Iron (Fe) is the fourth most abundant element in the Earth's crust (about 6.7%; Rudnick & Gao, 2003) but it is present at sub-nanomolar concentrations in seawater (<10 −9 mol L −1 ; Johnson et al., 1997). Yet, Fe is a key micronutrient for the growth and metabolism of all living organisms and especially phytoplankton for which it is essential for the proper functioning of the photosynthetic system (Behrenfeld & Milligan, 2013;Raven et al., 1999). Consequently, Fe has a direct influence on primary production (Martin et al., 1994;Sunda & Huntsman, 1995) and thus plays an important role on carbon export and sequestration in the ocean interior (Martin, 1990). Numerous natural fertilization studies have investigated the importance of iron, primarily in the Southern Ocean (Blain et al., 2007;Pollard et al., 2007), and have reported enhanced primary production rates and particulate organic carbon export efficiencies, which may influence the biological carbon pump (Morris & Charette, 2013). In the context of climate change (IPCC, 2021), characterizing the elements governing the efficiency of this pump is of great interest. Due to its importance, the number of dissolved Fe (DFe) concentration measurements has increased impressively in recent years thanks to the GEOTRACES program (https://www. geotraces.org/), particularly in the deep ocean. However, there is still a lack of data for some key ocean regions, such as the Western Tropical South Pacific (WTSP) Ocean.The WTSP Ocean (160°E-160°W) has recently been identified as a hotspot of dinitrogen (N 2 ) fixation with some of the highest rates recorded in the global ocean (Bonnet et al., 2017). Diazotrophy is a process favored in phosphorus-rich, nitrogen-poor waters, that fuels the ocean with new nitrogen, helping to maintain ocean productivity and carbon sequestration (Caffin et al., 2018). This region is characterized by two biogeochemical

Shallow hydrothermal plumes of the Tonga-Kermadec arc are not transported over long distances, as previously reported for deep plumes.• Surface scavenging and photoreduction of stabilizing-complexes mediate the low spatial dispersion of shallow hydrothermal dissolved iron.• Nevertheless, the cumulative impact of multiple sources along the Tonga-Kermadec arc fertilizes the entire Lau Basin with dissolved iron.

Diazotrophs are often limited by iron (Fe) availability in the oligotrophic ocean. The Western Tropical South Pacific (WTSP) ocean has been suggested as an intense N2 fixation area due to Fe fertilizations through shallow hydrothermal activity. Yet, the Fe demand of diazotrophs in their natural habitat, where they cohabit with other microbial organisms also requiring Fe, remains unknown. Here we develop and apply a method consisting of coupling 55Fe uptake experiments with cell-sorting by flow cytometry, and provide group-specific rates of in situ Fe uptake by the microbial community in the WTSP, in addition to bulk and size fractionation rates. We reveal that the diazotrophs Crocosphaera watsonii and Trichodesmium contribute substantially to the bulk in situ Fe uptake (~33% on average over the studied area), despite being numerically less abundant compared to the rest of the planktonic community. Trichodesmium had the highest cell-specific Fe uptake rates, followed by C. watsonii, picoeukaryotes, Prochlorococcus, Synechococcus and finally heterotrophic bacteria. Calculated Fe:C quotas were higher (by 2 to 52-fold) for both studied diazotrophs compared to those of the non-diazotrophic plankton, reflecting their high intrinsic Fe demand. This translates into a diazotroph biogeographical distribution that appears to be influenced by ambient dissolved Fe concentrations in the WTSP. Despite having low cell-specific uptake rates, Prochlorococcus and heterotrophic bacteria were largely the main contributors to the bulk Fe uptake (~23% and ~12%, respectively). Overall, this group-specific approach increases our ability to examine the ecophysiological role of functional groups, including those of less abundant and/or less active microbes.

In the Western Tropical South Pacific (WTSP) Ocean, a hotspot of dinitrogen fixation has been identified. The survival of diazotrophs depends, among others, on the availability of dissolved iron (DFe) largely originating, as recently revealed, from shallow hydrothermal sources located along the Tonga-Kermadec arc that fertilize the Lau Basin with this element. On the opposite, these fluids, released directly close to the photic layer, can introduce numerous trace metals at concentrations that can be toxic to surface communities. Here, we performed an innovative 9-day experiment in 300 L reactors onboard the TONGA expedition, to examine the effects of hydrothermal fluids on natural plankton communities in the WTSP Ocean. Different volumes of fluids were mixed with non-hydrothermally influenced surface waters (mixing ratio from 0 to 14.5%) and the response of the communities was studied by monitoring numerous stocks and fluxes (phytoplankton biomass, community composition, net community production, N2 fixation, thiol production, organic carbon and metal concentrations in exported material). Despite an initial toxic effect of hydrothermal fluids on phytoplankton communities, these inputs led to higher net community production and N2 fixation rates, as well as elevated export of organic matter relative to control. This fertilizing effect was achieved through detoxification of the environment, rich in potentially toxic elements (e.g., Cu, Cd, Hg), likely by resistant Synechococcus ecotypes able to produce strong binding ligands, especially thiols (thioacetamide-like and glutathione-like compounds). The striking increase of thiols quickly after fluid addition likely detoxified the environment, rendering it more favorable for phytoplankton growth. Indeed, phytoplankton groups stressed by the addition of fluids were then able to recover important growth rates, probably favored by the supply of numerous fertilizing trace metals (notably Fe) from hydrothermal fluids and new nitrogen provided by N2 fixation. These experimental results are in good agreement with in-situ observations, proving the causal link between the supply of hydrothermal fluids emitted at shallow depth into the surface layer and the intense biological productivity largely supported by diazotrophs in the WTSP Ocean. This study highlights the importance of considering shallow hydrothermal systems for a better understanding of the biological carbon pump.

Iron (Fe) is an essential micronutrient for phytoplankton, particularly diazotrophs, which are abundant in the Western Tropical South Pacific Ocean (WTSP). Their success depends on the numerous trace metals, particularly iron, released from shallow hydrothermal vents along the Tonga Arc. This study aimed to explore the impact of hydrothermal fluids on particulate trace metal concentrations and biological activity. To identify the composition of sinking particles across a wide area of the WTSP, we deployed sediment traps at various depths, both close and further west of the Tonga Arc. Seafloor sediments were cored at these deployment sites, including at a remote location in the South Pacific Gyre. The sinking particles were composed of a large amount of biological material, indicative of the high productivity of the Lau Basin. A significant portion of this material was lithogenic of hydrothermal origin, as revealed through Al-Fe-Mn tracing. The sinking material showed similar patterns between lithogenic and biogenic fractions, indicating that hydrothermal input within the photic layer triggered surface production. A hydrothermal fingerprint was suggested in the sediments due to the high sedimentation rates and the presence of large, heterogeneous, trace metal-rich particles. The presence of nearby active deep hydrothermal sources was suspected near the Lau Ridge due to the large particle size and the significant enrichment of Fe and Mn. Overall, this study revealed that deep and shallow hydrothermal sources along with submarine volcanism have a significant influence on the biogeochemical signature of particles in the Lau Basin at large spatial and temporal scales.