Personal view: The science behind climate change forecasts adds up to a lot of hot air, says Martin Agerup

By Martin Agerup

The Daily Telegraph
May 3, 2004

HOW would you react if fellow members of your profession were violating good practices of that discipline? What if the same people had massive amounts of taxpayers' money to develop flawed ideas, and wielded influence over public policies? Most people would be appalled. Yet this is exactly what is happening with the climate scenarios developed by the Intergovernmental Panel on Climate Change (IPCC), the body charged with advising governments on the causes and consequences of climate change, which met in Geneva last week. The IPCC has received much attention for its projections, which suggest that the earth's climate will warm by anywhere between 1.4°C and 5.8°C in the next century. These projections are largely based on scenarios about how people will use energy in the future, which in turn determine future emissions of greenhouse gases. These scenarios are developed by economists (not climate scientists, as many are led to believe by misleading press releases issued by the IPCC).

The scenarios are presented as an exercise in free thinking about the future. The Special Report on Emission Scenarios (SRES) published in 2000 describes them as "images of the future or alternative futures" which should not be seen as predictions or forecasts, rather as "computer-aided storylines." Unfortunately, they appear to present a very concrete result in terms of numbers for emissions. When fed into a computer, the results are more numbers, the 1.4°C and 5.8°C range for temperature rises. The SRES tries to have it both ways: a noncommittal scenario process and a clear numerical result. The SRES claims that the scenarios "are not assigned probabilities of occurrence, neither must they be interpreted as policy recommendations", but since each scenario gives a result which translates into a number, there is an implicit bias towards the extremes.

In the real world, scenario builders would normally treat an extreme outcome as less likely and therefore assign it less importance. It is normal not to assign probabilities when working with scenarios, since the point is to cover the full range of possible futures. By the same token, scenarios should not be used as forecasts, because a forecast makes no sense without a discussion of probability. Meteorologists only make weather forecasts three to five days into the future because of the tiny probability of being right with longer forecasts. Yet the SRES scenarios are indeed used to make forecasts, far into the future. Two out of the six so-called "marker scenarios" push up the IPCC's temperature range by 2°C, but neither is likely given historical trends. The first, which produces global warming of 5.8°C, assumes that by 2100, emissions will be more than twice what the historic growth trend would imply. To this must be added the least likely of the climate models, which assumes a very high sensitivity to increased carbon dioxide, to get to the apocalyptic number of 5.8°C. This scenario is entirely unrealistic.

The second absurd scenario, which produces global warming of 4.8°C, uses a projection for world population growth which is the UN's top estimates, while assuming a reversal of the global trend of the past 100 years towards less carbon-intensive energy sources. Again, this is extremely unlikely. Others have criticised the SRES methodology and practices. John Reilly of the MIT Joint Program on the Science and Policy of Global Change, calls the SRES approach an "insult to science". Reilly suspects that the scenario teams started with an emissions projection, estimated the relationship between emissions and growth, and finally calculated the growth rate needed to achieve the desired emissions projection. Ian Castles, former president of the International Association of Official Statistics, and David Henderson, former chief economist of the OECD, have criticised the SRES for using market exchange rates instead of purchasing power parity. This results in models that overestimate future growth rates in poor countries. Some of the SRES scenarios produce a world economy which, by 2100, is up to 25 times larger than it is today. Few would complain if this happens, but history suggests that it won't. Since 1975, world GDP growth per capita has averaged 1.2% annually, a rate which would produce a figure 3.7 times today's GDP per head by 2100. Even if world population doubles (a high estimate), the global economy would be only 7.4 times larger than at the base year (1990). With such sloppy practices, ignoring historical data and trends, and turning scenarios into forecasts, the SRES has misled the public and policy makers. This is serious because the results are being used as a basis for policies to regulate energy consumption - of which the Kyoto Protocol is the most obvious - that will cost billions of dollars and harm economic growth. Worse still, there is no proof they will have any measurable effect on the world's climate.

Martin Agerup is a Danish economist, economic historian and scenario expert, and president of the Danish Academy for Future Studies.