House of Lords - Economic Affairs

Select Committee on Economic Affairs Written Evidence

Memorandum by the International Policy Network

THE SCENARIOS UNDERLYING CLIMATE CHANGE "PREDICTIONS"

MARTIN ÅGERUP

Scientists are in broad agreement that human activities have some influence on global mean temperatures and climate, but they disagree about the extent of this influence. Absent the influence of humanity, the earth's climate would not be stable—it experiences extreme natural changes, and small manmade temperature increases are likely not to be a huge problem.

It is almost as uncontroversial that uncontrolled human emissions of greenhouse gases (GHGs) will result in an increase in global mean temperatures—all other things being equal. This does not in itself justify political action in general or climate mitigation in particular. The earth's climate would not be stable without human influence. In fact climate has always changed and will always change. Therefore, small manmade temperature increases are not a problem.

Some greenhouse gases are more effective per molecule when it comes to trapping infrared radiation than others. In order to simplify a matter slightly, the concentrations of the different GHGs in the atmosphere are converted into CO2-equivalents on the basis of their relative radiative potency. The CO2-equivalent atmospheric concentration is now about 32 per cent higher than the pre-industrial level. The concentration is measured in parts per million (ppm) air molecules. The pre-industrial concentration of CO2-equivalent was approximately 280 ppm while the current level is approximately 370 ppm. The IPCC projects a concentration in the year 2100 that ranges from 540 to 970 ppm.

Other things being equal, this development will enhance the natural greenhouse effect and cause the global mean temperature to rise. Most scientists agree on this. What is unresolved is the question of how much temperatures will rise. That is not just an academic issue. If human emissions of GHGs have a small effect on climate, it may prove too costly to curb these emissions compared to the benefits. The issue is further complicated by the fact that small increases in temperatures may in fact be an advantage.

It is important to establish by how much temperatures can be expected to rise over a relevant time scale because of the influence of human activities. The Intergovernmental Panel on Climate Change (IPCC) has attempted this exercise for the past 15 years. The IPCC was established in 1988 by the World Meteorological Organization (WMO) and the United Nations Environment Programme (UNEP). The role of the IPCC is to assess

the scientific, technical and socio-economic information relevant to understanding the scientific basis of risk of human-induced climate change, its potential impacts and options for adaptation and mitigation. The IPCC does not carry out research nor does it monitor climate related data or other relevant parameters. It bases its assessment mainly on peer reviewed and published scientific/technical literature.

In its third assessment report (2001), the IPCC predicted that the global mean surface temperature would rise by 1.4 to 5.8ºC[47] and came to the widely-quoted conclusion that

There is new and strong evidence that most of the warming observed over the last 50 years is attributable to human activities.

However, the IPCC's conclusions are highly controversial. The predictions depend on two sets of computer models. One set predicts how much CO2 and other greenhouse gasses (GHGs) that humanity will emit into the atmosphere. The other set of models predict how the climate system will react to such increases in GHGs.

In other words, the quality of the predictions depends completely on the ability of the models to produce reliable results. In fact, there are problems with both sets of computer models. These are the two main uncertainties with these models:

1. We don't know how sensitive the Earth's climate system is to increased levels of CO2.

2. We don't know how much CO2 we will emit—the subject of this paper.

THE IPCC'S EMISSIONS SCENARIOS

The IPCC's climate models calculate the consequences of increasing atmospheric GHG concentrations—typically at a level equivalent of a doubling of atmospheric carbon dioxide compared to the pre-industrial level. However, in order to predict expected warming, modellers need estimates of the rate of CO2 increase in order to forecast how long it will take for concentration to double. In other words, they need forecasts of human emissions of CO2 through the 21st century.

The IPCC has made such long range forecasting exercises for all of their three assessment reports since 1992. The latest results were published in 2000 in the Special Report on Emissions Scenarios (SRES).[49] The SRES working group produced 40 scenarios out of which six were chosen as "marker scenarios" and fed into the climate models. In other words the emissions scenario exercise is a crucial step in creating the above mentioned temperature growth range for 2100 of 1.3 to 5.8ºC. In fact, most of the span in this temperature range is produced by the huge difference in emissions across the marker scenarios.

The scenarios are based on parameters influencing emissions such as:

— GDP per capita growth on a country, regional and global level.

— Population growth.

— Energy efficiency (how much GDP does one unit of energy produce).

— Composition of fossil fuel consumption (coal, oil, gas).

— Non-carbon fuel share of total energy production (nuclear, wind, solar etc.)

EVALUATION OF THE EMISSIONS SCENARIOS

Obviously the quality of the emissions forecasts depend completely on the way these parameters are treated and the values given to them in the different scenarios. Over a time span of a century the degree of uncertainty for each parameter is enormous and as explained below, some of these uncertainties compound. However, this does not mean that anything should be considered plausible. Unfortunately the SRES scenario group does a very poor job of creating plausible scenarios.

For those who use scenario methods in their professions, the SRES scenario effort is problematic. There are three basic reasons for concern:

— Sloppiness.

— Bad methods.

— Scenarios being used as forecasts—the most important criticism.

The criticism concerning sloppiness includes treating the period 1990—2000 as part of the forecast period without taking into consideration that we now have data for that period. Apparently the SRES has been reusing data from previous scenario exercises. The problem is that time and real world observations have proven these projections wrong. For instance the scenario figures used for the increase in world GDP between 1990 and 2000 vary between 20.6 per cent and 35.4 per cent. However, IMF data for most of that period was available in 1999 and showed growth of 36.5 per cent.[50] Probably this error doesn't have a substantial impact on model results by 2100. Nevertheless, it's amazing that such sloppy practices are allowed as part of an input to a modelling exercise that involves the use of supercomputers and costs millions of Euros.

The IPCC also suffers from poor methodology. One leading economic modeller, John Reilly of the MIT Joint Program on the Science and Policy of Global Change calls the SRES approach to scenario building an "insult to science."[51] According to Reilly, the scenario teams have worked backward from a desired end result in terms of emissions and temperature increases. In other words, the IPCC has allegedly started with an emission projection then made an estimate of the relationship between emissions and growth and finally calculated the growth rate needed to achieve the desired emissions projection.[52]

Another criticism of methods has been advanced by Ian Castles, former President of the International Association of Official Statistics, and David Henderson, former Chief Economist of the OECD. They accuse the SRES team of using inappropriate exchange rates which exaggerate current global economic inequality. Since the scenarios assume that income equality will improve over the course of the century, models overestimate growth rates in low income countries.[53]

While the SRES methods concerning growth projections are technically unsound, the scenario models will have to be modified before we can conclude if the exchange rate error should lead to substantial reductions in total global growth rate projections. However, the "end result" in terms of world GNP/GDP does seem very high. Many of the scenarios give us a world economy which by 2100 is up to approx. 25 times larger than it is today.[54]

In order to achieve that multiplication the world economy would have to grow 3 per cent a year. This does not seem to be a realistic assumption, even for a high end scenario, given the historical performance of the world economy. But then, the SRES does not seem to take historic trends much into consideration. As David Henderson points out:

a surprising feature of the SRES that in a document surveying the long term future, which contains over 300 pages of main text and presents 40 different scenarios prepared by six different modelling groups, there is no chapter which systematically reviews the evidence of the past. The starting point for any such quantitative future-oriented inquiry should be a clear and careful survey of earlier developments and trends, going right up to the present.

What such a survey could have considered is the fact that since 1975 world GDP per capita has grown at an average 1.2 per cent per annum.[56] Extending this growth rate for 110 years would give us a world per capita income which was 3.7 times larger than the present. If we assume that population doubles over the same period (and that would be a high estimate) we get a global economy which is 7.4 times bigger than the present. That's quite far from a 25 time increase.

The only economy which has produced average annual growth rates comparable to the IPCC assumption over a sustained period is Japan. Between 1913 and 1996 Japan grew at an annual rate of 3.36 per cent.[57] But it seems doubtful that the entire world economy would be able to do the same for as long as a century.

THE USE AND ABUSE OF SCENARIOS

The most serious objection is the IPCC's use of scenarios. The scenarios are presented as an exercise in "free thinking" about the future. The SRES states that they are "images of the future or alternative futures" and should not be seen as predictions or forecasts. They are in fact "computer aided storylines". But at the end of the day the scenarios end up presenting a very concrete result in terms of numbers (emissions). These numbers are fed into a computer, which in turn produces a temperature. The SRES tries to have it both ways: a non-committal scenario process and a clear result in terms of a number.

The SRES claims that the scenarios "are not assigned probabilities of occurrence, neither must they be interpreted as policy recommendations."[58] But since each scenario gives a result which translates into a number there is in fact an implicit bias in favour of extreme scenarios. An extreme scenario extends the temperature range. A less extreme scenario doesn't. Therefore, an extreme scenario is implicitly given more importance. In the real world we would normally treat an extreme outcome as less likely and therefore assign less importance to it.

But the SRES scenarios are not assigned probabilities. That's normal when working with scenarios. The whole point of a scenario exercise is to cover the full range of possible futures. But this also implies that scenarios should not be used as forecasts, because a forecast doesn't make sense without a discussion of probability. Meteorologists only make weather forecasts extending 3-5 days into the future. The reason why they don't make longer forecasts is that the probability of being right gets too small. The SRES scenarios are used to make forecasts. In other words, the SRES team abuses the scenario method.

IMPROBABLE HIGH END SCENARIOS

The following section shows that the high end temperature projection (which is being used as a forecast) is based on two scenarios which again are based on assumptions that have a very low probability, especially combined.

The six marker scenarios were based on very different assumptions about the parameters listed above. It is therefore not surprising that they produce very different result in terms of carbon emissions as shown in Figure 1.[59]

At the low end, the B1 scenario produces cumulative emissions by 2100 of 983 gigatons. At the high end the A1FI scenario produces emissions of 2.189 gigatons, more than twice as much.

How do those results compare with historic emissions? Figure 2 shows emissions per capita over the past half century.[60]

Until the beginning of the seventies carbon emissions per capita grew, but since then they have stabilised and even declined slightly. This has happened despite considerable economic growth.

Ross McKitrick, Associate Professor at the University of Guelph, concludes that:

there is reason to believe that per capita CO2 emissions are somewhat invariant to economic growth, at least at a globally averaged level.

It would therefore seem reasonable that per capita carbon consumption in at least some scenarios was near the current level of 1.1 tonnes of carbon. Nevertheless, the SRES scenarios are all above 1.2Ct per capita by 2020. Not one scenario follows the current trend of no growth!

The high end scenarios produce some results which are improbable. Figure 3 is based on two historic trend lines of carbon consumption per capita. One is a projection of the high growth trend from 1950-73. It's certain that no time period in human history has had a higher growth rate in carbon consumption per capita than this period. The other line shows the lower growth trend for the whole period from 1950 to 1999.

Figure 3 shows how the A1FI scenario[62] overtakes even the fast growth trend by 2050. By 2100 , per capita emissions in the A1FI scenario are more than four times the current level, and 25 per cent above the high growth historic trend which was only sustained for a couple of decades in reality.

Not surprisingly the scenario also arrives at very high estimates of atmospheric concentration of CO2. Currently, CO2-levels are increasing by approximately 0.4 per cent a year.[63] Figure 4 shows how, by 2100, atmospheric CO2 concentrations at the high end of the A1FI scenario reach a level 2.27 times what the current trend would imply.[64] The fact that the 1990s had both high global economic growth and probably the highest nominal population growth ever (including all of the 21st century[65]) suggests that annual CO2-growth may not increase much above the current 0.4 per cent.

The current trend for the second most important GHG, methane, is that the rate of growth has been decreasing and the actual concentration is currently falling. Yet none of the scenarios follow that trend. Figure 5 shows how the A1FI scenario predicts a concentration more than double the current trend.[66]

It is reasonable to conclude that the A1FI scenario depicts an extremely unlikely future. Nevertheless, it is this scenario—and this scenario alone—which is responsible for the high end of the IPCC's Third Assessment Report temperature range of 1.3 to 5.8ºC for the year 2100. The 5.8º forecast is arrived at by running the GHG concentration figures of the A1FI scenario through the climate model with the highest sensitivity to increased CO2. If it wasn't for this scenario, the top temperature would be 1ºC lower.[67]

If we ignore the 1ºC which is produced by the A1FI scenario, we get a high temperature estimate of approx. 4.8ºC. This is generated by running the A2 scenario through the most GHG sensitive climate model. The A2 scenario predicts yearly CO2 emissions by 2100 which are almost as exorbitant as the A1FI figures. But this scenario gets there in a different way. One of the tricks of the A2 scenario is a very high population projection.

Figure 6 shows the 2002 UN Population projection[68]. The medium population projection for 2050 is 8.9 billion. The high estimate is 10.6 billion. Nevertheless, the A2 scenario has 11.3 billion people in 2050, a figure above the UN's high estimate.

Unfortunately, the latest UN projection doesn't go beyond 2050. But the 2000 median projection estimated that world population would be 10.4 billion in 2100. In the new projection, the median figure for 2050 is revised downward from 9.3 to 8.9 billion[69] which implies that the forecast for 2001 would also be lower today. The A2 scenario assumes a world population in 2100 of 15.1 billion[70]—50 per cent above the median estimate. The A2 population is below the UN high projection of 18 billion, but few people consider that to be a likely outcome. It's not really a projection but a kind of "business-as-usual" scenario.

Secondly, and much more importantly, as figure 7 illustrates, the A2 scenario has a much higher share of coal than any of the other scenarios. This is significant since coal emits more carbon per unit of energy than oil and especially gas. On top of that A2 has a much lower share of zero carbon energy (renewable sources, nuclear etc.).[71] In short, the A2 scenario assumes virtually no advances in energy technology over the next century. In fact, the scenario assumes that a trend towards less carbon intensive energy sources—a trend which has persisted for over a century—is actually reversed.

It is interesting to compare the A2 scenario with the A1T scenario. By 2100, the A2 has the second lowest world GDP of all six marker scenarios: 243.000 billion (measured in USD at 1990). In comparison the A1T scenario forecasts a world economy of 550.000 billion (1990 USD)—more than twice that size.[72] Still, the A2 world emits 28.9 Gt of carbon per year, which is 6.7 times more than the 4.3 Gt emissions that the A1T world produces. In other words, the carbon intensity of the A2 economy is 15 times higher—this world needs fifteen times more carbon per unit of GDP.

This difference illustrates the importance of expectations about future energy technology. The A2 and the A1T scenarios represent two extremes in this respect. In A2 there is very little technological development, so the world returns to coal as the primary source of energy. In A1T there is rapid technological development, and non-carbon energy technologies begin to replace fossil fuels by the middle of the century. Are both scenarios likely? The answer is probably no. The A1T scenario is likely, whereas the A2 scenario is not. The reason for this is that the A2 scenario reverses a very long term trend whereas the A1T scenario merely continues (and maybe accelerates) an existing long term trend.

Despite the fact that the share of coal in world energy supply has been decreasing over the whole of the 20th century[73] and no reversal of that trend seems in sight, the share of coal increases in the A2 scenario from the current level of 23.5 per cent[74] to 53 per cent.

For more than a hundred years the global carbon dioxide intensity of energy has been decreasing steadily, by almost a third[75]. The A2 scenario inexplicably reverses this trend of decarbonisation.

Most of the literature on the subject of energy technology expects that one or more non-carbon technologies will become competitive in the course of the 21st century. Both photovoltaic sources (solar energy) and windmills show historic learning curves with a learning rate of approximately 20 per cent. In other words, a doubling of cumulative installed capacity gives a 20 per cent reduction in costs. At that rate, wind energy will be competitive within a decade or two and photovoltaic energy around 2050. Even models which are less optimistic about the learning curve have alternative non-carbon energy sources overtaking fossil fuels before the end of the 21st century.[76]

For the time being, gas is becoming an increasingly attractive alternative in much of the world and the price of gas-generated electricity is falling. This is another trend which is reversed in the A2 scenario.

While nuclear energy has stumbled in Europe for political reasons, other parts of the world are expanding their nuclear energy base rapidly, particularly Asia. Nuclear energy's per centage share of global energy supply has expanded from 0.9 per cent in 1973 to 6.8 per cent in 2000. There is also a learning curve for nuclear energy, so the expansion of this energy source results in it becoming ever cheaper, so nuclear energy could be expected to play a much larger role in the future.[77]

To summarise, while both the A1FI and the A2 scenarios are possible scenarios and maybe even "good" scenarios within a framework of six very different marker scenarios, they are not suitable for the forecast exercise that the emissions output constitutes. Both scenarios have a very low probability since they are based on a number of assumptions which seem unlikely given historical trends.

CONCLUSION: HOW MUCH WARMER WILL THE CLIMATE BE BY 2100?

Table 1 summarises the basic steps that the IPCC needs to undertake in order to achieve the expected warming interval (left hand column). For each step it lists what are widely considered to be the principal uncertainties given current scientific understanding (right hand column).

  Table 1:

Forecast exercise Principal uncertainties

Create scenarios for future emissions of CO2 —World GDP growth per capita and its distribution among   Low, Middle and High income countries.

—Population growth

—Composition of different fossil fuels in total consumption

—Technological change, including shifts to non-carbon or low-  carbon energy sources, energy efficiency, carbon sinks etc.

Convert emissions to atmospheric concentrations —The life time of different GHGs in the atmosphere

Model radiative forcing and convert this forcing to a projected temperature —The sensitivity of the climate system to increased CO2   (feedback effects of clouds, aerosols etc).

  Natural climate effects enhancing or counterbalancing   manmade effect

Through the forecast exercise, total uncertainty compounds as each new uncertainty is added. Each of the parameters has a factor 2 uncertainty. The highest estimate is double the lowest estimate. But the result has a factor 4 uncertainty. The highest estimate is four times higher than the lowest.

Adding a third parameter with a factor 2 uncertainty would gives us a high estimate which was 8 times higher than the low estimate. Uncertainty compounds.

The short answer is that we simply do not know how much warmer climate will be in 2100. In fact, the degree of (compound) uncertainty is so large that merely by providing temperature intervals, the IPCC is extremely misleading. For many of the parameters even the degree of uncertainty is controversial, despite the IPCC's phoney confidence intervals. Climate science is not at a stage where it is capable of providing confidence intervals, especially not for the earth's climate almost one hundred years into the future. In fact even the term "uncertain" is often misleading when it comes to climate science. Many things are not uncertain but simply unknown.

The emissions scenarios produce growth rates of carbon emissions which are not in line with recent history. Especially the high end scenarios, such as the A1F-scenario, seem completely unrealistic. Nevertheless, the A1F-scenario provides the basis for the high end of the IPCC temperature range. According to the guidelines for scenario use that most forecasters and futurists refer and adhere to, a scenario must be both possible and probable. While the A1F-scenario is theoretically possible (and even that is debatable), it simply isn't probable.

As mentioned, temperature intervals do not make sense given current knowledge about the climate system. However, since this is the game that the IPCC has forced upon us, we should examine the interval of 1.5—5.8ºC temperature increase which is the central IPCC projection. This interval is almost certainly way too high. The upper estimate should be decreased by about 1ºC because it is based on the unrealistic emissions of the A1F-scenario, and by a further 1ºC which results from the unrealistic A2 scenario.

Figures

28 February 2005

References

Chakravorty, Ujiayant and Tse, Kin-Ping (no date). "Transition from Fossil Fuels to Renewable Energy", http://www.earthinstitute.columbia.edu/events/econSeminar/Chakravorty.pdf

Commission of the European Communities (2002), COM 702. Available at http://europa.eu.int/eurlex/en/com/rpt/2002/com2002—0702en01.pdf (last visited 13 July 2003).

Corcoran, Terence (2002). "An `insult to science'", National Post, 14 December. Available at

http://www.canada.com/national/nationalpost/financialpost/story.html?id=%7b54b082f6-100c-4ada-9012-edf7ec30dd03%7d

Den Elzen, M.G.J. and Both, S. (2002). "Modelling emissions trading and abatement costs in FAIR 1.1—Case study: The Kyoto Protocol under the Bonn-Marrakech Agreement", RIVM (National Institute of Public Health and the Environment, Denmark) report 728001021, available at

http://www.rivm.nl/bibliotheek/rapporten/728001021.html

The Economist (2003), "Hot Potato", 13 February.

Eurostat (2002), Yearbook, Brussels, European Commission.

Gray, Vincent (2002). The Greenhouse Delusion. Brentwood, Essex, UK: Multi-Science.

Grubb, Michael (forthcoming) "The Economics of the Kyoto Protocol" in Owen, Anthony D (Ed.) The Economics of the Kyoto Protocol.

Henderson, David (2003), "SRES and IPCC: Further Concerns", available at

http://www.economist.com/displaygeneric.cfm?pageheadgif=FinanceandEconomics&key=efhp1

pp. 7-10. See also: Castles, Ian and Henderson, David (2003). Correspondence with the SRES

http://www.economist.com/displaygeneric.cfm?pageheadgif=FinanceandEconomics&key=efhp1;

Castles, Ian and Henderson, David (2003). "The IPCC Emission Scenarios: An economic-statistical critique" in Energy and the Environment, Vol.14, Nos.2&3, pp.159-185.

International Energy Agency (IEA)(2002a), Beyond Kyoto: Energy Dynamics and Climate Stabilisation, Paris, OECD.

IEA (2002b), Key World Energy Statistics, http://www.iea.org/statist/keyworld2002/key2002/keystats.htm

Intergovernmental Panel on Climate Change (IPCC)(1996) Climate Change 1995—Scientific-Technical Analyses of Impacts, Adaptations and Mitigations of Climate Change. Report of IPCC Working Group II. Cambridge, Cambridge University Press.

IPCC (1998), The Regional Impacts of Climate Change: An Assessment of Vulnerability

IPCC (2000), Special Report on Emissions Scenarios, Summary for Policymakers, http://www.ipcc.ch/pub/sres-e.pdf

IPCC (2001), The Scientific Basis: Technical Summary, Climate Change 2001, Working Group I, available at http://www.grida.no/climate/ipcc—tar/wg1/010.htm

Landes, David (1998), The Wealth and Poverty of Nations, London, Abacus.

Lomborg, Björn (2001), The Skeptical Environmentalist, Cambridge, Cambridge University Press.

Maddison, Angus (1991). Dynamic Forces in Capitalist Development: A Long-Run Comparative View (New York: Oxford University Press). See also Maddison, Angus (2002). The World Economy: A Millennial Perspective (Paris: OECD Development Centre)

Magné, Bertrand and Moreaux, Michel (2002) "Long Run Energy Trajectories: Assessing the Nuclear Option In Response To Global Warming" (Preliminary Draft). February. http://weber.ucsd.edu/carsonvs/papers/617.pdf

Marland, Gregg (no date). Tom Boden Carbon Dioxide Information Analysis Center, Link:

http://cdiac.ornl.gov/ftp/ndp030/global99.ems

McKitrick, Ross (2003). "Emission Scenarios & Recent Global Warming Projections", Fraser Forum, January.

Nakicenovic, N. and Riahi, K (2002) "An Assessment of Technology Change", International Institute for Applied Systems Analysis, 2002. Available at http://www.iiasa.ac.at/Publications/Documents/RR-02-005.pdf

National Academy of Sciences (NAS)(2001) Climate Change Science: An analysis of some key questions, Washington, DC, National Academy Press. Available at http://www.nap.edu/html/climatechange/

Nordhaus, William D. (2001) "Global Warming Economics", Science. 294, pp.1283-4. (9 November).

Nordhaus, William D. and Boyer, Joseph G. (1999). "Requiem for Kyoto: An Economic Analysis of the Kyoto Protocol", (February 8), available at http://www.econ.yale.edu/nordhaus/homepage/Kyoto.pdf

Nordhaus, William D. and Boyer, Joseph G. (1999) "Roll the Dice Again: Economic Models of Global Warming", MIT Press (25 October), internet version:

http://www.econ.yale.edu/%7Enordhaus/homepage/dice—section—I.html

UNDP (2002), Human Development Report, Oxford, Oxford University Press.

http://hdr.undp.org/reports/global/2002/en/

UNPP (2002), World Population Prospects: The 2002 Revision. http://esa.un.org/unpp/

Webster, et. al (2001). "Uncertainty in Emissions Projections for Climate Models", MIT Joint Program on the Science and Policy of Global Change, Report No. 79 (August)

Available at http://www.mit.edu/afs/athena.mit.edu/org/g/globalchange/www/MITJPSPGC—Rpt79.pdf