RealClimate logo

Reactions to tighter hurricane intensity/SST link

Filed under: — gavin @ 20 March 2006

There was another twist to the hurricanes/global warming issue in Science Express on Friday where a new paper from the Webster/Curry team just appeared. This study, lead by Carlos Hoyos, crosses a few t’s and dots a couple of i’s on the connection of increasing numbers of intense hurricanes (Cat. 4 and 5) to sea surface temperatures (SST). Basically, they looked at a number of other key variables for hurricane intensity (like wind shear and humidity) and examined whether there was any pattern to those variables across the different ocean basins that they study. Bottom line? None of the other variables have as much explanatory power for the long term trends as SST which is the only consistently trending constituent in the mix. So far, so un-surprising. However, one interesting aspect of this story is that almost all the key players in the ongoing debate were interviewed by different journalists in various media and those comments are probably more useful for gauging the state of play than the details of the new paper itself.

First off, Chris Landsea, a serious researcher in the ‘dissenting’ camp, in the WSJ story (reprinted here) and in a Newsday article is quoted as saying that the data (particularly from the early 70s and the Northern Indian Ocean) aren’t good enough to come to these conclusions. However, Curry riposts that for the doubling of Cat 4 + Cat 5 storms to be an artefact, half the Cat 1 + Cat 2 storms in the early records would have had to have been mis-classified Cat 4 storms instead and she thinks that unlikely. More oddly, Landsea also makes the point that the sensitivity of the hurricane intensity disagrees with model predictions and theory. This is odd, not because it’s not true (and is the principle reason why the attribution of more intense hurricanes to GW is not yet set in stone), but because Landsea has previously been much more of a champion of favoring observations over modelling.

In the same National Geographic piece, Roger Pielke Jr. is broadly supportive and makes his standard comment that the increasing damages from hurricanes is mostly related to increased development rather than changes in intensity, but that obviously, future potential increases in intensity will be an additional factor.

Fox News interviewed Kerry Emanuel (author of a related study last year, which was slightly modified following comments by Landsea in Nature), who is also unsurprised that the longer term trends are related to SSTs and not to any of the other factors. Kevin Trenberth (also in the Fox News piece) made a good point though. The authors of the study used the NCEP reanalysis as the source of their data. A reanalysis is a re-running of the current state-of-the-art weather forecasting model for all the sources of data that were available in the past (i.e. a hindcast of what all the 6 hour weather forecasts would have been if they had used today’s model). These projects (and there are two main ones – NCEP and ERA-40) have a problem in that the amount of useful data increases as you go along – most significantly around 1979 when satellite data starts to be significant. So estimates of key quantities are likely to be worse prior to 1979. Not mentioned, but conceivably important is that the NCEP reanalysis is tied in some respects to the radiosonde data, which, as we discussed last year, may have some spurious trends. This doesn’t obviously affect the results significantly, but it does suggest that doing the analyses again using the ERA-40 data might be a useful check.

So where does that all leave us? Basically, although everyone acknowledges that there are data problems early in the record, it seems clear that there has been a global rise of the most intense hurricanes over the last 30 years and the most obvious explanation is that this is related to the contemporaneous increases in tropical SST in each basin. However, the magnitude of the correlation cannot yet be explained in terms of our basic theoretical understanding, and is significantly stronger than some modelling work has suggested it should be. Possibly the theory needs work (hurricanes are a complicated business!) or there are other factors at play that haven’t yet been considered. Since the SST changes are global, and almost certainly tied to greenhouse gas driven global warming, there are the beginnings of a corroborated link between increases in hurricane intensity and GW – however, so far there are only a couple of ducks in a row.

136 Responses to “Reactions to tighter hurricane intensity/SST link”

  1. 51
    Steve Bloom says:

    Re #48 comment: Rasmus, farther up in the comments someone had the idea of measuring SSTs in the wake of cyclones as well as in their path. From your more detailed description of the dynamics, that sounds as if it might be somewhat helpful (although still not quite a replacement for a depth profile of heat content). Is anyone looking at such an approach?

  2. 52
    Hank Roberts says:

    Measuring SSTs in the wake is mentioned “farther up” at #8, #9 and #13

  3. 53
    Kenneth Blumenfeld says:

    In response to Pat Neuman and Tom Huntington:

    As Tom probably knows, climatologists studying extreme rainfall use techniques that make it almost impossible to tease out meaningful temporal trends in extreme precipitation (and resulting local/regional flooding). Extreme rainfall, on an event-level is driven by the prevailing meteorological conditions…deep moisture, weak upper level winds etc. One could argue pretty effectively that a 100-year type storm requires a very special set of rarely-occurring conditions, and that even in the face of strong AGW, the change in probablility of one of these storms at any point is going to be so small that it will be almost immeasurable…How would you know, for example, that the probablility of such a storm has increased from .01 annually to .02 anually (a huge increase)? You can’t yet, since these probabilities are estimated from observed data, rather than changes in important meteorological factors, like precipitable water.

    Furthermore, analyses of extreme rainfall have been constrained by two important factors: the spatial dimension–that is, not enough gauges to measure the true rainfall maxima for most events–and also selection criteria–how many events are analyzed.

    In any case, I would argue that the methods used for extreme rainfall frequency analyses cannot actually answer the questions about temporal changes in local storm frequency, intensity, spatial coverage and seasonality. So it does not surprise me that the literature has not yet uncovered any significant changes…that is not really what folks have been looking for.

  4. 54
    Lewis Cleverdon says:

    Something I’d be grateful to see clarified is the impact on sea-temperatures at depth of the recently reported 30% decline in the Gulf Stream’s warming northward flow.

    Can anyone say whether a corresponding lack of cooling flow southward yet show identifiable effects on ‘STADs’, and if so, where ?



  5. 55

    Re: 51


    What do you want to know about the pre- and post-storm SSTs? I have the weekly average SSTs in 1×1 lat/lon grids for weeks prior to the storm arrival, the week that the storm was there, and the week after it departed (for every 6-hr storm position from 1982-2005). In our paper, we use the SSTs in the week prior to the arrival of the storm in order to avoid the storm mixing problems. There is a clear difference between pre- and post-storm SSTs. Do you want to know the average difference by year, by storm, by region? I’d be glad to tally those things–given that I can find the time, and someone thought this would be a useful exercise.


  6. 56
    pat neuman says:

    re 53

    I think this is a good summary of the Great Flood of 1993 (not statistical). “The Great Flood of 1993 on the Upper Mississippi River â?? 10 Years Later”, 4 pages.

    Statistics and historical flood data can be found through this link.

  7. 57
    pat neuman says:

    re 50. … EOS article, so I’ll have to look into it … -Chip

    I think this is the EOS article (as mentioned in 50.)

    > Satellite Altimetry and the Intensification of Hurricane Katrina
    > Eos, Vol. 86, No. 40,
    > News 4 October 2005
    > PAGE 366

    > Instead, Katrina intensified most rapidly
    > when she was over areas of anomalously high
    > dynamic topography, as measured by altim-
    > eters (Figure 2b). Katrina intensified first over
    > a warm-core eddy east of Florida as she grew
    > from a tropical depression to a Category 1
    > hurricane.Then, over the Loop Current and a
    > warm-core ring in the Gulf of Mexico, she inten-
    > sified from Category 1 to Category 5.
    > These dynamic topography highs are a proxy
    > for the vertically integrated heat content within
    > the water column.The depth of the warm water
    > pool, and not merely the temperature at the
    > surface, provides the reservoir of energy to
    > intensify a storm [Shay et al.,2000].

    The above text was copied from a post (Oct 4, 2005) at:

  8. 58
    pat neuman says:

    Judith Currea: … “frightening prospect” …
    Jim Acosta: … “government hurricane forecasters have heard these dire warmings before and they’re not impressed” …
    Max Mayfield: … “natural variability ALONE is what this can be attributed to” …
    CBS Video, aired March 20: CBS Evening News with Bob Schieffer
    Superstorms And Global Warming, Jim Acosta (CBS) at:

    Also, FYI:
    Sep 20, 2005

  9. 59
    llewelly says:

    There are several posts about hurricanes and their effect on SSTs. I found this paper on hurricane Edouard useful for understanding the basics, although a bit old (1997). I searched google scholar with ‘SST hurricane wake’ and got a number of similar studies.

    Vaguely related, this page has computer-modeling results that forecast SSTs, and potential ocean temperatures down to 3000m deep. Currents, salinities, and other modeled ocean properties are also available. I gather that there are relatively few direct measurements of ocean temperature profiles, but there are reasonably successful efforts extrapolating ocean temperature depth profiles using computer models. Of course this is all intended for use in real-time forecasting, not long-term trend analysis, or climate change studies, but I believe the underlying model is more flexible, although I don’t know. (I *do* know that plenty of other models intended primarily for short-term forecasting have been successfully used in climate change studies; see for exampleTom Knutson’s papers ). has more information about the model.

  10. 60
    Eli Rabett says:

    Pat Neuman in #57 has the correct cite for the EOS article I was remembering. Thanks

  11. 61
    Mauri Pelto says:

    #58 Pat thanks for the link to the testimony. Dr. Mayfield is unequivocal in his emphasis that it is not global warming that is causing the higher activity. This strikes me as an unusually emphatic response if solely from a scientific viewpoint. The debate that is ongoing reminds me of the debates around global warming when I first attended a science conference on it in 1981, the details all fuzzy due to the complicated world, but the underlying physics clear as can be. The higher sst-hurricane relationship is the same providing an undeniable clear physical mechanism for enhancing hurricane development, but associated with many other variables that make is difficult to pin down the exact impacts that have and will occur from higher sst. This makes NHC stand on the sole cause being natural variability out of place.

  12. 62
    pat neuman says:

    re 58. Correct spelling is: Judith A. Curry, on CBS Evening News with Bob Schieffer, Superstorms And Global Warming audio video at:

    Dr. Judith Curry gave excellent discussion aired on CBS March 20. the reporters on CBS Evening News with Bob Schieffer

    Reporters on the CBS Evening News with Bob Schieffer made an attempt tto ridicule what Dr. Curry had to say by using out dated material, saying that: “NOAA states on its website that increased hurricane activity is not related to global warming”. Dr. Curry followed by saying. “That is misleading”.

    re 61. In reply to comment by Mauri that “This makes NHC stand on the sole cause being natural variability out of place.”, it should be understood that NHC is under the NWS umbrella, and NWS is under the NOAA headquarters umbrella. NWS postion has been that global warming is too political and controversial to comment on, but off the record metorologists with NWS have been telling the public that there is no global warming or global warming problem. That was explained to me a couple years ago by senior meteorologists at NWS Forecast Offices in Minneapolis, Detroit and Chicago.

    Also see post #10463 at yahoogroup called globalwarming, below.
    Feb 25, 2006 9:19 pm
    Subject: Fwd: Re: [Global Warming] Re:PS: Blowing the whistle on climate change: articles

  13. 63
    Roger Pielke, Jr. says:

    Re: 58, 61, others

    [unnecessarily provocative intro deleted – moderator]

    Have a look at what Max actually said in Congressional testimony:

    “The increased activity [in the Atlantic] since 1995 is due to natural fluctuations/cycles of hurricane activity, driven by the Atlantic Ocean itself along with the atmosphere above it and not enhanced substantially by global warming.” Mayfield did not express a view on global trends. (Webster et al. had just come out the week before.)

    Why might he think this?

    Well, this was the consensus at the time. For example, MIT’s Kerry Emanuel said as much on his WWW site last fall. Since then, this view has certainly not been rejected as untenable, see for instance the agreed-upon Atlantic PDI dataset of Emanuel/Landsea — figure 1 on p. E12 in Landsea’s comment on Emanuel (2005) here:

    As Gavin suggests (response to 49) there are legitimate differences of opinions here, and Max’s views are certainly within the boundaries of valid scientific debate.

    It is unfortunate to see RealClimate add to the intensity of this issue by putting people into “camps” using words like “dissenters” to characterize them. For example, Trenberth and Emanuel favor radically different theoretical explanations for the role of global warming in hurricane behavior, but they are lumped together in a single “camp” because they are on one side of the global warming “yes/no” debate. Why isn’t, for instance, Trenberth called the “dissenter” and Emanuel/Landsea put into the same camp, based on their views on theory? The particular framing of scientific discourse into mainstream/dissenters based on the litmus test of global warming yes/no is grounded completely in politics not science, as there are many other scientific taxonomies one might choose.

    The reality is that when Mayfield testified his views faithfully represented the mainstream scientific community. Consider what Emanuel said at the time of his own work, and how he characterized the views of his colleagues, which would include Mayfield:

    “Working with both data and models, I see a large global warming signal in hurricanes. But it remains for me to persuade you and other of my colleagues of this, and it is entirely reasonable for you all to be skeptical…it is, after all, very new. It is not surprising, therefore, that what I have come to believe is at odds with any reasonable consensus.”

    If people want to see Max’s views on the science they should have a look at what he has published in the peer-reviewed literature and stop trying to parse individual statements they might come across. Here is a recent paper:

    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.

    NOAA certainly erred when it released that one press release with some unfortunate language about a “NOAA position” – there was no such position. But NOAA as an agency is chock full of outstanding scientists with a diversity of views, and Max is one of them. Under his leadership the NHC last year probably contributed to the saving of thousands of lives. Now people are routinely mischaracterizing his scientific views for what are certainly non-scientific reasons. RC can be a valuable site for scientific discourse, but when it allows unanswered in its comments mischaracterizations and untruths about individual scientists, it takes away from the value of the site. Thanks.

    [Response: Roger, thanks for your clarifications. One interesting point that could be made concerns what the ‘consensus’ was a few years back. I don’t think it was ever that “there is no influence of GW on hurricanes” – that would have been a positive statement of some strength. Instead, the consensus was more like “there is no evidence for the effect of GW on hurricanes” – subtly different, but at the time, much more scientifically defensible (because you never know what the next study will show). The problem with the original “NOAA position”, and with the responses at the end of season press conference was that they specifically used the first statement which I found rather odd and too definitive to be a ‘consensus’. Almost all ‘consensus’ statements go out of their way to ensure that they encompass the possibility that reasonable-yet-unproved hypotheses might at some point be shown to be correct. – gavin]

  14. 64
    Tom Huntington says:

    Reply to Kenneth Blumenfeld No. 53

    Actually, as Kenneth Blumenfeld probably knows, people have been looking very hard for trends in extreme rainfall. For example, see the publication by Groisman et al. (2004, J. Hydrometeorolgy 5:64-85) that summarizes some of his group’s work in this area; also consult the IPCC TAR. I think that you will find that, indeed there are published papers reporting trends towards increasing precipitation in the highest precipitation quantiles at least for the continental US. However, others have shown repeatedly (as I cited in J. Hydrol 2006 319:83-95) that increasing precipitation in the highest quantiles does not necessarily translate into increased flooding. As you correctly point out, to defensibly argue for an observed increase in the frequency of intensity of flood events, that by definition are very rare events, requires long-term records – perhaps longer than we currently have.

    That said, that certainly has not stopped people for looking for trends in frequency and intensity of flooding. The studies that I know of generally have not found any evidence for increases in flooding (again see J. Hydrol 2006 319:83-95; and Kundzewicz, Z.W., D. Graczyk, T. Maurer, I. Piskwar, M. Radziejewski, C. Svensson, and M. Szwed. 2005. Trend detection in river flow series: 1. Annual maximum flow. Hydrol. Sci. J. 50:797-810).

  15. 65
    pat neuman says:

    re. 63, 31

    Tom Huntington,

    Have you reviewed the NCDC documents (titles below) as part of your work and conclusions in your paper “Century of Data Shows Intensification of Water Cycle but No Increase in Storms or Floods”?
    Are you intending to reply to other questions I asked, in 31?

    Observed Variability & Trends in Extreme Climate Events (PDF)
    Changes in the Probability of Heavy Precipitation (PDF)
    Heavy Precipitation & High Streamflow in the Contiguous U.S. (PDF)
    Precipitation Trends Over The Russian Permafrost-free Zone (PDF)
    Climatic Extremes & Weather Events (Web System)
    Worldwide Weather & Climate Events (Numerous Special Reports)
    The Global Climate of 1997 (Special Report)
    Climate Monitoring Reports & Products (Special Reports, 1998-present)
    U.S. Billion Dollar Weather Disasters (Special Report)
    Climate Variations Bulletin (Monthly Reports, 1994 to Present)
    Climatic Extremes & Weather Events (Web System)
    NNDC Climate Data Online (Interactive Long-Term Climatic Data)
    Storm Events Database (Severe Weather Reports, 1995 to Present)
    Historical Global Extremes

  16. 66
    Kenneth Blumenfeld says:


    Right. Everything you said.

    My definition of “extreme” is more “extreme” than most. I am really talking about the low-frequency, large return-period events (50-year, 100 year), for which it is very difficult to establish trends because of their inherent rarity. Here space (IMHO) is more important than time, and long data series won’t tell you as much as a bunch of tightly-packed 30-year stations. Hershfield (1961) warned of this in the precipitation frequency atlas (US Weather Bureau Technical Paper 40)…yet the focus is generally on getting longer records rather than using space as a surrogate for time.

    I am eager to read your paper and get some of the more recent citations also (my reading on this stopped in late 2003). I appreciate that a comprehensive, up-to-date review is finally out there. Thanks for doing it.

  17. 67
    Nicholas says:

    Lynn Shay has been working on Oceanic Heat Content for hurricanes at the University of Miami. Here’s the website link:

    It is interesting to go back and look at the ocean data last year. The loop current/eddy is very pronounced prior to Katrina, especially when looking at the depth of the 26°C isotherm.

  18. 68
    Hank Roberts says:

    Re “no evidence”
    I wish (pigs could fly and also that) whenever an agency comments on trends, said agency (or some commenter) would include something informative like:

    “Using statisticics, to say, at the 5% confidence level, that a trend is observed (either increasing or decreasing) requires collecting ____ observations in a year, or ____ observations in ten years, or ____ observations in a century. To find only an up _or_ down trend (but not looking for both)p requires half as many observations, or the number of observations made over half the time period.”
    [Check me on the one and two tailed test requirements, I’m barely an amateur here….)

    There are web pages that allow field workers to find the grim news — how many samples or transects they must obtain, over how long a period, to have any hope of detecting a trend.

    Almost no environmental restoration projects are done with proper baselines. I suspect that’s because — given a good baseline and followup — it’s possible to say for sure that nobody will know what worked for 50 years or so. If you don’t do the baseline you can just do handwaving “isn’t this pretty” to show off the project, ignoring the “and what was it like before you started” issues.

  19. 69
    Roger Pielke Jr. says:

    Gavin (repsonse to 63)-

    I pretty much agree, though one slight clarification to your comment about whether the prior consensus was “no global warming” or “no evidence of global warming”. My understandng as an outsider to this literature is that it would be more accurate to say that the immediate prior consensus that the effects of global warming would be real but undetectable. I don’t think anyone has said “no effect” although I recognize that “no detectable effect” might sound that way. Kerry Emanuel is as good as any a spokeperson for the consensus view of the field at that time, and he wrote in 2004 (!!!):

    “Can one detect an actual increase in global tropical cyclone intensity? “Since 1950″ one would expect to have observed an average increase in intensity of around 0.5 m/s or 1 knot. Because tropical cyclone maximum wind speeds are only reported at 5-knot intervals and are not believed to be accurate to better than 5 to 10 knots, and given the large interannual variability of tropical cyclone activity, such an increase would not be detectable. Thus any increase in hurricane intensity that may have already occurred as a result of global warming is inconsequential compared to natural variability.”

    This is a pretty strong, unambiguous statement, and illiustrates how quickly things have changed in this community over the past year. It also helps to put the comments of some scientists who receive lots of unwarranted criticism these days into better context. A consensus does not turn on a dime. The pro- and anti- global warming cheer squads might like it to, but that is just not how science works.

    The reality is that there are today different expectations about what future hurricane/climate research will reveal, and those of us not conducting primary research in this area will have to follow along as those who do hurricane/climate science do their jobs.

    I know of many papers currently in the publication queue that will add diversity to our understandings, and unfortunately no single consensus. I hope that RC will cover the diversity of perspectives on this issue with the same effort as has been focused focused on the Webster group. It is valuable in this instance, and will be in those as well.


    Emanuel, K., 2004. in Hurricanes and Typhoons: Past, Present and Future, R. J. Murname and K.-B. Liu, Eds., (Columbia University Press, New York), pp. 395-407.

  20. 70
    alisa brooks says:

    I realize that water vapor has been discussed at length (April 6, 2005), and I understood the argument that water vapor within lower altitudes had a minimal impact on GW. That being said, I was wondering if someone could point me to an article (or discussion) that examined water vapor within the mesosphere that has been contaminated by other greenhouse gases, and any impact this would have on overall global warming or increased hurricane severity.
    I know my naivete is showing, but I am very interested in learning!

  21. 71
    Steve Bloom says:

    Re #69: It should be added that KE held to that view right up until he saw the striking results from his study, and that this is when he dropped out as a co-author on your BAMS paper. The consensus may not turn on a dime, but KE’s views sure did.

    One point I want to emphasize is that Max Mayfield/NHC’s understandable focus on the North Atlantic tends to make them want to discount global trends. Add to this the obvious fact that, in the aftermath of Katrina, Adm. Lautenberger and the rest of the Bush administration would really, really, really prefer that U.S. government hurricane forecasters deny any possible current North Atlantic GW enhancement of hurricane intensity (just imagine those headlines), and I think it’s clear that we’ll have to wait until the next administration for a change of tune. The forthcoming season may change all of this, though. While it’s true that another record or near-record season that includes a major strike on the U.S. coast won’t mean much scientifically, the political implications are a different matter. [unnecessary personalisatiom deleted – moderator]

  22. 72
    Steve Bloom says:

    Re #55: Chip, the altimetry data sounds (to my amateur ears) like it’s exactly what is needed, although if the data is available I have to imagine someone is already working on it. OTOH, thinking about it a little more, the altimetry data might be less than perfect for these purposes if it incorporates warm water at depths that might put it out of reach of a cyclone. But maybe some combination of altimetry data with fore and aft SSTs? The task begins to sound large.

  23. 73
    llewelly says:

    Alisa: The mesosphere has very little effect on climate, mostly
    because it is optically thin. ‘Optically thin’ means that it has very
    little effect on light (including infrared light) passing through
    it. It is optically thin primarily because it has a very low
    density. Due to this low density, a light wave passing through the
    mesosphere will interact with very few molecules. Water in the
    mesosphere has very little ability to reflect infrared light back down
    to Earth, because it is so sparse. The greenhouse effect occurs when
    outgoing infrared radiation is reflected back to Earth. Without that
    reflection, the infrared radiation simply escapes, and the Earth is
    not warmed. Mesospheric gasses are also unlikely to have much cooling
    effect on Earth for similar reasons.

    Note: I don’t have the climate expertise many others here do. Corrections appreciated!

  24. 74
    Dano says:

    RE 68:

    Almost no environmental restoration projects are done with proper baselines. I suspect that’s because — given a good baseline and followup — it’s possible to say for sure that nobody will know what worked for 50 years or so.

    Hank, I enjoy your comments and I think you’re a real plus on this comment board.

    And I have an old GF who did env. restoration for a living, and for years I listened to her concerns (and did some fieldwork in her plots and did grunt labor when needed). Where do I start to explain your statement? The utter and complete lack of money for a typical project (except some Fed projects)? The likely no data on a particular site (pick a site, any site)? The change in focus to population ecology and less atomistic study since the 60s? Reg. changes?



  25. 75
    Tom Huntington says:

    Re: No. 65.

    I don’t know that I can answer all the questions that you pose. I am sure that I have not reviewed every document on you list. Please read the paper in question and see what I base the conclusion about no increase in floods on. Unlike in the case of the recent papers on hurricane strength, I don’t know of recent papers that report increases in flooding.

    The final wording of news releases are beyond my control. In this instance I argued for a simpler statement that was closer to the title of the paper. Frankly, I was concerned that what I thought was the more influential, other 95% of the paper, would be diluted with the other half of the news release title. I may have been overruled so that the agency could not be accused of choosing which results of the study to emphasize and which not to. Ultimately, I agreed to the wording that was used.

    I have not encountered any censorship in reporting on my research findings if that is what you are getting at.

  26. 76
    alisa brooks says:

    Thanks llewelly. I really appreciate the information!

  27. 77
    pat neuman says:

    re 75.

    Tom Huntington,

    I will read your paper on my next trip to the public library. The NCDC website which I made reference to in comment 65 includes a link (below) to the 2001 testimony by Thomas Karl, Director of NOAAâ??s National Climatic Data Center (NCDC).

    Karl’s testimony didn’t mention increased floods, but included these statements:
    “It is likely that the frequency of heavy and extreme precipitation events has increased as global temperatures have risen. This is particularly evident in areas where precipitation has increased, primarily in the mid and high latitudes of the Northern Hemisphere.”

    Before 2001 Karl and other scientists at NCDC did research and made public radio interviews on climate change and work related to heavy precipitation, as you can see by the reports referenced in 65, at link:

    However, in 2001 the research on climate change and the public interviews by NCDC staff stopped, for the most part. I conclude that NOAA headquarters stopped NCDC from doing climate change research, shortly after the change in presidential administrations. Based on what you said in 75, I wish more of us could have been with USGS during the last five years. I think it’s pretty obvious what I am getting at here.

    I’m including a few links (below) on floods in the Red River basin (northern US and southern Manitoba), and North Dakota’s previously closed drainage system called Devils Lake.

    Since 1995, Devils Lake has grown to nearly the size of a Great Lakes.

    Before 1979, floods in the Red River of the North basin were determined mainly by snowmelt runoff in spring. However, the big floods in 1979, 1989, 1997 and 2001 at Fargo, Grand Forks and northward were due to nearly equal portions of snowmelt runoff and rainfall, probably an increasing share being rainfall runoff vs snowmelt runoff, and the snowmelt runoff coming 3-4 weeks earlier (see yellow plot labeled Fargo Red at:

    Flooding and Climate Change in Southern Manitoba

    Devils Lake near Devils Lake, North Dakota:

    Lake Elevation for the Period of Record

    Satellite image, data sources,1,1,1,1,1

  28. 78
    pat neuman says:

    Off topic (but I don’t remember there being a RC topic on flooding).

    DEVILS LAKE OUTLET: Outlet remains issue in Canada

    Posted on Thu, Mar. 23, 2006
    DEVILS LAKE OUTLET: Outlet remains issue in Canada

    Manitoba premier discusses concerns with prime minister
    Associated Press

    ND State officials still plan to start up the Devils Lake outlet in a little more than a month, but Manitoba officials are still pushing for a more sophisticated water filter.

    BTW, Lake Ontario is similar to Lake Erie in length and breadth (193 miles by 53 miles).
    At 1459.0 ft above mean sea level, Devils Lake/Stump Lake spills into the Sheyenne River through Tolna Coulee, at a lake surface area of about 450 sq mi. So, Devils Lake is about 3 percent of the area of Lake Ontario.

  29. 79
    Lynn Vincentnathan says:

    Is what we’re talking about more “frequency of hurricanes (& floods) has not been proven, partly because of meager stats and/or a lot of noise,” rather than a flat definite statement, “there has been no increase in frequency of hurricanes (& floods).”

    This makes a big difference, because I’d like to see the proof at the .05 or less level of significance that there is NO INCREASE IN FREQUENCY OF HURRICANE (OR FLOODS). IE, can it definitely be proven beyond a shadow of a doubt that hurricanes and floods have not become more frequent? (I know science doesn’t work that way, so I don’t expect an answer here, but many lay people don’t know that science is erring on the side of not risking statements about associations…)

    The problem is the media run with, “GW is NOT increasing hurricane frequency or flood frequency,” when is should more accurately state, “scientists have not yet established with high certainty that hurricane and flood frequency is increasing, but climate scientists in general do predict such increases due to GW.”

    I would also add, “Many lay people pretty much guess these are increasing in frequency already.”

  30. 80
    Tim Jones says:

    WMO sees link between global warming and hurricanes
    Fri Mar 24, 2006 4:58 PM GMT

    GENEVA (Reuters) – There is growing evidence of a link between
    global warming and natural disasters such as droughts and flooding,
    the head of the World Meteorological Organization (WMO) said
    on Friday.

    But Michel Jarraud, secretary general of the United Nations weather
    agency, said more research was needed into the links between global
    warming and extreme conditions like hurricanes.

    Jarraud told a news briefing: “We know for certain that there is an
    intensification of the hydrological cycle, which translates into
    greater risk in some areas of a rain deficit and accentuated problems
    of drought linked to climate change.”

    “In other regions there is a higher risk of flooding and in others a
    risk of greater frequency of heat waves,” he said.

    The WMO said last week that greenhouse gases including carbon
    dioxide — blamed for global warming and climate change — had
    reached their highest levels in the atmosphere.

    Scientists warn that greenhouse gas emissions must be slowed and
    reduced if the earth is to avoid climatic havoc with devastating heat
    waves, droughts, floods and rising sea-levels sinking low-lying
    island states and hitting seaboard cities.

    Carbon dioxide, which the WMO says accounts for 90 percent of
    warming over the past decade, is largely generated by human activity
    involving the burning of fossil fuels.

    “We must accentuate research efforts to better understand the links
    between climate change and a certain number of extreme phenomena,”
    Jarraud said.

    He noted 2005 was a record year for hurricanes in the Atlantic Ocean,
    including Hurricane Katrina which devastated New Orleans.

    “There is not yet a consensus in the scientific community on the link
    between hurricanes and global warming, but there are leads. I am
    fairly confident that in two or three years we will have more
    credible answers,” Jarraud said.

    Research into the link between climate change and El Nino could take
    five years, he added.

    El Nino, caused by interaction between abnormally warm or cool seas
    and the atmosphere, typically triggers drought in eastern Australia
    and Southeast Asia, and floods in western parts of North and South


  31. 81
    Steve Sadlov says:

    Being a Pacific guy and not all that familiar with the Atlantic, is there something in the Western Atlantic analogous to the warm pool in the Western Pacific (during the La Nina phase of the ENSO) and to the El Nino “surge” that breaks out when the warm pool has become too massive? Like the Pacific I’d imagine that any warm pool’s build up is driven by the trades and by the wanderings of the ITCZ. One other thing I’d wonder is, of the high intensity hurricanes, which would be thought to be Cape Verde storms? Also, what is predicted to happen with Sahara dust over the time frame in question? And finally, I must wonder about what overall behavior will be expected in the Sargasso Sea over the next 10, 100 and 1000 years, based on the model used for this study, and other models.

  32. 82
    Hank Roberts says:

    Steve, I recall this recent article; found with Google, along with many more that appear related:

    Atlantic-Pacific Ocean Asymmetry and Past Global Climates II …
    This is thought to account for lower salinity in the North Pacific relative to the North Atlantic, which in turn drives northward flow through the Bering …

  33. 83
    Mike Atkinson says:

    With global warming increasing the sea area over which hurricanes may form and the length of the season in which sea temperatures are high enough for hurricanes, I would expect hurricanes to form further north and south in the Atlantic and over a longer season.

    Given the difficulty in measuring hurricane strength might it not be easier to detect a global warming signal in hurricane dates and tracks? If it is has any such signal been found? Do climate models make predictions about dates and tracks that may be tested?

  34. 84
    pat neuman says:

    re 79, 77

    Flooding and Climate Change in Southern Manitoba

    “Gage records for the Red River in Manitoba have been kept for a little more than 100 years. During that time, the valley has experienced 4 large floods (greater than 3000 m3/s): 1950, 1979, 1998 and 1997. The Red River valley has also experienced considerable climatic variability during the 20th century, including long term temperature increases of roughly 1.5 C. However the influence of these climatic changes on the flood risk in the Red River valley is unclear. Large floods occur very rarely and, given that only four significant floods have occurred in the Red River over the last century, it is difficult to determine if the recent increase in flood frequency represents a real response to regional climatic change or if it simply reflects random year to year fluctuations”….

    I wonder if the authors noticed that the 4 large floods all occurred in the second half of the century? Maybe they just figured that fact wasn’t significant enough to mention.

  35. 85
    David B. Benson says:

    Re #84 — pat neuman, under some simplifying assumptions, the probability of 4 floods all occuring in the second half of the century is nearly the same as the probability of tossing a fair coin 4 times and getting heads each time, 1/16 = 6.25%. I don’t think this is all that improbable myself, just “unlucky”. Notice that the report states “it is difficult to determine…”

    I opine that it is only by using all the records, wherever obtained, that one might be able to say something about extreme events.

  36. 86
    pat neuman says:

    … “On April 18, 1997, flow on the Red River at Grand Forks, ND, was 136,900 cfs, and the stage was 52.21 ft, which is more than 2 ft higher than the record set in 1897. The recurrence interval for this peak flow was between 200 and 500 years.” …

    … “The 1826 flood is believed to be the greatest flood in the Red River valley in the last 200 years. The 1826 flood probably was the greatest on many streams in the RRB. From the accounts of the 1826 flood, several conclusions can be can be reached: (1) the fall of 1825 was extremely wet and most of the lakes and wetlands were overflowing, (2) a major snowstorm occurred in late fall, (3) a cold, snowy winter permitted an exceptionally deep snow pack to develop over much of the RRB, and (4) the coldest estimated March-April mean temperature at Winnipeg since 1815, was in 1826.” …

    Excerpt from the Red River Basin Boards Hydrology Team Inventory Report-December 2000, Flood History of the Red River Basin

    Also, correction (re 84): 4 large floods on the Red River in southern Manitoba during the last 100 years are 1950, 1979, 1996 and 1997.

  37. 87
    Thomas Palmer says:

    I have one question, as a lay-person who’s just plain interested in listening to the discussions on RC…

    Has anyone along this thread mentioned the Thermo-Haline cycle in the Atlantic Ocean? I only ask because I’m curious what the science community (eg: y’all) have to say about it’s effects on hurricane intensity. It would seem to me that water that is slowing down in movement has a greater chance of holding solar energy at a much higher level. This, coupled with GW (no, not George W. Bush, but please, someone else MUST be thinking that every time they see that acronym), would increase the surface temp of the ocean at a faster rate, right?

    Maybe I’m way off with my questions, and I do not know nearly enough about the science of climatology to fully understand some of the discussion, but I’m a curious bystander, who is happy that discussions such as these on RC are taking place. As an educator in the public schools (albeit, an English teacher), it’s important for me to impress such information upon the students I interact with on a daily basis. Thanks for all of your collective hard work.

    -Thom Palmer
    Dover, NH

    [Response: We discuss this a little here: -gavin]

  38. 88
    pat neuman says:

    David, I posted 86 then read your 85. I wouldn’t disagree with you on the coin flip example except that I’ve reviewed monthly temperatures in ND and MN since 1880, and temperatures from 1820-2006 at Minneapolis (1903 to 2006 dewpoints). I’ve also done research on the Devils Lake basin. The 1826 very large flood in the Red River basin occurred during a May snowmelt runoff. Floods on the Red River were heavily influence by rainfall. The hydrologic cycle in the Red River basin has undoubtedly changed, regardless of what kind of statistics one chooses to go with on flood frequency and magnitude. Same with Devils Lake. How high will Devils Lake get before it levels off or falls (on a decade basis)?

  39. 89
    llewelly says:

    NHC definition of Cape-Verde type hurricanes. Note while it is ‘normal’ for an Atlantic hurricane season’s strongest storms to be Cape-Verde type, there are some notable exceptions, such as the 2005 season. Of 2005’s 4 cat 5 hurricanes, only one, Emily was a Cape-Verde hurricane. Katrina, Rita, and Wilma were not.

    With an average of only 2 per year, and a high variance – some years have none, some have as many as 5 – it would much, much harder to detect a climate disruption signal in Cape-Verde hurricanes, than it would be to detect a climate disruption signal in all tropical storms.

  40. 90
    pat neuman says:

    To clarify my comment in 88.


    The hydrologic cycle in the Red River basin has changed because of regional climate change (temperature and humidity increasing), especially winter-early spring. The 1800s-1978 floods were driven by late April and May snowmelt runoff. Beginning in 1979, the hydrologic cycle changed in that snowmelt occurred 2-4 weeks earlier and flood episodes became dominated more by rainfall runoff than snowmelt runoff.

    Same kind of changes are taking place in the Devils Lake basin except for the fact that the body of water has grown. A larger body of water warms and cools more slowly than a smaller body of water, which changes seasonal evaporation rates. Transpiration rates are increasing with the basins due to the longer growing seasons. The changing hydrologic cycle elements (rates of snowmelt and evapotranspiration, volume of rainfall, intensity of rainfall?) that are part of a changing climate and hydrologic cycle within the basin have been ignored by NWS hydrologists in their modeling, public probabilistic products and flood predictions. From Jan 2000-July 2005 I tried to change that, but failed.

  41. 91
    Kenneth Blumenfeld says:


    If you want to track down the coop observer sites in the basin (DL or RR), plus any other dense networks that may be out there( for example, the soil-water conservation districts often have gauges; the Future Farmers of America used to have a pretty extensive network too), I’d be happy to take a look at the data with you to see if indeed there have been significant changes in precipitation extremes. Anything on the MN side is available from the Minnesota State Climatology office. I think you have my email so we could take this off-list.

  42. 92
    stefan says:

    To conclude that anthropogenic warming substantially enhances the strength of hurricanes involves three logical steps:

    (1) Warmer SST significantly enhances the strength of hurricanes.
    (2) Anthropogenic warming significantly raises SST in the relevant regions.
    (3) Anthropogenic warming has no other effects (like reducing vertical gradients) which mostly or fully cancel the SST effect.

    To confidently claim that the rise in hurricane energy over the past decades is entirely due to a natural cycle (as the NHC has done), you have to be confident that at least one of the three points above is wrong.

    Hurricane forecasters like the NHC people know a lot about (1), much more than I do, and they don’t seem to question this statement (note that the statement is not that SST is the only or dominant effect – just that it is important).

    Point (3) is a possibility raised e.g. by Isaac Held in our comments section last year, and it can and needs to be discussed by serious researchers. However, at this stage there is little evidence to suggest that the SST warming effect is compensated in this way, and the data of Emanuel (2005) and Hoyos et al. (2006) speak against it. Therefore, we can’t rule out that (3) is wrong, but I don’t think that anyone could claim with any confidence that point (3) is wrong.

    That leaves us with point (2), and indeed it is the one questioned by those who claim the rise in hurricane activity is natural. They claim that the relevant increase in SST is entirely part of a natural cycle, namely an increase in thermohaline ocean circulation in the Atlantic. As an oceanographer I do not think this is a supportable scientific position; it certainly is not a position that was consensus just a couple of years ago (see comment #63), and it goes against everything we know about global warming. This position would require that the effect of greenhouse gas warming somehow magically leaves out the tropical Atlantic (no model suggests this), and that the tropical Atlantic by chance has naturally warmed just by the same amount as the rest of the global SST but for a different reason. Have a close look at the comparison of tropical Atlantic SST changes with northern hemisphere temperature changes by Kerry Emanuel here. Or look at a global map of SST trends over the past decades: you don’t see anything special going on in the tropical Atlantic – similar warming happens in the tropical Pacific and Indian oceans, and most of the rest of the world. Also, the supposed increase in thermohaline circulation is not a measured phenomenon (quite unlike the rise in greenhouse gas concentrations) so one cannot be very confident that it is even happening, let alone that it is the sole effect that explains the SST rise in the tropical Atlantic. Therefore, in my view the press conference of the NHC, where they claimed with confidence that the rise in hurricane activity is natural, will be remembered as one of the darkest hours in the history of NOAA.

  43. 93
    pat neuman says:

    re 90 91

    Surface and interflow runoff are the largest components of runoff contributing to river flooding. Baseflows and reservoir regulations affect streamflow to a smaller degree. Antecedent storm conditions influencing the amount of surface and interflow runoff include: states of soil moisture and soil temperature, vegetation and snow cover, and the intensity, amount, and type of precipitation from the storm.

    In the Upper Midwest, climate change is increasing the amount of surface and interflow runoff because heavy rainfall is occurring earlier in the spring, before the vegetation is established. Establishing vegetation depends on the amount of daylight and sun angle, which won’t change with time, and also depends on day-to-day air temperatures for a season. Rainfall from February to mid-May has been increasing before vegetation gets established, leading to larger degrees of surface and interflow runoff, more erosion and more flooding. Based on operational experiences … not allowed during my career with NOAA NWS NCRFC to do research for peer reviewed papers that were related to political and controversial subjects like regional climate change and global warming.

  44. 94
    Lynn Vincentnathan says:

    RE #85 & coin-flips. You certainly cannot disprove that the coin is loaded for heads, esp if 4 came up in a row.

    So what is better: assume GW is not happening & not causing increase havoc, & do nothing; or think that GW might be happening & wreaking havoc, and, say, turn off lights not in use, etc.

  45. 95
    pat neuman says:

    re 93. should read “Establishing vegetation depends on the amount of daylight and sun angle, which won’t change with time” from year to year”, …

    Also, for example of discussion in 93, may see:

    “The Spring Flood Outlook products are seasonal products issued in February and March which provide forecasts of the flooding expected due to snow melt for river basins within the North Central River Forecast Center area of responsibility.”


    “North Central River Forecast Center Long Range Probabilistic Outlook Summary Valid February 23, 2006 – May 24, 2006


    Featured Products from NOAA/ National Weather Service
    National Centers for Environmental Prediction
    Hydrometeorological Prediction Center

  46. 96

    Re #49 and “Does all of this really matter? AGW is happening and mitigation activities are needed regardless of what is going on with hurricanes. Shades of the hockey stick!

    Every intuition tells me that GW should affect hurricanes. However, my intuition is not science. I am concerned with the possibility that some of the very excellent scientists on this site may have foregone their science and succumbed to a similar intuition.”

    BPL replies:

    I took the satellite temperature anomalies (differences from mean world temperature over the period) from December, 1978 to August, 2005. My source was NASA’s Global Hydrology and Climate Center (on the web at

    This gave me 321 data points. I divided them into two groups of as nearly equal size as I could get, December 1978 to March 1992 (N = 160) and April 1992 to August 2005 (N = 161). I calculated descriptive statistics for each group:

    number: mean: std. dev:
    160 -0.0230 0.106
    161 0.167 0.216

    The equality of two standard deviations can be tested for with a Fisher’s F test:

    F = s1^2 / s2^2

    where s1 is the larger standard deviation and s2 the smaller. The respective sample standard deviations each have N – 1 degrees of freedom. In this case, I found F(160,159) = 1.67. For the sample sizes involved, this result is significant at the 99.9% level. It is overwhelmingly likely that the later period is more variable than the early period.

    As the atmosphere heats up, it carries more energy, so we can expect weather to be more variable.

  47. 97
    pat neuman says:

    re 91.


    To look at changes in precipitation intensity, we’d want hourly time steps from radar/gage mosaic-ed HRAP cells (idea used to justify money for NWS AHPS).

  48. 98
    David B. Benson says:

    Re #95: Lynn, my point was only that attempting to make predictions based on only 4 extreme events doesn’t give one much confidence. My apologies for not being sufficiently clear about that. BPL’s #96 involves enough data to be meaningful. Even better, see all the commentary regarding Tom’s paper in J. Hydrology.

  49. 99
    pat neuman says:

    re 96. 98.


    As the atmosphere heats up and becomes more humid, variability in average temperatures is less (not more), based on climate station data since 1880 in the US and shown by time series plots at:

  50. 100

    Stefan, about point 2:
    As far as I have read, in recent years it seems that the THC has slowed down, which may lead to an increased area of higher SST’s in the (sub)tropical North Atlantic (due to less poleward heat movement), which may increase hurricane frequency and intensity (for the latter, if there is an increase in temperature in the deeper oceans), but the temperature gradient between tropical and extra-tropical SST also may play a role too (which may be a part of point 3).
    If the slowdown of the THC is a result of increased GHGs or part of a natural cycle is a matter of debate, but more important is that there is another natural component that has a huge impact: the change in (low) cloud cover in the (sub)tropics in the past decade(s).

    According to Wielicki ea. and the accompanying Chen ea. paper, there is an increase of ~2 W/m2 in direct insolation over the whole (sub)tropics in the period 1985-2000. This is confined to the tropics (30N-30S), while the influence of GHG’s is more evenly distributed over the whole earth’s surface. The increase in ocean heat content is specifically high in the NH subtropics, see Fig. 2 in Levitus ea., and coincides with the decrease in cloud cover (and the increase in insolation), which points more to cloud/insolation influences.
    Thus even if the NOAA points to the wrong cause, natural influences can’t be ruled out and may be more important than increased GHGs.

    It may be interesting to look in more detail at the (deep) ocean temperature trends and the cloud cover trends in the main area of tropical cyclone development in the North Atlantic…

    [Response: One of the nice things of our RC website is that the layperson can lecture the expert who’s been working for his entire career on THC. But let me return the favour: your idea that the North Atlantic tropics warm up if the THC slows down is a common misconception, see the graphs in this post. The idea of a number of hurricane researchers, namely that the tropics warm up when the THC speeds up, is physically correct, except that this cannot explain much (let alone all) of the observed warming of the Atlantic tropics of the past decades. For many papers on THC and its effects on climate, see the publications list on my home page. -stefan]