LOMBORG-ERRORS                       Home
                        List of chapters
Chapter 24:

Global warming

                       

 

MAIN ISSUE:

Lomborg´s chapter on global warming is, to a large extent, based on a review of the 2001 reports from the International Panel on Climate Change (IPCC). However, his review is biased. Lomborg systematically picks every little piece of information which may serve to downplay the importance of manmade effects on the world´s climate, whereas all information that points in the opposite direction is either neglected or heavily criticized. Also the economic discussions are extremely biased, and, furthermore, confusing and inconsistent. The text is marred by frequent attacks on the experts, i.e. attempts to undermine the reader´s trust in their expertise and their ability to make balanced judgments. Lomborg, who is a non-expert in this field, and whose ideas have not been subject to any relevant peer review, thus presents himself as having a better understanding of the field than the thousands of international experts whose meticulous works are subjected to three or four rounds of peer reviews before they are published. The only positive thing that can be said about this chapter is that it must have taken a lot of work to produce it - so much that one may wonder if Lomborg has really produced it on his own. Although there may be a few paragraphs here and there that are not very biased, the overall impression of the chapter is that it is altogether deliberately misleading. Anyone wanting information on the subject would have no reason to read Lomborg instead of the original IPCC reports.

Recommended references:

The IPCC reports from 2001 and later, to be downloaded here.

The book "Sceptical Questions and Sustainable Answers", with a chapter by Jesper Gundermann on the physical aspects of climatology, and chapters by Jesper Jespersen, Anders Christian Hansen and Alex Dubgaard on the economic aspects of the greenhouse effect. It may be downloaded here.

 

P. 259 left: (COMMENT)

"This chapter accepts the reality of man-made global warming but questions the way in which future scenarios have been arrived at and finds that forecasts of climate change of 6 degrees by the end of the century are not plausible." Comment: No scientists have made a forecast that the climate change will be 6 degrees. Scientists have not forecasted definite values, but have described a range of possible values. Thus, IPCC (2001) says that the global mean temperature is expected to rise by 1.4° to 5.8° C from 1990 to 2100, and that this temperature interval is the result of 35 scenarios fed to several different climate computer models. As a person teaching statistics, Lomborg should have had a basic understanding that there exists a likelihood distribution with an average and a standard deviation, and that in principle, those values that fall at the extreme ends of the interval are no less "valid" than those near the average. The term "plausible" is imprecise and undefined and should not be used by a statistician in such a context.

P. 261 left: ERROR OF BIAS

"The records seem to indicate substantial temperature swings throughout the Holocene on a millennial scale, and some indicators even show changes of 5 to 8° C over 1,500 years." Error: Although a temperature rise of this magnitude happened just before the start of the Holocene, there is little to indicate that such swings happened throughout the Holocene. Actually, the Holocene seems to be have been a period of unusually stable temperatures. It is true that reference 2122 (IPCC 2001) describes a single drill sample that shows large temperature fluctuations, but this deviates relative to the much larger number of other drill samples mentioned in the same reference. When Lomborg refers to a single deviating sample as "the records", and neglect all others, this is a severe bias. As Lomborg does not have any scientfic insight which allows him to prefer the deviating sample to the others, this bias is deliberate.

P. 261 right: (COMMENT)

"This graph allows IPCC to conclude that not only was the last century the warmest in the millennium but the 1990s the warmest decade . . " Comment: As so often, Lomborg transforms tentative conclusions to definite conclusions, in this case by using the phrasing "allows to conclude". Actually, IPCC makes due reservations and says that it is likely that things are as stated. It is unacceptable that Lomborg omits this reservation, because afterwards, Lomborg criticizes IPCC for making statements that are not plausible.

P. 261, text to figure 134: FLAW

Flaw: Lomborg fails to mention that whereas Pollack´s curve covers the whole of the globe, Mann´s curve covers only the northern hemisphere, Jones´curve covers primarily the northern hemisphere outside of the tropics, and Briffa´s curve covers exclusively the northern hemisphere north of 30° N. Thus, the curves are not directly comparable.

P. 261-262 and figure 134: FLAW

Lomborg has a detailed discussion of the difference between the Mann curve and the other curves, especially the Jones curve. Flaw: As the Jones curve fits better than the Mann curve into the picture that Lomborg wants us to see, he tries to downplay the validity of the Mann curve. He does this mainly by describing that, for the period 1000-1400, it relies mainly on tree rings, which means that it reflects just summer temperatures and does not say anything about temperatures in the sea. However, the Jones curve is based even more on tree rings, but is not criticized for this. Thus, the text is biased. The discrepancy between the two curves is seen mainly during the period 1500-1800, and in this period the Mann curve is not mainly based on tree rings and thus not subject to Lomborg´s points of criticism. Furthermore, Lomborg says (262 right) that "there is considerable disagreement" between the curves. This is not true. Mann has published his curve with clear indication of the uncertainties, and from this it appears that the two curves are not significantly different. Lomborg, who should be able to understand statistics, should not have concluded that the curves disagree, when in fact one curve lies within the uncertainty interval of the other. As there is no obvious reason to criticize the Mann curve and not the Jones curve, Lomborg´s text is deliberately biased.

P. 261-262 and figure 134: FLAW

Lomborg discusses the difference between the Mann curve and Pollacks borehole measurements. On p. 262 right he writes: "The only data estimated from direct measurements of temperatures come from boreholes and these show a dramatic rebound from the Little Ice Age." Flaw: As seen in figure 134, the Pollack borehole curve is linear for each century. This is because it is defined for each century by an average and a linear trend. If one wants to compare it with the other curves, then its average for the 20th century has to be the same as the average for the 20th century of the other curves. In Pollack´s original paper, the curve is indeed placed according to this reasoning. However, Lomborg has chosen to adjust it to the average for 1961-90. Thereby, he gives the Pollack curve a different average from that of the other curves during the 20th century: the Pollack curve is displaced by 0.2° C downwards. In this way, the Pollack curve is made to show the opposite of what it did before. Without displacement, it confirms the recent temperature rise. With displacement downwards, the recent temperature rise becomes less noticeable, and instead, the coldness of the Little Ice Age becomes considerably exaggerated. As such a displacement was not made in the original paper, this change is deliberately misleading.

P. 262 right: ERROR

"The only data estimated from direct measurements of temperatures come from boreholes and these show . . . ." Error: The borehole data are certainly not direct temperature measurements. They are proxy data as least as much as the data behind the other curves are. The previous temperatures are deduced from the drill samples by complicated indirect calculations. The drill samples tell about the heat flux from the surface downwards, but this heat flux depends not solely on the air temperatures, it also depends on moisture, ground colour, vegetation etc. If these parameters change over time, then the relationship between the drill sample measurements and air temperatures will change. The impression Lomborg gives, viz. that Pollack´s curve is probably more reliable than the others, is false.

P. 263 left: FLAW

" . . the Mann data impart an impression of an almost stable climate system . . " Flaw: Lomborg fails to mention that the Mann curve was published with indications of such an uncertainty that temperature fluctuations of the size order seen e.g. in the Jones curve can be contained within the uncertainty interval. He is not therefore justified in giving the impression that the Mann curve is unreliable.

P. 263 bottom left: FLAW

"The IPCC finds that some of the increase [i.e. from 1910 to 1945] can be explained by a substantial and natural increase in solar irradiation from 1700 onwards, which however is still only poorly quantified." Flaw: This is an understatement. Actually, when the combined effects of variable solar irradiation, vulcanic eruptions, and greenhouse gases are combined, the agreement with the observations is quite good, which is evident e.g. from Figure 4, p. 11 in IPCC (2001), and also from figure 12.7 in the same report - a figure which Lomborg refers to in note 2172.

P. 264 top left: (COMMENT)

" . . . with none considered more plausible than others . . ". Comment: Lomborg seems to have a problem with this. Apparently Lomborg has misunderstood the whole purpose of the scenarios, see the page on scenarios.

P. 264 left: (COMMENT)

"This makes all the scenarios somewhat artificial and certainly worst-case". Comment: Lomborg seems to have misunderstood the whole purpose of the scenarios, see the page on scenarios.

P. 266 right: MISLEADING STATEMENT

" . . the result of simulations depends entirely on the parameters and algorithms with which the computer is fed." Flaw: While this statement - as it stands - is true in a very general sense, it carries the implication that the IPCC models are arbitrary, depending entirely on the correlations between the data fed to them. This is not so, however. Contrary to models e.g. in economics, the climate models incorporate known physical constants and known physical laws, which allows predictions outside of the range experienced up to now. So the results do not depend entirely on the data fed to the computers. Lomborg´s text makes the computer models appear less reliable than they actually are.

P. 267 left: (GROUNDLESS DEROGATION)

" . . one of their surprisingly overlooked statements:" Comment: The wording suggests that something important has been overlooked or neglected. It has not. At present, it is standard practice to involve the effect of sulphate aerosols in the climate models.
Read more elsewhere on Lomborg-errors about the concept of groundless derogation.

P. 267 bottom left: FLAW

" - either it is not going to warm as much as previously claimed or . . " Flaw: By using the words "going to", Lomborg indicates that he is writing about the future. The reader gets the impression that the temperature rise in the future will not be so large as previously claimed. However, the quoted text from the 1996 IPCC report deals only with trends over the last century, not with the future.

P. 267 right, figure 138 and note 2172: ERROR

" . . although the rapid temperature increase from 1910 to 1945 is still left unexplained." . - In the note: "Some of this increase might be explained by . . ". Error: Lomborg´s text is deliberately misleading. IPCC´s Figure 12.7, which is referred to in Lomborg´s note, has three parts (a similar figure is in Stott et al. (2000): Science 290: 2133-2137). The first part shows that man-made changes cannot fully explain the temperature course. The second part shows that natural fluctuations cannot fully explain the temperature course. And the third part shows that the combination of man-made and natural causes may give a practically full explanation of the temperature course. Lomborg deliberately neglects the combination of the two causes. He maintains that the man-made cause and the natural cause are two alternative hypotheses, and refuses to combine them. Only by refusing this can he maintain that the models do not fit reality fairly well. If he allowed the combination of causes which every sensible person would allow, he would have to admit that the drastic temperature rise during the latest decades is practically exclusively due to man-made effects, and apparently, Lomborg is willing to do anything to avoid this conclusion. The inclusion of his figure 138, which demonstrates that models of man-made influences cannot explain the temperature course very well, instead of IPCC´s figure 12.7, which he has seen, and which demonstrates that man-made influences can explain the temperature course (when combined with natural fluctuations), is deliberately misleading.

P. 268 left: FLAWED ARGUMENTATION

" . . . picking just one cooling effect of many cooling and warming effects . . " Flaw: This is a flawed argumentation. Out of the effects shown in Lomborg´s figure 139, the sulphate aerosol effect is the largest of those that have a reasonable degree of certainty. Therefore, if just one of the effects is included in a model, it has to be this one. If one more is included, this would reasonably be ozone, which is actually the case, as evident from note 2178. The original figure, as presented in Summary for Policy Makers by IPCC, indicates that, for all other parameters than the ones here mentioned, the level of scientific understanding is "very low" - an obvious reason not to include them. On the other hand, if the many very poorly understood positive and negative effects were included, one would be able to get just the degree of fit that one wanted; thus, by including many effects, the scope for manipulation would be greater, not less. As said in note 2178: "Most integrators do not include the less dominant or less well understood forcings . . ". Thus, the scientists apply fairly objective criteria to avoid manipulating, in direct contrast to the impression that Lomborg tries to impart.

P. 268 bottom: FLAW

" . . whether it is sulfate particles that are temporarily cooling an otherwise relentless CO2 warming . . " Flaw: The non-expert may from this paragraph get the false impression that because of the effects of particles, which are not properly understood and hence not properly included in the models, the future rise in temperature may be less than predicted. What is actually implied, however, is the opposite: If it turns out that the cooling effects of particles are larger than assumed in the models up to now, then this means the model predictions of the effect of CO2 alone should be larger in order to balance the large cooling effect. As the number of particles will probably not rise in the future, whereas CO2 concentrations will continue to rise, such modified models would in the long term predict much larger temperature increases. To obtain Lomborg´s alternative, viz. that "the climate sensitivity to CO2 truly is smaller than previously expected", the cooling effects of particles would have to be smaller than believed at present.

P. 268 right, note 2182: (CORRECTION)

Correction: The cited conclusion is on p. 9 in the summary for policy makers, not on p. 5.

P. 268 bottom - 269 top, notes 2182 and 2183: ERROR

"For instance, studies making optimistic conclusions . . " Error: The reference in note 2182 is from 2001, whereas that in note 2183 is from 1999. Thus, the latter cannot possibly conclude anything about the former. And indeed, it does not. When IPCC in 2001 wrote about a large-scale consistency between models and observations, they referred inter alia to a work by Stott et al. (Science 290 (2000): 2133-2137), which incorporates a series of aerosol effetcts, including indirect effects on cloud reflectivity. The quote from ref. 2183, on the other hand, deals exclusively with such works which do not include the indirect effects of sulphate aerosols. Thus, the quote "would be significantly modified if . . " does precisely not deal with the type of studies referred to in the preceding sentence. At best, this error is due to gross negligence.

P. 269 top left and note 2185: ERROR OF BIAS

"Moreover, recent estimates seem to indicate that the direct cooling effect from particles may be much smaller than previously expected." Error: Here, Lomborg focuses on an estimate that the total forcing of particles should be +0.55W/m². This value is taken from Jacobson´s paper in 2001 in Nature. We get the impression that this is a later correction to the IPCC survey that is presented in Lomborg´s figure 139. But this is a deception. Actually, Jacobson (2001) is one out of several papers listed in IPCC 2001a in their section 6.7.3, table 6.5. This table contains 18 estimates of radiative forcings due to different kinds of aerosols, and Jacobson´s value of +0.55 or +0.54 W/m² is the highest (most positive) of all. Even relative to other estimates of the same category of aerosols, it is higher. When Jacobson´s (negative) figure for organic carbon forcing is added to his figure for black carbon forcing, the total forcing from these two types of aerosols becomes about +0.5 W/m². In comparison, the same parameter would be -0.16 W/m² if figures from Hansen et al. (1998) were used in the same way. Thus, here again, out of a range of values, Lomborg has simply taken the highest figure he could find, and then claimed that this figure represents "the newest study" using "better mixing models". When other, equally valid estimates are not mentioned, there seems to be an intentional bias. - The most widespread contention among experts seems to be that the warming due to black carbon only partially neutralizes the cooling effect of other particles. For an unbiased presentation of the subject, Lomborg should have included references to other literature, such as U. Lohmann & J. Feichter ( 2001): Geophys. Res. Lett. 28, 159-161.

P. 269 left: (COMMENT)

"The IPCC therefore estimate that the effective warming . . ". Comment: The word "therefore" is somewhat misleading, because IPCC does not conclude that more water vapor is the only factor that increases the warming over what is predeicted by the simple effect of CO2.

P. 269 top right: FLAW

"Basically, the feedback only works effectively if the entire lower atmosphere warms up . . ". Flaw: This is not quite correct. The crucial issue is to what extent the concentration of water vapor increases in the middle and upper troposphere. As this part of the troposphere is far from saturated with water vapor, an increase in moisture is possible in principle even if the cold air here does not get warmer. It would suffice that the earth´s surface and the very lowest part of the troposphere got warmer, because this would lead to increased rising of air from the surface, and hence increased convection that could bring moisture to higher levels. It should be added that the study by Hall & Manabe (note 2191) is based only on a model of the climate system, and therefore their study does not necessarily describe what actually happens. Thus, their description may be correct, but an increase in water vapor without much heating remains a possibility.

P. 269 right: FLAW

"Numerous studies have tried to identify the causes of the difference between surface and tropospheric temperature . . " Flaw: Lomborg does not cite these "numerous studies", so the reader will not know which they are. He is misleadingly imprecise here and gives the impression that the causes have not been identified, partly by distracting the reader´s attention to the question of satellite orbits. Actually, the IPCC report refers to two main papers (Bengtsson, 1999 and Hansen, 1997) which give detailed accounts of how one may explain the disagreement between observed trends on the surface and in the troposphere. These explanations include the eruption of the Pinatubo volcano in 1991 and the effect of changes in ozone concentrations due to ozone decomposition, which is appreciable in the troposphere, but not very appreciable at the surface. Lomborg cannot possibly have found the information necessary for his presentation without consulting the relevant parts of the IPCC report, and in doing so he must necessarily have become aware of the Bengtsson and Hansen papers. But Lomborg neglects the explanations presented to him, and writes in his own text that there is an unexplained lack of temperature increase in the troposphere.

P. 269-270 and figure 140: ERROR

Lomborg describes that whereas the surface temperature has risen approximately in accordance with the predictions, the troposphere temperature has risen very little, although the models predict that it should have increased even faster than the surface temperature. Error. The four curves in the figure are not adjusted to the same base line. Whereas the model simulations and "observed surface" are drawn relative to the mean temperature for 1950-1960, the curve for "observed troposphere" is drawn relative to the temperature around 1992 (cf. fig. 141). For the period since 1958, IPPC (2001) concludes on its p. 102: "Analysis of global temperature trends since 1958 in the low to mid-troposphere from balloons shows a warming of about +0.1° C/decade, which is similar to the average rate of warming at the surface." As Lomborg has read the IPPC report, he must know this. He should therefore have placed the troposphere average for 1979-2001 about +0.2° C above the base line, or - if not - have stated explicitly that this curve is not drawn at a level comparable to the others. Furthermore, the curves have not been smoothed in similar ways. Whereas "observed troposphere" is not smoothed, "observed surface" is smoothed over time, and the model curves are the average of results from two models. In this way, the "observed troposphere" gets a much more fluctuating appearance than the three others.- That the actual temperature increase is concealed by the displacement of the troposphere curve, seems to be a deliberate manipulation.

P. 270, bottom left: (COMMENT)

" - a difference which if persistent should lower the water feedback . . " Comment: This argumentation is probably faulty. If the rate at which the temperature drops with height becomes larger, this could have the opposite effect of what Lomborg claims. Firstly, it could increase convection and thus lift more water vapor upwards. Second, it would mean that the temperature at the surface would have to rise even more before the warming at higher levels were sufficient to bring about the necessary increase in heat radiation.

P. 270, bottom right: FLAW

"The programmers then improved the cloud parameterization in two places, and the model reacted by reducing its temperature estimate . . ". Flaw: First, Lomborg should have used the word "revised" instead of "improved", because the authors ask their readers to "note that although the revised cloud scheme is more detailed it is not necessarily more accurate than the less sophisticated scheme." Second, he refers to relatively old models, not to the newest revisions of the Hadley Center´s model. If his text had been unbiased, he would have cited the newest relevant literature on this point.

P. 271, figure 141: FLAW

Flaw: As in figure 140, the two curves are not placed relative to the same baseline, thus exaggerating the divergence between them. The figure is therefore deliberately misleading.

P. 271, figure 141: FLAW

Flaw: The curves for the troposphere show the average of the data for the 850, 500 and 200 mb pressure surfaces, with equal weighting to each. This gives far too much weight to the upper surfaces, especially the 200 mb surface which extends into the stratosphere. The point is that the human-induced ozone losses in the lower stratosphere have induced a strong cooling trend, which is assumed to have an effect down to the middle troposphere. This "top-down" cooling effect provides a consistent explanation of the "discrepancy" between the surface and tropospheric warming trends.

P. 271 right: (COMMENT)

Lomborg presents the iris hypothesis advanced by Lindzen, Chou & Hou in 2001. Comment: The iris hypothesis was, when presented, weakly founded, based on only a few years´data from a small part of one ocean. Already at the beginning of 2002, papers appeared that demonstrated logical flaws in Lindzen´s hypothesis. Actual measurements, reported later in 2002, indicate that the magnitude of the iris effect is at most 1/10 of what was claimed by Lindzen et al., and it may possibly be close to zero (Chambers, Lin & Young (2002): J. Climate 15: 3719-3726, and refs. therein). This is an example of why it is premature to spread "hot" new research results.

P. 271 bottom right: (COMMENT)

"This means that throughout the past 25 years the basic range of estimates of global warming from CO2 has not improved. " Comment: See comment to p. 317 left.

P. 272 bottom left: (COMMENT)

"(the only scenarios the IPCC had time to run)". Comment: This is a misunderstanding. The IPCC does not run model simulations, nor engage themselves in doing it. The sentence should have read: "the only scenario simulations made available for the IPCC (2001) assessment".

P. 272: FLAW

On p. 272 (bottom left and top right) Lomborg wants to make the point that the uncertainty due to differences between climate models is larger than the uncertainty due to emission scenarios. Flaw: In figure 137, 6 curves span a range of 2.5°C. From this, everything else being equal, we would expect that 9 curves would span a range of 3.06°C (2.5 mulitiplied by the square root of 9/6). Now, in figure 142 left and right, we have 9 curves which span 3.6° and 2.6° C, respectively, i.e. on average 3.1°C. this is as close as one can get on the predicted 3.06. Thus, considering the different number of curves, the uncertainty is exactly the same. Lomborg makes two flaws: He disregards the different number of curves, and he focuses on just that half of figure 142 that supports his point, neglecting the other half which does not. It is embarrassing that a person claiming to be a statistician makes such a mistake.

P. 272 right: ERROR OF QUOTATION

"Discouraging but honest, the IPCC concludes that . . " Error: The quoted sentence does not exist in the reference. On the contrary, section 9.3.2.1 says: "Thus, the choice of model and the choice of scenario are both important." And later, at the end of their section 11.5.1.2, they write: "That is, for sea level rise . . the uncertainty in climate sensitivity . . is more important than the uncertainty from choice of emissions scenario. This is different . . from the case of global average temperature change (section 9.3.2.1.), where the scenario and modelling uncertainties are comparable." It seems impossible that Lomborg´s distortion of the original text could be anything but deliberate.

P. 272 right and note 2218: FLAW

". . the IPCC seems to have chosen a rather pessimistic simple model . . " And in note 2218: ". . the simple model has been calibrated with a different set of AOGCMs, some of which surprisingly have not even run the new scenarios . .". Flaw: The models (the word "scenarios" is a mistake here) referred to in the note are not new. Of course, the newest models have the highest code numbers. Thus, GFDL_R30_c in figure 142 is newer than GFDL_R15_a in the note, and HadCM3 in figure 142 is newer than HadCM2 in the note. The "new" models preferred by Lomborg are actually outdated, and this is why they were not used. He tries to imply a bias in the IPCC work. But instead, this attempt demonstrates his own bias. The word "surprisingly" is an example of unjustified derogation.

P. 272 bottom right and note 2220: FLAW

". . this indicates that the IPCC simple model systematically overestimates the warming . . " Flaw: No, it does not. The simple model clearly lies well within the range spanned by the other models. For instance, in the right half of figure 142, the standard deviation of the average of +2.17°C in 2100, is 0.75, which means that 2.69° is well within one standard deviation from the average. Thus, the hypothesis that the simple curve does not deviate from the average of the other 9 curves, cannot be rejected. As a statistician, Lomborg should have been able to understand this. From an expert´s point of view, the models ECHAM and HadCM3 seem to have the most realistic description of the underlying physical processes, which is why it is reassuring that the climate sensitivity of the simple model is so close to just these two models.

P. 273 top right: FLAW

" . . still makes the models more noisy than the climate response." Flaw: First, the sentence is nonsense. Instead of "climate response", there should have been the words "emission scenarios". Second, as stated in the comments to p. 272 right, even when the sentence is corrected, its content is simply wrong.

P. 273 top right: FLAW

"Moreover, the simplistic models used by the IPCC appear to overestimate the climate sensitivity." Flaw: As stated in the comments to note 2220, this conclusion is not warranted. Altogether, the whole section on clouds is written and structured as if it were a scientitic discussion of the issue and will be understood as such by any non-expert. In reality, however, it is systematically biased, stressing every uncertainty that points in the direction of less warming, and minimising every uncertainty that points in the direction of more warming. Thus, it cannot be used as a survey of the existing state of knowledge.

P. 273 left, box on ozone hole: FLAW

"Ozone depletion is also linked in several ways to climate change, but the links are fairly weak and can be disregarded here." Flaw: Lomborg refers to a sentence in the paragraph "Frequently asked questions about ozone" in the executive summary of the WMO/UNEP scientific assessment of ozone depletion (1998). Here, the sentence is: "Ozone depletion and climate change are linked in a number of ways, but ozone depletion is not a major cause of climate change." The context is a discussion of the relative importance of CO2 increase and ozone depletion for temperature changes, and the point is that the former has a greater effect on the irradiation received at the earth´s surface than has the latter. In addition it is explained that climate changes can affect the time necessary for the ozone layer to recover. Overall, WMO/UNEP concludes that ozone depletion and climate changes are linked, but the effects go more in one direction than in the other. This is distorted by Lomborg, however, so that the reader believes that climate change has only negligible effects on the ozone level. The probable effect of CO2 on ozone is described e.g. in Nature 392: 589-592 (1998) and Geophys. Res. Lett. 25: 2141-2144 (1998). To this must be added newly published research which indicates that ozone depletion may contribute to climate changes at the earth´s surface (Shindell et al. J. Geophys. Res. 106: 7193-7210 (2001)).

P. 274, bottom left: FLAW

". . ozone depletion is at its maximum and it will recover within the next 50 years." Flaw: This does not quite agree with what is said in the executive summary formulated by WMO/UNEP, which Lomborg has used as a source. Here, we read: "The maximum ozone depletion is estimated to lie within the current decade or the next two decades, but its identification and the evidence for the recovery of the ozone layer lie still further ahead."

P. 276 left and note 2259: FLAW

"Since IPCC mentions solar influences only briefly . . ". Flaw: IPCC mentions solar influences not only in sections 6.11 and 12.2 as stated by Lomborg, but also in sections 6.15, 8.5, 8.6 and 12.4, among others, and the possible effect is discussed at length, not just briefly.

P. 276 left: FLAW

" . . it is probable that including indirect solar activity will also lead to a lowering of the estimates of the CO2 warming effect." Flaw: It is wrong to use the word "probable" here. The proper word would be "possible", i.e. the likelihood that this could be so is only small, cf. the uncertainty range for solar forcing in Lomborg´s fig. 139.

P. 276 bottom left - top right: FLAW

"Probably, solar brightness has increased about 0.4 percent over the past 200-300 years, causing an increase of about 0.4°C. ". Flaw: Although it may be correct that changes in solar irradiation have contributed with 0.4°C over the last 300 years, Lomborg´s reasoning is not correct. The change over the last 250 years indicated in figure 139 is about 0.3 W/m², which is 0.13 % of 235 W/m². Thus, a change of 0.4 % does not agree with figure 139. Also, the climate sensitivity indicated by Wilson (1997) is not 1°K/(W/m²) as claimed by Lomborg, but much less (about 0.4). So, on the basis indicated by Lomborg, the temperature increase would not be as indicated. However, going further back to the minimum in sunspot activity around 1650-1700, one does find a considerable rise in solar forcing from then and until now. In addition, most models assume a climate sensitivity somewhat larger than Wilson´s indicated above, which would mean that the temperature rise due to increased solar irradiance since about 1700 might be around 0.4°C.

P. 276 right and note 2261: FLAW

"A recent AOCGM study showed that the increase in direct solar irradiation over the past 30 years is responsible for about 40 percent of the observed global warming." Flaw: In the main text, the reader only hears about what points to such an effect. Only if he consults the note will he see that others think that there is very little effect. The "recent" study referred to by Lomborg is from 1997, whereas it is contradicted by the later and more authoritative report from IPCC (2001), which Lomborg has read. On p. 706 in the IPCC report is the following statement: "All reconstructions indicate that the direct effect of variations in solar forcing over the 20th century was about 20 to 25% of the change in forcing due to increases in the well-mixed greenhouse gases." Also, the authors of the 1997 study (Cubasch et al.) mention that the solar forcing used in their study may have been unrealistically large.

P. 276 bottom - 277 top and note 2263: FLAW

" . . but the researchers have actually achieved an impressive correlation going back as far as 1550." Flaw: In the note, Lomborg says that the paper referred to had a small problem with joining the two temperature curves. This is an understatement. Because the temperature curve for 1579 to 1860 is displaced upward by about 0.1° C, the sunspot curve was made to coincide rather well with the temperature curve. However, when the temperature curve is placed correctly, all data points on the sunspot curve before 1800 fall above the temperature curve, indicating that there has been a general temperature rise which cannot be accounted for by the sunspots. So, when this "small problem" is corrected, the conclusion is greatly changed. This was pointed out to Lomborg in 1999, and relative to the Danish version of his book, he has revised the text and put the reservation in the note. However, this reservation is minimised to such an extent that the text remains deliberately misleading.

P. 276 bottom - 277 top and notes 2263-2264: INCONSISTENCY

Here Lomborg talks about the "measured average temperature" and "an impressive correlation". Flaw. The so-called measured temperatures are actually, to a great extent, proxy data, especially when we go as far back as 1550. It is inconsistent that Lomborg presents these temperatures without any reservations and apparently with great trust, although the same kind of proxy data were severely criticized by him on pp. 260-263. This indicates a bias in his presentation.

P. 277 left: (COMMENT)

" . . new research seems to have found the connection . . ". Comment: A proper wording would have been: "new research has suggested a possible connection . .".

P. 277 right and figure 147: FLAW

"But Svensmark, along with several others, has shown that there seems to be a clear connection between global low-level cloud cover and incoming cosmic radiation." Flaw: First, there is no mention of who the alleged "several others" are. Second, in a complementary paper (Svensmark & Friis-Christensen (2000): J. Atmospheric and Solar-terrestrial Physics 62: 79-80), the authors downplay the strength of their theory by writing: "a possible effect of the cosmic ray flux cannot be ruled out". And Marsh and Svensmark have in 2001 on www.dsri.dk/~ndm/CLOUD_UPDATE/UPDATE.html concluded: " . . the uncertainties in cloud amount for the period after 1994 are clearly too large at present to draw any conclusion on either the absence or presence of a correlation with the cosmic ray intensity." In fact it is known today that the correlation between cloud cover and cosmic radiation falls apart after 1994. It is therefore not true that there is a "clear" connection.

P. 277 bottom - 278 top: ERROR

"This theory also has the tremendous advantage, compared to the greenhouse theory, that it can explain the temperature changes from 1860 to 1950, which the rest of the climate scientists with a shrug of the shoulders have accredited to `natural variation´ ". Error: First, it is quite wrong of Lomborg to formulate this as if the theory on impacts via cosmic radiation could be put instead of the greenhouse theory. Second, if the two theories are combined, it turns out that greenhouse gases explain about twice as much of the temperature rise from 1860 to 1950 as do the changes in cosmic radiation. Third, in its summary for policy makers, which Lomborg has read, IPCC (2001) presents a figure 4 showing that it is possible to explain the changes since 1860 exclusively by solar irradiance, vulcanic activity, and greenhouse gases. What Lomborg writes here is thus directly at variance with what he has read, which means that the text is deliberately misleading.

P. 278 left: ERROR

". . the fact that the emerging greenhouse gas signal only appears now . . ". Error: It is not a fact that the emerging greenhouse gas signal only appears now. At most, this postulate could be called a theory. Lomborg knows very well, e.g. from Laut and Gundermann (1998) cited by him, that his statement is flawed. There is a certain fit between solar cycles and temperature in fig. 146. Another fit could be made of greenhouse gases and temperature. Both fits are moderately good, but if the effect of greenhouse gases and the pattern of solar cycles are combined in one correlation, then the fit with the temperature curve is improved considerably. This suggests that the development of the temperatures is a combined effect of the two causes, an effect that becomes more convincing when they are combined than if either is considered alone. In this way, inclusion of the solar cycles actually gives further corroboration of the greenhouse effect, precisely opposite to the sarcastic comments made by Lomborg in his note 2271. An estimate of the relative contribution of the two causes up to now is that the greenhouse effect explains about 70 % of the total temperture increase. (Gunderman, in "Sceptical Questions and Sustainable Answers). As all this is well known to Lomborg, and as he nevertheless omits crucial arguments advocated by his opponents, his text is deliberately misleading.

P. 278 left: ERROR

"One such IPCC-loyal study finds that the solar hypothesis explains about 57 % of the temperature deviations . . ". Error: This is a misunderstanding. What the authors of the study wrote was that if we explain a part of the variance in temperatures as due to solar cycles, then there remains a certain residual variance. If we then include the correlation with greenhouse gases, the residual variance is further reduced to 43 % of its former value. Thus, if the solar cycles are entered first, then greenhouse gases explain 57 % of what remains. Lomborg, who claims to understand statistics, has completely failed to understand this.

P. 278 left: (COMMENT)

"Are the scenarios realistic? ". Comment: This is the wrong question, because the scenarios are per definition not realistic. The baseline scenarios explicitly exclude any efforts to mitigate climate change. The reason for this is that one of their purposes is to function as a tool - researchers are meant to take the baseline scenarios and then add specific climate mitigation efforts to study what the effects of these would be on environment and economy. If some mitigation were already included, then it would not be possible to find out e.g. what the total costs of mitigation would be. So the question is not if the scenarios are realistic or likely; the question is if they are possible. See the page on scenarios

P. 278 right - 279 left: ERROR

"Here, the IPCC have assumed that the CO2 concentration will grow by 0.64 percent per year . . ". Error: The argumentation that follows from here contains a mathematical blunder. Lomborg assumes that the CO2 concentration grows exponentially; that is not so. The additions to the natural background concentration may grow exponentially. At present, the yearly additions are on average1.6 ppm, i.e. about 1.6 % of the total 100 ppm addition up to now. If this rate continues, the total addition by the year 2100 will be roughly 490 ppm, i.e. the concentration will be 280 + 490 ppm = 770 ppm, which is roughly a doubling relative to present levels. Or we can get the same result by adding about 1.6 ppm per year to the present concentration of about 375 ppm , which, as stated by Lomborg, is an addition of about 0.43 % per year. The CO2 emissions have grown by about 1.6 % per year during the period 1970 to 2000. If emissions go on to grow exponentially like that, and if the concentrations increase in proportion to the emissions, then the yearly additions expressed in percent of the present CO2 concentration will increase gradually from the present 0.43 % to just above 2 % in the year 2100. The sum of these one hundred additions, starting with 0.43 % and ending with just above 2 %, will amount to a total of just above 100 %. That is, the CO2 concentration will slightly more than double in 100 years, precisely as assumed in the IPCC scenarios.

P. 279 bottom right: ERROR

"Yet, most standard computer simulations use an even higher value for CO2 increase, namely 1 percent." Error: The emission scenarios were formerly the IS92 scenarios, and from 1996 onwards they have been the so-called SRES, i.e. special report emission scenarios. There are six of these, which are presented in Lomborg´s fig. 136, together with one IS92 scenario. These are the ones used as a basis for runs of calibrated climate models. The type of scenario containing a 1 % per year increase in CO2 equivalents (not in CO2, as Lomborg claims), is a type of standard model that is used simply for calibrating and comparing different climate models in a more preliminary stage. The 1 % is an arbitrary standard, chosen to lie in the high end relative to actual emission increases, in order to reduce the influence of "noise" on the results. It is a standard in the same way as a calorie is a standard measure of the amount of heat required to warm one gram of water by one degree. In the same way as the 1° increase is a part of the definition of a calorie, the 1 % per year increase in this case is a part of the definition of the concept of a "transient climate response" (TCR). The executive summary in chapter 9, which Lomborg refers to, explains this as follows: "It [i.e. TCR] is defined as the globally averaged surface air temperature change for AOGCMs (models) at the time of CO2 doubling in 1%/yr CO2 increase experiments." Considering that Lomborg knows very well that the forecasts he refers to are based on IS92s and SRESs, which clearly do not have a 1%/year increase, and considering that he has read the quoted sentence defining a TCR, the presentation, if not deliberately misleading, is at least an indication of gross negligence.

P. 279 bottom right: (GROUNDLESS DEROGATION)

"It is truly hard to understand why modelers . . ". Comment: Yes, and the explanation is that they don´t.

P. 280 left: FLAW

"Typically, the models that we are presented with in the press are exactly these sorts of models that run much faster than the IPCC scenario . . " Flaw: This is not true. The simulations referred to by the press are nearly always based on IS92A or SRES scenarios. In this context, the article in Scientific American is a rare exception.

P. 280 left: (GROUNDLESS DEROGATION)

" . . the modelers have explicitly abandoned the idea of predicting the future . . " Comment: Lomborg derides the experts and describes them as persons who have "abandoned" the goal of predicting the future. He fails to recognize that they have never set this goal. The scenarios are just plausible "stories", with no likelihood ascribed to them, except that the likelihood is above zero. The purpose of the scenarios is to be able to say: If emission trends were like this, then the climate would probably become like that. So their purpose is not, and has never been, to predict anything. At least Lomborg acknowledges on p. 280 that the present aim is not to make predictions. But on p. 282 he has already forgotten this and treats the scenarios as predictions, which he then criticizes for being unrealistic.

P. 282 left: FLAW

"Remarkably, the scenario authors, without any explicit criteria, rank the four scenarios . .  " Flaw: In the cited paper, the authors do give criteria for the ranking: It expresses a combination of the vulnerability for adverse climate change impacts and connditions for mititgation and/or adaptation markers, based on the three measures "potential damage", "population" and "income".

P. 284 bottom left and note 2335: FLAW

"Here, it is not technical progress and economic reasons that limit the use of fossil fuels but rather "environmental concerns."" Flaw: This is not correct. The cited paper says: "The main driving force for the penetration of nonfossil options is the highly successful development of solar, wind, safe nuclear, and others: their costs fall to levels of 3-5 US cents/kWh by the middle of next century. This makes them increasingly competitive as coal, oil, and gas prices continue to rise." After this sentence, the authors put a footnote which is cited by Lomborg in his note 2335. The purpose of that note is to treat the issue that whereas nobody doubts that oil and gas will become expensive, it is not given that coal will necessarily become expensive. Therefore the authors mention possible causes that might make coal less competitive. The issue is not that whether this will happen, but whether it possibly could happen.

P. 284 top right: (COMMENT)
" . . a surprising number of somewhat naïve and clichéd statements. Summing up, the modelers tell us that the B1 scenario is not value-free . . ". Comment: Lomborg is apparently provoqued by the description of a future world where words such as peace, reconciliation and mutual trust enter into the description. And he is provoqued that the shift away from fossil fuels (coal) is not necessarily caused by price mechanisms, but be ethical values in the population. One may say that Lomborg is a market fundamentalist: he accepts no other way that energy use can shift than price mechanisms. In this way, Lomborg himself is not value-free. He is unable or unwilling to accept that people could make choices that are motivated by ethical rather than economic considerations. For those of us who are not market fundamentalists, it is conceivable that such a world could possibly arise. There are no statements as to how likely it is; it is just stated that it could conceivably happen. And it is interesting to ask: "What if . . .". It is sensible to include such a scenario into the whole array of scenarios, because we would want to know what would happen if future world society chose to be guided by these ethical values.

P. 285 bottom left: FLAW

"" . . it seems plausible that renewable energy will by itself or with fairly little "nudging" be competitive before mid-century." Flaw: Lomborg assumes that solar energy will become economically competitive by 2030-40 just by itself or by "fairly little nudging". However, where are the economic resources behind this development of the technology ? Who is making proper investments in new technology that will not start to pay off until 40 years ahead ? This would seem to be the sort of enterprise in which the state should be engaged, not private companies. As the citizen of a country where state support has been decisive for the successful development of the technology that harvests the renewable wind energy, it is strange that Lomborg belittles the importance of the state. Wind energy has been developed to its present stage because of pioneers driven by worries about future climate changes and energy reserves, and because of state support. This support has given the growing Danish wind turbine industry the necessary basis in the form of a large home market, which has given sufficient experience and economic turnover to turn Danish wind energy into an industry that is also able to compete internationally. However, the recent change of government in Denmark has - until recently - led to a change in environmental policy along Lomborg´s lines, almost stopping the home market in its tracks and thus making the industry much more vulnerable on the international market. - See also comments to p. 287 left.

P. 285 right: (GROUNDLESS DEROGATION)

". . hydrogen fusion, which incidentally the IPCC also leaves out of consideration . . " Comment: It remains doubtful whether hydrogen fusion technology can be made to work in the foreseeable future, and therefore it seems that IPCC are justified in leaving this technology out of consideration.

P. 286 top left: (COMMENT)

" . . if the price of solar energy continues its present decline of 50 percent per decade . . ". Comment: Actually, the figure of 50 % per decade is an expectation of future declines, not a reported trend. Lomborg´s source (Chakravotry et al.) states: "Ahmed reviews cost projections estimated by numerous engineering studies . . ".

P. 286 left: (COMMENT)

"Actually, 98.5 percent of all coal reserves will never be used . . ". Comment: What Lomborg forgets to mention is that in all scenarios, oil and natural gas reserves are completely exhausted.

P. 286 left: FLAW and INCOSISTENCE

" . . local transition only happens when the worn-out technology gets replaced." Flaw: Lomborg´s reference states that if users do not have perfect foresight about future prices, transitions to new energy sources will be more costly. "As a result," they say, "transition from one fuel to another will be much slower than predicted by the programming model used in this study." So, when Lomborg says what will happen when solar power "will begin to be used in the 2040s . . taking over all sectors by 2105" etc. he forgets to mention that although this is what happens in the model, in reality things will happen much slower (according to his source). Another point is Lomborg´s inconsistency. The issue here is what is termed "no-regret options" by the scientists, i.e. the fact that because of the inertia in society we do not use the newest and most efficient technology . What Lomborg does here is to mention the importance of no-regret options in order to paint a rosy picture of the future for solar energy. But on pp. 312-313, he wants to tell how expensive it will be to halt energy use, and here he does all he can to downplay the importance of no-regret options. So, Lomborg is not consistent.

P. 286 bottom left: (REMARK)

" . . tax on carbon emissions of US$100 per ton . . ". Remark: It should be noted that a tax of this size means that present energy prices are mulitplied by 4 or 5.

P. 286 right and note 2356: BIAS

"The consequence for global temperature development is shown . . ". Bias: Lomborg does not make proper reservations as to the uncertainty of these temperature prognoses. The paper by Chakravorty et al. does not actuallt state what climate model has been used to compute the trends in temperature, or even what factors have been considered. Evidently the model must be much older than those used by now by IPCC, and much more poorly based. Considering the extraordinary scepticism that Lomborg has towards the most elaborate climate models used by the world´s leading experts, it is remarkable that he has no reservations at all regarding the older model used by non-climatologists (Chakravorty et al.). On this point, the text is obviously biased.

P. 286 right: (COMMENT)

"This more realistic model contains . . " Comment: The model assumes large changes in the transport sector (electric vehicles), but no changes at all in the utilization of energy in households, i.e. no further insulation of houses, no use of more energy efficient white goods, etc. Is that supposed to be a "more realistic model" ?

P. 286 right and p. 287, note 2359: ERROR OF OMISSION

"In fact, under any reasonable scenario of technological change . . ". In note 2359: " . . the impact of an across-the-board reduction in all conversion costs will postpone the solar power transition, simply because fossil fuels get to be very cheap to use." Error: Here, Lomborg refers briefly - and only in the note - to the so-called DCC scenario in Chakravorty et al. They write about this (their p. 1219): "In the absence of any compelling reasons why the conversion cost for a particular resource - end use combination should decrease at a different rate than others, we assume that they all decrease at an equal rate over time." That is - why should it be only photovoltaic (solar) technology that should continually become cheaper ? Why should not other technologies also improve energy efficiency at the same rate ? So, they also run the DCC scenario. In the case of no or moderate energy taxes, this will cause the consumption of fossil fuels to rise even faster than in the base scenario with no technological change. Oil and gas will be exhausted soon, followed by exhaustion of coal reserves nearly 100 years earlier than in the base scenario. So even though solar energy becomes cheap, it will not be used before about 2250, and by that time, all fossil fuels will have been burnt, with the consequence that temperatures will rise greatly. The authors write: "What seems to be important in affecting emissions is not technological change in the energy sector as a whole, but the magnitude of cost reductions in the backstop technology [i.e. solar] relative to that of fossil fuels." Now Lomborg´s text in note 2359 is not a fair description of this. He forgets to mention that if improvements happen not only in solar technology, but in all energy technologies, all fossil fuels will be burnt, and we will get a tremendous greenhouse effect. Considering that Lomborg does not favor public direction of investments, but wants to leave all to the market forces, he has no reason to disregard technological improvements in other sectors. It, therefore, technology change is "across-the-board" in all sectors, the whole point of his text on pp. 285-287 is lost. It seems that Lomborg does not want to undermine his own argument, and therefore - i.e. deliberately - leaves out this point.

P. 287 left: INCONSISTENCY

" . . the truly important point is to make sure these renewable energy sources rapidly decrease in price, and this again requires substantially increased funding for much more research." Flaw: According to Lomborg, USA could easily afford to spend $2 billion a year on development of renewable energy resources. This is just 1/100 of the revenue from a $100 carbon tax. Maybe this is what he means by "just a little nudging?". However, Lomborg is inconsistent because in the box on p. 308 he claims that any additinal tax lowers the total welfare. On this whole issue, his text is confusing and illogical. Whenever taxes are detrimental to the free enterprise of the oil industry, he is against them. Whenever taxes help to ensure that the greenhouse effect will never become serious, he is for them.

P. 287 left: (GROUNDLESS DEROGATION)

"To sum up, the previous IPCC business-as-usual scenario and its usage as background for a 1 percent CO2 increase lead to overestimates in the speed of global warming." Comment: It seems here that Lomborg considers himself as approximately equal with God, i.e. as a person who knows exactly what will happen in the future. Being in this position, he is of course able to blame the thousands of experts who, unlike him, do not know how things actually will evolve.

P. 289 right: ERROR

". . global warming has often been connected to oceans that will rise several meters. . . There are, however, no grounds for these worries." Error: This is wrong. First, a part of the sea level rise is due to water expanding when it becomes warmer. At any given time, only the surface water is warmed up; the time it takes for all water in the oceans to mix, i.e. from deep layers to the surface and back again, is of the order of a thousand years. Warming up of the oceans and further sea level rise, will therefore continue long after the greenhouse gas concentrations in the air have stabilized. As is evident from Lomborg´s figure 137, when the sea level has risen by about 40 cm in 2100, the rise does not stop there. Various model simulations indicate that a doubling of the present greenhouse gas concentration will cause the ocean water to expand by 0.5 m to 2 m in the long term. To this must be added the contributions from melting inland ice and from the ice-sheets at Antarctica. If the temperature rise above Greenland amounts to 2.7° C or more (and most model simulations indicate that), then the whole of the inland ice there will become unstable. The average of the simulations is that nearly half of the ice will have melted in 1000 years, causing the oceans to rise by 3 m, and it is possible that nearly the whole will have melted in 1000 years, causing the oceans to rise by 7 m. Melting of Antarctic Ice is not predicted to contribute any larger net rise in sea level before 2100, but after that continued elevated temperatures may cause parts of the ice, especially at the West Antarctic Ice Shelves, to disappear. Such a process will most likely take many thousand years. If it happens, it will cause the level of the oceans to rise much more than will the melting of the inland ice on Greenland. All this is what can be foreseen if there is no further rise in air temperatures after 2100. See also comment to p. 315-316.

P. 293 left: (COMMENT)

"As a curiosity it is perhaps worth pointing out . . " Comment: Lomborg points out that the cyclone activity is weakest in El Niño years, and then goes on to discuss whether there could be a relation between El Niño and global warming. However, the reference for this (Landsea et al. 1999) also has another discussion, with a more obvious relation to global warming. The sea surface temperature of the North Atlantic has shifted as follows: 1869-1893: warm. 1894-1925: cold. 1926-1970: warm. 1971-1994: cold. 1995 onwards: ?. This shift between warm and cold temperatures coincides rather closely and significantly with cyclone activity: for instance there was a quiet period in 1971-1994, but very active periods just before and after that. So if we study a period that essentially ends around 1994, we will see a downward trend, which actually is not permanent. Now, global warming might conceivably affect the shift in North Atlantic sea surface temperatures, and therefore it would seem relevant to mention this shifting pattern. But if Lomborg had done that, he would have missed his point of falling cyclone activity. Instead, he uses the El Niño relationship from the same paper, a relationship that supports his main idea: El Niño is good for us in USA; we get fewer hurricanes, so don´t worry. Thus, this selection of what to present is slightly biased.

P. 293 right: (COMMENT)

"In the new 2001 report, IPCC has only found . . ". Comment: The reader gets no clear overview of what changes in weather are actually expected by IPCC. During the 21st century, they expect higher maximum temperatures, higher minimum temperatures, less difference between maximum and minimum, increased heat index over most areas, and more intense precipitation events over many areas. In addition, they consider it "likely" that there will be increased risks of droughts in the interiors of the continents, and increases in precipitation and peak wind intensities for tropical cyclones over some areas. Thus, for instance, they do not anticipate more or stronger gales outside the tropics. The only discrepancy between expected trends and observed trends up to now is that the "likely" changes in tropical cyclones have not been observed up to now (which does not preclude that these changes may appear in the future). Thus, by and large, there is fairly good agreement between what IPCC expects and what is actually happening. Lomborg´s implications that what the models tell us does not come true, is not correct. It is correct that some newspapers and magazines have written about more frequent storms and hurricanes, but who trusts what newspaper journalists write ?

P. 294 right and figure 152: (COMMENT)

"Particularly, it turns out that the number of intense cyclones . . has been declining, as has the number of cyclone days . . " Comment: Lomborg refers mainly to table 3 in Landsea et al. 1999. This table lists 13 indices of Atlantic all-basin tropical cyclones. For the period 1944-1996, 5 out of these 13 indices show a significant trend. Four show a significant downward trend. Lomborg cites three of these. One shows a significant upward trend. Lomborg does not cite that. This is somewhat selective, because he does cite three other downward trends which are not significant. These three have correlations of r = -0.14, -0.06 and -0.02, respectively. So, when Lomborg writes: "actually all three show a small, statistically insignificant decline", he includes one with r = -0.02; the regression coefficient in this case is given as 0.00 per decade. Nevertheless, the alleged statician Lomborg thinks that he can speak of this as a decline. All this, however, is what things looked like in 1996. But from then on, hurricane activity increased. There was a record of 33 hurricanes in the four years from 1995 to 1999. When these hurricanes are considered, the "significant" trends disappear. Thus, the downward trend in figure 152 has not continued. In 2002, Landsea and Goldenberg have stated (www.aoml.noaa.gov/hrd/tcfaq/G4.html) that "Instead of seeing a long-term trend up or down, we do see a quasi-cyclic multi-decade regime that alternates between active and quiet phases." The reader might consult a recent article which demonstrates that the average strength of hurricanes has been increasing in recent years; see P. J. Webster at el. (2005): Science 309 (5742): 1844-1846. See also the comment to p. 293 left.

P. 296 left: (COMMENT)

". . the catastrophe was in large measure caused by clear-cutting forests on the upstream slopes, causing more rapid runoff." Comment: In Lomborg´s view of the world, the problem is always somewhere else, never here. Thus, when IPCC expects more heavy precipitation events and hence more floodings, and when such floodings do indeed become more frequent, then, in Lomborg´s view, it is just because of something else. Here, on p. 296, it is because of deforestation (on p. 295 right, it is due to increasing populations etc., and on p. 299 left, it is due to lack of physical planning). Strangely, however, in the chapter on forests, Lomborg has written nothing about the need to protect forests in order to avoid flooding; in that chapter, the rate of deforestation in the tropics is considered unproblematic. Also, this issue demonstrates that Lomborg tries to deal as little as possible with actual problems (IPCC does expect more flooding, and this seems indeed to come true), and instead he uses a lot of space to shoot down imaginary problems (IPCC does not expect more hurricanes, and they have indeed not been observed).

P. 298 bottom right and notes 2508-2509: ERROR

"Generally speaking, it is much better that warming happens when it is cold . . ". Error: The literature available to Lomborg does not in general support his contention. The IPCC report on impacts contains a number of references, some of which support the notion that the reduction in deaths due to severe cold is larger than the increase in deaths due to severe heat, whereas others support the opposite notion. Thus, Lomborg could, referring to IPCC, have stated that the implications are mixed. He did not do that. Instead, he cites just one paper, Kalkstein & Davis (1989), which he has read thoroughly, according to notes 2508 and 2509. Unexpectedly, this paper (dealing with 48 US cities) states that hot weather events have a very significant impact on the number of deaths, especially in regions that are not particularly hot, whereas cold weather events have a less significant impact on the number of deaths, and again - this impact is felt most clearly in regions that are not particularly cold. All of this is clearly contrary to what Lomborg writes with reference to just this paper. As it is not okay to cite a paper for a contention that is clearly opposite to what the paper says, this will be counted as an error. It must be due either to gross negligence or to deliberate misleading.

P. 299 top left: FLAW OF OMISSION

Flaw: In a chapter titled "Consequences: present and future weather", one would expect an indication of whether there would be any future increase in the number of weather events with extreme temperatures. But all that Lomborg tells us is that up to now there has been no such increase. The future situation is not dealt with. Thereby, Lomborg avoids to tell us that according to the IPCC report from 2001, there will be an increase in the frequency and intensity of heat waves. As stated e.g. in section 9.3.6.1. in the report on the scientific basis, "Increased temperature variance adds to the probability of extreme high temperature events over and above what could be expected simply from increases in the mean alone." Furthermore, the temperature rises will be felt especially at night, and when the night temperatures stay high during a heat wave, this will add to the number of human deaths, according to the reference in Lomborg´s notes 2508 and 2509.

P. 300 right and note 2528: ERROR

"Such cost estimates have been made by many models, but . . the esitmates have been made using either the old IS92a business-as-usual scenario or a so-called 2XCO2 scenario." Error: Approximately as explained in a comment for p. 279 right, the 2XCO2 scenarios are used only at an intermediate step in the testing of the climate models (when making experiments how to model the interchange between atmosphere and sea), not in the final simulations. They are not used for cost estimates. It may be that Lomborg refers to extremely preliminary model runs (the 1996 references have not been checked), but it must have been clearly apparent from their descriptions that these were, of course, not actual simulations. The most easily available reference in Lomborg´s note 2528 (viz. IPCC 1997a: 31) has been checked. It is absolutely wrong to claim that this reference applies 2XCO2 models. Instead, it uses 7 different models with fairly moderate or very moderate increases in CO2 concentration, all of which lie below the IP92a scenario. Concerning this reference, Lomborg´s text is simply a lie.

P. 303 bottom left and figure 158: FLAW

"The cost of the Kyoto Protocol is depicted in Figure 158. . . " Flaw: What is depicted in Figure 158 is the gross cost, not the net cost. The amounts shown are the revenues of carbon taxes strong enough to cut carbon emissions to the desired level. It is assumed in the models that these revenues are then recycled into the economy in a "lump-sum fashion" whereby general reductions of all taxes e.g. on industries or private households are reduced with corresponding amounts. Thereby, the net cost e.g. for industry will be lower than the gross cost.

P. 304 left: FLAW

"The OECD estimates that the cost in 2050 will be around 2 percent of the OECD countries´ GDP . . " Flaw: The source for this is a working paper from 1994, describing what the OECD´s GREEN model can be used for in this context. Later runs of the model (Burniaux, 2000) have included for example technological advances concerning renewable resources, and they lead to the result that the cost of Kyoto Protocol could be as little as 0.13 % of GDP.

P. 305 bottom right: FLAW

". . they have all produced more or less the same results." Flaw: This is not true. In fact, estimates for the direct benefits of reducing greenhouse gas emissions levels range from as little as $5 to as much as $125 per ton in 1990 US dollars. See BJ. P. Bruce et al.(1996): Climate Change 1995; Economic and Social Dimensions of Climate Change [here cited from Kysar 2003].

P. 306 top left, note 2579: ERROR

". . . this model gives the same qualitative conclusions as all other integrated assessment models." In the note: ". . . as they do not look at the optimal course, and the economic models seem non-optimized . . . " Error: Lomborg postulates that any integrated assessment model (IAM) gives the same result as Nordhaus´  RICE model. But then he has the problem that models that were run by teams associated with IPCC have given very different results. He therefore seeks a reason to leave these latter models out of consideration. He does that by saying that they cannot be true IAMs, because they do not follow an optimization procedure.
    However, in the source referred to (IPCC2001c), it is evident that the models there are IAMs. They investigate how mitigation costs vary with the pathway to stabilization, and it is said directly that they identify a least-cost mitigation pathway (p. 545 in the source) and that the models impose a long-term cost minimization (p. 547). These model runs are also dealt with in another article: T. Morita et al. (2000): Environmental economics and policy studies 3(2). Here it is said specifically for some of the models that they are dynamic-optimization type models  (p. 80). So Lomborg´s reason to disregard these models is not valid.
    One of the models, called MARIA, which is a dynamic-optimization type model, and a few other models suggest that mitigation policies could lead to some unknown improvement of technologies that would not occur without mitigation, and that these technologies would reduce the energy system costs in the long term.  That is, the mitigation policy could give a net benefit to GDP growth. In Lomborg´s view, this suggests that there was unfulfilled optimization potential in the basic scenario, and this hidden possibility for a benefit should have been accounted for as pre-existing before the mitigation policy. This is an odd argumentation. It is generally acknowledged that technology developed for warfare or space travels may later turn out to have non-anticipated benefits to civil society. Nobody claims that because the possibility for these benefits existed already before the R&D investment in equipment for warfare or space travel, it is wrong to say that these are positive spin-offs from such endeavors.
    It should be noted that the model runs presented in IPCC2001c are used for a cost-effectiveness study, whereas Lomborg favours cost-benefit studies. The difference is that in cost-effectiveness, the starting point is to define a threshold beyond which damages from climate change become so large and unpredictable that they are unacceptable. Therefore, a stabilization target is defined, and the cheapest way to reach this target is found. The costs are, in most runs, a reduction in world GDP of between 0 and 3 %, as stated by Lomborg in his note. He forgets to mention that the actual costs are projected to be less than this, because the model runs do not include the economic effects of introducing new low-emission technologies, new revenue rising instruments or adequate international technology transfers. In any case, with costs of 0 - 3 % of GDP, the costs will more or less be comparable to the benefits, because avoided climate changes could easily amount to 0 - 3 % of GDP. For instance, one of Lomborg´s sources, Chakravorty et al. (1997), write: "According to detailed cost estimates of Cline (1992) and Nordhaus (1992), a 3-degree warming would cause [a damage of 1-2 % of global GDP]." So the cost-effectiveness calculations may be used to get a rough impression of cost/benefit ratios. The costs result inter alia from carbon taxes, typically at a level of about $100 per ton carbon. There may be an approximate cost/benefit balance at this level of carbon taxation, which is much higher than the  level of $7.5 / ton carbon which according to Lomborg´s Figure 160 and page 308 is the probable cost of an extra ton carbon in the atmosphere. Thus, the results form the RICE model differ widely from results of other types of model calculations. This may be the reason why Lomborg hides the crucial results from IPCC2001c away in a long note that few readers will study.

P. 306 top left, note 2579: (INCONSISTENCY)

In note 2579, Lomborg briefly refers to the economic calculations made in the IPCC mitigation report. (Flaw): His reason for not including these calculations in the main text is that they are allegedly not based on integrated assessment models, as explained above. The only policy that Lomborg accepts is that the costs of the greenhouse effect are given a price, and then the market will automatically adjust and find the optimal path. Thus, he feels able to reject the economic evaluation of IPCC´s so-called A1T scenario where renewable energy plays a dominant role, because IPCC writes that especially for this scenario, stabilization of carbon emissions lead to an increase in GDP compared to the baseline, due to apparent positive economic feedbacks of technology development and transfer. In other words, when it turns out under certain circumstances that it may actually pay to reduce carbon emissions, then this is used by Lomborg as a reason to exclude policies that deliberately favour such a scenario ! In this way, of course, he can make sure that a politically decided reduction of emissions will never pay.

P. 306 top left, note 2579: (CORRECTION)

Correction: Twice in the note, Lomborg refers to section 8.4.1.2. in the IPCC mitigation report. That section does not exist. The correct designation is section 8.4.3.

P. 306 left: FLAW OF OMISSION

"The advantage to these models is that they take into account both the costs and benefits . . " Flaw: The RICE model, as well as other similar models, fails to model greenhouse gases other than CO2. This omission is significant because these other gases - halocarbons, methane, and nitrous oxide - have been responsible for approximately 40 percent of the human-induced temperature change up to now, according to Lomborg´s own presentation in his Figs. 132 and 139. Most of these gases appear cheaper to control than CO2. Calculations that omit regulations of what causes 40 % of the greenhouse effect will of course give results that are far from the actual situation.

P. 306 left and figure 160: FLAW

". . the point is that the more CO2 we try to cut, the more expensive it becomes . . ". Flaw: In Lomborg´s text, there is no explanation of why it is costly to society to reduce carbon emissions. Actually, in the model used by Nordhaus and similar models, the only technological change that occurs is a pre-defined annual increase in energy efficiency. Apart from this, the economy can respond to carbon constraints only by switching between existing energy technologies. By and large, the model allows only one method of reducing carbon emissions, namely to cut the amount of energy used and hence to cut the amount of goods produced. It is therefore no wonder that in this model the reduction in emissions becomes so costly to society. As the precise size of the costs to society are so crucial to Lomborg´s argumentation, it is a flaw that he does not give a single hint as to how these costs are calculated, and what is included. It may be added that estimates of abatement costs of proposed environmental regulations historically have tended to be grossly overestimated (P. 268 in D. E. Kysar (2003): Ecology Law Quarterly 30: 223-278).

P. 306 bottom left: INCONSISTENCY

" . . the question is to establish what kind of damage the emitted CO2 will have on society . . on into the twenty-fourth century." Flaw: If Lomborg supports this as a relevant time horizon, why has he in all the preceding pages looked only at effects up to the year 2100 ?

P. 306 bottom left-top right: FLAW

"The model takes into account damage from . . human health and life . . " Flaw: In fact, Nordhaus and Boyer write on their page 80: "In the absence of systematic estimates of health impacts, we have relied on estimates based on the current prevalence of climate-related diseases." From Murray and Lopez (1996) they estimate the number of years of life that will be lost due to climate-related diseases, e.g. when the range of malaria expands. But they do not include any other sources of mortality, although IPCC reports with high confidence that there will be extra deaths due to more frequent heat waves, an increase in drowning, diarrhoeal and respiratory deaths and illnesses due to more frequent flooding, and in increase in malnutrition in those regions of the world where crop yields will be reduced.

P. 306 top right: ERROR OF OMISSION

"Converted to their present value . . " Error: At his point, Lomborg completely fails to explain how the loss of human lives is transformed into an economic value. He should have done this, because the method of calculation will probably to many people not appear acceptable. Nordhaus and Boyer explain that to put an economic value on the number of years lost, they assume that each year lost is worth two years of per capita income. But as the yearly income is different in the 13 different economic regions of the RICE model, the loss of a human life is not given the same value everywhere. In the USA, a life year lost is worth $ 68,200, whereas in Subsaharan Africa, a life year lost is worth $ 940. Now, because Subsaharan Africa will be severely affected by global warming, seventy percent of all life years lost are expected to happen there. But because these lives count with a weight of only 1.4 % of the weight of an American life, the many deaths in Africa will hardly be felt in the total balance. Grossly speaking, what Nordhaus and Boyer do, is to say: "Most deaths will occur in Africa, and as we don´t bother much about African lives, these effects are not very important." Thus, it is very debatable if the method of calculation is ethically acceptable. Lomborg avoids to mention this, and simply says that the value is $7.5. Such an omission is so serious that it is here counted as an error.

P. 306 top right: (COMMENT)

"Converted to their present value, the total and long-term damages from emitting an extra ton of carbon today is the equivalent of $7.5." Comment: This value, of course, depends very much on the rate of discount, which is established by a subjective choice. Therefore, the $7.5 is no objective measure of the costs.

P. 308 left: ERROR

". . . much academic analysis over the last decade has conclusively demonstrated that the strong double dividend is incorrect" Error: This is wrong, and Lomborg knows that it is wrong, because he has read the relevant section of chapter 8 in the source that he calls IPCC 2001c. Here one reads for instance in section 8.2.2.1.2: "While studies conclude that the swap between carbon and payroll taxes [i.e. the recycling of carbon tax revenue to reduce payroll taxes] . . . does not avoid net welfare losses in the USA . . . , a strong double dividend often occurs in Europe. As suggested by theoretical analyses . . , these differences can be explained by the differences both in the taxation systems and in the rigidites of the labour markets. Capros et al. . . demonstrate that the increase of employment in the EU countries due to payroll tax reduction is far higher under the assumption of wage rigidities than under the assumption of a classic flexible labour market. . . " This text is supplemented by a figure (their Figure 8.5) that gives an overview of what happens to the welfare of households or the whole society in European countries if  the revenue of a carbon tax is recycled to reduce `distortionary taxes´. In seven studies, there is a net loss to welfare (i.e. no strong double dividend) but in twenty seven studies, including studies made in Lomborg´s home country, Denmark, there is a net gain to welfare, i.e. a strong double dividend. The differences between results concerning the American and European economies are explained as follows in the text: "These discrepancies arise because labour taxes represent one of the most important sources of distortion in European countries as a result of the pre-existing tax structure and of the type of labour-market regulation that prevails in these countries."
Lomborg has read and seen this, but nevertheless he writes that the double dividend is incorrect. He even adds the word "conclusively" as if to signify that no doubt or ambiguousness remains. He knows that this is not so, and therefore his text is deliberately misleading.

P. 308 right and note 2600: (COMMENT)

"And the new environmental tax itself is distortionary as it discourages work effort . . " Comment: What is postulated here, is that if some sort of environmental ("green") tax is introduced, the costs of living for the workers are increased, i.e. they have less money left for consumption; this corresponds to a reduction in their "real wage", and this wage reduction will cause more persons to leave the labour market. This tax-interaction effect is postulated by those American economists referred to by Lomborg in his note 2600. This may appear sensible to certain economists in USA with its competitive labour market. But in European countries, with more regulated labour markets, it makes no sense. Experiencies from Lomborg´s home country, Denmark, are not that green taxes reduce the employment rate. If a worker gets less money left for consumption due to green taxes, he will probably work more, to compensate for the loss, and not less.

P. 308 right and note 2600: FLAW

"Unfortunately, most analytical and computable models show that the tax-interaction effect is bigger than the revenue-recycling effect." Flaw: This is not generally true; it is true for studies of the US economy, but most often not for studies of European economies. See the other comments above and below.

P. 308 bottom right: FLAW

". . simply because higher taxes will all in all lead to lower welfare." Flaw: Such a political statement should not be made by a person who allegedly just presents the facts and leaves the political judgment to the individual reader (cf. p. 5). It sounds more like the statement of a conservative economist employed by some American oil company to secure its short term interests. The idea seems to be that taxes will hamper investments and hence economic growth, and that economic growth is the only way in which welfare can increase. Even if we accept the importance of economic growth, this is wrong. For instance, the revenue from taxes could be used by the state to support development of technological solutions in the field of sustainable energy sources. This could bring about much needed improvements in wind and solar energy technology. As technological advance is a major impetus to economic growth, such investments might actually increase long term welfare.

P. 309 top right and note 2603: ERROR

"The IPCC generally accepts that there is no strong double dividend. . ." Error: This is wrong. The footnote has two quotes. The first is from chapter 7 of IPCC2001c, and it says that a strong double dividend presupposes that the original tax structure is seriously inefficient which in itself calls for a tax reform. However, this is written from an American perspective. In Europe, the taxation system is generally of a type which to an American liberalistic economist calls for a tax reform. The second quote is from chapter 8 of IPCC2001c. This quote is deliberately misleading, because Lomborg omits the crucial following sentences. The original text is " . . . is roughly zero. Thus, in some analyses the sum is positive, while in others it is negative. In economies with an especially distortive tax system (as in several European analyses), the sum may be positive and hence confirm the strong double-dividend hypothesis. IN economies with fewer distortions, such as in various models of the US economy the sum is negative. Another conclusion is that even with no strong double-dividend effect, a country fares considerably better with a revenue-recycling policy than with one that is not revenue-recycling, like grandfathered quotas."

P. 309 right: FLAW

". . today we know that the strong double dividend hypothesis is typically invalid." Flaw: Lomborg gives only one reference to this statement, which could hardly substantiate his claim that "we" know. Actually, the concept of the double dividend remains a part of mainstream environmental economics (see Jespersen, chapter 1 in Sceptical Questions and Sustainable Answers). Lomborg declares that any extra tax is negative (see comment for p. 308 bottom right). But the green tax is not an extra tax - it substitutes another tax and leaves the total tax pressure constant. So the negative effects on income and investments have already been accounted for and should not be double counted. The whole box on p. 308-309 is therefore faulty.

P. 309 right and note 2607: FLAW

". . because empirically, most carbon taxes have specifically not been recycled as reductions in the most distorting taxes . . ." Flaw: Lomborg gives only one reference to this, and this reference is clearly about a special trend for earmarking green taxes which is common in USA, but not in Europe. It is therefore misleading to say "most carbon taxes". 

P. 310, fig. 163: FLAW

Flaw: The size of the columns, and the adjoining figures, do not correspond to those in Nordhaus & Boyer. Especially the column for the 1.5 degree situation has been increased (by about $ 7 trillion). Lomborg does not indicate what might allow him to change the original figures.

P. 310 left: FLAW

"It shows that business-as-usual will present society with a total, one-time cost of $4,820 billion (rounded to $ 5 trillion)." Flaw: This statement is extremely flawed, because the figure with 4 digits implies a great certainty, while reality is that the figure could be anywhere from gains to catastrophic losses. Furthermore, the result is completely dependent on the choice of discount rate, which according to subjective judgments and priorities could be anywhere from 0 to 6 %. This flaw is extra important because Lomborg´s whole argument for not putting restrictions on oil consumption relies completely on the validity of this non-figure. A person who can cite such a figure without any indication of the degree of uncertainty whatsoever cannot call himself a statistician. Furthermore, Lomborg gives no reference to his figure. It does not tally with those found in the sources given on page 310. However, it seems that the source of the figure is as follows (as explained by Anders C. Hansen in Sceptical Questions and Sustainable Answers): Nordhaus & Boyer have calculated the net benefit of following a so-called geo-engineering strategy, which consists in neutralizing the heating due to greenhouse gases by injecting a layer of particles to the atmosphere that reject sunlight. According to Nordhaus & Boyer, the cost of geo-engineering is nil and has no adverse affects. The economic damage from global warming thus saved would amount to $3,901 billion. Converted to year 2000 prices that would be $ 4,820 billion. In Lomborg´s version, this is the cost of the business-as-usual approach. He just forgets to tell us that this price relates to a situation where the sunshine is always hazy. On p. 301 right, Lomborg cites estimates that the cost of the damage caused by global warming will be $ 480-640 billion a year. If he had recalculated this to yield a total present value, it would amount to $ 20-27 trillion. We cannot say whether this figure is more or less correct than the alternative figure of $ 5 trillion. It just illustrates the uncertainty. But the point is that with the higher figure, which is no less likely, Lomborg would obtain precisely the opposite conclusion from his cost-benefit ananlysis: It would more than pay to control global warming.

P. 305-312: FLAW OF OMISSION

Lomborg discusses the economic losses due to reductions in energy consumption in terms of trillions of dollars or % of GDP. Flaw: First, Lomborg forgets what he has read on p. 1204 in the paper by Chakravorty et al. (1997): "According to detailed cost estimates of Cline (1992) and Nordhaus (1992), a 3-degree warming would cause [a damage of 1-2 % of global GDP]. Given the considerable uncertainty . . even these authors admit that the numbers are rather speculative." So, it is misleading to draw a graph and read an intersect as if the figures were fairly exact. Second, Lomborg does not make it very clear to the reader how these so-called losses are calculated. It should be stated clearly that these losses are imaginary, because the amounts given are not to be subtracted from present GDP, but from some unknown future GDP. For instance, if we talk of a 2 % reduction in GDP, this may actually mean that a projected rise in GDP will not be 100 % relative to now, but only 98 %. Considering that we cannot predict the GDP in 100 years from now to a precision of less than 2 % (nor to a precision of 10 or 30 %), this means that the signal is much less than the noise in these calculations. To illustrate the size order of the so-called costs, we may here cite IPCC (2001c), section 8.4.3 (which Lomborg has read): "The annual 1990-2100 GDP growth rate across all the stabilization scenarios was reduced on average by only 0.003 % per year, with a maximum reduction reaching 0.06 % per year." Is that such a disastrous cutback in economic growth that we would rather prefer to live with the risk of climate change ? - What Lomborg does here is to restrict himself to the so-called top-down economic models. These models are built on the existing structure of society, e.g. the present transport systems, present consumer preferences, and present energy efficiency. On this basis, they calculate deviations from the projected course of development. The is also another option - to use the so-called bottom-up models. In bottom-up models, you ask questions like: How much can we expand production with only x % rise in energy consumption ? That is, you focus on technological solutions. As it turns out, such models lead to much more creative results, and they point in the direction of large increases in welfare with only modest increases in energy consumption. This is not to imply that bottom-up models are better or more useful than top-down models. The point is that Lomborg discusses only the latter, not the former. Or, more precisely, he accepts top-down models without reservations, at face value, whereas he does all he can to derogate the idea of bottom-up models.

P. 312-313: FLAW

Lomborg has a box on the so-called no-regret options. Flaw: Lomborg downplays the potential of the no-regret options, and exaggerates the costs of developing new, cleaner technologies. He admits that it might be possible to reduce energy consumption by up to 5 % without any consequences to material welfare - a very modest admission, considering that the total energy consumption of Danish households was reduced by 29 % from 1972 to 1990, without any decline in material welfare. If just the 5 % reduction admitted by Lomborg were utilized, it would have consequences for his figure 160. Here, the black curve should start at 5 percent reduction, rather than at zero or 1 percent, which would change the conclusion considerably. However, even though he admits this possibility, he does not modify his graph accordingly. Instead, he feels able to disregard the no-regret options altogether. But what, then, are the no-regret options that Lomborg disregards ? First, of course, it includes small everyday details, such as using ordinary light bulbs where we could just as well use energy-saving bulbs. Second, there are large scale options. For instance, the International Energy Agency (IEA) calculated in 1999 that 8 of the most energy-consuming countries outside the OECD are subsidizing energy consumption to such an extent that average energy prices are cut by 20 %. Abolishing these subsidies would reduce the world´s total energy consumption by 3.5 %, and the world´s total CO2 emissions by 4.6 %. By doing this the GDP of these countries would be improved by 1 % (IEA (1999): World Energy Outlook 1999. Insights. Looking at energy subsidies: Getting the prices right. OECD/IEA). Also, within the OECD countries, there is much room for more energy efficient production without reducing output. In countries such as the USA and Australia, energy consumption and associated emissions could be cut by 20-40 % at no net cost - more than enough to meet Kyoto targets (ref.: Clive Hamilton & Hal Turton, the Australia Institute). According to IPCC, working group III, 1996, the global potential for CO2 reduction without costs to production are 20 % now, 40 % in 2020 and maybe 70 % in 2030. Even if such figures are very uncertain, it is not honest to disregard them altogether.

P. 314 right: ERROR

"We expect that in 2035 the average American will be twice as rich as she is now . . . Both these arguments indicate that it is probably reasonable to have a discount rate of at least 4-6 percent." Error: A doubling in 35 years does not equivalate a rate of increase of 4-6 percent per year, but only 2.0 percent per year. Furthermore, the rate of increase in Lomborg´s Figure 149 is not quite that high. Taking the figure for the developed countries (the OECD countries) in the lower right hand figure, and using scenario B1, which up to 2035 is pretty much an average scenario, we find a growth rate of 1.5 % per capita up to 2035, and about 1.0 % thereafter (the original data are in Lomborg´s source IPCC 2000b). For Lomborg´s purpose, it would be more relevant to look at the growth rate for the whole world. In the B1 scenario, up to 2040, this growth rate is 2.5 percent per capita, or 3.4 percent if we consider the growth in total GDP. The rates gradually decline during the 21st century (which seems realistic), and on average for the whole century, they are 2.26 percent per capita, or 2.53 percent for total growth. - Therefore, the evidence presented here by Lomborg is no argument for choosing a rate of discount between 4 and 6 percent. If Lomborg wants to find a well-founded rate of discount for the period 2000 to 2100, he could choose any value between 1.0 percent and 2.53 percent, depending on what growth is indicated, but higher rates are unwarranted.
    The concept of discount rates is explained at length on this page of Lomborg-errors.  I
    One may remark how strange it is that one of the most crucial details of the whole book - the rate of discount - is so weakly founded, or rather, not founded at all. 

P. 314 right: FLAW

"Both these arguments indicate that it is probably reasonable to have a discount rate of at least 4-6 percent." Flaw: The discount rate is a subjective measure which expresses the personal view of how much weight the future should have relative to the present. The concept of discount rates is explained at length on this page of Lomborg-errors. As explained there, there are two approaches to setting the discount rate - the prescriptive and the descriptive approach. In the descriptive approach, you consider what rates of return people can obtain when they invest money in alternative projects. If you set the rate of discount in relation to that, you will end up with rather high rates, probably in the range 4-6 percent. However, this is much higher than the actual rate of increase in material welfare. Thus, the American economist W. Nordhaus, whom Lomborg supports, uses an estimate of about 1.3 % as the annual rate of increase of material welfare, whereas the social rate of discount is set by him at a level of 6 %, declining gradually over the years to 3 %. Setting the rate of discount so much higher than the rate of increase of material welfare, means that you accept an investment only if it gives much higher growth than seen in the rest of society. This  may be possible on a short time scale - within up to 30 years. But on longer time scales, the only acceptable method to set the discount rate is to use the prescriptive approach, where the rate is set in relation to the rate of increase of material welfare. That gives rates approximately in the ragne 1 to 2 %. If you consider time spans of a generation or more, then discount rates as high as 4 to 6 percent are unacceptable.
    If the discount rate is set by the descriptive approach, then it does not express our ability, with our material welfare, energy resources etc., to cope with larger and larger problems in the future relative to what we can afford to cope with today. Rather, it simply expresses the importance to us of the problems that future generations will face. If the discount rate is set as high as 5 %, this leads to absurd results. For instance, as explained by Dubgaard (chapter 10 in Sceptical Questions and Sustainable Answers), the present value of all real estate in Denmark is worth $238 billion. If sea levels rose so much that the whole of Denmark would have to be abandoned in 500 years, this would be no problem, because $238 billion, discounted 500 years back to its present value, would amount to $6, i.e. the equivalent of half a barbecued chicken with potato fritters. Even if the total value of Denmark had for example increased by a factor of 100 during the years, we would still be able to afford the equivalent of 100 half chickens. A more reasonable indication of where we should set the rate of discount would be 2 to 3 % for discount within generations, and 0.5 % for discount between generations (Dubgaard, loc. cit.). Chakravorty et al. (1997), on whom Lomborg relies on pp.285-287, use a rate of 2 %. Five percent, on the other hand, is equivalent to saying "To the hell with the future".

P. 315 top left: FLAW

" . . . projects that make a profit of more than 5 percent . . . although the 2 percent investments were more future oriented, they would in all probability leave our children and grandchildren with far fewer resources." Flaw: Lomborg mixes up things here.  Investments that make a yearly profit of 5 percent do not lead to a growth in resources of 5 percent per year. As already stated by Lomborg on page 314, the wealth of the average American will double in 35 years, corresponding to a growth rate of only 2 % per year - even though short-term investments give returns much higher than that.

P. 315 bottom - 316 top: FLAW

" . . a local warming of larger than 3° C, if sustained for millennia . . " Flaw: Once the concentration of greenhouse gases has risen, and the temperature has therefore also risen, it will last long before it drops again. According to IPCC, working group I, 2001, section 11.5.4.3 and figure 11.16, a scenario where the increase in greenhouse gases peaks at 2130 will, as the average of the range of possible climatic outcomes, cause the temperature at Greenland to stabilize from about 2130 onwards at about 5.5° above present. (At high latitudes the warming is greater than the global average). This would cause melting of inland ice on Greenland, which in turn would cause ocean levels to increase by 3 m in the course of 1000 years as a delayed effect. This is what happens if the greenhouse gas cencentration does not increase after 2130, i.e. approximately under the conditions that Lomborg refers to. Of course, once the temperature has increased by 5.5° C, and the ice and the ocean have been warmed up accordingly, the temperatures cannot suddenly drop again - for example by 2200 - as Lomborg suggests. Lomborg has read those parts of the IPCC report referred to above, but has not understood, or has not wanted to understand, what it says.

P. 317 left: ERROR

"The central climate sensitivity of 1.5-4.5° C has not changed over the past 25 years, indicating a fundamental lack of model adequacy. ." Error: This is not true. The US National Research Council report on global warming in 1979 was the first to offer the 1.5-4.5° C equilibrium warming range for doubled atmospheric CO2 concentrations. In that report, this range was determined by a simple one-standard-deviation analysis (two out of three chances it is within this range). Today, the same range is quoted, but now the chance of it lying in this range is roughly nine out of ten.

P. 317 bottom right: FLAW

" . . the total cost is about $5 trillion." Flaw: As explained in a comment on p. 310 left, this amount is fictitious and could just as well be anything else. To use the word "is" in the above sentence is greatly misleading.

P. 318 left: (COMMENT)

"" . . the effect of Kyoto on the climate will be minuscule . . ". Comment: Everybody knows that the intention with the Kyoto protocol is that it should just be a first step in a longer series of steps, and a text that disregards this aspect does not deal with the real issues.

P. 319 bottom - 320: FLAW

"Thus, climate policy may be used as a tool and a justification for charting an alternative course of development . . " Flaw: Lomborg thinks he has caught IPCC attempting to use the Third Assessment Report as a Trojan horse for promoting irrelevant political messages. He uses no less than half a page to cite IPCC as having made statements such as "we should build cars and trains with lower top speed . . we need to change individual lifestyles, and move away from consumption". However, those citations are not IPCC´s own, but only distorted versions of citations from the literature that IPCC has been set to review. IPCC has not proposed a single one of the things Lomborg claims the panel to have suggested.

P. 323- 324: INCONSISTENCIES

323 left: " . . the price will easily be 2 percent or more of world GDP per year. " 323 right: "Even stabilization of emissions will "only" cost 0.4 percent more of our overall wealth. . " 324 left: " . . 12 percent of the total . . " 324 right: "premium against global warming (at 2-4 percent of GDP)."  Flaw: The various estimates of what percentage of world GDP  will be lost due to climate mitigation differ, and there are no explanations of how these discrepancies arise.

P. 323 bottom right, Figure 166 and note 2738: (COMMENT)

"" . . the total income will be some $900 trillion." Comment: This is not a sum that one may found in the source. Rather, it is a sum calculated by Lomborg by discounting future costs at 7 % p.a.

P. 323 - 324: FLAW

"However, should we go down a path focusing on the enviornment . .  humanity will lose some $107 trillion . . . " Flaw: The reader will easily misunderstand this and believe that $107 trillion is the price for fighting climate change. This is not so. The difference between the A1 and the B1 scenario do not include costs of climate mitigation. The different concern for the environment regards non-climate issues, whose solutions according to Lomborg (p. 324 right) amount to $18 trillion. So Lomborg mixes up unrelated figures. See more on this page.

P. 324, Figure 166, note 2738: INCONSISTENCY

"The discount parameter used here is 7 percent. " Flaw: Lomborg argues (p. 314) that the discount rate should be at least 4 - 6 %, and Nordhaus uses a rate of 6 %, gradually declining to 3 % with time. There is no reason why Lomborg should then here use a rate of 7 %. One may suspect the actual reason to be that only thereby will the end result - $895 trillion - be sufficiently near to Nordhaus´ figure of $945 trillion.  This then makes the basic figures more similar than they actually are, and the extreme uncertainty on the figures is downplayed.