Polar Biology

, Volume 28, Issue 3, pp 238–249

Food web structure in the high Arctic Canada Basin: evidence from δ13C and δ15N analysis

Authors

    • Institute of Marine Science, School of Fisheries and Ocean SciencesUniversity of Alaska Fairbanks
  • BA Bluhm
    • Institute of Marine Science, School of Fisheries and Ocean SciencesUniversity of Alaska Fairbanks
  • R Gradinger
    • Institute of Marine Science, School of Fisheries and Ocean SciencesUniversity of Alaska Fairbanks
Original Paper

DOI: 10.1007/s00300-004-0669-2

Cite this article as:
Iken, K., Bluhm, B. & Gradinger, R. Polar Biol (2005) 28: 238. doi:10.1007/s00300-004-0669-2

Abstract

The food-web structure of the Arctic deep Canada Basin was investigated in summer 2002 using carbon and nitrogen stable isotope tracers. Overall food-web length of the range of organisms sampled occupied four trophic levels, based on 3.8‰ trophic level enrichment (δ15N range: 5.3–17.7‰). It was, thus, 0.5–1 trophic levels longer than food webs in both Arctic shelf and temperate deep-sea systems. The food sources, pelagic particulate organic matter (POM) (δ13C=−25.8‰, δ15N=5.3‰) and ice POM (δ13C=−26.9‰, δ15N=4.1‰), were not significantly different. Organisms of all habitats, ice-associated, pelagic and benthic, covered a large range of δ15N values. In general, ice-associated crustaceans (δ15N range 4.6–12.4‰, mean 6.9‰) and pelagic species (δ15N range 5.9–16.5, mean 11.5‰) were depleted relative to benthic invertebrates (δ15N range 4.6–17.7‰, mean 13.2‰). The predominantly herbivorous and predatory sympagic and pelagic species constitute a shorter food chain that is based on fresh material produced in the water column. Many benthic invertebrates were deposit feeders, relying on largely refractory material. However, sufficient fresh phytodetritus appeared to arrive at the seafloor to support some benthic suspension and surface deposit feeders on a low trophic level (e.g., crinoids, cumaceans). The enriched signatures of benthic deposit feeders and predators may be a consequence of low primary production in the high Arctic and the subsequent high degree of reworking of organic material.

Introduction

At high latitudes, the pronounced seasonality and the quantity of food resources are more likely to restrain growth and survival of the fauna than low temperature per se (Clarke 1998). In the Arctic Ocean, sea-ice cover and light availability cause short seasonal pulses of sea ice and pelagic primary production (Gosselin et al. 1997), which are the food sources for ice-associated (sympagic), pelagic and benthic organisms (e.g., Grebmeier and McRoy 1989; Smith and Schnack-Schiel 1990; Gradinger 2002).

The Arctic benthos is among the ultimate recipients of carbon fixed in primary production in the overlying water mass or within the sea ice, through the flux of sinking material from the euphotic zone (Grebmeier et al. 1989). The quantity and quality of food that is provided for the benthic community depend on nutrient availability and overall primary production in the water column, water stratification and mixed layer, midwater grazing rates and bacterial degradation (Fenchel 1988). This food input occurs in the form of direct sedimentation of phytoplankton, or through fecal pellets, zooplankton carcasses, molts and marine snow. In areas of high nutrient availability and high primary production, and simultaneously low zooplankton grazing, the coupling between the pelagic and the benthic system is tight, meaning that the benthos is provided with a rich food source (Dunton et al. 1989). In these high productivity systems, bottom-up processes are thought to drive the pelagic and benthic community and food-web structures (Dunton et al. 1989; Hunt et al. 2002).

Trophic structure and pelagic-benthic coupling is reasonably well understood for certain shallow shelf regions of the Arctic, such as the northern Bering and Chukchi Seas (Dunton et al. 1989; Grebmeier and McRoy 1989), the Canadian Arctic (Hobson and Welch 1992) and the Northeast Water Polynya (Hobson et al. 1995; Piepenburg et al. 1997). In contrast, we have very limited knowledge of energy flow and trophic structure of Arctic deep-sea regions, such as the Canada Basin. While we know from other deep-sea areas that diverse benthic communities can be sustained solely by short seasonal oceanic primary-production pulses (Mahaut et al. 1990; Rice and Lambshead 1994; Iken et al. 2001), it is unknown how the seasonal production pulses are linked to the deep-sea community in the Arctic. Previous isotope studies in Arctic Chukchi and Beaufort shelf regions (e.g., Saupe et al. 1989; Schell et al. 1998) found differences in the isotopic signatures of planktonic organisms with longitude. The present study presents the first, though incomplete, opportunity to revisit this longitudinal trend for plankton organisms from offshore locations and to include benthic organisms along the same longitudinal gradient.

The seasonality of Arctic marine food webs is largely driven by physical and chemical processes (Hunt et al. 2002). In such bottom-up driven systems, changes in productivity and grazing patterns in the overlying water column will have cascading and sometimes indirect effects on the community and food-web structure of the benthos. Changes in ice cover and primary productivity would be expected with the increasing trend of global climate warming (Stenseth et al. 2003). Hence, the analysis and understanding of Arctic marine food-web structure might reveal the impact of long-term productivity changes on energy flow patterns and processes. Therefore, stable isotope analyses were used to evaluate the food-web structure and the connection between the sympagic, pelagic and benthic systems in the deep Canada Basin.

Materials and methods

Investigation area

Samples were collected in August 2002 during a cruise with Louis St. Laurent to the Canada Basin (Fig. 1). Water depth at the stations sampled ranged from 625 m in the Amundsen Gulf, at the eastern end of the cruise track, to 3,398 m in the deep part of the Canada Basin. Depths between 800 and 2,082 m were sampled at Northwind Ridge at the western end of the cruise track (Table 1, Fig. 1). For description of the oceanographic features of the study area, refer to McLaughlin et al. (2004).
Fig. 1

Map of the study area with sampling stations marked along the cruise track

Table 1

List of stations sampled during the 2002 Canada Basin cruise in chronological order for food-web analysis (POM particulate organic matter, NA not applicable; source of material: B benthic, I ice-associated, P pelagic, S sediment sample)

Station no.

Date (dd/mm/yy)

Latitude (°N)

Longitude (°W)

Depth (m)

Gear

Sample type

Water depth sampled (m)

AG05

16/08/02

70°35′

122°59′

625

Box core

B

625

  

Bongo net

P

100

  

AL07

18/08/02

71°42′

134°42′

1,568

Box core

B

1,568

  

Niskin bottles

P-POM

50

750

1,568

AL09

19/08/02

72°38′

136°14′

2,575

Niskin bottles

P-POM

 

500

2,500

Ice sampling

I-POM, I

NA

  

AL10

20/08/02

73°29′

137°00′

3,170

Box core

S

3,170

  

Bongo net

P

100

  

Niskin bottles

P-POM

 

1,000

3,124

Ice sampling

I-POM, I

NA

  

RVB1

24/08/02

72°06′

139°50′

2,765

ROV

S

2,765

  

SB01

23/08/02

72°33′

141°00′

3,220

Niskin bottles

P-POM

50

600

2,500

SB02

25/08/02

72°44′

145°02′

3,550

Niskin bottles

P-POM

50

600

2,500

SB03

26/08/02

74°14′

148°22′

3,850

Bongo net

P

100

  

Ice sampling

I-POM, I

NA

  

Niskin bottles

P-POM

 

600

3,398

NW08

27/08/02

76°46′

148°57′

 

Bongo net

P

100

  

Ice sampling

I-POM, I

NA

  

NW07

30/08/02

75°33′

153°08′

3,900

Bongo net

P

100

  

Niskin bottles

P-POM

42

800

3,414

NW05

02/09/02

75°56′

155°39′

1,850

Box core

B, S

1,850

  

Bongo net

P

100

  

2,082

Niskin bottles

P-POM

 

800

2,082

NW03

02/09/02

75°59′

156°11′

1,220

Niskin bottles

P-POM

35

400

1,220

NW01

03/09/02

75°58′

156°49′

800

Box core

B, S

800

  

Bongo net

P

100

  

880

Niskin bottles

P-POM

34

400

878

NWR02

05/09/02

74°49′

161°38′

1,550

Niskin bottles

P-POM

41

375

1,534

NA05

06/09/02

74°20′

162°19′

1,350

Box core

B, S

1,350

  

NA09

06/09/02

75°09′

160°12′

2,000

Niskin bottles

P-POM

42

790

1,988

Sampling and sample treatment

The analysis of naturally occurring stable carbon and nitrogen isotopes in the tissue of organisms is an efficient tool to establish food-web relations between species on the scale of communities or ecosystems (Michener and Schell 1994). Stable isotopes show a stepwise enrichment between trophic levels of about 1‰ for δ13C and 3.3–3.8‰ for δ15N (DeNiro and Epstein 1981; Rau 1982; Hobson and Welch 1992; Post 2002), thus allowing identification of relative trophic positions among members of a food web. For the purpose of this study, we used a nitrogen enrichment factor of 3.8‰ to ensure comparability with other Arctic and deep-sea food-web studies (see Discussion). Stable isotope composition of body tissue reflects feeding patterns of the last several weeks to few months rather than just the most recent diet.

The overall number of samples and replicates collected was limited by logistical restrictions with ship time and winch equipment. Equipment failure precluded the collection of any larger and mobile epifaunal organisms by trawls or ROV manipulative arms, and the presence of epifaunal organisms could only be observed by ROV (see Bluhm et al. 2004). A list of stations is given in Table 1. Benthic samples were taken with a 0.25-m2 spade box core, and macrofaunal organisms were collected from approximately the upper 5-cm layer. Samples were sieved over a 500-μm sieve. At five stations we measured isotopic ratios of surface sediments collected from box core samples as the food source for deposit feeders (Table 1). Plankton organisms were taken from plankton bongo net samples (236 μm). Water-column particulate organic matter (P-POM) was obtained from water samples collected with Niskin bottle samplers to serve as baseline reference for the food web. These bottles were equipped with an in-situ fluorometer for sampling the sub-surface Chl a maximum layer, which was usually located between 30 and 50 m depth (Lee and Whitledge 2004). Samples were filtered onto pre-combusted 2.5-cm GF/F filters. Larger organisms (e.g., visible zooplankton specimens) were removed from the filters. The second baseline reference for trophic relationships was POM (mainly ice algae) from ice cores (I-POM). Ice cores from multi-year sea ice of 1.4–2.7 m thickness were collected with a Kovacs ice auger. The bottom 15-cm sections were melted in the dark and a sub-sample was filtered onto pre-combusted GF/F filters (see Gradinger et al. 2004). Both the pelagic and sea-ice POM filters were rinsed with deionized water prior to freezing. Where possible, pelagic and benthic animals were sampled individually and in replicates. In many cases, several individuals per species from the same station were combined to provide a large enough sample because of the generally small size of the organisms captured. Complete individuals instead of selected tissue types were measured for the same reason (Table 2).
Table 2

Mean δ13C and δ15N ratios of species or higher taxonomic groups in the Arctic Canada Basin. Where possible, organisms were measured individually; otherwise several individuals from the same station were pooled. Means±SE are given for three and more replicate measurements. Individual values are given for single and duplicate measurements. Feeding-type categories are based on isotopic signatures and literature references (see text) (FF filter feeder, SF suspension feeder, sSDF selective surface deposit feeder, uSDF unselective surface deposit feeder, SSDF subsurface deposit feeder, P predator, P/S predator/scavenger, NA not applicable). Uncertain designations are marked ‘?’

Taxonomic group/species

Station

Habitat

Individual/pooled

Replicates

δ13C (mean±SE)

δ15N (mean±SE)

Feeding type

Foraminifera

RVB1

Benthic

Pooled

1

−24.56

4.04

?

Cnidaria

 Nausithoe sp.

NW01

Benthic

Pooled

4

−19.99±0.71

13.34±0.23

FF

 Sminthea arctica

NW08

Pelagic

Individual

3

−24.22±0.47

15.79±0.41

P

Mollusca

 Bivalvia

  Bathyarca sp.

AG05

Benthic

Pooled

1

−22.34

12.66

uSDF

  Thyasiridae

AG05

Benthic

Pooled

1

−21.18

12.40

uSDF

  Dacrydium cf. viteum

AG05

Benthic

Pooled

1

−20.12

10.23

SF

  Yoldiidae

NA05

Benthic

Pooled

1

−18.79

14.18

SSDF

 Gastropoda

  Limacina helicina

NW07

Pelagic

Pooled

1

−22.39

5.91

FF

 Scaphopoda

  Siphonodentalium lobatum

NA05

Benthic

Pooled

1

−20.99

10.74

sSDF

Polychaeta

 Unidentified sp. 1

AG05

Benthic

Individual

1

−18.97

15.30

?

 Unidentified sp. 2

AG05

Benthic

Individual

1

−20.66

15.47

?

 Unidentified sp. 3

AG05

Benthic

Pooled

1

−21.29

10.21

sSDF ?

 Unidentified sp. 4

NA05

Benthic

Pooled

1

−19.48

13.68

sSDF ?

 Unidentified sp. 5

NA05

Benthic

Pooled

1

−19.66

13.41

sSDF ?

 Unidentified sp. 6

NA05

Benthic

Individual

1

−20.33

8.55

sSDF ?

 Ophelina cylindricaudata

AG05

Benthic

Pooled

1

−19.50

15.25

SSDF

AL07

Benthic

Pooled

1

−19.36

14.25

 

 Capitellidae

AG05

Benthic

Individual

1

−17.94

16.08

SSDF

 Dorvillea cf. rudolphi

AG05

Benthic

Individual

1

−20.72

16.16

SSDF

 Prionospio sp.

AG05

Benthic

Pooled

1

−19.82

14.50

sSDF

 Chaetozone setosa

AL07

Benthic

Pooled

1

−23.25

15.34

uSDF

 Glycinde wireni

NW05

Benthic

Individual

1

−21.37

17.67

P

 Nephtys cf. malmgreni

NA05

Benthic

Individual

1

−17.07

17.43

P

 Terebellides stroemi

NA05

Benthic

Individual

1

−19.04

16.48

SSDF

 Nereidae

NA05

Benthic

Individual

1

−17.41

17.37

P

Echiura

 Unidentified sp. 1

NA05

Benthic

Individual

1

−17.77

16.57

SSDF

Crustacea

 Copepoda

  Paraeuchaeta sp.

AG05

Pelagic

Individual

3

−25.39±0.05

11.85±0.07

P

NW07

Pelagic

Individual

3

−27.95±0.24

13.08±0.32

 

NW05

Pelagic

Individual

2

−26.58 and −27.33

12.69 and 13.04

 

  Metridia longa

NW07

Pelagic

Individual

8

−25.35±0.86

12.20±0.48

FF

NW01

Pelagic

Individual

 

−25.98±0.20

11.66±0.18

 

  Pseudocalanus sp.

AL09

Sympagic

Pooled

1

−29.64

5.51

FF

NW08

Sympagic

Individual

3

−27.21±0.57

5.48±0.83

 

  Calanus hyperboreus/glacialis

AG05

Pelagic

Individual

4

−26.67±0.19

9.44±0.22

FF

AL10

Pelagic

Individual

4

−25.91±0.54

10.52±0.62

 

NW08

Pelagic

Individual

3

−25.70±0.16

9.57±0.33

 

NW07

Pelagic

Individual

3

−26.37±0.11

9.27±0.34

 

NW05

Pelagic

Individual

3

−25.39±0.32

9.83±0.39

 

NW01

Pelagic

Individual

3

−25.68±0.08

10.15±0.09

 

NA05

Pelagic

Individual

7

−25.83±0.49

9.62±0.41

 

  Oithona similes and other small species

AL10

Pelagic

Pooled

1

−31.67

NA

FF

 Amphipoda

  Gammarus wilkitzkii (juveniles)

AL10

Sympagic

Individual

3

−24.65±0.29

7.10±0.14

FF

  Apherusa glacialis

NW08

Sympagic

Individual

3

−22.26±3.31

7.06±0.32

FF

SB03

Sympagic

Individual

4

−27.11±1.27

6.47±0.20

 

  Onisimus sp.

AL10

Sympagic

Individual

1

−23.58

12.44

P/S, SDF

  Parathemisto libellula

AG05

Pelagic

Individual

3

−24.57±0.47

10.81±0.19

P/S, FF, DF

NW07

Pelagic

Individual

1

−28.49

9.91

 

NW05

Pelagic

Individual

2

−28.03 and −27.09

11.01 and 11.02

 

  Hyperia sp.

AG05

Pelagic

Individual

2

−26.19 and −25.91

9.27 and 9.60

?

NW07

Pelagic

Individual

1

−26.57

12.08

 

NW05

Pelagic

Pooled

1

−27.22

11.13

 

  Pardaliscella cf. malygini

AG05

Benthic

Individual

3

−21.60±0.30

15.02±0.12

SDF

 Ostracoda

  Conchoecia borealis

AG05

Pelagic

Pooled

1

−26.94

12.66

P

NW07

Pelagic

Pooled

1

−28.17

11.85

 

NW08

Pelagic

Individual

3

−27.15±0.24

12.60±0.09

 

  Unidentified sp. 1

AG05

Benthic

Individual

1

−22.42

12.97

sSDF

 Tanaidacea

  Unidentified sp. 1

AG05

Benthic

Pooled

1

−20.90

11.80

sSDF

 Isopoda

  Gnathia sp. (male)

AG05

Benthic

Individual

2

−13.33 and −15.25

15.93 and 14.94

P/S

  Gnathia sp. (female)

AG05

Benthic

Individual

1

−21.13

17.23

P/S

 Cumacea

  Diastylis sp.

NA05

Benthic

Individual

1

−17.46

4.60

sSDF

  Eudorella sp.

NA05

Benthic

Individual

1

−19.47

5.37

sSDF

  Mixed species

NA05

Benthic

Pooled

1

−18.53

6.92

sSDF

 Decapoda

  Hymenodora glacialis

AL10

Pelagic

Individual

2

−22.36 and −21.35

16.57 and 15.22

P

NW05

Pelagic

Individual

2

−24.95 and −24.33

16.23 and15.82

 

Echinodermata

 Ophiuroidea

  Unidentified sp. 1

AG05

Benthic

Pooled

1

−23.52

11.70

sSDF

  Unidentified sp. 2

NA05

Benthic

Individual

1

−20.82

12.24

sSDF

 Crinoidea

  Unidentified sp. 1

AG05

Benthic

Individual

1

−17.97

6.86

SF

Chaetognatha

 Eukrohnia hamata

AG05

Pelagic

Pooled

1

−25.05

12.77

P

 

AL10

Pelagic

Pooled

3

−25.86±0.70

14.15±0.23

 
Samples were frozen at −20°C immediately after collection, and subsequently fumed with HCl to remove carbonates prior to analysis. Organisms with significant carbonate structures, e.g., ophiuroids, were soaked in 1 N HCl until bubbling ceased. All samples were dried prior to analysis. The stable isotope compositions of either a filter sample or approximately 0.3 mg homogenized organism sample were measured on a Thermo Finnigan Delta Isotope Ratio Mass-Spectrometer, with carbon PDB and atmospheric N2 as standards, at the stable isotope facility, at the University of Alaska, Fairbanks. Analytical error was 0.05‰ for 13C and 0.06‰ for 15N. Sample isotopic ratios are expressed in the conventional δ notation as parts per thousand (‰) according to the following equation:
$$ \delta X = [(R_{{\rm{sample}}} /R_{{\rm{standard}}} ) - 1] \times {\rm{1,000}} $$
where X is 13C or 15N of the sample and R is the corresponding ratio 13C/12C or 15N/14N.

Results

Isotopic composition of δ13C and δ15N values of P-POM identified at the Chl a maximum layer ranged between −27.7 and −23.0‰ (mean −25.8) and 4.2–5.9‰ (mean 5.3) (N=8), respectively. Variation between stations was fairly low, and there was no obvious longitudinal trend in P-POM isotopic ratios (Fig. 2). Isotopic composition of P-POM from intermediate and bottom water (see Table 1) was highly variable with δ13C values ranging from −27.4 to −18.9‰ and nitrogen isotope values ranging from 2.1 to 12.2‰. No distinct trends with depth or longitude were detectable. I-POM was slightly depleted compared to P-POM values with a range of −28.3 to −24.7‰ (mean −26.9) for δ13C and 2.3–6.5‰ (mean 4.1) for δ15N (N=4), also without any longitudinal trend (Fig. 2). Differences between mean P-POM and I-POM isotopic values were not statistically significant, using a t-test. Surface sediment (S-POM) values ranged from −25.1 to −21.7‰ for δ13C and 2.9–8.2‰ for δ15N (N=5). Mean isotopic values of S-POM (−22.9‰ for δ13C and 5.5‰ for δ15N) were slightly, but not statistically significantly (P>0.1, one-way ANOVA, SPSS 10.0 software), enriched compared to mean P-POM and I-POM values. There was a trend towards higher S-POM isotope values at the more western stations (Fig. 2, line included for indication of trend). P-POM, I-POM and S-POM values were single measurements per station so that no information about variability within a station could be obtained.
Fig. 2

Distribution of δ13C and δ15N values of ice POM, pelagic POM, surface sediments, and sympagic, pelagic and benthic organisms along a longitudinal gradient from west to east. Calanus spp. are included in the group of pelagic organisms and their isotopic composition is also depicted separately to show longitudinal relations within a narrow taxonomic group. The dashed line indicates the only noticeable trend in isotopic depletion from west to east in surface sediments

Sympagic organisms spanned about 6‰ (−29.6 to −23.3) in δ13C and about 8‰ (4.6–12.4) in δ15N values. Comparatively, pelagic species spanned a slightly wider isotopic range for both isotope ratios, with about 7‰ (−28.5 to −21.3) for δ13C and about 9‰ (7.8–16.6) for δ15N values. The largest range in isotope ratios was covered by benthic species with about 10‰ (−24.6 to −14.3) for δ13C and about 13‰ (4.6–17.7) for δ15N values. No longitudinal trends in isotope values could be deciphered in any of the three realms (Fig. 2). However, such a trend may be obscured by varying proportions of different feeding types and by differences in the type and amount of species measured along the latitudinal range sampled. Calanus hyperboreus/glacialis were consistently collected at most stations along the longitudinal gradient and, therefore, isotopic composition of these copepods were extracted from the pelagic-organism group and plotted again separately along the longitudinal axis (Fig. 2). As for all pelagic organisms, C. hyperboreus/glacialis did not show any trend with longitude in either δ13C or δ15N values. Sample coverage was too low to allow meaningful analysis of depth-related trends of benthic isotopic signatures.

Stable isotope ratios of individual species are given in Table 2. The nitrogen isotope ratios (Fig. 3) were used for identification of food-web relations between the collected taxa. Based on the mean P-POM δ15N value (5.3‰) as the reference, nitrogen isotope values span about 13‰. Based on the assumed 3.8‰ enrichment of δ15N per trophic level (TL) (Hobson and Welch 1992), the high Arctic food web consisted of four trophic levels for those species sampled within this study. Most sympagic organisms were located within the first TL (5.3–9.1‰), and most pelagic organisms were in the second (9.1–12.9‰) and a few also in the third TL (12.9–16.7‰). Most benthic organisms were placed in the second and third TL, and also the few animals occupying the fourth TL (16.7–20.5‰) were all benthic.
Fig. 3

Distribution of δ15N values of organisms from the Arctic Canada Basin by taxonomic groups. Trophic levels (TL1–4) are indicated in 3.8‰ enrichments steps with pelagic POM as baseline reference

Discussion

Constantly cold temperatures and high seasonality are biologically important characteristics of polar conditions (Clarke 1998). Seasonality is most noticeable in ice extent and light availability, both of which have significant influence on the ice-associated and pelagic primary production that form the base of the oceanic food web (Dayton et al. 1994). Most Arctic organisms are well adapted to the low temperature regime (Pörtner and Playle 1998). In addition, the seasonality and limited amount in food supply may pose constraints on the type of community and food-web structure that can be supported. Very similar conditions and constraints apply for deep-sea communities in general (Gage and Tyler 1991; Iken et al. 2001). Therefore, we compared food-web structure and prevalent feeding types in the Arctic deep Canada Basin with Arctic shelves and other deep-sea environments.

Sources of the high-Arctic food web

The two sources of marine primary production in the offshore high Arctic, sea-ice algae and phytoplankton (both measured here as POM), were not significantly different in their isotopic composition at the time sampled. This was surprising because recent studies (e.g., Schubert and Calvert 2001; Kennedy et al. 2002; Thomas and Papadimitriou 2003) found sea-ice algae to be isotopically heavier than phytoplankton. Limited CO2 availability within the brine channel network can make sea ice an isotopically closed system where CO2 is only slowly replenished (Thomas and Papadimitriou 2003). As a consequence, ice algae may be more enriched compared to the second source, phytoplankton (e.g., Hobson et al. 1995; Schubert and Calvert 2001). The lack of isotopic difference in ice algae and phytoplankton in this study may be because both biomass and productivity within the summer sea ice of the Arctic Ocean are low (Gosselin et al. 1997; Gradinger et al. 2004) compared to spring bloom values in Arctic fast-ice systems (Gradinger 1999). The positive correlation between δ13C and total POC, established for Arctic fast ice (R. Gradinger, B.A. Bluhm, M.R. Nielson, unpublished work), suggests that no isotopic enrichment of ice algae can be expected at times of low biomass such as in summer 2002 in the Canada Basin. In addition, the sea ice was very porous at the time of sampling in late summer, allowing increased water and CO2 exchange in this system (Thomas and Papadimitriou 2003). The similarity in water-column and sea-ice POM isotopic signatures prohibited the tracing of the relative importance of these two sources throughout the food web.

The benthic system is coupled to pelagic and ice-associated primary and secondary production through the sinking of organic material in the form of phytoplankton cells and fecal pellets and carcasses of sympagic and pelagic animals (Lampitt et al. 1993). Phytodetritus and other organic material is subject to bacterial degradation during sinking, altering the isotopic signature of this food source before its deposition on the seafloor (Mako and Estep 1984). The variation among pelagic POM values measured at various depths, however, was too large to detect any depth-related trends in isotopic signatures. Surface sediments as the repository of this phytodetritus and the food source for benthic deposit feeders were slightly enriched in carbon isotopes compared to water-column and ice-associated POM (Fig. 2), an indication of reworking by biological processes during downward transport. The relatively higher enrichment of surface sediments at the western stations may in part be attributed to the influence of the highly productive and isotopically enriched Chukchi Shelf waters (Dunton et al. 1989; Naidu et al. 2000). High primary productivity along the Bering Sea shelf break is discussed as a likely reason for the enrichment in isotopic signature in the northwest Bering Sea (Schell 2000). The nutrient-rich and isotopically enriched Bering Sea-Anadyr Water then crosses the Chukchi Shelf (Walsh et al. 1989, 2004). It may reach the deeper western Canada basin sediments and bottom communities by cross-shelf transport through Barrow Canyon and other down-slope pathways (Swift et al. 1997; Weingartner et al. 1998; Cooper et al. 1999; R.A. Woodgate, K. Aagaard, T. Weingartner, unpublished work). The eastern stations in the Canada Basin, in contrast, may be under the influence of the McKenzie River delta, which carries depleted, terrestrial isotope signatures compared to the marine system (Naidu et al. 1993; Macdonald et al. 2002). This could be one explanation for the higher isotopic signatures in western compared to eastern basin sediments. This west-east difference in S-POM, however, was not obvious in the benthic organisms sampled in this study.

Food-web structure

Nitrogen isotopes are a particularly good indicator of trophic levels, with higher δ15N values signifying higher trophic levels (Post 2002). Benthic animals as a whole had higher nitrogen isotope ratios and thus represented higher trophic levels than the pelagic fauna, which in turn was enriched compared to sympagic animals. Within the trends of these general groupings, there are distinct differences among organisms within these habitat groups. Following, organisms are discussed according to their position in the trophic system (Fig. 3).

Although P-POM and I-POM were not distinguishable in their isotopic composition, isotopic values of the sympagic organisms suggest a closer connection to ice-associated than pelagic production. Sympagic animals integrate isotopic ratios over longer time periods than I-POM, because of slower turnover rates in invertebrates compared to microalgae. Therefore, sympagic invertebrates may still reflect isotopic enrichment within the sea-ice system from a more productive and thus isotopically heavier situation in earlier spring/summer (R. Gradinger, B.A. Bluhm, M.R. Nielson, unpublished work), while I-POM had the already depleted signature of low productivity in late summer during the time of sampling. Among the ice-associated animals, the depleted δ15N of the copepod Pseudocalanus sp. indicated a herbivorous feeding strategy, presumably mainly on ice algae (Runge et al. 1991). The ice-associated amphipods within the first TL, adult Apherusa glacialis and juvenile Gammarus wilkitzkii were similar to each other and slightly enriched in δ15N compared to Pseudocalanus sp. This corroborates the herbivorous-detritivorous feeding mode suggested for A. glacialis (Werner 1997; Poltermann 2001) and herbivory of juvenile G. wilkitzkii (Scott et al. 2001) compared to the detritivorous and carnivorous adults in G. wilkitzkii (Poltermann 2001; Werner et al. 2002). The pelagic pteropod, Limacina helicina, was also part of the first TL. This confirms a predominantly herbivorous diet on phytoplankton (Falk-Petersen et al. 2001) rather than the carnivorous strategy proposed by Gilmer and Harbison (1991).

Other pelagic grazers were distinctly enriched and part of the second TL. C. hyperboreus and C. glacialis occurred regularly throughout the study area (Hopcroft et al. 2004) and are usually considered grazers on phytoplankton (Eilertsen et al. 1989; Hirche et al. 1994). Their position in TL2 suggests that, in addition to phytoplankton, they may also ingest smaller heterotrophic organisms and phytodetritus (Sato et al. 2002). The pelagic amphipod, Parathemisto libellula, occupied a similar omnivorous TL, suggesting a mixed diet on phytoplankton and small pelagic prey, which contradicts the proposed pure carnivory on Calanus spp. as based on laboratory experiments (Auel and Werner 2003). The sympagic amphipod Onisimus sp. is located towards the end of the second TL, which coincides more with a scavenging/predatory diet (Poltermann 2001; Borgå et al. 2002) than with a herbivorous feeding strategy proposed earlier (Werner 1997).

Distinctly predatory, pelagic species included the medusa Sminthea arctica, the decapod shrimp Hymenodora glacialis and the chaetognath Eukrohnia hamata, which were considerably enriched relative to the P-POM source and grouped in the third TL. The ubiquitous E. hamata is suggested to feed mainly on copepods (Froneman and Pakhomov 1998; Schell et al. 1998), most likely Calanus spp. (Eiane et al. 2002), and is, accordingly, enriched by one TL compared to Calanus spp. (Fig. 3).

Benthic species, in general, cover the full spread of trophic levels owing to their varying feeding strategies (Gage and Tyler 1991). There are only a few suspension feeders in most parts of the deep sea because of the limited availability of suspended material and low near-bottom flow (Thistle 2003). Suspension-feeders receive particles directly from the sedimentation process out of the water column or through resuspension. Sedimenting particles are richer in their nutritional content before they are deposited and reworked by benthic processes (Baird et al. 1985). It seems that the immediate connection to the sedimenting phytodetritus places suspension-feeders such as the mytilid Dacrydium cf. viteum, the crinoid and the cumaceans at a relatively low position in the food web (TL 1 and 2). Most surprising among these are the cumaceans with individuals measured close to and even below the POM δ15N baseline value. This very low TL cannot be fully explained at this time in terms of fractionation patterns or food sources, but it is an indication that very fresh material does reach the seafloor and can obviously be utilized by certain organisms. This is in agreement with field and experimental results in which cumaceans used phytodetritus as their major food source (L. Levin, Scripps Institute of Oceanography, personal communication).

As in most deep-sea benthic environments (Gage and Tyler 1991), the most common feeding mode in this study was deposit feeding. The isotopic spread of these deposit feeders may be explained by their differential selectivity towards certain particles. Selective surface feeders may select for the more freshly deposited particles on the seabed surface, which have not been reworked by biological processes numerous times and are thus isotopically lighter. Hence, selective deposit feeders would occupy lower trophic positions than unselective surface feeders, which ingest all particles regardless. In this study, organisms with the lighter isotopic signature of selective deposit feeders were part of the second TL and comprised a tanaidacean, two ophiuroids, a polychaete and several mollusc species (Table 2). Among these, for example, the scaphopod Siphonodentalium lobatum selectively feeds on benthic foraminiferans (Gudmundsson et al. 2003), which themselves can have distinctly depleted nitrogen isotope signatures (Fig. 3; also see Iken et al. 2001). In contrast, the majority of the benthic organisms clustered in the third TL are probably unselective surface deposit feeders or subsurface deposit feeders. Benthic feeding activities and bioturbation processes rework the same material continuously, thus enriching the deposited material even more and transferring it into deeper sediment layers where it becomes available to subsurface feeders (Clough et al. 1998). Unselective surface deposit-feeding polychaetes were represented by the cirratulid, Chaetozone setosa, and the spionid, Prionospio sp. (Fauchald and Jumars 1979). Subsurface deposit-feeding species included capitellids and the opheliid, Ophelina cylindricaudata (Fauchald and Jumars 1979; Holte and Gulliksen 1998). Similarly, the subsurface detritivorous bivalve of the family Yoldiidae (Holte and Gulliksen 1998) is a member of the enriched part of the third TL.

True benthic predators often are less common than deposit feeders in deep-sea systems because of the scarcity of benthic prey organisms (Gage and Tyler 1991). Likely predators in this study were the polychaetes Nephtys cf malmgreni, Glycinde wireni and the nereid polychaetes, which fell into the fourth TL. Members of these genera are known predators in shallower ecosystems (Fauchald and Jumars 1979).

In summary, the benthic food web was longer than the pelagic web owing to the continuous recycling and thus isotopic enrichment of food particles in the benthic system, while a more direct link to fresh phytodetritus existed in the pelagic system. However, several benthic species occupying relatively low trophic levels indicated that fresh phytodetritus reaches the Arctic deep-sea floor. Benthic deposit-feeding organisms had similar isotopic signatures as pelagic predators (TL3), while benthic predators occupied a higher TL than pelagic predators as they feed on an already enriched prey. Benthic epifaunal megafauna and the demersal ichthyofauna could not be sampled during our field expedition owing to equipment failure. Inclusion of these taxa could possibly result in an even longer benthic food web than elucidated in this study. Stable isotope composition of these more mobile organisms could yield additional valuable information on strategies of resource partitioning and alternative trophic pathways, e.g., a more direct link to the pelagic food web rather than having to rely on benthic refractory food sources (Iken et al. 2001).

Comparison with other Arctic and other deep-sea systems

The Arctic deep sea is ecologically particularly interesting because its fauna is exposed to both Arctic and deep-sea environmental conditions. Although the range of animals sampled in this study is not complete, some comparisons can be made with other food-web studies from the Arctic and the deep sea.

The data presented here do not confirm the longitudinal decrease of δ13C ratios in zooplankton, mainly copepods and chaetognaths, observed previously along a transect from the western Bering/Chukchi Sea to the eastern Alaskan Beaufort Sea (Saupe et al. 1989; Schell et al. 1998). In those studies, western locations in the Bering/Chukchi Sea were apparently under the influence of the high nutrient regime and enriched carbon isotope signal from the productive Bering Sea-Anadyr Water (Saupe et al. 1989; Schell et al. 1998). This signal caused δ13C enriched zooplankton communities at western but not at eastern locations. Unfortunately, no δ13C ratios for phytoplankton or POM are available from those two studies. Our zooplankton δ13C data from the Canada Basin were similar (Calanus spp.) or lighter by about 1% (Parathemisto sp., chaetognaths) than the most depleted values measured in the eastern Beaufort Sea by Saupe et al. (1989). This difference led us to the assumption that the influence of the high-nutrient Chukchi Shelf water with its enriched isotopic signatures may not reach the offshore epipelagic communities in the central Canada Basin. The western Chukchi Shelf surface water may be trapped to the west of the Northwind Ridge and over the Chukchi Cap, thus not influencing the surface waters in the Canada Basin or the eastern rise of the Northwind Ridge sampled in this study (Shimada et al. 2001). The sampled areas seem to be instead under the influence of the eastern Chukchi shelf water and the surface mixed layer water, carrying isotopically depleted water from the Alaska Coastal Current and the McKenzie River delta, respectively (Shimada et al. 2001; Macdonald et al. 2002).

Studies from the Northeast Water (NEW) Polynya east of Greenland (Hobson et al. 1995) and the Barrow Strait-Lancaster Sound region (Hobson and Welch 1992) analyzed the marine food-web structure on Arctic shallow continental shelf regions using stable isotopes. The fauna analyzed was comparable to that in the present study, facilitating a comparison of shallow and deep Arctic food webs. Pelagic POM δ15N values were similar in all studies (5.3‰ in Canada Basin, 5.4‰ in Lancaster Sound, 4.8‰ in NEW polynya); ice POM was only enriched in both shelf studies by 2–3‰ δ15N compared to pelagic POM. Overall food-web length, measured in 3.8‰ δ15N enrichment steps from pelagic POM to the most enriched benthic invertebrate species, was slightly shorter in both shelf studies compared to the present study: the food webs spanned three TL [8.4‰ in the NEW polynya (Hobson et al. 1995) and 10.7‰ in Lancaster Sound (Hobson and Welch 1992)] compared to four TL (12.4%) in the Canada Basin. The trophic levels of sympagic and pelagic organisms were comparable in all three studies. The difference in food-web length was mainly due to the benthic deposit-feeding and predatory species in the Canada Basin, which were considerably more enriched compared to the same feeding types at the shelf locations. The differences in food-web length may indicate a tighter coupling between pelagic production and benthic communities in shallow regions compared to the deep sea. In addition, this coupling seems to be tighter in high-productivity areas [e.g., NEW polynya (Hobson et al. 1995); Chukchi Shelf (Grebmeier and McRoy 1989; Grebmeier and Barry 1991)] compared to lower-productivity regions [Canada Basin and other central Arctic regions (Wheeler et al. 1996; Gosselin et al. 1997; Lee and Whitledge 2004; Gradinger et al. 2004)]. In the light of a changing environment, owing to globally changing climate and the expected strong impact on the Arctic system (Parkinson et al. 1999; Comiso 2002), it could be hypothesized that benthic shelf communities with tighter pelagic-benthic coupling would be more immediately impacted by changes in pelagic primary production than in the less tightly coupled Arctic deep-sea regions. More Arctic food-web studies with a wider sampling approach will be needed to gather evidence to test this hypothesis further.

A deep-sea food-web analysis similar to our study was conducted in the northeast Atlantic Porcupine Abyssal Plain (PAP) at about 5,000 m depth (Iken et al. 2001). PAP is a eutrophic system with a short seasonal pulse of phytodetritus as the main food input for the benthic system (Rice et al. 1994; Lampitt et al. 2001). Stable isotope analysis revealed three TL for the PAP food web (Iken et al. 2001) based on the same δ15N enrichment steps of 3.8‰ used in the present study. The δ15N of pelagic POM in this temperate system was much enriched (~8%) compared to the Arctic (~5.3%), emphasizing the importance of local baseline reference values. Most of the invertebrate benthos at PAP clustered in the second TL, with top predators in the third TL (Iken et al. 2001), while the Canada Basin benthic invertebrates clustered in the third and fourth TL, respectively. This suggests that the PAP benthic deep-sea system, despite its greater depth, may be more tightly coupled to phytoplankton production and less dependent on highly refractory material than the Arctic benthic system. High incidence of mobile epifaunal megafauna, e.g., holothurians and ophiuroids, at PAP are an indicator of the tight coupling to the primary production as these taxa effectively exploit fresh phytodetritus and thus remove a large part of the labile portion of annual deposition before it can be used by subsurface deposit feeders (Iken et al. 2001). In addition, total food supply to the PAP benthos is higher than in the Canada Basin. The seasonal primary production peak at PAP is on average about tenfold higher than in the Canada Basin (Lampitt et al. 2001; Gosselin et al. 1997; Lee and Whitledge 2004).

Acknowledgements

We acknowledge Fiona McLaughlin and Kathy Crane for leading the expedition. The captain and crew of the Louis St. Laurent provided important logistical assistance. Thanks go to Casey Debenham and Ian McDonald for unfaltering assistance in sieving box core samples for stable isotope samples. The Institute of Ocean Sciences group provided water samples, and Russ Hopcroft and Kevin Raskoff kindly provided the pelagic samples and their taxonomic identification. We would like to acknowledge the help of Nora Foster and Max Hoberg, both University of Alaska Fairbanks, for the identification of mollusc and polychaete samples, respectively. Tim Howe and Norma Haubenstock ran the stable isotope samples at the Alaska Stable Isotope Facility. Thanks go to Tom Weingartner for discussions on Arctic physical oceanography. Three anonymous reviewers provided excellent comments, which have greatly improved the quality of this paper. This project was funded through NOAA Ocean Exploration Office, grant no. NA16RP2627.

Copyright information

© Springer-Verlag 2004