Ned Nikolov: Dispelling the Milankovitch Myth

Posted: January 3, 2022 by tallbloke in climate, Critique, Cycles, data, Ice ages, Natural Variation, solar system dynamics, Uncertainty

Ned Nikolov, Ph.D.
Dec 30, 2021

There has been a long-standing belief in Paleoclimatology that orbital variations (a.k.a. Milankovitch cycles) have been responsible for the initiation and/or duration of glacial cycles (Ice Ages) over the past 800 Ky. Milankovitch cycles are often referred to as a pacemaker of the Ice Ages. This myth dates back to 1970s, when sediment cores revealed a weak correlation in the frequency domain between Earth’s 41-ky obliquity (axial-tilt) cycle and the periodicity of Ice Ages during the early Pleistocene (Quaternary). However, in the late Pleistocene, the frequency of glacial cycles better match the Earth’s 100-ky eccentricity cycle, which further fueled the confusion. Yet, no one has been able to demonstrate a meaningful relationship between glacial cycles and any of the Earth’s 3 orbital parameters obliquity, eccentricity and precession or combination thereof on a linear time scale. A physical causation requires a strong correlation between parameters in the time domain, not the frequency domain!

Using recent data describing the dynamics of global surface air temperature inferred from geological proxies and variations of Earth’s orbital parameters computed by the best available orbital models, we show here the lack of a physically meaningful relationship in the time domain between Milankovitch cycles and Ice Ages for the past 784 ky. Orbital data came from the state-of-the-art model by Laskar et al. (2004 and 2011) and were downloaded from a page on the official website of the Virtual Observatory Paris Data Center in France. The top-of-the-atmosphere (TOA) solar insolation on summer solstice at 65o N latitude was obtained from the Milankovitch Orbital Data Viewer of Colorado State University. A time series of global surface air temperature for the late Pleistocene (past 784 ky) was constructed from published reconstructions by Snyder (2016) and Friedrich et al. (2016). All data series used in the analysis share the same temporal resolution of 1,000 years. Standard Score (a.k.a. Z-Score) is used in some graphs to plot time series having different measurement units on the same axis.

Figure 1.

Changes in Earth’s mean annual distance to the Sun (measured in Astronomical Units, AU) resulting from variations of planet’s orbital eccentricity have been minuscule for the past 800 ky (Fig.1) causing only a ± 0.05 W m-2 variation in the Earth’s baseline TOA Total Solar Irradiance (TSI) (Fig. 2). According to recent solar reconstructions (e.g., Egorova et al. 2018), this variation is 50 – 100 smaller than the TSI fluctuations caused by Sun’s magnetic activity on centennial to millennial time scales. 

Figure 2.

Hence, orbitally induced TSI changes are bound to have an immeasurably small effect on Earth’s global surface temperature. The latter shows practically no correlation to TSI variations induced by the Milankovitch cycles (Fig. 3a and 3b).

Figure 3a

Figure 3b

Earth’s obliquity (axial tilt) varied narrowly between 22.3o and 24.5o for the past 800 ky (Fig. 4) while showing no relationship to global surface temperature over this time period (Fig. 5a and 5b). The linear correlation between obliquity and the global surface temperature is R2 = 0.053 (Fig. 5b).

Figure 4.

Figure 5a.

Figure 5b.

Earth’s orbital eccentricity also varied over a narrow range (from 0.004 to 0.05) during the past 800 ky (Fig. 6). Note a pronounced 400-ky cycle in the eccentricity time series, which is not found in the global temperature record (Fig. 7a). The correlation between global temperature and eccentricity is rather weak (R2 = 0.235) (Fig. 7b) albeit a bit better than the correlation between temperature and TSI or temperature and obliquity discussed earlier.

Figure 6.

Figure 7a.

Figure 7b.

Figures 8a and 8b depict the relationship between rates of change of global temperature and Earth’s orbital eccentricity. These time-derivatives are correlated with a coefficient R2 = 0.341, which is the strongest relationship found between global temperature and any orbital parameter of Earth! However, this correlation does not imply a physical causation, because eccentricity changes have a negligibly small effect on TOA TSI and surface temperature.

Figure 8a.

Figure 8b.

Paleoclimatologists claim that Milankovitch cycles affect Earth’s climate chiefly through changes in the solar flux reaching the top of the atmosphere at 65o N latitude on the day of June summer solstice. According to this concept, a decreased summer isolation at 65o N due to a specific configuration of obliquity, precession, and eccentricity impedes the summer melting of snow at high latitudes and promotes ice accumulation, which over time allows glaciers to grow, thus initiating an Ice Age. The reverse process is believed to take place, when the summer isolation anomaly at 65o N is above its long-term mean (baseline value). During such periods, more snow/ice is expected to melt in the summer, which is thought to initiate deglaciation… However, the data reveal a complete lack of a relationship between summer insolation at 65o N and the global surface temperature for the past 784 Ky (see Figs. 9a and 9b). The correlation coefficient between these time series is a meager R2 = 0.031 (Fig 9b).

Figure 9a.

Figure 9b.

Roe (2006) claims to have found a strong correlation between the rate of change of Earth’s ice volume and the TOA June insolation anomaly at 65o N latitude. However, in his study, the ice volume (measured in “arbitrary units”) was estimated by models relying on oxygen isotopes. Noticeably one of the models (SPECMAP) dating back to 1984 assumed a-priori that ice volume and orbital forcing were related. Hence, the ice volume estimates utilized by Roe (2006) appear to be biased toward orbital cycles and much less reliable than modern proxy-based reconstructions of global temperature. Also, planetary-level climate change is physically much better defined through variations of the absolute global surface temperature rather than fluctuations of some unitless ice-volume estimates. This is because temperature is a fundamental metric controlling the dynamics of icesheets and sea ice. In an effort to “mimic” Roe’s approach as closely as possible, we compared the rate of change of global surface temperature to the TOA June insolation anomaly at 65o N latitude for the past 784 Ky (Fig. 10a & 10b). The correlation coefficient between these time series is R2 = 0.254 (Fig. 10b), which suggests an effective lack of control by the orbitally driven high-latitude summer insolation changes on the global surface temperature of Earth.

Figure 10a.

Figure 10b.

CONCLUSION: The available data indicate that, in the time domain, the Milankovitch orbital cycles are poorly correlated (if at all) to changes of global surface temperature inferred from sediment- and ice-core proxies for the past 784 Ky. Hence, the geological record provides no evidence that Ice Ages of the past one million years were controlled or even influenced by known variations of Earth’s orbital parameters. Putting to rest the Milankovitch orbital hypothesis of climate change as an unsupported conjecture seems to be an important and necessary step toward developing a new and physically robust Paradigm of paleoclimate drivers as discussed in this video:

  1. Ned Nikolov says:


    What’s wise IMO is to stick to hard quantitative evidence, not prejudices. That’s my guiding principles. As I said in one of my comments above, in a scientific investigation, we should always remember the wise words of Arthur Conan Doyle:

    Once you eliminate the impossible, whatever remains, no matter how improbable, must be the truth.

  2. Pablo says:

    Brilliant thread…thanks everyone.

  3. Ned Nikolov says:

    Thank you, Pablo! This is dramatic storm video!!

  4. oldbrew says:

    Mars air pressure theory, originally from Dr. Jeffrey Kargel of the U.S. Geological Survey in Flagstaff, Arizona.

    The Obliquity of Mars

    As Mars’ obliquity — and thus the warmth of its polar summers — increase still more, however, the second stage of the process starts: some of the accumulated patches of surface ice become as much as 20 deg C warmer.

    And the increased warmth of the polar regions has another important effect (recognized by almost all Mars researchers): a great deal of carbon dioxide stored underground — either as frozen dry ice, a “clathrate” mixture of CO2 with water ice, gaseous CO2 that has been “adsorbed” by Mars’ soil (that is, chemically stuck to tis grains), and perhaps even underground liquid CO2 under pressure — evaporates and is released into the air.

    Mars’ somewhat cooling equatorial soil absorbs part of this new CO2 — but it is well outweighed by the thawing polar CO2, and so Mars’ air pressure significantly increases.

    This still leaves its air pressure only a tiny fraction of Earth’s (at most, about 4%) — but it produces enough of an added CO2 greenhouse effect to warm Mars by another 10 deg C. And it also raises Mars’ air pressure enough that liquid water can exist on its surface (indeed, it’s only a little too low for liquid water to briefly exist there right now).

    The result of these two factors — as you might expect — is that many of the accumulated thick patches of water ice on Mars’ surface are able to melt, producing the runoff gullies we see today.

    But then — as the next stage of the obliquity cycle occurs and Mars’ axial tilt starts to shrink again — these high-latitude areas cool down again, and the extra carbon dioxide is once again absorbed back into Mars’ soil and subsurface, lowering its air pressure again to its current tiny level (or even lower, during those period when Mars’ obliquity is less than its current 25 degrees).

  5. oldmanK says:

    Speaking of the obliquity of Mars, I recall reading this:

    Of course it had to be crustal shift, as otherwise it interferes with the dogma of the models.

  6. oldbrew says:

    A Causality Problem for Milankovitch

    Daniel B. Karner and Richard A. Muller
    Department of Physics
    University of California
    Berkeley, CA 94720

    Click to access Causality.pdf

  7. Ned Nikolov says:

    Thank you, Oldbrew!

    Many researchers have noticed over the years the inconsistencies between Milankovitch cycles and glacial fluctuations, but the “Milankovitch myth” continued to persist, because there was no other driver that science knew about (expect for the fictitious “CO2 radiative forcing”) that could explain the Ice Ages… until now!

    We now have a quantitatively verified alternative concept: Ice Ages were not a result of radiative forcing as everyone thought, but a consequence of a varying-through-time adiabatic forcing. This could have never been realized if it was not for our discovery about the adiabatic (pressure-induced) nature of the Atmospheric Thermal Effect. As it often happens in science, the discovery of a new paradigm in regard to one phenomenon leads to a complete overhaul of our understanding about other related phenomena…

  8. Paul Vaughan says:

    supplementary inclination

    “Figure 2. Earth’s orbital parameters from the La2010d astronomical solution […] (A) The orbital eccentricity variation, 0–10 Ma. (B) Periodogram of the orbital eccentricity variation shown in (A). (C) The orbital inclination variation relative to the invariable plane, 0–10 Ma (Note: Fig. 6 […] La2004 orbital inclination relative to the ecliptic plane). (D) Periodogram of the orbital inclination variation shown in (C). […]” — Cyclostratigraphy and Astrochronology – Linda Hinnov 2018

  9. Philip Mulholland says:

    Good digging, you’re mining a rich seam.

    δ18O is defined as the fractional change in 18O/16O in parts per thousand; its variations in the ocean are driven primarily by changes in the volume of global ice, which is depleted in 18O. We used two determinations for the depth of the samples. The first was to set the depth to the “driller’s depth” at the top of each core segment, assume no gaps between segments, and use linear interpolation in between. We call this the “adjusted depth”; there were no adjustable parameters in its determination. The second method was to adopt the scale developed by Berger et al. (1993), who argued that gaps were present between the core segments, and corrected for them by patching data from an adjacent core. The largest patch was for an assumed 75 cm gap near the depth 16 meters. We refer to this depth scale as the “patched depth.” The conclusions that we present in this paper are independent of which of these two depth scales we use.

    What is known in the geologic trade as “Gardening”. 🙂

  10. Ned Nikolov says:

    Trying to match frequencies in the sediment record to Milankovitch orbital frequencies and drawing “conclusions” from such matches is a dead-end exercise in terms of understanding the actual drivers of Ice Ages. Countless studies have run into such a dead end already. Yet, the Milankovitch myth continues to dominate scientific thinking suggesting an apparent lack of intelligence in the field of Paleoclimatology…

  11. oldmanK says:

    Part quote “δ18O is defined as the fractional change in 18O/16O in parts per thousand”.

    Taken from polar ice cores Vostok and Gisp2 agree, but are opposite to equatorial Kilimanjaro throughout the Holocene max. However all three increase abruptly at the end of the YD (definitely not a M causation).

  12. Philip Mulholland says:

    “but are opposite to equatorial Kilimanjaro throughout the Holocene max”
    Interesting point however we are there comparing a Hadley cell with the 2 Polar cells.
    Then there is the issue of vegetation water recycling (evapo-transpiration). Lots of vegetation in the tropics, zero for the polar icecaps.
    Too many confounding variables.

  13. Stephen Wilde says:

    ‘Many researchers have noticed over the years the inconsistencies between Milankovitch cycles and glacial fluctuations, but the “Milankovitch myth” continued to persist, because there was no other driver that science knew about.’

    I have seen it suggested that albedo variations are sufficient as a driver of ice ages since they mimic a change in distance from the sun. It is therefore untrue to suggest that science has not considered other drivers.
    Distance from the sun is a driver of the surface temperature enhancement along with atmospheric mass and the strength of the gravitational field (which together set surface pressure}. To that extent I agree with Ned.
    I find albedo variations pretty pursuasive because the phase change of water gives a potentially substantial and rapid response to Milankovitch cycles and the degree of response will be related to the configuration of the continents because ice forms more readily on land surfaces.
    If Ned wishes to insist on total global atmospheric mass variations as the only possible solution then I would like to hear of any mechanism that could change total atmospheric mass on the short timescale between glaciations and interglacials.

  14. I have been recording daily SOI figures (from the long paddock site and have noted a tidal effect related to the moon. As the tides at Tahiti are fairly constant the effect mainly relates to the tides at Darwin which gives a change to the atmospheric pressure.
    This paper “The 1,800-year oceanic tidal cycle: A possible cause of rapid climate change”
    Charles D. Keeling* and Timothy P. Whorf
    Scripps Institution of Oceanography, University of California, San Diego, La Jolla, CA 92093-0244
    Contributed by Charles D. Keeling, February 2, 2000; pp3814–3819 PNAS April 11, 2000 vol. 97 no. 8
    may interest Ned and those interested in the effect of the orbits of planets and the moon.
    The atmospheric pressures are related to oceans levels and ocean temperatures. The SOI is an indicator for ENSO. (El Nino/La Nina Southern Oscillation)

  15. Ned Nikolov says:


    I have made this point before, but here it is again: Albedo changes cannot account for the Ice Ages, because the role of albedo in controlling the global surface temperature is quite limited. The planetary albedo is not a free variable as many tend to think! Watch my albedo video presented at the 101st AMS conference in Jan of 2021 for a full quantitative explanation of how this works… Also, albedo changes cannot explain the observed polar amplifications during glacial cycles!

  16. Philip Mulholland says:

    I cannot agree with your proposition that albedo is a fixed property of a terrestrial body. This is demonstrably false by observation. For example, the albedo of Mars is higher during global dust storms. The Galilean satellites of Jupiter are vacuum worlds with albedo values that range from a dark 0.19 for Callisto (JIV) to a bright 0.68 for Europa (JII).

    I believe that you are confounding atmospheric albedo with surface albedo. The four Jovian Satellites clearly demonstrate a wide range of albedo for different surface types. The albedo of Earth is a composite of an atmospheric cloud veil albedo formed by the main condensing volatile, water (droplets and ice) and the surface visible properties, ocean, vegetated land, desert land and surface ice fields.

    Because planet Earth is a fast rotator the descending clear air zone of the tropical Hadley cell is located in the mid-latitudes. Depending on the surface properties these zones are either desert land or desert ocean. For desert ocean with its low albedo surface the daily solar energy is captured, stored and transported by ocean gyres to high latitudes. For desert land the daily solar insolation is not stored but is instead lost back to space each night by thermal radiant transmission through the clear sky atmospheric window of the descending limb.

    It is variations in the surface albedo that are related to the long-term geologic changes. The slow geologic change of ocean plumbing e.g., the continental rifting of the Southern Ocean and the barrier separation of compressional arc formation e.g., the Isthmus of Panama that caused major surface environment changes e.g., icecap formation in Antarctica that led to long slow changes in Earth’s total composite albedo.

    Using our DAET climate model I have established that for a constant pressure Earth, the short-term high frequency albedo changes will be from 0.306 for a 15 Celsius inter-glacial world to a bright 0.344 albedo for a 11 Celsius ice age world. This albedo change is caused by the 65-year climate cycle of path length changes in the reflective jet stream path sinuosity (short term), and associated surface elevation changes (long-term). This long-term change is caused by slow land ice build-up that expands the Polar cell, forces the Ferrel cell to adopt a lower latitude zone by ice elevation blocking and also generates the cold climate re-enforcing surface gravity winds e.g., Antarctica. This albedo climate change scenario is for a constant pressure Earth.

    Using our DAET climate model I have also established that for a constant albedo (your preferred model) 15 Celsius Earth, the long-term geologic pressure change back from our present low value of 1013 mb to a higher pressure of 1092 mb will generate a hot-house 24 Celsius Cretaceous world.

    Our range of values of long period global atmospheric pressure change derived from our DAET climate model is much smaller than those you suggest. I know that you will dismiss this as modelling and stories, but ultimately Geoscience relies on storytelling and modelling validation of data.

  17. Ned Nikolov says:


    As usual, you are jumping to conclusions before having actually understood the issue… I have never claimed that the “albedo is a fixed property of a terrestrial body“. Where did you get that crazy idea from?

    The role of planetary albedo in controlling surface temperatures is more complex than it meets the eye, and I tried to explain it in my video above. Please clear your mind of preconceived notions and watch the video 2-3 times while focusing on the the physics and numerical reasoning presented there…

  18. Philip Mulholland says:

    I have looked again at your video. I have a question – Did you study mean atmospheric molecular weight as one of your climate variables for Venus, Earth and Titan?

  19. Ned Nikolov says:


    I hope you understood our concepts of baseline albedo and albedo perturbations around the baseline, and why perturbations are necessarily small compared to the baseline albedo.

    Our model uses molar density of the atmosphere measured in mol m-3, but does not deal with the average molecular mass of the atmosphere. Why are you asking that question?

  20. Philip Mulholland says:

    I am interested in the relationship between molar mass, escape velocity and solar irradiance.

  21. Ned Nikolov says:

    I have not studied this problem yet…

Leave a Reply

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

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

Google photo

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

Twitter picture

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

Facebook photo

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

Connecting to %s