Climate Science Glossary

Term Lookup

Enter a term in the search box to find its definition.


Use the controls in the far right panel to increase or decrease the number of terms automatically displayed (or to completely turn that feature off).

Term Lookup


All IPCC definitions taken from Climate Change 2007: The Physical Science Basis. Working Group I Contribution to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Annex I, Glossary, pp. 941-954. Cambridge University Press.

Home Arguments Software Resources Comments The Consensus Project Translations About Support

Twitter Facebook YouTube Pinterest MeWe

RSS Posts RSS Comments Email Subscribe

Climate's changed before
It's the sun
It's not bad
There is no consensus
It's cooling
Models are unreliable
Temp record is unreliable
Animals and plants can adapt
It hasn't warmed since 1998
Antarctica is gaining ice
View All Arguments...

New? Register here
Forgot your password?

Latest Posts


Confusing Greenland warming vs global warming

What the science says...

This argument uses temperatures from the top of the Greenland ice sheet. This data ends in 1855, long before modern global warming began. It also reflects regional Greenland warming, not global warming.

Climate Myth...

Most of the last 10,000 years were warmer

Even if the warming were as big as the IPCC imagines, it would not be as dangerous as Mr. Brown suggests. After all, recent research suggests that some 9,100 of the past 10,500 years were warmer than the present by up to 3 Celsius degrees: yet here we all are. (Christopher Monckton)

This argument is based on the work of Don Easterbrook who relies on temperatures at the top of the Greenland ice sheet as a proxy for global temperatures. That’s a fatal flaw, before we even begin to examine the use of the ice core data. A single regional record cannot stand in for the global record — local variability will be higher than the global, plus we have evidence that Antarctic temperatures swing in the opposite direction to Arctic changes. Richard Alley discussed that in some detail at Dot Earth last year, and it’s well worth reading his comments. Easterbrook, however, is content to ignore someone who has worked in this field, and relies entirely on Greenland data to make his case.

Most of the past 10,000 [years] have been warmer than the present. Figure 4 shows temperatures from the GISP2 Greenland ice core. With the exception of a brief warm period about 8,200 years ago, the entire period from 1,500 to 10,500 years ago was significantly warmer than present.

This is Easterbrook’s Fig 4:


It’s a graph he’s used before, in various forms, almost certainly copied and altered from the original (click image below to see source: the NOAA web page for Richard Alley’s 2000 paper The Younger Dryas cold interval as viewed from central Greenland, though DE credits it as “Modified from Cuffy and Clow, 1997″, misspelling Kurt Cuffey’s name in the process:

Easterbrook continues:

Another graph of temperatures from the Greenland ice core for the past 10,000 years is shown in Figure 5. It shows essentially the same temperatures as Cuffy and Clow (1997) but with somewhat greater detail. What both of these temperature curves show is that virtually all of the past 10,000 years has been warmer than the present.

This is his Fig 5:


Easterbrook plots the temperature data from the GISP2 core, as archived here. Easterbrook defines “present” as the year 2000. However, the GISP2 “present” follows a common paleoclimate convention and is actually 1950. The first data point in the file is at 95 years BP. This would make 95 years BP 1855 — a full 155 years ago, long before any other global temperature record shows any modern warming. In order to make absolutely sure of my dates, I emailed Richard Alley, and he confirmed that the GISP2 “present” is 1950, and that the most recent temperature in the GISP2 series is therefore 1855.

This is Easterbrook’s main sleight of hand. He wants to present a regional proxy for temperature from 155 years ago as somehow indicative of present global temperatures. The depths of his misunderstanding are made clear in a response he gave to a request from the German EIKE forum to clarify why he was representing 1905 (wrongly, in two senses) as the present. Here’s what he had to say:

The contention that the ice core only reaches 1905 is a complete lie (not unusual for AGW people). The top of the core is accurately dated by annual dust layers at 1987. There has been no significant warming from 1987 to the present, so the top of the core is representative of the present day climate in Greenland.

Unfortunately for Don, the first data point in the temperature series he’s relying on is not from the “top of the core”, it’s from layers dated to 1855. The reason is straightforward enough — it takes decades for snow to consolidate into ice.

And so to an interesting question. What has happened to temperatures at the top of Greenland ice sheet since 1855? Jason Box is one of the most prominent scientists working on Greenland and he has a recent paper reconstructing Greenland temperatures for the period 1840-2007 (Box, Jason E., Lei Yang, David H. Bromwich, Le-Sheng Bai, 2009: Greenland Ice Sheet Surface Air Temperature Variability: 1840–2007. J. Climate, 22, 4029–4049. doi: 10.1175/2009JCLI2816.1). He was kind enough to supply me with a temperature reconstruction for the GRIP drilling site — 28 km from GISP2. This is what the annual average temperature record looks like (click for bigger version):


I’ve added lines showing the average temperatures for the 1850s (blue) and the last 10 years (red), and the difference between those is a warming of 1.44ºC. I’ve also added the two most recent GISP2 temperature data points (for 1847 and 1855, red crosses). It’s obvious that the GRIP site is warmer than GISP2 (at Summit Camp). The difference is estimated to be 0.9ºC on the annual average (Box, pers comm).

Let’s have ago at reconstructing Easterbrook’s Fig 5, covering the last 10,000 years of GISP2 data. It looks like this (click for bigger version):


The GISP2 series — the red line — appears to be identical to Easterbrook’s version. The bottom black line shows his 1855 “present”, and it intersects the red line in the same places as his chart. I’ve added a grey line based on the +1.44ºC quantum calculated from the GRIP temperature data, and two blue crosses, which show the GISP2 site temperatures inferred from adjusted GRIP data for 1855 and 2009.

Two things are immediately apparent. If we make allowance for local warming over the last 155 years, Easterbrook’s claim that “most of the past 10,000 [years] have been warmer than the present” is not true for central Greenland, let alone the global record. It’s also clear that there is a mismatch between the temperature reconstructions and the ice core record. The two blue crosses on the chart show the GISP site temperatures (adjusted from GRIP data) for 1855 and 2009. It’s clear there is a calibration issue between the long term proxy (based on ∂18O measurement) and recent direct measurement of temperatures on the Greenland ice sheet. How that might be resolved is an interesting question, but not directly relevant to the point at issue — which is what Don Easterbrook is trying to show. Here’s his conclusion:

So where do the 1934/1998/2010 warm years rank in the long-term list of warm years? Of the past 10,500 years, 9,100 were warmer than 1934/1998/2010. Thus, regardless of which year ( 1934, 1998, or 2010) turns out to be the warmest of the past century, that year will rank number 9,099 in the long-term list. The climate has been warming slowly since the Little Ice Age (Fig. 5), but it has quite a ways to go yet before reaching the temperature levels that persisted for nearly all of the past 10,500 years. It’s really much to do about nothing.

1855 — Easterbrook’s “present” — was not warmer than 1934, 1998 or 2010 in Greenland, let alone around the world. His claim that 9,100 out of the last 10,500 years were warmer than recent peak years is false, based on a misunderstanding or misrepresentation of data.

The last word goes to Richard Alley, who points out that however interesting the study of past climate may be, it doesn’t help us where we’re heading:

"Whether temperatures have been warmer or colder in the past is largely irrelevant to the impacts of the ongoing warming. If you don’t care about humans and the other species here, global warming may not be all that important; nature has caused warmer and colder times in the past, and life survived. But, those warmer and colder times did not come when there were almost seven billion people living as we do. The best science says that if our warming becomes large, its influences on us will be primarily negative, and the temperature of the Holocene or the Cretaceous has no bearing on that. Furthermore, the existence of warmer and colder times in the past does not remove our fingerprints from the current warming, any more than the existence of natural fires would remove an arsonist’s fingerprints from a can of flammable liquid. If anything, nature has been pushing to cool the climate over the last few decades, but warming has occurred.

See also: MT at Only In It For The Gold. My thanks to Richard Alley and Jason Box for their rapid response to my questions.

NOTE: This rebuttal is an edited version of a blog post first published by Gareth Renowden at Hot Topic.

Intermediate rebuttal written by Gareth

Update August 2015:

Here is a related lecture-video from Denial101x - Making Sense of Climate Science Denial


Last updated on 4 August 2015 by MichaelK. View Archives

Printable Version  |  Offline PDF Version  |  Link to this page

Argument Feedback

Please use this form to let us know about suggested updates to this rebuttal.


1  2  3  Next

Comments 1 to 50 out of 102:

  1. Gareth: Kudos on a great article! Please make it more SkS user friendly by adding "Related Articles" and "Further Reading" tabs.
  2. 1. The "last word" from Richard Alley is truly pathetic. It is indisputable that in the past the earth has been significantly hotter than it is today. In those hotter periods there nwas no AGW. I agree that there are "finger prints" in the present but that is irrelevant to the basic that it was warmer in the past. 2. Dr Easterbrook simply illustrates that the earth has been warmer. That is not disputed.My understanding of Dr Easterbrooks research is that he "proves" there is no correlation between between temperature and CO2 levels.
  3. alecpiper@2 1. Richard Alley does not dispute that it has been significantly warmer in the past. Of course in the past the warming wasn't due to AGW (because the was no A to cause the GW), but that doesn't mean the temperature changes were not due to changes in CO2. The greenhouse effect is the same regardles of the source of the CO2. 2. Easterbrook does not "simply illustrate that the Earth has been warmer in the past", he attempts to show that the Earth has been warmer in the past 10,000 years, which is unlikely to be true and is not supported by the evidence. Dr Easterbrook's research cannot prove that there is no correllation between temperature and CO2 levels, simply because it isn't true, you only need to look at e.g. the Vostok ice core data to see that It is well known that CO2 generally lags CO2, but that does not mean that CO2 didn't cause most of the warming as it has generally been a feedback mechanism rather than a forcing. However, there have been occasions where CO2 has been a forcing rather than a feedback, for instance the emergence from the snowball Earth conditions of the late Ordovician. However more appropriate places for discussion of the relationship between CO2 and temperature would be CO2 lags temperature or Does CO2 always correlate with temperature or CO2 was higher in the Ordovician. I'd be happy to discuss these issues with you on a more appropriate thread.
  4. You guys still miss the basic point. Over the past 25,000 years there have been major changes in temperature (up to 14 deg C) over 40 years without any anthropogenic influence. Recent temperature changes pale into insignificance when compared to the past. Studies also show a close correlation between GISP2 (Greenland Ice Core project) and other glaciers. Samples from GICP are representative of world changes. See Dr Easterbrooks readable paper.
    Response: No, you are missing that basic point. For education, see "Climate’s changed before," "It’s only a few degrees," and "It’s not bad," and if you want to argue, do so on those relevant threads, not this one.
  5. alecpiper@2 Here is a hint, you are likely to be taken more seriously if you are able to accept when you are incorrect, for instance the assertion that there is no correllation bewtween temperature and CO2. You are also shifting the goalposts from 10,000 years (i.e. the current interglacial) to the last 25,000 years. Of course there has been significant warming since the height of the last ice age, and of course it was non-anthropogenic. However that does not mean that changes in carbon dioxide were not implicated in the warming. Of course recent temperature changes are small compared to the difference between glacial and interglacial conditions, but they don't have to be that large to cause significant hardship, so that is a silly argument. Mentioning Easterbrook's WUWT article suggests you didn't read the article to which you were responding. You do know that it is Easterbrooks work that is the subject of the article?
  6. The temperature proxy record extracted from the Vostok ice cores appears to show the same thing, a peak about 10,000 year ago, and a gradual cooling trend since then. Look at the right side of this graph. The furthest right sawtooth is our current interglacial. It shows a temperature spike, then appears to show a gradual temperature decline. It also shows that this interglacial is colder than all of the previous interglacials over the past 420,000 years Chris Shaker
  7. Chris: Not all interglacials are created equal. Looking at the past 10,000 years showed an overall decline in temperature until the age of industrialization, when the decline halted and temperatures began their climb We have now equaled the highest temperatures of this interglacial (the period known as the Holocene Maximum or the Holocene Altithermal), with yet more warming in the pipeline; even if CO2 emissions are held to zero for the next 40 years, we're still going to be on the receiving end of another 0.6°C warming. The Yooper
  8. It appears that the modern temperature record being posted above may not be very accurate. The Berkeley Earth Surface Temperature project is incorporating criticism of data collection sites from Anthony Watts and his team ( "NEW - There have been many criticisms of station quality. How can you be sure that your results will be good if you are including stations that do not meet NOAA’s criteria for station quality? One of the elements that we plan to study is temperature records from just the very best sites (as classified by Anthony Watts and his team) contrasted with the poorer sites. We will include this comparison when we release our analysis. Each of our 39,028 sites has also been classified as urban or rural using the map published by the Modis satellite team, and we will use that classification to look for differences. " The most recent post I've found on the topic "In the next picture, Steve shows what Briffa and Osborn did – not only did they truncate their reconstruction to hide a steep decline in the late 20th Century but also a substantial early segment from 1402-1550:" [SNIP] Chris Shaker
    Response: [Dikran Marsupial] Quoted accusation of dishonesty snipped. Lets keep it to the scientific issues only please.
  9. Reply to cjshaker #8 on Surface temperature record thread.
  10. This very misleading graphical debunking of Easterbrook fails to switch to the required anomaly scale instead of an absolute scale to deal with the mismatch between his chosen recent temperature reconstruction and the long ice core. He plots two 1855 temperatures instead of one. If you actually match up the 1855 temperatures, as any sincere effort would require, you get exactly what skeptics claim history is like: a just as hot MWP and a hotter Roman period. Plot "It’s also clear that there is a mismatch between the temperature reconstructions and the ice core record. ... How that might be resolved is an interesting question, but not directly relevant to the point at issue." The way you resolve it is to use anomalies like the pros do. You have left the graphical impression that the present is hotter than ever. At best you've has corrected an error that the ice core ends in 1855 instead of 2000 (or 1950), even though this claim is not in any primary article I can find. I had been posting a GIF animation of the Greenland ice core far and wide and this "debunking" gave me pause as I prepared to yank it from my arsenal based on this post. I may have to edit it a bit now though to increase the instrumental "hockey stick" blade from the animation I have. Hereis another blogger combining two Greenland data sets into a single chart without matching them up properly.
  11. Oh wow, if I now look at my correction of your final chart, now that the modern warming is about twice as much as Easterbrook's, it becomes suggestive that there is a very good chance that recent warming may be a peak that is about to plunge back down, masking greenhouse warming for up to a century or more. That the AMO correlates nearly perfectly with decadal variation in the global average T also suggests there's about 40 years of downswing in the works that will also mask CO2 warming. It looks like it might not have been aerosols after all which caused the last downswing but was due to regular ocean oscillations.
  12. Nik "... there is a very good chance that recent warming may be a peak that is about to plunge back down..." and "...there's about 40 years of downswing in the works that will also mask CO2 warming..." I know you refer to oscillations - but what would be the physics driving such phenomena? At least the physics of aerosol effects on climate are pretty straightforward conceptually (even if the measurements are as difficult as most others are). And if it wasn't the aerosols, how do you account for the physics of excluding them (or double counting them if there's a 'cycle', 40 years or otherwise, with the same characteristics).
  13. "if it wasn't the aerosols, how do you account for the physics of excluding them" They might not have any truly discernibl­e and certainly not modelable cause at all if in fact climate represents a chaotic system on century time scales. And what system might be less expected to be prone to chaos than a huge land/air/o­cean biosphere subject to solar output and orbital variations­? The argument from ignorance of "forcings" is thus even less supportive of CO2 as the only alternativ­e than I had thought. Forcings might not even be required if suddenly the ocean currents shift drastically to alter now heat is released or withheld by them for a century or two at a time. There's too many formal mathematical Platonists in climatology and not enough dynamicists, I suggest. It's assumed that weather is chaotic, but why not climate too? I suddenly have a hilarious vision of a watchmaker riding a bull.
  14. Nik, aerosols "They might not have any truly discernibl­e and certainly not modelable cause at all.." I realise your response slipped straight over to models and maths and never got back to the physics. But I'm still concerned about the physics and observable responses to aerosols. Just looking at the histories of calamitous volcanic eruptions near the equator which spread dust and all manner of gunk across the skies worldwide. We know from written records about the effects on temperatures, colder, and agriculture, ruinous, of these events. Why would sustained, slower, less spectacular, releases of aerosols have different physical characteristics?
  15. NikFromNYC @ 10: 1) The "very misleading" graph shows the actual temperatures as the GRIP site with the small crosses, and the anomaly relative to the GISP2 core with the higher of the two horizontal lines. This can only be misleading to those who read neither the information on the graph, nor the article, which states immediately below the graph:
    "The GISP2 series — the red line — appears to be identical to Easterbrook’s version. The bottom black line shows his 1855 “present”, and it intersects the red line in the same places as his chart. I’ve added a grey line based on the +1.44ºC quantum calculated from the GRIP temperature data, and two blue crosses, which show the GISP2 site temperatures inferred from adjusted GRIP data for 1855 and 2009."
    It is hard to be clearer than that. Further, to be "mislead", a reader would have to also ignore the information in the graph plotting the temperature change at the GRIP site from 1855 to the present. 2) The clarity and openness with the information in the above article contrasts sharply with the practise of deniers. They either treat the temperature rise evident in the GISP2 core just prior to 1855 as the 20th century temperature rise (as does Easterbrook), or add a bar representing global temperature increase from (typically) 1905. The global temperature increase is significantly smaller than the local temperature increase, and comparing the two is inevitably misleading. Curiously, you find neither of those practises misleading, but find a clear debunking of the worst of them to be misleading because it clearly presents the relevant data. You must be using some non standard definition of "misleading" such as "shows my views on climate to be a house of cards". 3) In addition, Easterbrook's graph is is misleading because it treats a regional temperature index as a global temperature index. As just noted, regional temperatures have greater fluctuations in temperature than do global temperatures; a consequence of the fact that regional temperatures do not vary in synch. As can be seen from this graph of Holocene temperature proxies, regional temperature vary widely, but their average shows little variability: Indeed, the GISP2 record (light blue on the chart) shows more variability than most regional proxies, a fact that should be well known to any frequent commentator on climate. Even the average on this chart probably shows more variability than the true global mean temperatures because of the low number of proxies, and because the proxies in this chart have a Northern Hemisphere bias (with half the the proxies coming from the NH extratropics). As clearly indicated on the chart, 2004 temperatures are significantly above the average of even the Holocene Warm Period. Of course, it is rather difficult for you to comment on Easterbrook's misleading practise of treating a regional temperature proxy as a global record given that you do the same thing in your comment. 3) It is a bit rich you commenting on "misleading ... debunking[s]" when you claim that "If you actually match up the 1855 temperatures, as any sincere effort would require, you get exactly what skeptics claim history is like: a just as hot MWP and a hotter Roman period." As can be clearly seen from your plot, modern regional temperatures on the Northern Greenland ice cap are about as much warmer than the MWP as they are cooler than the Roman WP at that location. And honest description, then, would be that modern temperatures are hotter than the MWP at that location, though the Roman WP was hotter, at that location. Of course, globally, both where probably warmer than the 1950's, but cooler than the last decade.
  16. Adelady @14
    "Why would sustained, slower, less spectacular, releases of aerosols have different physical characteristics?"
    Because it is convenient for some to believe so, of course.
  17. How has this typo survived? Paragraph 2, sentence 2 shouldn't "...brief warm period 8,200 years ago..." be "...brief cool period 8,200 years ago...", or am I just losing it?
    Response: You are misreading the graphic. It shows a brief cool period just prior to 8,000 years ago, followed by a brief warm period about 8,200 years ago.
  18. @Tom Curtis #15 What is the citation for the "Holocene Temperature Variations" graphic?
  19. Steve, from another thread - remember that climate changes only with a forcing. The following illustrative image, created by Robert Rohde, helps describe temperature variations since the end of the last glacial stage: The forcings that drove the glacial cycles, and that drove our last deglaciation (~10000 years ago) reached a peak at the Holocene Climate Optimum, about 8,000 years ago. These forces are now operating in reverse, driving a slow trend towards glaciation. Small variations on that slow trend to glaciation have given rise to periods of regional warmth and cooling, popularly known by monikers such as the Roman Warm Perod, the Medieval Climatic Anomaly (in some areas, known as the Medieval Warm Period), and the Little Ice Age. Common causes for these small variations are increased periods of volcanic activity and reduced solar activity. The Little Ice Age was no "Ice Age", but just the latest episode of slightly cooler climate within this overall trend, most noticeable in northern Europe, and associated with both reduced solar activity and increased volcanic activity. When looked at over the whole Holocene, the LIA and MCA are just part of the overall trend towards slightly cooler conditions, largely driven by orbital forcing and most noticeable over NW Europe. But of course, the world is not actually cooling any more... Release of geologically-stored greenhouse gases by humans has given the climate an almighty kick upwards from that slow trend to cooling. As you can see from the inset figure, global temperatures have shot upwards towards the Holocene Opitmum levels, and are on a trajectory to go a very great deal higher than that. At the scale of the main graph, a conservative projection of 2C warming by the end of the century (similar to present warming rates) would have us off teh top of the graph only a couple of pixels to the right of the Y-axis - a nearly vertical rise on this graph scale. It is perhaps the fastest known warming in geological history, much faster than deglacial warming. That is due to the forcing of the CO2 and other greenhouse gases we have injected into the atmosphere. I would second Bernard J's suggestion (other thread) that you spend some time perusing this site for informative articles (Eric linked to a couple on the other thread), and perhaps even reading the linked peer-reviewed papers. That the slow cooling through the Holocene abruptly ended as a consequence of the Industrial Revolution is no coincidence: it was predicted as long ago as the dawn of the 20th Century by Arrhenius, and is a consequence of the inescapable physics of triatomic molecules that do not precipitate out of our atmosphere.
  20. Here is my very crude attempt to recreate the graph to modern times, and I would very much like to hear what mistakes I have done along the way.

    FWIW I'm not a scientist, so there...

    Anyway, the reason for my effort is that I have for some time been a bit annoyed that the current temperature (2009) grafted on the last graph in the text is not directly from the GISP/dome, rather it is derived from GRIP. This IMO makes it a bit weaker as a argument when presented to a layman/'skeptic' even if it does have a sound scientific methodology behind it.

    So after some hunting I found some near surface temp measurements here; from the summit, which AFAIK is where GISP2 was drilled. The station numbers are documented here.

    I threw the data in a spreadsheet and took the average from (after converting each timepoint to decimal) 2005.813 - 2011.813, which gave me -28.3°C. Incidently, and not necessarily meaningfully, the max temperature of the last 3 years are all above 0°C and 'rising'.

    Anyhow. I take that average and add it as a data point in to the Alley record linked in the text as for 2008 and get the following graph:

    The trend is for Alley data only (-9000 - 1880), just as reference.

    I'm sure I have made some grave errors in the process, and would appreciate constructive input. Thanks.

  21. Lanfear @20, the GRIP and GISP2 icecores where both drilled at the Summit.  Hence the difference in temperature as measured at GRIP is a reasonable proxie for the difference  at GISP2.

    Your chart shows the difference between the absolute temperature in 1895 as measured using the GISP2 ice core proxy, and the absolute temperature as measured at a nearby location using the thermometers in the 2000s, ie, the difference between the end of the GISP2 icecore and the higher of the two blue crosses in last graph in the original post.  However, as can be seen in the second last graph, GISP2 temperatures are cold relative to GRIP site temperatures in 1895, so it is far better to take the anomaly as is done in the OP or in the second graph @15.  Even that should be taken as indicative only, rather than an exact measure.

  22. Tom Curtis@21

    Yes, I agree with you that to properly compare (or splice in this case) apples with apples, using GRIP to project the temperature into current time.

    So, is the problem (with my attempt at extending the GISP2) to modern time, that GISP2 is showing a different temperature than what the AWS a the summit is producing, perhaps the actual surface (or should that be the firn) temperature vs. 2m above?

    And to get my facts correct, the proper splicing shows that the current temperature is higher than any point in the MCA, but we have some spikes going above current temperature BCE?

    PS. Is there a reason why this article does not show up in the argument list?

  23. Lanfear @22, AWS?

  24. Tom Curtis@23

    AWS - Automated Weather Station. More information is in the last link of my message @20.

  25. Lanfear @25, I apologize for the delayed response.  It is with good purpose, however.  I have had the good fortune of rediscovering a recent paper (Kobaski et al 2011) showing a reconstruction of GISP2 data using a different method from Alley et al (2000).  That paper, based on the same ice core as used by Alley et al, carries the reconstruction through to about 1990, something they are able to do because of a higher resolution sampling.  Further, they directly compare their reconstruction to the Box 2009 reconstruction of temperatures since 1840 that is used in the main post; and with the AWS temperature record.

    The main post here uses the Box data under the assumption that temperature differences determined by the same method are likely to be correct in terms of relative position, but when compared to records determined by other methods there may be biases (possibly unknown) which distort the result.  Based on this principle, it adds the temperature difference between the decadal average of 1850-59 (corresponding to the end of the Alley et al temperature proxy) and 2000-09 (the present) to the Alley proxy.  Decadal differences are used because of the low resolution of the proxy.

    Kobaski et al discuss the issue directly, saying:

    "Before the two records are combined with the 4000‐year temperature record, it is necessary to make an adjustment, as firn temperatures are colder than air temperatures by 0.2°C to 2.6°C in Greenland [Steffen and Box, 2001] as a result of the surface radiative cooling and inversion as noted above. As the mean difference between the decadal average reconstructed Summit temperatures and the reconstructed Greenland temperature for the 1845–2005 period is 1.75°C, the AWS and reconstructed Summit temperatures are reduced by 1.75°C (Figure 1)."

    (My emphasis)

    So you are correct that the cause of the difference is the difference between firn temperature and the surface air temperature.

    As to the question about GISP2 temperatures and the MWP, that is more complex.  The Kobashi 2011 reconstruction shows greater variability in temperature due to its higher resolution.  As a result it shows some temperatues exceeding modern values in the MWP.  It also shows a significant peak in temperatures around 700 AD which significantly exceed modern values, although that predates the traditional dating of the MWP:

    (Kobashi et al 2011, figure 1.  Click on image for higher resolution.)

    I have attempted to overlay the Alley 2000 proxy data with the Kobaski 2011 reconstruction for comparison.  Doing so clearly shows the difference in resolution.  It shows some differences in the result as well.  Most notably, the 700 AD peak in Kobashi 2011 corresponds to a trough in Alley 2000.  I am certainly not expert enough to say which is likely to be more accurate where the two disagree.

    Despite the differences, the curves track each other reasonably closely closely to show.  The match should put paid to any doubts that the termination of Alley 2000 is around 1855, and that the 1855 temperature in Alley et al is approximately 1.5 C lower than modern temperatures at the same location:

    (Alley 2000 overlayed with Kobaski 2011.  Alignment was achieved based on the axis.  Click on image for higher resolution.)

  26. Tom@25

    "I apologize for the delayed response."

    No worries, as you may have noticed, I take this with a slow pace, what with the holidays and everything.

    "I have had the good fortune of rediscovering a recent paper"

    Ahemm.. And I noticed that that paper already was in my archives, so I should have been aware of it. Problem being that the archives seem to grow more than what I am able to digest, let alone recall.

    "As to the question about GISP2 temperatures and the MWP, that is more complex."

    Yes, I should have been more explicit in this case. The root of this is, of course, the denialist fallacy where he/she makes the mistake of a) taking a single point of (proxy) measurement and then generalizing that to the whole globe as such, and b) not consider the fact that the y-axis is between -32 and -28°C.

    Thanks for your effort, and the reminder of Kobashi et al. 2011.

  27. This is quite a thread, I just spent 45 minutes reading and studying through it.

    Post 15 brings the GRIP2 chart to modern times using information from the GRIP site, by adding 1.44 C to 1855, but doesn't move the chart over to the left 50 years. I did that myself just now, lining up the chart so the end of the GISP2 data correctly lines up over 1855 rather than 1905.

    What I find by doing that is, the modern warming period must have started quite a bit earlier than we normally understand, at least in Greenland. I've been under the assumtion that modern warming started around 1830 to 1840 or so, as many believe. But the data presented on this thread shows 1780 or 1790. The beginning of the modern hockey stick on the revised Easterbrook (editied for 1855-2013 by Tom Curtis here and moved over 50 years by myself (sorry I don't know how to post that here) the temperature appears to bend upwards in the late 1700's, perhaps 1780.

    So, we have a long hockey stick with a 1.44 degree C rise from 1780 to now, with roughly the same slope throughout the whole 230 year period. 1780 is long before humans started making much CO2. This thread debunks the 10,000 claim of Monckton, Esterbrook, etc., but adds a new talking point for the deniers to use.....why did modern warming start in 1780?

  28. Sorry I meant GISP2 not GRIP2. 

    Also, I realized the 1.44 is not the full extent of the rise. That's only 1855 to now. If we go back to the beginning of the bend it's closer to 2 whole degrees C.

    I understand polar temp rise will be higher than global. But I think the 0.8 degree C figure that we always hear for global warming period is low if we include the entire rise, if taken back to 1780 like this. I have many documents, written in the early 1800's, that describe a warming climate, melting glaciers, rising sea levels, etc. in the early 19th century.

  29. steven foster @27 & 28:

    1)  The chart in my post @ 15 does in fact place the end of the GISP2 data in 1874 (based on pixel count), or five pixels prior to 2010.  Placing it six pixels prior to 2010 would have placed it in 1847, a little more accurate but hardly consequential.  I doubt it makes any visual difference.

    2)  I am not sure why you are making this point on the original Alley et al (2000) data when we have the Kobashi et al (2011) data from the same ice core that is:

    • Carried through to a more recent date;
    • At a higher resolution;
    • Includes reconstructed site temperatures from nearby instrumental readings back to 1840
    • Includes on site instrumental readings back to 1985
    • Is correctly alligned.

    Using the Kobashi data we see the warming at the GISP2 site started circa 1750, and was quite slow till the early twentieth century.  It then shows a rapid increase, and decline, followed by another rapid increase of equal size.

    3)  All of this is largely irrelevant because local proxies are not global proxies.  Anybody using the denier talking point you mention must be ignoring that basic fact.  What the Kobashi reconstruction shows us are the temperatures at one site in Greenland.  The do not show us Arctic temperatures in general, and certainly not NH or global temperatures in general.  In fact, we do in fact have NH and global temperature reconstructions over that period, which show the twentieth century warming to be much faster than the prior warming as the Earth exited the LIA.

    Faced with that fact, the question those rejecting AGW need to ask themselves is, why, after the Earth warmed enough to exit the LIA did it keep on warming?  Indeed, why did it not only keep on warming, but warm faster?  If the late twentieth century just represented a "recovery from the LIA", why did it not slow down as it approached the prior equilbrium level, but instead accelerated past it?

  30. I assure you I didn't count pixels or use a graphics program. I plotted GISP2 for the whole Holocene and plotted GRIP on a uniform timeline from 1855 to 2009 by simply starting with the 1857 GRIP temp which was quite a bit warmer than the 1855 GISP2 temp, and plotted from there. I generally don't like to splice different data sets together but this is only a blog so who can it harm? The resolution is low and this is not publishable but gives us something better than the useless chicken scratch graphics I've seen elsewhere. This differs little from what you did on post 15.  My attempt at doing the same thing.

  31. steven foster @30, I did not suggest you counted pixels.  Rather, I did, and by doing so showed the difference to not be appreciable.  For those unsure on that point, here is the graph with the data misplaced on the x-axis as noted by Foster @27:

    And here is Foster's version:

    Can you see a significant difference?

    That aside, you (Foster) do not address my point (2) above, ie, that we should be using Kobashi et al (2011) for investigating north central greenland temperatures.

  32. I see a significant difference in that in my chart, we can see the slope of the line in the 1700s and 1800s is equal to that in the 2000's. Your chart merely connects 2 points and has a straight line from 1855 to 2009, the slope in each century not being discernible.

    I already addressed Kobashi et all in post 30, when I mentioned useless chicken scratch graphics.


    [JH] Please lose the snark.

  33. steven foster @32, the problem is that the GRIP temperature you plotted onto GISP2 is that from Box (2009) as displayed by Kobashi et al.  So, when we compare Box (2009) in Kobashi et al (middle panel, black) with your extension of GISP2 we see the uniform rise in temperature is not a feature of the data, but only a consequence of the very low resolution of your graph.  The data clearly shows a rise in temperature to mid twentieth century, a fall, then a rise again, details that are missing from your graph.  Therefore your extension does not show the details significantly better than the straight line, and the supposition that it might is deceptive.

    And that is why there is no substitute for showing the "chicken scratch graphics".  They actually show the data accurately.  If you want to trim it down, show only the middle panel, but don't pretend that showing a less accurate ice core temperature series together with an instremental based series at two low a resolution to show relevant detail is a substitute for showing an accurate graph.  It isn't.  

  34. Yeah the resolution is really low because all I did was capture Easterbrook's graph and use photoshop to clone the entire thing over by 50 years so it correctly ended at 1855, and then I added the modern part by plotting about 30 data points from GRIP, and then played connect the dots with photoshop. I'm thinking of redoing it by plotting all the data on a spread sheet, but there is a lot of work involved. 

    Yeah, the rise is so steep you can't really see the sawtooth pattern from the falling temps and rising again, which happens twice, first from around 1850 to 1910, then from 1942 to 1975 (steep rise 1780 to 1850, 1910 to 1942, and 1975 to 2000).

    To be useful and scientificaly rigorous I should replot all the data, with all the GRIP data in a different color, with the y axis marked on the right side with actual GRIP temperatures, and GISP2 temperatures marked on the left side, lined up so the 1855 data in both sets was the same on the y axis. For this discussion, I thought what I did was good enough, because everything is so compressed.


    [JH] Duplicate post deleted. 

  35. steven foster @34, where you to take the effort with the actual data, a very useful graph would be the full holocene record from Alley et al overlaid with the 100 year average from Kobashi et al for the last 4000 years, and with an inset or second panel showing the last 2000 years from Kobashi et al overlaid with the 10 averages from Box (2009).  That would display the full range of data with a minimum of visual clutter and with appropriate resolutions for comparison.  The first graph could also have modern temperatures marked on the graph for comparison with the full holocene record.

  36. Thank you. I'm going to use the NOAA GISP2 data which has a resolution of 40 data points for every 1000 years for the whole Holocene. It looks like (so far) that the GRIP data has a lower resolution. I will look into the Kobashi suggestion. I did a quick low resolution thumbnail of what it will look like:

    I differ with you about Greenand ice core data being a local proxy only. The H and O isotopes in the water represent temperatures of the water from around the world that winds up in Greenland, and don't represent the temperatures on Greenland itself. Water in the oceans has H and O isotopes of a known concentration. But the boiling point of heavy water, as well as heavy oxygen water (H2O^18 rather than H2O^16) are higher than that of normal water, and are found in water vapor in the atmosphere at lower concentrations when the global temperature is low. When the global temperature is high, more heavy water evaporates around the world than when it's cold. But when the water vapor travels to Greenland, it is assumed that it all freezes, regardless of the temperature in Greenland. So the higher the temperatures around the world, the more O^18 and H^2 we find in the ice. So, it's a global proxy, more suggestive of northern hemisphere temperatures, of course, but water vapor doesn't always commit to one hemisphere. The other argument might be that the FREEZING point of heavy water is also higher than regular water, so more of it would snow out of the air before it gets to Greenland, so the temperature of Greenland matters. But snow crystals don't care what isotope of water....actually I need to think about this. Ok I thought about it. None of the water freezes in Greenland. It all freezes long before it gets there, probably as cirrus clouds in the stratosphere, which are already frozen long before. I don't know.

  37. steven foster @36, the low resolution thumbnail you linked to certainly is superior to the image from post 15.  I look forward to seeing the completed high resolution version.

    With regard to icecore temperature proxies, yes the temperature of the water from which the water in the snow originates does influence the isotope ratios, as also does the air temperature and distance travelled as the water vapour is transported to the site.  Lower air temperatures and longer travel distances resulting in a lower proportion of heavy isotopes.  As a result, ice cores are a regional (not hemispheric, and certainly not global) temperature proxy.

    The reason they are neither hemispheric nor global proxies relates to the pattern of wind circulation:

    As you can see, air near the equator is lifted high into the troposphere, effectively drying it by precipitating out nearly all of the water vapour.  The return to the surface of this very dry air is the cause of the bands of deserts at certain latitudes.  Air from the mid latitude cell, on the other hand procedes north at the surface (in the NH) before again rising to great altitude, drying it.  That air column allows mixture with air from the polar cell, but the water picked up will be precipitated out before the polar cell can cause it to circulate over greenland.  Consequently, the water vapour in air over Greenland predominantly is picked up from polar easterlies as they travel south.  This overall pattern means very little water vapour will reach Greenland from mid-latitude or tropical oceans.  At best Greenland ice cores would be an Arctic proxy.

    It is not even that, however, during the Holocene.  The local pattern of circulation means the watervapour precipitated out over Greenland comes almost entirely from the Atlantic north of Iceland and the Barents Sea:

    (Red arrow indicates wind pattern, red dots the location of Greenland ice cores)

    You will notice the large difference in circulation patterns between "Present day" and "Last Glacial Maximum".  That difference means that Greenland ice core data are not even proxies of the same region at different points of time.  It further means that they are far closer to Arctic proxies rather than Atlantic Arctic proxies durring glacials.  It should be noted that the AO and the AMO will also shift the winds over Greenland, slightly changing the region for which the ice core is a proxy over time.  That, however, does not make those ice cores proxies for a larger region.  Rather, it merely introduces noise, making the record more erratic and amplifying warming or cooling trends in the proxy record.  We have to accept that for Greenland, for we have no better proxies for Greenland temperatures.  But it is further reason to not treat the Greenland record as a global or hemispheric proxy.

  38. You are confused on several points. First, the surface wind patterns you describe are tropospheric, and have little to do with the transfer of mosture from equator to poles. I could cite a reference but that would just clutter this blog because I can simply use the illustrations you already provided. See in your post 37, the first illustration. Look at the Hadley cells in the tropical bands, blowing equatorial moisture into the temperate bands, and from there into the polar cells. Secondly, you state that equatorial moisture only reaches the troposhere and precipitates out in that band. Water vapor at the equator goes far higher into the stratosphere than at any other place on earth. Thirdly, you seem confused about the concept of dry air in desert bands. Air that may seem dry in a warm desert would be extremely moist at the poles. Warm desert air might have a dew point of +40F while the air in Greenland has a dew point of maybe -40 (C or F same).  The Rockey Mountains are covered in snow which came from desert air....just think of air in the Mohave Desert moving east and rising over Humphrey's Peak in Arizona. All the snow they get in Flagstaff comes from the desert air. Fourthly, you say snow in polar bands comes from moisture picked up by polar easterlies. This is impossible. In the antarctic polar band (60 and south) there is no ocean at all, all water vapor coming in from the stratospheric conveyor belt (Hadley cell to temperate cell to polar cell), and in the north, there is an icy ocean mostly covered with floes and fast ice. If you want to argue that an appreciable amount of water vapor comes from this icy northern ocean, you need to show that there's a difference in snow patterns between that pole and the other pole that doesn't have an ocean. There isn't. Fifthly, next post.

  39. Fifthly, because the Arctic Ocean is always frigid and full of ice, it is always around the same temperature....that of ice water (ice seawater)....during the entire Holocene. Because of this, arctic easterlies would always be the same temperature every year and if the snow came from arctic seawater it would not show any difference in isotopes in the bands. But in both arctic and antarctic ice cores, we see similar bands, because the snow in Greenland all comes from the stratosphere, not the Arctic Ocean. Sixthly, during the ice age, which amounts to 80 or 90% of the time, there is no Arctic Ocean at all, the entire north pole covered in ice several miles thick. How do you explain Greenland ice cores from 30,000 years ago when there was no liquid water north of Chicago? (tongue in cheek on Chicago). Seventhly, next post.

  40. Seventhly, the summit of Greenland is 10,000 feet in elevation, making it closer to the stratosphere that to sea level in terms of elevation, because the troposphere is much thinner at the poles than elsewhere, extending to maybe 15 or 20 thousand feet max and 30 or 40 k at the equator. And since snow falls downward, not upward, and the Arctic Ocean is at sea level, not much snow which precipitates from the Arctic Ocean would ever reach the summit. In terms of geographical distance, the summit is hundreds of miles from the Arctic Ocean and only 2 miles from the stratosphere. Therefore, in summary of all 7 of my points, most of the snow in Greenland comes from water that arose in the tropics, and therefore is a global or hemispheric proxy, not a local. This applies to MOST of the water vapor during the Holocene, and ALL of it during an ice age. Sorry, I don't mean to shout, but I've had a lot of coffee this morning.


    [JH] There's no need for you to apologize for your tone. It is quite reasonable.

    For future reference, it would be easier for people to follow your train of thought if you were to use (1), (2), (3), etc., rather than "firstly", "secondly", "thirdly", etc.

  41. Steven Foster, as I recall, the lowest known tropopause height is about 8 km, which is approximately 26000 ft. Your tropopause height numbers are way off, where did you get them? In tropical regions, it can be as high as 17 km. The top of Greenland is not closer to the Stratosphere than to sea level, that can be verified with 2 minutes on Google. 

    As for this statement "the snow in Greenland all comes from the stratosphere" what do you have for substantiation? It is such a strong statement, and so contrary to all the research on air dehydration at/actross the tropopause that you should present some very serious backup.


  42. Also, where is the reference for this?

    "All the snow they get in Flagstaff comes from the desert air"

    The college of agriculture and life sciences of U. of Arizona says that it comes from Pacific storms and their associated fronts. They also concur with Tom Curtis on the desert bands, as does essentially everything I've read about it.



    [RH] Shortened link that was breaking page format.

  43. Steven Foster @40...  Richard Alley, who is probably the leading expert on the Greenland ice core data, repeatedly states that the Greenland summit data is a regional, not hemispheric, record of temperature.

    DotEarth-Richard Alley


    [DB] Steven Foster was a sock-puppet, fake user ID deployed by the departed Morgan Wright.  Both have had user privileges revoked, as will any subsequent iterations.

  44. This has come up before but I cant find it. There is a paper demonstrating that it is regional. Fractionation during precipitation is reason. Anti-phased cooling periods between Antarctica and Greenland would also not be observed in ice core if the proxy was global (and would be challenging comparing GISP2 to tropical glacier icecore and insisting that they represented a single global proxy!).

  45. Found the paper I was looking for: "Validity of the temperature reconstruction from water isotopes in ice core", Jouzel et al 1997 (the et al including Johnson, Alley, Cuffey and Dansgaard).

  46. DB...  Why am I not surprised?

  47. re: steve foster @38-40, further research has shown that I have over interpreted the global circulation diagram and the data from one study shown above. It does not, however, support "Foster's" claims. I will discuss this in detail in a follow on post, but first go point by point through "Foster's"' claims:

    1) The troposphere includes the entire atmosphere from the base of the lowest blade of grass to (approximately) 150 hpa.  As such, it certainly includes winds over the ocean surface, and over Greenland, and hence it is tropospheric winds that account for the snowfall in Greenland.

    2)  Water vapour is carried far higher into the atmopshere in the tropics than anywhere else, but not in any significant amount into the stratosphere.  That is because the troposphere reaches to a far higher altitude in the tropics than elsewhere on the planet.  Because water vapour precipitates out with the cold, the stratosphere in fact has very little water vapour compared to the troposphere:


    3) You are missing the point.  The dry band at the northern edge of the NH Hadley zone results means nearly all water vapour from the tropics has been precipitated out.  Only a portion of it then travels further north, with most returnd to the tropics by the trade winds.  Of that which travels north, much of it will be precipitated out, with some of that being replaced by newly evaporated water vapour from mid-latitude oceans.  This is particularly the case as it travels along the surface.

    4)  Actually, Antarctica (168 mm year continental average) does recieve less precipitation than the Arctic (<500 mm per year average), with the highest most central regions of Antarctica receiving 50 mm per year compared to <100 to 200 mm per year for central Greenland (described as the driest part of the Arctic).

    5) The Arctic ocean varies quite significantly in winter temperatures, and away from the ice edge varies also in summer temperatures.

    6) Foster completely ignores the second figure in my preceding graph which shows the large difference in circulation between Holocene and glacial conditions, something I specifically commented on.

    7) As Phillipe Chantreau @40 points out, your maths on distances is just wrong.  More crucially, the stratosphere has very little moisture compared to the troposphere as can be seen by comparing the figure above to the zonally averaged specific humidity:


    Comparison shows a specific humidity of at least 1 g/Kg, ie, 1000 ppm by mass, through out the surface troposphere.  For comparison, stratospheric values are around 4 ppm by volume.  As H2O is a light molecule, that drops to closer to 2 ppm by mass.  Therefore the stratosphere is several orders of magnitude too dry to be the major source of precipitation at any location in the troposphere.  It is also far too dry to be a channel through which tropical (still less SH) moisture is tranferred to the poles.

  48. As noted above, I have overinterpreted the data available to me.  This is always a risk when interpreting scientific data as a non-expert.  In particular, not being expert I have not always read the range of scientific papers on a subject to put the particular papers on which I rely into context.  In this case I accepted too readilly the data from Langen and Vinthner, (2009), or more properly, their webpage on that paper.  Prompted by Foster/Wright above, I have since looked at a greater range of the literature and found a significantly more southerly source of Greenland precipitation in most papers (again without claiming an exhaustive search, so I may be missing something.  There are three key papers in this reassessment.

    1)  Johnsen et al (1998) find that most of Greenland precipitation, particularly on the west coast comes from "the subtropical part of the North Atlantic Ocean", with the best modelled results with an assumed source close to weather ship E (Echo), ie, just north of the Mid Latitude High:


    2)  Of particular interest in this discussion is Werner et al (2001), which examines source water for precipitation in both Greenland and Antarctica.  The find the only overlapping source for Summit at Greenland, and Vostok in Antartica is the tropical Indopacific, which provides just 10% of Greenland's precipitation, but 44% of Vostoks. Polar Seas provide 15% of Greenlands moiture, and the North Altantic 28%.  In all, 72% comes from the extratropical NH, and the remainder from the tropics, with none from the SH.


    3)  Also of interest is Sodemann et al (2008a), which discusses the sources of winter precipitation in Greenland.  In particular, the examine the effect of the North Atlantic Oscillation (NAO) on the sources:


    Again, the dominance of the North Atlantic as a source is evident; but so to is the large shift in source with a shift from postive to negative NAO.

    Overall, a clear picture that most precipitation from Greenland comes from the extratropical North Atlantic and Polar seas.  There is almost no moisture from the SH, and none from the SH extra-tropics.  There is an almost complete divergence between sources for Greenland and Antarctic ice, showing that neither can be considered a global proxy.  Further, Greenland ice shows a significant shift  in source based on dominant weather states (NAO) making it even less suitable as a global proxy.

    So, while my over interpretation has lead to a clear error in relationshipt to the dominant source of Greenland precipitation by placing it to far north, it does not effect the conclusion that Greenland ice cores are a regional, not a hemispheric or global proxy.  Curiously, the 44% contribution from tropical waters to the Vostok precipitation makes it, though clearly not a global proxy, a significantly better approximation than Greenland ice cores. 

  49. I normally just read this blog for information and am now posting for the first time. Morgan Wright is an active denier who trolls many climate blogs and sometimes pretends to be a climate scientist to state his denialist lies. He actually has his own denier blog at so be aware because he recently used a lot of information he got from this thread on his blog...he spun the bit about 1855 and GISP2/local proxy his way. Don't fed the troll.


    [JH] What exactly do you mean by your statement, "Don't feed the troll"? As best as I can determine, Morgan Wright has not posted a comment on this thread.

    Upon further review, I see that Morgan Freeman did in fact post on this comment thread under a different user name, Steven Foster. Morgan Freeman and his sock puppet have been banned from posting on SkS.  

  50. the last 10,000 years of GISP2 data

    If you have done any form of stock chart or commodity chart analysis you will se that the past 10,000 years shows no up or down trend in global temperatures.(or temperatures of the Greenland Ice Core)

    What it does show is Temperature Range of approx 5 degrees - or a sideways trend. That is the Long Term Trend.

    Its completely irrelvant as to whether or not Eastwards "present" is 1855.

    Focusing on short term trends are extemly misleading and in the case of stocks or commodities can be very expensive.


1  2  3  Next

Post a Comment

Political, off-topic or ad hominem comments will be deleted. Comments Policy...

You need to be logged in to post a comment. Login via the left margin or if you're new, register here.

Link to this page

The Consensus Project Website


(free to republish)

© Copyright 2022 John Cook
Home | Links | Translations | About Us | Privacy | Contact Us