April 21, 2004

Conflicting “Science” in Nature

The prestigious science journal Nature concurrently runs an article promising total melting of Greenland’s ice (applying climate models to this region) and another on how it is “next to impossible” to do accurate regional modeling.

Without acknowledging the incongruity, Nature’s editors include in their April 8, 2004, edition an article that relies on climate models to forecast certain elimination of Greenland’s ice-sheet a thousand years from now and another in which climate scientists say climate models applied at the regional level are not yet able “to predict what will happen in the next 20 years.”

In his Greenland study, Jonathan Gregory forecasts the virtual certainty (we choose our words precisely) that almost all Greenland’s ice will melt because of human activity. “This would mean a global average sea-level rise of seven meters [23 feet] during the next 1,000 years or more,” he says. That would be a serious threat. The trouble is that almost every assumption Gregory uses to reach this dramatic conclusion is wrong because he and his fellow researchers apply a global climate model to the regional level. As Swedish climate modeler Colin Jones says in the same edition, “The natural variability of climate makes [regional modeling] next to impossible.”

As reported in the same issue of Nature, the consensus of scientists is that this simply doesn’t work.
Reporting on a March 29, 2004 regional climate modeling workshop in Lund, Sweden, Nature’s Quirin Schiermeier writes, “Participants admitted privately that the immediate benefits of regional climate modeling have been oversold in exercises such as the Clinton’s administration’s U.S. regional climate assessment, which sought to evaluate the impact of climate change on each part of the country.”

At about 0.4% of the globe, Greenland is by definition “regional.” By way of comparison, the forty-eight contiguous United States occupy 2% of the planet’s surface — which leads us to the real bombshell in this edition of Nature. Climate scientists at least privately now admit that the “U.S. National Assessment” of global warming was inappropriate.

(In fact, they knew this at the time. As has been documented in sworn testimony before the House Oversight and Investigations Subcommittee on July 25, 2002, both reviewing scientists and those who produced the “Assessment” knew the models were worse than a table of random numbers when applied to the U.S. temperature history).

If they can’t accurately model climate for 2% of the world (the United States), then they can’t accurately model 0.4% (Greenland). But that didn’t stop Nature from publishing Gregory’s piece in which the researchers conclude, “[T]he Greenland ice sheet is likely to be eliminated by anthropogenic climate change unless more substantial emission reductions are made than those envisioned by the [United Nations’] IPCC [Intergovernmental Panel on Climate Change].”

Contrast this assertion with Georgios Amanatidis’ statement at the Lund gathering (he is a science officer at the European Commission’s Environment and Climate Program). “We’re not yet at the promised level where regional climate models can really influence policy-making,” he said. Yet, we note, they can and do, despite what Sweden’s Jones goes on to add: “Government and business would like us to predict what will happen in the next 20 years. Unfortunately we can’t.”

Other modelers tell Schiermeier their understanding of regional climate change “will remain uncertain” and suggest “policymakers’ expectations of precise local projections need to be dampened down.”

Nature fails to heed that advice. In fact, the lead sentence in Gregory’s article is so misleading we pause to wonder — yet again — what is happening to the scientific review process and editorial oversight at what arguably is the world’s most prestigious science periodical. It read:

The Greenland ice-sheet would melt faster in a warmer climate and is likely to be eliminated — except for residual glaciers in the mountains — if the annual temperature in Greenland increases by more than about 3°C. This could raise the global average sea-level by seven meters or more over a period of 1,000 years or more.

A reader can conclude from that opening passage that raising the temperature of Greenland by 3°C is going to melt virtually the entire ice mass in a millennium. A reporter who wants to foment a little global warming hysteria need only quote that opening passage. But is the paragraph an accurate representation of the paper’s findings?

Deeper in the article the authors state, “For a warming of 3°C, the ice sheet loses mass slowly, and over millennia might approach a steady state in a smaller inland form. With greater warming, mass is lost faster and the ice-sheet is likely to melt away. The most extreme scenario considered in the third assessment report of the Intergovernmental Panel on Climate Change (IPCC) involves a warming of 8°C; in this case, most of the ice-sheet disappears over the next 1,000 years.”

It seems to us the opening paragraph is profoundly misleading. The passage we just quoted doesn’t claim a temperature rise of “more than about 3°C” will melt virtually all of Greenland in a millennium. Rather, it says, a rise of 8°C — which is the most extreme among the IPCC’s scenarios — will result in a millennial meltdown.

In fact, by their own admission, the authors’ virtual certainty results from assuming virtual reality. Gregory et al. write that the threshold for net melting is passed in thirty-four of the thirty-five emissions scenarios published by the IPCC. The threshold, they explain, was “derived assuming uniform warming throughout the year.” Only then do they add, as any climate scientist and most other scientists know (and as the IPCC repeatedly states), greenhouse warming is much greater in winter when, in the authors’ words, “no melting takes place.”
Using summer warming (they don’t say how much) “and allowing for uncertainty in the threshold” (again unquantified and undefined), the threshold for net melting (not a complete meltdown) is missed in more than 30 percent of the scenarios.

We choose our words carefully but must tell it like it is: Nature, the most prestigious science magazine in the world, is publishing science that the scientific consensus says is “oversold.”

We raised this issue in a 190-word letter to Nature (reproduced below) on April 12, 2004. We have been informed in a telephone conversation that Nature has chosen not to run our letter, and will be sending us a formal letter describing the reasons why. Stay tuned to these pages as we will respond in detail to Nature’s rejection letter, for there can be no denying that the use of global models scaled to the regional level, which was criticized at the Lund meeting, is precisely the methodology used in the Greenland paper.

April 12, 2004 Letter to Nature:

Sir-The April 8 issue of Nature contains two contradictory articles. One, by Schiermeier (1) summarizes a conference on regional climate modeling held in Lund, Sweden. It stated that “participants admitted privately that the immediate benefits of regional climate modeling have been oversold in such exercises as the Clinton Administration’s US regional climate assessment” and that “policy-makers’ expectations of precise local regional projections need to be dampened down.”

Later in the same issue, Gregory et al. (2) “conclude that the Greenland ice sheet is likely to be eliminated by anthropogenic climate change unless more substantial emission reductions are made than those envisioned by the IPCC.” This projection was made with a regional model for climate-induced changes in ice balance.

The US assessment applied to 2% of the world’s surface area. Greenland occupies 0.4%. Models whose value is “oversold” over the United States are surely worth even less over Greenland.

The simultaneous appearance of these two articles is troubling, as one represents the consensus of a community of scientists, while the other was peer-reviewed by members of that same community.

Patrick J. Michaels
Department of Environmental Sciences
PO Box 400123
University of Virginia
Charlottesville, VA 22904-4123

No Comments

No comments yet.

RSS feed for comments on this post. TrackBack URI

Sorry, the comment form is closed at this time.

Powered by WordPress