9.3.3.1 Implications for temperature of stabilisation of greenhouse gases

Figure 9.16: Simple model results: Projected global mean temperature changes when the concentration of CO2 is stabilised following the WRE profiles. For comparison, results with the original S profiles are also shown in blue (S1000 not available). The results are ensemble means produced by a simple climate model tuned to seven AOGCMs (see Appendix 9.1). The baseline scenario is scenario A1B, this is specified only to 2100. After 2100, the emissions of gases other than CO2 are assumed to remain constant at their A1B 2100 values. The projections are labelled according to the level of CO2 stabilisation (in ppm). The broken lines after 2100 indicate increased uncertainty in the simple climate model results beyond 2100. The black dots indicate the time of CO2 stabilisation. The stabilisation year for the WRE1000 profile is 2375.

The objective of Article 2 of the United Nations Framework Convention on Climate Change (United Nations, 1992) is �to achieve stabilisation of greenhouse gas concentrations in the atmosphere at a level that would prevent dangerous anthropogenic interference with the climate system.� This section gives an example of the possible effect on future temperature change of the stabilisation of greenhouse gases at different levels using carbon dioxide stabilisation as a specific example. The carbon dioxide concentration stabilisation profiles developed by Wigley et al. (1996) (see also Wigley, 2000) commonly referred to as the WRE profiles, are used. These profiles indirectly incorporate economic considerations. They are also in good agreement with observed carbon dioxide concentrations up to 1999. Corresponding stabilisation profiles for the other greenhouses gases have not yet been produced. To illustrate the effect on temperature of earlier reductions in carbon dioxide emissions, results are also presented for the original stabilisation profiles referred to as the S profiles (Enting et al., 1994). The S profiles are, however, unrealistic because, for example, they require emissions and concentration values during the 1990s below those actually observed.

In order to define future radiative forcings fully, it is necessary to make assumptions about how the emissions or concentrations of the other gases may change in the future. In addition, it is necessary to have a base scenario against which the effect of the different stabilisation pathways may be assessed. The state of the science at present is such that it is only possible to give illustrative examples of possible outcomes (Wigley et al., 1996; Schimel et al., 1997; Mitchell et al., 2000).

To produce these examples, the SRES scenario A1B is used as the base scenario. CO₂ concentrations for this scenario are close to the WRE CO₂ profiles in terms of their implied past and near-future values, so our choice satisfies the underlying WRE assumption that emissions should initially follow a baseline trajectory. This is not the case for the S profiles, however, because as pointed out above, present day CO₂ concentrations already exceed the values assumed for the S profiles. Note that the baseline scenario (A1B) is specified only out to 2100. For stabilisation cases, emissions of non-CO₂ gases are assumed to follow the A1B scenario out to 2100 and are thereafter held constant at their year 2100 level. For scenario A1B, this assumption of constant emissions from 2100 leads to stabilisation of the other gas concentrations at values close to their 2100 values. For gases with long lifetimes (such as N₂O) it takes centuries to reach stabilisation. In all cases, however, the net radiative forcing changes for the non-CO₂ gases are small after 2100 and negligible after about 2200. Note that, in comparing the baseline case with the various stabilisation cases, the only gas that changes is CO₂.

The models used to calculate the other gas concentrations and to convert concentrations and sulphur dioxide emissions to radiative forcing are the same as those used in Section 9.3.3. The simple climate model used is again that based on Wigley and Raper (1992) and Raper et al. (1996), tuned to the different AOGCMs using the CMIP2 data set (see Appendix 9.1).

The temperature consequences of the five WRE stabilisation profiles used, based on the assumptions described above and using the simple model ensemble (the average results from tuning the simple model to several AOGCMs), are shown in Figure 9.16. The temperature results for the S profiles are also included for comparison. The simple climate model can be expected to give results in good agreement to those that would be produced by the AOGCMs up to 2100. Thereafter the agreement becomes increasingly less certain and this increasing uncertainty is indicated on the graph by the graduated broken lines. Indeed it has been shown in a comparison of results from the simple model and HadCM2 that the simple model under-estimates the temperature change compared to HadCM2 on longer time-scales (Raper et al., 2001a). This is at least in part due to the fact that the HadCM2 effective climate sensitivity increases with time (see Section 9.3.4.1). The results in Figure 9.16 are consistent with the assumption of time-constant climate sensitivities, the average value being 2.8°C.

Figure 9.17: Simple model results: Temperature change (a) from 1990 to 2100 and (b) from 1990 to 2350, resulting from the five WRE stabilisation profiles, using the simple climate model tuned to different AOGCMs as indicated on the bottom axis. (c) and (d) show the corresponding results for the four S profiles. The underlying assumptions are the same as those for Figure 9.15. For comparison, temperature changes for the base scenario A1B are also shown in (a) and (c).

Since sulphur dioxide emissions stabilise at 2100, the forcing from sulphate aerosols is constant thereafter. CH₄ concentrations stabilise before 2200, and the forcing change from N₂O concentration changes after 2200 is less than 0.1 Wm^-2. The continued increase in temperature after the time of CO₂ stabilisation (Figure 9.16) is in part due to the later stabilisation of the other gases but is primarily due to the inertia in the climate system which requires several centuries to come into equilibrium with a particular forcing.

Temperature changes from 1990 to 2100 and from 1990 to 2350, for the simple climate model tuned to seven AOGCMs, are shown in Figure 9.17. These Figures give some indication of the range of uncertainty in the results due to differences in AOGCM response. Figure 9.17a also shows the temperature change for the baseline scenario, A1B. The percentage reductions in temperature change relative to the baseline scenario that the WRE profiles achieve by 2100 are given in Table 9.3. These range from 4 to 6% for the WRE1000 profile to 39 to 41% for the WRE450 profile. Note that these reductions are for stabilisation of CO₂ concentrations alone.

Although only CO₂ stabilisation is explicitly considered here, it is important to note that the other gases also eventually stabilise in these illustrations. The potential for further reductions in warming, both up to 2100 and beyond, through non-CO₂ gases, depends on whether, in more comprehensive scenarios (when such become available), their stabilisation levels are less than the levels assumed here.

Only one AOGCM study has considered the regional effects of stabilising CO₂ concentrations (Mitchell et al., 2000). HadCM2, which has an effective climate sensitivity in the middle of the IPCC range (Table 9.1), was run with the S550 ppm and S750 ppm stabilisation profiles (�S profiles�; Enting et al., 1994; Schimel et al., 1997). Simulations with a simple climate model (Schimel et al., 1997) indicate that the global mean temperature response in these profiles is likely to differ by no more than about 0.2°C from the equivalent WRE profiles (Wigley et al., 1996; see Figure 9.16), though the maximum rate of temperature change is likely to be lower with the S profiles. Global mean changes in the AOGCM experiments are similar to those in Schimel et al. (1994). Note that the AOGCM experiments consider stabilisation of CO₂ concentrations only, and do not take into account changes in other gases, effectively assuming that concentrations of other gases are stabilised immediately. To allow for ongoing increases in other greenhouse gases, one would have in practice to reduce CO₂ to even lower levels to obtain the same level of climate change. For example, in the IS92a scenario, other trace gases contribute 1.3 Wm^-2 to the radiative forcing by 2100. If the emissions of these gases were to continue to increase as in the IS92a scenario, then CO₂ levels would have to be reduced by about 95 ppm to maintain the same level of climate change in these experiments.

Changes in temperature and precipitation averaged over five sub-continental regions at 2100 were compared to those in a baseline scenario based on 1%/yr increase in CO₂ concentrations from 1990. With both stabilisation profiles, there were significant reductions in the regional temperature changes but the significance of the regional precipitation changes depended on location and season. The response of AOGCMs to idealised stabilisation profiles is discussed in Section 9.3.4.4.

Table 9.3: Reduction in 1990 to 2100 temperature change, relative to the A1B scenario, achieved by five WRE profiles across all seven simple model AOGCM model tunings. WRE1000 refers to stabilisation at a CO2 concentration of 1,000 ppm, etc.
Profile	WRE1000	WRE750	WRE650	WRE550	WRE450
Percentage reduction in temperature change relative to A1B	4 � 6%	9 � 10%	14 � 15%	23 � 25%	39 � 41%

IPCC Homepage