Population dynamics of modern planktonic foraminifera in the western Barents Sea

. This study reports on diversity and distribution of planktonic foraminifera (PF) in the Barents Sea Opening (BSO). Populations of PF living in late summer (collected by means of stratified plankton tows) and recently deposited individuals (sampled by interface corer) were compared. High abundances reaching up to 400 ind.m -3 in tow samples and 8000 ind.cm -3 15 in surface sediments were recorded in the centre of the studied area while low abundances were observed in coastal areas, likely due to continental influences. The living and core-top assemblages are mainly composed of the same four species Neogloboquadrina pachyderma, Neogloboquadrina incompta , Turborotalita quinqueloba and Globigerinita uvula . The two species G. uvula and T. quinqueloba dominate the upper water column, whereas surface sediment assemblages display particularly high concentrations of N. pachyderma . The unusual dominance of G. uvula in the water sample assemblages 20 compared to its low proportion in surface sediments might be the signature of 1) a seasonal signal due to summer phytoplankton composition changes at the BSO, linked to the increase of summer temperature at the study site, and/or 2) a signal of a larger time-scale and wider geographical reach phenomenon reflecting poleward temperate/subpolar species migration and consecutive foraminiferal assemblage diversification at high latitudes due to global change. Protein concentrations were measured on single specimens and used as a proxy of individual carbon biomass. Specimens of all species show the same 25 trend, a northward decrease of their size-normalized-protein concentration. This suggests that foraminiferal biomass is potentially controlled by different constituents of their organelles (e.g. lipids). The coupling of data from plankton tows, protein measurements and surface sediments allows us to hypothesise that PF dynamics (seasonality and distribution) are decoupled from their metabolism.


Introduction
The introduction is somewhat mixed-up. Paleoceanographic information is mixed with modern assemblage studies, habitat depth mixed with seasonal variability, foraminifera with other organisms... In some places, too much details is given (e.g., 104 tow hauls in Greco et al. 2019). Suddenly the Southern Indian Ocean pops out... Please consider rewriting to better structure this section.
A : we worked on the introduction in order to structure it better and keep what we consider important information.
A : we deleted the sentence 54: Change "observed PF by the means of plankton tows," to "analyzed PF collected with plankton tows". Please note that Pados and Spielhagen (2014) analyzed both forams living in Polar and Atlantic waters and used both plankton tows and core top samples.
A : we replaced the sentence 73: Change "planktonic foraminifera" to PF (be consistent in using the abbreviations that you introduced) A : correction made 79: Change "planktonic foraminifera from a same species" to "PF of the same species" A : correction made 2. Oceanographic setting 84: Change "Spitsbergen Banken and shallow Bjørnøya; Storfjordrenna and Bjømøyrenna glacial troughs" to "Storfjordrenna and Bjørnøyrenna glacial troughs separated by shallow Spitsbergen Banken". Please verify if it's Spitsbergen Banken or Spitsbergenbanken.
A: we made the corrections L.91 and after verification in the literature, decided to use Spitsbergenbanken (correction also made L.102 for this term) 87: Nothing is written about the currents carrying Arctic Water to the study area.
A : Indeed we didn't mention Arctic Water in the paper as none of the discussed samples were collected in this water layer. Arctic Waters were only encountered twice during the cruise and mixed with NwAW (Giraudeau et al., 2016).
From 88 onwards: To my understanding, the Oceanographic section should only contain the state-of-the-art on the subject. If the authors performed some oceanographic measurements, please move the information to "Material and Methods" and "Results" sections.
A : we changed these sections of the manuscript accordingly 98: "above described" => "described above" A : correction made 3. Material and Methods 112 and elsewhere: You use either "Spitsbergen", "Spitzbergen" or "Spitzberg". Please unify. To my knowledge, "Spitsbergen" is the English spelling, while "Spitzbergen" is German and "Spitzberg" French.
A : we thank the reviewer for this observation and unified using Spitsbergen 118: Change "collection" to e.g., "sampling strategy" A : we amended following the reviewer's comment 124: "All living PF" -if they were preserved with ethanol, they were not living anymore. Change to, e.g., "All foraminiferal tests containing coloured cytoplasm ("living")...". Where the samples stained with Rose Bengal? If so, this should be mentioned. Otherwise, how were they coloured?
A : the net samples were not stained and individuals that was living at the moment of the sampling (the one we call "living") were distinguished by the natural coloration of their cytoplasm. We therefore choose not keep this sentence intact. 130: "separately numbered" => "counted separately" A : correction made 137: "Individual" => "The individuals" A : correction made 146: Referee#2 suggested using "core-top" instead of "subfossil". I think it also concerns the term "fossil".
A : we disagree and would prefer to keep fossil here.

147: INSU 1 -please check if the journal accepts footnotes.
A : we removed the footnote and place detailed the accronym INSU directly in the text. 149: Delete "the more horizontal" -something is horizontal or not, it can't be more or less horizontal.
A : we replaced horizontal by "even" 149-150: Change "The core-top sediment (0-0.5 cm slice)" to "The uppermost 0.5 cm of the core" A : we replaced following the Reviewer's suggestion 152-156: Please rewrite the two sentences so that they are more related to each other. A : the paragraph has been slightly rewritten but kept as we think it provides guidance to the reader.
169: The highest PF concentrations were found at the edge of the NwCW range (station 7) so I would refrain from saying that the highest concentrations were found in NwCW.
A : we changed the sentence to precise highest concentrations were found at the edge of the NwCW.
173: It should be specified which species are considered polar and which subpolar by the authors.
A : We used the SCOR WG138 taxonomy for this study where only N. pachyderma is described as a polar species. As the used taxonomy is mentioned in the Material and Methods, we would rather not comment on it in this section.
188: "analyse" => "analysis" or "analyses" A : corrected 194-195: The information in brackets was already given in the methods and is unnecessary here.
A : we removed the breckets 195: "successfully" -I assume you wouldn't mention them at all in the manuscript if they were unsuccessful.
A : we removed "successfully" 204: Is it exactly equal (down to 0.00000001 μm) or close enough to saz that the siye distribution of the picked tests is szmmetric?
A : we added precisions to the text 206: μm => μg A : correction made 207: μg => μg/μm (or μg*μm -1 ) A : correction made 209: "slightly (but not significantly)" -this is not very specific, please rephrase by, e.g., giving some numbers A : we mentioned the Figure 6 once again here. A : we prefer to keep Figure 1 as simple as possible and would rather not add the location of the Polar front. We cited Oziel et al., 2017 here, presenting very accurate data for it. Moreover, our transect don't cross the polar front.
A : we provided this additional information 375: "is" => "in" A : we did the correction Correspondence to: Julie Meilland (jmeilland@marum.de) Abstract. This study reports on diversity and distribution of planktonic foraminifera (PF) in the Barents Sea Opening (BSO).
Populations of PF living in late summer (collected by means of stratified plankton tows) and recently deposited individuals (sampled by interface corer) were compared. High abundances reaching up to 400 ind.m -3 in tow samples and 8000 ind.cm -3 15 in surface sediments were recorded in the centre of the studied area while low abundances were observed in coastal areas, likely due to continental influences. The living and core-top assemblages are mainly composed of the same four species Neogloboquadrina pachyderma, Neogloboquadrina incompta, Turborotalita quinqueloba and Globigerinita uvula. The two species G. uvula and T. quinqueloba dominate the upper water column, whereas surface sediment assemblages display particularly high concentrations of N. pachyderma. The unusual dominance of G. uvula in the water sample assemblages 20 compared to its low proportion in surface sediments might be the signature of 1) a seasonal signal due to summer phytoplankton composition changes at the BSO, linked to the increase of summer temperature at the study site, and/or 2) a signal of a larger time-scale and wider geographical reach phenomenon reflecting poleward temperate/subpolar species migration and consecutive foraminiferal assemblage diversification at high latitudes due to global change. Protein concentrations were measured on single specimens and used as a proxy of individual carbon biomass. Specimens of all species show the same 25 trend, a northward decrease of their size-normalized-protein concentration. This suggests that foraminiferal biomass is potentially controlled by different constituents of their organelles (e.g. lipids). The coupling of data from plankton tows, protein measurements and surface sediments allows us to hypothesise that PF dynamics (seasonality and distribution) are decoupled from their metabolism.

Introduction
Polar areas are sensitive to global temperature changes, particularly in the Arctic where warming occurs faster than in the rest of the world and has accelerated over the past 50 years (Shepherd, 2016). This Arctic amplification appears to be mainly caused by sea-ice loss under increasing CO2 (Dai et al., 2019). Recently increased advection of Atlantic Water into the Barents 50 Sea modifies its physico-chemical properties (Smedsrud et al., 2013), which directly affect the entire ecosystems in the region.
Higher temperatures lead to increased rates of planktonic primary production (Vaquer-Sunyer et al., 2013) and increased CO2 concentrations are expected to have a fertilization effect on marine autotrophs (Holding et al., 2015). Not only is productivity increasing, but spring and summer blooms are also occurring earlier in the European Arctic Ocean (Oziel et al., 2017). As a response, some taxa of calcifying groups (i.e. foraminifera, coccolithophores, molluscs and echinoderms; Beaugrand et al.,55 2013) exhibit a poleward movement in agreement with expected biogeographical changes under sea temperature warming.
Both satellite images (Smyth et al., 2004;Burenkov et al., 2011) and in situ measurements (Dylmer et al., 2013;Giraudeau et al., 2016;AMAP 2018) have shown rapid expansion of temperate species of coccolithophores in the Arctic. For example, Emiliania huxleyi shows a striking poleward shift (>5°) in the distribution of its blooms (Neukermans et al., 2018). This phenomenon, called "atlantification" (Årthun et al., 2012), is expected to impact every trophic level of the food web, from 60 small phytoplankton species (Neukermans et al., 2018) to larger organisms (Dalpadado et al., 2012). Recent studies have investigated the ecology and biodiversity of planktonic foraminifera from the high-latitude North Atlantic (e.g. . The species N. pachyderma comprises more than 90% of recent assemblages (i.e. found in surface sediments) from the Polar Region, North of Iceland (Kucera et al., 2005). Rather few studies on living planktonic foraminifera (PF) communities have focussed on (sub-) Arctic regions. Plankton tows also show that N. pachyderma is dominant species in Arctic (~90%) 65 followed by T. quinqueloba (5%) (Volkmann, 2000;Pados and Spielhagen, 2014). Through the compilation of population density profiles from 104 stratified plankton tow hauls collected in the Arctic and the North Atlantic Oceans, Greco et al. (2019) investigated the ecology of N. pachyderma. In particular, the variability of its habitat depth, and underlined the knowledge gap on its ecological preferences. In the western subpolar North Atlantic (Irminger Sea), the production of N. pachyderma shows two peaks, in spring and late summer, while winter shows a low production (Jonkers et al., 2010;. 70 The diversity of PF has increased in polar waters over the past decades, even though it remains low in comparison to lower latitudes, due to the poleward migration of warm-water species . A similar process appears to occur in the southern hemisphere (Meilland et al., 2016). Some species from lower latitudes are described as new components of poleward assemblages. The shift of PF assemblages to warmer conditions, since the pre-industrial stage, has been very recently highlighted more globally in the Northern hemisphere . These major modifications in PF distribution 75 patterns display changes more related to primary production than to water temperature itself (e.g. Jonkers et al., 2010;   . More studies on living PF communities in the Arctic regions are needed to assess the spatial and temporal variability in their population dynamics and to better constrain the today's polar and subpolar species ecological preferences. Taking the opportunity of a cruise dedicated to the exploration of the physical oceanography of the western Barents Sea 125 (MOCOSED 2014 cruise), we investigated the connections between the spatial variability of living PF, phytoplankton communities (Giraudeau et al., 2016), and the hydrological system through a South-to-North transect, between Northern Norway and Spitsbergen [68-76]°N. Along this transect, we compared PF living faunas (from plankton tow) to the assemblages found on the sea floor (from core-top sediments) in order to investigate eventual recent changes in their population dynamics.
This latitudinal transect also gave us the opportunity to quantify protein concentrations of individual living PF in this area for 130 the first time and along a physico-chemical gradient to see if and how it varies and explore how planktonic foraminifera from a same species may adjust to different environments.

Oceanographic setting
The studied area covers the western Barents Sea margin, i.e. Barents Sea Opening (BSO), where the surface and intermediate Prymnesiophytes (13 to 24%; major component Emiliana huxleyi) (Giraudeau at al., 2016). Three other features are noteworthy (Figure 2 d): i) the dominance of dinoflagellates (24%) at the southernmost station of the transect (close to the Norwegian coast) contrasted with their total absence in the well mixed NwAW, North of 74.5; ii) the presence of diatoms (10-20 %) in the surficial NwCW, but rare (<5%) to the North; iii) the constant increase in relative abundance of Cyanobacteria from < 5% to more than 15% along the South-to-North transect. 175

Material and Methods
In late summer 2014, from August 8 to September 20, the SHOM (French Hydrographic Office) operated the oceanographic cruise MOCOSED 2014, on board the "R/V Pourquoi pas?". Along a 700 km South-to-North transect from the Norwegian (68°N) to the Spitsbergen (76°N) coasts, investigations of hydrological processes at the BSO were carried out coupled with the exploration of the phytoplankton and foraminiferal communities using a total of 32 vertical casts deployed ≈ 20 km apart 180 from each other (Figure 1 and 2).

Living planktonic foraminifera from stratified plankton samples (MultiNet)
Living PF were collected at 7 of the 32 CTD South-to-North transect stations (#3 to #9), and at 2 stations (#1 and #2) located West-to-East ≈ 20 km apart from the central point of the main South-to-North CTD transect ( Figure 1; Table 1), using a stratified plankton tow (MultiNet Hydro-Bios type Midi, opening of 0.25 m 2 ) equipped with five nets (mesh size 100 μm to 185 avoid nets clogging in case of intense phytoplankton bloom). This sampling strategy was used in order to observe the potential effect of latitudinal changes but also of bathymetry, longitudinally, on PF distribution. At each station, one vertical haul sampled five successive water layers from the sea surface to 100 m depth. A second hauls has been deployed for the stations 1, 3, 4, 5, 6, 7 and 8, to collect material down to 700 m depth (Table 1) with ethanol (90%) buffered with hexamethylenetetramine until processing at the land-based laboratory. Back at the laboratory, MultiNet samples were washed over a 100 µm mesh, all foraminifera were removed from the sample and dried in an oven at 50 °C. All living PF, distinguished by their coloured cytoplasm visible through the shell, were picked, stored in counting cells and identified at the species level, following the SCOR WG138 taxonomy as implemented in Siccha and Kucera (2017). Empty 195 tests, considered as dead individuals were counted separately. Correlations following a non-metric multidimensional scaling ordination (NMDS) were carried out with the R package Vegan (Oksanen et al., 2013). Using the Bray-Curtis distance these correlations were tested between PF species absolute concentrations, the latitude of the station and parameters of the ambient waters (temperature, salinity, Chl-a concentration). Results are given in relative abundances (% of the total, live or dead fauna) or in absolute abundances in number of individuals per m 3 of filtered water (ind.m -3 ). 200 For protein extraction and measurement, a few living individuals (≈60) were picked on board from out of the shallowest water samples at stations 3 to 9 immediately after sampling. Only shells that were completely filled with cytoplasm were selected.

Protein biomass and test size measurements
After picking, individuals were immediately cleaned with a brush and filtered seawater to remove all particles, including organic matter, that were stuck to the test. The individuals were stored in a 1.5 mL Eppendorf vial and analysed on board, 220 using the bicinchoninic acid (BCA) method as explained in Meilland et al., (2016). Morphometric analyses on single foraminiferal tests were carried out at the University of Angers with an automated incident light microscope (Bollmann et al., 2004;Clayton et al., 2009) at a resolution of 1.4 μm 2 . Images were analysed for their two-dimensional (silhouette) morphometry (Beer et al., 2010), including minimum test diameter, which is the shortest distance from wall to wall that passes through the centre of the proloculus (the initial chamber of a foraminifer). Protein-to-size relations were determined for the 225 minimum diameter of each test providing size-normalized protein content (SNP) for data analyses. Foraminifera protein concentrations were linearly normalized to 1 μm minimum test diameter, being aware of any unavoidable errors related to nonlinear increments of biomass at volumetric test growth (cf. Beer et al., 2010).

Fossil planktonic foraminifera assemblages from core-tops (Multitube)
At 5 sites of the main CTD transect, an interface corer (Multitube type Oktopus GmbH, Institut National des Sciences de 230 l'Univers division of Brest, France) was used to obtain simultaneously 8 short sediment cores (less than 1 m in length) ( Figure   1; Table 1). At each station, the core with the more even and undisturbed water-sediment interface was selected. The uppermost 0.5 cm of the core was sampled and fixed with 95 % ethanol. Samples were stained with Rose Bengal, reacting with cytoplasm, to distinguish PF still bearing cytoplasm (fresh or in degradation) and thus very recently deposited from fossil PF without cytoplasm. Stained shells of foraminifera probably reflect the spring and summer population of the year of sampling, even 235 though the exact degradation time of cytoplasm is still poorly constrained (Schönfeld et al., 2013). The core-top sediments were wet sieved using a 100 µm mesh and the foraminifera were identified using the same SCOR WG138 taxonomy in order to be directly comparable to the plankton tow samples. Supprimé: thus we cannot be precise on the time-scale pointed out by Rose Bengal-stained specimens. Based on the hydrology, sites depth (Table 1), and sediment oxidation over the studied area, we can reasonably think that Rose Bengal coloration is in our situation highlighting spring and summer population that recently felt. The 280 discussion will be based on this assumption. ¶ For the purpose of this study, the core-top sediment has been wet sieved on a 100 µm mesh (same as the plankton net mesh size), and analysed for the planktonic foraminiferal assemblages. Every picked planktonic foraminifer has been identified consistently with those 285 collected by plankton tows. ¶ The NMDS analysis of species abundances with regard to environmental parameters (latitude of the station, temperature, salinity, Chl-a concentration) indicates that none of the species-specific distribution displays a significant correlation to any 310 of the tested variables (p-values > 0.1). NMDS documents distributional affinity ( Figure 5), with N. pachyderma and N. incompta plotting in the same area and T. quinqueloba and G. uvula plotting separately from each and also from the N. pachyderma/N. incompta area.

Planktonic foraminifera protein biomass
Individual protein content (BCA method) and associated test minimum diameter were measured for a total of 272 specimens 315 of the 4 major species, including 32 specimens of Neogloboquadrina pachyderma, 58 Neogloboquadrina incompta, 72 Globigerinita uvula and 110 Turborotalita quinqueloba. A 5 to 25 individuals per species were selected at each sampled depthinterval of the 7 stations along the S-N transect, paying careful attention to sample the whole size range of the populations. At station 7, the protein extraction was successful for only one specimen of N. pachyderma. Therefore, no value is display for this species at this station in Figure 6. 345 Minimum diameters of the 272 selected tests cover a large size range, from 65 to 315 µm with a median value of 160 µm. N.
incompta is the biggest species with a median of 200 µm, and G. uvula the smallest with a median of 110 µm (Table 2). For each studied species, the mean size is close enough to the median size to say that the size distribution of the picked tests is symmetric, thus making us confident that our test selection represents properly the natural test size range of each studied species. The biomass of a single individual normalized by its test size (SNP), averages out about 0.0055 µg of protein per µm 350 of foraminiferal shell diameter. It varies depending on species from 0.0004 (G. uvula) to 0.0426 µg.µm -1 (T. quinqueloba).
The SNP of all 4 species displays a northward decrease from 70 to 74° ( Figure 6). T. quiqueloba and G. uvula have slightly (but not significantly) higher relative protein concentrations than N. pachyderma and N. incompta ( Figure 6).

Planktonic foraminifera diversity and distribution in surface sediments
Concentrations of planktonic foraminifera with colourless empty tests varied from a maximum of 6200 ind.cm -3 at station 4 355 (71.3°N) to a minimum of 200 ind.cm -3 at the septentrional station 9 (Figure 7 a). Neogloboquadrina pachyderma was the most abundant species (31 to 59%) along the entire transect. Assemblages were more mixed at the two ends of the transect where N. pachyderma was less abundant. The assemblage at the southernmost point also contained Turborotalita quinqueloba (33%) and Neogloboquadrina incompta (24%). While at the northernmost point, station 9, T. quinqueloba (23%) co-occurred with Globigerinita uvula (25%). 360 Concentrations of planktonic foraminifera bearing a coloured cytoplasm (Figure 7 b) varied from 100 to 300 ind.cm -3 . All along the transect, the relative abundance of coloured N. pachyderma remained between 10 and 26 %. The species T. quinqueloba occurred everywhere above 20% and up to 40% South of 72°N. The central station 6 was dominated by G. uvula (38%). North of 74°, the fauna was balanced between N. incompta (33 and 9 %) and G. uvula (8 and 34%).

Distribution pattern of living planktonic foraminifera at the Barents Sea Opening
In the late summer of 2014 the hydrology at the BSO was characterised by a strong water stratification with a 30 m thick Chla enriched lens of NwCW that to the north overlapped with the NwAW (saltier and colder) from 69.8°N to 74.5°N. Further north, a well-mixed water column with characteristics of the NwAW occupied the Storfjordrenna Trough. Here a coccolithophore bloom and the highest concentration of cyanobacteria were recorded in the upper water column (Giraudeau 370 et al., 2016). Despite these marked features the pattern of planktonic foraminifera abundance did not correlate with any of the studied environmental parameters ( Figure 5). These observations confirm the low influence of commonly imputed parameters  (Giraudeau et al., 2016) Supprimé: global 8 such as temperature, salinity and primary production on PF density . In accordance with the conclusions 405 of Retailleau et al., (2018) conclusions, multiples indications however suggest a possible role of water turbidity in PF abundance variation. The highest densities of PF occurred in the 0-20m upper water layer between 70.5 and 74.5°N. Their very low abundances (total concentration < 5 ind.m -3 ) below 100 m depth suggest a shallow depth habitat for individuals in the region, especially for N. pachyderma which was recently reported to live between 25 and 280 m depth in the north Atlantic Arctic region (Greco et al., 2019). Very low abundances were also recorded nearby the Norwegian and Spitsbergen coasts. 410 The low abundances at the two ends of the studied transect could reflect planktonic foraminifera patchiness pattern of distribution (Meilland et al., 2019) or highlight the fact that waters under continental influences (nutrient-enriched, more turbid) likely hamper the foraminiferal production. In line with this, the abrupt decrease in abundances from West to East (stations 2, to 6, to 1) may be ascribed to the decrease in depth of the Bjømøyrenna Trough up to the Barents Sea shelf (from 1850 to 430 m), as foraminifera are suspected to avoid neritic waters over continental shelves (Schmuker, 2000). 415 The remarkable point of our results is the dominance of Globigerinita uvula. This species, described as a temperate to polar species (Schiebel and Hemleben, 2017), is known to account for less than 2% of the assemblages in marginal Arctic Seas based on material collected with a 63 µm plankton net mesh size (Volkmann, 2000). Neogloboquadrina pachyderma is considered the dominant species in polar regions, making up more than 90% of the total planktonic foraminifera assemblages (e.g. . The high densities of G. uvula recorded at the BSO in 2014 seem to contradict the former statements but are 420 consistent with a recent study reporting G. uvula as one of the dominant species in southern high latitudes, South of the Polar Front (Meilland et al., 2017). A possible explanation could be the warming experienced by the western Barents Sea (SST anomalies ≈ +2°C) and its increase in salinity (SSS anomalies ≈ +0.3) over the last decades (Dobrynin and Pohlmann, 2015).
These hydrological changes impact the plankton dynamics and biogeography, with a northward shift of the natural range of biological communities (Barton et al., 2016). Thus, the species distribution of planktonic foraminifera could be affected by an 425 eventual expansion of subpolar/temperate species towards high latitudes leading to phytoplankton composition changes, in response to sea temperature warming under global climate change. Our observations from the North Polar Region support the shift of planktonic foraminifera assemblages to warmer conditions already asserted from North Atlantic  and from the southern Indian Ocean data (Meilland et al., 2017). However, a single observational dataset is the Barents Sea is not sufficient to robustly validate this assumption and a second hypothesis for the dominance of G. uvula in our sampling area 430 could be a response to specific phytoplankton composition and ambient water conditions by pulsed reproduction events only in summer conditions. This seasonal pattern is known to occur in polar regions for Turborotalita quinqueloba (Schiebel and Hemleben, 2017). In fact, this species is the second dominant one in our late summer 2014 samples. As observed in this study, T. quinqueloba is also known to display high concentrations in the Barents Sea and western Spitsbergen (Volkmann 2000) and to co-occur with the typically polar species Neogloboquadrina pachyderma in the high-latitude cold-water assemblages 435 (Volkmann, 2000;Eynaud, 2011). Discrepancy between the species-specific distribution patterns was observed in late summer 2014 at the BSO. The low abundances of Neogloboquadrina pachyderma and Neogloboquadrina incompta consistent over the studied area versus the patchy distribution and high densities of Globigerinita uvula and Turborotalita quinqueloba, suggest differences in the 455 ecological strategy and behaviour between these two pairs of species. The patchy pattern of planktonic foraminifera distribution has been observed before (Boltovskoy, 1971;Siccha et al., 2012;Meilland et al., 2019) suggesting that high densities are not exclusively constrained by the physical structure of the (sub-) surface layers.
Potential differences in diet preferences could explain the observed species distribution in late summer 2014 at the BSO. Both G. uvula and T. quinqueloba are thought to follow food availability and primary production (Volkmann 2000, Schiebel and 460 Hemleben 2017). However, we observed no correlation between their distribution and Chl-a concentrations ( Figure 5). In late summer 2014, G. uvula and T. quinqueloba showed high concentrations especially at station 7, located at the crossroads of the Atlantic (NwAW) and Arctic waters flowing out of the Storfjordrenna (Figure 1), at the edge of the polar front (Oziel et al., 2017). From this location to the northern station, the concentration of phytoplankton was relatively low and the phytoplankton community showed singular characteristics, in comparison to the southern part of the transect: fuco-flagelattes became 465 dominant and diatom concentrations decreased. The fuco-flagelatte blooms (mainly Phaeocystis pouchetii in late summer 2014; Giraudeau et al., 2016) are well known to occur in the Barents Sea (Wassmann et al., 1990;Vaquer-Sunyer et al., 2013).
Our hypothesis is thus that high densities of G. uvula and T. quinqueloba are due to food composition (quality) rather than food concentrations (quantity). This also implies that satellite-derived chlorophyll concentrations, considered as potential indicator of algal bloom, may not always be good indicators to perceive foraminiferal concentration and distribution. 470

Planktonic foraminifera protein concentration, potential marker of their metabolism
Proteins are the main component of zooplankton biomass (Corg) in all oceanographic regions, from the tropics to polar areas (e.g., Percy and Fife 1981;Donnelly et al., 1994;Kumar et al., 2013;Yun et al., 2015). Their role is essential to organisms' growth and their concentration and composition likely reflect the environment individuals grew in and how well they adjust to it. Based on previous studies, the protein concentration of PF can be used as a proxy of its biomass (Corg) and foraminiferal 475 biomass should remain the same for a given test size (Schiebel and Movellan, 2012). However, in our study, the SNP (size normalized protein content) of PF decreases with higher latitude and hence with decreased in Chl-a concentration and temperature. The size of individuals picked for these analyses remains however constant. This observation suggests that foraminifera metabolisms (i.e. ability to consume/degrade food and to grow) is decreasing towards the north. This would be consistent with the observation of lower metabolism for zooplankton with decreasing salinity and temperature in the Arctic 480 (Alcaraz et al., 2010). Proteins are the main component of zooplankton biomass (Corg), closely followed by lipids. Lipids in zooplankton organisms are very variable geographically, showing a latitudinal pattern with high percentages in polar areas and low percentages in warm tropical waters. Lipids percentages also display seasonal features, with higher values in summer than in winter (Falk-Petersen et al., 1999;Mayzaud et al., 2011;Kumar et al., 2013). It is thus possible that a part of energy (biomass also by observations made on pteropods in the Arctic (Kattner et al., 1998;Phleger et al., 2001;Böer et al., 2005). The fact that higher SNP of foraminifera were observed where Chl-a is higher is compatible with the fact that polar organisms rely on their protein catabolism when food is easily accessible rather than on their lipid storage (Brockington & Clarke 2001). It has also been shown that a single organism in a cold environment is able to switch between predominantly protein or lipid 520 catabolism across its life (Mayzaud, 1976). This suggests that individuals from the same species can display more or less proteins for the same biomass in different locations. With the reduction of PF protein concentration (and likely metabolism) going north, one could expect lower abundances. However, we observe no link between PF concentrations, which appeared to be species specific, and protein concentrations (evolving similarly for all four species) suggesting a decoupling between individuals metabolism and densities. 525

Discrepancy between upper water column and interface sediment samples
The PF species compositions recorded during the late summer 2014 in the water column and in surface sediments are similar while species relative abundances are drastically different. Indeed, the living fauna (collected by plankton net) displays large relative and absolute abundances of the two species Globigerinita uvula and Turborotalita quinqueloba whereas the fossil 530 assemblages (found in core-tops) are largely dominated by Neogloboquadrina pachyderma or, at the southernmost station, codominated by T. quinqueloba and N. pachyderma (Figure 7 a). Affected by differential settling velocities (200 and 500 m.day −1 in normal conditions), water depth and test sizes of different species, the foraminiferal fluxes exported from the upper productive surface and reaching the sea bottom depend on direction and intensity of currents (Takahashi and Bé, 1984). Lateral advection may transport shells over long distances > 25 km for N. pachyderma and > 50 km for T. quinqueloba, respectively 535 (Von Gyldenfeldt et al., 2000). Lateral advection of shells is also strengthened by water stratification that increases resident time at the shear boundary between water masses (Kuhnt at al. 2013). The BSO has a complex hydrography with the buoyant NwCW flowing northwards above the NwAW that is entering eastwards the Barents Sea when cold BSAW is flowing westwards. In such areas, the PF settling velocities and extension of lateral advection poorly constrained. Consequently, the sediment core records cannot match exactly with the place and/or the intensity of production (Von Gyldenfeldt et al., 2000;540 Jonkers et al., 2015). However, considering potential lateral advection of shells, Pados and Spielhagen, (2014) observed from a study through the dynamic Fram Strait that the distribution pattern obtained by plankton tows was clearly reflected on the sediment surface, and that the assemblage on the sediment surface can be used as an indicator for modern planktonic foraminiferal fauna. This suggests that the large discrepancy between upper water column and interface sediment samples collected at the BSO in late summer 2014 should be taken into consideration. As a sedimentation rate of 1.3 ± 0.6 mm.yr -1 has 545 been recently measured in the Storfjordrenna outlet [76°N-17°E], close to our station 9 (Fossile et al., 2019), the core-top sediments may have recorded less than a decade. Even though sedimentation rates are likely to vary along the transect, we hypothesise that the sea surface-bottom differences in the foraminiferal assemblages along the South-to-North transect at the BSO might reflect a community change within a short period of time (about a decade).

565
Furthermore, the analysis of sediment from the 5 core tops demonstrated important differences between the assemblages of fossil fauna and recently settled tests (likely coming from surface Spring/Summer production), i.e. Rose-Bengal stained tests bearing not yet decomposed cytoplasm. For example, at 71.3°N, the percentages of coloured T. quinqueloba and G. uvula are twice as high as the ones observed for the fossil faunas (Figure 7). At 72.9°N in the surface sediment, G. uvula reaches up to 38% of the coloured assemblages (Figure 7 b) whereas it never exceeds 25% in the non-coloured ones (Figure 7 a). The high 570 abundance of the two species G. uvula and T. quinqueloba in the living fauna as well as in the recently settled shells, but not in the fossil faunas suggest that they may present a seasonal character with a production period focussed in late summer as a response to environmental and trophic conditions. This is supported by previous studies in the Arctic where T. quinqueloba has been found to dominate assemblages sampled in August (Carstens et al., 1997;Volkmann, 2000) but not in June/ early July (Pados and Spielhagen, 2014), and by sediment trap observations from the subpolar North Atlantic where T. quinqueloba 575 reaches its maximum in autumn (Jonkers et al., 2010). The dominance of N. pachyderma in the fossil faunas collected at the BSO and its low but constant presence in the coloured shells of surface sediment and plankton tow sampled in late summer 2014 suggests that this species may demonstrate a regular production throughout the year. Neogloboquadrina pachyderma production appears to be yearly sustained and constant in the area whereas other species clearly respond to a local seasonal signal. 580

Conclusion
The sampling and analytic approaches deployed during the MOCOSED14 cruise and combining the use of plankton net, coretop, molecular biology (protein measurement), environmental parameters and phytoplankton characterisation provides us with a unique dataset to better constrain the distribution of planktonic foraminifera within the highly complex studied area of the western Barents Sea. 585 The observed abundances of PF in the studied area are high offshore and the lower densities were recorded nearby the Norwegian and Spitsbergen coasts. These observations highlight the fact that waters under continental influences (nutrientenriched, more turbid) are rather inhospitable for PF production. The PF species composition observed at the BSO is diverse, with more than 10 different species in the net samples including Globigerinita uvula (45%), Turborotalita quinqueloba (26.2%), Neogloboquadrina incompta (15%) and Neogloboquadrina pachyderma (8.9%). The two species G. uvula and T. 590 quinqueloba dominate the living (water sample) population and display highly patchy abundances suggesting they occur in late summer in response to physico-chemical conditions and related specific primary productivity. The dominance of G. uvula in water samples could also be a signal of the temperature increase experienced over the last decades in the Barents Sea and the North Atlantic Ocean. Further sampling in the area is thus needed to test this hypothesis. The species N. pachyderma and N. incompta show low densities but a continuous distribution pattern in the water samples. They also dominate the core-top assemblages suggesting that both species present a more consistent production over the course of spring-summer season. 620 Unlike their species-specific abundances pattern of distribution, size-normalized protein concentrations of all four major species decrease with the increasing latitude (and a decrease in temperature and Chl-a concentration). This observation leads us to hypothesise that 1) PF abundance and metabolism are decoupled and 2) foraminifera metabolism in the North of the studied region is lower than in the South. It opens the following question: Can individuals of the same species balance the ratio between their protein and lipid concentrations (major components of zooplankton Corg) in order to adapt to environmental 625 conditions (e.g. temperature)? Further analyses on planktonic foraminifera lipid concentration and composition are thus needed and would help us to better understand the metabolism of these organisms and their fate in a context of climate change.

Data availability
Data will be made available on request to the main author until their online publication on PANGAEA (https://pangaea.de/). 20 Table 1: Location (Latitude and Longitude), sampling date and water depth, of the 9 MultiNet and 5 Multitube stations, incremented from South to North (stations 1 and 2 being positioned aside the main transect mid-point #6). Stations where phytoplankton analyses were performed are also indicated.