Delayed Oscillator model suggests NAO forces AMOC/AMO on a 60 year cycle

Posted: January 6, 2015 by tallbloke in Analysis, Cycles, Ocean dynamics

Glimmers of understanding are percolating through into mainstream climate science, this time through the journal Climate Dynamics. I can’t remember if Marcia Wyatt and Judy Curry explicitly linked these oscillations in their stadium wave paper, but it’s more evidence that our cycles driven theory of climate is correct, and that the 1976-2005 warming was mostly a natural phenomenon. It is likely to be followed by a 2006-2035 cooling phase, possibly accentuated by the lowest solar activity levels in two centuries or more. Unfortunately, the luni-solar dimension to the multidecadal variability is not explored. Nonetheless, this paper represents some joined up thinking in terms of the cyclic chain of cause and effect which connects the northern hemisphere oceanic oscillations.

nao-amocfigs10-11

A delayed oscillator model for the quasi-periodic multidecadal variability of the NAO
Cheng Sun, Jianping Li, Fei-Fei Jin Date: 06 Jan 2015
Abstract
Wavelet analysis of the annual North Atlantic Oscillation (NAO) index back to 1659 reveals a significant frequency band at about 60 years. Recent NAO decadal variations, including the increasing trend during 1960–1990 and decreasing trend since the mid-1990s, can be well explained by the approximate 60-year cycle.

This quasi 60-year oscillation of the NAO is realistically reproduced in a long-term control simulation with version 4 of the Community Climate System Model, and the possible mechanisms are further investigated. The positive NAO forces the strengthening of the Atlantic meridional overturning circulation (AMOC) and induces a basin-wide uniform sea surface temperature (SST) warming that corresponds to the Atlantic multidecadal oscillation (AMO). The SST field exhibits a delayed response to the preceding enhanced AMOC, and shows a pattern similar to the North Atlantic tripole (NAT), with SST warming in the northern North Atlantic and cooling in the southern part. This SST pattern (negative NAT phase) may lead to an atmospheric response that resembles the negative NAO phase, and subsequently the oscillation proceeds, but in the opposite sense. Based on these mechanisms, a simple delayed oscillator model is established to explain the quasi-periodic multidecadal variability of the NAO. The magnitude of the NAO forcing of the AMOC/AMO and the time delay of the AMOC/AMO feedback are two key parameters of the delayed oscillator. For a given set of parameters, the quasi 60-year cycle of the NAO can be well predicted. This delayed oscillator model is useful for understanding of the oscillatory mechanism of the NAO, which has significant potential for decadal predictions as well as the interpretation of proxy data records.

From the discussion:

There is some consensus
in observational studies that the NAO leads to the
NAT nearly instantaneously through surface heat flux
changes (Wu et al. 2009; Deser et al. 2010) and that the
AMO follows the NAO with a 15–20 year delay through
AMOC changes (Latif et al. 2006a, b; Latif and Keenlyside
2011; Li et al. 2013). These two mechanisms may provide
an alternative interpretation for the oscillatory sequence
and could partly explain the lead–lag relationships found
in the present study. However, these two mechanisms cannot
explain the quasi-60 year period of the NAO without
a feedback from the ocean onto the NAO. Previous studies
have suggested that the oceanic feedback is difficult to
detect in the observations and model dependent. For example,
Scaife et al. (2009) have shown in a multi-model study
that the late twentieth century NAO trend cannot be reproduced
in models forced by the observed SSTs. This indicates
either that the SSTs may be a weak factor in the NAO
variability or that these models may show much too weak
impact of SSTs on the NAO. The present analysis with the
CCSM4 model provides some modeling evidence for the
NAT feedback on the NAO, which is critical for reproducing
the NAO multidecadal variability. Since the performance
of a numerical climate model is always limited, a
limitation of our study is that the results are obtained only
by one single model. Thus, a multi-model comparison of
the NAO multidecadal variability for a better understanding
of how robust the mechanisms are among different climate
models warrants further investigation.

Previous studies have suggested several potential
sources of predictability for the NAO. For instance, one
potential source of predictability comes from the downward
control exerted by the stratospheric polar vortex on
the tropospheric NAO (Baldwin et al. 2003). It has also
been shown that the autumnal snow extent over Eurasia
significantly influences the winter NAO, with an above
(below) normal October snow extent leading to a negative
(positive) phase of the winter NAO (Cohen 2003). These
sources of predictability mainly affect the NAO variability
at timescales ranging from intraseasonal to interannual. The
present study has implications for predicting the NAO over
interdecadal timescales. Assuming that the approximate
60-year cycle in the NAO continues, we should be entering
a negative phase of the cycle, as suggested by Fig. 1a.
This may correspond to more negative NAO events during
the next decade. Considering that the negative NAO usually
leads to cooling over most of the NH continent (Hurrell
1995, 1996), the anthropogenic warming effect may be partially
offset. A limitation in this study is that we do not take
into account the impact of “weather noise” in the delayed
oscillator model. Previous studies have suggested that such
weather noise is needed to excite and energize the North
Atlantic multidecadal variability (Delworth and Greatbatch
2000). In fact, within a linear framework, the noise simply
deflects the solutions, and in this sense, the solutions
presented here can be regarded as the ensemble mean of a
large ensemble driven by noise. The ensemble mean is the
predictable part of the multidecadal variability, and thus it
is of special interest here.

Comments
  1. oldbrew says:

    ‘ the anthropogenic warming effect may be partially offset’

    Where are the actual measurements of this alleged effect?

  2. tallbloke says:

    Boilerplate kowtow to the gatekeepers of AGW required to get the paper past the editors and through peer review. Ignore it.

  3. oldbrew says:

    TB: agreed it’s just a routine cliche, but let’s call them out on it anyway. Where’s the logic?

    If they can’t quantify attribution, they are just arm-waving and we should keep telling everyone that.

    Btw if they say ‘partly offset’ they assume there will still be some increase of warming, which hasn’t been seen for a decade or more now. So if temperatures drop they will have got that wrong.

  4. sirgeogy says:

    So the NAO might actually be a driver of North Atlantic weather and not simply a casual observer? rgs.to/weather

  5. The correlation of the AMO with NH temperatures is perfectly clear.

    http://notalotofpeopleknowthat.wordpress.com/2014/08/22/correlation-of-the-amo-with-nh-temperatures/

    The key to the AMO is not so much much whether it is “warm or cold”, but the direction of travel. AMO hit rock bottom around 1976, and as it started to rise afterwards, so did temperatures.

  6. Edim says:

    Paul, exactly! Furthermore, the correlation of the AMO exists with SH temperature too (and other indices). AMO is a misnomer – it’s a global oscillation.
    http://www.woodfortrees.org/plot/hadcrut4sh/from:1850/mean:12/detrend:0.8/plot/esrl-amo/from:1850/mean:12/plot/hadcrut4sh/from:1850/mean:12/trend/detrend:0.735

    The secular trend is just another ‘oscillation’.

  7. tallbloke says:

    OB: Agreed. Call em out.

    Paul: Yes, and when you combine detrended AMO and SOI with integrated solar, you get all the history of C20th climate change near enough. Not much room left for CO2 and ‘adjustments’. 0.15C over 140yrs maybe.

  8. p.g.sharrow says:

    WOW! does that mean the climate mavens will accept that WEATHER follows a 60 year cycle?

    Climate change can only be see over 5 cycles, 300years. A 15 or 30 year hockystick is the work of a hockypuck! pg

  9. Jaime says:

    Finding it a little difficult to accept the notion of a quasi-periodic atmospheric weather pattern/mode (NAO) actually forcing a strengthening of the thermo-haline circulation which thereby induces a positive or negative mode of the truly oscillatory AMO. Even accepting that there is an observational 15 year delay between NAO and AMOC going back over the centuries, it is still a stretch to attribute NAO as the forcing agent in my opinion.

    NAO is a NH atmospheric mode which exhibits variability over both the short term (days/weeks) and the long term (multi-decadally). Undoubtedly, it must in some manner be affected by and, in turn affect other major oceanic/climate cycles/modes of variability, most particularly the AMO and the AO. But I cannot see it as a ‘forcing’ agent in its own right. Indeed, I would hesitate to describe any natural oscillatory or quasi-oscillatory cycle as a forcing agent. AMOC is ‘hardwired’ into the global geophysical environment and is, in a sense, self-sustaining, but must still be moderated to some degree by external forcings. Natural climate forcings are external and independent of the ocean/atmosphere circulation system – solar/lunar/planetary/interstellar. Even volcanic activity/aerosol forcing cannot be considered strictly independent in that sense.

  10. oldbrew says:

    Paul Homewood says: ‘AMO hit rock bottom around 1976, and as it started to rise afterwards, so did temperatures.’

    But the whys and wherefores are still out there somewhere😉

  11. vukcevic says:

    North Atlantic SST and N. Hemisphere temperatures followed the same trend until about 1970, when the two trends departed with the NH T acquired a steeper rising trend. However, the AMO (de-trended NA SST) and the de-trended NH T maintain strong correlation across time span of the data amiability.
    Both (AMO and NH T) 60 year quasi-cycles appear to be followers (with some delay) of another natural cycle (result of solar and earth magnetic fields interactions)

    NAO-AMO correlation is often described as non-stationary (i.e. in and out). NAO is a much older index, but its value is out of balance. Its two natural components, the southern (Bermuda, Azores or Lisbon) is given equal weight to the much stronger and more influential Northern (Stykkisholmur/Reykjavik). Taking this into account (re balancing of component weighting) a more meaningful and potentially correct NAO-AMO relationship can be established.

  12. Jaime says:

    The climate is in some ways analogous to a simple vibrating circular membrane (drum head) which has many distinct resonant modes of vibration, all of which can be excited by a single drum stroke.

    http://www.acs.psu.edu/drussell/Demos/MembraneCircle/Circle.html

    A small variation in the placement and the force of the stroke (climate forcing) elicits a different vibrational response from the head – the ‘noise’ generated by off-target drumming sometimes obscures the underlying resonant frequencies (climate cycles) but they are there still, hardwired into the design of the drum head. Simplistic analogy I know, but one which I find useful as a mental tool.

  13. okulaer says:

    Edim says, January 6, 2015 at 4:57 pm:

    “Paul, exactly! Furthermore, the correlation of the AMO exists with SH temperature too (and other indices). AMO is a misnomer – it’s a global oscillation.”

    That’s because it’s all controlled from the tropical Pacific, not from the northern Atlantic. The AMO is basically a teleconnected outward expression of a global, Pacific-driven variability in the ocean/atmosphere regime.

    If one is searching for the origin of some global climate effect, look where the dynamic processes and the amounts of energy handled at any one time are (by far) the biggest. Simple logic.

    What in turn controls the Pacific regime variability is clearly a different question. But one thing is for certain: It is not some other (and necessarily much smaller) ocean basin doing it.

  14. The paper is interesting.
    Just it does not reference any of my papers on related topics such as

    Scafetta N., 2014. Multi-scale dynamical analysis (MSDA) of sea level records versus PDO, AMO, and NAO indexes. Climate Dynamics 43(1-2), 175-192.

    Mazzarella A. and N. Scafetta, 2012. Evidences for a quasi 60-year North Atlantic Oscillation since 1700 and its meaning for global climate change. Theoretical and Applied Climatology 107, 599-609.

    In both cases the quasi 60-year oscillation of NAO is extensively discussed as well as the phase shift with the other indexes.

  15. ren says:

    Our previous study showed that the response of tropospheric pressure to variations of solar activity (SA) and
    galactic cosmic ray (GCR) fluxes reveals a regional structure determined by the positions of the main
    climatic atmospheric fronts, as well as it strongly depends on the epochs of the large-scale circulation
    [Veretenenko and Ogurtsov, 2012]. In the epochs of increasing intensity of the meridional circulation (the
    form C according to Vangengeim-Girs classification [Vangengeim, 1952; Girs, 1974]) an increase of GCR
    fluxes at minima of the 11-year cycle is accompanied by an intensification both of extratropical cyclones at
    Polar fronts of middle latitudes and Arctic anticyclones at high latitudes of the Northern hemisphere, as well
    as by a weakening of the equatorial trough at low latitudes. In the epochs of decreasing meridional
    circulation the troposphere response to SA/GCR variations reveals a similar regional structure, i.e., the
    regions of most pronounced effects are closely related to the climatic atmospheric fronts, but the sign of
    SA/GCR effects in these regions is opposite.
    It was also detected [Veretenenko and Ogurtsov, 2012] that the evolution of the meridional circulation
    is characterized by a roughly 60-year periodicity which, in turn, influences the sign of SA/GCR effects on
    troposphere pressure. Indeed, the reversals of the correlations between sea-level pressure at high latitudes
    and sunspot numbers occurred in the 1890s, the early 1920s, 1950s and the early 1980s and coincided well
    with the changes in the evolution of the C-type meridional circulation. Hence, the aim of this work is to
    study what processes may influence the evolution of the large-scale circulation and, then, the character of
    SA/GCR effects on troposphere pressure.
    http://geo.phys.spbu.ru/materials_of_a_conference_2012/STP2012/Veretenenko_%20et_all_Geocosmos2012proceedings.pdf

  16. Paul Vaughan says:

    A signal of persistent Atlantic multidecadal variability in Arctic sea ice
    http://folk.uib.no/ngftf/CV/PDF_Furevik/miles_et_al_grl_2014.pdf
    ftp://starfish.mar.dfo-mpo.gc.ca/pub/ocean/ingrid/Gary/Miles_grl51241.pdf (alternate link to same pdf)

    “Here we integrate and synthesize a set of multicentury historical records of Atlantic Arctic sea ice, supplemented with high-resolution paleoproxy records […] We establish a signal of pervasive and persistent multidecadal (~60–90 year) fluctuations […]”

  17. Unfortunately, the luni-solar dimension to the multidecadal variability is not explored. Nonetheless, this paper represents some joined up thinking in terms of the cyclic chain of cause and effect which connects the northern hemisphere oceanic oscillations

    That is to be expected from mainstream which they both are part of.

  18. I thought it might apply twice to both articles.

  19. But I am alright with it appearing one time. thanks