LOMBORG-ERRORS                       Home
                        List of chapters
Chapter 10:

Forests - are we losing them?




The chapter on forests is one of the most biased chapters in Lomborg┤s book. The whole text aims at making forest losses appear as small as possible, and Lomborg tries to make the reader believe that the total forest area in the world is constant or slightly rising, when in fact it is clearly declining. FAO recently published an authoritative overview of what is known on this subject. Although Lomborg has read this report, he repeatedly advances postulates that are directly contrary to what the report says.

Recommended references:

FAO┤s report from 2001: The global forest resources assessment 2000 (updated version relative to what was available in 2001). To be downloaded here.

The book "Sceptical questions and sustainable answers" has chapters on deforestation by Kňre Fog and Thorkil Casse. It may be downloaded here.


P. 110 right and p. 111 left: ERROR

"Globally, the overall area covered by forest has not changed much since 1950, as can be seen in Figure 60" . . . "Globally, forest cover has remained remarkably stable over the second half of the twentieth century. With the longest data series, global forest cover increased from 30.04 percent of the global land area in 1950 to 30.89 percent in 1994 . ." Error: "The longest data series" is the data series from the FAO Production yearbook. This series cannot be used as evidence for an increase in forest cover. Firstly, FAO states in the introduction of each yearbook: " It should be borne in mind that definitions used by reporting countries vary considerably and items classified under the same category often relate to greatly differing kinds of land . . . Thus, the area specified is not intended to refer to or delineate `forest coverage┤ ". When Lomborg uses the data, he clearly violates this reservation. Secondly, Lomborg does not adjust for inclusion of more and more countries and for revisions in particular countries. Thus, the "increase" from 1958 to 1959 is mainly due to a changed conception of what is designated as forest in Canada plus the inclusion of Mongolia which had not been included before, and the "increase" from 1960 to 1961 is due to the inclusion of the eastern part of New Guinea which had not been included before. With the data from 1971 onwards, the FAO yearbooks bring revised data where such irrelevant spurious changes are compensated for. With this revision, all its data from 1971 onwards show consistenly a yearly decline in global forest cover, in most periods a decline of c. 0.2 % per year, which agrees remarkably well with later estimates. It is evident from Lomborg┤s text that he is aware that these revisions have been made. This is evident from note 770, p. 375 right, where he cites two different estimates of the 1976 forest area. Thus, Lomborg has indeed studied the later revisions which show that the forest area is actually declining. In spite of this, he does not use them. Instead, he uses the unrevised figures, even though FAO itself states that the figures cannot be used for Lomborg┤s purpose. When Lomborg neglects that, it must be a case of deliberate misleading.

P. 111, legend to figure 60: FLAW

The figure shows "forest and woodland" 1948-94 and more restrictive definitions of forest from 1980 onwards. In note 767 , the newer definitions of forest are explained very briefly, but there is no explanation what is understood by "woodland". In note 770 he writes that "forest and woodland" counts everything with regular tree trunks. Flaw: Actually, however, FAO┤s definition of "other wooded land" comprises not only areas with regular tree trunks down to where only 5 % of the area is covered by canopies, but also areas with small trees, bushes or scrub, as long as this vegetation covers at least 10 % of the area. However, nowhere in Lomborg┤s text can the reader see that even bush steppe, maquis or savanna is included in the "forest area", and nowhere can he read that if forest is degraded and turned into scrub or bush steppe, this will not show up in the statistics. Right from the beginning in January 1998, this flaw was pointed out to Lomborg, but he has consistently refused to change his presentation. He continues to hide from the reader what the upper curves in fig. 60 actually show. If he had explained this to his reader, it would have destroyed his whole claim that forest area is increasing. Considering that it has been pointed out to him that his presentation is misleading on this point, it is remarkable that his text is still identical to what it was in 1998. The conclusion from this is that the presentation is deliberately misleading.

P. 111 right: ERROR OF BIAS

"Thus, with these considerable short-term uncertainties it seems necessary to focus on the longest possible time periods." Also in note 770: "It has therefore been important to employ the longest time-span series available, and FAO┤s long series from 1950 is the only one available. " Error: This argumentation is wrong. The long data series does not just eliminate short-term fluctuations. As time goes by, it includes more and more countries and areas, and therefore there will be a systematic bias in direction of greater forest areas in later years. Remarkably, however, at the end of note 767 Lomborg writes that the data for forest area in 1994 in Russia is clearly too low, and he has himself made an upwards correction. Thus, when spurious variation causes the data to go in the direction that Lomborg wants, he uses un-revised data. But when spurious variation causes the data to go in the opposite direction, he does revise them. This is an unacceptable bias.

P. 112 left: FLAW

Lomborg has read the FAO report from March 2001: "The global forest resources assessment 2000. Summary report." This report clearly states, on the basis of more thorough investigations than before, that the global forest area has been declining at a rate of 0.2 % per year as far back as the data go (Table 1 in the report). On p. 112, Lomborg just briefly mentions this report, stating that here FAO has once again changed its forest definition, and made a new estimate showing a small decline. Flaw: The 2001 FAO report is the most recent, thorough, reliable and authoritative source to date on deforestation. When this new evidence appears and contradicts what Lomborg has said until then, he should change his conclusions. He does not. He still sticks to older yearbooks with data that are notoriously unreliable, and only briefly states that the new evidence points to "a small decline". This presentation of the evidence is so distorted that it is deliberately misleading.

P. 112 right: ERROR

On regional forest loss: "Southeast Asia . . has only lost 7 percent . . " Lomborg has this information from chapter 10 in the book by B.L.Turner et al. (1990): The earth as transformed by human action. Error: Although chapter 10 of this book does have this information in its table 1 on global land use, the figure can impossibly be true. Chapter 10 is not on forests; it would have been more reasonable to take the information from chapter 11 in the same book (Williams 1990), which is specifically on forests. It is evident from table 11-5 in this book that forest loss in Burma and Malaysia alone from 1880 to 1980 is greater than Lomborg┤s alleged total forest loss in southeast Asia during the last 300 years. Where Lomborg indicates a loss of 7 %, several other lines of evidence actually point to a loss of somewhere around 50 %. However, this error seems to have been involuntary; many other figures in the table that Lomborg has cited are clearly wrong, and Lomborg has not systematically used those data that err to the low side.

P. 112 right and p. 117 right: FLAW

"Globally it is estimated that we have lost a total of about 20 percent of the original forest cover since the dawn of agriculture." (p. 112). "In a historical perspective, about 20 percent of all forest has been lost." (p. 117). Flaw: The estimate of 20 percent is very uncertain, and relies mostly on the estimate of how large was the original forest cover. Lomborg gives the impression that 20 % is the best estimate of the experts, and downplays the uncertainty associated with this figure. For instance, Goudie (1993) does not indicate a precise figure; he writes: "since preagricultural times world forests have declined approximately one fifth". And the references cited by Lomborg are not especially reliable, e.g. the data from table 10-1 in Turner et al (1990), referred to above. - Furthermore, it is unlikely that the figure should be as low as 20 %. For instance, Goudie┤s figure of one fifth is taken from the WRI (1991), whereas WRI in later reports has increased this figure. A report by the WRI, "World resources 1994-95", gives much higher figures: about 49 % for the tropics, and a figure at least as high as that for the temperate regions. To this must be added the effect of continued deforestation since 1994. A newer report from WRI, which Lomborg has also seen (World Resources 2000-2001), has the following text: "Using this approach, Matthews (1983 . . ) estimated that as of the early 1980s, humans had reduced global forest cover about 16 percent. Updating . . brings the total loss . . to roughly 20 percent. Historical forest loss could be much higher, however. A 1997 study by WRI, which used a higher resolution map of potential forest than the Matthews study, estimates that original forest cover has been reduced by nearly 50 percent." As Lomborg has seen these reports, it would have been fair of him to indicate a range of estimates from a low of 20 percent up to about 50 %, with 50 % being the newest estimate. He has chosen not to do this, and his text is therefore deliberately misleading.

P. 112 right and note 789: FLAW

"Globally it is estimated that we have lost a total of about 20 percent of the original forest cover since the dawn of agriculture." Flaw: Several of the figures cited by Lomborg in note 789 do not refer to the time since the dawn of agriculture. Richards gives an estimate for the last 300 years, and IPCC gives an estimate just for one hundred years, from 1850 to 1950. However, as Lomborg knows well from his reading of the chapter on forests in Goudie (1993), forest clearance starter much earlier, not only in Europe and the Mediterranean region, but also in Africa (3000 BP), South and Central America (7000 BP), and India and New Guinea (9000 BP). So, when postulating a loss of only 20 %, he neglects all losses that happened more than a few hundred years ago.

P. 112 right: FLAW

Lomborg writes: "The WWF, for example, claims that we have lost two-thirds of all forests since agriculture was introduced . . . , although there is no evidence to support this claim." Flaw: The reader gets a false impression of the vast difference between WWF┤s two thirds and the "true" 20 %. If the high estimate of 50 % had been mentioned, WWF┤s exaggeration would not appear grave. The main text of the WWF press release that Lomborg refers to gives figures that amount to a 62 % loss, which is not very far from 50 %. To this may be added that FAO and WRI may very well have underestimated recent forest loss in some regions. Considering the very large uncertainty about the size of the original forest cover, considering that the loss is likely to be around 50 %, and considering that WWF themselves claim to have made an original, new evaluation which increases the percentage relative to earlier estimates, it is not fair to say that there is no evidence for what WWF claims. And WWF┤s bias to the high side is probably smaller than Lomborg┤s bias to the low side.

P. 113 left: ERROR OF BIAS

"In the late 1970s it was feared . . . " In this paragraph, Lomborg gives the impression that estimates of the tropical deforestation rate have decreased steadily. For instance, Norman Myers┤ estimate of 2 % annual deforestation is contrasted with later, much lower estimates. Error: Myers┤ estimate concerns transformation of primary closed forest to any other formation, whereas the lower estimates cited refer to total clearance of forest. Thus, the figures are not comparable. The rate of disappearance of primary, i.e. undisturbed forest, could be expected to be larger than the rate of total disappearance of forest. In addition, many sources in the early and mid 1980┤es gave estimates for total deforestation of about 0.6 % annually, which is lower than comparable later estimates of 0.7-0.8 % annually. Thus, there has not been a clear trend for lowering the estimates. Lomborg knows this, because it is explained in Williams (1990), which Lomborg has read and cites extensively. It is only by omitting those estimates that do not fit into his pattern that he can claim that estimates have become ever lower. This bias is therefore deliberate.

P. 113 left: FLAW

In the same paragraph as above he cites FAO as having estimated annual deforestation at 0.8 % in the 1980s, falling to 0.7 % in the 1990s, and in the 2001 report the estimate declined even further to 0.46 %. Flaw: The text is ambiguous. It may either be understood in the way that estimates of deforestation have recently fallen, or that the actual rate of deforestation has recently fallen. In note 801 he writes: "The loss of tropical forests is . . ", implying that he talks of actual rates. If the reader understands the text in this way, he will imagine that the actual rates have changed along the lines of 0.8 - 0.7 - 0.46. Lomborg could have prevented this misunderstanding if he had said that the most recent estimates for the 1980s and the 1990s, are 0.8 and 0.7 respectively when we use ordinary methods, and 0.47 and 0.46 respectively when we use satellite imagery. By omitting the figure of 0.47 from the main text, he makes it possible to believe that deforestation is much less now than before. Also, the calculations in the footnote are confusing and incomprehensible when 0.47 is omitted from the main text.

P. 114 right: ERROR

Concerning how much tropical forest has actually disappeared "IUCN estimates that 80 percent of the original forest cover is still in place", i.e. just 20 % has disappeared. Error: When the average forest coverage today in countries that lie wholly within the rain forest climate zone is about 50 %, and when probably about 15 % of all tropical forest has disappeared from 1980 to 2000, it seems impossible that only 20 % should have disappeared since early civilisation. Lomborg has just a single source for this claim, and this is not an official IUCN estimate. It is a rough calculation in a symposium article dealing with a different subject. It is clear from the original source that the calculation is extremely rough in that the remaining amount of all tropical forest, wet as well as dry, is calculated as a percentage of the original amount of wet forest only. This reservation in the original source has been left out by Lomborg. This error must, at best, be due to gross negligence.

P. 114 right, note 812: FLAW

Although it is hardly possible to estimate the original amount of all tropical forest, it is possible from climatological data to estimate the original amount of rainforest, and hence the loss of rainforest, up to now. Note 812 has: "Several sources state that we should have lost more than 50 percent of the rainforest . . Unfortunately, there are no references." Flaw: The sources mentioned are popular texts where you would not expect references. References for the claim of c. 50 % loss do in fact exist, e.g. the WRI report for 1994-1995. Lomborg knows this, because these references were pointed out to him in 1999. By omitting these authoritative sources, and citing only popular texts, he casts undue suspicion on the 50 % estimate. As Lomborg has seen the relevant information, but omits it, his text is deliberately misleading.

P. 115 right: ERROR

"But as we have pointed out, there has not been a fall in global forest area during this period.", i.e. during the last 50 years. Error: As already explained in relation to Fig. 60, Lomborg┤s "long data series" is not reliable, especially not if one uses the unrevised data, as Lomborg does. If, however, in spite of this, we for a moment were to accept this unrevised data series up to 1980, then it would indicate an increase in global forest area of 75 m hectares up to then. From 1980 onwards, we have more reliable data indicating a loss of c. 150 m hectares from 1980 to 2000. Thus, the net trend over the last 50 years is a decrease. Furthermore, Lomborg knows very well from the FAO 2001 report that during the last 20 years net deforestation has been approximately 2/3 of gross deforestation - in other words that increases in forest area, for example due to new forests in Siberia, in no way compensate for losses elsewhere. The expression "No fall during this period" does not seem warranted when there has been a fall at least during the last 40 % of the period.

P. 115 right, 116 top left and 117 right: ERROR

On the net effect of plantations on the amount of natural forest, Lomborg writes p. 115: "Similarly, many allege that although forest cover has remained constant, this is because we have less natural forest and more plantations.", p. 116: " . . they reduce the economic pressure on other natural forests. As a result, these forests are better shielded . . ", and p. 117: "Plantations . . . actually help relieve pressure on natural forest." Error: The text gives no quantitative indication of the postulated "relief" of pressure on natural forest. Plantations are to a great extent eucalyptus and poplar. Many of the products from natural forests cannot be replaced by such wood, and there is no documentation that the establishment of plantations is correlated with less clearing of natural forest. In addition much forest is cleared to make way for agricultural land, and that pressure is not relieved by plantations. Lomborg┤s text on p. 115 is very unclear: he presents somebody else┤s point of view that a constant total forest area is obtained by plantations replacing natural forests; he does not take exception to this concept, only to the concept that this replacement will reduce biodiversity. When he postulates a constant forest area, and when he later concludes that plantations actually help relieve pressure on natural forest, the reader needs to understand that Lomborg agrees with this concept. One could always postulate that if there had been no plantations, deforestation would have been even greater, but as this is pure speculation, Lomborg cannot claim this to be true. Lomborg┤s purported aim is to just present the facts and official figures, and leave it to others to draw subjective conclusions. Now, what are the facts here? According to the FAO 2001 report on the world┤s forests, which Lomborg has read, succesful new plantations amount to approximately 3 m hectares/year. Half of these plantations are replanted on recently cleared forest areas, and the other half are plantations on former agricultural land. Thus, the growth in total forest area due to plantations established on open land is 1.5 m hectares annually, which is little compared to the c. 15 m hectares of forest cleared annually. Furthermore, 1.5 m hectares of socalled natural forest is also cleared annually to give place to plantations. So the "relief" given by plantations in open country is eaten up by plantations in former forest. There is no net relief. The point is that Lomborg has read the FAO report where these data are presented, but still he fails to utilize these data to adjust his unwarranted claim. At best, this is due to gross negligence.

P. 116 left and note 828: (COMMENT)

". . plantations make up just 3 percent of the world┤s forest area." Comment: Notice that according to newer data cited in the note, the figure is actually around 5 percent.

P. 116: ERROR

Lomborg claims that forest fires in the El Ni˝o year 1997 in Indonesia were not especially large. Thus he says: "The independent fire expert Johann Goldammer said that `there is no indication at all that 1997 was an extraordinary fire year for Indonesia og the world at large┤ " and "In conclusion, 1997 was in no way the year in which fire burned more forests than at any other time in history." Error: As stated by T. Lovejoy (Scientific American, Jan. 2002), Lomborg fails to mention that the first official Indonesian estimates of burned area were not in the least credible and later turned out to be far too low. In 1999 the Indonesian government and donor agencies, including the World bank, signed off on a report that the real number was 4.6 m hectares. A paper in Nature (F. Siegert et al. (2001): Nature 414 (6862): 437-440) refers to the same report and says that the fires in 1997 were "the largest fire disaster ever observed" and that "the 1997-98 fires by far surpassed the 1982-83 disaster." In note 835, Lomborg says that he had a personal communication with Goldammer in 1998. How strange then, that the same Goldammer contributed data to the official report from 1999, and later was co-editor of a report which confirms the extent of the fires (J. G. Goldammer & R. W. Mutch (2001): Global forest fire assessment 1990-2000. Working paper 55. FAO, Rome, 200a. Available at www.fao.org.80/forestry/fo/fra/docs/wp55_eng.pdf. In this report, the chapter on Indonesia was written by Goldammer himself. Here he states that from medio 1997 to medio 1998, the area burnt was 9.7 m hectares, of which 4.6 m hectares were forests and a further 0.2 m hectares were plantations. He also writes: "In 1997-1998 Indonesia experienced a fire episode that exceeded the size and impact of the 1982-83 fires (Goldammer et al. 1999 . . . )." So what Goldammer published in 1999 was exactly the opposite of what he allegedly said to Lomborg just one year before, in 1998. How strange.

Should Lomborg have known that he was wrong ? Yes. Firstly, before making his 2001 edition of his book, it would have been natural for him to check on the internet if there were any new data from Goldammer; in that case, he would have found his report. Second, we know that Lomborg had read the 2001 FAO report on the world┤s forests: this report states that the fires in 1997-98 were unusually extensive, and that "Fires were widespread in Indonesia in 1999 and 2000, but not on a scale comparable to 1997-98." Knowing this, it was inadmissible for Lomborg to write that 1997 was in no way the year in which fire burned more forests than usual, without checking this. And even if Lomborg were to dispute - for whatever reason - the most reliable estimates of 4.6 m hectares, his postulate that WWF has exaggerated the extent of the fires cannot possibly be retained. The misleading text on Indonesian forest firests is at least due to gross negligence, and most likely, considering the role of Goldammer, deliberate.

P. 116 right: ERROR

Lomborg postulates that the fires in 1997 in Brazil were not especially severe, or were mostly on already burnt land. And he concludes that the global area of tropical forest fires in 1997 is nowhere near the area burnt on Borneo in 1983-84, and well below the 13 million hectares that burnt in China and the former USSR in 1987. Error: Lomborg has read the 2001 FAO report on the world┤s forests: this report states that "Comprehensive global statistics on wildland fires required to make a reliable comparison of global fire occurrence in the 1980s and the 1990s do not exist." Thus, Lomborg should have been aware that his postulates could not be verified. Furthermore, Lomborg has read the WRI report from 2001: "World resources 2000-2001". Here he could read: "Tropical forest fires were unusually severe in 1997-98, following less-than-average rainfalls due to El Nino. The number of forest fires in Brazil increased dramatically between 1995 and 1998 . . . ". Thus, Lomborg must have known that what he wrote about Brazil, was not correct. His misleading is deliberate. Concerning China and USSR, it has - at least later - turned out that Lomborg was wrong. In the report by Goldammer et al. (2001), referred to in the previous section, it is stated that the area burnt in China in 1987 was "only" 1.3 m hectares. Concerning USSR, we may read that "official statistical data on forest fires before 1988 were deliberately falsified for political reasons", and that two different estimates on the fires in the Russian Far East and eastern Siberia in 1987 were 14 m ha. and 6 m ha., respectively; the latter was based on satellite data. We also read that this was much higher than in most years. Later, we read: "The strong negative impact of forest fires is particularly evident during the years when catastrophic forest fires are driven by extremely unfavourable weather conditions during the fire season. During the last 15 years [i.e. the period 1985-2000], Russia faced such years in 1987 and 1998." The area burnt in 1998 was 9.4 mio. ha., based on satellite data, which is more than the 6 m ha based on satellite data in 1987. Thus, when this is combined with data from Indonesia and Brazil, the claim that worldwide, 1997-1998 saw the burning of greater forest areas than ever reported before, seems warranted. Lomborg┤s claim that 1997 was "in no way" different, is completely wrong.

P. 117 left: ERROR

Lomborg writes that protection of forests is facilitated by economic growth: "Exploitation is due both to individual poverty and to poor government finances. Both problems are really rooted in poor economic conditions, and solutions therefore need to include solid, economic growth, in order to ensure that, in future, developing countries will be able to afford the resources to establish a broader perspective on forest development". Error: Lomborg does not provide any evidence for a positive correlation between economic growth and forest protection, and he neglects at least three investigations that contradict him. One is Koop & Toole (1999): Is there an environmental Kuznets curve for deforestation ? J. Development Economics 58: 231-244. Another is the investigation initiated by the World Bank that Lomborg refers to on p. 176-177 (Shafik 1994). Here, among the 6 indicators of the environment that were studied, one was forest preservation, and this indicator was not positively correlated with income. A third is the investigation presented in the FAO 2001 report on the world┤s forests, which Lomborg has read. In table 4 of this report the correlation between the rate of deforestation and several parameters is presented, among them the gross national product per capita. The correlation between GNP/capita and deforestation is +0.21, which is not significant. In correspondance on this item, Lomborg has claimed that the correlation is positive, so he is right in claiming that better economy reduces deforestation. However, in saying so, Lomborg dismisses the central theorem of statistics, viz. that a correlation which is not significant, could be due to mere chance, i.e. there is no proof that a positive correlation exists. Thus, what Lomborg writes is not only contrary to evidence which Lomborg has seen, but is also contrary to fundamental principles of statistics. Thus, Lomborg┤s text is deliberately wrong.