Unprecedented sea-ice minima enhances algal production deposited at the Arctic seafloor

Sea-ice in the Arctic is declining, with 2018 a particularly low year for ice extent, driven by anomalously warm atmospheric circulation in winter 2017/18. This is consistent with a multi-decadal trend to an earlier ice-free Barents Sea as climate change rapidly warms the Arctic. Here we investigate a N–S transect in the Barents Sea, crossing the Polar Front from Atlantic waters in the south to Arctic waters in the north, focusing on the organic geochemical signature (pigments and lipids) in surface sediments sampled in summer, between the years of 2017–19. Early ice-out in summer 2018 was confirmed by satellite imagery, tracking the evolution of Arctic sea-ice extent between years. Consistent with less extensive sea-ice cover in 2018 we found increases in multiple chlorophyll and carotenoid pigments as well as fatty acids (reflecting recent phytoplankton delivery) in the northern part of our transect at the seafloor. We attribute this to nutrient and organic matter release from earlier 2018 ice-out leading to stratification, post-melt phytoplankton blooms and the deposition of organic matter to the seafloor, evidenced by pigments and lipids. Organic matter delivered to the seafloor in 2018 was reactive and highly labile, confirming its deposition in the most recent season, pointing to rapid deposition. Correlations were found during ice-free periods between satellite-derived chlorophyll a and multiple indicators of water column productivity deposited at the seafloor. We also found convincing evidence of multi-year biogeochemical change across the Polar Front, where sedimentary change is marked by chlorophyll degradation products providing evidence of grazing, indicative of a tightly coupled ecosystem close to the marginal ice zone. Overall, our results show the tight coupling of Arctic productivity with the delivery and quality of organic matter to the seafloor and how this varies across the Barents Sea. More frequent early summer sea-ice loss driven by climate warming in the Barents Sea will have consequences for the delivery of organic matter to the seafloor with impacts for benthic organisms, microbiology and the sequestration of carbon.


Introduction
Climate change is leading to a longer ice-free period and earlier onset of sea-ice melt in the Arctic (Stroeve et al 2012, Notz andStroeve 2016).This has consequences for ocean stratification, currents and temperature, which can shift the timing and abundance of primary productivity, altering the carbon and nutrient sources available to pelagic and benthic organisms (Ke ˛dra et al 2015, Post 2017).Sea-ice has a critical role in the regulation of the climate system and covers between 5% and 8% of the ocean surface (Comiso 2003), moderating feedback as ice has higher albedo than water (Wadhams 2000) and it helps maintain deep-water formation, vital to retain stratification in polar oceans (Bretones et al 2022).Since ∼1980 increased inflow of Atlantic waters to the Barents Sea has contributed to warming of 1.4 • C (Stocker et al 2014), stimulating phytoplankton (Neukermans et al 2018).This is likely to be accelerated; modelling suggests the Barents Sea will be ice-free by 2050 with up to 4 • C of summer warming (Smedsrud et al 2013).
Satellite imagery has been used to detect and quantify reductions in Arctic sea-ice over decadal and sub-annual timescales (Serreze et al 2003, Comiso et al 2008), with sea-ice presence datasets available, based on satellite imagery at daily timescales (Spreen et al 2008).However, by solely looking at ice cover from space, it is difficult to ascertain the extent and timing of primary productivity changes taking place on the ice and in the water column and the effects of organic material on benthic productivity.Algal material and phytodetritus are key food sources for benthic animals (Boetius et al 2013) and microbial activity at the seafloor (Ravenschlag et al 1999, Orlova et al 2015), which ultimately mean that Arctic ocean sediments act as a carbon sink (Faust et al 2020, Stevenson et al 2020).
Phytoplankton and phototrophic bacteria are key sources of pigments to Arctic ocean sediments and have been used to understand phytoplankton communities, changes in primary production and alterations to the food web (Morata and Renaud 2008).The analysis of fatty acids and n-alkanes have been used to assess the quality of Arctic sedimentary organic carbon linked to marine productivity (Schubert and Stein 1997).Sterols are effective indicators of phytoplankton; the susceptibility of these compounds to degradation aids understanding of diagenetic processes (Volkman 1986, Belicka et al 2002, Belt et al 2015).Together, these markers can help assess the source of organic matter, as well as its abundance, quality and composition at the time of sampling.
Summer 2018 was a low year for Arctic sea-ice extent (Sumata et al 2022), with the Barents Sea icefree earlier than usual (Downes et al 2021, Sumata et al 2022).To the west of the Barents Sea, there was an unprecedented decline in Arctic sea-ice outflow in 2018 of less than 40% of the annual mean (average September to August) between 2000 and 2017, a result of anomalously warm atmospheric circulation in winter 2017/2018 (Sumata et al 2022).Other parts of the Arctic had similar trends, with very little winter ice formation in the Bering Sea in 2018 (Stabeno andBell 2019, Siddon et al 2020).In the Barents Sea, oceanographic observational data confirmed summer 2018 as being ice-free (Downes et al 2021), with diluted Atlantic water found close to the surface during Autumn 2018, following a long ice-free period (Lundesgaard et al 2022).This trend is consistent with the general trend towards an earlier ice-free Barents Sea which is becoming more frequent (Onarheim and Årthun 2017).
The Barents Sea is also influenced by varying Arctic and Atlantic currents which structure the supply of nutrients, phytoplankton and dependent consumers in the water column and at the benthos.
The transect investigated in this study crosses the Polar Front where warmer Atlantic water from the southwest, meets cooler Arctic water from the north, maintaining water column stratification (Barton et al 2018).Although one source of organic matter to the water-column is direct from sea-ice algae when it melts, also of importance are ice-edge phytoplankton, which take advantage of stable nutrient-rich conditions adjacent to ice due to stratification (Syvertsen 1991, Boetius et al 2013).Under-ice phytoplankton blooms where ice thins and meltwater ponds develop enabling light penetration have also been noted (Arrigo et al 2012).Water-column phytoplankton blooms occur shortly after ice-out, typically linked with its timing (Strass andNöthig 1996, Dong et al 2020).Here, we use measurements of pigments, fatty acids, sterols, n-alkanols and n-alkanes in surface sediments (0.5 cm depth), reflecting most recent deposits and benthos from a N-S transect in the Barents Sea, replicated in the summers of 2017, 2018 and 2019, investigating the effect of the 2018 unprecedented sea-ice minima on the quality and quantity of phototrophic derived carbon to the seafloor.Additionally, a N-S transect across the Polar Front provides the opportunity to explore phototrophic indicators of transition to help distinguish the signatures of Arctic versus Atlantic water productivity.

Study locations and sampling
Sediment cores were retrieved in the months of July and early August in the Barents Sea aboard the RRS James Clark Ross in 2017 (JR16006), 2018 (JR17007) and 2019 (JR18006), from seven stations to the east and south of Svalbard.Most of the sampled area (except in the vicinity of B03 and B13, table 1) is seasonally ice covered (maximum = April; minima = September) (Vinje and Kvambekk 1991).Although most of the ice is locally formed, there is also seasonally variable multi-year drift ice from the Kara and Laptev Seas (Pavlov et al 2004, Hop and Pavlova 2008, Tamelander et al 2009).Station B14 marks the mean position of the Polar Front which inflexes north due to the effect of Atlantic currents (Fossheim et al 2006, Oziel et al 2016).Sampling sites north of B14 mainly feature cooler Arctic water overlain by seasonally variable less saline meltwater.Warmer more saline Atlantic water enters the Barents Sea from both the north (after circulating clockwise around Svalbard with the West Spitsbergen Current) and from the south (Loeng 1991, Sundfjord et al 2007).
Stations B13 to B17 reflect the primary 30 • E longitudinal transect across the continental shelf (average depth 287-359 m below sea level (mbsl)), with station B18 included as a deeper continental shelf-edge comparison (average depth 3001 mbsl).
Sediments from the southern Barents Sea at station B03 (72.6 • N, 19.3 • E; average depth 367 mbsl) enable comparison with Atlantic waters.Replicate measurements (displayed as standard deviations to indicate reproducibility) were obtained each year at stations through deployment of a multicorer three times within the same local area (within 50 × 50 m grid), except where ice floes prevented redeployment within the gird, extending grid area slightly.Due to ice thickness B18 was only sampled in 2017 and B17 in 2017/18.Cores were sectioned using a stainless-steel spatula into foil (pre-ashed) and frozen at −80 • C. Pigment samples remained frozen at −80 • C until analysis, biomarker samples were stored at −20 • C. Surface sediments for this study comprise the upper 0.5 cm of each core sampled.

Satellite imagery of sea-ice cover (SIC) and chlorophyll a
SIC was obtained from Pan-Arctic AMSR2 seaice concentration data at 3.125 km grid resolution (2017-2019) from Universität Bremen (https://seaice.uni-bremen.de/sea-ice-concentration)(Spreen et al 2008) based on publicly available data from GCOM-W1 satellite (https://gportal.jaxa.jp/gpr).Images were obtained to show seasonal sea-ice retreat from April to July across the Barents Sea (SI figure 1).SIC closest to sampling (table 1) is displayed in figure 1.

Chlorophyll and carotenoid pigments
Surface sediments (2-3 g) were thawed and centrifuged (Eppendorf, 4000 rpm, 5 min, 4 Chlorophyll and its degradation products were quantified according to response factors from standards (DHI, Denmark).For quantification of hydroxychlorophyllone (no standard available), the response factor for pheophorbide a was used as the two components have similar UV/vis absorbance spectra (Airs et al 2001).Selected sediment samples were analysed by LC/MS n for assignment of components (Airs et al 2001).LC/MS n was performed using an Agilent 6330 ion trap mass spectrometer via an atmospheric pressure chemical ionisation source (m/z 400-1100).Post column addition of acid was used to aid ionisation of metallated components (Airs and Keely 2000).Components were assigned based on relative retention time, on-line UV/vis spectra, protonated molecule and fragmentation data (Airs et al 2001).Pigments are expressed as µg g −1 TOC and listed with key ratios in SI table 1.Where possible, samples were analysed in triplicate at each station from separate multicore deployments (Error bars: ±1 SD).Measurements of pigments on surface sediments (0.5 cm) reflect primarily the most recent seasons deposition as the half-life of many labile pigments is approximately 3 weeks in Barents Sea sediments (Morata and Renaud 2008).
1 µl aliquots in DCM were injected onto a 7890B gas chromatograph (GC) coupled to a 5977B MSD mass spectrometer (MS) (Agilent, Santa Clara, CA, USA) operated in full scan mode (70 eV; source temp 230 • C; helium flow 1 ml min).A 60 m HP1-ms fused silica capillary column (0.25 mm × 0.25 µm) (J&W Scientific, Folsom, CA, USA) was used for separation with a gradient of 6 • C min −1 to 17 • C, followed by a hold of 1 min and then a slower ramp of 2.5 • C min −1 to 315 • C and a final hold of 22 min.Retention time and mass spectra were used to identify organic compounds and quantification was relative to the peak area of 5α-androstane (internal standard, assumption of linear response).Error estimations are the result of relative percentage standard deviation calculations of a sample analysed in triplicate, applied on a percentage basis.Selected fatty acid methyl esters (FAMEs), n-alkanols, n-alkanes and sterols and ratios used in the study are listed in SI table 1.

Comparing sea-ice and satellite-derived chlorophyll a cover in the Barents Sea between 2017, 2018 and 2019
Summer SIC was less extensive in July 2018 compared with July 2017 and 2019 (figure 1).In July 2017 there was a broken array of moderate-low concentration SIC between station B15 and B16, with higher concentration SIC (up to 90%) in the vicinity of stations B17 and B18 (figure 1(a)).In contrast, in July 2018 the sea-ice margin was north of B18 (82 • N) and ranged between 60% and 90% SIC (figure 1(b)).Thickest and most extensive SIC was in July 2019 when 90% SIC approached station B15, extending eastwards towards Svalbard and northwards to the Nansen Basin (figure 1(c)).
In 2017 SIC evolved from covering as far south as station B14 at 31 May 2017, with only a broken array of moderate-thin SIC present between B15 and B16 by 31 July 2017 (SI figure 1(a)), evidencing a northwards retreat scenario (table 1).In 2018 SIC extended as far as B14 at 30 April 2018.However, due to low SIC north and north-east of Svalbard and associated weather conditions, SIC was less fixed and became more mobile enabling extensive break-up by 31 May 2018, except for localised cover in the proximity of B16 (SI figure 1(b)).Subsequently, by 30 June 2018 all stations were free of SIC, meaning all sampling was conducted in ice-free conditions, as confirmed by SIC at 21 July 2018 (SI figure 1(b); table 1).In 2019 SIC retreat was less dynamic with extensive thick (>90%) extent as far south as B15 at 31 May 2019, which despite thinning (∼60%-80%) between B14 and B15 by 30 June 2019, became more consolidated north of B15 by 14 July 2019 due to weather conditions (SI figure 1(c)).Therefore, 2019 had extensive SIC during sampling (table 1).
Satellite-derived chlorophyll a during ice-free periods was noticeably higher in concentration and earlier in timing in the Barents Sea during 2018, spanning the full transect by May/June compared with only partial cover in 2017 & 2019.

Indicators of algal production
Algal production was marked by consistently higher concentrations of pigments in 2018 at stations B16 and B17 when the northern Barents Sea was ice-free.Total chlorins were higher at B16 in 2018 (885 µg g −1 TOC) compared with five-and eight-fold lower amounts in 2017 and 2019 (185 and 109 µg g −1 TOC respectively) (figure 2(a)).At B17 chlorins were on average 20-fold higher in 2018 compared with 2017 (average of 5119 µg g −1 TOC, compared with 253 µg g −1 TOC respectively) (figure 2 There was a marked but smaller secondary maxima in total chlorins around station B14 (averages 800 µg g −1 TOC in 2017, 762 µg g −1 TOC in 2018, 962 µg g −1 TOC in 2019, figure 2(a)), in proximity with the mean position of the Polar Front (Fossheim et al 2006, Oziel et al 2016), which were consistently between stations B14 and B15 during sampling (SI figure 18).Similar trends were observed in chlorophyll a (figure 2(b)), fucoxanthin (figure 2(d)) and additional indicators in SI figure 2(a)-(c).The general position of the benthic Polar Front providing evidence of water mass separation between stations B14 and B15 was confirmed to be stable from bottomwater temperature versus salinity plots across all years (SI figure 18).Bottom water temperature at B15 in 2018 was slightly higher than 2017/19, when the Barents Sea was seasonally ice-free earlier than usual.

Indicators of fresh algal material
Both total FAMEs and summed short-chain (C 14 , C 16 and C 18 ) FAMEs were markedly higher in 2018 ice-free season compared with 2017 and 2019 iceabundant seasons at stations B15-B17 (figures 3(a) and (b)), with 2018 concentrations rising consistently from station B03 to B17.Short chain FAMEs at B17 were more than four-fold higher in 2018 (average 7683 µg g −1 TOC) compared with 2017 (average 1681 µg g −1 TOC).Unaltered chlorophyll a ((chlorophyll a/identified chlorins) * 100) was also higher in 2018 reaching 43% at station B16 (compared with 14% in 2017 and 2019) and 60% at station B17 (compared with 25% in 2019) (figure 3(c)).The ratio of sterols campesterol/campestanol was also significantly higher in 2018 at stations B16 and B17, with a 15-fold difference at B17 between 2018 (ratio of 35) and 2019 (ratio of 2.3) (figure 3(d)).High fatty acid n-C 14/16 ratios were present in 2018 at stations B15 to B17 compared with 2017 and 2019 sampling (figure 3(e)), whereas the autochthonous/allochthonous ratio of FAMEs was most clearly elevated in 2018 at station B17 (ratio 27) compared with 2019 (ratio of 12) (figure 3(f)).There were marked secondary increases in 2019 at station B14 in total FAMEs, short chain (C 14 , C 16 and C 18 ) FAMEs and % unaltered chlorophyll a (figures 3(a)-(c)) close to the Polar Front (Fossheim et al 2006), which was between B14 and B15 during sampling (SI figure 18).Complicating trends, in 2019 the ratio of campesterol/campestanol was highest at B03 and the fatty acid n-C 14/16 ratio was highest at B13 (figures 3(d) and (e)).

Indicators of transition across the Polar Front
A number of pigments and alteration products were especially good indicators of changes in primary producer community and variations in degradative susceptibility across the Polar Front, evidenced between B14 and B15 (Fossheim et al 2006) and from consistent benthic observations (salinity and temperature, SI figure 18).The carotenoid 19'hexanoyloxyfucoxanthin was markedly higher in all three years at station B14 with averages ranging between 120 and 153 µg g −1 TOC, an ∼50-fold increase from minima at B15 (∼3 µg g −1 TOC) (figure 4(a)).Similarly, total carotenoids were also elevated (up to an average of 387 µg g −1 TOC) in 2017 at B14 (figure 4

Indicators of Arctic versus Atlantic water productivity
Differences between the southern-most stations and the main transect (B03 and B13) suggest differences in quality of organic matter in regions of Arctic versus Atlantic water productivity.Total sterols were six-fold higher at B03 compared with B13 in 2019 (6075 and 1011 µg g −1 TOC respectively), but in 2017 and 2018 there was no major difference (figure 5(a)).For the ratio of C 14 , C 15 and C 16 iso FAMEs relative to n-C 16:0 FAME, values were higher in 2017 and 2018 in both B03 and B13 compared with B14-B17 but in 2019 it was only higher at B03 (figure 5(b)).C 14 , C 15 and C 16 iso FAMEs relative to n-C 16:0 FAME was also markedly higher at the deeper shelf edge B18 site (when measured in 2018), in proximity to the northerly Atlantic influence around Svalbard.The ratio of dinosterol:dinostanol was significantly higher in 2017 at both B03 and B13, compared with B14-B18, but markedly higher only at B03 in 2019 (figure 5(c)).Stigmasterol:stigmastanol was higher at B03 in all years compared with the B13 to B18 transect (figure 5(d)).

Associations between indicators, samples and satellite-derived chlorophyll-a
Differences in 2018 ice-free sampling season at B16 and B17 highlighting high material quantity and quality are reflected in the upper-right part of the biplot (figure 6) and include total FAMEs, campesterol/campestanol, short-chain FAMEs, cholesterol/brassicasterol, % unidentified carotenoids, CPI T FAMEs, FA short/long, fatty acid autochthonous/allochthonous, hydroxychlorophyll a/chlorophyll a, total SCEs and OEP 17-21 n-alkanes (figure 6).In contrast, in the opposite part of the biplot variables are present which highlight differences between the southern part of the transect B03 and B13 compared with the rest of the transect (e.g.C 14 , C 15 , C 16 iso/16:0 FAMEs and dinosterol/dinostanol) and complex relationships across the transect (e.g.pyropheophytin a/chlorophyll a and % unidentified chlorins).Variables such as short chain n-alkanols and total sterols are positioned midway between two groups and reflect variability at both parts of the transect.
Chlorophyll a concentration extracted from satellite imagery was markedly higher surrounding stations B17 and B18 in 2018 at a 25 km 2 resolution and higher also in 2018 surrounding stations B16, B17 and B18 at a 10 km 2 resolution (SI figure 6 & 7).There were good associations between satellitederived chlorophyll a and in 2018 with chlorophyll a, diadinoxanthin, total chlorins, fucoxanthin, % unaltered chlorophyll a and short-chain FAMEs (SI figure 8).

Discussion
The markedly less extensive SIC in July 2018 (compared with 2017 & 2019) was contemporaneous with increases in multiple chlorophyll and carotenoid pigments at stations B16 and B17, indicating higher levels of photosynthetic production, characteristic of an extensive phytoplankton bloom.Mechanistically, this can be attributed to earlier nutrient release from ice-out and stratification, enabling extensive phytoplankton blooms and subsequent rapid vertical flux of chlorophylls and carotenoids to the seafloor.Shifts in the timing of primary productivity pulses may also generate a mismatch between pelagic grazers (e.g.copepods such as Calanus spp.) and their food sources (Ji et al 2013).Changes in surface sediment geochemistry, predominantly from fatty acids, sterols and pigment ratios (e.g. % unaltered chlorophyll a) at the same stations in 2018 provide insight into the delivery of fresh labile (more reactive) material to the seafloor in a summer of minimal SIC in the Barents Sea.Underlying the predominant 2018 increased production signature at the northerly    Our observations are also consistent with phytoplankton extent, sea-ice and temperature trends in the northern Barents Sea which show more productive phytoplankton blooms in warm years (Dong et al 2020).Earlier ice retreat in the northern part of the Barents Sea in 2018 (see figure 1 and SI figure 1) may also reduce grazing pressure on phytoplankton resulting in greater vertical flux of pigments (and resulting carbon) to the seafloor (Renaud et al 2007), supported by a low proportion of steryl chlorin esters in 2018 (SI figure 5) and a high proportion of unaltered chlorophyll (figure 3(c)).Timing of the ice algal biomass peak in 2018 may also have been earlier (Ji et al 2013), although this may only partially explain the high algal production at B16 and B17 as sea-ice is known to typically contribute ∼20% of total primary production in the marginal sea-ice zone of the Barents Sea (Tamelander et al 2009).
The detection of multiple pigments at the sediment surface in 2018 at northerly stations B16 and B17 suggests a wide algal response with total chlorins, chlorophyll a and its early derivative hydroxychlorophyll a (Walker et al 2002) all following similar increasing trends (figure 2).Pronounced increases in fucoxanthin and diadinoxanthin point to siliceous diatoms which are especially abundant in both underice phytoplankton blooms and at the ice-edge margin (Ardyna et al 2014).We know that the surface sediments deposited in summer 2018 are mostly related to recent events as the half-life of some labile pigments is approximately 3 weeks in polar sediments (Morata and Renaud 2008).In the marginal ice zone (north of station B15) due to the likely tight pelagic to benthic coupling (Tamelander et al 2006) we posit that around this location advection will be minimal compared with relatively local or sub-regional scale sedimentation of algal particles, which is supported by the Barents Sea being dominated by planktonic algae with small cells (Wassmann et al 2006).Additionally, it is known benthic sedimentary mixing is relatively slow and shallow from excellently preserved sediment cores in the region (Stevenson et al 2020, Faust et al 2021) and surface sediment radiocarbon dates on foraminifera have post-bomb 14 C ages (Faust et al 2023).

Evidence of highly reactive algal derived labile organic matter
Lipid signatures highlighted that delivery of organic matter in 2018, especially at stations B16 and B17 was fresh and highly labile (figure 3).Fatty acids are labile and susceptible to diagenesis, and so their higher abundance at the surface sediment under conditions of early ice retreat and bloom, support delivery of fresh phytoplankton material to the sediment surface.Fatty acids in sediments of the central Arctic have been linked to marine productivity (Schubert and Stein 1997) and are known to be high in both ice algae and phytoplankton in the Barents Sea marginal ice zone (Falk-Petersen et al 1998).High short-chain fatty acids could also result from the degraded remains of Calanus sp. which feed on phytoplankton blooms in the marginal ice zone of the Barents Sea (Falk-Petersen et al 2009).High percentage of unaltered chlorophyll a in 2018 also points to freshly deposited pigments, rather than degradation products (Furlong and Carpenter 1988).Additionally, high campesterol (relative to campestanol) at B16 and B17 in 2018 point to fresh delivery of diatom remains, dinoflagellates and green algae (Volkman 1986, Volkman et al 2008) corroborating deposition of recent material.Direct, rapid deposition of algal biomass such as the filamentous Melosira arctica from melting sea-ice could also contribute to the highly labile signature, especially during initial ice-out (Boetius et al 2013).Together, this evidence highlights the importance of labile chloropigment and organic matter fluxes to the seafloor between and including the end of ice algal production and the open water phytoplankton bloom (Fortier et al 2002).Earlier 2018 sea-ice loss in the Barents Sea will have also altered the timing of the spring bloom, with potentially cascading effects on mid-summer production (Gaffey et al 2022), likely bringing forward deposition of the most labile bloom closer to the timing of sampling.

Biogeochemical change across the Polar Front
Increases in multiple markers around station B14 point to evidence of a distinct region of productivity and changing organic matter composition deposited across the Polar Front where Arctic and Atlantic water masses converge (figure 4).Changes in organic matter across the Polar Front in the Barents Sea have previously been evidenced by changes in the signature of sedimentary pyrolysis products, pointing to distinct differences in the composition of the source of OM (e.g.fish, phytoplankton, ice-age, zooplankton) (Stevenson and Abbott 2019) and by contrasting nutrient conditions and phytoplankton community composition (Downes et al 2021).The transition across the Polar Front is also marked by transition in the bioturbation activity of benthic animals (Solan et al 2020) and their reproductive and population dynamics (Reed et al 2021), with potential consequences for the character and degradation of organic matter at the seafloor.The Polar Front is a key boundary in the spring between highly productive Arctic and Atlantic waters with comparatively lower phytoplankton abundance (Makarevich et al 2021).High hydroxychlorophyllone, a chlorophyll transformation product at station B14 suggests that phytoplankton organic matter delivery is relatively recent, fresh and well-preserved and could additionally point to protist grazing activity (Kashiyama et al 2012, Tait et al 2015) and therefore, the presence of a productive and tightly-coupled pelagic food-chain.High 19'-hexanoyloxyfucoxanthin at B14 point to the presence of prymnesiophytes (Haberman et al 2003), a class of haptophyte algae and some dinoflagellates (Tangen and Björnland 1981), reinforcing the Polar Front as a zone of high productivity.
Higher productivity in 2019 at stations B13 and B14 (figures 2 and 3) close to the Polar Front also point to a larger bloom further south in the vicinity of the zone where Atlantic and Arctic water masses converge (Oziel et al 2016).In 2019 advection of phytoplankton cells from the Atlantic sector of the Barents Sea could have been higher, concentrating larger cells (which are more easily advected (Wassmann et al 2006, Font-Muñoz et al 2017)) around this area, stimulating production.This highlights heterogeneity in spatial and temporal variability between years.

Tracing Atlantic versus Arctic water productivity at the periphery of the Barents Sea
Geochemical markers also highlight the secondary signature of peripheral contrasting Atlantic productivity among sampled stations.For example, higher total sterols at the westerly station B03 in 2019 point to a pulse of productivity entering the shelf from the Atlantic, consistent with the direction of the North Atlantic Current into the Barents Sea, a phenomenon which has been previously associated with warming (Neukermans et al 2018) (figure 5(a)).Higher ratios of the microbial C 14 , C 15 and C 16 iso FAMEs relative to the saturated C 16:0 FAME at stations B03 and in 2017/18 at B13 point to enhanced sedimentary bacterial activity (Perry et al 1979), probably reflecting greater degradation in highly reactive sediments enabled by warmer bottom temperatures and/or differences in the delivery and quality of phytoplankton to the seafloor (figure 5

Implications in a changing Arctic
With a long-term trend to reduced summer seaice extent in the Arctic, together with an increasing likelihood of more frequent early ice-out (as in 2018 (Sumata et al 2022)), the future delivery of organic matter to the seafloor is likely to change, with earlier and more productive phytoplankton blooms.This will have consequences both for the benthic animals which feed on detritus at the seafloor (Solan et al 2020) and on the microbiology and resulting organic matter processing in sediments beneath (Stevenson et al 2020).This could result in increased feeding in areas of the seafloor beneath early ice-out scenarios, but reduced feeding in areas no longer proximal to ice-out scenarios, disrupting previously anticipated Arctic benthic animal phenology (Górska and Włodarska-Kowalczuk 2017).However, as the zone of most-labile organic matter moves northwards, southern sectors of the Barents Sea, areas close to the Polar Front will likely receive less phytoplankton-derived organic matter and of a less nutritious quality, impacting benthic animal activities, sedimentary microbial decomposition of organic matter and reductions in carbon storage at the seafloor.

Conclusions
Consistent with the wider recorded decline in Arctic sea-ice in 2018, we find marked increases in chlorophyll and carotenoid pigments and fatty acids deposited at the seafloor in the northern part of our transect in the Barents Sea, in contrast to more conventional deposition scenarios in 2017 and 2019.Importantly, our analyses indicate earlier ice-out in 2018 enabled conditions supportive of deposition of labile algal-rich organic matter to the seafloor, facilitated by post-melt phytoplankton blooms or deposition of degraded ice-attached algae directly to the sediment-water interface.The labile, highly reactive signature of this material deposited to surface sediments in 2018, confirm its recent seasonal deposition.Close relationships between satellite derived chlorophyll a and multiple indicators of productivity highlight the links between post-ice out pelagic algal blooms and deposition of this organic matter to the seafloor.Additionally, this study found evidence of a differing organic matter signature in the vicinity (both north and south) of the frequently observed Polar Front system in the Barents Sea, plus secondary evidence of Atlantic water mass influence in the most southerly and westerly sampling stations.Continued retreat of Arctic sea-ice, exemplified by early ice-out scenarios such as in 2018 driven by warming, together with gradual movement of the Polar Front northwards will alter the quality, abundance and position of organic matter delivery to the seafloor.This has clear consequences on benthic biology, microbial degradation of organic matter and likely changes in carbon storage at the seafloor.

Data availability statements
Sea ice cover maps (figure 1 (a)).Trends were reflected by changes in key pigments such as chlorophyll a, hydroxychlorophyll a and fucoxanthin (figures 2(b)-(d)) and additional indicators in SI figures 2(a)-(c).

Figure 1 .
Figure 1.Sea ice cover in the Barents Sea derived from satellite imagery for scenarios in July 2017 (a), 2018 (b) and 2019 (c).Mean position of the Polar Front shown in dotted lines (Fossheim et al 2006).
(b)), while continuing to show elevated concentrations at B16 and B17 in 2018.For pheophytin b and pyropheophytin b elevated concentrations at B14 were most marked in 2017 (figure 4(c)), with highest hydroxychlorophyllone concentrations in 2017 and 2018 at B14 and 2019 at B13 (figure 4(d)).

Figure 2 .
Figure 2. Key pigment metrics (stations B03-B18) highlighting high algal production in surface sediment pigments in ice-free 2018 sampling years at stations B16 and B17.Parameters include: (a) total chlorins, (b) chlorophyll a, (c) hydroxychlorophyll a and (d) fucoxanthin.Upward facing arrows indicate maximum concentration among replicates from B17 in 2018.Error bars indicate standard deviation of analyses in triplicate.Additional metrics following similar trends are presented in SI figure 2.

Figure 4 .
Figure 4. Key pigment metrics (stations B03-B18) highlighting transition of the Polar Front around station B14.Plots also highlight enhanced production in 2018 ice-free summer at stations B16 and B17.Parameters include: (a) 19'-hexanoyloxyfucoxanthin, (b) total carotenoids, (c) pheophytin b & pyropheophytin b, (d) hydroxychlorophyllone.Upward facing arrows indicate maximum concentration among replicates from B17 in 2018.Error bars indicate standard deviation of analyses in triplicate.Additional indicators highlighting complex trends across the transect are presented in SI figure 4.

Figure 6 .
Figure 6.Principal components analysis (PCA) biplot on log10 transformed, centred and standardised pigment and biomarker metrics.Samples from B16 and B17 in 2018 are in the upper right quadrant in proximity to indicators which show clear responses in the same year (see figures 2 and 3).
(b)).A slight pulse in the same indicator at the most northerly station B18 (sampled 2017) could indicate the northern signature of Atlantic water masses and the effect on sediment around the north of Svalbard.Enhanced preservation of key sterol ratios (dinosterol/dinostanol & stigmasterol/stigmastanol) (figures 5(c) and (d)) at westerly and southerly stations B03 and B13 point to unique source or productivity from the Atlantic sector, with enhanced dinosterol probably reflective of a dinoflagellate source (Boon et al 1979).

Falk-
Petersen S, Sargent J R, Henderson J, Hegseth E N, Hop H and Okolodkov Y B 1998 Lipids and fatty acids in ice algae and phytoplankton from the Marginal Ice Zone in the Barents Sea Polar Biol.20 41-47 Faust J C, Ascough P, Hilton R G, Stevenson M A, Hendry K R and März C 2023 New evidence for preservation of contemporary marine organic carbon by iron in Arctic shelf sediments Environ.Res.Lett.18 014006 Faust J C, Stevenson M A, Abbott G D, Knies J, Tessin A, Mannion I, Ford A, Hilton R G, Peakall J and März C 2020 Does Arctic warming reduce preservation of organic matter in Barents Sea sediments?Phil.Trans.R. Soc.A 378 20190364 Faust J C, Tessin A, Fisher B J, Zindorf M, Papadaki S, Hendry K R, Doyle K A and März C 2021 Millennial scale persistence of organic carbon bound to iron in Arctic marine sediments Nat.Commun.12 275 Fischer A D, Moberg E A, Alexander H, Brownlee E F, Hunter-Cevera K R, Pitz K J, Rosengard S Z and Sosik H M 2014 Sixty years of Sverdrup: a retrospective of progress in the study of phytoplankton blooms Oceanography 27 222-35 Font-Muñoz J S, Jordi A, Tuval I, Arrieta J, Anglès S and Basterretxea G 2017 Advection by ocean currents modifies phytoplankton size structure J. R. Soc.Interface 14 20170046 Fortier M, Fortier L, Michel C and Legendre L 2002 Climatic and biological forcing of the vertical flux of biogenic particles under seasonal Arctic sea ice Mar.Ecol.Prog.Ser.225 1-16 Fossheim M, Nilssen E M and Aschan M 2006 Fish assemblages in the Barents Sea Mar.Biol.Res. 2 260-9 Furlong E T and Carpenter R 1988 Pigment preservation and remineralization in oxic coastal marine sediments Geochim.Cosmochim.Acta.52 87-99 Gaffey C B, Frey K E, Cooper L W and Grebmeier J M 2022 Phytoplankton bloom stages estimated from chlorophyll pigment proportions suggest delayed summer production in low sea ice years in the northern Bering Sea PLoS One 17 e0267586 Górska B and Włodarska-Kowalczuk M 2017 Food and disturbance effects on Arctic benthic biomass and production size spectra Prog.Oceanogr.152 50-61 Haberman K L, Ross R M and Quetin L B 2003 Diet of the Antarctic krill (Euphausia superba Dana): II.Selective grazing in mixed phytoplankton assemblages J. Exp.Mar.Biol.Ecol.283 97-113 Holtvoeth J, Vogel H, Wagner B and Wolff G A 2010 Lipid biomarkers in Holocene and glacial sediments from ancient Lake Ohrid (Macedonia, Albania) Biogeosciences 7 3473-89 Hop H and Pavlova O 2008 Distribution and biomass transport of ice amphipods in drifting sea ice around Svalbard Deep Sea Res.II 55 2292-307 Ji R, Jin M and Varpe Ø 2013 Sea ice phenology and timing of primary production pulses in the Arctic Ocean Glob.Change Biol.19 734-41 Kahru M, Lee Z, Mitchell B G and Nevison C D 2016 Effects of sea ice cover on satellite-detected primary production in the Arctic Ocean Biol.Lett.12 20160223 Kashiyama Y et al 2012 Ubiquity and quantitative significance of detoxification catabolism of chlorophyll associated with protistan herbivory Proc.Natl Acad.Sci.109 17328-35 Ke ˛dra M et al 2015 Status and trends in the structure of Arctic benthic food webs Polar Res.34 23775 Loeng H 1991 Features of the physical oceanographic conditions of the Barents Sea Polar Res. 10 5-18 Lundesgaard Ø, Sundfjord A, Lind S, Nilsen F and Renner A H H 2022 Import of Atlantic water and sea ice controls the ocean environment in the northern Barents Sea Ocean Sci.18 1389-418 Makarevich P R, Larionov V V, Vodopyanova V V, Bulavina A S, Ishkulova T G, Venger M P, Pastukhov I A and Vashchenko A V 2021 Phytoplankton of the Barents Sea at the polar front in spring Oceanology 61 930-43

Table 1 .
Locations of stations, depth to seafloor and date of samples used in this study.
and SI figure 1) are based on the AMSR2 ASI Sea Ice Algorithm sea ice concentration data product (https://doi.org/10.1594/PANGAEA.898399)(Spreen et al 2008) which is publicly available (https://seaice.uni-bremen.de/data-archive) (Remote Sensing for Polar Regions working group from the Remote Sensing department at the Institute of Environmental Physics, Universität Bremen) and is based on publicly available data from satellite GCOM-W1 (https://gportal.jaxa.jp/gpr).Sea-surface chlorophyll a in ice-free water maps (SI figures 9-17) use publicly available OC-CCi v5.0 data (https://climate.esa.int/en/projects/ocean-colour/news-and-events/news/ ocean-colour-version-50-data-release/) and were processed by NEODAAS (NERC Earth Observation Data Acquisition and Analysis Service) (www.neodaas.ac.uk/).Geochemical seafloor data analysed for this study are included in supplementary information file 2. Temperature and salinity data (SI figure 18) from CTD profiles are compiled from Changing Arctic Ocean (CAO) cruises (2017-2019) available publicly from the British Oceanographic Data Centre (BODC) (www.bodc.ac.uk).Kvile Ø K, Stenseth N C and Stige L C 2020 Associations among temperature, sea ice and phytoplankton bloom dynamics in the Barents Sea Mar.Ecol.Prog.Ser.635 25-36 Downes P P, Goult S J, Woodward E M S, Widdicombe C E, Tait K and Dixon J L 2021 Phosphorus dynamics in the Barents Sea Limnol.Oceanogr.66 S326-42 Falk-Petersen S, Mayzaud P, Kattner G and Sargent J R 2009 Lipids and life strategy of Arctic Calanus Mar.Biol.Res. 5 18-39