Climate change could irreversibly modify Southern Ocean ecosystems. Marine ecosystem model (MEM) ensembles can assist policy making by projecting future changes and allowing the evaluation and assessment of alternative management approaches. However, projected future changes in total consumer biomass from the Fisheries and Marine Ecosystem Model Intercomparison Project (FishMIP) global MEM ensemble highlight an uncertain future for the Southern Ocean, indicating the need for a region-specific ensemble. A large source of model uncertainty originates from the Earth system models (ESMs) used to force FishMIP models, particularly future changes to lower trophic level biomass and sea ice coverage. To build confidence in regional MEMs as ecosystem-based management tools in a changing climate that can better account for uncertainty, we propose the development of a Southern Ocean Marine Ecosystem Model Ensemble (SOMEME) contributing to the FishMIP 2.0 regional model intercomparison initiative. One of the challenges hampering progress of regional MEM ensembles is achieving the balance of global standardised inputs with regional relevance. As a first step, we design a SOMEME simulation protocol, that builds on and extends the existing FishMIP framework, in stages that include: detailed skill assessment of climate forcing variables for Southern Ocean regions, extension of fishing forcing data to include whaling, and new simulations that assess ecological links to sea-ice processes in an ensemble of candidate regional MEMs. These extensions will help advance assessments of urgently needed climate change impacts on Southern Ocean ecosystems.

The biological pump, which removes carbon from the surface ocean and regulates atmospheric carbon dioxide, comprises multiple processes that include but extend beyond gravitational settling of organic particles. Contributions to the biological pump that arise from the physical circulation are broadly referred to as physical particle injection pumps; a synthetic view of how these physical pumps interact with each other and other components of the biological pump does not yet exist. In this study, observations from a quasi-Lagrangian float and ocean glider, deployed in the Southern Ocean’s subantarctic zone for one month during the spring bloom, offer insight into daily-to-monthly fluctuations in the mixed layer pump and the eddy subduction pump. Estimated independently, each mechanism contributes intermittent export fluxes on the order of several hundreds milligrams of particulate organic carbon (POC) per day. The float and the glider produce similar estimates of the mixed layer pump, with sustained weekly periods of export fluxes with a magnitude of 400 mg-POC-m-2-day-1. Export fluxes from the eddy subduction pump, based on a mixed layer instability scaling, occasionally exceed 500 mg-POC-m-2day-1, with some periods having strong inferred vertical velocities and others having enhanced isopycnal slopes. Regimes occur when a summation of the two pump estimates may misrepresent the total physical carbon flux. Disentangling contributions from different physical pump mechanisms from sparse data will remain challenging. Insight into how mesoscale stirring and submesocale velocities set the vertical structure of POC concentrations is identified as a key target to reduce uncertainty in global carbon export fluxes.

Deep Chlorophyll Maxima (DCMs) are ubiquitous in low-latitude oceans, and of recognized biogeochemical and ecological importance. DCMs have been observed in the Southern Ocean, initially from ships and recently from profiling robotic floats, but with less understanding of their onset, duration, underlying drivers, or whether they are associated with enhanced biomass features. We report the characteristics of a DCM and DBM (Deep Biomass Maximum) in the Inter-Polar-Frontal-Zone (IPFZ) south of Australia from CTD profiles, shipboard-incubated samples, a towbody, and a BGC-ARGO float. The DCM and DBM were ~20 m thick and co-located with the nutricline, in the vicinity of a subsurface ammonium maximum characteristic of the IPFZ, but ~100 m shallower than the ferricline. Towbody transects demonstrated that the co-located DCM/DBM was broadly present across the IPFZ. Large healthy diatoms, with low iron requirements, resided within the DCM/DBM, and fixed up to 20 mmol C m-2 d-1. The BGC-ARGO float revealed the DCM/DBM persisted for >3 months. We propose a dual environmental mechanism to drive DCM/DBM formation and persistence within the IPFZ: sustained supply of both recycled iron within the subsurface ammonium maxima and upward silicate transport from depth. DCM/DBM cell-specific growth rates were considerably slower than those in the overlying mixed layer, implying that phytoplankton losses are also reduced, possibly as a result of heavily silicified diatom frustules. The light-limited seasonal termination of the observed DCM/DBM did not result in a ‘diatom dump’, rather ongoing diatom downward export occurred throughout its multi-month persistence.

In the Southern Ocean, subsurface chlorophyll maxima (SCMs) can indicate deep accumulations of phytoplankton. Recent observations of subsurface chlorophyll fluorescence maxima (SFMs) from a large network of biogeochemical Argo (BGC-Argo) floats suggest that Southern Ocean SCMs are widespread. However, the attribution of SFMs to SCMs is not trivial, and SFMs are often observed without the presence of subsurface biomass maxima (SBMs), where biomass is quantified by particulate organic carbon. Consequently, it is questionable if these widespread SFMs represent increased phytoplankton biomass or if they are formed by intracellular processes that alter chlorophyll fluorescence, without a concurrent increase in biomass, such as photo-acclimation or non-photochemical quenching. This study builds confidence in the interpretation of SFMs as SCMs and finds their widespread occurrence of SCMs in the Southern Ocean during summer. We identify SCMs from ship-based chlorophyll sampling and SFMs from fluorometers using a distributional shape-based clustering method which achieves consistent results between ship and BGC-Argo float datasets. Ship data reveal a 15 % disagreement in the identification of SFMs as SCMs. We attribute these uncertainties to non-photochemical quenching corrections and increases in chlorophyll fluorescence yields with depth. In the overlying waters above these SCMs we find increased non-algal contributions to bio-optical POC in the upper mixed layer. These non-algal stocks obscure deep accumulations of phytoplankton biomass and result in the decoupling of SBMs from SCMs in a way that cannot be explained by increases in intracellular chlorophyll fluorescence with depth.

Subsurface accumulations of chlorophyll, also known as deep chlorophyll maxima (DCMs), have been studied in the tropical and temperate oceans for decades, but they have received less attention in the Southern Ocean. Their formation and maintenance are still under debate, as is their contribution to phytoplankton biomass and net primary productivity (NPP). Recently, the application of satellite-based NPP algorithms to data from biogeochemical (BGC)-Argo floats has improved vertically-resolved NPP estimates. Using this new approach on 247 BGC-Argo floats, we report (1) subsurface (below the mixed layer) estimates of NPP, (2) the contribution of subsurface NPP to total NPP, and (3) the influence of DCMs and deep biomass maxima (DBMs, based on particulate backscatter) on (1) and (2). We compare and contrast trends in adjacent latitudinal bands in the southern hemisphere, south of 30°S, from nitrate-limited oligotrophic waters to iron-limited high-nutrient, low-chlorophyll (HNLC) regions. This comparison of pervasive DCMs in oligotrophic waters with the same features in HNLC waters reveals differences in seasonality of DCM occurrence and their contribution to total NPP. Unlike oligotrophic DCMs, HNLC DCMs occur only during spring and summer, and their contribution to total NPP decreases from ~40% to ~25% through the productive season, likely linked to the availability of iron and silicate. When DCMs are present but not accounted for, up to 45% of NPP is not quantified. Our results highlight the importance of understanding the vertical structure of phytoplankton stocks and productivity, with direct impacts on global NPP estimates and, ultimately, the biological carbon pump.

Abstract Ocean Iron Fertilization (OIF) aims to remove carbon dioxide (CO2) from the atmosphere by stimulating phytoplankton carbon-fixation and subsequent deep ocean carbon sequestration in iron-limited oceanic regions. Transdisciplinary assessments of OIF have revealed overwhelming challenges around the detection and verification of carbon sequestration and wide-ranging environmental side-effects, thereby dampening enthusiasm for OIF. Here, we utilize 5 requirements that strongly influence whether OIF can lead to atmospheric CO2 removal (CDR): The requirement (1) to use preformed nutrients from the lower overturning circulation cell; (2) for prevailing Fe-limitation; (3) for sufficient underwater light for photosynthesis; (4) for efficient carbon sequestration; (5) for sufficient air-sea CO2 transfer. We systematically evaluate these requirements using observational, experimental, and numerical data to generate circumpolar maps of OIF (cost-)efficiency south of 60°S. Results suggest that (cost-)efficient CDR is restricted to locations on the Antarctic Shelf. Here, CDR costs can be <100 US$/tonne CO2 while they are mainly >>1000 US$/tonne CO2 in offshore regions of the Southern Ocean, where mesoscale OIF experiments have previously been conducted. However, sensitivity analyses underscore that (cost-)efficiency is in all cases associated with large variability and are thus difficult to predict, which reflects our insufficient understanding of the relevant biogeochemical and physical processes. While OIF implementation on Antarctic shelves appears most (cost-)efficient, it raises legal questions because regions close to Antarctica fall under 3 overlapping layers of international law. Furthermore, the constraints set by efficiency and costs reduce the area suitable for OIF, thereby likely reducing its maximum CDR potential.

Ocean phytoplankton play a critical role in the global carbon cycle, contributing ~50% of global photosynthesis. As planktonic organisms, phytoplankton encounter significant environmental variability as they are advected throughout the ocean. How this variability impacts phytoplankton growth rates and population dynamics remains unclear. Here, we systematically investigated the impact of different rates and magnitudes of sea surface temperature (SST) variability on phytoplankton community growth rates using surface drifter observations from the Southern Ocean (> 30oS) and a phenotype-based ecosystem model. Short-term SST variability (<7 days) had a minimal impact on phytoplankton community growth rates. Moderate SST changes of 3-5oC over 7-21 days produced a large time lag between the temperature change and the biological response. The impact of SST variability on community growth rates was nonlinear and a function of the rate and magnitude of change. Additionally, the nature of variability generated in a Lagrangian reference frame (following trajectories of surface water parcels) was larger than that within an Eulerian reference frame (fixed point), which initiated different phytoplankton responses between the two reference frames. Finally, we found that these dynamics were not captured by the Eppley growth model commonly used in global biogeochemical models and resulted in an overestimation of community growth rates, particularly in dynamic, strong frontal regions of the Southern Ocean. This work demonstrates that the timescale for environmental selection (community replacement) is a critical factor in determining community composition and takes a first step towards including the impact of variability and biological response times into biogeochemical models.

Phytoplankton indirectly influence the climate, through their role in the ocean biological carbon pump. Hence factors limiting phytoplankton growth directly impact the strength of the biological carbon pump and consequently climate. In the Southern Ocean, the subantarctic zone represents an important carbon sink, yet variables limiting phytoplankton growth are not fully constrained. Co-limitation by iron (Fe) and manganese (Mn) has recently been observed in the coastal and offshore Southern Ocean, but very few studies have focused on the subantarctic zone. In addition, no study has investigated the seasonal variability of Mn (co-)limitation of phytoplankton growth in the Southern Ocean. Using three shipboard bioassay experiments, we evaluated the seasonality of Fe and Mn co-limitation of subantarctic phytoplankton growth, south of Tasmania. We observed a strong seasonal variation in phytoplankton Fe limitation, and that the response of phytoplankton to Mn was subtle and thus readily masked by the responses to Fe. Combined addition of Fe and Mn enhanced carbon uptake of nanoeukaryotes in spring and microeukaryotes in summer while the addition of Mn alone stimulated the growth of picocyanobacteria in autumn. These results suggest the importance of Mn may vary seasonally and its control on phytoplankton growth may be associated with specific taxa.

Southern Ocean eddies shape the foraging ecology of marine apex predators such as marine mammals and seabirds. A growing number of animal tracking studies show that predators alter their swimming, diving, and foraging behavior in mesoscale eddies. However, little is known about how Southern Ocean eddies influence the distribution of mesopelagic micronekton (fish, squid, and crustaceans), which are major prey items of megafauna. Studies in other parts of the world have found that eddies can impact the abundance and community composition of micronekton. Here, we analyze acoustic observations from a 14-day survey of a mesoscale eddy, its surrounding waters, and the Sub-Antarctic frontal waters where the eddy originated. We report and interpret spatial patterns of acoustic backscattering at 18 kHz, a proxy indicating combined changes in species, size, and abundance of micronekton. We find that the vertical distribution of Deep Scattering Layers matched the underwater light conditions characteristic of the eddy core, periphery, and surrounding waters, at scales smaller than 10 km. Furthermore, the average water-column integrated acoustic backscattering values in the eddy core were only half of the values measured in the Sub-Antarctic Zone waters surrounding the eddy. By contrast, the acoustic properties of the eddy core were similar to those measured in the Polar Front Zone, where the eddy originated 27 days before our sampling. These results show that, as for physical and chemical tracers, the eddy maintained its biological characteristics from its source waters creating a unique habitat compared to its surrounding waters.

In the Subantarctic sector of the Southern Ocean, vertical entrainment of dissolved iron (DFe) triggers the seasonal productivity cycle. However, diminishing physical supply of new Fe during the spring to summer transition rapidly drives epipelagic microbial communities to rely upon recycled DFe for growth. Hence, subpolar waters evolve seasonally from a high fe ratio system (i.e., [uptake of new Fe]/[uptake of new+recycled Fe]) to a low fe ratio system. Here, we tested how resident microbes within a cyclonic eddy respond to different Fe/ligand inputs which mimic entrained new DFe (Fe-NEW), diffusively-supplied regenerated DFe (Fe-REG), and a control with no addition of DFe (Fe-NO). After 6 days, 3.5 (Fe-NO, Fe-NEW) to 5-fold (Fe-REG) increases in Chl a were observed despite ~2.5-fold range between treatments of initial DFe. Marked differences were also evident in the proportion of in vitro DFe derived from recycling to sustain phytoplankton growth (Fe-REG, 30% recycled c.f. 70% Fe-NEW, 50% Fe-NO). This trend supports the concept that DFe/ligands released from subsurface particles are more bioavailable than new DFe collected at the same depth. This additional recycling may be mediated by bacteria. Indeed, by day 6 bacterial production (BP) was comparable between Fe-NO and Fe-NEW but~2 fold higher in Fe-REG. Interestingly, a preferential response of phytoplankton (haptophyte-dominated) relative to bacteria was also found in Fe-REG. In contrast, in Fe-NEW and Fe-NO the proportion of diatoms increased. Hence, different modes of Fe/ligand supply modify BP and Fe bioavailability to phytoplankton that may drive distinctive floristic shifts and biogeochemical signatures.

The necessity to understand the influence of global ocean change on biota has exposed wide-ranging gaps in our knowledge of the fundamental principles that underpin marine life. Concurrently, physiological research has stagnated, in part driven by the advent and rapid evolution of molecular biological techniques, such that they now influence all lines of enquiry in biological and microbial oceanography. This dominance has led to an implicit assumption that physiology is outmoded, and advocacy that ecological and biogeochemical models can be directly informed by omics. However, the main modelling currencies continue to be biological rates and biogeochemical fluxes. Here we ask: how do we translate the wealth of information on physiological potential from omics-based studies to quantifiable physiological rates and, ultimately, to biogeochemical fluxes? Based on the trajectory of the state-of-the-art in biomedical sciences, along with case-studies from ocean sciences, we conclude that it is unlikely that omics can provide such rates in the coming decade. Thus, while physiological rates will continue to be central to providing projections of global change biology, we must revisit the metrics we rely upon. We advocate for the co-design of a new generation of rate measurements that better link the benefits of omics and physiology.

1. Detritivores need to up-cycle their food to increase its nutritional value. One method is to fragment detritus promoting the colonisation of nutrient-rich microbes, which consumers then ingest. This is known as microbial gardening. Observations and numerical models of the detritus-dominated ocean mesopelagic zone have suggested microbial gardening by zooplankton is fundamental process in the ocean organic carbon cycle, as it leads to increased respiration of carbon-rich detritus. However, no experimental evidence exists to prove microbial respiration is higher on smaller, fragmented detrital particles. 2. Using aquaria-reared Antarctic krill faecal pellets we showed fragmentation increased microbial particulate organic carbon (POC) turnover by 70 %, but only on brown faecal pellets of low nutritional value. Microbial POC turnover on un-and fragmented green faecal pellets of higher nutritional value was equal. Thus we find particle size alone is not enough to determine microbial activity, and the nutritional value and age of the particle are important. 3. We estimate mesopelagic zooplankton can potentially increase the proportion of essential nutrients (e.g. unsaturated fatty acids) in their food by at least 11 %. In addition we propose ‘communal gardening’ may occur whereby other mesopelagic organisms consume the particle and microbes gardened by a neighbouring detritivore. 4. Increases in microbial turnover of detrital POC reduces the sink of organic carbon in the ocean. Thus microbial gardening should be represented in models forecasting the future carbon cycle. Model parameterisations will require further understanding of the energetic gains to zooplankton communities, how microbial gardening influences other sinking particles such as detrital aggregates, and the relative importance of biological (i.e. particle lability, size and age) vs. physical (i.e. temperature and oxygen) constraints on gardening.