Guest Post by Willis Eschenbach
Over at Bishop Hill, the Bish has an interesting thread about a new proxy reconstruction by Rob Wilson et al. entitled “Last millennium northern hemisphere summer temperatures from tree rings: Part I: The long term context” , hereinafter Wilson 2016. The paper and the associated data are available here. They describe the genesis of the work as follows:
This work is the first product of a consortium called N-TREND (Northern Hemisphere Tree-Ring Network Development) which brings together dendroclimatologists to identify a collective strategy for improving large-scale summer temperature reconstructions.
At first I was stoked that they had included an Excel spreadsheet with the proxy data. Like they say in the 12-step programs, Hi, my name’s Willis, and I’m a data addict … anyhow, here’s a graph of all of the data, along with the annual average in red.
Figure 1. Plot of the proxy data from the Wilson 2016 Excel worksheet. All proxies cover the period 1710 – 1988, as indicated by the vertical dotted lines. Note what happens to the average at the recent end.
But as always, the devil is in the details. I ran across a couple of surprises as I looked at the data.
First, I realized after looking at the data for a bit that all of the proxies had been “normalized”, that is to say, set to a mean of zero and a standard deviation of one. This is curious, because one of the selling points of their study is the following (emphasis mine):
For N-TREND, rather than statistically screening all extant TR chronologies for a significant local temperature signal, we utilise mostly published TR temperature reconstructions (or chronologies used in published reconstructions) that start prior to 1750. This strategy explicitly incorporates the expert judgement the original authors used to derive the most robust [temperature] reconstruction possible from the available data at that particular location.
So to summarize the whole process: for most of the data used, it started out as various kinds of proxies (ring width, wood density, “Blue Intensity”).
Then it was transformed using the “expert judgement of the original authors” into temperature estimates in degrees celsius.
Then it has been transformed again, this time using the expert judgement of the current authors, into standard deviations based on the mean and standard deviation of the period 1750-1950. Why this exact period? Presumably, expert judgement.
Finally, it will be re-transformed one last time, again using the expert judgement of the current authors, back into temperatures in degrees celsius
This strikes me as … well … a strangely circuitous route. I mean, if you start with proxy temperatures in degrees C and you are looking to calculate an average temperature in degrees C, why change it to something else in between?
It got odder when I analyzed the authorship of the temperature records reconstructed from the 53 proxies. In 48 of the proxy, the original lead author is also an author on this paper. It is true that they said this study is the result of a consortium of dendroclimatologists. However, I had expected them to look at more tree ring temperature reconstructions from other authors. So when they say they depend on the expert judgement of the authors of the proxies, in more than 90% of the proxies studied, they are merely saying they trust their own judgment.
And indeed, they think quite highly of their own judgement. rating it “expert”.
But since that is the case, since they are depending on their own prior transformation of a record of, e.g., tree ring width in mm into an estimated temperature in degrees C, then why on earth would they convert it out of degrees C again, and then at the end of the day convert it back into degrees C? What is the gain in that?
My second surprise came after I’d messed with the actual data for a few hours, when I got around to looking at their reconstruction. They describe their method for creating their reconstruction as follows:
3. Reconstruction methodology
A similar iterative nesting method (Meko, 1997; Cook et al., 2002), as utilised in D’Arrigo et al. (2006) and Wilson et al. (2007), was used to develop the N-TREND2015 NH temperature reconstruction. This approach involves first normalising the TR data over a common period (1750 – 1950), averaging the series to derive a mean series and iteratively removing the shorter series to allow the extension of the reconstruction back (as well as forward) in time. Each nest is then scaled to have the same mean and variance as the most replicated nest (hereafter referred to as NEST 1) and the relevant time-series sections of each nest spliced together to derive the full-length reconstruction. For each nest, separate average time series were first generated for 4 longitude quadrats (Fig. 1). These continental scale time series were then averaged (after again normalising to 1750 – 1950) to produce the final large-scale hemispheric mean to ensure it is not biased to data rich regions in any one continent. 37 backward nests and 17 forward nests were calculated to produce the full reconstruction length from 750 to 2011.
Like the song says, “Well, it was clear as mud but it covered the ground” … I was reminded of a valuable insight by Steve McIntyre, which was that at the end of the day all these different systems for combining proxies are simply setting weights for a weighted average. No matter how complex or simple they are, whether it’s principal components or 37 backwards nests and 17 forwards nests, all they can do is weight different points by different amounts. This is another such system.
In any case, that explained why they put the normalized data in their spreadsheet. This normalized data was what they used in creating their reconstruction.
I got my second surprise when I plotted up their reconstruction from the data given in their Excel worksheet. I looked at it and said “Dang, that looks like the red line in Figure 1”. So I plotted up the annual average of the 53 normalized proxies in black, and I overlaid it with a regression of their reconstruction in red. Figure 2 shows that result:
Figure 2. Annual average of 53 proxies (black), and linear regression of Wilson 2016 iterative nested reconstruction. Regression is of the form Proxy_Average = m * Reconstruction + b, with m = 1.25 and b = 0.54.
All I can say is, I hope they didn’t pay full retail price for their Nested Reconstruction Integratomasticator. Other than the final data point, their nested reconstructed integrated results are nearly identical to a simple average of the data.
Finally, as you can see, in recent times (post 1988) the fewer the proxies, the higher the estimated temperature. This is abated but not solved by their method. We can see what this means by restricting our analysis to the time period when all of the proxies have data.
Figure 3. As in Figure 2, annual average of 53 proxies (black), and linear regression of Wilson 2016 iterative nested reconstruction (red). Blue lines show proxy data.You can see how the number of proxies drops off after 1988 by the change in the intensity of the blue color.
Again you can see that their reconstruction is scarcely different from the plain old average of the data. As you can see, according to the full set of proxies the temperature in 1988 was lower than the temperature in 1950, and there is no big hockey-stick in the recent data up to 1988. After that the number of proxies drops off a cliff. By 1990, we’ve already lost about 40% of the proxies, and from there the proxy count just continues to drop.
In closing let me add that this post is far from an exhaustive analysis of difficulties facing the Wilson 2016 study. It does not touch any of the individual proxies or the problems that they might have. I hope Steve McIntyre takes on that one, he’s the undisputed king of understanding and explaining proxy minutiae. It also doesn’t address the lack of bright-line ex-ante proxy selection criteria. Nor does it discuss “data snooping”, the practice of (often unconsciously or unwittingly) selecting the proxies that will support your thesis. I can only cover so much in one post.
My conclusions from all of this:
• Transforming a dataset from tree ring widths in mm to temperatures in degrees C, thence to standard deviations, and finally back to degrees C, seems like a doubtful procedure.
• Without seeing the underlying data, it is hard to judge the full effects of what they have done. While having the normalized datasets is valuable, it cannot replace the actual underlying data.
• Whatever their iterative nested method might be doing, it’s not doing a whole lot.
• I do not know of any justification for normalizing the proxies before averaging them. They are already in degrees C. In addition, normalization greatly distorts the trends a time series, in a manner that depends on the exact shape and variance of the time series.
Finally, to a good first approximation their reconstruction is the same as the annual average of the normalized data. That means their method uses the following process:
Transform the "expert judgement" proxy temperature estimates from degrees C to units of standard deviations. Average them. Transform them back to degrees C using linear regression.
I’m sorry, but I simply don’t believe you can do that. Well, you can do it, but the result will have error bars from floor to ceiling and will have little to do with temperature.
El Niño rains here tonight. We’ve gotten four inches (10 cm) in the last four days, and it’s supposed to rain on and off for a week … great news here in drought city.
Best of life to all, sun when you need it, rain when it’s dry, silver from the five-day-old moon far-reaching on the sea …
w.
My Usual Request: If you disagree with me or anyone, please quote the exact words you disagree with. I can defend my own words. I cannot defend someone’s interpretation of my words.
My Other Request: If you think that e.g. I’m using the wrong method on the wrong dataset, please educate me and others by demonstrating the proper use of the right method on the right dataset. Simply claiming I’m wrong doesn’t advance the discussion.
Thanks Willis, as usual an excellent analysis. ” Then it was transformed using the “expert judgement of the original authors” into temperature estimates in degrees celsius.” This means the author’s opinion is passing as data. They, the authors apparently live in a virtual reality world.
Willis may I suggest an even better new word? Aproxymachination? A+proxy+machination-
machination- “a scheming or crafty action or artful design intended to accomplish some usually evil end”
I’m not convinced there’s any conscious scheming going on here. There remain, however, severe doubts concerning the validity of dendroclimatology in se., combined with deep statistical jiggeripokeriation, possibly immense error limits, and (those dread words) weighted averages.
That was my first thought upon seeing this new hockey stick. Mann got his hockey stick by weighting certain defective series by a factor, IIRR, of 350, an obscenity, though derived via such a complex statistical rigamarole that even Mann may not have realized the magnitude of it, until it was pointed out by McIntyre/McKitrick, leading to much crying ‘aieee!’ and weeping and gnashing of teeth, I’m sure.
I’m hoping that the esteemed Dr. Wilson and his consorts will be left with something robust when the dust settles. They have certainly behaved admirably in their willingness to engage at Bishop Hill, among other things.
this is exactly why the word mathturbation was created.
No, not again! Amateur stats for CAGW data dredging. Plz go back to Stats 101.
OK Willis, we already knew that one can use tree rings to tell accurately what past Temperatures were; I think you may even have told us how that is done; even with just a single tree; so long as it is scrawny enough.
Well Charlie Brown’s Christmas tree would be a perfect thermometer of the past.
But what I would like to know is:
What do think of the possibility that one might even be able to use tree rings; given enough trees, to estimate roughly what the age of a tree might be ??
Anyhow, welcome to 2016 Willis.
G
Looks like the dendros are being too clever by half…. Trying to blind us with science???
unless the samples are all the same length, combining them is generally going to generate rubbish, because some will be from warm regions and others from cold regions. the smaller the sample size, the less random the distribution, the greater the garbage.
OK, I averaged all 54 data series from 1710 to 1988. The period for which they all have data. Then I plotted the results without any statistical nonsense. Here is the result. As can be seen temperatures have been rising steadily since the early 1800’s, long before CO2 was an issue.
This is VERY STRONG evidence that CO2 is not the cause of the current warming!!
http://oi67.tinypic.com/14b6g6d.jpg
To Michael Janowski –
If you plot is from this dendro paper’s data, they are suggesting a 3-4C rise in NH temps over the period 1840-2010.
Amazingly, they have not used “It’s worse than we thought” as a rise of this magnitude would seem to be much more than any other reconstruction. Most thermometers must also require calibration.
they are suggesting a 3-4C rise in NH temps over the period 1840-2010
===================
That is not supported by the data. If you look at my graph above the range on the 30 year average (the common measure of climate) is approximately 1.2 C.
You might be able to get a 3-4C rise if you take the min of the early 1800’s and the max of the early 2000’s, but that is misleading at best, because then you are measuring weather, not climate.
correction from 1710 to 1998.
http://oi63.tinypic.com/2mpzadu.jpg
In a tree ring proxy how do the adjust for CO2 increase?
Well, you know the old saw, “If you can’t dazzle them with your brilliance, baffle them with your bullsh-t.”
Colour me stupid, but I thought the thunderingly clear lesson from the divergence problem was “Trees aren’t thermometers”. Dendrochronology makes sense to me. I really loved the book “Exodus to Arthur” by Mike Baillie, but he made very clear that trees responded to temperature, or moisture, or basically whatever they felt like responding to, and that there wasn’t one fixed rule. A signal from the trees over a wide area is a sign that *something* happened, but not an unambiguous sign of *what*, and in that book he looked to ice cores and historical records to find out what. After reading some of Douglass Keenan’s papers at http://www.informath.org/ I am obliged to take even dendrochronology somewhat more warily; even if the trees might on occasion be reporting the temperature their reports may well be ascribed to the wrong times. (Not through malice. This kind of stuff looks to be pretty hard to do right.)
Their opinions are FULL of opinions. They might be full of themselves. But they are definitely full of something…
It’s all just slight of hand isn’t it? Pontification for obfuscation? It’s thousands of hours and dollars and words that end up just being one more example of “science done backwards/incorrectly/inaccurately”.
I’m sure a lot if sincere work has gone into this original paper. I always, however, have a huge issue with using tree ring data somehow to estimate historical temperatures, especially over the long term. To state the bleedin’ obvious, it’s just not as simple as that – it cannot be.
The main determinant of tree growth where I live (Surrey, south of London) is, I am certain, water supply, not temperature and, no, rainfall cannot be used as a proxy for temperature (certainly not around here anyhow). Further, what about the effects of, inter alia, soil fertility, parasite and general invertebrate activity, tree diseases and, dare I say it, air fertility including varying levels of that dreadful poison CO2? How are these taken account of in tree rings?
I’m sure tree rings are terribly interesting and indeed you may be able to infer some isues from them but, as Willis so succinctly evaluates, if you look at one factor like temperature, the error bars must be so huge as to render the (reconstructed) “data” meaningless.
Proper science please – not starting with the desired solution and corrupting data relentlessly until you have got a correlation.
I can see how you can use tree rings to correlate major climatic events, such as the slowdown in growth caused by a major eruption, like a Krakatoa magnitude eruption.
However, to try and define temperatures from tree rings on an annual basis is just plain voodoo, because of:
1. Rainfall variations.
2. Shade variations from other trees.
3. Fertilisation from nearby forest fires.
4. Fertilisation from the occasional animal dump.
5. Nitrate boost from lightning storms.
6. Variations in growth depending on the side of trunk.
7. Occasional insect infestations on leaves.
8. Variations in competition from roots of nearby trees for water and nutrients.
9. Variation in ring widths depending on the age of the trees.
10. Variation in soil temperatures depending on slope direction.
I have seen much more comprehensive lists, but the bottom line is this: the only place you can magically transform BS readings into ‘hard facts’ is in climate science, where the use of voodoo statistical techniques is considered both acceptable and normal.
Rats!
And, of course, the increasing effect of CO2 fertilisation in recent decades.
Peter
Your list plus co2 is a good one. To it can be added the trees relative position in a forest which over centuries might range from the edge to the centre to the edge again as other trees grow and die. That all effects shade, water supply, competition for nutrients etc. Let us also remember that this is for summer growth only.
Rob has made a good effort and he is to be congratulated for defending the article over at bishop hill.
However after having read many papers, technical books and attended a short course I still cannot see how the data can be translated to a reliable temperature proxy accurate to tenths of a degree Celsius. Some sort of approximation to moisture, yes, but temperature?
However I am prepared to be convinced so feel it would be interesting if one of the authors could be persuaded to supply an article for discussion here that clarifies how tree ring data can be reliably translated to a useable and useful northern hemisphere summer temperature proxy
tonyb
Not to mention that the growth of trees can fundamentally change the environment where the trees are growing, and that it changes the growth of the trees, which again changes the environment, (just look at what happens to all the abandon Christmas tree farms scattered across the U.S.), I think you would need about 5 or 6 tree generations, in a specific location, to even get a handle on the way that trees can change their own environment. Since Mann started the tree-ring-schtick 15-20 years ago, and a generation of trees can be anywhere from 80 to 400 years this research only has about 2380 years to go until the “science is settled.”
Many claim that CO2 was constant before the industrial age began. It is well known that CO2 is a restricting factor on tree growth.
It is therefore a given that tree growth, represented by ring width, increases as the CO2 concentration in the atmosphere increases, a fact exploited by modern horticulture.
The increase in tree ring width caused by the additional CO2 has not been subtracted from the proxy result. This biases the ‘temperature’ upwards because both the temperature and the CO2 are known to have risen since 1750, but the combined effects are being attributed to only the temperature.
It is well known that both temperature and CO2 concentration combine to create the tree rings. How then can the temperature signal be extracted directly from the resulting series of widths if the CO2 concentration changed? It cannot. For any period of tree ring data for which there is a known CO2 data set, the effect on growth by rising CO2 must first be subtracted from the ring-only result in order to obtain the temperature.
This is so obvious I shouldn’t have to make analogies.
The most important contributors to a ring width are temperature, CO2 level and water. Let’s assume the rainfall averages over time to be constant. We know full well the CO2 has risen and when and approximately by how much. The fact that the number of proxies drops off at the end is noted. But the jump in the ring width signal must first be corrected for CO2 before acceptance as a temperature proxy.
Michael Mann’s infamous ‘hockey stick chart’, even if badly constructed, is a CO2 concentration plot more than it is a temperature chart, particularly in the modern era. If the effect of CO2 fertilisation is subtracted from the ring widths, the ‘temperature blade’ is truncated. Similarly if the upturn in this study’s ‘temperature chart’ at the end is first corrected for CO2 fertilisation (using available records) there is a ‘temperature’ deduction throughout the modern age.
Whether wonky statistical techniques are used or not, a tree ring width has to be corrected for CO2 availability before the temperature signal can be extracted. Once the CO2 hockey stick growth is subtracted from the tree ring width hockey stick, the remaining basically flat line proves that tree rings provide a poor record of temperature.
Inspired by other contributions I can add:
The tree ring width minus CO2 fertilisation gives a pretty flat line which represents the rainfall. I said above the rainfall could be assumed to be constant. A warm, wet, sunny summer completely offsets a long cold winter. There is no meaningful temperature signal in them thar rings. It is rainfall plus CO2. Subtract the effect of known CO2 variation and the result is the available moisture.
I live alongside a lake, on low, though forested property. The river running through the lake was dammed in the 50s, raising the lake level somewhat. That, plus increased boating since then caused the shoreline to erode and move inland some 30, or so, feet, causing some large trees to fall into the water. In an effort to stabilize the shore, I have removed several dozen trees from the lake, and lined the shore with more rocks. The tree rings are quite interesting.
Some of the larger trees had as many as 100 rings. They all showed variable growth, until the last couple of decades. Then the rings got closer and closer together, until they finally ended. Looking at the rings only, several conclusions could be made. 1) We are entering a period of Catastrophic Global Cooling, and it’s now too late. 2) We are entering a period of Catastrophic Global Warming, causing sea levels to rise, and it’s now too late. 3) Waves from water skiing kills trees. It’s politics, not science that determines the correct answer.
Excellent summary of some of the reasons to be very skeptical that the width of tree rings can give a value for the temp of the Earth.
To these I would add a few thoughts.
It is rather well established that most trees grow best when there is a cold winter with no big warm ups, followed by a wet spring with few extremes.
And extreme and fluctuating conditions is known to hinder growth as well…extremes such as late frosts and freezes, early hot periods with insufficient soil moisture, excessively wet and cloudy conditions that persist for an extended period of time (which can cause root die-off due to soil fungi), etc.
In short, the way that moisture, temperature, periods of clouds and sun, and other such factors interact is more important than any single variable or even combinations of variables.
In a given area, the Goldilocks formula for optimum growth is likely to be average conditions (in other words, zonal flow in the jet stream) over long periods. Extreme conditions of any sort have the tendency to interrupt this optimum growth.
Heat can hinder growth of trees, and thus be negatively correlated with growth, under several distinct and separate circumstances.
Where is the logical justification given that backs up such studies?
Do the papers themselves contain such justification, or are these results given based on the assumption that wider tree rings mean warmer overall average temps in that season?
As Ian noted, moisture is the primary determinant of growth in many areas, including arid climate zones.
If you think this stuff is good, you should look into the CO2 proxies Mann used. They’re based on boron isotopes in fossil marine plankton (foraminifera to be exact). A relationship between atmospheric CO2 and dissolved oceanic CO2 is proposed, then a relationship between boron isotopes and foraminifera shells is proposed. Not only are there errors introduced in each level of abstraction, we know from Henry’s law that oceanic CO2 concentration is dependent on temperature. Yet Mann and others saw fit to produce atmospheric gas concentration estimates purported to be precise to parts per million, which is plainly absurd.
This is really the fundamental and inarguable failure of paleo climate modeling; there are no measurements available to support it and there never will be. Anyone who tells you they know what the temperature of the planet was before the thermometer was invented, with any precision approaching 1 degree centigrade, is lying.
Peter Miller: And fertilization from increasing levels of atmospheric CO2.
What seems to be ignored in most paleoclimatology work is the recorded human observations that relate to any given period. Humans have had to pay attention to changes in other living things in order to survive. This is particularly true of all efforts relating to agriculture and fisheries, because these directly impacted food supply. I don’t know (there is a lot I don’t know) whether or not anyone has tried to assess what, if any, relationship (or association) there might be between recorded observations other than climatological (or meteorological) and dendrochronology. This is all interesting “stuff”, but hardly the fodder for policy.
Continuing from Peter Miller….
sunshine/shade variation due to clouds
variation due to diseases
effect of varying humidity
effect of different soils/ geological substrate
fault density and orientation- effects on groundwater supply………………so it continues
My difficulty with tree rings is tht when one looks at a complete cross-section in the majority of cases there is no consistency in the width of the rings. As the results from a core taken from one part of a tree are likely to differ widely from another core. That cross-section that Mann proudly leans on in the “classic” portrait of him illustrates the problem beautifully.
This one?
?uuid=R4kDLKaoEeGec_Tjh5s0ow
You mean this one, I think, as modified by Yrs. Truly.
?w=640
w.
“the error bars must be so huge as to render the (reconstructed) “data” meaningless.”
Which is exactly what is wrong with the current surface temperature data, and yet climate scientist still believe it can be used to seed models and spot trends…
I did a small study of tree ring data on my tree farm in Alaska (pines) and knew when they were planted, the monthly precipitation, and the monthly temperature. There was almost no correlation with yearly temperature and the highest correlation was with precipitation in the spring growing period.
Charles, I wrote my comment above before reading your comment. I have also been involved with commercial plant and tree production.
I believe your simple and understated comment far more readily than I believe that tree rings from a selection of separate locations can be used to infer the temperature of the Earth (even if there was such a thing).
Charles,
Please put together a paper and publish it in a journal!!
…but before submittal, be sure to put the phrase “and could be related to man-made affects on the climate” at the very end, regardless of the non-sequitur it creates.
Be sure to refer to it as “robust.” There are still a few people who haven’t tumbled to the fact that “robust” means balderdash.
And when the paper is done, go the your local toy store and buy the latest model of a WWII airplane. Then go home, put the paper on top of the model, and add a line saying your paper is based on the “latest model”.
This study was done several years ago. When I return home from my seasonal migration to warmer climes, i will see if i can find the data.
Now that would be any interesting study: tree ring width as a function of both temperature and CO2, other factors held constant. I wonder if anyone’s done that?
Don’t you think there’s a bit of hypocrisy in writing:
“I’m sorry, but I simply don’t believe you can do that.”
And:
“If you think that e.g. I’m using the wrong method on the wrong dataset, please educate me and others by demonstrating the proper use of the right method on the right dataset. Simply claiming I’m wrong doesn’t advance the discussion.”
Sammydj:
You ask
There is a lot of hypocrisy – and falsehood – in your selective quotation out of context!
Willis Eschenbach (WS) wrote
That is very, very different from his having only written “I’m sorry, but I simply don’t believe you can do that” as you have suggested he did.
And WS explained that what he “don’t believe you can do” is to “transform” (i.e. alter) data according to “expert judgement” (i.e. personal opinions) without introducing very large uncertainty (i.e. “error bars from floor to ceiling” for temperature indications because the alterations are to fulfill opinions and not measured temperatures). Every scientist would share that ‘belief’.
So, if you disagree that altering data according to opinions would provide “error bars from floor to ceiling and will have little to do with temperature” then say why because your saying that would fulfill the second statement of WS that you have quoted.
Richard
Well said Richard.
+100
Seriously? That’s the best you can come up with? Pathetic troll.
Sammy, since you seem to have asked this question of the air, rather than a particular person, I will answer as if asked of me.
My answer is: No, I do not believe these statements are in any way hypocritical.
As usual, a good post Willis. It seems these guys think they can baffle the critics by making things much more complex than they really are. Given the tenacity with which guys like you and Steve Mc attack data you’d think they would ensure that their methods add something to the debate. And if you are play the expert judgment card it must be the first card revealed, not half way thru the hand – in other words, they have to quantify in advance the rules of applying their expert judgment.
Interesting, the sudden and steep post-war drop . The only thing is that it happens in 1955 and not 1946. Looks like trees must have started using engine room intakes somewhat later than merchant shipping fleets.
More seriously that is an interesting independent witness that there was some climatic change at that time . There is an anomalous kink in the temperature record at that time. This suggests it is real and not something that needs “correction”.
… It would be good to remove at least part of the 1940s blip, but we are still left with “why the blip”.
Tom Wigley to Phil Jones Sep 27, 2009
The first thing one sees is the original “hockey stick” graph–a long period of (at most) narrow and superficial change, leading up to a final open-ended increase. Knowing what we know of the Medieval and other warm periods, to say nothing of the Little Ice Age, isn’t it time simply to drop tree-rings as a temperature proxy? I agree with Willis. If anything, he is too kind. I agree with Ian also, that rainfall, not temperature, is the most likely factor generating tree-ring widths.
Exactly. When I look at tree stumps, I see narrow rings on one side of the stump but the same rings are thick on the other side. The next tree stump over can show varying ring thickness on the reverse sides, then in another order again on the next stump. Ring thickness variability appears completely random to me.
How can we have confidence in tree rings as an estimate of past temperature, then extend the estimates to a global scale? It’s voodoo science.
If I had tried to do this in an undergrad science paper, they would have tossed me out on my ear.
Wind is why one side of the tree has bigger rings. If you look at prevailing winds, you will note that the rings on the side of the tree which is stretched by the bending of the tree are thinner and more dense. While those on the opposite side of the tree (compressed side) are wider and less dense. This varies a bit with species of tree, but wind has a lot to do with it. Years ago, as a kid, I worked with an arborist (tree specialist). The old man knew his stuff and could tell you how the wind affected large areas by felling a few trees. Trees in the center of a large group would have rings even all the way around while on the edges they would be offset. SO many things affect the tree while it grows that trying to glean temperatures from the rings is a tough sell.
“If I had tried to do this in an undergrad science paper, they would have tossed me out on my ear.”
Not today. In undergrad engineering, maybe, but not in science as currently practiced in academia. Lysenko lives!
I’d have to disagree with I agree with Ian also, that rainfall, not temperature, is the most likely factor generating tree-ring widths.
Tree ring widths are an indicator of growing conditions. All of them put together. For instance, in one year the water could have indeed been low causing ring growth restriction. In another year, water could have been abundant, but the temperature could have been too low for the plant to grow much. Another year, the windows could have kicked up and created dust storms that blocked out sunlight intermittently during the growing seasons, even with adequate water and temperature.
Biological systems are very rich, very complicated interplays between the organism and the environment the organism is growing in. The only way you can gauge one of these factors is to know all of the others that may have contributed to the organisms’ growth. Assuming a static climate for tree rings with the only variable as temperature is a ridiculously naive.
Now, don’t get me wrong: Tree rings have value. But they are circumstantial evidence that supports stronger evidence. For instance, if you are looking to see if 18167 (Year without a summer) was global, using tree rings chronologically matched to time would let you know that “something” happened during that growing season. But this can’t be immediately linked to the eruption of Tambora without other evidence, such as temperature readings, correlating ash, and similar.
Those poor people in 18,167 AD. I’ve predicted their doom to famine. (Think I can get some money by saying it’ll be caused by climate change?)
I meant, of course 1816.
As mentioned elsewhere, rings within individual trees vary greatly from side to side. Why? Prevailing wind pressure will cause a tree to grow larger/stronger rings on one side. Non-beneficial nematodes or other pests in only part of the roots system. Shade, soil differences, competition from other trees or plants, If you’ve spent time in the woods, especially logging or cutting firewood these things are apparent with minimal opposition.
In a mixed forest, live oak trees will starve other trees like pines and firs of water and nutrients. Do the smaller rings in the pine trees indicate a declining temperature or a healthy live oak tree? Without access to the original data, including mapping of the trees used for proxies in relation to other growing vegetation or evidence of now dead vegetation, there can be no confidence in tree ring proxies. My experience, and as noted by others, says that water in the growing season is the most critical factor in wild tree growth.
pbh
“Dendroclimatologists” are aware of all these deficits and attempt to offset them by using only certain trees in certain parts of the forest in certain regions. There’s a likelihood that these selections incorporate other unknown errors, and don’t fully compensate for many factors, but yes, the pig does have lipstick.
the temperature could have been too low for the plant to grow much
======================================
it could also have been too high.
There has been sufficient comment on the ‘hide the decline’ incident to make even the bravest of dendrodisasterologists head for the hills.
Many other workers have needed to cross train due to their industries failing or as technology makes them redundant, why not them?
One could list quite a few new endeavors for dendro retreads, but witch doctors, and the like, are in such low demand today. I’d suggest they take a look at becoming ‘men of the cloth.’ Fustian would be most appropriate.
Two thoughts. No matter how much the data from 710 to 2011 is tortured it represents a virtual zero in the geological time span of Earth’s climate so is fairly meaningless. A bit like saying that if I hiccup twice in succession in a 24 hour period then my life is made up of continuous hiccups.
The span of the blue area in figures 1 and 3 represent a span of standard deviation between plus and minus 4 or so. Without being a statistical guru, surely that alone tells us that the data is meaningless.
What is the physical or mathematical basis for adding ( or averaging ) std devs?
If we were measuring rainfall at two sites and added the “normalised” std dev records, what would this represent? What would it tell us about the amount of water?
Very little I suspect.
Sadly it seems that the data is only supplied in SD units so it is not possible to profit from the original authors inestimable expertise in converting the proxies to temperature estimations. Maybe some leg work into the original papers could find the actual proxy temperatures.
I take it that YAM is our old friend the Yamal proxy.
This looks like a reworking of Mann’s hokey stick. [sic]
This is just more home-spun, non-validated “innovative” methods with no justification or auditing . Typical dendro-astrology.
Surely we have temperature records which would have been used to calibrate this proxy so it could make temperature predictions before those records began. Meaning that during the calibration period it should either reproduce what we all ready know or prove itself incorrect.
Yeah, that’s Mike’s nature trick, again.
calibration is better known as “selecting on the dependent variable”. It is a HUGE statistical NO-NO.
Google the term. The other soft sciences have finally recognized the problem after it generated tons of faulty conclusions. Which probably explains why it is so popular in Climate Science.
“Any ‘science’ with an adjective in front of it isn’t Science.” –attributed to Feynman
This is an excellent and important point.
consider this. you have a drug test (or a proxy) that is 95% accurate. You have 100 people (trees) in your sample. You test these people and 5 show positive for drugs.
what are the chances that these 5 people are using drugs? If you didn’t know that these people were drawn from a sample of 100, you would say the chances were pretty good, that you were 95% sure they were using drugs.
But you would be wrong. Because in a sample of 100 people you would expect to get 5 false positives even if none of the original 100 are taking drugs. The 5 people might be taking drugs, or might not.
And this is the problem with calibrating tree rings. Unless trees are 100% reliable proxies, then you are going to get a lot of false positives which will throw off your conclusions. And if the trees are 100% reliable, there would be no need to calibrate them in the first place.
And unless you have a statistician on your team, you probably wouldn’t even know that, so the authors might claim ignorance.
Unfortunately, the general public doesn’t know that and it’s counter-intuitive to some extent, so a devious person with a political agenda would have no trouble at all duping them with a paper like this. It’s truly disgusting in my opinion.
Unless you blindly build a test which is purely binary in output, false positive come at the expense of false negatives (non detection of a true positive).
Basic radar signal detection builds on false positive (false alarms) tradeoffs with false positive (missed detection) as gain and detection thresholds are adjusted. Thus if the signal strength is sufficiently above the noise defined threshold to declare a detection, false alarms that degrade confidence in the system are virtually eliminated. This true for any system. If an operator merely gets a binary output, false alarms strongly degrade operator confidence. Sophisticated thieves use this strategy to defeat the human element by systematically inducing false alarms in an alarm system over weeks or months until the operators get tired of false alarms and either turn off the system or ignore alarms.
So going back to your drug analogy, suppose a hypothetical test can discern 1 ng/mL (+/- 0.1) in a bodily fluid sample. Would it be reasonable to accept any sample at > 1.3 ng/mL as positive? If one decides as a management strategy to avoid the legal implications of a false positive, setting the positive threshold at 2 ng/mL would ensure the positive detection threshold is high to ensure it catches true users. In the world of cycling where doping has always been a problem, two samples are always taken, A and B, so that if A is positive, B can be tested to confirm. If B doesn’t confirm then the result from A must be discarded.
What you really want is a system that of say 100 true users tested, less than 5 would go undetected. For what you MOST want is a system that deters use. That’s the definition of a 95% detection rate. Not a system that scares 1,000 nonusers by producing 50 false positives (maybe because they ate a poppyseed muffin a day earlier for example).
“a devious person with a political agenda would have no trouble at all duping them
Especially if said devious persons already had buy-in from science-illiterate (or politically motivated) media gatekeepers and “science”
activistsjournalists.Joel, I by no means want to take away from your very good example of statistical error and what it means, but I do want to mention that, in this example, I don’t think it’s relevant, only because what’s being done with trees and temperatures isn’t necessarily what’s being done with the testing of blood for the presence or absence of a chemical substance.
Tree rings are being used as an analog for the thing being tested for, they aren’t a direct measure of it. In the situation you describe (which was based on Ferd’s attempt to explain a statistical artifact, and was valid in that use) there is a direct measure of the thing being detected and there are errors in that measure arising from imprecision of the instruments used. In the tree ring example, we have instrument error coupled with the error introduced when we derive the relation between the thing we’re measuring and the thing we intend to measure, which can be as large as 100%.
Proxy measures are never accurate, at best they’re representative analogs. In the case of tree rings, they can’t ever be used outside their so called “calibration range”, since they’re entirely dependent on empirical regression models and we can’t ever legitimately use an empirical model outside of its calibration range.
So while your discussion is very informative from a statistical perspective (and also useful to folks interested in defeating alarm systems), I don’t think it does much to support the use of tree rings to measure temperature before the thermometer was invented.
In the case of tree rings, they can’t ever be used outside their so called “calibration range”
============
It is worse than that. Calibration violates the statistical assumption of random sampling. So if you calibrate the tree rings, any subsequent statistical analysis of the rings to see how well they behave as a temperate proxy will be misleading.
Hidden in the calibrated tree rings are rings that are not responding to temperature at all. Rather they are responding to chance occurrences of other factors, but your calibration exercise gives the FALSE impression that they are responding to temperature. As a result, when you do temperature studies using calibrated tree rings, they are prone to giving FALSE results, because your underlying assumption, that the trees are responding to temperature, is FALSE.
Rob, there’s a fundamental problem with the method you propose; we can’t extrapolate from empirical data. In the absence of a well proven and accepted physical model of how tree rings relate to temperature, “calibration” is nothing more than a best fit least squares regression on observed temperature and tree ring width of some period of time. It doesn’t tell you anything at all about that relationship outside the observation period.
What these people are doing is completely bogus. There is absolutely no reason at all to believe anything they say. The paper is useless. A waste of time. A known lie. Junk. Without any merit.
I agree 100%.
I agree also. To me they seem desperate to get the hockey stick back and that’s all.
As usual, clear and easy to follow.
Then it has been transformed again, this time using the expert judgement of the current authors, into standard deviations based on the mean and standard deviation of the period 1750-1950. Why this exact period? Presumably, expert judgement.
At a guess, the 1950 cutoff might have to do with the “divergence problem” https://en.wikipedia.org/wiki/Divergence_problem ? The DP being that tree rings purportedly correlate to (apr-sep mean) temperatures up til 1950 and then show temperatures to decline. Seems like they’ve fixed THAT somehow.
I find this whole thing to be opaque and on the whole I think dendrochronology can indicate which years were good ones for a given tree and which were bad. Can they identify droughts, fires, insect infestations? Probably. But I remain skeptical that they are much good at thermometry.
Dendrochronology, the way I learned it way back when, was a method of dating, not a way to determine precise values of growth parameters. The logic behind it is that, in a given area, trees will tend to have good growth years and poor growth years. This can be used to construct a timeline of such years, but only for that area. This timeline can then be used to date certain strata if it is found to contain tree remains that grew during that timeline.
So yes, dry periods can perhaps be identified by using other information to infer correlations.
But to jump from that to using the width of the rings to estimate actual rainfall amounts seems more a misplaced leap of faith than science.
Jumping all the way to global temperatures would likely make that average voodoo practicing, head-shrinking, bones rolling, haruspex shake his head with disbelief.
Yes. Dendrochronology in archeology uses trees as a clock, one tree ring tick per year. Very different than pretending treemometers in dendro climatology. Doesn’t work since even at the extremes (tree lines on mountains, boreal forest/tundra lines), many factors influence growth besides temperature, and no tree grows uniformly in a stand.
The authors are not talking about dendrochronology, why are you?
That is a precise and well established science and has nothing to with dendroclimatologists.
Some climato-astrologists , like Mann, wilfully use dendrochronology when referring to dendroclimatology because the former has some credibility that they would like to steal
“…why are you?”
I was responding to the other comments that use that word.
And the point I made was just what you said, but I said it differently.
See this comment below by Joel, which I had not seen yet when I made my comment above. It appears that my recollection was correct:
http://wattsupwiththat.com/2016/01/16/aproxymations/#comment-2121076
Hi Willis,
When I read the extracts at BH it raised questions in my mind and I emailed Rob Wilson:
“I have just read the brief comment on your recent paper at BH and had a quick question from that.
I understand that you have used the datasets which you considered most robust going back over past centuries. I know from my own experience and observation from growing timber that temperature is just one of a number of factors which influence the rate of growth in trees and thus tree rings. Two simple examples are rainfall and cloud cover. Records of those seem, at best, patchy and generally non-existent for most of the period your study covers.
My main question is did you, in selecting datasets to use, assess the original studies to be certain that they had taken or were even able to take full account of all of those other factors which I believe have until quite recently been considered largely irrelevant in dendrochronology?
Are you able to point me towards any work which is able to explain how the growth rate of a tree 500 or 1000 years years ago can determined as being a result of temperature as opposed to rainfall levels, or reduced / increased sunlight from higher / lesser cloud cover, or a change in wind levels (perhaps due to abnormal jet stream positions) causing a degree of stunting / or preventing wind stunting for a year or two, or perhaps a 1 in 100 year flood that deposits nutrients enabling higher growth rates for a period of years thereafter ? The reason I ask is because data on those causes is typically unavailable and if there is a robust methodology for determining this I would be very interested to know and understand it.”
he kindly replied very promptly :
“You misunderstand the basics of dendro [climatology].
All trees are sampled in locations where tree growth is predominantly controlled by summer temperatures.
The residuals from the local scale modelling may represent “other” factors but are random over space and time When averaging over large regions, the “other” factors are further minimised by averaging
You cannot expect to get a temperature from a tree that is not growing in a temperature limited location.”
To which I replied and asked ” many thanks for such a prompt reply which is helpful as I try and get my head around some of this !
Temperature limited locations controlled by summer temperatures makes sense but is there a way to identify and exclude the other variants of cloud cover and rainfall at that location going back over the last thousand years ?”
[ I did not say that I actually do not understand how ‘locations controlled by summer temperature’ are determined – but you may have come across this….]
Rob Wilson again replied promptly :
“There is some debate about the effects of clouds on photosynthetic rates and a comparison between ring-width/density parameters and carbon isotopes is addressing that situation somewhat.
Brief take home story is that one of the reasons that ring-width is generally a poor proxy of temperature in this moist-temperature limited environments is that other factors are also impacting growth.
Density and related parameters are controlled almost entirely by summer temperatures in these environments.
It is far from perfect, but we state this in the paper – there are good reasons why we don’t explain 100% of the temperature variance.”
I didn’t have time to pursue this further but did briefly check out a few papers on variations in rates of photosynthesis [PS] related to cloud cover – there have been a number of studies. One using artificial light at times of cloud cover found a 25% increase in PS in the illuminated leaves; conifers may respond differently to deciduous.
Anyway the above may be of interest.
Thanks for posting. Your questions to him are almost too simple and logical. How can you basically use an uncontrolled site with no reference data on other conditions for 500 years and have the balls to say that the only thing that changed was temperature? I will give these guys one thing. They know how to play the game and manipulate the gullible people with the purse strings. Make it ‘sciencey sounding’ and make the graph ramp up at the end and presto! You can afford a new car!
Hi Chilemike,
They seem to prefer to ignore studies which show that leaves are able to alter and maintain optimum temperature of around 21 deg C regardless of the ambient air temperature. (see my replies to Bartleby below).
Of course that means that the rate of growth, size and density of tree rings has nothing to do with the ambient temperature and so you can’t even hazard a guess at what the temperature might or might not have been from studying tree rings.
Old England, all I can say is I think you’re far too polite.
Hi Bartleby,
I took the view that if I were critical I would receive no response at all ! But having just searched for “temperature limited locations” which intrigues me, I stumbled across a 2008 science article in the Daily Telegraph and have quoted parts below:
“The temperature inside a healthy tree leaf is affected much less by outside temperature than originally believed, from England to the Caribbean, according to biologists at the University of Pennsylvania.
Surveying 39 tree species ranging in location from subtropical to northerly climates, researchers found a nearly constant temperature in tree leaves.
The conversion of light into chemical energy – photosynthesis – most likely occurs when leaf temperatures are about 21°C, and the outside temperature plays little, if any, role. This means that in colder climates leaf temperatures are elevated and in warmer climates tree leaves cool to keep the temperature just right.
The research contradicts the longstanding assumption that temperature in a healthy leaf are coupled to ambient air conditions. For decades, scientists studying climate change have measured the oxygen isotope ratio in tree-ring cellulose to determine the ambient temperature and relative humidity of past climates.
This new work challenges the potential to reconstruct climate through tree-ring isotope analysis, since it suggests the method does not provide direct information about past climate, providing misleadingly warm estimates.”
From that I take it that the rate of photosynthesis is not dependant on the ambient temperature and so the rate of growth of the tree, and tree ring size and density is decoupled from temperature.
Thanks very much for that Old England, I didn’t know it but, now you’ve found it, it doesn’t surprise me. I’m one of those unfortunates who entertain the idea natural selection and evolution have played a role in the development of life, and I’ve also heard that temperature plays a significant role in the rate of chemical reactions (this results from an exposure to physical chemistry at an early age), so it makes quite a bit of sense to me that trees would learn to optimize energy production by controlling the temperature of those reactions. I believe the authors are suggesting trees are “warm blooded” 🙂
That sort of tosses the idea of using trees to measure temperature; it would be very similar to using a rectal thermometer on a horse to estimate temperature of the surrounding atmosphere. I think another person commenting on this thread suggested that, after removing the effect of CO2 on tree ring metrics, the only remaining signal was annual rainfall, which connects very well with the report you quote. Funny how things start linking up with each other like that isn’t it?
Hi again Barleby,
The researchers aren’t stating that trees are as you say, tongue in cheek, ‘warm blooded’ but that they are able through physical means control the leaf temperature. Below is a link to the article which touches on the mechanisms – apologies I should have put that up with the quotes as it is worth reading.
http://www.telegraph.co.uk/news/science/science-news/3344206/All-tree-leaves-have-thermostat-that-maintains-temperature.html
From the limited stuff I have read, and thus limited knowledge I have, there are a few things which strike me:
From the studies on photosynthesis [PS] rates depending on the amount of sunlight and reduced light from cloud cover it seems that light is a key driver of the rate of PS – but is not necessarily the dominant one as trees are able to moderate and control the PS rate in response to various changes. (artificial light at times of cloud cover increased PS by around 25% compared to adjacent trees at same temp and soil / water conditions).
It seems from some studies that trees can reduce / increase the rate of PS in lower leaves when higher leaves are at high PS rate in high / low ambient sunlight. Couple that with the ability to control temperature in leaves to maintain them within an ‘optimum range’ (circa 21 deg C) and it begs the question “is there a mechanism by which trees are able to control their rate of growth to optimum levels ?”.
Given the first two facts it could be part of a Darwininan ‘survival mechanism’ – trees which grow too rapidly produce less dense and thus weaker wood which means the trunk is more likely to snap in high winds and die off. So there appears to be a sound reason for trees to limit PS rates which would result in that happening.
When plants or trees are grown in a shaded position they grow tall and ‘leggy’ to reach the light – generally with a thin and weak stem – as soon as they have sufficient light available the growth is put into thickening / strengthening the stem / trunk and expanding the canopy with side branches and more leaves. On reaching adequate light the leaf canopy is exposed to wind and requires a stronger trunk. In mature trees they stop height-growth when they reach their normal / optimum level and then extend the side branches and leaf canopy at the same time as thickening the trunk. It seems to make sense that the tree would still need to control PS to avoid over-extending branch growth which would leave it exposed to damage from normal winds – let alone storm conditions.
Trees grown in windy conditions grow denser trunks, often slanted in response to prevailing winds. Lets assume that a mature forest is subject to a 1 in 100 year storm (as happened in the UK in 1998) and that the bulk of the forest is knocked down. The surviving trees will lay down denser wood for many years after that in response to new exposure to wind until forest regrowth (20-70++ years depending on species) means they are becoming more sheltered again. A dendrochronologist sampling those oldest surviving trees trees 200 or 300 years or more later will conclude that the multi-decade denser wood is as a result of temperature. A complete nonsense.
CO2 levels are a massive influence on rates of PS – and I suspect this is greater than light levels – because in commercial greenhouses where CO2 levels are highly elevated the rate of PS and growth is significantly increased compared to same plants in same type of greenhouse next door and all other conditions the same but without elevated CO2. (Thus identical sun or artificial light, growing medium, available water and temp to both). We know from NASA satellite studies that the amount of green vegetation globally has increased by ~ 15% in recent years solely due to the small increase in atmospheric CO2.
We also know that in elevated CO2 levels plants require less water – they obtain higher levels of CO2 for PS with a reduced number of stoma in the leaves and that reduces the transpiration rate and thus the plants water requirements. Hence why deserts are shrinking (acknowledged by NASA, from satellite imagery studies) as plants are able to spread into them despite the limited water available.
I don’t know if increased CO2 levels produce denser wood and haven’t searched for any studies on that.
Where sunlight and CO2 levels are the same then as any gardener knows the availibility of water is a key part of determining PS and growth rates – too little or too much (waterlogged) reduce growth.
All in all I tend to think of ‘dendrochronology’ as more of an aspirational art form than science; creative thinking is the order of the day in the absence of any knowledge on what were the key determinants of rates of tree growth at the time the sample comes from – rainfall, sunlight and cloud cover etc. In the light of trees ability to regulate the temperature of their leaves then the temperature appears to have a minimal effect and cannot begin to be determined from tree rings.
Btw – always use a thermometer with a large bulb on the end you hold when taking a horse’s temperature – that way you won’t lose it !
I very much like the term “aspirational art” and understand the implied difference with a science. Recently I attempted to make the same distinction on the grounds of it being experimental with repeatable results. I hold that climatology is still in the early throws of being a science and best categorized, as you suggest, an art until further developments take place. Climate may be one of the more complex sciences ever attempted by humans, it’s my personal opinion it’s already demonstrated the limitations of differential calculus and may require a new branch of mathematics to succeed (I kid you not 🙂
I promise to take that under sincere advisement!
Very interesting comments, Old England.
I have a question which I wonder if you may have some insight into: What sort of trees are they using to construct temperature estimates going back over a thousand years? The top graph shows steady temps over a period that looks to start around AD 700. Is it safe to assume that there are few places where large numbers of living trees of such age exist? Is it therefore safe to think they are using old stumps and such from long dead forests, which may well have had completely different conditions than exist now, even if you did believe that areas could be identified that summer temperature controls tree growth?
Besides for all that, does anyone think summer temps have good correlation with average annual temp?
Hi Menicholas,
they tend to use a variety of trees from what I can make out. I think the most discredited study was on bristlecone pines which made fairly wild claims based on a minute number of samples.
The University of Arizona has a Laboratory of Tree Ring research at their School of Earth and Environmental Science. They now run the Aegean Dendrochronology Project which includes tree ring samples dating back “8,000 – 10,000 years”.
Willis Eschenbach:
Your first diagram shows “Standard Deviations” as its y-axis for the “53 proxies used in Wilson 2016” for the “Northern Hemisphere, 750 to 2011”.
It is labelled to say “Red line shows annual average of all proxies” but the blue lines are not labelled.
I am tempted to think the red line is actual values of the average for each year and the blue lines are the standard deviations of the averages shown red. If so, then what are the average values shown in red; e.g. are they annual average temperatures in unstated units? If not, then what are the red and blue indications on the graph?
I ask for this clarification because I am trying to make sense of the relationship between the two plots. Generally, the red line has ‘high’ value when the blue lines show wide spread, but the spread of blue lines is greatest around ~1300 when the red line shows a rise that is much less than the rise after 1988.
Richard
Sorry for my lack of clarity, Richard. The blue lines are the individual proxies, the red line is the average.
w.
Willis:
Thankyou for the clarification.
Richard
Expert = Farmer from the next county but one.
Professional = Someone who makes a living doing something and sometimes makes a mistake.
I know which one I listen to.
And before anyone else says it, the other definition of “Expert” is “Drip under pressure”.
There are 21 proxies in year 2000 and 8 in 2010.
Visually, the blade sticks up because most of the 8 for 2010 and nearby years point up.
This small number allows that the pattern has a throw of the dice component, like a string of straight sixes.
It might be invalid to use the post 2000 data because it is so sparse in relation to the rest.
Might be interesting to look at the changes in the final graphs after successively holding back these 8 recent ones.
Likely that the blade would get smaller.
But that blue mist of individual proxies will still be there, even though some of the higher and lower meet the borders of the graph box before they stop.
Geoff
One thing to do would be to just use the proxies that exist in the end portion. Apples-apples.
Looks like there’s a bit a ‘plateau’ in there.
Thanks Willis for this article, And I really love this one!
“Well, it was clear as mud but it covered the ground” ( can I use that one liner?)
And this ; “Well, you can do it, but the result will have error bars from floor to ceiling” ,
as graph # 3 shows!
Thanks, Tobias. It is a line from a Belafonte song, viz:
Well it was clear as mud but it covered de groun’
The con-fu-sion made the brain go round.
So I went and asked a good frined of mine
Known to the world as Albert Einstien”
w.
Thanks willis, LOL
climate scientists have been very successful in limiting the discussion to temperature although the real issue in AGW is not the temperature but the relationship between fossil fuel emissions and temperature.
JM,
I watched your video, which has no explanation. Then I saw another one (time series), so I watched that, too.
What are you attempting to show? The videos are both confusing because there’s no explanation or conclusion.
I’ve only skimmed through the original paper but I wonder if there is any correlation between increased tree growth and increased CO2. So far as I can see they don’t seem to have looked at that, although there are an awful lot of links and it is quite possible I missed it.
As usual, there are simply too many ambiguities for me to begin to understand the “Reconstruction Methodology” without going through all the original data myself. The key to clarifying this analysis–and many of Steve McIntyre’s analyses–for those like me who rarely have the inclination to slog through the data is found in this statement:
To the extent that the foregoing statement is really true, then a table of the proxy names and weights (for each sub-period over which the weights are used) would slice through all the inscrutable verbiage. (Yes, I know, there are 53 proxies times 55 sub-periods, but a 53 x 55 spreadsheet accompanied by a verbal summary would boil the technique down nicely.)
Most likely I’ve simply failed to notice that there already are such analyses somewhere in Mr. McIntyre’s or Andrew Montford’s oeuvre. Or maybe some of the methods really aren’t linear weightings after all.
Willis. Another excellent article. Two questions:
1) What is the global coverage of these proxies? In particular, do any of them cover the oceans (which are 2/3 of the Earth’s surface)?
2) Accuracy: What are the confidence limits of these proxies? What does the scale on the y-axis of Fig 3 actually represent?
Thanks,
Walt.
Walt, look at the paper, there is a map showing the locations and some lists giving dates, locations and occurrences.
How to fix a Nested Reconstruction Integratomasticator:
In other news, what will Dana do now … Oh! Dear!
The Recipe:
1. Take a bowel of spaghetti.
2. Feed it into the Hockeystickernator.
3. Add a dash of “Expert Opinion”
4. Feed to the press.
Robert: Hardest I have laughed in weeks!
Thank you!
Why, oh why do proxy data series never extend to today? Why do they always end 20 to 30 years ago, thus inviting the splicing of thermometer data onto proxy data?