5.5. Aerosols and Ozone Precursors

In addition to the GHGs discussed above, aerosol particles and tropospheric ozone also change the radiative balance of the atmosphere, albeit in a spatially heterogeneous manner. Sulfate aerosol particles, which form as a consequence of SO₂ emissions, act as a cooling agent. Their net effect is quite uncertain, but is thought to offset the forcing from all non-CO₂ GHGs to date (Houghton et al., 1996). As interest in the role of sulfur has increased the since the previous IPCC assessment, and to encapsulate recent trends and expectations, sulfur emissions are discussed here in substantial detail. Nitrates, ammonia, organic compounds, and black carbon also contribute to the formation of atmospheric aerosols. Carbonaceous aerosols exert a small positive forcing effect, while the effects of other compounds and aerosols are less clear. Tropospheric ozone is a GHG also, with a small net positive forcing effect. Future tropospheric ozone levels will be determined by emissions of CH₄, CO, NO_x , and NMVOCs. The last three groups are reported and discussed here in a more aggregated and stylized form only, because these gases are short lived, their potential to form ozone is highly non-linear; NMVOCs are not distinguished by their reactivity, and data problems associated with including key sources in aggregated long-term models are large.

5.5.1. Ozone Precursors: Nitrogen Oxides, Non-Methane Volatile Organic Compounds, and Carbon Monoxide

5.5.1.1. Nitrogen Oxides

Emissions of NO_x primarily result from the combustion of fossil fuels. The NO_x concentration in exhaust gases depends on combustion conditions (temperature, residence time, air-to-fuel ratio, mixing) and varies widely across different applications. In particular, internal combustion engines used in road vehicles and ships have very high emissions, although new designs and exhaust-gas treatment offer much lower specific emission levels. Recent research (Davidson and Kingerlee, 1997; Delmas et al., 1997; Mosier et al., 1998) indicates that soil may be a significant source of NO_x emissions also. This source, however, is not included in the models used in the current report.

The 1990 NO_x emissions in the six SRES models range between 26.5 and 34.2 MtN, but not all the models provide a comprehensive description of NO_x emissions. Some models do not estimate NO_x emissions at all (MARIA, MiniCAM⁸), whereas others only include energy-related sources (MESSAGE) and have adopted other source categories from corresponding model runs derived from other models (i.e. AIM). Standardized (see Box 5-1 on Standardization) 1990 NO_x emissions in the SRES scenarios, measured as nitrogen, amount to 31 MtN (Figure 5-9).

As mentioned in Chapter 4, the volume of fossil fuels used for various energy purposes varies widely in the SRES scenario families. In addition, the level and timing of emission controls, inspired by local air quality concerns, is assumed to differ. As a result the spread is largest within the A1 scenario family, in which it is almost as large as the range across all 40 SRES scenarios. Up to the 2020s, all scenarios project rising NO_x emissions (Figure 5-9). The 25^th and 75^th percentile emissions corridor spans between 40 and 60 MtN by the 2020s, which can be interpreted as a "central tendency" among the entire spectrum of the 40 SRES scenarios. Beyond 2030, uncertainties in emission levels increase significantly. By 2100, the SRES range is between 16 and 150 MtN (i.e. emissions decrease by a factor of two or increase by a factor of five compared with 1990 levels). The median and mean emissions are tracked by a number of scenarios, most notably by B2- MESSAGE (B2 family marker) and A1B-IMAGE. In these scenarios, NO_x emissions tend to increase up to 2050 and stabilize thereafter, the result of a gradual substitution of fossil fuels by alternatives as well as by the increasing diffusion of NO_x control technologies. Low emission futures are described by various B1 family scenarios, whereas the upper bound for future NO_x emissions is represented by scenarios of the fossil fuel intensive A1 scenario groups (e.g. A1C- and A1G-MESSAGE) and the high population, high fossil energy A2 scenario family (A2-ASF, A2-MESSAGE, or A2G-IMAGE) (Figure 5-9).

The fossil fuel dominated A2-ASF (A2 family marker) with limited environmental concern has a rapidly increasing NO_x trajectory (Figure 5-9). Emissions in other A2 scenarios also continue to grow, except in A2-AIM for which emissions level off by the last decades of the 21^st century. In the A1B marker (A1B-AIM), the emissions growth is initially about as strong as in A2-ASF, but emissions peak in 2030, and decline as the fossil fuel share of total primary energy falls and the remaining fossil fuel technologies become more advanced (Figure 5-9). Scenarios from other A1 family groups that assume a much larger and continued role of fossil fuels yield much higher NO_x emissions, which reach 150 MtN by 2100 in the coal-based A1C-MESSAGE scenario. Emission growth in the B2 family scenarios is less steep than in the A1 family, but persists throughout the entire period, albeit at a declining rate. By 2100, emissions in the B2-MESSAGE scenario (B2 family marker) are about twice as large as in 1990 (Figure 5-9). B2-ASF has a similar trajectory, while the B2-AIM scenario has essentially constant NO_x emissions over the entire period. Emissions in the B1 marker (B1-IMAGE) are among the lowest of all the 40 scenarios (Figure 5-9). In this scenario, emissions increase stops around 2050 and subsequently declines toward the end of the 21^st century to 60% of the current level. Other scenarios within the B1 group coincide well with the B1 marker in 2100, although the maximum emission levels in these scenarios are much higher than in B1-IMAGE (Figure 5-9).

5.5.1.2. Non-Methane Volatile Organic Compounds

Figure 5-10: Standardized global emissions of NMVOCs for SRES scenarios, classified into four scenario families (each denoted by a different color code - A1, red; A2, brown; B1, green; B2, blue). Marker scenarios are shown with thick lines without ticks, globally harmonized scenarios with thin lines, and non-harmonized scenarios with thin, dotted lines (see Table 4-3). Black lines show percentiles, means, and medians for SRES scenarios. For numbers on the two additional illustrative scenarios A1FI and A1T see Appendix VII.

NMVOCs arise from fossil fuel combustion (as with NO_x , wide ranges of emission factors are typical for internal combustion engines), and also from industrial processes, fuel storage (fugitive emissions), use of solvents (e.g., in paint and cleaners), and a variety of other activities. As the chemical reactivities of the various substances grouped under the NMVOCs category are very different, so are their roles in ozone formation and the (potential) health hazards associated with NMVOCs. In this report NMVOCs are reported as one group. In 1990, the estimated NMVOC emissions range between 83 and 178 Mt, which after standardization (see Box 5-1) translates into 140 Mt (Figure 5-10). As discussed above for NO_x emissions, not all models include this emissions category or all of its sources; the most detailed treatment of NMVOC emissions is given in the ASF model.

A relatively robust trend across all 40 scenarios (see Figure 5-10) is a gradual increase in NMVOC emissions up to about 2050, as indicated by the 25^th and 75^th percentile corridor, with the range between 190 and 260 Mt by that year. Beyond 2050, uncertainties increase with respect to both emission levels and trends. As for NO_x emissions discussed above, the upper bounds of NMVOC emissions are formed by fossil fuel intensive scenarios within the A1 scenario family (e.g. A1B-ASF), and the lower bounds by the scenarios within the B1 family (with an important alternative higher scenario B1-ASF). Characteristic ranges are between 60 and 90 Mt by 2100 in the low emissions cluster and between 370 and 550 Mt in the high emissions cluster. All other scenario families and individual scenarios fall between these two emissions clusters, with the B2 marker scenario (B2-MESSAGE) closely tracking the median of global NMVOC emissions from all the SRES scenarios.

In the B1 family marker (B1-IMAGE) emissions gradually decline to 60% of the 1990 level by 2100 (Figure 5-10). The B1-AIM and B1-MESSAGE trajectories are similar, but differ somewhat from the B1 marker. They increase until the 2020s (AIM) or even the 2050s (MESSAGE) at modest rates, but subsequently decline to around or even below the B1 marker in 2100. The B1-ASF profile, however, is radically different, growing faster and continuously throughout the 21^st century. By 2100, the NMVOC emissions in B1-ASF increase to 350 Mt, or 2.5 times the 1990 level. This indicates the adoption in this scenario of fairly different assumptions with respect to technological development and emission controls as compared to the rest of B1 scenarios. In the A2 family marker (A2-ASF) and B2 family marker (B2-MESSAGE), emissions grow steadily until 2050, by which time they are around 50% higher than today. Thereafter, emissions in B2-MESSAGE decline gradually, while in the A2 marker the growth continues to reach 2.5 times the current level by 2100 (similar to B1-ASF). Emissions in A2-MESSAGE are fairly similar to those of the A2 marker, while emissions in A2-AIM decline to 170 Mt by 2100, only half of the A2 marker level. The trajectory in the A1B marker (A1B-AIM) is very distinct - up to 2060 a fast growth is observed to more than twice the 1990 level, after which emissions decline to the B2 family marker level (Figure 5-10). A decrease of NMVOC emissions in A1B-AIM after 2060 is explained mainly by the substitution of fossil fuels with renewables, especially in the transport sector. A very similar trajectory emerges in the A1B-MESSAGE scenario, while the A1B-ASF scenario emissions grow continuously up to 2100, by when they are the highest of the set at 550 Mt, almost four times the 1990 level.

5.5.1.3. Carbon Monoxide

CO emissions in 1990 are estimated to range between 752 and 984 MtCO across the models used to derive the SRES scenarios. The same caveats as for NO_x and NMVOC emissions (see above) also apply to CO emissions - the number of models that represent all the emission source categories is limited and modeling and data uncertainties, such as emission factors, are considerable. As a result, CO emission estimates across the scenarios are highly model specific, as indicated by the overlapping ranges of the four scenario families (Figure 5-11). From a standardized (see Box 5-1) 1990 level of 880 MtCO the range of future emissions is rather wide for both medium-term and long-term time horizons. By 2020, emissions range from 630 to 1550 MtCO, by 2050 they range between 470 and 2300 MtCO, and by 2100 the range is between 360 and 3760 MtCO (i.e. one order of magnitude difference between the highest and the lowest projections). Focusing on the 25^th and 75^th percentile intervals reduces uncertainty ranges somewhat, but nonetheless they remain substantial - between 1260 and 2300 MtCO by 2100. The median of all 40 scenarios is tracked quite closely by the B2-ASF scenario at the global level.

Emissions of CO follow rather different trajectories in the SRES markers than emissions of NMVOCs, except the A2 marker (A2-ASF). Starting from the standardized (see Box 5-1) level of 880 MtCO in 1990, emissions in the A1B marker (A1B-AIM) grow continuously and almost double between 1990 and 2100 (Figure 5-11). Emissions in A1B-MESSAGE (Roehrl and Riahi, 2000) increase at a higher rate than in the marker and reach thrice the current level by 2100. Emissions in A1B-ASF grow as fast as in A1B-MESSAGE until 2040 (1700 MtCO versus 1650 MtCO for A1B-MESSAGE), when the growth stops and emissions more or less stabilize. Emissions in the B1-IMAGE (B1 family marker) gradually decline to 40% of the current level by 2100 (Figure 5-11). Again, other B1 family scenarios have quite different trajectories and the patterns generated by AIM, ASF, and MESSAGE models for the B1 scenario family are similar to those derived by the same models for other storylines. The shape of the A2 family marker (A2-ASF) trajectory is very similar to that for NMVOCs, which suggests that these two substances are governed mainly by the same drivers and that similar assumptions with regard to improved emissions factors are adopted. The emission trajectory in the B2 marker (B2- MESSAGE) is just below the A2 marker trajectory and is about 2000 MtCO by 2100, more than twice the 1990 level. An increase of biomass energy use after 2050 becomes an important factor that affects emissions in this scenario. Other B2 family scenarios produced using the AIM and ASF models closely track the B2 marker, although emissions in these scenarios are somewhat lower, with those of B2-ASF at 1700 MtCO by 2100 and of B2-AIM falling just below 1500 MtCO.

IPCC Homepage