“First of all, the data in question weren’t the CRU’s data in the first place — they came from various national met offices, and came with restrictions on redistribution.”

Well first, CRU had already redistributed earlier versions of it several times to other researchers, second, when challenged to produce these restrictions (they had said they were prevented from redistributing to non-academics by these conditions), CRU were unable to find any, and third, most of the met services were not seeking to protect confidentiality (you could in most cases download the same data for free), they were attempting to protect the data’s integrity. By insisting that all copies came through them, they ensured that no modified or corrupted versions of their data got into circulation.

Why? Well, we get a hint by reading Phil Jones’ begging email asking for the release of the data. He points out that the CRU data has been modified, and is no longer identical to the original. Unfortunately, Phil can no longer remember what modifications were made or why. The met services didn’t object to Phil sharing the data, they objected to him publishing his modified data and calling it theirs.

“But that’s water under the bridge”

I’m sure you’d like to think so.

“But it has been close to a month since the CRU released all of the data, and where are the skeptics’ results? The people who have been demanding access to the entire CRU data-set sure don’t seem to be all that eager to do anything with it now that it has been released.”

The entire CRU data set hasn’t been released. Of the 20 Polish stations known to have been in earlier versions, 8 have been released; also apparently missing are Magdeburg, Brocken, Skopje, Fairbanks, Corpus Christi, New York, Ridgetown, Belleville, Nitchequon, Morden, Waseca, Chesterfield, Tacubaya, and Titizema/Chichijima. Surely you noticed? The sceptics certainly did.

“That being said, Phil Jones was working on securing permission to release all of his data long before “climategate” blew up — you can confirm that yourself with a little research.”

He started asking permission shortly after Climategate, when he realised he was going to lose the FOIA fight. And the nasty emails started long before even the FOIA requests.

#36,

High sea surface temps cause thunderstorms – the higher the temperature, the earlier, heavier, and denser the storms.

During the Eemian, ocean temperatures were probably around 2-3 C higher than today. (Lea 2000, Pelejero 2003, Martrat 2004.) Significant, but not catastrophic.

“Other research groups (and individuals as well) have not had any trouble replicating/confirming the CRU’s work, be it instrumental or paleoclimate work.”

Some have got similar results, by using equally dubious methods (e.g. reliance on bristlecones). Plenty of others have got entirely different results, and had just as much difficulty replicating the earlier studies. It’s quite easy to get no Hockeystick. (Zorita said they had tried to replicate MBH98 before McIntyre and had found a can of worms. Loehle, Ljungqvist, McIntyre, and even Mann himself if you count the contents of the BACKTO_1400-CENSORED directory have all done so.)

“Furthermore, models are *not* the primary evidence for climate sensitivity re: CO2 — the most important supporting evidence for that comes from various lines of paleoclimate data.”

It comes from attempts to get the models to emulate paleo data – based on all the assumptions the models make about what affects climate. Paleo data alone can provide no more than a correlation, it cannot prove causation.

“We could throw out every single modeling result and still be left with a mountain of evidence linking CO2 to major climate change. There is no doubt that CO2 has been the primary climate driver for the Earth over most of its history.”

Excellent! Why don’t you publish it? Why didn’t the IPCC publish it? Then we could stop all these arguments about what the actual evidence for it really is.

“Anyone familiar with Mann’s technique knows that the output of his processing will tell you whether you have a good temperature signal or just random noise.”

No it won’t. Anyone familiar with basic statistics will know that the best you can hope for is to distinguish a signal due to something from random noise following a particular assumed model. It won’t tell you it is a temperature signal. And in fact, we know it isn’t temperature – it’s Bristlecone pines and their 20th century growth spurt.

The random model McIntyre used was selected to be sufficient to show the flaw in the algorithm, it was not meant to be fully realistic. It is not being claimed that MBH98 actually was the result of random red noise processed by his algorithm, it is being claimed that the flaw in the algorithm over-emphasises any hockeystick shapes in the data, to such an extent that even random data with no such signal shows a hockeystick shape.

Yes, a lot of sceptics have done the same. As you’ll have noticed, the results are not identical, with your version giving higher values in earlier years. The main issues, however, are with matters such as corrections for UHI, site changes, instrument changes, transcription errors, missing values, etc. What changes were made in between the raw data, and the version CRU just released? It’s validation of the adjustments and corrections that is desired.

In your version, what values did you use to correct for UHI, for example? What correction did you use for the transition from cotton region shelters to MMTS?

Oh, and the CRU data just released was as a result of an FOIA request that CRU rejected, rejected again at appeal, fought at the ICO, and lost. They released the data because they were threatened with a contempt of court action if they didn’t.

And again, it’s not a question of whether the conclusion turned out to be right or not, it’s a matter of the proper scientific process being to do it out in the open so that people can check that it’s been done right. Show your working. Secret science is no good.

It depends which one. Some can be replicated, some can be replicated now as a result of sceptics making a fuss, others cannot as yet.

“If not, what additional information would I need that NASA/CRU/etc. don’t already supply?”

I’m not the best person to talk to on that. I could give examples, but the I’m not up to date on which ones people are most interested in.

But if you want to have a go, there’s an outstanding request for the 2006 Yamal regional chronology (a digital version of this series together with a list of all the measurement data sets used to make this composite, denoting each data set by ITRDB identification or equivalent), referred to in email 1146252894.txt as URALS and supporting Briffa et al. 2008. Good luck!

I should say before I start that I’m pleased to see your comments here of a higher calibre than is usual. I look forward to a good debate.

“Of course, you are well aware that results generated from temperature stations confirmed to have none of the above problems are virtually identical to results you get when you include *all* of the temperature stations.”

Absolutely not true.

I presume you are talking about Menne et al., and the mean trend calculated over a particular period for the continental United States – where the errors approximately cancel. However, if you look at absolute temperatures, the trends in maxima or minima, or diurnal range, pick a different time interval, or pick regional subsets, then the errors do not cancel and you get different numbers. If you ask the question of the whole world, then nobody knows because – apart from a few individual cases – nobody has done such an audit outside the United States.

I’m afraid this is a case of the classic “the errors don’t matter” trick. The fact is, temperatures from a thermometer sited within a few feet of a barbecue were used for years without anybody noticing – even to a non-scientist that must look bad, irrespective of whether it “mattered”.

The question is not “did it matter?”, the question is: “did they do the science competently, so that we can trust the results?” And in this case, the question I’d like to draw your attention to is “can I induce you to defend rather than condemn practices like not checking to see if the thermometer you’re relying on is sat right next to an artificial heat source?”

You were happy to accuse Roy Spencer of not being a real scientist for not quadruple-checking his results, but you’ll cheerfully make excuses for a whole catalogue of unchecked errors when made by others. Why?

And can you not see that’s exactly the sort of behaviour that has so damaged the reputation of climate science?

We are on the same page as far as the variation in global temperature in paleoclimatology. The temps have been quite a bit higher and quite a bit lower than our standard “normal.” In the hotter periods, there has been more CO2. Many times the CO2 level was considerably higher than present levels. Whether it was a leading cause or trailing result makes little difference except for semantic wrangling. Each time it has gone up previously, natural mechanisms sequestered the increase of CO2. Why do we necessarily leap to the conclusion that this time, the natural mechanisms which have handled higher levels repeatedly in the past will be unable to at this time? Whether it is anthropogenic, brought down in bubbles by little green men from mars or whatever source, CO2 is the same regardless of it’s origin. We are still well within the the historic “normal” range of CO2 concentrations. We are still with a fraction of a degree of “normal” global temperature averages. I see cause for concern and study. I don’t however see any justifications of exaggerated claims of our imminent danger (No not from you, I have not seen you make any such claim).

The PETM is a bit of a stretch when that was a 6C change and we are looking in the range of 0.3C in most of the studies I have seen. As far as the sea surface temps near 100F we have places pretty close to that on occasion now. Down here, we have a wet bulb temperature of 84.9 already (103F with RH of 51%). It is not even an unusual day.

The PETM is perhaps the best analog to what may happen in the next couple of centuries if we don’t get CO2 emissions under control — and a PETM-like event would pose severe challenges to human civilization.

Go back to the Eocene hothouse period; temperatures/humidity then certainly got high enough to make many lower latitude regions of the Earth uninhabitable by modern humans (i.e. with periods heat/humidity high enough to kill everyone), regions where billions of people now live. Sea surface temps got as high as 100F (possibly even higher) in the tropics. Sea surface temps like that would produce levels of heat/humidity in the atmosphere that would induce heat stroke in anyone exposed to those conditions for more than a few hours. Start getting wet-bulb temperatures in the mid to high 90′s (F) and even healthy, fit people will start dropping like flies.

First of all, the data in question weren’t the CRU’s data in the first place — they came from various national met offices, and came with restrictions on redistribution.

But that’s water under the bridge; All of the disputed data have now been released, and guess what? The new data produce the same results that NASA/CRU/etc have been publishing all along.

It took me just a few hours to modify code that I had written and produce my own initial temperature results based on the CRU data, results that I generated and uploaded to the imageshack image-hosting site the very same day I downloaded the data.

But it has been close to a month since the CRU released all of the data, and where are the skeptics’ results? The people who have been demanding access to the entire CRU data-set sure don’t seem to be all that eager to do anything with it now that it has been released.

BTW, it took the CRU a couple of years to secure the permission to redistribute the proprietary raw data that came with nondisclosure restrictions. Many national met offices view their temperature data as “cash cows” and thus place restrictions on data redistribution by third parties. Don’t blame scientists for that; blame conservatives/libertarians who insist that governments be run like profitable businesses. Businesses generally don’t make money by giving their goods away, so running government like a business means restrictions on the distribution of government-produced data.

That being said, Phil Jones was working on securing permission to release all of his data long before “climategate” blew up — you can confirm that yourself with a little research. He was working hard to secure permissions from various and sundry government bureaucracies, and understandibly got frustrated with foreign nationals who were demanding that he release the data immediately (before he had permission to do so)! Hence the nastily-worded emails.

“All you are doing here is repeating talking-points put out by others who haven’t any serious analytical work themselves.”

No, actually I am repeating the findings of the investigation. Let me be quite clear about my stance so you don’t have to keep making guesses as to what I am saying.

I do not know for certain if the analysis of the CRU is correct or incorrect. I have never made any statements either way. What I do have an issue with (supported by the findings of the investigations and my own observations of the news as it has played out) is that the CRU was trying every way they could to keep their data away from any detractors and to thwart anyone from raising any questions. To me this raises flags which makes their results suspect to me as it doesn’t conform to the norms expected in every other field of scientific inquiry.

Actually, that is a nice explanation of the steps you took with the data. If only everyone were so open in their methodologies.

Not being argumentative, just curious. You have the instrument data for this time period. What steps would you choose to to make the extrapolations that are the real point in question?With this time period of data, how far out would you comfortably make predictions of the extensions of any trends you see? (Really, I am not trying to trip you up. I am sure that there are some statistical rules of thumb to get things specifics. I am just interested in the discussion)