I picked a good weekend to be out of cell phone range and unconnected to the internet – and judging from how the rest of the week has gone, I’d have been minded to stay there…
As most readers are probably aware, there was an op-ed in the Saturday New York Times from Richard Muller announcing the Berkeley Earth team’s latest results. It was odd enough that a scientific paper was announced via an op-ed, rather than a press release, odder still that the paper was only being submitted and had not actually been accepted, and most odd of all was the framing – a ‘converted skeptic’ being convinced by his studies that the planet has indeed warmed and that human activity is the cause – which as Mike and Ken Caldiera pointed out has been known for almost 2 decades.
Not wanting to be upstaged, plenty of ‘unconverted skeptics’ – including Anthony Watts and Ross McKitrick decided to stage dramatic press events and release barbs of their own. This was followed by a general piling on of commenters and bloggers trying to spin the events in their preferred direction combined with plenty of cluelessness in the general media about exactly who these people are (no-one special), what earth-shattering discovery had been made (none) and what it all means (not a lot).
The ‘best’ response to this circus is to sit back and see how pretzel-like the logical justifications can become. I particularly like the recent twist to the “No true scotsman” post-hoc rationalisation. Since the ‘converted skeptic’/prodigal scientist meme is a very powerful framing for the media, the obvious riposte for the ‘skeptics’ is to declare that Muller was not a true skeptic. But since these terms have become meaningless in terms of any specific position, this ends up as a semantic argument that convinces no-one but the faithful.
The actual trigger for all this hoopla is the deadline for papers that can be cited in the Second Order Draft of the new IPCC report. They needed to have been submitted to a journal by Tuesday (31 July) to qualify. Of course, they also need to be interesting, relevant and known to the IPCC lead authors. But there seems to be far too much emphasis being put on this deadline. The AR5 report is pretty much 90% written, and the broad outlines have been known for ages. Very few of the papers that have been submitted this week are anything other than minor steps forward and only a small number will be accorded anything other than a brief mention in AR5, and most not even that.
Furthermore, once the SOD is finalised (Aug 10), Tuesday’s deadline becomes moot, and the only thing that matters for the final report is whether papers are accepted by March 2013. (In a spirit of full disclosure, I should mention that I was working on a couple of papers with an eye to making this deadline, but in the end decided it was preferable to take the time to do a good job on the papers than to submit something shoddy).
The only worthwhile substance to any of this is the work that has mostly been done by Robert Rohde on the Berkeley Earth code and database as we’ve noted previously – and once this week’s drama has faded into the distance overshadowed by some new blog-storm, this work will still be a useful advance.
But still the games go on. Senate hearings are one of the longest running games of political theatre going – where the Senators pretend to listen to the panelists and the panelists pretend that this is an efficient way to inform policymakers. This week’s was little different from the ones in the past – some earnest submissions from the mainstream, and a cherry-pickers delight of misinformation from the Republican invitee, John Christy, who even quoted the woefully inept first draft of the Watts paper as if it meant something.
To confuse the metaphor even further, Roger Pielke Sr loudly declared that whatever the results of the Watts paper it will end up being a game changer:
The TOB effect could result in a confirmation of the Watts et al conclusion, or a confirmation (from a skeptical source) that siting quality does not matter. In either case, this is still a game changing study.
If only people would change the games they play…
My inclination is just to sit back and watch the spectacle, admire the logic-defying leaps, marvel at the super-human feats of hubris and, in two weeks time, remark on how little actually changed.
Just in case there was anyone here who didn’t realise how little Anthony Watts understands about science, have a look at the nonsense he has just written about the new Hansen paper.
(The actual Hansen paper is out and the full text with supporting information is available for download here.)
With regard to MangoChutney – apparently Watts is not wanting informed comment from science people. He refused to publish a comment from Dana N alerting him to the suggestions on skepticalscience, and a mod wrote a scathing reply saying to go post elsewhere when someone else pointed to the skepticalscience artile.
Mango Chutney is a fairly regular Contrarian contributor to Richard Black’s blog at the BBC.
Given his track record at Auntie Beeb, it seems entirely possible (in the light of his peevish response to the cogent points raised in reply to his first post) that he was mainly interested in satisfying his preconceived prejudices about supposed ‘Pal’ review in the Scientific community, conspiracies, etc…
Clarification of my last post – Mango Chutney is a fairly regular contributor to the comments section of Richard Black’s blog at the BBC.
9 Eli the Rabett said: “It was a good week for popcorn.”
Where’d ya get that? You must be eating old corn, last years crop. I was just in corn country: Detroit, Chicago, & points in between. As far as I can tell there’s negligible corn on the stock throughout most of the grain belt. The combination of drought and heat got it.
I am grateful to Robert Rohde for making the best single graphic I have seen showing how complex life on this planet relates to lower levels of CO2 over half a billion years. And I am grateful to him for his work on the GeoWhen database (“an attempt to make sense of…the geologic timescale”).
I wonder why the links I have to his website globalwarmingart.com appear not to work. But the page http://en.wikipedia.org/wiki/User:Dragons_flight/Images is not affected.
I first saw the key graphic here: http://www.nytimes.com/imagepages/2006/11/06/science/earth/20061107_CO2_GRAPHIC.html
But it was in an article with garbled language which became the source of a lot of confusion.
John,
Recent rains and cooler temps may have salvalged much of the corn crop in the Northern states (NY, MI, WI, and MN). In some cases, this was too late, but others, not so. The output will still be low, but not a disaster. The same cannot be said for the I states further south, where much of the feed corn is grown. Expect further increases in meat prices, but corn may fall as the farmers begin to harvest.
Although I have read comments insisting that Watts include TOBs in his assessment of his data, I haven’t seen anyone suggesting the Muller do the same. Is that because he did and I’ve missed it or because he is now a convert?
The worst drought in more than half a century in America’s Corn Belt has slashed the corn crop to the lowest in five years, leading to a plunge of corn supplies to the smallest in 17 years by next summer, a Reuters poll of 21 analysts showed on Monday.
That would result in the third year in a row of razor thin corn stocks, keeping prices at record highs and rationing demand for the world’s most popular feed grain, analysts said.
http://in.reuters.com/article/2012/08/07/us-usa-agriculture-corn-idINBRE87603P20120807
From the same article:
Keeney said the rains came too late to help the corn crop but some of the late planted soybeans may benefit.
http://in.reuters.com/article/2012/08/07/us-usa-agriculture-corn-idINBRE87603P20120807
Ian @ 57, Muller took a different approach. When there was a change in protocol, he treated it as two different data series, in effect two different stations.
Ian,
Zeke Hausfather, commenting above, would be the best person to explain but the BEST method uses a breakpoint detection algorithm to seek out any discontinuities in temperature records. When a breakpoint is found the offending record is split into two, each segment treated as an independent station.
Assuming the breakpoint detection algorithm works as designed it should remove any inhomogeneties, whether from TOBS or siting changes or anything else, from the database. TOBS is adjusted for implicitly through this method, even thought there is no explicit TOBS adjustment procedure.
For no other reason than the joy of stringing two unrelated words together, reCaptcha: Blackburn’s gespati
Ian:
Muller et al detect discontinuities in the data (such as those caused by a change in TOBS) and splits the data at that point, treating the two halves as being separate stations. Since trends are computed for the split stations separately, they don’t need to explicitly account for changes in TOBS or thermometers, etc.
The BEST team ran some comparisons of station data adjusted via the traditional homogenization techniques (which account for changes in TOBS, etc) and their algorithmic approach, and when they’ve done so they’ve found the two almost perfectly match.
The fact that two such separate techniques yield almost indistinguishable results is a strong argument that both approaches are robust.
There is absolutely no excuse for Watts refusing to account for changes in TOBS. His argument that observers ignore directions to change observation times then LIE ON THEIR DATA SHEETS is crap.
Ian:
By the way, the implication of your second clause is extremely insulting. Maybe you should’ve stopped at “… I’ve missed it” because, judging by your posts here, you’ve missed a lot.
@Ian #57 – the short answer is that the BEST team did make adjustments that effectively allowed for changes to time of observation.
The slightly longer answer is that the BEST team used a different approach to their analysis of raw data, which is described in their ‘methods’ paper. AFAIK they treated discontinuities as new records.
http://berkeleyearth.org/pdf/methods-paper.pdf
has anyone taken a look at Arctic sea ice lately. Apparently a record low cyclone (not a polar low, but a synoptic scale system with lowest pressure of 963-hPa) has severely churned up the Beaufort Sea and is currently at about 75-80°N latitude. Unbelievable preliminary ice extent losses of as much as 285,000 km2 area over a single day. It’s hard to judge what the *quantitative* result of all this wind, advected warmth, and exposure to warmer, salty water will be yet, but the preliminaries are unprecedented in the satellite record.
See http://neven1.typepad.com/blog/2012/08/arctic-storm-part-2-the-color-purple.html for a great summary of day 2 of this summer storm. Incidentally, for those who are Arctic sea ice aficionados, it appears that after this storm winds down, we may go to the warm Arctic dipole pressure pattern that results in significant sea ice loss (U.S. global forecast model and ECMWF are both indicating this).
Ian@#57, AUIU Muller et al. uses a ‘scalpel’ approach that splits any site where TOB data varies into separate sites. Watts didn’t do that, or anything else, to account for TOB.
“Converted skeptic” is a rather odd title. I do science. I’m a skeptic. That what scientists do (including Gavin, other contributors to this site and many readers too). If Muller was genuinely skeptical of published research in climate science, all he had to do was read the published research and look for flaws. His exercise in checking the temperature record was obviously a waste of time, as no one has picked up a serious methodological flaw in existing practice.
The fact that he painted himself as one of the Watts camp, took money from a tainted source and generally had a belligerent attitude to climate scientists doesn’t define him as a skeptic. Possibly something else.
If he had the doubts he claimed he had, and went about checking the results, using a new methodology, or whatever, then published in the academic literature, maybe he wouldn’t have attracted as much media. Possibly there is some value in someone shouting from the rooftops something like “It’s all a fraud,” then recanting after checking. But I doubt very much that the hardcore deniers will take anything from this but confirmation that there’s some sort of weird conspiracy going on. You can’t make a head case sane by agreeing with them then changing your mind. If it was that easy, a lot of therapists would be out of work.
> grateful … Rohde … globalwarmingart
Grateful indeed, and for a long time. He kept that site up all through getting his PhD, when time must have been scarce.
It’s down for everyone right now:
http://www.downforeveryoneorjustme.com/globalwarmingart.com
And speaking as always from the cheerleading section of the bleachers:
remember to encourage the other folks out there making gifts of good science like globalwarmingart, and woodfortrees, and Tamino’s page, and so many others; some have contribution links or info, all have contact info.
Those of us who neither do, nor teach, can applaud and say thank you.
(Speaking up before I go mostly offline for a few weeks, to revegetate)
Philip,
I am a little skeptical about Muller’s conversion also.
Oh come on Dhogaza that’s a bit thin skinned but as I have no wish to insult or offend please would you accept my unreserved apologies. As I am a biochemist/molecular biologist there is a lot I’ve missed, as you so cogently observe, in climate science and am trying, with help from the replies here to catch up. I would note however that Philip Machanick seems a lot more critical than I of Riichard Muller.
Hi guys,
Can anyone help me understand the difference between the bottom row of Figure 4 in the new Hansen et al. PNAS paper and Figure 9 of the same paper. I think the left-hand panels of each figure are identical, but I don’t quite understand why the right-hand panel of Figure 4 is as different from the right-hand panel in Figure 9. How does a shift in the mean impact the shape of the distribution? I understand the behavior in Figure 4, but don’t understand Figure 9 (a behavior which Hansen et al. think is important).
Thanks for setting me straight.
-Chip
> It’s down for everyone right now
Thank you, Hank.
Ian:
Then you have no right to make veiled accusations of hypocricy.
Apology accepted, hopefully no more apologies will be necessary.
A lot of people climate scientists and interested and somewhat knowlegeable layfolk (like me) are highly critical of Muller. Your comment didn’t target Muller, though, you accused *us* of possibly letting him off the hook because he’s a “convert”.
That attitude isn’t going to win you any friends around here.
Ian @70
You say “As I am a biochemist/molecular biologist there is a lot I’ve missed…” Is this what makes you thick-skinned and needing to catch up? And I thought you guys used fume cupboards!
Philip Machanick questions Muller’s ‘conversion’ from denialist/skeptic to born-again climatologist. You, however, question whether this ‘conversion’ sort of allows Muller into some ‘club that can do no wrong.’ In my view, that is quite a different thing to be “a lot more critical” about. Is not one questioning Muller and his motives while the other is questioning climatology in its entirely with its motives?
Me? Well I just question you!
Chip, one is relative to a constant mean (1951-1980) the other is relative to the indicated mean. The captions are your friend.
I lok forward to Boris Johnson’s debut in the 75 meter turbine blade toss at the forthcoming Special Wind Olympics
#66.
The berkeley method does not “slice” records based on metadata. It slices based on the characteristics of the time series.
That is why, for example, zeke’s finding, that the berkeley method using no metadata matches the method using metadata is noteworthy. Some ( skeptics) have suggested the TOBS metadata is corrupt. A comparision of the two methods, argues against such a conjecture
@ 77: good point. Has the correspondence of the splicing with metadata been checked?
I heard Muller tooting his horn on Democracy Now a few days ago (a good show, IMHO, which followed him with Bill McKibben; even the Mullers of the world are allowed their say without bullying or interruption). What was hair-raising for me was realizing that (a) he still does not not admit that anybody was scientifically justified in accepting the reality of climate change until this very moment, with the completion of his own magnificent study, i.e., does not admit that the “problems” with data analysis were ever adequately addressed before this, and (b) he is as staunchly denialist as ever about the _effects_ of climate change. E.g., dismisses the present US drought and heat wave on the ground (spurious even if true) that at this very moment, below-average temperatures elsewhere in the world happen to level out the instantaneous global average. Also scoffs at the idea of polar-bear endangerment (ha ha, but see, e.g., http://www.esajournals.org/doi/abs/10.1890/09-1641.1 ). Et cetera.
In short, even to layperson like myself, his ongoing intellectual buffoonery is painfully apparent.
I hope I’m definitively clear of letting Muller off any hooks just because he’s a “convert” . . .
Larry
http://www.larrygilman.net
[Response:Larry, you might want to see what had to say about the 8 Muller “Democracy Now” interview fibs here at my Facebook page: here. –mike]
Re: Zeke Hausfather @ 50,
Thanks very much. This should keep me busy for a while!
Larry Gilman,
It may perhaps be that Muller’s inflated ego makes him ideal for playing “useful idiot”…at least in the unlikely event that there are any true skeptics among the denialists.
#75,
Thanks, Rattus. I understand that the means are different, but don’t understand how adding a constant to the anomalies (i.e. means calculated from different periods) results in a different shaped distribution.
-Chip
[Response: With respect to a different mean, the variations are different. – gavin]
I left a detailed comment, moderated out ,regarding the difference of scientific opinion on the paper by Dr Hansen reported upon in Wednesday’s NYT. In a nutshell I asked which of those if those agreeing with or dissenting from the conclusions of that paper were correct. I have no idea why this comment seeking information from experts in the field was not published. Are the critics correct or are they not?
[Response: Not. – gavin]
Great Observation!
Chip @ 71 and 82 – No doubt I’d have a better idea what you’re asking if I read the paper, but just in case, have you noted that Fig. 4 is Global, Fig. 9 is NH Land?
#79, Larry Gillman said, ” What was hair-raising for me was realizing that (a) he still does not not admit that anybody was scientifically justified in accepting the reality of climate change until this very moment, with the completion of his own magnificent study, i.e., does not admit that the “problems” with data analysis were ever adequately addressed before this,”
That was the same vibe I got from listening to Muller on an NPR Science Friday (via the podcast) a few days ago.
RE the diurnal range increasing, it seems only to be a slight upswing and not to the level of 1960s and before. See p. 13 of http://berkeleyearth.org/pdf/results-paper-july-8.pdf
Also, it is only one study. What do the other sources of climate data show re this?
Gavin,
re: the endless senate hearings, you may enjoy the first part of this article
http://westernaustralianwarp.wordpress.com/the-warped-science-of-political-success/
George
Gavin
re: the endless senate hearings, you may anjoy this
http://westernaustralianwarp.wordpress.com/the-warped-science-of-political-success/
George
Re 82 Chip Knappenberger, (re 75 Rattus Norvegicus)
about figs 4 and 9 of Hansen et al “Perception of climate change”
captions:
“Fig. 4. Frequency of occurrence (y axis) of local temperature anomalies (relative to 1951–1980 mean) divided by local standard deviation (x axis) obtained by counting gridboxes with anomalies in each 0.05 interval. Area under each curve is unity.”
“Fig. 9. Frequency of occurrence (y axis) of local temperature anomalies divided by local standard deviation (x axis) obtained by counting gridboxes with anomalies in each 0.05 standard deviation interval. Area under each curve is unity. Standard deviations are for the indicated base periods.”
See also discussion p.4 and p.7 (on the pdf file) – if I’ve understood correctly, indicates that all anomalies in fig. 4 are relative to the same base period, although the different frames use different standard deviation values – the middle frame uses detrended data, so the larger standard deviation there is not from the ongoing change in climate but a characteristic of the climatic state. The larger standard deviations compress the graphs horizontally and thus increase the height – but should otherwise preserve the shape, so far as I know. I could imagine some shape changes would occur because of the binning process (into 0.05 units of standard deviation intervals) – so when a different standard deviation is chosen, some data points that were in the same interval are now split up and vice versa (although the interval is small enough that I woulnd’t think this would have a large effect on the appearence of the graphs)..
Fig 9 uses anomalies relative to different base periods, so the means shift. The related section in the text also implies that it uses the different standard deviation values as well. You would think that the shape should be the same then, compared to the corresponding frame in fig. 4 (second row) (although the binning process could make some small alterations as before) – noting that the middle frames in each do not correspond (fig. 4 uses detrended sigma for 1981 – 2010, while fig 9 uses the entire period (1951 – 2010) and doesn’t specify (at least not in the parts I’ve read) whether the sigma is a detrended value or not. If it is not detrended, then all the graphs in that frame will individually be that much narrower with that much higher peaks then the normal distribution; I’d guess they could average to a normal distribution (?) if the trend had the right shape, but I’m thinking this case won’t quite fit (with the earlier decades being more similar, there should be an off-center peak with a longer tail on the other side, right?). It’s interesting to note that the middle decade of 1981 – 2010 approximates the normal distribution for the sigma of that period as shown in the last frame of fig. 9.
I think the off-center curves may sometimes appear to tilt away due to an optical illusion?
Re 84 Rick Brown – actually the second row of fig 4 is for NH land. The anomalies are relative to the same base period in fig 4 and different base periods in fig 9; the histogram shows anomalies in units of standard deviation, for standard deviation values from different periods – in each figure, though the middle frames of each row don’t correspond between the figures and I’m not quite sure about the right frames.
Actually I am noticing a few things that I couldn’t explain (compare decade 1951-1961 second two frames of 2nd row fig 4 to right frame of fig 9 – the lowness of the peak in fig 9 suggests that the sigma there is the detrended one, but that wouldn’t account for the higher peak of 2001-2011. But I haven’t read the whole paper either. The more general point about different base period sigmas tending to compress horizontally and stretch vertically, or vice versa, makes sense.
‘…remark on how little actually changed’ – i absolutely understand your chagrin at this guy turning up and claiming he’s made a major contribution to climate science….clearly he hasn’t. but what he has done is made a major contribution to the pseudo-debate in the media. the storyline of independent, sceptical, einstein looking physicist committing to the reality of agw is worth much more to the campaign for political committment (sad to say) than a hard-working climate scientist.
so, i think something has changed. not in the science but in the public debate (as demonstrated by the buzz in the denial-o-sphere about their betrayal).
Just piling on……what a crock of sh***t. Actually, I find myself in an odd agreement with Judith Curry when she says:
“If the attribution problem was as simple as Muller makes it out to be (curve fitting to CO2 concentration), then why are others wasting all their time with complex modeling studies, data analyses etc as described above?”
Septics have been ranting and raving about how complex attribution studies are prone to uncertainty, “unknown unknowns”, “butterfly effects” and other clichés for quite some time, and Muller himself has been disputing facts about measured temperatures and proxy reconstructions which, to my knowledge, are at least as solid as the attribution studies (and certainly less based on the favourite septic target of complex modelling). That Muller then suddenly would declare himself convinced by simple, non-peer reviewed curve fits (which most competent undergraduates in any scientific discipline could do) is, to my mind, the strangest part of Muller´s “conversion”. He certainly leaves himself open to completely justified criticism from both standard septics and climate scientists alike and appears to neither convince nor impress anybody in the process. Is the whole purpose of this exercise really just simple self-promotion and then to heck with his scientific reputation?
But what the hey, perhaps the reality based community should just laconically accept Mooneys conclusion:
“while in a scientific sense Muller’s conversion is quite insignificant […] in a political sense, his recent arrival is all that matters. So just declare victory, my scientific friends.”
I thought there was something interesting.
What about the increasing diurnal range since the 1990
On the last rollout of BEST, wood for trees quickly had the data and making BEST graphs went viral. This time, nothing. WFT does not have it, yet. Is there an obvious reason?
[Response:Aren’t the data the same data as before, just analyzed in a new way?–eric]
Gavin (#82),
Thanks. But so that I better understand… in Figure 9, the sample standard deviation is not calculated with respect to the sample mean, but wrt the mean from a different period?
[Response: The figures shows a distribution of anomalies with respect to a baseline, not a standard deviation. There are three baselines used (title of each plot), and the anomaly distribution in each figure for each decade uses the sd from the baseline to calculate the z-score. – gavin]
But it in Figure 4 the sample standard deviations are calculated in the normal way (i.e. from the sample mean)? Figure 4 seems to make sense to me, but I am still struggling with the significance of Figure 9–which Hansen uses to argue against a shifting baseline. Perhaps an interesting argument, but not one that I am completely following as of yet.
Thanks again,
-Chip
[Response: Figure 4 shows the impact of different sd, not different baselines. – gavin]
Yes I suppose, but they started ~47 years earlier, and ended about a year later.
Thanks for the succinct response Gavin it is appreciated
JCH,
The data is here for anyone who wants it: http://berkeleyearth.lbl.gov/auto/Regional/TAVG/Text/global-land-TAVG-Trend.txt
There is similar data available for any continent, country, U.S. state, or specific lat/lon location on earth: http://berkeleyearth.lbl.gov/city-list/
Eric,
The main differences were a few bug fixes, extending the data back to 1750, and the regional/local results.
#87
Yes, the change in the diurnal range is interesting. Gavin always seemed to say that the reason more warming was occurring in the northern hemisphere is because there is more land. I guess he was right. I’m surprised the authors didn’t include the change in diurnal range for each hemisphere. That would have been…better. In which case you would hope to see decreasing diurnal range for the northern hemisphere and no upswing? Or not.