Field observations of the ocean-atmosphere exchange of ammonia: Fundamental importance of temperature as revealed by a comparison of high and low latitudes Martin T. Johnson,1 Peter S. Liss,1 Thomas G. Bell,1 Timothy J. Lesworth,1 Alex R. Baker,1 Andrew J. Hind,1 Timothy D. Jickells,1 Karabi F. Biswas,2 E. Malcolm S. Woodward,3 and Stuart W. Gibb4 Received 22 June 2007 revised 24 September 2007 accepted 10 October 2007 published 16 February 2008. [1] Simultaneous measurements of NH3 in the atmosphere and NH4 + in the ocean are presented from fieldwork spanning 10 years and 110 degrees of latitude, including the first such simultaneous measurements in the remote marine environment at 55��N. At high latitudes, fluxes were almost exclusively from air to sea, in contradiction with previous lower-latitude studies, which have suggested that the open oceans are predominantly sources of ammonia to the atmosphere. Sensitivity analysis demonstrates that the direction and magnitude of the ocean-atmosphere NH3 exchange is highly dependent on water temperature. This temperature effect is sufficiently strong to outweigh the effects of variability in concentrations in seawater and atmosphere in many parts of the (open) ocean. This is highlighted in data from the Atlantic oligotrophic gyres, where fluxes were found to be predominantly out of the ocean despite extremely low dissolved ammonium concentrations in surface waters. Citation: Johnson, M. T., P. S. Liss, T. G. Bell, T. J. Lesworth, A. R. Baker, A. J. Hind, T. D. Jickells, K. F. Biswas, E. M. S. Woodward, and S. W. Gibb (2008), Field observations of the ocean-atmosphere exchange of ammonia: Fundamental importance of temperature as revealed by a comparison of high and low latitudes, Global Biogeochem. Cycles, 22, GB1019, doi:10.1029/2007GB003039. 1. Introduction [2] Ammonia (NH3) is important in the biogeochemical cycling and geographic redistribution of nitrogen and in its potential climate forcing role through its connections to the sulphur cycle and cloud formation [Liss and Galloway, 1993]. In the atmosphere and ocean, ammonia and its protonated form, ammonium (NH4) + are ubiquitous. Natu- rally and anthropogenically produced NHx (NH3 + NH4) + is transported through the atmosphere and generally occurs in decreasing concentration in air with distance from land. It has been suggested that in preindustrial times the oceans were probably a net source of NHx to the continents [Duce et al., 1991], but this is not the case today. Generally speaking, seawater NHx concentrations are lower in regions of low productivity nutrient-limited communities being more efficient at utilising recycled nitrogen and thus main- taining a lower ambient concentration. Thus high latitudes tend to have substantially greater NHx concentrations than low latitudes in the open ocean, with high-productivity coastal and shelf seas tending to have highest concentra- tions, irrespective of latitude [Johnson, 2004]. NHx is produced in surface waters by the biological reduction of nitrate (either directly or via the degradation of biologically synthesized organic nitrogenous material). In solution it is partitioned between ammonium and ammonia according to equilibrium thermodynamics: The proportion of NHx that occurs as NH3 (dependent on pH, temperature and ionic strength of the medium) is available for emission to the atmosphere the phase partitioning being dependent on the Henry���s Law coefficient. Ammonia is also emitted to the atmosphere by plants and animals in terrestrial environ- ments (both directly and through breakdown of organic nitrogen) by soil microorganisms and by various industrial and agricultural processes, including the direct volatilization of solid ammonium nitrate salts in fertilizer. There is also evidence of a volcanic source of NHx to the atmosphere [Uematsu et al., 2004], and of substantial ammonia emis- sions from seabird and seal colonies [Blackall et al., 2007 Theobald et al., 2006]. [3] Although in most Earth-surface environments the aqueous-phase partitioning between NH3 and NH4 + is dom- inated by the latter species (NH3 only becoming dominant at pH 9���10 under typical environmental conditions), the majority of emission to the atmosphere from terrestrial and marine surfaces is as NH3(g). Once in the atmosphere, ammonia reacts readily with acid gases and acidic aerosols and enters the particulate phase as NH4(p). + This is a slight simplification: In aerosol with an aqueous component a GLOBAL BIOGEOCHEMICAL CYCLES, VOL. 22, GB1019, doi:10.1029/2007GB003039, 2008 Click Here for Full Article 1School of Environmental Sciences, University of East Anglia, Norwich, UK. 2 New York State Department of Health, Albany, New York, USA. 3 Plymouth Marine Laboratory, Plymouth, UK. 4 Environmental Research Institute, UHI Millennium Institute, Thurso, UK. Copyright 2008 by the American Geophysical Union. 0886-6236/08/2007GB003039$12.00 GB1019 1 of 15

small proportion of the particulate NHx will be in the form NH3 (albeit 1% in low-pH aerosol solutions). However, it is a convenient simplification to consider NHx in the atmosphere to be partitioned between gas-phase NH3 and particle-phase NH4, + which convention we adopt following previous workers in the field, e.g., Gibb et al. [1999]. [4] While NH3(g) can be deposited from the atmosphere by wet deposition processes and direct uptake onto aqueous surfaces, deposition is dominated by the particulate form [Warneck, 1988], suggesting net transformation of NH3(g) to NH4(p). + However, this process is reversible, at least under some conditions [Milford et al., 2000]. It has been sug- gested that in the remote marine environment (i.e., away from significant terrestrial influence), NHx should be in thermodynamic equilibrium between gas and particle phases in the atmosphere, and also in thermodynamic (Henry���s Law) equilibrium with respect to NH3 across the ocean-atmosphere interface [Quinn et al., 1992 Johnson, 2004], but field data do not always support this [Johnson, 2004]. We do not investigate particle-phase data in this work because it does not directly influence gas exchange rates across the air-sea interface and should be studied in its own right. [5] Figure 1 summarizes previously published observa- tions of sea-air ammonia exchange. Note that throughout this paper, we use the convention of a positive flux being from sea to air (i.e., oceanic emission). Figure 1 only includes fluxes calculated from simultaneous measurements of NHx species in the atmosphere and ocean. Fluxes estimated from nonsimultaneous measurements and early estimates using only an atmospheric or seawater concentra- tion term [e.g., Georgii and Gravenhorst, 1977 Ayers and Gras, 1980] often give fluxes orders of magnitude greater than those observed in studies presenting simultaneous measurement, and are therefore discounted. We suggest that the reasons for the differences between simultaneous and nonsimultaneous observations are (1) the high variability and rapid changes in concentrations of NHx in all phases in space and time and (2) the low quality and high detection limit of some of the early methods for measurements of NHx, particularly in the atmosphere [Johnson, 2004]. [6] With the exception of the data of Asman et al. [1994], which were collected in the eutrophic and highly terrestri- ally and anthropogenically influenced North Sea, all previ- ous studies suggest that ocean-atmosphere ammonia exchange in the remote marine and pristine coastal environ- ments is predominantly from sea to air. In this paper we present the first open ocean data set showing predominantly downward (into ocean) fluxes, from fieldwork at 55��N. The reasons for such downward fluxes are identified and are compared to our data from two recent Atlantic Meridional Transect (AMT) cruises. 2. Fieldwork and Methods [7] Data presented here originate from six research cruises, summarized in Table 1 and Figure 2. These cruises were ships of opportunity for the study of ocean-atmosphere ammonia exchange and as such, cruise track and sampling opportunities were not necessarily optimal for the work undertaken. Methods used and participating personnel dur- ing each cruise are listed in Table 2. 2.1. Seawater Measurements 2.1.1. NHX(sw) [8] Total seawater ammonium concentration [NHx(sw)] was measured by flow injection (gas diffusion) ion chro- matography (FIGD-IC) [Gibb, 1994] during cruise PS211, an automated fluorescence method described by Jones [1991] during cruise AMT17 and the manual fluorescence method of Holmes et al. [1999], during the remaining cruises. [9] Sampling was conducted from Niskin bottles attached to the conductivity-temperature-depth profiler (CTD) frame (all cruises) and from the ships��� nontoxic supply (NTS), intake at approximately 5 m depth (S1801, JR75, D267 and Figure 1. Summary of data presented by previous studies of ocean-atmosphere ammonia exchange. Only data from simultaneous measurements of seawater and gas-phase species are considered. Positive values represent a flux from ocean to atmosphere. GB1019 JOHNSON ET AL.: OCEAN-ATMOSPHERE EXCHANGE OF AMMONIA 2 of 15 GB1019

AMT14). In general, excellent agreement was found be- tween samples taken from Niskin bottles fired at the surface (0���3 m depth) and simultaneously collected NTS samples from cruises S18/01, JR75 and D267 (Figure 3). These data suggest that sampling from the NTS is representative of the conditions in bulk surface seawater, at least during these cruises. 2.1.2. FIGD-IC [10] All samples were analyzed unfiltered, within 6 h of collection using flow injection gas diffusion���ion chroma- Table 1. Summary of Research Cruises Undertaken Cruise Ship Date Region PS211 RV Poseidon Sep 1995 NE Atlantic transect: Rekjavic to Lisbon Institut fur Meereskunde (now IFM-GEOMAR, Kiel), Germany 39 to 65��N, 32 to 10��E S18/01 RV Scotia Dec 2001 northern North Sea (survey) Fisheries Research Servies, Aberdeen, United Kingdom 57.5 to 63��N, 6 to 5.5��E JR75 RRS James Clark Ross Jun���Jul 2002 Arctic Ocean / Norwegian Sea (survey) British Antarctic Survey, Cambridge, United Kingdom 58.5 to 81.5��N, 7 to 13.5��E D267 RRS Discovery Nov���Dec 2002 Irminger Basin / NE Atlantic (survey) Natural Environment Research Council, United Kingdom 52.5 to 65.5��N, 35.5 to 6��E AMT14 RRS Discovery Apr���Jun 2004 Atlantic Meridional Transect (AMT): Falklands to United Kingdom Natural Environment Research Council, United Kingdom 37 to 49��N, 36 to 6��E AMT17 RRS Discovery Oct���Nov 2005 Atlantic Meridional Transect (AMT): United Kingdom to Cape Town Natural Environment Research Council, United Kingdom 36 to 56��N, 39 to 19��E Figure 2. Tracks of the six cruises from which data are presented. Table 1 provides further details. Plot produced using the freely available Ocean Data View Software (http://odv.awi-bremerhaven.de/ home.html). GB1019 JOHNSON ET AL.: OCEAN-ATMOSPHERE EXCHANGE OF AMMONIA 3 of 15 GB1019