nology or a geopolitical shift, could occur and radically alter future emissions.
Most emissions scenarios are developed using the IPAT model (Ehrlich and Holdren, 1971) in which environmental impact is the result of a multiplication of three driving forces: population, affluence per person, and technological impact per unit of affluence. When applied to greenhouse gas emissions, the impact is the rate of greenhouse gas emissions, while the technological factor is the rate of greenhouse gas emissions per unit of gross domestic product (GDP). The IPAT model has a long history in environmental studies, and there has been much debate over whether it is the proper approach (Chertow, 2001). As a purely mathematical multiplicative identity, it must yield correct emissions rates if all of the PAT factors are well known.
For example, world population growth today is approximately +1.3 percent per year (UN, 2002). Per capita GDP varies widely over time and by country, but recent estimates of world GDP per capita growth are approximately 1.1 to 1.4 percent per year (Maddison, 1995; World Bank, 2004). CO2 emissions per unit of world GDP output display a well-documented long term decline, persistent over the past century, of about −1.3 percent per year (Nakićenović, 1996). The IPAT model indicates that adding these three rates should yield CO2 emissions growth rates. The sum is about 1 to 1.5 percent per year growth in world CO2 emissions, and this is what is seen, albeit with significant interannual variations, when using estimates of CO2 emissions inventories (CDIAC, 2004).
Although the IPAT approach is reliable for estimating present-day emissions—as it must be if given good estimates for the main driving forces—using it to develop estimates of future emissions is by no means so easy. In principle, given a range of long-term projections for population, economic growth and technological changes, one could combine such projections to yield a range of future emissions rates. Some aspects of this approach are discernible in the predecessor scenarios to the Intergovernmental Panel on Climate Change (IPCC) Special Report on Emissions Scenarios (SRES) report: the IPCC IS92 emissions scenarios (Leggett et al., 1992).
Without doubt, the most sophisticated long-term projections exist for world and regional population changes, because these have been an area of intensive demographic research for many decades (Bongaarts and Bulatao, 2000; Lutz, 1996). Projections are routinely made for the world, regions, and countries by the United Nations (UN, 2002) and the International Institute for Applied Systems Analysis (IIASA; Lutz et al., 2004; Lutz, 1996). Less formal, or unpublished, projections are also developed by the