Seasonal and interannual changes in particulate organic carbon export and deposition in the Chukchi Sea

Authors


Abstract

[1] Particulate organic carbon (POC) export fluxes were estimated in the shelf-slope region of the Chukchi Sea using measurements of 234Th−238U disequilibria and the POC/234Th ratio in large (>53-μm) particles. These export fluxes were used in conjunction with rates of primary productivity and benthic carbon respiration to construct a POC budget for this shelf-slope region. Samples were collected along a series of shelf-basin transects in the spring (May–June) and summer (July–August) of 2004. These stations were previously occupied during the ice covered (spring) and open water (summer) seasons of 2002, allowing for an interannual comparison of export flux. In contrast to 2002, when open water POC fluxes were significantly higher than in the ice-covered period, POC export fluxes in 2004 were similar during the spring (average = 19.7 ± 24.8 mmol C m−2 d−1) and summer (average = 20.0 ± 14.5 mmol C m−2 d−1). The high POC fluxes measured during the spring are attributed to a plankton bloom, as evidenced by exceptionally high rates of primary productivity (average = 124.4 ± 88.1 mmol C m−2 d−1). The shelf-slope budget of particulate organic carbon indicates that 10–20% of primary productivity was exported below 50 m but was not consumed during benthic carbon respiration or burial and oxidation in underlying sediments. Furthermore, a water column−sediment budget of 234Th indicates that particulate material is retained in shelf sediments on a seasonal basis.

1. Introduction

[2] A fundamental question in studies of the Arctic Ocean carbon cycle is the extent to which organic carbon produced over the continental shelves is exported to the slope and interior basins. The continental shelves constitute nearly 50% of the total area of the Arctic Ocean and contribute up to 85% of primary productivity [Stein and Macdonald, 2004]. Strong gradients exist in both primary productivity and sinking particulate organic carbon (POC) fluxes between the shelves and interior basins. Thus shelf-basin exchange of POC has the potential to significantly increase the seasonal flux of carbon to the slope and interior basin of the Arctic Ocean. The magnitude of POC export from the surface waters, its consumption in underlying sediments, and off-shelf export are not well constrained, despite being critical to constructing an accurate Arctic carbon budget.

[3] A primary goal of this study is to provide a comparison between the export flux of POC and its consumption in underlying sediments in the shelf and slope region of the Chukchi Sea. Shelf POC that is exported from the euphotic zone and not remineralized in the water column or buried in underlying shelf sediments is potentially available for off-shelf export to the slope and interior basin. The fraction of organic carbon available for off-shelf export can be estimated by determining the rate of primary productivity, the timing and duration of spring or marginal ice zone blooms, water column and benthic carbon mineralization, and patterns of particle deposition [Gosselin et al., 1997; Grebmeier, 1993; Grebmeier and McRoy, 1989; Hargrave et al., 1994; Liu et al., 2000; Olli et al., 2002; Pomeroy, 1997; Sherr et al., 2003; Wheeler et al., 1996, 1997]. Recent work in the Chukchi Sea reported a seasonal decoupling between POC export fluxes and benthic carbon respiration [Moran et al., 2005]. Well-preserved marine organic carbon present in shelf and slope sediments provides evidence that this export carbon can become buried in the deeper shelf and slope sediments of the Chukchi Sea [Belicka et al., 2002, 2004].

[4] In this study, we derived a shelf-slope POC budget for the western Arctic Ocean that was constrained using estimated POC export fluxes determined from measurements of 234Th−238U disequilibria and the POC/234Th ratio on sinking particles. Seasonal patterns of particle deposition and export were determined by comparing water column deficits and sediment inventories of 234Th. POC export fluxes measured in 2004 were compared to fluxes measured in the spring and summer of 2002 [Moran et al., 2005] to assess seasonal and interannual changes in the carbon budget of this shelf-slope regime.

2. Methods

2.1. Study Area

[5] A key feature of circulation in this region of the western Arctic Ocean is the ∼0.8 Sv transport of nutrient-rich Pacific source water that flows through the Bering Strait and into the Arctic Ocean [Roach et al., 1995; Weingartner et al., 1998; Woodgate et al., 2005]. The biochemical properties of this nutrient-rich water are modified during transit over the wide Chukchi shelf by photosynthesis, respiration, ice formation and melt, and physical mixing [Melling and Moore, 1995; Weingartner et al., 1998]. This shelf water can be incorporated into the halocline of the Canada Basin and is an important source of nutrients, organic carbon, and suspended particulate matter to the interior basin [Carmack et al., 1997; Ekwurzel et al., 2001; Jones and Anderson, 1986; McLaughlin et al., 1996, 2004; Melling and Moore, 1995; Ostlund et al., 1987; Roach et al., 1995; Schlosser et al., 1995; Swift et al., 1997; Weingartner et al., 1998].

[6] The euphotic zone, defined as the depth at which solar irradiance is 1% of the surface value, averaged 23 ± 11 m over the shelf during the spring and deepened over the slope and basin to 42 ± 25 m. In the summer, the euphotic depth increased to 37 ± 16 m over the shelf and 56 ± 26 m over the basin. During the spring cruise, extensive ice cover existed at all stations, whereas during the summer cruise, shelf and slope stations were largely ice-free. Hydrographic properties for 2002 are described elsewhere [Moran et al., 2005].

2.2. Sample Collection

[7] Samples were collected in the western Arctic Ocean as part of the Shelf-Basin Interactions (SBI) Phase II field program aboard the USCGC Healy. An overview of the goals and approaches used in the SBI program is available elsewhere [Grebmeier and Harvey, 2005]. Samples were collected in the spring (15 May to 23 June) and summer (17 July to 26 August) of 2004 along a series of transects progressing from the shelf to the basin. Transects are labeled as follows: West Hanna Shoal (WHS), East Hanna Shoal (EHS), Barrow Canyon (BC), and East Barrow (EB) (Figure 1).

Figure 1.

Map of the Chukchi Sea, Arctic Ocean, showing sampling stations along two shelf-basin transects occupied during the spring of 2004 (15 May to 23 June 2004) and four transects occupied during the summer of 2004 (17 July to 26 August 2004). Transects are identified as BC (Barrow Canyon), EB (East Barrow), EHS (East Hanna Shoal), and WHS (West Hanna Shoal).

[8] Large-volume samples (200–1000 L) were collected using battery-operated in situ pumps (Challenger Oceanic Systems and Services, Surrey, U.K., and McLane Laboratories, Falmouth, USA) at a flow rate of 2–4 L min−1. Seawater was passed sequentially through one (53-μm) or a series (100-μm, 53-μm, 20-μm) of 142 mm Nitex screens, a 1-μm prefilter cartridge (7.6 × 7.6 cm), and two MnO2 impregnated cartridges (7.6 × 7.6 cm) connected in series to scavenge dissolved 234Th. Prefilter and MnO2 cartridges were dried at 60°C and ashed at 500°C at sea. Because 234Th−238U equilibrium was observed in deep-water samples collected during these and previous cruises, we assumed that this sequential cartridge collection method was accurate.

[9] In addition to large-volume samples, 2–4 L samples were collected using the CTD-rosette for measurement of total 234Th [Buesseler et al., 2001].

[10] Relatively undisturbed sediment core samples were collected with a Haps multicorer and sectioned into 1–2 cm intervals to a depth of 4 cm. During the spring, sediment samples were collected at 8 stations at the EHS and BC transects. During the summer, sediment samples were collected at 23 stations at the WHS, EHS, BC, and EB transects. Samples were stored frozen in plastic vials and returned to the shore-based laboratory for radiochemical and elemental analysis.

2.3. 234Th and 238U Analysis

[11] Dissolved and small-particle 234Th activities were measured by gamma spectrometry. The cartridge ash was packed into appropriate geometries for counting on a Canberra pure Ge well detector (GL20203, 150 cm3) or Canberra pure Ge planar detector (GCW3023, 2000 mm2). 3.5 g of cartridge ash were packed to 4 mL in polystyrene vials or in 4 oz. polypropylene jars to be analyzed with the well and planar detectors, respectively. 234Th activities were determined by gamma emission at 63.3 keV and decay-corrected to the midpoint of sample collection [Buesseler et al., 1992b]. When possible, 234Th on MnO2 cartridges was counted at sea using the planar detector. Owing to the large number of samples collected during the summer cruise, some of the MnO2 and small-particle samples (37 samples; 10% of samples collected during this cruise) were analyzed for 234Th by beta emission after purification by ion exchange chromatography [Buesseler et al., 1992b]. A remaining 41 samples from the upper 50 m were not analyzed owing to limited detector availability. On these samples, the average MnO2 cartridge collection efficiency from both cruises (75 ± 15%) was used, and the small-particle 234Th activity was assumed to be 10% of the dissolved activity.

[12] 234Th activities on large particles (>53-μm, >100-μm, 53–100-μm, and 20–53-μm size fractions) were determined by beta emission of the 234Pa daughter using a RIS∅ National Laboratory low-background beta detector. Particles collected on 100-μm, 53-μm, or 20-μm Nitex screens were resuspended by sonication and collected on a Whatman GF/F filter [Charette and Moran, 1999; Cochran et al., 2000]. The filter was mounted on an acrylic planchet and covered with clear plastic (1.5 mg cm−2) and Al foil (4.5 mg cm−2) to shield alpha particles and low-level beta emitters. Samples were counted three times, with counting times separated by at least seven days. Data was fitted to the 234Th decay curve and decay-corrected to the time of sample collection.

[13] Sediment 234Th activities were determined at 1–2 cm intervals down to a depth of 4 cm by gamma or beta spectrometry. 234Th activities were measured on all 8 sediment cores collected at the EHS and BC transects during the spring. Of the 23 sediment cores collected during the summer, 234Th was measured at 3 stations along the BC transect. Samples from the BC transect were counted using a pure Ge planar detector according to the same procedure as for dissolved and small-particle 234Th. Samples collected at EHS during the spring were determined by beta emission after purification by ion exchange chromatography [Buesseler et al., 1992b]. Uncertainties for all 234Th measurements were calculated as the 1-σ counting error.

[14] 238U activities were calculated from salinity according to 238U (dpm L−1) = 0.0708 × S (‰), which has been confirmed for the Arctic Ocean [Chen et al., 1986; Moran et al., 2005]. Salinity values were obtained from the calibrated CTD.

2.4. Analysis of Particulate Organic Carbon and Nitrogen

[15] Particulate organic carbon and nitrogen were analyzed on GF/F filter subsamples containing >53-μm, >100-μm, 53–100-μm, and 20–53-μm particles. Subsamples of known weight were placed in a dessicator with concentrated fuming hydrochloric acid for 24 hours to remove inorganic carbon and then dried for 24 hours at 60°C. Particulate organic carbon and nitrogen were measured using a CE-440 Elemental Analyzer (Exeter Analytical, Inc.) [Pike and Moran, 1997]. Blanks were determined for shipboard procedures by passing ∼100–200 mL GF/F-filtered seawater through a precombusted GF/F filter. Blanks determined for the summer cruise averaged 9.3 ± 3.7 μmol C and 1.1 ± 0.5 μmol N. Organic carbon was measured in surface sediments (0–1 cm) at the EHS and BC transects during the spring, and at the WHS, EHS, BC, and EB transects during the summer.

3. Results

3.1. Water Column 234Th and 238U Activities

[16] 234Th and 238U activities measured on samples collected using in situ pumps during two 2004 cruises are listed in auxiliary material Table S1. 234Th−238U disequilibria were evident at all stations, indicating particle export occurring on a timescale of days to weeks. During the spring, the lowest 234Th/238U activity ratios (∼0.6–0.5) were measured in shelf bottom waters and at ∼200 m over the slope and basin (Figure 2). Owing to the retreat of sea ice into the basin, the transects occupied during the summer cruise extended farther into the basin than during the spring, with maximum water depths of 3825 m, 3850 m, 3799 m, and 3250 m at the WHS, EHS, BC, and EB transects, respectively. During the summer, the lowest activity ratios (∼0.2–0.5) were measured over the shelf and in the surface waters of the slope and interior basin (Figure 3). Activity ratios of ∼0.8–1.0, indicative of minimal particle export, were measured below ∼200 m in the slope region and below ∼50 m in the interior basin. The presence of 234Th−238U disequilibria in deep waters (>300 m) is attributed to 234Th scavenging in shelf and slope bottom waters and subsequent transport off-shelf.

Figure 2.

Total 234Th/238U activity ratios (AR) measured during the spring 2004 SBI process cruise in the Chukchi Sea. Plots were constructed with a variable resolution rectangular grid (VG Gridding) using Ocean Data View Version 3.0-2005 (R. Schlitzer, Ocean Data View, 2004, http://www.awi-bremerhaven.de/GEO/ODV).

Figure 3.

Total 234Th/238U activity ratios (AR) measured during the summer 2004 SBI process cruise in the Chukchi Sea. Plots were constructed with a variable resolution rectangular grid (VG Gridding) using Ocean Data View Version 3.0-2005 (R. Schlitzer, Ocean Data View, 2004, http://www.awi-bremerhaven.de/GEO/ODV).

3.2. POC, PON, and POC/234Th Ratios

[17] POC/234Th ratios on >53-μm particles ranged from <1 μmol dpm−1 to ∼120 μmol dpm−1, and decreased with depth (auxiliary material Table S2). 234Th and POC measurements on >100-μm, 53–100-μm, and 20–53-μm sized particles at 50 and 100 m permitted a comparison of POC/234Th ratios among different size fractions (Figure 4). Size-fractionated POC/234Th ratios ranged from <1 to ∼50 μmol dpm−1. Ratios in the three size fractions at matching depths were similar to within a factor of two.

Figure 4.

The 53–100-μm POC/234Th ratios (x axis) are plotted against 20–53-μm (triangles) and >100-μm (circle) POC/234Th ratios (both on the y axis). Most POC/234Th ratios in the 53–100-μm size fraction are similar to ratios on 20–53-μm and >100-μm particles to within a factor of 2.

3.3. Sediment 234Th Activities and OC Concentrations

[18] Sediment 234Th activities and OC concentrations from the spring and summer of 2004 are listed in auxiliary material Table S3. Sediment inventories of excess 234Th were calculated using 234Th activities measured throughout a 4 cm core and the sediment density. Sediment density was calculated from wet and dry sediment weights [Baskaran and Naidu, 1995]. During the spring, excess 234Th inventories were confined to the upper 2–3 cm and ranged from 1.7 ± 0.3 dpm cm−2 to 10.5 ± 3.5 dpm cm−2 (average = 7.7 ± 4.3 dpm cm−2). During the summer, 234Th activities were higher and inventories ranged from 8.2 ± 2.4 dpm cm−2 to 77.7 ± 21.6 dpm cm−2 (average = 47.3 ± 35.6 dpm cm−2).

[19] Sediment organic carbon concentrations in surface sediments (0–1 cm) ranged from <0.1 mmol g−1 to 1.8 mmol g−1 (average = 1.2 ± 0.4 mmol g−1; 1.5% ± 0.5% by weight). Sediment organic carbon concentrations (mmol cm−3) were multiplied by a 210Pb-derived average sedimentation rate (3.4 × 10−4 ± 1.5 × 10−4 cm d−1) for the shelf-slope region of the Chukchi Sea to calculate the rate of OC loss to the sediments K. Lepore et al., 210Pb as a tracer of shelf-basin transport and sediment focusing in the Chukchi Sea, submitted to Deep Sea Research, Part II, 2006). The product of the organic carbon concentration in surface sediments and the shelf sedimentation rate represents the loss of OC from surface sediments by both burial and remineralization in the sediments below 1 cm. This term is added to the rate of benthic carbon respiration in order to calculate the total loss of organic carbon in shelf and slope sediments. Fluxes measured during the spring and summer of 2004 ranged from 0.1 mmol C m−2 d−1 to 4.5 mmol C m−2 d−1 (average = 1.6 ± 0.9 mmol C m−2 d−1).

4. Discussion

4.1. 234Th Fluxes Estimated Using One- and Three-Dimensional Models

[20] Because of its half-life (24.1 days) and high particle-reactivity relative to its soluble 238U parent, 234Th is a useful tracer of particle scavenging and export on a seasonal timescale. The radiochemical balance of 234Th is given as

equation image

where ATh is the activity of 234Th (dpm m−3), AU is the activity of 238U (dpm m−3), λTh is the decay constant of 234Th (0.0288 d−1), PTh is the sinking flux of 234Th (dpm m−2 d−1), Uh and Vh (m d−1) are the horizontal advective velocities in the x and y directions, Ka (m2 d−1) is the horizontal apparent diffusivity calculated according to Okubo [1971], and z is the depth of the euphotic zone [Bacon et al., 1996; Buesseler et al., 1995; Charette et al., 2001; Dunne and Murray, 1999; Gustafsson et al., 1998; Murray et al., 1996]. Because sufficient time series data were not available, a steady state must be assumed and equation (1) becomes

equation image

The steady state assumption is often used, although the assumption may be violated during periods of rapid particle export such as phytoplankton blooms and sediment resuspension events [Buesseler et al., 1992a; Kaufman et al., 1981; Ku and Musakabe, 1990; Rutgers van der Loeff et al., 1997; Minagawa and Tsunogai, 1980; Tanaka et al., 1983]. As a result, 234Th fluxes may be underestimated or overestimated when steady state is assumed under prebloom or postbloom conditions, respectively.

[21] The horizontal transport term for 234Th may be neglected when the water mass transit time between stations is long relative to the half-life of 234Th. However, it is important to consider horizontal advection and diffusion of 234Th in the overall calculation of 234Th fluxes in this study because high current velocities throughout the Chukchi Sea can lead to significant lateral transport between stations. The relative importance of advection and diffusion of 234Th to the 234Th deficit over the length scales and timescales relevant to this study may be estimated using the apparent diffusivity,

equation image

where Ka is the apparent diffusivity (cm2 s−1) and L is the length scale (cm) [Okubo, 1971]. The average distance between stations is ∼30 km along a transect line and ∼120 km between transects. Using these distances, Ka is 3.12 × 105 cm2 s−1 and 1.38 × 106 cm2 s−1, respectively. The length scale over which diffusive and advective transport of 234Th between stations can affect the measured 234Th deficit is given as

equation image
equation image

where Uh is the horizontal advection (cm s−1), and t is the average life-time of 234Th (35 d). Using the estimated values for Ka cited above, the length scale for diffusion ranges from 10–20 km. Using advection rates of 1 cm s−1 and 15 cm s−1, the length scale for advection ranges from 30–450 km. Although these estimates are simplistic, the implication is that advective transport of 234Th is significant between most stations occupied during the spring and summer and the data must be interpreted carefully within those limitations.

[22] Thus a steady state three-dimensional model that includes lateral advective transport was used to calculate the vertical flux of 234Th,

equation image

where the 234Th and 238U activities are integrated over the upper 50 m, and Uh and Vh are horizontal advective velocities averaged over the upper 50 m. Lateral transport of 234Th was calculated both along a transect (x) and between transects (y). Horizontal advective velocities were estimated using 2002 model output from a coupled ice-ocean model (U.S. Navy Polar Ocean Ice Prediction System; PIPS 3.0) [Clement et al., 2005; Maslowski et al., 2004]. These three-dimensional 234Th fluxes were only calculated at transects where 234Th gradients could be observed; namely BC in the spring, and EHS, BC, and EB in the summer.

[23] The one-dimensional and three-dimensional 234Th fluxes are listed in Table 1. At most stations, the one-dimensional 234Th fluxes were lower than the three-dimensional fluxes. The underestimation of the one-dimensional fluxes is due to west–east current velocities, especially between the BC and EB transects, and to the 234Th gradient throughout the study area, which ranged from higher 234Th activities in the west to lower 234Th activities in the east. The differences between the 1-D and 3-D model results reported here are generally greater than those reported by other studies [Benitez-Nelson et al., 2000; Charette et al., 2001; Gustafsson et al., 1998]. This is attributed to the high advective velocities throughout the Chukchi Sea, as well as large along- and cross-shelf gradients in 234Th activities.

Table 1. One-Dimensional and Three-Dimensional 234Th Fluxes Measured During the Spring and Summer of 2004
Station1-D 234Th Flux at 50 m, dpm m−2 d−1Velocity (x),a cm s−1Velocity (y),a cm s−13-D 234Th Flux at 50 m, dpm m−2 d−11-D Flux/3-D Flux
  • a

    Values were derived from 2002 model output and represent monthly averages throughout the upper 50 m of the water column.

Spring 2004
34-BC21660.82 ± 1181.43-3.001147.85 ± 816.531.45
31-BC31077.81 ± 317.9318.583.92−3875.12 ± −1143.08-
24-SB51903.71 ± 530.0314.554.5414106.92 ± 3927.640.13
28-BC4998.73 ± 212.959.925.86−5126.68 ± −1093.12-
26-BC51019.35 ± 402.1215.992.571107.24 ± 436.800.92
 
Summer 2004
21-BC31762.92 ± 413.6325.205.20287.91 ± 67.556.12
22-BC41749.89 ± 541.3420.968.311930.68 ± 597.270.91
23-BC51895.62 ± 1555.1714.728.766732.28 ± 5523.170.28
35-BC61740.67 ± 477.9312.0811.16−94.36 ± −25.91-
37-BC8195.62 ± 60.625.170.69−767.06 ± −237.71-
25-EB11686.46 ± 452.92-4.512756.68 ± 740.350.61
26-EB22594.23 ± 559.284.9310.916750.70 ± 1455.360.38
29-EB32520.65 ± 714.262.0711.474028.28 ± 1141.460.63
34-EB42290.82 ± 1087.293.5516.165279.22 ± 2505.680.43
33-EB52230.34 ± 1271.374.9317.981426.94 ± 813.401.56
32-EB62203.27 ± 785.144.3610.426627.30 ± 2361.660.33
31-EB71299.96 ± 1092.311.870.831019.26 ± 856.451.28
42-EHS41147.68 ± 803.38-8.01585.38 ± 409.761.96
44-EHS5872.64 ± 610.852.298.59−3952.94 ± −2767.06-
47-EHS61313.28 ± 449.221.9714.935656.26 ± 1934.790.23
48-EHS71230.48 ± 861.341.367.332224.64 ± 1557.250.55
49-EHS91356.48 ± 368.271.075.402947.05 ± 800.090.46
51-EHS12596.88 ± 417.82--596.88 ± 417.821.00

[24] The contemporaneous deployment of in situ pumps and sediment traps at selected stations along the EHS and BC transects allows for a direct comparison of 234Th fluxes determined using methods that differ in sampling technique and timescale of collection. Methods used for sediment trap deployment and subsequent sample processing are reported elsewhere [Lalande et al., 2006]. During the spring and summer of 2004, in situ pump- and sediment trap−derived 234Th fluxes were in agreement to within a factor of two for 70% of the measurements [Lalande et al., 2006]. These results suggest that the one-dimensional in situ pump-derived 234Th fluxes may be considered representative of the flux of 234Th at the time of sample collection. It should be noted that sediment traps were not deployed at the EB transect, where one-dimensional 234Th fluxes were consistently lower than three-dimensional fluxes. The results of the three-dimensional 234Th flux model indicates that one-dimensional fluxes may underestimate the true 234Th flux at most stations, particularly at the EB transect. However, because three-dimensional fluxes can only be calculated at one transect from the spring cruise, and at three transects from the summer cruise, one-dimensional fluxes were used to calculate the POC export fluxes discussed in the following sections.

4.2. POC Export Fluxes

[25] POC export fluxes previously reported for the interior basins of the Arctic Ocean are low [Anderson et al., 2003; Gosselin et al., 1997; Moran et al., 1997; Moran and Smith, 2000; Pomeroy, 1997; Wassmann et al., 2003; Wheeler et al., 1996], but high fluxes over the continental shelves have been observed [Amiel et al., 2002; Coppola et al., 2002; Hargrave et al., 2002; Moran et al., 1997, 2005; Olli et al., 2002]. In addition to spatial heterogeneity, the marginal areas of the Arctic Ocean exhibit intense seasonal variation in primary productivity and particle export. Marginal ice edge and spring blooms occur with increased sunlight in the early spring. These blooms may promote high POC export efficiencies when a time lag exists between the appearance of primary and secondary producers [Pomeroy and Deibel, 1986; Stein and Macdonald, 2004].

[26] POC fluxes determined here are comparable to those previously reported in the shelf-slope region of the western Arctic Ocean [Chen et al., 2003; Moran et al., 1997, 2005; Moran and Smith, 2000; Trimble and Baskaran, 2005]. However, this study provides new information on the seasonality of 234Th fluxes as well as size-fractionated POC/234Th ratios. The flux of POC was calculated by multiplying the flux of 234Th by the POC/234Th ratio on large particles that are assumed to be representative of the sinking flux,

equation image

where PPOC is the flux of POC (mmol C m−2 d−1), PTh is the flux of 234Th (dpm m−2 d−1) at 50 m, and POC/234Th is the ratio (mmol dpm−1) at 50 m [Moran et al., 2005, 2003]. A depth of 50 m was used by Moran et al. [2005], and is used here to allow a direct comparison between fluxes measured during 2002 and 2004. Furthermore, this depth is similar to the average depth of the euphotic zone (1% light level) at most stations. The POC/234Th ratio was similar to within a factor of 2 among the >100-μm, 53–100-μm, and 20–53-μm size fractions. Because POC/234Th ratios were similar among size fractions, and in order to be consistent with previous studies, the POC fluxes were calculated using the POC/234Th ratio on 53–100-μm or >53-μm sized particles [Moran et al., 2005] (Table 2).

Table 2. POC Export Fluxes, Calculated at 50 m in the Spring and Summer of 2004a
StationLatitude, °NLongitude, °WBottom Depth, m1-D 234Th Flux, dpm m−2 d−1POC/234Th>53−μm, 53−100−μm, μmol dpm−1POC Flux, mmol m−2 d−1
  • a

    The 234Th flux was calculated at 30 m.

Spring, 15 May to 23 June 2004
9-EHS0a72.007159.75539871.20 ± 190.675.87 ± 0.375.12 ± 1.17
16-EHS472.633158.6871101149.12 ± 238.173.14 ± 0.283.60 ± 0.81
17-EHS572.729158.4222541575.36 ± 369.880.63 ± 0.060.99 ± 0.25
19-EHS672.885158.26413421618.56 ± 349.05- ± -1.20 ± 0.26
20-EHS773.163157.8202500495.36 ± 395.581.33 ± 0.140.66 ± 0.53
24-SB571.779155.011741903.71 ± 530.03- ± -54.32 ± 19.22
26-BC570.045154.56814261019.35 ± 402.1221.50 ± 1.8121.92 ± 8.84
28-BC471.920154.927574998.73 ± 212.956.03 ± 0.366.03 ± 1.34
31-BC371.646155.8641751077.81 ± 317.9332.20 ± 1.6734.71 ± 10.40
34-BC271.395157.6671161660.82 ± 1181.4341.40 ± 4.9368.75 ± 49.59
Average     19.73 ± 24.83
 
Summer, 17 July to 26 August 2004
15-BC271.409157.4821261959.55 ± 1737.1425.31 ± 2.6049.60 ± 44.27
21-BC371.643156.0131721762.92 ± 413.6320.81 ± 2.2436.69 ± 9.47
22-BC471.902154.8605421749.89 ± 541.3413.13 ± 1.3922.97 ± 7.51
23-BC572.013154.65310281895.62 ± 1555.177.05 ± 0.8913.36 ± 11.09
35-BC672.156153.99018651740.67 ± 477.935.75 ± 0.4010.01 ± 2.84
37-BC872.732153.0643769195.62 ± 60.623.49 ± 0.380.68 ± 0.22
25-EB171.295152.532491686.46 ± 452.9210.79 ± 1.2618.20 ± 5.33
26-EB271.467152.3721092594.23 ± 559.288.89 ± 0.7723.07 ± 5.36
29-EB371.573152.4551902520.65 ± 714.265.68 ± 0.5714.31 ± 4.30
34-EB471.643152.3004222290.82 ± 1087.2918.65 ± 1.6842.73 ± 20.64
33-EB571.660152.20214732230.34 ± 1271.3719.14 ± 2.1442.68 ± 24.80
32-EB671.965152.13921872203.27 ± 785.1420.67 ± 2.4145.55 ± 17.08
31-EB772.412152.20336211299.96 ± 1092.313.12 ± 0.294.06 ± 3.43
42-EHS472.652158.5951672295.36 ± 803.3812.38 ± 1.1314.20 ± 10.03
44-EHS572.689158.467226872.64 ± 610.8516.77 ± 1.8514.64 ± 10.37
47-EHS672.837158.3403631313.28 ± 449.2212.36 ± 1.3616.23 ± 16.23
48-EHS772.909158.39312271230.48 ± 861.3415.77 ± 2.0819.41 ± 13.82
49-EHS973.031158.08919891356.48 ± 368.2716.80 ± 1.4722.79 ± 6.50
51-EHS1273.804156.8353626596.88 ± 417.826.53 ± 0.763.89 ± 2.76
59-WHS273.001160.7231601437.12 ± 474.3011.68 ± 1.3316.79 ± 5.86
58-WHS373.102160.4012163918.24 ± 1944.697.88 ± 0.9230.89 ± 15.74
55-WHS573.309160.1321197722.16 ± 221.654.00 ± 0.442.89 ± 0.94
54-WHS673.484159.6232137760.32 ± 231.9016.32 ± 2.5712.41 ± 4.26
52-WHS1273.925157.8263801336.96 ± 94.675.57 ± 0.641.88 ± 0.57
Average     20.00 ± 14.47

[27] Results from this study differ from Moran et al. [2005], where a threefold increase in POC export was observed between the spring and summer of 2002. In 2004, average POC export fluxes were similar during the spring (19.7 ± 24.8 mmol C m−2 d−1) and summer (20.0 ± 14.5 mmol C m−2 d−1) (p value = 0.97). Stations were occupied during bloom conditions during both sampling periods at BC, and under prebloom or early bloom conditions at EHS in the spring. POC fluxes measured in the spring averaged 37.2 ± 25.0 mmol C m−2 d−1 at BC and 2.3 ± 2.0 mmol C m−2 d−1 at EHS. During the summer, POC fluxes were similar at BC (average = 22.2 ± 18.2 mmol C m−2 d−1), and increased to an average of 15.2 ± 6.4 mmol C m−2d−1 at EHS (Figure 5).

Figure 5.

Spatial distribution of POC export fluxes measured during the 2004 spring (black) and summer (white) SBI process cruises.

[28] POC fluxes calculated from one-dimensional 234Th fluxes at the East Barrow transect were similar to or exceeded the rate of primary productivity, which averaged 28 ± 18 mmol C m−2 d−1 during the summer of 2004. However, previous studies have indicated that 234Th-derived POC export fluxes measured under postbloom conditions can often overestimate the POC flux [Buesseler et al., 1992a; Ku and Musakabe, 1990; Rutgers van der Loeff et al., 1997; Minagawa and Tsunogai, 1980; Tanaka et al., 1983]. Alaskan coastal water flowing from the Bering Sea along the Alaskan coast to the eastern Chukchi Sea becomes nutrient-depleted in early summer owing to blooms in the Bering and Chukchi Seas [Walsh et al., 1989]. At the time of sample collection, the East Barrow transect was characterized by low chlorophyll and nutrient concentrations, suggesting postbloom conditions. Although the one-dimensional 234Th fluxes may be underestimated owing to advective transport of 234Th to East Barrow, the POC flux measured at EB may be regarded as an upper estimate due to postbloom sampling.

4.3. POC Export, Primary Productivity, and Benthic Carbon Respiration

[29] The fraction of primary productivity that leaves the euphotic zone as the 234Th-derived POC export flux is identified as the Th-E ratio [Buesseler, 1998]. This expression of the efficiency of POC export determines the amount of reduced carbon available for respiration in the deep water and in the benthos. Rates of primary productivity were measured using 14C uptake experiments. A detailed description of the methods can be found elsewhere [Hill and Cota, 2005]. Low POC fluxes and Th-E ratios between 2% and 10% were observed during the spring of 2002 [Moran et al., 2005]. During the summer of 2002, POC fluxes increased by a factor of 3 compared to the spring, and export efficiencies ranged from 10% to 60% [Moran et al., 2005]. By comparison, export efficiencies ranged from 1% to 5% at EHS, and from 10% to 50% at BC during the spring of 2004 (Figure 6a). Low export efficiencies at EHS indicate prebloom or early bloom conditions, similar to those observed during the spring of 2002. During the summer, export efficiencies were between 10% and 70% at most stations (average = 40 ± 30%), similar to Th-E ratios reported for the summer of 2002 (Figure 6a). However, rates of primary productivity and POC export fluxes were higher in 2004 than in 2002 by a factor of ∼10 in the spring and ∼2–3 in the summer (Table 3).

Figure 6.

(a) POC export plotted against primary productivity measured during the 2004 spring and summer SBI process cruises. (b) Benthic carbon respiration plotted against POC export during the 2004 spring and summer SBI process cruises.

Table 3. Seasonal and Interannual Changes in Organic C Fluxes in the Chukchi Sea
FluxSpringSummer
mmol C m−2 d−1 % PPmmol C m−2 d−1 % PP
  • a

    Calculated as the difference between the POC export flux and the sum of the benthic C respiration and sediment OC; it represents the imbalance between POC export and loss to the sediments.

  • b

    These fluxes represent the average values of all stations (shelf and slope).

2002
Shelf
   PP15.1 ± 7.9-35.9 ± 23.8-
   POC export3.0 ± 3.120%13.5 ± 13.038%
   Benthic C respiration3.3 ± 1.522%7.2 ± 6.220%
   POC export - benthica−0.3 ± 0.30%6.3 ± 8.118%
   DOC accumulation  2.3 ± 1.66%
 
Slope
   PP8.9 ± 4.1-31.6 ± 21.9-
   POC export1.1 ± 1.612%8.9 ± 6.228%
   Benthic C respiration1.5 ± 1.017%2.7 ± 2.39%
   POC export - benthica−0.4 ± 0.60%6.2 ± 6.820%
   DOC accumulation  1.7 ± 1.15%
 
Totalb
   PP12.0 ± 6.9-32.9 ± 21.9-
   POC export1.7 ± 2.314%10.8 ± 9.533%
   Benthic C respiration2.3 ± 1.519%4.7 ± 5.014%
   POC export - benthica−0.6 ± 0.90%6.1 ± 8.419%
   DOC accumulation  1.9 ± 1.36%
 
2004
Shelf
   PP117.6 ± 35.9-158.1 ± 206.3-
   POC export27.9 ± 29.124%23.5 ± 12.015%
   Benthic C respiration7.1 ± 6.56%9.1 ± 9.16%
   Sediment OC2.3 ± 1.72%1.6 ± 0.41%
   POC export - benthica18.5 ± 25.616%12.8 ± 14.38%
   DOC accumulation  4.4 ± 4.63%
 
Slope
   PP132.9 ± 137.1-50.1 ± 86.1-
   POC export7.4 ± 9.96%17.5 ± 16.035%
   Benthic C respiration3.2 ± 2.32%4.0 ± 2.68%
   Sediment OC1.4 ± 0.51%1.3 ± 0.53%
   POC export - benthica2.9 ± 4.32%12.2 ± 13.724%
   DOC accumulation  3.0 ± 1.96%
 
Totalb
   PP124.4 ± 88.1-89.9 ± 146.9-
   POC export19.7 ± 24.816%20.0 ± 14.522%
   Benthic C respiration5.7 ± 5.65%6.5 ± 7.07%
   Sediment OC2.0 ± 1.42%1.5 ± 0.42%
   POC export - benthica12.0 ± 19.210%12.0 ± 15.513%
   DOC accumulation  3.6 ± 3.94%

[30] POC fluxes exceeded rates of benthic carbon respiration at most stations during the spring and summer of 2004 (p value = 0.08 during the spring; p value <0.01 during the summer) (Figure 6b). POC export fluxes were exceptionally high at Barrow Canyon, and exceeded rates of benthic carbon respiration by a factor of 2–10. Benthic carbon respiration was measured from oxygen uptake during sediment incubation experiments conducted during each cruise [Cooper et al., 2002]. The imbalance between benthic respiration and POC export fluxes is probably due to a time lag for a benthic response to seasonally high POC export [Moran et al., 2005]. However, benthic carbon respiration rates measured in the spring (average = 5.7 ± 5.6 mmol C m−2 d−1) were similar to rates measured during the summer (6.5 ± 7.0 mmol C m−2 d−1) (p value = 0.74), despite excesses in POC export averaging 12 ± 19 mmol C m−2 d−1 during the spring and 12 ± 16 mmol C m−2 d−1 during the summer (Table 3). The implication is that the rate of benthic respiration does not match the POC export flux during periods of high productivity. However, it is possible that this excess POC is remineralized in the water column below 50 m, or is consumed by the benthic community on longer, i.e., annual, timescales. It should be noted that these benthic carbon respiration rates are calculated from oxygen consumption, and neglect nonoxygen respiration. Therefore these respiration rates should be regarded as lower estimates of the total respiration that occurs in the sediments.

4.4. Seasonal Patterns in Particle Export and Deposition

[31] The water column-sediment balance of 234Th provides an indication of the pattern of particle export and sedimentation occurring on a seasonal timescale. If deposition of 234Th occurs directly to underlying sediments, then the sediment inventory should be equal to the deficit measured in the water column. An excess or deficiency of sedimentary 234Th is indicative of sediment focusing or export, respectively [Cochran et al., 1995], and can be quantitatively expressed as the focusing factor (FF) [Cochran et al., 1990],

equation image

where a focusing factor >1 indicates sediment focusing, and a focusing factor <1 indicates sediment export on a timescale of days to weeks. Sediment 234Th activities and inventories, focusing factors, and POC concentrations are listed in auxiliary material Table S3.

[32] Despite mixing and bioturbation in shelf and slope sediments, some seasonal changes in 234Th inventories were observed between the spring and summer of 2004. During the spring, focusing factors were near 1 at shallow shelf stations, and decreased with increasing water depth toward the slope and basin (Figure 7). Extremely low focusing factors calculated at EHS6 and BC5 are attributed to a time lag between the scavenging removal of 234Th from the water column and its deposition in underlying sediments, rather than a net export of particles from the region. During the summer, the focusing factor was ∼0.5 at BC2, but increased to 3.1 and 1.9 at BC3 and BC4, respectively (Figure 7). The observed increase in the focusing factor at BC indicates that particles are transported from the shelf to Barrow Canyon, implying that this is a region of net particle deposition. High-current velocities and concentrations of suspended particulate matter suggest that sediments deposited in Barrow Canyon may be transported to the interior basin [Moran et al., 2005; Weingartner et al., 1998]. However, focusing factors calculated along the BC transect indicate that this export is small relative to the deposition of particles within Barrow Canyon.

Figure 7.

Sediment focusing factors calculated using the water column and sediment 234Th inventories during the spring (black) and summer (white), 2004. Hatch marks are placed on the bars at intervals of 1.

4.5. Shelf-Slope POC Budget

[33] A shelf-slope budget of POC was constructed using rates of primary productivity, POC export, benthic carbon respiration, and sedimentary organic carbon loss in the spring and summer of 2004. These fluxes are compared to those measured in 2002, and are listed in Table 3. The sedimentary loss of organic carbon was calculated as the product of the organic carbon concentration in surface sediments and a 210Pb-derived sedimentation rate, and represents the loss of OC to burial and remineralization below 1–2 cm in shelf and slope sediments. The net community production (NCP) was calculated by adding the average POC export flux for the summer to the rate of production of suspended POC in the upper 50 m of the water column. DOC accumulation was assumed to equal 15% of NCP, and is compared to the POC fluxes listed in Table 3 [Hansell et al., 1997a; Hansell and Carlson, 1998; Hansell et al., 1997b; J. T. Mathis et al., Interannual variability of net community production over the Northeast Chukchi Sea Shelf, submitted to Deep Sea Research, Part II, 2006]. In the spring of 2004, an average of 12 ± 19 mmol C m−2 d−1 was exported from surface waters but not consumed immediately during benthic carbon respiration or burial and oxidation in shelf and slope sediments (Figure 8a). The imbalance between the POC export flux and respiration and burial in underlying sediments was more pronounced over the shelf (<400 m water depth) than the slope (>400 m water depth), with an excess of POC export flux reaching 19 ± 26 mmol C m−2 d−1 during the spring (Table 3). In contrast, POC export fluxes were balanced by benthic carbon respiration during the spring of 2002 [Moran et al., 2005].

Figure 8.

Seasonal fluxes of POC (mmol C m−2 d−1) measured during the (a) spring and (b) summer of 2004 in the shelf-slope region of the Chukchi Sea. All fluxes are reported in units of mmol m−2 d−1. The ΔPOC value represents the difference between the POC export flux and the sum of the benthic carbon respiration and sedimentary OC loss.

[34] The imbalance between POC export fluxes and consumption in underlying sediments measured in the summer of 2004 (12 ± 16 mmol C m−2 d−1; p value < 0.01) was similar to that measured in the spring of 2004 (12 ± 19 mmol C m−2 d−1; p value = 0.09) (Table 3 and Figure 8). More stations were occupied over the shelf than the slope, so the average ΔPOC listed in Figure 8 is weighted toward the excess POC export flux measured over the shelf. In 2002, the imbalance between POC export and benthic carbon respiration increased to nearly 20% of primary productivity between the spring and summer (6.1 mmol C m−2 d−1) [Moran et al., 2005]. Because sampling was conducted during the spring bloom in 2004, seasonal changes in POC fluxes were not as pronounced as in 2002. However, the shelf-slope POC budget from 2004 indicates that particulate organic carbon exported from surface waters was not immediately consumed during benthic respiration, nor was an appreciable amount buried or oxidized in shelf and slope sediments.

[35] The imbalance between POC export fluxes and benthic carbon respiration has been attributed to off-shelf export [Moran et al., 2005]. However, 234Th inventories measured in this study indicate that sediment focusing results in retention of particulate matter in the shelf-slope region (Figure 7). During the spring and summer of 2004, 12 ± 19 mmol C m−2 d−1 and 12 ± 16 mmol C m−2 d−1, respectively, was exported from surface waters but not accounted for by benthic carbon respiration or burial and oxidation in underlying sediments. It is most likely that much of this excess POC was remineralized in the water column below 50 m, especially at the deeper slope stations. In addition, POC that reaches the surface sediments could be consumed during anaerobic respiration, or stored as benthic biomass. A benthic macrofaunal biomass of up to 30–60 g C m−2 has been reported in the southern Chukchi Sea [Grebmeier, 1993]. Integrated over a 100 d growing season, 30–60 g C m−2 corresponds to a carbon flux of 25–50 mmol C m−2 d−1. While this flux is an upper estimate due to the uncertainty associated with extrapolation of a few measurements to the entire shelf region, the implication is that carbon storage in benthic biomass may be an important component of the carbon budget in this Arctic shelf-slope region.

[36] Sediment focusing measured in Barrow Canyon increased by a factor of 2–3 from spring to summer in 2004. Owing to limited sediment 234Th data during the summer, particulate transport into or out of Barrow Canyon cannot be determined definitively. However, rates of benthic carbon respiration up to 30 mmol C m−2 d−1 within this canyon indicate enhanced consumption of organic carbon. The increase in sediment focusing in Barrow Canyon during the summer could be due to the transport of particulate material, including POC, from the EHS and WHS transects, which would make Barrow Canyon an important region of sediment deposition in the Chukchi Sea.

5. Conclusions

[37] Despite the strong seasonality of primary productivity and POC export in the Arctic Ocean, 234Th-derived POC export fluxes measured in the shelf-slope region of the Chukchi Sea were similar during the spring and summer of 2004. Higher POC fluxes observed in the spring of 2004 compared to the spring of 2002 are attributed to a plankton bloom. POC export fluxes averaged 19.7 ± 24.8 mmol C m−2 d−1 in the spring and 20.0 ± 14.5 mmol C m−2 d−1 in the summer of 2004. In contrast, a threefold increase in POC export was observed between the spring and summer of 2002 (spring = 2.9 ± 5.3 mmol C m−2 d−1; summer = 10.5 ± 9.3 mmol C m−2 d−1) [Moran et al., 2005]. A three-dimensional advective transport model indicates that a one-dimensional model may underestimate the true 234Th flux as a result of significant cross-shelf transport of 234Th at the East Barrow transect.

[38] An organic carbon budget for this shelf-slope region indicates that up to 60% of POC exported from the surface water is not rapidly consumed during benthic carbon respiration or burial and oxidation in shelf sediments. Additionally, sediment focusing in Barrow Canyon during the summer indicates that sediments, and associated OC, may be transported to this region on a seasonal timescale. This freshly produced POC may be remineralized in the water column below 50 m, consumed during benthic respiration throughout the year, or incorporated into the benthic biomass and thereby retained in shelf sediments over biological timescales. Alternatively, this excess POC, which comprises up to 10–20% or primary productivity, may be transported to this canyon and exported to the interior basin over longer (i.e., interannual) timescales.

Acknowledgments

[39] We thank the Captain, officers, marine science technicians, and crew of the U.S.C.G.C. Healy for their assistance with field sampling. We also thank J. N. Smith for assistance with sample collection. Two anonymous reviewers are thanked for providing comments that improved an earlier version of this manuscript. This work was supported by the NSF (OPP-0124917 to S. B. M.; OPP-0125082 to J. M. G. and L. W. C.; OPP-0124943 to W. M.; OPP-0124868 to N. R. B.; OPP-0124900 to D. A. H.).

Ancillary