Detritus from Ice and Plankton Algae as an Important Food Source for Macroinfaunal Communities in the Canadian Arctic

Abstract

Most deep-sea organisms feed on the organic matter produced in surface waters and settle on the seafloor. In polar regions, sea ice algal detritus and phytoplankton detritus are the main food sources for benthic fauna that reach the seafloor in pulses. Climate change affects the extension and duration of sea ice cover, which may affect the quantity and quality of food reaching the seafloor, resulting in less ice algae and more phytoplankton biomass. We conducted onboard pulse-chase experiments using sediment cores collected from Baffin Bay, Amundsen Gulf, and the Beaufort Sea to study how macroinfaunal communities in the Canadian Arctic use both food sources. Dual-labeled (¹³C and ¹⁵N) diatoms, Thalassiosira nordenskioeldii (phytoplankton treatment) and Synedra hyperborea (ice algae treatment), were used as tracers of food consumption by macroinfaunal groups. Community structure was analyzed in each region and differences were found among sites. The total uptake of both food sources was higher in Baffin Bay; the macroinfaunal biomass was the highest, with facultative filter/surface-deposit feeders accounting for more than 70% of the total biomass. The Baffin Bay station was the only location where there were notable variations in the biomass-specific uptake of ice algae and phytoplankton detritus by the bivalves and polychaetes, as well as by the community as a whole. At the same time, both food sources were consumed in equal quantities at the Amundsen Gulf and Beaufort Sea stations. This suggests that ice algae are not preferentially uptaken, and macroinfaunal communities may be resilient to a decrease in ice algal input to the seafloor inflicted by sea ice reduction.

1. Introduction

The sedimentation of phytodetritus is recognized as the main source of food supply for deep-sea fauna [1,2,3,4,5]. In Arctic ecosystems, algae occurring in sea ice (ice algae) and the water column (phytoplankton) are known as the main primary producers [6], and they form the phytodetrital pool that reaches the sea floor. Ice algae (mostly pennate diatoms) grow at the bottom and within the sea ice column and can reach substantial biomass, particularly during spring [7]. At the same time, phytoplankton blooms normally occur in early summer after ice breakup [8,9].

In seasonal ice zones (SIZs), where sea ice is present during some part of the year, the underlying benthos receives both types of phytodetritus. Nevertheless, the amount, type and timing of sedimenting autotrophic material varies depending on the sea-ice cover, current velocities, water depth, mass of phytodetrital aggregates, and zooplankton grazing rates [10], among other factors. Although phytoplankton represents the major part of total annual primary production in SIZs [11,12,13], high ice algal biomass could also be observed during the winter –spring transition period [14]. Some studies have shown that a significant portion of ice algae production can rapidly descend to the seafloor [15,16,17,18] due to minimal grazing pressure during peak bloom periods [19]. This is further supported from the fact that ice algae aggregates form substantial clusters on the seafloor, typically measuring up to 10 cm in length [18]. This suggests that the incoming organic matter, which can significantly contribute to the total biogenic carbon production in the Arctic, ranging from 1 to 26% [6,11,20,21,22], could be an important contribution of an early-season food source for benthos after the food shortage during the winter [16,18,23,24]. Among all the effects of climate change in the Arctic, the decrease in the extent, thickness, and duration of seasonal sea ice is the most evident and accurately registered [25,26]. With a new climatic regime, less ice cover and more ice-free periods could drastically change the timing, quality, and amount of phytodetritus released to bottom sediments. Phytoplankton production could increase during longer ice-free periods [27,28], a response already observed in some Arctic Sea areas [29]. Consequently, the biomass of phytoplankton detritus could become higher than that of ice algae due to the loss of sea ice. This could provide less nutritious organic matter fluxes to the seafloor since phytoplankton species are considered a low-quality food item compared to ice algae, which generally have a higher relative polyunsaturated fatty acid (PUFA) content (36.9–43.3% vs. 22.6% [30]; 17–18% vs. 5% [31]; 9% vs. 1% [32]). PUFAs have important roles in the reproduction and growth of marine fauna [33,34,35], and many benthic organisms must obtain PUFAs from their food. Therefore, the loss of ice algae could impact Arctic food webs [6,36], as sea ice algal-derived lipids are important for zooplankton [37] and are also transferred to higher trophic levels [38].

Stable isotope analysis has shown that benthic macroinfauna assimilate ice algae and phytoplankton [36,39,40,41]. However, the stable isotope compositions of ice algae and phytoplankton could be similar and vary among regions, making them indistinguishable in many cases [19,42]. It is, therefore, not always possible to discriminate both carbon sources from the consumer’s tissues, e.g., [43]. Over the last years, experimental approaches have been developed using ¹³C-enriched algae to trace algal assimilation by benthic organisms [44,45,46,47]. This labeled algae can be detected in the consumers, allowing estimations of total uptake by the entire community. The first feeding experiments with shallow benthic communities in the Arctic were made by McMahon et al. [36] in the Svalbard Archipelago to investigate the consumption of fresh-frozen ice algae, ¹³C-enriched ice algae, and fresh-frozen phytoplankton by the whole community or by individual macrobenthic organisms. Sun et al. [32] utilized similar techniques with two common benthic species of the Arctic (Macoma balthica and Monoporeia affinis) collected from the Kotzebue Sound Estuary, Chukchi Sea (2–18 m depth). These authors also performed onboard experiments with communities collected from a ~200 m depth in northwest Svalbard and the western Barents Sea, adding natural phytoplankton and ¹³C-labeled ice algae (see [31]). Mäkelä et al. [48] presented the first results (North Water and Lancaster Sound Polynyas) of a series of pulse-chase experiments performed along the Canadian Arctic, using dual-labeled ice algae and phytoplankton to compare uptakes of deep-sea macroinfauna. In summary, some of these studies found a preference for ice algae among benthic species [32,36], while others did not [31,48], indicating variability in organisms’ responses to available food sources. Our study provides additional results from three pulse-chase experiments conducted with different benthic communities, in terms of species assemblage, abundance, and biomass, from Baffin Bay, Amundsen Gulf, and the Beaufort Sea regions, considered as less productive areas compared to North Water and Lancaster Sound Polynyas [49]. The general objective was to utilize dual-labeled (¹³C and ¹⁵N) diatoms Thalassiosira nordenskioeldii (phytoplankton treatment) and Synedra hyperborea (ice algae treatment) to simulate an external food pulse and measure the macroinfauna responses. This method is considered a powerful tool for comparing the uptake of ice algae and phytoplankton by macroinfauna across various regions of the Canadian Arctic. Additionally, it allows us to assess how community characteristics reflect the quantity and quality of organic matter in these different areas. We hypothesized that ice algae uptake would be considerably higher than phytoplankton uptake by macroinfauna in all sites, as organisms would prefer a better-quality food source such as ice algae. We also expected that macroinfauna would have lower responses and uptakes in the less productive areas (Beaufort Sea station) compared to highly productive regions.

2. Materials and Methods

2.1. Field Sampling

Onboard pulse-chase experiments were carried out using sediment push cores collected from boxcore casts at three stations (Stn 435, Stn 407, and Stn 177) during the ArcticNet 2015 cruise aboard the research icebreaker CCGS Amundsen (Figure 1). Stn 435, hereafter referred to as the Beaufort Sea station, was located on the slope of the Mackenzie Shelf, where seasonal sea ice typically begins to form in October, reaches maximum thickness around March–April, and generally is melted by mid-September [50,51]. Primary production in the Beaufort Sea presented low values compared to other Arctic shelves, ranging from 30 to 70 g C m⁻² yr⁻¹ [52,53], with annual particulate organic carbon (POC) fluxes estimated between 1 and 1.7 g C m⁻² y⁻¹ [50]. Stn 407, referred to hereafter as the Amundsen Gulf station, is situated within the boundaries of the Cape Bathurst polynya. This polynya, which spans the western part of the Amundsen Gulf and the eastern Beaufort Sea Shelf (Mackenzie Shelf), has moderate primary production, ranging from 90 to 175 g C m⁻² yr⁻¹ (compared to the North Water polynyas, which range from 76 to 254 g C m⁻² yr⁻¹). Annual particulate organic carbon (POC) fluxes at a 200 m depth in the Cape Bathurst Polynya have been estimated at 2.4 g C m⁻² yr⁻¹. Typically, this polynya undergoes a rapid and sustained expansion beginning in June, eventually covering an open water area of at least 6000 km², which generally persists for approximately four months [54]. On the other side, Stn 177 located near Broughton Island along the southeast coast of Baffin Island, referred to hereafter as the Baffin Bay station, is within an area where annual primary production ranges from 60 to 120 g C m⁻² yr⁻¹ [52] and is mostly covered by sea ice from October to July [55]. For a nearby station, the POC was estimated at 3.7 ± 0.2 μM C L⁻¹ and presented a low value compared to the other areas sampled in Baffin Bay by Fox and Walker [56]. All sampling stations were selected because of the presence of seasonal sea ice, where benthic communities can receive ice algae and phytoplankton food sources.

Near-bottom dissolved oxygen (O₂), temperature (T), and salinity were obtained from the ship’s CTD profiler (

Table 1. Summary of hydrographic and sediment characteristics for each station.

	Beaufort Sea	Amundsen Gulf	Baffin Bay
ArcticNet 2015 station number	435	407	177
Latitude	71°04.74′ N	70°59.62′ N	67°28.430′ N
Longitude	133°37.96′ W	126°03.39′ W	63°41.526′ W
Date sampled	27 August	23 August	25 October
Depth (m)	300	382	376
Bottom T (°C)	0.49	0.37	0.78
Bottom dissolved O2 (mL/L)	6.3	6.4	5.2
Bottom salinity (psu)	34.08	34.83	34.11
MAX Chl a in column water (mg/m3)	0.7	3.34	0.53
Chl a in sediments (mg/m3)	0.96	0.27	0.48
Sediment OM content (% DW)	8.53	10.60	5.57
Surface sediment δ13C (n = 3)	−25.20 ± 0.09	−23.78 ± 0.20	−22.20 ± 0.28
Surface sediment δ15N (n = 3)	6.17 ± 1.39	7.09. ± 0.49	7.88. ± 1.19
C:N (w/w, n = 3)	7.56 ± 0.12	6.96 ± 0.08	7.03 ± 0.84
Median grain size (µm)	15.488	7.299	10.417

).

2.2. Culture of Labeled Phytodetritus

Axenic clones of ice algal species Synedra hyperborea (CCMP 1422, Bigelow Marine Laboratories) and phytoplankton species Thalassiosira nordenskioeldii (CCMP 995, Bigelow Marine Laboratories) were cultured in a laboratory at 0 °C (light:dark = 12:12 h) in F/2 artificial seawater medium [57] amended with 50% ¹³C-bicarbonate and 50% ¹⁵N-nitrate over 21 days. Algae were harvested by centrifugation, freeze-dried, and stored at −80 °C to produce phytodetritus. The produced algae consisted of 22.3% ¹³C and 3.0% ¹⁵N (C:N ratio of 3.95 w:w) for Synedra hyperborea and 21.6% ¹³C and 4.5% ¹⁵N (C:N ratio of 3.52 w:w) for Thalassiosira nordenskioeldii. These algal species were selected as a food source because they occur in all the study sites (Northern Baffin Bay [58,59,60]; Amundsen Gulf [61], and Beaufort Sea [62,63]). Synedropsis hyperborea was also reported as an epiphyte on Melosira arctica, a centric diatom that dominates the sub-ice community [64,65].

2.3. Incubations Set-Up

Identical experimental protocols were performed for each station following Mäkelä et al. [48]. Undisturbed cores (9.4 to 10 cm internal diameter) with about 20 cm of sediment were transported into a dark and controlled-temperature room at 3 °C, gently filled with ambient bottom water, sealed off with lids containing a magnetic stirrer, and allowed to acclimatize for five hours before starting the experiment. Cores were randomly distributed into three treatments with five replicates each: (1) ¹³C-¹⁵N enriched ice algae detritus addition (freeze-dried Synedra hyperborea), (2) ¹³C-¹⁵N enriched phytoplankton detritus addition (freeze-dried Thalassiosira nordenskioeldii), and (3) controls without food addition. To simulate a phytodetritus pulse, each treatment core received 425 mg C m⁻² in the Beaufort Sea, 545 mg C m⁻² in the Baffin Bay, and 600 mg C m⁻² in the Amundsen Gulf stations. These concentrations were determined based on the POC flux estimations in each area, corresponding to the published data available during the incubations. They accounted for approximately 25% of the total annual POC flux at a depth of 200 m [66]. In the case of Baffin Bay station, where POC estimations were unavailable, an intermediate quantity was used. The amount of algal N added to treatment cores was calculated using the C:N ratios. For ice algae cores, these amounts were 57 (Beaufort Sea station), 73 (Baffin Bay station), and 81 (Amundsen Gulf station) mg N m⁻², while in phytoplankton cores, these amounts were 88 (Beaufort Sea station), 113 (Baffin Bay station), and 125 (Amundsen Gulf station) mg N m⁻². After adding food, all cores were incubated for four days in dark conditions.

2.4. Sample Processing

The overlying water was removed at the end of the 4-day incubation period. The first top 10 cm of sediment in each core was extracted and was horizontally sectioned by intervals of 0–5 cm and 5–10 cm. Each layer was divided in two; one half was kept for phospholipid fatty acid (PLFA) analyses, while the other half was used for macroinfauna extraction using a 500 µm mesh sieve. Samples were preserved in buffered 4% formaldehyde–seawater solution. Organisms were rinsed in ultra-pure water, counted, and categorized into taxonomic groups (mostly family level) under a dissecting microscope. Each core (n = 15 per station) was used as a replicate for abundance and biomass of macroinvertebrates. Determination of the feeding guilds of all identified taxa was accomplished using the available bibliography [67,68,69], and five functional groups were determined: facultative filter/surface- deposit feeder (FF/SDF), obligate filter feeder (FF), predator/scavenger (P/S), subsurface deposit feeder (SSDF), and obligate surface deposit feeder (SDF). To avoid contamination, different picking instruments were used to work with labeled and unlabeled sample cores. Identified specimens were oven-dried overnight at 60 °C, encapsulated in pre-weighed tin cups, and stored in a desiccator until isotopic analyses. Bivalves were decalcified by adding drops of 1 M HCl until bubbling ceased [70,71] and then dried as mentioned above without rinsing. Several organisms of each taxonomic group from Baffin Bay station were encapsulated together, while specimens from Beaufort Sea and Amundsen Gulf stations were individually encapsulated for isotopic analysis. In some cases, same species individuals found in different cores were pooled together to obtain sufficient sample mass for isotope analysis. In these cases, the total uptake was divided by the number of pooled cores. Due to the small size of organisms, gut contents were not emptied, so the isotopic values represent assimilation and ingestion, which is called uptake in this work. Samples for sediment δ ¹³C/C¹² and ¹⁵N/N¹⁴ analysis were collected from three control cores and stored at –80 °C before processing. Sediments were acidified two times with 1 M HCl drying between acid additions. All samples were simultaneously analyzed for ¹³C/C¹²-¹⁵N/N¹⁴ at the UC Davis Stable Isotope Facility using a PDZ Europa ANCA-GSL elemental analyzer (for macroinfauna samples) and a Micro Cube elemental analyzer (Elementar Analysensysteme GmbH, Langenselbold, Germany) (for sediment samples), both interfaced to a PDZ Europa 20–20 isotope ratio mass spectrometer (Sercon Ltd., Cheshire, UK). The mass spectrometer was calibrated against several replicates of four or more National Institute of Standards and Technology (NIST) standards (bovine liver, glutamic acid, enriched Alanine, IsoLife Enriched Maize, Nylon 5, Nylon 6, Glutamic Acid (GLU), REF5 (KN5)).

2.5. Calculations of Carbon and Nitrogen Uptake

Isotope ratios are expressed in the delta notation (δ ‰) as Equation (1).

$$\delta \chi (‰)=\left[\right(R_{\mathit{sample}}-R_{\mathit{reference}})-1]\times 1000$$

(1)

where X is ¹³C or ¹⁵N, R_sample is the ¹³C/¹²C or ¹⁵N/¹⁴N of the sample, and R_reference is the ¹³C/¹²C of the Vienna Pee Dee Belemnite standard (RVPDB = 0.0112372) or atmospheric air isotopic ratio (R_atmN = 0.0036765) for ¹⁵N/¹⁴N. The total amount of C or N of algae uptaken (I algae) by macroinfauna taxa was calculated by multiplying the excess ¹³C or ¹⁵N (above natural isotope signature of taxa) by specimen C or N content and then adjusted to account for phytodetritus labelling, as shown in Equation (2).

$$I_{\mathit{algae}}={\displaystyle \frac{E\times \left(\mathit{Specimens\; C\; or\; N\; content}\right)}{\mathit{At\; \%\; phytodetritus}/100}}$$

(2)

where e xcess (E) is the difference in the fraction (F_sample) and the background (F_background), F is equal to R/(R + 1), and R = (δ¹³C/1000 + 1) × (RVPDB = 0.0112372) or R = (δ¹⁵N/1000 + 1) × (R_atmN = 0.0036765). At % phytodetritus represents the percentage of ¹³C or ¹⁵N atoms of the total C or N present in the added labeled algae, which was 53.8 atom % ¹³C and 45.9 atom % ¹⁵N for Thalassiosira nordenskioeldii and 60.1 atom % ¹³C and 36.5 atom % ¹⁵N for Synedra hyperborea. F_background values were calculated using macroinfauna in control cores. A s pecimen’s C or N content expressed as a unit of mass was calculated using the output of the isotopic ratio mass spectrometry (IRMS) and the dry mass in mg. The total uptake (I_algae) by core was measured by adding up all uptakes of organisms present in each core. To get an indicator of uptake efficiency, the biomass-specific uptake was calculated by dividing the (I_algae) by the total sample weight of the different families [48].

2.6. Statistical Analyses

Since the amounts of C and N added in the cores differed in each experiment, the stations were considered independent samples. Statistical comparisons among treatments for the total uptake and biomass-specific uptake were made using t-test analysis (Welch approximation used). The Shapiro –Wilk test was used to test normality, and the homogeneity of variance assumption was visually examined using residuals vs. fitted (predicted) values. Data were log-transformed when necessary. When assumptions were not met, the Mann –Whitney U-test (nonparametric test) was used. To assess differences between stations, the multivariate community data at the family level (Bray –Curtis similarity matrix created from the fourth-root transformed matrix) were analyzed using permutational multivariate analysis of variance (PERMANOVA). For univariate analysis of abundance and biomass among stations, a Euclidean distance matrix was used to calculate resemblance before performing PERMANOVA. For both multivariate and univariate analysis, ADONIS (Permutational Multivariate Analysis of Variance Using Distance Matrices) and post hoc-ADONIS (pairwise-ADONIS [72]) included in the vegan package of R [73] were used. The significance level was set at p ≤ 0.05 for most of the statistical tests. However, due to the high variability in the uptake rates, the significance for these cases was reported at p ≤ 0.10; in these cases, the exact p-value is explicitly stated in the text. All statistical analyses and plots were carried out using R [74].

3. Results

3.1. Macroinfaunal Community Composition

Benthic community assemblages were significantly different among stations (PERMANOVA, p(perm) < 0.01,

Table 2. PERMANOVA results for community assemblages.

Pairs	F.Model	R2	p.Value	p.Adjusted
Baffin Bay vs. Amundsen Gulf	34.621470	0.5528690	0.001	0.003
Baffin Bay vs. Beaufort Sea	26.253965	0.4839087	0.001	0.003
Amundsen Gulf vs. Beaufort Sea	6.047447	0.1776182	0.001	0.003

). Beaufort station and Baffin Bay station, which exhibited the most distinct assemblages, only shared one family (Chaetodermatidae), whereas Beaufort and Amundsen Gulf stations, the most similar stations, shared eleven families; finally, Amundsen Gulf and Baffin Bay stations shared five families. At the Baffin Bay station, small Thyasiridae bivalves (mostly smaller than 1.5 mm) and Spionidae polychaetes contributed to the highest proportion of the macroinfaunal assemblage. In comparison, large- sized polychaetes and nemerteans (>10 mm) in low abundances dominated at Amundsen Gulf and Beaufort stations, which were more similar among them (Figure 2). The family of bivalves Thyasiridae was the most abundant at Baffin Bay station (46%) but only contributed to 8% of total C biomass (Figure 3). The polychaetes Spionidae were the second most abundant group (14%) and dominated in biomass, representing 28% of total biomass (Figure 3). At this station, large agglutinated foraminifera (>1 mm) comprised 70 and 51% of the total biomass and abundance, respectively. However they were excluded from the abundance and biomass calculations because it was difficult to determine if they were alive during sampling. At the Amundsen Gulf station, the polychaetes Maldane sarsi/arctica dominated with 9% of the total abundance and 19% of the overall biomass (Figure 3). At the Beaufort Sea station, the most abundant family was the Crustacean family Diastylidae (6%); however, all species occurred at low abundances (between 5 and 3% of the total abundance, Figure 3). One large Gnathidae crustacean (4 mm) and six nemerteans accounted for 52% of the total biomass at this station, and Lumbrineridae was the polychaete presenting the highest biomass (8% of total).

Univariate PERMANOVA results indicated that the Baffin Bay station had a significantly higher total C biomass compared to the Beaufort Sea and Amundsen stations (Pseudo-F = 9.3177, p(perm) = 0.001) in the top 0–5 cm of sediment (Figure 4B). Different distribution patterns were also observed in the sediment depth used by organisms. At the Baffin Bay station, 91% of the total biomass was observed in the surface layer (0–5 cm), while in Amundsen Gulf and in Beaufort Sea stations, organisms were more equally distributed in both layers (0–5 and 5–10 cm; Figure 4B). Pair wise ADONIS comparisons for biomasses between layers 0–5 and 5–10 cm showed no significant differences in the Amundsen Gulf station (Pseudo-F = 0.002, p(perm) = 0.867) or the Beaufort Sea station (Pseudo-F = 0.016, p(perm) = 0.711). Different feeding modes dominated in each station. In the Beaufort Sea station, the predators/scavengers (P/S) dominated with 67.2% of the total biomass. In the Amundsen Gulf station, 43.5% of the organisms were SSDF, 27.7% SDF, and 16.1% P/S, and in the Baffin Bay station, 73.9% of the total biomass was represented by FF/SDF organisms.

3.2. Ice Algae and Phytoplankton Total C and N Uptake Rates

Since natural δ¹³C and δ¹⁵N values higher than −17.7‰ and 18.2‰, respectively, were not found, δ¹³C > −15‰ and ¹⁵N > 19‰ values were used to determine which organisms were labeled (evidence of incorporation of ¹³C or ¹⁵N from the labeled phytoplankton or ice algae). Macroinfauna incorporated ¹³C- and ¹⁵N-labeled algae after four days of incubation in all experiments. In Baffin Bay station, 97% of the organisms in the surface layer had enriched δ¹³C and δ¹⁵N values, while in the Amundsen Gulf and Beaufort Sea stations, it was only 52% and 35%, respectively. In some cases (one at Baffin Bay station, ten at Amundsen Gulf station and eight at Beaufort Sea station), organisms were only ¹⁵N labeled. Only 20% of the organisms presented evidence of labeled algae incorporation for all stations in the deepest layer of sediment (5–10 cm).

Total C and N uptakes by macroinfauna in ice algae and phytoplankton treatments for the three stations are represented in Figure 5A,B. Carbon and nitrogen total uptakes were higher in phytoplankton cores for all stations. However, a statistically significant difference was only found in Baffin Bay station (

Table 3. Statistical analysis comparing ice algal and phytoplankton treatments for the total C and N uptake of the whole community, polychaetes, and bivalves. p ≤ 0.10 significance.

	Total C Uptake			Total N Uptake
Beaufort Sea Station	t	df	p	t	df	p
Whole community	−0.269	5.640	0.798	−0.446	6.348	0.671
Polychaetes	0.683	3.784	0.534 a	0.375	3.776	0.728 a
Bivalves	−3.038	2.009	0.092	−2.932	2.01	0.099
Amundsen Gulf station	t	df	p	t	df	p
Whole community	0.103	4.889	0.922 a	0.042	5.186	0.968 a
Polychaetes	0.233	5.261	0.825 a	0.151	5.353	0.885 a
Bivalves	U = 14		0.110	U = 15		0.059
Baffin Bay station	t	df	p	t	df	p
Whole community	−2.020	5.697	0.092	−1.913	6.23	0.100
Polychaetes	−2.392	5.658	0.056	−2.11	6.127	0.078
Bivalves	0.248	5.122	0.814	−0.425	5.736	0.686

a = log transformation, U = Mann –Whitney U-test.

). The macroinfaunal community in Baffin Bay station took up 0.23 ± 0.04 and 0.31 ± 0.06% of the total C and N added in ice algae cores, and in phytoplankton cores the C and N uptakes represented 0.42 ± 0.08 and 0.35 ± 0.07% of the total added, respectively. At the Amundsen Gulf station, macroinfauna took up to 0.09 ± 0.03 of the total C and 0.13 ± 0.05% of the total N added in ice algae cores, and 0.34 ± 0.20 and 0.33 ± 0.19% of the total C and N added in phytoplankton cores. Two large individual Amage sp. Polychaetes (Ampharetidae) present in two phytoplankton cores at this station took up ~4 mg C m⁻² and ~2.5 mg N m⁻², respectively. They were the individuals that presented the highest uptakes in all our experiments. These polychaetes were responsible for the high variability in the total uptake of phytoplankton. Without considering them, the average uptake was 0.32 ± 0.20 mg C m⁻² and 0.06 ± 0.04 mg N m⁻². Finally, macroinfauna uptook 0.07 ± 0.04 and 0.11 ± 0.06% of the total C and N, respectively, added to ice algae cores and 0.08 ± 0.03 (C) and 0.10 ± 0.03% (N) in phytoplankton cores for the station in the Beaufort Sea.

At all stations, polychaetes and bivalves were responsible for the major part of the C and N uptake (Figure 6A,B). At Baffin Bay station, 70% of the phytoplankton- derived C and N was taken up by polychaetes belonging to Spionidae and 15% by Lumbrineridae families. Also, in ice algae cores, Spionidae were responsible for the major fraction (57%) of the total C and N uptake, and bivalves from the family Yoldiidae were second in order with 16%. All polychaetes registered a more significant C and N total phytoplankton uptake than ice algae at the Baffin Bay station (Table 3). The total phytoplankton C and N uptake at the Amundsen Gulf station was dominated by two large individual Ampharetidae polychaetes (Amage sp.), which picked up 76%, followed by Spionidae with 17%. For ice cores at the same station, the contributions to total C and N uptakes were 35% for Maldanidae, 22% for Spionidae, 22% for Ampharetidae, and 14.7% for Yoldiidae. All bivalves together showed a significantly higher total N uptake of ice algae than phytoplankton at this station (p = 0.059, Table 3). However, ice algae core average biomass values were considerably higher 167.42 vs. 43.93 mg dry weight (DW) m⁻² (but not significantly different, t = 1.425; df = 6.381; p ≥ 0.05). The lowest total uptakes were observed at the Beaufort Sea station, where Spionidae was almost the unique family that took up C and N in the ice algae cores with 94% of the total uptake followed by 3% for Mytilidae bivalves. The total phytoplankton-derived C and N uptake was produced by Yoldiidae (41%), Mytilidae (23%), and the soft coral Alcyoniidae (17%). The bivalves presented a significantly higher C and N uptake of phytoplankton at the Beaufort Sea station (p = 0.092, Table 3), although biomass was significantly higher in this treatment (t = −2.725; df = 3.822; p ≤ 0.05), which could influence the uptake results.

3.3. Biomass-Specific Uptake of Ice Algae and Phytoplankton C and N

Biomass-specific uptake of phytoplankton- derived C and N was on average greater than the ice algae counterpart at all stations (Figure 5C,D). This difference was only significantly higher at the Baffin Bay station (

Table 4. Statistical analysis comparison between ice algae and phytoplankton treatments for the biomass-specific C and N uptake of the whole community, polychaetes, and bivalves. p ≤ 0.10 significance.

	Biomass-Specific C Uptake			Biomass-Specific N Uptake
Beaufort Sea station	t	df	p	t	df	p
Whole community	0.372	5.603	0.724	−0.357	6.906	0.732
Polychaetes	0.451	4.459	0.673 a	0.582	4.762	0.587 a
Bivalves	−1.761	3.320	0.168	−1.744	3.111	0.176
Amundsen Gulf station	t	df	p	t	df	p
Whole community	−0.003	7.242	0.997 a	U = 14		0.841
Polychaetes	0.087	5.679	0.933 a	0.112	5.787	0.914 a
Bivalves	0.636	3.144	0.568	0.712	3.078	0.526
Baffin Bay station	t	df	p	t	df	p
Whole community	−3.457	5.241	0.017	−4.719	7.401	0.002
Polychaetes	−2.191	5.881	0.071 a	−2.415	4.391	0.067
Bivalves	−2.3924	7.343	0.046	−2.330	6.812	0.054

a = log transformation, U = Mann –Whitney U-test.

). The highest average values of biomass-specific uptake were found at the Amundsen Gulf station (2.42 ± 1.14 µg C mg⁻¹ and 2.46 ± 1.28 µg N mg⁻¹ for ice algae and 4.16 ± 2.39 µg C mg⁻¹ and 3.99 ± 2.21 µg N mg⁻¹ for phytoplankton). Spionidae polychaetes and Yoldiidae bivalves were amongst the top three dominant taxa regarding biomass-specific uptakes at all stations. Only Spionidae at Beaufort Sea station and Yoldiidae at Amundsen Gulf station showed higher biomass-specific uptakes in ice algae cores vs phytoplankton; however, in most cases, stable isotope value measurements were only available for a few samples due to the low biomass of the organisms, which prevented any statistical comparison. The highest C and N biomass-specific uptakes were 20.34 ± 2.71 µg C mg⁻¹ and 20.65 ± 12.09 µg N mg⁻¹, registered by the Spionidae and Ampharetidae polychaetes, respectively, in phytoplankton treatment at Amundsen Gulf station. Finally, polychaetes and bivalves at Baffin Bay station had significantly higher biomass-specific C and N uptake in phytoplankton cores (Table 4).

4. Discussion

Our findings underscore the adaptability and dietary plasticity of benthic communities in response to a food pulse, which is important for predicting their resilience to changing primary production and organic matter fluxes in a warming Arctic. These results suggest that benthic communities can withstand a decrease in ice algae abundance and rely on phytoplankton in future scenarios. Contrary to our initial hypothesis and previous studies suggesting a preference for ice algae [32,36], we found that macroinfaunal communities consume both ice algae and phytoplankton, with phytoplankton being consumed more extensively in some cases. The differences in consumption levels across sites were linked to the distinct benthic communities of different sampling sites. Baffin Bay, which potentially receives the highest quantity of organic matter, supported a community dominated by facultative filter feeders and surface- deposit feeders. This contrasted with the Amundsen Gulf and Beaufort Sea regions, where food scarcity or episodic supply forced organisms to migrate deeper into the sediment. Additionally, the high abundance of Spionidae polychaetes in Baffin Bay aligned with regions experiencing a significant organic matter enrichment.

However, it is worth acknowledging the limitations of our study. First, we relied exclusively on axenic cultures of Synedropsis hyperborea and Thalassiosira nordenskioeldii to represent ice algal and phytoplankton communities, making it difficult to assess to which extent the observed patterns are influenced by the specific algal species, their degradation state, or other factors that could potentially impact the carbon cycling rates. Future research should, therefore, explore how variations in the composition and state of ice algal and phytoplankton species affect the carbon cycling rates reported here. Due to sampling methods and the presence of large-size foraminifera, whose viability at the time of sampling was uncertain, our study focused on the macroinfaunal community feeding ecology, excluding mobile epifauna. Consequently, the C and N uptake measured may have not shown the full picture for both infauna and epifauna communities, particularly in Baffin Bay, where high abundance and biomasses of large foraminifera were observed. Foraminifera are known to consume phytodetritus rapidly [75] and could have significantly contributed to the C and N uptake in this area.

Working with ex situ incubations implies that the sediment communities’ functioning may have been affected from decompression and warming effects during extraction. Yet, decompression effects in sediments above a 1000 m depth are not pronounced [76], and none of our cores were sampled deeper than a 400 m depth, minimizing these effects. The incubation temperature (3 °C) was slightly higher than the ambient temperature at the seafloor (Table 1), which may have influenced the processing rates [23,77]. Nevertheless, since this temperature was uniform across all incubation cores, it did not affect the experimental design or compromise the validity of the hypotheses tested. While in situ labeling experiments are more precise, the artifacts and high logistical costs in Arctic studies make using in situ benthic landers challenging. Finally, the macroinfaunal uptake measured in our experiments likely reflects more the faunal gut content’s C and N isotopic values than those in their tissues. The short duration of our experiments may have not allowed the complete incorporation of C and N into the organism’s tissues.

4.1. Macroinfaunal Community

Macroinfauna abundance and biomass values recorded in this study are consistent with other studies on the same regions (

Table 5. Macroinfauna abundance, biomass, and C uptake of ice algae (IA) and phytoplankton (PP) ± (SE) at five different stations in the Canadian Arctic using the same sampling protocol.

REGION (Station)	Abundance (Ind. m−2)	Biomass (mg C m−2)	Depth (m)	C Added (mg C m−2)	IA C Uptake (mg C m−2)	% of the Total C Added	PP C Uptake (mg C m−2)	% of the Total C Added
North Baffin Bay (Stn 124) a	10,952	3190 ± 432	709	1475	7.9 ± 1.8	0.5 ± 0.1	4.3 ± 0.9	0.3 ± 0.1
Lancaster Sound (Stn 323) a	8355	2110 ± 345	794	600	3.1 ± 1.0	0.5 ± 0.2	3.3 ± 0.4	0.6 ± 0.1
Baffin Bay (Stn 177) b	14,000 ± 795	2208 ± 549	376	545	1.25 ± 0.21	0.2 ± 0.03	2.27 ± 0.45	0.4 ± 0.08
Baffin Bay c	1482		251/500	-	-	-	-	-
Amundsen Gulf (Stn 407) b	2644 ± 191	1890 ± 1087	382	600	0.53 ± 0.16	0.1 ± 0.03	2.02 ± 1.17	0.3 ± 0.2
Amundsen Gulf d,e	2578 ± 730	100 to 49,000	259	-	-	-	-	-
Beaufort Sea (Stn 435) b	2444 ± 297	1234 ± 1297	300	425	0.29 ± 0.16	0.1 ± 0.04	0.34 ± 0.11	0.1 ± 0.03
Beaufort Sea d,e	828 to 2041	100 to 49,000	70/440	-	-	-	-	-

a = Mäkelä et al. [48], b = This study, c = Thomson [100], d = Conlan et al. [80], e = Conlan et al. [101].

), despite the different methodologies and sampling efforts. Macroinfaunal assemblage was different among the studied stations (Table 2). Due to the distances among the sampling locations and biogeographic information [78,79], this was expected for Baffin Bay station vs. Amundsen Gulf/Beaufort Sea stations (~2500 km separation). However, the Amundsen Gulf and Beaufort Sea stations (~300 km separation) were expected to be more similar considering that other studies did not find differences between the community structure of Amundsen Gulf and the Beaufort slope at similar depths [80]. The most remarkable difference was the occurrence of abundant species belonging to Thyasiridae, Spionidae, and Yoldiidae families with low individual biomass at Baffin Bay station. This contrasts with the low abundance of large deep burrowing Maldanidae polychaetes (Maldanidae sarsi/artica) at Amundsen Gulf station and Crustacea and Nemertea at Beaufort Sea station. Although many factors affect benthic abundance, biomass and structure (e.g., water depth, grain size, current velocity, oxygen concentration, sedimentation rate, competition, predation, etc.), some studies have concluded that food supply is among the most important factors [81,82,83]. Pearson and Rosenberg [84] used a succession model and predicted that sites with high OM concentrations lead to increased abundances, dominance of a few species (<diversity) and low individual biomass per benthic macrofaunal species. Ruhl et al. [85] also found that increases in POC flux could decrease equitability (communities where all species have similar abundances). Thus, it could be speculated that the macroinfauna community found at Baffin Bay station (see Figure 3) could be structured by a higher quality or amount of supplied OM than the Amundsen Gulf and Beaufort Sea stations. The dominance of FF/SDF organisms in Baffin Bay station, which could selectively feed and take advantage of high-quality food that settles to the seafloor, may also be a consequence of the greater amounts of OM that arrives at this station. High macroinfaunal abundances in the uppermost sediment layers have also been suggested to indicate a large quantity of OM transported by currents [86,87]. In contrast to the Baffin Bay station, where almost all the organisms were on the first 5 cm of sediment, in the Amundsen Gulf and Beaufort Sea stations, the biomass was distributed along both layers of sediment (0–5 and 5–10 cm). This could indicate sites with scarce or episodic food supply (oligotrophic sediments), as it has been proposed that organisms may pull OM to deeper sediment layers to avoid competition [88,89]. The high abundances of Spionidae polychaetes at Baffin Bay station may also indicate OM enrichment, as this taxon was dominant in deep-sea regions subjected to high quantities of OM inputs [90,91]. Finally, the abundant large- size agglutinated foraminifera (>5000 ind. m⁻²) in the epifaunal habitat of Baffin Bay station compared to very low abundance at the other two stations may be indicative of higher OM inputs at Baffin Bay station, as several studies demonstrated coupling between phytodetritus and foraminifera abundances [10,92,93,94,95,96,97]. Initially, for the planning of the experiments, Amundsen Gulf station inside of Cape Bathurst Polynya was considered as the station with the highest POC fluxes at the 200 m depth (2.08–12.8 g C m⁻² y⁻¹; [50]), which is close to the POC fluxes registered in the North Water Polynya (1.0~13.8 g C m⁻² y⁻¹; [98]). However, Forest et al. [12] concluded that the planktonic community could retain ~97% of the primary-produced C in the water column, which is the reason why the sedimentation and benthic C demand at a 400 m depth in the central Amundsen Gulf (near our station) are low, and benthic community composition seems not to be altered by the polynya [80]. Baffin Bay station, where POC fluxes estimations were not available in the area at the moment of the experiment, was considered as an intermediate station concerning POC fluxes, but this station actually could be the one that receives the highest quantity and/or quality of OM [56]. Another explanation for these contrasting macroinfaunal assemblages may be related to the sedimentation and deposition of mineral material coming from glacial bays near Baffin Bay station. Wlodarska-Kowalczuk and Pearson [99] found that glacial bays had more abundant Yoldiid and Thyasirid bivalves. In contrast, tube-building polychaetes replaced these bivalves in the central basin with lower inorganic sedimentation.

4.2. Ice Algae and Phytoplankton C and N Uptakes

Our results showed that after four days of sediment incubations, macroinfaunal organisms had uptaken both ice algae and phytoplankton at all the stations. This is consistent with other studies that found that arctic benthic organisms can rapidly respond to phytoplankton and ice algae inputs [31,32,36,48,102]. However, the low incorporation rates observed—less than 1% of the added carbon and nitrogen—may be attributed to the fact that the majority of the ingested material is typically directed towards respiration [103,104,105] or is consumed by bacteria [32,106,107,108]. Furthermore, responses were not the same at each station: at the Baffin Bay station, the macroinfauna community obtained the highest total uptakes, phytoplankton detritus consumption was significantly higher, and almost all the organisms in treatment cores ingested the added food. The higher total uptakes found at the Baffin Bay station compared to the Amundsen Gulf and Beaufort Sea stations could be associated with the higher biomass and abundance in the surface layer (0–5 cm). This observation agrees with previous research showing that consumption of OM is related to consumer biomass, e.g., [104,109]. In addition, the different community structures found in each station may explain the different total uptakes and percentages of labeled organisms [44,104,110,111]. The high abundance of FF/SDF organisms at Baffin Bay station, which are normally the first to access and process the deposited OM [44,112,113], contrasted with the dominance of P/S at Beaufort Sea station and SSDF at Amundsen Gulf station. The presence of these organisms, which could need more than 4 days to access labeled food, may explain the different percentages of organisms that consumed phytodetritus. Total uptake is strongly correlated with the biomass in each station, a pattern also shown in most of the feeding experiments that were developed using the same type of labeled algae in the Canadian Arctic (Table 5). In addition, the comparison of Lancaster Sound station (Stn 323, Mäkelä et al. [48]) and Amundsen Gulf station (Stn 407), even though there are different depths, could reflect the importance of the community structure in the C and N uptake. These stations received the same amount of C and presented similar biomass; however, the uptake in Stn 323 was higher (Table 5). This could be associated with the dominance of FF/SDF in Lancaster Sound station vs. the dominance of SSDF in Amundsen Gulf station.

Biomass-specific uptake was used to evaluate our original hypothesis that ice algae were consumed in a larger fraction than phytoplankton. Contrary to the total uptake, this measure allows us to interpret results without being influenced by the different community core biomass. Our results showed that both ice algae and phytoplankton were consumed equally by the whole community at two of the three stations. In addition, at the Baffin Bay station, the C and N biomass-specific uptake of phytoplankton was significantly greater than that of ice algae, which led to the rejection of our initial hypothesis that ice algae are consumed preferentially. Comparable results were found for a similar community by Mäkelä et al. [48] at a deeper station in the Lancaster Sound polynya, 900 km north of Baffin Bay station, where phytoplankton detritus also presented higher total and biomass-specific uptake (Table 5). Possibly, the organisms in those areas are used to consuming this type of food that was naturally available during the sampling season (summer). This may be reflected in the feeding experiments. In addition, it has been proposed that suspension feeders could be more efficient at consuming suspended phytoplankton than ice algae, which typically forms aggregates [36]. However, it was not possible to show, in our experiments, the exact reason for this preference. Feeding experiments with both labeled phytodetritus in the same treatment are warranted to verify whether there is a preference, e.g., [32].

Among the taxa that presented high biomass-specific C and N uptakes, Ampharetidae and Spionidae polychaetes were the ones with greater values, and this agrees with other studies that classified them as efficient feeders of ¹³C-labeled algae [44,48,109]. Two large individual Ampharetidae polychaetes (Amage sp.) were responsible for 76% of total phytoplankton C and N uptake. In contrast, two small Prionospio sp. uptake 17% of the total phytoplankton derived C and N for the same treatment in Amundsen Gulf station. However, their biomass-specific uptakes were similar, demonstrating that size does not affect their efficiency in algae uptake. Levin et al. [113] also found that small and large infauna have similar access to phytodetritus. Prionospio sp. was registered at all stations, although at Baffin Bay station it was six times more abundant. Nevertheless, the highest biomass-specific uptakes were found when this species was present in low abundance (Amundsen Gulf station), suggesting that they may become more efficient when they are in small numbers. Prionospio sp. is present throughout the Arctic and has been proposed as an important ecosystem engineer in the processing of OM in deep-sea sediments [48]. In this study, we showed that even if they are in low abundances, they still play an important role in OM processing. Our feeding experiments showed that Spionidae at Baffin Bay station utilized ice algae and phytoplankton equally, which agrees with the results of Mäkelä et al. [48]. Spionidae showed higher biomass-specific uptake of ice algae in the Beaufort Sea station, but we cannot make any conclusions due to the low number of samples (n = 2 for each treatment). Lumbrineridae is the only family defined as carnivores [68] that presented an important biomass-specific C and N uptake; however, it has been proposed that this family can change their feeding mode to deposit feeding depending on the available food [114]. In this study, Lumbrineridae polychaetes at the Baffin Bay station were deposit feeders and became significantly labeled- algae consumers. This is also supported by the low δ¹⁵N values (~12‰) found for these polychaetes in the control cores. This is in accordance with Link et al. [58], who identified the Lumbrineridae species Lumbrineris tetraura as a major organism that explains variation in benthic function.

5. Conclusions

The primary aim of our study was to assess the consumption patterns of ice algae and phytoplankton by macroinfaunal communities across three regions in the Canadian Arctic. Contrary to the null hypothesis, we did not observe a preference for ice algae detritus over phytoplankton among benthic macroinfauna. This underscores the notion that these organisms opportunistically exploit both organic matter (OM) sources, highlighting their vulnerability to fluctuations in OM quantity rather than quality. Consequently, our findings suggest that these communities’ potential resilience to projected reductions in ice algae is high.

Furthermore, our investigation revealed distinct macroinfaunal community structures, particularly at the Baffin Bay station. Here, a dominance of small facultative filter/surface-deposit feeders with high trophic diversity appears closely linked to substantial OM input. In contrast, the Amundsen Gulf station exhibited a community adapted to lower or sporadic food inputs, characterized by shallow-subsurface deposit feeders penetrating deeper to exploit OM resources. This aligns with the theory by Forest et al. [12] suggesting retention of primary production in the water column (due to zooplankton ingestion), reducing OM supply to the seafloor. Similarly, the Beaufort Sea station’s community structure reflected a reliance on refractory OM due to limited OM reaching the seafloor, indicating a dependence on the quantity of OM reaching these environments.

In conclusion, our study underscores the crucial role of OM supply in shaping macroinfaunal community structures in Arctic ecosystems. Any future alterations in primary production and subsequent OM export to the seabed could profoundly impact these communities, highlighting the importance of understanding and mitigating potential consequences of environmental changes in these sensitive regions.