Guest Post by Javier
It is a well-known feature of climate change that since 1850 multiple climate datasets present a ~ 60-year oscillation. I recently wrote about it in the 7th chapter of my Nature Unbound series. This oscillation is present in the Atlantic Multidecadal Oscillation (AMO), Arctic Oscillation (AO), North Atlantic Oscillation (NAO), Pacific Decadal Oscillation (PDO), Length of Day (LOD), and Global (GST) and Northern Hemisphere (NHT) temperatures, with different lags (figure 1).
To me this oscillation is not a cycle because prior to 1850 it had a more variable period and it is not well identified in LIA records. Since the origin of this oscillation is unknown, models have a hard time reproducing it and it is all but ignored by the IPCC. It is a big oscillation with an amplitude of ± 0.3 °C in NHT (0.1-0.2°C in GST; figure 2). While the long-term temperature trend is unaffected by it, there is a large effect on the 30-year trends. If this oscillation is considered, most of the climate alarmism vaporizes.
This oscillation was first detected by Folland et al. (1984) in global SST and nighttime marine air temperature records, and later correlated to precipitation records in the Sahel (Folland et al., 1986). The multidecadal oscillation was isolated by Schlesinger and Ramankutty (1994) in the global mean instrumental temperature record, as a 65-70-year northern hemisphere periodicity, and attributed to internal variability of the coupled ocean-atmosphere system. It was termed the Atlantic Multidecadal Oscillation (AMO) by Kerr (2000). Scafetta has published several articles on it since 2010 (Mazzarella & Scafetta, 2012, for example). Among skeptics it has been featured prominently, for example here at WUWT:
In favor:
Scafetta on 60-year climate oscillations. Anthony Watts
New paper in GRL shows that a 60-year oscillation in the global tide gauge sea level record has been discovered. Anthony Watts
Claim: Solar, AMO, & PDO cycles combined reproduce the global climate of the past. Guest essay by H. Luedecke and C.O.Weiss
Models overestimate 60-year decadal trends. Guest essay by Clive Best
Against:
The Elusive ~ 60-year Sea Level Cycle. Guest Post by Willis Eschenbach
It can be reasonably postulated that the famous pause is nothing more than the manifestation of the recent end of the ascending phase of the 60-year oscillation.
On examination of figure 2 we observe two prominent peaks at 2016 and 1876, separated by 140 years and thus at a similar point in the AMO oscillation. Both also took place at the end of a solar cycle. Perhaps the 1876 and 2016 El Niño events can be considered analogs, but clearly the 1876 peak shows a bigger NHT deviation and a much stronger effect on AMO.
We may remember that as the time the Challenger Expedition took place. It was the first fully scientific oceanographic expedition and one of the most successful ones. Among its achievements we can count (Steven Varner):
– The birth of oceanography as an independent scientific field.
– The first systematic plot of currents and temperatures in the ocean.
– A map of bottom deposits that has not been changed much by more recent studies.
– An outline of the main contours of the ocean basins.
– The discovery of the mid-Atlantic Ridge (which baffled scientists at the time).
– The recording of the 26,900 feet (8,200 meters) Challenger Deep, a new record ocean depth, in the Mariana Trench.
– The discovery of 715 new genera and 4,717 new species of ocean life forms.
– The discovery of prodigious life forms even at great depths in the ocean (refuting earlier hypotheses of lifeless bottoms).
The expedition departed England in December 1872 and returned in May 1876.
Recently scientists from the Scripps Institution of Oceanography (US) and the National Oceanography Centre (UK) took the data from the Challenger expedition and compared it to the Argo data from the same locations 135 years later (Roemmich et al., 2012). The warming observed was consistent with current knowledge, but they found something very interesting:
“The 0.33 °C ± 0.14 average temperature difference from 0 to 700 m is twice the value observed globally in that depth range over the past 50 years, implying a centennial timescale for the present rate of global warming.”
In other words, the warming for the first half of the period (mostly natural) is about the same as for the second half (including the anthropogenic contribution). They conclude that the warming rate of the oceans has not accelerated with the addition of anthropogenic GHGs.
For at least 4 years (1872-1876), and during all the time the Challenger was at sea, the world was experiencing La Niña conditions. It is also probable that 1871 was a La Niña year, making it one of the longest La Niña periods in recorded history.
Most people have the idea that La Niña means cooling and El Niño means warming when it is just the opposite. When strong La Niña conditions dominate, the Pacific accumulates more and more thermal energy due to higher insolation produced by the reduction of clouds due to lower evaporation. The planet thus acquires more thermal energy in the Pacific Ocean subsurface. Then it suddenly exploded in 1876 producing the largest known El Niño in historic times. A monster El Niño right in the middle of the pre-industrial IPCC baseline period (1850-1900). This puts to shame the notion that pre-industrial climate was more congenial. It was a complete catastrophe. Terrible multi-year droughts took place in Brazil, India, China, European Russia and many other places, claiming the lives of an estimated 20-50 million people, or at the time ~ 3% of the world’s population. The world’s worst natural disaster ever (not counting pandemics). We can’t even imagine it. China lost 13 million people. In India the death toll is estimated at 5.5 million, with 58.5 million people distressed by hunger. This occurred while the British colonial government exported food and reduced relief help, due to criticisms of excessive expenditure, prompting modern accusations of a colonial genocide.
So that is the human meaning of the spike at the left of figure 2. The 1876-78 El Niño was so big that it spread over all the oceans, causing a corresponding spike in the AMO. Afterwards AMO and temperatures started going down and the world recovered. El Niño accomplished its mission of releasing the excess energy accumulated during the La Niña years.
Looking at AMO data we can see that it has another interesting decadal periodicity. It is so clear that it is visible in unsmoothed monthly data, but it is better seen with a 4.5-year moving average (figure 6).
The decadal periodicity is also present in hemispheric and global temperatures, and, in an article in 2009, Anthony Watts with Basil Copeland defended a lunisolar influence behind it:
Anthony and Basil used HadCRUT3 global data, but since AMO and temperature are so correlated (see figure 1) and AMO has less noise, I am going to stick to AMO.
The decadal periodicity in AMO has a frequency of 9.0-9.1 years (Manzi et al., 2012; figure 7)
Due to its periodicity, it has been suggested numerous times that the 9-year peak corresponds to a Lunar tidal frequency. The nodes where the orbit of the Moon crosses the Earth’s ecliptic are two points where a maximal alignment of the Moon-Earth-Sun takes place. They half rotate around the Earth every 9.3 years producing higher tides at that period when they reach optimal alignment. Also, the elliptical orbit of the Moon rotates around the Earth, placed at one of the foci, every 8.85 years. Higher tides also take place when the perigee-apogee axis is properly oriented towards the Sun. The closeness of these two periods suggests that a 9.1 period could result from their interaction.
Scafetta (2010) ingeniously demonstrated using the JPL ephemeris that the speed of the Earth around the Sun is perturbed by the presence of the Moon at a frequency of 9.1 years (figure 8).
Although this does not demonstrate that the 9-year periodicity in AMO is due to the Moon, it does build a case. The effect of the Moon’s gravitation on atmospheric tides and oceanic tides has enough energy to produce the observed effect. Half of the vertical mixing in the oceans is due to tides, and the other half to wind. In addition, tides affect oceanic currents by sloshing huge amounts of water from one place to another. The expected effect is that stronger tides should produce cooling by enhancing the upwelling and mixing of colder, deeper water. It is important to realize that the tidal forcing is thus inverted with respect to AMO temperature anomaly, and higher tidal forcing should produce temperature troughs (for example in figure 6), not peaks.
Some people have suggested that longer cycles could be the result of a modulation between lunar and solar cycles. For example Greg Goodman (climategrog) in a comment in 2014:
“If we do the same process with 9.08 and 10.4 it gives a modulation frequency of 143 years so the “beat” period of each bulge in amplitude is 71.5. So, it is possible for an interplay of lunar and solar forces to produce the kind of long cycles seen in the climate record.”
Prior to that, in 2011, Clive Best explored in an article in his blog the possibility that the 60-year oscillation was produced by the combined effect of both the solar variability and the tidal variability: A 60-year oscillation in Global Temperature data and possible explanations.
Alas, he couldn’t find convincing evidence:
“There is no single astronomical effect which can explain the 60-year time period. I have looked into the possibility that a superposition of both the 11-year solar variability and the 18.6 year lunar tide could produce the observed 60 year oscillation. There is no convincing evidence that this is the case.”
So, I decided to revisit the 60-year oscillation to see if it is possible that the modulation between the 9-year frequency in AMO and the 11-year solar cycle could be responsible for the emergence of the 60-year oscillation through constructive and destructive interference. In principle the period of the beat from a 9-year period (T(1)) and a 10.9-year period (T(2)) is too short. T(beat) = 1 / (1/T(1))–(1/T(2)) = 52 years
However, since the solar cycle is quite variable I decided to plot it anyway. The result is most interesting (figure 9).
The non-stationary correlation between the two cycles produces a periodicity that is compatible with the ~ 60-year periodicity in AMO. Periods of high correlation between the 9-year AMO and 11-year solar periodicities correspond to cold 60-year AMO periods, while periods of high anti-correlation correspond to warm 60-year AMO periods.
Mechanistically, times of high correlation between the 9-year AMO and 11-year solar periodicities correspond to times when the highest tidal forcing (AMO cooling) coincide with the times of lowest solar activity (solar minima), which could explain why the AMO displays cooling. Times of high anti-correlation between the 9-year AMO and 11-year solar periodicities correspond to times when the highest tidal forcing (AMO cooling) coincide with the times of highest solar activity (solar maxima), which could explain why the AMO does not display cooling.
The irregularity of the 11-year solar cycle period could explain why the ~ 60-year oscillation is also irregular, and the low level of solar activity during the LIA could also explain why the 60-year oscillation is not apparent or weaker at that time.
Regardless of the 60-year oscillation being due or not to the modulation of a lunar tidal 9-year cycle and a solar activity 11-year cycle, the observation of the interplay between these two cycles leads to two conservative predictions that do not rest on any hypothesis. As we are in a period of high anti-correlation and as Solar Cycle 25 increases its activity over the next 5-6 years the AMO should experience a decrease associated with its 9-year periodicity, putting additional downward pressure on surface temperatures.
The second prediction has been proposed multiple times: the downward phase of the ~ 60-year AMO oscillation should cause a reduction in global temperatures of ~ 0.1-0.2 °C over the next 20-30 years, all other things being equal.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
Chimp April 28, 2018 at 10:04 pm
You’re so cute when you’re angry …
“Cycles” do emerge in chaotic systems, Chimp. Then they morph, or change periods, or simply disappear. Which makes them something very different from the daily or lunar cycles you discuss. They are predictable. Day and night don’t disappear. The yearly variations don’t suddenly change to a period of three years. Lunar cycles don’t appear and then vanish.
So yes, pseudocycles do emerge. Here are some in the Figure 5 from your own link.
And as your Figure 5 clearly demonstrates, these pseudocycles can be of any period length, they can appear at any time, they can disappear at any time, they can last any amount from a few cycles to a few years to a few decades to a few centuries, and they are totally unpredictable … in other words, they are as chaotic as the substrate from which they emerge.
As a result, you can’t use them to predict the future, or understand the present, or hindcast the past. So … what use are they?
w.
Not angry, just sad that you are so blind to reality and impervious to scientific facts.
Sad. But then you’re old and at the end of your pointless existence, desperately trying to puff yourself up to fend off the reality of your having achieved nothing in your aimless existence.
You’ve got nothing. You’re just digging your hole deeper and making people who previously laughed at you just embarrassed for you now.
As I said. Pathetic.
@ur momisugly Chump,
” … But then you’re old and at the end of your pointless existence, desperately trying to puff yourself up to fend off the reality of your having achieved nothing in your aimless existence. …”
—-
lol … that’s pretty solid projecting. 😊
Javier April 29, 2018 at 4:10 am
Javier, Figure 5 is from Chimp’s link to a paper that Chimp said showed the AMO over the period of the Holocene. Which makes sense, because the title of the paper is:
So if you think that Figure 5 has no “relationship with the AMO oscillation”, you’ll have to take it up with Chimp and/or with the authors and peer-reviewers of the paper. I’m not the one making the claim. They are. Go talk to them.
Regards,
w.
I have noticed some dismissing the AMO, but the fact this is part of the global thermohaline circulation that significantly affects global climate should be taken into account. It has been responsible for most, if not all the warming and cooling since the 1940’s and earlier. This is a natural way the global oceans redistribute energy from shortwave radiation.
The AMO is not just a regional phenomenon as it is part of the main ocean conveyor. It is the area where proxy data has found it to have the biggest influence on world climate throughout history including the ice ages. It has shown to also have the largest influence on global temperatures over the recent decades despite the alarmist agenda.
The Atlantic meridional overturning circulation (AMOC) is a system of currents in the Atlantic Ocean, characterized by a northward flow of warm, salty water in the upper layers of the Atlantic, including the Gulf Stream, and a southward flow of colder, deep waters that are part of the thermohaline circulation. The AMOC is an important component of the Earth’s climate system.
This ocean current system transports a substantial amount of heat energy from the tropics and Southern Hemisphere toward the North Atlantic, where the heat is then transferred to the atmosphere. Changes in this ocean circulation could have a profound impact on many aspects of the global climate system.
There is growing evidence that fluctuations in Atlantic sea surface temperatures, hypothesized to be related to fluctuations in the AMOC, have played a prominent role in significant climate fluctuations around the globe on a variety of time scales.
Measurements across the North Atlantic suggest multidecadal swings in sea surface temperatures that may be at least in part due to fluctuations in the AMOC. The figure below describes this variation in North Atlantic sea surface temperatures for the period 1856 to 2009. The repetitive cycle obvious in this figure is known as the Atlantic Multidecadal Oscillation (AMO). Evidence from paleorecords suggests that there have also been large, decade-scale changes in the AMOC, particularly during glacial times. These abrupt changes have had a profound impact on climate, both locally in the Atlantic and in remote locations around the globe.
At its northern boundary, the AMOC interacts with the circulation of the Arctic Ocean.
https://en.wikipedia.org/wiki/Atlantic_meridional_overturning_circulation
If you are to dismiss the AMO, then you are also dismissing the AMOC.
Perhaps interesting and still on topic?
I just heard on the radio (here in South Africa) that the difference in the tides is currently at its highest. They told us to be aware of unusual high and low tides.
Now, if my / our (?) theory is correct there will be more upwelling of cold waters during this time. It works like an extra stirrer effect. This means that we can expect a cold winter here in the SH…… and possibly more sunny (dry) weather in the NH……
Bartemis April 29, 2018 at 7:43 pm
Thanks, Bartemis. Let me see if I can explain the problem by example. Suppose that we have the HadCRUT annual temperature data up to 1991. That’s 85% of the full dataset.
We analyze it, and we find that there is a strong 60-year cycle in it, just as Javier said in the head post. Here’s the CEEMD analysis of that dataset:
A solid 60-year cycle, no question. So Javier appears to be correct.
Next, we take the dataset to 1991, and we determine the best-fit 60-year cycle for the period and overlay it on the data. Here’s that graph:
And indeed, that’s a good fit. Everything is going swimmingly, cycles rule.
Now, you say above that:
So let’s do that very thing … here is the HadCRUT dataset including the last 15% of the data, a small look outside the in-sample data used to determine the underlying pseudocycle, along with the projection of what you call the “most likely path” to cover that last 15% …
I’m sure you can see the problem … despite your assurances, we cannot “project the most likely path going forward”. I’ve tried this same thing on a number of climate datasets (as well as on stock market datasets), analyzing the first part and projecting the last part, and I can assure you that what works well in-sample often totally falls apart out-of-sample.
So your claim simply doesn’t hold up, another lovely theory gone aground on a reef of hard facts … and this is why I call them “pseudocycles”. True cycles allow us to project the future. We can say with great accuracy what the tides will be in ten years because the motions of the moon and earth are true cycles.
But we can’t do the same thing with Javier’s “60-year oscillations”, because they are only what I call pseudocycles.
You are correct that in your work you can do such extensions, for a simple reason. The springs and shock absorbers or the airplane frame members and such have constant unchanging underlying resonant frequencies, and what you see is some combination of those true cycles. So you can analyze and extend them.
But the climate is different. It doesn’t have such constant unchanging underlying resonant frequencies. And this is why it is so difficult to predict, and why you can’t simply extend the pseudocycles into the future.
Sadly, you close by saying …
Dang … and you were doing so well sticking to the science up to that point. But then you had to descend into ad hominems. Now in the past I’d likely have slapped you up ‘longside the head for your ugliness, but this is the new me. Kanye says to love everyone, Anthony Watts says be cool, and I’m doing my best to take their advice.
So let me just say that such unpleasant personal attacks do nothing to my reputation, but they don’t do yours any good … I’m not untutored, I’m self-tutored, but that doesn’t mean I’m wrong. Here’s a list from Google Scholar of over a hundred citations to my work in scientific journals.
Not bad for a self-tutored man with only two college science courses to my name, Intro to Physics and Intro to Chemistry …
Best regards,
w.
The notion that cycles need to be strictly periodic (line spectra) to be “real” is severely misguided scientifically. This is especially true in the geophysical context, where random cycles of continuously varying amplitude and phase (power density spectra) predominate. Such is the nature not only of ubiquitous random ocean waves, which are never strictly periodic, but of temperature signals as well.
There are well-established methods of exploiting narrow-band random cycles for useful prediction over limited horizons, (Wiener filters, Kalman filters) that do NOT rely upon the simplistic fitting of sinusoids to empirical data, as is often assumed in “climate science.” Sadly, Willis’ attempt to refute such admittedly-limited predictability suffers further from choosing the highly manufactured, non-stationary HADCRUT series as a putatively representative example. In reality, it merely shows lack of basic proficiency in geophysical signal analysis.
1sky1
Your comment was timely. It reminded me of something.
If you are doing an analysis of an induction motor at load the speed of the rotor will vary as does the slip of the motor. The FFT results would clearly indicate the rotational frequency of the rotor but upon examination of the signal it might be a little broad because the speed was varying. Early desktop analyzers did not have this capability but later ones implemented order tracking. I believe they worked by altering the timing signal as needed. Now when you examined the rotational signal it was very discrete and a greater S/N. An added benefit was that any signal that was rotationally ordered also had an improved signal.
Whether that could be employed on these datasets I do not know. However, the fact that the rotational frequency did very was no bother; We knew what the rotational was and where to find it. That begs the answer to this question. If the frequency is varying slightly what does that change or does it invalidate you analysis?.
Charles May April 30, 2018 at 4:13 pm
The problem is not that the frequency is “varying slightly”. It is that the signal disappears entirely, sometimes for thousands of years, which never occurs in your example.
w.
Willis
we are talking about the oceans, e.g.
http://www.woodfortrees.org/plot/esrl-amo/from:1850/to:2019
clearly you can see that what goes up must [also] come down?
Hardcrut has multiple problems, not least because of changes in calibration and recording techniques since about the 70’s and as I explained before, the set is not balanced NH/SH
Henry, that may indeed what YOU are talking about. But the head post is talking about, inter alia, the “Global (GST) and Northern Hemisphere (NHT) temperatures”.
Second, whether or not HadCRUT has been “adjusted” or not makes no difference to my analysis.
Third, as I replied to your earlier claim, HadCRUT first averages the NH and the SH separately, and then averages the two averages, for exactly the reasons you point out. This makes it “balanced NH/SH”.
w.
Willis.
Thx.
On that last issue of the balancing act. Do you have some proof of that?
Anyway. My set of 27 stations each hemisphere is balanced to zero latitude which is still different to what you are saying.
Hence. I get something totally different….
Henryp April 30, 2018 at 2:06 pm Edit
Henry, of course I have proof of that. Unlike many commenters here, I do my utmost to only make assertions for which I have evidence.
Regards,
w.
The number of stations used in each hemisphere are far from balanced especially before the 1960’s. There are many more used in the northern hemisphere than the southern hemisphere, so despite (NH+SH)/2 being correct it’s weighted very poorly.
https://crudata.uea.ac.uk/cru/data/temperature/crutem4/station-data.htm
To detrend the data you have to remove all the trend and in the above post the trend in the graph is still not completely removed.
I had to weight the detrend by 0.9 to remove all the trend and this graph below matches that of the AMO even over recent years.
http://www.woodfortrees.org/plot/hadcrut4gl/detrend:0.9
Compared with the AMO they match each other cycles.
http://www.woodfortrees.org/plot/hadcrut4gl/detrend:0.9/plot/esrl-amo/offset:-0.5
It’s actually more like 0.8 to remove all the trend shown below.
http://www.woodfortrees.org/plot/hadcrut4gl/detrend:0.8
Despite the slight change the result is the same as the previous conclusion.
http://www.woodfortrees.org/plot/hadcrut4gl/detrend:0.8/plot/esrl-amo/offset:-0.5
1sky1 April 30, 2018 at 3:13 pm
This issue is not that the claimed 60-year cycle is not periodic. The issue, as Figure 5 clearly shows, is that it disappears for thousands of years at a time.
I called it a “pseudocycle” to distinguish it from cycles like the tidal cycles, which are not line spectra but are predictable years in advance.
w.
Not entirely so! All that Figure 5 shows is that the power content of proxy data in the vicinity of 60-year periods drops below the (unspecified) color graphic threshold. Band-limited processes cannot be time limited.
BTW, tidal cycles are the classic geophysical example of processes specified by line spectra. Nearly a century ago Doodson identified over a hundred tidal constituents, i.e., discrete spectral lines representing pure sinusoids of finite amplitude, It’s only the fact that these lines do not occur in harmonic sequence, but at incommensurable frequencies, that prevents composite tidal records from manifesting strict periodicity over any finite time-interval.
Charles/ Matt
The point Willis is making is that things are not so straight forward with the weather as in normal physics….-. Which is true. For example, not even the primary solar cycle (Schwabe) has a constant length and the Hale cycles can differ between 20 and 23 years.On the Gb cycle it can happen that you miss a whole beat/ quadrant [probably related to the DeVries cycle]. In that case you actually can get an extended maximum or extended minimum leading to a much warmer or cooler period, respectively.
Namely, these things do have an effect on the amount of energy going into the oceans (the amount UV and IR is important here – you cannot heat water with a torch light?)
So, indeed, I agree with him that what looks like a sinusoidal relationship can sometimes summarily brake off and go into a different direction. All that is again related to other cycles which in turn run on much longer time scales.
I must say here again that I think Javier has done some outstanding work on pointing out to us those longer [solar] cycles.
The reason for the appearances of these longer term solar cycles are still a bit murky but I have found correlation between the position of the planets and the relevant SC.
Obviously, correlation does not yet prove causation.
In the meantime things stay the same, more or less, as it says in the bible: who knows which way the wind will blow?
By my results, we did make the turning point on the Gb cycle (2014) without any problems – no extended minimum expected – and we have started on the sinusoidal moving up again.
Pity nobody ever noticed that cooling has already set in. The problem here is that our pension funds and other markets will crash if they learn that AGW is just a hoax and a scam.
Hence, everybody is feeding on the scam and, if necessary,
adding ‘corrections’ to the data…..
We are talking about a few tenths of a degree K, globally, and who would even notice this, if the differences just between the rooms of house can be a few degrees K?
BW
Henry
henryp
You touched upon a subject…
“The reason for the appearances of these longer term solar cycles are still a bit murky but I have found correlation between the position of the planets and the relevant SC.
Obviously, correlation does not yet prove causation.”
…that I wrote a comment days ago that most ignored. You can find it if you look for it. Basically I was pointing out that when you do not have answers for something, look somewhere else.
http://spaceweather.com/glossary/imf.html
http://planetfacts.org/the-solar-system/
http://solarviews.com/eng/jupiter.htm
https://www.universetoday.com/34076/planetary-alignment/
https://www.fourmilab.ch/cgi-bin/Solar
The Sun and all our planets and their moons in orbit, all affect our climate by their positions in relation to Earth. If you have Jupiter and Saturn on the celestial opposite of the Earth from the Sun and, Mercury and Venus between the Earth and the Sun it changes our orbit and distance from the Sun. Each having their own Magnetic Fields influences to our Solar System. You can use the last link I provided and type in the dates of each Geomagnetic Reversal of the Sun to see the positions of the planets. I just do not have the time to put this theory together. Yet I have toyed around with it and found a few correlations to cycles we’ve had.
Mod
I am missing one of my comments. Do you still have it?
Thx. It is up now. Warning again for me to always keep a copy of a comment.
I think you can predict the AMO in future and in my view it won’t start the negative phase until close to 2035. Therefore we should have at least more than 15 years of the positive phase left and after this global temperatures will decline with it. Over the next 15 years or so the alarmist crowd will be making a big deal by claiming global temperatures beating records by barely 0.1 c with the next strong El Nino. This will end with the incoming negative phase of the AMO with a turn around trend of Arctic sea ice with it.
The negative phase will last around 30 years until 2065 then the next positive phase will finish around the beginning of the next century close to 2105. The reason for the slight variation in the 60 year cycle is because it really is in two parts, a 40 year warm phase than a 30 year cool phase repeating itself.
Regarding the D/O cycles in glacial intervals and Bond cycles in interglacials these have been found to complete the change to a significant new climate in less than 70 years. This period is very close to the warm and cool phases of the AMO and don’t think they are a coincidence.