1592

1.02

0.28

(16%)

(7%)

(4%)

12% (21%) 29%

Table 1. Relative contributions to total global radiative forcing change over 1990-2100 of different gases under the IS92a, c and e emissions scenarios. The forcing values here are those used in SAR WGI (Section 6.3). The low, mid and high suiphate aerosol forcing values are based on 1990 forcings of: direct aerosol forcing: -0.2, -0.3, -0.4 W m2; indirect aerosol forcing: -0.4, -0.8, -1.2 W m2 (the full range of aerosol forcing uncertainty is larger than this; see SAR WGI, pp. 113-115). Only the mid-aerosol forcing values were used in SAR WGI (Section 6.3). Forcing values are given in W m2; non-CO2 gas forcing values are also given as percentages of the CO2 value. CH1 forcing includes the related effects of tropospheric ozone and stratospheric water vapour changes. Halocarbon forcing includes the effects of stratospheric ozone changes.

2.2.3 Reference Stabilization Scenarios

Given the very large uncertainties in the roles of the non-CO2 gases relative to CO2 under an "existing policies" assumption, and given that no comprehensive studies have been carried out to examine their roles under the assumption of concentration stabilization, we can only consider them in a sensitivity study context. We, therefore, begin with a set of reference cases in which the emissions of CO2 follow a range of stabilization pathways, the emissions of CH4, N2O and SO2 are assumed to remain constant at their 1990 levels, and halocarbons follow the Montreal Protocol scenario used in the SAR WGI (Section 6.3) global mean temperature and sea level calculations.

For halocarbons in the reference scenarios we assume that the Montreal Protocol applies strictly (see SAR WGI: Chapters 2 and 6) so that there is only a single future scenario for these gases. Because the total forcing for these gases over 1990-2100 (accounting for the effects of stratospheric ozone changes) is relatively small (0.3 W m-2), uncertainties due to incomplete compliance with the Protocol and/or future emissions of substitute (hydrofluorocarbon, HFC) or non-controiled gases may be even smaller. In the context of global climate change, therefore, and given that they are not addressed by SAR WGI, we have chosen not to include these uncertainties. However, should a comprehensive (multi-gas) framework for stabilization be adopted, a more detailed gas-by-gas assessment of halocarbon forcing may be required at a specific country level.

Because the calculations performed here run beyond 2100, some assumption must be made regarding halocarbon emissions after this date. If these emissions remain constant at their 2100 level, the forcing level would remain close to 0.3 W m-2. This would stabilize halocarbon (primarily HFCs)

concentrations at relatively high levels. For the reference cases we assume that halocarbon emissions remain constant at their 2100 levels. Hence, eventually, concentrations will remain constant in accordance with Article 2. We note, however, that the constant-2100 emissions assumption leads to a potential global mean forcing overestimate after 2100 of, eventually, up to 0.4 W m-2.

For tropospheric ozone, in the absence of any projections, and again following SAR WGI (Section 6.3), we assume that the only forcing changes are those that arise from the ozone that is produced by methane induced changes in tropospheric chemistry. This term amounts to around 0.15 W m-2 by 2100 under IS92a, but is much less for the reference case of constant CH emissions. Our assumption here may be unrealistic if nitrogen, hydrocarbon, or other ozone precursors associated with ozone concentrations increase due to a rise in anthropogenic pollution.

It should be noted that we are not suggesting that the reference cases in any way reflect predictions of the future, especially with regard to future SO2 emissions, nor that they should be a target for policy. The point of the reference cases is to help assess the relative importance of CH4, N2O and SO2 emissions in determining future global mean temperature and sea level change.

To quantify the sensitivity of equivalent CO2 to other gases we consider perturbations from the reference cases in which annual CH4 emissions increase or decrease linearly over 1990-2100 by a total of ±100 Tg(CH4) (i.e., ±75 TgC) relative to 1990 and remain constant thereafter, annual N2O emissions increase or decrease linearly over 1990-2100 by a total of 12 Tg(N) relative to 1990 and remain constant thereafter, and annual SO2 emissions increase or decrease linearly over 1990-2100 by

2.2.4

Stabilizing the atmospheric concentrations of greenhouse gases. an explicit goal of Article 2, would not necessarily result in stabilizing the human caused perturbation in radiative forcing. This is because aerosols, which are not explicitly addressed by Article 2, also have radiative effects. If concentrations of both greenhouse gases and aerosols are stabilized, this would stabilize the human perturbation in global mean radiative forcing6. Note also that because aerosols are not uniformly mixed gases, the geographical distribution of emissions of aerosols and their precursors can have important effects on regional climate. Stabilizing the human perturbation in global mean radiative forcing is clearly different from stabilizing CO2 concentration alone. Thus, while mitigation efforts may target members of a suite of greenhouse gases, impact studies must consider climates influenced by multiple gases and aerosols. "Equivalent CO2" is a technique for considering multiple radiative forcing components in the aggregate.

In the calculations of future global mean temperature and sea level change given in SAR WGI (Chapters 6 and 7), the models were driven by the total radiative forcing, which

6 This will not eliminate climate variability because the climate system exhibits considerable natural variability, beyond anthropogenic influences.

was obtained by summing the forcings due to all anthropogenic trace gases (see Table 1). In global mean terms, this total forcing can be treated as if it came solely from changes in CO2; i.e.. from an "equivalent CO2 concentration". The equivalent CO2 concentration, Ceq. can be defined, therefore, from the relationship between actual CO2 concentration and radiative forcing. In SAR WGI, the relationship used was that from the First IPCC Assessment Report (IPCC, 1990). The uncertainty in this relationship may be up to approximately ±20 per cent (see IPCC TP SCM, 1997).

Although the equivalent CO2 concept is pedagogically useful and provides a means to compare the effects of CO2 with other gases, it does have disadvantages. An important disadvantage arises from the non-linear relationship between radiative forcing and CO2 concentration. This nonlinear relationship means that, at higher CO2 levels, it requires a larger CO2 change to increase radiative forcing by the same amount. Because of this, radiative forcing changes can be added, but CO, equivalents can not be. We have therefore retained the use of radiative forcing as our primary variable.

A further disadvantage of the equivalent CO2 concept is that, in the context of impact assessments, it addresses only the climate change aspect. Other impacts of increasing CO2 (e.g., fertilization), sulphate aerosol (acidification), and ozone may also be important. Also with the equivalent CO2 concept, as with radiative forcing, a global aggregate measure subsumes information about regional aspects of climate change that are critical in assessing impacts. It would be possible, for example, to impose a forcing pattern on the climate system that had zero global mean forcing, but which would lead to large changes in regional climate.

We now give equivalent CO2 results for different concentration stabilization levels. We consider S350, S450, S550, S650 S750, and WRE1000, together with the constant 1990-level emissions reference cases for CH4. N2O and SO2, and halocarbon emissions following the Montreal Protocol (see Section 2.2.2). To illustrate the dependence of equivalent CO2 level on the pathway to CO2 stabilization, we also consider WRE550. These reference case results are given in Figure 7, where the forcing values are given relative to 1990 (some 1.3 W m2 above the pre-industrial level). In the year 2500, close to the point of equivalent CO2 stabilization, the equivalent CO2 concentrations vary from 26 ppmv (S350) to 74 ppmv (WRE1000) above the actual CO2 level. In all cases, the forcing difference due to gases other than CO2 is the same: 0.66 W m2 over 1990 to 2500. As noted above, this is equivalent to differing amounts of CO2 at different concentration levels because of the non-linearity of the equivalent CO2/radiative forcing function.

Note that here the mid-1990 equivalent CO2 level is 342 ppmv, slightly below the actual CO2 level (354 ppmv). This is

8.50

8.97

(2375)

1072

1 154

Table 3. Equivalent CO2 (ppmv) and radiative forcing (from 1765) (AF) at the point of CO2 stabilization, for various assumptions about nonCO2 greenhouse gases and aerosols. The reference case assumes constant emissions for SO2, N2O and CH, after 1990. The "CO2 only" column assumes changes after 1990 are in CO2 only (as in SAR WGI). Note that the equivalent CO2 level at CO2 stabilization in these cases differs from the CO2 stabilization level because of differences between the 1990 CO2 and equivalent CO2 levels.

2.2.5

Equivalent CO2 Sensitivities

Section 2.2.4 provides estimates of equivalent CO2 that include the collective effects of CH4, N2O, SO2 and the halocarbons. Here we consider the influences of CH. N2O, and SO2 separately. To do this, we use emissions perturbations about the constant 1990 emissions reference cases.

For CH4 (Figure 9a), a perturbation in annual emissions from 1990 to 2100 of ±75 TgC (±100 Tg(CH4)) changes radiative forcing by approximately ±0.20 W m2 at concentration stabilization. This translates to equivalent CO2 differentials of approximately ±15 ppmv for S450 and ±22 ppmv for S650. For annual N2O emissions, a perturba tion of ±2 Tg(N) from 1990 to 2100 changes forcing by ±0.16 W m2 at concentration stabilization, and gives concentration differentials of ±12 ppmv for S450 and ±18 ppmv for S650 (see Figure 9b).

Sulphur dioxide sensitivities occur in two ways. First, there is the basic sensitivity to emissions uncertainties (Figure 10a). At concentration stabilization, perturbations of ±50 per cent

relative to 1990 in annual SO2 emissions (i.e., ±37.5 TgS) lead to forcing differentials of -0.37/+0.45 W m-2, which translates to equivalent CO2 concentration differentials of -27/+36 ppmv for S450 and -40/+52 ppmv for S650 (note that the sign of the forcing or concentration differential is opposite to the sign of the emissions perturbation).

In addition to the influence of emissions uncertainties, the effect of SO2 on equivalent CO2 concentrations is sensitive to the highly uncertain relationships between SO2 emissions and radiative forcing. SO2-derived sulphate aerosol affects radiative forcing both directly, under clear-sky conditions, and indirectly, through changes in cloud albedo. The central estimate of direct sulphate aerosol forcing for 1990 was calculated in SAR WGI as -0.4 W m2, an estimate of -0.8 W m2 was used in Section 6.3 of SAR WGI for the indirect forcing. When combined with a carbonaceous (soot) aerosol forcing of +0.1 W m22, this gives a total sulphate aerosol forcing of -1.1 W m2. To assess the sensitivity to uncertainties in this quantity, we use the range of ±0.1 W m2 for direct forcing and ±0.4 W m2 for indirect forcing (giving a total sulphate (plus soot) aerosol forcing range of -1.1 ± 0.5 W m2).