Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

3 Assessing emissions: five approaches compared

Comprehensiveness compared
Accuracy by category
Regional and national emissions by source
Appendix A: Estimates of greenhouse gas emissions
Appendix B: Calculating cumulative and current emissions

Susan Subak


In this chapter, I present a variety of ways to assess responsibility for greenhouse gas (GHG) emissions. The parameters that could define responsibility from a polluter pays perspective include: which greenhouse gases are counted; which sources are included; and what time frame is used for estimating them. A New Zealander who lives in a country with twenty methane emitting sheep for every person may prefer to keep the gases limited to carbon dioxide only. A Swiss citizen mostly emits carbon dioxide by burning fossil fuels, and may be unhappy if only this gas is controlled. And someone from a recently industrialized country such as Singapore might feel justified in pushing for the inclusion of historic emissions in global greenhouse negotiations. The definition of GHG emissions, therefore, has great practical impact on each country's relative responsibility for emissions. The feasibility of controlling emissions sources, linking national abatement actions efficiently with global targets, and verifying emissions after targets have been set are other important considerations that policy makers must take into account when assessing responsibility for emissions.

In the following analysis, five approaches for assigning responsibility among countries for greenhouse gas emissions are examined. They comprise two historical and three current emissions assessments which vary by level of coverage of sources (Table 3.1):

1 cumulative CO2, energy only;
2 cumulative CO2, energy and biota (including CO2 from both fuels and land clearance);
3 CO2, energy only (current);
4 partial CH4 and CO2 (including current emissions of CO2 from energy consumption and deforestation, and methane from energy production and landfills);
5 comprehensive (current emissions of CO2, CO, CH4 and N2O from energy, industrial, biotic and agricultural sources).

Table 3.1 Sources included in selected cumulative (1860-1986) and current (1988) emissions

1 Cumulative CO2, energy only X      
2 Cumulative CO2, energy & biotab X X    
3 COT, energy (current) X      
4 Partial CH4 and CO2 (current) X' X X  
5 Comprehensive (current) X X X X

a 'Other' includes cement production, and agricultural sources, including livestock, rice cultivation, fertilizer consumption, and biomass burning apart from deforestation. The gases include CO2, CH4, CO, and N2O.
b Includes estimated net CO2 release from soil carbon and from above-ground biomass in areas converted from forests to agricultural uses only.
c CO2 and CH4 emissions.

All of the approaches have already entered discussions, either in a political or an academic context. Most of the approximately two dozen countries that have pledged thus far to meet specific national targets to stabilize or control greenhouse gas emissions have focussed on the control of CO2 emissions from energy consumption. Setting targets for CH4 (methane) from energy and industrial sources and CO2 from biotic sources, in addition to CO2 from fossil fuel combustion is being seriously explored by several industrialized countries. The Framework Convention on Climate Change signed at Rio de Janeiro in June 1992, which requires developed country Parties to submit plans for stabilizing emissions, can be interpreted to apply to all greenhouse gas sources with the exception of halocarbons controlled by the Montreal Protocol. Allocating future emissions based on historical release of greenhouse gases has been proposed by a number of researchers (Krause et al. 1989; Smith 1991; Gruebler and Fuji 1991).

Any of these source categories could form a broad basis for resource transfers from North to South to fund technology transfers or greenhouse gas abatement projects. But as the baseline against which national targets or the allocation of traceable emissions permits are set, the national inventories must be accurate and verifiable. A consensus is more likely to be reached over setting targets for sources and gases that can be measured with confidence. Although in the past, regional environmental agreements have been signed before baseline national emissions estimates were completed, in the case of greenhouse gas emissions where the differences in countries' emissions rates are so great, nations are unlikely to favour setting specific targets for controlling sources for which accurate baseline inventories at the country level are not yet available and cannot yet be monitored.

In the following analysis, the relative comprehensiveness of the different source categories is briefly summarized, followed by a discussion of the problems in estimating emissions from these different sources and time frames. In addition, the implications of the five emission categories is illustrated for a selection of the major emitting countries. A brief description of the emissions totals used, and the method for calculating national inventories appear in Appendix A and Appendix B.


Comprehensiveness compared

Cumulative CO2, energy only

As it takes many decades for CO2 to be removed from the atmosphere, the increase in concentration of CO2 from pre-industrial levels is largely due to CO2 emitted in past decades. In this respect, historical CO2 emissions are much more relevant to the level of committed atmospheric warming than are current emissions. Emissions from past energy use, however, make up a smaller portion of total CO2 release than today's fossil fuel emissions because CO2 from land clearing may have been roughly comparable with fossil fuel related CO2 until the middle of this century (see Figure 3.1). Emissions from fossil fuel combustion since the start of the industrial revolution are estimated to be 175 to 215 gigatonnes (GT) of carbon (C), representing between about 55 and 70 per cent of total anthropogenic CO2 release (IPCC 1990). Contributions to warming, however, are considerably lower because CO2 is but one of the gases contributing to the heating effect. Considering both the change in fossil fuel emissions over time (Keeling 1973; Marland et al. 1990) and the estimated contribution of CO2 to total warming (IPCC 1990), it is calculated that cumulative CO2 emissions from energy may contribute about 40 per cent of the warming effect of the trace gases now in the atmosphere.

Figure 3.1 Historic CO2 emissions compared

Cumulative CO2, energy and biota

About 60 per cent of the warming effect of anthropogenic greenhouse gas emissions in the atmosphere is thought to be from CO2 or cumulative CO2 emissions from energy and biota (IPCC 1990).

CO2, energy only (current)

Although current CO2 emissions from energy is the least comprehensive of the five categories considered here, CO2 from energy alone is the major warming source from current emissions. Emissions from energy consumption contribute about 65 per cent of the expected warming effect of the trace gases now being emitted, if halocarbons are excluded from the total. This CO2 share of warming reflects the use of the IPCC's GWP for a 100 year period, published in 1992 (IPCC 1992). The estimated warming is only 60 per cent of the total if the IPCC's 1990 GWPs are employed and no doubt will change as the IPCC revises in the future.

Partial CH4 and CO2

This category, as defined above, covers about 80 per cent of the warming effect of current greenhouse gas emissions excluding halocarbons. Halocarbons have been omitted from the current emissions total because they are already being phased out under the Montreal Protocol.

Comprehensive emissions

The comprehensive approach to emission measurement theoretically represents 100 per cent of current greenhouse gas emissions.

The influence of time horizon

The relative comprehensiveness of the energy and modified comprehensive approaches varies considerably depending on how much of the heating effect of the different gases is taken into account. In Figure 3.2, current emissions are compared using three CO2 equivalence indices including the potential heating effect of the gases over a 20, 100, and 500 year time horizon. In the 20 year time horizon, CO2 from energy contributes only about 45 per cent of the total heating effect because the index based on the shorter time horizon does not capture the ultimate heating effect of CO2, which continues many decades beyond the atmospheric residence time of CH4, the next most important greenhouse gas. Accordingly, the proportion of the total heating contribution due to CO2 from energy is much higher about 70 per cent - over the longer time horizon.

Figure 3.2 Contributions to total emissions by source


Accuracy by category

Table 3.2 summarizes the difficulties in estimating emissions from each of the source groupings; it includes the IPCC's ranges of uncertainty in estimating emissions by source and gas globally for each of the five emissions groups.

Cumulative CO2, energy only

CO2 emissions from energy use have been estimated at the country level between 1950 and 1988 (Marland et al. 1989) and between 1860 and 1950 (Subak and Clark 1990). Marland et al. estimate that the uncertainty of their inventory is 6-10 per cent at the country level (Marland et al. 1988). The accuracy of the pre-1950 data set is limited because of changes in geographical borders and sovereignty, lack of information on the type of coal used in the past, and because data on fossil fuels traded in certain regions are incomplete or unavailable. To relate historical emissions to current concentrations, a coefficient or 'discount rate' must be applied to adjust for the CO2 that has been removed from the atmosphere over time.

Table 3.2 Estimated accuracy of GHG emissions accounts

Estimated range
(IPCC 1990 unless noted)
1 Cumulative CO, Energy (1860-1986)
Fossil Fuel Consumption (CO2) Medium 175-215 GT C (10%)
2 Cumulative CO2 (1860-1986)
Fossil Fuel Consumption (CO2) Medium 175-215 GT C (10%)
Land Use Changes (CO2) Lowa 82-152 GT C (30%)
3 CO2 Energy (Current)
Fossil Fuel Consumption (CO2) High 5.4 GT C (Marland et al. 1990) (5%)
4 Partial CH4 CO2 (Current)
Above Plus:      
Landfills (CH4) Medium 20-70 MT CH4 (50%)
Land Use (CO2) Medium 1.1-3.6 GT C (Houghton 1991) (50%)
Energy Prod. and Distribution (CH4) Medium 44-100 MT CH4 (40%)
5 Comprehensiveb (Current)
Above Plus:      
Fossil Fuel Combustion (N2O, CO) Medium 0.5-1.4 MT N2O (IPCC 1992) (50%)
Cement Production (CO2) High    
Biomass Burning (CH4, CO, N2O) Low 20-180 MT CH4 0.3-1.6 MT N2O
(IPCC 1991)
Enteric Fermentation (CH4) Medium 65-100 MT CH4 (20%)
Animal and Human Wastes (CH4) Low    
Rice Cultivation (CH4) Low 25- 170 MT CH4 (80%)
Fertilizer Consumption (N2O) Low 0.01-2.20 MT N2O (100%)
Halocarbons (CFCs, Halons, HCFCs) High    
Nylon Production (N2O) Medium 0.6-0.9 MT N2O (IPCC 1991) (20%)

a In this cave, the uncertainty at the country level is far greater than the estimated global range.
b This inventory coos not include stratospheric water vapour, which is thought to contribute about 4 per cent of the climate forcing (IPCC 1990), and O3 precursors-NO and volatile organic compounds.

Despite these accounting difficulties, estimates of CO2 from energy use since 1860 are more accurate than those of current emissions from land use changes and agriculture. Keeling (1973) has estimated the uncertainty of historical global estimates for CO2 emissions from fossil fuels at 13 per cent. This range of uncertainty compares with 100 per cent or more for CO2 emissions from biota in the current period and equivalent or greater uncertainties in estimating CH4 from rice cultivation and biomass burning, and N2O, NOx, and CO from all sources (IPCC 1990; Logan et al. 1981).

Cumulative CO2. energy and biota

Estimating national historical emissions from energy and biota involves all of the technical difficulties of estimating cumulative emissions from fossil fuels outlined above, plus the challenge of estimating biotic contributions. While data are available to calculate emissions from fossil fuels on an annual basis, comprehensive international forest surveys are conducted less frequently, generally every decade since 1949. Much of the pre-1950 data are for changes in area devoted to agricultural uses only and therefore omit forest conversion to other uses such as settlements, etc. For all periods, many of the forest surveys are considered unreliable. It is unlikely that additional scientific research will significantly improve the accuracy of these estimates on the national level as investigations of historical trace gas concentrations such as ice core and tree ring analyses shed light on global historic concentrations rather than on nation-specific emissions.

Energy, CO2 (current)

This is the most practical, that is, measurable and verifiable approach of the five. Carbon dioxide emissions from current energy consumption are estimated to be accurate at the country level within an error range of only about 6-10 per cent (Marland et al. 1989). A comparison of the (ORNL) Marland and Rotty inventory (Marland et al. 1988), which is based on United Nations energy statistics with a new inventory of CO2 release (don Hippel et al. 1992) from energy consumption that was derived from OECD/IEA statistics (OECD/IEA 1990a) suggests that the error range may be higher for some countries. Regardless, the level of uncertainty in estimating emissions from this source is far lower than the uncertainty associated with inventories of the other gases and sources and should improve in the near future as a number of agencies are refining emission factor estimates and end-use data at the country level.

Partial CH4 and CO2

This approach is midway in practicality between the CO2/energy only approach and the comprehensive approach. The additional sources - land use changes, landfills, and fossil fuel extraction - cannot be estimated as accurately as energy consumption. The error range for estimating CO2 from land use changes and CH4 from landfills and fossil fuel extraction is + 40 to + 50 per cent at the global level (IPCC 1990), with developing countries generally at the higher end. Nevertheless, these sources of CO2 and CH4 should be easier to monitor than the agricultural sources and remaining gases. The landfill and coal mine sources of CH4 are also potentially important sources of natural gas (US/Japan Working Group on Methane 1992). Employing technology to recover and utilize natural gas from these sources should eventually enhance our capacity to control and monitor CH4 release.

For a number of countries in the tropics where CO2 emissions from deforestation far outweigh emissions from energy consumption, per capita estimates change a great deal depending on the assumptions used to estimate land clearing and biomass levels. As the FAO's once-a-decade study of tropical deforestation and tree plantation establishment and the Brazilian Space Institutes (INPE) detailed remote sensing survey of the Amazon Basin are due to be published in the next few years, estimates of emissions from land use changes should improve significantly. In addition, new international statistics on forest growth in temperate countries recently completed by the FAO/ECE, as well as new country studies for Northern and Central Europe, provide further information on the magnitude of CO2 uptake in northern forests.

Comprehensive emissions

The additional sources and gases not included in the above list are far more difficult to inventory. Generally, emissions from the minor greenhouse gas N2O, and CO - which oxidizes to become CO2 and affects the atmospheric residence time of CH4 - are highly uncertain. All of the agricultural sources are included in this approach. Of these, the factors that determine the release of CH4 from livestock enteric fermentation may be the best understood. But even in this case, the accuracy of national estimates for many countries is doubtful at present, because the controlling factors, which include livestock diet, breeding, and management practices, vary from country to country and accurate data are not available for many countries, particularly in the developing world. Measured CH4 release from rice cultivation varies widely according to soil type, fertilizer application, climate, and irrigation regime, but the net effect of all these conditions on emissions is not yet understood. Calculation of CH4 release from animal and human wastes has only started to be undertaken in the last two years, and estimates are rough, reflecting extrapolations based on only a few site-specific studies. Emissions of CH4, CO, and N2O from biomass burning vary with the extent of crop or forest burning, and the moisture and carbon and nitrogen content of the biota. Emissions of N2O and CO from the remaining sources are all highly uncertain.

Unlike the sources covered only in the partial CH4 and CO2 approach (energy, deforestation, and landfills), the additional sources covered here (livestock, rice cultivation, cement production, and fertilizer consumption) pose greater problems as abatement targets because their control would likely entail directly curtailing economic activities rather than reducing the residuals stemming from these activities. The agricultural and industrial activities they represent may be considered essential subsistence activities by many countries (Parikh et al. 1991), although in the case of livestock management for some animals, reducing CH4 emissions through changes in diet and breeding may be compatible with development goals (Leng 1991).

Unlike the three CO2 approaches, the partial and fully comprehensive approaches require an index to compare the heating effect of CH4 and CO2 emissions. The problems involved in evaluating the relative warming contribution of the gases include the difference in estimating the atmospheric lifetime of gases (particularly CO2), calculating indirect effects of the emitted gases, and specifying the most appropriate time period for which to calculate the warming effect (IPCC 1990). In practice, however, the choice of CO2 equivalent applied to these sources may have little effect on most countries' relative ranking by warming contribution.


Regional and national emissions by source

The difference in relative emissions contribution from industrialized and developing countries is summarized in Table 3.3.

It is clear from Table 3.3 that the more current time frame and the addition of the non-energy sources increases the emphasis of emissions from developing countries. This overall pattern holds true for emissions from selected countries. Relative per capita emissions from ten countries that account for 60 per cent of current CO2 emissions from fossil fuels appear in

Table 3.3 Emissions from industrialized and developing countries (% of world total CO2 equivalent)

Emissions category Industrialized Developing
1 Cumulative CO2, energy 86 14
2 Cumulative CO2, energy and biota 68-80a 32-20
3 CO2, energy (current) 72 28
4 Partial CH4 and CO2 (current) 57 43
5 Comprehensive (current) 52-57b 48-43

a This range is based on alternate assumptions of historical land clearing rates.
b This range is calculated based only on differences between the short and longer GWPs. If CFCs are excluded from the totals, industrialized countries' emissions comprise 52 per cent of the total assuming the 100 year GWP.

Figure 3.3. The group includes the eight greatest emitters, in addition to Mexico and Nigeria.

The bars in Figure 3.3 show countries' emissions levels relative to the global mean. For example, the white bar for Germany indicates that per capita CO2 emissions from fossil fuels are three times the global per capita mean. The per capita emissions patterns illustrated fall into three general patterns:

Figure 3.3 Per capita emissions from selected countries

1 An upward slope from the bars sequenced first as cumulative emissions followed by current emissions and fossil-related CO2 followed by emissions from all greenhouse gases and sources. While the shape of curve is tentative for some countries, it is clear that emissions in the current period and emissions from biotic and agricultural sources emphasizes the contributions of these developing countries (Brazil, China, India, and Mexico).

2 The corresponding downward sloping pattern for industrialized countries is more dramatic. The scale of per capita emissions in the historic period and from fossil fuel emissions is significantly greater for these countries (Germany, UK, and USA).

3 A horizontal pattern emerges for the more recently developed countries Japan and USSR) where per capita emissions by time period is relatively constant and the biotic component minor.

Population-weighted emissions were selected as the most compelling form of comparison. A per capita emissions criterion is intuitively equitable in a spatial sense because it assesses individual responsibility regardless of political borders, although in practice the evenhandedness may be diminished because of individuals' disparate emissions release. Alternative allocators have major flaws. Per land area introduces undeserved entitlements to countries with large uncultivatable or uninhabitable regions. Per GDP is regressive in that late-developing countries tend to have high emissions levels per unit of output, as many developing economies are especially energy intensive. Several regions have advanced a population-based approach. Japan recently pledged to cap future CO2 emissions at current per capita levels. In preliminary discussions on approaches to meeting its overall CO2 stabilization goal, the European Community indicated a preference for national targets based on per capita CO2 emission levels. Analysts and scholars have also favoured the per capita approach, although with added variations and qualifications, e.g. considering cumulative population (Smith et al. 1990), weighting by adult population (Grubb et al. 1992), crediting carbon sinks on a per capita basis (Agarwal and Narain 1991) and designating an intergenerational per capita emissions allotment (Gruebler and Fujii 991).

In Figure 3.4, I compare three sets of data - greenhouse gas emissions, population, and GDP - for the selected countries. Two general patterns emerge. For the developed countries, the share of the world's total for each of the indicators forms an inverse U-shape, with the share of GDP (and UN contributions) greater than the countries' contribution to world population and emissions. For instance, Japan makes up 2 per cent of the world's population, produces 9 per cent of the global GDP, and releases only 5 per cent of annual fossil-fuel related CO2. In contrast, the pattern for developing countries, although not as pronounced, tends to be a U-shape, with relative GDP less than the other indicators. One implication of this pattern found among the ten countries examined here is that greenhouse gas intensity (GHG/GDP) is greater in developing countries relative to the corresponding ratios in developed countries than is the GHG/population ratio. Developing regions often release more CO2 (energy related) per unit of output than do developed countries (Grubb et al. 1992). In Figure 3.4, a comparison is also made with relative GDP, or purchasing power parity adjusted GDP, a newer index that attempts to adjust for differences in purchasing power in different parts of the world and for fluctuating exchange rates (Summers and Heston 988).

Figure 3.4 Socioeconomic assessments compared with GHG emissions

While analysts are already discussing ways of adjusting assessments in order to find an equitable solution and the right incentives, in practice the differences between alternate indicators such as population and adult population are often not as great as the differences between the levels of responsibility implied in the alternative emissions source categories discussed above. By way of comparison, Figure 3.4 shows that the difference between developing countries' share of the world's population versus adult population is not as great as the difference between countries' share of emissions from all greenhouse gases versus CO2 from energy only.



The choice of emissions source category can bear upon countries' implied responsibility for emissions if an agreement is based on some form of the polluter-pays principle. The emissions categories involving the greatest uncertainties in measurement - the more comprehensive approaches - also place relatively greater emphasis on emissions from developing regions, where a high proportion of emissions from agricultural and biotic sources originate. Per capita emissions of CH4 and N2O are frequently greater in developing countries.

Cumulative CO2 emissions from fossil fuels can now be estimated with more precision than can most of the sources in the comprehensive source category. However, the uncertainty associated with estimating current emissions from biotic and agricultural sources should diminish with time whereas flaws in historical data, in particular those relevant to land use, may be immutable. CO2 from fuel combustion is the most verifiable and comprises about 65 per cent of current contribution to global warming.

The partial CH4 and CO2 category, which includes CO2 and CH4 from energy, industrial, and biotic sectors (and excludes the more difficult-to-measure agricultural activities and the minor trace gases) makes up about 80 per cent of the total warming effect from current emissions (excluding halocarbons). It would be significantly more difficult to monitor emissions from these sources than to monitor CO2 emissions from fossil fuel combustion, but for a variety of reasons, economic as well as environmental, it may be time to develop emissions assessments applied to this more inclusive, but not comprehensive, approach.

Contents - Previous - Next