The ice cores are often lauded as evidence of the effects of carbon dioxide. Frank Lansner asks a pointed question and goes hunting to find any effects that can be attributed to carbon.
Where is the data that actually shows a strong and important warming effect of CO2? If CO2 has this strong warming effect, would not nature reflect this in data?
He has collected together the data from the last four warm spells (the nice interglacials between all the long ice ages) into one average “peak”. The common pattern of the rise and fall has already been recorded in many scientific papers. Orbital changes trigger the temperatures to rise first and about 800 years later (thanks to the oceans releasing CO2), carbon dioxide levels begin to climb. At the end of a patch of several thousand warm years, temperatures begin to fall, and thousands of years later the carbon dioxide levels slowly decline. No one is really contesting this order of things any more. What is contested is that those who feel carbon is a major driver estimate that the carbon dioxide unleashed by the warming then causes major amplification or “feedback”, making things lots warmer than they would have been if there was no change in carbon. Since most skeptics (but not all) agree that there is probably some warming due to extra CO2, the real question is “how much”.
Lansner points out that counter to the amplification theory, temperatures return most of the way back to their starting level (ice age temperatures) even while CO2 levels are elevated. If the CO2 can’t prevent the temperatures falling, it’s effect is anything but major.
Estimates of climate sensitivity and support for the “feedbacks” comes from models which depend on water vapor increasing high over the tropics. The radiosondes show that the models are wrong.
Frank graphs the change in temperatures and CO2, and finds a slight positive trend which is predictable (we know oceans release CO2 as they warm, so there would be a correlation). But then he plots the changes in CO2 against changes in the rate of temperature change, and finds no correlation at all (if CO2 was a major forcing, it would force or accelerate temperature change, which would show as the rate of temperature change). The data is limited to 1500 year blocks, so the time-frame is less than ideal, but the best available in the Petit data.
Thanks to Frank for his work
Searching for carbon’s effect in the ice cores…
An edited version of Frank Lansner’s post on Hidethedecline
In January 2009 Frank discussed the apparently missing CO2 warming signal in data at Watts Up With That in: “CO2, Temperature and Ice Ages”
I added most of the four large interglacial temperature peaks into one peak for a closer look.
The pro-CO2 argument goes, that 1) CO2-levels at only 210-240 ppm “must” be the reason that temperature boosts and thus could change 6-7-8 Kelvin or more on Earth.
However, in ice core graphs it’s clear that 2) CO2 concentrations far higher, 250-280 ppm, occurs while the temperature declines during 15-20 thousand years after the interglacial temperature peaks and thus temperatures returns almost to start level.
This doesn’t prove that CO2 has no effect, but nor does it suggest it has a big effect.
Furthermore, if some scientists thinks that CO2 just need to have a minor effect, it is as though they forget the basis of the CO2 theory that actually demands the CO2 effect to be dominating and strong:
The CO2 increase is supposed to increase Earth temperatures to a far higher level than seen on Earth in a million years or more in just few centuries. So obviously we are entitled to see actual data showing a significant CO2 effect dominating the natural mechanisms of temperature regulation on Earth.
I have downloaded the available CO2/temperature data from NASA (petit et al 1997-99):
I then to plotted data from sample to sample to seek for an actual CO2 warming signal.
(The Petit data has far less CO2 measurement data points than temperature points, so I have matched each CO2 measurement with the temperature data taken nearest possible to the CO2 sample. This gave, in average periods of approximately 1500 years length, the mismatch in sample data is approximately 52 years, but the difference has random direction and thus the CO2 data points are just 2,4 years later than the temperature points in average.)
Here’s how a scatter plot of dT/ dCO2 appears:
dT and dCO2 are calculated simply as:
The outcome showed a visible trend: A larger dCO2 is accompanied by a larger dT.
We know temperatures raise CO2 levels. So we’d expect to see a correlation.
It doesn’t help us with the question: Can CO2 itself increase temperatures? Or are higher CO2 concentrations mostly a result of temperatures?
Does a change in CO2 change the trend of temperatures?
Frank plots the derivative (which means he is looking not at the absolute highs and lows but at the rate that the slope of the graph changes.)
I made a plot ddT vs dCO2, the change of temperature trend pr 1000 years in kelvin as a function of change in CO2 concentration:
And again I made a scatter plot, and this time when highlighting the effect of CO2 change on change in temperature trend per millennium I found…
Nothing. The ability of CO2 to change the temperature trend from one period to the next is extremely hard to pinpoint. I just see white noise.
Is it wrong to expect a visible CO2 warming signal from these data? If CO2 was really capable of making global temperature much warmer than seen in a million years in nature — should there not be a visible connection between CO2 change and change in temperature trend?
I have thus failed to show a CO2 effect for these intervals in average 1500 years long. So perhaps the CO2 effect is supposed to have an effect only using some other length of periods? or? No doubt, data appears very noisy as CO2 trends and temperature trends are rather irregular appearing over the years — but still, should a strong CO2 effect not be visible from this noise? I think so.
Where is the data that actually shows a strong and important warming effect of CO2?
See Franks full post on his blog.
UPDATE: The slope of temperatures does not appear related to CO2 concentration in Vostok data
Frank compared the trend of temperatures as a function of the CO2 concentration:
dT = TB – TA
as a function of the average CO2 for the period:
CO2 = (CO2B+ CO2A)/2
It appears that there is no visible connection between CO2 concentration and the slope of temperatures in these Vostok data. In fact, if I insert a trend line it appears slightly negative.
Again, these are the data that should show the warming effect of CO2. Im not sure how much I can conclude on this, but i believe the CO2 – warming message would stand stronger if one could see bigger warming trends for bigger CO2 concentrations.