I had a brief email exchange with Professor
Dear James,
This is fascinating. I had no idea these emails were in the public domain.
In general Steve has gotten most of this right. There really is a problem
with the trees not being sensitive to temperature after about 1950. My
current best guess is that the higher CO2 since then has caused greater
warming at night (which is corroborated by minimum temperature trends,
since minimum temperatures usually occur at night). Trees respire more
at higher temperature, so they lose carbon when nights are warmer
than average. So their ring width has not increased as much as it would
have if the warming had been uniformly distributed over the diurnal cycle.
I think this is all published now so it should be possible to set the whole
record straight. But I did indeed feel at the time that Mike Mann had not
given me a straight answer. So if there is a response written, it won’t be
one defending Mike.
Jeff
Cheers,
James Padgett
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
I am not sure I buy the tmin argument. Does not altitude change the Tmin rather alot?
Are all the tress that are merged to create a reconstruction all at the same altitude?
CheshireRed says:
November 28, 2011 at 2:48 pm
Daft question; has anyone else tried to re-create Manns work? Perhaps not the HS 1000 years, but what about say, the 20th century? How accurate or otherwise would other tree rings be in replicating known temparatures?
====================================================
Steve McIntyre spent the better part of a decade trying to get Mann to give him the data and methodologies. Steve Mac showed the methods used would tend to give a HS graph regardless of the data. McShane and W. showed that the statistics and data used were not correct and couldn’t reasonably detect a sharp upturn in the temps. See Phil’s comment @ur momisugly November 28, 2011 at 2:47 pm for a likely reason why.
As to the question of accuracy in the 20th century, they throw out the last 40 years of the century. See my comment @ur momisugly November 28, 2011 at 1:31 pm for the email stating that they do exactly that. Of course, all of this assumes one can actually detect a temp signal from tree rings. You can’t. See Doug Badgero’s comment @ur momisugly November 28, 2011 at 1:45 pm for various effects on tree ring growth…also, see the growth season for the various trees. In high latitudes they don’t grow in the extended winter, regardless of the temps. So, even if one could extract a temp signal from a tree ring, it would only be representative of a partial year. So, even if you could, you can’t.
Tea leafTree ring reading is cli-sci’s answer to phrenology.To follow up on John’s very cogent comment (in case Dr. Severinghaus is listening in, as he might reasonably do after posting or permitting a repost of a letter of his), there are a number of very reasonable objections one might raise to this hypothesis. One is the rather long list of things given above that are known to strongly affect tree growth and ring formation. The assumption is that everything but temperature is on average not varying, so that variation with temperature only is resolvable from the “noise” of the confounding effects across the record. This by itself seems to be a very dubious assumption, especially with regard to drought.
Second, regarding the idea itself — that trees respire more and hence “lose more carbon at night” when night time temperatures are warmer, resulting in smaller growth rings. On the face of it, this hypothesis makes little sense. The temperature anomalies we’re talking about are what, order of 1C? Less than 1% of the overall absolute temperature. Compensating for this are: a growing season that is correspondingly longer. More moisture in the atmosphere (according to dogma associated with the high sensitivity) and less loss of water. And the really damning one — higher partial pressure of CO_2. After all, the temperature variation is a 1% effect, but the CO_2 partial pressure has increased by what, 20-30%? Isn’t that what the fuss is all about? So not only is CO_2 a fertilizer that is presumably taken up a significantly higher rate during the warmer, sunnier DAYS, but loss at night should be DIRECTLY INHIBITED by the higher partial pressure in the atmosphere compared to (say) the MWP or just the year 1900.
You might say (Dr. Severinghaus) that these effects won’t matter, but you cannot know this without direct experimental support. So I ask — are there greenhouse studies — anywhere — that support the hypothesis that warmer greenhouse nighttime temperatures in a CO_2-forced atmosphere inhibit tree growth? I rather think not — when I look at manuals produced for actual greenhouse utilization growing plants, they do not instruct one to cool the trees — rather they claim that photosynthesis rates double with a temperature rise of 10C. This means that indeed, any “loss” of carbon at night should be more than compensated for by the 10% increased growth rate from the 1C bump average temperature, the longer growing season, and the 20% bump in available CO_2. If this is not true, we’re right back to the confounding variable problem — why is there a downtrend in growth now — it cannot just be the order 1% or so higher diffusive losses at night given the order of 10% or more gains all of the rest of the time. You might see a bit less growth than a linear hypothesis would admit, but a flatlining? An actual decrease?
Best to say “we don’t know why the rings get smaller, but it is sufficient reason to reject them as a reliable proxy of temperature alone as there are clearly other confounding variables capable of creating natural variations in tree ring growth rates that overwhelm any possible thermal signal.”
Otherwise, the temptation to indulge in confirmation bias is almost impossible to resist. You exclude trees from the record when they disagree with what you want them to say and include them when “they behave”. Behave according to what standard? If nothing else, if included they would CORRECTLY increase the noise on in the input data on the end you are fitting to the proxy record. This would presumably weaken the reliability of the fit throughout — if large errors can occur in trees in the fit period (grounds for rejection from the fit) they can equally well occur in the trees you include, only not during the fit period! You end up with an entirely fallacious idea of the precision and reliability of the fit if you throw out the noise.
I encounter the same problem in random number generator testing (one of my specialities). Curiously, one of the best ways to guarantee that a given sequence of numbers is non-random is to only use it if it passes all the tests in a good random number generator tester. A perfect random number generator sometimes produces sequences that are quite ordered, ordered enough to marginally fail. What matters is the distribution and frequency of failures (at any given threshold of p for rejection) as much as successes. In order to get maximally random numbers, you have to validate the generation process, not any particular output result from that process.
You do precisely the same thing when you cherrypick the data in any way while building proxies out of hand-picked series selected by any means other than flipping a coin or rolling dice! — by trying to pick things that “look random” around the answer you want, you guarantee that your answer is in fact neither random where you need it to be (averaging over the confounding variables) nor accurate (you’ve introduced bias) nor honest (you underreport the true variances of the result and hence overestimate the reliability of the result.
Just a friendly thought. If there is an actual double-blind controlled greenhouse-based experiment that demonstrates that 1C warmer nighttime temperatures (all things being equal) can on average suppress growth rates of trees for the species involved, I’d be perfectly thrilled to hear about it; otherwise you are indeed simply asserting an unsupported hypothesis that might or might not be correct, where the correct default belief is incorrect, not correct.
rgb
The fact that biomass is increasing shows the flaw in the “cornerstone” of climate “science” – a radiative balance for Earth – there is no such thing – never was.
Large amounts of energy (and CO2) are stored as biomass – if there are not significant amounts of energy stored in this manner then explain fossil fuels.
I simply refuse to believe that a few zealots can make up some computer algorithms, squeal about doubting thomases in a vindictive manner and have any chance of being right.
They are unco=operative with their data and methods because they know they made ridiculous mistakes and tried to cover it up with lies, deceit and venom.
Instead of acting grown up they have put their reputations into disrepute and I for one cannot wait for the discrediting to begin. This may take some time as the MSM and politicians have a lot to lose as well but I remain convinced it will happen.
Scepticism seems to be catching on like wildfire globally.
Did climex ever get published (see foia/documents)
HUH?
The Professor says:
“I think this is all published now so it should be possible to set the whole
record straight.”
I would sure like to see where this is all published. In fact, I would like to see some references to his other statments about Tmin, more respiration at night, effects of more respiration on growth rate, etc. He said in the email that he was not an expert on tree rings. Is he now saying that he IS an expert??
Memo to the Team: My vote is “No confidence” in existing tree ring proxies.
Your mission is to restore confidence in tree ring proxies. To do that, you must conduct a rather extensive experiment in remote locations of your choice, where all the significant variables are measured and /or controlled over time. Specifically grow your own trees from seedlings. This should take several decades, maybe close to a century. Report back when you can show the signature of each variable on tree rings/wood independent of the others. If you leave anything out you have to start over. ;<)
His best guess ??? really ? seems like a WAG to me …
ahhh … didn’t it get cooler from 1950 – 1970 ?
I am not sure anything has been done with temperatures but I do know that tests have been done on elevated CO2. My guess is this is more speculation and computer model stuff and no empirical experimental work has been done to back it up.
I know around here we put our plants in greenhouses to stunt their growth so the tomatoes don’t grow so large as to consume the entire neighborhood and we all know that trees get much bigger the farther North you go where temperatures are good and cold [/sarc]
Thirty peer reviewed studies show that tree rings are correlated much more closely to CO2 than to temperature:
http://climatesanity.files.wordpress.com/2010/02/correlation-coefficents-sorted-by-co2.gif
People seem to be forgetting that regardless of Tmin being a hypothetical mechanism, the divergence is real (so far as we can tell with the data, hidden or otherwise). This means either trees are not thermometers, thermometers are not thermometers, or both.
Folks: If you plot growth rate vs temperature for a tree, you will obtain a quadratic function roughly the shape of an upside-down U, which means that growth rate increases with temperature up to an optimum (around 25 C for most plants, IIRC), and then growth rate declines very rapidly with temperature after that point. SO, unless the temperatures are always below the optimum, growth rates are cannot be used as some proxy for temperature. Maybe some groves of trees somewhere have stayed below that optimum for a thousand years, but I doubt it. Then there are all the other variables that affect growth, ESPECIALLY soil moisture, which is usually much more important to growth than temperature.
The use of tree ring width (or ring density) as a temperature proxy is so fraught with problems that it should be discontinued, IMO (maybe it has!).
Severeinghaus knows about the non-linear relationship, because he mentioned it in the first Climategate email posted at ClimateAudit. Has he changed his tune? Professor?
What!? Trees respire at night? You would bring biology into an important dendro analysis? Shame on you, what are you? A real scientist?
/sarcasm /joke **(sarcasm and joking off)
The answer, obviously, is YES! Professor Severinghaus is a real scientist.
Let’s see if I understand this night emission theory though. The tree respires more at night (oxygen intake, CO2 expelled) because the night is warmer… Than what? Or perhaps this a suggestion of identifying the whole diurnal temperature profile and that warmer nights might cause trees to grow less than expected? Somehow, that answer leaves me trying to understand the chemistry and biology involved. Expelled CO2 is pulling carbon from the tree so the growth ring is smaller? Trees after millions of years are so inefficient in their biology that they will fail to grow if temperatures are “off balance” and they themselves become a polluting CO2 source Someone should notify Lisa Jackson, she needs to implement immediate tree controls.
Not to mention, it is my understanding in the nursery trade that the plants have been healthier and many growers use CO2 increased air percentages to help speed up growth. For this idea (above) to be valid, all plants would have to suffer this night time reduced growth pattern also.
Seriously, I do not expect you to defend the theory as it sounds, to me, more like an off the top of the head strawman idea. An idea that is intended to start the thought process, not be the final opinion, at least not till after proper research is done.
Thank you for your adherence to science!
I think Dr. Severinghaus is a nice guy, who can’t be brutal enough to come out and say that Mann’s dendro reconstructions are worthless. Unfortunately, people like Mann and Jones just walk all over nice guys like Dr. Severinghaus.
Continental US summer temperatures (the only temperatures that really matter for tree growth) from 1950 to 1975 show a downward trend of -0.16 degF / Decade according to NCDC.
But the more significant problem is mentioned above. Summer temperatures do not give you a picture of the overall annual temperature. For example: contiguous US winter temperatures from 1995 to 2010 have a down trend of -1.92 degF / Decade. Summer temperatures (which would be shown in tree ring growth) trend UP at 0.63 degF / Decade. Overall annual temperatures over that period trend nearly flat at 0.09 degF / Decade. So the tree rings are only really a proxy for SUMMER temperatures and summer temperatures might be trending exactly the opposite of winter temperatures, as they are over that period in the US. Tree rings can not show whether climate overall is warming or cooling.
I believe it is stated in several places in the emails that dendrochronological data are only useful for June-July temperatures and that is what temperature data they are calibrated against. So it is quite possible for the June-July climatology for an area to be quite different from, even they opposite of, the annual climatology. So plotting tree ring proxies against anything other than June-July temperatures in any kind of a graph is intellectually dishonest. Trees don’t grow much in spring, fall, or winter but those seasons still have climate. Tree rings basically discount the climate of 10 months out of the year.
The argument that carbon dioxide caused the nights to be warmer and thereby influenced tree ring growth does not hold water. That is because Ferenc Miskolczi has determined that the transmittance of the atmosphere to outgoing infrared radiation has been unchanged for the last 61 years. During that same time the amount of carbon dioxide in the atmosphere increased by 21.6 percent. This means that the contribution of this added carbon dioxide to the enhanced greenhouse effect was exactly zero. Hence, increasing the amount of carbon dioxide did not warm the night air. They better start looking at the measurements that purport to show this warm night air with a more critical eye. I would like to see proof that it even exists.
JeffC says:
November 28, 2011 at 4:07 pm
ahhh … didn’t it get cooler from 1950 – 1970 ?
==============================================
Well, during and about 1950 was marked by a particularly cold period. La Nina?……. But, you’re right, for 1940 to about the mid 70s we had a cooling trend which cause the future ice age scare….. http://www.woodfortrees.org/plot/hadcrut3vgl/from:1940/to:1977/plot/hadcrut3vgl/from:1940/to:1977/trend …… Which does call into question all the blathering about CO2. While I wasn’t alive for all of that period, I can affirm that I was alive for a good part of it. As I recall, we were industrializing even before my memory. CO2 must have been on a holiday for much of that 30+ years. It decided to quit effecting the climate then, even though it was the earlier part of the logarithmic curve. In a post I did, I “backcasted” CO2 levels to 1944 and compared them to the temps to 1977….. the assertions about CO2 sensitivity simply do not hold. Not even the most recent paper that states we overestimated the sensitivity. http://suyts.wordpress.com/2011/11/01/historical-correlations-between-temps-and-co2/ Consider the increase of CO2 emissions for over 30 years…. or nearly 40 going from under 300ppm to over 330. If the CO2 logarithmic sensitivity is correct, that should have increased the earth’s temps more than the next 30ppm, but it didn’t, the opposite occurred. And then going from 360 ppm to present, we see another gradual decrease in temps. The fact of the matter is, our temps DO NOT CORRELATE WITH CO2 EMMISIONS. For the last increase of 100 ppm ….. the first 33ppm temps went down, the next 33ppm temps went up, the next 34ppm temps are going down……
In 2009 I had a paper in Climatic Change Loehle, C. “A Mathematical Analysis of the Divergence Problem In Dendroclimatology” Climatic Change
Volume 94, Numbers 3-4, 233-245, DOI: 10.1007/s10584-008-9488-8
And analyzed the complexities of tree responses at length. The assumption of linear response MIGHT be ok for a few decades, but not when trying to detect possibly warmer periods like the MWP or colder periods like the LIA (in which more snow could cause counter-intuitive effects).
The warmest places on the planet have no trees. The coldest places on the planet have no trees.
Obviously, each tree species has evolved to prefer a certain temperature, a certain level of precipitation, a certain level of nutrients, a certain level of desease/insect levels, a certain level of forest fire prevalence, and all these also vary depending on the time of the year or the seasonality that each happens.
Tree-ring widths or ring density can tell you absolutely NOTHING about temperature unless you have controlled for all the other variables and have ascertained how each one of those other variables (and temperature) can influence the data collected on ring widths or density. In March and September and for every single time of the year.
This is, simply, an impossible task.
Especially going back 10 years or and even more especially, 1000 years. We know nothing about desease or insects 1000 years ago for example. If you think that is not important, just search for the impacts of mountain pine beetle in the last 10 years.
The fact that tree-rings are being used AT ALL to diagnose historical temperatures is quite frankly, RIDICULOUS.
How this became mainstream climate science and defended to the point of actually getting people FIRED, is even more RIDICULOUS.
Craig Lohle says:
“In 2009 I had a paper in Climatic Change Loehle, C. “A Mathematical Analysis of the Divergence Problem In Dendroclimatology” Climatic Change
Volume 94, Numbers 3-4, 233-245, DOI: 10.1007/s10584-008-9488-8”
Yes. Some good evidence for obvious flaws in the ill-conceived field (?) of “dendropaleoclimatology” (funny, haven’t heard that word lately)….It is just amazing in my mind how some really, really dumb concepts gets into the mainstream thought in scientific journals. Not just in climate science. Just look at the windmill and solar cell crap!
Huh??? Okay, the Prof just confused me.
Our high school biology must have missed something. Don’t trees respire Oxygen? How do trees lose Carbon more at night when they respire more?
As I see it, it would be coincidence, at best, if a tree ring chronology anywhere matched “annual” temperature averages over any serious length of time. Imagine that a tree is a perfect thermometer for a moment. Imagine that its rings are directly related to temperature. As mentioned before, those rings are pretty much growth in only June and July.
So instead of a tree, imagine we have several weather stations in a grid cell that only record temperatures in June and July and no other months. Those for the sake of this argument represent individual trees. Now we look at all these records and find the ones whose June-July temperatures correspond in general slope closest to ANNUAL AVERAGE temperatures over some period of time. There might actually be a few of those but that will be mostly coincidence and the fact that those stations in June-July have a series that produces a plot of about the same shape as that of average annual temperatures is chance. What are the chances that they will continue to closely match global average temperatures as we go forward in time or look at earlier times than our calibration period? Probably not very good.
I am not saying that dendrochronology isn’t an important field of study. I am saying that plotting dendro time series data on the same graph as annual average temperatures doesn’t mean anything. They are apples/oranges. Now in the “hide the decline” graphic, Briffa’s and Mann’s dendro data first of all carry two different version of “summer” and so (I think) would have different calibration criteria. Then annual observation data are appended to the end of the graph (not “summer” observational data). So what we have is prior to the instrumentation data there are several “summer” proxies and annual instrumentation data tacked on at the end.
Imagine taking a census of the number of students attending classes in June and July and graphing that over time and then at the end of the graph you append the annual average of the number of students that attend classes. The graph would be meaningless, wouldn’t it? Because the early part is a graph of apples and the late part is graph of oranges. But they merge them together because the picture it presents is what they want people to believe. It really, really, gets the “point” across even if it is fiction.
What I find hard to believe is that anyone would take the graphic seriously in the first place let alone dissect it to any degree. The whole premise of the graphic is preposterous. It might have SOME modicum of validity if only June-July NH average temperatures were plotted. But the caption of the original “hide the decline” graphic said:
That right there is reason to demand that the instrument series be removed from the graphic. You can not compare summer only data with annual data (well, you CAN, if you are trying to show a difference between them). Why didn’t that get ripped out in review? Why were they surprised that they diverged? Why were very specific regional data (the dendro time series) completed with annual averages for the entire NH? I would have chosen the instrumentation data for the grids where the trees were located for June and July and then maybe I would have had something, unless the June-July temperatures in those actually declined, too (as happened in North America after 1950). Then maybe the “divergence” problem goes away if you do that but the graph doesn’t have the desired trend anymore.
At best tree rings can be a June-July proxy and that’s it. They reflect growing conditions during the part of the year when growing happens. They can’t much reflect conditions occurring when there isn’t much growing going on.
DocMartyn said:
I’m Glad you bring that up. Fairly early on at Climate Audit, Steve Mc had a post about some (Russian?) researchers who did a transit of a mountain to get cores from a group of trees at various altitudes to see how that affected the tree rings. I don’t remember the results, but the post should be looked up and we might be able to judge how Dr. Severinghaus’ idea holds up.