Quelccaya Ice Cap, first look

Posted: April 8, 2013 by tchannon in Analysis, climate, data, Dataset, Natural Variation


This is the relatively reliably dated section of the now archived Quelccaya ice core.

# Quelccaya Ice Cap, Peru 1800 Year Oxygen Isotope, Dust, and Major Ions Data
# Original_Source_URL: ftp://ftp.ncdc.noaa.gov/pub/data/paleo/icecore/trop/quelccaya/quelccaya2013.txt
# Contribution_Date
# Date: 2013-04-04
# Study_Name: Quelccaya Ice Cap, Peru 1800 Year Oxygen Isotope, Dust, and Major Ions Data
# Investigators: Thompson, L.G.; Mosley-Thompson, E.; Davis, M.E.; Zagorodnov, V.S.; Howat, I.M.; Mikhalenko, V.N.; Lin, P.-N.

The data shows a great deal of detail which mght be of interest to various Talkshop contributors, solar and or ocean index related.

Steve Mcintyre in his dry style noted the claim by some of d18O being a temperature proxy is somewhat dissed by 1998 showing in the wrong direction, the ups and downs of the proxy business.

d18O is not actually a temperature proxy, maybe more accurately described as an evaporation proxy, hence has far more to do with wet/dry. (same will apply to all ice cores, no news in this)


Given the very strange location of the icefield near desert, pampas and Amazon basin, plus near one of the critical coastal points to do with El Nino, this is an important data

Image right courtesy Google.


Image courtesy of University of Illinois. here


The as archived data is somewhat uneven showing the unreliable nature of ice cores unless that is you think this is really how things have changed.

On reading the notes it turns out the annual layering only applies to a subset so I cut off the early part where the published used spline adjustment. We don’t seem to get shown the actual problem.


A provisional look seems to hint at various things but I have no conclusion. This post is more about altering others to look.

[update] A relevant post by Steve Mcintyre is here.[/update]

Unofficial variance adjusted data for head plot and spectra, here.

Post by Tim Channon, co-moderator

  1. The Quelccaya has recently been throwing up plant remains dating back to 3000BC.

    And there’s more! Lonnie Thompson came up with a graph in 2003 showing warmer temperatures there back in the MWP, and commented

    For the Quelccaya Ice Cap (13.93°S, 70.83°W), this work revealed that peak temperatures of the MWP were warmer than those of the last few decades of the 20th century.

    And yet more!

    The fact that the Little Ice Age (about A.D. 1500 to 1900) stands out as a significant climatic event in the oxygen isotope and electrical conductivity records confirms the worldwide character of this event.


  2. Roger Andrews says:

    Here’s a plot of raw Quelccaya d18O against UAH TLT in the ten degree lat-long square around Quelccaya (9-19S, 66-76W) between 1979 and 2009. TLT shouldn’t be that far off air temperatures on the icecap, but no guarantees given:

    There’s zero correlation between TLT and d18O between 1979 and 2009 (R^2 = .002).

  3. tchannon says:

    Seems to cross off one link.

  4. vukcevic says:

    I did quick comparison of mid range frequencies from the CET and Peru ice data.
    Down on the favoured solar periods 60 (by Scafetta), 80 (by Gleisberg) and 105 (by Svalgaard) years.
    Not really: North and the South hemisphere marching in step, warmth in the north with ice in the south.

  5. Roger Andrews says:

    Cross off another one. The ice ain’t melting. Like Kilimanjaro, temps on top of Quelccaya never go above freezing.

    They have a pretty sophisticated weather station there too. Certainly free of urban warming impacts.

  6. Roger Andrews says:

    The more I look at data like these the more I become convinced that the Little Ice Age wasn’t so much a period of anomalously low temperatures as a period of anomalously high precipitation.

  7. Doug Proctor says:

    How do we go from isotope variations as noted to temperature variations in the world by such a proxy to a 0.1C or less uncertainty? And are the isotope variations measurement problems, preservation problems or source problems or causitive problems?

    And what really is the time resolution for what we measure, not what is deposited?

  8. Doug Proctor says:

    Actually, I thought that Antarctic and Greenland studies said that the top of the ice “breathes” for about 70 years (due to barometric changes) before becoming isolated from the atmosphere. That would depend in some part on precipitation rate, compaction and frequency of large, resetting events I suppose.

  9. tchannon says:

    The Hurst exponent is 0.55 – 0.65 which suggests not far off gauss random, however I have a twist to this, the data is too short. This came to light in a CSIRO tome where the usage of annual data killed the effect, which is looking for long term regime change, so whilst this data seems long, it isn’t really.Nevertheless the effect is merely maybe either way.

  10. tchannon says:

    Quite so Doug, lot of stuff which has never been properly done in controlled lab experiments.

    That said, this is about the water itself, nothing to do with trapped gas.

    How the ice moves is pertinent and so is any melt.

    Just seeing the diminution in amplitude with depth is a worry if this is merely about frozen water. Perhaps someone knowledgeable can explain.

  11. Joe Lalonde says:

    Roger Andrews says:
    April 8, 2013 at 10:57 pm
    The more I look at data like these the more I become convinced that the Little Ice Age wasn’t so much a period of anomalously low temperatures as a period of anomalously high precipitation.

    This is what I have been trying to show among the climate scientists ignoring precipitation completely for the strict use of temperature data. After all, an Ice Age IS the build-up of precipitation.
    Funny how low level cloud-cover NEVER crosses the equator and water on our planet surface changes directions at the 48 degree latitude.
    Even tree rings need water to grow and would distort temperature data if too much or too little.

  12. Roy Martin says:

    Very interesting data. Has been filed for future use. Thanks.

    Just eyeballing the adjusted plot, there appear to be some close correlations between the d18O record and climatic variations in southern Australia over the 20th. century.

    “Just seeing the diminution in amplitude with depth is a worry if this is merely about frozen water.”

    Same effect is quite pronounced in the Vostok records, compounded by a step transition during the inter-glacial periods. No real theory about why, just assumed that it was due to compaction under pressure – earlier layers are progressively thinner per time period – and possible bulk rise in temperature during each of the repeated inter-glacial periods. Also gives rise to questions about diffusion effects over millennium time scales, and the early firn type snow compaction phase. Expect there a lots of studies about both.

  13. Roger Andrews says:

    Joe Lalonde:

    Here’s a graphic illustration of the relationship between precipitation and ice formation that I posted on an earlier thread. It seems to fit here too:

  14. crosspatch says:

    “The more I look at data like these the more I become convinced that the Little Ice Age wasn’t so much a period of anomalously low temperatures as a period of anomalously high precipitation.”

    Or, put another way, and increase in clouds in the temperate regions. Some areas such as the Caribbean seemed to have had warmer than normal temperatures during the LIA.

  15. tchannon says:

    Updated post with a link to a relevant new post by Steve Mcintyre at Climateaudit.

  16. Joe Lalonde says:

    Roger Andrews,

    Oceans are a wide area where the oceans salinity does effect the rate of evaporation.
    Changes to these can have a very profound effect on the areas of concentrations.

  17. Roger Andrews says:

    A few parting comments on Quelccaya, summarized in the three graphs below:

    First Graph: There’s no correlation between Quelccaya d18O and observed TLT or SST over the period of instrumental record, but Thompson claims d18O correlates with ENSO events. According to the graph there is a positive correlation with Niño 3.4 since 1860 but it’s very weak (R^2=0.12). There’s no correlation between Niño 3.4 and ice accumulation (R^2= 0.02).

    Second Graph: When did the LIA occur at Quelccaya? According to ice accumulation data it peaked in the early 1600s. According to d18O it peaked over a hundred years later. (Note that d18O is 10-year smoothed.)

    Third Graph: Tim Channon noted that the variance of the annual d18O readings decreases as we go back in time. According to the graph the variance in fact peaks around 1720 and decreases before and after that. Why? One interpretation is more extreme weather events during the LIA.

    Quelccaya d18O doesn’t record any cooling effects from major volcanic eruptions – not Pinatubo, nor Krakatoa, nor Tambora nor the supposed 535AD mega-eruption we were discussing the other day. Other variables (dust, SO4 etc.) don’t respond either.

    The question seems to be; exactly what is d18O a proxy for?

  18. tchannon says:

    Fascinating. I completely agree, what does d18O do but this has been my mostly unspoken view for quite some time.

    Proxy need a great deal of caution, what exactly, what omitted variable(s).

    A nice one where I can’t say much involves a draft paper. This makes mincemeat of what was supposed to be so during the last ice age where perhaps the most telling evidence is dead bodies where there was supposed to be ice. Can’t have both.

    Your NINO plot does eyeball as similar.

    Bottom plot, another problem where answers are needed.

    Be a lot of eyes on the data and no doubt papers.