- Image by Getty Images via Daylife
Here’s a personal beef about a small point in the hurricane debate which CA readers may recall, which came to mind out of the EPA finding and Technical Support Document, which have a LOT to say about hurricanes and which rely relying on CCSP 3-3. (Perhaps I’ll review this on some occasion.)
CCSP 3-3 reviewed a 2005 dispute between Emanuel and Pielke over the seeming inconsistency between the lack of trend in US landfalling hurricanes and increased basinwide trends. Emanuel (2005b) responded to Pielke’s criticism by arguing that the US landfalling data set was only 1% of the total data set and that greater reliance should be placed on the larger data set.
In comments at CA in 2007, I observed that this effect was not limited to the US landfalling dataset. When the entire HURDAT set (the one used by Emanuel) was analyzed in spatial quintiles, there was no trend in a range of statistics (hurricane-days, storm-days) in the western quintiles adjacent to landfall, with the entire increase taking place in the eastern and mid Atlantic remote from habitation.
This was written up by Roger Pielke and I and submitted to GRL. Here is a relevant quote:
Upward trends in storm-days, hurricane-days, and power dissipation occur only in the east Atlantic (east of 73W), while west of 63W, these metrics have no trend or declining trends, consistent with a similar lack of trend in U.S. landfall …
Emanuel 2005b argued that lack of a trend in U.S. landfall PDI (identified by Landsea 2005) corresponding to the overall increase in NATL PDI was likely a random fluctuation, suggesting that the HURDAT track data contained “about 100 times more data” than the landfall data set and that his results accordingly had “a signal-to-noise ratio that is ten times that of an index based on landfalling wind speeds.” In other words, the landfall data might simply reflect the randomness of a small subset of the overall HURDAT basin data. However, our analysis shows that there is no inconsistency between the lack of landfall trend and the lack of trend in the western quintiles using the same HURDAT data employed by Emanuel.
Reviewers were violently opposed to the article, stating on the one hand that the statistical analysis was “fraudulent” and on the other hand that the results were already well known in the literature. (I think that Holland of UCAR, whose work was criticized) was one of the reviewers.) The editor, Famiglietti, said that there was a consensus and rejected the article without offering any opportunity for remediation.
I was quite shocked by the reviews and sent the correspondence to Emanuel. Emanuel replied very cordially:
I do think you have found something interesting that deserves to see the light of publication, and you should try again. I know that Christina Holland and Rob Scott at U. Texas, who found a similar eastward trend in genesis, have also had great difficulties getting their paper published…I do not know if they ultimately succeeded. You might try the QJRMS..in my opinion a very good journal, which being based in England may be far enough from the craziness to get a fair hearing.
We also sent a copy to Jim Kossin who also replied cordially:
Roger/Steve,
Please keep me posted on the progress of your manuscript dealing with the eastward shifting of activity. I’m revising our AMM/genesis region manuscript and I’d like to reference your paper if possible. I am referencing Landsea’s latest Eos paper, so either way, the issue of missing data versus climatic modulation will be addressed with appropriate citations, but it would be good to include your paper too for better balance.
Regards, Jim
Roger suggested submitting it elsewhere, but by this time I was working on other topics and didn’t pursue it. Many people criticize me for not “publishing” more, by which they of course mean “publishing more” in the academic journals as, of course, it’s not like I’m just scratching notes to myself; I “publish” something on the blog nearly every day, though I realize as well as anyone that the blog corpus is unruly and unindexed.
However, I also take the position that I’ve pretty much dropped out of left field into the climate debate. If my observations are correct, then they are correct whether or not I publish them; there are lots of smart people in the world and climate scientists with their billions of dollars of research funding should be able to get these things right whether or not I publish something in an academic journal or not. I realize that I might get more personal approbation if I did so, but I derive a lot of personal enjoyment out of investigating new issues and, last time I looked, personal enjoyment was why I’m doing this.
Anyway, on to CCSP 3-3.
The above small issue gets a mention in CCSP 3-3, which adopted the signal-noise argument of Emanuel (2005b) as follows:
The Power Dissipation Index for U.S. landfalling tropical cyclones has not increased since the late 1800s (Landsea 2005). Pielke (2005) noted that there are no evident trends in observed damage in the North Atlantic region, after accounting for population increases and coastal development. However, Emanuel (2005b) notes that a PDI series such as Landsea’s (2005), based on only U.S. landfalling data, contains only about 1 percent of the data that Emanuel’s (2005a) basin-wide PDI contains, which is based on all storms over their entire lifetimes. Thus a trend in basin-wide PDI may not be detectable in U.S. landfalling PDI since the former index has a factor of 10 advantage in detecting a signal in a variable record (the signal-to-noise ratio).
They did not refer to the result of our article observing the incorrectness of the Emanuel 2005b signal-noise observation, said by the GRL reviewer to be “well known in the literature”. BTW, lead authors of the relevant chapter of the CCSP Report included both Kerry Emanuel and Jim Kossin, both of whom had read and seemingly approved of Pielke and McIntyre (2007 rejected.)
Convening Lead Author: Kenneth Kunkel, Univ. Ill. Urbana-Champaign, Ill. State Water Survey
Lead Authors: Peter Bromirski, Scripps Inst. Oceanography, UCSD; Harold Brooks, NOAA; Tereza Cavazos, Centro de Investigación Científica y de Educación Superior de Ensenada, Mexico; Arthur Douglas, Creighton Univ.; David Easterling, NOAA; Kerry Emanuel, Mass. Inst. Tech.; Pavel Groisman, UCAR/NCDC; Greg Holland, NCAR; Thomas Knutson, NOAA; James Kossin, Univ. Wis., Madison, CIMSS; Paul Komar, Oreg. State Univ.; David Levinson, NOAA; Richard Smith,Univ. N.C., Chapel Hill
I’m mulling over re-submitting this article, updating to 2008 – which would please Roger enormously. This might be the easiest of my many pending chores. I’m also reflecting on whether an author has a confidentiality obligation in respect to review comments and may visit this issue separately.
![Reblog this post [with Zemanta]](https://i0.wp.com/img.zemanta.com/reblog_e.png)
Climate Audit 
77 Comments
In my experience, when one has made a mistake or done an incomplete job on a paper that is a regular paper, one gets very detailed but mostly polite comments with citations, corrections, etc. BUT if you stray into theology-land (not just on climate, try spotted owls or women in engineering or lead paint), then you can almost see the frothing at the mouth of the reviewers. Really, reviewers should get a grip on themselves…
I appreciate that it is a bore resubmitting to different journals, but every academic has to do it. I think you have to accept that for any one individual journal, you can get biased reviews; and editors get so many papers, they are happy just to be able to reject. The main thing is to get it out in the peer-reviewed press, because no matter where it is, it must be considered when it is published !
I would not place any emphasis whatsoever on the fact that people have been kind to your face; the whole point of anonymous peer-review is that you can reject someone’s paper (on scientific grounds) without fear of retribution, or personal confrontation.
I appreciate that you do this work on your own dime, and I would be delighted if you published more in the peer-reviewed press. That way, your work could only be overlooked with considerably more difficulty.
best wishes
per
Re: per (#2),
Per per, a man who knows, and I agree whole-heartedly with this comment.
And “with more difficulty” does not mean impossibly or even improbably.
Re: Steve McIntyre (#5),
Which would be more satisfying and fun, the young way or the current way?
Re: per (#2), Relative to published work being less easily overlooked: not quite so. Being published makes it more official when someone cites you, but IPCC has been quite able to ignore published work that don’t “fit”, including Steve’s publications.
Rejection of manuscripts happens all the time; sometimes you just have to persevere. One approach I have found helpful with “controversial” material is to submit to a minor but solid general purpose journal: specialist journals can be too wed to the consensus, and high status journals can be over picky. Being completely open with the editor about the controversial nature of the conclusions can also help, as long as the result is simple and clear.
Maybe it would be easier to get it accepted if you could get Emanuel to co-author it with you and Roger.
There are lots of things that I could have done to “get it accepted”, like re-submitting somewhere else. If I were young and depending on these things to get ahead or if I’d been publicly funded on the basis that I would publish in an academic journal, you can be assured that I would have. But I’m not.
As I said above, people in the field are amply funded to get things right and that’s their responsibility, which shouldn’t be in any way dependent on me.
Re: Steve McIntyre (#5),
Clearly, they feel no responsibility to “get things right”. If they go unchallenged they can claim consensus.
Perhaps worth considering resubmission part of your ongoing set of experiments demonstrating strengths/weaknesses of The System. A bit like our test of the Starbucks Hypothesis.
In this case, poking at the Peer Review system to see if an unusual yet accurate and not really inflammatory paper can make it into The System. As several have noted, one rejection doesn’t say much one way or ‘tother.
Re: Publication
Have you thought about dropping a version of the paper
into a preprint repository like the Los Alamos arXiv?
They have a section on Atmospheric and Oceanic Physics.
The advantage of the ArXiV is that it is an institutional
repository and so the article can be referenced with the
article URL. It is very common to drop manuscripts onto
the arXiv and this in no way prejudices future publication
in a regular journal (at least in physics). The general
quality of articles on the arXiv is quite good.
Things in the arXiv are not refereed, but it is
understood they are working their way through the
system. This gets the article into circulation.
Re #s 2,3,4,7,8 on the subject of publishing in the peer-reviewed journals:
Here’s what Steve said:
Let Steve be Steve. He’s providing an invaluable archive of data here, which is what distinguishes him from the so-called climatology experts who receive favorable treatment from the science journals because they are approaching these issues from a conformist point of view. The research money is not now flowing to those who hold contrarian, inconvenient views. It will take time, several years perhaps, before that changes and people like Lindzen, Christy, Spencer and yes, McIntyre are given the funding and/or the column space to present what appears to be the more rational viewpoint.
It’s not Steve’s responsibility to rescue climatology from itself. It’s the responsibility of those who practice in the field. All of them. Steve has shown that their methods, if not their ethics, are at best suspect. That in itself is a worthy contribution to this debate, which goes to the very core of how human beings should conduct themselves on earth.
As an outsider, I am not sure that the freedom to anonymously reject a paper, without fear of retribution or personal confrontation, is a worthy one. If the reviewers are entitled to anonymity, perhaps so are the authors during the process. Specifically, making anonymous and apparently unsubstantiated accusations of scientific fraud is not ethical in my view.
“..far enough from the craziness to get a fair hearing.”
? I’d like some elaboration there.
One thing that bothers me in the whole Hurricane debate is that increase damage amounts are being used to show increased intensity of the storms. If we had another Hurricanes that wiped Galveston back to the sand bar as the early one did, would the increased dollar amounts of damage show any change in the intensity of the storm that produced it?
By the way, if you count the original analyses Steve has done here with some sort of metric (autopsies performed? gaffes uncovered? stats audited?) he is one of the most prolific scientists around. Just saying.
Re: Craig Loehle (#14), Yeah, I mean, it takes a whole Team to try and keep up with him. 😉
just thought I would point to some of the comments made by referees.
Strangely, the finally published title and conclusions were substantively unaltered…
The editor also formed a view on the arguments, and I suspect that the Editor’s view on the arguments informed his view on this referee 🙂
One of the real difficulties is having an Editor who understands the issues. If the Editor does not understand, then they have no way to arbitrate or evaluate the arguments.
per
Re: per (#16),
Per, you are indeed a funny man. The layperson equivalent of those comments would end with “and the horse you rode in on”.
Re: per (#16),
Per:
The editor means everything, absolutely everything. Just think of Nature during the Maddox years and today.
Maybe we need to find someone to act as Steve’s secretary to pursue these submittals.
snip – [Steve: while I appreciate the support, it’s a little grandiose]
The science on this blog — that which I can understand — is prolific and amazing. It stands on its own and Steve should do whatever he wants with his time; whether young or not, there is never enough of it. That being said, I wish Steve had a research associate who would take not only the rejected article, but the many more here just waiting for a greater readership, and do the basic work of resubmission and submission all with the author(s) oversight. Steve, your work is a valuable contribution to climate science and especially to the accurate modelling of it. Thank you.
Steve,
Could you put a link to the 2007 remarks you mention above? And any other thread which is relevant?
Since I’m sure far more than 1% of the NA storms make landfall, I’m trying to see how the claim by Emanuel could have any sort of prima facie validity. Sure he could have as much data on as many storms as he wants, but the landfall data doesn’t have to have as much data to evaluate the storms. It just needs a (relatively) few pieces of data for each storm which touches land. Having huge amounts of data only makes it easier for bias from improved methods of data collection to make the conclusions invalid. I believe this was discussed in detail in the previous threads.
“All that is necessary for the triumph of evil is that good men do nothing.”
(Edmund Burke)
Pause a bit on that sentence. Let it sink in. Then tell me that it’s alright no one publishes McIntyre’s findings because, oh we’re on the intertubes!
Re: Luis Dias (#23),
Atlas Shrugged.
Re: Kenneth Fritsch (#27) said “Atlas Shrugged”, so said a philosopher that most other schools of philosophy do not recognise today, 80 years after she founded the Objectivist school.
I would caution Steve and everyone else here to not take the GW skeptics into that ghetto. Play the game as it is set up, find journals that are more open, and publish negative reviews of those that are not.
Re: Mike Lorrey (#30), A wise man once said, better to be right than President. Another responded he would “never be either”.
I put this wrongly on the thread below. It belongs here.
If Steve had chosen from the beginning only to make public his efforts, together with Ross and others, in the peer reviewed journals, and even assuming a portion of it had been accepted, my guess is that the majority of individuals, including many of the scientists, who visit this site on a regular basis, would never have heard of him or his work. As an attorney, I certainly never would have. Think what would have been lost.
What effect would it have had on the Team or the IPCC? Probably nothing. As it is, his work has been and is being widely disseminated and a hell of a lot of people know about it, including the climate science community. I think he made an extremely wise choice
I take it the Famiglietti that rejected you is the same editor of “Jesus” paper fame?
Recently, and this includes my day to day work, I can see the merit of the peer-reviewed system and how it relates to science. Basically, if your paper passes a continuity of argument check then it should be submitted. The conclusions may be incorrect and unrepeatable, which will then mean your paper doesn’t remain high on the citations list (unless its being cited as an example of a bad conjecture/results) or it may get repeated and be independently reinforced. The point is that it is out there for debate and possible repetition.
So with regards to Steve’s experience, it is a little annoying that a paper didn’t pass through because a reviewer appeared to mistake the ‘is the content and argument presented in a logical way’ benchmark for the ‘is this actually correct in its conclusion detail’. That is not the job of the reviewer; its the job of the scientific debate. A balance has to be struck.
As a last thing, and I have experienced this in my own career in physics and engineering, many people use the argument that if its peer-reviewed then automatically the findings are correct. No, this is wrong.
It just means that your argument is out there in a forum that is considered by convention and consensus to be the principal forum for scientific debate. Though it doesn’t mean that science needs to be disseminated through this process alone, it does have more impact going through the peer-review process.
I suspect that it is the rule rather than the exception that over history the scientifically minded sceptics and dissenters have not and do not get published in the recognized organs of science. But it is precisely those deniers and dissenters who over the years have made hundreds, perhaps thousands, of historically important scientific discoveries. After all, “discovery” describes what happens when an individual finds or finds out something that nobody else has noticed, figured out, or believed in.
If this sounds romantic, so be it. Living in the world ought to be a romance.
Please – while I appreciate the support and advice, readers are assuming that I’m complaining about being denied an outlet. That’s not what I was doing. Yeah, I was a bit annoyed with GRL, but I understand full well that there’s more than one way to skin a cat and until I made a more concerted effort to place the article, I’m not complaining about that – though I may at some time use the reviews in a thread on reviewing.
The issue that I’m interested in here is entirely different.
Our email correspondence indicated that two CCSP coauthors accepted our counter-argument on the signal-noise argument of Emanuel 2005b and thought that the GRL reviews were unreasonable. For the sake of argument, let’s stipulate that their email correspondence represented their actual thoughts and they weren’t tricking us (and I have no reason either from the tone of the correspondence or the reputation of the individuals to expect anything other than candor in their correspondence.)
Is it still acceptable for them to acquiesce in the problematic signal-noise argument of Emanuel 2005b in the CCSP review?
Yeah, it would be easier for them if the article had appeared somewhere. But what are they supposed to do if they know that there’s probably a defect in this part of the argument without the article having appeared anywhere?
Steve, please take note of the comments of per and Jonathan #2 and #3 (and Roger, Jim and Kerry!). It is quite common for papers to be rejected in controversial fields, particularly for ‘letters’ journals. GRL has a fairly high impact factor and rejects a lot of papers (though this is hard to believe in view of for example the acceptance of Easterling and Wehner). The best response is to revise the paper and submit it to a different journal. And if you write a criticism of Dr X in your paper, be aware that the editor is quite likely to choose Dr X as a referee. Also you need to be aware of the attitude of the editor of the journal. Sad but true. I’m sure there are many who would like to read your paper. So please do resubmit it, and put it on ArXiv in the Atmospheric and Oceanic Physics section.
I’m afraid I must disagree with Kerry Emanuel’s remark “which being based in England may be far enough from the craziness …” – there is just as much craziness here, if not more, as a glance at the web pages of our formerly respected scientific institutions such as the Royal Society and the Met Office confirms.
Steve: Once again, I’m well aware of issues in paper acceptance. And once again, I’m not “complaining” or seeking advice on what to do. I’m inviting to people to comment on potential obligations of CCSP authors aware of probable but not journal-published defects in parts of their argument.
I’m mystified by Emmanuel’s claim that the S/N ratio of his data set is 10 times greater than that of others. Since that ratio is purely a function of the measurement apparatus or method–and not of sample size–is he actually claiming such superiority? In actuality, it appears he’s just using that term in an empty display of “erudition.”
Re: sky (#33),
the issue is different. It’s the observation that landfalling stats are “consistent with” a wide variety of hurricane indices in the western quintiles, with the overall increase being contributed entirely by increases in east and mid-Atlantic. NOTHING to do with signal.
Re: Steve McIntyre (#35), Surely it should be possible to estimate how likely such a thing as an increase in hurricanes but not increase in landfalls? It just comes across as very implausible.
Great stuff here.
Steve, I hope you are taking care of yourself because if something happens to you, the rest of us are in trouble.
Thanks again for all you do. I am forever indebted.
I am unaware of any confidentiality issues on reviewers’ comments, except to preserve the identity of reviewers who have not chosen to id themselves.
James Lindgren
Professor of Law
Northwestern University
I do not think that these authors are comfortable with comments required for refutation outside the realm of peer-review and most often peer-review publishes positive results and seldom negative ones or critiques. The peer-review system would appear not so much interested in going back and pointing out past errors, but in moving on with ever more papers that might or might not at some point in time make the science forget the past errors.
Then we come to the review processes of peer-reviewed papers like the IPCC and CCSP who can in the end selectively use any peer-reviewed data they see fit – there is no objective algorithm to apply.
So what is the alternative for thinking people? Do your own analyses and reviews or find trusted parties to do them for you. I guess the only personal obligations in these cases would have to be self-imposed and I strongly suspect that those obligations resulting in action or reaction will be more oriented towards policy than science.
Steve, I think most of the studies in the hurricane/climate department are positive result studies, and are likely reflective of positive-outcome or publication bias. While this may be partly an issue with the journal, it is usually a function of researchers not submitting work (such as your Landfall non-signal) and leaving it in their desk-drawers. With the media’s religious devotion to the AGW meme, researchers are more likely to avoid publishing contrary results, and journals are less likely to see such submissions — a positive feedback.
Re: Ryan Maue (#39),
A perceptive observation IMO.
Re: Ryan Maue (#39),
Hence scientists’ search for metrics that confirm the AGW meme, even if they lack face validity. %Cat 4&5 Hurricanes is a pretty good example.
Re: Mike B (#43), the Webster et al. (2005) paper really says nothing about global warming causing the increase in the Cat 4/5 ratio. As there was no rigorous attempt to ascribe causality back in 2005, now 4-years later, the paper’s statistical findings can only be described as “not-inconsistent” with other research presented since. I believe Dr. Curry on other threads has intimated that a more robust explanation for the statistical findings is forthcoming in as-yet-unpublished (or unsubmitted) papers. This area of research is crucial to understanding future changes in TC intensity in the different basins on the globe, and I look forward to reading any new scripture on the subject.
Re: Ryan Maue (#44),
Ryan, what I read into Dr. Curry’s comments (particularly several back-and-forths with Kenneth F.) is that her group has found a trend in Cat 4/5 ratio. Now I don’t know if the denominator of this ratio is total hurricanes or total named storms or something entirely different. In my view any ratio of this sort is invalid on its face. What you’ve presented using ACE is far more comprehensive and makes much more sense.
Clarifying my point, the metric should be chosen a priori on the basis of a) validity for measuring total TC activity, intensity, and liklihood of damage and b) statistical sensitivity. ACE is a hands down winner over %cat4/5 on both criteria. Then let the data tell the story.
I think there is a problem, because there are an infinite number of things that may be wrong with a piece of research, or may supercede the research. I think that there is an important role for the gatekeeping function of peer-review, because that limits your commentary, or what you must cover.
There is actually quite an interesting literature on this issue, because there are a number of papers which get (tremendously exciting) result A. These papers are then followed up by proper papers, which show that (not A) is in fact correct. Yet frequently, the original A paper is the paper that is referred to, or cited, and not the subsequent papers that show it to be wrong. (Title: Persistence of contradicted claims in the literature Author(s): Tatsioni A, Bonitsis NG, Ioannidis JPA Source: JAMA-JOURNAL OF THE AMERICAN MEDICAL ASSOCIATION Volume: 298 Issue: 21 Pages: 2517-2526 )
So a paper that is not in the literature is even more problematic, and even if the author of a review knows about an unpublished paper, how can they communicate that to their readership ?
per
Re: per (#40),
per, I understand the point. However, I’m trying to consider what I would do if someone had sent me a paper that raised a relevant issue about a statement that I’d made, as we had done with the SNR argument in Emanuel 2005b. Personally, in either situation, I wouldn’t keep making the original statement without caveats even if the other guy hadn’t placed his observation in the peer-reviewed literature.
There are shades here and I don’t know whether Emanuel agreed with our point or thought it was interesting and worth publishing, without taking a final position on the matter. And by reputation (and from my brief emails), Emanuel is very cordial. It’s not that I feel slighted here, I’m just mulling over how people should handle this sort of situation.
I would personally have done something different than was done here. In business prospectuses, you can’t ignore information because it isn’t generally known (this is what gives rise to causes of action against people who breach these rules.) So my perspective on such obligations is different than an academic’s.
The climate science literature is littered with references to (submitted) and even (in prep) citations and even workshops. One alternative would be to email the guy, tell him that I wanted to refer to the matter and ask him to re-submit it somewhere so I could quote it.
In this case, I’d presented the point at an AGU conference. Another alternative would be to cite the AGU presentation. Maybe a pers comm. In the CCSP case, an option would have been to avoid the argument entirely if one were aware of a caveat against the argument that one couldn’t cite.
Re 43 & 44 above:
As I got Curry to admit about her papers with Webster, their data show no increase in cat 5 hurricanes, only cat 4 hurricvanes. See: http://volokh.com/posts/1157158094.shtml
– snip-
Steve: Judith Curry posts here from time to time. Please restate your point in better accordance with blog policies.
Hi Steve
“Yeah, it would be easier for them if the article had appeared somewhere. But what are they supposed to do if they know that there’s probably a defect in this part of the argument without the article having appeared anywhere?”
This is always problematic. I personally find it a pest to try
and reference non-published work, for the simple reason that no-one
else can read it.
So lets suppose there is a point of contention about your work
(I am being neutral). You want to refer to this point, maybe
to agree, or disagree, or partly agree/disagree and elaborate.
It really is a nuisance if article is not published in some form.
OK, information has been published in the blog, but how do readers
who want to dig into the point of contention dig the information
out of the blog. But blog discussions can be rambling, and it might
be necessary to wade through a lot of dross and extraneous material
to get the complete picture. Furthermore, the blog is not
institutional, maybe the economic crisis means SMs retirement
fund evaporates, so SM has to go back to his old day job, and
the blog stagnates and disappears.
One of SMs gripes has been the inability to get at raw data
used in analyses. And a common theme is that the best place
for this data is an institutional repositry. So, Steve has
has an unpublished paper, and it is a hassle to get it
published. Why not drop it into an institutional repository
like the arXiv so everyone can read it? No refereeing,
no fuss, but with a small amount of effort required to
learn how to do the mechanics of uploading.
So if I am author, and I know there is a point of contention.
Maybe I do some more work on my own, and think this point has a
flaw but is nevertheless valuable in the interplay of ideas.
So what to do? I might ask the person to place it into an
institutional repository. If they did not do that, then may be
it is simply not worth bringing up the point of contention in
the first place. Or I could really stir up the antheap by
asking if I could include the unpublished manuscript in
supplmentary material. And all of this depends on whether
the point is central, or a peripheral issue. There is a
very large patch of grey area for this topic.
BTW, about the impact of blogs. I have hit “McKitrick McIntrye”
with google scholar. The 2004 GL pdf proof at climateaudit gets 74
citations. The 2003 Energy and Environment gets 83. The 2005
E&E gets 34. Google scholar has some vagaries, but these are
eminently respectable citation counts. It is interesting the google
scholar points to the Climate-Audit proof pdf and not to the actual
GL article. So maybe blogs are better than regular articles
after all.
Mike, you are missing the point. The hypothesized link between global warming and hurricane activity is in the increase in intensity, which is reflected specifically by a greater proportion of hurricanes reaching the status of cat 4,5. No one is hypothesizing an increase in total global number of hurricanes, and no one is hypothesizing an increase in the number of hurricane days. The problem with ACE in terms of relating it to global warming is that it is dominated by number of TCs and number of hurricane days. Further, you need to square the wind speed, magnifying any error in wind speed.
The metrics are chosen based on physical understanding and theory, in this instance related to hurricane intensity. The data has told the story, over on the other active hurricane thread, and Ken has confirmed our calculations of increase in % cat 45
No one has presented a theory as to why ACE in particular would be expected to increase with global warming. So the fact that it doesn’t increase, tells us exactly what re the impact of global warming on hurricanes? nothing.
Re: Judith Curry (#48),
-snip-
The time series for annual global Cat45 counts using the IBTracs data series shows a positive trend significantly different than zero from 1981-2007 that drops off to not significantly different than zero from 1984-2007. Using the ratio of Cat45 counts to total TC counts with the same data shows a positive trend from 1981-2007 significantly different than zero that drops off to not significantly different than zero from 1990-2007.
The reanalysis TC data of Elsner and Kossin show no trends in the annual global Cat45 counts or CAT45 ratios from 1981-2006.
The cat45 counts for the period 1984-2008 fit reasonably well to Poisson distributions for the six TC basins indicating that they result from a chance combination of conditions.
As I recall, but have not verified personally, someone noted that the Cat45 trends are dominated by the Cat4 hurricanes without a trend in the Cat5. The average intensity of TCs shows no trend from 1981-2006.
Finally the regression of annual global Cat45 counts and ratios versus SST for the global zone from 24S to 24 N shows a trend slope not significantly different than zero for the counts from 1981-2007 and from 1984-2007 for the ratios.
Re: Judith Curry (#48),
Replace ACE with PDI (Emanuel 2005)?
Re: Judith Curry (#48),
Relating the intensity of discrete, highly localized events to global warming seems dicey at best. After all, what is dynamically essential for the development of very intense tropical cyclones is high water temperature under the track and strong shear in the atmosphere above. Neither of these factors is at all well-modeled by climate GCMs, which are notorious for regionally unreliable results. In fact, no hurricanes develop in the model runs. The claim that the proportion of Cat 4-5 hurricanes is a more telling metric than ACE strikes me as wishful handwaving, rather than proven science.
snip
Steve- your points were OK, but people end up piling on in these sorts of exchanges. Let Judy respond to the points on the table if she wants to do and I request that people hold their horses in belaboring Judy. I think that readers will get more response by piling on less. Anyway, that’s what we’re going to try.
Re: Andrew (#50),
Amen to avoiding piling on. I would much rather read a dialog between two knowledgeable people than to read one comment from someone knowledgeable and then a dozen counter-attacks from a dozen amateurs. Visits by representatives of the conventional wisdom should be welcomed and treated with respect.
ryanm: appeal to authority much?
Re: frost (#53),
ryanm said:
No, and I’m not doing that here. I just want to see a dialog, which is unlikely when a dozen enthusiasts jump on a post by a representative of the scientific establishment, many of them making the same points and many being strident in tone.
Re: frost (#63),
Whenever I visit blogs like this, I find myself wading through the comments “cherry-picking” the meaningful ones and ignoring the rest. I am quite sure Curry, et.al., recognizing ” the nature of the beast,” are capable of doing the same. Curry visits here fairly often. I doubt she does so for the abuse.
Climate change as it relates to MPI theory is not about more intense TCs, rather it is about a TC becoming more intense. There is a subtle, but important, distinction between the two that most of this thread appears to have missed.
Re: Jim Elsner (#51), Now that does sound subtle! By putting the emphasis on more, do you mean more as in a greater number or more as in greater intensity? I can hardly follow these discussions on hurricanes sometimes, everyone is saying different, and at times ambiguous things.
Re: Jim Elsner (#51),
I had excerpted the statement presented below from a link to the Elsner paper here on another thread at Post #198
http://www.climateaudit.org/?p=5449#comment-338368
The link to the Elsner paper and SI are here:
http:/myweb.fsu.edu/jelsner/PDF/Research/ElsnerKossinJagger2008.pdf
http://myweb.fsu.edu/jelsner/extspace/globalTCmax4.txt
The Elsner data, by my analysis, shows no trend of the median maximum wind speed for the 1981-2006 period studied nor would it show a trend for Cat45 counts or % Cat 45. What Elsner et al. (2008) show is a an increasing trend for the TCs at the higher maximum wind speed levels but a trend that has a slope significantly different than zero for the 90th quantile (upper 10 %) only occurs for the NATL basin– of all the global basins studied. The data were from a reanalysis procedure described in the paper.
Jim Elsner, there are two results from two different TC basins in your paper linked above in a graph that appear to look identical as noted by David Smith on an earlier analysis of your paper. Look at Figure 2 at the top for WP and EP where, while the x coordinates are different, the curves look exactly the same.
Also when we analyzed your paper we found that there was a very noticeable difference between the Best Tracks and your reanalysis when looking at the lower end of the maximum wind TC distribution. Would you care to comment on what you might propose as causing this difference? As I recall in the reanalysis compared to Best Track, the distribution started quite low and built up to a maximum slower.
I do not believe that most commenting here doubt that the maximum intensity of storms (wind sppeed) can theoretically increase with SST, the point is by how much and can it be detected to date in the noisy TC data and are the methods used in papers attempting to detect it proper and well understood.
Also I have a question about the theory of maximum TC intensity as it relates to the lower wind speed TCs that dissipate before reaching the higher quantiles of maximum wind speeds. At higher SSTs would not the extra energy “momentum” push these TCs to higher maximum wind speeds prior to dissipation assuming the dissipation factors remain the same?
I guess that’s what I get for being a bit pushy with my questions. :Blush:
Wouldn’t an increase in % Category 4/5 hurricanes imply an increase in average wind speed if the total counts and hurricane days remained the same?
Mark
Re: Mark T (#56), and an increase in ACE.
It’s… it’s… like you read my mind. You don’t need a separate theory for why ACE should increase with GW, you need a theory as to why it hasn’t increased in spite of GW. 😉
Mark
Re: Mark T (#58), what inevitably gets lost in all of the MPI theory and discussions about counts and ACE is the truly remarkably natural variability in the current climate system that absolutely dominates anything AGW related. Elsner et al. (2008) explicitly says that they didn’t consider it.
It is clear from the recent tropical cyclone literature and discussions here and elsewhere that knowledge and application of Maximum Potential Intensity (MPI) theory has been a little shaky — to be generous. I wish I understood 10% of what Kerry Emanuel knows or intuits about the machinery or nuts-and-bolts of tropical cyclone dynamics. It is far from a hand-wavy theory that gets thrown around indiscriminately today.
Just saw this at Prometheus: Inexpert Elicitation by RMS on Hurricanes
Re: John Baltutis (#62),
The world is a SCAAARY place. Insurance, anyone?
Re: bender (#68),
It certainly is such a place. More chills and spills if one navigates the atmosphere in a flying machine, but not as terrifying as your city and country roads in a gasoline, carbon-dioxide generating machine.
I think that this thread is as good as any for putting my calculations of the Cat5 hurricanes for the 1981-2007 time period trend using the IBTracs data series (greater than 119 knots). Below I show the trend slope, standard error for the trend slope, the adjusted R^2 and the p value for the slope being different than zero. I used the entire 6 basin counts and then those six basin counts without the NATL – as a sensitivity test.
The results show a significant positive trend for the six basin count but not one when the NATL Cat5 hurricanes are removed. The NATL shows long period cycles for TC activity that could yield trends that are sensitive to starting dates even over relativly long periods from 1981-2007. That does not say that the other six basins do not have cyclical TC activity characteristics and even in opposition to the NATL. Opposition could effect the sensitivity results for the case excluding the NATL counts. I am just saying that that situation would need an explanation.
I think all the papers that I have linked here and on other threads at CA on the same subject are all in agreement with one another and with my layperson’s attempts at analysis. What I think is sometimes lost in the reviews of these papers is the details of all these papers and the sensitivity testing issues that put the evidence in better perspective – or at least for me.
An issue in the Elsner paper linked in my previous post has a subtle statistic that pops out of their graphs that is not directly discussed in any detail in the paper. The graphs show a black zero line for a zero trend in the trend in maximum wind speed versus maximum wind speed quantiles and a red line that shows the average trend for all maximum wind speeds over the period of interest (1981-2006). That red line allows one to view whether, given an overall trend for maximum wind speeds, the quantiles shown in the graph are significantly different than the overall trend, i.e. can we show given the average trend that the higher quantiles trends are significantly different. Notice that the significance hurdle for that comparison is more difficult to overcome than that of comparing the trends to a zero trend line.
For the individual basins the red line shows that the average trend is approximately 0.1 meters per sec per year above the zero trend line for the EP, WP and SI basins (EP and WP appear in the graph to be identical and presenting an unlikely case), while NATL is 0.3 units above a zero trend and SP is nearly 0.2 units beneath the zero line and finally NI is close to the zero line. This comparison also shows the NATL as the only basin with a 90th quantile trend apparently significantly different than the average trend of the basin.
All six TC basins for Cat 5 hurricane counts:
Trend slope = 0.094; Std Error = 0.042; Adj R^2 = 0.13; p = 0.035
Five basins without NATL for Cat5 hurricane counts:
Trend slope = 0.045; Std Error = 0.033; Adj R^2 = 0.03; p = 0.182
Will our grandchildren run out of hurricanes?
These time series are based on global best-track data over the last two decades.
Over the same period the global temperature rose, so these trends are indeed consistent with global warming.
Since global warming drives everything on the planet and is projected to continue, and possibly accelerate, our progeny are “very likely” to experience the following extrapolation:
The last hurricane winds down around 2017.
Our offspring may never watch a Weather Channel meteorologist lose a hairpiece or blown down a beach live. Other employment may suffer, requiring stimulus:
My point in this simple-minded goofiness is to affirm what we all know: analysis of non-linear processes, like tropical cyclones, where understanding is poor and historical data quality and completeness vary over time, require caution, modesty and more than a few grains of salt.
Re: David Smith (#66),
David, I see we will be projected to have no hurricanes while we continue to experience Cat45 hurricanes.
Seriously though, I have seen conjectures from climate science about hurricanes transporting heat polewards as a negative feedback in a world of higher SSTs. Without those hurricanes and more intense ones, we will experience higher GW, so however you choose to play it, the outcome will be bad.
Re: David Smith (#66), nice analysis. I favor a 12-24 month running count when doing global counts of hurricanes. Webster et al. (2005) used pentads or 5-year chunks without explanation, but fortunately the climate signal popped right out.
Re Dave Smith’s #66, I’ve been trying to reconstruct the original /numbers/ that underly Dave’s plot, so far without real success. One problem is that the x axis, though labelled, does not identify which of the ticks corresponds to which year. A heavy tick at 5 year intervals would be useful. Alternatively a direct reference to the source would be a great help. Any chance?
The extrapolations were of course jokes, but it would nevertheless be interesting to indicate the approximate confidence intervals, say 90%, for the the predictions. Confidence interval tend to shed some realistic illumination onto extrapolations in general.
Re# 71 Robinedwards, send an e-mail to me ( mndsmith33 AT earthlink.net ) and I’ll send the spreadsheet to you.
I marvel at the stupidity of having the National Hurricane Center based in Miami, the Hurricane capital of the world! Using that logic we would create a National Tornado center in Kansas, A national Earthquake center in San Francisco, National Flood Center in New Orleans and lets give the National Volcano Center a front row seat at the base of Mauna Loa volcano.
You know its not about science or public safety, Its about egos and photo ops for an otherwise dull career.
ryanm: What is the inspiration for this weird comment? BTW, the National Severe Storms Lab is in Norman Oklahoma.
I got a chuckle from the “For Sale, by Acme Realty CO.” sign in front of the National Hurricane Center building in comment #66 above. Then it stuck me what a bad idea the location was.
Now that you mention it, National Severe Storms Lab is in Norman Oklahoma is another bad idea. Its a silly notion that they have to “Be There” to do their job.
I think that they just love the drama. But at what cost? Putting these facilities in Harms way means each needs to be built like Bunker, doubling or tripling the build and operating costs. Why go to the expense to make a facility and the supporting community and families disaster proof when all you have to do is plan ahead and build in a less disaster prone area. Its all about the Drama.
JohnT — “Now that you mention it, National Severe Storms Lab is in Norman Oklahoma is another bad idea. Its a silly notion that they have to ‘Be There’ to do their job.”
Unless the ol’ memory is failing, it was called the National Severe Weather Center several (4+?) decades ago and was located in Kansas City.
My coordination visit was so long ago that I don’t even remember the details any more, but the point is that this is a longstanding condition.
I just downloaded the 2008 Atlantic Best tracks database from NHC. I compared it with the 2007 database and found 9 newly added storms in the period from 1921-1925. Part of their ongoing reanalysis I guess. About 185 tracks added with only an increase in ACE of about 15. A couple years the ACE actually decreased with added storms. They must have made numerous adjustments in wind speeds.
Here is a comparison of the storm counts.