Geological setting

Continental margin strata in southern Alaska are created from sediment derived from the Yakutat terrane and several antecedent Mesozoic-to-modern accreted terranes that compose much of the northern North American Cordillera (Fig. F5) (Plafker, 1987; Plafker et al., 1994). The Yakutat terrane was likely excised from western Canada and translated to the northwest along the dextral Queen Charlotte-Fairweather Fault system (Plafker, 1987; Plafker et al., 1994; Landis, 2007; Perry et al., 2009) (Fig. F5). The age of initial Yakutat-North America subduction, when the leading edge of the microplate encountered the Aleutian Trench, and thus initiation of flat-slab subduction, is poorly constrained but may have occurred as early as ~40 Ma. Increased uplift has occurred in the past 10 m.y. associated with the convergence of increasingly thick crust and the formation of a syntaxial bend in the Pacific/North American plate boundary (Plafker et al., 1994; Rea and Snoeckx, 1995; White et al., 1997; Enkelmann et al., 2010; Finzel et al., 2011). Ongoing collision and flat-slab subduction of a thick (up to 35 km) oceanic plateau and cover strata (Christeson et al., 2010; Worthington et al., submitted) is constructing the present high topography of the Chugach-St. Elais Ranges (Pavlis et al., 2004; Eberhart-Phillips et al., 2006; Gulick et al., 2007), with active tectonic deformation spread throughout southern Alaska and northwestern Canada (Fig. F4) (Plafker et al., 1994; Mazzotti and Hyndman, 2002; Pavlis et al., 2004; Spotila et al., 2004; Eberhart-Phillips et al., 2006; Meigs et al., 2008; Spotila and Berger, 2010; Enkelmenn et al., 2010; Worthington et al., 2010; Finzel et al., 2011). Focused deformation occurs in two indentor corners (Fig. F4), one in the west that is experiencing thin-skinned young deformation (Bruhn et al., 2004) and the other in the east (Seward corner) (Spotilla and Berger, 2010; Enkelmann et al., 2010) where there is intense strain and high exhumation rates (“tectonic aneurysm”) associated with the change from strike-slip to collision (Enkelmann et al., 2008, 2010; Elliot et al., 2010). The active deformation front for this convergence cuts diagonally from the eastern syntaxis near Mt. St. Elias along the Malaspina Fault, reaching Icy Bay and then across the shelf as the Pamplona zone, and down the slope to the Aleutian Trench, thereby linking Yakutat-North America deformation structures with the Pacific-North America Faults (Bruns, 1983; Worthington et al., 2010). The northward boundary, or backstop, of deformation within the St. Elias orogen is debatable because of extensive ice cover and may include the Chugach-St. Elias Fault or the Contact fault (Fig. F6) (Spotila and Berger, 2010; Enkelmann et al., 2010); however there are suggested far-field effects more than 1000 km away (Mazzotti and Hyndman, 2002; Redfield et al., 2007).

The Yakutat terrane consists of Eocene to modern sedimentary rocks of the Kulthieth, Poul Creek, and Yakataga Formations that are primarily siliciclastic marine and glacimarine strata interbedded with volcanics and coal beds (Fig. F6) (Plafker et al., 1994). The Kulthieth Formation sediments are the oldest, deposited at ~60–35 Ma, and are composed of nonmarine to shallow-marine deltaic feldspathic and micaceous sandstones and siltstones formed during a general relative sea level regression (Risley et al., 1992; Plafker et al., 1994; Perry et al., 2009). The Poul Creek Formation appears to conformably overlie the Kulthieth, ranging in age from late Eocene to Oligocene (Risley et al., 1992; Plafker et al., 1994). This formation is characterized by a high abundance of argillaceous sediment that is in part glauconitic and organic rich, representing deposition during a general marine transgression in outer shelf/slope environments. It also contains waterlain basaltic tuff, andesitic breccia, pillow lava, and related deposits of the intertonguing Cenotaph volcanics (Plafker, 1987; Risley et al., 1992; Plafker et al., 1994). The Yakataga Formation represents glacimarine siliciclastic shelf/slope deposits dating from the onset of St. Elias glaciation in the late Miocene (~5.5 Ma) to the present (Plafker et al., 1994; Lagoe et al., 1993; Risley et al., 1992; Plafker, 1987; Eyles et al., 1991). It is composed of oldest diamict, possibly debris flows, transitioning upward into turbidites, diamictites, mudstones, and shallow-water sandstones with occasional coquinas and boulder pavements. The uppermost Yakataga mostly comprises facies that indicate significant lateral translation of the ice margin (paraglacial mudstones to boulder pavements). It is exposed onshore along the leading edge of the fold and thrust belt and is >5 km thick on the continental shelf.

Volumetrically, the major potential contributors of sediment to the Yakutat terrane during the Cenozoic are the Wrangellia, Alexander, Chugach, and Yukon-Tanana terranes (Fig. F5). The Kulthieth and Poul Creek Formations are likely derived from a parent source in British Columbia (Cowan, 1982; Plafker, 1987; Plafker et al., 1994; Landis, 2007; Perry et al., 2009). At the time of Kulthieth and Poul Creek Formation deposition, sediment transport pathways likely originated from the east, deriving material from the Intermontaine Belt (Cache Creek, Nisling, and the Stikine terranes), Coast Belt (Coast Plutonic Complex, Central Gneiss Complex), and Insular Belt (Wrangellia, Alexander, Chugach, and Yakutat terranes) in British Columbia during the latter stages of uplift (Fig. F5) (Gehrels and Berg, 1994; Plafker et al., 1994). The Wrangellia terrane consists of a lower section of Permian to Middle Triassic limestone, chert, and pelitic strata. These units are overlain by 4–5 km of mid-Triassic mafic lava flows and associated mafic and ultramafic rocks, providing a diagnostic isotopic and compositional marker of sediment derived from the Wrangellia terrane (Jones et al., 1977; Nokleberg et al., 2001; Trop et al., 2002). The Alexander terrane consists of Precambrian(?) through Middle(?) Jurassic sedimentary, metamorphic, and plutonic rocks (Gehrels and Berg, 1994) with a distinctive magnetic mineralogy (Cowan et al., 2006). The Chugach terrane consists mainly of highly deformed, weakly metamorphosed Upper Cretaceous graywacke and slate. These accretionary prism strata are interpreted to have been mainly derived from plutonic and volcanic rocks of the Coast Plutonic Complex in British Columbia (Dumoulin, 1987). The final major terrane that may have contributed sediment to the Yakutat terrane is the Yukon-Tanana terrane (Fig. F5). Polydeformed metamorphic rocks of this terrane consist mainly of quartz-mica schist, quartzite, metarhyolite, and gneissic plutonic rocks (Foster et al., 1994). Protolith ages are only partly known but include middle Paleozoic radiometric ages for some of the igneous rocks and Devonian paleontologic ages for some of the carbonate strata (Gehrels and Saleeby, 1987; Plafker, 1987; Haeussler et al., 2006).

Yakataga Formation and Surveyor Fan provenance is likely a combination of material supplied mostly from the exhuming fold and thrust belt (Kultieth, Poul Creek, and recycled Yakataga terranes) and the potential backstop of accretion provided by the Chugach and Prince William terranes, with minor contributions from more inboard terranes (Fig. F6). The Chugach-Prince William terranes are a subduction complex welded onto the continent when the Kula-Farallon Ridge subducted beneath the margin, resulting in high-temperature, low-pressure metamorphism (Plafker et al., 1994; Pavlis et al., 2004). The Chugach and Prince William terranes are lithologically similar and have been considered one composite terrane (Kusky et al., 1997). Age distributions of detrital zircons suggest metasedimentary flyschoidal rocks of the Chugach terrane are derived from inboard accreted terranes of British Columbia (Haeussler et al., 2006; Perry et al., 2009). The lithologies of the Chugach terrane are dominated by graywacke flysch and mélange units with substantial basaltic constituents. These mafic units have undergone very low grade metamorphism in the westernmost terrane, increasing in metamorphic grade to amphibolite-facies and phyllitic units of the Orca and Valdez groups in the Prince William terrane in the St. Elias region (Fig. F6) (Plafker et al., 1994; Haeussler et al., 2006). The Valdez group of the Prince William terrane north of the St. Elias Range is part of a mélange that is characterized by metavolcanic rocks and weakly foliated, green, glassy tuff, together with volcaniclastic graywacke and argillite (Plafker et al., 1994). Volcanic rocks of the southern margin of the Valdez group are dominantly tholeiitic pillow basalts with island-arc to mid-ocean-ridge basalt (MORB) compositions (Plafker et al., 1994). The Orca group dominates the Prince William terrane within the St. Elias Range. As with the Valdez group, this group is a deep-sea flysch complex with abundant oceanic basaltic rocks, sheeted dikes, and gabbroic intrusions of an ophiolite complex and underlies the Bagley Icefield along the contact fault system, from which the Bering Glacier sources (Plafker et al., 1994; Richter et al., 2006). Yakataga Formation strata accumulated in the later stages of the accretionary history of the Yakutat terrane. Sedimentary petrography of Yakataga lithic fragments reveals sedimentary, metasedimentary, and volcanic rock fragments. Zircon fission track, U-Th/He, and U/Pb analysis of the Yakataga Formation indicates intermixing of sediment derived from two or more sources, likely the Chugach-Prince William composite terrane and recycling of Poul Creek and Kulthieth rocks (Plafker et al., 1994; Enkelmann et al., 2008; Perry et al., 2009; Witmer et al., 2009).

Physical and oceanographic setting

The morphology of the Gulf of Alaska shelf seabed has been strongly influenced by active tectonics and glacial strata formation overprinted by glacial erosion (Carlson et al., 1982). The local bathymetry has been shaped by glacial and tectonic forces that produced five distinct across-shelf valleys in the study area: the Alsek Sea Valley, the Yakutat Sea Valley, the Bering Trough, the Kayak Trough, and the Hinchinbrook Sea Valley (Figs. F2, F7). These valleys are U-shaped, contain poorly sorted, glacially derived diamict along their flanks, and presumably formed during the advance of major ice streams and glaciers (Carlson et al., 1982). The extent of grounded ice cover on the shelf during the last sea level lowstand is poorly constrained, but it may have ranged from only the across-shelf valleys to the entire shelf (Fig. F8) (Molnia, 1986; Manley and Kaufmann, 2002). The glaciers that reached the sea were probably grounded tidewater, not floating ice sheets (Powell and Molnia, 1989). Little sampling has occurred on the continental slope or on Surveyor Fan. Grab samples recovered diamicton from the continental slope, which likely date to the LGM (Molnia and Sangery, 1979). The recently acquired jumbo piston Core EW0408-85JC on the continental slope at Site KB-2A reveals slowly accumulating (<1 mm/y) Holocene-age silty clay in the vicinity of the Alaska Coastal Current (ACC) (Davies et al., 2011). Piston Core EW0408-87JC on the proximal Surveyor Fan (Site GOAL-16B) also contains slowly accumulating (<1 mm/y), likely Holocene-age silty clay that may be derived from intermediate nepheloid layers supplied with sediment from the Alaska Coastal Current (Jaeger et al., 2008).

The modern oceanographic environment in the Gulf of Alaska is characterized by strong wintertime wave and wind energy, pronounced coastal currents, and deepwater gyres (Stabeno et al., 2004). The regional meteorology of the Gulf of Alaska is chiefly affected by energetic storms associated with the Aleutian Low Pressure System (ALPS). The Gulf of Alaska experiences mean wind speeds and frequency of gale-force winds similar to those of the western and central North Pacific (Stabeno et al., 2004). Cyclonic motion of the subarctic gyre drives circulation in the outer Gulf of Alaska with the inner Gulf dominated by the Alaska Coastal Current, a wind- and buoyancy-forced coastal jet (Fig. F7) (Stabeno et al., 2004). The southern boundary of this subarctic gyre, the North Pacific drift, diverges as it impinges on North America, with the northward branch becoming the Alaska Current (Fig. F7). The Alaska Current dominates flow along the southwestern and southern Alaska continental slope, eventually transitioning into the Alaskan Stream farther to the west. Several large semipermanent to seasonal eddies form within the Alaska Current, leading to important mixing of water masses across the continental margin (Stabeno et al., 2004).

Water and sediment are dispersed on the continental shelf by the Alaska Coastal Current (Feely et al., 1979; Stabeno et al., 2004; Weingartner et al., 2005). The current is generally confined to within 40 km of the coast and has flow velocities occasionally in excess of 50 cm/s and mean annual transport of ~106 m3/s (Royer, 1982; Stabeno et al., 1995). Although primarily wind driven, the Alaska Coastal Current is enhanced by a baroclinic response to coastal freshwater discharge to the Gulf of Alaska (Royer 1981, 1982). This freshwater flux is delivered via a series of small mountainous drainages that experience high precipitation rates (2–6 m/y) due to adiabatic cooling of warm, moist air associated with the cyclonic storm systems of the ALPS (Weingartner et al., 2005). Freshwater and sediment discharges are lowest in winter when much precipitation is stored as snow and peak in late summer/fall when meltwater and precipitation rates are greatest (Stabeno et al., 2004; Neal et al., 2010). Runoff creates a sharp shallow (<50 m) halocline of salinity contrast >3 over the shelf in the fall but during the winter is mixed (>100 m) and much more subdued (contrast of ~1) as a result of strong storms (Stabeno et al., 2004). Strong cyclonic winds dominate from fall through spring, peaking in December and January at 9 m/s, and intense (>14 m/s) easterly coastal jets forced by mountain topography can occur (Stabeno et al., 2004). These winds lead to downwelling conditions and near-bed (3 m above bed) nontidal maximum currents of 0.15 m/s in the summer and >0.3 m/s in the winter (Hayes and Schumacher, 1976; Hayes, 1979). Wave energy is highest in the winter months, decreasing in the summer. Monthly mean significant wave heights at National Oceanic and Atmospheric Administration (NOAA) Buoy 46001 (Fig. F7) averaged over 25 y are 3.5 ± 2 m (1σ), and maximum significant wave heights are 14 m in November and December (Gilhousen et al., 1983). Summer significant wave heights average 1.5 m at Buoy 46001, but maxima of 6–10 m can occur.

Regional productivity in the Gulf of Alaska is strongly influenced by the interaction of high freshwater runoff and regional meteorology dictated by the ALPS. During fall through spring, strong cyclonic winds associated with the ALPS support onshore surface Ekman transport and downwelling on the shelf, along with storm-induced vertical mixing (Stabeno et al., 2004; Childers et al., 2005). During summer the onshore winds and subsequent downwelling conditions relax, allowing occasional brief periods of coastal upwelling in this dominantly downwelling system (Stabeno et al., 2004). Primary productivity in the Gulf of Alaska in the winter is inhibited by low insolation and enhanced vertical mixing that limits the near-surface residence time of algae. Algal blooms occur over the shelf in the early spring due to the increased solar irradiance, wintertime replenished nutrient supply, and the onset of water column stratification leading to enhanced cell residence time in surface waters. Productivity remains relatively high through early summer but is followed by a reduction in summer due to nutrient limitation created by the strong halocline (Childers et al., 2005; Stabeno et al., 2004). Key nutrients (nitrate, silicic acid, and phosphate) are derived mostly from the subsurface ocean, via open-ocean upwelling, onshore Ekman transport, tidal pumping, and storm or eddy mixing (Childers et al., 2005). Nutrients delivered by the fluvial system include iron and silicic acid (Stabeno et al., 2004). Relatively few data are available on the cycling of iron in this system, although it appears shelf processes and surface water discharge may play a role in regulating surface-ocean iron concentrations (Stabeno et al., 2004; Schroth et al., 2009; Wu et al., 2009; Davies et al., 2011; Addison et al., submitted). In contrast to the productive coast, the central Gulf of Alaska is a high-nitrate low-chlorophyll (HNLC) region (Stabeno et al., 2004), and primary productivity is likely limited by micronutrients such as iron (Boyd et al., 2004; Tsuda et al., 2005). Sources of iron to the central basin include curl-driven upwelling, aeolian dust (Mahowald et al., 2005), advection of dissolved iron from the continental shelf and slope (Chase et al., 2007; Lam and Bishop, 2008), and terrestrial runoff (Stabeno et al., 2004; Royer, 2005). The Expedition 341 drill sites span from the productive nitrate-limited shelf system to the iron-limited open-ocean system.

Northwestern Cordilleran ice sheet dynamics

The glacial history of the Gulf of Alaska margin has been constructed through a combination of surface outcrop sampling (e.g., Lagoe et al., 1993; Lagoe and Zellers, 1996; White et al., 1997), scientific drilling (Rea, Basov, Janecek, Palmer-Julson, et al., 1993; Rea and Snoeckx, 1995; Prueher and Rea, 1998), and industry well cuttings (Lagoe et al., 1993; Zellers, 1995; Lagoe and Zellers, 1996). The chronology of these events has been established through paleomagnetic and diatom/radiolarian biostratigraphic control at Ocean Drilling Program (ODP) Site 887 (Barron et al., 1995), more limited tephrochronology, paleomagnetic, and biostratigraphic control at Deep Sea Drilling Project (DSDP) Site 178 (Von Huene et al., 1973; Lagoe et al., 1993), and foraminiferal biostratigraphy in industry wells on the shelf (Zellers, 1995) and in the Yakataga Formation outcrops (Lagoe et al., 1993). A summary of glacial history and the associated sediment record is shown in Figure F3. Alpine glaciation along the margin may have initiated as early as ~7 Ma (Lagoe et al., 1993) and was well under way by 5.5 Ma (Lagoe et al., 1993; Rea and Snoeckx, 1995; White et al., 1997), when elevation of the Chugach-St. Elias mountain belt was sufficient to trap precipitation from storms generated in the Gulf of Alaska. Initial onset of tidewater glaciation, Lagoe et al.’s (1993) “Glacial Interval A,” is linked to the appearance of ice-rafted debris (IRD) in the Yakataga Formation at the Yakataga Reef outcrop at 5.5 Ma and at Site 887 from ~5 Ma (Krissek, 1995) to 4.3 Ma (Rea and Snoeckx, 1995) (Fig. F3). A reduction in glacimarine sedimentation correlating with the ~4.5–2.8 Ma mid-Pliocene warm period (MPW) (Shackleton et al., 1995) is observed in marine and nonmarine records, though timing varies between different locales in the Gulf of Alaska region. In outcrop and continental shelf samples, the MPW lasts from 4.2 Ma to 3.5–3.0 Ma (Lagoe and Zellers, 1996). At Site 887, the MPW lasts from 3.6 to 2.8 Ma (Rea, Basov, Janecek, Palmer-Julson, et al., 1993; Rea and Snoeckx, 1995). Renewed onset of intense glaciation after ~3 Ma, “Glacial Interval B” (Lagoe et al., 1993), is characterized by an increase in IRD accumulation at 2.6 Ma within deep-sea records (Lagoe et al., 1993; Prueher and Rea, 1998) and by thick successions of diamictite in outcrop (Lagoe et al., 1993).

At ~1 Ma, the rate of terrigenous sedimentation doubles, likely due to widespread glacial advance associated with the mid-Pleistocene transition (MPT) that carved a series of U-shaped sea valleys to the shelf edge (Carlson, 1989; Lagoe et al., 1993; Rea and Snoeckx, 1995). This glacial intensification is referred to as “Glacial Interval C” (Berger et al., 2008a). Since the onset of Glacial Interval C, a series of 100 k.y. glacial–interglacial cycles characterize the late Pleistocene climate signal. Recent high-resolution seismic reflection profiles in the Bering Trough image glacial erosion surfaces that extend to the shelf edge, likely correlating with widespread Pleistocene glacial advances potentially associated with the onset of Glacial Interval C (Berger et al., 2008a). The northwestern lobe of the Cordilleran ice sheet was present on the shelf during the LGM (Fig. F8) (Davies et al., 2011), although the full extent of ice cover is poorly known (Molnia, 1986; Manley and Kaufman, 2002). Terrestrial records suggest that the regional LGM expression lasted from 23,000 to 14,700 cal y BP (calendar years before present), with evidence for millennial-scale cooling and transient glacial re-advances during deglaciation (Engstrom et al., 1990; Mann and Peteet; 1994, Briner et al., 2002; Hu et al., 2006; Davies et al., 2011). Retreat of the Bering Glacier off the continental shelf following the LGM likely occurred after 16,000 cal y BP, as indicated by peat accumulation in parts of the Bering foreland, and had apparently retreated well onshore by 14,700 cal y BP (Peteet, 2007; Davies et al., 2011).

Continental margin and Surveyor Fan stratigraphy

Varying degrees of glacial erosion, tectonic deformation, and rock exhumation in the St. Elias Range in southern Alaska and northwestern British Columbia since the Miocene (Lagoe et al., 1993; Rea and Snoeckx, 1995; Enkelmann et al., 2010; Spotila and Berger, 2010) supplied sediment into the Gulf of Alaska, leading to periodic significant increases in growth of the continental margin and Surveyor Fan (Stevenson and Embley, 1987; Lagoe et al., 1993; Rea and Snoeckx, 1995; Zellers, 1995; Worthington et al., 2010; Reece et al., in press).

The seismic stratigraphy of the Bering Trough between the Kayak Island zone (KIZ) and Pamplona zone (PZ; Fig. F4) has been imaged at several scales. Regional seismic surfaces, deformation structures, and seismic sequences are observed (Table T1; Figs. F9, F10), but age control is coarse and limited to cuttings from industry wells (Zellers, 1995). In high-resolution seismic Line GOA2505 and coincident crustal scale Line STEEP09, a series of erosive surfaces is imaged between the seafloor and Horizon 1; these surfaces are the signature of glacial advance–retreat cycles as outlined in Berger et al. (2008a), Willems (2009), and Worthington et al. (2008), with Horizon 1 being hypothesized as the first glacial advance to the edge of the modern continental shelf (Figs. F9, F11) at the start of the MPT. Proposed Site GOAL-15B will sample across Horizon 1 to test this hypothesized timing.

These seismic profiles also cross active faults on the slope (BT1, BT2) and abandoned faults beneath the current shelf (BT3, BT4, BT5; Fig. F10). Structures BT3, BT4, and BT5 are currently buried by more than ~1500 ms two-way traveltime of undeformed sediments and have gradually been rendered inactive since before the early Pleistocene deposition of Horizon 2 (Zellers, 1995; Berger et al., 2008a; Worthington et al., 2008). Site GOAL-15B is located adjacent to BT4 with the goal of determining the age of cessation of deformation as indicated by the absence of growth strata above Horizon 2 (Worthington et al., 2010). Lack of significant deformation in the sequences above Horizon 2 indicates that the underlying faults were abandoned prior to the MPT, possibly due to loading by sediments. On the forelimb of fold BT5, shelf-break seismic facies are present between 1.0 and 1.5 s two-way traveltime (TWTT), suggesting a previous depositional shelf break at this location subsequent to the MPT. Truncations of seismic strata, the presence of growth stratal packages on the backlimb, and overall geometry of BT5 provide evidence that this structure accommodated Yakutat-North America convergence as a growth fold in addition to acting as the former shelf edge. The overall architecture of the continental margin is thus the product of coupled depositional and tectonic processes.

At the southeastern end of the STEEP09 seismic profile (Fig. F10), two currently active faults (BT1 and BT2) are present on the continental slope exhibiting less burial than the structures on the shelf. Scarps ~750 m and ~300 m high associated with the active slope structures are visible on high-resolution bathymetry of the continental slope (Worthington et al., 2008). Site GOAL-17B is located just landward of BT2, which initiated after the Pliocene–Pleistocene transition, given the lack of growth strata observed below Horizon 3 (Fig. F10). The presence of two distinct sedimentary packages on BT2 is indicative of either a decrease in slope sedimentation during the early Pleistocene or an increase in deformation rate across BT2. Between Horizons 1 and 2, the angle of the observed growth strata becomes less pronounced, indicating a gradual decrease in fault growth rate during the early–mid-Pleistocene. Above Horizon 1, sediments are truncated by the anticline and are very slightly tilted toward the shelf, indicating minimal deformation on BT2 from ~1 Ma to the present. Taken together, the overall geometries of the upper and lower sedimentary packages within the Bering Trough suggest a fundamental shift in margin architecture from primarily tectonically influenced to primary depositionally influenced (Fig. F12).

Depositional basins on Khitrov Ridge along the continental slope west of the Bering Trough contain a sedimentary record of glacial–interglacial sedimentation overprinted by active tectonic deformation (Fig. F13). The seismic facies at this site are interpreted to represent contrasting hemipelagic (interglacial) and glacimarine (glacial maximum) cyclicity in sediment lithofacies. Processed CHIRP images, coincident with multichannel seismic (MCS) profiles, reveal that an upper postglacial transparent layer on the CHIRP profile that corresponds to the upper ~8 m of the sediment in Core EW0408-85JC dates to younger than 14.7 ka (Davies et al., 2011). The strong reflections in the CHIRP line and near the sediment/water interface to ~0.03 s (~8–25 m) in MCS Line GOA3201 likely represent the glacimarine sediments associated with the local LGM (likely 15–30 ka at an average accumulation of ~1.2 m/k.y., although with extremely high rates for brief intervals; Davies et al., 2011). It is hypothesized that the less reflective layered sediments in the MCS profile represent interglacials or interstadials, when the Bering Glacier terminus was in a greatly retreated position relative to the shelf break and ice-rafting of sediment was much reduced to absent. In contrast, the highly reflective intervals indicate times when ice rafting was active and there were higher accumulation rates of coarser glacigenic sediment. Active faulting is imaged in high-resolution seismic Profile GOA3101, showing surface deformation indicative of significant amounts of extension (Fig. F13). The faults in this extensional array, however, merge toward a common position, suggesting an underlying transtensional flower structure and a possible structural link between active structures within the offshore Yakutat block and the Aleutian Trench (Worthington et al., 2008) (Fig. F4).

The Neogene record of sedimentation on the Surveyor Fan, a terrigenous depocenter that comprises the majority of the Alaska Abyssal Plain (Fig. F7), comes from DSDP Leg 18 and ODP Leg 145 drilling. Leg 18 consisted of five sites drilled and interval-cored across the southwestern corner of the Surveyor Fan, the Aleutian Trench, and up the slope of the accretionary prism (Fig. F2) (Kulm, Von Huene, et al., 1973). In 1992, ODP Leg 145 occupied an additional site (887) in the far southwestern Gulf of Alaska on the Patton-Murray Seamounts (Fig. F2) (Rea, Basov, Janecek, Palmer-Julson, et al., 1993). Terrigenous turbidites, gravelly to diatomaceous mud, and claystone of the Surveyor Fan overlie marine chalk, barren clay, and basaltic basement of Pacific plate crust (Fig. F14) (Kulm and von Huene, 1973). The Chirikof and Surveyor Channel systems control present-day Surveyor Fan morphology and sediment distribution, but unlike other large deep-sea channels, they are not associated with a major fluvial system or submarine canyon (Fig. F7) (Ness and Kulm, 1973; Stevenson and Embley, 1987; Carlson et al., 1996; Reece et al., in press). The Surveyor Channel is >700 km long with three major tributaries (Ness and Kulm, 1973; Stevenson and Embley, 1987; Carlson et al., 1996; Reece et al., in press). Early studies divided the Surveyor Fan into two major sequences (Ness and Kulm, 1973; von Huene and Kulm, 1973; Stevenson and Embley, 1987), termed “upper” and “lower,” that were based on sedimentation rates and differences in acoustic facies imaged in two-dimensional (2-D) seismic reflection profiles. The boundary between the two sequences hypothetically represented a shift from a lower coarser grained facies to an upper finer grained facies possibly associated with Surveyor Channel inception and its control on fan sediment distribution during deposition of the upper sequence (Ness and Kulm, 1973; Stevenson and Embley, 1987). Reece et al. (in press) used reprocessed U.S. Geological Survey (USGS) and recently acquired high-resolution and crustal-scale seismic reflection data to correlate stratigraphic changes and fan morphology through time. In contrast to the previous interpretation of two seismic sequences within the fan, they recognized three sequences that are regionally extensive deposits likely related to increases in exhumation on land and regional response to global changes in climate (Figs. F14, F15). Sequences I and II exhibit laminated, laterally semicontinuous reflectors consistent with turbiditic deposition (Reece et al., in press) (Fig. F15). Sequence III is thinly laminated and contains reflectors that are laterally continuous, flatter, and smoother than those in the other sequences (Fig. F15). This seismic facies is especially prominent on the bathymetric high at Site GOA16-1A (Fig. F16). Stratal relationships at the sequence boundaries are highly variable and greatly influenced by basement topography and the presence of a mass transport deposit at the base of Sequence III in the northwestern portion of the fan (Fig. F15). Sequence II onlaps Sequence I in the in areas where Sequence I exhibits topography but is conformable in other locations. Sequence III onlaps sequence II in the proximal fan and downlaps it in the distal fan, where both sequences pinch out farther from the sediment source (Reece et al., in press). TWTT thickness (isopach maps) for the three sequences show a varying depositional history on the Surveyor Fan (Fig. F17) (Reece et al., in press). Sequence I depocenters are prevalent in topographic lows between basement highs, showing no significant spatial variation, which reflects infilling of preexisting Pacific plate topography. Deposits of Sequences II and III exhibit a distinct change in the locus of accumulation to shelf proximal depocenters that thicken into the Yakutat slope, with Sequence III thicker and covering a much larger area.

The correlation between seismic reflection profiles projected into the stratigraphy at DSDP Site 178 places tentative ages on sequence boundaries (Fig. F14). The Sequence I/II boundary occurs at ~330 m depth at Site 178, within a section of fine-grained sand to silty turbidites and interbedded diatomaceous ooze and mud with increasing diamictite upsection (Reece et al., in press; The Shipboard Scientific Party, 1973). The Sequence I/II boundary is placed at ~5 Ma, near the beginning of Glacial Interval A, based on 40Ar/39Ar dating of ash layers (Hogan et al., 1978) (Fig. F14). At 130 m depth, the Sequence II/III boundary lies within an interval of changing fan lithology. The section from 96 to 141 m contains abundant diamicton interbedded with silty clay and diatom-rich intervals, whereas the section from 141 to 280 m contains less diamicton, much more silty clay, and a fewer diatoms (The Shipboard Scientific Party, 1973; Reece et al., in press). The Sequence II/III boundary is tentatively dated ~1 Ma based on correlation with a magnetic polarity reversal identified at Site 178 (von Huene et al., 1973) (Fig. F14), making it coincident with the onset of Glacial Interval C. Both sequence boundaries are synchronous with a doubling in terrigenous sediment flux observed at ODP Site 887 at ~5 Ma and ~1 Ma, but no regional sequence boundary projected into Site 178 correlates with the onset of Glacial Interval B (Reece et al., in press).

Spatial variability in seismic facies and stratigraphy reveals the temporal evolution of fan stratigraphy. The thickening of Sequences II and III into the Yakutat terrane continental slope is evidence of the long-term connection of the Surveyor Fan to the Yakutat shelf (Stevenson and Embley, 1987; Reece et al., in press). Due to dextral transform motion of the Pacific plate and Yakutat terranes along the Queen Charlotte-Fairweather Fault, Surveyor Fan provenance likely varies from southern Coast Mountains sources in older fan sediment to St. Elias Range in younger fan sediment, similar to the sedimentary strata on the Yakutat microplate (Fig. F18) (Perry et al., 2009). The onset of Glacial Interval A led to a reorganization of fan sedimentation by spurring Surveyor Channel genesis (Reece et al., in press). The youngest phases of the Surveyor Channel created shelf-proximal depocenters at the base of the Yakutat terrane slope. Glacial Interval C, with its corresponding ice advances to the shelf edge, extended the Surveyor Channel across the Alaskan Abyssal Plain and markedly increased sediment flux to the Surveyor Fan. The Surveyor Channel system is a unique deepwater sediment delivery pathway because of its glacial source and trench terminus, both of which may contribute to the Surveyor’s ability to maintain a major channel and evade avulsion over long periods of time (Stevenson and Embley, 1987; Reece et al., in press).

Tectonic-climate interactions

The climatic influence on the width, structural style, and distribution of deformation in mountain belts is well established through analog, numerical, and analytical modeling studies based on critical wedge theory (Willett, 1999; Roe et al., 2006; Stolar et al., 2006; Whipple, 2009). Generally, an increase in erosional intensity through glacial or fluvial processes is predicted to accelerate rock uplift and decrease orogen width and relief (Whipple and Meade, 2004; Roe et al., 2006). In the Chugach-St. Elias mountains, the observed spatial patterns along the windward side of the orogen of increased exhumation rates and more deeply exhumed rocks, a relative deficiency of activity along the leeward side, and relatively shallow particle exhumation pathways are all indicative of a coupled tectonic-climate “wet prowedge” system (Fig. F6) (Berger et al., 2008a, 2008b; Meigs et al., 2008). Based on apatite (U-Th)/He thermochronometry, in conjunction with offshore seismic data and modeling results, Berger et al. (2008a) proposed that a structural reorganization of the St. Elias orogen occurred associated with the onset of Glacial Interval C and the MPT. The proposed structural reorganization includes initiation of a large-scale backthrust onshore and deactivation of faults in the offshore frontal portion of the wedge (Fig. F6). However, offshore faulting has remained active in the St. Elias Range, primarily associated with the Pamplona zone fold-and-thrust deformation front (Bruns and Schwab, 1983; Chapman et al., 2008; Plafker et al., 1994; Meigs et al., 2008). Recent modeling (Malavieille, 2010; Simpson, 2010) suggests that the extent of active faulting and folding in a frontal wedge is highly dependent on the details of mass redistribution by climate drivers and the magnitude of incoming sediment load.

However, the localization of exhumation solely along the windward equilibrium line altitude (ELA) as indicated by apatite (U-Th)/He thermochronometry (Berger et al. 2008b) has been questioned based on reinterpretation of bedrock samples and observations of detrital thermochronometry from glaciers draining from the St. Elias Range (Enkelmann et al., 2008, 2010). A purely tectonic explanation has been proposed for the observed patterns of exhumation. These patterns are interpreted not to reflect a temporal increase in exhumation rates over the Pleistocene, but rather are simply driven by a southward progression of the Yakutat fold and thrust belt, perhaps influenced by the westward arrival of the leading edge of the thicker sedimentary cover and crust of the Yakutat terrane (Meigs et al., 2008; Enklemann et al., 2009, 2010; Christenson et al., 2010; Worthington et al., submitted). Although, onshore thermochronometry data alone cannot uniquely distinguish between orographically versus tectonically controlled temporal changes in erosion (Meigs et al., 2008).

Gulf of Alaska paleoceanography

The Gulf of Alaska, located in the subarctic northeast Pacific Ocean, is an important component of Northern Hemisphere climate variability. Modern observations indicate linkages between ALPS atmospheric conditions, North Pacific circulation, and marine ecosystem productivity, yet paleoceanographic data describing past changes in this system are sparse. Previous paleoceanographic studies of the Gulf of Alaska have been limited to lower temporal resolution records retrieved from lower sediment accumulation rate locales that were sampled to avoid dilution by turbidites and remain above the regionally high carbonate-compensation depth (CCD) (Zahn et al., 1991; McDonald et al., 1999; Galbraith et al., 2007; Gebhardt et al., 2008). The late Pleistocene millennial-scale climate change typical of the North Atlantic has been inferred from earlier work in the Gulf of Alaska but has only recently been confirmed for the Expedition 341 region (Barron et al., 2009; Davies et al., 2011; Addison et al., submitted).

Core EW0408-85JC, collected on Khitrov Ridge at Expedition 341 proposed Site KB-1A, has provided a detailed record of the last deglacial period to the present based on lithofacies analyses, siliceous microfossils, organic matter composition and biogenic silica concentrations, redox-sensitive metals, and oxygen isotope data from planktonic and benthic foraminifers, all tied to a high-resolution age model (44 14C dates spanning 17,400 y) (Fig. F19). Surface water freshening likely due to glacial meltwater input began at 16,650 ± 170 cal y BP during an interval of relatively ice proximal sedimentation probably sourced from the Bering Glacier (Davies et al., 2011). A sharp lithofacies transition from diamict to laminated hemipelagic sediments indicates retreat of regional outlet glaciers onto land or into coastal fjords at 14,790 ± 380 cal y BP. A sudden warming and/or freshening of the Gulf of Alaska surface waters corresponds with this lithofacies transition and coincides with the Bølling interstadial of Northern Europe and Greenland. Cooling and/or higher surface water salinities returned during the Allerød interval, coincident with the Antarctic Cold Reversal, and continued until 11,740 ± 200 cal y BP, when onset of warming coincided with the end of the Younger Dryas (Davies et al., 2011). Two laminated opal-rich intervals (deglacial Bølling-Allerød [B-A] and the early Holocene) reveal discrete periods of enhanced water column productivity that likely correlate to similar features observed elsewhere on the margins of the North Pacific and are coeval with episodes of rapid sea level rise (Barron et al., 2009; Davies et al., 2011; Addison et al., submitted). Proxies for Holocene productivity are consistently higher than during the colder periods of expanded regional glacial activity. The finding of low productivity during the glacial and stadial intervals is consistent with previous findings from the open Gulf of Alaska, but are inconsistent with the hypothesis that such changes are the result of higher upper-ocean stratification during cold intervals (Sigman et al., 2004; Jaccard et al., 2005). The B-A interval is laminated and enriched in redox-sensitive metals, suggesting productivity-driven dysoxic-to-anoxic conditions in the water column. Enriched sedimentary δ15N ratios are present in these laminated intervals, suggesting a link between productivity and N cycle dynamics (Addison et al., submitted). Remobilization of iron from newly inundated continental shelves may have helped to fuel these episodes of elevated primary productivity and sedimentary anoxia (Davies et al., 2011; Addison et al., submitted).

A temporal correspondence exists between water column and sedimentation events observed in Core EW0408-85JC and global last deglacial–early Holocene climate. In addition to the Bølling interstadial diamict–laminated facies transition, an early termination of the Bølling-Allerød warm interval observed in Core EW0408- 85JC relative to the North Atlantic appears in a number of high-latitude North Pacific records (Davies et al., 2011). When compared to the δ18O ice core records from Antarctica (Ruth et al., 2007) and Greenland (Andersen et al., 2006; Rasmussen et al., 2006; Svensson et al., 2006), the planktonic oxygen isotope pattern of Core EW0408-85JC bears some similarity to both of these records, lending support to the idea that North Pacific climate records reflect both North Atlantic and Southern Ocean forcing (Mix et al., 1999; Davies et al., 2011).

Site survey data acquisition

In 2004, 1800 km of high-resolution MCS reflection profiles were collected in the Gulf of Alaska aboard the R/V Maurice Ewing (Fig. F2). The sources were dual 45/45 in3 GI (generator-injector [GI]) air guns with 3–5 m vertical resolution. For the lines over the Surveyor Fan, dual 105/105 in3 GI guns were used. Processing included trace regularization, normal move-out correction, bandpass filtering, muting, f-k (frequency-wave number) filtering, stacking, water-bottom muting, and finite-difference migration (Gulick et al., 2007; Berger et al., 2008a).

In 2008, ~1250 km of MCS reflection data and ~500 km of wide-angle seismic refraction data were acquired in the Gulf of Alaska as part of the St. Elias Erosion and Tectonics Project (STEEP; Fig. F2). The primary tectonic survey targets included the offshore Yakutat-North America deformation front, the offshore St. Elias orogenic wedge, known as the Pamplona zone fold-thrust belt, and the Dangerous River zone (DRZ). The survey also targeted the transition fault (results presented in Christeson et al., 2010) and the offshore zone of seismicity in the Pacific plate known as the Gulf of Alaska shear zone (Gulick et al., 2007; preliminary STEEP results presented in Reece et al., 2009). Primary stratigraphic targets included mapping the shelf and Surveyor Fan sediments to basement.

In 2011, ~3022 km of MCS reflection and ~600 km of wide-angle seismic refraction data were acquired with supporting sonobuoy refraction data. The purpose of the Gulf of Alaska seismic mapping program was to image the distal parts of the Surveyor and Baranof Fans and the northern Pacific plate crustal structure that these sediment bodies interact with. The program was funded through the USGS as part of the United States assessment of Article 76 of the Law of the Sea Convention. The survey design targeted two areas for assessment and a series of science targets to better understand the tectonic-sedimentary system within the Gulf of Alaska. The two regions where the Alaskan deep-sea fans cross the U.S. Exclusive Economic Zone are where the Surveyor Fan system interacts with Aja Fracture Zone, potentially forming a thickened deposit, and the distal Baranof Fan. Science targets include the Surveyor Channel and its interaction with the Aleutian subduction zone, the Aja Fracture Zone, the Chirikof Channel, and the channels of the Baranof Fan (Horizon, North, etc.). Expedition 341 Sites GOAL-18A, GOA18-1A, and GOA18-2A are in the imaged distal part of the Surveyor Fan to the north and west of Surveyor Channel.

Acquisition parameters for both the 2008 and 2011 seismic reflection data included a seismic source of 36 Bolt air guns with a total volume of 6600 in3 fired every 50 m. Receivers were located in an 8 km long solid streamer at 12.5 m spacing. Common midpoint spacing was 6.25 m. Seismic data processing included trace regularization, normal move-out correction, bandpass filtering, muting, stacking and frequency-wave number migration using Paradigm Geophysical FOCUS software (detailed seismic processing work flow in Worthington, 2010). Vertical resolution at the seafloor for this data set is 20–30 m.

Wide-angle reflection and refraction data were recorded along two profiles: STEEP01 is oriented west-east, crossing the offshore Yakutat microplate from near the Bering Glacier to east of the DRZ (Fig. F4); STEEP02 is oriented north-south and crosses the Yakutat shelf, the Transition fault, and the adjacent Pacific plate (Fig. F2). For Profile STEEP01, 25 ocean bottom seismometers (OBS) were deployed at ~15 km spacing across the profile. Data acquisition was simultaneous for the MCS and wide-angle data across STEEP01, with shot spacing of 50 m. Data were recovered from 21 instruments. Processing and survey details for Profile STEEP02 are presented in Christeson et al. (2010).

Locations of a 1975 USGS survey and a 1979 survey by Western Geophysical are shown in Figures F2 and F17. These seismic profiles provide ~30 m vertical resolution and reliably image up to 4 s TWTT of the subsurface, ~3.5 km at 1750 m/s seismic velocity. These profiles image major faults, fault-related folds, and unconformities, roughly illustrating stratigraphic and structural relationships along the margin (Bruns, 1983, 1985; Bruns and Schwab, 1983; Lagoe et al., 1993; Zellers, 1995). Processing of the 1970s USGS data used in Reece et al. (in press) and Worthington et al. (2010) included bandpass filtering, muting, normal move-out correction, and stacking. The 1980s USGS data processing included trace editing and balancing, muting, and bandpass filtering.

In addition to MCS data, high-resolution subbottom profiles and multibeam bathymetry exist for the Expedition 341 region. High-resolution subbottom profiles were collected during the 2004 R/V Maurice Ewing cruise using a Bathy 2000-P CHIRP subbottom profiler. High-resolution (5–20 m2) multibeam sonar data were collected at the proposed drill sites. At shallower depths (<800 m) a SIMRAD EM1002 midwater high-resolution multibeam sonar was used, and in deeper water (i.e., Site GOAL-16B), a STN ATLAS Hydrosweep DS-2 multibeam sonar was used. Additionally, in 2005, more than 162,000 km2 of high-resolution (~100 m2) multibeam sonar data were collected along the base of the Yakutat slope in the Gulf of Alaska in support of the United Nations Law of the Sea extended continental shelf investigation. Data were collected aboard the R/V Kilo Moana and post-processed at the University of New Hampshire Center for Coastal Studies (Mayer et al., 2005; Gardner et al., 2006). Vertical accuracy is ~0.3%–0.5% of the water depth. Additional high-resolution SeaBeam multibeam data were collected in 1988 by NOAA on the continental slope and proximal Surveyor Fan southwest of Kayak Island.

The supporting site survey data for Expedition 341 are archived at the IODP Site Survey Data Bank.