Spatial extent of rapid denudation in the glaciated St. Elias syntaxis region, SE Alaska

Authors


Abstract

[1] The ongoing collision of the Yakutat Terrane with the North American Plate formed the St. Elias Range in southeastern Alaska. Previously published detrital zircon fission track dating of sand-size material detected very rapid rock denudation within the syntaxis region of the St. Elias Range, but the size and spatial extent of the rapid exhumation are elusive due to the glacial cover that inhibits direct bedrock studies. Two possible mechanisms were suggested to explain the rapid denudation occurring underneath the Seward-Malaspina Glacier, which differ in their spatial extent of rapidly exhumed rocks. This study evaluates the spatial extent of rapid denudation at the St. Elias syntaxis region. We investigate the cobble-sized material of the Malaspina Glacier. Petrological classifications are used as a provenance tool and combined with zircon and titanite (U-Th)/He analysis to reveal cooling rates of the glacial detritus. In total, 1998 cobbles are classified in the field. Petrographic characterization is conducted on 91 thin sections and the (U-Th)/He ages of 59 cobbles are measured. Fourteen cobbles reveal young cooling ages of 3–2 Ma originating from various lithologies of undeformed and deformed magmatic and metamorphic rocks. Older age populations peak at 12, 27, and 36 Ma. The cooling ages and the mapped known extent of the different lithologies suggest that a larger then previously assumed area of the St. Elias syntaxis region experienced rapid exhumation. Rapid rock exhumation extends beyond the Contact Fault and comprises most of the Seward-Malaspina Glacier catchment.

1 Introduction

[2] Recent work has documented rapid denudation and deep-seated rock exhumation at the plate corners of convergent orogens such as the syntaxes of the Himalaya [e.g., Schneider et al., 1999; Zeitler et al., 2001a, 2001b; Koons et al., 2002, 2013; Vannay et al., 2004] and the St. Elias Mountains of Alaska [e.g., O'Sullivan et al., 1997; Pavlis et al., 2004; Berger et al., 2008a, 2008b; Enkelmann et al., 2008, 2009, 2010]. The St. Elias Range is of general interest because it gives insight into the early-stage processes of plate corner collision, which are mostly overprinted in the much more mature collision zones like the Himalaya. Another unique situation of the St. Elias syntaxis region is that the rocks of the downgoing Yakutat Terrane basement and the overriding North American Plate are exposed on land. The processes driving rapid rock exhumation in the St. Elias syntaxis region are currently not well understood because direct field observations are hindered by the thick ice coverage (Figures 1 and 2). Two models of rapid denudation and deep-seated rock exhumation have been proposed for the St. Elias syntaxis region: (1) rapid exhumation of a thin transpressive sliver along a reactivated suture zone (Contact Fault) at the syntaxial bend [Enkelmann et al., 2008; Spotila and Berger, 2010] and (2) the development of a coupling mechanism between efficient surface erosion and localized strain concentration at the syntaxis region that results in flow of weak middle and lower crustal material to the surface [Enkelmann et al., 2009; Koons et al., 2013]. The second model has been dubbed as a “tectonic aneurysm” and was proposed for the Himalaya and the Pamir [Pavlis et al., 1997; Zeitler et al., 2001a, 2001b; Koons et al., 2002, 2013; Thiede et al., 2004; Montgomery and Stolar, 2006; Kirby et al., 2007; Cao et al., 2013], and the European Alps [Garzanti and Malusa, 2008]. An incipient tectonic aneurysm has been suggested to explain the observation of young (3–2 Ma) zircon fission track (FT) cooling ages that were found in the Malaspina Glacier outwash material originating from the syntaxis of the St. Elias Range [Enkelmann et al., 2009, 2010; Koons et al., 2013].

Figure 1.

Simplified tectonic setting of southern Alaska and accreted terranes after Lahr and Plafker [1980] and Plafker et al. [1994]. Yakutat Terrane motion vector is taken from GPS measurements of Elliott et al [2010], Pacific Plate motion vector is from Plattner et al. [2007]. The white box marks the study area shown in Figures 2, 3, and 7. KIZ: Kayak Island Zone; PZ: Pamplona Zone; MTF: Malaspina Thrust Fault; ECF: Esker Creek Fault; TF: Totschunda Fault; COF: Totschunda-Connector Fault.

Figure 2.

ASTER GDEM (product of METI and NASA) of the study area. Color code highlights the occurring terranes within the Seward-Malaspina catchment. Grey shaded areas show the ice cover of the major glacial systems covering the St. Elias syntaxis region. Red line marks the catchment area of the Seward-Malaspina Glacier. Stars mark the seven counting and sampling locations of this study. Fault locations after Lahr and Plafker [1980], Plafker et al. [1994], and Bruhn et al. [2004]. Glacier cover shapefiles are used from GLIMS Glacier Database, provided by the National Snow and Ice Data Center [GLIMS and NSIDC, 2005]. BRF: Border Ranges Fault; CF: Contact Fault; CSEF: Chugach-St. Elias Fault; MTF: Malaspina Thrust Fault; ECF: Esker Creek Fault; FF: Fairweather Fault; COF: Totschunda-Connector Fault.

[3] One of the testable differences between the two suggested models is the different spatial extents of rapidly exhumed rocks underneath the Seward-Malaspina Glacier. While the transpressive sliver model envisions a narrow band of rapid exhumation along the Contact Fault, the tectonic aneurysm model predicts a larger region of rapid exhumation that is not restricted to the Contact Fault. This study investigates the spatial extent of the rapidly exhumed rocks in the Seward-Malaspina catchment by analyzing the cobble-sized detritus around the Malaspina Glacier lobe. We combine field observations, petrographic classifications on thin sections, and zircon and titanite (U-Th)/He thermochronometry to map out the rapidly exhumed rocks underneath the ice.

[4] Our results show a population of young cooling ages (~2.7 Ma) that occur in a large variety of lithologies including undeformed and deformed magmatic and metamorphic rocks. Based on the new results, we suggest that the region of rapid and deep-seated rock exhumation is not limited to the Contact Fault but extends over most geological units within the catchment. The results favor the model of a tectonic aneurysm. Additionally we suggest an alternative model of localized rock denudation that can also explain the observed data. In this alternative model, we propose a wider region of rapid denudation between the westward bending suture zone and the northwestward structural continuation of the Fairweather transform fault, the Connector Fault. Besides the tectonic implication for the denudation at the St. Elias syntaxis region, this study demonstrates that the approach of examining the cobble-sized detritus of glaciers is a valuable alternative to investigate the cooling record of an otherwise inaccessible bedrock region located underneath large alpine glacial systems.

2 Tectonic Setting

[5] The oblique subduction-collision of the Yakutat Terrane with the North American Plate formed the St. Elias Mountain Range that stretches along the southeastern margin of Alaska and the westernmost Yukon Territory (Figure 1). The St. Elias Range comprises mountain peaks of more than 5000 m adjacent to the North Pacific Ocean and is characterized by high precipitation rates and large glacial systems (Figure 2). The syntaxis of the St. Elias Range is the region where the main structural, geological, and geomorphic features bend 90° as the dextral transform motion along the Fairweather Fault transitions into predominantly thrust tectonics of the Chugach-St. Elias fold and thrust belt (Figure 1). The St. Elias syntaxis region is covered by high elevation (~1600 m) ice fields including the ~65 km long and ~20 km wide Seward Ice Field that is surrounded by the highest peaks of the mountain range (Figure 2). The Bagley, Malaspina, and Hubbard Glaciers are the main glacial systems that erode and transport material from the syntaxis region to the northwest, southwest, and southeast, respectively, and ultimately into the Pacific Ocean (Figure 2).

[6] The southern margin of Alaska is composed of several terranes that have been accreted to the North American Plate since Mesozoic times [Plafker et al., 1994]. The northernmost terrane is the Wrangellia Composite Terrane, which is subdivided into three distinct terranes (the Peninsular, Wrangellia, and Alexander Terranes) [Plafker et al., 1994]. The Wrangellia Composite Terrane mainly consists of Paleozoic to Jurassic crystalline basement, which is regionally overlain by Cretaceous fore-arc basin strata and associated with a variety of mafic intrusive rocks [Plafker et al., 1994; Trop and Ridgway, 2007].

[7] The Border Ranges Fault is the suture between the Wrangellia Composite Terrane and the Chugach Terrane (Figure 1). The Chugach Terrane is composed of volcaniclastic flysch and a Late Triassic to Cretaceous schist and mélange sequence associated with oceanic basaltic rocks. During the Eocene, the Chugach Terrane was intruded by the Sanak-Baranof plutonic belt and experienced metamorphism due to ridge subduction [e.g., Sisson et al., 1989, 2003; Hudson and Plafker, 1982; Pavlis and Sisson, 1995]. The ridge subduction and accompanying deformation formed the Chugach Metamorphic Complex that consists of a migmatitic gneiss zone located under the Seward Ice Field, surrounded by a schist zone composed of mainly biotite-plagioclase-quartz schist and a metabasite belt [Hudson and Plafker, 1982; Bruand et al., 2011; Gasser et al., 2011].

[8] The Contact Fault is the Paleogene suture between the Chugach Terrane and the Prince William Terrane and is located underneath the Seward-Bagley Ice Field. The Contact Fault is also known as the Bagley Fault due to its location underneath the Bagley Ice Field that drains east of the Seward Ice Field (Figure 2). In the St. Elias syntaxis region, the Prince William Terrane crops out as a thin band of Eocene deep-sea fan deposits, interbedded with pelagic sediments [Plafker et al., 1994]. The Chugach-St. Elias Fault is the suture between the North American Plate and the actively colliding Yakutat Terrane (Figures 1 and 2). The Yakutat Terrane is interpreted to be either overthickened oceanic crust or an oceanic plateau with increasing thickness from the northwest toward the southeast, based on seismic velocity profiles [Ferris et al., 2003; Christeson et al., 2010; Worthington et al., 2012]. In the eastern part of the terrane the Mesozoic Yakutat Group, a flysch and mélange sequence intruded by small granitoid plutons of Eocene age is overlaying the oceanic basement [Plafker, 1987]. Cenozoic cover strata of varying thickness (~0 km east of Seward Throat and >10 km toward the west) overlay the Yakutat basement. These rocks include the Kulthieth Formation (early Eocene to Oligocene), the Poul Creek Formation (late Eocene to late Miocene), and the Yakataga Formation (late Miocene to Pleistocene) [Plafker, 1987]. The Yakutat Terrane has been transported north along the Queen Charlotte-Fairweather transform [Plafker, 1987] resulting in flat-slab subduction since the Eocene [Finzel et al., 2011]. Subduction of an increasingly thicker and more buoyant southeastern part of the Yakutat Terrane resulted in subduction collision, which may have started in middle Miocene time [Plafker et al., 1994], or in late Miocene/Pliocene time (<7 Ma). The later onset of collision is based on paleoclimate reconstructions [White et al., 1997] and the occurrence of glacial sediments suggesting that the St. Elias Mountains were high enough to form alpine glaciers that reached to the coast [Lagoe et al., 1993]. The ongoing convergence of the Yakutat Terrane with North America is expressed by high seismic activity at the syntaxis region with several large earthquakes of M > 7 recorded in historic times [e.g., Plafker and Thatcher, 2008; Bruhn et al., 2012, and references therein].

[9] Low-temperature thermochronometric methods have been used to study the long-term denudation history of the Yakutat-North American collision zone. Previous denudation studies using fission track and (U-Th)/He analysis concentrated on the fold and thrust belt at the western part of the St. Elias Range [Spotila et al., 2004; Berger et al., 2008a, 2008b; Meigs et al., 2008; Enkelmann et al., 2008; Perry et al., 2009] and on the Fairweather Range that formed along the dextral Fairweather Fault located to the southeast [O'Sullivan et al., 1997; McAleer et al., 2009]. Recent work also revealed that the most rapid exhumation of deep-seated rocks occurs under the Seward-Malaspina Glacier at the St. Elias syntaxis region [Enkelmann et al., 2009, 2010]. The combination of fission track and U-Pb dating on detrital zircons from the Malaspina outwash revealed that Yakutat basement and Chugach Terrane rocks cooled rapidly at rates of 150–300°C/Myr due to denudation [Enkelmann et al., 2009]. However, the location, spatial extent, and mechanism of rapid rock denudation are not well known. This study aims to explore the spatial extent of the rapid denudation in the Seward-Malaspina Glacier catchment by investigating the cobble-sized detritus.

3 Methods

3.1 Sampling Strategy

[10] To overcome accessibility problems due to the heavy ice coverage of the St. Elias syntaxis region, we studied cobbles of different lithologies and identified cobble compositions to estimate their relative abundances at seven locations around the Malaspina lobe (Figure 2). It is important to note that the Malaspina Glacier comprises numerous folded ice-cored moraines that are recognizable by their different colors caused by varying abundances of different lithologies. Farther downstream, the Malaspina Glacier is characterized by a 6–10 km wide rim consisting of debris-covered ice that becomes increasingly vegetated (bushes and trees) toward the ocean, and a supraglacial lake and river system at the exterior part of the rim. The debris-covered ice of the Malaspina reaches to the Pacific, which prevents sampling of ice-free moraine material or material from the proglacial fluvial system. We sampled at various locations of the nonvegetated debris-covered ice around the entire lobe of the Malaspina Glacier and chose locations that were colored differently to collect representative lithologies present across the Malaspina Glacier lobe. Another reason to sample various locations was also to avoid landslide deposits [e.g., Benowitz and Layer, 2011] and to account for the large size of the glacier lobe and limited mixing of detritus during glacial transport. One sample location is located at the beach (MAL3). Similarly sized large cobbles (10–30 cm) like those from the moraine locations were analyzed at the beach location MAL3, the cobble size thereby excludes the possibility that those cobbles were transported long distances and originate from a different source than the Malaspina Glacier.

[11] Our sampling strategy was to (1) sample at least one cobble of each lithological group identified at the site to conduct petrographic analysis on thin sections and (2) sample additional cobbles from those lithologies that potentially contain zircon. As a consequence, our petrographic data analysis of the Malaspina cobbles can be used for a quantitative analysis. This is not the case for the zircon (U-Th)/He age analysis, which is strongly biased due to the fact that many lithological groups did not contain zircon (e.g., phyllites, fine-grained sediments, schists, some mafic rocks).

3.2 Petrographic Analysis

[12] The various lithologies present at each sample location were identified in the field and their abundances were counted at each sampling point (150 to 350 cobbles were counted at each location). In total, 1998 cobbles were counted. Following this, 91 thin sections were prepared and analyzed using a Zeiss Axiophot microscope.

3.3 (U-Th)/He Analysis

[13] In total, 79 cobbles were processed for zircon and titanite separation using a jaw crusher, sieves (350 µm), and standard magnetic, specific gravity, and heavy liquid separation procedures. We used (U-Th)/He analysis on zircons to identify the cobbles that experienced rapid cooling due to denudation to compare them with previously published 3–2 Ma detrital zircon FT results [Enkelmann et al., 2009]. Thermochronometric cooling ages represent the time elapsed since a rock cooled from an effective closure temperature to the surface. For the zircon FT and (U-Th)/He system, the closure temperatures are 210–290°C [Brandon et al., 1998] and 170–190°C [Reiners et al., 2004], respectively. Due to practical reasons of limited zircon abundances in some lithologies, we used zircon (U-Th)/He dating rather than zircon FT dating because only a few crystals are needed. In contrast several hundreds of grains are needed to prepare proper zircon FT mounts and conduct track etching. The zircon yield varied across lithologies and 24 of the 79 processed cobbles did not have zircons (mainly the fine-grained metasediments and mafic rocks). However, four of those non zircon-bearing rocks yielded titanite and in this case we used titanite (U-Th)/He dating, which has a similar closure temperature of 190–210°C [Reiners and Farley, 1999]. Sample location MAL5 was predominantly characterized by very fine-grained metasedimentary cobbles, and for these cobbles we did not yield any zircons. Therefore, there are no zircon (U-Th)/He ages reported from this sample location.

[14] Clear zircon grains without cracks were selected using a Leica M205C binocular microscope. The grain dimensions were measured for the calculation of the alpha-correction factor [Farley et al., 1996]. Afterward, the single grains were packed in Nb-tubes for analysis. In general, we analyzed 2 grains per sample, 119 grains in total from 59 Malaspina cobbles. Analyses were conducted with the Patterson helium-extraction line at the University of Tübingen, which is equipped with a 960 nm diode laser to extract the helium gas. Zircon grains were heated for 10 min at 20 A to extract the helium. Each grain was reheated to re-extract any remaining gas and verify that the grain was entirely degassed in the first heating. The re-extracts generally showed <1% of the first degassing. After helium analysis, the grains were sent to the University of Arizona at Tucson for U and Th measurements with an inductively coupled plasma–mass spectrometer.

[15] The four samples that did not contain zircons but suitable titanite for (U-Th)/He age determination were analyzed using the same analytical procedures as for zircons. Titanite has no regular grain shape and broken fragments cannot be discriminated from whole grains. For that reason, we packed the largest titanite fragments found in the sample and report the uncorrected age, based on the assumption that small fragments originate from the grain rim and large fragments from the grain core, as suggested by Stockli and Farley [2004].

4 Results

4.1 Petrographic Classification

[16] At every sampling spot (Figure 2), four to seven different lithologies were classified and counted in the field, and the raw data from the field are shown in Table 1. Based on the observations in the field and the classification of 91 thin sections, we regrouped the classified lithologies from the field and found six main lithologies for the whole Malaspina Glacier catchment (Figure 3). The lithologies can be assigned to different terranes that occur in the catchment based on existing knowledge of the bedrock geology of the St. Elias syntaxis region (Figure 4). These six lithological groups are as follows.

Table 1. Raw Data of Cobble Counting in the Field From 7 Sampling Locations Around the Malaspina Glacier Lobea
Sample locationCoordinatesLithologyRegrouped LithologyCountsAbundance (%)Total Counts
  1. a

    The cobbles classified as gabbro in the field were later found to contain quartz (visible in thin sections) and were therefore regrouped into the granitoid group.

MAL159°51′33.12″NGranitoidGranitoid6827 
 140°53′45.06″WMyloniteMylonite219 
  AmphiboliteAmphibolite14157 
  GreenschistAmphibolite156 
  GranuliteGneiss31 
      248
MAL259°46′37.38″NGranitoidGranitoid2516 
 140°47′22.26″WGneissGneiss3926 
  GreenschistAmphibolite32 
  PhyllitePhyllite, Micaschist4529 
  AmphiboliteAmphibolite4127 
      153
MAL359°42′2.10″NConglomerateConglomerate, Sedimentary Mixture4814 
 140°24′14.16″WGneissGneiss7522 
  PhyllitePhyllite, Micaschist16348 
  QuartziteConglomerate, Sedimentary Mixture72 
  GreenschistAmphibolite154 
  GabbroGranitoid113 
  GranitoidGranitoid237 
      342
MAL459°44′57.96″NGranitoidGranitoid4914 
 140°29′1.08″WGneissGneiss14342 
  PhyllitePhyllite, Micaschist12236 
  MyloniteMylonite93 
  GabbroGranitoid206 
      343
MAL559°52′15.00″NPhyllitePhyllite, Micaschist17655 
 140°6′51.72″WQuartziteConglomerate, Sedimentary Mixture124 
  RadiolariteConglomerate, Sedimentary Mixture144 
  BasaltAmphibolite9530 
  GneissGneiss52 
  SandstoneConglomerate, Sedimentary Mixture175 
      319
MAL659°49′1.50″NGneissGneiss26189 
 140°18′8.82″WMicaschistPhyllite, Micaschist217 
  AmphiboliteAmphibolite124 
  GabbroGranitoid10 
      295
MAL759°52′0.84″NGranitoidGranitoid4415 
 140°6′30.48″WGabbroGranitoid93 
  MetapeliteConglomerate, Sedimentary Mixture134 
  ConglomerateConglomerate, Sedimentary Mixture18161 
  PhyllitePhyllite, Micaschist269 
  QuartziteConglomerate, Sedimentary Mixture134 
  BasaltAmphibolite124 
      298
Total     1998
Figure 3.

The six main lithologies (color code is the same as in Figure 4) of the normalized Malaspina Glacier cobbles and their abundances in percent from cobble counting for all seven individual sampling locations, and for the whole Malaspina Glacier catchment (red outline). BRF: Border Ranges Fault; CF: Contact Fault; CSEF: Chugach-St. Elias Fault; MTF: Malaspina Thrust Fault; ECF: Esker Creek Fault; FF: Fairweather Fault; COF: Totschunda-Connector Fault.

Figure 4.

Correlation of the zircon (U-Th)/He cobble ages (circles), colored for their particular lithology and the inferred source terrane. Note the time axis has three breaks. The legend shows microscope images of representative thin sections of every lithology group, taken under a Zeiss Axiophot microscope with crossed analyzers and an image width of 4.6 mm. Color codes are the same as in Figure 3.

4.1.1 Granitoid Intrusive Rocks

[17] Typical granitoid rocks can be described as of two different types. One is on average made up of 40% quartz, 48% feldspar (potassium feldspar and plagioclase), and 12% biotite and is associated with accessory minerals, such as garnet and muscovite. Epidote and chlorite occur as alteration products from feldspar. The second granitoid type originates from a very undifferentiated melt and was therefore classified as a gabbro in the field. It consists of 42% feldspar (plagioclase, microcline) cumulate with 8% quartz, 12% orthopyroxene, 10% clinopyroxene, and 28% amphibole (hornblende) and constitutes a granitoid. The fraction of the opaque phase varies between samples. Often, dissolution or mineral reactions can be observed, especially within the mafic minerals and the unstable microcline grains. Both granitoid types show a medium- to coarse-grained fabric and, in some cases, slight deformation. The possible sources of the granitoid rocks are intrusions within the Wrangellia Terrane and the Chugach Terrane (Logan massif). In addition, it is possible that granitoid intrusions associated with Eocene magmatism are located under the Seward Glacier. Such intrusions are widely exposed west of the syntaxis and are therefore expected to occur also underneath the Seward Glacier.

4.1.2 Amphibolite and Metabasite

[18] The amphibolite rocks appear nearly monomineralic and are composed of ~90% amphibole (hornblende) and 10% plagioclase accompanied by a varying fraction of opaque phases. Grain sizes vary from medium- to very coarse-grained. Greenschist samples are composed of ~92% actinolite, 8% plagioclase, and accessory phases, such as tremolite and opaque phase (pyrite). Epidote can occur as an alteration product of feldspar, and a minor component of biotite occurs as a recrystallization phase from orthopyroxene. Fine- to medium-grained greenschist rocks show a clear foliation and, in contrast to amphibolite samples, a variety of deformation features. The amphibolites originate from the amphibolite facies metabasite belt that forms the southern margin of the Chugach Terrane (Figure 2) and stretches along the entire Chugach-St. Elias Range [Bruand et al., 2011].

4.1.3 Gneiss and Granulite

[19] (Ortho-) gneisses are composed of 37% quartz, 53% feldspar (potassium feldspar and plagioclase), and 10% biotite. In some cobbles, muscovite and garnet occur as accessory minerals in a small fraction. Chlorite and epidote, as alteration products from feldspar and biotite, are absent. The orthogneissic rocks are mainly coarse grained and are moderate to highly deformed and show foliation and recrystallization of quartz. The medium-grained granulite rocks are composed of 68% feldspar, 26% quartz, and 6% garnet. The gneisses are part of the Chugach Metamorphic Complex, located in the central part of the Chugach Terrane, which is overlain by the Seward Ice Field.

4.1.4 Phyllite and Schist

[20] Schistose and phyllitic rocks are composed of quartz (20%), feldspar (28%), and biotite (35%). Furthermore, accessory minerals such as muscovite (4%) and garnet (3%) occur as well as centimeter-sized crystals of andalusite in sample MAL6-22. Weathering products, such as epidote and chlorite, are observed in varying percentages. The fabric of the schist and phyllite samples is usually fine grained with a clear foliation (Figure 4). Phyllitic samples show a lower grade of alteration than the micaschist samples, which are commonly heavily weathered. The possible sources for the phyllites and schists can be the schist zone of the Chugach Metamorphic Complex, which surrounds the central gneissic core, as well as the Prince William Terrane, which is composed of metamorphic marine fan deposits and sediments (Figure 3).

4.1.5 Mylonite

[21] The sampled mylonitic rocks can be divided into two types based on the parent rock material. The first type originates from a gneissic or granitoid source rock. The observed mineral fractions are 28% quartz, 39% feldspar, 30% biotite, and 3% garnet. Sample MAL4-19 also contained pyroxene. The second mylonite type is based on a metabasite source rock. The mineral assemblage is made up of 70% amphibole (hornblende, actinolite) and 30% plagioclase. Sample MAL2-25 contained 35% quartz, 55% amphibole, and 15% feldspar. Small grains of epidote or biotite occur as alteration phases. Samples from the second category are heavily deformed and can be described as ultramylonitic rocks. Mylonites are formed in fault zones under ductile or brittle-ductile conditions. The most likely source for mylonites in the study area is the Contact Fault that represents the suture between the Chugach and the Prince William Terranes in the west and to the Yakutat Terrane in the eastern part of the Seward Ice Field (Figures 1 and 2). The metabasite belt of the southern Chugach Terrane most likely constitutes the parent rock of the second mylonite group.

4.1.6 Conglomerate and Other Sedimentary Rocks

[22] The conglomerate samples contain poorly sorted edge rounded to rounded feldspar (15%) and quartz (30%) grains (up to 4 mm), which are embedded into a dense, fine-grained matrix (55%; Figure 4). The other sedimentary samples encompass sandstones and metapelites as well as quartzite and radiolarite. Sample MAL7-22 is a pelagic sediment composed of 17% calcite, 10% feldspar, and 15% quartz grains interbedded in a very dense, fine-grained matrix (52%). The conglomerates and sedimentary mixture rocks likely originate from the Yakutat Terrane with the flysch and mélange units of the Yakutat Group that are exposed east of the Seward Throat and the unmetamorphosed Cenozoic cover sequences exposed west of the Seward Throat (Figure 2).

4.2 (U-Th)/He Results

[23] The zircon and titanite (U-Th)/He cooling ages range from 185 to 2 Ma, whereby most ages are between 40 and 2 Ma (Table 2 and Figure 4). The mean ages are only calculated for samples that reproduce with standard deviation <45%. There are several circumstances that can result in a large variation of (U-Th)/He ages between sample aliquots. For zircons, the most important are (1) zoning of U and Th causing an over/underestimation of the alpha correction factor, (2) parentless helium in fluid inclusions, (3) grain size variations, or (4) inherited helium due to partial resetting of metasedimentary rocks [Reiners, 2005; Fitzgerald et al., 2006]. This study is primarily focused on identifying those rocks that cooled extremely rapid, i.e., <3 Ma cooling age, rather than obtaining a precise cooling age for each cobble. Therefore, additional aliquots for those problematic samples were not measured. Instead, we took the younger single grain age as the cobble age as proposed by Fitzgerald et al. [2006]. The measured cooling ages from the four titanite samples range from 66 to 2 Ma (Table 2).

Table 2. Zircon and Titanite (U-Th)/He Ages for Malaspina Cobblesa
Sample4-He (mol)238-U (mol)235-U (mol)232-Th (mol)α-Uncorrected Age (Ma)Ftα-Corrected Age (Ma)Mean Age (Ma ± StD)Rock Type
  1. a

    Ft: α-ejection correction factor. Mean ages are calculated for samples with a standard deviation <45%.

  2. b

    Marks titanite samples. Labeling of samples: MAL(sampling location)-cobble number_(zircon/titanite grain number).

MAL1-2_11.484 × 10−134.864 × 10−123.597 × 10−141.023 × 10−1222.500.82727.19-Granitoid
MAL1-4_12.570 × 10−121.354 × 10−111.001 × 10−135.326 × 10−12133.300.744178.45  
MAL1-4_21.447 × 10−127.670 × 10−125.673 × 10−142.898 × 10−13142.840.742191.63185 ± 9.32Granitoid
MAL1-8_34.489 × 10−161.507 × 10−131.114 × 10−151.874 × 10−142.240.8212.73  
MAL1-8_45.249 × 10−162.301 × 10−131.702 × 10−158.583 × 10−141.630.8171.992.36 ± 0.52Mylonite
MAL1-12_17.412 × 10−142.230 × 10−121.650 × 10−142.965 × 10−1324.910.67436.92-Granitoid
MAL1-13_11.975 × 10−141.894 × 10−121.401 × 10−143.055 × 10−137.780.6312.34  
MAL1-13_23.367 × 10−134.355 × 10−123.221 × 10−141.864 × 10−1254.280.65782.41-Amphibolite
MAL1-14_12.430 × 10−149.469 × 10−127.004 × 10−141.316 × 10−111.510.7751.95  
MAL1-14_26.522 × 10−142.069 × 10−111.530 × 10−131.928 × 10−112.010.8442.392.17 ± 0.31Amphibolite
MAL1-19_12.141 × 10−159.649 × 10−137.137 × 10−151.127 × 10−131.670.8262.03  
MAL1-19_32.099 × 10−158.666 × 10−136.410 × 10−157.586 × 10−141.840.6812.70  
MAL1-19_44.843 × 10−152.287 × 10−121.692 × 10−141.660 × 10−131.610.7912.042.25 ± 0.38Gneiss
MAL2-4_15.560 × 10−133.367 × 10−112.490 × 10−138.562 × 10−1212.070.83414.47  
MAL2-4_24.985 × 10−133.661 × 10−112.708 × 10−138.184 × 10−1210.020.82112.2013.3 ± 1.60Granitoid
MAL2-5_12.860 × 10−138.531 × 10−126.310 × 10−141.312 × 10−1225.020.83429.98  
MAL2-5_22.724 × 10−135.250 × 10−123.883 × 10−141.059 × 10−1238.270.78948.4639.2 ± 13.1Granitoid
MAL2-10_13.697 × 10−142.608 × 10−121.929 × 10−148.897 × 10−1310.170.812.71  
MAL2-10_21.846 × 10−138.535 × 10−126.313 × 10−141.533 × 10−1216.060.77820.6316.7 ± 5.60Gneiss
MAL2-16_18.284 × 10−134.775 × 10−113.532 × 10−138.337 × 10−1212.900.85915.02  
MAL2-16_25.606 × 10−132.961 × 10−112.190 × 10−131.023 × 10−1113.570.8515.9615.5 ± 0.67Phyllite, Micaschist
MAL2-18_15.762 × 10−133.943 × 10−112.917 × 10−139.687 × 10−1210.700.85112.57  
MAL2-18_26.811 × 10−134.711 × 10−113.484 × 10−139.694 × 10−1210.680.86212.3912.5 ± 0.13Granitoid
MAL2-20_18.394 × 10−146.669 × 10−124.933 × 10−142.312 × 10−129.020.76611.78  
MAL2-20_21.222 × 10−131.110 × 10−118.210 × 10−148.771 × 10−138.360.73711.3411.6 ± 0.31Gneiss
MAL2-23_1b4.944 × 10−162.658 × 10−141.966 × 10−162.102 × 10−1412.19112.19  
MAL2-23_2b1.666 × 10−165.819 × 10−144.304 × 10−167.412 × 10−152.1512.15-Amphibolite
MAL2-24_1b5.965 × 10−164.666 × 10−143.451 × 10−161.629 × 10−149.160.85910.66  
MAL2-24_2b1.022 × 10−154.686 × 10−143.466 × 10−163.916 × 10−1414.180.88516.0213.3 ± 3.79Amphibolite
MAL2-26_11.631 × 10−138.147 × 10−126.026 × 10−141.253 × 10−1214.950.82118.21  
MAL2-26_21.156 × 10−136.409 × 10−124.740 × 10−148.883 × 10−1313.520.83416.2017.2 ± 1.42Gneiss
MAL2-27_11.343 × 10−147.834 × 10−135.794 × 10−159.867 × 10−1412.890.72917.67  
MAL2-27_35.937 × 10−146.709 × 10−124.962 × 10−141.199 × 10−126.580.7199.1513.4 ± 6.03Gneiss
MAL3-2_12.315 × 10−132.940 × 10−112.175 × 10−131.282 × 10−115.540.8016.92  
MAL3-2_23.322 × 10−135.330 × 10−113.942 × 10−132.320 × 10−114.390.7995.496.2 ± 1.01Conglomerate, Sedimentary Mixture
MAL3-6_12.124 × 10−141.183 × 10−128.750 × 10−152.431 × 10−1313.260.73618.01  
MAL3-6_24.144 × 10−141.330 × 10−129.837 × 10−151.797 × 10−1323.350.69533.5725.8 ± 11.0Gneiss
MAL3-7_11.088 × 10−122.560 × 10−121.893 × 10−141.276 × 10−11152.970.861177.41  
MAL3-7_21.669 × 10−124.788 × 10−113.541 × 10−135.891 × 10−1226.180.88429.61-Gneiss
MAL3-8_12.384 × 10−141.149 × 10−118.501 × 10−141.026 × 10−121.570.6862.29  
MAL3-8_22.368 × 10−149.668 × 10−127.151 × 10−141.863 × 10−121.820.6412.83  
MAL3-8_45.291 × 10−152.921 × 10−122.160 × 10−145.190 × 10−131.350.5112.642.59 ± 0.27Gneiss
MAL3-9_17.421 × 10−133.433 × 10−112.539 × 10−133.201 × 10−1216.360.78320.88  
MAL3-9_22.352 × 10−136.336 × 10−124.687 × 10−145.703 × 10−1328.080.75737.0629.0 ± 11.4Granitoid
MAL3-15_11.745 × 10−146.066 × 10−124.487 × 10−142.038 × 10−122.070.762.72  
MAL3-15_29.260 × 10−153.445 × 10−122.548 × 10−147.470 × 10−131.980.7432.672.69 ± 0.04Gneiss
MAL3-16_11.580 × 10−122.772 × 10−112.050 × 10−135.380 × 10−1242.090.75655.61  
MAL3-16_22.620 × 10−134.664 × 10−123.450 × 10−141.836 × 10−1239.770.77951.0053.3 ± 3.26Amphibolite
MAL3-18_11.259 × 10−133.434 × 10−112.540 × 10−138.711 × 10−122.680.8213.27  
MAL3-18_28.342 × 10−138.626 × 10−126.380 × 10−143.647 × 10−1267.880.82981.78-Amphibolite
MAL3-19_12.388 × 10−142.574 × 10−121.904 × 10−144.985 × 10−136.870.7139.64  
MAL3-19_24.643 × 10−144.350 × 10−123.218 × 10−146.587 × 10−137.980.64912.2911.0 ± 1.87Granitoid
MAL3-20_1b2.362 × 10−125.775 × 10−114.271 × 10−131.919 × 10−1129.36129.36  
MAL3-20_2b4.810 × 10−132.546 × 10−111.883 × 10−137.736 × 10−1213.66113.66-Granitoid
MAL3-21_1b2.857 × 10−112.254 × 10−101.667 × 10−121.278 × 10−1086.28186.28  
MAL3-21_2b3.599 × 10−113.034 × 10−102.244 × 10−121.363 × 10−0945.31145.3165.8 ± 29.0Granitoid
MAL4-3_12.267 × 10−125.594 × 10−114.137 × 10−138.819 × 10−1230.200.75639.91  
MAL4-3_24.532 × 10−139.902 × 10−127.324 × 10−141.869 × 10−1233.860.7942.8341.4 ± 2.06Granitoid
MAL4-5_16.790 × 10−133.551 × 10−112.627 × 10−135.978 × 10−1214.230.82117.33  
MAL4-5_24.520 × 10−132.527 × 10−111.869 × 10−139.966 × 10−1212.690.81815.5116.4 ± 1.29Gneiss
MAL4-6_19.085 × 10−133.037 × 10−112.246 × 10−134.499 × 10−1222.360.74230.11  
MAL4-6_21.548 × 10−125.042 × 10−113.729 × 10−139.327 × 10−1222.760.82727.51  
MAL4-6_36.083 × 10−132.061 × 10−111.524 × 10−132.162 × 10−1222.270.74629.8429.2 ± 1.43Gneiss
MAL4-8_12.105 × 10−147.508 × 10−135.553 × 10−151.072 × 10−1320.970.71829.19  
MAL4-8_21.234 × 10−133.337 × 10−122.468 × 10−146.021 × 10−1327.430.76435.8832.5 ± 4.73Gneiss
MAL4-9_11.317 × 10−132.924 × 10−122.163 × 10−146.292 × 10−1333.130.83839.51  
MAL4-9_29.972 × 10−143.541 × 10−122.619 × 10−146.795 × 10−1320.850.82825.1732.3 ± 10.1Gneiss
MAL4-10_11.065 × 10−143.857 × 10−132.853 × 10−151.400 × 10−1319.710.56734.71  
MAL4-10_32.281 × 10−147.011 × 10−135.185 × 10−159.990 × 10−1424.340.68435.5435.1 ± 0.59Gneiss
MAL4-11_11.744 × 10−141.263 × 10−129.340 × 10−152.625 × 10−1310.200.75913.43  
MAL4-11_21.015 × 10−133.267 × 10−122.416 × 10−144.517 × 10−1323.270.77130.16-Gneiss
MAL4-16_16.255 × 10−133.866 × 10−112.860 × 10−134.940 × 10−1212.160.8115.00  
MAL4-16_27.650 × 10−133.697 × 10−112.734 × 10−138.776 × 10−1215.180.81418.6416.8 ± 2.57Gneiss
MAL4-19_11.023 × 10−143.276 × 10−132.423 × 10−156.791 × 10−1316.420.82519.89  
MAL4-19_22.922 × 10−152.400 × 10−131.775 × 10−152.173 × 10−137.810.8938.75-Mylonite
MAL4-20_12.204 × 10−127.832 × 10−115.793 × 10−131.138 × 10−1121.040.79426.49  
MAL4-20_21.234 × 10−125.313 × 10−113.929 × 10−134.796 × 10−1217.590.82221.3923.9 ± 3.61Gneiss
MAL4-21_19.639 × 10−144.518 × 10−123.341 × 10−141.491 × 10−1215.340.78119.64  
MAL4-21_28.831 × 10−144.013 × 10−122.968 × 10−141.044 × 10−1216.060.7521.4020.5 ± 1.25Mylonite
MAL4-22_26.364 × 10−138.345 × 10−126.172 × 10−146.940 × 10−1249.440.76864.29-Phyllite, Micaschist
MAL6-5_12.117 × 10−136.299 × 10−124.659 × 10−148.550 × 10−1325.170.83929.99  
MAL6-5_22.153 × 10−134.959 × 10−123.668 × 10−141.059 × 10−1333.340.75943.8836.9 ± 9.82Gneiss
MAL6-7_11.485 × 10−133.769 × 10−122.788 × 10−143.407 × 10−1329.810.82236.24  
MAL6-7_26.852 × 10−134.161 × 10−113.077 × 10−131.568 × 10−1111.730.84313.91-Gneiss
MAL6-8_11.093 × 10−133.652 × 10−122.701 × 10−143.525 × 10−1322.620.7828.98  
MAL6-8_21.458 × 10−133.341 × 10−122.471 × 10−148.256 × 10−1331.880.75842.0235.5 ± 9.22Gneiss
MAL6-12_12.333 × 10−137.234 × 10−125.351 × 10−141.425 × 10−1223.840.80629.57  
MAL6-12_21.068 × 10−134.263 × 10−123.153 × 10−147.413 × 10−1318.620.80623.0926.3 ± 4.58Gneiss
MAL6-15_15.702 × 10−132.268 × 10−111.677 × 10−133.274 × 10−1218.810.80223.44  
MAL6-15_21.585 × 10−124.932 × 10−113.648 × 10−131.535 × 10−1123.180.85727.0425.2 ± 2.54Phyllite, Micaschist
MAL6-16_12.242 × 10−132.068 × 10−111.529 × 10−135.847 × 10−127.880.7979.88  
MAL6-16_23.538 × 10−132.710 × 10−112.004 × 10−135.099 × 10−129.680.8211.8110.8 ± 1.36Phyllite, Micaschist
MAL6-21_17.646 × 10−145.851 × 10−124.327 × 10−141.791 × 10−129.450.78412.05  
MAL6-21_23.500 × 10−132.402 × 10−111.777 × 10−135.956 × 10−1210.660.81413.1012.6 ± 0.74Phyllite, Micaschist
MAL6-22_11.573 × 10−133.872 × 10−122.864 × 10−141.239 × 10−1229.240.75938.49  
MAL6-22_21.404 × 10−134.274 × 10−123.161 × 10−142.068 × 10−1222.860.76729.7934.1 ± 6.15Phyllite, Micaschist
MAL6-23_12.773 × 10−121.620 × 10−101.199 × 10−122.459 × 10−1112.790.88414.46  
MAL6-23_29.851 × 10−135.032 × 10−113.722 × 10−131.156 × 10−1114.380.87116.5115.5 ± 1.45Gneiss
MAL6-24_11.646 × 10−131.075 × 10−117.949 × 10−142.018 × 10−1211.360.75315.08  
MAL6-24_21.019 × 10−136.894 × 10−125.099 × 10−141.781 × 10−1210.800.77313.9714.5 ± 0.78Gneiss
MAL7-2_13.341 × 10−141.190 × 10−118.804 × 10−148.050 × 10−121.880.7582.48  
MAL7-2_24.302 × 10−141.041 × 10−117.700 × 10−148.274 × 10−122.710.6873.943.21 ± 1.03Granitoid
MAL7-3_11.652 × 10−145.514 × 10−124.078 × 10−147.101 × 10−132.250.7872.86  
MAL7-3_27.265 × 10−152.242 × 10−121.658 × 10−143.565 × 10−132.420.7653.163.01 ± 0.21Amphibolite
MAL7-6_15.505 × 10−134.907 × 10−113.629 × 10−132.716 × 10−117.710.7889.78  
MAL7-6_21.040 × 10−131.311 × 10−119.698 × 10−146.097 × 10−125.550.7867.068.42 ± 1.92Conglomerate, Sedimentary Mixture
MAL7-8_13.416 × 10−131.910 × 10−111.413 × 10−137.691 × 10−1212.670.74916.91  
MAL7-8_26.039 × 10−149.270 × 10−126.857 × 10−145.383 × 10−124.450.7166.22-Conglomerate, Sedimentary Mixture
MAL7-14_18.246 × 10−153.524 × 10−122.606 × 10−141.882 × 10−121.620.7132.26  
MAL7-14_21.012 × 10−143.909 × 10−122.891 × 10−141.950 × 10−121.800.7092.542.40 ± 0.19Granitoid
MAL7-15_14.593 × 10−141.270 × 10−119.393 × 10−149.550 × 10−122.390.7593.15  
MAL7-15_21.204 × 10−132.458 × 10−111.818 × 10−131.726 × 10−113.270.7534.343.74 ± 0.84Granitoid
MAL7-17_13.488 × 10−149.162 × 10−126.776 × 10−146.014 × 10−122.560.7663.35  
MAL7-17_21.087 × 10−144.770 × 10−123.528 × 10−141.955 × 10−121.610.7292.212.78 ± 0.80Granitoid
MAL7-19_11.043 × 10−144.291 × 10−123.174 × 10−141.925 × 10−121.710.7072.41  
MAL7-19_22.088 × 10−146.988 × 10−125.169 × 10−144.139 × 10−122.040.6992.922.67 ± 0.35Granitoid
MAL7-20_15.306 × 10−141.585 × 10−111.172 × 10−135.812 × 10−122.390.8882.69  
MAL7-20_22.028 × 10−147.457 × 10−125.515 × 10−143.433 × 10−121.910.8392.272.48 ± 0.30Granitoid

[24] The mean cobble (U-Th)/He ages are plotted in a histogram along with a kernel density estimation calculated using DensityPlotter [Vermeesch, 2012] (Figure 5). Three cobble ages of >60 Ma were not plotted but are shown in Figure 4. Figure 5 reveals four dominant age peaks that together comprise 98% of the entire data set. The main age populations peak at 2.7 ± 0.03, 12.2 ± 0.01, 27.1 ± 0.5, and 35.8 ± 0.5 Ma and a smaller peak at 53.2 ± 3.3 Ma. The lithologies of the 2.7 Ma peak are granitoids, gneiss, amphibolites, and mylonites. The 12.2 Ma peak is a mixture of different lithologies, with a predominant fraction of gneiss and granitoids and a subordinate fraction of amphibolites, phyllites, mylonites, and sedimentary mixture rocks. The 27.1 and 35.8 Ma peaks are dominated by gneiss, granitoid, and phyllites.

Figure 5.

Histogram of 56 zircon/titanite cobble ages (2 Myr bin size); the fractions of the different lithologies are stacked on top of each other. The lithology color code is the same as in Figures 3 and 4. The black line is a kernel density estimation (KDE), calculated with the standard deviation and a bandwidth of 1.43 in DensityPlotter [Vermeesch, 2012]. Peaks were calculated with the implied mixed population determination tool, which gives the peak ages. Black circles display the data points, n is the amount of zircon/titanite cobble ages per lithology. One gneiss with 66 Ma, one phyllite with 64 Ma, and one granitoid with 185 Ma zircon (U-Th)/He ages plot outside of the x axis and are not shown.

[25] The (U-Th)/He ages of the low-grade or unmetamorphosed sedimentary rocks are younger than the depositional ages and are therefore interpreted as reset ages. A reset (U-Th)/He cooling age means that postdepositional burial was sufficient to heat the sediment sample well above the closure temperature (>180–200°C), which causes complete resetting of the cooling age signal due to the diffusive loss of helium. The result is a cooling age that is younger then the sediment deposition age, reflecting the time of cooling after the maximum heating. If the sediment was heated to temperatures around the closure temperature, only some of the accumulated helium will be lost, but some will also be retained. Those samples are considered partially reset, resulting in ages that are younger and older than deposition age. An unreset sample is a sediment sample that has not seen high temperatures and the cooling ages predate the time of deposition.

5 Discussion

5.1 Cobble Provenance and Cooling Ages

[26] The zircon (U-Th)/He cooling ages of the Malaspina cobbles do not provide provenance information. To address this problem, we use the petrology as a provenance tool to allocate cooling ages to certain geological units in the catchment. Knowledge about the geology in the catchment is of course limited due to the ice coverage and is based on observations made on the surrounding ice-free mountain ridges. For this study, we base our interpretation on the geologic map of the U.S. and Canadian Geological Survey [Journeay et al., 2000; Richter et al., 2006], on published mapping results of Gasser et al. [2011], and our own field observations on both the Alaska and the Yukon sides of the range. We assume that subglacial geology is the same as the ridge-top geology, which is supported by the fact that observed cobble lithologies from the Malaspina Glacier detritus can be correlated with observations from exposed bedrock in the catchment region. Granites, gneisses, and granulites occur in the center of the Chugach Metamorphic Complex and are surrounded by phyllites and micaschists [Hudson and Plafker, 1982]. The Chugach Terrane rocks comprise 33% of the entire Seward-Malaspina catchment. Amphibolites and metabasites are typical lithologies reported from the ultrabasite belt that forms the southern boundary of the Chugach Terrane [Hudson and Plafker, 1982; Bruand et al., 2011; Gasser et al., 2011]. Phyllite and schists are also typical rocks of the Prince William Terrane that forms only a small band (6%) in the upper part of the Seward Throat. Granitoids intruded into metasediments were also mapped at the northern ridges surrounding the Seward Ice Field and belong to the Wrangellia Composite Terrane (4% of catchment). The mylonites originate from the Paleogene suture zone (Contact Fault) between the Chugach Terrane in the north and the Prince William and Yakutat Terranes in the south and have been observed on nunataks within the Seward Ice Field [Bruhn et al., 2004] and along the Fairweather Fault farther east. We do not expect mylonites from the Border Ranges Fault, since this suture zone occurs only in small parts of the mountain ridges forming the northernmost boundary of the Seward Ice Field, and we did not observe mylonites during field studies in this region. Metasedimentary flysch and mélange rocks are typical for the Cretaceous Yakutat Group and sedimentary rocks for the Cenozoic cover of the Yakutat Terrane, which both underlie the Seward Throat and the Malaspina Glacier (Figure 2).

[27] The oldest (185 ± 9.3 Ma; MAL1-4) zircon (U-Th)/He cooling ages are typical for the Wrangellia Composite Terrane granitoids that were dated north of the study area [O'Sullivan and Currie, 1996; Enkelmann et al., 2010]. In addition, Jurassic granitoid samples of the Wrangellia Terrane exposed at the southern flanks of Mt. Logan, and part of the Seward catchment, yielded zircon FT ages of 44–41 Ma [O'Sullivan and Currie, 1996]. Two granitoid cobbles yielded similar zircon (U-Th)/He ages (41–39 Ma; MAL2-5 and MAL4-3; Table 2) and may also originate from the Logan massif. The rocks of the Chugach and southern Wrangellia Terranes were affected by the Eocene ridge subduction and associated metamorphism that was complete in this region by ~50 Ma [Pavlis and Sisson, 1995; Sisson et al., 2003]. We dated one amphibolite sample (MAL3-16) that falls into that time span and may reflect the change in the geothermal gradient due to ridge subduction.

[28] The zircon cooling ages from the phyllite and mica schist samples as well as some gneissic and granitoid samples range between 35 and 10 Ma and fit into the time span of the Chugach Terrane denudation described farther west and associated with deformation induced by Yakutat flat-slab subduction [Enkelmann et al., 2008, 2010; Finzel et al., 2011]. The granitoid rocks of the Chugach Terrane are not exposed within the Malaspina catchment, but it is likely that granitoid intrusions, associated with the Eocene magmatism, are located under the Seward Ice Field, similar to those exposed farther west in the Chugach Terrane. The conglomerate and sedimentary-mixture cobbles originate from the Yakutat Terrane and, in particular, the Yakutat Group that underlay the Seward Throat and the Malaspina Glacier. The zircon (U-Th)/He cooling ages of the sedimentary cobbles range between 8 and 6 Ma, which is remarkable given that most of the zircon (U-Th)/He and FT cooling ages reported from the Yakutat Terrane are either unreset or partially reset and are therefore much older (usually >30 Ma) [Berger et al., 2008a; Meigs et al., 2008; Enkelmann et al., 2010]. The young zircon (U-Th)/He cooling ages of the sedimentary mixture of cobbles and a published 3.5 Ma zircon (U-Th)/He sample from the bedrock within the Seward Throat [Enkelmann et al., 2010] indicate that Yakutat Terrane rocks are exhumed rapidly not only in the Seward Throat but also from depths that exceed those exhumation depths of the surrounding mountain ridges. The cooling ages of the mylonite cobbles range from ~20 to 2 Ma (Table 2), suggesting that denudation along the Contact Fault is not uniform or indicating exhumation along undocumented faults or the Border Ranges Fault. It is remarkable that even cobbles of undeformed amphibolites, undeformed granitoids, and gneisses yield very young cooling ages of 3–2 Ma (Figure 4 and Table 2).

[29] In summary, the sampled cobbles originate from the entire Seward-Malaspina catchment and comprise a wide range of zircon (U-Th)/He ages. Some of the older cooling ages have been recognized previously in exposed bedrock of the syntaxis region, but some cooling ages must originate from bedrock underneath the ice, as they have been documented before only in sand-size material of the Malaspina detritus.

5.2 Cooling Signals of the St. Elias Syntaxis

[30] To provide an overview of the cooling signal for the entire Seward-Malaspina catchment, we plotted the mean zircon (U-Th)/He ages in a histogram (Figure 5). An interpretation of the four major cooling age peaks at 2.7, 12.2, 27.1, and 35.8 Ma and a subordinate peak at 53.2 Ma is given in this section. The early Eocene peak (~53 Ma) probably represents cooling following ridge subduction and metamorphism of the Chugach Metamorphic Complex accompanied with a change in geothermal gradient. The four younger cooling age peaks can be related to the subduction of the Yakutat slab and subsequent terrane accretion. Denudation in southern Alaska started in the late Eocene to Oligocene with the activation of the Fairweather Fault that lead to the northward transport of the Yakutat Terrane and induced flat-slab subduction in southern Alaska [Plafker et al., 1994; Enkelmann et al., 2008, 2010; Finzel et al., 2011]. The early Oligocene age peak at 27.1 Ma probably also relates to the same process of flat-slab subduction and the northward transport of the Yakutat Terrane along the Fairweather transform. However, these two peaks need to be treated carefully due to the bias of our results related to the presence of some nonzircon-bearing lithologies. It is possible that the two peaks actually belong to one wide population that peaks ~30 Ma and is interpreted to reflect denudation due to Yakutat subduction. A large age population in our data set is composed of Miocene cooling ages with a peak at ~12 Ma (Figure 5). Interestingly, this cooling age signal occurs in all lithological groups including rocks that have been assigned to the unsubducted thicker part of the Yakutat Terrane, which is characterized by increasing crustal thickness from ~15 km in the west to ~30 km in the east [Ferris et al., 2003; Christeson et al., 2010, Worthington et al., 2012]. The simplest interpretation of this age population is that it results from denudation associated with Yakutat Terrane collision and identifies a minimum age for the onset of collision. The ongoing collision of the increasingly thickened eastern part of the Yakutat Terrane induced deformation and denudation at the southeastern Alaska margin as well as the Yakutat Terrane itself since at least middle Miocene time. The time when collision started in the St. Elias Range was long debated and is complicated by the transitional nature of the long-lasting subduction of a wedge-shaped buoyant and increasingly thickening Yakutat Terrane. Plafker et al. [1994] suggested that subduction collision started in the middle Miocene, but other studies suggested a later time at the end of the Miocene. White et al. [1997] investigated the vegetation and climate change over the past 18 Myr and suggest uplift of the St. Elias Mountains after 7 Ma. Planktic foraminiferal and paleomagnetic data from the Yakataga Formation suggest that the St. Elias Mountains were high enough to host alpine glaciers starting at the latest Miocene 6–5 Ma [Lagoe et al., 1993].

[31] The youngest cooling age peak of 2.7 Ma comprises 14 cobble ages and is thus a dominant cooling signal of the young rapid denudation process occurring under the Seward-Malaspina Glacier. This result confirms the 3–2 Ma zircon FT age populations found in the sand size material of the Malaspina Glacier [Enkelmann et al., 2009] and supports the idea of deep-seated rapid denudation due to interactions of tectonics and surface processes. Interestingly, the stacked histogram shows that these young ages occur in four different lithologies (undeformed granitoids, gneiss, amphibolite, and mylonite) that are typical of the Chugach Terrane and Wrangellia Composite Terrane (Figures 4 and 5 and Table 2) and that cover a large region within the glacial catchment (Figure 2).

5.3 Denudation Intensity

[32] The intensity of denudation can be inferred from the zircon (U-Th)/He cooling ages but also from our quantitative classification of cobble lithologies in the field. To convert cooling ages into denudation dates, the geothermal gradient must be estimated and a continuous denudation rate assumed. Several different Pliocene geothermal gradients were suggested for the St. Elias Range, which range from 25 to 50°C/km [e.g., O'Sullivan and Currie, 1996, Berger et al., 2008a, 2008b, Meigs et al., 2008]. Assuming a low geothermal gradient of 25°C/km and a closure temperature of 170–190°C, the rocks composing the 2.7 Ma zircon (U-Th)/He age population were exhumed from 6.8 to 7.6 km depths since 2.7 Ma, which would amount to denudation rates of 2.5–2.8 mm/yr. Assuming a higher geothermal gradient of 50°C/km, the closure temperature depths would be much shallower at 3.4–3.8 km, suggesting denudation rates of 1.3–1.4 mm/yr. The geothermal gradient underneath the Seward-Malaspina Glacier is not known, but is probably elevated as a result of rapid erosion that causes an upward bending (compression) of the isotherms. Using a one-dimensional denudation model and assuming a steady state thermal field, the 2.7 Ma zircon (U-Th)/He ages suggest denudation rates of 2 mm/yr [Reiners and Brandon, 2006]. Similar denudation rates of 2.0–2.7 mm/yr are suggested by the 3–2 Ma zircon FT age peaks found in the sand-size fraction of the Malaspina Glacier detritus [Enkelmann et al., 2009].

[33] Using the abundances of different lithologies observed in the field, we present a relative comparison between the denudation intensities of the Chugach and Yakutat Terranes within the catchment. The normalized number of cobbles counted around the Malaspina lobe (Table 1) was divided by the percent area of the assigned source terrane within the Malaspina catchment (Figure 6). The Chugach Terrane occupies ~33% of the Malaspina catchment and the Yakutat Terrane ~57%. Because only 10% of the catchment area belongs to the Prince William and Wrangellia Terranes, we do not consider them for this comparison of denudation intensity between the Chugach and Yakutat Terrane. Of the cobbles counted in the field, 1430 (319 granitoids, 614 gneisses, 459 amphibolites, and 38 mylonites) could originate from the Chugach Terrane, while only 340 (sedimentary mixture rocks, conglomerate) probably originate from the Yakutat Terrane. Figure 6 shows that the Chugach Terrane rocks produce cobble-sized detritus more than two times faster relative to the Yakutat Terrane rocks. This becomes even more significant when considering that the Chugach Terrane consists of hard crystalline and higher-grade metamorphic rocks, while the Yakutat Terrane is mainly composed of softer (meta-) sedimentary rocks. The counting observations might be influenced by the higher propensity of the Chugach Terrane rocks to produce cobble-sized clasts; however, the Chugach Terrane cobbles were transported over a longer distance than the Yakutat Terrane cobbles (mean transport distance ~ 92 km compared to ~ 57 km, respectively). Based on the difference of the transport distance and lithology, and the assumption that eroded basal bedrock stays in the basal zone of the glacier where comminution is high, we assume that the comminution effect balances each other [Boulton, 1978]. The comparison of the relative denudation rate between the Yakutat and the Chugach Terranes supports the fact that a fast exhuming region exists under the Seward Ice Field, as previously stated in Enkelmann et al. [2009].

Figure 6.

Comparison of the relative denudation rate of the Yakutat Terrane and the Chugach Terrane calculated by dividing the amount of the normalized cobbles counted by point counting analysis by the percentage of the area of the assigned terrane within the Seward-Malaspina Glacier catchment area.

5.4 Tectonic Implication

[34] Two models have been proposed explaining the rapid and intense denudation at the St. Elias syntaxis. Model 1 suggests rapid rock denudation occurring in a thin transpressional sliver that extrudes along the Contact Fault, and model 2 suggests denudation due to the development of a localized strong feedback between surface processes and tectonics as suggested in the tectonic aneurysm model (Figure 7). In the following, we evaluate these two models with respect to previously published results and our new data.

Figure 7.

Landsat image of the St. Elias Syntaxis region showing the area affected by rapid denudation based on the two suggested models. The terrane color code is the same as in Figure 2. The black arrows show the transport direction of detritus with young cooling signal. Right side shows sketches of the two suggested models for rapid denudation in the St. Elias syntaxis. Model 1: denudation of a transpressive sliver (modified after Spotila and Berger [2010]). Model 2: denudation of a bigger area of localized rapid denudation according to the tectonic aneurysm model of Zeitler et al. [2001a] (modified after Koons et al. [2013]). CF: Contact Fault; CSEF: Chugach-St. Elias Fault; MTF: Malaspina Thrust Fault; ECF: Esker Creek Fault; FF: Fairweather Fault; COF: Connector Fault.

[35] The Contact Fault is the old suture between the Chugach and the Prince William Terranes and is located underneath the >300 km long Seward-Bagley Ice Field that covers the orogenic spine (Figures 1 and 2). Bruhn et al. [2004] suggested that this suture zone has been reactivated in the late Cenozoic as a transpressive dextral fault forming the northwestward continuation of the Fairweather Fault. Evidence for active reverse faulting underneath the Seward Ice Field comes from seismic studies of locating the rupture zone of the M 7.4 St. Elias earthquake in 1979 [Estabrook et al., 1992], and the occurrence of a shallow, fault-bounded, bedrock ridge underneath the ice [Bruhn et al., 2012]. Rapid denudation of a thin transpressional sliver along the Contact Fault was first suggested by Enkelmann et al. [2008] based on a small population of ~6 Ma zircon FT ages found in the Yakataga Formation that were deposited 4–3 Ma. Zircon U-Pb crystallization ages of those young zircon FT grains suggested the Chugach Metamorphic Complex as the source, which supports a rapidly exhuming sliver along the reactivated suture zone [Enkelmann et al., 2008]. The transpressional sliver model was suggested by Spotila and Berger [2010] based on a review of bedrock cooling ages from the entire Chugach-St. Elias Range. Although the region of the Seward Ice Field experienced the highest uplift rates in the syntaxial bend, Spotila and Berger [2010] estimated that rock denudation rates are <1 mm/yr and only a very small sliver underneath the Seward Ice Field is exhuming much faster, sourcing the 3–2 Ma zircon FT ages found in the modern Malaspina sand-size detritus by Enkelmann et al. [2009]. Spotila and Berger [2010] estimated the areal size of the rapidly exhuming sliver underneath the Seward Ice Field to be ~100 km2, which would approximate to ~1.7 km width, given the ~60 km length of the fault (Figure 7). Based on the fact that this estimated area is 10–50 times smaller than the rapid denudation areas at the Himalayan syntaxes, the development of a strong feedback mechanism as suggested by the tectonic aneurysm model was considered as unlikely [Spotila and Berger, 2010].

[36] Model 2 envisions a larger region of intense rock denudation that centers on the St. Elias syntaxis and comprises most of the Seward Ice Field (Figure 7). This second model suggests a larger pop-up structure similar to those observed at the Himalayan syntaxes [e.g., Zeitler et al., 2001a, 2001b; Schneider et al., 2001; Butler et al., 2002; Seward and Burg, 2008; Koons et al., 2013], and the development of a positive feedback mechanism known as the tectonic aneurysm model (Figure 7). While the Himalayan syntaxes are hypothesized to represent mature and fully developed stages of a tectonic aneurysm, it is suggested that an incipient stage of aneurysm development may be observed at the St. Elias syntaxis [Enkelmann et al., 2009, 2010; Koons et al., 2013].

[37] In the Himalaya, the Indus and the Tsangpo-Brahmaputra Rivers flow from the high Tibetan Plateau toward the west and east, respectively, where they incise the high mountain massifs of Nanga Parbat (8125 m) and Namche Barwa (7782 m), creating extremely high local relief and high stream power, and transport material efficiently to the plains. The concept of the tectonic aneurysm model is that local rheological variations will develop from deep and rapid incision [Zeitler et al., 2001a]. As the strong upper crust is removed from above by efficient erosion, hot and weak material is uplifted from below resulting in compressed isotherms under the topography [Zeitler et al., 2001a]. Provided that this weakening occurs in a region where the crust is already close to failure and that efficient erosion continues, the uplift and removal of material will be concentrated locally and develop a positive feedback mechanism [Zeitler et al., 2001a]. This model has been developed based on various geological and geophysical observations from the Himalayan syntaxes and a comprehensive review can be found in Koons et al. [2013], although additional studies exploring alternative geodynamic mechanisms for aneurysm development are needed. One of the characteristic features at the Himalayan syntaxes is the occurrence of high-grade metamorphic rocks and magmatic intrusions that were exhumed from the lower crust (~40 km) within 10–3 Ma [e.g., Burg et al., 1998; Booth et al., 2004; 2009]. Biotite Ar-Ar cooling ages of exposed bedrock are <1 Ma delineating an outcrop area of metamorphic culmination that is small, with ~50 km diameter or less (<2000 km2) [Koons et al., 2013]. An area of 3300 ± 550 km2 of rapid rock denudation was mapped by <2 Ma zircon (U-Th)/He and FT bedrock ages at the Namche Barwa massif [Burg et al., 1998; Seward and Burg, 2008; Stewart et al., 2008], which was later revised to 4800 ± 550 km2 based on detrital zircon FT ages [Enkelmann et al., 2011]. The exceptional efficiency of eroding and evacuating material from this localized area at the Namche Barwa massif is best displayed by the study of the Tsangpo-Brahmaputra River detritus. When the Brahmaputra River exits the Himalayan foothills, 60–70% of the entire load is sourced by the Namche Barwa massif, which makes up 3% of the entire catchment [Enkelmann et al., 2011]. Farther downstream, in the Brahmaputra Basin in Bangladesh, 40% of the detrital zircon (U-Th)/He ages yielded ages between 0.4 and 1 Ma and originate from the Namche Barwa massif [Tibari et al., 2005]. The long-term (>106 year) denudation rates for the outlined rapid denudation area at Namche Barwa massif have been estimated to range between 2 and 9 mm/yr, whereby the mountainous regions surrounding it yield rates between 0.2 and 1 mm/yr [e.g., Burg et al., 1998; Booth et al., 2009; Seward and Burg, 2008; Enkelmann et al., 2011].

[38] Several observations consistent with the tectonic aneurysm model are present in the St. Elias syntaxis: (a) The region is characterized by high peaks and local relief of >5000 m (Figure 2). (b) The Seward-Malaspina Glacier and the other large glacial systems like the Bagley-Bering Glacier or the Hubbard Glacier present an efficient erosion and excavation agent that is a crucial requirement for the development of a positive feedback between tectonic and surface processes. (c) Glacial denudation is long lasting in the St. Elias Range and ice-rafted debris has been deposited in the Pacific starting 5.6 Ma [Plafker, 1987; Lagoe et al., 1993]. (d) Even if the long-term denudation rates determined by thermochronometry in the St. Elias Range are not as high as those reported from the Himalayan syntaxes, the rates for shorter time scales are higher. Denudation rates were estimated to be ~5 mm/yr for the past 104 years based on offshore sedimentation rates across the entire orogen [Sheaf et al., 2003]. Even higher denudation rates of 9–11 mm/yr for the last centuries were determined for the catchment of the Tyndall Glacier in Icy Bay and the Hubbard Glacier (Figure 2) [Koppes and Hallet, 2006; Trusel et al., 2008]. (e) Concentration of crustal strain at the syntaxis region is evident by numerical modeling [Koons et al., 2010] and GPS measurements that show that most of the current convergence between the Yakutat Terrane and North America are accommodated in a narrow region between Seward Ice Field and Yakutat Bay [Elliott et al., 2010; Elliott, 2011]. Taken together, these observations are consistent with the aneurysm model for rapid denudation and led to the suggestion that there may be an incipient stage of that model, or some similar model, in the St. Elias syntaxis [Enkelmann et al., 2009, 2010].

[39] One of the testable differences between the two suggested models for rapid denudation in the St. Elias syntaxis is the different size of the area affected by rapid denudation (Figure 7). If rapid denudation occurs only as a thin transpressional sliver along the reactivated suture zone between the Chugach and Prince William Terranes [Enkelmann et al., 2008; Spotila and Berger, 2010], the young (3–2 Ma) zircon (U-Th)/He cooling signals should only occur in mylonitic rocks and the ultrabasite belt of the Chugach Terrane (Figures 4 and 7). In contrast, model 2 would result in a much larger region undergoing rapid denudation and we expect young cooling ages in most lithologies occurring underneath the Seward Glacier and probably in the region to the north.

[40] Given that 3–2 Ma zircon (U-Th)/He cooling ages were obtained in a large variety of rocks including gneisses, granulite, schist, amphibolite, and undeformed granitoid rocks, but only in one out of three mylonite samples, we argue for a larger region (Figure 7) of rapid rock denudation that is not limited to a narrow band along the Contact Fault. We suggest that the region of rapid rock denudation extends over most of the Seward catchment and probably north of it. Evidence for such a northward extension of the rapidly exhuming rocks is given by the ~2 Ma detrital zircon FT age population peak found in the Chitina River detritus, which drains the Logan Glacier north of Mt. Logan massif (Figure 7) [Enkelmann et al., 2010]. Such a northward extent of rapid denudation would also result in young cooling ages in the Hubbard Glacier sediments transported to the east into Yakutat Bay (Figures 2 and 7), which was recently confirmed by a dominant ~3 Ma detrital zircon FT age population peak found in Hubbard Glacier outwash material [Enkelmann et al., 2012]. Together, these observations lead us to suggest that the area of rapid denudation is much larger than the Seward Ice Field, probably >2000 km2. We thus exclude the possibility of model 1 and favor model 2 or a third alternative model, explained in the next paragraph that produces a similar set of observations as the aneurysm model.

[41] This alternative model explains rapid denudation that occurs in a larger area along and between two transpressive fault zones that result from the north and northwestward translation of stress that is brought into the syntaxial region by the dextral Fairweather Fault. Rapid denudation would occur along the Contact Fault whereby the region has to be much wider as suggested in model 1 to justify the variety of observed lithologies with 3–2 Ma zircon (U-Th)/He ages. The second region of rapid denudation would stretch along the Connector Fault, which was suggested as the northward continuation of the Fairweather Fault, connecting with the active Totschunda Fault located north of the Wrangell Mountains (Figures 1 and 7). Rapid denudation on this latter fault zone would be necessary to explain the observed young zircon FT cooling ages in the Chitina and Hubbard Glacier detritus, as this fault stretches through the high ice field region.

[42] Several questions need to be addressed to support the idea of the tectonic aneurysm or the development of an alternative model for the St. Elias syntaxis. That is, (a) what is the structural relationship between the rapidly exhumed rocks and the surrounding, less rapidly cooled rocks. What is the role of the Border Ranges Fault and Connector Fault? (b) The seismic structure underneath the Seward Ice Field is also unclear. How thick is the crust? Is the brittle ductile transition zone bent upward as a result of localized denudation? (c) How much is the thermal structure disturbed by localized rapid denudation? Is rapid denudation also reflected by the cooling ages of higher closure temperature systems like Ar/Ar dating on biotite, muscovite, and hornblende? (d) What is the exact spatial extent and patterns of rapid denudation, and how far does it reach to the north and east? (e) When did this process of rapid denudation start? And finally, (f) what is the role of the subducting plate geometry on the pattern of surface deformation and denudation? Answering these questions is beyond the scope of observations presented in this study and is the target of future investigations.

5.5 Surface Processes Implication

[43] Another interesting aspect of our results is the large range of cooling ages including very rapidly exhumed rocks (3–2 Ma zircon (U-Th)/He age) in the cobble size detritus of the Malaspina Glacier, which were collected at the top of the 5–6 km wide rim of debris-covered ice. There are no comparable young zircon (U-Th)/He or FT ages reported from any exposed bedrock in the neighborhood and thus the material must originate from the glacier bed at low elevations where younger ages are expected. That observation gives evidence that the debris on top of the ice at the glacier terminus represents material that originates from rockfall but also from underneath the glacier, and thus that vertical sediment mixing must occur.

[44] The ~2 mm/yr denudation rates indicated by the youngest zircon (U-Th)/He age population are much higher than the rates estimated based on age-elevation relationship of bedrock cooling ages from the surrounding ice-free ridges (~ 0.3 mm/yr) [O'Sullivan and Currie, 1996; Spotila and Berger, 2010]. This difference in denudation rates suggests more efficient erosional processes acting on the ice-covered valley bottom than on the mountain ridges. This observation has been made in other glacial catchments in the St. Elias Mountains where a large discrepancy between cooling ages derived from bedrock and glacial detritus has been observed [Enkelmann et al., 2008, 2010].

[45] Glacial erosion is generally assumed to scale with the ice velocity, suggesting that erosion is more efficient when the glacier ice velocity is high. Headley et al. [2013] showed that other factors related to the fracturing of the rocks at the glacier bed are very important as well and may be the main reason explaining the rapid denudation processes underneath the Seward Ice Field. Measurements of surface ice velocities at the Seward Ice Field reveal ice flux is slow (<100 m/yr) and erosion is expected to be less efficient in comparison to the Seward Throat where ice velocities are much higher (~1800 m/yr) [Headley et al., 2012, 2013]. However, our petrographic classification and cooling age analysis clearly show that erosion is very efficient underneath the Seward Ice Field, probably as the results of the interplay between focusing strains at the syntaxial bend that results in faulting and making bedrock more susceptible for plucking, combined with the efficient removal and transport of the material through the Seward Throat.

6 Conclusion

[46] Based on the cobble lithologies and the measured zircon (U-Th)/He cooling ages of this study and those of Enkelmann et al. [2009, 2010], we suggest that rapid denudation occurs over a larger area of the St. Elias syntaxis region than was previously thought. This interpretation differs from previous interpretations of denudation being focused along a thin transpressional sliver along the Contact Fault. The observations presented here for the St. Elias syntaxis are consistent with observations suggested in the tectonic aneurysm model in the Himalayan syntaxes, based on the coincidence of high local relief, rapid deformation, and efficient removal of the strong upper crust by glacial denudation. An alternative model of localized rock denudation can also explain the existing data set, proposing a wider region of rapid denudation between the west and northwestward structural continuation of the Fairweather Fault, the Contact Fault, and the Connector Fault, respectively. The uniqueness of this interpretation needs to be tested by future geologic and geophysical studies of the dynamics of coupled erosion-tectonic processes in this region.

Acknowledgments

[47] This project was funded by the German Research Foundation (DFG grant EN-941) to EE and TE. We thank Ann-Kathrin Schatz for her help during fieldwork, and Sarah Falkowski and Rachel Headley for fruitful discussions. We are grateful for the constructive reviews of Nicole Gasparini, Emily Finzel, Jeff Benowitz, and one anonymous reviewer.

Ancillary