Since there has been a huge amount of interest in Steve McIntyre’s recent discovery of what happens to the “hockey stick” when all of the Yamal tree ring data is used rather than a special subset of 10, I thought it might be a good time to talk about how tree rings are grown.
Several readers have correctly pointed out that trees don’t respond just to temperature. The also respond to available water, available nutrients in soil, and available sunlight. For example, the only time in modern history where some trees did not produce summer growth rings occurred in the year 1816, also known as the Year Without a Summer:
Many consider the Year Without a Summer to have been caused by a combination of a historic low in solar activity (The Dalton Minimum) and a volcanic winter event; such as the powerful Mount Tambora eruption in 1815 which had four times the power of the famous Krakatoa eruption and hurled huge amounts of aerosols into the atmosphere.
Obviously, sunlight was a limiting factor for tree growth then, reduced temperature also figured in.
I touched on the idea of trees used for dendroclimatology being rain gauges before: Bristlecone Pines: Treemometers or rain gauges ? There has obviously been years of drought when trees also did not grow as much, so how do we separate temperature and moisture in the growth analysis?
But, right now what I really want to introduce readers to is Leibig’s Barrel.
While I don’t like to use Wikipedia that much, this page doesn’t seem to be controversial so appears safe enough:
Liebig’s Law of the Minimum, often simply called Liebig’s Law or the Law of the Minimum, is a principle developed in agricultural science by Carl Sprengel (1828) and later popularized by Justus von Liebig.
It states that growth is controlled not by the total of resources available, but by the scarcest resource (limiting factor). This concept was originally applied to plant or crop growth, where it was found that increasing the amount of plentiful nutrients did not increase plant growth. Only by increasing the amount of the limiting nutrient (the one most scarce in relation to “need”) was the growth of a plant or crop improved.This principle can be summed up in the aphorism, “The availability of the most abundant nutrient in the soil is as available as the availability of the least abundant nutrient in the soil.”
Liebig used the image of a barrel—now called Liebig’s barrel—to explain his law. Just as the capacity of a barrel with staves of unequal length is limited by the shortest stave, so a plant’s growth is limited by the nutrient in shortest supply.
Liebig’s Law has been extended to biological populations (and is commonly used in ecosystem models). For example, the growth of an organism such as a plant may be dependent on a number of different factors, such as sunlight or mineral nutrients (e.g. nitrate or phosphate). The availability of these may vary, such that at any given time one is more limiting than the others. Liebig’s Law states that growth only occurs at the rate permitted by the most limiting. For instance, in the equation below, the growth of population O is a function of the minimum of three Michaelis-Menten terms representing limitation by factors I, N and P.
It is limited to a situation where there are steady state conditions, and factor interactions are tightly controlled.
In the book: Forest dynamics: an ecological model, 1993, Oxford University Press, By Daniel B. Botkin the relationship between all of the different growth factors is modeled per Liebigs Law. In Chapter 3, page 64 the author talks about Liebigs Law and how it operates on forest populations. He then goes no to describe a model of forest growth.
Page 65 has a very interesting passage which I’ve highlighted:
For our readers reference, JABOWA is his forest modeling software, now in version 3 which you can read about here and download a free trial. It was designed for a specific region, the Northeastern United States. From JABOWA another similar forest growth simulator FORET was derived, which has been adapted for modeling diverse forests of the southern United States. Obviously modeling different forest regions requires specific model programming.
The point I’m making with all this is: If “the total growth response of a tree is the product of all environmental factors”, and forest modelers have to separate temperature and precipitation diameter increments, plus create different models for different forest regions, how can then one accurately divine temperature over millenia from width analysis of tree ring growth from trees in a single region?
The task is huge.
It seems to me that temperature is intertwined with so many other factors that it can’t easily be separated without at least knowing how they also changed over time. Further on in the book the author goes on to talk about other limiting growth factors, such as soil moisture, available light, nutrients etc, but a couple of pages, 71 and 72, really described how temperature (as degree days) is figured into forest growth estimates. I was surprised to see that photosynthesis is not linear with temperature, but parabolic! I’ve excerpted the pages below.
Another interesting thing I learned was that to accurately gauge forest growth, you should have nearby weather records. Missing that, you can estimate the temperature profile from January and July average temperatures. OK but how do you do that for 200, 500, or 1000 years ago?
Barring having any temperature information from the region at all, the possibility exists for doing lab growth experiments, though I don’t recall seeing anything about that in MBH98 or similar papers.
I make no claims of being an expert in either forest growth or dendroclimatology. I’m simply presenting interesting information here for discussion.
But, I am amazed at the nonlinearity and interactivity of all limiting growth factors, and especially the parabolic response to temperature.
Thus in my opinion, it seems rather difficult to estimate forest growth when parameters are known, and even then the outcome is uncertain.
Imagine trying to reverse engineer temperature from forest growth (using tree rings) with all of the simultaneously acting growth factors, population density and shade limiting, the parabolic response curves, and you realize that divining the temperature out of trees over millenia is a very difficult, nonlinear, and uncertain task.
If you see a wide tree ring, you can safely conclude the tree had a good year. If you see narrow rings you can conclude a poor growth year. But was that poor year a product of an unfavorable temperature range alone? Or was it due to lack of moisture or lack of sunlight or both? Not having local records for those, how would you know?
It seems to me that dendroclimatology has a lot of uncertainty.
Dave Wendt (02:55:14) :
Your scenario is called “peer pressure” and “finding a postdoc for the student pressure”.
I agree that it is a fascinating subject that has been swept under the rug by the consensus team.
There would appear to be abundant literature available that enables anyone with an open mind to question the validity of tree ring data used to produce estimates of temperatures. In addition, many of the contributors to this thread (most of whom, I would guess, are not professionals in the field) have highlighted additional factors.
The question that arises in my mind is why the peer review process has failed to raise all these very valid doubts. This would appy not only to Mann’s use of tree ring proxies in his (peer reviewed) works but also the authors of studies which Mann used to justify his use of temperature proxies.
I am not a scientist, so don’t fully appreciate how peer review works, but I would expect it to raise questions and demand answers before publication is permitted. How can all these undoubtedly intelligent people fail to ask such pertinent questions?
Anthony, the word “fraud” does not seemed to be allowed here.
Does it finally apply now??
OT but worth a laugh – the UK Met Office is predicting that this winter will be milder that last. I must remember to order my thermal underwear.
Oops !! the post was to go to “http://wattsupwiththat.com/2009/09/28/update-a-zoomed-look-at-the-broken-hockey-stick/” comments.
Are these tree rings from larches growing in permafrost? Wouldn’t the depth of the active layer of permafrost be a big factor in amount of growth?
An interesting article is “The Effect of Permafrost on the Northern Treeline” by V. V. Kryuchkov presented about 1973. Larches have some interesting growth patterns that maybe some botanists reading here could explain better than I (if indeed these rings are from Siberian Larches).
Nice book find. This was discussed in depth by many people at Climate Audit several years ago including the “parabolic response” to growth factors, the mathematical product model for growth, and the fact that tree ring growth is not symmetrical around the circumference of the tree. Somewhere in CA archives there’s a very funny annotated photograph of Michael Mann holding up a section of a tree that demonstrates asymmetric growth perfectly.
It’s good to see this technical but very clear enumeration of the the biological factors involved. The climatologists frequently enter the field with little if any appreciation for biological principles. Besides what has been presented here, there are influences on individual trees that cause them to produce different responses to their environments: 1) genetic variation, 2) mechanical damage (eg, strip-bark bristle cone pines), 3) tree anatomy (individual roots translocate nutrients vertically very well but not so well horizontally in some species which causes the trunk to grow unevenly if that root should tap into a pocket of minerals or water).
Given that trees have leaves which photosynthesise with CO2 ……. I guess its possible that they grow better with higher concentrations of it in the air. Which as a fact just by itsellf, makes a mockery of the idea that the growth was caused by the temperature …. which happened to have been caused by the CO2 ….
…..even if there was any correlation …. which there doesn’t appear to be …..
Given that trees have leaves which photosynthesise with CO2 ……. I guess its possible that they grow better with higher concentrations of it in the air. Which as a fact just by itself, makes a mockery of the idea that the growth was caused by the temperature …. which happened to have been caused by the CO2 ….
…..even if there was any correlation …. which there doesn’t appear to be …..
I have said Douglas Fir as an example, on the west side of the Cascades, it is a primary
Tree,in the Forest succession. Look at the tree rings of a 40-year old Doug Fir, and for the most part,wide,healthy, growth. Over here, on the east side of those cascades,that same 40 year old Doug Fir would be a fair Christmas tree.Moisture is the key…
I am just speechless as to how an entire field of researchers can look at a mountain of data, pick the 10 pieces they like while ignoring the copious volume of data they don’t like, make a graph, have that graph reviewd and published without anyone saying a damn thing. And they are PROUD of this workd and call it “science.”
I would have failed my 8th grade science project if I pulled a stunt like that.
Repeat after me: the chronologies that DO NOT correlate with temperature are JUST AS VALID as the ones that do. You DO NOT get to ignore them just because you feel like it.
Gary (06:07:03) wrote:
“…..there are influences on individual trees that cause them to produce different responses to their environments: ….. 3) tree anatomy (individual roots translocate nutrients vertically very well but not so well horizontally in some species which causes the trunk to grow unevenly if that root should tap into a pocket of minerals or water).”
I have noted in the lower trunk of some dead poplar trees that the wood discolored differently in various “zones” across the cross-section of the trees, and there was often a black “line” separating the zones. It appeared that the zones were determined by the root that fed that zone.
That’s why it’s called Climate “Science”.
Upon reflection, perhaps the method of tree ring data collection could be improved. Perhaps some improved variant of a UDD (Ultrasonic Decay Detector) see link below, might improve the quality of the source data by allowing non invasive and more comprehensive tree ring mapping.
http://www.treesearch.fs.fed.us/pubs/4429
Such a method of data collection might go some way to preventing future misinterpretations.
I worked on a similar model a few years ago. We looked at net productivity at all latitudes for forages and foraging lifeforms.
What we found was that the forage produced was the same regardless of latitude given the same growing degree days – but the parabola was taller and skinnier the higher in latitude one went. The brix (and related higher quality content) in the forage was much more sustained at higher latitudes, due to the constant solar insolation, leading to a much faster growth curve for the forager. The other thing we discovered was that the growth at higher latitudes was more efficient because it did not have to stop and start every day especially during the Boreal day which was almost constant. This meant that the forage needed less resources per unit of growth.
Very good read. I know that temp was only one factor in growth, but this sheds alot of light on the many veriables. I found it very interesting that a small tree ring, among other things, may indicate either cold or hot temperatures.
Maybe a better way of doing tree ring analysis is for each year to see what spiecis of tree grows the best, then look at those parabola tables to see what temp that spiecis grows best at.
Perhaps you know this, but Professor Botkin is a rather interesting fellow. I corresponded with him briefly regarding an hysterical column on climate by Paul Krugman, of the NYTimes. If you visit Botkin’s website, you will find an article he wrote about his skeptical attitude on AGW.
In our exchange, he made one excellent point relevant here as well. Computer models are great for gaining understanding of systems, how things may change if some other things change. They are not good for making predictions, unless the system is totally defined, like the orbit of a satellite, for example!!
If you start with the assumption that climate was constant within a narrow range and then hindcast the temperture and rainfall readings, using a temperature and rainfall profile from the 1800’s, for the period before records, then you get what you look for. That way, you know that if you don’t get the expected results from one sample, its obviously flawed, so you choose another one that is more compliant to your hypothesis.
Its far more difficult to explain why your original hypothesis was wrong than it is to ensure that your study comes up with the “correct” answer. The denro people aren’t the first, nor last, group to fall into this experimental fallicy. And its very difficult for most of us to admit that we were wrong, especially if there was great fanfare given to what we did, its easy to climb the tree, much more difficult to climb down.
Living things respond to a miriade of environmental conditions, they are not computer programs that operate in a controlled manner reliably time after time.
I have often wondered why they don’t look at O2 and CO2 isotope ratios in the tree rings (if that’s possible) to determine a better temperature proxy.
I’d like to thank Anthony for this article and for all his hard work.
jnicklin (08:55:40) – I’m sure that changes in air pressure allows gases to move through wood, so there are not many gases trapped in trees. Carbon-14 dating is based upon carbon in solid form, which does not migrate.
Anthony – Maybe we keep thermometers away from trees because the trees set such a bad example of how to measure temperature. 🙂
Anonymoose
I agree that wood is not so solid as it may seem, but I was considering the O2 and C bound up in the cell structures, not the air spaces. In either case, it may not be possible.
I couldn’t search CA images but did find this image: http://homepage.mac.com/williseschenbach/.Pictures/Mann_explains_treemometer.jpg
Quote Tenuc (00:11:29) :
“So I don’t think using a few trees from a few areas can give any useful information about temperature, and will at best only give a very broad and unquantifiable indication about climate, even when well calibrated to actual climate conditions.”
I absolutely agree that using a few trees from a few regions will not provide conclusive results on global temperature trends, and I wasn’t advocating doing only that. As important as the large- and long-scale trends are the uncertainties associated with them. I believe a careful scientist would try to estimate these uncertainties by (a) taking tree-ring samples from many places around the world as an attempt to piece together the more global picture, (b) taking many samples from the same location to investigate how much variation there is in that particular location, and (c) doing related experiments on the side, as you described doing.
I my view if tree rings can provide a useful proxy for historical temperature, then that would be exceedingly useful, and is worth trying to see if it can work. But quantifying the uncertainties in the measurement is an important piece of that work, and needs to be done with care.
Great description of forest growth factors – very instructive.
I think there still may be some temperature information to be gathered from tree ring data, once we get past the species growth curve standardization and the problems of defining the actual diameter of an asymmetrical tree ring.
Assume you have a valid growth curve for the species you’re studying. Identify three points on the curve: the ‘max growth value’ (MGV), a ‘smaller value’ (SV), and a ‘larger value’ (LV) at the same height on the curve as the smaller value.
If the tree ring you’re examining has a diameter exactly on the MGV, then the temperature can be determined directly from the curve, since no factor is the limiting factor. This condition is extremely unlikely.
If the tree ring has a diameter exactly as high as the SV, then we can determine a minimum value for the temperature, since a lower temperature would have produced a narrower ring. But the actual temperature may have been as high as the LV, which would limit the ring diameter to the same value, according to the standardized growth curve.
We can’t tell which factor was the limiting factor, but if it was the temperature, it can’t be any lower nor any higher than the SV or LV. If another factor is the limiting one, the temperature would fall somewhere in between.
So, tree ring data should be able to provide a temperature range which could have produced the given tree ring, even though a definitive, specific temperature reading can’t be determined.