that those who claimed this over the last decade indeed were correct. Mitchell et al. then ran two types of climate models: one ("A") in which only the greenhouse effect was changed (similar to all transient models that based the 1992 Framework Convention), and another ("B") in which a compensating effect of anthropogenerated aerosols was included.

Even though model "A" produced a warming of 2.5°C for doubled carbon dioxide, a relatively low figure, Mitchell et al. stated that it was "significantly different from the observed climate of the last quarter century"-in other words, it was predicting too much warming. They did find that model "B", which appears to calculate only 1.7°C of warming, was consistent with the climate of the last forty to fifty years.

Figure 2. Comparison of land-based and modelled temperature of the Northern Hemisphere

in the model of Manabe et al., (1991). This type of model was prominently featured in the 1992 Supplementary Report of the U.N. IPCC, which was produced specifically for meeting at which Framework Convention on Climate Change was adopted.

In the most cited ground-based temperature history, that of the U.S. Department of Energy, there is a warming trend of 0.4°C in the same period. Thus model "B" is only projecting an additional warming of 1.3°C (1.7-0.4°) for doubled CO2, which is exactly the value estimated by the "minority" scientists over the last decade.

Nonetheless, I have reason to suspect that this model, too, is not correct. I published two recent papers (Michaels et al, 1994, 1995) that specifically tested the hypothesis that the non sulfate greenhouse effect models would perform best where sulfates would not compromise their forecast (such as the polar regions, particularly Antarctica), and perform worst where the sulfate effects were greatest. If these hypotheses were supported, then it seemed plausible that sulfates were a sufficient explanation for the planet's failure to warm a la the models that based the 1992 Framework Convention update.

The opposite was true. Where there were no sulfates to harm the model, the forecasts were worst. Where the sulfate density was highest (and therefore should have compromised model behavior), the model performed best.

Because model "B", the sulfate + greenhouse model, is so heavily cited in the 1995 IPCC report, it requires close examination in any peer review of that report. From illustrations presented in the draft for reviewers, it became apparent to me that this model might still be reproducing the major error of the circa-1992 models, namely, that it was predicting that too much warming of the high latitudes of the Northern Hemisphere had already taken place. To determine whether or not this is the case required an analysis of the behavior of the high latitude grid cells.

I requested this data for my review of the 1995 IPCC document in an e-mail to the United Kingdom Meteorological Office (UKMO) on May 10, 1995. The next day John Mitchell, one of the lead authors of the IPCC draft, replied that he "felt it inappropriate to send [me] gridpoint data at this time." I replied, on the same day, "I do not understand your statement that it is inappropriate...Science is a cooperative effort in which information should be freely shared...It is, in my mind, not proper to withhold scientific information to a colleague who has been asked by the IPCC itself to review its own work, when that information is critical to the review". The data were not forthcoming.

I then telephoned Dr. Michael MacCracken, head of the U.S. Global Change Research Program, and also head of U.S. review team of the IPCC, stating that I was being refused data absolutely critical to a proper peer review of this important document. He stated that he would be contacting IPCC, but the important data never appeared.

I specifically requested the data required to review this model in other correspondence. making five separate inquiries to IPCC.

There is no doubt their refusal to supply me the model results seriously compromises the scientific review process of climate models that now form the basis for policy proposals to

reduce greenhouse emissions.

I have appended my review of the new IPCC document (dated June 5, 1995) as well as that of the IPCC "Synthesis" of recent findings (September 14, 1995) for this Record. Throughout these reviews I have noted that I was denied critical results required to perform an adequate review. Several of my review comments, especially on Chapter 6 in the June review, speak directly to problems in the climate models. My comments consist of 4,639 words and resulted in not one discernable change in the text of the IPCC drafts.

What does all of this say about repeated reports that there will be grave ecological consequences from global warming, large sea level rise, and spread of tropical diseases, and other apocalyptic scenarios?

Simply this: Such reports cannot be trusted. Many, including the recent EPA report on sea level rise, are based upon models that are now known to have been greatly overpredicting warming. And prediction of effects that are based upon the newer, cooler, models, such as the ones described here, are simply premature, because data required to analyze those models has not been provided to those who are known to be critical reviewers.

REFERENCES

Intergovernmental Panel on Climate Change (IPCC), 1992. Climate Change 1992: The Supplementary Report to the IPCC Scientific Assessment. J.T. Houghton, Senior Editor. 198 pp.

Manabe S., et al., 1991. J. Climate 4, 785-818.

Michaels, P. J. et al., 1994. Technology: J. Franklin Inst. 331A, 123-133.

1995. Atmospheric Research 37, 257-266.

Mitchell, J.F.B., et al., 1995. Nature, 376, 501-504.

Mitchell, J.F.B. et al., 1995. J. Climate, 8, 2364-2385.

Patel 1. Manders

Nov 14 1.995

TO: Mike MacCracken, U.S. Review Coordinator, IPCC 1995 Summary.

FROM: Patrick J. Michaels, University of Virginia

Following are my reviews of four sections of the IPCC 1995 draft, Chapters 3, 6, and the Policymakers Summary.

I could not provide a thorough review of Chapter 6 without the gridded transient output of sulfate/greenhouse models, particularly the one from UKMO. You know the story there.