Marcott et al claim of 'unprecedented' warming compared to GISP ice core data

The very first sentence of the Marcott et al (which is getting heavy press) abstract says:

Surface temperature reconstructions of the past 1500 years suggest that recent warming is unprecedented in that time.

Okay, let’s have a look at the claim. First this graph from the publication:

marcott-A-1000[1]

Seems reasonable when you look at that data, right? But let us examine a well known reconstruction from GISP2 ice core data in Greenland. Here’s a section from Dr. Richard Alley’s reconstruction:

Greenland ice core isotope past 4000 yrs

Now here is a simple scaling of the Marcott et al graph to get an approximate match for the temperature and time scales:

Marcott-Greenland_ice core_isotope

Note that this is just a simple visual comparison, with a rough match of the data for time and temperature scales – it isn’t intended to be anything else.

The full un-cropped Alley GISP2 plot can be seen below:

gisp2-ice-core-temperatures

In my overlay above, the Marcott et al graph full time scale on the x axis is 2000 years, and its temperature full scale on the y axis is two degrees C. The scaled overlay to the Alley GISP2 plot is a reasonably close match to the GISP2 plot scale units. The centerlines don’t match, but they can’t with this sort of comparison.

The idea here is simply to compare magnitudes of the data on the same time scale.

Clearly, the GISP2 data has greater magnitudes in the past 1500 years, and at longer time scales, the GISP2 temperature reconstruction dwarfs the magnitude of the Marcott et al temperature reconstruction. Dr. Don Easterbrook has a good synopsis of GISP 2 temperature reconstruction magnitude on WUWT here.

This simple visual comparison suggests that their “unprecedented” claim for the 1500 years BP is unlikely to hold up when examined against other reconstructions. As they say in the big leagues, more study is needed.

Marcott et al alludes to the warmer temperatures of the past in this paragraph:

Our results indicate that global mean temperature for the decade 2000–2009 (34) has not yet exceeded the warmest temperatures of the early Holocene (5000 to 10,000 yr B.P.). These temperatures are, however, warmer than 82% of the Holocene distribution as represented by the Standard5×5 stack, or 72% after making plausible corrections for inherent smoothing of the high frequencies in the stack (6) (Fig. 3). In contrast, the decadal mean global temperature of the early 20th century (1900–1909) was cooler than >95% of the Holocene distribution under both the Standard5×5 and high-frequency corrected scenarios.

Perhaps this weekend when I have more time, either I can do a proper plot of the data in a similar fashion to see how well they match when plotted side-by-side in the 1500 year time frame. Unfortunately I have other work to do today, so I can’t at the moment, and I’m traveling again tomorrow. Posting will be light.

Advertisements

  Subscribe  
newest oldest most voted
Notify of
thelastdemocrat

The trick is the “filter.”

Seth is propagating the story: http://www.pddnet.com/news/2013/03/recent-heat-spike-unlike-anything-11000-years
Every time I see one of these “studies” I think we have finally reached the pinnacle of stupidity, but then there’s another.

Modoki

[snip – if you have a valid criticism, post it, snark helps nobody -mod]

Richard

Anthony
I think it might be a good idea to look at your “years before present” they look rather out of alignment to me.
Richard
REPLY: That was done for a reason, so that the Y axis scales would both be visible. Note that I said “rough” alignment. – Anthony

Reblogged this on If You Voted For It — You Own It and commented:
Nevada County’s local left is promoting the Marcott etal Study, and this is reblogged is to provided for some perspective for the more rational thinkers.

Paul Sheraton

Now that the original hockey stick graph (and it author) cannot be ignored as anything other than a joke.
We now bring you… the new hockey stick, Hockey stick 2!

I’m assuming Marcott et al are still tacking on the thermometer record for 20th/21st century temperatures. If so, then there is still the apples/oranges comparison of thermometer and paleoclimatoligical records, where the statisical methods used to generate the paleo record wash out the short term variations that the thermometer records captures.
I also cannot see that the error bars 10K years ago will be the same as 500 years ago.

elftone

Ooh, look – they’ve got rid of the MWP. Again. Do they really think people will forget if they keep trying that trick?

Nullius in Verba

I posted this on the earlier Marcott thread, but it may be more appropriate here. Apologies.

I’ve had a quick look at the data, and it seems to be a mixture of different behaviours.
One set of five series shows a peak around 9k years ago, a dip from 7k-3k and a rise to the present day.
http://i49.tinypic.com/lbogh.jpg
Another set of around thirty five series shows pretty much of nothing. It looks like random noise.
http://i49.tinypic.com/t68r4x.png
And the third set of around twenty series shows the holocene optimum we know, with a peak around 2 C warmer than today.
http://i47.tinypic.com/2uylqh3.png
(There were a few others that didn’t cover a long enough interval to make a proper comparison, I think.)
The ‘Holocene Optimum’ series have been mixed in with a bunch of ‘random noise’ series and some ‘Holocene dip’ series which has the effect of bringing the Holocene peak down to below a degree. All they have to do then is paste the instrumental temperature series on the end. (Note, the end spike is not in any of the 73 series.)
Of course, that might be saying that the Holocene peak did not occur everywhere, and that in many places the temperature didn’t change, which is how they’re interpreting it. You would need to look at the individual data sources to tell.
Whether the Holocene Optimum was global or not, it’s definitiely very naughty to splice on the instrumental record and talk about anything being ‘unprecedented’. The proxy data smooths anything going on at less than 300-3000 years (figure S18 in the SI). Apples and oranges.

Nullius in Verba

I also noted from their map that most of their sites seemed to be on or near the coast. A maritime climate might not necessarily reflect changes in either ocean or continental climates.

SCheesman

This may also be an apples/oranges comparison. IF it is generally true that north-polar temperatures fluctuate more than global averages (as currently appears to be case), then an additional vertical scaling adjustment might need to be applied, even to make a qualitative match, and this would tend to reduce the high temperature fluctuations of the ice core data w.r.t. the global pollen reconstruction. Is there not a comparable global reconstruction over a longer time span that the comparison can be made with?

William McClenney

What might be truly unprecedented is if the Holocene DOES NOT have from 1 to 3 strong thermal excursions right at it’s typical 1/2 precession age end.
http://wattsupwiththat.com/2012/03/16/the-end-holocene-or-how-to-make-out-like-a-madoff-climate-change-insurer/

RockyRoad

The way temperatures have been “homogenized” in the past 100 years (do I hear a “model” echo?), they can’t honestly be compared to deterministic temperatures from ice cores. That’s like comparing apples and oranges.
So Marcott is right–his bunch of CAGWs only look at models and their models are designed to show that the last 100 years have the largest spike. I’m glad he took the time to tell us.
Marcott should have added that caveate to his statement; had he looked at the Ice Core data (which some alledge is smeared by diffusion so it’s undoubtedly less pronounced that reality), such a daring accuasation couldn’t have been made.
(How dare any of you look at anything but models…)

When is ‘present’? The acronym BP [before present] normally means before 1950 so that the time axis stays fixed.

Modoki

“if you have a valid criticism, post it, snark helps nobody -mod”
One study compares local temperature response at an acknowledged labile location (Greenland).
The other study compares global temperature response.
Apples and oranges.
Then again (to his credit) Mr. Watts stated:
” Note that this is just a simple visual comparison, with a rough match of the data for time and temperature scales – it isn’t intended to be anything else.”
It is a rough match of data measuring 2 different metrics which would be expected to be different.

Tom Jones

Seth’s article is everywhere, as opposed to the report, which is behind a paywall. So, this comment is informed by the former. They claim that the dramatic turnaround in temperature began about 100 years ago, which inspired two thoughts.
One is that temperature was amazingly sensitive to CO2 concentration, but became much less so as time went on. (Interesting physics) Two is that it’s a good thing the industrial revolution halted the increasingly rapid slide into the next ice age. Maybe Watt saved civilization

Phil.

The last datapoint of the Alley data is 1855, there’s been some warming since then.

Both could be right.
GIPS2 is a history from one location with a couple of proxies for temp.
Marcott uses one or more different proxies, from 73 ice and sediment core monitoring sites around the world [Exactly how much overlap in time between site samples? How noisy are the proxies?]
It might be a better representation than a Greenland ice core. In fact, it ought to be better. Whether it really is better is the open question. Mann sampled from more than one location, too, but the proxy is very noisy and the sample weighting preposterous.
Marcott may have done better, but getting a pay-walled paper in just under the wire for IPCC AR5 does not add to the paper’s credibility.

You do know that a number of Greenland ice cores are among the proxies used by Marcott et al, right? Its not that surprising that localized variability (especially in high-latitude areas) would be much higher than in global reconstructions.

utterly misses the point.
1. The Marcott analysis has a time resolution of around a century.
2. The marcott analysis is “global”
The ice core has much better resolution. That means it has a better chance of picking up high frequency variation. marcotts proxies are basically “smooths”
As an analogy if you took an average of say san francisco over a year suppose you had an answer of 55F on an annual basis. Then somebody came along and said “Hey on labor day it was 75!!!. well DUH!, one is a annual smooth and the other is a high frequency spike
For the spatial issue, imagine Anthony Posted CRN average for the US and said “the average was 76F” and imagine some warmist said ‘ hey, it was 95F in new mexico”
in both cases the person who cites higher frequency data to counter lower frequency claims is really OFF TOPIC. and the person points to high temps in new mexico to counter Anthony’s claims about a US average is also OFF topic.
In short. you cannot refute a spatial average ( Anthony says the US is 76F) by citing a point measurement ( LA was 92! anthony must be wrong ) and you cannot refute a low frequency measure ( the annual SF temp is 55F) by citing a high frequncy measure ( Sat, sept 5th at 1PM was 81F).
What you can do, is read Rhode closely. Then respond
Oh and here is another thing to ponder: Imagine the Holocene was warmer than they suggest.
That would imply a more sensitive climate since the change in forcing between Holocene and LIA isnt particularly big.

Claude Harvey

I see Michael Mann’s “endorsements” peppered all over some of the breathless news accounts of the study. That alone should raise suspicions.

DayHay

http://i.imgur.com/s19MOMd.jpg
Scale and time are so important. Perhaps this needs labeling “Your so called hockey stick is here”? What is actually stunning is only a 3 degree fluctuation over 10k years. So I would say that as soon as man arrived in numbers, everything got fixed, so everyone, pat yourselves on the back. If Man and CO2 were involved, they were the fix, not the cause. It is also very obvious that civilization on earth is much more at risk of freezing to death.

glen martin

link to abstract at top is busted

Matt in Houston

Perhaps Marcott et al confused the temperature data from their rectal thermometer and mixed the signals during an illness and drug induced deluded analysis? It seems likely that their intellectual grasp of reality is rather lacking and this would be a plausible mistake.
/sarc

Marcos José - Brazil

Using the full extent of the Marcott graph + GISP: http://img703.imageshack.us/img703/9075/marcottgisp.jpg

Matt Skaggs

Thanks for doing the heavy lifting NiV.

The modern anomaly is more likely due to the low resolution and dating uncertainty in the proxy data than unprecedented warming. There’s a reason why we don’t correlate well logs based on radiometric dating or even on paleontology. We correlate them based on log characteristics (kind of like wiggle matching).

Leif: “When is ‘present’? The acronym BP [before present] normally means before 1950 so that the time axis stays fixed.”
That’s a common criticism of Easterbrook graphics (which these appear to be; they’re not from Alley’s 2000 QSR article — although that’s somewhat implied). The first datapoint appears to be 95 BP, which is 1845 — “hiding the rise” during the past 150 years.
Is this correct? If so, there should a large note in the article text.
Also, it would be nice to have a citation (or link) to the source of these graphics.

Louis

Past warming won’t show up in this study if it wasn’t “sustained for at least 100 years” according to Marcott:
“The resolution of the reconstruction is averaged into 100-year segments, which means yearly or every-decade variability fails to show up in the new study. There could have been a period sometime in the past 11,000 years that was warmer than today, but if so it wasn’t sustained for at least 100 years.
‘That is a drawback,’ Marcott acknowledged, though he said the resolution is sufficient to show the rapid change between 1900 and 2000 and the change projected by climate models between 2000 and 2100.”
What do climate model projections have to do with it? Does their new and improved hockey stick include model projections?

Richard

“That was done for a reason, so that the Y axis scales would both be visible”
Ah yes Anthony, I see what you have done there.
Richard

FerdiEgb

Dear Anthony,
One need to be cautious by using the GISP2 data: that is one point on the map that reflects the temperature on the summit of the Greenland ice core, nothing else. It may reflect more or less the whole Greenland and beyond area, but in general temperature fluctuations of Greenland and NW Europe follow opposite swings, together with changes in the NAO.
More interesting would be the d18O/dD data from the same ice core, as that reflects the temperature of a large part of the North Atlantic Ocean, where most of the precipitation on the Greenland summit originates. That says more about the past temperatures over a larger area…

A “reconstruction” that misses the Minoan Warming Period 3x longer than the claimed resolution is not in need of fine tuning, it is in need of scraping.

Henry Clark

Part of how the Marcott graph fudges history is by flattening inconvenient variation in prior centuries to falsely suggest the difference between the peak of the MWP and bottom of the LIA was only around 0.4 degrees Celsius. In contrast, a more honest source, Loehle data at http://www.drroyspencer.com/global-warming-background-articles/2000-years-of-global-temperatures/ , shows such was around 1.2 degrees Celsius difference.
Every warm period in the Holocene (and before) has had more relative temperature change near the arctic than in the tropics (simply because the tropics never change much in temperature), as is so for the Modern Warm Period (“global warming”) as well as all of the others. The general picture of current temperatures not being much compared to past millennia is seen both in the GISP2 data as well as illustrated in as wide reconstructions as Loehle.
There are loads of other examples, including higher global sea levels than now during the Medieval Warm Period and still more so during the Holocene Climate Optimum. As another illustration, http://www.sciencedirect.com/science/article/pii/S0009254199000297 references how near-Arctic Eurasian vegetation was 20% more than now during the Holocene Climate Optimum and 55% more than now during the Eem interglacial optimum (due to those times being much warmer than now, with less frozen ground).

FerdiEgb,
Both hemispheres appear to correlate well.

TomRude

Mosher: “in both cases the person who cites higher frequency data to counter lower frequency claims is really OFF TOPIC”
How about the person who splice indirect proxies and the measured temperature record and claim unprecedence? How about the person who thinks a heterogeneous global temperature sampling on continents, weighted with statistical methods has a climatic relevance? Ignoring processes is what the AGW crowd loves the most. No need to understand meteorology and climatic evolution, just measure and compute and tada! It’s there.

The money for promoting man-made global warming hoax and global governance is going bye bye.
China Backing Away From Carbon Tax Start in 2013, Official Says
“China will wait until after this year to introduce a tax on carbon, deferring to concern that economic growth might suffer, a government researcher said. ”
“China first said it would implement an environmental tax in 2011, when the country released its 12th Five Year Plan, a list of policies to be carried out by 2015. ”
http://www.bloomberg.com/news/2013-03-06/china-backing-away-from-carbon-tax-start-in-2013-official-says.html

rogerknights

DayHay says:
March 8, 2013 at 11:47 am
What is actually stunning is only a 3 degree fluctuation over 10k years.

That should be an immediate red flag.

apachewhoknows

mod/mods
Thanks for not deleteing me per my hot headed request.
Cooling off time over.
Working on another blog just now and it jumped into my strange brain.
Why not attempt to bait and switch aka trick the Democrats in the U.S. Senate to put up some CO2/Climate Change/Global Waming , big spending bill as in a tax on milage on auto/truck use and or some such tax and spend boondogle.
Once it comes to the floor have Ted Cruz and or Rand Paul or some other U.S. Senator well schooled on the facts known.
Have that person up and ready for a 13 to 20 hour filabuster of facts.
Seems to some a nice cheap way to bypass the msm blockade.
apachewhoknows

Henry Clark

Looking closely at the plot posted by DayHay, as I like its helpful grid scaling:
http://i.imgur.com/s19MOMd.jpg
That graph is particularly illustrative of how transitions from the warm state of climate to the cold state have occurred with around 10 degrees temperature drop in GISP2 over as little as around 300 to 400 years sometimes (even if not counting at all the Younger Dryas event which included an even faster drop section, just in case it was different from the others by being influenced by a comet impact, though there seems no way that the bulk of the others could be comet impacts). That would be proportionally like every several decades or faster a magnitude of change equivalent to the difference between the Little Ice Age and now, again and again multiple times over, an extremely rapid change in climate terms, vastly beyond the comparatively tiny fraction-of-a-degree changes in modern history. That fits with what I have heard elsewhere from trustworthy sources, whereas in contrast CAGW propagandists saying any transition to an ice age could only be gradual over tens of thousands of years are just lying as usual.

FerdiEgb

D.B. Stealey says:
March 8, 2013 at 12:42 pm
Both hemispheres appear to correlate well.
Yes, but that are indeed the d18O and dD temperature proxies. Anthony used the GISP2 borehole temperature, which only reflects the temperature of the Greenland summit at the moment of deposit. The d18O and dD proxies reflect a much wider area of the oceans where the water vapour in the ice originates…

Ian W

Modoki says:
March 8, 2013 at 11:16 am
“if you have a valid criticism, post it, snark helps nobody -mod”
One study compares local temperature response at an acknowledged labile location (Greenland).
The other study compares global temperature response.
Apples and oranges.
Then again (to his credit) Mr. Watts stated:
” Note that this is just a simple visual comparison, with a rough match of the data for time and temperature scales – it isn’t intended to be anything else.”
It is a rough match of data measuring 2 different metrics which would be expected to be different.

But the paper does not say that recent warming in Greenland are unprecedented – it says:
Surface temperature reconstructions of the past 1500 years suggest that recent warming is unprecedented in that time.
So your criticism should be leveled at the paper not Anthony

FerdiEgb
David A. Evans

Steven Mosher says:
March 8, 2013 at 11:27 am

As an analogy if you took an average of say san francisco over a year suppose you had an answer of 55F on an annual basis. Then somebody came along and said “Hey on labor day it was 75!!!. well DUH!, one is a annual smooth and the other is a high frequency spike

So today is Labo(u)r day?
DaveE.

Alex

“Oh and here is another thing to ponder: Imagine the Holocene was warmer than they suggest.
That would imply a more sensitive climate since the change in forcing between Holocene and LIA isnt particularly big.” or you just think that c02 is important when it might not be. In fact loocking at the long perspective several million years temp and c02 curves are all over the place so it cant be that co2 is as important as you think.

AndyG55

And remember that MOST of the spike at the end was CREATED by man.. ie Hansen et al.

Sorry. As soon as you flatten out the MWP to next to nothing I smell a rat.
Rather than headlines there should be deep suspitions, at best, and scorn if you were not in the mood to be polite. It is as if you came out with a paper that said that George Washington didn’t exist

RockyRoad

Steven Mosher says:
March 8, 2013 at 11:27 am

utterly misses the point.
1. The Marcott analysis has a time resolution of around a century.
2. The marcott analysis is “global”
The ice core has much better resolution. That means it has a better chance of picking up high frequency variation. marcotts proxies are basically “smooths”

But diffusion of elements in the ice core is also postulated to “smooth” the temperature variations, so your supposition that “ice core has much better resolution” isn’t justified, Steven. The “high frequncy variation” you mention has likely been lost.
But back to the original posting, Marcott should show his temperature graph of the past 10,000 years so we can all see if it’s valid or not. To what does he compare the last 100 years (of which at least half isn’t due to much anthropogenic influence at all)? Inquiring minds want to know.

Jaime Jessop

What may be unprecedented, that which has certainly spiked markedly in the late 20th/early 21st Century, is culpable stupidity overlapping considerably with outright scientific deception. There is no way that this smoothed foraminifera derived proxy data of low resolution should be juxtaposed with very high resolution modern instrument data. The press are having a field day – EPIC global warming screams one particularly nauseatingly uninformed rag. Furthermore, the whole story is conflated with virtual reality IPCC computer predictions of extreme warming which, without any scientific justification, is being touted as a natural consequence of the trend shown by this ‘new’ data. And why did the study stop at 11,300 years. had they gone back further, they would have encountered the Younger Dryas which might have really screwed up their nice smooth pre-industrial revolution temp data, even at low resolution. Could have ended up with a double-ended hockey stick!

FerdiEgb

Steven Mosher says:
March 8, 2013 at 11:27 am
The ice core has much better resolution. That means it has a better chance of picking up high frequency variation. marcotts proxies are basically “smooths”
Agreed, but that is also the case for the last 150 years. The resolution for that part is yearly (even with a lot of caveats) in Marcott e.a. The current warming even is not sustained over 100 years. With the same resolution as the proxies, this still is one of the coldest periods. And similar warmings of the past may have been hidden in the lower resolution proxies…
Oh and here is another thing to ponder: Imagine the Holocene was warmer than they suggest.
That would imply a more sensitive climate since the change in forcing between Holocene and LIA isnt particularly big.

A common problem in the models: a huge climate sensitivity for natural forcings (solar, volcanoes) means a huge climate sensitivity for all types of forcings (including GHGs and anthro aerosols), according to the models. But that is very questionable. Do you think that 1 W/m2 more IR absorption by CO2 has the same effect as 1 W/m2 more insolation? I don’t. See e.g.:
http://climate.envsci.rutgers.edu/pdf/StottEtAl.pdf
Further, a lot of (mostly European) researchers disagree too:
http://www.sciencedirect.com/science/article/pii/S0277379105001964
(the link to the free copy doesn’t work anymore)
Here the citation:
So, what would it mean, if the reconstructions indicate a larger (Esper et al., 2002; Pollack and Smerdon, 2004; Moberg et al., 2005) or smaller (Jones et al., 1998; Mann et al., 1999) temperature amplitude?
We suggest that the former situation, i.e. enhanced variability during pre-industrial times, would result in a redistribution of weight towards the role of natural factors in forcing temperature changes, thereby relatively devaluing the impact of anthropogenic emissions and affecting future predicted scenarios.

Rob JM

How do you get error bars of of +/- 0.2 deg C from data with a spread of between 2 and 6 deg C?
How can you justify splicing the high resolution air temp data onto an ocean reconstruction?Surely you would need to use ocean temp.
Talk about pseudoscience!