Coupling between dimethylsulfide emissions and the ocean–atmosphere exchange of ammonia

DMS is a biogenic trace gas whose flux is consistently from ocean to atmosphere. Its emission is of particular interest owing to the potential climate-forcing effect of its oxidation products, particularly H₂SO₄, the dominant source of strong acidity in the atmosphere and key species in new particle formation and growth. This is the basis of the ‘CLAW’ hypothesis,^{[ 1 ]} which suggests a biological climate control (i.e. negative feedback) mechanism, although this is still open to debate (see ref. [2] and ‘opinion’ articles immediately following in the same issue). What is known with certainty is that there is a marine source of DMS^{[ 3 – 6 ]} and that a substantial proportion of fine-mode aerosol in the marine boundary layer (MBL) is composed of non-sea salt sulfate (nss-SO₄ ^2–), which in remote regions is derived from the oxidation of DMS and is the predominant source of aerosol acidity.^{[ 7 ]}

The ocean–atmosphere flux of DMS is unidirectional because (i) DMS is a relatively insoluble gas (Henry’s law constant, K_H ≈ 0.5 M atm^{–1 [ 8 ]}), and (ii) it is relatively short-lived in the atmosphere (lifetime ≈ 1 day^{[ 9 ]}) with relatively rapid removal via oxidation by various pathways.^{[ 10 , 11 ]} These factors combine to maintain a strong concentration gradient (and thus flux) across the ocean–atmosphere interface.

Once in the atmosphere, DMS will oxidise on a timescale of hours by reaction with NO₃ radical (at night) and BrO radical,^{[ 11 ]} or days with oxidation by OH radical.^{[ 10 ]} The end points of these oxidation reactions are MSA (methanesulfonic acid, HOSO₂CH₃) and sulfuric acid (H₂SO₄), the partitioning between these species being dependent on temperature and the relative concentrations of available oxidants.^{[ 11 ]} These processes are effectively irreversible in Earth’s oxidising atmosphere.

A proportion of the H₂SO₄ will be taken up by coarse mode sea salt aerosol,^{[ 12 ]} this proportion being strongly dependent on wind speed and the enrichment of CaCO₃ alkalinity in the coarse mode aerosol.^{[ 13 ]} The remaining H₂SO₄ can condense spontaneously to form sulfate particles (classically considered to occur via binary nucleation with water molecules) and will also condense onto or dissolve into existing fine mode particles in the atmosphere.^{[ 14 , 15 ]} Whereas the propensity for H₂SO₄ to enter the particulate phase is governed by equilibrium thermodynamics (higher temperatures favouring the gas phase), it is generally thought that it will not re-enter the gas phase under normal atmospheric conditions owing to its extremely high solubility (in aqueous solutions) and ‘stickiness’ to dry surfaces.^{[ 16 ]} Thus the gas-to-particle conversion of sulfuric acid can also be considered as an effectively irreversible process.

NH₃ and its protonated equivalent (ammonium, NH₄ ⁺), together referred to henceforth as NH_x, are ubiquitous in the ocean and in the atmosphere. In the latter, NH₃ is the dominant gaseous base and neutraliser of strong acidity. NH₃ is emitted to the atmosphere in large quantities by anthropogenic activities, both industrial and agricultural,^{[ 17 , 18 ]} as well as by natural biogenic production in the terrestrial and marine environments. Anthropogenic emissions have led to perturbations in the global NH_x cycle many times greater than the pre-industrial fluxes.^{[ 19 , 20 ]} However, NH_x is relatively short-lived in the atmosphere (of the order of days), so the majority of NH_x from strong terrestrial sources tends to be deposited from the atmosphere before reaching remote marine environments.^{[ 17 , 21 , 22 ]}

In contrast to DMS, all of the processes linking dissolved ammonium in the marine surface layer (NH₄ ⁺ _(sw)) with particulate ammonium in the atmosphere (NH₄ ⁺ _(p)) are completely reversible. NH₄ ⁺ _(sw) exists in thermodynamic equilibrium with NH_3(sw) (dissolved, non-solvated NH₃), the protonation reaction occurring near-instantaneously to maintain equilibrium.^{[ 22 ]} Under typical seawater conditions, ~1–10% of the total NH_x(sw) exists as NH_3(sw).^{[ 23 ]}

Ammonia is considerably more soluble than DMS (K_H ≈ 60 M atm^{–1 [ 8 ]}). Its solubility means that for typical surface seawater and MBL concentrations, the ocean and atmosphere are generally close to equilibrium with respect to NH₃ ^{[ 21 , 24 – 27 ]} and thus its ocean–atmosphere exchange can be considered a bidirectional process that is highly sensitive to temperature, pH and concentrations of NH₃ in both surface ocean and MBL.^{[ 24 ]}

Once in the atmosphere, NH_3(g) will react readily with acidic gases and particles to enter the particulate phase as NH₄ ⁺ _(p). The direction of NH_3(g) flux between the gas and aerosol phases is determined by the difference in concentration between NH_3(g) and pNH_3(g) (the partial pressure of NH₃ over the aerosol phase). In aqueous aerosol, this process is reversible,^{[ 22 ]} as it is with dry ammonium salts of nitrate and chloride.^{[ 21 , 28 ]} The lower partial pressure of NH₃ over the much less volatile ammonium sulfate ((NH₄)₂SO₄) and ammonium bisulfate (NH₄HSO₄) means that any NH₄ ⁺ _(p) in dry salts formed with nss-SO₄ ^2– are likely to be irreversibly reacted. However, these salts are highly deliquescent (i.e. water-attracting), so are unlikely to remain in the solid phase in the relatively humid MBL.^{[ 21 ]}

Quinn et al.^{[ 22 ]} considered all components of the ‘multiphase’ ammonia system (NH₄ ⁺ _(sw) ↔ NH_3(sw) ↔ NH_3(g) ↔ NH₄ ⁺ _(p)) in the remote marine environment. They found that the characteristic time for equilibration between the surface ocean and NH_3(g) in the atmosphere is of the order of a few days to a week (dependent on wind speed), whereas the equilibration between gas and aerosol (or cloud) waters occurs on a timescale of hours. Oxidation of NH_3(g) by OH radical is rather slow, and considered insignificant relative to other removal processes (wet and dry deposition, mostly from the particulate phase).^{[ 17 , 22 ]} There is generally a disequilibrium observed between the atmosphere and ocean with respect to ammonia,^{[ 22 , 24 , 27 , 29 , 30 ]} supporting the suggestion that air–sea equilibration is a relatively slow process in the system. Therefore, the major control on the concentration of NH_3(g) in the remote MBL (neglecting advection from other sources) must be the equilibrium between gas and particle phases; i.e. on timescales of hours, NH_3(g) will tend to reach equilibrium with pNH_3(g).

Quinn et al.^{[ 29 ]} observed average NH₄ ⁺ : nss-SO₄ ^2– molar ratios of 1.5 ± 0.44 in air of remote marine source (and substantially different and more variable ratios in air of terrestrial and volcanic sources). In addition, relatively consistent NH₄ ⁺ : nss-SO₄ ^2– molar ratios (typically between 1 and 2) have been observed in marine-source aerosols sampled in the remote MBL of the Atlantic (T. G. Bell, A. R. Baker and T. D. Jickells, unpubl. data). An exception to this trend in clean marine air is the data of Savoie et al.,^{[ 31 ]} who observed substantially lower ratios (annual minimum of 0.6) at Mawson, Antarctica. In addition, their data suggest strong coupling of the seasonal trends in nss-SO₄ ^2– _(p) and NH_x(p) concentrations over a 5-year period at several Antarctic and Southern Ocean sites.

Quinn et al.^{[ 29 ]} propose that the relative constancy of the NH₄ ⁺ : nss-SO₄ ^2– ratio in clean marine air may indicate a biogeochemical linkage between the processes that produce DMS and NH₄ ⁺ in the surface ocean, although they do not observe any relationship between the concentrations of these two species in seawater. Links at the biochemical level have been proposed, for example the possible coproduction of the osmolytes DMSP (dimethylsulfoniopropionate, the precursor to DMS) and glycine betaine (or 2-trimethylammonioacetate, the nitrogenous analogue to DMSP) (see ref. [32], after ref. [33]).

Liss and Galloway^{[ 32 ]} propose that the linkage is more likely to be in the atmosphere and driven by the scavenging of NH₃ by acidic sulfate particles, and that the resulting NH₄ ⁺ : nss-SO₄ ^2– ratio would be dependent on the relative amounts of NH₃ and H₂SO₄ available from various sources (marine emissions, transport from continental regions, etc.). There is substantial evidence for an atmospheric coupling through the titration of aerosol acid sulfate with ammonia. ^A

Following the hypothesis of Quinn et al.,^{[ 29 ]} Liss and Galloway^{[ 32 ]} invoke ‘coproduction’; i.e. they assert that a given NH₄ ⁺ : nss-SO₄ ^2– ratio in the remote marine environment (away from continental sources) must represent the ratio of the production of ammonia and DMS in the surface ocean. We argue that this is unlikely, because the source regions of the NH₄ ⁺ and nss-SO₄ ^2– found in a given aerosol particle will be spatially separated owing to the timescale of DMS oxidation. Thus the likelihood of uniform coupling between the N and S cycles in the marine environment over such scales of space and time is unlikely. Second, only a tiny proportion of the NH₄ ⁺ produced by biological activity in the ocean will be emitted to the atmosphere; relatively constant surface seawater concentrations are a result of the tight coupling between rapid uptake and regeneration by the mixed plankton community.^{[ 34 – 36 ]} In other words, sea-to-air flux is not the dominant loss of NH_x from seawater and the same appears to be true for DMS.^{[ 37 ]} Finally, only a proportion of the DMS emitted from the ocean will be oxidised to nss-SO₄ ^2–.^{[ 10 ]}

However, in a system where all of the atmospheric N and S is derived from the surface ocean, it must be the case that a given NH₄ ⁺ : nss-SO₄ ^2– molar ratio in the aerosol will have resulted from emission of a stoichiometrically equivalent ratio of ammonia and DMS-derived sulfate. Therefore, there must be a process by which there is atmospheric control on the rate of emission of one or both of the gases in question and we invoke the concept of ‘co-emission’, a recognised process in the terrestrial environment, to explain this.

We hypothesise that away from significant terrestrial influence, any un-neutralised sulfate acidity in the atmosphere will scavenge NH₃ towards a ‘titration end-point’. This depletion of gas-phase NH₃ will favour the ocean-to-atmosphere flux of NH₃, which, in the case of the remote marine atmosphere where nss-SO₄ ^2– is predominantly derived from DMS emissions, can be considered a DMS-driven ‘co-emission’ of ammonia. ^B Ammonia, however, cannot have the same driving effect on DMS emission owing to the irreversible nature of the processes transforming DMS_(sw) to nss-SO₄ ^2– _(p).

This hypothesised neutralisation mechanism for atmospheric acidity is consistent with the observed constancy in NH₄ ⁺ : nss-SO₄ ^2– ratios in the marine environment. Whenever atmospheric acidity increases, pNH_3(g) over the aerosol decreases, leading to enhanced uptake of NH_3(g) onto the aerosol and consequent increased flux of NH₃ from the ocean to compensate the change.

As ocean–atmosphere exchange is the slowest process in the multiphase ammonia system in the marine environment, we predict that factors affecting the flux of ammonia from the ocean will be key controls on the NH₄ ⁺ : nss-SO₄ ^2– ratio. We have identified in Johnson et al.^{[ 24 ]} that temperature is a fundamental constraint on the magnitude and direction of ocean–atmosphere ammonia exchange such that, at a global scale, it outweighs the effects of biological activity. We predict that temperature is therefore likely to be a key control on the NH₄ ⁺ : nss-SO₄ ^2– ratio. However, on a regional scale, where temperature is relatively invariable, biological activity and associated concentration and pH changes are likely to be important determinants of the NH₄ ⁺ : nss-SO₄ ^2– ratio.

An analogous process is well established in studies of terrestrial trace gas biogeochemistry – that of ‘co-deposition’ of NH₃ and SO₂ to wet and dry surfaces.^{[ 38 , 39 ]} The concept was first suggested by Brimblecombe,^{[ 40 ]} where NH₃ was invoked as a neutralising agent facilitating the dissolution of SO₂ onto the ‘dew’ on the surface of leaves. Subsequent laboratory studies observed enhanced efficiency of deposition of both NH₃ and SO₂ in the presence of both gases,^{[ 38 ]} particularly under high relative humidities. The mechanism invoked to explain this co-deposition is the ‘suppression of the pH limitation to solubility for either gas in the presence of the other’.

In data presented by Sutton et al.,^{[ 38 ]} we see a process more directly analogous to that which we propose for the marine environment (Fig. 1). The process, which they term ‘co-emission’, is driven by SO₂ release due to drying on the leaf surfaces in a wheat field. Prior to drying, the atmosphere and wet leaf surfaces were close to equilibrium. Following drying and subsequent emission of SO₂, NH₃ was depleted by gas-to-particle conversion to below a ‘compensation point’ i.e. pNH_3(g) over the substomatal cavities of leaf surfaces.^{[ 41 ]} This led to an emission of NH₃ from the surface as a direct result of the sulfur emissions. This is a process that occurs on a small timescale relative to our hypothesised process, but evidence for longer-term connections between atmospheric ammonium and sulfur in the terrestrial realm are demonstrated by Fowler et al.,^{[ 42 ]} who find strong relationships between deposited ammonium and SO₂ surface equilibrium concentrations over a 2-year period.

Fig. 1.  Evidence for co-emission of SO₂ and NH₃ from a wheat field, presented by Sutton et al.^[38] Reproduced with permission. See text for details.

We focus in the present manuscript on the remote marine environment, where the system of interest is relatively close to equilibrium and thus our hypothesised process should be most easy to identify. However, the implications of the proposed mechanism also apply to more complex (e.g. terrestrially influenced) systems. For instance, wherever there is a net flux of sulfuric acidity from the land surface over the ocean (the norm in many regions of the world^{[ 17 , 43 ]}), gas-phase ammonia concentrations are likely to rapidly fall under control of the aerosol phase (i.e. tend towards pNH_3(g)) as the air moves away from terrestrial source regions, owing to the relative slowness of air–sea exchange and oxidation by OH. Low aerosol pH will result in low pNH_3(g) and thus enhance the flux of NH₃ from the ocean, or inhibit influx, depending on ambient conditions and concentrations. Zhuang and Huebert^{[ 44 ]} present data from a Lagrangian experiment, which show an increasing NH₄ ⁺ : nss-SO₄ ^2– ratio as sulfate-rich air is advected over the ocean, with evidence of rapid conversion of ocean-emitted NH₃ to the particulate phase, providing strong circumstantial evidence of sulfate-driven ammonia emission in continental air masses.

The data of Savoie et al.^{[ 31 ]} show strong evidence of a coupling between the emissions of NH₃ and DMS, with strong synchronicity between nss-SO₄ ^2– _(p) and NH₄ ⁺ _(p) (Fig. 2a). This synchronicity cannot be explained by the seasonal cycles of NH_x(sw) observed in the Southern Ocean, or the Antarctic coastal seas (Fig. 2c, d). The Joint Global Ocean Flux Study (JGOFS) Southern Ocean dataset (http://usjgofs.whoi.edu/jg/dir/jgofs/southern/, accessed April 2008) and the Rothera Antarctic Time Series data^{[ 45 ]} show NH_x(sw) at its most depleted during the months where the data of Savoie et al.^{[ 31 ]} demonstrates the greatest aerosol ammonium (December, January). This is consistent with our current understanding of the seasonal cycle of NH_x(sw) in temperate and high-latitude oceans where the nutrient-limited months of high summer tend to efficiently re-use remineralised N, keeping concentrations low.^{[ 21 , 34 ]} The peaks in NH₄ ⁺ _(p) also occur out of phase with surface temperatures, which reach their annual maximum in March.

Fig. 2.  Antarctic data: (a) Monthly averaged aerosol ammonium and non-sea salt sulfate ( nss-SO₄ ^2–) digitised by these authors from 5 years of measured data presented by Savoie et al.^[31] at Mawson, Antarctica; (b) NH₄ ⁺ : nss-SO₄ ^2– molar ratio calculated from the data in (a); (c) monthly averaged surface seawater NH_x concentrations from the Joint Global Ocean Flux Study (JGOFS) Southern Ocean dataset (http://usjgofs.whoi.edu/jg/dir/jgofs/southern/, accessed April 2008), and (d) from the Rothera Atlantic Time-Series.^[42]

We therefore argue that the only likely explanation for the synchronicity between NH₄ ⁺ _(p) and nss-SO₄ ^2– _(p) in Antarctic air is our hypothesised process of coupled DMS and NH₃ fluxes from the ocean. In doing so we make some assumptions: (i) windspeed and ocean pH are relatively small controls over the ocean–atmosphere NH₃ flux over commonly observed ranges^{[ 21 , 22 ]} and therefore the seasonal changes of these parameters in the Antarctic are unlikely to be responsible for the observed changes in NH₄ ⁺ _(p); and (ii) the data of Savoie et al.^{[ 31 ]} are (as indicated by the authors) under strong influence of seawater emissions, i.e. not dominated by advection or local point sources such as penguin colonies, which are substantial sources of ammonia^{[ 46 ]} and can affect local NH₄ ⁺ : nss-SO₄ ^2– ratios substantially.^{[ 47 ]}

Below we present the details of a simple zero-dimensional finite difference box model of the ocean–atmosphere NH_x–SO₄ ^2– system, which is intended to test the hypotheses: (i) that stoichiometric production of DMS and NH₄ ⁺ is not required in seawater to give a constant NH₄ ⁺ : nss-SO₄ ^2– ratio in aerosol; and (ii) that acidic DMS oxidation products added to the atmosphere can drive NH₃ emission from the ocean by depletion of NH_3(g). Only the nitrogen side of the system is explicitly modelled, i.e. we do not attempt to investigate the effects of the relative timescales of DMS and NH₃ release and subsequent incorporation into aerosol, as this is not practicably possible in a model with no horizontal extent. In brief, the model is driven by input of SO₄ ^2– directly into the aerosol phase, and the response of the NH_x system and the NH₄ ⁺ : nss-SO₄ ^2– ratio are investigated.

Input parameters, variables and boundary conditions are listed in Table 1 and full details of the model can be found in the Accessory publication. Ocean–atmosphere exchange of NH₃ is calculated using the thin film model (concentration difference) approach^{[ 48 ]} using a scheme identical to that presented by Johnson et al.^{[ 24 ]} Equilibration time across the air–sea interface according to this scheme is on the order of 100 h under typical conditions.^{[ 21 ]}

Table 1.  Input parameters, boundary conditions, variables and processes in the box model of the NH₄ ⁺ _(sw) ↔ NH_3(sw) ↔ NH_3(g) ↔ NH₄ ⁺ _(p) system
Aqueous phase (seawater) concentrations in this table are in brackets, atmospheric concentrations (i.e. per m³ of air) are in parentheses

This is much slower than the characteristic time of the other modelled process, the interconversion between NH_3(g) and NH₄ ⁺, which occurs on a timescale of between 0.3 and 7 h.^{[ 22 ]} We employ an equilibrium scheme where the rate of flux between gas and particle phases is proportional to the difference between the concentration of NH_3(g) and pNH_3(g) (Eqn A6, Accessory publication). We use our own parameterisation of pNH_3(g), based on output of the Aerosol Inorganics Model (AIM, http://www.aim.env.uea.ac.uk/aim/aim.html, accessed September 2007)^{[ 49 ]} (Eqn A7 and associated text, Accessory publication).

The NH₄ ⁺ : nss-SO₄ ^2– ratio in the model is found to be very insensitive to the rate of equilibration between gas and aerosol phases except at very low rates (equilibration times of tens to hundreds of hours), where it approaches the characteristic time of the ocean–atmosphere flux. We adopt a time constant for this process that results in 95% equilibration over ~1 h, in keeping with the observations and models of Harrison and Kitto^{[ 50 ]} and Quinn et al.^{[ 22 ]}

Initial conditions and default parameter values are selected to represent typical ‘average’ conditions in the global remote ocean and MBL, particularly the Southern Ocean. The seawater ammonium concentration is taken as a high-latitude average concentration of 150 nM after Johnson^{[ 21 ]}; the sulfate input is considered as a per-unit area flux and is taken as the mean value of the Southern Ocean DMS flux observed by Shon et al.^{[ 51 ]} (1.3 μmol m^–2 day^–1), multiplied by their mean observed SO₂ conversion factor of 0.6, giving a sulfate input of 0.5 nmol m^–2 min^–1. The aerosol loss term is selected to maintain a constant sulfate loading, although the NH₄ ⁺ : nss-SO₄ ^2– ratio is entirely insensitive to this term as ammonium and sulfate are removed from the model atmosphere in the same ratio as they exist in the aerosol. The model is found to be very insensitive to initial values of NH_3(g), NH₄ ⁺ _(p) and nss-SO₄ ^2– _(p) concentrations over a wide range of values, so arbitrary values of 1, 2 and 3 nmol m^–3 respectively are adopted as defaults, based on previous observations.^{[ 19 , 22 , 28 ]} Sensitivity of the model to other factors is considered below.

The model is run from starting conditions (concentrations of NH_3(g), NH₄ ⁺ _(p), and SO₄ ^2– _(p)) until a dynamic steady-state (constant values of all fluxes and concentrations) is reached. Generally, such a state is reached within 48 h of model time, and in all cases within 7 days. The results of the model over a range of values of input variables and boundary conditions are presented in Fig. 3. Note that the NH₄ ⁺ : nss-SO₄ ^2– ratio rarely exceeds a value of just over 2. This is due to the rapid increase in pNH_3(g) when the NH₄ ⁺ : nss-SO₄ ^2– ratio approaches or surpasses neutralisation, inhibiting further gas-to-particle flux.

Fig. 3.  Model output showing the effect of various parameters on the NH₄ ⁺ : non-sea salt sulfate (nss-SO₄ ^2–) molar ratio (R) and the sea–air NH₃ flux (F, in nmol m^–2 min^–1). Note that points of F and R at each x-axis parameter value represent the dynamic steady-state solution to a discrete model run.

Changes in [NH_x(sw)], temperature (t), wind speed and pH all lead to a directly proportional response in NH₄ ⁺ : nss-SO₄ ^2– ratio and sea–air NH₃ flux. This is because all of these parameters directly affect the rate of sea–air NH₃ flux and thus the NH₄ ⁺ : nss-SO₄ ^2– ratio, in agreement with our hypothesised effect, sea–air flux being the rate-limiting step in the system.

The effect of sulfate input rate on the NH₄ ⁺ : nss-SO₄ ^2– ratio and sea–air NH₃ flux is different. As would be expected, increasing the sulfate input to the system causes the NH₄ ⁺ : nss-SO₄ ^2– ratio to decrease, as the rate-limiting sea–air NH₃ flux struggles to compensate. However, the sea–air NH₃ flux increases (up to a maximum value), because the increase in aerosol acidity is leading to uptake of NH_3(g) and consequent enhancement of sea–air NH₃ flux, demonstrating our hypothesised DMS-driven co-emission.

To further investigate this interaction between sulfate input and sea–air NH₃ flux in the model, we compare the effect of sulfate input and temperature (as a proxy for all of the parameters that enhance sea–air ammonia flux) on the NH₄ ⁺ : nss-SO₄ ^2– ratio (Fig. 4). The maximum value of air–sea flux is reached in each case when the sulfate loading is sufficient to consume effectively all of the available NH_3(g) at each timestep. This demonstrates that although DMS-driven co-emission of DMS and ammonia may occur over the range of conditions encountered in the remote marine environment, low sulfate emission rate and/or a combination of factors favouring enhanced sea–air NH₃ flux are required to ensure substantial neutralisation of the aerosol through emission of ammonia.

Fig. 4.  Model output showing the effect of sulfate input on (a) NH₄ ⁺ : non-sea salt sulfate ( nss-SO₄ ^2–) molar ratio (R) and (b) sea–air NH₃ flux, over a range of temperatures. This highlights that sulfate input drives sea–air NH₃ flux up to a maximum value that is strongly dependent on temperature (and other factors influencing the strength of ammonia emission from the ocean, not shown).

Our results indicate that in the remote marine environment, any DMS-derived sulfate can enhance the rate of sea–air NH₃ flux, as hypothesised. Furthermore, it appears that the physical factors controlling ocean–atmosphere ammonia exchange are of fundamental importance to the NH₄ ⁺ : nss-SO₄ ^2– molar ratio in the marine atmosphere, i.e. biological production alone cannot account for high ratios in the marine environment. Although DMS emission is clearly a significant control on the NH₄ ⁺ : nss-SO₄ ^2– ratio in remote marine environments through increasing the concentration of nss-SO₄ ^2–, this is somewhat moderated, particularly at higher temperatures and wind speeds, by the enhanced flux of NH₃ from the ocean. Thus, physical and chemical forcings (temperature, wind speed and pH) have a greater than expected control on the NH₄ ⁺ : nss-SO₄ ^2– ratio and consequently any climate–biogeochemical feedbacks in the Earth-system involving the reduced N and S cycles. For instance, the continuous resupply of NH₃ to the remore MBL through co-emission may be important in enhancing new particle formation through ternary nucleation^{[ 52 , 53 ]} or other mechanisms.^{[ 15 , 54 ]} This is not currently considered in models of particle formation over the ocean. ^C In this role, our proposed mechanism provides a possible positive reinforcement of the CLAW hypothesis. Thus long-term changes in ocean pH, temperature or wind fields may have significant effects on the radiative budget in Earth’s past or future climate.

These same factors (t, pH and wind) and consequent changes in atmospheric acidity through changing neutralisation by NH₃ may have important effects on aerosol iron processing and consequent availability of Fe to plankton in iron-limited regions of the World’s oceans. Whether these potential feedbacks are positive or negative at a given point in Earth’s history (or future) in part depends on the relative changes in magnitude of t and pH with changing atmospheric CO₂ (pCO₂) as they tend to act against each other in their effect on NH₃ flux (i.e. an increase in pCO₂ will tend to decrease ocean pH, thus inhibiting marine NH₃ emission, while increased temperature will favour NH₃ emission).

We have proposed a mechanism by which observed NH₄ ⁺ : nss-SO₄ ^2– ratios can be explained without the need to invoke the stoichiometric coproduction of the precursor gases NH₃ and DMS in seawater. Our simple model has demonstrated that in the remote MBL, where the predominant sources of NH₄ ⁺ _(p) and nss-SO₄ ^2– _(p) are marine and where the system is relatively close to equilibrium with respect to NH_x, the hypothesised mechanism appears to exert a significant control on the NH₄ ⁺ : nss-SO₄ ^2– ratio. However, the model only solves for a dynamic steady-state and therefore accounts for neither fine-scale variation in marine emissions nor the advection of air masses enriched in either NH_x or SO₄ ^2– from the terrestrial realm. For large tracts of the contemporary ocean, these are likely to alter the signal that might be expected from our proposed process.

However, in pre-industrial times, when anthropogenic and natural terrestrial emissions were minor contributors to the reduced N and S global biogeochemical cycles, our hypothesised coupling mechanism is more likely to be applicable over most of the open ocean. Therefore, we argue that the suggestion to assume that marine NH₃ emissions scale with DMS emissions globally^{[ 17 ]} is probably reasonable for pre-industrial times but is not likely to be applicable in the modern era owing to large anthropogenic perturbations of the S and reduced N cycles.

To further test the real-world significance of co-emission of DMS and ammonia from the ocean, observational evidence must be combined with analysis using detailed box models of DMS oxidation and the next generation of ocean-coupled atmospheric chemical transport models, which will explicitly model ammonia and DMS fluxes and their interactions.

Biographies