Monday
May282012
by 
Bishop Hill
Bishop Hill Tying the threads together
May 28, 2012 Climate: other In a post that rather ties together several of the recent threads at this site, Anthony Watts looks at the recent interest in Myles Allen's talk, the importance of the Hockey Stick and discusses a recent paper that looks at the tendency to see expertise where there is none.
[The Hockey Stick] was a good gamble at the time, but as Climategate has shown us, it may have been a winning hand with a one time jackpot, but they are losing the card game as the other players slowly realize they have a cheat in their midst.
Reader Comments (13)
Myles
My conclusion from a lot of study of the Climategate e-mails is very simple.
I don't trust anybody who claims to be a climatologist. I view the phrase 'Trust Me I'm A Climatologist' is a contradiction in terms. If you told me the time, I'd need three totally independent verifications before believing a word you said.
And I'd be very happy to get those verifications from a second-hand car dealer, an estate agent and a journalist since at least members of those occupations have some credibility
I hope that you understand. Climategate was not primarily about science. It was about honesty, integrity and credibility. IMO you guys - and your 'field of study' - have none.
Loved Anthony's reply to Richard Courtneys "illuminated admonishment" in the comments.
What a good choice Anthony made in featuring Matt Ridley's post on the Hockey Stick. Matt Ridley (Oxford alumnus) certainly has no difficulty in communicating his climate message with crystal clarity.
The Hockey Stick] was a good gamble at the time, but as Climategate has shown us, it may have been a winning hand with a one time jackpot, but they are losing the card game as the other players slowly realize they have a cheat in their midst.
======================
That says it all.
Myles's ears are doubtless burning. He can regain a shred of credibility by stating his falsifiability criteria for AGW.
Come on, Myles, you can do it. Under what circumstances would you concede that AGW is refuted? Here's a multiple choice question for you:
Myles Allen will declare AGW to be false in the event that:
(i) Snowball Earth returns.
(ii) The southern icecap once extends to South Georgia
(iii) The Greenland Strait is once iced over
(iv) The Thames freezes at Tower Bridge
(v) The monthly UAH-MSU LT anomaly stays below 0.8C from 2013 to 2023
(vii) The UAH-MSU LT 13-month running average stays below 0.5C from 2013 to 2023.
Or, Myles, you may choose the default option:
(viii) Under no circumstances whatsoever.
Choosing option (viii) will go down badly here, but will gain you great kudos among you co-religionists.
After moving the pea around to discuss temperature records instead of the various problems that climategate revealed, Myles then misrepresented the MET office error as 'a problem with input files'. Actually the problem was with averages, which are not input, they are computed. But as usual in climatology, Myles distorted the facts to suit the PR message.
As a previous poster pointed out, simply nothing from climate science can now be believed. Even if there are individuals you trust, they are entangled in the groupthink of a subject that is corrupt
Which is a tragedy, as the earth's climate is a thing deeply deserving of study. We are far worse off now than if, for some strange reason, the study of the climate had been forbidden for the past fifty years.
Which you could almost say it has been.
The “Watts Up With That” article suggests it’s all about cheating at cards, but Myles seems to me to be playing a rather more childish game, a version of “O’Grady Says”:
“Scientists say: ‘Do this’.”
“Now, do this. Ha ha, caught you out there. I didn’t say ‘Scientists say’.”
The problem with the "Harry" file is that I don't see how the problems it revealed could be fixed, and certainly haven't seen anything from them to show that they have been.
How could they identify problems with the data when they didn't even know what it was? They appeared to have a chaotic load of files full of numbers which might or might not have been what they were guessing it was, whether it was adjusted or not, whether it was real or the output of a model.
I have seen it suggested that this doesn't matter because "Harry" referred to an old series, but I don't see how the organisation and methods revealed could ever produce anything reliable.
I think that some of the enquiry conclusions boiled down to saying that what was going on met expected standards for academic research, which wouldn't be a surprise. I have seen the results of research commissioned by a manufacturer on a process from a university regarded as a leader in the field and saw the familiar story; poor experimental practice leading to a limited number of results of dubious accuracy which were nevertheless processed, interpolated and generally tortured to reach the desired conclusion.
I have also been through the process which teaches you to do this, and seen fellow students get good marks for submitting textbook results of an experiment which due to deficiencies in the apparatus could never produce those results. Trying to get the thing fixed so that it could be done properly doesn't make you popular!
To show how appallingly bad climate science has been managed, have a look at the latest report of Graeff's experiments on the true origin of warming in the atmosphere: http://tallbloke.wordpress.com/2012/05/29/lucy-skywalker-graeffs-second-law-seminar/#more-6549
The adiabatic atmospheric lapse rate appears to be about a tenth of the gravitationally-induced heating in a column with no convection. The implications of this are profound, particularly in mines where one has to ask whether the temperature rise is from the heating of the crust by the interior or whether the crust is heated by the air: http://en.wikipedia.org/wiki/TauTona_Mine
'The mine is so deep that temperatures in the mine can rise to life threatening levels. Air conditioning equipment is used to cool the mine from 55 °C (131 °F) down to a more tolerable 28 °C (82 °F). The rock face temperature currently reaches 60 °C (140 °F).'
The average subsurface geothermal gradient is about 25 to 30 degC/km [15 degF/1000 ft], but higher in volcanically active and oceanic crustal regions, lower in stable continental plates. Every log run in every deep well takes a bottom hole temperature reading. There are countless thousands of well records. The heat source is geothermal.
Pharos, I agree that the source heat is geothermal, but that does not mean that there is not a gravitational influence on the second law, as Graeff's experiments appear to suggest.
Let me try to sum up what I’ve got out of the past few days. I suggested in the Communicate2011 talk that the 0.02K revision to the HadCRUT temperature record was the only change to any published dataset used in the detection and attribution of human influence on climate to have resulted from the UEA e-mail affair, and that this was not generally appreciated by the public. I was using this as an example of how things have gone wrong in communicating climate science: this was not a talk about “climategate” per se.
On whether the “only change” statement was strictly correct, Steve McIntyre has pointed out that the e-mails raised new questions with the treatment of paleoclimate records, Judith Curry has observed that these records are needed to check our estimates of internal climate variability and Ross McKitrick has argued that some of the e-mails showed an improper dismissal of his paper on the correspondence between patterns of warming in the instrumental record and patterns of economic development. Many other points have been raised, but I would like to address these three.
I accept Steve’s point that paleoclimate reconstructions continue to evolve and fresh sources of uncertainty continue to emerge, although my impression is that they were evolving anyway before release of the e-mails and would have continued to do so regardless. This continued uncertainty is a key factor making it difficult for scientists like myself, outside the dendroclimatology community, to make use of tree-ring based data. In trying to cope with multiple blog threads simultaneously, I probably went too far in disparaging tree-ring data, for which I apologize to any dendroclimatologists who might be reading these threads. I do believe efforts to reconstruct pre-instrumental climate represent an interesting and worthwhile challenge: my point was simply that many people seem to think it is the main point of climate research, which it is not.
In response to Judith Curry’s point about the need for proxy reconstructions to test model-simulated internal variability, again, this is a question of “it would be nice if only we were able to do so.” In my personal view, the uncertainties and potential biases in the spectrum of variability that must arise from the process of stitching together multiple tree-ring records (many of which have to be individually detrended), and the fact that we know GCM-simulated variability is deficient on the small scales that the trees are responding to, make it difficult to use proxy records to falsify GCM-simulated large-scale variability. If a GCM disagrees with a paleo-record, do we reject the GCM’s internal variability, the forcing data used to drive the GCM, or the paleo-record itself? We do have observations of variability on sub-century timescales through the instrumental record and new products like the 20th century reanalyses: I think, in the short term at least, these potentially provide more information on internal variability than the millennial reconstructions.
Since the key question for attribution is the origins of the surface warming over the past 30 years, that being the only period for which we have direct observations of forcing, it is the spectrum of internal variability on 20-100 year timescales that is essential. Variability on longer timescales is less important for attribution of causes for the current warming trend. I stress this statement applies to surface temperature. Sea level responds on different timescales, making attribution correspondingly harder.
In response to Ross McKitrick’s point (apologies for being slow on this one), I suspect what Phil Jones was referring to in the “no need to calculate a p-value” remark (although you should really ask him) was the danger of over-interpreting chance covariation. The only p-values that mean anything are those that derive from physically-based hypotheses. It is all too easy to find a high p-value from a chance correlation (sunspots and number of Republicans in the US Senate is the classic example). I wasn’t involved at all in that IPCC chapter, but I would be inclined to agree with their assessment that what you were seeing in that paper was an example of such an acausal covariation, for which the p-value of a pattern-correlation is indeed meaningless.
Then there is the much more general point, raised by Lucia, Rhoda and many others, that my talk was misleading, because “climategate” was not about the data at all, but rather about scientific process and the probity of climate scientists. As Mike Hulme observed, “climategate” meant different things to different people: for me, the implications for the instrumental record were all-important, which is why I was castigating the British press for paying far less attention to the fact that the instrumental record got an almost (in deference to Ross) completely clean bill of health than it paid to the initial allegations. There was an interesting side-thread on why the HadCRUT got dragged into this in the first place, to which I don’t have much to add apart from reassuring everyone that I don’t blame the bloggers for this confusion.
Many people have asserted that the main impact of “climategate” is that we can no longer say “trust me, I’m a climate scientist” until we all come out and condemn CRU, Muir-Russell, Oxburgh, etc. “Trust me, I’m a climate scientist” is not a phrase I have ever used, and I hope I never will. I teach a 12-lecture course to our 3rd year physics students (open to the public if anyone is interested) that starts from the premise “Don’t trust climate scientists” – the point being that, as physicists, they should be able to understand the problem for themselves, and not be expected to take the IPCC’s word for it.
The only basis of trust in science is the reproducibility of results. This is why availability of data and model source code is so important: I have always supported open data, although I have also consistently said that I don’t think Freedom of Information requests are the right way to enforce it. Journal editors can and should enforce a simple “disclose or retract” policy if a result is challenged, and almost all of them do: if any don’t, then the solution is to name and shame them, not set up a parallel enforcement system. I also think it is always better to reproduce results from equations (and, where possible, independent models and observations) rather than “auditing” computer code.
Finally, on the “bad for democracy” remark that upset a lot of people. I don’t want to suppress discussion of the Medieval Warm Period, but everything has an opportunity cost. Time spent arguing over paleoclimate research is time not spent on, for example, the merits of the two degree “goal” agreed in Copenhagen and Cancun, with remarkably little scientific justification. Yet whether we aim to limit anthropogenic warming to two, three or four degrees has far bigger implications for climate policy than the existence or otherwise of the Medieval Warm Period. Why is this not a hot topic in the blogosphere?
Ironically, this whole discussion started from a throw-away post by Paul Matthews in a discussion of a lecture I recently gave on whether it would be possible to frame an effective climate mitigation policy that did not extend the reach of the State in the way that cap-and-trade, carbon rationing or geo-engineering clearly will. Paul has apologized, which is much appreciated, but the damage may be done, Paul. If the European Commission decide to impose carbon rationing in 2020 after another record-breaking warm decade, because we spent this past week discussing Myles Allen’s interpretation of climategate (not to mention his, admittedly poor, choice of shirts) rather than coming up with a less intrusive policy alternative, your grandchildren shall know the reason why.
Apologies for cross-posting on various threads.