Chemical Forms Sources And Concentration Levels

As shown in Figure 2, some I I different sulfur compounds define over 98% of the sulfur speciation in the atmosphere. Those in which sulfur is found bonded to either

2 CHEMICAL FORMS, SOURCES, AND CONCENTRATION LEVELS 129

TABLE 1 Global Sulfur Emission Inventory

DMS

so2

so42-

Other Reduced Sulfur

Total

Northern Hemisphere

Combustion of fossil fuel

0.37-0.42

65-90

1.8-2

1.1-1.4

68-94

Oceans

5.8-9.7

0.1-0.4

5.9-10

Volcanoes

2.4-6.6

1.4-2.9

0.47-1.2

4.3-11

Other0

0.032-0.6

1.3-1.6

1.3-2.5

0.18-1.6

2.8-6.2

Anthropogenic total

0.37-0.42

65-90

1.9-2.1

1.1-1.5

70-96

Natural total

5.8-10

2.4-6.6

2.6-5.3

0.7-3

12-25

N.H. total

6.2-11

69-98

4.5-7.4

1.8—4.5

81-121

Southern Hemisphere

Combustion of fossil fuel

0.02

7.1-9.2

0.2

0.12-0.13

7.4-9.6

Oceans

9.2-15

0.04-0.36

9.2-15

Volcanoes

1-2.6

0.6-1.1

0.13-0.43

1. T—4.1

Other0

0.021-0.2

1-1.3

0.8-1.6

0.096-0.53

2-3.7

Anthropogenic total

0.02

7.1-9.5

0.24

0.13-0.18

8.5-10.9

Natural total

9.2-15

1-2.6

1.4-2.7

0.26-1.3

12-22

S.H. total

9.2-15

9.1-13

1.6-2.9

0.39-1.5

20-33

Global total

15-26

78-111

6.1-10

2.2-6

102-154

" Includes biomass burning.

" Includes biomass burning.

For DMS, the data indicate that 97% of the global flux of 15 to 26 Tg S/yr results from emissions from the ocean (Berresheim et al., 1995). Of this marine total, 61% is from the SH and 39% from the NH. The next largest contributor is split between wetland sulfur releases and those from anthropogenic/industrial emissions. By contrast, for S02 the global flux is largely defined by NH emissions (e.g., ~ 88%). This reflects the major contribution made from fossil fuel burning in the highly industrialized NH [for details see Spiro et al. (1992) and Hameed and Dignon (1992)]. Anthropogenic emissions from the SH make up still another 9% of the global total for S02 with volcanic emissions making up most of the remainder. This means that volcanic emissions define the second largest primary S02 global source but comprise, on average, only ~7% of the total. (Note, during years involving major eruptions, this source is substantially larger.) As noted earlier in the text, a very recent addition to the global inventory of S02 are emissions from ships. This source is currently estimated to be 2 to 4% of the total. In the case of other reduced sulfur (i.e., H2S, CS2, and OCS), the fluxes from the NH and SH are within a factor of 3 of each other and are made up of significant contributions from both natural and anthropogenic sources.

Overall, Table 1 clearly indicates that insofar as gross amounts of sulfur are concerned, fossil fuel combustion in combination with industrial emissions represent the single largest sulfur source in the NH. This is followed by nearly equal contributions from volcanoes and marine emissions. Still smaller emissions can be attributed to biomass burning and wetlands, and to direct release from plants and soils. By contrast, in the SH ocean emissions of sulfur are nearly 2 times larger than those from fossil fuel combustion, the latter being followed by volcanic and ship emissions. [For a more in-depth survey see Bates et al. (1992b).]

A centrally important conclusion that can be extracted from Table 1 is that anthropogenic sources of sulfur have overtaken natural sources in the NH. For example, of the lOlTgS/yr (on average) released in the NH, nearly 83TgS/yr (i.e., 82%) can be assigned to human activities. By contrast, for the SH only 37% can be similarly assigned. The fact that human-related activities are now overshadowing the natural sulfur cycle in the NH raises some serious questions as to what environmental price tag is being paid for such a transgression? Shifts in atmospheric acidity and atmospheric turbidity in the NH have now been documented, and new concerns are being voiced about the impact of elevated sulfur on regional weather patterns and in long-term climate changes [Charlson et al. (1992)]. Thus, sulfur, like other trace chemical substances in our environment, when present in too large amounts has the potential for creating deleterious consequences for humankind.

The global source strength of a sulfur species, in combination with its areal source distribution and atmospheric lifetime, typically define the species concentration level at a given location. This being true, given that several sulfur species have the same integrated global emission fluxes, the species having the more regionally focused source tends to generate the highest concentration levels. On the other hand, the longer the lifetime of a sulfur species, all other things being equal, the higher its concentration and the lower its variability. For example, as shown in Table 2, the very long lived species OCS (i.e., 2 to 4 years) has a global median concentration

TABLE 2 Observed Mixing Ratio of Atmospheric Sulfur Species

Background

TABLE 2 Observed Mixing Ratio of Atmospheric Sulfur Species

Background

Sulfur

Background Marine

Continental

Polluted Continental

Species

Boundary Layer

Boundary Layer

Boundary Layer

Typical

Typical

Typical

Range

Median

Range

Median

Range

Median

h2s

2-30

10

5-150

60

80-810+

365

DMS

15-300

65

1-20

8

0-10?

<5?

OCS

400-800

600

300-7000

550

300-1800

545

CS2

1-35

10

15-50

30

65-370

190

so2

20-50

35

20-1000

500

150-6000+

1500

h2s

7-13

9

1-7

6

IDTA

IDTA

DMS

0-20

2

IDTA

IDTA

IDTA

IDTA

OCS

1-8

4

<3-18

7

IDTA

IDTA

CS2

IDTA"

IDTA

IDTA

IDTA

IDTA

IDTA

SO/

10-80

30

60-260

100

IDTA

IDTA

" IDTA, Insufficient data to assess.

" IDTA, Insufficient data to assess.

centered around 500pptv and varies by no more than a factor of 2 on a regional and global scale. At the opposite end of this scale, Table 2 reveals that S02, which has both highly focused continental sources and a relatively short lifetime (i.e., 0.5 to 9 days), displays some of the largest gradients of any sulfur compound with concentrations ranging from 35 to 5000 pptv. Of particular significance is the gradient between background continental regions and remote marine areas where factors of nearly 15 are seen. By contrast, DMS, which has a somewhat similar lifetime to S02, typically shows far more modest boundary layer concentration gradients. This is in keeping with DMS having a far less focused source region. Interestingly, due to the combination of its short lifetime, efficiency of vertical mixing, and the absence of high altitude sources, DMS unlike S02 displays very significant altitudinal gradients. Similar arguments to those given for OCS, DMS, and S02 can be used to explain the concentration levels and gradients observed for other sulfur species.

3 TRANSFORMATIONS

As stated earlier, three of the major players in the atmospheric sulfur cycle are DMS, S02, and S042~. Reflecting this conclusion, the present section on transformations will primarily focus on the processes by which DMS and S02 undergo further oxidation to reach the final oxidation state of sulfur +6. Of special significance will be the +6 sulfur forms H2S04(g), S042~(non-sea-salt sulfate, NSS), and methane sulfonate (MS).

DMS Oxidation

Some of the earliest studies that attempted to define the oxidation products of DMS were those carried out by Niki et al. (1983), Hatakeyama et al. (1985), and Grosjean (1984) in the early 1980s. These studies can best be labeled as "chamber studies" in that they typically involved filling a large multi-liter vessel with air mixtures containing DMS, NO, and other trace species (i.e., HONO), and then activating the system with solar or artificial radiation to produce OH radicals. There have been many different versions of the chamber-type study [see reviews by Yin et al. (1990), Turnipseed and Ravishankara (1993), and Berresheim et al. (1995)], some starting with sulfur in the form of DMS while others have used intermediate oxidation products like DMSO. The early studies as well as those that have followed have been quite revealing in demonstrating that among the important oxidation products generated from DMS are S02 and MSA, with lesser amounts of DMSO and DMS02. In fact, all of these products have now been directly measured in the atmosphere using modern instrumental techniques.

Although qualitatively revealing, chamber studies have also had their limitations. This reflects the fact that the gas mixtures employed have been significantly different chemically than that which is typically found in a marine boundary layer (MBL) environment. In this case two of the more important species involved have been DMS itself and the radical scavenging species NO. Both have typically been present in chambers studies at concentration levels several orders of magnitude higher than those found in the marine boundary layer. In addition, chamber studies have inherently been flawed due to their inescapably large surface-to-volume ratios (STVR). These have also been several orders of magnitude higher than those found in a marine environment (e.g., as aerosol surface area) and, thus, have led to greatly enhanced heterogeneous wall reactions. Since both concentration levels of DMS and NO as well as STVR factors impact on the DMS oxidation mechanism, not surprisingly, product distributions from individual chamber studies have been found to deviate significantly from study to study. They have thus left unanswered many of the quantitative details of the DMS product distribution within the MBL.

Among the more informative studies that have helped unravel aspects of the DMS oxidation mechanism have been those involving detailed laboratory kinetic investigations. These studies have focused on examining individual elementary reactions, the sum total of which, if available, would serve to define the overall DMS oxidation mechanism. One of the more pivotal of these was a study reported by Hynes et al. (1986). This study revealed that the reaction of OH with DMS proceeds not by a single reaction pathway but rather by two independent channels labeled kinetically as abstraction and addition, i.e., the reactions

As shown in Figure 3, the abstraction channel is nearly temperature independent; whereas, the addition reaction reveals a very significant negative temperature dependence. The crossover point for near equal contributions from both channels is seen as near 285 K. The early thinking on this mechanistic finding was that the OH abstraction channel was the channel that predominantly led to the formation of S02, while products such as DMSO, DMS02, and MSA were believed to be associated with the OH addition channel.

Evidence supporting the above position has included extensive field observations in which the stable end products MS and non-sea-salt sulfate (NSS) were measured and the value of their ratio then examined as a function of the ambient temperature. It was argued that S02 could be expected to undergo reasonably fast oxidation in the MBL via heterogeneous processes (see discussion under S02 Oxidation), thus forming NSS. On the other hand, MSA(g), formed from the addition channel, would be quickly scavenged by sea-salt aerosol to form MS. If indeed the above processes collectively define the mechanism by which both products are formed, as noted above the measured ratio of MS to NSS could be expected to provide a good chemical reflection of the average temperature at which the DMS oxidation occurred. In fact, extensive field measurements that have evaluated this ratio over a range of latitudes and altitudes have shown that the lowest values (e.g., 0.07) occur at tropical latitudes and that some of the highest values (e.g., >0.34) tend to be found at much higher latitudes (e.g., Berresheim, 1987; Savoie and Prospero, 1989; Bates et al, 1992a). However, a more limited but still quite significant number of ch3sch3 + OH CH3S(OH)CH3 =>• CH2SCH3 + H20 ch3sch3 + OH CH3S(OH)CH3 CH3S(OH)CH3 + 02 =>• products

(Abstr.) OH + CH3SCH3—►CH2SCH3 + H20 (Add.) OH + CH3SCH3 —►CH3SCH3 ^»-products

250 260 270 280 290 300 310

Temperature (K)

250 260 270 280 290 300 310

Temperature (K)

Figure 3 Temperature dependence of the rate coefficients for the OH/DMS addition and abstraction reaction channels as well as the total k value (modified from Berresheim et al., 1995).

observations have also been reported that do not follow this simple trend (e.g., Berresheim et al., 1995; Davis et al., 1998). Since these observations appear to be equally valid, they most likely point to a DMS oxidation mechanism that is more complex than originally thought. (Note, the potential importance of the MS/NSS ratio as defined by DMS oxidation rests in the fact that this ratio, if well understood, could be used to apportion the DMS contribution to total NSS. Perhaps, more importantly, it could be used as an indicator of the temperature environment under which the DMS oxidation process took place.)

One rendition of the DMS oxidation mechanism that reflects the thinking that the overall process is actually quite complex is that shown in Figure 4. This mechanism (shown here in abbreviated form) has folded in the most recent results from both field studies as well as laboratory kinetic investigations. Quite significant is the clear indication that not only is the OH abstraction channel a source of S02 but that the addition branch, in several different steps, also can form S02 as a product. Equally important, the stable product MS is shown as a product of both addition and abstraction channels. Its production efficiency is shown as being even further convoluted as

Figure 4 Abbreviated DMS oxidation scheme (modified from Davis et al., 1999).

a result of it being formed through competing gas and heterogeneous processes involving the intermediates DMSO and MSIA. Although the mechanism shown is still speculative (e.g., many of the elementary reactions have not yet been fully characterized), recent sulfur field studies, covering a wide range of latitudes, have provided evidence that strongly supports key aspects of this mechanism.

Unlike some of the earliest sulfur field studies, more recent investigations have reported a significant coupling between DMS and S02. Given that both DMS and S02 typically have MBL lifetimes of 0.5 to 2 days, if DMS is a significant source of S02, one would expect that these two species should be anticorrelated when measured at a rate significantly shorter than their respective lifetimes. Alternatively, if measured at a time resolution significantly longer than their respective lifetimes, one would expect a positive correlation. Thus, depending on the sampling rate, the appearance of either a correlation or anticorrelation in field data would signal there being a significant oxidative pathway from DMS to S02. As cited above, in the earliest studies for which simultaneous measurements of DMS and S02 were recorded, no relation between these two sulfur species was found. However, with improvements in instrumentation and the more judicious selection of field sites (e.g., free of anthropogenic pollution sources), a quite different picture is now emerging. Among the more significant studies has been that reported by Bandy et al. (1996), which took place in 1994 at Christmas Island. Located at 2°N in the middle of the Pacific Ocean, Christmas Island defines an ideal setting for studying DMS oxidation chemistry. Situated near the middle of the equatorial upwelling, it experiences near year-round elevated levels of DMS with no evidence of significant other sources of S02. In addition, being located well within the strong trade wind regime, it typically experiences stable meteorological conditions for several days at a time. Finally, due to the high solar flux and water vapor levels present, it also defines an environment where very high levels of the critical boundary layer oxidizing agent OH can be found. Reflecting these optimum conditions, Bandy and co-workers (1966) reported the first convincing field data showing a strong diel relation between DMS and S02. These investigator's high temporal resolution data were recorded over a 9-day time period and revealed a clear and convincing anticorrelation between these two sulfur species. The estimated DMS to S02 conversion efficiency was reported as 62 ± 6%.

In an airborne follow-up study at Christmas Island in 1996 [part of the National Aeronautics and Space Administration's (NASA's) GTE PEM-Tropics A program, Hoell et al. (1999)], the sulfur database reported was even more revealing. During this investigation direct observations were recorded of both DMS and S02 as well as the oxidizing agent OH. Equally significant was the availability in this new study of meteorological and chemical data as a function of altitude. As shown in Figure 5a, the profiles for DMS, S02, and OH make for a very convincing case that DMS is a major source of S02 and that DMS oxidation predominantly occurs via OH radicals. An analysis of these new data by Davis et al. (1999) resulted in an overall DMS to S02 conversion efficiency of 72 ± 22%, well within the range reported by Bandy et al. (1996). Given that the abstraction channel at the temperatures of Christmas Island (298 K) represented 70% of the total OH/DMS reaction rate, together with the conservative estimate that at least 8% of the product yield from DMS forms species

other than S02, one can estimate that the likely range for S02 formation from the abstraction channel is between 0.6 and 0.9. This suggests that the contribution of S02 from the addition channel is probably not much lower than 0.4; however, the issue of S02 contributions from the addition channel is more fully and better explored in the text below based on field studies conducted at much lower ambient temperatures.

Shifting to the recent National Science Foundation/National Oceanic and Atmospheric Administration (NSF/NOAA) program ACE-l [Aerosol Characterization Experiment, Bates et al. (1998)], the opportunity again presented itself to examine DMS oxidation chemistry under remote conditions, but with the field site being defined as the Southern Ocean, just to the south of Tasmania. Thus, it provided an environment having much lower ambient temperatures. Recall that because of the strong negative temperature dependence of the OH/DMS addition channel, the addition channel becomes more prominent under these conditions. In fact, based on the average temperature recorded on the ACE-l aircraft sampling platform (i.e., 280 K), both channels are estimated to be of near equal importance. Davis (unpublished results) in his analysis of the resulting DMS and S02 data has estimated that the overall DMS-to-S02 conversion efficiency to be still quite high, e.g., 0.7 to 0.9. Thus, to be consistent with the measured S02 levels and the previously cited average efficiency for S02 production from the abstraction channel of 0.75, he assigned a range of 0.7 to 0.9 to the addition channel. Although seemly quite high, the latter range is seen as being consistent with the previously discussed S02 tropical analysis.

In yet another study at still higher latitudes (i.e., 66°S), the land-based NSF SCATE Antarctic Program (Berresheim and Eisele, 1998), the average temperature recorded was only 273 K. In this case the addition channel is estimated to be 65% of the total OH/DMS kinetic rate. Although S02 measurements were not made during this campaign, direct observations of OH, DMS, and H2S04 suggest that the overall S02 conversion efficiency required to support the observed H2S04 levels would need to be well above 60%. Thus, these data again point toward an S02 efficiency for the addition channel of 50% or higher (Davis et al., 1998).

Quite interestingly, in all of the analyses cited above, involving high conversion efficiencies of DMS to S02, the measurements being used have been those reported by Bandy and co-workers (1996). The latter group pioneered the use of the isotopic-dilution gas-chromatographic mass-spectrometric technique (IDGCMS) for both field measurements of DMS and S02 (Bandy et al., 1993). On the other hand, in two recent ship-based studies, one in the tropical South Pacific [MAGE (Marine Aerosol and Gases Experiment), Yvon et al. (1996)], the other at high latitudes as part of the ACE-l program (De Bruyn et al., 1998), a quite different technique was employed in the measurement of S02. In both of these cases S02 measurements were made using the aqueous-phase fluorescence method (Saltzman et al., 1993). In both studies, evidence was found of diel trends in DMS and S02, further confirming the important role of photochemical oxidation of DMS. The reported overall conversion efficiencies for DMS to S02, however, were estimated to be significantly different from those cited above. In each of the ship studies, the conversion effi ciencies ranged from 30 to 50%, a factor of 1.5 to 2 lower than those already cited. It seems unlikely at this juncture, even though the latter studies were conducted on different platforms, that the results would differ by this amount. Both the average temperature and levels of other critical chemical species appear to have been similar for the tropical studies and the high latitude investigations. Whether the reported difference is due to S02 measurement difficulties or to yet unknown factors cannot be determined at this time.

Evidence that oxidizing agents (i.e., N03 and CI) other than OH are important in converting DMS to S02 has been primarily based on results from laboratory kinetic studies [see, e.g., Berresheim et al. (1995) and references therein]. In combination with model estimated levels of N03 and CI, the tentative conclusion has been reached that only the N03 mechanism is significant and that for most marine areas even this oxidant is unimportant. The exception would be for highly populated coastal influenced regions where major sources of N0X could be expected. As related to the importance of CI atom oxidation DMS, even though some recent sulfur field data (e.g., MAGE study) suggest that the impact from CI might be significant, other results reveal a different picture. For example, the results from the previously discussed field studies at Christmas Island as well as those in the Southern Ocean suggest that CI atom DMS oxidation is less than 15%. An independent study by Singh et al. (1996), in which C2C14 budget arguments were used to evaluate the significance of boundary layer CI atom oxidation, also resulted in a similar conclusion, namely, that the latter chemistry is of negligible importance in remote marine regions.

In addition to the pivotal question related to the oxidative conversion efficiency of DMS to S02, significant other DMS oxidation issues also continue to be the subject of continuing research. These include identifying the DMS oxidation intermediate(s) responsible for gas-phase H2S04(g) formation, the elucidation of the pathways by which DMSO, DMS02, MSA(g), and MS are formed, and the identification of the factors controlling the MS/NSS ratio. In the text that follows these DMS issues are further explored in the context of both recent laboratory kinetic investigations as well as the results from recent marine sulfur field studies.

As related to H2S04(g) formation, a review of the literature points to two major reaction sequences as being of potential importance. These include (3a) to (3d) and (4a) and (3d), e.g.,

Two of the most revealing recent field studies that have examined this issue are the previously discussed PEM-Tropics A Christmas Island airborne study and the

DMS + OH multisteps => S02 + other products

DMS + OH multisteps => S03 + other products S03 + (H20)x =► H2S04(g) + H20

ground-based SCATE study in Antarctica. Recall, that during the PEM-Tropics A field investigation direct observations of DMS, OH, S02, H2S04, and total aerosol surface area were simultaneously recorded. As shown in Figure 5b, using the known rate coefficients for processes (3a) to (3d), together with recently measured aerosol sticking coefficients for H2S04, Davis et al. (1999) concluded that the observed profile for H2S04(g) could be convincingly explained in terms of the observed diel profiles for S02 and OH. Since, as previously discussed, the observed S02 profile from this field study was also explicable in terms of OH/DMS oxidation, these new results are consistent with the idea that S02 is the critical DMS intermediate leading to gas-phase H2S04 formation. Taking a different approach, Jefferson et al. (1998), not having direct observations of S02, used the observations of DMS, H2S04, OH, and total aerosol surface area from the SCATE program to evaluate the two quantities /c0H [OH][DMS] and £surf[H2S04]. It was argued that if the direct formation of S03 from DMS were important, given the reasonably short lifetimes for both H2S04 and S03 in the Antarctic environment (i.e., <l h), one would expect a significant correlation between these two quantities. In fact, the R2 value was less than 0.2, indicating no relationship. Although still lacking the finality that comes from having a comprehensive set of elementary rate constants for each step in a mechanism, the collective results cited above strongly suggest that S02, not S03, is the dominant intermediate from the oxidation of DMS that leads to the gasphase formation of H2S04.

Field observations bearing on the mechanistic details surrounding the formation of DMSO and DMS02 from DMS oxidation have been limited in number and conflicting in their results. They include results from a sulfur field study near the Washington coast, the previously discussed Antarctic SCATE program, and finally the 1994 Christmas Island study. During the SCATE program, there were ~6 days of near continuous recording of DMSO and DMS02 (Berresheim et al, 1998). However, in the analysis of these data Davis et al. (1998) could find only 1 day out of the 6 sampled in which it appeared that both DMSO and DMS02 levels were controlled by local photochemical production. For all other days DMSO and DMS02 were shown to be controlled by transport processes, wherein large quantities of ocean-released DMS were initially carried aloft into the lower free troposphere, oxidized, and then returned in the form of intermediate as well as +6 oxidation state sulfur. But, on January 19, 1994, it appears that there was a significant break in this cycle in that background levels of both DMSO and DMS02 were found to be a factor of 10 lower than during the other 5 sampling days. Only on this day, was there any evidence of a diurnal profile for DMSO that tracked the measured ultraviolet (UV) solar irradiance. From their analysis of these data, Davis and co-workers (1998) estimated that the DMSO formation efficiency from the OH/DMS addition channel could range from 0.5 to 1.0, a value well within the limiting value assigned to this branching ratio based on two independent laboratory kinetic investigations. The DMS02 data, although considerably more noisy, were found to be most consistent with a branching efficiency for DMSO/OH to DMS02 of ~0.3.

In the field study near the Washington coast (Berresheim et al, 1993), 2 to 3 days of DMS and DMSO data were collected, but with very little ancillary data to facil itate defining the photochemical environment for this investigation. On April 14, however, sunny conditions prevailed nearly all day, and DMS and DMSO were sampled while air was advected in from the Pacific Ocean. For this specific case Berresheim et al. (1995) were able to estimate a branching ratio for the DMS/DMSO addition channel of ~0.5 but with a large uncertainty. This result may again be viewed as in good agreement with the above-cited results by Davis et al. (1998), but both have large uncertainties associated with them. In the 1994 Christmas Island study the measured levels of DMSO were found to be incredibly large relative to the median values of DMS observed, i.e., median DMSO 25 pptv, median DMS 200 pptv. Chen et al. (2000) in their analysis of this data quickly concluded that the two observations were totally irreconcilable in terms of any known DMS oxidation mechanism. This led them to put forward two possible hypotheses: (1) There were possibly unknown difficulties in the measurements of DMSO or (2) yet unknown sources of DMSO may exist. Still more recently NASA's PEM Tropics B Field study, direct airborne measurements of DMS, DMSO, and OH from sunrise to 1pm local time indicated that the highest levels of DMSO were at or near sunrise. Concentrations were found to decrease throughout the remainder of the measurement period (Nowak et al, 2001). The authors have pointed out that the temporal behaviour of DMSO is totally contrary to that expected if DMSO was formed only from the reaction of DMS with OH. Thus, these new data also suggest an additional source of DMSO, one that operates at night as well as possibly during daylight hours. Quite clearly if the latter hypothesis is correct, Table 1, as related to global sources of sulfur, would require further modification.

The issue of how efficiently DMSO might be formed through the OH/DMS addition channel raises the equally important question: Does the further oxidation of this species provide an effective pathway for formation of MSA(g)? This +6 oxidation state sulfur compound is shown in Figure 4 being formed from the oxidation of methane sulfinic acid (MSIA), another intermediate from the OH/DMS addition channel. Recent laboratory kinetic data (Hynes et al, 1996; Urbanski et al, 1998) would seem to support this notion in that they found the OH/DMSO reaction to be very fast and that the reaction appears to lead to near unity yields of the CH3 radical. The product CH3 radical is one that would be expected if the initial adduct formed from the reaction of OH with DMSO broke apart to form MSIA. Even so, in both of the airborne field studies previously cited, as well as during project SCATE, direct observations of gas-phase MSA have revealed a very low production efficiency for this species, e.g., typically <1%. The latter result is significant in that it translates to our explaining no more than 2 to 5% of the observed MS aerosol loading from the condensation of MSA(g). This means that both under tropical as well as the low-temperature conditions of the Antarctic, gas-phase production of MSA is not the major source of MS (the latter being a frequently cited measurement in much of the older literature involving DMS field studies).

The above MSA(g) results again focus our attention on the question touched on earlier in the text: How well do we really understand the factors controlling the value of the much cited MS/NSS ratio? Recall earlier (bottom of page 132) in the text we discussed the fact that some observations have shown a trend of increasing values in this ratio with decreasing temperatures (i.e., increasing latitude) but that notable exceptions had also been seen in this trend. One recent explanation for both the observed low yield of MSA(g) and yet significant yields of MS has been that proposed by Jefferson et al. (1998). To explain the MSA(g)/MS SCATE results, these investigators proposed that the rather high median levels of 1 to 2pptv of DMSO observed on the Palmer Peninsula (see earlier discussion in text related to DMSO formation at Palmer) could only be accounted for if heterogeneous DMSO reactions were occurring on sea salt aerosols. This hypothesis is supported by the fact that in several independent aqueous phase kinetic studies, DMSO has been shown to react in the aqueous phase (in the presence of oxidizing agents such as OH radicals) to form MSIA. This species was observed to subsequently undergo further oxidation to yield MS. Davis et al. (1999) came to a similar conclusion when analyzing the PEM-Tropics A data; however, these investigators noted that both gasphase DMSO and MSIA(g) would be equally good candidates as a heterogeneous source of MS. In yet another laboratory kinetic study, Lee and Zhou (1994) examined the aqueous-phase reaction of DMS with 03 as a possible source of aerosolphase oxidized sulfur. What they found was that because of the very low Henry's law constants for both DMS and 03, the probability of this aqueous process is rather unfavorable. However, under the most favorable conditions involving heavy clouds and relatively high 03 levels, it could prove to be a significant source of oxidized sulfur. Collectively, the above findings would seem to suggest that the formation of MS and possibly other sulfur species must be viewed in the context of both gasphase and heterogeneous chemistry in the atmosphere.

The implications of the above findings are quite significant in that they clearly point to the possibility that the MS/NSS ratio depends not only on the temperature of the environment where DMS oxidation occurs, but is equally, if not more, influenced by the nature of the aerosol environment, e.g., sea salt loading, cloud density, and liquid water content. For very low aerosol loadings a substantial fraction of the DMSO and MSIA(g) from the OH/DMS addition channel would most likely react with OH in the gas phase to produce S02. Most of this S02 would subsequently be converted into NSS. On the other hand, for very high aerosol loadings nearly all DMSO and MSIA would likely be scavenged, producing MS as a final product. Thus, for a given sampling location where the aerosol loading might vary from day to day, one could expect to find a range of MS/NSS values. In this context, one of the most stable environments in which MS/NSS values would be rather constant would be that defined by the tropical marine BL. Here the abstraction branch would strongly dominate DMS oxidation and the sea-salt aerosol loading would remain both relatively high and reasonably constant. Indeed, some of the most consistent values for the MS/NSS ratio have been those measured in the tropics (e.g., Saltzman et al., 1986; Savoie and Prospero, 1989; Berresheim et al., 1995). However, in spite of what appears to be a reasonably well documented environment, one should not lose sight of our earlier discussion that hinted at the strong possibility that there may be a significant and yet unidentified source of DMSO. If so, considerable rethinking of tropical marine sulfur chemistry may be necessary.

Thus far our DMS discussions have been primarily focused on the marine boundary layer (MBL). It may be asked, therefore, how dramatically does this picture change if the oxidation of DMS were to occur in the free troposphere (e.g., above 2 km)? In fact, as hinted at in our earlier discussions of the temperature dependence of the OH/DMS reaction, quite significant changes can occur. In the free troposphere three major physical changes occur in the environment: the temperature drops (e.g., 6.5°C/km), the pressure drops (i.e., exponentially), and the average aerosol surface area drops by at least one order of magnitude. It is the first and third of these shifts that potentially could have the most significant impact on DMS oxidation chemistry. Recall, that the OH/DMS addition channel has the strongest dependence on temperature (increasing with decreasing temperature), and therefore this channel becomes the dominant one with increasing altitude. On the other hand, laboratory studies suggest that this channel probably also has the greatest diversity in oxidation products. Equally important, the oxidation product distribution from this channel appears to have the greatest dependence on aerosol surface area, i.e., heterogeneous reactions. Thus, speculating on the net effect of these factors might point toward enhanced levels of both DMS02 and MSA(g). It could also mean that a much larger fraction of the DMSO and MSIA would be oxidized via OH, leading to higher yields of S02 or new products like sulfiirous acid (H2S03) from this channel. This sequence of reactions, in turn, could lead to the higher yields of H2S04(g) which under the cold temperatures of the upper troposphere could form the basis for new aerosol particle formation as suggested by Clarke (1993). Still another interesting result from this high-altitude DMS chemistry would be its impact on the MS/NSS ratio. For example, with an enhancement in the yield of S02 from the addition channel, the value of this ratio might remain reasonably low even though the temperature at which the oxidation occurred was quite low. Suffice it to say, both new laboratory kinetic studies as well as field observations will be required to actually quantify this chemistry.

S02 Oxidation

As both a primary source species (e.g., combustion and volcanoes) and as one of the major products from DMS oxidation, the atmospheric fate of S02 represents a major component of the atmospheric cycling of sulfur. It is estimated that between 40 and 60% of this S02 is directly deposited to either land or ocean surface areas (Berresheim et al., 1995). The remainder is believed converted into sulfur +6, although the detailed mechanisms by which this final oxidation state is reached continues to be the focus of ongoing research. What is now reasonably clear is that there are at least two general pathways by which this is achieved: one involving gas-phase chemistry, the other involving heterogeneous reactions. The gas-phase process is now relatively well understood, involving the ubiquitous oxidizing agent OH, e.g., reactions (3b) to (3d). The first two steps were reasonably well established by the mid 1980s (Finlay-son-Pitts and Pitts, 1986); however, only recently have the details of step (3d) been established (Lovejoy et al., 1996). This process has now been shown to involve a quadratic dependence on H20. But, considering the amount of H20 in the atmo sphere, this step is rarely if ever the rate-limiting step. In virtually all cases step (3b) is rate limiting.

As related to the gas-phase oxidation of S02, the importance of this process must be viewed both from the perspective of converting bulk atmospheric S02 to sulfate and from the point of view of its role as a major source of gas-phase H2S04. Current evidence suggests that the gas-phase oxidation of S02 is probably no greater than 20% of the total, and in the final analysis it could be no more than 5 to 10% (Lelieveld and Heintzenberg, 1992). On the other hand, the gas-phase production of H2S04 now appears to represent a critical step in the formation of new particles (via heterogeneous nucleation) that ultimately leads to cloud formation (e.g., Kreidenweis and Seinfeld, 1988). Thus, in the absence of this source, it would be difficult to explain how the atmosphere resupplies itself with CCN. CCN are routinely removed by both wet and dry deposition. This suggests then that the gas-phase oxidation of S02 is of primary importance in the atmosphere defining the strong link between sulfur emissions and climate effects.

Of the 80 to 90% of the S02 that is oxidized by non-gas-phase pathways, both heterogeneous reactions involving cloud droplets as well as sea salt aerosols are considered important. [For highly industrialized regions the influence of soot particles, trace metals such as Fe(+3), Mn(+2), and Cu(+2), and organic carbon reactions must also be included.] In remote marine areas, the heterogeneous oxidation process typically involves several steps. The first of these involves the critical equilibria shown in (5a) and (5b):

The presence of these equilibrium reactions means that the dominant form of sulfur +4, in the bulk aqueous phase, depends very much on the acidity of the aerosol species. For the most typical range of acidity in the troposphere, the dominant form of sulfur is the bisulfite ion (HS03 ). However, because of shifts in the levels of the individual forms of sulfur with changing pH, as well as the dependence of the reaction coefficients on pH, the most important aqueous-phase pathway for oxidation of sulfur +4 can be a strong function of pH, and therefore on the total amount of sulfur converted (e.g., Martin, 1984). This point is illustrated in Figure 6. Here it can be seen that for pH values above 5, the oxidation by 03 represents the dominant pathway; whereas for pH values less than 5, oxidation via H202 becomes the major source of sulfur +6. Other investigators (e.g., Chameides and Stelson, 1992; Siever-ing et al, 1992) have proposed that the sensitivity of the aqueous-phase oxidation of sulfur +4 to percent sulfur converted might be much smaller than originally thought. It has been suggested that this would be particularly true when the aerosol species is sea salt. The above group of investigators have argued that sea salt contains a natural buffering capacity involving the bicarbonate/carbonate system. Thus, seawater aerosol might be able to sustain a high rate of conversion of +4 sulfur to +6 through the 03 oxidative pathway for extended periods of time.

indicator of humankind's influence on the natural sulfur cycle. At present the database for both species is still quite limited with vertically resolved data now being available for no more than 20% of the global atmosphere. Most of the latter data is also limited to no more than one or two seasons of the year with the geographical coverage being confined largely to the North and South Pacific Oceans. Representative of these data are the latitudinal plots of S02 shown in Figures la and lb. For clarification purposes, these data have been binned for the altitude ranges of 0 to l and 2 to 12 km. Several points made earlier in the text, concerning anthropogenic effects, are clearly revealed in these plots. For example, the Northern Hemisphere is seen as having an average mixing ratio for S02 that is nearly five times higher than that for the Southern Hemisphere. Equally significant is what appears to be direct evidence for the focused release of S02 in the highly industrialized midlatitude region of 30 to 55°N.

Thus, given the limitations of current field data, one must turn to models to explore in greater depth the global atmospheric picture of sulfur. In this case the goal is that of gaining further insight into the distributions and variations in sulfur compounds and the processes that regulate their concentration levels. Several global-scale chemistry transport models have been developed in the past 7 years for just this purpose. These models endeavor to place available input sulfur data, as related to sources, sinks, and concentration levels, into a comprehensive global sulfur cycle (e.g., Langner and Rodhe, 1991; Pham et al., 1995; Feichter et al., 1996; Chin et al., 1996, 2000; Chuang et al., 1997; Koch et al., 1999; Barth et al., 2000). All include tropospheric sulfate and its major precursors (i.e., DMS and S02) and contain modules designed to handle anthropogenic and natural emissions, chemical transformations, advection/convection, and dry and wet deposition. Illustrative of the output from these models, we show in Figure 8a and 8b the annually averaged global surface-air distributions for S02 and sulfate based on results from Chin et al.'s (2000) model. This model includes sulfur from fossil fuel and biofuel combustion, shipping and aircraft emissions, biomass burning, volcanoes, and biogenic sources. Here it can be seen that the maximum S02 concentrations are clearly located at latitudes between 30 and 75°N, corresponding to the major industrial source regions of eastern United States, Europe, and eastern Asia. The levels of S02 are seen ranging from 1 to over 10 ppb. Interestingly, significant surface S02 concentrations are also shown to be present over southern Africa and Chile, largely reflecting ore smelting operations. The distribution of surface-air sulfate over the continents is found to be very similar to that for S02, although the gradients are clearly smaller. These observations reflect the fact that sulfate is the primary product of S02 oxidation and that transport as well as dry and wet deposition represent major losses for S02 (see discussion later in chapter). The model results also reveal that sulfur concentrations in the Arctic and near coastal regions in the Northern Hemisphere tend to be heavily influenced by anthropogenic releases. Returning to the field data shown in Figure la, the model values found at these near continental locations appear to be substantial larger than those reported in the latitudinal plots taken from the data of Thornton et al. (1999). Recall, however, that most of these data were recorded over remote regions of the Pacific. The sulfate distribution (i.e.,

A GLOBAL DISTRIBUTIONS OF SOE AND SULFATE

A GLOBAL DISTRIBUTIONS OF SOE AND SULFATE

180 120W 60W 0 60E 120E 180 Contours: 30 200 1000 5000

180 120W 60W 0 60E 120E 180 Contours: 30 100 500 1000

Figure H Global sulfur mixing ratios in the lowest 500 m as derived from the global chemistry transport model of Chin et a!. (1999): («1 S02 and (b) S04-~.

Figure 8/)) which also is shown failing off like S02 as one moves from continental regions to the open ocean, is similarly in good agreement with observational data when one considers the geographical location of the surface sampling sites [see, e.g., the data of Savoie et al. (1989) over the North Atlantic, that of Savote and Prospero (1989) over the North Pacific, and that in the Arctic reported by Barrie et al. (1989)].

As seen in the SCK data of Thornton et al. (1999) and in the model results shown m Figure 9a and 9b, the impact from anthropogenic emissions of sulfur is significantly attenuated at altitudes well above the boundary layer. This is due both to a large fraction of the combustion-based SO; and sulfate being deposited within the continental boundary layer and also to the more efficient dispersion of these species once at higher altitudes. Chin and Jacob (1996) have estimated that about 40% of the Northern Hemisphere industrial source of S02. is transported out as SCK or sulfate

90S 60S 30S 0 30N 60N 90N Contours: 20 50 200 500 1000

Sulfate (ppt)

Sulfate (ppt)

90S 60S 30S 0 30N 60N 90N Contours: 20 50 100 300 500

Figure 9 Global altitudinal and latitudinal distribution of the sulfur mixing ratio based on the global chemistry transport model of Chin et al. (1999): (a) S02 and (b) S042 .

90S 60S 30S 0 30N 60N 90N Contours: 20 50 100 300 500

Figure 9 Global altitudinal and latitudinal distribution of the sulfur mixing ratio based on the global chemistry transport model of Chin et al. (1999): (a) S02 and (b) S042 .

to the neighboring oceans and to the free troposphere, while the rest is removed by dry and wet depositions within the source region itself. These authors conclude that dry deposition takes up nearly one third of surface S02 emissions directly in the polluted region itself. Thus, although global anthropogenic emissions of S02 account for about 70 to 80% of the total emission of sulfate precursors, their contribution to the total sulfate burden is likely to be substantially less.

Yet another interesting result from the model studies is their assessment of the importance of volcanic emissions to the global sulfate burden in the troposphere. Chin and Jacob (1996) have found that this source is a significant contributor to the

4 GLOBAL DISTRIBUTIONS OF S02 AND SULFATE 149

TABLE 3 Ranges of Sources, Sinks, Total Mass and Lifetimes of S02 and Sulfate from Seven Global Sulfur Models

Ranges Median"

Guide to Alternative Fuels

Guide to Alternative Fuels

Your Alternative Fuel Solution for Saving Money, Reducing Oil Dependency, and Helping the Planet. Ethanol is an alternative to gasoline. The use of ethanol has been demonstrated to reduce greenhouse emissions slightly as compared to gasoline. Through this ebook, you are going to learn what you will need to know why choosing an alternative fuel may benefit you and your future.

Get My Free Ebook


Post a comment