Scope 48 - Sulphur Cycling on the Continents

2.1 INTRODUCTION

This chapter has two objectives. The first is to describe briefly the atmospheric sulphur cycle-its processes, pathways and fluxes-to provide a context for the following chapters on sulphur cycling in terrestrial and wetland ecosystems. The second is to identify areas of difficulty in the study of the sulphur cycle which may arise, usually at the interface between different reservoirs in the cycle, and thus, at the interface between the different disciplines  involved. These may include a lack of understanding of an exchange process or flux between two reservoirs, or a mismatch between the needs and capabilities of two disciplines, as happens, for example, in the choice of spatial scales of study. Identification of such issues may lead to new opportunities for interdisciplinary cooperation, aiding in their resolution.

The chapter is selective in its emphasis, rather than being a comprehensive, in-depth treatise on the atmospheric sulphur cycle. A number of these already exist (e.g. Galloway et al., 1985; Ivanov and Freney, 1983). The intended audience is primarily non-atmospheric-those studying sulphur cycling in various terrestrial and wetland ecosystems, those concerned with nutrient or harmful acidic atmospheric inputs to a surface ecosystem, or those engaged in integrated ecosystem modelling.

The contents are arranged in the traditional sequence, proceeding through the parts of the atmospheric sulphur cycle-emissions, transport, transformations, and deposition-with emphasis on the first and last. Brief discussions of atmospheric sulphur budgets and atmospheric inputs to ecosystems are included to serve as linkages with following chapters.

2.2 EMISSIONS

On a global basis, natural emissions of sulphur into the atmosphere from biogenic sources and from volcanoes are perhaps some 60 Tg (S) a^-1 (2.0 Tmol a^-1), with an estimated range of 46 to 124 Tg (S) a^-1 (1.4 to 3.9 Tmol a^-1) (Andreae, 1989; Andreae and Jaeschke, this volume). The inclusion of sulphate produced by seaspray increases this estimate by a highly uncertain 35 to 315 Tg (S) a^-1 (1.1 to 9.8 Tmol a^-1), but some 90% of this is cycled directly back into the ocean. The approximate percentage contributions of the main sources of natural sulphur emissions are: for volcanoes, 18% ; open-ocean biogenic production, 46%; coastal zone and wetland biogenic sources, 3% ; terrestrial plants and soils, 13% ; biomass burning, at least 4%; and wind-raised dust, 16%. The magnitudes of all natural sources are highly uncertain, but aside from seaspray, the terrestrial plant and soil term is least well known. In recent years the estimated magnitude of this source has been decreasing as more representative measurements have become available. In large part uncertainties arise from the extrapolation of few point measurements to the global scale.

Natural emissions exhibit high spatial and temporal variability. Approximately one-half occur in the marine environment. Terrestrial biogenic emissions are correlated with temperature, and thus emission rates tend to be higher towards the tropics. As well, emissions from plants and soils are highly dependent on specific plant and soil type, and thus are highly variable geographically, often on a small scale. Volcanic emissions are of two kinds: those during non-eruptive periods contribute about 90% of the total, while those during eruptions contribute significantly less, on average.

Estimates of global anthropogenic sulphur emissions range from 60 to 110 Tg (S) a^-1 (2 to 3 Tmol a^-1) for the years 1976 to 1985 (Moller, 1984; Hameed and Dignon, 1988), comparable to but probably somewhat larger than natural emissions (excluding seasalt spray). The combustion of fossil fuels accounts for 80 to 85% of the total, with the remainder coming from the smelting of ores and other industrial processes and burning. The approximate 1980 regional apportionment of fossil-fuel sulphur emissions is: China and Japan, 16%; USA and Canada, 28%; Europe, 34%; the remainder of the Northern Hemisphere, 15% ; and the Southern Hemisphere, 7% (Hameed and Dignon, 1988). Annual emissions of sulphur in North America and Europe east to the Ural Mountains are known with an uncertainty of about 10-15% as a result of detailed emission inventories developed in recent acid rain programmes (Dovland and Saltbones, 1986; NAPAP, 1990). In these regions, emissions are resolved into a grid of approximately 100 km on a side. For other regions of the world, the spatial distributions of anthropogenic sulphur emissions are not as well known.

Historically, sulphur emissions in North America and Europe have shown a general upward trend from the early part of this century into the decade of the 1970s (Figure 2.1). As a result of slowing economies and a growing awareness of sulphur's role in acid deposition, sulphur emissions have been decreasing- from an estimated high of 17 Tg (S) a^-1 (0.53 Tmol a^-1) in North America in 1970 down to 12 Tg (S) a^-1 (0.34 Tmol a^-1) in 1985, and in western Europe (those countries participating in the European Monitoring and Evaluation Program, EMEP), from a high of about 26 Tg (S) a^-1 (0.81 Tmol a^-1) in 1979 down to 21 Tg (S) a^-1 (0.66 Tmol a^-1) in 1988 (Iversen et al., 1989).

Figure 2.1. (a) Sulphur dioxide emissions (i) for the United States to 1980 (solid line) and (ii) for the United States plus Canada (broken line). (b) Sulphur dioxide emissions for (i) Europe from Field (1975) updated by Semb (pers. comm. , 1986) (solid line) and (ii) for EMEP Europe for Iversen et al. (1989) (broken line). Units are Tg (S) a^-1

Of great interest now is how much sulphur emissions in different regions of the world will increase in the coming decades. Most members of the Economic Commission for Europe signed a Convention on Long-Range Trans-boundary Air Pollution, one of whose protocols stipulated a decrease of at least 30% in sulphur dioxide emissions from 1980 levels by 1993. Even though the full reduction has not yet been achieved, evidence is accumulating that ambient concentrations and sulphur deposition are decreasing in Western Europe (Leck and Rodhe, 1989; Mylona, 1989) and in North America (Dillon et al.,1988).

Emissions from Asia, South America, and Africa may well increase dramatically. Galloway (1989) used United Nations estimates of population growth for 1980 to 2020 to estimate future sulphur emissions. In one scenario, the per capita emissions were held constant at 1980 levels while population alone increased. In a second, per capita emissions were allowed to increase by a factor of four relative to 1980 values, along with population growth. Figure 2.2 shows increases from 1980 for the two scenarios of approximately 2 and 6 times for South America, 2 and 7 times for Africa, and 1.5 and 6.5 times for Asia. The conclusion is that dramatic increases in sulphur emissions are likely from these continents in the next few decades, and, as a result, globally. As Galloway points out, the consequences are likely to be severely increased acidification in tropical regions, and significant disruption of the biogeochemical sulphur cycle, particularly in the tropics and sub-tropics. With reference to biogeochemical cycling, it is informative to compare emission densities and the relative importance of natural and anthropogenic emissions on different spatial scales. On a global basis, the average emission density for sulphur is about 0.4 g (S) m^-2 a^-1 (0.01 mol m^-2 a^-1), and the ratio of anthropogenic to natural emissions is of order unity or slightly greater. By contrast, the average emission densities in the highly industrialized areas of Europe and North America may be higher by an order of magnitude, and the ratio of anthropogenic to natural emissions is approximately ten. In the high-emission grid squares of Europe or North America, the emission density may reach more than 40 g (S) m^-2 a^-1 (1.3 mol m^-2 a^-1), and natural emissions are likely to be less by a factor of 100 to 1000. Although these comparisons are illustrative only, it is important to note that much of the current sulphur cycling work is being done on continents, often at sites in close proximity to areas of high emission density where anthropogenic emissions greatly outweigh natural ones.

 Figure 2.2. The emission of sulphur to the atmosphere in 1980 compared to 2020 with no change and with a change in per capita energy consumption (Galloway, 1989). Units are Tg (S) a^-1

2.3 TRANSPORT AND TRANSFORMATION

From the perspective of the study of chemical cycling and surface-atmosphere exchange in terrestrial and wetland ecosystems, the atmosphere serves as a delivery system to bring material to and take material away from the immediate vicinity of an ecosystem. The transport of chemical constituents through the atmosphere is effected by the several different scales of motion in the atmosphere, each dominated by different physical processes.

On the global scale, transport is controlled by the general circulation of the atmosphere, the westerlies of the mid-latitudes and the easterlies of the subtropics, for example, and the quasi-stationary high and low pressure systems such as the Bermuda-Azores high and the Greenland or Aleutian low. The time and distance scales for such transport are several days and several thousand kilometres, respectively. On the next smaller scale, transport is governed primarily by synoptic disturbances such as the travelling low-and high-pressure systems that migrate across the mid-latitudes of the Northern Hemisphere. Characteristic time and distance scales are of the order of a few days and a few thousand kilometres, respectively. At increasingly smaller scales-regional, local, and micro, which are characterized, respectively, by time and space scales of several hours and hundreds of kilometres, an hour and a kilometre, and minutes and metres-transport is increasingly influenced by orography and the nature of the underlying surface. These also have a strong influence on the vertical temperature structure of the lower atmosphere and thus on the amount of convective activity and turbulent mixing, which in turn affect transport on these smaller scales. Extensive reference material is available on meteorological regimes and factors affecting the transport of chemical constituents through the atmosphere (e.g. Hasse, 1983; Merrill, 1986; Whelpdale and Moody, 1989).

Several factors can influence the effectiveness of atmospheric transport of material to a receiving ecosystem. Most important are the magnitude and location of upwind sources of emissions. On larger scales these may be large areas of biomass burning in the trade-wind belt in either hemisphere, or emissions from the large industrialized regions of the North American and European continents. Both prevailing wind direction and distance to sources are important. In addition, seasonal changes in meteorological regime can result in the transport of material from completely different source regions. On the smaller scales, in addition to wind speed, direction, and distance to upwind sources, local influences on transport may become dominant: examples are areas of high convective activity, regions of extensive turbulent mixing, land-sea breeze circulations, or orographic channelling. Even the diurnal changes in atmospheric stability, with efficient day-time vertical mixing and strong night-time stratification may result in two distinct transport regimes for a particular location.

The same meteorological factors which influence the delivery of chemical constituents to a terrestrial or wetland system also control the dispersal and transport of natural emissions from the system away through the atmosphere. Of initial importance is the efficiency of vertical mixing which takes substances emitted at the surface to regions higher up in the atmosphere where more efficient transport occurs.

As noted earlier, the main anthropogenic sulphur species released into the atmosphere are gaseous sulphur dioxide and particulate sulphate. Species emitted from natural sources are more numerous, and include primarily the reduced sulphur species hydrogen sulphide, dimethyl sulphide, dimethyl disulphide, carbonyl sulphide, carbon disulphide and methyl mercaptain, in addition to sulphur dioxide and particulate sulphate.

The reduced sulphur compounds undergo odixation in the atmosphere to sulphur dioxide, primarily through reaction with the hydroxyl radical. The expected atmospheric residence time of most is a day or less. Sulphur dioxide undergoes oxidation in the gas phase, primarily with the hydroxyl radical, on the surface of particles, or in precipitation elements, through oxidation with the peroxyl radical or ozone, or in the presence of metal catalysts. The product in all cases is the sulphate form either in particles or in cloud and precipitation droplets.

Sulphur dioxide has an atmospheric residence time of about one day; for sulphate it is three to four days; and for total sulphur it is about two days (Rodhe, 1978). Residence times vary depending on the chemical and meteorological regime in the region of the world in question. The chemical form in which sulphur exists in the atmosphere has a significant influence on the efficiency with which it is removed to the surface. Consideration of meteorological regime and residence time for sulphur in the troposphere provides some insight into the 'zone of influence' that sulphur emissions may have. For example, in the mid-latitudes where the average lower troposphere wind speed may be about 10 m s^-1, sulphur will be deposited over a range of 1000 to 2000 km. Such distances are, in fact, comparable to the scale of current North American and European acidification problems, to which sulphur is the major contributor.

2.4 DEPOSITION

Chemical constituents in the atmosphere can be brought to the surface by a variety of processes (Figure 2.3). Those for which precipitation is the delivery mechanism are termed wet deposition processes. Gases, such as sulphur dioxide, can dissolve in cloud and rain drops or adsorb on to frozen precipitation elements. Sulphate particles are efficient condensation nuclei and are incorporated into precipitation by nucleation or as a result of scavenging by cloud droplets and falling drops. The efficiency of wet removal of sulphur species depends on the form (sulphate is more efficiently removed than sulphur dioxide), and on the characteristics of the precipitation in a given location: type, intensity, duration, frequency, etc. In principle, wet deposition is easily measured. With proper attention to sampling location, cleanliness, sample handling and chemical analysis procedures, reliable measurements of wet sulphur deposition can be routinely obtained.

Figure 2.3. Schematic representation of the deposition processes

Those deposition processes which do not involve precipitation are collectively termed dry deposition. Particles larger than about 10 µm in diameter may be removed by gravitational sedimentation. Smaller particles and gases, however, are more efficiently brought to the near-surface region by turbulent atmospheric motions, where they may subsequently be brought into contact with surface elements by molecular-scale processes. Actual uptake is accomplished by chemical reaction, dissolution, adsorption, etc. The efficiency of transfer to the near-surface region depends primarily on turbulence intensity in the several metres above the surface. Transfer to the surface depends on the characteristics of the surface (e.g. roughness, wetness, chemical properties) and on the properties of the constituent (e.g. state, reactivity, solubility). Sulphur dioxide is dry deposited more efficiently than particulate sulphate because it is more readily taken up at the surface. In contrast to wet, dry deposition is not easily measured directly. It can be measured in a research mode by the flux-gradient or eddy-correlation approach; but these techniques are not used for routine network measurements. The more usual approach is to parameterize the dry deposition process using a 'deposition velocity', and calculate deposition as a product with a measured or modelled ambient concentration.

A third category of deposition process includes mass transfer to the surface by impaction of fog or cloud droplets, and by riming. These are not usually included as either wet or dry deposition. However, in high-elevation forested ecosystems or in areas with frequent fog, they can be very efficient. For example, Lovett, Reiners and Olson (1982) estimated that chemical inputs to a balsam fir forest in New Hampshire by this mechanism exceeded those by ordinary precipitation by 50 to 300% .Deposition by these processes and by dry deposition may be extremely efficient locally, for example on the wind-ward edge of a forested area. Techniques to quantify such deposition are limited, although some progress is being made by examining throughfall in forests.

In addition to the measurement of deposition, progress has been made in calculating deposition using atmospheric models. The important processes of deposition are reasonably well understood, but due to their complexity, they are parameterized in rather simple terms in most regional and larger scale models. In the case of wet deposition, the limitations to more sophisticated simulation lie in the fine temporal and spatial detail required in describing the cloud microphysics, and in the limited resolution available in the input fields (such as precipitation amount). Commonly, scavenging (or washout) ratio or scavenging (or washout) coefficient parameters are used in connection with rainfall amount and atmospheric concentrations to model wet deposition. Modelling limitations in the case of dry deposition are the detail required in the various process descriptions and the characteristics of the underlying surfaces (e.g. roughness, moisture, temperature, etc.). The deposition velocity parameter is derived on the basis of measurements, and is a function of atmospheric stability, underlying surface characteristics, and chemical species. A typical range of sulphur dioxide deposition velocity values for summer season day-time at mid-latitudes is 0.5-0.8 cm s^-1. For particulate sulphate, values under comparable conditions would be 0.1-0.4 cm s^-1. Values could well be an order of magnitude lower at night and in winter. Only limited modelling of fog, cloud, and rime deposition inputs has been carried out.

Estimates of wet and dry deposition of sulphur are available for most parts of the world. In much of the marine environment, in the Southern Hemisphere, and in other remote areas, such estimates are often based on very limited precipitation and ambient concentration measurements. In many developed countries of the Northern Hemisphere the phenomenon of acid deposition has prompted intensive routine measurement programmes which produce reliable wet deposition information and rather more limited estimates of dry deposition of sulphur. In eastern North America and west- central Europe annual wet deposition may range up to more than 4 g (S) m^-2 (0.13 mol m^-2), and annual dry deposition may exceed 5 g (S) m^-2 (0.16 mol m^-2). In contrast, Figure 2.4 (Galloway, 1985) shows that annual non-sea-salt sulphur deposition in remote continental areas of the world is approximately 0.2 and 0.1 g (S) m^-2 (6 and 3 mmol m^-2) for wet and dry deposition, respectively, and in remote marine areas, 0.1 and 0.02 g (S) m^-2 (3 and 0.6 mmol m^-2) for wet and dry deposition, respectively.

For the heavily industrialized regions of the world like eastern North America and west-central Europe, wet and dry deposition are approximately equal on an annual basis as an area average. Dry deposition is greater close to source regions where sulphur dioxide is the prevalent species, usually in relatively high concentrations (Figure 2.5). Wet deposition tends to dominate further from source regions as sulphate becomes dominant. Wet deposition is the main pathway to the surface for sulphur in remote regions of the world. As noted above, in special circumstances, deposition by the direct impaction of droplets on to vegetation may far outweigh the two major pathways.

Sea-salt sulphur is a major constituent of the global, and particularly the marine, sulphur cycle. However, most sea-salt sulphur (an estimated 90%) is more or less directly redeposited to the ocean (Andreae and Jaeschke, this volume). Of that which does reach the continents most is deposited within about 100 km of the coast. In many uses of atmospheric-sulphur data, sulphate concentrations in precipitation and air are corrected for the sea-salt component by subtracting off an amount determined by ratio to the sodium ion which is assumed to have an exclusively seawater origin.

Figure 2.4. Wet and dry deposition of sulphur species in remote areas (Galloway, 1985). Units are g (S) m^-2 a^-1 

 Figure 2.5. Monthly variation of sulphur deposition: (a) its wet and dry components; (b) the dry fraction; and (c) the fraction of dry as sulphur dioxide. The left panel is for Long Point, Ontario, a site close to strong emissions, and the right panel is for Kejimkujik, Nova Scotia, a site distant from sources (Sirois and Barrie, 1988)

In most well established networks which measure wet sulphur deposition on a routine basis, station density is in the range of one station per 10 000 to 100 000 km². Better resolution is occasionally achieved in special studies. Networks for ambient concentration measurements, from which dry deposition estimates are obtained, generally have a lower density of stations (an exception is the EMEP network in Europe). Temporal resolution in wet deposition measurements is usually daily or weekly; in ambient concentration measurements it is usually daily. Most regional-scale models which have been developed to address the problem of long-range transport of sulphur and acid deposition have a spatial resolution of 80 to 150 km which is dictated by the resolution of input data, and a temporal resolution of not much less than a day. However, with the exception of the complex grid models under development and testing at the present time, regional models are most suitable for use on a seasonal or annual basis, thus providing an effective temporal resolution of this degree.

A characteristic of both wet and dry deposition is their high variability. Data from several years of observation at Canadian stations show that both wet and dry deposition are highly episodic. It is not unusual for 50% of the annual deposition (wet or dry) to come from 20% of the events (or daily measurements) (Figure 2.6; Sirois and Barrie, 1988). Depending on possible ecosystem  response, this observation might lead to some concern about temporal and spatial scales of deposition that are smaller than those resolved by current measurements. Differences of an order of magnitude are not uncommon between consecutive daily measurements, although this may be less in remote regions. The variability arises primarily from differences in meteorological and scavenging processes, and the location of the measurement site with respect to sources. One implication of the high temporal and spatial variability in deposition measurements is that care must be taken to ensure that single sites often used in ecosystem studies are representative of the area for which the deposition data are being used. For example, dry deposition determined from ambient concentration measurements made near the edge of a forest may not be representative of a complete watershed; wet deposition measurements made on the crest of a hill with frequent upslope flow may not be representative of an entire catchment. A comprehensive short-term testing programme is often required to establish the representativeness of a proposed site.

In summary , for many parts of the world where the sulphur cycle has been studied in the context of the acid rain problem, extensive deposition data bases are available as are numerous model results. These provide useful large-scale definition of the sulphur deposition fields in these regions. However, in most other regions of the world, sulphur deposition estimates are based on limited measurements at relatively few sites. In the context of sulphur cycling in terrestrial and wetland ecosystems, one would want to have reliable, representative measurements of wet deposition, and ambient concentration at least, in the particular ecosystem of interest. Often this requires that such studies establish their own atmospheric measurement stations. In these cases, comparison studies should be undertaken with neighbouring network stations and use made of available modelled deposition fields.

 Figure 2.6. Total dry sulphate deposition at Chalk River, Ontario, ranked by percentage of total events (dark histogram), and percentage cumulative deposition (solid line). The horizontal dashed line shows the fraction of total deposition contributed by 20% of events (Sirois and Barrie, 1988)

2.5 BUDGETS

Freney, Ivanov and Rodhe (1983) summarized previous work on the global sulphur cycle. They showed that the atmosphere is one of the smaller reservoirs for sulphur, containing less than 5 Tg (S) (0.16 Tmol) at anyone time. Figure 2.7, from Freney, Ivanov and Rodhe (1983), is a schematic representation of the major fluxes of the global sulphur cycle. As more recent information has become available, individual flux terms have been updated. For example, the review by Andreae and Jaeschke (this volume), which summarizes most recent information on natural sources within the global sulphur cycle, suggests that volcanic and terrestrial biogenic emissions are lower by a factor of two, while marine biogenic emissions are twice as large, compared to values given in Figure 2.7.

Figure 2.7. The major natural and anthropogenic (circled) fluxes of the global biogeochemical sulphur cycle (Tg (S) a^-1) (Ivanov and Freney, 1983). The Roman numerals denote: I, output of sulphur-containing minerals; II, industrial treatment of the sulphur-containing raw materials; III, inland water bodies; and IV, volcanoes. The fluxes are denoted by: P1, fuel combustion and metal smelting; P2 and P17, volcanic emissions; P3, aeolian dust; P4, biogenic emission from land; P5, sea-air-land transport; P6, deposition of large dust particles; P7, wash-out and dry deposition; P8, land- air-sea air transport; P9, weathering; P10, river runoff to oceans; P11, transport to ocean in underground water; P12, runoff to inland water bodies; P13, input in fertilizers; P14, leaching of fertilizers; P15, efflux from chemical industries; P16, efflux of acid mine water; P18, abrasion of shores; P19, sea spray; P20, wash-out and dry deposition; P21, sedimentation of reduced sulphur; P22, sedimentation of sulphate; P23, biogenic emission from oceans

A budget of the inflows and outflows for a specified region is an instructive means of organizing knowledge of the cycling of sulphur through the atmosphere and other compartments of the environment. Ideally, all the important pathways into and out of the region of study would be known; all the individual processes that comprise each pathway would be understood; and some quantitative measure of the importance of each pathway, i.e. the flux of sulphur, would be available. All this information is usually not available, but budgets help to account for the various inflows and outflows, and help to identify and determine the importance of missing pathways.

The major input terms in an atmospheric sulphur budget are usually natural and anthropogenic inputs from the earth's surface, and a horizontal advection flux from upwind sources. Outflow terms are usually wet and dry deposition, and a horizontal advection outflow. On occasion, exchange at the upper surface of the atmospheric 'box' is taken into account. Uniform conditions are assumed within the atmospheric region chosen; i.e. smaller scale processes and variability within the region are not considered. For horizontal advection, the transit time for the region of the atmosphere chosen must be large in comparison to the atmospheric residence time of sulphur, in order not to have a unacceptably large error in the difference between inflow and outflow. This requirement means that global and regional budgets are the most suitable for this approach (Rodhe, 1978).

Figure 2.8 is an example of a regional atmospheric sulphur budget for eastern North America (Galloway and Whelpdale, 1980). The individual fluxes are shown, and it is of interest to note that anthropogenic emissions exceed natural ones by a factor of about twenty. Dry and wet deposition are approximately equal, and, as well, are similar to the horizontal outflow. That is, one-third of eastern North American emissions are deposited wet, one- third are deposited dry, and one-third flow out of the region to the east with the prevailing winds. The atmospheric sulphur budget for northern Europe gives similar results (Rodhe, 1978), as does a recent budget for southeastern Europe (Katsoulis and Whelpdale, 1990). A budget for the North Atlantic Ocean demonstrates another use of this approach (Figure 2.9; Galloway and Whelpdale, 1987). In this case the wet deposition to the ocean is known reasonably well, but dry deposition is quite uncertain (0.9 ± 0.9 Tg (S) a^-1; 0.03 ± 0.03 Tmol a^-1), pointing to the need for additional work to determine this term more accurately. In this case, the eastward flux from the region is determined by difference using all other budget terms. Regional budgets provide a useful approach to organizing and summarizing knowledge on this scale. This has been done for several regions of the world now, but additional work is needed for the Mediterranean, the eastern Atlantic, southeast Asia and the southwest Pacific.

On the spatial scale of many countries, resolution of the inflow and outflow advection fluxes becomes difficult because the transit time across the country is less than the atmospheric residence time of sulphur. The recent use of long- range transport models has provided a remedy to this limitation. Country budgets have been constructed for Canada, the United States, the Soviet Union, Hungary, Czechoslovakia, Poland and the United Kingdom. Within the EMEP programme, budgets have been determined for all European countries using a modelling approach (Eliassen et al. , 1988). Country budgets are useful in an accounting sense, identifying those terms in which there is most confidence and those in which there is least.

Figure 2.8. Schematic representation of the atmospheric sulphur budget for eastern North American c. 1980 (Tg (S) a^-1) (Galloway and Whelpdale, 1980)

 Figure 2.9. Schematic representation of the amospheric sulphur budget for the western North Atlantic Ocean atmosphere (Tg (S) a^-1) (Galloway and Whelpdale, 1987)

Budgets for terrestrial ecosystems and wetlands usually include reservoirs in addition to the atmosphere: for example, biota, surface waters, soil. The main atmospheric fluxes of interest are usually natural emissions ( and anthropogenic ones if present), wet deposition, and dry deposition. Unless the ecosystem is extensive the advection flux terms are not considered. Identification of the specific processes of natural emission, and quantification of these three flux terms is the subject of much of the remainder of the volume, and this topic will be touched on only briefly here.

2.6 ATMOSPHERIC INPUTS TO ECOSYSTEMS

In this section the magnitude and possible significance of atmospheric sulphur deposition to a variety of natural and managed ecosystems are considered. Of primary interest are the several terrestrial systems and wetlands discussed in subsequent chapters. These include natural systems such as temperate and tropical forestland, savanna, freshwater lake ecosystems, and freshwater and saline wetlands. Managed systems include rice paddies, managed forestlands and many other agricultural systems. Possible inputs of sulphur to these systems, in addition to the atmosphere, include weathering, groundwater, runoff, flooding, river inflow, fertilizer application, irrigation water, and mine drainage.

It is difficult to associate a spatial scale with many of these systems. Some are extensive, such as the temperate and tropical forest systems, covering large portions of continents. Some are numerous, such as temperate lake and freshwater wetland systems, but they not necessarily contiguous, being more 'patchy' in nature, and thus of a smaller characteristic size. Systems in regions remote from industrial emissions are likely to experience sulphur deposition of uniform magnitude, while those close to highly populated and industrialized areas, such as much of the inland lakes and freshwater wetlands in the temperate Northern Hemisphere, may experience a large range of atmospheric inputs. In the latter case, where these systems are studied as 'calibrated watersheds' or 'selected forest ecosystems', the magnitude of sulphur deposition and its role in biogeochemical cycling may differ significantly from location to location.

Andreae and Jaeschke (this volume) have reviewed wet deposition fluxes in tropical regions, primarily areas of rainforest and moist savanna. Wet sulphur fluxes range from 0.07 to 0.68 g (S) m^-2 a^-1 (2 to 21 mmol m^-2 a^-1), with a median value of 0.17 g (S) m^-2 a^-1 (5.3 mmol m^-2 a^-1). These values are in agreement with the estimate of Galloway (1985), 0.2 g (S) m^-2 a^-1 (6 mmol m^-2 a^-1), for wet deposition in remote continental areas. For such systems, atmospheric deposition is usually the only sulphur input to the system, aside from internal cycling.

A second example of atmospheric sulphur inputs is for northern temperate lakes, where once again, the atmosphere is usually the only source of sulphur . Although such lake systems cover a small portion of the earth's surface, there are hundreds of thousands of them in the northern temperate zone, typically of area up to some 100 ha. Atmospheric inputs, direct and from watershed runoff, are the major sources of sulphur to freshwater lakes (Table 2.1) ; these amount to between 0.6 and 2.6 g (S) m^-2 a^-1 (19 to 81 mmol m^-2 a^-1) for the examples given. The table also contains two unusual cases: one where sulphuric acid was added in a whole-lake acidification experiment, and another which is an acid-mine drainage reservoir. In such non-natural situations, these other inputs far outweigh the natural ones.

Table 2.1. Sulphur inputs to lakes, expressed per unit area of lake surface. Atmospheric inputs are those directly to the lake surface. Runoff is via overland (not channel) flow. Seepage is the estimated inflow by groundwater; only a small fraction (< 30%) of the total input of water to Little Rock Lake is derived from seepage 

An interesting comparison (R.B. Cook, pers. comm.) is the contribution of atmospheric and land sulphur to rivers, based on global river data from Holland (1978, pp. 67, 96, and 140). Holland calculates that of the global average river flow to the oceans of 2 g (S) m^-2 a^-1 (63 mmol m^-2 a^-1), about 40% was derived from the atmosphere, and about 60% was derived from terrestrial sources, primarily weathering of sulphate and sulphide minerals.

Some measurements of atmospheric sulphur deposition and comparisons with other inputs have been made for freshwater and saline wetlands (Giblin and Wieder, this volume). Freshwater wetlands cover about 5.3 x 10⁸ ha of the earth, and saline wetland about 0.5 x 10⁸ ha. Atmospheric inputs to freshwater bogs, compiled by Giblin and Wieder (this volume), range from 0.3 to 3.2 g (S) m^-2 a^-1 (9.4 to 100 mmol m^-2 a^-1). Some receive other inputs, such as groundwater or overland flow, but their magnitudes are not known. In saline wetlands, the atmospheric inputs are insignificant, because the salt-water source of sulphur dominates.

The importance of atmospheric inputs of sulphur to rice paddies and agricultural ecosystems is dependent on the degree of management of the system. In the case of systems that are neither intentionally flooded nor fertilized, the atmosphere may be the only source of sulphur (and other nutrients). Rice paddies in southeast Asia, for example, receive between 0.5 and 2.3 g (S) m^-2 a^-1 (16 and 72 mmol m^-2 a^-1) (Lefroy et al. , this volume). In some systems the fertilizer application may be calculated to supplement that delivered by the atmosphere. In others, where large amounts of fertilizer sulphur are applied, atmospheric inputs may not be important.

This section has dealt only with sulphur inputs from the atmosphere to various types of ecosystems. Emissions of reduced sulphur compounds from these various ecosystems to the atmosphere have been measured much less frequently. Estimates for sulphur gas fluxes from wetlands and terrestrial ecosystems are discussed elsewhere in this volume (Andreae and Jaeschke, this volume; Giblin and Wieder, this volume). Emissions to the atmosphere are generally small compared to rates of cycling within the ecosystem. With the possible exception of marine wetlands (saltmarshes and mangrove swamps) , emissions to the atmosphere are also small compared to atmospheric inputs to terrestrial and wetland ecosystems.

2.7 CONCLUDING REMARKS

This summary of the atmospheric sulphur cycle has attempted to present what is known about the cycle that is of most relevance to the study of sulphur cycling in terrestrial systems and wetlands.

Emission rates of sulphur to the atmosphere from anthropogenic sources are known with reasonable certainty for most industrialized regions of the Northern Hemisphere. Most problematic at present are those from southeast Asia, Africa and South America. Natural emissions are less well known because estimates are derived from scattered point measurements of sulphur fluxes; these demonstrate high temporal and spatial variability in both marine and terrestrial systems.

Extensive networks for the routine measurement of wet deposition are in place in many areas of the world. Data quality ranges from very good in some of the 'research-grade' and intensively managed networks, to less than acceptable in others. One must exercise caution about the quality of much of the available data. Dry deposition is determined from measurements of ambient concentration combined with a parameter, the deposition velocity. Such measurements are fewer than are those for wet deposition, and in general, the derived dry fluxes are less certain. Deposition by other processes, such as impaction of cloud droplets and riming, is determined only on a limited basis, even though the inputs may be very important in certain ecosystems. Models of atmospheric transport and deposition have improved greatly over the past decade. They are able to provide reliable calculated estimates of wet and dry deposition for large portions of Eurasia and North America. These models are perhaps under-used by the ecosystem community at present. Results from both networks and model applications are limited in spatial and temporal resolution. The characteristic variability of wet and dry deposition in time and space requires that care be taken in interpolating to small space and time scales for small-scale ecosystem studies.

In the context of using information from the atmospheric portion of the sulphur cycle for ecosystem studies, the more important unknowns are the magnitude of other than wet deposition fluxes, the magnitude of natural emissions in a specific location, the lack of comprehensive modelled deposition values for small scales, and-more a problem than an unknown-the mismatch between the scales of interest for much of the atmospheric and ecosystem work. Improved dry and other deposition flux values will require that additional effort be devoted to the development of measurement techniques for reliable field operation, and measurement strategies to elucidate deposition in special situations where such inputs may dominate. More reliable natural emission fluxes will also require improved techniques and many more measurements over a variety of representative surfaces. The two remaining issues, both pertaining to the scales of interest, require the development of common interest and understanding in the concerned disciplines.

The following are among the important future challenges for interaction between the atmospheric scientist and those working on other portions of the biogeochemical cycle of sulphur:

(1) Obtaining reliable estimates of likely changes in anthropogenic sulphur emissions-decreases in North America and Europe, and probable large increases in southeast Asia, Africa and South America.

(2) Evaluating the implications of climate warming for the sulphur cycle- changes in variability of atmospheric inputs, increased stresses on marginal systems, and increases in natural emissions as a result of temperature rise.

(3) Making more extensive use of 'integrated models', i.e. models which couple simulations of processes in all reservoirs of the sulphur cycle-atmosphere, biosphere, soil, waters-to enhance understanding of linkages within the cycle and the overall functioning of the biogeochemical sulphur cycle.

(4) Assessing the impact of current and planned reductions in anthropogenic sulphur emissions in regions of the world affected by acid deposition, to determine whether these reductions are adequate to protect the biogeochemical sulphur cycle from unacceptable perturbation.

REFERENCES

Andreae, M. O. (in press). The global biogeochemical sulfur cycle: a review. In: Moore, B., Schimel, D. and Eddy, J. Trace Gases and the Biosphere. Office for Interdisciplinary Earth Studies, UCAR, Boulder, CO.

Andreae, M. O. and Jaeschke, W. A. This volume.

Baker, L. A., Urban, N. R., Brezonik, P. L. and Sherman, L. A. (1989). Sulfur cycling in a seepage lake. In: Saltzman, E. and Cooper, W. (Eds) Biogenic Sulfur in the Environment, American Chemical Society, Washington, pp. 79-100.

Cook, R. B., Kelly, C. A., Schindler, D. W. and Turner, M. A. (1986). Mechanisms of hydrogen ion neutralization in an experimentally acidified lake. Limnology and Oceanography, 31,134-148.

Cook, R. B. (1981). The biogeochemistry of sulfur in two small lakes. PhD Thesis, Columbia University, New York, 248 pp.

Dillon, P. J., Lusis, M., Reid, R. and Yap, D. (1988). Ten-year trends in sulphate, nitrate and hydrogen deposition in central Ontario. Atmos. Environ., 22, 901-5.

Dovland, H. and Saltbones, J. (1986). Emissions of sulphur dioxide in Europe in 1980 and 1983. EMEP/CCC-Report: 1/86, Norwegian Institute for Air Research, Lillestrøm, 34 pp.

Eliassen, A., Hov, Ø., Iversen, T. Saltbones, J. and Simpson, D. (1988). Estimates of airborne transboundary transport of sulphur and nitrogen over Europe. EMEP/ MSC-W Report 1/88, Norwegian Meteorological Institute, Oslo, 80 pp.

Fjeld, B. (1975). Forbruk av fossilt brensel i Europa og utslipp av SO₂ i perioden 1900-1972. NILU Teknisk Notat Nr 1/76, 6pp.

Freney, J. R., Ivanov, M. V. and Rodhe, H. (1983). The sulphur cycle. Section 2.5. In: Bolin, B. and Cook, R. B. (Eds). The Major Biogeochemical Cycles and their Interactions. Wiley, New York, pp. 56-61.

Galloway, J. N. (1985). The deposition of sulfur and nitrogen from the remote atmosphere: background paper. Chapter 8. In: Galloway, et al. (Eds). The Biogeochemical Cycling of Sulfur and Nitrogen in the Remote Atmosphere, pp. 143-72. Reidel, Boston.

Galloway, J. N. (1989). Atmospheric acidification: projections for the future. Ambio, 18, 161-6.

Galloway, J. N. and Whelpdale, D. M. (1980). An atmospheric sulfur budget for eastern North America. Atmos. Environ., 14, 409-18.

Galloway, J. N. and Whelpdale, D. M. (1987). WATOX-86 overview and western North Atlantic Ocean S and N atmospheric budgets. Global Biogeochem. Cycles, 1, 261-81.

Galloway, J. N., Charlson, R. J., Andreae, M. O. and Rodhe, H. (1985). The Biogeochemical Cycling of Sulfur and Nitrogen in the Remote Atmosphere. Reidel, Boston, 249pp.

Giblin, A. and Wieder, D. (1990). This volume.

Hameed, S. and Dignon, J. (1988). Changes in the geographical distributions of global emissions of NO_x and SO_x from fossil-fuel combustion between 1966 and 1980. Atmospheric Environment, 22, 441-449.

Hasse, L. (1983). Introductory meteorology and fluid mechanics. Chapter 1. In: Liss, P. S. and Slinn, W. G. N. (Eds), Air-Sea Exchange of Gases and Particles. Reidel, Boston, pp. 1-51.

Herlihy, A. T., Mills, A. L. Hornberger, G. M. and Bruckner, A. E. (1987). The importance of sediment sulfate reduction to the sulfate budget of an impoundment receiving acid mine drainage. Water Resources Research, 23, 287-292.

Holland, H. D. (1978). The Chemistry of the Atmosphere and Oceans. Wiley, New York, 351pp.

Ivanov, M. V. and Freney, J. R. (1983). The Global Biogeochemical Sulphur Cycle. Wiley, New York, 470pp.

Iversen, T., Saltbones, J., Sandnes, H., Eliassen, A. and Hov, Ø. (1989). Airborne transboundary transport of sulphur and nitrogen over Europe-model descriptions and calculations. EMEP MSC-W Report 2/89, Norwegian Meteorological Institute, Oslo, 92pp.

Katsoulis, B. D. and Whelpdale, D. M. (1990). Atmospheric sulfur and nitrogen budgets for southeast Europe. Atmos. Environ., 24A, 2959-70.

Leck, C. and Rodhe, H. (1989). On the relation between anthropogenic SO₂ emissions and concentration of sulfate in air and precipitation. Atmospheric Environment, 23, 959-66.

Lefroy, R. D. B., Mamaril, C. P., Blair, G. J. and Gonzales, P. B. (1990). This volume.

Lovett, G. M., Reiners, W. A. and Olson, R. K. (1982). Cloud droplet deposition in subalpine balsam fir forests: hydrological and chemical inputs. Science, 218, 1303-4.

Merrill, J. T. (1986). Atmospheric pathways to the oceans. Chapter 2. In: Buat- Ménard, (Ed.). The Role of Air-Sea Exchange in Geochemical Cycling. Reidel, Boston, pp. 35-63.

Moller, D. (1984). Estimation of the global man-made sulphur emission. Atmos. Environ., 18, 19-27.

Mylona, S. N. (1989). Detection of sulphur emission reductions in Europe during the period 1979-1986. EMEP MSC-W Report 1/89, Norwegian Meteorological Institute, Oslo, 149pp.

NAPAP (1990). Emissions involved in acidic deposition processes. State-of-Science/ Technology Report 1, January 1990 draft, National Acid Precipitation Assessment Program, Washington DC.

Rodhe, H. (1978). Budgets and turn-over times of atmospheric sulfur compounds. Atmos. Environ., 12, 671-80.

Sirois, A. and Barrie, L. A. (1988). An estimate of the importance of dry deposition as a pathway of acidic substances from the atmosphere to the biosphere in eastern Canada. Tellus, 40B, 50-80.

Whelpdale, D. M. and Moody, J. L. (1989). Large-scale meteorological regimes and transport processes. Chapter 2. In: Knap, A. H. (Ed.). The Long Range Atmospheric Transport of Substances from Continents to Continents and Continents to Oceans, Reidel, Boston, pp. 3-36.