CO₂ is the largest contributor to anthropogenic radiative forcing of the atmosphere. As described in more detail in Chapter 3, the main sources of anthropogenic CO₂ emissions are fossil fuel combustion and the net release of carbon from changes in terrestrial ecosystems, commonly referred to as land-use changes. To a lesser extent, CO₂ is emitted by industrial activities, in particular by cement production (Table 5-3).

5.3.1. Carbon Dioxide Emissions from Fossil Fuels and Industry

As shown in Table 5-3, fossil fuels were the main source of CO₂ emissions in 1990. Therefore, it is expected that future CO₂ emission levels will depend primarily on the total energy consumption and the structure of energy supply. The total energy consumption is driven by population size, level of affluence, technological development, environmental concerns, and other factors. The composition of energy supply is determined by estimated reserves of fossil fuel and the availability, relative efficiency, and cost of supply technologies.

Emissions from gas flaring and industrial emissions are much lower in comparison with energy-related emissions; for simplification, in this discussion they are added to the latter. In 1990, the global emissions from cement made up about 2.5% of the total global CO₂ emissions (Houghton et al., 1995).

Figure 5-2 shows standardized carbon emissions from fossil energy and industry for the 40 SRES scenarios. Sample statistics (in terms of percentiles, means, and medians) are indicated against the background of 40 individual scenarios that make up the SRES scenario set. The figure also presents emissions ranges spanned by each of the four scenario families in 2100.

SRES scenarios cover a wide range of annual emissions, and the uncertainties in future emission levels increase with time. Up to about the 2040s and the 2050s, emissions tend to rise in all scenarios, albeit at different rates. Across scenarios this reflects changes in the underlying driving forces, such as population, economic output, energy demand, and the share of fossil fuels in energy supply. By 2050, the emissions range covered by the 40 SRES scenarios is from about 9 to 27 GtC, with the mean and median values equal to about 15 GtC. The range between the 25^th and 75^th percentiles of emissions (the "central tendencies") extends from 12 to 18 GtC (i.e. from twice to thrice that in 1990). Within this interval lie three of the four marker scenarios. However, a fair number of scenarios (eight out of 40) also indicate the possibility of much higher emissions (in the 18 to 27 GtC range) that reflect an increase by a factor of up to 4.5 over 60 years (1990 to 2050). Another eight SRES scenarios have 2050 emissions below the 25^th percentile (Figure 5-2).

Beyond 2050, the uncertainties in energy and industrial CO₂ emissions continue to increase. By 2100, the range of emissions across the 40 SRES scenarios is between 3 and 37 GtC, which reflects either a decrease to half the 1990 levels or an increase by a factor of six. Emissions between the 25^th and 75^th percentiles range from 9 to 24 GtC, while the range of the four marker scenarios is even wider, 5 to 29 GtC. The 2100 median and mean of all 40 scenarios are 15.5 and 17 GtC, respectively.

As time passes in the scenarios, uncertainties not only increase with respect to absolute levels of CO₂ emissions, but also with respect to their trajectories. Scenarios portray different emission patterns that range from continuous increases up to 2100, through emissions that gradually level off by 2100, to trend reversals in which emissions begin to decline in the second half of the 21^st century. These dynamic emission patterns diminish the significance of emission levels in any particular year, such as 2100.

Over a long time horizon, it also becomes increasingly difficult to perceive the future in terms of "central tendencies." For instance, between 2050 to 2100, up to eight different scenarios from all four SRES scenario families have CO₂ emission levels within 10% of the median of all 40 SRES scenarios. Thus, there is no single scenario family or individual scenario that has "median" emissions with respect to the entire uncertainty space described by the 40 SRES scenarios. Similar emission levels can arise from very different combinations of driving-force variables that are embedded in the SRES scenario families and groups.

The wide ranges of energy and industry-related CO₂ emissions in the SRES scenarios reflect the fact that the "best" or the "most likely" quantifications are nearly impossible to identify. The following discussion suggests that even scenarios with very similar input parameters (e.g., population and GDP) may produce a large variation in resultant CO₂ emissions.

IPCC Homepage