MAY 2000 NOTES AND CORRESPONDENCE 725
NOTES AND CORRESPONDENCE
Nitric Acid±Sea Salt Reactions: Implications for Nitrogen Deposition to Water Surfaces
S. C. PRYOR Atmospheric Science Program, Department of Geography, Indiana University, Bloomington, Indiana
L. L. SùRENSEN Department of Wind Energy and Atmospheric Physics, Risù National Laboratory, Roskilde, Denmark
23 August 1999 and 20 October 1999
ABSTRACT
Many previous studies have indicated the importance of nitric acid (HNO3) reactions on sea salt particles for
¯ux divergence of HNO3 in the marine surface layer. The potential importance of this reaction in determining the spatial and temporal patterns of nitrogen dry deposition to marine ecosystems is investigated using models of sea spray generation and particle- and gas-phase dry deposition. Under horizontally homogeneous conditions Ϫ1 with near-neutral stability and for wind speeds between 3.5 and 10 m s , transfer of HNO3 to the particle phase to form sodium nitrate may decrease the deposition velocity of nitrogen by over 50%, leading to greater horizontal transport prior to deposition to the sea surface. Conversely, for wind speeds above 10 m sϪ1, transfer of nitrogen to the particle phase would increase the deposition rate and hence decrease horizontal transport prior to surface removal.
1. Introduction spheric ¯ux is uncertain and highly variable spatially, Paerl (1995) summarizes data that suggest dry deposi- a. Nitrogen deposition to aquatic ecosystems tion is a signi®cant contributor to the atmospheric ¯ux Most oceanic primary production by phytoplankton to all of the major oceans (i.e., dry deposition is greater is limited by nutrient availability. The limiting nutrients than 20% of the total ¯ux), and model calculations for commonly are nitrogen (N) and phosphorous (P). In- the seas around Denmark indicate that one-third of the creasing N and P concentrations have profound impli- total N comes from the atmosphere and approximately cations for ecosystem health and productivity. For ex- three-®fths of the atmospheric deposition (via wet and ample, Paerl (1995) suggests ``nitrogen-limited estuar- dry pathways) comes from particulate matter (Asman ies, shallow coastal and continental shelf waters account et al. 1995). It also has been suggested that atmospheric for nearly half the global oceanic primary production.'' ¯uxes to midlatitude coastal areas may be largest in the ϪϪ Typically, absorption of nitrate (NO32 ), nitrite ( NO ), late spring and summer months, when nutrient concen- ϩ and ammonium (NH4 ) from ocean waters supplies N trations are generally at a minimum and hence the effect required by organisms. Human activity has increased of this ¯ux may be disproportionately large (Fisher and the concentration of these and other N compounds in Oppenheimer 1991; Rendell et al. 1993). coastal waters, leading in some cases to greatly en- Nitrogen deposition is determined largely by emis- hanced primary production or eutrophication (Paerl sions of the following N-containing gases: nitric oxide
1995). A number of studies indicate that between 10% (NO), nitrogen dioxide (NO2), ammonia (NH3), and ni- and more than 50% of coastal N loading is attributable trous oxide (N2O). Nitric acid (HNO3) is formed prin- to atmospheric ¯uxes (e.g., Paerl 1995; Rendell et al. cipally as a result of NO and NO2 and has been shown 1993). Although the contribution of wet and dry de- to make a signi®cant contribution to total N deposition position and gas- and particle-phase N to this atmo- in a number of diverse terrestrial and marine environ- ments [e.g., in the studies of Sievering et al. (1992),
Pryor et al. (1999b), and Asman et al. (1995), HNO3 deposition contributed approximately 10% of the total Corresponding author address: S. C. Pryor, Atmospheric Science N ¯ux]. Program, Dept. of Geography, Indiana University, Bloomington, IN 47405. Nitrogen emissions to the atmosphere are currently E-mail: [email protected] twice ``natural background'' levels (Vitousek et al.
᭧ 2000 American Meteorological Society
Unauthenticated | Downloaded 09/29/21 08:33 AM UTC 726 JOURNAL OF APPLIED METEOROLOGY VOLUME 39
1997) and are predicted to increase by 25%±50% in less-developed regions by 2020 (Galloway et al. 1994). These projections indicate the importance of improved understanding of atmosphere±surface exchange and the potential for an enhanced role of atmospheric deposition in eutrophication problems in the twenty-®rst century. b. Dry deposition processes Because of dif®culties in direct measurement of dry deposition ¯uxes of particles and some gases, dry de- position often is determined by the ``concentration method'' in which the ¯ux F is given by multiplication FIG. 1. A schematic of the particle deposition model, illustrating of the concentration at some level above the surface by that the model is composed of two layers, a well-mixed turbulent a deposition velocity: layer in which gravitational settling and turbulent transport are re- sponsible for the transfer of particles toward the surface, and the F ϭϪ d(z)[C(z) Ϫ C(0)], (1) quasi-laminar surface layer in which gravitational settling of ``wet'' particles and diffusional transport are the primary transport mecha- where C(z) is concentration at measurement height z; nisms. This ®gure is adapted from Hummelshùj et al. (1992). C(0) is concentration at the surface (which may be zero depending on surface uptake); d(z) ϭ deposition ve- locity, which may be computed as the inverse of the to the ¯ux at the surface). Flux divergence (i.e., violation sum of a number of resistances plus a gravitational set- of this assumption), however, may occur because of tling term for particles: advection, storage effects, or chemical reactions be- 1 tween the measurement height and the ground (Kramm dgϭϩ . (2) and Dlugi 1994). Reaction of HNO3 on sea salt (NaCl) rabcϩ r ϩ r particles to yield sodium nitrate (NaNO3) in the particle In Eq. (2) the following are true. phase and volatize hydrochloric acid vapor (HCl) is a
potentially important source of ¯ux divergence of HNO3 1) ra is aerodynamic resistance. This term represents the resistance to transport by turbulence to the quasi- over the sea surface (Geernaert et al. 1998). This chem- laminar surface layer. ical reaction and others on and in sea salt aerosols also have been shown to play critical roles in halogen (par- 2) rb is quasi-laminar surface layer resistance. Very close to the surface, a laminar boundary layer forms ticularly chlorine and bromine) release in the marine (depth ഠ 100±1000 m), which essentially is free boundary layer (Sander and Crutzen 1996), which has of turbulence. Transport across this layer is largely been linked to ozone depletion [e.g., the rapid decrease the result of Brownian diffusion. in ozone concentrations in Arctic air masses during the spring (Barrie et al. 1988)]. 3) rc is surface resistance. For highly reactive or soluble gases the surface resistance is small. We assume a In this note, the reaction of HNO3 on sea salt is con- surface resistance of zero and 100% ef®cient surface sidered from a different perspective. The rate constant capture for particle deposition to water surfaces. for the HNO3±NaCl reaction is thought to be rapid (Fen- ter et al. 1994), and the reaction has a large equilibrium 4) g is gravitational settling velocity. For coarse-mode particles, the downward gravitational force exceeds constant (Seinfeld and Pandis 1998), so, given suf®cient the drag force from the viscosity of air, and hence particle surface area (and time), the reaction is expected to reach completion with near total transfer of N to the g is important. This transport process is negligible for gases and ®ne particles, however. particle phase. Here a ®rst analysis of the potential role of this reaction in modifying atmosphere±surface ¯uxes Additional processes that may be of importance to of N is presented. It is demonstrated that this transfer particle dry deposition but are not considered here in- of N from the gas to the particle phase has the potential clude interception and inertial forces, electrical migra- to modify substantially the spatial and temporal patterns tion due to surface charge, and thermophoresis caused of the ¯ux of N to the sea surface. by temperature gradients (see Zufall and Davidson 1998). 2. Modeling particle dry deposition to water surfaces c. The focus of this note a. Description of the model The majority of studies of dry deposition ¯uxes to surfaces explicitly or implicitly rely upon the constant Here the model of particle dry deposition velocity ¯ux layer assumption (i.e., that the measured or deter- described by Pryor et al. (1999) and based on work by mined ¯ux at a given height above the surface is equal Williams (1982) and Hummelshùj et al. (1992) is used.
Unauthenticated | Downloaded 09/29/21 08:33 AM UTC MAY 2000 NOTES AND CORRESPONDENCE 727
A schematic of the model is shown in Fig. 1. The form of the model is
( hϩ g,d)( ␦ ϩ g,w) d ϭ , (3) h ϩ ␦ ϩ g,d where h is transfer velocity in the layer dominated by turbulent transfer, g,d and g,w are gravitational settling transfer velocities (subscripts d and w indicate dry and wet particles, respectively, where dry particles are those present in the marine atmosphere above the quasi-lam- inar surface layer, and wet particles refer to particles that have taken up water as they approach the surface and are subject to higher humidities), and ␦ is transfer velocity across laminar surface layer. These terms are given by the following equations. 1) Transfer velocity due to turbulent processes: 1 h ϭ , (4) 1 zzz ln rr0m Ϫ⌿hh ϩ⌿ FIG. 2. Modeled particle as a function of wind speed and particle u* zLL d []0m diameter for NaCl particles. Note that u and z0m for the speci®ed U are obtained by iterative solution of the* logarithmic wind pro®le where is the von KaÂrmaÂn constant, u* is friction 10 velocity, z is reference height, z is momentum and Charnock formula (the Charnock constant is assumed to have a r 0m value of 0.0185). roughness length, ⌿h is the stability correction, and L is Monin±Obukhov length.
2) Transfer velocity due to gravitational settling: which particles are collected by the spray drops, zdrop is the average height that spray drops reach, rdrop is g p 2 ϭ dC, (5) the average radius of spray drops, qdrop is the ¯ux of g,x18 x []air spray drops from the surface, and ␣ is the area of sea surface covered by whitecaps: where g is gravity, is dynamic viscosity, p is Ϫ6 3.75 particle density, air is air density, dx is particle di- ␣ ϭ 1.7 ϫ 10 U 10 , (7) ameter (x ϭ d or w), and C is the Cunningham slip correction factor {C ϭ 1 ϩ (/d)[2.514 ϩ 0.8 from Wu (1988). exp(Ϫ0.55d/)]; is the mean free path of air; Sein- Transfer across the quasi-laminar surface layer repre-
feld and Pandis 1998}. For g,w, d is adjusted for sents the major limitation on ®ne particle deposition. hygroscopic growth in the quasi-laminar surface lay- The two components of (6) pertain to diffusional trans-
er [i.e., x ϭ w, so d ϭ dw, where dw is dependent fer through the quasi-laminar surface layer (also referred on relative humidity (Յ98.3% over salt water) and to as the viscous sublayer) and to the increase in down- particle composition using the approximations of ward movement of particles due to bubble burst activity Fitzgerald (1975)]. [see discussion in Pryor and Barthelmie (2000)]. 3) Transfer velocity across the quasi-laminar surface Deposition velocities computed using this model are layer: shown in Fig. 2 for pure NaCl particles and near-neutral atmospheric strati®cation for four different wind speeds. ϭ (1 Ϫ ␣ )(cu*ScϪ0.5 ReϪ0.5 ϩ u* ϫ 10Ϫ3/St ) ␦ bob The characteristic form of the graph with lowest de-
u* qdrop position velocity for particles with d ഠ 0.1±0.4 m 2 ϩ ␣bobϩ Eff(2r drop)(2z drop), (6) re¯ects higher deposition velocity for larger particles U ␣ []10 for which gravitational settling is nonnegligible and for
where ␣bob is relative area with bursting bubbles, c smaller particles for which Brownian diffusion is more ϭ 1/(c1͙c 2), c1 ϭ f(thickness of the molecular dif- ef®cient. This ®gure also emphasizes the importance of fusion layer), c 2 ϭ f(ratio of the height of roughness wind speed in determining deposition velocity (e.g., as Ϫ1 elements to the aerodynamic roughness length), Sc U10 increases from 5 to 15 m s , d for d ϭ 0.2 m is Schmidt number [Sc ϭ /D, where is the ki- increases more than fortyfold). As wind speeds increase, nematic viscosity of air, diffusivity D ϭ the size dependence of deposition velocity for particles kTC/(3d), k is the Boltzmann constant, and T is below 1 m decreases as enhancement of deposition temperature], Re is Reynolds number (Re ϭ due to bubble burst activity increases, and so the lim- u z /), St is Stokes number [St ϭ u2 /(g)], U itation placed on deposition by particle diffusivity is * 0m * g,x 10 is wind speed at 10 m, Eff is the effectiveness with effectively removed.
Unauthenticated | Downloaded 09/29/21 08:33 AM UTC 728 JOURNAL OF APPLIED METEOROLOGY VOLUME 39
TABLE 1. Comparison of model performance and observations of particle dry deposition to water surfaces.
Ϫ1 Ϫ1 Particle Observed d Modeled d (cm s ) Modeled d (cm s ) Ϫ1 Ϫ1 z0 (m) u* (m s ) description (cm s ) without bubbles with bubbles Sehmel and Sutter (1974) 2 ϫ 10Ϫ5 0.11 d ϭ 2.5 m 6.0 ϫ 10Ϫ3 3.2 ϫ 10Ϫ2 4.6 ϫ 10Ϫ2a cited in Slinn et al. ϭ 1.5gcmϪ3 1.2 ϫ 10Ϫ2 (1978) 2 ϫ 10Ϫ5 0.11 d ϭ 1.0 m 2.0 ϫ 10Ϫ2 8.7 ϫ 10Ϫ3 1.1 ϫ 10Ϫ2a ϭ 1.5gcmϪ3 2 ϫ 10Ϫ4 0.44 d ϭ 7.0 m 2.0, 4.5 1.5 1.5a ϭ 1.5gcmϪ3 2 ϫ 10Ϫ4 0.44 d ϭ 3.0 m 1.0 ϫ 10Ϫ1 4.5 ϫ 10Ϫ2 1.0 ϫ 10Ϫ1a ϭ 1.5gcmϪ3 Moller and Shumann ഠ5 ϫ 10Ϫ4 ഠ0.4 d ϭ 0.03 m 4.0 ϫ 10Ϫ2 1.0 ϫ 10Ϫ2 3.4 ϫ 10Ϫ2a (1970) cited in Slinn et NaCl al. (1978) Larsen et al. (1995) Not speci®edc d ϭ 0.2 m No bubbles: 2.8 ϫ 10Ϫ2 4.6 ϫ 10Ϫ2 U ϭ 3msϪ1 MgO 3.4 ϫ 10Ϫ2 Bubbles: Ϯ30%b Not speci®edc d ϭ 0.2 m No bubbles: 5.6 ϫ 10Ϫ2 1.1 ϫ 10Ϫ1 U ϭ 6msϪ1 MgO 6.4 ϫ 10Ϫ2 Bubbles: Ϯ30%b Not speci®edc d ϭ 0.2 m No bubbles: 9.5 ϫ 10Ϫ2 1.8 ϫ 10Ϫ1 U ϭ 9msϪ1 MgO 1.2 ϫ 10Ϫ1 Bubbles: Ϯ30%b
a Here U10 is inferred from the logarithmic wind pro®le and observed z0 and u*, ␣ was calculated based on U10. b Bubble burst simulated with bubblers. Bubblers set to give alpha ϭ 0.08 (representative of a wind speed of 17 m sϪ1). c z0 calculated from the Charnock formula and u* from the logarithmic wind pro®le.
b. Evaluation of the model assumptions are made about the uptake of HNO3 on sea Only a few direct measurements of particle dry de- salt and the prevailing atmospheric conditions. With re- position to water surfaces are available in the literature spect to the prevailing meteorological conditions, we for evaluation of mathematical models of deposition ve- assume near-neutral stability and horizontal homoge- neity. We also assume that, since reaction of HNO with locities, and the majority derive from wind-tunnel ex- 3 NaCl particles is a surface reaction (Mamane and Got- periments (Slinn et al. 1978; Larsen et al. 1995; Zufall tlieb 1992; Fenter et al. 1994), the size distribution of and Davidson 1998). Table 1 summarizes a comparison the resulting NaNO can be characterized by particle of the model performance relative to three wind-tunnel 3 diameters associated with the maximum surface area of experiments for a range of particle and environmental NaCl particles. parameters. As shown, the model performs well relative To calculate the surface-area size distribution for sea to the measurements, particularly in light of the exper- spray droplets, the following equations developed by imental error and uncertainties in initializing the model. Monahan et al. (1986) to describe spray generation in The model is within a factor of 2 for all comparisons open ocean conditions are used. The generation of sea except the experiment reported in Slinn et al. (1978) for salt droplets per unit area of sea surface [by bubble burst a particle diameter of 2.5 m. For this experiment the (F ) and spume (F )], per increment of droplet radius model underestimates observations, but this underesti- 0 1 r per second is given by mate may be caused by the model assumption that the particles are hydrophobic (the particle composition is dF 2 0 ϭ 2.373Ur3.41 Ϫ3(1 ϩ 0.057r 1.05) ϫ 10 1.19eϪB , (8) not speci®ed in the experimental description). If only dr 10 10% of the particles were ammonium nitrate (NH4NO3) (or another hydroscopic compound) the model would 0.380 Ϫ logr B ϭ , (9) predict the deposition velocity correctly. 0.65 and spume (r Ͼ 10 m): 3. The role of particles in modifying gas ¯uxes: dF1 HNO3±sea salt interactions ϭ 8.6 ϫ 10Ϫ6er 2.08U10 Ϫ8, (10) dr To illustrate the potential importance of HNO3±sea salt reactions on the spatial and temporal patterns of N where F is the ¯ux due to bubble burst (subscript 0) or ¯uxes to water surfaces, it is useful to compare the spume (subscript 1). Note that (10) is known to over- deposition velocity for HNO3 in the absence of reactions estimate the ¯ux of large droplets, but this overesti- with sea spray with the deposition velocity for the par- mation does not affect the analysis because the surface ticles that might act as the sink for HNO3 (via formation area of sea spray is dominated by r less than 10 m of NaNO3). To make this comparison, some simplifying (Smith et al. 1993).
Unauthenticated | Downloaded 09/29/21 08:33 AM UTC MAY 2000 NOTES AND CORRESPONDENCE 729
FIG. 4. Modeled particle d for an NaNO3 particle spectrum pre- FIG. 3. The normalized surface area of NaCl droplets computed us- sented as a function of particle d. Here, U10 was assumed to be 7 m ing (8)±(10). For clarity only data for r Յ 10 m are shown here. Ϫ1 s , and u and z0m are obtained by iterative solution of the logarithmic wind pro®le* and Charnock formula.
Figure 3 shows the normalized surface area of NaCl droplets as calculated using (8)±(10) and indicates a From (11)±(14), the deposition velocity of HNO3 for Ϫ1 de®ned maximum for diameters between 1.0 and 6.4 near-neutral stability and a wind speed of 7 m s is Ϫ1 m. For this analysis, we assume that the size distri- 0.48 cm s . bution generated by the ¯ux equations [(8)±(10)] rea- From these calculations one can observe that, for this Ϫ1 sonably represents ambient marine particles in this size scenario (i.e., U ϭ 7ms and near-neutral stability), range. transfer of HNO3 to the particle phase would lead to a If it is assumed that pure NaNO is formed by the reduction in the deposition velocity of N of over 50% 3 (from 0.48 to 0.2 cm sϪ1) leading to greater horizontal reaction of HNO3 on sea salt particles, a deposition velocity for those particles then can be calculated using transport of N prior to deposition. In this way a larger (3)±(6) for any wind speed. The deposition velocity for geographic area could be affected by in¯ux of anthro- Ϫ1 pogenic N compounds. Alternatively stated, this result NaNO3 particles for a wind speed of 7 m s is shown in Fig. 4. This ®gure indicates that, for diameters be- implies that, for the stated conditions, the atmospheric tween 1.0 and 6.4 m, the deposition velocity is 0.02± contribution to the N ¯ux in the coastal zone would be 0.8 cm sϪ1, with a value of 0.2 cm sϪ1 for the diameter reduced, but the ¯ux to open waters would be increased of maximum surface area (d ϭ 3.2 m). by reaction of HNO3 on sea spray particles. To investigate whether this result may be generalized The deposition velocity for HNO3 can be calculated using the resistance analogue [(11)±(14)] with the as- for varying conditions, the model described above was used to calculate deposition velocities for HNO3 and sumption that the surface resistance is zero (i.e., rc ϭ 0): NaNO3 particles (with d ϭ 1.0±6.4 m) for a range of wind speed conditions and near-neutral stability. The results, shown in Fig. 5, indicate that at all wind speeds 1 zr ra ϭ ln , (11) the deposition velocity for HNO3 is within the envelope u*z0 de®ned for particle NaNO3 deposition. However, at low Ϫ1 Ϫ1 (U10 Ͻ 3.5 m s ) and high (U10 Ͼ 10 m s ) wind 1 z0 r ϭ ln , (12) speeds the deposition velocity for HNO3 is lower than b u* z 0c that for an NaNO3 particle of 3.2-m diameter, and, for Ϫ1 wind speeds between 3.5 and 10 m s , transfer of HNO3 2/3 Re Ͻ 0.15: z ϭ 30 eϪ13.6Sc , and (13) to the surface is more rapid than it is for particle NaNO3, 0c u* indicating that transfer of N to the particle phase would 1/4 1/2 decrease the deposition rate and hence increase hori- Re Ն 0.15: z ϭ 20zeϪ7.3Re Sc , (14) 00cm zontal transport of N prior to deposition. The source of wherez0c is the roughness length for the chemical of the wind speed dependence in the ratio of HNO3:NaNO3 interest [from Joffre (1988) and Asman et al. (1994)]. deposition velocities implied in Fig. 5 is principally that
Unauthenticated | Downloaded 09/29/21 08:33 AM UTC 730 JOURNAL OF APPLIED METEOROLOGY VOLUME 39
ary layer, that for the conditions considered here this reaction is irreversible, that chloride de®cits in sea salt
aerosols are principally the result of HNO3 reaction (or at least that uptake of other acidifying gases does not
limit HNO3 uptake), and that NaCl aerosol contains sub- stantial water (Weiss and Ewing 1999) from the rela- tively high relative humidity of the marine atmospheric boundary layer [Beichert and Finlayson-Pitts (1996) have demonstrated the importance of surface water in
surface uptake of HNO3]. Further, we assume that the NaNO3 formed will be concentrated at the peak in the surface-area size distribution of sea spray as described by the algorithms of Monahan et al. (1986) (i.e., d ϭ 1.0±6.4 m). In addition, only horizontally homoge- neous conditions with near-neutral stability are treated. Within the limits imposed by these assumptions, it is
demonstrated that, at low and high wind speeds (U10 Ͻ Ϫ1 Ϫ1 3.5ms and U10 Ͼ 10ms ), the deposition velocity for HNO3 is lower than that for the ``average'' NaNO3 particle generated by surface reaction with HNO3, and FIG. 5. Modeled particle d for HNO3 and an NaNO3 particle spec- hence transfer to the particle phase would lead to in- trum presented as a function of U10. The particle spectrum was cal- creased N deposition rates. For wind speeds between culated from (8)±(10), and the envelope of d presented represents Ϫ1 3.5 and 10 m s , however, transfer of HNO3 to the the range of d for particles with diameters that have a surface area equal to or greater than 60% of the maximum surface area (i.e., d ϭ surface is more rapid than for particle NaNO3, indicat-
1.0±6.4 m). Also shown is the d for the diameter associated with ing that, at these wind speeds, transfer of N to the par- the peak in the surface area distribution (d ϭ 3.2 m). As in all the ticle phase would decrease the deposition rate and hence simulations, u and z are obtained by iterative solution of the log- * 0m increase horizontal transport of N prior to deposition. arithmic wind pro®le and Charnock formula. The step change in This effect may lead to a larger offshore area being HNO at U ϭ 3.5msϪ1 is due to a change in the formulas used 3 d affected by the products of emissions of oxides of ni- to calculate z0c [from (13)±(14)] as the Reynolds number increases above 0.15 [i.e., in the transition between smooth and rough condi- trogen from coastal cities. tions; see discussion in Joffre (1988)]. The results presented here emphasize the importance of particle diameter in determining particle deposition. If the role of ¯ow divergence in the coastal zone and the increase of particle deposition with wind speed is development of internal boundary layers is neglected, greater than that of HNO . Particle deposition to water 3 these results also imply that the numerical models cur- surfaces commonly is limited by the rate of transfer rently used to estimate N deposition to the coastal zone across the quasi-laminar surface layer. Resistance to this that do not treat explicitly the HNO3±sea spray reactions transfer is dependent on white cap activity [(6)]. As may overestimate the N ¯ux under moderate wind wind speed increases, white cap activity increases to the speeds and underestimate the N ¯ux under high wind 3.75 power [(7)], and hence particle transfer across the speed conditions. Production of sea spray increases with quasi-laminar surface layer increases in a nonlinear wind speed, hence, assuming HNO3±sea spray reaction fashion, leading to larger deposition velocities. This in- is not limited kinetically, high wind speeds would lead crease in deposition velocity with wind speed continues to enhanced transfer of HNO3 from the gas phase and until the point at which transfer of particles with d ഠ increased deposition of N close to the source of HNO3. 1.0±6.4 m across the quasi-laminar surface layer ceas- es to be the major limitation on deposition. Acknowledgments. The authors acknowledge discus- sions with Rebecca Barthelmie and Sùren Larsen of Risù National Laboratory and Michael Schulz of the Uni- 4. Discussion and implications versity of Hamburg. This research received funding
An examination of the importance of HNO3±sea salt from the EU project ``BASYS,'' the Danish Research reactions on the deposition of N to marine ecosystems Council, Environment Canada (KE501-8-0722EDM) has been presented. Models of sea spray generation and and NSF (ATM 971755). The authors also express particle- and gas-phase dry deposition were used to thanks to the three reviewers for their comments and compare deposition velocities for HNO3 and NaNO3 constructive criticisms. formed by reaction on sea spray. For this study, a number of a simplifying assumptions REFERENCES were made. The reaction of HNO3 on sea salt was treated Asman, W. A. H., 1994: Estimation of the net air±sea ¯ux of ammonia in a highly simpli®ed manner. We assume that this re- over the southern bight of the North Sea. Atmos. Environ., 28, action dominates HNO3 chemistry in the marine bound- 3647±3654.
Unauthenticated | Downloaded 09/29/21 08:33 AM UTC MAY 2000 NOTES AND CORRESPONDENCE 731
, and Coauthors, 1995: Atmospheric nitrogen input to the Kat- Oceanic Whitecaps, E. C. Monahan and G. Mac Niocaill, Eds., tegat. Ophelia, 42, 5±28. D. Reidel, 167±174. Barrie, L. A., J. W. Bottenheim, R. C. Schnell, P. J. Crutzen, and R. Paerl, H. W., 1995: Coastal eutrophication in relation to atmospheric A. Rasmussen, 1988: Ozone destruction and photochemical re- nitrogen deposition: Current perspectives. Ophelia, 41, 237±259. actions at polar sunrise in the lower Arctic atmosphere. Nature, Pryor, S. C., and R. J. Barthelmie, 2000: Particle dry deposition to 334, 138±141. water surfaces: Processes and consequences. Seas at the Mil- Beichert, P., and B. J. Finlayson-Pitts, 1996: Knudsen cell studies of lennium, C. Sheppard, Ed., Vol. 3, Elsevier Science, 127±139. , , L. L. S. Geernaert, T. Ellermann, and K. D. Perry, 1999a: the uptakes of gaseous HNO3 and other oxides of nitrogen on Speciated particle dry deposition to the sea surface: Results from solid NaCl: The role of surface-adsorbed water. J. Phys. Chem., ASEPS '97. Atmos. Environ., 33, 2045±2058. 100, 15 218±15 228. , , and B. Jensen, 1999b: Nitrogen dry deposition at an Fenter, F. F., F. Caloz, and M. J. Rossi, 1994: Kinetics of nitric acid AmeriFlux site in a hardwood forest in the Midwest. Geophys. uptake by salt. J. Phys. Chem., 98, 9801±9810. Res. Lett., 26, 691±694. Fisher, D., and M. Oppenheimer, 1991: Atmospheric nitrogen de- Rendell, A. R., C. J. Ottley, T. D. Jickells, and R. M. Harrison, 1993: position and the Chesapeake Bay estuary. Ambio, 20, 102±108. The atmospheric input of nitrogen species to the North Sea. Fitzgerald, J. W., 1975: Approximation formulas for the equilibrium Tellus, 45B, 53±63. size of an aerosol particle as a function of its dry size and Sander, R., and P. J. Crutzen, 1996: Model study indicating halogen composition and the ambient relative humidity. J. Appl. Meteor., activation and ozone destruction in polluted air masses trans- 14, 1044±1049. ported to the sea. J. Geophys. Res., 101, 9121±9138. Galloway, J., H. Levy, and P. Kasibhatla, 1994: Year 2020: Conse- Seinfeld, J., and S. Pandis, 1998: Atmospheric Chemistry and Physics: quences of population growth and development on deposition From Air Pollution to Climate Change. Wiley±Interscience, of oxidized nitrogen. Ambio, 23, 120±123. 1326 pp. Geernaert, L. L. S., G. L. Geernaert, K. Granby, and W. A. H. Asman, Sievering, H., D. Burton, and N. Caine, 1992: Atmospheric loadings of nitrogen to alpine tundra in the Colorado front range. Global 1998: Fluxes of soluble gases in the marine atmosphere surface Biogeochem. Cycles, 6, 339±346. layer. Tellus, 50B, 111±127. Slinn, W., and Coauthors, 1978: Some aspects of the transfer of at- Hummelshùj, P., N. O. Jensen, and S. E. Larsen, 1992: Particle dry mospheric trace constituents past the air±sea interface. Atmos. deposition to a sea surface. Fifth Int. Conf. on Precipitation Environ., 12, 2055±2087. Scavenging and Atmosphere±Surface Exchange Processes, S. E. Smith, M. H., P. M. Park, and I. E. Consterdine, 1993: Marine aerosol Schwartz and W. G. N. Slinn, Eds., Hemisphere, 829±840. concentrations and estimated ¯uxes over the sea. Quart. J. Roy. Joffre, S. M., 1988: Modelling the dry deposition velocity of highly Meteor. Soc., 119, 809±824. soluble gases to the sea surface. Atmos. Environ., 22, 1137± Vitousek, P.M., J. D. Aber, R. W. Howarth, G. E. Likens, P.A. Matson, 1146. D. W. Schindler, W. H. Schlesinger, and D. G. Tilman, 1997: Kramm, G., and R. Dlugi, 1994: Modelling the vertical ¯uxes of Human alteration of the global nitrogen cycle: Sources and con- nitric acid, ammonia and ammonium nitrate. J. Atmos. Chem., sequences. Eco. Appl., 7, 737±750. 18, 319±357. Weiss, D. D., and G. E. Ewing, 1999: Water content and morphology Larsen, S. E., J. B. Edson, P.Hummelshùj, N. O. Jensen, G. de Leeuw, of sodium chloride aerosol particles. J. Geophys. Res., 104, 21 275±21 285. and P. G. Mestayer, 1995: Dry deposition of particles to ocean Williams, R. M., 1982: A model for the dry deposition of particles surfaces. Ophelia, 42, 193±204. to natural surface waters. Atmos. Environ., 16, 1933±1938. Mamane, Y., and J. Gottlieb, 1992: Nitrate formation on sea-salt and Wu, J., 1988: Variations of whitecap coverage with wind stress and mineral particlesÐA single particle approach. Atmos. Environ., water temperature. J. Phys. Oceanogr., 18, 1488±1453. 26A, 1763±1769. Zufall, M. J., and C. I. Davidson, 1998: Dry deposition of particles. Monahan, E. C., D. E. Spiel, and K. L. Davidson, 1986: A model of Atmospheric Particles, R. M. Harrison and R. E. van Grieken, marine aerosol generation via whitecaps and wave disruption. Eds., John Wiley and Sons, 425±473.
Unauthenticated | Downloaded 09/29/21 08:33 AM UTC