It is some 40 years since the last explosive development in understanding the way the world works. In 1976 verification of Milutin Milanković’s astronomical theory to explain cyclical climate change as expressed by surface processes has had a similar impact as the underpinning of internal processes by the emergence of plate tectonics in the preceding decade. Signals that match the regularity of changes in the Earth’s orbital eccentricity and the tilt and precession of its axis of rotation, with periods of roughly 96 and 413 ka, 41 ka, 21 and 26 ka respectively, were found in climate change proxies in deep-sea sediment cores (oxygen isotope sequences from benthonic foraminifera) spanning the last 2.6 Ma. The findings seemed as close to proof as one might wish, albeit with anomalies. The most notable of these was that although Milanković’s prediction of a dominant 41 ka effect of changing axial tilt, the strongest astronomical forcing, had characterised cooling and warming cycles in the early Pleistocene, since about a million years ago a ~100 ka periodicity took over – that of the weakest forcing from changing orbital obliquity. Analysis of sedimentary cycles from different episodes in earlier geological history, as during Carboniferous to Permian global frigidity, seemed to confirm that gravitational fluctuations stemming from the orbits of other planets, Jupiter and Saturn especially, had been a continual background to climate change.

All manner of explanations have been offered to explain why tiny, regular and predictable changes in Earth’s astronomical behaviour produce profound changes in the highly energetic and chaotic climate system. Much attention has centred on the mathematically based concept of stochastic resonance. That is a phenomenon where weak signals may be induced to show themselves if they are mixed with a random signal – ‘white noise’ spanning a great range of frequencies. The two resonate at the hidden frequencies thereby strengthening the weak, non-random signal. Noise is already present in the climate system because of the random and highly complex nature of the components of climate itself and the surface processes that it induces.

The latest development along these lines suggests that something quite simple may be at the root of inner complexities in the climatic history of the Pleistocene Epoch: the larger an ice sheet becomes and the longer it lasts the easier it is to cause it to melt away (Tzedakis, P.C. et al. 2017. A simple rule to determine which insolation cycles lead to interglacials. Nature, v. 542, p. 427-432; doi:10.1038/nature21364). The gist of the approach taken in the investigation lies in analysing the degree to which the onsets of major ice-cap melting match astronomically predicted peaks in summer insolation north of 65° N. It also subdivides O-isotope signals of periods of sea level rise into full interglacials, interstadials during periods of climate decline and a few cases of extended interglacials. Through time it is clear that there has been an  increase in the number of interstadials that interrupt cooling between interglacials. Plotting the time of peaks in predicted summer warming closest to major glacial melting events against their insolation energy is revealing.

Before 1.5 Ma the peak energy of summer insolation in the Northern Hemisphere exceeded a threshold leading to full interglacials rather than interstadials more often than it did during the period following 1 Ma. Although Milanković’s 41 ka periodicity remained recognisable throughout, from about 1.5 Ma ago more and more of the energy peaks resulted in only the partial ice melting of interstadial events. The energy threshold for the full deglaciation of interglacials seems to have increased between 1.5 to 1.0 Ma and then settled to a ‘steady state’. The balance between glacial growth and melting increasingly ‘skipped’ 41 ka peaks in insolation so that ice caps grew bigger with time. Deglaciation then required additional forcing. But considering the far larger extent of ice sheets, the tiny additional insolation due to shifts in  orbital eccentricity every ~100 ka surprisingly tipped truly savage ice ages into warm interglacials.

Resolving this paradox may lie with three simple, purely terrestrial factors associated with great ice caps: thicker and more extensive ice becomes warmer at its base and more prone to flow; climate above and around large ice caps becomes progressively colder and drier, so reducing their growth rate; the more sea level falls as land ice builds up, the more the vertical structure and flow of ocean water change. The first of these factors leads to periodic destabilisation when ice sheets surge outwards and increase the rate of iceberg calving into the surrounding oceans. Such ‘iceberg armadas’ characterised the last Ice Age to result in sudden irregularly spaced changes in ocean dynamics and global climate to return to metastable ice coverage, as did earlier ones of similar magnitude. The second factor results in dust lingering at the surface of ice caps that reduced the ability of ice to reflect solar radiation back to space, which enhances summer melting. The third and perhaps most profound factor reduces the formation of ocean bottom water into which dissolved carbon dioxide has accumulated from thermohaline sinking of surface water. This leads to more CO₂ in the atmosphere and a growing greenhouse effect. Comforting as finding simplicity within huge complexity might seem, that orbital eccentricity’s weak effect on climatic warming – an order of magnitude less than any other astronomical forcing – can tip climate from one extreme to the other should be a grave warning: climate is chaotic and responds unpredictably to small changes …