[amplexity]

Measurements from ice cores indicate that atmospheric CO2 levels have widely fluctuated in the past (going back 600,000 years, see http://climate.nasa.gov/evidence/).  What has caused these changes - especially the increases in CO2?  (keep in mind that I'm not talking about the current increase in CO2).

[CliffordK]

A large part of the increases is caused by a shift of CO₂ out of the oceans.

Think of partial pressures of two liquids.  The warmer the liquids, the higher the partial pressures, and thus the less CO₂ that is dissolved in the oceans, and the more in the atmosphere.

Likewise, the cooler the oceans, and the more CO₂ that is dissolved in the oceans.

The oceans, of course, have different layers.  The wave action, and the "surf" is responsible for much of the gas being dissolved in the water, and thus gases are distributed early near the surface.  The surface layers are also most affected by temperature changes.  The deeper ocean layers may be far more stable in temperature, and slower to be affected by gas concentration changes.

However, there may be shallow arctic regions that behave a bit as both the cool deep, as well as the fluctuating shallows (especially important with methyl hydrates). 

Also, keep in mind that the primary temperature proxies are the ¹⁸O and ²H concentrations being evaporated, and falling as rain which can be affected by some of the same pressures as the CO₂ concentrations.

[amplexity]

But that seems to imply that temperature changes are driving changes in CO2.  Climate scientists are arguing the opposite, that changes in CO2 are driving change in temperature.  So CO2 levels are first to change, not temperature, right?

[CliffordK]

That is precisely the problem that is often found between correlation and causation.

In fact, the temperature proxies seem to indicate frequently the temperatures beginning to rise slightly before the CO₂ levels begin to rise.  And, in many cases there is a significant delay between the fall of the temperatures, and the later fall of the CO₂ levels.

The theory is that there is an external event causing a slight perturbation in temperatures.  This in turn causes a shift in CO₂ levels, which then amplifies the temperature shifts.

One of the leading theories is that slight perturbations in Earth's orbit caused by Milankovitch cycles causes the temperature shifts which are then amplified by the atmospheric CO₂ in a feedback loop.  Again, one has problems with correlation vs causation.  The average amount of sunlight striking Earth is relatively unaffected by the Milankovitch cycles.  What they may change is the distribution of the sunlight to the northern hemisphere or southern hemisphere.  The north has more land and the south has more water so it is possible that the slight perturbations by the Milankovitch cycles would in fact cause global changes in heat distribution.

One of the hotly debated subjects is the variable output of our sun.  It is known that the sun's output varies on about a decade cycle.  It is also believed that there are longer term fluctuations in the solar output.  In particular, about 400 years ago the sun went through a low activity period called the Maunder Minimum which also correlated with a cool period often termed the Little Ice Age.  Many climate theorists believe that the fluctuations in solar output alone were not sufficient to drive the temperature shifts of the past glacial cycles.  And, again, other factors such as CO₂ are believed to have contributed, or amplified the changes.

Certainly, what few people debate is that the recent atmospheric CO₂ increase is caused by humans burning fossil fuels.  And, indications seem to indicate a slight increase in global temperatures over the last few decades. 

Anyway, much of this is still theory in which we're conducting an experiment with the world as our petri-dish.

[amplexity]

thanks for your replies

[damocles]

But that seems to imply that temperature changes are driving changes in CO2.  Climate scientists are arguing the opposite, that changes in CO2 are driving change in temperature.  So CO2 levels are first to change, not temperature, right?

The historical temperature changes are driving the changes in CO2 levels.
The present changes in CO2 levels are driving temperature changes, according to some very solid science and good evidence.

Amplexity it is very important to note that the present change in CO2 is quite different in character to anything else that has happened in the last half million years.

Certainly in the ice core record there have been about four major fluctuations in the last half million years, with temperature leading carbon dioxide by about 50 to 200 years. The major effects have involved long cold ice ages and shorter warmer interglacial periods, which have started and ended fairly abruptly.
But in a scale of half a million years, 'abruptly' might mean anything up to 5000 years.

The more important thing is that a pattern has been established over the last half million years with carbon dioxide levels ranging between about 200 parts per million in the ice ages and 280 parts per million in the interglacial periods. In a mere 160 years since 1850, the carbon dioxide has risen from what was already a peak interglacial value around 290 parts per million to 395 parts per million at present. This has not followed any temperature rise of significance -- certainly nothing like the 5-10°C that might typically lead the 200-280 ppm increase in carbon dioxide in the natural cycle of the last half million years. We are just starting to see a global temperature rise over the last two decades that is probably, but not necessarily, associated with the unprecedented increase in atmospheric carbon dioxide to 395 ppm, and that is certainly following the carbon dioxide buildup. There are good reasons in theory and in atmospheric modelling to believe that a temperature increase as a result of the carbon dioxide increase, and following it, will be increasingly apparent from now on. The evidence is strong, but perhaps not absolutely compelling. It is quite a different effect than what has prevailed previously -- we are breaking totally unexplored ground.

[Mazurka]

But that seems to imply that temperature changes are driving changes in CO2.  Climate scientists are arguing the opposite, that changes in CO2 are driving change in temperature.  So CO2 levels are first to change, not temperature, right?

Exactly, the paleoclimatic record generally suggests that temperature drives C02 concentrations.  This has been picked up on a number of occasions by anthopogenic climate change sceptics.

However, there is one exception to this called the PETM event - the "Paleocene - Eocene Thermal Maximum". 

The evidence in relation to this event suggests greenhouse gas lead warming.  Whilst our understanding of the PETM is incomplete, one hypothesis put forward relates to the destabilisation and consequent release of methane clathrates, possibly linked to volcanic activity north and west of Shetland (where there is a region of hydrothermal vents and intrusive igenous rocks appear of about the right age appear on seismic surveys).  The nice thing about the hypothesis is that it does not contradict well established theories relating to Milankovitch cycles.