Musings on the Vostok Ice Core Record

(edited 11 July 2009 for clarity in discussion of lead-lag relationship between Vostok temperature and CO2)

Since I get asked so often what I think of the Vostok ice core data that James Hansen uses as ‘proof’ that CO2 drives temperature, I decided to spend a few days analyzing that data, as well as the Milankovitch forcings calculated by Huybers and Denton (2008) which (allegedly) helps drive the global temperature variations seen in the Vostok Data.

The following graph shows most of the 400,000 year Vostok record of retrieved temperature and CO2 content. Assuming that these estimates really are what they are claimed to be, what can they tell us — if anything — about the role of carbon dioxide in climate change?

First off you need to realize there has been a long history of debate in the climate community about the 3 Milankovitch cycles in solar forcing being the driving force for the Ice Ages. As far as I can discern, there have been at least three main problems with the Milankovitch theory.

First, the huge Vostok cycle at about 100,000 years remain unexplained because the Milankovitch cycles show no amplification for those events, just a regular series of small forcings which tend to be correlated with the smaller bumps on the curves of Vostok temperature and CO2 in the above graph.

A second problem has been that the positive correlation with Vostok (at the South Pole) came from NORTHERN Hemispheric forcing, not from the Southern Hemisphere. In the south, the Milankovitch forcings were out of phase with the temperature response in the Vostok ice core record. This has presented the problem of how Northern Hemispheric changes in solar forcing could cause such huge changes in Antarctic climate.

The third problem is that the Milankovitch forcings are so small it was difficult to see how they could cause even the smaller temperature changes in the ice core record – UNLESS climate sensitivity is very high (that is, feedbacks are strongly positive). This appears to be James Hansen’s view: the small Milankovitch forcings caused small increases in temperature, which in turn caused increases in carbon dioxide, which then took over as the forcing mechanism.

The recent paper by Huybers and Denton claims to have alleviated the 2nd and 3rd problems. If one assumes it is the length of summer in the Southern Hemisphere (rather than average yearly solar intensity) which is the main driving force for changing the Antarctic ice sheet, and also accounts for the fact that it takes many thousands of years for the ice sheet (and thus the temperature) to change in response to that summertime forcing, then the Southern Hemisphere Milankovitch cycles really do line up pretty well in time with the smaller Vostok events (but still not the huge 100,000 year events), and the “forcing” involved also becomes larger.

The Role of CO2 in the Vostok Record

So, where does CO2 fit into all of this? Well, at face value Hansen’s theory does require that temperature drives CO2…at least to begin with. Hansen claims the Milankovitch cycle in Earth-sun geometry caused a small temperature change, which then led to a CO2 change, and then the CO2 took over as the forcing. Or, maybe you could say the CO2 provided a strong positive feedback mechanism on temperature.

But it has often been noted that the Vostok CO2 record lags the temperature record by an average of 800 years, which is somewhat of a problem for Hansen’s theory. After all, if CO2 “took over” as the main driver of temperature, why do the CO2 changes tend to come after the temperature changes? But then there have also been uncertainties in dating the CO2 record due to assumptions that have to be made about how far and how fast the CO2 migrates through the ice core, giving the appearance of a different age for the CO2. So the lead-lag relationship still has some uncertainty associated with it.

But even if the CO2 and the temperature record lined up perfectly, would that mean Hansen is correct? No. It all hinges on the assumption that there was no forcing other than CO2 that caused the temperature changes. Hansen’s theory still requires that the temperature variations cause the CO2 changes to begin with, which begs the question: What started the temperature change? And what if the mechanism that started the temperature change is actually responsible for most of the temperature change? In that case, the climate sensitivity inferred from the co-variations between temperature and CO2 becomes less. The fact that even the recent work of Huybers and Denton does not address the major question of what caused the huge 100,000 year cycle in the Vostok data suggests there might be a forcing mechanism that we still don’t know about.

If Hansen is correct, and CO2 is the main forcing in the Vostok record, then it takes only about 10 ppm increase in CO2 to cause 1 degree C temperature change. The full range of CO2 forcing in the Vostok record amounts to 1.6 to 2 Watts per sq. meter, and if that caused the full range of temperature variations, then today we still have as much as 10 deg. C of warming “in the pipeline” from the CO2 we have put in the atmosphere from fossil fuel burning. That’s because 1.6 Watts per sq. meter is about the same amount of manmade forcing that supposedly exists in today’s atmosphere.

But if some other forcing is responsible for the temperature change, that necessarily implies lower climate sensitivity. And the greater the unknown forcing involved, the smaller the climate sensitivity you will calculate.

The Central Question of Causation

I believe that the interpretation of the Vostok ice core record of temperature and CO2 variations has the same problem that the interpretation of warming and CO2 increase in the last century has: CAUSATION. In both cases, Hansen’s (and others’) inference of high climate sensitivity (which would translate into lots of future manmade warming) depends critically on there not being another mechanism causing most of the temperature variations. If most of the warming in the last 100 years was due to CO2, then that (arguably) implies a moderately sensitive climate. If it caused the temperature variations in the ice core record, it implies a catastrophically sensitive climate.

But the implicit assumption that science knows what the forcings were of past climate change even 50 years ago, let alone 100,000 years ago, strikes me as hubris. In contrast to the “consensus view” of the IPCC that only “external” forcing events like volcanoes, changes in solar output, and human pollution can cause climate change, forcing of temperature change can also be generated internally. I believe this largely explains what we have seen for climate variability on all time scales. A change in atmospheric and oceanic circulation patterns could easily accomplish this with a small change in low cloud cover over the ocean. In simple terms, global warming might well be mostly the result of a natural cycle.

The IPCC simply assumes this internally-generated forcing of climate never occurs. If they did, they would have to admit they have no clue how much warming in the last 50 years is natural versus anthropogenic.

Fears of substantial manmade warming are critically dependent upon the assumption that the climate system never changes all by itself, naturally, or that there were not other external forcing mechanisms at work we are not aware of. Given the complex nonlinear behavior of the climate system, it seems to me that the assumption that things like global average low cloud cover always stays the same is unwarranted. In the end, in scientific research it’s the scientific assumptions that come back to bite you.

Where Does this Leave Carbon Dioxide?

About the only thing that seems like a safe inference from the Vostok record is that temperature drives CO2 variations – even Hansen’s theory requires that much — but the view that CO2 then caused most of the temperature variations seems exceedingly speculative.

And, if you are thinking of using the Vostok record to support warming driving today’s CO2 increase, you can forget it. In the Vostok record it amounts to only about 8 to 10 ppm per degree C, whereas our ~1 deg. C warming in the last 100 years has been accompanied by about 10 times that much CO2 increase.

The bottom line is that fears of substantial manmade climate change are ultimately based upon the assumption that we know what caused past climate change. For if past climate changes were caused by tiny forcings (too tiny for us to know about), then the climate system is very sensitive. Of course, those forcings might have been quite large – even self-imposed by the climate system itself — and we still might not know about them.

Finally, as a side note, it is interesting that the modern belief that our carbon emissions have caused the climate system to rebel are not that different from ancient civilizations that made sacrifices to the gods of nature in their attempts to get nature to cooperate. Technology might have changed over time, but it seems human nature has remained the same.