RELATED ARTICLES, highly suggested:
Clarification on BEST submitted to the House
Pielke Sr. on the Muller testimony
Independent company station siting analysis demonstrates the problem
Quote of the Week – other scientists weigh in
Guest Post by Willis Eschenbach
Well, I had hoped for the best from BEST, the new Berkeley Earth Surface Temperature project looking at the global temperature record. I was disheartened, however, by the Congressional testimony of Dr. Richard Muller of BEST.
Photo Source He said (emphasis mine):
Global Warming
Prior groups at NOAA, NASA, and in the UK (HadCRU) estimate about a 1.2 degree C land temperature rise from the early 1900s to the present. This 1.2 degree rise is what we call global warming. Their work is excellent, and the Berkeley Earth project strives to build on it.
Human caused global warming is somewhat smaller. According to the most recent IPCC report (2007), the human component became apparent only after 1957, and it
amounts to “most” of the 0.7 degree rise since then. Let’s assume the human-caused warming is 0.6 degrees.
The magnitude of this temperature rise is a key scientific and public policy concern. A 0.2 degree uncertainty puts the human component between 0.4 and 0.8 degrees – a factor of two uncertainty. Policy depends on this number. It needs to be improved.
Why do I think his testimony doesn’t help in the slightest? Well, to start with, I’ve never heard anyone make the claim that the land surface air temperature (excluding oceans) of the earth has warmed 1.2°C since 1900.
He cites three land temperature datasets, NOAA , NASA (GISTEMP), and HadCRU (he presumably means CRUTEM, not HadCRU).
Here’s the problem. The actual land surface air temperature warming since 1900 according to the existing datasets is:
NASA GISTEMP: 0.72°C
NOAA NCDC: 0.86°C
CRUTEM: 0.92°C
So Dr. Muller, in his first and most public appearance on the subject, has made some of the more unusual claims about the existing temperature datasets I’ve heard to date.
1. Since the largest temperature rise in the three datasets is 30% greater than the smallest rise, their work is not “excellent” in any sense of the word. Nor should the BEST team “strive to build on it.” Instead, they should strive to understand why the three vary so widely. What decisions make the difference? Which decisions make little difference?
2. Not one of the three datasets shows a temperature rise anywhere near the 1.2°C rise Muller is claiming since 1900. The largest one shows only about 3/4 of his claimed rise.
3. He claims a “0.2 degree uncertainty”. But the difference between the largest and smallest calculated warming from the three datasets is 0.2°C, so the uncertainty has to be a lot more than that …
4. He says that the land warming since 1957 is 0.7°C. The records beg to differ. Here’s the land warming since 1957:
NASA GISTEMP: 0.83°C
NOAA NCDC: 1.10°C
CRUTEM: 0.93°C
Note that none of them are anywhere near 0.7°C. Note also the huge difference in the trends in these “excellent” datasets, a difference of half a degree per century.
5. He fails to distinguish CRUTEM (the land-only temperature record produced by the Climategate folks) from HadCRU (a land-ocean record produced jointly by the Hadley folks and the Climategate folks). A minor point to be sure, but one indicating his unfamiliarity with the underlying datasets he is discussing.
It can’t be a Celsius versus Fahrenheit error, because it goes both ways. He claims a larger rise 1900-present than the datasets show, and a smaller rise 1958-present than the datasets.
I must confess, I’m mystified by all of this. With his testimony, Dr. Muller has totally destroyed any credibility he might have had with me. He might be able to rebuild it by explaining his strange numbers. But to give that kind of erroneous testimony, not in a random paper he might written quickly, but to Congress itself, marks him to me as a man driven by a very serious agenda, a man who doesn’t check his work and who pays insufficient attention to facts in testimony. I had hoped we wouldn’t have another temperature record hag-ridden by people with an axe to grind … foolish me.
Perhaps someone who knows Dr. Muller could ask him to explain his cheerleading before Congress. I call it cheerleading because it certainly wasn’t scientific testimony of any kind I’m familiar with. I hear Dr. Muller is a good guy, and very popular with the students, but still … color me very disappointed.
w.
PS – Muller also said:
Let me now address the problem of
Poor Temperature Station Quality
Many temperature stations in the U.S. are located near buildings, in parking lots, or close to heat sources. Anthony Watts and his team has shown that most of the current stations in the US Historical Climatology Network would be ranked “poor” by NOAA’s own standards, with error uncertainties up to 5 degrees C.
Did such poor station quality exaggerate the estimates of global warming? We’ve studied this issue, and our preliminary answer is no.
The Berkeley Earth analysis shows that over the past 50 years the poor stations in the U.S. network do not show greater warming than do the good stations.
Thus, although poor station quality might affect absolute temperature, it does not appear to affect trends, and for global warming estimates, the trend is what is important.
Dr. Muller, I’m going to call foul on this one. For you to announce your pre-publication results on this issue is way, way out of line. You get to have your claim entered into the Congressional Record and you don’t even have to produce a single citation or publish a paper or show a scrap of data or code? That is scientific back-stabbing via Congressional testimony, and on my planet it is absolutely unacceptable.
That is taking unfair advantage of your fifteen minutes of fame. Show your work and numbers like anyone else and we’ll evaluate them. Then you may be able to crow, or not, before Congress.
But to stand up before Congress as an expert witness and refer solely to your own unpublished, uncited, and un-verifiable claims? Sorry, but if you want to make that most public scientific claim, that bad siting doesn’t affect temperature trends, you have to show your work just like anyone else. If you want to make that claim before Congress, then PUBLISH YOUR DATA AND CODE like the rest of us mortals. Put your results where your mouth is, or if not, leave it out of your Congressional testimony. Why is that not obvious?
Anthony’s unpublished and unverifiable claims are as strong as your similar claims. That is to say, neither have any strength or validity at all at this point … so how would you feel if Anthony trotted out his unverifiable claims before Congress to show that Dr. Richard Muller was wrong, and didn’t show his work?
Like I said … color me very disappointed, both scientifically and personally. Dr. Muller, I invite you to explain your Congressional testimony, because I certainly don’t understand it. I am totally confident that Anthony will be happy to publish your reply.
I also urge you to either a) publish the data and code that you think shows no difference in trends between good and poor stations, or b) publicly retract your premature and unverifiable claims. You don’t get to do one without the other, that’s not scientific in any sense of the word.
PPS – Does any of this mean that the BEST analysis is wrong or their numbers or data are wrong or that the BEST folks are fudging the results? ABSOLUTELY NOT. I am disappointed in Dr. Muller’s claims and his actions. The math and the data analysis is an entirely different question. Theirs may be flawless, we simply don’t know yet (nor would I expect to, it’s early days). I look forward to their results and their data and code, this kind of initiative is long overdue.
I want to be very clear than the validity of their actual methods depends only on the validity of their actual methods. The problem is, we don’t even know exactly what those methods are yet. We have rough descriptions, but not even any pseudocode, much less code. Which in part is why I find Dr. Muller’s testimony unsettling …
RELATED: See the rebuttal letter to congress:
Clarification on BEST submitted to the House
Pielke Sr. on the Muller testimony
=========================================================
UPDATE: in apparent response to Willis Eschenbach, BEST has added this below to their FAQs page. For fairness, I reproduce it here. – Anthony
NEW (4/1) – It appears that in Dr. Muller’s testimony he shows a temperature rise greater rise than others had previously published. Is this so? Can you explain?
The Berkeley Earth plot is for the land data only, since we have not yet begun analysis of ocean temperatures. Because we only analyze land, that was the fair comparison to make. The ocean temperature rise is less, and when included in, it reduces the value of the total temperature rise.
We started with the land data for several reasons:
- It is the data that is affected most by the most contentious issues: data selection bias, urban heat island, and station integrity issues. These are big concerns and we wanted to address them.
- The temperature rise on land is greater than on the oceans, mostly because the ocean distribute the heat over the mixed layer and thereby reduces the temperature rise. Land keeps the heat mostly on the surface. So the land temperature is actually more sensitive to greenhouse gases than is the world temperature.
- The land issue, with 1.6 billion measurements, was a huge one to tackle. It made sense to divide the effort into two stages.
The land only data for the other three groups are available on their websites, and agree with the plot provided in Dr. Richard Muller’s testimony.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
Has anyone here mentioned the NY Times article quoting Anthony on this whole BEST analysis:
http://www.nytimes.com/cwire/2011/04/01/01climatewire-experts-heat-up-over-berkeley-lab-scientists-q-490.html?amp=&pagewanted=all
Sorry if this is a repeat from somewhere else…
Re the Scalpel and Low-Cut Filter.
The concept of the scalpel has a siren song to it. If you cut out all the bad stuff, then all you have left is good stuff. Splice them together and you have a long good thing. A whole bunch of short positive slopes will result in one long positive slopes. What could go wrong?
I have been recalling my college and graduate work that involved tensor analysis, coordinate systems and Fourier analysis. While the details are buried in the cobwebs of my mind, one point is clear: the deft use of the right coordinate system can make the math much easier.
Posit: The whole issue of Global Warming is to be found in the low frequency part of a temperature time series.
Convert the temperature records in to the Fourier space. The key theorem is that the frequency resolution dw/2pi Hz = 1/(N*dt) . (where dw is the frequency resolution and dt is time sample frequency) Usually we are interested in the high frequency stuff and aliasing. But this time, we are acutely interested in the very lowest frequency, the dw (which is in Hz, cycles/unit-time).
Let’s invert. We want resolution time per cycle (2pi/dw) = N*dt. And we want LONG time per cycle, like a cycle time > 100 years to confirm or disprove the GW hypothesis. Well it is staring us in the face: N*dt is the total length of the temperature record.
Suppose now that we take temperature records and using a scalpel of any kind, we take N*dt and make it into n1*dt and n2*dt were n1+n2 = N and n1 < N, n2 < N.
The each of the parts now have a LARGER dw, higher minimum frequency, which means a SHORTER resolution time per cycle than the original. The lowest dw from the original series is now in the bit bucket.
It is possible to stitch together the shorter fragments back into a longer time series. But since the fragments do not contain the lowest dw frequencies, those low frequencies cannot exist in the spliced time series. “Yeah. But there are now low frequencies to be seen in the spliced time series.”
The only low frequency data you see are purely contributed by the mental model of how you spliced them together. You have chased your tail. In the use of scalpel and suture you have thrown away the important stuff and substituted your preconceptions of how they should fit.
It is quite possible that BEST was in the best of intentions using human genome DNA mapping mathematical techniques. These have been quite successful in patching together small DNA sequences into a huge strand. But the DNA mapping is a high-frequency problem. It is great at eliminating errors at the base-pair level. But it is inappropriate for finding 100 year Global warming signal in daily temperature records.
Transforming the data into a Fourier space makes it clear that the scalpel is the wrong tool to analyze the problem.
Poetic justice is so sweet.
Off a link from Climate Etc., I found this letter from Muller to Marc Marano.
http://revkin.tumblr.com/post/4261516714/in-the-olden-days-people-would-have-checked-with
\\Muller’s full comment:
•Some of the bloggers seem to be confused between land temperature data (which shows a greater rise) and the World temperature data (which includes the oceans). It is remarkable how rapidly they accuse me of error or of lying. In the olden days, people would have checked with me before accusing me of wrongdoing. The pressure to be the first to blog is apparently winning. Rich //
(Cough!)
The pressure to be first in the Congressional Record seems to have won hands down.
Anthony and Willis and others have a very simple retort to Muller concerning his Congressional Testimony on data not yet shared for review, comment, or replication.
“You’re Not Allowed to Do This in Science”
http://wattsupwiththat.com/2011/03/18/you%E2%80%99re-not-allowed-to-do-this-in-science/
(and many others via Google)
I contacted Dr. Muller again and this was his response.
“My statement says only that without corrections the data agree with the 1.2 increase. I did not say that our work supports the prior groups.”
I still don’t trust him
Stephen Rasey says:
April 2, 2011 at 11:58 am
Re the Scalpel and Low-Cut Filter.
Thanks so much for this valuable explanation. I hope everyone reads it. I hope Muller reads it.
In addition, I really like your statement to Muller:
“You’re Not Allowed to Do This in Science!”
Stephen Rasey says:
April 2, 2011 at 11:58 am
“Transforming the data into a Fourier space makes it clear that the scalpel is the wrong tool to analyze the problem.”
Amen! For many decades, I’ve used a panoply of proven spectral analyis techniques, including multi-dimesionmal ones, to analyze a great variety of geophysical time-series. Two observations about temperature records stand out on the basis of that experience: 1) data quality is very poor generally and particularly inconsistent at the lowest frequencies and 2) would-be climate scientists are woefully unaware of them and make amateurish mistakes in analyzing the data. They even seem unaware of the categorical difference between a genuine secular trend and low-frequency oscillations that appear trend-like in records of inadequate duration. Sound scientific inferences can be made only on the basis of analyzing long, intact records from a fixed set of stations at environmentally unchanged sites, instead of indiscriminately slicing and dicing measurements from an ever-changing set of stations into a data sausage that is falsely claimed to indicate the average temperature of the Earth’s surface.
I might add that data quality is as poor as my typing skills.
[Already fixed. ~dbs, mod.]
sky says:
April 2, 2011 at 2:17 pm
“Sound scientific inferences can be made only on the basis of analyzing long, intact records from a fixed set of stations at environmentally unchanged sites, instead of indiscriminately slicing and dicing measurements from an ever-changing set of stations into a data sausage that is falsely claimed to indicate the average temperature of the Earth’s surface.”
Take that, BEST! Climate scientists fear reality and will approach it only when protected by a huge mesh of novel statistical methods.
All datasets point to the same thing. So what’s the only possible conclusion? It’s a collusion to hide the non-warming we’ve never seen but we know it’s there!
@Michael Irwin. OK, I understand the preliminary data and algorithms were not shared, but weren’t these deemed as preliminary? He seems to be behind his timeline of algorithms and data posted by last year’s end, so perhaps he had to do something. There are many notes the results are preliminary:
We have done an initial study of the station selection issue.
Our key caveat is that our results are preliminary and have not yet been published in a peer reviewed journal.
Based on our initial work at Berkeley Earth, I believe that some of the most
worrisome biases are less of a problem than I had previously thought.
I see a reaffirmation of openness in the data:
Berkeley Earth has assembled 1.6 billion temperature measurements, and will soon make these publicly available in a relatively easy to use format.
I assume, as promised, the algorithms and code too will be open for analysis. It would look mighty foolish for Muller to make these claims and either not fulfill them or fulfill them with egregious errors in the algorithms or data.
I would recommend patience, while at the same time contacting Muller to understand why he has not posted the data and algorithms prior to giving the preliminary report. If there is a good reason, and the data, code, and algorithms will be forthcoming shortly, I say it’s OK. There are many outs for Muller in a final report. Yet he has to have the credence to continue the work too. It takes money, money I hope he obtains based on his promises to create an open system that anyone can review.
Alexandre says:
April 2, 2011 at 4:43 pm
“All datasets point to the same thing. ”
Rubbish! Vetted small-town station records with consistent low-frequency spectral content show a very much different picture than major city records. They agree only on the subdecadal temperature variations. It doesn’t even require any signal anlaysis expertise to see the difference in “trends”; a twelve-year old has posted a You Tube video on the stark discrepancy. The bureaucratic objective of using all scraps of variously corrupted data in the GHCN dataset only serves to obfuscate the issue.
The fact that a truly discerning study has never appeared in the peer-reviewed literature speaks volumes about the analytic skills of climate scientists. You have to be blind to reality to imply that it’s never been seen by experieced professionals. Academics are good at making such presumptions.
Steve Keohane says (April 2, 2011 at 4:51 am)
Richard Carr & Willis, I thought the phrase containing “yardarm”, referred to one being hung from same.
And the twist, or “mashup” of the conventional was the creatively clever wordsmithing I was congratulating Willis for, Steve.
Desperate people do desperate things. That includes PhDs. This admittedly trite slogan may well explain all the contortions we are observing from the “Berserkely brains.” The noble CAGW crowd has been crowded out of most folks minds by REAL issues, like a paycheck. But these elitists still don’t get it. It is actually very funny, if you sit back and watch from “afar!”
Thanks, Willis, for highlighting this latest episode of CAGWology!
Anyone interested in contacting the David H. Koch Foundation.
It seems like a different independent organization needs to be set-up which is composed of sceptics this time.
For posterity, in 2040, someone needs to have gathered up all the Raw data before it is adjusted out of existence.
With the adjustments made to date, it was so cold in Europe in 1900 (and 1957) that no agricultural crop should have been able to grow. No Wheat was grown in Western Canada then (which is still to this day after all this warming, just barely on the precipace of being able to produce a crop despite intensive selective breeding which has produced more hardy varieties – without a hot October this year, production would have been half of normal – There is no way it was that cold in 1900. The farmers would have been wiped out – along with the rest of our great-grandparents).
I nominate Willis, Ross McKittrick, Steve McIntyre, Roger Pielke Sr., and Anthony to be on the board.
First of all, the raw data needs to be collected from independent sources before it is gone.
Willis,
Simply eyeballing the chart does make it look like the rise was 1.2 C. In fact, its a bit less, though some specific start dates could result in a 1.2 C rise, just not 1900 (or 1880 for that matter). As I mentioned in my post over at Lucia’s, one should be a bit more careful in congressional testimony, but I’d consider this something of a nit-pick that will mostly be problematic to those of us who work with the temperature record than the average observer (who has little frame of reference to figure out how a specific rise in the land temperature record relates to the larger issue).
I think Muller’s mention of his results analyzing Anthony’s data set was poorly done. It is worth noting, however, that they are somewhat consistent with the statement in Anthony’s abstract that “the overall mean temperature trends are nearly identical across site classifications”. Menne (and myself and JohnV) came to similar conclusions using a portion of the data. Of course, this is not the complete story, and there is plenty of interesting aspects of the min and max temps to dive into.
Bill Illis,
For the most part (at least pre-1992), GHCN data is the most raw data that is available for those stations. However, they supplement GHCN with 28000 additional stations or so.
They probably mistakenly used 1.2°C, when they meant 1.2°F.
Roger Carr says: April 2, 2011 at 6:37 pm Thanks Roger, I could only think of silly cartoons, eg. using the yardarm as a catapult.
Bob says:
April 2, 2011 at 1:59 pm
I, on the other hand, don’t understand the response. He said:
That was the statement with the problematic claim of 1.2°C. The other one is the claim of a “0.7 degree rise since” 1957.
His answer is about the agreement of his data with one of those two claims … but says nothing about the incorrect claim.
What am I missing here?
w.
Zeke Hausfather says:
April 2, 2011 at 7:16 pm
“Eyeballing the chart”? I posted the links to all the relevant datasets, and gave the results. I don’t find anything like that, unless he’s using °F … but then the other error, the post 1957 error, is clearly not °F.
“Poorly done?” If you had shared some information in confidence with Mullen and he had used it to attack you, with no supporting data, code, or even general description in Congressional testimony that was immediately reported around the planet, I doubt you’d be so blasé about it … so when anyone googled your name, it comes up with “Mullen tells Congress Zeke is Wrong”, you’d say it was “poorly done”? Somehow I don’t think so.
To me, Zeke, that’s underhanded and deceptive and bad behaviour on a host of counts, personal and professional.
w.
@Michael Irwin. One thing that may be missing is that the target dates for the data and time lines for the completion are past due.
There are two possibilities:
a) Muller lied about his objectives. I refuse to believe this without further evidence.
b) Muller is trying to continue his work that will provide a real, open platform to examine surface temperatures.
I like to think b) is the most likely. Unfortunately, a lot of funding seems controlled by a group of people who want to guilt American society into things that don’t make sense. They seem more related to personal enrichment and political goals than actually solving social problems. that’s why I really appreciated Muller’s talk.
I was so irritated by Al Gore and lemming like response to him that I decided to educate myself with what was actually going on. I probably as guilty as Phil Jones, in that I DON’T want global warming to be real if only to smash the smuggies! But until there are good measuring sticks (I think this particular site focuses on the measurement problem), I don’t know what to think. And even then, I don’t know what to think because there is so much money to be made during any proposed change.
I think the best approach is to assume Muller is a person who really does want to get the measurement part of it right, and to try to work with him. If he opens the Camino, we will know he is honest. If he doesn’t, then we will continue on our way with innuendo and garbage. He seems like the real McCoy, let’s trust he delivers.
I’ve read the previous 272 responses (well, there were 272 when I started). Several things stand out. Several of the responders have no statistical knowledge whatsoever. My favorite:
The more substantial cries were over Muller’s lack of EVIDENCE! Where’s the data? Where’s the code? You sniveling, backstabbing librul! And then dozens and dozens of criticisms of his methodology and data. Rather pointless conjecturing since BEST hasn’t yet released their dataset and code. I especially liked the refutations of the ‘scalpel’ – since none of the commenters actually know the implementing code their criticisms are really the most trenchant. Yeah, that’s sarcasm.
And this whole, “How could he go in front of Congress …. blah, blah” schtick is particularly amusing. One does not just show up in Washington, knock on the Capitol door and say, “I’d like to testify at your hearing today.” Well, one can I suppose, but there’s a glacier’s chance in Glacier National Park of surviving the 21st century that you’ll be ushered in the door. You are requested/invited to testify. It’s an honor, a privilege, and not one easily turned down. Congresscritters don’t take ‘no’ for an answer very well.
Given that, Muller did the best he could do: state his research is not finished, that they’ve done some preliminary sampling (with caveats) and report that the results are in broad agreement with previous studies. The only way he could have pleased most of the commenters here would have been to LIE and say the preliminary data disagreed with previous studies.
That then is the crux of the matter. It isn’t rigorous and accurate science performed to the best of the scientist’s abilities that you’re after – all you want is an end result that agrees with your beliefs. Muller didn’t provide that, therefor he’s fair game to be trashed, pilloried, insulted, demonized, whatever. And the odd thing is I came here not to praise Muller – cuz the last blog comment I wrote (several hours ago) was castigating him on a related but different matter.
I’d like to thank izen and Zeke Hausfather for interjecting a moment or two of sanity into the last hour and a half of my life.
Kevin, you don’t need statistical knowledge to spot BS and false statements. Willis clearly stated where all Dr.Muller was wrong. He went before congress and testified on a 1.2 degrees C rise from 1900, stating it is as per HADCRUT. It is not. If it is land data alone, it should be CRUTEMP. And if you don’t know the difference between these two, you have no business to open your mouth in the first place, that too, in front of Congress, where your testimony is being placed on record.
Second, he underhandedly dissed Anthony’s work given to him in confidence and trust, without showing any methods, data or code, which is dishonest.
These are the crux of Willis’ and many others’ comments. Those are in plain english and anybody with knowledge of english, common sense and a sense of fairness can see that. Which part of those didn’t you get?
Kevin O’Neill says:
April 3, 2011 at 1:41 am
Muller had no reason at all to start a fight. He had no reason to say a word about Anthony’s data, which he had accepted in confidence. He could have said nothing about Anthony’s work. Which would have been appropriate, given that he has presented no code, data, or even a simple explanation of his claimed research that he said shows Anthony is wrong.
Your claim that he had no options is just a sad attempt to rescue him from his own stupidity. He opened Pandora’s box, when there was no need to touch it at all. And while you might think that’s appropriate, I find it a massively stupid action that has come back to bite him savagely. Google his name today, you won’t find folks patting him on the back.
So no, Muller didn’t do “the best he could”. He could have shut his mouth about Anthony and and done much better. He should have waited until his results were published before publicly using them to attack someone, that’s just dirty and sleazy.
w.
PS – you say:
Bullshit. What I want is for Muller to act like a scientist, and publish his work before he stands up in public and attacks someone with it. The fact that you clearly don’t understand either that Muller has done anything wrong, or why people around the globe are objecting to what he did, should give you serious pause, my friend …