Wind Speed in Holland and Webster

Perhaps the key issue in Emanuel 2005 (as observed in Landsea’s 2005 comment) was the adjustment to the ATL data set performed by Emanuel (citing an old publication, Landsea 1993). We’ve snickered a bit recently at Martin Juckes using older data rather than newer data – only in climate science would this seem possible. And we see another such incident in Holland and Webster – this time with wind speed methodology – where they use a method of Landsea 1993 rather than the recommendations of Landsea 2005.
Holland and Webster stated:

We use the “best-track” tropical cyclone data base from the National Hurricane Center (Jarvinan et al. 1984). The only changes to the data set data have been to include the intensity corrections recommended by Landsea (1993).

Sounds innocent enough. But underneath this is the raging debate between Emanuel and Landsea about adjustment of wind speeds in the Hurdat database. I reviewed Landsea 1993 and Landsea 2005 here which, upon re-reading, stands up as a pretty reasonable representation of the debate. As noted in my earlier post, Landsea 2005 stated:

It is now understood to be physically reasonable that the intensity of hurricanes in the 1970s through to the early 1990s was underestimated, rather than the 1940s and 1960s being overestimated. To examine changes in intensity over time, it is therefore better to use the original hurricane database than to apply a general adjustment to the data in an attempt to make it homogeneous.

Despite this explicit statement from Landsea, Holland and Webster apply the method of Landsea 1993 without even citing Landsea 2005. Whether Landea 2005 is right or wrong, this adjustment is a topic that any competent data analyst needs to deal with before analyzing trends; however, Holland and Webster simply ignore the issue.

I noticed this matter when I experimented with a plot of the average length of a ATL cyclone and the average length of an ATL hurricane (which I examined in the course of examining the supposed constancy of proportion of hurricane-days to cyclone-days.) The average length of an ATL hurricane declines in the late 1960s, coinciding with the issue raised in Landsea 1993. Holland Webster presumably reduce wind speeds prior to 1970, but don’t provide any details on the adjustment other than what I’ve quoted. Landsea 1993 is specific to the aircraft reconnaissance period – but one would need to check whether H and W adjusted wind speeds for earlier periods as well.
The early years with high hurricane-days mentioned previously – 1886, 1887, 1933 – were not done with aircraft reconnaissance and would not, on the face of it, be subject to Landsea 1993 adjustments in any event. But H and W would need to be checked somehow to see if they adjusted data prior to the aircraft reconnaissance period as well. The results that I’ve presented so far have been based on the methodology recommended in Landsea 2005 – using the original Hurdat data – and will therefore not reconcile precisely to HW results (but are obviously a valid sensitivity study that the original data analysts should have done as well). I urge people interested in the topic to read or re-read my earlier post on Landsea 1993 versus Landsea 2005, as this affects H and W as well.


  1. jae
    Posted Jan 4, 2007 at 12:59 PM | Permalink

    Don’t these inhomogeneities make it impossible to use standard statistical techniques?

  2. Judith Curry
    Posted Jan 4, 2007 at 3:32 PM | Permalink

    Emanuel describes his discussion with landsea on this topic, you can see this about halfway down the page
    This presents the chief diagram supporting the landsea correction.

    Here is Emanuel’s critique of Landsea’s Science Perspective re the correction:

    Click to access E3-Critique-of-Landsea-etal.pdf

    Note: I do NOT find this critique convincing.

    Apart from Landsea and Emanuel, who seem to be having a little war on the subject, what do other hurricane scientists have to say about this? Webofscience shows 95 citations for Landsea (1993), with all but 13 of them occurring prior to Emanuel’s 2005 paper (other than Landsea’s science perspective, it does not appear that any of the 13 most recent papers refute the correction in any way). I cite below a few of the papers that used the Landsea correction, including a paper of Landsea’s as recent as 2001. There was also a paper by Emanuel published in 2000 that used the Landsea (1993) correction, which did not elicit any complaints from Landsea. Landsea himself in an essay written in 2000 explicitly supports use of this correction

    Goldenberg SB, Landsea CW, Mestas-Nunez AM, et al.
    The recent increase in Atlantic hurricane activity: Causes and implications
    SCIENCE 293 (5529): 474-479 JUL 20 2001

    Emanuel K
    A statistical analysis of tropical cyclone intensity
    MONTHLY WEATHER REVIEW 128 (4): 1139-1152 APR 2000

    Pielke RA, Landsea CN
    La Nina, El Nino, and Atlantic hurricane damages in the United States

    Landsea CW, Pielke RA, Mestas-Nunez A, et al.
    Atlantic basin hurricanes: Indices of climatic changes
    CLIMATIC CHANGE 42 (1): 89-129 MAY 1999

    Landsea CW, Bell GD, Gray WM, et al.
    The extremely active 1995 Atlantic hurricane season: Environmental conditions and verification of seasonal forecasts
    MONTHLY WEATHER REVIEW 126 (5): 1174-1193 MAY 1998

    Most interestingly, Bill Gray supports the use of the Landsea (1993) correction. See for Bill Gray’s critique of the Emanuel (2005) paper. While Gray is highly critical of Emanuel’s paper, he explicitly supports the use of Landsea’s correction, stating in his analysis that “The maximum sustained winds from 1950-1964 have been adjusted downward using the Landsea (1993) adjustment factor.”

    So that is the evidence that supports use of Landsea’s (1993) correction. The evidence against it is a perspective piece (not a scientific research article) that criticizes Emanuel for using this correction, without providing any rigorous analysis or other scientific argument for refuting Landsea’s own correction.

    What to make of this? Landsea’s sudden dislike of his own correction that had been used for over a decade right after Emanuel’s publication and Katrina seems curious, especially since Landsea has not rigorously refuted his correction in any way.

    With regards to H/W. I would have to say that the Landsea correction stands until someone refutes this. Should H/W have acknowledged this “controversy” that has mostly been playing out in the media? Given the body of evidence supporting this correction and the tradition of using the correction (and in the absence of any rigorous refutation of the correction), I would say that it is still appropriate to use this correction. That said, I think the intensity data prior to 1970 is not very good in any event, and much further work is needed in this regard.

  3. Steve McIntyre
    Posted Jan 4, 2007 at 4:57 PM | Permalink

    I’ve compared my counts of hurricanes and cyclones using the Unisys data without adjustment and your counts as provided in natldata.xls and found no differences in the period 1945-1970 covered by the Landsea 1993 correction.

  4. Roger Pielke Jr.
    Posted Jan 4, 2007 at 5:10 PM | Permalink


    This information also from Emanuel seems relevant as well:

    “As for that correction, it was based entirely on work by Chris Landsea (Landsea, 1993). Chris and I have talked about this since then, and we both agree that the correction was probably a little too large. I have since implemented a different, smaller correction.”

    “In correcting for biases in the original
    Atlantic tropical-cyclone data, I relied on a
    bias correction applied by Landsea6, presented
    as a table. I had fitted a polynomial to that
    correction, as I felt that a continuous rather
    than discrete correction was more defensible.
    Landsea believes that this had the effect of
    overcorrecting the most intense storms in the
    pre-1970 record, and I accept his revision to
    my analysis (Fig. 1b of ref. 3).”

    Click to access resource-1890-2005.48.pdf

    In the face of such uncertainties and disagreement, the prudent thing to do is to perform the analysis with error bars on the data, and see if the errors make a difference for the findings.

    Should H/W have considered such uncertainties in their analysis? I can’t imagine why anyone would not. At a minimum it results in a more comprehensive and defensible analysis.

  5. David Smith
    Posted Jan 4, 2007 at 6:15 PM | Permalink

    Several comments on storm intensity estimates:

    1. Even today, in 2007, there is uncertainty in current storm windspeed estimates. All one has to do is to read the discussion section of 2006 hurricane advisories to get a sense of the often-conflicting factors (satellite appearance, Dvorak estimate, recon data, was the recon sampling representative, perhaps incorporate a ship or buoy report, estimated windfield, need for continuity in public advisories, etc ) that the analysts have to gell into a single number. It’s an educated guess.

    As one goes back in time the pieces of data become more sparse and the interpretations vary. The combined mess is like Frankenstein’s monster – not suitable for climatology before about 1980, when satellite data became good enough for apples-to-apples comparisons.

    2. Compare, for a moment the recorded wind history of an older storm (1955’s Hurricane Connie ) to that of a modern storm ( 2003’s Isabel ). Note the relative stability of Connie’s reported winds (a steady 125 knots for over four days!) versus the up-and-down of Isabel’s winds. Satellite and recon have taught us that a storm’s windspeed varies quite a bit in its lifetime whereas in 1955 people likely believed that storms were smooth and made their estimates based on that belief.

    My belief is that, today, we capture storms which briefly reach hurricane windspeed whereas in older days smoothness was assumed and it was assumed that storms never made it to hurricane strength. Result: likely undercount of hurricanes in the past. They also likely missed peak windspeeds, with the result of underestimating maximum storm strength.

    For fun, look at the 6hour-to-6hour variability in windspeeds of storms in the last twenty years versus the 6hour-to-6hour variability fifty to sevety years ago: apples-to-oranges. Today’s storms, which reflect reality, are choppy while older storms are smoothed.

    3. Here’a a hoot from the historical record: 1960’s Hurricane Ethel . According to the data, it went from tropical storm strength to a category 5 hurricane in 18 hours, only to collapse back to tropical storm strength just 12 hours later. Do I have any belief whatsoever that this is accurate? Nope. Yet, it’s part of the record.

  6. David Smith
    Posted Jan 4, 2007 at 7:07 PM | Permalink

    Steve, you may want to take a look at histograms of the “reported 6-hour change in windspeed” (e.g., -15 knots, -10 knots, -5 knots, no change, +5 knots, + 10 knots, etc) for cyclones in three periods: 1930-1950 (pre-recon active period), 1960-1980 (early satellite) and 1985-2005 (modern satellite).

    This may confirm, or disprove, my sense that older data was smoothed, and some important stort-term strength changes were likely missed. This would tend to reduce peak windspeeds and the count of hurricanes.

  7. Judith Curry
    Posted Jan 4, 2007 at 7:09 PM | Permalink

    Roger, Emanuel originally overestimated Landsea’s correction by something less than 10%, by using an slightly erroneous polynomial approximation to Landsea’s correction. He has subsequently used the “corrected’ correction in his analyses. His corrected analysis is far closer to his original analysis than to the uncorrected data.

    Re H/W, it would have been rather astonishing not to have used the Landsea correction. but I agree, I would have said more about the uncertainties in the intensity data.

    David, I’m hoping that the effort led by Jim Kossin can go further back in time, before 1983, to assess and possibly improve some of the earlier intensity data in the satellite era.

    Also, I’ve sent Steve M a file that illustrates how much difference the landsea correction makes to NCAT45, PDI, and ACE calculations, hopefully he will post that soon

  8. David Smith
    Posted Jan 4, 2007 at 7:59 PM | Permalink

    RE #6 On second thought, a histogram probably doesn’t illustrate what I’m trying to understand. I’ll play with the data and look for a way to illustrate the problem.

%d bloggers like this: