USGS Logo Geological Survey Circular 838
Guides to Some Volcanic Terrances in Washington, Idaho, Oregon, and Northern California


Edward M. Taylor
Department of Geology
Oregon Slate University


This geologic field trip guide is a revision of "Roadside Geology, Santiam and McKenzie Pass Highways, Oregon" (Taylor. 1968) with additions pertinent to geology of the Bend-Sisters area. Checkpoint mileages express cumulative distance from Central Oregon Community College in Bend, over a route which leads to Sisters, around the McKenzie Pass-Clear Lake-Santiam Pass highway loop, and back to Bend (see Index Map, Figure 1). Persons using this guide should anticipate one full day of at least 140 miles, and should check accessibility to McKenzie Summit during winter, spring, and late fall.

FIGURE 1. Index map. Roadside geology, Bend, Sisters, McKenzie Pass, and Santiam Pass. (click on image for an enlargement in a new window)


A calc-alkalic volcanic arc has been intermittently active during the last 10-15 m.y. along the eastern part of the central Cascade Range in Oregon. The late Pleistocene record of this volcanic activity is well preserved on the crest of the High Cascades; the best exposed record of early Pleistocene, Pliocene, and late Miocene Cascade volcanism is found in volcanic and volcaniclastic deposits on the east flank of the range and in the adjacent Deschutes Basin. In the following discussion, structural, stratigraphic, and magmatic features of the Western Cascade, High Cascade, and Deschutes Basin subprovinces are described and their interrelationships are briefly summarized.

Central High Cascade Province

The central High Cascade Range in Oregon is chiefly a Pleistocene volcanic platform of overlapping basalt and basaltic andesite lava flows whose aggregate thickness is generally unknown but probably exceeds 4,000 feet locally. This platform is elongate north-south and is 20-30 miles wide. A typical volcano of the platform is a broad shield of light-colored, vesicular basaltic andesite with a cinder cone core that has been invaded by plugs and radial dikes. Pleistocene examples exposed in cross-section by glacial erosion include Sphinx Butte south of Separation Creek canyon, Deer Butte north of Lost Creek canyon, and Bald Peter east of Jefferson Creek canyon; a perfectly preserved Holocene example is Belknap Crater on the McKenzie Pass summit. Some of the basaltic andesite volcanoes developed large composite structures reaching 10,000 feet elevation on a shield base 10 miles wide. Examples include The Husband, North Sister, Mount Washington, and Three Fingered Jack. In contrast, many volcanoes of the platform were active for only a brief time and produced small cinder cones with or without narrow lobes of lava. Holocene examples abound; they include Yapoah Cone, Twin Craters, and Sims Butte near McKenzie Pass, and Nash Crater, Lost Lake Cones, and Blue Lake Crater near Santiam Pass. Pleistocene cinder cones are no less abundant but they are not as well preserved. Examples of glaciated remnants of Pleistocene cones include Bluegrass Butte, Condon Butte, and Scott Mountain near McKenzie Pass and Maxwell Butte, Hoodoo Butte, and Cache Mountain near Santiam Pass.

A systematic temporal inhomogeneity exists with in the High Cascade platform. Early Pleistocene lavas were predominantly high-alumina olivine tholeiites in vesicular, thin, widespread units, commonly with pronounced diktytaxitic textures. Later Pleistocene lavas were predominantly high-alumina basaltic andesites in thick, platy units, generally with pilotaxitic textures. The early basalts crop out in greater abundance and variety along the western and eastern margins of the platform and in the walls of deeply glaciated canyons; however, identical basalts do occur at higher levels. Later basaltic andesites cover most of the platform but they are also found at lower stratigraphic levels. Examples of early basalts are well exposed along the west margin of the platform at Cupola Rock in Lost Creek canyon, in cuts of Highway 126 north of Trailbridge Reservoir, and in the valley of Hackleman Creek west of Fish Lake. Early basalts along the east margin of the platform are widespread in the upper Metolius River valley and in the vicinity of Sisters.

A systematic spatial inhomogeneity also exists within the High Cascade platform. West of Bend, in the vicinity of South Sister, silicic volcanic rocks are interbedded with and rest upon the mafic platform rocks. A silicic highland of some 15 miles breadth was produced by the development of rhyolite, rhyodacite, and dacite domes surrounded by andesite, dacite, and rhyodacite lavas and ash-flow tuffs. Most of this highland is mantled by mafic cinder cones and lavas of the Triangle Hill group and by composite volcanoes such as Broken Top. Examples of interbedded andesites, dacites, and rhyodacite lavas are common in upper Squaw Creek and Tumalo Creek canyons. Two units of black andesitic ash-flow tuff (Century Drive Tuff and Shevlin Park Tuff) and one unit of pink, devitrified, dacitic ash-flow tuff (Desert Spring Tuff) were erupted from the highland and are well exposed west of Bend. Two rhyodacite ash-flow tuffs (Tumalo Tuff and Lava Island Tuff) and one extensive lapilli-fall pumice deposit (Bend Pumice) of High Cascade origin are also exposed near Bend but were probably erupted from a vent south of the silicic highland.

South Sister volcano is chiefly andesite with minor dacite and rhyodacite. Broken Top (east of South Sister) is basaltic andesite with minor interbedded dacite and rhyodacite lavas and small-volume ash-flow tuffs. Middle Sister (north of South Sister) is basalt with minor basaltic andesite, andesite, dacite, and rhyodacite. South Sister and nearby volcanoes are relatively late products of long-continued, compositionally diverse, and localized silicic magmatism.

In summary, the central High Cascade Range is not the simple Pliocene-Pleistocene belt of andesite volcanoes commonly depicted in geology textbooks; instead, it is a broad Pleistocene platform of mafic composition in which open-textured basaltic lavas were at first predominant, then became subordinate to basaltic andesite. Silicic magma has invaded this platform throughout its development but only in isolated regions.

Deschutes Basin Province

The eastern margin of the central High Cascade platform is marked by a very irregular contact with the late Miocene and Pliocene Deschutes Formation. Early platform intracanyon lavas extend as much as 5 miles east of the Cascade foothills and isolated Pleistocene volcanoes of basalt and basaltic andesite rest on Pliocene rocks of the Deschutes Basin. Examples of High Cascade intracanyon lavas occur in lower Metolius River canyon, Deep Canyon, and near Squaw Creek, Bull Flat, and Tumalo Creek between Sisters and Bend. Isolated volcanoes of High Cascade affinity include Squaw Back Ridge, Long Butte, and Pilot Butte and Awbrey Butte near Bend.

The Deschutes Formation contains stream-deposited silt, sand, and gravel, andesitic-to-rhyodacitic ash flow and ash-fall tuffs, and interbedded basalt flows. The basaltic lavas were erupted from cinder cones and fissure vents within the Deschutes Basin but the epiclastic and volcaniclastic rocks were chiefly of Cascade provenance. Close to the Cascades, the formation becomes thicker, basaltic andesite lavas predominate, and the volcaniclastic rocks become discontinuous interbeds. Indicators of transport direction within the volcaniclastic sediments point eastward. Deschutes Formation source volcanoes were coincident with, or not far removed from the High Cascade axis. One deeply dissected remnant of an andesitic Deschutes Formation source volcano is located at the bend of Metolius River, 12 miles east of Mount Jefferson.

Strata of the Deschutes Formation between Warm Springs and Bend are generally flatlying except where they are offset by north-northwest-trending normal faults of small displacement. These faults are part of the Brothers-Sisters Fault Zone and are best exposed in cross-section along cuts of Highway 126 in Deep Canyon east of Sisters. The Deschutes Formation is unconformably underlain by folded and faulted basalts of the mid-Miocene Columbia River Group on the north and by silicic domes, lavas, and tuffaceous rocks on the south, variously ascribed to Clarno and John Day Formations of middle Tertiary age.

Parts of the High Cascade - Deschutes Formation contact are fault controlled. This is especially obvious at the west base of the Green Ridge escarpment, a 20-mile-long north-south fault block in which upper Deschutes Formation rocks stand 2,000 feet above the east edge of the High Cascade platform. K-Ar ages of rocks on the crest of Green Ridge and of lavas emplaced against the base of the scarp indicate that faulting occurred during the interval 2.5 to 4.5 m.y. ago (Armstrong and others, 1975). A closely related but much less spectacular feature is the Tumalo Fault between Sisters and Bend. The Tumalo Fault can be traced without interruption for 15 miles; north and south segments extend its length another 10 miles. Parts of the Tumalo Fault were reactivated during the Pleistocene.

Green Ridge probably marks the east flank of a prominent Pliocene volcanic complex. North and south of Green Ridge the trace of a major fault is obscure; isolated hills of Deschutes Formation are surrounded by lavas of the platform and a previously existing fault-controlled topography appears to have been extensively eroded and almost completely buried.

A correct interpretation of the east-margin fault system is of great significance in understanding High Cascade and Deschutes Basin geology. My interpretation can be summarized as follows:

1. Deschutes Formation basaltic andesite lavas, andesite lavas, and ash-flow tuffs were derived from volcanoes near the present High Cascade axis and flowed eastward into the Deschutes Basin more than 4.5 m.y. ago. Rock units deposited by this process have been traced from Green Ridge to the Deschutes River. They reveal a continuous, gentle eastward dip of only 1-2 degrees. Therefore, Green Ridge cannot represent a tilted-up fault block. It is, instead, a remnant of stable eastern Cascade foothills and its rocks still rest on an initial paleo-slope.

2. The maximum age yet obtained by radiometric dating of central High Cascade platform rocks is 3.9 m.y. Therefore, Deschutes Formation rocks and source volcanoes must lie beneath the Cascade platform and displacement of the east-margin fault system probably exceeds (and might greatly exceed) 3,000 feet.

3. If Green Ridge has not been displaced upward relative to the Deschutes Basin, the Cascade axis has been displaced downward. This presumably occurred about 4.5 m.y. ago, terminating deposition of Deschutes Formation rocks.

Is it possible that a whole range of Pliocene composite volcanoes foundered and was buried beneath the Pleistocene High Cascade platform? Such an interpreration is strongly suggested by available field evidence. With appropriate informality, this hypothetical assemblage of volcanic rocks might be called the "Plio-Cascades."

East Margin of the Central Weatern Cascade Province

Stratigraphic and structural relationships at the western margin of the central High Cascade platform are obscured by more extensive erosion, thicker alluvial cover, and more luxuriant vegetation than along the eastern margin. However, striking similarities are evident. Isolated Pleistocene volcanoes of basalt and basaltic andesite occur in the Western Cascades at least 20 miles west of the platform. Examples include Harter Mountain, a cinder cone and flow near Quartzville, and Battle Ax Mountain. Early High Cascade platform basalts occur as intracanyon lavas in the adjacent Western Cascades. Examples include diktytaxitic basalts of Foley Ridge in the canyon of McKenzie River and similar rocks in the canyon of North Santiam River.

Rocks of the Western Cascades adjacent to the central High Cascade platform are predominantly late Miocene and Pliocene mafic lavas (Armstrong and others, 1975; Sutter, 1978) with subordinate ash-flow tuffs and silicic volcanic domes. These rocks have been included in the Sardine Formation by Peck and others (1964) but should be assigned to the Outerson Formation of Thayer (1937). They are equivalent in lithology and age to the Deschutes Formation. They are generally flat-lying except where they are offset by northwest-trending normal faults of small displacement. Although many units of the Outerson Formation were vented along the eastern edge of the Western Cascades, several andesitic ash-flow tuffs and one basaltic andesite ash-flow tuff appear to have moved westward from Plio-Cascade volcanoes. Examples of Outerson Formation rocks can be seen along the south-to-north crestline of mountains which includes Frissell Point, Bunchgrass Mountain, Browder Ridge, Iron Mountain, Echo Mountain, Crescent Mountain, Three Pyramids, and Coffin Butte.

Outerson Formation rests with angular and/or erosional unconformity upon a complex and poorly exposed assemblage of moderately deformed, altered, silicic volcaniclastic rocks with subordinate lavas and intrusive bodies. These rocks are part of the Sardine Formation. Examples of pre-Outerson rocks are best seen west of the Frissell Point-Coffin Butte crest line in canyons tributary to McKenzie River and South, Middle, and North Santiam Rivers. Many of these rocks have been assigned K-Ar ages between 14 and 20 m.y. by Sutter (1978).

Although field evidence is still inconclusive, it is likely that the eastern part of the central Western Cascades was displaced several thousand feet down on the east side of a north-south fault system during the interval 4 to 5 m.y. ago. Outerson Formation rock units exceed 3,000 feet in thickness and are approximately horizontal in attitude along the Frissell Point-Coffin Butte crestline. In the east face of a 30-mile-long deeply eroded north-south escarpment, they "sky-out" over the younger High Cascade platform. A major fault has been found at the base of this escarpment in a few places. For example, in the vicinity of Belknap Hot Springs, rocks on the crestline west of the fault are 6.2 m.y. old; rocks of the same age east of the fault occur 2,000 feet lower (Armstrong and others, 1975). Throughout its length, this fault system has been obscured by glaciation of the escarpment and by lavas of the High Cascade platform deposited against the base of the escarpment.

If it can be demonstrated that the High Cascade axis was displaced downward along marginal boundary faults relative to the Deschutes Basin and Western Cascades during the interval 3 to 4.5 m.y. ago, the High Cascade platform should be viewed as e Pleistocene fill within a Pliocene graben (Figure 2). On the basis of my own limited experience, I believe that this condition prevailed north at least as far as Mount Adams and south at least as far as Crater Lake, but was not necessarily continuous in time or space. It is likely that such a structure, 20 miles wide, would subside along many fractures, chiefly trending north-south. Intragraben faults could have served as channelways for ascending magma. This might explain why volcanoes on the High Cascade platform commonly occur within long north-south alignments and why volcanoes adjacent to the platform tend not to do so. The early flood of gas-rich diktytaxitic basalts might be related to unusually rapid ascent of magma during a time of relative crustal tension. Correspondingly, the later dominance of basaltic andesite might reflect a slower ascent and greater opportunity for evolution of magmas.

FIGURE 2. Diagrammatic cross section of Western Cascade, High Cascade, and Deschutes Basin, Oregon. (click on image for an enlargement in a new window)

Considerable evidence now suggests that a broad calc-alkaline volcanic field consisting of Deschutes Formation, "Plio-Cascades," and Outerson Formation rocks covered the eastern half of the central Cascade Range during late Miocene and Pliocene time. If it is assumed that a subduction system was responsible for this volcanism, it might also be assumed that the subduction process became inactive or modified approximately 4.5 m.y. ago. This might have led to relaxation of the crust, subsidence along the volcanic axis, less frequent ascent of andesite-dacite magmas from more restricted, residual reservoirs, and much increased outpouring of basaltic magma from relatively shallow levels, probably associated more with Basin and Range magmatism than with a subduction system. However, it is easy to propose models. Testing their validity will require years of intense effort on many fronts.

<<< Previous <<< Contents >>> Next >>>

Last Updated: 28-Mar-2006