Nature has published another remarkable example of academic check kiting by Michael Mann et al, this time “Atlantic hurricanes and climate over the past 1,500 years”. (Prior examples of academic check kiting discussed at CA are Ammann and Wahl, the story of which is well told by Bishop Hill’s Caspar and the Jesus Paper and “Mann, Bradley and Hughes 2004”, cited in Jones and Mann 2004.)
Mann et al 2009 reconstructs Atlantic tropical cyclone counts resulting in a curve that looks pretty much like every other Mannian curve. Atlantic tropical cyclone counts as a linear combination of reconstructed Atlantic SST in the east tropical Atlantic “main development region” (MDR), reconstructed El Nino and reconstructed North Atlantic Oscillation, using a formula developed in (3,16) – which surprisingly enough turn out to be articles by Mann himself (Mann and Sabatelli, 2007; Sabatelli and Mann 2007) previously discussed at CA here. This is summarized in the article as follows:
An independent estimate of past tropical cyclone activity was obtained using a statistical model for Atlantic tropical cyclone counts. This previously developed and validated 3,16 statistical model conditions annual Atlantic tropical cyclone counts on three key large-scale climate state variables tied to historical variations in Atlantic tropical cyclone counts: (1) the SST over the main development region (MDR) for tropical Atlantic tropical cyclones, which reflects the favourability of the local thermodynamic environment; (2) the El Nino/Southern Oscillation (ENSO), which influences theamount of (unfavourable) vertical wind shear; and (3) the North Atlantic Oscillation (NAO), which affects the tracking of storms, determining how favourable an environment they encounter. The statistical model was driven by proxy-based reconstructions17,18 of these three state variables (Fig. 2), yielding a predicted history of Atlantic tropical cyclone counts for past centuries.
One doesn’t necessarily expect much clarification from Mannian methodology and this time Mann surpasses himself. Remarkably the Methods Summary is almost word-for-word the same as the article. It’s actually a little less. Mann explains once again that they reconstructed hurricane counts by using reconstructions of the SST in the Atlantic Main Development Region, El Nino and the North Atlantic Oscillation, with the only new information being that they only used the North Atlantic Oscillation reconstruction over the past 500 years from Luterbacher (18), but it didn’t matter. Mann again cites (3,16) Mann and Sabatelli; Sabatelli and Mann.
We used a statistical model of tropical cyclone counts as conditioned on3,16: the MDR SST, the ENSO (measured by the boreal winter Nino3 SST index), and the boreal winter NAO index The statistical model, which is trained on the modern historical record, has been shown in independent statistical validation experiments3,16 to resolve roughly 50% of the interannual and longer-term variations in Atlantic tropical cyclone counts. The model, in this study, was driven by decadally smoothed proxy reconstructions of the three required climate indices to yield predictions of tropical cyclone activity over past centuries. The MDR SST and Nino3 reconstructions were derived from proxy-based surface temperature patterns spanning the past 1,500 years17. Though an NAO reconstruction was available only for the past 500 years18, the NAO influence was found to be very minor (Supplementary Information).
The Full Methods in the online version adds little additional information. Again we are told that they used a statistical reconstruction using Atlantic SST and El Nino from the enigmatic ref 17. Confidence intervals appear to be done using the recipes of Mann et al 2008 with lots of “decadally smoothed” series.
Statistical prediction of tropical cyclone counts using proxy reconstructions. Here the model was applied to decadally resolved reconstructions of MDR SST and Nino3 described by ref. 17 and the decadally smoothed winter NAO index of ref. 18. For the instrumental interval (1851 to present), standard errors due to uncertainties in the model coefficients were calculated from the residual decadal variance diagnosed from the validation residuals (standard errors were averaged for the early and late intervals of the split calibration/validation procedure). For the pre-1851 statistical model estimates, which are driven by reconstructed climate indices, there is an additional component of uncertainty due to the uncertainties in the climate indices themselves. This contribution was estimated by Monte Carlo simulations in which the statistical model was driven with an ensemble of 2000 randomly perturbed versions of the statistical predictors consistent with their estimated uncertainties17, and an additional random term due to the uncertainties in the model coefficients.
The Supplementary Information sheds no light on the methodology or the proxies.
The Supplementary Information contained no data sets. The proxies used for the Mann et al submission are not even listed.
The edifice is built on the SST and Nino3 reconstructions, both of which are references to the enigmatic reference 17, which turns out to be an unpublished submission of Mann et al.
17. Mann, M. E. et al. Global signatures of the Little Ice Age and the medieval climate anomaly and plausible dynamical origins. Science (submitted).
At the time that Nature published this article, there was precisely NO information available on what proxies were used in the reconstruction of Atlantic SST or El Nino or how these reconstructions were done. Did any of the Nature reviewers ask to see the other Mann submission? I doubt it. I wonder if it uses Graybill bristlecone pines.
UPDATE: Roger Pielke Jr observes below that Mann has provided a “grey” Supplementary Information at his website here http://holocene.meteo.psu.edu/~mann/Nature09/ , which commendably includes source code and data from the check-kited paper on i.e. we still don’t know anything about the reconstruction proxies for Atlantic SST or El Nino. To my knowledge, there is no reference in the original article or Nature SI to the Supplementary SI at Mann’s website; there is no link on Mann’s website to the Supplementary SI and the directory hosting the Supplementary SI is not readable or searchable. Unless you know the precise name of the subdirectory, you can’t there. Right now, I don’t know how Roger found the Supplementary SI, but once located, the documentation looks at first glance to be very commendable.
References:
Mann, M.E. , Jonathan Woodruff, Jeffrey P. Donnelly & Zhihua Zhang. NAture 2007. Vol 460| 13 August 2009| doi:10.1038/nature08219 http://www.meteo.psu.edu/~mann/shared/articles/MannetalNature09.pdf
3. Mann, M. E., Sabbatelli, T. A. & Neu, U. Evidence for a modest undercount bias in early historical Atlantic tropical cyclone counts. Geophys. Res. Lett. 34, doi:10.1029/2007GL031781 (2007)
16. Sabbatelli, T. A. & Mann, M. E. The influence of climate state variables on Atlantic Tropical Cyclone Occurrence Rates. J. Geophys. Res. 112, doi:10.1029/2007JD008385 (2007).
17. Mann, M. E. et al. Global signatures of the Little Ice Age and the medieval climate anomaly and plausible dynamical origins. Science (submitted).
18. Luterbacher, J. et al. Extending North Atlantic Oscillation reconstructions back to 1500. Atmos. Sci. Lett. 2, 114–124 (2002)
Climate Audit 
181 Comments
Who are the reviewers?
I think reviewers are typically anonymous.
snip – policy
The issues as always with Mann are
1)Are the proxies reliable and what is the evidence this is the case
2)Are the statistical methods correctly applied.
3)Is the model of predicting hurricanes from the data sources valid.
The third of these can be most easily checked over the next year by simply using real SST data with ENSO etc data to predict the number of storms. These effects should be evident over shorter time scales than other climate models predictions he has made and as such are open to testing ovber shorter time scales.
The power of any sciencetific model is in the accuracy of its predictions.
It is the stated policy of the Americam Meteorological Society that one cannot cite unpublished
manuscripts. Evidently this rule is ignored by lesser quality journals and reviewers?
Note that I am not stating that the AMS is perfect. 🙂
Jerry
The claim is made that storm tracks can be reconstructed 14 centuries after the event? And also which environments were favourable encounters?
I doubt this can be shown.
There’s a writeup of this article here. I’m not sure that it fully fits preconceptions. He seems to have discovered a Medieval Windy Period.
Re: Nick Stokes (#8),
No, it exactly fits preconceptions. Mann has “discovered” a Medieval Windy Period, but that only supports the point he is trying to prove… that warmer weather creates more hurricanes. Note the article sub “Historical estimates suggest that global warming could boost the number of hurricanes.” If you are Michael Mann and observations do not support your alarmism, what do you do scientifically? You create proxies and then claim the proxies prove the observations are wrong. This is not science. This is embarrassing.
Re: Nick Stokes (#7),
Nick, of course, this fits the preconception. In reading the paper, it is quite clear that Mann et. al. are only referring to the North Atlantic. And the team has openly accepted that there was a medieval warm period (or should I use the politically correct term of “medieval climate anomaly”?) in the northern hemisphere. This point is however a sideshow (straw man?) to the aim of the posting.
Nick, while we all appreciate your intelligent defenses of some rather poor mainstream AGW articles, you would gain a fair amount of credibility if you demonstrated less partisan blindness. The point of Steve’s posting was not that Mann et. al. is a faulty analysis (or sound, for that matter) but that it relies on unpublished articles, provides no data, no code and includes limited description of methodology. Even Dr. Schmidt appears to have undergone a recent miraculous conversion on the road to Damascus on the need for this supplemental information. Will you accept that Steve’s comments are valid and join the group in calling for full data and code to be released as standard practice?
Re: compy (#49),
So here’s something I don’t understand. How come the Little Ice Age is named as such but the MWP is known as the medieval climate anomaly?
Also, separately I couldn’t help but note that in the nature news post regarding the Mann paper) there is an apporving comment from one of his coauthors Urs Neu in citation 3
Re: compy (#49),
The data relied on was transmitted to the Crown by pirates (aar mate) including records from Margarito of Brindisi’s and is covered by confidentiality agreements, rare historical documents that have either been lost, stolen, or cannot be shared due to their physical fragility and importance to international relations1.
1. Mann, M. E., Jones, P. Pirate tales and climate records. Nature. (thinking about submitting )
Re: steven mosher (#54),
Mann/CRU must be losing their touch – they forgot to include that the data could be made available only to academics specialising in British maritime history.
(soley for the protection of non-academics, of course, who might use the data improperly and hurt themselves)
Re: compy (#49),
Compy, my remark on preconceptions was a bit tongue in cheek. But as Roger Pielke, Jr. (#51), has pointed out, Mann has released the data and code. I haven’t yet read the paper, but hope to do so today (I’m in a different time zone).
Re: Nick Stokes (#73),
Nick, not quite accurate. As Steve points out:
So, Nick, you have no excuse to evade the question: Will you accept that Steve’s comments are valid and join the group in calling for full data and code to be released as standard practice?
Re: compy (#128), (#49), On accuracy, I was responding to your statement that Mann had “provided no data, no code”, which is false. And on “partisan blindness”, I noted others making similar assertions, which suggests at least a possibly partisan-influenced failure to look.
But yes, I agree with the principle that where possible relevant code and data should be made available.
On your quote from Steve, he also said, rather puzzlingly, that “The Supplementary Information contained no data sets”, which suggests that he wasn’t seeing the part that Roger pointed to. These include, for example, “MDR AMO SST” and Nino3, which seem to cover what you are referring to. So despite your claim, I don’t see anything missing.
I do agree that it is undesirable when two linked papers do not appear simultaneously. Sometimes, however, there are reasons. There may actually be two distinct scientific projects, with different co-authors who do not want to merge their work. That seems to be the case here; the full citation of Ref 17 is:
Mann M.E., Z. Zhang, S. Rutherford, R.S. Bradley, M.K. Hughes, D. Shindell, C.M. Ammann, and F. Ni, in review: Global Signatures of the Little Ice Age and Medieval Climate Anomaly and Plausible Dynamical Origins. Science. Manuscript Number: 1177303
Folks here might look forward to reviewing that.
Re: Nick Stokes (#129),
Nick, you misread what Steve said. Note that “Supplementary Information” is capitalized, so Steve is clearly referring to the Supplementary Information referenced in the paper which can be found here.
Steve is right, you are wrong.
Re: compy (#132), Well, this is getting a bit technical. The page Roger linked to is headed Supplemental Information (yes, a subtle difference). But what matters is that all the information that you refer to was made available at the time of publication, so I don’t see any point in harassing Mann about it.
It’s outrageous that Mann should use the words “Science (submitted)”. The editors at Science should immediately reject the submitted paper because of the pressure Mann’s action puts on reviewers to lean towards acceptance.
But close to complete with a few minor details to tie up. And robust of course.
“The statistical model, which is trained on the modern historical record, has been shown in independent statistical validation experiments3,16 to resolve roughly 50% of the interannual and longer-term variations in Atlantic tropical cyclone counts.”
Am I missing something? Surely a science based paper would not use “roughly”. Also what happened to the 50% it didn’t resolve, might as well just toss a coin.
Historical British naval logbooks in which you can find detailed and well kept records of past cyclonic activity from 17th century onwards (the British were very pedantic) not good enough for Mann et co?
The model knows better?
Although it perhaps better not let them near those logs, don’t think claims of “missing data” would be warmly welcomed.
I’m curious about the reaction of the Pielke’s….
Interesting…. nothing at RC about BS09 error (he did send one post to Lucia’s), then end of story. They will not post on the error just you watch.
As we should expect the BBC have grabbed this story and given it as much coverage as they can. Everything Mr Mann publishes must be true, he’s proved that in the past.
http://news.bbc.co.uk/1/hi/sci/tech/8197191.stm
Life is easy when the gatekeepers hold the gate open for you…
I scanned the paper briefly. The information on what was done was very sparse. It appears to be a Poisson regression with a variety of Mannian methods scattered throughout.
I found this quote indicative of the quality of statistical insight in the paper:
Climate science statistics at work …
Will look at the document further later today.
Re: RomanM (#18),
Often I find CA posts amusing, but this is just depressing. I first became aware of Steve’s work well before CA existed, because PCA was my bread and butter at the time, and the debate about MBH 98/99 was something that I could readily understand. Not to mention that the “hockeystick” scared the hell out of me when I first saw it on the front page of my national newspaper via the IPCC.
I guess I look forward to Roman’s comments on the new paper, but I’m tired of Mann’s novel statistical methods… self snip.
Re: RomanM (#17),
Briggs, where are you? You need to have a look at this as well.
Steve-
Thank you for all your postings. They are the first that I look for each day, and I feel let down when I don’t find something new. I don’t claim to understand all the statistical analysis (but I am learnng a little from your work), but I can follow your train of thought, and it is usually amusing as you slice, chop, and dissect bogus material from warmist sources in a very quiet voice. I am only one of many who appreciate your work, but an additional pat on the back never goes astray, I am sure!
IanM
“…using a formula developed in (3,16)…”
Instead of holding up signs saying “John 3:16” at sporting events, we can have “Mann 3,16!”
Re: Mihcael Jankowski (#19),
That has T-shirt potential. Maybe on the same T-shirt as the flying spaghetti monster and the business about pirates and temperature.
As Professor Wegman pointed out, it is unfortunate that Mann and his cohorts have steadfastly refused the help of people who actually understand statistics. Of course, in addition to refusing their help, he has worked to make it impossible for them to even review his work.
snip – take a valium
snip – diversionary
The author list is almost identical: Mann, M. E. et al. reads Mann M.E., Z. Zhang, S. Rutherford, R.S. Bradley, M.K. Hughes, D. Shindell, C.M. Ammann, and F. Ni. I suppose there is yet-another-independent-study-confirming-hockey-stick soon to be found in Science 😉
If you apply Mann’s smoothing methods to the temperature records for the moon and for the sun it would have s a statistical significance of P = 0.0 for the statistical model itself (that is, a 0% chance that we would be incorrect in rejecting the hypothesis that the model coefficients for the three predictors are all zero).
Mann: “The statistical model, which is trained on the modern historical record.” He is apparently using the uncorrected (raw)hurricane counts to calibrate his model. NOAA just showed that the raw counts are incorrect because they missed many storms. It appears that Mann’s study is nonsense.
Steve: NOAA’s recent position is not the last word on the topic. The reconstructions are linear combinations of the proxies. If the NOAA counts change somewhat, it would have some effect on the reconstruction, but not necessarily as much as you think. I do not think changes in these counts would imply automatically that the study is “nonsense”, though it may well be nonsense on other grounds.
“He is apparently using the uncorrected (raw)hurricane counts to calibrate his model. NOAA just showed that the raw counts are incorrect because they missed many storms.”
snip it – diversionary
snip-
Folks, there’s a difference between check kiting, data being unavailable and a study not holding up.
Just because Nature hasn’t properly reviewed the paper doesn’t mean that the paper is garbage. It may or may not be. The check kiting and unavailability of data is not encouraging, but please don’t go a bridge too far.
NPR had it this morning.
http://www.npr.org/templates/story/story.php?storyId=111809658
Re: MB (#29),
If this methodology is an adequate description of what they actually did then the data should be closely scrutinized.
As Geoff and others indicated, there are very few locations used in the study, moreover these locations are subject to all kinds of sea level influences that would have to be “corrected” before they could derive any information concerning storm surges (UHI anyone?).
For example, according to the Geological Survey of Canada:
http://gsc.nrcan.gc.ca/coast/sealevel/hist_e.php
The sea level at Halifax has been rising fairly consistently over the last six thousand years at a rate of something like 33 cm per 100 years. (Based on shell, peat, salt marsh, mud, seaweed, tree and foraminifera deposites). Over the period of 1,000 years that’s more than 3 metres. What does that do to your assume storm surge?
And these sealevel changes have not been consistent in the Atlantic basin, for example Connecticut is more like 1 metre over the last 1,000 years.
And these sealevel changes have not been consistent over time, for example, in Chesapeake Bay, “sea level rise was relatively low during the 17th and 18th centuries”.
http://www.agu.org/revgeophys/dougla01/node4.html
Re: Jeff Norman (#106),
Jeff, storm surge is noted as a potential bias in one of the underlying Woodruff, Donnelly papers( I think G3 2008).
It would be interesting to see if the paper in submission to Science has any references to other papers in submission. Any way to check this?
Does anyone have a link to this article, I don’t have access to the journal myself. It’s Mann so I’m interested in the math particularly in the fact that Medival warming actually has occurred now??! In the past his math suppressed historic signals in relation to the calibration period so it’s confusing that this wouldn’t be the same. Perhaps it’s just less suppressed.
Mann keeping it interesting.
Its amazing that nothing could be listed except a conclusion. It’s like the birds shrinking by 2% due to global warming.
So perhaps the reviewers also reviewed the science submission and determined that method is acceptable for the data which apparently wasn’t worth listing?
Yow!
Climate Scientology seems to be polluting everything it touches. I used to have a high regard for Nature.
The opening line in the NPR story is a perfect example of – well — NPR science reporting, which accepts all warming stories uncritically, even whenthey fly directly opposite one’s accumulated, direct, personal observations:
“The Atlantic Ocean is experiencing the most intense period of hurricane activity in 1,000 years, according to a study in the journal Nature.”
I’m always reminded of Groucho Marx’s denial, when caught in the arms of another woman by Margret Dumond: “Who ya gonna believe, me or your lyin’ eyes?”
So if reference 17 is still in submitted stage, it hasn’t been peer-reviewed and therefore is invalid as a reference according to the loudly proclaimed climate science “standards.” Why not just summarize the results and note them as currently unpublished data instead of doing this little dance? If Nature is going to allow this sort of thing, then it ought to make the submitted paper available for scrutiny or require the supporting data be available in an SI. But the ruts are deep in that road, aren’t they?
Steve, you are in the news:
“Global Warming ate my data”
http://www.theregister.co.uk/2009/08/13/cru_missing/
Also cited: Sabbatelli, T. A., Mann, M. E. & Miller, S. K. Semi-empirical projections of future Atlantic tropical cyclone activity. J. Geophys. Res. (submitted)
Science now reported the following.
Re: Jeff Id (#40),
Just wow. I don’t think an accurate picture of hurricane activity can be formed by using landfall data even if I assume using this sediment technique is valid. Most storms don’t hit land. Using only 8 sites is really stretching things. I would think the hurricane strength to move measurable sediment is only localized to a very small area in most storms. I know using only US areas is pushing all credibility. Of the storms that make landfall, many (I’d say most, but I don’t have the data in front of me) hit Mexico, Cuba, and Hispaniola.
Re: Scott B (#59),
See mann’s comments to objections. at the link roger Jr. posted.
Re: Jeff Id (#40), Down Under they are called tropical cyclones. Here is a map of tracks for years 1950 to 2006, about 50 years. Now multiply the tracks by 30 to get 1,500 years of data. Ask yourself if you can sample 8 lagoons along the coast and reconstuct anything at all about these cyclones. Image from
http://www.bom.gov.au/cgi-bin/silo/cyclones.cgi?region=aus&syear=1950&eyear=2006&cyclone=all&loc=0
Re: Geoff Sherrington (#97),
Yes, trying to correlate sediment cores from lagoons across 8 locations with that density of cyclone debris is a job for Supa-Geologist. I’ve been correlating core for over 30 years and against that experience such a task may be considered somewhat challenging
Re: Geoff Sherrington (#97), Geoff.
I see no mathematical obstruction to using 8 lagoons to form an “estimate”. On the surface it seems rather straightforward. So, the questions would turn to..
1. Is the collect mechanism biased toward over/undercount
2. What kind of error bounds do you get
3. Whats the resolution of the proxy ( measuring accuracy)
4. How good is the count in the “instrumented period”
The first chart I made was the one that compares the standard case ( mann) versus Landsea. The difference in count ( L -M) was always positive ranging from about 1-3 extra storms per decade. shrugs.
Re: steven mosher (#115),
Statistically the construction of an estimate is an interesting sort of inversion problem. IT would have been nice to see a reference to actual statistical literature, but it will be an interesting project trying to figure out how it might have done properly.
Aside from these sorts of problems, the effect of rising sea levels on storm surges needs to be considered in the estimate and it doesn’t seem like there is an allowance for that.
Also we haven’t seen ANYTHING on the proxies used in the other half of the paper – the ones used to estimate Atlantic SST and Nino. Those are in the check-kited paper.
Re: Steve McIntyre (#116),
Agreed. My main point against geoff is that I see no absolute bar preventing one from taking up the question.
The check kiting, of course, cuts off any inquiry into one of the key aspects of this however. On the other hand
one can well imagine that people with Geoff experience would be highly skeptical of the results. That is, he would
not recommend expending company dollars to dig for gold on such an analysis, much less change policies WRT
GHGs.
Re: Steve McIntyre (#116), Ditto on the rising sea level issue. Don’t have a clueByfour on how one adjusts for that.
Re: steven mosher (#115),
Steven, did you not read my post #100 on correlation of cores ?
I sometimes despair at the level of geological illiteracy in the general population
Correlation across sample sites is important (critical) as without it you are haplessly counting the same cyclone over and over. That is the point of the track map that Geoff Sherrington posted
Re: steven mosher (#115),
I cannot fault your qualitative logic. OTOH, as one who has worked for years in this territory and sat through cyclones, I can see and read evidence of the changes in the courses of rivers, the natural relocation of lagoons, relative changes in sea level, the effects of fire/vegetation modification on runoff and sediment composition, and about n other factors. Apart from which the natural tidal variation at Broome WA, to take but one example, can be many metres, so that the time of day and part of the seasonal cycle of the cyclone landfall can make a difference to lagoon proxies.
So, to conclude this answer, quantitatively impossibe to reconstruct with any useful confidence.
In this part of the relaxed world, they use a word like the Mexican “manyana”, except that it is not so positive. It’s called “dreamin'”.
Re: steven mosher (#115),
On further reflection, might I please assume that if one of your points 1-4 fails, the whole concept fails. If so, I will concentrate on one specific effect re point 4, for tropical Australia. It addresses the “bar” question in your #124, how good the count is in the instrumented period. The example I shall give does not mean that I cannot provide others for consideration.
The water buffalo Bubalus bubalis is thought to have been imported from Indonesia to Australia about 1820. There were about 400,000 of them by 1980 when Australia had a rash encounter with the UN’s World Heritage List and shot them out in the wild, using machine guns from choppers at times. Some remain, but behind fences. I made policy submissions about them from the early 70s to the early 90s and was frequently sharing habitat.
The buffalo frequented waterholes, including precisely the class that an unsuspecting later researcher might investigate for sedimentology. The buff would stick a hoof as deep as a foot or two into the mud, covering large areas as water receded in the dry season. There was thus a period of thorough mixing of the top part of the lagoon sediment profile that coincides with most on the instrumented period.
A decision to accept or reject a candidate lagoon for cyclone surge research would require a measure of whether the lagoon sediment was buff homogenised (if I might use the term), or is an “interesting sort of inversion problem” to borrow a Steve term from # 116. No absolute measure comes to mind unless there are certain offshore islands known to have been free of buffs – and there probably are. Downside and quantitatively, the mainland coastal lagoons must be suspect in the instrumented (calibration) period. This restricts the scope of study, or its value if one does not include buff homogenisation.
As an addendum, if SST is so important in cyclone activity, why does the map of # 97 indicate so much activity over land? I have read some papers about mechanisms that keep them spinning after landfall, but the combination of Coriolis, Hadley cells, El Nino, SOI and the water rotating unexpectely clockwise in SH is a bit hard to spin together – especially as the western land they traverse is mostly very dry desert. The spin that I mention includes a mechanism that storm rain falling ahead of the storm over land is taken up again to sustain it, like perpetuum mobile.
Experiences – some of the world’s largest ships visit the NW of Australia. We had to have very large tugs to handle them and emergency plans involving cyclone forecasting to protect them, not always with success.
Here is a serious substantive problem with the paper:
http://rogerpielkejr.blogspot.com/2009/08/changing-perilous-assumptions-to-suit.html
Re: Roger Pielke, Jr. (#41),
As Roger Pielke Jr. points out, Mann (2009) is inconsistent with Mann (2007)’s proposition about stationarity of data. Yet no-one pulls up Mann and asks him to square the circle.
Re: Steve McIntyre (#28),
Well Steve its not exactly unexpected, because as you know Mann has “form”, as they say in horse racing.
On the other hand, just because Nature won’t do even a half-arsed job of actual peer-review and Mann himself has admitted he is no statistician and has followed previous form with respect to data, code and check-kiting, does not mean that by complete chance he hasn’t screwed up.
But I’m not betting on it.
Godel warned us to watch out for this stuff.
Has anyone alerted Mann to the fact that the author of the Harry Potter series surpasses him in sheer fantasy? But he’s gaining fast!
People may wish to express their thoughts about this work at Nature’s Climate Feedback – see the entry after the one about the CRU data spat.
How??? How is it even remotely possible for this to “pass” a peer review and go on to publication without any data on the reconstructions what-so-ever??
How?!?
I weep for the future of science….
Now, that’s climate change I can believe in. I think there’s profuse evidence that we’re well into another windy period right now.
Mann et al:
So, they argue that medieval level of hurricane activity exceeded recent level, so this is in some way opposite to Hockey Stick. When seen headlines, I expected new Hurricane HS, but this is quite the opposite. I can understand Steve’s skepticism toward anything that Mann does (taking into account previous experiences) but what we find in this paper is very different from what most of us would expect – not claims of “unprecedented” or “worse than we thought” conditions today compared with the past, but on the contrary, claim that MWP levels of tropical storminess was higher than today. It seems like far-reaching revision of HS picture of global climate if anything (high level of MWP activity was in large part according to Mann consequence of “relative warmth of the Atlantic”, which is to say that according to him tropical Atlantic in MWP was as warms as or warmer than today).
Certainly Mann still could be wrong, taking into account recent analysis by Lindesa et al that shows that increase in number of short term tropical storms in North Atlantic in recent decades is likely a consequence of better monitoring. So difference between MWP and recent times could be even larger that Mann supposes, but all in all, “big picture” he presents in the new paper doesn’t seem to look controversial whatsoever, and certainly is at odds with alarmist notion of how warming tropical atmosphere would behave in the global warming.
Data, Code:
http://holocene.meteo.psu.edu/~mann/Nature09/
Re: Roger Pielke, Jr. (#51),
thanks
I must have slept through a scene. Last time I read anything about Mann, he was denying the Medieval Warm Period. Now he’s claiming a Medieval Windy Period? How does he get windy without warm? Are the talking points mangled, or did I miss something?
Re Michael Jankowski (aka Mihcael Jankowski),#19: Yeah, “Mann 3,16” signs behind the goal posts: Gotta love it! (But, can you really trust somebody who does not know how to spell his own name?)
Re Michael Mann (aka Michael-the-Magician Mann): You (and yours) are SO GOOD at scamming the system. Gotta hand it to ya.
I guess one way to bring credibility to a preposterous paper (MBH hockey stick) is to write another that is far more absurd, thus making the previous one look sane in comparison.
Medieval Windy Period? Got a nice ring to it.
Right now sometimes I think we’re living in the Medieval Hot Air Period, at least when it comes to climate science. There they go again!
Everybody here knowas about the just outcoming study from Christopher W. Landsea ?
Impact of Duration Thresholds on Atlantic Tropical Cyclone Counts
More:
http://rogerpielkejr.blogspot.com/2009/08/mann-et-al-unsmoothed-landsea07.html
Re: Roger Pielke, Jr. (#60), Can one of you smart guys explain to me (and perhaps one or two others) what Dr. Pielke is saying in his post referenced in #60?
Published 13 August 2009, where do potential misconceptions come from ? Referees ?
Re: UC (#61),
That’s what I thought
.. Ah, page titled ‘Response to Reviewer Comments’ . Unprecedented (to me) tone in the responses.
Re: UC (#62), Re: steven mosher (#64),
I thing CA has had an impact on Mann practices. He’s published the code and data for the second time now on a paper he was lead on.
For Gore [Sr.] so loved the world that he gave his one and only Son, that whoever believes in him shall not perish [under rising sea], but have eternal [carbon neutral] life.
– Mann (3,16)
Re: Calvin Ball (#63),
Make that “Sun,” rather than “Son,” and I think you’ve got a keeper!
Where are you, Judith Curry, to remind us that Mann’s TC counts are pass`e and mean nothing in the bigger picture when associating TCs with SST. Remind us all, one more time, that we need to look at the %Cat45 metric.
Re: Kenneth Fritsch (#66), I recall Dr. Curry saying that no one says that frequency has or will increase in the future (except for Holland and Webster and Mann, of course). Roger Pielke Jr. has this nailed.
Re: Ryan Maue (#70), The silliness of this all is breathtaking.
Mark
If you want some grins compare these two files
Proxy-Driven Predictions [Year | Predictions ]
Standard case (uses Mann et al ’07 adjustment of historical TC series) (AD 500-1850)
TCStatModelRecon.dat
Alternative Case (uses Landsea ‘07 adjustment of historical TC series) (AD 500-1850)
TCStatModelReconLandsea.dat
Re: steven mosher (#67), I’m guessing Mann made some “adjustments” that just happen to reinforce his methodological result?
Mark
From one version of data mining to another… ahem. 😉
Mark
Two points,
First, if there were more hurricanes in the ‘Medieval Windy Period’ then presumably it was warmer then than it is now. But Mann rose to fame on denying there was a Medieval Warm Period, although he later had to concede there was a ‘Medieval Climate Anomaly’ (MCA). Does the MW(indy)P now mean the MCA was really a MW(arm)P?
Second, the ‘misconceptions’ referred to above by UC can surely only relate to queries raised by the papers reviewers. One wonders how kindly they will take to their queries being labelled ‘misconceptions’. Mann’s smugness may yet be his undoing.
Re: Dave Andrews (#72), I asked that same question, but after reflection, I remember the Team regrouping and claiming that there was a North Atlantic Medieval Warm Period, but it wasn’t global (as evidenced by the fact that California was in the “sweet spot” that represents the world, and the bristlecones where there).
So the Team story is that 1) the world temperature was perfectly flat from ~1000 to ~1900, 2) there was a Medieval Warm Period and a Little Ice Age, but they were both limited to the North Atlantic, and 3) this North Atlantic Medieval Warm Period caused the Great Atlantic Medieval Windy Period.
The only piece missing is the connection between the North Atlantic and the Central Atlantic (which is what actually causes the ‘canes). If the Team has demonstrated how that happened, I’d be interested.
This all would make for great material for a Saturday Night Live skit, bringing back John Lovitz as the Pathlogical Liar.
Trowser Trumpet
If I recall correctly, Richard Lindzen has stated that there would be fewer hurricanes and similar major weather events under global warming as the latter would tend to decrease the temperature difference between the poles and the equator, that being the main(?) driver for these storms.
If he is correct, that doesn’t fit well with Mann’s logic here. Has Mann dug himself into another hole?
Fingers crossed.
Mann’s predictions of increasing hurricanes can only be believable if you don’t actually notice in the news that we’ve had no named atlantic storms since the start of hurricane season on June 1. His theory looks more accurate if you don’t actually involve real world evidence.
How do we know that the overwash events were not caused by extratropical cyclones? Why does he assume the origin of the sediment events to be solely tropical? The big winter storms can have winds equivalent to category 2 or 3 hurricanes and can move large amounts of sediment.
A paper by Besonen et al is reviewed at http://www.worldclimatereport.com/index.php/2009/02/09/1000-years-of-boston-hurricanes/#more-359
where it is stated:
Besonen et al. conclude “Hurricane frequency, as recorded at LML, has not been constant over the last millennium; the 12th –16th centuries had a significantly higher level of hurricane activity (up to 8 extreme events occurring per century) compared to the 11th and 17 th–19 th centuries when only 2–3 per century was the norm.” Similarly, they conclude “The LML sedimentary record provides a well-controlled and annually-resolved record of category 2–3 hurricane activity in the Boston area over the last millennium. The hurricane signal shows centennial-scale variations in frequency with a period of increased activity between the 12 th–16 th centuries, and decreased activity during the 11th and 17 th–19 th centuries.”
Seems to be the opposite of what Mann is claiming.
Besonen, M.R., et al., 2008. A 1,000-year, annually-resolved record of hurricane activity from Boston, Massachusetts, Geophysical Research Letters, 35, L14705, doi:10.1029/2008GL033950.
My problem with Climate Audit:
My comments here do not refer to Mann’s paper but to the tenor of CA. There are two issues of note. First, I believe that scrutiny of all science and statistical rigor is important. And CA does this. But it is a tarnished scrutiny because it comes from only one view: that of the climate change skeptic. Simply, it is not an unbiased view. In my visits to CA I feel to see any substantial discussion of the scientific arguments of those who tend to dismiss climate change and global warming as either ‘hoax” science or simply bad science. In fact I have mentioned my concern about selective scrutiny to both Steve a number times. There has been little or no discussion of the papers presented at the Heartland Conferences, for example. To my recollection there has been no scrutiny of the science in error presented at the HC (e.g., Gray, Kinninmonth, …..) or cherry-picking (Idso, …..). Perhaps the HC is off limits because Steve was a major presenter.
I believe that strict scrutiny is very valuable to science. I believe that CA could play a more important role if it were not perceived as a blog with an agenda. But the bias (to me) is very evident and, unfortunately, I see this as a major opportunity lost.
Peter W
Re: Peter Webster (#81),
I’ve never met, nor heard of, any climate change skeptics. Moreover I don’t think there is any such beast. Everyone I know has faith in the existence of climate change.
On the other hand I know a few AGW sceptics.
Have you, or will you, express a similar view over at [sur]realclimate, or do you detect no “agenda” there?
Re: Peter Webster (#81),
This blog serves an almost singular purpose in the entire climate field, which needs a more thorough auditing than Steve McIntyre has time. ..
Re: Peter Webster (#81),
You should visit more often. It is obvious you have missed the theme of CA. It is not that of a climate change skeptic as you suggest. Steve’s main purpose is duplication of results and validity in the use of statistics. He does not express a view on climate change. Maybe there is a reason he has been requesting data and requesting the archiving of programs.
Re: Peter Webster (#81),
I tend to agree with Peter Webster. . The AGW field has been politicised and this is a detriment to all of us because of the importance of the issue. However, I believe that his description of Climate Audit is misplaced. Because of Steve Mcintyre, Climate Audit is an exception. The comments may stray to closely to advocacy but this is not because of Steve Mcintyre. he always trys to retain the focus on the science. Unfortunately, this is not the case with other blogs – even very noted ones
Re: Peter Webster (#81),
Peter, I understand and sympathize with your objection. I like crossfire and find it stimulating.I’m not trying to persuade people of things – other than the value of full, true and plain disclosure and careful due diligence.
Editorially, I would like more crossfire quite frankly. From time to time, I’ve tried to solicit critics of this site to provide articles of the type that you urge. I offer total editorial freedom to such authors.
For example, after Michael Tobis, for example, slagged me on his blog,making some repugnant accusations, I sent him a password providing him author privileges here if he wished to co-post threads here. He hasn’t availed himself of the opportunity. I’ve made similar offers to others and had negligible response.
As a writer and researcher, as opposed to an editor, I have only so much time and energy and am not finishing many core objectives. Plus I get tired more quickly than I used to and I can’t do all the things that I’d like. For example, I’m regularly criticized for not formalizing some of the ideas here into journal publications. For example, if I look back at the topics that were on my mind when I visited Georgia Tech, there are many points that are not reflected in journal articles, contained a somewhat original perspective, and, in my opinion, had statistical interest. If I could clone myself, I’d spend time on those sorts of articles.
While editorially, I would like more crossfire, I’m reluctant to spend my own time generating it. There are people who are quite energetic in dissecting “skeptical” articles. Schmidt and Benestad 2009 is a recent example, Santer et al 2008 is another. Any of them is welcome to present their views here without editorial control.
But in allocating my own time, I like to do things that no one else is doing and no one seemed to be doing the sorts of things that I’ve been doing.
I was only at the Heartland Conference for one day, as I had already planned a trip to Thailand to visit my son. HAd I not gone away, I might have covered it a bit more. I was interested in a couple of water cycle feedback presentations, but I’m not familiar enough with the issues to really cover the topic.
I tend to be attracted to articles that are in the news, particularly if they involve proxy reconstructions, which is an area that I have some expertise. Because of that, I’m now in a position that I can write up a comment on a proxy article quite efficiently and place them within a context. I’m unable to do that for an article on cloud feedback or something like that. But I’d very much appreciate any contribution that you or any of your associates were able to make (And have provided Author privileges to you, if you wish to use them – just sign in: sign in on the right panel.)
Regardless, thanks for stopping in. I understand your comment and why it was made.
Cheers, Steve
Re: Peter Webster (#81),
Hi Dr. W. Thanks for dropping by. let’s roll tape.
1. I see no logical argument whereby you get from
A.) there is one point of view … to
B.) therefore, it is a tarnished scrutiny
I fail to see the logic here. For example, we could visit a site that worked to combat Creationist arguments. There would be one point of view. Would that therefore on websterian logic be a tarnished scrutiny? The scrutiny stands or falls on the facts of the case, not on the motives of the people presenting those facts or their proclivity to take up certain publications. Let me put it another way, the ESTABLISHED science of climate studies provides steve with enough fodder for his interests. If the best science gets things wrong why waste your time shooting fish in a barrel.
Let Tamino go beat on them, he does a better job attacking sceptics than he does defended the mann.
2. Your premise is wrong. There are plenty of lukewarmers here. People ( yours truly) who believe in AGW and demand Open data and open code and open debate in business government and science.
3. There is one point of view on Real climate, one point of view on open mind.. take your pick. and so according to websters logic that means……..
4. Did you miss the scrutiny craig L took here?
Why would we waste our time on “science” that is out of the mainstream. As a believer in AGW I want to expend my energy in places where I think I can make a contribution. Simply, lobbying for open data,code and debate. I can’t do that on RC. Been there, tried that. I can do that here and nobody flames me for believing in AGW.
We try to admonish people for motive hunting.
Well, sack up, pick your favorite bad science from HC and request that you be allowed to demolish them here.
Don’t ask others to do what you are unwilling to do yourself.
Re: Peter Webster (#85), Well Dr. Webster, the onus is on those who are making alarming claims. If it were not for CA, Steve McIntyre, who would have had the guts to go after your 2007 co-author Michael Mann and his Hockey Stick? You?
snip – OT
Re: Peter Webster (#85),
Another post by a climate scientist and another scolding without imparting any science. Talk about opportunities lost.
Should we always look to the scientist as unbiased or without an agenda because, well, because scientists cannot have agendas? It is rather apparent to those that read papers on TC activity that certain authors’ papers are going to have the same general conclusion. That situation cries out for analyses of their work.
Those scientists’ agendas or that of a blog or even the perception of an agenda has little or nothing to do with the validity of their conclusions and analyses.
Re: Kenneth Fritsch (#105),
At the risk of running out into the OT weeds,
– snip. Yup.
None of what you mention is being published in the MSM or journals, either. Skeptical articles routinely get critiqued here, maybe you just don’t prefer to notice that.
Tell me, Peter Webster, do you post the same concern over at RC?
Mark
I think post #81 should be deleted for having editorialistic content!
Wait… message coming in on the headset… … …
Oh, it’s THAT Peter Webster. Aw, OK then.
A thousand pardons. I truly did not know who that was. 😉
Andrew
Peter, who better than yourself to fill the gap you identify? I’m sure Steve would be happy to turn over some blog space to you for analysis posts, and I think it would make a welcome addition to the blog.
I’m a neuroscientist. If I’d stick eight electrodes into the brain, all of them somehow located in the visual cortex or close to it (equivalent to “Eight sites along the U.S. East Coast and Puerto Rico provided a reliable record of the number of significant hurricanes going back about 1500 years.” a la Mann et al.), and then claimed to know what the brain does, or even better what it will do in ten minutes, or what it did ten minutes ago… I’d be considered out of my mind.
All I see in Mann’s publications are sampling problems, smoothing problems, oversimplifications of a nonlinear time-variant system.
“Medieval Windy Period”… sounds like a lot of “Hot Air” to me.
I just read that wind speeds in the US are declining steeply since 1973 thanks to global warming:
“It’s a very large effect,” said study co-author Eugene Takle, a professor of atmospheric science at Iowa State University. In some places in the Midwest, the trend shows a 10 per cent drop or more over a decade…”
I think you are remiss, Steve, if you do not point out that you tend to concentrate on articles relied upon the IPPC (or written by those more heavily involved with the IPCC), and in that subset, you tend to focus more on those that withhold data and/or code. How many of the Heartland presenters have work that the IPCC relied upon? In that small group, how many are likewise withholding data and/or code?
.
Also, as I recall, you began this site almost exclusively to counter the constant smearing you received (and still receive) from RC, so it is hardly a surprise that they also wind up the focal point of many, if not most, of your critiques!
.
Perhaps Peter should concentrate on cleaning up his own backyard before telling you how to clean up yours.
.
Mark
Can I second the motion for Peter Webster to contribute articles here. Peter? How about it?
Hi All,
Thanks for the comments. I think I finally have made it: “Websterian logic”! The important thing is to get something named after you no matter what it is named for! I remember once walking into an AGU talk on mars and seeing a slide with Hadley Circulation (Earth) on one side and Webster Circulation (Mars) on the other. Alas, that didn’t stick. Hadley 1, Webster 0.
But your comments are fair, especially about “cleaning up one’s own backyard”. I think that we have published 4 or 5 papers on tropical storms and climate including the last on El Nino flavors and NATC hurricanes. I stand by the results of the first two (Webster et al. 2005, Hoyos et al. 2006: paper can be downloaded at http://webster.eas.gatech.edu). These talk about an increase in intensity and not in number of hurricanes. If the former is wrong and there is no trend, the error in count would have to be -50% in 1970-75 and (if I recollect properly) +40% in 2000-2004 in major tropical storms. Possible but not likely. Carlos Hoyos’s work points out quite elegantly that the majority of information is shared between increase in TC intensity and warming SSTs. The Holland-Webster paper dealt with NATL storms and suggested an increase in number during the last 150 years but also (as per the earlier two papers) a change in the proportion of major storms. To be honest, I have not revisited the paper relative to the undercounting that Chris Landsea suggests. My main interest remains with intensity changes.
Actually, I would prefer that there has been no change in number of TCs globally. This fits well with a developing hypothesis about the role of TCs in the heat balance of the planet. Carlos Hoyos and I have shown that the area of convection in the tropics remains the same even though (i) SST increases so that the threshold from subsidence to convection occurs at higher and higher SSTs and (ii) the integrated heating within the convective area increases exponentially. I would like to point you to a reference but, alas, not all papers are immediately published. Point is that the development area of TCs stays the same but the threshold to higher intensity is reached more quickly as background SST increases. Back to the NATL. This is the ONLY ocean where the area of convection has increased as SSTs increase. Is this consistent with the possibility of more TCs in the NATC? Not sure but I am not quite prepared to clean up my Holland-Webster backyard just yet. But if we are wrong, we will be happy to do so.
Of course, the bad science I refer to does not get published (just as some good science doesn’t either!). But it is out there and picks up a lot of press coverage although not as much as the Mann et al paper. Therefore I suppose I can understand why CA concentrates on those papers (leaving aside issues such as data access for a while). So is it worth time showing that Bill Gray’s hypothesis is unphysical and incorrect (all he needs is for saturated vapor pressure to decrease with increasing SST (contra-Clausius-Clapeyron) and for evaporation to decrease with increasing SST and wind). Is it worth pointing out publically that Bill Kinninmonth leaves out a vital feedback in his calculations (he and I have communicated extensively on this issue and, to be fair, I don’t think I have convinced him). Judy C thinks it is tilting windmills and not worth the time and effort.
For what it is worth
Peter Websterian (I like that!)
Re: Peter Webster (#101), Hehe.
Glad you liked that Dr. W. Hang around, you and fellow hornet Dr. C are always a blast of fresh air and some of the venom here convects away when youall visit. Even the french dude, whatshisname, is welcome.
Re: Peter Webster (#101),
Peter, I find that you and Judith Curry, when putting forward % increases TC activity, apparently do not warn of the probable cyclical nature of TC activity that might be putting your start time at a valley and end time at a peak. Do you judge that this activity is cyclical and thus sensitive to the time interval used? I am referring here to TC counts and to %Cat45.
Sure it is. I and others here would be interested.
Peter, let me add a longstanding complaint of mine about IPCC presentations.
The terms of reference as implemented by WGs are, in effect, 5-year literature reviews. As a result, in no IPCC report, is there an exposition of the infrared radiation physics and how increasing CO2 actually increases temperature beyond a grade-school cartoon. The usual response to this observation from climate scientists is – take an undergraduate Atmospheric Radiation class. I don’t think that this is an acceptable answer.
Water cycle/cloud feedback is clearly the biggest question mark in whether increased CO2 is a small, medium or huge problem. If I were framing IPCC reports, I would require large sections on the issues that contribute most to the scientific uncertainty.
From a policy management point of view, policy makers should be clearly informed about which science issues were most uncertain and what steps or programs are required to reduce that uncertainty. The AR4 section on cloud feedbacks, for example, is disappointingly brief and, based on the relatively small number of current references, one would be forced to conclude that this was a backwater of present day research efforts.
Websterian sounds very Oxbridge – Bodleian, Websterian.
Off topic, but brought up by a commenter: (can move to Hurricane thread)
Re: Peter Webster (#101), Kenneth Fritsch, myself, among others are waiting for the physical explanation of why the % of category 45 / category 1 has increased during the past 30-years. This is not consistent with MPI theory and cannot be — please do a little back of the envelope calculation to show yourself that it isn’t.
As per Hoyos et al. (2006) confirming anything: it does not confirm Webster et al. (2005) in terms of % of category 4-5/total hurricanes. Pretty simple to see this: NCAT45 or frequency is used in the Hoyos paper for the mutual information analysis. Apples and oranges…
snip – venting
Re: Tom C (#107), Tom don’t speculate on the motives of Dr. W or Dr. C
WRT Mann.
#101 Peter Webster
Quite to the contrary, it is naive in the extreme to base the edifice of a GCM model on the simplistic assumption that higher temperatures will lead to an increase in humidity in strict accordance with Clausius-Clapyron. This might be true in a beaker in the lab, but in the real world, higher temperatures might well lead to secondary physical processes that would provide a countervailing effect. Then again, maybe the humidity increase would be amplified by secondary processes. snip
Steve: Let’s wait for a proper exposition of the issue rather than debating passim comments.
After presenting my non technical scoldings above, I would like to make a few observations about the paper in question in this thread.
Figure 3 in the paper is most revealing in that, when one looks at the reconstructed area, the statistical model and the regional sediment composite reconstruction do not show good agreement. To which the authors note that:
On the face of it the authors are only reporting what the graph in Figure 3 shows, but what is really evident is the general patterns of the model and composite do not match and the overlap of the uncertainties would not be expected to fail if the uncertainties are sufficiently large.
What is impressive, on the face of it, is the match of the instrumental TC count to that of the model and sediment composite record during the instrumental time period (1850-present). I know from the work that I have done on fitting the TC counts of the NATL to Poisson distributions that the SST variable does not improve the fit if one acknowledges the confounding of the change in dection capalities over time with changes in SST. I did this by using the Easy to Detect TC counts that were developed by David Smith. I agree that compensating for the phases of the AMM, ENSO and NAO indexes and the like can improve the Poisson fits.
What really puzzles me and I need to look at it in more detail is how the authors obtain such a good fit with sediment composites for landfalling hurricanes in the instrumental time period (1850-present) when overall the landfalling events have been shown to have no statistically significant trend during that period. Notice, though, that the authors weight the regional composites (landfall events to TC counts) and I believe change those weightings in one region for a given reason that could be rather arbitrary on further investigation. I suspect it will be these weightings that translate an overall no landfall trend into an increased trend for the corresponding TC counts.
Re: Kenneth Fritsch (#109), Thanks Kenneth, When you read Mann response to “objections” you’ll find he addresses the issue of weighting the collecting sites. His defense struck me as weak and I wanted to see a more ….here it comes… ROBUST defense of the weighing scheme as the selection of those weights seemed to be the handle that cranked the rest of the machine. That just a “perception” without going through the maths.
The 8 coastal sites that Mann has in his data have only 46 hurricanes recorded in the last 1000 years (actually less due to dual reporting in New England). It seems a bit much that one can deduce yearly counts from such sparse data.
Even with a small number of sampling sites concentrated in one area collecting sparse data you could gather enough to project a pattern for that area. I think the danger would be projecting the valid trend for the sample area onto a larger map. What if Cuba’s storm landings followed an inverse trend? You really wouldn’t know.
If he hasn’t played games with his data, its a valid piece of the puzzle.
Nature published this in the “Letters” section. What are the characteristics of a Letter in Nature?
Dr Webster,
May I make a suggestions for you first thread on CA?
As you are now legitimately (it would seem) in possesion of Dr Phil’s Wizardian station data, could you please spend some of your no doubt valuable time auditing it please?
In particular we would like to know what adjustments/correction the good doctor makes for non-climatic effects e.g. UHI when he derives his anomaly index. I’m sure Steve will be happy to provide you with any data he has that should hopefully help you in your task. Steve can then audit your work and we can then all comment on it and give you some valuable feedback (just as we wid for Craig L.
Since you are a bonefide climate scientist I’m sure you won’t have any problem in getting any paper you write that documents your analysis published in Nature or Science after Drs Phil nd Mike have peer-reviewed it.
KevinUK
My first guess, that the apparent good fit of composite landfall sediment proxy to the instrumental TC count was the result of the Mann weighting of the composite regions, has been shown on further reading to be incorrect.
Those composite to instrumental measures to model fits look suspicious to me and I am now thinking that the 40 year smooth and possible end point manipulation may play a role. Since I know from Mann’s SI that the regional normalization does not materially affect the apparent fit for the composite sediment proxy in the instrumental time period (1851 to present), I can simply look at the landfall record and compare it to the TC counts for this period – without the smooth or other Mann handling.
The smooth is referenced to Mann, M.E. Smoothing of climate time series revisited. Geophys. Res. Lett. 35 doi:10.1029/2008GLO34176(2008).
So having broken Mann’s hockey stick, now you’re out to break Mann’s wind?
Nick, you appear quick to cry that things are getting technical when you are proved wrong, rather than admit your error. The difference is really not that subtle – there is a clear distinction between Supplementary Information that is required to be submitted as part of the publication, is referenced in the publication and is maintained on the journal’s website as compared to data the author voluntary places on his website. Mann is to be commended for having released this data, but it should have been part of the SI and, as Steve points out, even that data is significantly incomplete. Why not just accept this?
Re: compy (#134), Compy, on the contrary you have been proved wrong. You said that Mann had provided no data or code (and I should join in badgering him to do this), then when I pointed out that he had, you said no, there was nothing about SST or Nino3. Then when I pointed out that yes, that was there, you take refuge in technicalities in what the files are called. What’s the point?
Do you know whether Nature is prepared to host data and code repositories? I suspect that they won’t, so Mann has to provide it from his own site, as he did.
Incidentally, after all these aspersions cast at Mann for alleged failure to provide code and data, would it be too much to ask that you acknowledge what he has provided?
Re: Nick Stokes (#135),
I’ve added the following in the thread:
I had written:
Since the publication of this post, as noted above, a Supplementary SI has been identified at Mann’s website – that was not linked in the article or SI. Before writing the above paragraph, I looked at Mann’s website and was unable to locate any reference to the Supplementary SI.
Re: Steve McIntyre (#136), Thanks, that is helpful. Do we know if Nature will host code on its SI? It seems to me that it would create legal difficulties. Nature might acquire liability for the performance of the code. And then they might have to license its use, which could then lead to restrictions on the use others make of it (maybe even including Mann).
Re: Nick Stokes (#137),
Econometrics journals require authors to archive working source code as a condition of review. There is no reason why climate science cannot adopt a “best practices” standard.
I don;t want to waste time discussing such an absurd legal fantasy and I don’t want readers to discuss it either.
Re: Steve McIntyre (#136), This page at the ESSC at Penn State has a collection of information links, including the “grey” SI page with data and code, and also a pdf of the article.
Nick, it is getting tiresome discussing this issue with you so I promise to stop after this. There is nothing I said that is incorrect. I summarized Steve’s point as being that Man et.al “relies on unpublished articles, provides no data, no code and includes limited description of methodology.” As Steve has ably pointed out, this was completely accurate. Steve also specifically referred to Supplementary Information which is not only a known term of art but specifically described in the article. So there can be no ambiguity about what Steve was referring to. A subsequent posting of limited data on Mann’s website in no way invalidates the comment.
As regards your comment:
I already said: “Mann is to be commended for having released this data.”.
As I noted in a previous post the impressive part of the Mann reconstruction was the pattern matches of the sediment proxy, the Poisson model and the observed TC counts for the instrumental period of 1851 to present. The matches for these time same series, excluding, of course, the observed, for the period from 500-1850 was very unimpressive and appropriate to the Steve M comment from G. Marx on whether to believe Mann when he says the “..two entirely independent estimates of past cyclone activity were found to be statistically consistent..” or my lyin’ eyes that see no pattern matching of the series and notes that if the CIs are sufficiently large any two time series can be said to be statistically consistent. Therefore, my interest zeroed in on the 1851 to present period.
In Mann’s grey SI posted at his site and linked here:
http://holocene.meteo.psu.edu/~mann/Nature09/
one can see that the regional weighting, or lack thereof, does not materially affect the fit and shapes of the three time series in the instrumental period. One can also see that removing the non US part of the sediment proxy does not materially affect the shape or match of the 3 three time series in the instrumental period. As a result, one should be able to note a similar pattern for the TC and US land falling hurricane in the instrumental period. To this end, I plotted the Mann TC counts on an annual, decadal and multi-decadal basis per his SI and present the three time series below. It shows the cyclical nature that Mann shows in the 2009 paper for TS counts along with an upward trending count over the instrumental period. For comparison I show the same information (and again from Mann’s SI) using Landsea TC counts. Again one can see the cyclical nature of the counts, but with Landsea one sees very little to no trend in counts.
The next step in my analysis was looking at the US land falling hurricanes as referenced in the unpublished paper by Pielke Jr and Steve M on trends of land fall and nearer to land fall TC activity over the instrumental paper linked here:
Click to access herrick.pdf
and using a link from that source listed here:
http://www.aoml.noaa.gov/hrd/hurdat/ushurrlist18512005-gt.txt
I reproduced a time series of observed land falling hurricanes for the US and present it in the third and final graph below. What I cannot comprehend is how Mann was able to smooth his series (US and Caribbean) for land falling sediment composites to be so different than the actual land falling hurricane frequencies, i.e. it would appear that the land falling sediment hurricane proxy does not correlate with the observed overall land falling hurricanes.
I need to look in more detail at the Mann grey SI, but at this point Mann and authors in the 2009 paper evidently did not consider an overall land fall hurricane to TC count correspondence. Rather the authors are looking strictly at the land fall sediment evidence at their rather limited selection of sites and relating that very limited land fall hurricane activity data to the TC activity in the instrumental period and thus into the the 500-1850 reconstruction period.
If this is in deed what they did, one then must have major concerns about the susceptibility of this whole process to cherry picking those sites and ask the obvious question why the TC counts have a better correspondence with the selected sediment sites than to the overall observed land fall hurricanes. Landsea assumes in some of his analyses that the overall land fall frequency should correlate with TC counts.
Interesting comment by Mann at RC; he apparently wants everyone to only get their information from sources he controls:
http://www.realclimate.org/index.php/archives/2009/08/resolving-technical-issues-in-science/
14 August 2009 at 6:22 PM
[…So any ‘inconsistencies’ claimed by detractors are either imagined, or manufactured in an intentional effort to deceive readers about what the study actually shows and claims. I would encourage any readers to get their information from the paper (and supplementary information), the various press releases, interviews (including ones I did for NPR and PRI), and a video conference I did for NSF. That can all be found here. In addition, corrections of specific misconceptions about the study (such as some of those described above) are available here. -mike]
After reread of Mann (2009) and the review comments, I am fairly certain that, indeed, the authors used the sparce regional composite sediment data in matching and displaying that proxy result with the TC count instrumental record.
I was thinking: why did it take so long to see what the authors did? The answer is that Mann and coauthors are using a correspondence of land fall hurricanes to TC counts in the sparce sediment proxy regions, i.e. they are assuming that land fall hurricanes have a direct relationship with TC counts in the regions of interest at the sparce sites. Yet when one compares the observed land fall hurricanes, in total, with the Mann TC counts in the instrumental period the trends and patterns do not match.
Therefore, the ever more burning question has to be how can one make a correspondence for the regions used and ignore the overall lack of correspondence of land fall hurricanes with the Mann TC counts. A couple of alternative answers seem apparent: (1) the Mann TC counts are in error or (2) the observed hurricane land fall records for the instrumental period are in error and those from the sediment deposit proxy are correct or at least more correct.
If my notions of what the Mann authors are saying and/or implying here are correct, I think we have fodder for further discussion and analyses.
http://www.essc.psu.edu/essc_web/research/Hurricane2009.html is an undated forecast of 2009 Atlantic hurricanes. Did this just go up recently?
Re: Steve McIntyre (#144),
From your link we have:
Since they have published their model, one can go back and make a hind cast for 2008 using the actual climate variables as input.
Re: Steve McIntyre (#144)
The Page Freshness Bookmarklet says that the linked page was last modified on 31 July 2009. You can get the javascript for this bookmarklet here:
http://www.bookmarklets.com/tools/data/index.phtml
It may be useful in your forensic investigations.
Re: frost (#147), tried out your bookmarklet. May be occasionally useful.
I think for the page Steve is curious about it is likely inaccurate.
I went up one folder to ..\research and opened links to a few pages that return the same modified date and the same time within a few seconds. The bookmarklet may only be indicating when the pages were moved into the research folder. What is really needed is the creation date.
The first paragraph of their prediction page ends with “… hurricane season, which starts on June 1st.”
I expect they would have used the word “started” instead of “starts” if the prediction was after June 1st. Though with these guys you can’t take anything for granted.
Their linked SST anomaly image is from 27 Apr 2009. It’s in a folder of similar images created every few days up to present.
Re: frost (#147),
Nifty link, thanks!
I have now looked at the hurricane land fall events from the regional sediment proxy in the 1851 to present as presented in the grey Mann SI. They are as follows:
New England: 1991, 1960, 1954, 1938
Mid Atlantic: 1893
SE Coast: 1989, 1954
Gulf: none
Caribbean: 1989, 1928, 1899
I am once again confused. How does one get any reasonable “calibration” with this sparse data?
I need to look at the next step to see how the authors worked their magic on these data – perhaps by using the pre-instrumental data. Anyway, we have 10 land fall events here from the sediment proxy and over 250 from the observed records.
Peter Webster’s detection of an “agenda” signal here at CA is as mistaken as the notion that proxies showing a reliable climate signal abound in nature. The only thing that most of us agree upon is that “climate science” is not being done rigorously. And this cuts across all the disciplines involved and the different backgrounds of the individual contributors to the discussion, which centers mainly on statistical issues. Although our host discourages discussions of physics, those of us trained in that discipline are perhaps the most outspoken. When Tom Vonk calls certain core tenets of AGW alarmism “bonkers,” that is the exasperation of a working scientist who knows his stuff. As I recall, far worse tones of discussion could be heard daily in faculty club preening contests.
From a physical standpoint, one can scarcely think of a worse climate proxy than estuarine sedimentation in an area visited by tropical storms. Littoral processes are enormously complex, with a whole panoply of factors (waves, tidal streams, run-off, particle size) affecting the sedimentation rate, which is poorly understood to begin with. The entire layered structure of sediments accumulated over years can be easily disturbed by unpredictable tropical cyclones and storm surges. One may scour away the sediments, another may overlay them with a fresh layer. Much depends upon the angle of incidence of the severest winds and storm waves, as well as the tidal stand. The claim that discrete hurricanes can be reliably identified from the surviving muddle of sediments is based not science, but on the agenda-driven desire to know more than is possible.
Re: John S. (#148),
Well said sir.
Re: John S. (#148), John, you wrote, “Although our host discourages discussions of physics, those of us trained in that discipline are perhaps the most outspoken.”
.
Physics is the most analytical of sciences and the most mathematical. Its methods and conclusions are the most clearly delineated and open to critical inspection. In that regard, after relatively modest study, it becomes clear that climate models can’t sustain a claim of AGW, because they can’t make falsifiable predictions and can’t pass perfect model tests. There isn’t even one valid uncertainty propagation study in the literature. Even phenomenologically, Demetris Koutoyiannis has shown in spades that models tuned to reproduce global temperature trends can’t get regional temperature trends right; i.e., their predictions are invariably falsified. In paleothermometry, physicists have granted physical meaning to principal components merely by fiat. And it’s not just that no one even protests. It’s that the entire discipline falls into line and propagates the abuse. No physicist speaks up. It’s incredible.
.
It’s a puzzle to me, then, why so many physicists have bought into the alarm. Notable skeptic Robert Parks, past president of the APS, regularly disparages “deniers” and invokes oil-company conspiracies. John Holdren and Steven Chu are high-profile alarm-preaching physicists, and there is no shortage of others. Don’t these people read the primary literature?
.
As a physicist, do you have any understanding of this — to me amazing — suspension of critical faculty among so many physicists? I am really bemused by this state of affairs. It’s almost as though AGW has trumped their sense of professional integrity.
Re: Pat Frank (#151),
I too am often mystified by physicists who uncritically accept GCM outputs or arbitrary PC imputations as representing physical reality. This may be the consequence of physics being highly specialized in a wide variety of subdisciplines. Specialists often know little about the state of knowledge outside their own narrow field. They simply assume that “climate science” is as rigorously based on first principles as their own research, which in more cases than not is done under controlled laboratory conditions rather than those of the chaotic real world. Certainly the high priests of AGW do everything to perpetuate the illusion of “settled science.” And, sensing a groundswell of public attention (and funding), human frailure puts the critical analytic faculty of some physicists to rest.
Re: John S. (#153),
I was a chemist in a former life, but what you say, John, about physicists, I think applies to other fields of science as well and including climate science.
Re: Pat Frank (#151),
Most physicist assume geophysics is also an exact science.
It isn’t: it’s an art.
Hans Erren
(Geophysicist)
Re: John S. (#148),
Obtaining a better understanding the Mann paper of interest to this thread has forced me to review some literature sources. The papers linked below do give methods and models for using sparse instrumental data to infer regional hurricane/TC activity. These papers do, and more in contrast to the Mann paper, warn of the uncertainties involved in applying these models. Mann claims to have taken these methods a step further in inferring TC activity for the whole of the NATL basin from 5 regional results and then assuming regional landfall events to TC activity for the region as a constant ratio. I have yet to pin down precisely the methods and models used by Mann (2009), but I think I have a basic understanding of the general models from which he had to chose.
From the paper linked at:
Click to access ElsnerJaggerLiu2008.pdf
We have the statement that “Over the past decade or so, several dozens of coastal lakes and marshes along the Gulf and Coasts have been cored with the hope of better understanding prehistoric hurricane activity”. That several dozen sites evidently available contrasts greatly with the 8 sites (divided into 5 regions) that Mann selected to estimate NATL wide TC activity.
The paper linked here:
Click to access Woodruff_Gcubed2008.pdf
describes the model developed looking at the overwash at Laguna Playa Grande, a site that Mann used in his estimations of TC activity. The paper linked directly above goes to great pains to point the absents of any overwash evidence for periods of 1100, 500 and 250 years in the approximately 5500 year old record as an indication that sedimentation rates (that can be shown to be changing over this time period) can make distinguishing individual overwash events difficult to impossible where the rates were lower. As a matter of fact, the Mann period of interest of 1500 years includes the 500 and 250 year stretches without a single overwash event. None of these periods involve the instrumental period of 1851 to present.
Re: Kenneth Fritsch (#160),
The two papers you reference certainly are more thorough and transparent in methodolgy and more circumspect in their conclusions than is Mann’s latest opus. The estimation of the average “return periods” of Gulf hurricanes of various intensities would have benefitted from using the Weibull extreme value distribution, rather than the exponential, but that is just a minor reservation.
The major reservation concerns using overwash deposits as a proxy for hurricane frequency. Supposed proxies should be no less thoroughly benchmarked than a scientific intrument before going into the field. The basic requirements of fidelity of sensing, linearity of response, and immunity from extraneous influences can scarcely be met by overwash sedimentation. I suspect that very long, sporadic gaps between registered events appear in the record because the highly intense hurricanes have scoured away the deposits of lesser ones. The morpholgy of the Gulf littoral has undergone some severe changes in recent millenia. Thus the record is incomplete. The estimation of average return periods is quite critically dependent upon having an exhaustive record of extremes. Even then the estimates can be quite tenous, because natural extremes tend to cluster temporally at the very highest levels.
I have yet to see a climate proxy record with truly impressive coherence at centennial and millenial time-scales. Mann’s claim of explaining 50% of the variance of NATL hurricanes, even if sustained, would be tantamount to a S/N ratio ~1. Good science weaves threads of evidence that harmonize like the strains of a symphony. With that level of noise, even Mozart would sound harsh.
Re: Kenneth Fritsch (#160),
I have perused Dr. Mann’s Nature letter. Seven of his nine sites are subject to winter storms known as nor’easters. I have a question. How would you separate the effects of nor’easters from hurricanes given that nor’easters are often as strong as hurricanes and occur much more commonly along the US Eastern coast than hurricane landfalls do?
I have another question. Why would he only use one Florida panhandle site (near Panama City) to be a proxy for the Gulf of Mexico when there are differences in structure between the Eastern Gulf and the Western Gulf and the middle area where the Mississippi River delta sediment is deposited?
Re: cdquarles (#166),
I do not have a good sense of the frequencies of northeastern storms or their intensities, but I suppose one would look at the over wash deposits (if any) coinciding with these storms during the instrumental period and determine whether they can be distinguished from a hurricane event.
As I recall I have read some concerns about storm surges, other than those from hurricanes, potentially muddying the waters, but do not recall if the issues were arm waved away or met head on.
It would appear that Mann here has done what he often does with his reconstructions and that is take proxies that the original developers have reservations about and “spin” them into something original and without emphasizing the reservations. He is also adept at the selection process and I need to dig deeper to see if the dozens of other over wash deposit sites would be eligible for his study and what a sensitivity test might reveal.
Of course, what is put into the background is that Mann’s empirical TC counts with minor adjustments show a significant trend in the instrumental period while the one he reproduced in his grey SI from Landsea, which has major adjustments, shows no trend during that time period. When one looks at the observed land falling hurricanes one sees no trends. Since Mann associates landfall events with TC activity in the surrounding region, why would he not concede the Landsea empirical TC count series since they appear to associate well with land fall events in the instrumental period and then go from there with his proxy using over wash deposits.
But wait a minute, that would be a Mann proxy without a hockey stick shape and it would mean that in his TC count model he might well be confounding changes in detection capabilities with changes in SST.
Re: Kenneth Fritsch (#168),
Thanks Kenneth. That’s what I thought. I have lived in the southeastern corner of the US for all but one year of my life. Nor’easters occur quite often. When the atmospheric conditions are right, you can get them repeatedly every three to five days; and this is a yearly or nearly yearly thing. Big nor’easters are a winter storm phenomenon for the most part, and they are quite destructive when they dump large amounts of frozen dihydrogen monoxide along with the high winds. Remember the super storm of 1993? That one even affected my neck of the woods. Thundersnow occurred at 32.5N 88.5W. Think about that one for a minute, and this one was in mid March.
About the muddying of the waters, so to speak, the high frequency of nor’easters might just make the background more uniform over time than if they weren’t so regular. Nevertheless, I don’t think that you can assume uniformity for the frequency of storms in the time series for any weather phenomenon. Might be wrong about that belief, but I’d like someone produce accessible data to test the hypothesis.
Re: cdquarles (#169),
My reading of the papers that I linked above is that the most favorable conditions for detecting individual hurricane over wash deposit events, i.e. separating near time events, are those where the background deposit rates are the greatest. I think this means that an intervening deposit layer is needed to distinguish individual hurricane over washes events that might occur with only a few years of intervening time.
The Woodruff paper linked above mentions that the lower background deposit rate in certain historical time periods explained 65% of the drop in apparent hurricane activity at the LPG site. I tend to get a little concerned about how these estimates are made.
Anyway I get the idea that more frequent northeasterners deposits, if distinguishable from hurricane over wash, would increase the resolution of hurricane events.
Re: Kenneth Fritsch (#170),
Hi Kenneth. I have perused the two linked papers you provided. What a refreshing contrast to the Mann letter. I see that some still observe the “Don’t make vast conclusions from half vast data!” rule.
Re: Kenneth Fritsch (#168), Why indeed? I think that summarizes the issues well.
Interestingly, I don’t see the same problem among engineers. If anything, by and large, they are skeptical; from the first-year new-grad through the seasoned PhD, skepticism reigns. Of course, engineers tend not to have high profile positions, either, they tend to work in arenas that have nothing to do with any sort of public policy, at least, not the sorts that would concern AGW theory.
Mark
Re: Mark T (#152),
Thank God it’s engineers who design bridges.
Re: Mark T (#152), Not only that, when engineers are big-time wrong, they get big-time fired. Kind of makes you want to make sure you’re right. 😉
Re: Mark T (#152), Among my colleagues who are applied foresters/wildlife biologists (ie, working for industry–almost engineers that is) I detect 90% sceptical and almost 0 foaming at the mouth AGWers. Again, it is a practical field.
Radar…
Mark
I’m not sure if it is by training, or because our minds started out that way. In other words, are we engineers because we are skeptics, or did becoming engineers make us skeptics? This isn’t limited to AGW issues, btw.
^Actually, rarely does an engineer get fired. Typically an engineer would get hassled to the point that he left on his own. 😉
Mark
Re: Mark T (#159), Yes, because that way the employer doesn’t have to pay unemployment benefits. 😉
Dr. Mann, I have a small request:
Although I appreciate the fact that you have released some data and code, would it be too much asked if you also released the rest of data (e.g., *.dat files in the directory “Data”)? I mean, it would be nice to actually be able to run the code you released. Thanks!
The sad part, cdquarles, is that your questions are likely rhetorical.
Mark
Anyone who has spent time along the southeastern coast (on foot) will notice that big storms change the beach landscape. Typically, a storm surge will flood the near-coast, but as it retreats it will carve out new channels as the water returns to the ocean. New dunes will also be created. The next time, a surge will respond to the new topography. For Mann’s analysis to work, the topography would have to remain constant for 1500 years so that a storm surge would overtop a barrier dune and leave a deposit in the same manner each time. This seems impossibly uniformitarian. That is not even counting dune migration and sea level rise (which Steve mentioned) over such a period. 1500 yrs ago it could have taken one heck of a storm to overtop dunes which were many feet farther above sea level.
Florida looks the way it does now because of the storms. The coastline is the path they take. It isn’t changing further now because of the continual efforts to prevent erosion (by shipping sand in.)
Mark
Re: Mark T (#173), Visit the gulf coast where Ivan or Katrina hit and tell me it isn’t still changing (in spite of our best efforts)
Oh, I know it is. Even my comment that the Florida coast isn’t changing should be stated as “sort of.” In spite of our best efforts, attempts to constrain nature are largely in vain. Keep in mind, I was also living in St. Louis during all the flooding in the early 90s.
This just serves to point out that an assumption of all other factors being constant w.r.t. climate is typically nonsense. The concept of “constant” is man made in this regard, and actually requires man’s intervention in the first place.
Mark
I couldn’t find the article, but there is a nice article where they use the historic maps of Charleston SC to show the progression of barrier islands at the mouth of the harbor. They used maps centered to make an old style cartoon. One can see an island at the top-north forming and progressing southward until it is reabsorbed at the bottom-south.
I also could not find the article that had an estimate that barrier islands, and the marshes they create, are on the average destroyed every 500 years due to erosion and hurricanes.
More reasons to choose proxies most carefully, and word the publications more so with caveats, assumptions, and large error estimates.
There’s a new paper in Nature on “2,000-year-long temperature and hydrology reconstructions from the Indo-Pacific warm pool“. From the abstract
and from the main text
See also the Editor’s summary Taking the long view of temperature.
RE Jonathan #178
This sounds like a good candidate for Craig’s next update, assuming it’s really calibrated to temperature. Of course, however, the inclusion criterion shouldn’t be that it finds a MWP, but rather that it meets Craig’s criteria of peer review, temperature calibration, etc. This one is bimillenial, which is great, but I don’t think it’s really essential for Craig’s methodology that all go back to near year 1.
I’m on the road and can’t access the paper right now.
Re: Hu McCulloch (#179), my interest was more in the editorial line being taken.
Re: Jonathan (#180), it is also notable that the supplementary information is substantial and seems to include raw data. People seem to be learning at last.
Academic check-kiting: the best & fastest way to erect a house of cards hypothesis.
Possible parallels for “Nature” and other leading journals…. New allegations (in a different discipline) may feature serious numerical and methodological errors, a highly uncooperative principal author hiding behind “this work was peer reviewed” and “the original data has been lost” types of evasions. Will “Nature” require corrections or retractions to improve the scientific record?
2 Trackbacks
[…] M has already read the paper and noticed what is apparently a complete and utter lack of detail More Check Kiting at Nature. By check kiting Steve is pointing out that the only reference to methods points to an as yet […]
[…] McIntyre of Climate Audit writes about “check kiting” related to this […]