4. A STRATEGY FOR UNCERTAINTY

For all the discussion of final concentration levels of carbon dioxide, we have no firm idea of what constitutes a "safe" level. Scientists have shied away from quantifying safe levels because of inherent uncertainties and because of reluctance to make broad value judgments, a task more suited to politics than to science (Azar and Rodhe, 1997). The unfortunate consequence is that arbitrary targets have emerged unchecked to fill this vacuum.

The central IPCC projection suggests that concentrations of carbon dioxide under a business-as-usual scenario will be over 700 ppmv in 2100, compared with the preindustrial level of 280 ppmv. A target concentration of 550 ppmv has gained prominence in the literature for the simple reason that it represents an approximate doubling of preindustrial carbon concentrations.⁸ Worse, because most published ranges use 550 ppmv as a mid-point prediction, it appears as a reasonable compromise. Yet, a concentration of 550 ppmv is estimated to lead to a rise in global average temperature of between 2C° and 5C° by 2150. By contrast, the natural fluctuation over the last 10,000 years has been only 1C° (Azar and Rodhe, 1997; Gleick and Sassin, 1990).

8 When the growth of other greenhouse gases is taken into account, some of which are far more potent at trapping heat, anything over 500 ppmv is equivalent to a doubling of CO₂ alone.

Despite the uncertainty regarding targets, many economic analyses of delay attempt to plot cost-effective paths to reach targets that are regarded as safe from the outset and remain fixed. They aim to lay out the best way of getting from here to there as if we knew with certainty where there was.⁹

9 Even in studies which consider alternative final concentrations, the implication is often that the choice of final target is resolved at the outset, not that we proceed towards one target at the start only to find that we have to change targets later as better information becomes available.

Of course, we cannot be so certain about final desirable concentrations. Some have argued that we should not act until we know more. In their eyes, it would be wrong to impose immediate economic change to reach a low target if further research reveals that higher concentrations and temperatures are perfectly safe. Clearly the uncertainty may work in the other direction as well: if we set 550 ppmv as our target now, we may later discover that a much lower target is required. We would then be forced to make further cuts in emissions than we had predicted. These cuts would be even more severe if we do nothing in the interim.

Even if the probabilities of getting it wrong were equal in both directions, the expected costs of excessive abatement at one extreme may be considerably less than the costs from climate change-induced catastrophes at the other extreme. Economic costs should have a lower bound than climate change costs given the immediate political pressure that keeps economic costs in check. Costs from severe climate-related damage would be largely beyond our control and may be difficult to avoid if they appear, given the time-lags in the system. Economic costs on the same scale could only occur if we set out immediately on a path far more stringent than any that have been suggested and adhere to it in the face of high costs - something that would fail to pass muster politically (Grubb, 1997).

In a framework of uncertainty, a sensible strategy is to adopt a position from which it is possible to meet different contingencies. Taking early incremental action would allow us to meet more stringent cuts if necessary without compromising our ability to revert quickly to today's emission paths if they are confirmed as safe. Alternatively, failure to act now may foreclose our ability to meet low stabilization targets without severe economic disruption, if we later realize that they are necessary. Models which incorporate uncertainty tend to advocate more near-term action as a type of hedging strategy (Grubb, 1997; Lempert et al, 1997; Manne, 1996).