TCS Daily

Climate Models and Consensus Science

By Garth W. Paltridge - May 11, 2005 12:00 AM

It is fairly easy to calculate the likely rise of global average temperature DT for the purely theoretical situation where atmospheric carbon dioxide (CO2) is doubled but nothing else about the atmosphere is allowed to change. The answer is about 1.2 degrees Celsius, and it would take a couple of hundred years to complete the change. The problem is that, in the real world, all sorts of other atmospheric and oceanic processes that depend on surface temperature are happening. Many of them amplify or reduce the original change of 1.2 degrees caused directly by the CO2. They are called feedback processes, and in principle their effects need to be added up to give an overall value for the total feedback factor F in the bottom line of the equation shown here for DT. Thus the calculated temperature change may be greater or less than 1.2 degrees -- this depending entirely on the value of F and on whether it is positive or negative.


All the vastly complicated numerical climate models floating around these days are machines which (in effect, if not explicitly) calculate their own particular values of feedback factor F by simulating as many as possible of the detailed processes going on in the earth-atmosphere system. The value of F for each model, and consequently the forecast by each model of the rise in global average temperature, depends largely on the particular set of tuneable parameters that have been chosen to make the models simulation of present climate look reasonable.


The basic point to remember is that we are not even sure in most cases whether the real values of the individual feedback factors making up the total F are positive or negative -- let alone knowing their values to a couple of decimal places. As a consequence, the actual change of global temperature for a doubling of CO2 might lie anywhere on the solid black line of the attached diagram, which is a graph of the equation for DT above -- that is, a graph of the change in temperature as it depends on whatever is calculated as the total feedback factor F.  



Considering all the fairly respectable models of the last twenty years or so, their answers cover the extensive range along the solid line from about F = -0.3 to F = 0.95. If we look only at the particular models selected by the UN IPCC (the Inter-governmental Panel on Climate Change) in its last report in 2001, the answers lie within the more restricted range indicated by the edges of the cross-hatching. That is, to a range of F between (roughly) 0.4 and 0.8, and to a corresponding rise of temperature between (roughly) 2 and 5 degrees.


The apparent convergence of the predictions of the IPCC models into a narrower range of possible temperature rise has to be taken with a considerable grain of salt. It is not necessarily an indication we are getting better at forecasting the future.


Because the modern models are so incredibly complex, it is very rare that a group of researchers develops a new climate model from scratch. They take some or all of the code from an existing model of another group, and slightly modify those bits of it that are relevant to their particular interest and expertise. The overall process ensures that there can be a gradual, and largely unconscious, move to a situation where all the supposedly independent models have common physics and common values for their tuneable parameters. They will quite naturally -- but not necessarily for a good physical reason -- begin to tell the same story. The danger is that the narrowing range of answers from the various models will be interpreted by scientists as an indication that the accuracy of their forecasts is improving.


Indeed the climate models are now so complex that no scientist outside the closed engineering shop of the numerical modelling community can ever really hope to assess whether or not the physical representations within them are acceptable. The normal and necessary process of scientific criticism cannot take place. All that can happen is for sceptics to take pot-shots at the modellers purely on the basis of intuition, and for the modellers to defend themselves by saying that no-one outside the modelling game has the knowledge to make sensible comment.


It is not the way to run the scientific railroad.


The very least that needs to be done if the IPCC is to maintain its credibility is to insist that all models used in IPCC assessments must calculate and publish the implicit individual feedback factors built into their calculations of the total feedback F. Then (at last?) the climate scientists of the outside world will have some understandable physics on which their intuition can work, and perhaps also a guide to the design of real-world experiment and observation so as to improve the modellers arbitrary selection of tuneable parameters.   


In other words we have to get away from simply running models and comparing their final output in some sort of search for a consensus on the results. Consensus is not science. Consensus tends to the politically correct. Consensus is not the sort of thing on which sensible people put their money.


There is an important aside to all this. Whatever is the total feedback factor in the real world, the temperature change is always going to be positive. In this restricted sense the bald statement that "the science of greenhouse warming is proven" is indeed acceptable. Increasing CO2 will certainly lead to higher temperatures than would have occurred otherwise. What is not acceptable is the corollary rather naughtily implied by those who loudly proclaim the statement in public -- namely, that the rise in temperature will be so large as to be disastrous, or even so large as to be noticeable. 


Emeritus Professor Garth Paltridge is retired Director of the Institute of Antarctic and Southern Ocean Studies, University of Tasmania, and former Chief Research Scientist of the CSIRO Division of Atmospheric Research. This article is based on a paper given at a Tech Central Station-sponsored conference, Managing Climate Change -- Practicalities and Realities in a post-Kyoto Future held last month in Canberra.



TCS Daily Archives