Miocene basin development and volcanism along a strike-slip to flat-slab subduction transition: Stratigraphy, geochemistry, and geochronology of the central Wrangell volcanic belt, Yakutat–North America collision zone

The evolution of continental-margin volcanic belts is typically difficult to reconstruct from stratigraphic studies. Erosion of high-standing volcanic belts removes much of the volcanic record, leaving only an expression of the subvolcanic plumbing of eruptive centers. Alternatively, subsidence of volcanic centers and subsequent burial by younger volcanic rocks obscures older phases of the stratigraphic record. Thus, exhumed sedimentary basins formed within volcanic belts offer valuable records of the long-term evolution of processes, environments, and climates. However, basins associated with such volcanic belts have not received the same rigorous analysis as other basin types, partly due to the paucity of well-exposed outcrops, postdepositional modification by deformation and intrusive activity, and abrupt lithofacies changes that complicate stratigraphic analyses (e.g., Smith and Landis, 1995; Bassett and Busby, 2005; Busby and Bassett, 2007). Previously unstudied sedimentary and volcanic strata exposed in high-relief regions of southern Alaska and adjacent parts of Canada offer insight on volcanic-related basin development. In particular, Holocene–recent deglaciation of the eastern Wrangell Mountains provides exceptional cross-sectional exposures of Neogene sedimentary and volcanic strata in the central part of the >500-km-long Wrangell volcanic belt (WVB) (Figs. 1–3). The depositional, geochronologic, petrologic, and geochemical characteristics of these strata are undocumented beyond the original regional-scale (1:63,360–1:250,000 scale) geologic mapping and a few K-Ar ages (Richter, 1976; MacKevett, 1978; Richter et al., 1990, 2000, 2006).

New geologic data from the WVB also contribute to reconstruction of the tectonic evolution of the northern Pacific margin, including subduction collision of the allochthonous Yakutat terrane. A >600-km-long segment of the Yakutat terrane was subducted at a shallow angle beneath south-central Alaska, prompting high rates of seismicity, crustal shortening, and erosion (Ferris et al., 2003; Eberhart-Phillips et al., 2006; Haeussler, 2008). Recent studies provide major advances in our understanding of Neogene deformation, exhumation, and sediment accumulation along the suture zone between the Yakutat terrane and the former continental margin of Alaska (Sheaf et al., 2003; Bruhn et al., 2004; Pavlis et al., 2004; Gulick et al., 2007; Berger et al., 2008a, 2008b; Chapman et al., 2008?; Meigs et al., 2008; Enkelmann et al., 2008, 2009, 2010; McAleer et al., 2009; Perry et al., 2009; Witmer, 2009, Christeson et al., 2010; Koons et al., 2010; Worthington et al., 2010, 2012). In contrast, links between the subducted northeastern edge of the Yakutat terrane and deformation, volcanism, and sediment accumulation in the adjacent WVB remain largely unexplored, especially compared with continental U.S. standards.

In this paper, we investigate basin development and volcanism in the central WVB through stratigraphic observations, sediment provenance, geochemistry, and geochronology of Miocene sedimentary and volcanic strata exposed in the eastern Wrangell Mountains (MacKevett, 1978; Richter et al., 1990). This study reports (1) detailed stratigraphic and sedimentological data characterizing conditions of deposition and volcanism; (2) sedimentary provenance data (conglomerate and sandstone modal analyses, paleocurrents, and detrital zircon ages); (3) major and trace element geochemical compositions of lavas; and (4) ⁴⁰Ar/³⁹Ar and U/Pb ages of tuffs and lavas that bracket the timing of volcanism and basin development. Integration of these data sets with published studies from the eastern and western WVB provides the basis for understanding basin development and volcanism of the entire WVB in a regional tectonic context and offers an exceptional example of the stratigraphy and long-term evolution of a volcanic-related sedimentary basin.

The northern Pacific plate margin is characterized by a west to east transition from normal subduction to flat-slab subduction to transform tectonics (Chapman et al., 2008; Haeussler, 2008). In the western region, subduction of the Pacific oceanic plate along the Aleutian megathrust produces a moderately dipping Wadati-Benioff zone that reaches depths of 100–150 km within ∼400 km of the trench and active volcanism in the Aleutian Arc (Fig. 1). The central region is distinguished by relatively high topography, a lack of active volcanism, and a shallowly dipping Wadati-Benioff zone produced by flat-slab subduction of thickened oceanic crust, referred to as the Yakutat microplate or Yakutat terrane. Following Haeussler (2008), we use Yakutat terrane when the allochthonous nature of the crust is being emphasized and Yakutat microplate when plate kinematics is being highlighted. Currently subducting northwestward at a rate of 4.0–4.9 cm/yr, the shallow slab extends nearly horizontally for ∼250 km northwestward beneath Alaska at a subduction angle of ∼6° before reaching a depth of 150 km more than 600 km inboard of the Aleutian Trench (Eberhart-Phillips et al., 2006; Leonard et al., 2007). The microplate is separated from the North American plate by the Fairweather right-lateral fault to the east, the Aleutian megathrust to the northwest, and the Chugach–Saint Elias thrust to the north (Fig. 1). The Chugach–Saint Elias Range records deformation and exhumation within the zone of convergence between the Yakutat microplate and

North American plate (Bruhn et al., 2004; Pavlis et al., 2004; Haeussler, 2008; Chapman et al., 2008). Regional tomographic (Eberhart-Phillips et al., 2006) and receiver function studies (Ferris et al., 2003) image the subducted portion of the Yakutat crust as a 15–20-km-thick low-velocity zone. The crystalline crust composing the unsubducted portion of the Yakutat terrane increases in thickness west to east from 15 km to 30 km (Worthington et al., 2012). Based on the crustal thickness, velocity structure, and shallow subduction angle, the Yakutat terrane is interpreted as an oceanic plateau composed of anomalously thick, and therefore more buoyant, oceanic crust (e.g., Bruhn et al., 2004; Pavlis et al., 2004; Gulick et al., 2007; Worthington et al., 2012).

The eastern region is marked by active volcanism in the WVB, shallow seismicity (<50 km), and right-lateral displacement along the Denali, Totschunda, and Fairweather faults (Page et al., 1991; Eberhart-Phillips et al., 2003; Preece and Hart, 2004). The Wrangell Mountains record volcanism, deformation, and exhumation along the northeastern edge of the subducted microplate (Miller and Richter, 1994; Preece and Hart, 2004; Enkelmann et al., 2010). Seismicity studies record few earthquakes along the eastern Denali fault, suggesting that this segment of the fault is relatively inactive (Page et al., 1991; Matmon et al., 2006). Instead, the Totschunda and Fairweather faults cut the WVB and transfer right-lateral shear from the Queen Charlotte fault in southeastern Alaska and British Columbia northward to the Denali fault in central Alaska. The Fairweather and Totschunda faults may be joined by an unnamed southeast-striking fault (Richter and Matson, 1971; Kalbas et al., 2008); alternatively, the Fairweather fault may merge westward with numerous thrust faults (Doser and Lomas, 2000). The Fairweather fault, Totschunda fault, and possibly a Fairweather-Totschunda connecting fault constitute a northward extension of the Queen Charlotte–Fairweather transform system that accommodates right-lateral motion of the Pacific plate and Yakutat microplate relative to North America (Kalbas et al., 2008).

Southern Alaska consists of a collage of accreted terranes, magmatic belts, exhumed sedimentary basins, and subduction complex rocks that amalgamated during Mesozoic to Cenozoic time (Nokleberg et al., 1994; Trop and Ridgway, 2007). Three composite terranes make up much of the crust. From north to south, these include the Yukon, Wrangellia, and Southern Margin composite terranes (Fig. 1). The Yukon composite terrane consists of Proterozoic–Paleozoic metamorphic rocks (Yukon-Tanana terrane) and igneous rocks (Stikine terrane). The Wrangellia composite terrane consists of Mesozoic island-arc assemblages (Wrangellia and Peninsular terranes) and Paleozoic–Precambrian rifted continental-margin assemblages (Alexander terrane) that accreted to the Yukon composite terrane by Cretaceous time (Rioux et al., 2007; Trop, 2008). Mesozoic basinal strata in the suture zone record collision and exhumation of the Wrangellia and Yukon composite terranes (Hampton et al., 2010). The Denali fault bisects the suture zone and accommodates as much as 400 km of Cretaceous–recent right-lateral displacement (Nokleberg et al., 1994; Matmon et al., 2006). The Wrangellia composite terrane is juxtaposed against the Southern Margin composite terrane along the Border Ranges fault system, which records as much as several hundred kilometers of Late Cretaceous–Paleogene right-lateral displacement (Pavlis and Roeske, 2007). From north to south, the Southern Margin composite terrane includes the Chugach, Prince William, and Yakutat terranes. The Chugach and Prince William terranes consist chiefly of Cretaceous–Eocene metasedimentary strata interpreted as offscraped seafloor strata associated with subduction of oceanic crust beneath the Wrangellia composite terrane prior to collision of the allochthonous Yakutat terrane (Plafker et al., 1994). The Yakutat terrane is sutured against the Prince William terrane along the Chugach–Saint Elias fault (Fig. 1; Chapman et al., 2008).

The WVB extends >500 km across eastern Alaska, northwestern British Columbia, and southwestern Yukon Territory, making up much of the high relief of the Wrangell Mountains, including ice-capped volcanoes with elevations exceeding 4600 m (Figs. 1–3). The volcanic belt consists of Miocene–Holocene lava flows, lava domes, and pyroclastic deposits erupted from shield volcanoes, stratovolcanoes, caldera complexes, and cinder cones along with subordinate siliciclastic strata that accumulated in volcanic-related sedimentary basins (Skulski et al., 1991, 1992; Richter et al., 2006). WVB deposits unconformably overlie the Wrangellia composite terrane, exhibit maximum preserved thicknesses >3 km, and yield geochronologic ages ranging from 26 Ma to 1500 ka (Fig. 4; Richter, 1976; MacKevett, 1978; Richter et al., 2006).

The ages, geochemical compositions, and structural framework of WVB strata vary from east to west. WVB deposits in the Yukon Territory and northwestern British Columbia crop out discontinuously for >300 km along the Duke River fault splay of the Denali fault (Figs. 1–3). Individual volcanic centers include the Stanley Creek, Alsek, Nines Creek, and Saint Clare centers (Fig. 1). Previous studies define two stratigraphic packages: a lower succession of Eocene–Oligocene sedimentary strata (Amphitheater Formation) and an upper succession of chiefly Miocene lavas and subordinate pyroclastic, intrusive, and sedimentary rocks (Wrangell Lavas; Skulski et al., 1991, 1992; Cole and Ridgway, 1993; Ridgway and DeCelles, 1993a, 1993b). The Duke River fault cuts the lower part of the WVB lavas and possibly related faults cut the entire lava pile. Eastern WVB volcanism is interpreted chiefly as the product of eruptions along leaky strike-slip faults, based on the dominance of alkaline to transitional geochemical compositions and close spatial association of the volcanic rocks with strike-slip faults that cut the volcanic pile (Souther and Stanciu, 1975; Skulski et al., 1991, 1992).

The central WVB, the focus of this study, extends from the Alaska-Yukon border region to Mount Blackburn, a Pliocene volcano in the Wrangell Mountains (Figs. 2 and 3). Previous studies define two stratigraphic packages: a lower succession of Miocene sedimentary strata with subordinate lavas, pyroclastic rocks, and intrusive rocks (Frederika Formation) and an upper succession of chiefly Miocene lavas and subordinate pyroclastic, intrusive, and sedimentary rocks (Wrangell Lavas; Figs. 4 and 5; Denton and Armstrong, 1969; MacKevett, 1978; Eyles and Eyles, 1989; Richter et al., 2000). Both successions are cut by north-striking high-angle faults that exhibit small amounts of normal separation (mostly <10–20 m) and the northwest-striking Totschunda fault, which reveals ∼10 km of right-lateral offset (Richter and Matson, 1971; Plafker et al., 1977). The central WVB includes at least three known eruptive centers: Sonya Creek, Castle Mountain, and Mount Churchill (Figs. 2 and 3). Miocene intrusions within the belt likely represent the dissected subvolcanic roots of remnant eruptive centers (Porphyry, Sourdough, Twarpies intrusions in Fig. 3).

The western WVB consists of Pliocene–Holocene shield volcanoes, stratovolcanoes, and cinder cones in the western Wrangell Mountains (Fig. 2; Richter et al., 1990, 1994, 2006). Extensive icefields and glaciers partially cover Quaternary volcanoes with summit calderas at elevations >3500–4900 m above sea level. Western WVB volcanism is attributed to subduction-related processes based on transitional to calc-alkaline geochemical compositions and the spatial association of shield and stratovolcanoes above a geophysically imaged northward-dipping subducting slab (Page et al., 1989; Miller and Richter, 1994; Preece and Hart, 2004; Fuis et al., 2008). In summary, the western WVB records Miocene–Holocene subduction-related volcanism, whereas the eastern WVB records Oligocene–Miocene volcanism and basin development associated with displacement along strike-slip faults. New geochronologic, geochemical, and stratigraphic data from previously undocumented Miocene strata of the central WVB permit comprehensive evaluation of the tectonic evolution of the entire 500-km-long length of the volcanic belt, including the transition zone from strike-slip to subduction tectonics.

Central WVB deposits are well exposed in recently deglaciated cirques below the Nabesna icefield, plateaus between the Chitistone and Nizina Rivers, and foothills along the White River (Fig. 3). Bedded strata exhibit consistently low-angle bedding dips (<15°) except for outcrops cut by the Totschunda fault. Figure 5 shows generalized sections for most of the studied outcrops. The Frederika Formation overlies, in angular unconformity, Mesozoic rocks with onlap relationships and an altitudinal range for the basal contact, reflecting an incised paleotopography with local relief >400 m (Figs. 3, 7A, and 7B). The maximum preserved thickness of the Frederika Formation in our measured sections is ∼600 m (section 5 in Fig. 5). Upsection transitions from mostly sedimentary strata of the Frederika Formation to chiefly volcanic strata of the Wrangell Lavas vary from abrupt (sections 3–5 in Fig. 5) to gradational (sections 9, 13, 15 in Fig. 5). Uppermost strata are largely inaccessible due to steep, glaciated relief, so our measured sections span only several hundred meters of the upper succession (e.g., section 18 in Fig. 5). The maximum preserved thickness of the Wrangell Lavas is estimated as >3000 m (MacKevett, 1978).

Detailed sedimentologic observations were made while measuring a total of >5.5 km of stratigraphic sections on a bed-by-bed basis using a Jacob staff. Sections were correlated by tracing marker beds in the field and applying chronostratigraphy from isotopic ages of volcanic rocks. Complex lateral facies changes, paleotopography preserved along the basal erosional surface, and modern high-relief topography precluded more detailed stratigraphic correlations.

The Frederika Formation and lower Wrangell Lavas can be divided into four main lithofacies associations on the basis of grain size, lithology, and sedimentary structures. The lithofacies record deposition in proximal volcanic aprons and alluvial fans and more distal braided streams, anastomosing streams, and floodplains. Key aspects of each lithofacies association are summarized in the following. Table 1 summarizes common lithofacies and corresponding standard interpretations of depositional processes and environments. Key sedimentological features are shown in detailed sections in Figure 6 and Supplemental Figures 1¹, 2², 3³, and 4⁴, and photographs in Figures 7, 8, and 9.

Description. This association consists chiefly of moderately to well-organized, moderately sorted, clast-supported pebble-cobble conglomerate (Gcm, Gcmi), and subordinate massive to cross-stratified sandstone (Sm, St, Sp), tuffaceous sandstone (Smv), and sparse siltstone (Fsm) (see Fig. 5). Strata comprise broadly lenticular (several to <15 m) upward-fining bodies with erosional bases that scoured tens of centimeters deep. Internally, beds are amalgamated and marked by reactivation surfaces, reworked organic debris, and clast imbrication. Lithofacies association 1 makes up the lower 100–130 m of sections 13 and 15 in the northeastern part of the outcrop belt (Figs. 5 and 6; Supplemental Figs. 3 and 4 [see ^{footnotes 3} and ⁴]). Representative photos are shown in Figures 8A and 8B.

Interpretation. Lithofacies association 1 reflects subaqueous gravel deposition in braided stream channels and bar tops by tractive forces capable of routinely transporting pebble- and cobble-sized clasts. Evidence for fluvial bedload deposition includes the erosive bases, scours, upward-fining trends, unidirectional paleocurrent indicators, and laterally discontinuous sandstone and conglomerate. Sparse interbedded siltstones record overbank and waning-flood deposits. Characteristics of lithofacies association 1 are diagnostic of gravelly braided streams characterized by streamflow and flood flow (Lunt and Bridge, 2004; Miall, 2006). The moderate degree of sorting and/or rounding, lack of mass flow deposits, and upsection transitions to fine-grained fluvial-lacustrine strata (sections 13, 15 in Figs. 5 and 6 and Supplemental Figs. 3 and 4 [see ^{footnotes 3} and ⁴]) indicate that braided streams formed within alluvial plain or distal alluvial fan environments as opposed to proximal alluvial fans.

Description. Stratified volcanic-lithic sandstone (Sm, Sp, St, Sr), carbonaceous siltstone (Fsc), mudstone (Fsm), laminated shale (Fsl), bioturbated limestone and marl (Ls), coal, volcaniclastic strata (Smv, Vb, Vts, Vt), and subordinate pebble-cobble conglomerate (Gcm, Gcmi) typify association 2 (see Fig. 5). Lobate to wedge-shaped bed geometries dominate sandstone units, although coarser grained sandstone and conglomerate beds exhibit channel-form geometries. Many sandstone interbeds exhibit nonerosive bases, horizontal stratification, load marks, current ripple lamination, and convolute lamination. Mudstones and siltstones exhibit mottling, variegated colors, and bioturbated and/or disrupted horizons, root traces, plant fossils, coalified wood, and petrified logs. Laminated mudstone and fissile shales contain abundant disseminated organic material and fish fossils (Fig. 3; Berry, 1928). Reworked pyroclastic and epiclastic deposits along with primary pyroclastic material are also present locally. Facies association 2 typifies the lower half of stratigraphic sections in the central and northwestern parts of the outcrop belt (sections 3–6, 11–15 in Figs. 5 and 6 and Supplemental Figs. 3 and 4 [see ^{footnotes 3} and ⁴]). Representative photos of common lithofacies are shown in Figures 9A–9F.

Interpretation. Lithofacies association 2 reflects the deposits of anastamosing stream channels and vegetated floodplains. Channel-form sandstone and conglomerate record deposition by subaqueous tractive transport in fluvial channels. Finer grained detritus was deposited from suspension as overbank and lacustrine sediments on floodplains, including crevasse splays, levees, vegetated overbank areas, mires, and ponds and/or lakes. Abrupt upsection transitions from channel sandstones to overbank deposits are comparable with multichannel anastomosing fluvial environments in humid high-latitude settings (Smith, 1983; Nadon, 1994; Morozova and Smith, 1999; Makaske, 2001). Breaching of channel margins prompted waning-flow and crevasse-splay deposition of massive to horizontally stratified sandstone with lobate, wedge, and channel-form geometries. Overbank mudstones with mottling, variegated colors, and bioturbated and/or disrupted horizons (Fsm) (see Fig. 5 for unit abbreviations) suggest pedogenesis between events of stream migration and/or avulsion and flooding. Floodplains were partly forested, based on the presence of coal and mudstone with root traces, plant fossils, coalified wood, and petrified logs (Fsc). Coal and carbonaceous shale with nodules indicate accretion of organic material under suboxic to anoxic conditions in floodplain bogs, fens, moors, muskegs, and/or swamps. Subaqueous suspension settling in lakes and ponds deposited laminated mudstone and black shale with fish fossils (Fsl, Fsc). Bioturbated limestone and marl (Ls) reflect subaqueous inorganic carbonate precipitation in lakes and ponds under suboxic conditions permissive of infaunal activity. Lower oxygen bottom conditions in deeper ponds and lakes enabled deposition of finely laminated organic-rich shale and tuff that imply the absence of burrowing infauna or resuspension of bottom sediment by currents. Sandstone interbeds with nonerosive bases, horizontal stratification, load marks, current ripple lamination, and convolute lamination record episodic higher energy lacustrine deposition from sediment-laden unidirectional flows derived from lake-margin fluvial-deltaic environments during floods (Johnson and Graham, 2004; Pietras and Carroll, 2006).

Description. Association 3 consists of >200-m-thick successions of matrix- to clast-supported pebble-boulder conglomerate and breccia (Gmm, Gcm, Gcmi), tuffaceous sandstone (Smv), and lavas (see Fig. 5 for unit abbreviations). Subordinate massive to cross-stratified sandstone (Sm, St, Sp) and sparse massive siltstone (Fsm) are also present. Bedding contacts are dominantly subhorizontal erosive surfaces. Sandstone and conglomerate lithofacies typically comprise broadly lenticular (several meters to <15 m), 2–20 m-thick, upward-fining bodies with nonerosive to erosional bases that scoured tens of centimeters deep. Facies association 3 characterizes the lower and middle portions of sections in the northwestern and northeastern part of the outcrop belt (sections 3–5, 7–10, 20, and 21 in Figs. 5 and 6 and Supplemental Figs. 1 and 2 [see ^{footnotes 1} and ²]). Figure 8C shows a representative photo of this association.

Interpretation. Facies association 3 records reworking of volcanic debris and lava deposition on distal volcanic aprons flanking active volcanoes. Poorly sorted, matrix-rich conglomerate and breccia indicate emplacement by debris flow, high-concentration flood flow, and stream-bank collapse. Clast-supported and imbricated conglomerates are best interpreted as waterlain bedload deposits within broad channels. Subordinate sandstone and siltstone reflect lower energy bar deposits formed between channels. Alluvial fans flanked active volcanoes judging by the presence of andesite and/or basalt lavas, dominance of basalt and/or andesite lithic clasts in conglomerate and sandstone (Figs. 10 and 11), and detrital zircon ages from sandstone that overlap the age of intrabasinal eruptive centers (Fig. 12). Juvenile volcanic detritus was mixed, remobilized, and transported downslope over relatively short distances (<10 km from active vents).

Description. This lithofacies association consists of thick successions of interbedded lavas, volcanic breccia (Gmv), lithic-vitric tuff breccia (Vts), vitric-crystal tuff (Vt), and unsorted to poorly sorted conglomerate (Gmv), and spatially associated aphanitic dikes and sills (e.g., section 18 in Fig. 6; see also Fig. 5). Many lavas grade upward from lower massive units to vesicular and/or amygdaloidal flow tops. Volcanic breccias are massive, unorganized, thick bedded, and poorly sorted with angular to subangular clasts of exclusively volcanic lithologies. Subordinate tuff breccia, tuff, and unsorted conglomerate are massive to horizontally stratified. Representative photos of common lithofacies are shown in Figures 8D–8F.

Interpretation. Volcanic breccias reflect remobilization of unconsolidated volcanic material by volcanic debris avalanche and debris flow (lahar) on high-gradient flanks of volcanic edifices (Waythomas and Wallace, 2002; Busby and Bassett, 2007). Tuff breccia and tuff record episodic pyroclastic flows, ground-surge deposits, and air fallout (Fisher and Schmincke, 1984; Freundt et al., 2000). These strata are interpreted as near-vent eruptive deposits (<2 km from central vents), primarily effusive eruptions and sparse pyroclastic eruptions and associated intrusive feeder bodies. Volcanic apron facies consist chiefly of event deposits related to stripping and light reworking of volcanic debris from the steep flanks of volcanoes during or soon after an eruption (Smith, 1987).

Quantitative provenance data have not been reported previously from the central WVB. New clast counts from conglomerate, modal analysis of sandstone thin sections, U/Pb geochronology of detrital zircons, and paleocurrent measurements permit reconstruction of the lithologies, ages, and locations of sediment sources.

Paleocurrent data were collected by measuring the orientation of imbricated clasts in conglomerate and cross-beds in sandstone, totaling 163 measurements. Corrections for bedding dip were not needed because sampled beds dip <10°. Rose diagrams summarizing paleocurrent directions are shown in Figure 3. Paleocurrents reflect southwest- to southeast-directed sediment transport in the northern part of the outcrop belt, northward transport in southern outcrops, and west-directed paleoflow in western outcrops.

Compositional data were obtained from counts of 1316 individual clasts at 19 pebble-cobble conglomerate beds in the field by tabulating the lithology of all pebble-, cobble-, and boulder-sized clasts within a 1–5 m² outcrop face, yielding a minimum population of 80–100 clasts per bed to provide statistical significance (Van der Plas and Tobi, 1965). Figure 10 summarizes conglomerate compositional data; raw data were reported in Tidmore (2004) and Delaney (2006). The dominant volcanic category accounts for 91% of clasts counted. Black, reddish-brown, greenish-gray, and gray volcanic clasts with aphanitic, aphanitic-porphyritic, and amygdaloidal textures are most common (andesite and basalt 54%). White, green, and tan laminated tuff, crystal-vitric tuff, welded tuff, and pumice make up 37% of the counted clasts. Minor sedimentary (3%) and metavolcanic (5%) clasts include greenstone, chert, and volcanic-lithic sandstone.

Standard petrographic thin-sections were made from 60 medium-grained sandstone samples that were obtained while measuring stratigraphic sections. Thin sections of the least altered samples (n = 34) were stained for K-feldspar and Ca-plagioclase feldspar, and point counted (400–450 grains per section) according to a modified Gazzi-Dickinson method (Dickinson, 1970), in which crystals larger than silt sized in lithic fragments are counted as monocrystalline grains (Ingersoll et al., 1984). Point-counted sandstones are texturally immature with moderately to poorly sorted, angular to subrounded framework grains contained in a matrix of clay, silt, and devitrified glass. Figure 11 summarizes key compositional trends in the point-counted samples. Sandstones are compositionally immature with mainly unstable lithic fragments and plagioclase feldspar (quartz, feldspar, lithics, Q:F:L—14:24:62). Volcanic fragments are especially common (volcanic lithics, metamorphic lithics, sedimentary lithics, Lm:Lv:Ls—6:84:9) and include mainly basalt and andesite. Argillite, shale, quartzofeldspathic siltstone, and sparse carbonate grains make up the sedimentary lithic grain population. Metamorphic lithic grains include mica schist, quartz-mica schist, gneiss, and slate.

Two samples of moderately sorted, medium- to coarse-grained volcanic-lithic sandstone were processed for detrital U/Pb geochronologic analyses. Figure 5 shows the stratigraphic positions of the samples. Spot analyses were collected from 39–41 randomly selected zircon grains per sample (80 ages total) using the Stanford–U.S. Geological Survey sensitive high-resolution ion microprobe (SHRIMP; Fig. 12). Supplemental Table 1⁵ presents U/Pb isotopic age data; analytical details are described in Appendix 1. U/Th ratios are <9 for all analyzed zircons and <5 for 73 of 80 grains indicative of an igneous origin (Supplemental Table 1 [see ^{footnote 5}]). Five major age populations are documented: 12–9 Ma (Late Miocene, 40% of grains analyzed), 25–14 Ma (Late Oligocene–Middle Miocene, 21%), 89–70 Ma (Late Cretaceous, 8%), 118–98 Ma (Late Early Cretaceous, 4%), and 181–135 Ma (Early Jurassic–Early Cretaceous, 24%). Single Eocene (53 Ma), Permian (289 Ma), and Devonian (374 Ma) grains are present. Sample PIL yields Late Oligocene to Late Miocene grains (Fig. 12A), whereas sample DBT exhibits a broader distribution of ages (Fig. 12B).

We report isotopic ages from 12 lavas, 2 tuffs, and 2 sandstones in Table 2 and Supplemental Tables 1 (see ^{footnote 5}), 2⁶, and 3⁷; analytical details are described in Appendix 1. The U/Pb zircon age determinations were conducted at the Stanford–U.S. Geological Survey Microisotopic Analytical Center using the SHRIMP reverse geometry. Individual zircon crystals were separated from tuffs and analyzed through individual spot analyses. Results for individual crystals were used to calculate a weighted mean average age for each sample (Fig. 13). The ⁴⁰Ar/³⁹Ar age determinations were performed at the Lehigh University (Bethlehem, Pennsylvania) geochronology laboratory on plagioclase feldspar crystals separated from porphyritic lavas and groundmass separated from aphanitic lavas (Fig. 14).

The sampled volcanic rocks yield isotopic ages that range from ca. 11.2 to 5.3 Ma. Two tuffs from the middle of eastern sections yield 11.2–11.1 Ma U/Pb zircon ages (Fig. 13); 10 lavas from both eastern and western sections yield 10.0–8.3 Ma ⁴⁰Ar/³⁹Ar ages (Fig. 14). Two lavas from western sections yield ⁴⁰Ar/³⁹Ar ages of 5.39 ± 0.70 Ma and 5.29 ± 1.24 Ma (Fig. 14). Detrital zircons from sandstone exhibit minimum age clusters that are consistent with the isotopic ages obtained from volcanic rocks. The minimum U/Pb detrital zircon age population in sample DBT-207 centers at 10.6 Ma (Fig. 12B), compatible with the 11.2 Ma U/Pb zircon age of a tuff located 104 m lower in the stratigraphy (Figs. 5 and 13). Zircons from sandstone sample PIL-72 yield a minimum age population centered at 10.7 Ma (Fig. 12A), consistent with the 8.4 Ma ⁴⁰Ar/³⁹Ar age of a lava exposed 268 m higher in the section (Figs. 5 and 14). Ages young westward from 11.2–9.5 Ma near Castle Mountain, to 9.6–8.3 Ma at Chimney Mountain, to 8.7–8.3 Ma near the West Fork Glacier. Two samples from sections 3 and 6 in the western and southwestern part of the study area yield even younger ages of ca. 5.3 Ma. The 5.3 Ma ages are statistically indistinguishable from with 5.7–5.4 Ma ⁴⁰Ar/³⁹Ar ages previously reported from two sills that intrude Mesozoic strata a few kilometers southwest of section 3 (Figs. 3 and 4). In summary, integration of our new ages with previously reported ages support progressive westward migration of volcanism from ca. 11.2 to 5.3 Ma (Figs. 3 and 4; Table 2).

In order to characterize the range of lava compositions present in the central WVB and to compare magmatism in this region with existing data for the western and eastern WVB and the nearby Sonya Creek volcanic field, lavas and a few hypabyssal intrusive samples were collected from across the area shown in Figure 3. The vast majority of these were collected in the context of the measured sections detailed in Figures 5 and 6 and Supplemental Figures 1–4 (see ^{footnotes 1–4}). General descriptions of the igneous units are given in Table 1, and the geochemical data keyed to section and location within section are provided in Table 3. Analytical methods are described in Notes section in Table 3. As illustrated in Figures 4–6 and Supplemental Figures 1–4 (see ^{footnotes 1–4}), lava flows are most common in the upper portion of the Frederika Formation (lithofacies 3) and the overlying Wrangell Lava (lithofacies 4). Because no clear geochemical distinctions are identified between lavas from these lithofacies associations, and no systematic geographically controlled chemical variation is observed (Fig. 15A), for geochemical and petrologic purposes the samples are treated as a single time-transgressive suite.

Although considerable care was taken in the field to collect the freshest possible volcanic materials, it is clear from LOI (total volatiles lost on ignition at 950 °C for 45 min) values in excess of 5 wt% in some cases and some analytical totals outside a generally accepted range of 98.5–101 wt% that a number of samples have been affected by secondary hydrothermal and/or meteoric fluid migration. Based on a combination of (1) petrographic characteristics including high abundances of secondary phyllosilicate and/or carbonate phases, (2) LOI > 3.5 wt%, and (3) anomalous (nonigneous) relationships between Mg, Fe, Ni, Cr, V, and Sc, 17 samples are interpreted to possess nonprimary geochemical traits and thus are eliminated from further geochemical and petrologic discussion. Figure 15A illustrates that the bulk compositions of these excluded samples generally are within the range of other samples in the suite; we therefore believe that our interpretations are based on a representative data set. Data for the excluded samples are presented in Supplemental Table 4⁸.

Figures 15B–15D portray important basic classification and nomenclature attributes of the Frederika and Wrangell magmatic suite. For these and all other geochemical plots, major element data have been normalized to 100% anhydrous prior to plotting. Using a variety of widely accepted elemental parameters and relationships, the data define a subalkaline basaltic andesite through rhyolite suite with predominantly calc-alkaline affinity, in line with a convergent plate margin setting. Figure 16 includes selected major (wt%) and trace (ppm) element concentrations plotted versus wt% SiO₂. Many of the trends observed are common for fractionation-dominated differentiation in igneous suites, for example, the overall decreases in MgO, TiO₂, Ni, and Sr and increase in K₂O with increasing SiO₂. Diverging trends in Y and Zr with increasing SiO₂ coupled with broad ranges in concentration for a number of elements at the more mafic (basaltic andesite) end of the compositional spectrum are characteristics that the Frederika-Wrangell suite share with eruptive products of the younger than 5 Ma Wrangell Volcanic Field (predominantly western WVB). As discussed in Preece and Hart (2004), these traits may suggest one or more of variable source, melting, or early differentiation histories.

Active intrabasinal volcanoes were the primary source of sediment deposited in the central WVB, based on provenance data. Texturally immature volcanic-lithic clasts, plagioclase feldspar, embayed monocrystalline quartz, and unstable accessory minerals (Figs. 10 and 11) reflect erosion of volcanic edifices and limited transport through depositional environments (e.g., Critelli and Ingersoll, 1995). Mafic- to intermediate-composition volcanic clasts in conglomerate are texturally and mineralogically identical to intrabasinal lavas that crop out in the central WVB. The youngest detrital zircon age population (11.7–9.3 Ma) closely matches the age of volcanic interbeds in the Frederika Formation and Wrangell Lavas (11.2–8.4 Ma; Figs. 3 and 12; Table 2), and most detrital zircons exhibit euhedral to subhedral grain shapes, oscillatory zoning, and U/Th ratios <6, indicative of an igneous origin (Williams, 2001; Rubatto et al., 2001). Subordinate detritus was eroded from inactive eruptive centers that remained topographically elevated in the eastern part of the central WVB. Sandstone sample PIL-72 yields 25–14 Ma detrital zircons (Fig. 12A) that overlap the age of volcanic rocks and intrusions exposed <20 km northeast of the central WVB near the Alaska-Yukon border in the Sonya Creek and Rocker Creek area (Figs. 1 and 4; Skulski et al., 1992; Richter et al., 2000).

Local Mesozoic–Paleozoic igneous and sedimentary sources contributed subordinate detritus to the central WVB. Late Cretaceous detrital zircons (89–70 Ma) match the age of plutons that crop out across southern Alaska and western Canada (Breitsprecher and Mortensen, 2004a, 2004b), including <50 km north of the central WVB near the Denali fault (Richter, 1976). Mid-Cretaceous detrital zircons (118–97 Ma) overlap the age range of 123–85 Ma plutons exposed along the north flank of the Wrangell Mountains (Snyder and Hart, 2007) and more distant areas of eastern Alaska and Yukon Territory (Richter et al., 1975; Foster et al., 1994; Israel et al., 2006; Israel and Cobbett, 2008). Middle Jurassic–Early Cretaceous detrital zircons (169–136 Ma) overlap the age of 170–130 Ma plutons that crop out along the south flank of the Wrangell Mountains (Roeske et al., 2003; Richter et al., 2006). Detrital zircon grains with Permian (289 Ma) and Devonian (374 Ma) ages match the ages of plutonic-metamorphic belts that crop out across eastern Alaska and adjacent parts of Canada (Aleinikoff et al., 1988; Nokleberg et al., 1994). In the central WVB, this belt includes 320–282 Ma plutons and metavolcanic rocks, including the Barnard and Ahtell plutons (Gardner et al., 1988; Richter et al., 1975) and gneiss with 380–340 Ma isotopic ages that crops out >50 km north of the WVB in eastern Alaska and Yukon Territory (Dusel-Bacon and Williams, 2009). Jurassic–Cretaceous detrital zircons may have been recycled from Jurassic–Cretaceous sedimentary strata that underlie the central WVB and crop out north and south of the central WVB (Figs. 3 and 7B). Mesozoic strata exposed <5–15 km south of the central WVB contain 153–152 Ma granitic clasts (Trop et al., 2002); similar strata exposed <30 km to the north yield 159–147 Ma detrital zircons (Manuszak et al., 2007). However, Mesozoic sedimentary strata were subordinate sources compared to igneous sources, given the paucity of sedimentary lithic grains in conglomerate and sandstone together with the dominance of Miocene detrital zircons in central WVB strata. Potential sources exposed north of the Denali fault would have contributed metamorphic clasts and Precambrian–Paleozoic detrital zircons (Foster et al., 1994; Dusel-Bacon and Williams, 2009; Hampton et al., 2010) that are not observed in central WVB provenance data. Similarly, potential sources exposed south of the Border Range fault (Chugach terrane, Fig. 1) would have provided metasedimentary clasts and a broader population of Early to middle Cretaceous detrital ages (Amato and Pavlis, 2010; Kochelek et al., 2011).

Stratigraphic and geochronologic data demonstrate that the depositional and stratigraphic evolution of the central WVB was closely linked to construction of intrabasinal volcanic centers, active crustal subsidence, and extensional deformation. Volcanic eruptions routinely influenced environments of deposition based on the intercalation of volcanic breccia, lahars, volcaniclastic and/or pyroclastic deposits, and lavas with alluvial-fluvial-lacustrine siliciclastic deposits. Stratigraphic data demonstrate that volcanic aprons and alluvial fans prograded across fluvial-lacustrine environments (Figs. 17 and 18). Fluvial-lacustrine strata with thin volcanic interbeds (associations 1 and 2) are progressively replaced upsection by volcanic apron strata with thick-bedded lavas and lahars (associations 3 and 4) that increase in abundance and thickness upsection until only volcanic strata are preserved (e.g., sections 4 and 5 in Fig. 5 and Supplemental Figs. 1 and 2 [see ^{footnotes 1} and ²]). We attribute progradation to topographic expansion of volcanic centers coeval with basin subsidence. The vertical stratigraphic pattern is consistent with tectonically controlled models of basin stacking patterns (e.g., Blair and Bilodeau, 1988). Volcanism was characterized by outpouring of lavas interrupted by less common, small-volume pyroclastic eruptions. The predominance of lavas compared with pyroclastic deposits suggests a volcanic field characterized by multiple vents scattered along fissures and hypabyssal intrusions that fed into volcanic deposits (Condit and Connor, 1996; Walker, 2000; Cole et al., 2006). The basin fill is not interpreted as one or more calderas given that lowermost strata overlie Mesozoic strata along a basin-wide unconformity, thick silicic ignimbrites typical of caldera fills are not evident, and ring faults and ring fracture intrusions are not observed. Miocene dikes strike parallel to north-striking normal faults that cut Miocene sedimentary and volcanic strata throughout the study area (Figs. 3 and 17; MacKevett, 1970a, 1970b), indicating that normal faults may have been conduits for magma ascent. The north-striking normal faults dip steeply and strata do not exhibit substantial postdepositional rotation (note dominance of bedding dips from 0° to 10° in Fig. 3), indicating limited displacement along nonlistric faults. Localized thickening of volcanic and sedimentary strata across some faults (Fig. 7A) indicates syndepositional displacement, but growth strata are not well developed; for example, intraformational unconformities are not evident. The north-striking orientation of normal faults is consistent with syndepositional extension in a zone of right-lateral shear between northwest-striking strike-slip faults that cut strata along the northeastern and southwestern margins of the outcrop belt (Totschunda fault and an unnamed fault that truncates the northeast side of Porphyry and Sourdough Mountains in Fig. 3). Alternatively, extensional deformation may have been unrelated to strike-slip tectonics and initiated mainly by thermal weakening from ascending magmatic bodies. Unfortunately, the timing and kinematics of faults that cut the basin fill are not well established. Additional field studies of faults and their relationship with Miocene volcanic and sedimentary rocks are needed to accurately and completely understand magma ascent mechanisms and basin development.

Stratigraphic evidence, including paludal coal seams and lacustrine deposits, indicate that sediment accumulation was accommodated by active subsidence rather than being simply a response to ponding by volcanic stratigraphy. Subsidence is attributable to extensional deformation along intrabasinal normal faults and heat flow associated with intrabasinal volcanic centers. Well-developed progradational stacking patterns (Figs. 5, 17, and 18) together with the limited thickness of coal and lacustrine strata indicate that subsidence was outpaced by supply of sediment and volcanic deposits. Topographic expansion of intrabasinal volcanoes may have eventually prompted development of an internally drained intramontane basin through damming of fluvial-lacustrine systems, but erosion of basin margin outcrops hampers detailed reconstruction of the basin paleogeography.

Our new stratigraphic, geochronologic, and lithofacies data from the central WVB (Fig. 5) document localized distribution of vent-proximal volcanic-sedimentary strata (<3200 km²), high accumulation rates of sedimentary-volcanic strata (>0.5 mm/yr), brief duration of sedimentary-volcanic accumulation (<3 m.y.), complex lateral lithofacies patterns, and evidence for lateral shifts in volcanism and basin development. These attributes are consistent with models of intraarc basin development (e.g., Smith and Landis, 1995). Similar intraarc successions are reported from continental arcs in Oregon-Washington-Idaho (Smith et al., 1987; Carlson and Moye, 1990; Smith and Landis, 1995; Janecke et al., 1997), Mexico (Righter et al., 1995), New Zealand (Stern, 1985), and Arizona (Bassett and Busby, 2005; Busby and Bassett, 2007). The depocenter likely originated as a transtensional intraarc basin if basin-bounding right-lateral faults and intrabasinal normal faults were active during sediment deposition (e.g., Polliand et al., 2005; Busby et al., 2005). With oblique convergence of only 10° from orthogonal or more, strike-slip faults may form in the upper plate (Jarrard, 1986), especially in thermally weakened crust of volcanic belts (Cole and Lewis, 1981; Geist et al., 1988; McCaffrey, 1992; Weinberg, 1992; Bellier and Sebrier, 1994; Van Dijk, 1994; Israde-Alcantara and Garduno-Monroy, 1999), resulting in transtensional intraarc basins (e.g., Polliand et al., 2005; Busby et al., 2005). Establishment of the timing and kinematics of central WVB deformation by future structural studies will help us to better evaluate the tectonic setting of basin development.

The previous section outlined a scenario whereby the volcanic-sedimentary stratigraphy, complex lithofacies changes, and overall structural setting argue for volcanism and basin development in an intraarc setting. Using the chemistry of the lavas in conjunction with all of the other data presented herein can place first-order constraints on tectonic interpretations, particularly when compared to other WVB geochemical information. The older eastern WVB and the younger western WVB have been examined in some detail down to a volcano by volcano or at least volcanic field by volcanic field basis (e.g., Skulski et al., 1991, 1992; Richter et al., 1994, 1995; Preece and Hart 2004; and numerous references therein). We utilize the larger geochemical data sets and tectonomagmatic interpretations from these studies as a backdrop to our understanding of the central WVB Frederika-Wrangell magmatism.

Figure 19 uses major element data recalculated to cation percent values in order to compare central and western WVB data to those presented for the eastern WVB by Skulski et al. (1991). Plots A and B illustrate Na + K and Fe/Mg versus Si relationships for the central (this study and Sonya Creek field) and eastern (Saint Claire Creek field) WVB, and plots C and D illustrate the same parameters for the central and western (young western Wrangell and Mount Churchill) WVB. These plots convey the same general information depicted in Figures 15B and 15C. In addition, the Frederika-Wrangell data show little if any overlap with transitional subalkaline to alkaline (eastern Wrangell, EW-T) and more strongly alkaline (EW-A) lavas associated with leaky transform magmatic activity in the Saint Claire field or to lavas with strong intraplate geochemical signatures from the northwestern Canadian Cordillera (Skulski et al., 1991; Thorkelson et al., 2011). In contrast, note the strong overlap of the Frederika-Wrangell data with calc-alkaline and hybrid suites from Saint Claire Creek (EW-C; eastern), Sonya Creek (SC; central) and the young western (WW-1 and combined WW-2a, WW-2b) WVB. Using the interpretations of Skulski et al. (1991, 1992) and Preece and Hart (2004), these overlaps suggest magma genesis and evolution associated with subduction and intraarc extensional processes. A subduction as opposed to intraplate association for the central WVB is supported by the recent regional geochemical synthesis forwarded by Thorkelson et al. (2011).

In order to further explore these associations and their implications, more detailed comparisons are drawn between the central (this study) and western (Preece and Hart, 2004) WVB in Figures 20 and 21. Both figures illustrate the three suites or trends defined in Preece and Hart (2004): WW-1—trend 1 calc-alkaline to tholeiitic suite; WW-2a—trend 2a strongly calc-alkaline suite; WW2b—trend 2b dominantly adakitic suite. Along with TiO₂, Y concentration strongly differentiates between the trend 1 and trend 2 lavas, particularly at increased levels of differentiation (Fig. 20A). This distinction at least in part reflects a role for early amphibole fractionation in trend 2 materials, implying higher fluid concentrations in the trend 2 magma sources and/or chambers. All lavas from the Frederika-Wrangell suite can be grouped with the trend 1 (WW-1) and trend 2a (WW2a) materials of the younger than 5 Ma western WVB, suggesting a tectonomagmatic setting akin to either the interior (intraarc extensional regime and decompression melting of subduction modified mantle) or back side (trench distal portion of a magmatic arc and/or melting of recently slab fluid–enriched mantle) of the western region illustrated in Figure 2 (Fig. 20B).

Figure 21 illustrates a series of element and element ratio versus Y (ppm) plots that are employed to place additional qualitative constraints on Frederika-Wrangell tectonomagmatism. Figure 21A highlights the WW-1 and WW-2a association and also serves as a reminder that even the low SiO₂ members of the suite presented in this study have undergone considerable differentiation, so care must be taken when commenting on mantle sources. The K/P ratio versus Y plot (Fig. 21B) illustrates the dominance of trend 1 characteristics (low K/P), suggesting a less prominent role for slab fluid and/or felsic crustal influences (e.g., Carlson and Hart, 1987) than is common for the majority of trend 2a and 2b materials. Similarly, the Sr/P ratio (normalized to primitive mantle; Borg et al., 1997) versus Y plot (Fig. 21C) offers additional evidence for subduction-related but possibly not active subduction-driven magma genesis; for example, the dashed line at (Sr/P)_PM = 5.5 is suggested to distinguish between magmas strongly enriched in components derived from subducting lithosphere (>5.5) from those enriched in these components (1–5.5). The Sr/Y versus Y plot (Fig. 21D) clearly distinguishes the lavas of this study from adakitic lavas that often are interpreted to result from partial melting of subducted mafic oceanic crust (e.g., Defant et al., 1991) due to various factors, including a young, hot slab with a shallow dip, plate margin effects, and development of a slab window (e.g., Yogodzinski et al., 2001). Such factors seem reasonable for present-day southern Alaska, where shallow slab geometries are documented. By analogy and/or comparison to studies in the western and eastern WVB, these geochemical signatures are taken to indicate that Miocene central WVB magmatism was a result of variable degrees of partial melting of a heterogeneous slab component–enriched mantle wedge source locally in response to development of an intraarc extensional regime.

Neogene deformation, volcanism, and basin development in south-central Alaska are linked to oblique convergence and flat-slab subduction between North America and the Yakutat microplate (Plafker, 1987; Plafker et al., 1994). In earlier tectonic models, oceanic crust along the inboard (northern) part of the Yakutat microplate was inferred to have been subducted beneath eastern Alaska and adjacent parts of Canada from Early Oligocene (ca. 30 Ma) to Middle Miocene time, with the onset of subduction based on the age of the oldest WVB lavas (Rocker and Sonya Creeks in Figs. 1, 4, 22A, and 22B). These early models infer that more buoyant continentalized crust of the southern part of the Yakutat microplate was subducted at a shallow angle beneath southern Alaska starting in Middle Miocene time (Plafker et al., 1994). Recent geophysical data, however, indicate that both the subducted and unsubducted parts of the Yakutat microplate consist of thick buoyant crust. The subducted northwestern portion of the microplate is 11–22 km thick, the unsubducted portion is 15–30 km thick, and both exhibit high ratios of P-wave to S-wave velocities typical of overthickened oceanic crust (Ferris et al., 2003; Eberhart-Phillips et al., 2006; Gulick et al., 2007; Christeson et al., 2010; Worthington et al., 2012), possibly representing an oceanic plateau (Pavlis et al., 2004). Recent integration of geologic data sets from south-central Alaska documents crustal shortening, regional exhumation, cessation of magma tism, and inversion of sedimentary basins above the flat-slab region since Oligocene time (Enkelmann et al., 2008, 2010; Benowitz et al., 2011; Finzel et al., 2011). Sedimentary basins located along the western and northern perimeter of the flat-slab region record enhanced sediment accumulation rates and sediment delivery from bedrock sources exhumed above the flat-slab region beginning in Late Oligocene–Middle Miocene time (Cook Inlet and Tanana basins; Ridgway et al., 2007; Finzel et al., 2011). Together, these studies indicate shallow subduction of the Yakutat microplate since Oligocene time. Along the eastern perimeter of the flat-slab region, diachronous Oligocene–Quaternary volcanism and basin development in the WVB are consistent with progressive northwestward insertion of the northward-tapering slab into the arcuate continental margin of eastern Alaska. Oligocene–Miocene basin development and volcanism in the eastern WVB are linked to right-lateral displacement along the Duke River and Dalton faults in Yukon Territory and British Columbia (Figs. 22A, 22B). Eocene–Oligocene alluvial fan, fan-delta, and lacustrine strata >900 m thick record strike-slip basin development (Ridgway and DeCelles, 1993a, 1993b; Ridgway et al., 1992, 1996). Progressive extensional deformation within strike-slip basins permitted construction of intrabasinal volcanic centers and progradation of >1000-m-thick successions of 18–15 Ma lava flows and pyroclastic rocks (Wrangell Lavas) upon underlying sedimentary successions (Miocene Stanley Creek, Alsek, Nines Creek fields in Figs. 22A, 22B; Skulski et al., 1991; Cole and Ridgway, 1993). The predominance of transitional to alkaline geochemical compositions of volcanic rocks together with their limited spatial distribution along strike-slip faults indicates volcanism related to extensional strike-slip faulting (i.e., leaky transform volcanism; Skulski et al., 1991, 1992). Age-equivalent volcanic centers in the Saint Clare field (Yukon Territory) and Sonya Creek–Rocker Creek area record a spatial transition from strike-slip to subduction tectonics (Fig. 22B). The presence of shield volcanoes that produced materials with transitional to calc-alkaline geochemical compositions indicates derivation from a mid-oceanic ridge basalt (MORB) like mantle wedge heterogeneously enriched via the addition of slab-derived components. We interpret transtensional basins and volcanic centers along northwest-striking strike-slip faults as the products of initial orogen-parallel translation of the leading edge of the Yakutat microplate along the eastern, northwest-striking part of the arcuate continental margin (southern Alaska syntaxis). Coeval subduction-related volcanic centers are interpreted to represent northwestward subduction of oceanic crust (the inboard margin of the Yakutat microplate) beneath the central part of the syntaxis (Saint Clare and Sonya Creek fields in Figs. 22A, 22B).

Sedimentary and volcanic strata exposed in the Wrangell Mountains record a progressive northwestward shift in the position of volcanism and basin development ca. 12–8 Ma (Fig. 22C; this study). Sampled volcanic suites exhibit transitional to calc-alkaline geochemical compositions that suggest generation from subduction-modified mantle sources. Basinal strata record progradation of volcanic strata (Wrangell Lavas) across alluvial-lacustrine strata (Frederika Formation) in response to topographic expansion of eruptive centers. We interpret this spatial-temporal shift in the location of volcanism and basin development as a consequence of insertion of progressively thicker crust (documented by recent geophysical studies; e.g., Worthington et al., 2012) into the eastern part of the syntaxis synchronous with a documented change in plate motion (Fig. 22C; Stock and Molnar, 1988). Volcanic centers originating from subduction-related processes and sources would be expected to migrate northwestward as thicker, more buoyant crust entered the subduction zone from the southeast along the Queen Charlotte–Fairweather transform fault and the leading edge of the subducting slab migrated northwestward. We infer that intraarc basin development and volcanism were accommodated by extensional deformation along north-striking normal faults and, more speculatively, that transtensional basin development was associated with basin-bounding, northwest-striking, right-lateral strike-slip faults.

Volcanism continued to migrate northwestward during Late Miocene–Holocene time based on the spatial distribution of younger than 5 Ma to Holocene volcanoes in the western Wrangell Mountains (Figs. 4 and 22D; Richter et al., 1990, 2006; Preece and Hart, 2004). Continued northwestward migration of magmatism is consistent with thermochronologic data that document exhumation in the central WVB ca. 5–4 Ma (Enkelmann et al., 2010). An exception to the northwestward-younging pattern of volcanism is Mount Churchill, a Holocene stratovolcano in the central WVB that erupted large-volume silicic fallout deposits as recently as 1147 yr ago (Clague et al., 1995; Richter et al., 1995). It is not known whether sedimentary basin development preceded volcanism in the western WVB, as documented in the central and eastern WVB, because the subvolcanic stratigraphy remains largely obscured by Holocene volcanoes (Richter et al., 2006). Western WVB volcanism records subduction-related processes based on complex spatial-chemical variations above a modern geophysically imaged, north-dipping slab (Preece and Hart, 2004; Fuis et al., 2008). All sampled western WVB volcanic suites contain mafic parental compositions consistent with derivation from a MORB-like mantle wedge that has been heterogeneously enriched via the addition of slab-derived components. Large shield volcanoes and stratovolcanoes (Fig. 2; front-side and back-side in Fig. 20) surround the interior region and exhibit calc-alkaline mafic suites that indicate higher degrees of partial melting catalyzed by active fluid fluxing into the mantle wedge. High-TiO₂ transitional tholeiitic eruptive products emplaced along the inboard part of the volcanic field (interior in Figs. 2 and 20) indicate relatively low degrees of partial melting resulting from an intraarc extensional regime.

Northwestward migration of subduction-related magmatism is supported by the spatial extent of a present-day subducted slab deeper than 50 km beneath only the westernmost WVB (Fuis et al., 2008), although some studies indicate that a present-day deeply subducted slab (>50 km) is not present beneath any part of the WVB (Eberhart-Phillips et al., 2006; Qi et al., 2007). Subduction-related fluids implicated in magma generation and evolution processes in the central and western WVB were likely derived from shallowly subducting portions of the Yakutat slab. Steady dehydration of a shallow slab would release fluids that would be expected to migrate laterally from the slab and into the western WVB (Gutscher et al., 2000; Preece and Hart, 2004; Eberhart-Phillips et al., 2006). Adakitic, amphibole dacite magmas documented in samples from Mounts Drum and Churchill (Fig. 2) are consistent with partial melting of the Yakutat slab (Preece and Hart, 2004). Adakite melts commonly form by slab melting when the downgoing slab is young and at shallow depths, the dip of the downgoing plate is very shallow, or the plate margin is characterized by elevated shear stresses (Drummond et al., 1996; Gutscher et al., 2000; Yogodzinski et al., 1995, 2001), conditions that characterize the present-day flat-slab region of southern Alaska (Ferris et al., 2003; Eberhart-Phillips et al., 2006).

Northwestward insertion of progressively thicker crust into the syntaxis during Miocene to Holocene time is also consistent with the evolution of deformation and sediment accumulation outboard (southward) of the WVB. Thermochronologic data from the Chugach, Saint Elias, and Fairweather Ranges document exhumation starting ca. 26 Ma with pulses ca. 20 Ma and ca. 11 Ma followed by rapid, deep-seated, and focused exhumation in the corner of the syntaxis beginning ca. 6–5 Ma (Berger et al., 2008a, 2008b; McAleer et al., 2009; Enkelmann et al., 2008, 2010). Rapid exhumation was synchronous with Late Miocene to Pleistocene crustal shortening and synorogenic sedimentation in the adjacent Gulf of Alaska (Lagoe et al., 1993; Gulick et al., 2007; Witmer, 2009; Worthington et al., 2010). The Yakataga Formation consists of a siliciclastic succession as much as 10 km thick with intraformational angular unconformities reflecting syndepositional deformation; the succession reflects enhanced sediment accumulation rates following deposition of the Eocene–Early Miocene Poul Creek and Kulthieth Formations (Plafker, 1987; Witmer, 2009; Worthington et al., 2010). Magnetostratigraphic data suggest that the base of the Yakataga Formation was deposited during Late Miocene time (later than 11.6 Ma; Lagoe et al., 1993; Zellers and Gary, 2007), concurrent with the shift in WVB volcanism and basin development from the eastern part of the syntaxis to the central WVB.

The Miocene–recent evolution of the southern Alaska syntaxis has been compared with syntaxes that define the Himalaya-Tibetan orogenic belt (Enkelmann et al., 2010). The Alaskan and Himalaya-Tibetan syntaxes are similarly characterized by extreme local relief resulting from rapid exhumation and deformation at the spatial transition from strike-slip tectonics to convergence (Zeitler et al., 1993, 2001; Stewart et al., 2008). Both regions exhibit Miocene–Quaternary volcanic centers with complex spatial-geochemical-age patterns, including calc-alkaline volcanism linked to slab dehydration and diachronous alkaline volcanism and basin development associated with transtensional deformation along regional strike-slip faults (e.g., Maury et al., 2004; Wang et al., 2006, 2007; Yang et al., 2007; Lee et al., 2010). We speculate that similar patterns of deformation, volcanism, and basin development shaped pre-Cenozoic syntaxial margins, especially during assembly of supercontinents, and may be distinguishable from typical continental margin processes through integrated geochemical, geochronological, and stratigraphic data sets.

Sedimentology, provenance, geochemistry, and geochronology document construction of the central WVB and link this event to Neogene tectonic processes that shaped the northern Pacific plate margin. Lithofacies associations identified in the Frederika Formation and lower Wrangell Lavas indicate deposition by volcanic aprons, alluvial fans, and more distal braided streams and floodplain channels, lakes, and mires. Stratigraphic and topographic relationships along the basal unconformity show that accumulation occurred on a high-relief (400+ m) erosional surface developed on deformed Mesozoic marine strata. Upward-coarsening stratigraphic successions record progradation of volcanic aprons onto fluvial-lacustrine deposystems in response to topographic growth of intrabasinal volcanic centers. Paleoclimate conditions were sufficiently warm and moist to support abundant vegetation, peat formation, and perennial aqueous depositional processes. Compositional, paleocurrent, and detrital ages link deposition of the Frederika Formation chiefly to erosion of active intrabasinal volcanoes. Local eruptive centers provided juvenile volcanic detritus, including volcanic-lithic clasts, plagioclase feldspar crystals, tuffaceous matrix, and 12–9 Ma detrital zircons. Subordinate sediment was derived from remnant volcanoes currently exposed along the Alaska-Yukon border as well as Mesozoic plutons and marine strata that underlie or crop out adjacent to central WVB strata. Central WVB lavas exhibit calc-alkaline geochemical compositions ranging from basaltic andesite to rhyolite. The sampled lavas broadly overlap Pliocene–Quaternary phases of the western WVB interpreted as the product of subduction-related volcanism. In contrast, Miocene volcanic rocks from the eastern WVB are attributable to magmatism along leaky strike-slip faults and exhibit higher Fe/Mg ratios and total alkalies at a given silica content. The U/Pb and ⁴⁰Ar/³⁹Ar age analyses of tuffs and lavas indicate that sediment accumulation and volcanism in the central WVB commenced by ca. 12.5–11.2 Ma and shifted westward to younger vents from ca. 8.3–9.6 Ma. Samples from two western sections yield even younger ages of ca. 5.3 Ma. Integration of these new ages with published ages supports northwestward migration of volcanism and basin development across the eastern (ca. 26–10.4 Ma), central (12.5–5 Ma), and western (5 or later to 0.8 Ma) WVB. Stratigraphic evidence, including coal and lacustrine deposits, indicates active tectonic subsidence during basin development. Likely subsidence mechanisms include heat flow from intrabasinal volcanoes and extensional deformation linked to intrabasinal normal faults. The orientation of the normal faults is consistent with east-west extension associated with right-lateral displacement along basin-bounding northwest-striking strike-slip faults that cut the basinal strata. WVB basin development and volcanism initiated in western Canada during northward translation of the Yakutat terrane along orogen-parallel, northwest-striking strike-slip faults. Volcanism and basin development shifted northwestward into eastern Alaska in response to insertion of progressively thicker crust into the southern Alaska plate corner synchronous with a shift in plate motions to more northwestward-directed convergence. Volcanism persists today in the westernmost WVB above a weakly imaged north-dipping subducted slab along the northeastern edge of the subducted portion of the Yakutat microplate. The stratigraphic, geochemical, and geochronologic attributes of the WVB and their links to regional tectonic elements in southern Alaska contribute to our growing understanding of the complex processes responsible for crustal accretion and recycling. In particular, WVB strata offer an exceptional example of diachronous volcanism, basin development, and deformation associated with subduction of thickened oceanic crust into a syntaxial plate margin characterized by a spatial transition from strike-slip deformation to shallow-slab subduction. Our analysis highlights the importance of integrated multidisciplinary data sets to distinguish ancient stratigraphic successions associated with slab-edge volcanism and deformation generated by subduction-collision of thickened oceanic crust along a syntaxial plate margin from volcanic-arc complexes produced by subduction of typical oceanic crust.

We completed 107 zircon U/Pb sensitive high-resolution ion microprobe (SHRIMP) analyses from 2 tuff samples and 2 sandstone samples. (Refer to Supplemental Tables 1 and 2 [see ^{footnotes 5} and ⁶] for U-Th-Pb data for all analyses.) Zircons were separated using standard sample-preparation methods, and U/Pb SHRIMP analyses, data reduction and plotting using Squid and Isoplot, respectively (Ludwig, 2006). No morphologic or color differentiation was made during hand-picking for the sample mount. Zircon mounts were examined through scanning electron microscope-cathodoluminescent imaging to detect structures correlated with zircon trace element concentrations. U-Th-Pb isotopic ratios and elemental abundance of zircon grains were determined by Trop at the Stanford-U.S. Geological Survey Microisotopic Analytical Center using the reverse geometry SHRIMP, following standard operating procedures. The primary beam spot size was 20–40 μm in diameter, and the spot was placed over the middle of the grain, avoiding cracks, inclusions, and broken edges of grain fragments. Analyses consisted of 4–6 scans through each isotope mass, counting 2 s on Zr₂O, 7 s on background, 12 s on ²⁰⁶Pb, 16 s on ²⁰⁷Pb, 10 s on ²⁰⁸Pb, 5 s on ²³⁸U, 3 s on ²⁴⁸(ThO), and 3 s on ²⁵⁴(UO). Pb/U ratios were calibrated with reference to standard zircon R33, which was analyzed after every fourth or fifth analysis of an unknown zircon. Reported ages are based on ²⁰⁴Pb/²⁰⁶Pb-corrected ²⁰⁶/²³⁸U ratios because errors of the ²⁰⁷Pb/²³⁵Ua and ²⁰⁶Pb/²⁰⁷Pb ratios are significantly greater. The ²⁰⁶Pb/²⁰⁷Pb ages for these Paleozoic–Cenozoic zircon grains are less reliable, given the relatively low concentration of ²⁰⁷Pb. The Pb/U ratios were calibrated with reference standard zircon R33 (419 Ma), which was analyzed after every fourth or fifth unknown analysis.

Ages for 12 basalt samples were determined by the ⁴⁰Ar/³⁹Ar method at Lehigh University (Bethlehem, Pennsylvania). Plagioclase phenocrysts were separated for dating from seven samples, and ground mass material was prepared from five aphanitic lavas that exhibit relatively little evidence of alteration and interstitial glass. Olivine and pyroxene phenocrysts were removed from the groundmass separates prior to irradiation in order to reduce trapped ⁴⁰Ar contamination. The samples were irradiated two batches. Samples BK1–654, CHM-38, CHM-39, FDN-3, FDN-8, SKO3, SKO-16, and SKO-28A were irradiated in position 5C of the McMaster University reactor for 6 h. The Oregon State University (OSU) TRIGA reactor was used for samples 57–6877, CHM-4, MGA-1, NZM-66, and PIL-335, which were subjected to a 10 h irradiation in the CLICIT (cadmium-lined, in-core irradiation tube) facility. The neutron flux within the irradiation containers was monitored with GA1550 biotite (98.79 ± 0.54 Ma; Renne et al., 1998). CaF₂ and K₂SO₄ were also included in the irradiation packages to monitor neutron-induced interferences from Ca and K, respectively.

Argon was extracted from the samples by stepwise heating in a double-vacuum resistance furnace, and purified using a cold finger cooled to liquid nitrogen temperature and two SAES ST101 getters. The analyses were performed with a VG3600 noble gas mass spectrometer equipped with an electron multiplier operated in the pulse-counting mode. The mass spectrometer sensitivity during the dating experiments was ∼1.154 × 10⁻¹⁹ mol/count/second ⁴⁰Ar. Extraction line blanks were typically ∼2.5 × 10⁻¹⁵ mol ⁴⁰Ar at 1350 °C and <1 × 10⁻¹⁵ mol ⁴⁰Ar at temperatures <1000 °C, and were approximately atmospheric in composition. The isotopic data were corrected for extraction line blank, mass spectrometer background, mass discrimination, radioactive decay of ³⁷Ar and ³⁹Ar, neutron-induced interferences, and atmospheric contamination prior to calculation of the ages. The neutron interference correction factors determined for the McMaster irradiation were (³⁶Ar/³⁷Ar)_Ca = 0.000280, (³⁹Ar/³⁷Ar)_Ca = 0.000699, and (⁴⁰Ar/³⁹Ar)_K = 0.026; for the OSU irradiation the corrections were (³⁶Ar/³⁷Ar)_Ca = 0.000264, (³⁹Ar/³⁷Ar)_Ca = 0.000680, and (⁴⁰Ar/³⁹Ar)_K = 0.00075. Mass discrimination ranged from 0.06 to 0.19%/atomic mass units over the course of the experiments (average measured atmospheric ⁴⁰Ar/³⁶Ar = 296.26 ± 0.75% for the McMaster samples and 297.76 ± 0.23% for the OSU samples). Ages were calculated using the decay constants and isotopic abundances of Steiger and Jager (1977). A complete tabulation of the analytical data is presented in Supplemental Table 3 (see ^{footnote 7}).

The final results of the dating experiments are summarized in Table 2. We report plateau ages for 8 samples yielding acceptable plateaus and atmospheric trapped ⁴⁰Ar/³⁶Ar. These plateau ages are the inverse-error weighted mean ages of contiguous steps that compose 50% or more of the gas released from a sample and define individual ages statistically indistinguishable from the weighted mean at the 2σ analytical error level. Inverse isochron ages are reported for samples FDN-8, MGA-2, NZM-66, and SKO-28A, which contain trapped argon of nonatmospheric composition. The isochron ages were determined by regressing blank-corrected isotope data using the method of York (1968).

This project was supported by the Donors of the Petroleum Research Fund, administered by the American Chemical Society (41464-B8 to Trop) and student grants (to M.R. Delaney, R. Tidmore, J. Witmer) from the Geological Society of America and the Bucknell University Program for Undergraduate Research. Undergraduate theses by Delaney and Tidmore provided important preliminary data sets important to this manuscript. We thank J. Morton for assistance with geochemical analyses; Delaney, Tidmore, and Witmer for helpful contributions in the field and laboratory; National Park Service personnel D. Rosenkrans and E. Veach for support to conduct field work in Wrangell Saint Elias National Park; Wrangell-Mountain Air pilots for helping us access remote outcrops; J. Wooden, F. Mazdab, and D. Bradley for assistance with U-Pb geochronologic analyses; T. Ager for palynological analyses; and B. Jordan for lab support. This study would not have been possible without the original careful mapping of the Wrangell volcanic field by D. Richter, E. MacKevett, and G. Plafker, among others. Constructive reviews by two anonymous reviewers and Associate Editor Terry Pavlis helped us improve the manuscript.