The Southern Ocean is a globally important site of carbon sequestration (DeVries, 2022). This is largely achieved by cool waters with higher carbon dioxidesolubility, deep water formation and a strong biological carbon pump (BCP), the process by which organisms mediate the transfer of carbon from surface waters into the deep ocean (Boyd et al., 2024; Volk and Hoffert, 1985). In the Southern Ocean, most total organic carbon is produced as particulates rather than dissolved material (Carlson et al., 2000). These particulates consist of protists, detritus, aggregates and zooplankton faecal pellets, moults and carcasses (Alldredge and Silver, 1988; Turner, 2015; Halfter et al., 2022). The euphotic zone determines the potential upper limit of the biological carbon pump through primary production; however, as these particles sink out of the euphotic zone, they are progressively modified (e.g. by fragmentation or aggregation), solubilised, remineralised and repackaged. This results in the flux attenuating with depth, often rapidly, as particulate organic carbon (POC) is converted into dissolved inorganic carbon (DIC) and dissolved organic carbon (DOC) (Goldthwait et al., 2005; Martin et al., 1987; Le Moigne, 2019; Volk and Hoffert, 1985). The particle flux can experience intense remineralisation in the upper ocean, releasing dissolved carbon towards the surface and fuelling the microbial loop wherein heterotrophic microbes consume dissolved organic matter and funnel this carbon back to higher trophic levels (Azam et al., 1983). However, ephemeral high production events or zooplankton driven modifications can bypass rapid attenuation near the surface, increasing the proportion of dissolved carbon which is eventually released into deeper waters (Buesseler and Boyd, 2009). High production events may arise after the collapse of a bloom causing rapidly sinking pulse of particulates (Lacour et al., 2023; Manno et al., 2020) and zooplankton can significantly impact the flux through production of faecal pellets and detritivory. Faecal pellet production repackages material into dense, fast sinking particles, contributing to the flux (Cavan et al., 2015; Liszka et al., 2019; Manno et al., 2015). Conversely, detritivory and sloppy feeding can fragment particles, reducing their size and sinking rate, contributing to flux attenuation (Mayor et al., 2014). The balance of these processes is highly dependent upon the resident zooplankton community as well as the composition and bloom stage of the primary producers (Lacour et al., 2023; Le Moigne et al., 2015; Steinberg and Landry, 2017).

Within the Southern Ocean, the Weddell Sea is simultaneously highly productive and a site of deep-water formation, helping to drive the global thermohaline ocean circulation (Talley, 2013). Consequently, it is of global biogeochemical significance. Indeed, the Weddell Sea accounts for ∼ 25 % of Southern Ocean primary productivity, averaging ∼ 477 Tg C a⁻¹ (Arrigo et al., 2008). However, this productivity is largely focused on the shelf and the marginal ice zone (Arrigo et al., 2008; El-Sayed and Taguchi, 1981). Iron and light availability typically limit productivity in the Southern Ocean, creating regions of high nutrient availability but low chlorophyll concentration however, sea ice, ice shelf associated melt waters and shelf waters often contain elevated nutrient concentrations due to greater terrigenous inputs and a stronger influence of resuspension, fuelling phytoplankton blooms (Arrigo et al., 2008; Klunder et al., 2014; Middag et al., 2013; Smith and Nelson, 1985). This productivity supports a wealth of secondary consumers, including Antarctic krill (Euphausia superba) that directly impact the BCP via the production of dense, fast sinking faecal pellets (FPs), as well as rapid grazing rates exerting top-down pressures on phytoplankton blooms (Schmidt et al., 2018; Whitehouse et al., 2009). Although estimates of krill biomass from the Weddell Sea are scarce as the area is difficult to access due to its remote location and consistent sea ice cover (Atkinson et al., 2017), observed values range from 1–68 g m⁻² wet weight in open water and 10–100 g m⁻² under ice (Brierley et al., 2002; Daly and Macaulay, 1988; Guihen et al., 2014). Model based estimates of krill contributions to the BCP suggest the northern Weddell Sea marginal ice zone may be one of the most productive regions in the Southern Ocean, with a potential contribution of up to 325–3350 mg C m⁻² d⁻¹ in spring (Belcher et al., 2019). This high productivity coupled with its hydrography makes the Weddell Sea a site of global biogeochemical importance (Brown et al., 2015; Hoppema, 2004).

As a major contributor to deep water formation, the Weddell Sea generates up to 60 % of Antarctic Bottom Water, the most voluminous and densest water mass on earth (Johnson, 2008; Orsi et al., 1999). Carbon sequestration periods can be increased when POC is remineralised in regions where deep water is formed (Baker et al., 2022), and as such these physical processes combined with the intense biological particulate production make carbon sequestration highly efficient in the Weddell Sea (Hoppema, 2004; Nissen et al., 2022; Volk and Hoffert, 1985). The Weddell Sea is experiencing rapid climate change, with a recent shift towards decreasing sea ice extent and thickness (Joshi et al., 2024). Warming, changes to sea ice and ice shelf melt dynamics have potential to impact both deep water formation and the resident plankton community, which in turn may impact the capacity of the Weddell Sea to act as a carbon sink (Nissen et al., 2022; Zhou et al., 2023). Consequently, improving our understanding of the mechanisms regulating the magnitude of the carbon pump in the Weddell Sea shelf and marginal ice zone, including characterisation of the vertical carbon flux itself, is of vital importance.

This study quantifies and characterises the variability of daily vertical particulate organic carbon and particulate organic nitrogen fluxes across 50, 100 and 150 m in the western Weddell Sea (4–27 February 2024), using a custom-built drifting sediment trap (DST). The DST was deployed on and off the shelf, within sea ice and in open water. This study aims to characterise the major inputs into the biogeochemical particle flux, and how these change with the environment and across depths. Results provide insight on Southern Ocean carbon cycle drivers by capturing the flux both where the upper limit of carbon flux is determined and the greatest influence on attenuation is exerted (Buesseler et al., 2020).

The POC flux that we measured in the western Weddell Sea varied from 42.5–364.1 mg C m⁻² d⁻¹ (mean of 123.2 mg C m⁻² d⁻¹), with higher fluxes observed on the shelf (S1, S4, S3, S5) than off the shelf (S2), correlating with the magnitude and depth of the chlorophyll maximum (Figs. 4 and S1A, B). Despite the importance of the Weddell Sea in global biogeochemical cycling and climate change disrupting sea ice production, contemporary direct particulate flux measurements in the Weddell Sea remain scarce (Joshi et al., 2024). Globally, 75 % of POC fluxes at 150 m are ≤90 mg C m⁻² d⁻¹, indicating the majority of our stations were well above this metric (Mouw et al., 2016). Previous studies in the northern and eastern Weddell Sea have reported POC fluxes ranging from 24 to 112 mg C m⁻² d⁻¹ at depths of 150–250 m, with occasional extreme values exceeding 1000 mg C m⁻² d⁻¹ during spring bloom conditions (von Bodungen, 1986; Cadée, 1992; Dunbar, 1984; Isla et al., 2006; Wefer et al., 1990). The fluxes observed in our study fall well within this broad range, although the absence of extreme values likely reflects sampling late in the productive season (end of February), as the high values (> 1000 mg C m⁻² d⁻¹) reported by Cadée (1992) and von Bodungen (1986) both occurred in spring (November/December). There are only a few observations of PON fluxes within the Weddell Sea. Previous studies reported PON fluxes ranging from 0.51–20.86 mg N m⁻² d⁻¹ (10–230 m) (Thomas et al., 2001; Michels et al., 2008) which is the same order of magnitude as our observations (3.4–51.9 mg N m⁻² d⁻¹).

In order to prevent the trap drifting into rough weather, the trap was tethered to the ship on its first deployment at S1, the station closest to the coast. Tethered traps have the potential to behave differently from free drifting traps, with lateral advection across the trap potentially reducing the collection efficiency of the trap (Buesseler et al., 2000). To help mitigate the impact of tethering, the ship was allowed to drift alongside this trap where reasonably possible, however collection efficiency may still have been impacted. Despite tethering, patterns in the magnitude and characterisation of the flux present at S1 appear to well reflect the productive conditions present at this site. Station S1 also had the highest chlorophyll concentration observed (2.73 µg L⁻¹, Fig. 3), the highest POC and PON fluxes (Fig. 4) and a mean POC : PON ratio of 7.45; close to the Redfield ratio of phytoplankton (6.63, Redfield, 1934). Together these values suggest high productivity and export with relatively fresh particulate material which had experienced little nitrogen remineralisation (Copin-Montegut and Copin-Montegut, 1983; Wakeham et al., 1997) (Fig. 5). Indeed, the chlorophyll maximum at S1 is in line with bloom-like conditions observed within the wider Weddell Sea (von Berg et al., 2020), however in this particular region of the shelf, chlorophyll concentrations > 5 µg L⁻¹ have been observed (Salyuk et al., 2025). This suggests S1 was productive, but the bloom was not as strong as is possible on the shelf, perhaps a result of sampling towards the end of the productive season. The higher POC fluxes observed consistently across the shelf (S1, S3, S4, S5) likely reflects the higher concentrations of (micro)nutrients (such as iron and manganese) compared to the Weddell Sea open waters (Balaguer et al., 2023; Klunder et al., 2014; Middag et al., 2013). Moreover, the shallower water column of the shelf stations supports mixing and resuspension that facilitates nutrient resupply to the surface (Isla et al., 2006; Pudsey and King, 1997; Semper and Darelius, 2017).

Although the sea ice edge can also promote productivity, there were no strong patterns observed in the particulate flux between sea ice and ice-free stations (Arrigo et al., 2008; Lannuzel et al., 2016; Smith and Nelson, 1985). This likely reflects the highly dynamic nature of sea ice during the sampling period. Stations were classified as ice-free based on conditions present during deployment, however, as the ice was constantly shifting, we suggest that the particulate flux integrated longer-term patterns in sea ice coverage which were minimised when we only accounted for deployment conditions. For instance, according to satellite images and coverage data, S2 (classed as “ice free”) was ice-covered two days prior to sampling (Figs. S2, S3), and S5 had a variable sea ice coverage throughout February and January (Fig. S5). The likelihood that most of our stations are heavily influenced by sea ice is supported also by patterns in POC : PON. With the exception of S1, the POC : PON ratio of the flux varies from 9.35–18.84. These ratios exceed both classic Redfield ratio for phytoplankton (6.63, Redfield, 1934) as well as the Southern Ocean specific ratio (mean of 6) (Tanioka et al., 2022). Sinking particulate material commonly has a higher C : N relative to fresh phytoplankton, as nitrogen is preferentially absorbed by grazers and microbes with depth (Atkinson et al., 2012; Copin-Montegut and Copin-Montegut, 1983; Wakeham et al., 1997). For example, fresh Krill FP C : N has been shown to range between 3.2–13.2 (Atkinson et al., 2012; Manno et al., 2024). However, our range of ratios exceed values stated above, suggesting the potential for an additional influence of ice associated biota across the stations. Ice associated microplankton often have elevated and more variable C : N ratios as compared to pelagic phytoplankton, with ice associated microplankton C : N ranging from 6.2–12.6 (Cozzi and Cantoni, 2011; Kennedy et al., 2002; Niemi and Michel, 2015) and pelagic Southern Ocean phytoplankton broadly ranging between ∼ 6–8 (Arrigo et al., 1999; Hoppe et al., 2017). Indeed, microplankton community data from stations S2–S5 indicate that ice associated microplankton comprise and average 11.5 %–26.4 % (mean of 18.2 %) of protists up to the 1 % photosynthetically available radiance depth (Table S6). In addition, the mean contribution of ice associated microplankton at S1 is lower (7.7 %), reflecting POC : PON patterns in our data. These findings highlight the limitations of categorising flux observations solely based on instantaneous ice conditions in dynamic polar environments.

Antarctic krill exerted a dominant control on the particulate carbon flux across shelf stations. Cylindrical FP, reliably identifiable as krill-derived, accounted for 98 % of total FP volume (with round and oval pellets accounting for ∼ 1 % of total FP volume each). 88 % of cylindrical FP were wider or equal to 80 µm, indicating that they likely originated from E. superba (Atkinson et al., 2012). The dominance of different FP types can be highly variable, dependent upon food availability and the zooplankton community composition (Manno et al., 2015; Perhirin et al., 2025). However, in regions where krill dominate, krill FP have been found to represent > 80 % of all sinking particles (Cavan et al., 2017), in good alignment with our data. Our traps were deployed between 11.4 and 22.4 h, and shorter deployments may favour faster sinking larger pellet sizes such as krill FPs (Baker et al., 2020; Perhirin et al., 2025). However, our results are also in good agreement with longer-term moored sediment traps within the Weddell Sea, where pulses of krill FP flux dominate the particulate flux on the shelf (Bathmann et al., 1991; von Bodungen, 1986; Dunbar, 1984).

Using a conversion factor of 0.08 mg C mm³ (Pauli et al., 2021), cylindrical FP contributed between 17.4 % and 99.0 % of carbon to the total POC flux (with a median of 47.9 %), with the remaining carbon attributed to detritus (largely phytodetritus) (Fig. 5A). The carbon content of krill FP is not fixed, varying with grazing rates and food availability (Atkinson et al., 2012). Indeed, Pauli et al. (2021) reported a standard deviation of ±0.044 C mg mm³ (equivalent to ±55 %), which could extend our estimated range of FP contribution to the POC flux, with the median value ranging between 26.3 %–74.2 %. Despite this variability, stations with higher POC fluxes generally exhibited higher faecal pellet fluxes, indicating a strong coupling between krill activity and flux magnitude. These observations reinforce previous findings that krill-mediated repackaging of organic matter into large, dense, fast-sinking particles is a key pathway for carbon export in the Southern Ocean (Belcher et al., 2017, 2019; Cavan et al., 2019; Manno et al., 2022; Trinh et al., 2023).

POC fluxes increased or remained relatively stable with depth across most stations (S1, S2, S4, S5) rather than attenuating rapidly. The absence of attenuation in our data may be explained by pulses of productivity in the recent past, lateral advection of material, zooplankton producing particulates at depth and or a lack of zooplankton grazing on/fragmenting particles (e.g. Cavan et al., 2015; Freudenthal et al., 2001; Le Moigne et al., 2015; Smith et al., 2018). While temporal mismatches between production and collection cannot be fully excluded, physical processes such as lateral advection should have a minimal impact on “catch efficiency” given the sediment trap was free drifting for stations S2–S5 (Buesseler et al., 2000). We suggest instead that the consistent vertical patterns in faecal pellet flux are strongly indicative of zooplankton mediated control. Across most stations (S1–S4), krill faecal pellet flux peaked at 100 m, irrespective of surface chlorophyll concentration and environment. We propose this result is indicative of krill vertical migration or movement, and production of pellets at depths between 50 and 100 m (below the MLD). Patterns in krill vertical distribution are both highly variable and seasonal, although in summer krill often migrate to depths of ∼ 100 m during diel vertical migration, krill are strong swimmers and can also disperse throughout the water column, even migrating to the seabed (Bahlburg et al., 2023; Clarke and Tyler, 2008; Schmidt et al., 2011; Smith et al., 2025; Tarling et al., 2018). Vertical movements of krill (i.e. krill spending time at depths of 50–100 m) well reflect the trends in our data. In addition, individual FP volumes remained largely consistent with depth across stations (Fig. 5B), suggesting that relatively little FP fragmentation took place between 50 and 150 m. As discussed above, this may support the hypothesis that krill are defecating at depth, meaning less opportunity for FP to fragment before they are collected at 100 or 150 m. Consistent FP volumes with depth may also reflect rapid sinking speeds as well as low fragmentation rates (either from higher particle integrity or minimal interaction with detritivores) (Atkinson et al., 2012; Briggs et al., 2020; Mayor et al., 2014). Indeed, at S1 and S5, cylindrical faecal pellet volumes significantly increase with depth, and have a lower POC : PON ratio which is relatively consistent with depth, likely indicating a pulse of relatively fresh material is caught in the trap, with rapid sinking of larger pellets. However, given that krill faecal pellet flux peaks at 100 m across stations S1–S4 (regardless of whether patterns in the POC : PON ratio indicate a recent “pulse” of material or otherwise) we suggest a behavioural component (i.e. vertical migration or movement) is also exerting a significant influence on patterns in particulate fluxes and that overall, krill FP have potential to strongly impact export as well as upper ocean fluxes.

By producing faecal pellets below the mixed layer rather than at the surface, migrating krill effectively bypass near-surface remineralisation, enhancing the efficiency of carbon export (Liszka et al., 2019; Steinberg and Landry, 2017). Given estimated sinking rates of krill pellets (27–1218 m d⁻¹, mean of 309 m d⁻¹; Atkinson et al., 2012) a substantial fraction of this material could reach the seafloor on the shelf (average depth of 421 m) within < 1–15 d, contributing directly to benthic carbon supply and enhancing long-term sequestration. Carbon-specific krill pellet degradation rates have been estimated as a 23 %–30 % loss at 300 m and up to a 35 % loss at 500 m (Pauli et al., 2021). Consequently, across the shelf stations (S1, S3, S4, S5), between 19.8–111.2 mg C m² d⁻² could reach the seafloor. The contribution of krill to the particulate carbon flux was not limited to FPs. Krill eggs present at stations S1 and S3 potentially contributed (2.9 %–6.7 %) to the POC flux (Fig. 6). The available food environment has a significant impact on maternal condition, egg quality and therefore hatching success (Quetin and Ross, 2001; Steinke et al., 2024; Yoshida et al., 2011). A substantial but undetermined fraction of these eggs will lead to deep carbon export, either through failure to hatch or through hatching to larvae which are non-viable or eaten while still at depth (Perry et al., 2020). Overall, our results highlight that krill eggs have the potential to contribute significantly to the particulate carbon flux, although their contribution is yet unquantified at an ocean scale.

This study provided observations of the particulate flux in the upper western Weddell Sea across multiple depths, a region of global importance for carbon sequestration and deep-water formation. We demonstrated that particulate fluxes in the upper 150 m are strongly structured by shelf processes, ice-associated production and zooplankton activity. POC and PON fluxes were higher on the shelf than open ocean, reflecting enhanced nutrient supply and biological productivity in shelf waters. In contrast, no clear distinction was found between sea ice and ice-free stations, reflecting the constant movement of the ice and that often our “ice free” sites were still influenced by ice.

Krill exerted a strong control on the carbon flux across all shelf sites, with over 98 % of all FP belonging to krill, and contributions of up to 99 % of the POC flux emphasising the role of zooplankton repackaging in mediating carbon export. The peak in FP flux occurred consistently at 100 m across the shelf, suggesting krill are producing FPs at depth. Krill eggs are a poorly constrained component of the flux but our results highlight their potential role as a significant pathway for carbon transfer, particularly towards the end of the productive season where they may contribute up to 7 % of POC flux.

As climate change influences sea ice dynamics, the Southern Ocean is subject to dramatic change, and it is unclear how zooplankton communities may react. Our findings show the importance of zooplankton mediated processes in driving carbon export within the upper 150 m, a region where particulate modification is most intense. Consequently, it is vital that biogeochemical studies in the euphotic zone, incorporate zooplankton processes (particularly vertical migration or movements) to obtain a representative, process-based understanding of carbon cycling in a rapidly changing Southern Ocean.