Toward a new theory of ice-ages I (Introduction)

This was a quick outline of a theory I’ve been developing since the summer on the cause and progression of ice-ages. As I started describing the theory and the background, it mushroomed and spawned into several articles and then  … I hit a snag.
Crucial to my theory was that temperature was controlled by CO2 (but not in the normal way). I knew CO2 lagged temperature, but I kept going hoping I could explain this. Then it all came unstuck, not on the CO2 lag, but when I realised that there were very significant falls in temperature occurring around 16,000 years after the peak of each inter-glacial. (See I’m a CO2 denier) and CO2 did not fall in the same way. So, CO2 could not possibly be causing this cooling, so, CO2 clearly did not directly cause the ice-ages.
(Note in my defence: this presumed CO2 induced positive feedback during the ice-ages is one of the main reasons for believing in positive feedbacks in today’s climate. So, it never occurred to me that it would be so obviously flawed).
I’ve now been thinking about this since Xmas and not really making much progress. So, I have decided to start publishing my articles, knowing the conclusion is wrong, but hoping that the discussions might suggest a way forward to me, or that at worst, they prove useful to others.

Fig 1.1 A view into Glencoe Valley showing glacial U-shape (source: Dave souza)

Current state of Knowledge

In 1742 Pierre Martel, an engineer and geographer saw erratic boulders in the valley of Chamonix in the Alps and the locals said these erratic boulders were from when the glaciers extended much farther. From that first step, the idea gradually developed that much of the North of Britain, Europe, Asia and America were once under huge ice sheets and that these ice-sheets had fundamentally moulded the character of northern places like Scotland. Today we now accept that U shaped valley, with wide bottoms, steep sides and broken rocky tops that are characteristic of Scotland are a consequence of the Ice-age. However, a few centuries on there is still no real understanding of what causes ice ages.
Many ideas have been proposed as to why ice-ages may come and go: changes in the Earth’s orbit around the Sun known as Milankovitch cycles; the motion of tectonic plates resulting in changes in the relative location and amount of continental and oceanic crust on the Earth’s surface, which affect wind and ocean currents; variations in solar output; the orbital dynamics of the Earth-Moon system; and the impact of relatively large meteorites, and volcanism including eruptions of super-volcanoes, and atmospheric composition.
But whilst the cause and driver of ice-ages is not known, thanks to ice-cores we can understand at least some of the details of their progression.

Ice Cores (Temperature and atmospheric composition)

Ice cores can be used to reconstruct climate change over many thousands of years giving both atmospheric gas concentrations and temperature. Small bubbles of air are trapped in the ice when it freezes. From these gas concentrations can be measured, and although there is some doubt whether these accurately reflect the gas composition at the time** changes in the gas composition in the ice will reflect changes in the atmosphere.
Like gas composition changes in temperature can be estimated from the ice core but this time from the isotopic composition of the water molecules in the ice.  The isotopes used are ₁₆O that makes up 99.76 percent of the oxygen in water  and ₁₈O. Likewise ₁H is 99.985%  of the hydrogen in water but there is a heavier isotope ₂H, known as deuterium.
Sensitive mass spectrometers are used to measure the ratio of the isotopes of both oxygen and hydrogen in the ice cores. The water molecules in ice cores always have more of the lighter isotopes than sea water particularly in cold periods. This is because it takes more energy to evaporate the heavier water molecules with heavy isotopes and also because water molecules containing heavier isotopes are more easily precipitated. The mechanism is complex and far from a perfect thermometer, but in general, the isotope level is temperature dependent and so can be used to estimate temperature.

Vostok Ice Core

Shown below is just such a reconstruction of some of the various gases and temperatures from the Vostok ice core. This represents over 400,000 years of the earth’s history captured as changes in the composition of gas and water in the ice.  Dates are shown with the modern day at the right and measurements from the deepest and so oldest ice at the left. The three top traces (CO2 in blue, Temperature in red & CH₄in green) show a strong similarity to each other with a series of marked warming phases rising to a peak around 125,000, 240,000 & 320,000 and another around ~420,000 before present. After the peak we see a slow drop in values for around 100,000 years culminating in another sharp rise.

Fig. 1.2 Vostok ice core showing various proxies for climate over last 420,000:
CO2 (blue), proxy temperature (red), CH4 (green), δO18 ( brown), Solar Insolation (Inverted mauve).

Relationship of ice-core to solar insolation (sunshine) is shown by dash green lines.
There is a sharp change in relationship between solar and ice-core at 238,000 years BP (Highlighted by yellow shading)

Ice cores timing & melting

However, like temperature and the gases, the date shown is not a direct measurement. Instead, the age is counted like counting the rings on a tree but with one huge proviso: If the climate warmed enough to cause melting, then whole years, decades or even centuries could be removed from the record.
And there is very strong evidence that such ice-core melting has taken place. If we look at the orbital cycles of the earth as shown on the lowest trace, and the rate of change of ₁₈O (i.e. the proxy for temperature) we see that there is a strong correspondence between the rate of change of temperature and solar insolation as shown by the green dashed lines showing that the highest and lowest rates of temperature change correspondence with the falling and rising phases of insolation. The correspondence is close toward the right (modern day), but that there is an increasing offset between the two as we go back in time and at around 238,000BP, there is a marked change in this correspondence.
This is very strong evidence that part of the ice core has melted. That means the temperatures was above that of the present day when such ice-cores remain frozen continuously throughout the year.

Evidence for Ice Core melting up to 238,000BP

By comparing solar insolation with ₁₈O between 230,00 – 320,000 it appears readings in this period are being reported around 11,000 years too early compared to the calculated insolation. As there is no reason to believe the insolation is incorrect, this suggests that around 11,000 years of ice core missing. If so, the most likely time this occurred is just prior to 238,000 years so that 238,000BP represents the first time in 11,000 years that yearly snow accumulated. This is confirmed by other changes:
1. Narrower Peak: The peak of CO2 and temperature and methane is much shorter around 238,000 years.
2. Shorter interval to last peak: The length between the peak at 238,000 and 320,000 is shorter than the rest. We can see from the table below that if we increase it by the 11,000 years it is much closer to the other inter-peak times:

3. Peaks fail to reach typical peak value:  Corroboration evidence is shown by the horizontal green lines in Fig 1.2. all the CO2 levels except those at 238,000 years rise just above this line and then start. This is highly indicative of some kind of change that occurs in the climate when this threshold is reached. If so, we would also expect to see the period just before 238,000 years cross this line as well, but apparently the record for this is missing from the ice. Likewise, temperature and Methane levels are not as high as other periods – with the remarkable exception of present temperature which appears to be around 2-4C lower than expected.

Notes:
**Salby (see lecture) gives a compelling argument based on the “signature” of the gas, which more or less proves that the gas composition in the bubbles is likely to have been distorted in some way.  Therefore the composition of the gas bubbles will not be the same as the composition the the atmosphere. More research needs to be done in this area, but like so much in climate, such research is not welcome as it could only reduce the (false) certainty ascribed to ancient gas levels. Whilst Salby does not himself suggest a mechanism, the three obvious ones (for those who have looked) are:

1. Clathrates: clathrates form when CO2 interacts with water at high pressure. Then as the ice is de pressurised as it is brought up, micro-fractures occur and the air pressure in the bubbles forces out the liquid clathrates through these micro-cracks thereby reducing the CO2 in the core.
2. As anyone looking at melted snow can see, snow is not 100% pure. If contaminated by algae or other photosynthesising life, then these will convert CO2 to hydrocarbons. Whilst few in number, the next effect could be significant over the years it takes to fully enclose the air. These may be consumed or be decomposed to volatile hydrocarbons which acting as an anti-freeze can melt the ice and leach out.
3. The third effect is intra-ice diffusion. This may reduce the peaks and increase the troughs.