The Intergovernmental Panel on Climate Change (IPCC)

Second Assessment Report, Working Group I, Chapter Eight

NOTE: The original Working Group I report was approved by the IPCC in December, 1995. Subsequent to that approval, IPCC has apparently allowed additional edits to the document. Some changes are editorial, serving to add clarification or to correct sentence structure. However, other changes appear to go beyond that and have the effect of changing the substance and tone of this chapter. The most significant edits are identified below. New material is italicized, deleted material has a strike through it.

Summary

" Many but not all The Majority of these studies show that the observed changes in global-mean, annually-averaged temperature over the last century is unlikely to be due entirely to natural fluctuations of the climate system."

deleted:

"The evidence rests heavily on the reliability of the (still uncertain) estimates of natural variability noise levels."

new:

"Furthermore, the probability is very low that these correspondences could occur by chance as a result of natural internal variability. The vertical patterns of change are also inconsistent with the response patterns expected for solar and volcanic forcing."

"Viewed as a whole, these results indicate that the observed trend in global warming mean temperature over the past 100 years is larger than our current best estimates of natural climate variations over the last 600 years. unlikely to be entirely natural in origin."

Section 8.1

"The attribution of a detected climate change to a particular causal mechanism can be established only by testing involves tests of competing hypotheses."

"The claimed statistical detection of an anthropogenic signal in the observations must always be accompanied by the caveat that other explanations for the detected climate-change signal cannot be ruled out completely, unless a rigorous attempt has been made to do so."

new: "There is, however, an important distinction between achieving 'practically meaningful' and 'statistically unambiguous' attribution. This distinction rests on the fact that scientists and policymakers have different perceptions of risk. While a scientist might require decades in order to reduce the risk of making an erroneous decision on climate change attribution to an acceptably low level (say 1-5%), a policymaker must often make decisions without the benefit of waiting decades for near-statistical certainty."

Section 8.1.3

"We now have: * more relevant model simulations, both for the definition of an anthropogenic climate change signal and for the estimation of natural internal variability. * more relevant simulations for the estimation of natural internal variability, and initial estimates from paleoclimatic data of total natural variability on global or hemispheric scales; * more powerful statistical methods for detection of anthropogenic change, and a better understanding of simpler statistical methods and increased application of pattern-based studies with greater relevance for attribution."

Section 8.2.2 Inadequate Representation of Feedbacks

new: "Deficiencies in the treatment and incorporation of feedbacks are a source of signal uncertainty."

Section 8.2.5

"Current pattern-based detection work has not attempted is now beginning to account for these forcing uncertainties."

Section 8.3.2

"Initial attempts are now being made For these reasons and many others, scientists have been unable to use paleoclimate data in order to reconstruct a satisfactory, spatially-comprehensive picture of climate variability over even the last 1,000 years. Nevertheless, The process of quality-controlling paleoclimatic data, integrating information from different proxies, and improving spatial coverage should be encouraged. Without a Better paleoclimatic data bases for at least the past millennium, it will be difficult are essential to rule out natural variability as an explanation for recent observed changes, or and to validate coupled model noise estimates on century time scales (Barnett et al., 1995)."

Section 8.3.3.3

deleted: "While such studies help to build confidence in the reliability of the model variability on interannual to decadal time scales, there are still serious concerns about the longer time scale variability, which is more difficult to validate (Barnett et al., 1995). Unless paleoclimatic data can help us to 'constrain' the century time scale natural variability estimates obtained from CGCMs, it will be difficult to make a convincing case for the detection and attribution of an anthropogenic climate change signal."

Section 8.4.1

deleted : "While none of these studies has specifically considered the attribution issue, they often draw some attribution-related conclusions, for which there is little justification."

Section 8.4.1.1

"The conclusion that can be drawn from this body of work, and earlier studies reported in Wigley and Barnett (1990) is that the warming trend to date is unlikely to have occurred by chance due to internally-generated variability of the climate system, although this explanation cannot be ruled out. This, however, does not preclude the possibility that a significant part of the trend is due to natural forcing factors. Implicit in such studies is a weak attribution statement--i.e., some (unknown) fraction of the observed trend is being attributed to human influences. Any such attribution-related conclusions, however, rest heavily on the reliability of our estimates of both century time-scale natural variability and the magnitude of the observed global warming mean trend. At best, therefore, trend significance can only provide provides circumstantial support for the existence of an anthropogenic component to climate change, but does not directly address the attribution issue."

Section 8.4.1.3

"These empirical estimates of In summary, such studies offer support of a DT2x are subject to considerable uncertainty, as shown in a number of studies (see, e.g., Wigley and Barnett, 1990; Wigley and Raper, 1991b; Kheshgi and White; 1993b). In summary, such studies offer support for a DT2x value similar to that obtained by GCMs, and suggest that human activities have had a measurable impact on global climate, but they cannot help to establish a unique link between anthropogenic forcing changes and climate change."

Section 8.4.2.1

new: "Implicit in these global mean results is a weak attribution statement--if the observed global mean changes over the last 20 to 50 years cannot be fully explained by natural climate variability, some (unknown) fraction of the changes must be due to human influences".

deleted: "None of the studies cited above has shown clear evidence that we can attribute the observed changes to the specific cause of increases in greenhouse gases."

Section 8.4.2.3.

new: "To date, pattern-based studies have not been able to quantify the magnitude of a greenhouse gas or aerosol effect on climate. Our current inability to estimate reliably the fraction of the observed temperature changes that are due to human effects does not mean that this fraction is negligible. The very fact that pattern-based studies have been able to discern sub-global-scale features of a combined CO2 + aerosol signal relative to the ambient noise of natural internal variability implies that there may be a non-negligible human effect on global climate."

Section 8.5.2

new: "Simultaneous model-observed agreement in terms of changes in both global means and patterns, as in the recent study by Mitchell et al. (1995a), is even less likely to be a chance occurrence or the result of compensating model errors."

Section 8.6

"Finally we come to the most difficult question of all: 'When will the detection and unambiguous attribution of human-induced climate change occur ?' when the detection and attribution of human-induced climate change is likely to occur. The answer to this question must be subjective, particularly in the light of the very large signal and noise uncertainties discussed in this Chapter, it is not surprising that the best answer to this question is 'We do not know'. Some scientists maintain that these uncertainties currently preclude any answer to the question posed above. Other scientists would and have claimed, on the basis of the statistical results presented in Section 8.4, that confident detection of a significant anthropogenic climate change has already occurred. would and have claimed, on the basis of the results presented in Section 8.4, that detection of a significant climate change has already occurred. As noted in Section 8.1, attribution involves statistical testing of alternative explanations for a detected observed change and Few if any would be willing to argue that completely unambiguous attribution of (all or part of) this change to anthropogenic effects has already occurred, or was likely to happen in the next several years."

new: "However, evidence from the patterned-based studies reported on here suggests that an initial step has now been taken in the direction of attribution, since correspondences between observations and model predictions in response to combined changes in greenhouse gases and anthropogenic sulphate aerosols:

• have now been seen both at the surface and in the vertical structure of the atmosphere;

• have been found in terms of complex spatial patterns rather than changes in the global mean alone;

• show an overall increase over the last 20 to 50 years;

• are significantly different from out best model-based estimates of the correspondence expected due to natural internal climatic variability.

The body of statistical evidence in Chapter 8, when examined in the context of our physical understanding of the climate system, now points toward a discernible human influence on global climate. Our ability to quantify the magnitude of this effect is currently limited by uncertainties in key factors, including the magnitude and pattern of longer-term natural variability and the time-evolving patterns of forcing by (and response to) greenhouse gases and aerosols."

Section 8.7