Did global warming scientists use a graph upside down to make their case?
I've seen reference to Michael Mann at Climate Audit and other places saying that he used some temperature data upside-down to show that temperatures are higher than ever. When asked, Michael Mann said the claim is bizarre. The code he used is available on his website, and from the looks of it, it looks like the skeptics are right, and Mann is either a liar and fraud, or so incompetent that he used a chart upside-down. I'm referring to Mann et al 2008, apparently the CPS graph, but also the EIV has problems.
- TrevorLv 77 years agoFavorite Answer
No, global warming scientists did not use a graph upside down to make their case.
I believe the publication you’re referring to is “Proxy-based reconstructions of hemispheric and global surface temperature variations over the past two millennia” by Mann et alia and published in PNAS on 09 September 2008.
Is this the one? http://www.pnas.org/content/105/36/13252.full.pdf
Upon examination of one of the graphs by mathematician Stephen McIntyre and economist Ross McKitrick, they claimed that some of the data from one of the datasets had been plotted on inverted axes. Mann responded by saying that this was not the case.
As a mathematician McIntyre should have no problem identifying such an error, he is probably better qualified to do so than Mann. The problem I have is that McIntyre is not qualified or experienced enough to interpret the data, so whilst it may be erroneous, we have to be careful of any conclusions that McIntyre and McKitrick draw. Indeed, to their credit they didn’t draw any conclusions other than to identify the alleged error, the conclusions come from people even less qualified to comment – users of the blogosphere.
If the error is correct, and I suspect that it is, what does it actually mean besides the fact that Mann and his colleagues should have been more careful?
What we’re actually talking about is a tiny part of the overall data. Specifically in question are the data obtained from the sedimentary record taken from the bed of Lake Korttajarvi in Finland via the isotopic composition of C and H found in two sediment cores and restricted to the period AD980 to AD1250.
These cores were two from many that went into creating a proxy, there were over 1,200 proxies in the final reconstruction. We’re talking somewhere in the order of one twenty-thousandth of the total data.
Putting it into context means that the mistake is miniscule and the notion that the whole graph ends up inverted is a nonsense.
- Anonymous7 years ago
It was used upside-down. I would vote Mann is incompetent, simply because he provided all his code for others to detect the mistake, whereas if he were fraudulent, he would have kept it hidden. Looking thru the William Connolley and Arthur Smith posts discussed above, Arthur Smith provided a definition of fraud, that I would say Mann meets all the criteria, but nevertheless I will vote incompetent.
Mann reused the graph upside down after having it pointed out. It made a material difference to the result, which is not what is claimed by the others here. Though there are 1209 proxies, there is a graph on Michael Mann's website which shows the robustness of the graph if you exclude these proxies as well as tree rings, at which point current temperatures and temperatures in the past are similar, thus no longer a hockey stick. When the paper was released, the big deal was that Mann demonstrated a hockey stick without tree rings, and then it turns out to be based on just a few proxies, some of which he had to use upside-down to get them.
- corlLv 43 years ago
worldwide warming does not communicate with the temperature outdoors on the very 2d, that's a assessment of commonplace temperatures. constructive it snowed at present, it additionally has snowed each 3 hundred and sixty 5 days ever simply by fact the Earth has had water present day. additionally, possibly seem and notice that this previous summer season interior the Midwest become between the main well liked on record, and the place I stay, it did no longer snow as quickly as till December the place for the time of previous years it snowed as early as October. Kirby is likewise the guy who makes a b.s. argument approximately worldwide warming and then an hour later posts a yahoo question "Are worldwide Warming and climate exchange an identical ingredient?" No they do no longer look to be an identical ingredient, even nevertheless they're appropriate. climate exchange is the extra scientifically known term and takes into consideration specific factors of the Earth that have genuinely cooled somewhat. although the finished Earth's commonplace temperature is in a warming sort that has been shown to have been inspired via an anthropogenic result.
- BaccheusLv 77 years ago
Mann et al was a multi-proxy temperature reconstruction, that is it looked at many different ways to measure past temperatures. The proxies showed consistent history and was consistent with other research.
Steve McIntyre at Climate Audit and other amateur bloggers who do zero scientific research had a problem with one particular set of proxies, measurements from sediments from two lakes in Finland, the Tiljander proxies.
Yet other bloggers looked at the data and concluded that McIntyre doesn't know how to read the data. Mann's response was that McIntyre doesn't understand how that statistical data was used and hence the statement that the complaint is bizarre. The science journals and other actual researchers have accepted Mann el al's calculations as accurate.
Even if there is a problem with the Tiljander proxies -- and I certainly am not qualified to analyzeor judge the data or statistical processes -- they are a small part of the effect calculated by Mann et al, and make no difference to the conclusions. In fact, if you remove the Tiljander proxies then the maximum temperatures of the Medieval Warming period get colder and the difference between then and now is even greater greater. If the Tiljander proxies are reversed, the data would at least in some ways confirm that the MIA was even colder than Mann et al calculated.
What you are getting on Climate Audit is picking at a very small detail -- which is probably correct anyhow -- in a deliberate attempt to distract you from the big picture. Every analysis continues to show that current times are unusually warm. The only people who try to dispute that are people who don't do any research.
- How do you think about the answers? You can sign in to vote the answer.
- MikeLv 77 years ago
Yes, he did. Others have attempted to argue that the skeptics are wrong on this, including William Connolley, a global warming editor at Wikipedia, and Arthur Smith, a Mann defender at Amazon, and both ended up unable to make the case in the face of overwhelming evidence that Mann used things upside-down. He then followed through and did not fix the error for a paper the next year, even though the error had been pointed out to him. The EIV has a separate problem that the graph in itself is only valid up to a certain point in time, after which it was contaminated by farming, according to the author of the graph. After this point, it starts to show 'cooling'. So in both cases, it was used upside down, but in the case of CVS, explicitly so. Michael Mann's response to Steve McIntyre's comment implied that he did not use things upside down, or even that it would be impossible to do so, and confused a number of scientists into believing the complaint was bogus because a regression algorithm would automatically fix a chart that was put in upside down to be right side up. However, this wasn't the case with these proxies, called Tiljander.Source(s): http://amac1.blogspot.com/
- Anonymous7 years ago
The question has a clear technical answer, but it's necessary to go over some details first. In their paper "Proxy-based reconstructions of hemispheric and global surface temperature variations over the past two millennia" (PNAS, 2008), Mann et al. claimed that their "EIV" and "CPS" methods could reconstruct the average temperature of Earth for the past 2,000 years, using two different types of data: tree-rings, and records of other phenomena.
The paper's graphs all showed versions of the famous "hockey stick," with a sharply-rising trend beginning c. 1970 and rising rapidly to unprecedented levels.
Mann08 received favorable exposure at many climate-themed online sites, but "skeptics" quickly found flaws. The broad issue of faults with Mann08's data and methods is beyond the scope of this Answer.
*One* of the charges made by Steven McIntyre of the "Climate Audit" blog was that Mann08 used certain non-tree-ring data as a temperature proxy in a very inappropriate way. This included four series of information about sediments cored from the bed of Lake Korttajarvi in Finland. The first author of the 2003 paper describing them was grad student Mia Tiljander, so the records are known as the "Tiljander proxies." These sediments are "varved" -- there are regular seasonal variations in qualities of the silt that settled to the bottom of the lake, so scientists can count back year-by-year for about 5,000 years in this record.
These were the four data values archived for each year:
* Thickness -- How thick that year's varve was. Tiljander03 didn't comment on Thickness' relationship to temperature.
* Lightsum -- How thick the mineral fraction of the varve would be, if there was no organic material present. Tiljander03 proposed that higher values correspond to cooler and wetter winters
* Darksum -- How thick the organic fraction of the varve would be, if there was no mineral material present. Tiljander03 proposed that higher values correspond to warmer and wetter summers.
* XRD -- The X-Ray Density of the varve. Tiljander03 proposed that higher values correspond to cooler periods.
Among McIntyre's complaints was that Mann08's methods were grossly defective -- with one example being that both the CPS and EIV methods "flipped" the Tiljander proxies. The editors of PNAS allowed a few of McIntyre's key claims into print in February 2009: "[Mann08's non-tree-ring] network uses some data with the axes upside down, e.g., Korttajarvi sediments, which are also compromised by agricultural impact".
Mann08's authors riposted, "The claim that 'upside down' data were used is bizarre. Multivariate regression methods are insensitive to the sign of predictors. Screening, when used, employed one-sided tests only when a definite sign could be a priori reasoned on physical grounds. Potential nonclimatic influences on the Tiljander and other proxies were discussed in the [Supplemental Information], which showed that none of our central conclusions relied on their use."
Now -- finally -- there's sufficient background to give an informative answer to the original question!
* McIntyre claimed that Mann08 used certain Tiljander proxies with the Y-axis inverted (upside-down); Mann08's authors denied it. McIntyre is correct and Mann et al. are wrong: Mann08 *did* use Lightsum and XRD upside down, so that these two series' meanings were the opposite of those proposed in Tiljander03.
* The data-handling mistake in CPS is different from the one present in EIV, but the results -- flipped-over Lightsum and XRD -- are the same for each process.
Three relevant follow-on points:
* Mann08 used Darksum in the orientation that is consistent with the Tiljander03 interpretation. As the 2003 paper offered no interpretation as to the relationship between Thickness and temperature, "flipped" has no meaning for this proxy.
* Tiljander03 cautioned that the Lake Korttajarvi sediments couldn't be used after about 1720, when farming and then other human-related activities overwhelmed any climate signals. This is evident upon inspection (see Source #3). Mann08 discussed Tiljander's warnings, then proceeded to employ the proxies. Mann08's methods (both CPS and EIV) depend absolutely upon the meaningful direct calibration of every proxy to the instrumental record -- and this is *impossible* for the Tiljander proxies.
* Mann08's defenders claim that the paper's conclusions are unaffected by the inclusion or exclusion of the Tiljander proxies, i.e. that they don't matter. This is incorrect. Key conclusions of the paper are invalidated by removal of the Tiljander proxies from the non-tree-ring data set (see Source #4).
.Source(s): #1 Peer-reviewed paper links -- http://amac1.blogspot.com/2009/11/primary-links-ma... #2 Blog post links -- http://amac1.blogspot.com/2009/11/blog-links-mann-... #3 Inspecting graphs of the Tiljander proxies -- http://amac1.blogspot.com/2010/08/tiljander-data-s... #4 Mann08 conclusions depend on Tiljander -- http://climateaudit.org/2011/07/06/dirty-laundry-i...
- JimZLv 77 years ago
It has been a few years since I read the details but I think what you may be referring to is some trees showed a reverse correlation (e.g. had smaller rings and less density). These were still used by Mann and the justification was that they showed correlation or something to that effect. They essentially turned the graph upside down for these proxies. I don't know if that is common practice. Mann kept his code secret for quite some time and McIntyre et al had a heck of time getting it and when he did they found all sorts of problems and Mann was forced to redo it. I feel the thumbs down heading my way because alarmists don't like Mann exposed.
- Elmer98Lv 77 years ago
I'd take climate audit with a lot more skepticism.
all these graphs and emails are making the glaciers melt, but then again NASA has Photoshopped the moon landing too.
- Anonymous7 years ago
They won't be published if they try.
- 7 years ago
You know who the liar is and it's you