Guest post by Nancy Green
Tamino claims he has added 3 spikes to the Marcott et al proxy data and the Marcott et al process detects them.
Source: http://tamino.files.wordpress.com/2013/04/many_vs_unpert.jpg
This, he then proposes, is proof that there are no 20th century spikes in the Holocene. This claim appears to run counter to a prediction I made recently in a WUWT post; that as you increase the proxy resolution you are more likely to find spikes.
See:
http://wattsupwiththat.com/2013/04/03/proxy-spikes-the-missed-message-in-marcott-et-al/
Having had my reply disappeared at Tamino’s site, I thought readers at WUWT might be interested. I don’t believe Tamino’s conclusion follows from his results. Rather, I believe he has demonstrated the truth of my original prediction. What needs to be understood is that adding a spike to the proxy data is not the same as adding a spike to the proxies. This is where people get confused.
The proxies are ocean cores or similar sitting in some repository. They are real, physical objects. To truly add a spike to the proxies you would need to travel back in time and change the temperature of the earth. This would then affect the proxies in some fashion, depending on the resolution of the proxies, how they respond regionally, including lags, gain or damping. The proxy response might also be affected by other unknown factors at the time that are not visible in the proxies. In other words, the spikes that you add to the proxies would have all the resolution problems that the proxies themselves have.
However, adding spikes to the proxy data is an entirely different animal. The proxy data is an abstract representation of the proxy. It is numbers drawn on a sheet of paper or electronic equivalent. Now you are adding (drawing) high resolution spikes onto low resolution proxy data, with no accounting for regional affects, lag, gain, damping or confounding factors. It should be no surprise at all that these high resolution spikes jump out. If they didn’t, it would point to a serious flaw in Marcott et al.
An analogy might help better understand the problem. Imagine for a moment that we are not dealing with temperature, but rather trying to detect planets around stars. We have before us a photograph of a star taken by a telescope on Earth. We look at this under the microscope. However, we find no planets because the telescope lacks the angular resolution to distinguish them from the star itself.
Now let’s go out to the star in question and add planets around the star and take more photos with our telescope. These planets are real objects. We know they exists. However, it will make no difference; we still can’t see the planets with our telescope. In this example we have added a spike to the actual proxy and it has made no difference.
Now let’s add a spike to the proxy data. Instead of placing planets around the star, take the photo from the telescope and draw a picture of a planet on it. This is an example of adding a spike to the proxy data. The photo is an abstract representation of the star and its planets, equivalent to the proxy data. Now examine the photo under a microscope and voila, the planet (spike) will now be visible.
What we are seeing in action is actually a form or misdirection used in stage magic. It fools us on the stage just as it does in science. It is our minds that create the confusion (illusion) between what the proxies actually are and what the proxy data actually is. The proxies are ocean cores – they are real objects. The proxy data is an abstract representation of the real object. However in our minds we are so used to dealing with real objects as abstract representations that we are fooled into thinking they are one and the same.
If anything, what Tamino has actually done is to prove the point of my original article. He has added high resolution spikes to the low resolution data and as predicted they are detectable. To conclude however that this somehow proves there are no 20th century type spikes in the Holocene makes no sense. As we have seen in this example, no matter how many planets you physically add around a star it makes no difference if you lack the resolution to detect them. This is no proof that they don’t exist. It is only after you examine them at sufficiently high resolution that they become visible.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
Ask Tamino why the proxies don’t pick up the spike in the 20th century in Marcott’s PHD thesis. Without the redating to create the ‘scythe’ the proxies can’t find the current spike so how would they detect past spikes?
Leo Geiger says:
April 8, 2013 at 8:59 am
———
Maybe there’s a dichotomy between the primary conclusions of the paper and why the paper is important and interesting press wise. If Marcott et al doesn’t speak to the absence of spikes in the past, then the claims about the uniqueness of the current temperature spike lose their basis. Maybe the primary conclusions stand; the ‘popular’ conclusions don’t.
Ryan:
I am replying to your post at April 8, 2013 at 5:53 am.
I am writing this reply on the assumption that your questions are genuine. However, I state that your iteration on WUWT of your questions without reference to the full answers you have been given to those questions on WUWTT implies that your post is trolling.
Your questions were answered fully and in detail to you when you and Thomas raised those issues on the previous thread where Nancy Green first provided her analogy (prior to the above refinement).
That thread is at
http://wattsupwiththat.com/2013/04/03/proxy-spikes-the-missed-message-in-marcott-et-al/
Answering your questions would require copying and pasting the answers you were given to those questions, and that would be wasted effort because you have failed to understand those answers.
So, I ask you to read that thread again and to itemise the issues you failed to understand. Having done that then you can post the matters you have failed to understand to here for additional explanation.
Richard
Good point Gary; and I see that one or more above agree with you; I would also like to point out that one should not take the absence of posts specifically indicating agreement (or disagreement) as _not_ reaching more than just a few folk. Many like myself will ‘take in’ such an observation and mentally file it away for future reference.
.
Mark Bofil says (similar to markx):
April 8, 2013 at 9:37 am
“Uniqueness” is too strong a word. What the Marcott reconstruction provides is context:
It can’t be said if this is unique based solely on Marcott, but we have clearly gone from near the bottom to near the top of the 11,000 year range in a short amount of time. It becomes one more piece added to the puzzle, and helps us understand the broader picture being built up. It is a shame that this is getting lost in the eagerness by some to tear Marcott down.
I imagine the question of ‘uniqueness’ gets into what higher resolution proxies (with more limited time spans or spatial coverage) can say directly, statistical estimates of the probabilities of ‘spikes’ based on signals in smoothed proxies, limits on plausible natural physical mechanisms for temperature changes, and other things which I don’t know enough about to comment further on in any detail without doing some careful reading.
Leo Geiger
And we went from near the coldest to near the warmest in a few decades from 1690 to 1710
And near the warmest to near the coldest from 1560 to 1590
And near the warmest to near the coldest from 1220 to 1250
And….but you get the picture, there are numerous drastic swings in temperature which can be picked up in climate observations that may not be seen in the low resolution proxies that such as marcott use.
Tonyb
@Jit 1:33 am
I agree that the proxies grow in the shallower, photic zones. But just what is the mean and std dev of the distribution of descent times to the ocean floor 3 km down of a proxy bug? The mixing of waters in the deep is surprising. Just how broad is the atomic testing spike?
Furthermore, as you point out, the temperature that is recorded by the proxy is in large measure weighted toward the conditions of the maximum growth period rather than an unbiased average temperature throughout the year. The maximum growth period will be a function of fertile conditions based upon non-temperature factors such as nutrients from run-off and floods or even schooling of fish.
Finally, I wanted to point out that these proxies were a mixed bag of ultra-deep water, Shelf, photic, lacustrine, and high-Antarctic desert plateaus. Each one of them trying to measure a 1 part in 100 anomalous temperature of the surface air. I’m skeptical of the 8000 year temperature decline much less the 100 year spike.
climatereason says: April 8, 2013 at 11:02 am
Which proxy reconstruction stacks show decadal global mean temperature changes of similar magnitude that are reliable enough after accounting for errors? Or to borrow Nancy Green’s analogy, which telescope is resolving those planets?
How about the 8.2k event in the Marcott data?
http://www.geo.arizona.edu/palynology/geos462/8200yrevent.html
It should have been a 3°C sudden difference much above the modern warming swing?
Leo Geiger:
It seems that you have joined this discussion late.
Nancy Green first raised her analogy for a previous thread where the limitations of the paper by Marcott et al. were discussed. This thread is about Foster’s incompetent assessment of the paper by Macott et al..
I respectfully suggest that you read the previous thread. It is at
http://wattsupwiththat.com/2013/04/03/proxy-spikes-the-missed-message-in-marcott-et-al/
The main point you seem to not understand is that
absence of evidence is not evidence of absence.
Also, ‘one black swan demonstrates that not all swans are white’.
As Lars P. says at April 8, 2013 at 12:54 pm, we know of a ‘black swan’ that Marcott et al. fails to detect.
I commend you to read the previous thread because I see nothing in your posts which was not covered in the previous thread.
Richard
I have read Green’s earlier post. The problem with argument by analogy is analogies are imperfect and they can be pushed too far. This
is doing that. A “spike” in the 20th century, by itself, tells you nothing about whether or not more would be expected earlier.
The point is that Marcott is able to make a reliable statement based on a temperature distribution that has been widened to account for lost high frequencies. Using that approach is relatively insensitive to whether or not there are individual spikes. I realize this thread is about Foster, but none of the threads seem to acknowledge the actual approach Marcott used.
It is hard to keep up with the quantity of posts being churned out here. They have the shelf life of a souffle left in the sun on a hot summer day.
You seem unable to understand that the three spikes shown in the graph at the head of this post are “fakes”, deliberately introduced by Tamino in his bumbling attempt to prove the proxy records of Marcott are capable of resolving 20th century-type temperature excursions. There is nothing to explain.
Leo Geiger says: April 8, 2013 at 10:50 am
“….we have clearly gone from near the bottom to near the top of the 11,000 year range in a short amount of time….”
Leo, how can we possibly know that from Marcott?
This is the point of the whole debate. Do Marcott’s smoothed and smeared and then averaged proxies show the range? In a further 10,000 years the 20th century uptick very well may be invisible in similar proxies.
Lars P. — Thank you for the link to the 8200 year event. I realize I am not explaining this clearly, since it does not seem to be understood that I agree Marcott’s reconstruction would not fully resolve such an event. That is what Marcott said too. I am pointing out that the method they used to make their concluding statements takes into account their inability to resolve this type of event. Green and others posting here never seem acknowledge this, always giving the impression that Marcott used the unmodified low resolution signal for their statements.
As an aside, I see from that link there is a physical mechanism to explain that particular event:
Obviously something like that isn’t happening all the time. It certainly doesn’t make for a very good ‘natural’ explanation for the modern “spike”…
Regardless, the presence of unresolved spikes is accounted for in their widened temperature distribution histogram.
markx says:
Look at Marcott Figure 3. Do you understand why there is a distribution with a dashed black line and another with a solid black line, what the difference between the two is, and which one Marcott used to make their statements?
Leo Geiger says:
April 8, 2013 at 3:37 pm
A “spike” in the 20th century, by itself, tells you nothing about whether or not more would be expected earlier. The point is that Marcott is able to make a reliable statement
==========
The spike is not “by itself”. I have already explained why the circumstances surrounding the spike allows one to made predictions. Tamino has demonstrated that my prediction is correct.
Marcott has used an unproven statistical method without citation. The divergence between his thesis and later paper in Science suggests the results are anything but reliable. In this regard I am in strong agreement with the Dr Pielke.
http://rogerpielkejr.blogspot.com/2013/03/fixing-marcott-mess-in-climate-science.html
There have been at least three well known warm spikes in the earth’s very recent past; the Minoan Warm Period, the Roman Warm Period and the Medieval Warm Period.
No one knows the cause of these warming and there is no indication they are substantially different than the Modern Warm Period. As with our current time, they represent a period of food abundance and relative prosperity for the average person. These warm period were well documented by previous generations of researchers.
note: Contrast this to the miserable existence of humans 90k years ago, where the genetic record shows we were the endangered species of the day, Things got so bad that polar bears were out taking donations to “save the humans”,
However, the recent generation of researchers has sought to minimize these warm periods and the prosperity they brought. Instead they have claimed these warming spikes were regional, or not significant. However, I submit to my readers that it is a simple matter to test the accuracy of these claims for yourself.
Ask yourself this simple question. What would happen if a proxy was found that showed that the previous proxies had been wrong, and CO2 levels were not stable during those times? Rather that CO2 had increased rapidly around the time of these past warming spikes – perhaps due to a great belch of CO2 released every 1500 years or so by the deep oceans.
Would climate science continue to claim that the these warming spikes didn’t exist? Or would climate science overnight change its story? That in fact these past spikes were global and in combination with the CO2 record were certain proof that CO2 was driving climate?
What would Climate Science do?
There might also be a dampening of ancient data by osmosis of the various chemical markers to layers of surrounding ages. Chemicals constantly move toward equilibrium.
If I understood the original paper correctly; A single proxies datum of information, represented a climate temperature smeered over a 300 year period.
So adding a spike to the proxie data and testing their algorithm, does nothing to show how sensitive the original proxie data is to detect spikes.
When a single proxie datum represents 300 years of weather at a location, there will be no spikes in the data. It is self-smoothing by vertue of what it is.
Nancy Green asked: “Evidently medical students don’t take much math as part of their curriculum. Might Climate Science suffer from the same problem?”
CRU climatologists do have a math teacher cum aide, whom I’ve met. At least, I presume this person is still on the staff – by tacit mutual agreement we have not mentioned the job since Climategate reared its head; our friendship is in another context.
Nancy Green:
You said this in your earlier post:
Fine, except they did not use the low resolution signal to make their concluding statements (warmer than 72%…cooler than 95%…). But that is what is being implied over and over again in posts like yours. They modified their low resolution distribution into a wider one to account for the lost resolution.
I suppose the next step, after acknowledging this, is then to suggest they didn’t widen it properly, which I see you have now done:
But at least that implicitly recognizes they weren’t using the low resolution signal directly (I hope). I am not sure what criteria you have used to decide that their method is unproven and without citation, since it is explained over multiple pages, including various tests, all with citation, in the supplementary material. Do you have a specific issue with what they did? That would be more interesting to hear than a casual dismissal.
That depends entirely on understanding why there are ‘divergences’. Yes, there was a problem with how the ends of the proxies were handled in the stacking, creating that exaggerated proxy uptick at the end. No, that can not be an issue in the rest of the reconstruction where most records are not ending. Nor does the uptick change the overall distribution. Bootstrap arguments of the form “if we can’t trust that, how can we trust anything” are inherently weak.
Simple? I would think testing this is a not so simple matter of doing a careful analysis of the individual proxies, the methods used to convert them to global means, and the errors present in that procedure right from data collection at the start to stacking into global means at the finish. I don’t think rhetorically asking about some scenario and presuming cynical flip-flops yields any real answers.
Oddly enough, I don’t actually disagree that the general media coverage could have been done much better on this one. But that applies to what this site is doing too.
Leo Geiger says:
April 8, 2013 at 9:39 pm
Fine, except they did not use the low resolution signal to make their concluding statements (warmer than 72%…cooler than 95%…).
============
It is a mathematical nonsense to try and compare a high resolution spike against the low resolution average. It makes the conclusion extremely misleading.
In this regard I am in strong agreement with the Dr Pielke.
http://rogerpielkejr.blogspot.com/2013/03/fixing-marcott-mess-in-climate-science.html
If anything, what Marcott shows is that we are at risk of slipping into an ice age and this risk is accelerating.
Leo Geiger says:
April 8, 2013 at 5:11 pm
Lars P. — Thank you for the link to the 8200 year event. I realize I am not explaining this clearly, since it does not seem to be understood that I agree Marcott’s reconstruction would not fully resolve such an event. That is what Marcott said too.
I am not sure I understand what do you understand under partially resolved? Is it 0.05%? It should be a 3°C that goes way out of the scale.
I am pointing out that the method they used to make their concluding statements takes into account their inability to resolve this type of event. Green and others posting here never seem acknowledge this, always giving the impression that Marcott used the unmodified low resolution signal for their statements.
From the above it is clear that a 200-300 years excursion are not captured by their method, however the chart is justified to defend the hockey stick:
http://pressrepublican.com/0205_columns/x1916524832/Climate-change-A-clear-and-present-danger
“Both the new Marcott data and the Mann data, where they overlap on the right, agree very well. Note that the temperatures today are higher than at any time in at least the past 4,000 or more years”
As an aside, I see from that link there is a physical mechanism to explain that particular event:
…
Obviously something like that isn’t happening all the time. It certainly doesn’t make for a very good ‘natural’ explanation for the modern “spike”…
How that event happened is irrelevant to this conversation.
There are “could have produced” in the explanation that you embrace so gladly which may very well be wrong. It is the explanation that current models can give, but we know that these models cannot reproduce even the warming from the first part of the century (1910-1940) which is actually ironically the warming part in Marcott’s not robust hockey stick.
As Marcott’s chart missed that excursion it is clear that it misses also the MWP and LIA which are similar in time frame but not so dramatic in amplitude, therefore any discussion about grafting actual temperature data on that chart is totally irrelevant and leads to wrong conclusions.
We know from a great deal of data that MWP and LIA were global events
http://www.co2science.org/data/mwp/mwpp.php
Leo Geiger says: “Fine, except they did not use the low resolution signal to make their concluding statements (warmer than 72%…cooler than 95%…). But that is what is being implied over and over again in posts like yours. They modified their low resolution distribution into a wider one to account for the lost resolution.”
They correcly used red noise in Fig 3, but drew Fig 1B with incorrect white noise to create the blue error bar. Using a correct error bar (e.g. like Fig 3) the 1 std dev blue area would make the 20th century rise invisible. Obviously depicting white noise there is deceptive since few people read the number and understand the context: 72% is less than 84% which is 1sd, thus making the rise and current temperatures completely normal for the Holocene. OTOH many people look at the hockey stick with the incorrect white noise 1sd shaft and incorrectly conclude “abnormal warmth compared to the last few thousand years”. Along with the problems with the final proxy data point, it is likely that the penultimate 1900-1909 point is too low due to proxy deposition and measurement problems.
Like MBH98, this paper is flawed and breaks new ground in incorrect analysis and depiction of data for purely political purposes.
eric1skeptic:
I write to support your point in your post at April 9, 2013 at 3:45 am.
And I hope to ‘clear the mist’ for those who do not understand the correct technical statements you make.
The method adopted by Marcott et al. cannot determine short-term variations in global temperature such as the recent rise in global temperature. But a ‘statistical trick’ was used to create a recent ‘up-tick’ in their graph. If that trick were valid then the graph would also have provided the ‘down’tick’ reported by Lars P. (at April 8, 2013 at 12:54 pm), but it does not.
Marcott et al. have admitted in their FAQ that the ‘up-tick’ in their graph is “not robust”; i.e. it is an artifact of their statistical trick.
Leo Geiger, Tamino and others are trying to obscure the fact of that statistical trick which can only be either gross incompetence or a deliberate attempt at deceit.
Richard
Thank you a lot for giving everyone such a superb opportunity to read critical reviews from this site. It is usually so excellent and as well , packed with a good time for me and my office colleagues to visit your blog more than 3 times weekly to see the fresh issues you have got. And of course, I’m so usually motivated considering the amazing points you give. Some two facts in this article are unequivocally the simplest we have had