Which is correct. Glacial-interglacial transitions are correlated with Milankovitch cycle (solar insolation) triggering/forcing*. I.e., when solar insolation is at a minima or maxima due to superposition of the Milankovitch cycle maximas (or minimas), this marks the transition point. Following maximas, there will then be a rapid decrease in solar insolation, triggering entry into a glacial period. Following minimas, there will then be a rapid increase in solar insolation, triggering entry into an interglacial period. Once either entry is underway, there are two main feedbacks: CO2 increase or decrease due to warming or cooling oceans, respectively, and retreat or advance of continental ice sheets. Both are positive feedbacks augmenting the transition direction. For a warming phase, increasing atmospheric CO2 creates more warming, releasing more CO2 from warming oceans, etc.
There is insufficient energy available for the solar insolation variability to induce the full range of temperature change observed. Atmospheric CO2 is the only known, quantified factor with sufficient energy to cause the full range of temperatures. There have been some theories about other solar cycle causes, but they will require much more evidence to effectively supplant the current understanding of glacial/interglacial variability.
* Not all the details of the glacial/interglacial record correlate with Milankovitch cycle solar insolation variability. Most of the variability does correlate well with Milankovitch.
cogitator says, “There is insufficient energy available for the solar insolation variability to induce the full range of temperature change observed. Atmospheric CO2 is the only known, quantified factor with sufficient energy to cause the full range of temperatures.”
First, this confuses the meaning of energy. Every change requires some energy exchange, but energy is not the primary driver in models for solar insolation. Secondly, CO2 does not inherently possess any available energy.
Solar insolation is the suns energy at the surface, and it is the result of the so-called solar constant, the earth's orbital position, the sun angle, the earth’s albedo above the surface, and the absorption of the atmosphere in the short wavelengths. The albedo is almost an unknown, and is usually modeled to one significant figure as 0.3. It varies with cloud cover, and all the things that affect cloud cover, including especially ocean temperature.
coagitator says, “Not all the details of the glacial/interglacial record correlate with Milankovitch cycle solar insolation variability. Most of the variability does correlate well with Milankovitch.”
There's more. Consider the following by Robert Ehrlich, Solar Resonant Diffusion Waves as a Driver of Terrestrial Climate Change, 1/4/06. http://arxiv.org/abs/astro-ph/0701117
>>VI. PROBLEMS WITH MILANKOVITCH THEORY AND CONCLUSION
>>In Milankovitch theory past glaciations are assumed to arise from small quasiperiodic changes in the Earth's orbital parameters that give rise to corresponding changes in solar insolation, particularly in the polar regions. A brief discussion of five problems with this theory are listed below, and a more detailed description of some of them can be found elsewhere. (Karner, 2000)
>>(a) Weak forcing problem: The basic problem with the theory is that observed climate variations are much more intense than the insolation changes can explain without postulating some very strong positive feedback mechanism.
>>(b) 100 ky problem: The preceding basic problem can be illustrated for the case of one particular parameter - the orbital eccentricity. The dominant climate cycle observed during the last million years has a roughly 100 ky period, which in Milankovitch theory is linked to a 100 ky cycle in the eccentricity. However, the effect of this eccentricity variation should be the weakest of all the climate-altering changes, in view of the small change in solar insolation it would cause. For example, consider the Earth's orbital eccentricity, e, which has been shown to have several periods including one of 100 ky during which e varies in the approximate range: e = 0.03 ± 0.02. (Quinn, 1991) The resultant solar irradiance variation found by integrating over one orbit for each of the two extreme e-values is about ±0.055%, or ±0.17 W/m^2 difference at the top of the Earth's atmosphere. Given that climate models show that a one percent change in solar irradance would lead to a 1.80C average global temperature change, then the change resulting from a ±0.055% irradiance change would be a miniscule 0.1 DC hardly enough to induce a major climate event - even with significant positive feedback.
>>(c) 400 ky problem: The variations in the Earth's orbital eccentricity show a 400 ky cycle in addition to the 100 ky cycle, with the two cycles being of comparable strength. Yet, the record of Earth's climate variations only shows clear evidence for the latter.
>>(d) Causality problem: Based on a numerical integration of Earth's orbit, a warming climate predates by about 10,000 years the change in insolation than supposedly had been its cause.
>>(e) Transition problem: No explanation is offered for the abrupt switch in climate periodicity from 41 ky to 100 ky that is found to have occured about a million years ago. Of these five problems with Milankovitch theory, the current theory clearly shares only (c).
Now how all of that mitigates the fact that CO2 lags temperature in the ice cores is something I'm not clear on. Al Gore tells me that anthropogenic increases in CO2 levels are the cause of Global Warming yet the data does not support any such assertion, in fact it flies in the face of such an assertion.
Another question, how do Malinkovitch Cycles affect the surface area of the ocean acting as a CO2 source?