Lomborg-errors: "Cool it!"

Extreme weather, extreme hype
    Home                                                                                                                                   Cool it

 Comments to pages 94 - 116 in Cool it.


Chris Mooney criticises Lomborg´s chapter on hurricanes here.


Page 95: "It makes three strong and specific points".
Flaw: Lomborg refers to a summary statement on tropical cyclones and climate change issued by the World Meteorological Organization in November 2006. This summary contains not three, but ten clearly formulated points. But Lomborg cites only the first three of these points - which happen to fit his overall agenda - and leaves out those that do not fit him. Most importantly, he omits the point stating that scientists agree that we should expect a rise in peak wind-speed as the sea surface warms. Instead, he focuses on the point stating that "The recent increase in societal impact from tropical cyclones has been largely caused by rising concentrations of population and infrastructure in coastal regions." Few scientists would oppose against this statement, but Lomborg claims that this is all there is to it - as if we had to choose whether damages were due  to denser infrastructure or to increasing wind speeds, instead of admitting that both factors contribute in tandem, and that the intensity of cyclones is expected to rise in the future, wherefore the necessary protective measures will be ever more costly - especially as there is at the same time a rising of sea level.

Page 95: "When Al Gore tells us that . . . it is incorrect".
Flaw: Lomborg refers to a summary statement on tropical cyclones and climate change issued by the World Meteorological Organization. He focuses on point 1 in this statement, that a detectable human-caused signal is controversial, but he omits the following points:
"6. It is likely that some increase in tropical cyclone peak wind-speed and rainfall will occur if the climate continues to warm. Model studies and theory project a 3-5% increase in wind-speed per degree Celsius increase of tropical sea surface temperatures.
7. There is an inconsistency between the small changes in wind-speed projected by theory and modeling versus large changes reported by some observational studies."
Point 6 says that a human-caused signal is likely to appear in the future, and point 7 says that unexplained increased wind-speeds have been observed already. Therefore, when Gore says that "global warming is making hurricanes more powerful", he is right. this is considered a deliberate flaw, as Lomborg has read points 6 and 7, but neglects them.

Page 100 bottom and note 507: " . . . or almost twice the damage of Katrina".
Flaw: The source given for this statement is NOAA 2006c. But according to this source, the economic costs of the Great Miami hurricane would today have been $ 90 billion, which is only 20 % higher than the estimate given in the same source for Katrina ($75 billion). The figure cited by Lomborg for the Great Miami hurricane originates form another source, Pielke et al. 2007.

Page 100 bottom : "Had it hit today, it would have caused damage of about $100 billion . . ".
Flaw: Lomborg´s source, Pielke et al. 2007, gives a figure of  $78 billion. "

Page 100 bottom, note 509: In the note: "Notice that both hurricanes . . . caused many deaths . . ".
Flaw: The important omission is that hurricane Katrina caused the death of 1,200 - 1,300 people, making it the second most fatal hurricane after the Galveston hurricane. Estimates for the Great Miami hurricane range from 373 to 800 dead. That better warning systems today would have evacuated much of the city is at odds with what actually happened when Katrina struck in 2005.

Page 101, figure 29:
Flaw: The figure is taken from Pielke et al. 2007, but the columns for 1900 and 1915 are considerably higher than in the source.

Page 102 top: "Katrina ranks third at $81 billion."
Error: That source (Pielke at al. 2007) which is used here, has the following estimates of damages: Galveston $72 -78 bn, Great Miami $140 - 157 bn, and Katrina $81 bn. So Katrina ranges second, not third. Although Lomborg states at the bottom of page 100 that the Galveston hurricane would have cost $100 today, he gives no source for this claim. It must be added that we are talking about hypothetical damages, and that the calculation of these depends inter alia on the economic growth rate from 1900 to 1925, which is not precisely known. But when Lomborg´s source has Katrina ranging as second, it is a manipulation to write without any reservation that it ranges third.
    The economic damages from Katrina are estimated at more than $100 bn by the federal organisation NOAA.


General comments:
One of Lomborg´s main references is Kundzewicz et al. (2005) (note 535).
This reference says: "Out of more than a thousand long time series made available to the authors, a data set consisting of 195 long series of daily mean flow records was selected. " Out of these 195, 140 were from Europe and North America. The criteria for choice of stations included "catchment size, with priority given to smaller catchments which are more likely to have less strong anthropogenic influence . . ". None of the time series went further than 2001; that is, the extremely large floods in east-central Europe in 2002 do not enter the analysis.
The exclusion of catchments with much anthropogenic influence is in a way sensible, but also has consequences. Those rivers which up during history have relatively often produced disastrous floodings are more likely than other rivers to have been affected now by the building of dikes. Also, large rivers are more likely to gather precipitation from a large area and hence to produce large floodings. The exclusion of such river systems, and the fact that all time series stop no later than 2001, all may contribute to leave out precisely those situations where large floodings occur.
After the publication of that paper, some other papers have been published with opposite results. Schmocker-Fackel & Naef (2010) report increasing floods in parts of Switzerland (Journal of hydrology 381: 1-8). (Petrow & Mertz (2009) : Journal of hydrology 371: 129-141 analysed data from 145 river discharge gauges from all over Germany for the period 1952 - 2002 and found a general rising trend in extreme events for the western, southern and central parts of Germany. These authors also publish a survey of previous studies. Many of these indicate no general trend, or decreasing trend in one region simultaneous with increasing trend in another region. A few indicate marked increases in flood events, for instance in Australia: Franks (2002): Hydrology and Earth System Sciences 6 (1), 11–16.

Page 109 top: "This is well correlated with the historic evidence that shows much greater flood risks in the colder cliamtes  of the Little Ice Age."
Flaw: At least for the Czech Republic, this is not true. The floodings in August 2002 were the largest ever since the start of recordings in 1500 (Brazdil et al. (2006), cited by Lomborg). 

Page 109 top: "This pattern was the main cause of flooding on the lower Rhine . . ."
Flaw: The pattern described by Lomborg does not hold true for the period 1951-2002 (Petrow & Mertz, cited above). They found that the mainpart of stations in the Rhine catchment area showed significantly increasing trends in flood events. For Germany in general, they state that the trend analyses for the winter maxima gave similar results
as the analyses for the annual maxima. Significant upward trends in winter maxima were identified at 23% of all sites. No significant downward trends were detected.

Page 109 : "For the river Vltava in the Czech Republic, floods have decreased over the past century. "
Comment: This is technically correct, when `the past century´ means the 20th century. However, it is also misleading, because the largest flood ever recorded there was in 2002.


Page 115: "Essentially, we see that serious tornadoes have declined substantially over the past half-century."
Flaw: According to the paper by Verbout et al., cited by Lomborg, the trend for the period 1954-2003 has been as follows for the different categories of tornadoes:
F0 and larger (all reported tornadoes): More than a doubling
F1 and larger: No change
F2 and larger: More than 50 % decrease
F3 and larger: More than 50 % decrease
F4 and larger: Slight decrease, probably not significant
Lomborg refers to those categories that show a significant decrease (F2+ and F3+), but not to those that show no significant trend (F1+ and F4+). The authors of the paper write: ". . the F4 and greater record has also remained relatively consistent throughout the time period, yet the F2 and greater and F3 and greater records have a declining trend over this period, with a possible disontinuity in the late 1970s. . . suggesting overrating problems with the F2 rated series." Notice that the F-scale was invented in 1971, and it was probably not used for post-hoc classification of tornadoes until some years later. The discontinuity in the late 1970s, especially for the F2 series,  suggests  that former tornadoes were relatively often classified as F2 or F3, whereas similar recent tornadoes might only be classified as F1 or F2, respectively. In conclusion, there is probably no true trend at all, implying that Al Gore is not right (no real increase) and Lomborg is not right (no real decline). This might not have been counted as a flaw in Lomborg´s book, had he not stressed the point with italics and not called this "an important additional piece of information."

Page 116, figure 34.
The data source is no longer at the web address given by Lomborg, but is found here.
Flaw: The decline in fatalities in the most recent years is exaggerated. Actually, the average yearly number of fatalities due to hurricanes (before Katrina) and floods is approximately the same now as in the 1940s. As the US population has doubled (from 140 millions to 280 millions), the number of fatalies per million is now half of what it were (hurricanes without Katrina) or slightly above half (floods). The curves in the figure show a continued decline after 2000. This must be an artifact due to the smoothing procedure; with 11-year smoothing, the curves should have stopped in 1999.