Nature Unbound VII – Climate change mechanisms

by Javier

Climate variations that alter the angular momentum of the atmosphere modify the speed of the Earth’s rotation, which affects the length of day (LOD). Alterations in LOD integrate different climate-affecting phenomena, and can anticipate turning points in climate.

Several pathways and feedback mechanisms have been identified that carry and amplify the signal produced by solar cycles. The stratospheric top-down mechanism relies on the quasi-biennial oscillation to affect the northern hemisphere winter climate. A bottom-up mechanism acts mainly in the Pacific Ocean, through changes in SST and cloud cover.

The existence of an ice-ocean-atmosphere ~ 60-year oscillation of unknown origin, might explain a significant part of the observed climate variability.


The Holocene prior to 1950 underwent episodes of profound and abrupt climate change that significantly altered the vegetation of entire ecozones at times of very modest GHG changes. This Holocene variability was reviewed earlier (Part A, and Part B). Thus, besides GHGs and volcanic eruptions, other poorly identified climate change mechanisms must be at work. Proxy analysis has allowed the identification of several climate cycles, some of solar origin (Bray, Eddy, and de Vries cycles), and one of non-solar origin (1500-year cycle). The mechanisms by which they act on climate are poorly understood, and additional mechanisms producing non-cyclical climate changes must also exist.

Several climate change mechanisms have been thoroughly reviewed elsewhere and will not be reviewed here. The effect of GHG concentration changes in the atmosphere is postulated as one of the main mechanisms for climate change, since these gases absorb and emit in the IR band (Hansen et al., 2007). Despite intense efforts, the magnitude of this mechanism’s contribution to climate change is still uncertain. The effect of volcanic eruptions on global climate is due to the absorption and scattering of radiation by sulfate aerosols produced by the vast amounts of sulfur dioxide injected into the stratosphere during strong volcanic eruptions. While they have a high latitude surface warming effect in the Northern Hemisphere winter by warming the stratosphere and disrupting its circulation (Robock and Mao, 1992), they have a stronger surface cooling effect in the lower latitudes and tropics during spring and summer, that can last for a few years (Robock, 2000). Surface cooling is because the sulfate aerosols are more efficient at scattering incoming solar radiation than absorbing Earth’s surface radiation. The aerosols participate in the destruction of ozone through their interaction with anthropogenic chlorine, and in the reduction of atmospheric water vapor (Soden et al., 2002). There is also speculation that volcanic eruptions might trigger or imitate El Niño conditions through their effect on trade winds, partially ameliorating the cooling effect.

Length-of-day and climate

Our capability to measure very slight fluctuations in the speed of rotation of the Earth, that result in micro-second changes in the length-of-day (LOD), has produced some interesting evidence on how climate-related phenomena affect the rotation of the planet. The rotation of the Earth is being constantly slowed by the friction of the tides with the ocean bottom, but once this secular increase in LOD is accounted for, excessive variations of the order of 1 millisecond in LOD are taking place in a scale from hours to decades. The excessive LOD (∆LOD) has been decreasing during the last forty years (figure 90 a), which is the opposite trend to the secular deceleration of the Earth. This multi-decadal decrease in ∆LOD (rotation increase) is attributed to tiny differential rotation in the Earth’s core associated with the geomagnetic field. Considering the Earth a closed system, the conservation of the angular momentum requires that any change in the angular momentum of the solid Earth is matched by an opposite change in the angular momentum of the fluid Earth.

The evidence shows that atmospheric angular momentum (AAM) changes are responsible for LOD changes at certain time scales, but not others. Two mechanisms are proposed to explain the imposed torques that transfer the angular momentum between the atmosphere and the solid Earth. One is due to surface wind tangential stresses across the surface, causing friction torques. The other one is due to mountain torques caused by surface pressure variability near areas of high topography.

Figure 90. Variation of LOD since 1981. a) From top, excessive LOD time data with its decadal trend (T9), components of period between 500 and 2000 days (T6-T8), between 100 and 500 days (annual and semiannual components, T4-T5), and less than 100 days (T1-T3). Four lower curves shifted for convenience, but all in same vertical scale. b) Fourier amplitude spectrum of the excessive LOD time series between 1981 and 2012. The dotted line is the 95% significance level. Source: S.-H. Na et al., 2013. J. Astron. Space Sci. 30, 1, 33-41.

The variations in LOD can be subdivided according to their periodicity (figure 90; Na et al., 2013). The annual (T5) and semiannual (T4) seasonal components match corresponding wind signals in the troposphere and stratosphere that peak 6 months out of phase and are of stronger amplitude due to Northern Hemisphere jets during the boreal winter, while the Southern Hemisphere winter peak is of lower amplitude. The long period band (500-2000 days; T6-T8, figure 90) corresponds in its biannual component to the quasi-biennial oscillation, resulting from the reversal of the zonal winds in the tropical stratosphere, while the 3-4-year component matches the ENSO signal (figure 91; Haas & Scherneck, 2004). During periods of El Niño, the tropospheric zonal winds have westerly anomalies. At the peak of the westerly anomaly period, the globally integrated AAM is notably strong, driving a slowing of the Earth’s rotation. During the 2015-16 winter season, El Niño produced a LOD excursion reaching 0.81 ms in January 2016.

Figure 91. Excess in length of day and El Niño Southern Oscillation. Blue curve, left scale: ∆LOD in milliseconds, long-term detrended (minus T9) and seasonal removed (minus T4-T5). Data from IVS, International VLBI (Very Long Baseline Interferometry) Service for Geodesy & Astrometry, treated as described in J.M. Gipson & C. Ma, 1999. IERS Technical Note 26. Red curve, right scale: Multivariate ENSO index (MEI). There is a clear correlation between the two time series and both show the ENSO events during the 23 years period. Source: R. Haas & H.-G. Scherneck, 2004. IVS 2003 Annual Report.

In 1976, Lambeck and Cazenave reported on the similarity between the trends of numerous climate indices for the past two centuries and changes in ∆LOD, in particular surface temperature and pressure, were related to wind strength. They concluded that periods of increasing zonal winds correlate with an acceleration of the Earth while periods of decreasing zonal circulation correlate with a deceleration of the Earth. They found a lag of 5-10 years in the climatic indices. Their result has been reproduced multiple times, and an example is shown with SST and ∆LOD (figure 92; Mazzarella, 2013). Lambeck and Cazenave considered the wind changes to be too small to account for the change in ∆LOD, and pointed instead that both might share a common origin. They ended with an interesting prediction, as the article was written after several decades of decreasing temperatures:

“if the hypothesis is accepted then the continuing deceleration of [speed of rotation] for the last 10 yr suggests that the present period of decreasing average global temperature will continue for at least another 5-10 yr. Perhaps a slight comfort in this gloomy trend is that in 1972 the LOD showed a sharp positive acceleration that has persisted until the present…”

As they suggested, 4 years after the 1972 sharp decrease in LOD took place, current global warming started, coinciding with the publication of the article. So, the gloomy cooling trend was substituted by what to some is an even gloomier warming trend. Except for the short interval 1987-94, ∆LOD decreased from 1972 to 2004, and has been slightly increasing since.

Figure 92. Earth rotation and sea surface temperature anticorrelation. Continuous line, detrended yearly values of ∆LOD with a 5-year running mean smoothing, shifted ahead 4 years. Dotted line, detrended yearly values of Northern Hemisphere SST, from HadSST3 with a 5-year running mean smoothing. Source: A. Mazzarella, 2013. Nat. Sci. 5, 149-155.

The close correlation between SST and the AAM (and LOD) has been known for a long time. The correlation is explained as due to ocean-atmospheric coupling where upwelling and downwelling depend on wind strength, and atmospheric pressure correlates with SST. Salstein (2015), one of the foremost experts in AAM, explains that the atmosphere has been simulated by a large number of models that are driven solely by the temperature of the underlying ocean surface. Based on these models, AAM has been calculated since the late 19th century from available SST data, and checked against LOD estimations based on lunar occultation measurements.

The nature of the postulated common cause that affects LOD and climate remains obscure. As the decadal trend in LOD is attributed to very small changes in the geomagnetic interaction between core and mantle, either internal solid Earth changes or external magnetic influences could be responsible for the effect on climate that accompanies decadal changes in LOD. Given the importance of stratospheric winds in AAM changes and the disproportionate effect of solar variability on the stratosphere, the Sun is a possible candidate to cause changes in the AAM, thus affecting LOD. Le Mouël et al. (2010) reported a close correlation between the amplitude of the semi-annual variation in LOD (∆LODsa) and the 11-year solar activity cycle. The sunspot number, a proxy for solar activity, led by one year and explained ~ 30% of the amplitude of ∆LODsa (figure 93). A correlation without lag was found between the amplitude of ∆LODsa and galactic cosmic rays. ∆LODsa is due mainly to the 6-month out of phase variation in zonal wind intensity caused by the difference in insolation between hemispheres. It is therefore linked to a fundamental feature of the climate system: the latitudinal distribution and transport of energy and momentum, as the Earth equilibrates the net radiative flux distribution balance between the equator and the poles, establishing the equator-to-poles temperature gradient (figure 66).

Figure 93. Modulation of the semi‐annual LOD variation by the solar Schwabe cycle. In blue, long term variation in the amplitude of the semiannual oscillation in LOD. The amplitude of the Fourier coefficient of the 6‐month spectral line of LOD is computed in a 4‐yr centered sliding window. In red, the sunspot record, inverted and offset to the right by one year. Source: J.-L. Le Mouël et al. 2010. Geophys. Res. Lett. 37, L15307.

Solar signal pathways

The Earth’s upper atmosphere and magnetic field form a coupled system with the Sun and geospace (the space inside the Earth’s magnetic field), that is connected by the solar wind. This coupled Sun-Earth system is responsible for the maintenance of life-compatible conditions on the Earth for thousands of millions of years, since without it, the oxygen would have been stripped from the atmosphere by the solar wind. The Sun is the source of nearly all the energy that drives the climate on Earth, but the Sun’s light is only a part of that energy, the other part corresponds to solar particles and fields. The flow of mass, momentum, and energy from the Sun’s interior through the interplanetary medium into the geospace environment is represented in figure 94 (Baker, 2000). Some of the effects of this flow within the coupled system reveal the effect of solar variability on the atmosphere. I have already discussed the effect of the 11-year Schwabe solar cycle on ozone formation and geopotential height (figure 65), that globally affects the Hadley circulation, causing an extension or contraction of the tropics. Let’s briefly review some of the phenomena that are also likely to be involved in the solar variability regulation of the climate.

Figure 94. The geospace environment engine. The flow of mass, momentum, and energy from the Sun’s interior through the interplanetary medium into the geospace environment. Both normal solar wind flows and transient events are indicated. Source: D.N. Baker. 2000. J. Atmos. Sol.-Terr. Phys. 62, 1669–1681.

The quasi-biennial oscillation (QBO) is a most remarkable atmospheric phenomenon and a major determinant, with ENSO, of seasonal and inter-annual weather variability. In the equatorial stratosphere, strong zonal winds circle the Earth. They originate at an altitude of 10 hPa (~ 35 km) and migrate downward at ~ 1 km/month until they dissipate at the base of the stratosphere at 80 hPa (~ 20 km). As the new zonal wind belt originates to replace the downward migrating previous one, it moves in an opposite direction, alternating easterly and westerly winds (Baldwin et al., 2001; figure 95). The QBO is usually defined at 30 hPa, where winds in one direction will start and increase in strength, and then decline and be replaced by winds moving in the opposite direction. The easterly and westerly phases of the QBO alternate every 22-34 months with an average of 28 months, but the periodicity is tuned to the yearly cycle, so the phase reversal occurs preferentially during the Northern Hemisphere late spring. The signature of the QBO in angular momentum, rather than having only a single spectral frequency peak at ~ 28 months, includes two additional spectral peaks at the annual frequency plus or minus the QBO frequency. In a breakthrough at the time, Lindzen and Holton (1968) proposed, and it was later demonstrated, that convection-originated vertically-propagated gravity waves provided the necessary wave forcing (momentum) for the QBO generation and maintenance (figure 95).

Figure 95. Dynamical overview of the QBO during northern winter. The propagation of various tropical waves is depicted by orange arrows, with the QBO driven by upward propagating gravity, inertia-gravity, Kelvin, and Rossby-gravity waves. The propagation of planetary-scale waves (purple arrows) is shown at middle to high latitudes. Black contours indicate the difference in zonal-mean zonal winds between easterly and westerly phases of the QBO, where the QBO phase is defined by the 40-hPa equatorial wind. Easterly anomalies are light blue, and westerly anomalies are pink. The mesospheric QBO (MQBO) is shown above ~80 km, while wind contours between ~50 and 80 km are dashed due to observational uncertainty. Source: M.P. Baldwin et al. 2001. Rev. Geophys. 39, 2, 179-229.

The QBO is a tropical phenomenon that affects the global stratosphere through the modulation of winds, temperatures, extra-tropical waves, meridional wind circulation, the transport of chemical constituents, and the distribution of ozone. One of the most puzzling aspects of the QBO is that it also modulates the Northern Hemisphere Polar Vortex, a persistent, large-scale, mid-troposphere to stratosphere, low pressure winter zone that when strong contains a large mass of very cold, dense Arctic air, and when weak and disorganized allows masses of cold Arctic air to push equatorward, causing sudden temperature drops in ample regions of the Northern Hemisphere.

In a series of seminal articles Karin Labitzke with Harry van Loon (1987; 2006) established that the QBO modulates the effect of solar activity on the stratosphere and the Polar Vortex. With great insight Labitzke, who was aware of the state of the solar 11-year cycle through time, unlocked a problem that had occupied researchers for centuries when she decided to segregate the data on stratospheric polar temperatures according to QBO phase (Kerr, 1987; figure 96). The very low correlation when all the data is considered, becomes very high using the segregated data, and Labitzke became the first to identify a strong sunspot-weather correlation.

Figure 96. The effect of solar activity on winter North Pole stratospheric temperature. A) Dash-dotted line, January-February average temperature at 30 mbar over the North Pole. Solid line, the 10.7 cm solar flux, a measure of solar activity proportional to the number of sunspots. A correlation cannot be seen. B) Dash-dotted line, same as in A, but only for the years with QBO in west phase (squared in A). Solid line, same as in A. The correlation becomes obvious. East phase QBO values show anti-correlation (not shown). Source: R.A. Kerr. 1987. Science, 238, 479-480.

The data segregation, pioneered by Labitzke, has been used very successfully to establish the relationship between phenomena that present phases, like the QBO, solar variability, and ENSO, and their effects on the Polar Vortex, stratospheric sudden warming events, the PDO and the NAO. Northern Hemisphere winter weather forecasts rely on the QBO phase, solar activity level, and ENSO state, and this has been a very active area of research for the past 30 years. The difficulty is great because despite an improving understanding of the physical basis, weather and climate models have problems reproducing a realistic QBO. For example, only 4 of more than 30 models used for the last IPCC report (AR5) had any sort of QBO. However, reanalysis readily displays the statistical association between the QBO phase and solar activity with stratospheric temperature and geopotential height (figure 97).

Figure 97. The effect of QBO phase and solar activity on Northern Hemisphere winter stratospheric temperature and geopotential height. a) Composite December-January 30 mbar temperature anomaly (°C, 1981-2010 baseline) for seven QBO east years. The situation corresponds to a disorganized polar vortex with more frequent cold Arctic surface air incursions at lower latitudes. b) Same as in a, for five QBO west years. A well-organized polar vortex keeps Arctic air trapped underneath. c) Composite January-March 500 mbar geopotential height anomaly (m, 1981-2010 baseline) for eighteen solar minimum years. A high winter North Pole geopotential is associated to a negative phase of the Arctic Oscillation. d) December-February correlation index between solar index and geopotential height for the 1980-2014 period. High solar activity correlates with low geopotential height over the Arctic. Source: NCEP/NCAR Reanalysis.

The current understanding, supported by observations, reanalysis, and modeling, is that the energy and momentum for the generation and maintenance of the QBO, and the stratospheric effects of the solar cycle and ENSO are provided by different kinds of gravity waves that originate from convection and weather phenomena in the tropics and propagate vertically (figures 95 & 98). Although the stratospheric effects propagate globally and affect both polar annular modes, the geographic asymmetry with most land masses and mountain ranges in the NH, creates hemispheric asymmetry. Planetary-scale Rossby waves that originate at northern mid-latitudes propagate vertically and reach the stratosphere, and the state of the stratosphere determines what happens next with those waves.

The QBO, solar activity, and ENSO act as gate keepers by determining the conductivity of the stratosphere to planetary waves. Combinations of these three factors during the winter cause a constructive or destructive interference with the vertical planetary waves, and in the first case the waves are deflected poleward transmitting heat and momentum to the stratospheric North Arctic Mode and Polar Vortex where they break (figure 98). This selective interaction with planetary waves maintains and even enhances tropical stratospheric anomalies, due to changes in solar activity, as they migrate poleward in the NH and downward to the polar troposphere, during certain winters. On the surface they determine the state of the dominant mode of variability, the Arctic Oscillation, and extend their influence to the North Atlantic Oscillation.

Figure 98. Summary of proposed top-down solar variability effects on climate. Only the Northern Hemisphere is represented, with the left and right halves showing the differences between summer and winter. The effects of solar wind induced magnetic and/or electric coupling, and the effects of cosmic rays are still quite unknown and thus not considered. Energetic particle precipitation at the pole produces odd Nitrogen and Hydrogen species in the upper atmosphere, that are more efficiently transported downward by the winter stratospheric vortex, reducing polar ozone levels. UV solar irradiation, variable with the solar cycle, is responsible for the ozone layer and its temperature gradients. Different types of tropical waves (orange) originating from convection, are responsible for the creation and maintenance of the Quasi-Biennial Oscillation (QBO), that together with the Brewer-Dobson circulation is responsible for the poleward transport of ozone. The position of the Tropical Jet Stream is determined by the Hadley circulation, while the strength and position of the Jet Stream and the Polar Night Jet depend on the strength of the Polar Vortex. Depending on stratospheric conditions, planetary-scale Rossby waves (red) can be deflected during the winter, causing stratospheric warming and a weakening of the Polar Vortex. The Polar Vortex determines the winter state of the Arctic Oscillation (AO), which strongly influences the North Atlantic Oscillation (NAO). Solar activity level, through its effect on stratospheric conditions, influences Northern Hemisphere winter weather far more than its small change in irradiation suggests. The ITCZ, the Inter-Tropical Convergence Zone, is the climatic equator. ENSO, El Niño Southern Oscillation.

While the East and West phases of the QBO have an opposite influence in modulating the effect of solar activity (figure 99), it is the easterly phase of the QBO in combination with low solar activity that shows the larger departure from average conditions. In QBOe winters, during low solar activity, the polar geopotential height is higher, polar stratospheric temperatures are higher, sudden stratospheric warming events occur earlier, the Polar Vortex is more frequently weaker and disorganized, the polar Jet Stream forms meanders that extend into lower latitudes, there is a higher frequency of blocking days, and the AO and NAO tend to be in negative phase (figure 99). This results in winters that are colder in mid-high Northern Hemisphere latitudes. Meteorologists have learned that NH winters with QBOe and low solar activity (like 2017-18 winter) tend to be cold and with more snow, especially in non-La Niña years, unless a recent stratospheric-reaching volcanic eruption interferes and produces a warmer winter.

Figure 99. Conceptual model of potential drivers of winter Jet Stream and NAO variability. A red arrow indicates a strengthening of the target box, while a blue arrow indicates a weakening. NAO can be used as a surrogate for jet stream variability. Arrows are not proportional to strength or confidence attached to the potential forcing. The main drivers of winter jet stream and NAO variability are believed to be the QBO, solar activity, ENSO, and Autumn Eurasian snowpack. Tropical volcanic eruptions can strongly interfere with the other factors driving a stronger vortex and positive NAO. Source: R. Hall et al. 2015. Int. J. Climatol. 35, 8, 1697-1720.

Given the complexity of the solar signal transmission through this indirect pathway, that is both conditional and seasonal, we do not have yet a good quantitative understanding of the mechanism. A further complication comes from the inability of most models to include or realistically reproduce stratospheric phenomena. However, the qualitative knowledge is solidly grounded in observation, reanalysis, and modeling (Baldwind & Dunkerton, 2005; Gray et al., 2010). During solar grand minima, solar activity gets stuck in low mode, and the frequency of cold winters in the NH multiplies, although warm winters can still occur, especially during QBOw phases. A significant deviation towards AO/NAO negative conditions is accompanied by a general decrease in winter temperatures and an increase in snow precipitation, that result in glacier advances. These were the conditions observed during the Maunder period of the Little Ice Age that have been reproduced in models with a clear solar attribution (Shindell et al., 2001).

A complementary pathway for the solar signal has been described for the rest of the solar spectrum that reaches the ocean surface, warming it (Meehl et al., 2009; figure 100). It has been shown to act mainly on the Pacific Ocean, which has the largest oceanic tropical surface. In the relatively cloud-free areas of the subtropics this “bottom up” mechanism determines the amount of evaporation, and by increasing moisture transport, cloud cover and precipitation at the convergence zone at times of higher solar activity, it expands the Hadley circulation and increases trade winds. The strengthened atmospheric circulation and moisture creates a feedback that enhances warm humid air subsidence at the sub-tropics further reducing cloud cover and enhancing the effect. The tropical effect is transmitted to the stratosphere by gravity waves, and carried to mid-latitudes by the Brewer-Dobson circulation. The increase in trade winds associated with the expansion of the Hadley circulation produces a negative anomaly in Eastern Pacific sea surface temperatures a year before the solar peak, that transforms into a positive anomaly with a two-year lag to the solar peak (Meehl et al., 2009).

Figure 100. Summary of proposed bottom-up solar variability effects on climate. Only the Pacific during the northern-winter is represented. In less clouded subtropical areas, peak solar activity increases evaporation. The enhanced moisture is transported by trade winds to convergence zones increasing precipitation, and strengthening Walker (not shown) and Hadley circulations. Intensified trade winds increase equatorial ocean upwelling reducing equatorial SST and driving a cloud reduction, through enhanced subsidence, that acts as a positive feedback similar to La Niña conditions, allowing more solar radiation to reach the surface. The Hadley circulation expands poleward increasing the area of the tropics. After the solar peak the eastern equatorial surface transitions to higher SSTs a couple of years later. The effect is stronger during the northern-winter. ITCZ, Inter-Tropical Convergence Zone, the climatic equator.

As we have seen, there are several pathways by which the solar signal is transmitted to the atmosphere-ocean coupled system and amplified through several feedback mechanisms, affecting climate (Roy, 2013; figure 101). Although the solar changes are small, the amplification energy is provided by the climate system. The pathways are complex, involving some of the lesser known climate phenomena, and act in a phase, seasonal, and latitudinal dependent way, often with opposite results. This is why the solar variability signal, whose effect can be seen so clearly in paleoclimatic records, is so hard to see in real time. The existence of more than one pathway also multiplies the signal, and further enhancement is attained through the different lags that allow an accumulation of the effect.

Figure 101. Flow chart of the Sun’s influence on climate. The three major climate variabilities, solar, QBO and ENSO are shown with oval outlines, whereas, the major circulations, responsible for modulating the effect of major variabilities are shown by non-rectangular parallelograms. The climatic effects are shown in blue-line boxes, with the direction of change shown by + (for increase) or − (for decrease). Source: I. Roy. 2014. Int. J. Climatol. 34, 3, 655-677.

The 50-70-year oscillation and the Stadium Wave hypothesis

The existence of a multidecadal mode of climate variation was first detected by Folland et al. (1984) in global SST and night marine air temperature records, and later correlated to precipitation records in the Sahel (Folland et al., 1986). This 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).

In the following years the 50-70-year oscillation was observed in North Atlantic sea level pressure and winds (Kushnir, 1994), North Pacific and North American temperatures (Minobe, 1997), length of day and core angular momentum (Hide et al., 2000), fish populations (Mantua et al., 1997: Klyashtorin, 2001), Arctic temperatures and sea ice extent (Polyakov et al., 2004), ENSO events relative frequency (Verdon & Franks, 2006), and global mean sea level (Jevrejeva et al., 2008).

Most of these records display also a ~ 20-year periodicity that is apparent in Greenland δ18O ice core data (Chylek et al., 2012). Differences between this periodicity, that is most apparent at mid-latitudes subsurface temperatures, and the 50-70-year oscillation, most apparent at high latitudes deep water salinity levels (Frankcombe et al., 2010), preclude a direct harmonic relation between them. Additionally, proxies indicate the ~ 20-year periodicity was more intense during the LIA, while the 50-70-year oscillation appears more intensely in 20th century records, and might not have been present, at least with that interval, during the LIA (Gray et al., 2004).

It is generally believed that the 50-70-year oscillation originates from internal ocean-atmosphere variability, rather than being externally forced or random generated. There remains the possibility that multidecadal external solar and tidal forcings set in motion the transfer of heat between different oceanic basins, and the 50-70-year oscillation could be an emergent temporal resonance from the intrinsic delays in the oceanic and atmospheric heat transmission. In this regard, the Bjerknes hypothesis establishes that hemispheric anomalies in the transport of heat by the atmosphere and the oceans should be of equal magnitude and opposite sign (Bjerknes compensation). The Bjerknes compensation has not been measured in nature due to our inability to properly sample heat transport by the oceans, but interestingly the Bergen climate model, when reproducing the Bjerknes compensation under constant forcing, generates a 60-80-year periodicity reminiscent of the AMO (Outten & Esau, 2017).

Whatever its cause, a climate quasi-periodicity leads to a climate quasi-predictability. Not as good as deterministic predictability, but much better than chaotic unpredictability. Divine and Dick, in their 2006 study of the historical variability of sea ice edge position in the Nordic Seas, correctly identified the effect of the 50-70-year oscillation over any putative anthropogenic effect, and ended with the conclusion that “during decades to come, as the negative phase of the thermohaline circulation evolves, the retreat of ice cover may change to an expansion.” It must have taken courage to predict a sea ice expansion in 2006, when essentially everybody else was predicting a sea ice collapse, yet since 2007 Arctic sea ice has been showing a, still non-significant, modest growth in September extent.

In her 2012 PhD thesis, Marcia Wyatt developed a hypothesis on the dynamic transfer of a climate signal between the different ocean basins, Arctic sea ice, and the atmosphere by the 50-70-year oscillation, that she termed “the Stadium Wave.” The hypothesis neatly links all the different manifestations of the 50-70-year oscillation, accounting for their lags, and produces a complete set of “quasi-predictions” that should be good for as long as the oscillation maintains its periodicity (figure 102; Wyatt & Curry, 2014). Wyatt and Curry, 2014, is one of the few articles (with Divine & Dick, 2006) that correctly predicted the current pause in Arctic sea ice melting at a time when the data suggested the opposite: “this [sea ice decline] trend should reverse… Rebound in West Ice Extent, followed by Arctic Seas of Siberia should occur after the estimated 2006 minimum of West Ice Extent and maximum of AMO” (Wyatt & Curry, 2014).

Figure 102. The 50-70-year oscillation and the Stadium Wave hypothesis. a) Simplified Stadium-Wave Wheel cartoon showing a 60-year cycle from 1976 to 2036. Red color indicates the high warming phase, and blue color the low warming/cooling phase. AMO, Atlantic Multidecadal Oscillation. AO, Arctic Oscillation. NAO, North Atlantic Oscillation. PDO, Pacific Decadal Oscillation. LOD, Length of Day. NHT, Northern Hemisphere Temperature. Modified from: M.G. Wyatt & J.A. Curry. 2014. Clim. Dyn. 42, 9-10, 2763-2782. b) Length of Day in milliseconds, daily data (grey) and long-term average smoothed (black). Source: IERS EOP. c) Atlantic Multidecadal Oscillation in °C detrended, monthly data unsmoothed (grey) and long-term average smoothed (black). Source: NOAA. d) Northern Hemisphere Temperature in °C anomaly (1961-90 baseline), monthly data (grey) and long-term average smoothed (black). Source: Hadley Climate Research Unit. e) Arctic Sea Ice extent in million km2, September data (grey) and long term average smoothed (black). Source: 1962-1978, M.A. Cea Piron & J.A. Cano Pasalodos. 2016. Rev. Clim. 16. 1979-2017 NSIDC. f) Sea Level rate of change in mm/yr. Average of Church & White 2011, Ray & Douglas 2011, and Jevrejeva et al. 2014. Source: S. Dangendorf et al. 2017. PNAS. 114, 23, 5946-5951. Orange and blue bars are inflection points when a phase might have changed. A decrease in Sea Level rate is anticipated by the hypothesis.

The most important consequence of the 50-70-year oscillation is that ~ 30 year warming and cooling phases and their associated effects on pressure, winds, precipitation, sea ice, and sea levels, should be properly accounted for by any global climate theory and models. It is clearly not the case of the leading CO2 hypothesis of climate change and the models that support it, where the last cooling phase of the oscillation was assigned to anthropogenic aerosols, and the last warming phase to anthropogenic greenhouse gases, leading to a completely unexpected, albeit predictable, warming pause 30 years later.



1) Excessive variation in the speed of rotation of the Earth, as measured by the length of day (∆LOD), is affected by alterations in the angular momentum of the atmosphere due to changes in tropospheric and stratospheric winds that are driven among other things by ENSO and insolation variations.

2) ∆LOD can be considered a proxy for zonal wind strength that acts as a leading indicator for major climatic turns.

3) The solar cycle, Quasi-Biennial Oscillation, and ENSO, determine the status of the ozone layer in the winter Northern Hemisphere stratosphere and through it, the status of the Polar Vortex and winter weather.

4) The solar cycle also affects SST and Hadley circulation strength through a bottom-up mechanism mediated by evaporation/precipitation and wind-induced upwelling.

5) A 50-70-year oscillation has been observed in multiple climate-related phenomena that can explain a significant part of climate variability. The Stadium Wave hypothesis proposes that dynamic energy-transfer through different ocean-atmospheric-ice compartments can explain the timing of climatic changes.


I thank Andy May for reading the manuscript and improving its English.

References [Bibliography ]

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

via Climate Etc.

January 21, 2018 at 05:43PM

Leave a Reply

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

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

Google+ photo

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

Twitter picture

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

Facebook photo

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

Connecting to %s

%d bloggers like this: