Climate Feedback

The Hurricane-Global Warming Debate, No Clarity Yet

Posted by Oliver Morton on behalf of Roger Pielke Jr.

Last week Kevin Trenberth offered a perspective on Nature Climate Feedback where he suggested that “clarity is actually emerging” in the scientific community on the relationship of greenhouse gas emissions and hurricane activity. My perspective on this debate differs from Kevin’s: Those in the tropical cyclone research community with different views on this subject remain as far apart as ever, and a community consensus beyond that presented in last year’s WMO report has yet to emerge.

Here are a few additional comments following Kevin’s piece:

1) The recent Holland (2007) paper in EOS dipped back into the same North Atlantic dataset that is the subject of intense debate to find that the area where storms have formed appears to have changed over time, observing that the dataset shows “the eastward spread of cyclone developments into the equatorial Atlantic.” Of course, those who argue that changes in observational systems are responsible for the increase in storm frequency found in the dataset can equally apply this logic to the location where storms are first observed. That is, they would argue that as observations have improved, more storms have been observed to have formed in the eastern Atlantic. Holland’s paper therefore provides ample support for both sides of the debate, which is not surprising since its analysis depends upon the exact same dataset that has been contested. Whether or not landfall rates are a constant proportion of basin activity remains a scientific frontier in hurricane climatology.

2) What has been neglected in the debate – by both sides – and is implicit in Holland (2007) is that whatever the cause of increased activity in the North Atlantic, the increase has taken place in the eastern Atlantic far from where landfalls occur (and this holds for storm counts and measures of intensity). On this point everyone seems to agree (see this exchange between ClimateAudit’s Steve McIntyre, who pointed out this trend on his blog and Georgia Tech’s Judy Curry, who is a collaborator with Holland). What this means is that if global warming has indeed led to an increase in North Atlantic hurricane activity, then it has necessarily (and ironically) had the effect of making the United States relatively safer from hurricanes as the proportion of total storms that makes landfall has decreased dramatically. (That is, landfalls show no trend while basin activity has increased due to global warming according to Mann et al. 2007 cited by Kevin). Imagine that headline – “Global Warming Protects U.S.from Increasing Landfalls”! Strange but true, but that is exactly the right message to take if Holland (2007) is taken at face value.

3) Kevin writes that none of the studies that he mentions cites Chang and Guo (2007), when in fact both studies that he references cite Chang and Guo. As I am sure that Kevin is aware, there are several other major studies of historical storm overcounts in the North Atlantic presently available in the community, which arrive at a range of conclusions which can be used to support either side of this debate. And there is also a major reanalysis project ongoing at NOAA as well. So it is perhaps a bit misleading to select one study as being “most definitive” at his point. On the historical undercounts there are various ways to assess whether they “matter” or not, and the Mann et al. (2007) paper cited by Kevin provides only one approach to this question. I have pointed out that if one accepts Chris Landsea’s estimates of historical storm undercounts, then one gets 529 total storms from 1905-1955 and 527 from 1956-2006, so it would seem to matter from that perspective. Obviously, the community’s lengthy arguments over this point would suggest at a very basic level that it does matter.

Debate over climate change and hurricanes exemplifies what Dan Sarewitz calls the “excess of objectivity” in science. Simply put, this means that in most any political issue involving science, there is an ample reservoir of knowledge from which partisans can selectively draw from to support a wide range of positions. Those skeptical of the human role in climate change are often criticized for cherrypicking science and selective presentation of science. And appropriately so. But these standards should be applied across the board, even to those who argue strongly for a human role in the climate system.

On the issue of climate change and hurricanes, clarity on the relationship of greenhouse gas emissions and hurricane behavior is not yet emerging, but the science is continuing and evolving in rapid fashion. The most recent study (what ever it is) will add to knowledge, but will not end debate. That will take some time.

Meantime, you can take a look at an exchange on this subject that Kevin and I were involved in last year, where we had similar views – Kevin’s that the debate was decided, and mine that it was still ongoing:

Pielke, Jr., R. A., C. Landsea, M. Mayfield, J. Laver and R. Pasch, 2005.

Hurricanes and global warming, Bulletin of the American Meteorological Society, 86:1571-1575. (PDF)

Comment by Anthes et al. 2006, Hurricanes and global warming: Potential linkage and consequences, BAMS, Vol. 87, pp. 623-628. (PDF)

Pielke, Jr., R. A., C.W. Landsea, M. Mayfield, J. Laver, R. Pasch, 2006.

Reply to Hurricanes and Global Warming Potential Linkages and Consequences, Bulletin of the American Meteorological Society, Vol. 87, pp. 628-631, May. (PDF)

— Roger Pielke Jr.


  1. Report this comment

    Sylvain said:

    Climate feedback is doom to go back on the blacklist with this post.

  2. Report this comment

    Urs Neu said:

    Concerning point 1: The main conclusion of Holland 2007 is: “This decrease [of landfall ratio] arose largely from a similar decrease of tropical cyclone developments in the well-observed western Caribbean and southern Gulf of Mexico regions, and it cannot be ascribed to unobserved eastern region tropical cyclones. It thus appears to be a real feature and not an analysis artifact.” – The eastward spread of cyclone development is discussed, but was not the main point.

    Concerning point 3: The Landsea 2007 estimation of underreporting is based on the decrease of the percentage of landfalling hurricanes between the periods 1900-1965 and 1966-2006. However, if his assumption, that this decrease is due to underreporting, was true, then the percentage should be even higher before 1900, because ship tracks were even more rare at that time. However, the opposite is true. The percentage 1851-1900 in the observational record is lower than 1900-1965, and the percentage 1851-1886 is about the same as 1966-2006. Thus the basic assumption of Landseas estimation obviously is wrong.

  3. Report this comment

    James Elsner said:

    The logic escapes me. What does the ratio of landfalls to total storms have to do with the actual risk of getting hit along the coast?

    While there is no significant trend in U.S. landfall statistics, the probability of a strong hurricane affecting the continental United States increases (decreases) in anomalously warm (cool) years by an amount that is consistent with climate model projections of overall hurricane activity (Jagger and Elsner 2006).

  4. Report this comment

    Steve Reynolds said:

    Re Urs point 3 about pre 1900 data:

    ’In 1870 there were 65,000 vessels in the world fleet, as compared with 24,000 in 1914. The world fleet did not build back up to the 1870 figure until 1973.

    With the introduction of iron and then steel in ship construction ships got larger (both sail and steam) and so fewer were needed. It took the vast increase in world trade after WWII to overwhelm the effect of this factor.’

    From comment 94:

  5. Report this comment

    Phillip Huggan said:

    The connection between increased hurricane strength/incidence and Global Warming is nearly 100% certain. Only thing left to argue over is timeline; whether we’re already into the big cyclones. Given that hurricane costs scale at the cube of windspeed not accounting for increased urbanization in the future (60% of cities are coastal), the prudent course of action for civic politicians is to fund (via tax/rebate incentives) relavent homeowner Home Depot purchases and the prudent course of action for federal politicians is to keep National Guard reserves where they are desired and helpful.

  6. Report this comment

    JamesG said:

    James Elsner

    Correct – the risk of being hit would be hugely increasing even if the sea cooled, owing to the huge increase of people to the area (around 20% per year). An argument for adaptation?

    Usually no significant trend means no increase in probability. To obtain an increased probability therefore you have to massage the data. Enter Bayesian analysis. Warning though: Bayesian methods can be biased by initial assumptions. Your bald statement above contains none of the many caveats of your paper.

  7. Report this comment

    George B said:

    Hurricane landfalls show no trend according to your post. Therefore, the probability of a hurricane striking the U.S. coastline is unchanged.

    Now exactly what do the additional hurricanes on the other side of the Atlantic have to do with this?

    Elsner is correct. The statement “What this means is that if global warming has indeed led to an increase in North Atlantic hurricane activity, then it has necessarily (and ironically) had the effect of making the United States relatively safer from hurricanes as the proportion of total storms that makes landfall has decreased dramatically.” is deceptive, weasel worded propaganda, not scientific analysis.

  8. Report this comment

    JamesG said:

    To be more specific about assumptions influencing results: The initial inference that high absolute temperature causes tropical storms is what forces both the models and any Bayesian filter to agree on increased hurricane likelihoods. However, using an alternative inference: that all storms are caused by temperature (or potential) difference, would force models to agree with the actual landfall trend, because in a warming world the difference doesn’t actually increase. The landfall record – being quite accurate – can also be extrapolated to give a good estimate of the total number of storms unless the locus of storm activity has actually changed. Has it?

Comments are closed.