Beyond Milankovitch

by Donald Rapp

On the terminations of Ice Ages.

Terminations occur on solar up-lobes

There is no doubt that there is merit in the widely accepted Milankovitch theory that Ice Ages and their terminations are controlled by solar input to the NH in mid-summer. It is also clear that relying on the solar input to the NH alone, does not adequately account for the occurrence of terminations of Ice Ages. The variation of solar input to high latitudes is modulated by precession, which produces continual up-lobes and down-lobes in solar input with a ~ 22,000-year period. While every termination is accompanied by the 5,500-year rising portion of an up-lobe in the solar input to high latitudes, many strong up-lobes do not produce a termination.

Many solar up-lobes do not produce terminations

While essentially all terminations occur soon after a significant increase in solar intensity, there are many such increases in solar intensity that do not produce terminations, and the termination around 400 kya is not explainable at all by this argument. Figure 1 shows the relative peak summer solar intensity at 65°N vs. year, along with the occurrences of the various terminations. We may regard the pattern of variability of peak solar intensity in Figure 1as a rapidly varying (~22,000-year period) contribution due to precession of the equinoxes, with an envelope whose amplitude depends on the more slowly varying obliquity and eccentricity. In a sense, this is like an amplitude-modulated radio wave where a high-power “carrier wave” is modulated in amplitude by the signal (in our case eccentricity and obliquity). The key fact is:

There are many sharp up-lobes, but only a fraction of them are associated with terminations. Hence, a sharp up-lobe in solar intensity might be necessary for a termination, but it is certainly not sufficient.

Figure 1. Solar input to 65°N at midsummer shown along with the termination ramps from Rothlisberger et al.(2008).

Need for an X-factor to accompany solar up-lobes at termination

At least one other factor must be compounded into the overall model for why and when terminations occur. After reading many dozens of learned papers on terminations of Ice Ages, I have come to the tentative conclusion that over the past 800,000 years, the natural state of the Earth was that of what we call an “Ice Age”. Apparently, Ice Ages occurred because the energy balance of the Earth in pre-industrial times favored production of ice sheets in the North, and the ice sheets were enhanced when the solar input to the NH was unusually low. Even during periods with rising lobes in the pattern of solar input to the NH, Ice Ages persisted in the absence of another factor(s) that forced origination of a termination. After a termination, an Interglacial ensued. Apparently, this “X-factor(s)” diminished and eventually disappeared during an Interglacial. When the X-factor(s) was no longer operating, the gradual growth of ice sheets resumed from their depleted condition after a termination. The new Ice Age took many tens of thousands of years to mature, but ultimately, a new global maximum of the ice sheets was reached.

So, what we had was not, as one might tend to perhaps assume, unusual Ice Ages that interfered with natural periods of relative warmth. Instead, we had persistent Ice Ages that were intermittently terminated when the X-factor(s) arose, as an exception to the rule, rather than as a state of normalcy. The terminations were the exceptions.

Therefore, the search for the holy grail of Ice Ages is essentially the search for the X-factor(s) that causes terminations, along with rising solar input to the NH.

It is interesting that Wolff et al.(2016)understood the need for the X-factor(s). They said:

Milanković theory would suggest that terminations (and associated interglacial onset) should occur on the rising limb of NH summer insolation. As not every precessional cycle leads to an interglacial, there must be another factor that leads to Interglacials occurring in some precessional cycles and not others.

The “doom” blog (one of the more credible climate discussion blogs),[1]reviewed many theories and models of Ice Ages in detail and particularly ice sheet dynamics. This blog concluded:

… another ingredient is needed to explain the link between insolation and termination.

And what is the missing ingredient that turned the rise in northern insolation around 20,000 years ago into the starting gun for deglaciation, when higher insolation at earlier times failed to do so?

The termination of the last Ice Age is a fascinating topic that tests our ability to understand climate change.

On other climate blogs, writers and commenters seem very happy that climate scientists have written a paper that “supports the orbital hypothesis” without any critical examination of what the paper actually supports with evidence.

Returning to the question at hand, explaining the termination of the last Ice Age – the problem at the moment is less that there is no theory, and more that the wealth of data has not yet settled onto a clear chain of cause and effect. This is obviously essential to come up with a decent theory.

Solar cycle from glacial maximum to termination to Interglacial to new Ice Age

Ideally, the description of a termination is as shown in Figure 2. In this figure, the blue and red curves depict the noon solar intensity at 65°N on June 21 for the last two termination/Interglacial periods with superimposed time scales. In both cases, the longstanding Ice Age had been maturing for many tens of thousands of years. While the ice sheets grew, periodic up-lobes of solar input inhibited ice sheet growth for a time, while down-lobes enhanced ice sheet growth. But on balance, the ice sheets grew for many tens of thousands of years. Eventually, the Ice Age reached its ultimate depth of cold, CO2was depleted to less that 200 ppm, desertification generated large sources of dust, and winds transferred the dust to the ice sheets. This tipped the scales of the energy balance, and now the ice sheets rapidly ablated as the albedo decreased. The ~5,500-year up-lobe was enough to produce a termination. An Interglacial was produced. It lasted around ~5,500 years (or so). At this point, with resurgent plant life, dust sources dried up, and the solar curve turned to a down-lobe. A new Ice Age began from humble beginnings. In this idealistic picture, the total duration of termination ramp plus Interglacial is 11,000 years (give or take a few thousand). Note that the current Interglacial seems to be lasting too long to fit this simple picture. That remains difficult to explain, although soot, dirt, ash and dust generated by human activity over the last 500 years has dirtied the residual ice sheets to the point that no new Ice Age seems possible in the millennia to come.

Figure. 2. Idealized concept of transitions in Ice Ages

Duration of a termination

Reviewing all the data as far back as 8000,000 years, the durations required for the termination ramp from glacial to interglacial conditions in the last nine deglaciations are listed in Table 1. On average, the duration of the transition from glacial to deglacial conditions took roughly 6,000 years (about ¼ of a precessional period).

Table 1.  Durations required for the transition from glacial to interglacial conditions.

Variation of physical variables through the termination cycle

Data on the patterns of the transition from glacial maximum to termination ramp to Interglacial are shown in Figures 3 to 5.

Figure 3. Superposition of nine temperature curves around the last termination (WAIS, 2013).

Figure 4. Temperature change through the current Interglacial. (Masson-Delmotte et al.(2011))

 

Figure 5. Temperature change before and through the previous Interglacial. (Masson-Delmotte et al.(2011))

It has been found that sharply rising dust levels preceded termination in every case over the last 800,000 years. Data on the last two terminations are given in Figure 6.

Figure 6. Comparison of dust deposition rate and temperature change through the last and previous terminations. Dust data from Schneider et al.(2013). Temperature data from Masson-Delmotteet al.(2011)or WAIS (2013).

Dust deposition as a trigger to initiate terminations of Ice Ages

Ellis and Palmer (2016)emphasized the importance of dust deposition as a trigger to initiate terminations of Ice Ages.

As Ellis put it (private communication):

“Almost everyone agreed that Milankovitch cycles controlled the glacial cycle. But they were unable to explain why some cycles failed to produce an interglacial while others did, and during subsequent research it became apparent that there was no accepted answer to this troubling but central question. A theory is not a theory, if it has a thumping great lacuna in the middle of it. This led me into a detailed study of the glacial cycle, and the revelation dust was at a peak just before each interglacial.

It was only when Michael Palmer sought to refine my rough and rugged draft paper, that the prior research of Mahowald and Ganopolski and many others was discovered. And it was surprising that all of these papers danced around what I saw as the central agent of Ice Age modulation, without identifying and explaining it as such. Ganopolski, for instance, identified a link between ice sheet volume and dust and presumed that the volume of ice was causing the dust – in other words this must have been glaciogenic dust caused by ice-rock erosion. But previous papers had already identified the source of the dust as the Gobi and Taklamakan deserts, excluding the possibility that the dust was glaciogenic.”

Ellis and Palmer (2016)said:

“When CO2reaches a minimum and albedo reaches a maximum [at the peak of an Ice Age], the world rapidly warms into an interglacial. A similar effect can be seen at the peak of an interglacial, where high CO2and low albedo results in cooling. This counterintuitive response of the climate system also remains unexplained, and so a hitherto unaccounted-for agent must exist that is strong enough to counter and reverse the classical feedback mechanisms.”

They proposed:

The answer to both of these conundrums lies in glacial dust, which was deposited upon the ice sheets towards the end of each glacial maximum… during the glacial maximum, CO2depletion starves terrestrial plant life of a vital nutrient and causes a die-back of upland forests and savannahs, resulting in widespread desertification and soil erosion. The resulting dust storms deposit large amounts of dust upon the ice sheets and thereby reduce their albedo, allowing a much greater absorption of insolation.

They asserted that their proposal:

… explains each and every facet of the glacial cycle, and all of the many underlying mechanisms that control its periodicity and temperature excursions and limitations.

The paper by Ellis and Palmer then discussed variability of solar input to high latitudes, which is very heavily traveled ground, and we need not elaborate on this here. But one point they raised is worth emphasizing: One cannot invoke rising solar input to high latitudes as the sole cause of terminations of Ice Ages since many such increases in solar input do not produce terminations. Increased solar input might be necessary for terminations but it is clearly not sufficient.

The theory of Ellis and Palmer explains how terminations occur, but it does not necessarily explain why Interglacials terminate. Once an Interglacial is established, history shows that that a new Ice Age typically begins in less than ten thousand years. A likely explanation for this, as an addendum to the theory of Ellis and Palmer, is to note that during the course of an Interglacial, CO2rises rapidly toward about 280 ppm, the plant life of the Earth resuscitates, and dust levels drop precipitously. While the ice sheets are greatly reduced from the levels of the Ice Age, substantial amounts of remnant snow and ice remain at high latitudes. Lacking dust on these ice sheets, the energy balance favors expansion of the ice sheets. The 11,000-year up-lobe in solar input to high altitudes that coincided with the termination and Interglacial now turns sharply downward, thus supporting expansion of the ice sheets as the outcome of the Interglacial. A nascent new Ice Age begins. This Ice Age continues for up to about 100,000 years until the CO2concentration drops so far that plant life is severely affected, dust becomes prevalent, and the cycle continues.

Measured dust levels in Antarctic ice cores are summarized in Figures 7 and 8, using the Coulter method and the laser method.

Figure 7. Antarctic dust concentration in ice cores as measured by the Coulter method. (Lambert, et al.(2008))

Figure 8. Antarctic dust concentration in ice cores as measured by a laser optical method. (Lambert, et al.(2008))

The essential basis for the hypothesis advanced by Ellis and Palmer is illustrated in Figures 9 and 10. The lower graph in these figures shows the dust loading in the ice core at Antarctica. The middle graph shows peak solar intensity at 65°N on June 21. The top graph shows the Antarctic temperature anomaly. Vertical red dashed lines are drawn at each major sharp rise in Antarctic temperature, presumably initiating a termination.

From the data, we immediately observe:

  • Sharp dust peaks and/or dust accumulations immediately precede all terminations.
  • Up-swings of the oscillatory solar intensity occur at the start of each termination.

The inevitable conclusion must be:

Terminations are essentially always preceded by a buildup of dust and mainly occur on a sharply rising lobe of the solar oscillation.

Therefore, the inference made by Ellis and Palmer is that two situations are necessary precursors to a termination of an Ice Age:

(1) There must be a sharp maximum in dust loading. (But this typically occurs after there has been a long, extended Ice Age).

(2) The sharp dust maximum must coincide with sharply rising solar intensity.

Note particularly that sharply rising solar input to high latitudes by itself does not necessarily lead to a termination. Many such high amplitude lobes of solar input to high latitudes do not lead to termination. Termination only occurs with rising solar input to high latitudes, after reaching the depth of the greatest CO2reduction in a very mature Ice Age, combined with peak dust. Most high amplitude lobes of solar input to high latitudes have only moderate effect on the growth of ice sheets. Therefore, one might infer that peak dust is even more important than solar input to high latitudes in termination of Ice Ages. While these inferences do not in themselves prove a cause–effect relationship, they are highly suggestive.

Figure 9. (top) Antarctic temperature. (middle) Solar intensity at 65°N on June 21. (bottom) Dust loading in Antarctica ice core using the Coulter counter.

Figure 10. (top) Antarctic temperature. (middle) Solar intensity at 65°N on June 21. (bottom) Dust loading in Antarctica ice core using the laser counter.

Dust fluxes at LGM and solar absorption

Lambert et al.(2015) presented a new global dust flux data set for Holocene and LGM based on observational data. They created a new interpolation of the unevenly spaced Holocene and LGM dust flux measurements compiled in the DIRTMAP database to two global grids. As in most studies, they were primarily interested in the impact of enhanced glacial oceanic iron deposition on the LGM-Holocene carbon cycle, rather than dust deposition on ice sheets. The only place where any information was provided for deposition at Greenland or North America during the LGM was in a small color-coded map with rather crude resolution. A rough reading of this map indicated the rates of deposition of dust as follows:

Greenland (core drilling region)        0.3 g/m2/yr

Greenland (southern tip)             0.8 g/m2/yr

Great Lakes area:                           32 g/m2/yr

Laurentide ice sheet:                    12 g/m2/yr

Maher et al.(2010)and Lambert et al.(2015) made estimates of dust deposition rates for the Laurentide ice sheet and the Great Lakes area at the LGM. An average of these values is:

Great Lakes area:                        20 g/m2/yr

Laurentide ice sheet:                   7 g/m2/yr

Following Mahowaldet al.(1999), it is assumed that the average diameter of a dust particle is 2.5 microns. The geometrical blocking area of a dust particle is therefore estimated to be

3.14 ´(1.25 ´10-6)2m2~ 5 ´10-12m2.

And the mass of a dust particle is estimated to be

4/3 ´3.14 ´(1.25 ´10-6)3m3´1.5 ´106 g/m3= 1.2 ´10-11 g

Based on purely geometrical effects, the geometrical blocking area of surface dust is estimated to be:

5 ´10-12m2/ 1.2 ´10-11 g = 0.4 m2of blocking area per g/m2of dust

Hence in the geometrical approximation, a layer of 2.5 g/m2of surface dust completely blocks a surface optically. A layer of 1 g/m2of surface dust blocks 40% of the surface.

The above estimates are based on geometry, treating a dust particle as a sphere of diameter 2.5 microns. However, the effective cross-sectional area of a particle for light scattering and absorption, Asis related to the geometrical area of a particle, Ad, by the relation:

As= AdQext

where Qextis the so-called extinction efficiency or scattering efficiency. It turns out that because of diffraction effects, Qextis >1. In analyzing dust on Mars, Tomasko (1999)estimated Qext~ 2.6. Thus, the true optical effect of 1 g/m2of dust particles is to optically block 100% of the surface. Mars dust is roughly similar to Earth dust in regard to average diameter.[2]

When light interacts with a single particle, some light can be scattered into the forward cone, some light can be scattered into the backward cone, and some light can be absorbed. According to studies of Mars dust, the best estimate is that of the total amount of light scattered and absorbed, 7% is absorbed, 77% is forward-scattered, and 16% is back-scattered. Note that most of the scattering is in the forward direction. Thus, the predominant effect of dust particles embedded in ice is to forward scatter the light deeper into the ice and to a lesser extent, absorb the light.

Source of the dust

Ellis and Palmer (2016)reviewed at length, data and models on the source of dust and albedo effects of dust during Ice Ages. I will be content with only a very brief mention of only a few points they raised. There seems little doubt that the combination of low temperature and low CO2at the LGM was detrimental to plant life. They argued that the principal impact was on high altitude regions, normally arid regions, and northern regions during summer. The source determined by isotopic analysis was attributed to the Gobi and Taklamakan deserts, and there was little evidence of glaciogenic dust. As Ellis put it (private communication):

The question is how these peaks in dust flux were generated during the latter millennia of each Ice Age. If it was not the grinding of ice sheets that caused the dust, then what did? The likely answer was that the low CO2concentrations at each glacial maximum led to plant extinction in high altitude arid regions, turning them into new CO2deserts. And this was a scenario that fitted the source region for Greenland dust very well – the Gobi. The Gobi is mostly steppe grassland, but under the low CO2conditions of a glacial maximum the entire Gobi became a shifting-sand desert that created vast dust clouds, as is proven by the massive dust deposits upon the Loess Plateau in China.

And so, the beauty of this theory … is that it is a simple thought-experiment that can be followed by anyone, and yet in my view it explains every facet of the glacial-interglacial cycle. And it even explains the previously inexplicable – the reason why some strong precessional insolation maxima failed to produce much in the way of melting and warming, let alone an Interglacial.

In short:

Ice Ages cause ice sheet extension Þoceanic cooling Þoceanic CO2absorption = plant asphyxiation on the Gobi plateau Þnew CO2deserts Þdust generation Þice sheet contamination for 10 kyrs.

Then:

Rising NH Milankovitch insolation Þice sheet surface ablation and melting Þice sheet dust exposure and concentration Þice sheet albedo reduction Þincreased insolation absorption Þincreased ablation and melting ÞInterglacial.

It is a simple feedback system that is very powerful and operates regionally, unlike CO2which is a global feedback agent of indeterminate strength. And yet we know that Interglacials are regional phenomena rather than global phenomena, because they only coincide with increased Milankovitch insolation in the northern hemisphere, and never with increased insolation in the south. Ergo, the primary feedback controlling interglacial inception must be regional to the NH, rather than global.

Summary

I think that Ellis makes some powerful arguments here; yet uncertainties remain about how much dust was deposited at the LGM and how effective that dust was in increasing the absorptivity of the ice sheets. The Antarctic data support the hypothesis but Antarctica is a long way from the ice sheets. The Greenland data are more enigmatic. It is widely agreed that dust deposition on the ice sheets was much greater than it was at Greenland or Antarctica, but so was the rate of snow accumulation.

The hypothesis put forth by Ellis and Palmer has considerable potential merit. As the Ice Age deepens, the combination of low CO2levels, cold, and wind turns originally marginal, semi-desert regions into major sources of wind-blown dust. The evidence from ice cores indicates that the LGM was a period with relatively high levels of atmospheric dust blown by strong winds. Some models estimate relatively high levels of dust deposition on the ice sheets at the LGM. These levels of dust deposition would significantly darken the upper levels of the ice sheets. Thus, as Ellis and Palmer suggested, dust levels on the ice sheets peaked prior to and at the LGM, providing a mechanism to initiate a termination when synchronized with an up-lobe in summer solar input to high northern latitudes. The role of dust in terminations of Ice Ages is probably far more important than many realize. It appears likely that unusually high dust levels coupled to sharply rising solar intensity at high latitudes was a major factor in initiating termination of Ice Ages. However, as always in such speculations, uncertainties remain.

Finally, it is noteworthy that because Ellis and Palmer (2016)was published in an obscure journal, it seems to have been somewhat ignored by the science community. My hope is that my book will give their theory additional exposure.

Bonehead model for Ice Ages:

In the following, I propose a modification to the Paillard (1998)model.

I hypothesize four states of the Earth system:

I= interglacial

g= mild glacial

G= glacial maximum

U= termination

Each of these is characterized by an ice volume parameter vRand a time constant for ice volume change, TR, where subscript Rcan be I, g, Gor U. At any time t, v= the current ice volume and dv/dt = rate of change of ice volume.

There is a constantly acting solar forcing term F(t) with amplitude that oscillates with the precession frequency. The time starts at some negative value (years) and works its way forward via integration of differential equations. We can gain insight into the magnitudes of the vRas follows.

The basic equation is

dv/dt = (vR– v)/TR– F(t)

in which the rate of change of ice volume in any state of the system is a maximum when v is small and is a minimum when vis large. The solar forcing term F(t) oscillates with a ~22,000 year period due to precession. F(t) is measured from its average, so it can be negative or positive.

When the first term on the right side is large (v<< vR), the first term outweighs F(t) and the ice sheets are growing. As F(t) oscillates, it can add or subtract from the first term’s contribution to dv/dt, but dv/dt remains substantial. This is the “g” state.

After passing through the “g” state for about four periods of precession (roughly 90,000 years), vreaches a level that we characterize as the “LGM” or in this model, the “G” state. In the “G” state, the first term becomes small compared to F(t). As long as F(t) remains in a down-lobe (about 11,000 years) the “G” state persists. But when F(t) enters an up-lobe, dv/dt turns negative and the system enters the “U” state – termination. The “U” state persists for about half of the up-lobe, and when most of the ice is gone (in about 5,500 years), the system enters the “I” state – an Interglacial – during which the ice volume remains low but does not change. The Interglacial lasts another 5,500 years as the solar up-lobe diminishes. Finally, F(t) enters a down-lobe and dv/dt turns positive and a new “g” state begins.

Figure 11shows a simplistic picture of the variation of vand dv/dtwith t. In the plot of dv/dt vs. t, the area under the curve in the “g” region must equal the area of the rectangle in the “U” region (to conserve mass). Figure 11does not show the effect of variation of insolation. In region “g”, up-lobes and down-lobes of solar due to precession add ups and downs to the smooth curve shown.

Figure 11. Schematic variation of vand dv/dt vs. t.

Consider a system at t= 0 early in the Interglacial state. We require that v< vIand F(t)is large enough that dv/dtis slightly less than 0.

dv/dt = (vI– v)/TI– F(t)is slightly less than zero at start of “I” state with F(t)on an up-lobe of precession.

Furthermore, the magnitudes must be chosen such that when F(t)heads downward due to precession, a point will be reached where dv/dtturns positive. At that point, the system switches to the “g” state.

Now the system switches to a new differential equation:

dv/dt = (vg– v)/Tg– F(t)

Here, vgis large, so that the rate of buildup of ice volume remains > 0 throughout. However, as v builds up, the rate of accumulations decreases. Furthermore, the ups and downs of F(t)due to precession adds up and down perturbations to the progress of the ever increasing ice volume.

At some point in time, when vgets large enough, the system switches to the “G” state. It is not clear a priori how to set this transition. Perhaps when v reaches say, 90% of vg? At such a point, the rate of accumulation has slowed down considerably and the ice volume is large. In the “G” state,

dv/dt = (vG– v)/TG– F(t)

The parameters must be chosen so that dv/dt ~ 0 for moderate F(t). At some point, an up-lobe in the precession cycle occurs, and F(t) increases. When dv/dtturns positive, the system enters the “U” state. In the “U” state:

dv/dt = (vU– v)/TU– F(t)

Here, dv/dt << 0 and the system undergoes a rapid termination. When v reaches a critically small level, the system enters the “I” state and the cycle begins over again. Determining appropriate values (and units) of the parameters is likely to be a tricky business. It seems likely that vU~ vG~ vgare large whereas vIis small. Clearly, TUmust be the smallest time parameter.

Figure 12illustrates the theory further. Note that in actuality, the amplitude of the solar oscillations varies from cycle to cycle but they are drawn in the figure (for simplicity) as if the amplitude never changes. Figures 11 and 12 are highly idealized, and each glacial-interglacial cycle will have its own unique character. Ups and downs (as illustrated by the blue curves) might be quite different in reality from this idealized picture of events.

Figure 12. Schematic representation of (dv/dt) in the transitions to and from Ice Ages.

In this model, there are four periods within the Ice Age cycles. The “g” period (white background) involves long-term buildup of the ice sheets over many decades. In this glacial period, the rate of accumulation decreases with time. After about 4 solar precessional cycles (~88,000 years), the ice sheets approach their greatest extent. The system then enters the “G” state – a global ice maximum (brown background). This global glacial maximum lasts for about 11,000 years, and occurs during a solar down-lobe, and the CO2concentration drops to its lowest level (less than 200 ppm). This impacts plant life, and heavy dust deposition occurs on the ice sheets. With the next solar up-lobe, the combination of heavy dusting and rising solar intensity produces a termination (“U” state – green background). The ice sheets dissipate in a mere ~5,500 years. As termination proceeds, the CO2concentration rises, plant life recovers, and dust levels drop precipitously. When the ice sheets reach a minimum, the system enters the Interglacial state (“I”) (blue background) which lasts another ~5,500 years, completing the solar up-lobe. The Interglacial state has a climate not unlike that of today. When the solar curve turns downward, the value of dv/dt turns positive and a new Ice Age begins (“g” state begins anew). The total length of the “U” and “I” states is ~11,000 years.

The black curve in Figure 12shows a simplified dv/dt (where v is the total ice volume) neglecting solar variations. The solar cycle relentlessly oscillates with the ~22,000 period due to precession. The blue curve shows the effect of solar perturbations on dv/dt. Each up-lobe and down-lobe produces a corresponding “bump” in the dv/dt curve but a termination cannot ensue until the “G” state, when dust deposition decreases the albedo of the ice sheets.

Figure 13shows the corresponding curves for ice volume v(t). In this model, the variability of the solar input to high latitudes is inconsequential during the long-lasting “g” state. In the “g” state, dv/dt is dominated by the first term on the right side of the equation. The Solar input comes into play only near the end of an Ice Age. The amplitude of the solar oscillations (due to precession) is not important; only the phasing matters. A down-lobe in solar late in the Ice Age produces a glacial maximum. During that time, CO2sinks to its lowest values and large amounts of dust are generated and transported by winds. As the precessional cycle turns upward, a termination sets in. An Interglacial follows until precession drives the solar curve downward. Solar precession acts as the pacemaker for transitions from gto Gto Uto Ito g.

Figure 13. Schematic representation of v(t)in the transitions to and from Ice Ages.

While the models as shown in Figures 12 and 13are very satisfying in their simplicity, nature is not so cooperative. This simple picture does not adequately describe the reality in detail, although it does have some characteristics that resemble reality in very rough outline.

Addendum

This write-up is abstracted from chapters 8 and 11 of the third edition of my book on Ice Ages which is now in press and might be available in early 2019. I am willing to send pdf versions of the chapters to serious scientists on request.

References

Ellis, Ralph and Michael Palmer (2016) “Modulation of Ice Ages via precession and dust-albedo feedbacks” Geoscience Frontiershttp://www.sciencedirect.com/science/article/pii/S1674987116300305

Lambeck, Kurt et al.(2014) “Sea level and global ice volumes from the Last Glacial Maximum to the Holocene” PNAS111, 15296–15303.

Lambert, F. et al.(2008) “Dust-climate couplings over the past 800,000 years from the EPICA Dome C ice core” Nature452, 616-619

Lambert, Fabrice (2015) “Dust fluxes and iron fertilization in Holocene and Last Glacial Maximum climates” Geophys. Res. Lett.,42, 1-10.

Maher, Barbara et al.(2010) “Global connections between Aeolian dust, climate and ocean biogeochemistry at the present day and at the last glacial maximum” Quaternary Science Reviews99, 61-97.

Mahowald, N.et al.(1999) “Dust sources and deposition during the last glacial maximum and current climate: A comparison of model results with paleodata from ice cores and marine sediments” JGR104, 15895-15916.

Masson-Delmotteet al.(2011) “A comparison of the present and last interglacial periodsin six Antarctic ice cores” Clim. Past7, 397–423.

Paillard, Didier (1998) “The timing of Pleistocene glaciations from a simple multiple-state climate model,” Nature 391, 378-381.

Rothlisberger, R., M. et al.(2008) ‘‘The southern hemisphere at glacial terminations: Insights from the Dome C ice core,’’ Climate of the Past Discussions, 4, 761–789.

WAIS (2013) “Onset of deglacial warming in West Antarctica driven by local orbital forcing”Nature500, 440-446.

Wolff, E. W.et al.(2016) “Interglacials of the past 800,000 years” Past Interglacials Working Group of Interglacials of the last 800,000 years, Reviews of Geophysics54, 162–219.

[1]https://scienceofdoom.com/2014/04/14/ghosts-of-climates-past-nineteen-ice-sheet-models-i/

[2]Rapp, D. Human missions to Mars, 2ndedition, Springer-Verlag, 2016.

Moderation note:  As with all guest posts, please keep your comments civil and relevant.

375 responses to “Beyond Milankovitch

  1. This is superb writeup. Thank you.

  2. Roy W. Spencer

    Very nice!

  3. In a sense, this is like an amplitude-modulated radio wave where a high-power “carrier wave” is modulated in amplitude by the signal (in our case eccentricity and obliquity).

    I think you can see that rather clearly here:

    • Clive.

      On your graph the grey and blue plots represent a combination of both precession and obliquity. However, the precession element is in opposition in the N and S hemispheres, while the obliquity element fluctuates simultaneously. So when summing Milankovitch insolation at both poles, precession cancels out, leaving just obliquity – so the red plot represents the 41 ky periodicity of obliquity.

      P.S. Thanks for your marvelous graph in the paper.

      Ralph Ellis

  4. I believe much of this is simply overthinking the dominant control of water vapour from the oceans. By the time interglacials reach their peak temperatures, after 7Ka of steady warming, the dominant effect of water vapour from the warming oceans has become cloud formation, which provides 90W/M^2 of ocean cooling by evaporation, cloud formation and precipitation, and 50W/m^2 from albedo. With a lot of control in reserve, as the satellite images show.

    This 140W/m^2 is not static, it rises inexorably with warming oceans, and is plenty to shut down the relatively small interglacial warming effect by reducing the incomimg solar insolation of 340W/M^2 insolation to match, by the fractions of a W/M^2 the effect actually is, c.0.001 deg pa over 7Ka, but steady over that 7Ka and beyond, also overpowering any small CO2 effect, from warming oceans and hom sap, which was still burgeoning when cloud control ended the last interglacial warming, crushing any small CO2 effect as the rains finish off the ice remnants.

    The Milankovitch cycle effects ARE important to the initiation of interglacials, but wholly insignificant wrt the strong cloud control that ends the warming effect they create, in which the planet needs no help. That response is automatically delivered by the warming oceans, in the dominant control of increasing evaporation of water vapour as ever increasing cooling cloud cover in the smart planetary lagging of the atmosphere, water vapour that also offers warming the opposite warming feedback from GHE at the colder, less humid, blue sky climate of the ice age temperature floor. Simples!

    Hence the narrow temperature range of the cycle, closely controlled by water vapour’s modification of solar insolation, at a few degrees within 300 degrees K. This is my theory, which is mine, as Anne Elk once said. Isn’t it a lesson in the blindingly obvious? Discuss.

    • I think this is incorrect, and the evidence for this comes from the obliquity-led interglacials MIS-11 400 ky ago and the Holocene now. These interglacials remained in their warm mode for some 22 ky and 12 kyr, simply because obliquity insolation remained high and there was no precessional NH winter (no Great Winter) to force NH cooling. Water vapour and clouds were unable to do anything about it, and temperatures remained high as long as there was no NH Great winter (which is normally generated by precession).

      Clouds do have an influence on temperatures, as Willis Eschenbach has demonstrated with his Thunderstorm Thermostat theory, but those influences are smaller and are concentrated around the tropics instead of at the NH pole.

      The Thunderstorm Thermostat.
      https://www.researchgate.net/publication/249883480_The_thunderstorm_thermostat_hypothesis_How_clouds_and_thunderstorms_control_the_Earth%27s_temperature

      Image:
      How the obliquity-led Holocene temperature has followed obliquity.
      The Holocene is obliquity-led, because eccentricity is low at present, and therefore precession is weak. So the Holocene is an obliquity-led interglacial, just like MIS-11, and will turn into glacial condition shortly.

      • You have to put W/M^2 on your unquantified assertion for it to be credible.. The study I have made of this shows very clearly that the effect of the oceans on climate through water vapour is 100 times greater than the level of effects you are guessing may have something to do with interglacial warming (NOT the end of the interglacial BTW). Show me some numbers that remotely compare with the power of current water vapour insolation reduction effects of c.140W/m^2, that you can support with research by more than one person? References required. Thanks

      • Surface albedo changes can provide a 300 W/m2 change in insolation absorption – far greater than you suggest for water vapour. The albedo difference between fresh snow and bare earth can be up to 0.75.

        BTW – how do you get a 140 W/m2 increase from water vapour? We know that latent heat takes a lot of energy from the lower atmosphere, but an INCREASE of 140 W/m2 over the standard radiation budget seems excessive. Please provide a revised radiation budget that demonstrates this increase, between ice age conditions and interglacial conditions.

        Thanks, Ralph.

  5. Simple and complete though I intuitively want to integrate volcanism, both as a source of dust and GHGs, into the relationship.

    From a climate policy standpoint, it would seem that land use should be given higher priority in climate action.

    • Unfortunately for the vulcanism theory, it does not seem to fit the data. Vulcanism does not fit the periodicity of the Milankovitch cycle, while the interglacial cycle does. And if the dust-ice-albedo theory is correct, the dust that precipitated each interglacial was from the Gobi desert, rather than being volcanic.

      Ralph Ellis

  6. Dimitris Poulos

    Since I made the explanation of the solar wind phenomenon and theoretically calculated all it’s properties it is clear that the change between glacial ages is because of the solar wind pause. Since solar wind is the result of planetary resonance to produce this infinite resonant oscillation, it is normal that glacials correlate to Milankovic cycles as do planetary orbits and their resonance. I have the collection of all my papers in my blog dimispoulos.wixsite.com/dimis and till now quite a few people have read them in ResearchGate and I have some very positive feedback.

  7. Dimitris Poulos

    here is the link to my paper that proves that ice ages are the result of solar wind pause http://www.gpcpublishing.org/index.php/gjp/article/view/443

  8. Salvatore del Prete

    Do not by into the dust theory much.

  9. Salvatore del Prete

    buy

  10. The paper by Ellis and Palmer can be found here. There is a feee pdf download too.

    Modulation of Ice Ages via dust and albedo
    https://www.sciencedirect.com/science/article/pii/S1674987116300305

    Ralph Ellis

  11. Dimitris Poulos

    here is the link to ResearchGate with a collection of my papers too that prove between others that ice ages are the result of solar wind pause https://www.researchgate.net/profile/Dimitris_Poulos

  12. I’m a lay person, so I’m sure I’m not understanding.
    1. Solar changes are necessary but not sufficient to explain terminations of ice ages?
    2. Changes in dust are always necessary–more reliably associated with terminations than solar changes?
    3. Does the termination at about 400 kya , where there was no solar change, suggest that changes in dust are both necessary and sufficient?
    4. Some of your early sentences: the ice state is normal, it is terminations of an ice state that require explanation? It is (mainly) dust that explains how the earth sometimes moves away from the normal ice state, and then moves back again?

    • Points 1 and 2 – yes and yes.
      Point 3 – there WAS a solar change 400 ky ago, but it was smaller because it was led by obliquity not precession. There is a trade-off between these two solar influences – precession is short (say 8 kyr summer) and strong and only operates in one hemisphere – while obliquity is long (say 20 kyr) and muted, and operates in both hemispheres simultaneously. Obliquity also operates for the entire annual summer, whereas precession can only be at a maximum in the annual summer and either spring or autumn, but not for both. So the total insolation delivered by the weaker obliquity can be significant.

      It is apparent that obliquity on its own CAN initiate an interglacial, without much precession. Our current interglacial is also an obliquity led interglacial (precession reduces every 400 ky due to reduces eccentricity).

      Ralph

  13. Does the dust theory say anything about the transition problem?

    In fact, from 1–3 million years ago, climate cycles did match the 41,000-year cycle in obliquity. After 1 million years ago, this switched to the 100,000-year cycle matching eccentricity. The transition problem refers to the need to explain what changed 1 million years ago.
    https://en.wikipedia.org/wiki/Milankovitch_cycles#Transition_problem

    • Ha, ha, now that’s a tricky question.
      Trust Oldbrew to come up with that one…. ;-)

      The MPT is a tricky problem that no current theory fully explains. On the surface it seems clear that the climate (the oceans) were gently cooling for millions of years and then passed some kind of transition point (the oft-cited tipping point) some 800 ky ago. And this changed the interglacial cycle from an obliquity-led 41 kyr cycle, to a predominantly precession-led 90 ky or 115 ky cycle.

      As an initial response, we might surmise that during the weak ice ages, prior to the MPT, the ice-sheet extent and thickness was so small that obliquity was able to clear the ice sheets every 41 ky. Simple. Ah, yes, but why did precession not clear the ice-sheets every 22 ky? Precession is normally stronger than obliquity, and should have easily cleared the ice sheets. Ah yes – problem.

      This is work in progress – but I think the answer lies in the limited extent of the ice-sheets prior to the MPT. Perhaps obliquity can operate more effectively in the far north than precession can. Not sure if the math fully substantiates this, but it is true that a high obliquity does indeed present more of the northern hemisphere towards the sun – while precession can only place the earth closer to the sun during the annual summer season. So perhaps obliquity can indeed have a greater effect on ice sheets that are limited to the far north.

      After the MPT, the climate was now so cold that the ice sheets extended rapidly during just one 5 ky precession-led Great Winter. And the resulting NH ice sheet was SO large, with such a high albedo, after just 5 ky, that the next obliquity maxima or precession maxima were unable to melt it. Thus the ice age continued to deepen and extend throughout many Milankovitch maxima – until the ice sheets were finally covered in dust and their albedo reduced.

      Any thoughts gratefully received and considered.

      Ralph

  14. A nice essay and I agree with 70% of it !

    The basic metronome over the last 5 million years has always been obliquity. Obliquity changes the total annual solar energy received at both polar regions simultaneously. The Arctic and Antarctic warm and cool in synchrony. Eccentricity simply modulates the precession of the equinoxes (distance to perihelion), which then give an additive assymetric summer boost to obliquity in the Arctic (grand summer maximum).

    800k years ago the dynamics of ice ages changed as ice sheets grew too large. Yet even now when eccentricity is high the obliquity timing can still reasserts itself, because the extra boost in summer insolation from precession alone is sufficient to trigger an interglacial. An example is this is the jagged glaciation 600,000 years ago. I think it is also why the Epica interglacial warmed early with dust levels actually falling.

    When eccentricity is low however, the dust albedo effect exactly as proposed by Ellis, and described here, is the missing X-factor needed to trigger an interglacial. I fully agree!

    When all else fails Gaia terminates an ice age.

    Note that for at least the next 100,000 years we have low eccentricity!

    • What is for sure is that whenever CO2 falls below 210ppm dust accumulates on ice sheets increasing albedo. The only data we have are those from ice cores in Antarctica and Greenland which didn’t melt. The dust that really matters now lays in deep sediments or under the Great Lakes.

      I wonder if anyone is looking for it?

      • Indeed. However, we do have dust records from the source of that dust, as deposited upon the Loess Plateau in China. They have over a million years of dust layers there. Ok this is the source of the dust, rather than the actual data from the Laurentide ice sheet, but it will give further clues to the extent of the ice age dust-flux.

        Ralph Ellis

      • That is a very good point. Indeed, all your points have been well taken. There are models out there (many references in my book) that estimate that dust levels were much higher (factors of 5 to 20 or more) at the ice sheets than at Greenland, and much higher at the southern edge of the ice sheets than further north. These models are highly approximate but if taken at face value, and combined with the estimated blocking area of a dust particle, it would imply that at the LGM, the lower ice sheets were blanketed with dust. Some scientists have explored the margins of the remnant ice sheets in Greenland where old ice from the LGM is exposed but unfortunately, this source of old dust data seems to be corrupted by recent deposition of ash, soot, dust and dirt. Again, references are given in my book.

    • >>When eccentricity is low however, the dust albedo
      >>effect exactly as proposed by Ellis, and described
      >>here, is the missing X-factor needed to trigger an
      >>interglacial.

      But there are dust increases before every interglacial, not just the low eccentricity ones. So it is likely that EVERY interglacial requires a dust-albedo effect.

      Remember that whether you support the idea of obliquity-led interglacial initiation, or precession-led interglacial initiation, a number of those orbital cycles are missing from the interglacial record. For example, the obliquity maxima 90 ky and 50 ky ago did not create an interglacial.

      Why?

      Because there was no dust or insufficient dust on the ice sheets to absorb that extra insolation. There is no point increasing NH insolation, if 85% of it is being reflected and rejected by high albedo ice sheets. And so the ice age continued to grind on to even deeper levels, until the conditions for an interglacial were just right. And those conditions included at least 10 ky of dust, plus a large increase in NH insolation provided by either precession or obliquity. Now the increased insolation could get a grip on those ice sheets, because of their lower albedo, and the ice could be ablated and melted.

      Cheers,
      Ralph

      • Ralf,

        There are two examples where dust does not play a role, both of which are coincident with maxima in eccentricity – 200k years ago and 600k years ago. The cycle almost reverts back to 41k interval interglacials. The reason for this is the boost in summer insolation from the earth being closer to the sun at perihelion. This was enough to melt back ice sheets without a large increase in albedo.

      • Well there were small increases in dust just before those semi-interglacials, but obviously not enough to create a full interglacial. Do remember that these dust records come from Antarctica, while the actual warming was caused by a NH Great Summer (Milankovitch maximum) warming NH ice-sheets.

        It may well be that there was greater dust in the NH than was recorded in Antarctica. I will see if I can find a Loess Plateau dust record that will give further information on this.

        Cheers,
        Ralph

      • >>Dust does not play a role – 200k years ago and 600k years ago.

        Donald Rapp’s graphs do show small dust peaks just before those smaller proto-interglacials. See his figs 9 and 10. Again this is Antarctic data, and the Arctic may have received more Gobi dust than this.

        Unfortunately we do not have more Arctic data, unless the Loess Plateau can provide more information. But I am struggling to find a decent record of Loess dust that is pertinent to this research. Lots of graphs giving proxy indications of temperature, but not dust flux.

        Ralph

    • Excerpt from my book: Tzedakis et al. (2012) attempted to fill the need for a theory of Interglacials. They pointed out once again, that there were significant differences in the duration of Interglacials in the past, and that “a theoretical framework with predictive power for interglacial duration has remained elusive”. The duration of an Interglacial is related to the period of high sea level, and this can be interpreted in various ways. They proposed “that the interval between the terminal oscillation of the bipolar seesaw and three thousand years (kyr) before its first major reactivation provides an estimate that approximates the length of the sea-level high-stand, a measure of interglacial duration.” They went on to elaborate at length on this approach, but I can’t make much sense of the methodology. However, as I pointed out several times, using sea level as the measure of duration of an Interglacial will yield longer durations than using a measure based on the resurgence of accumulation of ice at high altitudes.

  15. “Ice age CO2 reductions coincide with an increase in ice sheet extent and therefore an increase in global albedo, and this should result in further cooling of the climate. But what actually happens is that when CO2 reaches a minimum and albedo reaches a maximum, the world rapidly warms into an interglacial. A similar effect can be seen at the peak of an interglacial, where high CO2 and low albedo results in cooling. This counterintuitive response of the climate system also remains unexplained, and so a hitherto unaccounted for agent must exist that is strong enough to counter and reverse the classical feedback mechanisms.” Ellis and Palmer 2016

    Although the dust mechanism appears robust – much more germane is what happens at the other end of the cycle. Why when greenhouse gases are abundant and ice volume is at its lowest is there rapid cooling from runaway ice sheet feedbacks? There are memes that say it can’t happen. No great NH winter for 50,000 years at least – or that anthropogenic greenhouse gases mean that statistical stationarity is dead – it’s a whole new world. Both ideas are no more than shibboleths. We are in a great NH winter – albeit not as deep as some – and I am reserving judgement on the end of history.

    https://upload.wikimedia.org/wikipedia/commons/thumb/5/53/MilankovitchCyclesOrbitandCores.png/861px-MilankovitchCyclesOrbitandCores.png – borrowed from Wikipedia.

    It has been suggested that AMOC is involved in glacial inceptions. That this may involve the AO/NAO – that in turn are modulated by solar UV/ozone chemistry. Whither to with a dimming sun?


    Fig 7 Smeed et al 2014

    Then there is the mysterious anti-phase relationship of heat storage at the poles. Could a Bond event trigger runaway ice? There are many unanswered questions.

    • Interglacials always terminate at the next obliquity minimum. The Holocene would naturally terminate within 15k years from now. Maybe our ancestors by then will be thanking us for delaying it through CO2 emissions.

      • I think the jury is out on whether precession or obliquity is the primary cause of cooling and the next glacial era. As per the diagram below, precession and obliquity have been in synch most of the last 400 ky, so it is difficult to tell which is dominant.

        … Certainly the glacial era after MIS-11 was led by obliquity, because there was no precession (low eccentricity). The precessional NH Great Winter ~420 ky ago was so weak you can hardly see it, and so the interglacial extended right to the end of the obliquity cycle.
        … MIS-9 seems to have slipped into glacial conditions long before obliquity fell to a minimum, so this appears to have been a precessional glacial era.
        … The MIS-7 interglacial appears to have been triggered by precession, although the subsequent glacial era again had precession and obliquity falling together.
        … The initial fall and rise of MIS-6 was certainly triggered by precession, as obliquity was sitting at a maximum during this time.
        … MIS-5 interglacial looks to be precession led, but precession and obliquity once again fell together.

        Upper plot: Orange = Milankovitch insolation (precession and obliquity combined).
        Middle plot: Black = obliquity, Green = eccentricity.
        Lower plot: Red = Antarctic temperature.

        (It might be nice to redo this plot with raw precession, instead of the orange Milankovitch plot, it would give a better idea of the relative influences.)

    • I don’t think the question will be put to them. Carbon sequestration on grazing and cropping lands is needed to feed the world this century, reclaiming desserts, restoring grassland and forests for meeting environmental challenges and the rest is just technology across sectors – which is good for business and at which we are very good.

      There are 10’s of millions of people with genius level IQ’s today. The 21st century is when technology evolution accelerates again

    • >>Why when greenhouse gases are abundant and
      >>ice volume is at its lowest is there rapid cooling
      >>from runaway ice sheet feedbacks?

      This happens…

      Firstly – because CO2 feedbacks are irrelevant during ice ages. When measured regionally, abedo feedbacks – whether due dust or bare earth – can generate changes in insolation absorption of 200 W/m2 – some two orders of magnitude greater than CO2. And we know that interglacials and glacials are generated by regional feedbacks, rather than global feedbacks, because they always happen during changes in NH Milankovitch insolation. (Not during SH changes.) So CO2 is irrelavant as an ice age feedback.

      Secondly – the next ice age is generated by NH summer insolation falling below a certain level. This is normally triggered by a NH precessional Milankovitch minima (a NH precessional Great Winter), where NH summer insolation can plunge by 100 W/m2. That is a huge loss of energy to the high latitudes of the NH. So a normal interglacial, which is normally dependent upon precessional insolation, can only last for the 5 or 6 ky of a precessional Great Summer – before the climate is plunged into the Great Autumn and the Great Winter and the ice sheets extend.

      Surprisingly, the ice extent record demonstrates that a full 25% or 30% of the total ice-age ice-sheet volume is built up during that first 5 or 6 ky Great Winter. This demonstrates how devastating a strong Great Winter can be, and how fast these ice sheets can extend and build.

      Ralph Ellis

      • The question was why glacial inception when the planet is warmest and ice extent lowest. MH summer insolation is a very partial answer – it happens only sometimes when insolation is low. There are other – internal – factors and glacials happen when the ducks line up.

        The global change in energy from glacial max to min is some 8W/m2 for a 10K temperature change – e.g. https://www.ipcc.ch/publications_and_data/ar4/wg1/en/ch9s9-2-1-3.html – and I don’t btw think much of your numbers.

        But the IPCC numbers imply a system that seems a lot too sensitive to me. I think the answer might lies in bi-stable, asymmetric closed and open cell cloud formation. Closed cell cloud over cooler oceans – with high albedo – are less energetic and persist for longer than closed cells over warmer water surfaces. Closed cells rain out from the center to form open cells. This is not counter intuitive just counter the naive assumption that warmer means more cloud cover.

      • I rather liked your paper btw – and Clive Best’s graphic.

      • >>The question was why glacial inception when
        >>the planet is warmest and ice extent lowest.

        I think it is clear that every glacial inception (during an interglacial) occurs when Milankovitch insolation reduces below a certain level. And when there was no precessional insolation decrease – due to eccentricity being low, like 420 ky ago – there was no ice-age inception and the interglacial continued for the full obliquity cycle.

        Perhaps you could show us a deep NH insolation minima that did not initiate an ice age.

        Ralph

      • Still not the problem. Low NH summer insolation may be a necessary but not sufficient condition for glacial inception.

        “The climate system has jumped from one mode of operation to another in the past. We are trying to understand how the earth’s climate system is engineered, so we can understand what it takes to trigger mode switches. Until we do, we cannot make good predictions about future climate change… Over the last several hundred thousand years, climate change has come mainly in discrete jumps that appear to be related to changes in the mode of thermohaline circulation.” Wally Broecker

      • Ralf,

        When you say “NH precessional Great Winter” you are really talking about a drop in the “NH Summer” (June) insolation. Inside the Arctic circle the winter (January) insolation is always zero.

        However it is more complicated than that. Kepler’s Law ensures that the integrated “NH summer” insolation hardly changes at all during a precession cycle. The summer period is simply shortened at perihelion.

        The only Milankovitch cycle that increases net “NH summer” insolation is the obliquity. This is obvious since if obliquity was zero there would be no seasons and the insolation at the poles would be precisely zero for ever.

      • >>you are really talking about a drop in the
        >>“NH Summer” (June) insolation Inside the
        >>Arctic circle the winter (January)
        >>insolation is always zero.

        Indeed, but I think the summer melt season is more important than the winter deposition season. If you have a weak and cool summer, it may not melt the winter snows. And the albedo cooling feedback of the remaining snows will be considerable, perhaps allowing a build-up of snow and ice in subsequent years.

        .

        >>Kepler’s Law ensures that the integrated “NH summer”
        >>insolation hardly changes at all during a precession cycle.
        >>The summer period is simply shortened at perihelion.

        Indeed, but a short sharp NH summer will still melt NH ice sheets, even if total summer insolation remains much the same. For instance, the MIS-7 interglacial 240 yr ago was too late for simple obliquity warming, and was likely forced by precession. Conversely, the small warmings 320 kya, 265 kya, 200 kya, 110 kya and 80 kya appear to have been forced by precession, not obliquity.

        Conversely you might argue that MIS-11 435 kya generated an interglacial, even without very much precessional input. Although the lack of a precessional Great Winter 420 kya seems instrumental in allowing MIS-11 to extend for the whole obliquity cycle.

        What is needed here, is a comprehensive analysis of precessional and obliquity insolation inputs – as separate elements – for cycles that DID produce an interglacial (ie: therefore presuming that dust-albedo feedbacks remain a constant in these cycles). This may help decide which element is the more influential.

        The present combined Milankovitch insolation plot masks the relative influence of precessiona nd obliquity, by considering them to be equals. But they are not equal in their influences at all.

        Precession, as you say, is a short sharp summer influence, plus it operates in alternate hemispheres. While obliquity is sustained for a longer period during the summer, and operates at both poles simultaneously. Precession is redistributing energy from pole to pole, while obliquity is redistributing energy from the tropics to both poles. These energy transfer systems are quite different, so they should not be summed within a singular Milankovitch insolation plot.

        Food for thought.

        Ralph

      • Robert I. Ellison | September 8, 2018 at 4:34 pm |
        “The question was why glacial inception when the planet is warmest and ice extent lowest.”

        A bit like why do teams on top of the ladder beat the teams below them?
        Whatever happens at inception, which may be well before the peak, can only be noticed once it’s effects, the dropping heat and increasing ice extent become obvious.

        The valid point is less sun input.
        This can be orbital in that the earth gets dragged further away from the sun in its orbit around it by the proximity of other large planets.
        This can be source derived in that for whatever reason the sun temporarily puts out less energy.
        This could be, no one has mentioned it but, space dust drifting through causing a shadow effect. Would not be seen at times like now if it has gone away in an extended orbit.
        At the earth level albedo changes.
        Cloud.
        Dust.
        Algal or vegetable growth,
        Volcanoes and other magma eruptions.
        Meteorite impacts.
        Many of these would have feedbacks on cloud formation.

        Clive Best makes a very salient point re Milankovitch
        “-However it is more complicated than that. Kepler’s Law ensures that the integrated “NH summer” insolation hardly changes at all during a precession cycle. The summer period is simply shortened at perihelion.”

  16. Donald Rapp, thank you for the essay.


  17. It certainly seems that when it gets cold enough an interglacial is triggered, so some idea of dust due to the extinction of vegetation is plausible for the ones that follow the coldest periods. But we also see interglacials like at 700k, 550k and 200k that don’t follow such cold periods. However, these have stronger solar forcing which seems to sometimes be sufficient by itself. So I would say that there are two causes. Dust is the one that explains the apparent bottom limit to the global temperature, but strong enough solar forcing can be sufficient sometimes without the added dust.

    • Evidently, termination is caused by solar power input to the ice sheets. If the solar power is high enough that it can force a termination without dust, that is perhaps conceivable in a few cases. However, the data suggest that, at least in almost every case, dust enables the solar power to melt the ice sheets.

      • I think the exceptions are the double-interglacials like at 550k and 200k where there is not a dip to very cold temperatures after the first of the pair so the second one probably is not helped by dust. These are also marked by higher than usual NH solar forcing.

      • Jim.
        Donald Rapp’s graphs do show small dust peaks just before those smaller proto-interglacials. See his figs 9 and 10. Again this is Antarctic data, and the Arctic may have received more Gobi dust than this. Donald was using some alternate dust data to the series I used, derived from laser counter and Coulter counter techniques.

        Unfortunately we do not have Arctic data going back that far, unless the Loess Plateau can provide more information. But I am struggling to find a decent record of Loess dust that is pertinent to this research. Lots of graphs trying to give proxy indications of temperature, but not dust flux.

        Ralph

      • There is always going to be more dust at colder temperatures, and the question is how much colder does it need to be for that dust to matter. It seems to matter less when the albedo forcing is larger.

      • Jim.
        It is not the cold that produces dust at the LGM, it is the lack of CO2.

        Image: Dust vs CO2.

  18. The sentence that says:

    Ice Ages cause ice sheet extension Þoceanic cooling Þoceanic CO2absorption = plant asphyxiation on the Gobi plateau Þnew CO2deserts Þdust generation Þice sheet contamination for 10 kyrs.

    Should read:

    Ice Ages cause ice sheet extension = oceanic cooling = oceanic CO2 absorption = plant asphyxiation on the Gobi plateau = new CO2 deserts = dust generation = ice sheet contamination for 10 kyrs.

    Ralph

  19. How can we tell what the temperature was eons ago ?
    I understand carbon dating. But temperature recording somehow ?

    • The oxygen isotope ratio in Arctic and Antarctic ice cores can be used as a proxy for temperature, because the isotope ratio in rain changes with temperature. Similarly, the same oxygen isotope also changes in foraminifera (plankton) shells, and can be extracted from deep sea cores. And the results from these two methods agree very well with each other, so we can be fairly sure the results are correct.

      There are also other methods like the dust layers on the Loess Plateau in China, which record differences in annual dust production and grain size, which must be linked to climate and agree well with the oxygen isotope ratio proxy. Plus there are stalactites and mites that can provide similar data. So the temperature of the last few million years is fairly well established.

      Ralph

  20. I only briefly discussed the topic of duration of Interglacials in this posting. The behavior of the earth system after an Interglacial has lasted some 5,500 years and the solar precession starts downward is difficult to characterize and probably varies significantly from cycle to cycle. It seems likely that dv/dt (rate of ice volume) could turn positive, but sea level remains high for several thousand years as ice slowly accumulates in the incipient emerging new Ice Age. Based on ice volume or sea level, the duration of an Interglacial might be closer to 20,000 years than 5,500 years based on dv/dt turning positive.

  21. >>Note that the current Interglacial seems to be
    >>lasting too long to fit this simple picture.
    >>That remains difficult to explain.

    The extension of the Holocene interglacial, is because orbital eccentricity is low, and so this interglacial is obliquity-led (~41 ky) rather than precession-led (~22 ky). So the NH is getting extra insolation for longer.

    Milankovitch insolation comprises obliquity insolation and precession insolation. This resolves into a combined graph that looks like a simple 22 ky cycle, but is not really, it is a combination of two cycles. But because precession is (normally) much stronger than obliquity, the graph primarily follows the ~22 ky precessionary cycle.

    I call this precession-dominated Milankovitch cycle the Great Year, because that is what the ancient Greeks called it, and the precessional Great Year does indeed have seasons just like an annual year. Just like the annual seasons, a NH Great Summer occurs at the same time as a SH Great winter. During a NH Great Summer each annual summer will provide increased summer insolation in the Arctic region – some 6 kyrs of increased summer insolation in the north. (But it will also provide colder annual winters in the north).

    Note that all interglacials are generated by NH Great Summers, not SH Great Summers. So we know that interglacials are produced by regional conditions and feedbacks in the NH, and not by a global feedback like CO2. If CO2 were the primary feedback for interglacials, then half the interglacials should occur during a SH Great Summer, and they do not. This is a good indication that ice ages are modulated by albedo feedbacks and not CO2 feedbacks, because all the major land masses and ice-age ice-sheets are in the NH.

    However – there is always a however – precessional insolation is reduced considerably when orbital eccentricity is low, as it was 400 ky ago and is now. During these low eccentricity eras, obliquity insolation provides a much larger share of the total Milankovitch Great Summer insolation. This is why the MIS-11 interglacial 400 ky ago, and the Holocene interglacial now, were both extended. Precession was low in these eras, and so there was no precessional-led NH Great Winter to cool the Arctic and force a new ice age (a strong NH Great Winter can reduce the NH summer insolation by 100 W/m2).

    So this is why our current Holocene interglacial is happily following the reduction in obliquity – as per the diagram below – because there is no deep precessional Great Winter to force NH cooling and force the next ice age. But the evidence from MIS-11 demonstrates we can get an obliquity-led ice age. The only caveat being that eccentricity will remain low for some considerable time. Our current orbital conditions are just about the same as when MIS-11 fell into an ice age, and so we are currently sitting on a knife-edge (yet another tipping point) between interglacial and ice age conditions.

  22. While every termination is accompanied by the 5,500-year rising portion of an up-lobe in the solar input to high latitudes, many strong up-lobes do not produce a termination.

    It takes a long time for ice sheets to thaw and thin, termination occurs when the ice runs out. One of the up-lobes will correlate close to that time.

    • >>It takes a long time for ice sheets to thaw and thin,
      >>termination occurs when the ice runs out.

      Not true. Nearly all up-lobes (Great Summers) are of the same duration, because they are nearly all based upon the same ~22 ky precessional cycle. And yet some Great Summers will make no impression upon ice extent whatsoever, while others will melt ALL the ice sheets in their entirety – in just 5 or 6 kyr. There is a huge difference in the response of the ice sheets, to very similar NH insolation increases.

      See Clive Best’s graph below. See how some very strong insolation increases (upper dark blue) made no impression on ice extent (middle light blue).

  23. And what is the missing ingredient that turned the rise in northern insolation around 20,000 years ago into the starting gun for deglaciation, when higher insolation at earlier times failed to do so?

    It ice that was sequestered on northern continents, more than a hundred thousand years ago, finally thawed and thinned and retreated. Again, warming occurs when the ice age ice thaws and runs out of ice to thaw.

    High insolation does not end an ice age when there is still too much ice. An ice chest does not warm when there is still too much ice. When the ice is depleted enough, warming can occur.

    • >And what is the missing ingredient that turned the
      >>rise in northern insolation around 20,000 years ago
      >>into the starting gun for deglaciation, when higher
      >>insolation at earlier times failed to do so?

      Dust.

      Please read the paper. That is the whole point of the paper – that dust on the ice sheets will lower their albedo, and allow more insolation absorption, and therefore allow them to melt (in the presence of higher Milankovitch Great Summer insolation).

      Modulation of ice ages via precession and dust-albedo feedbacks
      https://www.sciencedirect.com/science/article/pii/S1674987116300305

      Ralph

  24. Need for an X-factor to accompany solar up-lobes at termination

    Ice depletion is the X-factor that always accompanies the beginning of warming. It is not a result, it is a cause, it is THE cause.

    • >>Need for an X-factor to accompany solar up-lobes at termination.

      That is the whole point of the article.
      Read it again.

      R

  25. the longstanding Ice Age had been maturing for many tens of thousands of years. While the ice sheets grew, periodic up-lobes of solar input inhibited ice sheet growth for a time, while down-lobes enhanced ice sheet growth. But on balance, the ice sheets grew for many tens of thousands of years.

    It takes that long to build the ice sheets and deplete the energy and water stored in the oceans to produce the ice. The ice volume maxes and the ice extent grows and causes cooling even after ice volume has maxed and is decreasing. Thawed ice produces water and evaporation and snowfall and redistribution of the ice as snowfall elsewhere.

  26. Great article again!

    “The point of no return” the Termination, i think there is at least one more evolution to think over.
    After many milenia of ice age the ice shields reach very low latitudes at NH. They are thick, but at times their aera is still growing they began to thin out.
    Than, if the named parameters fit together, the ice shields and glaciers over lower latitudes collapse rapidly and the Termination is “unstoppable”…?

    • After the ice sheets thin, they retreat! YES!

    • That hardly makes sense.
      The ice sheets are at a stable extent. Then they grow to a more unstable extent. And this forces a collapse – but not back to the previous stable extent, but right back to nothing. Why should that happen?
      And there is no evidence that ice sheets thin as they grow.

      R

      • And there is no evidence that ice sheets thin as they grow

        Ice in ice sheets and glaciers do flow. debris was pushed from the North to middle latitudes by flowing ice, the ice sheets thinned as they flowed.

        There is plenty of evidence of this.

      • >>Ice in ice sheets and glaciers do flow.

        We know ice sheets flow, but where is the evidence that they thin with time? The upper layers of the high ice sheets are constantly being refreshed with new winter snows, as the ice cores from the high Greenland plateau demonstrate.

        R

  27. I have a copy of the phd dissertation of prof Hans Oerlemans done in 1980
    http://www.staff.science.uu.nl/~oerle102/
    He describes a mechanical component of the termination of the ice age when the ice flows southward under its own weight and then looses height in the northern feeding area, which then accellerates the slumping.

    • Well yes, just like glaciers flow and thin and deplete and retreat when they run out of ice.

    • I have not actually read that but you described it very well.

    • Thanks, an additional trigger is the isostatic adjustment of the mantle. The weight of the icecaps lowers the top height, which additionaly prevents the icecap from growing further. Post glacial rebound lifts the area up again so the area can be seeding ground for a new icecap in the next jce age.

      Iff this mechanical see-saw synchronises with the milankovitch see-saw there is an amplifying effect. Evolution of elastic mantle properties could explain the observed harmonic effects in the pleistocene ice ages.

  28. In dealing with phenomena in the distant past, there will always be some uncertainty. Many individuals have their pet theories. We’ve heard them before. Several times. Maybe more than several. But only theories deeply imbedded in the data are worth considering.
    The ice sheets melted and an energy source was needed. Figure 1 shows that each termination ramp occurred on an up-lobe in the persistent oscillations due to precession. The evidence is overwhelming that solar energy melted the ice sheets. But why did some specific solar up-lobes cause termination when others didn’t? Because the variations of solar input to high latitudes from cycle to cycle are actually relatively small percentage-wise, and apparently, a small increase in solar alone does not get the job done. A much bigger factor is needed: solar absorptivity. Figures 9 and 10 provide very strong implications that dust generated near the end of a long Ice Age when CO2 dips below 200 ppm is the factor that increases the absorptivity by far more than a few percent.

    • Some uncertainty indeed.
      A simple mechanical analogy (Source: NAS Committee on Abrupt Climate Change, 2002)

      Many simple systems exhibit abrupt change. The balance above consists of a curved track on a fulcrum. The arms are curved so that there are two stable states where a ball may rest. ‘A ball is placed on the track and is free to roll until it reaches its point of rest. This system has three equilibria denoted (a), (b) and (c) in the top row of the figure. The middle equilibrium (b) is unstable: if the ball is displaced ever so slightly to one side or another, the displacement will accelerate until the system is in a state far from its original position. In contrast, if the ball in state (a) or (c) is displaced, the balance will merely rock a bit back and forth, and the ball will roll slightly within its cup until friction restores it to its original equilibrium.’(1)

      In (a1) the arms are displaced but not sufficiently to cause the ball to cross the balance to the other side. In (a2) the balance is displaced with sufficient force to cause the ball to move to a new equilibrium state on the other arm. There is a third possibility in that the balance is hit with enough force to cause the ball to leave the track, roll off the table and under the sofa.

      “Now imagine that you have never seen the device and that it is hidden in a box in a dark room. You have no knowledge of the hand that occasionally sets things in motion, and you are trying to figure out the system’s behavior on the basis of some old 78-rpm recordings of the muffled sounds made by the device. Plus, the recordings are badly scratched, so some of what was recorded is lost or garbled beyond recognition. If you can imagine this, you have some appreciation of the difficulties of paleoclimate research and of predicting the results of abrupt changes in the climate system.” https://www.nap.edu/read/10136/chapter/3#13

  29. Reviewing all the data as far back as 8000,000 years, the durations required for the termination ramp from glacial to interglacial conditions in the last nine deglaciations are listed in Table 1. On average, the duration of the transition from glacial to deglacial conditions took roughly 6,000 years (about ¼ of a precessional period).

    When the ice age ice depletes and the ice sheets thin, the ice retreat happens relatively rapidly. About 10 thousand years for the last major ice age, from 20 to 10 thousand years ago.

  30. There is no doubt that there is merit in the widely accepted Milankovitch theory that Ice Ages and their terminations are controlled by solar input to the NH in mid-summer.

    There is doubt that there is merit in the not accepted theory that ice ages end when the ice is depleted and the ice sheets thin and retreat.

    Ice ages are caused by ice and they end when the ice runs out.

    BUT THAT IS WHAT HAPPENS AND WHAT CAUSES ICE AGES TO END, RUNNING OUT OF ICE. Think about this!

  31. I read your entire explanation and think I undrstood some of it. To me this is just simple science of heat teransfer, the fact that water expands as it goes from 39’F to 32’F, and water reflects radiant heat. If I understood what you said, I thing you agreed that the new Ice Making stage of the new Ice Age began about 18 thousand years ago. That is enough for me.

  32. Salvatore del Prete

    MY QUESTION FOR ANYONE ON THIS BOARD.

    How does any of the possibilities we are discussing explain all the abrupt climatic changes that took place especially 20,000 to 10,000 years ago ending with the YD, which was not unique to that period of time?

    All the things being discussed take forever to happen and yet we had the earth going from inter- glacial to at least half full blown glaciation in periods of thousands of years if not hundreds of years or even less during that time frame.

    What was behind these large swings back and forth ?

    • “Recent scientific evidence shows that major and widespread climate changes have occurred with startling speed. For example, roughly half the north Atlantic warming since the last ice age was achieved in only a decade, and it was accompanied by significant climatic changes across most of the globe. Similar events, including local warmings as large as 16°C, occurred repeatedly during the slide into and climb out of the last ice age. Human civilizations arose after those extreme, global ice-age climate jumps. Severe droughts and other regional climate events during the current warm period have shown similar tendencies of abrupt onset and great persistence, often with adverse effects on societies.” https://www.nap.edu/read/10136/chapter/2

      In the words of Michael Ghil (2013) the ‘global climate system is composed of a number of subsystems – atmosphere, biosphere, cryosphere, hydrosphere and lithosphere – each of which has distinct characteristic times, from days and weeks to centuries and millennia. Each subsystem, moreover, has its own internal variability, all other things being constant, over a fairly broad range of time scales. These ranges overlap between one subsystem and another. The interactions between the subsystems thus give rise to climate variability on all time scales.’

      The theory suggests that the system is pushed by greenhouse gas changes and warming – as well as solar intensity and Earth orbital eccentricities – past a threshold at which stage the components start to interact chaotically in multiple and changing negative and positive feedbacks – as tremendous energies cascade through powerful subsystems. Some of these changes have a regularity within broad limits and the planet responds with a broad regularity in changes of ice, cloud, Atlantic thermohaline circulation and ocean and atmospheric circulation.

      I always heavily discount X factors – and the large majority of change in the system is internally generated.

    • >>How does this explain all the abrupt climatic changes
      >>that took place ?

      If you read the paper, Salvatore, you will find out.

      The theory is that the dust concentrates on the surface of the ice, lowering the albedo from 0.85 to perhaps 0.5 or less. This can increase insolation absorption by 200 or 300 W/m2 over the northern ice sheets. That is two orders of magnitude greater than is ascribed to CO2 feedbacks.

      Not only that, but the W/m2 figure ascribed to CO2 is only valid after a full 5 kyr, while the value for dusty-ice-albedo applies from almost day one. So after 10 years of interglacial warming the albedo feedback might be absorbing 200 W/m2 regionally, while the CO2 feedback will only be adding just 0.006 W/m2 globally to increased warming. The CO2 figure would not even power a honey bee in flight.

      The 200 to 300 W/m2 provided by albedo is a vast amount of extra energy, that ablates and melts the massive ice sheets within the 5 ky of a precessional Great Summer. While the reverse process is, of course, much slower – building an ice sheet takes up to 115 kyr.

      Ralph Ellis

      • Perhaps I am wrong but again this doesn’t seem to be the question that was asked.

      • When the massive ice sheets thawed, melt-water was trapped on land.
        The melt-water broke out in surges that caused rapid cooling of oceans and there was rapid warming as this ice cold melt-water mixed with warmer oceans. This caused the temperature spikes in temperature data and spikes in sea level rise data between 20 and 10 thousand years ago.

    • >>How does this explain all the abrupt climatic
      >>changes that took place especially 20,000 to
      >>10,000 years ago ending with the YD

      Do you mean Dansgaard–Oeschger fluctuations? They took place a lot earlier than 20 ky ago.

      In truth even though the science is settled, nobody knows what causes Dansgaard–Oeschger events – but the dust-albedo theory would suggest a possible answer. Remember that these are very rapid and transient warmings, that take place in less than 100 years (I saw one paper that said less than ten years). So that is an 8ºc warming inside ten years, leading to up to 10% of NH ice sheet loss and the rafting of icebergs, and then a more gradual re-freeze. They are mini interglacials, but are hardly visible in the SH.

      So what event in the NH could cause massive warming, and is linked to albedo? How can we suddenly darken the northern ice sheets, while CO2 remains high and there is no chance of the Gobi becoming a CO2 desert?

      The answer is forest fires, which coat the ice sheets with black soot, which is the most powerful of the albedo agents. And all the D-O events are linked to forest fire combustion products. But since the fires only coated a thin layer of ice, say a couple of decades or centuries, the ice-melt was limited and the ice age returned.

      However, while this may be a compelling theory, nobody has looked for this ash and soot within the ice core record, to differentiate it from the usual Ca dust readings.

      Ralph

  33. So desert dust may be the “x” factor. It should be pretty straightforward measuring the amount of dust present in the layers of ice cores – you mentioned some studies that have. Now, can someone please do an experiment to verify the amount of albedo change and increase in absorption of ice with different levels of dust present. Then we will know whether to start looking for “y”.

    • As I said in an earlier posting, we don’t know the dust levels at the ice sheets during the LGM very well. Several published models indicate that dust levels at the ice sheets far exceeded Greenland dust levels. I can give you references. In my original posting I estimated that the true optical effect of 1 g/m2 of dust particles is to optically block 100% of the surface. All the models indicate much higher dust levels on the ice sheets at the LGM.

    • Actually, we know the Greenland dust contamination levels very well, because we have the layers still in situ. See image below. What we have not had as yet, is any serious study of the albedo changes in these layers, nor any studies on how much the dust will concentrate on the surface during ablation and melting.

      This is because nobody has previously thought or argued that dust may be responsible for interglacial modulation. So the dust studies that have been performed on this Greenland LGM dust were designed to see if modern industrial soot was increasing Greenland melt (to see if the Greenland ice sheet would melt away and cause sea level rises). So the experiment was completely wrong, and did not determine what was happening during the LGM, nor at the beginning of the Holocene interglacial.

      Before the right experiment can be designed, the right questions have to be asked. But now the theory is out there, and the pdf has been read by 20,000 people so far, designing the experiment is fairly simple. All we need is grant. And for that we need an educational establishment that is prepared to think the unthinkable – that CO2 is not the primary feedback agent controlling ice ages – and we will be away.

      And therein lies the problem. The CO2 bandwagon has sucked up all the grants and demonised all the other ideas and theories, so the alternate experimentation cannot be done. Indeed the Royal Society declared the Dust-Albedo paper to be a Denier Paper, and treated it with derision – so I don’t think that funding will come from that source.

      Mind you I did retort that the Royal Society was inhabited by imbeciels. One of the paper’s reviewers said that plants at high altitude could not be starved of CO2, because the concentration of CO2 at altitude is the same as at sea level. And so the entire theory was wrong, and the paper was thrown out because of this. I kid you not. I did ask in return why airliners have emergency oxygen for passengers, since the concentration of O2 at altitude is the same as at sea level…! That is the Royal Society for you – a bastion of settled science based upon kindergarten education.

      • Ralf, I think your problems with getting published except for in this Chinese journal relate to some wrong statements about CO2 forcing. Usually forcing is expressed as a global average because it is the global temperature that responds. Also dividing the CO2 forcing by 5000 years to get your “honey bee” quote is plain wrong. 3 W/m2 is already a rate. You don’t divide that by time to get an average. It applies over the time over which CO2 changes, so yes it is 3 W/m2 and yes that is comparable with the global average albedo forcing change, and yes Hansen was more right than you. Your attempts to dismiss CO2 were probably your downfall, and were totally unnecessary for the dust argument anyway.

      • In Jiminy’s bizarro world there are no planetary responses and the TOA forcing just keeps increasing. Glacially slowly in this context of course. The total change in forcing is some 8W/m2 btw – glacial max to min – still glacially slow. And just how definitive can that be?

        Better off with temperature – at equilibrium –

        ΔTs = −F/λ – where F is 8W/m2 and net feedback – now and then is -2W/m2 per K.

        => ΔTs = 4K

        Wait – that can’t be right.

      • >>Also dividing the CO2 forcing by 5000 years to get
        >>your “honey bee” quote is plain wrong. 3 W/m2 is
        >>already a rate. You don’t divide that to get an average.

        3 W/m2 is the value of forcing that will occur with a doubling of CO2. But if the CO2 has not doubled, then how can you get that much forcing?

        Perhaps you could calculate the increase in forcing at the beginning of the interglacial, when CO2 concentrations had only increased by 2 ppm – when they went from 180 ppm to 182 ppm. What was the extra forcing caused by that small change in CO2 concentration, during the first 100 years of the interglacial? (CO2 concentrations rose at about 2 ppm per century, during interglacial warming.)

        Thanks,
        Ralph

      • >>Usually forcing is expressed as a global average
        >>because it is the global temperature that responds.

        But we know that interglacial warming is a regional effect, not a global effect, because it only ever happens during NH Great Summers (NH Milankovitch maxima). If interglacials were being triggered by a global feedback agent like CO2, then we would see some interglacials triggered during SH Great Summers. But we don’t.

        So interglacial warming is a regional event in the NH, probably linked to the NH’s large continents and to their large ice sheets – ie: it is linked to NH albedo. Yet traditional climate science is so wedded to the the concept of global influences, that Hansen spread ice-sheet albedo forcing out across the globe, achieving something like 4 W/m2 globally – much the same as the feedback-forcing of CO2. But that is not what happens.

        Ice ages are just like normal annual years. That is why I like the terms Great Summer and Great winter, because they are very descriptive of what happens during the ice age cycle. So what melts the ANNUAL snow and ice in Canada, is it the return of strong insolation to the far north during the Annual Summer, or is it the ambient temperature in Argentina? Clearly northern annual snow and ice is cleared by regional effects in the north, and the clearance of northern ice-age ice-sheets is no different.

        When expressing the albedo effect regionally and seasonally, we do not get Hansen’s 4 W/m2, we actually get 200 or 300 W/m2 – a huge increase in regional insolation absorption and warming. This is the effect that warms and melts the northern ice sheets, not the trivial global effects of CO2.

        Ralph

      • Hansen has 3 W/m2 from CO2 and related GHG changes and 3.5 W/m2 from albedo over a transition period (in his case it was the cooling phase, but it should still apply to warming) making the GHG effect of comparable importance in the global warming to the albedo effect when accounting for the final temperature change that he has as 5 C globally from the total 6.5 W/m2 forcing change. This is comparable with sensitivities expected given how water vapor adds to the response. You can’t dismiss the CO2 effect with the “honey bees” comment when it is comparable with albedo for the warming.

      • The pattern of insolation affects the global energy budget via its mean effect on the global albedo. Large ice sheets have an effect on the global albedo, and less energy is absorbed by the earth as a whole. This has global influences, even at Antarctica, but that influence is proportional to how much less energy is absorbed as a global average. This is where the feedbacks from CO2 and other GHGs come in, because those respond to the global temperature. Apples with apples is using global averages for forcing. The net insolation change is nothing globally. Its global effect is only seen through the changing albedo pattern, so this is albedo forcing not insolation forcing, and that is how Hansen views it.

      • >>You can’t dismiss the CO2 effect with the “honey
        >>bees” comment when it is comparable with albedo
        >:for the warming.

        When making a new proposal, it is usual to challence the old assumptions. And this is one of the old assumptions that needs challenging, because it is being used in an incorrect manner.

        So you did not answer my question. What will the CO2 forcing-feedback when CO2 concentrations have risen from 180 to 182 ppm, during the first century of the interglacial. Because each century needs to force the warming to achieve an increase in temperature into the next century.

        It is not a difficult question. Do you think the answer is 3 W/m2?

        R

      • I could answer that, but then you have to answer how much the global albedo changed in that century for it to be apples to apples. What matters more is the total forcing change from beginning to end because we are only interested in the temperature step in the end. That’s what affects the climate. Hansen has it at 3 W/m2 from non-H2O GHGs and 3.5 W/m2 from albedo. You seem to be disputing this, but it is unclear from the paper.

      • >>I could answer that, but then you have to answer how
        >>much the global albedo changed.

        This is another thing Hansen got wrong – albedo is regional, not global. It is the local insolation absorption on the ice sheets that is critical, not a smeared-out global albedo. I ask you, what melts the annual winter ice in Canada – is it the albedo-insolation-absorption on the ice in Canada, or is it the albedo in Argentina…?

        Besides, even if you smeared the ice sheet albedo response all over the earth, it would all act from day one. The dust does not build up on the ice sheets slowly over the length of the interglacial like CO2 does – it is there from the very beginning. Interglacials only occur after a few millennia of dust has already fallen, so the dust is already in the ice sheet and will be exposed by a small increase in warming (and stay on the surface while the ice sheet melts).

        Ralph

  34. sheldonjwalker

    ❶①❶①❶①❶①
    ❶①❶①❶①❶①
    ❶①❶①❶①❶①
    ❶①❶①❶①❶①
    .

    “Beyond Milankovitch” is an excellent article. It is well presented, easy to follow, and very believable.

    Just like my articles.

    Global Warming – Did we Pass or Fail?

    A detailed analysis of global warming, in the different regions of the Earth.

    – the Arctic region

    – the Antarctic region

    – the Land

    – the Oceans

    – the entire Earth

    This article is very relevant to the problem of global warming.

    Can we save the Earth, and the human race?

    Have the 1.5 and 2.0 degrees Celsius temperature limits, become irrelevant?

    https://agree-to-disagree.com/global-warming-did-we-pass-or-fail

  35. The dust peaks around 665 kyr and between 450-470 kyr BP are a problem for this postulate.

    • Those dust peaks are not a problem. The theory requires that two key elements have to be in place before an interglacial can be initiated – some 10 ky of dust, and a large increase in NH insolation due to a Milankovich maximum (a Great Summer).

      Dust on its own is not sufficient to generate an interglacial, so if there is a large dust deposit and a weak Great Summer, like 270 ky ago, the ice age grinds on until the right conditions are in place.

      This dust can remain in situ for a few thousand years, waiting for an increase in insolation. As long as the next Great Summer is strong enough to burn away a few upper layers of less dusty ice, it will eventually reach the highly contaminated layers that lie below. Remember that the dust concentrates on the surface of the ice sheets, and so any ablation or melting will reduce the albedo yet further, and create a runaway effect.

      Ralph

    • The two dust spikes that I referred to both appear have higher summer energy. And some dust spikes at low summer energy are at glacial terminations.
      The only thing here that we can be certain of, is that colder periods are dustier.

  36. If readers are looking for what the ice sheets would have looked like at the LGM, when covered with Gobi desert dust, then here is an image of that very dust – still in situ after some 20 kyr.

    This is a remnant of LGM ice way up in the north of Greenland, that has been extruded and thinned as it flowed some 20 km from where the snows originally settled. You can imagine that if there was this much dust way up in the north, then the Laurentide ice sheet over the Great Lakes must have been almost black with dust. The albedo of the southern ice sheets may have reduced to 0.5 or less – there was more dust and greater albedo reductions than our paper has allowed for.

    Image:
    The dusty ice on the right hand side of this image represents some 35 kyr ago, during the dusty era before the interglacial was initiated. Moving to the left along the blue arrow takes just beyond the start of the Holocene interglacial, which began some 17 ky ago. NH dust eventually stopped some 14 ky ago, as CO2 concentrations rose, the CO2 deserts turned green again. Then there is another thin line of dust, caused by the Younger Dryas cool period which started some 12 ky ago, when CO2 levels fell again and the dust returned. And then the dust stops again, as we enter the Holocene interglacial.

    • RalfEllis,

      The last time Earth was in a cycle of ice ages and interglacials was some 300 Ma ago. That lasted about 70 Ma. Eventually it terminated at GMST increased by about 15C over the next 40 Ma, Can you suggest the likely cause(s) that enabled Earth to escape that 70 Ma period of ice ages and interglacials?

      • An interesting question that we have not looked at. Quite obviously there is a forcing-feedback element that controls the long term climate. Classical science would say it is CO2 that causes long-term fluctuations, but I am not sure that CO2 has the power to create such dramatic temperature changes.

        Certainly during the recent ice age cycle, CO2 appears to be relatively impotent in the face of large swings in temperature. The D-O events of the last ice age gave 8ºc of NH polar warming inside 10 years, which has to be well outside the scope of CO2 influences. So we know there are other effects that can cause dramatic temperature changes.

        I have not researched this, but other very long term climate forcings could include solar cycles or perhaps orbital changes. Current orbital calculations by Laskar demonstrate orbital stability back to 25 million years, but prior to that it is difficult to tell exactly what happened.

        Ralph

      • Ralf Ellis,

        An interesting question that we have not looked at. Quite obviously there is a forcing-feedback element that controls the long term climate.

        Thank you for your reply, and for explaining that my question is not something you have researched or can answer.

        I was prompted to ask this question by something Javier said in his last post https://judithcurry.com/2018/08/14/nature-unbound-x-the-next-glaciation/ . He said that once a certain obliquity and some other condition is reached, Earth is committed to proceed to the next glacial period. He said there have been no cases in the past 800,000 years where another glacial period has been avoided once the condition has been reached (from memory). So, I wondered how Earth escaped from the Permian glaciation. The relevance of this question is that it might help us to understand what conditions would be required for Earth to escape the current icehouse phase.

        Shaviv, Vesier and others have suggested there is a 150 Ma cycle of icehouse periods, and every second one is colder – hence the 70 Ma Carboniferous-Permian icehouse at around 300 Ma ago and the one we are in now. Between these was the less severe Silurian ice age around 450 Ma and mild period 170 Ma ago (with minimum about 4 C warmer than present). Shaviv offers and explanation for the 150 Ma cycle is an increase in cosmic ray flux caused by … :

        Shaviv (2007). The Milky Way Galaxy’s Spiral Arms and Ice-Age Epochs and the Cosmic Ray Connection http://www.geosociety.org/gsatoday/archive/13/7/pdf/i1052-5173-13-7-4.pdf . [I notice you are aware of this because you posted a comment on it 2 years ago]

        Shaviv and Veizer (2003). Celestial driver of Phanerozoic climate? http://www.geosociety.org/gsatoday/archive/13/7/pdf/i1052-5173-13-7-4.pdf

        However, this does not explain what conditions existed that allowed Earth to escape the 70 Ma icehouse period, with its sequence of glacial and interglacial periods similar to those occurring since the start of the Pleistocene.

      • Peter Lang. Your question is one I have not answered in my upcoming paper, but the thought certainly arose during the writing of it. How do you get out i of the ice planet state? My paper is nearly finished and quantifies a clear and straightforward probable cause of the interglacial heating event, and also how the warming, perhaps not the effect, is quickly halted a few degrees higher, in the straightforward physics manner I already suggested above, with numbers. by cloud formation, variable with temperature and humidity, that controls 140W/m^2 of insolation during interglacials. Far more powerful than the forcing effects of the small changes in our orbit and spin. The warmer it gets, the less the surface insolation. But that is at the end of the warming.

        I suggest the answer to the root cause of an interglacial thaw and escaping an ice planet state is the same, and the only way the Earth can escape an ice planet state, once the ice covers the sea, albedo halves from 0.3 to 0.15, which I assume means the surface absorbs half today’s heat? and the water vapour in the atmosphere must drop to a very low number, as must other gases from the oceans, as there is much less evaporation from the ice, plus some sublimation? So that is a seriously latched state. I am going to have to publish somewhere soon so people can check my numbers and throw rocks, but where? I’d like to get my name on it as an original work….

      • Another long-term climate forcing I liked, was an old paper that suggested the slow rise of the Himalaya was the cause of cooling over the last few million years. Because of its tropical location, the Himalayan plateau and its high altitude, high albedo ice-sheet, can cause a disproportionate amount of insolation reflection and cooling.

        I was not entirely convinced that this forcing effect was strong enough to give the observed results, but it does demonstrate the wide variety of forcing and feedback agents than can effect the long-term climate.

        Ralph

      • I hadn’t previously seen the Himalayas as an explanation. But I have seen the joining of North and South America and blocking of the ocean circulation as a cause. Recent evidence suggests it closed about 10 ma ago (i.e. earlier than commonly claimed). This coincides with an abrupt cooling

        Source: Source: IPCC AR4 WG1 Chapter 6 Figure 6.1 https://www.ipcc.ch/publications_and_data/ar4/wg1/en/figure-6-1.html

  37. Herman Alexander Pope

    Go look at the tail of a glacier. There are dark streaks in the ice but there are white streaks above. Melt water evaporates and new snow covers layers of dust. The dust layers combine when the ice thaws. Dust causes less thawing than it gets credit for.

    • Yes, but it would appear that interglacial warming required 10 ky of dust layers, before it could achieve the full melting of the ice sheets. When only a small layer of ice was contaminated, the result was a transient D-O event. Plus if you have a large thickness of ice sheet contaminated by dust, the dust will concentrate on the surface, and reduce the albedo of the ice even further.

      So the conditions for precipitating a full interglacial are quite specific — some 10 ky of dust, contaminating a substantial thickness of the ice sheet, plus a NH Great Summer (NH Milankovitch maximum).

      And dust-soot can be quite influential on even small glaciers.
      See: End of the Little Ice Age in the Alps forced by industrial black carbon, by Painter.
      http://www.pnas.org/content/110/38/15216

      Ralph

  38. Maybe its time to mention that there doesn’t seem to be a well defined “Milankovitch Theory”. At least, when I punch “Milankovitch Theory” into Google, I get a whole bunch of vague and unclear responses that more or less merely say that variations in the earth’s orbit affect solar input to high latitudes and this is somehow related to occurrence of Ice Ages. Exactly how solar input creates and terminates Ice Ages is left as an exercise to the reader.
    The dust concept by Ellis and Palmer does not contradict Milankovitch Theory because there is no Milankovitch Theory as such. The dust concept by Ellis and Palmer for the first time, clarifies what the Milankovitch Theory is. It is a theory that says over the past 800,000 years or so, the earth would have remained in an Ice Age despite variable solar input to high latitudes, were it not for occasional events that caused temporary interruptions. And when we look at the data, lo and behold we find not solar wind, not humidity, not warmth generated by declining ice, nor volcanoes; but we find high dust levels and rising solar intensity at each instance.

    • There is a Half-Milankovitch theory.

      All ice ages and interglacials are linked to Milankovitch orbital cycles, but not all of them work. So as Donald Rapp mentioned in his article here – there needs to be a Factor-X that can differentiate between Milankovitch forcings.

      Factor-X has remained elusive for many decades, but it does seem likely that Factor-X is actually dust and albedo.

      Ralph

    • Maybe its time to look into the foundations of the Milankovitch theory. It relies on an assumption that Earth obliquity does not change more than 2deg from a fixed mean value (of 1850).

      The source of that is in the link/statement here: https://malagabay.wordpress.com/2015/04/14/celestial-crystal-balls-and-the-temple-of-amen-ra/j-n-stockwell-1872-obliquity-statement/ It refers to secular changes. How that became dogma is told here: https://malagabay.wordpress.com/2015/04/14/celestial-crystal-balls-and-the-temple-of-amen-ra/ In the second link reference is made to an observation of JF Dodwell in 1936 re abrupt obliquity changes.

      JH Lieske in “THE EVOLUTION OF ADOPTED VALUES FOR PRECESSION” says “The most thorough observational investigation to date was done by Wittmann (1979, 1985) who ultimately found by careful re-reduction of older data that the earlier observational result probably was fictitious in nature.” A. Wittmann in “The obliquity of the ecliptic” had noticed anomalies between theoretical and ancient measurements of obliquity in 1979, but curiously in ’85 put doubt in his earlier opinion.

      The same questioning sentiment is evident in the 2004 “PROGRESS REPORT OF THE INTERNATIONAL ASTRONOMICAL UNION DIVISION I WORKING GROUP ON PRECESSION AND THE ECLIPTIC” J.L. HILTON

      Today there is strong evidence in support of Dodwell’s abrupt changes, whereas the idea or ‘incredible certitude’ of an unchanging obliquity remains afaik- vague. Abrupt changes in obliquity put Milankovitch theory in serious question.

  39. Very interesting post. Thank you.

    It’s fascinating that terminations occur when CO2 concentrations are lowest, and interglacials end when CO2 concentrations are high. This suggests that estimating climate sensitivity on the assumption that warming is mainly due to change in CO2 concentration is deeply flawed.

    • The oceans are huge carbonated drinks and the vapor pressure of CO2 in the atmosphere gets higher as ocean temperature gets higher and the vapor pressure gets lower as temperatures get lower, just like in carbonated soft drinks. CO2 changes follow temperature changes, they do not lead.

    • >>It’s fascinating that terminations occur when CO2
      >>>concentrations are lowest, and interglacials end
      >>when CO2 concentrations are high.

      Indeed – that was the very conundrum that started this research. It did seem that CO2 was NOT controlling temperature – but if it did not, then what else could? The answer had to lie in the NH, because all ice ages are terminated by rising insolation in the NH – a NH Great Summer. So what feedback agent resides only in the NH, and not in the south? The answer had to be northern ice sheets, and their high albedo.

      And the theory fitted the evidence like a glove. High albedo ice sheets will indeed reduce insolation absorption and increase cooling, perhaps for thousands of years, forming an ice-age. And then we need an agent that can reduce ice-sheet albedo, to produce an interglacial. And quite conveniently, every interglacial is preceeded by 10 ky of dust, which we know did coat the ice-sheets and reduce its albedo.

      So the theory was logical and supported by real-world evidence. And it also explained why there was no runaway temperature increase during interglacial warming – because in an albedo-controlled world, warming feedbacks cease when all the ice has gone.

      Ralph

      • Thank you for your clear and concise explanation. This was explained some time ago on Climate Etc, perhaps buy you, along with some excellent photos of almost black ice in Greenland. The black is from black carbon.

    • Ironic if peak dust from low CO2 conditions brings on warming.

      • Indeed. This was one of the primary things that initiated this research, as it seemed to eliminate CO2 as a cause. And yet eminent scientists continued to champion CO2 as the feedback agent, despite the obvious confliction. There had to be a more logical reason for ice age modulation.

        Ralph

  40. “Finally, it is noteworthy that because Ellis and Palmer (2016)was published in an obscure journal, it seems to have been somewhat ignored by the science community. ”

    There is a reason for that. I mean seriously they cite Willis’ conjectures about tropical clouds as if it were a viable climate theory.

    If ya’ll want to be taken seriously about dust and ice ages then you have a long way to go.

    • Steven Mosher: There is a reason for that. I mean seriously they cite Willis’ conjectures about tropical clouds as if it were a viable climate theory.

      What evidence can you cite that they are not viable hypotheses?

      What exactly do you think happens on the ocean surface when the incident downwelling LWIR increases? And the evidence for that is what — that nobody has studied it yet?

      O’Gorman and others reviewed the research on the relationship of tropical rainfall to surface temperature increase and found a 6% increase in rainfall with per 1C increase in surface temp. Is there some evidence somewhere that an increase in rainfall over the ocean surface can increase necessarily without an increase in cloud cover?

    • >>I mean seriously they cite Willis’ conjectures about tropical
      >>clouds as if it were a viable climate theory.

      Oh, look, its Moshy with his standard drive-by snark. Tell us, Moshy, what exactly is wrong with Eschenbach’s Thunderstorm Theory of climate regulation?

      And if you read my paper, which I am sure you have not, you would see that the Eschenbach paper represents one line of text, and was only included because one reviewer wondered how the climate was regulated after all the ice sheets had melted. Scientists have to keep the reviewers happy, but I don’t expect you would know about that, as you have not published yourself….

      But the Thunderstorm Theory takes no part in ice age modulation, according to this theory. Thunderstorms may play a minor role, but ice sheet albedo trumps everything.

      Ralph

    • >>If oul want taken seriously about dust and ice
      >>ages then you have a long way to go.

      Ok, Moshy, so what is wrong with the dust-ice-albedo theory? Give us your ever so worthy opinion of its strengths and failings.

      R

  41. A wonderful paper, easy to read.
    I like the dust theory as it provides a mechanism required for the relative speed of the glacial termination to interglacial. A significant rise in temperature over a short period. Retained energy provides the kick start.

    Dust settling on ice affecting albedo would not produce that speed of change.

    Perhaps the dust caused the following.

    Ice sheets expanding from high latitudes require water vapour in atmospheric transport from tropical latitudes. Increased dust aerosols cause tropical evaporation to fall out of the atmosphere closer to the evaporative source as warm rain, retaining heat at low latitudes. Dryer atmosphere reaches the higher latitudes reducing ice sheet formation.

    The reintegration of higher volumes of warm rain accelerates ocean surface temperatures, increasing evaporation increasing the cycle. We have just witnessed that very same cycle. Earths cycle of evaporation to absolute condensation (full loss of energy) is far from perfect, highly variable and prone to choking. That 100% efficiency exists is an assumption.

    Dust increases aerosol cloud and rain formation, increasing the efficiency of local repatriation of evaporated energy, ceasing ice sheet growth. Only retained energy could create the rapid rise in temperature. Dry warmer winds accelerate the ice sheet decline.
    Regards

  42. Since capitalism (Americanism in particular) is the sole cause of global warming — in a roundabout way (we moderns generate too much CO2) — the alarmist must now abandon the greenhouse effect and adopt,
    plant asphyxiation on the Gobi plateau
    , as the cause and blame it all on our living too well? Sounds like a tough job.

    • Indeed. The theory suggests that CO2 does indeed cause interglacial warming – but only by getting so low that all high-altitude plant life dies…! Oh, dear, that is not going to go down well in some quarters.

      R

      • Ralfellis, maybe the model should include two additional items:

        1. a geomechanics module which estimates changes in surface elevation as a function of ice loads. Several years ago I did an estimate using excel and it seemed that increased ice load dropped the ground surface enough to (maybe) change the weather. The changes in ground level are not linear, but the geomechanics equations are fairly easy to use in a coupled dynamic model.

        2. The dust over the ice field may be one of two drivers, the other may be the effect of dust on icean temperature. Not being a climatologist i wouldnt know what a dust cloud blowing from the Gobi over the North Pacific could do. I only have personal observation, where I live we get clouds of sahara dust and it seems to me the air gets hotter, dryer, dustier, AND SUNLIGHT IS DIMMED.

        Hope this helps a little bit.

      • Then there’s the problem of whether it is the dust that is the cause or is it really the wind and the dust merely evidence that there was such a wind? For example…

        Every first of the month the Mullah would cross the border with thirty donkeys with two bails of straw on each. Each time the custom person would ask the Mullah’s profession and the Mullah would reply, “I am an honest smuggler.”
        So each time The Mullah, his donkeys and the bails of straw would be searched from top to toe. Each time the custom folk would not find anything. Next week the Mullah would return without his donkeys or bails of straw. Years went by and the Mullah prospered in his smuggling profession to the extent that he retired.
        Many years later the custom person too had retired. As it happened one day the two former adversaries met in a country far from home. The two hugged each other like old buddies and started talking. After a while the custom person asked the question which had been bugging him over the years, “Mullah, please let me know what were you smuggling all those years ago?”
        The Mullah thought for a few seconds and finally revealed his open secret, “Donkeys.”

      • Wagathon, I’ve always heard that tale as bicycles from France to Spain.

  43. If I didn’t miss something, there isn’t a special hypothesised mechanism for the termination of an interglacial. So there must be some lagging slow moving parameter that controls entropy removal from the, lets say “gaia-sphere”. The same lagging parameter seems to contribute to the inception. The actual mechanisms of cause will probably be different.
    The atmosphere empirically seems to be approaching some kind of maximum entropy producing state. Since the system is highly nonlinear this is to be expected. So the “gaia-sphere” may actually have an own frequency of internally generated glacial fluctuations that happen to be locked to the insolation cycles? And the missing parameter is internal? May the dust e.g. trigger a termination of its own at some point, even without insolation ramp?
    Or some other mechanism?

  44. The glacial cycle is so powerful in its effects that a lot of things oscillate with the same frequency, CO₂, ice, sea levels, albedo, dust levels, temperature gradients, and probably a lot more that we haven’t been able to measure through proxies.

    It is tempting to assign a causal effect to any of them, like dust or CO₂, forgetting that they are oscillating at orbital (Milankovitch) frequencies, and that the only possible reason they are doing so is because they are a consequence, not a cause. That is the reason why Ellis & Palmer 2016, that blames dust, has been pretty much ignored, and why the much better published Snyder 2016, that blames CO₂, was immediately contested by Schmidt et al., 2017.

    This article has several serious problems, and the first one is that it never clearly states that high dust levels are a consequence of orbitally-induced deep glacial conditions, and consequently it never discusses the problem of establishing a correct causal relationship. It presents only the evidence in favor of the hypothesis, while carefully avoiding to discuss the evidence against.

    The article also has several things incorrect. We read things like:

    “It has been found that sharply rising dust levels preceded termination in every case over the last 800,000 years.”

    Every case?

    “Terminations are essentially always preceded by a buildup of dust”

    From “every case” to “essentially always”. What does “essentially always” mean when we are discussing just 11 terminations in 800,000 years?

    Then in the comments we read from the author, Donald Rapp:

    “If the solar power is high enough that it can force a termination without dust, that is perhaps conceivable in a few cases.”

    So we learn that the hypothesis has serious problems because there are several (of 11) interglacials that are not preceded by high dust levels. Obviously they are the ones that are not preceded by deep glacial conditions which greatly strengthens the argument that dust is a consequence, and not a cause. As in the case of CO₂, being a consequence does not preclude being an important feedback that contributes to interglacials, but it rules it out as the main factor, as let’s remember, the glacial cycle oscillates at Milankovitch frequencies and interglacials do occur at times when orbital conditions are favorable.

    Is there a need for dust, or other factor X to explain glacial terminations? Donald says yes because:

    “One cannot invoke rising solar input to high latitudes as the sole cause of terminations of Ice Ages since many such increases in solar input do not produce terminations.”

    This is a strawman argument. Donald claims to have read “many dozens of learned papers on terminations of Ice Ages.” He seems to have missed some very important ones, and my articles at this very same blog were I discussed them.

    The 65°N 21st June insolation as the sole cause of terminations has been challenged multiple times in the literature. It must be remembered that it was not the proposed solution to the glacial cycle by Milutin Milanković, who proposed the caloric half-year, so saying that it constitutes the essence of Milankovitch theory is misleading.

    Donald should read more carefully:
    Huybers, P. 2006. Early Pleistocene glacial cycles and the integrated summer insolation forcing. Science, 313, 5786, 508-511.
    Tzedakis, P. C., Crucifix, M., Mitsui, T., & Wolff, E. W. (2017). A simple rule to determine which insolation cycles lead to interglacials. Nature, 542 (7642), 427.

    By abandoning the wrong 65°N 21st June insolation parameter that was not an original part of it, Milankovitch theory can fully explain each and every glacial termination, not only of the past 800 Kyr, but of the past 2.6 Myr.


    Figure 5. Effective energy at each insolation peak during the past 2.6 Myr. The effective energy is defined in equation (1), and the discount rate of 0.002107 GJ m⁻² kyr⁻¹ is chosen as the maximum a posteriori probability estimate of the statistical model calibrated over the past 2.6 Myr (Methods). Each insolation peak is plotted according to the classification as the onset of an interglacial (red circles), a continued interglacial (black diamonds) or an interstadial (light blue triangles), as in Supplementary Table 1; open symbols correspond to uncertain assignments. The two horizontal lines with a ramp in between show that the simple model is very successful in separating complete deglaciations from incomplete and no deglaciations (see text). The ramp is drawn according to the maximum a posteriori probability estimate of the statistical model (Methods and Extended Data Fig. 10). The labels on the data correspond to either MIS or ages of insolation peaks (kyr bp). Tzedakis et al., 2017.

    So to conclude:

    1. There is no need for a factor X to explain glacial terminations. Milankovitch theory, as originally formulated, or as revised by Huybers 2006, or Tzedakis et al., 2017, provides an explanation for every termination in orbital terms.

    2. High dust levels are the consequence of deep glacial conditions. As such high dust levels follow the 100 kyr eccentricity frequency the same as ice levels. Interglacials (and terminations) do not follow the 100 kyr frequency (11 interglacials in 800,000 years).

    3. High dust levels fail to explain 3 glacial terminations out of 11 (27%). An explanation that fails to explain almost 1/3 of the cases cannot replace a better explanation (Milanković’s caloric half-year, Huyber’s summer energy, or Tzedakis’ effective energy).

    4. Elucidating dust levels contribution to glacial terminations (as well as CO₂ levels contribution) is an interesting scientific problem. No doubt the role of feedbacks is very important to explain how so much ice is melted in so little time. The rest of the feedbacks, like rising sea levels or decreasing albedo, must be added to explain that problem that solar forcing cannot explain by itself.

    Bibliography mentioned:

    -Ellis, R., & Palmer, M. (2016). Modulation of ice ages via precession and dust-albedo feedbacks. Geoscience Frontiers, 7 (6), 891-909.
    -Huybers, P. 2006. Early Pleistocene glacial cycles and the integrated summer insolation forcing. Science, 313, 5786, 508-511.
    -Schmidt, G.A., et al. 2017. Overestimate of committed warming. Nature, 547, 7662, E16.
    -Snyder, C.W. 2016. Evolution of global temperature over the past two million years. Nature, 538, 7624, 226.
    -Tzedakis, P. C., Crucifix, M., Mitsui, T., & Wolff, E. W. (2017). A simple rule to determine which insolation cycles lead to interglacials. Nature, 542 (7642), 427.

    • Javier,

      The Tzedakis et al. paper is interesting and predictively it ‘works’. I also agree that the 65N June insolation is a bit of a red herring, and total caloric summer insolation is better. However they don’t really explain why it works.

      However, as a glacial lengthens, the energy needed for deglaciation decreases. A statistical model that combines these observations correctly predicts every complete deglaciation of the past million years and shows that the sequence of interglacials that has occurred is one of a small set of possibilities.

      Also they look back only as far as 2.5 Mya. An exact match of the current eccentricity cycle with that of today can be found ~2.8 Mya. I suspect their model doesn’t work then as the 41 ky cycle skips a couple of beats.

      • Clive, Huybers 2006 does a better job at explaining with a very similar parameter, summer energy. Melting depends on the number of hours spent above a certain threshold of insolation capable of warming enough to melt the ice.

        When precession factor is very high insolation is very high on the 21st June, but the summer is shorter as the Earth is closer to the Sun and moves faster.

        Obliquity takes energy from the equator and puts it at the poles creating the conditions that up to 0.8 Myr ago were enough to produce an interglacial. Since then obliquity requires also high precession factor at the same time. This creates certain orbital windows when interglacials can take place, defined by high summer energy.

        The requirement for large ice levels build up is an observation that suggests that ice sheet instability is dominated by feedback factors that are larger the bigger the ice sheets. I have no problem with dust being one of those feedback factors that promote melting once it starts due to orbital factors.

    • >>The hypothesis has serious problems because
      >>several (of 11) interglacials that are not preceded
      >>by high dust levels.

      Which ones? All interglacials seem to have ‘high’ dust levels prior to initiation. And do remember that the dust data in these graphs is from Antarctica, because that is just about all there is. Arctic dust levels might have been much higher, but we have no record of that, unless data from the Loess Plateau can come to the rescue.

      .

      >>let’s remember, the glacial cycle oscillates at
      >>Milankovitch frequencies

      No it does not, that is the whole problem. Orbital cycles vary with ~22 ky and ~41 ky cycles, while ice age cycles over the last 40 ky span either 90 ky to 115 ky. This is not the same cycle at all. So you then have to explain why the ice age cycle misses out so many beats in the orbital cycle. The dust-albedo theory can do this very easily.

      .

      >>Milankovitch theory can explain each and every
      >>glacial termination, over the past 2.6 Myr.

      That graph you reproduce is deliberately misleading, which is why I did not like this paper one little bit. They started by plotting the strength of Milankovitch insolation maxima (Great Summers) against their effectiveness in causing an interglacial or interstadial (a proto-interglacial). And the sad result was that there was NO correlation whatsoever. This is the initial result:

      As you can see, there is no correlation whatsoever. However, undeterred, they then introduced a fudge-factor based upon the length of the glacial period. And make no mistake, there is no good reason for this fudge factor, it is merely there to produce the result they wanted. If you torture the data enough, it will give you the result you want. Having applied some thumb-screws to the data, they then achieved this graph:

      And while this fiddled graph looks better, it is still deliberately misleading. Here, they classify the Great Summers numbered 5a, 5c, 6e, 9a and 11a as producing interstadials. But they were hardly interstadials at all. Numbers 5a and 6e are especially dubious, as they were among the strongest Great Summers over the last 800 ky and yet they produced no warming response whatsoever. These Great Summers should really be marked with an unequal cross, marking the grave of a completely failed Great Summer that produced no warming.

      So they now draw a random wiggly line through the graph and declare Great Summers above it are significant and produce interglacials. But there is no reasoning why the strength required for interglacial initiation should suddenly in rise, nor is there any explanation for why Great Summers like 5a and 6e should fail completely. And if a theory cannot explain major anomalies like failed Great Summers, it is not much of a theory.

      But of course the dust-albedo theory CAN explain why Great Summers 5a and 6e failed, it is because dust production had only just begun at that stage, or had not begun at all. The dust theory requires a substantial thickness of contaminated ice, normally 10 ky of dusty ice, before an interglacial can proceed to completion. Great Summers 5a and 6e and many others did not have that layer of dust, so the northern ice sheets remained high albedo white, and the increased insolation of the Great Summer was reflected and rejected.

      A theory that can explain all eventualities, is a successful theory, so the dust-ice-albedo theory is a success.

      Ralph

      • “Which ones? All interglacials seem to have ‘high’ dust levels prior to initiation.”

        By no means. High dust levels are only present at the throughs in eccentricity, coinciding with high ice levels.
        This is an example where it doesn’t work (Clive Best figure).

        “Arctic dust levels might have been much higher”

        You can’t claim evidence you don’t possess. We are going from it is there to it would be there if we have the right record.

        Let’s remember, the glacial cycle oscillates at Milankovitch frequencies

        “No it does not, that is the whole problem.”

        Yes it does. I have showed that:

        It has always been tied to the obliquity cycle and it continues being so. We need to explain why since the Mid-Pleistocene transition is skipping more beats, but we don’t need dust to explain why interglacials take place when they do.

        “As you can see, there is no correlation whatsoever. However, undeterred, they then introduced a fudge-factor based upon the length of the glacial period.”

        No, that is not what I see. What I see is that caloric summer insolation alone explains very well interglacials up to the Mid-Pleistocene transition. But we know that the world has been cooling, and ice sheets became so large in the last 800,000 years that caloric summer insolation was no longer enough by itself to produce interglacials at every obliquity oscillation. That does not mean that it became irrelevant and dust took over the job as you defend. What you call a fudge factor is an observation grounded on evidence. For the past 800 kyr interglacials tend to happen when large ice build up has taken place. This observation has been in many articles for a long time like Paillard 1998.

        “Here, they classify the Great Summers numbered 5a, 5c, 6e, 9a and 11a as producing interstadials. But they were hardly interstadials at all.”

        As I understand the marine isotope stages were defined a long time before Tzedakis et al., 2017, and MIS 5a, 5c, and 5e are interstadials, while MIS 5b and 5d are stadials, so here you are charging against the people that defined marine core stages, like Cesare Emiliani in the 1950’s.

        “But of course the dust-albedo theory CAN explain why Great Summers 5a and 6e failed, it is because dust production had only just begun at that stage, or had not begun at all.”

        You have it backwards, dust levels were low because deep glacial conditions that promote dust production had not been reached. Dust is a consequence, not a cause.

        “A theory that can explain all eventualities, is a successful theory, so the dust-ice-albedo theory is a success.”

        The success of a hypothesis is never in the hands of those promoting it, or they will all succeed. An the ability to explain is a pre-requisite, not a confirmation. Sadly the dust hypothesis is another case of cause-effect confusion.

      • stevefitzpatrick

        Javier,
        “For the past 800 kyr interglacials tend to happen when large ice build up has taken place.”
        .
        Indeed. I have long suspected that the “x-factor” is the slow isostatic adjustment under thick ice sheets, which gradually reduces the altitude of the ice sheet front (increasing it’s summertime air temperature), while at the same time thickening the atmosphere over the oceans (by 100-150 meters). The ever higher ice sheet surfaces become gradually colder, reducing heat loss to space, while more heat is retained by the thicker atmosphere over the oceans. The rise in summertime insolation (the ~22,000 year cycle) then initiates a rapid ice sheet collapse… while isostatic rebound lags thousands of years behind the collapse. When slow rebound has raised surface altitudes sufficiently during an interglacial, and summertime insolation drops (after ~10,000 years), ice sheets begin to form again, re-initiating the cycle.

      • Steve, I liked also your past idea that lower sea levels, coupled with glacier enhanced elevations, acted as a positive feedback, (keeping growing glaciers ever colder.) Do you still hold that thought?

        Overall, the dust theory makes sense and has physical evidence supporting it and a substantial negative feedback theory is required to explain, in addition to orbital M-cycles, the glacial cycle pattern since the MPT.

      • >>High dust levels are only present at the troughs
        >>in eccentricity, coinciding with high ice levels.

        Yes, dust is only generated when CO2 is low – coinciding with high ice levels. All very true.

        .

        >>Example where it doesn’t work (Clive Best figure).

        But it does work in your Clive Best diagram. All we need is dust before the interglacial or proto-interglacial temperature rise, and there was dust just before the temperature rise in both cases. Take a look at Donald Rapp’s Fig 10, where the laser-counter method shows up the dust peaks much better.

        And do remember that this is an Antarctic dust record. We do not have a record of what happened in the NH, as that record does not exist. And we know from the D-O temperature fluctuations during the last ice age, that the SH does not react so quickly to NH temperature fluctuations. Many of the dramatic D-O events are completely missing from the SH record. So a small dust-blip in the SH, as recorded here, may actually mask a very large dust-blip in the NH.

        Ralph

      • >>I understand the marine isotope stages were defined
        >>a long time before Tzedakis et al.

        Yeah, but these are not MIS numbers – they have attached them to the Milankovitch insolation plot, not the LR04 marine isotope plot. Look at numbers 9a and 9e, which are nowhere near the temperature maxima. So they have generated their own notation, which bears no relationship to MIS. And as I said before, 5a and 5c do not equate to any warming whatsoever.

        .

        >>You have it backwards, dust levels were low because
        >>deep glacial conditions that promote dust production had
        >>not been reached. Dust is a consequence, not a cause.

        Have you even bothered to read my paper, Javier..? This is EXACTLY what the paper says.

        Yes, low CO2 is a consequence of cooling, and high dust is a consequence of low CO2. But interglacial warming is a consequence of high dust – so in the final analysis dust is causal of warming, not consequential.

        Ralph

      • stevefitzpatrick

        Ron Graf,
        Yes, the the 100 K-year pseudo-cycle seems to me likely due to a combination of these effects. (Heck, maybe dust even contributes!) When Ice sheets become high enough, their surface temperatures drop and heat loss to space also drops, while at the same time oceans have more atmosphere above them and heat loss is reduced. Combine this with the multi-thousand year lag in isostatic response, and you have a a process likely to oscillate. Of course, the warming influence due to CO2 from fossil fuels will probably delay the next ice age by some centuries (or even millennia)…. which may turn out to be a good thing for humanity in the long run.

      • “But interglacial warming is a consequence of high dust – so in the final analysis dust is causal of warming, not consequential.”

        Interglacials preceded by low dust contradict that statement.

      • Javier, Ralph, Clive, are their any examples after the MPT where the the following obliquity insolation peak is accompanied by NH summer precession peak and interglacial fails? And, if there are such examples then doesn’t it mean we need to exclude that M-cycle insolation can independently trigger interglacial? And, if that is true then M-cycle’s reinforcement would need to be a negative feedback of reaching a glacial peak, or CO2 minimum. And if CO2 minimum’s direct effect is known certainly to be positive, it would need to be an indirect effect, like dust. Right?

        Yet, I feel that the dust does not need to exclude the possibility of other indirect negative feedbacks, like gravity causing a critical pressure at the base of the glaciers, leading to collapse due to ice-skater effect (melting at point of ground contact). These seems less supported then the dry dust theory, which directly affects albedo, and for which there is now physical evidence. I applaud Ralph’s paper.

    • Thanks Peter.

      You can see:
      How low the albedo of ice sheets can go, given a little dust.
      How much extra insolation can be absorbed.
      The tenacity of dust and soot to stay on the surface, rather than being washed away.

      And so this increased insolation absorption will continue, decade after decade, century after century.

      R

      • The first scientific expedition to the Arctic was by Scoresby in 1820, sponsored by the Royal society in order to discover why the arctic ice was melting so substantially.

        In his journals he writes that he could observe large amounts of black soot covering the snow and ice. He attributed it to the developing economy of America

        tonyb

      • Javier,

        I asked you on your last post if you could explain what conditions existed at the end of the Permian ice age that allowed Earth to escape from the 70 Ma icehouse phase, so that GMST then warmed from 13C to 28C over the next 40 Ma. Perhaps you missed my comment. Can you suggest an explanation? It is relevant because it might help us to understand what conditions would be required for Earth to escape the current icehouse .., and calculate when it might happen. I suspect the locations of the continents and ocean gateways prevents us getting out of the icehouse phase for tens of millions of years.

    • “Some photos of black soot on Greenland”

      Yet Greenland ice-sheet surface mass balance has increased greatly over the past two years and increases nearly every year. Greenland ice mass loss is through icebergs.

      This should open some eyes. The purported effect doesn’t work as advertised. Despite all the dust there is no Greenland surface melting.

      • [Repost in correct place]

        Javier,

        I asked you on your last post if you could explain what conditions existed at the end of the Permian ice age that allowed Earth to escape from the 70 Ma icehouse phase, so that GMST then warmed from 13C to 28C over the next 40 Ma. Perhaps you missed my comment. Can you suggest an explanation? It is relevant because it might help us to understand what conditions would be required for Earth to escape the current icehouse .., and calculate when it might happen. I suspect the locations of the continents and ocean gateways prevents us getting out of the current icehouse phase for tens of millions of years.

      • Black soil spread on snow melts the snow. Canadians do this on tennis courts to make the snow melt quickly in spring. So of course the black soot on the ice in Greenland is melting it.

        Judith Curry’s post Sea level rise acceleration (or not). Part V: detection & attribution explains that most of the sea level rise is due to melting of Greenland. Curry says:

        Figure 4 from Chen et al. shows the time series of the budget components. The steric component of the rate of sea level rise (related to ocean temperature increase) shows a lower value in 2014 than in 1993 (apparently associated with the lingering effects from the Pinatubo eruption). The increase in the rate of sea level rise is largely attributed to the Greenland ice sheet.

      • >>Greenland ice mass loss is through icebergs.
        >>This should open some eyes. The purported effect
        >>doesn’t work as advertised. Despite all the dust
        >>there is no Greenland surface melting.

        You are flailing wildly, once again, and entirely wrong. You are in danger of making all your clams and arguments appear suspect. Back in the real world, this recent paper says:

        Quote:
        The relative contribution of ice discharge to total loss decreased from 58% before 2005 to 32% between 2009 and 2012. As such, 84% of the increase in mass loss after 2009 was due to increased surface runoff. These observations support recent model projections that surface mass balance, rather than ice dynamics, will dominate the ice sheet’s contribution to 21st century sea level rise.

        Quote:
        As such, 84% of the increase in mass loss after 2009 was due to increased surface runoff. These observations support recent model projections that surface mass balance, rather than ice dynamics, will dominate the ice sheet’s contribution to 21st century sea level rise.
        https://agupubs.onlinelibrary.wiley.com/doi/epdf/10.1002/2013GL059010
        https://agupubs.onlinelibrary.wiley.com/doi/full/10.1002/2013GL059010

        In other words the amount of Greenland ice sheet melting and ablation has INCREASED, relative to iceberg calving. The paper does not ascribe a reason for this, but one might postulate this is due to Chinese industrial dust and soot, because there has been an observed reduction in ice sheet albedo recently.

        Ralph

      • “You are flailing wildly, once again, and entirely wrong.

        Back in the real world, this recent paper says

        In other words the amount of Greenland ice sheet melting and ablation has INCREASED

        Apparently you have a problem distinguishing between the relative contribution of Greenland’s surface mass balance to sea level rise, and its change.

        At the Danish Meteorological Institute (DMI), they are experts in ice and they determine the surface mass balance every day of the year. They start the season on September 1st and on average they end the season with ~ 400 Gt of accumulated surface mass balance (see figure in my post above). So they have this to say about it:

        “Note that the accumulated curve does not end at 0 at the end of the year. Over the year, it snows more than it melts, but calving of icebergs also adds to the total mass budget of the ice sheet.”
        https://www.dmi.dk/en/groenland/maalinger/greenland-ice-sheet-surface-mass-budget/

        That surface melting is increasing and will contribute more to sea level rise does not mean that surface mass balance is negative or will be so within the 21st C. Greenland surface mass balance is very POSITIVE (~ 400 Gt/year). The great amount of dust and carbon observed in Greenland is not melting the surface of the GIS. It is being buried by snow.

        For having published an article on this issue it is surprising how little you know about Greenland surface mass balance. It casts a doubt on what you say in the article that you don’t know the effect of dust on Greenland’s ice sheet surface mass balance today.

      • >>Over the year, it snows more than it melts, but calving
        >>of icebergs adds to the total mass budget of ice sheets.

        Yes, we know that Javier. So berg calving is one component – a smaller component – of Greenland ice sheet mass balance. So how does that square with your claim that:

        Quote:
        Greenland ice mass loss is through icebergs.

        Clearly the greater ice sheet loss is through ablation and melting at the periphery of the ice sheet, not through iceberg calving. And while you are at it, look at the dates of these graphs. The paper I quoted gave the two decades up to 2014. Your graph is for mass balance increase is for the year 2016-2017. So the Greenland mass balance rose for one year. Wonderful – but one year proves nothing.

        Ralph

      • You still have failed to notice the grey line that shows the mean (1981-2010) surface mass balance. On average Greenland surface gains ~ 400 Gt of ice every year. The final balance is positive because of iceberg calving (as the DMI clearly states in their page).

        Without iceberg calving Greenland would be gaining ice every single year, because despite all the dust its surface mass balance is very positive.

        Dust doesn’t appear as the hand that moves the planet out of glacial periods. Too weak for that. I concede that it might have an important contribution when it is present, that is not always.

      • “Quote: The relative contribution of ice discharge to total loss decreased from 58% before 2005 to 32% between 2009 and 2012. As such, 84% of the increase in mass loss after 2009 was due to increased surface runoff.”

        This seems to be peculiar to that 2009-12 period, given what happened 2012-18. Unless it can somehow be shown 2012-18 period had far less dust, this hypothesis doesn’t seem to have held up any better than the “imminent ice-free Arctic summers” predictions of a decade ago.

      • Javier says:

        The great amount of dust and carbon observed in Greenland is not melting the surface of the GIS. It is being buried by snow.

        Clearly, that statement is nonsense. Look at the photos. That black soot on the surface did not accumulate in one year. It accumulated over many years or decades.

        Also look at the third photo. It shows a fissure into which melt water has been flowing. There is no signs of annual layers of black soot like the layer at the surface. Any trained observer should be able to recognise and correctly interpret what this means.

      • Ralph, am I correctly following your hypothesis?

        Glacial advance: NH snow > (glacial ablation + calving)
        therefore increasing albedo, therefore cooling GMST, therefore declining precipitation

        Glacial maximum: NH snow = (glacial ablation + calving)
        therefore stable albedo, therefore GMST minimum, therefore precipitation minimum and CO2 minimum, therefore life minimum, therefore desert dust maximum. Stability continues until NH insolation maximum due to M-cycle.

        Glacial collapse: NH snow << (glacial ablation + calving)
        therefore dust (soil) remains exposed on surface keeping albedo low, therefore continued positive feedback for warming through ~5000 year M-cycle peak where increasing precipitation can be overpowered by amplified summer insolation.

        Interglacial: NH snow < (glacial ablation + calving), gradual decline in albedo matched by gradual decline in NH insolation, therefore stable warm GMST, therefore CO2 recovery and increased precipitation, therefore life recovery, therefore low dust.

        Interglacial termination:
        The low dust and declining NH summer insolation increases the chances for circumstances, (i.e. large volcano or asteroid strike,) where a fresh, high albedo snow makes it through a summer without completely melting leading to ever higher albedo and GMST drop.

        Question for anyone: during declining GMST (or negative radiative imbalance) shouldn't there be more precipitation due to warmer ocean oceans putting out more vapor than the cold atmosphere wants? Wouldn't this also be a positive feedback for snow accumulations?

      • More or less correct. Except that interglacial termination is caused by reducing M-insolation in the NH (a new Great Winter). This is why it is uncertain we will get a new ice age niw, because there is no deep NH Great Winter for the next 100 ky or so. It depends on what insolation levels trigger a new ice age. We are working on that now.

        See also the short summary of the cycle in Rapp’s article at the top of the page.

        Modulation of Ice Ages via Dust and Albedo
        https://www.sciencedirect.com/science/article/pii/S1674987116300305

        R

      • “Clearly, that statement is nonsense. Look at the photos. That black soot on the surface did not accumulate in one year. It accumulated over many years or decades.”

        Not sure you’ve thought this through. How can a surface that is gaining an average of 400Gt of snowpack a year accumulate dust on top of the snowpack over years or decades? Seems vastly more likely that the dust in those photos was gradually exposed during a seasonal melt (snow melts and evaporates, dust does not).

        If you don’t live in a snowy area, you might not realize that as snow melts, it gets dirtier, even though you can’t see “layers” of dirt inside.

    • I think a missing factor here is weather variability. These soot covered ice images were taken after a major melt event, but Greenland receives regular snowfalls in summer and I would bet that this soot was covered in white high albedo snow multiple times ceasing melting. Perhaps the answer is precipitation falls so low at the end of glaciations that these snow covering events are not enough to cover the melting events?

      • “These soot covered ice images were taken after a major melt event, but Greenland receives regular snowfalls in summer and I would bet that this soot was covered in white high albedo snow multiple times ceasing melting. ”

        Yes, the soot would be covered almost every year. But clearly, it has melted back to this level many times and the soor collecting at all levels above is collecting at this level when the snow and ice melts back to this level … and continues to melt down as the soot stays on top. Clearly this soot has been collecting for many decades, or centuries, and clearly there are no soot layers below, so clearly, the ice is melting down further each year as a consequence of this black layer.

  45. Ralph Ellis: In truth even though the science is settled, nobody knows what causes Dansgaard–Oeschger events

    You might be interested in my blog post on this…
    https://tallbloke.wordpress.com/2018/05/19/possible-origin-of-dansgaard-oeschger-abrupt-climate-events/

    • Indeed, there are a few resonances and coincidences like that. The ancient Egyptian Sothic Cycle was also 1460 years, but that is calendrical, not orbital. But the bottom line is, how can an orbital alignment trigger a climatic event? How can it effect climate?

      The theory that D-O events are caused by continent-wide wildfires can indeed explain a climatic effect. Continental wildfires will provide rapid Arctic warming and melting, via the ice-sheet albedo effect. And they will also require time to rebuild a fuel-load before another fire can erupt. However, the possibility that there may be a steady cycle underlying D-O events is more problematic. If that cycle is real, and not another artificial construct of a fourier analysis (like the supposed 100 ky ice age cycle), then you also need a trigger than can cause wildfires. And that is more difficult.

      R

      D-O events closely linked to fire combustion products.
      (bottom purple plot)

    • Oldbrew,

      The 1470-year periodicity in Dansgaard–Oeschger events is based mainly on the original GISP2 timescale. When using the more recent GICC05 age model that is considered superior, the strong 1470-peak weakens considerable and shifts to ~ 1500 years. Nowadays most authors do not accept the 1470-year periodicity.

      • Citing IPCC author Stefan Rahmstorf may not be to everyone’s taste here, but never mind…

        Timing of abrupt climate change: A precise clock —
        Potsdam Institute for Climate Impact Research, Potsdam, Germany

        An analysis of the GISP2 ice core record from Greenland reveals that abrupt climate events appear to be paced by a 1,470-year cycle with a period that is probably stable to within a few percent; with 95% confidence the period is maintained to better than 12% over at least 23 cycles. This highly precise clock points to an origin outside the Earth system; oscillatory modes within the Earth system can be expected to be far more irregular in
        period.

        http://www.pik-potsdam.de/~stefan/Publications/Journals/rahmstorf_grl_2003.pdf

      • I wonder if the authors that don’t ‘accept’ 1470 years D-O period have just given up hope of finding a reason for it. I’ve offered a simple enough possible mechanism for theorists to look at.

        This paper has a good discussion of the problem, and 79 references to ‘1470’ !
        On the 1470-year pacing of Dansgaard-Oeschger warm events
        Michael Schulz
        Institut fu¨r Geowissenschaften, Universita¨t Kiel, Kiel, Germany

        https://www.geo.uni-bremen.de/geomod/staff/mschulz/reprint/1470y_pacing.pdf

      • Yes, but Rahmstorf is wrong on that as on so many other things. I also cited him when I wrote my article on DO-events, but have learned since that the original GISP2 timescale was wrong.

        This is what Stephen Obrochta had to say about this issue in 2015:
        “GICC05 and the North GRIP (NGRIP) ice core strongly suggest that the GISP2 chronology contains significant inaccuracies, and thus NGRIP superseded GISP2 as the key last-glacial reference series. GISP2 on it’s original chronology (Meese-Sowers) exhibited an inconsistent climate/annual layer thickness relationship; Svensson et al. (2006) note that: “the existence of the proposed 1,470yr cycle depends on the exact timing and phasing of the onset of D–O events, and … this is exactly where we believe that the GISP2 time scale is inaccurate.” Furthermore, the GISP2 age scale places the Toba Eruption about 4,000 years too early (e.g., Svensson et al., 2013).
        The GICC05 chronology has been transferred to GISP2 though high-precision volcanic synchronization with NGRIP (Seierstad et al., 2014). This largely corrects the inconsistent climate-layer thickness relationship of GISP2 (Obrochta et al., 2014a) and dras- tically changes the characteristics of the time series.”

        It is a problem when a hypothesis depends on the timescale of a specific ice core that is seriously questioned. Besides the Holocene and Late Glacial evidence is consistent with ~ 1500 years, not ~ 1470 years. The 300 years accumulated difference over the entire Holocene is too large to be accommodated.

      • This highly precise clock points to an origin outside the Earth system; oscillatory modes within the Earth system can be expected to be far more irregular in

        Internal natural frequencies are determined by internal mass and spring rate, these are expected to correlate with internal events much better than external forcing that has no way to get feedback from the internal system. Correlations are mostly just sometimes as they resonate with and against internal cycles. Ice cycle frequencies change when the mass of ice that is sequestered does not all return to the oceans or the spring rate changes as warmer oceans cause more snowfall.

    • Ice age ice sheets extended into middle latitudes. The ice does thaw and meltwater builds. The meltwater gathers enough water and energy to promote snowfall and it snows again and rebuilds ice from the meltwater. These cycles occur in the Northern Hemisphere and shows up in Greenland ice core data. These cycles do not show up in Antarctica ice cores. This is not from external causes this is from internal dynamics on the Northern Hemisphere continents. Ice core data in Greenland recorded the events. More snowfall in Greenland had to come from meltwater because the oceans were still low and cold. Meltwater is being transferred to Greenland and some to Antarctica. Ice is sequestered and not returned to the oceans and is not there for another major warm cycle or major ice cycle.

      Oldbrew, think about this. Your friend, Alex Pope, London conference, two years ago.

  46. The only thing I haven’t explained is how they get the date of the end of the Ice ages. It has to do with the ice melting from the top down. They say the last Ice Age ended about 18 thousand years ago. I say that point was ice deposited before the peak of the ice making and which was about 80+ thousand years ago.
    At the peak of the rise of the ice layer, when the Ice melting stage began, the melting ice left all the solids laying on the ice as the water and CO2 departed into the atmosphere. This was a sentence in my explanation. The Ice melting stage began lowering the ocean and the land surface area began to groe. The foliage was not enough to overcome that CO2 produced by nature. Another few thousand years later enough green foliage was killed and the CO2 level in the ice gegan to drop.

  47. Tzedakis et al. (2009) tried to find an explanation for the diversity of past Interglacials in terms of solar input alone but no correlation could be found. They concluded:
    What emerges is that astronomical forcing alone cannot explain the difference in interglacial intensity before and after the mid-Brunhes event (430 kya).
    Tzedakis et al. (2009) compared properties of ten presumed Interglacials but no definitive results were found, other than the obvious distinction between those prior to the mid-Brunhes event (MBE) being less warm and of greater duration.
    Tzedakis et al. (2012) attempted to fill the need for a theory of Interglacials. They pointed out once again, that there were significant differences in the duration of Interglacials in the past, and that “a theoretical framework with predictive power for interglacial duration has remained elusive”. The duration of an Interglacial is related to the period of high sea level, and this can be interpreted in various ways. They proposed “that the interval between the terminal oscillation of the bipolar seesaw and three thousand years (kyr) before its first major reactivation provides an estimate that approximates the length of the sea-level high-stand, a measure of interglacial duration.” They went on to elaborate at length on this approach, but I can’t make much sense of the methodology.
    Tzedakis et al. (2017) observed that many peaks in solar input to high latitudes do not produce Interglacials. Although they did not cite Ellis and Palmer (2016), their paper seems to be the only one after Ellis and Palmer to come to this obvious realization. Evidently, Tzedakis et al. (2017) understood that there is something missing in the conventional Milankovitch theory which attributes terminations solely to increases in solar input to high latitudes. They attempted to adapt the conventional Milankovitch theory to the data so that the occurrence of Interglacials falls out naturally from the theory, based solely on solar input to high latitudes. Dust was not included as a factor.
    Tzedakis et al. (2017) based their analysis on integrated summer solar input to 65°N latitude (S1). This varies with the 22,000-year cycle of peak summer solar input, but its pattern of variation is different from the peak solar insolation in mid-summer (S2). While the two measures of solar input to high latitudes are similar in some ways, the relative heights of various peaks are different in the two cases.
    Tzedakis et al. (2017) argued that the use of S1 is more appropriate than S2 for such a study but their justification for this conclusion is not clear to this writer. Evidently, S1 and S2 provide independent, but related measures of solar input to 65°N, and it does not appear to this writer that one or the other has exceptional a priori merit. If S1 is used, it represents most of the caloric input from the Sun during summer for one year. Yet, the effects of dust on albedo are more likely to be important at midsummer when the Sun is highest in the sky. The problem with using the integral of solar input is that the oscillations are so small that one must invoke extremely small variations above a hypothetical threshold to obtain any differentiation between events. There simply isn’t enough variability in the integral from cycle to cycle to cause glacial cycles.
    Like essentially all papers on Milankovitch theory, Tzedakis et al. (2017) is not an analysis of physical phenomena. No description of energy flow, ice sheet energy balance, or ice sheet dynamics is included. Nor is any account taken of ice sheet surface conditions, albedo or absorptivity. Rather, it seeks to find statistical correlations between the timing of the solar time-pattern with the ice volume-time pattern, and if such a correlation can be made, they interpret that as a probable cause-effect relationship.
    The fundamental postulate of Tzedakis et al. (2017) is that along the timeline where S1 varies across paleo-time, when a peak in S1 exceeds a very narrow threshold, a termination will occur. Since S1 varies cyclically with a roughly 22,000-year period, the width of the up-lobe in S1 from half height to half height is roughly 8,000 years. There might be some merit to this viewpoint if the magnitudes supported it, for it might be roughly hypothesized that when the summer melt passes above a significant level during the ~8,000 years of higher S1 during an up-lobe in solar input to high latitudes, the ice sheets cannot recover in the following winter, and a termination is initiated. And if the summer melt during such an ~8,000-year period is determined by the magnitude of S1 during that period, this would generate a model for solar-induced terminations without necessarily invoking the role of dust on ice sheets. But the problem with this, is the fact that S1 changes only slightly from cycle to cycle, which then requires an extreme sensitivity of the ice sheet threshold to extremely small changes in S1, which does not make sense physically.
    While Tzedakis et al. (2017) attempted to correlate all the peaks in S1 over the past 2.7 million years, I only concentrate here on the last 800,000 years. In the initial correlation derived by Tzedakis et al. (2017) for the past 800,000 years (their Figure 2d and part of Figure 2c) they indeed found that the most of the highest peaks in S1 were associated with terminations. However, many high peaks in S1 did not lead to terminations. These included the following peaks – as labeled by Tzedakis, et al. (2017):
    3, 5a, 5c, 6e, 7a, 9a, 11a, unlabeled at 462 and 449 kya, 13c, 15c, unlabeled at 571, 712, 737, 747 kya.
    Based on their results for the last 800,000 years, they somewhat arbitrarily set a threshold for termination at S1 = 5.945 GJ/m2. The differences in S1 from cycle to cycle seem to be too small for the physics of melting ice sheets. It is difficult to imagine the physics of a geological ice sheet system that is so extremely sensitive that the stability of the ice sheets would distinguish so drastically between such small variations in S1 to determine in its geological outcome. That an input of 5.95 GJ/m2 would lead to a termination but an input of 5.94 GJ/m2 would not, seems quite incredible to this writer. And if that hypothesized sensitivity were real (which I strongly doubt) why do some sub-threshold values of S1 lead to terminations?
    Tzedakis, et al. (2017) attempted answer this question by refining their model by making the assumption that the ice sheets become less stable as they expand with time. However, this notion is rather vague, and based on unproven hypothetical models. According to this hypothesis, the threshold for a termination is reduced at each peak in S1 by a factor proportional to the time elapsed since the previous interglacial, assuming the total ice buildup is proportional to that elapsed time. This assumption automatically divides the peaks in S1 into two categories:
    (1) One group has long time periods since the last deglaciation – but these are just the ones that produce terminations. These are assigned a lower threshold for termination since by assumption, the threshold is reduced by a factor proportional to the time elapsed since the previous termination.
    (2) Those peaks that occur between terminations are assigned higher thresholds since the time elapsed since the previous threshold is less. Hence, they don’t produce terminations.
    These assumptions have some merit (in principle) because they provide a clearer delineation between solar peaks that cause termination and those that don’t. As I pointed out previously, Tzedakis et al. (2017) almost stumbled on the answer, but they were so locked in to the belief that solar input to the NH (and only that) controls the Ice Age cycles, that they missed the only credible choice of the X-factor: dust. The whole notion of a threshold for solar is misleading. If it were only a matter of solar input to the NH exceeding a solar threshold, there would be no explanation for the fact that so many solar peaks do not produce terminations. Realizing this, they invoked the assumption that the duration of time since the last termination was important in setting the threshold. In doing this, they were rubbing elbows, so to speak, with a more credible answer. The duration of time since the last termination is indeed important, but not because it lowers an ephemeral solar threshold by a tiny amount, but rather because it leads to the depth of the glacial maximum, which in turn, develops high dust levels that get deposited on the ice sheets, thus increasing the solar absorption. Furthermore, by using the integral of solar intensity over a summer instead of peak solar intensity at mid-summer, they muted the range of solar variations from cycle to cycle.
    But what does it mean to say that the ice sheets become “less stable” as they grow? There are some speculative theories about this. Perhaps a simpler (yet unproven) explanation is this. As an Ice Age matures, and the ice sheets expand, at least two things occur on a broader geological scale:
    (1) As Ewing and Donn (1956) suggested, an ice covered Arctic Ocean severely reduces evaporation and sublimation, thus greatly reducing precipitation in polar areas, allowing high-latitude snowfall to melt during the summer.
    (2) Dust builds up sharply in the atmosphere leading to great increases in deposition on the ice sheets.
    During the course of an Ice Age, dust levels in the atmosphere build up with time. Toward the end of an Ice Age, in the depths of the coldest decade or two, when CO2 levels dip below 200 ppm, precipitation is reduced and dust deposition is sharply increased on the ice sheets. This makes the ice sheets more susceptible to extreme summer melt.
    Paillard (2017) lauded Tzedakis, et al. (2017) for “discovering a rule that links astronomical forcing to the timing of the Quaternary Interglacials” and indeed I suppose they have, but it is a statistical manipulation lacking a physical basis. Paillard indicated that at least three physical mechanisms have been proposed to explain the statistical finding that as Ice Ages get older, they become more susceptible to termination. According to Paillard, these include:
    1. Instabilities in large ice sheets: for instance, the bases of such ice sheets might become warm, inducing faster ice flow and increased iceberg calving.
    2. Ice-sheet albedo (reflectance) in a cold and dry glacial climate: for large ice sheets, less snow and more dust increases the surface absorption of solar heat and enhances ice-sheet melting.
    3. Bottom-water formation and ocean carbon storage, which might be directly linked to the rise in atmospheric CO2 that precedes deglaciations.
    It is also possible that all three of these mechanisms act together to ultimately lead to the simple dynamics outlined in Tzedakis and colleagues’ paper.
    I believe that the evidence strongly points to the second possibility.

    • R
      One thing that Javier’s articles here have established clearly is an obliquity based explanation of interglacial timing since the MPR (mid Pleistocene revolution) which is compelling and sufficient.

      Javier showed that, since the MPR, there is an exact correlation between the the obliquity cycle and temperature lagged by 6,500:

      Every 2 or 3 obliquity peaks – after a 6,500 year delay from the ocean’s thermal inertia, an interglacial begins. No exceptions.

      And it turns out that the question as to why not every obliquity peak induces an interglacial is not horrifyingly intractable as sometimes claimed, but is very straightforward. It is simply predictable from precession and eccentricity. Regarding precession, it is not precession per se but the modulation of precession that is the critical factor, that is, the oscillation in the amplitude of the precession peaks. This modulation follows eccentricity – the maximal peaks of precession modulation occur at the peaks of eccentricity.

      It is obvious from orbital considerations why precession modulation should follow eccentricity. And in turn, insolation at 65N follows precession and the modulation of precession. In fact it simplifies analysis of Milankovich forcing to consider three parameters, precession modulation, eccentricity and insolation at 65N as one and the same phenomenon. We could call this the “triumvirate” Milankovitch forcing. When an obliquity peak is close to a triumvirate peak, an interglacial is triggered. If the peaks are not close, then there is only an abortive “stump” of an interglacial, while essentially glaciation persists.

      And that’s all it is. An obliquity peak – thermally lagged by 6,500 years, coinciding with a peak of precession modulation/eccentricity/65N insolation causes an interglacial. Due to phase shifting, sometimes a peak of this triumvirate will fall exactly half way between two obliquity peaks. In this case you get a double-headed interglacial, as occured 200k and 600k years ago and will happen again 200k years in the future.

      It is clear that nonlinear and chaotic dynamics play a role in the glacial cycle. The timing of interglacials indicates that the system is a periodically forced nonlinear oscillator, with the dominant forcing being obliquity. This nonlinearity is enough to explain why exact amplitudes of forcing don’ always match. As Tsedakis pointed out, for example, the 400kya MIS 11 interglacial terminated with insolation still relatively high due to a low amplitude node of eccentricity, as now – but glacial inception followed anyway: the interglacial “was not prolonged by weak insolation forcing”. That’s nonlinear for you.

      Here is Javier’s recent article on glacial inception:

      https://judithcurry.com/2018/08/14/nature-unbound-x-the-next-glaciation/

      • Obliquity doesn’t melt ice. Solar power does. Solar power depends on obliquity, eccentricity, and precession. They all act together, not singly.

      • Obliquity doesn’t melt ice. Solar power does. Solar power depends on obliquity, eccentricity, and precession. They all act together, not singly.

        Major ice ages have ice sheets that are extended into the middle latitudes. There is always enough solar power there to thaw ice every summer. It does not matter which mode Milankovich is in. Ice ages end when enough ice has thawed to thin the ice sheets enough that they can retreat. Last Glacial Max was max ice extent but it was long after max ice volume. Max ice volume was when ice extent was increasing the fastest.

      • Javier showed that, since the MPR, there is an exact correlation between the the obliquity cycle and temperature lagged by 6,500:

        That is really very funny, none of these cycles look anything like the ice age profiles.

      • exact amplitudes of forcing don’ always match

        Yep, when there is enough ice an ice age happens, when there is not, it does not happen.

        When enough ice thaws and depletes, and ice age ends, when there is not, it does not happen.

        An ice age is caused by ice and an ice age ends due to not enough ice.

      • popesclimatetheory

        That is really very funny, none of these cycles look anything like the ice age profiles.

        EPICA dome 3. Yes, the one they’re trying to bury (according to Stephen Macintyre).

    • 1. Instabilities in large ice sheets: for instance, the bases of such ice sheets might become warm, inducing faster ice flow and increased iceberg calving.
      2. Ice-sheet albedo (reflectance) in a cold and dry glacial climate: for large ice sheets, less snow and more dust increases the surface absorption of solar heat and enhances ice-sheet melting.
      3. Bottom-water formation and ocean carbon storage, which might be directly linked to the rise in atmospheric CO2 that precedes deglaciations.

      While agreeing with you that the second possibility is the most likely, I want to point out that there is a possible 4th effect.

      4. Increasing tides combined with falling sea levels help break up Arctic ice sheets.

      The increase in obliquity has another effect. It changes the amplitude of the lunar orbital precession. The lunar orbit is inclined at 5 degrees to the solar ecliptic and precesses with a Lunar standstill period of 18.6 years. The maximum declination of the moon at ‘lunar standstill’ is currently 28.5 degrees and this occurs every 18.6 years. However 7000 years ago the earth’s tilt was 24.3 degrees compared to the current 23.5 degrees. The obliquity is also currently decreasing so that in 12000 year time it will be only 22.6 degrees.
      This means that lunar standstills reached 29.1 degrees around LGM, while in 12000 years time they will reduce to only 27.4 degrees. This significantly changes the maximum strength of tidal flows at large latitudes during a lunar standstill. It is the tractional (horizontal) component of the moon’s gravitational tide which draws the oceans into a tidal bulge. This maximum tractional force occurs at about 45 degrees to the tidal bulge, thereby increasing the average traction acting on polar regions. With sea levels 100m lower than today at LGM causes much stronger tidal currents acting under and upon ice sheets.

      see Ocean Tide Influences on the Antarctic and Greenland Ice Sheets

  48. The dust theory is imaginative and innovative, and regular readers here will already be familiar with its energetic promotion by Ralf. I have to say that while the explosions of Dustin was at glacial maxima are incontrovertible, I am skeptical of dust as a necessary and sufficient cause of glacial termination.

    In this I tend to agree with Javier’s long post just above. It is far from proved that abrupt glacial termination needs dust.

    It is already well established that ice sheets become unstable as they get very large. For instance:

    https://iri.columbia.edu/~blyon/REFERENCES/P86.pdf

    Dozens more such papers are out there.
    Runaway albedo driven glacial termination can be started due to stability (e.g. Lyapunov) dynamics alone, with no need for dust.

    The dust doesn’t require CO2 starvation. Yes if we are confident that CO2 dipped below 200ppm then there was indeed some starvation. However extreme cold climate will bring aridity and dust with or without CO2 starvation.

    And finally, I am skeptical that airborne dust would change albedo enough to cause glacial termination. Dust is deposited in discrete weather patterns, not continuously. Any snow will bury it. Albedo warming of a dusty snow surface would melt the surface and largely remove the dust – or at least diminish its surface coverage – as an immediate negative feedback. I would be surprised if even global dustiness, specifically in its effect on snow and ice, would change global albedo by as much as 1%.

    Thus in summary the dust theory of glacial termination is useful to inform people about glacial termination aridity and CO2 starvation (how can there be “starvation” of a pollutant? :-) but ultimately insufficient and unnecessary for glacial termination.

    • Erratum
      First paragraph
      Explosions of dustiness

    • >>It is already well established that ice sheets become
      >>unstable as they get very large. For instance:

      … But why would instability lead to a collapse of an ice sheet, rather than a retreat back to a previously stable dimension?
      … Why are (all) ice age terminations coincident with dust deposition?
      … Why do many orbital insolation maxima produce no warming or ice dissipation at all?
      … How can ‘instability’ terminate an entire ice sheet in just 5 ky?
      … Many of these ‘instability’ theories cite slippage over the surface, when it is a fact that large ice sheets actually flow over mountains. So should surface friction and ‘slippage’ calculations be replaced by viscosity?

      So many questions.

      .

      >>The dust doesn’t require CO2 starvation … extreme cold
      >>climate bring aridity and dust with or without CO2 starvation.

      But much of the Gobi was actually wetter during the LGM than now. Look at the PMIP3 climate models for the LGM in the paper. This is not simply model data, as there is evidence that presently dry basins in the Gobi region were shallow lakes during the LGM.

      Plus we have the required ingredients for desert formation here. Low CO2 concentrations combined with high altitude and low temperatures, which is detrimental to C3 flora and not conducive to recolonisation by C4 flora. These conditions can indeed form CO2 deserts, so why dismiss them in favour of traditional desert formation?

      R

  49. This is an impressively clear, persuasive, and interesting explanation for someone like me, outside the field. Many thanks to Ellis, Palmer, Rapp, and Curry. It is great that the Ellis & Palmer 2016 paper is open-access.

    I have to admit to being confused by a minor point, which is the cause of low CO2 concentration during glacial maxima. The Ellis & Palmer paper says it is poorly understood:

    “But the reason for the CO2 reduction and its possible connection to the increase in dust is poorly understood.”

    In contrast, Rapp’s writeup states the cause quite definitively as “oceanic cooling = oceanic CO2 absorption”.

    Beyond that, the anti-correlation between antarctic CO2 and log(Dust) is amazingly good, as shown in Fig. 9 of Ellis & Palmer, 2016. The anti-correlation is so good that it seems like it must be either an artifact of the measuring instruments, or a law of climate science.

    • Yes, it is well understood. The ocean (and biosphere) maintain an equilibrium atmospheric CO2 level that depends on the temperature. The framing of the role of CO2 in the Ice Ages in this paper is rather deficient (the honey bee stuff), and it didn’t need to be to make the dust argument.

    • >>I have to admit to being confused by a minor point, which
      >>is the cause of low CO2 concentration during glacial maxima.

      It is likely due to oceanic cooling and absorption. But you cannot easily put that in a peer review paper, because it is unproven and lies outside the concensus.

      Most papers indicate that oceanic cooling only caused 30ppm of the CO2 reduction, and then they scrabble around looking for other causes. A favourite is that the high dust-flux caused oceanic fertilisation and a large increase in marine plankton, which sucked up all the extra CO2 and deposited it on the ocean floor. To which I say “great – but how did all that ocean-floor CO2 get back into the atmosphere within the short 5 ky of an interglacial?”. To which the answer is “err, ahh, errr, ahhh ummmm.”.

      For the purpose of this paper the mechanism behind ice age CO2 sequestration and release was largely irrelevant – it happens. It is likely due to something very simple, like greater oceanic overturning allowing larger volumes of CO2 to be absorbed. But simple answers have gone out of fashion in the modern era.

      Ralph

    • “This is an impressively clear, persuasive, and interesting explanation for someone like me, outside the field.”

      That’s how it works. You first fool yourself, then you fool those that don’t know much about the issue. Fooling the experts is the tough part. They usually start nitpicking about pesky details that don’t match and contrary evidence.

  50. I would like to correct several assertions of one “Javier”.
    (1) Ellis’ theory does not “blame dust” for disintegration of the ice sheets. It blames solar power, with enhanced absorption due to dust.
    (2) How the heck would you know that the Ellis-Palmer paper “has been pretty much ignored”? Did you take a poll? Or do you mean by you?
    (3) If you want to quibble about “every case” vs. “essentially always” you establish yourself as a champion nitpicker. The theory works in every case but it works more emphatically in most cases.
    (4) Your claim: “there are several (of 11) interglacials that are not preceded by high dust levels” is simply and demonstrably wrong.
    (5) As to whether “I missed important papers” (e.g. Javier’s) my book has 505 references to the literature. I have read all the references you mentioned and they are covered in my book.
    (6) Your evident satisfaction with Tzedakis et al. (2017) suggests that you are prone to accept gimmickry. I put that to bed in a previous post.
    (7) I have read a couple of your postings and indeed it is evident that you (whoever you are) are intelligent and knowledgeable. But you seem to be married to the idea that solar, and solar alone dictates the glacial cycles (without factor X) despite the fact that the data that clearly indicate otherwise. With that idee fixe, you seem unable to grasp what seems to me to be the obvious. I’ll let this report stand on its own merits, and let posterity reach its own conclusions.

    • And I might add that the dust-ice modulation paper has been downloaded 20,000 times. So people are certainly looking at it.

      R

    • rappolini

      I replied to your comment but my reply went accidentally to the wrong place, upthread:

      https://judithcurry.com/2018/09/08/beyond-milankovitch/#comment-880450

    • Donald,

      (1) Whatever.

      (2) In the almost 2 years since Ellis & Palmer, 2016 was published it has been cited 4 times according to Google Scholar, and only two of them correspond to actual scientific articles. By way of comparison, the Snyder 2016 article that blames CO₂, has been cited 51 times. I consider one citation per year as being pretty much ignored.

      (3) “The theory works in every case.” No it doesn’t. Show how it works in MIS 17, MIS 15a, and MIS 7c.

      (4) Then show me wrong.

      (5) If you read those that I mentioned, you failed to discuss that they propose a much better parameter than 65°N 21st June insolation that solves the problem that according to you needs solving.

      (6) Tzedakis hypothesis has less gaps than yours, so if you put Tzedakis to bed, your hypothesis goes to comma by the same criteria.

      (7) Your accusations don’t fly. You have chosen your favorite hypothesis, and I have chosen mine. I have no bone in this fight, but Huybers and Tzedakis are heavy weights in this field and their arguments, in my opinion are better than yours or Ellis’.

      I’ll use Clive Best figure to show why the dust hypothesis has serious problems:

      That’s 150,000 years. Dust levels are terribly inconsistent with ice melt, as the arrows (mine) show.

      Or we can use this other time during MIS 5.

      The dust levels are off respect the ice melt. The go down thousands of years before the melting starts. As ice accumulates every single year, how deep is that dust after 5000 years of ice accumulation when the melting starts? An if dust levels can act many thousands of years after they occur, why would then matter when they occur?

      • Javier…

        >>Dust levels are terribly inconsistent with ice
        >>melt, as the arrows (mine) show.

        A correlation with minimum ice is not what we are concerned with here, it is a correlation with temperature rise that matters – the beginning of the interglacial warming. And I have to say your arrows are in completely the wrong position. If you align your arrows with the start of the melt, they match very nicely with dust production.

        And you label your arrows as ‘big dust little melt’ and ‘small dust large melt’. Yeah, but cast your eyes upwards and you will see that ‘little dust’ coincided with a much stronger Milankovitch Great Summer. It is a combinatiuon of dust (albedo) and insolation that will generate an interglacial.

        However, the ice volume in your second image, at the start of the more recent (previous) interglacial, is probably correct. Just like with the Holocene, this interglacial received obliquity warming some 10 or more millennia before it received any precessional warming. So you can see there was a small amount of obliquity warming from your dotted line onwards, while the ice volume did not change. This is not unexpected – ice has a very large latent heat of melting and evaporation, and is very difficult to melt without substantially increased insolation.

        Dust production reduced at this point because CO2 levels were rising in line with temps, as they will. So all of that is very logical. Then we have precessional warming at the beginning of your solid line, resulting in a sudden increase in temperature and matching reduction in ice volume, just as we would expect.

        You seem concerned that dust production reduced before the ice melt. I am not sure why. The previous 20 ky of dust has not gone anywhere – most of it is still there within the ice sheets. All we need is precessional warming strong enough to ablate the upper layers of the ice sheet (which are still contaminated by small dust levels), and the thick dust below will be exposed. And then as the ice ablates further, the dust will concentrate on the surface so that the ice sheet will become black, just like in the images above.

        Now we have all the requirements for complete ice sheet dissipation. We have strong Milankovitch Great Summer insolation (with obliquity and precession together), plus low albedo ice sheets that can absorb that insolation. No mathematical fiddle-factors are required. All we need is several millennia of dust, and NH insolation above a certain level – and voila….!!

        The simple explanations are always the best.

        Ralph

      • >>In the almost 2 years since Ellis & Palmer, 2016 was
        >>published it has been cited 4 times, and only two of
        >>them correspond to actual scientific articles.

        How many times has your paper been cited, Javier? Oh, wait a minute….

        And you fail to recognise that the dust-ice-albedo paper runs against the consensus – it actually states that CO2 is not the primary feedback mechanism for interglacial warming and cooling. So how is any climate scientist going to get a grant, if they quote a denier paper?

        You may scoff at such a conspiratorial notion, but that is what happens. One of my Royal Society reviewers refused to pass the paper because I quoted a known denier – Willis Eschenbach’s Thunderstorm Thermostat Theory (he writes articles for WUWT). Yup, this is what happens.

        Ralph

      • “How many times has your paper been cited, Javier? Oh, wait a minute….”
        My best cited paper has over 450 citations, and I’ve got also a couple with well over 100 citations. None of them in climate science as I don’t do research on climate science. There are excellent specialist climate scientists doing an outstanding job, like Judith Curry or Peter Huybers. What is the point in getting into a field one is not an expert, and get a mostly wrong article in an obscure journal than then gets ignored by everybody? Perhaps you can tell me.

      • Sorry, the second interglacial in that diagram does not have a stronger Great Summer. My bad.

        Ralph

  51. Reply to ralfellis | September 9, 2018 at 9:44 am |
    Thank you for answering my maverick piece. You have referred to this work https://pdfs.semanticscholar.org/b571/0d2bfefe772dfb958e6a012e6b0d01740ede.pdf

    I have no doubt on the competence there. But that is for secular changes. Look at page 16 at this section “—-where ε0= 23.254500 degrees, and the coefficients—-“. Epsilon suffix 0 is the ‘as is’ value; empirical.

    It has been assumed all along since JN Stockwell that it is unchanging, when in fact evidence shows that it changed in 2345bce and at other times before that, within the last 8k years. Once that empirical value changes substantially (it was ~14.5 around 3200bce) then the whole story changes. The devil is in the detail, and this is one detail that has been overlooked for too long. Both Dodwell and Wittmann noticed that discrepancy. I found that same anomaly from a different field altogether.

    • Obliquity was 14.5 degrees in 3200 BC?
      Where is the evidence for that?

      R

      • You may find the evidence and a lot of related information here; https://melitamegalithic.wordpress.com/2018/03/26/mnajdra-south-calendar-design/

        An 18deg equinox to solstice angle, at latitude 35.8 gives an earth obliquity of about 14.5deg. Not precise, but way beyond 22. This unit is one of several, most are at 18deg, but this one was converted to ~29.5; for today’s obliquity. The conversion corroborates Dodwell’s statement and theory. That particular unit still works; I tested a part model of it to predict the solstice day (QED, to the nearest 3hrs, repeatable), see https://melitamegalithic.wordpress.com/2017/02/24/first-blog-post/

        These megalithic calendar due to their design have retained a record of earth tilt at time of build/operation. But no precise date. That was subsequently found from the several proxies that all show abrupt changes at near identical dates. Particularly the sudden changes of temperature trends seen in ice-cores from both poles and opposite, at equator.

        The mathematical analysis, that dates back to JN Stockwell, analyses anticipated secular change. But it cannot tell where the mean was at any one time, or what abrupt change took place. Dodwell’s analysis was the keen insight of an open mind. The above is solid proof he was right.

  52. It is really difficult to read all this stuff and understand the truth and try to understand the wrong ideas.

    An ice age is caused by ice. Snow is produced using moisture that evaporates from oceans and energy that was stored in oceans. In warm times with deep warm oceans, the moisture and energy from the oceans produces ice until it is enough to cause an ice age.

    In cold times with shallow cold frozen oceans, there is no moisture or energy from the oceans and ice in ice sheets and glaciers thaw and deplete. It takes a long time to thaw and deplete a lot of ice, the ice age ends when the ice runs out.

    This is not considered in most everything all of you write.

  53. Donald Rapp, thank you for this interesting post, and thank you, Ralf Ellis and Clive Best for the informative comments and discussion.

    I am persuaded that dust on ice sheets, together with relatively high insolation in the Northern Hemisphere, were the main cause of glacial terminations.

    I am also persuaded we need to do all we can to lift Earth out of the current ice age. It will take thousands of years to do so. I doubt there is enough fossil fuel to provide enough CO2 and black carbon to free us from the current ice age. However, I have an idea as to how it could be done.

    We need to melt the ice on Antarctica too. Here’s how:

    We build huge coal fired and nuclear power stations in Antarctica. The electricity is used to produce hydrocarbon fuels (petrol, diesel, jet fuel, etc.) from sea water (ref. to US Navy and Audi Research on this). These fuels are exported to the world for transport fuels and other uses. Diesel could also be used in power stations in Antarctica to produce more CO2 and black soot.

    Voila!

    • >>I am also persuaded we need to do all we can to
      >>lift Earth out of the current ice age.
      >>It will take thousands of years to do so.

      Oh, I don’t think it is that difficult.

      Firstly, orbital eccentricity is low for the next 100 ky or more, and so there will be no very deep NH Great Winter to force the NH (and the world) into a new ice age. Having said that, we do have the same orbital configuration now, that has caused glacial initiation in the past – I estimated 500 or 1,000 years before the new ice age will arrive. But orbital conditions then improve again, so it is difficult to know if this new ice age would be sustained anyway.

      But if it did grow, through high albedo ice-sheet cooling, then all we would need to do is spray the expanding Canadian and Siberian ice sheets with black carbon each spring. You would probably only need a fleet of 50 – 100 747s to do this. It is not beyond our abilities, by any means.

      Ralph

  54. One thing is for certain, CO2 has nothing to do with the cycle. There is no mechanism by which CO2 would increase before the end, and decrease before the start of an ice age. The AGW doesn’t even attempt to explain the most important events in the cycle.

    Why CO2 is Irrelevant to the Earth’s Lower Atmosphere; You Can’t Absorb More than 100%
    https://co2islife.wordpress.com/2018/09/08/why-co2-is-irrelevant-to-the-earths-lower-atmosphere-you-cant-absorb-more-than-100/

    • >>There is no mechanism by which CO2 would increase
      >>before the end, and decrease before the start of an ice age.

      Indeed. They say that CO2 concentrations increase at a glacial maximum, due to increased Great Summer insolation warming the climate. And then this extra CO2 helps the interglacial warming. But if the world was warming anyway, without this extra CO2, then why bother invoking CO2 assistance at all?

      And why do some Great Summer Insulation maxima produce no warming at all? Why did temperatures fail to respond. And why, when CO2 levels are at a peak during an interglacial, does the world cool? The CO2 feedback system cannot explain any of this.

      But the albedo theory can. At a glacial maximum the ice sheets are already dusty, with lower albedo, because of the previous 10 ky of dust. All the climate system is waiting for, is some extra NH insolation to kickstart the process. And as soon as more insolation arrives, the interglacial begins.

      Ralph

      • The problem with those theories is the their is a cycle, 100,000 yr ice age, 10,000 yr inter glacial. It is hard to explain a cycle like that based on CO2. Also, CO2 never gets low enough that the marginal impact of CO2 really matters. All additional CO2 does is lower the level of 100% saturation, it doesn’t increase the energy of the system. You can only capture 100%, and CO2 does that at relatively low levels, so more CO2 just captures it slightly closer to the surface. Lastly, CO2 traps 13 to 18 micron LWIR. You can shine those wavelengths on ice and it won’t melt. In fact ice emits LWIR of a shorter wavelength. LWIR between 13 and 18 is consistent with a black body temp -50 to -110 degree C. They don’t melt ise.

      • According to Lamb et al., 2008 dust inversely correlates with temperature during glacial periods.

        “Comparing dust and stable isotope (dD) profiles, there is a significant correlation during glacial periods (Fig. 2), and up to 90% of the dust variability can be explained by the temperature variations.”

        The correlation is really good and we all know that CO₂ and temperature also correlate. So you cannot work out a causality between three things that correlate. Temperature could drive the other two, or dust could drive down temperature through irradiation absorption and reflection, and temperature decrease drive down CO₂.

      • >>Temperature could drive the other two, or dust
        >>could drive down temperature through irradiation
        >>absorption and reflection, and temperature
        >>decrease drive down CO₂.

        You are beginning to flail, Javier, instead of think rationally.

        How does a linear temperature drive a logarithmic dust flux?
        How does a logarithmic dust flux drive a linear temperature?
        You see the problem?

        But the dust-albedo paper explains everything, because the land area exposed below the CO2 plant-death-zone increases exponentially, as the graph in the paper demonstrates. (M Palmer kindly graphed the land area of the Gobi for each altitude, and the result matches the logarithmic dust production.)

        But you would know that, if you had bothered to read the paper.
        Really, Javier, you are like a blind man on a driving test…!

        Ralph

  55. Would not land exposed on the continental shelves and shallow seas by the 400 foot drop in sea level at the glacial maximum have been a major contributor to the world’s dust load? Has anyone taken this into account? The ice free exposed land off Alaska and Eastern Siberia would have been devoid of terrestrial plants and once dried (residual salt would have inhibited colonization by plants) would have been a large source of dust very close to both major ice sheets.

    • That should promote dust throughout the cold period with low oceans and would not explain dust cycles.

    • Not really.

      I saw a paper that espoused this idea – but as you know, any exposed land will be covered by plant growth very quickly. Grasses and weeds would colonise the newly exposed land within a couple of years, preventing any dust production.

      In addition, dust production is initiated at a sudden point during the ice age, while the sea level reduction is reasonably constant. There is no sudden decrease in sea levels that coincides with greatly increased dust production.

      However, dust production IS highly correlated with CO2 decrease. In fact, they almost match each other. And this is just what we would expect, as the CO2 flora-death-zone reduces in altitude down the upland regions of the world – exposing greatly increased (logarithmic, in fact) land areas as it goes.

      Image: Dust vs CO2.
      Dust plot is inverted and logarithmic.

  56. So dust may contribute to terminations, but… We know earth’s climate is governed by overall (global!) negative feedback. The climate’s response to changes in insolation changes and other signals or forcings doesn’t happen instantaneously but with delays. This causes ringing and oscillations in the earth’s climate system. These may synchronize with external signals (insulation cycles) or not. The periods of cycles are largely a characteristic of the earth system, not external forcing. When looking at earth’s cycles, think of physical characteristics like harmonics and hysteresis. I did this on my phone so be kind.

    • Nuts, I meant insolation, not insulation. Really this virtual keyboard is tiny and I can’t use my thumbs to type

  57. We clearly remain in the Pleistocene. As such, the CURRENT “natural” condition is to be glacial, but the Pleistocene is only the third significant glacial period in the Phanerozoic, and one of five that we know of in the 4.6 billion years of earth history. The accounting becomes sketchy, but it is safe to say that overall that these macro glacial periods represent only a small fraction of earth history.

    The “natural” condition overall is clearly NOT to be in a macro glacial period.

    We come now to the fibrillations between “glacials” and “interglacials” within the macro scale glacial periods you deal with here. Very noticeably, the transitions from glacials to interglacials takes place very rapidly, while the planet must be dragged kicking and screaming from an interglacial back to a glacial. A reasonable hypothesis would be that interglacials more closely represent the natural state and that a “z” factor generally suppresses temperature during macro glacial periods.

    It seems reasonable to suspect that whatever the Z factor is, it also drives the glacial oscillations with macro glacial periods.

    • >>Transitions from glacials to interglacials takes place
      >>very rapidly.
      >>A reasonable hypothesis would be that interglacials
      >>more closely represent the natural state.

      Or you could suppose that the interglacial generation mechanism is much stronger than the glacial generation mechanism. And that is probably so.

      The glacial mechanism depends on (say) 50 cm of snow falling each year on an ice sheet. So building a 3,000 m ice sheet is by necessity a slow process – dependent upon snow deposition and ablation rates.

      Conversely, if an interglacial can reduce ice sheet albedo from 0.85 to 0.35, then the ice sheets can increase insolation absorption by 250 W/m2, over the entire annual summer season. That is a lot of energy, and a lot of melting.

      R

      • The reduction in albedo will also be slow when there are 3k meters of ice between incident light and less reflective earth.

        I see no alternative to the z factor plunging the planet from paradise to ice 5 times in 4.6 billion years; in wildly different configurations of continents. No known orbital variations have periods of hundreds of millions of years. CO2 merely follows the falling temperature like a poodle on a leash, just like in the ice cores and benthic cores; as the cooling ocean draws down atmospheric CO2.

  58. I’m bored but not surprised. Dust on ancient ice seemed like an interesting mechanism at the time. Less so now given the hedgehog tendency of the author explaining everything in terms of one big idea. He doesn’t seem to think too fast on his feet either. The 2.8 W/m2 CO2 forcing change – for instance and I did reference the IPCC – was from the LGM to pre-industrial – where the total forcing including ice and land effects is some 7.8 W/m2. Although the engineer in me loathes such fake precision. The 4 W/m2 is for CO2 doubling. There is a simple formula for temperature change.

    ΔTs = −F/λ

    Where delta Ts is about 10K and F is about 8W/m2. Modern feedbacks are very uncertain – but I suggest based on the very simple formula that net λ is around -1 W/m2 per K. If it is constant. This is one of the benefits of being an engineer trained in Fermi problems – although some engineers seem to have been in the bar that day.

    It seems everyone has a big idea. The factor X – whether it is running out of ice, dust, demented deniers, something incomprehensible involving black and white 1’s or certifiable like melting Antarctica. I don’t have a big idea – I have a paradigm that is likely science consensus. The US National Academy of Sciences (NAS) defined abrupt climate change as a new climate paradigm as long ago as 2002. A paradigm in the scientific sense is a theory that explains observations. A new science paradigm is one that better explains data – in this case climate data – than the old theory. The new theory says that climate change occurs as discrete jumps in the system. Climate is more like a kaleidoscope – shake it up and a new pattern emerges – than a control knob with a linear gain.

    Everything is chaotic from orbits to snail populations. The only I know to calculate orbits is with a massively complex n-body Hamiltonian. This is – theoretically – a continuous function where gravity effects are integrated over small time intervals. It would all be deterministic – but you would need Laplace’s demon to do the calculation. There are simpler methods but they do not capture the abrupt departure of orbits into new fractionally dimensioned state. Just like the three body problem of Poincare suggests – but with a massive central body.

    I have been wondering for some time now how the reality of orbits might diverge from the usual calculations. But only instrumental data over very long periods could tell – and that seemed unlikely. But I have been reading today the Dodwell manuscript – The Obliquity of the Ecliptic – and was shown a fascinating world of ancient instrumental observations – from all over the world – that diverges from a modern science with short periods of observations and mathematical models. The document dates from the mid 20th century and is an utter pleasure to read.

    “I was aware that Professor Drayson’s theory was regarded as a paradox, contrary to sound astronomical doctrine, by high astronomical authorities; and I wished to find out where it was in error. It soon became clear that the paradox lay in the fact that for this proposed physical and cyclical movement of the earth, which he believed to be still in progress, there was apparently no known or possible physical cause.

    Nevertheless, although the theory was thus untenable, it seemed to me worthwhile to trace out more clearly just how much, and why, the ancient and mediaeval observations of the obliquity of the ecliptic, on which Professor Drayson based his conclusions, differed from Newcomb’s internationally accepted formula for the secular, or age-long, variation of the obliquity. These observations went back to values given by Strabo, Proclus, Ptolemy, and Pappus in the early centuries of the Christian era. They indicated a consistent and increasing divergence in past ages from the values calculated by means of Newcomb’s formula.”
    http://www.setterfield.org/dodwell_manuscript_1.html

    I suggest that in the hyper complex Earth system that the search for an X factor is a forlorn exercise. There are more likely to be x, y and z factors in any process. Multiple and changing positive and negative feedbacks that we have barely scratched the surface of. If you are looking for simple rules they can be found elsewhere

    https://wordpress.com/post/watertechbyrie.com/2918

    • >>The 2.8 W/m2 CO2 forcing change was from
      >>the LGM to pre-industrial.

      Indeed, and that is the problem, because 2.8 W/m2 is the total forcing over 5 or 6 ky. But at the start of the interglacial we do not get that amount of forcing. In the first century of the interglacial we only get 0.045 W/m2 of additional forcing, from a paltry 2ppm rise in CO2. And it is the feedback assistance from that miniscule 0.045 W/m2 that must force the climate to warm, so that the next century is warmer than the previous. Quoting 2.8 W/m2, as is so often done, is meaningless because that is not what we have.

      Conversely, if the albedo of the ice falls from 0.85 to 0.35, we will see a 250 W/m2 increase in insolation absorption across the ice sheets, for each and every annual summer. And that is two orders of magnitude greater than the feedback from CO2.

      R

      • ….and the 250 W/m2 regional albedo feedback is some 5,000 times greater than the 0.045 W/m2 CO2 feedback, during the first century of the interglacial. So what is more likely to result in rapid warming and rapid ice-sheet melting – dust-albedo or CO2…?

        R

      • I’m ignoring this with the ‘contempt it deserves’.

      • >>I’m ignoring this with the ‘contempt it deserves’.

        Sounds like you lost the argument.
        But the usual response is ‘sorry, you were right…’

        R

      • Quotation marks because I was quoting Ellis as he full well knows.

      • The CO2 point was made because Ellis once again leapt to a wrong interpretation of a simple point. CO2 follows temperature – sans not too much volcanism – from Henry’s law but mostly as a biokinetic feedback. It is still happening today. No one is suggesting that CO2 led the way – at least I am not.

        e.g. http://environmentportal.in/files/Temperature%20associated%20increases%20in%20the%20global%20soil.pdf

        Temperature shifts the balance between respiration and photosynthesis. It added substantially to global warming between the LGM and preindustrial.

        Ellis’ numbers are complete nonsense. Whatever the albedo of ice is – it is mostly obscured by cloud – cloud is an anti-correlated feedback to surface temperature – and thus a positive feedback to warming whatever the source. It is a known unknown. And while dust should melt ice – the global energy budget depends on global average albedo change – not on regional changes in albedo. The 4 /m2 change from doubling CO2 is a different problem entirely.

      • >>Ice … is mostly obscured by cloud – cloud is an
        >>anti-correlated feedback to surface temperature
        >> and thus a positive feedback to warming.

        Err, not on the LGM ice sheets it is not. Precipitation dramatically fell over the Laurentide and European ice sheets at the LGM. In part, but not exlusively, because of the strong katabatic winds that flowed to the south, down the cold ice sheets. And as you know, descending air does not condensate.

        .

        >>And while dust should melt ice – the global energy
        >>budget depends on global average albedo change –
        >>not on regional changes in albedo.

        Ok, so what melts the annual winter snows in Canada. Is it the regional albedo-driven insolation and absorption in Canada, or the albedo in Argentina?

        I keep asking this question, but nobody want to answer. Face facts – Hansen got this wrong and led everyone astray. He said:

        a. Surface albedo could be averaged across the globe to derive an global 3 W/m2 forcing. But as I have argued above, it is regional albedo that matters, not global albedo.

        b. The value for CO2 forcing is another ~3 W/m2, when that is the total value at the end of 5 ky of interglacial warming. The CO2 forcing during the first century of interglacial warming is only ~0.06 W/m2.

        Of what value was Hansens’ paper?
        (Climate Sensitivity Estimated From Earth’s Climate History. Hansen and Makiko)

        Ralph

      • What little direct evidence for Arctic cloud is that cloud cover declined during recent times. It is not linearly related to precipitation as is the common but misguided skeptic meme – it is non-linearly anti-correlated with surface temperature especially over oceans. In Rayleigh–Bénard convection closed reflective cloud cells tend to form over cold surfaces. With sufficient collision energy and nucleation sites closed cells rain out from the center to form open cells. Closed cells persist for inordinate periods. High northern latitudes were cloud free at the LGM? You are not notably credible.

        a) What the hell does melting ice in Canada have to do with anything. Of course ice melts in high latitudes – where else would you find it. Other than at high altitude in Papua New Guinea. The albedo effects of ice melting in one region must be globally averaged to understand changes to the global energy budget. This is something you inexplicably deny for misguided reasons of denying the geophysics of carbon dioxide it seems.

        b) Again irrelevant – CO2 added to warming as a positive feedback in the transition. But what you appear to be arguing is that it doesn’t lead deglaciation. An argument with yourself it seems because no one credible says it does.

        C) I stopped reading Hansen probably about 1988. Far to simple to be credible. As is this. I don’t know who you are arguing with but it is not me.

    • OK. Now we are out of the box. Have just answered ralfellis above. There is the solid proof that Dodwell was right all along. Had he known of that he might not have been ignored. But such is life.

  59. Salvatore del Prete

    Question let’s take the last great cooling the YD. Is there evidence that dust played a role in the rapid ending of this cold period?

    • I go out on a limb with the YD, because I think it was caused by a meteorite impact on the Laurentide ice sheet.

      The YD is the opposite of the D-O warming events, and plunged the NH back into an ice age in less than a century. So a normal ice age takes 100 ky to develop, and the YD took less than a century – and it also caused the extinction of the megafauna and the demise of Clovis Man. This was a truly catostrophic climate change.

      As a modification of what has been said before on this topic, I think a meteorite struck the Laurentide ice sheet and spread ballistic slush-balls all over North America. These caused the half million Carolina Bay impact craters that litter North America, and are all orientated towards the Great Lakes region.

      This is not a popular theory, but it does explain all the known facts and is difficult to refute, and I find that appealing.

      The Carolina Bays and the Destruction of North America
      https://www.academia.edu/20051868/The_Carolina_Bays_and_the_destruction_of_North_America

      Some of the half million Carolina Bays that litter North America.

      https://www.academia.edu/20051868/The_Carolina_Bays_and_the_destruction_of_North_America

  60. An observation: A full deglaciation does not appear to happen unless similar low temperatures are reached, and low eccentricity, only then is the effective obliquity phase functional. This suggests a particular combination of ‘events.’ One is that ocean levels are at lowest. Salinity would expected to be higher than now – how significant? At deep glaciation there would be extensive sea ice, such that the Southern Ocean circulation may be minimal or cease, changing ocean circulation patterns to favour rapid warming during the ‘up-lobe’ conditions. Are there any clues that this may be significant?
    Some comments have discounted the role of dust. The presence of dust in ice cores prior to deglaciation is not an issue, as dust on snow may be buried in the next snow fall, but when on a snow surface and warm enough, it would tend to remain on top while the snow melts below. When melting exposes the frozen ice surface, dust would most likely be washed away, any the dust within the ice would then aid the melting process, absorbing more heat than pure ice.

    • >>An observation: A full deglaciation does not appear to
      >>happen unless similar low temperatures are reached,

      Yes, but according to the dust-ice-albedo theory the primary criteria for interglacial initiation is actually low CO2 and therefore high dust. Temperature is not relevant, excepting that it plays a role in reducing CO2.

      .

      >>dust would most likely be washed away.

      Both experience and experimentation suggests that the dust particles warm and embed themselves into the ice surface, like the grains on sand-paper. They are quite difficult to wash awy.

      R

  61. Pingback: Bits and Pieces – 20180910, Monday | thePOOG

  62. Javier: “Huybers and Tzedakis are heavy weights in this field”. Indeed. But are they are “heavy weights” rather than “heavyweights”. Members of the establishment might serve to add inertia to the prevailing views? A theory has to stand on its own, regardless of the credentials of those who propose it. I don’t know Ellis. I never met him. I think he is an airline pilot who happens to have a deep interest in ice ages as an avocation. Amazing that he would come up with this, but he seems to be very smart. He is apparently regarded as an interloper by the establishment? Me too? But at least I have some credentials. See http://home.earthlink.net/~drdrapp/

    • You keep going in the wrong direction with me, Donald. I don’t care who you are or what are your opinions on unrelated matters like biblical studies or creationism. All I care is about the scientific product. Good scientific products tend to end being published and discussed in scientific journals and tend to convince other scientists when the evidence is good. This doesn’t mean there are no exceptions, but the dust hypothesis doesn’t appear to be one.

      The quality of the work of Peter Huybers and Christos Tzedakis is outstanding. That is why they publish so often in the best journals. When I looked into this issue by myself by looking at the data without any assumptions I reached independently the same conclusion so that gives me a lot of confidence that in my opinion they are correct.

      By contrast when I look at the dust data I reach the opposite conclusion you and Ellis reach. As I have shown, the data does not support the evidence. You have the cause/effect relationship reverted. Just flipping the temperature data as I have done in the graph below shows that.

      So it is the science that doesn’t fly, not the people, and attacking Tzedakis and Huybers will not make it fly.

      • >>The quality of the work of Peter Huybers and
        >>Christos Tzedakis is outstanding. That is why
        >>they publish so often in the best journals.

        The work of Huybers was deliberately misleading. They published their obliquity paper, wrapped it in Christmas paper, and tied it up beautifully with a silk ribbon. But failed to mention that half of their obliquity maxima did not produce an interglacial. And a theory that only explains half the problem, is half baked.

        And, as I said before, the Huybers obliquity theory did not work when using the standard paleoclimatic chronology, so they made their own – which did match the obliquity cycle. Yet more Christmas wrapping paper.

        .

        And your ‘flipping the temperature data’ demonstrates that you still do not understand the theory. A theory is not simply a correlation, because correlation does not mean causation. And your championing of obliquity forcing without an X-factor is patently false, because of the missing interglacials. Just like Huybers, you cannot explain why the obliquity maxima 370, 290, 170, and 50 ky ago, failed to produce an interglacial. Where is your X-factor, that turns a failed obliquity maxima into a successful maxima?

        Instead of looking for correlations, you should be looking for mechanisms. And there is a mechanism here. Low temperatures = low Co2 = increased dust = low albedo ice = greater insolation absorption. Plus increased insolation = new interglacial. It is not difficult to understand, nor is it difficult to show the data that supports the theory.

        Do temperatures decrease during ice ages? = yes.
        Does CO2 reduce during ice ages? = yes.
        Does reduced CO2 cause C3 plants to fail and die? = yes
        Does CO2 partial pressure reduce with altitude? = yes.
        Did LGM treelines lower in a manner that cannot be ascribed solely to temperature? = yes.
        Is the Gobi steppe grassland at high altitude? = yes.
        Did the Gobi become a desert during the LGM? = yes.
        Was there increased dust production from the Gobi? = yes.
        Did the NH ice sheets get covered in dust? = yes.
        Is that dust from the Gobi? = yes.
        Does this dust reduce the albedo of ice sheets? = yes.
        Does this dust-albedo happen just before interglacials? = yes.
        Does NH insolation rise before each interglacial? = yes.
        Do interglacials proceed at a rate that is difficult to ascribe to insolation alone? = yes.
        … … (because other insolation maxima fail to produce interglacials).
        Is it likely that dust-albedo was a decisive feedback that precipitated interglacial warming? = yes.

        So yes, you have me here, Javier.
        Sorry – apart from all that, I ‘lack evidence’ and cannot prove that dust is the feedback agent for interglacial modulation… ;-)

        Ralph

      • “The work of Huybers was deliberately misleading.”

        Your opinion only. I agree with the reviewers of the Science article.

        “And your ‘flipping the temperature data’ demonstrates that you still do not understand the theory.”

        It is not a theory, but a hypothesis, that has failed to garner much support.

        “And your championing of obliquity forcing without an X-factor is patently false, because of the missing interglacials. Just like Huybers, you cannot explain why the obliquity maxima 370, 290, 170, and 50 ky ago, failed to produce an interglacial.”

        I have absolutely no problem in explaining that on orbital terms and ice volume alone. No dust needed.


        Figure 133. The timing of Pleistocene glaciations as a function of summer energy, ice-volume and eccentricity.
        https://judithcurry.com/2018/08/14/nature-unbound-x-the-next-glaciation/

        “Instead of looking for correlations, you should be looking for mechanisms.”

        I asked for the evidence and I didn’t get it. Which means that you are talking big without the evidence to back up what you say. Where is the evidence that the plants in the Gobi plateau died from low CO₂? Nowhere. It doesn’t exist. That’s how serious is your hypothesis.

        “you have me here, Javier.”

        Yes. In the unsupported claims category.

      • Javier: You (not me) were the one who introduced the notion of “heavy weights” in the field. Publishing stuff on ice sages and climate change in the best journals proves nothing to me after reading half a thousand of them. Good scientific products are good whether the in–group accepts them in journals or not. I have 100 publications and I’ve been through the mill for 60 years of postdoctoral research and I have noted that a lot of stuff gets published that is junk, while interlopers from outside the in–group have difficulty getting in. To me, it is a wonder that none of the heavy weights you cite could clearly discern the obvious.
        As I pointed out in a previous, detailed posting, Tzedakis (2017) is simply not science. First of all it utilizes the integral of solar input over a year which varies far less from year to year than the midsummer solar intensity. Then it supposes that there is a threshold for solar input, above which terminations occur. In this model, because he used the yearly integral, the sensitivity of the ice sheets is so incredibly fine that a minuscule difference in solar input distinguishes between ice ages and interglacials. Since that did not work as well as they hoped, they further presupposed that some invisible little demon reduces the threshold when an ice age has been around for a long time. Changes of less than 1% in solar do not cause terminations. Changes of tens of percent in absorptivity do cause terminations. It is not a change in threshold that produces a termination. It is a change in solar energy absorbed. There is no physics in Tzedakis (2017). It is all more or less curve fitting using the wrong dynamic variables.

      • You are entitled to your opinion, but Didier Paillard has this to say about the Tzedakis et al., 2017 article:

        “Tzedakis and colleagues have made progress in this direction by discovering a rule that links astronomical forcing to the timing of the Quaternary interglacials. In the earlier part of the Quaternary, interglacials occurred when the amount of summer solar radiation was above a given threshold; in the later part, this threshold was a linearly decreasing function of the time elapsed since the previous interglacial. The authors’ idea is not completely new, but their demonstration is enlightening because it focuses on the key difficulty — the occurrence of interglacials. Their statistical analysis shows that the rule is robust, and that the observed sequence of interglacials is the most probable one, among a small set of possible histories. This is a strong indication that ice ages are indeed predictable, and that the underlying mechanisms behind the 100,000-year climate cycle are probably quite simple.”

        You might not like it, but an article with a statistical analysis is a valid mathematical approach that helps distinguish between alternative possibilities.

        Your problem with the paper probably comes because it leaves your pet hypothesis hanging out dry, unnecessary, and unable to pass such a strict statistical test.

      • >>Where is the evidence that the plants in the Gobi
        >>plateau died from low CO₂? Nowhere. It doesn’t exist.

        Read the paper.
        You are really obtuse sometimes.

        R

      • “You are really obtuse sometimes.”

        No. I wanted you to say it, because I know you don’t have any evidence supporting what you say.

        From your article:

        “although there were changes in precipitation patterns during the glacial maximum, the world was far from becoming a barren and dusty desert due to a lack of precipitation. Confirmation can be seen in the PMIP3 precipitation map in Fig. 7b,

        As these two charts demonstrate, the glacial maximum world was far from becoming a cold and barren desert due to either precipitation or temperature reductions.”

        I understand you are not a scientist, but by now you should know that climate model output does not constitute a demonstration of anything except computer skills, and paleoclimate models trying to reproduce conditions 20,000 years ago are the worse of the lot.

        What you have done there is state your belief supported by a global model composite with appalling resolution that the plants could not have died from anything other than low CO₂.

        Knowing you I understand that it is useless to try to present evidence that does not support what you say as you will reject it. But any interested reader can go to check what the proxy evidence says.
        Jasechko, Scott, et al. “Late-glacial to late-Holocene shifts in global precipitation δ18O.” Climate of the Past 11.10 (2015): 1375-1393.
        https://www.clim-past.net/11/1375/2015/cp-11-1375-2015.pdf
        They show in their figure 1 that the Northern China-Mongolia area was -5 to -10°C colder during the LGM. Their figures 2 and 6 show Northern China having a precipitation deficit.
        Conditions there are now difficult for plants with historical draughts seriously affecting the region. Lowering the temperature and precipitation should have made conditions a lot more difficult for plants, so the hypothesis that they died due to low CO₂ is not supported.
        Figure 5 shows 5 different precipitation models and their composite, illustrating the huge differences between models that prevent getting a significant answer from them.

      • >>In the later part, this threshold was a linearly decreasing
        >>function of the time elapsed since the previous interglacial.

        But they fail to explain why a larger ice sheet is more prone to instant collapse, instead of just retreating back to a more stable configuration. (There is not even any convincing explanation as to why a large ice sheet is more vulnerable to anything. This is a convenient meme, that has been built into many models to get them to work.)

        What they actually did, as I demonstrated earlier, is find zero correlation between insolation maxima and interglacial temperature. (As per the first graph I showed.) So they invented a fiddle-factor that would place most of these insolation maxima in the right order – raising some up, and lowering others down. And voilla – a contrived correlation.

        >>Their statistical analysis shows that the rule is robust

        But shows that their explanation of an underlying mechanism is weak.

        R

      • “But they fail to explain why a larger ice sheet is more prone to instant collapse, instead of just retreating back to a more stable configuration.”

        I don’t know what they explain about it, but they don’t need to. Multiple studies indicate that when glaciers and ice sheets grow too big, the contact surface with the rock can get into a state of flux causing a rapid collapse. Evidence for this type of phenomena exists. For example Heinrich events produced sudden catastrophic release of iceberg armadas at intervals when the ice sheets have become large enough. Heinrich events are evidence of instant collapse.

        So the fact you ignore something doesn’t mean the evidence for it doesn’t exist.

      • >>I have absolutely no problem in explaining that on
        >>orbital terms and ice volume alone. No dust needed.

        Your graph is marked ‘Global Ice Volume Proxy’. But that is not an ice volume proxy at all, it is a benthic temperature proxy. So what is this diagram supposed to show us?

        Reading this diagram in the normal sense, your graph graph (which contains a temperature plot, not an ice volume plot) fails to explain why the obliquity summer energy peaks 160 ky and 40 ky ago failed to generate a temperature rise, let alone an interglacial. Again this looks like a half-baked argument.

        And simply using the caloric value is also a mistake, because the total energy of the summer is not what ice sheets can absorb. Much of the obliquity insolation is at a shallow spring or autumn inclination, which is easily reflected by snow and ice. The most effective insolation for ice sheet absorption is midday midsummer insolation – and that is precessional.

        Ralph

        What does this graph purport to show us…?

      • “Your graph is marked ‘Global Ice Volume Proxy’. But that is not an ice volume proxy at all, it is a benthic temperature proxy. So what is this diagram supposed to show us?”

        I see you ignore everything about δ¹⁸O isotopic proxy. For your education, from Lorraine Lisiecki web page:
        “The LR04 stack [Lisiecki and Raymo, 2005] spans 5.3 Myr and is an average of 57 globally distributed benthic d18O records (which measure global ice volume and deep ocean temperature) collected from the scientific literature.”
        http://lorraine-lisiecki.com/stack.html

        “What does this graph purport to show us…?”

        As you told me:
        “Read the article.
        You are really obtuse sometimes.”

      • >>I understand you are not a scientist, but by
        >>now you should know that climate model
        >>output does not constitute a demonstration of
        >>anything except computer skills.

        You are flailing wildly again, Javier. If you had read the paper fully, instead of skimming it, you would see it contains real physical evidence of a wetter Gobi during the LGM.

        Ralph

      • >>What does this graph purport to show us…?
        >>Ralph

        Ok, I sort of see what you are doing there. But it still makes no sense. What you are saying in that graph is that:

        a. Small ice sheets can melt a little bit. (green band)
        b. Large ice sheets will not melt at all. (purple band)
        c. Very large ice sheets will melt completely. (blue band)

        Where is the sense in that? What is the mechanism that will make very large ice sheets melt, but large ice sheets continue to grow – when subjected to the same obliquity-inspired insolation values? And don’t just say ‘instability’ because that explains nothing. What type of instability? And why does the ice sheet not just retreat to is previously very stable condition?

        .

        What you really need on that graph is a dust plot. So what this wonderful graph really shows us is that:

        a. Small ice sheets can melt a little bit. (green band)
        … … Because there is still some exposed ground in high
        … … latitudes, that can absorb that extra insolation.
        b. Large ice sheets will not melt at all. (purple band)
        … … Because all the high latitudes are covered in high
        … … albedo ice, that can reflect all that extra insolation.
        c. Very large ice sheets will melt completely. (blue band)
        … … Because they are now covered in low
        … … albedo dust, that can absorb all that extra insolation.

        Ok, Javier, now you have a real theory. Well done – your graph fully supports the dust-ice-albedo theory. Thanks for that, I will buy you a beer sometime. Do you want a citation, on any future papers…?

        Ralph

      • I see you are narrative-driven. You need a narrative that makes sense to you. It doesn’t matter that your explanation is made up, and once you attach to one the evidence that goes against it is ignored or dismissed summarily.

        That graph shows that ice volume (and dust production) depend on eccentricity. High eccentricity promotes lower ice volume because it prevents ice accumulation from one summer energy oscillation to the next, while low eccentricity promotes high ice volume because it allows ice accumulation from one summer energy oscillation to the next. At the LGM the ice had accumulated for 100 kyr.

        But interglacials depend primarily on summer energy, as they can only take place during the high summer energy windows. Once obliquity gets very low interglacials are over. No exceptions.

        The problem for your hypothesis is that high eccentricity, that favors low ice accumulation and low dust production, also favors interglacials, shooting your hypothesis in the foot.

        Under high eccentricity, low ice volume, low dust production conditions, a high summer energy window will always produce an interglacial. That is why we get two interglacials in MIS 15 and in MIS 7, that are separated by 400 Kyr in two high eccentricity periods.

        So by accounting for ice volume, eccentricity, and summer energy, all interglacials are perfectly explained. Tzedakis et al. only considered ice volume (through ice age), and summer energy, but their formula works because ice age depends on eccentricity, so the third factor is to a certain extent accounted for.

        You attached to your hypothesis early on, and so you failed to notice that dust production is directly proportional to ice volume and inversely proportional to temperature. Lambert et al., 2008 show that but you preferred to ignore it and run with CO₂. So you had two consequences of ice accumulation and temperature decrease, dust and CO₂ levels, trying to act as causes.

        And you have decided to ignore, despite being repeatedly told, that at times of high eccentricity your hypothesis breaks down.

      • >>High eccentricity promotes lower ice volume because it
        >> prevents ice accumulation from one summer energy
        >>oscillation to the next.

        Hold on a minute. Explain that assertion in more detail.

        Obliquity insolation is not effected by eccentricity, unless you include an element of precession. What you are doing here is including precessional insolation into your theory, while denying that precession plays a role in ice age modulation.

        .

        >>Interglacials depend primarily on summer energy, as they
        >>can only take place during the high summer energy windows.
        >>Once obliquity gets very low interglacials are over.
        >>No exceptions.

        Except the precessional peak 320 ky ago did produce an interstadial, even though obliquity was at a minimum. As did the precessional peak150 ky ago. So you are wrong. Look, obliquity tilting the earth over further may well produce greater insolation in high latitudes during the summer. But placing the Earth closer to the Sun at the same time (precession + eccentricity) enhances that summer insolation by a considerable margin.

        .

        >>Under high eccentricity, low ice volume, low dust
        >>production conditions, a high summer energy window
        >>will always produce an interglacial.

        If you have low ice volume, you already have an interglacial. I presume you mean high eccentricity produces higher summer insolation, and a low ice volume. Indeed, but it only does so through precessional effects, not simply by obliquity. If there were no precession, summer insolation would not necessarily increase with eccentricity – spring or autumn insolation might increase instead so there would be no increased summer insolation. Just admit that precession is a key element in your theory.

        And you forget (yet again) that your theory has some serious difficulties. 290 and 175 ky ago obliquity and precession were high. Eccentricity was reducing, but the precessional insolation was still 580 W/m2 measured at 65ºN in both cases – two of the highest recorded. And yet the temperature and ice volume did virtually nothing. Why? Because there was no dust or dust had only just started. There was insufficient dust to lower ice sheet albedo.

        Or are you saying that dust has no effect on ice sheet albedo?

        And that a lower ice sheet albedo will have no effect on insolation absorption?

        And that increased absorption will have no effect on ice sheet melting and volume?

        .

        >>You … failed to notice that dust production is directly
        >>proportional to ice volume and inversely proportional
        >>to temperature.

        Because I require causation, not simple correlations.

        How does ice volume create dust?
        How does temperature create dust, when the Gobi was wetter?
        How does temperature create dust, when CO2 was the limiting factor on LGM treelines?

        You have a theory based upon a correlation and fresh air. And you cannot change your mind, because you have invested too much time on a failed theory.

        In reality we know that increased dust was caused by CO2 reductions rather than temperature reductions, because CO2 was the limiting factor on upland treelines during the LGM. Ergo, it was dust wot dun it guv…! (sic)

        Ralph

      • This discussion ends here. It is really funny that you will tell me that I am too invested when you are the one that has an article in a Chinese journal that is defending to the death by ignoring the evidence that contradicts it.

        Even Donald Rapp acknowledged that “If the solar power is high enough that it can force a termination without dust”, making dust an non-required factor in several terminations, as I have pointed to you.

        I have no interest in convincing you of anything. Nor do I want to pooh-pooh your scientific efforts, as that is not my job. I have presented the evidence that contradicts your thesis for anybody to look up. Everybody can believe in the power of dust for all I care. People that tend to believe in things unsupported by evidence, like the Young Dryas impact theory, are inconvincible. No amount of evidence will ever convince them since they didn’t require much evidence to believe in the first place. No point in discussing with them.

        Scientists however are a different breed, as they are trained to weight the evidence. No doubt they can be mistaken by the evidence as is the case with CO₂ and climate, but they usually don’t believe in hypothesis that lack substantive evidence. I predict your hypothesis is going to remain in scientific limbo being pretty much ignored while the field advances in a different direction. But you can still be happy by convincing a bunch of people over the internet that your hypothesis has merit. Keep at it and good luck.

      • >>This discussion ends here.

        So you don’t want to acknowledge that you were using precessional insolation in your theory all along. Well, that matters not.

        I think most people could see that interglacial initiations were caused by high Milankovitch insolation – which happens to be a combination of obliquity and precession when eccentricity is high. As I explained in Fig 2 of my paper.

        R

  63. Ellis,

    All we need is dust before the interglacial

    a small dust-blip in the SH, as recorded here, may actually mask a very large dust-blip in the NH.

    So the amount of dust (strength of purported cause) is unrelated to the effect (melting at termination). The fact that the cause takes place thousands of years before the effect is no impediment. And if the amount of dust detected in Antarctica is small, it doesn’t matter to the hypothesis because the unrecorded dust in the NH could still be very large.

    I didn’t know we were dealing with pixie dust that is invisible and can perform magic. If that is the case I withdraw my objections. Pixie dust is really powerful stuff and can drive terminations.

    The real problem for the dust hypothesis, besides being weak on evidence is that the available evidence does not support a causal effect for dust on terminations.

    In just 11 terminations we have 5 of them without high dust levels (purple marks), 6 instances of high dust levels that did not produce an interglacial (blue marks), and one where the purported cause an effect are separated by many thousands of years (orange mark).

    However if we just flip the temperature data we discover the correct correlation:

    High dust levels correlate very strongly with deep glacial, very cold, conditions.

    There are two non-exclusive explanations, both of which are presented by Lambert et al., 2008 who are the scientists that obtained the dust data from the Epica ice cores. One is that dust levels are a consequence of deep glacial conditions that favor the production, transport and accumulation of dust. The other is that dust can affect the radiative balance of the atmosphere by absorbing or reflecting incoming solar radiation. This second one would actually go against the proposal that dust levels favor glacial termination. They might produce mainly a cooling effect and that is why they correlate so well with the inverse of temperature.

    • Javier…

      You don’t seem to understand this theory at all.

      Of course you will get high dust when temperatures are low, but that is not directly due to temperature, it is due to low CO2. It is low CO2 that causes dust, by killing all the flora on the Gobi plateau.

      And…

      Yes of course you can have dust without an interglacial, because you also need coincident high insolation.
      Yes of course you can get a delayed interglacial, because the dust is still there in the ice sheets. As long as the top few layers can be ablated by a new insolation maximum, the old dust will be exposed.
      And as I said before, if you look at Donald Rapps laser counter dust dataset, you will see that there was dust before all those proto-interglacials.

      As to the dust datasets, yes it is a shame that we do not have a full dust record for the NH. However, the NH record we do have from the last glacial period conforms very closely with SH data – excepting that NH dust levels were ten times heavier or more, than SH dust.

      And transient temperature fluctuations, like the many NH D-O events, did not show up in the SH. Nor, to a large extent, did the YD cooling, which only shows as a small dip instead of catastrophic cooling. So yes, the SH temperature and dust record can and does mask much larger variations in the NH. That is a fact that you can confirm yourself.

      Sorry I cannot conjure up a pristine NH dust dataset for you, but some things are even beyond my considerable abilities…. ;-)

      Ralph

      • “It is low CO2 that causes dust, by killing all the flora on the Gobi plateau.”
        What evidence do you have that the flora of the Gobi plateau was killed by low CO₂? Because alternatively it could have been killed by low temperature or low precipitation, or both. Evidence, please. Making unsupported statements actually damages the standing of a hypothesis. How do you distinguish between those possibilities to blame CO₂?
        “Yes of course you can have dust without an interglacial, because you also need coincident high insolation.”
        Show that there was no high insolation at those times. It is going to be tough as you claim the dust can wait in place for thousands of years waiting for insolation to become high.
        “As long as the top few layers can be ablated by a new insolation maximum, the old dust will be exposed.”
        Are you sure the dust can wait 5000 years for the insolation? Because a P-38 that landed in Greenland in 1942 was under 100 m of ice just 50 years later.
        https://www.popularmechanics.com/flight/a22575917/wwii-p-38-discovered-under-300-feet-of-ice-in-greenland/
        How many meters of ice need to be melted to reach the 5000-years old dust that is supposedly driving the melting?
        “Sorry I cannot conjure up a pristine NH dust dataset for you”
        Oh the problem is for you, that lack the evidence to back your hypothesis. The Antarctic dust records support that it is not dust.

      • A problem with the tagging. Here is again:

        “It is low CO2 that causes dust, by killing all the flora on the Gobi plateau.”

        What evidence do you have that the flora of the Gobi plateau was killed by low CO₂? Because alternatively it could have been killed by low temperature or low precipitation, or both. Evidence, please. Making unsupported statements actually damage the standing of a hypothesis. How do you distinguish between those possibilities to blame CO₂?

        “Yes of course you can have dust without an interglacial, because you also need coincident high insolation.”

        Show that there was no high insolation at those times. It is going to be tough as you claim the dust can wait in place for thousands of years waiting for insolation to become high.

        “As long as the top few layers can be ablated by a new insolation maximum, the old dust will be exposed.”

        Are you sure the dust can wait 5000 years for the insolation? Because a P-38 that landed in Greenland in 1942 was under 100 m of ice just 50 years later.
        https://www.popularmechanics.com/flight/a22575917/wwii-p-38-discovered-under-300-feet-of-ice-in-greenland/
        How many meters of ice need to be melted to reach the 5000-years old dust that is supposedly driving the melting?

        “Sorry I cannot conjure up a pristine NH dust dataset for you”

        Oh the problem is for you, that lack the evidence to back your hypothesis. The Antarctic dust records support that it is not dust.

      • >>What evidence do you have that the flora of the
        >>Gobi plateau was killed by low CO₂? Because
        >>alternatively it could have been killed by low
        >>temperature or low precipitation. Evidence, please.
        >>Making unsupported statements actually damage
        >>the standing of a hypothesis.

        If you would do us the courtesy of reading the dust-albedo paper, it would save us all a great deal of time.

        As the paper explains in great detail, much of the Gobi was wetter during the LGM than now. And the temperature-treeline was much higher than the CO2-treeline, during the LGM, therefore CO2 was the limiting factor for high altitude C3 plant life. Ergo: Co2 caused the Gobi to become a shifting-sand desert, not temperature or precipitation. Many climate scientists have failed to spot this basic truism about CO2, so I am not surprised you missed it too.

        .

        >>Because a P-38 that landed in Greenland in 1942
        >>was under 100 m of ice just 50 years later.

        Not during the LGM it would not have been – precipitation was much less during the LGM than now. Besides, ice is a fluid and a heavy object will sink into it, just as it will sink in water. (You can pass a loop of thread around an ice block, with a weight attached, and after a couple of days the loop of thread will pass clean through the ice block.) Have you researched the sinkage rate of metallic objects on ice sheets?

        In the Antarctic, which is a better comparison with LGM precipitation rates and conditions than the Arctic, the ice has the following age-depth:

        20m = 55 yrs. 30m = 85 years.
        90m = 1,200 years. 130m = 2,400 years.

        Depth and Density of the Antarctic Firn Layer,
        by van Broeke.

        Your P-38 challenge does not pass the first hurdle.

        Ralph

      • “In the Antarctic, which is a better comparison with LGM precipitation rates and conditions than the Arctic, the ice has the following age-depth:

        20m = 55 yrs. 30m = 85 years.
        90m = 1,200 years. 130m = 2,400 years.”

        The P38 example does illustrate the problem. After 5000 years your dust is buried under hundreds of meters of ice and cannot trigger a termination that has to melt a huge amount of ice first.

  64. Cut Milankovitch some slack.
    We’re talking about a nonlinear oscillator, which is periodically forced.
    So there’ll be some chaos thrown in.
    Not everything will match to the 5th decimal place.
    But it doesn’t mean that Milankovitch insolation cycles are not the driver.
    They clearly are – or the alignments (e.g. with obliquity with 6500 year lag) are too good a coincidence.

    • >>But it doesn’t mean that Milankovitch insolation cycles are not the driver.

      I think the concensus of opinion here and in most quarters, is that insolation cycles ARE the driver. But not all maxima drive – which is a bit of a problem. Hence Donald Rapp’s search for the X-factor that will feedback and enhance the insolation driver.

      Ralph

      • “I think the concensus of opinion here and in most quarters, is that insolation cycles ARE the driver.”

        The consensus appears to be changing, and obliquity is gaining in importance at the expense of 65N summer insolation.

        Huybers 2006 has almost 300 citations, and Tzedakis, Crucifix, and Wolff are all members of the Past Interglacials Working Group of PAGES.

        Huybers, P. (2006). Early Pleistocene glacial cycles and the integrated summer insolation forcing. Science, 313 (5786), 508-511.

        The field is evolving due to better evidence, but not in the direction where you are.

      • Hybers is another highly misleading paper.

        Huybers created his own paleoclimate chronology, without reference to any other outside chronological pegs. (The benthic record does not record the volcanic pegs in the ice-core data.). And his resulting chronology was much closer to his obliquity-cycle theory than the standard ice-core chronology. Which was handy.

        As mentioned to Clive Best earlier, obliquity and precession work together in ice age modulation, but should not be summed (as in Milankovitch data) as they are quite different in their effects. Precession transfers energy from one pole to the other, and operates in alternate hemispheres. Obliquity transfers energy from the tropics to the poles, and operates in both hemispheres simultaneously. They are quite different.

        What is required is a model that can tune obliquity and precession independently, giving different emphasis to each, in search of the solution that best matches the temperature data. And this can be done independently of dust-albedo feedbacks, if only successful terminations are modeled. By doing so, we could assume that dust-albedo feedbacks were the same in each case.

        Ralph

      • >>The consensus is changing, and obliquity is gaining
        >>in importance at the expense of 65N summer insolation.

        Not really sure what you mean by that. Obliquity is an essential component in 65N summer insolation, especially during eras of low eccentricity.

        Ralph

      • I think the concensus of opinion here is that there is no consensus because we do disagree with each other.

  65. Dimitris Poulos

    Most people, around climate especially, like adventure, they would prefer they were living in an adventurous world, fighting against the evil. This is what they pretend they do with their fight against agw. But this thing agw does not exist but in fantasy. Solar activity has been decreasing since 1950 but temperatures have risen considerably since, indeed. What people forget to look at and I have pointed out in my research and papers is that solar wind is strengthening since 1930 and is the reason for nowadays high temperatures. Between others in my papers I have completely explained the solar wind phenomenon. You can find a collection of my papers in ResearchGate https://www.researchgate.net/profile/Dimitris_Poulos for complete explanation of the solar activity and solar wind phenomena and the derived climate variability. Climate variability has been analyzed and explained for some decades to tens of thousands of years and till the solar wind pause that leads earth to the ice ages for each certain long interval it takes place.

    • “Solar activity has been decreasing since 1950 but temperatures have risen considerably since, indeed.”

      This is not correct.

      Solar activity did not start decreasing until the late 1990’s. The big decrease in solar activity coincides with the pause.

      • Dimitris Poulos

        this is correct of course, why not? the diagram shows it well, solar activity was increasing since 1950 and is decreasing since (except for a steep drop in 1970), the decrease in 2000 for example lead to high temperatures, there was no pause yet. If it was a matter of solar activity alone we would be experiencing 1940 temperatures…

      • Dimitris Poulos

        p.s. in my papers I have even predicted that warming would end in 2000, this has been the only climate model that projects dropping temperatures. CO2 effect on climate has currently been minimal, almost zero, contrary to what most believe.

      • “Solar activity did not start decreasing until the late 1990’s.”

        So this is wrong then?
        Hint: Try extrapolating Leif’s line back to SC 19.
        Which was at a max in the late 1950’s.

        https://i1.wp.com/www.leif.org/research/Ren-Denial.png?zoom=2

      • “So this is wrong then?”

        It is not wrong. The sunspot number is a proxy and it has its problems. For example sunspots are counted and that gives a number, but the area of the sunspots is ignored, assuming that over time the number/area remains constant. That has been shown not to be the case.

        Many authors have observed that sunspots come in different sizes at different times of the solar cycle, and even between different solar cycles. So counting their number is an arbitrary decision that leads to a different result than measuring their area. Robert Wilson has a good article on that:
        Wilson, R. M. (2014). Comparison of the Variations of Sunspot Number, Number of Sunspot Groups, and Sunspot Area, 1875-2013.
        https://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/20140006388.pdf

        His conclusion is:
        “In conclusion, had Wolf used SSA [sunspot area] or G and SSA rather than G and N for describing solar activity, the record would look slightly different, especially as related to the amplitudes and occurrences of SSNmax for the years 1957, 1978 1980, and 1982.”

        Ouch! The proxy looks different if we use sunspot area instead of number. Something that Leif Svalgaard knows perfectly well but ignores because of his strong bias against any solar effect on climate.

        OK. SSN is an imperfect proxy. There is nothing new on that. All proxies have problems that need to be considered. But for the time discussed (late 20th C), we don’t need proxies, as we are measuring solar activity. One of the best ways is 10.7 cm solar radio flux, as direct TSI measurements have a lot of instrumental problems.

        10.7 cm radio flux agrees with SSA. The purported decrease in solar activity over the period of temperature increase is a lie.

        For the entire warming period 1975-2000, solar activity remained at a very high level, and it only started to decline in the late 1990’s, followed by the reduction in the speed of warming that we know as the pause.

      • Salvatore del Prete

        The solar activity needed to cause cooling did not start until late 2005. Prior to that the solar activity was causing warming.

        Now finally in late 2017 it looks like the lag times are in and average value solar parameters are low enough to start a cooling trend which has started as expected and will be continuing for many years moving forward.

      • “For the entire warming period 1975-2000, solar activity remained at a very high level, and it only started to decline in the late 1990’s,”

        And, again, no it didn’t.

        Your graph still shows that SC19 was the peak of solar TSI and every SC since has been less intense.
        Ergo: Solar TSI has been decreasing since the 1950’s (when it was at a peak).
        Othewise we are in Moncktonian logic.

      • “Solar TSI has been decreasing since the 1950’s (when it was at a peak).”

        No. Solar activity decreased from SC19 to SC20, but increased from SC20 to SC21. It wasn’t decreasing from SC19.

        And this behavior breaks your logic because between 1950-1975 temperature decreased, and increased from 1975, as solar activity did.

        Solar activity relationship with temperature rate of change has its problems, which should be expected since temperature is affected by many other factors, but it compares very favorably to CO₂, that has shot through the roof since 1960 while temperature has pretty much kept doing what it was doing.

        So if solar variability is a bad explanation, imagine CO₂ that doesn’t look similar neither in its long term trend, nor in its multidecadal oscillations.

      • Salvatore del Prete

        Javier you are exactly correct. Tony is exactly wrong as usual

      • Solar activity and RF from GHGs are not mutually exclusive.

        But as far as solar activity:

      • “Solar activity and RF from GHGs are not mutually exclusive.”

        Of course not, but there is a competition to be declared the main agent of climate change, and more specifically the main cause of modern global warming. The Sun is supported by proxy evidence, while CO₂ is supported by modern instrumental measurements and mechanism understanding. Professional judges have prematurely declared CO₂ as victor and due to powerful interests and political interference the judgement is currently being upheld. Also, if the Sun is the main agent they will have to recognize that they don’t understand much about climate change, and they rather die (literally, but of old age). But the pause showed that the competition hasn’t ended. As the current solar minimum revives the solar hypothesis we can expect renewed attacks with false memes like solar activity and temperature going opposite ways.

  66. So I simply took a combination of Obliquity and Eccentricity and then played around to get this result when compared to Antarctica temperatures.

    Obliquity defines net annual insolation near the poles, while eccentricity modulates increased June insolation.

    • >>Obliquity defines net annual insolation near
      >>the poles, while eccentricity modulates increased
      >>June insolation.

      Not sure exactly what this does, Clive. Eccentricity does not modulate obliqity, as the red plot in your first graph at the top of this page demonstrates. Eccentricity only modulates with precession. So to plot obliquity insolation only, your red plot would be better.

      https://judithcurry.com/2018/09/08/beyond-milankovitch/#comment-880281

      I think something more complex than this is required. Obliquity and precession insolation should be plotted separately somehow.** Their influence adjustable separately, but in some sort of parallel fashion, so the relative influence of the two can be tweaked to find the perfect coupling. A simple summation of the two is an injustice to the different nature of these forcings, but I am not entirely sure how else they can be represented in parallel.

      Ralph

      ** Trouble is, Laskar does not give separate insolation data for obliquity and precession. But your red plot has already separated out obliquity insolation, so if you subtract the red plot from the basic Laskar-Milankovitch insolation data, you should get precession insolation only.

      • Ralf,

        It doesn’t really ‘mean’ anything. I was just playing with linear combinations of Obliquity (in degrees) and Eccentricity of the earth’s elliptical orbit. Fourier analyses of LR04 and temperature show the two main components are Obliquity and Eccentricity, while Precession is small. This plot just highlights that effect. The plot does not contain any insolation data. However, Eccentricity is the volume control for 65N insolation.

      • I still think dust is the crucial factor once one or two obliquity cycles fail because the ice sheets have become too large.

      • >>Fourier analyses of LR04 and temperature show the
        >>two main components are Obliquity and Eccentricity,
        >>while Precession is small. This plot just highlights that effect.

        That is the trouble with fourier analyses, because the ice age cycle over the last 400 ky is either 90 ky or 115 ky. It is never 100 ky (see my table 1). And I tabulated that data manually myself, from the Laskar and Epica data. So these fourier analyses suggesting a 100 ky ice age cycle have taken many researchers down the wrong path.

        In addition, eccentricity on its own has no forcing insolation effect whatsoever. Eccentricity can only have an effect when it is combined with precession, and so if you want to include the 100 ky influences of eccentricity it HAS to be done via a precessional input.

        Cheers,
        Ralph

    • >>Obliquity defines net annual insolation near the poles

      Clive. While I am thinking about this, there other oddities of orbital insolation.

      Obliquity will warm the NH for the whole of the summer six months, while precession will not (as you noted). But because much of this warming is in the spring an autumn, it is not as effective at ice sheet absorption as precession. Ice tends to reflect the low incident insolation of a NH spring and autumn, and absorb much more during the midsummer midday. So while the obliquity summer is longer than the precessional equivalent, the latter is more effective at melting ice than low-angle obliquity.

      Precession has another oddity. It will warm the end of the spring, for say 3 ky. Then warm the summer, for say 6 ky. Then warm the beginning of the autumn, for say 3 ky. In that order. And I do think I have found evidence for that autumn warming in the record. So simply taking midsummer Milankovitch insolation may be leaving out a vital element of glacial and interglacial warming.

      And finally, precessional insolation is much more effective when obliquity is high, because the Earth’s angle of inclination is much higher. So precession in combination with high obliquity delivers the highest insolation to the NH high latitudes. But this particular component is already included in the Laskar insolation data.

      So any analysis of insolation forcing needs not only to separate obliquity and precession, but possibly also separate the precessional spring, summer and autumn.

      Cheers,
      Ralph

      • Ralf,

        Obliquity warms both NH and SH summers equally. The effect of precession depends on eccentricity and is zero for a circular orbit. So maximum 65N insolation always coincides with a local maximum in eccentricity. Kepler’s second law ensures that the annual change in net insolation due only to eccentricity is essentially zero. However it still plays a role in the timing ice ages because it modulates the grand summer heating. If eccentricity is large the glacial cycle reverts partially back to the obliquity cycle.

        cheers

        Clve

  67. I think this is worth reiterating as a separate post, for discussion. James Hansen wrote that :

    a. Surface albedo should be averaged across the globe, based upon sea-level data, and so he derived a global 3 W/m2 forcing. (See his fig 5.) But as I have argued previously, it is regional albedo that matters in interglacial modulation, not global albedo. Question – what melts the annual winter snows in Canada – is it the regional albedo-driven insolation absorption in Canada, or the albedo and ambient temperature in Argentina? And why would ice age ice sheets be any different?

    b. The value for CO2 forcing is ~3 W/m2, a figure we are perhaps all familiar with. However, that is the total forcing value at the end of 5 ky of interglacial warming. The CO2 forcing during the first century of interglacial warming is only ~0.06 W/m2. Does anyone think that ~0.06 W/m2 is going to drive the temperatures up, so that the next century in the interglacial will be warmer than the last?

    Climate Sensitivity Estimated From Earth’s Climate History.
    http://www.columbia.edu/~jeh1/mailings/2012/20120508_ClimateSensitivity.pdf
    Hansen and Makiko

    Ralph

    • Albedo is regional but has a global effect in proportion to its impact on the net incoming solar radiation. The other parts of earth would not cool in Ice Ages unless it had a global effect. It is not just a regional cooling. To calculate the global effect you need to look at what it does to the global solar input. This is where 3.5 W/m2 comes in. And this is apples with apples when compared to other radiative forcings like CO2. If somehow CO2 rose as albedo rose they would offset instead of adding, and a case can be imagined where they cancel and the global temperature does not change.
      Also, it doesn’t matter that they both take 5000 years to change, 3/5000 and 3.5/5000 are comparable global forcing changes per century if you want to look at it that way.

      • >>Albedo is regional but has a global effect in
        >>proportion to its impact on the net incoming
        >>solar radiation. The other parts of earth would not
        >>cool in Ice Ages unless it had a global effect.

        But we are primarily interested in ice sheet dissipation here (being the definition of an interglacial). And it is regional insolation absorption on the NH ice sheets that controls ice sheet ablation and melting.

        Increasing global albedo by 3 W/m2 is not going to melt an ice sheet 3,000 m thick, but adding 250 W/m2 to its entire surface for each annual summer season will. So why is anyone interested in an artificially smeared-out global albedo value of 3 W/m2 ?

        It makes me think that the big picture has been lost – the big picture being albedo reductions of ice sheets caused by 10 ky of glacial-maximum dust. And of course when you do a word-search of Hansen’s paper for ‘dust’, there are no hits. So the largest of all the ice age feedback systems, is completely missing from Hansen’s influential paper.

        You see what I mean by losing the big picture.

        R

      • Hansen is more interested in the global average temperature that changes by 5 C with a 6.5 W/m2 global averaged forcing change. And it is this quantity for which the CO2 change is also important. Taking other parts of the world into account, the Ice Ages would only have been half as cold without the CO2 amplification effect. So, on the contrary, Hansen has the big picture, the global picture, in mind with his numbers that explain why it got so cold globally.

      • “Insolation at the time of the LGM (21 ka) was similar to today. Nonetheless, the LGM climate remained cold due to the presence of large ice sheets in the Northern Hemisphere (Peltier, 1994, 2004) and reduced atmospheric CO2 concentration (185 ppm according to recent ice core estimates, see Monnin et al., 2001). Most modelling studies of this period do not treat ice sheet extent and elevation or CO2 concentration prognostically, but specify them as boundary conditions. The LGM radiative forcing from the reduced atmospheric concentrations of well-mixed greenhouse gases is likely to have been about –2.8 W m–2 (see Figure 6.5). Ice sheet albedo forcing is estimated to have caused a global mean forcing of about –3.2 W m–2 (based on a range of several LGM simulations) and radiative forcing from increased atmospheric aerosols (primarily dust and vegetation) is estimated to have been about –1 W m–2 each. Therefore, the total annual and global mean radiative forcing during the LGM is likely to have been approximately –8 W m–2 relative to 1750, with large seasonal and geographical variations and significant uncertainties (see Section 6.4.1).”

        This is AR4 – I didn’t read AR5 for reason. The average temperature change was some 10K. With immense regional differences and great uncertainty as stated. Based on current feedbacks I’d guess that forcing change is at least 100% underestimated. Probably to do with cloud geophysics.

        Here is a much better ‘big picture of the LGM than anything found in this post.

        http://www.dandebat.dk/eng-klima5.htm

        Jiminy is as usual completely ordinary.

      • So they say that dust has a net negative effect which is interesting. If it is net negative it can’t get us out of Ice Ages. Discuss.

      • I have a standard reply for people who say discuss. They are inevitably dogmatic pedants and I don’t.

      • Likewise.
        We are not kids in class.

      • >>Ice sheet albedo forcing is estimated to have
        >>caused a global mean forcing of about –3.2 W m–2

        Indeed, that is all very standard. But that is not necessarily what happens on the ground. What we need for ice sheet melt is not global temperature, but insolation absorption on the very ice sheets themselves.

        So these papers that merely quote 3.2 W/m2 are not explaining anything. 3.2 W/m2 is not going to melt the surface of the Laurentide ice sheet, some 2,000 m above the Great Lakes (an increase of 0.6%).

        R.

      • What you need for CO2 levels to change is for the global temperature to change. That is the connection. CO2 is a globally well mixed gas and its concentration depends on the global mean temperature. The global mean temperature depends in turn on the global mean albedo and GHGs. If you ignore the global energy budget, you are missing a key connection that explains the CO2 levels.

    • >>Hansen said:
      >>Surface albedo should be averaged across the globe,
      >>based upon sea-level data, and so he derived a
      >>global 3 W/m2 forcing.

      And here is another error in the Hansen paper. By taking sea levels as a proxi for albedo, during the interglacial warming period, he is only calculating the reduction in albedo due to ice sheet retreat (exposing new bare ground). He is not including a value for dust-albedo on the surface of the ice sheets (Hansen does not mention dust once).

      This makes a large difference, because at the very beginning of the interglacial the value for ice-sheet retreat albedo is minuscule, and increases as it proceeds, as does the CO2 feedback (see his Fig 5). In contrast, the dust-albedo feedback would be there in full force from the very beginning of the interglacial. So in rough figures we get the following feedback values, for the first century of the interglacial warming period:

      Hansen:
      CO2 ~0.06 W/m2
      Albedo ~0.06 W/m2
      Total ~0.12 W/m2

      Ellis:
      Albedo ~250 W/m2 (measured regionally on the ice sheets).

      So which scenario will precipitate a interglacial ice sheet melting, the Hansen theory or the Ellis theory? As I said previously, an increase of ~0.12 W/m2 per century is not going to warm that particular century, to continue the global warming into the next century. Nor is it going to melt an ice sheet some 3,000m thick.

      Ralph

      • You seem to be conflating the Milankovitch mechanism with global forcing and these are not contradictory. Milankovitch explains local albedo changes, but the global response in temperature depends on the global mean albedo change. No one disputes that the albedo change is local but it does impact the global energy budget.

  68. Ralf Ellis, @ https://judithcurry.com/2018/09/08/beyond-milankovitch/#comment-880546

    Firstly, orbital eccentricity is low for the next 100 ky or more, and so there will be no very deep NH Great Winter to force the NH (and the world) into a new ice age. Having said that, we do have the same orbital configuration now, that has caused glacial initiation in the past – I estimated 500 or 1,000 years before the new ice age will arrive. But orbital conditions then improve again, so it is difficult to know if this new ice age would be sustained anyway.

    Thank you. I am thinking about this. But I am sceptical. I think the next ice age is virtually inevitable. Reasons:

    1. We are past the peak of the current interglacial. I understand every interglacial has been followed by a glacial period within about 40 to 120 ka. What is the maximum length of an interglacial?

    2. Human caused increase in CO2 emissions cannot continue for much more than a century or two. This is a tiny blip in geological time. It is insignificant, in the context of the 40 to 120 Ma cycle of glacial and interglacials.

    3. It seems the location of the tectonic plates and ocean gateways will not allow much warming. It will be tens of millions of years until their realignment is suitable for warming. See Scotese 2018 https://www.researchgate.net/publication/324017003_Phanerozoic_Temperatures_Tropical_Mean_Annual_Temperature_TMAT_Polar_Mean_Annual_Temperature_PMAT_and_Global_Mean_Annual_Temperature_GMAT_for_the_last_540_million_years charts of:

    a. the pole to equator temperature gradients at various times from now until 540 Ma ago

    b. the locations of the continents at various times from now until 540 Ma ago. Note that, from 43 to 233 Ma ago there were no ice sheets at either pole. South America was connected to Antarctica, so there was no current circulating Antarctica preventing warm water from the tropics warming the south polar regions. Australia was also connected to Antarctica during the warmest times (before 70 Ma ago), and there was no ice mass at the south pole before 140 Ma ago.

    You can move the slider to see the configuration of the continents plates better here: https://www.youtube.com/watch?v=tObhGzHH2aw

    The Antarctic current won’t be blocked off for some 300 Ma. However, Antarctica moves away from the South Pole by about 80 Ma in future and the perhaps the configuration is better suited for ocean currents to transport heat from equator to poles around 80 Ma in future. See (move the slider): https://www.youtube.com/watch?v=bQywDr-btz4

    In short, Earth could be stuck in an icehouse phase for 80 Ma.

    Please discuss.

    • I meant to include this:

      Restrictions on ocean currents. Continents act as barriers to the flow of oceanic currents. The oceans transport a substantial amount of heat poleward (a maximum of 2–4 x 10^15 watts at latitudes of 20–30 degrees, representing one third to one-half of the total poleward heat transport by the oceans and the atmosphere). The distribution of land can block poleward heat transport by the ocean and may influence polar climates and the subsequent extent of ice and snow cover. Thus, the shape and size of the ocean basins become a factor in controlling global climate.

      https://pdfs.semanticscholar.org/2aee/e8766625b0dc837384e9cb1b25c429990538.pdf pp.6-7

      • Thus, the shape and size of the ocean basins become a factor in controlling global climate.

        This is exactly correct! Shape and size determines how much warm tropical water can be available for snowfall in Polar and near Polar places and determines how much ice gets sequestered on land in cold places.

        Polar oceans are covered with sea ice while the oceans build depth and energy. When the oceans are deep enough and warm enough, the sea ice is thawed and the evaporation and ice production is started and continued long enough to turn the warm ocean water into ice on land. When the oceans are low and cold enough, the sea ice is formed again and the extra snowfall is turned off while the ice reflects and thaws and as the cold ice age plays out.

      • Ice formed from evaporation of warm water in cold places produces ice for ice ages. Ice thaws faster than it is replaces by snowfall during cold times, it is cold because the ice is thawing and reflecting. Cold times continue to get cold and stay cold because the ice is still flowing fast enough to maintain ice extent.

    • >>We are past the peak of the current interglacial.
      >>I understand every interglacial has been followed by
      >>a glacial period within about 40 to 120 ka.

      Yes, but you don’t mention eccentricity. Our current low eccentricity means the Great Winter will be very mild. That is why I like this terminology, because it is very descriptive. And in a mild Great Winter, the chances of a full glacial period are much lower.

      See the colourful Clive Best graph above, that includes an eccentricity plot.

      Ralph

      • Ralph

        Thank you for your reply. I do not have a detailed understanding of the subject as you, Clive Best and Donald Rapp have. To help me understand, can you tell me:
        1. How long do you estimate the interglacial will last?
        2. What has been the longest interglacial in the past 800,000 years?
        3. If you are projecting that this interglacial will be much longer than any previous interglacial, what is causing this one to be so much longer?
        4. Is it credible, given the positions of the continents, the ocean gateways, and the Antarctic current?

        I’ll add another comment on the Antarctic current.

      • The longest interglacial was MIS-11, some 400 ky ago, which lasted nearly 35 ky. This is almost the full width of the obliquity cycle, because eccentricity was low and there was no strong precessional Great Winter to interrupt it. The Holocene is similar to MIS-11 because it too is a low eccentricity interglacial, which is why it is also extended.

        The end of the Holocene depends upon a reduction in insolation values for a combination of precession and obliquity. We will stand at the point of glacial inception in about 1 ky. But as I said before, it is uncertain if we will tip over the edge because this is (will be) a very weak Great Winter.

        R

      • Ralph,

        Thank you. For this clear explanation. I understand as follows (plus my own interpretations of its significance):

        The longest interglacial was about 35 ka. It occurred about 400 ka ago. Obliquity and precession were similar to now. Therefore it is quite possible that the current interglacial could extend to 35 ka (about another 25 ka from now?)

        The point of glacial inception is about 1 ka from now. Thereafter, Earth may or may not cool to the next glacial maximum or the interglacial may continue for about another 25 ka, before tipping over to descend into the next glacial maximum.

        Either way the next glacial maximum is likely to occur about 82 or 123 ka since the last glacial maximum (which was about 21 ka ago).

        The ocean gateways and the positions of the continents have changed little since the start of the current icehouse phase, so these cannot provide conditions to lift Earth out of this icehouse. As mentioned in a previous comment, projections of the future locations of the continents suggest conditions to end the current icehouse phase will not occur for at least another 80 Ma (c.f. the last severe ice house period lasted 70 Ma) https://www.youtube.com/watch?v=bQywDr-btz4 .

        The CO2 contribution from burning fossil fuels over 2 to 3 centuries is an insignificantly short time, so cannot lift Earth out of the current severe icehouse phase.

        Therefore, there is negligible likelihood of catastrophic warming. But future cooling could be devastating for civilisation.

    • Peter
      1. We are past the peak of the current interglacial. I understand every interglacial has been followed by a glacial period within about 40 to 120 ka. What is the maximum length of an interglacial?

      Correct. Since the system is a periodically forced nonlinear oscillator, the precise values of insolation that everyone agonises over, are relatively unimportant. It is the position of the whole system in its phase space that matters. Tzedakis noted that the interglacial 400,000 years ago was, like ours, at a node of low eccentricity. So the insolation did not decrease as much at the end compared to other interglacials that took place with more eccentricity variation. But this did not stop glacial inception. (“The interglacial was not prolonged by subdued insolation forcing” to use the author’s words.) This is an important insight. “Subdued insolation forcing” won’t prolong this interglacial either. The current permanent cold anomaly in SSTs around Antarctica could indicate the very early stage of glacial inception. Glacial inceptions and terminations are led from Antarctica.

  69. Quote from a friend:

    “… has sent around an article trying to calculate warming in the upper ocean as one of the key effects in global warming. This article omits the key thermo-dynamics of Antarctic Bottom Water. It is hard to produce a full mathematical thermodynamic model for the oceans using extensive observations as the Argo Buoy system of a few thousand buoys is only giving us a profile of the upper 2 kms.

    South America fully broke away from the West Antarctic Peninsula forming the Drake Passage less than 20 m yrs ago. This event caused the circumpolar current to form; a current of around 100 million cubic metres of water moving per second and the largest ocean current on Earth. The boundary separating the Earth’s oceans to the north from the Southern Ocean is called the Antarctic Convergence. At this circular boundary around 45 degrees S to 50 degrees S the temperature drops from 5.6C to less than 2.0C within about 50 kms (quite a spectacular change forming one of the largest ecological barriers on Earth). Below this boundary any pack ice that forms is less salty than the ocean water it comes from so the resultant surface waters are then more salty and heavier and sink to form the largest single body of ocean water on Earth.This “bottom water” that sinks and creeps north is initially around minus 0.5C. Over millions of years it has continued to form the bottom 2kms of deep water in the Atlantic, Pacific and Indian Oceans as far north as 10 degrees N!!!

    Since the formation of the Antarctic Convergence the Antarctic Bottom Water has been not only cooling the ocean, but probably has been the main factor in cooling the Earth over the last 6 + million years and forming a cooling trend over-riding the shorter temperature changes caused by recent ice ages. As the Earth cooled it first reacted to slight extra cooling effects of the 41,000 yr changes in axial tilt and then, when even cooler, it was sensitive to extra cooling variations caused by the 100,000 odd yr Milankovich cycles due to variations in the shape of the elliptical orbit. Note from the figure below that in the paleo-ocean temperature graph (of O18/O16 isotope variations) the 41,000 tilt cycles appear nearly 2 million years before the 100,000 year cycles in the last 1 million years. (It is not as if these cycles have not been around for millions and millions of years, but rather the Earth has to get to certain lower temperatures before their cooling effects are noticeable in the paleo-temperature record).

    Finally, I see no evidence that the oceans have been a key leading driver in modern global warming but only that variations in upwelling and sinking of ocean water can cause short-term cyclical climatic changes due to interchanges of heat with the atmosphere. The clearest cycle being the Pacific Decadal Oscillation of roughly 60 years with half cycles around 30 years. In Australia we are very aware of the more unpredictable El Nino-La Nina variations. (And there are other similar cycles in the Atlantic and Indian Oceans).

  70. Here is something else that may need highlighting. There is a common belief and argument in many recent paleoclimatology papers, that ice sheets become ‘unstable’ as they grow larger. Javier has used this argument in one of his graphs above.

    This ‘instability’ is apparently due to the weight of the ice sheet pressure-melting the ice at its base, so that it forms a low-friction semi-liquid basal region that can easily slide over the underlying silts (if present) or rock surface. This causes ‘instability’, so presumably we are to imagine that the ice sheet slips (southwards) and into the sea or to lower latitudes, where it melts.

    But for some unspecified reason this sudden slippage does not end when the ice sheet reduces in thickness and weight, it just keeps on slipping and sliding until the whole ice sheet slides into the sea. Or something like that. And there are a whole raft of science papers that quote this ice sheet ‘instability’ including this one.

    https://www.researchgate.net/publication/327112341_A_theory_of_Pleistocene_glacial_rhythmicity

    As a theory this lacks evidence and reasoning. The greatest problem, as far as I can see, is that large ice sheets do NOT simply run over the surface of the silts or rocks below – riding on a layer of semi-liquid ice and crushed rocks. They actually flow over the mountain tops, as the radar image below demonstrates. So a friction value for basal sliding (ice sheet instability) is of little or no value here, because the ice is not flowing over the rock surface. It is flowing over the mountain-tops. What is more important here is the viscosity of the ice, as it flows and extrudes due to the weight above.

    But even if the weight of ice above made the ice flow quicker, there is no reason why that flow would not slow when the weight is removed. In short, there is no reason to presume that an ice sheet will become ‘unstable’ as it grows, and collapse completely. It is far more likely that if viscosity and basal friction were a problem, the ice sheet would merely oscillate around an optimum size and thickness.

    Ralph

    The plume erupting in the first image (below) is due to super-cooled high-pressure water erupting in a fountain through ice fissures, and refreezing in the upper layers. Yet the ice layers above simply flow over this obstruction, just as the layers in a stable atmosphere will flow over a rounded hill – producing a lenticular cloud. So, just as the upper layers in a stable atmosphere are not retarded by ground contact (the geostrophic wind), so the upper layers of the ice sheets are not retarded by the underlying rock surface. (Lets call this the ‘geostrophic ice flow’, as ice sheet flow will be effected by Coriolis as much as the wind is…)

    • What is more important here is the viscosity of the ice, as it flows and extrudes due to the weight above. This is an important point.

    • Ice sheets never become unstable. Ice sheets thaw, based on sunshine and exposure (ice extent) and the ice thaws until it is gone because it does not snow enough to replace the ice that thaws every year when oceans are low and cold. After the ice sheet thins, it retreats because it ran out of ice. Thick ice sheets do not retreat or become unstable.

    • Ralf
      What is missing from your consideration is the chaotic-nonlinear nature of the climate system, including glaciation, and it’s switching between a limited number attractors within its phase space. It’s not just you – this paradigm is conspicuously absent from this whole discussion. And further – from climate science in general.

      This disappoints me – chaotic-nonlinear dynamics are well enough understood but the wider science community seems satisfied to bottle it up is small fields such as chemical engineering. Everyone is scared to let the genie of chaos-nonlinearity out of the bag and acknowledge its central role in complex systems such as climate.

      With a chaotic-nonlinear paradigm, it’s not so hard to envisage dramatic system changes between attractors, triggered by movement of the system to a particular part of the phase space. In this paradigm, the multidimensional phase space needs to be considered. Then big transitions are no longer about mechanistic discussions of single factors – my factor against your factor. It’s all factors together. This isn’t just “why-cant-we-all-just-get-along” – this is a more relevant and fruitful chaos-nonlinear paradigm.

      Here is a good paper showing how long term ice build up over several precession (or obliquity) cycles can lead to instability and set the stage for rapid reversal and glacial termination (whthout necessarily acknowledging the chaos-nonlinear dynamics underlying it):

      http://moraymo.us/wp-content/uploads/2014/04/abe-ouchieetal_20131.pdf

      • Philsalmon.
        The whole idea of the paper was to strip out the myriad complexities, and see if ice age modulation could be brought down to a couple of parameters. And it can, because the dust-ice-albedo feedback theory is entirely logical and rational.

        Of course we could add complexity, to explain every lump and bump in the cycle, like all the 65 yr and 1,000 yr cycles we see in the midern era. But that is not important. What matters is that the big ice age picture can be explained just with orbital cycles, CO2, dust, and albedo.

        R

      • The chaos perspective is about complexity but actually provides simplicity.
        See the book “Deep Simplicity” by John Gribbin.

      • I am hardly one to decry deterministic chaos – in the scientific sense something very different to mere complexity. In principle the system behaves in ways diagnostic of complex interactions of powerful sub-systems. It includes SAM and NAM, ocean gyres, regimes of atmospheric circulation, shifting patterns of rainfall, AMOC, heat transport and perhaps dust. Perhaps wild cards like asteroid impacts that shift the Earth itself like a child’s spinning top knocked of course before regaining its equilibrium.
        While 65N insolation may set the scene – there are powerful and diverse geophysics at play in both glacial inception and termination. In a dynamically complex system there is no simple cause and effect and the whole is more than the sum of the parts. These are simple rules at the heart of Earth’s complexity.

  71. Nice work, Donald Rapp, thanks for sharing — I’ve been curious about the role of dust since encountering the subject in the Last Glacial Maximum wiki.

    • But you will not find anything about the dust-albedo theory on the Wiki page. This is because the Wiki moderators said:

      … My paper was not peer-reviewed by a proper university (Beijing).
      … That the theory was a denier theory.

      If anyone would like to add components of this theory to the Wiki page I would be grateful, but I don’t suppose they will stay there too long.

      https://en.wikipedia.org/wiki/Last_Glacial_Maximum

      Ralph

  72. Here is the ice dynamics that needs explaining in Europe in an animation

    • That animation is far too fast.
      Cannot see what is going on at all.

      R

    • This is REALLY interesting for me. It appears the ice cap did not start to build in Northern Europe, from the coastal Western Regions, until 40Ka BP and started to recede at 22Ka BP as the interglacial sets in. Finally gone by 8Ka BP. So the ice cover of Northern Europe only occurs for some 32 Ka, even in the most extreme Norther and coastal areas? Much less for the more inland southerly parts of Northern Europe. Is that correct?

      Proably worth considering against the below 4 ice age overlay. As are statements here about how ice ages vary. Not really……They have perturbations, yes, but the strong basic controls of the cyclic periodicity pull them back onto the steady track of the unperturbed cycle. Stable ice age floor, then 100Ka perturbations over 7Ka, stopped by clouds from the oceans almost as suddenly as an electric kettle when it gets too warm = nowish, cooling back to ice age as the effect(s) disipate. Just does. The puny 1.6 W/m^2 in atmospheric insolation variation hom sap may be producing with CO2 is noise in the water vapour dominance, or not, is not going to change that, it is itself negated by the natural response of clouds to decrease insolation and stabilise surface temperatures by increased evaporation, as other small perturbations on the main cycle are regulated to the interglacial set point while the warm phase is predominant, probaly while the lingering drivers persist. NB: The insolation effects on the NASA energy budget are not fixed, they vary with the absolute conditions. That’s the control system between key influences I think Peter Lang suggests precipitates and also limit the changes.

      When I find a way to publish my paper around the PALS review mafia, a solid mechanism that can deliver this repeatable MIlankovitch cycle effect will be revealed for your discussion. Any real scientists out there have ideas for open minded sceptical review that are not the OMICS peer review cowboys? I can always be messaged on Twitter as catandman. My email is not hard to find.

      Anyway, below is the static picture showing how Ice ages really are consistent in general form, with excursions. I can even explain some very good reasons why the temperatures and sea levels of the Younger Dryas happened as they did, and how the cycle recovered to track others. If you look at such a system with the telescope back to front, you will see all sorts of noisy nonsense, and the big picture if you use it the right way around, over many lifetimes. Real change does not happen on human time scales, so academics have to scam it up for grants and to support CO2 reducing laws that make cynical people lots of money to no useful or measureable end.

  73. Or simply save it on the desktop and it appears as separate still gifs? Click and drag works on a Mac.

  74. Donald Rapp, Ralph Ellis, Clive Best and CE Denizens,

    I circulated Ellis and Palmer (2016) to an email group today. I have just received the following email from Dr Howard Brady, and post it with his permission (see link to his latest book below):

    1. I notice correspondence re the Ellis et al., paper: some comments:
    2. In the first place there is a present fad to interpret even small climate change in terms of C02 change which even if one followed the logarithmic calcs on the effect of small changes in C02 on absorption do not amount to anything significant; these C02 explanations do not amount to anything other than story telling;
    3. One has to explain that C02 lags T in ice cores by 600 years from 0-400,000BP and about 1200 years between 400,000 BP and 800,000 BP. This is not a small lag easily explained away and suggests major C02 changes in these cores are driven by temperature change;(we cannot go beyond 800,000 as it looks like any older ice is deformed or not there in the ice caps);
    4. I note also the comments on dust. I think the amount of polar dust in ice ages was huge. During my M.Sc. and Ph.D. that were based studying Antarctic sediment cores during 4 mainland expeditions, when we had strong ice ages conditions the drop in sea level was so strong that thousands of sq km of shallow sea floor must have been exposed to wind. Even in glacial deposits 200kms inland we were picking up by microscopic analysis large amounts broken diatoms that must have come from marine sediments as part of wind-blown debris; and of course there was also dust in the ice cores;
    5. Leaving Antarctica if one, for example, goes to the Jervis Bay region(which I know) the drop in sea level per ice age was around 140 metres and again the sea floor was exposed for over 20 kms seaward. The winds must have been stronger than today as we had the development of dunes hundreds of feet high at this time; so as the larger sand grains were deposited, there must have been finer dust blown all over the place, that may have been a feedback adding to cooling;
    6. I note also a comment re cosmic rays. Ever since the work of Bond we have been studying sediment cores stretching back over a million years with layers of high Be10 and high C14. These layers correspond strongly to cold events. There is too many of them not to see a link with cosmic radiation which alone has the power to create muons to hit molecules in the lower atmosphere or to collide with surface sediments to create such isotopes;
    7. Over a longer time scale we now have over 25,000 microfossil analyses over the last 500,000 years where O18/016 temperature data shows cold ocean temperatures that coincide with each transit of the solar system over arms of the Milky Way. This evidence requires explanation and certainly points to a cosmic wave relationship;
    8. While major fluctuations within the recent ice ages can be linked to Milankovich curves, these are really small-time changes within a much larger non-linear system with many, many other cycles at play;
    9. Finally before getting totally wrapped up in Milankovich curves, there are also variations of the Sun’s distance from Earth up to 2 solar diameters due to the gravitational effect of Jupiter and Saturn; and the cycle is 179 years!! These 1-2% variations in distance are enough to have small modulating effects. The great Australian scientist Pro Rhodes Fairbridge who worked mainly in New York complained that these effects should not be overlooked; as has Nicola Scafetta a well known Italian physicist at the present time.

    best wishes,

    Howard Brady
    Author of Mirrors and Mazes: A Guide Through the Climate Change Debate
    http://www.mirrorsandmazes.com.au/

  75. Thanks for those comments by Dr Brady. Yes, I would concur on his assessments here. With a couple of caveats:

    Exposed sea floor will not stay exposed for long, as flora will recolonise within a couple of years and prevent increased dust-flux. Besides, the isotopic analysis of Greenland dust and the evidence from the Loess Plateau have demonstrated that the dust came from the Gobi. So now we have to explain how the Gobi was producing copious amounts of dust, even though the Gobi was wetter at the LGM.

    The answer I discovered, which seems to be new to all of palaeoclimate science, is that CO2 became the limiting factor for high altitude C3 treelines. If you read my paper, a number of respected treeline studies got into a right muddle, because they could not understand why the treeline came down so far during the LGM. They presumed temperature was the cause, and ended up with nonsense lapse rates to achieve this. These nonsense LGM lapse rates and temperatures have ended up in all the climate models, and so all the climate models are wrong at their very foundations.

    The true answer is that C3 treelines declined because of low CO2, not low temperature. And if we include low CO2 in our climate calculations, then the reason for the high Gobi plateau becoming a desert is obvious. The Gobi became a CO2 desert, not an aridity desert, and C4 plants were unable to recolonise because it was too cold up there (C4 plants tend to be tropical).

    So the fundamental driver of interglacial warming is LOW CO2. It is low CO2 that causes the dust; it is the dust that lowers ice abedo; and it is the low abedo that causes the interglacial warming. So climate scientists were right, CO2 does cause global warming – but only by getting so low that all upland plant life dies…

    Modulation of ice ages via precession and dust-albedo feedbacks
    https://www.sciencedirect.com/science/article/pii/S1674987116300305

    Ralph

    • I will show a different and deterministic physics approach to how the reliable in amplitude and period interglacial is triggerred and driven over 7Ka, once I find a non PALS publisher that isn’t a dodgy rip off business. This approach includes the psoitive warming effects of both dust and ice albedo reduction, maybe a little CO2 lagging GHE – while it is still cold, but only as positive feedback after the warming has begun and is picking up energy. But dust and albedo are not considered the causal force that steadily drives the climate to the repeatable interglacial plateau over 7Ka, and is limited there, a few degrees higher, by burgeoning clouds, which can keep forming until the insolation is reduced enough to stop further surface warming. I don’t understand the interglacial heat balance equilibrium but it is clear, there is a lot more of the sky available for clouds so feedback is not out of range by any measure, and the climate holds that strongly latched level until the interglacial energy dissipates and the long term stable ground state is resumed. Simples. Need to publish and be damned now…. just actual data and deterministic physics, no statistical computer models are harmed in my approach. Just BS (bad science)

    • Ralph,

      Thank you. I sent Howard Brady the link to your comment. He is currently moving house so he doesn’t have time to reply to your comment. You and he agree on most points. He had a few things to add, but it would be better if he wrote them himself, rather than me re-posting them. He has a lot to contribute, but just doesn’t have time at the moment. I suspect he would not be prepared to make comments without having read what has already been discussed and carefully preparing a comment.

  76. We need go way beyond Milankovic. There are two or three fundamental ideas of how the Earth system works. Unless forced it doesn’t change. Most of these forcings are anthropogenic with a smidgeon for solar. To that can be added the purely periodic component – cancelling out in short order. The random can be dismissed out of hand – it doesn’t add to zero either. Here’s one of my favorite pictures from one of my very favorite scientist and hydrologist Dimitris Koutsoyiannis.

    The other idea is abrupt climate change. It was popularized by a committee of illustrious climate scientists way back in 2002 – in the US NAS report “Abrupt climate change – inevitable surprises. Climate happens in regimes and shifts even at decadal scales. Over much longer periods it gets more extreme – as in the Nilometer data.

    Computers can be used for many things. There is or was a twitter feed above on modelling that wetlands could expend under sea level rise if there is enough room to expand. I came across a copy that fell of the back of a truck. As a coastal engineer of many years standing I can only point out that this was known 100 years ago at least. I wouldn’t surprised if it is even ancient. Our Australian original peoples has intergenerational quality controlled stories going back to the last marine transgression.

    Computers cannot of course spit out a single deterministic solution on a centennial scale – on account of they are chaotic and the initial conditions imprecisely known. They can spit out 1000’s of solutions – but that’s not much use. And a waste of time when they could be doing so many other fun things with them.

    And it’s got my ire up. It’s all the fault of AGW pissant progressives. The leaders are from fields outside of Earth sciences. Their followers are scientifically naive and technically untrained. This – and a manufactured consensus around the most basic of geophysics – gives them carte blanche apparently to hector and tell eminent scientists even – how they should talk about climate change.

    But if you are talking talking predictable – we get back to inevitable surprises. As I said – in such dynamical complexity multiple causes and effects should be ruled out before settling on a singular cause and effect. This is hard to do with paleodata.

  77. Yes let’s talk cloud. There is a geophysics mechanism in Raleigh-Benard convection in a fluid heated form below. “Marine stratocumulus cloud decks forming over dark, subtropical oceans are regarded as the reflectors of the atmosphere.1 The decks of low clouds 1000s of km in scale reflect back to space a significant portion of the direct solar radiation and therefore dramatically increase the local albedo of areas otherwise characterized by dark oceans below.2,3 This cloud system has been shown to have two stable states: open and closed cells. Closed cell cloud systems have high cloud fraction and are usually shallower, while open cells have low cloud fraction and form thicker clouds mostly over the convective cell walls and therefore have a smaller domain average albedo.4” https://aip.scitation.org/doi/10.1063/1.4973593

    Closed cell cloud persists disproportionately longer in cooler conditions – before raining out from the center.

    • “Most of these forcings are anthropogenic with a smidgeon for solar.”

      Did you mean that, because it is obviously wrong if as it reads. What is the technical basis for this assertion? No they are not. Most change in the record is natural, human’s have only ever inhabited a very few degrees range for short few Ka sitting on top of one short interglacial event, initiated by a serious perturbation that has occured every 100Ka Milankovitch eccentricity cycle years for the last 1 million years to lift the planet off the stable ice age floor temperature by a few degrees at equable latitudes and raise sea levels temporally 100 metres, and which has occured every 41Ka year Milankovitch obliquity cycle for 2.5 Ma before that. The small changes within this cycle are the hard to predict or model perturbations from all the random odds and sods, but are insignificant in our time and limited by the main controls of periodicity and range of the dominant cycles and controls of their their causes.

      The Younger dryas and other multi Ka perturbations in the cycles were non trivial, but the main cycle pattern still recovers and repeats because the main enrgy sources and controls are so dominant. What happens in human lifetimes is trivial and also immaterial. Not what I think, what the record shows.

      I prefer nature and deterministc physics to the predictions of statistical non linear models built on unprovable and plain wrong assertions, designed to prove a problem exists from one source rather than find out whether there is a problem and test what caused it. etc. The IPCC has simply become the propaganda arm of the renewable energy subsidy racket, avoidably impoverishing people everywhere to no environmental gain by regressive law. Follow the money, its not really about the climate.

      • I was going to say that’s the IPCC shibboleth – as in their well known forcing graphic – but frankly I didn’t think I needed to in a sophisticated forum like this.

        But perhaps more pertinently – there is a climate shift due this next decade – if it is not already happening. Perhaps they could try predicting that… and if they get that right… they can think about decadal forecasts. But that would require 1000’s of times more computing power, the electrical power of 12 coal powered generators and a heat sink the size of Antarctica. Hey that’s an idea…

      • Ok. Accept you menat the opposite, problem is its hard to be less than rigorous in science, especially with trolls around every corner….totally agree about the models. A smiley would have done the trick. Models are complex with their inputs pre-biased to a preset agenda and not accurately reflecting natural systems they pretend to emulate, partly because of built in bias and partly because of the impossible complexity vs modern computer power…GIGO. That OK?

      • FYI Peter Huybers has done a lot of useful work that relates to m own approach to causes of change in my own paper, in progress, and was responsive and helpful when I contacted him regarding details recently.

  78. ralfelis:

    >>The quality of the work of Peter Huybers and
    >>Christos Tzedakis is outstanding. That is why
    >>they publish so often in the best journals.

    The work of Huybers was deliberately misleading… a theory that only explains half the problem, is half baked.

    Palaeoalbedo remains a comparative blank on the geophysical map, because it ( theres no avoiding the word) reflects both palaeecology and the past history of aerosols. Huybers and his coauthors have made this known set of unknowns perfectly clear , and one expects that the subject will become clearer as more time and firld work are invested in uits resolution.

    • “Palaeoalbedo remains a comparative blank on the geophysical map”

      Not just paleo.

      No one knows how albedo varied in the past and the error bars are ? probably large around even today’s point satellite measurements.

      Unknown, presumed constant appears to be more than just Monty Python humor.

      • >>No one knows how albedo varied in the past and
        >>the error bars are?

        Well, we could find out. The image below is of actual LGM dust still in its original ice layers, and so we could do some albedo tests on this.

        This ice has travelled 30 km, and been extruded and thinned. Plus this is northern Greenland, and not the Laurentide Great Lakes where dust levels were ten time higher. But we could still get a good idea of LGM albedo – before and after surface concentration by ablation and melting.

        But nobody has done the research as yet, because nobody thought that LGM dust was important.

        Ralph

      • Well, we could find out.

        No doubt dust is important, but no one will ever no the most important factor of albedo which is cloudiness.

        And I was referring to the fact that this is true for modern times as well.

      • What do you mean, never know? Assertion over fact is the tool of the climate alarmist. Not a real sceptical scientist. At interglacial conditions cloud albedo is estimated to be around 50W/m^2, to add to the evaporation and cloud formation cooling to space of 90W/m^s, so 140W/m^2 of cooling in total from clouds currently, variable in response to temperature change. A strong control at interglacial temperatures. GHE warming may dominate water vapour effect at glacial temperatures – looks like it does but I don’t have that info. . Stephens, G. L., D. O’Brien, P. J. Webster, P. Pilewski, S. Kato, and J.-l. Li (2015), The albedo of Earth, Rev. Geophys., 53, 141–163, doi:10.1002/2014R G000449.

        ftp://ftp.cira.colostate.edu/ftp/CloudSat/cloudsat_pubs/800-899/ID%20872%20Stephens_et_al-2015-Reviews_of_Geophysics.pdf

      • TE,
        thanks for always trying to knock us back on the tracks during the runaway CAGW train.

        You don’t think newer models are better at cloud cover prediction and rain prediction? Beyond 10 days I guess?
        Scott

      • What do you mean, never know? … albedo is estimated …

        And how would you validate estimates of past albedo?

        Albedo is only now being measured from orbiting point observations.

        But reflection is anisotropic, that is it occurs at different amounts depending on angle. So even a point source has margin of error.

      • You don’t think newer models are better at cloud cover prediction and rain prediction?

        I’m not intimately familiar with latest models and albedo estimates, but clouds are relatively very small compared to model grids, making them subject to non-physical parameterization.

        This being the case, I don’t believe there’s any way albedo can be accurate and precise.

        FWIW, here’s a comparison of CERES estimated planetary albedo and that of the GISS models:

        Model is very much in error in absolute terms and the variance of observation is larger than indicated by the model.

        Again, there is uncertainty with the observational estimate.

        But it’s a dirty secret of the uncertainty of the models because we don’t know with precision what the planetary albedo was or is.

      • >>But reflection is anisotropic, that is it occurs
        >>at different amounts depending on angle.

        Yeah, we know. And lots of experimentation has been done on that. But nobody has done it on the actual LGM dust, as depicted in the image above.

        Ralph

      • Anisotropic may be random – like the toss of a coin – so the error must sum to zero.

      • … anisotrophy… ?… may be random…

      • Robert I Ellison: Anisotropic may be random – like the toss of a coin – so the error must sum to zero.

        Where do you get this idea that random variation must sum to 0 over any finite observation intervals? When fitting linear models to data the residuals will sum to 0, but that is true even if the residuals contain non-random bias. When fitting nonlinear models, the residuals will not generally sum to 0 with or without non-random bias. But even if the random variation has expected value of 0, the finite series of observed values will seldom sum to 0.

      • Re: “You don’t think newer models are better at cloud cover prediction and rain prediction?
        I’m not intimately familiar with latest models and albedo estimates”

        You’ve been corrected on this many, many times. You simply make sure to ignore the evidence on the subject. There’s really no excuse, at this point. Once again:

        https://judithcurry.com/2018/07/22/the-perils-of-near-tabloid-science/#comment-877593

        Clouds reflect solar radiation into space and thus can act as a negative feedback in response to warming; clouds also reflect/absorb radiation emitted by the Earth and thus can act as a positive feedback in response to warming. Lower level clouds tend to act as a negative feedback, while higher level clouds tend to act as a positive feedback. Climate models predict a net positive feedback from clouds due to increases in higher level clouds and reductions in lower level clouds in response to warming. That positive cloud feedback has been observed. For further evidence and discussion, see:

        “Clearing clouds of uncertainty”
        “Evidence for climate change in the satellite cloud record”
        “A net decrease in the Earth’s cloud, aerosol, and surface 340 nm reflectivity during the past 33 yr (1979–2011)”
        “A determination of the cloud feedback from climate variations over the past decade”
        “Cloud feedback mechanisms and their representation in global climate models”
        “New observational evidence for a positive cloud feedback that amplifies the Atlantic Multidecadal Oscillation”
        “Impact of dataset choice on calculations of the short-term cloud feedback”
        “Long-term cloud change imprinted in seasonal cloud variation: More evidence of high climate sensitivity”
        “Thermodynamic constraint on the depth of the global tropospheric circulation”

        As noted in the second paper listed above:

        The primary drivers of these cloud changes appear to be increasing greenhouse gas concentrations and a recovery from volcanic radiative cooling. These results indicate that the cloud changes most consistently predicted by global climate models are currently occurring in nature.
        […]
        Our results suggest that radiative forcing by a combination of anthropogenic greenhouse gases and volcanic aerosol has produced observed cloud changes during the past several decades that exert positive feedbacks on the climate system. We expect increasing greenhouse gases will cause these cloud trends to continue in the future unless offset by unpredictable large volcanic eruptions.”

        https://www.nature.com/articles/nature18273

  79. Ralph Ellis,

    Discussion on this thread seems to have stalled. I, for one, am persuaded by your evidence and arguments. You have been very patient answering questions and responding to comments. I have learnt a lot. I really appreciate your contribution. I also thank Donald Rapp for writing the post and others who have contributed to the discussion.

    My main interest is in the policy implications. I am convinced the world is making a major mistake by implementing policies to reduce global warming. From my perspective we have a rare opportunity to try to get Earth to escape from the current icehouse conditions. Humans should implement policies to try to grab this opportunity. It seems there is 1000-2000 to try extend the current interglacial to 35 ka, and 35 ka to lift Earth out of the current icehouse.

    I am sure it could be achieved given the time scales involved and the rates of technology development humans have demonstrated can be achieved. But I suggest we need to reverse policies from trying to reduce global warming to trying to maximise it.

    • Peter: My opinion is that the deposition of soot, dust, dirt, ash, and other debris on northern ice and snow is a greater factor than rising CO2 and in keeping with our theory, will prevent the advent of a new ice age for the foreseeable future. Perhaps in the hypothetical Nirvana when the world runs on solar and wind energy, this might be abated?

      • Donald Rapp,

        Thank you for your reply. I am certainly persuaded that tens of thousands of years of dust dust accumulating each year in the accumulated ice, accumulates at the surface as the ice melts, and becomes denser and absorbs more heat each year as it gets denser. This lifts the planet from glacial maximum to interglacial in around 10,000 years compared with around 80,000 years to cool from interglacial to glacial maximum. I agree CO2 concentration in the atmosphere is a follower of the temperature changes, not the root-cause, and it plays only a small part in glacial teminations.

        I am not yet convinced that the soot that has accumulated over the past 1 or 2 centuries will be effective over a long enough period (thousand of years) to prevent the planet’s descent to the next glacial maximum. Likewise, I doubt the CO2 emissions over the past century will be effective for long enough to prevent the cooling to the next ice age. Ralph Ellis said the tipping point is around 1000 years from now. So, I doubt either the soot or CO2 that has accumulated over the past 200 years, or until we replace fossil fuels with nuclear power this century, will have much effect in 1000 years from now. However, I am certainly open to persuasion on this.

        Thanks again for your reply.

  80. Even if considerably more sanguine about David’s prognostications propensities – I am less than sanguine about his policy prospects. First – if it is so technology driven I can’t see it waiting around for an Earth warming first nostalgia. Secondly – the future is conservation and restoration – on cropping and grazing lands, in grasslands, woodlands, deserts – by adding carbon. At the base is increasing carbon content as organic material. Something of 500 GtC lost since the advent of agriculture – says Professor Rattan Lal.

  81. TE,
    Thanks for the response about clouds and precipitation above.
    Scott

  82. Richard Lindzen

    You might want to look at the following two papers:

    Roe, G. 2006, In Defense of Milankovitch, GRL 33 doi: 10.1029/2006GL027817

    Edvardsson, et al, 2002, Accurate spin axes and solar system dynamics: Climatic variations for the Earth and Mars, Astronomy and Astrophysics, 384, DOI: 10.1051/0004-6361:20020029

    Both make a strong case that Milankovitch works very well.

    • They don’t seem to. First paper says ” The Milankovitch hypothesis as formulated here does not explain the large rapid deglaciations that occurred at the end of some of the ice age cycles” (pg 7). The second paper: “The main period in glacial data (around 100 000 years) is, however, hard to explain in terms of orbital variations.” (pg 1).

      The major problem is that the evidence indicates otherwise. This paper https://agupubs.onlinelibrary.wiley.com/doi/pdf/10.1029/95PA00578 describes a hypothetical mechanism that alters obliquity over long periods. Evidence seem to fit such a theory, except on much shorter time spans and with abrupt changes (one and a half cycles are evident).

      The last 100ky, according to this: https://ars.els-cdn.com/content/image/1-s2.0-S0277379114003485-gr1.jpg have not been smooth.

    • Dear Professor Lindzen: I am not sure what you mean by the assertion that Milankovitch “works well”? Does Milankovitch tell use a priori when an ice age will terminate? Does it explain why termination occurs typically in a mere 5500 years? Does it explain why dust buildup sharply maximizes prior to a termination?
      I have quite a bit to say regarding Roe’s work, but I don’t know how to post graphs on this site. A summary is this… At first glance, the apparent agreement between the slope of the SPECMAP curve and the solar curve is quite impressive. Nevertheless, a few caveats are in order. First, the fact that the SPECMAP was tuned to the solar intensity curve forces the oscillations in dv/dt to match the oscillations in the solar curve. Second, the SPECMAP curve is not very precise, and assessing slopes on the nearly vertical trends can be tricky. Some considerable chartsmanship is involved. Note that there are several places where the dv/dt curves seem to match the sharp oscillations of the solar curve, yet the v curve has very small oscillations. Something doesn’t compute? Is v equal to the integral of dv/dt? Indeed, there are six important time periods in this figure where the apparent slopes (dv/dt) match the solar curve nicely but the ice volume (v) does not match at all. The integral of Roe’s dv/dt does not equal v. Third, the time scale for the SPECMAP is believed to contain serious errors prior to 600 kya. Fourth, Lisiecki, Raymo and Curry (2008) discussed the assumption that benthic delta18O represents the phase of ice volume change despite the fact that benthic delta18O is also affected by deep water temperature change. They also said: “Generating a robust age model for benthic delta18O or ice volume without the assumptions of orbital tuning remains an important, unsolved problem.” Roe’s figure shows that the growth and diminution of ice volume is responsive to solar input to high northern latitudes, but it does not define where Ice Ages should begin and end.
      Roe (2006) also carried out a similar analysis for the ocean sediment record developed by Huybers and Wunsch (2004) that did not utilize orbital tuning. In this case however, the comparison of dv/dt with the solar curve is not so good. At first glance, there seems to be fair agreement between the slopes of the HW04 curve and the solar intensity curve. However, when one expands the graphs to larger size, discrepancies appear. Chartsmanship hides the discrepancies. As before, deriving dv/dt from v is a subjective process.
      Roe’s final conclusion sums the situation well. Roe realized that the Milankovitch theory was highly relevant, but at the same time very incomplete. Roe said:
      The Milankovitch hypothesis as formulated here does not explain the large rapid deglaciations that occurred at the end of some of the ice age cycles: many studies point to the need to invoke internal dynamics of ice sheets as a mechanism for occasional rapid collapses if a threshold size is exceeded. Nor do the results explain the mid-Pleistocene transition between an earlier interval characterized by 40 kyr durations of ice ages and a later interval with 80 kyr to 120 kyr durations… The prevailing view to date has been that ice sheet volume is the most important variable to consider. While this is obviously the case for global sea level, it is ice sheet extent that matters most for albedo, and ice sheet height that matters for atmospheric circulation… Ice sheets are dynamic systems and these properties can vary quite differently from each other. However, the results presented here demonstrate the critical physical importance of focusing on the rate of change of ice volume, as opposed to the ice volume itself. The available evidence supports the essence of the original idea … [that] (1) the strong expectation on physical grounds that summertime insolation is the key player in the mass balance of great Northern Hemisphere continental ice sheets of the ice ages; and (2) the rate of change of global ice volume is in anti-phase with variations in summertime insolation in the northern high latitudes that, in turn, are due to the changing orbit of the Earth.

    • Richard,

      Like Donald, I too was not persuaded by the Roe paper (In Defense of Milankovitch).

      By representing deglaciations via the rate of change in ice sheets, rather than the volume of ice sheets, the graph highlights very minor transitions simply because they occurred rapidly. This is similar to saying that a car changing speed from 50 mph to 55 mph is much more significant than a car changing speed from 20 mph to 70 mph, because the former occurred more rapidly. Yet clearly the energy required to generate these changes was much greater in the latter than the former.

      Nobody is doubting that a strong Milankovitch maximum can (often) generate a reduction in ice sheet volume, as most maxima do indeed generate some kind of ice sheet response. But what the majority of these maxima do not do, is generate a full-blown interglacial. If Milankovitch was the sole answer to this conundrum, we should get an interglacial every Milankovitch maxima, and quite clearly we do not. Since the forcing element of the ice age cycle (ie: Milankovitch) remains fairly consistent, the method of modulation MUST reside within the response of the climate system to this forcing (ie: feedbacks).

      It seems clear that CO2 cannot be the feedback mechanism, because when CO2 concentrations are high the world cools, and when CO2 is low the world warms. This counter-intuitive response is very difficult to explain via CO2 feedbacks. But what other feedbacks can give this strong response to Milankovitch polar warming, a response that can melt vast ice sheets in just 5 kyr?? And what feedback system is selective – only being present just before major deglaciations?

      The answer is surely dust. Dust is only present on the ice sheets after several Milankovitch maxima have come and gone, and is always present just before each interglacial. And the dust albedo feedback effect can be two orders of magnitude stronger than the CO2 feedback, when measured regionally. In fact, the albedo feedback is much greater than even this, when measured century by century, because it is permanently infused within the ice sheets and operates from century-one of the interglacial – rather than having to build up its feedback response slowly over several millennia, as CO2 must do.

      In short:
      Milankovitch must work alongside a feedback system.
      And that feedback is highly likely to be dust-ice-albedo.

      Regards,
      Ralph Ellis

    • Richard,

      Another advantage of dust feedbacks over CO2 feedbacks, is the magnitude of their response to temperature changes.

      CO2 has a linear relationship with palaeoclimatic temperature, so CO2 and temperature are obviously correlated but not necessarily causal. In great contrast, dust has an exponential relationship with both temperature and CO2. So the response of dust to changes in temperature (actually responding to changes in CO2) is far greater than the response of CO2 itself.

      So which feedback system is more likely to modulate ice volume and temperature: the small changes in CO2 cincentrations, or the exponential changes in dust flux…?

      Graph:
      Co2 (blue) vs Dust (green)
      The dust plot is inverted and logarithmic

  83. Pingback: Weekly Climate and Energy News Roundup #328 | Watts Up With That?

  84. Pingback: Weekly Climate and Energy News Roundup #328 |

  85. Pingback: Ice Age insights | Clive Best

  86. Pingback: Ice Age insights – Climate Collections

  87. nickreality65

    The point of this second experiment is to demonstrate that a surface with multiple outgoing heat transfer pathways cannot radiate as a BB. Just as reflected, transmitted, absorbed incoming radiation must equal 1.0 the outgoing radiative and non-radiative heat transfer processes must equal 1.0. Radiation does not function independently from the non-radiative processes.

    The immersion heater is feeding 1,180 W of power into the insulated pot of water which is boiling at an equilibrium temperature of 200 °F. (6,300 feet) The only significant pathway for energy out of this system is through the water’s surface.

    Any surface at 200 °F radiates at 1,021 W/m^2. This is 2.38% of the 42,800 W/m^2 power input to the system. That means 97.6% of the power input is carried away by non-radiative heat transfer processes, i.e. conduction, convection and evaporation.

    Likewise, the significant non-radiative heat transfer processes of the atmospheric molecules render the 396 W/m^2 LWIR radiation upwelling from the surface impossible.

    No 396 W/m^2 upwelling BB LWIR means there is
    No energy to power the 333 W/m^2 GHG out-of-nowhere perpetual energy loop,
    No energy for the CO2/GHGs to “trap” or absorb and re-radiate “warming” the atmosphere/surface,
    No RGHE or 33 C warmer and
    No man-caused climate change.

    This second experiment validates the findings of the modest experiment.

    Modest experiment:
    https://www.linkedin.com/feed/update/urn:li:activity:6394226874976919552

    Annotated TFK_bams09
    https://www.linkedin.com/feed/update/urn:li:activity:6447825132869218304

  88. Pingback: Journey Into The Underworld: 4 | Frank Davis