Abstract

The accreted Wrangellia flood basalts and associated sedimentary rocks that compose the prevolcanic and postvolcanic stratigraphy provide an unparalleled view of the architecture, eruptive environment, and accumulation and subsidence history of an oceanic plateau. This Triassic large igneous province extends for ∼2300 km in the Pacific Northwest of North America, from central Alaska and western Yukon (Nikolai Formation) to Vancouver Island (Karmutsen Formation), and contains exposures of submarine and subaerial volcanic rocks representing composite stratigraphic thicknesses of 3.5–6 km. Here we provide a model for the construction of the Wrangellia oceanic plateau using the following information and visualization tools: (1) stratigraphic summaries for different areas of Wrangellia; (2) new 40Ar/39Ar geochronology results; (3) compilation and assessment of geochronology and biostratigraphy for Wrangellia; (4) compiled digital geologic maps; (5) an online photographic archive of field relationships; and (6) a Google Earth file showing the mapped extent of Wrangellia flood basalts and linked field photographs.

Based on combined radiometric (U-Pb, 40Ar/39Ar, K-Ar), paleontological, and magnetostratigraphic age constraints, the Wrangellia flood basalts were emplaced during a single phase of tholeiitic volcanism ca. 230–225 Ma, and possibly within as few as 2 Myr, onto preexisting submerged arc crust. There are distinct differences in volcanic stratigraphy and basement composition between Northern and Southern Wrangellia. On Vancouver Island, ∼6 km of high-Ti basalts, with minor amounts of picrites, record an emergent sequence of pillow basalt, pillow breccia and hyaloclastite, and subaerial flows that overlie Devonian–Mississippian (ca. 380–355 Ma) island arc rocks and Mississippian–Permian marine sedimentary strata. In contrast, Alaska and Yukon contain 1–3.5-km-thick sequences of mostly subaerial high-Ti basalt flows, with low-Ti basalt and submarine pillow basalts in the lowest parts of the stratigraphy, that overlie Pennsylvanian–Permian (312–280 Ma) volcanic and sedimentary rocks. Subsidence of the entire plateau occurred during and after volcanism, based on late-stage interflow sedimentary lenses in the upper stratigraphic levels and the presence of hundreds of meters to >1000 m of overlying marine sedimentary rocks, predominantly limestone. The main factors that controlled the resulting volcanic architecture of the Wrangellia oceanic plateau include high effusion rates and the formation of extensive compound flow fields from low-viscosity, high-temperature tholeiitic basalts, sill-dominated feeder systems, limited repose time between flows (absence of weathering, erosion, sedimentation), submarine versus subaerial emplacement, and relative water depth (e.g., pillow basalt–volcaniclastic transition).

INTRODUCTION

Approximately 3% of the ocean floor is covered with oceanic plateaus or flood basalts, mostly in the western Pacific and Indian Oceans (Fig. 1; Coffin and Eldholm, 1994). These regions can rise thousand of meters above the ocean floor and rarely exhibit magnetic lineations like the surrounding seafloor (Ben-Avraham et al., 1981). Oceanic plateaus are produced from high volumetric output rates of magma generated by high-degree melting events that are distinct from melting beneath mid-ocean ridges. Most of what we know about the architecture of oceanic plateaus is based on obducted portions of oceanic plateaus, drilling of the volcanic sequences of extant oceanic plateaus, and geophysical studies of oceanic plateaus. Study of oceanic plateaus improves our understanding of the relationship between large igneous provinces (LIPs) and large melting events, mass extinctions, and continental growth (e.g., Kerr, 2003; Wignall, 2001).

Oceanic plateaus form near spreading ridges, extinct arcs, fragments of continental crust, and in intraplate settings. They form crustal emplacements 20–40 km thick with flood basalt sequences as much as 6 km thick, with individual plateaus extending as much as 2 million km2 in area (e.g., Ontong Java Plateau; Coffin and Eldholm, 1994). Current understanding of the construction of oceanic plateaus is based on the age, composition, and stratigraphy of the volcanic rock in conjunction with observations of interrelationships between sedimentation, erosion, and magmatism (Saunders et al., 2007). Stratigraphic and geochronological studies of an obducted oceanic plateau, where the base and top of the volcanic stratigraphy are exposed, as well as the underlying and overlying sediments, provide a means for evaluating the construction of an oceanic plateau (e.g., emplacement of flows, eruption environment, tectonic setting during formation, time scale of volcanism, paleoenvironment preceding and following eruption, uplift and subsidence history).

Wrangellia flood basalts represent one of the best-exposed accreted oceanic plateaus on Earth. Wrangellia is a rare example of an accreted oceanic plateau where parts of the entire volcanic stratigraphy are exposed, as well as the prevolcanic and postvolcanic stratigraphy. This oceanic plateau is exposed in numerous fault-bound blocks in a belt extending from Vancouver Island, British Columbia, to south-central Alaska. In this contribution we integrate new observations on the volcanic stratigraphy and prevolcanic and postvolcanic stratigraphy of Wrangellia with previously published data, including geochronology and biostratigraphy, to evaluate the construction and age of the Wrangellia oceanic plateau. This material is presented as descriptions of Wrangellia stratigraphy, compiled geologic maps, photographic databases, an interactive Google Earth file, and a review and compilation of previous research on Wrangellia. The maps, photographs, and archiving of information offer tools to visualize and fully explore the large body of information about Wrangellia, bringing together past and present research to provide an overview of the architecture of the Wrangellia oceanic plateau (Greene et al., 2008, 2009a, 2009b). This contribution highlights the prominent geologic features of the Wrangellia oceanic plateau, and compares them to other oceanic plateaus, to provide information that can be used to test and refine models on the genesis of oceanic plateaus.

WRANGELLIA FLOOD BASALTS: THE VOLCANIC STRATIGRAPHY OF AN OCEANIC PLATEAU

A large part of Wrangellia formed as an oceanic plateau, or transient LIP, that accreted to western North America. Wrangellia flood basalts are the defining unit of Wrangellia, in conjunction with underlying and overlying Triassic sediments with age-diagnostic fossils (Jones et al., 1977). The flood basalts are defined as the Karmutsen Formation on Vancouver Island and the Queen Charlotte Islands (Haida Gwaii), and the Nikolai Formation in southwest Yukon and south-central Alaska (Fig. 2). Smaller elements of Middle to Late Triassic basalt stratigraphy in southeast Alaska are also believed to correlate with the Wrangellia flood basalts (Plafker and Hudson, 1980).

Geographic Distribution and Aerial Extent of the Wrangellia Flood Basalts

The Wrangellia flood basalts in British Columbia, Yukon, and Alaska have been identified by geologic mapping and regional geophysical surveys. Exposures of Wrangellia flood basalts extend ∼2300 km from Vancouver Island to south-central Alaska (Fig. 2; Supplemental Google Earth File1). Exposures of the Karmutsen Formation cover ∼58% of Vancouver Island, British Columbia. In southern Alaska, elements of Wrangellia have a well-defined northern boundary along the Denali fault and extend southward to the outboard Peninsular terrane (Fig. 2). In southeast Alaska and southwest Yukon, Wrangellia is mostly limited to slivers of intensely dissected crustal fragments with minimal aerial extent (Fig. 2).

From recently compiled digital geologic maps, the aerial exposure of the Wrangellia flood basalts is ∼20 × 103 km2 on Vancouver Island, ∼800 km2 in southwest Yukon and southeast Alaska, and ∼2000 km2 across southern Alaska (Table 1). The original areal distribution was considerably greater, and these estimates of outcrop extent do not consider areas of flood basalt covered by younger strata and surficial deposits. The boundaries and crustal structure of Wrangellia in southern Alaska, as defined by recent magnetic and gravity surveys, indicate a distinct magnetic high interpreted as an expression of the presence of thick dense crust from Triassic mafic magmatism (Glen et al., 2007a, 2007b; Saltus et al., 2007). Wrangellia crust beneath Vancouver Island (∼25–30 km thick) has seismic properties corresponding to mafic plutonic rocks extending to depth that are underlain by a strongly reflective zone of high velocity and density (Clowes et al., 1995). This reflective zone was interpreted by Clowes et al. (1995) as a major shear zone where lower Wrangellia lithosphere was detached.

Geologic History of Wrangellia

In this paper Wrangellia (or the Wrangellia terrane) refers to fault-bound sections of the upper crust that contain diagnostic successions of Middle to Late Triassic flood basalts and Triassic sedimentary strata that overlie Paleozoic units (as originally defined by Jones et al., 1977). The areas of Wrangellia in Alaska and Yukon are referred to as Northern Wrangellia and areas in British Columbia are referred to as Southern Wrangellia. The Wrangellia composite terrane refers to three distinct terranes (Wrangellia, Alexander, Peninsular; Fig. 2) that share similar elements or have a linked geologic history (as defined by Plafker et al., 1989, 1994; Nokleberg et al., 1994; Plafker and Berg, 1994). The connections between the Wrangellia, Alexander, and Peninsular terranes are not well established, although a single Pennsylvanian pluton in Alaska is proposed to link the Proterozoic–Triassic Alexander terrane to Wrangellia by the late Pennsylvanian (Gardner et al., 1988). This paper specifically focuses on Wrangellia and does not examine the relationship of Wrangellia to the Alexander and Peninsular terranes.

Wrangellia has a geologic history spanning a large part of the Phanerozoic prior to its accretion with western North America. The geologic record beneath the flood basalts comprises Paleozoic oceanic arc and sedimentary sequences with a rich marine fossil assemblage. Paleontological studies indicate that Southern Wrangellia was located in cool-temperate waters at northern paleolatitudes (∼25°N) and not far from the North American continent during the Pennsylvanian–Early Permian (Katvala and Henderson, 2002). The geologic record is sparse to absent during the Middle Permian–Early Triassic throughout Wrangellia. A major, short-lived phase of tholeiitic flood volcanism occurred in the Middle to Late Triassic in submarine and subaerial environments. Paleomagnetic studies of Wrangellia flood basalts indicate eruption in equatorial latitudes (Irving and Yole, 1972; Hillhouse, 1977; Hillhouse and Gromme, 1984), and Late Triassic bivalves indicate an eastern Panthalassan position in the Late Triassic (Newton, 1983). A mantle plume origin for the Wrangellia flood basalts was proposed by Richards et al. (1991) and is supported by ongoing geochemical and petrological studies (Greene et al., 2008, 2009a, 2009b). Late Triassic–Early Jurassic arc magmatism is preserved as intrusions within and volcanic sequences overlying Wrangellia flood basalts throughout areas on Vancouver Island. Paleobiogeographic studies indicate that Wrangellia was located in the northeast Pacific Ocean during the Early Jurassic (Smith, 2006). Wrangellia probably accreted to western North America in the Late Jurassic–Early Cretaceous (e.g., Csejtey et al., 1982; McClelland et al., 1992; Nokleberg et al., 1994; Umhoefer and Blakey, 2006; Trop and Ridgway, 2007), and was caught up in the Cretaceous–Eocene thermal events of the Coast Mountains, as well as partly covered by younger sediments (Hollister and Andronicos, 2006).

SUMMARY OF THE STRATIGRAPHY OF WRANGELLIA

A full description of prominent features of Wrangellia stratigraphy and a summary of previous research are presented in the supplemental figure, map, methods, and data files. Table 2 provides a synthesis of Wrangellia stratigraphy. Supplementary Data Files 12 and 23 also contain a compilation of ∼500 references related to Wrangellia.

Northern Wrangellia

The accretion and northward tectonic migration of parts of Wrangellia, followed by the oroclinal bending of Alaska, has left Wrangellia flood basalts exposed in an arcuate belt extending ∼450 km across south-central Alaska (Fig. 3). Wrangellia stratigraphy forms most of the Wrangell Mountains in the eastern part of southern Alaska and extends westward in a wide belt immediately south of the Denali fault, along the southern flank of the eastern Alaska Range and in the northern Talkeetna Mountains (Fig. 3). The northwestern boundary of Wrangellia corresponds with a prominent steeply dipping structure (Talkeetna suture zone), which may represent the original suture between Wrangellia and the former continental margin (Glen et al., 2007a). Mesozoic and Cenozoic sedimentary basins along the inboard margin of Wrangellia, adjacent to and overlying the Talkeetna suture zone and Denali fault, record the uplift and collisional history of Wrangellia and the former continental margin, north of the Denali fault (Trop et al., 2002; Ridgway et al., 2002; Trop and Ridgway, 2007).

Some of the best exposures of Wrangellia flood basalt stratigraphy are preserved as part of an east-west–trending synform underlain by mafic and ultramafic plutonic rocks in the Amphitheater Mountains (Figs. 3 and 4). There are ∼3 km of subaerial basalt flows overlying <800 m of submarine stratigraphy in the Amphitheater Mountains. Wrangellia flood basalts also form two prominent northwest- to southeast-trending belts along the northeast and southwest margins of the Wrangell Mountains, between the Totschunda fault and Chitina thrust belt (Fig. 3). The Nikolai Formation in the Wrangell Mountains is estimated to be ∼3.5 km in total thickness and is composed almost entirely of subaerial flows (MacKevett, 1978). The Nikolai Formation consists of low-titanium basalts that form the lowest ∼400 m of volcanic stratigraphy in the Alaska Range, and the remainder of the volcanic stratigraphy in the Alaska Range and all of the sampled stratigraphy in the Wrangell Mountains are high-titanium basalt (Greene et al., 2008). A cumulative thickness of >3.5 km of marine sedimentary rocks, which range in age from Late Triassic to Late Jurassic, overlies the Nikolai basalts in the Wrangell Mountains. From the Nutzotin Mountains along the Alaska-Yukon border to southeast Alaska, Wrangellia forms a thin northwest- to southeast-trending belt in the southwest corner of Yukon (Fig. 2). The best exposures of Wrangellia in Yukon are in the Kluane Ranges, where the stratigraphy is similar in most aspects to stratigraphy in the Wrangell Mountains, and it has been described using the same nomenclature (Muller, 1967; Read and Monger, 1976). Triassic basaltic and sedimentary rocks with similarities to Wrangellia sequences in southern Alaska are exposed as elongate fault-bound slivers within a large and complex fault system in the Alexander terrane in several areas of southeast Alaska.

Southern Wrangellia

Northern and central Vancouver Island is underlain by Wrangellia, which forms the uppermost sheet of a thick sequence of northeast-dipping thrust sheets that constitute the upper crust of Vancouver Island (Fig. 5; Monger and Journeay, 1994; Yorath et al., 1999). Wrangellia lies in fault contact with the Pacific Rim terrane and Westcoast Crystalline Complex to the west, and is intruded by the predominantly Cretaceous Coast Plutonic Complex to the east (Wheeler and McFeely, 1991). The cumulative thickness of Wrangellia stratigraphy exposed on Vancouver Island is >10 km (Yorath et al., 1999). Two prominent northwest- to southeast-trending anticlinoria (Buttle Lake and Cowichan) are cored by Paleozoic rocks, which are not exposed on northern Vancouver Island (Fig. 5; Brandon et al., 1986; Yorath et al., 1999). Wrangellia also forms a large part of the southern Queen Charlotte Islands, where it is mostly unsampled.

The deepest stratigraphic levels of Wrangellia, which are exposed in the Buttle Lake and Cowichan anticlinoria, comprise the lower to middle Paleozoic Sicker Group and the upper Paleozoic Buttle Lake Group (Fig. 5). The combined total thickness of the Sicker and Buttle Lake Groups is estimated to be ∼5000 m (Massey, 1995; Yorath et al., 1999; Fig. 5).The basalt stratigraphy around Buttle Lake is proposed as the type section for the Karmutsen Formation, as this is where the most complete stratigraphic section is preserved (∼6 km thick; Figs. 5 and 6; Yorath et al., 1999). Recent mapping on northern Vancouver Island has delineated the three-part volcanic stratigraphy of pillowed lava sequences, hyaloclastite, and subaerial flows of the Karmutsen Formation (Fig. 6; Greene et al., 2009b; Nixon et al., 2008). Picritic pillow basalts occur near the top of the submarine basalt stratigraphy on northern Vancouver Island (Greene et al., 2006, 2009b). A photographic archive of field relationships in Wrangellia is available as an abbreviated version in the Supplemental Figure File4 and in full online at http://www.eos.ubc.ca/research/wrangellia/. Additional geologic maps for areas of Alaska, Yukon, and Vancouver Island can be found in the Supplemental Map File5.

AGE OF WRANGELLIA FLOOD BASALTS

Previous Geochronology for Wrangellia Flood Basalts and Related Plutonic Rocks

Samples of Wrangellia flood basalts and related plutonic rocks from British Columbia, Yukon, and Alaska were previously dated in eight separate studies; these results are summarized in Table 3 (all ages below are quoted with 2σ analytical uncertainty). A total of 15 ages (4 U-Pb, 9 40Ar/39Ar, and 2 K-Ar) from the literature include 4 basalts and 11 plutonic rocks. In Southern Wrangellia, 3 U-Pb ages from gabbroic rocks on southern Vancouver Island are available, including (1) a single concordant analysis of a multigrain baddeleyite fraction that yielded a 206Pb/238U age of 227.3 ± 2.6 Ma (Parrish and McNicoll, 1992), and (2) two unpublished 206Pb/238U baddeleyite ages of 226.8 ± 0.5 Ma (5 fractions) and 228.4 ± 2.5 (2 fractions) (Table 2; Sluggett, 2003). A single whole rock of Karmutsen basalt from Buttle Lake on Vancouver Island yielded a 40Ar/39Ar plateau age of 224.9 ± 13.2 Ma (Lassiter, 1995).

In Northern Wrangellia, the U-Pb results for zircon separated from a gabbro sill possibly related to the Nikolai basalts in southwest Yukon yielded an age of 232.2 ± 1.0 Ma (average 207Pb/206Pb age of 3 discordant [1.6%–2.4%] analyses from multigrain zircon fractions) (Table 3; Mortensen and Hulbert, 1991). K-Ar dating of biotite from peridotite in the Kluane mafic-ultramafic complex provided ages of 224 ± 8 Ma and 225 ± 7 Ma (Campbell, 1981). In Alaska, three samples of Wrangellia flood basalts from the Wrangell Mountains yielded whole rock 40Ar/39Ar plateau ages of 228.3 ± 5.2 Ma, 232.8 ± 11.5 Ma, and 232.4 ± 11.9 Ma (Lassiter, 1995). Five 40Ar/39Ar plateau and isochron ages of variable precision have been determined for hornblende and biotite separates from mafic and ultramafic plutonic rocks in the Amphitheater Mountains in the Alaska Range (Table 3). Three of these samples, with ages of 225.2 ± 6.5 Ma, 225.7 ± 2 Ma, and 228.3 ± 1.1 Ma, are from the Rainy Creek area, which is to the north across a major fault from typical volcanic stratigraphy of Wrangellia flood basalts (Bittenbender et al., 2007). The Rainy Creek area is a steeply dipping sequence of picritic tuff and volcaniclastic rocks, mafic and ultramafic intrusive rocks and dikes, and limestone that is distinct from the stratigraphy of the Nikolai Formation; these units may be older than the Wrangellia flood basalts (Bittenbender et al., 2003) or may be younger (Nokleberg et al., 1992). Two gabbros related to Wrangellia flood basalts in the Tangle Lakes area of the Amphitheater Mountains have reported 40Ar/39Ar ages of 230.4 ± 2.3 and 231.1 ± 11 Ma (Bittenbender et al., 2003; Schmidt and Rogers, 2007); however, analytical information is not available for these samples.

40Ar/39Ar Geochronological Results

Analytical methods for 40Ar/39Ar dating are described in Supplemental Methods File6. Petrographic textures and major and trace element chemistry for all of the geochronological samples are listed in Supplemental Data File 37. Age spectra are shown in Figure 7 and the analytical results are summarized in Table 4; the analytical data are available in Supplemental Data File 48. All ages indicated below are interpreted to represent cooling ages that correspond to the bulk closure temperature (Tcb) of the different minerals to Ar diffusion (∼200 °C plagioclase; ∼550 °C hornblende; ∼350 °C biotite; as summarized in Hodges, 2003). We processed 20 mineral separates, including 14 plagioclase separates prepared from the basalt and intrusive samples, 5 hornblende separates, and 1 biotite separate, from 19 samples from throughout Wrangellia for 40Ar/39Ar dating; 13 samples were Wrangellia flood basalts or intrusive equivalents and 6 were from younger, crosscutting intrusive rocks. Of the 13 Wrangellia flood basalts or intrusive samples, 9 were basalt flows and 4 were mafic sills or gabbroic rocks. One biotite separate was from an ultramafic plutonic rock from the Kluane Ranges, Yukon (sample 05-SIS-751). The five hornblende separates were all from younger dikes and intrusions that crosscut Wrangellia basalts and were selected as they provide minimum ages for eruption. The samples that are younger crosscutting intrusive units are clearly distinguishable from the Wrangellia flood basalts by their different textures and whole-rock chemistry.

Plagioclase separates from three basalt flows (two submarine and one subaerial flow) and one gabbro from Vancouver Island were analyzed. The incremental heating data of the three basalt flows form plateaus over 67%–96% of the 39Ar released with ages of 72.5 ± 1.2 Ma, 180.9 ± 3.1 Ma, and 161.1 ± 7.3 Ma (Fig. 7A; Table 4); inverse correlation diagrams (39Ar/40Ar versus 36Ar/40Ar) yield isochron ages that are concordant with the plateau ages. The results for a gabbro from the Alice–Nimpkish Lake area display a disturbed saddle-shaped age spectrum and may have been affected by excess 40Ar (e.g., Lanphere and Dalrymple, 1976; Harrison and McDougall, 1981). The crosscutting mafic dike (sample 4718A3) yields a plateau age of 193 ± 26 Ma (54% of the 39Ar released) and a broadly concordant isochron age of 186 ± 13 Ma (Fig. 7A; Table 4).

Plagioclase separates from five basalt flows and one mafic sill in Alaska, including two samples from the Wrangell Mountains and four samples from the Alaska Range, yield a total of three interpretable ages based on plateau ages (80%–99% of the 39Ar released): 191 ± 11 Ma, 160.7 ± 1.3 Ma, and 169.0 ± 2.4 Ma (Fig. 7B; Table 4). For sample 5810A4, the plateau age (137.3 ± 8.5 Ma) is considerably younger than the integrated age (160.57 ± 8.64 Ma), reflecting the younger ages of the lower temperature steps included in the plateau age. The individual step-heating results for sample 5719A5 are associated with relatively large errors, and two samples of basalt flows (5715A1, 5810A10) show strongly disturbed 40Ar/39Ar systematics (Fig. 7B).

The biotite separate from a peridotite in Yukon yields a plateau based on results from the higher temperature steps (60% of the 39Ar released), with an age of 227.5 ± 1.2 Ma and an inverse isochron age of 226.1 ± 3 Ma (Fig. 7C; Table 4). Plagioclase separates from two basalt flows from different levels of the volcanic stratigraphy of the Nikolai Formation in Yukon display strongly disturbed age spectra (Fig. 7C).

Hornblende separates from four younger intrusive samples from Alaska, including two samples from the Wrangell Mountains and two samples from the Alaska Range, yield plateau ages (54%–95% of the 39Ar released) of 29.7 ± 1.1 Ma, 123.13 ± 0.77 Ma, 148.80 ± 0.83 Ma, and 149.9 ± 0.93 Ma (Fig. 7D; Table 4); the plateau ages for these samples correspond with their respective isochron ages. The results for two hornblende analyses from an amphibolite dike in the Rainy Creek area of the Amphitheater Mountains show disturbed 40Ar/39Ar systematics (Fig. 7D).

Interpretation of New 40Ar/39Ar Geochronology

Among the 13 samples of Wrangellia flood basalt that were analyzed in this study, the results for eight samples satisfy the spectra and isochron criteria to be geologically interpretable 40Ar/39Ar ages (see Supplemental Data File 4 [see footnote 8]). The only sample inferred to have retained a magmatic age is the peridotite from Yukon for which the biotite separate yields a plateau age of 227.5 ± 1.2 Ma. This age, which corresponds to cooling of the ultramafic intrusion through the Tcb of biotite (∼350 °C) and is thus a minimum age of crystallization, is consistent with the range of previously published and reported ages from 227 to 233 Ma for emplacement of the Wrangellia flood basalts described above. The remaining seven ages from the Wrangellia basalts range from 191 to 73 Ma, indicating open-system behavior of the 40Ar/39Ar systematics (Table 4). The Karmutsen basalts on Vancouver Island underwent prehnite-pumpellyite facies metamorphic conditions (1.7 kbar, ∼300 °C) and higher metamorphic gradients where proximal to granitoid intrusions (Cho and Liou, 1987), and the mineralogy of Nikolai basalts in Alaska indicates a similar degree of metamorphism. Plagioclase, which has a low Tcb for Ar diffusion (<200 °C; Cohen, 2004), was analyzed from each of these samples, thus these samples likely degassed 40Ar during thermal resetting at some period after their emplacement.

Despite resetting, the 40Ar/39Ar ages of these basalts are geologically meaningful and provide temporal constraints on the geologic history of the Wrangellia terrane. Three of the reset 40Ar/39Ar ages (191, 181, and 161 Ma) of Karmutsen basalts from Vancouver Island are within the age range of Bonanza arc intrusions and volcanic sequences (197–167 Ma) that intrude and overlie the Karmtusen basalts on Vancouver Island (Table 4; Nixon and Orr, 2007). The three reset ages of Nikolai basalts from the Amphitheater Mountains (169, 161, and 161 Ma) are similar to ages of felsic plutonic rocks of the Early to Middle Jurassic Talkeetna arc, in close proximity (<30 km) to the south (168–150 Ma; Rioux et al., 2007). In the Amphitheater and Talkeetna Mountains, Schmidt et al. (2003a) reported that reset plagioclase 40Ar/39Ar ages (174–152 Ma) for eight Nikolai basalts and gabbroic rocks are similar to the crystallization ages for these felsic plutonic rocks, which are assigned to the Peninsular terrane. A K-Ar isochron age of 112 ± 11 Ma was determined from seven Nikolai basalts in the Wrangell Mountains, indicating resetting of K-Ar systematics during tectonism related to northward transport of Wrangellia (MacKevett, 1978; Plafker et al., 1989). The hornblende 40Ar/39Ar ages are coincident with regional magmatic events reported from other studies in the areas in which these samples were collected. Two Late Jurassic ages of 148.8 ± 0.83 and 149.9 ± 0.93 Ma from mafic dikes in the Wrangell Mountains correspond with ages of Late Jurassic plutons of the Chitina arc (140–160 Ma; most ages between 145 and 150 Ma), ages that are synchronous with a major regional orogeny related to subduction (Grantz et al., 1966; MacKevett, 1978; Hudson, 1983; Dodds and Campbell, 1988; Plafker et al., 1989; Roeske et al., 2003).

In summary, the ages of Wrangellia flood basalts and intrusive rocks with associated analytical errors that are <±10 Myr (n = 9) are in the range from 222 to 233 Ma (Fig. 8 inset; Tables 3 and 4). The three U-Pb ages from mafic sills from Vancouver Island are within error of the four 40Ar/39Ar ages of basalts from Alaska that meet this precision criterion, as well as the 40Ar/39Ar plateau age of 227.5 ± 1.2 Ma given by biotite from a peridotite in this study. The slightly older age of 232.2 ± 1 Ma for a gabbro from Yukon (Mortensen and Hulbert, 1991) is the only age that is outside the range for the Vancouver Island samples, thus this gabbroic sill may represent an earlier phase of magmatism. The relatively narrow range of ages for Wrangellia flood basalts and associated intrusive rocks indicates that the duration of the majority of the magmatic activity was likely <5 Myr, occurring between ca. 230 and 225 Ma.

Paleontological Age Constraints

Fossils in sedimentary strata directly underlying and overlying the Wrangellia flood basalts also provide constraints on the age and duration of volcanism. MacKevett et al. (1964) reported that shale directly beneath Nikolai basalts on Golden Horn Peak in the Wrangell Mountains contained abundant Middle Triassic index fossils identified as Daonella frami Kittl. In 1971, a mineral exploration party discovered a sedimentary layer (<2 m thick) of fissile black shale between sills of Karmutsen basalts on Mount Schoen on Vancouver Island that contained imprints of Daonella tyrolensis (Carlisle, 1972). As part of the present study, samples of Daonella were collected from Golden Horn Peak in Alaska and on Mount Schoen on Vancouver Island. The Daonella specimens from both localities appear to have similar forms and are closely related to Daonella frami (C.A. McRoberts, 2006, personal commun.). The Daonella specimens appear older than Upper Ladinian forms, and are likely of middle Ladinian age (Poseidon Zone, ca. 235–232 Ma; C.A. McRoberts, 2006, personal commun.). Recovery of the conodont Neospathodus from these Daonella beds on Vancouver Island agrees with this age.

Magnetostratigraphic Age Constraints

Magnetic reversals were common throughout the Late Triassic; however, only a single magnetic polarity reversal is preserved in the Wrangellia flood basalts (Hillhouse, 1977). Magnetostratigraphy in the Newark rift basin of eastern North America has provided a high-resolution geomagnetic polarity time scale that reflects nearly idealized geomagnetic field reversal behavior in the Late Triassic (Kent and Olsen, 1999). A total of 59 polarity reversals occurred between 233 and 202 Ma; the mean duration of polarity intervals is estimated to be 0.53 Myr; the longest polarity interval is ∼2 Myr and the shortest polarity interval is ∼0.02 Myr (Kent and Olsen, 1999). All of the recorded polarity intervals between 225 and 230 Ma were at least ∼0.5 Myr in duration. Based on this, the Wrangellia basalts likely erupted within a period of only 2 Myr and possibly in a considerably shorter time span.

DISCUSSION

Comparison of Northern and Southern Wrangellia

The stratigraphy and age of different areas of the Wrangellia oceanic plateau provide constraints on the construction of the volcanic stratigraphy, the paleoenvironments existing at the time of eruption, and the duration of volcanism. To aid in the following discussion, a summary of observed and previously reported field relationships of Wrangellia flood basalts, and the prevolcanic and postvolcanic rock record, is presented in Table 5, and a compilation of ages and biostratigraphy for Paleozoic–Triassic rocks of Wrangellia is presented in Figure 8 and Supplemental Data File 59. We establish the similarities and differences in the stratigraphy of Northern and Southern Wrangellia and use them in subsequent sections to interpret the eruption environments and accumulation subsidence histories of different parts of this vast accreted oceanic plateau.

The basement of Wrangellia has different age strata in Alaska and Yukon than on Vancouver Island (Fig. 8, Supplemental Data Files 610 and 711). The basement of Wrangellia was originally defined as a Pennsylvanian–Permian volcanic arc sequence that may have been deposited on oceanic crust (Jones et al., 1977), and this has been maintained by numerous authors (Smith and MacKevett, 1970; MacKevett, 1978; Coney et al., 1980; Monger et al., 1982; Saleeby, 1983; Beard and Barker, 1989). However, whereas Pennsylvanian–Permian volcanic arcs are preserved in Alaska (Smith and MacKevett, 1970; MacKevett, 1978; Beard and Barker, 1989), older Paleozoic rocks make up much of Vancouver Island (Muller, 1977; Brandon et al., 1986), and there is no significant arc volcanism in the Pennsylvanian–Permian of Vancouver Island (Fig. 8; Yole, 1969; Massey and Friday, 1988; Yorath et al., 1999). The Paleozoic volcanic sequences on Vancouver Island are considerably older (ca. 380–355 Ma) than dated volcanic sequences in Alaska (ca. 312–280 Ma, mostly from the Wrangell Mountains; see Fig. 8 and references therein). Paleozoic limestone beneath Wrangellia basalts in Alaska contains Early Permian bryozoans, brachiopods, foraminifera, and corals (Smith and MacKevett, 1970). On Vancouver Island, conodonts in the Buttle Lake Group indicate Mississippian–Permian ages (Orchard fideBrandon et al., 1986; Orchard, 1986; Fig. 8; Supplemental Data File 5 [see footnote 9]).

The volcanic stratigraphy on Vancouver Island consists of a tripartite succession of submarine flows (50%–60% of total thickness), volcaniclastics, and subaerial flows, whereas volcanic stratigraphy in Alaska and Yukon is predominantly subaerial flows (>90%; Table 5; Greene et al., 2008, 2009a, 2009b). In areas of central Vancouver Island, the stratigraphically lowest pillow basalts were emplaced on unconsolidated fine-grained sediments, and mafic sills intrude marine strata with Daonella beds. In other areas of central and southern Vancouver Island, basal basalt flows overlie Permian limestone and sills intrude this limestone (Fig. 6). The submarine section on Vancouver Island (∼3000 m) is substantially thicker than in the Alaska Range (<800 m). In the Alaska Range, the stratigraphically lowest submarine flows were emplaced on nonfossiliferous black shale. In the Wrangell Mountains, a basal flow conglomerate directly overlies shale with Daonella beds and contains erosional remnants from the underlying Paleozoic units. Laterally discontinuous zones of conglomerate along the base of the Nikolai Formation in Yukon have been interpreted as syntectonic deposits and flows related to the formation of grabens (Israel et al., 2006). The flood basalts in Northern Wrangellia are low-Ti basalts (<400 m) overlain by high-Ti basalts, whereas almost all of the basalt stratigraphy in Southern Wrangellia is high-Ti basalt and includes an area 30 km in diameter with abundant picritic pillow basalts (Greene et al., 2008, 2009a, 2009b).

The overlying limestone and interflow sedimentary lenses in Northern and Southern Wrangellia are lithologically similar and have a similar range of fossil ages (Table 5; Fig. 8). In Southern Wrangellia, interflow sedimentary lenses are common in the upper parts of the Karmutsen Formation and the overlying limestone contains late Carnian–Norian fossils. Within upper Nikolai Formation stratigraphy in Northern Wrangellia, interflow sedimentary lenses occur in southwest Yukon and the Clearwater Mountains, and Nikolai basalts are overlain by limestone with age-diagnostic late Carnian–early Norian fossils. Sedimentary strata extend up through the Triassic-Jurassic boundary in Northern and Southern Wrangellia. The interflow limestone lenses and overlying sedimentary rocks are important time and depth markers for estimating the subsidence and mantle thermal history of Wrangellia (see following discussion). Based on the similar age, high effusion rate, tectonic history, and trace element and isotopic geochemistry of Wrangellia basalts (Greene et al., 2008, 2009a, 2009b), the flood basalts of Northern and Southern Wrangellia are interpreted to have formed from the same plume-related magmatic event.

Eruption Environment for Wrangellia Flood Basalts

Northern Wrangellia

The Nikolai basalts in Alaska were emplaced as effusive eruptions in a shallow marine and subaerial environment (Fig. 9). Sedimentary rocks directly beneath the flood basalts in the Alaska Range are mostly siliceous argillites, carbonaceous black shales, and mudstones; the upper part of the 200–250 m sequence has a higher proportion of black carbonaceous shale and carbonate (Blodgett, 2002). In this area, Blodgett (2002) interpreted the total absence of fossils and biogenic sedimentary structures (trace fossils), and the even, parallel laminations (indicating lack of bioturbation), as indicative of deposition in a starved, anoxic shallow submarine environment. The higher proportion of black carbonaceous shale and calcareous component for sediments higher in the sequence may indicate a shallower depositional environment for the younger sediments than for the older sediments, possibly due to uplift prior to eruption (Blodgett, 2002). The pillow basalts in the Alaska Range (∼500 m) are highly vesicular, consistent with eruption in shallow water (Jones, 1969; Kokelaar, 1986). Volcaniclastic flows intercalated with pillow basalt also indicate eruption in shallow water; this transition typically occurs in <200 m water depth for tholeiitic magmas (Kokelaar, 1986). The subaerial flows (>3000 m) were emplaced as inflated compound pahoehoe flow fields during prolonged, episodic eruptions similar to those in most continental flood basalts (e.g., Self et al., 1997). The flows erupted from a limited number of eruption sites that are rarely observed, except for a large sill-dominated eruptive center in the Amphitheater Mountains.

In the Wrangell Mountains and Yukon, the basal flow conglomerate and pillow breccia (<100 m thick) erupted in shallow water (<200 m; Fig. 9). Rounded clasts in the basal flow conglomerate indicate an area of relief near sea level in the Wrangell Mountains. Above the basal flow unit in the Wrangell Mountains are ∼3500 m of subaerial sheet flows that lack features of submarine emplacement. The proportion of amygdules in the massive flows is variable, but generally high, similar to many continental flood basalts. In contrast, submarine sheet flows exposed in accreted portions of the Ontong Java Plateau in the Solomon Islands of the western Pacific Ocean rarely have vesicles (Petterson, 2004); pillowed and massive flows are preserved throughout the stratigraphy of the Ontong Java Plateau.

Southern Wrangellia

The tripartite Karmutsen stratigraphy on Vancouver Island formed as an emergent basalt sequence that passed through a deeper water, shallow water, and subaerial stage, similar in ways to those described in the formation of emergent seamounts (Fig. 9; Schmidt and Schmicke, 2000) and Hawaiian volcanoes (Garcia et al., 2007). Sills obscure relationships at the sediment-basalt interface at the base of the Karmutsen Formation, and there may have been sediments on top that are no longer preserved, but most of the fine-grained strata underlying the Karmutsen were deposited below storm wave base. Carlisle (1972) reported that pre-Karmutsen sediments show a progressive change from coarse bioclastic limestone to laminated and silicified shale, indicating a transition from an organic-rich, shallow-water environment to a starved, pelagic deeper water depositional environment prior to initiation of volcanism. Daonella fossils in fine black shale from near the top of this unit imply dysoxic bottom waters typical of mud dwellers that float on soupy sediments (Schatz, 2005).

The deeper water stage (>200 m water depth) of the Karmutsen Formation was dominated by effusive activity that formed pillowed and massive flows (Greene et al., 2009). The pillowed flows are interconnected tubes and lobes that contain large-diameter pillows (>2 m) and have low abundances of amygdules. The massive flows may represent some of the master tubes for delivery to distal parts of flow fields or locally increased effusive rates, due to topography, as evinced by concave basal contacts. The basalts increase in vesicularity, as does the proportion of volcaniclastics, upward in the submarine stratigraphy (Nixon et al., 2008).

The shallow-water stage of the Karmutsen preserves an increasing proportion of volcaniclastic units (pillow breccia and hyaloclastite) conformably overlying mostly close-packed pillows. The pillow breccias are commonly associated with pillowed flows and contain aquagene tuff (Carlisle, 1963) or redeposited hyaloclastite. The transition from close-packed pillowed flows to pillow breccia and hyaloclastite probably occurred in <500 m water depth; however, in certain areas on northern Vancouver Island the volcaniclastic unit is >1500 m thick (Nixon et al., 2008). Sedimentary structures (e.g., graded bedding, fluidization structures) are present locally and indicate resedimentation processes. Pyroclastic deposits containing lapilli tuff and volcanic bombs do not appear to be common (Carlisle, 1963). The volcaniclastic rocks likely formed primarily via cooling-contraction granulation, magma-water-steam interaction, autobrecciation, and mass wasting, rather than pyroclastic eruption.

The emergent subaerial stage is marked by the relative absence of volcaniclastic and pillowed flow units and dominance of massive amygdaloidal sheet flows. The sheet flows were emplaced as inflated compound pahoehoe flow fields atop an enormous oceanic plateau. There are isolated sections of submarine flood basalts (<200 m thick) within the uppermost subaerial Karmutsen stratigraphy (Surdam, 1967; Carlisle and Suzuki, 1974); these units form a volumetrically minor component of the subaerial stratigraphy. The intra-Karmtusen sedimentary lenses formed in isolated, low-lying areas in a predominantly subaerially exposed plateau (Carlisle and Suzuki, 1974).

There are similarities between the stratigraphy of Karmutsen basalts and the stratigraphy of the Hawaii Scientific Drilling Project core (HSDP2; Garcia et al., 2007). In both the Karmutsen basalts and HSDP2, predominantly pillowed flows in the lower parts of the submarine stratigraphy give way to increasing proportions of volcaniclastic units upsection, below the submarine-subaerial transition. Vesiculated pillows of the Karmutsen Formation on Northern Vancouver Island are more common near the top of the pillow unit and in the hyaloclastite unit. HSDP2 drill core shows an increase in vesicularity with decreasing depth in the submarine lava flows (Garcia et al., 2007). Intrusions are more common in the lower parts of the stratigraphy within both submarine sections and, although more difficult to identify, they appear to be less common within the subaerial sections. In contrast, flow morphologies and thicknesses are distinct in Karmutsen stratigraphy and Hawaiian volcanoes.

Accumulation and Subsidence of the Wrangellia Basalts

The geology, age, and biostratigraphy of Wrangellia can be used to estimate the rate of accumulation of Wrangellia basalts and the subsidence of the Wrangellia oceanic plateau. Carlisle and Suzuki (1974) originally estimated an accumulation rate for the Karmutsen basalts of 0.17–0.27 cm/yr over 2.5–3.5 Myr. This yielded a total erupted volume of basalt of 3.7–4.0 × 105 km3 (they assumed an area 400 km × 150 km, roughly the size of Vancouver Island, or 60 × 103 km2, and a stratigraphic thickness of 6 km) and a volumetric output rate of 0.10–0.16 km3/yr (Carlisle and Suzuki, 1974). The area of exposure of Karmutsen basalts in this study was calculated using digital geology maps for exposures of the Karmutsen Formation on Vancouver Island and the Queen Charlotte Islands, and represents a minimum estimate of surface area. The estimated total erupted volume and volumetric output rate were calculated using a stratigraphic thickness of 6 km and duration of volcanism of 5 Myr. Our estimate of minimum volcanic output rate is ∼0.03 km3/yr and that of minimum total erupted volume of Karmutsen basalts is 1.4–1.5 × 105 km3. Even using this very conservative estimate for the area of exposure and age, the volumetric output rate is comparable to recent estimates of long-term volumetric eruption rates for ocean islands such as Iceland (0.02–0.04 km3/yr) and Hawaii (0.02–0.08 km3/yr) (White et al., 2006).

Subsidence of the Karmutsen basalts during volcanism was recorded by the deposition of interflow sedimentary lenses between the upper flows during the waning stages of volcanism as low-lying areas of the plateau were submerged. The occurrence of interflow lenses indicates that by the end of Karmutsen flood volcanism most of the top of the basalt plateau had subsided and submerged below sea level. This implies that over the duration of subaerial volcanism, the rate of accumulation of basalt flows was comparable to the rate of subsidence. Carbonate deposition was preserved during the waning stage of volcanism, and there was no significant break after volcanism ceased. There are few signs of erosion between the Quatsino limestone and Karmutsen Formation (in places only a <25 cm thick siltstone and/or sandstone layer is preserved), and interflow lenses have fossils identical to those of the lower part of the Quatsino limestone.

Postvolcanic subsidence of the Wrangellia oceanic plateau is recorded by hundreds of meters to >1000 m of Late Triassic marine sedimentary rocks overlying the basalt stratigraphy. The Quatsino limestone was deposited on top of the plateau as it began to submerge beneath sea level, and while volcanism waned. Deposition continued as the plateau became fully submerged and the sea transgressed over the entire plateau. Initially, intertidal to supratidal limestones were deposited in shallow-water, high-energy areas, some of the limestones reflecting quieter, subtidal conditions (Carlisle and Suzuki, 1974). Upsection, the Quatsino Formation reflects a slightly deeper water depositional environment and abruptly grades into the overlying Parson Bay Formation (Carlisle and Suzuki, 1974; Nixon and Orr, 2007).

In Alaska and Yukon, the carbonate rocks overlying the Nikolai basalts (>1000 m; Chitistone and Nizina limestones) preserve a record of the gradual submergence of the extensive Nikolai basalt platform. There are only rare occurrences of thin (<0.5 m) weathered zones or discontinuous, intervening clastic deposits at the top of ∼3500 m of subaerial basalt flows in the Wrangell Mountains (Fig. 3; Armstrong et al., 1969). The absence of significant erosion or deposition of clastic sediment and the age of the Chitistone limestone (Fig. 8) indicate only a brief interval of nondeposition between the end of volcanism and carbonate deposition. Following volcanism, several cycles of shaley to argillaceous limestone were deposited in a high-energy, intertidal to supratidal (sahbka) environment, similar to the modern Persian Gulf, and form the lowest 100 m of the Chitistone limestone (Armstrong et al., 1969). Limey mudstone and wackestone with abundant disintegrated shelly material (∼300 m thick) indicate a gradual transition to low-energy shallow-water deposition with intermittent high-energy shoaling deposition. The upper part of the Chitistone and overlying Nizina limestones reflect deeper water deposition on a drowned carbonate platform (Armstrong et al., 1969). Gray to black shale and chert of the overlying McCarthy Formation represent submergence of the carbonate platform below the carbonate compensation depth (Armstrong et al., 1969). A thin tuff bed in the lower part of the McCarthy Formation has been dated at 209.9 ± 0.07 Ma (J. Trop, 2006, personal commun.), and the Triassic-Jurassic boundary is preserved in the upper McCarthy Formation (Fig. 8). Between the end of basaltic volcanism and the end of the Triassic (∼25 Myr), ∼2000 m of shallow- to deep-water marine sediments accumulated on top of the Nikolai basalts. Neglecting sediment compaction, this indicates a minimum subsidence rate of ∼80 m/Myr. This subsidence decreased substantially, to <20 m/Myr, in the Early Jurassic (Saltus et al., 2007). Initial subsidence rates of 50 m/Myr were also estimated by Richards et al. (1991) based on the thickness and age of carbonate rocks overlying the Nikolai Formation.

Several oceanic plateaus and ocean islands worldwide preserve evidence of rapid subsidence after their formation (Detrick et al., 1977), and mantle plume models predict subsidence following the formation of oceanic plateaus (e.g., Campbell and Griffiths, 1990). Subsidence may result from dispersion of the mantle buoyancy anomaly above a plume; decay of the thermal anomaly and cooling and contraction of the lithosphere; removal of magma from the plume source, causing deflation of the plume head; and/or depression of the surface from loading of volcanic and plutonic material on the lithosphere (e.g., Detrick et al., 1977; Campbell and Griffiths, 1990). Hawaiian volcanoes undergo rapid subsidence during their growth due to the response from loading of intrusive and extrusive magmas on the lithosphere (e.g., Moore and Clague, 1992). Pacific Cretaceous plateaus, such as Ontong Java, Manihiki, and Shatsky Rise, underwent significantly less subsidence than predicted by current models, with post–flood basalt subsidence comparable to that of normal ocean seafloor (Ito and Clift, 1998; Roberge et al., 2005). Drilling of the Kerguelen Plateau indicates that subaerial flows, perhaps originally 1–2 km above sea level, subsided below sea level and paleoenvironments of the overlying sediments changed from intertidal to pelagic (Coffin, 1992; Frey et al., 2000; Wallace, 2002). Subsidence estimates for the Northern Kerguelen Plateau (Ocean Drilling Program Leg 183, Site 1140) indicate that ∼1700 m of subsidence occurred since eruption of the basalts at 34 Ma (∼50 m/Myr; Wallace, 2002). The subsidence history of Wrangellia may be more fully reconstructed by future studies that estimate the depth of sedimentation and age from microfossils collected from sediments overlying the Karmutsen and Nikolai Formations. The well-preserved and accessible sedimentary strata overlying the accreted Wrangellia plateau has the potential to affect our understanding of the thermal history of oceanic plateaus from plume-controlled magmatic events that are not associated with mid-ocean ridge processes, such as several oceanic plateaus in the Pacific (e.g., Shatsky Rise; Sager, 2005).

IMPLICATIONS FOR THE ARCHITECTURE OF OCEANIC PLATEAUS

The architecture of the erupted products of oceanic plateaus provides one of the most direct ways of evaluating the largest magmatic events on Earth. Construction of oceanic plateaus has implications for the age and duration of magmatism, the association of plume- or plate-controlled processes, uplift and subsidence history related to large magmatic events, mantle melting processes and thermal evolution, and the effect of eruptions on the environment (e.g., Ito and Clift, 1998; Wignall, 2001; Sager, 2005; Saunders et al., 2007). The Wrangellia plateau provides an unmatched view of the preeruption, syneruption, and posteruption phases of the growth of plume-related oceanic plateaus on Earth. Here we compare Wrangellia to other oceanic plateaus, and to continental flood basalts and modern ocean island hotspots, to highlight the conditions and processes that control the formation of oceanic plateaus.

Based on the stratigraphic and temporal constraints outlined in this study, and the petrology and geochemistry of the Wrangellia basalts from Greene et al. (2008, 2009a, 2009b), the main factors that controlled the construction of the Wrangellia oceanic plateau include high effusion rates and the formation of extensive compound flow fields from low-viscosity, high-temperature tholeiitic basalts, sill-dominated feeder systems, low volatile content, protracted eruption (perhaps >10 yr for individual flow fields), limited repose time between flows (absence of weathering, erosion, sedimentation), submarine versus subaerial emplacement, and relative water depth (e.g., pillow basalt–volcaniclastic transition). Combined, these factors contributed to the production of 4–6 km of basalt stratigraphy with limited magnetic reversals, a lack of fossil remains in associated sediments, and minimal evidence of eruptive explosivity. In Table 6, the major features of the Wrangellia, Ontong Java, and Caribbean oceanic plateaus are compared. From this comparison, several commonalities are evident, including high effusive rates, sill-dominated feeder systems, largely submarine eruptions (but varying amounts of subaerial eruptions), presence of high-MgO lavas, absence of alkalic lavas during the main eruptive phase, and relatively short periods of repose. Short repose intervals are indicated by the absence of weathering, erosion, sedimentation, and magnetic reversals, in addition to age constraints.

Distinct phases of volcanism, characterized by changes in the chemistry of the basalts, occur during the construction of some oceanic plateaus and many continental flood basalts. In the latter, three main phases of volcanism include initial low-volume transitional to alkalic eruptions, a main phase of large volume of tholeiitic flows that produced large flow fields from fissure eruptions, and a waning phase of decreased-volume eruptions from widely distributed, localized volcanic centers with higher silica and/or alkali lavas and explosivity (Jerram and Widdowson, 2005). In contrast, oceanic plateaus appear to largely form during a single, main tholeiitic phase, such as observed for Wrangellia, although both the Ontong Java and Caribbean Plateaus reveal distinct phases of late-stage volcanism (Table 6). The evolution of the Ontong Java, Manihiki, and Hikurangi plateaus (Fig. 1), which are perhaps dispersed parts of a single giant oceanic plateau (Taylor, 2006), show similarities with a main tholeiitic stage (ca. 126–116 Ma) followed by late-stage volcanism lasting more than 30 Myr (Davy et al., 2008). The late-stage volcanism on Hikurangi and Manihiki includes seamounts and volcanic ridges that are mostly alkalic in composition, whereas late-stage volcanism on Ontong Java is mostly tholeiitic (although alkalic rocks are present; Hoernle et al., 2008). To date, there is no evidence for a clear gap in volcanism in the formation of the Wrangellia oceanic plateau, although additional high-precision geochronology is required to better evaluate the onset and termination of magmatism. There is also no evidence for, or trend toward, alkalic compositions from basaltic sections sampled on Vancouver Island and in Yukon and Alaska (Greene et al., 2008, 2009a, 2009b), even for samples taken from the uppermost stratigraphic sections just below covering sedimentary rocks. Perhaps parts of the Wrangellia plateau are missing or are not exposed, or Wrangellia may represent a purely tholeiitic large igneous province formed during a single-stage (<5 Myr) interval of time.

The high eruption rates, relatively short periods of repose, and basaltic lava compositions of the Wrangellia and other oceanic plateaus are comparable to those of continental flood basalts, which have estimated eruption rates orders of magnitude greater than present-day hotspots such as Hawaii (e.g., Thordarson and Self, 1996; White et al., 2006). The volumetric output of flow fields in continental flood basalts, which is likely comparable for oceanic plateaus, is on the order of hundreds to thousands of cubic meters per second, and individual flow fields persist for years to perhaps more than a decade (although volumes of individual eruption events can be highly variable; Jerram, 2002). For example, the flow rate estimated for the Roza flow field in the Columbia River Basalt Group (∼4000 m3/s) is orders of magnitude higher than the flow of lava during the past 26 yr on Kilauea Volcano on the island of Hawaii (∼2 m3/s; <6 m3/s; Heliker and Mattox, 2003). The higher sustained flow rates for the tholeiitic basaltic lavas that dominated oceanic plateaus produce compound flow fields as opposed to shield-like features that form from the lower volume eruptions on Hawaiian, and other hotspot, volcanoes. The morphology of flows in oceanic plateaus is also distinct from flows on Hawaiian volcanoes. Oceanic plateaus consist of thick (10–50 m), tabular and laterally extensive flows extending tens of kilometers, whereas Hawaiian volcanoes form thin (<5 m) lobes with large local variation and features like volcanic cones, craters, and channels. Despite comparably high estimates of eruption rates for modern eruptions at ocean island hotpots over brief periods (e.g., Laki, 1983–1984; Self et al., 1997), there is no modern analogue for the formation of submarine sections of oceanic plateaus.

CONCLUSIONS

The volcanic stratigraphy of the obducted Wrangellia flood basalts records the construction of a major oceanic plateau and provides a rare view of the preeruption, syneruption, and posteruption phases and paleoenvironments of an oceanic plateau that originated from a plume-related magmatic event not associated with mid-ocean ridge–controlled processes. The Wrangellia basalts are now exposed over ∼27 × 103 km2 in Alaska, Yukon, and British Columbia (Vancouver Island and Queen Charlotte Islands) along the western edge of North America, although the original extent was significantly greater. Geochronologic, paleontological, and magnetostratigraphic constraints indicate that volcanism in Northern and Southern Wrangellia occurred ca. 230–225 Ma.

In Northern Wrangellia, the basalts in Alaska (∼3.5 km thick) and Yukon are bounded by Middle to Late Triassic marine sediments and unconformably overlie Pennsylvanian and Permian marine sediments and volcanic arc sequences (ca. 312–280 Ma). The earliest flows were emplaced in a shallow-marine environment, but the main phase of volcanism consisted of compound pahoehoe flow fields that form a tabular, shingled architecture with high-Ti basalts overlying low-Ti basalts. Grabens formed along the base of Wrangellia basalts in Yukon.

In Southern Wrangellia, basalts on Vancouver Island (∼6 km thick) are bounded by Middle to Late Triassic marine sediments and unconformably overlie a basement of Devonian–Mississippian arc rocks and Mississippian–Permian marine sedimentary strata. Early growth of the volcanic stratigraphy on Vancouver Island was dominated by extrusion of pillow lavas and intrusion of sills into sedimentary strata. The plateau grew from the ocean floor and accumulated >3000 m of submarine flows, which were overlain by 400–1500 m of hyaloclastite and minor pillow basalt before the plateau breached sea level. The hyaloclastite formed primarily by quench fragmentation of effusive flows under low hydrostatic pressure. The plateau then grew above sea level as >1500 m of subaerial flows were emplaced. The plateau subsided during its construction and intervolcanic sedimentary lenses formed in shallow water in local areas as eruptions waned. After volcanism ceased, the plateau continued to subside for more than 25 Myr and was overlain by hundreds of meters to >1000 m of limestone and siliciclastic deposits.

Wrangellia holds keys to understanding some of the important questions about the development of oceanic plateaus, the time scale of eruptions, mantle properties during large-scale melting events, and possible effects on the environment.

This manuscript was influenced by discussions with Travis Hudson and material that he provided. We are grateful to Jeff Trop for his advice. Jeanine Schmidt and Peter Bittenbender were very helpful with data and ideas about Alaskan geology. Don Carlisle thoughtfully provided maps and notes that helped with field work on Vancouver Island. Andrew Caruthers and Chris Ruttan helped with field work on Vancouver Island. David Brew was very helpful with information about southeast Alaska. Journal reviews by Lincoln Hollister and Colin Shaw were especially useful in helping us focus the information and ideas presented in this paper. Funding was provided by research grants from the British Columbia Geological Survey and Yukon Geological Survey, the Rocks to Riches Program administered by the British Columbia and Yukon Chamber of Mines, and Natural Sciences and Engineering Research Council Discovery Grants to James Scoates and Dominique Weis. A. Greene was supported by a University Graduate Fellowship at the University of British Columbia.

1Supplemental Google Earth File. ALL Wrangellia files. This .kmz file consists of georeferenced information that is designed to be viewed with the satellite imagery in Google Earth. The Google Earth application is available for free download at http://earth.google.com/. Information about Google Earth can be found at this web address. Please use the most current version of Google Earth for viewing these files (version 5.0 as of June 2009). The Supplemental Google Earth File contains ten separate folders: (1) Mapped Wrangellia flood basalts. This is a red transparent layer that shows the distribution of the Wrangellia flood basalts in Alaska, Yukon, and British Columbia. The map was derived from data in Wilson et al. (1998, 2005), Israel (2004), Massey et al. (2005a, b), and D. Brew (2007, written commun.). (2) Major faults in Alaska and Yukon. (3) Major faults in southwest BC. These folders show the location of faults in parts of Alaska, Yukon, and BC. These files were filtered from original files found at the following locations: Alaska from http://www.asgdc.state.ak.us/; Yukon from in Israel (2004); BC faults from Massey et al. (2005a, b). (4) Alaska sample locations. (5) Yukon sample locations. (6) Vancouver Island sample locations. These files show the locations, sample numbers, and flow type of samples of Wrangellia flood basalts collected during field work. (7) Alaska Range photograph locations. (8) Wrangell Mountains photograph locations. (9) Yukon photograph locations. (10) Vancouver Island photograph locations. These four folders contain small versions of georeferenced photographs. Multiple photographs are referenced to a single coordinate. Therefore, in order to view all of the photographs from a single coordinate it is necessary to open the subfolders for each region in the My Places menu so individual photographs can be selected and viewed. Zooming in also helps to distinguish photograph locations. These photographs and others can be viewed in higher resolution online at the following link: http://www.eos.ubc.ca/research/wrangellia/. If you are viewing the PDF of this paper or reading it offline, please visit http://dx.doi.org/10.1130/GES00212.S1 or the full-text article on www.gsapubs.org to view the Supplemental Google Earth File.
2Supplemental Data File 1. Endnote database for Wrangellia. This Endnote library is a compilation of approximately 500 references mostly related to research of Wrangellia or parts of western North America. This is a zipped file containing an .enl database file and a folder named SD1_Endnote_database_Wrangellia.Data, which is needed to open the .enl file. The .enl file was created with Endnote X1.0.1 (Bld 2682) and should be compatible with older versions of Endnote. If you are viewing the PDF of this paper or reading it offline, please visit http://dx.doi.org/10.1130/GES00212.S2 or the full-text article on www.gsapubs.org to view Supplemental Data File 1.
3Supplemental Data File 2. Reference list for Wrangellia. This Microsoft Word file is an exported version of the Endnote file in Supplemental Data File 1. If you are viewing the PDF of this paper or reading it offline, please visit http://dx.doi.org/10.1130/GES00212.S3 or the full-text article on www.gsapubs.org to view Supplemental Data File 2.
4Supplemental Figure File. This PDF file contains nine figures with captions. These figures accompany Supplemental Data File 7 and illustrate some of the exceptional exposures of the volcanic stratigraphy of the Wrangellia plateau in Alaska, Yukon, and British Columbia. Additional georeferenced photographs are part of the Supplemental Google Earth File and also can be found at the following website dedicated to research on Wrangellia: http://www.eos.ubc.ca/research/wrangellia/. If you are viewing the PDF of this paper or reading it offline, please visit http://dx.doi.org/10.1130/GES00212.S4 or the full-text article on www.gsapubs.org to view the Supplemental Figure File.
5Supplemental Map File. This file contains four maps of different parts of Wrangellia in a single pdf file with references. Two maps are simplified regional geologic maps of southwest Yukon and southeast Alaska showing the distribution of Triassic flood basalts. Two maps are detailed geologic maps of areas in the Amphitheater Mountains in the Alaska Range and on northern Vancouver Island. If you are viewing the PDF of this paper or reading it offline, please visit http://dx.doi.org/10.1130/GES00212.S5 or the full-text article on www.gsapubs.org to view the Supplemental Map File.
6Supplemental Methods File. This Microsoft Word file summarizes the analytical methods used for 40Ar/39Ar analyses. If you are viewing the PDF of this paper or reading it offline, please visit http://dx.doi.org/10.1130/GES00212.S6 or the full-text article on www.gsapubs.org to view the Supplemental Methods File.
7Supplemental Data File 3. Geochemistry for 40Ar/39Ar samples. This Microsoft Excel file contains the analytical data for whole-rock analyses for the 19 samples dated by the 40Ar/39Ar dating method. Analyses were performed by ActLabs and the analytical methods are summarized in Greene et al. (2009b). Additional information in this table includes: UTM coordinates, geographic location, lithology, mineral proportion, and texture. If you are viewing the PDF of this paper or reading it offline, please visit http://dx.doi.org/10.1130/GES00212.S7 or the full-text article on www.gsapubs.org to view Supplemental Data File 3.
8Supplemental Data File 4. 40Ar/39Ar analytical data. This Microsoft Excel file contains the analytical data for the 40Ar/39Ar dating method for the 20 mineral separates analyzed at the Noble Gas Laboratory in the Pacific Centre for Isotopic and Geochemical Research at University of British Columbia. Age spectra and isochron diagrams for each of the samples are shown along with all the analytical data from incremental step-heating. The analytical methods are summarized in the Supplemental Methods File. Where multiple analyses of a single sample were made, all the results for each sample are included in a single worksheet in this workbook. If you are viewing the PDF of this paper or reading it offline, please visit http://dx.doi.org/10.1130/GES00212.S8 or the full-text article on www.gsapubs.org to view Supplemental Data File 4.
9Supplementary Data File 5. Wrangellia ages and biostratigraphy. This Microsoft Excel file contains three worksheets. (1) The Age worksheet contains ∼750 isotopic ages for units assigned to Wrangellia. Most of this data was extracted from the CordAge 2004 database and was supplemented by ∼50 ages not in the CordAge database, mostly from Alaska. Samples of the Coast and Insular Belts are included, based on their location, and caution should be used in deciding whether samples are actually located within Wrangellia. CordAge 2004 (a database of isotopic age determinations for rock units from the Canadian Cordillera) is an MS-Access based database, consisting of the merged datasets of the publically available products BCAge 2004A-1 (released October 2004; Breitsprecher and Mortensen (2004a)) and YukonAge 2004 (released July, 2004; Breitsprecher and Mortensen (2004b)). The compilation contains all reported non-proprietary isotopic age determinations for bedrock units from British Columbia and Yukon Territory, respectively: 9321 age determinations from 5997 rock samples, summarizing 778 published articles, theses, reports or unpublished sources. Katrin Breitsprecher offered assistance with extracting this information. The user is referred to Breitsprecher and Mortensen (2004a, 2004b) for information about the rating system of ages (Rel_rating column; column H) and other details. The database should not be cited as the source of the age. Age determinations should be cited to the original source, which is provided in each record of the database. (2) The Age_refs worksheet contains the information of the references for the age compilation listed in the Age worksheet (described above). The ref no column (column A) in the Age_refs worksheet refers to the Ref No column (column L) in the Age worksheet. (3) The Biostratigraphy worksheet is a compilation of 75 fossil age determinations from published literature related to Wrangellia. The full reference where each determination is published is included in column B and the region is shown in column E. All of the age ranges are plotted in Figure 8 according to the region indicated in column E. The low and high ages for each age range use the age boundaries for epochs and stages from Gradstein et al. (2004). Age boundaries for the Triassic are from Ogg (2004) and revised based on Furin et al. (2006). If you are viewing the PDF of this paper or reading it offline, please visit http://dx.doi.org/10.1130/GES00212.S9 or the full-text article on www.gsapubs.org to view Supplemental Data File 5.
10Supplemental Data File 6. Previous research on Wrangellia. This Microsoft Word file is a summary of previous research on parts of the Wrangellia terrane in Alaska, Yukon, and British Columbia that includes many references. These references and others are listed in Supplemental Data Files 1 and 2. If you are viewing the PDF of this paper or reading it offline, please visit http://dx.doi.org/10.1130/GES00212.S10 or the full-text article on www.gsapubs.org to view Supplemental Data File 6.
11Supplemental Data File 7. Stratigraphy of Wrangellia. This Microsoft Word file presents a detailed description of the stratigraphy of Wrangellia. See the Supplementary Figure File for nine figures that accompany this summary of the stratigraphy. If you are viewing the PDF of this paper or reading it offline, please visit http://dx.doi.org/10.1130/GES00212.S11 or the full-text article on www.gsapubs.org to view Supplemental Data File 7.