Earth’s obliquity and temperature over the last 20,000 years

Recently, there was a timeline comic from XKCD about the history of Earth’s temperature over the Holocene which got quite a bit of play, due to the fact that it had the inevitable “hockey stick” splice onto the end. Josh came up with a much less dogmatic and more detailed version, which we covered here.

That led Javier, who is a PhD level scientist in molecular biology and biochemistry who recently published an essay at Dr. Judith Curry’s website here, to come up with a similar but even more technical cartoon. We may have more on his essay later, but for right now I present his excellent timeline cartoon, that includes Earth’s obliquity and temperature over the last 20,000 years. Changes in Earth’s obliquity and thus solar insolation angles is believed to be the main driver of long term climate change, initiating and ending ice ages. As seen in the diagram below, the angle between Earth’s rotational axis and the normal to the plane of its orbit (obliquity) oscillates between 22.1 and 24.5 degrees on a 41,000-year cycle. It is currently 23.44 degrees and decreasing at a rate of about 0.013° per century.

22.1–24.5° range of Earth's obliquity Image: Wikimedia. CC2.0

22.1–24.5° range of Earth’s obliquity Image: Wikimedia. CC2.0

An often-cited 1980 study by Imbrie determined that,

“Ignoring anthropogenic and other possible sources of variation acting at frequencies higher than one cycle per 19,000 years, this model predicts that the long-term cooling trend that began some 6,000 years ago will continue for the next 23,000 years”

With that in mind, have a look at Javier’s cartoon:


NOTE: an updated version of this cartoon timeline was posted on Oct 1, 2016 to fix a couple of typos in text labels pointed out by commenters.


123 thoughts on “Earth’s obliquity and temperature over the last 20,000 years

  1. Hmmm
    so axis tilt has insignificant effect on temperature
    so quiet sun tilt has insignificant effect on temperature

    • Your interpretation is way off the mark.

      However, the author does acknowledge that we cannot know the future because we don’t know what disrupting events will take place. To express that view in Latin: Ignoramus.

  2. Memo to all planners who want to anticipate future climate impacts. What are your contingency plans for the end of the inter-glacial?

    • Surely you mean “You have to ignore a lot of History to claim that today’s Climate status is somehow unusual.”

      But gee, all that history is kinda boring, isn’t? What do I need to know all that stuff for?

      • Martin A says: September 30, 2016 at 12:11 am

        History is just one damn thing after another.

        History is a lot like climate science. There are patterns but the details are unpredictable because history and the climate are both chaotic. On the one hand you have the arrogant buffoons who claim to understand it all and pretend that they can make valid predictions. On the other hand there are those who are more humble. link

  3. Beautiful cartoon. Had missed it over at CE. It actually shows Earths millennial thermal inertia well.
    For the first half of the Pleisocene Ice Ages (~2 million BCE to ~ 1 million BCE, the cycle was 41000 years. Obliquity. Speculation is that onset was caused by closure of the Panama isthmus and changes in ocean circulation.
    But Then it transitioned to ~120000 years, ‘two obliquities’ . The ‘two obliquities’ says something else is now going on. My personal current speculation is ocean currents. At the LGM, sea level was ~120 meters lower. That means places like the Sundra Straight were dry land, and ocean circulations had to be very different. Something other than obliquity alone has to explain the glacial/interglacial shift from one obliquity cycle to two.

    • Not just the ocean circulation. smaller oceans. less depth.
      Christ on a crutch. how to model such world. Do the ice sheets suppress volcanoes? What of the oceans, with less water pressure are volcanoes more active? As ice sheets develop do you get crustal depression and what are the seismic effects? With smaller oceans less depth is there a change in the wave energy, How are tides effected. I could go on, but ahem, I am out of my depth


      • One missing ingredient is altitude. Glaciers form at a a combination of high Northern or Southern latitude plus altitude. During major glaciations the ice sheets grow latitudinally as they create their own local temp reduction. They also grow higher, with the Northern ice sheets achieving 2 kms. thickness. This greatly reduces the temp at the top due to lapse rate effect. This makes the glaciation self reinforcing and extends ice ages well into periods which are, technically, warming periods in terms of orbital characteristics.

    • “But Then it transitioned to ~120000 years, ‘two obliquities’. The ‘two obliquities’ says something else is now going on. … Something other than obliquity alone has to explain the glacial/interglacial shift from one obliquity cycle to two.”


      There is no need to invoke any special mechanism because in reality there was no Mid-Pleistocene Transition. What happened was caused by the progressive cooling of the planet as the oceans were becoming cooler over hundreds of thousands of years of Quaternary Ice Age.

      This graph doesn’t have time scale, but it is 5 million years (200,000 per tick). It nicely shows the cooling of the planet.

      It reached a point in which obliquity alone was not capable of getting the planet out of glaciation and it needed a set of favourable conditions that include high precession factor and something else that adds a very fast positive feedback to ice melting. Some say albedo, others say dust, I believe it is low sea levels coupled with ice sheets extending to those low levels, so when the melting starts rising oceans are extremely effective at melting a lot of ice very quickly.

      You can see that in this other graph were the red line shows the interglacial level and the blue line shows the failed interglacial level. The blue line is dropping as the planet cools:

      There is no 100-kyr cycle. Interglacials are sometimes skipping one obliquity cycle, and since Eemian two were skipped (123 kyr). The 100 kyr cycle is an artifact of the irregular distribution of interglacials. If you count from one to the next you almost never count 100 kyr.

      The planet has become so incredibly cold that if it becomes any colder we run the risk that obliquity is not able to reach interglacial temperatures even with those additional factors and the planet gets locked in glacial conditions until the Ice Age starts to pass several millions of years in the future.

      Our only hope is that the cooling has bottomed as the first graph suggests. The oceans are already at 3.9°C and it is possible that they won’t get much colder. But in any case do not count on an interglacial coming in 30,000 years because there won’t be one. When the glaciation comes it will be for at least 70,000 years. That’s about 2800 human generations to ponder why we were so worried about global warming.

      • Javier, thanks for responding. I hear your, but am not convinced. Obliquity I get. Two obliquities I dont. Why two instead of 1.5 or 2.5 or whatever. I hear you loud and clear. But remain skeptical.
        FWIW, your graphic is better than SKCD and Josh put together.

      • I’ll write an article on that, expanding my arguments and showing further evidence, although I might not be able to convince you. This is a very divisive issue between experts. They almost got to a real fight over the Devil’s cave data.

      • The supposed Milankovitch solar variations should speak to the amplitude as well as the timing. If you will argue that the timing is similar in the Pliocene, (by no means certain) why is the amplitude so much smaller? One can imagine the power of obliquity creating equal amplitude variations during warmer periods. Clearly that did not happen.

      • gymnosperm,

        “The supposed Milankovitch solar variations should speak to the amplitude as well as the timing.”

        Not necessarily if there are feedbacks whose strength depends on temperature. Dansgaard-Oeschger evens are a good example of very abrupt wide changes in temperature that take place only during the very cold stage. Once the planet gets on an interglacial they don’t take place even though the same mechanisms are probably still taking place. They just don’t produce the same effect because the conditions are different. In the same manner planetary conditions are not the same now that during the early Pleistocene, so the response to the same input has different amplitude.

      • Javier

        Just a general comment about your contributions of late. I’ve learned much and enjoyed your posts. English might not be your first language but your communications are coming through loud and clear. That is what counts. Thanks for your efforts.

      • “Once the planet gets on an interglacial they don’t take place even though the same mechanisms are probably still taking place.”

        That’s like trying to open a door when it’s already open.

      • cerescokid says, “Javier, … I’ve learned much and enjoyed your posts. English might not be your first language but your communications are coming through loud and clear. That is what counts. Thanks for your efforts.”

        I’ll take the compliment of Javier’s English skills a couple steps further. First, he makes very few mistakes most native English speakers make, no doubt in part for not having had the decided disadvantage of attending American (or British, I imagine) K-12. Second, were it not for a couple tiny, telltale habits, I might not suspect a non-native speaker at all. For example:

        Because the word “entre” serves in Spanish (as well as French and Portuguese, FWIW) as both “between” AND “among,” Javier commits the common mistake of using “between” when “among” should be chosen. (e.g., “This is a very divisive issue BETWEEN experts.”)

        Also, because the Spanish (and French and Portugese) word “que” is most commonly translated as “that” or “than,” but is used to mean “as” in some comparative phrases – e.g., “igual QUE yo” in Spanish means “same AS me” – he says things like “In the same manner planetary conditions are not the same now THAT during the early Pleistocene…” instead of “same now AS during the early Pleistocene.”

        If we wanted to get pathetically pedantic, we could flag the use of “to” instead of “into,” as in, “They almost got TO a real fight over the Devil’s cave data.” Even so, compare these relative nothings to the garbled grammar and punctuation we see so often on these pages and across the entire Web, by native English speakers, and Javier’s English skills come into much better perspective.

        All of this is just a long-winded way of saying that not only is Javier’s writing content superb, his mastery of English is, as well, IMNVHO ;)

        Javier: te felicito, y te agradezco mucho tus excelentes aportes.

        (PD: Ojalá que sí hables español – si no, ¡que pena!)

        Brad Crawford

    • The exact opposite of reality Salvatore.
      Yes solar governs climate (ultimately).
      However it is the Earth’s positioning in relation to the Sun that alters TSI absorbed, crucially at the most sensitive part of Earth’s geography in regard to temp and feedbacks. ~ 65 deg N.
      The fact is that currently that geographical location is not at a time when forcings have/are increased/ing ( especially not in the last 150 years – that is hardly “geological” time).
      So absolutely “this period in the climate” IS unique in that regard.

  4. Every period of warming seems to have been beneficial. It took warming to raise CO2 to the point where agriculture was possible.

  5. Hypothesizing that the ice temperature proxies,
    …which are 2000 year low averages
    are the same as the last 200 years of temporal temperature measurement
    is a bit less than stellar in my book.

    It is great up to about 1500 yag even if averaged over 2000 years,
    the most recent 1500 years, not so much.

    • “Hypothesizing that the ice temperature proxies,
      …which are 2000 year low averages”

      Do you have a reference for this, or you just made it up? Some ice cores have a yearly resolution.

      • First, look up the ice sintering process for the top 100-150m of snowfall as it compacts into ice. UMich has some analysis of the Vostok ice core data.

        Add to that the “Sources of Uncertainty in Ice core data” from NOAA or the like.

      • Mr. anonymous “Javier” person,
        …….”The Modern Warming Period cannot disappear because
        …….. we are significantly warmer than during the Little Ice Age.”
        is certainly a telling position for an anonymous person.

    • Susan, as much As I have criticised Shakun, Marcott, and others, there are non-treemometers that do enable approximate temperatue resolution on ~centennial scales. Don’t throw the baby out with the bathwater.

      • Key words are “approximate” and “centennial scales”.
        Note, additionally, the “assumptions” about O2 ratios and associated error bars (missing) and then glomming/hacking the 20th century temperatures, suitably adjusted, right onto a graph with error bars of, likely, multiple degrees.

      • I tend to agree with Susan. Splicing a thermometer record onto proxy data is just not very convincing to me. For all we know, the H-max was way warmer than is reported by proxy data. … and of course, there is that data manipulation issue by the warmista. Lastly, it smacks of Mikes’ “Nature Trick”.

        This whole historical temperature sector of climate science is way to academic for the purpose of making any future predictions or adopting policy. Rough estimates?? … sure, I can buy that. Accurate .. nope.

      • “I tend to agree with Susan. Splicing a thermometer record onto proxy data is just not very convincing to me.”

        Except that I did not use thermometer data in that graph. It is all coming from proxy reconstructions. So you both are agreeing on something that is not relevant for this cartoon. And quite frankly, it is a cartoon, not a Nature article. Somebody here needs to relax a little.

  6. The Dire Wolf went extinct without any dire warnings about Man’s (Caveman’s) campfire CO2 emissions. Yet Man lived on and learned to make even bigger “campfires”.
    There are a whole bunch more of us now than there were then. Maybe even more animals and types of animals.
    Maybe not Dire Wolfs but there are still a lot wolves. And a whole, whole bunch of dog breeds that didn’t exist and wouldn’t have existed without Man.
    When they talk about what types of critters Man has driven to extinction (in one way or another), today’s meme seems to have shifted from hunting and fishing to “Climate Change”. Why don’t they mention or include the types of critters that Man has introduced?
    Are they claiming that you are not “natural”? Then what about them?

  7. But we do know what the future climate holds. This interglacial wii end. It will get cold and the ice will return. We are already in the orbital configuration for it and the changes can not be stopped.

    As axial tilt decreases, summer warming decreases in the N.H.

    Eventually that reaches a point where ice melt doesn’t complete, and an albedo feedback kicks in as the ice caps grow. This growth continues for 100,000 years as eccentricity and other orbital parameters are shifting so that the next obliquity cycle can’t melt the poles.

    In about 110,000 years, we get the next interglacial as all those changes reverse.

    The unknown part is start date. 300 years? 1000 or 3000? Or did it start with the Little Ice Age, but has a 1400 year short cycle on top of it, so we got this last Hurrah! of almost normal before the next cycle down in 2040 from which we can not recover…

    • “As axial tilt decreases, summer warming decreases in the N.H.”

      Summer warming decreases in both hemispheres with reducing obliquity,

      But knowing that a glacial period is coming is not the same as “knowing” future climate. Nobody has a clue how the climate is going to be. Not in 30 years, not in 100 years and not in 1000 years. A lot of people pretend they know, but they don’t.

      • But in 5000 years there is little doubt about the ice… and in 3000… and only a little in 2000.

        BTW, an important point not mentioned in the article but in “Ice Age” by Mary & John Gribbin, is that length of N.H. summer is a key point.

        The S.Pole always stays frozen, so it isn’t critical. The N.Pole can melt, so it enables the exit from a glacial. Precession and eccentricity controls this. Distance from the sun is unimportant compared to many more days of summer and fewer of winter. Interglacials come when the Earth in N.H. summer is at aphelion, so there are more days of summer to melt N.Pole ice, and added eccentricity enhances this. So yes, both poles are cooling now with reducing tilt, but only the N.Pole matters and an interglacial only comes when it is warmed longest and at high tilt, and glacials return when it no longer is… as is happening now (but takes thousands of years to progress and we don’t know if it started in 1300, or starts in 3600 AD…)

      • “But in 5000 years there is little doubt about the ice…”

        You and I have little doubt, but there are papers calling for an abolition of glaciations or at least a delay of tens of thousands of years due to high CO2 levels. So opinions abound.

        “Interglacials come when the Earth in N.H. summer is at aphelion”

        No, I don’t think so. That would link them to the precession cycle, and interglacials are very clearly linked to the obliquity cycle. While precession redistributes the annual insolation between the different seasons, obliquity actually changes the annual amount of energy that the poles receive and they always receive it maximally during the summer. And that small amount that increases every year added over thousands of years is what determines when interglacials can happen. Whether they happen or not requires other factors.

      • 30 years may be doable if we can figure out ocean cycles. For the longer time points (>60yr), I know I’ll be long dead. And anything past 2100 is policy irrelevant today because we don’t know what the state of mankind’s science-tech will be at that time.

        We may know by 2030 if the sun is indeed entering another Maunder-like minimum, or just a Dalton. And the only thing we can do about that possibility is build resilience with nuclear power and coal. Solar panels covered in ice and iced-up wind turbines won’t cut it to get through the brutal winters of a Wolff-Sporer-Maunder trifecta.

      • According to the EPICA ice cores, [CO2] has followed temperature over the last 850,000 years which makes one wonder what caused the temperature changes given that [CO2] did not.

        This post gives a plausible answer but it is still a long way from presenting a mathematical equation with predictive skill. We should be able to predict when the next glaciation will start but instead we have guesses ranging from 300 to 5,000 years.

    • One thing is for sure Solar cycles, Milankovitch cycles or any other cycle had no effect for periods of tens of millions of years during periods when the climate was in it’s hot phase – i.e. when it was 10-12 degrees C higher than today. Something else was trumping these cycles.
      I know. Lets dump all of the fossil carbon back into the atmosphere and see what happens.

    • I believe in the cyclic nature of the icefields. As the northern cover grows, the southern having more energy input shall shrink.
      Ancient maps show islands in the South, that are under ice, why? Some old data suggests northern routes for trade existed prior to the current ice. And we’re used extensively. How old would those stories and fables be if such were true today.
      Other interesting modeling I’ve seen, explain parts due to our orbit, and changes done to our orbit, by outside passing gravity fields, and not just our giant neighbors. And explanations of the sun’s workings, and that a reduction is occurring,
      But, it’s a pretty day, the sun is still warm, the sky is blue. Picnic?

  8. There is a nit to pick with the Javier cartoon, as well as Josh’s and the xkcd charts. There is no indication of uncertainty (i.e. error bars) with the temperature line. There is an assumption of certainty built into those wavy lines that may not be warranted.

  9. They all make the same error of comparing 100 year averaged temperatures with annually averaged temperatures in the last 150 years. Average the last 100 years and suddenly the hockey stick disappears.

    • The Modern Warming Period cannot disappear because we are significantly warmer than during the Little Ice Age.

    • Great graph. And now I have a clear idea of the older and younger Dryas.

      Couple of spelling slips. “Woolly Rhino”, not “Wolly”. (Presumably went extinct because people kept shearing them to make sweaters, and they froze to death.) “Hun invasion .. triggered by a draught.” Bit of draught-proofing and we’d have been spared that. Unless it means draught beer.

      Yes, I know you wouldn’t want to try shearing a rhino. Neither would I. But people were tougher in those days.

  10. [SNIP waaaaaaaaaaaaaaaayyyyyyyyyyyyyyyyy waaaayyyyy WAAAAAAAAYYYYYY!!! off topic – election politics do NOT belong on this thread /mod]

    • Milankovich Cycles are an input, and the solar cycles are an input. But plate tectonics can have effects, also. Moving Antarctica over the South Pole, moving continents around the North Pole, block ocean currents from transporting warm water to the poles, and cold water to the equator, making extremes of cold worse. This is probably a necessary pre-condition for glacial eras to occur. And who knows what effects changing the water depth in the ocean basins does to currents. Still lots of unknowns yet.

  11. [2] Older Dryas cold period: At the end of Bølling, temperatures suddenly plummeted about 20° F (~11°C) from the Bølling maximum in a few hundred years (Figs. 1, 2), initiating the Older Dryas cold period, which lasted from about 14,300 to 14,000 years before present (BP). Temperatures returned to near full glacial levels and glaciers halted their rapid retreat.

    An example . Why Javier ? Why did it happen not only that timer but over and over again as the historical climatic record shows from 20000-10000 years ago. The YD not the exception at all.

    • We think we understand how Dansgaard-Oeschger events manifest, we just have no idea of what causes them, but hypothesis abound.

      Personally I think there’s plenty of evidence that they are not of solar origin.

      My bet is that it is an oceanic oscillation that it is paced by an extraterrestrial force. But it is not going to be easy to prove that.

  12. Just watched two episodes of a documentary on “apocalypses”, and although they concentrated on Mayan and Egyptian, it was a science show, no ooga booga, aliens or mysterious stones. Mostly they focused on droughts.

    Both shows had scientists who presented evidence that said droughts were natural, some due to ice ages that were somewhat predictable (in 1,500 year cycles for a few hundred years), and that the climate could change very quickly, in decadal time spans.

    So none of this is a shock to non-political scientists. And I don’t mean “political scientists”…

  13. Just wanted to note there are 4 different changes in the Milankovitch Cycles.

    But it doesn’t matter. It is the combination of all 4 and then how much ice really builds up and then how long it takes to break the back of the glaciers that matters. When the glaciers cover the northern hemisphere to 50N, it doesn’t even matter if there is an upturn in solar energy for 22,000 years, The glaciers just reflect the majority of that extra energy back to space and it has no effect because it is not long enough.

    Secondly, 65N is not really the latitude to focus on. Solar energy at 65N is ALWAYS high enough to melt the winter snow at this latitude. It is really 75N where the make-or-break action happens. At this latitude, solar radiation can fall below 440 W/m2 in the summer and the snow doesn’t get to completely melt out.

    At 65N, the lowest the solar radiation gets in a Milakovitch is something around 500 W/m2 in a summer period and this is more than enough to melt out ALL of the snow at that latitude by April or early May.

    It is 75N, where the Arctic sea ice may not melt out or the snow on land doesn’t get melted out.

    Then Albedo takes over and the glaciers can slowly move south.

    Even in the deepest downturns of the Milankovitch, the glaciers are melting furiously at their southern edges in the summer because the solar energy is still very high. Summer in Chicago? Summer in Churchill? There is NO WAY the glaciers can build up at these latitudes in the summer. They build up at 75N and that latitude only until they reflect so much sunlight that the glaciers can push south.

    It takes a good upturn and several upturns to break the back of the Ice-Albedo sunlight reflectance of the glaciers. There are models to show how this happens but just review this chart and notice how chaotic it really is.

    Summer solar radiation is not going below the magic 440 W/m2 at 75N for 130,000 years. This interglacial is going to be longest one yet. Milankovitch Cycles are not at all regular.

    • Just wanted to also note that continental drift pays a major part in these cycles.

      If there is not enough land mass at 75N or 75S, the Milankovitch Cycles have no impact. Really large continental scale glaciers cannot build up on the ocean at 75N, only land-based glaciers can.

      One issue with the ice ages which is also important, is that if glaciers build up on land at 75N or 75S, they can depress the land enough so that is now becomes Ocean for up to 1.0 million years at a time and this can suppress glacial activity for a long period of time.

      For example, we have Hudson Bay, and the Barents Sea and the Kara Sea and East Siberia Sea which were all above sea level until 2.0 million years ago or so. The glaciers that built up starting at 2.6 million years ago, pushed them below sea level and this completely changed the impact of the Milankovitch Cycles. The timelines of glaciation then do not compare to what really happened with solar radiation at 75N because now we are dealing with ocean conditions which were previously land conditions.

      As North America has slowly rotated more north-west, this left more and more landmass above sea level so that that Milankovitch Cycles could begin to have an impact in the last 2.7 million years. Before this time, it was too far south for glaciation. As northern Eura-Asia dropped lower and lower because of previous glaciations starting at 2.7 million years ago, this also affected the timeline of Milankovitch Cycle impacts.

      When Gondwana moved across the South Pole between 460 million years ago and 290 million years ago, there are periods when glaciation could not happen because all of the landmass was pushed below sea level due to previous glaciations.

      • This image shows the paleocoordinates and vectors for the SW and NE corners of Colorado since the breakup of Pangea, according to Christopher Scotese. The intervals are 10 million years. The last two points beyond the vectors represent the present positions.

        Nothing here suggests a clockwise rotation to my eye. One might even argue for modest anticlockwise rotation over the last 40 million years, but the overall trend since the Cretaceous is certainly NE to SW. This direction fosters closing of the Isthmus of Panama (not that I believe this saw either), but moves North America away from 75 North.

      • Current Continental drift according to GPS stations.

        Since the driver of these movements (for North America, sea floor spreading at the mid-Atlantic ridge) has been in operation for the past 55 million years, one can assume at least North America has been following this pattern for a good 20 million years or so.

    • Bill,

      “Summer solar radiation is not going below the magic 440 W/m2 at 75N for 130,000 years. This interglacial is going to be longest one yet.”

      Analogy with MIS19, the closest astronomical analog to the Holocene suggests otherwise.

      MIS19 started its accelerated temperature decline in what would be 1000 years from now with exactly the same NH summer insolation we have.

      No magical wattage, just continuous decline in obliquity.

      • Radiation measures from the radiation tower at Eureka, Canada at 80N.

        The SW Down is solar radiation which gets to about 520 W/m2 at the June equinox.

        The SW UP is how much of that solar radiation is reflected back up by the snow. Very high until early June when the snow completely melts out. It doesn’t really start reflecting again until late September (snow is back) but by then, the Sun in getting low and by late October they don’t get any sunlight, several hours of sunrise only.

        This is 80N and the forecasts from Milankovitch modelling is that the summer solar radiation will decline here by about 0.5 W/m2 over the next 2,000 years and then go back up again. It will stay high until 52,000 from now when it will drop considerably but probably stay above the 440 W/m2 I was talking about and then go back up again. The big drop does not happen until 128,000 from now. 65N is always much higher than these numbers. So if the snow melts at Eureka Canada by early June, there is no ice age coming.

      • Bill,

        You say it is not possible, yet I have showed you a precedent. I don’t know how it happened 777,000 years ago, but it clearly happened, so it can happen again. It will most probably happen again. Even now in the midst of the current warming there are permanent ice patches in places in Canada. They were much increased during the Little Ice Age. They will increase and expand and create cold micro-environments, and that Sun will not melt them despite our calculations. Every year for thousands of years they will grow, with temporary set backs.

      • Just 2,000 years ago, the active Glacial field we see today at Mt. Logan,was nearly absent. The same with A Glacier in Wyoming,that exist today,but absent before the LIA came along.

        John Kehr talked about this and more in his worthy book,

        He like you Javier,showed that we have a new ice age phase ahead of us. He thinks we are in the late Autumn phase in the Holocene.

    • Bill, I like the graph, but it’s showing summer insolation at 75N compared to an Antarctic ice-core. I realize the Antarctic cores give a generally good global temp, but prb’ly better to compare 75N w/a Greenland core.

  14. Great graphic Javier. It should be used in schools, 1st year Uni, and elsewhere.

    “Wolly” the Rhino is fine by me :-)

  15. “The Hun invasion may have been triggered by a draught…”

    If correct, this gives hooliganism a significant historical precedent, and throws a fly into the “Climate-Change-Driven Drought” theory of the Huns’ predations. ~_^

  16. If precession doesn’t matter, then why didn’t maximum obliquity melt the NH glaciers two cycles ago? Maximum obliquity must coincide with NH summer solstice at perihelion along with lower eccentricity in order to melt the ice sheets. These cycles coincide just right approximately every 100,000 years.

    • Precession does matter, but obliquity is the leading factor. That’s why the spacing was 41 kyr at early Pleistocene and not 23 Kyr. The 100 kyr cycle is a myth. Analyze every peak from sedimentary record (not ice core) for the past 1 million years and its distance to the next and you will see.

      • Since I am not capable in creating a graph with proper info, perhaps someone could chart the obliquity cycle for the last 800,000 years and overlay by single points the time when NH summer solstice was at perihelion. That would show if or when the two cycles lined up to cause and interglacial.

  17. Nice graphic for sure. As has been said here before…basing any kind of correlation on half of one cycle is pointless. Does the correlation hold for the last million years? Not at all.

    • The correlation between obliquity and temperatures holds for the last million years. It is surprisingly strong.

      Prof. John Baez run a Gabor transform modulus, which is a time-windowed Fourier analysis on both 65°N summer insolation from calculations, and on temperature data from Epica Dome C ice core record. The first one shows the expected strong bands corresponding to Precession (19 and 23 kyr) and obliquity (41 kyr), while the last showed very weak bands at precession and a strong band at obliquity, as well as at 100 kyr.

      The conclusion is clear. The correlation between obliquity and temperatures is much stronger than between precession and temperatures.

      • This correlation is largely spurious.

        Obliquity only comes to the fore when eccentricity is low – so in the current era and 400 ky ago. Thus interglacials are realigned with obliquity at these times, which skews the frequency analysis towards obliquity even though orecession is the dominant factor.

        In the meantime, interglacials are predominantly modulated by precession. But since the seasonal precession cycle (the Seasonal Great Year) has an average periodicity of 22ky, it is substantially similar to the obliquity cycle. So what you are seeing in that frequency analysis is an average 22 ky precessional mosulation of ice ages, that is held in the close embrace of obliquity every 400 ky, and therefore has to remain in step with it.

        So the ice age cycle is much more complex than you imply. This is why recent ice ages have lengths of 90 ky and 115 ky, because they are primarily driven by precession, not obliquity.

        Precessional SGY lengths (in kyr) for the last 5 glacial cycles. They equate precisely with the ice age cycle. Total ice age cycle lengths for these five groups are: 117, 115, 91, 90, and 95 ky. The last ice age is at the top.

        23, 21, 26, 22, 25

        23, 22, 23, 24, 23

        21, 21, 27, 22

        16, 22, 15, 17, 22

        25, 20, 22, 17, 21

      • Ralph,

        “This correlation is largely spurious.”

        So what you are seeing in that frequency analysis is an average 22 ky precessional mosulation of ice ages, that is held in the close embrace of obliquity every 400 ky, and therefore has to remain in step with it.”

        No, that is what you see. I don’t see that at all. For a start a 22 kyr cycle and a 41 kyr cycle cannot be kept in step. They just drift all over the place and when they do temperatures always follow more strongly obliquity than precession. I pick that by eye when I analyze the data over colored backgrounds and it is the same the Gabor transform is picking mathematically. So no, it is not spurious. The lack of a 23 Kyr band in the temperature frequency analysis is also very significative and difficult to dismiss as also spurious.

        “This is why recent ice ages have lengths of 90 ky and 115 ky”

        I believe you are not counting correctly. Oz4caster has done this beautiful alignment of the Epica data:

        Now look at the frequencies considering the value given at zero by Oz4caster as the termination time:

        T5 to T4 82900 = 41000 x 2
        T4 to T3 91700
        T3 to T2 112400
        T2 to T1 119700 = 41000 x 3

        T4 to T2 204100 = 41000 x 5
        T5 to T1 406700 = 41000 x 10

        Close to multiples of 41,000 except for the deviation at Termination 3.

        So the answer is that we have interglacials at multiples of obliquity but with a frequency of two and a half cycles of obliquity which is the average value of 100 kyr.

      • By the way, look at what happened after Termination 3. That termination came 10,000 years too late according to the obliquity cycle. It is the green interglacial. It appears that something delayed it, and you can see that this interglacial was cut short by 10,000 years when the obliquity cycle turned down according to schedule.

        If termination 3 is where it should be according to your model, then you don’t have an explanation for why it was such a short interglacial, while I do.

      • >>The lack of a 23 Kyr band in the temperature frequency analysis is
        >>also very significative and difficult to dismiss as also spurious.

        As my list demonstrates, the seasonal precessional cycle has varied from 15 ky to 25 ky, over the last 400ky, so how are you going to pull a 22 ky frequency analysis from that? It is not as stable as the obliquity cycle.

        Besides, the precessional-obliquity cycle has little influence on temperatures, if the ice sheets are not loaded with dust. You can see that in the orbital cycle record, where significant numbers of precessional and obliquity cycles have no effect on ice volume and temperatures.

        What is your reasoning for many of the obliquity cycles demonstrably having no effect on temperature?


      • >>If termination 3 is where it should be according to your model, then
        >>you don’t have an explanation for why it was such a short interglacial.

        I fail to see the problem. The interglacial 240 ky ago was cut short because precessional insolation at 65 degrees north plunged very rapidly into a very deep Great Winter. And that cooled the climate very quickly. You can see that all these Milankovitch cycles are dominated by precession. Count them, and count the cycle lengths.

        You will note that interglacial warm periods during periods of high eccentricity (350 ky to 50 kyr ago) are all quite short, because they are primarily modulated by precession which has a shorter cycle length (about 23 ky). The only long interglacials are 400 ky and 10 ky ago, because eccentricity was low at this time, and obliquity was more dominant, and the interglacial was able to expand to almost the full length ofnrhe 41 ky obliqiity cycle.

        Temperature vs precession.

      • Ralph,

        “the seasonal precessional cycle has varied from 15 ky to 25 ky, over the last 400ky, so how are you going to pull a 22 ky frequency analysis from that? It is not as stable as the obliquity cycle.”

        The variable seasonal precession cycle is picked very well by the Gabor transform module of Prof. Baez in the left panel of the figure showed. The failure to pick it in the temperature data is not a problem of the method. It is simply not in the data.

        “the precessional-obliquity cycle has little influence on temperatures, if the ice sheets are not loaded with dust.”

        This is just an hypothesis and thus cannot be used as an argument.

        “What is your reasoning for many of the obliquity cycles demonstrably having no effect on temperature?”

        All obliquity cycles have an effect on temperatures, as the Gabor transform module and analysis of the data clearly show, just not all produce an interglacial, because although it is the main factor, it is not sufficient by itself since the mid-Pleistocene. Before it was, because the Earth was warmer.

        “The interglacial 240 ky ago was cut short because precessional insolation at 65 degrees north plunged very rapidly into a very deep Great Winter.”

        Same happened at the end of Eemian and the interglacial was not cut short as your figure shows. Glacial inceptions at 120 kyr, 330 kyr and 400 kyr responded very differently to the drop in insolation that at 240 kyr. You don’t have an explanation for that, and I do. It was obliquity.

        “The only long interglacials are 400 ky and 10 ky ago, because eccentricity was low at this time, and obliquity was more dominant, and the interglacial was able to expand to almost the full length ofnrhe 41 ky obliqiity cycle.”

        Current interglacial is not long, it is standard so far, as the graph from Oz4caster above shows. Right on schedule to end in 1000-2000 years as MIS19 also did. You are making a rule from one exception, MIS11.

  18. There will never be another ice age. The soot produced by fossil fuel burning will prevent ice accumulation. When people come agreement on that concept they will just spread black soot over ice in the NH to prevent any buildup. Simple really.

    • That is left for future generations, if there is anything left to burn at that time. In any case I wouldn’t hold my breath for that solution working, because it assumes that ice albedo is a major feedback factor, and perhaps it is not.

  19. Javier. Actually, precession is key, with obliquity being secondary.

    A small fact. …. Obliquity cannot warm the globe, it merely borrows warmth from the tropics and gives it to the poles. And both poles simultaneously. (Unlike precession, which warms the poles alternately.). This is how world glacial temperatures are modulated.

    This is instructive. It tells us that ice age modulation is not a global phenomina, it is a regional phenomina. And interglacials only happen when precession provides a Great Summer (increased insolation) in the NH (often in conjunction with obliquity). So a global gas like Co2 is going to be far less influential, than regional insolation and albedo fluctuations in the NH.

    It also tells us that another feedback agent must be operating. If global insolation remains constant, while the world warms into an interglacial, something else must be driving this change. And that feedback agent is albedo. The increased NH insolation melts the ice caps, and it is the reducing albedo that absorbs the insolation and warms the world. So albedo is key, while Co2 does – well – not a lot. (The SH is not so important, because it has smaller landmasses and no great tropic-encroaching ice sheets. Somit has smaller albedo changes.)

    In addition, obliquity is only more influential now because eccentricity and therefore precession are both at minimums. And will be so for the next 100 ky. Normally it is precession that has the greatest influence on NH insolation, and therefore upon glacial and interglacial temperatures.

    Modulation of Ice Ages.


    • Ralph,

      “A small fact. …. Obliquity cannot warm the globe”

      Evidence shows the opposite. Temperatures follow obliquity

      “And that feedback agent is albedo.”

      That’s your proposed feedback agent. I disagree, but to my knowledge there is no clear evidence to resolve this issue.

      “In addition, obliquity is only more influential now because eccentricity and therefore precession are both at minimums.”

      It is actually the opposite. Obliquity has less effect when eccentricity is at a minimum and gets less help from precession. The result is that interglacials are more spaced when eccentricity is low. The next interglacial is going to be a long wait.

      • >>“A small fact. …. Obliquity cannot warm the globe”
        >>Evidence shows the opposite. Temperatures follow obliquity.

        Total global insolation does not change at all, during the obliquity cycle. So please tell me how a steady global insolation value can change temperatures, if you do not invoke another feedback agent or causal agent.

        >>It is actually the opposite. Obliquity has less effect when eccentricity
        >>is at a minimum and gets less help from precession.

        Completely wrong. The regional insolation variations caused by obliquity do not change significantly with changes in eccentricity. But the regional insolation provided by precession does reduce significantly with reducing eccentricity. Ergo, the proportion of obliquity insolation gets larger, with decreasing eccentricity. Plot the graphs, and see for yourself,

      • “Total global insolation does not change at all, during the obliquity cycle. So please tell me how a steady global insolation value can change temperatures”

        Total global insolation was essentially the same at the Last Glacial Maximum and at the Holocene Climatic Optimum, so clearly the difference is made by when and where that insolation is falling.

        ” Ergo, the proportion of obliquity insolation gets larger, with decreasing eccentricity.”

        I did not make myself clear. Obliquity needs the help of Precession to reach sufficient warming, so without enough Precession factor it doesn’t matter that obliquity becomes the dominant factor. Interglacials are more spaced when Eccentricity is low, like now, and less spaced when Eccentricity is high, like 200 kyr ago.

  20. I thought ice core data showed that the medieval warming was warmer than now. Same for Roman. It was supposed to be world wide.

    • Nobody knows for sure. Most proxies usually do not include very recent dates so we don’t have an apples to apples comparison. Glaciers all over the world and some biological data, like the tree line in altitude, suggest that we are currently warmer than the medieval warm period.

      Greenland ice cores have lots of issues, and they don’t reflect global climate very well.

    • It was supposed to be world wide.
      Regional warming is a red herring bit of nonsense. If only Europe warms, the average temperature of the earth increases. You have global warming. The only time this is not true is if somewhere else cools at the same time. At which time the average temperature of the earth is unchanged.

  21. Javier,
    A great presentation!
    But I have to disagree on some issues.

    It is very important to keep things in perspective.
    The earth is a sphere and receives the same amount
    of direct light regardless of angle.
    Especially keep in mind just how small a 2.4 degree
    shift really is.

    The earth’s iron core rotates slightly independent of the
    outer layers and outer crust. The importance of obliquity
    is not angle, but the fact that the earth is constantly shifting
    in directions that are contrary to the motion of its iron core.
    During an inter-glacial obliquity is a magnetic field amplifier.

    During a glacial period the earth’s iron core loses its wobble
    because the sun has suffered a substantial loss of acceleration.

    Please visit . . Weathercycles.wordpress
    ” Fibonacci and climate “

  22. At the risk of repeating myself I will again assert that the primary factors in climate change here on our world are solar, orbital and planetary dynamics. I proved it to myself through a simple mathmatical comparison of the relative energies. You can do thr same if you simply gather and objectively analyze the relevant data that has been available for decades.

  23. I have look at all the post but no one is able to address the abrupt climatic changes that took place especially 10000 years ago to 20000 years ago. Especially the sharp down turns in the temperature.

    The YD being a prime example but there are so many similar events all I would have to say are not solved.

    The change to abrupt cold periods such as the YD remain a mystery in contrast the abrupt warm ups do have some good theories one I will pass along in my next post.

  24. Humans are estimated to have arrived in Australia about – 50,000 years ago – However, the extinctions of the megafauna being wiped out by this was actually a hypothesis of Tim Flannery – the ex- climate change commissioner – effectively knocked off his post by Tony Abott (yeah cheers finally) . Thus he even got his own science wrong – all scientists who still work in the field Tim Flannery abandoned for his new opportunistic field – now agree that for the 1st 25,000 years the indigenous Australians lived alongside the megafauna – and believe it not – how ironic – it is now theorised (thus evidence to support it) that the Australian megafauna were wiped out by – yes you guessed it – climate change!!!

  25. Sirocko and Seelos, 2005, A late Eemian aridity pulse in central Europe during the last glacial inception, nature, vol. 436, 11 August 2005, doi:10.1038/nature03905, pp 833-836, state:

    “Investigating the processes that led to the end of the last interglacial period is relevant for understanding how our ongoing interglacial will end, which has been a matter of much debate…..

    “The onset of the LEAP occurred within less than two decades, demonstrating the existence of a sharp threshold, which must be near 416 Wm2, which is the 65oN July insolation for 118 kyr BP (ref. 9). This value is only slightly below today’s value of 428 Wm2. Insolation will remain at this level slightly above the [glacial] inception for the next 4,000 years before it then increases again.”

    “The possible explanation as to why we are still in an interglacial relates to the early anthropogenic hypothesis of Ruddiman (2003, 2005). According to that hypothesis, the anomalous increase of CO2 and CH4 concentrations in the atmosphere as observed in mid- to late Holocene ice-cores results from anthropogenic deforestation and rice irrigation, which started in the early Neolithic at 8000 and 5000 yr BP, respectively. Ruddiman proposes that these early human greenhouse gas emissions prevented the inception of an overdue glacial that otherwise would have already started.”

    conclude Muller and Pross (2007)

    “Here, we study the questions why we still live in an interglacial world and when we should expect the end of the Holocene under natural conditions (no anthropogenic influence) or under anthropogenic perturbations (also referred to as “Anthropocene”), questions which attracted considerable interest in recent years. It was argued that without earlier anthropogenic activity we would live already in glacial world (Ruddiman’s hypothesis). Tzedakis et al. (Nature Geoscience, 2012), using MIS 19 as the best analogy in terms of the orbital parameters for the Holocene, suggested that the new glacial inception would start within the next 1500 years, assuming natural CO2 level of 240 ppm. However, 240 ppm is much lower than preindustrial CO2 level and CO2 concentrations during several most recent interglacials (starting from MIS 11). Here, using the comprehensive Earth system model of intermediate complexity CLIMBER-2, carefully calibrated for the simulations of the past eight glacial cycles, we show that (i) although climate conditions during late Holocene were very close to the bifurcation transition to the glacial climate state (Calov and Ganopolski, Geophys. Res. Lett., 2005), it is very unlikely that under pre-industrial CO2 level (280 ppm) glacial inception would occur within the next several thousand years; (ii) it is likely that the current interglacial, even without anthropogenic CO2 emission, would be the longest interglacial during the past million years; (iii) current CO2 level makes new glacial inception virtually impossible within the next 50,000 years; (iv) in agreement with earlier result of Archer and Ganopolski (Geochem. Geophys. Geosyst., 2007) based on a conceptual model of glacial cycles, we found that consumption of a large portion of available fossil fuel could postpone the next glacial inception by hundreds of thousand years.”

    • I am aware of those studies. They are proposing that the current interglacial will be unprecedentedly long based on certain assumptions.

      Ruddiman’s hypothesis is highly controversial and heavily contested by many of the experts in carbon stores and methane sources. It has captivated the press and the public but failed to make a similar impact in the scientific community. He has done some serious cherry picking of data that leads to questions about his integrity in acknowledging his hypothesis’ weaknesses.

      The CO2 hypothesis of a delayed glacial start relies on two unproven assumptions, a high climate sensitivity for CO2 and a very long residence time for CO2. There is evidence that both assumptions are incorrect.

      The insolation hypothesis relies on the hypothesis that glaciations main determinant is northern summer insolation and on the assumption that current levels are enough to prevent a glacial inception. while the first is under discussion and without agreement, the second is clearly false. Proponents of this hypothesis have failed to do their homework as there have been previous glacial inceptions with current insolation levels, for example MIS19.

      While northern summer insolation is going to remain the same for the next few thousands of years, obliquity is going down at its fastest rate, which is going to make both poles colder with time, and the colder the poles the more heat that the Earth loses, independently of insolation. Failure to see this important fact is leading these people to the incorrect assumption that we are going to have an unprecedented interglacial. Luckily for us all we are not going to be around to see the failure of that prediction.

    • Really nice chart showing 65N summer solar insolation for the past 200,000 years and then going out for the next 135,000 years. [Note that this modelling can be very accurate for +/- 5 million years and outside of this large boundary, the uncertainties get too large].

      Basically we are already at the end of a low cycle right now and the next low cycle is 52,000 years out. But I don’t think a 15 W/m2 drop at that time versus today is enough to put us into an ice age. The next big drop is 130,000 years from now which itself may not be enough for ice age as well (probably close though).

      • “Basically we are already at the end of a low cycle right now”

        End of a low cycle in northern summer insolation. The end of a low cycle in obliquity is 10,000 years away. Both poles are going to continue cooling until then, and the rest of the planet likely too.

  26. This is the best post ever! Sorry I missed it as I was travelling. Thanks Javier, I am impressed by your understanding of Milankovitch cycles.

    I have often wondered if one of the important triggers for glaciation might actually be the moisture content of the arctic air masses rather than the temperature, especially in the Canadian archipelago region. The region is fairly dry today, and its difficult to imagine enough snowfall in this region to initiate glaciation no matter how cold it is. Is there a possible mechanism where influx of warmer water into the arctic periodically shifts the climate to a much greater maritime influence, transforming it from a relatively dry climate to one of heavy snowfall? Ive always wondered if it really needed to be much colder to initiate glaciation, or even if glacier formation could be starved of snowfall to bring about the beginnings of interglacial.

    Thanks again for a great post and great comments.

Comments are closed.