New ties between the Alexander terrane and Wrangellia and implications for North America Cordilleran evolution

Western North America is characterized by the Cordilleran accretionary mountain belt that has seen episodic plate convergence since the early Paleozoic (Fig. 1; Colpron et al., 2007; Coney et al., 1980; Jones et al., 1977; Monger et al., 1982). The long-lived accretionary history of the northern Cordillera has resulted in a collage of terranes and overlap assemblages that seemingly have quite disparate geologic histories. Recent syntheses on terrane affinities has seen the collage transformed into a more coherent picture with geologic ties between terranes that were previously thought of as complete separate entities (cf. Colpron et al., 2007). This view of terranes adheres to the more traditional ideas of possible links between Laurentia and the accreted terranes with a few seemingly truly exotic pieces caught up in the collage (Coney et al., 1980; Jones et al., 1977; Rubin et al., 1990). Building upon these ideas, much of the current research in Cordilleran geology is focusing on those terranes that appear to show a more complex relationship with respect to Laurentia.

Wrangellia and the Alexander terrane are the two largest outboard terranes with characteristics that indicate a seemingly exotic nature with respect to the Laurentian margin (Fig. 1). The two terranes, along with smaller outboard terranes and the larger Peninsular terrane, comprise the Insular terranes and are separated from more inboard Intermontane terranes by Mesozoic and younger magmatic rocks and faults (Fig. 1). Current thinking indicates that the Insular terranes were amalgamated to the western margin of the Intermontane terranes as one tectonic entity by at least the Middle Jurassic (McClelland et al., 1992; van der Heyden, 1992).

Recent studies into the origin and tectonic history of the Alexander terrane have increased our knowledge significantly, suggesting early Paleozoic stratigraphic and tectonic linkages with the Timanide orogen of northeastern Baltica and Caledonian orogen of northern Laurussia (Bazard et al., 1995; Beranek et al., 2012; Beranek et al., 2013a, 2013b; Nelson et al., 2013). However, little is known about the relationship between the Alexander terrane and Wrangellia prior to accretion to the Laurentian margin. Gardner et al. (1988) showed that the two terranes were stitched together by a 309 Ma pluton, but only briefly speculated on the tectonic and stratigraphic significance of this relationship. This revelation was in direct contrast to earlier studies that suggested the two terranes were separate tectonic entities until their mutual accretion to the North American margin in the mid-Cretaceous. These studies, however, were based primarily on seemingly different Triassic stratigraphy present on each of the terranes (Jones et al., 1977).

We present new high-precision U-Pb zircon CA-TIMS (chemical abrasion isotope dilution thermal ionization mass spectrometry) ages and geochemical and stratigraphic data that indicate that Wrangellia and the Alexander terrane shared a tectonic setting and geologic history by at least the latest Devonian. The new data indicate that the Alexander terrane, at least in part, formed the basement to a portion of Wrangellia and thus, the two terranes must be thought of as one entity by the latter part of the Paleozoic when considering paleogeographic reconstructions.

Wrangellia encompasses rocks from eastern Alaska and Yukon (northern Wrangellia) to Vancouver Island and northwestern United States (southern Wrangellia; Fig. 1). It was first described as a terrane by Jones et al. (1977), primarily based upon a thick package of Triassic basalts that is ubiquitous to the whole terrane, but also the Paleozoic basement underlying the basalts.

Paleozoic rocks in southern Wrangellia are characterized by Devonian through Permian volcanic, volcaniclastic, siliciclastic, and carbonate rocks of the Sicker and Buttle Lake groups (Massey, 1995; Yorath et al., 1999). Recent studies on Vancouver Island have identified several episodes of arc and back-arc cycles beginning in the Late Devonian and continuing into the Early Permian (Ruks et al., 2009, 2010). Paleozoic rocks are unconformably overlain by Middle Triassic marine shale and a thick package of Upper Triassic mafic volcanic rocks and carbonate (Nixon and Orr, 2006).

Northern Wrangellia is characterized by a thick succession of Carboniferous volcanic and volcaniclastic rocks conformably overlain by Permian coarse- to fine-grained marine siliciclastic rocks and carbonates. These rocks comprise the Skolai Group with the volcanic member assigned to the Station Creek Formation and the sedimentary member to the Hasen Creek Formation (Fig. 2; Read and Monger, 1976; Smith and MacKevett, 1970). A thick succession of mafic flows, pillow basalt and hyaloclastite interbedded with chert and thin crystal tuffs are found at the base of the Station Creek Formation (Israel and Cobbett, 2008; Read and Monger, 1976). The Hasen Creek Formation in most places is found to gradationally overly the volcanic rocks of the Station Creek Formation. The Skolai Group is unconformably overlain by Middle Triassic marine sedimentary rocks and thick accumulations of Upper Triassic mafic volcanic rocks, carbonate and calcareous to carbonaceous turbidites.

The Alexander terrane comprises much of coastal British Columbia and southeast Alaska and extends northwest through British Columbia, Yukon and southern Alaska (Fig. 1). It is a composite terrane divided into the Craig and Admiralty subterranes (Gehrels and Saleeby, 1987) that were likely amalgamated during the late Paleozoic (Karl et al., 2010). The Craig subterrane underlies all of southwest Yukon and more than 90% of the rest of the Alexander terrane (Gehrels and Saleeby, 1987). In southeast Alaska, the Craig subterrane is composed of Neoproterozoic metamorphic rocks of the Wales Group overlain by Upper Cambrian to Lower Silurian mafic and felsic volcanic rocks, shale and limestone of the Descon Formation and Upper Silurian to Lower Devonian carbonate and red beds of the Heceta and Karheen formations, respectively (Beranek et al., 2012; Gehrels and Saleeby, 1987). The Craig subterrane in southwest Yukon and northwestern British Columbia is characterized by volcanic, carbonate, and siliciclastic rocks assigned to several informal assemblages, including the Cambrian to Ordovician Donjek assemblage, the Ordovician to Silurian Goatherd Mountain assemblage, the Devonian to Triassic Icefield assemblage, and the Silurian to Devonian Bullion Creek limestone (Fig. 2; Dodds and Campbell, 1992).

In southwest Yukon, rocks in the Alexander terrane and Wrangellia are traditionally shown to be separated by the Duke River fault (Fig. 1). This fault has a protracted deformation history that includes oblique Early Cretaceous thrusting, possible latest Cretaceous to Paleocene strike-slip displacement and Miocene to present northeast directed thrusting (Cobbett, 2011). It is a young feature that is not representative of the original terrane boundary.

On the north side of the Duke River fault, Lower Permian conglomerates of the Hasen Creek Formation in Wrangellia overlie previously undated gabbro of the Steele Creek complex along a pronounced local unconformity (Fig. 3A; Read and Monger, 1976; Sharp, 1943). The unconformity can be tracked to the east for up to 20 km where it terminates against the Duke River fault. The unconformable relationship is unusual in that elsewhere in northern Wrangellia the contact between the Station Creek and the Hasen Creek formations is transitional from volcanic dominated to sedimentary dominated stratigraphy (Read and Monger, 1976; Smith and MacKevett, 1970).

In the Alexander terrane, on the south side of the Duke River fault, near the Alaska border, gabbros assigned to the previously undated Mount Constantine complex intruded Silurian to Devonian carbonate rocks of the Bullion Creek limestone (Figs. 1, 3B). Field observations show that this gabbro, although altered, is lithologically similar to the gabbro of the Steele Creek complex, north of the Duke River fault. Several other gabbro bodies, the largest near Mount Vulcan (Fig. 1), are interpreted as having intruded rocks assigned to the Alexander terrane.

Devonian and older rocks in the Alexander terrane, were intruded by swarms of mafic sills and dikes up to several meters thick. These sills are found throughout southwest Yukon and northwest British Columbia (Fig. 3C; Dodds and Campbell, 1992; Mihalynuk et al., 1993). The age of the dikes is unknown, but they are spatially associated with the gabbro bodies that intruded the Alexander terrane and have not been reported to have intruded rocks younger than Devonian.

We used U-Pb zircon CA-TIMS geochronology to constrain the timing of gabbro intrusion and deposition of a felsic crystal tuff in the lower Station Creek Formation. One gabbro sample is from the Steele Creek complex in Wrangellia and another is from the Mount Constantine complex in the Alexander terrane near the Klutlan Glacier area. The tuff and the Steele Creek gabbro were analyzed at the Pacific Centre for Geochemical Isotopic Research at the University of British Columbia (PCIGR). The Mount Constantine gabbro was analyzed at the Isotope Geology Laboratory at Boise State University. Details of methods are given in the Data Repository¹. Dates are weighted mean ²⁰⁶Pb/²³⁸U dates from 3 to 6 analyses of single zircon grains that are equivalent in age.

Gabbro from the Steele Creek complex in Wrangellia yielded a date of 363.47 ± 0.92 Ma (Fig. 4A; Table 1). Gabbro from the Mount Constantine complex in Alexander yielded a date of 363.53 ± 0.12 Ma (Fig. 4A; Table 2). Dates are identical and interpreted as igneous crystallization ages. A crystal-rich felsic tuff interbedded with pillowed mafic flows and chert from the lower Station Creek Formation yielded a date of 352.84 ± 0.30 Ma (Fig. 4B; Table 1), interpreted as the deposition age.

Geochemical data were obtained from the two dated gabbros, one sample from the Mount Vulcan body, two samples of sills that intruded the Alexander terrane, and several of basalt from the lower Station Creek Formation to determine their composition and tectonic setting (Table 3). The basalts show primitive mantle-normalized trace element plots typical of back-arc basin basalt to N-MORB and E-MORB affinity (Fig. 5A). The sills and gabbro show non-arc like geochemical signatures (Fig. 5B). The sills are nearly identical with slightly depleted N-MORB characteristics and the Mount Vulcan gabbro is very slightly depleted similar to the sills. The Steele Creek gabbro is similar to the N-MORB characteristics of the Mount Vulcan gabbro, but shows elevated Ti, Eu, and Nb concentrations. The Mount Constantine gabbro has the most depleted values of all the gabbros (Fig. 5B). The pattern shows a non-arc affinity but with strongly depleted values for almost all elements.

The U-Pb zircon and geochemical data provide the first evidence for a latest Devonian to Early Mississippian rifting event that affected the Alexander terrane and Wrangellia. We propose that this rifting was related to back-arc development along the margin of the Alexander terrane within the Skolai/Sicker arc system (Fig. 6A). A subduction zone would have been present at the margin of the Alexander terrane and extended beyond the terrane limits, into the oceanic realm, such that the Skolai portion of the arc was interacting with the Alexander terrane and the Sicker portion was not (Fig. 6A). In southern Wrangellia, arc activity during the Late Devonian to earliest Mississippian is well documented (Massey, 1995; Yorath et al., 1999). Extension of the arc-system began by the latest Devonian, likely due to slab rollback, and is manifested by the intrusion of non-arc gabbro complexes and voluminous sills and dykes at the margin of the Alexander terrane (Fig. 6A). The back-arc was fully developed by the Early Mississippian with deposition of the lower Station Creek Formation basalt and chert. The timing of back-arc development in northern Wrangellia corresponds with a similar tectonic setting in southern Wrangellia, documented by the presence of bimodal volcanic rocks and volcanogenic massive sulfide–style mineralization (Massey, 1995; Yorath, 1991) indicative of an arc under extension (cf. Piercey et al., 2004).

Collapse of the back-arc in northern Wrangellia in the latest Mississippian to latest Pennsylvanian was likely facilitated by a reversal of subduction polarity and accompanied by a resurgence of arc volcanic activity built primarily upon Wrangellia (Fig. 6B). Final closure of the back-arc led to the collision between Alexander and the arc system, clogging and subsequently shutting down subduction. We propose that the collision led to exhumation of the gabbro complexes along the margin of the Alexander terrane and the local deposition of conglomerates of the Hasen Creek Formation across the gabbro complexes and surrounding basement (Fig. 6B). Arc-related melts intruded across the Alexander-Wrangellia boundary, including the Barnard Glacier pluton (Dodds and Campbell, 1992; Gardner et al., 1988). Early Permian volcanic rocks are interbedded with the sedimentary rocks in both southwest Yukon and southeast Alaska, indicating that the arc in northern Wrangellia was slowly dying (Karl et al., 2010; Read and Monger, 1976). In southern Wrangellia, latest Mississippian to Early Permian volcanic and volcaniclastic rocks are interbedded with marine sedimentary rocks and capped by carbonates (Massey, 1995; Ruks et al., 2010), suggesting the Skolai/Sicker arc system also died out in the south by the Permian.

Wrangellia and the Alexander terrane relationships have broader implications for how tectonic terranes are perceived. Jones et al. (1977) defined Wrangellia largely on Triassic stratigraphy and thought the older portions of the terrane were separate from the Alexander terrane. Our data show a clear relationship between the two terranes, which were thought exotic to one another; they actually have ties that go back as far as their inception. This is a provocative idea that blurs the definition a terrane by showing that many of the observations of Jones et al. (1977) can actually be attributed to evolving tectonic processes rather than disparate geologic histories.

Although the Alexander terrane and Wrangellia are seemingly exotic to Laurentia, we show they are not exotic from one another. From at least the late Devonian, the two can be considered as being tectonically and stratigraphically linked. Furthermore, the latest Devonian to Mississippian rifting is similar to other events recently identified along the Laurentian margin. The opening of the Slide Mountain basin in the Devonian to Mississippian involves the Intermontane terranes of the northern Cordillera, mainly the Yukon-Tanana and Stikine terranes (see Fig. 1 for terrane relationships), separating from the Laurentian margin (Colpron et al., 2007). Faunal and magmatic similarities between Wrangellia and the Stikine terrane existed in the late Paleozoic, suggesting that they were in close proximity (Belasky et al., 2002). It may be that the rifting observed within the Alexander terrane and Wrangellia is part of the system that opened the Slide Mountain–Angayucham ocean basin (Fig. 6C).

The authors thank M. Colpron and J. Nelson for discussions on the Paleozoic paleogeography of the Alexander terrane and comments on earlier versions of the manuscript. Reviews by G. Gehrels and J. Goodge helped to improve the manuscript. This paper is Yukon Geological Survey contribution number 017.