4.2.1 The Role of Trace Gases in Climatic Studies

The direct radiative effects of trace gases in the atmosphere are primarily due to their absorption in the infrared part of the spectrum. Most effective in this regard are gases with absorption in the parts of the spectrum where water vapour and CO₂ are almost transparent (813 µm), the atmospheric window. If absorption bands for different gases overlap, the effects of a changing concentration on the radiative transfer are reduced. Gases with no change in their abundance do not directly contribute to a climatic change, but may affect the magnitude of the role of other constituents due to such overlapping of absorption bands. They must therefore be included in radiative transfer computations with their fixed concentrations.

Measurements of concentration trends and sensitivity studies by means of radiation models have resolved that the strongest climatic change signals are to be expected from CFCl₃ (F 11), CF₂Cl₂ (F12), CH₄, N₂O and O₃ (see further Section 4.5.1). A number of other halogenated compounds and hydrocarbons altogether probably are less important than anyone of the above mentioned constituents and will not be considered specifically in the present context (cf Ramanathan et al., 1985).

In the following subsections we shall consider in some detail our knowledge about present atmospheric concentrations of these gases and the processes that regulate them. It should be stressed that the natural as well as anthropogenic sources and sinks of methane and nitrous oxide are still quite uncertain. In the account given below an attempt is made to be as internally consistent as possible, but considerable modifications may turn out to be necessary as more data become available.

4.2.2 Methane

4.2.2.1 Past and Present Concentrations

The presence of CH₄ in the atmosphere has been known since the 1940s when Adel (1939) and Migeotte (1948) observed strong absorption bands in the infrared region of the solar spectrum which they attributed to the presence of atmospheric CH₄. The first insitu measurements of tropospheric CH₄ mixing ratios were made in the late 1960s when sensitive analysis techniques had become available. Measurements indicated average tropospheric CH₄ mixing ratios of about 1.41 ppmv for the Northern Hemisphere and about 1.30 ppmv for the Southern Hemisphere (Ehhalt and Schmidt, 1978). Significantly higher values (1.61.8 ppmv) were observed in 1976 in the free atmosphere above Europe by Seiler et al. (1978), who attributed the difference to a possible ongoing increase of CH₄. The existence of an upward trend was clearly established by Rasmussen and Khalil (1981a) and has lately been confirmed by numerous measurements at different baseline stations (Fraser et al., 1981; Fraser et al.,1984) and by measurements of the meridional CH₄ distribution between the two hemispheres (Blake et al., 1982; Mayer et al.,1982). Figure 4.1 shows the longterm trend of atmospheric CH₄ in clean air at mid latitudes in the Northern Hemisphere. The individual data points have been obtained from measurements during aircraft and ship missions between 4060° N over Europe and the Northern Atlantic (W. Seiler, unpublished). Data obtained by Blake et al. (1982) and Blake (1984) at similar latitudes on the west coast of the United States in clean air from the Pacific Ocean have also been included. Despite the different sampling locations and sampling altitudes, these observations agree very well suggesting that they are representative for middle latitudes in the Northern Hemi sphere.

Figure 4.1 Trend of tropospheric CH₄ mixing ratios observed at midlatitudes of the Northern Hemisphere. Data are obtained from measurements carried out on aircrafts (circles), on ships (squares), and at different land based stations during clean air conditions (triangles). The figure includes data (dots) measured by Rowland and colleagues (see Blake, 1984) at similar latitudes in air from the Pacific Ocean

The data summarized in Figure 4.1 clearly demonstrate the existence of an average temporal increase of atmospheric CH₄ during the last ten years of about 18 ppbv per year or 1.1% per year. The deviation of the individual values from the average increase is probably due to seasonal variations (Rasmussen and Khalil, 1981b) and to meridional transport. Unpublished CH₄ data over the Atlantic from ship cruises between 50° N and 40° S (W. Seiler) show an average increase of 23 ppbv per year between 1977 and 1984 north of the InterTropical Convergence Zone (ITCZ) and 18 ppbv per year for the region south of the ITCZ, i.e. an increase of 1.4% and 1.2% per year, respectively. Similar trends have been found by Rasmussen and Khalil (1981a), Fraser et al. (1984), Blake (1984) showing that the long-term trend of atmospheric CH₄ is a global phenomenon.

Information on CH₄ trends cannot be obtained from direct measurements prior to 1975, because of a change of calibration (Heidt and Ehhalt, 1980). Using spectroscopic data, Ehhalt et al. (1983) concluded that the average CH₄ increase in the troposphere between 1948 and 1975 cannot have been larger than 0.5% per year. Considering the large uncertainties involved in the calculation of absolute CH₄ mixing ratios from absorption in the solar spectrum (see Fink et al.,1964), this conclusion should be considered with caution.

Better information on longterm trends of CH₄ can be obtained from analysis of air bubbles trapped in the ice sheets of Greenland and Antarctica. Figure 4.2 summarizes results reported by Robbins et al. (1973), Craig and Chou (1982) and Rasmussen and Khalil (1984). The age of the trapped air has been corrected for the time needed to close the firn by densification, i.e. about 90 years. Despite the uncertainties involved in the analysis of very small air volumes, the individual data sets reported by the different groups agree reasonably well and show relatively constant CH₄ mixing ratios of about 0.7 ppmv before 1700 AD. Samples from polar ice cores taken in Greenland (e.g. at Camp Century, Crete, Dye 33) show an approximately exponential increase of the tropospheric CH₄ abundance during the last 300 years (Figure 4.3). This increase correlates remarkably well with the increase of human world population which is shown in Figure 4.3 by filled dots, which indicates that the increase of the tropospheric CH₄ abundance is most likely related to anthropogenic activities, primarily agriculture.

Figure 4.2 CH₄ mixing ratios measured in air trapped in ice cores as function of time. Dots and filled triangles are data taken from Rasmussen and Khalil (1984) and represent values obtained from ice cores in Greenland and Antarctica, respectively. Open circles are data published by Craig and Chou (1982) and squares are data published by Robbins et al. (1973)

 Figure 4.3 Growth of human population and increase of atmospheric CH₄ mixing ratios during the last 600 years

4.2.2.2 Sources and Sinks

Methane is produced by microbial activities during the mineralization of organic carbon under strictly anaerobic conditions, e.g. in waterlogged soils and within the intestines of herbivorous animals. CH₄ is also released by anthropogenic activities such as exploitation of natural gas, biomass burning, coal mining. The CH₄ emission rates have been estimated by different authors, e.g. Koyama (1963), Ehhalt (1974), Ehhalt and Schmidt (1978), Sheppard et al. (1982), Khalil and Rasmussen (1983a), Seiler (1984), Blake (1984), and Crutzen (1983, 1985). The emission rates of the individual sources reported in the literature differ by more than one order of magnitude (see Table 4.1) which reflects the large uncertainties in estimating production rates. This uncertainty is primarily due to the limited data base from the individual ecosystems but also to the insufficient information on the size of individual ecosystems that contribute to the CH₄ emissions. Estimates are also difficult because of the complexity of methanogenic food chains which implies large spatial and temporal variations of the CH₄ emission rates within individual ecosystems. Thus, the CH₄ budget estimated for 1980 and shown in the last column of Table 4.1 should still be considered as tentative.

Most of the biogenic methane is released by ruminants, from rice paddies, and freshwater swamps and marshes. Only the releases due to ruminants appear to be reasonably well known. Early estimates by Hutchinson (1949) yielded a value for emissions by large herbivores in the 1940s of 45 Tg per year. Seiler (1984) has recently estimated that the production was 7299 Tg per year in 1975. This figure is based on the world population of domestic and nondomestic ruminants and accounts for different emission rates for different species.

Table 4.1 CH₄ emission rates from individual ecosystems (in Tg per year = 10¹² g per year). The last column presents new estimates made in this volume (W. Seiler)

Sources	Sheppard et al. (1982)		Khalil and Rasmussen (1983a)		Seiler (1984)	Blake (1984)	Crutzen (1985)	Seiler (this volume)
Ruminants	90		120		7299	71160a	60	70100
Paddy fields	39		95		3075	142190	120200	70170
Swamps/Marshes	39		150		1357	121190	7090	2570
Ocean/Lakes	65		23		717	1834			1535
Other biogenic	817b		100c		615	60397d
Biomass burning	60		25		5397	25110	2070	55100
Natural gas	50				1829		33	3040
Coal mining			40		30	62100	34	35
Other nonbiogenic	50				12			12
Total	1210		553		225395	5001160	400	300550
(a) including herbivorous insects
(b) including CH4 production from organic solid waste and natural ecosystems
(c) including CH4 production by termites
(d) including CH4 production from seasonal and tropical rain forests

The estimates of the CH₄ emission from rice paddies show large variations. The maximum value, i.e. 220 Tg per year (Ehhalt and Schmidt, 1978) was based on emission rates obtained from laboratory experiments (Koyama, 1963), whereas the minimum values, 3075 Tg per year (Seiler , 1984) were based on insitu measurements of the CH₄ release rates in Spanish and Californian rice paddies (Seiler et al., 1984a; Cicerone et al., 1983). Higher emission rates (70170 Tg per year) have recently been obtained by HolzapfelPschorn and Seiler (1985) by evaluating semicontinuous, insitu measurements of CH₄ fluxes from Italian rice paddies during a whole vegetation period. The lower CH₄ emission rates observed in the Spanish rice paddies are explained by the inflow of Mediterranean water containing sulphate which may have caused a reduction of the methanogenesis. Since more than 95% of the total harvested rice paddy area of 1.5 x 10¹² m² (FAO, 1983) is located in the Far East from where data on CH₄ emission rates are not available, the figure of 70170 Tg per year remains uncertain.

Estimates of the CH₄ emission from swamps and marshes are even more uncertain and in some cases speculative. Most of the reported figures are based on CH₄ emission rates measured during short time periods from small eutrophic ponds or swamps at midlatitudes which may not be representative of the large areas of swamps and marshes in tropical regions. The few measurements which have been carried out in natural, undisturbed freshwater wetlands show CH₄ emission rates that differ by more than two orders of magnitude (see e.g. Harriss et al., 1982) which makes a reliable estimate of the global CH₄ emission difficult. A more reliable figure may be obtained by making use of the annual CH₄ flux rates measured in rice paddies which on the average are 45110 g per m. This seems reasonable since most rice paddies and natural freshwater wetlands are located in similar climatic regions and show comparable net primary production rates.

The total area of swamps and marshes in 1980 was about 1.6 x 10¹² m² (Clark, 1982), which is comparable with the total harvested rice paddy area. About 75% of this area or 1.2 x 10¹² m² is located in the tropical and subtropical regions. These areas are only temporarily flooded and, therefore, only active in methanogenesis during part of the year. During the dry season, some fraction of these areas may even act as sinks for atmospheric CH₄ (see Harriss et al., 1982). The duration of the flooding period and the extension of the flooded area depend on the annual rainfall pattern and differ considerably between climatic regions. Large parts of the marshlands in the Amazon and Congo Basins are only flooded during half of the year. Even shorter flooding periods may occur in some swamp areas (e.g. Pentanal, Brazil; Okawengo, Botswana; Junk, private communication). As a first approximation, it is assumed that the total freshwater swamp and marsh areas are flooded on the average during 56 months, equivalent to an area of 0.50.6 x 10¹² m² being flooded throughout the year. Applying the average CH₄ emission rate of 45110 g m² per year, we derive a total CH₄ production from swamps and marshes of 2570 Tg per year. This figure does not take into account the fact that significant portions of these areas consist of unvegetated open water (Walter, 1973), which show considerably lower CH₄ emission rates than vegetated areas (Cicerone et al.,1983; Seiler et al.,1984a).

Blake (1984) recently reported an average CH₄ emission from swamps and marshes of 121 Tg per year. This value was calculated by applying the temperature dependence of CH₄ emission from a swamp at northern midlatitudes to a total swamp area of 2.6 x 10⁶ km². Using rather the value of 1.6 x 10⁶ km² reported by Clark (1982) and assuming that tropical swamps and marshes are only flooded during 56 months per year, the annual emission from these ecosystems is reduced to 3137 Tg per year which is in good agreement with our estimate of 2570 Tg per year. Similarly, figures ranging between 150 and 300 Tg per year (Koyama, 1963; BakerBlocker et al.,1977; Ehhalt and Schmidt, 1978) seem too large.

Methane is also formed by methanogenesis in oceans and lakes, in waterlogged tundra soils, and in human intestines. Together with the production of CH₄ by herbivorous insects these sources emit 722 Tg per year (Seiler, 1984, 1985). Another source is the production and release of CH₄ from decomposition of solid waste which might contribute as much as 10 Tg per year. This figure is based on an estimated total disposal of 250 x 10⁶ ton per year of organic municipal wastes and a CH₄ release rate of about 40 g CH₄ per kg (EPA, 1973; Cheremisinoff et al.,1980). Methane may also be released from digestors which are commonly used in Asian countries for production of biogas. Production rates and emission factors, however, are not known, so that their contribution to the global CH₄ budget cannot be quantified. Sheppard et al. (1982) proposed an additional CH₄ production from natural aerobic soils and estimated the total CH₄ emission from this source to be about 700 Tg per year. This figure is much biased by the experimental procedure used, which does not allow extrapolation to natural conditions. There is rather clear evidence that natural aerobic soils do not produce but decompose atmospheric methane (Keller et al.,1983; Seiler et al., 1984a).

Methane is also produced by abiogenic processes such as biomass burning (55100 Tg per year), leakage and venting of natural gas (3040 Tg per year), coal mining (about 35 Tg per year), fossil fuel combustion (12 Tg per year), i.e. 120180 Tg per year (Seiler, 1984). This figure agrees reasonably well with other estimates as summarized in Table 4.1.

The present total CH₄ release rate into the atmosphere in 1980 was thus 300550 Tg, the average being about 425 Tg per year. The global average tropospheric CH₄ mixing ratio of 1.55 ppmv in 1980 corresponds to a total mass of CH₄ of about 4,400 Tg, which implies an average residence time for CH₄ of about 10 years. This production must be equal to the total sink strength plus the amount of CH₄ needed to raise the CH₄ abundance at the observed rate. The following approximate balance between sources and sinks seems possible. The ongoing rise corresponds to about 50 Tg per year so that the required sink strength to balance the CH₄ cycle is about 375 Tg per year. The most important sink mechanism is the photochemical oxidation of CH₄ by tropospheric OH. Adopting an average OH concentration of 5 10⁵ molecules per cm³ (Crutzen and Gidel, 1983) yields a sink strength in the troposphere of 260 Tg per year. About 60 Tg CH₄ is transported annually into the stratosphere where it is oxidized. Furthermore, CH₄ is decomposed in the soil by methanotrophic organisms (Keller et al.,1983; Seiler et al.,1984a). The decomposition rate has been estimated to be about 20 Tg per year (Seiler et al.,1984a). These CH₄ sinks altogether account for 340 Tg per year, which is about the destruction required to balance the CH₄ budget.

4.2.2.3 Temporal Changes of Sources and Sinks

The observed increase of atmospheric concentrations of CH₄ due to an imbalance between sources and sinks has apparently increased over time. The sinks may to a first approximation be considered as first order reactions, i.e. they are approximately proportional to the atmospheric CH₄ concentrations. In the case of methane, the most dominant sink mechanism is tropospheric oxidation by the OHradical. It is possible that the OHabundance has decreased with time, particularly due to increasing concentrations of carbon monoxide, which has a significant influence on the distribution of tropospheric OH (Crutzen, 1983). Measurements carried out during more than one decade in the Southern Hemisphere (Seiler et al., 1984b) and in the free Northern Hemisphere troposphere over Germany and at Mauna Loa Observatory at 20° N (Seiler, 1985, unpublished data) indicate an upward monoxide trend of 0.61% per year which may have caused an average reduction of OH by about 0.2% per year. This is much less than the observed increase in atmospheric CH₄ indicating that increasing CH₄ production probably is the dominant cause for the observed increase. This also seems plausible in view of the observed ¹³C concentrations in atmospheric CH₄ (Craig and Chou, 1982).

There is strong direct evidence that the total CH₄ source has increased during the last decades due to increasing anthropogenic activities (see e.g. Seiler, 1984). Table 4.2 summarizes the production from different CH₄ sources calculated for selected years. The accuracy of the individual values in the table is of course less than indicated by the digits retained, which is, however , done to illustrate the trends as well as possible.

Table 4.2 Estimated trend of CH₄ emission rates from individual ecosystems (Tg per year)

The trend of CH₄ emission by ruminants and rice paddies has been based on data for the increase of the ruminant population and the annually harvested rice paddy area, respectively, as published in the yearbooks of the Food and Agriculture Organization (FAO). In the case of swamps and marshes it has been assumed that the total surface area of these wetlands has decreased with time due to draining and dredging. We assess that the total area has decreased by about 1.2% per year from about 2.5 x 10⁶ km² in the middle of the 1940s (Twenhofel, 1951) to 1.6 x 10⁶ km² in 1980 (Clark, 1982). This may have caused a reduction of the CH₄ emission from swamps and marshes from about 79 Tg per year in 1940 to the present value of 47 Tg per year.

The emissions from 'other biogenic sources' (cf Section 4.2.2) were 722 Tg per year in 1980 and probably changed only slightly during previous decades. The CH₄ emission from waste disposal on the other hand may have tripled between 1940 and 1980.

The temporal change of the CH₄ emission by biomass burning is based on statistics published by the Food and Agriculture Organization (F AO, 1983), and data reported by Seiler and Crutzen (1980). We conclude that the burning of industrial and fuel wood has increased by about 1.1% per year. The burning of agricultutal wastes probably increases at the same rate as food production (1.5% per year) and the burning due to shifting agriculture and deforestation may change at the rate of increase of rural population in tropical areas, i.e. 1.2% per year. In contrast, burning of savannahs and wild fires in temperate and boreal forests have been assumed not to change with time. On this basis, we deduce that the CH₄ emission from biomass burning has changed from about 49 Tg per year in 1940 to about 79 Tg per year in 1980.

The estimate of CH₄ emission because of transmission losses during distribution and use of natural gas has been based on the rate of natural gas production (Clark, 1982) assuming loss rates of natural gas to be 34%. Similarly the CH₄ emission by coal mining is assumed to be related to the total coal production. Increasing usage of natural gas and coal has then resulted in an increasing CH₄ emission from these two abiogenic sources from 21 Tg in 1940 to about 70 Tg in 1980.

The trend of the total CH₄ emission is shown in Figure 4.4. Although the absolute values of the individual annual CH₄ emission rates are uncertain, there is no doubt that the CH₄ emissions have increased during the last 40 years by about 140 Tg which corresponds to an average rate of about 3.5 Tg per year or 1% per year. The rate of CH₄ emission increase was highest between 1960 and 1975, when the cattle production, the harvested rice paddy area as well as the consumption of natural gas increased exponentially but has declined in recent years. If we assume that the atmospheric OH number density has not changed significantly during this period we may compute tropospheric CH₄ concentrations by assuming an approximate steady state at any time. The fact that the methane turnover time, about 10 years, is small compared with the doubling time of atmospheric CH₄ concentration implies that these values should be a reasonable first approximation of reality. We obtain values of about 1.04 ppmv for 1940, 1.12 ppmv for 1950, 1.26 ppmv for 1960 and 1.42 ppmv for 1970. The last figure is in fair agreement with the CH₄ amount given by Wilkniss et al. (1973), indicating that the calculated CH₄ emission rates as shown in Figure 4.4 are reasonable. It should be pointed out, however, that the turnover time for CH₄ may be changing with time due to changes of the abundance of the OHradical. Extrapolation of the CH₄ emission rates to times which were anthropogenically undisturbed yields a value of about 120 Tg per year which corresponds to a CH₄ mixing ratio of 0.5 ppmv compared to 0.7 ppmv as obtained from the analysis of air trapped in ice cores (Figure 4.3). Although approximate the above scenario of past changes is an internally consistent one.

Figure 4.4 Trend of estimated total CH₄ emission rates between 1940 and 1980 (for detail see text)

4.2.2.4 Longterm Trends

A continued increase of tropospheric CH₄ may have an important influence on tropospheric and stratospheric chemistry by changing the distributions and concentrations of other gases, e.g. ozone which is another greenhouse gas. On the other hand other gases emitted into the atmosphere by man may affect the abundance of the OHradical and thereby the methane concentration by decreasing the rate of methane destruction (cf Crutzen, 1985).

The reaction pathway for oxidation of CH₄ is strongly affected by the mixing ratio of NO. At NO mixing ratios higher than about 10 pptv in the lower troposphere, the oxidation pathway of CH₄ yields an average net gain of 2.7 ozone molecules and 0.5 OHradicals per methane molecule oxidized (Crutzen, 1985). With an atmosphere containing less than about 10 pptv NO the main oxidation pathway is rather one which may destroy about 3 OHradicals and 1.5 O₃ molecules per methane molecule oxidized to CO. The influence of NO may, therefore, lead to the interesting situation of increasing OH and O₃ concentrations in air with high NO mixing ratios, e.g. in the polluted atmosphere of the Northern Hemisphere and decreasing OH and O₃ concentrations in the cleaner Southern Hemispheric air. In fact, measurements indicate a significant increase of the tropospheric O₃ in the Northern Hemisphere in recent decades, whereas the O₃ abundance in the Southern Hemisphere seems to be constant (cf Section 4.2.5.2).

Increasing CH₄ mixing ratios might also have an impact on stratospheric chemistry, particularly due to the reaction with Cl forming HCl which is a stable compound in the lower stratosphere. HCl diffuses into the troposphere where it is removed by precipitation. Increasing CH₄ mixing ratios would lead to a reduction of the Cl abundance in the stratosphere and thus reduce the effect of active chlorine species which destroy ozone.

The oxidation of CH₄ in the stratosphere also provides a source of water vapour in the stratosphere. Increasing CH₄ mixing ratios in the troposphere will therefore cause higher stratospheric water vapour concentrations. Increasing water vapour mixing ratios might in turn cause a temperature increase in the stratosphere due to the absorption of infrared radiation.

4.2.3 Nitrous Oxide

4.2.3.1 Past and Present Concentrations

A secular increase of tropospheric N₂O abundance has been observed although the rate is considerably less than that of CH₄. The first reliable data showing such an increase were reported by Weiss (1981) who measured the N₂O mixing ratios in a large number of air samples collected between 1976 and 1980 at different locations in both the Northern and Southern Hemispheres. These data show an increase from 298 to 301 ppbv in the Southern Hemisphere and 299 to 302 ppbv in the Northern Hemisphere during this fiveyear period, corresponding to an average annual increase of 0.2% per year. Rasmussen and Khalil (1981b) reported an increase of 0.3% per year and most recently, Khalil and Rasmussen (1983b) also determined an increase of 0.3% per year based on continuous measurements at Cape Meares (45°N) and Tasmania (42°S).

These results agree very well with measurements by Seiler (unpublished data; 0.33% per year), obtained from continuous measurements at Cape Point (35 °S). Additional information is provided by measurements at 5 GMCC (Global Monitoring for Climatic Change) stations in both hemispheres which show an increase of 0.51.1 ppbv per year or 0.20.3% per year during a 6year period between 1977 and 1982 (Harriss and Nickerson, 1984).300 ppbv of N₂O is equivalent to 1.500 Tg N.

4.2.3.2 Sources and Sinks

Nitrous oxide is destroyed in the stratosphere almost exclusively by photolysis and reaction with 0(¹D). NO is formed, which in turn plays a major role in regulating the stratospheric ozone concentration and distribution. No tropospheric sink mechanisms of significance have been detected so far . The stratospheric destruction has been determined from the observed vertical N₂O profile in the stratosphere to be 8.49.4 Tg N per year (Johnston et al., 1979; Crutzen, 1983), corresponding to an average residence time for N₂O of about 170 years. A similar value (159 years) has been derived by Ko and Sze (1982) on the basis of twodimensional model calculations. The annual increase of N₂O by 0.20.3% implies that the total source strength should be 1214 Tg N per year. Estimates of N₂O emission rates published during the last decade are summarized in Table 4.3.

The flux of N₂O into the atmosphere is primarily due to microbial processes in soil and water and is a part of the nitrogen cycle. Until recently, denitrification (i.e. reduction of NO₃ to N₂) has generally been considered to be the major mechanism for N₂O production. Recent experiments, however, have shown that nitrification (i.e. oxidation of NH₄⁺ to NO₃) also plays an important role or may even be the dominant source. Furthermore, N₂O is not only produced but also destroyed in the soil and water, whereby an N₂O equilibrium value is established which is determined by parameters such as temperature, redox potential, pH value, etc. (see e.g. Seiler and Conrad, 1985). Therefore, reliable flux rates between ocean/soil and atmosphere can only be obtained by insitu measurements under natural conditions.

Table 4.3 Summary of N₂O emission rates from individual ecosystems (in Tg per year). The last column presents new estimates for this volume (W. Seiler)

Source	Hahn and Junge (1977)	Khalil and Rasmussen (1983b)	Crutzen (1983)	Seiler (this volume)
Ocean/Freshwater	16185	9.0	12	2
Natural soils	665	13.4	(?)	6
Fertilizer	620		<3	0.62.3
Gain of cultivated land		6.6	13	0.20.6
Fossil fuel burning	14		1.8
Biomass burning			12	12
Lightning	1055			<0.1
Total	80300a	29	812	1215
(a) includes N2O production from unknown sources

There is presently little information on the N₂O flux from natural, undisturbed soils. Measurements on unfertilized agricultural soils show high diurnal variations (Conrad et al., 1983; Slemr et al.,1984) and high spatial and seasonal variability (see e.g. Mathias et al., 1980; Bremner et al., 1980). N₂O emissions obtained by different groups vary between 2 and 36 g N m² h¹ in temperate climates (Keller et al., 1983; Seiler and Conrad, 1981; Goodroad and Keeney, 1984) and 435 g N m² h¹ in subtropical climates (Slemr et al., 1984). Based on these data, Slemr et al. (1984) estimated the global N₂O flux from unfertilized soils to be about 4.5 Tg N per year. This figure increases to about 6 Tg N per year if we account for new information on the N₂O production in tropical rain forests which indicate flux rates of about 43 g N m² h¹ (Keller et al., 1983).

The application of mineral nitrogen fertilizers leads to enhanced N₂O flux rates indicating that part of the applied fixed nitrogen is converted into N₂O and released into the atmosphere. Until recently it was generally assumed that about 1015% might be lost as N₂O. New data show, however, that this figure is much too high and that the loss as N₂O is strongly dependent on the mode of application and type of fertilizer. The highest loss rates were observed for anhydrous ammonia and ammonium fertilizers which supports the view that nitrification is the dominant N₂O production process. Results published by Breitenbeck et al. (1980), Conrad et al. (1983), Slemr et al. (1984) show average N₂O loss rates of 0.04% for nitrate, 0.150.19% for ammonium and urea, and 5% for anhydrous ammonia. These values seem to be independent of climate (Slemr et al.,1984) and may thus be representative for global conditions. Based on the total production rates of the different types of fertilizers, the global loss of mineral fertilizers in the form of N₂O is estimated to 0.52.0%. The same amount may be emitted from denitrification and/or nitrification of mineral fertilizers leaching from the fields into groundwater or surface freshwater ecosystems (Conrad et al.,1983; Kaplan et al.,1978). The total loss rate of fertilizer nitrogen then becomes 14%. With a total mineral nitrogen fertilizer consumption of about 60 Tg in 1980, the total N₂O emission due to application of nitrogen fertilizers amounted to 0.62.3 Tg N per year in 1980 with an average of 1.5 Tg N per year.

N₂O may also be released from soils due to the increase of land area used for agriculture. During the last decade the cultivated land area has increased by about 0.3% per year (FAO, 1983) or 5 x 10¹⁰ m² per year which compares well with the area of seasonal and tropical forests (5.3 x 10¹⁰ m²) cleared annually (Bolin et al., 1983). Revelle and Munk (1977) estimated the total land area converted into cultivated land during the last 100 years to be on the order of 8.5 x 10¹² m² which probably has caused an annual loss of soil carbon into the atmosphere of 0.10.4 x 10¹⁵ g C. With a nitrogen to carbon ratio of about 1:20 for soil organic carbon, this carbon loss is equivalent to a nitrogen loss of 520 Tg per year of which about 3% (Terry et al.,1980) or 0.20.6 Tg N may be lost as N₂O.

N₂O is also formed by nitrification in ocean water and emitted into the atmosphere. First estimates by Hahn and Junge (1977) indicated the total oceanic N₂O source to be 16160 Tg N per year. Much lower values (410 Tg N per year) have been obtained by Cohen and Gordon (1979) and more recently Hahn (1981) arrived at a value of 14 Tg N per year by reconsideration of his earlier data. Extensive oceanic observations in the Atlantic Ocean between 50° N and 40° S (W. Seiler, unpublished data) indicate an average N₂O flux from the ocean into the atmosphere of about 2 Tg N per year which supports an earlier suggestion by Weiss (1981) that the oceanic N₂O source is small compared to the stratospheric sink.

N₂O is also emitted into the atmosphere by anthropogenic activities such as fossil fuel and biomass burning. The first estimates of the N₂O emission from fossil fuel burning were reported by Weiss and Craig (1976) and Pierotti and Rasmussen (1976) who deduced emissions of 1.6 and 2.2 Tg N per year, respectively, at the beginning of the 1970s. Production rates for 1980 can be calculated by applying the N₂O/CO₂ ratios reported by Weiss and Craig (1976) to be 2.05 x 10⁴ for coal burning and 2.25 x 10⁴ for oil burning. The ratio for burning of natural gas is about 1 x 10⁴ and is calculated from data published by Pierotti and Rasmussen (1976). With an average figure of about 2 x 10⁴, the total N₂O emission from fossil fuel burning in 1980 becomes about 2.9 Tg N.

The N₂O emission from biomass burning was originally estimated by Crutzen et al. (1979) to be 8 Tg N per year, but was reduced to 12 Tg N per year when new N₂O emission factors from biomass burning became available (Crutzen, 1983).

Summarizing the N₂O production from all potential sources yields a likely total N₂O emission of 1215 Tg N per year. Considering the uncertainties of the estimates of the individual sources, the agreement of these production rates with the total source strength deduced from the rate of increase of atmospheric N₂O may be fortuitous. The contributions of the different sources may turn out to be markedly different as more data become available.

4.2.3.3 Temporal Changes of Sources and Sinks

Using the N₂O/CO₂ ratio in emissions from fossil fuel burning (Weiss and Craig, 1976) we estimate that the N₂O emission from fossil fuel burning may have increased from about 0.6 Tg N per year in 1940 to the present value of about 1.9 Tg N per year which corresponds to an annual increase of about 0.03 Tg N or 3% per year as an average during these 40 years.

Based on the production rates of mineral fertilizers (FAO, 1983), the annual N₂O emission rate has probably increased from about 0.01 Tg in 1950 to about 1.4 Tg in 1980. Because of the upward trend of biomass burning, the N₂O emission from this source may have almost doubled during the last 40 years from 0.61.1 T g N per year in 1940 to about 12 T g N per year in 1980.

Increasing amounts of N₂O should also have been released due to the increasing demand of food and correspondingly increasing area of cultivated land. Because of the growth of world population, the N₂O emission of this source may have doubled during the last 50 years from 0.3 Tg N in 1930 to about 0.6 Tg N in 1980.

Summarizing all N₂O emission rates, the total N₂O flux into the atmosphere may have changed from 89 Tg N per year during anthropogenically undisturbed conditions to the present value of about 14 Tg N per year (see Figure 4.5). The upward trend of N₂O emission rate increased from about 0.1% per year at the beginning of the century to about 1.3% per year during the last 10 years, primarily due to the rapidly increasing emissions from fertilizer application and combustion of fuels. If we accept a residence time for N₂O in the atmosphere of about 170 years and first order sink mechanisms, we can deduce that the emission scenario shown in Figure 4.5 should have caused an increase of the N₂O mixing ratio by about 8% since late last century and a present annual increase by somewhat less than 0.3%. The latter figure is in good agreement with observations. The undisturbed N₂O mixing ratio therefore probably was about 280 ppbv as compared to 301 ppbv in 1980.

Figure 4.5 Trend of estimated total N₂O emission rates between 1880 and 1980 (for detail see text)

4.2.3.4 Interaction with Other Atmospheric Trace Gases

In contrast to CH₄, the influence of increasing N₂O on the atmospheric chemistry seems to be restricted to the stratosphere where N₂O is destroyed by reaction with atomic oxygen. This leads to formation of nitric oxide which in turn reacts with stratospheric O₃ by a catalytic reaction sequence, leading to a reduction of the overall O₃ abundance in the stratosphere by 35% for a doubling of the N₂O mixing ratio, assuming otherwise present atmospheric conditions.

4.2.4 Chlorofluorocarbons (CFCs)

Past and Present Concentrations

The presence of chlorofluorocarbons in the atmosphere was detected in the early 1970s and caught considerable attention in 1974 because of their being a possible source for chlorine in the stratosphere and accordingly a possible threat to the ozone layer (Molina and Rowland, 1974). Soon thereafter, Ramanathan (1975) and Wang et al. (1976) pointed out the possibility that due to their infrared absorption bands the CFCs enhance the atmospheric opacity and contribute to the greenhouse effect. Ramanathan et al. (1985) have recently considered a series of chlorinated and/or fluorinated hydrocarbons in the atmosphere, almost all of them exclusively anthropogenic, with regard to their role in the radiative balance of the atmosphere. It is clear that CFCl₃ (F11) and CF₂Cl₂ (F12) are the most important ones and we are justified in primarily considering these two gases in the following analysis of possible present and future changes of climate due to their emissions.

The CFCs are produced for a variety of uses such as solvents, refrigerator fluids and spray can propellants. Although atmospheric measurements are available for only about a decade, past concentrations can be deduced with reasonable accuracy on the basis of production and emission figures as provided by the Chemical Manufacturing Association (CMA, 1982). Figure 4.6 shows past emissions of F11 and F12. A rapid increase until about 1970 changed to a decline during the latter part of the 1970s, which was caused by restrictions on the use of CFCs introduced by some countries because of their possible threat to the ozone layer. It should be noted, however, that during this period the nonpropellant use has continued to increase by about 4% per year, while the propellant use has decreased from 56% to 34% of the total CFC production. A marked increase of the total CFC use has been reported for 1983.

Figure 4.6 Historical emissions of CFC11 (CFCl₃) and CFC12 (CF₂Cl₂), (CMA, 1982). In the left figure the slopes corresponding to 5%, 10% and 20% annual increase are indicated (from CMA, 1982)

The CFCs are photochemically decomposed almost exclusively in the stratosphere. On the basis of the temporal changes observed in recent years in relation to emissions as well as overall budget considerations the atmospheric lifetimes of F11 and F12 have been estimated to be about 80 and 170 years respectively (Ko and Sze, 1982; Cunnold et al., 1983a, 1983b).

At the beginning of 1980 the average mixing ratio of F11 in the lower troposphere is estimated to have been 168 pptv and to have been increasing at an annual rate of 5.7%. Midlatitude values for the Northern and Southern Hemispheres seem to have been about 10 pptv higher and lower respectively than the global average which is due to the fact that most emissions occur in the northern middle latitudes (Cunnold et al., 1983a). The Fl2concentration is estimated to have been 285 pptv with an annual increase of 6% and a difference between the Northern and Southern Hemispheres by about 30 pptv (Cunnold et al., 1983b).

4.2.5 Ozone

4.2.5.1 Formation and Destruction

Atmospheric ozone varies considerably both in space and time as a result of interactions between atmospheric motions and chemical reactions. We know these reasonably well as a result of the development of complex models of relevant chemical reactions and long series of observations from the Earth's surface and since 1978 also from satellites (WMO, 1981; Mateer et al., 1980; Zerefos and Ghazi, 1985; IAMAP, 1985; NASA, 1984; McPeters et al., 1984).

It is now accepted that the concentration of tropospheric ozone is increasing due to photochemical processes. As was already discussed in Section 4.2.2.4, methane, carbon monoxide and nitrogen oxides (NO_x) play important roles in this context. Their increasing concentrations and reactions with the hydroxylradical are of prime importance for the ozone chemistry in the troposphere (Crutzen, 1985). Evidence for production of background O₃ in the troposphere has been reported. by Fishman and Seiler (1983). Quantitative information is, however, difficult to obtain because of our present inadequate knowledge about the relative importance of the processes involved. From analysis of photochemical sink processes for ozone Fishman et al. (1979) conclude that production of ozone probably is more important for the ozone budget of the troposphere than the influx of ozone from the stratosphere.

4.2.5.2 Tropospheric Ozone

Hartmannsgruber, Attmannspacher and Claude (1985) have analysed the ozone sonde data from Hohenpeissenberg (Federal Republic of Germany) during the period 1967 to 1980 and found an increase of about 60%, i.e. an average annual increase of 4%. Feister and Warmbt (1985) have observed a 61% increase of tropospheric ozone between 1956 and 1983, i.e. 2.3% per year. Dütsch (1985) has found an annual increase of 0.7% per year from an analysis of ozone sonde ascents at levels between 1.5 and 7 km over Payerne (Switzerland) during the period 19681984. Reiter et al. (1983) have observed an increase of the ozone volume fraction by 32% from 1973 to 1982 for GarmischPartenkirchen (Federal Republic of Germany), most of which occurred between 1980 and 1982.

Bojkov and Reinsel (1985) and Logan (1985) have recently analysed the numerous observations (those quoted above and by others) from many parts of the world made by ozone sondes and Umkehr measurements. The uncertainties of these observations are still considerable (Hilsenrath et al.,1985; WMO, 1982). Instrumental corrections, in exceptional cases more than 30%, need to be applied (De Muer, 1985). Many series of observations are from regions with considerable industrial activity and the global representativeness may be questioned.

Since the lifetime of the ozone molecule in the troposphere is comparatively short (a few weeks), considerable spatial variations occur and it is difficult to assess the possible global change of ozone. An increase has, however, clearly taken place at middle and high latitudes of the Northern Hemisphere during the last 23 decades, particularly during the summer months. Changes by 2050% have been recorded and the present rate seems to be 12% per year. Few long records are available from the tropics, but in areas of extensive biomass burning (South America, Africa, India) an increase seems also to occur. Although probably not representative for the hemisphere as a whole (and certainly not for the globe), the average annual change of up to 100 mb (»16 km) at Resolute (Canada) and Hohenpeissenberg shows the typical upward trend of ozone as a function of elevation at middle and high latitudes during the last few decades (Figure 4.7).

4.2.5.3 Stratospheric Ozone

Dütsch (1985) has extracted the following trends from ozone soundings over Payerne for the period 19681984;0.7% per year for the 12.516 km level,0.5% per year at 1921 km, +0.2% per year at 24 km, zero change at 27 km, and +0.1% per year at 30 km.

Measurements from 15 Umkehrstations (see Figure 4.8) published by Reinsel et al. (1984) indicate a reduction of the ozone amount above 35 km altitude and maybe a small increase below this level. A statistically significant trend for the period 19701980 of 0.2 to0.3% per year is indicated in the layer 3448 km. Changes at lower stratospheric levels do not seem to be statistically significant. In the data from Hohenpeissenberg a slight decrease is found between 14 and 24 km and a substantial decrease by 34% from 1977 to 1982 above 28 km corresponding to 0.7% per year. Satellite measurements between 1979 and 1982 also show a distinct change of roughly 5% at 1 mb (45 km) (Fleig et al., 1985; Planet et al., 1985). It should be emphasized, however, that rather rapid changes can occur at these upper levels, which should not necessarily be interpreted in terms of a trend.

Figure 4.7 Vertical distribution of the trend of ozone (in % per year) at Resolute (75°N, Canada, 19661979) and Hohenpeissenberg (47°N. Germany, 19701983). The horiwntal bars give the 90% confidence interval. The shaded area shows the range of the tropopause heights (from Logan, 1985)

Figure 4.8 Estimated trends (in % per year) over the period 19701980 at 13 Umkehrstations for each Umkehr layer, 59 using model which includes smoothed atmospheric transmission data. (Overall 95% confidence interval estimates are also indicated on diagrams.) From Reinsel et al. (1984)

4.2.5.4 Tolat Ozone

The world's longest Dobson observation record at Arosa indicates an increase of the total ozone by 23% between 1925 and 1942, a maximum around 1942, and a decrease by 12% between 1960 and 1981 before the El Chichon eruption suddenly decreased the total ozone by about 5% (Dutsch, 1985). From the Payerne soundings a change between 1940 and 1980 by about 3%, i.e.0.75% per decade can be deduced. Komhyr et al. (1985) report a change of about 3% per decade since 1970 at Mauna Loa. Since about 90% of the ozone is found in the stratosphere, the trends of total ozone primarily reflect changes at these higher levels.

4.3 LIKELY FUTURE CONCENTRATIONS OF ATMOSPHERIC GREENHOUSE GASES

4.3.1 General Guidelines

Projections of likely future atmospheric greenhouse gas concentrations are very uncertain, partly because their generation, circulation and destruction in the atmosphere are not well understood, partly because the way man will disturb these natural cycles in the future cannot be foreseen very well. Although we shall attempt to define as 'realistic' scenarios as possible, these should be considered to be serving as a basis for assessments of the sensitivity of the climate system to anthropogenic disturbances, rather than describing most likely future changes.

As has been pointed out above there are interactions between the key greenhouse gases which we are considering, not only directly by chemical interaction, but also indirectly in that many natural processes are temperature dependent and thus sensitive to an induced climatic change. This fact needs to be carefully considered.

A few general guidelines for the estimation of likely future concentration should be noted

• The longer the mean atmospheric lifetime of an emitted gas is, the more probable it is that it will be accumulating in the atmosphere and that the increase will remain about proportional to the emissions.

• Gases with short residence times in the atmosphere will reach an equilibrium concentration comparatively soon, which is approximately proportional to the rate of emission.

• lthough positive or negative feedback mechanisms often cannot be determined quantitatively, they may be considered qualitatively in assessing likely future changes.

In a recent paper Ramanathan et al. (1985) have considered altogether 47 atmospheric trace gases with regard to their possible contribution to the atmospheric greenhouse effect and its future change. Many of these only play a very minor role and will not be considered in any detail. We list in Table 4.4 the most important ones and also the estimates of future concentrations as given by Ramanathan et al. (1985). These will be discussed further below. Scenarios for CH₄, N₂O and CFCs have also been presented by Hoffman et al. (1983). Wuebbles et al. (1984) have published a Proposed Reference Set of Scenarios for Radiatively Active Atmospheric Constituents in which timedependent as well as steady state scenarios are considered. For model studies it is important that some standardized scenarios are developed in order to intercompare the relative accuracy of different radiation models. The reader is referred to the report of Wuebbles et al. (1984) for detailed description of the proposed scenarios.

 Table 4.4 Pre-industrial and present volume fraction, residence time and estimated future volume fraction of atmospheric greenhouse gases (based on Ramanathan et al., 1985)

	Tropospheric volume  fraction		Tropospheric  residence		Projected volume  fraction
Gas	Pre-industrial (ppbv)	1980  (ppbv)	time  (yrs)		2030  (ppbv)
CO2	275.103	339.103	(»7)		450.103
CH4	700	1550	10		2340	18503300
N2O	280	301	170		375	350450
CFCl3 (F11)		0.17	80		1.1	0.52.0
CF2Cl2 (F12)		0.28	170		1.8	0.93.5
CHClF2 (F22)		0.06	20		0.9	0.41.9
CH3CCl2		0.14	8		1.5	0.73.7
CF3Cl (F13)		0.007	400		0.06	0.040.1
CF4 (F14)		0.07	>500		0.24	0.20.31
		Projected change in per cent from preindustrial period
O3 Tropospheric			+12.5
O3 Stratospheric:	10km		+3.8
	22km		+4.5
	26km		+ 2.0
	30km		6.1
	34km		22.6
	40km		37.9
	50km		5.5

We make the following specific comments on the estimates of probable future concentrations of the greenhouse gases as given in Table 4.4.

4.3.2 Methane, CH₄

There is no doubt that the tropospheric CH₄ mixing ratio will continue to increase mainly due to anthropogenic activities in agriculture to meet the increasing demand for food. The rate is, however, difficult to predict as CH₄ is emitted by various sources whose individual future trends may differ considerably (cf Section 4.2.2). In addition, the future change of atmospheric CH₄ is also dependent on the anthropogenic emission of other trace constituents such as CO, NO_x, nonmethane hydrocarbons, etc., which will affect the abundance and distribution of OH.

We can deduce an approximate value for future mixing ratios by extrapolating the linear regression between tropospheric CH₄ abundance and total human population as observed in the past (Figure 4.3) using estimates for the future world population. In this way we find that the CH₄ mixing ratios may reach values of 2.0 and 2.5 ppmv, for 2000 and 2050 respectively. At steady state these mixing ratios require CH₄ emission rates of 550 to 700 Tg per year. Whether or not these figures will be reached obviously depends on the population growth rate and on the possible introduction of new varieties of rice plants with higher yields, the influence of application of mineral fertilizer to rice paddies, increase of fossil fuel burning and future CH₄ destruction rates in turn depending on emissions of air pollutants.

The value given by Ramanathan et al. (1985) in Table 4.4 is in close agreement with the projection above. A more rapid increase has been proposed by Harriss and Nickerson (1984) which leads to a doubling (i.e. about 3.3 ppbv) in 2050. Also Hoffman et al. (1983) project a more rapid increase. An argument for a faster increase is the possible release of continentalslope sediment methane clathrates due to oceanic warming (Revelle, 1983). Since the arguments for increasing CH₄ production are not based upon very accurate knowledge of all processes involved, the projection by Ramanathan et al. (1985) seems reasonable, although possibly somewhat low.

4.3.3 Nitrous Oxide, N₂O

As was shown in Section 4.2.3, N₂O emissions seem to have increased rather rapidly during the last few decades from merely about 0.1% per year at the beginning of this century to about 1.3% per year during the last 10 years. This has been predominantly caused by the increase of N₂O from fertilizer application and combustion of fuel. If we accept a continued exponential growth of emissions we derive a value of 375 ppbv for 2030 and 500 ppbv for 2050 (Weiss, 1981). There is, however, strong evidence for believing that the upward trend of anthropogenic N₂O emission rates will slow down because of less rapid increase of the use of fossil fuels and limitations of the total acreage of land used for cultivation. We must keep in mind, however, that the atmospheric residence time of N₂O is long (about 170 years) so that the N₂O mixing ratios will approach constant steady state mixing ratios only after a period of about 200 years even if the N₂O release rates would stay constant. Such a steady state mixing ratio would be about 500 ppbv if the present production remains constant at 14 Tg N per year. If we assume a doubling of the fossil fuel consumption within the next 100 years (cf Chapter 2) and a likely maximum fertilizer consumption of 250 Tg N per year, the total N₂O production rate may become as high as 20 Tg N per year. The atmospheric concentration would still only reach about 360 ppbv in 2030 assuming a residence time of 170 years (cf Section 4.2.3). We accept the projections by Ramanathan et al. (1985) as reasonable, but consider the upper limit (450 ppbv) as unrealistically high.

4.3.4 Chlorofluorocarbons, CFCl₃, CF₂Cl₂

Because of the comparatively long residence times of these CFCs only about 10% of the amounts emitted into the atmosphere has so far been decomposed. During several decades to come a large fraction of the accumulated emissions will remain in the atmosphere. Because of the continued increase of the non-propellant use of CFCs by about 4% per year (cf Section 4.2.4) the slowdown of emission increase during the last decade most likely is only a temporary one, if no new restrictions on future uses of CFCs will be agreed upon. The projections made by Ramanathan et al. (1985) correspond to an annual emission increase by about 3% from prevailing conditions in 1980 and seem in this sense reasonable, although this is much less an increase than during the 1960s and early 1970s (cf Figure 4.6). Wuebbles et al. (1984) have used a model of Wuebbles (1983) to estimate the steady state CFC concentrations which would gradually be reached with constant emissions at 1983 rates and find the values 0.6 ppbv and 1.3 ppbv for F11 and F12 respectively. An exponential increase starting with the observed rates of emission increase in 1980 (about 5% per year) yields on the other hand the extreme values of 2.1 and 3.0 ppbv in 2030 and 5.6 and 7.8 ppbv in 2050, respectively.

A rapid increase of CFCs might gradually have a profound influence on stratospheric ozone. Concentrations above 12 ppbv may reduce ozone concentrations by 10% or possibly more. In view of the recent progress in agreement on an international ozone convention, it seems plausible that some restrictions on the use of CFCs will be agreed upon and that the emission scenarios given in Table 4.4 exaggerate likely future increases. It is still of considerable interest to carry out sensitivity analyses with regard to the role of F11 and F12 in changing the global radiation climate using the scenarios given in Table 4.4.

A few other halocarbons that may have a slight effect on radiative transfer through the atmosphere have also been listed in Table 4.4. Their likely effects are comparatively small and we need not consider their role further in the present context. The interested reader is referred to Ramanathan et al. (1985).

4.3.5 Ozone, O₃

The assessment of likely future changes of ozone is difficult because of the complex photochemical interactions that occur. All the gases discussed above play a role, and in addition other chlorinated species, since chlorine formed in their photolytic decomposition interacts with the ozone molecule. There is also a feedback through the temperature changes that the different greenhouse gases bring about. The results obtained by DeRudder and Brasseur (1985), Owens et al. (1985) and Ramanathan et al. (1985) differ in some important respects. The former two arrive at considerably larger enhancement of tropospheric ozone than do Ramanathan et al. (1985). The differences depend both on the treatment of the chemistry and transport of the nitrogen catalysts and other possible chemical interactions, as well as different assumptions about likely future emissions of the whole series of the other trace gases that are of importance. We shall adopt their scenario (Table 4.4) for the following analysis. Tropospheric ozone is accordingly considered to increase on the average by 0.25% per year, which is considerably less than has been observed during the last two decades (cf Section 4.2.5). The upward trend is assumed to decrease above 9 km and becomes zero at 27 km. At higher elevation a substantial decrease is foreseen, primarily induced by CFCs and other chlorinated trace gases.

4.4 RADIATIVE EFFECTS OF GREENHOUSE GASES

4.4.1 Some Principal Considerations

The greenhouse gases we are concerned with affect the radiation field principally in the same way as does CO₂, i.e. increased amounts reduce the direct emission of infrared radiation from the Earth's surface to space and increase the emission from stratospheric levels. There are, however, some differences. Due to photochemical reactions the mixing ratio of most of these greenhouse gases decreases more rapidly above the tropopause than does CO₂. This leads to a more rapid decrease of the cooling rate between 15 and 35 km.

Ozone has a vertical distribution that is very different from those of the other greenhouse gases, which must be accounted for. Ozone also has an absorption band in the solar part of the spectrum. This absorption has the opposite effect of that due to radiation transfer in the infrared part.

In assessing the temperature changes that may result from changing concentrations of these greenhouse gases their contributions must be dealt with simultaneously because their absorption bands partly overlap. A radiation model that can be used for such computations necessarily becomes complex. It has not yet been possible to incorporate at the same time possible feedback mechanisms that are present in the real climate system, and which have been shown to be of importance by experiments with general circulation models (GCMs). So far only one -dimensional models have been employed. By comparison of experiments to assess the effects of increasing concentrations of CO₂ using GCMs and one-dimensional, radiative-convective models the latter can be corrected for ice-albedo and cloud feedback mechanisms not explicitly accounted for (cf Section 4.4.3 below).

4.4.2 Model Computations and the Importance of Spectral Overlapping

A most detailed analysis of the changes to be expected from the simultaneous increase of a number of greenhouse gases in the atmosphere has been presented by Ramanathan et al. (1985). They employ a one-dimensional, radiative-convective model which includes latent heat flux between the Earth's surface and the atmosphere, they improve the methods for transmission computations, and analyse in some detail the errors introduced by using different resolutions in describing the spectra for the different gases. In their analysis of the role of overlapping absorption bands, first suggested by Tannhuser (1968), they make use of the Malkmus narrow band model. They have, however, omitted the temperature dependence of hot band strength for gases other than CO₂, CH₄, and N₂O.

The effect of band overlapping on the computed temperature increase at the Earth's surface is shown in Table 4.5 for a series of gases considered in this analysis. Two different temperature increases are presented: the ('skin') temperature of the Earth's surface (T_g) and the air temperature at the surface (T_s). The table also shows T_g computed for each gas separately, i.e. without consideration of overlapping. T_g is generally about 10% larger than T_s which presumably is due to more direct heating of the atmosphere by radiative processes than by sensible and latent heat flux from the Earth's surface to the atmosphere. We further notice the marked differences between different gases with regard to overlapping absorption bands. The roles of CH₄ and N₂O are reduced by about 50% particularly due to overlap with water vapour. The CFCs on the other hand, have primarily absorption bands in the 'window region " where absorption by H₂O and CO₂ is weak. There is only 1020% reduction of their role due to overlapping. For further details reference is made to the original publication by Ramanathan et al. (1985).

Table 4.5 The effect of overlapping of absorption bands on the computed surface warming. T_s and T_g are the changes of atmospheric temperature at the Earth's surface and of ground temperature respectively

Constituent	Change of  greenhouse	With Overlapa		Without overlap
	gas	Ts  (ºC)	Tg  (ºC)	Tg  (ºC)
CH4	1.25x	.09	.08	.19
N2O	1.25x	.12	.10	.20
CF2Cl2 (F12)	01 ppbv	.16	.15	.19
CFCl3 (F11)	01 ppbv	.14	.13	.16
CF4 (F14)	01 ppbv	.06	.05	.12
CF3Cl (F13)	01 ppbv	.22	.20	.25
CH2Cl2	01 ppbv	.03	.02	.04
CHCl3	01 ppbv	.06	.06	.09
CCl4	01 ppbv	.08	.07	.11
C2H2	01 ppbv	.02	.02	.07
C2F6 (F16)	01 ppbv	.13	.12	.20
CH3CCl3	01 ppbv	.02	.01	.03
PAN	01 ppbv	.04	.03	.06
a) the numbers have been rounded off. When the T is less than .02
K the result may be uncertain by as much as ± 30%.
Ts is surface-air temperature change and Tg is surface (or ground) temperature change.

4.4.3 Climatic Effects of Projected Increases of Other Greenhouse Gas Concentrations

The analysis of likely climatic effects of increasing greenhouse gas concentrations is a complex problem and requires the use of general circulation models (GCMs) as discussed in Chapter 5. No such experiments with enhanced concentrations of CH₄, N₂O or CFCs have, however, been made. We shall therefore express the climatic effects due to increasing concentrations of these gases in comparison with those deduced for increasing CO₂ concentrations by using the results from experimentation with a one-dimensional model of the atmosphere-ocean system (Ramanathan, 1980). Cloud and icesnow-albedo feedbacks as well as other more complex interactions can, however, only be analysed by using three-dimensional general circulation experiments. As will be described in Chapter 5, such more elaborate experiments have been conducted to determine the likely importance of such processes in assessing the global temperature change due to doubling of CO₂. It is found that the most likely change becomes 3.5± 2.0 °C, which we compare with a warming by merely 2.1 °C, if employing a one-dimensional model (see Chapter 5). Based on this comparison and the assumption that similar modifications of the climate system would be found, if changing concentrations of the other greenhouse gases were introduced into GCMs, we shall adopt a factor 1.7 ± 0.9 to translate the results from one-dimensional models to changes expected in more complex GCMs, to take into account in an approximate way the feedback processes mentioned above.

The model results referred to above concern the likely change into a new equilibrium following an instantaneous change of greenhouse gas concentrations in the atmosphere. As is also pointed out in Chapter 5, a considerable delay of such a change can be expected because of the inertia of the climate system primarily due to the oceans. In the present situation, with gradually increasing concentrations of the greenhouse gases, merely about half of the equilibrium change may have been realized. For a more extensive discussion of this problem reference is made to Chapters 5 and 6.

With the aid of a one-dimensional model Ramanathan et al. (1985) have shown that CH₄, N₂O, CFCl₃, CF₂Cl₂ and ozone account for more than 90% of temperature changes due to other greenhouse gases than CO₂. The most important other gases have been listed in Table 4.4 with rough estimates of plausible future concentrations.

It is first of interest to assess the temperature changes that may have been induced by past increases of the concentrations of these other greenhouse gases. On the basis of the changes given in Table 4.4. (except that the preindustrial CH₄ concentration has been put equal to 1.15 ppmv) Ramanathan et al. (1985) deduce with their one-dimensional model an increase of 0.27 °C as compared with a CO₂ -induced increase of 0.52 °C, i.e. a total increase of 0.79 °C. Referring to our discussion above, this would correspond to a total change of 0.62.1 °C if employing a GCM, of which 0.31.1 °C should have been realized if accounting also for the inertia of the climate system (cf Chapters 5 and 6). About one third of this change is due to the increase of the concentrations of other greenhouse gases.

Ramanathan et al. (1985) next compute the likely changes of the equilibrium temperature between 1980 and 2030, using the projected greenhouse gas concentrations (including uncertainties) as given in Table 4.4. The results Ramanathan et al. obtained with their one-dimensional model are shown in Figure 4.9. The increase of CO₂ from 340 ppmv to 450 ppmv would bring about a temperature change of 0.7 °C and that due to the increase of other greenhouse gases becomes 0.31.8 °C with a most plausible value of 0.8 °C. If we add past changes as computed above and also consider the uncertainty of future CO₂ changes (cf Chapter 3, Figure 3.2) we obtain:

Computed equilibrium temperature change due to increasing greenhouse gases 18902030 (onedimensional model). The indicated uncertainty is due to uncertainty of future projections of gas concentrations.

Carbon dioxide	1980	0.5 °C
	19802030	0.7 °C	(0.31.1°C)
Other greenhouse gases	1980	0.3 °C
	19802030	0.8 °C	(0.31.8 °C)
Total		2.3 °C	(1.43.7 °C)

Figure 4.9 Cumulative equilibrium surface temperature warming due to increase in CO₂ and other trace gases, from 1980 to 2030. After Ramanathan et al., 1985

We note that the value 2.3 °C is about the same as the expected temperature change for doubling of CO₂ (as deduced with the aid of the one-dimensional model). The uncertainty due to the inadequate knowledge of future greenhouse gas concentrations is considerable. Accounting also for other feedback mechanisms in the climate system that present GCMs are able to describe and as discussed above we deduce an increase of the global equilibrium temperature of at least 1.1 °C (slow increase of CO₂ as well as other greenhouse gases), more probably about 3.54.0 °C and possibly even more (if the high scenario projections of greenhouse gas concentrations are realized). The uncertainty range and delay due to the inertia of the climate system is discussed further in Chapters 5 and 6.

A few additional comments with regard to the role of the other greenhouse gases are of interest. CFCs may have a warming effect at the tropical tropopause which in turn might lead to an enhanced flux of water vapour into the stratosphere. The vertical distribution of the temperature change is also different from that due to CO₂ alone (Figure 4.10), and the implications should be further analysed.

Figure 4.10 Change in the vertical distribution of temperature due to an increase of CO₂ alone and CO₂ with all other trace gases listed in Table 4.5. (After Ramanathan et al., 1985)

The most important contributions to the heating come from CFCl₃, CF₂Cl₂, N₂O, and CH₄