Earth and Planetary Science Letters 452 (2016) 238–246

Contents lists available at ScienceDirect

Earth and Planetary Science Letters

www.elsevier.com/locate/epsl

Biospheric and petrogenic organic carbon flux along southeast Alaska

Xingqian Cui a,∗, Thomas S. Bianchi a, John M. Jaeger a, Richard W. Smith b

a Department of Geological Sciences, University of Florida, Gainesville, FL, USAb Global Aquatic Research LLC, 6276 Ridge Rd., Sodus, NY 14551, USA

a r t i c l e i n f o a b s t r a c t

Article history:Received 4 March 2016Received in revised form 27 July 2016Accepted 1 August 2016Available online xxxxEditor: D. Vance

Keywords:carbon cyclefjordssoutheast Alaskapetrogenic OCglaciers

Holocene fjords store ca. 11–12% of the total organic carbon (OC) buried in marine sediments with fjords along southeast (SE) Alaska possibly storing half of this OC (Smith et al., 2015). However, the respective burial of biospheric (OCbio) and petrogenic OC (OCpetro) remains poorly constrained, particularly across glaciated versus non-glaciated systems. Here, we use surface sediment samples to quantify the sources and burial of sedimentary OC along SE Alaska fjord-coastal systems, and conduct a latitudinal comparison across a suite of fjords and river-coastal systems with distinctive OC sources. Our results for SE Alaska show that surface sediments in northern fjords (north of Icy Strait) with headwater glaciers are dominated by OCpetro, in contrast to marine and terrestrially-derived fresh OC in non-glaciated southern fjords. Along the continental shelf of the Gulf of Alaska, terrestrial OC is exported from rivers. Using end-member mixing models, we determine that glaciated fjords have significantly higher burial rates of OCpetro (∼1.1 ×103 g OC m−2 yr−1) than non-glaciated fjords and other coastal systems, making SE Alaska potentially the largest sink of OCpetro in North America. In contrast, non-glaciated fjords in SE Alaska are effective in burying marine OC (OCbio-mari) (13–82 g OC m−2 yr−1). Globally, OC in fjord sediments are comprised of a mixture of OCpetro and fresh OCbio, in contrast to the pre-aged OC from floodplain river-coastal systems. We find that there may be a general latitudinal trend in the role of fjords in processing OC, where high-latitude temperate glacial fjords (e.g., Yakutat Bay, SE Alaska) rebury OCpetro and non-glacial mid-latitude fjords (e.g., Doubtful Sound, Fiordland) sequester CO2 from phytoplankton and/or temperate forests. Overall, we propose that fjords are effective in sequestering OCbio and re-burying OCpetro. Based on our study, we hypothesize that climate change will have a semi-predictable impact on fjords’ OC cycling in the near future.

© 2016 Elsevier B.V. All rights reserved.

1. Introduction

Over geological timescales, atmospheric CO2 drawdown can be achieved in part through the net burial of biospheric OC (OCbio) in sedimentary basins (e.g., Berner, 2004). OCbio consists of car-bon that is fixed via photosynthesis in terrestrial (OCbio-terr) and marine environments (OCbio-mari). Petrogenic OC, or OC stored in rocks (OCpetro), represents the largest global pool of OC and in-teracts with the modern active carbon cycle through rock uplift and weathering (Hedges, 1992). While some fraction of OCpetro can be degraded by microbes during transport and re-burial, allowing components of OCpetro to contribute to atmospheric CO2 produc-tion (Berner, 2004; Galy et al., 2015; Petsch et al., 2001), much of it escapes oxidation and is re-buried in marine sediments, having relatively minimal effects on the modern C cycle (Berner, 2004;Galy et al., 2015). Overall, the burial of OCbio and oxidation of

* Corresponding author. Fax: +1 352 392 9294.E-mail address: cuixingqian@ufl.edu (X. Cui).

http://dx.doi.org/10.1016/j.epsl.2016.08.0020012-821X/© 2016 Elsevier B.V. All rights reserved.

OCpetro are important processes influencing atmospheric CO2 over geological timescales (Mackenzie and Lerman, 2006).

Coastal sediments are major carbon depo-centers that account for as much as 90% of Holocene OC burial in the global ocean (Berner, 1982; Burdige, 2005; Hedges, 1992). In particular, fjords have been recognized as “hotspots” of carbon burial (Nuwer and Keil, 2005), with more recent estimates suggesting fjord surface area-normalized OC burial rates are at least five times greater than other marine systems and one hundred times greater than the en-tire ocean average (Smith et al., 2015). More specifically, fjords accumulate only ∼5% of total sediment input from the continents and yet account for 11–12% (∼17–20 × 1012 g OC yr−1) of the total OC buried in marine sediments. This implies that they represent a potentially important CO2 sink over glacial/interglacial timescales.

Fjords occur over a wide range of latitudes (40◦–80◦) and are consequently subjected to a broad spectrum of climatic and land-ocean processes. For example, in high latitudes, glacial erosion contributes to significant input of OCpetro to fjords (Syvitski et al., 1987; Walinsky et al., 2009). Moreover, glacial retreat and ex-

X. Cui et al. / Earth and Planetary Science Letters 452 (2016) 238–246 239

pansion of vegetation at lower latitudes can result in a change from OCpetro-dominated export to one more comprised of OCbio. Fjords from southeast (SE) Alaska, British Columbia and Fiordland, New Zealand reveal latitudinal gradients in concentrations and ac-cumulation rates of sedimentary OC from glacier- to vegetation-dominated watersheds (e.g., Cui et al., 2016; Hinojosa et al., 2014;Hood et al., 2009; Smith et al., 2015; Smittenberg et al., 2004;Walinsky et al., 2009). We believe that Holocene fjords provide natural laboratories for the examination of global gradients in the fate of OCpetro vs. OCbio.

While past work has explored the sedimentary dynamics be-tween glaciers and fjords (e.g., Jaeger et al., 1998; Syvitski et al., 1987; Ullrich et al., 2009), linkages between biochemical processes and carbon storage in fjord sediments have received less attention (Hood and Scott, 2008; Hood et al., 2009; Walinsky et al., 2009). Specifically, SE Alaskan fjords are believed to have the highest OC accumulation rates (Hallet et al., 1996; Walinsky et al., 2009)among global fjords (12 to 265 g OC m−2 yr−1), with this region possibly accounting for as much half of the total global OC burial in fjords (Smith et al., 2015). Yet, less is known about the poten-tial burial and export of OCpetro and OCbio from within SE Alaskan fjords to the open marine environment. While the relationship be-tween glacial versus non-glacial watersheds and sedimentary OC properties in SE Alaskan fjords has been examined (Walinsky et al., 2009), details on the fate of changing OCpetro inputs are still lacking.

In this study, we posit that SE Alaska re-buries a significant amount of OCpetro through bedrock erosion but also represents a significant modern-day CO2 sink by burying OCbio. In addition, SE Alaska shows distinct sedimentary OC differences between vege-tated and glaciated watersheds. Building on the previous work in fjords of SE Alaska (e.g., Hood et al., 2009; Walinsky et al., 2009), British Columbia (Smittenberg et al., 2004) and New Zealand (Cui et al., 2016; Hinojosa et al., 2014; Smith et al., 2015), we conduct a multi-tracer comparison of sedimentary OC composition in surface sediments from these regions (with the addition of radiocarbon measurements in SE Alaska) to obtain a more global perspective on 1) differences in how fjords bury carbon, 2) the potential effects of climate change on carbon cycling in fjords, and 3) the potential role of fjords in the global sequestration of CO2.

2. Site description, sampling, and methods

2.1. Site description

SE Alaska covers the coastal region from Prince William Sound to the SE Panhandle, with latitudinal ranges between 61◦N and 54◦N. SE Alaskan fjords span a range in watershed characteris-tics from tidewater glacial in the northern Gulf of Alaska (GOA) to glacier-free and well-vegetated in the southern Panhandle. Relative to Icy Strait (58◦N), northern areas are largely covered by glaciers, while the southern panhandle has a significant amount of bedrock exposure with variable vegetation cover following early Holocene glacial retreat. The bedrock geology of SE Alaska is dominated by accreted tectonic terranes. Coastal watersheds are composed of the Chugach, Yakutat, and Alexander terranes (Plafker et al., 1989). The watersheds in this study contain mostly marine and marginal marine sedimentary rock of the Chugach/Yakutat terranes, with some contribution of Alexander terrane (schist, marble, greenstone, graywacke, phyllite, and argillite; Winkler et al., 2000) from the Hubbard Glacier in Disenchantment Bay.

2.2. Sampling

Eight sampling locations cover a range from glaciated to vege-tated watersheds (Fig. 1). More specifically, multi-core samples are

from sites EW0408 76 (multi-core MC 2), 73 (MC 2), and 68 (MC 2) in/near Yakutat Bay (north), and sites 32 (MC 8), 36 (MC 3), and 39 (MC 4) in/near Sitka Sound (south). These two transects repre-sent the inner-to-outer reaches of fjords within distinct watersheds that are currently glaciated in the north (Yakutat) or non-glaciated in the south, which is partially-covered by vegetation and exposed bedrock. In addition, southern sites 43 (MC 4) and 19 (MC 2) are within the relatively small Deep and Crawfish Inlets, respectively, which are both highly vegetated. We use previously published sediment accumulation rates at each site (Ullrich et al., 2009;Walinsky et al., 2009) except for station 73, where a revised 210Pb geochronology is established (Table S5; Fig. S3).

2.3. Methods

Bulk sediment total organic carbon (%OC), total nitrogen (N), C/N ratios, stable carbon isotopes (δ13C) are measured after re-moval of inorganic carbon with acid fumigation, according to the method of Harris et al. (2001). Samples for OC radiocarbon anal-ysis are combusted and converted into graphite using the zinc reduction method (Xu et al., 2007). Radiocarbon analyses are made using an Accelerated Mass Spectrometry (AMS) at the Na-tional Ocean Sciences Accelerator Mass Spectrometry (NOSAMS) laboratory, using the general radiocarbon methods described by McNichol et al. (1992). Radiocarbon results are reported as ages, �14C values and fraction modern (Fmod) values. Fmod values are based on �14C values using the equation Fmod = (�14C/1000 +1)/[e(−λ∗(collection year–1950))]) (see supplemental data). Biomarker analyses, which included lignin phenols, fatty acids, and glycerol dialkyl glycerol tetraethers (GDGTs), are measured according to the methods of Feng et al. (2013), Goñi and Hedges (1992), and Hopmans et al. (2004), with details provided in the supplemental data.

Eight lignin phenols, normalized to OC (Λ8 in mg 100 mg OC−1), are used to trace OC contributions from vascular plants (Goñi et al., 1998). The compound 3,5-dihydroxybenzoic acid (3,5-Bd), forms largely in soils by bacteria from the breakdown of plant-derived materials (Houel et al., 2006), is used as a soil tracer. The ratio of 3,5-Bd/vanillyl phenols (3,5-Bd/V) is used as a proxy of relative soil OC versus vascular plant contributions (Houel et al., 2006). The ratio of vanillic acid to vanillin ((Ad/Al)V) is used as an index for lignin oxidative degradation. Branched/isoprenoid tetraether (BIT) is used to indicate soil OC inputs relative to marine primary production (Hopmans et al., 2004). Fatty acids are another group of biomarkers for tracing a diversity of OC sources (Bianchi and Canuel, 2011; Waterson and Canuel, 2008). For example, short-chain fatty acids (SCFAs; C12–18) are considered to have a diver-sity of OC sources, including terrestrial plants, marine algae, and microbes. In contrast, the even-numbered long-chain fatty acids (LCFAs; C24–30), a major component of leaf waxes, are produced exclusively by terrestrial sources (Feng et al., 2013). The ratios of terrestrial to aquatic fatty acids (TARFA; C24+26+28/C12+14+16) have been widely used to indicate relative terrestrial OC inputs (Waterson and Canuel, 2008).

A ternary mixing model is applied to determine the background OCpetro in sediments by assuming a constant OCpetro background level in fjord sediments (see supplemental data) (Galy et al., 2008;Leithold et al., 2006). The OC buried in these fjord sediments is comprised of a mixture of OCpetro and non-rock-derived OCbio, which consists of OCbio-mari and OCbio-terr. Then the radiocarbon composition of the bulk OC can be expressed as follows:

Fmod × %OC = Fmod-bio × %OCbio + Fmod-petro × %OCpetro (1)

where, Fmod, Fmod-bio and Fmod-petro are the radiocarbon composi-tions of the bulk OC, OCbio and OCpetro in sediments. %OC, %OCbio,

240 X. Cui et al. / Earth and Planetary Science Letters 452 (2016) 238–246

Fig. 1. Terra/MODIS imagery with sample locations along SE Alaska. Samples in Yakutat/Disenchantment Bay (76, 73, 68) and southern fjords (Sitka Sound: 32, 36, 39; Deep Inlet: 43; Crawfish Inlet: 19) are shown in A and B, respectively. Water depth is ∼200 m along the continental shelves and Yakutat Bay, and <200 m in the sampled southern fjords. Glacier coverage is extensive and predominant in the northern region north of Cross Sound/Icy Strait. The coastal regions in southeastern Alaska have no glacier coverage (seasonal snow pack is present), but more landslides (pink polygons; Tongass National Forest Landslides data from Southeast Alaska GIS Library). Imagery from NASA Worldview (https :/ /worldview.earthdata .nasa .gov). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

and %OCpetro are the concentrations of bulk OC, OCbio, and OCpetroin sediments. The Fmod of the OCpetro is taken as 0 due to the ancient age of bedrock along SE Alaska. Equation (1) is further modified as (details in supplemental data):

Fmod × %OC = Fmod-bio × (%OC − %OCpetro) (2)

By assuming Y = Fmod × %OC and X = %OC, then

Y = Fmod-bio × (X − %OCpetro) (3)

Thus, by plotting %OC × Fmod versus %OC on a X–Y plot, the slope of the plot represents Fmod-bio and the intercept on X axis represents %OCpetro (Galy et al., 2008; Leithold et al., 2006). In this model, we have two key assumptions. The first is that the background concentration of OCpetro is consistent across all sam-ples, with no assumed dilution from phytodetrital inputs. Phytode-trial inputs would contribute inorganic particles having no OCpetro, thereby diluting the initial concentration of OCpetro in sediments. Station 68-2 is excluded in this model because of its location on the continental shelf where marine phytodetrital material is a likely source, as well as containing bedrock-derived material that may have experienced more significant hydrodynamic sorting dur-ing cross-shelf transport. The second assumption is that Fmod val-ues of OCbio are uniform throughout the SE Alaska system. In SE Alaska, OCbio is mainly derived from marine (OCbio-mari) and ter-restrial (OCbio-terr) primary production. OCbio-mari is newly formed and therefore has an Fmod value of ∼1.0. OCbio-terr is from terres-trial sources, which likely have a short residence time on land (due

to a steep topography), so the Fmod of OCbio-terr might be close to 1.0. Hence, we propose that the Fmod of OCbio, which is a combina-tion of both OCbio-mari and OCbio-terr, is in general consistent across all stations.

Concentrations of OCbio are calculated by subtracting OCpetrofrom total OC at each site, and then further separated into OCbio-terrand OCbio-mari based on terrestrial and marine δ13C binary end-member model (see supplemental data). The relationships can be expressed as follows:

%OC × δ13C = %OCpetro × δ13Cpetro + %OCbio-terr × δ13Cbio-terr

+ %OCbio-mari × δ13Cbio-mari (4)

%OC = %OCpetro + %OCbio-terr + %OCbio-mari (5)

where, δ13C, δ13Cpetro, δ13Cbio-terr, and δ13Cbio-mari are the bulk OC, OCpetro, OCbio-terr and OCbio-mari δ13C values. In this study, we used −20� and −26� as the δ13C end-member values for OCbio-mariand OCbio-terr based on the ranges of sedimentary OC δ13C values and proposed end-member values (Walinsky et al., 2009). A value of −25.4� was used as the OCpetro δ13C end-member value by averaging bedrock samples and river suspended particles from SE Alaska (Walinsky et al., 2009) (see supplemental data).

For statistical analysis, a one-way ANOVA is performed to test for any significant differences between independent groups of vari-ables at the 95% confidence interval. Principle component analysis (PCA) also is performed to discriminate for any other controlling

X. Cui et al. / Earth and Planetary Science Letters 452 (2016) 238–246 241

Fig. 2. The major parameters discussed in this study, including bulk proxies (bulk organic carbon: %OC; OC stable carbon isotopes: δ13C; and OC radiocarbon fraction of mod-ern: Fmod), lignin phenol proxies (sum of eight lignin phenols normalized to the OC: Λ8; and the ratio of 3,5-Bd over vanillyl phenols) and lipid biomarkers (even-numbered long chain fatty acids: LCFAs; terrestrial to aquatic ratios of fatty acids: TARFA). Numbers along x-axis represent coring stations. Samples from Deep Inlet (43), Crawfish Inlet (19), Sitka Sound (32, 36, and 39) and Yakutat Bay (76, 73, and 68) are shown in different colors and patterns. Error bars are presented for some of the parameters. Lignin phenol (Λ8) and fatty acid (LCFAs) proxies are in the units of mg 100 mg OC−1 and mg g OC−1, respectively. The full list of parameters is attached in Fig. S1.

variables linked with bulk and biomarker patterns in sediment samples. All parameters are mean-normalized.

3. Results and discussion

3.1. OC sources to fjord sediments

3.1.1. Autochthonous source inputsBulk sediment OC contents range between 0.3% and 7.2%, with

samples from northern Yakutat Bay having significantly lower %OC than all other samples (p < 0.001). Overall, fjord sediments in both northern and southern fjords are relatively lower in C/N ratios (8.7 to 12.8), and are more enriched in δ13C values (−21.0� to −24.9�; Figs. 2, S1, Table S1) than most other coastal systems (Goñi et al., 1998; Smith et al., 2015). These values further sup-port the importance of marine (algal) inputs to SE Alaskan coastal sediments (Addison et al., 2013). There is evidence of higher in-put of algal OCbio-mari to sediments in the southern fjords com-pared to the northern fjords, as reflected by more enriched δ13C values of −21.7 ± 0.7� compared to −23.7 ± 1.6; p = 0.05, re-spectively (Fig. 2; Table S1). This supports the work by Walinsky et al. (2009), which had even greater sampling spatial resolution of these SE Alaskan fjords. One possible reason for such relatively high OCbio-mari concentration in these panhandle fjords is due to the occurrence of strong nutrient-rich seawater that intrudes into the fjords and to the greater stratification that occurs in the south-ern fjords (Addison et al., 2013). The lower turbidity in the south-ern fjords (Fig. 1), which have glacier-free watersheds, also may increase the depth of euphotic zone. In addition, higher nutri-ent supply, such as dissolved inorganic nitrogen and phosphorus, in glacier-free drainage basins may also stimulate primary pro-duction (Hood and Scott, 2008). As suggested by Walinsky et al.(2009), there can be significant changes in the composition and overall abundance of phytoplankton in these waters due to a com-plex suite of oceanographic controls on the nutrient supply (e.g., Weingartner et al., 2009). The relatively low sedimentary OC con-

centrations in the glaciated fjords likely represent low primary productivity because of high turbidity and exceptionally high in-organic sediment accumulation rates due to glacial runoff (Cowan et al., 1997).

3.1.2. Allocthonous source inputsBulk OC radiocarbon Fmod values in southern fjord sediments

range between 0.93 and 0.97, which are equivalent to radiocarbon ages of 576 and 205 yr, respectively (Fig. 2, Table S1), suggesting minimal inputs of OCpetro relative to OCbio. This is consistent with the observation that the watersheds in these southern fjords are more vegetated than in the northern fjords. In contrast, Fmod val-ues in the northern fjord sediments range between 0.10 and 0.68 and increase seaward along the Yakutat Bay transect, accompanied by lower C/N ratios and more enriched δ13C values. This suggests that the sedimentary OC along Yakutat Bay is a mixture of OCpetroand OCbio-mari (Fig. S2).

OC-normalized lignin contents (Λ8, sum of eight lignin phe-nols) in the southern fjords range between 0.68 and 1.79 mg 100 mg OC−1 and is negatively correlated with δ13C and C/N values (R2 = 0.88, 0.91; p < 0.001) (Fig. 2; Table S2), indica-tive of greater vascular plant input. In addition, (Ad/Al)V ratios (0.27–0.32) in southern fjord sediments are low relative to large river-dominated shelves (Goñi et al., 1998), suggesting that the vascular plant input in these fjords is relatively fresh. Together, this indicates that fresh vascular plants are a major source of sed-imentary OC and its relative abundance in sediments is largely influenced by dilution from marine primary production. Similarly, the dissolved organic matter (DOM) of streams in these water-sheds (with minimal glacier coverage) tends to be dominated by aromatic compounds from vascular plants (Hood et al., 2009). Vas-cular plant material is likely transported into southern fjords from landslides (Fig. 1) and/or fluvial inputs by rivers draining these forested drainage basins (Addison et al., 2013; Hood et al., 2009). In contrast, samples along the Yakutat Bay transect (sites 76, 73, 68) reflect minimal input of vascular plant material, based on the

242 X. Cui et al. / Earth and Planetary Science Letters 452 (2016) 238–246

Fig. 3. Principle component analysis plot of sediment samples from SE Alaska from compiling all components analyzed in this study. The black axes on the bottom and left side are the scales for the components (e.g., SCFA, δ13C), while the red axes represent scales for the samples (e.g., 32). PC1 and PC2 explain >90% of the total variance. Λ8 clusters with all the terrestrial-OC indicators (

∑8, BIT and LCFA) in

the upper right quadrant, opposite (PC1 axis) from the other terrestrial OC degrada-tion indices ((Ad/Al)V, 3,5-Bd/V, S/V). ∑8: sum of eight lignin phenols normalized to sediment weight; TFA: total fatty acids (mg g OC−1); CPI: carbon preference in-dex calculated as C23+25+27+29/C24+26+28+30; ΛP: p-hydroxyl phenols normalized to OC (mg 100 mg OC−1). (For interpretation of the references to color in this fig-ure legend, the reader is referred to the web version of this article.)

low Λ8 values, which ranged between 0.02 and 0.09 mg 100 mg OC−1. The high (Ad/Al)V ratios (0.42–1.00) indicate the lignin ma-terial that is present is more degraded. Interestingly, these results conflict with earlier work that estimated ca. 25–50% of the OC is terrestrially-derived in the northern fjord sediments (Walinsky et al., 2009). These differences are largely due to the end-member model used by Walinsky et al. (2009), which was based on vas-cular plant and marine δ13C values, without considering OCpetroinputs. The δ13C value of OCpetro is between the values of vascular plants and marine OC, which could cause an underestimation of terrestrially-derived OC, once again revealing the need for obtain-ing better constraints on OCpetro inputs in these glaciated fjords.

BIT and 3,5-Bd/V values range between 0.08 and 0.23, and 0.04 and 0.09 in southern fjord sediments, respectively (Figs. 2, S1), indicating significant soil inputs, likely through landslides and/or river transport (Addison et al., 2013). In contrast, samples along the northern Yakutat Bay transect cluster with the soil indices and lignin degradation indices in the PCA plot (Fig. 3) and have rel-atively high 3,5-Bd/V (0.24–0.31) and (Ad/Al)V (0.42–1.00) values and low BIT values (0.05–0.08), indicative of some OC input from soils. This suggests that soils are more poorly developed in the northern glaciated fjords than in the southern non-glaciated ones, which is likely related to the more recent retreat of glaciers in the northern fjords.

LCFAs, derived primarily from terrestrial plant leaf waxes, range between 1.06 and 0.52 mg g OC−1 (Fig. 2; Table S2) in south-ern fjord sediments, indicative of inputs from non-woody vascu-lar plant inputs. Although upper Sitka Sound receives significant amounts of OCbio-terr from rivers, as indicated by the high TARFAratio (0.50), the highest %OC and OC accumulation rates are found in the middle reach of Sitka Sound (station 36). Enriched δ13C and high LCFAs suggest that OC at station 36 is derived from both terrestrial and marine inputs. In contrast to the highly tur-

bid Yakutat Bay to the north, it is likely that reduced terrigenous particle inputs to Sitka Sound result in minimal water-column par-ticle turbidity that limits primary production (Addison et al., 2013;Weingartner et al., 2009). Interestingly, the two southernmost sta-tions (43 and 19) have significantly higher abundances of SCFAs (9.06 and 4.21 mg g OC−1, respectively) than stations along the Sitka Sound transect (Fig. S1; Table S2). When considering that water temperatures in the southern fjords (5–15 ◦C) are favorable for salmon spawning (Hood and Berner, 2009), there may be ad-ditional sources of SCFAs in Deep and Crawfish Inlets from salmon metabolic products and/or microbial biomass decomposition. Yaku-tat Bay head (i.e., Disenchantment Bay; 76) sediments have an SCFA content comparable to southern fjord sediments, which may be related to inputs of microbial detritus from glaciers. Microbes living on glaciers have been shown to be a potential OC reser-voir (Stibal et al., 2012). Moreover, microbial-sourced DOM was found to be important in glacial meltwaters in streams (Hood et al., 2009). While we can only speculate at this time, the SCFAs found at station 76 may in part be derived from microorganisms associated with glaciers. The strong tidal currents that mix water masses from the Yakutat Bay up to Disenchantment Bay could be another source of SCFAs (Ullrich et al., 2009). The unique compo-sition of OC in Yakutat Bay and the southern fjords (Sitka Sound, Crawfish Inlet and Deep Inlet) suggests that relative glacier water-shed coverage and corresponding glacigenic sediment delivery are the dominant factors in controlling composition and accumulation rate of OCbio-terr within SE Alaskan fjords.

3.2. Offshore transport of terrestrial OC

Station 68, located on the continental shelf adjacent to the Yakutat Bay and under the influence of the Alaskan Coastal Cur-rent (ACC), has higher OCbio-terr inputs (i.e., BIT, LCFAs, Λ8) and a fresher OC signal ((Ad/Al)V, 3,5-Bd/V) than stations within Yakutat Bay (76 and 73). This suggests that OCbio-terr at station 68 is not directly exported from Yakutat Bay but may have been from river discharge (e.g., Alsek; Fig. 1), which contributes large amounts of fresh water to the ACC (Weingartner et al., 2009). Furthermore, fjords have been proposed to be effective in trapping sediments (Smith et al., 2015; Syvitski et al., 1987) and are therefore un-likely to be significant sources of OCbio-terr to shelves. The relatively low terrestrial inputs and degradation signal at station 68, relative to southern fjord sediments, are most likely affected by retention of OCbio-terr in proglacial lakes, such as Harlequin Lake and Tanis Lake (Gustavson, 1975), during their fluvial transport to the coast (Fig. 1).

3.3. Pathways for the deposition and accumulation of OC

Based on the binary mixing model, the overall OCpetro back-ground concentration in SE Alaskan fjord sediments (0.34%) (Fig. 4) is comparable to the OC content (0.33%) of Chugach terrane bedrock (Walinsky et al., 2009), indicating minimal degradation of OCpetro during transport and deposition. Stations 76 and 73 have mass accumulation rates of 33 and 0.9 g cm−2 yr−1, which result in OCpetro accumulation rates of 1113 and 30 g OC m−2 yr−1

(Fig. 5; Table S3), and are higher than any organic carbon accumu-lation rates reported previously in fjords and continental shelves (Smith et al., 2015). Extremely high accumulation rates of OCpetroin the northern fjords are believed to be largely driven by tem-perate glacial denudation of sedimentary and metasedimentary bedrock (Hallet et al., 1996; Plafker et al., 1989). For example, glacier erosion rates in SE Alaska are 10 to 100 times higher than in Norway (0.4 mm yr−1) and the Swiss Alps (1.0 mm yr−1) (Hallet et al., 1996). Total sediment accumulation in Yakutat Bay has been estimated to be 190 × 106 t yr−1 (Hallet et al., 1996), which we

X. Cui et al. / Earth and Planetary Science Letters 452 (2016) 238–246 243

Fig. 4. The content of modern OC that is calculated as bulk OC times fraction mod-ern values of bulk OC (%OC × Fmod) in fjord sediments collected from SE Alaska system as a function of bulk OC (%OC). The solid black line represents the best fit, while the black dashed lines represent 95% confidence intervals. The slope repre-sents the fraction of modern value of biospheric OC (Fmod-bio), while the intercept on x-axis (ca. 0.3404/1.0089) represents the concentration of petrogenic OC (OCpetro) in sediments.

used to estimate an OCpetro burial rate of ca. 570–760 × 109 gOC yr−1. This rate of OCpetro burial is very high for a system of this size and to put this into perspective, this is about 2 to 3 times the OCpetro burial rate of Bengal fan system (Galy et al., 2015). To take this a step further, when considering the total sediment accumulation for a number of different regions in SE Alaskan (including the continental shelf, Muir Inlet, Icy, and Yaku-tat Bay), estimates exceed 410 × 106 t yr−1 (Hallet et al., 1996;Jaeger et al., 1998). Once again for perspective, this is 2 and 4 times the sediment loads from the Mississippi River (210 ×106 t yr−1; Jaeger et al., 1998) and all North American west coast rivers (107 × 106 t yr−1; Milliman and Syvitski, 1992). This makes SE Alaska potentially the largest sink of OCpetro in North America, highlighting the importance of glacigenic sediment inputs along this coastline. The glaciers in this region appear to be efficient in eroding OCpetro from bedrock and re-accumulating it in coastal ma-rine sediments for long-term carbon burial, further enhancing the role of exhumation and erosion in the long-term storage of carbon

in the lithosphere. However, due to negligible OCpetro degradation during transport and deposition, this process does not contribute to the atmospheric CO2 rise, and therefore has minimal net effect on the modern C cycle.

Using end-member mixing models, we estimate that the con-centrations of OCbio-mari in the southern vegetated fjords range between 2.2% and 4.8%, which translate to OCbio-mari accumula-tion rates ranging from 13 to 82 (avg. 39 ± 26) g OC m−2 yr−1

(Fig. 4). Interestingly, this range is comparable to the total OC ac-cumulation rates in New Zealand fjords (Hinojosa et al., 2014) and several times higher than global continental shelves (Smith et al., 2015). Once again, Walinsky et al. (2009) report strong evidence for high diatom productivity in these SE Alaskan fjords, particularly in the southern fjords. These high rates of primary production may be linked with intrusions of nutrient-rich waters from the Gulf of Alaska, followed by vertical mixing of nutrients to the surface layer (Addison et al., 2013). They may also be related to local fluvial inputs of silica and iron, as well as efficient recycling of phos-phorus and nitrogen within poorly circulated fjord waters, possibly subsurface open-ocean sources from strong tidal mixing that can entrain nutrients into the euphotic zone (Hood and Scott, 2008;Stabeno et al., 2004). Thus, the higher productivity in the non-glaciated southern fjords, along with their higher accumulation rates of OCbio-mari suggest that these systems may be hotspots for rapid burial of phytodetritus, which may also suggest efficient sequestration of CO2 from the atmosphere. If so, these highly pro-ductive fjords may act as coastal net sinks for CO2, further adding to other coastal sinks, such as blue carbon habitats (e.g., wetlands and sea grasses) (e.g., Chmura et al., 2003).

3.4. Latitudinal comparisons of OC accumulation in fjords in the northern and southern hemispheres

Fjords from high latitude SE Alaska (62◦–55◦N), mid-latitude British Columbia (55◦–48.5◦N), and mid-latitude New Zealand (46◦–44.5◦S), have distinctly different watershed types, such as glaciated, non-glaciated/bare bedrock, and non-glaciated-highly forested, respectively. Along this high-to-mid latitudinal gradient, sedimentation rates decrease dramatically (Smith et al., 2015). In contrast, mass accumulation rate-normalized OC content in fjord sediments increases from SE Alaska (0.4 ± 2.4%; calculated based on measurements in this study and the data reported in Walinsky et al., 2009) south to British Columbia (1.0 ± 0.2%; Smittenberg et al., 2004), and then to New Zealand (2.7 ± 1.8%, Smith et al., 2015; Fig. 6, Table S4) – largely due to bedrock inorganic dilu-

Fig. 5. The accumulation rates (AR) of three major types of OC at each station in SE Alaska and the averages in three fjord systems: SE Alaska (SEA); British Columbia (BC); and Fiordland, New Zealand (NZ). Three major types of OC include: terrestrial biospheric OC (OCbio-terr); marine biospheric OC (OCbio-mari); and petrogenic OC (OCpetro). Numbers along X-axis represent station names. Values at stations within fjords are calculated by multiplying OC AR at each site with fractions of each type of OC using mixing models. Station 68 on the shelf is calculated using a revised three end-members model based on δ13C and Fmod values due to phytodetrital dilution.

244 X. Cui et al. / Earth and Planetary Science Letters 452 (2016) 238–246

Fig. 6. The contrasting OC concentrations and compositions in southeast Alaskan fjords (SEA), British Columbia fjords (BC), Fiordland, New Zealand fjords (NZ), US West Coastal (USWC) margins dominated by small mountainous rivers (SMRs) and passive margins in North and South America dominated by extensive lowland floodplain rivers, including Mackenzie River shelf (MRS), North Gulf of Mexico (NGoM), and Amazon River Coast (AMC) shelf. Fractions of petrogenic OC (OCpetro) and biospheric OC (OCbio) including marine (OCbio-mari) and terrestrial (OCbio-terr) are calculated using the mixing models (see supplemental data). The fraction of modern (Fmod) values on top of each bar represents the Fmod values of OCbio estimated using the slope of the binary model.

tion (Syvitski et al., 1985). OC in the sediments of these three regions is comprised of a mixture of OCpetro and OCbio vary-ing in 14C ages that range from about 18 000 yr (SE Alaska) to less than 200 yr (Fiordland, New Zealand) (Smith et al., 2015;Smittenberg et al., 2004).

Based on our end-member mixing models (supplemental data), we estimate OC concentrations in different latitudinal regions and correct for the age of OCbio (Fig. 6). There is minimal variation in the background content of OCpetro, which ranges between 0.26% and 0.34% by weight for the three fjord regions (Fig. 6, Table S4). It is also important to note that the general OCpetro background should be affected largely by bedrock type. For example, fjord drainage basins dominated by volcanic rocks (e.g., Patagonia fjords, Chile, not included in this study) may have negligible OCpetro back-ground levels, whereas fjords within uplifted marine sedimentary rock (e.g., SE Alaska) will have higher values. The average 14C age of OCbio in these three fjord systems varies from modern to about 320 yr old (Table S4), suggesting that the majority of the OCbiois modern in age. The fractions of OCbio (defined as OCbio/total OC) are positively correlated with bulk %OC and are higher to-wards the mid-latitude fjords (British Columbia and New Zealand) (Fig. 6). While the dataset is limited, there may be a general trend from high latitude glacial to mid latitude non-glacial fjords in the role of OCpetro reburial and sequestration of CO2 via phytoplank-ton and/or temperate forests (Fig. 6). In the case of the forested mid-latitude fjords of New Zealand, previous work has shown rel-atively consistent inputs of freshly-produced OCbio from the water-shed to fjord sediments; this is further supplemented by periods of rapidly-injected inputs of fresh vascular plant detritus due to seismic and mass-wasting events (Cui et al., 2016). These inputs of fresh terrestrially-derived plant detritus to fjord sediments re-sult in high burial efficiency that we believe is analogous to rapid burial of phytoplankton described earlier in the southern fjords of SE Alaska and British Columbia (Fig. 5). Unfortunately, significantly more radiocarbon and biomarker data are needed to further val-idate this potential latitudinal pattern on a global scale. Previous studies proposed that fjords were effective in regulating CO2 levels over glacial–interglacial timescales, since fjords bury OC during in-terglacial periods but mobilize and remineralize pre-deposited OC during glacial advances (Smith et al., 2015).

Large-river and small mountainous river-dominated continen-tal shelves and fjords appear to be the primary centers for the burial of OC in the coastal margins (Blair and Aller, 2012; Smith et al., 2015). To compare the importance of fjord systems in bury-

ing OC relative to large river-dominated open shelf systems and small mountainous river (SMRs)-dominated active margin systems, we summarize surface sediment data from selected regions where such comparison can be made such as US Western Coastal (USWC) margins, Mackenzie River shelf (MRS), Northern Gulf of Mexico (NGoM), and Amazon River Coastal (AMC) shelf (Fig. 6, Table S4). We estimate the Fmod values of OCbio and calculate the concen-trations of the three types of OC in each system based on the end-member mixing models described earlier (see supplemental data). Based on the aforementioned datasets, fjords appear to be more comparable to USWC in OCbio age (modern – 322 yr and 558 yr, respectively), but their OCbio is significantly younger (p =0.01) than that found along large-river-dominated coastal systems (1430–7760 yr). Although river suspended particles in SMRs have been shown to have a high OCpetro background (ca. 0.41%; Leithold et al., 2006), OCpetro content in the adjacent shelf sediments is only ca. 0.19% (Fig. 6; Table S4), indicative of possible OCpetro degra-dation during transport and deposition in such systems, such as the Ganges–Bramaptura (Galy et al., 2015). In contrast, the high OCpetro background found in fjords fed by SMRs are likely related to intense deep physical weathering of petrogenic OC that occurs in these systems, usually during short-term erosive events (Blair and Aller, 2012). Due to the young age of OCbio and high OCpetroconcentrations found in fjords, these systems may be as or even more efficient than SMR-dominated shelves in storing biologically-fixed CO2 and re-burying OCpetro. In contrast, large river-dominated open shelf systems have much lower concentrations of OCpetroin sediments (0.03%–0.17%) than active margins (Fig. 6, Table S4) – likely due to extensive oxidation and dilution of OCpetro dur-ing sediment transport and temporary storage (Galy et al., 2015). Together, these findings further support the notion that fjords and SMRs-dominated shelves are more effective than large-river floodplain systems in sequestering OCbio and re-burying OCpetro– largely due to shorter turnover times (Blair and Aller, 2012;Galy et al., 2015).

3.5. Future implications

Sediment flux into glaciated Alaskan fjords occurs primarily during the summer, when glaciers melt and transport bedrock sediment into fjords (Cowan et al., 1997). Bedrock erosion rates and the subsequent sediment deposition are mainly controlled by glacial ice flux (Jaeger and Koppes, 2015). Glacial melting and re-treat has been found to increase over the past century and is still

X. Cui et al. / Earth and Planetary Science Letters 452 (2016) 238–246 245

currently accelerating (Arendt et al., 2002). If ice flux increases dur-ing periods of glacial retreat (Jaeger and Koppes, 2015), this could likely increase glacier erosion and glacigenic sediment accumula-tion rates, which may also enhance OCpetro re-burial.

In SE Alaska, runoff into fjords reflects the glacial cover-age in the surrounding watersheds. In the northern Gulf region, temperature-driven glacial meltwater dominates input to fjords, whereas the southern fjords are more precipitation-driven rainfall dominated systems (Neal et al., 2010). Precipitation in the south-ern fjords is >5 m yr−1 and has increased by 10% over the past 50 years (Stafford et al., 2000). These systems may behave similar to the high rainfall New Zealand fjords in their ability to rapidly bury modern OC, and possibly even sequester CO2. Similar to the role of mass-wasting events in New Zealand fjords, an increase in extreme precipitation events in southern Alaskan fjords likely has the potential to enhance OCbio transport via surface runoff and mass-wasting events (Hood and Berner, 2009).

Although climate change and the associated receding of glaciers may result in an initial increase in OCpetro and OCbio export in SE Alaska, long-term and extended glacial melting during the An-thropocene may result in an eventual reduction in total sediment loads at high latitude fjords and continental shelves as they shift from glacier covered landscapes, to non-glaciated and then vege-tated landscapes (Cowan and Powell, 1991; Hallet et al., 1996). The properties of OC within other glaciated fjords may differ from SE Alaska, based on several factors such as OCpetro maturity, glacier areal extent, recession rates, ice flux, regional climate, and local vegetation types. Therefore, extending the estimates of OCpetro and OCbio burial in SE Alaska to glaciated fjords requires sampling and characterization of fjord sedimentary OC in other regions of the world.

Acknowledgements

We thank the core repository at Oregon State University, who helped to prepare and ship the samples. We acknowledge Jon L. and Beverly A. Thompson Endowed Chair of Geological Sciences held by T.S. Bianchi in the Dept. of Geological Sciences at Univer-sity of Florida, and China Scholarship Council for supporting part of this research. All original data are attached as supporting infor-mation to this manuscript. We also thank Editor Derek Vance and three anonymous reviewers for valuable comments.

Appendix A. Supplementary material

Supplementary material related to this article can be found on-line at http://dx.doi.org/10.1016/j.epsl.2016.08.002.

References

Addison, J.A., Finney, B.P., Jaeger, J.M., Stoner, J.S., Norris, R.D., Hangsterfer, A., 2013. Integrating satellite observations and modern climate measurements with the recent sedimentary record: an example from Southeast Alaska. J. Geophys. Res., Oceans 118, 3444–3461.

Arendt, A.A., Echelmeyer, K.A., Harrison, W.D., Lingle, C.S., Valentine, V.B., 2002. Rapid wastage of Alaska glaciers and their contribution to rising sea level. Sci-ence 297, 382–386.

Berner, R.A., 1982. Burial of organic carbon and pyrite sulfur in the modern ocean: its geochemical and environmental significance. Am. J. Sci. 282, 451–473.

Berner, R.A., 2004. The Phanerozoic Carbon Cycle: CO2 and O2. Oxford University Press.

Bianchi, T.S., Canuel, E.A., 2011. Chemical Biomarkers in Aquatic Ecosystems. Prince-ton University Press.

Blair, N.E., Aller, R.C., 2012. The fate of terrestrial organic carbon in the marine en-vironment. Annu. Rev. Mar. Sci. 4, 401–423.

Burdige, D.J., 2005. Burial of terrestrial organic matter in marine sediments: a re-assessment. Glob. Biogeochem. Cycles 19.

Chmura, G.L., Anisfeld, S.C., Cahoon, D.R., Lynch, J.C., 2003. Global carbon sequestra-tion in tidal, saline wetland soils. Glob. Biogeochem. Cycles 17.

Cowan, E.A., Cai, J., Powell, R.D., Clark, J.D., Pitcher, J.N., 1997. Temperate glacima-rine varves: an example from Disenchantment Bay, southern Alaska. J. Sediment. Res. 67.

Cowan, E.A., Powell, R.D., 1991. Ice-proximal sediment accumulation rates in a tem-perate glacial fjord, southeastern Alaska. Spec. Pap., Geol. Soc. Am. 261, 61–74.

Cui, X., Bianchi, T.S., Hutchings, J.A., Savage, C., Curtis, J.H., 2016. Partitioning of or-ganic carbon among density fractions in surface sediments of Fiordland, New Zealand. J. Geophys. Res., Biogeosci. 121, 1016–1031.

Feng, X., Benitez-Nelson, B.C., Montluçon, D.B., Prahl, F.G., McNichol, A.P., Xu, L., Repeta, D.J., Eglinton, T.I., 2013. 14C and 13C characteristics of higher plant biomarkers in Washington margin surface sediments. Geochim. Cosmochim. Acta 105, 14–30.

Galy, V., Beyssac, O., France-Lanord, C., Eglinton, T., 2008. Recycling of graphite during Himalayan erosion: a geological stabilization of carbon in the crust. Sci-ence 322, 943–945.

Galy, V., Peucker-Ehrenbrink, B., Eglinton, T., 2015. Global carbon export from the terrestrial biosphere controlled by erosion. Nature 521, 204–207.

Goñi, M.A., Hedges, J.I., 1992. Lignin dimers: structures, distribution, and potential geochemical applications. Geochim. Cosmochim. Acta 56, 4025–4043.

Goñi, M.A., Ruttenberg, K.C., Eglinton, T.I., 1998. A reassessment of the sources and importance of land-derived organic matter in surface sediments from the Gulf of Mexico. Geochim. Cosmochim. Acta 62, 3055–3075.

Gustavson, T.C., 1975. Sedimentation and Physical Limnology in Proglacial Malaspina Lake, Southeastern Alaska. Spec. Publ. SEPM.

Hallet, B., Hunter, L., Bogen, J., 1996. Rates of erosion and sediment evacuation by glaciers: a review of field data and their implications. Glob. Planet. Change 12, 213–235.

Harris, D., Horwáth, W.R., van Kessel, C., 2001. Acid fumigation of soils to remove carbonates prior to total organic carbon or carbon-13 isotopic analysis. Soil Sci. Soc. Am. J. 65, 1853–1856.

Hedges, J.I., 1992. Global biogeochemical cycles: progress and problems. Mar. Chem. 39, 67–93.

Hinojosa, J.L., Moy, C.M., Stirling, C.H., Wilson, G.S., Eglinton, T.I., 2014. Carbon cycling and burial in New Zealand’s fjords. Geochem. Geophys. Geosyst. 15, 4047–4063.

Hood, E., Berner, L., 2009. Effects of changing glacial coverage on the physical and biogeochemical properties of coastal streams in southeastern Alaska. J. Geophys. Res., Biogeosci. 114.

Hood, E., Fellman, J., Spencer, R.G., Hernes, P.J., Edwards, R., D’Amore, D., Scott, D., 2009. Glaciers as a source of ancient and labile organic matter to the marine environment. Nature 462, 1044–1047.

Hood, E., Scott, D., 2008. Riverine organic matter and nutrients in southeast Alaska affected by glacial coverage. Nat. Geosci. 1, 583–587.

Hopmans, E.C., Weijers, J.W., Schefuß, E., Herfort, L., Sinninghe Damsté, J.S., Schouten, S., 2004. A novel proxy for terrestrial organic matter in sediments based on branched and isoprenoid tetraether lipids. Earth Planet. Sci. Lett. 224, 107–116.

Houel, S., Louchouarn, P., Lucotte, M., Canuel, R., Ghaleb, B., 2006. Translocation of soil organic matter following reservoir impoundment in boreal systems: impli-cations for in situ productivity. Limnol. Oceanogr. 51, 1497–1513.

Jaeger, J.M., Koppes, M.N., 2015. The role of the cryosphere in source-to-sink sys-tems. Earth-Sci. Rev. 153, 43–76.

Jaeger, J.M., Nittrouer, C.A., Scott, N.D., Milliman, J.D., 1998. Sediment accumulation along a glacially impacted mountainous coastline: north-east Gulf of Alaska. Basin Res. 10, 155–173.

Leithold, E.L., Blair, N.E., Perkey, D.W., 2006. Geomorphologic controls on the age of particulate organic carbon from small mountainous and upland rivers. Glob. Biogeochem. Cycles 20.

Mackenzie, F.T., Lerman, A., 2006. Carbon in the Geobiosphere: Earth’s Outer Shell. Springer Science & Business Media.

McNichol, A., Gagnon, A., Jones, G., Osborne, E., 1992. Illumination of a black box: analysis of gas composition during graphite target preparation. Radiocarbon 34, 321–329.

Milliman, J.D., Syvitski, J.P., 1992. Geomorphic/tectonic control of sediment discharge to the ocean: the importance of small mountainous rivers. J. Geol., 525–544.

Neal, E.G., Hood, E., Smikrud, K., 2010. Contribution of glacier runoff to freshwater discharge into the Gulf of Alaska. Geophys. Res. Lett. 37, L06404. http://dx.doi.org/10.1029/2010GL042385.

Nuwer, J.M., Keil, R.G., 2005. Sedimentary organic matter geochemistry of Clayoquot Sound, Vancouver Island, British Columbia. Limnol. Oceanogr. 50, 1119–1128.

Petsch, S., Eglinton, T., Edwards, K., 2001. 14C-dead living biomass: evidence for microbial assimilation of ancient organic carbon during shale weathering. Sci-ence 292, 1127–1131.

Plafker, G., Nokleberg, W., Lull, J., 1989. Bedrock geology and tectonic evolution of the Wrangellia, Peninsular, and Chugach Terranes along the Trans-Alaska Crustal Transect in the Chugach Mountains and Southern Copper River Basin, Alaska. J. Geophys. Res., Solid Earth 94, 4255–4295.

Smith, R.W., Bianchi, T.S., Allison, M., Savage, C., Galy, V., 2015. High rates of organic carbon burial in fjord sediments globally. Nat. Geosci., 450–453.

246 X. Cui et al. / Earth and Planetary Science Letters 452 (2016) 238–246

Smittenberg, R.H., Hopmans, E.C., Schouten, S., Hayes, J.M., Eglinton, T.I., Sin-ninghe Damste, J.S., 2004. Compound-specific radiocarbon dating of the varved Holocene sedimentary record of Saanich Inlet, Canada. Paleoceanography 19.

Stabeno, P., Bond, N., Hermann, A., Kachel, N., Mordy, C., Overland, J., 2004. Mete-orology and oceanography of the Northern Gulf of Alaska. Cont. Shelf Res. 24, 859–897.

Stafford, J., Wendler, G., Curtis, J., 2000. Temperature and precipitation of Alaska: 50 year trend analysis. Theor. Appl. Climatol. 67, 33–44.

Stibal, M., Šabacká, M., Žárský, J., 2012. Biological processes on glacier and ice sheet surfaces. Nat. Geosci. 5, 771–774.

Syvitski, J.P., Asprey, K., Clattenburg, D., Hodge, G.D., 1985. The prodelta environment of a fjord: suspended particle dynamics. Sedimentology 32, 83–107.

Syvitski, J., Burrell, D., Skei, J., 1987. Fjords: Processes and Products. Springer-Verlag, New York.

Ullrich, A.D., Cowan, E.A., Zellers, S.D., Jaeger, J.M., Powell, R.D., 2009. Intra-annual variability in benthic foraminiferal abundance in sediments of Disenchantment

Bay, an Alaskan Glacial Fjord. Arct. Antarct. Alp. Res. 41, 257–271.Walinsky, S., Prahl, F., Mix, A., Finney, B., Jaeger, J., Rosen, G., 2009. Distribution and

composition of organic matter in surface sediments of coastal Southeast Alaska. Cont. Shelf Res. 29, 1565–1579.

Waterson, E.J., Canuel, E.A., 2008. Sources of sedimentary organic matter in the Mis-sissippi River and adjacent Gulf of Mexico as revealed by lipid biomarker and δ13CTOC analyses. Org. Geochem. 39, 422–439.

Weingartner, T., Eisner, L., Eckert, G.L., Danielson, S., 2009. Southeast Alaska: oceano-graphic habitats and linkages. J. Biogeogr. 36, 387–400.

Winkler, G.R., MacKevett, E., Plafker, G., Richter, D., Rosenkrans, D., Schmoll, H., 2000. A geologic guide to Wrangell–Saint Elias National Park and Preserve, Alaska. U. S. Geol. Surv. Prof. Pap. 1616, 166.

Xu, X., Trumbore, S.E., Zheng, S., Southon, J.R., McDuffee, K.E., Luttgen, M., Liu, J.C., 2007. Modifying a sealed tube zinc reduction method for preparation of AMS graphite targets: reducing background and attaining high precision. Nucl. In-strum. Methods Phys. Res., Sect. B, Beam Interact. Mater. Atoms 259, 320–329.