Scientific forecasting methods should be used to project climate change.  The methodologies used should be those shown empirically to be relevant to the particular types of problems involved in climate forecasting.  However, the evidence shows that the IPCC forecasts are instead based on opinions.

Many public-policy decisions are based on unaided expert judgments.  The experts may have access to empirical studies and other information, but they often make predictions, or judgmental forecasts, without the aid of scientific forecasting principles.  Research on persuasion has shown that people have substantial faith in the value of such forecasts, and this faith increases when experts agree with one another.  However, the opinions of experts are an invalid basis for public policy.

“The opinions of experts are wrong as often as the opinions of nonexperts.”

Judgmental Forecasts.  Comparative empirical studies have routinely concluded that judgmental forecasting by experts is the least accurate of the methods available to make forecasts.⁵ For example, Professor Phil Tetlock, of the University of California at Berkeley, conducted a major study in which he recruited 284 participants whose professions included, “commenting or offering advice on political and economic trends.”⁶ He asked these experts to forecast the probability that various events would or would not occur, picking areas (geographic and substantive) within and outside their areas of expertise.  By 2003, he had accumulated over 82,000 forecasts.  The experts barely outperformed nonexperts and neither group did well against simple forecasting rules that extrapolate from the past to predict the future.⁷

Examples of expert climate forecasts that turned out to be completely wrong are also easy to find.  [See the sidebar on “Climate Forecasts Based on Expert Opinion.”]

Computer Modeling versus Scientific Forecasting.  Over the past few decades, the methodology used in climate forecasting has shifted so that expert opinions are informed by computer models.  Advocates of complex climate models claim that they are based on well-established laws of physics.  But there is clearly much more to the models than physical laws, otherwise the models would all produce the same output, which they do not, and there would be no need for confidence estimates for model forecasts, which there certainly is.  Climate models are, in effect, mathematical ways for experts to express their opinions.⁸

There is no empirical evidence that presenting opinions in mathematical terms rather than in words improves the accuracy of forecasts.  In the 1800s, Thomas Malthus forecast mass starvation.  Expressing his opinions in a mathematical model, he predicted that the food supply would increase arithmetically while the human population would grow at a geometric rate and go hungry.  Mathematical models have not become much more accurate since.

“Computer modeling does not improve the accuracy of opinion.”

Two international surveys of climate scientists from 27 countries, in 1996 and 2003, show increasing skepticism over the accuracy of climate models.⁹ Of more than 1,060 respondents, only 35 percent agreed with the statement, “Climate models can accurately predict future climates,” whereas 47 percent disagreed.  [See Figure I.]¹⁰

Problems with Climate Models.  Researchers who have examined long-term climate forecasts have concluded they are based on nothing more than scientists’ opinions expressed in complex mathematical terms, without valid evidence to support the chosen approach.¹¹

For example, when computer simulations project future global mean temperatures with twice the current level of atmospheric CO2, they assume that the temperature forecast is as accurate as a computer simulation of present temperatures with current levels of CO2.¹² Yet it has never been demonstrated that temperature forecasts are as accurate as simulations of current conditions based on actual temperature data.  Indeed, there are even serious questions surrounding model simulations of current temperature with current CO2 levels.  According to the models, the earth should be warmer than actual measurements show it to be, which is why modelers adjust their findings to fit the data.

“Climate models based on historical data don’t accurately predict current temperatures.”

The models do not represent the real world sufficiently well to be relied upon for forecasting.¹³ As physicist Freeman Dyson concluded, climate models “do a very good job of describing the fluid motions of the atmosphere and the oceans,” but “they do a very poor job of describing the clouds, the dust, the chemistry and the biology of fields and farms and forests.”¹⁴

The climate models’ simulations do not correspond to past or present temperatures, rainfall patterns, tropical cyclones or plant responses.  Professor Bob Carter, of the Marine Geophysical Laboratory at James Cook University in Australia, examined evidence on the predictive ability of the general circulation models (GCMs) used by the IPCC scientists.  He found that while the models included some basic principles of physics, scientists had to make a number of “educated guesses” because knowledge about the physical processes of the earth’s climate is incomplete.¹⁵ Thus, in practice:

• The GCMs failed to predict recent global average temperatures as accurately as fitting a simple curve to the historical data and extending it into the future.
• The models forecast greater warming at higher altitudes in the tropics, when the greatest warming has occurred at lower altitudes and at the poles.
• Furthermore, individual models have produced widely different forecasts from the same initial conditions, and minor changes in their assumptions can produce forecasts of global cooling.¹⁶

When models predict global cooling, the forecasts are rejected by modelers as “outliers” or “obviously wrong.”  This suggests that when the models are averaged together to create consensus estimates of temperature change, the results are biased due to the omission of models that show cooling.¹⁷

Researchers have found serious deficiencies in the GCMs on which the IPCC based its previous Third Assessment Report.  For example, David Bellamy and Jack Barrett found: ¹⁸

1. The models produced very different cloud distributions and none came close to the actual distribution of clouds, which is important because the type and distribution of clouds can either enhance or reduce the Earth’s temperature.  Some clouds tend to trap more of the sun’s radiant heat, while others reflect the sun’s rays away from the Earth’s surface, mitigating the effect of increased greenhouse gases.
2. Assumptions about the amount of radiation from the Sun absorbed by the atmosphere and the Earth’s surface varied considerably from model to model, yielding widely varying temperature forecasts.  This is important because absorption of heat energy from sunlight is the primary mechanism of global warming.
3. The models did not accurately represent the known effects of CO2, much less the uncertain possible feedbacks that reduce or enhance those effects, and as a result their climate change forecasts cannot be relied upon.

The review by Bellamy and Barrett concluded: “The climate system is a highly complex system and, to date, no computer models are sufficiently accurate for their [IPCC] predictions of future climate to be relied upon.”¹⁹ And since climate model forecasts for periods of up to five years have proven to be inaccurate, the review concluded there is little basis upon which to make accurate projections for 50 to 100 years.²⁰

Referring to the GCMs used for the IPPC forecasts, a lead author of Chapter 3 of the Fourth Assessment Report wrote that “… the science is not done because we do not have reliable or regional predictions of climate.”²¹

“Climate models based on recent data can’t accurately predict temperatures five years in the future.”

Other agencies’ attempts to forecast for shorter periods and smaller geographical areas have also been unsuccessful.  For instance, annual forecasts by New Zealand’s National Institute of Water and Atmospheric Research (NIWA) are no more accurate than chance.²² NIWA’s low success rate is comparable to other forecasting groups worldwide, according to New Zealand climatologist Jim Renwick, a member of the IPCC Working Group I and a coauthor of the Fourth Assessment Report.  (Renwick also serves on the World Meteorological Organization Commission for Climatology Expert Team on Seasonal Forecasting.)  He concludes that current GCMs are unable to predict future global climate any better than chance.²³