sea Exchange of Aerosols
SCOPE 21 -The Major Biogeochemical Cycles and Their Interactions
16
|
Biogeochemical Cycles and the airsea Exchange of Aerosols
|
|
R. A. DUCE |
ABSTRACT
The seaair exchange of aerosols plays an important role in the global biogeochemical cycles of carbon, phosphorus, nitrogen, and sulphur. In this paper a brief general discussion of sea-to-air and air-to-sea transport of aerosols is given followed by a discussion of the sources, transport, and
seaair exchange of compounds of each of these elements in marine aerosols. In terms of the interaction of these cycles, the clearest case for marine aerosols is for N and S, with the likelihood that most of the
NH4+ and
SO42- on the smallest particles are present as a result of the different acid/base characteristics of NH3, SO2/SO3 and their condensed phases. Phosphorus transport from sea to air is clearly related to organic carbon transport, and it is possible that a significant quantity of marine aerosol nitrogen may also be associated with organic carbon. Since there is considerable recycling of trace susbtances across the airsea interface, evaluation of the net (vs gross) input rate of C, P, N, and S compounds from the continents via the atmosphere to the ocean is quite difficult but is necessary to evaluate accurately the importance of airsea exchange in the biogeochemical cycles of these elements.
16.1 INTRODUCTION
In terms of mass, the ocean is probably the major natural source of aerosols. It is also the ultimate sink, either directly, or indirectly via rivers and other inputs, for a significant fraction of continentally derived aerosols. Aerosols over the ocean, which we will refer to as marine aerosols, are thus not only derived from the ocean. Primary and secondary aerosols from the continents can be transported thousands of miles over ocean waters before being removed by wet and dry deposition processes. Sea salt aerosols produced
by the ocean do not necessarily have the same relative chemical composition as the bulk sea-water from which they are formed. Thus airsea exchange of the marine aerosol involves input of continental material to the ocean and input of modified marine source material to the atmosphere. Since there is a considerable amount of recycling of trace substances across the
airsea interface, evaluation of the net input of these substances from the continents via the atmosphere to the ocean is often extremely difficult.
The purpose of this paper will be to point out what we know, what we think we know, and what we almost certainly do not know about exchange of aerosols in both directions across the airsea interface. A discussion will then follow on carbon,
sulphur, phosphorus, and nitrogen in marine aerosols, possible relationships among these biogeochemical cycles relative to marine aerosols, and the airsea exchange of these elements.
16.2 SEA-TO-AIR TRANSPORT OF AEROSOLS
It has been estimated that the sea produces between 1000 and 10,000 Tg/yr of atmospheric sea salt particles with radii less than ~ 20
µm (Eriksson, 1959, 1960; Blanchard, 1963;
Petrenchuk, 1980). Most of the atmospheric sea salt particles in this size range are produced by whitecap bubbles breaking at the sea surface. When a bubble bursts at the seaair interface, atmospheric particles are produced both from a central jet and from the shattering bubble film cap. When the bubbles break, they skim off a thin layer of the sea surface to form the film and jet drops. MacIntrye (1968) has investigated this
`microtome' effect for jet drops and has shown that the material present in the top jet drop was originally spread over the interior of the bubble surface (both bubble cap and the portion submerged) at a thickness equal to approximately 0.05 percent of the bubble diameter. Bubbles in breaking waves range in diameter from a few tens of micrometres to a few millimetres with most being between 100
µm and 1 mm. Blanchard (1963) and Cipriano and Blanchard (1981) found that the diameter of the jet drop was about 10 percent of the diameter of the bubble producing it. Thus a 100
µm bubble will produce a 10 µm atmospheric sea salt particle which is composed of material originally present in the top 0.05
µm of the bubble (or ocean) surface. Similarly the 100
µm jet drop from a 1 mm bubble is derived from the top 0.5 µm of the airsea interface. These relationships for jet drops are summarized in
Figure 16.1.
As expected, the airsea interface has been found to be highly enriched in surface active organic materials and other substances associated with them (MacIntyre,1974;
Liss, 1975; Duce and Hoffman, 1976). Thus it is no surprise that these materials can be considerably enriched on the jet drops relative to their concentration in bulk sea-water a few cm below the surface. This `chemical fractionation' process is very difficult to evaluate by collecting and analysing aerosols over the ocean surface because the ocean is not the only source for particles found there. However, careful laboratory studies and
carefully controlled studies of artificial bubbling and aerosol sampling over the ocean have shown clearly that such substances as iodine, phosphorus, probably organic nitrogen, many organic carbon compounds, and certain heavy metals are enriched on the sea salt particles produced by bursting bubbles. In many cases, however, this enrichment in the ambient marine atmosphere is swamped by the high concentration of these substances already present in the atmosphere from other sources.
Figure 16.1 Relationships among bubble diameter, jet drop diameter, and microtome depth (After
MacIntyre, 1974). Reproduced by permission of John Wiley & Sons, Inc.
We know considerably less about film drops than jet drops. Considering the size of most of the bubbles in breaking waves, it is clear that relatively few jet drops less than
510 µm diameter will be produced, yet there is a considerable number (and mass) of sea salt particles smaller than this size in the marine atmosphere. Cipriano and Blanchard (1981) have shown that most of the smaller sea salt particles are film drops, and that these are derived primarily from the very large bubbles, perhaps 1 mm and larger, which may produce as many as 100 or more film drops per bubble. Submicrometre size film drops are commonly produced. In general, the results of Cipriano and Blanchard (1981) suggest that most of the atmospheric sea salt particles smaller than
510 µm diameter originate as film drops from bubbles larger than 1 mm, while most of the sea salt particles in the atmosphere larger than ~ 20 µm originate as jet drops derived from bubbles larger than ~ 200 µm (see
Figure 16.2).
Figure 16.2 Relative jet and film drop contributions to atmospheric sea salt particles. This is a first approximation only and should not be used in an exact sense (After Cipriano and Blanchard, 1981). Reproduced by permission of American Geophysical Union
Very little is known about chemical fractionation on film drops, although it almost certainly occurs. Clearly, if the data of Cipriano and Blanchard (1981) are correct, most atmospheric sea salt particles with any significant lifetime in the atmosphere (i.e., with r
20 µm) are derived from film drops. Considerable research is needed on the physics and chemistry of film drop production to understand properly sea-to-air chemical exchange of aerosols.
The size of atmospheric sea salt particles is directly related to the ambient relative humidity. For example an atmospheric sea salt particle with a radius of 10 µm when ejected from the ocean surface would have a radius of 4 µm when in equilibrium with an ambient relative humidity of 80 percent and only 2 µm when in equilibrium with a relative humidity of 70 percent, the relative humidity at which solid precipitates may begin to form in the droplet. Note that even at relative himidities of 50 percent or lower, the hygroscopic nature of salt particles assures at least a film of water on the particles. Thus at the relative humidities of 75 to 85 percent often found in the lower marine boundary layer, solution chemistry and
gasliquid interaction processes apply relative to ocean derived aerosols. This must be kept in mind when considering the atmospheric chemistry of marine aerosols.
16.3 AIR-TO-SEA TRANSPORT OF AEROSOLS
Marine aerosols are transported to the sea via wet and dry removal processes. Estimates of the wet (rain) removal of aerosols, or material present on aerosols, are often made utilizing the wash-out factor or scavenging ratio,
W.
where:
|
|
CRD |
|
|
W = |
 |
|
|
|
CA |
|
and CR
is the concentration of material in rain (e.g., µg kg-1), CA
is the concentration of material in air (e.g., µg m-3),
D is the density of air (e.g., ~ 1.2 kg m-3 at 20°C, 1 atm), and
W is dimensionless. Values for W generally range from a few hundred to a few thousand, which roughly means that one gram (or one
cm3) of rain scavenges about one m3 of air. W is, of course, dependent upon a number of factors, including particle size, chemical composition, vertical concentration distribution of the aerosol, vertical extent of the precipitating cloud, etc. Gatz (1977) and Duce
et al. (1979) have shown that W decreases with decreasing particle size for certain trace metals at urban and continental coastal sites. Recent studies by Ng and Patterson (1981) indicate that
W for Pb, a small particle element, is greater than that for continental dust at a mid-Pacific Ocean site. Thus effects of size, particle chemistry, and other factors make it difficult to predict
W more accurately than in the range between a few hundred and a few thousand.
Dry deposition of aerosols is often estimated utilizing the deposition velocity,
vd,
where:
vd = F/M
and M is the mass of aerosol (or material on aerosol) (e.g., µg
cm-3), F is the flux of particles to the surface (e.g., µg cm-2 s-1),
and vd is the deposition velocity (e.g., cm s-1).
vd can be estimated from theoretical considerations or it can be measured in laboratory or field situations. Considerable controversy exists about the validity of dry deposition velocity values determined from measurements to buckets or dry plates (see for example
Sehmel, 1980). Effective deposition velocities for particles in the stable aerosol size range near ground are often found to be near 1 cm s-1,
but this varies considerably with particle size, wind speed, and surface roughness.
In a laboratory wind tunnel experiment Sehmel and Sutter (1974) investigated aerosol deposition velocity over a water surface as a function of wind speed and particle size. Their results are presented in
Figure 16.3. Slinn and Slinn (1980) developed a theoretical two-layer model for the prediction of particle dry deposition to natural waters. They pointed out that if particle growth by water vapour condensation occurs in the humid regions near an air-water interface, the deposition velocity of certain types of particles (they used
(NH4)2SO4 as an example) with dry radii near 1 µm is nearly independent of particle size and can be approximated by
vd = 1.3 x 10-3
, where
is the mean wind speed in cm s-1.
At winds of 510 ms-1, this results in an aerosol vd of ~ 1 cm
s-1 (see Figure 16.4).
Figure 16.3 Atmospheric particle dry deposition velocity,
vd, to a water surface as a function of particle size and wind speed from wind tunnel studies (After Sehmel and Sutter, 1974). Reproduced by permission of J.
Rech. Atmos.
Aside from the size range ~ 1 µm described above, both Sehmel and Sutter (1974) and Slinn and Slinn (1980) predict a large difference in
vd for very small (e.g., 0.1 µm) vs very large (e.g., 20 µm) particles. For example, for a 20 µm and 0.1 µm radius particle at a wind speed of 10
m s-1,
Sehmel and Sutter (1974) predict 40 cm s-1 and 0.01 cm s-1
respectively and Slinn and Slinn (1980) predict 10 cm-1 and 0.01 cm
s-1 respectively. Thus, if even a very small percentage of a substance is present over the ocean on the larger particles, these larger particles may dominate the dry deposition to the ocean.
As an example, recent studies of lead in the atmosphere at Enewetak Atoll in the mid-Pacific Ocean and its deposition to the ocean surface during the SEAREX (Sea/Air Exchange) Program have shown that the mass median radius of the Pb containing particles is ~ 0.2 µm) (R. Duce, unpublished data).
However, the measured dry deposition velocity of the Pb to a flat plate several metres above the ocean surface was ~ 0.7 cm s-1
(C. C. Patterson, unpublished data), much greater than that expected from the mass median radius. However, if the measured Pb particle size distribution was applied to the theoretical deposition velocity relationships derived by Slinn and Slinn (1980), the predicted total Pb dry deposition agreed well with that observed. The small percentage of Pb on the largest particles was controlling the Pb deposition. Thus the size distribution of any substance of interest on aerosols
and the deposition velocity as a function of particle size must both be known to adequately approximate the dry deposition of any substance to the ocean.
Figure 16.4 Theoretically derived particle dry deposition velocity,
vd, to a water surface as a function of particle size and wind speed (After Slinn and Slinn, 1980). Reproduced by permission of Pergamon Press Ltd.
Perhaps the most serious problem in interpreting wet and dry deposition calculations for any substance entering the ocean is evaluating gross
vs net deposition. In evaluating the airsea exchange portion of biogeochemical cycles, we are most often interested in the net input of material into (or in some cases out of) the ocean. Measurements of rain concentrations and measurements or calculations of dry deposition from concentration and dry deposition velocities only give us gross input values. From this we must remove the recycled component coming from the ocean itself. If atmospheric sea salt
particles had the same chemical composition as bulk sea-water, this would be simple, as we could simply use the
X:Na ratio in the atmospheric sample and in the ocean and factor out the marine component. If, however, the marine component does not have the composition of sea-water, this approach will not
workit will over-estimate the net flux into the sea. There are several approaches that give some promise of being able to evaluate this problem, and these will be discussed in the following sections on C, S, P, and N exchange across the seaair interface.
16.4 CARBON
Carbon can be present on marine aerosols in essentially three forms-carbonate-bicarbonate, organic, and elemental carbon or soot. Except in areas downwind from continental or island limestone or coral deposits, carbonate is a small fraction of marine aerosol carbon and will not be considered further here.
Hundreds of different organic compounds are present on atmospheric particles and each has its own characteristic physical and chemical properties and associated atmospheric sources, residence times, and sinks. Data are very sparse on individual compounds in the marine aerosol, however, so we will restrict our discussion to total organic carbon.
Organic carbon in the marine aerosol has been measured by several investigators in the past few years. In many cases the concentrations reported for organic carbon probably include elemental carbon as well. Reported values in the near surface marine atmosphere
(1090 metres) are presented in
Table 16.1. Note that the mean concentrations observed are very similar at the various marine sites in the northern and southern hemispheres, with a range of mean concentrations of only 0.2 to 0.9 µg m-3.
The overall mean concentration at all eleven sites is 0.48 ± 0.20 µg m-3.
Thus, a concentration of about 0.5 µg m-3 is representative of the organic (or total) carbon content of the marine aerosol over much of the world ocean. Hoffman and Duce (1977) found that about 80 percent of the carbon was on particles with radii < 0.5 µm at Bermuda, Hawaii, and Samoa sites (see
Figure 16.5). They suggested this was the result of gasparticle conversion reactions in the atmosphere. The distribution of organic carbon and Na as a function of particle size for laboratory generated sea salt particles and for a typical atmospheric sample from Bermuda is represented in
Figure 16.6. While the Na size distribution is quite similar in the ambient and laboratory aerosol, the organic carbon distribution is quite different. The organic carbon in the laboratory generated sea salt particles is enriched several hundred-fold over the sea-water concentration (relative to Na) and is present on generally the same size particles as the Na. This suggests that the large-particle organic carbon is probably present on the sea salt aerosols when they are produced by the ocean, while the small particle, or dominant, organic carbon comes from some other source.
Table 16.1 Organic carbon concentration in the marine aerosol
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
0.150.47 | |
|
|
|
|
|
|
0.331.6 | |
|
|
|
|
|
|
0.200.86 | |
|
|
|
|
|
|
0.150.78 | |
|
|
|
|
|
|
0.380.48 | |
|
|
|
|
|
|
0.360.43 | |
|
|
|
|
|
|
0.731.2 | |
|
|
|
|
|
|
0.220.74 | |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
0.130.41 | |
|
|
|
|
|
|
0.070.53 | |
|
Barger and Garrett (1976) | |
|
|
|
|
|
|
|
Figure 16.5 Organic carbon concentration as a function of marine aerosol particle size (After Hoffman and Duce,
1977). Each symbol indicates one sample. Reproduced by permission of American Geophysical Union
Recent SEAREX studies by Chesselet
et al. (1981) utilizing stable carbon isotopes lend support to the suggestion that the small particle organic carbon is not of marine origin.
Figure 16.7 shows the Na and organic carbon concentration and the 
13C value for a size-separated aerosol sample collected at Enewetak Atoll.
13C
values for the smallest particle are 26‰ to -28‰. Chesselet et al. (1981) point out that this range is similar to
13C values for continental vegetation, coal, and the products of petroleum combustion,
26 ± 2
‰, suggesting that the small-particle carbon is of continental origin. The 13C values for the large-particle carbon are
18 to -22‰. This is similar to 813C of marine organic carbon, which is generally
21 ± 2‰ in low latitude regions (40°S50°N), suggesting that the large-particle organic carbon in the atmosphere is of marine origin.
Figure 16.6 Comparison of organic carbon and sodium concentration as a function of particle size on ambient marine aerosols and laboratory generated sea salt particles (After Hoffman and Duce,
1977). Reproduced by permission of American Geophysical Union
Figure 16.7 Sodium, organic carbon, and 13C as a function of marine aerosol particle size.
13C is defined in the usual notation:
(From data reported in Chesselet et al.,
1981). Arrows indicate less than or equal to.
The global production of aerosol organic carbon is rather poorly known. Duce (1978) estimated a global source strength of organic carbon on primary aerosols of ~ 56 Tg yr-1, of which about half was from natural sources
(the ocean, crustal weathering, forest fires) and about half from pollution sources. Duce (1978) estimates that about 80 to 160 Tg yr-1 of particulate organic carbon results from gas to particle conversion processes, resulting in a total global production rate of ~ 140
220 Tg yr-1. This agrees well with the estimate of Jaenicke (1978) of ~ 200 Tg yr-1, ~ 150 Tg yr-1 from gas-particle conversion and 50 Tg yr-1 from direct production.
A complication here is elemental carbon. Elemental carbon has been found in marine sediments in the mid-Atlantic and Pacific Oceans (~ 0.02
0.1 percent dry weight; Smith
et al., 1973; Griffin and Goldberg, 1975). This material originates from the combustion of coal, wood, and petroleum products (Griffin and Goldberg, 1979). It has been measured extensively in urban areas, but few data are available in remote regions. Rosen
et al. (1981) find elemental carbon concentrations ranging from ~ 0.1 to 0.9 µg m-3
at Barrow, Alaska, with the higher concentration being observed in the winter months. The high concentrations coincide with apparent long range transport of pollution across the Arctic during the winter (Rahn and McCaffrey, 1980) and indeed are near typical concentrations in many urban areas (Rosen
et al., 1981). No data for elemental carbon are available from remote marine regions, but the concentrations are probably considerably less than at Barrow in the winter since filters are typically still white in mid-Pacific locations even after the passage of 104
m3 of air.
The global production of elemental carbon by fires may be very high. Ryan and McMahon (1976) indicate that as much as 40 percent of the particles produced during burning of temperate forests and agricultural wastes may be elemental carbon. Seiler and Crutzen (1980) estimated that as much as
90180 Tg yr-1 of aerosol carbon may be produced in this way. Note this is in the range of the total production of aerosol organic carbon estimated by Duce (1978) and Jaenicke (1978). During discussions at this SCOPE meeting it was suggested that perhaps 10 percent of the
90180 Tg yr-1 of aerosol carbon may be emitted as fine particles that can be transported great distances. Elemental C may thus play a very important role in the global aerosol carbon budget. The historical record of burning is maintained in coastal marine and in lake sediments. As emphasized by UNESCO (1980), the fate of this elemental carbon may be important for the global carbon budget, and its production and distribution should be investigated in more detail.
Information on the net input of organic carbon to the ocean from the atmosphere is essentially nil. We have virtually no data on organic carbon in marine rains. Neumann
et al. (1959) reported 1.73.4 mg kg-1 organic carbon over Sweden, while Williams (1967) reported 0.7 mg kg-1 just north of Samoa. Gagosian (1981) found a mean organic carbon concentration of 0.64 ± 0.48 mg kg-1 in 10 samples collected during the SEAREX Program at Enewetak Atoll. If we assume a global value of
0.51 mg kg-1 organic carbon (i.e., a
W of 1000 to 2000) and a rainfall over the ocean of 3.9 x1017 kg yr-1 (Baumgartner and
Reichel, 1975), there is a total annual input of ~ 200
400 Tg yr-1. This is very large, about 24 times the estimates of the riverine input of organic carbon to the ocean (Duce and
Duursma, 1977) and higher than the estimates of global organic carbon aerosol production described above (which may well indicate that the production estimates are too low). This calculated input is, of course, a gross input to the ocean and is likely to be considerably higher than the net input because much of it is probably recycled carbon from the ocean surface, present on the largest and most scavengeable sea salt particles.
We have a similar problem distinguishing net from gross dry deposition, as it is very possible that the gross deposition will be dominated by the relatively small concentration of organic carbon on the large particles, which have high dry deposition velocities. For example, if we assume we have 0.4 µg m-3
of organic carbon on particles with radii ~ 0.20.3 µm with a
vd of ~ 0.05 to 0.5 cm s-1
(see Figure 16.4), the global dry deposition of this fraction would be ~ 2 to 20 Tg yr-1. If we assume the other 0.1 µg m-3
of organic carbon is on 15 µm particles with a
vd of 13 cm s-1,
this size fraction would contribute 1030 Tg yr-1 to the ocean. There are no field data on dry deposition of organic carbon to the ocean at this time.
The use of carbon isotopes offers an excellent opportunity for untangling the relative role of small (net input) particles and large (recycled material) particles and may allow us to evaluate the net input of organic carbon to the ocean. Experiments to do this are planned by R. Chesselet and his group at the SEAREX Program study in Samoa during the summer of 1981. This group will measure organic carbon concentrations and
13C values in rain, in dry deposition, and in the ambient aerosol as a function of particle size. Results of the analyses of these samples should indicate clearly which size particles are primarily involved in the wet and dry deposition processes and thus what the net input of organic carbon is to the ocean.
16.5 SULPHUR
Sulphate in the marine aerosol is generally present in concentrations higher than expected for particles generated from sea-water, i.e., the
SO42-:Na ratio in marine aerosols is almost always higher than the value of 0.25 found in sea-water
(Buat-Menard
et al., 1974; Gravenhorst, 1978; Bonsang et al., 1980). Several investigators in the 1950s and 60s suggested this `excess' sulphate was the result of chemical fractionation processes during bubble bursting and sea salt particle production, but it has generally been accepted that the excess sulphate has a different source. In a recent paper, however, Garland (1981) found small enrichments
(1030 percent) of sulphate during bubble bursting using radiotracers in a laboratory study. Thus this process cannot be ruled out completely as a potential source for some of the excess sulphate observed.
Several studies have shown that the sulphate size distribution in the marine aerosol is bimodal, with one fraction present on sea salt size particles and the other on submicrometre size particles. Gravenhorst (1978) found that the mass median radius of the fine-particle sulphate on marine aerosols over the North Atlantic was ~ 0.4 µm while that for sea salt sulphate was ~ 2.5 µm. Lawson and Winchester (1978, 1979) found similar results in the southern hemisphere. In a review of recent studies of
SO42- over the Atlantic, Georgii (1978) points out that the contribution of non-sea salt sulphate to total sulphate in marine aerosols may be 60 to 80 percent, with as much as 1 µg m-3
of excess sulphate over the North Atlantic. Huebert and Lazrus (1980b) found many of their samples
over the North and South Pacific had SO42-:Cl ratios similar
to sea-water, suggesting the excess SO42- was rather small in this area. Bonsang
et al. (1980) have shown that the SO42-:Na ratio on particles with radii > 1 µm is the same as in sea-water, and these larger sulphate particles are clearly derived from the ocean, similar to organic carbon, but in this case with no chemical fractionation. Thus it is possible that `excess sulphate' and 'fine-particle sulphate' are essentially synonymous in the marine aerosol. Some recent data on fine-particle or excess sulphate are presented in
Table 16.2. Note that excess
SO42- is generally 0.5 to 1 µg m-3 over the North Atlantic and apparently somewhat lower over the North Pacific. It is considerably lower in the southern hemisphere, generally
0.050.1 µg m-3.
The data of Bonsang et al. (1980) in the southern Indian Ocean are an exception to this. These data were obtained from 20°S. to 60°S., and there is no explanation at
this time for the high excess sulphate concentration observed in that region. Lawson and Winchester (1979) point out that fine-particle sulphate in marine air at Punta Arenas, Chile is much lower in the winter (0.01 µg
m-3)
than the summer (0.05 µg m-3), suggesting possible seasonal (and perhaps biological) effects. Meszaros (1978) suggested that there may be a
SO42-maximum over the North Atlantic at ~ 40°N. and that this may be a result of conversion of high concentrations of SO42-, from pollution origin in the
westerlies.
Table 16.2 Fine particle or excess sulphate in marine aerosols-some recent data
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
0°10°N. | |
|
|
|
|
0.81 ± 0.49 |
|
|
|
|
|
Meinert and Winchester (1977) | |
|
|
|
|
|
|
|
|
|
|
0.41 ± 0.28 |
|
|
|
|
|
|
|
|
|
| Central East Pacific, |
|
|
|
0°15°S. | |
|
|
|
|
|
0.050.1 | |
|
Darzi and Winchester (1981) | |
|
|
|
|
|
|
|
|
Lawson and Winchester (1979) | |
|
Punta Arenas, Chile |
|
|
Lawson and Winchester (1979) | |
|
|
|
|
|
The source of the small-particle SO2 is
a matter of considerable interest. That much if not most of it is derived from
gasparticle conversion involving SO2, is almost certain, but the source of the SO2,
is not. The sulphur cycle, and gasparticle sulphur exchange, has been discussed at great length by others and will not be developed here.
Bonsang
et al. (1980) contend that in marine regions far from continents the SO2, is not continentally derived but is probably from the oxidation of marine derived reduced forms of sulphur such as
dimethylsulphide. Maroulis
et al. (1980) suggest that much of the background SO2, results from oxidation of COS. The higher concentrations of fine-particle
SO42- over the North Atlantic support a continental, and probably anthropogenic, source for the SO2, that is converted to SO42-
in this region.
Taylor et al. (Chapter 4, this volume) have also discussed the bimodal size distribution of aerosol sulphate and their conclusions are similar to those reached in this paper. Of particular interest in the paper by Taylor
et al. (Chapter 4, this volume) is their discussion of the relationships among
NH4+ , SO42-, and rain-water acidity. Even in remote marine areas there do not appear to be sufficient basic substances to neutralize the fine-particle H2SO4;
the only basic material evaluated to date is NH3. No data are available on the presence, for example, of basic amines or other basic gases in the marine atmosphere.
In terms of seaair exchange of SO42- consideration of the primary sea salt sulphate would appear to be simple and straightforward. That may not be the case, however. For example, Lawson and Winchester (1979) point out that in the southern hemisphere marine atmosphere, the highest fine particle
SO42- concentrations were observed along the coast during the most intense periods of large-particle sea salt
sulphate. This suggests there may be some direct contribution to fine-particle
SO42- from the ocean, and agrees with Cipriano and Blanchard's (1981) recent findings that the ocean can apparently produce reasonably large concentrations of
sub-micrometre size aerosols.
With no intention of summarizing the literature, it is interesting to comment briefly on sulphate in marine rains. Numerous investigators have found excess sulphate in marine rains (e.g., Eriksson, 1957;
Junge, 1963; Nyberg, 1977; Asman and Ridder, 1981, among others). Perhaps the most geographically
extensive set of data to date were reported by Kroopnick (1977). Rain from 19 stations in the North and South Atlantic and Pacific Oceans was analysed for
SO42- and Cl-. The mean SO42-:Cl- ratio observed
for precipitation at each sea level marine site at which eight or more samples were collected is presented in
Table 16.3. Data are also given for Mauna Kea, Hawaii, 3500 m above sea level. The sea-water
SO42-:Cl- ratio is 0.14 and the mean rain-water
SO42-:Cl ratio at all the sites (except Mauna Kea) is 0.20 ± 0.09. While excess
SO42- is certainly observed, the data from this study suggest it is not observed at all sites all of the time. Note also that there is independent evidence for CI loss from marine aerosols, probably partly as HCl due to acidification of the marine aerosols by, among other things, H2SO4 formation.
Thus Cl itself is not always a conservative indicator of the sea salt contribution for any substance.
Table 16.3 The SO42-:Cl ratio in marine rains (From data in Kroopnick, 1977). Reproduced by permission of University of Hawaii Press
|
|
Sampling |
Number of |
Mean SO42-:Cl- |
|
site |
observations |
ratio |
|
|
|
|
|
|
|
|
|
0.27 + 0.27
|
|
|
|
|
0.20 ± 0.08
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Excess sulphate in rain should be a maximum in areas where the excess sulphate in the marine aerosol is highest, e.g., over the North Atlantic. This is not apparent in the data of
Table 16.3. However, Asman and Ridder (1981) found a mean excess
SO42- concentration of 2.6 mg litre-1 for a series of samples collected over the North Atlantic from a Dutch weathership at 60°N., 2°E. In one low-pH sample (3.8), the excess sulphate was 16 mg litre-1. Nyberg (1977) found a mean excess
SO42- concentration of 1.2 mg litre-1 for 26 rain samples collected from ships in the North Atlantic from 45° to 65°N. and 33°W. to 4°E. In contrast, Pszenny and Duce (1981) collected 33 rain samples under onshore wind conditions from a tower about 50 metres above sea level on the windward shore of American Samoa in the South Pacific. They found `a hint of an excess
SO42- component' of ~ 0.04 mg litre-1. With total
SO42- concentrations averaging 1.3 mg litre-1, this was essentially a statistically insignificant enrichment. Thus, there is some evidence that excess sulphate in rain is considerably higher in some marine regions that others, as expected. A number of detailed studies of rain chemistry over the world ocean are presently underway, and considerably more data should be available soon.
Interestingly in eleven global cycles of atmospheric sulphur published since 1963, all use the same value for the sea spray sulphate-sulphur source strength, 44 Tg yr-1. This number is derived from the excellent pioneering papers of Eriksson (1959, 1960) in which he derived the global sea salt flux (~1200 Tg yr-1). A fundamental assumption in the derivation of the 44 Tg yr-1 figure was that the amount of sea salt
SO42- deposited on continents was 10 percent of that redeposited directly in the ocean from the atmosphere. The 10 percent figure was derived from the total chloride entering the ocean each year via river run-off (it was all assumed to be of sea salt origin deposited on the continents via rain and dry deposition) and Eriksson's calculation of the annual deposition of atmospheric sea salt (and thus chloride) directly to the ocean. Using the
Cl-:SO42- ratio in sea-water and the riverine chloride input to the sea, it was calcualted that 4 Tg yr-1 of sea salt
SO42--S entered the ocean via rivers. This results, then, in a global production of 44 Tg yr-1 of atmospheric sea salt
SO42-S. Clearly the 44 Tg yr-1 is directly related to the assumption that 10 percent of the atmospheric sea salt is deposited on land. If the percentage is less than 10 percent, the 44 Tg yr-1 figure is too low. Similarly, if the 1200 Tg yr-1 figure for global atmospheric sea salt production is too low, the estimate of 44 Tg yr-1
SO42-S from sea salt is also too low.
Blanchard (1963) re-evaluated the global sea salt production figures using the general approach of Eriksson, but utilizing actual measurements of the temporal and geographical variation of wind speeds over 5° to 20° latitude/longitude squares over the global coean (Eriksson assumed a 12 kt wind over the entire ocean). Blanchard estimated a global sea salt production of 10,000 Tg yr-1, about 8 times that of Eriksson. Petrenchuk (1980) calculated a value of 1300 Tg yr-1 salt removal by rainfall over the ocean each year. Relatively little consideration was given to dry deposition of sea salt, which may be equal to or greater than rain removal for sea salt. Thus it is quite possible that the commonly used estimate of 44 Tg yr-1
SO42--S introduced into the atmosphere on sea salt particles may be low by a factor of 2 to 8.
16.6 PHOSPHORUS
Much less is known about phosphorus in the atmosphere than the other three elements we are considering. In his discussion of the global phosphorus cycle, Pierrou (1976) devotes one paragraph to atmospheric phosphorus, and concludes `Unfortunately no measurements or estimates seem to have been published on this subject.' There are at least a few more data available now as a result of several recent papers on the global atmospheric phosphorus cycle (Graham and Duce, 1979) and on the sea as a source for atmospheric P (Graham
et al., 1979).
There is no information on the presence of any significant concentration of vapour phase phosphorus compounds in the atmosphere, although certainly some do exist, for example certain pesticides. It is assumed, however, that the atmospheric phosphorus cycle is largely controlled by aerosol P. Graham and Duce (1979) calculated a global atmospheric P cycle, which is shown in
Figure 16.8. The primary sources of atmospheric P are crustal weathering, the ocean, and pollution. The major pollution sources are from phosphate manufacturing and processing, from the soil as a result of man's activities, and from coal combustion. A detailed breakdown of sources, sinks, and the entire cycle is given in the original paper.
Phosphorus concentrations in the marine aerosol have been measured by Graham and Duce (1979) and by Graham
et al. (1979). These data are summarized in Table
16.4. The effect of the continents, particularly the Sahara dust plume, is clearly evident in the data. Concentrations over the mid-Pacific appear to be ~ 0.5 ng m-3.
Concentrations north of'- 45°N. over the Atlantic are similar to this, but concentrations closer to North and South America are typically 5-10 ng m-3,
with 50 ng m-3 common in the Sahara dust plume. Graham and Duce (1981) found the major mass of total P in the marine aerosol to be on particles with radii of
1 3µm, with perhaps
1020 percent of the mass on submicrometre particles. The calculated gross input of atmospheric P to the ocean is ~ 1.4 Tg yr-1 with a net input of ~ 1.1 Tg yr-1. The global annual mobilization of aerosol phosphorus is ~6 Tg yr-1 (see
Figure 16.8).
Figure 16.8 The atmospheric phosphorus cycle. The numbers in parentheses are the estimated inputs of sea-water soluble phosphorus through rivers and through the atmosphere (After Graham and Duce, 1979). Reproduced by permission of Pergamon Press Ltd.
Graham and Duce (1982) have considered the relative importance of the atmospheric transport of phosphorus to the phosphorus content of the waters in the western North Atlantic. They found that about 35 ± 15 percent of the total phosphorus present in marine aerosols over the western North Atlantic was soluble in sea-water within 12 hours. A somewhat lower percentage was found for Sahara dust aerosols. Since the total phosphorus input into an area of the ocean bounded by the North American coast, 25°N., and 65°W. was calculated to be 0.5 to 1.0 x 1010g yr-1, at least 0.2 to 0.4
x 1010 g yr-1 of phosphorus available for rapid nutrient use was deposited. This compares with ~ 3 x 1010 g P yr-1 which is estimated to enter estuarine areas through east coast North American rivers, i.e., atmospheric input is at least 10 percent of riverine input to this region. However, the riverine input is added to estuaries and coastal waters, and it is uncertain how much of it may be deposited there. Atmospheric deposition occurs over a wide area, including much of the nutrient-poor Sargasso Sea. Thus the atmosphere may be a more important source of nutrient P to open ocean regions than the gross figure would indicate.
Table 16.4 Phosphorus concentrations
in the marine aerosol (From data in Graham and Duce, 1979)
|
|
|
|
|
|
|
|
|
|
Western North Atlantic (30°45 °N.) | |
|
1.0 15 | |
|
|
|
0.6
9.2 | |
|
Near Shore North America
Atlantic | |
|
15
60 | |
|
|
|
3 80 | |
| Central North Atlantic (45°60 °N.) |
|
0.6
1.0 | |
|
West coast of South America | |
|
6 20 | |
|
|
|
0.4
0.9 | |
|
|
|
0.4
0.8 | |
|
Central Pacific (20 °N15 °S.) | |
|
0.04 0.9 | |
|
Sutcliffe et al. (1963), Bruyevich and Kulik (1967), and MacIntyre and Winchester (1969) all found that in laboratory bubbling studies, phosphate was significantly enriched on sea salt aerosols produced by bursting bubbles compared with sea-water. Using a bubble generator in coastal sea-water, Graham
et al. (1979) found that total phosphorus enrichment ranged from 4 to 170, i.e., P:Na on the sea salt aerosols was 4 to 170 times that of the sea-water. In their studies, Graham
et al. (1979) measured total phosphorus (using persulphate oxidation), and reactive phosphorus (that soluble in distilled water, primarily
PO43- , HPO42-, etc.). The difference between these two was termed `organic' phosphorus. It was found that the enrichment of reactive phosphorus was only a factor of
28 above sea-water whereas that of organic phosphorus was often
100200, suggesting most of the enriched phosphorus is associated with surface-active organic substances. Factor analysis and multi-variate regression analysis of sodium, aluminium (an indicator of crustal sources), excess vanadium (an indicator of anthropogenic sources), reactive P, and organic P in a number of samples collected over the eastern equatorial Pacific strongly suggested that crustal weathering was the primary source for the reactive P, the ocean was the primary source for organic P, and pollution apparently represented a minor source for both forms in this area. The mean atmospheric phosphorus concentrations observed in this region were 4.1 ± 1.2 ng m-3
for reactive P and 2.8 ± 1.6 ng m-3 for organic P, i.e., ~ 60 percent reactive P. In Hawaii and Samoa the reactive P, the form clearly from continental areas, was ~ 35 percent and 20 percent respectively, reflecting the increasing distances from continental regions (Graham and Duce, 1981). On the basis of these studies it would appear that organic P over the ocean is largely recycled from the ocean whereas reactive P is from continental areas. The input of reactive P to the ocean is thus of primary interest in the P biogeochemical cycle. More detailed measurements of the wet and dry input to the ocean of atmospheric reactive phosphorus must be made before an accurate assessment can be made of the importance of atmospheric P transport to the marine chemistry of surface waters. At present it would appear from
Figure 16.8 that perhaps ~10 percent of the dissolved form of phosphorus entering the oceans each year is via atmospheric transport.
16.7 NITROGEN
Nitrogen in marine aerosols is somewhat similar to sulphur, i.e., gas-particle conversion plays a significant, in this case major, role in determining the
NH4+ and NO3 content of marine aerosols. Evidence suggests that the ocean plays a relatively minor role in the direct production of marine aerosol
NH4+ and NO3, although data from mid-ocean regions are sparse.
In the case of NO3, much of the early aerosol
NO3 data are
suspect due to artifact NO3 formation from NO2 on the filters utilized to collect particulate
NO3
(Appel
et al., 1979). However the early data of Junge (1957) from Hawaii appear to show little evidence of this problem. Huebert and Lazrus (1980a) showed there was no correlation between aerosol
NO3 and Cl collected from an aircraft in the marine boundary layer over the North and South Pacific, indicating that sea salt aerosols were not a significant source for
NO3. Similarly Huebert (1980) found no correlation between aersosol
NO3 concentration and sea state in the central Pacific. Pszenny and Duce (1981) found no correlation between
NO3 and
Na+ in marine rains collected in American Samoa (see Figure
16.9). Huebert and Lazrus (1980a) point out that, assuming there was no chemical fractionation of sea-water
NO3 during sea salt particle production, sea spray alone could be expected to contribute no more than
0.10.2 ng m-3
of particulate NO3, whereas measured marine aerosol
NO3 concentrations are typically 0.1 to 0.3
µg m-3
(see Table 16.5). NO3 showed no significant correlation with HNO3,
SO42-, NH4+ or Rn in the studies of Huebert and Lazrus (1980b).
Figure 16.9 Nitrate versus sodium concentration in rain samples collected in Samoa (After Pszenny and Duce, 1981). Arrows indicate less than or equal to.
On the other hand Sequeira (1981) found a significant correlation
(r = 0.81, 68 data pairs) between NH4+ and NO3 in precipitation collected at Valentia on the Irish coast. He also found a stoichiometric ratio for
NH4+ : NO3 of near unity and suggested NH4+ and
NO3 ions are tied up as NH4+NO3. Other evidence does not support this, however. Parungo et al. (1981) and Gravenhorst
et al. (1979, 1981) found that most of the aerosol NO3 was concentrated in the sea salt size range
(110 µm). Gravenhorst
et al. (1979, 1981) found most of the NH4+ on particles with radii < 0.5 µm. the same size which contains the excess sulphate. They suggest that this indicates that conversion of gaseous NH3 and
HNO3 over the ocean does not result in NH4NO3.
Table 16.5 Nitrate concentrations in marine aerosols (µg m-3
STP)