Professors Michaels and Singer Call on Science Editors to Withdraw Kerr-McElroy Paper and Issue a Correction

CHARLOTTESVILLE, VIRGINIA, JANUARY 11, 1994---In an unprece- dented challenge to a published scientific research paper, Profes- sors Patrick J. Michaels and S. Fred Singer and research associate Paul C. Knappenberger, all of the University of Virginia's Department of Environmental Sciences, today issued a Technical Comment to the editors of the journal Science, calling the conclusions of the just-published paper by James Kerr and Thomas McElroy of the Canadian Atmospheric Environment Service "so flawed as to require a formal withdrawal from the scientific literature."

The Kerr-McElroy paper, "Evidence for Large Upward Trends of Ultraviolet-B Radiation Linked to Ozone Depletion," received extensive newspaper and television coverage as the result of a press release issued by Science prior to the paper's appearance in the November 12 issue. Focus was put on the authors' claim of detecting as much as a 35 percent per year increase in surface ultraviolet radiation over Toronto between 1989 and 1993, and their contention that the large upward trends of surface UV were due to the depletion of stratospheric ozone over populated regions of North America.

Both statements are incorrect, according to Prof. Michaels, who also directs the Virginia State Office of Climatology at UVA, and Prof. Singer, who also directs The Science & Environmental Policy Project in Washington, D.C. Specifically, the UVA scien- tists challenged the Canadian study on the following points:

• The results quoted did not show error bars (i.e. margins of error), which are invariably attached to every scientific result. When the statistical analysis was repeated at the Virginia State Office of Climatology, using the Canadian data, the margins of error were found to be so large as to cancel out most, if not all, of the claimed increase in surface ultraviolet radiation.

• An examination of the published paper revealed that a significant number of data points were mislabeled and therefore overlooked by Kerr and McElroy in their analysis. The inclusion of these points reveals that in four of the five years studied there was no increase in surface UV whatsoever.

• The alleged upward trend in surface UV was based entirely on just four high readings--out of more than 300--all of which occurred in late March of 1993, a period when ozone over eastern North America was strongly affected by the largest winter storm of the century. Eliminate these four readings and the "trend" is essentially zero.

  "The absence of error bars should have immediately raised an alarm to the editors at Science," said Prof. Singer. "The fact that the paper was published casts doubts on the adequacy of the peer review process in this case."

  Prof. Michaels noted that F. Sherwood Rowland, a professor at the University of California at Irvine and coauthor of the CFC- global ozone depletion theory, has gone on record in support of the Kerr-McElroy paper. Quoted in the Los Angeles Times on November 12, Rowland said, "Now we have good data to point to."

  That comment, said Prof. Michaels, raises serious questions about possible internal pressure to rush the study into publica- tion. Michaels pointed out that Rowland is the most recent past president of the American Association for the Advancement of Science (AAAS), the professional organization that publishes the journal Science. Both Michaels and Singer noted that firm evidence for long-term global ozone depletion due to CFCs is still lacking.

  The UVA scientists contend that since the ozone content in the stratosphere did diminish in 1992 and 1993 (most likely as a result of the Mt. Pinatubo eruption in 1991), the absence of an expected UV trend shows that a depletion of global ozone does not necessarily enhance the UV intensity at the earth's surface.

  This lack of correlation, in their view, confirms the well- known work of Joseph Scotto and colleagues at the National Cancer Institute, who measured surface UV at eight U.S. locations between 1974 and 1985. Scotto detected no long-term increase in surface UV.

  ---

  Copies of the Technical Comment sent to the editors of Science can be obtained by contacting the Virginia Office of Climatology, University of Virginia (add address, phone, and fax)

  ---

  January 7, 1994

  To: Colleagues
  From: FRED SINGER

  THE WIDELY REPORTED UV TREND IS A HOAX

  Pat Michaels and I are sending a Technical Comment to Science . The attached press release draft tells the story. We call for a formal retraction of the conclusions of the Kerr-McElroy paper published in the Nov. 12, 1993 issue.

  Using simple regression analysis, K-Mc essentially drew a straight line through their data points, but gave no error bars, confidence limits, or any other indication of the statistical robustness of their results. By just "eye-balling" their published data points we could see that any trend would be zero, if the 1993 data points, at the end of their record, were omitted. Specifically, by removing the quite unusual "winter" points of March 25 to 29, 1993, the UV-B trend (at 300 nanometers) would become zero instead of their claimed--and widely quoted--35% per year!

  After repeating the analysis, using their actual data, we found, first of all, that their claimed trends had hardly any statistical significance, especially when we added the data points from two winter seasons that the authors for some reason had mislabeled and did not use in their statistical analysis.

  Others had noticed some of the same shortcomings of the K-Mc paper. But it was widely praised by environmental activists when it appeared. And, thanks to a press release issued by Science that stressed the significance of the paper, it produced favorable stories in major newspapers and on CBS News.

  For many years now, activists had been claiming the existence of global depletion of stratospheric ozone due to CFCs, usually in press releases stating that the depletion was "worse than expect- ed." [Expected from what? An inadequate theory?] Their big disappointment had been that there was no corresponding upward trend reported for UV-B. Hence their great joy and their eagerness to adopt the K-Mc paper as the "smoking gun" that finally validated not only a UV-B trend but also ozone depletion.

  F. Sherwood Rowland, a professor at the University of California at Irvine and coauthor of the CFC-global ozone depletion theory, has gone on record in support of the Kerr-McElroy paper. Quoted in the Los Angeles Times on November 12, Rowland said, "Now we have good data to point to."

  The upshot of this episode will likely be a black eye for hyped science and for environmental extremists, and a victory for basing environmental policies on scientific results that can withstand scrutiny.