Chapter 2. the Atmospheric Sulfur Cycle in the Tropics

Acidification in Tropical Countries Edited by H. Rodhe and R. Herrera @ 1988 SCOPE. Published by John Wiley & Sons Ltd

CHAPTER 2 The Atmospheric Sulfur Cycle in the Tropics

ROBERT DELMAS AND JEAN SERVANT

2.1 Introduction...... - ...... 43 2.2 The Atmospheric Sulfur Cycle in Equatorial Regions...... 4S 2.2.1 The Tropical Environment...... 47 2.3 Production of Atmospheric Sulfur...... - . 48 2.3.1 General Survey of the Transformation of Sulfur...... 48 2.3.2 Influence of the Environment on Biogenic Sulfur Sources. . . 48 2.4 Biogenic Sources...... 51 2.4.1 Biomass Burning...... 53 2.4.2 Marine Sources...... 55 2.4.3 Soil-derived Sources...... 56 2.4.4 Other Sources...... 56 2.4.5 Conclusions...... 56 2.5 Sulfur Compounds in the Tropical Atmosphere...... 57 2.5.1 Concentrations of Sulfur Compounds...... - . 57 2.5.2 Concentration of Sulfur Compounds in the Equatorial Atmosphere...... 57 2.6 Sinks of Sulfur Compounds in Intertropical Regions...... 60 2.6.1 Sinks...... 60 2.6.1.1 Dry Deposition...... 60 2.6.1.2 Wet Deposition...... 60 2.6.1.3 Seasonal Variations in Dry and Wet Deposition. . . 61 2.6.1.4 Acid Rainwater and Influence of Sulfur Compounds. 61 2.6.1.5 Transport-Influence of Equatorial Regions on the Global Atmospheric Sulfur Cycle...... 66 2.7 Conclusions...... - ...... 66 References...... 67

2.1 INTRODUCTION

Sulfur is ubiquitous in nature: it exists in soils mainly as organic compounds, but also as sulfate or sulfide (0.5 mg S/g); in sea water as sulfate (2.65 mg

43 44 Acidification in Tropical Countries

S/g); in plants as sulfate, sulfide, or organic compounds (0.1 to 2 mg/g of dry matter); and in the atmosphere in gaseous and solid states (0.3 /Jg to 1 /Jg/kg). The exchange of sulfur between these different media generally takes place in the atmosphere. Sulfur is emitted into the atmosphere in a more or less reduced state as hydrogen sulfide, organic sulfides, or as sulfur dioxide. Much of it is then oxidized into sulfuric acid or its salts and removed mainly by precipitation. In the absence of major climatic changes this natural sulfur cycle is stable, disturbed mainly by volcanic eruptions. A recent perturbation is the result of industrialization, mainly as a consequence of increasing energy demand. Based on evidence from Arctic polar ice samples (Herron, 1982; Murozumi et at., 1969), this perturbation became significant around 1900 and is mainly confined to the Northern Hemisphere. Assessments of sulfur fluxes from both natural and anthropogenic sources vary (Kellog et at., 1972; Friend, 1973; Robinson and Robbins, 1975; Granat et at., 1976; Cullis and Hirschler, 1980; Sze and Ko, 1980; Muller, 1984a, b). In these studies the natural emissions were poorly known owing to the small number of reliable measurements. Natural processes have become better understood during the past decade thanks to data obtained by Aneja et at. (1979a, b), Adams et at. (1979, 1980), Delmas et at. (1980a), Delmas and Servant (1983). These studies, carried out in the eastern US and in equatorial Africa, showed that a positive gradient of the S-flux is to be expected with decreasing latitudes and that the maximum occurs in equatorial regions. The global natural and anthropogenic emissions based on the most recent estimations are presented in Table 2.1. The production of sulfur by sea- spray has been omitted. The natural emissions emanating from land and sea

Table 2.1 Global emissions of sulfur compounds from natural and anthropogenic sources

Source Intensity References 1012g(S)/year Natural Marine 40-48 Andreae and Raemdonck, 1983 (revised by Savoie, 1984) Continental 35-50 Moller, 1984a Volcanic 2-12 Moller, 1984a; Cadle, 1980

Anthropogenic in 1975 69-104 Moller, 1984b; Cullis and Hirschler, 1980 The Atmospheric Sulfur Cycle 45 are approximately equal, and man-made emissions are of the same order of magnitude. Sulfur compounds in the atmosphere are rapidly removed or oxidized, except for carbonyl sulfide (CaS). According to Granat el at. (1976), the production of submicron particles is considered the most important reason for the elaboration of a global sulfur budget. These submicron particles consist partly of sulfuric acid and sulfates. Sulfuric acid is probably the main agent of rainwater acidification even in the presence of other strong (e.g. nitric) or weak (e.g. formic and acetic) acids. However, in the ionic state of sulfate, sulfur is a natural fertilizer in air and in some soils (Tamm, 1976). This chapter describes the sulfur cycle in tropical regions in order to provide data on the state of acidification of air and rainwater in a tropical environment before further industrialization.

2.2 THE ATMOSPHERIC SULFUR CYCLE IN EQUATORIAL REGIONS In order to describe the effects of the atmospheric sulfur cycle on a particular ecosystem, three components should be considered: (i) the external or internal sources, (ii) the chemical and physical transformations of the different compounds emitted to the atmosphere, and (iii) the sinks that include deposits and transport outside the ecosystem. Different sources of atmospheric sulfur are found in equatorial regions. The main source is associated with the biological transformations of sulfur in soil and in the biosphere. The reduction of sulfates to sulfides by the mineralization of organic matter under special conditions supplies the atmosphere with gaseous sulfur compounds. These compounds are emitted in reduced forms such as hydrogen sulfide (H2S) and different methyl and carbon sulfides. The complexity of the cycle and of the processes concerned are illustrated in Figure 2.1. In the total sulfur cycle, reactions in the soils are bidirectional. In the strongly oxidizing environment of the earth's atmosphere, however, natural conversion from reduced to oxidized forms is the rule; there is no important reverse reaction from a higher oxidation state to a lower one. While the direction of the reaction does not vary during the day, reaction rates do. The processes in the soils are continuous, but their intensity is modulated by the type of soil and vegetation, and by climatic variations. Biomass burning is the other source linked to the biosphere; it includes bushfires in dry savannas, and fires used in agriculture or for deforestation. This source is mainly of anthropogenic origin and is sporadic, occurring only during the dry season. ~ 0'

Solid or CH3S03 H liquid Soils' ------phase Volcanoes wind S02 erosion Gaseous phase

Oceans' Man-made -Bubbl ing emissions -Diffusion of methy I 1 suIf ides 2

. Bushfires in -Deforestation savanna due to regions shifting agriculture -Savanna and 1 = natural forest 2 = anthropogenic burnings

1 Biological transformations of S in soils 2 Biomass burning

Cont inental biosphere

Figure 2.1 Simplified sulfur cycle in the tropics The Atmospheric Sulfur Cycle 47

In the equatorial regions, atmosphere-biosphere exchanges are the main sources of atmospheric sulfur, although other sources exist. The main sources are:

(a) The oceans: all the continental zones in the intertropical regions are influenced by oceanic air in the lower layers of the atmosphere, either permanently (South America, equatorial Africa, Oceania), or sporadi- cally, according to the latitudinal position of the meteorological equator (West Africa, tropical Asia) (Riehl, 1954). (b) The soils: important quantities of dust, moved into the air by strong winds, are a source of particulate sulfur. This phenomenon is particularly important in the Asian and African desert zones.

Other sources are industrial and domestic activities and volcanic emissions. Some compounds introduced into the atmosphere are stable, such as particulate sulfates (oceanic, soil-derived); others are unstable. The unstable compounds can be reduced gases such as H2S (biosphere), methyl sulfides (ocean, biosphere), or carbonyl sulfides (biosphere, fires); oxidized gases such as sulfur dioxide (S02), emitted during combustion or derived from the oxidation of reduced gases; and some other species with short lifetimes. Sulfur compounds are eliminated from the atmospheric reservoir by dry and wet deposition, with different speeds at different steps in the oxidation chain. Wet deposition arises from in-cloud scavenging and below-cloud scavenging. Rain acidity is due to the incorporation of acidic compounds (S02, H2SO4)' Vertical and long-range horizontal transports could also be considered as sinks, particularly for carbonyl sulfide (COS).

2.2.1 The Tropical Environment The previous discussion described how the natural environment influences the atmospheric sulfur cycle in equatorial regions. The two main biomes in these regions are the forests and the savannas. It is difficult to estimate the surface of the equatorial forests in the world, as they are evolving constantly and decreasing rapidly as a result of deforestation for cultivation. Lanly and Clement (1979) give a value of 17.6 x 10 6 km 2 for all intertropical forests, with primary forests representing less than 3 x 10 6 km2. The total biomass of the equatorial forests comprises about 50% of the continental biomass of the planet. The equatorial forested zones have a high rainfall, usually estimated at 1,400 mm/year (Richards, 1964), and some regions receive more than 3,000 mm/year. Savannas cover about 20 x 106 km2 and receive an annual rainfall of between 120 and 900 mm. The biomass of the savannas decreases with rainfall amount, but no linear relation between these two parameters exists. 48 Acidification in Tropical Countries

2.3 PRODUCTION OF ATMOSPHERIC SULFUR

2.3.1 General Survey of the Transformation of Sulfur The natural cycle of sulfur in the soils includes a complex chain of oxidation and reduction reactions controlled by microorganisms. In the soils, most sulfur (75% to 99.5%) appears in organic compounds. The propor- tions of organic sulfur depend largely on the nature of the soil; the inorganic fraction remains low. Plants absorb most of the sulfur in the form of sulfate ions. Organic sulfur compounds then become mineralized due to the action of microorganisms such as bacteria, heterotrophic organisms or fungi. Stud- ies have shown that methyl mercaptan (MeSH), dimethyl sulfide (DMS), dimethyl disulfide (DMDS), and carbonyl sulfide (COS) are produced under aerobic conditions, and carbon disulfide (CS2), MeSH, DMS, and COS formation prevails under anaerobic conditions (Katoda and Ishida, 1972; Banwart and Bremner, 1975). Sulfates can be fixed in soil by organisms that use them for cellular synthe- sis, or absorbed by plants and eventually animals. They can also be reduced to sulfides. Sulfate reduction is the most important cause of the emissions of biogenic sulfur compounds into the atmosphere. One type of sulfate reduction is the assimilatory pathway, in which organisms reduce sulfate in order to meet their nutritional requirement for sulfur. The formation of volatile sulfur by this process is believed to be limited. A second type of sulfate reduction is the dissimilatory pathway. Sulfates serve as a terminal electron acceptor in anaerobic respiration for bacteria of the genera Desul- fovibrio and Desulfotomaculum. As a consequence, large amounts of sulfide accumulate in some soils, the main product being H2S. Numerous authors have studied the sulfur cycle in soils: Asami and Takai (1963), Postgate (1969), Trudinger (1969), Katoda and Ishida (1972), Schlegel (1974), Banwart and Bremner (1975, 1976) are some of them. Other works present some global analyses (Dommergues and Mangenot, 1970; Alexander, 1977; Adams et at., 1980).

2.3.2 Influence of the Environment on Biogenic Sulfur Sources Many factors have a direct influence on the emission of gaseous sulfur compounds from the soils: the S-content and composition of the organic matter; the distribution of bacteria; the type and hydrology of the soils, which de- termine their pH and Eh; the temperature; and the atmospheric proximity of the production zone of S-gases. The composition of organic matter influences the nature and quantity of sulfur gases emitted. Emission rates were linked positively to soil organic matter by Farwell et at. (1979). However, there seems to be a threshold of The Atmospheric Sulfur Cycle 49

about 2% organic matter below which sulfur volatilization does not occur. Two conditions are necessary for sulfate reduction: the source of sulfates must be unlimited and the content of electron donors (organic acids or hydrogen) must be sufficient. Decomposing litter could be an important site for volatilization (Haines, 1983). The hydrous state of the soils determines the oxygen content and the emission of sulfur compounds (Table 2.2). Water-saturated soils promote sulfate reduction through dissimilation. A laboratory study on different types of soils carried out by Farwell el al. (1979) showed that emissions of H2S and CS2 greatly increase with water content in swamp soils. The most important sulfur species emitted to the atmosphere is H2S. Aneja et al. (1980) found that about 50% to 70% of the total measured sulfur flux in the eastern United States consists of H2S. Saline marshes are strong emitters compared with other types of soils.

Table 2.2 Sulfur fluxes of different soils in North Carolina, USA (Adams et ai., 1979

Soil order AverageS flux (glm2/year) n H2S OCS CH3SH DMS CS2 DMDS L:S Alfisol 9 0.010 0.002 - 0.001 0.002 0.002 0.017 Inceptisol 19 0.0028 0.0022 - 0.0002 0.0012 0.0014 0.008 Histosol 11 0.0178 - - 0.0007 0.0001 0.001 0.019

Saline swamp 1 0.0194 0.0016 - 0.0069 - - 0.028 Saline marsh 7 139.50 6.36 6.56 - - - 152.40 Alluvial clay 11 0.0004 0.004 - 0.0001 - - 0.0002

Important variations of H2S emissions with different soils have been observed by Delmas el al. (1980a) and Jaeschke et al. (1978). The intensity of the emission was much more important under anaerobic conditions, thus confirming the dominance of dissimilatory sulfate reduction as a biogenic source of sulfur. Soil temperature influences the rate of microbial transformation of sulfur in soils. Enzyme activity is multiplied by a factor 2 to 3 for every 10 °C up to 50°C. Above this threshold, enzymes in solution are rapidly inac- tivated (Katoda and Ishida, 1972). This dependence has been observed by several authors (Hill et al" 1978;Jaes~hket;Cal,! 1978),A linearrelationship with the mean ambient temperature appears, with a correlation coefficient of 50 Acidification in Tropical Countries

250

~ 200 /,' 0 .- Q) /' >- .-..- ...... /' II) /' c: 150 +-0

"U ~ 01 100 >- .J:I x ::3 -~ ::3 -::3 (/) 0 0 5 15 T °C Figure 2.2 Regression of mean ambient temperature by East-West Sure Grid tiers versus sulfur fluxes (from Adams et at., 1980). Copyright 1980, Electric Power Research Institute (EPRI) report EA-1516, Biogenic Sulfur Emissions in the SURE Region. Reprinted with permission

0.85, from 158 S-flux measurements performed in the eastern United States (Adams et ai., 1980) (Figure 2.2). Gaseous sulfur compounds are emitted into the atmosphere if the production zone is near the surface. In an oxygen zone, sulfides can be oxidized into sulfates by bacteria such as Beggiatioa and Thiorix. In this environment, gaseous sulfides cannot reach the atmosphere, which is why hydromorphic soils such as flooded forest soils, swamps, and paddy fields are important sources. This is not the case for lakes and lagoons except near the shores or under eutrophication conditions. The sulfides produced in the sediment layer are oxidized before reaching the surface, as exhibited in a lake of the Amazon basin by Brinkmann and Santos (1974) (Figure 2.3). These results were confirmed recently by Staubes et at. (1986). It has been believed for years that H2S formed in anaerobic sites in culti- vated soils could not escape oxidation by FeH ions, but this assertion was disproved recently (Delmas et at., 1980a). From a general point of view, any type of soil can be a source of reduced sulfur compounds either during mineralization of organic matter under aerobic or anaerobic conditions, or during sulfate reduction. This last source is probably the most important, especially under anaerobic conditions. The Atmospheric Sulfur Cycle 51

m Lake surface 9 8 (504 " ~ " -.- / I ' 6 1/ y' .', /" , 4 ,_/ , I I I 2 !.." ,\ \ 0 0 .2 .3 .4 :5 .11 mg/liter

Figure 2.3 Distribution of oxygen (02), sulfates (SO~-) and hydrogen sulfide (H2S) in an Amazonian flood plain lake (from Brinkmann and Santos, 1974). Copy- right 1974, Tellus. Reprinted with permission

2.4 BIOGENIC SOURCES

Despite its probable importance, this source is not well known. Detailed studies, such as those carried out by Adams el al. (1980) and Aneja el al. (1979a, b) in midlatitude zones in the USA, have not been undertaken in the tropics. A study of H2S emissions from the humid tropical forests of the Ivory Coast was conducted by Delmas el at. (1980a) and Delmas and Servant (1983). The flux was measured using two methods. The first was by accumulating H2S at night under a forest canopy. The second method in- volved H2S vertical profile measurements using aircraft at different distances from the coast, in the direction of the horizontal advection of air masses in the monsoon layer « 2,000 meters alt.). The first method yielded a 12.5pg S/m2/hour flux for dry soil and 300 pg S/m2/hour for partly flooded soil. The second method yielded a similar value: 40 pg S/m2/hour above swamps and flooded forests (Delmas el aI., 1980a). Soil emissions from forested zones were also calculated using an atmospheric circulation model in which the planetary boundary layer (the monsoon layer in that case) was considered to be a box at equilibrium. The concentration of sulfur in particulate matter (SO~-), S02 and H2S were measured over a one-year period at the entry and exit of the box. The results obtained for the exchanges of sulfur compounds allowed the calculation of soil emissions (Figure 2.4). Sulfur emissions showed pronounced seasonal changes. Values were low during the dry season, 55 pg S/m2/hour; and high during the wet season, 270 pg S/m2/hour. The annual mean value was 170 pg S/m2/hour or 1.5 g 52 Acidification in Tropical Countries

... 400 ::> 0 ~ '"...... E ...... CJI :t 300

200, Annual mean value

100

0 J F M A M A SON D

S/m2/year. These values have the same order of magnitude as those deducted from measurements and are in accordance with the relationship between S- flux and latitude, increasingtoward the equator. Adams et ai. (1980) predict an annual average H2S flux ranging from 0.5 to 1.0 g S/m2/year near the equator; the total suifur flux might be as high as 2.3 g S/m2/year at the equator. The above approach to estimating fluxes accounts for emissions from different types of soils and shows seasonal variations in intensity of the source. However, this determination is not precise due to uncertainties regarding chemical and meteorological parameters, and all S-compounds were not taken into account. The flux of organic sulfides is probably low: during the dry season, (CH3h SH + C2Hs SH emissions reach only 1 to 70 X 1O-4g S/m2/yearin the La Selva rainforest in Costa Rica (Haines et ai., 1984), about 20 x 10-4 the average value calculated for the Ivory Coast. In a forest at equilibrium, external inputs and outputs of sulfur must The Atmospheric Sulfur Cycle 53 be equal. Outputs are the sum of emissions into the atmosphere from litter, soils, and river run-off. In the tropical rainforests of Costa Rica and Venezuela, net exchanges of sulfur are 1.17 and 0.4 g S/m2/year respectively (Haines, 1983). If river run-off is taken as negligible, these values represent maximum emissions of gaseous sulfur to the atmosphere. They are compa- rable to the mean H2S fluxes measured in the Ivory Coast, and correspond to a global emission for the tropical forests of 21 and 7 Tg S/year compared to 27 Tg S/year from the data obtained in the Ivory Coast study. Clearly, any extrapolation from available data presents large uncertainties. Further research is needed in order to evaluate this source accurately. Such studies present difficulties because source intensities vary with the nature of the soils and with soil water status. For example, data obtained for water-saturated soils are an order of magnitude greater than those for dry soils. A detailed study of all sulfur compounds emitted is needed. Studying fluxes of COS and CS2 in relation to the global S-cycle would be interesting even though these gases are less important on a regional scale due to their longer residence time in the atmosphere.

2.4.1 Biomass Burning Biomass burning is an important source of aerosols and gases in tropical areas. Primarily anthropogenic, these burnings occur for three reasons: deforestation for cultivation; clearing associated with shifting agriculture; and annual bushfires, which are traditional in the dry savannas. This phenomenon is particularly important in Africa where 40% to 60% of the savannas located between 5° and 10° N are burned each year (Deschler, 1974). Seiler and Crutzen (1980) (see Table 2.3) estimate the spreading of these fires and the biomass consumed. In their approach the total amount of the biomass burned annually (M) in a biome is:

M = A x B x Q X f3 (gram of dry matter per year) = g dm/year where:

A = total land area burned annually (m2/year) B = the average of organic matter per unit area in the individual biome (g dm/m2) Q = fraction of the average above-ground biomass relative to the total biomass B f3 = the burning efficiency of above-ground biomass.

In orderto infer thequantitiesemittedinto the atmospherefrom bioma~~ burning, the sulfur content of the biomass and the amount of sulfur 54 Acidification in Tropical Countries volatilized during the fires must be determined. Such an estimation is difficult because the S-contents of various plant species in forests and savannas vary, ranging from 0.2 to 2 mg/g (0.02-0.2%). However, the fraction of volatilized sulfur depends on the type of plant and on temperature of the combustion. Two examples are given for a savanna in the Ivory Coast (Table 2.4). The fraction of volatilized sulfur is 67% for a Loudetia savanna and 35% for an Andropogon savanna. Clearly, it is difficult to evaluate accurately the intensity of this source; more data are needed. An estimation of the magnitude of the source can be made by considering the mean biomass burned annually, 3.6 Tg dm/year (Seiler and Crutzen, 1980), with a maxi-

Table 2.3 Amounts of biomass burned annually in tropical and subtropical areas (Seiler and Crutzen, 1980) (unit: 100 Tg dm/year)

Activity Burned area Biomass exposed Annually burned (million hectare) to fire biomass

Burning due to 21-62 24-72 9-25 shifting agriculture (42) (48) (17) Deforestation due to 8.5-15.1 16-25 5.5-8.8 population increase (12.0) (20.5) (7.2) and colonization

Burning of savanna 12.2-23.8 4.8-19 and bush lands (600) (18) (11.9)

Table 2.4 Sulfur emissions during bushfires in two types of savanna in the Ivory Coast (Delmas, 1982)

Loudetia Andropogon savanna savanna

Total above-ground biomass before 7 9 the fire (T(dm)/ha)

S content in the dry matter (mglg) 0.29 0.33

S amount in above-ground biomass 2.03 3.0 before the fire (kglha)

Average percentages of sulfur volatilized 67 35

S fluxes (kglha) 1.40 1.05 The Atmospheric Sulfur Cycle 55 mum S-content of 0.2% (Bowen, 1969; Shriner and Henderson, 1978) and an S-volatilization efficiency of 50%. This yields a maximum emission of 3.6 Tg S/year. In the intertropical zone as a whole, this source is weaker than the biological source, but it may be significant in some regions. Our estimate is lower than that of Andreae (1985), which is about 7 Tg S/year, corresponding to an annual biomass burning of 6,800 Tg dry matter. Our calculation concerns only the tropics and does not include the use of wood as fuel, which represents about 1,000 Tg dry matter/year. Compounds formed during combustion have not been analyzed. Measure- ments in the Ivory Coast showed that S02 was dominant in the smoke plume and that H2S was also formed (Delmas, 1982). Very hot air displaced by the wind before the flames causes the vegetation to parch and become partially pyrolized before burning. This is likely the cause of the emission of reduced gases such as H2S. Emissions of COS have also been observed by Crutzen et ai. (1979). The ratio of COS/C02 concentrations = 6 x 10-6 obtained from aircraft measurements corresponds to a global source equal to 4 x 1010 g (COS-S)/year, or one-tenth of the total emission of COS into the atmosphere (Crutzen et ai., 1985).

2.4.2 Marine Sources

Oceans are the major natural source of sulfur in the troposphere, with SO~- a ubiquitous trace species in the oceanic atmosphere. A large fraction of this SO~- is derived from sea-spray and is not discussed further here. An ex- tensive survey on sulfate concentrations from nonmarine origins over 25 regions of the world ocean concluded that the average S-concentration is about 0.7 j.tg/m3.This concentration differs across regions, e.g. by a factor of 30 to 50 for polluted regions of the Mediterranean Sea and the pristine areas of the southern oceans (Savoie, 1984). In tropical areas, the highest concentrations were recorded to the west of the continents: 1.6 j.tg/m3 in the Gulf of Guinea, 5 j.tg/m3 west of South America, 1.1 j.tg/m3 in the North Atlantic Ocean, 2 j.tg/m3 in the Gulf of Mexico, and 1.5 j.tg/m3 in the Bay of Bengal. In the intertropical oceans, values of 0.3 to 0.6 j.tg/m3 were observed. Higher concentrations are often linked to terrigenous dusts transported from the deserts of Arabia, India, Pakistan, and Sahara. Near the west coast of South America a high sulfate concentration results from industrial activity in Chile. Near the northwest coast of South America the high sulfate concentration can be attributed to either pollution or a natural source. The S02 concentration in air is not necessarily linearly correlated with oceanic productivity (Nguyen et ai., 1978; Bonsang, 1980; Andreae et at., 1983; Cline and Bates, 1983). However, such a relation between the final products of S-compounds and the emission of DMS has been demonstrated. 56 Acidification in Tropical Countries

2.4.3 Soil-derived Sources

Sulfur derived from soils is associated with wind erosion. At high wind speeds, dusts are suspended in the atmosphere, and the finest aerosol particles are carried by air masses over thousands of kilometers during dry haze episodes (Prospero and Bonatti, 1969; Bertrand et al., 1974; Rahn, 1981; Duce et al., 1980; Bravo et aI., 1981; Darzi and Winchester, 1982; Ue- matsu et aI., 1983; Parrington et aI., 1983; Ryaboshapko, 1983). In Africa, this source represents 15% of the atmospheric SO~- in the equatorial zone (Delmas et aI., 1978; Delmas and Servant, 1983). This value was estimated from chemical criteria and explained by the presence of gypsum in the zones where the dry haze originates.

2.4.4 Other Sources Volcanic emissions and man-made pollution are other potential sources of sulfur compounds in intertropical zones. Active volcanoes in tropical zones are numerous, but most of the sulfur emitted during eruptions, about 2 Tg S/year globally (Muller, 1984a), probably enters the stratosphere directly. Estimations from Naughton et al. (1975) and Cadle (1980) are higher, at 23.5 and 20-30 Tg S/year. If we consider the amount of sulfur in the oceans, sedimentary rocks, and evaporites, these values are too high on a long-term basis. Volcanic emissions into the troposphere during noneruptive periods appear to dominate and a global release rate of 12 Tg S/year is probably a higher limit for the volcanic flux. This source may be important on the regional scale during volcanic episodes. The present anthropogenic sulfur emissions over the world are estimated as 104 Tg S/year (Cullis and Hirschler, 1980) and 75 Tg S/year (Muller, 1984b). The difference between estimates is due to the higher emission factors for coal and lignite used in Cullis and Hirschler's work. Global S02 emissions are expected to increase between 1975 and 2000, mainly due to increased coal combustion. In 1979, man-made sulfur emissions are estimated to have been 1.1, 2.1, and 0.35 Tg S/year for Africa, Central and South America, and Oceania, respectively. The total emissions of these regions represent 8% of the world emissions (Varhelyi, 1985).

2.4.5 Conclusions The major component of the atmospheric sulfur cycle in the temperate latitudes is anthropogenic. In inter-tropical regions, however, the main terrestrial source is natural biomass decomposition. Recent studies indicate that The Atmospheric Sulfur Cycle 57

most soils and litter decomposition can be major sources of sulfur compounds for the atmosphere, with the greatest emissions occurring under anaerobic conditions. In humid equatorial regions where rains are abundant and anaerobic sites are common, the high temperature and organic content of the soils present favorable conditions for greater S-emissions. Flooded soils, probably the greatest source, depend on the annual rainfall distribution. The emission of sulfur compounds in savanna soils has not been measured, although one might assume that its value is lower than in humid areas because anaerobic conditions are less common. One important source could be biomass burning, but this is a difficult source to evaluate as it is sporadic and occurs during the dry season. In some tropical countries, particularly those with high population densities in tropical Asia, anthropogenic emissions might exceed natural ones. In China, very high S02 concentrations are recorded south of the Yangtze River as a result of coal consumption (Zhao and Sun, 1986).

2.5 SULFUR COMPOUNDS IN THE TROPICAL ATMOSPHERE

2.5.1 Concentrations of Sulfur Compounds The data on concentrations of sulfur compounds are limited. Most measurements concern total S or SO~- concentrations in particulate matter (Table 2.5). No measurements of carbon sulfide concentrations are available for the tropical zone.

2.5.2 Concentration of Sulfur Compounds in the Equatorial Atmosphere Several authors have measured the size spectrum of particulate SO~-. The results obtained in the Amazon basin by Orsini et at. (1982) are shown in Table 2.6. The SO~- particles were collected with a cascade impactor six- stage Battelle Model); the analysis of sulfur was carried out by Pixe (particle induced X-ray emission). The spectrum obtained differs from that of oceanic and continental atmospheric particles. The maximum mass of SO~- corresponds to the submicron particles. Similar results were obtained in the Ivory Coast (Delmas, 1980) and in the Congo (Delmas et at., 1986). The cascade impactor used in the latter study was a six-stage high-volume sampler (Weather Measure model), and the analysis was carried out by atomic absorption. About 70% of the SO~- mass corresponded to the submicronic fraction. On the other hand, chemical criteria for marine- and soil-derived origins have shown that 70% of the massof SO~- in parti~ulate matter correspondsto airborne sulfates in the Ivory Coast (Delmas el at., 1978). V1 00

Table 2.5 Atmospheric concentrations of sulfur compounds in the intertropical zone (average values in parentheses)

References Location Concentration (nglm3) S particulate SOz SO~- HzS DMS

Lawson and Bolivia, mountain 51 Winchester, 1979 Manaus, Brazil 131 Salvador, Brazil 260

Orsini el ai., 1982 Amazon Basin 215-342

Lodge el ai., 1974 Panama, 850-3,150 equatorial forest

Penkett el ai., 1979 Sudan 500-9,000

Delmas el ai., Ivory Coast, 500-2,000 500-3,800 50-3,500 2.7-7.1 1978 and equatorial forest (650) 1,700) (600) Delmas and Servant, 1983 n- - - -

The Atmospheric Sulfur Cycle 59

Table 2.6 Granulometric mass-spectrum of sulfur in atmospheric aerosol in the Amawnian basin (Orsini et al., 1982) (units: ng(S)/m3)

IC stage 0 I 2 3 4 5 (S) Sample Size range (jlm) <0.25 0.25-0.5 0.5-1 1-2 2-4 >5 IC AM 4A 27 160 92 23 14 26 342

IC AM 4C 19 157 132 23 24 20 375

IC AM 3A - 136 57 12 7.8 2.0 215

IC AM 3C 5.8 96 106 - 8.5 3.0 219

IC AM 2A 3.1 138 48 5.7 12 - 235

IC AM 2C - 127 74 9.3 12 3.0 226

This particle-size distribution, mainly in the submicronic range, implies that SO~- is formed primarily from homogenous gas-phase reactions. If the formation had been the absorption and subsequent reaction of vapor-phase H2SO4 on seasalt or soil-derived particles, the SO~- mass distribution would have been similar to the surface distribution of these particles. In that case, SO~- would have shown a mean mass diameter of 4 J.l.m. No in situ study on the mechanism of S-oxidation in equatorial regions has been reported, but results were obtained using a budget method (Delmas and Servant, 1983). Above a forested zone in the Ivory Coast, the calculated monthly mean values of the oxidation rate show a variation from 5.3% to 12.9%/hour with an annual arithmetic mean of 8.4%/hour. This value is an order of magnitude higher than those observed at midlatitude rural zones. It was obtained through an indirect budget method, so precision is likely to be low. However, it seems to account for the concentrations of sulfates observed on the continent. A direct measurement of this oxidation rate in Australia (Ayers, 1986) yielded a lower value of O.25%/hour, but the photochemical characteristics of the atmosphere in Australia and equatorial Africa may have been different. Rapid gas-phase oxidation suggests that the equatorial atmosphere presents favorable features, in particular a relatively high concentration of OH-radicals. Between 15 x Nand 15 x S, the photochemical model of Logan €l at. (1981) givesan OH-concentration of about 2 x 106 atoms/em3 at ground level. Crutzen and Gidel (1983) obtained similar results. 60 Acidification in Tropical Countries

2.6 SINKS OF SULFUR COMPOUNDSIN INTERTROPICAL REGIONS

2.6.1 Sinks

Sulfur compounds leave the atmosphere by dry and wet deposition. The elimination of sulfur in an ecosystem also occurs via horizontal advection and flux of matter outside the limits of the ecosystem. Deposition processes depend on the chemical and physical characteristics of the compounds, such as size spectrum for aerosols and solubility for gases. The residence time of a compound in the atmosphere depends on all these parameters.

2.6.1.1 Dry Deposition Gases or particles are deposited by mechanisms that depend on near-surface concentrations and on the deposition velocities of the elements. During the day, the near-surface concentration is close to the mixing layer value due to atmospheric turbulence. At night, under stable atmospheric conditions, the near-surface concentration decreases. Deposition velocity is independent of concentration but depends on aerodynamic conditions, the size distribution of the particles, and the solubility (or chemical reactivity) of the gases. For the S02 example, the Henry's coefficient is 1.3 M/atm, but the effective H- coefficient varies from pH 2 to pH 6 (Schwartz, 1984). Deposition of soluble gases on vegetation depends on whether the stomata are open or closed. Because of pollution, S02 is the most extensively studied gas in the industrialized countries (Belot et at., 1974; Garland, 1974; Petit et ai., 1976; Garland and Branson, 1977; Fowler, 1978; Shreffler, 1978), where deposition speeds have been measured between 0.3 and 3.2 em/see, averaging around 1 em/sec. No similar study has been carried out in the intertropical regions. The deposition velocity of SO~- particles is low « 0.1 em/see) and does not depend on gravity due to the particles' submicronic diameter. In the continental zones, Cawse (1974) and Muller and Schumann (1970) indicated a minimum in the 0.1-1 Jim range, and measured a deposition velocity of 0.01 em/sec. Based on past estimates of dry deposition, Galloway (1985) gives values for dry deposition rates of 0.02 and 0.02 g S/m2/year for S02 and excess SO~- in remote marine areas, and 0.07 and 0.03 g S/m2/year in remote continental areas.

2.6.1.2 Wet Deposition Wet deposition is proportional to the total annual rainfall and concentration of sulfur compounds in the rainwater. The compounds are mainly sulfates and sulfites, which could result from the dissolution and oxidation of S02 The Atmospheric Sulfur Cycle 61 in the liquid phase or from rainout and washout of particulate SO~-. In most experiments, the collectors are not closed between rain events, thus the dry and wet deposits are analyzed together. The data obtained by authors at different sites are displayed in Table 2.7. Concentrations vary by a factor 5, the lowest corresponding to the highest rainfall intensity. Delmas, Mathieu, Galloway et al. and Lacaux et al. obtained data under good conditions for collection and analysis. In particular, Galloway et at. (1982) used collectors that opened and closed before and after the rain, to preclude dry deposition. The analysis was done by an ionic separation followed by conductivity measurement (Dionex apparatus) giving good precision and reproducibility in the sub-ppm concentration range. Lacaux et at. (1986) used an automatic sampler and a Dionex apparatus for SO~- analysis. They found an SO~- washout ratio (the ratio of concentration of an element or ion in one liter of rainwater to its concentration in one cubic meter of air) of 103. This is compared to 3 x 102, the geometric mean of the ratios at Samoa, Miami, and Bermuda, which have been used to calculate the SO~- deposition flux over the world oceans (Andreae and Raemdonck, 1983).

2.6.1.3 Seasonal Variations in Dry and Wet Deposition The study undertaken in the Ivory Coast by Delmas (1980) on particulate SO~- and on the main gaseous S-species H2S and S02 affords the oppor- tunity to compare the relative importance of dry and wet depositions in a tropical atmosphere. Studied over a one-year period, the seasonal variations are apparent (Figure 2.5). Dry depositions are calculated from average monthly concentrations of S02, H2S, and SO~- measured in a forested zone, with deposition velocities of 1 em/see for H2S and S02, and 0.1 em/see for SO~-. Wet deposition estimates are based on the weighted means of data obtained at 11 stations in this zone. The wet sulfur deposition, representing about two-thirds of the total, is the highest. This coincides with the assessment of sulfur deposition in remote continental areas given by Galloway (1985). Wet deposition modulates seasonal variations, and total deposition is highest during the wet season due to an increased source of biogenic sulfur in the atmosphere, and due to washout.

2.6.1.4 Acid Rainwater and Influence of Sulfur Compounds In pristine areas without anthropogenic sources of sulfur, the pH of rainwater is thought to be controlled by the carbonate system buffer CO2/H2C03 and has a value of 5.6. The natural parts of the atmospheric cycles of sulfur, nitrogen, and carbon introduce strong acids (H2SO4, HN03) and weak acids (HCOOH, CH3COOH) sufficient to lower the pH of natural rainwater by 0\ N

Table 2.7 Concentrations of SO;- in rainwater in the equatorial zone (average values in parentheses)

References Location Remarks Concentration Rainfall Deposits (mg(SO;-S)/liter) (mm/year) (g(S)/m2/year)

Visser, 1961 Kam pala (0.6) (0.8) Uganda Reponed by Granat et at., 1976 Hesse, 1957 Kiyuhu 0.1 Rodhe et at., 1981 Kenya, Av. 9 locations, 3 years Tanzania Monthly samples Uganda 0.21 Eriksson, 1966 Zaire 0.28 Bromfield, 1974 Nigeria N. Av. 11 locations, 2 years 0.13 877 0.11 Mathieu, 1972 Ivory Coast Av. 1 location, 2 years (0.26) Delmas, 1980 Ivory Coast Av. 11 locations, 1 year (0.43) 1,604 (0.69) Lacaux et at., (unpublished) Lamto, Av. 1 location, 1 year 687 (0.35) Ivory Coast

Reponed by Haines, 1983 Johnson et at. 1979 La Selva, Av. 1 location, 1 year (0.35) 3,400 1.25 Costa Rica Jordan and Murphy, 1982 San Carlos, Av. 1 location, 1 year (0.42) 3,500 1.52 Venezuela Galloway et at., 1982 San Carlos. Av. 1 location, 6 months (0.17) 3,9100 Venezuela The Atmospheric Sulfur Cycle 63

... N E ~ ..... :::: c> c> E E E E 200 400

150 300

100 .5 200

0 0 M A M A 5 0 N D

Figure 2.5 Monthly variations of sulfur deposition in a forested zone in the Ivory Coast (from Delmas, 1980). (1) Wet deposition (mg(S)/m2/month); (2) dry deposition (mg(S)/m2/month); (3) concentrations of SO~- -S in rainwater (mg/liter); (4) monthly rainfall (mm) about one unit from that value. The impact of sulfur compounds alone might cause pH values to range from 4.5 to 5.6 (Charlson and Rodhe, 1982). At the most central locations of the Antarctic plateau, before industrial pollution and far from the marine influence, ancient ice cores up to 30,000 years old were found to be acidic at pH 5.2 to 5.5 (Legrand, 1980; Delmas el ai., 1980b, 1982). Sources of base-line acidity are probably the biogenic activity of the Antarctic Ocean and atmospheric HN03. The volcanic contribution to the total H2SO4 fallout on the Antarctic plateau is estimated to be 30% for the last 100 years (Delmas el ai., 1982). Over the oceans, where pollution is low (the estimated total deposition of SO~- of nonmarine origin equals the estimated rate of DMS emission), the pH values are 5.4 over the temperate southern Atlantic zones, 5.0 over the northern Atlantic tradewind zones and 5.3 over the north Pacific (Savoie, 1984). Over the continents, pH values may decrease partly due to the S- emission. However,low valuesareobservedoverthe equatorialbelt despite high ammonia (NH3) emissions and in contrast with the predictions by Daw- 64 Acidification in Tropical Countries son (1977)basedon the acidityof tropicalsoils.These NH3 emissionshav@ been inferred from the NH4+ concentration in air and rainwater in the Ivory Coast (Servant et al., 1984). As NH3 is a base, it neutralizes the acidity of H2SO4. Examples of acidification of natural rainwater are given by the data obtained in the Global Precipitation Chemistry Project (GPCP) (Galloway et al., 1982). Precipitation from five remote areas of the world is analyzed. Two areas are located in the intertropical zone; Katherine, Australia, and St Carlos de Rio Negro in the Venezuelan tropical rainforest (Table 2.8). The average pH values at the different stations were near 4.8. At times the pH was quite acidic, below 4.5, which is the value Charlson and Rodhe (1982) quote as the lowest expected natural value. One interesting new result of this study is the evidence of a fairly constant background of strong mineral acidity with most of the pH-variability associated with organic acids. This finding was suggested by a preliminary analysis of weak organic acids, but requires confirmation at other sites to ensure its generality. This study shows that air and rainwater acidities are linked to the cycle of major elements in the atmosphere of which the sulfur cycle is one. Such a result seems to be confirmed by our recent measurements in the Congo where we have determined the chemical composition of aerosols in a tropical rainforest site (Dimonika, 400 km west of Brazzaville). Organic carbon represents about 70% to 80% of the total mass of particles (Table 2.9). The size spectrum of particulate organic carbon (POC) corresponds to submicron particles (diameter ::; 1 J.lm),suggesting a gas-particle conversion from gaseous organic compounds and probably the presence of an acidic component. In the Ivory Coast, at a site at the border of the savanna forest 200 km from the coast, rainwater was collected for the BAPMON Project. The sampler was opened during rain only, and the water was frozen immediately and sent

Table 2.8 pH and contribution of acids to free acidity (%) in rainwater (Galloway et al., 1982)

St. Georges, Poker Flat, Amsterdam Katherine, St. Carlos, Bermuda Alaska Island Australia Venezuela pH 4.8 4.9 4.7 4.7 4.8 (4.1-6.0) (4.7-5.4) (4.0-5.5) (4.0-5.4) (4.0-5.4)

H2SO4 111 65 73 33 18

HN03 .::;35 17 '::;14 26 17

HX* >0 >18 >13 >41 >65

HX* could be HCl, organic acids or H3P04; the authors believe it was an organic acid. The Atmospheric Sulfur Cycle 65

Table 2.9 Chemical composition of the atmospheric aerosol in a tropical forest site, Dimonika in the Congo (Delmas et al., 1986)

Dimonika, Concentration (nglm3) Congo POC SO- NH: NO) K+ Na+ End of rainy season 6,850 1,709 422 322 212 155

End of dry season 7,000 207 385 18 180 247 to French laboratories. The pH, conductivity, and concentrations of SO~-, NO.3, PO~-, Cl-, NH: (by ion chromatography) and of Ca2+, Mg2+, K+, Na +, FeH (by atomic absorption) were measured. From May 1983 to May 1984, 97 events were recorded with pH values between 4.3 and 6.2 (Figure 2.6). About 19% of the rain events have a pH < 5.0, 46% a pH between ~ 5.0 and < 5.6, and 35% a pH ~ 5.6. The more acidic events represent 10% of the total rainfall. In the ionic balance, sulfates and nitrates account for about 70% of the free acidity of rainwater. The remaining 30% might come from organic acids, such as formic acid, the result of isoprene oxidation by OH radicals leading to the formation of formaldehyde. The pH shows slight seasonal variations: values are higher during the dry season when soil-derived particles bring alkaline material from sahelian zones (Lacaux el ai., 1986).

en 20 -c: Q) > Q) 15 -0

Q) 10 .D E ::J z 5 0 4.4 4.7 5.0 5.3 5.6 5.9 6.2 pH

Figure 2.6 Histogramof the pH of precipitationcollectedat the Bapmon site of Lamto, Ivory Coast (from Lacaux et al., 1986) 66 Acidification in Tropical Countries

2.6.1.5 Transport-InfluenceofEquatorialRegionsontheGlobal Atmospheric Sulfur Cycle The horizontal transport of trace compounds by air masses depends on their atmospheric residence time. In the continental zones with high precipitation and humidity, the oxidation of S-gases, with the exceptions of CS2 and COS, is rapid and their residence time is probably less than one day. In these zones, the major source of sulfur is biomass decomposition and most airborne sulfates stay in the planetary boundary layer. This layer presents very efficient rainout with rainfall between 1,500 mm/year and 2,000 mm/year. In the Ivory Coast, the residence time of aerosols in the atmosphere is estimated at several hours in the rainy season and about two days in the dry season. Under such conditions, transport of the sulfur compounds will be limited mainly to a few hundred kilometers, again with the exceptions of CS2 and COSo Most of the sulfur emitted is recycled locally or in nearby ecosystems, such as the savanna. On the other hand, bushfires occur during the dry season, and particles formed in the plumes by gas-particle conversion are mostly in the submicronic range. Their residence time in the atmosphere likely approaches two or three weeks, the value usually given for the free troposphere. These particles can be carried away by air masses and dispersed over regions far away from the generating zones. They may, for example, supply the oceanic atmosphere with SO~- from nonmarine sources. In the equatorial regions, a vertical transport of sulfur compounds into the stratosphere is conceivable. These regions are located at latitudes where the tropospheric air enters the stratosphere through the ascending branches of the Hadley cells. Most of the sulfates of the stratospheric layer appear to originate in the COS photolysis over a 20 km altitude during periods with low volcanic activity (Crutzen and Schmaitzl, 1983; Servant, 1986).

2.7 CONCLUSIONS Attempting to estimate the sulfur budget over the intertropical regions is difficult due to limited and often unreliable atmospheric chemistry data for these regions. According to current knowledge, the sulfur cycle in the tropics still appears to be basically natural. This assertion does not mean that anthropogenic pollution in tropical regions is always negligible, but rather that the background of sulfur in the tropical atmosphere is natural. In the industrialized countries of the Northern Hemisphere, the background of sulfur is anthropogenic. However, man-made emissions may already exceed natural emissions in tropical countries with high population densities and these are expected to increase in the future. The natural part of the sulfur cycle in the tropics presents some typi- cal features: reaction speeds in the atmosphere and pedosphere are very The Atmospheric Sulfur Cycle 67

fast; the main source of sulfur compounds is the biodegradation of organic matter in humid zones; and the turnover of the cycle is dependent on natural and climatic conditions on a regional scale. Rapid degradation of litter due to luxuriant micro-fauna, fungal, and bacterial activity favored by constant high temperatures and soil humidity, and the presence of numerous hydromorphic zones favorable to S-emissions in relation with high rainfall affect source intensity. Features of the tropical sulfur cycle also influence the atmospheric oxidation of sulfur compounds and the recycling of these compounds through wet deposition associated with high rainfall. With a nearly constant year-round air temperature, the atmospheric fraction of the sulfur cycle is modulated by the annual rainfall distribution for both emissions and depositions. One difficulty in studying the sulfur cycle over tropical regions is the regional variation in climatic conditions. For example, during monsoons the ocean could have a great impact in some regions. In addition, the sulfur cycle should not be studied alone: atmospheric oxidation rates for sulfur are influenced by the nitrogen, carbon cycles through their impact on the concentrations of the hydroxyl radical. The acidification of air and rainwater is caused by sulfur compounds. In the tropics, the pedosphere seems to be a natural source of ammonia, which partly counterbalances the acidifying influence of sulfur compounds. This effect needs further study. Nevertheless, with low anthropogenic emissions, the pH of rainwater can be reduced to below 4.5. One reason could be the presence of organic acids, as is probable in the forested zone, with the atmospheric oxidation of the gaseous organic compounds emitted by the biosphere. Despite the great intensity of sulfur sources in intertropical regions, they exert only a small influence on the global sulfur cycle. Most of the compounds emitted are recycled locally and contribute to the stability of the primary ecosystems. One exception is carbonyl sulfide (photolized in the stratosphere), which could be emitted in the forested zones where conditions are favorable. This compound might be the source of the sulfate layer in the stratosphere during volcanic 'quiescent' periods.

REFERENCES Adams, D.F., Farwell, S.o., Pack, H.R. and Bamesberger, W.L. (1979). Preliminary measurements of biogenic sulfur-containing gas emissions from soils. APCA Jour- nal, 9, 380-383. Adams, D.F., Farwell, S.O., Robinson, E. and Pack, H.R. (1980). Biogenic sulfur emissions in the Sure region. Final report Pullman, Wash.: Washington State University, EPRI No. EA-1516,EPRI Project No. RP 856-1. 68 Acidification in Tropical Countries

Alexander, H. (1977). Soil microbiology 2nd edition, Chapter 21. John Wiley and Sons, New York. Andreae, M.D. (1985). The emission of sulfur compounds to the remote atmosphere. Background paper in Galloway, J.N., Charlson, RJ., Andreae, M.D. and Rodhe, H. (Eds.) The Biogeochemical Cycling of Sulfur and Nitrogen in Remote Atmosphere: 5-25. D. Reidel, Dordrecht, Holland. Andreae, M.D. and Raemdonck, H. (1983). Dimethyl sulfide in the surface ocean and the marine atmosphere: A global view. Science, 221, 744-747. Andreae, M.D., Barnard, W.R and Ammons, J.H. (1983). The biological production of dimethysulfide in the ocean and its role in the global atmospheric sulfur budget. Ecol. Bull., (Stockholm), 35, 167-178. Aneja, V.P., Overton, J.H. Jr, Cupitt, L.T., Durham, J.L. and Wilson, W.E. (1979a). Carbon disulfide and carbonyl sulfide from biogenic sources and their contribution to the global sulfur cycle. Nature, 282, 493-496. Aneja, V.P., Overton, J.H. Jr., Cupitt, T.L., Durham, J.L. and Wilson, W.E. (1979b). Direct measurement of emission rates of some atmospheric biogenic sulfur compounds. Tellus, 31-2, 174-178. Aneja, V.P., Aneja, AP. and Adams, D.F. (1980). Biogenic sulfur compounds and the global sulfur cycle. J. Air Poll. Cont. Assoc., 32, 803-807. Asami, T. and Takai, Y. (1963). Formation of methyl mercaptan in paddy soils. Soil Sci. Plant Nutr., 9,23-27. Ayers, G. (1986). Personal communication. Banwart, W.L. and Bremner, J.H. (1975). Formation of volatile sulfur compounds by microbial degradation of sulfur-containing amino-acids in soils. Soil BioI. Biochem., 7, 359-364. Banwart, W.L. and Bremner, J.H. (1976). Evolution of volatile sulfur compounds from soils treated with sulfur-containing organic materials. Soil BioI. Biochem., 8, 439-443. Belot, Y., Bourreau, J.e., Dubois, M.L. and Pauly, e.S. (1974). Measurement of deposition rate of sulfur dioxide on leaves of plants with sulfur 34. FAO/IACA, Isotope Ratios as Pollutant Source and Behaviour Indicators. Vienna, November 18-22. Bertrand, J., Baudet, J. and Drochon, A (1974). Importance of natural aerosols in West Africa. J. Atm. Res., July-December, 845-860. Bonsang, B. (1980). Atmospheric cycle of sulfur from marine origins. Ph.D. thesis, University of Picardie, St Quentin, France. Bravo, J.L., Salazar, S. and Muhlia, A (1981). Mineral and sea salt aerosol concentrations over the low latitude tropical Atlantic and Pacific oceans during F.G.G.E. Geofisica Intern., 20, 303-317. Brinkmann, W.L.F. and Santos, U. de M. (1974). The emission of biogenic hydrogen sulfide from Amazonian flood plain Pakes. Tellus, 26, 261-267. Bromfield, AR (1974). The deposition of sulphur in the rainwater in northern Nigeria. Tellus, 26, 408-411. Cadle, RD. (1980). A comparison ofvo1canic with other fluxes of atmospheric trace gases. Rev. Geophys. Space Phys., 18,746-752. Cawse, P.A (1974). A survey of atmospheric trace elements in the U.K. (1972- 73). AERE Report R 7669. Atomic Energy Research Establishment, Harwell, Berkshire, UK. Charlson, R.J. and Rodhe, H. (1982). Factors controlling the acidity of natural rainwater. Nature, 295, 683-685. The Atmospheric Sulfur Cycle 69

Cline, J.D. and Bates, T.S. (1983). Dimethyl sulfide in the equatorial Pacific Ocean: A natural source of sulfur to the atmosphere. Geophys. Res. Leu., 10, 949-952. 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, CH3Cl and the effect of various NO sources on tropospheric ozone. J. Geophys. Res., 88, 6641-6661. Crutzen, P.J. and Schmaitzl, U. (1983). Chemical budgets of the stratosphere. Planet Space Sci., 31, 1009-1032. Crutzen, P.J., Heidt, L.E., Krasneck, J.P., Pollock, W.H. and Seiler, W. (1979). Biomass burning as a source of atmospheric gases CO, H2, NO, CH3Cl and COSo Nature, 282,253-256. Crutzen, P.J., Delany, AC., Greenberg, J., Haageson, P., Heidt, L., Rueb, R., Pol- lock, W., Seiler, W., Wartburg, A and Zimmerman, P. (1985). Tropospheric chemical composition measurements in Brazil during the dry season. J. Atm. Chem., 2, 233-256. Cullis, C.F. and Hirschler, M.M. (1980). Atmospheric sulfur: Natural and man- made sources. Atm. Environ., 14, 1263-1278. Darzi, M. and Winchester, J.M. (1982). Aerosol characteristics at Mauna Loa Obser- vatory, Hawaii after Asian dust storm episodes. J. Geophys. Res., 87,1251-1258. Dawson, G.A (1977). Atmospheric ammonia from undisturbed land. J. Atm. Res., 20,3125-3133. Delmas, R. (1980). Contribution to the study of equatorial forests as natural sources of atmospheric sulfur compounds. PhD thesis No. 952, University Paul Sabatier, Toulouse, France. Delmas, R. (1982). On the emission of carbon, nitrogen and sulfur in the atmosphere during bushfires in intertropical savannah zones. Geophys. Res. Leu., 9, 761-764. Delmas, R. and Servant, J. (1983). Atmospheric balance of sulfur above an equatorial forest. Tellus, 35B, 110--120. Delmas, R., Baudet, J. and Servant, J. (1978). Evidence of natural sources of sulfates in the humid tropical environment. Tellus, 30, 158-168. Delmas, R., Baudet, J., Servant, J. and Baziard, Y. (1980a). Emissions and concentrations of hydrogen sulfide in the air of the tropical forest of the Ivory Coast and of temperate regions in France. J. Geophys. Res., 85, 4468-4474. Delmas, R., Aristarain, A and Legrand, M. (1980b). The acidity of polar precipitation: A natural reference level for acid rain. In Drablos. D. and Tollan, A (Eds.) Ecological Impact of Acid Precipitation: 104-105. Proceedings of an International Conference, Sanderfjord, Norway, 11-14 March. Delmas, R., Briat, M. and Legrand, M. (1982). Chemistry of south polar snow. J. Geophys. Res., 87, 4314-4318. Delmas, R., Clairac, B., Cros, B., Servant, J. and Cachier, H. (1986). Chemical composition of atmospheric aerosols in an equatorial forest area. Second Inter- national Symposium on Biosphere: Atmosphere Exchanges. Mainz FRG, 16-22 March. Deschler, W.W. (1974). An examination of the extent of grass fires in the savannah of Africa along the southern side of Sahara. 9th International Symposium on Remote Sensing of the Environment. Environ. Res. Inst. Mich., Ann Arbor, Michigan, April 15-19, 1974. Domrnergues, Y. and Mangenot, F. (1970). Microbial Ecology of the soils. Masson and Company, Paris. Duce, R.A., Unni, C-K, Ray, BJ., Prospero, J.M. and Terril, J.T. (1980). Long- 70 Acidification in Tropical Countries

rangeatmospherictransportof soil dustfrom Asia to the tropical North Pacific: Temporal variability. Science, 209, 1522-1524. Eriksson, E. (1966). Air and precipitation as sources of nutrients. In Linser, H. (Ed.) Handbook of the Nutrition of Plants and Fertilizer: 774-792. Wien, Springer- Verlag. Farwell, S.O., Sherrard, AE., Pack, M.R and Adams, D.F. (1979). Sulfur compounds volatilized from soils at different moisture contents. Soil Bioi. Biochem., 11,411-415. Fowler, D. (1978). Dry deposition of SOz agricultural crops. Atm. Environ., 12, 369-373. Friend, J.P. (1973). The global sulfur cycle. In Rasool, S.I. (Ed.) Chemistry of the Lower Atmosphere: 177-201. Plenum Press, New York. Galloway, J.N. (1985). The deposition of sulfur and nitrogen from the remote atmosphere. In Galloway, J.N., Charlson, RJ., Andreae, M.O. and Rodhe, H. (Eds.) The Biogeochemical Cycling of Sulfur and Nitrogen in the Remote Atmosphere: 143-175. D. Reidel, Galloway, J.N., Likens, G.E., Keene, W.C and Miller, J.M. (1982). The composition of precipitation in remote areas of the world. J. Geophys. Res., 87, 8771- 8786. Garland, J.A (1974). Deposition of gaseous sulfur dioxide to the ground. Atm. Environ., 18, 75-79. Garland, J.A and Branson, J.R (1977). The deposition of sulfur dioxide to pine forest assessed by radioactive method. Tellus, 29, 445-454. Granat, L., Rodhe, H. and Hallberg, RO. (1976). The global sulfur cycle. In Ni- trogen, Phosphorus and Sulfur Global Cycles, SCOPE 7, Ecol. Bull., (Stockholm) 22,89-134. Haines, B. (1983). Forest ecosystem S04-S input-output discrepancies and acid rain: Are they related? Gikos, 41,139-143. Haines, B., Black, M. and Fail, J. Jr. (1984). Potential dimethyl sulfide and ethyl mercaptan emissions from the floor of a tropical forest. Personal communication. Herron, M.H. (1982). Impurity sources of F-, Cl-, NO;- and SO~- in Greenland and Antarctic precipitation. J. Geophys. Res., 87, 3052-3060. Hesse, P.R (1957). Sulphur and nitrogen changes in forest soils of East Africa. Pl. Soil, 9, 86-96. Hill, F.B., Aneja, V.P. and Felder, R.M. (1978). A technique for measurement of biogenic sulfur emission fluxes. J. of Environ. Sci. Health, 3, 199-225. Jaeschke, W., Georgii, H.W., Malewski, H.C and Malewski, H. (1978). Contribution of HzS to the atmospheric sulfur cycle. Pageoph., 116,465-475. Johnson, D.W., Cole, D.W. and Giessel, S.P. (1979). Acid precipitation and soil surface absorption properties in a tropical and in a temperate forest soil. Biotropica, 11, 38-42. Jordan, CF. and Murphy, P.G. (1982). Productivity and mortality of two forest types in the Amazon territory of Venezuela. In Haines, 1983. Katoda, H. and Ishida Y. (1972). Production of volatile sulfur compounds by microorganisms. Ann. Rev. Microbial., 26, 127-138. Kellog, W.W., Cadle, RD., Allen, E.R., Lazrus, AL. and Martell, E.A (1972). The sulfur cycle. Science, 175,587-596. Lacaux, J.L., Baudet, J. and Servant, J. (1986). Unpublished results. Lanly, J.P. and Clement, J. (1979). Present and future forest and plantations areas in the tropics. FAO Report, Rome. Lawson, D.R and Winchester, S.W. (1979). Atmospheric sulfur aerosol concentra- The Atmospheric Sulfur Cycle 71

tions and characteristics from South American Continent. Science, 205, 1267- 1269. Legrand, M. (1980). Measurement of the acidity and electrical conductivity of Antarctic precipitation. PhD thesis No. 316, Laboratory of Glaciology, Greno- ble, France. Lodge, J.P., Machado, P.A, Pate, J.B., Sheesley, D.C and Wartburg, AF. (1974). Atmospheric trace chemistry in the American humid tropics. Tel/us, 26, 250-259. Logan, J.A, Prather, M.J., Wofsy, S.C and McElroy, M.B. (1981). Tropospheric chemistry: A global perspective. J. Geophys. Res., 86, 7210-7254. Mathieu, P. (1972). Chemical supply by rainwater in savannas and forests: Influence of the intertropical forested environment on ion migrations and solid matter transports. PhD thesis No. AO. 7838, Nice, France. Moller, D.M. (1984a). On the global natural sulphur emission. Atm. Environ., 18, 19-27. Moller, D.M. (1984b). Estimation of the global man-made sulphur emission. Atm. Environ., 18, 29-39. Moller, U. and Schuman, G. (1970). Mechanism of transport from the atmosphere to the earth's surface. J. Geophys. Res., 75, 3013-3019. Murozumi, M., Chow, T.J. and Patterson, C (1969). Chemical concentrations of pollutant lead aerosols, terrestrial dusts and sea salts in Greenland and Antarctic snow strata. Geochim. Cosmochim. Acta, 33, 1247-1284. Naughton, J.J., Lewis, V., Thomas, D. and Finlayson, J.B. (1975). Fume composi- tions found at various stages of activity at Kilauea volcano, Hawaii. J. Geophys. Res., 80, 2963-2966. Nguyen, B.C, Gaudry, A, Bonsang, B. and Lambert, G. (1978). Reevaluation of the role of dimethyl sulfide in the sulfur budget. Nature, 275, 637-639. Orsini, CO., Netto, P.A and Tabacniks, M.H. (1982). Preliminary data on atmospheric aerosols of the Amazon basin. Atm. Environ., 16-9, 2177-2181. Parrington, J.R, Zoller, W.H. and Aras, N.K (1983). Asian dust: Seasonal transport to the Hawaiian Islands. Science, 220,195-198. Penkett, S.A, Atkins, D.H.F. and Unsworth, M.H. (1979). Chemical composition of the ambient aerosol in the Sudan Gezira. Tel/us, 31, 295-307. Petit, C, Trinite, H. and Valentin, P. (1976). Study of turbulent diffusion above and within a forest: Application in the case of S02. Atm. Environ., 10-12, 1057-1063. Postgate, J.R. (1969). Recent advances in the study of sulfate reducing bacteria. Bacteriol. Rev. 29-4, 425-441. Prospero, J.M. and Bonnatti, E. (1969). Continental dust in the atmosphere of the eastern equatorial Pacific. J. Geophys. Res., 74, 3362-3371. Rahn, KA (1981). Relative importance of North America and Eurasia as sources of Arctic aerosols. Atm. Environ., 15, 1447-1456. Richards, P.W. (1964). The Tropical Rain Forest. Cambridge University Press, Cam- bridge. Riehl, H. (1954). Tropical Meteorology. McGraw Hill, New York. Robinson, E. and Robbins, RC (1975). Gaseous atmospheric pollutants from ur- ban and natural sources. In Singer, S.F. (Ed.) The Changing Global Environment: 111-123. D. Reidel, Dordrecht, Holland. Rodhe, H., Mukolwe, E. and Soderlund, R (1981). Chemical composition of precipitation in East Africa. Kenya Journal of Science and Technology (A) 2, 3-11. Ryaboshapko, AG. (1983). The atmospheric sulfur cycle. In Ivanov, M.V. and Freney, G.R (Eds.) The Global Biogeochemical Sulfur Cycle: 203-296. Wiley, New York. 72 Acidification in Tropical Countries

Savoie, D.L. (1984). Nitrate and non-seasaItsulfate aerosolsover major regions of the world ocean: Concentration, sources. PhD thesis, University of Florida, USA. Schlegel, H.G. (1974). Production, modification and consumption of atmospheric trace gasesby microorganisms. Tellus, 26, 11-19. Schwartz, S.£. (1984). Gas-aqeousreactions of sulfur and nitrogen oxides in liquid water clouds. In Calvert, I.G. (Ed.) SOz, NO and NOz Oxidation Mechanisms: Atmospheric Considerations. Butterworth, Boston. Seiler, W. and Crutzen, P.J. (1980). Estimate of gross and net fluxes of carbon between the biosphere and the atmosphere from biomass burning. Climatic Change, 2,207-247. Servant, J. (1986). The burden of sulphate layer of the stratosphere during volcanic quiescent periods. Tellus, 38B, 74-79. Servant, J., Delmas, R., Rancher, J. and Rodriguez, M. (1984). Aspects of the cycle of inorganic nitrogen compounds in the tropical rain forest of the Ivory Coast. J. Atm. Chern., 1, 391-401. Shreffler, S.H. (1978). Factors affecting dry deposition of SOz on forests and grass- lands. Atm. Environ., 12, 1497-1503. Shriner, D.S. and Henderson, G.S. (1978). Sulfur distribution and cycling in a deciduous forest watershed. J. Environ. Qual., 7,392-397. Staubes, R., Ockelmann, A. and Georgii, H.W. (1986). Emissions of biogenic sulfur compounds from various soils. Second International Symposium on Biosphere Atmosphere exchanges. Mainz, FRG, 16-22 March; Seiler, W., organizer. Sze, N.D. and Ko, M.K.W. (1980). Photochemistry of COS, CSz, CH3SCH3, HzS: Implications for the atmospheric sulfur cycle. Atm. Environ., 14, 1223-1239. Tamm, CO. (1976). Acid precipitation: Biological effects in soil and on forest vegetation. Ambio, 5, 235-238. Trudinger, P.A. (1969). Assimilatory and dissimilatory metabolism of inorganic sulphur compounds by micro-organisms. Adv. Microbiol. Physiol., 3, 111-118. Uematsu, M., Duce, R.A., Prospero, J.M., Chen, L., Merrill, J.T. and McDonald, R.L. (1983). The transport of mineral dust from Asia over the North Pacific Ocean. J. Geophys. Res., 88, 5343-5352. Varhelyi, G. (1985). Continental and global sulfur budget: Anthropogenic sulfur dioxide emissions. Atm. Environ., 19, 1029-1040. Visser, S. (1961). Chemical composition of rain water in Kampala, Uganda, and its relation to meteorological and topographical conditions. J. Geophys., 66, 3759- 3765. Zhao, O. and Sun, B. (1986). Air pollution and acid rain in China. Ambio, 15,2-5.