Optimizing C02 Emissions

We turn now to the problem of coupling the impulse-response model summarized above to a simple socioeconomic model. The coupled model will then be applied to determine the optimal C02 regulation policy that would minimize the net impact of climate change. This can be represented generally as the sum of two contributions: the direct and indirect costs of climate change itself (termed climate "damage costs" in the following), and the costs incurred in reducing C02 emissions ("abatement costs"). We restrict the discussion here to the impact of C02 alone, as the dominant anthropogenic greenhouse gas, but note that the same approach can be applied to the climate change induced by other greenhouse gases also.

A number of cost studies and optimized cost-benefit analyses of the economic impact of CO, emissions have been published in the literature (e.g., Manne and Richels, 1991; Peck and Teisberg, 1992; Nordhaus, 1993; Richels and Edmonds, 1995; Tahvonen and Storch, 1994; Nordhaus and Yang, 1996; Wigley et al„ 1996; see also the review by Fankhauser, 1995). Normally, the climate models are reduced to box-type models, while the economic models are represented as aggregated dynamic-growth models,

Emissions [GtC yr1] 10Q0

Emissions [GtC yr1] 10Q0

Concentration [ppm]

Temperature change [ C]

Concentration [ppm]

Temperature change [ C]

1900 1950 2000 2050 2100 1900 1950 2000 2050 2100 1900 1950 2000 2050 2100

1900 1950 2000 2050 2100 1900 1950 2000 2050 2100 1900 1950 2000 2050 2100

Emissions [GtC yr1]

Emissions [GtC yr1]

1800 2000 2200 2400 2600 2800 3000 [yr]

6000

4000

2000

Concentration [ppm]

Temperature change [°C]

Concentration [ppm]

Temperature change [°C]

4000

2000

1800 2000 2200 2400 2600 2800 3000

1800 2000 2200 2400 2600 2800 3000 [yr]

FIGURE 9 Evolution of the atmospheric CO concentration and the global mean temperature, computed with the impulse-response climate model of Hooss et ul. (2000), for a BAU scenario and a frozen-emissions scenario F over the next 100 years (upper panels) and the next 1000 years (lower panels). The long-term climate change in the lower right panel is seen to greatly exceed the predicted climate change in the next 100 years (indicated also by boxes in the lower panels).

1800 2000 2200 2400 2600 2800 3000 [yr]

1800 2000 2200 2400 2600 2800 3000

1800 2000 2200 2400 2600 2800 3000 [yr]

FIGURE 9 Evolution of the atmospheric CO concentration and the global mean temperature, computed with the impulse-response climate model of Hooss et ul. (2000), for a BAU scenario and a frozen-emissions scenario F over the next 100 years (upper panels) and the next 1000 years (lower panels). The long-term climate change in the lower right panel is seen to greatly exceed the predicted climate change in the next 100 years (indicated also by boxes in the lower panels).

dependent on the distribution of the total output production between consumer goods, investments in capital and technological development, abatement measures, etc. as control variables. To account for the influence of the long time-scales imposed by the climate system, and the issues of intergenerational accounting and equity that these raise, we apply in the following the more realistic nonlinear impulse-response climate model of Hooss et al. (2001). While radically reducing the economic model to simple price expressions for the climate damage and abatement costs, in accordance with Hasselmann et al. (1997). The principal conclusions drawn from our discussion will be independent of the details of the economic model.

The global climate-damage costs are taken proportional to the sum of the squares of the change in global mean near-surface temperature and the rale of change in the global mean temperature. This corresponds to the assumption that any change in the present climate, to which humans and ecosystems have had time to adapt, is detrimental, and that the damages increase nonlinearly both with the change in global mean temperature and with the rate of the temperature change. The global mean temperature is regarded here as a proxy for all climate change variables, such as precipitation, cloudiness, the frequence and strengths of El Nino, the intensities of storms, droughts and other extreme events, and the rise in sea level. This is dynamically consistent with atmospheric climate variables, since the atmospheric response to greenhouse forcing can generally be well described in numerical climate simulations, as mentioned above, by a few dominant EOF patterns, whose coefficients are diagnostically coupled to the global mean temperature. However, the projection onto global mean temperature is more questionable for climate properties related to the ocean, such as El Niño and the rise in sea level, since the time scale of the ocean response to external forcing differs from that of the atmosphere (cf. Fig. 4).

The expression for the abatement costs is based on the assumption that any deviation r= (e — e0)/e0 of the CO, emissions e(t) from the emissions e0(t) of the BAU economic development path, in which all climate change impacts are ignored, incurs costs. For small deviations, the costs are assumed to be quadratic in r. Quadratic-cost penalties are also introduced for the first and second time derivatives of r to parametrize the effects of economic inertia (capital losses, development costs, etc.)

The optimal C02 emission path is the one that minimizes the total time-integrated sum of the climate damage and abatement costs. In intertemporal economic accounting, all costs are traditionally discounted at the same rate. Theoretically, this is equal to the inflation-adjusted market interest rate. However, the appropri

Sb: Baseline optimized scenario Sz: Zero economic inertia

Sd: Damage & abatement costs both discounted (Ta=rd=50y) FIGURE 10 Optimized CO, emission paths with resulting changes in C02 concentration and global mean near-surface temperature. S(„ baseline scenario: discounting of abatement costs only, finite economic inertia; S., same as baseline scenario without economic inertia; S,(, same as baseline scenario but with equal discounting of climate damages and abatement costs.

ate discount rate for nonmarket values, which comprise a large fraction of the climate damage costs, is the subject of considerable debate (cf. Hasselmann et al, 1997; Nordhaus, 1997; Heal, 1997; Brown, 1997; Hasselmann, 1999). It has been argued on ethical grounds, and also on the basis of economic, time-dependent relative-value reasoning, that the appropriate discount rate for such values should be smaller than for market goods or even zero. We have accordingly applied separate discount rates for the climate damage and abatement costs. In our baseline optimization run S(, we have assumed a zero discount rate for the climate damage costs and a discount rate of 2% for the abatement costs.

Figure 10 shows the optimal CO, emission paths and the associated atmospheric CO, concentrations and global mean-temperature evolution for the baseline case Sb and two further cases S, and Slt. The scenario Sz is identical to the baseline case except that the economic inertia is set equal to zero. Although the emission paths for the solutions Sh and S, differ significantly in the first few decades, the differences in the long-term climate impact are minor. This demonstrates that for an effective climate mitigation policy, long-term emission abatements far outweigh the impact of short-term reduction measures. Essential to averting major climate change in the long term is the gradual but complete replacement of fossil fuels by carbon-free energy technologies.

This is further illustrated by Fig. 11, which compares the cases Si, and Sz with the emission reductions agreed to by the industrialised countries in the Kyoto protocol. The Kyoto curve lies between the cases Sh and Sz and thus appears quite acceptable from the viewpoint of these computations. From the long-term perspective imposed by the memory of the climate system, however, the details of the Kyoto compromises over emission-reduction percentages appear rather irrelevant compared with the central challenge of establishing an effective long-term post-Kyoto mitigation strategy that will gradually but surely lead to a restructur ing of the present energy technology from fossil fuels to carbonfree energy generation.

The third case, Sd, in Fig. 10 illustrates the strong influence of the discount rates on the computed optimal solutions. In contrast to the baseline scenario Sh and the zero-inertia scenario Sz, in which only the abatement costs were discounted, in scenario S,; both costs were discounted at the same rate of 2%, following standard economic practice. In this case, the optimal emission path leads to a climate "catastrophe" similar to the BAU case shown in Fig. 9. The explanation is simple: since major climate change develops only after several centuries, the associated discount factor is very small, and the discounted climate damage costs are negligible. Thus, there is only a small cost-penalty incurred in following the BAU path. This also explains why previous cost-benefit analyses (e.g. Nordhaus, 1993), based on the applicaton of uniform dis-

-With economic inertia (baseline)

--Zero economic inertia

Kyoto protocol

1950 today 2050 Ly'J

FIGURE 11 Comparison of the Kyoto protocol with the optimized solutions S{, and Sz of Figure 10, with and without economic inertia, respectively.

-With economic inertia (baseline)

--Zero economic inertia

Kyoto protocol

count factors for all costs, obtained optimal mitigation strategies requiring only minor C02 emission reductions.

0 0

Post a comment