2.4.1. Carbon Dioxide Emissions Ranges (continued)

Some additional information about the range of future emissions can be obtained by examining the results from detailed analyses (see, e.g., Alcamo et al., 1996, 1998) conducted by Nordhaus and Yohe (1983), Edmonds et al. (1986), de Vries et al. (1994), and Manne and Richels (1994). These analyses confirm that the range is very large. The analysis of Edmonds et al. (1986) shows a range of 87 to 2 GtC by the year 2070 for the 95^th and 5^th percentiles, respectively; the range is from 27 to 4 GtC for the 75^th and 25^th percentiles. Nordhaus and Yohe (1983) estimated the range to be from 55 to 7 GtC for the 95^th and 5^th percentiles, and from 27 to 12 GtC for the 75^th and 25^th percentiles, respectively. In contrast, de Vries et al. (1994) estimated a fundamentally smaller range of 24 to 11 GtC by 2050 for the 95th and 5th percentiles, respectively, but they standardized some of the scenario assumptions, such as population and economic growth, relative to a base case.

The SRES team also applied the criteria presented in Section 2.2 to identify intervention scenarios from among the CO₂ emissions scenarios in the database. Of the 190 scenarios that reported CO₂ emissions through the year 2100, 62 were classified as intervention scenarios and 88 as non-intervention scenarios. For 40 scenarios the information available was insufficient to determine whether or not they included any climate policies. These are referred to as non-classified scenarios. The statistics for these three scenario samples are given in Table 2-2. The medians for these samples are also shown in Figures 2-1c to 2-1f. The analysis indicates that many of the 88 non-intervention scenarios are emulations of IS92 scenarios, and many of these are emulations of IS92a. For example, 42 of the 88 non-intervention scenarios were produced by EMF-14, of which 25 are based on IS92a.

Figure 2-1c shows spaghetti curves for the non-intervention scenarios found in Table 2-2. In the year 2100 these scenarios cover almost the same range (1.2 GtC to 59.4 GtC) as does the entire sample of 190 scenarios (-2.1 GtC to 59.4 GtC). The sample of non-intervention scenarios includes some that have very low emissions, which suggests that emissions can be low even in the absence of explicit climate intervention policies. However, the non-intervention sample also has a higher median than the total sample - 21 GtC in 2100 compared with the 15.7 GtC median found in the total sample. The lower median of the total sample may result from downward pressure exerted by the inclusion of some intervention scenarios in the total sample. However, it could also result from the influence of IS92a-like scenarios (upward pressure) in the non-intervention sample (see also Figure 2-1f).

Figure 2-1d likewise depicts spaghetti curves for the intervention scenarios. The entire range for the intervention scenarios is small compared to those of the no policy scenarios and the total set of scenarios. In 2100 the maximum (14.4 GtC) and the median (6 GtC) of this sample are significantly lower than maximum and median values of the full and non-intervention sets.

Figure 2-1e shows the range for scenarios that could not be classified into these two groups because of insufficient information in the database. The range of scenarios in this category is similar to that of the total sample in that it is very broad. The median of this set in 2100 is also similar to that of the total sample (15.7 GtC) and follows a similar trajectory.

Finally, the IS92a scenario (Leggett et al., 1992; Pepper et al., 1992) appears to have influenced subsequent emissions scenarios in the literature. The median population, gross world product, and primary energy consumption trends in many scenarios in the literature track very closely the developments in the IS92a scenario (Morita and Lee, 1998). This is because IS92a constituted a reference baseline that was emulated by different modeling groups in a number of scenario evaluation and comparison activities. Figure 2-1f shows the set of 35 IS92a-like scenarios that could be classified from the set of non-intervention scenarios. As these scenarios appear to emulate IS92a, they show little variation around the median of 20.3 GtC, which is about the emission level in that scenario (20.4 GtC) in 2100.

The analyses in the following sections focus on the total set of scenarios in the database only. The distinction between intervention and non-intervention scenarios applies only to analyses of CO₂ emissions.

IPCC Homepage