rafted boulder is physical evidence of the cataclysmic Missoula floods that swept through this valley roughly 15,000 years ago. The specific rock type is unrecorded. Many erratics in the valley are so deeply weathered that their original minerals are masked beneath a rind of decomposition. At 240 feet elevation, this boulder was deposited as the great temporary Lake Allison filled the Willamette Valley to considerable depth, stranding icebergs and their rocky cargo across the landscape. Documented by Ira Allison, whose groundbreaking 1930s research first proved that massive floods from glacial Lake Missoula had inundated the Willamette Valley.

About this rock: Mixed lithology / Indeterminate

Primary source: Not a discrete source. This category covers (a) erratics whose lithology was not confidently identified in field surveys, (b) composite boulders with multiple rock types in a single hand sample (such as gneiss with granitic intrusions, or argillite with mafic dikes), and (c) erratics whose surface weathering has obscured the original rock type.

Maximum flood elevation marker: Distributed across the entire 0–400 foot Willamette Valley erratic envelope.

A boulder classified as "mixed" or "indeterminate" is doing legitimate scientific work even when its rock type cannot be pinned down. What its iceberg carried is telling us something either way. If the boulder is genuinely composite — gneiss veined by granite, or argillite cross-cut by a mafic dike — then the source area must include a contact zone where two rock types are in physical association. That narrows the provenance considerably: the iceberg can be traced not to "the Belt Supergroup in general" but to a specific structural setting within the Belt. Composite boulders are common around contacts between the Idaho Batholith and the Belt sedimentary cover (where granite intrudes argillite), around the Moyie Sills (gabbro-in-quartzite contacts), and around metamorphic core complex margins.

If the boulder is simply unidentifiable — too weathered, too fine-grained for field identification, or recorded ambiguously in older notes — then it joins the catalogue as a location and elevation data point even without a confirmed lithology. The Lake Allison high-water mark at 400 feet is reconstructed from all 400 erratics in the database, not just the ones with confirmed source-region identifications. A boulder of uncertain lithology in a Willamette Valley pasture at 380 feet elevation still tells us that floodwater reached at least that high at that location, regardless of where the rock originally came from.

Field practice for Piper's 1928–1929 USGS work was lithology-by-eye, with occasional sample collection. Many cards in the original survey notes record "rock" or "boulder" with no further detail. Modern revisits, where they have been done, sometimes upgrade the classification of these records to specific lithologies, but the bulk of the indeterminate inventory has never been re-examined. This is one of the larger gaps in the modern erratic database and a candidate for citizen-science crowdsourcing through projects like the Ice Age Floods Institute's documentation efforts.

Notable examples: Not applicable as a category.

Modern science: Future provenance work has the most leverage here, because re-classifying a single indeterminate erratic into a defined lithology category sharpens the provenance picture for the whole valley. The same techniques applicable to single-lithology erratics — detrital-zircon U-Pb geochronology, whole-rock geochemistry, paleomagnetic remanence on intact mafic material — apply to indeterminate erratics with the additional value of producing new information rather than confirming existing classifications.

Sources:

The Willamette Valley setting

Geographic scope

The Willamette Valley is a roughly 150-mile-long, 30-mile-wide structural lowland in northwest Oregon, bounded by the Coast Range to the west and the Cascade foothills to the east. For the purposes of this overview the region includes the entire valley floor from the Portland Basin south through Newberg, the Tualatin Valley, McMinnville, Salem, Albany, Corvallis, and Eugene, plus the foothill margins where flood deposits and ice-rafted erratics are found. Named features within scope include the Erratic Rock State Natural Site (Bellevue Erratic) west of McMinnville, Mount Pisgah and the Willamette Floodplain National Natural Landmark south of Eugene, Mount Sylvania at the southern edge of Portland, the West Hills / Tualatin Mountains, and the approximately 400 documented ice-rafted erratics scattered between roughly 50 ft and 400 ft above sea level across the valley floor and lower flanks.

The Ice Age Floods story here

The Willamette Valley is the southern end of the line for the Missoula Floods. After racing west down the Columbia River corridor from the failure of the glacial Lake Missoula ice dam in northern Idaho, floodwaters reached a constriction at the Kalama Gap near present-day Kalama, Washington — the narrowest point in the lower Columbia. Discharge volumes routinely exceeded what that constriction could pass. Water backed up, ponding behind the gap to depths of 120 to 150 meters in the Portland Basin and back-flooding roughly 200 km south into the Willamette Valley (Minervini, O'Connor, and Wells, USGS OFR 03-408, 2003).

The resulting temporary lake, named Lake Allison for Oregon State University geologist Ira S. Allison, repeatedly filled the valley to a maximum elevation of approximately 400 ft (120 m) above sea level — as far south as Eugene. At its full pool the lake was approximately 111 miles long, averaged 31 miles wide, covered an estimated 3,000 square miles, and reached maximum depths near 400 ft over the lowest parts of the valley floor. Unlike the violent erosion seen upstream in the Channeled Scabland, Lake Allison was a slackwater pond: floodwaters arrived turbulent, then slowed, suspended sediment dropped out, and the lake quietly drained back through the Columbia Gorge over the following weeks. Estimates of duration per event are on the order of one to several weeks before drawdown was complete.

The floods were not a single event. Rhythmite sequences in the Sanpoil Valley of eastern Washington record up to 89 distinct flood events; Waitt's classic work documented about 40 last-glacial jökulhlaups through southern Washington, separated by intervals of decades to about 50 years. In the Willamette Valley the stratigraphic record is more muted because each successive flood reworked or buried the prior deposit, but multiple flood couplets are preserved where stratigraphy is best exposed. The cosmogenic 10Be chronology of Balbas, Barth, and Clark et al. (Geology, 2017) anchors one of the largest floods at 18.2 +/- 1.5 ka, with the main sequence of later Missoula floods running from about 15.6 ka to a final event near 14.7 +/- 1.2 ka.

The cumulative depositional signature is enormous. As Lake Allison drained after each event, the suspended fines settled out as the Willamette Silt — a blanket of fine sand and silt 180 to 200 ft thick across the lower valleys, deeply weathered into the rich agricultural soils that underpin modern Willamette Valley viticulture and farming. Allison correlated these soils with Palouse-derived sediments scoured from eastern Washington, confirming the floods as the depositional source.

The most distinctive feature of the Willamette record, however, is the ice-rafted erratics. As icebergs calved from the disintegrating Purcell Trench Lobe and from blocks of the failed ice dam, they were swept downstream entrained in the flood. Reaching the slackwater of Lake Allison, the bergs floated until they grounded or melted, dropping their cargo of cobbles and boulders. The result is a scattered population of exotic rocks — argillites, quartzites, and granitic boulders that have no business being in western Oregon — sitting on hillslopes and valley floors hundreds of miles from any bedrock source. The maximum elevation of documented erratics — roughly 400 ft above sea level — is the most direct physical marker of how deep Lake Allison stood. Above that contour, none; below it, hundreds. The 400-ft line is, in effect, a fossil bathtub ring.

Piper's 1928-1929 USGS field survey

The foundational regional record was assembled by Arthur M. Piper, a USGS hydrogeologist who undertook systematic field mapping of the Willamette Basin in August 1929 (with related Oregon work running 1928-1929). Piper's primary mandate was groundwater and dam-site geology — his Willamette Valley work fed eventually into USGS Water-Supply Paper 890 — but in the course of mapping he logged the locations, elevations, and lithologies of every exotic boulder he encountered on the valley floor and foothills. He photographed many of them in place; the surviving plates are archived in the USGS Denver Library Photographic Collection (Piper, A.M. Collection).

What makes Piper's catalog exceptional is that he worked at the right moment. By the late 1920s most of the valley floor had been cleared and farmed, exposing erratics in plowed fields where they would have been invisible under native prairie a generation earlier; but the boulders had not yet been systematically removed by mid-century agricultural mechanization, urban development, and quarrying. Piper's notebooks and field maps capture roughly 200 erratics in their original positions, many of which have since been moved, buried, or destroyed. Ira S. Allison expanded the catalog in the 1930s, publishing "Glacial erratics in Willamette Valley" (GSA Bulletin, 1935). The combined Piper-Allison dataset forms the core of the modern compilation: J.M. Minervini, J.E. O'Connor, and R.E. Wells (USGS Open-File Report 03-408, 2003) digitized and integrated their notes with later additions to publish the definitive database of approximately 400 ice-rafted erratics for the Willamette Valley and Portland Basin, paired with maps showing inundation depths and sedimentary facies. No other Missoula Floods region has a comparably dense, century-long inventory of individually located boulders.

Modern science

Recent work has refined the chronology, provenance, and depositional model. Balbas et al. (2017) used cosmogenic 10Be exposure dating on flood-transported boulders to bracket the timing of the largest flood at 18.2 +/- 1.5 ka — older than the 15-13 ka window that earlier rhythmite work had defined — and to date the final Missoula flood at 14.7 +/- 1.2 ka, tying flood chronology directly to retreat of the Cordilleran Ice Sheet's Purcell Trench and Okanogan lobes.

On provenance, Bjornstad (2014, E&G Quaternary Science Journal) documented more than 2,100 erratics across a 60-square-kilometer area at Rattlesnake Mountain, Washington, establishing the lithologic signature for the regional erratic population: roughly three-quarters granitic, with the remainder a mix of Proterozoic quartzite, argillite, gneiss, diorite, schist, and gabbro. The argillites and quartzites are diagnostic of the Mesoproterozoic Belt-Purcell Supergroup of northern Idaho, northwestern Montana, and southeastern British Columbia — the same bedrock that lay beneath the Purcell Trench Lobe and the Lake Missoula ice dam. The lithology mix in Willamette Valley erratics matches the Rattlesnake Mountain population closely, confirming that Willamette boulders are the same population as Channeled Scabland boulders, simply transported the full 600+ km downstream as the bergs survived the journey through the Columbia Gorge.

J.E. O'Connor of the USGS has been central to the synthesis. O'Connor co-authored the 2003 OFR with Minervini and Wells and has continued to publish on flood hydraulics, deposit chronology, and the Willamette Valley record, most recently as co-author on O'Connor, Baker, Waitt, et al. (Earth-Science Reviews, 2020), "The Missoula and Bonneville floods — A review of ice-age megafloods in the Columbia River basin," which represents the current consensus synthesis. Earlier USGS Professional Paper 1620 (O'Connor, Sarna-Wojcicki, et al.) established the origin, extent, and thickness of Quaternary geologic units in the Willamette Valley including the Willamette Silt.

Visiting today

The single most visited erratic in the region is the Bellevue Erratic at Erratic Rock State Natural Site, on Oberson Road off Oregon 18 between Sheridan and McMinnville. The boulder is a roughly 40-short-ton (originally estimated near 90 tons before erosion) chunk of Belt Supergroup argillite, traceable some 400+ miles upstream to its source in northern Idaho or Montana. It sits at approximately 305 ft above sea level near the upper limit of Lake Allison inundation, on a 250-ft hilltop overlooking Yamhill Valley vineyards. The Oregon Parks and Recreation Department maintains a 0.2-mile paved interpretive trail, picnic area, and signage; the 4.4-acre site is day-use only.

In the Portland metro area, the City of Tualatin and the Tualatin Ice Age Foundation have built the most ambitious interpretive infrastructure: the Tualatin Ice Age Walking Trail and the planned 22-mile Ice Age Tonquin Trail linking Tualatin, Sherwood, and Wilsonville. The Tualatin Heritage Center grounds display labeled erratics; replicated mastodon and ground sloth footprints are imprinted in the trail pavement; and the Tualatin Public Library serves as the local interpretive center. The Lower Columbia Chapter of the Ice Age Floods Institute, which covers the Willamette Basin and lower Columbia, hosts lectures at the Tualatin Heritage Center on the third Thursday of each month except August and December (also offered via Zoom, archived on the IAFI YouTube channel), and runs at least one field trip per year. The Willamette Floodplain National Natural Landmark south of Corvallis preserves 713 acres of largely unplowed bottomland prairie on Willamette Silt — a glimpse of what the lake-bottom landscape looked like before settlement.

Sources

Capture roadmap

What this site looks like once Phase 1 lands.

Every site along the trail will receive the full Terrain360 capture treatment: ground-level 360° panoramas, drone aerial imagery, and photogrammetry-based 3D models that visitors can spin in their browser. This page reserves the slots; the imagery flows in as field capture completes.

360° panoramic

Walk the site in your browser

Ground-level 360° panorama, every step along the feature, captured by Terrain360 field crews.

Phase 1 target · June–July 2026
Drone aerial

Read the landscape from above

Drone flyovers reveal the geometry of catastrophe — ripple marks, gravel bars, and scour patterns invisible from the ground.

Phase 1 target · June–July 2026
3D photogrammetry

Spin the geology in your browser

Photogrammetry and Gaussian-splat models let visitors rotate, measure, and inspect features in detail-page WebGL viewers.

Phase 1 target · June–July 2026
Why this erratic matters

Icebergs traveled hundreds of miles

Ice-rafted erratics are the smoking gun of the Missoula Floods. Boulders this size, of this lithology, cannot have arrived in the Willamette Valley by any process other than transport inside Cordilleran icebergs floating on floodwaters deep enough to clear the Columbia Gorge constrictions. Each erratic is a depth gauge for a single flood event. The cluster of erratics across the valley records both the maximum flood elevation (~400 ft above sea level) and the approximate frequency of the largest events.