Antarctic krill sequester similar amounts of carbon to key coastal blue carbon habitats

Marine life plays a key role in oceanic carbon storage termed blue carbon. Coastal vegetation (seagrass, mangroves, salt marshes) dominates blue carbon research and policy, but the open-ocean biological carbon pump also stores vast amounts of carbon in the deep sea and is not routinely framed as blue carbon. Phytoplankton fix dissolved CO2 into organic carbon; upon death or grazing, organic matter and faecal pellets sink, and if not remineralised, carbon is sequestered for decades to centuries. Without the biological pump, atmospheric CO2 would be about 50% higher. Antarctic krill (Euphausia superba) are key to this pump; their faecal pellets, moults and carcasses can dominate Southern Ocean carbon fluxes during the Austral growth season (October–April). Krill face threats from climate-driven sea-ice loss and an expanding fishery lacking management tuned to ecological variability. Antarctic ecosystem services are globally important but comparatively under-quantified. The study’s purpose is to estimate carbon sequestration for the entire circumpolar Antarctic krill population to enable comparison with coastal blue carbon systems and to showcase the role of open-ocean animals in carbon sequestration, with an approach extendable to other pelagic organisms.

Prior work has emphasised coastal vegetated habitats as blue carbon stores, while the biological carbon pump’s role in open oceans has been less integrated into policy contexts. Antarctic krill have been identified as major contributors to Southern Ocean particulate organic carbon (POC) flux through faecal pellets, moults and carcasses, with observations showing krill-dominated fluxes in many regions and seasons. The KRILLBASE dataset (1926–2016) enabled large-scale estimates of krill distributions. Earlier modeling (e.g., Belcher et al.) suggested substantial krill faecal pellet carbon production and export in the marginal ice zone during spring–early autumn, but export is not equivalent to sequestration. Models of the ocean circulation (e.g., OCIM) provide time-to-surface metrics to identify sequestration depths (≥100-year isolation). Attenuation of sinking particles is often represented via Martin’s b parameter; krill pellets’ rapid sinking suggests lower attenuation (less remineralisation) than the global POC average, implying greater transfer to depth. There remain uncertainties in krill egestion rates, attenuation variability, and the contribution of krill life-history elements (moults, migrations, carcasses) not always captured in biogeochemical models, which often omit micronekton like krill.

Study period covered the Austral productive season (October–April). Krill density: Postlarval krill numerical densities were obtained from KRILLBASE (1926–2016) filtered for samples with top depth <20 m and bottom depth >50 m, standardised to January 1 using empirical algorithms, and extended to monthly means for October–April. Densities were projected onto a 2° latitude × 6° longitude grid. Densities >600 ind. m−2 (0.3% of data) were capped at 600 to reduce bias from extreme swarms. The circumpolar area considered is about 16 million km². Faecal pellet carbon production (FPCprod): For each cell and month, FPCprod at 20 m was calculated as krill density (N, individuals m−2) times egestion rate (E). A revised, conservative adult egestion rate of 0.46 mg C ind−1 d−1 (median; 5th–95th percentile 0.11–1.23 mg C ind−1 d−1) was derived via three independent methods and checked against plausible annual food consumption relative to regional primary production. Attenuation to depth: Mixed-layer export (at 20 m) was attenuated using a power-law (Martin’s b) based on krill pellet observations, with b = −0.30 as the best estimate, and a global comparison case of b = −0.86. This attenuation implicitly captures sinking, remineralisation, and fragmentation. Carbon sequestration depth: Sequestration depth was defined as the depth where the First Passage Time (FPT) to the surface is ≥100 years using the Ocean Circulation Inverse Model (OCIM; 2° resolution, 24 vertical levels). Sequestration depth varies spatially (135–760 m) with a circumpolar mean of 381 m. Sequestered flux (FPCflux): For each cell and month, FPCflux at the sequestration depth was computed by attenuating FPCprod from 20 m to FPT100 using the chosen b. Total sequestered carbon was obtained by summing FPCflux × cell area × days across all cells and months (October–April). Validation: Modeled fluxes were compared to sediment trap observations. At South Georgia (300 m), observed krill pellet flux in January was 46 mgC m−2 d−1 vs modeled 40 mgC m−2 d−1. At Palmer Station (170 m), observed total POC flux in January averaged 16 mgC m−2 d−1; modeled krill pellet flux was 6 mgC m−2 d−1 (≈38% of total), noting the model includes only adult krill (>40 mm). Sensitivity and uncertainty analyses: Parameters (krill density, egestion, attenuation b, sequestration depth) were varied by ±10% and across plausible ranges (b 25th–75th percentiles −0.61 to +0.13; global b −0.86; egestion 5th–95th percentiles; sequestration depth 5th–95th). Economic valuation: Total sequestered C was converted to USD using the Social Cost of CO2 (SCCO2) spanning USD$51–640 per tCO2, with mass conversion from C to CO2 using 44/12. Additional components and comparisons: Best estimates for krill moults and active transport (migration respiration) were included for context (moults ≈20 MtC yr−1; migration ≈26 MtC yr−1). Krill contributions were compared to modeled sequestration by phytoplankton and copepods and to global coastal blue carbon habitats (mangroves, seagrasses, salt marshes). Tracking krill-derived DIC: Using the OCIM transport matrix, steady-state dissolved inorganic carbon distribution and residence time from remineralised krill pellets were computed for b = −0.30 (and sensitivity with b = −0.86).

• Mean sequestration depth across the Southern Ocean was 381 m (range 135–760 m), implying pellets need only reach mid-mesopelagic depths for ≥100-year sequestration. Depths were shallowest in hotspots like the Scotia Sea and deepest in the Indian sector. - Mixed-layer krill faecal pellet export (FPCprod at 20 m) totaled 44 MtC yr−1 over October–April; 20 MtC yr−1 (≈45%) is sequestered at FPT100 depths. - Spatial fluxes: At 20 m, FPCprod ranged 0–276 mgC m−2 d−1; at sequestration depths, FPCflux ranged 0–120 mgC m−2 d−1. Highest average fluxes occurred in the Atlantic sector where krill densities are greatest. - Regional totals of sequestered C: Atlantic 13.7 MtC (USD$2.6–32.2B), Indian 4.1 MtC (USD$0.8–9.5B), Pacific 1.9 MtC (USD$0.3–4.4B); All regions sum to 19.7–20 MtC (USD$3.7–46.1B). - Efficiency and timing: On average ≈2.5% of regional NPP is routed to krill pellet sequestration, reaching up to 74% in swarm locations. Transit time from photosynthesis to sequestration is likely under a week due to rapid gut passage (<1 day) and pellet sinking (~300 m d−1). Transfer efficiency of krill pellets to sequestration depths (~45%) far exceeds commonly reported 1–10% for the biological pump defined at fixed deeper horizons (e.g., 1000–2000 m). - Contribution relative to plankton: Modeled sequestration by phytoplankton and copepods is ~160 MtC in the study region; krill pellets add ~20 MtC, i.e., ~12% of total plankton-mediated sequestration. Including moults (≈20 MtC) and active transport (≈26 MtC) yields a potential krill total of ~66 MtC yr−1. - Global storage and residence: For b = −0.30, the steady-state global DIC inventory from remineralised krill pellets is ~8.7 Gt C with a mean residence time of ~219 years; ~20% of export reaches the seafloor as particulate pellets, enabling benthic sequestration pathways. Using a global b = −0.86 reduces storage to ~1.7 Gt C and residence to ~58 years. - Model–data agreement: At South Georgia (300 m), modeled krill pellet flux (40 mgC m−2 d−1) matched observations (46 mgC m−2 d−1). At Palmer Station (170 m), modeled krill pellet flux was 6 mgC m−2 d−1 vs total observed POC 16 mgC m−2 d−1 (adult-only model likely underestimates juvenile-rich coastal areas). - Valuation: The 20 MtC sequestered per season corresponds to USD$4–46 billion depending on SCCO2. - Comparability to coastal blue carbon: Seasonal krill pellet sequestration (~20 MtC) is of the same order as global salt marsh (13 MtC yr−1), mangrove (24 MtC yr−1), and seagrass (44 MtC yr−1) sequestration rates, with krill’s vast habitat area compensating for lower per-area fluxes.

By integrating krill distribution, pellet production, particle attenuation, and circulation-based sequestration depths, the study quantifies how a single pelagic species can drive carbon sequestration on par with globally recognized coastal blue carbon habitats. Krill’s large biomass, rapid pellet sinking, and relatively modest sequestration depths in the Southern Ocean yield high transfer efficiency and substantial seasonal sequestration totals. Results imply that models and policies overlooking pelagic micronekton underestimate ocean carbon sequestration and misattribute fluxes to phytoplankton detritus or copepod pellets. Particle-type-specific attenuation is critical: adopting a krill-appropriate b (−0.30) increases both depth transfer and storage longevity compared to the global average b. The OCIM-based tracking highlights long residence times and widespread distribution of remineralised krill carbon, predominantly within the Atlantic sector but also reaching the Indian and Pacific basins, even into the Northern Hemisphere. In a blue carbon policy context, krill-dominated regions function as important carbon sinks while also supporting iconic biodiversity, strengthening conservation arguments. However, krill populations are vulnerable to climate-driven sea-ice loss and fishery pressures, and community shifts (e.g., toward salps) could alter sequestration efficiency due to different pellet properties. Incorporating krill and other micronekton, and behaviors like diel vertical migration, into Earth System Models should be a priority to improve projections of the biological carbon pump under climate change.

The study provides the first circumpolar estimate of Antarctic krill faecal pellet carbon sequestration grounded in observed krill distributions, particle attenuation tailored to krill pellets, and physically informed sequestration depths. Antarctic krill faecal pellets sequester approximately 20 MtC per Austral productive season, worth USD$4–46 billion at current SCCO2 ranges, comparable in magnitude to global sequestration by key coastal blue carbon habitats. Including krill moults and migration-related respiration could raise total krill-mediated sequestration to ~66 MtC yr−1. The findings underscore the need to explicitly represent pelagic micronekton and particle-type-dependent attenuation in models and to recognize pelagic blue carbon in conservation and management. Future work should reduce uncertainties in attenuation (b) via targeted observations across regions and seasons, refine egestion estimates and depth distributions of production, incorporate carcasses, juvenile krill, and winter processes, and leverage higher-resolution circulation models to better resolve sequestration depths and pathways.

• Attenuation uncertainty: Martin’s b shows the largest natural variability and observational spread. Using b within the 25th–75th percentile range (−0.61 to +0.13) changes total sequestration from ~9 to ~63 MtC; adopting the global b (−0.86) yields ~4 MtC. - Circulation model resolution: OCIM’s coarse resolution (2°) and largely monotonic FPT with depth may not fully resolve complex Southern Ocean water mass structure, potentially biasing sequestration depth estimates regionally. - Egestion and production assumptions: A conservative adult egestion rate (0.46 mg C ind−1 d−1) was used; actual rates vary with diet and season. The model assumes production at 20 m and attenuates to sequestration depth without explicitly resolving depth-varying egestion. - Biomass representation: Model focuses on adult krill (>40 mm) and may under-represent juvenile-dominated coastal regions. KRILLBASE densities are spatially and temporally uneven and highly skewed; extreme values were capped at 600 ind. m−2 to limit bias. - Seasonal and component scope: Estimates cover October–April and faecal pellets only for the main result; carcasses are not fully quantified in sequestration totals, and winter processes are not explicitly modeled. - Economic valuation: SCCO2 spans a wide range and is policy-dependent; valuation excludes co-benefits or potential leakage effects. - Validation is limited to select sites with sediment traps; broader observational coverage of krill pellet fluxes and attenuation is needed.