The Ice Man Cometh

Have you ever wondered when the next ice age will begin? According to Physicist and fellow blogger Clive Best we may already be past the optimum temperature of the Holocene Interglacial and be sliding back towards the next ice age. Clive has fitted the harmonics of combined Earth orbit cycles to a high resolution temperature record derived from carbonate microfossils from 57 ocean drilling sites[1]. A combination of the 100,000 year eccentricity cycle and 41,000 year obliquity cycle provides an excellent fit to the ocean microfossil temperature record (Figure 1). Since Earth’s orbital cycles are known with precision, this can be used to forecast what comes next. His conclusion is rather chilling.

Figure 1 The black line is based on a stack of 57 d18O temperature records for carbonate microfossils from Atlantic, Pacific and Indian Oceans [1]. The way this is plotted, cold is up and warm (interglacials) is down. The blue line is combined 100,000 year and 41,000 year orbit cycles [2,3]. The excellent fit supports the theory, first proposed by Milankovitch, that variations in orbital parameters have controlled the onset and termination of glacial cycles on Earth for the last 2.6 million years [4]. Projecting this into the future shows that Earth is close to the turning point of the cycle and it is downhill from here towards the next ice age. Chart from Clive Best.


  • The LR04 stack of 57 temperature records published by Lisiecki and Raymo [1] provides one of the highest resolution records for global ocean temperature variation over the last 5.3 million years (Ma). The data provide overwhelming evidence for a major cooling trend that resulted in the onset of extensive N hemisphere glaciation about 2.7 Ma ago [4].
  • Cyclical variations in the LR04 stack bear witness to complex interactions of geological (plate tectonic) and orbital cycles upon Earth’s climate and the relative importance of individual components vary with time.
  • Milankovitch orbital cycles have three components 1) eccentricity 100,000 years, 2) obliquity 41,000 years, 3) precession 23,000 years. Clive Best finds the precession cycle is absent from the marine LR04 stack but a combination of eccentricity and obliquity matches the observed LR04 record with some precision [2,3]
  • Global cooling resulted in the onset of glaciation about 2.6 million years ago. For 1.7 million years, glacial cycles were dominated by the 41,000 year obliquity orbital cycle. But then, about 900,000 years ago the 100,000 year eccentricity cycle kicked in resulting in longer ice ages, much larger oscillations in temperature with brief and fragile interglacials that last 10 to 20 thousand years. We are 12,000 years into the Holocene interglacial.
  • The empirical relationship between orbital cycles and climate change in the past can be used to forecast what will happen next and the result is chilling. It is possible that we reached peak temperature 2000 years ago in the Roman Warm Period [1] and that we are already heading towards the next ice age [2,3].
  • The overall situation is made more complex by the superposition of shorter time scale cycles of approximately 1500±500 years duration in the Holocene known as Bond cycles [5]. Bond cycles are modulated by solar geomagnetic activity and give rise  to what we know about more recent historical climate cycles such as The Roman Warm Period.
  • The most recent of these cycles is The Modern Warm Period. With solar geomagnetic activity in steep decline, the cold phase of the Bond cycle may now be aligned with the cold phase of the eccentricity and obliquity cycles although it is too early to say yet if the Modern Warm Period is over.

The LR04 benthic foraminifera stack

The Lisiecki-Raymo (LR) 04 stack of benthic foraminifera temperatures is one of the amazing data sets to emerge in recent decades that records details of Earth history. The data come from 57 records of variable duration from the Pacific, Atlantic and Indian Oceans (Figure 2).

Figure 2 The 57 d18O records analysed by Lisiecki and Raymo [1] which when added together produce the summary stack shown in Figure 3.

The records go back 5.3 million years, but it is really the last 500,000 years that are of greatest interest for understanding recent ice age cycles. Oxygen has three isotopes – 16O, 17O and 18O. In creatures with calcium carbonate skeletons the ratio of 16O and 18O measured relative to a standard (d18O) varies in response to changes in seawater composition and temperature. Temperature dominates. The d18O composition for calcium carbonate shells, therefore, provide a proxy temperature record (Figure 3).

Figure 3 The d18O temperature record for the benthic carbonate stack in blue, left hand scale [1] and the dD derived temperature record for ice from the Vostok ice core in red, right hand scale [6,7]. The alignment is quite striking but prior to 250,000 years there are clearly time scale calibration issues.

The records shown in Figure 2 have been merged to produce the composite stack shown in Figure 3 (blue line) that may be compared with the temperature record from the Vostok ice core (red line). The Vostok temperature record is based on the hydrogen and deuterium isotope composition of the ice (dD). The similarities of these two records in broad terms is striking although earlier than 250,000 years ago there are clearly some time scale calibration problems resulting in data offset on the x-axis.

Figure 4 The dD records for ice from the Vostok (and Epica) ice cores give the impression of a fairly flat temperature evolution for the Holocene. This observation is partly responsible for the notion that the Holocene climate has been uniform and that recent deviations from static climate may be attributed to Man. The evidence from LR04 is distinctly different with gradually warming ocean temperatures for the last 12,000 years. Note that the resolution on LR04 is 1,000 year intervals, too coarse to capture any recent warming that may have occured.

While the broad alignment of LR04 with Vostok is striking there is also one striking difference and that is during the all important last 12,000 years of the Holocene. Where Vostok (and Epica) show a largely flat temperature evolution for the Holocene, LR04 does not (Figure 4). The benthic carbonate record from 57 sites around the globe shows rapid ocean warming beginning 18,000 years ago. 12,000 years ago the rate of warming slowed but did not stop. The peak (so far) was during the Roman Warm period and there are signs that natural warming has slowed to a near stop in anticipation of the next turn downwards.

Milankovitch orbital cycles

Serbian physicist Milutin Milanković was among the first to suggest that temporal changes in Earth’s orbital geometry around The Sun gave rise to the cyclic rythm of ice ages and interglacials. Clive Best has found that of three orbital parameters it is the 100,000 year eccentricity cycle and the 41,000 year obliquity cycle that have greatest impact on the pattern of ice ages over the last 900,000 years.

Figure 5 Chart slightly adapted from Clive Best [2,3]. On this chart the present day is to the right. The green line shows 100,000 year eccentricity cycle. Zero eccentricity (left hand scale) = circular orbit. Note the 100,000 year cycle is superimposed upon a 400,000 year super-cycle. The black line is the LR04 benthic stack. The arrows show where we are on the eccentricity cycle today and where we were 400,000 years ago. The stage is set for a rapid decline into the next ice age.

Figure 6 Chart slightly adapted from Clive Best [2,3]. On this chart it is the lower blue line that traces out the obliquity cycle (left hand scale). Close scrutiny shows that rapid glacial terminations correspond to rising obliquity (increasing tilt of Earth’s axis) and that ice ages begin with falling obliquity. Note that the current obliquity cycle has already begun to fall.

Bond cycles

In addition to the orbital cycle modulation of ice ages and interglacials there are second order impacts upon Earth’s climate, potentially caused by cyclical change in solar activity itself on a time scale of 1500±500 years (i.e. a frequency that varies from 1,000 to 2,000 years). Cyclic variability to climate caused by the Sun is widely contested. These cycles are identified in Greenland ice cores where they are called Dansgaard-Oescheger events [8] and in the detrital content of North Atlantic ocean sediments where they are called Bond events [5]

Bond et al [5] measured the clastic component of deep ocean sediments from 4 sites in the North Atlantic, recognising 3 specific types of debris derived from specific localities and deposited by melting drift ice as it passed “overhead” (Figure 7).

Figure 7 Bond et al [5] recognise three types of lithic detritus in N Atlantic bore holes that have specific provenance, i.e. Icelandic volcanic glass, detrital carbonate and hematitie stained grains, that are interpreted to have been transported by drift ice. The stack of 4 wells is shown in the lowermost panel. 

A picture of cyclic change in drift ice emerges from the combined stack shown as the lowermost panel in Figure 7. In recent centuries, the 4 localities have been receiving virtually no ice rafted debris consistent with the observation that drift ice is largely absent from the North Atlantic today. An intriguing aspect of Bond et al’s work is that the observed cycles correlate with cycles in the cosmogenic nuclide 14C whose production is modulated by the strength of the solar wind and solar geomagentic activity.

This work has for many years intrigued me especially since Bond cycles are easily linked to what we know about historical climate change for the last 2,000 years or more as illustrated in Figure 8.

Figure 8 The Bond stack compared with CO2 from ice cores and Mauna Loa and a number of historical episodes. Skara Brae is a famous Neolithic archaeological site on the Orkney Islands that overlaps with the warm period between Bond cycles 4 and 3. RWP = the Roman Warm Period maps out as an extended anomalous warm period that is perhaps also picked out as a blip on the LR04 benthic stack (Figure 4). Deterioration of climate as NW Europe descended into the Dark Ages (DA) cold period may have contributed to the fall of The Roman Empire. The down pointing arrows mark the expedition of Eric the Red to found Viking settlements in southern Greenland (985 AD) and the loss of the Greenland Knarr in 1380 AD that marked the end of the Greenland settlements. The habitation of Greenland by the Vikings corresponded with the Medieval Warm Period (MWP) that gave way to The Little Ice Age (LIA) and ultimately the Modern Warm Period.

Bond et al interpret their data in terms of cyclical change in the pattern of atmospheric circulation and ocean currents that are linked to cyclical change in solar geomagnetic activity. The activity of the Gulf Stream is particularly important and Bond et al suggest that it is periodically cut off by the cold Labrador current allowing drift ice,  blown by northerly winds, to advance much further south into the N Atlantic than occurs at the present time. A larger than previously understood variance in the spectral output from The Sun, linked to changes in geomagnetic activity, may begin to provide understanding of the physical process [9].

The data and findings of Bond et al have been contested [10, 11]. The data shown in Figure 7 are either real in which case their findings are extremely important, or they are somehow false. Unfortunately Bond is dead and his co-workers have thus far failed to respond to my emails.

The cause of Pliocene-Pliestocene glaciations

The reason why the Earth entered a period of cyclic glacial activity 2.6 million years ago is poorly understood although as shown in Figure 9 this was in part due to a longer term cooling trend.

Figure 9 From Clive Best [3]. On this chart cold is up and warm is down, the present day is to the left.  In black is the remarkable LR04 d18O record. Hence this chart shows long term cooling and increasing amplitude of glacial oscillation. In blue are variations in insolation due to orbital parameters. From 5.3 million years to 900,000 years ago, the 100,000 year oscillation is all but absent in the LR04 data which oscillates solely in response to the 41,000 year obliquity cycle. But since 900,000 years ago the 100,000 year cycle kicks in together with the 41,000 year cycle.

One of the enigmatic aspects of the orbital coherence of ice ages is that the changes in insolation are too small to account for the substantial changes in temperatures that occur and so there needs to be some form of amplification mechanism. In ice cores, CO2 concentration, trapped in air bubbles in ice,  vary semi-synchronously with temperature. Warmists argue that this is the amplification mechanism hence imparting importance to CO2 as a controlling variable on Earth’s climate. Sceptics point out that CO2 lags temperature and is therefore a response to temperature change and not the cause of it. They also point out that turning points in the temperature record occur at maximum CO2, i.e. cooling phases begin when CO2 is at a maximum. Furthermore, the small changes in CO2 concentration are also insufficient to account for the large swings in temperature between ice ages and interglacials.

It has been suggested by a number of sources that the underlying geological causes responsible for the onset of glaciation are likely to be a combination of plate tectonic events: 1) the closure of the Panama Isthmus that created the American “super continent” believed to have a profound effect on ocean circulation and leading to the establishment of the Gulf Stream; 2) the drift polewards of Antarctica that will have impacted the amount of land ice, sea level and the distribution of sea ice in the southern hemisphere (Figure 9); 3) the uplift of the Tibetan Plateau. It is beyond the scope of this post to delve deeper into the underlying causes of the ice age cycles but informed input in comments would be welcome. Clive has already offered his explanation [12].

One observation I would make is that the 41,000 year temperature oscillation signal in the oceans is present all the way through the 5.3 million year LR04 record (Figure 9). Hence, rhythmic temperature oscillations were taking place long before the onset of N hemisphere glaciation 2.7 million years ago. Despite major plate tectonic adjustments over the past 5 million years this 41,000 year heart beat is pervasive.

When will the next ice age begin?

Clive Best has combined the eccentricity and obliquity cycles to provide the excellent fit to the observed LR04 stack shown in Figure 1. Since future orbital cycles are known with precision this algorithm can be used to forecast when the next ice age is due to begin. Looking at the detail (Figure 10) suggests we are right on the turning point of the combined orbital cycles. We are at the pinnacle of the climatic optimum and should be grateful for that.

Figure 10 Detail of the recent past and future from Figure 1. Note that in this case time is passing from left to right, cold is up and warm is down. Chart from Clive Best [2,3]

Clive has this to say:

The current glacial cycle has similar orbital parameters to the one occuring 5 glaciations ago. That interglacial warm period lasted 10,000 years or about the length of the current one. We are due for another ice age…  Next Ice Age due to start in ~1000 years time.

There is no immediate need to panic since natural climate variability proceeds as a series of bumps that last 100s to 1,000s of years. Northern Europeans should be concerned by the current solar slumber that is reminiscent of recent cold periods such as The Little Age and should be prepared for perhaps 30 years or more of periodic extreme cold winters and have the power generation infrastructure in place that is fit for that purpose.

[1] Lorraine E. Lisiecki and Maureen E. Raymo 2005. A Pliocene-Pleistocene stack of 57 globally distributed benthic D18O records. PALEOCEANOGRAPHY, VOL. 20

[2] Clive Best: A challenge for climate models

[3] Clive Best: Phenomenology of Ice Ages

[4] Mark Maslin & Jonathan Adams: The onset of Northern Hemisphere Glaciation during the Tertiary and Quaternary

[5] Gerard Bond et al. 2001. Persistent Solar Influence on
North Atlantic Climate During the Holocene. SCIENCE VOL 294

[6] Petit, J.R et al, 1999. Climate and atmospheric history of the past 420,000 years from the Vostok ice core, Antarctica. Nature 399: 429-436

[7] NOAA NCDC Vostok Ice Core

[8] Dansgaard–Oeschger event

[9] Sarah Ineson et al 2011. Solar forcing of winter climate variability in the Northern Hemisphere. NATURE GEOSCIENCE | ADVANCE ONLINE PUBLICATION

[10] JOHN T.ANDREWS 2009. Seeking a Holocene drift ice proxy: non-clay mineral variations from the SW to N-central Iceland shelf: trends, regime shifts, and periodicities. JOURNAL OF QUATERNARY SCIENCE (2009) 24(7) 664–676

[11] Andrews et al 2009. A robust, multisite Holocene history of drift ice off northern Iceland: implications for North Atlantic climate. The Holocene 19,1 (2009) pp. 71–77

[12] Clive Best: The real cause of Ice Ages – Resonant dust clouds?

This entry was posted in Climate change and tagged , , , , , . Bookmark the permalink.

43 Responses to The Ice Man Cometh

  1. Roger Andrews says:

    A few years ago I posted my pet theory on what causes ice ages in a comment at:

    Here’s the relevant part of the comment:

    “I have concluded that we’ve got everything backwards. We think ice ages are caused by climate change, but it’s actually the other way round. Climate change is caused by ice ages. And ice ages are controlled by the cyclic behavior of the NH ice sheets.

    “First, three governing assumptions:

    “1. For at least 400,000 of the last 500,000 years the Earth has been in an ice age condition. In other words, ice ages are the norm. We don’t have to explain them. What we do have to explain are the interglacial periods. These, not the ice ages, are the anomalies.

    “2. Interglacials occur only in the NH, where ice sheet extent ranges from maybe as much as 30 million sq km during glacial maxima down to a small fraction of that during interglacials. They don’t occur in the SH. Antarctica has been ice-bound for millions of years.

    “3. The problem of explaining what causes ice ages therefore comes down to explaining what causes NH interglacials.

    “Now the explanation, step by step:

    “Step 1: The NH exits an interglacial period and the NH ice sheets begin to advance again. (Note, we don’t have to explain why they advance. Ice ages are the norm. The Earth is simply returning to its natural climatic – or if you prefer, “balanced” – state.)

    “Step 2: The NH ice sheets continue to advance for tens of thousands of years. But eventually a point is reached where they become top-heavy, and then they begin to behave like surge glaciers, calving large quantities of ice into the sea once every five or ten thousand years. These “Heinrich Events” (HEs) are identified from ice-borne detritus in sea bed cores, and they are uncorrelated with temperature, CO2, dust, solar activity or any other climatic variable. They’re purely a result of ice sheet dynamics.

    “Step 3: HEs are usually not large enough to upset the balance between ice sheet stability and albedo, but eventually we get one that is. The last HE, which occurred about 15,000 years ago, was one. It calved off an enormous amount of ice at the margins of the ice sheets (maybe as much as 10,000 sq km) and the ice that was left couldn’t fill the hole fast enough to offset the warming impacts of the albedo change. So the climate warmed and the ice sheets continued to retreat, and now we are back in an interglacial.

    “Step 4: The interglacial ends. How? Straight answer, I don’t know, but the trigger is again probably physical rather than climatic. So for the time being I’m going to stick with the theory that the NH ice sheets start growing again when the Arctic Ocean becomes ice-free and/or when the Northern Baltic and Hudson Bay become dry land.”

    Now have at me 🙂

    • Euan Mearns says:

      Now have at me

      OK 🙂 Roger, on this occasion I do not agree with your thesis apart from the very first part – it is the interglacials that are currently anomalous – why not just one great big long ice age? I do not agree because it does not begin to address the alignment of orbital and solar variables presented here with the observed cyclic change in temperature.

      In interglacials, CO2 rises to a maximum and we get plunged back into an ice age. During ice ages, ice extent and albedo increases to a maximum and we get ice sheet collapse.

      My own hunch, that I was shy of putting in the main text, is that there are long time scale solar cycles, aligned with Earth orbital cycles (solar system gravitational forces causing both), that we have not observed directly. One thing for sure, observing the Sun with a satellite for 30 years, observing little change and applying that to all time is rubbish. I simply don’t know enough about the physics of Earth orbital cycles and the orbital cycles of the Planets etc and how this may stir the Sun. I have the luxury of being able to delete rubbish comments should the need arise 😉

      • Kit P says:

        Roger do not give up your day job, you are not a scientist. Neither am I but as an engineer I had to learn a lot of science. Knowing why is interesting but knowing what is essential. First you must carefully define terms. What do you mean by climate?
        Clearly 65 million years ago, the climate was different. There were no ice caps or glaciers. Then there was a cooling trend until about 2 million years ago. We are now in an ice age where the mass of water in glaciers and polar ice caps increases and decreases slowly.
        The climate is a stable system. To define that, a unstable system is a system where a small change causes a large effect. While we may not be able to correctly correlate the ’cause’ and ‘effect’, over the last 2 million years, it seems reasonable to predict that it is more likely glaciers will start to expand again in the next 1000 and man will not be able to prevent it by burning coal.

      • Roger Andrews says:

        Hi Euan:

        I too started out as a confirmed Milankovician but I’ve since concluded that there are just too many objections to the Milankovic cycles = ice ages theory, with probably the main one being that the cycles simply don’t fit the ice ages. Here’s a plot of LR04 against the 100,000 year cycle for the last million years, with the phase of the cycle adjusted to peak during the current interglacial. Some interglacials and glacial maxima miss the Milankovic peaks and troughs by tens of thousands of years and before 800,000 BP the correlation is negative:

        Carl Wunsch of MIT also doesn’t think Milankovitch had much to do with it:

        I hope this isn’t a rubbish comment 😉

        • Clive Best says:

          The best evidence that Milankovitz cycles must have some effect on NH glaciation is the paper by Gerard Roe, In defense of Milankovitch, Geophysical Research Letters, Vol 33, He plotted the change in insolation at 60N in summer against the rate of change in Ice Volume. There is a pretty good correlation. However this still does not explain the interglacials nor the 100 year cycle.

    • Clive Best says:


      Nice idea – and yes the natural climate of the Earth currently is an ice age with large ice sheets covering the N hemisphere. Somehow though we have to explain why such HEs occur in response to orbital forcing. Why should the threshold only be reached when the ellipticity is at a maximum? Summer temperatures in the N hemisphere vary with obliquity effecting seasonal melting. One can also imagine why the increase/reduction in ice albedo accelerates change. But the d18O is regular rather than chaotic or non-linear, so I suspect there is an external cyclic explanation. I don’t know the answer.

  2. Kit P says:

    OMG we are all going to die. One of the fun parts of working on the geological repository for spent nuclear fuel is being around geologists who tend to take the long view. Knowing how long pumps to remove decay can be off without a problems, is an important part of my job. At Yucca Mountain, geologists determined that the area had never been covered by glaciers although precipitation would be higher during glacial periods. Therefore they can predict, give or take 100,000 years, when trace amount of long lived fission products will reach ground water and what the does would be for those drinking the well water. I suggest that the scale of the curves be in ice ages not years. Three ice ages from now …..
    I used to worry about when a geological repository would be ready. Now I think it needs to be ready before the next ice age. This is after King Obama is dead. I also think we will reprocess spent fuel. I never understood why we can recycle beer cans and not spent fuel.

  3. zaphod42 says:

    Good stuff, all.

    Doesn’t it seem that there is an element missing from all of this, represented in the chart showing CO2 levels. One comment was that we could not forestall the coming ice age by burning coal. And, perhaps we could not if we were just starting. Since we have been burning fossil fuels in quantity for several hundreds of years, we have a condition never before seen. Outside forcing by homo sapiens sapiens.

    The question is not whether or not we should be entering an ice age. Almost certainly we should. And it would be a natural event – similar to other cycles for the past several millions of years. The question is, are we actually changing the planet sufficiently to alter the natural cycles?

    Perhaps the oceanic circulatory patterns are such that we cannot do that. Still, since we are in new terrritory, we must wait and see. Will our interference create a new status quo? Is it even possible to do so? The IPCC reports to date have indicated it may be, and that our actions might destabilize our climate, causing perhaps foreseen and perhaps unforeseen forcings. At this point it seems to be a crap shoot…

    As to the nuclear power debate, I have long held that such will be needed for the short to mid term, and in a closed system fails just as fossil fuels do when source rock becomes impractical or impossible to mine. How many times can nuclear fuel be recycled?


    • Kit P says:

      “How many times can nuclear fuel be recycled? ”

      About 5% of fissionable atoms in fuel assemblies actually fissions and become fission products.

    • Euan Mearns says:

      Hi Craig, welcome to Energy Matters. You should gather I have sceptic leanings 😉 but welcome reasoned debate from all sides. I currently sit on the low end of the IPCC spectrum of opinion on climate sensitivity somewhere in the vicinity of 1, which in my judgement means risks to climate stability from CO2 are manageable but in truth unknown. When climate “experts” tell you they can forecast climate hundreds of years into the future but the same experts cannot tell you what causes ice age cycles then you know straight away that said experts are buffoons.

      The question is, are we actually changing the planet sufficiently to alter the natural cycles?

      This is a very good question. Many of the climate experts I refer to above have done all they can to write natural climatic cyclicity out of history and my own position is that only through understanding past natural cycles and their cause can we begin to understand how Man may be impacting the natural rhythm. Attributing past natural climate variability to Man is pretty unscientific in my view and in this regard “climate science” has failed humanity.

    • Clive Best says:

      CO2 has a rather weak forcing on climate. The effective first order “warming’ is 3.5 Watt/m2 for each doubling in atmospheric CO2. In other words 3.5 watts/m2 of extra forcing for CO2 levels of ~600 ppm and 7 watts/m2 at levels of 1200ppm. This logarithmic dependence on CO2 means that most of the warming effect by man has probably already occurred. So the question is how much CO2 do we need to keep in the atmosphere to avoid the next Ice Age ? We would have to offset the reduction in summer insolation in the Arctic circle which has already started. As a wild guess we would probably need to keep CO2 levels above 600 ppm for the next 70,000 years!

  4. Clive Best says:


    Thanks for posting this. I have long thought that explaining the dynamics of Ice Ages is the key test as to whether scientists can seriously claim to understand the climate. While climate models remain unable to reproduce the glacial cycle, we should likewise remain sceptical of their ability to project anthropogenic global warming. In particular are human CO2 emissions the primary driver future climate change? This is blatantly not the case over the last 800,000 years because glacial cycles drive CO2 levels, not the inverse.

    It is true that given a set of initial conditions corresponding to the last glacial maximum (LGM) static energy balance models can be made to match, but the dynamics of why interglacials happen every 100,000 years remains totally unexplained. I am pretty sure that we are missing something fundamental yet to be discovered. This is likely to be an as yet unknown astronomical effect changing not just the distribution of solar radiation but also its (spectral) intensity. This must have either been amplified 5 million years ago by plate tectonics or some other as yet unknown geophysical phenomena.

    Scienceofdoom also has recently got his teeth into this subject with a series of posts which try to get to the bottom of the mystery. These are a bit academic but well worth a read.


  5. Euan Mearns says:

    Roger, thanks for the paper. Tried reading it last night, problems with comprehension that I put down to wine. Still struggling to fully understand it today 🙁 If i get the gist, Wunsch has two main complaints. 1) for the 100,000 year cycle the small number of events limits statistical significance and 2) the physical impact of orbital forcing (as it is understood) is too small to account for ice age cycles. I guess myself and Clive are both saying that orbital forcing is not properly understood and my hunch would be solar variability that is controlled by the same forces that controls the Earth orbital cycles, i.e. the gravitational / angular momentum rhythm of the solar system – though that is pure speculation.

    This leaves two questions, 1) what brought on the ice ages and 2) is the 41,000 / 100,000 year cyclicity real

    My own view is that plate tectonic events – Antarctica drifting poleward, piling up ice, shifting albedo, lowering sea level, creating land bridges could be significant; the closure of the Panama isthmus could be significant; uplift of Tibet could be significant and I’d add uplift of western cordillera (though i’m unsure on timing there); could all combine to create the conditions for ice ages. And this leaves the question are they cyclical and if so why?

    I’d really need to sit and play with data myself to answer this. Looking at Clive’s 5.3 Ma chart I see a high degree of coherency with the 41,000 year signal right the way through although there are some instances where it breaks down – but we are living in a multivariate natural world. But then we need to take into account that the data may have been tuned to the orbital signal – but they will only have adjusted the paleontological dates within limits I presume.

    In my judgement the 41,000 and 100,000 year signals do exist. The 41,000 year signal goes way back to before the time of ice ages and cannot be explained by rhythmic growth and collapse of ice. Something does happen about 900,000 years ago to change the character – not just the period but also the amplitude. But the causal relationship remains a mystery that can probably only be answered by hanging around for 10,000 years.

    I am hopeful that the current solar slumber will cast some light on this problem in my life time.

    • Clive Best says:

      The 41,000 year cycle in the tilt of the Earth’s rotation axis (obliqiuity) is the only Milankovitch cycle that has a clear signal throughout the last 5 million years. See for example this graph . One obvious reason why this must be the case is to consider what happens to the climate zones. The Arctic circle increases in size as the tilt increases. So during winter a larger area is in darkness with consequent increase in snow and Ice. Likewise the Tropics increase in size which squeezes out the central temperate zone. Probably this also then effects the ocean circulation.

      The real mystery is what caused the switch to 100,000 saw tooth oscillations ?

      • Euan Mearns says:

        Thanks for the chart Clive – it will save me having to plot it. The correlation is stunning and we need to recall that the comparison is with a proxy for “mean ocean temperature”. Either side of 2 million years there are two cycles that show opposite correlation. Understanding the cause of that is likely to help understanding the overall process. I like your explanation about increasing size of Arctic, presumably the Antarctic circle does the same. Would be good to see the extreme positions plotted on a sphere. This essentially takes you away from the insolation argument towards longer winters, more snow and ice that is not melted by the longer summers leading to accumulation. But I still think the temperature oscillation in LR04 has to be the Sun. In terms of the solar system, what is happening every 41,000 years to tilt and untilt Earth?

        Also, are you able to say how 1 per mill d18O translates to dT. I’m supposed to know this stuff since I used to sell this sort of service – but got lazy and employed others do the work 🙂

      • Roger Andrews says:

        “The real mystery is what caused the switch to 100,000 saw tooth oscillations?”

        Well, the earth continued on its path around the sun just as it always had, so whatever caused the switch must have been something that happened on Earth. But what?

        My guess is a transition from no ice ages to ice ages.

        How might this explain the switch? Here’s a wild and woolly hypothesis. Before the transition the natural oscillations in the Earth’s climate system would presumably have been determined by water slopping around in the oceans, but after the transition they would have been determined by the growth and decay of ice sheets, which takes a lot longer. This would have resulted in a comparatively short pre-ice “resonance period” followed by a longer ice-age resonance period

        And the pre-ice resonance period coupled preferentially with the 40,000 year Milankovic cycle and the ice-age resonance period coupled preferentially with the 100,000 year Milankovic cycle.

        How might terrestrial cycles couple with Milankovic cycles? In the 17th century Christiaan Huygens showed how two pendulums synchronize themselves when coupled together. Here’s a video showing how it works with metronomes. If you visualize the metronomes as different elements of the earth’s climate system and the cans as Milankovic cycles you should get the idea. And if you don’t the video is still interesting to watch 🙂

  6. A C Osborn says:

    Every 100,000 years the Aliens come back to earth with their Ice ray guns to ensure that the Intelligent Life clock gets set back to zero as they do not want any competition.
    Well it makes as much sense as CO2 molecules controlling the Climate.
    I was interested to see that in the AGU presentations the NASA Scientists could measure the Cooling affect of CO2 in the Thermosphere, but not the supposed warming in the lower atmosphere.

  7. A C Osborn says:

    Euan, you may be interested in this, with regards to Ice Ages.

  8. Roger Andrews says:

    Euan and Clive:

    I’m putting together a blanket response to your comments but it’s taking me a little while to get my stuff together. Please bear with me. Back ASAP. 🙂

  9. Kit P says:

    One of the reasons I am a skeptic is watching how people use science to deal with problems. Geology is a useful tool in identifying problems that could occur. Every nuke plant is built based on the results of geological studies. The nuke plant in the PNW is built on basalt built up by massive lava flows that occur every 600,000 years. The last one was 600,000 years ago. If you only visit one place in America, it should be Yellowstone NP. Better hurry it could go blow any day.
    The nuke plant is not designed for the hazards associated with 600,000 year natural disasters. We are all going to die anyway. Not from radiation but from starvation. I saw the movie.

  10. Kit P says:

    The PNW was mostly under a glacier 20,000 years ago. An ice dam formed Lake Missoula and subsequent periodic failures created some unusual formation that could not be explained by the scientific consensus. When geologist Harlen Bretz published his theory he was ridiculed by the scientific consensus. Of course the collection of over whelming evidence, proved that Bretz was correct.
    One of the reasons I am a skeptical of ‘consensus’ theories of trace amounts of CO2 or PM2.5 is that the theories are relatively new and the scientific community just has not had time to collect the evidence. The evidence collected so far is not very convincing. Disagree and you are labeled a denier or a troll.

    • Euan Mearns says:

      Kit, the general objective of this blog is to accumulate informed opinion from all sides of “the energy debate”. The only rules are that comments have to be evidence based and civil. I have not yet censored or edited any comments and hope to continue in this mode. I understand where you are coming from when you say:

      Disagree and you are labeled a denier or a troll.

      My immediate reaction to those who use these terms is they are either ignorant, weak or have vested interests. I like to argue. And the only way to have a good argument is for those who disagree is to bring it on – in a civil evidence based argument. E

      PS I am totally ignorant on the PM 2.5 debate

  11. Roger Andrews says:

    Back again.

    Euan, you say: “This leaves two questions, 1) what brought on the ice ages and 2) is the 41,000 / 100,000 year cyclicity real?”

    Well, the 41,000 obliquity and 100,000 year eccentricity cycles are real. Here they are, plotted over the last 100,000 years. (The data came from a NASA website that isn’t there any more but they should still be around somewhere).

    The question is, did these cycles bring on the ice ages?

    Here we run into problems. The first, as already noted, is that the irregular timing of the ice ages and interglacials doesn’t match the regular timing of the cycles.

    The second problem is that the Milankovic cycles are smooth and regular while the d18O records show a sawtooth pattern, with erratic and prolonged descents into glacial minima but (comparatively) rapid recoveries back into interglacials. The changes from advance to retreat are very sharp – at least when viewed on time scales of a few thousand years – and to all intents and purposes they can be regarded as “tipping points”.

    The third problem is; how can minuscule changes in Milankovic cycles – and over periods of a few thousand years the changes are really tiny, cause the Earth’s climate to tip over and head off in the opposite direction?

    I don’t think they can. Something else is doing it.

    Which I guess, Clive, represents a response to your question: “Somehow though we have to explain why (Heinrich Events) occur in response to orbital forcing. Why should the threshold only be reached when the ellipticity is at a maximum?” I don’t think this is the way it happens.

    As to how it does happen, consider the two graphs in the link below. The first shows the LR04 record over the last 100,000 years with the Milankovic cycles superimposed. It’s hard to see how eccentricity and obliquity could have had had anything to do with the abrupt recovery from the glacial minimum that began ~17,000 years ago.

    The second shows the LR04 record over the last 100,000 years with Heinrich Events 1 through 6 superimposed (Hemming 2004). The regular spacing of the HEs shows the ice sheet surging and calving about once every 7,000 years, and the coincidence between H6 and the onset of the recovery from the glacial minimum suggests that this is when the ice sheet finally “oversurged” itself and began to retreat.

    In short, it’s all ice sheet dynamics. Or at least mostly ice sheet dynamics. I think.

    • Euan Mearns says:

      Roger, I’m not buying your ice sheet dynamics story. The whole point of this post was to showcase Clive’s work combing eccentricity and obliquity cycles which I think does provide a good match to observations. But accepting that these cycles are too puny to cause ice ages and interglacials I will stick with my current hunch that unknown / never observed changes in the Sun play a role.

      What do you think causes the cyclical change in LR04 prior to 3 million years ago?

      Could you either send me or post a link to the orbit cycle data. Tks E

    • Clive Best says:


      The 41,000 year cycle is clearly present throughout the last 5 million years. You can even see it superimposed on recent glaciations. However, I agree that the last cycle was a saw-tooth form similar also to the one 5 cycles previously and Milankovitch can’t explain it. This sudden collapse from the last glacial maximum to the current interglacial is spectacular and yes – ice sheet dynamics must have played a leading role. The Younger Dryas is supposed to have ended with the collapse of the North American Ice sheets and formation of the great lakes. The global temperature of the Earth probably rose by over 4C and up to 10C at high latitudes. This is far greater than even the most alarming future AGW predictions.

  12. Roger Andrews says:

    A couple more points:

    Euan: I think Munsch is saying basically what I’m saying, that Milankovic cycles don’t correlate well enough with ice ages to demonstrate causation.

    Clive: Thanks for the heads-up on the Science of Doom threads. It’s interesting to note that there are a couple of models that can reportedly replicate some of the features of ice ages using natural forcings and no CO2. Maybe with a little bit of extra effort the modelers could replicate the 20th century temperature record with natural forcings and no CO2 as well. 🙂

  13. Roger Andrews says:

    Eaun and Clive

    Another blanket response to your comments.

    Euan; Clive and I actually aren’t very far apart on this. Clive agrees with me that Milankovic cycles can’t explain the ice age “sawtooth pattern” and that ice sheet dynamics must have played a leading role in the sudden collapse from the last glacial maximum, and I agree with Clive that his LR04 / Milankovic cycle comparisons make it difficult to escape the conclusion that Milankovic cycles play an important role in the long-term. (Although I admit to having no clue as to why the 100,000 year cycle replaced the 40,000 year cycle.)

    The two hypotheses are in fact complementary rather than contradictory. When we combine them we find we have plausible explanations for the ~100,000 year ice age periodicity (Milankovic cycles) and also for the departures from this periodicity and the “sawtooth” pattern of the d18O records (ice sheet dynamics).

    However, none of this answers the question posed at the top of this thread, namely when doth the ice man come next. According to the figures below, which superimpose the EPICA record for the current interglacial on the EPICA records for the last two interglacials, he’s already about 10,000 years overdue:

    Finally, another wild card:

  14. Hugh Sharman says:

    This is an excellent posting that deserves a wide readership.

    My eye was caught by Roger Andrew’s extraordinary but unambiguous assertion that “Interglacials occur only in the Norther Hemisphere, where ice sheet extent ranges from maybe as much as 30 million sq km during glacial maxima down to a small fraction of that during interglacials. They don’t occur in the SH. Antarctica has been ice-bound for millions of years.”

    Hmmm…can this really be true?

    With so many more erudite and expert commentators than I posting under the article, I hesitate to ask such a dumb question.

    But are “ice ages” really just a Northern hemisphere phenomenon? If not, are they synchronous? Evidence? Explanation?

    I’m confused!

    It seems to me that on a planet with up to 10 billion humans, evolving from just a few million at the start of this interglacial, a mere 15,000 years ago, the early arrival of the next ice age, possibly happening very fast, ought to exercise rather more interest than it seems to in this age of “catastrophic” AGW.

    Hugh Sharman

    • Euan Mearns says:

      Hugh, I don’t think interglacials are confined to the northern hemisphere. This has all to do with the distribution of land mass at high latitude. In the S hemisphere, Antarctica is over the pole and has been ice bound for millions of years. But large temperature swings are recorded at Vostok and presumably this may have impacted ice accumulation around the margin of the continent. Otherwise the only other bit of land in the right place is the southern tip of S America where in Chile the S and N Patagonia ice fields are remnants I believe from the last ice age. I’m unsure about the history of growth and decay of this ice sheet.

      Personally I think there is much greater risk of extreme cold events in the coming decades than CAGW and European governments would do well to build a power delivery system fit for that purpose. Energy decline in a freezing climate is a chilling prospect.

      • Willem Post says:

        …which would make distributed wind and solar energy even less viable. It appears nuclear, a la France, using thorium in the future, is a more secure answer.

        • Euan Mearns says:

          I wonder where the power is coming from today in NE USA? I have this image of PV under 2 feet of snow and wind turbines iced up. Under such severe winter conditions I know what I would prefer to rely on – backed up with a wood burner.

          • Willem Post says:


            In Vermont, I have a natural-draft (no electricity), propane stove for space heating in the basement, and another, forced-draft, propane, stove in the kitchen, also for space heating. Also, I have a forced-draft, propane, 95% efficient, furnace for heating the rest of the house.

            In case of power failure in winter, the basement stove is adequate to keep the house above 50F

            In case of power failure in summer, some of the food in the refrigerator may spoil. We may use our two-burner, propane heater for cooking, or eat out.

            Thank you for writing your articles. Even, if some of global warming is caused by mankind, the cycles you mention will be far more dominant regarding climate change.

            Any RE actions mankind takes are short-term, and largely futile, make matters worse, because of the resources required to build out inefficient RE systems, and eventually, the lack of resources due to replacement cycles.

            I wrote 3 articles on the subject. Here is one.

          • Euan Mearns says:

            Willem, thanks for the link – a very interesting read, but don’t have time to read it all right now. It seems we are pretty well on the same page. Not sure I agree entirely with need for “smog” to have a major cooling effect. I see late 20th Century warming as a combination of natural warming cycle overprinted by an anthropogenic component (perhaps). I think natural warming has now shifted to natural cooling, mitigated perhaps by an anthropogenic warming component.

            But I also believe that equating climate to global average temperature is a mistake. Natural changes in the pattern of atmospheric circulation are possibly more important at a local level. The meandering Jet stream we have had in recent years is likely linked to the Sun and can cause the extreme weather events we are seeing – nothing to do with global average temperature.

            You mention 8 warming phases in Holocene. What evidence is that based on? Bond / Neff? Or do you have other lines?

            Best Euan

    • Clive Best says:

      Interglacials occur globally not just in the northern hemisphere. Ice cores taken from Antarctica show the same temperature dependence as those from Greenland. The extent of the northern glaciers were far larger because there was more land area to cover across N. America, Europe and Asia, whereas Antarctica was already covered by ice. The ice just gets thicker over Antarctica and the sea ice spreads. Otherwise ice ages are a global phenomenon and tropical temperatures also dropped by about 4C.

  15. Roger Andrews says:

    The word “interglacial” means “between glaciations”, and I have therefore defined them as periods when there is little or no ice. We get such periods only in the Northern Hemisphere. Antarctica has been permanently glaciated for the last ~30 million years.

    All pretty straightforward. So why does Hugh Sharman find this “extraordinary”?

    Probably because d18O records are routinely cited as evidence that ice ages and interglacials were global events. Temperature-wise they may have been, but ice-wise they certainly weren’t.

    Thompson’s recent work at Quelccaya also raises questions as to what d18O records actually tell us, but I’m not going to get into that unless someone asks.

  16. Roger Andrews says:


    You may already know about this paper (I didn’t until just now) but if not it’s worth a look. The authors replicate the 100,000-year periodicity and the sawtooth patterns of the last four ice ages using a combination of “insolation and internal feedbacks between the climate, the ice sheets and the lithosphere–asthenosphere system.”

    I get the impression that this result may have been predetermined by the input (“atmospheric GCM experiments (were) forced with different insolation values ….. CO2 concentrations and ice-sheet sizes, calculated in advance”) but maybe it wasn’t.

    • clivebest says:


      This looks a very interesting paper. I am surprised this didn’t get wider publicity. What they are suggesting is that as ice builds up until it reaches a critical size and then collapses. They have a model whereby as the N. American glacier grows it eventually extends too far south for it’s own stability and then rapidly collapses during the next warm spell caused by orbital forcing. The largest warm spells due to precession tend to coincide with larger ellipticity. Once melting starts The N.American ice sheet rapidly collapses within a few thousand years ending an ice age. CO2 is not involved at all so long as it is not too low or too high. The Eurasian ice sheet instead is sensitive to 41,000 y cycle.

      Of course it is a model and they will have tuned it so as to reproduce the last glacial cycles. However it still is impressive.

    • Euan Mearns says:

      Roger, thanks for the paper – very interesting, not sure i understand it all. They seem to suggest that N American ice sheet operates on a 100,000 year cycle and the Eurasian ice sheet on 41,000 year cycle. This is interesting in itself since this could impart the temperature structure observed in LR04 and ice cores. But does this mean that Eurasian ice sheet can collapse independently of N American? i.e. periods of ice free in Eurasia with on going glaciation in N America? And is the current phase of ice free on both continents somehow anomalous? Or does that take us back to the position 400,000 years ago? Can’t help observing that 400,000 is divisible by both 100,000 and 41,000.

      Whilst I can see intuitively why isostacy (and eustacy) may play a role in the rhythm, I don’t really understand how this is supposed to work from what is written in the paper. Clive?

    • Roger Andrews says:


      Here’s the best I can do on your questions. Clive may want to add something.

      “does this mean that Eurasian ice sheet can collapse independently of N American? i.e. periods of ice free in Eurasia with on going glaciation in N America?”

      Apparently yes. The Laurentide ice sheet “covered most of northern North America between c. 95,000 and c. 20,000 years before the present day” but over this period in Eurasia there were four distinct glacial advances spaced about 40,000 years apart, apparently with not much ice left between them – see Figs 13 through 16 in

      “Is the current phase of ice free on both continents somehow anomalous?” We may in fact have more ice than is usual for an interglacial. The GRIP2 ice core hole in Greenland hit bedrock at 120,000 years BP, suggesting that Greenland may have been ice-free during the last one.

      “I don’t really understand how (isostasy) is supposed to work from what is written in the paper.” What they’re saying is that their model doesn’t work if isostatic rebound occurs immediately after the ice cap begins to shrink because this would cause the surface of the ice cap to bounce back to a higher elevation and shrink less rapidly. But I don’t know why they should mention this because isostatic rebound is delayed anyway. Hudson Bay and the Baltic are
      still rebounding even though the ice disappeared 15,000 years ago. I must be missing something.

      The more I think about it the more I begin to suspect that the model results were predetermined by the input. Sawtooth-shaped output with a periodicity of ~100,000 years isn’t all that surprising when the model is forced with sawtooth-shaped input with a periodicity or ~100,000 years. But again maybe I’m missing something.

Comments are closed.