What’s going on with the Sun?
In the last post in the climate research series we described David’s major finding that changes in total sunlight lead Earth’s temperature by one sunspot cycle. But what’s going on with the Sun — what is the mechanism? In this post David lays out four puzzling clues about solar influence on our global temperature, then puts forward a hypothesis. What force (or forces) are required to resolve all these odd points?
To recap: Both his Fourier analysis and many independent papers suggest there is a delay between total solar irradiation (TSI) and global temperature. David reasoned that the delay is a true delay, not just a smoothing effect while increased heat propagates around the planet. Because the timing is so tied to solar cycles, the trigger for the delay must start on the Sun, not on the Earth. This is not just a case of our oceans slowly absorbing the extra energy from the Sun — and there simply isn’t enough, in any case. Something quite different is going on. Something on the sun changes, in sync with the variation in sunlight, but the corresponding changes following about 11 years later and change the way the Earth responds to incoming energy. It modulates the Earth’s albedo, controlling the Earth’s temperature like a tap controls the flow of water through a pipe.
For the moment we’ll call this mysterious phenomenon Force X (think X-rays, or Planet X). Candidates include solar magnetic fluxes, solar wind changes, and shifts in the solar spectrum (during each solar cycle, the energy shifts from more UV to more infra red and back). Something going on in the sun changes things like clouds, aerosols or jet streams on planet Earth, and through these secondary changes the Sun apparently controls a lot of the variation in temperatures on Earth.
The clues — the four phenomena that emerged from Evans’ work:
1. The Notch. In the short run, extra sunlight does no warming. Over the last century about 10% of the warming on Earth is because the direct heating effect of the Sun shining a little more brightly. However, while there are small peaks and falls in sunlight as part of the eleven year sunspot cycle, mysteriously there are no corresponding rises and falls in global temperature. This is the “notch effect”, where the extra energy from the Sun doesn’t produce a spike of warmth (even a small one, which would be detectable easily with Fourier maths). So some effect (but what!?) is occurring at the same time as a peak in sunlight, which changes the way the Earth responds to the extra light.
2. Changes in albedo (presumably clouds) are more important than the direct heating effect of changes in sunlight. David looked at the data on albedo and the amount of sunlight arriving on Earth, and calculated that the effect of changes in externally driven albedo (EDA) on global temperature is at least two times greater than the direct heating effect of changes in TSI, and possibly much more. Clouds blanket 60% of the Earth, they are the gatekeepers of sunlight into our climate system. Yet no major model includes the influence of EDA.
3. The delay, of one sunspot cycle. There is a delay of one sunspot cycle (about 11 years, or one half of a full solar cycle of about 22 years) between a change in TSI and a change in temperature on Earth. Something other than the direct heating effect of TSI is driving that — there must also be an indirect effect of TSI, that is delayed. The delay originates on the Sun, not on Earth, because it is synced to the solar cycle. What is it about the Sun that changes one sunspot cycle after the sunlight (or TSI) peaks and falls?
4. In the long run extra sunlight does too much warming. For periods longer than 20 years, the Sun controls temperatures on Earth in a fairly predictable, linear way (this is the flat section of that Fig 2 in post 21). The low pass filter shows that over any long time period, and in every dataset, the “warming effect” is much the same. David calculated that the effect of a change in sunlight carries 14 times the influence of the actual change in energy itself. So bizarrely, the effect of the suns light is amplified somehow — which is consistent with albedo modulation (and really, it is hard to see what else it could be).
All these clues tell us that something much bigger and more important than mere fluctuations in the Joules of sunlight is controlling our climate. Solar TSI is an indicator, but not the agent. So figure that something coming off the Sun that isn’t sunlight is causing Earth to warm and cool, quite possibly through controlling clouds. But, under a single-agent hypothesis, whatever it is also stops Earth warming in the short period that sunlight peaks each cycle. Strangely the thing that seems to amplify the effect of sunlight is also delayed by a sunspot cycle, but in the long run a more active sun means a warmer Earth.
In this post David considers whether all four of the unexplained climate phenomenon can be wrapped into one hypothesis, one force, called “force X” for now. If there are moments when you read this single hypothesis and wonder if this is not an easy or ideal fit, rest assured, in the next post in this series he splits Force X into two separate processes, replacing the Force X hypothesis with the Force ND hypothesis. Two forces fit better, but are more complicated — Occam’s razor and all that. That’s up soon. — Jo
23. The Force-X Hypothesis
This post pieces together numerous clues about strong influences on surface warming into a working hypothesis about a single strong warming influence, called “force X”.
We have now identified four “strong” manifestations of one or more climate influences not included in conventional climate models, where “strong” means to have at least as much effect on surface temperature as the direct heating effect of changes in total solar irradiance (TSI):
- Externally driven albedo (EDA) is causing albedo modulation. Post 10.
- The notch implies a countervailing influence during TSI peaks. Post 21.
- The empirical transfer function implies an indirect solar sensitivity (ISS). Post 21.
- The delay implies an influence that lags TSI by ~11 years. Post 22.
The EDA and notch are new pieces of the puzzle; the ISS and delay have long been known but not necessarily connected. This theory is presumably original in combining all four.
It is well known that variations in direct heating by TSI are too small to explain global warming: the 11-year-smoothed TSI (at 1 AU) at rose ~0.7 W m−2 from 1900 to 2000, but that only caused ~0.08 [0.05, 0.25] °C (Eq. (1) of post 21) of the observed 0.8 °C of observed surface warming, or ~10%.
How many independent strong influences can there be? AR5, Table TS.6, lists only greenhouse gases, aerosols, and albedo changes due to land-use change as long-term strong influences, and volcanoes can be strong but are transitory; yet the four manifestations listed above are none of these.
The simplest explanation is that the four items above are not manifestations of four separate and previously unknown influences, but of one. Here we will be guided by Occam’s razor.
Let us assume there is only one influence, for the remainder of this post. We call it “force X” — because although the outline and some properties of the influence can be deduced at this stage, the exact mechanism is unknown. (There is an historical precedent for the “X”: Wilhelm Röntgen named x-rays thusly in 1895 because although he could demonstrate their presence and effects, he did not know what they were exactly. Hopefully when and if force X becomes completely known, it will be renamed.)
The notch shows that force X is synchronized to the Sun — if the delay was of constant duration then force X would get out of sync with the TSI peaks. Thus the delay is not a propagation delay for heat moving around on Earth. It is difficult to see how the relatively tiny changes in TSI during a TSI maximum could alter anything of significance on Earth, so it would appear that force X originates in the Sun, though it may act via agents on Earth. The EDA finding indicates force X acts via albedo modulation. The delay indicates that force X acts ~11 years after a corresponding change in smoothed TSI, implying that changes in TSI occur ~11 years before the corresponding changes in force X. The duration of the delay is suggestive of one sunspot cycle (a Schwabe cycle, ~11 years), or half of one full solar cycle (a Hale cycle, ~22 years) — a half-cycle delay in the Sun’s dynamo is perhaps the simplest and most natural lag in a rotating system. The ISS indicates that force X has an order of magnitude more influence on surface temperatures than the direct heating effect of TSI over the longer term.
We will assume without loss of generality that force X is a warming influence, rather than cooling. At the sunspot maximum each sunspot cycle the TSI peaks, whence force X must trough to counteract the effect of the peaking TSI in the surface temperature record. These TSI peaks occur just when the Sun’s magnetic field flips polarity and the solar polar field goes through zero (though many other aspects of the Sun’s magnetic field do not go through zero).
There exists an influence on the Earth’s mean surface temperature, called “force X”, such that:
- When force X increases, the surface starts warming immediately and becomes warm after a delay determined only by the thermal inertia of the Earth’s climate system (in the order of a year).
- Force X modulates the Earth’s albedo.
- Changes in force X occur half of a full solar cycle, or ~11 years on average, after corresponding changes in smoothed TSI (smoothed or averaged over at least an entire sunspot cycle).
- Force X is weaker when the Sun’s magnetic field is reversing its polarity.
- On decadal and centennial time scales, force X is such that the surface warming associated with a change in TSI is an order of magnitude greater than the direct heating effect of that change in TSI.
The observational evidence for a delay in post 22 is of a statistical nature, consistent with underlying changes in TSI found by smoothing TSI over a sunspot cycle or longer, sufficient to be independent of the TSI peak in the middle of the sunspot cycle. The delay is not sharply defined; TSI does not precisely foretell every small change in force X and surface temperatures that occur ~11 years later. Instead force X lags the underlying or smoothed TSI. Hence there is no contradiction in force X being at its weakest during a sunspot cycle just when TSI is peaking, which is of course about ~11 years after TSI last peaked. Force X appears to change on a decadal scale only in response to decadal changes in TSI. Also bear in mind that TSI is not force X, only an imperfect predictor of force X, and we have only the sunspot record and estimated sunspot-TSI relationships to work with.
The coincidence between peaks in TSI and the flipping of the solar magnetic field could explain the notching. As TSI peaks force X is in a trough — see Fig. 1 — and these countervailing influences cancel out sufficiently closely as to leave no trace in the surface temperature record. This begs the question of whether there might be some feedbacks or some other principle behind such a precise cancellation.
Figure 1: When TSI peaks, the solar magnetic field is at its weakest because it is reversing polarity. This figure merely illustrates the timing; the solar polar field is but one aspect of the Sun’s magnetic field. (It was Joanne Nova who first realized the significance of this coincidence of notching with the reversal of the Sun’s magnetic field.)
Many solar phenomena are related to the power delivered by the Sun’s electromagnetic field, and thus to the product of its electric field strength and magnetic flux — which correlate, so the power is roughly proportional to the square of the magnetic flux. The number of sunspots is such a phenomenon, so, because the square of the magnetic flux is indifferent to its polarity, the sunspots follow an apparent ~11 year cycle even though the full solar cycle is ~22 years. Some phenomena, such as hydrological cycles on Earth (which are mapped back thousands of years on the Nile delta), are sensitive to the full 22-year cycles rather than to the 11-year sunspot cycle.
The duration of the hypothesized delay is one sunspot cycle, or ~11 years on average. But, starting from the moment of a significant change in underlying or smoothed TSI, is the delay to the corresponding change in force X the duration of the current, previous, or next sunspot cycle, or maybe a weighted average of all of them?
An analogy may help to understand the delay.
A four stroke combustion engine has four phases: “suck, squeeze, bang, and blow”. If you know how much fuel and air is inhaled during the “suck” phase then you know how much power will be produced in the “bang” phase, which comes half a full cycle (two phases) later. Apparently something similar is happening with the Sun: the sunspots, or the tiny changes in TSI, tell us how much force X there will be half of a full solar cycle later.