In honor of a historic lapse in application of scientific practices, I bring you the “Hide the Decline” mug.
Commentary
We are entering the third week of “Climate gate”, and the slower-to-react elements of the blogosphere are now discussing reactions. At The Volokh Conspiracy,
Jim Lindgren, an attorney, correctly identifies the technical objections to “the trick” used to “hide the decline”.
The Briffa reconstruction mentioned is the one that prompted the “hide the decline†comment. Post-1960 tree-ring data was deleted because it did not match the temperature data, a discrepancy that both created an impression of warming rather than cooling and called into question the use of those tree rings in the first place.
Stripping away all the statistics, the problem with “the decline” is notthat it suggests recent temperature actually dropped. After all, we have real thermometers that indicate temperature rose after the 60s. The problem with “the decline” is that they show a smoothed tree ring reconstruction would indicate they dropped after 1960.
The obvious question which would arise in the minds of anyone who saw a figure that showed declining temperatures in the tree-ring reconstruction during a period of rising temperatures would be: How do we know the tree-rings did not indicate cooling during past warming periods before we had thermometers? For example: Assuming the Medieval Warm period was global, how do we know the tree rings would have shown warming? Why wouldn’t they show cooling, as they do during a sizeable fraction of during the thermometer record?
So, the problem with the decline is that its mere existence casts doubt on the accuracy of the reconstruction of historic temperatures: that is temperatures during periods that predate thermometers. Because the main scientific contribution of the hockey stick is to try to estimate historic temperatures, the mere existence of “the decline” casts doubt on any reconstruction in which a decline is seen. This severely undercuts our ability to judge whether the past temperatures were warmer or cooler than current temperatures.
Of course, the fact that we can’t make such judgments based on a particular proxy reconstruction does not disprove AGW. All it means is that we don’t happen to know whether the current temperatures are unprecedented based on that particular reconstruction.
What is alarming to those who read the CRU zip files in the unauthorized release that kicked off Climategate is the inclination for any group of scientists to conceal the fact that a particular analysis might contain an important uncertainty. Normal practice in science would be to show the post-60s decline in reconstructed temperature along side the known rise in temperature. Acknowledge that the mis-match in the two temperatures (i.e. the divergence) violates the main assumption underlying the reconstruction of historic temperature, and then explain why one thinks a similar divergence (or mismatch) did not occur during the Medieval period.
Under normal scientific protocols, it is not acceptable to perform “tricks” to “hide the decline” from those who will later read the paper, and who may not be privy to the learned discussion in the private emails exchanged by scientists.
Now, you can ponder all this as you sip coffee from your hide the decline mug, which is also available as a beer stein.
I wouldn’t say that the Volokh Conspiracy has been slow to respond. They’ve been consistently blogging about it since at least Nov. 20. http://volokh.com/2009/11/20/climate-scientists-unfiltered/
Mike– I know Jonathan Adler has been blogging from early one. Later EV and Posner added posts. But I think Jim Lindgren only jumped in today. Am I mistaken?
haha, I want one.
Your decline graph isn’t as severe as the MXD data actually is though. What happened?
Jeff–
That’s the data UC sent me. If you have other data, I can make a new graph. I can’t use graphs people are posting on blogs becuase they are using “pretty journal article” style with 1 pt fonts for the curves and 8pt numerals etc. You need BIG fonts for the mug or they will be smeary. Based on the previous mug, these should be large enough… but really, they need to be cartoon-size.
Lucia,
Hold onto your hat. There’s more decline news coming. Don’t ask, just wait. The material needs to be vetted.
Lucia:
“The problem with “the decline†is that they show a smoothed tree ring reconstruction would indicate they dropped after 1960.”
Aside from “the decline†in Briffa, there should also be focus on the reconstructions of Mann and Jones for which UEA, in its November 24 statement (“95% data availableâ€) (see: http://www.uea.ac.uk/mac/comm/media/press/2009/nov/CRUupdate) apparently shows no positive (warming) anomalies for the modern period. UEA’s publication is slightly more than two and one-half years after McIntyre’s similar findings in May, 2007 (see “The Maestro is in da Houseâ€: http://www.climateaudit.org/?p=1553). McIntyre’s observation then about the IPCC’s confidence intervals were also very important. How do develop such intervals about reconstructions in the latest years when the data employed are apparently drawn from thermometers and the like? What can be said about IPCC historical confidence intervals if confidence intervals for the modern period are inappropriately shown to be stronger.
This is the pertinent section addressing the divergence problem:
So, we know our tree-mometers don’t reflect current warming. We don’t know why. Therefore, we assume the problem away, and assume they are perfect tree-mometers at all other points throughout history.
How this ever got past peer review is beyond me. It certainly wouldn’t get by an engineering design review.
Thanks for providing more background on the decline chart. If I might make a point about the choices a scientist has when faced with such fact.
The assumption in using tree rings as temperature proxies is that a tree that is a temperature proxy today was also a temperature proxy yesterday. In the case of this Data briffa and others noted a Divergence in at some sites. That divergence is the object of ongoing studies.
When face with a data set like the decline set the scientist has 3 choices:
1. Reject the data outright, arguing that the site and species display a divergence and thus cannot be trusted to reconstruct the past in any reliable way. On this view the theory trumps the data and all the data is rejected as a freak sort of thing.
2. Reject the theory that trees can function as thermometers. That means no reconstructions from trees of any kind.
3. Modify the theory to account for the anomaly at this site or within this species. That is “point” to some modification of the theory that will allow one to keep bot the theory and the data.
Each of these is LOGICALLY possible when data contradicts theory: throw out the data, throw out the theory, modify the theory ( or data). In Briffa’s case he choose to “modify” the theory by lopping off the offending piece of data. But nobody has offerred up a modification to the theory that would support this. Just that such a modification ( explanation) is required.
Methodologically, I would argue that Briffa should throw ALL the data from that reconstruction out UNTIL a modification to the theory is offered and and confirmed. Once such a modification ( C02 fertilization for example) is confirmed by other examples of divergence THEN one could justify lopping off the data.
But the trick doesnt end there. As I read jeans, Mc and UC and others, another part of the problem is padding the truncated series with instrument data so that a smooth can be created.
I have an issue with smoothing data, especially using acasual filters on a casual process. Smooths make for pretty graphs, and tacking on instrument data or reflecting data or any of these endpoint tricks are mere chartsmanship.
Lucia,
What you are talking about here is the much discussed divergence problem. And it just isn’t true that:
“What is alarming to those who read the CRU zip files in the unauthorized release that kicked off Climategate is the inclination for any group of scientists to conceal the fact that a particular analysis might contain an important uncertainty.“. This is all based on a jocular remark by one scientist in a private email. But CRU scientists had already begun a series of papers, very publicly examining the divergence problem. It begins with the Briffa/Jones/Osborn landmark 1998 paper “Reduced sensitivity of recent tree-growth to temperature at high northern latitudes”.
In their
2004 review, Briffa/Osborn gave a very clear account of the divergence issue as you’ve raised it here:
The above facts seem to support an inference that some slowly varying factor began to exert a very widespread negative influence on the trend of these MXD data from around the middle of the 20th century, with effects at higher frequency also becoming noticeable in some high-latitude regions. For the time being, we circumvent this problem by restricting the calibration of the density data to the period before 1960. This reduces the potential overlap between temperature observations and density measurements and means that less data can be reserved for independent tests of the validity of predictive equations. This situation is far from ideal, but the alternative, using data after 1960 and thus incorporating non-temperature- related bias when fitting regression equations as a function of density variability, would invariability produce earlier estimates of past temperature that, to some extent, too warm.
No concealment here.
Hi
The UK met office now appears to have decided to completely revamp the (CRU) global temp series (see WUWT today) and make the data and code open source. I suspect that if this is carried out much of the supposed global temp rise will simply cease to exist. This is what the work of EM Smith on GISTemp is gradually revealing.
With a better global record the whole rise/decline problem might then morph into a new set of issues.
I might buy a Patrick Michaels erasing Hansen’s graphs mug. Wait, I’m not your audience.
“No concealment here.”
I mean really. If the denialists actually honestly discussed the divergence problem, they might have a valid point. I’m torn as to whether the problem is their inherent laziness, an inability to understand, or a breathless need for instant rhetorical gratification end attention. It’s probably a combination.
Could you explain which people you include in “denialist” here? I realize there are people who don’t understand what “the decline” is, and some who also don’t understand the divergence problem. But there are plenty of people honestly discussing both fairly at climate audit.
Nick Stokes
Yeah. Tell that to Jon Stewart of the daily show.
Look, that claim does not pass the laugh test, nor does the fact that the divergence problem is discussed. Sorry. No.
Nick Stokes: No concealment here.
No concealment at all. You are absolutely right. They stated, to everyone’s face (6 years later, mind you), that they had a theory, that data they had collected contradicted that theory, they had no plausible explanation for the contradictory data, so they were throwing the inconvenient data out.
The relevant quotations, from your passage, are:
… the alternative, using data after 1960 and thus incorporating non-temperature- related bias when fitting regression equations as a function of density variability, would invariability[sic] produce earlier estimates of past temperature that [are], to some extent, too warm.
and
For the time being, we circumvent this problem by restricting the calibration of the density data to the period before 1960.
One has to give them marks for chutzpah. Why this was considered publishable science is a good question.
Lucia
You are a denialist.
Absolutely, 100%.
This stupid mockery of the “Hide the decline” nonsense proves it. You know full well this has been discussed at length by Briffa and other dendrochronlogists. Go and look at the Delayed Oscillator blog.
Punch my ticket
“they had no plausible explanation for the contradictory data”
See this is where you have been led astray by the nonsense on this blog.
Briffa published has on the divergence problem, and it is well known to dendrochronolgists.
Nick Stokes,
you very clearly show that this issue was known and papers were published investigating it.
Why then, is this not mentioned and those papers referenced in the Hocky Stick Papers where its effect would be apparent if data wasn’t deleted and replaced?
Why would the differences in data sources not be clearly shown and explained with reference to the papers and to the reasons data after ~1960 shouldn’t be used for these reconstructions??
For experts in the field, I would still think that this would be important to keep in the forefront rather than just an accepted issue that no longer needs to be mentioned.
Until there is at least a working hypothesis as to what is happening this is simply not credible.
Take another swing Bro!!
Curse you Lucia!!
I am going to have to get a job so I can afford mugs for all the Warmers I know!!!
Nathan, the publications on divergence tend to be a ragtag collection of excuses. Briffa et al are trying desperately to save a theory by pulling possible explanations out of thin air, acid rain being the most prominent. It’s backed up by nothing, nothing at all, other than “We have a theory, the data correlated well for 100 years, now it’s uncorrelated if not negatively correlated for 50 more, and we think we can save the theory by pointing at X, Y, and Z (with little to no evidence for X, Y, or Z).”
I’m reminded of the furious attempts to add yet more epicycles to salvage Ptolemy.
Nobody proposes a mechanism for the diverence and shows that the mechanism applies to the series in question. The papers are a lot of arm waving. The best approach in my mind is Wilsons. You guys go find that paper.
Punch my Ticket
“Nathan, the publications on divergence tend to be a ragtag collection of excuses. ”
What, so you just decided it was bad? You just decided they weren’t plausible? Do you have any reason for this?
Remember it’s not just Briffa doing this. It’s the whole of dendrochronology. It’s not some secret problem, it’s not some subterfuge, it’s been documented many times.
Steven
“Nobody proposes a mechanism for the diverence and shows that the mechanism applies to the series in question. The papers are a lot of arm waving. The best approach in my mind is Wilsons. You guys go find that paper.”
Great Steven. You like that one the best…
This issue of ‘hiding the decline’ is about ridiculing dendrochonologists, and about attempting to make it sound like some sort of mysterious underhanded data fixing is going on.
It’s dirty politics, nothing more.
Have you been over at James Annan’s blog debating climate sensitivity with him?
Nathan, divergence has been documented many times and the end result is always, “Well, it could have been this or it could have been that”, i.e. more hypotheses of more epicycles, unsupported and quite possibly unsupportable.
It seldom occurs to anyone collecting a paycheck in this field (and writing grant proposals to ensure that continues) that 100 years of correlation followed by 50 years of no or negative correlation can also be explained with, “Hell, we have been barking up entirely the wrong tree because of sheer coincidence.”
Boris the “denialist” position on hide the decline is as annoying and stupid as the people who performed that data massaging on that particular chart. Here’s a question. How important was this chart? Don’t be too rash here. especially if you havent taken the time to read the emails WRT this, and if you don’t have access to all the material. hehe.
Nathan:
Nah, it’s just poking fun. Lighten up.
here’s another way to look at hide the decline. I think everyone acknowledges the existence of the divergence. Interpreting the cause of the divergence is less clear. Let’s say its unsettled. We kinda know its unsettled because there is an amount of literature attempting to explain it non of it generally accepted as THE definitive answer on the divergence.
So, the IPCC is charged with representing BOTH the consensus and the UNCERTAINTY.
Does the chart that hides the decline represent BOTH. Does that chart and that chapter represent BOTH the consensus and the uncertainty?
And who agrees with me?
“I agree very much with the above sentiment. My concern was motivated by the possibility of expressing an impression of more concensus than might actually exist . I suppose the earlier talk implying that we should not ‘muddy the waters’ by including contradictory evidence worried me . IPCC is supposed to represent concensus but also areas of uncertainty in the evidence. ”
Guess who wrote this?
Guess what he was talking about?
But wait, there’s more.. no you will have to wait.
There was a new paper a few days ago that showed Aspen Trees (which are direct competitors to the Pine Trees which most cores come from) have increased growth rates by 50% in recent decades due to increased CO2 levels.
http://www.physorg.com/news179118204.html
So if someone could show Pine Trees don’t respond as well to increased CO2, then that could be a real explanation. The others don’t cut it and they were truly “hiding the decline in the proxy versus temperature correlation” without a valid explanation.
I never use data that does not work backwards and forwards and through all possible values. It goes in the recycle bin if it doesn’t pass this test and so should tree-ring data if there is not a proven reason for the divergence since 1950.
Punch My Ticket (Comment#26328)
There are two debates going on here, which adds confusion. I actually share some of your doubts about proxy reconstructions. The divergence problem is real, and diminishes my faith in the result. And it doesn’t have much to do with AGW.
That’s issue 1. The other debate is about the dishonest attempts here to represent it all as underhand, “hidden”. It isn’t – it’s all out there. To quote Briffa’s review phrase again, “This situation is far from ideal“. It’s the current state of knowledge.
Bill Illis,
“So if someone could show Pine Trees don’t respond as well to increased CO2, then that could be a real explanation. The others don’t cut it and they were truly “hiding the decline in the proxy versus temperature correlation†without a valid explanation.”
Wouldn’t this still be a problem? That is, just because Pine Trees do not respond as well to increased CO2 doesn’t mean that they would show decreases in their growth indicators. They would have lower or flat growth rate indicators. Wouldn’t someone actually have to show that CO2 is a negative influence for Pine Trees to match the Decline??
Nick Stokes,
“The above facts seem to support an inference that some slowly varying factor began to exert a very widespread negative influence on the trend of these MXD data from around the middle of the 20th century, with effects at higher frequency also becoming noticeable in some high-latitude regions. For the time being, we circumvent this problem by restricting the calibration of the density data to the period before 1960. This reduces the potential overlap between temperature observations and density measurements and means that less data can be reserved for independent tests of the validity of predictive equations. This situation is far from ideal, but the alternative, using data after 1960 and thus incorporating non-temperature- related bias when fitting regression equations as a function of density variability, would invariability produce earlier estimates of past temperature that, to some extent, too warm.
No concealment here.”
lemme see, there is no hypothesis for the decline either. As there is no reasonable hypothesis for the decline, how can we then ignore the possibility that these trees in these areas are NOT reliable thermometers, if thermometers at all??
Do you REALLY want to keep discussing this sham??
Hey, thanks for a mug with which right-handers can show off a graph.
steven mosher (Comment#26331)
… Guess what he was talking about?
Yes, it’s Briffa having a tetchy argument with Mann, back in 1999 talking about AR3. And yes, they have results that show differences, and B is sticking up for his, and not wanting it to be said that they agree.
So what happened? Sec 2.3.2.2 of the AR3! It lays it all out – Mann’s result, Briffa’s, Fig 2.21 showing the difference – all discussed. No spurious consensus there.
There’s a moral here too. The tense statements that you pick out were based on a misunderstanding between them, for which Mann apologised. They realise that their emails are imperfect, iterative communications. Yet you folk think you can pick out odd phrases out of context, from a converstaion of which you were not part, and say they mean that climate science is undermined.
Nick:
“There are two debates going on here, which adds confusion. I actually share some of your doubts about proxy reconstructions. The divergence problem is real, and diminishes my faith in the result. And it doesn’t have much to do with AGW.
That’s issue 1. The other debate is about the dishonest attempts here to represent it all as underhand, “hiddenâ€. It isn’t – it’s all out there. To quote Briffa’s review phrase again, “This situation is far from ideal“. It’s the current state of knowledge.”
1. It doesnt have much to do with AGW?
Hmm. It doesnt have much to do with RADIATIVE PHYSICS. But you should read the assumptions of the 10th annual WMO congress ( 1987 I believe) which gave the rationale for the FOUNDING of the IPCC.. Something about the current warming being unprecedented in human history…
kuhnkat (Comment#26344)
lemme see, there is no hypothesis for the decline either.
Again, two separate debates. Yes, there are legitimate criticisms of proxies and dendro. I am sympathetic. So is Briffa. Actually they did have a hypothesis – ozone. It fills the next section.
But that has nothing to do with climategate, which is about allegation of scientific malpractice. There’s nothing wrong with openly doing the best you have with what you’ve got. The only debate is whether it is worth publishing.
The alleged issue in this post was “hiding”. And there isn’t any on divergence.
steven mosher (Comment#26347)
There were no major paleo temperature reconstructions in 1987. Indeed, GISS temp was very recent, and there was nothing else. The case for AGW was not built on this.
IPCC graphs did not append a temperature series to a proxy one. They did superimpose the series on the same graph.
Indeed, I think there has been much overanalysis of Jones “trick” crack. Whatever may have appeared in subsequent papers, I think the “trick” he was referring was just the addition of the instrumental curve to avoid the impression of global cooling.
kuhnkat (Comment#26344)
The idea is that Aspen Trees would out-compete Pine Trees and the Pines’ growth would decline relative to earlier periods.
The possible explanations for the divergence include pollution, acid rain, aerosols blocking sunlight, increased competition from other tree species, reduced Pine Tree growth from CO2 fertilization and increased fire-fighting efforts which increased competition from other species.
But generally, the NH forest cover and biomasss increased substantially after 1950. So it was not acid rain or pollution or increased aerosols.
[The fire-fighting effort is just one I like to bring up because if you ever look at a satellite picture of the northern forests, all you see is really big fire scars. They used to burn hundreds of kilometres across. There are very few areas where mature (older than 50 year) trees actually exist because they used to just die from the fires every 30 years or so.]
So increased competition from other species (plant or animal or us I guess) is the only possible explanation that holds up.
Nick
Hidden from whom? What was the effect of showing the actual decline vs. hiding it when these and similar were shown to the public? Whether or not we add loaded terms like “dishonest” or “underhanded”, using the trick to “hide the decline” hide the divergence from the public when the graphs were used as a message to communicate to them.
The effect of this graph was:
a) Members of the public who did not delve into the paper were not made aware of the uncertainty.
b) Members of the public who bothered to delve into the paper noticed the “trick” and lost confidence in the IPCC process because it presented this information in a way that concealed this uncertainty from the casual reader.
Once the emails were hacked the general members of the public became aware of the issue, and they too came to lose confidence. Moreover, those members of the public knew “the decline” had been hidden from them.
Really Nick?
http://www.grida.no/publications/other/ipcc_tar/?src=/climate/ipcc_tar/wg1/048.htm
Now.. Look at briffa’s figure 2.21.
Figure 2.21: Comparison of warm-season (Jones et al., 1998) and annual mean (Mann et al., 1998, 1999) multi-proxy-based and warm season tree-ring-based (Briffa, 2000) millennial Northern Hemisphere temperature reconstructions. The recent instrumental annual mean Northern Hemisphere temperature record to 1999 is shown for comparison. Also shown is an extra-tropical sampling of the Mann et al. (1999) temperature pattern reconstructions more directly comparable in its latitudinal sampling to the Jones et al. series. The self-consistently estimated two standard error limits (shaded region) for the smoothed Mann et al. (1999) series are shown. The horizontal zero line denotes the 1961 to 1990 reference period mean temperature. All series were smoothed with a 40-year Hamming-weights lowpass filter, with boundary constraints imposed by padding the series with its mean values during the first and last 25 years.
Care to comment?
Read this first.
http://camirror.wordpress.com/2009/11/26/the-trick/#more-62
Then Can you explain to me the bit about the filter?
What data was chopped off the proxy series?
To produce the smooth what 25 years were used?
The series ends in 1960. Talk to me about placing such a series based on its value in the 1961-90 period.
Nick
This is not “the nature trick”. It’s what some have describe as “the trick” on blogs, but it’s not “the nature trick”.
So let me get it straight… Nick doesn’t understand what “trick”, the “trick” refers to?
Either that or he knowingly misrepresented that trick himself…no doubt to “hide the decline” in Phil Jones’ reputation. >.>
Nick Stokes (Comment#26348)
Yes, there are legitimate criticisms of proxies and dendro. I am sympathetic. So is Briffa. Actually they did have a hypothesis – ozone. It fills the next section.
CA has covered this eg Osborne 2005
A number of factors were taken into account when selecting the most appropriate periods for calibration and verification of the gridded density data against observed temperatures. The most important factor is the identification by Briffa et al. (1998a) of a recent downward trend in the high latitude tree-ring density data, relative to (and apparently unrelated to) warm-season temperature. This density decline becomes large enough to impair the calibration after about 1960. For this reason, both Briffa et al. (2001) and Briffa et al. (2002a) used only pre-1961 data for calibration of their subcontinental, regional temperature reconstructions. This is a reasonable choice, provided that it is explicitly stated that this approach assumes the apparent recent density decline is due to some anthropogenic factor and that similar behaviour is assumed, therefore, not to have occurred earlier in the reconstruction period – which would otherwise introduce bias in the reconstructed temperatures. At present, no satisfactory explanation of the relative MXD decline has been identified, and further work must dictate whether this assumption will be supported or rejected (Briffa et al., 1998a, 2003, and Vaganov et al., 1999, discuss and investigate possible causes).
Due to some “unknown anthropogenic influence” is clearly a case of the mind projection fallacy eg Jaynes probability as logic.
(A) (My own imagination) – (Real property of Nature)
(B) (My own ignorance) – (Nature is indeterminate)
Similarly the ‘ozone hypothesis’ in the biological recorders is constrained by theory,this is well known eg Darwin
steven mosher (Comment#26356)
Really Nick?
No, Steven, you’re changing the subject. You were referencing (#26331) Briffa’s email about differences between the proxy representations. It was a different email, and there was no issue of endpoint conditions there. The basic issue was that Briffa’s curve was warmer overall (and did have a hint of MWP)..
Carrick (Comment#26361)
So let me get it straight… Nick doesn’t understand what “trickâ€, the “trick†refers to?
Would you like to offer your version?
Lucia,
I read all the major climate blogs, but I have to say that I find yours to be very much more interesting as time goes on. I read RC just for balance and I don’t really take anything posted there very seriously given the level of filtering and spin. Joe Romm is entertaining, but the shrillness and vitriol make it rather uncredible. WUWT is also entertaining, but comments tend to be very one-sided and anyone posting a warming oriented view is rather quickly quashed. CA is interesting and SM tries very hard to keep things focused on the narrow area he’s interested in. I’ve been snipped there multiple times when I ventured too far off topic. Fair enough, but his approach does tend to throttle the discussion a bit more than I’d like.
I’m probably close to the Denailist TM view of the world as the term is used by folks that believe that disaster is impending. I just don’t see credible evidence that we’re headed to hell in a hand basket. That said, the arguments that go on here driven by the Simons and Nathans make me think about views that are contrary to mine, and that’s a good thing (I disagree with them so strongly most of the time that I almost can’t believe that I’m saying this :-))
You demonstrate an openness to ideas, a balance (Nathan’s Denialist TM rant above notwithstanding) and a pure intellectual honesty that are approaching unique in the climate science blogosphere.
Thanks for that.
OA
Thanks Ozzie!
Nick Huh?
I’m talking about hiding the decline. which first came up in the discussion of 1999. two issues: the truncation and the smoothing.
I dont even care about the shaft
Nick, it’s been discussed ad nauseam. But here it is again:
1) delete e.g. the data points after e.g. 1961 (Briffa series),
2) replace with vertically shifted instrumental temperature series (to remove discontinuity where the instrumental data gets grafted in),
3) filter spliced data
4) truncate filtered data after 1961
The effect of this is to change a downward trend into an upward trend at the end point.
That is, to “hide the decline”.
There’s nothing wrong with showing the entire filtered, dendro series and overlaying it with the original time series, which is what you suggested was done.
As Lucia points out the problem with this is it hides the fact in that graph that there is a divergence problem.
The point is not whether the divergence problem is discussed elsewhere, and is simply whether this is an honest way to portray the data on this graph.
I can imagine what would happen to a grad student trying to present this chart as part of his thesis defense, were he “caught out” on this error.
The fact that Mann and Jones have collectively told a half-dozen different stories about “hiding the decline” doesn’t make them look much either.
McIntyre’s replication of Mann’s Nature trick is show here.
“lucia (Comment#26309)
December 5th, 2009 at 5:58 pm
If the denialists actually honestly discussed the divergence problem,
Could you explain which people you include in “denialist†here? I realize there are people who don’t understand what “the decline†is, and some who also don’t understand the divergence problem. But there are plenty of people honestly discussing both fairly at climate audit.”
LOL.
This whole topic epitomises your real attitude.
It’s not about science, it’s about a public pillory and humiliation. This is humorous, apparently, on a topic that has serious consequences for not just the human race, but most species on this planet. This is more important for you than it is to spread the word about than the science behind AGW, that is as sound today as it was yesterday.
Carrick (Comment#26381)
Thanks for spelling that out. It is indeed what was done in the WMO 1999 cover, at least according to CA’s replication. However, I think that what PJ was talking about may be different, and may indeed be what Gerry North believed. BTW, it’s not a big issue for me – just this version seems to me more probable.
1. As Lucia said, PJ referred to Mann’s Nature trick. But I don’t think what you’ve said describes what is done by PCA methods in MBH98. Nor does Mann, who famously said in response to a RC comment that this splicing had never been done. I believe he has now conceded that it was done in the WMO diagram.
2. Mann’s GRL 1999 paper was the ancestor of the diagram that appeared in the AR3, and that does do what I described, superimposing the instrumental on the multiproxy plot. My guess is that PJ just got the papers mixed up. I cannot see what the “Nature trick” could be.
3. In the recent CRU account, they show the final WMO plot, but also at the bottom an “alternative version”, which shows the superposition. My guess is that this was an earlier version, and is what PJ was referring to in his email.
Carrick (Comment#26381) December 5th, 2009 at 11:05 pm
Thank you Carrick. I Wasn’t Discussing the humps in the shaft but hiding the decline. You will note that nick, who often comments at CA about endpoint smoothing has nothing to say about the “tacking” on a shifted instrument series onto a proxy record so that a smooth can be drawn. Also, thanks for that clear exposition. I
Nick. perhaps we can address 1 point in that Ar3 figure.
First the description of the smoothing. can you just simply say that its misleading or incomplete or something that indicates that you would do it differently.
You have a proxy series that is cut off at 1960. What do you do if you want to smooth that?
Bill Illis,
thank you.
Nick,
I was going to respond, but, I am outclassed by the rest of your adversaries.
Have fun!!
“That is, to “hide the declineâ€.
There’s nothing wrong with showing the entire filtered, dendro series and overlaying it with the original time series, which is what you suggested was done.
As Lucia points out the problem with this is it hides the fact in that graph that there is a divergence problem.
The point is not whether the divergence problem is discussed elsewhere, and is simply whether this is an honest way to portray the data on this graph.”
They aren’t hiding any real decline, there is no argument that the temperature record did not dive like the proxies did. This is just another beat up over how to handle a problem with a proxy. If they had tried to hide a decline in the actual temperature record, I would expect the outrage that has been venemously expressed around the blogs. At worst, it is just a matter for slap on the wrist.
The fact that such an insignificant matter inspires such feigned outrage, while at the same time generating a coffee mug for laughs just leaves me wondering. Get your story straight.
bugs,
“They aren’t hiding any real decline, there is no argument that the temperature record did not dive like the proxies did.”
you are correct. They are not hiding a temperature decline.
The decline in the proxie record in comparison to the temperature record calls into question the reliability of those trees as temperature proxies.
The argument about whether there is a reasonable hypothesis for the decline that would not be applicable to the rest of the record is central to maintaining the usefulness of those trees as proxies.
If something that could only be attributed to modern humans or something that could only be reasonably attributed to the modern period is not found to be responsible for the decline you can not use those trees as proxies.
As Bill Illis pointed out, the data about modern tree growth would tend to exclude these possibilities.
“The decline in the proxie record in comparison to the temperature record calls into question the reliability of those trees as temperature proxies.
The argument about whether there is a reasonable hypothesis for the decline that would not be applicable to the rest of the record is central to maintaining the usefulness of those trees as proxies.”
Given that dendrochronology is only a new science, I’m not surprised there are problems with it that have not been resolved yet, and possibly may never be resolved. It could be a dead end. I don’t see why it’s made out to be the end of the world. It’s like bagging Einstein because he never could work out the GUT, and persisted in following what turned out to be scientific dead ends.
Dendrochronology is not a nascent science. If you mean dendroclimatology, or perhaps dendrothermometry I might agree. I have no issues with calling it that and leaving it to them to work out the kinks and come back in a few years when they’ve got things figured out. But that sure doesn’t sound like settled science to me. It almost sounds like reasons for healthy skepticism. Hmm, imagine that.
Bugs,
some of the folks here have worked in new science areas. I buy the theory that trees can act as a proxy for temperature, but in the case of the divergence I’d have to be conservative and do one of the following.
1. Leave the divergent part in and suffer the consequences because of a bad correlation with temp. OR
2. Toss the whole series out and live with wider confidence intervals on the reconstruction. OR
3. Show results for all cases ( with/ without/ truncated) and comment on all three and the pros and cons.
Usually in my work I just did #3 and managment decided what answer they “liked” and then we would fight over them hiding the data…hehe.
” But that sure doesn’t sound like settled science to me.”
The science is ‘settled’, which is a very simple way of saying the pieces mostly fit, which they do. The IPCC report looks at several lines of assesing the warming, so that we don’t rely on any one line of evidence. Explaining the full meaning what is behind ‘settled’ is complex, and the phrase is completely misrepresented as meaning ‘our knowledge is perfect’, which it isn’t and never will be.
Bugs,
all of a sudden we aren’t talking about a reason for a lynching.
steven mosher (Comment#26389)
“You will note that nick, who often comments at CA about endpoint smoothing has nothing to say about the “tacking†on a shifted instrument series onto a proxy record so that a smooth can be drawn.”
Not so. See this thread
#24601 and #24609.
When you refer to the AR3 figure, do you mean 2.20 or 2.21? The smoothing method is spelt out in detail for 2.21, and seems orthodox. In 2.20, AFAIK, the unsmoothed instrumental data is shown, but not involved in the smoothing. I presume smoothing is done as in 2.21. Do you have contrary info?
As to what to do with smoothing a cut-off series – as in say #24601 referenced above, my recommendation would be to smooth in the regular way, using the pre-1960 data, as described in Fig 2.21, or better, using linear extension (MRC). Don’t introduce new data.
Nature trick is shown in here http://www.climateaudit.org/?p=7810#comment-367024
Data is here http://www.climateaudit.org/wp-content/uploads/2009/11/mbhsmooths1.txt
with columns
% Year AnnualRecon Instrumental MB98SmoothTrick MBH98SmoothNoTrick MBH99SmoothTrick MBH99SmoothNoTrick ]
Mann admitted the MBH99 trick in comment at RC, but they still seem to deny Nature trick. I guess it is because gavin said there was no trick in the smoothed series of MBH98 5b
What about the ‘hide the PC1 residual’ shocker? If MBH99 had not forced PC1 to align with the NT series on the basis of the “suggestion” that they were influenced by CO2 fertilisation then the hockey stick would have had a still deeper blade!
Truly shocking – MBH99 hid the real extremity of 20th century warming by means of this trick! I’m looking for an email on the subject – “We have to get rid of the exaggerated 20thC warming – can anyone think up a way of hiding this?” 😉
‘Inferences, uncertainties and limitations……’
Simon Evans, the only effect of PC1 fixing was the increased millennial decline:
http://signals.auditblogs.com/files/2008/07/fig11.png
UC – ok, I’ll take your word for that – my bad. Fixing it increased the indication of the range in natural variability though 😉
“Nobody proposes a mechanism for the diverence and shows that the mechanism applies to the series in question. The papers are a lot of arm waving. The best approach in my mind is Wilsons. You guys go find that paper.”
Come on, steven, if you’ve read Wilson’s paper (and I’m sure you have–I believe we discussed it at CA a long time ago) you know he discusses several proposed mechanisms for divergence. Attribution is more difficult. However, the skeptics position that trees suddenly stop responding to warm temperatures is not well argued either. You guys treat that as if it were the null hypothesis, which is silly given that divergence is only seem in a limited latitude area and even then not at every site within that range.
“Could you explain which people you include in “denialist†here?”
Most of your commenters, I would say. The Ravens and kuhnkats and tetrises (tetri?) of the world. And perhaps every last Anthony on the planet at this point. I haven’t followed CA for years, so I wouldn’t know what is being discussed there. I wouldn’t consider either you or Steve Mc denialists, but you both take a paritsan angle toward AGW. Despite your one or two shots at Monckton, you focus on the skeptic/denial side of things.
Boris–
Ok. But its worth noting that even if you don’t follow climate audit, the blog posts there describe the divergence problem in exactly the way you find acceptable. Apparently I also apply the way you find acceptable.
I agree there may be people somewhere who may describe the divergence in the incorrect manner you suggest. I don’t happen to recall Raven, Kuhnkat or tetris doing so.
FWIW: The word “partisan” relates specifically to political parties (e.g. “Republican” or “Democrat”.) Skeptic/denialist/climate activiest/ etc. are not political parties. They can be points of view– and this is a different thing.
In any event, in my opinion, if you think I focus on the denial side of things, this means that you include discussions that entirely accept the truth of global warming in “denying global warming”. This is odd, but it is my impression that many people define “denying global warming” as “not constantly and actively working to convince others that run away global warming is imminent by constantly discussing the risk of the most alarming predictions and making it appear scientist think those most alarming, and unlikely, outcomes are more probable than not.”
In my opinion, when people define “denialism/skeptic” this way, they undercut the public’s confidence in the truth of AGW because they come to believe that non-denialism requires adherence to a creed of catastrophism.
Of course this is only my opinion, and it may be that representing a constant stream of unlikely horribles (which then don’t pan out) raises the public confidence. Who knows?
bugs:
In your fevered imagination anyway.
“the skeptics position that trees suddenly stop responding to warm temperatures is not well argued either”
Huh? I think the skeptics position is that the rings in question can’t be said to know the past temperatures if the disagree with recent temperatures. Pretty elementary. The position that trees suddenly stopped responding for the 20th century warming was Briffa’s.
The problem with all the possible explanations of the divergence problem is that there aren’t any ways to test them. Given this state of affairs it would be better to avoid such trees until an explanation which does not invalidate them is reasonably supported. Such an explanation would have to be exclusive to the late twentieth century, otherwise there is no valid calibration and no valid proxy based on them for the past.
Simon:
Which was seriously needed.
Even with that corrected, Mann’s reconstructions don’t exhibit the correct level of natural variations (e.g. see von Storch’s comments)
UC (Comment#26426)
But the problem with supposing that PJ was referring to that is that its effect is almost invisible. It’s not much of a “trick”. It shows up on your plot only because you’ve added the red and green lines.
And I don’t believe it is what is depicted on Lucia’s mug.
Carrick,
I don’t think they did either. I think Mann et al. 2008 is better, though personally I have limited confidence in reconstructing past temps in fractions of degrees . I have very, very much less confidence, though, in claims that the MWP was +3C warmer (or whatever Plimer or CO2 Science or whoever might claim), so I think these multi-proxy recons are useful guides as to whether projected future temps are taking us above what has been experienced.
Steven Mosher:
My own view is they generally respond to other phenomena such as precipitation changes, changes in average cloud cover, etc, and further that for most of these factors, there are levels which stimulate an optimal growth rate.
In fact, one explanation for the divergence is precisely this problem, namely that once the trees go above their optimal temperature (since say 1960), you see a decrease in growth rate after they’ve reached their “peak growth rate”… e.g. a “divergence”.
Why is this important? Well because these very same tree series are being used to argue against a MWP being comparable in temperature to the 1980-current period.
If they suffer a similar high temperature diverge during the MWP, as is observed since e.g. 1960, then it really is a big deal for example to “hide the decline”.
I’ve wondered why people are not considering teh ‘trick’ graphs as works of science. Just what merit do they have in regard to the scientific process.
If scientific merit is interpreted as the degree to which a result is useful (in the sense of its predictive power) then these graphs have no merit. Indeed, one could very well go further and say that these graphs are not works of science. They merge two results and then claim it is a third. The resulting graph lacks any predictive power. It is not science but is anti-science in the sense used today. It is a poltical statement cast in scientific terms
Simon, if you look at the spectral content of Mann 2008, it still utterly fails to give a realistic spectral response. I know you’ve seen this, but this is for others who haven’t:
In this figure you see a comparison of GISS (red), Mann during calibration (purple) and Mann during reconstruction period (blue).
Obviously something went seriously wrong during the the Mann reconstruction period.
Speaking of past reconstructions, according to IPCC AR4, during the the last interglacial period, the temperature was about 5°C warmer than current, and ocean levels were about 4-m above current. So simply having higher than current temperatures in itself isn’t unheard over, even in “recent” geological history.
If they were growing vines then it could be warmer than today. We don’t know what is “normal” for glacier ice, because all we have is data going out of a Little Ice Age to current, and don’t know what these glaciers looked like during the MWP (before the decent into the LIA).
Boris (Comment#26452) December 6th, 2009 at 10:43 am
“Nobody proposes a mechanism for the diverence and shows that the mechanism applies to the series in question. The papers are a lot of arm waving. The best approach in my mind is Wilsons. You guys go find that paper.â€
“Come on, steven, if you’ve read Wilson’s paper (and I’m sure you have–I believe we discussed it at CA a long time ago) you know he discusses several proposed mechanisms for divergence. Attribution is more difficult. ”
Isnt that what I said. I think we are discussing the same Wilson paper ( wilson esper millenium study?) anyways, people are proposing ideas and mechanisms but as you put it “attribution” is more difficult.
I’d propose this. Most of us here if given the choice would focus all of our attention on that divergence thing. If we wrote up papers or results for our bosses we would say : “look at the cool problem we found. Look something that doesnt make sense, lets go understand it” and we wouldn’t let somebody show a chart without showing our interesting problem.. YA BUT LOOK AT THIS!
wow this is cool, it could mean a lot of things or it could just be a bunch of stupid trees at this one site.
At least that’s the way I think.
I think I’ve said similar thing WRT YAMAL. I have no issue with briffa throwing in that huge ass 6 sigma tree. I just want to see what it looks like without that tree. That’s the whole point of all of this. Those of us who have looked at historical data or experimental data Have all faced those situations where the results didnt come out as expected. And then cutting and slicing and rejecting sub sections of the data starts to happen. That whole data selection process is key and needs to be open and transparent.
Nick Stokes (Comment#26465)
Oh well, you are playing one of these games once again. UC was not claiming that “fix the CO2 in PC#1″-trick was the one what is under discussion here (Mike’s Nature trick). I don’t even believe that you thought so, but you are once again trying to confuse and obscure uninformed readers — your favorite hobby.
And no, that fix trick is not something showing up only in UC’s figure. The trend line (which would be flat without this trick) is plotted in the original MBH99 hockey stick figure, and even discussed:
” The substantial secular spectral peak is highly significant relative to red noise, associated with a long-term cooling trend in the NH series prior to industrialization (\delta T = -0.02 ^{\circ}/century). This cooling is possibly related to astronomical forcing, which is thought to have driven long-term temperatures downward since the mid-Holocene at a rate within the range of -0.01 to -0.04 ^{\circ} C / century [see Berger, 1988]
…
Berger, A., Milankovitch theory and climate, Rev. of Geophys., 26,624-657, 1988.”
And that didn’t go unnoticed. Keith Briffa:
http://www.eastangliaemails.com/emails.php?eid=138&filename=938031546.txt
“I do not believe that global mean annual temperatures have simply cooled progressively over thousands of years as Mike appears to and I contend that that there is strong evidence for major changes in climate over the Holocene (not Milankovich) that require explanation and that could represent part of the current or future background variability of our climate.”
Simon Evans
“UC – ok, I’ll take your word for that – my bad.”
Here’s the code http://signals.auditblogs.com/2008/07/01/hockeystick-for-matlab/ , in case words are not enough 😉
Nick, my favorite denialist. Can you explain PJ’s trick to me, once more ? Adding a separate instrumental series to the figure ??
Jean S (Comment#26474)
Nick Stokes (Comment#26465)
Oh well, you are playing one of these games once again. UC was not claiming that “fix the CO2 in PC#1″-trick was the one what is under discussion here (Mike’s Nature trick).
Not at all. I was trying to track down what exactly was “Mike’s Nature trick” – something that was actually in MBH98 (Nature). UC pointed to something that is actually in the paper, but, as you say, doesn’t seem to be the “trick”. I’ve seen much discussion of this issue, but no answer. Do you have one?
My theory above is that PJ meant the GRL paper (’99).
steven mosher (Comment#26282)
care to shed some light on that comment: jean_sbls@yahoo.com
UC (Comment#26477)
Can you explain PJ’s trick to me, once more?
I’ve been trying to figure out PJ’s reference to “Mann’s nature trick”. I’m not sure what “PJ’s trick” you are referring to.
Nick Stokes (Comment#26478)
Come on, you know as well as I do that the trick was used in MBH98 Figure 5b:
http://www.climateaudit.org/wp-content/uploads/2009/11/mbh98smoothzoom.png
http://www.climateaudit.org/?p=7810#comment-367076
http://www.climateaudit.org/?p=7810#comment-366982
I can not give you the compliment UC did (“favorite”), but you have to be the worst denialist I’m aware of.
Jean S (Comment#26483) December 6th, 2009 at 2:29 pm
There was no manipulation of actual temperature records, only problems in how to deal with proxies that were clearly at variance with reality for a period of time late in the temperature record. Problems that are still being worked on in a field of science that is still in it’s early days.
For some reason, this is held up as a travesty of science, that an area of developing science is not yet prefect. Now who would have guessed that?
It is also being held up as evidence that the whole of AGW science collapses like a deck of cards, when it is not the underlying basis of AGW, there is a physical basis for it. The actual paleo evidence is only a small part of the case of the whole WG1 report, which relies on several independent streams of research to create the case for AGW.
For an alternative way of looking at errors, try this.
http://clearclimatecode.org/
Bugs in Gistemp that climateaudit missed! What sort of auditors are you? The horror.
Carrick,
I said “better”, not perfect 😉
according to IPCC AR4, during the the last interglacial period, the temperature was about 5°C warmer than current
Er, no – 3C to 5C warmer over Greenland and Antarctica is indicated, but “Simulated LIG annual average global
temperature is not notably higher than present, consistent with
the orbital forcing.” (6.4.1.6). Despite that, yes, sea level was 4 – 6m higher, which can probably not be accounted for by orbitally-induced melting of the Greenland ice sheet alone (this is not comforting!).
But anyway, I agree that the climate has changed significantly in the past in response to forcings. I’m not sure why that makes the changing of forcings ok!
If they were growing vines then it could be warmer than today.
Hmm… do we do the vines debate now? They’re growing some very decent wine here in the UK now, and we don’t drink it just because the water isn’t safe…. 😉
From your earlier post:
If they suffer a similar high temperature diverge during the MWP…
Well, sure. But two points, i) it is not all chronologies that have shown a post-1960s decline, and those that don’t correlate fairly well with those that do during the MWP and ii) we have, as you well know, reconstructions excluding all tree rings (which will, of course, have their own issues).
We end up aware of uncertainties. What we don’t end up with, it seems to me, is any restored certainty in the global or even just hemispheric warmth of the MWP, since there never was any such certainty in the first place. Many attack Mann’s work whilst at the same time pretending that there was some sort of general acceptance of the level of MWP warmth before 98/99. This is almost always ‘supported’ by reference to 7c in the FAR which, as I’m sure you’re well aware, was a schematic based on Lamb’s reconstruction of temperatures in Central England alone. People may have speculated about the extent of the MWP, but how can speculation without multi-proxy evidence be assumed to be a better view of past temperatures?
Jean S (Comment#26483)
Come on, you know as well as I do that the trick was used in MBH98 Figure 5b:
We’re going around in circles. You said above “UC was not claiming that “fix the CO2 in PC#1″-trick was the one what is under discussion here (Mike’s Nature trick).” But you are pointing again to what UC was linking.
As I said above, that endpoint treatment has a barely visible effect. It’s hard to imagine that it is what PJ had in mind. And there’s no indication that he used it in the WMO diagram.
bugs (Comment#26484) December 6th, 2009 at 2:48 pm
Bugs. the clearclimatecode people stopped working on the code a while back. Yes, they found some errors. At CA there were only a couple of us looking at the code. The big problem was the AIX environment. Early on some guys tried to get it compiled on Linux. There were some compiler flag issues which resulted in infinite loops and other nice things. To get it to compile guys started to change the code. Big problem there. Bascially, you can’t start your audit with a port. You need it running in its native enviroment to do the job right ( ie check the port)
Anyways, EM Smith has been working on the project in relative obscurity ( except for those of us who read him) finding bugs and stuff. At some point He and I will get together, he lives just down the road, purchase a workstation so we can get it up and running in a native environment and then do a proper I V&V. Maybe invite dan hughes and others in at that point.
You see my evil plan is to destroy CRU so that people have to rely on GISSTEMP and just when they do that… BAMM we pull that rug out from underneath them… Then they run to use the NOAA index and BAMM we hit NOAA with the FOIA and I get all the code for doing USHCN v2 and we go after the 3 historical sources for that data. BAMM. its turtles all the way down brother.
Opps I shouldn’t telegraph my evil plot. But there you have it.
Take down CRU, then everybody has to play on the GISSTEMP battlefield where EM smith has been documenting the code to death. Diabolical huh? I just made it up. or not.
Nick,
“And there’s no indication that he used it in the WMO diagram.”
and, once more, what trick he used then in the WMO diagram?
Nick Stokes:
“The above facts seem to support an inference that some slowly varying factor began to exert a very widespread negative influence on the trend of these MXD data from around the middle of the 20th century, with effects at higher frequency also becoming noticeable in some high-latitude regions. For the time being, we circumvent this problem by restricting the calibration of the density data to the period before 1960.”
If the relationship between temperature and MXD data is non-monotonic, this in itself vastly increases the uncertainty for determining temperature in the past- which is not taken into account. If some non-temperature effect can change MXD data now, why can it not do so in the past ?
Hiding the anomalous data is not “circumventing” the problem, it is HIDING the problem.
what carrick said.
per
Lucia, you need a new graphic for your mug. The decline that is being hidden is much worse than that.
Simon Evans (Comment#26486) December 6th, 2009 at 2:51 pm
Simon do I see you slouching toward the Lukewarmer position on the MWP, which is to say the CIs maybe too wide to say anything with confidence or to say something like the present warming is within 1 sd of the warmest period in the last 1000 years.
UC (Comment#26489)
You are just changing the subject. We still don’t have an answer to a point that Lucia originally raised – what “Mann Nature trick” could PJ have been referring to?
per #26490
which is not taken into account
Yes, it is. B says:
This reduces the potential overlap between temperature observations and density measurements and means that less data can be reserved for independent tests of the validity of predictive equations.
That is reflected in a greater error range. The possibility that a proxy may cease to track temp in the past, while tracking during the calibration period, is present in any proxy for any purpose, and is a real reason for caution. Observed modern divergence makes that more troublesome, as scientists are all too well aware. That is why it is so much discussed.
Steven – I’ve never thought we can (currently, at least) determine the global MWP temperature with such confidence. At the same time I have far less confidence in claims it was globally as warm or even warmer than today (not that it might have been, but that it was). I think those who claim it was, say, 3C warmer are charlatans whose dodginess exceeds anything ever charged upon the hockey team.
Either way, I think that increasing GHGs will warm the atmosphere and that we should assess the consequences of that and what, if anything, to do about it. So, I don’t think my views of confidence in multi-proxy studies makes me a luke-warmer, no. What sort of projection of likely future trends do lukewarmers consider? I’ll go along with the most probable equilibrium CS being c.3C, and that is entirely regardless of what the actual global MWP temperature was.
steven mosher (Comment#26488) December 6th, 2009 at 3:04 pm
bugs (Comment#26484) December 6th, 2009 at 2:48 pm
Bugs. the clearclimatecode people stopped working on the code a while back. Yes, they found some errors. At CA there were only a couple of us looking at the code.
I should get that engraved on a coffee mug. After all the rumpus that has been raised CA over the GISTEMP over the years, the perptetual digs made at it in the blogs still, the self righteuos indignation and demands for opening the code, freeing the data, the incomeptence of the science, what are they hiding, in the end “there were only a couple of us looking at the code”.
Simon Evans (Comment#26495)-Who exactly says that globally and on average the MWP was 3 degrees warmer than the present? No one I’ve seen. If you are talking about the Idso’s, they’ve cataloged a lot of studies on the on the MWP but they’ve never said that the figure works out to something like that. In fact, only a small number of studies show that large a difference and the median of those they look at is more like .75 degrees. I think that’s a little high, for the record and I would bank on it being closer to zero. But I don’t know anyone who says that the global average was more than a degree higher during the MWP-nobody, just nobody, says that.
Andrew_FL
Who exactly says that globally and on average the MWP was 3 degrees warmer than the present?
Ian Plimer, Professor of Mining Geology at the University of Adelaide and author of “Heaven and Earth” (which is currently selling very well. E.g. page 22 he says:
“during the MedievalWarming, the global temperature was a few degrees warmer than today.”
and page 490 he says present warming is:
“up to 3C below the Minoan, Roman and Medieval
warmings”
Or how about our friend Christopher Monckton:
Scores of scientific papers show that the mediaeval warm period was real, global and up to 3C warmer than now.
http://meteo.lcd.lu/globalwarming/Monckton/apocalypse_cancelled.pdf
So now you know of two people, at least, who say that, and they’re both having their influence upon public opinion.
Carrick (Comment#26461) December 6th, 2009 at 12:11 pm
bugs:
In your fevered imagination anyway.
No, I am taking this from the IPCC report and the peak science bodies around the planet. Not my imagination.
Nick Stokes
“That is reflected in a greater error range. The possibility that a proxy may cease to track temp in the past, while tracking during the calibration period, is present in any proxy for any purpose, and is a real reason for caution.”
truncating the period of comparison does indeed increase the uncertainty, in a way that is measurable. But it assumes a monotonic relationship with temperature; which is not true.
Including a period where the data has a completely different relation to temperature than heretofore, would increase the uncertainty enormously; so much so, that you wouldn’t even be able to do statistics on it. Your 95% confidence intervals would be laughable ! Removing the data that show that this is the case is HIDING the damaging data.
per
” Removing the data that show that this is the case is HIDING the damaging data.”
There is no damaging data. The warming during the sixties is real, it didn’t cool at all like the proxies indicated. The only issue is to what extent can we use trees as a temperature proxy. The issues are ongoing and have not yet been resolved. If we can’t use trees as a proxy, so what, other proxies, such as glaciers, indicate a ‘hockey stick’. It would be good to be able to use trees, as they can reach back much further in time.
Apart from all that, the paleo climatology part of AGW science is just part of multiple, parallel streams of research.
bugs:
There is a lot of hyperbole in the IPCC report “and peak science bodies”.
You hadn’t noticed?
If one reads them carefully and filters the BS, the take home message is things aren’t nearly as grave as the Rodney Dangerfield of Climate Science (ManBearPig aka Al Gore) would have us believe. Yes there are things that need to be addressed, but CO2 control isn’t a magic bullet for end the negative environmental effects of human activity. It is true and will remain so for an indefinite time that the main effect of human activity on the environment comes from the sum of local environmental impacts on the globe, rather than global variations impacting local environments.
In fact, as they are approaching it in Copehagen, it’s pretty much a complete wash. Even James Hansen agrees with that.
BUGS
“The issues are ongoing and have not yet been resolved. If we can’t use trees as a proxy, so what, other proxies, such as glaciers, indicate a ‘hockey stick’. It would be good to be able to use trees, as they can reach back much further in time.”
A HOCKEY STICK has a FLAT SHAFT ( no MWP) and a curved blade.
Let me break it down for you: In the instrument period we have a CURVED BLADE. Nobody in the Lukewarmer camp argues that instrument series from 1850 is flat. Warmists say its gone up about 1C or so, Luke warmers question the accuracy of that. You want me to put a number on it, I’d guess the blade is too steep by .1C or so. WRT to proxies we expect to see proxies track this blade. When we dont see them track the blade ( correlate) then we got issues.. partly with the steepness of the blade and partly with the validity of the proxy. To be sorted out….
When it comes to SHAFTS the issue is:
A. Have proxies been cherry picked for flat shafts?
B. Have confidence intervals been properly calculated?
C. Its the reconstruction spatially robust.
On the issues of blades and proxies some recent papers have even mentioned that errors in the instrument series may cause the lack of correlation. The point being like you say reconstructions are not settled science. There are right ways and wrong ways to reduce the uncertainty in reconstructions. Lopping off data from a proxy is not confidence inspiring. Flipping series is not confidence inspiring. Hiding raw data is not confidence inspiring.
Not updating proxies is not confidence inspiring. Peer review by your buddies is not confidence inspiring. poor Graphsmanship is not confidence inspiring. 40% of people dont trust climate scientists. All I can suggest to improve trust is to improve transparency. For me a conclusion that is open that asserts 67% confidence is MORE trustworthy than a solution that uses closed data and methods. I’ll give you a dumb way to look at it.
You tell me youre 80% confident in your conclusion, but you wont share your data and methods. Hmm in my book I discount that by the Probability youre bullshitting = 50% so I got 40% confidence in your result ( YMMV)
bugs (Comment#26497) December 6th, 2009 at 4:28 pm
Dude those of us looking at the code posting our progress on the blog. You can go back and read. The biggest issue was the Fortran, as most of us have moved on from fortran ( from scientific programming to commercial world) Then JohnV did his open temp so it got put on the backburner. Unlike CRU we dont get funded by Shell Oil so its a purely volunteer effort. But after sticking needles in my eyes for a few weeks I could report back the following. There was no huge smoking gun problem. Not that we expected one. In fact I told this to gavin before we even got the code. The big issues remain what we know from reading the papers:
How stations get combined ( GUESS WHAT? jones has the same problem with the giss approach, which is actually a variant of easterlings approach) and how rural stations get assessed. The other reason why we are not chomping at the bit is the need to get surface stations done as that plays a big role in the final assesment.
I am a lurker. I come out just to say that since I have started reading this blog last spring I have grown to respect it (and its host) very much. Please keep up your standards, you are doing a great public service.
Simon Evans (Comment#26501)-Technically neither of them quite say what you say they do, as far as I can tell. I don’t really pay much attention to Monckton, though.
Looking closely at Monckton’s claim the problem with it is that he is explicitly-and yes, disingenuously, but not as much as you suggest-stating the greatest values (up to being the operative words!). It is literally true, although misleading, to state that some studies of local proxies show that big of a difference. Remember exactly what I said:
“Who exactly says that globally and on average the MWP was 3 degrees warmer than the present?”
As for Plimer, I plead ignorance. Who? No need because I don’t care.
Simon,
We had some discussions about what levels of sensitivity qualifies one as a lukewarmer. Let me put it this way. The way things are currently assesed is via a “democracy” of models approach which is politically driven, you make a model it gets included. I’ve agreud before ( and ray and gavin did not throw fruit at me) that the IPCC should select models that perform BEST. They certainly use a selection of models when they do attribution studies ( only use models that dont have a lot of drift) so I’d feel much better about picking a sensitivity number after looking at the results of the models that had the best skill at hindcasting an accurate historical record.
Slippery enough for you?
“A HOCKEY STICK has a FLAT SHAFT ( no MWP) and a curved blade.
Let me break it down for you: In the instrument period we have a CURVED BLADE. Nobody in the Lukewarmer camp argues that instrument series from 1850 is flat. Warmists say its gone up about 1C or so, Luke warmers question the accuracy of that. You want me to put a number on it, I’d guess the blade is too steep by .1C or so. WRT to proxies we expect to see proxies track this blade. When we dont see them track the blade ( correlate) then we got issues.. partly with the steepness of the blade and partly with the validity of the proxy. To be sorted out….
When it comes to SHAFTS the issue is:
A. Have proxies been cherry picked for flat shafts?
B. Have confidence intervals been properly calculated?
C. Its the reconstruction spatially robust.”
The first time I saw the error bounds I wondered just how much use was the reconstruction. The ‘flat’ part hasn’t existed from the start, it’s more like a shaft made out of absorbant cotton. The latest reconstruction shows much more of what I expect it should look like. The idea that the shaft should be perfectly straight has never occured to me.
Scooter (Comment#26345)
A left handed mug would allow the thumb to conveniently rest over the most recent data. How about it Lucia?
“steven mosher”
RE: native code compilation.
I think reading through HARRY I can see problems where even he wasn’t sure he wasn’t seeing compiler artifacts in his programs when comparing them to previous CRUT output. It looks like for some of it he was using the GNU g77. This comment stands out, though:
So if someone who is working there and has free access to everything available can’t get output that matches, it seems doubtful that an outsider will be able to do so.
Wonder what compiler was used for the f90 progs.
Crosspatch.
I was refering to the GISSTEMP recompile. There the code had AIX dependencies and compiler dependencies. Some of the check for exiting loops were compiler flag dependent. Shit I had to figure out just by guessing while other guys would try to compile the stuff. ” Suffice it to say there were no porting guides. Which is why they are reluctant to supply code. It wasnt written to be read by others. hell, GISSTEMP has both PYTHON AND FORTRAN. they hired a new kid from someplace in france that was doing all their code in Python. good christ. EM smith has made some headway.. he’s looking to get an old sparc station so he doesnt have to handle the bigendian/littleendian issue.
WRT harry. I would not be surprised if there were compiler dependencies on the final outcome. None of these guys rig their code for test or do unit test so when you port if you cant match the answer at the global scale you probably have to go instrument the code in the native enviroment, check that your instrumentation doesnt eff up the answer. OPC other peoples code. sucks unless you like puzzles.
crosspatch (Comment#26577) December 6th, 2009 at 8:49 pm
I don’t wanna turn this into a bash fortran programmer session, but I won’t stop anyone who does. Imagine trying to take fortran programs from VMS into TSO and then into flavors of UNIX and then into the PC world and then into an RTOS world. after a while you wanna shoot people, but youre a geek so you just write nasty comments in their stupid code
Lucia:
Steven McIntyre said: “Lucia, you need a new graphic for your mug. The decline that is being hidden is much worse than that.”
Since I just ordered a mug, I hope you have the latest chart on because if the decline is hidden too well, I might not be able to see it.
steven mosher (Comment#26599)
I don’t want to bash any programmers either. I meant to point out how difficult porting something would be if they have trouble themselves. And I apologize for not getting the GISSTEMP angle, I thought you were considering getting the CRU stuff running.
By the way, Steve M has an approach that I have actually found useful myself in his rewriting stuff into R. If you try to translate a program into a different language, you can often gain insights into things and catch mistakes that would go unnoticed of you were simply trying to read through the code.
In my experience I have taken prototype code in java, for example, and had to port it to c or c++ to scale it. Completely translating it to another language does wonders for gaining an understanding of exactly what is going on as one must go through each line of code. I’ve found a few really doozies in my time by doing that including one case where I wondered how it had even run at all.
Crosspatch.
I agree 100% the first piece of scientific code I inherited was written in rocky mountain basic ( think basic with fortran flavoring)
Anyways, only by writing it in another language ( pascal– dont ask) was I able to understand the code. But then I had working code in the orginal language to always go back and step through. Same with fortran to c conversions.
Evil me, trying to hide the truth from Nick. RC:
http://www.realclimate.org/index.php/archives/2009/11/the-cru-hack/
“Phil Jones in discussing the presentation of temperature reconstructions stated that “I’ve just completed Mike’s Nature trick of adding in the real temps to each series for the last 20 years (ie from 1981 onwards) and from 1961 for Keith’s to hide the decline.†The paper in question is the Mann, Bradley and Hughes (1998) Nature paper on the original multiproxy temperature reconstruction, and the ‘trick’ is just to plot the instrumental records along with reconstruction so that the context of the recent warming is clear. ”
I.e., according to RC here’s the Mann’s trick. “plot the instrumental along with reconstruction”.
And later, in reply to comment #167, gavin clarifies:
“This has nothing to do with Mann’s Nature article. The 50-year smooth in figure 5b is only of the reconstruction, not the instrumental data.”
So, the RC-version of the trick is clearly to add a separate instrumental series to the graph. Did PJ do such trick in the WMO cover?
UC (Comment#26681)
I.e., according to RC here’s the Mann’s trick. “plot the instrumental along with reconstructionâ€.
And later, in reply to comment #167, gavin clarifies:
“This has nothing to do with Mann’s Nature article. The 50-year smooth in figure 5b is only of the reconstruction, not the instrumental data.â€
So, the RC-version of the trick is clearly to add a separate instrumental series to the graph. Did PJ do such trick in the WMO cover?
Yes, I agree with that RC version. And I’ve been saying that it isn’t the “Nature trick”. But it is in GRL, and I think PJ just mis-remembered.
PJ did do that in the draft version for the WMO cover, and I suspect that was what this email refers to. In the end the curves seem to have been merged in WMO1999.
Incidentally, that “Mann’s trick” seems to have caught on. Moberg, for example, did the same. It’s a fairly logical thing to do.
“Yes, I agree with that RC version. And I’ve been saying that it isn’t the “Nature trickâ€. But it is in GRL, and I think PJ just mis-remembered”
Separate instrumental series is in both MBH98 5b and MBH99 3a.
UC (Comment#26684)
Separate instrumental series is in both MBH98 5b and MBH99 3a.
Ah, indeed it is. I missed it in 5b because it isn’t as obvious (not so bladelike), and is called “actual data” rather than instrumental, or some such. My apologies to PJ’s memory. But that confirms my view that this is the trick that PJ refers to.
“But that confirms my view that this is the trick that PJ refers to.”
But the interesting thing is that both smooths include instrumental padding! You can take mike’s word for MBH99, and my code for MBH98,
http://www.climateaudit.org/?p=7810#comment-367353
Trick=MBHsmooths1(982:996,3); % pad with instrumental
out98t=flipud(filter(B98,A98,flipud([filter(B98,A98,[in98;Trick;zeros(100,1)])])));
out98t=out98t(26:576);
And how did PJ add separate instrumental series ‘to each series for the last 20 years’??
UC (Comment#26688)
But don’t you agree that the effect of this padding is quite small, and barely visible among the noise of the graph without magnification, as you used to show it? That’s why I don’t think that, even though it was done, that it qualifies as a “trick”.
I agree that the talk of adding 20/40 years of different data is unclear on this account. But it also is not a very natural way of describing padding.
Nick,
“But it also is not a very natural way of describing padding.”
To me it is very clear way to describe padding with instrumental. “adding in the real temps
to each series”. See also https://www.uea.ac.uk/mac/comm/media/press/2009/nov/CRUupdate “To produce temperature series that were completely up-to-date (i.e. through to 1999) it was necessary to combine the temperature reconstructions with the instrumental record, because the temperature reconstructions from proxy data ended many years earlier whereas the instrumental record is updated every month. The use of the word “trick†was not intended to imply any deception.”
And the effect is very visible if you look at the print version of MBH98. And this method can be used to hide the decline. Next.
I was wondering if Lucia was going to use a heat activated section. My wife has a nice cup that when cold the windows in the scene are dark. When you put something hot in the mug, the windows light up. This way we could have a snapshot look at “hiding the decline” while we drink hot tea or coffe and blog.
UC, no one figured out the confidence intervals for MBH9x did they?
John, AFAIK, no one has figured out CIs for MBH99, for MBH98 they are pretty straightforward. I read the documents and e-mails, and it seems that even Tim Osborne was unable to replicate them even though he had direct assistance from the mann himself!
“UC, no one figured out the confidence intervals for MBH9x did they?”
MBH98 are easy to figure out, and they are incorrect (scale factor errors are not taken into account). MBH99 CIs are still a mystery, but if one uses correct way (Brown, Sundberg) they’d be floor to ceiling.
Lucia,
Andrew Bolt is promoting your mug. You might get some orders from Aussie territory.
http://blogs.news.com.au/heraldsun/andrewbolt/index.php/heraldsun/comments/climategate_the_mug/
Did you use that heat sensitive paint so that when you pour in Hot coffee the end of the hockey stick disappears? Now, that would be funny!