Image Credit: Photobucket.com – GISP2 – Alley, 2000
By WUWT regular “Just The Facts”
In building WUWT Paleoclimate Reference Page during these crowdsourcing threads (1, 2) there have been a number disputes raised about various graphs. During this thread I am hoping that we can determine logical classifications for graphs based upon Alley, 2000. In future crowdsourcing threads we will have to address an array of other disputed reconstructions including Briffa et al., 1998, Jones et al., 1998, Mann et al., 1998, Pollack et al., 1998, Jones et al., 1998, Mann et al., 1999, Mann et al., 2000, Briffa et al., 2001, Esper et al., 2002 and Jones and Mann 2004, AR4 section 6.6.1.1 2007 and Marcott et al. 2013.
Thus far three methods have been used for classifying disputed graphs on the WUWT Paleoclimate Reference Page. The first is to add a “Graph Background” link to offer additional pertinent details. The second is to add a Disputed Graph label to the graph, along with a brief description of the nature of the dispute. The third is to place the graph in the section at the bottom of the Paleoclimate page titled Incorrect/Falsified Graphs along with a longer description of the reason the graph is incorrect or has been falsified. Your suggested improvements to our graph classification methodology are most welcome.
I this thread I’d like to solidify the classifications for the graphs based upon Alley, 2000. The dispute around Alley, 2000 has focused on the axis labels, as WUWT commenter Phil states here
Any graph that claims to use Alley’s GISP2 data must either finish at 95 years Before Present (BP=1950) or AD1855 because that is the final date in his database which is on-line and freely available to us all. Lappi’s graph mistakes Present for 2000 as does Easterbrook, they should have a note added pointing out their error or be excluded.
Based on Alley’s own Figure 1 from his 2000 paper;
the x axis label is Age (thousand years before 1950) or Years Before Present (1950AD) and the Alley, 2000 data clearly ends 0.0951409 thousand years Before Present. Taking that into account, the following are the graphs from the WUWT Paleoclimate Reference Page know to be based upon Alley, 2000 and their proposed Classifications:
2,500 Years – GISP2 – Alley, 2000 Classification: Correct – No Change
10,000 Years – GISP2 – Alley, 2000 Classification: Correct – No Change
10,000 Years – GISP2 – Alley, 2000 Classification: Correct – No Change
10,000 Years – GISP2 – Alley, 2000, Ljungqvist et al and HadCRUT3 Classification: Correct – No Change
10,000 Years – GISP2 – Alley, 2000 Classification: Confusing – Add Label: Disputed Graph – The x axis label should read Years Before Present (1950 AD)
10,000 Years – GISP – Alley, 2000 – Vostok – Petit et al. 1999 – Click for Animation – Classification: Confusing – Add Label: Disputed Graph – The x axis labels should read Years Before Present (1950 AD)
10,000 Years – GISP2 – Alley, 2000 Classification: Incorrect – Move to Incorrect/Falsified Graphs section – Label Incorrect Graph – The x axis label, “Years Before Present (2000 AD)”, should read Years Before Present (1950 AD)
10,000 Years – GISP2 – Alley, 2000 Classification: Incorrect – Move to Incorrect/Falsified Graphs section – Label Incorrect Graph – The x axis label, “Years Before Present (2000 AD)”, should read Years Before Present (1950 AD)
Please provide your thoughts and recommendations on these proposed graph classifications and labels for the Alley graphs, as well as any other improvements, corrections or other additions to the WUWT Paleoclimate Reference Page in comments below.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
The problem I have is using Greenland cores or Greenland temperature as a representation of global climate. Greenland is subject to extremely local weather events, some of which can be persistent on a decadal scale. Changes in persistent pressure areas and jet stream flows can cause significant local changes in Greenland temperatures. In fact, I saw a paper recently that compared O18 in Swiss speliothems with Greenland ice cores. While there were some correlations there were also some cold periods in Greenland that didn’t show up in Switzerland and even when they did show up together there was a considerable lag of time between them.
I’m just skeptical of using Greenland ice cores because there is no way to sort out local changes from global changes.
The idea of vetting and classifying seems to be good one. However, assimilating a graph covering 10k years with a careless error of 50 years “before present” issues with something that is falsified seems unreasonable and improper.
The axis labelling issue is worth flagging but this does not fall into the same category as falsification.
Things like Jones and Mann cropping , padding , grafting etc is clearly falsification.
It would be to dilute their sins to confound that will a rather insignificant technical error of not correctly noting what “before present” is defined as.
Yep, I like the Years Before Present (xxxx AD) much clearer for these kinds of graphs.
I have always been interested in all the creative ways scientists have gone about trying to infer the temperatures of the past, whether “relatively” recent or ancient. And given that; I would hope any would understand why such minor variations (above) are simply a “noise factor”, given; the flucuations over the “four” seasons for a particular longitude, latitude and altitude.
When I was young, a change in climate simply meant that the seasons were changing from spring to summer or from fall to winter, and for some of us that was “dramatic” enough.
How about changing the the x-axis labels to be correct.
Reblogged this on This Got My Attention and commented:
Disputes among “climate scientists”?
NOAA graph should be labelled as misleading for plotting temperature record along side and as continuation of GISP2.
The two datasets are incompatible and cannot be compared like that. The time resolution of the sampling interval is not the same and there is a physical averaging process on a decadal scale in ice/firn closure.
Plotting the two on the same graph clearly is intended to make that comparison and is thus misleading and not scientifically valid.
The only one that may pass the test is the second one showing 2500 y GISP2. But …
2500 y GISP2 , best of the bunch but though the 10 year running mean looks to have similar variance to the GISP data this is presumably more by luck than judgement and it still has unsmoothed data which is not compatible.
Also running mean appears to run up to the end of the data which means it’s been padded with something. Likely by duplicating last value. This will bias strongly towards one final value this is not representative of the data and also not compatible for reasons in previous post.
So why all these graphs?
The only thing I can see as being legit is a straight plot of Alley’s data.
If someone has studies the physical time constant of the damping effect of closure time and done an equivalent processing on temp record, that may be legit. I don’t see that being done anywhere here .
Gerg Goodman says: April 13, 2013 at 12:25 pm
The idea of vetting and classifying seems to be good one. However, assimilating a graph covering 10k years with a careless error of 50 years “before present” issues with something that is falsified seems unreasonable and improper.
The axis labelling issue is worth flagging but this does not fall into the same category as falsification.
Things like Jones and Mann cropping , padding , grafting etc is clearly falsification.
It would be to dilute their sins to confound that will a rather insignificant technical error of not correctly noting what “before present” is defined as.
But the graphs are incorrect, regardless of the the method and motivations. And 50 years is quite important when the warming of the second half of the 20th century forms the entire basis of the Catastrophic Anthropogenic Global Warming narrative.
Once we start to deal with Jones, Mann, Marcott, etc. we can consider whether it makes sense to break out the Incorrect/Falsified Graphs section into multiple sections, but for now I am inclined to leave it as is. What does everyone else think?
Splicing the Northern Hemisphere HadCRUT3 record onto the GISP2 reconstructions is awkward. If there is any Arctic amplification, one might expect temperature change in Greenland to be greater than the hemispheric mean.
And what is the reference for your identification of Bond peaks and troughs. They don’t exactly match the cycles Bond identified.
Greenland instrumental record is almost equal to the whole Arctic, that is very similar to CET and CEt is almost identical to NH. Today Greenland temperature is similar to 1930-40s.
One thing that seems a serious omission on the paleo page is the origin of all these plots. Many have a link to Alley’s paper but that is not the origin of the graph.
Anything provided on that page should have a clear and prominent link to the source of the graph.
Alley 2000 is flawed because the isotope paleothermometer is flawed. Every meteoroligst, knows that the most important factor that determines the “cloud temperature” is dew point. But it’s the cloud temperature and rayleigh rain out effect that is ultimately determines the isotope ratios on the ice cores. Hence we’re essentially looking at dewpoints, not at temperatures.
If you compare the i snow accumulation at the ice cores and the isotopes, it’s a near perfect match r2 96% that is simply because both are proxy of the same, wet or dry climate.
More here:
https://dl.dropbox.com/u/ 22026080/non-calor-sed-umor. pdf.
fixing the link (sorry) https://dl.dropbox.com/u/22026080/non-calor-sed-umor. pdf.
Steve Keohane says: April 13, 2013 at 12:30 pm
How about changing the the x-axis labels to be correct.
I am all for it. Craig Loehle, Ph.D. and J. Huston McCulloch, issued a “Correction to: A 2000-YEAR GLOBAL TEMPERATURE
RECONSTRUCTION BASED ON NON-TREE RING PROXIES stating that:
Craig Loehle even did a guest post on WUWT:
http://wattsupwiththat.com/2010/09/28/loehle-vindication/
Their corrected 200 year graph appears towards the top of the Paleoclimate page without label and the incorrect version appears at the bottom in the Incorrect/Falsified Graphs. If the authors of the incorrect Alley graphs would issue corrected versions of their graphs then we would handle them the same as we have the correction to Loehle, 2007.
Sorry some program put in spaces:
https://dl.dropbox.com/u/22026080/non-calor-sed-umor.pdf.
crosspatch says:
April 13, 2013 at 12:22 pm
“The problem I have is using Greenland cores or Greenland temperature as a representation of global climate. Greenland is subject to extremely local weather events, some of which can be persistent on a decadal scale.”
Isn’t the warming of a warm period exaggerated in Arctic-subarctic latitudes?
If I may be so bold. Every data set needs to be °K with the following parameters:
Resolution: Years for which individual measurements span. Is Marcott 300 years? ±150 years or what?
Graph span: A graph that starts 11k BPE and draws trend lines might need to discard several hundred years at both ends for the purpose of trend graphing.
Reference temperature: The temperature against which differences are measured. This makes for several important tests. One, if the reference temp can be verified by other methods. Second if better amplification factors can be deduced. Rather than just tossing out a dataset it can perhaps be rescued.
Resolution, span, reference point.
Gerg Goodman says: April 13, 2013 at 1:02 pm
NOAA graph should be labelled as misleading for plotting temperature record along side and as continuation of GISP2.
The two datasets are incompatible and cannot be compared like that. The time resolution of the sampling interval is not the same and there is a physical averaging process on a decadal scale in ice/firn closure.
Plotting the two on the same graph clearly is intended to make that comparison and is thus misleading and not scientifically valid.
Good point, and the graph;
http://wattsupwiththat.files.wordpress.com/2012/10/gisp2_10ke1.jpg
already has a Disputed Label due to the axis label. Thus I am inclined to move this one to the Incorrect/Falsified Graphs
So why all these graphs?
Facilitates the process of elimination.
The only thing I can see as being legit is a straight plot of Alley’s data.
If someone has studies the physical time constant of the damping effect of closure time and done an equivalent processing on temp record, that may be legit. I don’t see that being done anywhere here .
Perhaps we need to add a Disputed label to all of the graphs where disparate data sets are spliced together highlighting the issues associated with this approach?
Gerg Goodman says: April 13, 2013 at 1:31 pm
One thing that seems a serious omission on the paleo page is the origin of all these plots. Many have a link to Alley’s paper but that is not the origin of the graph.
The “origin” is listed under each graph, but many currently list where the image was found/hosted, not necessarily who created the image. Part of the goal of this crowdsourcing exercise is to identify the provenance of these graphs. Any help you can provide on this front would be most appreciated.
Anything provided on that page should have a clear and prominent link to the source of the graph.
Every graph is linked, just click on the pic and it will take you to where it was found/hosted.
richard telford says: April 13, 2013 at 1:26 pm
Splicing the Northern Hemisphere HadCRUT3 record onto the GISP2 reconstructions is awkward. If there is any Arctic amplification, one might expect temperature change in Greenland to be greater than the hemispheric mean.
Agree, we need to draft up a warning label for them.
And what is the reference for your identification of Bond peaks and troughs. They don’t exactly match the cycles Bond identified.
I didn’t identify anything, these are all 3rd party graphs. The goal here is to determine if the graphs are accurate or not. As such, please present evidence and links to support your assertion that the Bond Cycle Peaks and Troughs labeled on this graph;
http://i90.photobucket.com/albums/k247/dhm1353/Climate%20Change/Subatlantic_Had.png
“don’t exactly match the cycles Bond identified”?
justthefactswuwt says:
April 13, 2013 at 2:14 pm
I didn’t identify anything, these are all 3rd party graphs. The goal here is to determine if the graphs are accurate or not. As such, please present evidence and links to support your assertion that the Bond Cycle Peaks and Troughs labeled on this graph;
—————-
Without referencing the source of all of the information, the graph are almost useless. The best place to find the timing of Bond cycles is presumably Bond et al. (1997). The last two cold events listed are at 1.4 and 2.8 kyr BP (no little Ice Age, because of missing core tops).
http://rivernet.ncsu.edu/courselocker/PaleoClimate/Bond%20et%20al.,%201997%20Millenial%20Scale%20Holocene%20Change.pdf
Interesting to read that Briffa, Jones, Vinther et al paper again and look at their table 1.
If we look at the annual temp for Greenland in the decade 1851-60 we have a temp of minus -2.1c. ( annual)
Then if we look at the decade 1981-90 we have an annual temp of minus-2.5c
Then if we look at the decade 1991-2000 we have an annual temp of minus -2.1c.
Alley’s graph ended in 1855 and we can now show above that the period from 1980 to 2000 was 0.3 C colder than the decade 1851-60. That’s a period of twenty years.
Here is that study. http://www.cru.uea.ac.uk/cru/data/greenland/vintheretal2006.pdf The Table 1 is on page 11.
Interesting how the recent winters and springs ( 1980 – 2000 ) were colder than the 1851- 60 period as well.
PLEASE NOTE. This is a post from another blog I wrote about a couple of weeks ago..
How about making more categories, such as ‘Lie’;, ‘Misleading’, ‘Mistaken’ ? I would put the chart with the wrong BP date as ‘Mistaken’.
Seems to me that the graph as a faithful representation of the data is more critical than the correct labelling of its axes. The former determines whether or not the graph is “correct”; the latter is a hygene issue that can be amended.
I do think you have the right categories, only how to partition the graphs into these categories may take a bit of tweaking.
Personally, whether it is labelled so or not, the use of the word “Present” to denote the date 1950 is confusing. Further, other graphs use 2000 etc for “present”. Any permanent repository of graphs ought to have a more absolute standard of labelling. Somehow my eye missed the parenthetic base date in the titles and it took me a while to get oriented to reading them before trying to assimilate the graphic.
This is a lot of work — I’m very glad to see someone putting in the time to straighten it out. Thanks.
Let’s crowd source CG3.