Translator Disclaimer
1 January 2019 Erratics and Other Evidence of Late Wisconsin Missoula Outburst Floods in Lower Wenatchee and Adjacent Columbia Valleys, Washington
Author Affiliations +

The Pleistocene Missoula floods through eastern and central Washington are by peak fl ow rate (discharge) the greatest freshwater cataclysms known on Earth. Newly explored features along the Wenatchee reach of Columbia valley give new evidence and revise earlier interpretations of size, frequency, and routing of megafloods.

Crystalline-rock erratics derived far northeast lie scattered about the sandstone hills of lower Wenatchee valley and adjacent Columbia valley up to 495 m altitude, 320 m above Columbia River. They can only have been ice-rafted by flood(s) running down the Columbia. Before the late Wisconsin Okanogan lobe of Cordilleran ice blocked the Columbia, at least one monstrous Missoula flood poured down the valley past Wenatchee and backflooded Wenatchee valley.

Rhythmically bedded sandy silt in Columbia valley between Trinidad and Wenatchee records repeated silt-rich backfloods up the Columbia from Quincy basin—after Okanogan-lobe ice had blocked the Columbia upvalley. Rhythmically graded silt beds in Wenatchee valley at Dryden containing Columbia-derived dropstones record ten Missoula backfloods up the valley.

Thick silt farther up Wenatchee valley between Peshastin and Leavenworth had been thought deposits of a long-lived lake, dammed supposedly by the Malaga landslide. But the heights and distribution of provable lake beds now make Moses Coulee bar the only viable dam—and only up to altitude 275 m. The silt above Dryden lying at 315–385 m altitude must also have been laid by Missoula floods.


Glacial Lake Missoula floodwater flowed down the Wenatchee reach of Columbia valley deeper and faster than anywhere else in the famed Channeled Scabland. Clear evidence for this has been known informally for years and even decades. Here we present some of the unpublished earlier work and new details of a complicated series of Missoula floods through this valley reach.

In the 1960s William A. (“Bill”) Long had located many crystalline-rock cobbles and boulders scattered about lower Wenatchee and adjacent Columbia valleys. These erratics seemed explicable only by ice-rafting on a large flood. Large-area geologic mapping (Waitt 1982, 1987) elaborated this interpretation. In the late 1980s Long resumed his hunt for erratics and plotted their locations on topographic maps.

The erratics fall into two groups. One comprises diverse bright granite, gneiss, quartzite, and other cobbles and boulders apparently derived far up Columbia valley (subject of this report). The other group—darker, coarser-grained very large tonalite and granodiorite boulders from the Mount Stuart batholith in mountains to the southwest—lie on valley fl oor and sides farther upvalley, between Leavenworth and Peshastin (subject of Stanton et al. 2019).

Figure 1.

Map showing regional distribution of Cordilleran ice sheet at maximum stand (pale blue), glacial lakes (green), Missoula-flood erosional and depositional areas (brown), outlet of Missoula floods (purple arrow). Glacial Lake Columbia (simplified) depicted at maximum possible height. LPO, Lake Pend Oreille. (Modified from Waitt 2016.)


This report partly on the Columbia-derived erratics stems from Long's work, from recent fieldwork, and from Waitt's 1970s field notes that had been only partly explored for geologic-map reports (Tabor et al. 1982, 1987). Locations and descriptions of the erratics derive mostly from an unpublished 1990 Bill Long manuscript.

Possible Sources of Megaflood

Late Wisconsin megafloods past Wenatchee have for more than four decades been inferred to have burst from glacial Lake Missoula—maximum volume about 2500 km3and depth as much as 640 m. But what of other possible sources of Columbia River megaflood?

As the Okanogan and Columbia River lobes of Cordilleran ice (Figure 1) advanced and retreated, they dammed many successive lakes in tributary valleys. Deposits recording such lakes during the recession lie abundantly about the Methow, Okanogan, and upper Columbia valleys at low to high levels (Fulton 1969; Waitt 1972, 1984, 2016; Waitt and Thorson 1983). When earlier these ice lobes had advanced from British Columbia south, they must also have dammed surface lakes in tributaries—relatively large ones in a few northwest-trending valleys in southern British Columbia. Yet the geographic settings do not allow a truly gigantic lake. A hypothetical subglacial lake as large as 150 km3in Okanogan valley (Lesemann and Brennand 2009) also would be much too small and shallow to release a flood that could reach the high-level evidence (such as erratics) near Wenatchee. Both 1-D and 2-D hydraulic models peg peak discharge of flood down the Columbia to above 10 million m3s-1(see below). The only known Pleistocene lake with enough volume and depth to have released such cataclysms was glacial Lake Missoula.

Figure 2.

Index map of central Washington showing places named in text. Dotted red line is maximum extent of Cordilleran icesheet, simplified and slightly modified from Atwater (1986, plate 1), who adapted it from Waitt and Thorson (1983, Figure 3-1). Dotted boxes are areas of maps for Figures 4, 5, 15, and 17. Wen, Wenatchee; RI, Rock Island; SG, Sentinal Gap.


Glacial Limits and Missoula Floods near Wenatchee

The Purcell Trench lobe of the late Wisconsin Cordilleran ice sheet dammed Clark Fork River to impound glacial Lake Missoula (Figure 1) (Bretz 1959, Baker 1973, Waitt and Thorson 1983). Periodic outbursts from the lake—as much as 2500 km3—swept south down Grand Coulee and the Channeled Scabland farther east but at times also down the northwest “big bend” of Columbia valley (Figures 1, 2) (Waitt 1977b, 1982, 1985, 1987, 1994, 2016, 2017; Waitt et al. 2009). Any giant flood pouring down or backflooding up this reach of the Columbia also backfloods lower Wenatchee valley.

Early studies concluded that the Columbia had been glaciated to below Wenatchee. By fields of huge boulders both sides of the river near and below Wenatchee, Landes (1911) inferred “a great glacier . . . down the Columbia as far as . . . Moses Coulee.” Bretz (1930) interpreted undulating bouldery deposits in Columbia valley near Wenatchee and Moses Coulee as glacial moraines. The huge low-level boulders, a high-level granitic erratic, and what he thought was ice-striated and polished bedrock led Chappell (1936) to infer that Wenatchee had been glaciated. A state geologic map (Huntting et al. 1961) shows most of Wenatchee area as glacial drift.

But Waters (1933) had mapped the limit of the Okanogan lobe of late Wisconsin Cordilleran ice in Columbia valley near Chelan Falls 30 km above Wenatchee, and Flint (1935, 1937) agreed. In Wenatchee valley, Page (1939) delineated the late Wisconsin mountain-glacier limit near Leavenworth 20 km above Wenatchee.

J Harlen Bretz's theory of gigantic flood through eastern Washington and the lower Columbia, developed in eleven evidence-rich papers 1923–1932, had been controversial because he had no plausible water source. After Pardee (1942) identified Pleistocene glacial Lake Missoula as a viable source, Bretz et al. (1956) with new evidence and rationale proved great flood through the Scablands. Trimble (1963) found that only megaflood could explain great bars in the Portland basin.

It was US Forest Service geologist Bill Long who began to unravel a megaflood history of lower Wenatchee valley. Field photographs, unpublished manuscripts, and observations presented on a field conference (Porter 1969) show that by the late 1960s Long had located erratic stones at dozens of sites in lower Wenatchee and adjoining Columbia valleys. Surficial-geology mapping in 1976– 1977, aided by vertical aerial photographs and detailed topographic maps—revealed gigantic Pangborn bar in Columbia valley east of the river (Figure 3) and the megaflood origin of other bars near Wenatchee and downvalley (Waitt 1977a, 1982). Giant current dunes with downvalley-facing lee slopes embellish the tops of Pangborn and other bars. Gravel pits expose downvalley-dipping foreset beds in mixed-lithology bouldery gravel. Clearly giant downvalley-running floods had built these bars. Hundreds of boulders up to 12 m diameter both sides of the river at Wenatchee (most of them now gone by urban development) derive from bedrock sources 7–38 km upvalley—well below the glacial limit near Chelan Falls. The boulders thus came not as direct glacial deposits but as flood bedload. A westside bouldery bar 4 km long and 60 m high (Wenatchee bar) built across lowermost Wenatchee valley (Figures 4, 5) contains gneiss boulders up to 7.5 m. Its foreset beds dip gently west out of Columbia valley and up the Wenatchee (Figure 6). In the lower 37 m of Rock Island bar, upvalley-dipping foresets in basaltic sandy gravel record backflood up Columbia valley from Moses Coulee (Figure 7). Each of five such beds there capped by fine beds, including varves, record breaks as long as a few decades between Moses Coulee floods. Overlying this interval are 14 m of bedded sandy silt. Lacking finely laminated clayey beds characteristic of standing lakes, this deposit seems from backfloods coming from much farther downvalley (Waitt 1982).

Figure 3.

Oblique aerial photograph northward of Pangborn bar. Snow highlights giant current dunes on its top, as between the airfield runways. Their downvalley-facing steep sides show this bar formed beneath downvalley-running flood(s). The lower Malaga bar (M) topped by smaller giant dunes records a smaller downvalley flood. City of Wenatchee at left west of Columbia River. (Photo by R. B. Waitt, Feb 1972).


And so stratigraphy and geomorphology in Columbia valley near Wenatchee have revealed three episodes and patterns of late Wisconsin megaflooding from glacial Lake Missoula (Figure 8) (Waitt 1994, 2016, 2017):

  1. Before the Okanogan glacier lobe dams the northwest reach of the Columbia, flood rages down Columbia valley past Wenatchee as deep as 320 m and backfloods Wenatchee valley.

  2. Advancing Okanogan ice blocks the Columbia. At least five Missoula floods down Moses Coulee build a great bar (Moses Coulee bar) that dams the Columbia and backfloods upvalley past Wenatchee.

  3. Further advance of Okanogan ice blocks Missoula floods from the head of Moses Coulee. Many megafloods via Grand Coulee to Quincy basin backflood up Columbia valley.

Farther east the immense Missoula floods carved small valleys into gigantic scabland channels, crossed divides, and built huge gravel bars— famed features of the Channeled Scabland (Bretz et al. 1956, Bretz 1959, Baker 1973). Some 90 great Missoula floods poured out beneath the last-glacial ice dam (Waitt 1980, 1985; Atwater 1986; Waitt et al. 2009, 2016). Most of the water poured to the Scabland via Rathdrum valley (Figure 1), where peak discharge computes as high as 25 million m3sec-1(O'Connor and Baker 1992, Denlinger and O'Connell 2010). These are Earth's largest known freshwater discharges.

Figure 4.

Map spotting Cordilleran-ice erratics and other Missoula-flood features in Columbia and lower Wenatchee valleys as high as 495 m. Map largely traced from a Bill Long illustration of about 1990. Solid circles are Long's sites as he numbered them in the late 1980s. Solid squares with number prefix W are Waitt sites of Sept. 1976. Many of Long's numbered sites embrace two or more specific sites as much as 400 m apart. Table 1 lists only some. Sites plotted on topographic maps have been transferred to this small-scale summary without precise georeferencing. Waitt's W1, W3, and possibly W5 are the same stones that Long had located as L5, one of L6, and L2. Offsets between the W and L sites owe to errors in the analog transfers. OG, Ohme Gardens.


A regional upsection thinning and fining of floodlaid graded beds in stratigraphic sections across the region show the later Missoula floods became smaller (Waitt 1985, Atwater 1986). As the regulating Purcell Trench ice dam thinned during deglaciation, shallower lake levels would destabilize it. So later floods came smaller and more often—dozens of them too small to reach high and distant sites like lower Wenatchee valley.

Ice-Rafted Erratics near Wenatchee Carried by Missoula Floods

Unweathered crystalline-rock and metasedimentary-rock erratics scattered about Columbia and lower Wenatchee valleys as high as altitude 495 m seem inexplicable unless floated in icebergs that grounded near the edges of deeply ponded water. Evidence of deep flooding disappears upvalley at the glacial limit near Chelan Falls. So the largest last-glacial flood(s) down the Columbia predate the Okanogan lobe blocking the valley (Atwater 1987, Waitt 2016).

Bill Long had measured altitudes of erratics by a Taylor Pocket Altimeter model SMT-5-51 and later plotted most sites on US Geological Survey (USGS) 1:24,000 topographic maps (Table 1). Dan Long, Waitt, and Stanton have found erratics at several sites where plotted on Bill Long's maps. But such boulders in the way of orchards and building developments have been removed. Table 1 lists the best sites and largest stones—those Long had assigned numbers to (his maps spot erratics at thrice his numbered sites.) We add six sites from Waitt's 1976 fieldwork. Several of Long's numbered erratics between Peshastin and Dryden (Figure 5) remain ambiguous. His unpublished 1990 table suggests some of them are Columbia origin (subject of this report). But his map notations indicate that some of them came from Mount Stuart batholith via Icicle Creek glacier (Stanton et al. 2019). Those sites needing further investigation are not vital to this report or to Stanton et al. (2019), and we exclude them here.

Figure 5.

Map of lower Wenatchee valley spotting erratics derived from Cordilleran ice. Bill Long map-plotted many more sites than those in Table 1. Several numbered sites comprise two or more nearby erratics or clusters. Omitted from this map and Table 1 are 13 sites between Leavenworth and Peshastin of large boulders derived from Mount Stuart batholith addressed in a companion report (Stanton et al. 2019). Numbers not shown for 11 sites of erratics between Peshastin and Dryden that need reinvestigation and do not influence either report. OG, Ohme Gardens. Silt bodies simplified from map of Tabor et al. (1987).


Figure 6.

South-southwest-directed photograph showing railroad cut into upstream part of large floodbar built across mouth of Wenatchee valley. Faintly discernable foreset beds in upper third of cut dip gently west—out of Columbia valley, up Wenatchee valley (to right). (Photo by W. A. Long in 1960s.)



Erratics in Columbia and lower Wenatchee valleys. Preceding each site number, the L = W. A. Long stations and W = R. B. Waitt stations. Omitted numbers (e.g., L13, L31–37, L42–43) are sites superceded by later data or needing further fieldwork. Coordinates are rendered by GIS registration of points marked on paper maps. Altitude is interpolated from USGS 1:24,000 topographic-map contours. W. A. Long's altimeter-derived altitudes are within a few to several meters. Rock Type parenthetical notes from older data table. Diameter values recorded by W. A. Long were apparently routinely the largest diameter axis; Waitt routinely records intermediate diameter axis—designated here (ID).


Figure 7.

Measured section of Rock Island bar showing stratigraphy of sequence of floodbeds in Columbia valley: five basalt-gravel beds shed upvalley from Moses Coulee, overlain by bedded sandy silt inferred from Quincy basin backfloods, overlain in turn by duned mixed-lithology gravel from a downvalleyward late flood.


The erratics at diverse levels (Table 1) range from angular to round, a few glacially striated, a few barely modified joint blocks. Clasts of leucocratic feldspar-megacrystic granite, granite pegmatite, mesocratic granodiorite, dark-gray meta-argillite, and quartzite came from bedrock lithologies that do not crop out in lower Wenatchee valley. Many megacrystic-granitic boulders near Wenatchee (Figure 9) resemble bedrock along the shores of Lake Pend Oreille in Idaho where the Purcell Trench lobe had stood (Waitt 1977a, 1982).

At the south end of Jumpoff Ridge south of Malaga, a cluster of dark-gray meta-argillite, quartzite, and various granitics—mostly cobble sized—lies at altitude 470 m, or 295 m above the Columbia River (Figure 4, site 1). They rest on Columbia River Basalt bedrock that also underlies all areas upslope.

A huge fresh-looking landslide at Malaga heads along a 25-km-long curving scarp in basalt high on Jumpoff Ridge (Figures 4, 10) (Tabor et al. 1982). It is the most recent of many large landslides from basalt cliffs rimming both sides of Columbia valley that descend to the river. Unweathered granitic boulders lie on Malaga landslide up to at least 230 m above the river. For instance a glacially striated 2 m granitic boulder lies at altitude 378 m just east of Stemilt Creek (Figure 4, site 4). Farther east, unweathered granitic boulders up to 1.5 m lie up to at least altitude 427 m (Figure 4, sites 2, 3, W5, W6). So the landslide predates at least one erratic-rafting megaflood.

Across the Columbia above huge Pangborn bar, a large light-colored granitic boulder lies at altitude 460 m atop landslide deposits (Figure 4, sites 5, W1; Figure 9). Many other erratics overlie this east-side landslide debris in the 450–495 m altitude range (Figure 4, sites 6–12, W2–4).

In lower Wenatchee valley in a gully north of Ohme Gardens, a cluster of erratics at altitude 397 m comprises mostly light-colored granitic clasts but includes a fresh 1.5 m boulder of Columbia River basalt (Figure 4, site 14; Figure 11E). No bedrock basalt crops out in Wenatchee valley north of Wenatchee River, so the basalt boulder too is erratic. Another fresh erratic basalt boulder lies at altitude 340 m west of the mouth of Warner Canyon (unnumbered site 200 m northeast of site 29, Figure 5). Basalt boulders—abundant in drift of the Okanogan lobe but rare in that of the Purcell Trench lobe—reveal Missoula megaflood(s) down mainstem Columbia valley that may also have quarried ice from the Okanogan lobe.

Figure 8.

Summay seque tial routi gs of th ee Misso la-floods enarios rough nor hwest C lumbia va ley and arge-c ule drainageway (scabland t acts) dur ng ea ly to mi dl perio of la e isco sin megaflood-i gs. In ca ib ated radio arbon time s in er reted by Wait (201 , thi sequen e ru s from a out 8.9 a (A ) o ab ut 18.5 a (B t ab ut 18.0 a (C . ed l nes depic infer ed posit on of th O ano an lobe nd C lum ia River lobe ordi leran-ice termi i at the ffer t t mes. Ligh blue rrows show major nc omi ng route of Mi so la flood . Green arrow depic backflood t war Wenat hee.


Scores of megacrystic granite, granodiorite, and quartzite clasts about the north side of Wenatchee valley between Cashmere and the Columbia show that floodwater rafted icebergs upvalley (Table 1) (Figures 5, 11 A–D). A kilometer northeast of Monitor, granitic stones on a 470 m ridgecrest (unnumbered site about 28 m above site 26 on Figure 5) show that floodwater overran the spur. Many granitic boulders as large as 2.5 m lie near Cashmere (Figure 5, sites 27–30).

The clearly ice-rafted erratic granite, granodio-rite, gneiss, meta-argillite, quartzite, and other exotic boulders in Columbia valley near Wenatchee, like similar ones in Quincy basin and Pasco basin, are inferred to derive from the Purcell Trench ice lobe (Bretz 1930; Bretz et al. 1956; Waitt 1980, 1982, 2016; Bjornstad 2014). Bedrock similar to these erratics indeed crops out along Lake Pend Oreille.

Figure 9.

Photograph toward south-southwest across Columbia valley from basaltic landslide debris above Rock Island (town). The large unweathered angular megacrystic K-feldspar granite or granodiorite boulder has been ice rafted here to altitude about 457 m, some 281 m above Columbia River (site W1). A Bill Long photograph reveals this boulder as site L5. Partly by this very boulder, Chappell (1936, Fig. 77) had inferred the Wenatchee area glaciated. (Photo by R. B. Waitt, Sep 1976.)


It is tempting to infer the rare basalt erratics in lower Wenatchee valley to have rafted from the Okanogan ice lobe, which was rich with them. But Columbia River basalt also crops out in small areas just west of the southwest end of Lake Pend Oreille (Griggs, 1973, 1976)—in the terminal-most reach of the Purcell Trench ice lobe at its maximum late Wisconsin stand (Waitt et al. 2016). The basalt erratics might well have come from there.

Bedded Sandy-Silt Deposits in Columbia Valley

In valleys in southern Washington, ash beds, loess layers, and anciently filled rodent burrows within rhythmically graded sand-silt beds tell of lengthy nonflood periods between the graded beds laid by successive Missoula backfloods (Waitt 1980, 1985). Missoula floods also ran into glacial Priest Lake and glacial Lake Columbia in northern Washington, where varved lake beds separate the episodic flood beds. These definitive northern stratigraphic sections prove that each megaflood deposited only one major graded bed (Atwater 1984, 1986, 1987; Waitt 1984). Repeated major graded beds at any Missoula-backflood site imply repeated megafloods.

Rock Island bar in Columbia valley 20 km below Wenatchee consists largely of five basalt-gravel foreset beds shed up the Columbia from Moses Coulee (Figure 7). Sand and silt beds that overlie each gravel bed include varves—37 varves atop one of the gravel beds. Moses Coulee bar dammed the Columbia for several decades at least.

Above these foreset beds lie bedded floodlaid sandy silt at altitudes about 214–228 m. About 19 graded granule-to-silt beds overlie a huge eddy bar at Trinidad (mouth of Lynch Coulee). Bedded sandy silt floors Moses Coulee up to altitude 280 m, overlies the north end of Moses Coulee bar, and overlie the south end of Pangborn bar up to altitude about 323 m (Figure 12) (Waitt 1994, 2016). These deposits lack laminated clayey layers or varves that would signify a long-standing lake. They seem from silt-laden Missoula backfloods running upvalley from Quincy basin.

Erratic-Bearing Rhythmites at Dryden

About ten graded sand-to-silt beds 60 to 75 cm thick (Figures 6, 13) lie at altitude 310–317 m at Dryden in lower Wenatchee valley. They display cross laminations directed generally west, upvalley. Capped by only weak brown soil, these beds are late Wisconsin age. They overlie a thick mature reddish soil that caps gravel 20 m thick containing weathered stones of Mount Stuart tonalite-granodiorite. This older gravel seems outwash from a pre-Wisconsin glacier upvalley (Stanton et al. 2019).

Figure 10.

Cross-section of Columbia valley showing landslides overlain by Pangborn bar and other late Wisconsin megaflood deposits (from Waitt 2016).


Figure 11.

Photographs showing high-level erratics (see Figures 4, 5, and Table 1 for locations): A) light-colored granite, 2 m, site 26, 442 m altitude (by W. A. Long); B) quartzite, 1 m, site 25, 436 m altitude (by W. A. Long); C) striated granitic erratic (by W. A. Long); D) light-colored megacrystic granite, 1 m, Warner Canyon site 27, altitude 430 m (by W. A. Long); E) granite (1 m) and basalt (2 m), northwest of Ohme Gardens, site 14, altitude 397 m (by R. B. Waitt, May 2017).


In each bed of the Dryden rhythmites, a basal coarse to medium sand grades up through fine sand to silt. No delicate clayey laminae lie between the graded beds, nor varves, to suggest these floods had engorged a standing lake. Within the rhythmites lie rare angular dropstones as large as 40 cm (mostly much smaller) of light-colored granitic rocks like the erratics in lower Wenatchee valley (Figure 14). The Dryden rhythmites record at least ten separate backfloodings up lower Wenatchee valley by Lake Missoula megafloods. These silty rhythmites appear to be from the floods that deposited the layered sandy-silt in Columbia valley. The Dryden beds lie 30–95 m higher than the fine beds in Rock Island bar and in lower Moses Coulee but are at similar level to those overlying part of Pangborn bar (Figure 12).

Other Fine Valley-Floor Deposits

Sandy silt overlies gentle valley-floor surfaces in Wenatchee valley from Dryden up to glacial moraines near Leavenworth (Figure 5). This widespread but poorly exposed material appears to be an upvalley part of the beds at Dryden. The silt generally rises upvalley from altitude about 300 m west of Cashmere to 385 m north of Leavenworth (Figure 12). It thickly veneers older valley-floor deposits and rises upvalley with pre-existing topography. This resembles the upvalley rise of much thicker fine sand and silt (“slackwater deposits”) in deeply backflooded valleys in southern Washington like the Walla Walla, lower Yakima, and lower Tucannon (Figure 1) (Waitt 1980, 1985).

The silt concentrates on gentle slopes where invading silty water had ponded deepest. Once flood currents go fairly slack, fine sand and silt settle fairly fast. Such deposits would be naturally thickest where the silty water column had been deepest over the valley center. Silt must also have once veneered the valley sides. But these thinner deposits would readily wash off steeper slopes or mix with colluvium, and with time disappear.

Figure 12.

Longitudinal profile of Columbia valley from Rocky Reach down to Potholes cataract (at Quincy basin). Flood bars projected from both valley sides to centerline. Natural grade of river approximated by line drawn through tailraces of dams. Several late Wisconsin water levels shown: upper limit of erratics by downvalleyward megaflood; level of lake ponded by Moses Coulee bar; maximum level of upvalley backfloods via Quincy basin—as inferred by Waitt (2016), and as inferred in this report. Solid vertical bars plot altitude ranges of bedded silt apparently deposited by backflood: T, Trinidad (mouth of Lynch Coulee); MC, floor of lower Moses Coulee; MCB, upvalley part of Moses Coulee bar; RIB, Rock Island bar; PB, downvalley part of Pangborn bar; WV; Wenatchee valley from Cashmere (Ca) up to Leavenworth (Le), the Dryden rhythmites distinguished separately.


Figure 13.

Rhythmites at Dryden as better exposed in late 1960s than now. A) Eocene sandstone at river level overlain by 12 m of gravel capped by a reddish argillic paleosol. Atop this ancient gravel lies 15 m of rhythmically-bedded fine sand and silt capped by feeble soil; B) Detail of Dryden rhythmites. (Photos by W. A. Long.)


Numbers of Missoula Floods Recorded at Different Sites

Why only ten beds at Dryden when one counts many times that in other parts of the region? In lower Yakima and Walla Walla valleys where an ash couplet from Mount St. Helens marks a time horizon correlatable between sections, fewer beds overlie the ash at higher upvalley sites than at lower downvalley sites (Waitt 1980, 1985). Smaller, later Missoula floods couldn't reach higher sites. Dryden is a relatively high and distal site. Some of the smaller, later floods could inundate lower-level reaches like Pasco basin or Trinidad. Apparently only a few floods were large enough in volume and discharge to follow the valleys as far up as Dryden.

Physically Dammed Lake

Fine sand and silt in Wenatchee valley upvalley of Cashmere has been in-ferred to be lacustrine (Waitt 1977a, Whetten and Waitt 1978)—perhaps an upvalley extent of laminated deposits near Wenatchee that reveal a long-dammed lake. But later study of graded megaflood beds in southern Washington (Waitt 1980) showed rhythmically graded silty beds like those at Dryden to record repeated giant backfloods—not a long-standing lake (Waitt 1982, 1985).

Yet Waitt (1982) still entertained the idea of a lake in Wenatchee valley dammed to 320 m or high-er. The supposed damming agent is giant landslide(s) below Wenatchee (Figure 10) whose rocky toes still encumber the Columbia River.

Finely laminated, varved, clay-bearing lake sediment, as displayed in a tall roadcut section in Horse Lake Canyon (Figure 4), proves a long-standing lake in Columbia valley near Wenatchee. Fieldwork in 2017–18 and review of Waitt's 1976–77 field notes and maps prove such deposits no higher than about 275 m altitude. One possible damming agent is the Malaga landslide below Wenatchee (Gresens 1983, p. 50). Another is the great bar built across the Columbia by floods out of Moses Coulee farther downriver (Figure 15)—that had dammed a lake at about 275 m. Waitt (1982, 1994) attributed large hanging deltas at the mouths of Stemilt and Squilchuck creeks to this lake. Two areas of lake deposits—northern Rock Island bar and northern Moses Coulee bar (Figure 12)—lie downvalley of the Malaga landslide. Such a lake could have been dammed only by Moses Coulee bar.

Figure 14.

Granitic dropstones from northern Columbia River embedded in Dryden rhythmites (stone on the left about 30 cm high). (Photo by W. A. Long.)


None of the known lacustrine exposures near Wenatchee can now be uniquely tied to the Malaga landslide. The several known sites of varved clayey lake beds, all below 275 m, could all be damming by Moses Coulee bar. There is no proof known in field evidence that landslide had been a damming agent.

The silt lying at 280 to 385 m on the northeast side of Wenatchee valley between Cashmere and Leavenworth (Figure 5) seems not of long-term physical ponding but of repeated brief Missoula floods. These silt bodies may be the same sediment as the Dryden rhythmites.

Modeled Megafloods

A great flood down an ice-free Columbia valley can be reconstructed by one-dimensional step-backwater routine. Earlier 1-D modeling of Columbia valley through 154 detailed cross-sections from above Grand Coulee to northwestern Pasco basin (Harpel et al. 2000, Waitt et al. 2009, p. 806) has been re-run with the current HEC-RAS program (Waitt 2017). Peak model discharge of about 13 million m3s-1best fits field evidence (highest scabland, scarps cut into loess, and erratics) for peak floodfl ow at and below Wenatchee (Figure 16).

Previous two-dimensional hydraulic modeling of Missoula floods (Denlinger and O'Connell 2010) tested a routing scenario that blocks off the northwest part of Columbia valley. New 2-D modeling reconstructs Missoula megafloods along the Wenatchee reach under several geographic and ice-margin scenarios (Denlinger et al. 2017). The nonlinear shallow-water equations (Berger et al. 2011) describe the physics of water in open channels on digitized regional topography. Vetted results of ongoing model runs are months and years away.

Figure 15.

DEM image of Moses Coulee bar. Megaflood(s) down Moses Coulee built the huge bar into and across Columbia valley (Bretz 1930, Bretz et al. 1956, Hanson 1970) and in late Wisconsin time dammed Columbia River (Waitt 1977a, 1982). Including its severed upvalley part Rock Island bar, Moses Coulee bar is at least 13 km long—an earthen dam massive enough to block off the valley for some time. Tall cliffs cut into Columbia River Basalt bedrock on the east valley side show that the bar pushed Columbia River west from an earlier long-term course nearer the cliffs (dashed line). MCB, Moses Coulee bar (main part); RIB, Rock Island bar; RIR, former Rock Island rapids, since 1931 drowned behind Rock Island Dam (the dark line just below); T, Trinidad.


Yet early results help visualize and quantify the system seemingly responsible for the fine sediment in lower Wenatchee valley. One explored dambreak-model scenario is from a maximal glacial Lake Missoula at surface level 1295 m, lake volume about 2500 km3. In this scenario, a near-maximal Okanogan lobe blocks Columbia valley between Grand Coulee and Chelan, making Grand Coulee (assumed completely open, as now) the westmost floodway. About 14 hours after dambreak Lake Missoula floodwater via Grand Coulee and the Telford scabland tract is filling Quincy basin while outfl ow down lower Crab valley spills into Columbia valley at Beverly (Figure 17A). By 20 hours Quincy basin has filled to spill through western saddles, and the head of water at Beverly has backflooded up the Columbia to Wenatchee (Figure 17B). From 30 to 120 hours Quincy basin fills and begins to drain as Pasco basin also fills—and deepening water backfloods up the Columbia and up Wenatchee valley to above Dryden (Figure 17C).

Figure 16.

Longitudinal profiles (water-surface and energy-line values) of model discharge 13 million m3s-1down northwest ‘big bend' reach of Columbia valley from above Grand Coulee intake to below Sentinal Gap. Computations by Corps of Engineers HEC-RAS hydraulic program through 154 topographic cross-sections. Steep drops are where fl ow goes hydraulically “critical” through constrictions— the greatest discharge that can pass through such gaps. Just below these “critical” reaches water goes turbulent, slows, and surface level rises. Solid circles plot 24 of the highest erratics in Columbia and lower Wenatchee valleys (Table 1), projected to valley centerline.


Silt-Routing Enigmas

The silt beds in lower Wenatchee valley contain no clear evidence how Missoula floodwater routed: down mainstem Columbia valley, or up valley by way of either Moses Coulee or Quincy basin. Missoula floodwater bulked with loess along eastern scabland routes should be more silt laden than floodwater by western routes like Moses Coulee. Backfloods from the southeast—Quincy basin or partly Pasco basin—should be highly silty. Such water backflooding as high as 350 m accommodates all the bedded-silt deposits in the Wenatchee-Trinidad reach of the Columbia and the Dryden rhythmites (Figure 12). This backflood level is plausible from an apparent 419 m maximum level of ponding in Quincy basin—though this level drops in the 55 km course to the Columbia (Figures 12, 17). The 10 rhythmites at Dryden are consistent with the dozens of late Wisconsin Missoula megafloods proven to have engorged the eastern scablands including Quincy basin (Waitt 1985).

Figure 17.

Snapshots of 2-D hydraulic dambreak model flood from a maximum glacial Lake Missoula at level 1295 m as by Denlinger et al. (2017). In this scenario upper Grand Coulee is fully open but the mainstem Columbia below and intakes to Moses Coulee blocked by Okanogan-lobe ice. Equations and algorithm by David L. George resemble Denlinger and O'Connell's (2010), but modernized as by Berger et al. (2011). Wenatchee area and lower Wenatchee valley backflood long distance via Quincy basin. Time (t) in hours and minutes since dambreak. This run innundates all known areas of rhythmic silt in Columbia valley (Figure 12) including the floor of lower Moses Coulee. Wen, Wenatchee area; GC, lower Grand Coulee; MC, Moses Coulee; Quin, Quincy basin; Bev, Beverly at mouth of lower Crab Creek; Pas, northwest edge of Pasco basin; SG, Sentinal Gap. Dryden in lower Wenatchee valley lies above the “D” in Dry.


But a higher flood level—up to 385 m— is needed for silt lying between Peshastin and Leavenworth. This level appears rather high to have come from Quincy basin (Figure 12). The ongoing 2-D hydraulic modeling (Denlinger et al. 2017) should eventually limit what's hydraulically possible of megafloods routed via Quincy basin.

Alternatively the Dryden-Leavenworth silt came from floods via Moses Coulee. Such late Wisconsin megafloods likely entered the Columbia at a higher level than those from Quincy basin and backflooded Wenatchee valley higher. Yet there are ten beds at Dryden but no separate evidence of more than five late Wisconsin megafloods via Moses Coulee (Waitt 1985, 1994, 2016). Moses Coulee floods, having routed through a large glacial Lake Columbia in a region of relatively meager Pleistocene loess, should carry less silt (and fewer erratics) than Missoula floods along other routes.

Another alternative for the high silt between Peshastin and Leavenworth is Missoula flood down the mainstem Columbia. In this scenario there's no question that this floodwater was high enough to backflood Wenatchee valley to above 385 m (Figures 12, 16). Yet except for the Dryden rhythmites, there is no other known stratigraphic evidence to prove more than one such flood.


Despite lingering enigmas, several large late Wisconsin events in the Wenatchee area can be distinguished in sequence.

  1. Landslide into the Columbia at Malaga. It may have dammed a persisting lake, but at some level below 275 m.

  2. Enormous Missoula flood(s) ice-rafting diverse stones down the Columbia to about altitude 495 m near Wenatchee. Such flooding would erode a Malaga landslide and drain any lake.

  3. Okanogan ice lobe blocks the Columbia upvalley, the river thus diverted upstream into some eastern course. Great Missoula floods down Moses Coulee build giant bar out mouth of Moses Coulee that dams Columbia River into a provable lake to altitude 275 m. This natural earthen dam likely holds for several to many centuries, but the reduced river eventually cuts through and restores a low-level valley.

  4. Repeated Missoula backfloods along Columbia valley via Quincy basin. Sand-silt rhythmites in Wenatchee valley at Dryden (315 m altitude) are from backfloods. They could stem from huge early down-Columbia floods, from later Moses Coulee floods, or from yet later floods via Quincy basin. Their abundant silt and presence of similar silt beds in Columbia valley between Trinidad and Wenatchee suggest Quincy basin floods. Poorly exposed thick silt farther up Wenatchee valley between Peshastin and Leavenworth (330–385 m) may originate from the same backfloods as laid the Dryden rhythmites.


Dan Long, who had accompanied his father on some 1960s Wenatchee valley field excursions, supplied Bill Long's old field photographs and topographic-map plots with their marginal pencilled notes. Ken and Susan Lacy as generous hosts at Trinidad have facilitated Waitt's 2017 fieldwork. Jim E. O'Connor's discussions during fieldtrips 2015–2018 question certain of my earlier explanations and thus sharpen new observations. Using 1990s topographic cross-sections by Christopher J. Harpel, Austin T. Rains in 2017 ran the HEC-RAS 1-D hydraulic model of Figure 16. New 2-D hydraulic modeling by Roger Denlinger and David George help quantify Missoula floods under different geographic scenarios digitally mastered by Charles Cannon. George supplied output plots from which Figure 17 derives. Peer reviews by William Scott, Brian Atwater, and John Clague much improve the manuscript and some of the figures.

Literature Cited


Atwater, B. F. 1984. Periodic floods from glacial Lake Missoula into the Sanpoil arm of glacial Lake Columbia, northeastern Washington. Geology 12:464–467. Google Scholar


Atwater, B. F. 1986. Pleistocene glacial-lake deposits of the Sanpoil River valley, northeastern Washington. US Geological Survey Bulletin 1661. US Government Printing Office, Washington, DC. Google Scholar


Atwater, B. F. 1987. Status of glacial Lake Columbia during the last floods from glacial Lake Missoula. Quaternary Research 27:182–201. Google Scholar


Baker, V. R. 1973. Paleohydrology and sedimentology of Lake Missoula flooding in eastern Washington. Geological Society of America Special Paper 144. Geological Society of America, Boulder, CO. Google Scholar


Berger, M. J., D. L. George, R. J. LeVeque, and K. T. Mandli. 2011. The GeoClaw software for depth-averaged flows with adaptive refinement. Advances in Water Resources 34:1195–1206. Google Scholar


Bjornstad, B. N. 2014. Ice-rafted erratics and bergmounds from Pleistocene outburst floods, Rattlesnake Mountain, Washington, USA. E & G Quaternary Science Journal 63:44–59. Google Scholar


Bretz, J H. 1930. Valley deposits immediately west of the Channeled Scabland. Journal of Geology 38:385–422. Google Scholar


Bretz, J H. 1959. Washington's Channeled Scabland. Bulletin 45. Washington Division of Mines and Geology, Olympia. Google Scholar


Bretz, J H., H. T. U. Smith, and G. E. Neff. 1956. Channeled Scabland of Washington—New data and interpretations. Geological Society of America Bulletin 67:957–1049. Google Scholar


Chappell, W. M. 1936. Geology of the Wenatchee [15-minute] quadrangle, Washington. Ph.D. Dissertation, University of Washington, Seattle. Google Scholar


Denlinger, R. P., and D. R. H. O'Connell. 2010. Simulations of cataclysmic outburst flood from Pleistocene glacial Lake Missoula. Geological Society of America Bulletin 122:678–689. Google Scholar


Denlinger, R. P., D. L. George, C. M. Cannon, R. B. Waitt, and J. E. O'Connor. 2017. Modeling cataclysmic outburst floods from Pleistocene glacial Lake Missoula [abstract]. In Geological Society of America, 2017 Annual Meeting, Seattle, WA. Paper 110–5. Google Scholar


Flint, R. F. 1935. Glacial features of the southern Okanogan region. Geological Society of America Bulletin 46:169–193. Google Scholar


Flint, R. F. 1937. Pleistocene drift border in eastern Washington. Geological Society of America Bulletin 48:203–232. Google Scholar


Fulton, R. J. 1969. Glacial lake history, southern interior plateau, British Columbia. Geological Survey of Canada Paper 69–37. Department of Energy, Mines and Resources, Ottawa. Google Scholar


Gresens, R. L. 1983. Geology of the Wenatchee and Monitor quadrangles, Chelan and Douglas Counties, Washington. Bulletin 75. Washington Division of Geology and Earth Resources, Olympia. Google Scholar


Griggs, A. B. 1973. Geologic map of the Spokane quadrangle, Washington, Idaho, and Montana. Miscellaneous Geologic Investigations Map I-768. US Geological Survey, Washington, DC. Google Scholar


Griggs, A. B. 1976. The Columbia River Basalt Group in the Spokane quadrangle, Washington, Idaho, and Montana. Bulletin 1413. US Geological Survey, Washington, DC. Google Scholar


Hanson, L.G. 1970. The origin and deformation of Moses Coulee and other scabland features on the Waterville Plateau, Washington. Ph.D. Dissertation, University of Washington, Seattle. Google Scholar


Harpel, C. J., R. B. Waitt, and J. E. O'Connor. 2000. Paleodischarges of the late Pleistocene Missoula floods, eastern Washington, USA. [abstract]. Orkustofnun Rept. OS-2000/036. Extremes of the Extremes Conference, Reykjavík, Iceland. p. 21. Google Scholar


Huntting, M. T., W. A. G. Bennett, V. E. Livingston Jr. , and W. S. Moen (compilers). 1961. Geologic Map of Washington (1:500,000 scale). Washington Division of Mines and Geology, Olympia. Google Scholar


Landes, H. 1911. The road materials of Washington. Bulletin 2. Washington Geological Survey, Olympia. Google Scholar


Lesemann, J. E., and T. A. Brennand. 2009. Regional reconstruction of subglacial hydrology and glaciodynamic bahaviour along the southern margin of the Cordilleran ice sheet in British Columbia, Canada and northern Washington State, USA. Quaternary Science Reviews 28:2420–2444. Google Scholar


O'Connor, J. E., and V. R. Baker. 1992. Magnitudes and implications of peak discharges from glacial Lake Missoula. Geological Society of America Bulletin. 104:267–279. Google Scholar


Page, B. M. 1939. Multiple alpine glaciation in the Leavenworth area, Washington. Journal of Geology 47:785–815. Google Scholar


Pardee, J. T. 1942. Unusual currents in glacial Lake Missoula, Montana. Geological Society of America Bulletin 53:1569–1599. Google Scholar


Porter, S. C. 1969. Pleistocene geology of the east-central Cascade Range, Washington. Guidebook for Third Pacific Coast Friends of the Pleistocene Field Conference. University of Washington, Seattle. Google Scholar


Stanton, K.M., R. B. Waitt, and W. A. Long. 2019. Pre-Wisconsin valley-glacier erratics between Leavenworth and Peshastin, Wenatchee valley, Washington: Northwest Science 92(5):311–317. Google Scholar


Tabor, R. W., R. B. Waitt, V. A. Frizzell Jr. , D. A. Swanson, G. R. Byerly, and R. D. Bentley. 1982. Geologic map of the Wenatchee 1:100,000 quadrangle, central Washington. Miscellaneous Investigations Map I-1311. US Geological Survey Information Services, Denver, CO. Google Scholar


Tabor, R. W., V. A. Frizzell Jr. , J. T. Whetten, R. B. Waitt, D. A. Swanson, G. R. Byerly, D. B. Booth, M. J. Hetherington, and R. E. Zartman. 1987. Geologic map of the Chelan 30-minute by 60-minute quadrangle, Washington, scale 1:100,000. Miscellaneous Investigations Map I–1661. US Geological Survey Information Services, Denver, CO. Google Scholar


Trimble, D. E. 1963. Geology of Portland, Oregon, and adjacent areas. US Geological Survey Bulletin 1119. US Government Printing Office, Washington, DC. Google Scholar


Waitt, R. B. 1972. Geomorphology and glacial geology of the Methow drainage basin, eastern North Cascade Range, Washington. Ph.D. Dissertation, University of Washington, Seattle. Google Scholar


Waitt, R. B. 1977a. Guidebook to Quaternary geology of the Columbia, Wenatchee, Peshastin, and upper Yakima valleys, west-central Washington. Open-File Report 77–753. Available online at (accessed 18 February 2019). Google Scholar


Waitt, R. B. 1977b. Missoula flood sans Okanogan lobe [abstract]. Geological Society of America, Abstracts with Program 9:770. Google Scholar


Waitt, R. B. 1980. About forty last-glacial lake Missoula jökulhlaups through southern Washington. Journal of Geology 88:653–679. Google Scholar


Waitt, R. B. 1982. Surficial deposits and geomorphology. In R. W. Tabor, R. B. Waitt, V. A. Frizzell Jr. , D. A. Swanson, G. R. Byerly, and R. D. Bentley. Geologic map of the Wenatchee 1:100,000 quadrangle, central Washington. Miscellaneous Investigations Map I-1311. US Geological Survey Information Services, Denver, CO. Google Scholar


Waitt, R. B. 1984. Periodic jökulhlaups from Pleistocene glacial Lake Missoula—new evidence from varved sediment in northern Idaho and Washington. Quaternary Research 22:46–58. Google Scholar


Waitt, R. B. 1985. Case for periodic, colossal jökulhlaups from Pleistocene glacial Lake Missoula. Geological Society of America Bulletin 96:1271–1286. Google Scholar


Waitt, R. B. 1987. Erosional landscape and surficial deposits. In R. W. Tabor, V. A. Frizzell Jr. , J. T. Whetten, R. B. Waitt, D. A. Swanson, G. R. Byerly, D. B. Booth, M. J. Hetherington, and R. E. Zartman. Geologic map of the Chelan 30-minute by 60-minute quadrangle, Washington, scale 1:100,000. Miscellaneous Investigations Map I-1661. US Geological Survey Information Services, Denver, CO. Google Scholar


Waitt, R. B. 1994. Scores of gigantic, successively smaller Lake Missoula floods through Channeled Scabland and Columbia Valley. In D. A. Swanson and R. A. Haugerud (editors), Geologic field trips in the Pacific Northwest. Geological Society of America, Annual Meeting, Department of Geological Sciences, University of Washington, Seattle. v. 1, Chapter 1–K. Google Scholar


Waitt, R. B. 2016. Megafloods and Clovis Cache at Wenatchee, Washington. Quaternary Research 85:430–444. Google Scholar


Waitt, R. B. 2017. Pleistocene glaciers, lakes, and floods in north-central Washington State. In R. A. Haugerud and H. M. Kelsey (editors), From the Puget Lowland to East of the Cascade Range—Geologic Excursions in the Pacific Northwest. Geological Society of America Field Guide 49:175–205. Google Scholar


Waitt, R. B., and R. M. Thorson. 1983. Cordilleran Ice Sheet in Washington, Idaho, and Montana. In H. E. Wright, Jr. (editor). Late-Quaternary Environments of the United States; v. 1, The Late Pleistocene ( S. C. Porter, editor). University of Minnesota Press, Minneapolis. Pp. 53–70. Google Scholar


Waitt, R. B., R. P. Denlinger, and J. E. O'Connor. 2009. Many monstrous Missoula floods down Channeled Scabland and Columbia Valley. Geological Society of America Field Guide 15:775–844. Google Scholar


Waitt, R. B., R. M. Breckenridge, E. P. Kiver, and D. F. Stradling. 2016. Late Wisconsin Cordilleran Icesheet and colossal floods in northeast Washington and north Idaho. In E. S. Cheney (editor), The Geology of Washington and Beyond—From Laurentia to Cascadia. University of Washington Press, Seattle. Pp. 233–256. Google Scholar


Waters, A. C. 1933. Terraces and coulees along the Columbia River near Lake Chelan, Washington. Geological Society of America Bulletin 44:783–820. Google Scholar


Whetten, J. T., and R. B. Waitt. 1978. Preliminary geologic map of the Cashmere quadrangle, Chiwaukum lowland, Washington. Miscellaneous Field Studies Map MF-908 (scale 1:24,000). US Geological Survey, Reston, VA. Google Scholar
Richard B. Waitt, William A. Long, and Kelsay M. Stanton "Erratics and Other Evidence of Late Wisconsin Missoula Outburst Floods in Lower Wenatchee and Adjacent Columbia Valleys, Washington," Northwest Science 92(sp5), 318-337, (1 January 2019).
Received: 16 February 2018; Accepted: 3 December 2018; Published: 1 January 2019

Back to Top