Flicker: an explanation for the “D-O events” – rapid climate fluctuations of the last glacial period

The astonishing D-O fluctuations of >10 C in a few centuries: that’s real climate change

If the phrase “ice age” conjures up an idea of a uniformly frigid and icy climate, not changing for thousands of years – then that idea is wrong. Very wrong indeed. Glacial periods, such as the interval between our current Holocene and the previous Eemian interglacials, are characterized by rapid and even violent climate fluctuations. Most extraordinary are what are called the “Dansgaard–Oeschger Events” (D-O events) that happened during the final Pleistocene glacial stage. These episodes of rapid warming to nearly interglacial conditions occurred approximately every 3,000 years from 90 kyrs  to 12 kyrs ago. During these excursions, northern hemisphere temperatures as experienced in the North Atlantic increased by up to 10 degrees over one or two centuries only, before immediately cooling by a similar amount, again over just a few centuries.

Just think about that number for a moment. Ten degrees. The entire political juggernaut of climate change in the 21st century is focused on preventing only 1.5 degrees of warming, with shrill alarmist academics and other assorted opportunists earnestly warning of global catastrophe if temperature increases by just 2 degrees. Over the same one or two hundred year time frame of the D-O events. And we’re not talking about just one or two such events. In the latter part of the last glacial interval there were about 20 such D-O events, 20 massive excursions of climate change, from deep glacial to interglacial climate and back again, on just a few centuries. And note that in the context of the 100,000 year duration of the glacial-interglacial cycles in the late Pleistocene, a century is a very short time indeed.

There is a paradox in glacial climate dynamics. During the Pleistocene glacial period, especially in the last million years since the Mid Pleistocene Revolution (MPR), the climate spends 80-90% of the time in cold glacial conditions and 10-20 % of the time only in warmer interglacial conditions. In terms of analysis of stability and the attractor landscape analysis, this indicates that the glacial state is the preferred, more stable one. Except it is not more stable, but less. Climate temperature fluctuations are more frequent and of greater magnitude – notably the D-O events, during the cold glacial periods. By contrast during the warmer interglacials, while climate is still subject to significant variation, it is of lesser amplitude than during the glacial intervals. Climate is more stable. This is why human civilizations became established during the Holocene over the last 12,000 years, while behaviourally modern humans existed from 70,000 years ago. Take a human baby from 70,000 years ago and raise him or her in a family today, and he or she would function the same as out own children in school, exams, work and society. But the great climate instability of the glacial period forced humans to live a migratory and ever-adapting lifestyle, and nascent civilizations were no doubt cut short by a climate convulsion forcing dispersal in search of food and safety.

There is of course also fractality in the climate wavetrain shown in figure 1. The micro-interglacials within the late Pleistocene are miniature versions of full interglacials such as the Holocene, the Eemian and all the other previous ones.

The temperature wavetrain during the last glacial interval leading up to the Holocene interglacial is shown in figure 1.


fig 1 DO events last glacial interval

Figure 1. The temperature wavetrain measured by the GISP Greenland ice core (Alley 2000) showing the sharp spikes of rapid warming and cooling known as Dansgaard-Oeschger (D-O) events, numbered 1-20 going back in time. The “final” DO event, number 1, is given the name the “Bolling Allerod”.


The DO events are not evenly spaced in time but start slow in frequency, get faster then slower again. They have the highest frequency between 50 and 30 thousand years ago, becoming more spaced out both before and after this time. The most recent D-O event, number 1 in figure 1, has been given a name – the Bolling-Allerod event. The interval between the Bolling Allerod DO event and the subsequent jump in temperature to the Holocene interglacial (that we are now in) has also been given a name, the “Younger Dryas”, though why this was necessary is not clear since similar gaps exist between all 20 of the DO events in the last glacial interval.

The astrophysicist Ian Wilson made this comment on the D-O events on a WUWT blog post on May 14, 2018 ( https://wattsupwiththat.com/2018/05/14/poking-a-hole-in-the-latest-younger-dryas-impact-paper-uniformitarian-impact-craters-part-trois/#comment-2355409 )

A naive observer would look at figure 1 in your post and say that it looks like the signal that would be produced by a sputtering motor. At regular intervals, you have the Dansgaard-Oeschger (D-O) warm events firing up throughout the depths of the last Ice-Age. This is like someone repeatedly turning the ignition key on their car only to have the engine die every time. Using this analogy, you might suspect that something is trying to shift the Earth’s climate system from its glacial to its interglacial phases but there is not enough umph to get the transition to take place.

However, this all changes at the Bolling Allerod interstadial. A DO-event takes place upon a rising background temperature. This gives it that extra umph to temporarily kick the Earth’s climate from its glacial to its inter-glacial mode. Unfortunately, the engine sputters and the Earth’s climate system falls back into its glacial mode. It’s not until the next DO event at the end of the Younger-Dryas (Y-D) that key is turned and the engine finally fires up and begins running on all cylinders, entering the current interglacial period.


This is a good observation, the process Ian calls “sputtering” can also be called “flickering”. Ian went on to suggest that lunisolar tides with multi-millennial timescale played a role in forcing the climate system into these periodic sharp excursions.

In complex quasi chaotic systems such as the climate, there exist “attractors” that are preferred states. The system is attracted to those states so that the phase space of the system does not show uniform distribution between all possible states but clustering in space and time at a few preferred states called attractors (figure 2).


fig 2 attractorsfig 2b poincare maps

Figure 2. Of all the possible states in a multi-dimensional phase space, a system’s state will cluster around just a few states. These are called attractors (left). The phase evolution of the system is called the “Poincare map” (right).


At certain moments a system jumps from one attractor to another. However this transition is not always a clean process with a single jump. Sometimes as the system, located on one attractor, begins to feel the “pull” of another attractor, it behaves in a very undecided way, jumping between the two attractors until one or another of the attractors wins out, proving the strongest. Over the last 3 million years of the Pleistocene, the two main attractors have been glacial and interglacial.

The large scale timing of glacials and interglacials is under the controlling influence of the Milankovitch orbital cycles, of precession, obliquity and eccentricity. Interglacials happen (at least over the last million years) when all three of these cycles coincide to give the maximal climate warming effect. This brings the climate’s position in the phase space close enough to the interglacial warm climate attractor to allow it to make the jump successfully to the interglacial attractor. Without the climate being in this Milankovitch warm optimum, that happens only every 100,000 years, then the system still tries to jump to the interglacial attractor, but the jump doesn’t hold, and the system immediately falls back to the glacial attractor.



This abortive jumping between attractors is referred to as “flicker” in the literature of chaotic systems such as biological ecosystem dynamics. A good example of this is the following research paper by Dakos et al. (2013) whose abstract is given below:


Flickering is a signal of an impending alternate attractor

Flickering as an early warning signal. Dakos V, van Nes EH, Scheffer M. Flickering as an early warning signal. Theoretical Ecology. 2013 Aug 1;6(3):309-17.


Most work on generic early warning signals for critical transitions focuses on indicators of the phenomenon of critical slowing down that precedes a range of catastrophic bifurcation points. However, in highly stochastic environments, systems will tend to shift to alternative basins of attraction already far from such bifurcation points. In fact, strong perturbations (noise) may cause the system to “flicker” between the basins of attraction of the system’s alternative states. As a result, under such noisy conditions, critical slowing down is not relevant, and one would expect its related generic leading indicators to fail, signaling an impending transition. Here, we systematically explore how flickering may be detected and interpreted as a signal of an emerging alternative attractor. We show that—although the two mechanisms differ—flickering may often be reflected in rising variance, lag-1 autocorrelation and skewness in ways that resemble the effects of critical slowing down. In particular, we demonstrate how the probability distribution of a flickering system can be used to map potential alternative attractors and their resilience. Thus, while flickering systems differ in many ways from the classical image of critical transitions, changes in their dynamics may carry valuable information about upcoming major changes.


The flicker which accounts for the DO excursions during the last glaciation, show the inter-attractor behavior to have a fractal property. The DO events are abortive attractors, and could be called “micro-interglacials”. Chaotic systems as diverse as ecosystem predator-prey populations and climate systems are united in their chaotic dynamical behavior, in respect to attractor seeking behavior and flicker for instance.


Flicker can be influenced by periodic forcing from outside

Some scientists consider there to be solar or astrophysical forcing of climate changes. As just mentioned above, this is clearly true for the Milankovitch forcing of the glacial-interglacial cycle, most importantly by obliquity. However arguing that all climate variation is externally forced, whether by astrophysical effects or CO2, leads to the serious error of assuming or requiring the climate to be passive, only changing from outside forcing. However the climate’s clearly visible chaotic and attractor-seeking behavior makes it obvious that there is a strong internal dynamic to the climate and that the system is more than capable of “changing itself”. Now as I’ve said many times in discussions with such scientists convinced of strong astrophysical forcing of all climate events, it doesn’t have to be either-or in an exclusive way between astro forcing and internal oscillation. It can be both and. You can have external periodic forcing interacting with internal nonlinear oscillatory dynamics. It has a name – the “periodically forced nonlinear oscillator”.

There are well established experimental models in the literature of the periodically forced nonlinear oscillator. The BZ reaction (Belousov-Zhabotinsky) thin film chemical oscillator can be periodically forced with external light flashes. The same is true of the catalysed CO oxidation on a platinum surface. Even ocean tides can fall into this category when the coastal or estuarine geography is complex with a narrow neck. The heart beat (figure 3) is also generated from a network of cardiac muscle fibres acting as nonlinear oscillators and adding a pacemaker is an example of external periodic forcing of the system to try to get the beat regular. Here are papers describing such experimental and natural oscillatory systems:





figure 3a heartfig 3b heart traces

Figure 3. The heartbeat is a natural example of a nonlinear oscillator whose oscillation is of great importance to all of us!


Periodic forcing of a nonlinear oscillatory system can be either weak or strong. When it’s strong, then the system’s oscillation is the same as the external forcing oscillation. But when it’s weak the emergent system oscillation can be complex and differ considerably from the forcing oscillation. Then it becomes hard to link the system’s oscillation to its external forcing. But it still is externally forced. I suspect such scenarios of weakly periodically forced nonlinear oscillation are common in the ocean-climate system. There are hints at astrophysical periodicity but they are elusive and they come and go from the climate time record.

The advantage of the periodically forced nonlinear oscillator model in climate is that it does not require the astrophysical forcing to have all the energy needed to change the oceans for instance. The forcing can provide a pacing role only and the energy for the periodic transitions comes from the internal dynamics.

Milankovitch obliquity forcing of interglacial timing and duration is an example of strong periodic forcing. The cycles are slow enough for obliquity’s variation of solar input alone to change the ocean’s temperature, with the 6500 year thermal lag that you have pointed out previously.

However the DO events are much too rapid for external astrophysical forcing to provide all the energy. Here the internal dynamics and excitability caused by ocean-cryosphere interactions provides the energy, but there is still the possibility of weak external periodic forcing. How to prove that could be a challenge though.


Li and Born (2019) show data indicating that the DO events may well be unforced internal oscillations of the Northern Seas coupled system:


But we already know that systems with internal oscillatory dynamic can also be periodically forced, either strongly or weakly. So I think it’s a hybrid of both. Excitable media with internal oscillatory dynamic can be forced by external periodic forcing. From the sun, from lunisolar tides and so forth.



Have other glaciations in earth’s history has transitional flicker at their beginning and end? Yes they have.

I am proposing here that flicker is a transitional phenomenon that is always possible when the earth either enters, or leaves, a glacial state. For a borderline period the climate system will feel the pull of both glacial and non-glacial attractors sufficiently for chaotic flicker to occur. So – has flicker happened before or after any of the other deep glacial periods in earth’s history, of which there are several before the current Pleistocene.

Yes it has. Benn et al. 2015 discovered that after the deep Marinoan glaciation which lasted 15 million years between 650 and 635 million years ago, geological evidence of Milankovitch forced glacial-interglacial flicker was found. That’s the fun part of science – when a prediction arising from a hypothesis is confirmed! This is an amazing piece of geochronology, resolving changes at the level of thousands of years, hundreds of million years ago.


Benn, D., Le Hir, G., Bao, H. et al. Orbitally forced ice sheet fluctuations during the Marinoan Snowball Earth glaciation. Nature Geosci 8, 704–707 (2015). https://doi.org/10.1038/ngeo2502

Here is an extract from the abstract of this paper:

“Two global glaciations occurred during the Neoproterozoic. Snowball Earth theory posits that these were terminated after millions of years of frigidity when initial warming was amplified by the reduction of ice cover and hence a reduction in planetary albedo. This scenario implies that most of the geological record of ice cover was deposited in a brief period of melt-back. However, deposits in low palaeo-latitudes show evidence of glacial–interglacial cycles. Here we analyse the sedimentology and oxygen and sulphur isotopic signatures of Marinoan Snowball glaciation deposits from Svalbard, in the Norwegian High Arctic. The deposits preserve a record of oscillations in glacier extent and hydrologic conditions under uniformly high atmospheric CO2 concentrations.”


Note also, by the way: alternation between glaciation and interglacial, just like the Pleistocene, under “uniformly high CO2 concentrations”. Good to confirm that CO2 does not force warmth.

So next time you see something flickering, or sputtering, such as a car trying to start on a cold morning, or a neon light flickering when being turned on – remember that you are looking at a chaotic phenomenon that can explain systems as diverse as car engines, light bulbs, animal populations and glacial-interglacial climate changes.

A recent study by Mitchell et al. 2021 filed by Southampton University U.K. comes to the same conclusion as Benn et al. 2015. Mitchell and colleagues studied banded iron deposits in locations including Australia, China and Africa. They too uncovered the signature of Milankovitch forced oscillation between ice cover and open water near what was then the equator during the Sturtian glaciation 650-700 million years ago:


It thus seems that wherever glaciation is in a marginal, borderline state, such as at the equator in the “snowball earth” Cryogenian glaciations which (as Nick Eyles showed many years ago) were not really a solidly frozen earth – such borderline glaciation is entrained into alternate advance and retreat by Milankovitch orbital forcing.

2 thoughts on “Flicker: an explanation for the “D-O events” – rapid climate fluctuations of the last glacial period

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s