Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

On the natural spatio-temporal heterogeneity of South Pacific nitrous oxide

Abstract

Nitrous oxide (N2O) is a powerful greenhouse gas and ozone depleting substance, but its natural sources, especially marine emissions, are poorly constrained. Localized high concentrations have been observed in the oxygen minimum zones (OMZs) of the tropical Pacific but the impacts of El Niño cycles on this key source region are unknown. Here we show atmospheric monitoring station measurements in Samoa combined with atmospheric back-trajectories provide novel information on N2O variability across the South Pacific. Remarkable elevations in Samoan concentrations are obtained in air parcels that pass over the OMZ. The data further reveal that average concentrations of these OMZ air parcels are augmented during La Niña and decrease sharply during El Niño. The observed natural spatial heterogeneities and temporal dynamics in marine N2O emissions can confound attempts to develop future projections of this climatically active gas as low oxygen zones are predicted to expand and El Niño cycles change.

Introduction

Nitrous oxide (N2O) is a gas of significant importance to current and future climate change. As both a potent greenhouse gas1 and substantial agent of stratospheric ozone destruction2, it serves a dual role in regulating climate. Its natural sources are largely microbial: N2O is a product of the nitrification and denitrification metabolisms, occurring throughout both terrestrial and marine ecosystems. Atmospheric N2O concentrations are currently growing, mainly because of artificial fertilizer application3. The secondary effects of global change, from warming oceans and circulation changes, could further increase N2O concentrations. Heterogeneity in sources and sinks on land and in the ocean makes constraining the global budget difficult. The ocean source exhibits a large degree of spatial structure, with N2O production being potentially concentrated in the confines of the oxygen minimum zones (OMZs) of the eastern tropical Pacific and Arabian Sea4, because N2O production from both microbial metabolisms is amplified under low oxygen conditions. This localization to remote areas impedes the acquisition of directly-measured data required to understand the system. Correct attribution of the dynamic sources and their mechanistic controls is key to interpreting observations and predicting future feedbacks on climate.

Global estimates of marine N2O emissions span a wide range due to the dearth of direct measurements at the sites of significant production within the water column. The most recent 5th IPCC Assessment Report5 suggests magnitudes for total natural emissions from soils, ocean, and atmosphere between 5.4 and 19.6 Tg N of N2O yr−1. Estimates for oceanic emissions alone range from 1.8 to 9.4 Tg N yr−1, with the upper estimate increasing with each subsequent IPCC report as more measurements are made and higher precision models developed. Much of the uncertainty regarding the marine N2O efflux derives from the ability of localized highly productive, and by consequence, suboxic waters to dominate the net budget.

In the global ocean, N2O is generated much as it is on land: certain groups of microorganisms harness inorganic forms of nitrogen for energy and release N2O as an inefficient byproduct. One of these sets of organisms–the nitrifiers–oxidize ammonium to nitrite and nitrate released through respiration, of which a small fraction, ~1 out of 10,000 molecules6,7 is typically emitted as N2O. This minor global marine pathway may be swamped by production localized to the OMZs. When oxygen (O2) concentrations are highly reduced, as occurs in the OMZs, the production pathways diversify with dramatic effects on N2O production. The efficiency of nitrification plummets as O2 is consumed, observed both in cultures6 and in the environment8, increasing the N2O yield to as much as 1 molecule in 50. In addition, at very low O2 concentrations, another microbial pathway–denitrification–is enabled, which can rapidly produce N2O as an obligate metabolic intermediate9. The high productivity stimulating N2O production leads to shallow depths of maximal concentration, and thereby primes these regions with the potential to disproportionately impact the global N2O budget well beyond the minor basal term.

OMZs are linked to strong wind-induced upwelling, which brings nutrient rich water from depth to the surface10, fueling high rates of primary productivity. When this newly fixed organic matter sinks into the ocean interior, it enhances biological respiration and reduces O2 concentrations to below 0.01% saturation11,12 at shallow intermediate depths of ~100–500 m. Slow lateral ventilation of these regions restricts resupply of O2, resulting in permanently suboxic zones and the sustenance of anaerobic microbial metabolisms13. These waters tend to achieve their lowest O2 concentrations along a fairly conserved surface-normalized density layer14,15,16, of 1026.5 kg m−3 (Fig. 1a). Although marine measurements of N2O are still fairly limited, studies have found OMZs to harbor the greatest N2O concentrations and fluxes by far. N2O supersaturations by as much as 100-fold above atmospheric equilibrium in the Eastern Tropical South Pacific (ETSP) OMZ have been observed off the coast of Peru, with estimates that this region alone could account for nearly a quarter of global marine N2O emissions17. Similar measurements and flux estimates exist for the analogous Eastern Tropical North Pacific18 and Arabian Sea19, with comparably high surface supersaturations and atmospheric emissions20.

Fig. 1: Near-continuous nitrous oxide data across the Pacific basin.
figure1

a Spatial map of 15-day back trajectories from the four Pacific monitoring stations analyzed in this study, overlain on dissolved oxygen concentrations at the 1026.5 kg m−3 density horizon, with the 10 and 20% saturation contours displayed (dashed blue lines). A random selection of ten trajectories each from Hawaii, Tasmania, and California-originating back-trajectories are shown in gray, and ten randomly selected Samoa back-trajectories in color. Squares represent the positions of the trajectories 15 days prior to arrival at the Samoa station. Only the Samoa station back-trajectories frequently intersect the boundaries of the oxygen minimum zones. b De-trended time series of atmospheric N2O concentrations measured at the Samoa station, split among austral summer (orange) and winter (purple) along with a mean annual oscillation (black). This seasonal cycle is driven largely by the southward shift of the intertropical convergence zone in January bringing the influence of higher northern hemisphere concentrations to the south. c De-seasonalized N2O concentrations (light gray dots, with monthly means represented by the dark gray line) compared with a 3-month forward shifted Niño 3.4 index (orange) and a running integrated Niño 3.4 index (green). All time series are shown normalized. The Niño index time series is flipped vertically to facilitate visual comparison with N2O, i.e., high N2O aligns with negative La Niña excursions. The N2O concentrations align well with both ENSO metrics, with p < 0.0001.

The remoteness of these OMZs has hampered their direct study. Unsurprisingly then, the specialized processes of suboxic environments have yet to be fully incorporated into climate models and global budgets. Further, while OMZ regions are small on a global scale (accounting for only ~0.1% of ocean volume), the surrounding hypoxic waters are much larger (10% of ocean volume)4. In these hypoxic regions on the borders of OMZs where O2 is low, N2O production peaks at depths close to the surface18, between 50 and 200 m, which stand out as hotspots of N2O concentrations at depth14,21. The N2O produced by basal nitrification globally achieves maximum concentration at deeper depths14, between 500 and 1000 m, and is insulated from rapid atmospheric communication, even if nitrification rates peak shallower in the ocean22. As such, N2O produced in the mesopelagic deep would not be expected to display much spatial heterogeneity. In contrast, the shallower and steeper subsurface N2O maxima in the OMZs combine to enable rapid diffusive mixing into the surface ocean and establish these regions as hotspots of N2O efflux to the atmosphere. Critically, in OMZs, N2O exhibits low residence times and rapid turnover rates, permitting high variability, both over seasonal and interannual scales18,23,24 and as the OMZs expand into the future25,26. This feedback among OMZ size, N2O emissions, and climate emphasizes the importance of quantifying the N2O production of these regions.

Here we take a complementary approach to investigate the role of the OMZs in N2O emission compared with the open ocean, not by directly visiting these regions but by remotely analyzing their effect on atmospheric concentrations. Due to the short time scale of tropospheric transport relative to the ocean, these monitoring stations need not be located directly within or adjacent to the region of study. In so doing, we map N2O anomalies across the Pacific and, because of the time series nature of the dataset, investigate the interannual variability in association with El Niño and La Niña.

Results and discussion

AGAGE data for marine N2O studies

The network of Advanced Global Atmospheric Gases Experiment (AGAGE) stations has been quantifying greenhouse gas levels including N2O since the late 1970s27. This global network now produces high frequency, high precision measurements; the data improved substantially in the mid-1990s with the incorporation of higher precision methodologies28. N2O data from the last ~20 years are measured every 40 min (Fig. 1b, Supplementary Figs. 1 and 2) and have a precision on any individual measurement of 0.1 ppb but much higher confidence for the ensemble28,29. Combining these measurements with atmospheric back trajectories generated by the HYSPLIT4 model from monitoring stations (Fig. 1a, Supplementary Data 1), we map the relationship between atmospheric N2O concentrations and marine O2 content. Similar analyses have previously been conducted to investigate recent continuing sources of chlorofluorocarbons from land30. The marine environment is dynamic, and time-resolution is difficult to sample for many biologically-active chemical parameters requiring in situ point measurements. The AGAGE dataset, however, does not suffer the same restrictions, and the high-frequency measurements can be utilized to investigate variability in detail, including the important interannual dynamics that arise due to the influence of El Niño–Southern Oscillation (ENSO) on these regions.

Land sources could obscure the attribution of marine N2O, so all back trajectories that intersect continents over a 20-day model run are excluded in our analysis. This rigid protocol limits the choice of stations and marine areas. Multiple stations were considered for analysis, but eliminated as being ill-suited for identifying features in the OMZs. The land and monsoonal influence in the Arabian Sea obscures any marine signal in that OMZ. Similarly, the available high precision North Pacific atmospheric measurement stations are located along the California coast and in Hawaii, and do not communicate with the eastern tropical North Pacific OMZ before mixing homogenizes any advected signal (Supplementary Fig. 3). The station at American Samoa, however, permits excellent monitoring of the ETSP. The easterly trade winds rapidly transport air parcels intersecting the region of low O2 across the Pacific (scale of days to weeks) before turbulent mixing with other air masses is able to fully mask the OMZ signal (Fig. 2). Moreover, along this trajectory, air tends to remain in the troposphere (Supplementary Fig. 4), minimizing stratospheric impacts on surface conditions. The raw N2O data from Samoa, when de-trended for the long-term anthropogenic effect and de-seasonalized to minimize the intra-annual oscillation driven largely by interhemispheric transport from the north as the Intertropical Convergence Zone moves southward during boreal winter (Supplementary Figs. 1 and 2), reveal striking relationships with the ocean biogeochemical state (Fig. 3).

Fig. 2: Back trajectory of Samoa atmospheric N2O.
figure2

Locations of air parcels with their de-trended nitrous oxide measurements at Samoa are plotted (a) 5, (b) 10, (c) 15, and (d) 20 days prior to arrival at Samoa. De-trended nitrous oxide measurements are represented by the color of the corresponding dot, as anomalies relative to the station mean. The data maintain a clear spatial delineation for the full 20-day period, with much higher concentrations passing over the eastern tropical Pacific, and much lower concentrations arriving from the west and the Southern Ocean. Overlain on this map are the 10 and 20% oxygen saturation horizons at the 1026.5 kg m−3 potential density level (black contours). The orange circle indicates the location of the Samoa station.

Fig. 3: Gridded N2O anomalies relative to the Samoa mean.
figure3

De-trended and de-seasonalized N2O concentrations from Samoa (color) are gridded based on back-trajectory locations 15 days prior, and a mean value is calculated for air passing over each 5° grid cell at that time (color). Overlain on this map are the 10 and 20% oxygen saturation horizons at the 1026.5 kg m−3 potential density level (black lines). These two data sets come from entirely independent sources, and the coincidence between them is striking; high atmospheric N2O concentrations align with the oxygen minimum zone, extending westward into the tropical Pacific and down the Chilean coast of South America. The orange circle indicates the location of the Samoa station.

The seasonal modulation in Samoa N2O has been well-established as an effect of northern influence. However, across the Pacific sites, trajectories remain mostly well-confined to the hemispheres in which they originated (Supplementary Fig. 5). The air parcels that reach Samoa from the northern hemisphere however, do not have markedly higher N2O concentrations than their southern counterparts. Many high concentration trajectories do travel across the Isthmus of Panama from the Caribbean (Supplementary Fig. 6), but these are filtered out by the land filter. Notably, 20 days prior there are still a large number of high concentration trajectories in the southern hemisphere, which travel northward off the coast of South America before transiting to Samoa. Indeed, very few trajectories actually cross over the South American continent; most are deflected either north or south along its western edge. As a result, land emissions of N2O from South America are unlikely to have a large impact on the visible spatial signal.

Oxygen minimum zones emerge as N2O hotspots

A divergence in concentrations is visible just one-day prior, with higher concentrations arriving from the north than from the south (Supplementary Fig. 3). This pattern remains moving further back in time; the highest concentrations pass over the eastern tropical Pacific, while the lowest come from the west passing over the Southern Ocean. The pattern becomes slightly less defined between 15 and 20 days-prior, as parcels with high concentrations diverge again along the western coasts of North and South America. Given the velocity of the trade winds and the distance between Samoa and the ETSP, ~8 m s−1 and 104 km, respectively, this time scale is reasonable. Certainly the observed atmospheric concentration anomalies of N2O at Samoa are a combined result of intra-hemispheric biological production with interhemispheric transport, but the analysis presented herein aims to tease apart the biological impact from the tropical Pacific ocean from the additional seasonal supply from the N2O-enriched northern hemisphere (Supplementary Fig. 7).

In defining the OMZ boundary by the 20% O2 saturation contour on the 1026.5 kg m−3 potential density surface, the 15-day back trajectories identify higher than average N2O among the air parcels passing over the OMZ (Fig. 3). This isopycnal layer, generally at depths of ~200–500 m (shallower in the eastern Pacific), tends to co-locate the minimum O2 concentration within the water column31. Whereas water this deep does not tend to equilibrate rapidly with the atmosphere, this density surface reflects broader features in the water column, such as the existence of a shallower N2O maximum driven by microbial processes18. This shallower N2O maximum can be more easily transmitted to the surface waters from the interplay of upwelling and lateral and vertical mixing32. Mesoscale and sub-mesoscale eddies in particular, below the spatial resolution that can be resolved by these data, are likely pathways that promote mixing of deeper N2O into the surface waters33,34.

The air masses reaching Samoa that pass over the OMZ are ~0.4 ppb higher than the remainder of the South Pacific. This intense localization to the lowest O2 waters in the eastern tropical Pacific, extending southward along the Peruvian and Chilean margin, suggests a disproportionate role of OMZs and coastal upwelling waters in generating atmospheric N2O relative to the open ocean. Extension of higher N2O anomalies along the equator outside of the 20% O2 contour also remain visible west to 140° W (Fig. 3), albeit with reduced magnitude, indicative of a separate likely N2O source from equatorial upwelling and subsequent outgassing35,36. Further, the air overlying the coastal Pacific along the South American upwelling region is especially high in its N2O content, indicative of the heterogeneities that exist even within suboxic waters17. The coastal-most waters are the most productive and the resulting organic matter fuels greater microbial generation of N2O. Moreover, the depth of the anoxic onset and coincident N2O maximum is generally shallower toward the coast, increasing the communication between ocean and air. Narrow, shallow shelves within the OMZs in particular are important producers of N2O17,23,37,38, but local sources with such spatial resolution cannot be identified via our methods.

While the absolute difference between N2O mixing ratios in air that has passed over the OMZ versus those that did not is small, the abundances above the OMZs are likely much higher but are diluted to the observed anomalies after 15 days of transport and mixing. Notably, the air parcels will continue to gain N2O as they travel over supersaturated water masses or lose N2O over undersaturated areas. The coherence of the 15-day backtracked Samoan N2O signal and its clear demarcation with the ETSP highlight the significance of the OMZs for the marine N2O efflux. Indeed, the offset between the air passing over the OMZ grid cells and the rest of the South Pacific matches the average seasonal variability in the Samoan data. The seasonal cycle, which includes terrestrial sources and is heavily influenced by seasonal shifts in the intertropical convergence zone, peaks in Austral winter (January/February) with minima 0.50 ± 0.13 ppb (se) lower in summer (July/August). The rapid N2O cycling rates inherent to the highly productive OMZs cause these regions to respond quickly to time-dependent shifts in the underlying biogeochemistry and physical structure of the water column18,24.

Further modulation by El Niño and La Niña

It is known that ETSP biogeochemistry is influenced by the El Niño/La Niña oscillations of the tropical Pacific39,40. La Niña brings enhanced upwelling of nutrients to the surface, thereby increasing primary production, further de-oxygenating the subsurface region, and supplying more organic matter to drive greater denitrification and N2O production13. Conversely, during an El Niño decreased upwelling reduces the surface productivity in the eastern tropical Pacific and deepens the oxycline, thereby contracting the OMZ and decreasing N2O emissions. In this study, the near-continuous data over 21 years permits a more quantitative analysis. The period from 1996 through 2016 included both a number of strong El Niño (1997/98, 2002/03, 2009/10, 2015/16) and La Niña (1998/99, 2007/08, 2010/11) states. Our analysis consists of dividing the N2O dataset among ENSO states in two different ways, both of which produced equivalent results: First, by aligning the N2O oscillation against the Niño 3.4 index, resulting in a time lag of 3 months relative to the ENSO state; second, by producing a running integral of the Niño 3.4 index forward in time, which essentially represents a cumulative effect of prolonged El Niño/La Niña periods (Fig. 1c). The Niño 3.4 index describes a 5-month running mean of sea surface temperature anomalies in the central equatorial Pacific, and was used because it is indicative of the broader Pacific region.

Strikingly, during La Niña, air parcels passing over OMZ grid cells increase on top of their already higher than average concentrations, whereas parcels that do not pass over OMZ grid cells exhibit no systematic change (Fig. 4). This increase with La Niña is not uniform across the OMZ, however, as the lowest O2 waters (<10% saturation) increase by approximately one-third while those of hypoxia (<20% saturation) increase during a strong La Niña by more than double compared with neutral conditions (Table 1, Fig. 5). This occurrence implies the lowest O2 waters already exhibit their maximal production whereas the slightly more oxygenated waters on the periphery have potential for lower O2 states with La Niña to augment their N2O production. This decrease in O2 can amplify both nitrification-based production as well as expand the zone where denitrification-sourced N2O is permitted. During El Niño, while the parcels not passing over the OMZ do not change systematically from their mean state, those passing over the OMZ consistently see reduced N2O concentrations relative to their means (Table 1, Fig. 4) and lower than the long-term Samoan average. In all, the lowest O2 waters are by far the most variable temporally, subject to short-scale dynamics that affect these waters more than the open ocean. Such interplay confirms the importance of OMZs, with their shallow N2O maxima and fast generation times18 to marine N2O efflux. The cutoff of coastal upwelling associated with El Niño and the corresponding dramatic decline of primary productivity supplying organic matter to the interior curtail N2O production and eliminate the unique conditions required for OMZ-based production. The effects are even more pronounced during the strongest of El Niño and La Niña events, including the recent 2015-2016 El Niño, with greater shifts from the mean state during the most intense periods (Fig. 5). The Pacific-wide decrease in N2O observed in the AGAGE back-trajectory analysis is in line with ship-based observations from the eastern41, central35, and western42 Pacific basins.

Fig. 4: El Niño/Southern Oscillation effects on gridded N2O.
figure4

The time series was split among three states: La Niña, with a Niño 3.4 index < −0.5; Neutral, with an index between −0.5 and 0.5; and El Niño, with an index > 0.5. New gridded maps were calculated for each, by averaging only data points within the corresponding El Niño/Neutral/La Niña states. a The change in each grid cell’s mean value from a Neutral state to a La Niña state. b The change from Neutral to El Niño. During La Niña, the eastern tropical South Pacific oxygen minimum zone (OMZ) displays heightened N2O in air that has passed over the water with no consistent effect in air that has passed over non-OMZ waters. During an El Niño, the concentrations in air that has passed over the OMZ are substantially reduced with again no systematic effects observed in air that has not passed over the OMZ.

Table 1 De-seasonalized 15-day back trajectory nitrous oxide anomalies for select spatio-temporal regimes relative to the Samoan all-time mean.
Fig. 5: N2O anomalies depend on dissolved oxygen and El Niño state.
figure5

Climatological oxygen concentrations are at the density level of 1026.5 kg m−3 and El Niño/Southern Oscillation (ENSO) state is divided by neutral (|index| < 0.5), weak (0.5 < |index| < 1.0), and strong (|index| > 1.0) states. N2O concentrations were separated by ENSO states and gridded as in Figs. 3 and 4, then means were calculated for grid cells of each O2 saturation range. N2O in air that has passed over the oxygen minimum zones with O2 < 20% saturation is greater than average during neutral and La Niña states, with highest N2O values at lowest % O2 saturation. Air parcels that have passed over regions with lowest % O2 saturation have nearly equal N2O values during strong La Niña and neutral states, but parcels that have passed over regions with 10–20% saturation exhibit substantial changes in N2O between the two states. Warmer surface waters across the Pacific (El Niño states), coincide with reduced N2O, reducing more substantially with stronger El Niño conditions, especially evident among the 0–20% O2 saturation waters which even show a sign change in anomaly. This El Niño effect is likely due to increased stratification of the water column isolating the deeper N2O rich waters from the surface. Shown are the mean ± standard error for grid cells within a bin.

ENSO impacts atmospheric circulation patterns that might obscure attribution of the observations to differences in marine production and efflux29,43,44. Notably, N2O concentrations are higher in the northern hemisphere because of greater terrestrial and anthropogenic production. Similarly, sulfur hexafluoride (SF6) is a compound with comparable asymmetry in hemispheric concentration because it is an industrially produced volatile compound linked with anthropogenic emissions overwhelmingly in the northern hemisphere. Indeed, SF6 has an even greater interhemispheric gradient and a steeper growth rate than N2O. In repeating our analysis for SF6 for which AGAGE data at Samoa are available (2005 onwards), no apparent skew toward O2 deficient waters (Supplementary Fig. 8) or ENSO-rooted signal arises (Supplementary Figs. 9 and 10), suggesting that the spatio-temporal relationship in N2O concentrations is tied to autochthonous OMZ production rather than hemispheric exchange. In completing our 15-day back trajectory analysis, very few tracks cross over the inter-tropical convergence zone, largely keeping equatorward surface winds isolated in their respective hemispheres (Supplementary Figs. 3, 6 and 7). Still, during an El Niño, there is less influence of mixing from the northern hemisphere and increased south-easterly winds, resulting in a relative decrease in atmospheric N2O across the southern hemisphere45. Indeed, this broad decrease is observed in the data (Figs. 4 and 5). Increased inter-hemispheric exchange during La Niña events should increase concentrations across the southern hemisphere, but the observed changes are to a great extent localized to air passing over the OMZs. The localization again indicates the dominance of the local biological production term over perturbations in atmospheric transport. Impacts of ENSO are not limited to the Pacific Ocean; the inter-annual variability also causes changes to terrestrial emissions of N2O46,47. Impacts of ENSO on marine biogeochemistry can be substantial sources of this interannual variation in atmospheric N2O45. Importantly, over the period of 1999–2009, El Niño periods were indeed associated with low marine N2O fluxes and La Niña events with high anomalies46.

Summary

In developing models capable of predicting N2O under future climate scenarios, proper attribution of the natural sources and variability of this gas is of significant importance. The effects on these regions of anthropogenic climate change, and whether N2O from the OMZs indeed will increase into the future, remain open questions. Our results indicate that the variability of oceanic sources linked to OMZs can feedback substantially on the climate system as low O2 areas are predicted to expand25,26 and ENSO cycles change in intensity and frequency48. The enormous potential of OMZs to produce N2O and their direct response to El Niño/Southern Oscillation forcing, necessitates their explicit inclusion in N2O budget calculations and better monitoring of these regions from the eastern tropical Pacific to quantify the precise effect on N2O efflux before mixing reduces the local signal. The simple model presented here cannot precisely map or quantify these emissions because it does not take into account mixing or integrative effects along the transects. We also cannot directly attribute the role of transient sub-mesoscale eddies on enhanced N2O production33,49,50. While not individually observable from this dataset, such effects are encompassed in the ensemble average over the 21 years of measurements. Nonetheless, the high-frequency data from Samoa reveal not only that the ETSP is a large source of atmospheric N2O, but also that it is subject to such large changes in short-term budgets by ENSO that they are visible more than half-way across the Pacific Ocean.

Methods

Atmospheric data from AGAGE

Very few N2O data exist from oceanographic cruises, floats, or other ocean-going platforms, particularly from the OMZ regions of interest51. However, the AGAGE, maintains a network of stations around the world that have been monitoring atmospheric gases, including N2O, since the late 1970s. The analysis herein is limited by the locations of these stations and the frequency and precision of their measurements. Several stations around the Pacific Ocean are relevant for these inquiries into OMZs and N2O production: Cape Matatula in American Samoa, Trinidad Head in California, and Cape Grim in Tasmania, Australia, as well as one upper troposphere NOAA station at Mauna Loa, Hawaii. Their locations are marked on the map in Fig. 1. Some station records for N2O exist as far as the 1970s; however, the precision of the data was significantly improved in the mid-1990s with the implementation of new gas chromatograph multi-detector systems28. As a result, the analysis is limited to a recent subset of 21 years of data (1996–2016). At the AGAGE stations in the analysis, data were collected 36 times per day (a frequency of 40 min) whereas the Hawaii station logged data hourly.

To examine the overall spatial patterns and inter-annual variability in N2O (and sulfur hexafluoride) measurements, the long-term trend of rising anthropogenic N2O concentrations needs to be removed. De-trending was achieved by finding a two-degree polynomial fit and subtracting this general curve from the measured concentrations (Supplementary Fig. 1). De-seasonalization of the time series was achieved by averaging all de-trended data by month, interpolating these points to create a continuous sinusoidal seasonal cycle, and subtracting this oscillation from the de-trended series (Supplementary Fig. 2). This oscillation is largely driven by interhemispheric influence by the northern hemisphere during January/February. The de-seasonalized time series is utilized to assess interannual variability in atmospheric concentrations.

HYSPLIT back-trajectory modeling

In order to estimate from where air parcels arriving at the monitoring stations have traveled and help determine the influence of the OMZs, NOAA’s HYSPLIT4 model was used to calculate back-trajectories from the four Pacific stations (Supplementary Fig. 3). The model uses a combination of Lagrangian and Eulerian methodologies to compute air parcel trajectories, as well as more complex dispersion, atmospheric chemistry, and deposition dynamics52. Climatological data was retrieved from the NCEP/DOE reanalysis project, which uses grid cells of 2.0° latitude by 1.75° longitude, and temporal coverage four times daily. In setting up the model to calculate the trajectories, the latitude, longitude, and altitude of the stations were specified. Back trajectories were calculated four times per day for 21 years, from January 1, 1996 to December 31, 2016, resulting in a total of 30,672 trajectories analyzed for each station (26,953 with associated N2O measurements). Each trajectory was then matched with the corresponding station’s N2O measurement taken nearest to the time at which the back-trajectory arrived at the station.

It is unclear how far back in time the trajectories remain valid before inaccuracies lead to severe divergences between the calculated path and the actual, and both errors from the climatological data and from the HYSPLIT model itself must be considered in evaluating the certainty of the trajectories. Generally, HYSPLIT trajectories incur a distance error on order of 20% of the track length53, and both errors from the climatological data and from the HYSPLIT model itself must be considered in evaluating the certainty of the trajectories. However, a large number of trajectories calculated within an ensemble, as conducted here, can minimize the influence of any individual track54, and the emergent patterns indicate any error to be random such that its net effect can be minimized. In order to eliminate land-based sources of emissions, a filter was created to remove all trajectories that passed over continental land within the 20-day period of calculation. For the Samoa station, the island itself was not designated as land in this filter, under the assumption that it is too small to have substantial impacts on local atmospheric N2O concentrations. The stations nearer to the continents (California and Tasmania) proved more difficult to filter, and local emission sources can impact the analysis results at these locations.

Defining the ENSO strength during each trajectory

Several indices exist to measure ENSO-induced variability, and data of each is available from NOAA as monthly averages dating to 1950. Niño 3 and Niño 3.4 are now considered to be the most accurate indicators of the stage of the cycle55. These are based on sea surface temperature anomalies along the equator between 5° S and 5° N, from 150 to 90° W and from 170 to 120° W respectively55. NOAA defines an El Niño event as Niño 3.4 > 0.4, and a La Niña event as Niño 3.4 < −0.4. The Niño 1 + 2 index is measured as differences in far the eastern tropical Pacific, but has the highest variance of the indices and is not as indicative of the inter-annual cycle. The Multivariate ENSO Index version 2 (Niño MEI) was also considered; this index is the result of a combination of six variables—sea-level pressure, zonal and meridional components of the surface wind, sea surface temperature, surface air temperature, and total cloudiness fraction of the sky56. Niño indices also exhibit seasonal cycles, and were de-seasonalized by first subtracting monthly averages from the long-term series, then applying a Kolmogorov–Zurbenko (KZ) filter with six iterations of filters with length 2, 4, 6, 8, and 12 months to remove all variability on scales <1 year. Once de-seasonalized, the Niño 3.4 and MEI indices were very similar in phase and amplitude, so ultimately the Niño 3.4 index was used for the remainder of the analysis.

In evaluating the best metric of ENSO to use, it is also necessary to consider how biological productivity and N2O are expected to respond to climatic variability. It is reasonable that there should be a lag between changes in climate forcing and the biological response, though the length of lag is not apparent. Another consideration is that the biological response can be proportional to the integral of the ENSO index rather than the index itself. This supposition arises from the consideration that a long El Niño period, for example, may induce longer lasting additive impacts on OMZs and productivity than a short one, and so the cumulative state of the Pacific should be considered rather than discrete time points.

Preliminary comparisons of atmospheric data from Samoa and ENSO indices suggest that it may be necessary to incorporate this lag or integration into the analyses. The best fit for the de-seasonalized Niño 3.4 index to the Samoa data was determined to be a 3-month shift (Supplementary Fig. 5). The integral of the Niño 3.4 index, calculated with a cumulative trapezoidal numerical integral and then re-normalized, also fits the N2O data remarkably well (Fig. 1c). Ultimately the shifted Niño 3.4 index was chosen for analysis for simplicity and agreement with previous studies, but we note that the alignment with the integrated index implies an underlying feature of the system generating a period shift of 3 months.

Gridding marine oxygen concentrations

Dissolved O2 concentrations were taken from the World Ocean Atlas 2013 with the isopycnal surfaces calculated using the TEOS-10 equation of state16. The O2 content along the isopycnal surface with σ0 = 26.5 kg m−3 (Figs. 2 and 4) was used to evaluate the relationship between atmospheric N2O and marine O2 content; this value is commonly taken in literature and in practice to define the depth layer of the O2 minimum in the Eastern Pacific31,57, on the order of 200–500 m below the sea surface. These subsurface O2 concentrations were then gridded across the Pacific in boxes of 1° × 1° latitude × longitude, and trajectory endpoints were associated with the O2 values of the water below them.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability

All data derive from publicly accessible resources, i.e., the AGAGE network, World Ocean Atlas, and HYSPLIT model. The authors declare that the data supporting the findings of this study are available within the paper and its Supplementary Information files.

References

  1. 1.

    Montzka, S. A., Dlugokencky, E. J. & Butler, J. H. Non-CO2 greenhouse gases and climate change. Nature 476, 43–50 (2011).

    CAS  PubMed  Google Scholar 

  2. 2.

    Ravishankara, A. R., Daniel, J. S. & Portmann, R. W. Nitrous Oxide (N2O): the dominant ozone-depleting substance emitted in the 21st century. Science 326, 123–125 (2009).

    ADS  CAS  PubMed  Google Scholar 

  3. 3.

    Syakila, A. & Kroeze, C. The global nitrous oxide budget revisited. Greenh. Gas Meas. Manag. 1, 17–26 (2011).

    ADS  CAS  Google Scholar 

  4. 4.

    Codispoti, L. A. Interesting times for marine N2O. Science 327, 1339–1340 (2010).

    CAS  PubMed  Google Scholar 

  5. 5.

    Ciais, P. et al. Carbon and other biogeochemical cycles. in Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (eds Stocker, T. F. et al.) 465–570 (Cambridge University Press, 2013).

  6. 6.

    Goreau, T. J. et al. Production of NO2 and N2O by nitrifying bacteria at reduced concentrations of oxygen. Appl. Environ. Microbiol. 40, 526–532 (1980).

    CAS  PubMed  PubMed Central  Google Scholar 

  7. 7.

    Santoro, A. E., Buchwald, C., McIlvin, M. R. & Casciotti, K. L. Isotopic signature of N2O produced by marine ammonia-oxidizing archaea. Science 333, 1282–1285 (2011).

    ADS  CAS  PubMed  Google Scholar 

  8. 8.

    Ji, Q., Babbin, A. R., Jayakumar, A., Oleynik, S. & Ward, B. B. Nitrous oxide production by nitrification and denitrification in the Eastern Tropical South Pacific oxygen minimum zone. Geophys. Res. Lett. 42, 10755–10764 (2015).

    ADS  CAS  Google Scholar 

  9. 9.

    Zumft, W. G. Cell biology and molecular basis of denitrification. Microbiol. Mol. Biol. Rev. 61, 533–616 (1997).

    CAS  PubMed  PubMed Central  Google Scholar 

  10. 10.

    Paulmier, A. & Ruiz-Pino, D. Oxygen minimum zones (OMZs) in the modern ocean. Prog. Oceanogr. 80, 113–128 (2009).

    ADS  Google Scholar 

  11. 11.

    Revsbech, N. P. et al. Determination of ultra-low oxygen concentrations in oxygen minimum zones by the STOX sensor. Limnol. Oceanogr. Methods 7, 371–381 (2009).

    CAS  Google Scholar 

  12. 12.

    Tiano, L. et al. Oxygen distribution and aerobic respiration in the north and south eastern tropical Pacific oxygen minimum zones. Deep Sea Res. Part I Oceanogr. Res. Pap. 94, 173–183 (2014).

    ADS  CAS  Google Scholar 

  13. 13.

    Babbin, A. R., Keil, R. G., Devol, A. H. & Ward, B. B. Organic matter stoichiometry, flux, and oxygen control nitrogen loss in the ocean. Science 344, 406–408 (2014).

    ADS  CAS  PubMed  Google Scholar 

  14. 14.

    Suntharalingam, P. & Sarmiento, J. L. Factors governing the oceanic nitrous oxide distribution: simulations with an ocean general circulation model. Global Biogeochem. Cycles 14, 429–454 (2000).

    ADS  CAS  Google Scholar 

  15. 15.

    Key, R. M. et al. Global Ocean Data Analysis Project, Version 2 (GLODAPv2), ORNL/CDIAC-162, NDP-P093. (Carbon Dioxide Information Analysis Center, Oak Ridge National Laboratory, US Department of Energy, 2015). https://doi.org/10.3334/CDIAC/OTG.NDP093_GLODAPv2.

  16. 16.

    Olsen, A. et al. The global ocean data analysis project version 2 (GLODAPv2) - An internally consistent data product for the world ocean. Earth Syst. Sci. Data 8, 297–323 (2016).

    ADS  Google Scholar 

  17. 17.

    Arévalo-Martínez, D. L., Kock, A., Löscher, C. R., Schmitz, R. A. & Bange, H. W. Massive nitrous oxide emissions from the tropical South Pacific Ocean. Nat. Geosci. 8, 530–533 (2015).

    ADS  Google Scholar 

  18. 18.

    Babbin, A. R., Bianchi, D., Jayakumar, A. & Ward, B. B. Rapid nitrous oxide cycling in the suboxic ocean. Science 348, 1127–1129 (2015).

    ADS  CAS  PubMed  Google Scholar 

  19. 19.

    Naqvi, S. W. A. et al. Budgetary and biogeochemical implications of N2O isotope signatures in the Arabian Sea. Nature 394, 462–464 (1998).

    ADS  CAS  Google Scholar 

  20. 20.

    Naqvi, S. W. A. & Noronha, R. J. Nitrous oxide in the Arabian Sea. Deep Sea Res. Part A. Oceanogr. Res. Pap. 38, 871–890 (1991).

    ADS  CAS  Google Scholar 

  21. 21.

    Nevison, C. D., Weiss, R. F. & Erickson, D. J. Global oceanic emissions of nitrous oxide. J. Geophys. Res. 100, 15809–15820 (1995).

    ADS  CAS  Google Scholar 

  22. 22.

    Dore, J. E., Popp, B. N., Karl, D. M. & Sansone, F. J. A large source of atmospheric nitrous oxide from subtropical North Pacific surface waters. Nature 396, 63–66 (1998).

    ADS  CAS  Google Scholar 

  23. 23.

    Farías, L., Besoain, V. & García-Loyola, S. Presence of nitrous oxide hotspots in the coastal upwelling area off central Chile: an analysis of temporal variability based on ten years of a biogeochemical time series. Environ. Res. Lett. 10, 044017, https://doi.org/10.1088/1748-9326/10/4/044017 (2015).

  24. 24.

    Yang, S. et al. Global reconstruction reduces the uncertainty of oceanic nitrous oxide emissions and reveals a vigorous seasonal cycle. Proc. Natl Acad. Sci. USA https://doi.org/10.1073/pnas.1921914117 (2020).

    Article  PubMed  Google Scholar 

  25. 25.

    Stramma, L., Johnson, G. C., Sprintall, J. & Mohrholz, V. Expanding oxygen-minimum zones in the tropical oceans. Science 320, 655–658 (2008).

    ADS  CAS  PubMed  Google Scholar 

  26. 26.

    Schmidtko, S., Stramma, L. & Visbeck, M. Decline in global oceanic oxygen content during the past five decades. Nature 542, 335–339 (2017).

    ADS  CAS  PubMed  Google Scholar 

  27. 27.

    Prinn, R. G. et al. History of chemically and radiatively important atmospheric gases from the Advanced Global Atmospheric Gases Experiment (AGAGE). Earth Syst. Sci. Data 10, 985–1018 (2018).

    ADS  Google Scholar 

  28. 28.

    Prinn, R. G. et al. A history of chemically and radiatively important gases in air deduced from ALE/GAGE/AGAGE. J. Geophys. Res. Atmos. 105, 17751–17792 (2000).

    ADS  CAS  Google Scholar 

  29. 29.

    Nevison, C. D., Mahowald, N. M., Weiss, R. F. & Prinn, R. G. Interannual and seasonal variability in atmospheric N2O. Global Biogeochem. Cycles 21, GB3017 (2007).

    ADS  Google Scholar 

  30. 30.

    Rigby, M. et al. Increase in CFC-11 emissions from eastern China based on atmospheric observations. Nature 569, 546–550 (2019).

    ADS  CAS  PubMed  Google Scholar 

  31. 31.

    Karstensen, J., Stramma, L. & Visbeck, M. Oxygen minimum zones in the eastern tropical Atlantic and Pacific oceans. Prog. Oceanogr. 77, 331–350 (2008).

    ADS  Google Scholar 

  32. 32.

    Buitenhuis, E. T., Suntharalingam, P. & Le Quéré, C. Constraints on global oceanic emissions of N2O from observations and models. Biogeosciences 15, 2161–2175 (2018).

    ADS  CAS  Google Scholar 

  33. 33.

    Stramma, L., Bange, H. W., Czeschel, R., Lorenzo, A. & Frank, M. On the role of mesoscale eddies for the biological productivity and biogeochemistry in the eastern tropical Pacific Ocean off Peru. Biogeosciences 10, 7293–7306 (2013).

    ADS  Google Scholar 

  34. 34.

    Callbeck, C. M., Lavik, G., Stramma, L., Kuypers, M. M. M. & Bristow, L. A. Enhanced nitrogen loss by eddy-induced vertical transport in the offshore peruvian oxygen minimum zone. PLoS ONE 12, e0170059 (2017).

    PubMed  PubMed Central  Google Scholar 

  35. 35.

    Cline, J. D., Wisegarver, D. P. & Kelly-Hansen, K. Nitrous oxide and vertical mixing in the equatorial Pacific during the 1982–1983 El Niño. Deep Sea Res. Part A. Oceanogr. Res. Pap 34, 857–873 (1987).

    ADS  CAS  Google Scholar 

  36. 36.

    Weiss, R. F., Van Woy, F. A. & Salameh, P. K. Surface water and atmospheric carbon dioxide and nitrous oxide observations by shipboard automated gas chromatography: results from expeditions between 1977 and 1990. Rep. ORNL/CDIAC‐59, NDP‐044, Oak Ridge Natl. Lab., Oak Ridge, Tenn. (1992).

  37. 37.

    Naqvi, S. W. A. et al. Increased marine production of N2O due to intensifying anoxia on the Indian continental shelf. Nature 408, 346–349 (2000).

    ADS  CAS  PubMed  Google Scholar 

  38. 38.

    Morgan, E. J. et al. Air–sea fluxes of greenhouse gases and oxygen in the northern Benguela Current region during upwelling events. Biogeosciences 16, 4065–4084 (2019).

    ADS  CAS  Google Scholar 

  39. 39.

    Stramma, L. et al. Observed El Niño conditions in the eastern tropical Pacific in October 2015. Ocean Sci. 12, 861–873 (2016).

    ADS  Google Scholar 

  40. 40.

    Espinoza‐Morriberón, D. et al. Impacts of El Niño events on the Peruvian upwelling system productivity. J. Geophys. Res. Ocean 122, 5423–5444 (2017).

    ADS  Google Scholar 

  41. 41.

    Ji, Q. et al. Investigating the effect of El Niño on nitrous oxide distribution in the eastern tropical South Pacific. Biogeosciences 16, 2079–2093 (2019).

    ADS  CAS  Google Scholar 

  42. 42.

    Butler, J. H., Elkins, J. W., Thompson, T. M. & Egan, K. B. Tropospheric and dissolved N2O of the west Pacific and east Indian Oceans during the El Niño Southern Oscillation event of 1987. J. Geophys. Res. Atmos. 94, 14865–14877 (1989).

    ADS  CAS  Google Scholar 

  43. 43.

    Prinn, R. et al. Global average concentration and trend for hydroxyl radicals deduced from ALE/GAGE trichloroethane (methyl chloroform) data for 1978–1990. J. Geophys. Res. 97, 2445–2561 (1992).

    ADS  CAS  Google Scholar 

  44. 44.

    Hirsch, A. I. et al. Inverse modeling estimates of the global nitrous oxide surface flux from 1998-2001. Global Biogeochem. Cycles 20, GB1008, https://doi.org/10.1029/2004GB002443 (2006).

  45. 45.

    Nevison, C. D. et al. Exploring causes of interannual variability in the seasonal cycles of tropospheric nitrous oxide. Atmos. Chem. Phys. 11, 3713–3730 (2011).

    ADS  CAS  Google Scholar 

  46. 46.

    Thompson, R. L. et al. Nitrous oxide emissions 1999 to 2009 from a global atmospheric inversion. Atmos. Chem. Phys. 14, 1801–1817 (2014).

    ADS  Google Scholar 

  47. 47.

    Saikawa, E. et al. Global and regional emissions estimates for N2O. Atmos. Chem. Phys. 14, 4617–4641 (2014).

    ADS  Google Scholar 

  48. 48.

    Freund, M. B. et al. Higher frequency of Central Pacific El Niño events in recent decades relative to past centuries. Nat. Geosci. 12, 450–455 (2019).

    ADS  CAS  Google Scholar 

  49. 49.

    Altabet, M. A. et al. An eddy-stimulated hotspot for fixed nitrogen-loss from the Peru oxygen minimum zone. Biogeosciences 9, 4897–4908 (2012).

    ADS  CAS  Google Scholar 

  50. 50.

    Bourbonnais, A. et al. N-loss isotope effects in the Peru oxygen minimum zone studied using a mesoscale eddy as a natural tracer experiment. Global Biogeochem. Cycles 29, 793–811 (2015).

    ADS  CAS  Google Scholar 

  51. 51.

    Bange, H. W. et al. MEMENTO: a proposal to develop a database of marine nitrous oxide and methane measurements. Environ. Chem. 6, 195–197 (2009).

    CAS  Google Scholar 

  52. 52.

    Stein, A. F. et al. NOAA’s HYSPLIT atmospheric transport and dispersion modeling system. Bull. Am. Meteorol. Soc. 96, 2059–2077 (2015).

    ADS  Google Scholar 

  53. 53.

    Freitag, S. et al. Atmospheric Measurement Techniques Combining airborne gas and aerosol measurements with HYSPLIT: a visualization tool for simultaneous evaluation of air mass history and back trajectory consistency. Atmos. Meas. Tech. 7, 107–128 (2014).

    CAS  Google Scholar 

  54. 54.

    Cape, J. N., Methven, J. & Hudson, L. E. The use of trajectory cluster analysis to interpret trace gas measurements at Mace Head, Ireland. Atmos. Environ. 34, 3651–3663 (2000).

    ADS  CAS  Google Scholar 

  55. 55.

    Trenberth, K. Nino 3.4 Index using ersstv5 from CPC, https://psl.noaa.gov/data/climateindices/list/ (2020).

  56. 56.

    Wolter, K & Timlin, M. Multivariate ENSO Index Version 2, https://www.esrl.noaa.gov/psd/enso/mei (2020).

  57. 57.

    Evans, Z. et al. The Role of Water Masses in Shaping the Distribution of Redox Active Compounds in the Eastern Tropical North Pacific Oxygen Deficient Zone and Influencing Low Oxygen Concentrations in the Eastern Pacific Ocean. Limnol. Oceanogr. https://doi.org/10.1002/lno.1141 (2020).

Download references

Acknowledgements

We are indebted to Susan Solomon for her invaluable contribution in shaping this work. We further thank Ronald Prinn, Kerry Emanuel, Noelle Selin, Daniele Bianchi, and Ryan Woosley for fruitful discussions on this study. This work was funded by an MIT startup and Simons Foundation Grant 622065 to A.R.B., and by the David S. Bacon Fund for undergraduate research. We are additionally grateful for the financial support received from Dr. Bruce L. Heflinger. The AGAGE measurements at Samoa are supported by NASA’s Upper Atmosphere Research Program and by NOAA’s Global Monitoring Division.

Author information

Affiliations

Authors

Contributions

A.R.B. and E.L.B. devised and executed the study, and wrote the paper with contributions from all authors. E.L.B. designed the numerical approach and analyzed the data. J.M. and R.F.W. provided the atmospheric N2O and SF6 measurements used in this research.

Corresponding author

Correspondence to Andrew R. Babbin.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Peer review information Nature Communications thanks Carolin Löscher and other, anonymous, reviewer(s) for their contributions to the peer review of this work. Peer review reports are available.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Babbin, A.R., Boles, E.L., Mühle, J. et al. On the natural spatio-temporal heterogeneity of South Pacific nitrous oxide. Nat Commun 11, 3672 (2020). https://doi.org/10.1038/s41467-020-17509-6

Download citation

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Search

Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing