ence is that the sensitivity of the climate Stephen Schneider to carbon dioxide will turn out to be at the low end of the IPCC uncertainty range—which is for a warming of 1.5 to 4.5 degrees C if carbon dioxide were to GLOBAL WARMING: double and be held fixed over time. Emissions scenarios, according to NEGLECTING THE COMPLEXITIES the IPCC, fall into six “equally sound” alternative paths. These paths span a or three decades, I have Measuring the Real State of the World. doubling in carbon dioxide concentra- been debating alternative A “skeptical environmentalist” is cer- tions in 2100 up to more than tripling solutions for sustainable tainly the best kind, I mused, because and well beyond tripling in the 22nd development with thou- uncertainties are so endemic in these century. Lomborg, however, dismisses sands of fellow scientists complex problems that suffer from miss- all but the lowest of the scenarios: and policy analysts—ex- ing data, incomplete theory and nonlin- “Temperatures will increase much less F changes carried out in ear interactions. But the “real state of the than the maximum estimates from myriad articles and formal meetings. world”—that is a high bar to set, given IPCC—it is likely that the temperature Despite all that, I readily confess a lin- the large range of plausible outcomes. will be at or below the B1 estimate [the gering frustration: uncertainties so infuse And who is Lomborg, I wondered, lowest emissions scenario] (less than 2° the issue of climate change that it is still and why haven’t I come across him at C in 2100) and the temperature will cer- impossible to rule out either mild or cat- any of the meetings where the usual sus- tainly not increase even further into the astrophic outcomes, let alone provide pects debate costs, benefits, extinction twenty-second century.” confident probabilities for all the claims rates, carrying capacity or cloud feed- Cost-benefit calculations show that and counterclaims made about environ- back? I couldn’t recall reading any sci- although the benefits of avoiding cli- mental problems. entific or policy contributions from him mate change could be substantial ($5 Even the most credible international either. But there was this massive 515- trillion is the single figure Lomborg assessment body, the Intergovernmental page tome with a whopping 2,930 end- cites), this is not worth the cost to the Panel on Climate Change (IPCC), has notes to wade through. On page xx of economy of trying to constrain fossil- refused to attempt subjective probabilis- his preface, Lomborg admits, “I am not fuel emissions (a $3-trillion to $33-tril- tic estimates of future temperatures. myself an expert as regards environmen- lion range he pulls from the economics This has forced politicians to make their tal problems”—truer words are not literature). Asymmetrically, no range is own guesses about the likelihood of var- found in the rest of the book, as I’ll soon given for the climate damages. ious degrees of global warming. Will illustrate. I will report primarily on the The Kyoto Protocol, which caps in- temperatures in 2100 increase by 1.4 thick global warming chapter and its dustrialized countries’ output of green- degrees Celsius or by 5.8? The differ- 600-plus endnotes. That kind of dead- house gases, is too expensive. It would ence means relatively adaptable changes weight of detail alone conjures at least reduce warming in 2100 by only a few or very damaging ones. the trappings of comprehensive and tenths of a degree—“putting off the tem- Against this background of frustra- careful scholarship. So how does the re- perature increase just six years.” This tion, I began increasingly to hear that a ality of the text hold up to the pretense? number, though, is based on a straw- young Danish statistician in a political I’m sure you can already guess, but let me man policy that nobody has seriously Corbis science department, Bjørn Lomborg, give some examples to make clear what I proposed: Lomborg extrapolates the had applied his skills in statistics to bet- learned by reading. Kyoto Protocol, which is applicable ter determine how serious environmen- The climate chapter makes four ba- only up to 2012, as the world’s sole cli- tal problems are. Of course, I was anx- sic arguments: mate policy for another nine decades. ious to see this highly publicized contri- Climate science is very uncertain, Before providing specifics of why I bution—The Skeptical Environmentalist: but nonetheless the real state of the sci- believe each of these assertions is fatally RICHARD HAMILTON SMITH flawed, I should say something about ance with the IPCC, other national cli- tion would likely increase estimates of Lomborg’s methods. First, most of his mate assessments and most recent stud- climate sensitivity by a factor of several. nearly 3,000 citations are to secondary ies in the field of climate science. As a final example, he quotes a con- literature and media articles. Moreover, Now let us look in more detail at troversial hypothesis from Danish cloud even when cited, the peer-reviewed arti- the four major arguments he makes in physicists that solar magnetic events cles come elliptically from those studies this chapter. modulate cosmic rays and produce “a that support his rosy view that only the Climate science. A typical example clear connection between global low- low end of the uncertainty ranges will be of Lomborg’s method is his paraphrase level cloud cover and incoming cosmic plausible. IPCC authors, in contrast, of a secondary source in reporting a radiation.” The Danish researchers use Lomborg admits, “I am not myself an expert as regards ENVIRONMENTAL PROBLEMS”—truer words are not found in the rest of the book. were subjected to three rounds of review 1989 Hadley Center paper in the jour- this hypothesis to support an alternative by hundreds of outside experts. They nal Nature in which the researchers make to carbon dioxide for explaining recent didn’t have the luxury of reporting pri- modifications to their climate model: climate change. Lomborg fails to dis- marily from the part of the community “The programmers then improved the cuss—and I haven’t seen it treated by that agrees with their individual views. cloud parameterizations in two places, the authors of that speculative theory ei- Second, it is ironic that in a popular and the model reacted by reducing its ther—what such purported changes to book by a statistician one can’t find a temperature estimate from 5.2° C to this cloud cover have done to the radia- clear discussion of the distinction among 1.9° C.” Had this been first-rate scholar- tive balance of the earth. Increasing different types of probabilities, such as ship, Lomborg would have consulted clouds, it has been well known since pa- frequentist and Bayesian (that is, “ob- the original article, in which the conclud- pers by Syukuro Manabe and Richard jective” and “subjective”). He uses the ing sentence of the first paragraph pre- T. Wetherald in 1967 and myself in word “plausible” often, but, curiously sents the authors’ caveat: “Note that al- 1972, can warm or cool the atmosphere for a statistician, he never attaches any though the revised cloud scheme is more depending on the height of the cloud probability to what is “plausible.” The detailed it is not necessarily more accu- tops, the reflectivity of the underlying Third Assessment Report of the IPCC, rate than the less sophisticated scheme.” surface, the season and the latitude. The on the other hand, explicitly confronted In a similar vein, he cites Richard S. reason the IPCC discounts this theory is the need to quantify all confidence Lindzen’s controversial stabilizing feed- that its advocates have not demonstrat- terms. Working Group I, for example, back, or “iris effect,” as evidence that ed any radiative forcing sufficient to gave the term “likely” a 66 to 90 per- the IPCC climate sensitivity range match that of much more parsimonious cent chance of occurring. Although the should be reduced by a factor of almost theories, such as anthropogenic forcing. IPCC gives a wide range for most of its three. He fails either to understand this Emissions scenarios. Lomborg as- projections, Lomborg generally dismiss- mechanism or to tell us that it is based serts that over the next several decades es these ranges, focusing on the least se- on only a few years of data in a small new, improved solar machines and oth- rious outcomes. Not so much as one part of one ocean. Extrapolating this er renewable technologies will crowd probability is offered for the chance of a small sample of data to the entire globe fossil fuels off the market. This will be dangerous outcome, yet he makes a firm is like extrapolating the strong destabi- done so efficiently that the IPCC scenar- assertion that climate “will certainly” lizing feedback over midcontinental ios vastly overestimate the chance for not go beyond 2 degrees C warming in landmasses as snow melts in the major increases in carbon dioxide. How the 22nd century—a conclusion at vari- spring—such an inappropriate projec- I wish this would turn out to be true! But wishes aren’t analysis. One study is decision stopped IPCC from looking at rise in sea level driving small-island in- cited; ignored is the huge body of eco- the total cost-benefit of global warm- habitants from traditional homelands), nomics work he later accepts to estimate ing.” (As an aside, I should mention that and likely changes to climatic extremes a range of costs if we were to implement it is strange he chose to cite the penulti- and variability.