Guest Post by Willis Eschenbach
When folks tell me that the sunspot-related changes in total solar irradiance make changes down here on the surface of our amazing planet, I suggest that they take a look at the numbers.
From peak to trough over the sunspot cycle, the top-of-atmosphere total solar irradiance varies by about 1.2 watts per square metre (W/m2) … which, when averaged over the entire surface of the earth gives a change of about 0.28 W/m2. This is less than a tenth of one percent variation in total incoming energy.
But in fact, it’s less than that. Only about half of the sunlight makes it to the surface, so we’re down to 0.14 W/m2 change from peak to trough, less than a twentieth of a percent.
Now, downwelling radiation at the surface of the earth averages about 500 W/m2 on a 24/7 average basis. And out of that, we’re supposed to believe that a variation on the order of a tenth of a W/m2 is going to make a difference …
“Ah, you don’t understand”, folks inform me, “Yes, TSI only changes by a tenth of a percent over a solar cycle. But extreme ultraviolet (EUV) varies as a percentage much, much more than that!” … and you know what?
They’re right …
… but they’re also wrong. Let me explain why.
To start with, here’s the breakdown of the strength of the solar radiation by wavelength.
Figure 1. Spectrally resolved top-of-atmosphere sunshine. X-axis units are nanometres (nm).
In the middle is the visible spectrum, from about 380 nm to about 750 nm. Longer wavelengths than that are called “near infrared”. Wavelengths shorter than that are ultraviolet (UV).
And way over on the left, at 10 – 24 nm wavelength between the vertical red lines, is the tiny amount of extreme ultraviolet (EUV).
So that is the first problem. Even though it varies on a percentage basis more than the TSI, the EUV represents such a small part of the sun’s energy that it cannot be even seen at this scale.
The second problem is that the variation in EUV is much, much smaller in absolute terms than the variation in TSI. In Figure 2 I’ve compared the variations in the middle of the EUV spectrum (18 nm) to the variations in the blue part of the visible spectrum (~500 nm).
Figure 2. Monthly variations in solar output, measured in the EUV (red/yellow line) and in the visible spectrum (blue line)
As you can see, the variations in the EUV are very small compared to the variations in the visible spectrum.
In fact, the only reason that the percentage variations in EUV are greater than the percentage variations in TSI is that changes in EUV start from almost zero … so even a tiny absolute change in EUV is a large percentage change in EUV.
For those reasons, I hold that looking at EUV to explain surface climate variations is a blind alley … but as always, YMMV …
Best to everyone on a warm and quiet night,
w.
AS ALWAYS: I politely ask that you QUOTE THE EXACT WORDS THAT YOU ARE DISCUSSING so that we can all understand the subject of your thoughts.
DATA: I’ve used the solar data recommended here for use in the CMIP6. Yes, I know it has manifold problems, I pointed some of them out here on WUWT, but none of them affect these results.
Semi off-topic but is connected to previous comments regarding ionizing radiation and “we don’t know what we don’t know.”
I remember when SC24 was not progressing as predicted, particularly as compared to Dr. David Hathaway’s forecasts.
However, what I appreciated about him, is when the sun was blank. And blank for a long time; He came out with this quote:
“These findings have thrown our best computer models of the sun into disarray. “It is certainly challenging our theories,” says Hathaway, “but that’s kinda nice.””
To me, that quote captures the spirit of a true scientist.
SC24 humbled a lot of solar physicists. I suspect SC25 will too.
I stumbled across a phrase the other day watching a TedX talk on YouTube.
The term was “intellectual phase-locking.” Essentially that term explains the consensus on not just climate science, but on solar physics too. No one who wants to be seen as “credible” wants to buck the consensus. Thus everyone phase-locks their view of their field to the consensus.
Thus it often takes a young patent office clerk (or similar without vested baggage) to come along and disrupt the consensus.
Joel,
Nature has a way of doing that.
Mother Nature apparently agrees with Feynman that science is the belief in the ignorance of experts. And she delights in showing up those who presume to be experts in aspects of her doings.
And, Feynman famously described Joel’s phenomenon relating to Millikan’s oil drop experiment.
Inspired by an earlier comment:
Thank you for explaining in an analog the CO2 silliness.
As you said, EUV may change by a large percentage, but in absolute values it is really quite small.
For CO2, percentage wise, it has increased by 43% since the pre-industrial age. 280ppm to 400ppm.
But for actual change, it has only increased by .000120
Interestingly, in the combustion of fossil fuels, an approximate equal amount of water vapor is liberated. Couple that with changes of water usage in irrigation, home lawns and parks in places that grass didn’t exist, water is a heat exchange medium for power plants. Heck, where we live construction areas are required to spray the dirt so that it doesn’t kick up dust. In the desert. It seems to me that the amount of water vapor that humans add to the atmosphere dwarfs the amount of carbon added.
But carbon is much easier to regulate than water.
Willis,
Sorry, I did not see your small note about data source buried in your usual comments about stating differences by quotation of your original words.
The actual monthly UV flux was not measured for most of the time shown in your graph. It was reconstructed from proxies.
One proxy is the sunspot number. It is full of calibration difficulties back to your 1850 date, as you know. See for example http://www.leif.org/research/Defense-of-GN-Revisions.pdf
There are at least 2 sources of error to question here.
1. Is a historic sunspot number known to vary in harmony with the extreme UV flux? Unproven, I think. Probably not now possible to prove now. Problems with slow drifts.
2. Is the variation in sunspot number counted by different authorities large enough to significantly affect extreme UV flux interpretations? Probably yes.
Most worrying is the general treatment of the extreme UV intensity by the IPCC related authorities, in preparation for CMIP6, the periodic comparison of climate models, given in your reference here:
http://solarisheppa.geomar.de/solarisheppa/cmip6
There are two broad ways to consider this problem.
The first is to accept that the best and brightest researchers have brought their skills together and accepted that many assumptions are valid for future use, as here in calculating climate forcings. These assumptions include acceptance of the accuracy of past sunspot counts, the validity of corrections to them, the relation of the count to other factors such as solar irradiance, neutron counts, Be-10 isotope abundance etc. Because temperature is a primary parameter, the associations of these with temperature as appropriate has been studies and understood well enough to allow researchers to proceed with the IPCC based assumptions. The error terms at each stage for each parameter are well understood, reported and carried through with formal propagation mathematics where required; and they are deemed acceptable.
The second is to accept that in the absence of direct measurement of solar extreme UV (or indeed any solar UV) for many decades before say 1985, it is pissing in the wind to pretend that a reconstruction of any significance can be made of daily or monthly UV fluxes for model inputs, but hey! This is climate science.
When I started in measurement science with CSIRO in the 1970s, I would be sacked if I proposed to run models of global importance on input data that infilled decades at a time, in the absence of direct measurement. My standards have not changed. So, I feel confident in calling bulls*** on the modelling input data here. It is simply preposterous to claim that one can make reconstructions to the resolution given with the IPCC input data for CMIP6.
Willis, you are aware of snake oil data, so I wonder why you presented it without more visible caveats. I agree with your finding that total extreme UV variation is very small, plausibly too small to matter for global climate changes, but I disagree with part of the part you took to get there. Geoff.
Is a historic sunspot number known to vary in harmony with the extreme UV flux? Unproven, I think. Probably not now possible to prove now. Problems with slow drifts.
EUV creates and maintain the ionosphere at the altitude of 105 km. Solar heating and moves the air and the ions [and electrons] across the Earth’s magnetic field generating by dynamo action an electric current the magnetic effect of which we can detect at the surface as a diurnal variation of the geomagnetic field. This variation was discovered back in AD 1722 and we have kept track of ever since. The physics of this direct effect is well-known and is not controversial. This means that we have a reliable record of EUV stretching back that far.
http://www.leif.org/research/Reconstruction-of-Solar-EUV-Flux-1740-2015.pdf
It happens that this record validates the sunspot [group] number record, e.g. Slide 41 of
http://www.leif.org/research/Defense-of-GN-Revisions.pdf
So, we do know the run of EUV for more than 200 years.
There is much more that we do not know then know.
Yep, some people don’t know much…
I should have said we know it all.
What is the consensus (if any) about the effects of cosmic rays on ozone destruction (or production?) at different layers of he atmoshere? If there is indeed an effect, then variation of cosmic rays correlating with changes in solar magnetic fields would provide a more plausible cause than changes in UV radiation, since the correlation between solar variability and changes in cosmic rays is not disputed I think.
Currently Cosmic rays are up 18% since Dec 2014.
See
http://spaceweather.com/images2018/30jul18/stratosphere_california_strip.png
Willis I always like your articles because you provide data and often take a new approach to analysis. That said I think you need to write down and look at your initial assumptions and ask if there is evidence that contradicts those assumptions.
In this case you make the assumption that 1 watt of visible light radiation is equivalent to 1 watt of extreme UV. Since you show the total energy is very constant changes in solar radiation should have negligible effect. Except we know low orbit satellites are dragged out of orbit early when the upper atmosphere expands when the sun is active and there is more solar wind and extreme UV. This occurs even though the total energy changes very little. Your assumption fails. There are dramatic effects that can hit you in the head (literally in this case). Others have commented on UV- ozone and other effects that are not controlled by the IR through visible spectrum. Ozone is generated by UV but has significant IR absorption so we have a significant change in IR absorption in a total high energy portion of the spectrum controlled by changes in a very low total energy portion of the solar spectrum and it change a lot.
In the past you have used Fourier analysis on weather patterns. The people who write the math and statistics books usually are weak at emphasizing the assumptions or requirements for a particular technique. Fourier analysis is for stationary linear systems. Its great for looking at the performance of your radio circuit but that fails when your son is wildly twisting the tuning knob in the circuit. Weather is chaotic, not linear. The changing time lags of the yearly cycle means its not stationary. Fourier analysis will not work. It may kind of work for the fundamental frequency but the higher order frequencies will be smeared out into random noise.
Gary Palmgren
Thanks, Gary.
Since I’ve never denied that solar changes have an effect on the upper atmosphere, I fear you are mistaken about my assumptions …
Same objection. I know changes in the sun cause changes in the upper atmosphere and have said so many times. What I’ve never seen are any observations showing any such changes down at the surface.
Nonsense. The sunspot cycles are about as far from a “stationary linear system” as you can get, but Fourier analysis works just fine with sunspots.
I have no idea what you are calling the “changing time lags of the yearly cycle” … but whatever they are, they don’t obliterate the yearly cycle or make it unanalyzable.
Ah, yes, now you’ve moved on to science by assertion, my favorite kind. Please give us an example where Fourier analysis does NOT work on a weather-related dataset.
And in any case, I’m only looking for the fundamental frequency of ~ 11 years, not “higher order frequencies”.
Regards,
w.
UV radiation is resposible for more or less heating of the ozone layer, more UV, more heat and the polar jet stream is rotating well, less UV, less heat and the jet stream is meandering and produces blocking weather situations.
that’s what Joanna Haig found out, the results may be cold winters or hot summers
[url=https://www.nature.com/news/2010/100414/full/news.2010.184.html]One source[/url]
An other [url=http://www.landscheidt.info/?q=node/128]source[/url]
Link correction:
Source
Second link correction:
[url=https://www.nature.com/news/2010/100414/full/news.2010.184.html]One source[/url]
Sorry !
correction of the “correction”
One source
Willis’s head post is correct but irrelevant.
This is how it works via solar changes affecting global cloudiness:
http://joannenova.com.au/2015/01/is-the-sun-driving-ozone-and-changing-the-climate/
Looking at the spectral chart, something looks off. It shows 500nm as being ‘blue’, but it is not. That’s well within the ‘green’ part of the spectrum. Or is there a spectral shift when the spectrum is measured in vacuum as compared to at the bottom of the atmosphere?
That graph is a little off, but not by a huge amount. The boundary between baby blue and green lies around 500 nm. Green shades into yellow about 600 nm.
https://upload.wikimedia.org/wikipedia/commons/2/25/Electromagnetic-Spectrum.svg
TSI is a minor player but in the same direction as the dramatic influence of clouds on average global temperature. DIY on that; perhaps use this as a guide http://lowaltitudeclouds.blogspot.com. The time-integral of SSN is a good proxy for the influence of clouds and/or TSI on average global temperature.
Leif,
Re WUWT September 9, 2018 12:15 am
I study your writings with deserved respect, more closely than most writers, because you operate in a way similar to mine. Having studied a topic and assessed it, you then defend or reject it with the firm hand that comes from proper study and experience.
It was with surprise that you defended the proxy method of calculating past UV irradiance after I had earlier written with deliberate provocation that, in part, “These assumptions include acceptance of the accuracy of past sunspot counts, the validity of corrections to them, the relation of the count to other factors such as solar irradiance, neutron counts, Be-10 isotope abundance etc. … The error terms at each stage for each parameter are well understood, reported and carried through with formal propagation mathematics where required; and they are deemed acceptable. The second is to accept that in the absence of direct measurement of solar extreme UV (or indeed any solar UV) for many decades before say 1985, it is pissing in the wind to pretend that a reconstruction of any significance can be made of daily or monthly UV fluxes for model inputs, but hey! This is climate science”
This concerns your response including “So, we do know the run of EUV for more than 200 years.”
I have not seen any publication yet that covers the topic with “The error terms at each stage for each parameter are well understood, reported and carried through with formal propagation mathematics where required; and they are deemed acceptable”.
But there are many papers that I miss. There are so many papers, so few with high quality science. Can you support your assertion that the run of EUV is known over the past 200 years, with a statement that the errors involved in those measurements are also known?
And can you do so against the challenge that I consider the relevant IPCC science as pissing in the wind? Please convince me that your science has stayed hard and has not strayed into the realm of belief.
Best regards Geoff.
Figure 1 of http://www.leif.org/research/Reconstruction-of-Solar-EUV-Flux-1740-2015.pdf
shows the processes involved. Each ‘box’ represents a physical process that is well-known in quantitative detail [and most are verified by rocket launches]. The main error source is the variation from station to station due to differences in underground conductivity. Figures 11-13 summarizes the results and their errors. The errors are tiny compared to the signal. Figure 15 shows the relation between F10.7 [a good proxy for EUV] and the diurnal range. Again, the error is negligible.
Notice that the sunspot numbers are not used to infer the EUV. In fact we do it the other way around. The geomagnetic data let us infer EUV; it so happens that the EUV [the actually measured diurnal variation] can be used to validate the sunspot [group] numbers: section 11 of the link I gave you.
Leif,
Thank you for your responses.
Sadly, they seem to confirm what I have suspected for years, that the climate industry has next to no idea of how to properly handle overall mathematical uncertainty in the use of proxies. Geoff.
That seems to be so, but in no way casts doubt on the proxies themselves. The proxies are well-determined within their known uncertainties. That they are misused in climate ‘research’ is a problem of that field, not of the proxies.
“but in no way casts doubt “
Leif Svalgaard
Typo fixed.
w.
Leif.
When you write “That seems to be so, but in no way casts doubt on the proxies themselves” I am reminded of the MBH tree ring proxies, as is you are saying of them they are reliable. We can, with accuracy, count the number of rings, measure their density or width or whatever, all with good accuracy. But that is not the problem. The problem is to calibrate them against some other parameter, like temperature or rainfall. That cannot be done with accuracy. There are too many known unknows and you have to make too many assumptions of indefinable accuracy. Same with reconstructions of solar irradiance. It is the step of calibrating against a useful other parameter where the severe errors can lurk. Geoff
The tree ring proxies ar likely to be uncertain. However, the solar TSI reconstruction, in my view is on firm ground, c.f. http://www.leif.org/research/EUV-F107-and-TSI-CDR-HAO.pdf
There is general agreement that variations in TSI are due to variations of the solar magnetic field. The latter is now firmly constrained, thus also constraining TSI. There are still people that cling to the old obsolete group sunspot number [and thus TSI] for several reasons [e.g. funding = don’t rock the boat].