The continuous expansion of oxygen-deficient zones (ODZs) poses risks to marine ecosystems and societies dependent on fisheries for income and sustenance. However, the trajectory of this deoxygenation in response to 21st-century climate change remains uncertain. To gain a clearer understanding of future oxygen dynamics and processes leading to deoxygenation, we investigate the response of ODZs during Cenozoic global warming periods in the Miocene and Early Eocene, using a combination of oxygen-sensitive proxies including foraminifer-bound nitrogen isotopes (FB-δ15N) and iodine-to-calcium ratios in planktic foraminifer shells (I/Ca). Our findings reveal contracted, rather than expanded, tropical Pacific ODZs during all studied warm intervals. The increased oxygenation closely aligns with high-latitude warming and reduced meridional sea surface temperature gradients, indicating a climatic driver behind these observed changes. We discuss potential causes for the contraction of ODZs, including (i) diminished wind-driven equatorial upwelling and primary productivity, and/or (ii) increased deep-ocean ventilation. Finally, we compare the behaviour of Pacific vs. Indian Ocean ODZs during the Miocene and investigate potential teleconnections between these two wind-driven ODZs.

Ocean’s oxygen (O2) is essential to most marine life forms and represent a fundamental component of the biogeochemical cycling of nitrogen and carbon. Its inventory is declining in response to global warming. Contrasting predictions about the future of the tropical oxygen deficient zones (ODZs) in numerical simulations and palaeoceanographic evidence for contracted ODZs during Cenozoic’s warmest periods, make long-term predictions about the future of ocean O2 challenging. We present new evidence for tropical ocean oxygenation during the Paleocene-Eocene Thermal Maximum (PETM), a rapid warming event that serves as a geologic analogue to ongoing warming. Foraminifera-bound nitrogen isotopes indicate that the tropical North Pacific ODZ contracted during the PETM, implying higher O2. Metabolic modelling of aquatic ectotherms shows that a concomitant increase in planktic foraminifera size implies that seawater oxygen partial pressure (pO2) rose in the shallow subsurface throughout the tropical North Pacific, beyond the spatial extent of the ODZs. These findings call for an oceanographic mechanism capable of both enhancing subsurface oxygenation and operating beyond the regional scale of the North Pacific ODZs, on millennial timescales. These divergent changes are consistent with Ocean General Circulation Models under SSP5-8.5 scenario for 2300, in which a decline in biological productivity allows tropical subsurface oxygen to rise even as global ocean oxygen declines. The tropical upper ocean oxygen increase may have relieved physiological stress, helping to avoid a mass extinction in planktic organisms during the PETM, in spite of the largest benthic extinction of the Cenozoic.

Upwelling generates a nutrient-rich "cold tongue" in the eastern equatorial Pacific Ocean (EEP), with impacts on global climate, oceanic biological productivity, and the carbon cycle. The cold tongue was reduced during the Pliocene Epoch, a feature attributed to weaker upwelling and an associated deepening of the surface mixed layer in the EEP. Here, we report nitrogen-isotope evidence that modern-like upwelling occurred in the EEP during the Pliocene and has persisted over the past 5 million years. We explain the reduced Pliocene cold tongue as an expression of the reduced temperature difference between surface and subsurface waters in the tropical Pacific. The attendant reduction in the vertical density gradient may have maintained EEP upwelling despite the expected slackening of the trade winds under Pliocene warmth.

The marine nitrogen (N) cycle plays a central role in regulating ocean productivity, ecosystem structure, and biogeochemical feedbacks on climate. However, our understanding of historical changes in the marine N cycle is limited by the short time span and spotty spatial coverage of direct observations. This study explores the use of crustose coralline algae (CCA) as a novel proxy for reconstructing past changes in the marine N cycle. We use the biomineral-bound nitrogen isotopic composition of tropical CCA (δ15NCCA) and compare it to the δ¹⁵N of subsurface nitrate (δ15N Nitrate) and corals (δ15Ncoral). Our results show that CCA reliably track subsurface nitrate δ¹⁵N and, therefore, can be used to reconstruct past changes in the N cycle. Additionally, anchoring δ15N coral values to co-occurring CCA allows us to isolate the reliance on photosymbionts in coral species, revealing taxonomic and regional variation in trophic strategies. Therefore, δ15N CCA also provides a tool to investigate coral trophodynamics in the modern ocean as well as in the historical record.

Biological fixation of dinitrogen (N2), the primary natural source of new bioavailable nitrogen (N) on Earth, is catalyzed by the enzyme nitrogenase through a complex mechanism at its active site metal cofactor. How this reaction functions in cellular environments, including its rate-limiting step, and how enzyme structure affects functioning remain unclear. Here, we investigated cellular N2 fixation through its N isotope effect (15 epsilon fix), measured as the difference between the 15N/14N ratios of diazotroph net new fixed N and N2 substrate. The value of 15 epsilon fix underpins N cycle reconstructions and differs between diazotrophs using molybdenum-containing and molybdenum-free nitrogenases. By examining 15 epsilon fix for Azotobacter vinelandii strains with natural and mutated nitrogenases, we determined if 15 epsilon fix reflects enzyme-scale isotope effects and, thus, N2 use efficiency. Distinct and relatively stable 15 epsilon fix values for wild-type molybdenum- and vanadium-nitrogenase isoforms (2.5 parts per thousand and 5.8-6.6 parts per thousand, respectively), despite changing cellular growth rate and electron availability, support 15 epsilon fix as a proxy for isoform type among extant nitrogenases. Structural mutation of active site N2 access altered molybdenum-nitrogenase 15 epsilon fix (3.0-6.8 parts per thousand for alpha-70VI mutant). Structure-function and isotopic modeling results indicated cellular N2 reduction is rate-limited by N2 diffusion inside nitrogenase due to highly efficient catalysis by the active site cofactor, exemplifying 15 epsilon fix as a tool to probe N2 fixation mechanisms. Diffusion-constrained reactions could reflect structural tradeoffs that protect the oxygen-sensitive cofactor from oxygen inactivation. This suggests that nitrogenase function is optimized for modern oxygenated environments and that pre-Great Oxidative Event nitrogenases were less diffusion-limited and potentially exhibited larger 15 epsilon fix values.

Recent studies show that water column denitrification, the main oceanic nitrogen sink, decreased under warmer than present conditions, but the response N 2 fixation, the main oceanic nitrogen source, remains unresolved. Here we address this gap using novel foraminifera bound δ¹⁵N records from the Western and Eastern tropical Atlantic spanning the warm Pliocene, which is considered a geological analog for predicted warming. Our records reveal that N 2 fixation was substantially reduced during the Pliocene, likely due to decreased excess phosphorous supply, driven by lower global water column denitrification. After ~ 2.8 Ma, as glaciations intensified, obliquity paced variability in N 2 fixation emerged, reflecting sea level driven changes in continental shelf area and benthic denitrification. These findings unveil N 2 fixation’s sensitivity to both long term as well as glacial interglacial scale changes, which is critical because reduce N 2 fixation under warmer conditions could influence the oceanic nitrogen balance and global ocean productivity.

The formation of the Isthmus of Panama closed the Central American Seaway, severing the only Late Cenozoic low‐latitude connection between the Pacific and Atlantic Oceans. Here we clarify the Early Pliocene (5.3–3.6 million years ago [Ma]) sequence of events associated with the shoaling of the Central American Seaway based on differences in upper ocean biogeochemical properties between the eastern tropical North Pacific (ETNP) and the Caribbean Sea. Foraminifera‐bound nitrogen isotopes (FB‐ δ 15 N) are elevated in the ETNP relative to the Caribbean Sea throughout the Early Pliocene. Whereas ETNP FB‐ δ 15 N shows no long‐term trend across the Early Pliocene, FB‐ δ 15 N in the Caribbean Sea declines by ∼0.5‰ between 4.6 and 4.5 Ma, and by an additional ∼1‰ between 4.35 and 4.25 Ma. We interpret the divergence between ETNP and Caribbean Sea FB‐ δ 15 N to indicate progressive isolation of their subsurface nutrient pools due to CAS shoaling. The oxygen isotopic composition of seawater ( δ 18 O sw ) derived from planktonic foraminifer δ 18 O and Mg/Ca shows a small but variable gradient between the ETNP and Caribbean Sea over the Early Pliocene, with a trend toward a larger δ 18 O sw gradient after 4.25 Ma. We suggest that the development of persistent chemical differences in both thermocline nutrients and surface waters between the ETNP and Caribbean Sea after 4.1 Ma reflects the cessation of basin‐scale oceanic exchanges across the Central American Seaway.

The nitrogen (N) isotopic composition of coral tissue provides insight into N sources and cycling on reefs, and coral skeleton-bound organic matter (CS-δ15N) can extend these insights into the past. Across the Bermuda platform, we measured the δ15N of four coral species and their potential N sources, as well as an asymbiotic filter feeder as a comparative heterotroph and benthic macroalgae as a comparative autotroph. Organisms and organic N pools from the coral reefs exhibit a δ15N increase toward the Bermuda coast, likely due to anthropogenic N inputs. At all sites, the δ15N of bulk coral tissue is consistent with corals feeding dominantly on zooplankton-sized organic matter and some smaller suspended particulate N. The corals lack the trophic δ15N elevation that characterizes serpulids; this is consistent with internal recycling and retention of low-δ15N metabolic N by symbiont-bearing corals. The data are inconsistent with corals’ reliance on the dissolved inorganic N used by macroalgae at the same sites. Among coral species, two species with smaller polyps (1-2 mm) have ~1‰ lower bulk tissue δ15N than two counterparts with larger polyps (5-10 mm), perhaps due to differences in food source. Taxon-specific δ15N differences are also observed between coral tissue and skeleton-bound N, with larger differences in the two small-polyp species. In net, however, CS-δ15N mean values and spatial gradients were similar in the four species studied.

The biogeochemical processes shaping the fluxes of nitrogen (N) in the tropical to subarctic Pacific are illuminated by nitrate isotope ratios (δ 15 N and δ 18 O) along the GEOTRACES GP15 section. In the equatorial and tropical Pacific, nitrate δ 15 N and δ 18 O are 2–4‰ higher in the thermocline than in deeper waters. This widespread elevation is driven by both nitrate assimilation in Southern Ocean surface waters and denitrification in the eastern tropical Pacific. In addition to this, a poleward increase in the δ 15 N of surface nitrate from the Equator to 10°S is generated by the progressive consumption of nitrate upwelled at the Equator and transported southward. This process leads to increases in the δ 15 N of phytoplankton biomass and a strong meridional nitrate isotopic gradient in the thermocline due to regeneration of sinking N. North of the Equator, an analogous gradient is barely detectable due to weaker northward flow and a compressed spatial scale for nitrate drawdown. By contrast, between 5°N and 40°N, the nitrate isotope gradients are dominated by the processes driving the oceanic fixed N budget: denitrification and N 2 fixation. High‐δ 15 N and ‐δ 18 O nitrate in the tropics is advected from the oxygen deficient zone where denitrification occurs, whereas in the adjacent subtropics, low‐δ 15 N thermocline nitrate suggests a response by N 2 fixation that is not observed in the south. The asymmetry of the north and south tropical Pacific gradients in thermocline nitrate isotopes has implications for efforts to reconstruct the N cycle in the past.

Hypoxia is observed and projected to expand in many aquatic environments, largely due to excess anthropogenic nutrient inputs and climate change, thus influencing biogeochemical processes. Denitrification, generally an anaerobic process, removes bioavailable nitrogen and produces nitrous oxide (N2O). However, limited observations of the effect of oxygen on denitrification restrict our ability to estimate changes in the amount of bioavailable nitrogen and N2O emissions under anthropogenic perturbations and climate change. Here, we show that all denitrification steps increased, while the N2O production yield from denitrification decreased with decreasing oxygen in Chesapeake Bay - the largest estuary in the United States. The different steps of denitrification responded similarly to oxygen changes in Chesapeake Bay, unlike open ocean oxygen minimum zones, with implications for the accumulation or depletion of denitrification intermediates such as nitrite and N2O. Our observations also suggest that current model parametrizations of denitrification in Chesapeake Bay likely overestimate denitrification and nitrogen removal in the presence of oxygen, which would bias the evaluation of nutrient cycling, ecosystem productivity, and the extent of hypoxia. Overall, our newly derived oxygen sensitivities of denitrification could be used to improve model parametrizations of denitrification and constrain the nitrogen budget and N2O emissions in estuarine and coastal environments experiencing hypoxia.

Incorporation of animal-based foods into early hominin diets has been hypothesized to be a major catalyst of many important evolutionary events, including brain expansion. However, direct evidence of the onset and evolution of animal resource consumption in hominins remains elusive. The nitrogen-15 to nitrogen-14 ratio of collagen provides trophic information about individuals in modern and geologically recent ecosystems (<200,000 years ago), but diagenetic loss of this organic matter precludes studies of greater age. By contrast, nitrogen in tooth enamel is preserved for millions of years. We report enamel-bound organic nitrogen and carbonate carbon isotope measurements of Sterkfontein Member 4 mammalian fauna, including seven Australopithecus specimens. Our results suggest a variable but plant-based diet (largely C 3 ) for these hominins. Therefore, we argue that Australopithecus at Sterkfontein did not engage in regular mammalian meat consumption.

The regions near the Antarctic Peninsula in the Southern Ocean are highly productive, with notable phytoplanktonic blooms in the ice-free season. The primary productivity is sustained by the supply of nutrients from convective mixing with nutrient-rich subsurface waters, which promotes rapid phytoplankton growth as the sea ice melts in spring and summer. Surface waters are marked by the contrast between the warmer Drake Passage and the colder Weddell Sea, and seasonal duration of sea ice cover varies accordingly. Sea ice exerts multiple controls over primary production, by shading the light entering the ocean and stratifying the upper ocean with freshening by ice melt. However, the interaction between sea ice and productivity remains poorly characterized because satellites are unable to quantify biomass in partially ice-covered ocean, and direct measurements are too scarce to characterize the seasonally varying productivity. Here we evaluate productivity by assessing removal of nitrate from surface waters by biological nutrient utilization and study the associated change in δ15N of nitrate. We use a combination of bottle samples and in situ nitrate measurements from published databases, completed by two transects with isotopic measurements. The timing of sea ice melt date conditions the initiation of nitrate drawdown, but the annual minimum of nitrate only weakly correlates with sea ice concentration. As previously reported, we observe that δ15N of nitrate increases with nitrate depletion. Interestingly, the lowest nitrate depletion and δ15N values are found in the central region of N–S transects, where intermediate temperature and sea ice conditions prevail. Deeper mixing in waters that passed through the northern Bransfield Strait may explain higher nitrate concentration due to both a greater nitrate concentration at the beginning of the growth season and reduced productivity under light limitation in deeply mixed waters, confirmed by nitrogen isotope modeling. This highlights the importance of oceanographic controls on productivity patterns in sea ice regions in the Southern Ocean.

Previous studies have reported a nitrate 15 N depletion in the Mediterranean Sea compared to the global ocean, attributed to either N 2 fixation or atmospheric deposition of anthropogenic N. In this study, we report basin‐wide full‐depth profiles of nitrate δ 15 N (vs. Air) and δ 18 O (vs. Vienna Standard Mean Ocean Water, VSMOW) in the Mediterranean Sea. Our results confirm a consistent 15 N depletion across the entire Mediterranean Sea, with significantly lower nitrate δ 15 N values in the eastern basin (2.2 ± 0.2‰) than in the western basin (2.9 ± 0.1‰). In contrast, there is no significant difference in nitrate δ 18 O between the two basins (2.2 ± 0.3‰ and 2.1 ± 0.2‰, respectively). These observations point to a supply of low‐ δ 15 N N to the Mediterranean Sea, accumulating as regenerated nitrate, which is diluted by the nitrate in the Atlantic inflow, creating an west‐to‐east gradient in nitrate δ 15 N. A four‐box model reveals that, given a water residence time of 120–170 years in the Mediterranean, a modest input rate of 1–3 Tg N yr −1 —originating from N 2 fixation, atmospheric deposition, or their combination—is adequate to produce the observed low δ 15 N of Mediterranean nitrate. Additionally, partial degradation of dissolved organic nitrogen imported from the Atlantic may add low‐ δ 15 N nitrate to the Mediterranean, but it alone cannot explain the full isotopic signal. Distinguishing among these sources will be aided by the reconstruction of Mediterranean nitrate δ 15 N through time using either time‐series data of nitrate δ 15 N or calcareous fossil‐bound organic nitrogen isotope ratios.

In the Antarctic Zone (AZ), deep nutrient-rich waters ascend to the surface, feeding the Southern Ocean's overturning circulation cells. However, the rate of upwelling exceeds the capacity of phytoplankton to fully consume the gross nutrient supply to the AZ surface, leading to the release of previously sequestered CO2 into the atmosphere. During ice ages, enhanced nutrient utilization has been proposed as a mechanism that could contribute to lower atmospheric CO2 concentration. Fossil-bound δ15N records in the AZ point to a more complete nitrate consumption in surface waters during ice ages. This increase in nitrate utilization coincides with reduced export production, suggesting a lower gross nitrate supply to the surface and, therefore, a reduction in the exchange of water between the surface and the deep ocean. Preliminary reconstructions indicate more than a 5-fold reduction in the rate of gross nitrate supply to match paleo proxy data and near complete nitrate consumption at the surface. Model simulations are ambiguous, but none show more than a ≥ 2-fold reduction in water exchange in the AZ during ice ages.One hypothesis for this discrepancy is the progressive depletion (“mining-out”) of nutrients from the AZ upper ocean. Reduced glacial upwelling, combined with repeated summer nitrate consumption and the export of assimilated nitrate as sinking organic matter, followed by deep winter mixing, could gradually deplete the upper water column’s nutrient reservoir. This process would lower the shallow subsurface nutrient concentrations and elevate nitrate δ15N relative to the deep ocean. As a result, the nutrient supply per volume of upwelled water would decline, aligning better with model simulations. To test this hypothesis, we developed a 1D advection-diffusion-reaction model of the water column, accounting for surface nitrate consumption and isotope fractionation. The model was calibrated using Argo floats data and high-resolution hydrographic nitrate isotopes transect in the AZ (GO-SHIP SO4P 2018), successfully matching depth and seasonal profiles. We also applied the model to the western subarctic Pacific, which exhibits a similar observation pattern for fossil-bound δ15N and export production during ice ages but contrasts in the ratio between advective and diffusive nutrient supply.Our results highlight the critical role of nutrient mining in driving isotopic changes during ice ages. With reduced upwelling, nutrients are progressively depleted in the upper AZ. However, even under this mechanism, a substantial reduction in upwelling (more than a twofold decrease) is still required to achieve observed glacial δ15N values – though less extreme than previous estimates. Nevertheless, in reduced upwelling scenarios, the glacial surface nitrate concentration is significantly higher than previous estimates. This supports the potential of nutrient mining in matching paleo-data with less drastic changes to the Southern Ocean.