WUWT Opinion Poll – tree derived temperature data

This is for entertainment only. Given the week we’ve had, I thought it might be interesting to gauge some opinion about dendroclimatology. While we can certainly argue the merits of “who said what” etc. the question on my mind is what do people think of the technique of using tree rings for determining past climatic history?

Readers, please invite others at non skeptical blogs to participate, use the “share this” link. I’ll extend a blanket  invitation to anyone to participate, no matter what your view might be.

Since this is a highly polarized issue, I’ll note that the poll code is setup (by WordPress.com) to minimize the possibility of vote stuffing and encourage one vote per person. You’ll know you’ve hit that security feature if certain messages are displayed.

Here’s the poll question:

Of course I should add that no online poll is scientific, it is only an interesting and entertaining exercise in gauging the opinion of people who visit here.


newest oldest most voted
Notify of

My degree in botany taught me that plants have a variety of responses to environmental conditions. Tree-rings have proven good for archeological dating, but beyond that too many factors confound their clear interpretation as signals of climatic changes.

Doug in Seattle

I do not necessarily dismiss tree rings from consideration, but they appear to be of limited usefulness due to the many environmental factors that affect tree growth.
The use (abuse?) of tree rings in studies by Mann, Briffa and other members of the profession have not given me or many others great confidence in either the ability of tree rings to measure paleoclimate or in the competence (ethics?) of some researchers.


[snip – Tom, too many forbidden words here casting aspersions on scientists]

Joseph in Florida

I can not imagine what kind of a fool it would take to think that tree rings could tell us temperature. Think of the “scientists” who discounted the actual known historical record of the medieval climate optimum and also the little ice age based on what a few trees might reveal via their tree rings. Utterly amazing.

Antonio San

I think this poll is misleading as it almost advocates science by referendum. A careful line not to cross…
REPLY: Well that certainly was not the intent. I don’t think you’ll find many “science by referendum” studies or publications that have a picture of Snoopy and Woodstock in it. It was intended to gauge opinion of visitors to WUWT and to be entertaining and interesting, nothing more. It is the weekend after all.
But since you’ve demonstrated that some folks might get the wrong idea, I’ve made it clearer. Thanks. – Anthony


I have 2 identical species of trees planted at my house – one on the southeast corner in full sun that gets lots of water and one in the northwest corner in the shade that gets little water. After 8 years, the one in the sun with water is twice the size of the other. How could one make any interpretation of climate based on such trees?
True, these trees are affected by artificial (man-made) influences, but use of trees to estimate climate also seems fraught with a huge range of local environmental influences that simply cannot be accounted for.
My opinion: tree rings are interesting, but should be used with utmost caution and only when a very large sample of trees corroborates other sources of data with very little deviation.

Al Gore's Holy Hologram

The thickness of tree rings is heavily influenced by rainfall which can be heavy during cold or warm seasons or periods. Thus dendrochronology is a weak proxy for gauging temperature.

Willy Nilly

I collected some tree ring data as an undergraduate 30 years ago. There’s an inherent bias in that healthy trees are preferred, so all selected trees started life in good times. Bad times are gauged from those trees which successfully lived through them. Microclimates — such as if the tree is on a south-facing concave vale, or north-facing convex slope, or competition from neighboring trees — rule. You can only get climate information from features shared across most samples. So it is the median measure for each year which counts. You need to use all the data to get the median. Theoretically some selectivity could be needed, but only if the on-site collection was top-heavy by microclimate — you’d hope the sample collectors were better than that. Briffa’s concern that samples should be from multiple locations is correct, but one location is valid if it is a big enough “location”. A few trees from one location would be worth more per tree than many trees from one location, so you scale their importance as the square root of the number of samples, up to some maximum. So Briffa’s expressed concerns are good science, but not comprehensive I suspect.


FYI: I believe the correct term is dendrochronology. Never heard of nor could find anything on dendroclimatology.
Reply: Really? There is this thingy called Google you may want to try. ~ ctm


Plant development is a complex biologic process and small changes to a plants circumstances can have a big impact on rate of growth. Any group of plants growing in the same location, under similar conditions will show a range of growth rates due to genetic makeup and small natural variations – trees included.
Trying to use trees as proxies to measure small changes in temperature, which is then applied to produce global estimate of change is clearly risible. The Mann and Briffa hockystick graphs prove this to be the case as their temperature reconstructions bear no relationship to historic records.

Tim at 12:52 brought up a good point. A long time ago, my dad planted two oak trees about 7 feet apart. One tree grew 3 times as fast as the other tree. It grew 3 times as fast because it was getting the rain water run off from the house and was in a sunnier spot. Both trees are now cut down, but I remember my dad remarked about the differences in sizes despite the trees being planted at the same time.

Results 1 – 10 of about 154,000 for dendroclimatology. (0.47 seconds)
It may be a new, contrived, compound word. But 154,000 hits indicates it
is being used.


Its hard to see how tree rings were ever calibrated as a temperature proxy considering the number of confounding variables. The correct choice in the above poll therefore is “#2: the tree derived temperature data has been called into question.” There is nothing unsure about this.


My guess is that the whole problem is that trees are SOMETIMES halfway decent thermometers. But when which ones were, and when?

I hope this isn’t “off topic”. But I’m the orginal “contrarian” I believe when it comes to using O18 to 016 as a proxy for “temperature”.
I think the geophysics types and hydrologists would tell you that since the common availability of Mass Spec devices since WWII (probably even more so in the 60’s and ’70’s) O16/O18 ratios were used to trace water outflow from costal regions, with the concept that THUNDERSTORMS IN COSTAL REGIONS enriched the O18 isotope.
One can argue that a higher O18 number represents higher temperatures, as the O18 to O16 goes UP in even inland thunderstorms during the peak temperature summer months versus spring or fall.
But it should be obvious the NUMBER and INTENSITY of the TS’s will ALSO INFLUENCE THE AMOUNT OF O18 to 016.
I will submit, therefore, that the O18 to 016 ratio is a proxy for “atmospheric energy”, but that may or may not have any correlation with Troposheric temperatures (seasonal).

Peter Plail

You are not a warmist climatologist are you? You are certainly displaying the initiative and inquisitiveness of one.

R Pearse

I planted 12 boxwoods a couple of years ago in front of my house. Six get more sunlight and water because of a spruce tree and an overflowing eave trough, and they are 50% bigger on average. So I think tree rings are an unreliable proxy for temp


Besides the obvious complexities of other climate and botanical factors, the need for extensive metadata appears to be essential in order to control for the many environmental factors that can influence tree growth. Given that the environmental data changes through time as well as does climate, I am not very optimistic as to the robustness of trees as temperature proxies.


Treerings are not entirely useless from a climatological point of view. Most chronologies do reflect years with extremely bad weather – usually caused by major volcanic eruptions. Years like 1453, 1601, 1783, 1816 are found in most northern hemisphere treering series. Beyond that I am very skeptical.
It should be noted that treering chronologies were originally developed for archaeological dating, and in that field they have been brilliantly successful, as they have also been for calibrating radiocarbon dates.


There might be some correlation between tree ring widths and temperature, but surely you cannot get from them accurate measurements to one tenth of a degree. The attempt to created temperature “records” 2000 years back, with a resolution of one tenth of a degree is absolutely insane. When I first saw the Hockey Stick, it took me all of 10 seconds to deduce it was bunk. I just looked at the numbers on the vertical (temperature) scale, divided into tenth of degrees.
Being an engineer, I have some grasp of what numbers convey.


Well, I voted unsure. I agree with Bernie, the metadata is lacking and this seems to me to be a neccessary prerequisite for their use as temperature proxies.

K. Moore

I was under the impression that the only certain thing you could learn from tree rings was the tree’s age when it was cut down. Differences in width of the growth rings couldn’t possibly tell you anything with mathmatical certainty. They would only show that conditions were more or less favorable for growth during the tree’s lifetime.
Historical records and passages in literature of the day would seem to me the best indicator of the general temperature range during a particular period.


Evidence provided by an actual referendum, while very limited, still beats the blatant assertions of consensus and ‘settled science’.


A sole tree can certainly not tell of global past climate, however, the GLOBAL tree population, past and present, might tell us something useful.


I’d vote yes to,
With a sufficiently large sample of trees there should be an atmospheric temperature signal.

Doug in Seattle

It is my understanding that tree rings when correlated with temperature records can provide a reasonable proxy for temperature beyond the range of where correlation is possible.
What Mann does is to correlation a limited subset of the data (bristle comb pine from the US SW) with temperature and then adjust the rest of the tree ring data to fit the correlation. The result of his approach is that in forcing a fit he makes all the data prior to the instrumental record (and especially the data that is not the limited set he correlated) do weird things – like erasing the MWP. What Briffa appears to be doing is similar.
This approach is apparently only acceptable in the world of dendrochronology (or dendroclimatology for those that like the alternative term). Why it is acceptable there appears to relate to what in science is called “confirmation bias”.
Climate science isn’t the first field where confirmation bias has been allowed to run amok. That does not however make it acceptable.
Now that major government policy that can (and likely will) devastate the economies of the western world is being adopted (based in large part on the perceptions generated by the results of Mann’s and Briffa’s work), it has become imperative that the science community insist upon better proof.
In a perfect world where science is buffered from politics this is what would occur. But alas, we live in a world where science has bee hijacked by politics, so get ready for a wild ride down the economic slide.
Might be a good idea for younger readers to learn Chinese or Hindi as that is where the world economy appears to be headed.


Original, tree rings were used solely for dating purposes and for that they are very useful. Their use for temperature measurement, proxy, came much later and has always been somewhat dubious.
As our biology friends have already pointed out tree growth is subject to many environmental parameters and separating them from one another is extremely difficult and cannot be done by the method that ‘the team’ has used.

Stephen Skinner

Probably tree rings are probably better indicators of rain as others have stated, as sunlight on it’s own can bring about a plants demise. A drought can produce narrow tree rings which can happen at high or low temperatures.
Some years ago there was an incredible TV program called The Green Machine. One thing I remember from that was the sound of photosynthesis which appeared to be driven by sunlight. In which case I would say tree rings are an accurate indicator of sunlight and rain. That is NOT temperature and rain.
However, maybe there is a tree out there that grows in response to temperature alone. But to derive gloabl temperatures from that tree I would have thought you would then need to find the same kind of tree scattered evenly around the planet.
Another aspect of tree ring growth is what a trees growth profile would be over it’s lifetime. Do trees have the potential to have even growth rings throughout their lives?
I thought tree rings were used primarily as a means of dating historical artefacts, and have been a useful cross reference for carbon dating, and magnetic dating.

Frank Kotler

Vigorous, healthy tree growth in recent years (whether a proxy for temperature or not) is considered evidence of something disasterous coming our way.
A different “selection” of trees indicates that they’re not doing so well in recent years. This is considered “good news” (by some).
Are we in favor of tree growth, or opposed?
Watts up with that???

Stephen Skinner

Probably too many probablys.


I should add that an atmospheric temperature signal will only be found in cold climates.
I have only to look out my window here in Western Australia to see plants grow like crazy in the winter because it rains. Come summer almost all plant growth will stop because it doesn’t rain.

Jacob (13:36:02) :
The attempt to created temperature “records” 2000 years back, with a resolution of one tenth of a degree is absolutely insane.
Yes, it is absolutely insane; however, the purported resolution is not one tenth of a degree, but worst… one hundredth of a degree, i.e. index of growth*0.01 °C. It’s not science. I have used the formula for calculating the change of temperature based on treerings growth, then compared with solar irradiance and the resulted correlation is amazingly high. Then I plotted the raw data, without making the conversion to temperature and compared the results with the solar irradiance databases…
The correlation didn’t disappear, but reinforced. What does it mean? That the treerings growth depends mostly on insolation; the last assertion coincides with what we know about plants physiology. Siberian Larch Trees, or Russian Pines, slow their growth when the insolation is higher than 50% of the total available, so the conversion of treering growth index to change of temperature will always give false results and flattened lines. There is no doubt about it.
Bristlecone Pines and Siberian Larch Pines respond also to temperature in the same way. Their growth slows when temperature is above 23 °C. There is no way to know if the narrowing of the treerings growth was due to low or to high temperatures. The claim on associating the treerings growth with the environmental temperature is simply fraudulent.
More fraudulent is the fact that some dendroclimatologists have chosen C3 plants (pines) for their assessments when they perfectly know the treerings growth will give its maximum value at 23 °C, and lesser or minimum values at temperatures both below and above 23 °C. It would be fair if they would have chosen C4 trees from mid, tropical and equatorial latitudes, which respond proportionally to increases of insolation.
On the other hand, the growth of any tree depends on the healthy state of its leaves, which is where 95 % of photosynthesis occurs. However, if an army of ants decide to cut leaves of a group of trees in a given region, those trees will offer an extremely low growth even when the insolation index is 1, or if the temperature and humidity are optimal for the tree grows splendidly.
Conclusion, treerings are absolutely useless as proxies for assessing environmental temperatures.


The only thing tree ring widths tell you is whether conditions were favorable for a particular tree at a particular time. This might mean warm and/or wet weather, but what factors matter/mattered when and where and for which trees is…impossible. Dendroclimatology is, as far as I am concerned, a waste of time given the difficulties inherent which out weigh the dating advantages IMAO.


air temperature has little infuence on tree growt. if there is enough rain and lots off sunshine but colder temps then the tree will have a good year.

ian middleton

I’ve only ever considered tree rings to be useful in determining the trees age.
Wouldn’t hang my hat on the accuracy of any temperature data that may be implied by them.

Whether or not a given tree species grow in a certain area can be a temperature indicator, so the ‘tree line’ or limit seems to be a reasonable indicator of climate. Certainly no palm trees in Antarctica.

I think its clear that ring width is a function of many factors, temp being only one, and for that reason width should be discounted as a temperature proxy. I would suggest that a better means of analysing cores is to look at isotopic ratios, much as is done with speleo core data. Determining carbon isotope ratios of course is easy, of course, and I’m no expert, what possibly could be used here?


For sure not enough precision.


Tree rings are a possible source of temperature history.
Trees rings are a possible source of precipitation history.
Trees rings can possibly tell us about surrounding vegetation, particularly taller trees.
The trick is to figure out which of many factors the tree is trying to tell us about.
In the papers I’ve read, I didn’t see any reasonable methods being applied to try and separate out the various factors.
It would be interesting to take the various tree ring datasets and use CPS or other methods, but trying to extract a precipitation history. That should be as valid (or invalid) as trying to extract a temperature history.
All of the above comments also apply to sediment varves, such as used by Mann to determine the annual hurricane count back several hundred years. Of course the multiple factors affecting sediment varve thickness aren’t all the same as for tree rings, but the challenge is the same — to determine which of many different possible driving agents are the ones affecting varve thickness in different levels of the core. Just like with tree rings, we have no assurance that the reason for different thicknesses in one time frame is the same as in another time frame.
As one wag put it “Torture the data sufficiently and one can get any answer. “


How are these “other” hockey stick graphs addressed in this post at Real Climate.

a jones

Eh. What poll? Where’s the poll? And what’s the question?
Like global warming it isn’t showing up here.
Or am I looking in the wrong place?
Kindest Regards
REPLY: Try turn on photos for your browser or use a different browser.- A

John Cooper

What Wade (13:08:38) : said was exactly right.
Two years ago I planted about ten Dawn Redwood tree seedlings all within 100 feet of each other.. Some of them have done really well and others are puny. Whether it’s the soil, the sun, neighboring trees sapping their nutrients, I don’t know. Location, Location, Location…


From what I can see, studying tree rings requires even MORE random sampling than other proxies for the simple and obvious fact that each tree’s growth will depend mightily on rainfall, light, location (valleys and mountain shadows, other trees, underground water, etc.)
Trees are notoriously different between themselves, even identically aged trees next to each other in a forest.
I could see using trees for general, averaged climate as long as you gather and use data from a fairly large number of trees over a fairly large area. Anything else, any other use, any lesser method of gathering “data” will be meaningless. Or maybe, will be wishful thinking.


FYI, here is an email I have just sent to the NPR ombudsperson:
Dear Sir or Madam:
As a longtime environmentalist and fan of NPR, who does retain a scientific world view on controversial subjects like the widespread but increasingly religious belief in C02 induced “global warming” I have been appalled by the lack of objectivity apparent in almost all NPR shows that claim to treat this subject in an objective fashion. I’m sure that we can all agree that were the alarmist prognostications of the IPCC valid, then extreme measures would indeed be appropriate. The trouble is, they just aren’t. In fact, it should by now be evident to any reasonably objective person who has bothered to follow the real scientific debate on this topic, that the “consensus science” on which the public statements of Mr. Gore, Dr. Mann, Dr. Briffa, and other global warming advocates are based, is currently in the advanced stages of a complete melt down. This is particularly true now, given the manner in which Dr. Stephen McIntyre has recently revealed the dependence of Dr. Briffa’s new substitute “hockey stick” on tree ring data that is little less than bogus. If you don’t believe me, please visit wwww.wattsupwiththat.com and consider the recent relevant posts and discussion comments.
NPR staffers can’t even seem to get very basic facts, like the current state of arctic sea ice, correct in their news stories. It is quite untrue to categorically assert, as a gerundive fact, that “with arctic sea ice disappearing” anything else should be postulated. Arctic sea ice is NOT disappearing, notwithstanding the drumbeat of today’s mass media, including NPR. Here, for your information, is AMSR-E’s data on arctic sea ice for the past seven years: http://www.ijis.iarc.uaf.edu/en/home/seaice_extent.htm
Please note that we are the *third year* of a strong, apparently cyclic rebound, in arctic sea ice extent. The strength of the pattern over three years is particularly evident in this third season. This trend is consistent with the known reality that, for the last 8-10 years, there has not been ANY increase in global temperatures; if anything, there has been a slight decline, and many informed scientists expect this decline to continue for the next twenty years or more, at least.
The facts are diverging from global warming rhetoric. Its your job to notice that.
Given the extent of the denial and hysteria that mark the current state of discussion on global climate predictions, it is safe to predict at least one thing: whether the recent 8-10 year flatlining of the IPCC’s projected increases in global warming (and consequent divergence between the alleged cause and its predicted effects) continues, or merely represents a temporary divergence between theory and reality, the theory of CO2 induced global warming (which has always been based on the incorrect assumption that climate modellers actually understand the energy dynamics of earth’s system well enough to distinguish a real cause from a hypothetical one), seems destined one day to be a textbook example of the “extraordinary popular delusions” about which Charles Mackay first wrote in 1841.

Willis Eschenbach

First, Anthony, thanks for an interesting question. It is actually two questions, viz:
1. Does temperature leave a signal in the width of the tree rings?
2. Can we reconstruct the signal from the width of the tree rings?
The answer to the first question is definitely yes. A perfect temperature for that kind of tree will lead to wider rings.
The answer to the second question, however, is generally no.
Tree rings are not good thermometers. Trees grow poorly when it is too cold. But they also grow poorly when it is too hot. There is no way to tell which is which. A hot year is indistinguishable from a cold year.
This makes it theoretically impossible to extract the original signal. If a tree ring with a width of 2.7 mm could either represent 60°F or 80°F, we simply cannot reconstruct the original signal from the tree ring widths.
But wait, there’s more, as the pitchman says on TV … it gets worse!
Narrow rings can also occur because it is too dry, too windy, too much frost, late spring, waterlogged soil, and a host of other things.
Goldilocks in the story had three temperatures — too hot, too cold, and just right.
With trees, wide rings definitely mean “just right”. The timing of the frosts and the water when needed and the sunlight and the temperature and the nutrients and the clouds and the wind all conspired to make a good year for that tree. Things were just right. Not one thing was lacking.
But unlike Goldilocks, tree rings cannot tell too hot from too cold. Both of them make narrow rings. So does a lack of water. So does the wind blowing from the wrong quarter, and a host of other environmental factors.
The standard way to attempt to get around this problem is by simply interpreting narrow rings as representing cooler years. Inevitably, this pushes the calculated “temperature” downwards. It falsely represents the past peak temperatures as being lower than they actually were. Any event that slows tree growth, including too hot a temperature, is calculated as though it were a cooler temperature. You can see what that might do to e.g. the Medieval Warm Period.
So there are two huge problems with extracting the temperature signal from the tree rings:
1. A narrow ring can mean either too hot or too cold.
2. No one single factor (e.g. moisture, temperature, hours of sunshine) can cause a wide ring. All of them have to be right to get a wide ring. But a variation in any single factor can cause a narrow ring.
Problem #1 means that even in the best of cases, with all other factors being equal, our resulting reconstruction will be inaccurate and also biased downwards.
Problem #2 means in the real world, our resulting reconstruction will more inaccurate and will be biased even further downwards.
So my vote on the poll is no, at present the temperature signal can not be reconstructed through tree ring width.
My best to all,

Antonio San

Thanks Anthony.
I am always annoyed when I read in some media “that a majority of people believe in AGW” and feel that one single climatologist always trumps 100,000 uninformed people…

David Ball

I still trust the Hudson’s Bay Temperature records much more than I would trust dendro. Very close to 400 years of sub-artic temperature records that amount to millions of data points. Compared to that, the dendro stuff is very weak. No wonder there was such a concerted effort to marginalize the guy who was documenting that record. The Hudson’s Bay Company was probably in cahoots with big oil anyway, ……. 8^]


Some nice words here about dendrochronology being a better tool than dendroclimatology. I think that is true but there are limits. The same variations in growth of individual trees discussed here has the potential to make some wood samples simply unmatchable to a master chronology. Local climates and microclimates have similar effects.
There is however a natural human temptation to produce a match rather than no-match and a less than rigorous match is sometimes quoted as ‘the best available match’. But what is not always recognised or even mentioned when such dates are quoted is the very real possibility that the date produced might be wrong. Keenan has written about this.
Sometimes dendro dates are ‘supported’ by similar radiocarbon dates. However the master dendro sequences have been put together with the assistance of radiocarbon dating and radiocarbon dating is calibrated by information from the dendro sequences. I am not a scientist nor a historian but to me this means that the dates produced have a degree, perhaps a large degree, of uncertainty.
Years AD are not generally a problem because there is usually plenty of other evidence about dates so dendro dates are often simply used to refine an already reasonably well known date. But in years BC, particularly in pre-Classical times, there may be no available belt to go with the braces. There are therefore a lot of issues about chronology at this time and every reason for scepticism about the certainty of the scientific dates that are available.
All this is further confounded by the refusal of some dendrochronologists to publish or otherwise make available their data…

Jerry Haney

OT : I used to live in State College, PA (yes, home of M. Mann). On Mount Nittany there are two Hemlock trees, side by side, that are thousands, yes thousands, of years old growing in the middle of a small stream. They are extremely big and healthy and a sight to behold. If you ever get to that area, do visit them. It does require a bit of a hike, but knowing that you are touching something that is still alive and is that old is a very humbling experience, much like visiting a Stonehenge.
No, I do not believe our present science can determine past temperature using treerings.

Stephen Skinner

Sorry. OT and not entertaining.
Britons creating ‘more emissions’ http://news.bbc.co.uk/1/hi/sci/tech/8283909.stm
Greenhouse gas emissions created by Britons are probably twice as bad as figures suggest, says the government’s new chief energy scientist Professor David MacKay.
“Other countries make stuff for us so we have naughty, naughty China and India out of control with rising emissions but it’s because they are making our stuff for us now,” he said…
…”This not only means that the true scale of required emissions reductions in the Western world will be much higher but that the impact on economic growth and living standards there will also be more severe than so far believed.”
I wonder what support this viewpoint will have amongst those people that have seen their jobs go to China?