By P Gosselin on 19. April 2022
A 2021 study appearing in Nature Communications by Buentgen et al reports on the results of a double-blind experiment of 15 different groups that yielded 15 different Northern Hemisphere summer temperature reconstructions. Each group used the same network of regional tree-ring width datasets.
Hat-tip: Klimaschau 108
What’s fascinating is that all groups, though using the same data network, came up with a different result. When it comes to deriving temperatures from tree-rings, it has much to do with individual approach and interpretation. Sure we can follow the science, but whose results?
The 15 groups (referred to as R1–R15) were challenged with the same task of developing the most reliable NH summer temperature reconstruction for the Common Era from nine high-elevation/high-latitude TRW datasets (Fig. 1):
Cropped from Figure 1, Buentgen et al
The 15 groups who contributed independently to this experiment all had experience in developing tree ring-based climate reconstructions. But as the study describes, each group employed a distinct reconstruction approach. In summary, the results ranged by as much as 1°C.
How could the groups come up with different results?
The paper’s abstract summarizes: “Differing in their mean, variance, amplitude, sensitivity, and persistence, the ensemble members demonstrate the influence of subjectivity in the reconstruction process. We therefore recommend the routine use of ensemble reconstruction approaches to provide a more consensual picture of past climate variability.”
Did anyone payed attention that they are not using “Mike’s Nature’s trick to hide the decline”?
The “supposed” anomalies decline in the most recent decade at least 🙂
Clearly the self claimed “renowned” Mann has to use the “blue pencil” somewhere in this article.
They then “carefully” cut the latest decade in the “reconstruction”. Compare the end of the X scale of Figure 3 with Figure 4. I still have to figure out how these things pass “peer review” (pal review?).
Data scatter.
Put two humans in a room and show them the same quantitative information and it is guaranteed to produce two different opinions on what it means. Use 15 persons and you’ll get 15 different answers, but the answers will likely show a mean value with proportionally less data scatter. Ditto with any larger group. It’s natural, it’s mathematical.
Yes, two people playing the same computer game, will play it differently..
Your point is ??
Would you expect 15 different conclusions if “the science was settled”?
Duane, do you realize how hilarious is your answer!! One of the main pillars of Science is replicability. Using the same data and the same methodology, results have to be replicable for conclusions to be taken. What you say is not Science, it’s statistics of scattered values. Even so, with the same data and using the same sample, the statistical values obtained must be the same. I know that Climate Science, according to some alarmists, can benefit from your view, to be possible to blame any possible conclusion on humans. However that is not Science at all.
I asked the arborist who treats my huge 100’ maples about tree rings and temperature. He just laughed. The width of tree rings are determined by local precipitation, more gives fat rings, less gives thin rings. If the north side of the tree gets more water than the south side, the north side will have fatter rings. Ring width has nothing to do with temperature.
Of course. Anyone working with this kind of data knows that tree rings are hygrometer proxies, not thermometer ones. That’s why I always get suspicious when Mann (auto-proclaimed renowned climate scientist) writes something. He can only be one of two things: dumb or deceiving.
The only reason they got within a degree is because they knew what temperature they were shooting for.
You think they can really tell the difference between Douglas fir tree rings grown in Seattle at near sea level, where the average July high is 73, and Truckee at over 6000 feet where the July average high is 83?
I don’t.
How do these authors know that the temperature histories from these 9 scattered sites were so similar that they could be added with benefits?
From earlier tree ring proxies?
Circular?
Geoff S
“Differing in their mean, variance, amplitude, sensitivity, and persistence, the ensemble members demonstrate the influence of subjectivity in the reconstruction process. We therefore recommend the routine use of ensemble reconstruction approaches to provide a more consensual picture of past climate variability.” Sorry guys science by consensus is not science.
not science but opiance … opinion dressed up as “science” …
Time to throw these logs on the fire.
Nothing more needs to be said.
Tree rings, is it temperature or water?
Just look at time-lapse filming of desert regions.
One month is very hot, no growth
Annual rains come, amazing growth.
Next month still very hot.
What happened? It rained which caused explosive growth.
Tree rings do not show temps.
The conclusion that “Differing in their mean, variance, amplitude, sensitivity, and persistence, the ensemble members demonstrate the influence of subjectivity in the reconstruction process. We therefore recommend the routine use of ensemble reconstruction approaches to provide a more consensual picture of past climate variability.” is just plain wrong.
What they have found is that tree-ring reconstructions only vaguely represent past temperatures — probably just “better growing weather” instead — and cannot reliably tell us anything about temperature itself.
Averaging bad data — unsuitable data — from different groups does not produce good data.
Never did, never will.
(I will try to write a full essay on this as it violates the rules of scientific evidence.)
“We therefore recommend the routine use of ensemble reconstruction approaches”
Doesn’t that essentially amount to “let’s average a bunch of guesses that may or may not be right”?
First:
Tree ring width varies.
Second:
To what degree is it true that these trees add one ring per year? This can be observed. Take a known grove, and cut a couple trees per year, and count their rings.
Third:
For observable time spans, is there a mathematical correlation between ambient temperature and tree ring width? This can be observed. Take a locale with a 60-year temperature record, and correlate that with the tree ring width of a tree known to be 60 years old. Memory, or photos, can verify tree age.
Fourth:
Doesn’t it seem that as a tree grows, its inner rings might get compressed, to some degree? We get better violins with hardwood that has been “aged,” allowed to dry – the moisture in this wood may be left to age from 5 to 20 years, with apparent differences in 5yo wood versus 20yo wood – all things being equal.
This could be measured, empirically. Not exactly sure how. However, we can observe distortions in tree rings from knots.
Fifth:
This website says that as some trees age, the width of the ring it grows becomes smaller. So, trees may have a high-growth juvenile or youth phase, and a later phase of life with belly fat and hair growing out of its ears. And trouble getting off the ground from a sitting position. And increased interest in bowel movement patterns.
The data graph at this link really has quite an elbow – looks like a scree plot. You have to factor in the elbow at just the right time, or your measures will be thrown off.
Men are at a steadily increasing risk of heart attack across the lifespan. Women’s risk rises, also, but much more slowly. Until menopause. When their increasing risk rate per year climbs. Another data pattern with an elbow.
But you cannot simply cannot have some specific year be the break point for this elbow, if modeling heart attack risk. Because this elbow is menopause. Women go into menopause at differing chronological points. Average age of beginning is supposed to be 51, and most women have beginning of menopause from 45 to 55. Quite a range of elbows. 10 years. 1/7 or 1/8 of entire life span.
https://seattlecentral.edu/qelp/sets/056/056.html
Two problems are shown up here, one technical and one social/?political:
1. Technical. The real history of temperature was not an average of possible but one actual. Each run uses temperatures that are an average of both proxied temperatures and error bars for those temperatures. The ensemble averages out actual highs and lows that had zero error bars. The result is low frequency whereas actual events were high frequency. So the MWP etc easily disappears in the mix and error bars.
This is the same problem for the IPCC scenarios: the actual future will be one pattern, not necessarily anything like the ensemble average.
2. Social/political. The public, including politicians, misread the ensemble as an accurate representation of past temperatures, a specific rise and fall instead of an averaged, educated, or best guess. To them, it shows the Minoan, Roman and Medieval warm periods didn’t exist. They don’t know that one interpretation showed they or one did exist but uncertainties in the interpretations and error bars wiped them out.
This loss of probable short to middle term warm and cold periods in the record is then compared to detailed measured records of the last 160 years that has almost zero error bars and high frequencies. A 50 year high frequency warm event looks like a horrific anomaly and is touted as such.
Maybe not. If you look at the ensemble error bars you can imagine an actual temperature rise and fall that matches the last 160 years – or exceeds it – while staying true to the uncertainties of the data and interpretations.
You will notice that past excessive hot or cold periods are written off as local, while currently they are called global. That could simply be in the nature of data uncertainties and interpretation.
Averages hide data variance. If the real variance was shown, noone would believe the doomsaying!
I am impressed that this ensemble group came up with ONLY 15 schemes.
Recalling those few decades back, when beginning my career in the petroleum exploration business, the joke was that if you gave X geologists the same dataset and asked them to render an interpretive map, there would be X+1 or X+2 … or etc maps.
There is no common era.
If 15 groups come up with 15 completely different answers, then this is not science, these are just guesses. We can’t even test them vs the “correct” answer, so they have the added benefit of likely never being proven wrong. Sure there is measurement error, uncertainties from numerous other factors not modeled and/or no data available, etc., but if the science was sound then most if not all of the groups would be in general agreement. Obviously this technology is not fit for purpose of determining past temperature. It is of course fit for purpose of propaganda and grant farming.
“a more consensual” That’s stupid beyond belief.
Let me translate it:
We therefore recommend the routine use of ensemble reconstruction approaches to provide a more fallacious picture of past climate variability, by the bandwagon fallacy.
Because logic is overrated and climastology must crap on the scientific method with all its might.