National Assessment on Climate Change makes a poor first impression

National Assessment on Climate Change makes a poor first impression
September 1, 2000

After much fanfare, the U.S. National Assessment (USNA) on global warming—officially titled Climate Change Impacts on the United States: The Potential Consequences of Climate Variability and Change—was released for a “60-day public comment period” June 12 through August 11. Within days, however, the verdict was in.

Almost immediately, officials from the Environmental Protection Agency said the report had little in common with reality. EPA's Mike Slimak and Joel Scheraga said portions of the document had "a rather extreme/alarmist tone and do not appear to fairly reflect the scientific literature, the historical record, or the output of extant models." Hardly a rousing endorsement of the Administration's pet pony, and this from the pony's own groom!

Science luminaries as bright as satellite guru John Christy have deemed the USNA “an evangelistic statement about a coming apocalypse [and] not a scientific statement about the evolution of a complicated system with significant uncertainties.”

Not scientific? That’s pretty harsh language when used by an acclaimed scientist to describe a document whose authors claim it was subject to extensive scientific peer review. But Christy is exactly and precisely correct.

In science, we propose hypotheses and then test them against observations. If the data indicate the initial hypothesis cannot be supported, all calculations contingent upon that hypothesis are no longer supported either.

In the USNA, substitute “climate model” for “hypothesis,” because, scientifically speaking, that’s all a “model” is: a presumably consistent hypothesis about the way a system works. The “horror lines” (more words from Christy) in the USNA are largely generated by one of only two climate models that were used, a model produced by the Canadian Climate Centre (CCC).

The CCC is a global climate model, but the USNA applies it only to the United States, a relatively small (3 percent) region of that globe. But global climate models are generally not good enough to resolve the peculiarly violent American climate, and the CCC simply doesn’t work here in the United States.

The USNA’s plot of twentieth century temperature departures from the long-term average over the lower 48 states does not use annual values; instead, it uses a 10-year average centered upon the target year, which gives the CCC a fighting chance of accuracy by smoothing over the year-to-year variability caused by El Niños, volcanoes, and shifting jet streams.

You can see two equivalent warmings in this smoothed data, both of the same size. One ended 70 years ago, before greenhouse warming could have caused it, and the other ended yesterday. The overall variation is about ± 0.5°F.

The USNA data also show the differences between the CCC smoothed predictions and the observed 10-year smoothed temperatures (i.e., the “errors” in the model). The difference between the temperatures the CCC predicted and what has been observed are greater than the natural variability of climate!

That’s prima facie evidence that the CCC does not work when applied to the United States. It means the average temperature is a better predictor of the twentieth century’s climate than what’s generated by Canada’s complex, multimillion-dollar model. And if the CCC can’t tell us what’s happened in the known past, what possible good is it for predicting an unknown and unknowable future?

That stunning finding—that the model used to give the “horror lines” in the National Assessment is, in fact, worse than no model at all—is something the peer review process should have red-flagged, assuming the 400 USNA scientists had one another look at the document before releasing it for public comment. How could everyone have missed this glaring error? (As they say in math texts, “the proof is left to the student.”)

Perhaps worse, the CCC is an extreme outlier among climate models. The latest version of the upcoming United Nations “Third Assessment Report” on climate change shows how a large number of climate models predict different rates of warming in the next 100 years—depending upon the internal dynamics of the calculations and assumptions made about emissions of carbon dioxide, other greenhouse gases, and compensatory cooling.

All the models are pretty much linear—which means that once they start their human-induced warming, it proceeds at a constant rate. The simplest reason for that is that the temperature response to a given greenhouse gas is logarithmic (warming damps out as concentrations increase), while the growth in the main greenhouse gas, carbon dioxide, is exponential (rising at an increasing rate over time). The linking of a damped warming to an exponential cause yields a straight line.

Of the 26-odd global climate models we are aware of, only one—the CCC—produces a substantially exponential warming. Yet it was one of the two climate models chosen as a basis for the USNA.

The other model the USNA used, from the United Kingdom’s Hadley Centre, is more “midrange.” In other words, the USNA offers two forecasts: one that is pretty much the average of all climate models (the U.K. model), and the other which is extremely hot (the CCC). The folks who wrote the USNA know full well the press will zero in on the hot one.

Notably, there is no equivalently cold version. But such a model exists, right here in the good ol’ U.S. of A., at the taxpayer-supported U.S. National Center for Atmospheric Research (NCAR) in Boulder, Colorado. When fed the known increases in greenhouse gases for the last 15 years (which are lower than scientists had assumed 15 years ago) it produces barely 1°C of warming for this century, compared with 3°C for the Hadley model and 8°C for the CCC.

That the USNA chose a high and a midrange scenario and ignored a low one is incontrovertible evidence of a biased attempt to produce a biased document. And that the high scenario simply doesn’t work is testimony to Christy’s assertion that the USNA isn’t scientific at all.

According to Nature magazine, University of Virginia environmental sciences professor Patrick J. Michaels is probably the nation’s most popular lecturer on the subject of climate change. Michaels is the author of Sound and Fury: The Science and Politics of Global Warming.


National Assessment Synthesis Team, 2000. Climate Change Impacts on the United States: The Potential Consequences of Climate Variability and Change, draft report. United Nations Intergovernmental Panel on Climate Change, 2000. Third Assessment Report, in press.