A recent paper in the Journal of Climate by J.F.B. Mitchell shows that the models of that type are off by at least 1.3 and as much as 2.3 degrees Celsius today.

And I would like to show you a couple of slides that demonstrate this problem, if I could. I will need to borrow your microphone. [Pause.]

In 1979, we put up a series of satellites that measure the temperature with a great deal of accuracy, and the bottom line there, the one that does not change, is the temperature history for the southern half of the planet which, by the way, correlates at 97 percent with the mean temperature between 5,000 feet and 30,000 feet in the atmosphere. The satellite finds no warming in that layer.

The top graph is the warming predicted by the model that was most cited in the United Nations' document specifically designed to produce the backing for the Rio Climate Treaty.

And if we go to the northern hemisphere, the situation gets even worse because the northern hemisphere of course is predicted to warm more rapidly than the southern hemisphere because it has more land in it.

Now, if we run that model backwards-and this appeared in the refereed literature at the time that the treaty was being ratifiedyou can see the open circles, which are the observed temperatures of the northern hemisphere, and the closed circles are the modeled temperatures by that particular model.

Ladies and gentlemen, this is a large and propagating error that I believe should have been known to this Congress at the time of the 1992 Framework Convention, but it was not.

Anyway, these models also produce very, very large polar warming. This is the warming projected in the high latitude by that model, and here is the warming of the polar region that has been observed.

You can see in fact there is a warming. The only problem is it occurs before 1940 which is before the greenhouse effect changed very much. Since then, between then and now, there is very little. Well, the good news, and I will finish up with this, is that the so-called skeptics turn out to have been right.

Here is the first sentence from a paper by J.F.B. Mitchell that appeared August 10th in Nature Magazine. "Climate models suggest that increases in greenhouse gas concentrations should have produced a larger mean warming than has been observed."

That is a very polite way of saying that those people who were derided as a small minority in fact were correct. The models were too warm.

And Mitchell's paper is a very interesting one. I will finish on what it says because there is an interesting flip on the end of this discussion.

Mitchell's model incorporates the effect of sulfate aerosol along with greenhouse gases in the atmosphere. When he does not do that, his model is too warm, even though it only predicts a net warming of doubling carbon dioxide of 2.5 degrees.

You can take Mitchell's numbers and you will see that the net warming that he is predicting for doubling carbon dioxide is 1.7 degrees. And he states that his model correctly tracked the last 40

to 50 years, in which the surface records showed a warming of .4 degrees.

How much warming does that leave out to the end of the next century? 1.7 - 1.4 degrees or 1.3 degrees?

That is precisely the number that the so-called skeptics, whenever they were fortunate enough to be brought here in front of this Congress, had been saying for the last seven years-1 to 1.5 degrees.

Finally, I would just like to close and show you a little bit of this model.

The top graph is the one; the model without the sulfates in it for the year 2040 to 2050. You can see the consistent high warming of the high latitudes of our hemisphere.

And the middle graph is a sulfate plus greenhouse model. The middle graph clearly shows large warming in the high latitudes also.

And it became apparent to me, when I saw this chart-and we can turn the lights back on-when I saw this chart, which was in the United Nation's 1995 update document on climate change, that this model was making the same error that the other ones were making that it was producing too much warming in the higher latitudes.

So I requested from the United Nations that they send me the data that went into that model, and I was denied.

I wrote back, and I said this is a horrible breach of scientific ethic. You must send me the data because I have been asked by the United Nations to review their own work. I was denied.

I informed the Director of the U.S. Global Change Research Program. He said he would call them. The data never showed up.

There were six separate requests to provide a proper review of this document. They were all denied.

Therefore, any policy statements or any impact statements that are made on these new, more reliable models are based upon models that were not subject to review by those who were known to provide critical review in the process.

Thank you.

[The prepared statement and attachments of Dr. Michaels follow:]

23-558 96-2

Testimony of Patrick J. Michaels, Department of Environmental Sciences, University of Virginia to the Subcommittee on Energy and the Environment, U. S. House of Representatives, November 16, 1995. This testimony represents the no offfical position of the University or Commonwealth of Virginia.

My testimony centers around four issues of critical importance to the problem of global warming:

1. New calculations support the view of the scientists who predicted that global warming would be relatively modest.

2. Older calculations that based the 1992 Framework Convention on Climate Change were known to be greatly overestimating warming at the time the Convention was ratified.

3.

Critical scientists are still being denied data (by taxpayer-supported organizations) that are required to quantitatively review new syntheses on global warming.

4. Therefore, any studies of the impact of climate change on ecosystems, health, and the economy, based upon older models, are in error, and newer models have yet to be properly, reviewed.

Controversy surrounding the issue of global warming caused by an enhancing greenhouse effect is a classic example of the normal and creative scientific tension that exists between those who formulate hypotheses (i.e. "models") and those who evaluate such hypotheses with observed data. If this issue had not been politicized before the hypotheses and observations could naturally converge, global warming would just be another interesting example of applied physical science.

Unfortunately, that is not the case. For the last decade, a community of scientists, often referred to as a "small minority", had argued, based on the observed data on climate change, that the modelled warming was far too large, and therefore that any intrusive policy would not be based upon reliable models of global warming. This view, generated only by data, has been cast in a very negative political light, which has had a chilling effect on scientific free speech. At the same time, testimony has been repeatedly given before the

U.S. Congress that the observed and modelled temperatures are "broadly consistent". In terms of federal funding, this view has been amply rewarded. But, the two views have never been reconciled scientifically. In fact, while models suggest a warming of roughly 4°C for doubling atmospheric carbon dioxide, the data suggested a much lower number, or approximately 1.0-1.5°C of additional warming.

The most important development in the last two years is that it is now acknowledged that the community that argued for the lower numbers appears more likely to be correct. Moreover, it is apparent that the climate model most heavily cited by the United Nations in its 1992 "Supplementary Report" on Climate Change, prepared specifically to provide technical backing for the Framework Convention on Climate Change, was known to be making large errors in its forecast of present temperature at the time of the adoption of the Framework Convention, and yet this was never entered into the debate concerning the Convention by the community that developed the models.

This observation argues strongly that the scientific review process that bases such international agreements has been flawed. Alternatively, there may have been omissions in communicating to responsible individuals how large the errors in these calculations were.

In a recent paper in the Journal of Climate, J.F.B Mitchell et al. examined a climate model that was very similar to what was heavily cited by the United Nations in 1992. They found that such models predict that the atmosphere should have already warmed between 1.3 and 2.3°C as a result of changes in greenhouse gases; the observed warming is 0.5°C.

As a graphic example of this, note the difference between the greenhouse warming model of Manabe et al. (1991), which was heavily cited by the U.N. in 1992, and observed temperatures as sensed by satellite (Figure 1). The only assumptions used in calculating this difference are that the climate model is correctly estimating the temperature before the greenhouse effect had changed, and that the satellite is correctly measuring the temperature of the lower atmosphere. Comparision of balloon-measured temperatures with the satellite data ensures that this is the case.

Figure two shows an additional comparison between the same model and ground-measured temperatures over the last hundred years in the Northern Hemisphere. An earlier version of this plot appeared in Bulletin of the American Meteorological Society (Michaels and Stooksbury, 1992), prior to ratification of the Framework Convention.

The 1995 "Second Scientific Assessment" of the U.N. Intergovernmental Panel on Climate Change (IPCC) now states that the climate models that supported the 1992 Supplementary Report were in fact predicting too much warming. Rather, it states that the model that best tracks the past climate is one which includes the compensating effects of sulfate aerosols. This model was published by Mitchell et al. in Nature on August 10.

Mitchell et al. state at the outset that models similar to those that based the Framework Convention "have produced a larger mean warming than has been observed", an admission

-.4

-.6 -.8.

1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995

Figure 1. Satellite-sensed lower tropospheric temperature vs. surface temperature in the coupled ocean-atmosphere transient model of Manabe et al. (1991), for the period January. 1979 (When the satellite record begins), through October, 1995. Top: Northern Hemisphere; Bottom: Southern Hemisphere.