Climate’s uncertainty principle

Reposted from Dr. Judith Curry’s Blog, Climate Etc.

by Garth Paltridge

On the costs and benefits of climate action.

Whether we should do anything now to limit our impact on future climate boils down to an assessment of a relevant cost-benefit ratio. That is, we need to put a dollar number to the cost of doing something now, a dollar number to the benefit thus obtained by the future generations, and a number to a thing called “discount for the future”—this last being the rate at which our concern for the welfare of future generations falls away as we look further and further ahead. Only the first of these numbers can be estimated with any degree of reliability. Suffice it to say, if the climate-change establishment were to have its way with its proposed conversion of the global usage of energy to a usage based solely on renewable energy, the costs of the conversion would be horrifically large. It is extraordinary that such costs can even be contemplated when the numbers for both the future benefit and the discount for the future are little more than abstract guesses.

Assessment of the future benefit is largely based on two types of numerical modelling. First, there are the vast computer models that attempt to forecast the future change in Earth’s climate when atmospheric carbon dioxide has increased as a consequence of the human activity of burning fossil fuel. Second, there are the computer-based economic models which attempt to calculate the economic and social impact of the forecasted change of climate. Reduction of that impact (by reducing the human input of carbon dioxide to the atmosphere) is the “benefit” in the cost-benefit calculations.

Taking the climate change calculations first, it should be emphasized that in order to be really useful, the forecast must necessarily be of the future distribution of climate about the world—on the scale of areas as small as individual nations and regions. Calculating only the global average of such things as the future temperature and rainfall is not useful. The economic models need input data relevant to individual nations, not just the world as a whole.

Which is a bit of a problem. The uncertainty associated with climate prediction derives basically from the turbulent nature of the processes going on within the atmosphere and oceans. Such predictability as there is in turbulent fluids is governed by the size (the “scale”) of the boundaries that contain and limit the size to which random turbulent eddies can grow. Thus reasonably correct forecasts of the average climate of the world might be possible in principle. On the scale of regions (anything much smaller than the scale of the major ocean basins for example) it has yet to be shown that useful long-term climate forecasting is possible even in principle.

To expand on that a little, the forecasts of the global average rise in temperature by the various theoretical models around the world range from about 1 degree to 6 degrees Celsius by the end of this century—which does little more than support the purely qualitative conclusion from simple physical reasoning that more carbon dioxide in the atmosphere will increase the global average temperature above what it would have been otherwise. It does little to resolve the fundamental question as to what fraction of the observed rise in global surface temperature over the last thirty or so years (equivalent to a rise of about 1 degree Celsius per century if one is inclined to believe observations rather than the theory) is attributable to the human-induced increase in atmospheric carbon dioxide. There is still a distinct possibility that much of the observed rise in global temperature may be the result of natural (and maybe random) variability of the system.

While the forecasts of future global average climate are not really trustworthy and would probably not be very useful even if they were, the potentially much more useful forecasts of regional climates are perhaps just nonsense. A good example supporting this rather negative view of the matter is the variability of the set of hundred-year forecasts of the average rainfall over Australia. Each forecast was produced by one of the many climate models from around the world. The present-day measured average is about 450 millimetres per year. The forecasts for the next century range from less than 200 mm to more than 1000 mm per year. That sort of thing makes finding a model to support a particular narrative just too easy.

As a consequence, the economic models of the future of regions and nations are highly unreliable if only because their regional and national inputs of forecasted climatic “data” are unreliable. But to make matters vastly worse, the economic models themselves are almost certainly useless over time-scales relevant to climate. Their internal workings are based on statistical relations between economic variables devised for present-day conditions. There is no particular reason why these relations should be valid in the future when the characteristics of society will almost certainly have changed. As Michael Crichton put it: “Our [economic] models just carry the present into the future.” And as Kenneth Galbraith once remarked: “Economic forecasting was invented to make astrology look respectable.”

There is a lot of discussion among academics as to what should be an appropriate “discount for the future” to apply in the cost-benefit calculations associated with human-induced climate change. The discussion quickly becomes incomprehensible to the average person when phrases such as “cross generational wealth transfer” and “intergenerational neutrality” and so on appear in the argument. These are fancy terms supposedly relevant to what is essentially a qualitative concept of fairness to future generations. The concept is so qualitative that there is virtually no hope of getting general agreement as to how much we should spend now so as not to upset the people of the future.

There are two extremes of thought on the matter. At one end there are those who tell us that the present-day view of a benefit for future generations should be discounted at the normal rate associated with business transactions of today. That is, it should be something of the order of 5 to 10 per cent a year. The problem for the academics is that such a discount would ensure virtually no active concern for the welfare of people more than a generation or so ahead, and would effectively wipe out any reason for immediate action on climate. At the other end of the scale, there are those who tell us that the value of future climatic benefit should not be discounted at all—in which case there is an infinite time into the future that should concern us, and “being fair” to that extended future implies that we should not object to spending an unlimited amount of present-day money on the problem.

Academics tie themselves in knots to justify the need for immediate action on climate change. For example, we hear argument that “discounting should not be used for determining our ethical obligations to the future” but that (in the same breath) “we endorse a principle of intergenerational neutrality”—and then we hear guesses of appropriate discount rates of the order (say) of 1.5 per cent a year.

The significant point in this cost-benefit business is that there is virtually no certainty about any of the numbers that are used to calculate either the likely change of climate or the impact of that change on future populations. In essence it is simply assumed that all climate change is bad—that the current climate is the best of all possible climates. Furthermore, there is little or no recognition in most of the scenarios that mankind is very good at adapting to new circumstances. It is more than likely that, if indeed climate change is noticeably “bad”, the future population will adjust to the changed circumstances. If the change is “good”, the population will again adapt and become richer as a consequence. If the change is a mixture of good and bad, the chances are that the adaptive processes will ensure a net improvement in wealth. This for a population which, if history is any guide, and for reasons entirely independent of climate change, will probably be a lot wealthier than we are.

Perhaps the whole idea of being fair to the people of the future should be reversed. Perhaps they can easily afford to owe us something in retrospect.

The bottom line of politically correct thought on the matter—the thought that we must collectively do something drastic now to prevent climate change in the future—is so full of holes that it brings the overall sanity of mankind into question. For what it is worth, one possible theory is that mankind (or at least that fraction of it that has become both over-educated and more delicate as a result of a massive increase of its wealth in recent times) has managed to remove the beliefs of existing religions from its consideration—and now it misses them. As a replacement, it has manufactured a set of beliefs about climate change that can be used to guide and ultimately to control human behaviour. The beliefs are similar to those of the established religions in that they are more or less unprovable in any strict scientific sense.

This essay originally appeared in The Quadrant.