I have seen several graphs showing the prehistoric temperatures and CO2 concentrations derived from ice-core data. enter image description here My understanding is that CO2 and temperature correlate. I assume that Milankovich Cycles are what caused the variations observed in temperature in the ice record, but what specifically caused the CO2 to fluctuate?


2 Answers 2


Six plus years after the fact, I find the accepted answer to be a bit insufficient. With that,

but what specifically caused the CO2 to fluctuate?

Notice in the two graphs how the temperature drop and carbon dioxide drop gradually at the onset of a glaciation while the temperature rise and carbon dioxide rise at the end of a glaciation are comparatively sharp.

The onset of a glaciation was triggered by summertime temperatures in the far north being mild, mild enough for snow from the previous winter to last through the summer. A very widely used metric is whether the mean temperature over land at 65° north latitude was under 0° Celsius (32° Fahrenheit). Several things occur once snow can last over the summer:

  • The Earth's albedo increases due to the snow (and later, ice) coverage.
  • The oceans cool, thereby increasing the oceans ability to dissolve carbon dioxide as $\text{CO}_2$ dissolves much more readily in cold water as opposed to warm water.
  • Just as the Antarctic ice has trapped $\text{CO}_2$, so does the snow / firn / ice that is beginning to cover the far north.
  • The reduced $\text{CO}_2$ levels in the atmosphere result in even more cooling.

While this is a vicious loop, it nonetheless is a gradual process. This is why the onset of a glaciation is slow. Compare this with the end of a glaciation, which is very quick in comparison. Once conditions are ripe for warming

  • The melting glacial ice releases the $\text{CO}_2$ that it had trapped. At high enough pressures and low enough temperatures, the $\text{CO}_2$ embedded in the snow (and then firn, and then ice) forms clathrates. Clathrates are unstable not-quite molecules.
  • The warming temperatures cause the ocean temperatures to rise. Because $\text{CO}_2$ is less soluble in warm water than in cold water, the atmospheric $\text{CO}_2$ that had been absorbed by the oceans is released back to the atmosphere.
  • The increased $\text{CO}_2$ levels in the atmosphere cause the Earth to warm even more.
  • The melting ice exposes bare rock, which has a very low albedo (about 10%), compared with snow and ice (about 90%).

Like the onset of a glaciation, the end of a glaciation is also a vicious cycle. The vicious cycle at the end of a glaciation tends to be very vicious indeed as the melting glacial ice and warming oceans release their bottled-up $\text{CO}_2$ quickly once the process starts.

I assume that Milankovich Cycles are what caused the variations observed in temperature in the ice record

That is true, to some extent. One of the biggest mysteries with regard to paleoclimate is why the Earth switched from a 41000 year glaciation cycle to a 100000 year cycle. This happened about a million years ago. The problem is that while there is a roughly 100000 year cycle in one of the drivers of the Milankovich cycles, it supposedly is a very weak driver. The ~41000 year drivers are much stronger. While the onset of a glaciation is highly consistent Milankovich cycle theory (the conditions have to be ripe for temperatures at 65° latitude to remain below freezing all summer long), the end of a glaciation for the last million years is not. Why did ice sheets start persisting for ~100000 years?

This 100000 year problem remains a bit of a mystery, but one explanation I particularly like is that the ~41000 glaciations eventually scraped the regolith in the far north out to sea. The scraped clean rock gave the ice a nice solid surface to cling on to and thereby persist over what would have been the end of a weak ~41000 glaciation cycle. (Those older ~41000 glaciation cycles were weaker than are the current ~100000 year cycles.) The weaker ~100000 year signal eventually came to dominate due to hysteresis effects.

  • 2
    $\begingroup$ I'll eventually add references. It won't take ~100000 years, but it won't be today. $\endgroup$ Commented Mar 12 at 2:51

This article investigates the ocean's dissolved carbon content (ocean reservoir) and the seafloor's methane hydrates ("methane ice") as a source of CO2 fluctuations in ice cores. They discuss three ways the ocean may release CO2.

  • through physics (changes in temperature and ocean currents)

  • biologically or chemically

  • through general climate changes

Hope this helps.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.