Prominent bacterial heterotrophy and sources of 13 C-depleted fatty acids to the interior Canada Basin

In recent decades, the Canada Basin of the Arctic Ocean has experienced rapidly decreasing summer sea ice coverage and freshening of surface waters. It is unclear how these changes translate to deeper waters, particularly as our baseline understanding of organic carbon cycling in the deep basin is quite limited. In this study, we describe full-depth profiles of the abundance, distribution and carbon isotopic composition of fatty acids from suspended particulate matter at a seasonally ice-free station and a semi-permanently ice-covered station. Fatty acids, along with suspended particulate organic carbon (POC), are more concentrated and 13C-enriched under ice cover than in ice-free waters. But this influence, apparent at 50 m depth, does not propagate downward below 150 m depth, likely due to the weak biological pump in the central Canada Basin. Branched fatty acids haveδ13C values that are similar to suspended POC at all depths and are more 13C-enriched than even-numbered saturated fatty acids at depths above 3000 m. These are likely to be produced in situ by heterotrophic bacteria incorporating organic carbon that is isotopically similar to total suspended POC. Below surface waters, there is also the suggestion of a source of saturated even-numbered fatty acids which could represent contributions from laterally advected organic carbon and/or from chemoautotrophic bacteria. At 3000 m depth and below, a greater relative abundance of long-chain (C20−24), branched and unsaturated fatty acids is consistent with a stronger influence of re-suspended sedimentary organic carbon. At these deep depths, two individual fatty acids (C12 and iso-C17) are significantly depleted in 13C, allowing for the possibility that methane oxidizing bacteria contribute fatty acids, either directly to suspended particulate matter or to shallow sediments that are subsequently mobilized and incorporated into suspended particulate matter within the deep basin.


Introduction
In the past two decades, the Arctic Ocean's Canada Basin has seen both rapidly decreasing summer sea ice coverage (Maslanik et al., 2011;McLaughlin et al., 2011;Stroeve et al., 2007) and freshening of surface waters (Macdonald et al., 2002;McPhee et al., 2009;Yamamoto-Kawai et al., 2009).These changes have been accompanied by a deepening of the chlorophyll maximum depth (Jackson et al., 2010;McLaughlin and Carmack, 2010), and a trend towards smaller phytoplankton cell sizes and increased bacterial abundance in surface waters (Li et al., 2009).Combined with the observed decrease in primary productivity in the increasingly ice-free Canada Basin (Cai et al., 2010;Grebmeier et al., 2010;Lee et al., 2012), these trends will likely result in decreasing export of organic carbon to the deep basin and sediments.
Across ocean basins, sinking particulate organic carbon (POC) flux has been shown to correlate with bacterial abundance and productivity at deeper depths (Nagata et al., 2010;Yokokawa et al., 2013).Prior to the recent decline in summer sea ice, ice-tethered sediment traps in the Canada Basin documented a much smaller POC flux through the upper 200 m than is observed in other oligotrophic regions (Honjo et al., 2010).Despite a vanishingly small, and likely decreasing supply of autochthonous organic carbon exported from the sea surface, prokaryotic abundance at mesopelagic and bathypelagic depths in the Canada Basin (> 100 m) were found to be comparable to subtropical and equatorial regions (He et al., 2012;Uchimiya et al., 2013).This imbalance between prokaryotic abundance and sinking POC flux suggests the importance of an alternate carbon and energy source supporting microbial productivity in the deep Canada Basin.Heterotrophic bacterial productivity at mesopelagic depths could be supported by labile organic carbon produced by phytoplankton in the productive Chukchi Sea and laterally supplied as dissolved organic carbon (DOC) at mesopelagic depths (Davis and Benner, 2007;Mathis et al., 2007;Shen et al., 2012;Walsh et al., 1989).Long-distance lateral transport of re-suspended sedimentary particles has also been observed below 100 m depth (Honjo et al., 2010;Hwang et al., 2008;Jackson et al., 2010;O'Brien et al., 2013), delivering associated organic carbon to the interior Canada Basin.In addition to bacterial heterotrophy, significant bathypelagic chemoautotrophy is suggested by the radiocarbon and stable isotopic composition of dissolved inorganic carbon (DIC), sinking POC and suspended POC (Griffith et al., 2012).The dynamics between prokaryotic production and organic carbon cycling appears to be distinct from other oligotrophic ocean basins and the questions of what carbon and energy sources support bacterial production in the dark Canada Basin and how they change with time will be highly relevant to predicting the future of organic carbon sequestration and cycling in this deep Arctic Basin.
To advance our understanding of organic carbon cycling in Canada Basin, we investigated the distribution and isotopic composition of fatty acids from suspended POC collected at two deep-basin stations in 2008, a summer with recordlow sea ice coverage (Maslanik et al., 2011).Isotopic studies of bulk organic carbon pools have addressed the supply and fate of vertically and laterally delivered organic carbon to the interior Canada Basin (Griffith et al., 2012;Honjo et al., 2010;Hwang et al., 2008).The provenance and cycling of organic matter are better understood in the Mackenzie River and Beaufort shelf and slope because DOC and POC analyses have been complemented by isotopic studies of fatty acids and other biomarkers (Drenzek et al., 2007;Goñi et al., 2005;Tolosa et al., 2013).Fatty acids are integral components of bacterial and eukaryotic membranes.They can therefore derive both from microbial cells suspended at depth and from degraded organic carbon associated with sinking particles or re-suspended sediments when recovered from suspended POC.Although multiple sources are represented, isotopic analysis of fatty acids allows for more specific investigations of microbial processes which are difficult to resolve through bulk analyses.Here we present the results of our in- vestigations and discuss the multiple organic carbon sources and bacterial metabolic strategies that could contribute to the observed profiles.
2 Material and methods

Sampling
Suspended particulate matter was collected during the Joint Ocean Ice Study (JOIS) in 2008 from two stations in the Canada Basin of the Arctic Ocean: CB4 (seasonally icefree; 74 • 59.9980 N; 150 • 0.0020 W; 3825 m bottom depth) and CB9 (semi-permanently ice-covered; 77  Griffith et al. (2012).Filters were packaged into envelopes of pre-combusted aluminium foil, frozen at −20 • C, and were subsequently partitioned for radiocarbon analysis (Griffith et al., 2012) and for lipid analysis.Samples destined for lipid extraction were maintained at −20 • C for three years until processing.

Lipid extraction, identification and quantification
Total lipids were extracted from frozen filters in 50 mL of methylene chloride/methanol (9 : 1), assisted by the Microwave Accelerated Reaction System (MARS Xpress, CEM Corp) at 100 • C for 20 min.The filter extraction procedure was repeated with fresh solvents and both extracts were combined, passed through a 0.45 µm PTFE filter and evaporated to dryness under a stream of ultra-high purity N 2 .Combined extracts were saponified in 1 mL of 0.2 M KOH in methanol/water (4 : 1) at 80 • C for 2 h.Neutral lipids were recovered by three extractions with hexane after addition of 10 % NaCl.The pH of the remaining aqueous layer was lowered to ∼ 1 by dropwise addition of 6 N HCl, and the acid fraction was recovered by three additional extractions with hexane/methylene chloride (4 : 1).The distribution of total lipids in 10 % subsamples of neutral and acid fractions were determined by GC-TOF as trimethylsilyl ether derivatives.
The remaining 90 % of acid fractions were transesterified to convert fatty acids into their methyl esters by refluxing in 5 % HCl in MeOH (with known δ 13 C value) at 70 • C for 12 h.After adding MilliQ water to cooled vials, fatty acid methyl esters (FAMEs) were recovered by extracting three times with hexane/methylene chloride (9 : 1).Extracts were dried with sodium sulfate and concentrated under a stream of ultrahigh purity N 2 .FAMEs from trans-esterified acid fractions were identified by comparison of retention times with a 37component FAME mixture (Supelco part number 47885-U) and bacterial FAMEs mixture (Matreya part number 1114), and for a few representative samples, by GC-MS.Quantification of FAMEs was performed by GC-FID with an external calibration curve and assigned a conservative 5 % uncertainty.The contribution of fatty acids from the sorption of organic carbon to the GF/F filter was subsequently corrected for by subtraction of surface-area-normalized abundances recovered from a 142 mm GF/F filter lowered to 3805 m water depth but not pumped through (Table 1).

Isotopic analysis of FAMEs
Compound-specific δ 13 C analysis of FAMEs was performed at the Organic Mass Spectrometry Facility at the Woods Hole Oceanographic Institution (WHOI) on an HP-6890 GC coupled to a Finnigan-MAT DeltaPlus IRMS.FAME samples were injected using a Gerstel CIS-4 programmable temperature vaporizing (PTV) injector, separated on a Varian CP-Sil5CB column, and the GC was interfaced using a Finnigan-MAT GC Combustion Interface III (modified with an integral fused silica combustion system at 860 • C).GC-IRMS measurements were made in triplicate and uncertainty is reported as the larger of the standard deviation of three measurements or analytical uncertainty (0.3 ‰).Overall system accuracy and precision were confirmed to be better than 0.3 ‰ based on a suite of nine externally analyzed standards.Measured δ 13 C values of FAMEs are expressed relative to the PDB standard and corrected for addition of the methyl carbon during trans-esterification by mass balance (Supplement Table 1).Values reported in Table 2 are also corrected for fatty acids contributed by organic carbon adsorption.The adsorption blank was defined by the abundance and isotopic composition of fatty acids from the blank filter (Tables 1 and 2).

Bulk analysis of suspended POC
The concentration and isotopic composition of suspended POC was described by Griffith et al. (2012) and exhibits similar profiles at both stations.At near-surface depths, the semipermanently ice-covered station (CB9) has a higher concentration of suspended POC than the seasonally ice-free station (CB4).Despite similar radiocarbon ages, CB9 is also more 13 C-enriched at 9 and 50 m depth.But at deeper depths, suspended POC concentrations converge on similar, very low values (≤ 0.04 µM) with δ 13 C values of −25 to −23 ‰ (Griffith et al., 2012).The sorption of dissolved organic carbon onto glass fiber filters during their time immersed in seawater was also investigated.A 142 mm GF/F filter lowered to 3805 m water depth, but not pumped through, yielded 3.1 µmol of organic carbon from half of the filter area with a δ 13 C value of −25.1 ‰ (Griffith et al., 2012).The distribution of fatty acids recovered from the other half of the blank filter and their δ 13 C values are described in Tables 1 and 2 and discussed below.

Concentration of fatty acids
Analysis of fatty acids was performed on separate fractions of the same filters used for suspended POC measurements.
As with POC, fatty acid concentrations decrease with depth (Fig. 2a).Below 1500 m, the sill depth of the Canada Basin, fatty acid abundances at CB4 are < 10 ng L −1 .Despite these low abundances, the concentration of fatty acids recovered from the blank filter was less than fatty acids from filtered seawater at all depths (Table 1).
Although both POC and fatty acid concentrations are higher (per liter seawater) under ice cover than in open     water (Fig. 2a), fatty acids account for a greater fraction of suspended particulate organic carbon at station CB4 compared to CB9 at 50 m (Fig. 3a).POC-normalized fatty acids also reveal a subsurface enrichment of saturated, evennumbered fatty acids at 2500 m depth (Fig. 3a).This peak results from disproportionately low suspended POC concentrations combined with fatty acid concentrations on par with the sample from 2000 m (Table 1).At mesopelagic depths, between 150 and 1000 m, a more modest enrichment in POCnormalized bacterial branched fatty acids was revealed at station CB4 (Fig. 3).These subsurface enrichments are superimposed on a general decreasing trend of POC-normalized fatty acids with depth.Similar POC-normalized values were observed in the shallow profile from CB9 (Fig. 3b).In the abyssal Canada Basin, however, very low POC-normalized concentrations were found which did not decrease with depth, suggesting a different composition or source of suspended POC in the deepest 1000 m (Fig. 3a).

Distribution of fatty acids
At 50 m depth, the most abundant lipids at both stations were saturated C 14 and C 16 fatty acids (Fig. 2b).Monounsaturated (MUFAs) and polyunsaturated fatty acids (PUFAs) are only significant contributors to total fatty acids at station CB9 at 50 m depth.Appreciable concentrations of PUFAs were not recovered from either station.Below 50 m, C 16 and C 18 fatty acids make up the majority of total fatty acids, followed by iso-and anteiso-C 15 at both stations.There is a general pattern of decreasing relative C 14 abundance and increasing C 18 abundance with depth, and a proportionally important contribution from C 15 fatty acids in the shallowest 1000 m (Fig. 2a).A similar pattern of relative abundances is observed at station CB9, although with two important exceptions: the C 18 -dominated sample at 150 m where C 18 also drives the larger concentration of total fatty acids found at CB9 compared to the same depth at CB4, and more abundant unsaturated fatty acids at 50 m.At station CB4, Fig. 2b illustrates a distinction between the distribution of fatty acids found above 3000 m, and those found at and below this depth.In the deepest 1000 m, saturated C 16 and C 18 continue to be the most abundant fatty acids although C 14 is relatively more abundant than above.We also observe greater proportions of MUFAs and branched fatty acids (BFAs), as well as longchain C 20−24 fatty acids, supporting a distinct source of suspended POC at these depths.

Isotopic composition of fatty acids
At both stations in the Canada Basin, all individual fatty acids from 50 m depth are depleted in 13 C compared to the same lipids recovered from deeper depths (Fig. 4a and b).Saturated C 14 , C 15 and C 16 fatty acids display the largest δ 13 C deviations (e.g., 6-13 ‰ for C 16 fatty acid), while others are more modestly depleted.A similar pattern can be observed in suspended POC from the near-surface depths compared to below (δ 13 C deviation of 3-6 ‰; (Griffith et al., 2012).In all samples, odd-numbered and branched fatty acids (orange shades in Fig. 4) are generally enriched in 13 C compared to even-numbered fatty acids (blue shades in Fig. 4).The abundance-weighted average δ 13 C values of BFAs are within 1 ‰ of POC above 3000 m depth at station CB4 while the SFAs are depleted in 13 C by 4.3-5.8‰ (Table 2).This pattern is also reflected in an average 5 ‰ enrichment of BFAs compared to saturated fatty acids (SFAs) between 50 and 2500 m. δ 13 C values from station CB9 hint at a similar pattern although sampling limitations resulted in fewer possible comparisons.At 3000 m and below, there is a shift towards 13 C-enriched SFAs such that the δ 13 C values of SFAs more closely match suspended POC than the δ 13 C values of BFAs do (Table 2).The exceptions to this pattern are the C 12 and iso-C 17 fatty acids.C 12 fatty acids are depleted by 7-19 ‰ compared to average SFA values despite being within 1 ‰ higher in the water column.Uncorrected δ 13 C values of C 12 fatty acid are within analytical uncertainty between 3000 and 3750 m (Supplement isotopic variability emerged from mass-balance blank corrections (Fig. 4a).While it is possible that this variability is an artefact of our blank determination method, the C 12 fatty acid is unambiguously 13 C-depleted compared to the rest of the water column in the abyssal Canada Basin.Branched iso-C 17 fatty acids are also very depleted in 13 C at 3000 m and below.Its δ 13 C values are more negative at 3000 and 3500 m than any fatty acid observed higher in the water column (Fig. 4a).

Discussion
The profile of fatty acids at stations CB4 and CB9 show strong similarities with open-ocean settings although concentrations are much smaller.Saturated C 14 and C 16 fatty acids are the most abundant fatty acids in the epipelagic waters of the Canada Basin (Fig. 2b) as well as other coastal and open-ocean regions (Gutiérrez et al., 2012;Hamanaka et al., 2002;Loh et al., 2008;Schultz and Quinn, 1972;Tolosa et al., 2004Tolosa et al., , 2013;;Wakeham, 1995;Wakeham et al., 2010;Xu and Jaffé, 2007).In many shallow locations, unsaturated fatty acids follow C 14 and C 16 in abundance, but these are scarce at 50 m in the Canada Basin as they were in a high Arctic fjord during the summer months (Mayzaud et al., 2013).However, the absence of unsaturated fatty acids could also result from their degradation during sample storage or extraction and may not reflect the true distribution of fatty acids in the Canada Basin.We therefore interpret the absence of unsaturated fatty acids with caution, particularly as PUFAs were detected in suspended POC collected from the moreproductive coastal areas of the Beaufort Sea (Connelly et al., 2012;Tolosa et al., 2013).At deeper depths, the dominance of saturated C 16 and C 18 fatty acids mirrors other oligotrophic water columns (Loh et al., 2008;Wakeham, 1995).
Although we noted a greater concentration of total fatty acids under ice cover than in open water (Fig. 2a), this contrast is small compared to the difference between the Canada Basin and much-higher concentrations observed in low-latitude epipelagic regions (e.g., Hamanaka et al., 2002;Tolosa et al., 2004;Wakeham, 1995;Wakeham et al., 2010).But the overall decreasing trend of POC-normalized fatty acids above 3000 m (Fig. 3), also seen in the oligotrophic central Pacific and Sargasso Sea (Loh et al., 2008;Wakeham, 1995), suggests fatty acids in the Canada Basin behave similarly to low-latitude oligotrophic oceans where they are thought to be among the more labile components of POC.
The concentration and isotopic composition of fatty acids in suspended POC from the Canada Basin reflect vertically stratified organic carbon sources at surface depths, in the interior basin, and in the deepest 1000 m.In the following sections, we discuss the possible origin of fatty acids found on the blank filter as well as the phytoplanktonic, microbial, and advected sources of fatty acids to the stratified layers of the Canada Basin.

Sources of the organic carbon sorption blank
The sorption of DOC onto glass fiber filters is a widely recognized contributor of non-particulate organic carbon to suspended POC samples collected by in situ pumps (Gardner et al., 2003;Moran et al., 1999;Schultz and Quinn, 1973).As pumps are lowered into the water, filters likely accumulate this organic carbon near the surface where the highest concentrations of DOC, suspended POC and bacterial abundance are found.The sampling depth, at which the filter has the longest soaking time, is also a possible contributor to the sorption blank.At station CB4, the radiocarbon content of adsorbed organic carbon measured on a filter that traveled the length of the water column and was held at 3805 m depth while POC from other depths were sampled (−247 ± 9 ‰) is very similar to DOC at 20 m water depth (−234 ± 5 ‰) (Griffith et al., 2012) while the δ 13 C value of adsorbed organic carbon (−25.1 ± 0.1 ‰, Table 2) is intermediate between the δ 13 C values of DOC (−22.1 ± 0.1 ‰) (Griffith et al., 2012) and suspended POC (−29.5 ± 0.1 ‰, Table 2) at the surface.Interestingly, the 14 C and δ 13 C values from the blank filter are also similar to suspended POC at 3000 m and below ( 14 C = −227 ± 40 ‰, δ 13 C = −24.6 ± 0.1 ‰ at 3750 m; Griffith et al., 2012).We calculate the possible contributions of these sources following the dual isotope mass balance approach outlined by Griffith et al. (2012).These calculations indicate 27 % of the sorbed organic carbon blank could derive from surface DOC ( 14 C = −234 ± 5 ‰, δ 13 C = −22.1 ± 0.1 ‰ at 20 m), 35 % surface POC ( 14 C = +8 ± 9 ‰, δ 13 C = −29.5 ± 0.1 ‰ at 50 m), and 38 % deep DOC ( 14 C = −494 ± 2 ‰, δ 13 C = −23.1 ± 0.1 ‰ at 3807 m), assigning the majority of organic matter sorption to surface waters.
Fatty acids found in the adsorption blank do not resemble suspended POC in shallow waters, however, either in their distribution or isotopic composition.C 16 and C 18 fatty acids dominate, as they do in suspended POC from 150 m and below (Fig. 2b).Although fatty acids point toward a deeper source region for adsorbed organic carbon than bulk isotopic considerations do, it is possible that this discrepancy can be explained if adsorbed organic carbon preferentially incorporates attached bacterial biomass rather than total organic carbon in surface waters or if it incorporates bacterial biomass from 3805 m depth with a fatty acid distribution that echoes that found at bacteria-dominated mesoand bathypelagic depths.The weighted average δ 13 C value of fatty acids from the adsorption blank is similar to the δ 13 C value of suspended POC (Table 2), following the pattern observed in BFAs from suspended POC between 150 and 2500 m, consistent with a heterotrophic bacterial source (see Sect. 4.4).The 14 C value of the adsorption blank limits the contribution of surface-ocean bacteria incorporating phytoplankton-derived organic carbon with the 14 C value of DIC (+31 ± 4 ‰ at 20 m; Griffith et al., 2012), but not bacteria incorporating "aged" DOC which would have the same isotopic composition as the DOC sorption component of the blank.Bacterial cells that attach to the blank filter at bathypelagic depths would also contribute an "aged" 14 C value to the adsorption blank.Although the depth at which bacteria attach to glass fiber filters is ambiguous, we can assume that a contribution of fatty acids from non-particulate sources is present at all depths.Taking our seawater-soaked filter to be representative of this blank, we correct for their contributions by subtraction (Table 1) or mass balance (Table 2).

Semi-permanently ice-covered vs. seasonally ice-free surface waters
Previous work indicates that ice cover affects POC flux, bacterial abundance and bacterial productivity in the western Arctic Ocean, but these investigations often combine the influences of ice cover and seasonality (Honjo et al., 2010;Sherr and Sherr, 2003;Sherr et al., 2003) or compare the shallower, nutrient-rich and more ice-free Chukchi Sea with the western Canada Basin (He et al., 2012;Honjo et al., 2010;Moran et al., 2005;Rich et al., 1998).We find supporting evidence that ice cover affects the concentration and composition of epipelagic organic carbon within Canada Basin by comparing stations CB4 and CB9.At the time of our sampling, station CB9, which was ice-covered, hosted a greater concentration of suspended POC at 9 and 50 m depth which was more 13 C-enriched compared to ice-free station CB4 (3-6 ‰; Griffith et al., 2012).This contrast is unlike DIC and DOC which have δ 13 C values within 0.5 ‰ between stations at 20-25 m depth (Griffith et al., 2012).
The concentration and distribution of fatty acids also exhibit differences between the ice-covered and open-water stations and offer clues to their cause.As with suspended POC, the absolute abundance of total fatty acids is greater and individual fatty acids are more 13 C-enriched under ice cover (CB9) compared to open water (CB4; Figs.2a and 4).However, normalizing fatty acid concentrations to suspended POC reveals the opposite pattern (Fig. 3), indicating that the composition of suspended POC is different at the two stations.MUFAs and PUFAs are also only significant contributors to total fatty acids at station CB9 (Fig. 3).While SFAs have more negative δ 13 C values than BFAs at both stations, the isotopic contrast is much greater at station CB9 (8.4 ‰) than CB4 (5.3 ‰; Table 2).Combined, these differences suggest different dynamics between bacterial and other components of POC at the two stations.
It has been reported that before the recent decline in summer sea ice in the Canada Basin, ice algae contributed up to 57 % of total primary productivity and released a large fraction (31-65 %) of the resulting organic matter as DOC (Gosselin et al., 1998).This labile organic carbon is likely to be enriched in 13 C compared to organic carbon produced by phytoplankton in the water column (Belt et al., 2008;Forest et al., 2007).The relative 13 C-enrichment of BFAs at station CB9 compared to station CB4 (Fig. 4) could therefore reflect greater bacterial dependence on organic carbon derived from sea-ice algae because strong isotopic similarity is expected between fatty acids produced by bacterial heterotrophs and their organic carbon source (Blair et al., 1985;Hayes, 2001;Monson and Hayes, 1982).At station CB9, the average δ 13 C value of BFAs falls between suspended POC (−27.0 ‰) and DOC (−21.7 ‰; Griffith et al., 2012) supporting greater bacterial incorporation of DOC under ice cover compared to station CB4 where the average BFA δ 13 C value is more negative and very similar to suspended POC (Table 2).It does not appear that SFAs at 50 m have a 13 C-enriched, ice-algal source at either station.Instead they are likely derived from phytoplankton in the water column.
During a late summer expedition in 2008, differences in the vertical structure of phytoplankton and bacterial abundances were observed between CB4 and CB9 (He et al., 2012) although comparable bacterial abundances were found at our sampling depth of 50 m.This was the chlorophyll maximum depth during our sampling and it represented a maximum in both bacterial and phytoplankton abundance at station CB4 (He et al., 2012).However at station CB9, a larger concentration of bacteria was found closer to the sea surface (He et al., 2012).The larger isotopic contrast between SFAs and BFAs at station CB9 could thus be a result of a spatial decoupling between primary production and heterotrophic production.The focusing of organic carbon production near the sea surface under greater sea ice coverage (Gosselin et al., 1998) suggests that DOC released from ice algae could fuel bacterial secondary production throughout the upper 50 m under ice cover, while phytoplankton production at the chlorophyll maximum depth directly supports bacterial production in open water.It is also possible that the ∼ 2 ‰ enrichment in both suspended POC and SFAs at station CB9 is a reflection of larger overall contributions from heterotrophic bacteria compared to station CB4.

Sequestration of 13 C-depleted fatty acids in surface waters
Suspended POC and individual fatty acids are significantly depleted in 13 C in the near-surface compared to the water column below (Fig. 4) while DIC and DOC are not (Griffith et al., 2012).This could be explained by an allochthonous source of 13 C-depleted particulate carbon, such as terrestrial material exported by the Mackenzie River.Although neither station is proximal to a source of terrestrial organic carbon, isotopic and elemental evidence indicates a substantial fraction of riverine fresh water becomes stored in the surface layer of the central basin in some recent years (Guay et al., 2009;Macdonald et al., 2002;Yamamoto-Kawai et al., 2009).It has also been shown that surface-water bacterioplankton can access and re-mineralize the presumably refractory terrestrial DOC delivered with it (Hansell et al., 2004).However, the summer halocline particle trap www.biogeosciences.net/10/7065/2013/Biogeosciences, 10, 7065-7080, 2013 (Jackson et al., 2010) may isolate this influence from our 50 m samples, as these samples were collected below the halocline in the Pacific Summer Water (PSW) layer (Griffith et al., 2012).Isotopic evidence for proportionally greater terrestrial contributions to DOC in the surface layer in 2008 is also lacking at both stations because the δ 13 C values of DOC are similar throughout the upper 350 m (Griffith et al., 2012).
In addition, a δ 13 C transition from "light" terrestrial values to a more 13 C-enriched marine primary source has been documented in sedimentary organic carbon much closer to shore at the base of the Beaufort slope (Naidu et al., 2000).An alternative explanation for "light" suspended POC and fatty acids at 50 m depth is a large effective fractionation factor between DIC and phytoplankton biomass.This depth coincides with the chlorophyll maximum at both stations, which a previous study found to be dominated by diatoms (Gosselin et al., 1998).At this light-limited depth, most 13 C-depleted fatty acids (C 14 , C 15 and C 16 ) could be produced by slowly growing marine diatoms expressing their maximum expected fractionation between 12 C and 13 C (Popp et al., 1998), combined with an additional isotopic effect on the large end of that observed between algal biomass and fatty acids (Schouten et al., 1998).The "heavier" fatty acids (e.g., iso-C 15 , anteiso-C 15 and C 18 ) are likely to have greater contributions from heterotrophic bacteria.Two of these, iso-C 15 , anteiso-C 15 , have a known bacterial source (Kaneda, 1991).The isotopic similarity between bacterial fatty acids and suspended POC (at station CB4) or between bacterial fatty acids and a combination of DOC and POC (station CB9) is consistent with bacterial heterotrophs incorporating organic carbon from POC and DOC (Blair et al., 1985;Hayes, 2001;Monson and Hayes, 1982).
The strong isotopic contrasts between 50 m and below (Fig. 4) seem to indicate a weak coupling between surface waters and the basin interior, caused by the ineffective biological pump that operates in the Canada Basin.The flux of organic carbon through the shallowest 200 m is orders of magnitude smaller than that found in other ocean basins, and this small flux is accompanied by similarly minuscule fluxes of diatom frustules and coccoliths (Honjo et al., 2010).Conditions and productivity at the sea surface do not appear to have a controlling influence on suspended POC and fatty acids at depth in the Canada Basin and the effects of sea ice coverage discussed above are confined to surface waters, also suggesting that seasonal influences may not strongly affect the deep basin.Not only are the concentration and isotopic composition of suspended POC very similar below 150 m at stations CB4 and CB9 (Griffith et al., 2012), but a more extensive survey of suspended POC concentrations below 100 m has revealed a general absence of gradients in the interior Canada Basin (Jackson et al., 2010).Bacterial abundances are also similar below 50 m near stations CB4 and CB9 (He et al., 2012;Uchimiya et al., 2013).No significant differences can be observed in fatty acids recovered from both stations at 1000 m, either.Major differences do exist in the abundance and distribution of fatty acids at 150 m depth, but it is more likely that this sample represents an external influence such as POC delivered by a mesoscale eddy because fatty acids are both unlike station CB4 at 150 m (Fig. 2a) and unlike fatty acids above and below at station CB9 (e.g., small relative concentrations of C 14 , iso-C 15 and anteiso-C 15 fatty acids compared to 50 and 1000 m).Such eddies have been documented at approximately 150 m depth at locations more proximal to the Northwind Ridge (Fig. 1), and are known to transport nutrient-and organic carbon-rich Pacific-origin seawater to the western Canada Basin (Mathis et al., 2007;Pickart, 2004).

Bacterial heterotrophy in the interior Canada Basin (150-2500 m)
Multiple features in the concentrations and δ 13 C values of fatty acids indicate that heterotrophic bacteria suspended in the water column are important contributors compared to other sources of fatty acids in the interior Canada Basin.The broad subsurface peak in the POC-normalized abundance of BFAs in the shallowest 1000 m at station CB4 (Fig. 3a) may be produced by a relatively large proportion of productive bacterial heterotrophs synthesizing branched fatty acids at mesopelagic depths where Honjo et al. (2010) have proposed intense bacterial heterotrophy contributes to the weak biological pump.The isotopic divergence between BFAs and even-numbered SFAs (∼ 5 ‰, Table 2) is, arguably, the most consistent pattern in fatty acids above 3000 m depth and another indication of strong bacterial heterotrophy.Such divergent δ 13 C values between BFAs and other fatty acids are not reported in suspended POC from surface waters of the Mediterranean (Tolosa et al., 2004), near-surface, oxic depths of the Black Sea (Wakeham et al., 2007) or near-surface, oxic depths of the Cariaco Basin (Wakeham et al., 2010).Instead, δ 13 C values of fatty acids and suspended POC in the interior Canada Basin evoke the isotopic relationship observed in sinking particulate matter (Tolosa et al., 2004), as well as net-heterotrophic riverine (Zou et al., 2006) and estuarine (Boschker et al., 2005) settings where external input of terrestrial organic carbon supports bacterial secondary production.Unlike turbid estuaries and rivers, the Canada Basin is particle-poor (Griffith et al., 2012), but a significant external, and possibly episodic source of laterally transported organic carbon from the productive Chukchi Sea (Davis and Benner, 2007;Mathis et al., 2007;Shen et al., 2012;Walsh et al., 1989) or from re-suspended Beaufort Slope sediments (Honjo et al., 2010;Hwang et al., 2008;Jackson et al., 2010;O'Brien et al., 2013) has been shown.The close isotopic relationship between BFAs and suspended POC (< 1 ‰ except 50 m at CB9, Table 2) at each individual depth above 3000 m is also striking (Fig. 4).Because of the small expected isotopic fractionation between bacterial heterotrophs and substrate organic carbon (Blair et al., 1985;Monson and Hayes, 1982) are produced in situ by heterotrophic bacteria suspended in the meso-and bathypelagic Canada Basin.Bacterial incorporation of DOC can not be ruled out between the depths of 500 and 2500 m, however, due to the close isotopic similarity between POC and DOC (Fig. 4a; Griffith et al., 2012).Oddnumbered SFAs are also isotopically more similar to average BFAs than average SFAs below 150 m, suggesting that a significant fraction of straight-chain C 15 and C 17 fatty acids are produced by heterotrophic bacteria.
4.5 Sources of 13 C-depleted fatty acids in the interior Canada Basin (150-2500 m) The Canada Basin is vertically stratified, and the influence of the Pacific-origin, Atlantic-origin and isolated deep basin waters can be seen in the isotopic composition of DIC, DOC and suspended POC (Griffith et al., 2012).The vertical distribution and δ 13 C values of fatty acids, however, do not appear to be strongly influenced by water mass except for the possibility of Pacific-origin Summer Water (PSW), from which the 50 m samples were obtained (Griffith et al., 2012).Instead, fatty acid distributions and isotopic compositions suggest three distinct zones in the water column: 50 m, 150-2500 m, and 3000-3775 m.
Here we discuss the largest region, 150-2500 m, where a 13 C-depleted source of SFAs is suggested by the subsurface peak in the POC-normalized abundance of SFAs at 2500 m depth (Fig. 3a), 95 % of which is made up of C 16 and C 18 fatty acids (Fig. 2b).C 16 fatty acid shows a progressive 13 Cdepletion between 500 and 2500 m depth, opposite to the trend expected from decreasing contributions from surfacederived, "light" organic matter and increasing relative contributions from "heavy" bacterial heterotrophs.The δ 13 C value of saturated C 18 fatty acid also remains similar to its value at 50 m rather than trending towards a more positive value with depth as would be consistent with an increasingly important bacterial heterotrophic source.Near these depths, the possibility of significant microbial DIC incorporation was suggested by an isotopic mixing model of radiocarbon and δ 13 C values of bulk carbon pools (Griffith et al., 2012) pointing towards a third, chemoautotrophic source of fatty acids in the interior Canada Basin.
We identify four sources of organic matter that may contribute to the total fatty acids that we recovered between 150 and 2500 m: (1) phytoplankton-derived organic carbon produced in surface waters; (2) laterally delivered organic carbon originating either from primary productivity in the Chukchi Sea or from re-suspended shelf sediments; (3) suspended bacterial cells with unspecified metabolism; and (4) non-particulate organic carbon adsorbed onto the glass fiber filters during sample collection.We have explicitly corrected for adsorbed organic carbon (Tables 1 and 2) and only consider the first three fatty acid sources in the following model and discussion.
In order to explore the potential importance of the three fatty acid sources to the interior Canada Basin, we construct an isotopic mass balance model according to (1) where δ represents δ 13 C value.We focus on 1000-2500 m where the isotopic divergence between SFAs and BFAs is most pronounced at station CB4 (Fig. 4a).We also consider only C 16 and C 18 fatty acids, which represent 84-95 % of total fatty acids at these depths (Table 1).δ surface : although the sinking flux of organic carbon is known to be very small (Honjo et al., 2010), and the 13 Cdepleted isotopic signature of fatty acids in surface waters does not appear to form the major component of fatty acids in suspended POC below 50 m depth (Fig. 4), surface-derived organic matter delivered with sinking particles will still contribute some fraction of fatty acids to the interior Canada Basin.We assign δ surface as the average δ 13 C value of C 14 and C 16 at 50 m at station CB4.Because the δ 13 C value of C 18 is more similar to BFAs than SFAs, it likely represents greater contributions from heterotrophic bacteria and is therefore not included in our endmember value for phytoplankton-derived organic matter.Although 50 m may appear deep for a surface source, we believe our 50 m samples, taken at the chlorophyll maximum, to be a reasonable proxy for the isotopic composition of sinking fatty acids.The mechanisms of organic matter aggregation and the depths from which it is dominantly exported are not well-understood in the central Canada Basin, but the sequestration of nutrients below the summer halocline means that a large proportion of total water column primary productivity occurs at the subsurface chlorophyll maximum depth (Lee et al., 2012;Martin et al., 2012).A recent analysis of suspended particulate matter from the coastal Beaufort Sea also identifies a POC maximum and biomarker indications of "fresh" organic carbon at the chlorophyll maximum depth (Tolosa et al., 2013).This is unlike lower-latitude ocean basins where the subsurface chlorophyll maximum does not correspond to a biomass or productivity maximum.
δ advected : the δ 13 C values and distribution of fatty acids from allochthonous organic carbon in the interior Canada Basin are much more difficult to assign because of multiple and poorly defined possible sources.Organic carbon from primary productivity over the Chukchi (Davis andBenner, 2005, 2007;Shen et al., 2012) or Beaufort shelves (Ortega-Retuerta et al., 2012) could be incorporated into suspended POC following deposition to sediments, re-suspension and lateral transport (Mathis et al., 2007).Saturated C 14 , C 16 and C 18 fatty acids in sediments at the base of the Beaufort slope, presumably with a Beaufort Sea primary origin, have more enriched δ 13 C values relative to those between 150 and www.biogeosciences.net/10/7065/2013/Biogeosciences, 10, 7065-7080, 2013 2500 m (Drenzek et al., 2007;Goñi et al., 2005).Sea-ice algae will also contribute organic carbon that is 13 C-enriched (Belt et al., 2008;Budge et al., 2008).Bioavailable DOC with undefined δ 13 C value may also be incorporated directly into suspended POC by aggregation processes (Burd and Jackson, 2009;Engel et al., 2004).We use δ 13 C values from ice algal, phytoplankton and copepod fatty acids reported from samples collected near Barrow, Alaska (Budge et al., 2008) to represent δ advected values for fresh and re-worked sources of organic carbon from nearby productive waters (Table 3).Beaufort shelf sediments also host organic matter delivered by the Mackenzie River (Drenzek et al., 2007;Goñi et al., 2005;Yunker et al., 2005), and their mobilization could deliver marine, terrestrial and riverine organic carbon to the interior Canada Basin.We also use the δ 13 C value of fatty acids from Mackenzie River suspended POC (Goñi et al., 2005;Tolosa et al., 2013) and Beaufort slope sediments (Drenzek et al., 2007;Goñi et al., 2005;Tolosa et al., 2013) to represent δ advected values for sedimentary fatty acids (Table 3).
f bacteria : because the δ 13 C value of bacterial fatty acids will depend on an unknown fraction of bacterial heterotrophs vs. chemoautotrophs and on chemoautotrophic pathways, we allow δ bacteria to be an unknown variable and instead estimate f bacteria from prokaryotic abundances and a biomassto-phospholipid fatty acid conversion factor assuming phospholipid fatty acids are the major source of fatty acids in marine bacteria (Oliver and Colwell, 1973;Oliver and Stringer, 1984).Using prokaryotic abundances reported by Uchimiya et al. (2013), the carbon content of bacterial cells in oligotrophic waters (4-7 fg C cell −1 ; Christian and Karl, 1994;Gundersen et al., 2002), and considering that 49 % of microbial cells pass through GF/F filters (Lee et al., 1995), we calculate that 11-14 % of suspended POC captured in our samples could be attributed to prokaryotic biomass carbon between 1000 and 2500 m depth (Fig. 5).Although δ 13 C values of BFAs indicate that at least some of this prokaryotic biomass is heterotrophic bacteria (Sect.4.4), this range falls within the 10-22 % of suspended POC that Griffith et al. (2012) attribute to chemoautotrophic biomass based on the radiocarbon similarity between DIC and suspended POC and the modest 13 C enrichment in POC between 1000 and 2500 m.Total prokaryotic biomass reported by Uchimiya et al. (2013) includes both bacterial and archaeal cells while only bacterial cells will be represented in fatty acid abundances.Still, there appears to be a relatively constant offset between the fraction of POC from total prokaryotic biomass and from fatty acid carbon (Fig. 5) which supports suspended prokaryotes as a dominant source of fatty acids at these depths.In their intact phospholipid form, fatty acids are thought to be a constant fraction of bacterial biomass (Balkwill et al., 1988;Findlay et al., 1989;White et al., 1979).The ratio of fatty acid carbon to biomass carbon is poorly constrained for slowly growing deep ocean bacterioplankton, but for a sedimentary enrichment culture with large bacterial Proportions of total POC that can be attributed to prokaryotic biomass and organic carbon from fatty acids (FA).Prokaryotic organic carbon calculated from abundances reported by Uchimiya et al. (2013) with a conversion factor of 6.5 ± 2.5 fg carbon cell −1 (Christian and Karl, 1994;Gundersen et al., 2002) and assuming 51 % recovery on GF/F filters (Lee et al., 1995).Fatty acid abundances converted organic carbon equivalents using the weighted average % C for fatty acids at each depth.
cells (56 fg C cell −1 ), 21 % of biomass carbon could be attributed to fatty acid carbon from phospholipids (Findlay et al., 1989).It has also been observed that approximately half of total phospholipid is lost upon transitioning a marine bacterium from nutrient-replete culture conditions to suspension in seawater (Oliver and Stringer, 1984), so on this basis we assume fatty acid carbon is approximately 10 % of biomass carbon.An additional assumption about the proportional balance between bacterial and archaeal cells must be made to calculate f bacteria .A dominance of bacteria was reported below 500 m at the base of the Chukchi Slope (Kirchman et al., 2007), however this dominance would yield more bacterial fatty acid than we recovered from all depths between 500 and 2500 m.The maximum proportion of bacteria that fatty acid abundances allow is 54 %, closer to the fraction of bacteria found at meso-and bathypelagic depths of the oligotrophic North Pacific and Atlantic Oceans (Herndl et al., 2005;Karner et al., 2001).Conservatively, we assume that 50 % of prokaryotic cells are bacterial and that 10 % of their biomass carbon can be attributed to fatty acid carbon.Using an abundance-weighted fatty acid composition to calculate fatty acid carbon from observed fatty acid abundance, these assumptions yield an f bacteria value that averages 80 % between 1000 and 2500 m.Defining δ measured as the average δ 13 C value for C 16 and C 18 fatty acids between 1000 and 2500 m depth (−27.5 ‰), three unknown variables remain: f surface , f advected and δ bacteria .Combining equations ( 1) and ( 2) to eliminate f surface leaves: relating f advected to δ bacteria by a linear equation in the form y = mx + b. Figure 6 illustrates the relationship between f advected and δ bacteria for the full range of possible f advected values (0-20 %).Considering that f advected + f surface = 20 %, this corresponds to f surface values of 20 % to 0 % (Fig. 6, secondary horizontal axis).Each line in Fig. 6 represents a solution to the mass balance equation for a different advected component (summarized in Table 3).Bacterial heterotrophs will likely synthesize fatty acids with δ 13 C values similar to BFAs and POC at each depth (Sect.4.4): −21 to −23 ‰ between 1000 and 2500 m.Model results, however, are uniformly more negative than these values implying production of 13 C-depleted fatty acids by suspended bacteria at depth.Chemoautotrophic bacteria utilizing the RubisCO enzyme and Calvin Cycle to reduce inorganic carbon could contribute 13 C-depleted fatty acids to the interior Canada Basin while their biomass would be isotopically similar to heterotrophic bacteria (Hayes, 2001;Sakata et al., 2008).The genetic potential for this metabolic pathway has been identified at mesopelagic depths of other oligotrophic regions (Swan et al., 2011), and the expected δ 13 C value of chemoautotroph-derived fatty acids of ∼ −28 ‰ is similar to observed δ 13 C values for C 16 and C 18 between 1000 and 2500 m depth.This value is calculated from a likely isotopic fractionation between DIC and fatty acids of ∼ 29 ‰ (Sakata et al., 2008) and the δ 13 C value of DIC between 500 and 2500 m which is +1 ‰ (Griffith et al., 2012).With defined end-member δ 13 C values for bacterial heterotrophs and chemoautotrophs, it is also possible to calculate their proportions through two-component isotopic mix- ing based on model output δ bacteria (Fig. 6, secondary vertical axis).Chemoautotrophic production of fatty acids appears to be a significant contributor to the total fatty acid pool regardless of the advected source of fatty acids.A minimum contribution from chemoautotrophic bacteria of 67 % is obtained in the unlikely case of purely terrestrial advected material (δ advected = −35.0‰ representing Mackenzie River POC) with a maximum f advected value of 0.2.This result suggests that, at minimum, chemoautotrophic bacteria contribute 67 % of the total fatty acids recovered from suspended particulate material.

Allochthonous fatty acids between 3000 and 3775 m
In the deepest 1000 m, the abundance of total fatty acids no longer decreases with depth (Fig. 2a), exhibiting a similar trend to prokaryotic abundance (Uchimiya et al., 2013).
www.biogeosciences.net/10/7065/2013/Biogeosciences, 10, 7065-7080, 2013 The distribution of fatty acids is also distinct from depths above with a greater representation of long-chain fatty acids (C 20−24 ), BFAs and MUFAs (Fig. 2b).This compositional difference is likely a reflection of an allochthonous source of particulate matter to the deep basin.Re-suspension and lateral transport of sediments from surrounding margins dominates sinking POC at these depths (Hwang et al., 2008;Honjo et al., 2010), and this has also been identified as an important component of suspended POC (Griffith et al., 2012).The source of this allochthonous organic carbon could be the terrestrially dominated Mackenzie and Beaufort slopes to the south or the Chukchi slope and base of the Northwind Ridge to the west.At the base of the Beaufort Slope, shallow sediments host abundant short-chain SFAs and MUFAs, as well as proportionally important long-chain fatty acids (C 23−32 ) and BFAs (Belicka et al., 2004).Further investigations of long-chain fatty acids at the base of the Beaufort slope revealed a normal distribution of even-numbered SFAs centered around C 24 which derive primarily from plant waxes (Drenzek et al., 2007).Combined, these resemble the distribution of fatty acids in suspended POC (Fig. 2a) with the exception of the shorter maximum chain length.Although C 24 is the most abundant long-chain fatty acid in suspended POC, we do not detect C 26−32 fatty acids and suspect these may have been lost to degradation during their suspension in oxic bottom waters (cf.Rontani et al., 2012).Organic carbon in the sediments of the Chukchi slope reflects primarily a marine source from the productive Chukchi Sea (Belicka et al., 2002(Belicka et al., , 2004)).The distribution of fatty acids is subtly different than the base of the Beaufort slope (Belicka et al., 2004) with only trace C 14 saturated fatty acids, and a distribution of long-chain fatty acids centered around C 26 rather than C 24 (Belicka et al., 2002).These differences, combined with the greatest particle load observed in the southern Canada Basin (Jackson et al., 2010), make the Beaufort slope a more likely source region for suspended POC in the abyssal Canada Basin than the Northwind Ridge.
There are also isotopic differences between the deepest 1000 m and above.The δ 13 C value of DOC and suspended POC are different by ∼ 1.5 ‰, unlike the majority of the water column above (Fig. 4a; Griffith et al., 2012).But, in general, there is a trend towards 13 C-enrichment of SFAs resulting in more isotopic similarity between suspended POC and SFAs, as well as between SFAs and BFAs at these depths (Table 2; Fig. 4a).It is possible that intense bacterial heterotrophy supported by re-suspended sedimentary organic carbon contributes a larger proportion of SFAs at deep depths.Another possibility is that a larger proportion of total fatty acids is delivered to the suspended POC reservoir from a sedimentary source.The isotopic variability in fatty acids from the deepest depth, 3775 m, may reflect this greater sedimentary source.
The exceptions to this pattern are C 12 and iso-C 17 fatty acids, both of which have δ 13 C values more negative than any value higher in the water column (Fig. 4a).δ 13 C values of C 12 and iso-C 17 at 3000 m are also more negative than fatty acids in the Mackenzie River (Goñi et al., 2005).Although some of these values can therefore be explained by a terrestrial source, a more 13 C-depleted origin is required for iso-C 17 .Bacterial oxidizers of 13 C-depleted methane may be their origin since methane and gas hydrates have been observed in the coastal Beaufort Sea (Dallimore and Collett, 1995;Paull et al., 2007).Synthesis of iso-C 17 could occur in sediments hosting methane hydrates that are subsequently mobilized, or iso-C 17 could be produced in situ if there is a source of methane to the deep basin.

Conclusions
In the interior Canada Basin, suspended fatty acids show evidence of both advected and in situ sources.Surface waters host fatty acids that are more depleted in 13 C than in the dark basin below and also reflect contrasting ecological conditions under ice cover compared to open water.The 13 C depletion, not observed in DIC or DOC, are likely to result from slowly growing diatoms, but possibly also from the delivery of terrestrial organic carbon to the central Canada Basin.Deeper depths appear to be isolated from the effects of ice cover, but at all depths, a strong isotopic similarity between POC and branched fatty acids is apparent supporting the hypothesis that intense bacterial heterotrophy contributes to a weak biological pump in the Canada Basin (Honjo et al., 2010).An additional, 13 C-depleted source of saturated, even-numbered fatty acids between 1000 and 2500 m is also indicated, dominantly (> 67 %) of chemoautotrophic origin.Lateral advection of DOC from the Chukchi Sea, and POC re-suspended from sediments are also potential sources.We believe that the excess abundance of prokaryotes in the deep Canada Basin, compared to sinking POC flux (Nagata et al., 2010;Uchimiya et al., 2013;Yokokawa et al., 2013), could therefore be explained by a combination of bacterial heterotrophs supported by a lateral supply of organic carbon and chemoautrotrophic bacteria utilizing the Calvin Cycle to fix inorganic carbon with a yet unknown energy metabolism.Fatty acids in the deepest 1000 m reflect the contributions of re-suspended sediments (Hwang et al., 2008) with greater relative abundances of long-chain fatty acids (C 20−24 ), and isotopic dissimilarity compared to shallower depths.Two individual fatty acids recovered from 3000 m and below, C 12 and iso-C 17 are too depleted in 13 C to derive from the water column or shallow sediments and more likely are produced by methane oxidizing bacteria hinting at a poorly defined methane cycle in the abyssal Canada Basin.
Fig. 2. (a) Summed concentration of total fatty acids.(b) Relative abundances of individual fatty acids with blue shades representing even-numbered fatty acids and orange shades representing oddnumbered and branched fatty acids and purple shades representing long-chain fatty acids.

Fig. 4 .
Fig. 4. δ 13 C values of individual fatty acids with depth where SFAs are in blue shades, BFAs are in orange shades and open circles represent MUFAs and PUFAs for (a) Station CB4 and (b) Station CB9.POC and DOC(Griffith et al., 2012) are represented by black lines.
Fig. 5.Proportions of total POC that can be attributed to prokaryotic biomass and organic carbon from fatty acids (FA).Prokaryotic organic carbon calculated from abundances reported byUchimiya et al. (2013) with a conversion factor of 6.5 ± 2.5 fg carbon cell −1 (Christian andKarl, 1994;Gundersen et al., 2002) and assuming 51 % recovery on GF/F filters(Lee et al., 1995).Fatty acid abundances converted organic carbon equivalents using the weighted average % C for fatty acids at each depth.

Fig. 6 .
Fig.6.Results of isotopic mass balance model for δ 13 C bacteria as a function of f advected between 1000 and 2500 m depth.Each line represents a solution for isotopically distinct sources of advected organic carbon with f surface plotted on the secondary horizontal axis and the fraction of chemoautotrophic vs. heterotrophic bacteria on the secondary vertical axis.f bacteria is 0.8.

Table 1 .
Concentration of POC and FAMEs in the Canada Basin.

Table 2 .
Blank-corrected δ 13 C values of POC and FAMEs in the Canada Basin.

Table 3 .
δ 13 C values of organic matter sources that could be laterally supplied to Canada Basin.