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

STRATIGRAPHY

Formal stratigraphic subdivision of the Columbia River Basalt Group has recently been made (Swanson and others, 1979b) (fig. 2). considerable effort was expended in doing this, in order to provide a strong framework for topical studies. Some additional distinctive units have been found in Idaho and Oregon since the nomenclature for the Columbia River Basalt Group was established, but they can easily be given member rank and assigned to one of the three formations in the Yakima Basalt Subgroup.

Figure 2. Stratigraphic nomenclature, age, magnetic polarity, and chemical type for units in the Columbia River Basalt Group. N, normal magnetic polarity; R, reversed; T, transitional. Subscripts refer to magnetostratigraphic units of Swanson and others (1977; 1979b). Geologic time scale from Berggren and Van Couvering (1974). From Swanson and others (1979b). (click on image for an enlargement in a new window)

The criteria used to recognize specific stratigraphic units include megascopic and less commonly microscopic petrography, magnetic polarity, and chemical composition. Megascopic petrography and magnetic polarity can be determined in the field and, taken together in over-all stratigraphic context, are generally sufficient to identify a particular unit. Ambiguities commonly arise, however, and chemical analyses provide an invaluable and independent guide for checking and correcting field identifications. In fact, use of chemical analyses for correlation purposes is so rewarding that no study of the basalt requiring identification of flows should be undertaken without provision for chemistry. Physical characteristics such as weathering color, size and shape of vesicles, thickness, and type of columnar jointing have been used by some past workers as correlation criteria, but we have found them unreliable because of lateral variability except in some local areas.

The formal stratigraphic units of the Columbia River Basalt Group (fig. 2) have been described in detail by Mackin (1961), Bingham and Grolier (1966), Schmincke (1967b), Swanson and others (1979b), Swanson and Wright (1978), and Camp and others (1979); only general statements are made here.

Imnaha Basalt

Outcrops of the Imnaha Basalt, the oldest formation in the group, are confined to extreme southeast Washington, northeast Oregon, and adjacent parts of Idaho, where feeder dikes are known (fig. 3A). Whether the Imnaha occurs farther west beneath younger rocks is conjectural. It covers a surface of rugged local relief and has an aggregate thickness of more than 500 m. Most flows in the formation are coarse grained end plagioclase phyric (Hooper, 1974). Five chemical types have been distinguished (Holden and Hooper, 1976; Kleck, 1976; Vallier and Hooper, 1976; Reidel, 1978; table 1). Trace element compositions are given by Nathan and Fruchter (1974). Many flows contain zeolite amygdules (Kleck, 1976), and smectitic alteration is widespread. The Imnaha Basalt conformably underlies the Grande Ronde Basalt, and future work may find areas in which the two formations interfinger. Most of the Imnaha has normal magnetic polarity, but the oldest and youngest flows known have reversed polarity, based on measurements with a portable fluxgate magnetometer. Two samples of the Imnaha have ⁸⁷Sr/⁸⁶Sr initial ratios of .7044 and .7043 (McDougall, 1976).

Figure 3. Maps showing generalized distribution and feeder dikes for the formations in the Columbia River Basalt Group. From Swanson and Wright (1978, fig. 3.4). (click on image for an enlargement in a new window)

Picture Gorge Basalt

Much of the Picture Gorge Basalt is apparently coeval with the middle part of the Grande Ronde Basalt (fig. 2), as judged by the interfingering of the two formations (Cockerham and Bentley, 1973; Nathan and Fruchter, 1974) and reconnaissance magnetostratigraphic work (R.D. Bentley and D.A. Swanson, unpub. data, 1977) along lower Butte Creek, Oregon (fig. 1). The Picture Gorge crops out only in and surrounding the John Day Basin in north-central Oregon (fig. 3A), where feeder dikes comprise the Monument dike swarm (Waters, 1961; Fruchter and Baldwin, 1975). Possibly the ancestral Blue Mountains uplift kept most flows from spreading northward out of the basin. The formation is at least 800 m thick; Thayer and Brown (1966) report a thickness of more than 1800 m at one locality, although faulting may have duplicated part of the section. The Picture Gorge contains aphyric to highly plagioclase-phyric flows with compositions falling into a broad field known as Picture Gorge chemical type (Wright and others, 1973; table 1). Trace element data are given by Osawa and Goles (1970) and Nathan and Fruchter (1974). Some of the thicker flows contain pegmatoids (Lindsley and others, 1971). The formation can be subdivided into three informal units based on field characteristics and magnetic polarity, according to R.D. Bentley (Swanson and others, 1979b). The Picture Gorge at its type section has an average ⁸⁷Sr/⁸⁶Sr initial ratio of .7037 (McDougall, 1976).

Grande Ronde Basalt

The Grande Ronde Basalt is the oldest formation of the Yakima Basalt Subgroup (fig. 2) and the most voluminous and areally extensive formation in the entire Columbia River Basalt Group, underlying most of the Columbia Plateau (fig. 3B) with an estimated volume of more than 150,000 km³. Its thickness varies widely depending on underlying topography; the thickest preserved section exceeds 1000 m in drill holes in the Pasco Basin, and sections 500-700 m thick occur in the Blue Mountains and other uplifted or deeply incised areas. Most flows in the formation are very sparsely plagioclase-phyric to essentially aphyric, although a few of the oldest flows in and near the Lewiston Basin contain abundant large plagioclase phenocrysts. Major element compositions fall in a broad range termed Grande Ronde chemical type (table 1). Flows having different compositions within this range are interleaved throughout the section, although flows of high-Mg type rather consistently overlie flows of low-Mg type in the western part of the plateau. The Grande Ronde Basalt is subdivided into four magnetostratigraphic units on the basis of magnetic polarity (Swanson and Wright, 1976b; Swanson and others, 1979b); these units provide the only useful subdivisions of the formation on a plateau-wide basis. Feeder dikes occur throughout the eastern half of the plateau and are apparently not confined to distinct swarms as formerly thought (Waters, 1961). The Grande Ronde conformably overlies the Imnaha Basalt, intertongues with the Picture Gorge Basalt, and conformably underlies and locally interfingers with the Wanapum Basalt. Commonly a thick soil and, locally, weakly lithified clastic sediments occur on top of the Grande Ronde; they indicate a significant time break, although probably no longer than a few tens of thousands of years judging from the local interbedded relations between the Grande Ronde and Wanapum Basalts in southeast Washington (fig. 4). The ⁸⁷Sr/⁸⁶Sr initial ratios may increase slightly upsection, from values of about .7046 in the older flows to about .7052 in the younger flows (McDougall, 1976); this interpretation is tentative, as the older flows were sampled at different locations than the younger flows, so that variation due to lateral isotopic heterogeneity in the source rocks is also possible.

Table 1. Average major-element compositions for chemical types in the Colombia River Basalt Group
[Averages include analyses available through March 1977; see Wright and others (1979) for more recent averages and further discussion of chemical types ]

Chemical types (defined by method of Wright and Hamilton, 1978)

Figure 4. Schematic stratigraphic section in Benjamin Gulch, 3 km south of Pomeroy, Washington, showing chemical types for 13 basalt flows. Note interbedded nature of chemically different flows. Uncorrected for north dip of about 4 degrees. From Swanson and others (1979b).

Wanapum Basalt

The Wanapum Basalt is the most extensive formation exposed at the surface of the Columbia Plateau (fig. 3C) but is much less voluminous than the Grande Ronde, probably having a volume of less than 10,000 km³. On a local scale, the Wanapum conformably overlies the Grande Ronde, except for minor erosional unconformities or interbedded relations. On a regional scale, however, the Wanapum overlies progressively older basalt from the center toward the eastern margin of the plateau. Such onlap is not apparent along the northern and western margins, however. These relations suggest that the plateau had tilted westward before Wanapum time.

The oldest member of the Wanapum Basalt, the Eckler Mountain Member, occurs in the Blue Mountains and adjoining foothills of southeast Washington and northeast Oregon (Swanson and others, 1979b, 1980). The oldest flow in the member, the basalt of Robinette Mountain, is diktytaxitic and contains the lowest K₂O and incompatible trace element concentrations of any other flow in the Yakima Basalt Subgroup (tables 1 and 2); it was erupted from a long fissure south of Dayton, Washington. Several flows and dikes of the next youngest flow—the basalt of Dodge, a vary coarse-grained plagioclase-phyric, grusy-weathering unit—form excellent markers in the Blue Mountains. The basalt of Dodge is chemically similar to but much coarser and more porphyritic than some high-Mg flows of the Grande Ronde Basalt (tables 1 and 2). The basalt of Shumaker Creek, the youngest unit in the member, is neither widespread nor easily recognized in the field, although its high K₂O and P₂O₅ are distinctive (tables 1 and 2).

Table 2. Average trace-element compositions for chemical types in the Yakima Basalt Subgroup
[Instrumental neutron activation analyses by L.J. Schwarz under direction of J.J. Rowe]

The Frenchman Springs Member overlies and locally interfingers with the Eckler Mountain Member and crops out widely in the central and western parts of the plateau (fig. 5A). Its volume is probably 3000 to 5000 km³. Generally three to six flows, in places as many as ten, occur in any one section. Flows were erupted from north-northwest-trending dikes extending through the Walls Walla area of southeast Washington (Swanson and Wright, 1978; Swanson and others, 1979b, 1980). Highly porphyritic flows near Soap Lake at the southern end of Grand Coulee may have erupted along the northward extension of the known feeder system. Most flows of the Frenchman Springs Member contain rare to abundant glomerocrysts of plagioclase, although some are aphyric and indistinguishable in the field from some flows of Grande Ronde Basalt. The member has a high FeO and TiO₂ composition known as Frenchman Springs chemical type (tables 1 and 2). The ⁸⁷Sr/⁸⁶Sr initial ratio is about .7053 (McDougall, 1976). The Frenchman Springs is overlain by the Roza Member.

Figure 5. Maps showing generalized distribution and feeder dikes for the Frenchman Springs, Roza, and Priest Rapids Members of the Wanapum Basalt and the Pomona Member of the Saddle Mountains Basalt. From Swanson and Wright (1978, fig. 3.5). (click on image for an enlargement in a new window)

The Roza Member, a highly plagioclase-phyric unit with a volume of about 1500 km³, is well known, and readers are referred to papers by Lefebvre (1970) and Swanson and others (1975; 1979b) for details. The member was erupted from a linear vent system more than 165 km long in the eastern part of the plateau (fig. 5B). The Roza consists principally of two cooling units, although more thin units occur near the vent system. The composition of the Roza is similar to that of the Frenchman Springs Member although on the average slightly richer in MgO (tables 1 and 2). Its ⁸⁷Sr/⁸⁶Sr initial ratio is about .7054 (McDougall, 1976).

The priest Rapids Member overlies the Roza Member and is the youngest basalt throughout most of the northern part of the Columbia plateau. The member occurs as far southwest as the Columbia Gorge (fig. 5C). All known feeder dikes are confined to the far eastern part of the province. Several dikes occur near Orofino, Idaho, and along Slate Creek about 16 km east of Freedom, Idaho (W.H. Taubeneck, T.L. Wright and D.A. Swanson, unpub. chemical data, 1977; V.E. Camp in Swanson and others, 1979a), and probable vents are located near Emida, Idaho, and in Palouse, Washington. Other vents presumably exist in northern Idaho, as intracanyon flows of the member occur far up the ancestral St. Joe River valley. The estimated volume of the priest Rapids is 2000-3000 km³. In and near Spokane, flows of the Priest Rapids Member fill valleys as much as 100 m deep eroded into the main part of the Latah Formation, a sequence of fine-grained clastic sediments interbedded with and overlying the Grande Ronde Basalt. Flows of the Priest Rapids invade sediments of the Latah in many places near Spokane as well as other sediments near Orofino, Idaho. The Priest Rapids Member contains magnetically reversed flows of two distinctly different compositions, a very high FeO and TiO₂ type (Rosalia chemical type) and a high MgO type (Lolo chemical type) (tables 1 and 2). The flows of Rosalia chemical type are found throughout most of the extent of the member and are consistently older than those of Lolo chemical type, which are confined to the southern two-thirds of the member's outcrop area. A few thin flows of different compositions occur near vent areas in northern Idaho and adjacent Washington. McDougall (1976) obtained an ⁸⁷Sr/⁸⁶Sr initial ratio of .7053 on a flow of Rosalia chemical type near Frenchman Springs Coulee.

Saddle Mountains Basalt

This formation, the youngest in the Columbia River Basalt Group, is about 13.5 to 6 m.y. old and contains flows erupted sporadically during a period of waning volcanism, deformation, canyon cutting, and development of thick but local sedimentary deposits between flows. The Saddle Mountains Basalt has a volume of only about 3000 km³, less than one per cent of the total volume of basalt, yet contains by far the greatest chemical and isotopic diversity of any formation in the group.

The Umatilla Member is the oldest and one of the most extensive members in the formation. It occurs in extreme southeast Washington and northwest Oregon (the Troy and Lewiston basins and Uniontown Plateau) (Price, 1977; Ross, 1978; Swanson and others, 1980); vent areas and a feeder dike occur in the Puffer Butte area (fig. 1; Price, 1977). Remnants of the Umatilla fill a broad shallow paleovalley leading from the Troy basin across the present-day Blue Mountains in northern Oregon to the Milton-Freewater area, where the flow spread out of the paleovalley as a sheet flood covering much of south-central Washington (D.A. Swanson and T.L. Wright, unpub. map, 1978; Swanson and others, 1979a). Lava was channelled along some canyons eroded during post-Wanapum time in the western part of the Columbia Plateau, as along Yakima Ridge (R.D. Bentley in Swanson and others, 1979a). The distribution of the Umatilla provides the earliest evidence for extensive erosion and canyon-cutting of the Columbia River Basalt Group on the Columbia Plateau, although erosion of basalt was substantial in the Columbia Gorge before Priest Rapids time (Beeson and Moran, 1979). The Umatilla has an unusual chemical composition characterized by lower contents of CaO and MgO and higher contents of Na₂O, K₂O, and incompatible trace elements than most other flows in the group (tables 1 and 2). The content of Ba is 2000 ppm or more, sufficient alone to identify the member. P.R. Hooper (oral commun., 1979) has recognized two flows of slightly different K₂O, TiO₂ and Ba contents in the member. The ⁸⁷Sr/⁸⁶Sr initial ratio of the Umatilla is high, about .7092 (McDougall, 1976). The Umatilla underlies the Wilbur Creek Member.

The Wilbur Creek Member and the overlying Asotin Member, distinctly different flows but possibly chemically related, were apparently erupted in the Clearwater embayment of Idaho. From there, the flows advanced down valleys and gorges leading from the Uniontown Plateau to the central part of the Columbia Plateau; remnants of the valley-filling flows occur east and west of lower Cow Creek (fig. 1), near Warden and Othello, and elsewhere (Swanson and others, 1980). The flows crossed the northern part of the Pasco Basin (Myers and Price, 1979) and moved down a canyon along Yakima Ridge possibly as far west as Yakima (R.D. Bentley in Swanson and others, 1979a). The flows overlie quartzitic gravel of extra-plateau derivation along Yakima Ridge and trace a westward course of the ancestral Columbia River from the Pasco Basin to Yakima. The Wilbur Creek has a major element composition similar to that of the intermediate-Mg Grande Ronde chemical type, and the Asotin is similar to the basalt of Robinette Mountain (table 1); however, trace element compositions easily discriminate the flows (table 2).

The Weissenfels Ridge Member overlies the Asotin Member and contains several flows confined mainly to the Lewiston Basin and presumably erupted there. The basalt of Lewiston Orchards, one flow averaging 10-15 m thick, is sparsely plagioclase-phyric and contains groundmass olivine visible with a hand lens. It is relatively rich in MgO and poor in K₂O (table 1). The overlying basalt of Slippery Creek consists of several flows, at least one of which contains abundant groundmass olivine. Its chemical composition differs from other flows of the Saddle Mountains Basalt. The basalt of Anatone (Price, 1977) has a major element composition similar to the Lolo chemical type but is enriched in light REE.

The Esquatzel Member (fig. 2) occurs as isolated remnants of an intracanyon flow along and just north of the modern Snake River upstream from Devils Canyon (Swanson and others, 1980). It apparently was erupted within the ancestral Snake drainage, flowed downcanyon, and entered the ancestral Columbia River valley in the central part of the plateau. The Esquatzel then flowed along Yakima Ridge in a course similar to that of the Wilbur Creek and Asotin Members. The Esquatzel is distinguished petrographically by irregularly distributed phenocrysts and clots of strongly zoned plagioclase and clinopyroxene. Its major element composition can be confused with some low-TiO₂ flows of the Frenchman Springs Member of the Wanapum Basalt (table 1), but its trace element composition is distinctive (table 2) and its ⁸⁷Sr/⁸⁶Sr initial ratio of about .7146 extraordinarily high (the Mesa flow of Nelson and others, 1976).

The Pomona Member (fig. 2), well known through the work of Schmincke (1967b), occurs across the province from the Clearwater embayment in Idaho to southwest Washington near the coastline (the basalt of Packsack Lookout of Snavely and others, 1973), a distance of about 500 km (fig. 5D). The member, probably consisting of only one flow, was apparently erupted about 12 m.y. ago in the Clearwater embayment; V.E. Camp (in Swanson and others, 1979a) located feeder dikes northeast of Orofino, Idaho, for a flow probably correlative with the Pomona. It flowed out of the embayment down an ancestral Snake River canyon, virtually coincident with the modern canyon, to the central plateau, where it spread out as a broad sheet covering much of south-central Washington and extreme north-central Oregon. The member advanced along the ancestral Columbia River westward along Yakima Ridge to the site of Yakima. It can readily be traced as far west as Nosier, Oregon, in the Columbia Gorge. From there, its pathway to southwest Washington is unclear, but it presumably followed an ancestral Columbia drainage system much as earlier flows did. Peperites and invasive flows formed along the margin of the flow where it plowed into sediments, and a fused vitric tuff underlies the flow in many other places (Schmincke, 1967c). The Pomona, with a volume of more than 600 km³, is one of the most voluminous single flows in the group. It has a distinctive chemical composition (tables 1 and 2) and petrography (Schmincke, 1967b) and represents one of the best markers on the Columbia Plateau. Its ⁸⁷Sr/⁸⁶Sr initial ratio is about .7078 (Nelson and others, 1976; McDougall, 1976).

The Elephant Mountain Member (fig. 2) was probably erupted in part from a dike mapped by Ross (1978) in the Troy basin of northeast Oregon, where it is known as the Wenaha flow of Walker (1973b). Flows of the member advanced down the ancestral Snake River canyon as the Pomona Member had done about 1.5 m.y. earlier. The flows spread outward from the mouth of the canyon near Mesa, Washington, and covered much of south-central Washington, in many places capping volcaniclastic debris that had been erupted in the Cascades, carried eastward by rivers, lahars, and winds, and deposited on the Pomona Member. Recent mapping has defined the west and southwest margin of the member along a line extending approximately southward from Yakima to the Horse Heaven Plateau (Swanson and others, 1979a). The Elephant Mountain consists of several flows, all chemically similar, of normal and transitional magnetic polarity. Its major element composition is similar to that of Rosalia chemical type except for lower P₂O₅ (table 1), but its trace element content is distinct. The member has a ⁸⁷Sr/⁸⁶Sr initial ratio of about .7078 (McDougall, 1976; Nelson and others, 1976), similar to that of the Pomona.

The Buford Member (fig. 2), a single magnetically reversed flow 20-30 m thick, is the youngest known basalt on the plateau surface of extreme southeast Washington and northeast Oregon, where it is confined and presumably was erupted. Its major-element composition shows distinctly higher light REE contents (table 2).

The Ice Harbor Member (fig. 2), dated as about 8.5 m.y. old (McKee and others, 1977), was erupted from the central part of the Columbia Plateau, where dikes and remnants of vent areas have been recognized. The last previous eruptions from the central part of the plateau were those of the Frenchman Springs Member, about 6 m.y. before the Ice Harbor volcanism. Most flows are confined to the area of venting, but at least one flow spread westward to the Richland area and southwestward to Wallula Gap. The Ice Harbor Member can be subdivided into three readily mappable units of different compositions (tables 1 and 2). The lowest unit—the basalt of Basin City—contains plagioclase and olivine phenocrysts, has normal magnetic polarity, and is chemically distinct. The middle unit—the basalt of Martindale—carriers clots and single crystals of clinopyroxene, plagioclase, and olivine and has reversed magnetic polarity; two related compositions characterize the Martindale (Helz, 1978), the dominant of which is listed in table 1. The upper unit—the basalt of Goose Island—contains sparse plagioclase and magnetite phenocrysts, has normal magnetic polarity, and has the most FeO-rich composition of any known flow in the group (and one of the most FeO-rich compositions of any terrestrial basalt). The three informal units have a similar ⁸⁷Sr/⁸⁶Sr initial ratio of about .7077 (Helz, 1978). Helz (1978) has recently completed an exhaustive experimental and petrogenetic study of the Ice Harbor Member. The Ice Harbor vent system is about 90 km long and has strong aeromagnetic expression (Swanson and others, 1979c).

The Lower Monumental Member, about 6 m.y. old, is the youngest member in the Saddle Mountains Basalt. It is confined to the modern Snake River Canyon between Devils Canyon and Asotin, Washington, a distance of about 150 km. Its source was presumably near Asotin or farther east but has not yet been identified. Its chemical composition is similar to Lolo chemical type, although slightly higher in alkalies and markedly higher in light REE (tables 1 and 2). The Lower Monumental has a high ⁸⁷Sr/⁸⁶Sr initial ratio of about .7109 (Nelson and others, 1976).