Guest essay by Stan Robertson
In a recent post entitled “Changes in Total Solar Irradiance” (http://wattsupwiththat.com/2014/10/25/changes-in-total-solar-irradiance/ ), Willis Eschenbach showed a plot of the solar irradiance that impinges at the top of the earth’s atmosphere. I have borrowed that from his post and repeat it here for convenience:
Fig. 1 Variations of TSI
Willis asked a profound question about these results:
“If the tiny eleven-year changes in TSI of a quarter of a W/m2 cause an observable change in the temperature, then where is the effect of the ~ 22 W/m2 annual variation in the amount of sun hitting the earth? That annual change is a hundred times the size of the eleven-year TSI change. Where is the effect of that 22 W/m2 change?”
This is a great question, but it is really two questions. First, why don’t we see some significant annual cyclic variation of global mean temperature? This is a truly profound question! It ought to keep climate modelers awake all night, every night. Second, if 22 W/m2 variations peak to trough don’t produce noticeable temperature variations, why should the 0.25 W/m2 variations of TSI associated with solar cycles produce any measurable temperature variations?
Let’s take the first question first. TSI reaches a peak on January 3 when we are nearest the sun and drops to a minimum six months later. Now 22 W/m2 is comparable to the change of TSI at 60 degrees north or south latitudes between ice ages and interglacial times. On this basis, one might expect to see a fairly substantial annual cyclic variation in global mean temperature. I failed to recall any in the many plots of global temperature anomalies that I have seen, but thought perhaps that single years wouldn’t stand out clearly in long, noisy records. So I grabbed a quick ten year data plot that I happened to have on hand to see if it showed annual cycles. None were obvious, but just to be sure, I took another look at the (also-quickly-available) periodogram for sea surface temperatures that I had made for a previous WUWT article (http://wattsupwiththat.com/2014/07/26/solar-cycle-driven-ocean-temperature-variations.) Not only is there no significant temperature variation with a one year period, there IS a small amplitude oscillation (0.13 oC peak to trough, 2X amplitude) at the 11 year solar cycle period with oscillation peaks that are nicely in phase with the sunspot peaks.
Fig. 2. Amplitude Periodogram of sea surface temperature anomalies 1954 – 2014
One of the first suggested explanations for the lack of annual cycles that I recall was that the variations might occur too fast for the earth mean temperature to respond. Considering that temperatures of either the northern or southern hemispheres of earth respond dramatically on a seasonal time scale to changes of solar flux at the surface, this seemed unlikely to me. Nevertheless, I dusted off my old computer program for calculating ocean surface temperature changes for changes of impinging solar flux. Previous calculation results have been reported here: (http://wattsupwiththat.com/2013/10/10/the-sun-does-it-now-go-figure-out-how) and here: http://wattsupwiththat.com/2014/07/26/solar-cycle-driven-ocean-temperature-variations
In the first of these, I found that a thermal diffusivity of 1 cm2/s for upper ocean waters was needed to account for the ocean surface temperatures (HadSST3gl) and ocean heat content measurements since 1965. If there were no changes of cloud cover or evaporation, 70% of that 22 W/m2 or 15 W/m2 would enter the atmosphere. If it impinged on oceans, it would drive annual temperature variation of 0.45 oC peak to trough. The temperature oscillations would, indeed, be larger if the solar flux variations occurred over a longer time. With a ten year period, they would produce temperature oscillations of 2.25 oC. In either case, most of the variations of the peak heat flux would be taken into the oceans and eventually returned later. Nevertheless, annual oscillations of 0.45 oC ought to stick out like a sore thumb in Fig. 2. So why don’t they occur? The only plausible explanation is that increases of cloud cover prevent most of that 22 W/m2 variation from ever reaching the surface. If absorbed by atmosphere, land or ocean, large temperature changes would necessarily follow. The minimum temperature increases would be those of the oceans, due to their transparency and large heat capacity. But they don’t show!
We can make this a little more quantitative to show that there is reason to believe that most of the TSI variations are negated by changes of cloud cover. The variations of cloud cover should correspond to variations of the atmospheric water column, as shown here in this plot from http://www.climate4you.com .
Fig 3. Atmospheric water vapor column (thickness if subjected to 1 atm pressure)
The total water column varies annually by about 0.45 cm peak to trough, for about 19% annual variation. Taken as a sinusoidal oscillation, its peak to trough variation would be 0.45 cm and its rate of change would have a peak to trough variation of (2 π 0.45 cm/yr). This rate of change would need to be provided by the solar flux that evaporates water at the earth’s surface. It takes about 2260 joules per gram to evaporate water. Then neglecting the minor amount of energy needed to lift the water vapor up into the atmosphere, the peak to trough rate of energy change needed for evaporation at the earth surface would then be:
(2 π 0.45 cm/yr) x (1 gm/cm3 x ( 2260 j / gm) x (1 yr / (365 x 86400 s)) x 104 cm2/m2 = 2 W/m2
This shows that very little of the available TSI variation is needed to produce the annual changes of atmospheric water column and, presumably, the variation of cloud cover. But if earth albedo changes in proportion to the variation of the atmospheric water column, then reflected solar radiation would vary by 19% of the mean 101 W/m2, or 19 W/m2. That would leave only about 3 W/m2 of the 22 W/m2 of TSI variation available to heat the earth surface. Since about 2 W/m2 is needed to produce evaporation, that leaves only about 1 W/m2 to be absorbed and warm the surface. Using the same computer program that I mentioned previously, I calculated that 1.0 W/m2 annual variations at the ocean surfaces would produce surface temperature oscillations of about 0.037 oC peak to trough. This is too little to be reliably extracted from noisy sea surface temperature data, but this is about what is shown in Figure 2.
A careful examination of Fig. 3 shows that the water column peaks seem to occur about late October rather than Jan. 3. The early peak is thought to be due to the end of the vegetation growth season in the northern hemisphere. The larger land mass of the northern hemisphere allows it to contribute more to evaporation during its growth season than does the southern hemisphere. This puts the annual TSI variation and cloud cover variation slightly out of phase but that really doesn’t matter much as long as there is enough extra cloud in January to negate the peak TSI. Another point worth noting about Fig. 3 is the step change downward after the 1998 El Nino. Prior to that, the water column was increasing, presumably because of surface warming and increasing evaporation. The smaller water column since 1998 is consistent with some cooling and the flat global temperatures of this century.
The most significant result of the preceding analysis is that it is clear that evaporation of water vapor into the atmosphere and cloud formation must provide a very strong negative feedback to radiative forcing in the UV/Vis bands that deliver most of the solar energy to earth. Starting from the present near-equilibrium conditions, a decrease of albedo would let more solar radiation reach the surface of the earth. That should be able to evaporate more water, produce more clouds and raise the albedo. If the albedo were to increase a bit beyond equilibrium, the surface would receive less insolation, the upper oceans would cool and cloud cover would decrease until balance was restored. Considering that downwelling infrared radiation is absorbed essentially at the ocean surfaces, the only thing that it can do is produce evaporation. We have just seen that a radiative forcing of 22 W/m2 apparently produces only a few hundredths of a degree of ocean surface temperature change. It seems a bit absurd to think that the 3.7 W/m2 of IR forcing that is expected to accompany a doubling of the atmospheric concentration of CO2 might do more. CO2 is simply not the control knob for the earth’s temperature.
Since cloud cover is so exquisitely regulated that it maintains a steady mean temperature, it would appear to be necessary for climate models to handle clouds well. In fact, however, that is one of their weaknesses. In general, the models used by the IPCC do a miserable job of modeling rainfall. It is highly likely that they are doing an equally poor job of cloud cover and albedo. Until this situation is dramatically improved, the climate models will remain essentially useless for anything but scare tactics.
Moving on to Willis’ implied question: If 22 W/m2 produces no significant temperature variations, why should the 0.25 W/m2 associated with the approximately 11 year solar cycles have the larger effect shown in Fig. 2? Only about half of this small amount would even reach ground level anyway. So how is it that we see 11 year solar cycle period temperature variability in the 60 year sea surface temperature record of Fig. 2? There are several possible explanations here. Some folks claim that the solar cycle temperature oscillations are spurious, but that seems unlikely to me for several reasons. First, the temperature peaks match the sunspot peaks. Second, I showed that Willis’ slow Fourier transform technique is quite capable of pulling this small signal out of the noisy data. Additionally, Roy Spencer, Nir Shaviv and others have found temperature variations of similar magnitude using different methods and data sets. Leif Svalgaard thinks that ~ 0.1oC temperature variations are real; however, he mistakenly persists in thinking that TSI variations of order 0.1 W/m2 at the earth surface can cause such temperature changes in several tens of meters of upper oceans. (Bear in mind that the first 25 meters of ocean has about 10X the heat capacity of the entire atmosphere.) Others claim that the temperature variations are spurious due to significant volcanic eruptions having occurred with approximate solar cycle timing. I think this to be very unlikely on a 60 year data set.
So, let’s take the question and the result of Fig. 2 seriously for a moment. The TSI variations associated with the solar cycle are only about 0.25 W/m 2, averaged over the earth surface and daily cycles. About 70%, or 0.175 W/m2 enters the troposphere. About (160/340)x0.25W/m2 = 0.117 W/m2 reaches the surface at wavelengths below 2 micron. About half the difference between the 0.175 and 0.117 W/m2 reaches the surface at longer wavelengths and after scattering in the atmosphere. This gives a peak to trough variation of about 0.15 W/m2 that would reach the surface. This is only about 15% of the 1.0 W/m2 that would be needed to drive surface temperature oscillations of 0.13 oC. So without even considering the possibility that changes of albedo might prevent most of the solar flux variation from even reaching the earth, it is apparent that TSI variations associated with the solar cycle do not provide enough energy to produce the temperature oscillations shown in Fig. 2.
To make it even more certain that the TSI variations are not the direct cause of the surface temperature oscillations, recall that albedo variations of about 19% were sufficient to negate the 22 W/m2 annual TSI variation and that this required only about 2 W/m2 to evaporate the water. One would therefore expect that about one could negate 0.25 W/m2 variations with about (0.25/22)x2 = 0.023W/m2. This is only about 15% of the 0.15 W/m2, 11 year, TSI variation that would occur at ground level if there were no albedo change. So even though the TSI variations would be too small to produce the observed surface temperature changes, they should easily evaporate enough water for a nullifying negative feedback. So the tiny variations of TSI associated with the solar cycle should be just as effectively negated as the 22 W/m2 of the annual cycle. This leaves a very stark question: If the temperature oscillations of Fig. 2 at the 11 year period are real and if they are produced by the sun, then how could the sun do it?
To answer this we need to consider another point made by Willis Eschenbach here: http://wattsupwiththat.com/2013/12/28/the-thermostatic-throttle/ . He showed that the evaporative feedback that regulates Earth’s albedo and temperature functions most strongly near the equator. Oceans areas near the poles show the reverse behavior. Tropical albedo changes cool the tropics, but near the poles the albedo decreases with increasing temperatures. This has the effect of making the equatorial zone cooler than it would be otherwise, while making the poles warmer. There seems to be less of either positive or negative feedback in mid-latitudes. This is what allows volcanic eruptions and other atmospheric disturbances outside the equatorial regions to affect surface temperatures. If the sun contributes something other than the dinky TSI changes over solar cycles, and outside the equatorial zone, then it might be able to produce the oscillations shown in Fig. 2.
It is well known that large volcanic eruptions can cool the earth. Volcanic ash shades the earth and produces short term cooling, but the most significant and longer lasting effects occur due to aerosols. The USGS (http://volcanoes.usgs.gov/hazards/gas/climate.php) says: The most significant climate impacts from volcanic injections into the stratosphere come from the conversion of sulfur dioxide to sulfuric acid, which condenses rapidly in the stratosphere to form fine sulfate aerosols. [Cloud droplets grown on] the aerosols increase the reflection of radiation from the Sun back into space, cooling the Earth’s lower atmosphere or troposphere. Several eruptions during the past century have caused a decline in the average temperature at the Earth’s surface of up to half a degree (Fahrenheit scale) for periods of one to three years. The climactic eruption of Mount Pinatubo on June 15, 1991, was one of the largest eruptions of the twentieth century and injected a 20-million ton (metric scale) sulfur dioxide cloud into the stratosphere at an altitude of more than 20 miles. The Pinatubo cloud was the largest sulfur dioxide cloud ever observed in the stratosphere since the beginning of such observations by satellites in 1978. It caused what is believed to be the largest aerosol disturbance of the stratosphere in the twentieth century, though probably smaller than the disturbances from eruptions of Krakatau in 1883 and Tambora in 1815. Consequently, it was a standout in its climate impact and cooled the Earth’s surface for three years following the eruption, by as much as 1.3 degrees at the height of the impact. Sulfur dioxide from the large 1783-1784 Laki fissure eruption in Iceland caused regional cooling of Europe and North America by similar amounts for similar periods of time.
These comments show that naturally occurring variations of aerosols are capable of producing surface insolation changes that are NOT entirely killed by negative feedback.
As long-time WUWT readers are aware, the Danish researcher, Henrik Svensmark, in 1996 proposed that cosmic rays that enter the atmosphere can produce aerosol condensation nuclei. The flux of cosmic rays is modulated by the strength of the sun’s magnetic field that reaches the earth and this varies with the nominal 11 year solar cycle. Fewer cosmic rays reach earth at the solar cycle peaks than at minima. This has been confirmed by direct measurements of cosmic ray flux over several solar cycles. Recent studies also seem to confirm that condensation nuclei can be produced by cosmic rays. See, e.g., http://www.youtube.com/watch?v=sDo7saKaEys .
What remains to be seen is whether the amounts of cosmic ray produced condensation nuclei and their variations are capable of significantly modulating the amount and reflectivity of cloud cover. This should be settled by measurements within the next decade or two. It would take very little change of cloud cover to produce the 0.13 oC peak to trough temperature oscillations at the 11 year period shown in Fig. 2. In the WUWT article in which I first used Fig. 2, I showed that it would take peak to trough variation of solar flux of about 1 W/m2, averaged over the sea surfaces to produce this temperature oscillation. This solar magnetic field effect would presumably occur over all latitudes from poles to equator. It would need to produce an average of about 1% change of cloud reflectivity, which presently reflects about 100 W/m2 of the average TSI at the earth.
Conclusions: The feedback that negates the effect of 22 W/m2 should be of huge concern to climate modelers. The amounts, types, both vertical and horizontal distributions and albedo of clouds need to be accurately modeled in order to determine the patterns of surface temperature on the earth. In these regards, I think that the present models used by the IPCC are inadequate, misleading and lacking in any ability to predict global mean temperatures for the future.
Whatever one might think to be the cause of the temperature oscillations shown in Fig. 2 at the nominal 11 year solar cycle period, it should be very clear that the TSI variations over a solar cycle are completely incapable of producing them. If the sun really is responsible for producing those small temperature changes, then Svensmark’s cosmic ray modulation theory would seem to be our best hope for understanding how it does it. Think of the cosmic ray modulation as a small amount of jiggling of the earth’s cloud thermostat. About one percent modulation of cloud albedo over a nominal 11 year solar cycle is all that is required.
Or maybe I should just say:
I’ve looked at clouds from both sides now
From up and down and still somehow
It’s cloud’s illusions I recall
I really don’t know clouds at all
Biographical Note: Stan Robertson, Ph.D, P.E., is a physicist, retired from Southwestern Oklahoma State University.