SCOPE 21 - The Major Biogeochemical Cycles and Their Interactions, Chapter03, Atmospheric Interactions-Homogeneous Gas Reactions of C N, and S Containing Compounds

This article is a review of the most important photochemical processes that take place in the atmosphere and of the cycles of many C, N, and S containing atmospheric constituents. The emphasis is on the essential role of tropospheric ozone and how the distribution of this gas is influenced by photochemical interactions between carbon and nitrogen containing compounds, which are changing substantially due to man's activities. There is a large potential for tropospheric ozone formation, of which at most only 10% is realized due to generally low NO_x concentrations. The input of NO_x in the troposphere may be dominated by anthropogenic sources so that tropospheric ozone may have increased in the industrial era.

The budgets (sources and sinks) of many compounds are derived and it is shown that these are mostly dominated by tropospheric reactions. However, several gases are rather inert in the troposphere and their photochemical breakdown in the stratosphere leads to products that influence stratospheric ozone and sulphate aerosol. There often remain large uncertainties in the quantitative aspects of the atmospheric cycles of trace constituents.

3.1 INTRODUCTION

Of the elements carbon, nitrogen, sulphur, and phosphorus, gas phase reactions are by far the least important for phosphorus. There are some indications that phosphine (PH₃) is released from tropical forest soils (R. Herrera, personal communication), but the residence time of PH₃ is probably so short that little atmospheric transfer can occur. Phosphorus enters the atmosphere mainly on soil dust (Duce, Chapter 17, this volume). On the other hand, several C, N, and S compounds have considerable atmospheric concentrations, mainly in combination with the elements O and H. These compounds play an essential role in the gas phase photochemistry and the cycling of many volatile elements in the atmosphere, not only those containing C, N, and S, but also, for example, the photochemically important halogen compounds.

The composition of the natural atmosphere is to a considerable extent determined by biological processes. With the exception of CO₂, the typical natural cycle of an element consists of release at the earth's surface as a reduced, often hydrogen-bound gas, its subsequent oxidation in the atmosphere by gas phase photochemical reactions, and finally the removal from the atmosphere by precipitation scavenging, and deposition on the earth's surface. Here, such compounds may serve as electron acceptors in anoxic environments making possible the many microbiological reduction-oxidation processes in the soils and waters of the earth. As two examples we mention methane (CH₄), which is photochemically converted to carbon monoxide (CO), and hydrogen sulphide (H₂S), which is converted to sulphur dioxide (SO₂) in the atmosphere. Further oxidation leads next to carbon dioxide (CO₂) and sulphuric acid (H₂SO₄). Carbon dioxide is removed from the atmosphere mainly through photosynthesis in vegetation, while sulphuric acid is removed mainly through wet and dry deposition. I will note in the following that the natural sources of many biogenic gases are not well known. It should also be mentioned here that CO₂ does not play a significant direct role in the photochemistry of the earth's atmosphere, so that the photochemistry of CO₂ will not be discussed in the following. The reason for this is that CO₂ absorbs the photochemically active ultraviolet radiation in the same wavelength regions as molecular oxygen, which is a thousand times more abundant. The role of CO₂ in atmospheric chemistry is indirect by its influence on the temperature structure of the atmosphere and climate.

In contrast to natural, biological processes, man's activities cause the emissions of increasing amounts of oxidized gases into the atmosphere. The total global supply of anthropogenic fixed nitrogen, as NO, and of sulphur, as SO₂, are now an appreciable fraction of the total supply of these nutrient elements. In contrast to the biogenic emissions, industrial emissions occur mainly as point or regional sources, so that atmospheric concentrations of the various gaseous pollutants are non-uniformly distributed and their concentrations often strongly correlated. This leads to strong photochemical and chemical interactions, that may result in very different photochemical behavior to that which occurs in clean air masses.

In the following discussion I will review the homogeneous, atmospheric reactions and cycles of the most important gases that contain carbon, nitrogen, and sulphur in combination with hydrogen and oxygen. The gases that will be considered are important because of their role in biological, climatic and photochemical processes. I will emphasize the important role of tropospheric ozone in atmospheric photochemistry, how it influences the cycles of many trace elements and how man's activities affect tropospheric ozone. The experience of the last decade has already shown that there are many ways in which human activities may affect the balance between ozone formation and destruction in the stratosphere and thereby influence the intensity of ultraviolet radiation at the ground, and maybe climate, with possible repercussions for the biosphere.

In the following sections, brief reviews of the homogeneous photochemistry of the troposphere and stratosphere will be presented. Next, the budgets of the photochemically most important carbon, nitrogen and sulphur containing gases will be taken up for review. I will emphasize new developments since the SCOPE 7 study (Svensson and Söderlund, 1976), and interactions between cycles whenever they can be recognized.

3.2 TROPOSPHERIC PHOTOCHEMISTRY

Although only about 10% of all atmospheric ozone (O₃) is located in the troposphere, the lowest 10

17 km of the atmosphere, this small amount of ozone is nevertheless of fundamental importance for the composition of the earth's atmosphere. The reason for this is the production of the highly reactive hydroxyl (OH) radical by the two reactions:

(R1)	O₃ + hv		O(¹D) +O₂ (<310 nm;1 nm = l0^-9m)
(R2)	O(¹D) + H₂O		2OH

where O(¹D) denotes an electronically excited oxygen atom. It is the attack by OH that initiates the oxidation of many trace gases in the atmosphere, the most important examples of which are shown in Table 3.1 and Figure 3.1 (Levy,1971, 1974; McConnell et al., 1971). In the background troposphere, about 60% of the OH radicals react with CO, and about 40% with CH₄. Smaller fractions react with the other gases listed in Table 3.1.

The average concentration of hydroxyl in the atmosphere is now estimated to be about 6 x 10⁵ molecules cm^-3, with an uncertainty of about a factor of two (Crutzen et al., 1978; Derwent and Eggleton, 1981; Gidel et al., 1982). This estimated range is consistent with the global observations of methylchloroform, which is removed from the atmosphere by reactions with OH and which does not appear to have any atmospheric sources other than the well known industrial emissions. Most hydroxyl is located in equatorial regions, where the intensity of ultraviolet radiation is at a maximum and where absolute water vapour densities are highest (Crutzen et al., 1978; Derwent and Eggleton, 1981). The calculated OH distribution also explains most of the features of the C¹⁴O observations in the troposphere (Volt et al., 1979).

Although OH reacts overwhelmingly with CO and CH₄, these reactions do not necessarily lead to the removal of hydroxyl from the atmosphere, because hydroxyl will mainly act as a catalyst. For instance, in the presence of sufficient concentrations of another catalyst, nitric oxide, the oxidation of carbon monoxide will lead to the formation of tropospheric ozone, without loss of OH and NO, as follows:

(R3)	CO + OH	H + CO₂
(R4)	H + O₂ + M	HO₂ + M
(R5)	HO₂ + NO	OH + NO₂
(R6)	NO₂ + hv	NO + O (<400 nm)
(R7)	O+O₂+M	O₃+M

net:	CO + 2 O₂	CO₂ + O₃

(R3)	CO + OH	H + CO₂
(R4)	H + O₂ + M	HO₂ + M
(R8)	HO₂ + O₃	OH + 2O₂

net:	CO + O₃	CO₂ + O₂

again does not lead to loss of hydroxyl, and takes place whenever the ratio of the atmospheric concentrations of NO and O₃ is less than 2 x 10^-4. With ozone volume mixing ratios increasing from about 20 x 10^-9 (20 ppbv) at ground level to 100 ppbv at the tropopause, the break-even point between reaction chains (R3

R7) and (R3, R4, R8) is attained at nitric oxide volume mixing ratios of 4 x 10^-12(4 pptv) at ground level and 20 pptv at the tropopause. These are indeed very low concentrations, but they may nevertheless not be exceeded in extensive regions of the troposphere because of the very short residence times of the oxides of nitrogen, NO and NO₂, which are a result of the rapid formation of highly water-soluble nitric acid by the reactions:

(R9)	NO + O₃		NO₂ + O₂
(R10)	NO₂ + OH(+M)		HNO₃ (+M)

Through similar reactions ozone is also produced in the oxidation of CH₄ and other hydrocarbons if the mixing ratios of NO are sufficiently large. McFarland et al. (1979) have indeed indicated the possibility of very low background concentrations of NO by the measurements of volume mixing ratios of less than 10 ppt in the marine boundary layer of the tropical Pacific. Noxon (1981) measured profiles of NO_x and reported a mixing ratio for NO_x (NO + NO₂) of 30 pptv at 3 km altitude at Hawaii. There are, unfortunately, too few measurements to derive a typical distribution of NO in the troposphere, and to make a good estimate of the sources and sinks of tropospheric ozone. The only term that can be estimated with reasonable reliability is the global ozone loss by the reactions (R1) and (R2), amounting to about 8 x 10¹⁰ molecules cm^-2s^-1(Fishman et al., 1979a). This ozone loss may be compared with the estimated downward transport of about 6 x 10¹⁰ molecules cm^-2s^-1stratospheric ozone by meteorological processes and an ozone loss rate at the ground of the same magnitude (Fabian and Pruchniewicz, 1977). In order to balance the ozone budget of the troposphere, it is, therefore, quite feasible that photochemical interactions between carbon and nitrogen containing gases lead to tropospheric ozone production. This is particularly important as the atmospheric sources of NO, CO, CH₄ and other hydrocarbons are greatly influenced by anthropogenic activities, especially in the Northern Hemisphere.

Table 3.1a Budgets of carbon species; atmospheric lifetimes in hours, months, or years; diffusion distances in E

W, S

N and vertical directions (in km) over which concentrations are reduced to 30% by chemical reactions. Lifetimes and removal rates calculated with (OH) = 6 x 10⁵ molecules cm^-3. 1 ppmv = 10^-6; 1 ppbv = 10^-9; 1 pptv = 10^-12

Table 3.1c Budgets of sulphur species. For explanation see Table 3.1a. COS and CS₂ are not listed here, because too little is known about their sources and sinks. The industrial source of CS₂ is about 2 x 10¹¹g S per year

There are tropospheric ozone data available that indicate a possible increase of 20% in the middle troposphere (2

8 km) at north temperate latitudes over the time period 1967

1979 (Angell and Korshover, 1979), although such an increase seems not to be present in the Mauna Loa data of the past 6 years (Komhyr, personal communication).

To emphasize further the importance of tropospheric ozone I also point out its role in the earth's radiation budget, because of its strong absorption band located in the atmospheric `window' region near 9.6 µm. As a result of this, ozone absorbs, and radiates back to the earth's surface, terrestrial radiation that otherwise would escape to space. The particular efficiency of the tropospheric ozone fraction is caused by the pressure broadening of the absorption lines in the troposphere. A hypothetical doubling of tropospheric ozone has been calculated to lead to a surface temperature increase of about 0.7°C (Fishman et al., 1979b).

An approximate upper limit to the potential ozone production rates in the troposphere can be derived by assuming that there is enough NO present in the atmosphere that during the oxidation of a hydrocarbon molecule to CO, as with methane, there are 2

3 ozone molecules produced per carbon atom (Crutzen, 1973; Fishman et al., 1979a). The further oxidation of CO to CO₂ subsequently adds another ozone molecule to the atmosphere. Making use of the information on the sources of CO, CH₄ and non-methane hydrocarbons (Table 3.1), an average global, tropospheric ozone production potential of 2 x 10¹² molecules cm^-2s^-1can be calculated. The actual tropospheric ozone production is probably at most 10% of this potential, because ozone loss at the ground or by photochemical reactions is about 10¹¹ molecules cm^-2s^-1(Fishman et al., 1979a). This indicates that much oxidation of hydrocarbons, which occurs mainly in the tropics, must take place without the production of ozone. One may guess that this is so because the air above the tropical forests is at present practically devoid of nitric oxide. Future expansions of industrial and biomass burning activities in the tropics may, however, have large implications for the tropospheric ozone distribution, because of the enhanced supply of nitric oxide to the boundary layer of the atmosphere in which the oxidation of isoprene and terpenes to CO occurs. It should be mentioned here that most of the nitric oxide produced by lightning is deposited above the tropical boundary layer where the concentrations of these highly reactive hydrocarbons should be rather low. I will return to the possible implications of human activities on the air chemistry of the tropics when I discuss the role of nitric oxide in the production of carbon monoxide.

Figure 3.1 Compilation of the most important photochemical processes in the atmosphere, including estimates of flux rates expressed in moles per year between the earth's surface and the atmosphere and within the atmosphere. The processes are numbered and explained in the figure texts. The symbol 'E' means:

10^y where y is the number that follows E, e.g. 2E13 is 2

10¹³

As carbon monoxide is the main reactant of hydroxyl, an increase in the atmospheric carbon monoxide content could lead to lower tropospheric concentrations of OH and thereby cause an increase in the tropospheric abundances of those gases that are mainly removed by reactions with OH (Wofsy, 1976). This could have further chemical and climatological consequences, as for instance a larger transfer of gases to the stratosphere with possible effects on ozone. It is likely, however, that a rise in the tropospheric CO abundance by anthropogenic activities will be accompanied by increases in the tropospheric NO abundance. This would lead to higher ozone concentrations. The increases in NO and O₃ would tend to increase the OH concentrations through reactions (R1), (R2), and (R5), thereby counteracting the effect of CO.

Methane oxidation is not only a potentially significant source of CO, but the atmospheric concentrations of OH may largely be determined by the particular reaction paths that are followed during methane oxidation. Again, nitric oxide plays an important role in this. With enough NO present (>10 pptv) the reaction path, leading to formaldehyde (CH₂O), is as follows:

(R11)	CH₄ + OH	CH₃ + H₂O
(R12)	CH₃ + O₂ + M	CH₃O₂ + M
(R13)	CH₃O₂ + NO	CH₃O + NO₂
(R14)	CH₃O + O₂	CH₂O + HO₂
(R5)	HO₂ + NO	OH + NO₂
(R6)	NO₂ + hv	NO + O (<400 nm), 2x
(R7)	O + O₂ + M	O₃ + M, 2x

net:	CH₄ + 4 O₂	CH₂O + H₂O + 2 O₃

(R11)	CH₄ + OH	CH₃ + H₂O
(R12)	CH₃ + O₂ + M	CH₃O₂ + M
(R15)	CH₃O₂ + HO₂	CH₃O₂H + O₂
(R16)	CH₃O₂H + hv	CH₃O + OH
(R14)	CH₃O + O₂	CH₂O + HO₂

net:	CH₄ + O₂	CH₂O + H₂O

However, the photolysis of CH₃O₂H is slow, which results in a residence time of 1 week for this compound. Thus the methyl hydroperoxide may be rained out of the atmosphere or react with the earth's surface or aerosol particles. If this is the case, the oxidation of CH₄ will lead to a loss of two odd hydrogen radicals (OH and HO₂ and no formation of CO would occur in the atmosphere. It follows that the pathways of methane oxidation in the atmosphere are not yet satisfactorily resolved. More intricate questions will have to be addressed regarding the oxidation of higher hydrocarbons.

An interesting, and potentially important, interaction between carbon and nitrogen gases in the atmosphere occurs through the formation of certain organic nitrate molecules during the oxidation of hydrocarbons. The most important of these are probably the peroxy-acyl nitrates, especially peroxy-acetyl nitrate (PAN) with the chemical formula CH₃C(=O)O₂NO₂. PAN thermally decomposes in the atmosphere according to the reactions

(R17)	CH₃(C=O)O₂NO₂		CH₃(C=O)O₂ + NO₂
(R18)	CH₃(C=O)O₂ + NO		CH₃ + CO₂ + NO₂

(R19)	CH₃(C=O)O₂ + HO₂		CH₃ (C=O)O₂H + O₂
(R20)	CH₃(C=O)O₂H + hv		CH₃ + CO₂ + OH

depending on the concentration of nitric oxide. Based on information by Hendry and Kenly (1977) and Cox and Roffey (1977), the following values are calculated for the atmospheric residence time of PAN against photochemical destruction for several heights in the U.S. standard atmosphere: at z = 0 km, T = 288K, (PAN) = 3 days; z = 4 km, T= 262K, (PAN) = 1 month; z = 6 km, T = 249K, (PAN) = 1 yr; z = 8 km,

Because of the strong temperature dependence of reaction (R17), the lifetime of PAN against thermal decomposition increases dramatically with height. Likewise, higher PAN concentrations in the atmosphere may be favoured during the winter season (Hendry and Kenley, 1977). Although photolysis of PAN should be rather slow, the only ultraviolet spectrum of PAN published so far (Stephens, 1969) does not contain enough information to exclude significant photolysis in the middle and upper troposphere and stratosphere. Reaction of PAN with OH may likewise not entirely be neglected in these regions (Atkinson et al., 1979), but should not result in an atmospheric residence time shorter than several months.

Surprisingly, PAN is slowly lost from the atmosphere by wet and dry removal. A laboratory study by Garland and Penkett (1976) gave a deposition rate of PAN on water surfaces that was slower than that of ozone. Like ozone, PAN is not removed by rainfall, so that in the free troposphere above 3 km, the atmospheric lifetime of this gas is clearly much larger than that of NO_x. Long range transport of NO_x may, therefore, occur in the middle and upper troposphere with PAN as the vehicle. When PAN reaches the boundary layer it is thermally decomposed and NO_x is released to the atmosphere (Crutzen, 1979).

PAN is formed in polluted air by photochemical reactions, involving non-methane hydrocarbons and NO_x and has been observed in many urban environments (e.g. Stephens, 1969; Lonneman et al, 1976; Nieboer and van Ham, 1976; Penkett et al., 1977; Spicer, 1977; Tuazon et al., 1981). At high concentrations this gas is a major phytotoxicant and it affects health by causing eye irritation. In polluted air close to urban centres it is often present in concentrations similar to those of nitric acid and typically 10

20% those of NO_x (Spicer, 1977; Tuazon et al., 1981). PAN is formed from acetaldehyde (CH₃CHO), which is a photochemical intermediate in the photochemical decay of many non-methane hydrocarbons (Demerjian et al., 1974). The simplest chain of reactions leading to the formation of PAN occurs, following the oxidation of ethane, as follows:

(R21)		C₂H₆ + OH	C₂H₅ + H₂O
(R22)		C₂H₅+ O₂ + M	C₂H₅O₂+ M
(R23)		C₂H₅O₂ + NO	C₂H₅O + NO₂
(R24)		C₂H₅O + O₂	CH₃CHO + HO₂
(R25)		CH₃CHO + OH	CH₃CO + H₂O
(R26)		CH₃CO + O₂ + M	CH₃(C=O)O₂ + M
(R17)	CH₃(C=O)O₂ + NO₂ + M		CH₃(C=O)O₂NO₂ + M

C₂H₆ is present in the troposphere at sufficiently high volume mixing ratios (1

2 ppbv) (Singh et al., 1979; Rudolph et al., 1981) so that appreciable production of PAN from NO_xmay take place anywhere in the atmosphere (Singh and Hanst, 1981); this makes the possible role of PAN in global atmospheric photochemistry especially worthwhile to investigate. As observations at ground level in rural areas show (Penkett et al., 1977; Singh et al., 1979) PAN is not restricted to urban environments. In fact, PAN has been observed on several cruises in the subtropical and tropical Atlantic, whenever colder and more polluted air masses could reach the lower latitudes. Typical sea level concentrations during such episodes were of the order of 50 pptv (Guicherit, private communication). With more PAN expected in the background middle and upper troposphere (Crutzen,1979; Singh and Hanst, 1981), PAN could be an important reservoir species for the long range transport of NO_x.

The previously mentioned reactions are of critical importance in tropospheric photochemistry, because of their role in determining the concentrations of OH in background air. Most photochemistry of the gas phase in the troposphere is derived from this. For example, it is normally assumed that after the initial attack by hydroxyl, the reduced sulphur gases H₂S, (CH₃)₂S and CH₃SH are rapidly converted to SO₂. However, this has only been demonstrated in the laboratory for H₂S, while SO₂ formation did not take place for (CH₃)₂S (Cox and Sandalls, 1974), so that atmospheric reaction chains that do not produce SO₂, cannot be ruled out. In the presence of enough NO, we may even speculate about the possibility of formation of carbonyl sulphide (COS) via the reaction sequence:

(R27)	CH₃SCH₃ + OH	CH₃SCH₂ + H₂O
(R28)	CH₃SCH₂ + O₂ + M	CH₃SCH₂O₂ + M
(R29)	CH₃SCH₂O₂ + NO	CH₃SCH₂O + NO₂
(R30a)	CH₃SCH₂O	CH₃S + CH₂O (probable)
(R30b)	CH₃SCH₂O + O₂	CH₃SCHO + HO₂(not impossible)
(R31)	CH₃SCHO + OH	H₂O + CH₃ + COS

Although less likely, it may not be ruled out that the initial attack of OH on (CH₃)₂S occurs by addition of OH to S (Atkinson et al., 1978; Kurylo, 1978). Also, in the absence of enough NO, peroxides and many other compounds may be formed that can be removed heterogeneously from the atmosphere (Bentley et al., 1972; Panther and Penzhorn, 1980), so that only a fraction of the (CH₃)₂S may be converted to COS.

Because COS is a sulphur compound with a much longer atmospheric lifetime ( years) than (CH₃)₂S (days), it will be transported globally and reach the stratosphere (Hanst et al., 1975; Sandalls and Penkett, 1977; Inn et al., 1979; Torres et al., 1980). The photolysis of COS leads to the SO₂ and H₂SO₄ needed to explain the stratospheric sulphate layer during periods of low volvanic activity (Crutzen, 1976). Through its influence on the stratospheric aerosol layer, COS is of some significance for the earth's climate; the existence of any long term trends in the atmospheric abundance of COS should be ascertained (Turco et al., 1980; Hofmann and Rosen,1981).

Carbonyl sulphide, with an atmospheric volume mixing ratio of about 0.5 ppbv, is probably the most abundant sulphur species of the atmosphere. Recently, there have been some speculations that this gas (or CS₂) would also be the precursors of the SO₂ that has been measured in the middle and upper troposphere over the Pacific (Maroulis et al., 1980). However, the proposed reaction mechanisms via reaction with hydroxyl (Logan et al., 1979; Sze and Ko, 1980) have been shown to be incorrect (Atkinson et al., 1978; Ravishankara et al., 1980). The hypothesis that COS oxidation in the stratosphere is an important source of SO₂ and sulphate during periods with little volcanic activity has, however, probably been shown to be correct from measurements of COS and SO₂ in the stratosphere (Inn et al., 1979; Meixner, 1981).

A particularly interesting kinetic observation has recently been made by Jones et al. (1982), who showed that CS₂ reacts with OH only in the presence of O₂ via

These results have been confirmed by Niki and co-workers (Niki, personal communication).

Likewise, Jones et al. (1982) have reported that the photodissociation reactions occur with a quantum yield of 10^-3 in the 320 nm absorption band, making the two listed pathways to COS and SO₂ almost of equal importance. These reactions may be particularly important as sources of atmospheric COS. Recently, this gas has also been identified as a potential corrosion agent (Graedel et al., 1981).

The oxidation of SO₂ to H₂SO₄ in the presence of NO_x also involves initiation by reaction with hydroxyl, leading, for example, to the reaction sequence (Calvert et al., 1978):

(R33)	SO₂ + OH + O₂	HSO₅
(R34)	HSO₅ + NO	HSO₄ + NO₂
(R35)	HSO₄ + HO₂	H₂SO₄ + O₂

At low NO concentrations the oxidation steps are less certain (Davis and Klauber, 1975; Davis et al., 1979).

An interesting possible influence that NO_x might exert on the SO₂ oxidation cycle in industrial air masses has recently been pointed out (Rodhe et al., 1981). Because reaction (R10) between OH and NO₂ is about ten times faster than reaction (R33) it may act as a sink for OH when the mixing ratio of NO_x is of the order of a few ppb or larger. Under situations when photochemical oxidation is important, the oxidation of SO₂ may, therefore be delayed until the concentrations of NO₂ become smaller than about 1 ppbv. At higher NO_x concentrations, the production of H₂O₂ from will be suppressed because of reaction (R5). As H₂O₂ may be an important oxidant of SO₂ in water (Penkett et al., 1979), the oxidation of SO₂ in clouds likewise may be delayed.

Although mainly gas phase reactions leading to SO₂ oxidation have been discussed, it is clear that heterogeneous and aqueous phase reactions may be at least as important as homogeneous reactions for the oxidation of SO₂ to H₂SO₄, especially during winter time at middle and high altitude (see e.g. Shaw and Rodhe, 1981).

Under most atmospheric conditions the conversion of NO_x to HNO₃ via reaction (R10) occurs within a few days. The atmospheric residence time of highly soluble HNO₃ must be shorter than that of water vapour, which is about 9 days (Levine and Schwartz, 1981). As this is much less than the time scale of 2 months needed for HNO₃ to convert back to OH and NO₂, particularly through the photolysis reaction the formation of nitric acid by reaction (R10) provides an efficient sink for NO_x. In this way, an appreciable upward transport of NO and NO₂ to the stratosphere is made difficult.

It remains, however, of interest to explore the possibility of some leakage of NO_x and NO₂ (and other pollutant gases, such as SO₂) through the tropospheric water vapour and cloud filter to higher altitudes (middle and upper troposphere, lower stratosphere). One mechanism for such transport may be the rapid upward transport in frontal zones and especially thunderstorms so that reaction (R10) cannot be completed. Thunderstorms, which penetrate the tropopause, may directly transfer lightning-produced and boundary-layer NO_x into the lower stratosphere. As a consequence of this, outside polluted areas, such rapid transfer could bring about an increase of NO_x mixing ratios with altitude, providing a complementary interpretation to the NO_x profiles to the downward transport from the stratosphere proposed by Liu et al. (1980). This may be especially important for the tropical band and during summertime at mid-latitudes. Another mechanism may be transport at high latitudes during winter, when little or no OH is produced by reaction (R2) and when, furthermore, precipitation occurs in rather small amounts and more as snow, which may be less efficient than rain in scavenging trace gases. This possibility of global NO_x transport is speculative, though not entirely unreasonable, judging from observations of considerable pollution levels in the Arctic regions during winter (Rahn and Heidam, 1981; Shaw, 1981).

Observations of HNO₃ by Lazrus and co-workers (Lazrus and Gandrud, 1974; Huebert and Lazrus, 1978) and of NO₂ by Noxon (1978, 1979) in areas remote from pollution sources have shown volume mixing ratios of NO₂ and HNO₃ that are substantially larger in the stratosphere than in the troposphere. This indicates a net transfer of odd nitrogen from the stratosphere to the troposphere, but does not rule out the possibility of a significant transport of some tropospheric NO_x to the lower stratosphere at preferred locations. One-dimensional models are totally inadequate to address this issue and far too few data are available to make a reliable judgment on this interesting matter.

Finally, I shall also briefly discuss whether photochemical reactions may affect ammonia. For this gas, both uptake at the earth's surface and release by microbiological processes should be considered in order to establish the overall net source of NH₃ to the atmosphere. The only homogeneous gas phase reaction of ammonia is again reaction with OH:

This reaction is, however, rather slow (k₃₈ 1.5 x 10^-13cm³ molecule^-1s^-1), so that with average (OH) 6 x 10⁵ molecules cm^-3, NH₃ would remain in the atmosphere for a season, if reaction (R38) represents the only atmospheric loss process. In reality, ammonia will be removed from the atmosphere on the average during a period of about 9 days, which represents the average residence time of water vapour in the atmosphere. Thus, at most only a fraction of atmospheric NH₃, about 10%, could react with OH. In addition, the NH₂ radical that is formed can react with HO₂:

providing a pathway back to NH₃. Such reactions with radicals or minor atmospheric constituents are important because the reaction

is slow. It remains to be seen whether NH₃ oxidation would represent a source or a sink for atmospheric NO_x (McConnell, 1973) through reactions such as (Lesclaux and Demissy, 1977; Hack et al., 1978; Kurasawa and Lesclaux, 1980):

(R41)	NH₂ + NO_x		H₂O + N₂O_x-1(x = 1,2)
(R42)	NH₂ + O₃		NH₂O + O2

where further oxidation of NH₂O may lead to NO_x, probably via HNO formation. The break-even point between reactions (R41) and (R42) occurs at about 60 pptv of NO. Such mixing ratios of NO_x can indeed be reached in continental, agricultural areas where we expect substantial emissions of atmospheric NH₃ to take place. It may be that NH₃ oxidation implies a sink for some atmospheric NO_x and a source for N₂O. It is, however, not possible to make a numerical assessment because of insufficient information on the distributions of NO_x, OH and NH₃.

The great importance of NH₃ in atmospheric chemistry is due to its role in establishing the pH of rain and cloud water. A discussion of this may be found in the overview paper of Taylor et al. (Chapter 4, this volume).

3.3 STRATOSPHERIC PHOTOCHEMISTRY

The temperature structure and the dynamic processes in the stratosphere are to a major degree determined by the absorption of solar ultraviolet energy by ozone. The total amount of ozone in the stratosphere is nevertheless rather small; it corresponds to the number of air molecules that are contained in a 3 mm thick layer at standard temperature and pressure. As I have noted, the troposphere contains only 10% of the atmospheric ozone, so that most ozone is located in the stratosphere, which is between 10 and 50 km. Atmospheric ozone also plays an important ecological role. Most ultraviolet solar radiation between about 210 and 310 nm is filtered out by atmospheric ozone. However, penetration of UV radiation to ground level starts at about 300 nm and increases by orders of magnitude within the next 20 nm. It is this radiation that may be biologically harmful. The penetration: of this radiation to the ground is enhanced by reduction of the ozone column.

Since the beginning of the 1970s it has become increasingly clear that a number of human activities can lead to global changes in the amount of stratospheric ozone. Following suggestions by Johnston (1971) and Crutzen (1971), initial attention was directed to pollution of the stratosphere by direct injections of NO from high-flying aircraft. Earlier, Crutzen (1970) had proposed that NO_x (NO + NO₂) would catalyse the destruction of ozone and limit its stratospheric abundance by a simple set of photochemical reactions:

(R43)	O₃ + hv	O + O₂ (< 1140 nm)
(R44)	O + NO₂	NO + O₂
(R45)	NO + O₃	NO₂ + O₂

net:	2 O₃	3 O₂

(R46)	O₂ + hv	2 O (< 240 nm)
(R7)	O + O₂ + M	O₃ + M(2x)

net:	3 O₂	2 O₃

(R1)	O₃ + hv		O(¹D) + O₂ (>310 nm)
(R47)	O(¹D) + N₂O		2 NO

Stratospheric ozone may especially be affected by compounds that are relatively inert in the troposphere, because of a low solubility in water, a slow photolysis and a slow reaction with OH (e.g. nitrous oxide and several chlorocarbon gases, such as natural CH₃Cl and industrially produced CFCl₃, CF₂Cl₂, CCl₄, and CH₃CCl₃). Another way to influence stratospheric chemistry is by direct injection of material in the upper troposphere and stratosphere well above most atmospheric water and water vapour (e.g., volcanic eruptions, meteoritic impacts, nuclear weapons testing, and emissions from aircraft in the stratosphere).

The oxides of nitrogen, NO_x, play a remarkable catalytic role in the ozone balance of the atmosphere. Above about 25 km the net effect of NO_x additions to the stratosphere will be a lowering of ozone concentrations by the set of reactions present earlier (R43 + R44 + R45). However, below about 25 km in the stratosphere, NO_x protects ozone from destruction. An important reason for this is a set of reactions:

(R5)	HO₂ + NO	OH + NO₂
(R6)	NO₂ + hv	NO + O
(R7)	O+O₂+M	O₃+M

net:	HO₂ + O₂	OH + O₃

As we have shown before, this reaction set is also basically the cause of ozone production that takes place in the troposphere. In the lower stratosphere, the chain of reactions (R5 + R6 + R7) tends to counteract the destruction of ozone by the catalytic reaction pair:

(R48)	OH + O₃	HO₂ + O₂
(R8)	HO₂ + O₃	OH + 2 O₂

net:	2 O₃	3 O₂

(R48)	OH + O₃	HO₂ + O2
(R5)	HO₂ + NO	OH + NO₂
(R6)	NO₂ + hv	NO + O
(R7)	O+O₂+M	O₃+M

	no net chemical effect.

An additional role of NO_x in the stratosphere occurs through interactions with Cl and ClO. As with NO_x, chlorine atoms and chlorine monoxide molecules participate in an effective catalytic chain of reactions that converts ozone back to molecular oxygen. This cycle goes as follows:

(R43)	O₃ + hv	O + O₂
(R49)	O + CIO	Cl + O₂
(R50)	Cl + O₃	ClO + O₂

net:	2 O₃	3 O₂

transform the catalysts Cl and ClO into HCl, which does not react photochemically with ozone. Furthermore, since the reaction

ties up both some ClO and some NO₂ as non-reactive ClNO₃, it is clear that ozone removal by NO_x additions to the stratosphere is mitigated by the NO_x interference in the chlorine cycle. The photochemical chains of reactions required to explain the distribution of stratospheric ozone and to predict the further effects of human activities are thus quite complicated. However, additions of NO to the low stratosphere may tend to increase local ozone concentrations by reducing some of the ozone loss that would otherwise occur because of the catalytic action of OH and HO₂, and Cl and ClO.

The importance of the ozone production and protection by NO_x in the stratosphere was dramatically emphasized by discoveries of Howard and Evenson (1977) and Zahniser and Howard (1979), who found reaction (R8) and especially (R5) to be much faster than previously estimated. This finding resulted in substantial downward revisions of estimated total ozone-column reductions due to stratospheric NO_x additions from aircraft (Duewer et al., 1977; Crutzen and Howard, 1978; Logan et al., 1978). As NO_x is produced by the oxidation of N₂O via reaction (R47), the same conclusions are also partially valid regarding the possible effects of a future rise in the atmospheric content of nitrous oxide. In this case, however, more of the additional NO_x produced in reaction (R47) will reach the upper regions of the stratosphere, where the reaction set (R43

R45) is more important.

In 1974, theoretical predictions (Crutzen, 1974) of total ozone reductions that used the then recommended rate constants yielded the following approximate relationship between total ozone change (V₃) and a hypothetical increase in the volume mixing ratio of atmospheric nitrous oxide, µ(N₂O),

A doubling in the atmospheric abundance of N₂O was therefore expected to yield a 20% decrease in total ozone. New rate constants for reactions (R5) and (R8), determined by advanced laboratory techniques first led to substantial downward revisions in the ozone-reduction estimates. Actually, it was consequently estimated that an increase in the atmospheric N₂O abundance could lead to an increase in total ozone. Presently, however, it is calculated that an increase in the atmospheric N₂O abundance will lead to a decrease in total ozone. This author's own calculations indicate a loss in total stratospheric ozone of about 12% for a doubling of N₂O, keeping all other factors that affect ozone constant. The reason for this finding is that substantially smaller concentrations of OH in the lower stratosphere below about 30 km are now predicted through new calculations of the rate coefficients related to the formation and photolysis of HO₂, NO₂, and HNO₃. The role of NO₂ is that of a catalyst in the reaction cycles:

(R54)	HO₂ + NO₂(+M)	HNO₄ (+M)
(R55)	OH + HNO₄	H₂O + NO₂ + O₂

net:	OH + HO₂	H₂O + O₂

(R10)	OH + NO₂ (+M)	HNO₃(+M)
(R56)	OH + HNO₃	H₂O + NO₃
(R57)	NO₃ + hv	NO₂ + O

net:	2 OH	H₂O + O

These reactions enhance the loss of HO_x from the lower stratosphere, so that much less HNO_3. and more NO_x, is calculated than was the case before. The newer calculations are in much better agreement with observations (Coffey et al., 1981a). As a consequence, the normal ozone balance in the lower stratosphere is again much more affected by NO_x catalysis, i.e. reactions (R43

R45), and not so much by HO_x, and ClO_x,. Thus the compensation effects of NO_x, in the HO_x, and ClO₂ catalytic cycles, as discussed before, can not make up for the enhanced loss of ozone due to NO_x, catalysis in the entire stratosphere.

The importance of COS as a source of sulphur in the stratosphere, where this gas is efficiently photolysed by ultraviolet radiation has already been discussed. More detailed discussions on the many aspects of stratospheric photochemistry are available in recent reviews (e.g. WMO/NASA,1982).

3.4 THE MOST IMPORTANT CARBON COMPOUNDS

With volume mixing ratios of about 3.3 x 10^-4, carbon dioxide is by far the most abundant carbon-containing gas in the atmosphere. Its atmospheric concentration is increasing by about 0.5% or roughly 1.6 ppm per year (Freyer, 1979). This increase is mainly due to the burning of fossil fuels, which amounts to about 5.3 x 10¹⁵g C/yr (Rotty, 1981). Only about 58% of this input remains airborne, but a better estimate can not be made because the source or sink of CO₂, connected with changes in the global biomass is not well known (see, e.g., Seiler and Crutzen, 1980; Melillo and Gosz, Chapter 6, this volume).

Carbon dioxide is important for the radiative heat balance of the earth. A doubling of atmospheric CO₂, which may become established by the middle of the next century, can cause a climatic warming by 1.5

6^°C (Hansen et al., 1981). Contrary to the effects in the troposphere, an increase in atmospheric CO₂ leads to a cooling of the stratosphere, which leads to increased ozone concentrations, partially offsetting the ozone depletions caused by other anthropogenic activities, such as chlorofluorocarbon release. CO₂ is not a photochemically active gas, except insofar as it affects atmospheric temperatures. We will restrict the discussion of it to what is said above and refer to the extensive reports of the past years, e.g. the SCOPE 13 report (Bolin et al.; 1979).

This gas has somewhat variable concentrations in the atmosphere. High concentrations, about 150

200 ppbv, are found near the earth's surface at the middle and high latitudes of the Northern Hemisphere, which clearly suggests an anthropogenic source (Seiler, 1974). In the Southern Hemisphere the CO volume mixing ratios are near 50

60 ppbv southwards of 20^°S (Seiler, 1974; Heidt et al., 1980). Model calculations show, however, that the anthropogenic source of carbon monoxide cannot be the dominant one. Because of reaction with OH, a very large source of CO must be located in the equatorial regions that cannot be supplied from local industrial emissions or transported from mid-latitudes. Large natural, mainly tropical, sources of CO with a total strength of 2

3 x 10¹⁵g CO/yr must exist. The oxidation of methane can only contribute about 25% of this.

Two important mechanisms of CO production in the tropics have, therefore, been postulated. Zimmerman et al. (1978) proposed that the oxidation of isoprene (C₅H₈) and of terpenes (C₁₀H₁₆), which are emitted by trees, constitute the required source. They estimated worldwide isoprene and terpene emission rates of 3.5 x 10¹⁴ and 4.8 x 10¹⁴ g C/yr respectively, and derived a range of global CO production rates of between 4 and 13 x 10¹⁴g/yr from their oxidation. The other important source of carbon monoxide in the tropics comes from the burning of biomass, mainly due to a variety of land clearing operations, grass fires in savannas, agricultural waste and firewood burning (Seiler and Crutzen, 1980). The total estimated CO source from these fires was estimated at about 8 x 10¹⁴ g C/yr with an uncertainty of at least a factor of 2 (Crutzen et al., 1979). Although the carbon monoxide budget of the atmosphere may be explained in these terms, there are large uncertainties connected with these numerical estimates, especially concerning the tropical emissions.

Recent work by Seiler and Fishman (1981) has identified seasonal variations in mid-latitudes with CO maxima occurring during the winter and CO minima during the summer months. This behaviour can be explained by the much more pronounced photochemical destruction of CO by reaction with OH in summer, when this radical is calculated to be much more abundant than during the winter season. This seasonal behavior may prove to be very useful in deriving better information on the natural contributions of CO to the atmosphere from the oxidation of the non-methane hydrocarbons that are also emitted by trees at maximum rates during summer.

The annual destruction rate of methane in the atmosphere is calculated to be equal to 4 x 10¹⁴ g CH₄ with an uncertainty range of maybe 2

6 x 10¹⁴ g CH₄. This calculation is based on an average tropospheric OH concentration of about 6 x 10¹⁵ molecules cm^-3, which has an uncertainty of a factor of two (Lovelock, 1977; Singh et al., 1979; Derwent and Eggleton, 1981; Gidel et al., 1982) in accordance with the methylchloroform balance of the atmosphere. As uptake of methane by microbiological processes in soils and waters has never been found to be significant (Ehhalt, 1974; Seiler, private communication), this rules against the substantially larger sources of methane that have been proposed by Ehhalt (1974) and recently by Sheppard et al. (1981). From this I derived the tentative budget of the biological sources of methane, which is presented in Table 3.1. A substantial portion of the methane will be oxidized in the atmosphere to CO, but in the absence of enough NO there will be a rain-out of about 10¹⁴ g C as CH₃O₂H. It appears that several sources contribute comparable quantities of CH₄ to the atmosphere, and some of these sources are expanding world-wide. For the period 1972

1978 (FAO, 1975, 1977, 1980), the cattle population increased by 1.2% annually, the rice paddy area by 1.7% and the rice production by 4.6% per year. Of particular importance as an interaction between biogeochemical cycles are the observations by Cicerone and Shetter (1981) who observed a more than fourfold enhancement in CH₄ yields when rice fields were supplied with nitrogen fertilizer. Nevertheless, their global estimate of CH₄ emissions from rice fields is less than 6 x 10¹³ g per year. It is interesting to note that biomass burning is also a significant source of methane to the atmosphere, supplying between 30 and 110 Tg annually (Crutzen et al., 1979). It is quite likely that the extent of burning has also grown substantially during the past decade, but no reliable statistics are available. An additional production of methane of maybe 1.5 x 10¹⁴ g CH₄ per year takes place in the digestive systems of termites during the decay of wood (Zimmerman et al., 1982). This must still be shown by field measurements.

An important, abiological source of CH₄ is venting and leakage of natural gas. Current world-wide natural gas consumption amounts to about 10¹⁵ g CH4/yr. A world-wide average leakage rate of only 2% would supply 2 x 10¹³ g CH₄ annually to the atmosphere. A good statistical evaluation of the leakage rate has not been published, and the 2% number based on unofficial information from gas supply companies is rather uncertain.

Observations by Rasmussen and Khalil (1981) have shown, an annual global increase in methane by about 2% during 1978

1980. This increase has also been discovered by Seiler (private communication) from data extending back to 1976. Spectroscopic measurements from the Jungfraujoch in Switzerland seem to rule out an increase by more than 10% during the time period 1955

1977 (Zander, private communication), so that the increase in atmospheric methane may only have been of more recent dates. Carbon isotope ratio measurements of CH₄ and its potential atmospheric sources may be of great value in identifying the source of the increase in atmospheric CH₄ over the past years (Rust, 1981; Stevens and Rust, 1981). Because of the important role of methane in both stratospheric and tropospheric photochemistry, an increase in CH₄ will significantly affect air chemistry. Furthermore, methane plays a role in the earth's radiation budget (Wang et al., 1976).

Isoprene and terpenes seem to be the main volatile organic compounds that are emitted by trees (Went, 1960; Zimmerman et al., 1978). They are important because they may be converted to CO. Their residence time in the atmosphere is only a few hours. Extrapolating from field measurements of emissions made in the U.S.A., Zimmerman et al. (1978) estimated isoprene and terpene emission rates of 3.5

10¹⁴ and 4.8

10¹⁴ g C/yr respectively. These are uncertain by a factor of two and it may not be excluded that other hydrocarbon gases are also emitted in important quantities in the tropics.

Virtually nothing is known about the oxidation paths of these compounds in the atmosphere. As discussed for methane, it appears that the availability of nitric oxide is an important factor that determines the routes and time constants of their oxidation in the atmosphere. If too little NO is present in the air, it is possible that many photochemically rather long-lived and water-soluble gaseous intermediates (e.g. organic peroxides and alcohols) are formed. These intermediates may be removed by precipitation so that there is no guarantee that the efficiency of CO formation from each emitted carbon atom will be close to unity. Conversely, any future additions of NO from industry and fires to the boundary layer of the tropical forests could substantially shorten the time needed to break down C₅H₈ and C₁₀H₁₆ via aldehydes to carbon monoxide. This may have implications for the future atmospheric CO distribution. The accompanying effects on ozone formation have already been discussed. There may even exist a potential for photochemical smog formation under stable meteorological conditions during the dry season in the trophics.

3.5 THE MOST IMPORTANT NITROGEN COMPOUNDS

In the SCOPE 7 report the global source of NO was estimated at 23 Tg N in 1975 (Söderlund and Svensson, 1976). These estimates were obtained by an extrapolation of earlier 1965 emission estimates of 15 Tg N by Robinson and Robbins (1968). On the other hand, an analysis of NO_x emissions recently made by Böttger et al. (1980), estimates the global NO_x production from combustion engines to range between 8.2 and 18.5 Tg N/yr. For specific continents and regions more detailed analyses have been compiled by Smith (1980). For Western Europe, Smith reports an annual atmospheric emission rate of about 2 Tg N. For the U.S.A. and Japan, production rates were estimated at 6.6 and 0.7 Tg N/yr respectively. A consideration of these more recent figures casts some doubt on the emission estimates by Söderlund and Svensson (1976) and demonstrates clearly an unsatisfactory state of affairs that calls for more complete and up-dated analyses of the industrial NO_x production rates. Besides, because of the short lifetime of NO_x in the atmosphere, regional budgets are important for pollution studies. Globally, we will adopt here a range of 12

20 Tg N per year as the NO_x production rate estimate from industrial high temperature combustion processes.

Considerable attention has recently also been given to the emissions of NO_x by subsonic aircraft above 9 km in the Northern Hemisphere, which is estimated at 0.25 Tg N/yr. (Liu et al., 1980). Although this source of NO is much smaller than that produced by all other fossil fuel combustion processes, the NO_x will reside in the atmosphere much longer than the nitric oxide that is produced at low altitudes, because of a lower probability of uptake of HNO₃ in cloud water. Liu et al. (1980) postulated, therefore, that the aircraft nitric oxide would contribute significantly to the NO_x concentrations in the upper troposphere together with the nitric oxide that is of stratospheric origin. They also postulated that these sources of NO_x would exert a considerable effect on the tropospheric ozone. The validity of this argument also depends, however, on the efficiency of transfer of the industrial NO and NO₂ from the ground to the upper troposphere. As NO and NO₂ are not very soluble and do not react in water (Lee and Schwartz, 1981), such transport may occur by occasional vigorous overturning of the troposphere during frontal passages and thunderstorms. This cannot be estimated by the current 1-D and 2-D models of the atmosphere that consider only averaged motions and average deviations thereof, and that thereby may grossly under-estimate the transfer of photochemically reactive gases to the middle and upper troposphere.

The nitric oxide resulting from this process is mainly a product of the oxidation of the fixed nitrogen that was already bound in the biomas. This source of nitric oxide is hard to estimate, but based on the studies by Crutzen et al. (1979) and by Seiler and Crutzen (1980), it should be of the order of 20 Tg N/yr with about a factor of two uncertainty. This may only represent 30% of all fixed nitrogen that is contained in the burned biomass, so that there remains the possibility of large atmospheric sources of other nitrogenous gases. One of these gases is N₂O; other possibilities are ammonia, hydrogen cyanide (HCN), and other nitriles. The presence of HCN in the atmosphere was predicted by Crutzen et al. (1979), because of a long atmospheric residence time and the likelihood of a substantial atmospheric source from biomass burning (Schmeltz and Hoffmann, 1977). Its presence in the atmosphere above 12 km at an average volume mixing ratio of 0.17 ppbv has recently been reported (Coffey et al., 1981b). The reaction coefficient of HCN with OH varies from about 3 x 10^-14cm³ molecule^-1 s^-1at ground level to slower than 7 x 10^-15cm³ molecule^-1 s^-1above 12 km (Fritz et al., 1981). This means that the oxidation of HCN in the atmosphere can at most yield 0.1 Tg N/yr of NO_x, which is quite insignificant. Rain-out of HCN in the atmosphere may likewise be neglected because of insufficient solubility of HCN in water (Landolt-Börnstein, 1962), so that only uptake at the earth's surface can be a significant sink for atmospheric HCN. With a maximum deposition velocity of about 0.5

1 cm/s on the ocean and at land (Liss and Slater, 1974), the transfer of fixed nitrogen from the atmosphere is at most about 10 Tg N/yr, which may be of ecological significance provided HCN can be used efficiently as a fixed nitrogen source.

For this source of NO widely different estimates ranging from 4.2 TgN/yr (Tuck, 1976) to 40 TgN/yr (Chameides et al., 1977) were published several years ago. Most recent estimates are about 3

4 TgN/yr (Dawson, 1980; Hill et al., 1980), or even 1.8 TgN/yr (Levine et al., 1980). I estimate a wide range of possible values of 1

10 Tg N/yr, for which the more recent estimates, being based on better statistics on lightning stroke energy dissipation and lightning stroke frequency, are more heavily weighted.

This source still awaits careful evaluation. Laboratory studies have shown that emission of NO_x is much dependent on soil acidity (Nelson and Bremner, 1970). The SCOPE 7 report (Söderlund and Svensson, 1976) gives a range of 1

14 Tg N/yr for global emission, based on two Russian studies, while Galbally and Roy (1978) propose 10 Tg N/year as a global rate extrapolating from the results of their box measurements on some Australian soils. With so little new information available, I propose a range of 1

15 Tg N/yr for the NO_x release from soils, which includes the possibility of no appreciable emission at all.

Nitric oxide is also released from the ocean surface as a result of nitrite photolysis. The source strength has been estimated at 0.15 Tg N/yr (Zafiriou and McFarland, 1981)

There are many estimates available for this source of NO, depending on the assumptions which are made about the stratospheric distribution of N₂O, of which there are too few measurements. The assumptions about the tropical distributions of N₂O are especially critical in making these estimates. From the analysis of one balloon flight at 9°N and several observations at other latitudes, Schmeltekopf et al. (1977) derived an average global, vertically integrated, production rate of NO of 1.6 Tg N/yr. The measured profiles of N₂O with high stratospheric N₂O concentrations have been confirmed by three other equatorial flights by the same research group (Goldan et al., 1980). However, the estimate of Schmeltekopf et al. (1977) is too high by a factor of two, because these workers considered the photochemical process to last for 24 hours each day. The analysis by Johnston et al. (1979) avoided this error and led to a production of 1 Tg N/yr. This may also be somewhat on the high side, however, as no proper account could be taken of the temperature dependence of O(¹D) quantum yield in O₃ photolysis near 310 nm. If the estimate by Johnston et al. (1979) has an uncertainty of about 40%, the production of NO from N₂O oxidation may range between 0.7 and 1.4 Tg N/yr. This source determines the abundance of NO in the stratosphere and is critical for stratospheric ozone chemistry. In polar regions, galactic cosmic rays produce NO at rates between 0.024 and 0.036 Tg N annually, depending of the phase of the solar cycle. Sporadic solar cosmic ray events may occasionally produce almost ten times as much NO per event as galactic cosmic rays in a whole year, but on the average this source of NO is also small compared to that derived from N₂O oxidation. The possibility of some significant contribution to the stratospheric NO_x budget from higher layers in the atmosphere can, however, not be totally ruled out. In the ionosphere 200

400 Tg N/yr of NO_x are produced by the ionizing action of short wave ultraviolet radiation and 20 Tg N/yr by auroral activity at solar maximum (Crutzen, 1979). By far most of the fixed nitrogen produced in this way is removed by the reactions

(R58)	NO + hv	N + O (<191 nm)
(R59)	NO + N	N₂ + O

net:	2 NO	N₂+2ON₂+O₂

It is difficult to estimate how much fixed nitrogen may nevertheless escape this sink because it depends on vertical exchange between the stratosphere and the layers above 100 km, about which little is known. A rather small leakage would be sufficient to establish a significant effect, and this may especially take place during the winter at high latitudes when reaction (R58) does not take place.

Although ammonia may not participate in any significant homogeneous gas phase reactions in the atmosphere, it is nevertheless of much importance in the chemistry of the troposphere through its influence on rain-water chemistry (Taylor et al., Chapter 4, this volume). Most fixed nitrogen is circulated in the atmosphere in the form of ammonia. Except during the winter, ammonia is often present at ground levels over land with volume mixing ratios of 1

10 ppbv (Breeding et al., 1973; Shrendikar and Lodge, 1975; Georgii and Lenhard, 1978; Hoell et al., 1980). Over the oceans much lower NH₃ volume mixing ratios seem to be far more common (Ayers and Grass, 1980) . Normally, a very marked drop of NH₃ mixing ratios with height is observed, so that typically volume mixing ratios below 1 ppbv are observed above 3 km (Kaplan, 1973; Georgii and Müller, 1974; Georgii and Lenhard, 1978). Exceptions are the higher values which were measured in the eastern U.S.A. in March 1979 (Hoell et al., 1980).

The role and cycling of NH₃ in the atmosphere are complex. Plants can both absorb (Hutchinson et al., 1972) and emit NH₃ (Farquhar et al., 1979). A study by Denmead et al. (1976) showed emission of NH₃ from the soil, but re-absorption of this NH₃ within the plant canopy. As a consequence of this, an appreciable loss of ammonia to the atmosphere was detected from a grazed pasture, but only a small flux from an ungrazed one. Because of this role of vegetation it is particularly hard to estimate the atmospheric budget of NH₃. In the following we can, therefore, only show the gross inputs of NH₃ to the atmosphere, and even for these the uncertainties are very substantial:

Healey et al. (1970) listed animal urea excretion rates for most important domestic animals in the U.K. From this information and statistics on the population of domestic animals (FAO, 1975), one can extrapolate to an urea production of about 50 Tg/yr in the developed world.

In the developing world the food intake per animal is, however, only 30

60% of that in the developed world, so that urea production per head should be about half as large as in the developed countries. From the available statistics in the developing world one may calculate an additional urea production of 20 Tg/yr, so that the total world-wide urea production by domestic animals adds up to about 70 Tg or 35 Tg N/yr. Healey et al. (1970) assumed that 10% of the nitrogen in the urea would volatilize to NH₃. This release estimate is, however, very low in comparison to the results of field studies obtained by Denmead et al. (1974) and especially Porter (1975) that indicated a relative volatilization loss of 30%. Adopting this ratio as more representative, the global production of NH₃ from urea would be about 10 Tg N/yr. To this must be added the contribution from faeces, which may double the release of ammonia to 20 Tg N/yr (Böttger et al., 1980). This estimate is somewhat smaller than those derived by Söderlund and Svensson (1976) and Böttger et al. (1980). For the release of ammonia from wild animals we will accept the relatively small source of 2

6 Tg N as given by Söderlund and Svensson (1976).

The loss of NH₃ from fertilized agricultural fields is estimated to be about 5% (Bolin and Arrhenius, 1977) . With an annual fertilizer application rate of almost 60 Tg N, the release of NH₃ is equal to about 3 Tg N/yr.

This is an unknown quantity. Dawson (1977) made a detailed analysis of biomass decomposition and from a soil

atmosphere exchange model proposed an NH₃ source of 27 Tg N/yr, with most emission taking place at temperate latitudes. Even if this emission was roughly correct, substantial re-absorption of ammonia is probably going to take place as in the study by Denmead et al. (1976) in the ungrazed pasture. Böttger et al. (1980) actually rule out a significant source of this type on the basis of a few quoted measurements on NH₃ release from soils, and Söderlund and Svensson (1976) do not even discuss it. Here I will consider the value of Dawson (1977) as an upper limit. In particular, I do not rule out a significant release of NH₃ especially during early spring when fresh vegetation is not well developed and dead material on the ground is abundant. When the soil is warmed by solar radiation much release of NH₃ to the atmosphere may indeed occur.

Referring to the discussion given earlier regarding NO release, a volatilization of ammonia of up to 60 Tg N/year that takes place especially in the tropics may not be ruled out. This requires of course that much of the ammonia is not oxidized in the fires. Söderlund and Svensson (1976) mention, however, that ammonia is a surprisingly stable gas in combustion systems.

I will use the estimate of Söderlund and Svensson (1976), who proposed 4

12 Tg N as the global source.

During the years that have passed since the publication of the SCOPE 7 report, much knowledge has been collected about the questions related to N₂O. Atmospheric measurements and measurements of soil release now strongly indicate that the total global source of N₂O must be much smaller than some of the values that were listed by Söderlund and Svensson (1976). Photochemical reactions in the stratosphere may, therefore, represent the main sink for atmospheric nitrous oxide. For reasons stated before, the loss rate of 15 Tg N₂O-N/yr of Schmeltekopf et al. (1977) is probably too high by a factor of 2. Instead, Johnston et al. (1979) estimated a loss rate of 8.4 Tg N/yr. With this loss rate the atmospheric lifetime of N₂O is about 180 years.

The processes that add significant and similar quantities of N₂O to the atmosphere will be discussed below.

Weiss and Craig (1976) estimated that 1.6 Tg N would be produced annually by coal and fuel oil burning. To this must be added about 0.15 Tg N from natural gas burning as estimated by Pierotti and Rasmussen (1976). The fossil fuel combustion source of N₂O is increasing at a rate of 3.5% per year. The increase in N₂O that results from this source almost satisfies the observed world-wide increase in N₂O concentrations by about 0.2% per year (Weiss, 1981). Additional, potentially important production of N₂O may take place by the use of reducing catalysts in automobile engines to cut down on NO emissions (Pierotti and Rasmussen, 1976; Weiss and Craig, 1976). Weiss and Craig (1976) estimated a potential global rise in N₂O emissions rates by 2.1 Tg N/yr, if all 1976 year cars were to be equipped with Pt catalysts.

This source was estimated by Crutzen et al. (1979) at 8 Tg N/yr. More recent data (to be published) indicate, however, a lower emission rate of 1

2 Tg N/yr.

According to Cohen and Gordon (1979) this source of N₂O may range between 4 and 10 Tg N per year, whereby the production of N₂O would take place mainly during nitrification. Elkins et al. (1978) estimate an N₂O release of 7 Tg N/yr for the latitude belt 20°S to 20°N, and in a recent review Hahn (1981) derives about 14 Tg N/yr for the entire ocean. Seiler (1981) and Weiss (1981) from their extensive oceanic observations conclude, however, that the oceanic source of atmospheric N₂O is small compared to the stratospheric sink. Although significant quantities of N₂O may be released from estuaries (Kaplan et al., 1978), the N₂O production from oceans and estuaries is probably much lower than 10 Tg N/yr.

According to estimates by Bohn (1977) and Wilson (1978), during the past century the average annual, world-wide loss of organic carbon due to the ploughing of agricultural soils has amounted to 1

2 x 10¹⁵sg C. With an average C/N ratio of 20 (Delwiche and Likens, 1977), the annual loss of fixed nitrogen from these soils has amounted to 50

100 Tg N. In a recent study Terry et al. (1980) reported that 2.7% of the fixed nitrogen that is lost in the drained, cultivated soils of South Florida did appear as N₂O. If similar ratios applied for other ecosystems that lose organic nitrogen due to agricultural activities, the annual global source of N₂O would be equal to 1

3 Tg N. This source of N₂O may, therefore, have been significant during the past century and may have led to an increase in atmospheric N₂O.

Recent studies generally show relative emissions of N₂O nitrogen compared to fertilizer nitrogen input typically of the order of 1% or less (McKenney et al., 1978; Breitenbeck et al., 1980; Conrad and Seiler, 1980; Mosier and Hutchinson, 1981; Mosier et al., 1981a, b; Ryden, 1981; Seiler and Conrad, 1981). There is a tendency for the relative N₂O release to be higher for ammonium containing fertilizer, an indication for the importance of the nitrification process as a source of atmospheric N₂O. High yields of N₂O (4

6.8%) were measured by Bremner et al. (1981) after application of anhydrous ammonia. Altogether, the global source of N₂O from nitrogen fertilization is probably smaller than a few Tg N/yr. Thus the impact of the increased nitrogen fertilization on stratospheric ozone is probably not a very urgent matter. This view is supported by the observations by Weiss (1981) of a global upward trend in atmospheric N₂O of only 0.2% per year. In fact, Weiss explains that all the atmospheric N₂O increase is due to the increased rate of fossil fuel combustion.

About this very little information is available. Seiler and Conrad (1981) normally found even smaller N₂O release rates (0.1

4g N/ha/day) in natural soils than in fertilized soils near Mainz, Germany. Mosier et al. (1981b) derived an average N₂O flux of 2.3 g N/ha/day from a native prairie in N.E. Colorado during the summer months. In the latter case the loss of N₂O accounted for 10% of the N inputs from atmospheric deposition and N₂ fixation. Earlier studies quoted by Hahn and Junge (1977) gave release rates of N₂O from unfertilized or natural soils in the range 0.1

34 g N/ha/day. With such limited and highly variable information it is clearly possible only to guess at global N₂O releases from terrestrial ecosystems.

3.6 THE MOST IMPORTANT SULPHUR COMPOUNDS

A detailed compilation of the industrial atmospheric sources of SO₂ was made by Cullis and Hirschler (1980). The man-made sources for 1976 add up to 104 Tg S/yr, of which 62% comes from coal, 25% from petroleum, 11% from non-ferrous ores, and 2% from a variety of other industries.

Another potential source of sulphur dioxide could be biomass burning. With typical S/C atomic ratios of about 0.002 in biomass, the global emission of sulphur from this source could be about 6 Tg S/yr with an uncertainty of a factor of two. SO₂ may not be the dominant sulphur gas produced in fires, because an appreciable fraction of more reduced gases, such as H₂S and COS, may also be emitted.

An important natural source of SO₂ to the atmosphere is provided by volcanic eruptions. Starting from the very low production estimates of 1.5 Tg SO₂/yr by Kellogg et al. (1972), the estimated SO₂ inputs have been going up: 7.5 Tg SO₂/yr (Cadle, 1975), 47 TG SO₂/yr (Naughton et al., 1975) and maybe 100 Tg SO₂/yr (Cadle, 1980). An earlier estimate by Bartels (1972), based on elemental analysis of the Greenland icecap, was 34 Tg SO₂/yr. Therefore, volcanic emissions apparently contribute 10

30% of the anthropogenic SO₂ input to the atmosphere. During periods with intense volcanic activity the contributions may be substantially larger.

The SCOPE 7 compilation by Granat et al. (1976) gives a total annual emission rate of biogenic, reduced sulphur of 37 Tg S/yr, about 90% of which emanates from the oceans or coastal waters. Cullis and Hirschler (1980) give, however, appreciably higher biogenic sulphur fluxes of almost 100 Tg S annually, about half of which comes from the land. The latter result seems to have been obtained by a questionable extrapolation of some measured sulphur fluxes over fertilized agricultural fields in Sweden (Siman and Jansson,1976) to all ecosystems of the world. At this stage it is, therefore, hard to assign any significance to the H₂S release rates given by Cullis and Hirschler, except their value for agricultural soils in the developed world of about 4 Tg S annually. Biogenic sulphur flux estimates over non-saline soils in the eastern U.S.A. by Adams et al. (1980) show large variabilities by orders of magnitude, but average emission rates were estimated at 0.2 gm^-2yr^-1 for coastal wetlands, 0.3 gm^-2yr^-1for inland soils high in organic matter and 0.015 g^-2yr^-1for more `typical' inland soils. For all soils in the study area an average emission rate of 0.03 gm^-2yr^-1was calculated. This is about a factor of ten smaller than the sulphur emission of 0.2

0.3 kg m^-2yr^-1obtained by Siman and Jansson (1976) for fertilized Swedish fields, so that the contribution of agricultural sulphur emissions by Cullis and Hirschler (1980) should be considered as an upper limit.

Airborne observations by Crawford and Reisinger (1980) in the South-eastern United States have indicated the possibility of substantial biogenic sulphur fluxes from wetland areas of the order of 12 g Sm^-2yr-¹. If such a flux were characteristic for most terrestrial wetlands of the world with a total estimated area of 2 x 10¹²m²,an annual flux of 24 Tg S would result. Considering the much lower emissions given below, this estimate may be on the high side.

There exists a large variability in reported sulphur emissions from salt marshes and tidal flats, with emission values quoted as 0.5

1940 gm^-2 yr^-1 (Adams et al., 1980), 39 gm^-2yr^-1 (Steudler and Peterson, 1980) and <0.2

0.6 gm^-2yr^-1 (Aneja et al., 1979a). To extrapolate from these values to global estimates may be hopeless. More promising are methods that average over emissions from larger areas of an ecosystem. For a salt marsh at Wallops Island, Goldberg et al. (1981) estimated average H₂S emission rates of 0.4 g S m^-2yr^-1in December and 6 g Sm^-2yr^-1 in July, maybe due to the expected temperature dependence. The value of these estimates is, clearly dependent on good interpretation of the meteorology during the particular observation period. Only a limited analysis of this could be made by Goldberg et al. (1981). If I derive from their work average emission rates of 3 and 6 gm^-2yr^-1 for the temperate and the tropical latitudes respectively, considering that these fluxes would represent sulphur emissions in estuaries, in algal bed and reef areas and oceanic upwelling zones with a combined area of 2.4 x 10¹²m²(Whittaker and Likens, 1975), the production of sulphur from these ecosystems could almost reach 10 Tg S/yr.

Sulphur emission rates from the open oceans have been estimated by Bonsang (1980) to be in the range 31

42 Tg S annually. This range is larger than the estimate of 27 Tg S/yr of Granat et al. (1976). As only emissions of dimethylsulphide could be measured, this range may have to be shifted upwards.

Relatively large H₂S emissions may occur in the tropics, but much of the emitted sulphur may be recycled within the tropical forest stands (Brinkmann and Santos, 1974). Relatively large H₂S emissions of 16 Tg S/yr have, nevertheless, been proposed to occur in wet tropical forests (Delmas et al., 1980). These emissions give rise to precipitation with a pH range of 4.7

5.7 (Stallard and Edmond, 1981).

Clearly, the world-wide biogenic and volcanic emissions are poorly known, but together they could be in the range 70

100 Tg S/yr. Thus, on a global average, they may turn out to be quite comparable to the industrial emissions. In the heavily industrialized areas of the Northern Hemisphere, the industrial emissions should, however, be much larger than the natural emissions, (e.g. Granat et al., 1976; Galloway and Whelpdale, 1980).

The importance of COS for the stratospheric sulphate layer has been discussed before. Besides an estimated loss rate of about 5 x 10¹⁰ S/yr by photolysis in the stratosphere, little is known about the sources and sinks of this compound. Possible sources include emissions from wetland areas (Aneja et al., 1979a, b; Adams et al., 1980), coal processing, volcanic and fumarolic emissions (Crutzen, 1976), biomass burning (Crutzen et al., 1979) and oxidation of CS₂. Besides through photochemical breakdown in the stratosphere, COS may also be removed from the atmosphere by deposition on soils and vegetation and on the ocean surface. Decomposition and subsequent hydrolysis in the ocean may limit the lifetime of COS to 10 years (Johnson, 1981), which is much shorter than the lifetime of about 30 years due to stratospheric photolysis. It appears, however, that the ocean surface waters are always supersaturated with COS, so that they supply about 0.4 Tg S annually to the atmosphere (Rasmussen, private communication). Altogether, the global source may exceed 1 Tg S per year, so that COS must be taken up at the earth's surface, most probably by plants. Graedel et al. (1981) have pointed to the potential role of COS as a corrosive gas.

The industrial release of CS₂ to the atmosphere in the U.S.A. is equal to 1.35 x 10¹¹g/yr (Peyton et al., 1976), which may be extrapolated to a world-wide, annual release of about 4 x 10¹¹g. The resulting production of COS via reactions (R32a-c) is equal to 1.7 x 10¹¹g S/yr, emphasizing that COS may be influenced by anthropogenic activities. It is easy to show that the average measured volume mixing ratio of 30 pptv in the continental U.S.A. boundary layer (0

2 km) in May and June is compatible with a mean OH concentration of about 1.5 x 10⁶ molecules cm^-3, so that anthropogenic emissions may well explain much of the observations of Bandy et al. (1981).

3.7 CONCLUSIONS

There are many photochemical interactions between the atmospheric cycles of carbon, nitrogen, and sulphur containing species, so that these species cannot be studied independently of each other. In the troposphere the interactions between carbon and nitrogen compounds contribute significantly to the abundance of tropospheric ozone. The photolysis of ozone by ultraviolet solar radiation leads to the formation of the highly reactive OH radical, which is the only tropospheric compound capable of attacking such important gases as CO, NO_x H₂S, and the hydrocarbons, including those containing sulphur and halogens. Without the presence of ozone in the troposphere, the atmospheric abundance of all these gases would have been much higher than observed, in several cases by orders of magnitude. This is clearly of profound importance in air chemistry. Because of man's activities the factors determining the photochemistry of tropospheric ozone are changing. Thus, it is important to improve our understanding of the distribution of this gas in the troposphere, which also includes better knowledge about the flux of stratospheric ozone to the troposphere.

Using the currently available knowledge of tropospheric chemistry I have discussed important photochemical reactions in the atmosphere and have also derived estimates of the sources and sinks of several important atmospheric trace gases as compiled in Table 3.1 and Figure 3.1. The influence of man's activities on most of the processes is substantial. The fluxes of compounds to the stratosphere are relatively small, but they are, in several cases, of fundamental importance for the photochemistry of the stratosphere. Some compounds that are inert in the troposphere, such as N₂O, CFCl₃, CF₂Cl₂ and CCl₄, are of special importance as their photochemical breakdown in the stratosphere leads to the reactive radicals NO, NO₂, Cl and ClO that attack ozone through catalytic chains of reactions. Therefore, even small inputs of these and other stable compounds at the earth's surface may have profound effects on upper atmospheric chemistry.

Finally, due to expansions in agriculture during the past century the soils have lost about 1

2 x 10¹⁵g C/yr (Bohn, 1977; Wilson, 1978). With an N/C ratio of 0.05 (Delwiche and Likens, 1977), this implies a loss of 50

100 Tg N each year from agricultural soils, which is comparable to the present use of artificial nitrogen fertilizer. The fate of these two sources of fixed nitrogen in the environment may well have similarities. In the same process also about 35

50 Tg S have been affected annually, a substantial fraction of which has moved from the agricultural soils into different environments (see McGill and Christie, Chapter 9, this volume).

ACKNOWLEDGEMENTS

For helpful discussions and comments I thank Drs. R. Conrad, J. Lovelock, H. Niki, W. Seiler, Mr. R. Chatfield and the participants at this SCOPE meeting. Special thanks go to Dr. A. F. Tuck for this thorough and generously constructive review of this paper.

3.8 REFERENCES

Aneja, V. P., Overton, J. H., Cupitt, L. T., Durham, J. L., and Wilson, W. E. (1979a) Direct measurement of emission rates of some atmospheric biogenic sulphur compounds, Tellus, 31, 174-178.

Angell, J. K., and Korshover, J. (1979) Comparison of ozone variations derived from ozonosondes and Umkehr measurements for the period 1969-1976, Mon. Weather Rev., 107, 599-607.

Atkinson, R., Darnall, K. R., Lloyd, A. C., Winer, A. M., and Pitts, J. N., Jr. (1979) Kinetics and Mechanisms of the reaction of the hydroxyl radical with organic compounds in the gas phase, Adv. Photochem., 11, 375-488.

Bandy, A. R., Maroulis, P. J., Shalaby, L. and Wilner, L. A. (1981) Evidence for a short tropospheric residence time for carbon disulphide, Geophys. Res. Lett., 8, 1180-1183.

Bartels, O. G. (1972) An estimate of volcanic contributions to the atmosphere and volcanic gases and sublimates as the source of the radio isotopes ¹⁰Be, ³⁵S, ³²P and ²²Na, Health Phys., 22, 387-392.

Bolin, B., Degens, E. T., Duvigneaud, P., and Kempe, S. (eds) (1979) The Global Carbon Cycle, SCOPE Report No. 16, New York, Wiley.

Böttger, A., Ehhalt, D. H., and Gravenhorst, G. (1980) Atmosphärische Kreisläufe von Stickoxyden and Ammoniak, Report Jül-1558, Institut für Chemie 3, Kernforschungsanlage Jülich (FRG).

Breitenbeck, G. A., Blackmer, A. M. and Bremner, J. M. (1980) Effects of different nitrogen fertilizers on emission of nitrous oxide from soil, Geophys. Res. Lett. 7, 85-88.

Brinkmann, W. L. F. and Santos, V. de M. (1974) The emission of biogenic hydrogen sulphide from Amazonian floodplain Lakes, Tellus, 26, 262-267.

Cadle, R. D. (1980) A comparison of volcanic with other fluxes of atmospheric trace gas constituents, Rev. Geophys. Space Phys.,18, 746-752.

Cicerone, R. J., and Shetter, J. D. (1981) Sources of atmospheric methane: measurements in rice paddies and a discussion, J. Geophys. Res., 86, 7203-7209.

Coffey, M. T., Mankin, W. G., and Goldman, A. (1981a) Simultaneous spectroscopic determination of the latitudinal, seasonal and diurnal variability of stratospheric N₂O, NO, NO₂ and HNO₃, J. Geophys. Res., 86, 7331-7341.

Cohen, Y., and Gordon, L. I. (1979) Nitrous oxide production in the ocean, J. Geophys. Res., 84, 347-353.

Cox, R. A., and Roffey, M. J. (1977) Thermal decomposition of peroxyacetyl nitrate in the presence of nitric oxide, Environ. Sci. Technol., 11, 900-906.

Crawford, T. L., and Reisinger, L. M. (1980) Transport and transformation of sulphur oxides through the Tennessee Valley Region, EPA

600/7

126, June 1980.

Crutzen, P. J. (1970) The influence of nitrogen oxides on the atmospheric ozone content, Q.J.R. Meteorol. Soc., 96, 320-325.

Crutzen, P. J. (1973). A discussion of the chemistry of some minor constituents in the stratosphere and troposphere, PAGEOPH, 106-108, 1385-1399.

Crutzen, P. J. (1976) The possible importance of CSO for the sulphate layer of the stratosphere, Geophys. Res. Lett., 3, 73-76.

Crutzen, P. J., Fishman, J., Gidel, L. T., and Chatfield, R. B. (1978) Numerical investigations of the photochemical and transport processes which affect halocarbons and ozone in the atmosphere. Annual Summary of Research, Department of Atmospheric Sciences, Colorado State University, Fort Collins, Colorado.

Crutzen, P. J., and Howard, C. J. (1978) The effect of the HO₂ + NO reaction rate constant on one-dimensional model calculations of stratospheric ozone perturbations, PAGEOPH, 116, 497-510.

Davis, D. D., and Klauber, G. (1975) Atmospheric gas phase oxidation mechanisms for the molecule SO₂, Int. J. Chem. Kin. Symp., Series 1, 543-556.

Dawson, G. A. (1977) Atmospheric Ammonia from undisturbed land, J. Geophys. Res. 82, 3125-3133.

Delmas, R., Baudet, J., Servant, J., and Baziard, Y. (1980) Emissions and concentrations of hydrogen sulphide in the air of the tropical forest of the Ivory Coast and temperature regions in France, J. Geophys. Res., 85, 4468-4474.

Demerjian, K. L., Kerr, J. A., and Calvert, J. G. (1974) The mechanism of photochemical smog formation, Adv. Environ. Sci. Technol., 4, 1-262.

Denmead, O. T., Simpson, J. R., and Freney, J. R. (1974) Ammonia flux into the atmosphere from a grazed pasture, Science, 185, 609-610.

Duce, R. A. (1982) Biochemical cycles and air/sea exchanges of aerosols, Chapter 16 this volume.

Duewer, W. H., Wuebbles, D. J., Elsaeeser, H. W., and Chang, J. S. (1977) NO_x catalytic ozone destruction: sensitivity to rate coefficients, J. Geophys. Res., 82, 935-942.

Elkins, J. W., Wofsy, S. C., McElroy, M. B., Kolb, C. E., and Kaplan, W. A. (1978) Aquatic sources and sinks of nitrous oxide, Nature, 275, 602-604.

FAO (Food and Agricultural Organisation) (1975) Production Yearbook 1974, 28, Rome, FAO.

FAO (Food and Agricultural Organisation) (1980) Production Yearbook 1979, 33, Rome, FAO.

Fishman, J., Solomon, S., and Crutzen, P. J. (1979a) Observational and theoretical evidence in support of a significant in-situ photochemical source of tropospheric ozone, Tellus, 31, 432-446.

Freyer, H. D. (1979) Variations in the atmospheric CO₂ content, in: Bolin, B., Degens, E. T., Kempe, S., and Ketner, P., (eds) The Global Carbon Cycle, SCOPE Report No. 13, Chichester, Wiley, 79-99.

Galbally, I. E., and Roy, C. R. (1978) Loss of fixed nitrogen from soils by nitric oxide exhalation, Nature, 275, 734-735.

Garland, J. A. and Penkett, S. A. (1976) Absorption of peroxy acetyl nitrate and ozone by natural surfaces, Atmos. Environ., 10, 1127-1131.

Georgii, H. W., and Müller, W. J. (1974) On the distribution of ammonia in the middle and lower troposphere, Tellus, 26, 180-184.

Goldan, P. D., Kuster, W. C., Albritton, D. L., and Schmeltekopf, A. L. (1980) Stratospheric CFCl₃, CF₂Cl₂ and N₂O height profile measurements at several latitudes, J. Geophys. Res., 85, 413-423.

Graedel, T. E., Kammlott, G. W., and Franey, J. P. (1981) Carbonyl sulphide: potential agent of atmospheric sulphur corrosion, Science, 212, 663-665.

Hack, W., Schacke, H., Schröter, M., and Wagner, H. Gg. (1978) Reaction of NH₂-radicals with NO, NO₂, C₂H₂, C₂H₄and other hydrocarbons, 17th Symposium (International) on Combustion, 505-513.

Hahn, J., and Junge, C. (1977) Atmospheric nitrous oxide: a critical review, Z. Naturforsch., 32a, 190-214.

Hanst, P. L., Spiller, L. L., Watts, D. M., Spence, J. W., and Miller, M. F. (1975) Infrared measurement off fluorocarbons, carbon tetrachloride, carbonyl sulphide and other atmospheric trace gases, J. Air Pollut. Control Ass. 25, 1220-1226.

Heidt, L. E., Krasnec, J. P., Lueb, R. A., Pollock, W. H., Henry, B. E., and Crutzen, P. J. (1980) Latitudenal distributions of CO and CH₄ over the Pacific, J. Geophys. Res., 85, 7329-7365.

Hill, R. D., Rinker, R. G., and Wilson, H. D. (1980) Atmospheric nitrogen fixation by lightning, J. Atmos. Sci., 37, 179-192.

Hofmann, D. J., and Rosen, J. M. (1981) On the background stratospheric aerosol layer, J. Atmos. Sci., 38, 168-181.

Huebert, B. J., and Lazrus (1978) Global tropospheric measurements of nitric acid vapor and particulate nitrate, Geophys. Res. Lett., 5, 577-580.

Inn, E. C. Y., Vedder, J. F., and Tyson, B. J. (1979) COS in the stratosphere, Geophys. Res. Lett., 6, 191-193.

Johnston, H. S., Serang, O., and Podolske, J. (1979) Instantaneous global nitrous oxide photochemical rates, J. Geophys. Res., 84, 5077-5082.

Jones, B. M. R., Burrows, J. P., Cox, R. A., and Penkett, S. A. (1982) The atmospheric photo-oxidation of CS₂, Atmos. Environ. (submitted).

Kaplan, W. A., Elkins, J. W., Kolb, C. E., McElroy, M. B., Wofsy, S. C., and Duran, A. P. (1978) Nitrous oxide in fresh water systems: an estimate for the yield of atmospheric N₂O associated with disposal of human waste, PAGEOPH, 116, 423-438.

Kurasawa, H., and Lesclaux, R. (1980) Rate constant for the reaction of NH₂ with ozone in relation to atmospheric processes, Chem. Phys. Lett. 72, 437-442.

Kurylo, M. J. (1978) Flash photolysis resonance fluorescence investigation of the reaction of OH radicals with dimethyl sulphide, Chem. Phys. Lett. 58, 233-237.

Landolt-Börnstein (1962) Zahlenwerte und Funktionen, Vol. 5, Berlin, Springer-Verlag, 629.

Lee, Y. N and Schwartz, S. E. (1981) Evaluation of the rate of uptake of nitrogen dioxide by atmospheric and surface liquid water, J. Geophys. Res. (submitted).

Lesclaux, R., and Demissy, M. (1977) On the reaction of NH₂ radicals with oxygen, Nouveau J. Chim., 1, 443-444.

Levine, S. Z., and Schwartz, S. E. (1981) In-cloud and below-cloud scavenging of nitric acid vapor, Atmos. Environ. (in press).

Levy, II, H. (1974) Photochemistry of the troposphere, Adv. Photochem., 9, 5325-5332.

Liu, S. C., Kley, D., McFarland, M., Mahlman, J. D., and Levy II, H. (1980) On the origin of tropospheric ozone, J. Geophys. Res., 85, 7546-7552.

Logan, J. A., Prather, M. J., Wofsy, S. C., and McElroy, M. B. (1978). Atmospheric Chemistry: response to human influence, Phil. Trans. R. Soc. (London), 290, 187-234.

Lovelock, J. E. (1977) Methyl chloroform in the troposphere as an indicator of OH radical abundance, Nature, 267, 32-34.

McElroy, M. B., and McConnell, J. C. (1971) Nitrous oxide: a natural source of stratospheric NO, J. Atmos. Sci., 28,1095-1098.

McGill, W. B., and Christie E. K. Biogeochemical aspects of nutrient cycle interactions in soil and organisms, Chapter 9, this volume.

Maroulis, P. J., Torres, A. L., Goldberg, A. B., and Bandy, A. R. (1980) Atmospheric SO₂ measurements on project Gametag, J. Geophys. Res., 85, 7345-7349.

Mosier, A. R., and Hutchinson, G. L. (1981) Nitrous oxide emissions from cropped fields, J. Environ. Qual., 10, 169-173.

Mosier, A. R., Stillwell, M., Parton, W. J., and Woodmansee, R. G. (1981b) Nitrous oxide emissions form a native shortgrass prairie, Soil Sci. Soc. Am. J. (submitted).

Naughton, J. J., Lewis, V., Thomas, D., and Finlayson, J. B. (1975). Fume compositions found at varies stages of activity at Kilauea volcano, Hawaii, J. Geophys. Res., 80, 2963-2966.

Nicolet, M., and Vergison, E. (1971) L'Oxyde azoteux dans la stratosphere, Aeronomica Acta, A 91.

Noxon, J. F. (1978) Tropospheric NO₂, J. Geophys. Res., 83, 3051-3057, Correction: J. Geophys. Res., 85, 4560-4561 (1980).

Noxon, J. F. (1981) NO_x in the mid-Pacific troposphere. Geophys. Res. Lett., 8, 1223-1226.

Penkett, S. A., Sandalls, F. J., and Jones, B. M. R. (1977) PAN measurements in England - Analytical methods and results, VDI-Berichte, 270, 47-54.

Peyton, T. O., Steele, R. V., and Mabey, W. R. (1976) Carbon disulphide, carbonyl sulphide, Final Report EPA Contract No. 68-01-2940, Stanford Research Institute, Menlo Park, CA, USA.

Porter, K. S. (1975) Nitrogen and Phosphorus. Food production, Waste and the Environment, Ann Arbor, Michigan, Ann Arbor Science Publishers Inc.

Rasmussen, R. A., and Khalil, M. A. K. (1981) Atmospheric methane (CH₄): trends and seasonal cycles, J. Geophys. Res., 86, 9826-9832.

Robinson, E., and Robbins, R. C. (1968) Sources, abundance and fate of gaseous atmospheric pollutants, Final Report, Project P.R.-6755, Stanford Research Institute, Menlo Park, CA, USA.

Rotty, R. M. (1981) Distribution and changes in industrial carbon dioxide production, Reference IV-l, WMO/ICSU/UNEP Scientific Conference on Analysis and Interpretation of Atmospheric CO₂ Data, Report WCP-14, Geneva, Switzerland, September 1981.

Rust, F. (1981) Ruminant methane (¹³C/¹²C) values: relation to atmospheric methane, Science, 211,1044-1046.

Sandalls, F. J., and Penkett, S. A. (1977) Measurements of carbonyl sulphide and carbon disulphide in the atmosphere, Atmos. Environ., 11, 197-199.

Schmeltekopf, A. L., Albritton, D. L., Crutzen, P. J., Goldan, P. D., Harrop, W. J., Henderson, W. R., McAfee, J. R., McFarland, M., Schiff, H. I. Thompson, T. L., Hoffman, D. J., and Kjome, N. T. (1977) Stratospheric nitrous oxide altitude profiles at various latitudes, J. Atmos. Sci., 34, 729-76.

Seiler, W., and Conrad, R. (1981) Field measurements of natural and fertilizer-induced N₂O release rates from soils. J. Air Pollution Control Ass. (in press).

Seiler, W., and Fishman, J. (1981) The distribution of carbon monoxide and ozone in the free troposphere, J. Geophys. Res. (submitted).

Shaw, R. W., and Rodhe, H. (1981) Non-photochemical oxidation of SO₂ in regionally polluted air during winter, Report CM-53. Dept. of Meteorology, University of Stockholm.

Shrendikar, A. D., and Lodge Jr, J. P. (1975) Micro-determination of ammonia by the ring oven technique and its application to air pollution studies, Atmos. Environ., 9, 431-435.

Singh, H. B., and Hanst, P. L. (1981) Peroxyacetyl nitrate (PAN) in the unpolluted atmosphere: an important reservoir for nitrogen oxides, Geophys. Res. Lett., 8, 941-944.

Singh, H. B., Sales, L. J., Shigeishi, H., Smith, A. J., Scribner, E., and Cavanagh, L. A. (1979) Atmospheric distributions, sources and sinks of selected halocarbons, hydrobcarbons, SF₆, and N₂O, Environmental Protection Agency Report EPA

600/3

107.

Söderlund, R., and Svensson, B. H. (1976) The global nitrogen cycle, in, Svensson, B. H., and S6derlund, R. (eds) Nitrogen, Phosphorus, and SulphurGlobal Cycles, SCOPE Report No. 7, Ecol. Bull. (Stockholm), 22, 23-73.

Stallard, R. F., and Edmond, J. M. (1981) Geochemistry of the Amazon. I: Precipitation chemistry and the marine contribution to the dissolved load at the time of peak discharge, J. Geophys. Res., 86, 9844-9858.

Stevens, C. M., and Rust, F. E. (1981) The carbon isotope composition of atmospheric methane, J. Geophys. Res., (submitted).

Svensson, B. H., and Söderlund, R. (eds) (1976) Nitrogen, Phosphorus and SulphurGlobal Cycles, SCOPE Report NO. 7, Ecol. Bull. (Stockholm), 22.

Taylor, G. S., Baker, M. B., and Charlson, R. J. Atmospheric interactions of C, N and S cycles: the role of aerosols and clouds, Chapter 3, this volume.

Torres, A. L., Maroulis, P. J., Goldberg, A. B., and Bandy, A. R. (1980) Atmospheric OCS measurements on Project GAMETAG, J. Geophys. Res., 85, 7357-7360.

Tuazon, E. C., Winer, A. M., and Pitts, Jr, J. N. (1981) Trace pollutant concentrations in a multi-day smog episode in the California south coast air basin by long pathlength FT-IR spectroscopy, Environ. Sci. Technol. (in press).

Turco, R. P., Whitten, R. C., Toon, O. B., Pollack, J. B., and Hamill, P. (1980) OCS, stratospheric aerosols and climate, Nature, 283, 283-286.

Wang, W. C., Yung, Y. L., Lacis, A. A., Mo, T., and Hansen, J. E. (1976) Greenhouse effects due to man-made perturbations of trace gases, Science, 194, 685-690.

Weiss, R. F., and Craig, H. (1976) Production of atmospheric nitrous oxide by combustion, Geophys. Res. Lett., 3, 751-753.

Went, F. W. (1960) Organic matter in the atmosphere and its possible relation to petroleum formation, Proc. Nat. Acad. Sci., 46, 212-221.

Wilson, A. T. (1978) Pioneer agriculture explosion and CO₂ levels in the atmosphere, Nature, 273, 40-41.

Wofsy, S. C. (1976) Interactions of CH₄ and CO in the Earth's atmosphere, Ann. Rev. Earth Planet. Sci., 4, 441-469.

Zahniser, M., and Howard, C. J. (1979) The reaction of HO₂ with O₃. J. Chem. Phys., 73,1620-1626.

Zimmerman, P. R., Greenberg, J. P., and Crutzen, P. J. (1982) Termites: a potentially large source of atmospheric trace gases, Science 218, 563-565.

COMMENT TO CHAPTER 3

During the past decade the scientific understanding of the great cycles of the elements at the Earth's surface seems to have matured. The biosphere is now accepted as a participant and the stratosphere, once the exclusive domain of elitist aeronomers, is now the concern of us all. These are, I think, all changes for the good and it was therefore a pleasure to read Dr. Crutzen's paper on atmospheric interactions and have them confirmed.

Indeed I am so much in accord with the spirit of the paper that criticism would be difficult were it not for a few disagreements over matters of fact. There follows my comments on these disagreements and also on things we agree about.

Can we be sure that there are no gaseous phosphorus compounds in the atmosphere (section 3.1)? Among the compounds we have sought but not yet found are P(CH₃)₃ and NPO. They would be interesting even at the 10^-12ppv level in view of their probable brief residence times. The existence of As(CH₃)₃ and N(CH₃)₃ makes P(CH₃)₃ worth looking for.

One of the key steps in the better understanding of atmospheric chemistry was the discovery of the role of the hydroxyl radical (section 3.2). I would like to ask Dr. Crutzen if he is confident that ozone is the principal precursor of OH in the natural troposphere. There is some compelling although circumstantial evidence for active photochemistry involving oxides of nitrogen over tropical waters and in the sea surface of these regions.

The probable climatic consequences of the accumulation of CO₂ are well discussed. To many it must come as a surprise to hear that tropospheric ozone has a similar warming effect (section 3.2). Since the abundance of both of these gases is connected with combustion and since their effects are additive, we shall presumably hear more about tropospheric ozone in the near future.

I agree that the natural and man induced burning of tropical vegetation is an important source of atmospheric nitrogen compounds and indirectly as a source of tropospheric ozone (section 3.2 and section 3.5). Among the interesting candidate compounds known to be present in smoke but not discussed, are methyl nitrite and methyl nitrate.

Oceanic sources of NO and associated compounds seem to be dismissed in a single reference (McFarland, 1979). Yet Zafiriou, McFarland and Bromund (1980) propose that the tropical oceans are a source of tropospheric NO. They found that the photolysis of the nitrite ion in sea water by solar radiation in the near ultraviolet generates both OH radicals and NO. The NO concentration in equilibrium with the tropical ocean was up to 1000 times greater than its observed aerial concentration (10^-11 ppv).

During an expedition on the Meteor in 1973 to the tropical regions of the North Atlantic I found the production of high concentrations of PAN and other alkylperoxy nitrates. This observation, which was at the time a mere curiosity, takes on a new significance in the light of the Zafiriou and McFarland (1981) findings and also when the unexpected atmospheric stability of these nitrate esters is taken into account.

An interesting sulphur compound not here discussed is CH₃SO₃H. According to Cox (personal communication) methyl sulphonic acid is a significant product in the oxidation of CH₃SCH₃ and CH₃SH by OH radicals. CH₃SO₃H is physically and chemically rather like H₂SO₄ and may not be distinguished from it in the normal routine of sulphate aerosol measurement. The measurement of the ratio of CH₃SO₃H to H₂SO₄ in the tropospheric aerosol might provide a means of distinguishing the relative importance of biological and industrial inputs.

It is usually assumed that volcanos vent SO₂. This may be true of emissions from fumaroles at comparatively quiescent periods, but no one yet has been able to sample directly the gases coming from a volcano in full voice. This is the important time to sample, when the gases go straight to the stratosphere and in substantial quantity. There are some reasons to believe that the direct emissions are of reduced gases rather than SO₂.

The coexistence of methane (section 3.4.3) at 1.5x10^-6 ppv and oxygen in the current atmosphere is a chemical anomaly so profound that it would reveal to an observer with near certainty the presence of life on this planet. It is the item of evidence that most clearly justifies the new orthodoxy about the contemporary atmosphere: that which takes in the biosphere as an active participant.

The residence time assigned to methane has fluctuated with the estimates of the OH abundance over the range 2 to 20 years. Now at last we seem to have a fairly firm mid-range value of 7 years.

The potential significance of the claim that methane is rising in concentration by 2% per year (Rasmussen and Khali, 1981) demands confirmation. It may indeed be rising in concentration but to prove it requires an absolute accuracy of analysis, including the preparation of standards, of at least 0.5% and a great deal of faith in the constancy of the atmosphere being sampled.

Like methane, N₂O (section 3.5.3) at 0.33 x 10^-6 ppv is a chemical anomaly characteristic of the Earth and a consequence of its biosphere. It is good to know that it is not now regarded as a serious threat to stratospheric ozone.

I have been obliged professionally to follow the rapid unfolding of stratospheric chemistry as observations refined the various hypotheses about the chlorine catalysed depletion of ozone. I disagreed with some of the conclusions aeronomers drew about the consequences of ozone changes on our future health. This did not stop me from being deeply impressed with their professional competence. It is good to know that these powerful talents are now applied to the more important problems of tropospheric chemistry.

The stratospheric experience has revealed how important trace molecular species and radicals can be. Before the chlorine

ozone affair who would have thought of looking for chlorine nitrate or peroxynitric acid in the stratosphere? I know for certain that there are substantial quantities of as yet unidentified substances in the troposphere. It remains to be seen how much they may enlighten our views on the cycles of the elements.

REFERENCES

Zafiriou, O., McFarland, M., and Bromund, R. H. (1980) Nitric oxide in sea water. Science 207, 637-639

	Abstract
	3.1 Introduction
	3.2 Tropospheric Photochemistry
	3.3 Stratospheric Photochemistry
	3.4 The Most Important Carbon Compounds
		3.4.1 Carbon Dioxide
		3.4.2 Carbon Monoxide
		3.4.3 Methane
		3.4.4 Isoprene and Terpenes
	3.5 The Most Important Nitrogen Compounds
		3.5.1 Nitric Oxide, Nitrogen Dioxide and Nitric Acid
		3.5.2 Ammonia
		3.5.3 Nitrous Oxide
	3.6 The Most Important Sulphur Compounds
		3.6.1 Sulphur Dioxide
		3.6.2 The Reduced Sulphur Gases H₂S, (CH₃)₂S, CH₃SH, COS and CS₂
	3.7 Conclusions
	Acknowledgements
	References
	Comment to Chapter 3 J. E. Lovelock
	References


Gas	Direct source/year	Secondary source/year	Removal by	Atmospheric life-	Transport distances
	Source identification	Source identification		times	x, y, z (km);
					volume mixing ratios
					in unpolluted tropo-
					sphere

CO	416 x 10¹⁴g CO	3.7 9.3 x 10¹⁴g CO	30 x 10¹⁴g CO	2 months	4000, 2500, 10
	Biomass burning	methane oxidation	OH		50200 ppbv
	6.4 x 10¹⁴g CO	413 x 10¹⁴g CO	4.5 x 10¹⁴g CO
	Industry	C₅H₈, C_l0H₁₆	Uptake by soils
	0.22 x 10¹⁴g CO	oxidation
	Vegetation

CH₄	0.30.6 x 10¹⁴g CH₄		4 x 10¹⁴g CH₄	7 yrs	Global
	Rice paddy fields		OH		1.5 2.0 ppmv
	0.32.2 x 10¹⁴g CH₄
	Natural wetlands
	0.6 x 10¹⁴g CH₄
	Ruminants
	< 1.5 x 10¹⁴g CH₄
	Termites
	0.31.1 x 10¹⁴g CH₄
	Biomass burning
	0.2 x 10¹⁴g CH₄
	Gas leakage

C₅H8	8.3 x 10¹⁴g C		8.3 x 10¹⁴g C	10 hrs	400, 200, 1
C_l0H₁₆	Trees		OH		010 ppbv


Gas	Direct source/year	Secondary source/year	Removal by	Atmospheric life-	Transport distances
	Source identification	Source identification		times	x, y, z (km);
					volume mixing ratios
					in unpolluted tropo-
					sphere

NO_x	1220 10¹²g N	1.01.5 10¹²g N	2585 10¹²g N	1.5 d	1500,400, 10
(NO+NO₂)	Industry	Oxidation of N₂O	OH		1100 pptv
	1040 10¹²g N		Deposition
	Biomass burning		on soils and
	110 10¹²g N		oceans
	Lightning
	115 10¹²g N
	Soils
	0.15 10¹²g N
	Ocean
	0.25 10¹²g N
	Jet aircraft

HNO₃		1585 10¹²g N	Rain	3 d	3000, 600,1.5
		NO₂ + OH			10-300 pptv

N₂O	1.8 10¹²g N		611 10¹²g N	100200 yrs	global
	Fossil fuel burning		Stratospheric		300 ppbv
	12 10¹²g N		photolysis
	Biomass burning
	12 10¹²g N
	Oceans, estuaries
	13 10¹²g N
	Cultivation natural soils
	< 3 10¹²g N
	Fertilized fields
	?
	Natural soils

NH₃	1020 10¹²g N		Rain	<9d	< 9000, 1000, 3
	Domestic animals				03 ppbv
	26 10¹²g N
	Wild animals
	< 3 10¹²g N
	Fertilized fields
	<30 10¹²gN
	Natural fields
	412 10¹²g N
	Coal burning
	<60 10¹²g N
	Biomass burning


Gas	Direct source/year	Secondary source/year	Removal by	Atmospheric	Transport distances
	Source identification	Source identification		Life-times	x, y, z (km);
					volume mixing ratios
					in unpolluted tropos-
					sphere

SO₂	64 x 10¹²g S	40100 x 10¹²g S	OH	5 d	5000, 700, 2.5
	Coal burning	oxidation H₂S, DMS	Rain		10200 ppty
	26 x 10¹²g S
	Petroleum burning
	11 x 10¹²g S
	Non-ferrous ores
	1030 x 10¹²g S
	Volcanoes

H₂S	< 4 x 10¹²g S		OH	2 d	2000, 500, 1.5
(CH₃)₂S	Agricultural fields				0100 ppty
CH₃SH	3142 x 10¹²g S
	Open ocean
	10 x 10¹²g S
	Coastal waters
	16 x 10¹²g S (?)
	Tropical forests
	24 x 1012g S (?)
	Wetlands

(R32b)	CS₂ + hv		CS₂^*
(R32c)	CS₂^*+ O₂		COS + O