Roy Martin: Milankovitch cycles and Ice Age timings

Posted: June 16, 2011 by tallbloke in climate, Solar physics, solar system dynamics

Is this the real connection
between Milankovitch cycles and ice ages timing?
Guest post from Roy Martin

This informal study was prompted by the post entitled: “Ice ages driven by earth orientation changes not CO2” and the subsequent comments. A partial review of the literature shows that, in spite of the very extensive research done in the decades since Milankovitch first proposed the idea, there is still no clear understanding about just how the predicted variations in solar input can actually change the climate. The following is a back to basics attempt to clarify some of the issues, without recourse to sophisticated statistics or modelling, which arrives at a conclusion somewhat different from the common perceptions.

Figure 1.) This shows the composite of the variations in 650 N insolation, Benthic δ18O (location unknown), and global temperature inferred from the Vostock ice core data. Yes, the curves are a bit fuzzy because I got lazy and lifted them from an illustration in Wikipedia, but nicer looking plots would not change the conclusions drawn from them.

When you look objectively at the composite insolation graph, putting aside what we know of the cycles from which it is derived, the dominant feature is an almost regular oscillation of varying magnitude, averaging a 20.6 K yr period over 800 K yr s. The magenta coloured bars over the graph mark the high points of the ~100,000 yr eccentricity cycle. but there is no actual 100,000 yr cycle present. The influence of the 100 K yr cycle has morphed into a modulation of the magnitude of the 20.6 K yr cycle, with alternating high and low phases at about the same interval. There is a repeating pattern in the maxima of the 20.6 K yr modulation, with periods of ~ 400 K yr s, marked with blue arrows. Also, although the 41 K yr obliquity cycle seems to have disappeared, I believe it is responsible for the slightly higher and lower amplitudes of successive 20.6 K yr waves. At this stage a regression analysis on the resultant would derive the component cycles, but that is our own mathematical cleverness, and something the earth’s climate systems cannot do. I therefore contend, and I believe this is most important, we now have to forget what we know about how we built up the composite wave, and only look at the effect on the earth system of the variations in the resultant actual insolation.
The δ18O data has been taken as the proxy for global temperatures, because it seems that the Vostock data may not correctly represent the timing of temperature variations on the global scale. However, the Vostock data has been used as a guide to the interglacial temperatures. The dashed blue line approximates the lower temperatures during the Holocene, then the solid blue line was drawn across the δ18O data near the same level, thus defining the interglacial periods. Over the 800 K yr period there are eleven interglacial peaks, with the intervals arbitrarily numbered here from one to ten. The interglacial peaks are marked with solid red vertical lines, and the termination at the beginning of each preceding warming period (WP) are marked with solid green vertical lines. Each WP is further marked with solid grey bars. Intermediate WP s are marked with dashed green and red vertical lines.
From this we find that:-
1.Every single WP was initiated at a rise in the 20.6 K yr insolation cycle, and endured while the insolation was above the solid black median line. 16 of 16 events.
2.All but one of the subsequent cooling periods (CP) was initiated the next time the insolation fell to or below the median line. 14 of 15 events.
We note that the insolation at 650 N varies up to 20%, or 100 W/m2 from minimum to maximum. This clearly points to the 20.6 K yr insolation cycle as the primary driver of the interglacial WP s. The change in insolation is much higher than the level of forcing attributed to CO2 which is therefore unlikely to have ever been more than a bit player as a cause of warming, certainly if, as recent studies are suggesting, the level of sensitivity is really only a fraction of that derived from models. Many studies have shown that in any case CO2 levels do not rise until after the significant warming is under way.
The only exception to the otherwise very consistent pattern is the start of the ~ -700 K yr CP, which happened near a 20.6 K yr maximum. However, it is noted that a moderate minimum occurred during the WP, after which the rate of warming was much lower. That may also have altered the normally stable WP mode, facilitating the early change to a cooling mode even though insolation was still rather high.
The reversals into CP s are quite abrupt, initiated in all cases by a fall in insolation below the median level, and it would appear that a very stable internal mode must have also been established during cooling periods. The δ18O record shows that in periods 2, 4, 5 and 6 there was a largely uninterrupted fall in temperature, in spite of high phases in one, two or three intermediate 20.6 K yr periods. All of the significant intermediate WP s that occur in cycles 1, 4, 7, 9 and 10 also coincide with high phases in the 20.6 K yr periods.

Figure 2.)shows the lengths of the periods between terminations and the corresponding interglacials

Figure 2.) shows the lengths of the periods between terminations and the corresponding interglacials. This clearly demonstrates the tendency of the periods to occur at near multiples of 20.6 K yr, i.e., listing multiples and occurrences: 1,2; 2,3; 3,0; 4,12; 5,1; 6,2. There is very definitely no 100 K yr period in there. The changing insolation is acting on a complex earth system in which non linear responses occur internally. Thus the direct effect of the original eccentricity and obliquity cycles is completely missing from the record of glaciations and interglacials.
During the last two 400 K yr periods there was an approximate repetition of the pattern of cycle lengths. However, although the pattern may not be completely random, there is definitely no regular long term periodicity. We can expect that this, together with the non linear nature of internal climate variability, would confound any analysis aimed at uncovering 41 K or 100 K components in the outcome. The effects are irreversible.
The observation that the cycles of the last 800 K yr have been timed in some way by the 20.6 K yr periods is not new. It has been referred to directly or by inference in a number of papers. However, to the best of my current knowledge, the consistent close coupling of the WP s to the high phases of the cycles, as established here, has not been previously documented.
I contend that the coupling of the WP s demonstrates most clearly that the Milankovitch cycles do effect the timing of both the long and short term cycles of glaciation and interglacial warming, but not in the way commonly accepted.
Many possible consequences might arise from this hypothesis, but one that should give us cause seriously to think is that, falling from a peak early in the Holocene, the present insolation level is already below that of eight of the ten previous CP beginnings.

Comments
  1. Malaga View says:

    Excellent posting… wonderful analysis and clear thinking.

    Conceptually I can handle very happily that:
    1) When insolation rises above 497 W/m2 at 65N (give or take a bit) the earth is in warming mode.
    2) When insolation drops below 497 W/m2 at 65N (give or take a bit) the earth is in cooling mode.

    Fits very nicely with the analogy of how to control the heating / cooling of a coffee pot…

    The underlying graphics are a bit fuzzy… but I think that is a true reflection of the underlying quality of the data… calculating insolation back over 800,000 years only works if the system is constant… and it is not on that timescale… so clear thinking from first principles is important.

    However, it is possible that the equilibrium insolation varies over time as, in the words of Forest Gump, shit happens…
    Perhaps when the Bering Strait opened up the balance changed… or as sea levels rose / fell… or as the solar system evolved..

    So this time it might be a bit different…
    Our extended interglacial in the Vostok record shows it is longer than normal…
    But that might just a data quality issue…

    Conceptually, I can also handle this very happily because this is where I started from…
    Before my excursion into debunking the flawed logic of the CAGW faith system.

    But I am still very curious and would like to understand the Vostok dust record.

  2. tallbloke says:

    The thing about the precession cycle is that you’re always robbing Peter to pay Paul. As insolation increases at 65N it falls elsewhere. I still think there is a net background influence from eccentricity on around the 100,000 year timescale, but it may be insignificant if my ideas about the way energy gets released from the ocean stand up. Basically, I’m thinking the increasing 65N insolation plus the drop at the equator encouraging the ocean to go into a super-El Nino mode is what does the trick to lift Earth out of glacial conditions into the interglacial epochs. It’s all a question of whether we can get our heads round the possibility of the ocean storing heat on those kind of timescales. I think it can do it, because it took the ocean millions of years to cool down after it warmed all the way to the bottom at an earlier epoch.

  3. Roger Andrews says:

    I counted the warming periods on Figure 1 and noted where they coincided with 65N insolation peaks (without regard to amplitude). First time round I counted 17 warming periods, 14 of which coincided with insolation peaks. If we apply simple coin-toss statistics we find that the chances that this is a random event are about 200 to 1 against.

    Then I counted them again. This time I got 16 warming periods, 12 of which coincided with insolation peaks (probably because I double-counted first time around). Coin-toss statistics show that the chances that this is a random event are about 25 to 1 against.

    I could have counted a third time, but based on these two counts I think we can reasonably conclude that 65N insolation peaks and warming periods are related. (I didn’t do any counts for cooling periods because they are a lot less abrupt and therefore harder to define than the warming periods.)

    The problem, however, is that there are 37 insolation peaks in the record (again counted without regard to amplitude) and only about a third of them coincide with warming periods. Moreover, a number of the insolation peaks that don’t coincide with warming periods are as large as the insolation peaks that do. Why didn’t these peaks cause warming too?

    The most plausible explanation is that while increased insolation provides favorable conditions for warming, it isn’t what triggers it. As to what does, I’m still going with my theory that the abrupt warming that precedes interglacials is triggered by Heinrich events that calve huge amounts of ice from the NH ice sheet over a short period of time, and when a large Heinrich event coincides with an insolation increase at 65N what’s left of the NH ice sheet begins to shrink rapidly. The current interglacial in fact began with a large Heinrich event that coincided with a 65N insolation increase.

    Note that this mechanism – gradual ice sheet growth during the onset of an ice age and abrupt shrinkage at the end – also explains the sawtooth pattern we see in the benthic and ice core 18O records.

  4. Malaga View says:

    The thing about the precession cycle is that you’re always robbing Peter to pay Paul.

    Well that can be an apples and pears robbery….

    If 65N is in cooling mode then the three NH glacial nodes are cold: NE Canada, Greenland and Norway…

    If 65S is in warming mode then we will get offsetting ice loss in Antarctica… perhaps a new Southern Greenland…. and warmer Pacific currents that keep Alaska and Siberia out of the ice box…

    The maps and imagery regarding the last ice are deceiving….
    The ice did not cover the whole of the NH land masses… look carefully…
    The sea ice around Antarctica can only be pure speculation / artistic license.

  5. tallbloke says:

    MV: don’t get me wrong, I appreciate why 65N warming is associated with the move to interglacial rather than 65S due to landmass difference. I still think the whole process needs a kick up the jacksy from tropical heat release though. But I admit it’s just a strong hunch at the moment. I intend to follow it up once I’ve got the software I’m after.

  6. Thierry says:

    Anyone who wants to understand how and why ice ages happen should read Marcel Leroux books. He would have been happy with the article above (he died in 2008). He was one of the greatest climatologist, and was a sceptic. He is the inventor in the 80’s of the Polar Mobile High concept which is the definitive basis of meteorology and climatology finally reunified. According to Marcel Leroux, MPH which are large discs of cold air of high pressure, 1,5 km thick, 1000 to 3000 km wide originating regularly form the poles and going towards the equator following various paths. Depressions are just the vortex created by MPH as they move and leave some empty space behind them, with meridional humid air going up north at the leading edge (in northern hemisphere). In his books, he explains clearly how insolation deficit lead to ice age. It is amazing to see that most meteorologists today still believe that Jet Stream, or hypothetic precursor at high altitude creates depressions. Marcel Leroux explains that it is antiphysics, since very low pressure air at 30000 feet can not have any effect on the dense and humid air at the ground level.

    http://lcre.univ-lyon3.fr/climato/amp.htm

    It is in French, but his books are in English

    Marcel Leroux explains that when it is colder, more powerful MPH are moving south, and consequently, more humid air is arriving in the north as rain of snow. During ice ages, polar insulation deficit, which is even more important at 80°N than at 65°N, explains the increased number of MPH. Since they are thin, they are easily blocked or guided by the reliefs, explaining why they follow certain paths and why huge mountains of ice were located in the northeast of Canada and in Scandinavia. More snow means more difficulty for the sun to melt it all in summer season. One season after another, more ice contributes to increasing number of MPH. You also understand why temperature means are irrelevant, since it can be colder at certain locations, but warmer at others (especially in areas where warm and humid arrives in the north regions). He also explains a lot of other phenomenas with the MPH concept (El Nino, la Nina, monsoons, Sahel drought, Kilimanjaro).

    When I see a moving satellite map now, MPH appears so obvious, so that I do not understand why nobody cares about that.

    Remember that name : Marcel Leroux

  7. tallbloke says:

    Salut Thierry, and thanks for your comment. I see Marcel’s books command high prices on the secondhand market. That’s always a good sign for the quality of the content.

  8. Stephen Wilde says:

    Top down solar effects interacting with bottom up oceanic effects and manifesting the combined net outcome as per the Leroux MHP concept.

    I don’t have the time or mathematical expertise to do the proofs but we are looking at a decent start here.

    Until recently I hadn’t heard about the MPH idea but it certainly fits my thoughts.

    If a quiet sun affects the polar vortices then a more distant sun ( or a different angle of incidence) will do it too and the outcome is a change in global cloudiness and albedo.

    That change in global cloudiness and albedo alters solar shortwave input to the oceans thereby affecting the equilibrium temperature of the oceans and it is the requilibrium temperature of the oceans that matters and not the equilibrium temperature of the atmosphere.

    A substantial amplification of solar changes arises largely unrelated to TSI and arising from chemical rather than radiative processes in the atmosphere.

    Then, eventually, in fits and starts via a combination of ENSO, PDO and the 1000 to 1500 year thermohaline circulation (often involving multiple cycles) the variations in ocean energy content feed back to the troposphere.

    If one gets solar and oceanic phasing acting together to supplement each other’s effects then that is the trigger for abrupt changes whether towards warming or cooling.

    I suspect that warmer and relatively stable interglacials only occur when solar and oceanic cycles are phased so as to largely offset each other’s effects.

    Due to the current land/sea distribution that is only for about 10% of the time hence long glaciations compared to short interglacials.

  9. Tenuc says:

    Thanks, Roy, for a well analysed and interesting post.

    First thoughts are that perhaps we need to consider some of the following basics. These are in no particular order and I’d be grateful if everyone can add more to the list…

    Perhaps better progress could be made by bearing in mind Rogers inspired meme that glacials are the norm for our planet and interglacials are only infrequent, although beneficial, events. We need to look for what causes our normally stable cool climate to heat up rapidly every 100ky or so for a few thousand years, before its gradual drift to more normal climes begins once again.

    We know that orbital factors can alter the amount of thermal energy delivered by our sun, and that the output from sol also changes over time as it is a is a variable star.

    The whole solar system behaves as on integrated whole, with the planets seeming to influence the sun (and vice-versa) in a quasi-cyclical way.

    Arctic is an ocean mostly encircled with land – Antarctic is land encircled by and ocean. They appear to respond in very different ways to interglacial events. Why?

    We only have reasonable data for both orbital and solar output for the last few hundred years – prior to this we only have a poor quality and temporally smeared proxy record of what actually occurred.

    There is a paradox in that the onset of interglacials is rapid, while orbital and solar TSI variability seem to take place over long time periods. What other factors could provide a sudden energy release capable of triggering these sudden short-lived warming events? There are some known unknowns regarding quality of proxy data and orbital calculations. There could be some unknown unknowns, which perhaps some unconventional minds are beginning to grapple with…

    Gravity Energy – John Dodds
    https://tallbloke.wordpress.com/2010/09/22/john-dodds-planetary-gravitational-explanation-for-cyclic-warming-on-earth/

    Unified (Gravity/EM Charge) Field – Miles Mathis
    http://milesmathis.com/uft2.html
    http://milesmathis.com/ice.html

  10. Malaga View says:

    TB: I still think the whole process needs a kick up the jacksy…

    Another Bond Event would give the process a kick up the jacksy
    Whether we like it or not…

    I’d hoped to not last long enough to reach the next Bond Event, however, we have 3 nagging little points:

    1) It’s a 1470 year or so cycle and the last one started about 1470 years ago… take a look at what was happening in about 530 to 540 A.D. It was cold, and dark, and the sun wasn’t very bright… In fact, they called it The Dark Ages.

    2) The sun has gone very very quiet. Not pleasing in the context of #1.

    3) We’ve had a sudden onset of more cold and more snow at the poles with the oceans cooling starting in 2003 (it takes a while to cool a few gigatons of water…)

    http://chiefio.wordpress.com/2009/04/06/bond-event-zero/

    Thierry: Marcel Leroux explains that when it is colder, more powerful MPH are moving south, and consequently, more humid air is arriving in the north as rain of snow.

    That seems to fit the Super Storms….

    The Terracycles evidence of super storms is convincing with just 5 basic observations.

    1) In order for the Earth to rapidly cool, heat must be removed from the Earth at a rapid rate.
    2) The ice sheets only occurred on half of Earth’s northern hemisphere
    3) Glacial sketches made by scientists show ice sheets spread from centralized locations.
    4) The Atlantic Conveyor extended all the way to the Labrador and Norwegian Seas during the time of abrupt climate change.
    5) The Greenland Ice Cores show a brief period of exceedingly high winds

    http://www.terracycles.com/joomla/superstorms.html

    Lots of amazing ideas converging QUICKLY from my perspective….

  11. Tim Channon says:

    Indeed a nice post placing a different view on the table.

    Number of things I want to say.

    a) the Vostok timescale is deeply broken

    b) there is something very strange about the shorter periods such as ~21k which seem chameleon-like, may or may not figure. Say more about that in a moment.

    c) 100k and longer is a highly contentious matter

    d) I’ve broken off looking at the ice data for the time being after a shocking discovery to do with supposed orbitally tuned data, as shown in Nature etc. The data switched periods from how it is in the core, no idea why. Did a quick analysis anyway, total shock when the numbers appeared, seems I had decoded the numbers which had been fed in and many round numbers appeared, human fingers and none are actual orbital periods, merely shorthand for roughly thereabouts.

    From (b) when I worked using the sum of 65N and 65S it was because of a better fit. This would be far better if the timescale was corrected. You will also find the longer 41k is then being used. Spectrally it ought to be in there larger than 21k yet is not so obvious.

    I am thinking about the best way to do my own insolation tuning of the timescales. Some novel approaches have come to mind, especially on how to avoid biasing the real data.

    On (c) I have been trying to discover the origin of the opinion that this impresses as modulation. One work in a series of reports does raise a very novel idea which had not occurred to me. Unfortunately I cannot see how it can do what is claimed, suspecting the author has made a dire mistake, getting time and period mixed up, a very easy thing to do and the reports are not formally reviewed. That author is still around, you might have seen his name, seems a good person with genuine interest and is usefully cross discipline. Since it is as likely the mistake is mine, no name.

    Decoding what he has written, it is more my field, it is about a leaky sampling system. The earth spins once in 24 hours so any point on earth receives an integrating blast of solar irradiation. This is an improper sample, in effect a crude chopper. He is contending this will create longer periods from periodic variation in irradiation, in essence producing the non-linear process necessary for modulation (interaction of two components, a linear system has none). This is probably a variant on aliasing but seem to me to go in the wrong direction.
    I have also tried to model it to see what happens, no useful result.

  12. Malaga View says:

    Tenuc: normally stable cool climate…

    Take another look…
    It might be cold a lot of the time but it is not stable… its a rollercoaster…
    The temperature record has more bumps that I have had hot dinners….
    Lots of warming bumps that just fizzle out…
    The cold is there and explained…
    The warm is there are explained…
    So hang on to your hats… here we go…
    Wwwwwwweeeeeeeeeeeeee 🙂

  13. Thierry says:

    Indeed, the two main books of Marcel Leroux are quite expensive. But they are the discovery of my life, since I found there all the answer I needed for years. You can have some extracts there :

    http://books.google.fr/books?id=w9QVlNzn_asC&lpg=PP1&dq=marcel%20leroux&pg=PP1#v=onepage&q&f=false

    The most important ones are :

    Global warming: myth or reality : the erring ways of climatology
    Dynamic Analysis of Weather and Climate

    When you see an image like this one :

    or

    http://www.sat24.com/world.aspx

    you understand that mobile polar highs (MPHs) are the key actors of meteorology and climate. PDA, NAO, ENSO and so on are just the consequence (according to Leroux) of MPH dynamics and therefore have no effect on the climate. Quite counterintuitive heh ?. For instance, El Nino phenomena appears when the “meteorological equator” (ITCZ), as he calls it, moves a little bit south at west of South America, due to more numerous northerly MPH traveling south through North America. Easterlies originating from Gulf of Mexico crosses Panama into the pacific pushing away warm water south and prevents cold water upwelling along west south America coast. At the same time, strong MPH in Asia pushes more warm water towards the East in the equatorial counter current. As a consequence, El Nino is the consequence of a more MPH active northern hemisphere and a less MPH active than southern hemisphere. ENSO is therefore not an indication of a climate change.

    What does have a signification of the climate change, is the pressure evolution. According to Leroux, atmospheric pressure is the key parameter, not the temperature. Pressure tells you everything of MPH dynamics. Pressure is further more a far less falsifiable indicator than temperature. More pressure in the “anticyclone agglutinations” or AA (that is where all MPH ends up, for instance Azores high, of St Helene high, of Hawaii high), and less pressure at high latitude means that we are in a MPH “rapid mode” telling you that the poles are cooling. According to Leroux, significant rise in pressure in AA as well drop in pressure at high latitude has been observed since the beginning the 70s, showing that we may have entered a cooling phase since then. Again counterintuitive and not what we have been told ! Heat wave or cold wave on continent are caused by the same phenomena : MPH agglutination on continent, prevents, by air stability, cooling under great insolation in summer time, and warming from westerlies in winter times. Therefore, cooling eras mean more heat waves and cold waves due to more MPH originating from the poles and agglutinating on continents. It means also more violent weather, since cold MPH creates more powerful depressions.

    Now you understand why models are really useless : nobody has implemented MPH in models !

    and so on….

    As you see, Leroux has a quite iconoclast approach, based on a long life dedicated to observation and sane physics. His work deserve a very close attention.

  14. Tim Channon says:

    horizon.documentation.ird.fr/exl-doc/pleins_textes/…/010028119.pdf

  15. Malaga View says:

    Thierry: THANK YOU for the introduction to Marcel Leroux.

    There’s No Global Warming, Because There’s No Global Climate

    there is not one “global” climate, but a large variety of climates, depending on latitude, geographic conditions, and atmospheric dynamics.

    One must consider how all the parts of the climate fit together and interact, because the climate evolves steadily, but with diverse manifestations.

    While some parts of the globe have warmed, northeastern Canada and western Greenland (where the North American Ice Age glaciation originated) have been cooling considerably since 1960.

    http://www.21stcenturysciencetech.com/Articles%202005/NoGlobalWarm.pdf

  16. Roy Martin says:

    Thanks to all for the very positive responses. I might have expected that in this company someone would have found a big hole somewhere.
    I deliberately avoided all references to specific causes in the original post, but of course it needs the support of sound explanations. Picking up from on a critical point:

    Stephen Wilde says:
    “If one gets solar and oceanic phasing acting together to supplement each other’s effects then that is the trigger for abrupt changes whether towards warming or cooling.”

    As it reaches the glacial termination the state of the system must reach a point of near instability, in which any number of processes could be involved and interconnected. The breaking up of the edges of the ocean ice sheet extending from the continental glaciers seems to be pretty well established as one of them. I visualize it as “getting all of the ducks in line”. One starts walking, and the rest follow in turn, with a teleconnecting cascade that ends up in a new and stable set of dynamic sub-systems. I am coming to the view that the apparently innocuous rise in insolation just after the minimum of a 20.6 K wave is actually large enough to be effective if all other necessary conditions come into phase together. Solar, oceanic, ice and atmospheric conditions all have to be in the right state for the transition to occur.

  17. Roy Martin says:

    ….Picking up on a critical point.

  18. tallbloke says:

    Roy:
    “the present insolation level is already below that of eight of the ten previous CP beginnings. ”

    But looking on the bright side, nowhere else in the series is a major interglacial followed by such a small dip below the average insolation at 65N and a recovery follows relatvively quickly. Might we escape the decent to full glacial conditions for another 15,000-50,000 years? Or is the cooling ‘already in the pipeline’?

  19. […] Thierry says: June 16, 2011 at 11:04 pm […]

  20. Thierry says:
    June 16, 2011 at 11:04 pm

    “”you understand that mobile polar highs (MPHs) are the key actors of meteorology and climate. PDA, NAO, ENSO and so on are just the consequence (according to Leroux) of MPH dynamics and therefore have no effect on the climate. “”

    These pulses of air mass viewed in the picture in the second link are the physical manifestation of the Lunar declinational Atmospheric tides normal oscillating on their 27.32 day period, if you sort daily/hourly satellite photos into sequences of 27.32 day stacks you will see the repeating of patterns especially on the West side of the Andes, where the surges in the Humboldt current follows the declinational tides forces that drive the meridional flows at depth.

    http://research.aerology.com/aerology-analog-weather-forecasting-method/

  21. Roy Martin says:

    Malaga View said:

    “But I am still very curious and would like to understand the Vostok dust record.”

    A good question. In the Petit Et.Al article in Nature 399, 3June 1999, their Figure 2.) had a very good illustration of the inverse relation between the amount of dust in the cores, and ice volume and the inferred temperature. I cannot recall seeing any papers or articles on the subject since then. The Patagonian deserts were identified as the source of the dust, so if the rapid dispersal of the ash from the recent Chilian volcano is representative, it is probable that the Patagonian dust was distributed around most of the southern hemisphere over quite long periods late in the glaciation cycle. One would expect that to have a significant influence on the climate.

    Does anyone know of recent references?

  22. Ulric Lyons says:

    “…it is probable that the Patagonian dust was distributed around most of the southern hemisphere over quite long periods late in the glaciation cycle. One would expect that to have a significant influence on the climate.”

    Spring dust storms from deserts are more severe after colder winters.

  23. Malaga View says:

    Roy Martin: Patagonian dust was distributed around most of the southern hemisphere…
    Ulric Lyons: Spring dust storms from deserts are more severe after colder winters.
    Thank you… taking another look at the Vostok graphics…

    The dust generally seems to peak when the temperature has dropped by about 8C…
    So in the depths of a glacial period there are some very dry deserts and strong winds….
    then the dust levels drop off as the temperatures begin to rise and the rains return.

    Overall, this paints a far bleaker picture than I had ever imagined…
    Super Storms dumping huge ice sheets over Canada, Greenland and Norway…
    Accompanied by Super Niño and Super Niña effects around the world…

    I don’t think I wanted to know that answer 🙂

  24. Tenuc says:

    @Malaga View “…Overall, this paints a far bleaker picture than I had ever imagined…
    Super Storms dumping huge ice sheets over Canada, Greenland and Norway…
    Accompanied by Super Niño and Super Niña effects around the world…”

    It’s worst than we thought… :-))

    Although it’s pretty bad down south in the Antarctic at the moment – best not to try and survive down there!!!

    Regarding my comment on stability during glacials, it seems that our planet spends about 90% in ‘cold’ mode, but only 10% in ‘warm’ mode – cold climate is more stable.. I put the wiggles and bumps down to climate quasi-cycles, which seem fractal in nature and occur at all temporal and spatial scales.

    Perhaps it’s time to follow the swallows south…

  25. Roger Andrews says:

    “Patagonian dust was distributed around most of the southern hemisphere…”

    When I was working down there in 1993 a lot of Patagonian dust was distributed into my mouth and other bodily orifices. But it wasn’t ordinary dust, it was fine volcanic ash left lying around from the 1991 Cerro Hudson eruption.

    And with the recent Chaiten eruption and the ongoing Puyehue eruption there will be yet more volcanic dust lying around in Patagonia, ready for the wind to pick it up and distribute it somewhere else.

    And when Puyehue quits some other volcano will take over. After all, there are a lot of active volcanoes down there, and they’ve been erupting on and off for a long time.

    So maybe the Vostok dust is related to volcanic eruptions?

  26. Roy Martin says:

    In the Vostok dust record there is also a pronounced spike at ~ -70 Kyr. This is no doubt due to the Toba eruption on the island of Sumatra, which created a crater 100 Km long and up to 35 Km wide.The caldera is now occupied by Lake Toba, and is clearly visible in Google Earth. It took many centuries for the concentration of dust in the ice to return to background levels, There is a corresponding drop in the temperature record over the same period.

  27. Ulric Lyons says:

    Any 20.6 K yr cycle in here ?

  28. NikFromNYC says:

    I regretfully dispose myself to being a character who could give a shit about your blah blah blog.

    But you are on some boy’s blogroll.

    So you are a Firefox tab, tonight.

    I made art about the Green Bubble.

  29. Roger Andrews says:

    Roy; The spike in the Vostok dust record is at 60,000 years BP and Toba erupted around 73,000 years BP. But the ice core age dates could be off.

  30. Tim Channon says:

    Roy,
    I’ve just had a play with the ice data.

    For Vostok there is recent paper which goes into ice squirm, ie. it moved, flowed, messed up. The result moves the timescale pretty close to some of the orbitally timed timescales. There are three sets of movement to a pattern.

    On returning and looking at what you have shown I think you have been misled by the large 100ky component. I think I need to make available the filtered version. It might then become much clearer. It is so bad the ripples move completely out of phase.

    See if I can find the paper again and do something tomorrow.

  31. tallbloke says:

    I’m deliberately staying out of this one at the moment to allow other peoples ideas to develop. Keep it coming.

    Nik, you abrasive, creative New Yorker you. 🙂

    http://www.lyricsmania.com/smell_of_petroleum_lyrics_pogues_the.html

  32. Roy Martin says:

    Tallbloke said,
    June 17, 11:55 am:

    “But looking on the bright side, nowhere else in the series is a major interglacial followed by such a small dip below the average insolation at 65N and a recovery follows relatively quickly. Might we escape the descent to full glacial conditions for another 15,000-50,000 years? Or is the cooling ‘already in the pipeline’?”

    Yes, I pondered that point too, and it is one of the reasons why I put the 400 K yr marker arrows where they are. It emphasises that the pattern of oscillations just before and after the two previous nodes is almost identical, but with the following lows rather lower than in the present low phase. It certainly does open up the possibility that a similar deep cooling period will not ensue from where the system is now. However, my conclusion is that the probabilities are against that, firstly because of the 80% of past occurrences, and secondly because the insolation will stay below my nominal mean line for the next 16 to 18 K yrs. Deep cooling is the more likely scenario, sooner rather than later.

    From a reading of the evidence of temperatures for the whole of the Holocene, my own answer to your final question is a small yes. Measurements from many sources show that while temperatures in different zones have fluctuated madly, most of them actually peaked somewhere between – 10 and – 5 K yrs, consistent with a peak in 65N insolation at ~ – 9.5 K yr. My own guestimate is that the ‘average’ temperature has fallen about 0.5degC since ~- 5 K yrs. Already the beginning? Maybe, but to to echo Malaga View: I still think the whole process needs a kick up the jacksy…

  33. Roy Martin says:

    Roger Andrew said,
    June 18 at 1:21 am

    “Roy; The spike in the Vostok dust record is at 60,000 years BP and Toba erupted around 73,000 years BP. But the ice core age dates could be off.”

    Roger, look carefully: the start of the rise in dust in Vostok is just before – 70 K, right where you would expect it. It took some time to reach maximum concentration.

  34. Roy Martin says:

    Tim Channon said,
    June 18 at 3:01 am:

    “I’ve just had a play with the ice data.
    For Vostok, etc..”

    Tim, in the note I said: “The δ18O data has been taken as the proxy for global temperatures, because it seems that the Vostock data may not correctly represent the timing of temperature variations on the global scale. However, the Vostock data has been used as a guide to the interglacial temperatures.”

    Perhaps a little less ambiguously: “However, while the Vostok record has not been used for timing, because that is possibly questionable, It has been used as a guide to the level of interglacial temperatures.” Temperature changes during the cooling periods are another matter again.

    In the interpretation of Figure 2.), I also said: “There is very definitely no 100 K yr period in there.” An actual 100,000 year climate cycle has always been a phantom, and a myth.

  35. Roger Andrews says:

    Roy:

    All the Vostok data I’ve looked at show dust beginning to increase around 52,000 BP, peaking around 60,000-61,000 BP and returning to background levels around 64,000 BP. See

    http://eesc.columbia.edu/courses/ees/climate/labs/vostok/

    and

    http://iridl.ldeo.columbia.edu/SOURCES/.ICE/.CORE/.VOSTOK/.dstflx/T+exch+table-+text+text+skipanyNaN+-table+.html

    What data are you looking at?

  36. Tim Channon says:

    “In the interpretation of Figure 2.), I also said: “There is very definitely no 100 K yr period in there.” An actual 100,000 year climate cycle has always been a phantom, and a myth.”

    Spectra, non-linear scales, log amplitude.
    For full Berger91 5e6 years. Data violates Shannon, probably why it is so noisy (insufficient bits retained in the published data, truncation noise, undithered data).

    I could pick exact items out of that..

    Doss the 100k/400k seem related to the ice data? Yes. the unexplained is the amplitude.

  37. Tim Channon says:

    Does this explain the spikes? Nope.

    However, if as some have suggested there is a modulation mechanism, it might. The ice data seems to have factors which make sense in that context, odd harmonics, a train of which is how a square or similar wave manifests.

    I don’t know what is going on.

  38. Malaga View says:

    Ulric Lyons: Any 20.6 K yr cycle in here?

    Interesting….
    No obvious 20.6K time signature in the temperature record for Greenland.
    No obvious 20.6K time signature in the accumulation record either

    Also this NOAA data looks different from the graph you have referenced..
    ftp://ftp.ncdc.noaa.gov/pub/data/paleo/icecore/greenland/summit/gisp2/isotopes/gisp2_temp_accum_alley2000.txt
    In the depth of glaciation average temperatures range between -35C and -53C…
    Very different… but still no time signature…
    When average temperatures are below -40 C then snow accumulates at less than 10 cm/year.
    When average temperatures are about -30 C then snow accumulates at more than 20 cm/year.

    Unfortunately, both sets of data do not cover entry into the glacial period.

    So my guesses:
    1) There is rapid accumulation in Greenland as we enter a glacial period.
    2) Once Greenland is truly buried and surrounded by ice then the temperature dynamics change.
    3) Insolation only becomes important again in Greenland when the retreating ice front returns.

    But they are only guesses…

  39. Roy Martin says:

    Roger Andrews said:
    June 18, 2011 at 2:42 pm

    “All the Vostok data I’ve looked at…
    What data are you looking at?”

    Roger, I have commented on this in some detail, because it seems to illustrate the critical importance of good data.

    vostok dust
    From Petit et al, 1999, Figure 2.)

    The data is from the article by J. R. Petit et al: “Climate and atmospheric history of the past 420,000 years from the Vostok ice core, Antarctica”, Nature Vol 399, 3 June 1999. That was written by the original research team. It shows the start of the dust spike at -71 to -72 K yr. That it was due to the Toba eruption seems certain for the following reasons:

    The Toba eruption is variously dated from geological sources at -73 K +/- 3, -74 K +/- 4, and most recently 71.5 K +/- 3, which all bracket the Vostok date. One caveat on the Vostok data in the 1999 article was that the temporal accuracy in the <110 K yr period was only claimed to be “better than +/- 5 K yr”. I have used an update by the authors of the Deuterium/Temperature dataset, published in 2001 by the WDC, which shows exactly the same timings. My impression from a lot of playing around with the data is that the +/- 5 K yr range is very generous, and it is probably less than +/- 3 K yr. (For older periods the possible error range increases considerably.)

    As I read the Columbia data and plot you referenced, the rapid rise in dust flux occurred at ~ -67 K yr, the peak at ~ -60 K yr, and return to background at ~ -52 K yr. This compares with 72 K yr, 67 K yr and 57 K yr respectively from Vostok. It appears that they have plotted data derived from a 5 K yr trailing moving average, thus shifting the whole plot that much later in time. The data itself was obviously pre processed, because the intervals are at a regular 364.72 year spacing, whereas the original data point spacing from the cores, during that time interval, is between eighty and ninety five years. Also, the age is quoted to an accuracy of 0.01 yr, and the flux to four or five decimal places, misleading to say the least. Altogether a bit naughty when they are teaching future scientists.

    Going further, it has seemed improbable that the Toba eruption alone, as enormous as it was, could have caused enough dust to persist in the atmosphere for about fifteen thousand years. However, records from elsewhere show that by ~ – 70 K yr temperatures around the world had fallen so much from the interglacial peak that widespread deserts had already formed. Temperatures following Toba rapidly fell further, creating a Black Swan event that probably caused the addition of much more dust into the atmosphere, thus considerably extending the dust and cold excursion.

    So why did the Toba dust and cold excursion end? In the context of my post, it is noted that it occurred during a low phase of the 20.6 K yr cycle, and ended when the insolation again rose above the nominal median line. The pattern of small highs and lows in the last glacial period is roughly in line with the continuing 20.6 K yr cycle.

  40. Roy Martin says:

    Ulric Lyons said:
    June 17, 2011 at 11:48 pm

    “Any 20.6 K yr cycle in here ?”

    At first sight one would have to say no, however, for the hell of it I have sketched in about where the high and low points of the 20.6 K yr insolation cycle would fall. Not much joy there at all.roy-100k greenland

    But I am not really moved by that. The relative growth in ice coverage in the NH was much greater than in the Antarctic which seems to have created quite a different response to insloation between the two regions. For instance, while the maximum temperature anomaly in Antarctia was in the order of -8degC, Greenland regularly went below -20degC for about 50 K yrs. Antarctica temperatures appear to have remained somewhat responsive to ocean temperatures, while vast areas over the NH ice sheets were hardly affected at all. Coupled with higher albedo at lower lattitudes, a lack of NH response to the insolation changes is easy to understand.

    Until the oceananic extensions of the ice sheets started to break down…

  41. Tenuc says:

    I remembered I had an old paleo-sea level map of the last 140ky and thought I’d see how 20ky(ish) cycle correlated.

    Don’t know why the most recent cycle was so badly off, but the rest seem to hit some sort of peak. Can’t expect exact matches as I suspect the data is crap.

    Interesting that 20ky(ish) is supposed to be the length of the precession of the equinoxes – perhaps a galactic alignment also has a part to play??? Also strange that in 2020 we are due to start The Age of Aquarius, as the long-term alignment cycle of the planets Saturn and Jupiter will have a conjunction at the very beginning of the sign on that date, Perhaps Aquarius signals the start of the slow slide into glaciation… 🙂

  42. tallbloke says:

    Tenuc, don’t say that! I’m an Aquarian, and I can tell you my vocation is to turn up with pitchers of beer, not buckets of ice cubes. 🙂

  43. Tenuc says:

    Oops! Forgot the URL for the chart…


    sealevel-100k

  44. Tenuc says:

    tallbloke says:
    June 20, 2011 at 2:59 pm
    “Tenuc, don’t say that! I’m an Aquarian, and I can tell you my vocation is to turn up with pitchers of beer, not buckets of ice cubes. 🙂

    If it is ice cube, I’ll bring the scotch… 😉

  45. tallbloke says:

    Hmmm, interesting graph. Glaciation so thick it withstood increased insolation at -20kyears? Or eccentricity cycle sufficiently advanced that the precession cycle didn’t have a strong effect? But then if that’s so, it’s something other than the precession effect which lifts us out of glacial to the current interglacial. My mega El Nino effect?

    During the previous epoch when insolation cycles were low amplitude from -400k to -330k, the glacial was short, but deep. Hmmm.

  46. Roger Andrews says:

    TB

    “Glaciation so thick it withstood increased insolation at -20kyears?”

    I counted 37 instances of increased 65N insolation (without regard to the amplitude of the increase) in the graph at the beginning of this thread, and glaciation withstood the increase in about two-thirds of these cases.

    Why? My theory is that the NH ice sheets don’t begin to retreat until an insolation increase coincides with a Heinrich event, which is what happened at the beginning of the current interglacial. But no one seems to want to bite on this, so now I offer the following alternative hypothesis (from Nir Shaviv):

    http://www.sciencebits.com/ice-ages

  47. Roger Andrews says:

    Roy Martin:

    Thanks for your comprehensive reply on the Vostok dust-Toba connection.

    I’m struck by your comment: “It appears that they have plotted data derived from a 5 K yr trailing moving average, thus shifting the whole plot that much later in time.” If this is indeed what they’ve done then I’m not sure the word “naughty” goes far enough.

    If we are at liberty to shift Vostok dates back by 5,000 years then we can certainly relate the Vostok dust to the Toba eruption. (And if the dust didn’t come from Toba, where did it come from?)

    Another question is why the dust didn’t cause cooling. The closest association in the Vostok record is a temperature peak that lags the dust peak by about 2 k years. Or maybe the 18O data have been shifted too?

    I think the basic question you have raised – and it’s come up before – is whether the dates in the Vostok and all the other ice core records that we base so many of our ice-age analyses on are reliable. I know Tim Channon thinks they aren’t (are you there, Tim?)

    Another minor point. Some people plot ice core records with zero years at the right end of the X-axis and others with zero years at the left. This can get confusing. We need a standardized presentation mode.

  48. Stephen Richards says:

    Thierry

    Les scientifiques français sont les meilleurs du monde, je crois. The french have an independence in their DNA that’s why Courtillot and Allegre have been fighting against AGW. Everyone I know here in france say that AGW is a connerie, un blague. My neighbours, all farmers and gardeners (in the french sense not the english), say it’s the cycles! They just know.

    Tallbloke, love the matchless. One beat evey mile. Which one is it, 350 or 500.?

  49. tallbloke says:

    STephen, welcome.

    J’aime la France! Ancestors from both sides of my family are from there. They are free thinkers.It is indeed the cycles…

    It’s a 1949 G80S 500cc with the twin saddle option. 🙂

  50. Tim Channon says:

    “Roger Andrews says:
    June 20, 2011 at 5:18 pm

    [ice core questionable]”

    Tempted to say severe things and give examples. Better not.

    I completely agree a supposed volcanic event should not smear in time but it is not alone, there is instance after instance where conflicts occur, In this case there is dust; and the fingerprint data shows it came from where exactly?

    You are correct to point out plot inconsistency

    I deliberately plot conventionally left to right, top to bottom unless there is an overriding reason why not. Reason is very long engineering experience where many childish professionals are lazy and sloppy. They never think of why they are working, is not for their personal consumption. Such as giving a drawing to someone in a far flung country where their native language is different, now jump things in random directions.

    In this instance the paleo community have decided to use their own private standard. This poses a problem if the intended audience is not in that community. The practice of time left to right for public works has plenty of paleo examples (which I could show).

    This is also about the usage or not of signed dates. 220ky ago or -220k

    That is my opinion. Discuss and disagree freely. Doubt there is an easy answer. I could switch. With plotting it is only a matter of say reverse. Nah, got it, lets do bottom to top. 😉

    Reversing a timeseries can be trivially done using a spreadsheet. Can show how. Most plot software has a reverse switch or flip the time data sign.

  51. Tim Channon says:

    News: Had a look at the Devil’s Hole cave chronology, an infamous dataset which is supposed to contradict orbital ice age explanation.

    r2=0.42, sorry, there is a signature and at the same or slightly longer insolation lead as for Vostok. (my time math is known not quite right, was quick and dirty)
    Once again this is for the sum of 65N and 65S giving the circa 40ky wiggle. As with the Vostok data I exclude the long period. Cheating yes on r2 value but showing a clear linkage.

    The data I found is pretty awful. Apparently the dating which is supposed to be so accurate is only done for a few points and the rest interpolated.

    Rather obviously the cave is far from the north or south pole.

    There seems to be little supporting information.

    From a spectra point of view the data makes little sense. There is a lot of long period mostly a little below 100ky. it does contain orbitals, strange amplitudes.

    Amusingly, on modelling, the dataset ceases 60ky ago so I let it extrapolate. Ho ho, adds the dip and today, suggesting a valid long term characteristic.

    The ice ages are supposed to be for *both* poles, is not one at a time. The tropics etc. continue if a little cooler.

  52. Roy Martin says:

    Tenuc,

    Re your sea level chart. If you were to plot actual 20.6 K yr periods you would find that four of the high points would line up, but of course that would be quite odd, because we would be expecting the lines to be marking minima in the sea level, not maxima. Only the 20.6 K yr low is where it would be expected.

    IMHO, it is usually best to assume that the data is correct more or less correct, and look for a possible explanation on that basis. I have not really examined this issue closely before, but as a first thought: following the interglacial maximum we might be seeing a lagged response between insolation and ice volume change, of up to 10,000 years initially, reducing until they came into phase during the approach into the state at the LGM, causing ice volume to increase rapidly and sea level to fall to the stadial minimum. An ocean heat storage issue?

  53. Roy Martin says:

    To Tim Channon:

    As an old engineer, I agree whole heartedly with everything you have said about data plotting conventions, and the odd approach taken by paleantologists. It can lead to mistakes and unnecessary confusion.

  54. Roy Martin says:

    @ 9:06 am …data is more or less correct.

  55. Tenuc says:

    Roy Martin says:
    June 21, 2011 at 9:06 am
    “Re your sea level chart. If you were to plot actual 20.6 K yr periods you would find that four of the high points would line up, but of course that would be quite odd, because we would be expecting the lines to be marking minima in the sea level, not maxima. Only the 20.6 K yr low is where it would be expected…”

    Hi Roy. The problem is that we know the difficulty we have measuring sea level accurately even using the wealth of satellite data. I therefore expect the quality of paleo-sea level data to be far worse that is often pretended. At best we should hope to see a vague correlation with repeated events at long time scales; at worst even these correlations could be positively misleading. We don’t even know where the error bars should be!

    Same applies I expect to temp, insolation, orbital calcs etc. The reality is that it’s like looking through a thick fog at a landscape which nobody has previously trodden. Even the very structure of our planet (solar system?) has changed considerable at paleo time-scales and the further back we go the errors increase exponentially.

    I don’t think this means we should give up trying, rather all theories of past events need to recognise the high degree of uncertainty about observation and data. Any solutions requiring precision to elucidate an answer (e.g. like tiny amounts of CO2 warming) must always be viewed with a very high degree of scepticism.

  56. Roy Martin says:

    Roger Andrews said:
    June 20, 2011 at 4:25 pm

    “My theory is that the NH ice sheets don’t begin to retreat until an insolation increase coincides with a Heinrich event, which is what happened at the beginning of the current interglacial. But no one seems to want to bite on this,….”

    I agree, as I said in previous comments:-

    June 17, 2011 at 9:11 am: “As it reaches the glacial termination the state of the system must reach a point of near instability, in which any number of processes could be involved and interconnected. The breaking up of the edges of the ocean ice sheet extending from the continental glaciers seems to be pretty well established as one of them.”
    “I am coming to the view that the apparently innocuous rise in insolation just after the minimum of a 20.6 K wave is actually large enough to be effective if all other necessary conditions come into phase together. Solar, oceanic, ice and atmospheric conditions all have to be in the right state for the transition to occur.”

    June 20, 2011 at 8:54 am: In comment on Ulric Lyons entry on Greenland temperatures approaching the LGM: “Antarctica temperatures appear to have remained somewhat responsive to ocean temperatures, while vast areas over the NH ice sheets were hardly affected at all. Coupled with higher albedo at lower lattitudes, a lack of NH response to the insolation changes is easy to understand.
    Until the oceanic extensions of the ice sheets started to break down…”

  57. Roy Martin says:

    Hi Tenuc, Uncertainties about error ranges are always a potential hazard when you are trying to discover correlations. I would suggest that with recent data for sea levels good statistical methods make it possible to constrain the error bands quite well, regardless of the method of measurement, so it is possible to know how meaningful conclusions based on them will be. The paleo-data for sea level is measured differently, both temporally and spacially, so as you say, we do need to be realistic about what we can expect to learn when we look for correlations. I had a look back at several sources today, and it is obvious that there are significant issues with all of them. However, I think I take a more optimistic view of where that leaves us than you do. To be sure, possibly large errors can lead to only vague understandings of what really happened, if the data is looked at in isolation. One way through is to look sideways at all of the events that happened within the same time frame that could have been involved in any way, and take into account all of their interactions in the framework of known mechanics and physics. While a basic set of direct measurements may contain broad uncertainties, I suggest that constraints set by known concurrent events should substantially reduce the limits within which a given event could have occurred, thus greatly enhancing the possibility of drawing meaningful and justifiable conclusions from possibly questionable data. It is like having a jigsaw puzzle with very indistinct images on the pieces; at first you cannot see the picture, but in the end there can be only one way they will physically fit together.

  58. tallbloke says:

    Roy, I think that’s right, and offers a way to tighten constraints and improve timelines on proxies with temporal issues. If enough of these ‘data points’ can be established, and a pattern to errors established, for example temperature increase ‘compressing’ time in ice cores, then general filtering and processing rules can be established which can be used to improve datasets.

    The problem we have here is that we are uncertain as to the comparative strengths of various cycles and the lags between them and their associated end effects. Going back to the lineup of the precession cycles with Tenuc’s sea level, wouldn’t it be when there is less insolation at 65N and more at the equator that sea level will fall the most, due to extra snow accretion on the icepack?

  59. Tenuc says:

    Roy Martin says:
    June 21, 2011 at 12:59 pm
    “…To be sure, possibly large errors can lead to only vague understandings of what really happened, if the data is looked at in isolation. One way through is to look sideways at all of the events that happened within the same time frame that could have been involved in any way, and take into account all of their interactions in the framework of known mechanics and physics…”

    I agree that the more information we can get on temporally related events the better chance we have of seeing the correct picture. Unfortunately science today has become very narrow/specialised and often contradictions or conformations are swept under the carpet or simply go unrecognised. I think this is one of the reasons why it’s often ‘non-professional’ scientists who discover the links, rather than those professional scientists with a vested interest in supporting the status quo.

  60. Tenuc says:

    tallbloke says:
    June 21, 2011 at 1:33 pm
    “…Going back to the lineup of the precession cycles with Tenuc’s sea level, wouldn’t it be when there is less insolation at 65N and more at the equator that sea level will fall the most, due to extra snow accretion on the icepack?”

    I like your thinking, Rog…

    Glaciers / sea ice and snow-pack are always in a state of dynamic change and only when we have a situation where there is more ablation than accretion for long time periods do we end up with warming events. It requires the ingress of moist warm air particulates to produce snow, with the amount deposited depending on the efficiency of the transport system (wind and ocean currents) from equator to poles.

    Warm and wet and high particulates – amount of ice cover increases.
    Dry and cold (particulates insignificant) – cover decreases.

    After every 4 or so 20ky(ish) precession quasi-cycles conditions during the 5th quasi-cycle are such that ablation outweighs growth and we move back to inter-glacial conditions again. This warm period lasts for about 10ky(ish) before the slow decent back to normal glacial condition begins one again.

  61. Roger Andrews says:

    Roy Martin:

    “As it reaches the glacial termination the state of the system must reach a point of near instability, in which any number of processes could be involved and interconnected. The breaking up of the edges of the ocean ice sheet extending from the continental glaciers seems to be pretty well established as one of them. I am coming to the view that the apparently innocuous rise in insolation just after the minimum of a 20.6 K wave is actually large enough to be effective if all other necessary conditions come into phase together. Solar, oceanic, ice and atmospheric conditions all have to be in the right state for the transition to occur.”

    I missed that. But you said it better than I did. We need such a mechanism to explain why glacial events often don’t coincide with Milankovitch cycles and why the transitions to interglacials are so abrupt. However, it doesn’t imply that Milankovitch cycles are unimportant.

    Tenuc:

    “Glaciers / sea ice and snow-pack are always in a state of dynamic change and only when we have a situation where there is more ablation than accretion for long time periods do we end up with warming events.”

    I think you’re saying more or less the same thing?

  62. Roy Martin says:

    On June 21 @ 9:06 I said:
    “following the interglacial maximum we might be seeing a lagged response between insolation and ice volume change, of up to 10,000 years initially, reducing until they came into phase during the approach into the state at the LGM, causing ice volume to increase rapidly and sea level to fall to the stadial minimum. An ocean heat storage issue?”

    I went through a lot of data today, from papers already in file plus a few new ones, ending up convinced that my hunch was right. All of the data acknowledges a significant time lag in ice volumes and, by the inverse, sea levels in relation to inferred temperature and O18. On the chart put up by Tenuc on June 20 @ 3:01, if the sea level curve is shifted back ~10K Yr in time, it can easily be seen that it closely follows the 20.6 K yr insolation curve. Going back to the Gerard Roe 2006 paper; by taking the derivative dV/dt of the ice volume he in effect shifted the ice volume curve backwards in time by ~5.2 K yr arriving at the almost exact match with the 20.6 K yr period insolation curve. The two results differ by nearly 5 K yr, but I suggest are within the temporal limits of the rather loosely timed ice volume and sea level data sets, this gives us a good indication that the ice volume and hence sea level changes really are in lagged time with, and presumably being driven by, the insolation changes defined by the Milankovitch. Incidentally, I have a feeling that Roe’s interpretation of the dv/dt correlation may be a minor red herring. So what does happen?

    I suggest that the answer probably lies with ocean heat response times. This is the hypothesis: Additional heat is absorbed by the middle and lower latitude oceans during the high phase of the insolation cycle, reaching a maximum as the insolation falls back to about the median line. It has been pointed out that while the change in heat input is greatest at higher latitudes the variation in middle and lower latitudes is still quite significant. Heat transfer towards the poles takes some time, so ice melt does not begin until well after the heat input starts to rise from a periodic minimum, and maximum deglaciation and highest sea level are not reached until well after the maximum solar heat input peak. The reverse happens during during cooling, with the result that maximum glaciation and lowest sea level is not reached until well after the insolation has reached the lowest point and ocean heat content has already begun to rise. Thus we end up with an oscillating system between ocean heat content and ice volume, driven all the way by the 20.6 K yr period Milankovitch cycle. Changing levels of heat acting directly on the ice sheets, and albedo effects, have been considered as probably the dominant factors, but I suggest that changing ocean heat content is more likely to be consistent with the existence of the observed time lag,

    So far so good, by why does this interglacial sequence end so abruptly? The answer appears to result from that fact that all terminations are preceded by two quite low amplitude cycles, thus with lower heat imput for an extended period, and it is also observed that every maximum glaciation and lowest sea level point is exactly in phase with the next minimum in the insolation cycle. There is in these termination cycles no lag between minimum heat input and maximum glaciation. I am not sure exactly what to suggest happens next.

    The evidence that the cycles of glaciation are driven by the Milankovitch cycles is quite compelling, but again I emphasize that the 20.6 K yr cycle is the primary determinant of all timings, and that there is no sign of direct 41 K yr or 100 K yr forcings in the actual record.