SCOPE 35 - Scales and Global Change

ABSTRACT

The spatial and temporal variability of the atmospheric concentrations of individual gases is determined by the location of their production sources and strongly related to their photochemical lifetimes, which in many cases depends on reactions with hydroxyl radicals. Most chemical compounds are emitted into the atmosphere at very discrete locations. The photochemically most reactive ones are largely processed within the lowest levels (1-2 km) of the troposphere under variable meteorological conditions, resulting in a spatially and temporally very inhomogeneous system that is extremely difficult to model and explore experimentally. Fortunately, the photochemistry of the background free troposphere is much simpler. It is substantially determined by the far less variable compounds O₃, CO and CH₄, which have atmospheric lifetimes of the order of months to years. Here, the greatest difficulty is the correct modelling of the chemical transformations and transport of NO and NO₂ which act as powerful catalysts in the photochemistry of the atmosphere.  As the photochemistry of the background atmosphere is changing substantially due to growths in the atmospheric burdens and emissions of the important gases CH₄, CO, O₃, and NO_x, the development of global photochemical-meteorological models is an important enterprise. It is bound to be an essential ingredient in the future 'Global Change' programme.

Although the development of high-resolution, global photochemical transport models may become feasible some time in the future through increasing computer resources, 'off-line' low resolution models which utilize the essential information extracted from high resolution, large-scale meteorological models probably will be the most useful tools for a long time to come.

INTRODUCTION

The chemical composition of the atmosphere is the result of a great variety of processes and the field of atmospheric chemistry intersects with such broad and diverse scientific disciplines as meteorology, optical physics, chemistry, and biology. It requires high technology for measurements of many ultraminor constituents in an uncontrollable and greatly variable environment.

The most abundant atmospheric gases, molecular nitrogen and oxygen, may be considered to represent the equilibrium state of global biogeochemical processes that have operated on time scales of tens of thousands of years. Because of their very long chemical lifetimes, the volume mixing ratios of these gases are extremely constant in the atmosphere below 100 km altitude. Only above this height, photochemical processes driven by solar ultraviolet radiation become fast enough to affect the distribution of molecular oxygen. For molecular nitrogen photodissociation becomes effective at even higher altitudes. Gravitational separation becomes more important than meteorological mixing above about 100 km.

There is a general rule that the shorter the lifetime, the higher the variability of a trace gas. For this, Junge (1974) has derived a useful approximate empirical equation relating the spatial and temporal, relative standard deviation s ' (the ratio between standard deviation and arithmetic average) of the volume mixing ratio of a chemical constituent to its atmospheric lifetime (t_c/s):

s '= 0.02/ t_c ,

which is especially useful for compounds with average lifetimes of the order of a year or more, because it allows the mean atmospheric residence times of chemical compounds to be roughly estimated from the measured variability of their concentrations. With this information their emission rates into the atmosphere and their integrated atmospheric sinks can be approximately determined.

Other atmospheric chemical compounds are far less abundant than N₂ and O₂. The main reason for this, however, is not much smaller emissions, but much shorter atmospheric lifetimes, leading to much lower abundances. For instance, the net fluxes of CO₂ and O₂ at the earth's surface due to biological (and combustion) processes are roughly equal and opposite in sign. However , the total mass of atmospheric CO₂ is more than a thousand times less than that of O₂. Consequently, human activities affect the distribution of CO₂ much more than that of O₂. For many other chemical compounds the atmosphere is even more strongly affected by human activities.

The shorter-lived and less abundant trace gases are, however, not only important as sensitive indicators of environmental change. Their short atmospheric residence times also indicate, of course, that they are chemically very active. Because of their high reactivity many compounds are only present in the atmosphere at extremely low concentrations. In fact, several of these have so low concentrations that they can never be measured, such as electronically excited atomic oxygen. Many radical species (i.e. molecules with an odd number of electrons) are exceedingly important, because they serve as catalysts in photochemical chain reactions in the atmosphere. The most important among these is hydroxyl (OH). Despite the fact that the average volume mixing ratio of OH in the troposphere is only 3 x 10^-14, it is this radical, and not abundant molecular oxygen, which is responsible for the first critical step in the oxidation and removal of most compounds in the atmosphere. In its turn, because hydroxyl is born from photochemical interactions between ozone and water vapour, ozone too is clearly a very important tropospheric gas.

Most important chemical constituents are emitted into the atmosphere at the earth's surface by natural, mostly biological, processes and increasingly also by anthropogenic activities, whereby the role of man is not only resticted to technological, but also agricultural activities. Generations of humans, all over the world, have been changing the earth's biosphere over millenia, e.g. through forest removal. Tropical agricultural activities produce substantial quantities of important trace gases.

The results of human activities are clearly visible in the atmosphere, not only regionally but also globally. Of substantial importance from the point of view of global atmospheric photochemistry are the changes that are taking place in the abundance of methane, carbon monoxide, and the oxides of nitrogen. It is likely that this is leading to changes in the concentrations of tropospheric ozone and hydroxyl, and the cycling of many compounds in the background atmosphere. As we will discuss, a gradual shift of photochemical activity from clean to more polluted atmospheric environments may be the result of the these changes. It is, therefore, extremely important that long-term measurements are performed to document the changes that are occurring in ozone at various locations in the background troposphere. Field experiments to test atmospheric photochemical chemistry are likewise of major importance. Such efforts can only be carried out fruitfully if appropriate photo-chemical-meteorological models are also developed. This is exceedingly difficult, especially for the lower troposphere, because emission sources of most photochemically active constituents are distributed very heterogeneously at the earth's surface and occur during variable meteorological conditions. Fortunately, the modelling of the photochemistry of the background atmosphere is substantially dependent on the distributions of the more long-lived gases, carbon monoxide, methane, ozone, and water vapour, which are therefore more uniformly distributed and relatively well known from observations. The main difficulty is the appropriate modelling of the highly- variable, short-lived NO_x (NO and NO₂) gases.

In the following we will mainly discuss global scale atmospheric chemical processes, especially how they might be changing because of various anthropogenic activities and how they might be modelled. From the start it should be clear of course that the complexity of atmospheric chemistry and meteorology is so large that any modelling activity can at the best be only a gross approximation of reality. Fortunately, some recent efforts to develop global scale models for medium and long range weather forecasting at some meteorological research centres have become increasingly successful. The extraction of the essential meteorological data from such models for the purpose of photochemical modelling, therefore, presents itself as a promising approach.

PHOTOCHEMISTRY OF THE BACKGROUND TROPOSPHERE

The main actors: OH, O₃, NO_x, CO, and CH₄

Although only about 10% of all atmospheric ozone is located in the troposphere, the lowest 10-17 km of the atmosphere, this small amount of ozone, with volume mixing ratios of 20-100 ppbv (1 ppbv = 10^-9), is nevertheless of fundamental importance for the composition of the earth's atmosphere. The reason for this is the generation of electronically excited atomic oxygen through the absorption of ultraviolet radiation by ozone, and its reaction with water vapour:

O3+hn ®  0(1D)+02(<310 nm)	(R1)
O(1D) + H2O ®  2OH	(R2)

 It is the attack by OH that initiates the oxidation of most trace gases in the atmosphere (Levy, 1971; McConnel1 et al., 1971). This is of fundamental importance, because only a few of the many gases that are produced at the earth's surface can be removed by rainfall and none by reaction with molecular oxygen. According to current estimates, the average volume mixing ratio of hydroxyl radicals in the troposphere is only about 3 x 10^-14 (e.g. Crutzen, 1982). Therefore, although the atmosphere contains almost 21% molecular oxygen, it is the ultraminor constituent OH which starts almost all oxidation processes. In the background troposphere about two thirds of the OH radicals react with CO, and one third with CH₄. Smaller fractions react with various other gases. Both methane and carbon monoxide are increasing in the atmosphere (Rasmussen and Khalil, 1981; Rinsland et al., 1985; Blake et al.. 1982; Bolle et al.. 1985; Rinsland and Levine, 1985). This can clearly influence the background tropospheric concentrations of hydroxyl and, therefore, also those of many other important atmospheric constituents. It is, therefore, clear that understanding of tropospheric chemistry and estimations of the future impact of human activities require detailed knowledge of the photochemical reactions affecting ozone, carbon monoxide, and methane. We will also show how the concentrations of ozone and hydroxyl are strongly affected by catalytic reactions that depend on NO and NO₂.

Although OH reacts overwhelmingly with CO and CH₄ in the background troposphere, these reactions do not necessarily lead to its removal from the atmosphere. They are merely the starting points for various chains of reactions, which may compensate for the initial OH loss and which have important implications for the chemical composition of the troposphere. For instance, in the presence of sufficiently large concentrations of nitric oxide, the oxidation of carbon monoxide will lead to formation of tropospheric ozone, without loss of the catalysts OH, HO₂, NO, and NO₂, through the reaction chain:

In reactions (R4) and (R7), and in the following, the symbol M denotes a third molecule (mostly O₂ and N₂) which serves to stabilize the product which is formed by the association reaction.

A competing chain of reactions, leading to ozone destruction, which dominates in NO-poor environments:

likewise does not lead to the loss of OH and HO₂ radicals. The second reaction sequence is more important than the first one whenever the ratio of the 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 the earth's surface to 100 ppbv at the tropopause, the break-even point between the reaction chains (R3-R7) and (R3, R4, R8) is attained at nitric oxide volume mixing ratios of 4 x 10^-12 (4 pttv) at the ground and 2 pptv at the tropopause. Although these are very low concentrations, they may, nevertheless, not be exceeded in extensive regions off the troposphere in view of the very short residence times of NO and NO₂ due to the rapid formation of highly water-soluble and photochemically quite inactive nitric acid via:

 during daytime and:

followed by the deposition of NO₃ and N₂O₅ on cloud drops and wetted aerosol during night-time. During daytime, the NO₃ which is formed through reaction R11 is immediately photolysed to reproduce the original reactants NO₂ and O₃.

Because of the very short lifetime of NO_x in the troposphere, we may expect appreciable concentrations of NO_x only within, at most, a few weeks travel distance from where they are produced, i.e. mostly near the highly industrialized regions at mid-latitudes in the Northern Hemisphere, and the tropical upper troposphere, where significant amounts of NO are formed by lightning. In agreement with these thoughts. McFarland et al. (1979) have indeed measured background volume mixing ratios of NO of less than 10 pptv in the marine boundary layer of the tropical Pacific. Furthermore, Noxon (1981) has reported a mixing ratio for NO_x of 30 pptv at 3 km altitude at Hawaii.

Besides reacting with NO and O₃ (R5 and R8), HO₂ can also react with itself, especially in NO-poor environments, leading to the production of H₂O₂, which plays an important role in aqueous oxidation chemistry (e.g. Chameides and Davis, 1982; Graedel and Wechsler, 1981):

This reaction sequence leads to the oxidation of carbon monoxide without affecting tropospheric ozone. Because highly water-soluble H₂O₂ can be removed efficiently by uptake in clouds and by precipitation, these processes are important sinks for perhydroxyl (HO₂), and indirectly for OH. Furthermore, H₂O₂ is also involved in the reaction pair: 

which removes OH and HO₂. The oxidation of carbon monoxide in NO-poor environments, therefore, most likely leads to a loss of OH and O₃. In NO-rich environments ozone formation is strongly favoured (Crutzen, 1973).

The oxidation of methane is likewise of very large importance in troposopheric photochemistry. In the first place, about 30-40% of the hydroxyl radicals react with CH₄. Secondly, the oxidation chains of methane strongly affect the atmospheric budgets of hydroxyl and ozone (Crutzen, 1973). Again, nitric oxide plays an important role in determining the oxidation pathways. In NO-rich environments, almost certainly, rapid formation of formaldehyde (CH₂O) occurs, following the initial reaction between OH and CH₄:

Most important, this sequence of reactions leads to the net production of two ozone molecules with various intermediates, in particular NO an NO₂, serving as catalysts.

In NO-poor environments, the CH₄ oxidation steps may follow several pathways that may also lead to CH₂O formation.

However, the photolysis of methylhydroperoxide (CH₃O₂H) is slow, resulting in a relatively long mean atmospheric residence time for this compound of about a week. Consequently, methylhydroperoxide may be removed from the atmosphere by rainfall, or deposition on cloud droplets, aerosol particles, or on the earth's surface. Wherever this occurs, the oxidation of CH₄ will have led to a loss of two odd hydrogen radicals (OH and HO₂).

Another reaction mechanism which occurs in NO-poor environments and which leads to the net loss of OH and HO₂, is

Finally, the catalytic pair of reactions

is particularly important, because it destroys both OH and HO₂.

The oxidation pathways of methane in NO-poor environments are, therefore, far less predictable than in NO-rich environments and do not always lead to production of CH₂O. It is, however, likely that hydroxyl and per hydroxyl radicals are being destroyed rather efficiently.

Three pathways lead to the oxidation of formaldehyde to carbon monoxide with the overall net production of 0.8 HO₂ radicals per CH₂O molecule:

Because the average photochemical lifetime of CH₂O in the troposphere is only a few hours, the likelihood of precipitation scavenging of formaldehyde is rather low and formation of carbon monoxide efficient.

Changes in background atmospheric photochemistry due to human activities

As HO₂ is rapidly converted into OH by reactions R5 or R8, leading to net gain, or net loss of ozone, respectively, the oxidation sequences leading from CH₄ to CO, and from there to CO₂, have a strong effect on the abundance of both hydroxyl and ozone concentrations in the background troposphere.

Depending on the ambient concentrations of NO, the oxidation of one molecule of methane to one molecule of carbon dioxide yields the following astonishing net results:

(a) in NO-poor environments, i.e. especially in the unpolluted background troposphere: a net loss of 2-3.5 hydroxyl radicals and 0-1.7 ozone molecules.

(b) in NO-rich environments, i.e. especially in the highly industrialized, middle and high latitude regions of the northern hemisphere: net gain of 0.5 OH radicals and 3.7 ozone molecules.

The amount of methane in the atmosphere has been increasing for considerable time, in agreement with the growth of various human activities, which are summarized in Table 6.1. Long term observations during the past three decades indicate an average yearly increase by about 1.1% (Rasmussen and Khalil, 1981; Rinsland et al., 1985; Blake et al., 1982; Bolle et al., 1985). Furthermore, analyses of air trapped in ice cores have shown that the atmospheric methane content before 1650 was 2-3 times lower than at present (Rasmussen and Khalil, 1984; Craig and Chou, 1982). These global increases in methane have, therefore, probably caused substantial increases in ozone concentrations in NO-rich environments and decreases in hydroxyl concentrations in NO-poor environments. Over the past decades, the production of ozone at Northern hemisphere mid-latitudes has been strongly enhanced by concomittant increases in NO emissions from industrial activities (Table 6.2), and by the increases in carbon monoxide concentrations in the free troposphere by about 2% per year that have been deduced from solar photographic spectra taken on the Jungfraujoch in Switzerland during the past 30 years (Rinsland and Levine, 1985). Elsewhere, no long-term CO trends have yet been analysed, although air trapped in ice cores may contain relevant information. The observed rise in CO may have led to a decrease of OH also at mid-latitudes in the Northern Hemisphere through reaction R3. However, enhanced ozone and nitric oxide formation and the effects of reaction R5 may have counteracted this.

In the NO-poor atmospheric environments, the observed increase in background CH₄ concentrations have, however, most likely led to lower hydroxyl concentrations. Also in these regions, in agreement with photochemical theory, increases in carbon monoxide concentrations may have occurred, although no data are available. Through reaction R3, this would have led to a further reduction in the hydroxyl concentrations, causing a strong positive feedback. In fact, this raises questions about the stability of the chemistry of the background troposphere. Altogether, we conclude that, as time goes on, lesser quantities of industrial and natural gases may become oxidized in the tropics by reactions with hydroxyl, leading to a build-up of various important trace gases in the troposphere. Gradually, these gases are increasingly oxidized in the temperate latitude zones of the Northern Hemisphere, especially from April to October, leading to enhanced production of tropospheric ozone in these regions, which obtain much NO from fossil fuel combustion processes. Model calculations indicate the possibility of a doubling of average ozone concentrations in the lower background troposphere at mid-latitudes since the  start of the industrial period (Crutzen and Gidel, 1983). There are indeed  several indications of upward trends in tropospheric ozone. Angell and Korshover (1983) report increases in ozone concentrations by 20%  in the middle troposphere (2-8 km) at north temperate latitudes over the time period 1967-1979. Similar increases in free tropospheric ozone at Hohenpeissenberg in Southern Germany have also been reported by Attmannspacher et al. (1984). Observations of surface ozone at the clean air stations of Mauna Loa, Hawaii and Point Barrow, Alaska, may likewise suggest upward trends by »1% per year over the past decade (Harris and Nickerson, 1984). Several more examples are given by Logan (1985).

Table 6.1 Tropospheric sources of CO, CH₄ and C₅H₈ + C₁₀H₁₆, their average residence times, typical transport distances, and atmospheric concentration ranges. Derived from information in Crutzen et al. (1986), Crutzen (1983), Bolle et al. (1985), Seiler (1974), Zimmerman et al. (1978), and Bingemer and Crutzen (1986). Another significant source of methane may come from northern peatlands (Harriss et al., 1985). Note that most emission estimates are uncertain by at least a factor of two

Table 6.2 Tropospheric sources, average residence times, typical transport distances, and atmospheric concentrations ranges of NO_x. Derived from information in Crutzen (1983), Galbally and Roy (1978), and Borucki and Chameides (1984). Note the large uncertainties in estimates which must be resolved by future research

Geographical distribution of sources and sinks of CO and CH₄

So far only very few successful hyroxyl concentration measurements have been made in the troposphere, so that much of the current theory of atmospheric photochemistry is untested by observations. Indirectly, this can, however, be achieved by comparison of calculated and observed concentrations of trace gases that react with hydroxyl. Using photochemical-meteorological models, it is possible to estimate the global distribution of hydroxyl radicals in the background troposphere. These calculations require knowledge of the global distributions of O₃, H₂O, CO and CH₄, which are quite well known from observations. The calculated OH distributions are, however, also strongly determined by the global NO and NO₂ distributions, which are expected to be very variable of which very few observations have so far been made in the 'background' troposphere. Fortunately, the atmospheric sources of NO are much better known (see Table 6.2), so that these can be introduced at the appropriate locations in models. Photochemical reactions and atmospheric transport determine the distribution of NO and NO₂. The global distributions of NO and NO₂ (and associated products), and of OH can, therefore, in principle be calculated. An example of a possible distribution of OH concentrations, calculated with a two-dimensional (latitude and height dependent) model, is shown in Figure 6.1. We note that the highest OH concentrations appear in the tropics, which is to be expected, because it is there that the flux of solar ultraviolet radiation peaks. From these calculations, the average concentration of hydroxyl in the troposphere is estimated to be above 6 x 10⁵ molecules cm^-3. Few reliable direct observations of hydroxyl radical concentration are yet available, although efforts to improve the situation are under way. The calculated OH concentration distribution shown in Figure 6.1 is, however, roughly consistent with the global observations of methylchloroform (CH₃CCl₃), which is removed from the atmosphere by reactions with OH and which has no other sources than the rather well known industrial emissions at mid-latitudes (Crutzen, 1982; Crutzen and Gidel, 1983: Zimmermann, 1984). Methylchloroform is, therefore, a particularly suitable gas to validate tropospheric photochemical-meteorological models. Comparison between observations and model calculations is facilitated by the fact that the atmospheric lifetime of CH₃CCl₃ is about 10 years, so that its distribution is mainly latitude depedent (Prinn et a/., 1983). Of critical importance for a successful simulation of the methylchloroform distribution are the calculated hydroxyl concentrations in the tropics, and the exchange of air masses between the Northern to the Southern Hemisphere. Judging from comparisons between calculated and observed distributions of CFCl₃ and CF₂Cl₂, interhemispheric transport is likewise described very well (Zimmermann, 1984; Crutzen and Gidel, 1983). Both gases are produced in rather well known quantities by industry, but in contrast to CH₃CCl₃ they do not react with OH.

Figure 6.1 Calculated mean annual distribution of average daytime hydroxyl concentrations (units: 10₆ molecules/cm³). Note the occurrence of maximum concentrations in the tropics. This is due to a minimum in total stratospheric ozone in equatorial regions, which allows substantial penetration of photochemically active solar ultraviolet radiation (Crutzen, 1982. Reproduced by permission of S. Bernhard)

We must caution that the apparent success of the adopted model to simulate the global distributions of CH₃CCl₃, CFCl₃, and CF₂Cl₃ may have been quite fortuitous. As we will note in the following section, the design of an appropriate photochemical-meteorological model is an extremely difficult task, requiring the modelling of many processes that occur on space and time scales that are far smaller than can be resolved by models. The transport of NO_x is particularly affected by this, so that calculated NO_x and OH distributions should be very model dependent. Nevertheless, it is likely that the solution of Figure 6.1 contains the main features of the global OH distribution in the background troposphere and roughly the right numerical values. This information is extremely useful, because it allows estimation, for example, of the global distribution of the sources and sinks of CH₄ and CO, which are removed from the atmosphere by reaction with OH. The results are shown in Figures 6.2 and 6.3. It appears that the tropical regions do contribute strongly to the destruction and production of CO and CH₄. The most important tropical sources for carbon monoxide are biomass burning and the oxidation of hydrocarbons, especially isoprene (C₅H₈) and terpenes (C₁₀H₁₆), that are emitted from tropical vegetation, as indicated in Figure 6.4. Most biomass burning activities are related to tropical agriculture (forest clearing, shifting agriculture, agricultural waste burning). Most methane is probably formed by the decay of organic matter in the anaerobic sediments of natural marshlands and rice fields. The estimated source strengths of the various processes that contribute to the production of CH₄ and CO are shown in Table 6.1. Most of these are still very uncertain. Much research is necessary to establish them with greater certainty.

Figure 6.2 Calculated and estimated sources and sinks of methane in 10¹² g per year (Crutzen and Gidel, 1983). Note the importance of the tropics. The various sources of CH₄ are identified in Table 1. OH = destruction by OH (320 x 10¹² g); M = transport from northern to southern hemisphere (70 x 10¹² g); G = annual growth (90 x 10¹² g); E = total source (410 x 10¹² g)

Figure 6.3 Calculated and estimated sources and sinks of carbon monoxide in 10¹² g per year (Crutzen and Gidel, 1983). Note the importance of the tropics. The various sources of CO are identified in Table 6.1. OH = Destruction by OH radicals (2054 x 10¹² g); D = destruction by microbial action in soils (640 x 10¹² g); I = fossil fuel burning (640 X 20¹² g). CH₄ ®CO = oxidation of CH₄ to CO (570 x 10¹²); T = tropical sources of CO (1480 x 10¹² g), The transport of CO from the northern to the southern hemisphere is about equal to 190 x 10¹² g

Figure 6.4 Measured average CO volume mixing ratios over the tropical forest (jungle) and savanna (cerrado) regions of Brazil during the dry season (8 vertical profiles over cerrado, 10 profiles over the forests). The concentrations measured in the lower troposphere are as high as in polluted industrial regions in the northern hemisphere (Crutzen et al, 1985. Reproduced by permission of D. Reidel Publishing Company)

Sources and sinks of background tropospheric ozone

The term in the tropospheric ozone budget that can be estimated with the greatest confidence is the global ozone loss by the reactions (R1) and (R2), which on average amounts to about 8 x 1010 molecules cm^-2s^-1. The ozone loss rate at the ground is of similar magnitude (Galbally and Roy, 1980). The estimated downward transport of stratospheric ozone by meteorological processes is about equal to 6 x 10¹⁰ molecules cm^-2s^-1 (e.g. Galbally and Roy, 1980; Gidel and Shapiro, 1980; Fabian and Pruchniewicz (1977). It is, therefore, clear that the potential production of significant amounts of ozone in the troposphere must also be considered. As discussed earlier, catalytic reactions that result in ozone production in the background troposphere take place during the oxidation of CH₄ and CO in environments with sufficient concentrations of NO. According to the estimates presented in Figures 6.2. and 6.3, the average tropospheric destruction rates of CO and CH₄ by reaction with OH are about equal to 3 x 10¹¹ molecules cm^-2s^-1 and 8 x 10¹⁰ molecules cm^-2s^-1, respectively. If all oxidation of CO and CH₄ would occur in NO-rich environments, yielding one ozone molecule for each carbon monoxide, and 2.7 ozone molecules for each methane molecule oxidized, the average, global net ozone production rate would be equal to about 5 x 10¹¹ moelcules cm^-2s^-l. This is far larger than the ozone that can be destroyed by reactions R1 and R2 and by destruction at the earth's surface. Consequently, a substantial fraction of the background troposphere must contain so little NO_x that ozone is not produced but destroyed. Altogether, we may estimate that about half of the oxidation of CH₄ and CO in the background free troposphere occurs in NO-poor, the other half in NO-rich, environments. As a substantial fraction of NO is produced by anthropogenic activities (see Table 6.2), it is likely that substantial increases in tropospheric ozone may have occurred during the industrial era at northern mid-latitudes.

The potential for tropospheric ozone production

An approximate upper limit to the ozone production rates that could occur in the troposphere can be derived by assuming that there is always and everywhere enough NO in the troposphere, so that, as with methane, there are 2-3 ozone molecules produced per hydrocarbon molecule which is oxidized to carbon monoxide (Crutzen, 1973). Because most of these hydrocarbons, especially isoprene, are produced by tropical vegetation and very short-lived, their oxidation to carbon monoxide occurs mainly in lowest 2-3 km above the tropical forests, as is clearly shown in Figure 6.4. Making use of the information on the sources of CO, and non-methane hydrocarbons (Table 6.1), an average global, tropospheric ozone production of more than 10¹² molecules cm^-2s^-1 would be possible. As we have seen, the actual tropospheric ozone production rate could only be at most 10% of this amount. This indicates that most oxidation of the reactive hydrocarbons emitted from tropical vegetation must take place without the production of ozone. One may therefore guess that most air masses above the tropical forests contains subcritical NO concentrations. At the same time it is, however, also clear that future expansions of industrial and biomass burning activities in the tropics (and elsewhere), leading to larger NO_x concentrations, could cause large increases in tropospheric ozone, especially in the boundary layer above the tropical forests in which most of the oxidation of isoprene and terpenes to CO occurs. An example of ozone production in polluted tropical air masses is presented in Figure 6.5. In the this case, the necessary nitric oxide to cadse the photochemical production of ozone was supplied by the burning of biomass in the savanna regions of Brazil during the dry season.

Figure 6.5 Average profiles of ozone volume mixing ratios over the tropical humid forest (selvas) and savanna (cerrado) regions of Brazil during the dry season (8 vertical profiles over cerrado, 10 profiles over the forests). Note the much higher ozone concentrations over the savanna regions. This is due to the photochemical smog created by biomass burning effluents, especially NO (Crutzen et al, 1985. Reproduced by permission of D. Reidel Publishing Company)

THE MODELLING OF ATMOSPHERIC-CHEMICAL PROCESSES

The problem of spatial and temporal variability

The distribution of chemical compounds in the atmosphere is determined by complex photochemical and meteorological processes that occur over a wide range of time and space scales. The compounds that are produced at the earth's surface by biological and by anthropogenic processes are from the start not evenly distributed in the atmosphere. For example, the oxides of nitrogen NO and NO₂ are mostly produced from such discrete sources as power plants, highways, or industrial centres and cities, which together cover at most about one percent of the continents of the earth. As the average atmospheric lifetime of NO_x is only a few days, it is clear that atmospheric motions can not disperse these compounds evenly through a major portion of the atmosphere. The same holds for many other short-lived compounds, such as the reactive hydrocarbons which are released to the atmosphere mostly by forests, but also by industrial processes. As the sources of many short-lived, photochemically active atmospheric constituents are so discrete and often do not coincide in space, efforts to model the atmospheric chemical systems are faced with the immense problem to resolve or approximate the temporal and spatial discreteness of the distributions and sources of short-lived, reactive chemical compounds. They are also very variable in space and time, so that the overlap of chemical plumes can be highly variable, depending on the particular meteorological situation. The discontinuity of emission patterns and photochemical interactions between various classes of compounds is, therefore, a major obstacle in the design of appropriate photochemical-meteorological models that describe the transport and photochemistry of reactive compounds that are released to the atmosphere. It is, therefore, not surprising that up to now no high-resolution models have been developed that take care of these difficulties. In fact, the question may be asked as to what type of models are most appropriate to describe the highly space and time variable photochemistry that occurs in regions of intensive releases of trace constituents to the atmosphere. A good representation of these processes is also important for the correct description of the photochemistry of the background troposphere because this depends critically on the ambient concentrations of short-lived NO_x, which is produced only over a very small fraction of the earth's surface.

Coupling between boundary layer and free tropospheric chemistry

As discussed in section 2, 1he photochemistry of the background troposphere mainly involves water vapour and the other chemical constituents O₃, CO, CH₄, NO_x, and their photochemical reaction products, such as OH and HO₂. Fortunately, the spatial and temporal variability, especially of CH₄, CO and O₃, is much less than for most other gases that are primarily released from the earth's surface. The explanation for this is that CH₄, CO and O₃ have relatively long tropospheric lifetimes, ranging from months (CO and O₃) to more than ten years (CH₄). The main problem in photochemical modelling of the background troposphere is, therefore, caused by the high variability in the distribution of the NO_x gases and the randomness of the removal of highly water-soluble compounds, especially HNO₃, H₂O₂, and CH₃O₂H by precipitation processes. The latter requires the appropriate simulation of the effects of cloud processes. However, clouds not only are responsible for the removal of highly water-soluble compounds from the troposphere, but may also efficiently transport less water-soluble gases rapidly from the lower to the middle and upper troposphere. Most gases that are emitted at the earth's surface by natural or anthropogenic processes are insufficiently water-soluble to be affected by precipitation removal. This is e.g. also the case for NO and NO₂. For all those gases, efficient upward transport may occur in cloudy regions, especially in convective clouds or along frontal slopes. The more short-lived the gas, the more important is this transport.

To demonstrate the importance of the proper meteorological modelling of cloud transport processes, we compare in Figure 6.6 the calculated vertical distributions of SO₂ in the tropical marine atmosphere that are obtained with two very different types of models. In this part of the atmosphere, the release of dimethylsulphide (DMS) from the ocean and its oxidation, following reaction with hydroxyl, is the main source of SO₂ (Andreae and Raemdonck, 1983; Barnard et al., 1982; Cox and Sheppard, 1980; Lovelock et al., 1974). A typical conversion time from DMS to SO₂ is less than one day. The two profiles that show a rapid decrease in volume mixing ratios with altitude are calculated with so-called eddy-diffusion models, which until recently have been used most commonly for global atmospheric chemistry modelling, especially in the stratosphere. In analogy with molecular diffusion theory, in these models the net vertical flux of a gas is set proportional to the vertical gradient of its volume mixing ratio. As we note, with these assumptions, the calculated volume mixing ratios of SO₂ in the middle and upper troposphere are much less than those observed (Maroulis el al., 1980). The reason for this dissatisfactory behaviour is that the eddy diffusion formulation cannot resolve the effect of brief episodes of vigorous vertical exchange.
                         

Figure 6.6 Calculated distributions of average distributions of SO₂ in the tropical oceanic troposphere. The curves marked L and R are typical for results obtained with eddy diffusion models. A detailed meteorological model of the tropical circulations which includes rapid cloud transport, the so-called 'Staubsauger' model (Chatfield and Crutzen, 1984) gives totally different results that agree much better with the range of observations that are indicated with horizontal bars (Maroulis et al, 1980) 

Figure 6.7 A picture of the cycling of the most important sulphur compounds, dimethyl-sulphide (DMS), sulphur dioxide, and particulate sulphate, in the tropical marine atmosphere. Violent upward motions in thunderstorm cells rapidly deposit short-lived DMS to the tropical middle and upper troposphere, where it is oxidized to SO₂, which in turn is oxidized to H₂SO₄ and condensed to sulphate aerosol. Other sulphur compounds (SX) may also be involved. An important role in these oxidation processes is played by hydroxyl during daytime and by NO, during night-time. The latter is formed from the oxidation of NO which is produced by lightning (Chatfield and Crutzen, 1984)

On the other hand, very different results, which agree much better with the observations, are obtained with the so-called 'Staubsauger' (the German word for 'vacuum cleaner') model, in which the meteorological exchange processes that are typical for the tropical marine atmosphere are much better resolved in time and space (Chatfield and Crutzen, 1984). In this formulation the tropical troposphere is divided into five regions, ranging from a very turbulent central region with very strong upward motions occurring in intense thunderstorm cloud clusters over only a few percent of the tropics, to a stable outer region, where downwind motions dominate. These regions correspond most closely to the intertropical convergence zone and the horse latitudes. Zonally and temporally averaged, they constitute the well-known Hadley circulation of the tropics. The intense upward motions in the thunderstorms of the center region are fed by strong trade winds, which blow equator wards in the lowest 1-2 km of the marine atmosphere. In the upper troposphere and the outer subsidence regions the wind directions are in the opposite directions, producing a closed cycle. In this way tropical boundary layer air can be swept up to the middle and upper troposphere in less than an hour, too brief for photochemical processes to play a significant role. Figure 6.7 shows schematically how SO₂ formation and transport occur under such circumstances.

Other candidates for the 'Staubsauger' action are, for example, the reactive hydrocarbons that are produced by tropical vegetation. This is especially interesting, because tropical forests are located in regions with much rainfall which is caused by strong upward motions and intense thunderstorm activity. In these air masses a maximum in lightning activity and substantial production of nitric oxide may also be expected. The potential consequences of this for tropical air chemistry have so far not been explored by atmospheric observations.

High resolution, photochemical-meteorological models

For the troposphere it is clearly important to develop three-dimensional models which take into account the intricacies of the many scale's of atmospheric motions. Such models have been developed in the meteorological community for weather predictions and for climate studies and have been rather successful in describing (parameterizing) rather discrete meteorological processes, e.g. moist convection. On the other hand, such models are not faced with anything like the spatial discreteness of the sources of the chemical constituents and their variable lifetimes. Several efforts are indeed under way to introduce chemistry in elaborate meteorological models. For example, ambitious efforts take place at the US National Center of Atmospheric Research (NCAR, 1985) to introduce photochemistry in an existing mesoscale meteorological model (originally developed by Anthes and Warner, 1978) with parameterized subgrid scale cloud processes and detailed simulation of the boundary layer processes, both of which are of critical importance for a correct description of the long range transport of atmospheric pollutants. The vertical extent of the model covers the troposphere in 15 layers, while the horizontal domain may be chosen to cover the Western US or parts thereof. In fact, the model is generalized so that it can be adapted to simulate air pollution problems at any other location. For a typical version of the model, covering the Western US, the minimum horizontal grid is equal to 60 km x 60 km, which allows for up to a week simulation of pollution episodes with the most advanced computers that are currently available. A simplified reaction scheme describes the photochemistry of O₃, various classes of hydrocarbons, NO_x, SO₂, and their most important reaction products. The principal aim of the effort is to develop a regional 'acid deposition' model, which is suitable to assess source-receptor relationships and transport across national borders (NCAR, 1985). No doubt, the development of this model is a very interesting effort, which for the first time brings together expertise in such diverse atmospheric science subdisciplines as boundary layer and mesoscale meteorology, cloud physics, and atmospheric chemistry. The task is clearly enormous and it will be especially difficult to test both the photochemical and meteorological subsections of the models sufficiently so that undisputed advice can be given regarding choices between regulatory measures vis-á-vis a multitude of pollutant gases, such as SO₂, NO_x and hydrocarbons that often emanate from different industrial activities and nations. So far, in connection with the 'acid rain' problem, most regulations have been concerned with the emissions of SO₂. No doubt, NO_x and hydrocarbons will come increasingly in the forefront of interest. It should be noted here that for these gases not all emissions are anthropogenic. Nitric oxide is also produced by lightning discharges in quantities that are still very uncertain. Of even greater importance compared to anthropogenic inputs may be the emissions of reactive hydrocarbons, such as isoprene and terpenes, from forests (Zimmerman, 1977).

Finally we may note that even a rather tight grid spacing of 60 km x 60 km can of course not resolve the extreme heterogeneity of emission patterns at the earth's surface, because only the largest metropolitan areas occupy a major fraction of such a grid box. To resolve this problem, higher resolution, so-called nested-grid models, covering important subsections of the larger domain may be designed. Still, even such models cannot resolve the enormously complex patterns of emissions, ranging from point to city size sources.

Low-resolution, large scale models

From the previous discussion it is clear that the design of even a limited, subcontinental-size chemical-meteorological model constitutes an utterly complicated task which requires major team work and computer  resources.

The future of this kind of mesoscale, atmospheric modelling is hard to predict. One obvious way to proceed is to build on ever increasing computer power and to go for higher resolution. However, I believe that there is considerable room for alternative approaches, such as the development of models that allow for appreciable stochastic behaviour. In fact, the development of such models may be the only hope of arriving at a comprehensive global model of atmospheric chemistry for the foreseeable future.

Efforts which have been carried out so far in developing global scale models have been restricted to the modelling of some long-lived gases such as CFCl₃, CF₂Cl₂, and N₂O, mainly in order to test the dynamic behaviour of the three-dimensional model (Zimmermann, 1984). Although the development of more elaborate models might appear immature, simplified approaches to global scale modelling may, nevertheless, prove useful. For instance, in the previous section we have shown results from a two-dimensional (latitude, height dependent) model with which an average global OH distribution was calculated. Somewhat fortuitously, the calculated OH distribution was verified rather satisfactorily by the globally observed CH₃CCl₃ distribution. This OH distribution was subsequently used to derive important information regarding the global distribution of sources and sinks of methane and carbon monoxide. The future extension of such a model to three dimensions may, therefore, prove fruitful for some atmospheric-chemical applications, allowing, for example, the effects of continental emissions to be separated from those over marine areas. However, in such models it will be necessary to include even rougher parameterizations of the exchange processes between the boundary layer and the middle and upper troposphere than in mesoscale models. Fortunately, the requirements for global models are different than those for mesoscale models, because one is never interested the know the detailed distribution of chemical compounds around the globe at a particular time, but merely in the average distribution, and in variations there off .Most promising in the future development of three-dimensional models may, therefore, be the introduction of the stochastic simulation of boundary layer and convective processes, which may be derived from a thorough, statistical analysis of the results obtained by detailed large scale, meteorological models. Considering the strongly increasing capabilities of such global scale models in predicting global weather over periods approaching a week, there is considerable hope that the necessary parameters for such stochastic modelling efforts may be derived, so that global modelling with such large grid spacings as, for example, 500 km x 500 km can prove useful. In practice, the meteorological models are supplied with up-to-date information every six hours, so that appropriate statistics should be extractable for photochemical applications.

CONCLUSIONS

Atmospheric chemistry is the interplay between a large variety of environmental processes, such as industrial and agricultural activities, microbial activities in soils and waters, all sorts of meteorological phenomena, and the photochemistry of many trace gases.

Substantial atmospheric composition changes are occurring which affect the overall functioning of tropospheric chemistry. According to current photochemical theory, a gradual build-up of various trace gases that react with hydroxyl should be occurring in the global atmosphere. Because the observed increasing atmospheric concentrations of methane in the background atmosphere probably induce lower hydroxyl concentrations, these in turn would cause higher concentrations of methane and carbon monoxide, implying a potentially strong positive feedback. In fact, the current status of knowledge of atmospheric photochemistry allows a legitimate question about the overall stability of the chemical composition of the background atmosphere. Furthermore, as a result of the processes that are discussed in this paper, a gradual shift of the oxidizing power from the clean background atmosphere to more polluted regions can be postulated, leading especially to enhanced ozone concentrations at mid-latitudes. It is clear that these largely theoretical thoughts should be urgently tested by chemical observations in the background atmosphere and by the high-quality monitoring of various critical compounds at a sufficient number of stations around globe. At the same time, better estimates must be made of the sources of trace consitutents at the earth's surface. Biological emissions are by far most difficult to estimate.

The design of appropriate models is very important in order to study the combined effect of the many environmental factors that contribute to atmospheric chemistry. The development of such photochemical models is made exceedingly difficult by the great variability in the atmospheric concentrations of the most reactive gases, which are caused by the strong temporal and spatial heteorogeneity of the meteorological processes and the distribution of the sources of trace gases at the earth's surface. Fortunately, rather successful models of the large-scale meteorological processes have been developed over the past decade especially for the purpose of medium and long range weather predictions. The extraction of the essential meteorological data from such models for atmospheric chemistry applications, therefore, presents itself as most promising venue. Because for global tropospheric chemistry modelling, one is not interested in the three-dimensional chemical distribution at any particular time, but more on average distributions and variability, the development of global models containing a substantial element of stochasticity seems most appropriate.

REFERENCES

Andreae, M. O., and Raemdonck, H. (1983). Dimethylsulfide in the surface ocean and the marine atmosphere: a global view. Science, 221, 744-745.

Anthes, R. A., and Warner, T. T. (1978). Development of hydrodynamic models suitable for air pollution and other mesometeorological studies. Mon. Wea. Rev., 106, 1045-1078.

Angell, J. K. , and Korshover, J. (1983). Global variation in total ozone and layer mean ozone: An update through 1981. J. Climate Appl. Meteorol., 22, 1611-1626.

Attmannspacher, W ., Hartmannsgruber, R., and Lang, P. (1984). Langzeittendenzen des Ozons der Atmosphäre aufgrund der 1967 begonnenen OzonmeBreihen am Meteorolgischen Observatorium HohenpeiBenberg. Meteorol. Rdsch., 37, 193-199.

Barnard, W. R., Andreae, M. O., Watkins, W. E., Bingemer, H., and Georgii, H. W. (1982). The flux of dimethyl sulfide from the oceans to the atmosphere. J. Geophys. Res., 87, 8787-8794.

Bingemer, H. and Crutzen, P. J. (1986). The production of methane from solid wastes. J. Geophys. Res., 92 (D2), 2181-2187.

Blake, D. R., Mayer, E. W., Tyler, S. C. Montague, D. C., Makide, Y., and Rowland, F. S. (1982). Global increase in atmospheric methane concentration between 1978 and 1980. Geophys. Res. Lett, 9, 477-480.

Bolle, H. J., Seiler, W., and Bolin, B. (1985). Other greenhouse gases and aerosols. Assessing their role for atmospheric radiative transfer. In: WMO/ICSU/UNEP International Assessment of the Role of Carbon Dioxide and Other Radiatively Active Constituents in Climate Variation and Associated Impacts.

Borucki, W. J., and Chameides, W. L. (1984). Lightning: Estimates of the rates of energy dissipation of nitrogen fixation. Rev. Geophys. Space Phys., 22, 363-372.

Chameides, W. L., and Davis, D. D. (1982). The free radical chemistry of cloud droplets and its impact upon the composition of rain. J. Geopyhs. Res., 87 , 4863-4877.

Chatfield, R. B., and Crutzen, P. J. (1984). Sulfur dioxide in remote oceanic air: cloud transport of reactive precursors. J. Geophys. Res., 89(D5), 7111-7132.

Cox, R. A., and Sheppard, D. (1980). Reactions of OH radicals with gaseous sulphur compounds. Nature, 284, 330-331.

Craig, H., and Chou, C. C. (1982). Methane: The record in polar ice cores. Geophys. Res. Lett., 9, 1221-1224.

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

Crutzen, P. J. (1982). The global distribution of hydroxyl. In Goldberg, E. D. (Ed.) Atmospheric Chemistry, pp. 313-328. Dahlem Konferenzen, Springer-Verlag, Berlin.

Crutzen, P. J. (1983). Atmospheric interactions-homogeneous gas reactions of C, N, and S containing compounds. In Bolin, B., and Cook, R. B. (Eds.) The Major Biogeochemical Cycles and Their Interactions, SCOPE 24, pp. 68-112. John Wiley and Sons, New York.

Crutzen, P. J., and Gidel, L. T. (1983). A two-dimensional photochemical model of the atmosphere. 2: The tropospheric budgets of the anthropogenic chlorocarbons, CO, CH₄, CH₃CI and the effect of various NO. sources on tropospheric ozone. J. Geophys. Res, 88, 6641-6661. .

Crutzen, P. J., Delany, A. C., Greenberg, J., Haagenson, P., Heidt, L., Lueb, R., Pollock, W., Seiler, W., Wartburg, A., and Zimmerman, P. (1985). Tropospheric chemical composition measurements in Brazil during the dry season. J. Atmos. Chem., 2, 233-256.

Crutzen, P. J., Aselmann, I., and Seiler, W. (1986). Methane production by domestic animals, wild ruminants, other herbivorous fauna, and humans. Tellus, 38B, 271-284.

Fabian, P., and P. G. Pruchniewicz (1977). Meridional distribution of ozone in the troposphere and its seasonal variations. J. Geophys. Res., 82, 2063-2073.

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

Galbally, I. E., and Roy, C. R. (1980). Destruction of ozone of the earth's surface. Q. J. R. Meteorol. Soc., 106, 599-620.

Gidel, L. T., and Shapiro, M. (1980). General circulation model of the net vertical flux of ozone in the lower stratosphere and the implications for the tropospheric ozone budget. J. Geophys. Res., 85, 4049-4058.

Graedel, T. E., and Wechsler, C. J. (1981). Chemistry in aqueous atmospheric aerosols and raindrops. Rev. Geophys. Space Phys., 19, 505-539.

Harris, J. M., and Nickerson, E. C. (Eds.) (1984). Geophysical Monitoring for Climate Change No 12. Summary Report 1983, NOAA, Boulder, Colorado, 184 pp.

Harriss, R. C., Gorham, E., Sebacher, D. I., Bartlett, K. B., and Flebbe, P. A. (1985). Methane flux from northern peatlands. Nature, 315, 652-653.

Junge, C. E. (1974). Residence time and variability of tropospheric trace gases. Tellus, 26, 477-488.

Levy, H., II (1971). Normal atmosphere: Large radical and formaldehyde concentrations predicted Science, 173, 141-143.

Levy, H., II. (1974). Photochemistry of the troposphere. Advances in Photochemistry, 370-524.
Logan, J. (1985). Tropospheric ozone: seasonal behavior, trends and anthropogenic influence J. Geophys. Res., 90 (D6), 10463-10482.

Lovelock, J. E., Maggs, R. J., and Rasmussen, R. A. (1974). Atmospheric dim ethyl sulfide and the natural sulphur cycle. Nature, 248, 625-626.

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

McConnell, J. C., McElroy, M. B., and Wofsy, S. C. (1971). Natural sources of atmospheric CO. Nature, 233, 187-188.

McFarland, M., Kley, D., Drummond, J. W., Schmeltekopf, A. L., and Winkler, R. H. (1979). Nitric oxide measurements in the equatorial Pacific. Geophys. Res. Let., 6, 605-608.

NCAR. (1985). The NCAR Eulerian Regional Acid Deposition Model. Technical Note NCAR/TN-256 + STR, National Center for Atmospheric Research, Boulder, CO 80307, USA: 178 pages.

Noxon, J. F. (1981). NO. in the mid-Pacific troposphere, Geophys. Res. Lett., 7, 1223-1226.

Prinn, R. G., Rasmussen, R. A., Simmonds, P. G., Alyea, F. N., Cunnold, D. M., Lane, B. C., Cardelino, C. A., and Crawford, A. J. (1983). The atmospheric lifetime experiment. 5: Results for CH₃CCl₃ based on three years of data. J. Geophys. Res., 88, 8415-8426. 

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

 Rasmussen, A. N., and Khalil, M. A. K. (1984). Atmospheric methane in the recent and ancient atmospheres: concentrations, trends and interhemispheric gradient. J. Geophys. Res., 89 (D7), 11599-11605.

Rinsland, C. P., Levine, J. S., and Miles, T. (1985). Concentration of methane in the troposphere deduced from 1951 infrared solar spectra. Nature, 318, 245-249.

Rinsland, C. P., and Levine, J. S. (1985). Free tropospheric carbon monoxide concentrations in 1950 and 1951 deduced from infrared total column amount measurements. Nature, 318, 250-254.

Seiler, W. (1974). The cycle of atmospheric CO. Tellus, 26, 117-135.

Stockwell, W. R. (1985). A Homogeneous gas phase mechanism for use in a regional acid deposition model. Atmos. Environ. (submitted).

Zimmerman, P. R. (1977). Tampa Bay photochemical oxidant study, Rep. EPA 904/9-77-028, Environmental Protection Agency, Research Triangle Park, N. C. 1977.

Zimmerman, P. R., Chatfield, R. B., Fishman, J., Crutzen, P. J., and Hanst, P. L. (1978). Estimates on the production of CO and H₂ from the oxidation of hydro- carbon emissions from vegetation. Geophys. Res. Lett., 5, 679-682.

Zimmermann, P. H. (1984). Ein dreidimensionales numerisches Transportmodell für atmosphärische Spurenstoffe. Dissertation, University of Mainz, F. R. G.

 

Back to Table of Contents

The electronic version of this publication has been prepared at the M S Swaminathan Research Foundation, Chennai, India.