Buy

Books
Click images for more details

Support

 

Twitter
Recent comments
Recent posts
Currently discussing
Links

A few sites I've stumbled across recently....

Powered by Squarespace
« Money without accountability | Main | McKibben's mindpoop »
Thursday
Jan142016

Wilson trending

Rob Wilson emails a copy of his new paper (£) in QSR, co-authored with, well, just about everybody in the dendro community. It's a tree-ring based temperature reconstruction of summer temperatures in the northern hemisphere, it's called N-TREND and excitingly it's a hockey stick!

I gather that Steve McIntyre is looking at it already, so I shall leave it to the expert to pronounce. But on a quick skim through the paper, there are some parts that are likely to prompt discussion. For example, I wonder how the data series were chosen. There is some explanation:

For N-TREND, rather than statistically screening all extant TR [tree ring] 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 reconstruction possible from the available data at that particular location. 

But I'd like to know how the selection of chronologies used was made from the full range of those available. Or is this all of them?

If so, there are not many - only 54 at peak. Moreover, the blade of the stick is only supported by a handful. On a whim, I picked one of these series at random - a reconstruction of summer temperatures for Mount Olympus by Klesse et al, which runs right up to 2011. Here's their graph:

You can see the uptick in recent years, but you can also see an equally warm a medieval warm period. This is only one of the constituent series of course, but I think it will be worth considering this in more detail as it does raise questions over the robustness of the NTREND hockey stick blade. 

PrintView Printer Friendly Version

Reader Comments (171)

Not being an expert I just "feel" that the "hockey stick" at the right end is a product of the warm period after 1975 or so, mixed in with the steady growth of man made CO2 emissions, which stimulate growth and enable more growth even with less water available. So how do you "homogenize" down that blade to account for increased CO2 not there prior to the creation of said "hockey stick"? There must be an effect that should be "adjusted" for. If not, why not? Just curious.

Jan 14, 2016 at 3:21 PM | Unregistered CommenterJohn Peter

Is the graph #1 centred around an average? If so, how is it derived? I was wondering if the preponderance of red might make it seem more 'alarming' than it otherwise might be.

Jan 14, 2016 at 3:23 PM | Unregistered CommenterJJB MKI

 I have discussed this with the met office and David parker who as you know created the 1772 record which is daily as opposed to the monthly 1659 record created by Manley. The record is good and robust. 

After 1772, as I said, it improves markedly, before that Manley agrees with me about the usefulness for comparison. Amongst the 'other reasons' I mention he wrote …

Before 1752 monthly means capable of comparison with those of England today […] cannot be provided satisfactorily unless one has daily observations to cope with the change from the Julian to the Gregorian calendar.

Add to this the reading taken in rooms, the infilling from another country, non-instrumental readings (etc., etc) and you understand why temperatures from this period were rounded to the nearest degree or half-degree ….

Jan 14, 2016 at 3:36 PM | Unregistered CommenterPhil Clarke

JJB, they say it has been "normalised to the 1400–1850 period to enhance differences in the Medieval and recent periods".

But you are right, it's arbitrary and affects how the graph is perceived. It would be better to simply plot the graph. No red or blue colouring is required.

Also I agree with Bish and Brandon that it's very disappointing that they are so vague about how the series were selected. Their use of 'mostly' reminds me of the Fast Show.
They use the word 'robust' repeatedly but make no attempt to show that their result is robust with respect to their choice of series selection, which I expect it isn't.

Jan 14, 2016 at 3:38 PM | Registered CommenterPaul Matthews

Ironically this is part of the same climatological canon that just cannot display pre-1979 Arctic sea ice data, because it's considered obsolete...

Jan 14, 2016 at 3:41 PM | Registered Commenteromnologos

I note Phil Clarke has not answered as to whether he thinks tree rings are more accurate than CET (rounded - or not - to half a degree). Whatever, I have something else that confuses me...

MBH's HS blade was infamous for Mike's trick of infilling with real temp data so that the decline that his TR proxies was showing was hidden. That being the case, if Wilson's blade is made up of (only) TR proxies it would seem that either his paper refutes MBH, or MBH (TR only) refutes his. And that's not even taking into account the different ways the MWP was handled.

Jan 14, 2016 at 3:53 PM | Unregistered CommenterHarry Passfield

How does one check the provenance of any individual core sample taken from a tree? If they are all computerised, are they are open to 'homogenising', without detection?

The MWP has reappeared which has to be evidence of credibility over the Hockey Stick, and reduces the credibility of those who have argued that the MWP was a local issue. But how could there be so much discrepancy if dendro is a record of temperature?

When dendrochronology is used to date timber in buildings, structures archaeological artefacts etc, it is accepted, because it is about matching patterns in growth rings, not trying to diagnose the cause of different growing seasons.

Jan 14, 2016 at 4:03 PM | Unregistered Commentergolf charlie

Also kudos for Rob posting here.

My question is whether rainfall historians, squirrel experts, human migration professors, moose-sh!t experts also claim this squiggly graph as their own

who gets first dibs on the tree ring ?

Jan 14, 2016 at 4:08 PM | Unregistered CommenterEternalOptimist

Thank you Phil Clarke (12:01PM) and Martin (12:06PM) for trying to help me with news of a sidebar! But to no avail – I just don't get a sidebar appearing, nor any other bar. Just the 'About' info with no way to move on from it.

But the link you gave, Phil, did get me to the publications which is what I wanted. So thanks again.

Now I just need to find time to read it!

Jan 14, 2016 at 4:13 PM | Registered CommenterJohn Shade

The difficulty of attribution is squared here. Firstly, the attribution of tree-ring response to warming, or whatever; secondly, the attribution of warming to man, or whatever.

And to cube it, how can this result be attributed to reality, given the bias of the researchers, despite Rob Wilson's brief flirt with skepticism in 2005?
==================

Jan 14, 2016 at 4:18 PM | Unregistered Commenterkim

Phil Clarke is very sniffy about "temperatures from this period [being] rounded to the nearest degree or half-degree …."

For myself, I am a lot more sniffy about whether even the very latest, super-duper, state of the art, artfully adjusted and homogenised temperature data series can be relied upon to be accurate to within even a full degree. On a good day.

Jan 14, 2016 at 4:21 PM | Unregistered CommenterMartin Brumby

Martin

Hubert Lamb said about temperature reconstructions that 'we can understand the tendency but not the precision.'

I would not trust any temperature record to be more accurate than half a degree or so until we arrive at the modern era of digital weather stations, but then we can have a problem with siting.

CET is remarkably constrained within its temperature boundaries and it is relatively straightforward to pick out unusually hot or unusually cold years, so the 1659 record is a good one, supported as it is by other useful proxy records and by data from climatically similar areas.

I would invite Phil again to confirm why he dismisses interpolation from a climatically similar area a few hundred kilometres away, but accepts them for other temperature records with different climates from up to 1500 km?.

More interestingly, he can explain why he believes tree rings are a reliable summer proxy thermometer accurate to fractions of a degree? We are all waiting for that.

tonyb

Jan 14, 2016 at 4:50 PM | Unregistered Commentertonyb

Hi again,
Will be brief as this will all come to the surface again if SM has a look at the data.

Choice of data-sets:

The data used represent the latest PUBLISHED versions of local and regional tree-ring based reconstructions of summer temperatures. We wanted to use data that came as close (ideally beyond) to AD2000 as possible.

The data for ALL these studies were sampled predominantly in temperature limited environment. We are quite clear in the paper how much of the local temperature variance these different data-sets explain. Some are stronger than others although there are clear problems in the instrumental data in some locations (i.e. central Asia etc).

The residual error of the local scale calibration will represent “other” factors influencing growth, but these influences are random in space and time (easily testable by examining the residuals). When the data are averaged over such a large-scale, the consistent dominant temperature response dominates. Please look at Appendix Figure C3 – the good agreement between these two records is quite profound and says much about common large scale forcing.

Note – we carefully discuss the differences between ring-width (RW) and density based parameters. RW is a TR variable that integrates different environmental signals. The temperature signal can be optimised by careful site selections at tree-line (high latitude or elevation) but RW is still generally a weaker variable than density for reconstructing summer temperatures.

N-TREND is an evolving data and if you could make me aware of other data-sets (starting before 1750 and extending beyond 2000) not included I would be happy to post updates. There are for example, some new data-sets in Stoffel et al. (2015) which we could now include.

Rob

Jan 14, 2016 at 4:57 PM | Unregistered CommenterRob Wilson

Rob
Kudos for showing up and discussing your paper
Tonyb

Jan 14, 2016 at 5:06 PM | Unregistered Commentertonyb

For all those with questions about tree rings as temperature proxies, on his blog Ecologically Oriented, Jim Bouldin did a series of posts on theoretical problems with this subject, starting off with the observation that a small ring could derive from either a temperature too cool or too high for the tree to grow. Since there is no way of knowing, each data point has 2 solutions, even if you accept that it has been possible to select trees where temperature is the only variable factor in the whole life of the tree, eg that water or other factors never became key to the tree's growth. In short, the field of dendro thermometry is probably unsound scientifically.

Jan 14, 2016 at 5:10 PM | Unregistered CommenterDiogenes

Do the data sets control for CO2 concentration?

Jan 14, 2016 at 5:19 PM | Unregistered Commenterrhoda

While writing this, I see that Rob has already responded. Nevertheless, some brief responses from me:

1. Accessing paper and other materials.

The page at the link provided by Rob seems to adapt according to the width of your browser window. Wide enough and you get a sidebar with links to the data and publications. Narrower and the sidebar disappears, but is replaced by a drop-down list of links to the right of the main title. At least that's how it works for me on a desktop.

2. Tree-rings versus temperature versus other things like water, grazing, neighbouring trees falling down, etc.

This is obviously a big issue that I can't do justice to here. We cite some refs that should be a good introduction about climate influences in general. First, best to make clear that we do not claim that only temperature influences tree growth in general. Instead, we approach the issue as follows.

Many (though not all) of the tree-ring records we use were selected from sites near to the northern or elevational tree-lines. Their growth is limited by summer temperature because they are growing near to their limit for summer warmth. This increases the influence of summer temperature, decreases the influence of some other factors (e.g. if it is a non-arid site, moisture limitations will be reduced).

Next, remember that at each site we have multiple samples. Some other factors will not affect all trees at once, so will not affect all samples at once. Further, this can be quantified by looking at the agreement between samples and the sample size over time (in dendroclimatology that's often done using the Expressed Population Signal -- see Wigley et al. 1984).

Note that the new paper does not report this level of analysis at individual sites -- instead we build on such work already done for the individual chronologies or groups of chronologies that we have collated into the N-TREND network.

Empirical correlation against observed climate data is then another component of assessing temperature signal in the data (and whether other factors overwhelm it, at least during the instrumental period). See final column of our Table 1, though the authors of the original chronologies will have done this to varying extents too.

In the same way that agreement between individual tree samples can be assessed, so can agreement between the aggregate site records, or agreement between even larger aggregations of the data. See e.g. Appendix Figure C3 in our Supplementary information (also available at the N-TREND site) which compares the North American and Eurasian independent composites. Note that even if the temperature reconstructions were perfect (they're not), we wouldn't expect perfect agreement (correlation=1) because North American and Eurasian temperature variations are not identical, especially at interannual timescales. Correlations remain positive throughout, especially at timescales of 20-years and longer (see the bottom panel), though they weaken before 1400 and then weaken further before 1000. This demonstrates some degree of common forcing on the tree growth -- which is more likely temperature than, say, sunshine or precipitation, because the latter show more localised variability and (I expect) a weaker common signal between continents.

So, that is to give you a flavour of what can be done to demonstrate a significant temperature influence.

3. Data selection.

As Andrew picked out in the actual blog post, we don't rely on empirical screening for selection. There are pros and cons of course. Including that we can utilise the expert judgement in selecting records where a dominant temperature signal is expected (i.e. going near the northern or elevational tree-line), which can then be tested with empirical correlations.

Rob can say more about the database collation, since I wasn't directly involved in that, but don't get the impression that we have selected some special 54 from thousands of candidate records -- that would be the wrong impression. The pool of long records that have already been published with a temperature interpretation is simply not that big, especially when "duplicates" are removed (i.e. where a later update is available, perhaps with more data or alternative methods of removing the age signal). See the comment Rob just posted.

4. Difference from CET.

Why not? CET vs a mean over extra-tropical NH land will show some differences, regardless of CET reliability.

Whether CET is better than tree-rings is also a moot point, since we don't have CET-like records over the NH for >1000 years. Probably I'd prefer thermometers in unheated rooms if we had them. Or use both and compare. :-)

5. CO2 fertilisation effects.

We don't identify or remove such effects. The empirical evidence for sustained effects (i.e. over decades) on trees in cool, moist locations over long periods of time is scarce.

Jan 14, 2016 at 5:36 PM | Unregistered CommenterTim Osborn

Rob

It appears you "calibrated" the series using temperature data with the assumption that tree ring width was proportional to temperature with little error.

Basically you and your authors believe that tree ring width ~ f ( temperature). So no other factors have significant effect on a tree's growth?

Jan 14, 2016 at 5:59 PM | Unregistered CommenterMicky H Corbett

Tim
I have reconstructed CET to 1538 using observational proxies and science papers and it seems to stack up well with dutch research.

I am currently working on the 13th century version.it seems on the whole rather cool wet and stormy but with a notable heatwave around 1249.

I have been in contact with David parker and Dr McCarthy at the met office and they suggested I should contact you specifically regarding trying to update lambs south westerly winds index for England from 1340 but which ends in 1975. I wish to update it and wonder if you have the information to hand as the annual days of south westerly winds that lamb devised does not currently seem to have a direct equivalent these days.any suggestions?

tonyatclimatereasondotcom
thanks
Tonyb

Tonyb

Jan 14, 2016 at 6:13 PM | Unregistered Commentertonyb

That doesn't look like a hockey stick, and more like the corrected Pages 2K.

Brandon, it could be that there are works from the authors not because the authors pushed their own works, but because their works caused them to be authors.

Jan 14, 2016 at 6:18 PM | Unregistered CommenterMikeN

Large author list, but there are some prominent names missing...

Jan 14, 2016 at 6:21 PM | Unregistered CommenterMikeN

Sorry to bang on about this: If TR proxies were = to thermometers then studies based on them would surely offer the same outcomes. For example, I venture that a second CET thermometer would record similar readings to the first within a margin of error. In that respect - and given how much credence was given to Mann's HS - all other studies based on TRs should surely come to the same (not just similar) conclusion. But I don't see that here.

Jan 14, 2016 at 6:43 PM | Unregistered CommenterHarry Passfield

"However you can explain why you think tree rings have more scientific merit that CET"

Wider geographical distribution.

Jan 14, 2016 at 7:07 PM | Unregistered CommenterEli Rabett

So, The Church of Climatology has a Hockey Schtick resurrection story?
They appear to have 'fixed' the tree ring divergence problem, but it beats me, why use 'proxies' when there are adequately useful ways to reliably and validly measure temperature directly?
Of course the 'N' Trend has an exciting, impersonal pseudo-scientific ring about it, freed at last from the haunting debacle of MBH, designed entirely for the MSM, UN et al. to peddle and paraphrase sounding very knowledgeable to their wider congregation, eg. "we're N-Trending our way to whatever ... selected cause célèbre du jour."

Jan 14, 2016 at 7:19 PM | Unregistered CommenterManfred

Empirical correlations is a weak argument for creating a proxy. If you get 0.75 as your r2 what does that actually transfer to in terms of uncertainty? 20%? 25%? You need to be within 5% to be considered useful. That's a standard metric in engineering.

And how does the tree ring grow with regards to temperature, which is typical a temperature range above some threshold? If it saturates at say 25 degrees how do you tell if its 24 or 26 degrees without extensive lab characterisation.

And how accurate is your temperature reference? ± 0.1 degrees as per temperature anomalies? That value is the result of assuming source data uncertainty distributions not of repeatable characterisation.

There are so many metrology issues with trying to produce a tree ring width v temperature relationship, and a lot of confounding factors that any result is only valid academically.

That may be good for the field but it's not useful for real world applications.

Jan 14, 2016 at 7:24 PM | Unregistered CommenterMicky H Corbett

"This strategy explicitly incorporates the expert judgement the original authors used to derive the most robust reconstruction possible .... ". What a beautiful ringing phrase!! I wish I had used it in my thesis instead of the mundane: "According to Bloggs et al .... ".

Separately, I thought the whole point about "hiding the decline" was not the sleight of hand itself but the fact that the measured temperatures and the "TR" data diverged significantly enough to make the measured temperature graft necessary to tell the story. Hence the conclusion that TR data from past pre-thermometer eras could not be relied upon.

Jan 14, 2016 at 7:32 PM | Unregistered CommenterBudgie

Tonyb,

we use the Jenkinson & Collison (1977) scheme to calculate "objective" Lamb Weather Types (LWT) using reanalysis data for 1871 to present. See the second link in the first table of the Jenkinson section of our Lamb Weather Types page. i.e. this file: 20CR_1871-1947_ncep_1948-2016_12hrs_UK.csv

This is a csv file that should open in a spreadsheet. The final column is the LWT for each day (see the table on the above webpage). 15 are SW days. 5 and 25 are Anticyclonic and Cyclonic SW days. You could make counts of these three types for each month, season or year and see how they compare with the Lamb data during the overlap period. I don't know if Lamb used only pure SW days, or also included the hybrid ASW and CSW days. If you get a good match, you can use our counts to bring the series right up to date.

Jan 14, 2016 at 7:35 PM | Unregistered CommenterTim Osborn

Tonyb,

I meant to add that by "we" here, I mean CRU rather than me or N-TREND.
Jones, Harpham & Briffa (2013), referenced on that LWT webpage.

Jan 14, 2016 at 8:00 PM | Unregistered CommenterTim Osborn

The raving madness of climate heathens.

Temperatures have gone up. Stop freaking out about it.

The question is why.

Now, DiCaprio answers that humanity is fault. He owns one of the most expensive yachts in the world and sleeps with Victoria Secret models. The "scientist" Nic Lewis (for example) answers that its not humanity but he has never seen a yacht or a model.

It is evident who is correct and who is not.

I rest my case.

Jan 14, 2016 at 8:32 PM | Unregistered CommenterAila

Eli Rabbett:

Wider geographical distribution
Oh dear. You've heard one joke you've heard yamal.(Apologies to Josh)

(re: Tree rings are more accurate than CET)

Jan 14, 2016 at 8:36 PM | Unregistered CommenterHarry Passfield

Aila = Dork? (It's as good a fit as tree rings). Sheesh.

Jan 14, 2016 at 8:40 PM | Registered CommenterHarry Passfield

I know blogs are not, traditionally, the place that scholars are prepared to respond to criticism but I warmly applaud the authors for taking time to comment here. I look forward to reading the paper and the debate it will surely create.

Jan 14, 2016 at 8:48 PM | Unregistered CommenterProfessor Bob Ryan

I, too, would like to thank Tim Osborn and Rob Wilson for their participation here. It's much appreciated by all, I'm sure.

Jan 14, 2016 at 9:33 PM | Unregistered Commenterjorgekafkazar

Tim
Many thanks for this. I will try to correlate the data from the indexes. I do not know why lamb apparently used south westerly rather than any westerly type wind which are likely to be of a warm and wet nature:.

Tonyb

Jan 14, 2016 at 9:59 PM | Unregistered Commentertonyb

"They appear to have 'fixed' the tree ring divergence problem, but it beats me, why use 'proxies' when there are adequately useful ways to reliably and validly measure temperature directly? "

Before 1700? or even somewhat later? With good global coverage?

One of the mysteries of existence is why the Chinese never came up with good thermometers early on. Would have made things a lot easier.

Jan 14, 2016 at 10:02 PM | Unregistered CommenterEli Rabett

To begin with, my thanks to Rob for showing up to defend his work. He's one of the good guys. Also, my thanks to him and the authors for archiving at least the final form of the data used.

With that said, I have real problems with a few things. First is the selection of proxies. For a multi-proxy study to have a hope of being valid, it needs to have clear ex-ante criteria for the proxy selection, to avoid "data-snooping". I can find no record of such criteria. And this is doubly true when doing a "meta-analysis" of previous studies, as in this case.

Next is the blind acceptance of "the expert judgement the original authors". I have several real issues with that process.

One is the same general problem this study has—the underlying studies mostly have no clear ex-ante criteria for the proxy selection. In many cases, lots of data got left on the cutting room floor.

In addition, a careful count shows that out of the 53 previous proxy studies, no less than 48 of them are by one or more of the current authors ... I'm sure they trust their own expert judgement, but I'm less sanguine.

However, I also have more fundamental problems with the mystical transmutation of tree-ring width data into temperature estimates.

First, tree-rings are often out of agreement with themselves—cores taken from one side of the tree may have little to do with cores taken from the other side.

Next, without knowing the tree's natural history, we don't know if, how, or when the surrounding conditions changed. Neighboring trees fall, land slides, forests burn, insects selectively knock back some species and encourage the growth of other species ... and all of these affect tree ring widths.

Next, it is rare to see any tree-ring study of temperatures that even attempts to disentangle the temperature and the moisture signals. Remember that tree rings were first used to chronicle droughts ...

The researchers seem to think that this can be avoided by picking trees from certain areas like mountaintop tree lines. But again, this gets back into the problem of the lack of ex ante proxy selection criteria. The field of tree rings is rife with data snooping. Heck, it got so bad that it was touted as an advantage, that we could find and select "climate-sensitive" trees by the fact that they correlated well with the surrounding temperature record ...

Finally, the response of any plant to temperature is NOT a straight line. Instead, it is shaped like an upside-down "U". It's the old Goldilocks phenomenon. Some temperatures are too warm for any given plant, some too cold, and some temperatures are just right. So at both high and low temperatures we get stunted growth (narrow tree rings), with high growth (wide rings) only occurring in between the too hot and too cold ends.

This means that when we see a narrow ring in a given tree, we have at least four major environmental choices as to the reason:

1. The year was too cold. This is the assumption that is followed in all tree ring reconstructions, to the exclusion of all the others.

2. The year was too hot.

3. The year was too dry.

4. The year was too wet.

Again, the researchers seem to think that this can be avoided by choosing trees from the upper tree lines, where they think that the only issue is temperature. What they are overlooking is the interaction of temperature and moisture at the hot end of the scale.

If a plant has plenty of water, it can cool the leaves through transpiration. It also has adequate water for the internal process of the tree to function. As a result, it is able to continue to grow strongly in warm or hot periods.

But when water is short, it doesn't take much of a temperature rise to significantly slow down the growth process. And the upper tree line is generally somewhat arid even on a good day. As a kid we lived up near the tree line, and I used to see the trees suffering heat stress in dry summers, with the leaves at the tips of the limbs turning brown even though it was not all that warm.

So being up near the tree-line is not some magic bullet that eliminates the confounding moisture signal.

There is one more huge issue. In a generally warming period such as say 1850-present, almost any method for translating tree-rings (or any other proxy) into temperature will mine for hockeystick. Suppose I take a thousand random pseudo-proxies which represent yearly data of some sort for the last 1000 years. I use some method (e.g. linear regression) to transform each of them to best match modern temperatures from 1850 to the present, and average them.

Now, here's the thing. The've all been transformed based on their "calibration" with post-1850 temperatures. In general, that's the only way to convert a proxy measurement to a temperature measurement. So the average of the result of these random pseudo-proxies will increase from 1850 onwards.

But the period prior to 1850 will be totally random, so it will average to a straight line. We end up with an average which is basically a straight line prior to 1850, and then shows rising "temperatures" after that ... which might sound familiar, as it is also known as a "hockeystick" shape.

As a result of all of those issues, I gotta say that I am not going to put any weight on this study at all.

Regards to everyone,

w.

Jan 14, 2016 at 10:14 PM | Unregistered CommenterWillis Eschenbach

There is no point in using just any old statistics, as these could well end up producing the wrong conclusion. True evidence based science requires careful selection that will produce an outcome in line with funding and grant requirements. Unhappily in this case the resulting graph does not really look much like the desired hockey stick, but funds are now available for more work this year to refine the results.
As the Abstract says:
"Future dendroclimatic projects should focus on developing new long records from data-sparse regions such as North America and eastern Eurasia as well as ensuring the measurement of parameters related to latewood density to complement ring-width records which can improve local based calibration substantially."

Jan 14, 2016 at 10:56 PM | Unregistered Commenternicholas tesdorf

looking at the graph, at what time was the climate not changing? what caused temps to increase from 1800 onward? how come the big drop in temps around 1960 show up on the graph, also in the news of the time, but not on today's official temperature records?

Jan 14, 2016 at 11:03 PM | Unregistered Commenterferd berple

Thank you to Rob Wilson for referring people to the data, which is easily accessible.
There has been a lot of work gone into the compilation of the data set. What is remarkable is the 53 data sets from 1710 to 1988. Inevitably the further one goes back, the sparser the data going further, but there are still 23 data sets for the year 1000 and 30 for 1200 and 42 for 1510.
One point is the fall-off after 1988. In 1990 the count is 33, in 2000 24, in 2010 7 and 2011 just 3. A crude average of 2011 is 3.30, against 1.63 for 2010 and 1.79 for 2003 (the next highest). That said, the final M-Trends does not appear to include it. A crude 15 year centred moving average is excluding 2011 data is very close to the final N-TREND2015 graph. The only major differences are that N-TREND2015 seem to clip the C9th and C15th warming spikes and the C15th cooling spikes. No doubt this is due to adjustment for outliers.
Since 1900 the graph also very closely replicates the Gistemp Northern Hemisphere anomaly for 90N-23.6N. http://data.giss.nasa.gov/gistemp/graphs_v3/Fig.B.gif
The visual main difference is that in N-TREND2015 the post 1975 warming is about the same magnitude as the early twentieth century warming, whilst in Gistemp it is almost twice as large.

Jan 14, 2016 at 11:28 PM | Unregistered CommenterKevin Marshall

Is it just me, I seem to be getting an awful, very uneasy feeling of deja vu?

McIntyre therefore prepared a revised dataset, replacing Briffa's selected 12 cores with the 34 from Khadyta River. The revised chronology was simply staggering. The sharp uptick in the series at the end of the twentieth century had vanished, leaving a twentieth century apparently without a significant trend. The blade of the Yamal hockey stick, used in so many of those temperature reconstructions that the IPCC said validated Michael Mann's work, was gone.

Scroll down, for end figure and graphic.

Jan 15, 2016 at 12:14 AM | Unregistered CommenterAthelstan.

Kevin, one is limited on the current end by when the core was taken. The paper is a compilation of previously published results. It's a feature not a bug

Jan 15, 2016 at 12:21 AM | Unregistered CommenterEli Rabett

Mickey h corbett

I think that Craig Loehle managed to publish a paper through the inquisition about 10 years ago, which covers some of your areas of interest. The Jim bouldin blog might help you locate these heretical findings

Jan 15, 2016 at 12:48 AM | Unregistered CommenterDiogenes

Rob
respectfully


why so many co-authors ?

Jan 15, 2016 at 1:00 AM | Unregistered CommenterEternalOptimist

" .......The paper is a compilation of previously published results. It's a feature not a bug....."

Unless the bug is embedded in the previously published results?

Kevin Marshall, many thanks for your work. The following is of particular interest:-

.....In 1990 the count is 33, in 2000 24, in 2010 7 and 2011 just 3.

Ash die back?

PS, over the last few years our Ash (its her name) appears to be really enjoying life! Has no paleoclimatology trepanning scars, long live the old girl! Keeps us cool in rare full sun days and in awe of her ability to look scary during winter!

Jan 15, 2016 at 1:03 AM | Registered CommenterGreen Sand

RW, TO,

Thank you for yet another worthless piece. "Robust"? I think not.

A shameful waste of public money.

"The residual error of the local scale calibration will represent “other” factors influencing growth, but these influences are random in space and time (easily testable by examining the residuals)."

I could make a good suggestion as to where you might look for the "residuals".

Jan 15, 2016 at 1:41 AM | Unregistered Commenterjolly farmer

EternalOptimist, I guess that people whose tree-ring reconstructions were used were added as co-authors.

Jan 15, 2016 at 2:02 AM | Unregistered CommenterMikeN

Jan 14, 2016 at 10:02 PM | Eli Rabett aka Dr. Joshua Halpern

As you well know but disingenuously chose to ignore, the reference to direct temperature measurement was made wrt to the infamous Schtick as in:

So, The Church of Climatology has a Hockey Schtick resurrection story?

Pulpit pounding is poor practice.

Jan 15, 2016 at 2:55 AM | Unregistered CommenterManfred

“Many (though not all) of the tree-ring records we use were selected from sites near to the northern or elevational tree-lines …”.
==========================
It beats me why tree-rings are preferred temperature proxies as opposed to tree-lines.
There are many tree-line studies at both high latitudes and high elevations indicating a MWP, for instance G.M MacDonald, K.V Kremenetski, D.W Beilman: “… conifers have not yet recolonized many areas where trees were present during the Medieval Warm period (ca AD 800–1300) or the Holocene Thermal Maximum (HTM;ca 10 000–3000 years ago) …”.
http://rstb.royalsocietypublishing.org/content/363/1501/2283

Jan 15, 2016 at 3:15 AM | Unregistered CommenterChris Hanley

Willis says:

This is what concerns me in this sort of study. I've spot checked a few series so far: Gulf of Alaska, Labrador, Thelon, Coppermine and in each case, I've noticed that Wilson et al 2016 has purported to reduce divergence problems by replacing earlier series with alternate versions that go up more. There are a number of newish series and, as too typical with dendros, for many of these series, the underlying measurement data is unarchived. It took nine years for measurement data for D'Arrigo et al 2006 to be archived; hopefully the time will be shorter this time.

The long delays in archiving data for D'Arrigo et al 2006 meant that everyone had pretty much forgotten about the paper, by the time that the measurement data was made available. I have some interesting notes on the NWT series of D'Arrigo et al 2006 (Coppermine, Thelon), which I'll try to write up while I'm revisiting this, as there is some overlap between Darrigo et al 2006 series and Wilson et al series.

Jan 15, 2016 at 3:37 AM | Unregistered CommenterSteve McIntyre

"However you can explain why you think tree rings have more scientific merit that CET"

Wider geographical distribution.

Jan 14, 2016 at 7:07 PM | Unregistered CommenterEli Rabett

++++
Eli Rabett - are you saying that CET can buck global trends over 400 or so years? If so, what geographical (or other) features do you think England has which would keep it significantly out of step with the rest of the planet?

Jan 15, 2016 at 3:48 AM | Unregistered Commenternot banned yet

PostPost a New Comment

Enter your information below to add a new comment.

My response is on my own website »
Author Email (optional):
Author URL (optional):
Post:
 
Some HTML allowed: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <code> <em> <i> <strike> <strong>