30 June 2009

Tropical Cyclone Damages in China

A new paper has been published in the Bulletin of the American Meteorological Society titled 'Tropical Cyclone Damages in China: 1983 to 2006" by Zhang et al. available here in PDF. The paper finds no trends in either tropical cyclone landfalls or in normalized damage, as indicated in the following two figures from the paper.

The paper concludes:
The direct economic losses and casualties caused by landfalling tropical cyclones in China during 1983–2006 are examined using the dataset released by the Department of Civil Affairs of China. . . The direct economic losses trended upward significantly over the past 24 yr. However, the trend disappears if considering the rapid increase of the annual total GDP of China, suggesting that the upward trend in direct economic losses is a result of Chinese economic development. There is no significant trend in tropical cyclone casualties over the past 24 yr.
What does this mean? This means everywhere that scholars have looked and published results in the peer-reviewed literature (including the United States, Caribbean, Central America, Mexico, China, India, non-China East Asia, and Australia), there have been no trends identified over the periods of record in either landfalling tropical storms or their damage.

Even though this is what the peer reviewed literature says, acknowledging as much is enough to get you labeled a "denier." We do live in interesting times.


  1. I suspect that there is an error n Figure 7, as the graphs look the same to me, and China's wealth has increased faster than population.

  2. Roger,
    Graphs ‘a’ and ‘b’ of figure 7 are not the same. The losses in ‘b’ are higher in later years. Not much, but different.

    Oh, and only a denier would deny the difference in the dots detected with due diligence.

  3. How do they justify using GDP to ignore the trend? I've heard of blaming increasing population in the storm track, and that makes sense to me. But I don't see using GDP overall for such a reason. More development could just as easily mean more efforts to protect population and infrastructure, which would cause costs to decrease. Otoh, higher GDP could also mean more exposed development, which would cause even higher expense.

    So while I agree that there is no proof of increased losses due to AGW, I don't see this study as adding much to the discussion. And I'm not calling them (or you) deniers.

    Is there other literature that justifies using GDP for this purpose?

  4. Dean-

    GDP is appropriate to use, see Norhaus here:


    And also my discussion of Nordhaus at:


    I am out the door but happy to answer in more dpeth later if you have follow ups. Thanks!

  5. nor have major hurricanes increased in the US.

    Atlantic major hurricanes are not historically outside the mean.

    the global tropical cylcone index is at a 30 yr low.

    and we can add strong US Tornados as another measurement that is recently below normal.

  6. Roger,

    Regarding the first cite, two of the conclusions are:
    "there appears to be an increase in the frequency and intensity of tropical cyclones in the north Atlantic"
    "greenhouse warming is likely to lead to stronger hurricanes, but the evidence on hurricane frequency is unclear"

    along with his comments on GDP. And there seems to be some contradiction between those two conclusions.

    It seems to me that there is a lot of contention on this issue. I have seen reference to studies that find increases in extreme storms (not losses) and other that don't. There are many metrics - wind, mb, landfall, overall energy, etc.

    The studies you reference are part of the debate, but just one side of it.

  7. Dean
    Most folk define hurricanes as limited to the Caribbean so there isn't a contradiction. And the North Atlantic debate does not affect the non-trend conclusions of land-falling cyclones because these cyclones rarely land.

    As for contention, if you can't find a trend anywhere despite many, many efforts using increasingly creative stats then the debate should really be over by now.